Imaging peptide and protein chirality via amino ... - ACS Publications

1. Imaging peptide and protein chirality via amino acid analysis by chiral ... 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45...
0 downloads 0 Views 878KB Size
Subscriber access provided by Kaohsiung Medical University

Article

Imaging peptide and protein chirality via amino acid analysis by chiral × chiral two-dimensional correlation liquid chromatography Ulrich Woiwode, Roland Johann Reischl, Stephan Buckenmaier, Wolfgang Lindner, and Michael Lämmerhofer Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00676 • Publication Date (Web): 06 Jun 2018 Downloaded from http://pubs.acs.org on June 6, 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 25 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

Imaging peptide and protein chirality via amino acid analysis by chiral × chiral twodimensional correlation liquid chromatography Ulrich Woiwodea, Roland Johann Reischlb, Stephan Buckenmaierc, Wolfgang Lindnerd,e, Michael Lämmerhofera, * a

Institute of Pharmaceutical Sciences, Pharmaceutical (Bio-)Analysis, University of Tübingen,

Auf der Morgenstelle 8, 72076 Tübingen, Germany b

University of Salzburg, Department of Biosciences, Bioanalytical Research Labs,

Hellbrunnerstrasse 34, 5020 Salzburg, Austria c

Agilent Technologies, Research and Development, Hewlett-Packard-Str. 8, 76337 Waldbronn,

Germany d

Lindner Consulting GmbH, Ziegelofengasse 37, 3400 Klosterneuburg, Austria

e

Institute of Analytical Chemistry, University of Vienna, Waehringerstrasse 38, 1090 Vienna,

Austria

*Author for correspondence: Prof. Dr. Michael Lämmerhofer Institute of Pharmaceutical Sciences, Pharmaceutical (Bio-)Analysis University of Tübingen Auf der Morgenstelle 8 72076 Tübingen, Germany T +49 7071 29 78793, F +49 7071 29 4565 e-mail: [email protected] 1 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

Abstract The present contribution illustrates the utilization of a chiral × chiral two dimensional liquid chromatography (2DLC) setup with tert-butylcarbamoyl quinine chiral stationary phase (CSP) in the first dimension (1D) and tert-butylcarbamoyl quinidine CSP in the second dimension (2D) to analyze FMOC-derivatized D and L amino acids from peptide hydrolysates. Hereby, in the 1D and 2D chiral separation dimensions factors such as selector and immobilization chemistry of the CSPs, mobile phase, temperature, column hardware dimensions, stationary phase supports, particle type and packing were identical. Orthogonality between 1D and 2D CSPs was solely based on their stereochemistry i.e. their opposite configurations in two chiral centers of the selector molecules, which results in inversion of enantiomer elution orders in the two dimensions. Using Coreshell CSPs for fast chromatography allowed 2D-flow rates which were 60 times faster than the

1

D-flow rates to enable online comprehensive two-dimensional

chromatography (LC×LC). Due to very similar chemoselectivity, yet opposite elution orders of corresponding enantiomers in 1D and 2D, characteristic 2D-elution patterns for achiral and chiral components can be generated. Peaks of achiral components and impurities are lined up on the diagonal line in the 2D separation space (contour plot) and thereby removed from the chromatographic space of the target enantiomers avoiding overlaps with potential interferences. Corresponding enantiomers provide cross peaks on the 2D chromatogram. Moreover, enantioselectivity of both single CSPs is combined to result in an enhanced overall 2D enantioselectivity. The concept is illustrated for the therapeutic peptides gramicidin and bacitracin. Since all amino acids give a consistent elution order as FMOC-derivatives, all enantiomers of the same configuration are either above or below the diagonal line allowing straightforward imaging of the configuration of the amino acids in peptides by the 2D chromatogram.

2 ACS Paragon Plus Environment

Page 2 of 25

Page 3 of 25 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

Introduction Monitoring stereochemistry in peptides and proteins is an integral component in quality control and structure elucidation of synthetic peptides, therapeutic peptides, non-ribosomal natural peptide products, recombinant protein therapeutics and also for clinical analysis of protein modifications.1-6 The synthetic peptide market is steadily growing. Besides common quality control procedures, such as analyzing the correct molecular mass and sequence, absence of individual amino acids’ racemization in the course of peptide synthesis has to be ascertained to ensure full structural integrity.1,2 The same applies to therapeutic peptides, where D-amino acids are frequently incorporated to increase product stability towards proteolytic degradation thus improving their pharmacokinetics.3 Searching for new bioactive compounds to combat the shortage in new potent drugs and in particular antibiotics, which can overcome the problem of multiple resistances has led to the discovery of many non-ribosomal natural peptide products via identification of non-ribosomal peptide synthetase genes in bacteria and fungi by modern genome mining tools. These non-ribosomal peptides usually contain a significant fraction of Damino acids and uncommon (non-proteinogenic) amino acids, for which no enantiomeric standards are available.4,7 Further, racemization of individual amino acids in recombinant protein therapeutics is one of the critical quality attributes and difficult to detect.5 It requires specific methodologies. Likewise, aspartic acid racemization represents one of the major types of modification that leads to an age-dependent accumulation of abnormal protein in human tissues.6 Other amino acids such as serine have some tendency to racemize8 and hence amino acid racemization in proteins might be of particular interest in the quality control of protein-based therapeutics as well as in clinical diagnostics. The goal to determine the amino acid configuration within peptides and proteins is typically accomplished by enantioselective amino acid analysis which is done after full hydrolysis of the analytes. GC-MS with chiral columns e.g. ChirasilVal4,9-11 or indirect chromatographic approaches using chiral derivatizing agents (e.g. Marfey’s reagent) followed by achiral 3 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

chromatography (mostly RPLC) are common approaches.12-14 A number of CSPs have been reported to be able to separate amino acid enantiomers directly.15-19 Derivatization is often performed for the sake of enhancing detection sensitivity and lowering limits of detection in real samples e.g. clinical samples.20-22 Chemical impurities that may occur e.g. due to derivatization, or matrix components in biological samples may complicate the comprehensive enantioselective amino acid analysis. For this reason, 2D and 3D enantioselective chromatographic procedures have been developed, typically in a multiple heart-cut methodology.23-26 In fact, with the commercial availability of instrumentation 2DLC is becoming an established tool in both research as well as routine analysis.27-29 However, it is still rarely used for comprehensive enantioselective 2DLC. In one example, offline comprehensive RP × chiral weak anion exchange hyphenated to ESI-QTOF-MS was proposed for amino acid analysis.30 In another example, two polysaccharide type chiral columns were used for online comprehensive chiral×chiral 2DLC using ultrafast enantioselective chromatography in the second dimension. In addition to exploit the orthogonality between the two hyphenated CSPs (OD-3R and OJ-3R), the authors used highly efficient chiral selectors bonded to sub-2µm fully porous and 2.7 µm Coreshell in order to achieve high-speed separations in the second dimension. These ultrafast separations in the second dimension were successfully applied to the analysis of complex mixtures of closely related pharmaceuticals and synthetic intermediates, including chiral and achiral drugs and metabolites, constitutional isomers, stereoisomers, and organohalogenated species.

31

In this context, it is worthwhile mentioning that chiral columns can deliver orthogonal

selectivity compared to the conventional achiral columns. So, the approach of using chiral columns is being extensively used in the pharmaceutical industry to achieve selectivities complementary to RPLC in common 1-dimensional LC setups. 32,33 Hence, their combination in a 2DLC setup should be beneficial. As mentioned above, there are two particular problems arising in 1DLC applications of CSPs for the enantioselective analysis of natural compounds including peptides: i) Enantiomeric (stereoisomeric) impurities or matrix components in more complex 4 ACS Paragon Plus Environment

Page 4 of 25

Page 5 of 25 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

samples are difficult to distinguish from target enantiomers, in particular when no MS detector is used. This also includes peaks emerging from side reactions of derivatizing reagents in amino acid analysis. ii) Another problem is the lack of standards for the enantiomeric impurity in many cases of natural compounds including some uncommon amino acids present in non-ribosomal natural peptides. Gasparrini and coworkers solved these problems by the “inverted chirality columns approach”.34 The employed CSPs with fully synthetic chiral selectors having opposite configurations showed reversed enantiomer elution orders for the chiral natural compound, which allowed the assignment of the minor enantiomeric impurity peak. Achiral impurities (i.e. sample matrix components) revealed the same retention factors. Since they did not change their elution position, when the inverted chiral column was used, they could so be distinguished from minor enantiomeric impurities. Other options to solve the problem include enantioselective 1DLC with online hyphenated circular dichroism detector (1DLC-CD)35 , enantioselective 1DLC coupled to high-resolution mass spectrometry (1DLC-HRMS)4,36 and 1DLC with two instruments equipped with inverted chirality columns.34 1DLC-CD may give a rough idea of the chirality, yet assignments may be less safe if there are overlaps of distinct chiral compounds. Using enantioselective 1DLC-HRMS, one could improve the specificity (with the exception of co-eluting isobars), but enantioselective ion suppression could make the approach risky, in particular if the enantiomeric compound is present in trace amounts only. The inverted chirality column approach with two instruments will not work very well with complex amino acid mixtures due to too many peak overlaps in each of the two (as shown in the Suppl. Material of this paper). In this work, we propose chiral×chiral 2DLC for enantioselective amino acid analysis of peptide and protein hydrolysates. Instead of combining two orthogonal retention principles such as RPLC and enantioselective LC we herein use two CSPs with similar chemoselectivity but orthogonal stereochemistry (opposite chiral recognition due to inverted configurations in C8 and C9 of recognition site). The proposed setup consists of online hyphenated tert-butylcarbamoyl quinine (tBuCQN) and quinidine (tBuCQD) derived CSPs (Fig. 1), which exhibit reversed elution 5 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

order for N-derivatized amino acids (such as N-fluorenylmethoxycarbonylated amino acids). With this setup characteristic elution patterns and distribution of enantiomers in the resultant 2Dchromatograms are obtained. Coreshell materials were used as support to enable fast separations necessary in the second dimension for online-comprehensive LC×LC. Hereby we evaluate, how this new 2DLC methodology can be used for stereoconfiguration profiling of chiral amino acids in peptide therapeutics providing a clear visual image which is straightforward to interpret even in absence of enantiomeric standards. The proof of principle is documented by the analysis of D-amino acids from peptide antibiotic drugs gramicidin and bacitracin.

Experimental Section Materials. Gramicidin (composition: 73.2% Gramicidin A1 and 98.2% Gramicidin A1, A2, B1, C1 and C2, according to supplier) was from Alfa Aesar (Karlsruhe, Germany) and bacitracin was provided by Sigma Aldrich (Munich, Germany). Racemic amino acids were purchased from Sigma Aldrich and Roth (Karlsruhe, Germany). 9-Fluorenylmethyl chloroformate (FMOC-Cl) as well as other chemicals (amantadine, ADAM), solvents (HPLC-grade), and mobile phase additives (such as acetic acid and ammonium acetate) were obtained from Merck (Darmstadt, Germany). Ultra-pure water was obtained from an Elga PureLab Ultra purification system (Celle, Germany). Instrumentation. An Agilent 1290 Infinity II 2D-LC Solution from Agilent Technologies (Waldbronn, Germany) was used for online-comprehensive 2DLC. The first dimension consisted of the following components: Quaternary low-pressure gradient UHPLC pump (Flexible pump, G7104A), sampling device (Multisampler, G7167B), column compartment (Multicolumn Thermostat, G7116B), UV-detector (Variable wavelength detector, G7114B) with 14 µL flow cell (G1314-60186) and a pressure relief device (pressure release kit, G4236-60010) between UV6 ACS Paragon Plus Environment

Page 6 of 25

Page 7 of 25 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

detector and 2D-interface. The second dimension was composed of the following parts: binary high-pressure gradient UHPLC pump (High Speed Pump, G7120A), valve drive (G1170A) with two-position four port valve and two 60 µL loops as interface (items in G4236A), dedicated column compartment for 2nd dimension column (Multicolumn Thermostat, G7116B) and diode array detector (G7117B) with 1 µL flow cell (G4212-60008). The 2D-chromatographic data were processed with Open Lab CDS Rev. C.01.07SR3 and LC-Image Version 2.6b3 LC×LC-HRMS from GC Image (Lincoln, NE, USA). HALO fused-core silica with 2.7 µm diameter, 1.7 µm solid core, 90 Å pore size and 150 m2/g surface area was obtained from Advanced Materials Technologies (Wilmington, DE, USA). The resulting tBuCQN-CSP (0.18 mmol/g selector coverage) and tBuCQD-CSP (0.18 mmol/g) were packed into 50 x 4 mm ID stainless steel columns by Bischoff Chromatography (Leonberg, Germany)37. Peptide hydrolysis. Bacitracin (1 mg/mL) was dissolved in 1 mL 6 M hydrochloric acid. The glass vial was sealed under nitrogen and heat-treated at 110 °C for 16 h.38 Subsequently, the sample was completely evaporated in a SpeedVac (Thermo Savant ISS110 from Thermo Scientific, Holbrook, NY, USA). The residue was filled up to the original volume (1 mL) with 0.5 M borate buffer pH 7.7 (adjusted with 10 M NaOH), stirred and centrifuged for 60 sec at 13,200 rpm. The supernatant was used for derivatization of liberated amino acids. For gramicidin (1 mg/mL), the hydrolysis step was modified by having 30% glacial acetic acid (v/v) present in 6 M hydrochloric acid for reasons of poor aqueous solubility. Derivatization. Rapid pre-column derivatization was done directly before analysis using an adapted FMOC derivatization protocol (Suppl. Fig. S1):39 First, 200 µL of reference amino acid solution (achiral and DL-amino acids separately, as specified in Tab.1, and mixture as given in Fig. 3, with 0.15 mg/mL of each individual enantiomer, in 0.5 M borate buffer pH 7.7) or 200 µL gramicidin/bacitracin hydrolysate were mixed with 200 µL of FMOC-Cl solution (1 mM in 7 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

acetonitrile) directly in a glass vial. The reaction mixture was allowed to react for 1 min at room temperature. Excess of FMOC-Cl was then quenched by reaction with 200 µL ADAM (40 mM in acetonitrile-water, 3:1 (v/v)).

Online comprehensive 2DLC. Conditions for comprehensive chiral×chiral 2DLC were as follows: Dimension, 1; column, tBuCQN-CSP; mobile phase, methanol/acetic acid/ammonium acetate (98:2:0.5, v/v/w); flow rate, 0.05 mL/min; injection, 5 µL of derivatized amino acids samples; column temperature, 25 °C; detection, UV at 262 nm with 40 Hz sampling frequency; transfer volume, 60 µL (concurrent filling and injecting); modulation time, 60 seconds (83.3% loop fill); dimension, 2; column, tBuCQD-CSP; mobile phase, methanol/acetic acid/ammonium acetate (98:2:0.5, v/v/w); flow rate, 3 mL/min; column temperature, 25 °C; detection, DAD at 262 nm with 160 Hz sampling frequency.

Results and discussion 2DLC Setup. The comprehensive 2DLC setup40 with a UHPLC pump both in first (1D) and second dimension (2D), interfaced via a dual loop, two-position/four-port valve for modulation (Suppl. Fig. S2) was employed herein to enantioselectively separate amino acid hydrolysates. This interface collects 1D-effluent in one loop while simultaneously analysis of an earlier collected fraction through the other loop occurs. The valve was equipped with loops of 60 µL volume each. The 1D-chiral separation was performed with a mobile phase composed of methanol/acetic acid/ammonium acetate (98:2:0.5, v/v/w) at a flow rate of 0.05 mL/min and the modulation time (time between each valve switch and period available to develop the 2D-chiral separation, respectively) was adjusted to 60 s. This corresponded to a loop filling of 83.3% so that we can assume that no 1D-eluate fraction was lost, which makes this 2D-analysis truly comprehensive. 8 ACS Paragon Plus Environment

Page 8 of 25

Page 9 of 25 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

D and 2D separations. In the present comprehensive dual enantioselective LC×LC design, 1D

1

and 2D-columns are ideally packed with CSPs that are chemically completely identical, but differ solely in their stereochemistries. This can be accomplished when the chiral selectors of 1D and 2

D columns have opposite absolute configurations but are otherwise obtained by the same

immobilization

chemistry

and

surface

coverage

(achieved

through

identical

selector

concentrations in the reaction mixture). In the present example, this was realized by chiral weak anion-exchanger CSPs, which were based on tert-butylcarbamoyl quinine (tBuCQN) and quinidine (tBuCQD) (Figure 1). These two chiral selectors have opposite absolute configurations in position 8 and 9. Although the remaining stereogenic centers on QN (8S, 9R) and QD (8R, 9S) exhibit the same configuration (1S, 3R, 4S) and the two chiral selectors are actually diastereomers, chromatographically they behave like enantiomers (therefore often termed “pseudo-enantiomers”; note, the pseudo-enantiomeric behavior has its origin in the fact that chiral recognition occurs in a binding domain around C8/C9 where the two alkaloids have opposite stereochemistries).41 As a consequence, they exhibited reversed elution orders for Nderivatized amino acids such as FMOC-amino acids (Suppl. Table S1) with largely comparable separation factors on tBuCQN- and tBuCQD-based CSPs when they were run with the same mobile phase (herein composed of methanol/acetic acid/ammonium acetate; 98:2:0.5, v/v/w) (Table 1). Indeed, it is a significant advantage to use the identical mobile phase in 1D and 2D separation, like herein. It eliminates the incompatibility issue often observed between 1D- and 2

D-mobile phase such as in RPLC-HILIC or RPLC-chiral LC which can be a major obstacle in

2DLC separations. Peak distortions may occur when the 1D effluent that constitutes the sample matrix of the 1D fractions becomes injected into the 2D column and has higher elution strength than the (initial)

2

D mobile phase composition. This for example would complicate the

combination of the current enantioselective LC separation with its totally organic mobile phase in 1

D and a reversed phase separation in 2D. However, these shortcomings can be partially 9 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

circumvented by the application of modulation techniques between the two separation dimensions, such as stationary phase assisted modulation42 or the use of an active solvent modulation valve.43,44 Important parameters to be considered in comprehensive LC×LC are the dimensions of the 1D and 2D columns. Ideally, the 1D column has a narrower inner diameter, which can be run at a low volumetric flow rate (note, in the current embodiment the 1D column is run below the optimal flow velocity). In contrast, the 2D column should be short with a wider diameter so that the 2D separations can be fast, with typically high flow rates but without volume overloading from the 1D sample.31 In spite of these design rules, in the present proof of principle demonstration 1D and 2

D columns had identical dimensions (50 x 4 mm ID), simply because they were readily

available. The two utilized columns were highly comparable in terms of selector coverage and similarity of surface chemistry. In order to allow fast separations in the second dimension, 2.7 µm superficially porous silica particles (SPPs) were used as support of the CSPs in 1D and 2D columns.37 Moreover, the shallow van Deemter curves related to these SPP materials45 aided to ensure comparable separation performances over a wide range of linear flow velocities. Hence, the LC×LC experiments were established to enable a minute-timescale separation in the 1D- and a second-timescale in the subsequent 2D-separation, by applying a 60 times faster 2D-flow (1F, 0.05 mL/min; 2F, 3 mL/min). As outlined in earlier publications, tBuCQN- and tBuCQD-CSPs exhibit exceptional enantioselectivity for various N-derivatized amino acids (AAs) with consistent elution orders.20-22 Herein, we used FMOC-Cl to derivatize amino acids of peptide hydrolysates (Fig. S1). As documented in Table 1, the D-enantiomers of FMOC-AAs exhibited stronger retention on the 9(S)-configurated selector (tBuCQD) and L-enantiomers on the 9-(R)-configurated selector (tBuCQN).41 The chromatograms given in Fig. 2 reflect the utilization of this elution order inversion in a 2Dchromatographic environment. The chromatogram in Fig. 2a shows the detector signal of the 1D 10 ACS Paragon Plus Environment

Page 10 of 25

Page 11 of 25 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 of racemic phenylalanine after FMOC-derivatization. The marked (black, red, golden) parts of the chromatogram indicate fractions #15, #25 and #30 with the highest concentrations of the three analytes (FMOC-ADAM, D and L enantiomers of FMOC-Phe), which have been collected into the loops and transferred to the second dimension The corresponding

2

D

chromatograms are overlaid on the right-hand side (Fig. 2b). FMOC-D-Phe (#25, red colored) isolated from and eluting before FMOC-L-Phe (#30; golden colored) in 1D is stronger retained on the 2D column and vice versa. The peak of the achiral reagent adduct (used to quench the excess of reagent), FMOC-amantadine (FMOC-ADAM, #15, black), eluted close to t0 on both 1D and 2D columns.

Characteristic elution pattern in enantioselective 2D-correlation LC×LC. One of the fundamental theorems in comprehensive 2DLC is the quest for sufficient orthogonality in the two separation dimensions, in order to maximize the 2D-separation space and to accordingly increase the effective peak capacity delivered by the method. In the present case, however, we combined both orthogonality (related to stereoselectivity) and similarity (in terms of chemoselectivity) principles. As a matter of consequence, we observed a very useful complementarity in 1D and 2D retention in spite of a high correlation of the retention factors in the two dimensions as becomes clearly evident from the correlation matrix shown in Table 2. In fact, this particular 2D design yields very characteristic elution patterns, as schematically outlined in Figure 3a, which may be advantageous in several instances. This becomes particularly clear if a sample is to be considered, which contains a more complex matrix than just the target peptide. After hydrolysis, the resulting amino acids are derivatized (here with FMOC-Cl) for chiral separation. During the derivatization excess of reagent is used, which - for sake of reproducibility - is quenched at the end of the reaction with a component providing a non-reactive derivative. In a 1D chiral separation followed by UV detection, the reagent peak and reagent hydrolysate peaks as well as minor achiral matrix components 11 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

(impurities) can severely complicate the enantioselective analysis because these components may overlap with target enantiomer peaks. In the current enantioselective 2D-correlation LC×LC design, these achiral components, which are of no interest, are removed from the 2D separation space of the target N-derivatized chiral amino acids. Due to similar (equal) chemoselectivities on the 1D and 2D column, these elute on the diagonal line (parity line; indicating equal retention on 1

D and 2D column) and do not longer interfere with the enantioselective analysis of the chiral

amino acids. The same is true for achiral amino acids (e.g. glycine, β-alanine) and also chiral amino acids, which cannot be resolved into enantiomers (if there are any). In sharp contrast, complementarity exists in the 1D and 2D columns for the chiral amino acids for which the chiral selector system delivers enantioselectivity. Since their elution order is reversed in 1D and 2D separation dimensions (due to the use of columns with the opposite configurations), they spread out of the diagonal line into the 2D-separation space. This resolves them from the interfering peaks. Moreover, since typically the elution order is consistent within a particular series of N-derivatized amino acids (e.g. for FMOC-AAs D before L on tBuCQN-CSP and L before D on tBuCQD-CSP46; see Suppl. Table S1) a consistent pattern appears on the 2Dseparation space: for FMOC-amino acids in tBuCQN×tBuCQD 2DLC, L-amino acids yield peaks below the diagonal line, while D-amino acids exhibit peaks above the diagonal line. If a racemic mixture is present, we observe both cross peaks. If the columns in 1D and 2D are exchanged (reversed column sequence), enantiomeric peaks are flipped i.e. L-enantiomers above and Denantiomers below diagonal line with the column combination tBuCQD×tBuCQN (see Suppl. Fig. S6). This very characteristic elution pattern can be of particular utility for natural product analysis such as non-ribosomal peptides from various bacteria or fungi.4 Very often such natural peptides contain uncommon non-proteinogenic amino acids for which no commercial reference standards for D and L-amino acids are available. If the chiral recognition mechanism of Nderivatized amino acids is very consistent, which is the case for N-acyl amino acids on quinine and quinidine carbamate CSPs (Suppl. Table S1-a and S1-b)21, the amino acid configurations 12 ACS Paragon Plus Environment

Page 12 of 25

Page 13 of 25 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

can be simply assigned based on their position above the diagonal parity line (D-amino acid retention space) or below it (L-amino acids retention space). Such assignments of course have then to be validated by complementary techniques. This can be supported by exemplary retention data summarized in Table 1. It can be seen that, without exception, L-amino acids (S-enantiomers except for Cys for which it is the R-enantiomer) are stronger retained on the tBuCQN 1D-column, whereas they elute first on the tBuCQD 2Dcolumn, as outlined above. Table 1 also reveals that separation factors were not completely identical on tBuCQN and tBuCQD CSPs, which can be ascribed to their diastereomeric nature. However, this does not compromise the outlined principle of the characteristic elution patterns, as can be seen by the practical implementation depicted in Fig. 3b for a set of amino acid racemates (note, the same plot in terms of normalized retention factors47 gives an even more clear representation as can be seen in Fig. 3c). This test mixture contained 4 amino acid racemates (Leu L, Val V, Phe F, Trp W; dashed letter indicates D-enantiomer) as well as 2 achiral amino acids (Gly G, β-alanine βA). It was obvious that there was a pronounced correlation of both separations with this setup (strong retention on the 1D column was generally accompanied by strong retention in 2D) due to very similar chemoselectivity (Table 2). For this reason, the peaks for the achiral amino acids (G and βA) and the reagent peak lay on the diagonal parity line as prompted by the scheme in Fig. 3a. Chiral amino acids in contrast, did not elute on this parity line due to orthogonal stereoselectivity on the 1D and 2D column. Reflecting the inversion of elution orders in one-dimensional representations, D-enantiomers were distributed to the upper left and L-enantiomers to the lower right of the diagonal in the 2Dchromatogram. This opened up the possibility to easily recognize D-amino acids in complex samples, even when no enantiomeric standards were available. This 2DLC system is also sufficiently robust for quantitative analysis (which however was not the goal in this proof of principle study) as exemplified for a test mixture by a preliminary calibration and validation (Suppl. Material). Retention time stability was always better than 0.13% RSD both 13 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

in 1D and 2D (Table S2). Calibration in the range of 0.5 to 50 µM gave sufficient linearity (r > 0.9999) (Table S3), repeatabilities for peak volume for three QC samples with 0.5, 5, and 50 µM were always lower than 1% RSD, except for the 0.5 µM concentration for which they were < 4% RSD (Table S4). Accuracies (as measured by recoveries of these QC samples) were between 98 and 100%. Except for one 0.5 µM QC which showed a recovery of 116.5% (Table S4). LODs were around 0.15 µM (corresponding to about 1 pmol actually injected amino acids) (Table S5). In conclusion, this comprehensive enantioselective LC × enantioselective LC gave results that may be termed higher-order enantioselective 2D-chromatograms, as the distribution within 2D space provides key information. The 2D chromatogram showed some analogy to representations of 2D-NMR experiments (such as NOESY and COSY) as signals outside the diagonal provided information on specific features of the solutes, in this case chirality and configuration, respectively.

Application for stereoconfiguration monitoring of gramicidin. The practical applicability of the method is documented by the application to amino acid composition analysis of gramicidin A1 (Fig. 4a). It is an active pharmaceutical ingredient in various pharmaceutical formulations. Gramicidin is a linear membrane channel forming pentadecapeptide antibiotic that is produced by Bacillus brevis via the nonribosomal pathway. It consists of 15 hydrophobic amino acids with alternating L- and D-configuration forming a β-helix-like structure with an N-terminal formyl residue and a C-terminal ethanolamine. Other naturally occurring isoforms, gramicidin B and C, have either phenylalanine or tyrosine replacing tryptophan at position 11, respectively. The peptide was hydrolyzed in boiling acid similar to earlier published procedures48,49 with acetic acid to completely dissolve the considerably lipophilic peptide.50 Artificial racemization during 16 h hydrolysis is typically minor and minor traces of enantiomeric impurities created due to racemization in the course of hydrolysis do usually not negatively impact the stereoconfiguration determination, which is the focus of the present study (note, if the goal is determination of trace 14 ACS Paragon Plus Environment

Page 14 of 25

Page 15 of 25 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

D-amino acids, this minor racemization has to be corrected for e.g. through use of DCl/D2O and MS detection).51 Fig. 4b shows the 2D chromatogram for the hydrolyzed peptide gramicidin. It can be seen that the results are in agreement with the expected structure. Only two D-amino acids were found in the sample corresponding to FMOC-D-valine (V’) and FMOC-D-leucine (L’) (above the diagonal line spanned by reagent peak and Gly peak), while all other amino acids were present in the L-configuration. The additional unassigned peaks corresponded to the gramicidin isoforms which were observed in minor quantities only. In Suppl. Fig. S4, the chromatograms as they would appear on a 1DLC separation are displayed for comparison. It can be seen that overall enantioselectivity was greatly improved by the 2D setup. This example clearly demonstrated the applicability of this 2D method in practice.

Application for stereoconfiguration determination of bacitracin. Furthermore, the applicability has been tested for the antibacterial drug bacitracin (Fig. 5a). Bacitracin is a cyclic peptide produced by organisms of the licheniformis group of bacillus subtilis via nonribosomal peptide synthetases. It is widely used in topical applications as an antibacterial agent against a variety of gram-positive and a few gram-negative organisms. The 2D chromatogram of hydrolyzed bacitracin is shown in Fig. 5b. Once more it is striking that there are a number of D-amino acids contained in the sample. D-Orn (O’), D-Asp (D’), D-Glu (E’), and D-Phe (F’) peaks were expected (all above diagonal line). His (H), Leu (L), Ile (I), and Lys (K) were found in their L-configuration (below diagonal line). Furthermore, an L-aspartic acid (D) peak is visible in the 2D-chromatogram, although this amino acid was not present in the intact peptide. It may have originated from the deamidation of asparagine (N) during hydrolysis. Besides, there were some more uncommon patterns observed. The thiazoline ring was hydrolyzed releasing an Ile (I) and a Cys (C) at the N-terminus. It thus actually represented a masked Cys residue, which should have been present in the L-configuration (R-enantiomer). It was doubly labelled with FMOC-Cl and thus exerted strong retention. This caused some 15 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

problems, i.e. a so-called wrap-around in the 2D-chromatogram indicated as (wC) in Figure 5b. However, such a wrap-around can easily be recognized by larger peak-widths in the second dimension and an additional 60 seconds/modulation between absorbance maximum in first and second dimension detector. Again, when the diagonal line was shifted parallel to the first diagonal line, it became evident that the Cys peak emerged below this new diagonal line thus indicating L-configuration for the masked Cys, as expected. Somewhat unexpectedly we observed another intensive peak above the diagonal line, which was assigned to D-allo-isoleucine (aI’). Upon acidic hydrolysis, the N-terminal isoleucine residue of bacitracin A underwent epimerization, yielding D-allo-isoleucine besides L-isoleucine in the hydrolyzed mixture. This may be explained by the protonation of nitrogen in the thiazoline ring, subsequent deprotonation of the α-carbon of the masked L-isoleucine residue and flipping of the electron-pairs due to the electron-withdrawing effect of the protonated nitrogen. The retrograde steps yielded both epimeric forms, L-Ile (I) and D-allo-Ile (aI’).52 The latter appeared as an additional peak above the diagonal, as mentioned above.

Conclusions The presented chiral × chiral two-dimensional correlation liquid chromatography method with chiral stationary phases based on the same selector chemistry but with opposite stereoconfigurations (here quinine and quinidine carbamates) give higher-ordered chiral LC×LC 2D-chromatograms which can provide a vivid image of the stereochemistry of a peptide sample. It results in valuable and easily accessible stereochemical information on unknown and complex samples found during pharmaceutical development, impurity analysis and analysis of natural compounds such as nonribosomal peptides, synthetic peptides and racemization in therapeutic proteins. Proteins and common (e.g. synthetic) peptides typically are constituted by all-L-amino acids. Since it is considered that only specific amino acids have a tendency to racemize, the method will be well applicable to such more complex structures because the separation space of 16 ACS Paragon Plus Environment

Page 16 of 25

Page 17 of 25 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

D-amino acids will be not overcrowded. However, more problematic would be the simultaneous quantification of L-amino acids in addition to the D-amino acids. Yet, MS detection can resolve this problem and then this assay format can be more generic. The approach worked well in this case of amino acids, but some other analyte mixtures would require a different set of columns and eluents. In some cases, however, it is difficult to get commercially available columns with the opposite configurations, e.g. polysaccharide and macrocyclic antibiotics CSPs. This limits the generality of the approach a bit.

Supporting information FMOC-derivatization of amino acids (Fig. S1); Flow paths and valve positions of the dual 4port/2-position valve (Fig. S2); 1D- and 2D-traces of 2DLC separations (Figs. S3, S4 and S5) shown in Figs. 3, 4 and 5; Chromatographic data for FMOC-derivatives of all proteinogenic and several non-proteinogenic amino acids (Tab. S1a,b); Comparison of tBuCQN×tBuCQD, tBuCQD×tBuCQN and tBuCQN×tBuCQN column configuration (Fig. S6 a-f); Repeatability of retention times (Tab. S2); Preliminary validation and calibration (Tab. S3 and S4); LOD and LOQ (Tab. S5)

Notes The authors declare no competing financial interest.

Acknowledgements M. L. is grateful to Agilent Technologies for financial support of this research. 17 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

References (1) D'Hondt, M.; Bracke, N.; Taevernier, L.; Gevaert, B.; Verbeke, F.; Wynendaele, E.; De Spiegeleer, B. J. Pharm. Biomed. Anal. 2014, 101, 2-30. (2) Liang, C.; Behnam, M. A. M.; Sundermann, T. R.; Klein, C. D. Tetrahedron Lett. 2017, 58, 2325-2329. (3) Szajek, A. Pharm. Tech. 2014, 38. (4) Gerhardt, H.; Sievers-Engler, A.; Jahanshah, G.; Pataj, Z.; Ianni, F.; Gross, H.; Lindner, W.; Lammerhofer, M. J. Chromatogr. A 2016, 1428, 280-291. (5) Zhang, Q.; Flynn, G. C. J. Biol. Chem. 2013, 288, 34325-34335. (6) Ritz-Timme, S.; Collins, M. Ageing Res. Rev. 2002, 1, 43-59. (7) Süssmuth, R. D.; Mainz, A. Angew. Chem. Int. Ed. Engl. 2017, 56, 3770-3821. (8) Reischl, R. J.; Lindner, W. J. Pharm. Biomed. Anal. 2015, 116, 123-130. (9) Frank, H.; Nicholson, G. J.; Bayer, E. J. Chromatogr. Sci. 1977, 15, 174-176. (10) Schurig, V. J. Chromatogr. B 2011, 879, 3122-3140. (11) Pätzold, R.; Brückner, H. In Quantitation of Amino Acids and Amines by Chromatography – Methods and Protocols, Molnár-Perl, I., Ed.; Elsevier: Amsterdam, 2005, pp 98-118. (12) Bhushan, R.; Bruckner, H. J. Amino Acids 2004, 27, 231-247. (13) Ilisz, I.; Berkecz, R.; Peter, A. J. Pharm. Biomed. Anal. 2008, 47, 1-15. (14) Livnat, I.; Tai, H. C.; Jansson, E. T.; Bai, L.; Romanova, E. V.; Chen, T. T.; Yu, K.; Chen, S. A.; Zhang, Y.; Wang, Z. Y.; Liu, D. D.; Weiss, K. R.; Jing, J.; Sweedler, J. V. Anal. Chem. 2016, 88, 11868-11876. (15) Armstrong, D. W.; Liu, Y.; Ekborgott, K. H. Chirality 1995, 7, 474-497. (16) Hoffmann, C. V.; Pell, R.; Lammerhofer, M.; Lindner, W. Anal. Chem. 2008, 80, 8780-8789. (17) Hyun, M. H.; Song, Y.; Cho, Y. J.; Kim, D. H. J. Chromatogr. A 2006, 1108, 208-217. (18) Davankov, V. A. In Chiral Separations. Methods in Molecular Biology, vol 243., Gübitz G.; M.G., S., Eds.; Humana Press, 2004, pp 207-216. (19) Ilisz, I.; Péter, A.; Lindner, W. Trends Anal. Chem. 2016, 81, 11-22. (20) Ilisz, I.; Aranyi, A.; Peter, A. J. Chromatogr. A 2013, 1296, 119-139. (21) Lämmerhofer, M.; Lindner, W. In Adv. Chromatogr., Grushka, E.; Grinberg, N., Eds.; CRC Ress, Taylor & Francis Group: Boca Raton, FL, 2008, p 1− 109. (22) Oyama, T.; Negishi, E.; Onigahara, H.; Kusano, N.; Miyoshi, Y.; Mita, M.; Nakazono, M.; Ohtsuki, S.; Ojida, A.; Lindner, W.; Hamase, K. J. Pharm. Biomed. Anal. 2015, 116, 71-79. (23) Miyoshi, Y.; Oyama, T.; Itoh, Y.; Hamase, K. Chromatography 2014, 35, 49-57. (24) Ianni, F.; Sardella, R.; Lisanti, A.; Gioiello, A.; Cenci Goga, B. T.; Lindner, W.; Natalini, B. J. Pharm. Biomed. Anal. 2015, 116, 40-46. (25) Weatherly, C. A.; Du, S.; Parpia, C.; Santos, P. T.; Hartman, A. L.; Armstrong, D. W. ACS Chem. Neurosci. 2017, 8, 1251-1261. (26) Welsch, T.; Schmidtkunz, C.; Muller, B.; Meier, F.; Chlup, M.; Kohne, A.; Lammerhofer, M.; Lindner, W. Anal. Bioanal. Chem. 2007, 388, 1717-1724. (27) Stoll, D. R.; Li, X.; Wang, X.; Carr, P. W.; Porter, S. E.; Rutan, S. C. J. Chromatogr. A 2007, 1168, 3-43. (28) Iguiniz, M.; Heinisch, S. J. Pharm. Biomed. Anal. 2017, 145, 482-503. (29) Pirok, B. W. J.; Gargano, A. F. G.; Schoenmakers, P. J. J. Sep. Sci. 2018, 41, 68-98. (30) Woiwode, U.; Neubauer, S.; Kaupert, K.; Lindner, W.; M., L. Supplement to LCGC Eur./LCGC North Am. 2017, 30, 34-42. (31) Barhate, C. L.; Regalado, E. L.; Contrella, N. D.; Lee, J.; Jo, J.; Makarov, A. A.; Armstrong, D. W.; Welch, C. J. Anal. Chem. 2017, 89, 3545-3553. (32) Regalado, E. L.; Welch, C. J. Trends Anal. Chem. 2015, 67, 74-81. 18 ACS Paragon Plus Environment

Page 18 of 25

Page 19 of 25 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

(33) Barhate, C. L.; Breitbach, Z. S.; Pinto, E. C.; Regalado, E. L.; Welch, C. J.; Armstrong, D. W. J. Chromatogr. A 2015, 1426, 241-247. (34) Badaloni, E.; Cabri, W.; Ciogli, A.; Deias, R.; Gasparrini, F.; Giorgi, F.; Vigevani, A.; Villani, C. Anal. Chem. 2007, 79, 6013-6019. (35) Takatori, K.; Toyama, S.; Fujii, S.; Kajiwara, M. Chem. Pharm. Bull. 1995, 43, 1797-1799. (36) Cho, S.; Lee, I.; Kim, J.; Hwang, I. S.; Kim, S.; Han, K. M.; Kim, J. W.; Lee, J. H.; Park, Y. K.; Han, S. Y.; Chae, K. R. Bull. Korean Chem. Soc. 2011, 32, 4458-4461. (37) Reischl, R. J.; Hartmanova, L.; Carrozzo, M.; Huszar, M.; Frühauf, P.; Lindner, W. J. Chromatogr. A 2011, 1218, 8379-8387. (38) Matsubara, H.; Sasaki, R. M. Biochem. Biophys. Res. Com. 1969, 35, 175-181. (39) Gustavsson, B.; Betnér, I. J. Chromatogr. A 1990, 507, 67-77. (40) Stoll, D. R.; Carr, P. W. Anal. Chem. 2017, 89, 519-531. (41) Maier, N. M.; Nicoletti, L.; Lämmerhofer, M.; Lindner, W. Chirality 1999, 11, 522-528. (42) Vonk, R. J.; Gargano, A. F.; Davydova, E.; Dekker, H. L.; Eeltink, S.; de Koning, L. J.; Schoenmakers, P. J. Anal. Chem. 2015, 87, 5387-5394. (43) Gargano, A. F.; Duffin, M.; Navarro, P.; Schoenmakers, P. J. Anal. Chem. 2016, 88, 17851793. (44) Stoll, D. R.; Shoykhet, K.; Petersson, P.; Buckenmaier, S. Anal. Chem. 2017, 89, 92609267. (45) Patel, D. C.; Breitbach, Z. S.; Yu, J.; Nguyen, K. A.; Armstrong, D. W. Anal. Chim. Acta 2017, 963, 164-174. (46) Lajko, G.; Grecso, N.; Toth, G.; Fulop, F.; Lindner, W.; Peter, A.; Ilisz, I. Molecules 2016, 21, 1579. (47) Gilar, M.; Olivova, P.; Daly, A. E.; Gebler, J. C. Anal. Chem. 2005, 77, 6426-6434. (48) Christensen, H. N.; Hegsted, D. M. J. Biol. Chem. 1945, 158, 593-600. (49) Synge, R. L. M. Biochem. J. 1945, 39, 351-355. (50) Kelkar, D. A.; Chattopadhyay, A. Biochim. Biophys. Acta 2007, 1768, 2011-2025. (51) Frank, H.; Woiwode, W.; Nicholson, G.; Bayer, E. Liebigs Ann. Chem. 1981, 354-365. (52) Konigsberg, W.; Hill, R. J.; Craig, L. C. J. Org. Chem. 1961, 26, 3867-3871.

19 ACS Paragon Plus Environment

Analytical Chemistry

Figures:

3R 4S 3R 9R

1S

8S 1S

4S 9S 8R

tBuCQN-CSP

tBuCQD-CSP

Fig. 1: Structures of tert-butyl-quinine and -quinidine carbamate selectors immobilized to Coreshell silica.

#25 D

FMOC-ADAM

#30 L 2D-Absorbance

FMOC-ADAM

#15

tBuCQD-CSP

(b)

1t < 1t D L

2t < 2t L D L

D

2t o

#30 #25 #15

1t

[min]

2t

Injection of subsequent fraction

tBuCQN-CSP

(a)

1D-Absorbance

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

2t o

[sec]

Fig. 2: Switch of elution order exemplified by FMOC-Phe enantiomers. First eluting enantiomer is D on tBuCQN (1D) and L on tBuCQD (2D). Conditions are as specified in experimental section.

20 ACS Paragon Plus Environment

Page 20 of 25

60

(b)

W‘

50

60

(c)

W‘

40

W

V‘

30

L‘ βA

20

F

G

L

Achiral matrix components

V Achiral amino acids

Reagent peaks 10

F‘

40

V‘ L‘

30

βA

20

W

G

F

LV

L-amino acids

0

10

0 0

10 20 30 40 50 1D retention time [min] – “R-column”

1.0

W‘ 0.8

0.6

F‘

V‘ F 0.2

0.0 0

10 1D

20 30 40 50 retention time [min] – tBuCQN

W

L‘

0.4

2D

D-amino F‘ acids

retention time [sec] - tBuCQD

50

2D

retention time [sec] – “S-column”

(a)

2D

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

– normalized retention factor - tBuCQD

Page 21 of 25

*

G L V

0 1D

0.2 0.4 0.6 0.8 – normalized retention factor – tBuCQN

Fig. 3: (a) General characteristic elution pattern of chiral × chiral 2DLC of with chemically equivalent columns differing just in the absolute configuration of the chiral selectors in 1D and 2D columns. (b) Characteristic elution pattern exemplified for an amino acid test mixture (achiral beta-alanine (βA) and glycine (G); racemic Leucine (L), Valine (V), Phenylalanine (F) and Tryptophan (W); Dashed letter indicates D-enantiomer) using tBuCQN×tBuCQD online LC×LC 2D-setup. For corresponding 1D- and 2D-traces see Suppl. Fig. S3. (c) The same chromatogram presented with normalized retention factors (note, retention of FMOC-β-Ala –marked with asterism- was taken as minimal retention for normalization because retention times of this compound were more robust as compared to the earlier eluted reagent peak which is present in excess but shifted a little bit in dependence on concentration of amino acids in the sample).

21 ACS Paragon Plus Environment

1

Analytical Chemistry

(a) HCO

(b)

Gramicidin A1 V

G

A

L‘

A

V‘

V

V‘

W

L‘

W

L‘

W

L‘

W

NCH2CH2OH H

60

retention time [sec] - tBuCQD

50

40

W V‘ L‘

30

G AV

20

2D

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

10

0 0

10 1D

20 30 40 50 retention time [min] – tBuCQN

Fig. 4: Application of tBuCQN×tBuCQD 2DLC separation to gramicidin. (a) Structure of gramicidin and (b) 2D-chromatogram. Dashed letter indicates D-enantiomer. Conditions are as specified in experimental section. For corresponding 1D- and 2D-traces see Suppl. Fig. S4 (Note, the additional blops at 1t = 31 min / 2t = 31 min are L-Phe and L-Tyr originate from the minor Gramicidin components B1 (L-Phe) and C1/C2 (L-Tyr). The other blop at 1t = 33 min / 2t = 35 min is an oxidation product of tryptophan (oxyindolylalanine). The respective FMOC-derivative was detected in MS-analysis).

22 ACS Paragon Plus Environment

Page 22 of 25

Page 23 of 25

(a)

L

E‘

I

α K ε

O‘

I

H

F‘

D‘

N

(b) 60 O‘

retention time [sec] - tBuCQD

50

D‘

40

E‘ F‘

30

D K

aI‘

20

WC

2D

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

10

LI H

0 0

10

20 1D

30 40 50 60 retention time [min] – tBuCQN

70

80

Fig. 5: Application of tBuCQN×tBuCQD 2DLC separation to bacitracin. (a) Structure of bacitracin and (b) 2D-chromatogram. Dashed letter indicates D-enantiomer.

Conditions are as

specified in experimental section. For corresponding 1D- and 2D-traces see Suppl. Fig. S5 (note, the additional blop at 1t = 31 min / 2t = 21 min is L-Phe, probably generated by racemization of DPhe (F’) or due to a minor variant. D-Leu (L’, 1t = 18 min, 2t = 17 min) is present for the same reason. Bis-labeled L-Histidine (H, 1t = 59 min, 2t = 46 min) was identified by injecting FMOClabeled L-His and can be seen along with mono-labeled Histidine (H); The parallel is obtained as extension of the diagonal line above the 2D-contour plot. It is then just moved vertically down to the 1D chromatogram axis in parallel to the diagonal line. In general, such wrap arounds should be avoided which can be achieved here e.g. by use of an ionic strength gradient).

23 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 24 of 25

1

Table 1: Retention data obtained from injection of single DL-amino acid derivatives or from

2

achiral test compounds.

3 Dimension 1

1 D - tBuCQN 1 1 k or 1k D kL α DL

FMOC-derivative

2 D - tBuCQD 2 2 k or 2k D kL α DL

2

kL/kD

kD /kL

1 D/ 2D Diastereoselectivity 2

1

kL( D)/kD ( D)

2

tBuCQN × tBuCQD 12

α DL b

1

kD ( D)/kL( D)

TTBBa Ethanolamine

0.00

-

-

0.00

-

-

-

-

-

0.42

-

-

0.22

-

-

-

-

-

ADAM

0.38

-

-

0.27

-

-

-

-

-

beta -Alanine (βA) Glycine (G)

0.52

-

-

0.62

-

-

-

-

-

0.99

-

-

1.12

-

-

-

-

-

Alanine (A)

0.74

1.03

1.39

1.19

0.81

1.47

1.09

1.15

2.03

Aspartic acid (D)

2.36

2.72

1.15

3.20

3.14

1.02

1.33

1.18

1.54

Cysteine (C)

4.98

6.36

1.28

8.18

5.55

1.47

1.11

1.29

1.95

Glutamic acid (E)

1.54

2.27

1.47

2.69

1.85

1.46

1.20

1.19

2.07

Histidine (H)

0.67

1.03

1.54

0.84

0.57

1.47

0.85

0.82

2.13

Isoleucine (I)

0.73

1.33

1.82

1.56

0.73

2.14

0.99

1.17

2.81

allo -Isoleucine (aI)

0.75

1.42

1.89

1.56

0.74

2.09

0.99

1.10

2.82

Leucine (L)

0.65

1.07

1.65

1.20

0.66

1.81

1.02

1.12

2.45

Lysine (K)

2.29

3.20

1.40

3.44

2.29

1.50

1.00

1.08

2.05

Ornithine (O)

2.78

3.56

1.28

3.92

2.78

1.41

1.00

1.10

1.91

Phenylalanine (F)

1.44

2.00

1.39

2.40

1.60

1.50

1.11

1.20

2.05

Tryptophan (W)

2.38

3.59

1.51

3.81

2.52

1.51

1.06

1.06

2.13

Tyrosine (Y)

1.39

2.01

1.45

2.26

1.50

1.51

1.08

1.12

2.09

Valine (V)

0.71

1.29

1.84

1.41

0.72

1.96

1.02

1.09

2.69

4 5 6

a

1,3,5-Tri-tert-butylbenzene 0.1 mg/mL in mobile phase as t0 marker, not derivatized

7

b

calculated from 1αDL and 2αDL: 12αDL = ((1αDL)2 + (2αDL)2)0.5

8

24 ACS Paragon Plus Environment

Page 25 of 25

9

Table 2: Correlation matrix (Pearson correlation coefficient) of retention factors on tBuCQN and tBuCQD CSPs

10 11

k D (tBuCQN) k L (tBuCQN) k D (tBuCQD) k L (tBuCQD) 0.961 0.988 0.975 1.000

12

k D (tBuCQN) k L (tBuCQN) k D (tBuCQD) k L (tBuCQD)

13

** All correlations are significant on the P-level of 0.01 (2-side).

0.961

1.000

0.979

0.977

0.988

0.979

1.000

0.975

0.975

0.977

0.975

1.000

14 15 16

For TOC only

17 18

retention time [sec] – “S-column”

60

2D

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

W‘

50

40

W

D-amino L-amino acids acids

V‘

30

L‘

F

G

Achiral matrix components

βA

20

V L

Achiral amino acids Reagent peaks

10

0 0

19

F‘

Damino acids

Lamino acids

2D

1D

10 20 30 40 50 1D retention time [min] – “R-column”

20

25 ACS Paragon Plus Environment