Analysis of Conjugated Fatty Acid Isomers by the ... - ACS Publications

Subscriber access provided by UNIV OF LOUISIANA

Article

Analysis of Conjugated Fatty Acid Isomers by the PaternòBüchi Reaction and Trapped Ion Mobility Mass Spectrometry Xiaobo Xie, and Yu Xia Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.9b00374 • Publication Date (Web): 10 May 2019 Downloaded from http://pubs.acs.org on May 10, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Analysis of Conjugated Fatty Acid Isomers by the Paternò-Büchi Reaction and Trapped Ion Mobility Mass Spectrometry Xiaobo Xie† and Yu Xia*, † † MOE Key Laboratory of Bioorganic Phosphorus Chemistry & Chemical Biology, Department of Chemistry, Tsinghua University, Beijing 100084, China ABSTRACT: Fatty acids containing conjugated carbon-carbon double bonds (C=Cs), such as conjugated linoleic acids (CLAs), attract growing research interest due to their bioactivities against diabetes, cancer, and atherosclerosis. Analysis of conjugated fatty acid (CFA) is challenging for existing analytical techniques because it requires determination of geometry (cis (Z) vs. trans (E)) and location of individual C=C. In this study, we have developed a method to achieve confident, fast, and quantitative analysis of CFA isomers from mixtures. This method combines the strength of trapped ion mobility spectrometry (TIMS) for fast isomer separation and the Paternò-Büchi (PB) reaction followed by tandem mass spectrometry (MS/MS) for C=C location determination. Notably, the PB reaction of CFA is regio-selective to terminal C=Cs, thus forming diagnostic fragment ions unique to conjugated C=Cs from PBMS/MS. These fragment ions facilitate identification and quantitation of individual CLA isomers differing in C=C locations, affording limit of identification of 1 nM. Given that PB-MS/MS alone cannot identify the geometry of C=C, TIMS has been employed for characterizing C=C geometry. TIMS is capable to separate various C=C geometric isomers of CLAs, allowing visualization of C=C isomerization during the PB reaction. By coupling the PB-MS/MS with TIMS, two CLA isomers, CLA 18:2(9Z,11E) (46.9 ± 1.1%) and CLA 18:2(10E,12Z) (53.1 ± 1.1%), are quantified in a commercial CLA supplement.

Conjugated fatty acids (CFAs) containing two or more conjugated carbon-carbon double bonds (-C=C-C=C-) are less common in nature as compared to their non-conjugated analogues (-C=C-CH2-C=C-), a key structural component in polyunsaturated fatty acids (PUFAs). One type of better-known CFAs is conjugated linoleic acids (CLAs). CLAs contain a collective group of octadecadienoic acids (FA 18:2) bearing two conjugated double bonds differing in combinations of C=C locations and C=C geometries (cis (Z) or trans (E)). CLAs are biosynthesized from Δ9,12 linoleic acid or Δ9,12,15 linolenic acid in rumen via microbial biohydrogenation or they can be converted from Δ11 vaccenic acid by delta-9 desaturase in the mammary gland.1 While dozens of CLA isomers are present in ruminant fat, CLA 18:2(9Z,11E) accounts for 80-90% of total CLA, followed by CLA 18:2(10E,12Z) (~5%).2 Recently, CLAs have become a subject of increasing research interest due to potential health benefits against diabetes, cancer, and atherosclerosis.3-4 Different CLA isomers are reported to have distinct bioeffects while the mechanism remains elusive.5-6 Obviously, the capability to distinguish and quantify individual CFAs is highly needed to unravel the mechanisms. CFAs often exist as multiple isomers in natural or synthetic sources. Established analytical techniques mainly rely on gas chromatography (GC)7 or silver ion high-performance liquid chromatography (Ag+-HPLC)8 to resolve possible CFA isomers while mass spectrometry (MS) is employed for detection. Careful optimization and relatively long separation time are necessary to achieve desirable C=C positional and geometrical separations of CFAs. 9 Ion mobility spectrometry (IMS) is a gas-phase technique allowing quick separations (in milliseconds) of ions according to differences in their structures and shapes. A variety of IMS techniques have been coupled

with MS (IM-MS) to resolve isomeric structures that may be indistinguishable by MS alone.10-14 For isomers bearing small structural differences, such as different geometry of C=C in a lipid molecule, separations have been attempted by high resolving power IMS 15 or through derivatizations16 or adduction formation17 to increase conformation differences among isomers.17 For instance, ultra-high resolution IMS separations for two phosphatidylcholine lipid standards differing in C=C geometry have been realized by applying traveling waves in a serpentine multi-pass structures for lossless ion manipulations platform.18 Because of the need to compare retention time of an unknown sample to that of the synthetic standards, limited availability of commercial CFA standards becomes a bottleneck for methods based on separation. Tandem mass spectrometry (MS/MS) via collision-induced dissociation (CID) is a technique routinely used for lipid identification and quantitation.19 Distinguishing C=C positional and geometrical isomers via CID alone is difficult. To solve this problem, chemical derivatizations targeting C=Cs are often employed before or coupled to MS analysis. Among them, Diels−Alder reaction is specific to conjugated C=Cs and thus it allows high sensitivity and selectivity to CFAs in the presence of more abundant PUFA analogues.20 This method, however, has only been demonstrated to dienes, and it cannot provide geometry information of C=Cs. Ozonolysis cleaves C=C with high efficiency and provides signature fragments for straightforward C=C location identification.21 Curtis group has demonstrated ozonolysis coupled in-line with Ag+-LC-MS for the analysis of CLA isomers from milk and bacteria.22 Our group and others have utilized the Paternò-Büchi reaction, a [2+2] cyclo-addition reaction involving C=C and electronically excited carbonyl compounds, to selectively derivatize C=C.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Scheme 1. PB-MS/MS for isolated C=C and conjugated C=Cs using D6-acetone as the PB reagent. (a) For isolated C=C, the PB

reaction shows no regio-selectivity and thus a pair of diagnostic ions, viz. FA and FO are generated for each C=C bond from PBMS/MS. (b) For conjugated C=Cs, acetone shows regio-selectivity to terminal C=Cs, driven by the stability of radical intermediates. Consequently, PB-MS/MS only produces two diagnostic ions regardless of the total number (n) of conjugated C=C bonds. Isomerization of C=C and other minor side reactions also accompany the PB reaction. This derivatization when coupled with subsequent CID or ultraviolet photodissociation (UVPD) is capable of identification and quantitation of C=C location isomers from different classes of lipids containing non-conjugated C=Cs.23-31 Due to fast C=C isomerization during the PB reaction process, PB-MS/MS cannot differentiate C=C geometry.32-33 Several gas-phase reaction or dissociation methods have been developed, aiming to characterize lipids with information of C=C location or geometry. These include ozone-induced dissociation (OzID)34 , 193 nm-UVPD35 , electron impact excitation of ions from organics (EIEIO)36 , ion/ion reactions, and ion/molecule reactions37-38 . Above approaches are all capable of providing C=C location information; however, C=C geometry determination is less straightforward. Blanksby and co-workers show that the rate of OzID is faster for trans than cis C=Cs39 and it is about 200-times faster for conjugated C=Cs34 than non-conjugated systems. Based on careful measurement of reaction rate differences, differentiation of C=C geometry and thus identification of methylated CLA isomers have been achieved for dietary supplement.34 Obviously, solving challenging isomeric problems relating to CFAs requires combining the strength of separation techniques and detailed structural determination capability from MS/MS. We consider that the PB-MS/MS approach is a good candidate for C=C location determination due to its high sensitivity and

wide compatibility with commercial mass spectrometers.29 Although the PB reaction of conjugated C=Cs is not much utilized in organic synthesis, there is evidence supporting that excited carbonyl can add onto conjugated C=Cs to form oxetane rings, same as that observed for isolated C=Cs.40 Therefore, we are interested in exploring the PB reaction chemistry of conjugated C=Cs and its applicability for the analysis of CFA isomers. Regarding isomer separation, we choose to use the trapped ion mobility spectrometry (TIMS). TIMS is a relatively recent advance in the IMS field, featuring relatively high resolving power (up to 200), desirable for isomer seperations.41 In this study, we have paired TIMS and PB-MS/MS for the analysis of CFA isomers. The PB reaction chemistry of conjugated C=Cs is different from isolated C=Cs in that the PB reagent (acetone) shows high regio-selectivity upon addition. This unique chemistry allows detection of C=C diagnostic ions unique to conjugated systems from PB-MS/MS, thus facilitating localization of terminal C=Cs in CFAs. Complementary to PB-MS/MS, TIMS is capable of separating C=C geometric isomers as demonstrated by standards of CLA isomers. This capability enables visualization of C=C isomerization in CLAs during the PB reaction. By combining PB-MS/MS and TIMS separations, identification and quantification of CLA isomers in a commercial CLA supplement has been achieved.


Page 2 of 9

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Negative ion mode ESI MS spectra of the PB reactions of (a) LA 18:2(9Z,12Z) (5 μM), (b) CLA 18:2(10E,12Z) (5 μM), using D0acetone/H2O(v/v, 50/50) solvent system, and (c) CLA 18:2(10E,12Z) (5 μM), using D6-acetone/ H2O/MeOH (v/v/v, 68/22/10) solvent system. (d) Plots of relative ion abundances of CLA 18:2(10E,12Z) ([M-H]-, m/z 279.2) and the PB products ([PBM-H]-, m/z 343.3) as a function of UV exposure time using the same reaction condition as in (c). FA 13:0 was used as internal standard (IS, 5 μM). PB-MS2 CID spectra of (e) LA 18:2(9Z,12Z) ([PBM-H]-, m/z 343.3) and (f) CLA 18:2(10E,12Z) ([PBM-H]-, m/z 343.3) (CE = 35 eV).

EXPERIMENTAL SECTION Nomenclature. Lipid nomenclature is adopted from LIPID MAPS.42 Briefly, FA 18:2 (9Z,11E) denotes a 18-carbon fatty acid with two degrees of unsaturation. The position and geometry of C=C are defined in parentheses. Delta nomenclature is used for C=C location annotation (counted from carboxylic end); the geometry of a C=C is indicated by Z (cis) or E (trans). Chemicals. FA 18:2(9Z,11E) and FA 18:2(10E,12Z) were purchased from Nu-Chek Prep (Elysian, MN, USA); FA 18:2(9Z,12Z) and FA 18:2(9E,11E) were obtained from SigmaAldrich (St. Louis, MO). FA 18:2(9Z,11Z), methyl ester standards of FA18:2(10E,12Z), and 18:3(9Z,11E,13Z) were purchased from Matreya Inc. (Pleasant Gap, PA, USA). HPLC grade acetone, methanol, and water were purchased from Fisher Scientific Company (Ottawa, ON, Canada). CLA dietary supplements were obtained commercially (1500 mg soft gel containing 70-84% of CLAs, NOW Foods, Bloomingdale, IL, USA). Solutions of lipid standards and CLA dietary supplements were prepared at 1.0~10.0 μM in D6acetone/water/MeOH (v/v/v, 68/22/10) solvent system. Ammonia (1%), LiCl (1 mM), or 100 μM AgNO3 was added to the sample solution for enhancing electrospray ionization (ESI) in negative or positive mode. The PB reaction. A flow microreactor was constructed from UV transmitting fused silica capillary (363 µm OD, 100 μm ID; Polymicro Technologies/Molex; Phoenix, AZ, USA). The end of the flow path was connected to an infusing ESI source for online analysis or the reaction solution was pre-collected in a nanoESI tip for offline analysis. A low-pressure mercury lamp (BHK, Inc.; Ontario, CA) with emission band centered at 254 nm was used for initiating the PB reaction. The UV exposure time was tunable from 0 - 40 s and optimized for each reaction.

A schematic representation of the reaction setup is shown in Figure S1, Supporting Information. Mass Spectrometry. A TIMS-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) was employed for acquiring high resolution MS1, MS/MS and IMS data. TIMS analyzer was calibrated using reduced mobility (K0) of tune mix from the relationship: K0 ∝ 1/V, where V is the elution voltage. The settings for the mass spectrometer were as follows: ESI voltage of −3200 V for negative ion mode and 3500 V for positive ion mode, capillary temperature of 200 °C, dry gas at 3.5 L/min, and nebulizer gas at 0.3 bar. Spectra were acquired with a time resolution of 1 s when TIMS mode was off. When TIMS mode was on, the parameters for TIMS were as following: negative ion mode, 1/K0: 0.78-0.88 ramp time: 359 ms, accumulation time: 1 ms; positive ion mode, 1/K0: 0.82-0.90, ramp time: 470 ms, accumulation time: 1 ms. RESULTS AND DISCUSSION Regio-selective PB reaction for CFAs. Acetone-water (v/v, 50/50) binary solvent system has been effective for performing the PB reaction of different classes of unsaturated lipids containing one or multiple non-conjugated C=Cs.23-24 As shown in Figure 1a, the PB products of linoleic acid (LA, FA 18:2(9Z,12Z)) are clearly detected at m/z 337.3 ([PBLA-H]-) with minimum side reactions (ESI-MS in negative ion mode). This solvent system, however, led to prominent side reactions for conjugated FA 18:2(10E,12Z), producing PB products with conversion yield less than 10% (Figure 1b). Accurate mass measurement and MS/MS revealed that one of the most abundant side products at m/z 295.2 was oxidized CLA ([CLA + O - H]-), while others (m/z 395.3, 393.3, 379.3, 367.3) likely resulted from radical chain reactions or complex photochemical reactions. Elemental compositions of these products are shown in Table S1, Supporting Information. It is known that



conjugated dienes are more prone to peroxidation than isolated C=Cs owing to forming delocalized radical intermediates.43 This aspect may account for dominant oxidation observed in Figure 1b relative to non-conjugated ones even though all solutions were purged by nitrogen gas prior to UV irradiation. Obviously, the PB reaction condition needed to be optimized by reducing side reactions. We found that an addition of 10% methanol to acetone-water (v/v, 75/25) effectively reduced a majority of side reaction products (Figures S2-S3, Supporting Information). Furthermore, use of D6-acetone as the PB reagent provided relatively clean reaction spectrum (Figure 1c). Due to introducing 6 Da mass shift to the PB products, chemical interferences to the PB products was reduced, resulting in cleaner PB-MS/MS spectra. Therefore, D6-acetone was used as the PB reagent in all later studies.

Figure 2. (a) EIMs of [PBM + Li]+ (m/z 351.3) produced from CLA 18:2(10E,12Z), CE = 10 eV. (b) EIMs corresponding to remaining [PBM + Li]+ (m/z 351.3, black trace), 12FO (m/z 251.1, blue trace), and 10FA (m/z 193.1, red trace), CE = 35 eV. Extracted MS/MS spectra from different 1/K0 ranges: (c) 0.960-0.990, (d) 0.9330.955, and (e) 0.905-0.933, CE = 35 eV.

The progress of the PB reaction of CLA 18:2(10E,12Z) was investigated by monitoring relative ion abundances of intact CLA (m/z 279.2) and its PB products (m/z 343.3) as a function of UV irradiation time using FA 13:0 as internal standard. As shown in Figure 1d, formation of the PB products reaches a plateau around 8-second UV exposure, while the abundance of

intact CLA keeps dropping due to competing sides reactions. Although the conversion of CLA to the PB products is low to moderate (10-20%), contribution from side reactions is significantly reduced by choosing shorter reaction time, i.e. less than 8 s. This aspect leads to reduced chemical interference and more stable PB product ion signal, which are preferable for mixture analysis. Besides differences in the PB reaction phenomena between conjugated and non-conjugated C=C systems, their PB products also showed distinct dissociation behavior under low energy CID. It is worth noting that two regio-isomers of oxetane ring are formed for each C=C in non-conjugated system because acetone can add onto a C=C with oxygen pointed toward the carboxylic acid or away from it (Scheme 1a). Rupture of the oxetane ring forms fragment ion containing an aldehyde (FA) at the cleavage site from one regio-isomer and an olefin fragment (FO) from the other. Typically, there is no regio-preference upon acetone addition to an isolated C=C, leading to detection of one pair of C=C diagnostic ions for each C=C. For instance, four diagnostic ions are detected from PB-MS/MS of LA 18:2 (9Z,12Z), viz. 9FA/9FO, 12FA/12FO, in Figure 1e. For simplicity in annotation, the superscript, 9 or 12, denotes the cleavage site on the fatty acyl chain, corresponding to each C=C location, i.e., 9 and 12, respectively. PB-MS/MS of CLA 18:2(10E,12Z), however, only produces one fragment per C=C. That is, only 10FA (m/z 185.1) and 12FO (m/z 243.2) are present, while 10FO (m/z 217.2) and 12FA (m/z 211.1) are missing (Figure 1f). Notably, these two diagnostic ions are formed about five-times more abundantly than LA 18:2 (9Z,12Z), the non-conjugated analogue. Same types of fragment ions are observed from PB-MS/MS of lithium adduct ions of CLA 18:2(10E,12Z) ([PBCLA + Li]+, m/z 351.3, Figure S4, Supporting Information). These results are indicative of forming oxetane ring in a regio-selective fashion for conjugated C=C systems. In order to gather more evidence on the formation of regioisomers, the PB products of CLA 18:2(10E, 12Z) were subjected to TIMS separations. When the PB products were analyzed as deprotonated ions ([PBM - H]-, m/z 343.3), the two suspected regio-isomers products, each leading to the formation of 10FA (m/z 185.1) and 12FO (m/z 243.2) upon CID, were not resolved by TIMS (Figure S5a,b, Supporting Information). Metal ions form coordination with carboxylic acid head group and sometimes with C=C in lipids.44 This property leads to tighter distribution in CCS values of lipid metal adduct ions, thus better separations of lipid isomers by IMS.17, 45 Although Ag+ can greatly enhance separations of CLA isomers by TIMS as discussed in the later section, it was experimentally difficult to detect silver adduct ions of the corresponding PB reaction products (Figure S5c, Supporting Information). Li+ was found to provide reasonable ion signal of PB products ([PBM+ Li]+) as well as improved separations in TIMS, likely resulting from multiple interactions between Li+, oxetane and carboxylic adic group. Figure 2a and 2b compare extracted ion mobilograms (EIMs) of the lithiated PB products ([PBM+Li]+, m/z 351.3) of CLA18:2(10E,12Z) at low (10 eV, to minimize fragmentation) and high (35 eV) CID energies. Clearly, the PB products are a mixture of several isomers. CID of the early-eluted doublets (1/K0: 0.960-0.990) produced C=C diagnostic ions at m/z 193.1, corresponding to 10FA. This fragment ion can only result from cleavage of the oxetane ring regio-isomer shown in the inset of Figure 2c (D6-acetone added on 10 C=C, with the dimethyl


Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry group pointed away from carboxylic). The mobility doublets should arise from partially separated cis- and trans-stereo isomers of the oxetane ring; that is the two alkyl substitutions are on the same or opposite side of the ring. The two diastereomers produce identical fragment ion, 10FA, and thus cannot be distinguished by PB-MS/MS. However, by coupling PB-MS/MS with TIMS, diastereomers can be clearly detected, without requiring laborious separation procedures. The peak eluted in the middle (1/K0: 0.933-0.955) was difficult to fragment under 35 eV collision energy (Figure 2d), likely consisting of a six-membered ring structure. This side reaction product may be formed from Diels-Alder reaction or ring closure from a diradical-intermediate (Scheme S1, Supporting Information). The last eluting doublet (1/K0: 0.905-0.933) produced a prominent olefin fragment (12FO) at m/z 251.1. This fragment ion can only result from acetone adducts at 12 C=C, with oxygen pointing to the methyl end (Figure 2e). Identification of these PB product isomers was further supported by reverse phase liquid chromatograph RPLCMS/MS analyses (data provided in Figure S6, Supporting Information). The above set of results supports that the PB reaction is regioselective for conjugated C=Cs. The same regio-selectivity was reported for the PB reaction of penta-1,3-diene and alkyl ketones by Funke et al., in which the formation of two dominant oxetane regio-isomers was verified by GC-MS. Such selectivity was rationalized to be driven by the stability of radical intermediates (Scheme 1b). That is, the carbonyl oxygen in acetone prefers to attacking terminal carbon atoms in the diene to form resonance-stabilized allylic radical.33 This process is followed by a quick radical-radical recombination, forming oxetane ring products with regio-specificity. Consequently, only one type of C=C diagnostic ion (either FA or FO) is produced from CID at each C=C location. Moreover, FA is only generated from the C=C closer to the carboxylic head, while FO is formed at the methyl end C=C. Mass difference (Δm) between the two fragments is determined by the total number (n) of the conjugated C=C bonds and satisfies the following relationship: FO - FA = C2n+1H2n-2D6-O when using D6-acetone as the PB reagent. We further examined PB-MS/MS of several other CLAs as free acid and in the form of fatty acid methyl ester (FAME). In the case of FAME 18:2(10E,12Z) (Figure 3a), regio-specific C=C diagnostic ions, 10FA and 12FO, were formed abundantly at m/z 207.2 and 265.3, respectively. Figure 3b shows PB-MS/MS spectrum of FAME 18:3(9Z,11E,13Z) ([PBM + Li]+, m/z 363.3). Only two diagnostic ions from terminal C=Cs are detected, viz. 9F + 13F A at m/z 193.1 (C10H18LiO3 ) and O at m/z 277.3 (C17H22D6LiO2+), consistent with the proposed PB-MS/MS pathways shown in Scheme 1b. Moreover, mass difference between 9FA and 13FO, allows determination of n = 3 from solving the equation: FO - FA = C2n+1H2n-2D6-O; thus, one can confidently determine a conjugated triene structure in this lipid. Identification and Quantification of CLA Isomers. PBMS/MS of conjugated C=C systems is capable of distinguishing CFA and PUFA isomers of different C=C locations. Figure 3c shows PB-MS/MS spectrum of an equal molar mixture (3.3 μM each) containing three FA 18:2 standards: CLA 18:2(9Z,11E), CLA 18:2(10E,12Z), and LA 18:2(9Z,12Z). Based on the PBMS/MS pathways summarized in Scheme 1, unique fragment ions of each isomeric species are as follows: 11FO (m/z 229.2)

from CLA 18:2(9Z,11E), 10FA (m/z 185.1) from CLA 18:2(10E,12Z), and 9FO/12FA (m/z 203.2/211.1) from LA 18:2(9Z,12Z). A comparison to ions detected in Figure 3c therefore allows facile identification of each isomers. The unique diagnostic ions 9FO/ 12FA belonging to LA 18:2(9Z,12Z) are detected at much lower abundances than that of CLA isomers, consistent with the PB-MS/MS behavior described earlier (Figure 1e, f). Note that 9FA (m/z 171.1) results from CLA18:2(9Z,11E) and LA (9Z,12Z); while 12FO (m/z 243.2) derives from CLA 18:2(10E,12Z) and LA (9Z,12Z). These two ions thus cannot be used for independent identification but help confirming the structural assignments.

Figure 3. PB-MS/MS of (a) FAME 18:2(10E,12Z) ( [PBM + Li], m/z 365.4) and (b) FAME 18:3(9Z,11E,13Z) ([PBM + Li]+, m/z 363.3). (c) PB-MS/MS of an equal molar mixture of three FA 18:2 isomers (3.3 μM each): CLA 18:2(9Z,11E), CLA 18:2(10E,12Z), and LA 18:2(9Z,12Z) ([PBM - H]-, m/z 343.3). (d) Calibration curve of CLA 18:2(9Z,11E) by plotting % abundance of C=C diagnostic ions vs. % molar concentration from PB-MS/MS of binary mixture of CLA 18:2(9Z,11E) and 18:2(10E,12Z), with total concentration kept at 10.0 μM.

PB-MS/MS also facilitates quantitation of individual CFA positional isomer from mixtures. Using a mock mixture of CLA



18:2(9Z,11E) and CLA 18:2(10E,12Z) as an example, a series of solution containing different molar fractions of CLA 18:2(9Z,11E) were subjected to PB-MS/MS. First, relative diagnostic ion abundances of CLA 18:2(9Z,11E), % (I(9FA) + I(11FO)), were calculated from using the summed diagnostic ion abundances from both isomers as denominator. They were plotted against their corresponding concentrations (% CLA 18:2(9Z,11E)) in the binary mixture (Figure 3d). Such calibration curve showed excellent linearity (R2 = 0.999). Based on detection of diagnostic ions, limit of identification at 1 nM was achieved (Figure S7, Supporting Information). These results are comparable to conventional GC/MS analysis;46 however, analysis time is significantly shortened from hours to minutes.

Figure 4. EIMs of deprotonated ions ([M – H]-, m/z 279.2) derived from (a) CLA 18:2(9Z,11Z), (b) CLA 18:2(9E,11E), (c) CLA 18:2(9Z,11E), (and d) CLA 18:2(10E,12Z) when UV is off (black trace) and on (red trace). EIMs of silver adduct ions ([CLA + 109Ag]+, m/z 389.1.e) derived from (e) CLA 18:2(9Z,11Z), (f) CLA 18:2(9E,11E), (g) CLA 18:2(9Z,11E), and (h) CLA 18:2(10E,12Z).

Analysis of C=C Geometry by TIMS. One limitation of PBMS/MS is that it cannot distinguish or determine the geometry of C=C bonds due to fast C=C isomerization. Therefore, TIMSMS was tested for separating four CLA standards consisting of different combinations of C=C geometry. Because CLA 18:2(9E,11Z) was not commercially available, CLA 18:2(10E,12Z) was employed as a model for E,Z-configuration, while all the rest CLAs have C=Cs at 9 and 11. TIMS measurements showed that deprotonated CLAs ([M-H]-) having Z,E or E,Z configurations were more compact than their Z,Z or E,E isomers. CLA 18:2(9Z,11E) (1/K0: 0.840) and CLA 18:2(10E,12Z) (1/K0: 0.837) thus could be resolved from CLA 18:2(9E,11E) (1/K0: 0.849) and CLA 18:2(9Z,11Z) (1/K0: 0.847) (Figure 4a-c). However, separations between Z,E and E,Z or Z,Z and E,E are difficult for deprotonated ions. EIM of each deprotonated CLA species typically has a peak width in the range of 0.007-0.008 V·s/cm2 (full width at half-maximum). This requires a resolving power higher than 280 and 425 to separate deprotonated Z,E vs. E,Z isomers and Z,Z vs. E,E isomers, respectively, which is beyond the capacity of TIMS. Silver adduct ions ([M + Ag]+) of unsaturated lipids have been

shown to allow for better IMS separations as compared to deprotonated ([M - H]-) or alkali metal adduct ions (e.g. [M + Li]+).17, 45 Due to strong coordination between Ag+ and C=C, the overall shape of the silver-lipid ion complex becomes more dependent on the number, location, and geometry of C=Cs, thus resulting in larger difference in CCS values among isomers. Indeed, [M + Ag]+ afforded enhanced separations of the four CLA isomers (Figure 4e-h) with elution in the order of E,E-, Z,E-, Z,Z-, and E,Z-configurations. EIMs of six individual isomers of FA 18:2 and their mixtures are documented in Figures S8-S9, Supporting Information. These results suggest that TIMS-MS is effective in separating C=C geometric isomers in the form of silver adduct ions although careful calibration with CLA standards is required. Cis-Trans Isomerization of C=C of CLAs in the PB Reaction. It is well documented that during the PB reaction C=C isomerizes due to energy transfer from excited ketones or from addition-elimination of ketone to the C=C (Scheme 1b).3233 This process can now be visualized by online coupling the PB reaction with TIMS-MS. As shown in Figure 4a, before the PB reaction, only a single mobility peak at 1/K0 0.847 is observed for [M-H]- ions of CLA 18:2(9Z,11Z). Upon exposure of the same solution to 254 nm UV irradiation, a second peak appeared at 1/K0 0.839, likely consisting of CLA 18:2(9Z,11E) and/or CLA18:2(9E,11Z). Besides, the original mobility peak is slightly up-shifted, which may be due to formation of CLA 18:2(9E,11E) isomer. Other three CLA standards showed consistent C=C isomerization phenomenon after UV exposure (red traces in Figure 4). We monitored isomerization process by varying the online UV exposure time from 1 s to 8 s. No significant changes of the position or peak area of the mobility peaks were observed, consistent with fast kinetics reported for photo-chemical isomerization.47 Because the four CLA isomers produce identical mobility peak patterns after UV exposure, there is no obvious preference in forming a specific geometry of conjugated C=Cs. Another useful aspect for identification is that by comparing IMS peak positions before and after UV exposure, one can quickly determine if the CLA has Z,Z or E,E combinations (low mobility, larger 1/K0) vs. Z,E or E,Z combination (high mobility, smaller 1/K0).

Figure 5. (a) PB-MS/MS spectrum of CLA mixture ([PBM - H]-, m/z 343.3) from supplements (1 g/mL) and (b) EIM of [M +


Page 6 of 9

Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 109Ag]+

(m/z 389.1) of CLA mixture in methanol with 100 μM AgNO3 added.

Details for optimization of the PB reaction, characterization of side products, and TIMS of six CLA isomers as described in the text (PDF).

Identification of CLAs from supplements. A combined PBMS/MS and TIMS-MS approach was developed and applied for quantitative analysis of CLA isomers in a commercial CLA supplement. In this approach, PB-MS/MS of deprotonated CLA ions should be used because it is more sensitive than PBMS/MS of lithium or silver adduct ions. However, TIMS-MS analysis should be performed with the silver adduct ions because it allows best separation of CLA isomers and thus rendering more confident identification of C=C geometries. A commercially obtained CLA supplement (1 g/mL) was dissolved in the optimized solvent system for performing online PB reaction. Figure 5a shows the PB-MS/MS spectrum of CLA supplement in negative ion mode. Two pairs of conjugated C=C diagnostic ions are formed abundantly at m/z 171.1, 229.2 and m/z 185.1, 243.2, indicating that it contains Δ9, 11 and Δ10, 12 CLA isomers. The silver adduct ions of CLAs were subjected to TIMS-MS analysis. As shown in Figure 5b, two IMS peaks at 1/K0 0.856 and 0.863 are observed. By comparing with that of the standards (black traces in Figure 4g and 4h), these two peaks are assigned as CLA 18:2(10E,12Z) and CLA 18:2(9Z,11E), respectively. Using the quantitation procedure described for Figure 3d, the CLA supplement was determined to contain (53.1 ± 1.1) % of FA 18:2(10E,12Z) and (46.9 ± 1.1) % of FA 18:2(9Z,11E).

AUTHOR INFORMATION

CONCLUSION We have developed a method for rapid and sensitive quantitation of CFA isomers by merging analytical advantages from PB-MS/MS and TIMS-MS. Driven by the stability of allylic radical intermediates, the PB reaction of conjugated C=Cs exhibits high regio-selectivity, different from that of isolated C=Cs. This selectivity allows formation of fragment ions unique to conjugated C=Cs under lower energy CID conditions as verified by standards of CFAs containing two or three C=Cs. Based on this principle, PB-MS/MS has been developed for identification and quantitation of CLA positional isomers from mixtures. Complementary to localization of C=Cs, TIMS-MS is capable of separating CLA isomers according to the geometry of C=C. For instance, deprotonated ions of CLAs are separated into two ion mobility groups (Z,E/E,Z vs. Z,Z/E,E), while separation is enhanced by analyzing their silver adduct ions. Consequently, double bond isomerization can be visualized during the PB reaction condition. Ion mobility fingerprints before and after C=C isomerization can also be used as an isomer-specific feature for identifying Z,Z and E,E isomers from Z,E and E,Z isomers without the requirement of careful mobility calibrations or comparisons to CLA standards. The above approach has been successfully applied for quantifying two major CLA isomers in food supplement. Given the largely simplified procedure and fast analysis speed, we believe the paired PB-MS/MS and TIMS-MS method should be potentially applicable to other conjugated molecules and more complex biological samples.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website.

Corresponding Author †* E-mail: [email protected] ORCID Yu Xia: 0000-0001-8694-9900

Author Contributions All authors have given approval to the final version of the manuscript.

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT Financial support from National Natural Science Foundation of China (No. 21722506 and No. 21621003) is greatly appreciated.

REFERENCES (1) Kishino, S.; Ogawa, J.; Yokozeki, K.; Shimizu, S. Appl. Microbiol. Biotechnol., 2009, 84, 87-97. (2) Chilliard, Y.; Glasser, F.; Ferlay, A.; Bernard, L.; Rouel, J.; Doreau, M. Eur. J. Lipid Sci. Tech., 2007, 109, 828-855. (3) Belury, M. A. Annu. Rev. Nutr., 2002, 22, 505-531. (4) Bhattacharya, A.; Banu, J.; Rahman, M.; Causey, J.; Fernandes, G. J. Nutr. Biochem., 2006, 17, 789-810. (5) Pariza, M. W.; Park, Y.; Cook, M. E. Prog. Lipid Res., 2001, 40, 283-298. (6) Pariza, M. W.; Park, Y.; Cook, M. E. Proc. Soc. Exp. Biol. Med., 2000, 223, 8-13. (7) Christie, W. W. Lipids, 1998, 33, 343-353. (8) Sehat, N.; Yurawecz, M. P.; Roach, J. A. G.; Mossoba, M. M.; Kramer, J. K. G.; Ku, Y. Lipids, 1998, 33, 217-221. (9) Roach, J. A. G.; Mossoba, M. M.; Yurawecz, M. P.; Kramer, J. K. G. Anal. Chim. Acta., 2002, 465, 207-226. (10) May, J. C.; McLean, J. A. Anal. Chem., 2015, 87, 1422-1436. (11) Bowman, A. P.; Abzalimov, R. R.; Shvartsburg, A. A. J. Am. Soc. Mass Spectrom., 2017, 28, 1552-1561. (12) Dong, L.; Shion, H.; Davis, R. G.; Terry-Penak, B.; CastroPerez, J.; van Breemen, R. B. Anal. Chem., 2010, 82, 9014-9021. (13) Kyle, J. E.; Zhang, X.; Weitz, K. K.; Monroe, M. E.; Ibrahim, Y. M.; Moore, R. J.; Cha, J.; Sun, X.; Lovelace, E. S.; Wagoner, J.; Polyak, S. J.; Metz, T. O.; Dey, S. K.; Smith, R. D.; Burnum-Johnson, K. E.; Baker, E. S. Analyst, 2016, 141, 1649-1659. (14) Poad, B. L. J.; Maccarone, A. T.; Yu, H.; Mitchell, T. W.; Saied, E. M.; Arenz, C.; Hornemann, T.; Bull, J. N.; Bieske, E. J.; Blanksby, S. J. Anal. Chem., 2018, 90, 5343-5351. (15) Deng, L.; Ibrahim, Y. M.; Hamid, A. M.; Garimella, S. V.; Webb, I. K.; Zheng, X.; Prost, S. A.; Sandoval, J. A.; Norheim, R. V.; Anderson, G. A.; Tolmachev, A. V.; Baker, E. S.; Smith, R. D. Anal. Chem., 2016, 88, 8957-8964. (16) Ahonen, L.; Fasciotti, M.; Gennas, G. B.; Kotiaho, T.; Daroda, R. J.; Eberlin, M.; Kostiainen, R. J. Chromatogr. A., 2013, 1310, 133137. (17) Groessl, M.; Graf, S.; Knochenmuss, R. Analyst, 2015, 140, 6904-6911. (18) Wojcik, R.; Webb, I. K.; Deng, L.; Garimella, S. V.; Prost, S. A.; Ibrahim, Y. M.; Baker, E. S.; Smith, R. D. Int. J. Mol. Sci., 2017, 18, 183-194. (19) Murphy, R. C. Tandem Mass Spectrometry of Lipids: Molecular Analysis of Complex Lipids. Royal Society of Chemistry: London, 2015. 1-280. (20) Higashi, T.; Takekawa, M.; Min, J. Z.; Toyo'oka, T. Anal. Bioanal. Chem., 2012, 403, 495-502. (21) Thomas, M. C.; Mitchell, T. W.; Harman, D. G.; Deeley, J. M.; Murphy, R. C.; Blanksby, S. J. Anal. Chem., 2007, 79, 5013-5022.



(22) Sun, C.; Black, B. A.; Zhao, Y. Y.; Ganzle, M. G.; Curtis, J. M. Anal. Chem., 2013, 85, 7345-7352. (23) Ma, X.; Xia, Y. Angew. Chem. Int. Ed., 2014, 53, 2592-2596. (24) Cajka, T.; Fiehn, O. Metabolomics, 2016, 12. (25) Xu, T. F.; Pi, Z. F.; Song, F. R.; Liu, S.; Liu, Z. Q. Anal. Chim. Acta., 2018, 1028, 32-44. (26) Murphy, R. C.; Okuno, T.; Johnson, C. A.; Barkleyte, R. M. Anal. Chem., 2017, 89, 8545-8553. (27) Bednarik, A.; Bolsker, S.; Soltwisch, J.; Dreisewerd, K. Angew. Chem. Int. Ed., 2018, 57, 12092-12096. (28) Esch, P.; Heiles, S. J. Am. Soc. Mass Spectrom., 2018, 29, 19711980. (29) Zhang, W.; Zhang, D.; Chen, Q.; Wu, J.; Ouyang, Z.; Xia, Y. Nat. Commun., 2019, 10, 79-87. (30) Ma, X.; Chong, L.; Tian, R.; Shi, R.; Hu, T. Y.; Ouyang, Z.; Xia, Y. Proc. Natl. Acad. Sci. USA, 2016, 113, 2573-2578. (31) Waldchen, F.; Becher, S.; Esch, P.; Kompauer, M.; Heiles, S. Analyst, 2017, 142, 4744-4755. (32) Velosa, A. C.; Baader, W. J.; Stevani, C. V.; Mano, C. M.; Bechara, E. J. H. Chem. Res. Toxicol., 2007, 20, 1162-1169. (33) Fréneau, M.; Hoffmann, N. J. Photoch. Photobio. C, 2017, 33, 83-108. (34) Pham, H. T.; Maccarone, A. T.; Campbell, J. L.; Mitchell, T. W.; Blanksby, S. J. J. Am. Soc. Mass Spectrom., 2013, 24, 286-296. (35) Williams, P. E.; Klein, D. R.; Greer, S. M.; Brodbelt, J. S. J. Am. Chem. Soc., 2017, 139, 15681-15690. (36) Baba, T.; Campbell, J. L.; Le Blanc, J. C. Y.; Baker, P. R. S. Anal. Chem., 2017, 89, 7307-7315.

(37) Randolph, C. E.; Foreman, D. J.; Betancourt, S. K.; Blanksby, S. J.; McLuckey, S. A. Anal. Chem., 2018, 90, 12861-12869. (38) Alves, S. P.; Tyburczy, C.; Lawrence, P.; Bessa, R. J.; Brenna, J. T. Rapid Commun. Mass Spectrom., 2011, 25, 1933-1941. (39) Poad, B. L.; Pham, H. T.; Thomas, M. C.; Nealon, J. R.; Campbell, J. L.; Mitchell, T. W.; Blanksby, S. J. J. Am. Soc. Mass Spectrom., 2010, 21, 1989-1999. (40) Funke, C. W.; Cerfontain, H. J. Chem. Soc., Perkin Trans. 2, 1976, 1902-1908. (41) Michelmann, K.; Silveira, J. A.; Ridgeway, M. E.; Park, M. A. J. Am. Soc. Mass Spectrom., 2015, 26, 14-24. (42) Liebisch, G.; Vizcaino, J. A.; Kofeler, H.; Trotzmuller, M.; Griffiths, W. J.; Schmitz, G.; Spener, F.; Wakelam, M. J. O. J. Lipid Res., 2013, 54, 1523-1530. (43) Gu, J.; Wu, W.; Danovich, D.; Hoffmann, R.; Tsuji, Y.; Shaik, S. J. Am. Chem. Soc., 2017, 139, 9302-9316. (44) Grossert, J. S.; Cubero Herrera, L.; Ramaley, L.; Melanson, J. E. J. Am. Soc. Mass Spectrom., 2014, 25, 1421-1440. (45) Šala, M.; Lísa, M.; Campbell, J. L.; Holčapek, M. Rapid Commun. Mass Spectrom., 2016, 30, 256-264. (46) Johnson, D. W.; Beckman, K.; Fellenberg, A. J.; Robinson, B. S.; Poulos, A. Lipids, 1992, 27, 177-180. (47) Bull, J. N.; Scholz, M. S.; Coughlan, N. J.; Kawai, A.; Bieske, E. J. Anal. Chem., 2016, 88, 11978-11981.


Page 8 of 9

Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table of Contents (TOC)


9

Analysis of Conjugated Fatty Acid Isomers by the ... - ACS Publications

Recommend Documents