Chiral recognition of amino acid enantiomers using high-definition differential ion mobility mass spectrometry

Chiral recognition of amino acid enantiomers using high-definition differential ion mobility mass spectrometry

International Journal of Mass Spectrometry 428 (2018) 1–7 Contents lists available at ScienceDirect International Journal of Mass Spectrometry journ...

1MB Sizes 0 Downloads 36 Views

International Journal of Mass Spectrometry 428 (2018) 1–7

Contents lists available at ScienceDirect

International Journal of Mass Spectrometry journal homepage: www.elsevier.com/locate/ijms

Full Length Article

Chiral recognition of amino acid enantiomers using high-definition differential ion mobility mass spectrometry J. Diana Zhang, K.M. Mohibul Kabir, Hyun Eui Lee, William A. Donald ∗ School of Chemistry, University of New South Wales, Sydney, NSW, 2052, Australia

a r t i c l e

i n f o

Article history: Received 15 August 2017 Received in revised form 25 December 2017 Accepted 7 February 2018 Available online 13 February 2018 Keywords: Chiral recognition Enantiomer separation Differential ion mobility High-field asymmetric waveform ion mobility Ion mobility spectrometry Mass spectrometry

a b s t r a c t Enantiomeric analysis of small molecules is important in many research fields, including in drug development. Here, chiral recognition of amino acid enantiomers using differential ion mobility spectrometry (DMS) mass spectrometry (MS) is demonstrated. Diastereomeric proton bound complexes were formed between enantiomers of amino acids (tryptophan and phenylalanine) and N-tert-butoxycarbonylO-benzyl-l-serine (BBS) by electrospray ionization for analysis by DMS-MS and collision-induced dissociation (CID). If the DMS resolution is sufficiently high, ionic diastereomers (as opposed to enantiomers) can in principle be separated without the use of a chiral gas. Peaks corresponding to the Land D-enantiomers for both tryptophan and phenylalanine in the DMS-MS spectra were resolved by increasing the ratio of He in the carrier gas from 100% N2 to 50:50 He:N2 %. In contrast, CID spectra of the corresponding diastereomeric dimer complex ions were nearly identical, indicating that chiral recognition by CID was not possible under these conditions. For mixtures of L- and D-tryptophan, a linear calibration curve can be obtained by plotting the enantiomeric excess measured by DMS-MS vs. the known values in solution (slope of 1.000, intercept of −0.010 and R2 of 0.997). That is, enantiopurity can be quantified using a separation process that occurs in milliseconds. Thus, DMS-MS analysis of proton bound diastereomeric dimers is a powerful approach for the rapid enantiomeric analysis of relatively small molecules. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Chirality is naturally embedded within many biomolecules, including amino acids, sugars and proteins, and is essential to proper cellular and biological function [1,2]. Molecular chirality is of fundamental importance in drug discovery [1,3]. For example, more than 50% of drugs that are brought to market have chiral centers, and of these, ca. 50% are administered as racemates [4,5]. However, pure enantiomeric forms of drugs are often needed in order to produce a desired therapeutic effect because the biological activity of enantiomers can differ significantly in a chiral environment, including many ligand-protein binding motifs [6,7]. For example, R-thalidomide is an effective anti-nausea drug, whereas S-thalidomide can cause birth defects [4]. Thus, newly developed drugs, as recommended by the Food and Drug Administration (FDA), are required to undergo intensive enantiomeric analysis [8,9]. FDA guidelines emphasize the need for rapid, robust and sen-

∗ Corresponding author. E-mail address: [email protected] (W.A. Donald). https://doi.org/10.1016/j.ijms.2018.02.003 1387-3806/© 2018 Elsevier B.V. All rights reserved.

sitive methodologies in both qualitative and quantitative analyses during drug development [1,10] In order to achieve enantiomeric separation of chiral molecules, enantiomers need to be subjected to an asymmetric chiral environment; i.e., an environment that promotes stereoselective interactions. Enantioselectivity requires at least three-points of intermolecular interactions between an enantiomer and a chiral environment such that at least one interaction is stereochemically dependent, which is known as Pirkle’s rule [11]. One application of Pirkle’s rule is to form a diastereomeric complex by complexing a chiral selector molecule with an enantiomer. Thus, enantioselectivity can in principle be achieved by using analytical instrumental methodologies that are sensitive to the different thermodynamic energies and structures present in the binding of the ‘analyte’ enantiomer to the chiral selector [11,12]. Several mass spectrometry (MS) methods for chiral recognition have been developed for rapid, selective and sensitive detection of chiral analytes from chemical mixtures [1,2,10]. These MS chiral recognition methodologies can be divided into three main categories: (i) determining the relative abundance of diastereomeric adduct ions that are formed between an enantiopure reference compound and analyte enantiomers; (ii) using collision-induced

2

J.D. Zhang et al. / International Journal of Mass Spectrometry 428 (2018) 1–7

dissociation (CID) to identify any differences in the kinetic stability of diastereomeric complex ions; and (iii) using gas-phase ion molecule reactions to identify differences in thermodynamic or kinetic constants of reactions that depend on chirality [2,10,12,13]. For example, Yao et al. [14] measured the difference in the relative product ion abundances in the CID mass spectra of size-selected diastereomeric proton bound trimer complex ions for the chiral recognition of amino acids. However, such methods require the formation of relatively large and weakly bound complex ions for optimal chiral recognition. Moreover, the CID fragmentation patterns of many diastereomeric complex ions that differ by the presence of an L- and D-enantiomer are indistinguishable [14,15]. Ion mobility spectrometry (IMS) is an emerging technique for gas-phase chiral analysis [1]. In IMS, gas-phase ions can be separated and detected based on their mobility as they pass through a buffer gas under the influence of a constant weak electric field [12,16]. The mobility of an ion in a given carrier gas can depend on the mass (m), charge (z), and collision cross-section (CCS) of the ion, the carrier gas, and the electric field strength [2,13,17–19]. In conventional IMS, an electric field lower than 100 V/cm is used to ensure that ion mobility is independent of the electric field strength [20]. Thus, ions can be separated because the drift velocity of the ions depend on their m, z, and CCS. IMS based ion separations can be several orders of magnitude shorter than common methodologies for chiral separations, such as high-performance liquid chromatography [13,21]. Furthermore, as an orthogonal approach to MS, IMS can be used to separate ions prior to mass analysis, which can increase ion signal-to-noise and lower limits of detection [22,23]. In the last decade, there has been a strong interest in developing IMS as a robust method for enantiomeric analysis [1,22]. Recently, research in this field has primarily involved the analysis of relatively large biomolecules such as epimeric glycans [24], catechin epimers [17], carbohydrate anomers [16], and mono- and disaccharide isomers [25]. Studies regarding the enantiomeric analysis of small molecules such as amino acids and those relevant to drug development have been limited. To date, there has been four reports of IMS being used for chiral separations in the literature [2], two of which involves relatively well-established IMS based instruments; i.e., drift tube ion mobility-mass spectrometry (DTIMS) [3] and travelling wave ion mobility-mass spectrometry (TWIMS) [13]. By use of IMS, enantiomeric amino acids have been separated by using (i) an enantiopure amino acid as a reference compound complexed with a divalent metal, to form metal bound trimers [13,18,23]; and (ii) a chiral modifier that was doped into the drift gas, which resulted in different ion mobilities for each enantiomer ion upon ion-molecule clustering with the chiral gas modifier [3]. However, these IMS methods can be limited by (i) the formation of weakly bound and relatively large complex ions (i.e., trimers), which require the optimization of concentration ratios between a reference compound, metal ion, and chiral selector; (ii) relatively high analyte concentrations; (iii) the use of relatively ‘soft’ ionization methods that favor ion formation of polar molecules (e.g., electrospray ionization, ESI); and (iv) insufficient resolution of the L- and the D- enantiomers for direct and accurate quantitation [26]. In addition, the prolonged use of a suitable chiral buffer gas dopant can be relatively expensive compared to the use of solid chiral selector molecules, such as modified amino acids with a tert-butoxycarbonyl (Boc) protecting group (e.g., N-tert-butoxycarbonyl-O-benzyl-l-serine; BBS, where Bzl is a benzyl group) [14]. Thus, the exploration of alternative MS based methods for rapid chiral analysis is warranted. One type of IMS based methodology that has undergone considerable improvement in resolving power is differential ion mobility spectrometry (DMS), otherwise known as high-field asymmetric waveform ion mobility spectrometry or FAIMS (Fig. 1). DMS utilizes an alternating high and low electric field to separate gas-phase ions, in which the high-field portion of the electric field is greater

Fig. 1. Diagram of differential ion mobility and the ion funnel interface to a linear quadrupole ion trap mass spectrometer.

than 10 kV/cm [20]. Under the influence of a high electric field, the mobility of an ion depends nonlinearly on the electric field [27]. In DMS, a high frequency alternating asymmetric voltage is applied between two electrode plates and a longitudinal gas flow ‘carries’ ions between the two plates (Fig. 1) [18]. For the asymmetric voltage, the high-field is applied for a short duration and a lower voltage of opposite polarity is applied for a longer duration. The peak amplitude of the high-field portion is referred to as the dispersion voltage (DV). In the resultant ‘dispersion field’ (ED = DV/d where d is the distance between the DMS electrodes), ions migrate towards one electrode under the high-field portion of the waveform, and reverse direction towards the other electrode during the low-field portion of the waveform. By superimposing a constant and adjustable DC compensation voltage (CV), resulting in a ‘compensation field’ (EC ), onto the asymmetric waveform, the net displacement of the ion caused by the different ion mobilities in the high and low electric fields can result in the transmission of different ions between the plates. By scanning a range of compensation voltages, a spectrum of the relative number of ions that have traversed the gap can be obtained as a function of CV [18,22,28–31]. Some advantages of DMS include: (i) ion separation is more orthogonal to MS than other IMS based methods because ion separation depends less strongly on m and z, and more on the electric field [32]; and (ii) the peak capacity of DMS can be relatively high compared to other IMS based techniques [22]. In DMS, the resolving power for a given peak is defined as the ratio of EC and the full-width-at-half maximum of the peak. The resolving power of DMS-MS was initially limited to ∼10 [28], partly due to the use of curved (cylindrical or spherical) electrodes that resulted in inhomogeneous electric fields between the plates [31]. However, the resolving power values for multiply charged peptide ions as high as ∼500 has been reported by the use of ‘high-definition’ DMS-MS [31]. The improvement in resolving power in high-definition DMSMS has been attributed in part to the use of: (i) planar electrodes for more homogeneous electric fields than for curved electrodes (cylindrical FAIMS) [28]; (ii) relatively high concentrations of He and H2 in the carrier gas, which results in higher ion mobilities than by use of heavier carrier gases (and non-Blanc ion mobility behavior) [28,30,33,34]; and (iii) higher ED values than were used previously, which enhances ion separation [31,35]. The use of DMS-MS for chiral recognition has been reported twice by Reimann and co-workers [18,23], which corresponds to the third and fourth reported example of using any IMS based method for chiral analysis. In these studies [18,23], cylindrical FAIMS was used to partially separate metal bound trimer com-

J.D. Zhang et al. / International Journal of Mass Spectrometry 428 (2018) 1–7

3

plex ions that contained one divalent metal ion, one enantiomeric analyte molecule, and two reference amino acid chiral selector molecules. Here, high-definition DMS [31] is used for the first time for enantiomeric recognition and analysis of small molecules. By use of high He fractions in the carrier gas, proton bound dimers containing an enantiopure analyte molecule (tryptophan or phenylalanine) and the chiral selector, (N-tert-butoxycarbonyl-Obenzyl-l-serine), can be readily and completely resolved without the use of metal ions or ionic trimer complexes. These results indicate that DMS separations can be sensitive to subtle differences between proton bound diastereomeric complex ions that cannot be distinguished using CID. In addition, DMS-MS can be used to directly quantify enantiomeric excess of tryptophan in mixtures of L- and D-tryptophan, where enantiomeric excess is defined in terms of enantiomeric ratios [2]. 2. Experimental 2.1. Chemicals L-tryptophan was purchased from Alfa Aesar (Heysham, England). D-tryptophan, L-phenylalanine, D-phenylalanine and N-tert-butoxycarbonyl-O-benzyl-l-serine (BBS), were purchased from Sigma (St Louis, MO, USA), and used without further purification. Methanol was obtained from Scharlau (Sentmenat, Spain). Deionized water (18 M) was obtained using a MilliQ water purification system (Merck, Darmstadt, Germany). Stock solutions of the amino acids were prepared in 90:10 methanol:water, and the stock solution of BBS was prepared in 100% methanol. Sample solutions containing equimolar concentrations of the amino acid (200 ␮M) and the chiral selector (200 ␮M) were prepared in 99:1 methanol:acetic acid for mass spectrometry experiments. Racemic solutions were prepared using stock solutions of L- and D-enantiomers (tryptophan or phenylalanine), and BBS (400 ␮M) in 99:1 methanol:acetic acid. To obtain a calibration curve, concentrations of L- and D-tryptophan were adjusted such that the final total tryptophan concentration was 400 ␮M.

Fig. 2. (a) Tryptophan, (b) phenylalanine, and (c) N-tert-butoxycarbonyl-O-benzyll-serine. Asterisks denote chiral centers.

MA, USA), a minor portion of which is directed towards the ESI emitter to desolvate ions and the major portion carries ions between the two DMS plates towards the capillary entrance to the MS. The outlet of the DMS device was fixed ∼2.5 mm from the capillary entrance to the MS. A DC voltage of 150 V was applied to one DMS electrode. An asymmetric AC waveform (2:1 bisinusoidal, 860 kHz, and 4.1 kV dispersion voltage) and a DC voltage was applied to the other electrode (CV + 150 V). For all DMS spectra, the CV scan speed was 0.5 V/min. All data were collected in triplicate, and uncertainty values were obtained from the standard deviation of such replicates. Calculations and uncertainty values for the linear regression analysis were obtained using Igor Pro v 4.0. 3. Results and discussion 3.1. Chiral selector and limitations of CID-based chiral recognition

2.2. ESI–MS Mass spectrometry experiments were performed using a linear quadrupole ion trap MS (LTQ, Thermo Scientific, San Jose, CA, USA) that is equipped with a planar DMS device (see below), an electrodynamic ion funnel [36] (Heartland Mobility, Wichita, KA, USA), and an ESI source. For ESI, solutions were directly infused (5 ␮L/min) through a pulled borosilicate capillary emitter (76 ␮m inner diameter). A voltage of 3.5 kV was applied to the ESI emitter relative to the curtain plate entrance to the DMS device to initiate and maintain ESI. The capillary entrance to the ion funnel of the MS was heated to 120 ◦ C. For CID of [(L/D-tryptophan)(BBS) + H]+ , an isolation width of 3 m/z (centered on the m/z of the ion of interest) and a normalized collision energy of 1% was used. All measurements were obtained in triplicate to report the calculated average and standard deviation of these replicates. 2.3. DMS The planar DMS analyzer has been previously described by Shvartsburg and co-workers (Fig. 1) [28,37]. The DMS device consists of two 65 mm long and 35 mm wide stainless steel plates that are separated by an analytical gap of 1.88 mm. Ions formed via ESI are introduced orthogonally into the analytical gap through the inlets of a curtain plate (2 mm diameter; 1 kV vs. ground) and one of the DMS electrodes (1 mm diameter). A carrier gas (N2 mixed with 0–50% He) was introduced from the side of the curtain plate (1 L/min, controlled by flow meters; MKS Instruments, Andover,

Amino acids are typically used as benchmark standards in order to demonstrate the suitability of a method for the chiral analysis of small molecules. In this study, tryptophan and phenylalanine (Fig. 2) were chosen for analysis, to enable direct comparisons with previous work on the enantiomeric separations of amino acids using DTIMS [3], TWIMS [13] and cylindrical FAIMS [18]. Modified amino acids with a tert-butoxycarbonyl group have been identified as appropriate chiral ‘selectors’ for forming diastereomeric complexes (proton bound trimers) in ESI that can be used in CID for chiral recognition (see Introduction). Of three potential Boc-modified chiral selectors investigated by Yao et al. [14] [i.e., N-tert-butoxycarbonylphenylalanine, Ntert-butoxycarbonylproline, and N-tert-butoxycarbonyl-O-benzyll-serine)], N-tert-butoxycarbonyl-O-benzyl-l-serine (BBS) (Fig. 2c) was chosen in our work because proton bound trimers containing two BBS molecules and an enantiomer of an amino acid molecule could be used for chiral recognition by CID for a wider range of amino acids than by use of the other selector molecules. BBS contains a number of polarizable heteroatoms and two bulky substituents (benzyl and tert-butoxycarbonyl groups) to the chiral center (Fig. 2c), and such groups can facilitate multiple points of interactions with a chiral analyte that are stereochemically dependent. An ESI mass spectrum obtained for solutions containing tryptophan and BBS is shown in Fig. 3. Four ions (m/z 188, 205, 318 and 500) that correspond to [(D-Trp)-NH2 ]+ , [(D-Trp) + H]+ , [(BBS) + Na]+ , and [(D-Trp)(BBS) + H]+ respectively, were formed in

4

J.D. Zhang et al. / International Journal of Mass Spectrometry 428 (2018) 1–7 Table 1 Abundance of Fragment Ions Relative to the Precursor Ion for the CID of Proton Bound Dimers of Trp and BBS.a Relative Abundance (%)

L-Trp

D-Trp

[(Trp) + H]+ [(Trp)-NH2 ]+

58.04 ± 0.29 15.18 ± 0.18

58.45 ± 0.89 15.29 ± 0.12

a

Normalized CID energy of 1%, 30 ms activation, and isolation width of 3 m/z.

[(Trp) + H]+ , and [(Trp)-NH2 ]+ formed via the loss of BBS (Eq. (1a)), and BBS along with NH3 (Eq. (1b)), respectively. [(Trp)(BBS) + H]+ → [(Trp) + H]+ + BBS +

+

[(Trp) + H] → [(Trp)-NH2 ] + NH3 Fig. 3. Electrospray ionization mass spectrum of D-tryptophan (200 ␮M), and Ntert-butoxycarbonyl-O-benzyl-l-serine (200 ␮M) in 99:1 methanol:acetic acid.

[(L-Trp)(BBS) + H]+

(1a) (1b)

[(D-Trp)(BBS) + H]+ ,

and the relaFor CID of tive abundances of the product ions are essentially the same for both diastereomeric complex ions (Fig. 4 and Table 1). These data indicate that: (i) the activation free energies of ion dissociation are nearly the same for both diastereomeric cluster ions, and (ii) chiral recognition using CID for the dimeric cluster ion is not possible under these conditions. These data are consistent with the CID results reported for this same complex ion by Yao et al. [14] Vekey and Czira [15] also demonstrated that CID of proton bound dimers of alanine, proline, tryptophan, and valine, using tryptophan, phenylalanine, or proline as the chiral selector molecule resulted in nearly identical fragmentation patterns, indicating that CID is not sufficiently sensitive to slight differences in the binding energies of many proton bound diastereomeric dimers for chiral recognition. 3.2. Chiral recognition by DMS-MS

Fig. 4. Collision-induced dissociation mass spectra of (a) [(L-Trp)(L-BBS) + H]+ , and (b) [(D-Trp)(L-BBS) + H]+ .

relatively high abundances (>20%). In contrast, [(D-Trp)(BBS)2 + H]+ was formed in relatively low abundance (∼6%). The proton bound trimer is formed in relatively low abundance relative to the proton bound dimer likely as a result of the dimer having less steric hindrance in the complex ion than the trimer. Moreover, sequential ion-molecule binding energies tend to decrease as the number of neutral molecules in ionic clusters increase and charge densities decrease [38,39]. The formation of sodiated BBS in higher abundance than protonated BBS is consistent with the results reported by Yao et al. [14] [(D-Trp)-NH2 ]+ is a well-known collisional activation product of protonated tryptophan [40], with a reaction barrier of 97.5 kJ/mol [40]. The reaction barrier is sufficiently low that the product ion can be readily formed during ion transfer from atmospheric pressure to the mass analyzer, even under conditions that favor the formation of relatively weakly bound dimer and trimer ions. Collision-induced dissociation mass spectra of size-selected [(LTrp)(BBS) + H]+ and [(D-Trp)(BBS) + H]+ are shown in Fig. 4. For both precursor ions, the primary CID product ions correspond to

To separate diastereomeric cluster ions that cannot be distinguished by CID, ideally the DMS-MS resolving power should be as high as possible. One of the primary contributors to both the resolving power and resolution of DMS is the carrier gas composition. Particularly, the use of lighter carrier gases such as He can significantly improve resolving power because ions have higher mobility in lighter carrier gases than in heavier gases [31,33]. Moreover, the peak capacity of DMS can depend on the carrier gas composition to a greater extent than in drift tube based IMS owing to the non-Blanc dependence of ion mobility in gas mixtures at high electric fields [34]. Thus, DMS-MS spectra of [(Trp)(BBS) + H]+ formed from a solution containing a racemic mixture of tryptophan and BBS were acquired as a function of increasing He number density in a He:N2 carrier gas mixture (Fig. 5). As the extent of He in the carrier gas increased from 0:100 to 50:50 He:N2 , the spectral resolution increased dramatically. For a carrier gas composition of 0:100 He:N2 , at least four DMS spectral features were identified but not resolved from each other (Fig. 5a). In contrast, the use of 50:50 He:N2 resulted in the baseline resolution of five spectral features and resolving power for all five ions ranges from ∼15 to 41 (Fig. 5f). The abundance of the unlabeled peak at EC values of ∼28 V/cm (Fig. 5a), 33 V/cm (Fig. 5b) and 39 V/cm (Fig. 5c) decreases as the He fraction in the carrier gas increases, and the peak is not observed at higher He concentrations (Fig. 5d–f). This may result from higher radial diffusion of the ion packet in lighter gases than heavier gases during ion transit between the DMS electrodes, which leads to reduced ion transmission at relatively high He concentrations [33]. The use of higher He fractions and ED values resulted in electrical breakdown, which limited further improvements to resolution. Moreover, the CV values shifted to significantly higher values as the He fraction in the carrier gas increased from 0 to 50%. These data indicate that the ions are dispersed to a greater extent in the asymmetric electric field in the presence of higher than lower concentrations of He in N2 , which results in greater ion separation.

J.D. Zhang et al. / International Journal of Mass Spectrometry 428 (2018) 1–7

Fig. 5. Differential ion mobility spectra of [(Trp)(BBS) + H]+ formed from a racemic mixture using a carrier gas composition of (a) 100:0 N2 :He, (b) 90:10 N2 :He, (c) 80:20 N2 :He, (d) 70:30 N2 :He, (e) 60:40 N2 :He, and (f) 50:50 N2 :He.

5

In Fig. 6, representative DMS-MS spectra of [(L-Trp)(BBS) + H]+ , [(D-Trp)(BBS) + H]+ , and the racemate of [(Trp)(BBS) + H]+ that were acquired using the ‘optimal’ DMS separation conditions are shown. A dominant spectral feature at an EC value of ∼38 V/cm was observed in all three spectra. In contrast, spectral peaks at EC values of ∼31 and ∼51 V/cm were found in the spectra of [(L-Trp)(BBS) + H]+ (Fig. 6a) and the racemate of [(Trp)(BBS) + H]+ (Fig. 6c), but not in that for [(D-Trp)(BBS) + H]+ (Fig. 6b). In addition, two spectral peaks at EC values of ∼28 and ∼48 V/cm were obtained in the spectra of [(D-Trp)(BBS) + H]+ (Fig. 6b) and the racemate of [(Trp)(BBS) + H]+ (Fig. 6c), but not in that for [(L-Trp)(BBS) + H]+ (Fig. 6a). These data indicate that four peaks in the racemate DMS spectrum of [(Trp)(BBS) + H]+ can be used to distinguish the L- and D-enantiomers of tryptophan from each other. In Fig. 6, DMS-MS spectra are shown for [(L-Phe)(BBS) + H]+ , [(D-Phe)(BBS) + H]+ , and the racemic mixture for [(Phe)(BBS) + H]+ . For L-Phe (Fig. 6d), the primary spectral feature at an EC value of ∼46 V/cm is also in the racemate spectrum (Fig. 6f) but not present in the D-Phe spectrum (Fig. 6e). Likewise, the primary spectral feature in the D-Phe spectrum (∼39 V/cm; Fig. 6e) is also present in the racemate spectrum (Fig. 6f) but not in the L-Phe spectrum (Fig. 6d). These data indicate that high-definition DMS-MS can also be used in the chiral recognition of phenylalanine using a proton bound dimer of phenylalanine and BBS. Previously, Yao et al. [14] demonstrated that CID of proton bound dimers of phenylalanine and BBS could not be used in the chiral recognition of phenylalanine. These data indicate that high-definition DMS-MS can be significantly more sensitive in distinguishing diastereomers than CID for these proton bound dimers. Although the concentration of the analyte used in these experiments is relatively high (200 ␮M) compared to many mass spectrometry measurements, the concentration used here is an order of magnitude lower than that used previously [14] for chiral analysis of tryptophan and phenylalanine by the collisioninduced dissociation method. For each chirally pure [(Trp)(BBS) + H]+ and [(Phe)(BBS) + H]+ , there are three and two major peaks, respectively, that correspond

Fig. 6. Differential ion mobility spectra of [(Trp)(BBS) + H]+ (a)-(c) and [(Phe)(BBS) + H]+ (d)-(f), formed from solutions containing. L-BBS and (a) L-Trp, (b) D-Trp, (c) racemic mixture of Trp, (d) L-Phe, (e) D-Phe, and (f) racemic mixture of Phe. Asterisks denote peaks used for quantitative analysis (Fig. 7).

6

J.D. Zhang et al. / International Journal of Mass Spectrometry 428 (2018) 1–7

Fig. 7. Enantiomeric excess of tryptophan measured using DMS-MS of [(Trp)(BBS) + H]+ (open square) plotted as a function of the known enantiomeric excess in solution. The linear regression best fit curve (black line) has a slope, intercept and R2 value of 1.000 ± 0.022, −0.010 ± 0.012 and 0.997, respectively. The standard deviation in each DMS-MS measurement was obtained from three replicate measurements.

to at least three and two different protonation isomers and/or ion conformations. For small molecules, multiple peaks in IMS are generally attributed to different protonation site isomers (or protomers) in the gas phase [26,41]. The difference in the number of DMS spectral peaks for tryptophan vs. phenylalanine can be attributed to the different conformational space in the respective proton bound dimers. Tryptophan has a relatively basic indole group, in addition to the amine and carboxylic acid functional groups, which can interact with the proton in different conformations (and protonation isomers). In contrast, the non-polar side chain of phenylalanine is unlikely to interact strongly with the proton, which may restrict the number of possible conformations for [(Phe)(BBS) + H]+ that can be formed via ESI compared to that for tryptophan. 3.3. Quantification of enantiomeric excess DMS-MS spectra of [(Trp)(BBS) + H]+ were obtained for mixtures containing different ratios of L and D-tryptophan with BBS, in which the CV values were scanned from ∼45 to 56 V/cm at 0.5 V/min. In Fig. 7, the measured relative ratios of the integrated DMS peak areas corresponding to L- and D-tryptophan, [L]/([L] + [D]), are plotted as a function of the known ratios for solutions containing tryptophan from 10:90 to 90:10 of each enantiomer. The linear regression best fit line for nine different enantiomeric solution mixtures has a slope, intercept and R2 value of 1.000 ± 0.022, −0.010 ± 0.012 and 0.997, respectively. Because the slope and intercept values correspond to the ideal values of 1 and 0, and the R2 is relatively high, these data indicate that high-definition DMS of proton bound dimers can be used for accurately quantifying enantiomeric excess for unknown mixtures of L-tryptophan and D-tryptophan. 4. Conclusions High-definition DMS can be used for the rapid and quantitative chiral recognition of small molecules, including tryptophan and phenylalanine. Collision-induced dissociation of proton bound diastereomeric complexes of tryptophan and Ntert-butoxycarbonyl-O-benzyl-l-serine, and phenylalanine and N-tert-butoxycarbonyl-O-benzyl-l-serine result in essentially no difference in the ion fragmentation patterns for the L- vs. D-enantiomers of tryptophan and phenylalanine, respectively. However, by use of relatively high concentrations of He in the ion carrier gas, DMS-MS can resolve the enantiomers of tryptophan and phenylalanine in proton bound diastereomeric complex ions.

These data indicate that DMS can be significantly more discerning of slight differences between diastereomeric complexes than CID. Surprisingly, two diagnostic peaks for chirality corresponding to [(L-Trp)(BBS) + H]+ and two corresponding to [(D-Trp)(BBS) + H]+ can be baseline resolved from each other in racemic mixtures by DMS (i.e., four diagnostic peaks of chirality in a single DMS spectrum). This is the first time that diastereomeric dimer ions (as opposed to trimers) have been used for chiral recognition using any ion mobility mass spectrometry based method. Moreover, DMSMS can be used to directly measure enantiomeric excess values for mixtures containing L-tryptophan to D-tryptophan in ratios from 0:100 to 100:0%. Given that DMS-MS can be used for the chiral analysis of small molecules by measuring protonated dimers without the use of metal ions, coupling high-definition DMS-MS to soft plasma based ion sources, such as dielectric barrier discharge ionization [42–44], may be particularly well-suited to the chiral analysis of medium to low polarity small molecules [45]. It is anticipated that high-definition DMS-MS will be a powerful approach for the rapid chiral recognition and quantitation of enantiomeric excess for many other small molecules, including those relevant to drug discovery. Acknowledgements This work was supported by the Australian Research Council (DP160102681) and the NSW Smart Sensing Network (NSW Department of Primary Industries). References [1] L. Wu, F.G. Vogt, A review of recent advances in mass spectrometric methods for gas-phase chiral analysis of pharmaceutical and biological compounds, J. Pharm. Biomed. Anal. 69 (2012) 133–147, http://dx.doi.org/10.1016/j.jpba. 2012.04.022. [2] X. Yu, Z.P. Yao, Chiral recognition and determination of enantiomeric excess by mass spectrometry: a review, Anal. Chim. Acta 968 (2017) 1–20, http://dx. doi.org/10.1016/j.aca.2017.03.021. [3] P. Dwivedi, C. Wu, L.M. Matz, B.H. Clowers, W.F. Siems, H.H. Hill Jr., Gas-phase chiral separations by ion mobility spectrometry, Anal. Chem. 78 (2006) 8200–8206, http://dx.doi.org/10.1021/ac0608772. [4] L.A. Nguyen, H. Hue, C. Pham-Huy, Chiral drugs: an overview, Int. J Biomed. Sci. 2 (2) (2006) 85–100, PMC3614593. [5] G.Q. Lin, Q.D. You, J.F. Cheng (Eds.), Chiral Drugs: Chemistry and Biological Action, Wiley, New Jersey, 2011, http://dx.doi.org/10.1002/9781118075647. [6] H.Y. Aboul-Enein, I.W. Wainer, The Impact of Stereochemistry on Drug Development and Use, Wiley, New York, 1997, 9780471596448. [7] A.J. Hutt, in: H.J. Smith (Ed.), Smith and Williams’ Introduction to the Principles of Drug Design and Action, 4th ed., CRC Press, Florida, 2006, pp. 117–183, 9780415288774. [8] FDA’s policy statement for the development of new stereoisomeric drugs, Chirality 4 (5) (1992) 338–340, http://dx.doi.org/10.1002/chir.530040513. [9] W.H. De Camp, Chiral drugs: the FDA perspective on manufacturing and control, J. Pharm. Biomed. Anal. 11 (11–12) (1993) 1167–1172, http://dx.doi. org/10.1016/0731-7085(93)80100-F. [10] E.A.A. De Camp, A. El-Aneed, Enantioselectivity of mass spectrometry: challenges and promises, Mass Spectrom. Rev. 32 (2013) 466–483, http://dx. doi.org/10.1002/mas.21379. [11] W.H. Pirkle, T.C. Pochapsky, Considerations of chiral recognition relevant to the liquid chromatographic separation of enantiomers, Chem. Rev. 89 (2) (1989) 347–362, http://dx.doi.org/10.1021/cr00092a006. [12] J.R. Enders, J.A. McLean, Chiral and structural analysis of biomolecules using mass spectrometry and ion mobility-mass spectrometry, Chirality 21 (2009) E253–264, http://dx.doi.org/10.1002/chir.20806. [13] V. Domalain, M. Hubert-Roux, V. Tognetti, L. Joubert, C.M. Lange, J. Rouden, C. Afonso, Enantiomeric differentiation of aromatic amino acids using travelling wave ion mobility-mass spectrometry, Chem. Sci. 5 (2014) 3234–3239, http:// dx.doi.org/10.1039/c4sc00443d. [14] Z.P. Yao, T.S.M. Wan, K.P. Kwong, C.T. Che, Chiral analysis by electrospray ionization mass spectrometry/mass spectrometry. 1. Chiral recognition of 19 common amino acids, Anal. Chem. 72 (2000) 5383–5393, http://dx.doi.org/10. 1021/ac000729q. [15] K. Vekey, G. Czira, Distinction of amino acid enantiomers based on the basicity of their dimers, Anal. Chem. 69 (1997) 1700–1705, http://dx.doi.org/ 10.1021/ac960931m. [16] J. Hofmann, H.S. Hahm, P.H. Seeberger, K. Pagel, Identification of carbohydrate anomers using ion mobility-mass spectrometry, Nature 526 (2015) 241–244, http://dx.doi.org/10.1038/nature15388.

J.D. Zhang et al. / International Journal of Mass Spectrometry 428 (2018) 1–7 [17] A. Troc, M. Zimnicka, W. Danikiewicz, Separation of catechin epimers by complexation using ion mobility mass spectrometry, J. Mass Spectrom. 50 (2015) 542–548, http://dx.doi.org/10.1002/jms.3560. [18] A. Mie, M. Jörntén-Karlsson, B.O. Axelsson, A. Ray, C.T. Reimann, Enantiomer separation of amino acids by complexation with chiral reference compounds and high-field asymmetric waveform ion mobility spectrometry: preliminary results and possible limitations, Anal. Chem. 79 (2007) 2850–2858, http://dx. doi.org/10.1021/ac0618627. [19] R. Cumeras, E. Davis, C.E. Davis, J.I. Baumbach, I. Gracia, Review on ion mobility spectrometry. Part 2: hyphenated methods and effects of experimental parameters, Analyst 140 (2015) 1391–1410, http://dx.doi.org/ 10.1039/c4an01101e. [20] R.M. O’Donnell, X. Sun, P.B. de B. Harrington, Pharmaceutical applications of ion mobility spectrometry, Trends Anal. Chem. 27 (1) (2008) 44–53, http://dx. doi.org/10.1016/j.trac.2007.10.014. [21] J.N. Dodds, J.C. May, J.A. McLean, Investigation of the complete suite of the leucine and isoleucine isomers: toward prediction of ion mobility separation capabilities, Anal. Chem. 89 (2017) 952–959, http://dx.doi.org/10.1021/acs. analchem.6b04171. [22] F. Lanucara, S.W. Holman, C.J. Gray, C.E. Eyers, The power of ion mobility-mass spectrometry for structural characterization and the study of conformational dynamics, Nat. Chem. 6 (2014) 281–294, http://dx.doi.org/10. 1038/NCHEM.1889. [23] A. Mie, A. Ray, B.O. Axelsson, M. Jorten-Karlsson, C.T. Reimann, Terbutaline enantiomer separation and quantification by complexation and field asymmetric ion mobility spectrometry-tandem mass spectrometry, Anal. Chem. 80 (11) (2008) 4133–4140, http://dx.doi.org/10.1021/ac702262k. [24] P. Both, A.P. Green, C.J. Gray, R. Sardzik, J. Voglmeir, C. Fontana, M. Austeri, M. Rejzek, D. Richardson, R.A. Field, G. Widmalm, S.L. Flitsch, C.E. Eyers, Discrimination of epimeric glycans and glycopeptides using IM-MS and its potential for carbohydrate sequencing, Nat. Chem. 6 (2014) 65–74, http://dx. doi.org/10.1038/NCHEM.1817. [25] H. Yang, L. Shi, X. Zhuang, R. Su, D. Wan, F. Song, J. Li, S. Liu, Identification of structurally closely related monosaccharide and disaccharide isomers by PMP labeling in conjunction with IM-MS/MS, Sci. Rep. 6 (2016) 28079, http://dx. doi.org/10.1038/srep28079. [26] H. Tian, N. Zheng, S. Li, Y. Zhang, S. Zhao, F. Wen, J. Wang, Characterization of chiral amino acids from different milk origins using ultra-performance liquid chromatography coupled to ion-mobility mass spectrometry, Sci. Rep. 7 (2017) 46289, http://dx.doi.org/10.1038/srep46289. [27] A.A. Shvartsburg, Differential Ion Mobility Spectrometry: Nonlinear Ion Transport and Fundamentals of FAIMS, CRC Press, Florida, 2008, 9781420051063. [28] A.A. Shvartsburg, F. Li, K. Tang, R.D. Smith, High-resolution FAIMS using planar geometry analyzers, Anal. Chem. 78 (11) (2006) 3706–3714, http://dx. doi.org/10.1021/ac052020 v. [29] A.A. Shvartsburg, R.D. Smith, Ultrahigh-resolution differential ion mobility spectrometry using extended separation times, Anal. Chem. 83 (1) (2011) 23–29, http://dx.doi.org/10.1021/ac102689p. [30] A.A. Shvartsburg, R.D. Smith, Accelerated high-resolution differential ion mobility separations using hydrogen, Anal. Chem. 83 (23) (2011) 9159–9166, http://dx.doi.org/10.1021/ac202386w. [31] A.A. Shvartsburg, T.A. Seim, W.F. Danielson, R. Norheim, R.J. Moore, G.A. Anderson, R.D. Smith, High-definition differential ion mobility spectrometry with resolving power up to 500, J. Am. Soc. Mass Spectrom. 24 (1) (2013) 109–114, http://dx.doi.org/10.1007/s13361-012-0517-5.

7

[32] A.A. Shvartsburg, A.G. Anderson, R.D. Smith, Pushing the frontier of high-definition ion mobility spectrometry using FAIMS, Mass Spectrom. (Tokyo) 2 (Spec Iss) (2013) S0011, http://dx.doi.org/10.5702/ massspectrometry.S0011. [33] A.A. Shvartsburg, W.F. Danielson, R.D. Smith, High-resolution differential ion mobility separations using helium-rich gases, Anal. Chem. 82 (6) (2010) 2456–2462, http://dx.doi.org/10.1021/ac902852a. [34] A.A. Shvartsburg, K. Tang, R.D. Smith, Understanding and designing field asymmetric waveform ion mobility spectrometry separations in gas mixtures, Anal. Chem. 76 (2004) 7366–7374, http://dx.doi.org/10.1021/ac049299k. [35] A.A. Shvartsburg, D.C. Prior, K. Tang, R.D. Smith, High-resolution differential ion mobility separations using planar analyzers at elevated dispersion field, Anal. Chem. 82 (18) (2010) 7649–7655, http://dx.doi.org/10.1021/ac101413k. [36] J.S. Page, K. Tang, R.D. Smith, An electrodynamic ion funnel interface for greater sensitivity and higher throughput with linear ion trap mass spectrometers, Int. J. Mass spectrom. 265 (2-3) (2007) 244–250, http://dx.doi. org/10.1016/j.ijms.2007.02.032. [37] A.A. Shvartsburg, K. Tang, R.D. Smith, Optimization of the design and operation of FAIMS analyzers, J. Am. Soc. Mass Spectrom. 16 (2005) 2–12, http://dx.doi.org/10.1016/j.jasms.2004.09.009. [38] W.A. Donald, R.D. Leib, M. Demireva, B. Negru, D.M. Neumark, E.R. Williams, Average sequential water molecule binding enthalpies of M(H2 O)19-124 2+ (M = Co, Fe, Mn, and Cu) measured with ultraviolet photodissociation at 193 and 248 nm, J. Phys. Chem. A 115 (1) (2011) 2–12, http://dx.doi.org/10.1021/ jp107547r. [39] W.A. Donald, E.R. Williams, Evaluation of different implementations of the thomson liquid drop model: comparison to monovalent and divalent cluster ion experimental data, J. Phys. Chem. A 112 (16) (2008) 3515–3522, http://dx. doi.org/10.1021/jp711012b. [40] H. Lioe, R.A.J. O’Hair, G.E. Reid, Gas-phase reactions of protonated tryptophan, J. Am. Soc. Mass Spectrom. 15 (2004) 65–76, http://dx.doi.org/10.1067/j. jasms.2003.09.011. [41] J. Boschmans, S. Jacobs, J.P. Williams, M. Palmer, K. Richardson, K. Giles, C. Lapthorn, W.A. Herrebout, F. Lemiere, F. Sobott, Combining density functional theory (DFT) and collision cross-section (CCS) calculations to analyze the gas-phase behavior of small molecules and their protonation site isomers, Analyst 141 (2016) 4044–4054, http://dx.doi.org/10.1039/c5an02456k. [42] M.M. Nudnova, L. Zhu, R. Zenobi, Active capillary plasma source for ambient mass spectrometry, Rapid Commun. Mass Spectrom. 26 (2012) 1447–1452, http://dx.doi.org/10.1002/rcm.6242. [43] M.C. Dumlao, D. Xiao, D. Zhang, J. Fletcher, W.A. Donald, Effects of different waveforms on the performance of active capillary dielectric barrier discharge ionization mass spectrometry, J. Am. Soc. Mass Spectrom. 28 (4) (2017) 575–578, http://dx.doi.org/10.1007/s13361-016-1535-5. [44] E.R. Stephens, M. Dumlao, D. Xiao, D. Zhang, W.A. Donald, Benzylammonium thermometer ions: internal energies of ions formed by low temperature plasma and atmospheric pressure chemical ionization, J. Am. Soc. Mass Spectrom. 26 (12) (2015) 2081–2084, http://dx.doi.org/10.1007/s13361-0151272-1. [45] J. Abian, The coupling of gas and liquid chromatography with mass spectrometry, J. Mass Spectrom. 34 (1999) 157–168, 10.1002/(SICI)1096-9888(199903)34:3<157:AID-JMS804>3.0. CO;2-4.