The Reciprocal Principle of Selectand-Selector-Systems in ... - MDPI

0 downloads 0 Views 9MB Size Report
Nov 15, 2016 - can take one chiral stationary phase (CSP) to develop others: if a CSP ... reciprocal principle involved a series of CSPs developed for the ...
molecules Review

The Reciprocal Principle of Selectand-Selector-Systems in Supramolecular Chromatography † Volker Schurig Institute of Organic Chemistry, University of Tübingen, Auf der Morgenstelle 18, 72076 Tübingen, Germany; [email protected]; Tel.: +49-7071-62605 † This contribution is dedicated to Professor Stig Allenmark at his 80th anniversary. Academic Editor: Yoshio Okamoto Received: 5 October 2016; Accepted: 7 November 2016; Published: 15 November 2016

Abstract: In selective chromatography and electromigration methods, supramolecular recognition of selectands and selectors is due to the fast and reversible formation of association complexes governed by thermodynamics. Whereas the selectand molecules to be separated are always present in the mobile phase, the selector employed for the separation of the selectands is either part of the stationary phase or is added to the mobile phase. By the reciprocal principle, the roles of selector and selectand can be reversed. In this contribution in honor of Professor Stig Allenmark, the evolution of the reciprocal principle in chromatography is reviewed and its advantages and limitations are outlined. Various reciprocal scenarios, including library approaches, are discussed in efforts to optimize selectivity in separation science. Keywords: supramolecular chromatography; electromigration methods; selectand; selector; reciprocal principle; combinatorial libraries; cyclopeptides; single-walled nanotubes (SWCNTs); fullerenes; cyclodextrins; calixarenes; congeners; enantiomers; molecular recognition; chiral recognition; chiral stationary phases

1. Introduction Separation approaches in highly selective supramolecular chromatography [1,2], with their most refined expression of stereochemical resolution [3], rely on the fast and reversible non-covalent association equilibrium between functionally and/or structurally complementary partners. The designations selector and selectand were introduced by Mikeš into chromatography to avoid ambiguities that may arise from the use of the notations solvent-solute, ligand-substrate, or host-guest [4]. The new terms were coined in analogy to Ashby’s cybernetic operator-operand terminology [5]. This terminology extends to all supramolecular interactions such as antibody-antigen, receptor-substrate, metal ion-ligand etc. In chromatographic partitioning systems, selectors are employed to separate mixtures of selectands which are added to the mobile phase. The selector is either present as a stationary phase or as an additive to the liquid mobile phase. In order to find an optimized selector for a pair of selectands, the role of selector/selectand can be reversed by the reciprocal principle. One molecule of the selectand is now used as the stationary phase and an array of potential structural related selectors are now employed as analytes in the mobile phase. The best identified candidate is then used as the optimized lead stationary phase for the target selectands to be separated. Thus the reciprocal recognition principle applies to selective separation system when selectors and selectands change their role and belong to two or more forms of congeners, analogues, isomers etc. In the special case of chiral recognition, both enantiomers of selectand and selector must be available. This excludes proteins [6], carbohydrates [7] and antibodies [8], although they represent versatile chiral stationary phases (CSPs). Pertinent examples will be described in the following Sections.

Molecules 2016, 21, 1535; doi:10.3390/molecules21111535

www.mdpi.com/journal/molecules

Molecules2016, 2016,21, 21,1535 1535 Molecules

of35 35 22of

2. Origin of the Reciprocal Principle in Chromatography 2. Origin of the Reciprocal Principle in Chromatography Following the enantioseparation of [5]–[14]helicenes on the optically active charge-transfer agent Following the enantioseparation of [5]–[14]helicenes on the optically active charge-transfer agent R-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionic acid (TAPA, Figure 1) or its 2-butyric acid R-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionic acid (TAPA, Figure 1) or its 2-butyric acid homologue (TABA) [9–11], Mikeš speculated that the function of the selector and the selectand could homologue (TABA) [9–11], Mikeš speculated that the function of the selector and the selectand could be reversed, i.e., optically active helicenes could be used as resolving agents for racemic compounds be reversed, i.e., optically active helicenes could be used as resolving agents for racemic compounds [4]. [4]. Thus, in a reciprocal fashion, if the selectands A1 and A2 are chromatographically separated on the Thus, in a reciprocal fashion, if the selectands A1 and A2 are chromatographically separated on the selector B1 (or B2), the selectands B1 and B2 are separated on the selector A1 (or A2) as well, whereby selector B1 (or B2), the selectands B1 and B2 are separated on the selector A1 (or A2) as well, whereby the respective pairs A1/A2 and B1/B2 refer to isomers (e.g., stereoisomers) or belong to members of the respective pairs A1/A2 and B1/B2 refer to isomers (e.g., stereoisomers) or belong to members homologous series of compounds (Figure 2). Indeed, it was subsequently demonstrated that P-(+)of homologous series of compounds (Figure 2). Indeed, it was subsequently demonstrated that hexahelicene-7,7′-dicarboxylic acid disodium salt could be used as the chiral stationary phase (CSP) P-(+)-hexahelicene-7,70 -dicarboxylic acid disodium salt could be used as the chiral stationary phase as π-donor selector for the enantioseparation of N-2,4-dinitrophenyl(DNP)-α-amino acid esters as (CSP) as π-donor selector for the enantioseparation of N-2,4-dinitrophenyl(DNP)-α-amino acid esters π-acceptor selectands by HPLC [12]. as π-acceptor selectands by HPLC [12].

Figure (R)-(−)-2-(2,4,5,7-tetranitro-9Figure 1. 1. Enantiomeric Enantiomeric separation separation of of racemic racemic [5]–[14]helicenes [5]–[14]helicenes on on (R)-( −)-2-(2,4,5,7-tetranitro-9fluorenylideneamino-oxy)propionic fluorenylideneamino-oxy)propionicacid acid(TAPA) (TAPA) linked linked to to silica silica gel gel [4]. [4]. (a) (a) odd odd numbers; numbers; (b) (b) even even numbers. numbers.Insertions: Insertions:[7]helicene [7]heliceneand andTAPA. TAPA.

Figure Figure2.2.The Thereciprocal reciprocalsupramolecular supramolecularrecognition recognitionprinciple. principle.IfIfthe theselector selectorBBseparates separatesthe the selectands selectands A11and andA2A , than the selector A is expected to separate the selectands B and B . A and Bi are A , than the selector A is expected to separate the selectands B 1 and B 2 . A i and i are homologous 2 1 2 i homologouscongeners compounds, congeners or isomers (including enantiomers). compounds, or isomers (including enantiomers).

The principle of reciprocity by inverting the role of selectand and selector had first been demonstrated via the differentiation of enantiomers by chiral solvating agents (CSA) in NMR

Molecules 2016, 21, 1535

3 of 35

Molecules 2016, 21, 1535

35 The principle of reciprocity by inverting the role of selectand and selector had first3 of been demonstrated via the differentiation of enantiomers by chiral solvating agents (CSA) in NMR spectroscopy[13]. [13].Pirkle Pirkleetetal. al.stated: stated:the theenantiomers enantiomersofofsecondary secondaryand andtertiary tertiaryalcohols alcoholsmay maybebecaused causedtoto spectroscopy havenonequivalent nonequivalentnmr nmrspectra spectrathrough throughthe theuse useofofappropriate appropriateoptically opticallyactive activeamines aminesasassolvents. solvents.The Theconverse converse have of this phenomenon also obtains; amine enantiomers may have nonequivalent nmr spectra in optically active of this phenomenon also obtains; amine enantiomers may have nonequivalent nmr spectra in optically active alcohols [14] and Pirkle and Hoover concluded that the roles of an enantiomeric solute and a CSA may alcohols [14] and Pirkle and Hoover concluded that the roles of an enantiomeric solute and a CSA may interchangeablefor foraagiven givenpair pairofofcompounds compounds[15]. [15].The Thereciprocal reciprocalprinciple principlewas waslater laterextended extendedby by bebeinterchangeable Pirkle et al. to enantioselective liquid chromatography [16,17]. The researchers pointed out that one Pirkle et al. to enantioselective liquid chromatography [16,17]. The researchers pointed out that one cantake takeone onechiral chiralstationary stationaryphase phase(CSP) (CSP)totodevelop developothers: others:ififaaCSP CSPderived derivedfrom from(+)-A (+)-Aretains retains(+)-B, (+)-B, can thenaaCSP CSPcomprised comprisedofofimmobilized immobilized(+)-B (+)-Bshould, should,using usingthe thesame sameinteractions, interactions,selectively selectivelyretain retain(+)-A (+)-A[18,19]. [18,19]. then Or put in other words by Pirkle and Pochapsky: This ‘bootstrapping’ method of designing reciprocal CSPs Or put in other words by Pirkle and Pochapsky: This ‘bootstrapping’ method of designing reciprocal CSPs fromcompounds compoundsthat thatthemselves themselvesresolve resolveon onexisting existingCSPs CSPsisisbased basedon onthe thepremise premisethat thatififtwo twomolecules moleculesshow show from mutual chiral recognition, then it does not matter which of the two is bound to a stationary support for that mutual chiral recognition, then it does not matter which of the two is bound to a stationary support for that recognition to occur. This is in practice not strictly correct, for the nature of attachment to the CSP often affect recognition to occur. This is in practice not strictly correct, for the nature of attachment to the CSP often chiralchiral recognition. Within limits, however, reciprocity is is a auseful [20]. The Thelatter latter affect recognition. Within limits, however, reciprocity usefulguide guidetoto CSP CSP design design [20]. important restrictions was repeatedly stressed by Pirkle et al.: The two situations are not ‘mirror images’ important restrictions was repeatedly stressed by Pirkle et al.: The two situations are not ‘mirror images’ andthe therelationship relationshipis isonly only approximate. The manner of immobilization, for example, has some bearing and approximate. The manner of immobilization, for example, has some bearing uponupon the the efficacy of the chiral recognition process [19]. Thus deviations from the reciprocal principle may arise efficacy of the chiral recognition process [19]. Thus deviations from the reciprocal principle may arise from from the environment the selector the stationary e.g., the result of immobilization and the environment of the of selector in thein stationary phase,phase, e.g., as theasresult of immobilization and the the presence of a spacer linking the selector to a chromatographic matrix presence of a spacer linking the selector to a chromatographic matrix [21]. [21]. A vivid example of the de novo design of a CSP for a particular targetchiral chiralselectand selectandby bythe the A vivid example of the de novo design of a CSP for a particular target reciprocal principle involved a series of CSPs developed for the separation of the enantiomers of the reciprocal principle involved a series of CSPs developed for the separation of the enantiomers of the non-steroidalanti-inflammatory anti-inflammatorydrug drug(NSAID) (NSAID)naproxen naproxen[22,23]. [22,23].By Bythe theso-called so-calledimmobilized immobilizedguest guest non-steroidal method, a single enantiomer of naproxen (Figure 3a) was immobilized to form a naproxen-derived method, a single enantiomer of naproxen (Figure 3a) was immobilized to form a naproxen-derived CSP CSP (Figure 3b) which was used then used to identify a potential candidate an array of test racemates (Figure 3b) which was then to identify a potential candidate fromfrom an array of test racemates as as a de novo enantioselective naproxen selector [23]. This reciprocal study also identified the structural a de novo enantioselective naproxen selector [23]. This reciprocal study also identified the structural requirementsfor forenantioselective enantioselectiverecognition recognitionfor forthe thetarget targetracemate racematenaproxen. naproxen.The Thetailor-made tailor-made requirements Whelk-O1 CSP CSP (Figure only showed the highest enantioseparation factor for naproxen (α = 2.25) Whelk-O1 (Figure3c) 3c)not not only showed the highest enantioseparation factor for naproxen but also enantioseparated structurally related NSAIDs as well as a host of different racemates by a (α = 2.25) but also enantioseparated structurally related NSAIDs as well as a host of different combination of face-to-edge, and face-to-face π-π interaction as well as hydrogen bonding [22,23]. racemates by a combination of face-to-edge, and face-to-face π-π interaction as well as hydrogen The Whelk-O1 a broad spectrum for enantioseparations of aromatic selectands. The bonding [22,23]. CSP The offered Whelk-O1 CSP offered a broad spectrum for enantioseparations of aromatic enantioseparation of bis-, trisand hexakis-adducts of C 60 has also been achieved by HPLC on the selectands. The enantioseparation of bis-, tris- and hexakis-adducts of C60 has also been achieved by Whelk-O1-CSP (Section 4). (Section 4). HPLC on the Whelk-O1-CSP

(a)

(b)

(c)

Figure3.3.(a) (a)Structure Structureof ofnaproxen; naproxen;(b) (b)AAsingle singleenantiomer enantiomerofofnaproxen naproxenattached attachedcovalently covalentlyororionically ionically Figure onsilica; silica;(c)(c)The Theoptimized optimized CSP Whelk-O1 (commercialized Regis Chemical Morton Grove, on CSP Whelk-O1 (commercialized by by Regis Chemical Co.,Co., Morton Grove, IL, IL, USA). Reprinted (adapted) with permission from [23]. USA). Reprinted (adapted) with permission from [23].

Some preliminary work on the use of the reciprocal principle has been reported for the design of Some preliminary work on the use of the reciprocal principle has been reported for the design pyrethroid specific CSPs [24]. In a preliminary attempt to design novel CSPs for the resolution of the of pyrethroid specific CSPs [24]. In a preliminary attempt to design novel CSPs for the resolution stereoisomers of the insecticide cypermethrin, the principle of reciprocal interaction has been considered of the stereoisomers of the insecticide cypermethrin, the principle of reciprocal interaction has been via the following rationale: ‘If a CSP based on a pure enantiomer (+)-A resolves a racemate (+/−)-B, then a considered via the following rationale: ‘If a CSP based on a pure enantiomer (+)-A resolves a racemate single pure enantiomer (+)- or (−)-B should resolve (+/−)-A. This is true if no major point of interaction is (+/−)-B, then a single pure enantiomer (+)- or (−)-B should resolve (+/−)-A. This is true if no major point changed when attaching B to silica’. It should be mentioned that by employing the CSP (−)-A instead of of interaction is changed when attaching B to silica’. It should be mentioned that by employing the CSP

Molecules 2016, 21, 1535

4 of 35

(−)-A instead of (+)-A, a reversal of the elution order of (−/+)-B would obtain. Likewise by employing the CSP (−)-B instead of (+)-B, again a reversal of the elution order of (−/+)-A would obtain. The biologically most active (1R cis S) stereoisomer of cypermethrin has been structurally modified by an allylic group, followed by free radical addition or hydrosilylation reaction on two different silanes and linking (i) a short-chain monofunctional unit and (ii) a long-chain trifunctional unit, respectively, to silica. As cypermethrin can be resolved on N-(3,5-dinitrobenzoyl)-phenylglycine CSP, the reciprocal interaction of the cypermethrin CSP with racemic N-3,5-dinitrobenzoyl-phenylglycinepropylamide was claimed [24]. The actual objective to use the cypermethrin-derived CSPs to identify and optimize the interaction with racemic test compounds from which a single enantiomer will then be bound to silica to afford a cypermethrin-specific CSP still awaits its realization [24]. 2.1. Capillary Electromigration Methods (CE) In enantioselective capillary electromigration methods, the non-racemic chiral selector is added to the background electrolyte as part of the mobile phase. The method of enantioseparation of charged selectands with neutral selectors is referred to capillary zone electrophoresis (CZE), whereas the method of enantioseparation of neutral selectands with charged selectors, entailed with self-mobility, is referred to electrokinetic chromatography (EKC) [25]. Chankvetadze argued that from a mechanistic point of view there is no principle difference whether the selectand or the selector is charged in electromigration methods [26]. The philosophy behind this concept has been stated as follows: if a neutral chiral selector can resolve the enantiomers of a charged chiral analyte, then the enantioseparation of the neutral chiral analyte should also be possible with the charged chiral selector [26]. Thus another type of the reciprocal principle applies to electromigration methods. Whereas the enantioseparation of neutral chiral selectands on charged chiral selectors were at first considered not possible in earlier studies of EKC, it was later successfully realized based on the reciprocal concept. Thus positively charged cyclodextrin derivatives [25,27] and negatively charged cyclodextrin derivatives [27–30] were used for the enantioseparation of a multitude of neutral chiral compounds [31,32]. Another reciprocal principle in the achiral electromigration domain has not yet been realized, but may be anticipated. Native α-, β-, γ-cyclodextrin [33] and, more recently, δ-cyclodextrin [34] have been used as an added mobile phase selector for molecular association with various positively or negatively charged selectands in CZE. By inverting their role, a charged selector may be able to separate various cyclodextrin congeners preceding δ-CD (α-, β-, γ-) and, more importantly, congeners following up δ-CD (>δ). In enantioselective capillary electrochromatography (enantio-CEC), a non-racemic chiral selector is used as CSP coated on the inner capillary wall. As the first examples, racemic 1-phenylethanol [35,36] and some non-steroidal anti-inflammatory drugs (NSAIDs, Figure 4) were enantioseparated on Chirasil-Dex (permethylated β-cyclodextrin chemically linked to polydimethylsiloxane and thermally immobilized on the inner wall of a 80 cm (effective) × 50 µm I.D. capillary column). In order to find the best cyclodextrin selector for a given racemic analyte by the reciprocal approach, the target molecule may now be used as CSP, whereas the modified cyclodextrin is added to the mobile phase. However, as the cyclodextrins are only available in one enantiomeric form (comprised of D -glucose building blocks) two columns coated with the L - and D -target molecule, respectively, are required. As shown in Figure 5 the highest enantioselectivity for phenylalanilol was observed with hydroxypropyl-β-cyclodextrin due to a large retention time difference on the two mirror image CSPs [37]. Although not related to the reciprocal principle, the following approach is of interest in its own right. A dual chiral recognition system can be obtained when one cyclodextrin selector is used as the CSP whereas another one is added to the mobile phase (combination of CEC and EKC). As the enantioselectivities of the bonded and free CDs were opposite, the incremental addition of the sulfonated β-cyclodextrin to a Chirasil-Dex-coated capillary led to an inversion of the elution order of racemic 1,1’-bi-2,2’-naphthol-hydrogenphosphate (20 mM borate/phosphate (pH 7) buffer, at 30 kV) [29].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

5 of 35 5 of 35 5 of 35

Figure 4. Enantioseparation of profens by electrochromatography on Chirasil-β-Dex coated and Figure of profens profens by electrochromatography on Chirasil-β-Dex Chirasil-β-Dexcoated coatedand and Figure 4.4. Enantioseparation Enantioseparation of by electrochromatography on ◦ immobilized (d f = 0.20 μm) on an 80 cm (effective) × 50 μm I.D. fused-silica capillary at 20 °C and 30 kV. immobilized (d = 0.20 µm) on an 80 cm (effective) × 50 µm I.D. fused-silica capillary at 20 C and immobilized (dff = 0.20 μm) on an 80 cm (effective) × 50 μm I.D. fused-silica capillary at 20 °C and 30 kV. Buffer: 20 mM Tris/HCl at pH 7. Reproduced from S. Mayer, Doctoral Thesis, University of Tübingen, 30 kV. Buffer: mM Tris/HCl pH 7. Reproduced from Doctoral S. Mayer,Thesis, Doctoral Thesis, of University Buffer: 20 mM20 Tris/HCl at pH 7. at Reproduced from S. Mayer, University Tübingen,of Tübingen, Germany, 1993. Tübingen, 1993. Tübingen,Tübingen, Germany, Germany, 1993.

Figure 5. Retention time differences of three modified β-cyclodextrins (β-CD as mono-2.6Figure 5. time time differences of three β-cyclodextrins (β-CD as mono-2.6Figure 5. Retention Retention differences of modified three modified β-cyclodextrins (β-CD as D- and L-phenylalaninol by the CEC-experiment (50 mM dimethylphenylcarbamate) on silica-bound D and L -phenylalaninol by the CEC-experiment (50 mM dimethylphenylcarbamate) on silica-bound mono-2.6-dimethylphenylcarbamate) on silica-bound and L-phenylalaninol by the CEC-experiment ammonium acetate, pH 6.5, 15 kV, column: 68 cm eff.D×-50 μm I.D.) [37]. ammonium acetate, pH 6.5, 15 kV, column: 68 cm eff. × 50 μm I.D.) [37]. (50 mM ammonium acetate, pH 6.5, 15 kV, column: 68 cm eff. × 50 µm I.D.) [37].

2.2. Combinatorial Library Approaches in Achiral Supramolecular Chromatographic Systems 2.2. Combinatorial Library Approaches in Achiral Supramolecular Chromatographic Systems 2.2. Combinatorial Library Approaches in Achiral Supramolecular Chromatographic Systems The reciprocal principle has been used for the rational design of an optimized selector for The reciprocal principle has been used for the rational design of an optimized selector for The reciprocal principle has been used for the rationalisdesign optimized selector for molecular molecular recognition. Another optimization principle based of onan the use of combinatorial selector molecular recognition. Another optimization principle is based on the use of combinatorial selector recognition. Another optimization principle is based on the use of combinatorial selector libraries libraries aimed at identifying the prime selector for a given enantioselective separation system. This libraries aimed at identifying the prime selector for a given enantioselective separation system. This approach can also be combined with the reciprocal principle [38,39]. It has generally been accepted approach can also be combined with the reciprocal principle [38,39]. It has generally been accepted

Molecules 2016, 21, 1535

6 of 35

aimed at identifying the prime selector for a given enantioselective separation system. This approach can also be combined with the reciprocal principle [38,39]. It has generally been accepted that the screening of libraries of selectors anchored to solid supports allows the identification of key structural elements responsible for host/guest interaction and molecular recognition through binding assays with soluble selectands. Combinatorial methods have been classified into two major categories. One is based on the synthesis of mixtures and another is based on the parallel synthesis of pure library components (also called the high-throughput approach). In the former approach, a mixture of library components is synthesized and screened simultaneously for the desired property. In the parallel library approach, each library component is synthesized and screened individually. The advantage of the parallel approach is that the library is generally better characterized, whereas the random library technique is advantageous in that a large number of components can be synthesized and screened more efficiently. A classification of the various combinatorial approaches for the preparation and use of CSPs for enantioseparations has been summarized [40]. Peptide libraries have been used to optimize selectors that bind specifically to a desired target protein in the area of protein purification. Solid phase ‘one-bead-one-peptide’ (OBOP) parallel combinatorial libraries have been successfully employed to search for affinity ligands for large proteins. Thus the immobilized linear hexapeptide Try-Asn-Phe-Glu-Val-Leu was identified as a potential host for the purification of S-protein by affinity chromatography [41]. Also an affinity resin containing the peptide ligand Phe-Leu-Leu-Val-Pro-Leu has been developed for the purification of fibrinogen. The selector was identified by screening the solid-phase combinatorial peptide library using an immunostaining technique [42]. In a reversed reciprocal fashion, affinity chromatography employing the immobilized S-protein was used for the screening of affinity peptide selectands from a soluble peptide library consisting of octamers with glycine (Gly) at both termini of each peptide, i.e., Gly-X-X-X-X-X-X-Gly. The six variable centre positions were constructed using random sequences of the six L-amino acids Tyr, Asn, Phe, Glu, Val and Leu. Peptides that were retained specifically on the immobilized S-protein column were eluted by 2% acetic acid. The peptides in the acid eluate were further separated using reversed-phase HPLC. Each separated peptide fraction was collected and the peptide sequences deconvoluted by mass spectrometry (MS/MS). The screenings of the peptide library resulted in twelve affinity peptides and eight out of them contained the essential sequence Asn-Phe-Glu-Val [43]. 2.3. Combinatorial Cyclopeptide Library Approaches in Chiral Capillary Electrophoresis Combinatorial cyclohexapeptide libraries were used as novel multicomponent chiral additives for enantioseparation in capillary electrophoresis (CE) by Jung and Schurig et al. in 1996 [44]. It was inferred that the time-consuming screening of a multitude of individual potential chiral selectors can be avoided by employing a whole selector library. The random library approach is then followed by deconvolution steps. The sub-libraries with reduced heterogeneity can then be employed to identify the most enantioselective selectors. In general, cyclopeptide libraries of high molecular diversity can be obtained by including structural variations. By using the nearly 200 commercially available amino acids of defined chirality, 2006 = 64 × 1012 individual cyclohexapeptides are in principle accessible. The number of cyclopeptides can further be increased by varying the ring size and by including non-peptidic components, derivatives with side chain and backbone modifications, and selected building blocks with inverted chirality [44]. Three libraries consisting of 183 = 5832 individual cyclohexapeptides were synthesized as a random mixture and were characterized by electrospray mass spectrometry [44]. Each library had the composition cyclo(O-O-X-X-X-O) and consisted of three defined positions (O) and three randomized positions (X) which represented a mixture of 18 of the common L-α-amino acids except cysteine and tryptophan. To aid ring closure, one position (O) consisted of unnatural D-alanine (ala). When added to the mobile phase in CE, the libraries cyclo(Asp-Phe-X-X-X-ala), cyclo(Arg-Lys-X-X-X-ala) and cyclo(Arg-Met-X-X-X-ala), respectively, enantioseparated Tröger’s base and N-2,4-dinitrophenyl (DNP)- or Fmoc-D,L-glutamic acid (α = 1.04–1.13) (Figure 6) [44].

Molecules 2016, 21, 1535

7 of 35

Molecules 2016, 21, 1535

7 of 35

Figure 6. Enantioseparation of racemic Tröger’s base and DNP-D,L-glutamic acid with combinatorial

Figure 6. Enantioseparation of racemic Tröger’s base and DNP-D,L-glutamic acid with combinatorial cyclohexapeptides by CE. Conditions: 10 mmol cyclopeptide library cyclo(OOXXXO) in phosphate buffer cyclohexapeptides by CE. Conditions: 10 mmol×cyclopeptide phosphate pH 7.4 (20 mmol). Capillary: 50 cm (effective) 50 μm I.D., 20 library kV (left, cyclo(OOXXXO) middle) and 10 kV in (right), buffer pH 7.4 (20 Capillary: 50 (middle cm (effective) 50 µm I.D., 20 kV (left, middle) and 10 kV detection at mmol). 260 nm (left) and 340 nm and right)×[44]. (right), detection at 260 nm (left) and 340 nm (middle and right) [44]. The result raised the question whether all 183 individual components (which are present in the library at high dilution and probably in a non-equimolar ratio) possess the same or, more likely, different Theenantioselectivities result raised the question whether all winners’. 183 individual components (which aremembers present in the with individual ‘losers and It may even be conceived that some library at high dilution andorder probably in a non-equimolar the same or, more likely, exert an opposite elution of the selectand thus reducing theratio) overallpossess enantioselectivity. Cooperative between the individual selectors were believed be absent [44]. To identify theconceived prime selector differenteffects enantioselectivities with individual ‘losers andtowinners’. It may even be that some in the library, Chiari et al. used an interesting deconvolution strategy [45]. members exert an opposite elution order of the selectand thus reducing the overall enantioselectivity. approach was performed in two steps: (i) the synthesis of sublibraries with a progressively CooperativeThis effects between the individual selectors were believed to be absent [44]. To identify the increased number of defined positions and (ii) the evaluation of their ability to act as chiral selectors prime selector in the library, Chiari et al. used an interesting deconvolution strategy [45]. in CE for a set of DNP-D,L-amino acids. First of all, Chiari et al. deconvoluted the library cyclo(ArgThis approach wasby performed in Dtwo thelibrary synthesis of sublibraries with a progressively Lys-X-X-X-β-Ala) substituting -Alasteps: (ala) of(i)the cyclo(Arg-Lys-X-X-X-ala) of Jung and increased number of by defined positions(Table and1,(ii) the evaluation of their ability to actS.2.2 as chiral Schurig et al. [44] achiral β-alanine S.1.1) [45]. Then a number the sublibraries S.2.1, and S.2.3 were prepared, whereby proline was deliberately selected for one of the positions X as selectors in CE for a set of DNP-D,L-amino acids. First of all, Chiari et al. deconvoluted theit library facilitates hexapeptide cyclization. However, only proline in position (Table 1, S.2.2) exerted aof Jung cyclo(Arg-Lys-X-X-X-β-Ala) by substituting D -Ala (ala) of the library four cyclo(Arg-Lys-X-X-X-ala) high enantioselectivity of the sublibrary.

and Schurig et al. [44] by achiral β-alanine (Table 1, S.1.1) [45]. Then a number the sublibraries S.2.1, S.2.2 and S.2.3 were prepared, whereby proline was deliberately selected for one of the positions X as Table 1. List of sublibraries of cyclohexapeptides used in the deconvolution process. X stands for one it facilitates of hexapeptide cyclization. However, proline in position fourwith (Table 1, S.2.2) exerted the 18 of the common L-α-amino acids exceptonly cysteine and tryptophan. Reprinted permission from [45]. Copyright (1998) American Chemical Society. a high enantioselectivity of the sublibrary. Numbering

Cyclohexapeptide Library

S.2.3:

cyclo(Arg-Lys-X-X-Pro-β-Ala) cyclo(Arg-Lys-Val-Pro-X-β-Ala) Cyclohexapeptide Library cyclo(Arg-Lys-Met-Pro-X-β-Ala) cyclo(Arg-Lys-X-X-X-β-Ala) cyclo(Arg-Lys-Ile-Pro-X-β-Ala) cyclo(Arg-Lys-Leu-Pro-X-β-Ala) cyclo(Arg-Lys-Pro-X-X-β-Ala) cyclo(Arg-Lys-Tyr-Pro-X-β-Ala) cyclo(Arg-Lys-X-Pro-X-β-Ala) cyclo(Arg-Lys-Phe-Pro-X-β-Ala) cyclo(Arg-Lys-X-X-Pro-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Val-β-Ala) cyclo(Arg-Lys-Val-Pro-X-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Met-β-Ala) cyclo(Arg-Lys-Met-Pro-X-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Ile-β-Ala) cyclo(Arg-Lys-Ile-Pro-X-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Leu-β-Ala) cyclo(Arg-Lys-Leu-Pro-X-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclo(Arg-Lys-Tyr-Pro-X-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Phe-β-Ala)

Table 1. List of sublibraries of cyclohexapeptides used in the deconvolution process. X stands for one S.1.1: cyclo(Arg-Lys-X-X-X-β-Ala) of the 18 of the common L-α-amino acids except cysteine and tryptophan. Reprinted with permission S.2.1: cyclo(Arg-Lys-Pro-X-X-β-Ala) from [45]. Copyright (1998) American Society. S.2.2: Chemical cyclo(Arg-Lys-X-Pro-X-β-Ala) S.3.1: Numbering S.3.2:

S.1.1:S.3.3: S.2.1:S.3.4: S.2.2:S.3.5: S.2.3:S.3.6: S.4.1:

S.3.1:S.4.2: S.3.2:S.4.3: S.3.3:S.4.4: S.3.4:S.4.5: S.3.5:S.4.6: S.3.6: S.4.1: S.4.2: S.4.3: S.4.4: S.4.5: S.4.6:

cyclo(Arg-Lys-Phe-Pro-X-β-Ala)

cyclo(Arg-Lys-Tyr-Pro-Val-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Met-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Ile-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Leu-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Phe-β-Ala)

Molecules 2016, 21, 1535

8 of 35

The optimized library cyclo(Arg-Lys-X-Pro-X-β-Ala) (S.2.2) was then semipreparatively resolved Molecules 2016, 21, 1535 8 of 35 into three fractions by reversed-phase HPLC. The second fraction exerted the highest enantioselectivity Molecules 2016, 21, 1535 8 of 35 and amino acid revealed that this fraction (S.2.2) contained the hydrophobic amino Theanalysis optimized library cyclo(Arg-Lys-X-Pro-X-β-Ala) was then semipreparatively resolved intoacids Val, The optimized librarygave cyclo(Arg-Lys-X-Pro-X-β-Ala) (S.2.2) was semipreparatively resolved into three fractions reversed-phase HPLC. The synthesis second fraction highest enantioselectivity and (Table 1). Met, Ile, Leu and Phe.by This rise to the of exerted thethen sixthe sublibraries S.3.1–S.3.6. three fractions by reversed-phase HPLC. The second fraction exerted thefor highest enantioselectivity and amino acid analysis revealed that this fraction contained hydrophobic amino acids deconvolution, Val, Met, Ile, The sublibrary cyclo(Arg-Lys-Tyr-Pro-X-β-Ala) (S.3.5) wasthe selected the further since amino acid analysis revealed fraction contained the hydrophobic amino acids Val,sublibrary Met, Ile, Leu and Phe. This gave rise tothat thethis synthesis of the six sublibraries S.3.1–S.3.6. (Table 1). The Tyr contains an aromatic system useful for detection purposes. Now the six sublibraries S.4.1–S.4.6 Leu and Phe. This gave rise to the(S.3.5) synthesis the sixfor sublibraries (Table 1). The cyclo(Arg-Lys-Tyr-Pro-X-β-Ala) was of selected the furtherS.3.1–S.3.6. deconvolution, since Tyrsublibrary contains an were synthesized (Table 1). In a similar fashion position 5 was optimized by a preceding HPLC cyclo(Arg-Lys-Tyr-Pro-X-β-Ala) (S.3.5) was selected for the deconvolution, sincewere Tyr contains an aromatic system useful for detection purposes. Now the further six sublibraries S.4.1–S.4.6 synthesized aromatic system useful for detection purposes. Now the six sublibraries S.4.1–S.4.6 were synthesized (Table 1). In a similar fashion position 5 was optimized by a preceding HPLC separation into three separation into three fractions and again identifying hydrophobic amino acids as being essential (Table 1). In a similar fashion position 5 was optimized by aas preceding HPLC separation into three and hence fractions and again hydrophobic amino acids being essential for enantioselectivity. for enantioselectivity. Phe identifying and Tyr in position 5 provided the highest enantioselectivity fractions identifying hydrophobic amino enantioselectivity acids as being essential for enantioselectivity. Phe andand Tyragain in position 5 provided the highest and hence the final defined the final defined cyclohexapeptide selector (Figure 7)for has been selected for the enantioseparation of Phe and Tyr in position provided the highest enantioselectivity and henceof the final defined cyclohexapeptide selector5(Figure 7) has been selected the enantioseparation DNPD,L-glutamic DNP-D,L-glutamic acid (α = 8) 1.24) (Figure 8) [45]. cyclohexapeptide selector (Figure 7) has been selected for the enantioseparation of DNP-D,L-glutamic acid (α = 1.24) (Figure [45]. acid (α = 1.24) (Figure 8) [45].

Figure

cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) Figure 7. The optimized cyclohexapeptide library component, deconvoluted by Chiari et al. [45]. 7.Figure The optimized cyclohexapeptide component, deconvoluted byetChiari 7. The optimized cyclohexapeptide library library component, deconvoluted by Chiari al. [45]. et

al. [45].

Figure 8. Enantioseparation of D,L-DNP-amino acids in a running electrolyte containing cyclo(ArgLys-Tyr-Pro-Tyr-β-Ala) (10 of mM 20 mM sodium buffer, pH 7.0) containing as chiral mobile phase 8. Enantioseparation -DNP-amino acidsphosphate in a running electrolyte cyclo(ArgFigure 8.Figure Enantioseparation ofD,Lin D , L -DNP-amino acids in a running electrolyte containing additive (CMPA) in CE. cmsodium (effective) × 50 μm I.D.; pH −20 7.0) kV,as 15chiral °C. Reprinted with Lys-Tyr-Pro-Tyr-β-Ala) (10 Column: mM in 2057mM phosphate buffer, mobile phase cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) (10 mM in 20 mMChemical sodium phosphate buffer, pH 7.0) as chiral mobile permission from [45]. Copyright American Society. additive (CMPA) in CE. Column:(1998) 57 cm (effective) × 50 μm I.D.; −20 kV, 15 °C. Reprinted with phase additive (CMPA) in CE. Column: cm (effective) 50 µm I.D.; −20 kV, 15 ◦ C. Reprinted with permission from [45]. Copyright (1998)57 American Chemical× Society. The overall deconvolution process required the synthesis and the evaluation of 15 sublibraries permission from [45]. Copyright (1998) American Chemical Society.

The of overall process by required the synthesis the evaluation of 15 [46]. sublibraries instead the 54deconvolution syntheses considered a classical procedureand of serial deconvolution instead the 54 syntheses considered by a classical and procedure of serial deconvolution Inof a follow up study, several cyclohexapeptide cylcoheptapeptide analogues of[46]. the identified The overall deconvolution process required synthesis and the evaluation of[47]. 15 sublibraries In amembers follow upwere study, several and cylcoheptapeptide analogues of the identified library studied ascyclohexapeptide chiral selectors the in countercurrent capillary electrophoresis At were studied countercurrent capillary At the operating pH of 7.0, all of as thechiral selectors bear ainpositive charge,of while the electrophoresis analyte bears a [47]. negative instead oflibrary the 54members syntheses considered by selectors a classical procedure serial deconvolution [46]. the operating pH of 7.0, all of the selectors a positive charge, while the ether) analyte bears a negative charge. A poly(acrylamide-co-allyl amide ofbear D-gluconic was coated onto theidentified In a follow up study, several cyclohexapeptide andacid-co-allylglycidyl cylcoheptapeptide analogues of the charge. A poly(acrylamide-co-allyl amide of D-gluconic acid-co-allylglycidyl ether) was coated onto the

library members were studied as chiral selectors in countercurrent capillary electrophoresis [47]. At the operating pH of 7.0, all of the selectors bear a positive charge, while the analyte bears a negative charge. A poly(acrylamide-co-allyl amide of D-gluconic acid-co-allylglycidyl ether) was coated onto the capillary to suppress the EOF to ensure that the analyte and selector would migrate in opposite directions by counter-current CE. The following changes were made in the deconvoluted cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala). (i) Tyr in position five was

Molecules 2016, 21, 1535

9 of 35

substituted by Tyr(3-NO2 ), Phe, Phe(4-NO2 ) and Trp; (ii) Tyr in position three was substituted by positively charged Lys, negatively charged Glu and neutral Leu; (iii) Ala was inserted into cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) in the subsequent positions 1–6 to yield conformationally flexible cycloheptapeptides with positional variety. All positively charged cyclohexapeptides are effective Molecules 2016, 21, 1535 9 of 35 chiral selectors for the enantioseparation of DNP-α-amino acids with resolutions factors Rs = 1–5. The charge in position 3 did anythat influence onand enantioselectivity of DNP-AAs. However, all of capillary to suppress the not EOFhave to ensure the analyte selector would migrate in opposite directions by counter-current CE. to Theresolve following changes were made the finding deconvoluted cyclohexapeptide the cycloheptapeptides failed DNP-amino acids [47].inThis suggested that the size and (i) Tyrimportant in position five substituted by Tyr(3-NO 2), Phe, Phe(4-NO2) rigidity cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala). of the cyclopeptide system was forwas ensuring chiral discrimination. In subsequent and Trp; (ii) Tyr in position three was substituted by positively charged Lys, negatively charged Glu NMR studies it was recognized that the presence of aromatic moieties carrying electron-withdrawing and neutral Leu; (iii) Ala was inserted into cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) in the subsequent substituents, i.e., nitro groups, were essential for enantiorecognition via π-π stacking interactions and positions 1–6 to yield conformationally flexible cycloheptapeptides with positional variety. All amide/aromatic n-π interactions [47,48]. positively charged cyclohexapeptides are effective chiral selectors for the enantioseparation of DNP-αBased onacids the optimized cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclohexapeptide selector deconvoluted amino with resolutions factors Rs = 1–5. The charge in position 3 did not have any influence on enantioselectivity of DNP-AAs. However,etallal. of the cycloheptapeptides failed to resolve DNP-amino by Chiari et al. [45], Jung and Schurig also screened related cyclopeptides with small acids [47]. This finding suggested that the size and was rigidity of the cyclopeptide system wasand important structural variations [49]. Thus aromatic tyrosine replaced by phenylalanine tryptophane for ensuring chiral discrimination. In subsequent NMR studies it was recognized that the presence of in position 3 and 5. In both cases the enantioselectivity was reduced for several DNP-amino aromatic moieties carrying electron-withdrawing substituents, i.e., nitro groups, were essential for acids (Table 2, Figure 9)viaand choice of the aromatic amino acid had a pronounced effect on enantiorecognition π-π the stacking interactions and amide/aromatic n-π interactions [47,48]. enantioselectivity. When the five-membered ring of proline was opened, the deconvoluted rigid structure of Based on the optimized cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclohexapeptide selector by Chiari et al. [45], Jung and Schurig al. also screenedby related small structural the cyclohexapeptide was released andetaccompanied ringcyclopeptides extension. with Thus when proline was variations [49]. Thus aromatic tyrosine was replaced by phenylalanine and tryptophane in position 3 at all substituted by 6-aminohexanoic acid in cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala), no enantioseparation and 5. In both cases the enantioselectivity was reduced for several DNP-amino acids (Table 2, Figure 9) was observed for various DNP-amino acids [49]. Enantioselectivity also broke down when the amino and the choice of the aromatic amino acid had a pronounced effect on enantioselectivity. When the acid sequence in cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) was changed to cyclo(Arg-Pro-Lys-Tyr-Tyr-β-Ala) five-membered ring of proline was opened, the rigid structure of the cyclohexapeptide was released and cyclo(Arg-Pro-Tyr-Lys-Tyr-β-Ala). Thuswhen the sequence amino acid-proline-aromatic and accompanied by ring extension. Thus proline was‘aromatic substituted by 6-aminohexanoic acid in amino acid’ (Tyr-Pro-Tyr) was essential for chiral recognition. Finally the ring size of the cyclopeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala), no enantioseparation at all was observed for various DNP-amino acids [49]. Enantioselectivity broke down when the aminowas acid observed sequence inwith cyclo(Arg-Lys-Tyrwas probed (Table 3) [49]. Thealso highest enantioselectivity cyclohexapeptides. Pro-Tyr-β-Ala) was changed to cyclo(Arg-Pro-Lys-Tyr-Tyr-β-Ala) and cyclo(Arg-Pro-Tyr-Lys-Tyr-βIn agreement with results of De Lorenzi et al. [47], enantioselectivity was lost for conformationally Ala). Thus the sequence ‘aromatic amino acid-proline-aromatic amino acid’ (Tyr-Pro-Tyr) was essential flexible cycloheptapeptides. Surprisingly, this drop was even more pronounced with cyclopentapeptides for chiral recognition. Finally the ring size of the cyclopeptide was probed (Table 3) [49]. The highest (Table 3). When the cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) was opened up to the linear enantioselectivity was observed with cyclohexapeptides. In agreement with results of De Lorenzi peptideetArg-Lys-Tyr-Pro-Tyr-Lys by lost exchanging β-Ala for flexible Lys nocycloheptapeptides. enantioselectivity for DNP-amino al. [47], enantioselectivity was for conformationally Surprisingly, acids was [49].more Thus the cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) identified by thisobserved drop was even pronounced with cyclopentapeptides (Table 3). When the cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) opened upthe to the linear peptide Arg-Lys-Tyr-Pro-Tyr-Lys by not be Chiari et al. (Figure 7) [45] indeedwas represents most favourable selector which could exchanging β-Ala for Lys no enantioselectivity for DNP-amino acids was observed [49]. Thus the optimized further. cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) identified by Chiari et al. (Figure 7) [45] indeed represents the mostfactors favourable which could optimized further.by CE (conditions as in Table 2. Resolution Rs ofselector DNP-amino acids not on be cyclohexapeptides

Figure 6) [49].

Table 2. Resolution factors Rs of DNP-amino acids on cyclohexapeptides by CE (conditions as in Figure 6) [49].

Amino Acid/Cyclopeptide

Amino Acid/Cyclopeptide

cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) cyclo(Arg-Lys-Phe-Pro-Phe-β-Ala) cyclo(Arg-Lys-Phe-Pro-Phe-β-Ala) cyclo(Arg-Lys-Trp-Pro-Trp-β-Ala) cyclo(Arg-Lys-Trp-Pro-Trp-β-Ala)

DNP-Glu

DNP-Glu

4.97 4.97 2.33 2.33 1.69 1.69

DNP-Norleu

DNP-Norleu 3.80 3.80 1.63 1.63 1.31 1.31

DNP-Meth DNP-Meth 5.46 5.46 1.96 1.96 1.21 1.21

DNP-Ala

DNP-Ala 2.27 2.27 1.03 1.03 1.47 1.47

DNP-Norval

DNP-Threo

DNP-Norval DNP-Threo 3.64 2.10 3.64 2.10 1.66 0.59 1.66 0.59 1.93 0.90 1.93 0.90

Figure 9. Influence of change the change Tyrfor forPhe Phe in in the the cyclohexapeptide for enantioseparation of Figure 9. Influence of the of of Tyr cyclohexapeptide for enantioseparation of norvaline (conditions as in Figure 8) [49]. norvaline (conditions as in Figure 8) [49].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

10 of 35 10 of 35

Table Resolution factors RsDNP-amino of DNP-amino on cyclopeptides of ring varying ring by CE Table 3. Resolution factors Rs of acids acids on cyclopeptides of varying size by CEsize (conditions (conditions Figure 6) [49]. as in Figureas 6)in [49]. Amino Acid/Cyclopeptide DNP-Glu DNP-Norleu DNP-Meth Amino Acid/Cyclopeptide DNP-Glu DNP-Norleu DNP-Meth cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) 4.97 3.80 5.46 cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) 3.80 1.41 5.461.53 cyclo(Lys-Lys-Tyr-Tyr-Tyr-Tyr-Lys) 4.97 0.36 cyclo(Lys-Lys-Tyr-Tyr-Tyr-Tyr-Lys) 0.36 1.41 0.62 1.53 0 cyclo(Lys-Tyr-Arg-Tyr-β-Ala) 0 cyclo(Lys-Tyr-Arg-Tyr-β-Ala) 0 0.62 0 cyclo(Arg-Lys-Tyr-Tyr-β-Ala) 0.23 0.73 cyclo(Arg-Lys-Tyr-Tyr-β-Ala) 0.23 0.73 0.430.43

DNP-Ala DNP-Ala 2.27 2.27 1.05 1.05 0.35 0.35 0.34 0.34

DNP-Norval DNP-Norval 3.64 3.64 1.46 1.46 0.37 0.37 0.55 0.55

DNP-Threo DNP-Threo 2.10 02.10 00 0 00

2.4. Reciprocal Principle in Combinatorial Approaches Utilizing Cyclopeptides—A Failure and a Success 2.4. Reciprocal Principle in Combinatorial Approaches Utilizing Cyclopeptides—A Failure and a Success Instead of a deconvolution strategy, a reciprocal LC approach has been considered by Jung, Schurig Instead of a deconvolution strategy, a reciprocal LC approach has been considered by Jung, et al. to identify an enantioselective hit component in the deconvoluted library (Figure 10) [49]. Schurig et al. to identify an enantioselective hit component in the deconvoluted library (Figure 10) [49].

Figure 10. 10. Illustration Illustration of of the the identification identification of of an an optimal optimal selectand selectand in in aa small small library library exhibiting exhibiting the the Figure highest enantioselectivity toward a chiral selector used either in the Sor R-configuration as revealed highest enantioselectivity toward a chiral selector used either in the S- or R-configuration as revealed by the the largest largest difference differenceof ofits itsretention retention(earmarked (earmarkedby bythe thepale paleblue bluecolour colouratatthe theright). right). by

Since the cyclopeptide library exists only in one enantiomeric form (obtained from L-amino acids) Since the cyclopeptide library exists only in one enantiomeric form (obtained from L-amino acids) two different columns containing the target racemate separately in the L- and D-form, respectively, are two different columns containing the target racemate separately in the L- and D-form, respectively, are required. The sub-library component present in the mobile phase which exhibits the largest difference required. The sub-library component present in the mobile phase which exhibits the largest difference in its retention on the two mirror-image selectors, should also show the highest enantioselectivity in in its retention on the two mirror-image selectors, should also show the highest enantioselectivity the reciprocal system when used as chiral selector for the racemic non-bonded candidate used as in the reciprocal system when used as chiral selector for the racemic non-bonded candidate used as selectand (Figure 10). selectand (Figure 10). In initial experiments L- or D-DNP-alanine silica (Figures 11 and 12) was filled into In initial experiments L- or D-DNP-alanine silica (Figures 11 and 12) was filled into 100 µm 100 μm I.D. capillaries. Cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) was selected as the analyte as it shows high I.D. capillaries. Cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) was selected as the analyte as it shows high enantioselectivity toward racemic DNP-alanine (Figure 8) [45]. Micro-HPLC measurements were enantioselectivity toward racemic DNP-alanine (Figure 8) [45]. Micro-HPLC measurements were performed in the reversed phase mode (H2O), in the organic polar mode (methanol/acetonitrile) and performed in the reversed phase mode (H2 O), in the organic polar mode (methanol/acetonitrile) and in the normal phase mode (toluene, n-hexane/2-propanol). However, no measurable retention time in the normal phase mode (toluene, n-hexane/2-propanol). However, no measurable retention time differences for the two mirror image stationary phases were observed. Also the coating of 50 μm I.D. differences for the two mirror image stationary phases were observed. Also the coating of 50 µm I.D. capillaries with L- or D-5-fluoro-2,4-DNP-alanine in the presence of aminopropyltrimethoxysilane in capillaries with L- or D-5-fluoro-2,4-DNP-alanine in the presence of aminopropyltrimethoxysilane in toluene and the investigation of cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) by pressure-supported open-tubular toluene and the investigation of cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala) by pressure-supported open-tubular capillary electrochromatography (OT-CEC) did not reveal differences in retention times as expected capillary electrochromatography (OT-CEC) did not reveal differences in retention times as expected from the reciprocal principle [49]. This unexpected result reinforces the limitation of the reciprocal from the reciprocal principle [49]. This unexpected result reinforces the limitation of the reciprocal principle as enantioselectivity may be reduced by non-specific interactions on silica. This might be principle as enantioselectivity may be reduced by non-specific interactions on silica. This might be circumvented by end-capping of silanol groups or elongation of the linker categories [38,39]. circumvented by end-capping of silanol groups or elongation of the linker categories [38,39].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

11 of 35 11 of 35 11 of 35

Figure 11. Linking of L- or D-DNP-alanine to silica (5 μm) via the amine route [49]. Figure 11.11.Linking to silica silica(5 (5μm) µm)via viathe theamine amineroute route [49]. Figure LinkingofofLL- -or orDD-DNP-alanine -DNP-alanine to [49].

Figure 12. Linking of L- or D-DNP-alanine to silica (5 μm) via the epoxide route [49]. Figure LinkingofofLL- -or orDD-DNP-alanine -DNP-alanine to [49]. Figure 12.12.Linking to silica silica(5 (5μm) µm)via viathe theepoxide epoxideroute route [49].

Since the cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala), used as chiral mobile phase selector Since the cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala), used as chiral mobile phase selector in CE, enantioseparates DNP-glutamic acid (Figure 8), the reciprocal principle was also tested by adding Since the cyclohexapeptide cyclo(Arg-Lys-Tyr-Pro-Tyr-β-Ala), used as chiral mobile phase selector in CE, enantioseparates DNP-glutamic acid (Figure 8), the reciprocal principle was also tested by adding L-DNP-Glu in a first run, and D-DNP-Glu in a second run, to the principle background electrolyte inby theadding CE in CE, enantioseparates DNP-glutamic acid (Figure 8), the reciprocal was also tested L-DNP-Glu in a first run, and D-DNP-Glu in a second run, to the background electrolyte in the CE experiment. The corresponding time in of a the added run, cyclohexapeptide was determined by indirect L -DNP-Glu in a first run, and Delution -DNP-Glu second to the background electrolyte in the experiment. The corresponding elution time of the added cyclohexapeptide was determined by indirectCE UV-detection. also no striking difference of cyclohexapeptide the elution time of the cyclohexapeptide was experiment. TheUnexpectedly corresponding elution time of the added determined by indirect UV-detection. Unexpectedly also no striking difference of the elution time of was the cyclohexapeptide was observed for the mirror-image situation of the two experiments between 15–20 °C and at voltages of UV-detection. also no striking of the elution of°C theand cyclohexapeptide observed for Unexpectedly the mirror-image situation of thedifference two experiments betweentime 15–20 at voltages of 10–25 kV, again reinforcing the limitation the concept [49]. was observed for the mirror-image situationof thereciprocal two experiments between 15–20 ◦ C and at voltages 10–25 kV, again reinforcing the limitation ofofthe reciprocal concept [49]. For screening a mixture of a combinatorial library for its best selector, a reciprocal approach of 10–25For kV,screening again reinforcing of the reciprocal [49]. also a mixture the of alimitation combinatorial library for itsconcept best selector, also a reciprocal approach was developed by Li et al. in line with Figure 10 [50]. The model study concerned the optimized Fordeveloped screeningby a mixture combinatorial itsmodel best selector, also a reciprocal approach was Li et al. of in aline with Figure library 10 [50].for The study concerned the optimized enantioseparation of (1-naphthyl)leucine ester with a small peptide library (4 × 4 = 16) containing Leu, was developed by Li al. in line with Figure 10 [50]. The model study(4concerned the optimized enantioseparation ofet (1-naphthyl)leucine ester with a small peptide library × 4 = 16) containing Leu, Ala, Gly and Pro and four different N-acyl components. In a reciprocal fashion, L-(1-naphthyl)leucine Ala, Gly and Proofand four different N-acyl components. Inpeptide a reciprocal fashion, enantioseparation (1-naphthyl)leucine ester with a small library (4 ×L4-(1-naphthyl)leucine = 16) containing Leu, ester was first immobilized and used as CSP for the enantiomeric peptide libraries obtained from Lester and used as CSP for the enantiomeric peptide libraries obtained from LAla, Glywas andfirst Proimmobilized and four different N-acyl components. In a reciprocal fashion, L -(1-naphthyl)leucine and D-amino acids, respectively. L- and D-sublibraries, exhibiting the largest retention differences on and D -amino acids, respectively. L and D -sublibraries, exhibiting the largest retention differences on ester was first immobilized and used as CSP for the enantiomeric peptide libraries obtained

Molecules 2016, 21, 1535

12 of 35

from L- and D-amino acids, respectively. L- and D-sublibraries, exhibiting the largest retention Molecules 2016, 21, 1535 12 of 35 differences on the L-configurated CSP, were deconvoluted and the best selector for the racemate was identified [50]. The optimized peptide Leu-Gly wasbest then attached to silica. It enantioseparated the L-configurated CSP, were deconvoluted and the selector for the racemate was identified [50].racemic The (1-naphthyl)leucine ester with an enantioseparation factor α = 3. Thus it was indeed demonstrated optimized peptide Leu-Gly was then attached to silica. It enantioseparated racemic (1-naphthyl)leucine that CSPs cananbeenantioseparation developed by a reciprocal approach using enantiomeric ester with factor α chromatographic = 3. Thus it was indeed demonstrated that CSPslibraries. can be developed by a reciprocal chromatographic approach using enantiomeric libraries. 2.5. On-Bead Library Combinatorial Approaches 2.5.For On-Bead Library Combinatorial Approaches the enantioselective screening of a target racemate, Weingarten et al. used a resin library containing 60 members of a multimodal receptor wereWeingarten bonded to polystyrene (100 µm) For the enantioselective screening of a targetwhich racemate, et al. used a beads resin library containing 60 members of a multimodal receptor which weresupported bonded to polystyrene (100 μm) (Figure 13) [51]. The chemically encoded, solid-phase library wasbeads generated by a (Figure 13) [51]. The chemically encoded, solid-phase supported library was (Figure generated by The a mix-andmix-and-split protocol employing three classes of chiral building blocks 13). rationale split protocol employing three classes of chiral building 13). IfThe rationale behind behind this non-chromatographic screening method was blocks stated (Figure as follows: a tightly binding andthis highly non-chromatographic screening method was stated as follows: If a tightly binding and highly to enantioselective (e.g., Rs > 10) resin or other solid material were readily available, then it would be possible enantioselective Rs > 10)stirring resin orthe other solid material wereareadily available, thenThe it would be possible to resolve compounds(e.g., by simply racemate with such resin and filtering. screening method resolve compounds by simply stirring the racemate with such a resin and filtering. The screening method employed differently coloured enantiomeric target molecules that were labelled with differently employed differently enantiomeric targetthe molecules that wereacid labelled with differently coloured dyes. Thus acoloured quasi-racemate included blue L-α-amino derivative and the red coloured dyes. Thus a quasi-racemate included the blue L-α-amino acid derivative and the red D-αD -α-amino acid derivative. Proline was selected as the test compound. The idea was to treat an amino acid derivative. Proline was selected as the test compound. The idea was to treat an equimolar equimolar mixture of these coloured probe molecules with a library of chiral selectors on synthetic mixture of these coloured probe molecules with a library of chiral selectors on synthetic beads in beads in which each bead carried a different chiral selector (one-bead-one-selector, OBOS). It was which each bead carried a different chiral selector (one-bead-one-selector, OBOS). It was reasoned reasoned that a highly enantioselective binding would yield beads that were either red or blue, whereas that a highly enantioselective binding would yield beads that were either red or blue, whereas nonnon-enantioselective binding would yield beads that were brown. Thus the best selector was directly enantioselective binding would yield beads that were brown. Thus the best selector was directly visualized through a screening.The Thereddest reddestand and bluest beads found visualized through atwo-colour two-colourdifferential differential binding binding screening. bluest beads found with each probe to determine determinethe thestructures structuresofof their associated chiral with each probewere werethen thenpicked pickedand and decoded decoded to their associated chiral selectors. Finally the leading resolving resins were individually resynthesized on a gram scale selectors. Finally the leading resolving resins were individually resynthesized on a gram scale andand again treated enantiomers.Subsequent Subsequenttreatment treatment with NaOMe again treatedwith withexcess excessred redand andblue blue proline proline enantiomers. with NaOMe andand HPLC quantification determinationofofthe theenantiomeric enantiomeric excess HPLC quantificationofofthe thereleased released dyes dyes provided provided aadetermination excess (ee)(ee) of of thethe proline The identified identifiedenantioselective enantioselective resolving resin was then prolinederivatives derivativesbound boundto to the the beads. beads. The resolving resin was then tested via abilitytotokinetically kineticallyresolve resolve racemic racemic proline tested via itsitsability proline ester esterin insolution solution[51]. [51].

Figure Structureofofthe themultimodal multimodal receptor ofof differently dye-tagged Figure 13.13.Structure receptor array arrayand andkinetic kineticresolution resolution differently dye-tagged solid-supported amino acids from a 60-member library. Module A contains 15 different and solid-supported amino acids from a 60-member library. Module A contains 15 different D- andD-L-amino L-amino acids, module B contains RR or SS diamine enantiomers, and the receptor module C contains acids, module B contains RR or SS diamine enantiomers, and the receptor module C contains RRRR or RRRR or SSSS enantiomers × 2different × 2 = 60library different library [51]. members [51]. 13 SSSS enantiomers resulting inresulting 15 × 2 ×in2 15 = 60 members Figure 13 Figure reproduced reproduced with permission from [52]. with permission from [52].

Molecules2016, 2016,21, 21,1535 1535 Molecules

1313ofof3535

2.6. On-Bead Combinatorial Approach of Individual Library Members 2.6. On-Bead Combinatorial Approach of Individual Library Members Fréchet et al. used an on-bead combinatorial approach to the rational design of optimized CSPs FréchetAt et first al. used an on-bead combinatorial approach to the rational design optimized CSPs for HPLC. a library of 3 × 12 = 36 L-α-amino acid anilides (consisting of 3ofα-amino acids and for HPLC. At first a library of 3 × 12 = 36 L -α-amino acid anilides (consisting of 3 α-amino acids 12 aromatic amines) was synthesized in solution, then attached to functionalized macroporous and 12 aromatic amines) was synthesizedbeads in solution, then attached to functionalized macroporous polymeric methacrylate/dimethacrylate which were packed into a ‘library column’ and tested polymeric methacrylate/dimethacrylate beads which were packed into a ‘library column’ and tested for for the enantioseparation of the target analyte. The lead selector of the library, i.e., L-proline-1the enantioseparation of the target analyte. The lead selector of the library, i.e., L -proline-1-indananilide, indananilide, employed for the HPLC enantioseparation of the target racemate N-(3,5-dinitrobenzoyl)employed for the HPLC of the target racemate leucine diallylamide, was enantioseparation identified by a deconvolution process usingN-(3,5-dinitrobenzoyl)-leucine 11 ‘sublibrary columns’ of lower diallylamide, was identified by a deconvolution process using 11 ‘sublibrary of [53,54]. lower diversity, each of which containing a CSP with a reduced number of library columns’ components diversity, each of which containing a CSP with a reduced number of library components [53,54]. Fréchet et al. also combined the library approach and the reciprocal principle for the development of Fréchet et al. alsoCSPs combined the library and the for the development enantioselective for HPLC [55,56]. approach Thus according to reciprocal Figure 14,principle a single enantiomer (+)-T of the oftarget enantioselective CSPs for HPLC [55,56]. Thus according to Figure 14, a single enantiomer racemate was at first immobilized onto a polymeric support and the resulting CSP was(+)-T used offor the racemate was first immobilized onto a polymeric andprepared the resulting CSP thetarget HPLC screening of at sublibraries of racemic compounds thatsupport have been by parallel was used for the HPLC screening of sublibraries of member racemic compounds have beenprepared preparedas combinatorial synthesis. The best enantioseparated of this librarythat (±)-S is then by parallel combinatorial synthesis. The best enantioseparated member of this library )-S is the single enantiomer (−)-S, coupled to a solid support and used in the second step for the(± required then prepared as the single enantiomer (−)-S,This coupled to amethodology solid support used in the second enantioseparation of the target racemate (±)-T. reciprocal hasand been exemplified for the step for the required enantioseparation of the target racemate ( ± )-T. This reciprocal methodology target racemate N-(3,5-dinitrobenzoyl)-leucine and the parallel library components of racemic 4-aryl-1,4has been exemplified for the target and the parallel library dihydropyrimidines obtained in aracemate one-pot N-(3,5-dinitrobenzoyl)-leucine three-component reaction (Biginelli dihydropyrimidine components of racemic 4-aryl-1,4-dihydropyrimidines obtained in a one-pot three-component reactiona three-component condensation reaction) [56]. Hydrogen-bonding and π-π interactions played (Biginelli dihydropyrimidine three-component condensation reaction) [56]. Hydrogen-bonding and dominant role in chiral recognition. π-π interactions played a dominant role in chiral recognition.

Figure Reciprocal combinatorial approach to the to optimization of a CSP. of Reproduced with permission Figure14.14. Reciprocal combinatorial approach the optimization a CSP. Reproduced with ofpermission the Royal Society of Chemistry from [55]. from [55]. of the Royal Society of Chemistry

Librariesofofpotential potentialchiral chiralselectors selectorshave havealso alsobeen beenprepared preparedby bythe theUgi Ugithree-component three-componentreaction reaction Libraries β-lactams and and screened screened for using thethe reciprocal approach involving a CSP ofofβ-lactams for their theirenantioselectivity enantioselectivitybyby using reciprocal approach involving a comprised of the immobilized model target compound N-(3,5-dinitrobenzoyl)L -leucine [57]. The CSP comprised of the immobilized model target compound N-(3,5-dinitrobenzoyl)-L-leucine [57]. lead candidate identified from of 2-oxo-azetidine 2-oxo-azetidine acetic aceticacid acid The lead candidate identified fromthe thelibrary libraryofofracemic racemic phenylamides phenylamides of derivativeswas wassubsequently subsequentlysynthesized synthesizedininbulk. bulk.The Theup-scaled up-scaledracemate racematewas wasthen thenresolved resolvedby bychiral chiral derivatives preparative HPLC and one enantiomer was immobilized on a solid support. The obtained CSP showed preparative HPLC and one enantiomer was immobilized on a solid support. The obtained CSP showed enantioselectivities of α = 3 for various amino acid derivatives. Interestingly, the enantioselectivities

Molecules 2016, 21, 1535

14 of 35

enantioselectivities of α = 3 for various amino acid derivatives. Interestingly, the enantioselectivities observed were even higher than those obtained during the reciprocal screening on the immobilized target compound N-(3,5-dinitrobenzoyl)-L-leucine [57]. 2.7. Batch-Screening of Peptide Libraries and the Reciprocal Principle Previous studies by Pirkle, Hyun et al. had shown that the enantiomers of racemic N-(3,5-dinitrobenzoyl)-peptide selectands are enantioseparated on a number of CSPs [58]. It was therefore reasoned that in a reciprocal fashion N-(3,5-dinitrobenzoyl)-peptide CSPs should be capable of enantioseparating various racemic compounds [59]. In an off-column batch-screening approach, Welch et al. probed the selector-selectand interaction directly, rather than resorting to a bonded analogue of the selectand used as CSP and packed into a column [60,61]. Thus the individual components of a parallel library of 50 silica-supported N-(3,5-dinitrobenzoyl)-dipeptide CSPs were prepared by solid phase synthesis on aminopropylsilica (APS) particles [59]. Whereas the ‘aa1’ position (Figure 15, top) consisted of the five D-α-amino acids Phg, Val, Pro, Gln and Phe, the ‘aa2’-position (Figure 15, top) was comprised of the same five α-amino acids in the L- and D-form, respectively, giving rise to 50 different dipeptides. The members of the library were subsequently screened for the enantioselective recognition of the test racemate N-(2-naphthyl)alanine diethylamide by an efficient batch-screening process. Thus individual vials containing approximately 50 mg of the candidate CSPs were placed into an autosampler and a dilute solution of the target racemate was added to each vial. The concentration (1 mL; 10−5 M in 2-propanol/n-hexane (20/80, v/v)) of the analyte solution was low enough to avoid saturation of the available interaction sites on the CSP. During equilibration using a rotary platform shaker, the individual enantiomers were partitioned between the solid and liquid phases of the vial. After equilibration, enantioselective HPLC of 50 µL injections of the supernatant solution in each vial afforded two peaks for the individual enantiomers. A 1:1 ratio revealed no enantioselectivity of the adsorbent, whereas ratios up to 1:2.72 indicated the presence of an enantioselective adsorbent. It was found that the homochiral combinations (DD) of the library components exhibited a higher enantioselectivity than the heterochiral combinations (LD) [60]. The presence of steric bulk in the ‘aa 2’ position and of hydrogen-bonding groups in the ‘aa1’ position proved to be important for enantiomeric discrimination of the test racemate. APS-bonded D-Gln-D-Val-DNB, D-Pro-D-Val-DNP and D-Gln-D-Phe-DNP were identified as the leading enantioselective selectors for racemic N-(2-naphthyl)alanine diethylamide. Based on these findings, an improved library of individual dipeptides was designed. It included amino acids with side chain functional groups such as Glu, His, and Arg, Whereas the ‘aa1’ position (Figure 15, top) consisted of the seven L-α-amino acids Gln, Asn, Ser, His, Asp, Arg and Glu, the ‘aa2’-position (Figure 15, top) was comprised of the L-α-amino acids Val, Leu, Ileu, Phe, Tyr, tert-Leu and Trp and the D -α-amino acids Val, Leu, Ileu, Phe, Tyr and Trp, respectively, giving rise to 91 different dipeptides. N-(2-naphthyl)alanine diethylamide was enantioseparated on APS-bonded L-Glu-L-Leu-DNP with the high enantioselectivity of α = 20.74 (Figure 15, bottom) [61]. Up to 100 mg of the compound could be enantioseparated using an analytical column (4.6 mm × 250 mm I.D.) in a single injection. The versatile approach of Welch et al. permitted the identification of the necessary structural requirements for chiral recognition by comparing the relative performance of various CSPs in the library. The enantioselective library approach, commercialized jointly by Regis Technology and MediChem Research [reported in: Chem. Eng. News 1999, 11, p. 101], may also prove useful for the discovery of improved selectors for other types of supramolecular chromatographic interactions.

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

15 of 35 15 of 35

Figure Top: DNB-dipeptide libraries usedasasCSP CSP for for the of N-(2-naphthyl)alanine Figure 15. Top:15.DNB-dipeptide libraries used theenantioseparation enantioseparation of N-(2-naphthyl)alanine diethylamide. Bottom: The optimized aminopropylsilica-bonded L-Glu-L-Leu-DNB selector enantioseparates diethylamide. Bottom: The optimized aminopropylsilica-bonded L-Glu-L-Leu-DNB selector the test racemate with high enantioselectivity. Column: 4.6 mm × 250 mm I.D., eluent: 2-propanol/nenantioseparates the test racemate with high enantioselectivity. Column: 4.6 mm × 250 mm I.D., hexane (20/80 v/v). Reprinted with permission from [61]. Copyright (1999) American Chemical Society. eluent: 2-propanol/n-hexane (20/80 v/v). Reprinted with permission from [61]. Copyright (1999) American AnChemical improvedSociety. strategy for the evaluation of common CSPs by the straightforward single-vial-

equilibration-approach of Welch et al. was based on rapid LC/MS screening of isotopically labelled pseudoenantiomers present in the supernatant instead of the HPLC analysis of the enantiomers [62]. An improved strategy for the evaluation of common CSPs by the straightforward single-vialConcurrently, another batch-wise screening method was described by Wang and Li [63]. The equilibration-approach of Welch et to al.that wasofbased rapid screening of incubation isotopically approach was almost identical Welchon et al. withLC/MS the exception that the waslabelled pseudoenantiomers present in the supernatant instead of the HPLC analysis of the enantiomers performed with the selector attached to a polymeric resin, instead of a chromatographic support. The [62]. reliability of the established stepwise solid-phase synthesis considered Concurrently, another batch-wise screeningMerrifield-type method waspeptide described bywas Wang and Li [63]. as an advantage. However, it wastoalso recognized would not The approach was almost identical that of Welchthat et the al. screening with theresult exception thatnecessarily the incubation correlate with the chromatographic experiment as no chromatographic support was used for the batch was performed with the selector attached to a polymeric resin, instead of a chromatographic support. screening step [38,39]. Crucial elements of the approach were the reciprocity of chromatographic The reliability of the established stepwise solid-phase Merrifield-type peptide synthesis was considered separation and the use of enantiomeric libraries. Thus, a potential chiral selector from the parallel library as an advantage. However, it wasresin. alsoThe recognized that the screening result would not necessarily was linked onto a solid-phase racemic analyte in the proper solvent was then allowed to correlateequilibrate with thewith chromatographic experiment as noThe chromatographic was used this potential selector on the resin. enantiomeric ratiosupport of the analyte in the for the supernatant was analysed after the equilibration period by circular dichroism. A selective adsorption batch screening step [38,39]. Crucial elements of the approach were the reciprocity of chromatographic of one of the the two to the resin was indicative an efficient chiralselector selector. from It wasthe thenparallel separation and useenantiomers of enantiomeric libraries. Thus, aofpotential chiral linked onto a chromatographic support, and its enantioselectivity toward the target analyte was library was linked onto a solid-phase resin. The racemic analyte in the proper solvent was then allowed evaluated. The feasibility of the parallel library screening procedure was compared with the model to equilibrate this potential selector onenantioseparation the resin. The enantiomeric ratio of ester the analyte system with studied earlier of the chiral HPLC of N-(1-naphthyl)leucine using a in the supernatant was analysed after the equilibration period by circular dichroism. A selective adsorption 16-member small library [50]. al. simplified the reciprocal using onlyof one (all-L)chiral which was equilibrated of one of theBluhm two et enantiomers to the resinsystem was by indicative anlibrary efficient selector. It was then on two columns containing the immobilized target racemate (1-naphthyl)leucine ester either in linked onto a different chromatographic support, and its enantioselectivity toward the target analyte was the L- or enantiomeric D-form, followed by incubation of each CSP with a mixture library [64]. The evaluated. The feasibility of the parallel library screening procedure was compared with the model adsorbed members of the library on the two columns were washed off and analysed by a reversed system studied earlier of the chiral HPLC enantioseparation of N-(1-naphthyl)leucine ester using a phase HPLC assay. Components with the same retention time but different intensities in both 16-member small library [50]. chromatograms represented an optimized selector. Its chemical structure was then determined by Bluhm et oral.LC-MS-MS. simplified thereciprocal reciprocal system by using candidate, only onei.e., library (all-L)N-3,5which was LC-MS In the fashion, the optimized immobilized dinitrobenzoylL-leucinecolumns ester, wascontaining used as CSP to immobilized enantioseparatetarget racemic N-(1-naphthyl)leucine equilibrated on two different the racemate (1-naphthyl)leucine ester in with theLhigh α =followed 12 [64]. In an study, 200-member parallel ester either the - or enantioseparation enantiomeric Dfactor -form, byoptimized incubation of aeach CSP with a mixture

library [64]. The adsorbed members of the library on the two columns were washed off and analysed by a reversed phase HPLC assay. Components with the same retention time but different intensities in both chromatograms represented an optimized selector. Its chemical structure was then determined by LC-MS or LC-MS-MS. In the reciprocal fashion, the optimized candidate, i.e., immobilized N-3,5-dinitrobenzoyl-L-leucine ester, was used as CSP to enantioseparate racemic N-(1-naphthyl)leucine ester with the high enantioseparation factor α = 12 [64]. In an optimized

Molecules 2016, 21, 1535

16 of 35

study, a 200-member parallel dipeptide library was created and screened with 10-undecenyl-N(1-naphthyl)leucinate [65]. The library was prepared using Fmoc-protected α-amino acids with amino-methylated polystyrene (AMPS) resin as solid support. The library contained three modules. Modules 2 and 3 consisted of ten different L-α-amino acids, whereas the N-terminating module 1 was an acetyl group or a 3,5-dinitrobenzoyl (DNB) group. Each single library member was incubated with a solution of racemic 10-undecenyl-N-(1-naphthyl)leucinate. After equilibration, the supernatant was collected and assessed for enantiomeric excess ee by circular dichroism (CD). In another study, an 81-member dipeptide library was evaluated for the target compound containing a stereogenic phosphorus atom, i.e., tert-butyl-1-(2-methylnaphthyl)phosphane oxide [66]. The member with the highest enantioselectivity was DNB-D-Asn-L-Thr-Abu-AMPS (Abu = 4-aminobutanoic acid). It was attached to aminopropylsilica and used as CSP for the target racemate in HPLC. However, the enantioselcetivities between the batch-experiment and the chromatographic experiment did not match due to the different solid supports. Therefore a longer linker had to be employed to immobilize the chiral selector onto silica to observe an enantioseparation factor of α = 3.2 for the racemic phosphine oxide. A 121-member peptide library was also tested for the enantioseparation of atropisomeric 1,10 -bi-2,20 -naphthol. A trityl-protected di-asparagine selector (Fmoc-Asn(Trt)-Asn(Trt)) yielded a separation factor of α = 7.2 [67]. By the reciprocal principle, it can be assumed that the target racemate 1,10 -bi-2,20 -naphthol, for which a highly enantioselective peptide has been identified, could also be used as a CSP for the enantioseparation of the racemic peptides. Another high throughput screening protocol was proposed for chiral selector discovery. It was modeled after the protocol for biological screening of candidate drugs from chemical libraries. The procedure was based on target distribution between an aqueous phase and an organic phase [68]. 3. Single-Walled Carbon Nanotubes, SWCNTs In single-walled carbon nanotubes (SWCNTs), a single layer graphene sheet is rolled-up into a tubular structure. The use of SWCNTs as selective stationary phases for the chromatographic separation of various classes of achiral compounds has been reviewed [1]. SWCNTs exist in three forms, i.e., two achiral structures (armchair and zig-zag) and a chiral structure. Chiral single-walled carbon nanotubes (n,m)-SWCNTs are characterized by n,m- indices, where n and m are coordination numbers of the carbon atoms in the hexagonal network. By convention, when n is greater than m, the species is right-handed or P (for positive), and when m is greater than n, the species is left-handed or M (for negative). (n,m)-SWCNTs represent interesting supramolecular entities for the reciprocal chromatographic selector/selectand relationship. The enantiomers of nine pre-separated single chirality (n,m)-SWCNTs of varying enantiomeric purity were separated on Sephacryl gel containing allyl-dextran as the chiral selector by gel permeation chromatography (Figure 16) [69]. The successful enantioseparation was proved by their opposite CD-spectra. The enantiomeric purities are not known and they varied for different (n,m)-single chirality SWCNTs with the second eluted enantiomer being more enriched [69]. The fractionation of (n,m)-SWCNTs by countercurrent chromatography also led to chiral recognition [70]. The method was based on the partition of sodium deoxycholate (SDC) dispersed (n,m)-SWCNTs in a polyethylene glycol/dextran aqueous two-phase system. The (n,m)-dependent bimodal elution peak width of the chiral nanotubes were interpreted as enantiomeric recognition in the presence of chiral SDC and polymeric dextran [70]. Dextran represents a complex branched glucan (a polysaccharide made of D-glucose building blocks). By the reciprocal approach, nonracemic (n,m)-SWCNTs may be considered in the future as novel CSPs for the enantiomeric separation of lower linear dextrins (‘acyclodextrins’, maltooligosaccharides) which are readily available in both enantiomeric forms obtained from D- and L-glucose, respectively. Incidentally, acetylated/silylated dextrins with a different degree of oligomerization and devoid of a molecular cavity have been used as CSPs for the enantioseparation of derivatized amino acids and various underivatized racemic compounds by GC, thus highlighting the role of the polar external surface of the selector for chiral recognition in contrast to the well-established inclusion mechanism exerted by cyclodextrins [71].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

17 of 35 17 of 35

Figure 16. Illustration of the enantioseparation of P- and M-SWCNTs by gel permeation liquid Figure 16. Illustration of the enantioseparation of P- and M-SWCNTs by gel permeation liquid chromatography on an allyl-dextran stationary phase. Reprinted with permission from [69]. Copyright chromatography on an allyl-dextran stationary phase. Reprinted with permission from [69]. Copyright (2014) American Chemical Society. (2014) American Chemical Society.

Using molecular dynamics simulations, Raffaini and Ganazzoli investigated in a theoretical Using simulations, Raffaini andformed Ganazzoli in aacids theoretical study themolecular adsorptiondynamics and denaturation of an oligopeptide by 16investigated chiral L-α-amino and study the aadsorption and denaturation an oligopeptide formed 16 outer chiralsurface L -α-amino acids and having helical structure in the nativeofstate on both the inner andbythe of the chiral having a helical structure in thewith native state on handedness, both the inner the armchair outer surface of the chiral (10,20)and (20,10)-SWCNTs an opposite andand of the (16,16)-SWCNT nanotube a similar diameter [72]. The molecular indicated that the (10,20)and with (20,10)-SWCNTs withfor ancomparison opposite handedness, and ofsimulations the armchair (16,16)-SWCNT enantioseparation of the chiral carbon nanotubes[72]. by chromatographic methods are feasible that usingthe nanotube with a similar diameter for comparison The molecular simulations indicated oligopeptides of aofsufficient length. It was also suggested that natural oligopeptides could not only enantioseparation the chiral carbon nanotubes by chromatographic methods are feasible using be used to separate enantiomeric nanotubes, but in the reciprocal fashion also chiral nanotubes of be oligopeptides of a sufficient length. It was also suggested that natural oligopeptides could not only single handedness could be considered as chiral selectors for racemic oligopeptides (Giuseppina used to separate enantiomeric nanotubes, but in the reciprocal fashion also chiral nanotubes of single Raffiani, personal communication, handedness could be considered as2016). chiral selectors for racemic oligopeptides (Giuseppina Raffiani, Unspecified chiral (n,m)-SWCNTs were tested as CSPs for the enantioseparation of twelve classes personal communication, 2016). of racemic pharmaceuticals by nano-LC [73,74]. The finding would require that the commercially Unspecified chiral (n,m)-SWCNTs were tested as CSPs for the enantioseparation of twelve available chiral (6,7)-SWCNT (carbon >90%, 0.7–0.9 nm diameter by fluorescence) obtained from Sigmaclasses of racemic pharmaceuticals by nano-LC [73,74]. The finding would require that the Aldrich was nonracemic or even enantiomerically pure. Neither the mode of the required enantiomeric commercially available chiral (6,7)-SWCNT (carbon >90%, 0.7–0.9 nm diameter by fluorescence) enrichment of the synthetic racemic (6,7)-SWCNT nor its chiroptical data was reported. Some racemic obtained from Sigma-Aldrich was nonracemic or even enantiomerically pure. Neither the mode compounds showed a deviation of the expected 1:1 peak ratio and no enantioseparation was noticed for of the the chiral required enantiomeric of theofsynthetic racemic (6,7)-SWCNT nor its chiroptical (7,8)-SWCNTs. Alsoenrichment in another report the continuous-flow enantioseparation of carvedilol data was reported. Some racemic compounds showed a deviation of the expected 1:1 peak ratio with fluorescent detection on chiral carbon nanotubes, chiroptic data and enantiomeric compositions and no enantioseparation was noticed for the chiral (7,8)-SWCNTs. Also in another report of the of the chiral selector were not mentioned [75]. As chiral SWCNTs contain equal amounts of left- and continuous-flow enantioseparation of carvedilol fluorescent detection chiral carbon nanotubes, right-handed helical structures, their use as CSPwith requires the separation of theseon non-superimposable mirror chiroptic data and enantiomeric compositions of the chiral selector were not mentioned [75]. As image forms into single enantiomers of defined enantiomeric purity as outlined by Peng et al. [76]. chiral SWCNTs equal amountswhether of left- and helical their use as CSP requires In contain order to investigate the right-handed use of SWCNTs canstructures, improve enantioseparations on anthe separation thesecomprised non-superimposable mirror image forms intoliquid single(R)-N,N,N-trimethyl-2-aminobutanolenantiomers of defined enantiomeric existingofCSP, of the chiral nonracemic ionic bis(trifluoromethane-sulfon)imidate, purity as outlined by Peng et al. [76]. two capillary columns, i.e., one containing the chiral ionic liquid alone and the containing the chiral liquid together with the SWCNTs were tested by on GCan In order toother investigate whether the ionic use of SWCNTs can improve enantioseparations [77]. ItCSP, was shown that the column containing SWCNTs improved the enantioselectivity of the existing comprised of capillary the chiral nonracemic ionic liquid (R)-N,N,N-trimethyl-2-aminobutanolchiral ionic liquid. The ‘nonchiral’ role of capillary the SWCNT was due to one an increase of the surface ofliquid the bis(trifluoromethane-sulfon)imidate, two columns, i.e., containing the chiralarea ionic inner capillary wall by formation of a layer with a skeletal network, leading the enhanced retention alone and the other containing the chiral ionic liquid together with the SWCNTs were tested by GC [77]. times, whichthat improved the resolution Rs for the enantiomers [77]. the enantioselectivity of the It was shown the capillary columnfactor containing SWCNTs improved

chiral ionic liquid. The ‘nonchiral’ role of the SWCNT was due to an increase of the surface area of the 4. Chromatography of Fullerenes and the Reciprocal Principle inner capillary wall by formation of a layer with a skeletal network, leading the enhanced retention The molecular recognition of fullerene various selectors times, which improved the resolution factorcongeners Rs for the on enantiomers [77]. by liquid chromatography has been reviewed [1,78,79], including enantioseparation of chiral native fullerenes or its derivatives 4. [1]. Chromatography of Fullerenes and the Reciprocal For more selective molecular recognition leading toPrinciple large separation factors between C60 and C70 (α > 2), a supramolecular approach was first employed by Hawkins et al. [80]. As a strong π-acceptor The molecular recognition of fullerene congeners on various selectors by liquid chromatography has been reviewed [1,78,79], including enantioseparation of chiral native fullerenes or its derivatives [1].

Molecules 2016, 21, 1535

18 of 35

Molecules2016, 2016,21, 21,1535 1535 Molecules

18of of35 35 18

For more selective molecular recognition leading to large separation factors between C60 and C70 (α > 2), for afor supramolecular approach was first aemployed by Hawkins et al. [80]. As a strong π-acceptor selector theπ-donor π-donorfullerene fullerene selectands, aCSP CSPcontaining containing N-(3,5-dinitrobenzoyl)-phenylglycine selector the selectands, N-(3,5-dinitrobenzoyl)-phenylglycine selector for the π-donor fullerene CSP containing N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG), originally developed byselectands, Pirkleet etal. al.afor for enantioseparations [23],was wasemployed employed(Figure (Figure17). 17). (DNBPG), originally developed by Pirkle enantioseparations [23], (DNBPG), originally developed by Pirkle et al. for enantioseparations [23], was employed (Figure 17).

Figure 17. HPLC separation of (12.2 min) and (23.5 min) on Pirkle-type ionically DNBPG Figure17. 17.HPLC HPLCseparation separationof ofCC6060 (12.2min) min)and andCC C707070(23.5 (23.5min) min)on onaaaPirkle-type Pirkle-typeionically ionicallyDNBPG DNBPG Figure 60(12.2 containing column (250 × 10 mm I.D.) eluted with n-hexane at 5.0 mL/min and detected at 280 nm, containingcolumn column(250 (250× × mm I.D.) eluted with n-hexane mL/min detected at 280 nm, containing 1010 mm I.D.) eluted with n-hexane at at 5.05.0 mL/min andand detected at 280 nm, αα 2.25 (r.t.). Reprinted with permission from [80]. Copyright (1990) American Chemical Society. = 2.25 (r.t.). Reprinted with permission from [80]. Copyright (1990) American Chemical Society. ==α2.25 (r.t.). Reprinted with permission from [80]. Copyright (1990) American Chemical Society.

An unprecedented increase in retention of of C C6060 and and CC7070with with increasing increasing column column temperature temperature was was An Anunprecedented unprecedentedincrease increasein in retention retention of C 60 and C70 with increasing column temperature was observed by Pirkle and Welch on the covalently bonded π-acidic N-(3,5-dinitrobenzoyl)-phenylglycine observed observedby byPirkle Pirkleand andWelch Welchon onthe thecovalently covalentlybonded bondedπ-acidic π-acidicN-(3,5-dinitrobenzoyl)-phenylglycine N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG) containing column (250 × 4.6 mm I.D.) (Figure 18) [81]. Van’t Hoff plots revealed therare rarecase case (DNBPG) containing column (250 × 4.6 mm I.D.) (Figure 18) [81]. Van’t plots revealed (DNBPG) containing column (250 × 4.6 mm I.D.) (Figure 18) [81].Hoff Van’t Hoff plots the revealed the of positive enthalpy and entropy values, i.e., the gain in entropy is accompanied by an endothermic ofrare positive enthalpy entropyand values, i.e., the gain i.e., in entropy accompanied by an endothermic case of positiveand enthalpy entropy values, the gainis in entropy is accompanied by an adsorption. adsorption. endothermic adsorption.

Figure 18.N-(3,5-dinitrobenzoyl)-phenylglycine N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG) bonded toReprinted silica. Reprinted Reprinted with Figure 18. (DNBPG) bonded to silica. with permission Figure 18. N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG) bonded to silica. with from [81]. Copyright (1990) American permission from[81]. [81].Copyright Copyright (1990)Chemical AmericanSociety. ChemicalSociety. Society. permission from (1990) American Chemical

Tripodal 3,5-dinitrobenzoate ester and 2,4-dinitrophenyl ether stationary phases containing Tripodal Tripodal 3,5-dinitrobenzoate 3,5-dinitrobenzoate ester ester and and 2,4-dinitrophenyl 2,4-dinitrophenyl ether ether stationary stationary phases phases containing containing three π-acidic functionalities for simultaneous multipoint supramolecular interaction with fullerenes three threeπ-acidic π-acidicfunctionalities functionalitiesfor forsimultaneous simultaneousmultipoint multipointsupramolecular supramolecularinteraction interactionwith withfullerenes fullerenes (‘Buckyclutcher’) were developed and they provided the highest retention and the greatest separation (‘Buckyclutcher’) were developed and they provided the highest retention and the greatest separation (‘Buckyclutcher’) were developed and they provided the highest retention and the greatest separation factor for the /C 70 mixture mixture [82]. [82]. A remarkable remarkable separation of 15 15 fullerene congeners in the the range of factor /C 70 A separation of congeners in range of factorfor forthe theCCC606060 /C [82]. A remarkable separation of fullerene 15 fullerene congeners in the range 70 mixture C 60 –C 96 , including enantiomerically enriched specimens, by HPLC on a syndiotactic poly(methyl Cof 60–C , including enantiomerically enriched specimens, by by HPLC onona asyndiotactic C6096–C enantiomerically enriched specimens, HPLC syndiotacticpoly(methyl poly(methyl 96 , including methacrylate) selector in 45 min with chlorobenzene as eluent has been described [83]. methacrylate) selector in 45 min with chlorobenzene as eluent has been described [83]. methacrylate) selector in 45 min with chlorobenzene as eluent has been described [83]. (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionic acid(TAPA) (TAPA) (Figure 1b, and (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionic acid (TAPA) (Figure (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionic acid (Figures 1b1b, andand 19), Figure 19), a strong π-electron acceptor due to the tetranitrofluorenylidene moiety, forms supramolecular Figure 19),π-electron a strong π-electron acceptor due to the tetranitrofluorenylidene forms supramolecular a strong acceptor due to the tetranitrofluorenylidene moiety,moiety, forms supramolecular charge charge transfer complexes withappropriate appropriate π-donor molecules. TAPA wasto used toseparate separate andwith C70 charge transfer complexes with π-donor molecules. was used to CC6060C and transfer complexes with appropriate π-donor molecules. TAPATAPA was used separate C60 and 70 C70 with a separation factor α up to 2.6 and the same peculiar increase of retention with increasing with a separation α 2.6 upand to 2.6 same increase peculiarofincrease retention withtemperature increasing a separation factorfactor α up to the and samethe peculiar retentionofwith increasing temperature previously observed with π-acidic N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG) was temperature previously observed with π-acidic N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG) previously observed with π-acidic N-(3,5-dinitrobenzoyl)-phenylglycine (DNBPG) was foundwas and found and interpreted as arising from solute solvation at low temperatures [84]. The same authors found and interpreted as arising solute at solvation at low temperatures [84]. authors The same authors interpreted as arising from solutefrom solvation low temperatures [84]. The same mentioned mentioned that 60and and can behave as both both π-electron π-electron donor and π-electron π-electron acceptor (theaffinity electron mentioned that CC60can CC7070 can as and acceptor (the electron that C60 and C70 behave asbehave both π-electron donor anddonor π-electron acceptor (the electron of affinity of C 60amounts amounts to 2.6 eV). affinity of C 60 to 2.6 eV). C60 amounts to 2.6 eV). The enantioseparation of the the tris-adduct tris-adductC 60[C(COOEt) [C(COOEt)22]]]33 and and hexakis-adduct hexakis-adduct CC6060[C(COOEt) [C(COOEt)22]]66 The tris-adduct CC60 60 Theenantioseparation enantioseparation of of the [C(COOEt) 2 3 and hexakis-adduct C60 [C(COOEt)2 ]6 entailed with an inherent chiral substitution pattern (Figure 20) has also been described on the CSP entailed entailedwith withan aninherent inherentchiral chiralsubstitution substitutionpattern pattern(Figure (Figure20) 20)has hasalso alsobeen beendescribed describedon onthe theCSP CSP (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneaminooxy)propionic acid (TAPA) bonded to silica gel (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneaminooxy)propionic acid (TAPA) (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneaminooxy)propionic acid (TAPA)bonded bondedto tosilica silicagel gel (Figure 19) by micro HPLC [85]. (Figure (Figure19) 19)by bymicro microHPLC HPLC[85]. [85].

Molecules 2016, 21, 1535 Molecules 2016, 2016, 21, 21, 1535 1535 Molecules

19 of 35 19 of of 35 35 19

Structure of of (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionic (R)-(−)-2-(2,4,5,7-tetranitro-9-fluorenylideneamino-oxy)propionicacid acid(TAPA) (TAPA) Figure 19. Structure bonded to silica gel [4,84]. bonded to silica gel [4,84]. 60 The ofof C60 20,20, top) hashas been achieved by The enantioseparation enantioseparationof ofbis-, bis-,tristris-and andhexakis-adducts hexakis-adducts C(Figure top) been achieved 60 (Figure HPLC on the Whelk-O1-CSP [86]. As the Whelk-O1-CSP (Figure 3c) was originally optimized for by HPLC on the Whelk-O1-CSP [86]. As the Whelk-O1-CSP (Figure 3c) was originally optimized 60 may racemic naproxen [23], [23], it may be speculated that some bis-, trisof C60 be for racemic naproxen it may be speculated that some bis-,and tris-hexakis-adducts and hexakis-adducts of Calso 60 may enantioseparated by the silica-bonded naproxen selector (Figure 3a) via a twofold reciprocal recognition also be enantioseparated by the silica-bonded naproxen selector (Figure 3a) via a twofold reciprocal scenario. recognition scenario.

Figure of the tris-adduct C60 60[C(COOEt)22]33. Bottom: Enantioseparation of a hexakisFigure 20. 20. Top: Top:Structure Structure of the tris-adduct C60 [C(COOEt)2 ]3 . Bottom: Enantioseparation of a 60 [C(COOEt) 2 ] 6 on the CSP TAPA bonded to aminopropyl silica gel silica by micro HPLC (α HPLC = 1.1). adduct C 60 2 6 hexakis-adduct C60 [C(COOEt) ] on the CSP TAPA bonded to aminopropyl gel by micro 2 6 Packed capillarycapillary (20 cm ×(20 0.25 I.D.), acetonitrile/water (75/25, (75/25, v/v), room (α = 1.1).fused-silica Packed fused-silica cmmm × 0.25 mmμL/min I.D.), µL/min acetonitrile/water v/v), temperature. Reprinted with permission from [85]. room temperature. Reprinted with permission from [85].

In order to optimize supramolecular recognition between fullerenes and polyaromatic In order to optimize supramolecular recognition between fullerenes and polyaromatic compounds compounds (PAHs), the reciprocal principle of changing the chromatographic roles of selector and (PAHs), the reciprocal principle of changing the chromatographic roles of selector and selectand selectand has been suggested by Saito et al. [87]. In liquid chromatography, fullerene selectors can be has been suggested by Saito et al. [87]. In liquid chromatography, fullerene selectors can be used used as supramolecular stationary phases either as native crushed specimen packed into capillaries or as supramolecular stationary phases either as native crushed specimen packed into capillaries 60 phase (Figure 21) covalently linked to a functionalized silica gel [78,88,89]. The chemically bonded C60 or covalently linked to a functionalized silica gel [78,88,89]. The chemically bonded C60 phase [87] showed preferential interaction with triphenylene and perylene which possess partial structures (Figure 21) [87] showed preferential interaction with triphenylene and perylene which possess partial similar to that of C60 60. Interestingly, larger retention factors are observed for nonplanar vs. planar PAH structures similar to that of C60 . Interestingly, larger retention factors are observed for nonplanar 60 molecules of comparable molecular size. Thus effective molecular recognition occurs between the C60 vs. planar PAH molecules of comparable molecular size. Thus effective molecular recognition selector and nonplanar selectands with a similar molecular curvature of the fullerene [87]. Bonded C60 60 occurs between the C60 selector and nonplanar selectands with a similar molecular curvature of 60 and C70 70 [87,88]. As phases were also used for self-recognition, i.e., for the congener separation of C60 the fullerene [87]. Bonded C60 phases were also used for self-recognition, i.e., for the congener 60 phase entailed compared to an octadecylsilica (ODS) reversed phase devoid of aromatic moieties, a C60 separation of C60 and C70 [87,88]. As compared to an octadecylsilica (ODS) reversed phase devoid high retention factors for C60 60 and C70 70 and an enhanced separation factor (α = 2.9) [87]. of aromatic moieties, a C60 phase entailed high retention factors for C60 and C70 and an enhanced separation factor (α = 2.9) [87].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

20 of 35 20 of 35

Figure 21. 21. Synthetic Synthetic scheme scheme for for aaCC6060 bonded bonded silica silica stationary stationary phase. phase. Reproduced Reproduced with with permission permission Figure from [87]. [87]. Copyright Copyright Wiley-VCH, Wiley-VCH, Weinheim, Weinheim,Germany. Germany. from

Fullerenes have have also also been been used used as as stationary stationary phases phases in in gas gas chromatography chromatography (GC). (GC). A A glass glass capillary capillary Fullerenes column (15 m × 0.29 mm I.D.) was coated with pure C60 (film thickness 0.1 μm) using a high-pressure column (15 m × 0.29 mm I.D.) was coated with pure C60 (film thickness 0.1 µm) using a high-pressure static method method[90]. [90].The Thethermostable thermostablesolid solid C60 phase behaved liquid phase squalane its static C60 phase behaved likelike thethe liquid phase squalane in itsinvan vanWaals der Waals interaction n-alkanes. der interaction with with n-alkanes. C 60 linked to poly(dimethylsiloxane) (Figure 22, top) was coated on a 10 m × 0.25 mm I.D. fused C60 linked to poly(dimethylsiloxane) (Figure 22, top) was coated on a 10 m × 0.25 mm I.D. silica and used thefor selective separation of isomeric hexachlorobiphenyls differing in orthofused column silica column andfor used the selective separation of isomeric hexachlorobiphenyls differing in chlorine substitution pattern. Thermodynamic parameters were obtained by the retention-increment ortho-chlorine substitution pattern. Thermodynamic parameters were obtained by the retention-increment method RR′0 [91,92]. [91,92]. It It separates separates supramolecular method supramolecular interactions interactions from from non-specific non-specific interactions interactions of of the the linker. linker. It was shown that the retention behaviour of PCBs depends strongly on the degree of ortho-chlorine It was shown that the retention behaviour of PCBs depends strongly on the degree of ortho-chlorine substitution which which in in turn turn is is linked linked with with non-planarity non-planarity (Figure (Figure 22, 22, bottom). bottom). The The deviation deviation from from substitution planarity of PCBs strongly decreases the supramolecular interaction with C 60 [93]. Since planar PCBs planarity of PCBs strongly decreases the supramolecular interaction with C60 [93]. Since planar PCBs are more more toxic toxic than than non-planar non-planar PCBs, PCBs, aa tentative tentative link link of ofretention retentionbehaviour behaviouron onCC60 and and toxicity toxicity has has are 60 been advanced [93]. been advanced [93]. In Table Table44 the the selective selective interaction interaction of of aromatic aromatic compounds compoundswith withC C60 linked to poly(dimethylsiloxane) In 60 linked to poly(dimethylsiloxane) (Figure 22, byby thethe retention-increment R’, i.e., value two indicates that the retention (Figure 22,top) top)isisexpressed expressed retention-increment R0 ,a i.e., a of value of two indicates that the of the PCB congener on C 60 is twice as compared to a reference column devoid of the fullerene retention of the PCB congener on C60 is twice as compared to a reference column devoid of the selector [94]. fullerene selector [94]. Table 4.4. Retention-increments Retention-increments R’ for the the selective selective interaction interaction of of aromatic aromatic compounds compounds with with C C60.. Table R0 for 60 60 phase (0.25 μm). Complexation column (k): 10 m x 0.25 mm I.D. fused silica column coated with the C Complexation column (k): 10 m x 0.25 mm I.D. fused silica column coated with the C60 phase without (k C060, ):without 10 m × C 0.25 mm I.D. fused silica column coated with dimethylReference (k0, column (0.25 µm). column Reference 60 ): 10 m × 0.25 mm I.D. fused silica column coated acetylaminopropyl(methyl)polysiloxane (0.25 μm) [94].(0.25 µm) [94]. with dimethyl-acetylaminopropyl(methyl)polysiloxane

Analyte Analyte 1.2-Dichlorobenzene 1.2-Dichlorobenzene 2.6-Dichlorotoluene 2.6-Dichlorotoluene Nitrobenzene Nitrobenzene 2-Nitrotoluene 2-Nitrotoluene Aniline Aniline Phenol Phenol Naphthalene Naphthalene 1-Methylnaphthalene 2-Methylnaphthalene 1-Methylnaphthalene Indole 2-Methylnaphthalene Quinoline Indole Isoquinoline Quinoline 4-Chlorophenol Isoquinoline 2.4-Dichlorophenol 2.4.6-Trichlorophenol 4-Chlorophenol 2.4-Dichlorophenol 2.4.6-Trichlorophenol

T (°C) R’ = 0(k − k0)/k0 T (◦ C) R = (k − k0 )/k0 80 0.61 80 80 0.690.61 80 0.69 80 0.950.95 80 80 0.910.91 80 80 80 1.581.58 90 90 0.500.50 90 0.73 90 0.730.83 120 120 120 0.830.70 120 120 0.701.28 120 120 1.281.87 120 2.94 120 1.870.34 120 120 2.940.77 120 120 120 0.341.29 120 0.77 120 1.29

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

21 of 35 21 of 35

Figure C60 60 linked linked to to aminopropyl aminopropyl poly(dimethylsiloxane). poly(dimethylsiloxane). Bottom: Figure 22. 22. Top: Top: Synthesis Synthesis of of C Bottom: Gas Gas chromatographic elution order of hexachlorobiphenyl congeners with different degree of ortho-chloro chromatographic elution order of hexachlorobiphenyl congeners with different degree of ortho-chloro substitution. substitution.The Theincreased increasedretention retentiontime timeof ofthe thesecond secondeluted elutedenantiomer enantiomerisismarked markedwith withan anarrow. arrow. ◦ 10 10 m m ××0.25 0.25mm mmI.D. I.D.fused fusedsilica silicacolumn column coated coated with with the the C C6060phase phase(0.25 (0.25μm), µm),190 190 °C, C,carrier: carrier: 0.6 0.6bar bar (at (at gauge) gauge) helium, helium, electron electron capture capture detection. detection. Reprinted Reprinted with with permission permission from [93].

5. 5. The TheReciprocal ReciprocalSelectand/Selector Selectand/SelectorSystem SystemofofFullerenes/Cyclodextrins Fullerenes/Cyclodextrins n =n , A system parpar excellence is represented by fullerenes (Cn, (C A supramolecular supramolecularselectand/selector selectand/selector system excellence is represented by fullerenes number of carbon atoms) and cyclodextrins (CDn, n = number of D -glucose building blocks). It also n = number of carbon atoms) and cyclodextrins (CDn, n = number of D-glucose building blocks). lends the principle of reciprocal recognition. According to Figure 23, the23,role selector and It alsoitself lendstoitself to the principle of reciprocal recognition. According to Figure theof role of selector selectand can be reversed. and selectand can be reversed. The The non-chromatographic non-chromatographic supramolecular supramolecular recognition recognition phenomenon phenomenon between between fullerenes fullerenes and and cyclodextrins is well established. γ-Cyclodextrin (CD8) forms water-soluble bicapped 2:1 association cyclodextrins is well established. γ-Cyclodextrin (CD8) forms water-soluble bicapped 2:1 association complexes complexes with with [60]fullerene [60]fullerene (C (C6060))[95–101]. [95–101]. Also Also β-cyclodextrin β-cyclodextrin(CD7) (CD7)may may undergo undergo complexation complexation with C 60 [102,103]. with C60 [102,103]. Shape of C C60 and C70 on native γ-cyclodextrin chemically Shape selectivity selectivity governs governs the the HPLC HPLC separation separation of 60 and C70 on native γ-cyclodextrin chemically bonded to silica [104]. C 70 was much more strongly retained than C60 on this stationary phase whereas bonded to silica [104]. C70 was much more strongly retained than C60 on this stationary phase whereas no of the the two two fullerenes fullerenesoccurred occurredon onthe thecorresponding corresponding unmodified silica indicating no separation separation of unmodified silica indicating thatthat the the separation is due to the selective supramolecular interaction with the cyclodextrin selector [104]. separation is due to the selective supramolecular interaction with the cyclodextrin selector [104]. InIna areciprocal reciprocalstrategy, strategy,Bianco Biancoetetal.al.used usedaaCC60-bonded -bonded phase phase (Figure (Figure 24) 24) in in an an aqueous/organic aqueous/organic medium medium 60 for the high-efficiency HPLC separation of α-, βand γ-cyclodextrins (CD6-CD8) [105]. As expected, for the high-efficiency HPLC separation of α-, β- and γ-cyclodextrins (CD6-CD8) [105]. As expected, the the trace trace shows shows increased increased retention retention with with increased increased molecular molecular weight weight of of CDn CDn (Figure (Figure 25). 25). With the advent of the enzymatic access to large-ring cyclodextrins [106,107], separation With the advent of the enzymatic access to large-ring cyclodextrins [106,107], separationof of the the congeners CDn represented a considerable analytical challenge. Koizumi et al. separated CD6-CD85, congeners CDn represented a considerable analytical challenge. Koizumi et al. separated CD6-CD85, obtained by the action of cyclodextrin glycosyltransferase from Bacillus macerans, on synthetic amylose by high-performance anion-exchange chromatography (HPAEC) with pulsed amperometric

Molecules 2016, 21, 1535

22 of 35

obtained by the action of cyclodextrin glycosyltransferase from Bacillus macerans, on synthetic amylose Molecules 22 Molecules 2016, 2016, 21, 21, 1535 1535 22 of of 35 35 by high-performance anion-exchange chromatography (HPAEC) with pulsed amperometric detection using a 25 cm × 4 amm I.D. ×Dionex CarboPac PA-100 column (Figure 26) [108], whereas Bogdanski et detection detection using using a 25 25 cm cm × 44 mm mm I.D. I.D. Dionex Dionex CarboPac CarboPac PA-100 PA-100 column column (Figure (Figure 26) 26) [108], [108], whereas whereas al. Bogdanski separated CD6-CD21 by liquid-chromatography with electrospray ionization mass spectrometric Bogdanski et et al. al. separated separated CD6-CD21 CD6-CD21 by by liquid-chromatography liquid-chromatography with with electrospray electrospray ionization ionization mass mass detection (LC/ESI-MS) using a 25 cm ×using 4 mmaa 25 I.D. LiChrospher [109]. With one exception spectrometric detection (LC/ESI-MS) ×× 44 mm LiChrospher 22 column [109]. With 2 column NH spectrometric detection (LC/ESI-MS) using 25 cm cm mm I.D. I.D. NH LiChrospher NH column [109]. With (CD9 on HPAEC) the congeners CDn were eluted from the stationary phases according to the degree one exception (CD9 on HPAEC) the congeners CDn were eluted from the stationary phases according one exception (CD9 on HPAEC) the congeners CDn were eluted from the stationary phases according of polymerization (Figure 26). to the degree of polymerization (Figure 26). to the degree of polymerization (Figure 26).

Figure Schematic representation of the reciprocal principle cyclodextrins. Figure 23.23. Schematic betweenfullerenes fullerenesand and cyclodextrins. Figure 23. Schematicrepresentation representationof ofthe thereciprocal reciprocal principle principle between between fullerenes and cyclodextrins. Reproduced from Bogdanski, Doctoral Thesis, University Reproduced from A.A. Bogdanski, Tübingen,Tübingen, Tübingen,Germany, Germany,2007. 2007. Reproduced from A. Bogdanski,Doctoral DoctoralThesis, Thesis,University University of Tübingen, Tübingen, Tübingen, Germany, 2007.

Figure Synthetic pathway 6,6-[60]fulleropyrrolidine via Figure 24. Synthetic pathwaytoto toaaa6,6-[60]fulleropyrrolidine 6,6-[60]fulleropyrrolidine derivative derivative by aziridine ring opening viavia Figure 24.24. Synthetic pathway derivativeby byaziridine aziridinering ringopening opening 1,3-dipolar addition to C 60 and grafting the fullerene-containing triethoxysilane to silica microparticles 1,3-dipolar addition C60and andgrafting graftingthe thefullerene-containing fullerene-containing triethoxysilane 1,3-dipolar addition to to C60 triethoxysilanetotosilica silicamicroparticles microparticles leading to aa chemically homogeneous material with defined and high chromatographic leading chemicallyhomogeneous homogeneousmaterial material with with defined defined structure structure leading to to a chemically structure and andhigh highchromatographic chromatographic efficiency. Reprinted with permission from [105]. Copyright (1997) American Chemical Society. efficiency. Reprinted with permission from [105]. Copyright (1997) American Chemical efficiency. Reprinted with permission from [105]. Copyright (1997) American ChemicalSociety. Society.

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

Molecules 2016, 21, 1535

23 of 35 23 of 35

23 of 35

Figure 25. 25. Separation Separation of of α-, α-, ββ- and and γ-cyclodextrins γ-cyclodextrins (CD6, (CD6, CD7, CD7, CD8) CD8) on on immobilized immobilized C C60 60 by HPLC. Figure Figure 25. Separation of α-, β- and γ-cyclodextrins (CD6, CD7, CD8) on immobilized C60 by HPLC. 250 ××1.8 stainless steel column packed withwith [60]fulleropyrrolidine-based silica (Hypersil, 5 μm), 250 1.8mm mmI.D. I.D. stainless steel column packed [60]fulleropyrrolidine-based silica (Hypersil, 250 × 1.8 mm I.D. stainless steel column packed with [60]fulleropyrrolidine-based silica (Hypersil, 5 μm), 0.25 mL/min water/methanol/tetrahydrofuran (80/10/10 for 3 min and linear gradient to 5eluent: µm), eluent: 0.25 mL/min water/methanol/tetrahydrofuran (80/10/10 for then 3 min and then linear eluent: 0.25 mL/min water/methanol/tetrahydrofuran (80/10/10 for 3 min and then linear gradient to ◦ 40/30/30 in 40/30/30 25 atinat 25 permission from[105]. [105]. Copyright (1997) American gradient to 2525°C. min, at 25 C. with Reprinted with from permission from [105]. Copyright (1997) 40/30/30 in min, 25 min, °C.Reprinted Reprinted with permission Copyright (1997) American American Chemical Chemical Society. Chemical Society.Society.

Figure 26. Separation cyclodextrincongeners congeners (CD6–CD85) obtained by cyclodextrin Figure 26. Separation of of cyclodextrin (CD6–CD85)enzymatically enzymatically obtained by cyclodextrin glycosyltransferase from Bacillus macerans on synthetic amylose by high-performance anion-exchange glycosyltransferase from Bacillus macerans on synthetic amylose by high-performance anion-exchange Figure 26. Separation(HPAEC) of cyclodextrin congeners (CD6–CD85) enzymatically by cyclodextrin chromatography with pulsed amperometric detection using a 25 cmobtained × 4 mm I.D. Dionex chromatography (HPAEC) with pulsed amperometric detection using a 25 cm × 4 mm I.D. Dionex glycosyltransferase Bacillus macerans syntheticfrom amylose CarboPac PA-100from column. Reprinted withon permission [108].by high-performance anion-exchange CarboPac PA-100 column. Reprinted with permission from [108]. chromatography (HPAEC) with pulsed amperometric detection using a 25 cm × 4 mm I.D. Dionex In striking contrast, unusual with supramolecular selectivity CarboPac PA-100 column.an Reprinted permission from [108].pattern has been observed for the reciprocal separation of an CD6–CD15 on a fullerene C60 selector anchored to silica particles In striking contrast, unusualcongeners supramolecular selectivity pattern has been observed for the (Figure 27), which strongly deviates from the usually observed correlation of retention and molecular reciprocal separation of CD6–CD15 congeners on a fullerene C60 selector silica particles In striking contrast, an unusual supramolecular selectivity pattern anchored has been to observed for the weight [110]. Thus, whereas CD6 and CD7the exhibit a lowobserved retention correlation and are eluted ~5 min, CD8–CD10 (Figure 27), which strongly deviates from usually ofin retention reciprocal separation of CD6–CD15 congeners on a fullerene C60 selector anchored to and silicamolecular particles are eluted only after a long retention gap of about 15 min with the elution order CD8 < CD10 < CD9 in weight Thus,strongly whereas deviates CD6 andfrom CD7 the exhibit a low retentioncorrelation and are eluted in ~5 min, (Figure[110]. 27), which usually observed of retention andCD8–CD10 molecular ~20 min. The higher CD11-CD15 congeners are eluted together with CD6 in ~5 min (Figure 27). The are eluted only afterwhereas a long retention gap of abouta15 min with the elution orderinCD8 < CD10 < CD9 weight [110]. Thus, CD6 and CD7 exhibit low retention and are eluted ~5 min, CD8–CD10 drop of retention between CD10 and CD11 of 15 min is indeed remarkable. The unusual elution order in ~20 min. The higher CD11-CD15 congeners are eluted together with CD6 in ~5 min (Figure are eluted onlydetected after a long gap of about 15 min with the elution(ESI-MS) order CD8 CD10 10) [112] may reveal clues of molecular recognition pattern. molecular recognition pattern.

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

25 of 35 25 of 35 25 of 35

Figure 28. Structure of the silica-bonded spacer stationary phase (top) and the silica-bonded C60 stationary Figure of thethe silica-bonded spacer stationary phase (top) the and silica-bonded C60 stationary Figure 28. 28.Structure Structure silica-bonded spacer stationary phaseand (top) the silica-bonded C60 phase (bottom) [110].ofReproduced from A. Bogdanski, Doctoral Thesis, University of Tübingen, phase (bottom) [110]. Reproduced from A. Bogdanski, Doctoral Thesis, University of Tübingen, stationary phase (bottom) [110]. Reproduced from A. Bogdanski, Doctoral Thesis, University of Tübingen, Germany, 2007. Tübingen, Tübingen, Germany, 2007. Germany, 2007. Table 5. Apparent complexation constants Krel of cyclodextrins and silica-bonded [60]fullerene (C60) Table 5. Apparent Apparent complexation complexation constants constants K Krel rel of cyclodextrins and silica-bonded [60]fullerene (C60 60)) Table 5. related to CD6 (α-cyclodextrin) [110]. related to CD6 (α-cyclodextrin) [110].

Krel Krel

CD6 CD6 1 1

Krel

CD7 CD8 CD7 CD8 CD7 CD8 39 4.5 39

CD6 4.5 1

4.5

39

CD9 CD9 46

CD9 46 CD10 46

42

CD10 CD10 CD11 42 42 1

CD11 CD11 1

CD12 1 2

CD12 CD12 2 2

It has been speculated [110] that by using a single [78]fullerene selector (C78), the most strongly It has been speculated [110] that by using a single [78]fullerene selector (C78), the most strongly interacting CD-congener can be identified and, according to the reciprocal(Cprinciple, could then It has been speculatedcan [110] by using a single [78]fullerene strongly 78 ), the most interacting CD-congener be that identified and, according to the selector reciprocal principle, could then applied as a silica-bonded resolving agent for the separation of all five isomeric [78]fullerenes interacting can be identified and,for according to the reciprocal could[78]fullerenes then applied applied as CD-congener a silica-bonded resolving agent the separation of all principle, five isomeric including the direct enantioseparation of chiral D 3-[78]-fullerene [113]. as a silica-bonded agent for the separation of all five isomeric including the directresolving enantioseparation of chiral D3-[78]-fullerene [113]. [78]fullerenes including the direct enantioseparation of chiral D3 -[78]-fullerene [113]. 6. Chromatography of Calixarenes and the Reciprocal Principle 6. Chromatography of Calixarenes and the Reciprocal Principle 6. Chromatography of the Calixarenes and Reciprocal Principle Gutsche reviewed application ofthe calixarenes (Figure 29) as reciprocal selectors and selectands Gutsche reviewed the application of calixarenes (Figure 29) as reciprocal selectors and selectands [114].Gutsche reviewed the application of calixarenes (Figure 29) as reciprocal selectors and selectands [114]. [114].

Figure 29. 29. The structure of calix[4]arene. Reproduced with permission of the Royal Chemistry Society of Figure of calix[4]arene. Reproduced with permission of the Royal Society Figure 29.The Thestructure structure of calix[4]arene. Reproduced with permission of the Royalof Society of Chemistry from [114]. from [114]. from Chemistry [114].

Following the purification of C60- and C70-fullerenes with calix[8]arene [115,116], a reciprocal Following the purification of C -fullerenes with with calix[8]arene calix[8]arene [115,116], [115,116], a reciprocal C60 60-- and C70 70-fullerenes separation of calix[4]arene, calix[6]arene and calix[8]arene (Figure 30, top) by microcolumn liquid calix[6]arene and calix[8]arene calix[8]arene (Figure 30, top) by microcolumn liquid separation of calix[4]arene, calix[6]arene chromatography (μ-LC) on a chemically-bonded C60 silica stationary phase (Figure 21) has been chromatography (μ-LC) on a chemically-bonded C60 silica stationary phase (Figure 21) has been reported by Saito et al. [117]. An improved sseparation of the same calixarene congeners by μ-LC on reported by Saito et al. [117]. An improved sseparation of the same calixarene congeners by μ-LC on

Molecules 2016, 21, 1535

26 of 35

chromatography Molecules 2016, 21, 1535(µ-LC)

on a chemically-bonded C60 silica stationary phase (Figure 21) has26been of 35 reported by Saito et al. [117]. An improved sseparation of the same calixarene congeners by µ-LC a 6,6-[60]fulleropyrrolidine silica stationary phase (Figure waslater laterreported reportedby byBianco Biancoet et al. al. aon6,6-[60]fulleropyrrolidine silica stationary phase (Figure 24)24) was (Figure 30, bottom) [105].

Figure 30. Top: Structures of tert-butyl-calix[n]arenes. Bottom: HPLC trace for the separation of Figure 30. Top: Structures of tert-butyl-calix[n]arenes. Bottom: HPLC trace for the separation of tert-butylcalix[n]arenes on a 6,6-[60]fulleropyrrolidine stationary phase (Figure 24). 250 × 1.8 mm I.D. tert-butylcalix[n]arenes on a 6,6-[60]fulleropyrrolidine stationary phase (Figure 24). 250 × 1.8 mm I.D. stainless steel column. Eluent: CH2Cl2/2-propanol (99.5/0.5), flow rate 0.3 mL/min; T: 25 °C; UV detection stainless steel column. Eluent: CH2 Cl2 /2-propanol (99.5/0.5), flow rate 0.3 mL/min; T: 25 ◦ C; at 280 nm. Reprinted with permission from [105]. Copyright (1997) American Chemical Society. UV detection at 280 nm. Reprinted with permission from [105]. Copyright (1997) American Chemical Society.

Achiral chromatographic selectivity has been achieved for various amino acid esters on silica-bonded calix[4]arene tetraester [118,119]. The enantioseparation of binaphthyl atropisomers on Achiral chromatographic selectivity has been achieved for various amino acid esters on the chiral acylcalix[4]arene amino acid derivatives (N-l-alaninoacyl)calix[4]arene and (N-l-valinoacyl) silica-bonded calix[4]arene tetraester [118,119]. The enantioseparation of binaphthyl atropisomers on calix[4]arene) by electrokinetic chromatography (EKC) has been reported [120]. By enantioselective the chiral acylcalix[4]arene amino acid derivatives (N-l-alaninoacyl)calix[4]arene and (N-l-valinoacyl) liquid chromatography a chiral calixarene functionalized with (−)-ephedrine at the lower rim and calix[4]arene) by electrokinetic chromatography (EKC) has been reported [120]. By enantioselective chemically bonded to silica has been used to separate racemic 1-phenyl-2,2,2-trifluoroethanol [121]. liquid chromatography a chiral calixarene functionalized with (−)-ephedrine at the lower rim and A chiral resorc[4]arene selector [122] of the basket-type containing four ω-unsaturated alkyl chains chemically bonded to silica has been used to separate racemic 1-phenyl-2,2,2-trifluoroethanol [121]. was synthesized from resorcinol and 1-undecenal [123]. Afterwards chiral L-valine-tert-butylamide A chiral resorc[4]arene selector [122] of the basket-type containing four ω-unsaturated alkyl chains moieties were attached to the eight hydroxyl groups. The chiral resorc[4]arene was then linked via was synthesized from resorcinol and 1-undecenal [123]. Afterwards chiral L-valine-tert-butylamide the ω-unsaturated alkyl chains to poly(dimethyl-methylhydro)siloxane via platinum-catalyzed moieties were attached to the eight hydroxyl groups. The chiral resorc[4]arene was then linked hydrosilylation to yield chemically bonded Chirasil-Calix-Val (Figure 31) [123]. On a 20 m × 0.25 mm via the ω-unsaturated alkyl chains to poly(dimethyl-methylhydro)siloxane via platinum-catalyzed I.D. fused silica capillary column coated with Chirasil-Calix (0.25 μm) the N(O,S)-trifluoroacetyl-D,Lhydrosilylation to yield chemically bonded Chirasil-Calix-Val (Figure 31) [123]. On a 20 m × 0.25 mm α-amino acids Ala, Abu, Val, Ile, Leu, Pro, Phe, Glu, Tyr, Orn and Lys were enantioseparated by GC I.D. fused silica capillary column coated with Chirasil-Calix (0.25 µm) the N(O,S)-trifluoroacetyl-D,Lwith Rs > 2 [122]. α-amino acids Ala, Abu, Val, Ile, Leu, Pro, Phe, Glu, Tyr, Orn and Lys were enantioseparated by GC with Rs > 2 [122].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

27 of 35 27 of 35

Synthesis of the CSP Chirasil-Calix. Reprinted Reprinted with permission from [123]. Figure 31. Synthesis

In dual chiral chiral recognition recognition system system Chirasil-Calix-Val-Dex, Chirasil-Calix-Val-Dex, aa resorc[4]arene resorc[4]arene containing containing In the the dual L permethylated β-cyclodextrin L-valine -valine diamide diamide (Calix-Val) (Calix-Val) and and permethylated β-cyclodextrin (CD) (CD) were were chemically chemically linked linked to to poly(dimethylsiloxane) [124]. The selector system belongs to the category of mixed chiral stationary poly(dimethylsiloxane) [124]. The selector system belongs to the category of mixed chiral stationary phases CSP both N-TFA-valine ethyl ester (due(due the presence of Calix-Val) and phases [125]. [125]. On Onthis thisbinary binary CSP both N-TFA-valine ethyl ester the presence of Calix-Val) chiral unfunctionalized 1,2-trans-dialkylcyclohexanes (due(due to to thethe presence and chiral unfunctionalized 1,2-trans-dialkylcyclohexanes presenceofofthe the CD) CD) were were simultaneously enantioseparated (Figure 32) [124]. In order to account for matched and mismatched simultaneously enantioseparated (Figure 32) [124]. In order to account for matched and mismatched enantioselectivities imparted by by the thetwo twoselectors, selectors,the thechirality chiralityofofCalix-Val Calix-Val could changed from enantioselectivities imparted could bebe changed from to to L to D . When racemic Calix-LD-Val were used, any observed enantioselectivity could only arise L to D . When racemic Calix-LD-Val were used, any observed enantioselectivity could only arise from from the presence CD [124]. the presence of CDof[124]. In example above, above, the the chiral chiral resorc[4]arene resorc[4]arene is is used used as as an an enantioselective enantioselective selector. selector. In In the the example In aa reversed fashion the enantiomers of a chiral resorc[4]arene compound can also be studied as selectands. reversed fashion the enantiomers of a chiral resorc[4]arene compound can also be studied as selectands. Thus, tetrabenzoxacine resorc[4]arene CSP Thus, tetrabenzoxacine resorc[4]arene (Figure (Figure 33, 33, top) top) was was enantioseparated enantioseparated by by HPLC HPLC on on the the CSP Chiralpak Okamoto et et al. al. [126]. [126]. The The chromatograms chromatograms show show temperature-dependent temperature-dependent Chiralpak AD AD developed developed by by Okamoto interconversion profiles due to enantiomerization during enantioseparation (Figure interconversion profiles due to enantiomerization during enantioseparation (Figure 33, 33, bottom) bottom)[127]. [127]. By peak-form analysis featuring plateau formation [128], the inversion barrier was determined By peak-form analysis featuring plateau formation [128], the inversion barrier was determined to to ΔG = 92 ± 2 kJ/mol (298K) [127]. ∆G = 92 ± 2 kJ/mol (298K) [127].

Molecules 2016, 21, 1535 Molecules 2016, 21, 1535 Molecules 2016, 21, 1535

28 of 35 28 of 35 28 of 35

Figure 32. 32. The The mixed mixed CSP CSP Chirasil-Calix-Val-Dex. Chirasil-Calix-Val-Dex. Figure Figure 32. The mixed CSP Chirasil-Calix-Val-Dex.

Figure 33. Top: Interconverting enantiomers of tetrabenzoxacine resorc[4]arene by HPLC on the CSP Figure Top: Interconverting enantiomers tetrabenzoxacine resorc[4]arene HPLC CSP Figure 33.33. Top: Interconverting enantiomers of of tetrabenzoxacine resorc[4]arene byby HPLC onon thethe CSP Chiralpak AD. Bottom: Distorted peak profiles caused by enantiomerization (dotted line: expected Chiralpak AD. Bottom: Distorted peak profiles caused by enantiomerization (dotted line: expected Chiralpak AD. Bottom: Distorted peak profiles caused enantiomerization (dotted line: expected peak profiles devoid of interconversion). Reprinted withbypermission from [127]. peak profiles devoid of interconversion). Reprinted with permission from [127]. peak profiles devoid of interconversion). Reprinted with permission from [127].

This finding is recalled here, because one may speculate to employ interconverting calixarene This finding is recalled here, abecause one may speculate employofinterconverting calixarene enantiomers in a reciprocal fashion dynamic the spirit to of ato scenario column deracemization This finding is recalled here,as because oneCSP mayinspeculate employ interconverting calixarene enantiomers in a reciprocal fashion as aan dynamic CSP in the spirit of a scenario of column deracemization described byinWelch [129].fashion Thus when enantiomerically pure which interacts with high enantiomers a reciprocal as a dynamic CSP in the spirit of selectand, a scenario of column deracemization described by Welch [129]. Thus when an enantiomerically pure selectand, which interacts with high affinity and enantioselectivity with an interconverting selector present in the stationary phase, is added described by Welch [129]. Thus when an enantiomerically pure selectand, which interacts with high affinity and enantioselectivity with an interconverting selector present in the stationary phase, is added to the mobile phase, it may cause of the selector, thereby creating a nonracemic for affinity and enantioselectivity withderacemization an interconverting selector present in the stationary phase, isCSP added to the mobile phase, it may cause deracemization of the selector, thereby This creating a nonracemic CSP for subsequent enantiomeric separations of the selectand in the its racemic intriguing was to the mobile phase, it may cause deracemization of selector,form. thereby creating a scenario nonracemic subsequent enantiomeric separations of the selectand in its racemic form. This intriguing scenario was demonstrated by HPLC with a conformationally naphthamide atropisomer it was CSP for subsequent enantiomeric separations of labile the selectand in its racemic form.[129]. ThisFirst intriguing demonstrated by HPLC with a conformationally labile naphthamide atropisomer [129]. First it was shown that naphthamide can bewith enantioseparated on the Whelk-O1 CSP (Figure 3c). Under scenario was the demonstrated by HPLC a conformationally labile naphthamide atropisomer [129]. shown that conditions the naphthamide can be enantioseparated CSPWith (Figure Under stopped-flow [130], deracemization took placeon in the the Whelk-O1 mobile phase. the 3c). reciprocal stopped-flow conditions [130], deracemization took place in the mobile phase. With the reciprocal concept the racemic naphthamide was then linked to silica whereas the soluble analog of the concept the racemic naphthamide was then linked to silica whereas the soluble analog of the

Molecules 2016, 21, 1535

29 of 35

First it was shown that the naphthamide can be enantioseparated on the Whelk-O1 CSP (Figure 3c). Under stopped-flow conditions [130], deracemization took place in the mobile phase. With the reciprocal concept the racemic naphthamide was then linked to silica whereas the soluble analog of the enantiomerically pure Whelk-O1 selector was added to the mobile phase. After deracemization of the CSP, the racemic analog of the Whelk-O1 selector was enantioseparated, however, with the expected subsequent decrease of the separation factor α due to gradual racemization of the CSP [129]. The reciprocal concept of interconverting molecules in separation methods could also be extended to the liquid chromatographic investigation of conformational cis/trans-diastereomers of L-peptidyl- L-proline. Studies of proline-containing peptides separated on calixarene-bonded silica gels have been investigated by Gebauer et al. [131] and interconversion barriers of L-proline containing dipeptides by dynamic CE were determined [132]. 7. Conclusions Since the first reports by Mikeš and Pirkle on the use of a reciprocal approach in chromatography through changing the role of selectands and selectors, many examples have been reported in different areas of selective supramolecular interactions including combinatorial advances. Advantages and limitations of the reciprocal concept are outlined. The reciprocal strategy clearly aids the development of optimized selectors for difficult separations of selectands including enantiomers. Various scenarios of the interchanging role of selectors and selectands in separation science are discussed in the hope that this will stir further endeavors devoted to the principle of reciprocity in supramolecular recognition. The present account should be seen as part of the emerging field of supramolecular separation methods [1,2] with further emphasis also to molecular sensor approaches and surface related recognition phenomena. The optimization of selector/selectand systems is the precondition for meaningful mechanistic studies by various molecular modelling approaches. Acknowledgments: This work was partially funded by the German Research Council (DFG Schu 395/18-2) in 2005 on ‘Combinatorial Chiral Selectors for the Enantiomeric Separation by Chromatography and Electromigration’. Conflicts of Interest: The author declares no conflict of interest.

References 1. 2.

3. 4. 5. 6. 7. 8. 9. 10.

Schurig, V. Supramolecular Chromatography. In Applications of Supramolecular Chemistry; Schneider, H.-J., Ed.; CRC-Press: Boca Raton, FL, USA, 2012; pp. 129–157. Ciogli, A.; Kotoni, D.; Gasparrini, F.; Pierini, M.; Villani, C. Chiral Supramolecular Selectors for Enantiomer Differentiation in Liquid Chromatography. In Differentiation of Enantiomers I; Schurig, V., Ed.; Springer: Heidelberg, Germany, 2013; Volume 340, pp. 73–106. Allenmark, S.; Schurig, V. Chromatography on chiral stationary phases. J. Mater. Chem. 1997, 7, 1955–1963. [CrossRef] Mikeš, F. The Resolution of Chiral Compounds by Modern Liquid Chromatography. Ph.D. Thesis, The Weizmann Institute of Science, Rehovot, Israel, 1975. Ashby, W.R. An Introduction to Cybernetics; Chapman & Hall Ltd.: London, UK, 1956; p. 10. Allenmark, S. Separation of Enantiomers by Protein Based Chiral Phases. In A Practical Approach to Chiral Separations by Liquid Chromatography; Subramanian, G., Ed.; Wiley-VCH: Weinheim, Germany, 1994; p. 183. Okamoto, Y.; Yashima, E. Polysaccharide derivatives for chromatographic separation of enantiomers. Angew. Chem. Int. Ed. 1998, 37, 1020–1043. [CrossRef] Zeleke, Y.M.; Smith, G.B.; Hofstetter, H.; Hofstetter, O. Enantiomer separation of Amino Acids in immunoaffinity Micro LC-MS. Chirality 2006, 18, 544–550. [CrossRef] [PubMed] Mikeš, F.; Boshart, G.; Gil-Av, E. Helicenes, Resolution on chiral charge-transfer complexing agents using high performance liquid chromatography. J. Chem. Soc. Chem. Commun. 1976, 99–100. [CrossRef] Mikeš, F.; Boshart, G.; Gil-Av, E. Resolution of optical isomers by high-performance liquid chromatography, using coated and bonded chiral charge-transfer complexing agents as stationary phases. J. Chromatogr. 1976, 122, 205–221. [CrossRef]

Molecules 2016, 21, 1535

11. 12.

13. 14. 15. 16. 17. 18. 19.

20. 21. 22. 23. 24.

25. 26.

27. 28. 29. 30. 31. 32. 33.

34.

30 of 35

Mikeš, F.; Boshart, G. Resolution of optical isomers by high-performance liquid chromatography—A comparison of two selector systems. J. Chromatogr. 1978, 149, 455–464. [CrossRef] Kim, Y.H.; Balan, A.; Tishbee, A.; Gil-Av, E. Chiral differentiation by the P-(+)-Hexahelicene-7,70 -dicarboxylic acid disodium salt. Resolution of N-2,4-Dinitrophenyl-α-amino-acid esters by high performance liquid chromatography. J. Chem. Soc. Chem. Commun. 1982, 1336–1337. [CrossRef] Pirkle, W.H. The nonequivalence of physical properties of enantiomers in optically active solvents. Differences in nuclear magnetic resonance spectra. I. J. Am. Chem. Soc. 1966, 88, 1837. [CrossRef] Pirkle, W.H.; Burlingame, T.G.; Beare, S.D. Optically Active NMR Solvents VI. The determination of optical purity and absolute configuration of amines. Tetrahedron Lett. 1968, 9, 5849–5852. [CrossRef] Pirkle, W.H.; Hoover, D.J. NMR chiral solvating agents. Top. Stereochem. 1982, 13, 263–331. Pirkle, W.H.; House, D.W.; Finn, J.M. Broad spectrum resolution of optical isomers using chiral high-performance liquid chromatographic bonded phases. J. Chromatogr. 1980, 192, 143–158. [CrossRef] Pirkle, W.H.; Däppen, R. Reciprocity in chiral recognition: Comparison of several chiral stationary phases. J. Chromatogr. A 1987, 404, 107–115. [CrossRef] Pirkle, W.H.; Hyun, M.H.; Bank, B. A rational approach to the design of highly-effective chiral stationary phases. J. Chromatogr. 1984, 316, 585–604. [CrossRef] Pirkle, W.H.; Pochapsky, T.C.; Burke, J.A., III; Deming, K.C. Systematic Studies of Chiral Recognition Mechanisms. In Chiral Separations; Stevenson, D., Wilson, I.D., Eds.; Plenum Press: New York, NY, USA, 1988; pp. 23–35. Pirkle, W.H.; Pochapsky, T.C. Considerations of chiral recognition relevant to the liquid chromatographic separation of enantiomers. Chem. Rev. 1989, 89, 347–362. [CrossRef] Pirkle, W.H.; Readnour, R.S. Chromatographic approach to the measurement of the interstrand distance for some chiral bonded phases. Anal. Chem. 1991, 63, 16–20. [CrossRef] Pirkle, W.H.; Welch, C.J.; Lamm, B. Design, synthesis, and evaluation of an improved enantioselective naproxen selector. J. Org. Chem. 1992, 57, 3854–3860. [CrossRef] Welch, C.J. Evolution of chiral stationary phase design in the Pirkle laboratories. J. Chromatogr. A 1994, 666, 3–26. [CrossRef] Taylor, D.R.; Clift, E. The Use of Reciprocal Interaction in the Design of Pyrethroid Specific Chiral Stationary Phases. In Recent Advances in Chiral Separations; Stevenson, D., Wilson, I.D., Eds.; Plenum Press: New York, NY, USA, 1990; pp. 39–41. Terabe, S. Electrokinetic chromatography: An interface between electrophoresis and chromatography. Trends Anal. Chem. 1984, 8, 129–134. [CrossRef] Chankvetadze, B. Principles of Enantiomer Separation in Electrokinetic Chromatography. In Electrokinetic Chromatography—Theory, Instrumentation and Applications; Pyell, U., Ed.; John Wiley & Sons: Chichester, UK, 2006; pp. 179–206. Jakubetz, H.; Juza, M.; Schurig, V. Electrokinetic Chromatography employing an anionic and a cationic β-cyclodextrin derivative. Electrophoresis 1997, 18, 897–904. [CrossRef] [PubMed] Terabe, S.; Ozaki, H.; Otsuka, K.; Ando, T. Electrokinetic chromatography with 2-O-Carboxymethyl-βcyclodextrin as a moving “Stationary” Phase. J. Chromatogr. 1985, 332, 211–217. [CrossRef] Mayer, S.; Schleimer, M.; Schurig, V. Dual chiral recognition system involving cyclodextrin derivatives in capillary electrophoresis. J. Microcolumn Sep. 1994, 6, 43–48. [CrossRef] Tait, R.J.; Thompson, D.O.; Stella, V.J.; Stobaugh, J.F. Sulfobutyl ether β-cyclodextrin as a chiral discriminator for use with capillary electrophoresis. Anal. Chem. 1994, 66, 4013–4018. [CrossRef] Chankvetadze, B.; Endresz, G.; Blaschke, G. About some aspects of the use of charged cyclodextrins for capillary electrophoresis enantioseparation. Electrophoresis 1994, 15, 804–807. [CrossRef] [PubMed] Chankvetadze, B.; Blaschke, G. Enantioseparations in capillary electromigration techniques: Recent developments and future trends. J. Chromatogr. A 2001, 906, 309–363. [CrossRef] Guttman, A.; Paulus, A.; Cohen, A.S.; Grinberg, N.; Karger, B.L. Use of complexing agents for selective separation in high-performance capillary electrophoresis: Chiral resolution via cyclodextrins incorporated within polyacrylamide gel columns. J. Chromatogr. 1988, 448, 41–53. [CrossRef] Wistuba, D.; Bogdanski, A.; Larsen, K.L.; Schurig, V. δ-Cyclodextrin as novel chiral probe for enantiomeric separation by electromigration methods. Electrophoresis 2006, 27, 4359–4363. [CrossRef] [PubMed]

Molecules 2016, 21, 1535

35. 36. 37. 38. 39. 40.

41. 42.

43. 44.

45.

46. 47.

48.

49. 50.

51. 52. 53. 54. 55.

56.

31 of 35

Mayer, S.; Schurig, V. Enantiomer separation by electrochromatography on capillaries coated with Chirasil-Dex. J. Sep. Sci. 1992, 15, 129–130. Mayer, S.; Schurig, V. Enantiomer separation by electrochromatography on open tubular columns coated with Chirasil-Dex. J. Liq. Chromatogr. 1993, 16, 915–931. [CrossRef] Trapp, O.; Trapp, G.; Schurig, V. University of Tübingen, Tübingen, Germany. Unpublished work. 2003. Li, T. Peptide and peptidomimetic chiral selectors in liquid chromatography. J. Sep. Sci. 2005, 28, 1927–1931. [CrossRef] [PubMed] Li, T. Development of efficient chiral stationary phases from peptide libraries. LabPlus Int. 2005, 19, 15–18. Svec, F.; Wulff, D.; Fréchet, J.M. Combinatorial Approaches to Recognition of Chirality: Preparation and Use of Materials for the Separation of Enantiomers. In Chiral Separation Techniques, 2nd ed.; Subramanian, G., Ed.; Wiley-VCH: Weinheim, Germany, 2000; pp. 57–94. Huang, P.Y.; Carbonell, R.G. Affinity purification of proteins using ligands derived from peptide libraries. Biotechnol. Bioeng. 1995, 47, 288–297. [CrossRef] [PubMed] Kaufman, D.B.; Hentsch, M.E.; Baumbach, G.A.; Buettner, J.A.; Dadd, C.A.; Huang, P.Y.; Hammond, D.J.; Carbonell, R.G. Affinity purification of fibrinogen using a ligand from a peptide library. Biotechnol. Bioeng. 2002, 77, 278–289. [CrossRef] [PubMed] Huang, P.Y.; Carbonell, R.G. Affinity chromatographic screening of soluble combinatorial peptide libraries. Biotechnol. Bioeng. 1999, 63, 633–641. [CrossRef] Jung, G.; Hofstetter, H.; Feiertag, S.; Stoll, D.; Hofstetter, O.; Wiesmüller, K.-H.; Schurig, V. Cyclopeptide libraries as new chiral selectors in capillary electrophoresis. Angew. Chem. Int. Ed. 1996, 35, 2148–2150. [CrossRef] Chiari, M.; Desperati, V.; Manera, E.; Longhi, R. Combinatorial synthesis of highly selective cyclohexapeptides for separation of amino acid enantiomers by capillary electrophoresis. Anal. Chem. 1998, 70, 4967–4973. [CrossRef] [PubMed] Bluhm, L.; Huang, J.M.; Li, T.Y. Recent advances in peptide chiral selectors for electrophoresis and liquid chromatography. Anal. Bioanal. Chem. 2005, 382, 592–598. [CrossRef] [PubMed] De Lorenzi, E.; Massolini, G.; Molinari, P.; Galbusera, C.; Longhi, R.; Marinzi, C.; Consonni, R.; Chiari, M. Chiral capillary electrophoresis and nuclear magnetic resonance investigation on the structure-enantioselectivity relationship in synthetic cyclopeptides as chiral selectors. Electrophoresis 2001, 22, 1373–1384. [CrossRef] Marinzi, C.; Longhi, R.; Chiari, M.; Consonni, R. Capillary electrophoresis investigation on the structure-enantioselectivity relationship in synthetic cyclopeptides as chiral selectors. Electrophoresis 2001, 22, 3257–3262. [CrossRef] Schurig, V.; Wistuba, D.; Jung, G.; Wiesmüller, K.-H. University of Tübingen, Tübingen, Germany. Unpublished work. 2010. Wu, Y.; Wang, Y.; Yang, A.; Li, T. Screening of mixture combinatorial libraries for chiral selectors: A reciprocal chromatographic approach using enantiomeric libraries. Anal. Chem. 1999, 71, 1688–1691. [CrossRef] [PubMed] Weingarten, M.D.; Sekanina, K.; Still, W.C. Enantioselective resolving resins from a combinatorial library. Kinetic resolution of cyclic amino acid derivatives. J. Am. Chem. Soc. 1998, 120, 9112–9113. [CrossRef] Maier, N.; Franco, P.; Lindner, W. Separation of enantiomer: Needs, challenges, perspectives. J. Chromatogr. A 2001, 906, 3–33. [CrossRef] Murer, P.; Lewandowski, K.; Svec, F.; Fréchet, J.M.J. Combinatorial “Library on Bead” approach to polymeric materials with vastly enhanced chiral recognition. J. Chem. Soc. Chem. Commun. 1998, 2559–2560. [CrossRef] Murer, P.; Lewandowski, K.; Svec, F.; Fréchet, J.M.J. “On Bead” combinatorial approach to the design of chiral stationary phases. Anal. Chem. 1999, 71, 1278–1284. [CrossRef] [PubMed] Lewandowski, K.; Murer, P.; Svec, F.; Fréchet, J.M.J. Highly selective chiral recognition on polymer supports: Preparation of a combinatorial library of dihydropyrimidines and its screening for novel Chiral HPLC ligands. J. Chem. Soc. Chem. Commun. 1998, 2237–2238. [CrossRef] Lewandowski, K.; Murer, P.; Svec, F.; Fréchet, J.M.J. A combinatorial approach to recognition of chirality: The preparation of highly enantioselective aryldihydropyrimidine selectors for chiral HPLC. J. Comb. Chem. 1999, 1, 105–112. [CrossRef] [PubMed]

Molecules 2016, 21, 1535

57.

58. 59. 60. 61.

62. 63. 64. 65. 66.

67. 68. 69. 70.

71.

72. 73.

74.

75.

76. 77. 78.

32 of 35

Brahmachary, E.; Ling, F.H.; Svec, F.; Fréchet, J.M.J. Chiral recognition: Design and preparation of chiral stationary phases using selectors derived from Ugi multicomponent condensation reactions and a combinatorial approach. J. Comb. Chem. 2003, 5, 441–450. [CrossRef] [PubMed] Pirkle, W.H.; Alessi, D.M.; Hyun, M.H.; Pochapsky, T.C. Separation of some enantiomeric dipeptides and tripeptides on chiral stationary phases. J. Chromatogr. 1987, 398, 203–209. [CrossRef] Welch, C.J.; Bhat, G.; Protopopova, M.N. Silica-based solid phase synthesis of chiral stationary phases. Enantiomer 1988, 3, 463–469. Welch, C.J.; Protopopova, M.N.; Bhat, G.A. Microscale synthesis and screening of chiral stationary phases. Enantiomer 1988, 3, 471–476. Welch, C.J.; Bhat, G.; Protopopova, M.N. Selection of an optimized adsorbent for preparative chromatographic enantioseparation by microscale screening of a second generation chiral phase library. J. Comb. Chem. 1999, 1, 364–367. [CrossRef] [PubMed] Welch, C.J.; Pollard, S.D.; Mathre, D.J.; Reider, P.J. Improved method for rapid evaluation of chiral stationary phase libraries. Org. Lett. 2001, 3, 95–98. [CrossRef] [PubMed] Wang, Y.; Li, T. Screening of a parallel combinatorial library for selectors for chiral chromatography. Anal. Chem. 1999, 71, 4178–4182. [CrossRef] [PubMed] Bluhm, L.H.; Wang, Y.; Li, T. An alternative procedure to screen mixture combinatorial libraries for selectors for chiral chromatography. Anal. Chem. 2000, 72, 5201–5205. [CrossRef] [PubMed] Wang, Y.; Bluhm, L.H.; Li, T. Identification of chiral selectors from a 200-member parallel combinatorial library. Anal. Chem. 2000, 72, 5459–5465. [CrossRef] [PubMed] Blodgett, J.; Wang, Y.; Li, T.; Polavarapu, P.L.; Drabowicz, J.; Pietrusiewicz, K.M.; Zygo, K. Resolution of tert-Butyl-1-(2-methylnaphthyl)phosphine oxide using selectors identified from a chemical combinatorial library. Anal. Chem. 2002, 74, 5212–5216. [CrossRef] [PubMed] Huang, J.; Li, T. Efficient resolution of racemic 1,10 -Bi-2-naphthol with chiral selectors identified from a small library. J. Chromatogr. A 2005, 1062, 87–93. [CrossRef] [PubMed] Chen, Z.; Yang, Y.; Werner, S.; Wipf, P.; Weber, S.G. A screening method for chiral selectors that does not require covalent attachment. J. Am. Chem. Soc. 2006, 22, 2208–2209. [CrossRef] [PubMed] Liu, H.; Tanaka, T.; Kataura, H. Optical isomer separation of single-chirality carbon nanotubes using gel column chromatography. Nano Lett. 2014, 14, 6237–6243. [CrossRef] [PubMed] Zhang, M.; Khripin, C.Y.; Fagan, J.A.; McPhie, P.; Ito, Y.; Zheng, M. Single-step total fractionation of single-wall carbon nanotubes by countercurrent chromatography. Anal. Chem. 2014, 86, 3980–3984. [CrossRef] [PubMed] Sicoli, G.; Pertici, F.; Jiang, Z.; Jicsinszky, L.; Schurig, V. Gas-chromatographic approach to probe the absence of molecular inclusion in enantioseparations by carbohydrates. Investigation of linear dextrins (“acyclodextrins”) as novel chiral stationary phases. Chirality 2007, 19, 391–400. [CrossRef] [PubMed] Raffaini, G.; Ganazzoli, F. Separation of chiral nanotubes with an opposite handedness by chiral oligopeptide adsorption: A molecular dynamics study. J. Chromatogr. A 2015, 1425, 221–230. [CrossRef] [PubMed] Ahmed, M.; Yajadda, M.M.A.; Han, Z.J.; Su, D.; Wang, G.; Ostrikov, K.K.; Ghanem, A. Single-walled carbon nanotube-based polymer monoliths for the enantioselective Nano-liquid chromatographic separation of racemic pharmaceuticals. J. Chromatogr. A 2014, 1360, 100–109. [CrossRef] [PubMed] Alhassan, H.; Antony, V.; Ghanem, A.; Yajadda, M.M.A.; Han, Z.J.; Ostrikov, K.K. Organic/Hybrid nanoparticles and single-walled carbon nanotubes: Preparation methods and chiral applications. Chirality 2014, 26, 683–691. [CrossRef] [PubMed] Silva, R.A.; Talio, M.C.; Luconi, M.O.; Fernandez, L.P. Evaluation of carbon nanotubes as chiral selectors for continuous-flow enantiomeric separation of carvedilol with fluorescent detection. J. Pharm. Biomed. Anal. 2012, 70, 631–635. [CrossRef] [PubMed] Peng, X.; Komatsu, N.; Bhattacharya, S.; Shimawaki, T.; Aonuma, S.; Kimura, T. Optically active single-walled carbon nanotubes. Nat. Nanotechnol. 2007, 2, 361–365. [CrossRef] [PubMed] Zhao, L.; Ai, P.; Duan, A.-H.; Yuan, L.-M. Single-Walled carbon nanotubes for improved enantioseparations on a chiral ionic liquid stationary phase in GC. Anal. Bioanal. Chem. 2011, 399, 143–147. [CrossRef] [PubMed] Jinno, K. Molecular Recognition for Fullerenes in Liquid Chromatography. In Chromatographic Separations Based on Molecular Recognition; Jinno, K., Ed.; Wiley-VCH: New York, NY, USA, 1997; pp. 147–237.

Molecules 2016, 21, 1535

33 of 35

Baena, J.R.; Gallego, M.; Valcárcel, M. Fullerenes in the analytical sciences. Trends Anal. Chem. TRAC 2002, 21, 187–198. [CrossRef] 80. Hawkins, J.M.; Lewis, T.A.; Loren, S.D.; Meyer, A.; Heath, J.R.; Shibato, Y.; Saykally, R.J. Organic chemistry of C60 (Buckminsterfullerene): Chromatography and osmylation. J. Org. Chem. 1990, 55, 6250–6252. [CrossRef] [PubMed] 81. Pirkle, W.H.; Welch, C.J. An unusual effect of temperature on the chromatographic behavior of buckminsterfullerene. J. Org. Chem. 1991, 56, 6973–6974. [CrossRef] 82. Welch, C.J.; Pirkle, W.H. Progress in the design of selectors for buckminsterfullerene. J. Chromatogr. 1992, 609, 89–101. [CrossRef] 83. Kawauchi, T.; Kitaura, A.; Kawauchi, M.; Takeichi, T.; Kumaki, J.; Lida, H.; Yashima, E. Separation of C-70 over C-60 and selective extraction and resolution of higher fullerenes by syndiotactic helical poly(methyl methacrylate). J. Amer. Chem. Soc. 2010, 132, 12191–12193. [CrossRef] [PubMed] 84. Diack, M.; Compton, R.N.; Guiochon, G. Evaluation of stationary phases for the separation of buckminsterfullerenes by high-performance liquid chromatography. J. Chromatogr. 1993, 639, 129–140. [CrossRef] 85. Gross, B.; Schurig, V.; Lamparth, I.; Hirsch, A. Enantiomer separation of [60]Fullerene derivatives by Micro-column-HPLC using (R)-(−)-TAPA as chiral stationary phase. J. Chromatogr. A 1997, 791, 65–69. [CrossRef] 86. Gross, B.; Schurig, V.; Lamparth, I.; Herzog, A.; Djojo, F.; Hirsch, A. Enantiomeric separation of [60]Fullerene derivatives with an inherent chiral addition pattern. J. Chem. Soc. Chem. Commun. 1997, 1117–1118. [CrossRef] 87. Saito, Y.; Ohta, H.; Terasaki, H.; Katoh, Y.; Nagashima, H.; Jinno, K. Separation of polycyclic aromatic hydrocarbons with a C60 bonded silica phase in microcolumn liquid chromatography. J. Sep. Sci. 1995, 18, 569–572. [CrossRef] 88. Stalling, D.L.; Guo, C.; Kuo, K.C.; Saim, S. Fullerene-linked particles as LC chromatographic media and modification of their donor/acceptor properties by secondary chemical reactions. J. Microcolumn Sep. 1993, 5, 223–235. [CrossRef] 89. Stalling, D.L.; Guo, C.Y.; Saim, S. Surface-linked C60/70 -Polystyrene Divinylbenzene beads as a new chromatographic material for enrichment of Co-planar PCBs. J. Chromatogr. Sci. 1993, 31, 265–278. [CrossRef] 90. Golovnya, R.V.; Terenina, M.B.; Ruchkina, E.L.; Karnatsevich, V.L. Fullerene c60 as a new stationary phase in capillary gas chromatography. Mendeleev Commun. 1993, 3, 231–233. [CrossRef] 91. Schurig, V. Molecular Association in Complexation Gas Chromatography. In Chromatographic Separations Based on Molecular Recognition; Jinno, K., Ed.; Wiley-VCH: New York, NY, USA, 1997; pp. 371–418. 92. Schurig, V. Elaborate treatment of retention in chemoselective chromatography—The retention increment approach and non-linear effects. J. Chromatogr. A 2009, 1216, 1723–1736. [CrossRef] [PubMed] 93. Glausch, A.; Hirsch, A.; Lamparth, I.; Schurig, V. Retention Behaviour of polychlorinated biphenyls on Polysiloxane-Anchored C60 in gas chromatography. J. Chromatogr. A 1998, 809, 252–257. [CrossRef] 94. Gross, B. Enantiomerentrennung chiraler Fullerenaddukte mittels LC-Bestimmung von Retentionsinkrementen an Cyclodextrin-, Calixaren- und Fulleren-Stationärphasen. Ph.D. Thesis, University of Tübingen, Tübingen, Germany, 1998. 95. Andersson, T.; Nilsson, K.; Sundahl, M.; Westman, G.; Wennerström, O. C-60 embedded in γ-cyclodextrin: A water-soluble fullerene. J. Chem. Soc. Chem. Commun. 1992, 604–606. [CrossRef] 96. Sundahl, M.; Andersson, T.; Nilsson, K.; Wennerström, O.; Westman, G. Clusters of C60 -fullerene in a water solution containing γ-cyclodextrin—A photophysical study. Synth. Metals 1993, 55–57, 3252–3257. [CrossRef] 97. Priyadarsini, K.I.; Mohan, H.; Tyagi, A.K.; Mittal, J.P. Inclusion complex of γ-cyclodextrin-C-60-formation, characterization, and photophysical properties in aqueous solutions. J. Phys. Chem. 1994, 98, 4756–4759. [CrossRef] 98. Kanazawa, K.; Nakanishi, H.; Ishizuka, Y.; Nakamura, T.; Matsumoto, M. An NMR Study of the Buckminster Fullerene complex with Cyclodextrin in aqueous solution. Fuller. Sci. Technol. 1994, 2, 189–194. [CrossRef] 99. Yoshida, Z.; Takekuma, H.; Takekuma, S.-I.; Matsubara, Y. Molecular recognition of C60 with γ-Cyclodextrin. Angew. Chem. Int. Ed. Engl. 1994, 33, 1597–1599. [CrossRef] 100. Priyadarsini, K.I.; Mohan, H.; Mittal, J.P. Characterization and properties of γ-Cyclodextrin-C60 Complex in aqueous solution. Fuller. Sci. Technol. 1995, 479–493. [CrossRef] 79.

Molecules 2016, 21, 1535

34 of 35

101. Wang, H.M.; Wenz, G. Molecular solubilization of Fullerene C60 in water by γ-Cyclodextrin thioethers. Beilstein. J. Org. Chem. 2012, 8, 1644–1651. [CrossRef] [PubMed] 102. Samal, S.; Geckeler, K.E. Cyclodextrin-fullerenes: A new class of water-soluble fullerenes. J. Chem. Soc., Chem. Commun. 2000, 1101–1102. [CrossRef] 103. Yu, Y.; Shi, Z.; Zhao, Y.; Ma, Y.; Xue, M.; Ge, J. Highly Water-Soluble [60]Fullerene-ethylenediamino-betacyclodextrin inclusion complex: The synthesis and self-assembly with poly (acrylic acid). J. Supramol. Chem. 2008, 20, 295–299. [CrossRef] 104. Cabrera, K.; Wieland, G.; Schäfer, M. High-performance liquid chromatographic separation of fullerenes (C-60 and C-70) using chemically bonded gamma-cyclodextrin as stationary phase. J. Chromatogr. A 1993, 644, 396–399. [CrossRef] 105. Bianco, A.; Gasparrini, F.; Maggini, M.; Misiti, D.; Polese, A.; Prato, M.; Scorrano, G.; Toniolo, C.; Villani, C. Molecular recognition by a Silica-bound Fullerene derivative. J. Amer. Chem. Soc. 1997, 119, 7550–7554. [CrossRef] 106. Larsen, K.L. Large Cyclodextrins. J. Inclusion Phenom. Macrocycl. Chem. 2002, 43, 1–13. [CrossRef] 107. Ueda, H.; Endo, T. Large-Ring Cyclodextrins. In Cyclodextrin and Their Complexes: Chemistry, Analytical Methods, Applications; Dodziuk, H., Ed.; Wiley-VCH: Weinheim, Germany, 2008; pp. 370–380. 108. Koizumi, K.; Sanbe, K.; Kubota, Y.; Terada, Y.; Takaha, T. Isolation and Characterization of cyclic alpha-(1 ->4)-Glucans having degrees of polymerization 9–31 and their quantitative analysis by high-performance anion-exchange chromatography with pulsed amperometric detection. J. Chromatogr. A 1999, 852, 407–416. [CrossRef] 109. Bogdanski, A.; Larsen, K.L.; Bischoff, D.; Ruderisch, A.; Jung, G.; Süßmuth, R.; Schurig, V. Proceed. In Proceedings of the 12th International Cyclodextrin. Symposium, Montpellier, France, 16–19 May 2004; p. 171. 110. Bogdanski, A.; Wistuba, D.; Larsen, K.L.; Hartnagel, U.; Hirsch, A.; Schurig, V. Reciprocal principle of molecular recognition in supramolecular chromatography—Highly selective analytical separations of cyclodextrin congeners on a Silica-bonded [60]Fullerene stationary phase. New J. Chem. 2010, 34, 693–698. [CrossRef] 111. Braun, M.; Hartnagel, U.; Ravanelli, E.; Schade, B.; Böttcher, C.; Vostrowsky, O.; Hirsch, A. Amphiphilic [5:1]and [3:3]-Hexakisadducts of C-60. Eur. J. Org. Chem. 2004, 9, 1983–2001. [CrossRef] 112. Saenger, W.; Jacob, J.; Gessler, K.; Steiner, T.; Hoffmann, D.; Sanbe, H.; Koizumi, K.; Smith, S.M.; Takaha, T. Structures of the common cyclodextrins and their larger analogues-beyond the doughnut. Chem. Rev. 1998, 98, 1787–1802. [CrossRef] [PubMed] 113. Diederich, F.; Whetten, R.L.; Thilgen, C.; Ettl, R.; Chao, I.; Alvarez, M.M. Fullerene isomerism: Isolation of c2v,-c78 and d3-c78. Science 1991, 254, 1768–1770. [CrossRef] [PubMed] 114. Gutsche, C.D. Calixarenes Revisited. In Monographs in Supramolecular Chemistry; Stoddart, J.F., Ed.; The Royal Society of Chemistry: Cambridge, UK, 1998; p. 190. 115. Suzuki, T.; Nakashima, K.; Shinkai, S. Very convenient and efficient purification method for fullerene (C60) with 5, 11, 17, 23, 29, 35, 41, 47-Octa-tert-butylcalix (8) arene-49, 50, 51, 52, 53, 54, 55, 56-octol. Chem. Lett. 1994, 23, 699–702. [CrossRef] 116. Atwood, J.L.; Koutsantonis, G.A.; Raston, C.L. Purification of C60 and C70 by selective complexation with calixarenes. Nature 1994, 368, 229–231. [CrossRef] 117. Saito, Y.; Ohta, H.; Terasaki, H.; Katoh, Y.; Nagashima, H.; Jinno, H.; Itoh, K.; Trengove, R.D.; Harrowfield, J.M.; Li, S.F.Y. Separation of calixarenes with a chemically bonded C-60 silica stationary phase in microcolumn liquid chromatography. J. Sep. Sci. 1996, 19, 475–477. 118. Brindl, R.; Albert, K.; Harris, S.J.; Tröltzsch, C.; Horne, E.; Glennon, J.D. Silica-bonded Calixarenes in chromatography I. Synthesis and characterization by solid-state NMR spectroscopy. J. Chromatogr. A 1996, 731, 41–46. 119. Glennon, J.D.; Horne, E.; Hall, K.; Cocker, D.; Kuhn, A.; Harris, S.J.; McKervey, M.A. Silica-bonded calixarenes in chromatography. 2. Chromatographic retention of metal ions and amino acid ester hydrochlorides. J. Chromatogr. A 1996, 731, 47–55. [CrossRef] 120. Sanchez-Peña, M.; Zhang, Y.; Warner, I.M. Enantiomeric separations by use of calixarene electrokinetic chromatography. Anal. Chem. 1997, 69, 3239–3242. [CrossRef]

Molecules 2016, 21, 1535

35 of 35

121. Healy, L.O.; McEnery, M.M.; McCarthy, D.G.; Harris, S.J.; Glennon, J.D. Silica-Bonded Calixarenes in chromatography: Enantioseparations on molecular basket phases for rapid Chiral LC. Anal. Lett. 1998, 31, 1543–1551. [CrossRef] 122. Pfeiffer, J.; Schurig, V. Enantiomer separation of amino acid derivatives on a new polymeric chiral Calix[4]arene stationary phase by capillary gas chromatography. J. Chromatogr. A 1999, 840, 145–150. [CrossRef] 123. Ruderisch, A.; Pfeiffer, J.; Schurig, V. Synthesis of an enantiomerically pure resorcinarene with pendant L -Valine Residues and its attachment to a Polysiloxane (Chirasil-Calix). Tetrahedron Asymmetry 2001, 12, 2025–2030. [CrossRef] 124. Ruderisch, A.; Pfeiffer, J.; Schurig, V. Mixed chiral stationary phase containing modified resorcinarene and β-cyclodextrin selectors bonded to a polysiloxane for enantioselective gas chromatography. J. Chromatogr. A 2003, 994, 127–135. [CrossRef] 125. Allenmark, S. A Note concerning chromatography on mixed stationary phases. Enantiomer 1999, 4, 67–69. 126. Okamoto, Y.; Kawasaki, M.; Kawashima, M.; Hatada, K. Chromatographic resolution. 7. Useful chiral packing materials for high-performance liquid chromatographic resolution of enantiomers: Phenylcarbamates of polysaccharides coated on silica gel. J. Am. Chem. Soc. 1984, 106, 5357–5359. [CrossRef] 127. Trapp, O.; Caccamese, S.; Schmidt, C.; Böhmer, V.; Schurig, V. Enantiomerization of an inherently chiral resorcarene derivative: Determination of the interconversion barrier by computer simulation of the dynamic HPLC experiment. Tetrahedron Asymmetry 2001, 12, 1395–1398. [CrossRef] 128. Bürkle, W.; Karfunkel, H.; Schurig, V. Dynamic phenomena during enantiomer resolution by complexation gas chromatography. A kinetic study of enantiomerization. J. Chromatogr. 1984, 288, 1–14. [CrossRef] 129. Welch, C. Imprintible brush-type chiral stationary phase. J. Chromatogr. 1995, 689, 189–193. [CrossRef] 130. Reich, S.; Trapp, O.; Schurig, V. Enantioselective stopped-flow multidimensional gas chromatography—Determination of the inversion barrier of 1-Chloro-2,2-dimethylaziridine. J. Chromatogr. A 1999, 892, 487–498. [CrossRef] 131. Gebauer, S.; Friebe, S.; Scherer, G.; Gübitz, G.; Krauss, G.J. High-performance liquid chromatography on Calixarene-bonded Silica Gels—Separation of cis/trans isomers of Proline-containing peptides. J. Chromatogr. Sci. 1998, 36, 388–394. [CrossRef] 132. Schoetz, G.; Trapp, O.; Schurig, V. Determination of the cis-trans isomerization barrier of several L -Peptidyl- L -Proline dipeptides by dynamic capillary electrophoresis and computer simulation. Electrophoresis 2001, 22, 2409–2415. [CrossRef] © 2016 by the author; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

Suggest Documents