Sequential Hybrid Three-Dimensional Gas Chromatography with

7 days ago - The described integrated system can be used in a number of modes, but one useful approach is to target specific classes of compounds for ...
1 downloads 10 Views 1MB Size
Subscriber access provided by Kaohsiung Medical University

Sequential Hybrid Three-Dimensional Gas Chromatography with Accurate Mass Spectrometry: A Novel Tool for HighResolution Characterization of Multicomponent Samples Dandan Yan, Yong Foo Wong, Simon P. Whittock, Anthony Koutoulis, Robert A. Shellie, and Philip John Marriott Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00142 • Publication Date (Web): 26 Mar 2018 Downloaded from http://pubs.acs.org on March 26, 2018

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

Sequential Hybrid Three-Dimensional Gas Chromatography with Accurate Mass Spectrometry: A Novel Tool for HighResolution Characterization of Multicomponent Samples DanDan Yan,†,‡ Yong Foo Wong,ξ Simon P. Whittock,†,§ Anthony Koutoulis,# Robert A. Shellie,†,Ʈ,ǁ Philip J. Marriott,‡,* †

Australian Centre for Research on Separation Science, School of Natural Sciences, University of Tasmania, Hobart, TAS 7001, Australia ‡Australian Centre for Research on Separation Science, School of Chemistry, Monash University, Wellington Road, VIC 3800, Australia ξSchool of Chemical Sciences, Universiti Sains Malaysia,11800 Penang, Malaysia §Hop Products Australia, 446 Elizabeth St, Hobart, TAS 7000, Australia #School of Natural Sciences, University of Tasmania, Hobart, TAS 7001, Australia ƮTrajan Scientific and Medical, 7 Argent Place Ringwood 3154 VIC, Australia ǁSchool of Science, RMIT University, GPO Box 2476, Melbourne Victoria 3001, Australia

ABSTRACT: A novel sequential three-dimensional gas chromatography – high resolution time-of-flight mass spectrometry (3D GC−accTOFMS) approach for profiling secondary metabolites in complex plant extracts is described. This integrated system incorporates a non-polar first-dimension (1Dnp) separation step, prior to microfluidic heart-cut (H/C) of a targeted region(s) to a cryogenic trapping device, directly followed by rapid re-injection of trapped solute into a polar second-dimension (2DPEG) column for multidimensional separation (GCnp–GCPEG). For additional separation, effluent from 2DPEG can then be modulated according to a comprehensive 2D GC process (GC×GC), using an ionic liquid phase as third-dimension (3DIL) column, to produce a sequential GCnp– GCPEG×GCIL separation. Thus unresolved or poorly resolved components, or regions that require further separation, can be precisely selected and rapidly transferred for additional separation on 2D or 3D columns, based on the greater separation realized by these steps. The described integrated system can be used in a number of modes, but one useful approach is to target specific classes of compounds for improved resolution. This is demonstrated through the separation and detection of oxygenated sesquiterpenes in hop (Humulus lupulus L.) essential oil and agarwood (Aquilaria malaccensis) oleoresin. Improved resolution and peak capacity was illustrated through the progressive comparison of tentatively identified components for GCnp–GCPEG and GCnp–GCPEG×GCIL methods. Relative standard deviations of intra-day retentions (1tR, 2tR, and 3tR) and peak areas of ≤ 0.01%, 0.07%, 0.71% and 7.5%, respectively, were achieved. This analytical approach comprising three GC column selectivities, hyphenated with high resolution TOFMS detection should be a valuable adjunct for improved characterization of complex plant samples, particularly in the area of plant metabolomics.

The measurement and identification of plant secondary metabolites has always been a major challenge to plant analysts, with suggestion that approximately 200,000 metabolites exist in the plant kingdom.1 Several analytical methods have been suggested to offer improved coverage for the analysis of complex mixtures of plant secondary compounds.2 Hyphenated chromatographic–MS technologies have found great utility in addressing analytically demanding separations, particularly in the “omics” field, especially with respect to the ability to provide simultaneous separation, detection, and a measure of identification of metabolites.3,4 In complex plant extracts, effective molecular separation approaches are required to detect and propose the identity of trace or major components from a multitude of other matrix components. In particular, gas chromatography with mass spectrometry (GC–MS) facilitates these objectives for volatile or semi-volatile metabolites. GC–MS provides robust quantification of a multitude of volatile or semi-volatile metabolites in a single plant extract, allowing corresponding interpretation of the central pathways of secondary metabolism.5

However, metabolic compositions in natural plant extracts are of such diversity and complexity, not always anticipated by analysts, that peak capacity of a single-dimension (1D) separation will often be exceeded. The introduction of comprehensive two-dimensional GC (GC×GC) technology and the higher peak capacity that it offers, gives rise to significantly improved separations for complex samples.6-8 This technology enhancement is suited to hyphenation with MS for the untargeted analysis of complex matrices, providing valuable expansion of metabolic coverage in comparison with 1D GC–MS methods.9-12 Despite obtaining a peak capacity approximately the multiplication of that of the participating columns, GC×GC is still unable to provide complete separation (ideally as fully resolved single components) of all secondary metabolites, which comprise extensive chemical diversity and a variety of physicochemical properties. Development of higher order GC designs with expanded resolving power, able to cater for metabolites of widely different abundance, is desirable. Several recent studies have described advanced multidimensional GC (MDGC) designs comprising > two dimensions of GC separation, which permit a range of

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

new separation procedures.13-17 These include the recently proposed comprehensive three-dimensional GC, hybrid MDGC, and enantioselective four-dimensional dynamic GC as examples of higher order GC designs that have moved beyond conceptualization and implementation, to further enhance separation selectivity in ways not possible in 1D and 2D GC techniques. All these designs are directed to the goals of either providing improved analysis of either global (i.e., expanding the metabolite coverage for metabolomics analysis) or target detection and/or measurement of analytes. The added chemical selectivity available to a sequential 3D GC separation system based on available column types, may be explored by focusing on critical regions of interest in the first column separation, to improve analytical resolution, ideally accompanied by improved identification. Here, we describe a novel and functional automated threedimensional GC arrangement that incorporates aspects of 1D GC, heart-cut MDGC (GC–GC), and GC×GC with accurate mass time-of-flight MS detection (accTOFMS) in the one integrated instrument. The proposed system incorporates advantages of both GC–GC and GC×GC, by incorporating a first-dimension (1D) separation step, prior to microfluidic heart-cutting of a targeted region(s) into a second-dimension (2D) column for multidimensional separation (GC–GC). If further separation is required, the effluent from 2D can be further modulated to produce an added GC×GC separation. In practice, it operates with preliminary 1D separation of a complex matrix using 1D GC, followed by “extraction” of target regions into a second (GC–GC) and/or a third column (GC– GC×GC) for additional successive separations; i.e., operating as sequential GC–GC×GC. The analytical performance, applicability, and benefits of utilizing three dimensions of chemical selectivities with accTOFMS detection for high resolution separation and detection of oxygenated sesquiterpene components in hop (Humulus lupulus L.) essential oil and agarwood (Aquilaria malaccensis) oleoresin are demonstrated and discussed. This integrated three-dimension GC-accTOFMS system can serve as alternative contemporary tools for high resolution characterization of complex volatile and semivolatile plant derived mixtures. EXPERIMENTAL SECTION Chemicals and reagents. Analytical reagent grade dichloromethane was purchased from Merck (Darmstadt, Germany). A series of n-alkanes (C7–C30) was purchased from Sigma-Aldrich (St. Louis, MO). Samples. Hop (H. lupulus L.) cone samples were provided by Hop Products Australia (Hobart, Australia) and essential oil was obtained using hydro-distillation. Agarwood (A. malaccensis) oleoresin was provided by Green Agro Agarwood Products Sdn. Bhd. (Penang, Malaysia). Samples were stored at −20 ˚C when not in use. A 1:100 (v/v) dilution of the samples in dichloromethane were prepared prior to GC analyses. GC×GC–QMS system: GC×GC–quadrupole(Q)MS analyses were conducted on an Agilent 7890A GC system coupled with a 5975c QMS (MSD; Agilent Technologies), retrofitted with an Everest model longitudinal modulation cryogenic system (LMCS; Chromatography Concepts Ltd., Doncaster, Australia). Chromatographic separation was performed using a SUPELCOWAX®10 column (Supelco, Bellefonte, PA) of dimensions 30 m × 0.25 mm I.D. × 0.25 µm film thickness (df) as 1D column, and a SLB-IL59 (Supelco) or BPX5 (Trajan

Scientific and Medical, Ringwood, Australia) as 2D column (1.4 m × 0.1 mm I.D. × 0.08 µm df) connected by a Restek deactivated PressFit union (Restek Corporation, Bellefonte, PA). Modulation was performed at −10 ˚C with modulation period (PM) of 9.0 s. Chromatographic conditions were: oven temperature program, 50 ˚C (hold 1 min) heated at 3 ˚C/min to 245 ˚C (hold 20 min); injector temperature, 200 ˚C; carrier gas, helium at a constant flow rate of 1.0 mL/min; injection volume, 1 µL using a split ratio of 5:1. The MS conditions were: transfer line, 0.5 m × 0.1 mm I.D.; temperature, 250 ˚C; ion source temperature, 230 ˚C; electron ionization mode at 70 eV; mass scan range of 70-300 Da at 12,500 u.s−1 (giving an acquisition rate of 27.15 scans s−1).

Figure 1. Instrument schematic of sequential GCnp–GCPEG×GCIL– accTOFMS configuration. 1D: first, 2D: second, and 3D: third dimension columns; DS, Deans switch; EPC, electronic pressure control; SV, switching valve; CT, cryotrap; FID, flame ionization detector; LMCS, longitudinal modulation cryogenic system; C1, PressFit union; C2, capillary union; DFS, deactivated fused-silica transfer line; QTOFMS, quadrupole-time-of-flight mass spectrometer.

Sequential GCnp–GCPEG×GCIL–accTOFMS system: Integrated multidimensional three-dimension experiments were conducted on an Agilent 7890A GC system coupled with a 7200 series quadrupole-time-of-flight MS (QTOFMS; Agilent Technologies, Mulgrave, Australia), retrofitted with an Everest model LMCS; equipped with a flame ionization detector (FID) and PAL3 Auto Sampler (CTC Analytics AG, Zwingen, Switzerland). A DB-5ms Ultra Inert non-polar phase column (30 m × 0.25 mm I.D. × 0.25 µm df; Agilent Technologies) was used as 1Dnp column; a SUPELCOWAX®10 polyethylene glycol polar phase (30 m × 0.25 mm I.D. × 0.25 µm df; Supelco) as 2DPEG phase; and SLB-IL59 polar phase column (1.4 m × 0.1 mm I.D. × 0.08 µm df; Supelco) as the 3DIL column. A microfluidic Deans switch (DS; Agilent, part no. G2855B) for heart-cut (H/C) effluent switching was used to interface the end of 1Dnp to the start of 2DPEG with a deactivated fused-silica column (DFS; 1.8 m × 0.1 mm I.D.) as transfer line to a FID. H/C switching of effluent flow from 1Dnp to either 2DPEG or the FID was controlled through the events option in MassHunter software via a three channel auxiliary electronic pressure control (EPC) module (G1570A; Agilent Technologies). GC×GC experiments were performed by modulating the effluent from 2 DPEG into 3DIL, with modulation period (PM) of 5.0 s or 9.0 s, and modulation temperature (TM) of −10 ˚C. The larger PM setting was used to check for wrap-around. A deactivated PressFit union was used to connect the two columns. SilTite metal ferrules (Trajan Scientific and Medical) were used to connect columns to the DS device. An SGE liquid CO2 cryo-

ACS Paragon Plus Environment

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

Analytical Chemistry genic trapping device (CT; Trajan) was positioned at the beginning of 2DPEG, to trap and re-focus the H/C solutes from 1 Dnp to 2DPEG. A schematic of the system used here is given in Figure 1, for 1D GC mode (without heart-cutting, indicated by green arrow) or for transfer of heart-cuts to the 2DPEG column (indicated by red arrow), respectively. Programmed temperature gradient analysis from 50 ˚C (hold 1 min) to 245 ˚C (hold 20 min) at 3 ˚C/min was used for 1D GC experiments. To perform H/C MDGC operation, a 2-step sequence was programmed for the H/C event. The required switching time was entered into the events software in accordance with 1D GC-FID analysis for the target region. Cooling of the SGE CT was initiated 4 min prior to transfer of the target region H/C, to cryofocus H/C solutes. After H/C operation, the oven temperature (T) was immediately reduced to 50 ˚C (hold 1 min) at 25 ˚C/min, and then to remobilize the trapped H/C fraction from CT into 2DPEG the CT coolant supply was stopped, and T programming was recommenced. Helium was used as carrier gas (99.999% purity). A column head pressure of 33 psi (constant flow rate of ca. 0.82 mL/min) was applied for the 1Dnp column, whilst the EPC for the DS was set at 28.5 psi (constant flow rate of ca. 1.05 mL/min) for the 2DPEG column. System balance for DS flow switching without leakage to the other channel is confirmed (Figure S1). Injector T was set at 200 ˚C, and 1 µL was injected at 5:1 split ratio. The FID (250 ˚C) was operated at a sampling frequency of 100 Hz. The outlet of the 3DIL column was connected to the QTOFMS via a DFS column (0.6 m × 0.1 mm I.D.). The MS transfer line T was set at 250 ˚C. Since QTOFMS was operated in total transfer of ion (TTI) mode through the quadrupole sector, this is referred to hereafter as accTOFMS. Ion source T, emission current and electron ionization voltage were set at 280 ˚C, 4.6 µA and 70 eV, respectively, with a mass range of 40-350 Da. The MS acquisition rate (the maximum available) was nominally 50 spectra/s (experimentally determined to be 47.5 Hz) with acquisition time of 20 ms/spectrum, which generates 235 transients/spectrum. This system represents a 3D system, because the cryogenic modulation process mimics conventional GC×GC, whereas the DS at the outlet of the first column operates as it would in a MDGC system, with a CT device at the inlet of the second column functioning to trap, refocus, then introduce the compounds eluting from 1Dnp to 2DPEG. The GCnp−GCPEG experiments were conducted using the same column configuration but without the modulation process. Although GCnp−GCPEG is a composite column arrangement, the short 3D IL column has negligible effect on the separation of solutes in 2D.

the PEG phase column in MDGC and GC×GC operations were acquired RESULTS AND DISCUSSION Conceptually, this integrated system design (GCnp– GCPEG×GCIL–accTOFMS) is based on the body of knowledge of GC×GC and heart-cut MDGC, supported with precise control of microfluidic switching, solute trapping and remobilization operations in a single GC system. To illustrate the necessity for higher orders of chemical selectivity for the phytocharacterization of complex plant samples, 1D GC separation was applied to two reasonably complex plant extracts (H. lupulus L. and A. malaccensis). Figures 2Ai and 3Ai depict the classical GC–FID analysis of hop and agarwood essential oils using the sequential 3D-GC system, with DS device in its “OFF” position to direct the 1D effluent to the FID detector. In the current example, oxygenated sesquiterpenes (OS) were the main region of interest, due to its important relevance to the specific and unique beer flavor impression, and also its acceptance as a key chemical identifier groups for different hop cultivars or agarwood grades.18-20 Due to the lack of sufficient phase selectivity and/or peak capacity, a significant overlap of metabolites was observed for the OS cluster (Figures 2Aii and 3Aii). On the basis of these observations, GCnp–GCPEG and GCnp–GCPEG×GCIL approaches were devised to provide improved analysis and/or measurement of metabolites of interests. The target regions are now shown to be better separated on the 2DPEG column (Figures 2Bi and 3Bi). This illustrates the inability of the single (1D) dimension separation (essentially based on dispersive interaction) to adequately resolve a large number of components with similar physicochemical characteristics. The elution range on the 2DPEG column is expanded by some 5-fold, and this contributes to the greater separation noted.

Agilent MassHunter ver. B.06.00 (Agilent Technologies) was used for data acquisition and processing. NIST (National Institute of Standards and Technology) 11 MS spectrum library was used for spectrum searching and matching. Acquired MassHunter data were exported in csv format for data display using OriginPro 8 SR2 software Version 8.0891 (Origin, Northampton, MA). 2D and 3D plots were generated by directly importing MassHunter files to GC Image ver. 2.5 (GC Image LLC, Lincoln, NE). The MS library match of > 700 is generally chosen for 1D GCMS to ensure positive identifications are not overlooked, noting that interfering ions of low abundance compounds reduce match quality. The same match criterion is used for MDGC and GC×GC. The accTOFMS was tuned every 3 analyses. Retention indices for

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

Figure 2. Sequential GCnp–GCPEG×GCIL–accTOFMS analysis of oxygenated sesquiterpenes in hop essential oil: (Ai) 1Dnp FID response, the region to be H/C is denoted by the dotted rectangle, with inset (Aii) an expansion of target region 47.8–49.8 min; (Bi) Heart-cut MDGC–accTOFMS analysis of target region 47.8–49.8 min with inset (Bii) the 1Dnp FID response showing 47.8–49.8 min H/C to 2DPEG; (C) GCnp–GCPEG×GCIL–accTOFMS analysis of target region 47.8–49.8 min; red dots are auto-generated by software for detected peaks. The numbering of peaks in (Bi) and (C) is given in Table S1 and Table 1, respectively.

Figure 3. Sequential GCnp–GCPEG×GCIL–accTOFMS analysis of oxygenated sesquiterpenes in A. malaccensis oleoresin: (Ai) 1Dnp FID response, the region to be H/C is denoted by the dotted rectangle, with inset (Aii) an expansion of the target region 54–59 min; (Bi) Heart-cut MDGC–accTOFMS analysis of target region 54–59 min with inset (Bii) the 1Dnp FID response showing 54–59 min H/C to 2DPEG; (C) GCnp–GCPEG×GCIL–accTOFMS analysis of target region 54–59 min; red dots are auto-generated by software for detected peaks. The numbering of peaks in (Bi) and (C) is given in Table S2 and Table S3, respectively.

For 1D chromatographic regions that exhibit unresolved or poorly resolved component(s), a microfluidic heart-cut sampling strategy was implemented to excise and transfer target compound(s) or region(s) for further separation in a polar 2 DPEG phase. A CT installed at the inlet of the 2DPEG column effectively retains the H/C fractions. Here, a second dedicated oven and/or pressure program is implemented in conjunction with the remobilization event (i.e., when liquid CO2 supply is switched off) of the trapped species, which provides further improved separation on the 2D polar column. The cryotrapping procedure also minimizes the broadening or distortion of peak shapes (2wh ≈ 6 s) arising from the 1D apolar separation. Figures 2Bi and 3Bi show the GCnp–GCPEG–accTOFMS analysis of hop and agarwood extracts, where selected OS regions were successfully H/C to 2DPEG (Figure 2Bii and 3Bii). Improved

ACS Paragon Plus Environment

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 separation was achieved as evidenced by the number of resolved peaks as a result of both additional phase selectivity and the expanded elution space (increased peak capacity) on the 2D column.

The number of detected components increased by approximately two-fold in comparison to 1D GC analysis. Nevertheless, a proportion of the peaks are still observed to suffer a certain degree of overlap, particularly for agarwood oleoresin (Figure 3Bi). This illustrates the inability of the 2D (i.e., insufficient peak capacity of GCnp–GCPEG) separation process to adequately resolve the number of OS with anticipated different physiochemical properties. On the basis of these observations, a subsequent cryogenic modulation experiment (GC×GC) of the 2D elution profiles was devised to extend the peak capacity, and, hence metabolite coverage. To further enhance chemical selectivity for regions that show considerable peak overlaps (Figures 2Bi and 3Bi), the effluent from 2DPEG was cryogenically modulated to produce a GC×GC separation. In this instance, suitable selection of the 3D stationary phase will be important, with an aim to provide three complementary separation modes, i.e., separations that provide differences in chemical selectivity for each dimension.

Figure 4. Result of the GCPEG×GCIL stage of the multiple-column combination for the analysis of oxygenated sesquiterpenes in hop essential oil using sequential GCnp–GCPEG×GCIL: with (A) SLBIL111; (B) SLB-IL82; and (C) SLB-IL59 as 3D column coupled with DB-5ms 1D and SUPELCOWAX®10 2D columns.

Table 1. Τentatively identified oxygenated sesquiterpenes in hop essential oil sample by sequential GCnp–GCPEG×GCIL system. No.

Compound name

m/z of significant ions (relative ion abundance)

Base ion M.A.a (ppm)

Match factor

RIexpb

RIrefc

1

Alloaromadendrene oxide

79.0539 (63.3), 91.0545 (100), 105.0706 (91.4), 119.0847 (62.6), 147.1177 (65.0), 163.1126 (59.9) [C6H7]+, [C7H7]+, [C8H9]+, [C9H11]+, [C11H15]+, [C11H15O]+

[C7H7]+; -3

721

1998

NAd

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

Page 6 of 9

2

Caryophyllene oxide

79.0541 (75.1), 91.0539 (100), 105.0700 (77.6), 163.1120 (70.9), 178.1363 (27.6) [C6H7]+, [C7H7]+, [C8H9]+, [C11H15O]+, [C12H18O]+

[C7H7]+; 3.59

722

2012

2001

3

Aromadendrene oxide

79.0545 (38.5), 91.0538 (100), 105.0698 (94.8), 163.1118 (83.6), 178.1362 (46.0) [C6H7]+, [C7H7]+, [C8H9]+, [C11H15O]+, [C12H18O]+

[C7H7]+; 4.69

731

2021

NA

4

Humulene epoxide II

67.0539 (100), 79.0553 (49.9), 96.0570 (65.5), 105.0703 (43.0), 123.0812 (15.4) [C5H7]+, [C6H7]+, [C6H8O]+, [C8H9]+, [C8H11O]+

[C5H7]+; 4.87

762

2045

2044

5

Guaiol

67.0538 (43.9), 79.0541 (63.0), 93.0699 (59.6), 107.0851 (100), 121.1013 (34.7), 163.1478 (57.5) [C5H7]+, [C6H7]+, [C7H9]+, [C8H11]+, [C9H13]+, [C12H19]+

[C8H11]+; 3.99

770

2059

2064

6

Elemol

67.0538 (74.2), 79.0537 (60.6), 91.0540 (81.4), 105.0697 (100), 119.0852 (47.9), 161.1327 (52.9) [C5H7]+, [C6H7]+, [C7H7]+, [C8H9]+, [C9H11]+, [C12H17]+

[C8H9]+; 1.68

781

2088

2090

9

α-Acorenol

67.0540 (68.2), 81.0698 (100), 93.0700 (68.9), 105.0700 (51.5), 189.1641 (45.1), 204.1868 (23.2) [C5H7]+, [C6H9]+, [C7H9]+, [C8H9]+, [C14H21]+, [C15H24]+

[C6H9]+; 0.95

799

2142

2127

10

γ-Eudesmol

91.0544 (55.9), 105.0700 (74.2), 133.1013 (48.4), 161.1325 (100), 189.1639 (63.4), 204.1868 (38.3) [C7H7]+, [C8H9]+, [C10H13]+, [C12H17]+, [C14H21]+, [C15H24]+

[C12H17]+; -0.14

854

2172

2168

11

Guai-1(10)-en-11-ol

59.0488 (54.3), 91.0540 (66.0), 105.0695 (61.0), 119.0851 (49.2), 161.1320 (100), 189.1628 (21.0) [C3H7O]+, [C7H7]+, [C8H9]+, [C9H11]+, [C12H17]+, [C14H21]+

[C12H17]+; 2.96

805

2182

2180

12

tau.-Cadinol

79.0547 (42.2), 91.0549 (60.0), 105.0700 (100), 119.0855 (44.4), 161.1324 (72.0) [C6H7]+, [C7H7]+, [C8H9]+, [C9H11]+, [C12H17]+

[C8H9]+; -1.17

768

2192

2189

13

Ledene oxide-(II)

77.0382 (55.9), 91.0541 (100), 105.0695 (88.2), 131.0851 (75.1), 187.1475 (67.1), 202.1710 (64.2) [C6H5]+, [C7H7]+, [C8H9]+, [C10H11]+, [C14H19]+, [C15H22]+

[C7H7]+; 1.39

753

2212

NA

14

Ledol

67.0540 (96.3), 81.0708 (48.5), 93.0703 (91.4), 107.0851 (100), 119.0850 (48.4), 161.1315 (42.3) [C5H7]+, [C6H9]+, [C7H9]+, [C8H11]+, [C9H11]+, [C12H17]+

[C8H11]+; 3.99

818

2036

2035

15

γ-Gurjunenepoxide-(2)

67.0543 (88.6), 79.0545 (100), 93.0704 (83.3), 107.0851 (69.4), 205.1594 (21.3) [C5H7]+, [C6H7]+, [C7H9]+, [C8H11]+, [C14H21O]+

[C6H7]+; -3.46

760

2059

NA

16

Cubenol

91.0535 (30.1), 105.0697 (70.5), 119.0852 (100), 161.1320 (43.2) [C7H7]+, [C8H9]+, [C9H11]+, [C12H17]+

[C9H11]+; 2.74

796

2069

2065

17

α-Muurolol

67.0543 (64.4), 79.0554 (76.6), 91.0540 (90.1), 105.0699 (100), 119.0852 (65), 161.1327 (69.7) [C5H7]+, [C6H7]+, [C7H7]+, [C8H9]+, [C9H11]+, [C12H17]+

[C8H9]+; -0.22

754

2152

2170

742

2302

NA

67.0541 (66.2), 79.0543 (84.6), 91.0541 (100), 105.0696 (60.8), 136.0876 (97.1) [C7H7]+; 1.39 + + + + + [C5H7] , [C6H7] , [C7H7] , [C8H9] , [C9H12O] a M.A., mass accuracy calculated from accurate mass of the proposed base peak ion mass. b RIexp, RI calculated from 2D retention time for the identified component. c RIref, reported RI values on polar wax phase for the stated compounds. d NA, reference retention index value not available for wax or equivalent column phase. Compounds not identifiable are not added to the table, based on criteria of poor match quality and/or incorrect retention index. 18

Cedr-8-en-15-ol

Recent studies indicate the applicability of phosphonium and imidazolium-based ionic liquid (IL) phases for the separation of plant secondary components.21-23 Hence, the suitability of phosphonium and imidazolium-based IL phases were investigated as the 3D column phase, for constituents that comprise of polar solutes (i.e. oxygenated sesquiterpene analogs). Figure 4 shows the GC×GC elution patterns of oxygenated sesquiterpene classes using three different IL phases (SLB-IL59, SLBIL82, and SLB-IL111) with increasing hydrogen bond basicity constants (HBC) and McReynold’s constants (MC). For the tested IL columns, the cation structure is the key to performance (as the same anion is present in each IL) that lead to differences in the separation and selectivity properties. Whilst the SUPELCOWAX®10 × SLB-IL111 combination provided the largest phase selectivity differences, apparently less spread

of components over the 2D separation space was observed (Figure 4A). In the current examples, a phosphonium based IL was found to exhibit larger distribution constant differences for the oxygenated sesquiterpenes than that of imidazoliumbased IL phases (i.e., increased range of 3tR values), which can be rationalized as a progressive increment of solute solubility in the IL phases as polarity decreased. Considering the good separation performance of SLB-IL59 (Figure 4C), it was chosen as the 3D column for subsequent analysis. Without cryogenic modulation of the 2D effluents in the 3 DIL column, compounds are displayed as classical GCnp−GCPEG analyses. Subsequent modulation of the solutes provide an average peak width at half height of about 400 ms with sufficient data points (ca. 19) per peak acquired by the

ACS Paragon Plus Environment

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 accTOFMS which has a relatively slower acquisition, for effective peak reconstruction. The analytical benefits of the sequential chemical selectivities across the three columns (nonpolar and typical polar arrangements) in GCnp−GCPEG×GCIL is apparent. For instance, baseline separation of two compounds (compounds 5 and 15 in Figure 2C) in hop essential oil that were otherwise co-eluted in 2DPEG was achieved; the resolution (Rs) for these compounds is ca. 0.4 when applying GCnp−GCPEG–accTOFMS analysis, and was significantly enhanced by modulating the effluent from the 2DPEG, providing a Rs > 1.5. A more striking example is the sequential GCnp– GCPEG×GCIL–accTOFMS analysis of oxygenated sesquiterpenes in A. malaccensis oleoresin, with a large number of components well resolved (Rs > 1.5) in the GCPEG×GCIL chromatographic profile (Figure 3C). The gain in compound coverage is apparent; 1D GC analysis (for the target analytes region) of A. malaccensis oleoresin enables detection of ca. 13 compounds, while GCnp−GCPEG×GCIL analysis indicated ca. 38 compounds, corresponding to a threefold increase in detected components. Apparent from the result in Figure 3C is a structured result comprising perhaps 3 series of peaks, which evidently is the reason for the better 2D selectivity of the GC×GC operation. In addition to the high separation power, the accTOFMS dimension provides informative structural elucidation of compounds, with high mass resolution and mass measurement accuracies of < 6 ppm which increases confidence for attribution of ion heteroatomic formulae of compound fragmentations. As an example, the MS corresponding to peak 5 and peak 15 in hop essential oil, suggests guaiol and γgurjunenepoxide-(2) to be possible compounds with the highest match threshold (> 750) (Figure S2). However, the highest matching library entry does not guarantee correct identity due to the many similar fragmentation patterns of plant secondary compounds. Exact mass analysis of m/z 222.1984 and m/z 220.1833, suggested the formulae C15H26O (mass accuracy −2.62 ppm) and C15H24O (mass accuracy −5.15 ppm), respectively, which support the selection of library entries. To further ascertain peak identity, retention indices (RI; 2DPEG) were determined and compared with those reported in literature. Placing the CT at the inlet of the 2DPEG column allows multiple heart-cutting of each alkane from the 1D column to be cold trapped then eluted using a single chromatographic analysis on the 2D polar phase. Revolatilization (i.e., CO2 supply to CT switched off) of alkanes commences simultaneously with initiation of exactly the same second dedicated oven program used for analysis of hop and agarwood samples. Figure 5 illustrates the multiple H/C operations of alkanes, indicating good flow switching and balancing, and cold trapping efficiency. By applying RI matching, a large number of NIST library search entries could be filtered out (rejected). The 3D RI could also be determined to further improve the selection of accuracy threshold.24 Due to unavailability of standard reference RI for sesquiterpenes on ionic liquid phases, 3DIL RI was not possible. The oxygenated sesquiterpenes in hop essential oil and agarwood oleoresin by the sequential GCnp–GCPEG×GCIL system were tentatively identified based on the aforementioned selection criteria (Table 1 and Table S3).

Figure 5. Heart-cutting analysis of n-alkanes showing: (A) the separation of linear C12–C24 on the 1Dnp column including the 12 heart-cut windows indicated by red arrows. The 1Dnp FID response after the heart-cut operation is overlaid, showing effective system balance; (B) separation of the heart-cut alkanes on 2DPEG column; (C) separation of hop essential oil sample on 2DPEG column. Overlay of alkanes and hop compounds (i.e., plot B and C) permits calculation of index values.

To depict the increment of compound coverage (i.e. the number of components physically resolved and detected), the separation performance of GC×GC and sequential 3D−GC was compared (Figure S3). As expected, GC×GC analyses of hop (Figure S3 Ai and Aii) and agarwood (Figure S3 Bi and Bii) samples displayed fewer resolved peaks for the oxygenated sesquiterpene region, which is logical due to the lower peak capacity in comparison to the sequential 3D separation. Analytically, the chromatographic selectivity afforded through the use of three complementary stationary phases leads to considerable increase of peak capacity. The peak capacity for the proposed GCnp−GCPEG×GCIL may be represented as nc, 3D = 1nc + (2nc × x) × 3nc 1 2 where nc, nc and 3nc are the 1D, 2D and 3D peak capacities, respectively and x is the number of transferred fractions. The calculation of 2nc × 3nc accords with the equation defined by Synovec et al.17 nc, 2D = 1t / 1MR 2w, where 1t, 1MR and 2w are the 1 D runtime, modulation ratio, and the 2D peak width at base, respectively. The calculation of 1nc uses the equation nc = tR 25 (max) – tM / w, where tR (max), tM and w are retention time of the last peak, hold-up time and average peak width (4σ criterion), respectively. In the current example, nc for 1D GC analysis of agarwood oleoresin is ca. 400 (ca. 350 for hop), the nc for GCnp−GCPEG analysis of agarwood oleoresin is ca. 1100 (ca. 950 for hop), and GCnp−GCPEG×GCIL analysis of agarwood oleoresin provided the highest peak capacity of ca. 5000 (ca. 4500 for hop). The peak capacity production (i.e., peak capacity per separation time) of each system was determined, over the region of elution of the target oxygenated sesquiterpenes. Results indicated that 1D GC provides approximately 5 peaks/min, GCnp−GCPEG yields ~8.5 peaks/min, and the highest is GCnp−GCPEG×GCIL which provides ~40 peaks/min. If the number of peaks is considered with respect to the total analy-

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

sis time, the extended analytical time of the GCnp−GCPEG×GCIL experiment will yield significantly less peak capacity production, since only a limited target region is selected. The significant increment of peak capacity for GCnp−GCPEG×GCIL system also translated to greater resolving power as evidenced from the separation of three agarwood critical components, tentatively identified as santalcamphor (peak 15), estafiatin (peak 15a) and saussurea lactone (peak 15b) (Figure 3C), in 3D with Rs = 1.5, yielding three distinct “pure” accurate mass spectra (Figure S4). These compounds apparently were not readily resolved using the 2D polar column. The intra-day precision (% RSD) for the GCnp−GCPEG×GCIL system was assessed by five consecutive injections of hop oil sample on the same day (n = 5). 1tR, 2tR, 3tR and peak areas of 10 selected peaks were ≤ 0.01%, 0.07%, 0.71% and 7.5%, respectively (Table S4), suggesting good repeatability of the method. This also signifies that incorporating both GC−GC and GC×GC in a single GC system does not unduly affect compound retention or measurement precision provided that the system is well flow-balanced to enable precise control of flow switching, solute trapping and release, and cryogenic modulation. CONCLUSION This study demonstrates the development of an integrated GCnp−GCPEG×GCIL−accTOFMS system that provides three dimensions of separation selectivity of selected regions for high resolution GC analysis of complex plant extracts. The system provides initial assessment of phytocomplexity of the samples though 1Dnp analysis with FID. DS flow switching directs regions of interest (usually regions with multiple overlaps) for additional separation on 2DPEG, essentially a twodimensional separation, followed by detection with accTOFMS. For H/C solutes that still suffer significant coelutions on 2DPEG, a comprehensive 2D stage using GCPEG × GCIL separation was applied to the whole H/C 2D effluent to further improve molecular separation of the solutes. This integrates a decision process based on the complexity of the samples, and the degree of separation sought from either GCnp, GCnp−GCPEG or GCnp−GCPEG×GC separation. With higher dimensions of chromatographic separation (from 1D GC to GC−GC×GC), the extension of compound coverage for oxygenated sesquiterpenes was apparent, in addition to improved peak capacity and separation efficiency. A threefold increase in the number of detected compounds was experimentally obtained. The described methodology should be a valuable adjunct for the improved hybrid MDGC analysis of complex samples, offering high resolution chromatographic and accurate mass analysis of plant-derived extracts.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website. Additional Figures and Tables (pdf).

AUTHOR INFORMATION

Notes The authors declare no competing financial interests.

ACKNOWLEDGMENT DDY gratefully acknowledges the provision of a Tasmania Graduate Research Scholarship. Agarwood sample provided by Green Agro Agar-wood Products Sdn. Bhd. are also acknowledged. The authors would like to acknowledge support from the Australian Research Council (Linkage Project LP140100160: Hop Flavoromics for Distinctive Beer), and Discovery Project DP130100217.

REFERENCES (1) Fernie, A. R.; Trethewey, R. N.; Krotzky, A. J.; Willmitzer, L. Nat. Rev. Mol. Cell Biol. 2004, 5, 763–769. (2) Hall, R. D. New Phytol. 2006, 169, 453–468. (3) Fiehn, O. Plant Mol. Biol. 2002, 48, 155–171. (4) De Vos, R. C. H.; Moco, S.; Lommen, A.; Keurentjes, J. J. B.; Bino, R. J.; Hall, R. D. Nat. Protoc. 2007, 2, 778–791. (5) Lisec, J.; Schauer, N.; Kopka, J.; Willmitzer, L.; Fernie, A. R. Nat. Protoc. 2006, 1, 387–396. (6) Tranchida, P. Q.; Franchina, F. A.; Dugo. P.; Mondello, L. Mass Spectrom. Rev. 2016, 35, 524–534. (7) Prebihalo, S. E.; Berrier, K. L.; Freye, C. E.; Bahaghighat, H. D.; Moore, N. R.; Pinkerton, D. K.; Synovec, R. E. Anal. Chem. 2018, 90, 505–532. (8) Cordero, C.; Kiefl, J.; Schieberle, P.; Reichenbach, S. E.; Bicchi, C. Anal. Bioanal. Chem. 2015, 407, 169–191. (9) Wong, Y. F.; Perlmutter, P.; Marriott, P. J. Metabolomics 2017, 13, 46. (10) Magagna, F.; Cordero, C.; Cagliero, C.; Liberto, E.; Rubiolo, P.; Sgorbini, B.; Bicchi, C. Food Chem. 2017, 225, 276–287. (11) Magagna, F.; Valverde-Som, L.; Ruíz-Samblás, C.; CuadrosRodríguez, L.; Reichenbach, S. E.; Bicchi, C.; Cordero, C. Anal. Chim. Acta 2016, 936, 245–258. (12) Eyres, G.; Marriott, P. J.; Dufour, J. P. J. Chromatogr. A 2007, 1150, 70–77. (13) Wong, Y. F.; Kulsing, C.; Marriott, P. J. Anal. Chem. 2017, 89, 5620–5628. (14) Mitrevski, B.; Marriott, P. J. Anal. Chem. 2012, 84, 4837–4843. (15) Watson, N. E.; Bahaghighat, H. D.; Cui, K.; Synovec, R. E. Anal. Chem. 2017, 89, 1793–1800. (16) Watson, N. E.; Prebihalo, S. E.; Synovec, R. E. Anal. Chim. Acta 2017, 983, 67–75. (17) Siegler, W. C.; Crank, J. A.; Armstrong, D. W.; Synovec, R. E. J. Chromatogr. A 2010, 1217, 3144–3149. (18) Yan, D.; Wong, Y. F.; Tedone, L.; Shellie, R. A.; Marriott, P. J.; Whittock, S. P.; Koutoulis, A. J. Chromatogr. A 2018, 1536, 110– 121. (19) Wong, Y. F.; Chin, S. T.; Perlmutter, P.; Marriott, P. J. J. Chromatogr. A 2015, 1387, 104–115. (20) Praet, T.; Van Opstaele, F.; Steenackers, B.; De Vos, D.; Aerts, G.; De Cooman, L. J. Am. Soc. Brew. Chem. 2016, 74, 65–76. (21) Qi, M.; Armstrong, D. W. Anal. Bioanal. Chem. 2007, 388, 889– 899. (22) Ragonese, C.; Sciarrone, D.; Tranchida, P. Q.; Dugo, P.; Dugo, G.; Mondello, L. Anal. Chem. 2011, 83, 7947–7954. (23) Wong, Y. F.; Uekane, T. M.; Rezende, C. M.; Bizzo, H. R.; Marriott, P. J. J. Chromatogr. A 2016, 1477, 91–99. (24) Jiang, M.; Kulsing, C.; Nolvachai, Y.; Marriott, P. J. Anal. Chem. 2015, 87, 5753–5761. (25) Engewald, W.; Dettmer-Wilde, K. In Practical Gas Chromatography; Dettmer-Wilde, K.; Engewald, W., Eds.; Springer-Verlag: Berlin Heidelberg, 2014; pp 51.

Corresponding Author *Email: [email protected]

ACS Paragon Plus Environment

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

Analytical Chemistry

ACS Paragon Plus Environment