Nitrogen-containing cyclic compounds as iminium ion sources for selected reaction monitoring detection of derivatized analytes

Article abstract of DOI:10.1177/1469066719869817

Liquid chromatography–tandem mass spectrometry is one of the most sensitive tools for determination of trace amounts of analytes in metabolomics and proteomics. The highest sensitivity is achieved in selected reaction monitoring detection, which involves fragmentation of the molecular ion between two levels of mass selection. However, fragmentation of some compounds is complicated. Detection sensitivity of such analytes may be increased by derivatizing them with a specific moiety fragmentation of which results in product ion of high abundance. In this work, we reveal the influence of iminium ions' structures on their stability by comparing six nitrogen-containing cyclic compounds as derivatization reagents for tandem mass spectrometric analysis of amino group-containing analyte. Commercially available starting materials (piperidine, 2,6-dimethylpiperidine, 1-methylpiperazine, morpholine, pyrrolidine and 1-cyanomethyl-3-methylimidazolium ionic liquid) were used for the synthesis of corresponding carboxylic acids which were further used for derivatization of the model analyte tryptamine. Liquid chromatographic–mass spectrometric analysis of differently derivatized tryptamine was performed for the evaluation of release and stability of corresponding iminium ions under collision-induced dissociation conditions. As a result, morpholine moiety was shown being the most promising iminium ion source among tested compounds. Possible sub-fragmentation pathways of investigated iminium ions were discussed, and the structures of secondary product ions were proposed.

Full text of DOI:10.1177/1469066719869817

Original research paper
European Journal of Mass Spectrometry
0(00) 1–11
! The Author(s) 2019
Article reuse guidelines:
Nitrogen-containing cyclic compounds as
iminium ion sources for selected reaction
monitoring detection of derivatized analytes
sagepub.com/journals-permissions
DOI: 10.1177/1469066719869817
journals.sagepub.com/home/ems
ˇ
Andrius Zilionis
Abstract
Liquid chromatography–tandem mass spectrometry is one of the most sensitive tools for determination of trace amounts of
analytes in metabolomics and proteomics. The highest sensitivity is achieved in selected reaction monitoring detection,
which involves fragmentation of the molecular ion between two levels of mass selection. However, fragmentation of some
compounds is complicated. Detection sensitivity of such analytes may be increased by derivatizing them with a specific
moiety fragmentation of which results in product ion of high abundance. In this work, we reveal the influence of iminium
ions’ structures on their stability by comparing six nitrogen-containing cyclic compounds as derivatization reagents for tan-
dem mass spectrometric analysis of amino group-containing analyte. Commercially available starting materials (piperidine,
2,6-dimethylpiperidine, 1-methylpiperazine, morpholine, pyrrolidine and 1-cyanomethyl-3-methylimidazolium ionic liquid)
were used for the synthesis of corresponding carboxylic acids which were further used for derivatization of the model
analyte tryptamine. Liquid chromatographic–mass spectrometric analysis of differently derivatized tryptamine was per-
formed for the evaluation of release and stability of corresponding iminium ions under collision-induced dissociation
conditions. As a result, morpholine moiety was shown being the most promising iminium ion source among tested
compounds. Possible sub-fragmentation pathways of investigated iminium ions were discussed, and the structures of
secondary product ions were proposed.
Keywords
Cyclic iminium ion, selected reaction monitoring, collision-induced dissociation, amino group derivatization, amide bond
fragmentation, EDC/NHS coupling
Received 17 May 2019; accepted 23 July 2019
acts as collision cell where fragmentation occurs over
the chromatographic elution time.1,4 However, frag-
Introduction
Liquid chromatography with selected reaction monitor- mentation of some compounds is complicated because
ing (SRM) detection is a method of choice for deter- of their specific structure or the desired dissociation
mination of trace amounts of analytes in complex pathway giving the product ion of interest is only
samples because of high sensitivity and selectivity. one of many active dissociation pathways and thus
SRM is a non-scanning mass spectrometric technique the abundance of product ion is limited.2 Fortunately,
performed on triple quadrupole mass spectrometers a specific derivatization of such analytes may be
and it shows a linear response over a wide dynamic employed in order to make fragmentation of the result-
range up to five orders of magnitude.1 Moreover, a ing molecular ions effective, thus increasing the sensi-
very high signal to noise ratio is achieved with SRM tivity of SRM detection.
detection compared to scanning mass spectrometric
measurements because the two levels of mass selection
with narrow mass windows result in a very effective
reduction in chemical noise.1,2 As a result, an increase
in sensitivity of one to two orders of magnitude is
achieved.3 While the first and the third quadrupoles
act as mass filters to select a molecular ion and a spe-
Faculty of Chemistry and Geosciences, Vilnius University, Vilnius, Lithuania
Corresponding author:
ˇ
Andrius Zilionis, Faculty of Chemistry and Geosciences, Vilnius University,
Naugarduko Street, 24, Vilnius 03225, Lithuania.
cific fragment ion of analyte, the second quadrupole Email: andrius.zilionis@chf.vu.lt

2
European Journal of Mass Spectrometry 0(00)
Ross et al. described a multiplexed protein quantita- Fisher Chemical); triethylamine (ꢀ99.5%, Sigma-
tion strategy based on a set of isobaric reagents.5 These Aldrich); bromoacetic acid (99%, Acros Organics); etha-
reagents form amide bonds to primary amino groups of nol (ꢀ99.5%, Carl Roth); acetonitrile (LC-MS grade, Carl
the peptides, and fragmentation of amide bond during Roth); N,N-dimethylformamide (>99%, Fluka); thi-
collision-induced dissociation (CID) results in the neu- onyl chloride (ꢀ98%, Carl Roth); N-hydroxysuccini-
tral carbonyl moiety loss, while the charge is retained mide (NHS, ꢀ98%, Acros Organics); tryptamine
by the reporter fragment based on 1-methylpipera- hydrochloride (99%, Sigma-Aldrich); 1-cyanomethyl-
zine.5,6 Later, 2,6-dimethylpiperidine moiety was used 3-methylimidazolium bis(trifluoromethylsulfonyl)imide
in tandem mass tagging strategy for reporter ion gen- (for synthesis, Merck); sodium peroxide (96%, Acros
eration after amide bond cleavage in the same way.7 Organics); hydrochloric acid (32%, Fisher Chemical);
Although these two strategies did not employ SRM sodium hydroxide (ꢀ98%, Sigma-Aldrich); 2-(4-mor-
detection, the intense peaks of cyclic product ions ori- pholino)-ethane sulfonic acid (MES, ꢀ98%, Fisher
ginated from derivatized peptides dominated the mass BioReagents); phosphoric acid (ꢀ85%, Acros
spectra, while a simple trimethylamine-based quater- Organics); 1-(3-dimethylaminopropyl)-3-ethylcarbodii-
nary ammonium ionization tag used by Che and mide hydrochloride (EDC, ꢀ98%, Acros Organics);
Fricker was reported undergoing significant neutral tris-(hydroxymethyl)-aminomethane (TRIS, ꢀ99.9%,
loss of trimethylamine from labeled peptides during Carl Roth); formic acid (FA, LC-MS grade, Sigma-
CID.8 It was also reported that Hofmann elimination Aldrich). All solutions were prepared by dissolving
is the reason for low stability of the linear trialkyl qua- required amounts of reagents in bi-distilled water
ternary ammonium salts under CID conditions.9 unless otherwise noted. MES buffer solution (0.1 mol/l,
Taking this into account, we expect that the charge pH 6) was prepared by dissolving 1.066 g of 2-(4-mor-
retention on basic nitrogen atom after neutral carbonyl pholino)-ethane sulfonic acid in &40 ml of water,
moiety loss could be an attractive pathway to produce adjusting pH to 6 with a saturated sodium hydroxide
product ions of high abundance for SRM detection.
solution and adding water up to 50 ml final volume.
In this work, we reveal the influence of iminium ions’ Sodium phosphate buffer solution (0.1 mol/l, pH 7.5)
structures on their stability by comparing six nitrogen- was prepared by dissolving 0.5765 g of phosphoric
containing cyclic compounds as derivatization reagents acid (85%) in &40 ml of water, adjusting pH to 7.5
for SRM detection of amino group-containing analyte. with a saturated sodium hydroxide solution and
Piperidine, 2,6-dimethylpiperidine, 1-methylpiperazine, adding water up to 50 ml final volume.
morpholine, pyrrolidine and 1-cyanomethyl-3-methyli-
Synthesis of carboxylic acids of piperidine, 2,6-
midazolium ionic liquid were chosen as starting mater- dimethylpiperidine, 1-methylpiperazine, morpholine,
ials because of basic nitrogen atoms in cyclic structures pyrrolidine, diethylamine and di-n-propylamine was per-
and commercial availability. Corresponding carboxylic formed as described elsewhere.10 Briefly, 0.00864 mol of
acids were synthesized and further used for derivatiza- each base was dissolved in 4 ml of tetrahydrofuran. In
tion of the model analyte tryptamine which was chosen the cases of 2,6-dimethylpiperidine, diethylamine and
because of UV absorbance, retention on a reversed di-n-propylamine, 20ml of triethylamine was added.
phase liquid chromatography column and the presence Bromoacetic acid (0.4 g, 0.00288 mol) was dissolved in
of primary amino group. Liquid chromatographic– 10ml of tetrahydrofuran and added dropwise to a stirred
mass spectrometric analysis of differently derivatized solution of a base. The reaction mixture was stirred at
tryptamine was performed in SRM mode at different room temperature for three days. The white solids were
collision energies in order to compare the release and filtered, washed with tetrahydrofuran, recrystallized
stability of different iminium ions during CID. Positive from hot ethanol and dried under a nitrogen atmos-
electrospray ionization (þESI) efficiencies of trypta- phere. In the cases of diethylamine and di-n-propyla-
mine derivatized with 1-methylpiperazine and morpho- mine, the white solids were collected after washing
line moieties were compared using selected ion them twice with tetrahydrofuran and dried under a
monitoring (SIM) mode, while product ion scan mode nitrogen atmosphere. In the case of pyrrolidine, the
at relatively high collision energies was used to reveal oily liquid was collected after washing it twice with tetra-
possible sub-fragmentation pathways of different imi- hydrofuran and dried under a nitrogen atmosphere.
nium ions.
1-carboxymethyl-3-methylimidazolium
bis(trifluoro-
methylsulfonyl)imide was synthesized by the mild
hydrolysis of a cyano group of 1-cyanomethyl-3-methy-
limidazolium bis(trifluoromethylsulfonyl)imide using
Experimental
Chemical reagents were used as received without an sodium peroxide.11 Briefly, 1-cyanomethyl-3-methylimi-
additional purification: piperidine (ꢀ99%, Carl Roth); dazolium bis(trifluoromethylsulfonyl)imide (0.003mol,
2,6-dimethylpiperidine (ꢀ97%, Acros Organics); 1.2069g) was mixed with water (10 ml) at 40ꢁC.
1-methylpiperazine (99%, Acros Organics); morpholine Sodium peroxide (0.0045 mol, 0.3509g) was added very
(ꢀ99%, Carl Roth); pyrrolidine (ꢀ99%, Acros slowly to a stirred mixture and the resulting solution was
Organics); diethylamine (ꢀ99%, Alfa Aesar); di-n-pro- heated at 40ꢁC for 2 h. After that, the solution was acid-
pylamine (99%, Alfa Aesar); tetrahydrofuran (ꢀ99.8%, ified to pH &1 with concentrated hydrochloric acid. The

ˇ
Zilionis
3
oily liquid (bottom layer) was collected after a few min- left at room temperature for 2 h. The reaction was
utes using a pipette, washed with water twice, suspended quenched by adding 200 ml of sodium phosphate
in 2 ml of water and neutralized with a saturated sodium buffer. The efficiencies of tryptamine derivatization
hydroxide solution. The resulting solution was used as 1- using active NHS esters of (4-methyl-1-piperazinyl)ace-
carboxymethyl-3-methylimidazolium source. The yields tic acid and 4-morpholinylacetic acid were 63.1% and
of the synthesis varied between 18 and 83% (the yield of 77.4%, respectively, and derivatization conditions were
1-carboxymethyl-3-methylimidazolium synthesis was not further optimized.
not estimated). The presence of the desired compounds
Liquid chromatographic–mass spectrometric analysis
in the final products was confirmed by direct infusion was performed on Agilent Technologies 1290 Infinity
mass spectrometry: a small amount of each synthesized liquid chromatography system coupled to Agilent
product was dissolved in the mixture of water and aceto- Technologies 6410 Triple Quad mass spectrometer. For
nitrile (50:50, v/v) and injected directly to the ESI source a comparison of release and stability of iminium ions
of Agilent Technologies 6410 Triple Quad mass spec- during CID, the products of tryptamine derivatization
trometer operating in þESI scan mode.
were separated on reversed phase Waters Acquity UPLC
Synthesis of active NHS esters of (4-methyl-1- BEH Phenyl column (1.7 mm, 2.1 ꢂ 100 mm). Solvent A
piperazinyl)acetic acid and 4-morpholinylacetic acid (0.1% FA in water) and solvent B (0.1% FA in aceto-
was performed as described elsewhere.10 Briefly, 0.7 g nitrile) were used for gradient elution (0min – 10% B,
(0.0096 mol) of dimethylformamide and 1.14 g 6 min – 50% B, 8 min – 80% B). The column was equili-
(0.0096 mol) of thionyl chloride were dissolved in brated with starting mobile phase for 6 min before
12 ml and 8 ml of tetrahydrofuran, respectively. an injection. For a comparison of þESI efficiencies
Thionyl chloride solution was cooled in an ice bath of differently derivatized tryptamine, reversed phase
and dimethylformamide solution was added dropwise. Waters Acquity UPLC BEH C18 column (1.7mm,
The reaction mixture was left in an ice bath for 30 min. 2.1 ꢂ 100 mm) was used with isocratic elution mode
After that, the ice bath was removed and 0.8 g (15% B). In both cases, mobile phase flow rate was
(0.0068 mol) of pre-powdered NHS was added to a stirred 0.25ml/min, column temperature was 25ꢁC and detec-
solution, followed by the addition of (4-methyl-1-pipera- tion wavelength was 280 nm. þESI parameters were as
zinyl)acetic acid (1.004 g, 0.0064 mol) or 4-morpholinyla- follows: nebulizer gas temperature 200ꢁC, gas flow rate
cetic acid (0.93 g, 0.0064 g). The reaction mixture was 8 l/min and capillary voltage 4000V.
stirred overnight at room temperature. The solids were
filtered and washed with tetrahydrofuran. The presence
Results and discussion
of active NHS esters was confirmed by the reactions of
synthesized products with tryptamine.
Derivatization of analytes prior quantitation by liquid
Derivatization of tryptamine was performed in two chromatography with SRM detection should be fast,
ways. For a comparison of release and stability of imi- the desired products should be obtained in high yields
nium ions during CID, EDC/NHS strategy was and without side-products, the derivatives should not
employed. Briefly, 0.5 mol/l solutions of investigated require additional purification and the added chemical
carboxylic acids were prepared by dissolving required derivative group should be released effectively during
amounts of previously synthesized carboxylic acids in CID. In this work, six cyclic carboxylic acids were used
MES buffer. In the case of 1-carboxymethyl-3-methyli- as derivatization reagents (see Figure 1(a)), and frag-
midazolium bis(trifluoromethylsulfonyl)imide, 50 ml of mentation of amide bond with the neutral loss of car-
previously prepared neutral solution of carboxylic acid bonyl moiety was employed as the product ion
was diluted twice with MES buffer. Hundred micro- generation pathway (see Figure 1(b)). Several different
liters of each carboxylic acid solution were mixed strategies could be utilized for amide bond formation.12
with 100 ml of freshly prepared 0.5 mol/l EDC solution, Despite the fact that EDC/NHS strategy could result in
and 10 ml of 0.25 mol/l NHS solution (prepared in various side-products,13 this strategy was chosen for a
acetonitrile) was added. Hundred microliters of comparison of release and stability of iminium ions
0.0001 mol/l tryptamine solution (prepared in sodium during CID because of simplicity: the carboxyl groups
phosphate buffer) were added and the reaction mixture of synthesized derivatization reagents were activated
was left at room temperature for 1 h. The reaction was with EDC/NHS followed by covalent attachment of
quenched by adding 100 ml of 2 mol/l TRIS solution. tryptamine by its primary amino group, thus forming
The efficiencies of tryptamine derivatization using amide bonds.14 Differently derivatized tryptamine
EDC/NHS strategy were not estimated. For a compari- was separated on a reversed phase chromatography
son of differently derivatized tryptamine’s ionization column, and mass spectrometric detection was applied
efficiencies, active NHS esters of (4-methyl-1-piperazi- in both SRM and SIM modes. During SRM, an amide
nyl)acetic acid and 4-morpholinylacetic acid were used. bond cleaves and a proton remains associated with a
Briefly, 30 ml of 0.0005 mol/l tryptamine solution (pre- basic nitrogen atom of derivatization reagent moiety,
pared in sodium phosphate buffer) was mixed with 70 ml thus generating the iminium ion (1-carboxymethyl-3-
of freshly prepared 0.0214 mol/l active NHS ester solu- methylimidazolium moiety is an exception). The release
tion (prepared in ethanol). The reaction mixture was and stability of these product ions during CID are of

4
European Journal of Mass Spectrometry 0(00)
Figure 1. The structures of synthesized carboxylic acids (a) and the scheme of tryptamine derivatization and fragmentation (b) (note that
methylimidazolium moiety is an exception, and CID of correspondingly derivatized tryptamine is not illustrated by scheme b).
critical importance for detection sensitivity and were abundance at optimal collision energy (see
investigated by increasing collision energies and mea- Supplemental material S2; optimal collision energies
suring the peak areas of corresponding product ions in were determined using SRM data from Figure 2).
SRM mode, while SIM mode was required for normal- Nevertheless, the complete release of the added chem-
ization of SRM data of all six different product ions ical derivative group without any sub-fragmentation of
originated from six different molecular ions. It is worth the resulting ion is necessary for the maximum yield of
noting that dilution and injection volume of each corresponding product ion. Such fragmentation is pos-
EDC/NHS reaction sample were adjusted so that the sible when the unwanted secondary fragmentation
areas of SIM peaks were similar (normalization coeffi- occurs at significantly higher collision energy compared
cients varied between 0.904 and 1.090). Taking this into to the formation of desired product ion. As a result, the
account, we assume that similar amounts of molecular resistance of iminium ions to sub-fragmentation is of
ions physically passed the third quadrupole during great importance in our experiments.
SIM measurements and that normalized SRM data
Normalized SRM data demonstrate that the maxi-
are acceptable for tentative comparison of the yields mum yields of iminium ions generated from pyrroli-
of various iminium ions. The initial analysis of each dine, morpholine and piperidine moieties are similar
derivative was performed in product ion scan mode and they are significantly higher in comparison
using collision energy of 10 eV in order to select the with the maximum yields of iminium ions generated
product ions for SRM detection (collision energy of from 2,6-dimethylpiperidine, 1-methylpiperazine and
20 eV was used in the case of 1-carboxymethyl-3-methy- 1-carboxymethyl-3-methylimidazolium moieties (see
limidazolium moiety; see Supplemental material S1). Figure 2; note that CID of the latter moiety resulted
The structures of these product ions and normalized in different fragmentation which will be discussed sep-
areas of corresponding SRM peaks obtained at various arately). Dissociation of tryptamine derivatized with
collision energies are illustrated in Figure 2.
any pyrrolidine, piperidine, 2,6-dimethylpiperidine
or morpholine moiety resulted in ‘‘clean’’ release of
single product ion at collision energy of 10 eV. In the
case of 1-methylpiperazine moiety, corresponding imi-
nium ion demonstrated the highest abundance, but an
Release of different cyclic iminium ions during
CID of derivatized tryptamine
Fragmentation of tryptamine derivatized with any additional low-abundance ion of m/z 70 was also
investigated saturated nitrogen-containing compound observed. The maximum yield of 1-methylpiperazine-
resulted in corresponding iminium ion of high based iminium ion was obtained at collision energy of

ˇ
Zilionis
5
Figure 2. The structures of iminium ions for SRM detection and SRM peak area dependence on CID energy. The standard deviation is
indicated by error bars, n ¼ 3.
20 eV, and fragmentation using this energy resulted in of corresponding iminium ions at optimal collision
significantly increased abundance of an additional ion energies (20, 22 and 20 eV, respectively; see
of m/z 70. Moreover, the peak of m/z 301 shows that Supplemental material S2). Moreover, the peaks of
dissociation of molecular ions was not complete at opti- molecular ions were insignificant at these energies,
mal collision energy (see Supplemental material S2). thus confirming that nearly-complete release of imi-
These data confirm that the secondary low-energy nium product ions occurred. It is worth noting that
fragmentation pathway was active. Since the yield of tentative comparison of the yields of various iminium
product ion was limited by sub-fragmentation to ions became possible due to normalization of SRM
some extent even at low collision energy, it is not sur- data using SIM data. Since the accuracy of SIM meas-
prising that 1-methylpiperazine moiety experimentally urements of differently derivatized tryptamine was
demonstrated the lowest maximum yield of desired imi- impossible to evaluate, the accurate comparison of
nium ion among investigated saturated compounds.
similar normalized SRM values would be questionable
The second lowest maximum yield of saturated (note that the error bars indicate standard deviation of
iminium ion was obtained during CID of tryptamine SRM peak areas in Figure 2). Taking into account that
derivatized with 2,6-dimethylpiperidine moiety (see the model analyte containing any of the latter three
Figure 2). Interestingly, slightly higher collision moieties demonstrated
a preferred fragmentation
energy (24 eV) was required for the maximum yield of behavior, we conclude that the highest yields of product
2,6-dimethylpiperidine-based iminium ion in compari- ions could be expected using these moieties and our
son with the energies optimal for the rest iminium ions, experimental data support this conclusion.
and this could be explained by some steric hindrance at
the nitrogen provided by the two methyl groups which
Stability of iminium ions originated from pyrrolidine
resulted in lower availability of the electrons on nitro-
and piperidine moieties
gen for donation. Fragmentation of tryptamine deriva-
tized with 2,6-dimethylpiperidine moiety yielded In general, iminium ions are capable of further frag-
no additional product ion at collision energy of 10 eV, mentation reactions: McLafferty-type rearrangement
and only low-abundance additional product ion of m/z requires a c-H and thus is possible in the cases of at
58 was observed during fragmentation using optimal least one C3 alkyl substituent, while onium reaction
collision energy of 24eV (see Supplemental material requires at least one C2 alkyl substituent.15,16 The
S2). These data suggest that the maximum yield of 2,6- latter reaction was reported being a multistage process
dimethylpiperidine-based iminium ion was a bit limited that involves ion-neutral complex intermediate and
by sub-fragmentation, and the extent of this limitation H-transfer between the partners in this complex.17,18
was experimentally shown being clearly lower compared It was also reported that H-transfer occurs after rota-
to 1-methylpiperazine-based iminium ion.
tion of the partners in the complex relative to each
The highest maximum product ion yields were other to some degree.17 Since cyclic structure is more
obtained during fragmentation of tryptamine deriva- rigid compared to linear substituents, we hypothesize
tized with pyrrolidine, piperidine or morpholine moi- that the formation of a specific ion-neutral complex and
eties, and these moieties demonstrated ‘‘clean’’ release H-transfer could be restricted during low-energy CID

6
European Journal of Mass Spectrometry 0(00)
Figure 3. Comparison of stabilities of cyclic and linear iminium ions under CID conditions: (a) comparison of pyrrolidine- and diethy-
lamine-based iminium ions; (b) comparison of piperidine- and di-n-propylamine-based iminium ions. The standard deviation is indicated
by error bars, n ¼ 3.
of cyclic iminium ion. In addition, these ions should be based iminium ion lacks c-H, so the only possible low-
more resistant to sub-fragmentation via McLafferty- energy sub-fragmentation pathway of this iminium ion
type rearrangement because the carbon groups is onium reaction – the occurrence of which was con-
attached to a nitrogen atom are ‘‘tied back’’. Taking firmed by CID of correspondingly derivatized trypta-
this into account, we expect cyclic iminium ions being mine at collision energy of 40 eV (the losses of one and
more stable under CID conditions compared to linear two ethene molecules yielded ions of m/z 58 and 30,
iminium ions.
data not shown). Onium reaction also resulted in ring
The maximum yields of iminium ions originated from opening of cyclic iminium ion during CID of trypta-
pyrrolidine and piperidine moieties were among three mine derivatized with pyrrolidine moiety at collision
the highest cyclic ion yields obtained in our experiments. energy of 49 eV. McLafferty-type rearrangement (via
In order to evaluate the influence of cyclic structure C¼C bond) and imine elimination were enabled by dis-
of iminium ion on its stability under CID conditions, ruption of cyclic structure and yielded product ions of
pyrrolidine- and piperidine-based iminium ions were m/z 42 and 55, respectively (see Figure 4(a)). Proposed
compared to diethylamine- and di-n-propylamine- sub-fragmentation pathways of pyrrolidine-based imi-
based iminium ions, respectively. Diethylamine was nium ions are illustrated in Supplemental material S3.
chosen as linear analog of pyrrolidine lacking c-H (see It is noteworthy that the product ion of the second
Figure 3(a)), and di-n-propylamine was chosen as linear onium reaction (ion of m/z 30) was not observed, sug-
analog of piperidine containing flexible alkyl substitu- gesting that onium reaction requires higher collision
ents with c-H (see Figure 3(b)). Our results demonstrate energy compared to McLafferty-type rearrangement
that the maximum yield of cyclic iminium ion was or imine elimination in this case. Since opened ring of
similar to the maximum yield of corresponding linear pyrrolidine-based iminium ion is necessary for the latter
iminium ion in both cases. On the other hand, the two processes, onium reaction could be considered as a
yields of pyrrolidine- and piperidine-based iminium ‘‘bottle neck’’. Taking into account that cyclic iminium
ions were much higher than the yields of corresponding ions were shown being more resistant to sub-fragmen-
linear iminium ions at higher collision energies (roughly tation under CID conditions compared to their linear
at ꢀ35 and ꢀ23eV, respectively). These results show analogs and that onium reaction was a necessary step
that cyclic iminium ions were more resistant to sub- for sub-fragmentation of both pyrrolidine- and diethy-
fragmentation under CID conditions than their linear lamine-based iminium ions, we conclude that onium
analogs.
reaction of cyclic pyrrolidine-based iminium ion
In order to reveal likely sub-fragmentation pathways requires higher collision energy compared to diethyla-
of different iminium ions, mass spectrometric analysis mine-based iminium ion, and this supports the hypoth-
of differently derivatized tryptamine was performed in esis we previously made.
product ion scan mode using relatively high collision
Di-n-propylamine-based iminium ion contains c-H
energies which previously resulted in SRM peak areas and thus is capable of McLafferty-type rearrangement
similar to those what were obtained using collision and onium reaction. The occurrence of both these reac-
energy of 10 eV (see Figure 4(a) to (e)). Diethylamine- tions was confirmed by CID of correspondingly

ˇ
Zilionis
7
Figure 4. Product ion mass spectra of derivatized tryptamine obtained at relatively high collision energies and the proposed structures of
secondary product ions originated from pyrrolidine (a, 49 eV), piperidine (b, 51 eV), 1-methylpiperazine (c, 33 eV), morpholine (d, 43 eV),
2,6-dimethylpiperidine (e, 49 eV) and methylimidazolium (f, 40 eV) moieties.
derivatized tryptamine at collision energy of 30 eV (the iminium ion via onium reaction with subsequent
losses of ethene and propene molecules yielded ions of McLafferty-type rearrangement (via C¼N bond),
m/z 86 and 72, data not shown). McLafferty-type while additional double-bond migration with subse-
rearrangement was shown being more favorable frag- quent McLafferty-type rearrangement (via C¼C
mentation pathway of di-n-propylamine-based iminium bond) could yield the ion of m/z 70. The latter ion
ion at several different collision energies (10–40 eV, data could undergo imine elimination yielding the ion
not shown) strongly suggesting that this rearrangement of m/z 41. We also speculate that McLafferty-type
requires lower energy compared to onium reaction in rearrangement (via C¼N bond after double-
the case of flexible alkyl substituents containing c-H. bond migration) with subsequent imine elimination
Fragmentation of tryptamine derivatized with piperi- yielded the second most intense product ion of m/z
dine moiety at collision energy of 51 eV yielded various 55. Despite the fact that H-transfer via internal 6-mem-
product ions and the assumption that 6-membered bered ring forms an additional internal 4-membered
cyclic iminium ion undergoes double-bond migration ring and thus is expected being spatially restricted,
with subsequent retro Diels-Alder (RDA) reaction or this unlikely McLafferty-type rearrangement was con-
McLafferty-type rearrangement was necessary for sidered being the reason for ring opening in this case
explanation of the origin of these ions (the proposed because we were not able to explain the ion of m/z 55 in
structures are illustrated in Figure 4(b)). We speculate any other way. Proposed sub-fragmentation pathways
that the most intense product ion of m/z 70 was of piperidine-based iminium ions are illustrated in
obtained during RDA reaction of cyclic iminium ion Supplemental material S4. Although these fragmenta-
after double-bond migration, and two different struc- tion pathways were not proved and the proposed struc-
tures were possible depending on the exact position of tures of the secondary product ions were not confirmed,
double-bond prior to RDA reaction. Moreover, RDA our results in Figure 3(b) clearly demonstrate that
reaction after double-bond migration could also yield piperidine-based cyclic iminium ion requires higher
the product ion of m/z 44. On the other hand, this energy for the primary sub-fragmentation step
product ion could be obtained during ring opening of (or competing steps) compared to fragmentation of

8
European Journal of Mass Spectrometry 0(00)
linear di-n-propylamine-based iminium ion via conven- onium reaction. However, we speculate that cyclic
tional McLafferty-type rearrangement and onium structure was also disrupted during McLafferty-type
reaction.
rearrangement (via C¼N and C¼C bonds) after
double-bond migration, and the ions of m/z 69 and 55
were yielded during subsequent elimination of the neu-
tral imine. Although the second fragmentation pathway
is questionable (see discussion of piperidine-based imi-
nium ion fragmentation), we were not able to explain the
Stability of iminium ions originated from
1-methylpiperazine, morpholine and
2,6-dimethylpiperidine moieties
It was reported that sub-fragmentation of iminium ions origin of the latter two ions in any other way. Proposed
containing the piperazine ring could result in secondary sub-fragmentation pathways of 2,6-dimethylpiperidine-
product ions of m/z 70 (C4H8Nþ) and 98 (C5H10Nþ2 ).19 based iminium ions are illustrated in Supplemental
We have shown 1-methylpiperazine-based iminium ion material S7.
being of low stability under CID conditions and sub-
fragmentation of this ion yielded relatively intense
Fragmentation of 1-carboxymethyl-
product ion of m/z 70 at collision energy of 20 eV (see
3-methylimidazolium moiety
Supplemental material S2). This loss of 43 Da is one of
the typical losses from a methyl-substituted piperazine The lowest maximum yield of product ion for SRM
ring, and the lost molecule was reported being detection was obtained during fragmentation of trypta-
N-methylmethanimine.19 Taking this into account, we mine derivatized with 1-carboxymethyl-3-methylimida-
conclude that RDA reaction occurred after the double- zolium moiety (see Figure 2). While fragmentation of
bond migration. Another reported secondary product other investigated derivatives resulted in the formation
ion of m/z 98 was also observed using collision energy of expected iminium ions at collision energy of 10 eV,
of 33 eV (see Figure 4(c)), and the origin of this ion CID of derivatized model analyte did not yield a
could be explained by the neutral loss of methyl radical. double-charged iminium ion in this case because of
Proposed sub-fragmentation pathways of 1-methylpi- imidazolium p-electron system involving a pair of elec-
perazine-based iminium ions are illustrated in trons from N1 atom. Moreover, only negligible product
Supplemental material S5.
ions were observed using collision energy of 10 eV (data
Morpholine-based iminium ion was shown to be not shown). Since this cationic ring system was
among three the most stable iminium ions, and the reported to undergo multiple stages of CID,20 mass
stability of the rest two ions (piperidine- and pyrroli- spectrometric analysis of tryptamine derivatized with
dine-based) has been discussed above. Two intense sec- 1-carboxymethyl-3-methylimidazolium moiety was
ondary product ions of m/z 56 and 70 were observed repeated in product ion scan mode at higher collision
during CID of tryptamine derivatized with morpholine energies (20, 30 and 40 eV). Although various add-
moiety at collision energy of 43 eV (see Figure 4(d)). itional product ions were observed using any collision
The ion of m/z 56 is expected being yielded by energy, the product ion of m/z 83 was the most intense
McLafferty-type rearrangement (via C¼C bond) after in all cases (see Figure 4(f) and Supplemental material
ring opening via onium reaction, while the ion of m/z S1). This ion was reported being 1-methyl-3H-imidazo-
70 was likely obtained during RDA reaction after lium (C4H7Nþ2 ), and fragmentation pathway yielding
double-bond migration. Interestingly, the latter path- this ion was reported being similar to the cleavage of
way yielding product ion of m/z 70 was also active the alkyl C–N bond of N-alkyl pyrroles with hydrogen
during sub-fragmentation of piperidine- and 1-methyl- transfer from the leaving alkyl moiety to the nitrogen.20
piperazine-based iminium ions, but morpholine- and At least three additional product ions of m/z 95, 96 and
piperidine-based iminium ions were shown being sig- 123 were originated from 1-carboxymethyl-3-methyli-
nificantly more resistant to sub-fragmentation under midazolium moiety using collision energy of 40 eV,
CID conditions in comparison with 1-methylpipera- while the second most intense product ion of m/z 144
zine-based iminium ion. Since RDA reaction can be was originated from tryptamine moiety (see Figure 4(f);
regarded as a double a-cleavage, lower resistance of an excellent scheme of sub-fragmentation pathways of
1-methylpiperazine-based iminium ion to RDA reac- 1-carboxymethyl-3-methylimidazolium was reported in
tion could be explained by lower energy of C-N bond Lesimple et al.20). Several active dissociation pathways
in comparison with either C–C or C–O bond. Proposed yielding different product ions clearly lowered the
sub-fragmentation pathways of morpholine-based imi- abundance of the main product ion, thus limiting the
nium ions are illustrated in Supplemental material S6. application of 1-carboxymethyl-3-methylimidazolium
CID of tryptamine derivatized with 2,6-dimethylpiper- moiety as derivatization reagent for SRM detection.
idine moiety resulted in low-abundance secondary product
ion of m/z 58 at collision energy of 24 eV, and secondary
Differences in ionization efficiencies of derivatized
product ions of m/z 55, 58 and 69 were obtained using
tryptamine
collision energy of 49 eV (see Figure 4(e)). The ion of
m/z 58 was likely obtained during McLafferty-type The total amount of product ion depends not only on
rearrangement (via C¼N bond) after ring opening via the dissociation of molecular ion but also on the yield

ˇ
Zilionis
9
of molecular ion in electrospray ionization source. All concentrations and ratios of the reagents in order to
investigated saturated compounds are tertiary amines minimize side-products from the reaction.
containing basic nitrogen atom which could be proto- Unfortunately, significant amounts of UV-absorbing
nated during þESI. Despite the fact that 1-methylpi- side-products were obtained in all cases. Taking this
perazine moiety demonstrated the lowest iminium ion into account, EDC/NHS strategy was replaced by the
yield during CID of corresponding molecular ion, this active NHS esters. Such esters of (4-methyl-1-piperazi-
moiety contains an additional nitrogen atom possibly nyl)acetic acid and 4-morpholinylacetic acid were
increasing ionization efficiency. As a result, the total synthesized, reacted with tryptamine and the products
yield of 1-methylpiperazine-based iminium ion is not of interest were the same as synthesized using EDC/
necessarily the lowest. Morpholine moiety also contains NHS strategy because NHS esters are selective reagents
an additional charge-localizing oxygen heteroatom. for modification of primary amino groups.3,21 The
Moreover, the latter moiety demonstrated one of the injection volumes were adjusted, so that the areas of
highest iminium ion yields during CID of corresponding UV peaks of differently derivatized tryptamine were
molecular ion. Whereas pyrrolidine, piperidine and 2,6- approximately the same for both different reaction mix-
dimethylpiperidine moieties do not contain any add- tures (see Figure 5(a)) and we assume that very similar
itional basic atom, morpholine moiety was considered amounts of analytes were physically infused into ESI
being easier to ionize than these moieties. With this in source. RP-UPLC analysis with isocratic elution was
mind, we compared the effects of 1-methylpiperazine and chosen in order to maintain the same ionization condi-
morpholine moieties on þESI of derivatized tryptamine. tions for differently derivatized tryptamine.
An additional UV detection was necessary for nor-
Tryptamine derivatized with morpholine moiety
malization of SIM data in order to compare ESI yields. demonstrated significantly higher þESI efficiency
Model analyte tryptamine absorbs UV light at 280 nm. than tryptamine derivatized with 1-methylpiperazine
Since derivatization reagents do not absorb UV light at moiety (2þ ions were not obtained and þESI yields of
this wavelength and the effect of an additional peptide 1þ ions differed &1.2 times, see Figure 5(b)). Taking into
bond on UV absorbance is insignificant at 280 nm, deri- account that in general oxygen is somewhat less charge-
vatization of tryptamine was considered causing negli- stabilizing heteroatom than nitrogen, lower þESI effi-
gible changes in absorbance. EDC/NHS strategy was ciency of tryptamine derivatized with 1-methylpiper-
simple to perform, but it resulted in UV-absorbing side- azine moiety could be explained by some steric
products which were not completely separated from the hindrance at the nitrogen atom provided by the methyl
products of interest. Although co-eluting compounds group, which makes it more difficult for nitrogen atom to
do not interfere with normalization of SRM data approach a proton during ionization. The same RP-
using SIM peaks, the overlapped UV peaks are not UPLC analysis with SRM detection and UV normaliza-
acceptable for normalization of SIM data. Detailed tion was performed for a comparison of the total product
mass spectrometric analysis showed that co-eluting ion yields and &1.9 times higher product ion yield was
side-products were tryptamine coupled to one or sev- obtained using morpholine moiety in comparison with 1-
eral b-alanine molecules (data not shown). As b-alanine methylpiperazine moiety (see Figure 5(c)). It is note-
is a side-product of EDC reaction with NHS,13 the worthy that 1-carboxymethyl-3-methylimidazolium
coupling was performed using
a
variety of moiety contains a permanent charge and thus
Figure 5. The chromatograms of tryptamine derivatized with 1-methylpiperazine and morpholine moieties: (a) UV chromatograms;
(b) SIM chromatograms; (c) SRM chromatograms. The relative standard deviation was <1.7% in all cases, n ¼ 3.

10
European Journal of Mass Spectrometry 0(00)
2. Reid GE, Roberts KD, Simpson RJ, et al. Selective
identification and quantitative analysis of methionine
containing peptides by charge derivatization and
tandem mass spectrometry. J Am Soc Mass Spectrom
2005; 16: 1131–1150.
3. Zhou X, Lu Y, Wang W, et al. ‘Fixed charge’ chemical
derivatization and data dependant multistage tandem
mass spectrometry for mapping protein surface residue
additionally acts as an ionization enhancer of derivatized
analyte resulting in possibly the highest yield of molecu-
lar ion among investigated compounds. Unfortunately,
we were not able to synthesize corresponding active NHS
ester for a ‘‘clean’’ derivatization of the model analyte.
Since fully resolved UV peak was not obtained in this
case, þESI yield was not evaluated.
accessibility.
1339–1351.
J Am Soc Mass Spectrom 2010; 21:
Conclusions
4. Bantscheff M, Lemeer S, Savitski MM, et al. Quantitative
mass spectrometry in proteomics: critical review update
from 2007 to the present. Anal Bioanal Chem 2012; 404:
939–965.
5. Ross PL, Huang YN, Marchese JN, et al. Multiplexed
protein quantitation in Saccharomyces cerevisiae using
amine-reactive isobaric tagging reagents. Mol Cell
Proteom 2004; 12: 1154–1169.
6. Xiang F, Ye H, Chen R, et al. N,N-dimethyl leucines as
novel isobaric tandem mass tags for quantitative prote-
omics and peptidomics. Anal Chem 2010; 82: 2817–2825.
7. Dayon L, Hainard A, Licker V, et al. Relative quantifi-
cation of proteins in human cerebrospinal fluids by
MS/MS using 6-plex isobaric tags. Anal Chem 2008; 80:
2921–2931.
8. Che F and Fricker LD. Quantitative peptidomics of
mouse pituitary: comparison of different stable isotopic
tags. J Mass Spectrom 2005; 40: 238–249.
9. Cydzik M, Rudowska M, Stefanowicz P, et al.
The competition of charge remote and charge directed
fragmentation mechanisms in quaternary ammonium
salt derivatized peptides – an isotopic exchange study.
J Am Soc Mass Spectrom 2011; 22: 2103–2107.
10. Pappin DJC et al. Isobarically labeled analytes and frag-
ment ions derived therefrom. Patent US 8,273,706 B2,
USA, 2012.
11. Vaughn HL and Robbins MD. Rapid procedure for the
hydrolysis of amides to acids. J Org Chem 1975; 40:
1187–1189.
12. Montalbetti CAGN and Falque V. Amide bond forma-
tion and peptide coupling. Tetrahedron 2005; 61:
10827–10852.
In this study, six nitrogen-containing cyclic compounds
were examined as product ion sources for selected reac-
tion monitoring detection of derivatized tryptamine.
Nearly-complete release of the added chemical deriva-
tive group without any sub-fragmentation of the result-
ing product ion was observed during dissociation of the
model analyte derivatized with pyrrolidine, piperidine
or morpholine moiety, and the highest fragmentation
yields were confirmed experimentally. Cyclic iminium
ions were shown being more resistant to sub-fragmen-
tation under CID conditions in comparison with their
linear analogs likely due to spacial restrictions limiting
onium reaction and McLafferty-type rearrangement.
Although cyclic structures were shown being add-
itionally capable of RDA reaction, primary sub-frag-
mentation processes of cyclic iminium ions required
higher collision energy compared to conventional
McLafferty-type rearrangement and onium reaction of
linear iminium ions. Expected sub-fragmentation path-
ways of various iminium ions were discussed, and the
structures of experimentally obtained secondary prod-
uct ions were proposed. Considering both ionization of
derivatized tryptamine and fragmentation of molecular
ions processes, morpholine moiety was shown being a
promising product ion source for SRM detection of
derivatized analytes. However, further investigation of
permanent charge-containing moieties is required.
Declaration of conflicting interests
13. Totaro KA, Liao X, Bhattacharya K, et al. Systematic
investigation of EDC/sNHS-mediated bioconjugation
reactions for carboxylated peptide substrates. Bioconjug
Chem 2016; 27: 994–1004.
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
14. Fischer MJE. Amine coupling through EDC/NHS: a
practical approach. In: De Mol NJ and Fischer MJE
(eds) Surface plasmon resonance. 1st ed. New York:
Humana Press, 2010, pp.58–60.
Funding
The author(s) received no financial support for the research,
authorship, and/or publication of this article.
15. Gross JH. Mass spectrometry. 1st ed. Berlin: Springer –
Verlag Berlin Heidelberg, 2004, pp.292–294.
Supplemental material
16. Watson CH, Baykut G, Battiste MA, et al. Infrared
multiphoton and collision-induced dissociation studies
of same gaseous alkylamine ions. Anal Chim Acta 1985;
178: 125–136.
Supplemental material for this article is available online.
ORCID iD
17. Hudson CE and McAdoo DJ. Characterization by theory
of H-transfers and onium reactions of CH3CH2CH2Nþ
H¼CH2. J Am Soc Mass Spectrom 2007; 18: 270–278.
18. Hudson CE and McAdoo DJ. Ethene elimination from
CH3CH2NH¼CHþ2 : reaction pathways at the boundary
between stepwise and concerted processes. J Am Soc
Mass Spectrom 1998; 9: 138–148.
Andrius Zilionis
https://orcid.org/0000-0002-3358-2356
References
1. Lange V, Picotti P, Domon B, et al. Selected reaction
monitoring for quantitative proteomics: a tutorial. Mol
Syst Biol 2008; 4: 1–14.

ˇ
Zilionis
11
19. Niessen WMA. Fragmentation of toxicologically relevant
drugs in positive-ion liquid chromatography-tandem
mass spectrometry. Mass Spectrom Rev 2011; 30:
626–663.
20. Lesimple A, Mamer O, Miao W, et al. Electrospray mass
spectral fragmentation study of N,N0-disubstituted
imidazolium ionic liquids. J Am Soc Mass Spectrom
2006; 17: 85–95.
21. Schmidt A, Kellermann J and Lottspeich F. A novel
strategy for quantitative proteomics using isotope-coded
protein labels. Proteomics 2005; 5: 4–15.