Charge inversion under plasma-nanodroplet reaction conditions excludes Fischer esterification for unsaturated fatty acids: a chemical approach for type II isobaric overlap

Direct infusion ionization methods provide the highest throughput strategy for mass spectrometry (MS) analysis of low-volume samples. But the trade-off includes matrix effects, which can significantly reduce analytical performance. Herein, we present a novel chemical approach to tackle a special type of matrix effect, namely type II isobaric overlap. We focus on detailed investigation of a nanodroplet-based esterification chemistry for differentiating isotopologue [M + 2] signal due to unsaturated fatty acid (FA) from the monoisotopic signal from a saturated FA. The method developed involves the online fusion of nonthermal plasma with charged nanodroplets, enabling selective esterification of saturated FAs. We discovered that unsaturated FAs undergo spontaneous intramolecular reaction via a novel mechanism based on a carbocation intermediate to afford a protonated lactone moiety (resonance stabilized cyclic carbonium ion), whose mass is the same as the original protonated unsaturated FA. Therefore, the monoisotopic signal from any saturated FA can be selectively shifted away from the mass-to-charge position where the isobaric interference occurs to enable effective characterization by MS. The mechanism governing the spontaneous intramolecular reactions for unsaturated FAs was validated with DFT calculations, experimentation with standards, and isotope labeling. This novel insight achieved via the ultrafast plasma-nanodroplet reaction environment provides a potentially useful synthetic pathway to achieve catalyst-free lactone preparation. Analytically, we believe the performance of direct infusion MS can be greatly enhanced by combining our approach with prior sample enrichment steps for applications in biomedicine and food safety. Also, combination with portable mass spectrometers can improve the efficiency of field studies since front-end separation is not possible under such conditions.


Coconut oil analysis S21
Topic 14 References S24

Experimental section
Mass spectrometry.Thermo Fisher Scientific Velos Pro LTQ and LTQ Orbitrap (for highresolution detection) mass spectrometers (San Jose, CA, USA) were used for analysis and data collection.Unless otherwise stated, MS parameters applied were as follows: 400 °C capillary temperature, 5 mm distance from an ion source to MS inlet, 3 microscans, 100 ms ion injection time, 60% S-lens voltage for Velos Pro LTQ and 55V Tube Lens voltage for LTQ Orbitrap.Spray voltage in the range of 1.5 − 2 kV was used for contact and noncontact nano-ESI MS analysis.To induce plasma during nano-ESI, 6 kV was used, which enabled plasma-nanodroplet fusion.Headspace vapor analysis (in the absence of nano-ESI) was also performed using 6 kV applied to a silver (Ag) electrode in close proximity to the sample to be analyzed (See Fig S1A).Solids were also analyzed in a similar manner, by applying 6 kV to the Ag electrode, in the absence of nano-ESI.Mass spectra were recorded for at least 30 s, yielding an average of 300 individual scans.Data collection and processing for MS were performed with Thermo Fisher Scientific Xcalibur 2.2 SP1 software.Unless otherwise mentioned, Tandem MS with collision-induced dissociation (CID) was used for structural characterization of analyte.We used 30% (manufacturer's unit) and 1.5 Th (mass / charge units) of normalized collision energy and isolation window for CID experiments, respectively.

2.
Full MS analysis of saturated and unsaturated fatty acids.
Table S1.Full MS analysis and esterification of saturated fatty acids achieved using methanol vapor and the vapors of the carboxylic acids in the presence of positive and negative corona discharge.In the Table (-) and (+) denote fragments detected in negative and positive-ion mode.Fragments in (-) and (+) column are shown in order of decreasing MS relative intensity.Mass accuracy (in ppm) data are provided for high resolution MS data for some ester products.
Table S2.Full MS analysis and esterification of unsaturated fatty acids when using the vapors of methanol and carboxylic acids, in positive (+) and negative (-) plasma.Fragments are shown in order of decreasing MS relative intensity.

Fig. S1. (A)
Setup for contained nano-atmospheric pressure chemical ionization mass spectrometry (nAPCI MS) used for analyzing headspace vapors of compounds. [1]Corona discharge is produced at the tip of the Ag electrode during application of DC voltage >4 kV.Potential also causes reactive olfaction to electrostatically attract/adsorb the solid particles of an analyte (in the glass vial) on the electrode surface with subsequent APCI ionization.The PTFE container concentrates the vapor-phase molecules/particles into a small space and the glass capillary collimates the vapor toward the plasma.The valve on the side of the PTFE container increases vapor flow rate by reducing the saturation of air above the sample.For esterification, the methanol vapor was brought in between the nAPCI source and the MS inlet.Positive-ion mode mass spectra recorded for esterification reactions between headspace vapors of methanol and saturated fatty acids: (B) propionic acid, (C) butyric acid, (D) valeric acid, and (E) myristic acid.Nano-ESI was not used here.The experimental setup for analyzing vapors is shown below.The geometry optimization of the reagent, products and all intermediates along the transformation pathway was caried out by using B3LYP functional, [2,3] 6-311g(d,p) [4,5] basis and empirical dispersion at GD3 [6] level.All calculations were performed in a gas phase approximation.The energetics along transformation mechanism was calculated with accounting of zero point energy (ZPE) and Gibbs free energy corrections.All the simulations were realized by using Gaussian16 software. [7]g. S7.(A) DFT free energy diagram of (B) gas-phase lactone (9'') synthesis from oleic acid (1'') through intermediate hydroxycarboxylic acid (4'') formation.

Proton affinity discussion and calculations
Differences in proton affinity ruled out.Since protonation is the crucial step in Fisher esterification, our next investigation was related to probable difference in proton affinities for unsaturated versus saturated FAs.Unfortunately, the literature does not provide sufficient information about proton affinities of FAs to make any informed conclusions.However, it is known that measured pKa values in solution increase with increasing FA chain length but decrease with increasing number of double bonds. [8]f we assume the same acidity/basicity trend for gas phase species, then proton affinities of unsaturated FAs should be less than saturated species, which may prevent their protonation and subsequent esterification.Intuitively, we can expect a higher C=C bond influence on carboxyl group acidity when it is closest.But as we discussed above, proximity of double bond has the opposite effect on esterification; it is favored.Moreover, hydronium ions (PA 725.6 kJ/mol), thought to be present in our experiment , have sufficient ability to protonate any fatty acid, including the strongest carboxylic acids such as formic acid (PA 742.0 kJ/mol), acetic acid (PA 783.7 kJ/mol), and propanoic acid (PA 797.2 kJ/mol).Thus, it is unreasonable to believe that differences in proton affinities, if any, can suppress esterification of unsaturated FAs.
Proton affinities (PA) were calculated according to formula:

Coconut oil analysis
Coconut oil and hydrolyzed coconut oil were diluted in methanol (1:1000) prior to analysis.The hydrolysis of coconut oil was performed by solution of KOH in ethanol.2.5 mL of the melted oil was mixed with 3.0 mL of 15% KOH in ethanol and then heated till all ethanol evaporated.
MS analysis of saturated and unsaturated fatty acids S4

Fig. S2 .
Fig. S2.Positive-ion mode mass spectrum derived from the analysis of acetic acid in the presence of ethanol headspace vapor, which reacted via esterification mechanism to give expected product at m/z 103.(A) Full MS spectrum recorded in real-time during a plasma desorption/ionization experiment and (B) schematic illustration of expected reaction.

Fig. S4 .
Fig. S4.Positive-ion mode MS/MS spectra of (A) reaction product (MW 88 Da) generated from plasma ionization of propionic acid in the presence of methanol vapor, (B) standard butyric acid, (C) reaction product (MW 102 Da) generated from analysis of butyric acid in the presence of methanol, and (D) standard of valeric acid.The data excludes methylene insertion, which will afford the carboxylic acid with one carbon longer than the reactant.The fragmentation pattern of the observed ester products (A and C) is markedly different from the corresponding acids of the same molecular weight (B and D).

Fig. S5 .
Fig. S5.Positive-ion mode mass spectrum of palmitic acid recorded in the presence of MeOH headspace vapor, after exposure to corona discharge.

Fig. S13 .
Fig. S13.Positive-ion mode MS/MS spectra of (A) oleic acid (500 µM) and (B) oleic (500 µM)/stearic (125 µM) acids mixture at m/z 285 position where isobaric overlap occurs.It could be difficult to infer the presence of stearic acid without shifting the mass to another m/z position.

Fig. S17 .
Fig. S17.MS/MS data for stearic acid (A) in negative and (B) positive ion modes.

Fig. S18 .
Fig. S18.Tandem MS data for oleic acid (A) in negative and (B) positive ion modes.

Fig. S19 .
Fig. S19.MS/MS spectra of linoleic acid (A) in negative and (B) positive ion modes.

Fig. S22 .
Fig. S22.Negative-ion mode mass spectra of (A) diluted coconut oil and (B) hydrolyzed coconut oil at -2 kV nano-ESI, in the absence of plasma.

Fig. S23 .
Fig. S23.Positive-ion mode mass spectra of (A) diluted coconut oil and (B) hydrolyzed coconut oil at 2 kV nano-ESI, in the absence of plasma.Insert in (A) shows absence of fatty acids at lower m/z range in positiveion mode at low spray voltage.Insert in (B)shows molecular formula assignment based on previous report,[9] which also detected the same peaks.

Fig. S25 .
Fig. S25.Negative-ion mode MS analysis of Coconut oil using (A) -2 kV, in the absence of plasma and (B) -4 kV in the presence of plasma.Negative plasma fusion into negative droplets do not result in any noticeable change in mass spectra.

Fig. S26 .
Fig. S26.Positive-ion mode MS analysis of hydrolyzed coconut oil at (A) 4 kV and (B) 6 kV, all in the presence of plasma with +6 kV showing more reactive droplets for esterification.Peaks labeled with red font are ester products of the corresponding acid [C9 from C8:0 and C13 from C12:0].Literature confirms the odd number saturated fatty acids are not found in coconut oil.[10]