您的当前位置：首页正文

手性合成技术

来源：九壹网

Chiral Separation Techniques:A Practical Approach,Second,completely revised and updated edition

Edited by G. Subramanian

ISBNs:3-527-29875-4 (Hardcover); 3-527-60036-1 (Electronic)

3Combinatorial Approaches to Recognitionof Chirality:

Preparation and Use of Materials for theSeparation of Enantiomers

˘vec,Dirk Wulff and Jean M. J. FréchetFrantis˘ek S

3.1Introduction

The continuing trend to replace racemic drugs,agrochemicals,flavors,fragrances,

food additives,pheromones,and some other products with their single enantiomersis driven by their increased efficiency,economic incentive to avoid the waste of theinactive enantiomer,and regulatory action resulting from the awareness that indi-vidual enantiomers have different interactions with biological systems. There areseveral methods to obtain enantiomerically pure compounds. The most importantare:(i) syntheses based on chiral starting materials from natural sources such asamino acids; (ii) enantioselective reactions; and (iii) separations of mixtures of enan-tiomers using methods such as crystallization via diastereoisomers,enzymatic orchemical kinetic resolution,and chromatographic separation [1–3].

Although very efficient,the broad application of the direct preparation isrestricted due to the limited number of pure starting enantiomers. The design of amultistep process that includes asymmetric synthesis is cumbersome and the devel-opment costs may be quite high. This approach is likely best suited for the multi-tonscale production of “commodity”enantiomers such as the drugs ibuprofen,naproxen,atenolol,and albuterol. However,even the best asymmetric syntheses donot lead to products in an enantiomerically pure state (100%enantiomeric excess).Typically,the product is enriched to a certain degree with one enantiomer. Therefore,an additional purification step may be needed to achieve the required enantiopurity.The chromatographic methods that include gas and liquid chromatography,aswell as electrophoresis fit into the third group of methods designed for the separa-tion of individual enantiomers from their mixtures. These techniques,characterizedby the use of chiral stationary phases (CSP) and chiral additives,respectively,emerged more than three decades ago as valuable methods for analytical assays inacademic laboratories and clinical testing. Since chromatography can be used in vir-tually any scale,it has also been used for preparative- and production-scale separa-tions due to its high efficiency and ease of operation. For example,preparative liq-uid chromatography implemented in the simulated moving bed (SMB) formatenables the isolation of sufficient quantities of pure chiral compounds to carry outearly pharmacological and toxicological studies providing both enantiomers for

3Combinatorial Approaches to Recognition of Chirality:Preparation …

comparative biological testing [4]. Chromatographic separations continue to be usedfor the process and quality controls,even after a large-scale asymmetric technologyhad been developed and implemented.

3.2Engineering of a Chiral Separation Medium

A chiral separation medium is a complex system. Ideally,interactions that lead toenantioseparation are maximized while nonspecific interactions should be com-pletely suppressed. Typically,a medium for chromatographic separations involvesthe solid support,the selector,and the linker connecting the two,as shown in scheme3-1.

Scheme 3-1

Most commercial CSPs contain chiral selectors (vide infra) supported by poroussilica beads. Silica-based chromatographic supports have numerous advantages suchas broad range of different porosities,high mechanical stability,and resistance toswelling. Columns packed with these materials generally exhibit high efficiencies.However,residual silanol groups on the surface of the silica may contribute to non-specific interactions with the separated enantiomers,thereby decreasing the overallselectivity of the separation medium. This was demonstrated by the improved enan-tioselectivities measured for CSPs that had their residual silanol groups capped afterthe attachment of a chiral selector [5]. In contrast to silica particles,synthetic organicpolymers are more seldom used as a platform for the preparation of chiral stationaryphases. Their excellent stability over the entire range of pH,the variety of availablechemistries,and the more accurate control of both the functionality and the porousproperties make them a good alternative to the well-established silica matrices.Although porous polymer beads have been used successfully for a wide variety ofchromatographic separations including the size-exclusion chromatography ofsynthetic polymers,the normal-phase or reversed-phase separations of smallmolecules,and the ion-exchange and hydrophobic interaction chromatography ofbiopolymers,there are only a few examples of polymer-based chiral separationmedia with attached selectors [6–10]. In contrast,a larger number of chiral separa-tions has been demonstrated using polymer-based molecular imprinted separationmedia [5,11–13].

The use of a polymeric support also affords a unique opportunity to control inde-pendently the variables that may affect the chiral recognition process,which is hardto achieve with silica. For example,the type and number of reactive sites can be eas-ily adjusted with a polymer support. We recently reported an extensive study of the

3.3Chiral Selectors

effects of support chemistry,surface polarity,length and polarity of the tether,andselector loading [14]. Our group also demonstrated that separation media based onan organic polymer support provide enhanced enantioselectivities and reduced reten-tion times when compared to analogous silica-based chiral stationary phases,mostlyas a result of substantially decreased nonspecific interactions (Fig. 3-1) [8].

Fig. 3-1.Separation of racemic 3,5-dinitrobenzamido leucine N,N-diallylamide on silica and polymer-based chiral stationary phases. Conditions:column size 150 ×4.6 mm i.d.; mobile phase 20%hexanein dichloromethane; flowrate 1 mL min–1; injection 7 µg. Peaks shown are:1,3,5-tri-tert.-butylbenzene(1),R-enantiomer (2); S-enantiomer (2Ј). (Reprinted with permission from ref. [8]. Copyright 1997American Chemical Society.)

3.3Chiral Selectors

Enantioseparation is typically achieved as a result of the differences in interactionenergies ∆(∆G) between each enantiomer and a selector. This difference does notneed to be very large,a modest ∆(∆G) = 0.24 kcal/mol is sufficient to achieve a sep-aration factor αof 1.5. Another mechanism of discrimination of enantiomersinvolves the preferential inclusion of one into a “cavity”or within the helical struc-ture of a polymer. The selectivity of a selector is most often expressed in terms ofretention of both enantiomers using the separation factor αthat is defined as:

α= kЈ2/kЈ1

(1)

where kЈ1and kЈ2are the retention factors of the first and the second peak,respec-tively,defined as

kЈ= (tr– to)/to

(2)

3Combinatorial Approaches to Recognition of Chirality:Preparation …

where trand toare the retention times for the analyte and the unretained compound(void volume marker),respectively.

Chiral selectors are the most important part of the separation system. This is whymost attention during the development of new chiral separation media has alwaysbeen devoted to selectors. As a result of the growing interest in chiral chromatogra-phy,a large number of phases and additives have emerged to meet the challenge ofenantiomer separations [5,15,16]. For example,more than 90 CSPs were commer-cially available for separations in the liquid chromatographic mode in the early1990s [17].

The majority of currently known selectors can be divided into the following cat-egories:

1.Proteins.A chiral stationary phase with immobilized α1-acid glycoprotein on sil-ica beads was introduced by Hermansson in 1983 [18,19]. Several other proteinssuch as chicken egg albumin (ovalbumin),human serum albumin,and cellohy-drolase were also used later for the preparation of commercial CSPs. Their selec-tivity is believed to occur as a result of excess of dispersive forces acting on themore retained enantiomer [17]. These separation media often exhibit only modestloading capacity.

2.Modified polysaccharides.Although derivatives of microcrystalline cellulosehave been used for chiral separations since the 1970s [20],materials useful inhigh-performance liquid chromatography (HPLC) were only developed byOkamoto in the mid-1980s. CSPs involving various esters and carbamates of cel-lulose and amylose coated on wide-pore silica are currently the most frequentlyused chiral media for chromatographic separations in both analytical and prepar-ative scales [21,22]. Although they often do not exhibit very high selectivities,they separate an extremely broad range of different racemates.

3.Synthetic polymers.In the 1970s,Blaschke prepared several crosslinked gels fromN-acryloylated L-amino acids and a small percentage of ethylene dimethacrylateor divinylbenzene and used them for the low-pressure chromatographic resolutionof racemic amino acid derivatives and mandelic acid [23]. Another polymer-basedCSP was later prepared by Okamoto from isotactic poly(triphenylmethylmethacrylate). This material is the prototypical polymeric selector with a well-defined one-handed helical structure [24]. This polymer was prepared by anionicpolymerization using a chiral organolithium initiator,and then coated onto poroussilica beads. While these columns were successful in the separation of a broadvariety of racemates,their relative lack of chemical stability and high cost makethem less suitable for large-scale applications.

4.Macrocyclic glycopeptides. The first of these CSPs – based on the “cavity”of theantibiotic vancomycin bound to silica – was introduced by Armstrong [25]. Twomore polycyclic antibiotics teicoplanin and ristocetin A,were also demonstratedlater. These selectors are quite rugged and operate adequately in both normal-phase and reversed-phase chromatographic modes. However,only a limited num-ber of such selectors is available,and their cost is rather high.

5.Cyclic low molecular weight compounds. Chiral separations using chiral crownethers immobilized on silica or porous polymer resins were first reported in the

3.3Chiral Selectors

mid-1970s [6]. These highly selective and stable selectors have found only a lim-ited application for the separation of atropoisomers. In contrast,modified cyclicglucose oligomers – cyclodextrins – have proven to be very universal chiral selec-tors for chiral separations in electrophoresis,gas chromatography,and liquidchromatography [26]. In addition to the formation of reversible stereoselectiveinclusion complexes with the hydrophobic moieties of the solute molecules thatfit well into their cavity,they are often functionalized to further enhance hydro-gen bonding and dipolar interactions [27]. Attached covalently to porous silicabeads,they afford very robust CSPs with modest selectivities for a number ofracemates.

6.Metal ion complexes.These “classic”CSPs were developed independently byDavankov and Bernauer in the late 1960s. In a typical implementation,copper (II)is complexed with L-proline moieties bound to the surface of a porous polymersupport such as a Merrifield resin [28–30]. They only separate well a limited num-ber of racemates such as amino acids,amino alcohols,and hydroxyacids.

7.Small chiral molecules. These CSPs were introduced by Pirkle about two decadesago [31,32]. The original “brush”-phases included selectors that contained a chi-ral amino acid moiety carrying aromatic π-electron acceptor or π-electron donorfunctionality attached to porous silica beads. In addition to the amino acids,alarge variety of other chiral scaffolds such as 1,2-disubstituted cyclohexanes [33]and cinchona alkaloids [34] have also been used for the preparation of variousbrush CSPs.

3.3.1Design of New Chiral Selectors

CSPs with optically active polymers,such as modified cellulose,polyacrylates,andproteins,have been used successfully for a variety of enantioseparations [5,13,15,17]. Despite extended studies,the mechanism of separation for these CSPs is not yetcompletely understood,which makes it difficult to develop new media of this type.In contrast,bonded natural and synthetic chiral selectors such as substitutedcyclodextrins,crown ethers,and brush-type selectors have several advantagesincluding well-defined molecular structures and sufficiently developed enantiomer“recognition”models. For example,the separation of enantiomers with brush-typestationary phases is based on the formation of diastereoisomeric adsorbate “com-plexes”between the analyte and the selector. According to the Dalgliesh’s 3-pointmodel [35],enantiomer recognition is achieved as a result of three simultaneousattractive interactions (donor–acceptor interactions such as hydrogen bonding,π-stacking,dipole–dipole interactions,etc.) between the selector and one of the enan-tiomers being separated. At least one of these interactions must be stereochemicallydependent [36–39]. Compared to all other selectors,brush-type systems afford themost flexibility for the planned development of a variety of different chiral station-ary phases suitable for the separation of a broad range of analyte types [40,41].The majority of the original chiral selectors for brush-type CSPs were derivedfrom natural chiral compounds. Selectors prepared from amino acids,such as phenyl

3Combinatorial Approaches to Recognition of Chirality:Preparation …

glycine and leucine,[42–45] and quinine [46,47],are just a few examples of themost common chiral moieties. However,further development of highly selectiveCSPs requires the design of new types of synthetic receptors that will also make useof compounds outside the pool of natural chiral building blocks.3.4In Pursuit of High SelectivityAccording to Equation 3,the resolution Rsof two peaks in column separation is con-trolled by three major variables:retention defined in terms of the retention factor kЈ;column efficiency expressed as the number of theoretical plates N; and selectivitycharacterized by the selectivity factor α[48]:kЈNRs=(α−1)l41+klЈ (3)In this equation,kЈ1is the retention factor of the first peak. The most significant con-tribution to the overall resolution has the selectivity term (α–1) since the resolutionis a linear function of the selectivity factor. Obviously,an excellent separation canalso be achieved on columns with a high efficiency. However,the dependency of res-olution on efficiency is not linear,and levels off at high efficiencies thus making thequest for a further increase less useful. Since the technology of packed analyticalcolumns is well established and columns with very high efficiencies can be pro-duced,baseline enantioseparations are achieved even with selectors that have lowselectivity factors αclose to 1. This is why many commercial columns are very suc-cessful despite their modest selectivity factors for most racemates that typically donot exceed α= 3. In fact,very high selectivity factors are often not desirable for ana-lytical separations since the second peak would elute much later and the timerequired for the separation would be extended unnecessarily [49]. A highly desirablefeature for chiral columns is their broad selectivity,i.e. their ability to separate alarge number of various enantiomers.Most of the criteria and features outlined above for liquid chromatography mediaalso apply to the development of selectors for electrodriven separations such as elec-trophoresis and electrochromatography.Chromatographic separations in preparative columns and on preparative and pro-cess scale are based on the same concepts. However,packing large-scale columns toachieve efficiencies matching those of analytical columns remains a serious chal-lenge. Typically,preparative columns have much lower efficiencies even if they arepacked with analytical grades of stationary phases. Therefore,preparative columnshave to be much longer in order to obtain the same number of theoretical plates thatenable separations similar to those achieved in smaller columns. Unfortunately,theuse of longer columns substantially contributes to the costs of the equipment,and3.5Acceleration of the Discovery Process

their ultimate length is limited by the overall pressure drop that can be tolerated bythe system. SMB technology helps to solve both these difficulties [50].

A better solution for preparative columns is the development of separation mediawith substantially increased selectivities. This approach allows the use of shortercolumns with smaller number of theoretical plates. Ultimately,it may even lead to abatch process in which one enantiomer is adsorbed selectively by the sorbent whilethe other remains in the solution and can be removed by filtration (single plate sep-aration). Higher selectivities also allow overloading of the column. Therefore,muchlarger quantities of racemic mixtures can be separated in a single run,thus increas-ing the throughput of the separation unit. Operation under these “overload”condi-tions would not be possible on low selectivity columns without total loss of resolu-tion.

Another important issue that must be considered in the development of CSPs forpreparative separations is the solubility of enantiomers in the mobile phase. Forexample,the mixtures of hexane and polar solvents such as tetrahydrofuran,ethylacetate,and 2-propanol typically used for normal-phase HPLC may not dissolveenough compound to overload the column. Since the selectivity of chiral recognitionis strongly mobile phase-dependent,the development and optimization of the selec-tor must be carried out in such a solvent that is well suited for the analytes. In con-trast to analytical separations,separations on process scale do not require selectivityfor a broad variety of racemates,since the unit often separates only a unique mixtureof enantiomers. Therefore,a very high key-and-lock type selectivity,well known inthe recognition of biosystems,would be most advantageous for the separation of aspecific pair of enantiomers in large-scale production.

Despite continuing progress in the design of new selectors,the process is slow asit mostly involves a traditional one-selector/one-column-at-a-time methodology.

3.5Acceleration of the Discovery Process

3.5.1Reciprocal Approach

The first approach to the accelerated development of chiral selectors reported byPirkle’s group in the late 1970s relied on the “principle of reciprocity”[51]. This isbased on the concept that if a molecule of a chiral selector has different affinities forthe enantiomers of another substance,then a single enantiomer of the latter will havedifferent affinities for the enantiomers of the identical selector. In practice,a separa-tion medium is prepared first by attaching a single enantiomer of the target com-pound to a solid support that is subsequently packed into a HPLC column. Race-mates of potential selectors are screened through this column to identify those thatare best separated. The most promising candidate is then prepared in enantiopureform and attached to a support to afford a CSP for the separation of the target race-mate. This simple technique was used by several groups for the screening of various

3Combinatorial Approaches to Recognition of Chirality:Preparation …

families of compounds [52–55] The reciprocal method is particularly suitable for sit-uations in which the target enantiomer is known and its separation from a racemateis required. Since chromatographic techniques are readily automated,a broad vari-ety of novel chiral ligands may be considered.

3.5.2Combinatorial Chemistry

Although the reciprocal approach potentially enables the screening of large numbersof compounds,only the advent of combinatorial chemistry brought about the toolsrequired for the synthesis of large libraries of potential selectors in a very shortperiod of time. In addition,using the methods of combinatorial chemistry,novelstrategies different from those of the reciprocal approach could also be developed.Combinatorial chemistry and high-throughput parallel synthesis are powerfultools for the rapid preparation of large numbers of different compounds with numer-ous applications in the development of new drugs and drug candidates [56–58],metal-complexing ligands and catalysts [59–63],polymers [64],materials for elec-tronics [65,66],sensors [67],supramolecular assemblies [68–70],and peptidic li-gands for affinity chromatography [71]. The essence of combinatorial synthesis isthe ability to generate and screen or assay a large number of chemical compounds –a “library”– very quickly. Such an approach provides the diversity needed for thediscovery of lead compounds and,in addition,allows their prompt optimization. Thefundamentals of combinatorial chemistry,including rapid screening methodologies,are reviewed in numerous papers and books [72,73]. The following sections of thischapter will describe a variety of different combinatorial methods that have led toselectors for the recognition of chirality aiming mainly at the development of robustmedia for the separation of enantiomers.

3.6Library of Cyclic Oligopeptides as Additives to Back-ground Electrolyte for Chiral Capillary Electrophoresis

Enantioresolution in capillary electrophoresis (CE) is typically achieved with thehelp of chiral additives dissolved in the background electrolyte. A number of low aswell as high molecular weight compounds such as proteins,antibiotics,crownethers,and cyclodextrins have already been tested and optimized. Since the mecha-nism of retention and resolution remains ambiguous,the selection of an additive bestsuited for the specific separation relies on the one-at-a-time testing of each individ-ual compound,a tedious process at best. Obviously,the use of a mixed library of chi-ral additives combined with an efficient deconvolution strategy has the potential toaccelerate this selection.

3.6Library of Cyclic Oligopeptides as Additives to Background Electrolyte …

The power of a combinatorial approach to chiral additives for CE was firstdemonstrated by Jung and Schurig who used a library of cyclic hexapeptides [74].Since the number of hexapeptides representing all possible combinations of 20 nat-ural L-amino acids is 64 ×106,the first study involved only mixed libraries ofhexapeptides of the type c(OOXXXO) consisting of three fixed positions O andthree randomized positions X represented by any of 18 natural amino acids (cys-teine and tryptophan were not included into the scheme). Three cyclopeptidelibraries c(L-Asp-L-Phe-XXX-D-Ala),c(L-Arg-L-Lys-XXX-D-Ala),and c(L-Arg-L-Met-XXX-D-Ala),each consisting of 5832 members,were prepared andtested in chiral CE. When dissolved in an electrolyte to form 10 mmol/L solutions,all three libraries enabled the separation of racemates. For example,the first libraryfacilitated the baseline separation of racemic Tröger’s base in a 67 cm-long capillarywith a selectivity factor αof 1.01 and column efficiency of 360000 plates. Similarly,the second library helped to resolve the N-2,4-dinitrophenyl (DNP) derivative of glu-tamic acid in a capillary with the same length affording a selectivity factor of 1.13and a column efficiency of 79000 plates. These results indicate the presence of use-ful selectors in the mixed library. However,this brief study did not attempt thedeconvolution of the mixture and did not identify the best selector.

Scheme 3-2.

The deconvolution of a cyclic hexapeptide library to specify the best selector forthe target racemate has recently been reported by Chiari et al. [75]. Several librariesof linear hexapeptides with protected lateral chains were prepared using solid-phasesynthesis on Merrifield resin,and the cyclization reaction was carried out aftercleavage in solution. The study also started with a mixed library of 5832 compoundsconsisting of cyclic c(OOXXXO) hexapetides with 3 fixed ”O“ positions consistingof L-arginine,L-lysine,and β-alanine and 3 randomized positions (X) occupied byany of the 18 L-amino acids (Scheme 3-2a). Once again,cysteine and tryptophanwere not included. The substitution of D-alanine originally used by Jung and Schurig

3Combinatorial Approaches to Recognition of Chirality:Preparation …

fraction 1fraction 2fraction 3

Absorbance at 226 nm (–)80

0102030

Time (min)

4050

Fig. 3-2.Semipreparative RP-HPLC profile of cyclo(Arg-Lys-X-Pro-X-Ala). The crude sublibrary (160mol) was dissolved in 0.1% (v/v) TFA and applied to a Whatman Partisil 10 µm ODS-2 (1 ×50 cm)column. The peaks were eluted using a 40-min linear gradient of 0–25% acetonitrile in water at aflowrate of 7mL min–1. Fractions were collected every 2 min and pooled in three fractions as indicatedby arrows; 130 µmol of peptides was recovered (yield 81%). (Reprinted with permission from ref. [75].Copyright 1998,American Chemical Society.)

with β-alanine led to improved resolution of DNP-glutamic acid enantiomersachieved with the complete mixture.

Since the proline residue in peptides facilitates the cyclization,3 sublibraries eachcontaining 324 compounds were prepared with proline in each randomized position.Resolutions of 1.05 and 2.06 were observed for the CE separation of racemic DNP-glutamic acid using peptides with proline located on the first and second randomposition,while the peptide mixture with proline preceding the β-alamine residue didnot exhibit any enantioselectivity. Since the c(Arg-Lys-O-Pro-O-β–Ala) libraryafforded the best separation,the next deconvolution was aimed at defining the bestamino acid at position 3. A rigorous deconvolution process would have required thepreparation of 18 libraries with each amino acid residue at this position.

However,the use of a HPLC separation step enabled a remarkable acceleration of the deconvolution process. Instead of preparing all of the sublibraries,the c(Arg-Lys-O-Pro-O-β-Ala) library was fractionated on a semipreparative C18HPLCcolumn and three fractions as shown in Fig. 3-2 were collected and subjected toamino acid analysis. According to the analysis,the least hydrophobic fraction,whicheluted first,did not contain peptides that included valine,methionine,isoleucine,leucine,tyrosine,and phenylalanine residues and also did not exhibit any separationability for the tested racemic amino acid derivatives (Table 3-1).

Percentage of Acetonitrile (- - - -)3.6Library of Cyclic Oligopeptides as Additives to Background Electrolyte …

Table 3-1.Values of enantiomeric resolution of DNP-amino acids in a running electrolyte containingthe three fractions 1,2,and 3 of the cyclo(Arg-Lys-X-Pro-X-βAla) sublibrary separated by preparativeHPLC.Analytec(Arg-Lys-X-Pro-X-βAla)DNP-D,L-GluDNP-D,L-AlaDNP-D,L-Leu2.050.800ResolutionFraction 11.0900Fraction 24.055.543.53Fraction 31.692.450.89This led to the conclusion that these amino acids were essential for the resolutioncapability and only 6 new libraries of 18 compounds had to be synthesized withthese amino acid residues to define the position 3. Surprisingly,the separation abil-ities of all six libraries were very similar. Therefore,tyrosine was chosen for contin-uing deconvolution,since it is convenient as its aromatic ring can easily be detectedby UV spectrometry. The last step,defining position 5,required the synthesis andtesting of 6 individual hexapeptides.302520Resolution151050DNP-AlaDNP-n-ValDNP-h-SerRKYPYßARKYPXßARKXPXßARKXXXßADNP-ProDNP-GluDNP-GlnDNP-LeuFig. 3-3.Comparison of the values of enantiomeric resolution of different DNP-D,L-amino acids at dif-ferent deconvolution stages of a cyclic hexapeptide sublibrary. Resolution values in a cyclo(Arg-Lys-X-X-X-β-Ala) sublibrary,in the first line,are compared to those obtained in sublibraries with a progres-sively increasing number of defined positions. All the sublibraries were 30 mMin the running buffer whilethe completely defined cyclo(Arg-Lys-Tyr-P-Tyr-β–Ala) peptide is used at 10 mMconcentration. Condi-tions:cyclopeptide sublibrary in 20 mMsodium phosphate buffer,pH 7.0; capillary,50 µm i.d.,65 cmtotal length,57 cm to the window; V = –20 kV,I = 40; electrokinetic injection,–10 kV,3 s; detection at340 nm. (Reprinted with permission from ref. [75]. Copyright 1998,American Chemical Society.)DNP-Asp66

3Combinatorial Approaches to Recognition of Chirality:Preparation …

The improvements in resolution achieved in each deconvolution step are shown inFigure 3-3. While the initial library could only afford a modest separation of DNB-glutamic acid,the library with proline in position 4 also separated DNP derivativesof alanine and aspartic acid,and further improvement in both resolution and thenumber of separable racemates was observed for peptides with hydrophobic aminoacid residues in position 3. However,the most dramatic improvement and best selec-tivity were found for c(Arg-Lys-Tyr-Pro-Tyr-β-Ala) (Scheme 3-2a) with the tyrosineresidue at position 5 with a resolution factor as high as 28 observed for the separa-tion of DNP-glutamic acid enantiomers.

In addition to the development of the powerful chiral additive,this study alsodemonstrated that the often tedious deconvolution process can be accelerated usingHPLC separation. As a result,only 15 libraries had to be synthesized instead of 64libraries that would be required for the full-scale deconvolution. A somewhat simi-lar approach also involving HPLC fractionations has recently been demonstrated byGriffey for the deconvolution of libraries screened for biological activity [76].Although demonstrated only for CE,the cyclic hexapeptides might also be usefulselectors for the preparation of chiral stationary phases for HPLC. However,thiswould require the development of non-trivial additional chemistry to appropriatelylink the peptide to a porous solid support.

3.6.1Library of Chiral Cyclophanes

Inspired by the separation ability of cyclic selectors such as cyclodextrins and crownethers,Malouk’s group studied the synthesis of chiral cyclophanes and their interca-lation by cation exchange into a lamellar solid acid,α-zirconium phosphate aimingat the preparation of separation media based on solid inorganic-organic conjugatesfor simple single-plate batch enantioseparations [77–80].

Scheme 3-3.

3.6Library of Cyclic Oligopeptides as Additives to Background Electrolyte …

An example of the modular preparation of the cyclophane 3from the substitutedbipyridine 2and a general tripeptide 1is shown in Scheme 3-3. The host molecule3contains a pre-organized binding pocket. The overall basicity of such moleculesalso facilitates their intercalation within the lamellas of acidic zirconium phosphate,thus making this chemistry well suited for the desired application.

While the bipyridinium part of the cycle is fixed,the peptidic module is amenableto combinatorial variation. The ability of a library of cyclophanes 3containing 20dipeptides and one tripeptide to recognize chiral compounds was studied in deu-terium oxide solutions using the facile 1H NMR titration technique that requiredonly a small amount of the selector [77–79]. The chemical shifts of both protons ofthe –CH2– group linking the bipyridinium unit with the phenyl groups of the cyclo-phane were monitored as a function of concentration of the added guest molecules(Fig. 3-4) and used to calculate of binding constants K.

(a)

(b)(c)

0.0200.015

5.8

5.7ppm

5.6

R-DOPA

∆ppm0.010

0.0050.000

0.1

S-DOPA

0.20.3

Concentration (M)

Fig. 3-4. (A) Changes in chemical shift of protons of cyclophane –CH2– groups between bipyridiniumand phenyl in 1H NMR spectra of 3as a function of (R)-DOPA concentration (a) 0,(b) 0.111,and (c)0.272 mol L–1. (B) Change in chemical shift plotted against the analytical concentration of (R)- and(S)-DOPA. The solid line is calculated for 1:1 host – guest complexation. (Reprinted with permissionfrom ref. [79]. Copyright 1998,American Chemical Society.)

3Combinatorial Approaches to Recognition of Chirality:Preparation …

For example,cyclophane 3containing (L)-val-leu-ala tripeptide showed signifi-cant association with (D)-3,4-dihydroxyphenylalanine 4(DOPA) and the drugnadolol 5with Kvalues of 39 ±6 and 23 ±3 mol–1,respectively,demonstratingrather high selectivity of this cyclic selector. In contrast,only very low binding con-stants were observed for (L)-DOPA (K= 3 mol–1),Dand Ltryphtophan (K = 5 and6 mol–1),and N-2-naphthylalanine (K = 10 mol–1) [79]. A further increase in thehydrophobicity and,perhaps,enhanced π-acceptor ability of the lateral functionali-ties of the oligopeptide moiety may help substantially to improve the selectivity thatwould be required for their successful application in intercalated solids. It should benoted that this category of selectors is exceptional since,in contrast to the vastmajority of typical selectors,it operates in environment friendly aqueous media.Their solubility in water results from the cationic nature of the cyclophane.

3.6.2Modular Synthesis of a Mixed One-Bead – One-Selector Library

The screening of libraries of compounds for the desired property constitutes anessential part of the combinatorial process. The easier and the faster the screening,the higher the throughput and the more compounds can be screened in a unit of time.This paradigm has led Still’s group to develop a combinatorial approach to chiralselectors that involves a visual screening step by optical microscopy that enables themanual selection of the best candidates [81].

Scheme 3-4.

3.6Library of Cyclic Oligopeptides as Additives to Background Electrolyte …

In order to prove the concept,they prepared a library of 60 selectors using threedifferent building modules A,B,and C (Scheme 3-4). Module A consists of 15 dif-ferent D- and L-amino acids,while module C was a cyclic amide formed by the con-densation of two RRor SS1,2-diaminohexane molecules with one isophthalic acidand one trimesic acid unit. Two different stereoisomers (RRRRand SSSS) were pre-pared. These two modules A and C were linked through module B (3,4-diaminopy-rolidine) that serves as a turn element directing modules A and C toward one another.This module was also used for the attachment of the selector to the solid support.Aminomethylated Merrifield resin (100 µm polystyrene beads crosslinked with 2%divinylbenzene) modified with ε-aminocaproic acid was chosen as the support,andthe library was prepared using a split synthesis process that led to a mixture of dif-ferent beads with each bead containing only one selector [72]. Briefly,a batch ofreactive beads was split into four parts and reacted with small amounts of tag acidsto encode the stereoisomers used in the first reaction step and to enable decoding thesuccessful selectors. Each pool was then treated in a separate flask with modules(RR)-B-(RRRR)-C,(RR)-B-(SSSS)-C,(SS)-B-(RRRR)-C,and (SS)-B-(SSSS)-C,respectively. Once these reactions were completed,beads from all flasks were com-bined and then split into 15 reaction vessels. Each set contained beads with all fourcombinations of modules B–C. In the next step,the beads in each reactor werepooled,tagged and treated with 15 different activated amino acids (module A). Oncethis reaction was complete,the beads were combined again and used for the screen-ing.

To find the most efficient selectors in the library,blue and red dye-labeled enan-tiomericprobe molecules 6and 7were prepared by linking pentafluorophenyl estersof L- and D-proline with Disperse Blue 3 and Disperse Red 1,respectively,throughan isophthaloyl (shown in structures 6and 7) or a succinyl moiety. For detection,a

3Combinatorial Approaches to Recognition of Chirality:Preparation …

Fig. 3-5.Schematic of the visual screening of col-ored beads in the field of optical microscope. Opencircles red beads,close circles blue beads,dashedcircles brown beads.portion of beads was added to an excess of a mixture containing both proline – dyeconjugates and left to interact for 4h. The beads were then washed and observed byoptical microscopy. A typical picture showing the desired effect is presentedschematically in Fig. 3-5.Differently colored beads could be seen in the field of observation. Those beadsthat exhibited higher affinity to respective colored D- and L-proline – dye conjugateswere red and blue,while beads with no selectivity were brown. Beads with the purestblue or red color were manually picked and decoded to determine the structures oftheir selectors. Once decoded,gram quantities of beads bearing these selectors wereprepared and treated again with the mixture of dye-proline conjugates. Subsequentrelease of adsorbed molecules and determination of the enantiomeric excess (ee.)enabled direct comparison of the enantioselectivities of these selectors (Table 3-2).Enantioselectivities similar to those shown in Table 3-2 were also found for theracemic proline pentafluorophenyl ester,and confirm that there is no positive or neg-ative contribution to selectivity entailed by the large dye moiety.Table 3-2.Enantiodiscrimination of selected library members using the two-color assay with prolinederivatives.Library memberL-His-(SS)-B-(RRRR)-CD-His-(RR)-B-(SSSS)-CL-Asp-(SS)-B-(RRRR)-CD-Asp-(RR)-B-(SSSS)-CL-Asn-(SS)-B-(SSSS)-CD-Asn-(RR)-B-(RRRR)-CEnantiomeric excess (% ee)49 for D51 for L44 for D48 for L51 for L39 for DAlthough the preparation of the quite complex selector modules prior to the syn-thesis of the library represented a rather significant synthetic effort,this studyshowed clearly the potential of combinatorial chemistry in the early developmentstage of a chiral separation medium and demonstrated a novel approach to rapidscreening that might be amenable to full automation in the future.3.7Combinatorial Libraries of Selectors for HPLC

3.7Combinatorial Libraries of Selectors for HPLC

3.7.1On-Bead Solid-Phase Synthesis of Chiral Dipeptides

Several attempts to prepare efficient chiral stationary phases using Merrifield’ssolid-phase peptide synthesis have been reported in the past. For example,in 1977Gruska [82] prepared a tripeptide bound to a solid support using a sequence of pro-tection – coupling – deprotection reactions. This approach appeared to suffer fromincomplete conversion in the coupling steps,and the stationary phase exhibited onlya modest selectivity. A similar stationary phase was later prepared by attaching thepure tripeptide L-val-ala-pro,prepared in solution,to porous silica beads [83]. ThisCSP exhibited higher selectivity than that of Gruska thus indicating the possibledetrimental effect of undesired or uncontrolled functionalities on the recognitionability of the stationary phase. Recently,Welch analyzed these results and realizedthat the primary reason for the relative failure of the early approaches to on-beadsolid-phase synthesis of oligopeptide selectors could be traced to the relatively lowreactivity of the functional silane reagent,1-trimethoxysilyl-2-(4-chloromethyl-phenyl)ethane,used for the preparation of the original chiral stationary phases [84].As a result of steric constrains,selector surface coverage of only 0.3mmol g–1couldbe achieved after activation using this bulky silane reagent. In contrast,Welch easilyobtained CSPs containing at least twice as much selector using aminopropyltri-ethoxysilane activation. This group also optimized the reaction conditions to affordsilica beads with a high amine surface coverage and to realize their essentially quan-titative functionalization in the subsequent reaction step. This more successfulapproach enabled study of the effects of various variables on the separation proper-ties of chiral stationary phases thus prepared.

First,they compared CSPs 1 and 3 prepared by the two-step solid-phase method-ology with their commercially available counterparts (CSPs 2 and 4) obtained bydirect reaction of the preformed selector with a silica support. Although no exactdata characterizing the surface coverage density for these phases were reported,allof the CSPs separated all four racemates tested equally. These results shown in Table3-3 subsequently led to the preparation of a series of dipeptide and tripeptide CSPs5–10 using a similar synthetic approach. Although the majority of these phasesexhibited selectivities lower or similar to those of selectors built around a singleamino acid (Table 3-3),this study demonstrated that the solid-phase synthesis was a

3Combinatorial Approaches to Recognition of Chirality:Preparation …

Table 3-3.Enantioseparation of 2-methylnaphthoyl-N,N-diethylamide and naproxene methyl esterusing CSPs 1-11.CSP12a34a567891011akЈ15.676.4810.439.086.647.605.284.957.774.813.15α1.101.081.101.071.151.101.121.031.111.132.22kЈ13.193.643.162.541.812.181.455.301.871.358.06α1.331.331.001.001.001.001.001.291.001.001.84Commercial CSPs 2 and 4 from Regis Technologies,Inc. (Morton Grove,IL) have selectors close to those of CSPs1 and 3,respectively.Conditions:mobile phase 10%2-propanol in hexane,column 250 ×4.6 mm i.d.,flow rate 2mL min–1,UV detec-tion at 254nm.viable alternative to the more traditional approaches and opened the access tolibraries of chiral separation media.3.7Combinatorial Libraries of Selectors for HPLC

The experiments of the initial study were performed on a 5g scale. Although fullyfeasible,this quantity of functionalized beads required the use of very substantialamounts of reagents and solvents for both the preparation and the chromatographictesting. Therefore,the same group later developed a microscale methodology forscreening with only 50mg of the CSP [85]. Their approach assumed that if an effi-cient CSP is placed in a solution of a racemate,it should adsorb preferentially onlyone of the enantiomers,thus depleting it from the solution. To confirm this assump-tion,they placed a few milligrams of the beads in a vial and added a dilute solutionof a less than equimolar amount of the target racemate. After equilibration,a sampleof the supernatant liquid was injected into a commercial chiral HPLC column andthe peak areas were determined for both separated enantiomers to calculate theselectivity. If the areas were equal,the CSP did not exhibit any selectivity,while anychange in the amount of an enantiomer represented some level of selectivity.

After this feasibility test,a library consisting of 50 types of beads each contain-ing a different dipeptide selector attached through its C-terminal group was preparedand screened (Fig. 3-6) [84]. The first amino acid residue (aa 1) was chosen from a

Fig. 3-6. General structure of dipeptide CSPs.

group consisting of Denantiomers of phenylglycine,valine,leucine,glutamine,andphenylalanine. Both Dand Lenantiomers of the same set of amino acids were thenused as the second residue (aa 2),thus affording 50 different dipetides (5 ×10). TheN-termini of these dipeptides were then capped with the π-acidic dinitrobenzoyl(DNB) group. The selectivity of each member of this parallel library was screenedagainst the model racemate N-(2-naphthyl)alanine diethylamide (8).

A number of interesting conclusions could be drawn from the screening. Forexample,the amide hydrogen atom at the second amino acid residue close to the

3Combinatorial Approaches to Recognition of Chirality:Preparation …

DNB group appeared to be essential for high selectivity. Similarly,a large groupaffording steric hindrance at the same amino acid also improved the resolution. Incontrast,proline at that position did not lead to CSPs with good selectivities.Although the substituted amino acid-based selectors were thoroughly studied earlierusing various techniques including X-ray diffraction and NMR,this study broughtabout unexpected results. For example,the glutamine residue (gln) at the initial posi-tion was beneficial for selectivity. Similarly,homochiral (D–D) dipeptides affordedbetter selectivity than many heterochiral sequences. The best selectivity of this selec-tor library was observed for (L)-gln-(L)-val-DNB. Although successfully demon-strated,the ”in-batch“ screening is less sensitive than the direct separation in HPLCmode and its use appears limited to the discovery of selectors with selectivity factorsof at least 1.5. In addition,this evaluation allows only relative comparisons and exactnumerical values for the selectivity factors cannot be calculated easily.

To further extend this study,the authors expanded their selection of amino acidsincluding hydrogen bonding residues (L-isomers of glutamine,asparagine,serine,histidine,arginine,aspartic acid,and glutamic acid) in position 1 closest to the sur-face of the support and both Dand Lamino acids with bulky substituents (leucine,isoleucine,t-leucine,valine,phenylalanine,and tryptophan) in the position 2 [86].The terminal amine functionalities of these dipeptides were again capped by dini-trobenzyl groups. One sublibrary of 39 attached selectors could be prepared directly,while the second sublibrary involving 32 selectors required the Fmoc lateral chainprotection during its preparation. Hence,the complete library used in this studyincorporated 71 dipeptide selectors out of 98 possible structures. All of these CSPswere tested for the resolution of 8using the batch approach. Evaluation of resultsshown in Fig. 3-7 indicated that glutamic acid,aspartic acid,and histidine in posi-tion 1 and leucine,isoleucine,and phenylalanine in position 2 afforded selectorswith enantioselectivity far better than that of the gln-val-DNB selector lead identi-fied in the original library [84].

The usefulness of this solid-phase synthesis/screening was finally validated bysynthesizing 5g of beads with the (L)-glu-(L)-leu-DNB selector. These were packedinto a 250 ×4.6 mm i.d. HPLC column and evaluated using normal-phase chro-matographic conditions. The separation of racemic 8shows Fig. 3-8. This separationwas remarkable for several reasons:first,for its excellent selectivity factor (α=20.74) enabling an outstanding separation of both enantiomers with an isocratic mix-ture of 2-propanol-hexane; second,the kЈvalue for the second peak was 78.99 andthe peak did not elute until after almost 2h,indicating that a rather strong interac-tion is involved in the recognition process; and finally,the column afforded a highselectivity factor of over 18 even in pure ethyl acetate that might be a better solventfor many racemates than the hexane mixture and can easily be recycled. The large“distance”between the peaks of both enantiomers resulting from the high selectiv-ity was found extremely useful for separations under overload conditions. For exam-ple,Fig. 3-9 shows a remarkable enantioseparation of 100mg of the model racemateon the analytical size column that produced both enantiomers in optical purity of98.4 and 97%ee,respectively [86].

3.7Combinatorial Libraries of Selectors for HPLC

Enantioselectivity(R(S)) trp(S trp(R)) t-le(S tyu(R)) tyrr(S ph(R)) phee(S i-l(R)) i-leeu(S leu(R)) leuu(S) val valaa 2

(S) glu

(S) asp

(S) arg(S) his

aa 1

(S) gln

(S) ser(S) asn

Fig. 3-8. HPLC evaluation of a 250 ×4.6 mm i.d.analytical column packed with the selected dipep-tidic (S)-Glu-(S)-Leu-DNB CSP. Conditions:mobile phase 20% 2-propanol in hexane,flowrate2.0 mL min–1,UV detection at 280 nm. (Reprintedwith permission from ref. [86]. Copyright 1999,American Chemical Society.)

Fig. 3-9. Preparative HPLC of 100 mg of the test racemate 8in a single 2 mL injection using a 250 ×4.6 mm i.d. column containing (S)-Glu-(S)-Leu-DNB CSP. Conditions:mobile phase ethyl acetate,flowrate 2.0 mL min–1,UV detection at 380 nm. Injection 2 mL of 50 mg mL–1racemate solution. Frac-tions collected before and after the indicated cut point were 98.4% ee and 97% ee pure,respectively.(Reprinted with permission from ref. [86]. Copyright 1999,American Chemical Society.)

3Combinatorial Approaches to Recognition of Chirality:Preparation …

The use of silica beads as a support for the preparation of peptide libraries appearsto be somewhat problematical. While Welch could demonstrate the very satisfactoryresults described above [84–86],Li was not able to achieve complete coupling of thefirst amino acid to all of the surface functional groups of modified silica,thus leav-ing behind a number of functionalities that might be detrimental for the desired chi-ral separations [87]. Therefore,Li’s group used the organic polymer supports that aremost common in the field of solid-phase synthesis:aminomethylated Merrifieldresin (2%crosslinked polystyrene beads) and amine-terminated polyethylene glycolmodified polystyrene resin (NovaSyn TG amino resin). They prepared two 4 ×4 par-allel libraries using these supports,each consisting of 16 selectors obtained fromfour amino acids (L-leucine 9,L-alanine 10,glycine 11,and L-proline 12) and fourcarboxylic acids providing πblocks (3,5-dinitrobenzyl 13,benzyl 14,naphthyl 15,and anthryl 16) with varied aromatic moiety. The achiral glycine unit 11served as anegative internal control. The solid-phase synthesis of this 16-member library onresin was performed using a Hi-top filter plate manual synthesizer and Fmoc strat-egy.

The screening was performed in a way similar to that of Welch,except that itinvolved the use of a spectropolarimeter instead of chiral chromatography to deter-mine the selectivity. Equal amounts of the target racemate 17were added into eachof the 16 wells containing beads and the ellipticity of the supernatant liquid in eachwell was measured after equilibrating for 24h at the wavelength of the maximumadsorption (260nm). Knowing the specific ellipticity of one enantiomerically pure

3.7Combinatorial Libraries of Selectors for HPLC

analyte and the total concentration of enantiomers in the solution determined fromUV adsorption,the enantiomeric ratio of components in the supernatant liquid couldeasily be calculated. The results are summarized in Fig. 3-10. Obviously,no selec-tivity is expected for the glycine-based selectors. Thus the readings for these selec-tors set the accuracy limits of the circular dichroism method. Comparison of ellip-ticities measured with the same selector prepared on both different resins indicatedthat the more polar and hydrophilic TG resin afforded lower selectivities. However,the effect of the matrix did not change the fact that the highest ellipticity was foundfor beads with the same selectors. This suggested that the support chemistry mightaffect the screening,a fact,that was also demonstrated quite convincingly by ourresearch group [8,10].

Fig. 3-10. Ellipticities measured at 260 nm for separations achieved with the members of the parallellibrary of 16 dipeptide CSPs (Reprinted with permission from ref. [87]. Copyright 1999,AmericanChemical Society.)

The two best selectors resulting from Li’s screening,DNB-L-ala and DNB-L-leu,were then prepared on a larger scale,attached to silica beads modified with 3-amino-propyl-triethoxysilane,and the CSPs were packed into columns. Respective separa-tion factors of 4.7 and 12 were found for the separation of racemic naphthyl leucineester 17using these CSPs.

3Combinatorial Approaches to Recognition of Chirality:Preparation …

3.7.2Reciprocal Screening of Parallel Library

In our own research we have demonstrated the power of the reciprocal approach withthe screening of a library of substituted dihydropyrimidines. These compounds rep-resent a new family of potential selectors not derived from natural chiral blocks [55,88]. The library of dihydropyrimidines was prepared using a multicomponent con-densation strategy [89]. Multicomponent reactions involve two or more reactants ina single-step process that provides durable core structures and highly variable sidechains from simple starting materials. Therefore,these strategies are very useful forcombinatorial syntheses of libraries of small molecules. The Biginelli dihydropy-rimidine synthesis,first reported more than 100 years ago [90] is one of such multi-component reactions. It involves a one-pot cyclocondensation of β-keto esters 18,aldehydes 19,and ureas 20that affords dihydropyrimidine heterocycles 21(Scheme3-5) [91]. This simple approach was applied in our laboratory to create a library ofover 160 diverse racemic dihydropyrimidines [55]. Although this library was notvery large,its diversity was sufficient to obtain an efficient selector and to establisha correlation between structural features and enantioselectivity. The approach is alsonotable for its direct applicability to libraries of racemic molecules.

Scheme 3-5.

In order to perform such a correlation,our library was “screened”using a “recip-rocal”CSP with an arbitrary bound chiral target (L)-(3,5-dinitrobenzoyl) leucine(Fig. 3-11).

The target was immobilized on monodisperse macroporous poly((N-methyl)aminoethyl methacrylate-co-methyl methacrylate-co-ethylene di-

3.7Combinatorial Libraries of Selectors for HPLC

methacrylate) beads [10] affording CSP 11. Some results of the screening of thelibrary are shown in Fig. 3-12 and 3-13. They revealed that many racemic dihy-dropyrimidines are not resolved at all (separation factor α= 1.0),while rather highαvalues of up to 5.2 were achieved for top candidates such as 4-(9-phenanthryl)-dihydropyrimidine 22. Figure 3-14A shows the chromatographic separation of thisracemate.

We also performed a single-crystal X-ray structure analysis of this lead com-pound. The solid state structure of this compound depicted in Fig. 3-15 shows a half-boat-like (“sofa”) conformation with the 9-phenanthryl group in a quasi-axial orquasi-flagpole position,and the α,β-unsaturated exocyclic ester in a s-cis confor-mation. This cleft-like conformation is advantageous for the creation of centers witha high recognition ability,since one enantiomer “fits”in better than the other thusleading to selectivity.

Inspection of the large body of data collected from the separation experimentsalso revealed several structural requirements necessary for good chiral recognition.

6543210

3Combinatorial Approaches to Recognition of Chirality:Preparation …

Separation factor, α3-nitrophenyl3,5-dinitrophenyl2,4-dinitrophenyl4-cyanophenyl4-trifluoromethylR1

phenyl4-isopropylphenyl2,4-dimethylphenyl3-methoxyphenyl4-methoxyphenyl2,3-dimethoxyphenyl2,4-dimethoxyphenyl3,4-methylenedioxyphenyl3-hydroxyphenyl2,4-dichlorophenylR2

2-thiophene2-naphthyl1-naphthyl4-methoxy-1-naphthylMeH

1-pyrenyl5-nitro-1-naphthyl2-methoxy-1-naphthyl9-anthrylFig. 3-12. Selectivity factors for the separations of sublibraries of racemic ethyl (6-methyl-) and (ethyl 1,6-dimethyl) 2-oxo-4-substituted-1,2,3,4-tetrahydropyrimidine-5-carboxylates. (Reprinted withpermission from ref. [55]. Copyright 1999,American Chemical Society.)

8Separation factor, α6

R11-pyrenyl

4-methoxy-1-naphthyl9-phenanthryl1-naphthyl3,5-dinitrophenyl2,4-dimethoxyphenylphenyl

R2=HX=S

0R2=Me

R2=H

X=O

X=OR2=Me

X=S

Fig. 3-13. Selectivityfactors for the separationsof sublibraries of racemic(ethyl 6-methyl-2-oxo-),(ethyl 1,6-dimethyl-2-oxo-),(ethyl 6-methyl-2-thio-),and (ethyl 1,6-dimethyl-2-thio) 4-substi-tuted-1,2,3,4-tetrahy-dropyrimidine-5-carboxy-lates. (Reprinted withpermission from ref. [55].Copyright 1999,Ameri-can Chemical Society.)

9-phenanthryl3.7Combinatorial Libraries of Selectors for HPLC

Fig. 3-14. Separation of (A) (±)-4-(9-phenanthryl)-dihydropyrimidine 22 on chiral stationary phase CSP11and (B) racemic 3,5-dinitrobenzamidoalanine-N,N-diethylamide on chiral stationary phase CSP 12.Conditions:column 150 ×4.6 mm i.d.; mobile phase dichloromethane; flowrate 1 mL min–1.

Fig. 3-15. Spatial structure of (±)-4-(9-phenanthryl)-dihydropyrimidine 22determined by X-ray diffrac-tion. The hydrogen atoms are not shown for clarity.

For example,only those dihydropyrimidines that contained a hydrogen-bondingdonor at position 3 next to the chiral center were separated. Remarkably,dihy-dropyrimidines with non-substituted nitrogen atoms at positions 1 and 3 resulted inseparations with longer retention times and decreased separation factors α. Increas-

3Combinatorial Approaches to Recognition of Chirality:Preparation …

ing the π-basicity of the aromatic group at C4 resulted in higher separation factorsdue to a stronger interaction with the π-acidic 3,5-dinitrobenzoyl group of CSP 11used for the screening. Obviously,the substitution pattern as well as the π-basicityof the aromatic group at the C4 atom play essential roles. Dihydropyrimidines withortho-substituted aromatic groups show much higher enantioselectivities comparedto meta- and para-substituted groups. For example,the observed enantioselectivityfor the 1-naphthyl-dihydropyrimidines was almost twice that of the corresponding 2-naphthyl derivative (see Fig. 3-12). Addition of a second ortho substituent in the aro-matic ring of the dihydropyrimidines led to a dramatic deterioration of enantiosepa-ration as observed with the 9-anthryl or 2-methoxy-1-naphthyl substituted com-pounds. Finally,another important effect on enantioselectivity was observed when athiourea was used instead of urea for the preparation of 21.In general,the selectiv-ity factor for the 2-thio analogs increased two-fold compared to the corresponding2-oxo-DHPMs (see Fig. 3-13). For example,4-(9-phenanthryl)-2-thiodihydropyri-midine exhibited a selectivity of α= 11.7 instead of 5.2 for the best oxygenated ana-logue. However,the lability of the thio-containing compounds compared to the oxo-analogue led us to select the latter for the preparation of more rugged CSPs.

In the next step,the best candidate from the series 2-oxo-4-(9-phenanthryl)-dihy-dropyrimidine 22was prepared and isolated in enantiomerically pure form,thenattached to a macroporous polymer support. To attach the isolated selector to the aminofunctionalized macroporous polymethacrylate support,a suitable reactive “handle”hadto be introduced into the dihydropyrimidine. We chose to functionalize the methylgroup at the C6 carbon atom by a simple bromination to afford (-)-22.Coupling of thiscompound to the amino functionalized support then gave the desired chiral stationaryphase CSP 12(Scheme 3-6) containing 0.20mmol g–1of the selector.

Scheme 3-6.

3.7Combinatorial Libraries of Selectors for HPLC

CSP 12afforded good separations for a variety of racemic α-amino acid deriva-tives. Figure 3-14B shows an example of a typical separation of 3,5-dinitrobenz-amidoalanine-N,N-diethylamide enantiomers with a separation factor α= 7.7. Eventhough CSP 12was designed for the separation of derivatized amino acids,otherclasses of compounds such as dihydropyrimidines and profens could also beresolved. In addition,this phase could be utilized under reversed-phase separationconditions. Despite the suppression of the hydrogen bonding interactions betweenthe CSP and analyte,rather good enantioselectivities (α= 3.5) were obtained evenin a mobile phase consisting of 50%water. This confirmed that CSP 12was a ver-satile phase capable of enantioseparations in either reversed-phase or normal-phasemode.

The reciprocal strategy is best suited for typical situations encountered in theindustry that require the preparation of a highly selective CSP for the separation ofonly a single racemic product such as a drug. Since single enantiomers of that com-pound must be prepared for the testing,the preparation of a “reciprocal”packingwith a single enantiomer attached to the support does not present a serious problem.Once this CSP is available,a broad array of libraries of potential racemic selectorscan easily be screened. Although we selected a simple Biginelli three-componentcondensation reaction to prepare the library of selectors based on “non-natural”compounds,many other libraries of chiral organic compounds could also bescreened as potential selectors for the chiral recognition of specific targets. Thisapproach provides one more benefit:when such an extensive study is carried outwith structurally related families of compounds typical of chemical libraries,a bet-ter understanding of chiral recognition may quickly be generated and used for designof even more successful selectors.

3.7.3Reciprocal Screening of Mixed Libraries

The reciprocal screening of a mixed library described by Li’s group very recently[92] is an interesting variation of the approach outlined in the previous text. Follow-ing a standard procedure,a L-naphthylleucine CSP 13containing the target analytewas prepared first by attaching L-17 onto silica using a standard hydrosilylation pro-cedure.

A mixed 4 ×4 peptide library consisting of 16 members was again prepared fromthe earlier-shown two families of building blocks 9–12 (all Lenantiomers) and

3Combinatorial Approaches to Recognition of Chirality:Preparation …

13–16shown earlier,each terminated with a glycine unit,using a solid-phase syn-thesis followed by cleavage. This complete library was injected on a short columnpacked with CSP 13and eluted in a gradient of 2-propanol in hexane (Fig. 3-16a).The chromatogram features a number of peaks. Obviously the retention timesdepended on the interaction of individual library members with the immobilized tar-get as well as on possible interactions with both the support and the mobile phase.However,the retention times alone did not provide any information that might berelated to the chiral recognition since only one enantiomer of each potential selectoris present in the library. Therefore,an identical library was also prepared using all Denantiomers of the amino acid building blocks and the separation process wasrepeated on the same column (Fig. 3-16b).

Fig. 3-16. Chromatograms of mixed libraries of 16 L(a) and 16 D(b) selectors using reciprocal station-ary phase CSP 13.Conditions:column 50 ×4.6 mm i.d.; mobile phase gradient of 5–20% 2-propanolin hexane; flowrate,1.2 mL min–1; UV detection at 254 nm. (Reprinted with permission from ref. [92].Copyright 1999,American Chemical Society.)

The substantial difference between these two chromatograms was a clear proofthat CSP 13interacted differently with the mixtures of Land Denantiomers. Thisalso indicated the presence of at least one pair of enantiomers that interacted selec-tively with the CSP. Unfortunately,a tedious synthesis of 16 sublibraries (eight Landeight D) containing decreasing numbers of blocks had to be prepared to deconvolutethe best selector. A comparison of the chromatograms obtained from these subli-braries in each deconvolution step was used again,and those selectors for which nodifference was observed were eliminated. This procedure enabled the identification

3.7Combinatorial Libraries of Selectors for HPLC

of two powerful selectors,DNB-L-leu-gly and DNB-L-ala-gly in agreement with theresults of the parallel screening [87].

These two selectors terminated with a glycine were then prepared on a largerscale,their carboxyl groups reacted with 3-aminopropyltriethoxysilane,and the con-jugate immobilized onto silica. Each CSP was packed into columns and used for theseparation of racemic (1-naphthyl)leucine ester 17.Separation factors of 6.9 and 8.0were determined for the columns with DNB-ala-gly and DNB-leu-gly selectorrespectively. These were somewhat lower than those found for similar CSPs usingthe parallel synthesis and attached through a different tether [87].

The major weakness of this method appears to be the limited size of the librarythat can be screened. Although the authors believe that their method is well suited toscreen medium-sized libraries with up to a few hundred members,the most impor-tant limit is the requirement of having equal sets of both Land Dbuilding units. Thisis easily achieved with amino acids that provide relatively large diversity at a rea-sonable cost. However,this may be a serious problem with many other families ofchiral compounds. The other drawback of the current implementation is the largenumber of sublibraries that must be synthesized and screened to specify the bestselector. In fact,the number of sublibraries in the published procedure equaled thenumber of members of the original library thus making the expected accelerationeffect of this combinatorial approach questionable. However,the use of the massspectroscopic detection during the first two parallel screening separations of Fig. 3-16 would afford molecular weights of the separated compounds that are specific foreach individual selector. If the retention of any of the injected compounds is thesame for both D- and L-libraries,no selectivity occurs. In contrast,different retentiontimes in both runs indicate selectivity and even allow an estimation of selectivity fac-tors. Such an approach might totally avoid the tedious multistep deconvolution pro-cess and accelerate the screening procedures.

3.7.4Library-On-Bead

Our group also demonstrated another combinatorial approach in which a CSP carry-ing a library of enantiomerically pure potential selectors was used directly to screenfor enantioselectivity in the HPLC separation of target analytes [93,94]. The bestselector of the bound mixture for the desired separation was then identified in a fewdeconvolution steps. As a result of the “parallelism advantage”,the number of columnsthat had to be screened in this deconvolution process to identify the single most selec-tive selector CSP was much smaller than the number of actual selectors in the library.Our strategy consisted of the following steps:A mixture of potential chiral selec-tors is immobilized on a solid support and packed to afford a “complete-library col-umn”,which is tested in the resolution of targeted racemic compounds. If some sep-aration is achieved,the column should be “deconvoluted”to identify the selectorpossessing the highest selectivity. The deconvolution consisted in the stepwisepreparation of a series of “sublibrary columns”of lower diversity,each of whichconstitute a CSP with a reduced number of library members.

3Combinatorial Approaches to Recognition of Chirality:Preparation …

The feasibility of this approach was demonstrated with a model library of 36compounds prepared from a combination of three Boc protected L-amino acids(valine 23,phenylalanine 24,and proline 25) and 12 aromatic amines (3,4,5-trimethoxyaniline (26),3,5-dimethylaniline (27),3-benyloxyaniline (28),5-aminoindane (29),4-tert-butylaniline (30),4-biphenylamine (31),1-3-benyloxyani-line (28),5-aminoindane (29),4-tert-butylaniline (30),4-biphenylamine (31),1-aminonaphthalene (32),4-tritylaniline (33),2-aminoanthracene (34),2-aminofluorene (35),2-aminoanthraquinone (36),3-amino-1-phenyl-2-pyrazolin-5-one (37)). The complete library was prepared by a two-step procedure thatincludes the activation and coupling of the N-Boc-protected α-amino acids with thevarious amines followed by deprotection of the resulting protected amides (Scheme3-7). The mixture of deprotected amino acid derivatives in solution was then immo-bilized onto a polymeric solid support,typically activated 5-µm macroporouspoly(hydroxyethyl methacrylate-co-ethylene dimethacrylate) beads,to afford thechiral stationary phases with a multiplicity of selectors. Although the use of columns

3.7Combinatorial Libraries of Selectors for HPLC

Scheme 3-7.

with mixed selectors has not been recommended for actual enantioseparations [95],it is ideally suited for our combinatorial approach to optimized selectors.

As expected from the design of the experiment,the HPLC column packed withCSP 14 containing all 36 members of the library with π-basic substituents separatedπ-acid substituted amino acid amides. Although encouraging since it suggested thepresence of at least one useful selector,this result did not reveal which of the numer-ous selectors on CSP 14was the most powerful one. Therefore,a deconvolution pro-cess involving the preparation of series of beads with smaller numbers of attachedselectors was used. The approach is schematically outlined in Fig. 3-17.

Fig. 3-17. Schematic of the deconvolution process used in the library-on-bead approach.

3Combinatorial Approaches to Recognition of Chirality:Preparation …

In the next step,each single amino acid was coupled separately with the 12amines resulting in three new CSPs (CSP 15–CSP 17) each containing mixed li-gands. Once packed into HPLC columns these CSPs were evaluated. The highestselectivity factor of 13.7 in the first deconvoluted series of columns was found forthe proline-based CSP 17while the α-values of the two other columns were close to5. Further deconvolution of the proline-based column was carried out by splitting the12 members of the amine building blocks 26–37in two subgroups to afford a thirdset of columns (CSP 18and CSP 19). Thus,the first proline-based sublibrary col-umn (CSP 18) was prepared using 6 amines with smaller aromatic substituents(amines 26–31),and the second column (CSP 19) with amines having larger sub-stituents (32–37).These columns exhibited selectivity factors of 13.6 and 7.3,respectively. In the next step,the 6 amine-based groups present in the more selectivecolumn CSP 18were divided again into 2 groups (26–28 and 29–31),containing 3amines with mainly meta-substituted aromatic amines and another 3 with para-sub-stituted amines. The respective columns CSP 20and CSP 21exhibited rather highα-values of 17.4 and 14.9 for racemic (3,5-dinitrobenzoyl)leucine diallylamide andindicated that both groups involved at least one selector with a very high selectivity.This was not surprising for CSP 20,since we have previously demonstrated that theselector substituted with 3,5-dimethylaniline (27)that was intentionally used in ourexperimental design,was quite powerful [8]. Since the performance of these twocolumns was similar,we decided to further deconvolute CSP 21as it involved anentirely new set of selectors. Three columns CSP 22–CSP 24packed with beads con-taining only individual selectors were prepared with the amines 29,30,and 31,respectively. Although all these columns exhibited rather high selectivities,an α-value of 23.1 was achieved with CSP 22featuring 5-aminoindane 29 as a part of theproline selector. Figure 3-18 shows the changes in selectivity factors determined for(3,5-dinitrobenzoyl)leucine diallylamide on CSP 14–24.

Fig. 3-18. Selectivity factors αdetermined for (3,5-dinitroben-zoyl)leucine diallylamide on CSP15–24. Conditions:analyte (3,5-dinitrobenzoyl)leucine diallyl-amide; column 150 ×4.6 mm i.d.;mobile phase 20% hexane indichloromethane; flowrate 1 mLmin–1,UV detection at 254 nm.

3.7Combinatorial Libraries of Selectors for HPLC

Since this method of screening initially operated by selecting groups of moleculesrather than individual compounds,and since the difference between both CSP 20andCSP 21was small,it is indeed possible that our “best”CSP 22was not actually themost efficient selector of the original mixture. To confirm this,as well as to satisfyour curiosity to uncover which other selector was very powerful,we prepared threeadditional columns CSP 25–27containing single proline-based selectors withamines 26–28as a control experiment. As expected from the previous work [8],CSP26prepared with amine 27also exhibited a very high selectivity (α= 24.7 for (3,5-dinitrobenzoyl)leucine diallylamide) similar to that of CSP 22.Surprisingly,CSP 24and CSP 27,prepared with amines 26and 28respectively,afforded only modest α-values of less than 4.

The rapid increase in the separation factors observed for the individual series ofcolumns reflected not only the improvement in the intrinsic selectivities of the indi-vidual selectors but also the effect of increased loading with the most potent selec-tor. Although the overall loading determined from nitrogen content remained virtu-ally constant at about 0.7 mmol g–1for all CSPs,the fractional loading of each selec-tor increased as the number of selectors in the mixture decreased. Thus,the wholemethod of building block selection and sublibrary synthesis can be also viewed as anamplification process.

In the classical one-column-one-selector approach,the number of columns thathave to be tested equals the number of selectors. Using the chemistry describedabove,this would require the preparation,packing,and testing of 36 CSPs. In con-trast,our combinatorial scheme allowed us to obtain a highly selective CSP from thesame group of 36 selectors using only 11 columns (less than one-third). A simpletheoretical calculation reveals that the use of all 20 natural amino acids with 12amines would lead to a library of 240 selectors. While the preparation and testing of240 columns would be time consuming,a mixture of these selectors could be decon-voluted using our approach with only 15 columns or just 1/16 of the total number ofcolumns that would otherwise be required. The parallelism advantage of the“library-on-bead”approach with mixed selector column would be even moreimpressive with much larger libraries of selectors for which the deconvolution bysplitting the library in each step to two or three sublibraries would rapidly lead to themost selective CSP. Obviously,this approach can dramatically decrease the timerequired for the development of novel CSPs.

Although the power of this combinatorial approach was clearly demonstrated,ourmethod also has some limitations. For example,in a hypothetical situation in whichonly a single selector is active and all members of a much larger library are attachedto the beads in equal amounts,the percentage of the active selector in the mixture islow. Despite its possibly of high specific selectivity (selectivity per unit of loading),the actual selectivity of a mixed selector CSP may be rather small because of the lowloading of the specific selector. Accordingly,the peaks for both enantiomers mayelute close to each other and the actual separation may become impossible toobserve within the limits of experimental errors. Thus the sensitivity of the chro-matographic screening may somewhat limit this approach. However,the number ofselectors that can be screened in a single column remains impressive.

3Combinatorial Approaches to Recognition of Chirality:Preparation …

In contrast,there are fewer limitations from the chemical point of view. Thepreparation of large,well-defined,libraries that involve amino acid building blockshas been demonstrated many times. Carefully optimized reaction conditions for thepreparation of other mixed libraries can also ensure that each desired compound ispresent in sufficient amount. However,the reaction rates of some individual selec-tors with the activated solid support may be lower than that of others. As a result,themore reactive selectors would occupy a majority of the sites within the beads. Sincethe most reactive selectors may not be the most selective,testing of a slightly largernumber of specifically designed CSPs may be required to reduce the effect of false-negative results.

While the reciprocal approach is best suited for the development of a CSP for asingle,well-known racemic target,the library-on-bead technique is more useful forthe initial scanning of various targets to find a lead selector. It is easy to imagine adevelopment laboratory with a number of columns with immobilized libraries ofselectors used to screen target racemates in very rapid fashion. Such pre-screeningwould suggest the type of selector chemistry that may be best suited for a specifictarget. The next step would either involve deconvolution of the library on bead orreciprocal testing of parallel libraries of selectors with analogous core chemistries.

3.8Conclusion

Combinatorial chemistry,a powerful tool in many areas such as drug discovery,materials research,and catalysis,can also be used effectively in the area of molecu-lar recognition to discover new selectors for the recognition of chirality. To date,only a few combinatorial strategies leading to chiral selectors have been demon-strated in the literature. Other approaches such as combinatorial molecular imprint-ing [96–98] may soon emerge to expand the scope of combinatorial recognition pro-cesses. In the future,it is very likely that combinatorial methods will become awidely used tool,even for the development of effective selectors for specific targets.The power of combinatorial chemistry resides in both the large numbers of com-pounds that can be prepared within a very short period of time and the rapid assayand deconvolution techniques that may be used for testing to discover the optimal ornear-optimal selector within the library. This availability of libraries encompassinga broad diversity of ligand types enables rapid identification of suitable selector fam-ilies,their comparative screening,and the rapid preparation of custom-made separa-tion media for the resolution of specific racemates [99]. As an additional benefit,studies carried out with broad arrays of structurally related families of selectors canfurther improve the general understanding of chiral recognition.

Chiral separation media are quite complex systems. Therefore,neither combina-torial methods nor even the identification of the best selector can ensure that an out-standing chiral separation medium will be prepared. This is because some other vari-ables of the system such as the support,spacer,and the chemistry used for their con-

Reference

nection to the selector must also be taken into the account. It is likely,that combi-natorial methods will also soon be used for the optimization of these subunits of thesystem.

Acknowledgments

Support of this research by a grant of the National Institute of General Medical Sci-ences,National Institutes of Health (GM-44885) is gratefully acknowledged. Thiswork was also partly supported by the Division of Materials Sciences of the U.S.Department of Energy under Contract No. DE-AC03-76SF00098.

References|

[1]Seebach,D.; Sting,A. R.; Hoffmann,M. Angew. Chem.,Intl. Ed. 1996,35,2708–2747.[2]Gawley,R. E.; Aubé,J. Principles of Asymmetric Synthesis; Pergamon:New York,1996.

[3]Abdoul-Enein,H. Y.; Wainer,I. W. The Impact of Stereochemistry on Drug Development and Use;Wiley-VCH:New York,1997.

[4]Adachi,S. J. Chromatogr. A1994,658,271–282.

[5]Ahuja,S. Chiral Separations:Applications and Technology; American Chemical Society:Wash-ington D.C.,1997.

[6]Sogah,G. D. Y.; Cram,D. J. Am. Chem. Soc. 1976,98,3038.

[7]Hosoya,K.; Yoshizako,K.; Tanaka,N.; Kimata,K.; Araki,T.; Haginaka,J. Chem. Lett. 1994,1437–1438.

[8]Liu,Y.; Svec,F.; Fréchet,J. M. J. Anal. Chem. 1997,69,61–65.

[9]Sinibaldi,M.; Castellani,L.; Federici,F.; Messina,A.; Girelli,A. M.; Lentini,A.; Tesarova,E. J.Liq. Chromatogr. 1995,18,3187–3203.

[10]Lewandowski,K.; Murer,P.; Svec,F.; Fréchet,J. M. J. Anal. Chem. 1998,70,1629–1638.[11]Wulff,G. Angew. Chem.,Intl. Ed. Eng. 1995,34,1812–1832.

[12]Cormack,P. G.; Mosbach,K. React. Funct. Polym. 1999,41,115–124.

[13]Allenmark,S. Chromatographic Enantioseparations:Methods and Applications; Ellis Horwood:

New York,1991.

[14]Lewandowski,K.; Murer,P.; Svec,F.; Fréchet,J. M. J. Anal. Chem. 1998,70,1629–1638.

[15]Jinno,K. Chromatographic Separations Based on Molecular Recognition; Wiley-VCH; New York,

1997.

[16]Beesley,T. E.; Scott,R. P. Chiral Chromatography; Wiley-VCH:New York,1998.[17]Siret,L.; Bargmann-Leyder,N.; Tambute,A.; Caude,M. Analysis 1992,20,427–435.[18]Hermansson,J. J. Chromatogr. 1983,26,71.[19]Hermansson,J. Trends Anal. Chem. 1989,8,251.

[20]Hesse,G.; Hagel,R. Chromatographia 1973,6,277–280.

[21]Okamoto,Y.; Aburatani,M.; Hatada,K. Bull. Chem. Soc. Jap. 1990,63,955–957.[22]Okamoto,Y.; Yashima,E. Angew. Chem.,Intl. Ed. 1998,37,1021–1043.[23]Blaschke,G.; Schwanghart,A. D. Chem. Ber. 1976,109,1967–1975.[24]Okamoto,Y.; Nakano,T. J. Am. Chem. Soc. 1980,102,6358.

[25]Armstrong,D. W.; Tang,Y.; Chen,S.; Zhou,Y.; Bagwill,C.; Chen,J. R. Anal. Chem. 1994,66,

1473–1484.

[26]Harada,A.; Furue,M.; Nozakura,S. J. Pol. Sci. 1978,16,189.[27]Armstrong,D. W.; DeMond,W. J. Chromatogr. Sci. 1984,22,411.

[28]Rogozhin,S. V.; Davankov,V. A. Dokl. Akad. Nauk SSSR 1970,192,1288.[29]Humbel,F.; Vonderschmitt,D.; Bernauer,K. Helv. Chim. Acta 1970,53,1983.

3Combinatorial Approaches to Recognition of Chirality:Preparation …

[30]Yamskov,I. A.; Berezin,B. B.; Davankov,V. A.; Zolotarev,Y. A.; Dostavalov,I. N.; Myasoedov,N.

F. J. Chromatogr. 1981,217,539–543.

[31]Pirkle,W. H.; Finn,J. M.; Schreiner,J. L.; Hamper,B. C. J. Am. Chem. Soc. 1981,103,3964–3966.[32]Pirkle,W. H.; Hyun,M. H. J. Org. Chem. 1984,49,3043–3046.

[33]Gargaro,G.; Gasparrini,F.; Misiti,D.; Palmiery,G.; Pierini,M.; Villani,C. Chromatographia 1987,

24,505–509.

[34]Lämmerhofer,M.; Lindner,W. J. Chromatogr. 1996,741,33–48.[35]Dalgliesh,C. E. J. Chem. Soc. 1952,3940–3942.

[36]Pirkle,W. H.; Burke,J. A.; Wilson,S. R. J. Am. Chem. Soc. 1989,111,9222–9223.[37]Pirkle,W. H.; Pochapsky,T. C. J. Am. Chem. Soc. 1987,109,5975–5982.[38]Pirkle,W. H.; Pochapsky,T. C. Chem. Rev. 1989,89,347–362.[39]Davankov,V. A. Chromatographia1989,27,475–482.

[40]Pirkle,W. H.; Bowen,W. E. HRC-J. High Resolut. Chromatogr. 1994,17,629–633.[41]Pirkle,W. H.; Murray,P. G.; Burke,J. A. J. Chromatogr. 1993,641,21–29.[42]Pirkle,W. H.; Welch,C. J.; Lamm,B. J. Org. Chem. 1992,57,3854–3860.

[43]Gasparrini,F.; Misiti,D.; Villani,C.; Borchardt,A.; Burger,M. T.; Still,W. C. J. Org. Chem. 1995,

60,4314–4315.

[44]Cuntze,J.; Diederich,F. Helv. Chim. Acta 1997,80,897–911.

[45]Zimmerman,S. C.; Kwan,W. S. Angew. Chem. Intl. Ed. Eng. 1995,34,2404–2406.[46]Lämmerhofer,M.; Lindner,W. J. Chromatogr. A 1998,829,115–125.[47]Lämmerhofer,M.; Lindner,W. J. Chromatogr. 1999,839,167–182.

[48]Snyder,L. R.; Kirkland,J. J. Introduction to Modern Liquid Chromatography; Wiley:New York

1979.

[49]Pirkle,W. H.; Pochapsky,S. J. Chromatogr. 1986,369,175–177.

[50]Strube,J.; Jupke,A.; Epping,A.; Smidt-Traub,H.; Schulte,M.; Devant,R. Chirality 1999,11,

440–450.

[51]Pirkle,W. H.; House,D. W.; Finn,J. M. J. Chromatogr.1980,192,143–158.[52]Pirkle,W. H.; Burke,J. A. J. Chromatogr. 1991,557,173–185.[53]Pirkle,W. H.; Koscho,M. E. J. Chromatogr. 1999,840,151–158.[54]Welch,C. J. J. Chromatogr. A 1994,666,3–26.

[55]Lewandowski,K.; Murer,P.; Svec,F.; Fréchet,J. M. J. J. Comb. Chem. 1999,1,105–112.

[56]Terrett,N. K.; Gardner,M.; Gordon,D. W.; Kobylecki,R. J.; Steele,J. Chemistry-A European Jour-nal 1997,3,1917–1920.

[57]Thompson,L. A.; Ellman,J. A. Chem. Revs. 1996,96,555–600.

[58]Gordon,E. M.; Gallop,M. A.; Patel,D. V. Acc. Chem. Res. 1996,29,144–154.[59]Menger,F. M.; Eliseev,A. V.; Migulin,V. A. J. Org. Chem. 1995,60,6666–6667.

[60]Francis,M. B.; Finney,N. S.; Jacobsen,E. N. J. Am. Chem. Soc. 1996,118,8983–8984.[61]Gilbertson,S. R.; Wang,X. Tetrahedr. Lett. 1996,37,6475–6478.

[62]Cole,B. M.; Shimizu,K. D.; Krueger,C. A.; Harrity,J. A.; Snapper,M. L.; Hoveyda,A. H. Angew.

Chem.,Intl. Ed. Eng. 1996,35,1668–1671.

[63]Cole,B. M.; Shimizu,K. D.; Krueger,C. A.; Harrity,J. A.; Snapper,M. L.; Hoveyda,A. H. Angew.

Chem. Intl. Ed. Eng. 1996,35,1668–1671.

[64]Brocchini,S.; James,K.; Tangpasuthadol,V.; Kohn,J. J. Am. Chem. Soc. 1997,119,4553–4554.[65]Schultz,P. G.; Xiang,X. D. Current Opinion in Solid State & Materials Science 1998,3,153–158.[66]Xiang,X. D.; Sun,X. D.; Briceno,G.; Lou,Y.; Wang,K. A.; Chang,H.; Wallace-Freedman,W. G.;

Chen,S. W.; Schultz,P. G. Science 1995,268,1738–1740.

[67]Dickinson,T. A.; Walt,D. R.; White,J.; Kauer,J. S. Anal. Chem. 1997,69,3413–3418.[68]Burger,M. T.; Still,W. C. J. Org. Chem. 1995,60,7382–7383.

[69]Goodman,M. S.; Jubian,V.; Linton,B.; Hamilton,A. D. J. Am. Chem. Soc. 1995,117,

11610–11611.

[70]Carrasco,M. R.; Still,W. C. Chemistry & Biology 1995,2,205–212.[71]Huang,P. Y.; Carbonell,R. G. Biotechnol. Bioeng. 1995,47,288–297.

[72]Wilson,S. R.; Czarnik,A. W. Combinatorial Chemistry:Synthesis and Application;Wiley:New

York,1997.

[73]Terrett,N. K. Combinatorial Chemistry; Oxford University Press:Oxford,1998.

[74]Jung,G.; Hofstetter,H.; Feiertag,S.; Stoll,D.; Hofstetter,O.; Wiesmuller,K. H.; Schurig,V.

Angew.Chem. Intl. Ed. 1996,35,2148–2150.

Reference

[75]Chiari,M.; Desperati,V.; Manera,E.; Longhi,R. Anal. Chem. 1998,70,4967–4973.

[76]Griffey,R. H.; An,H. Y.; Cummins,L. L.; Gaus,H. J.; Haly,B.; Herrmann,R.; Cook,P. D. Tetra-hedron 1998,54,4067–4076.

[77]Garcia,M. E.; Gavin,P. F.; Deng,N. L.; Andrievski,A. A.; Mallouk,T. E. Tetrahedr. Lett. 1996,

37,8313–8316.

[78]Gavin,J. A.,Deng,N. L.; Alcala,M.; Mallouk,T. E. Chem. Mater. 1998,10,1937–1944.[79]Gavin,J. A.; Garcia,M. E.; Benesi,A. J.; Mallouk,T. E. J. Org. Chem. 1998,63,7663–7669.[80]Mallouk,T. E.; Gavin,J. A. Acc. Chem. Res. 1998,31,209–217.

[81]Weingarten,M. D.; Sekanina,K.; Still,W. C. J. Am. Chem. Soc. 1998,120,9112–9113.[82]Kitka,E. J.; Gruska,E. J. Chromatogr. 1977,135,367–376.

[83]Howard,W. A.; Hsu,T. B.; Rogers,L. B.; Nelson,D. A. Anal. Chem. 1985,57,606–610.[84]Welch,C. J.; Bhat,G.; Protopopova,M. N. Enantiomer1998,3,463–469.[85]Welch,C. J.; Protopopova,M. N.; Bhat,G. Enantiomer 1998,3,471–476.[86]Welch,C. J.; Bhat,G.; Protopopova,M. N. J. Comb. Chem. 1999,1,364–367.[87]Wang,Y.; Li,T. Anal. Chem. 1999,71,4178–4182.

[88]Lewandowski,K.; Murer,P.; Svec,F.; Fréchet,J. M.J.Chem. Commun. 1998,2237–2238.

[89]Armstrong,R. W.,Combs,A. P.; Tempest,P. A.; Brown,S. D.; Keating,T. A. Acc. Chem. Res. 1996,

29,123–131.

[90]Biginelli,P. Gazz. Chim. Ital. 1893,23,360–416.[91]Kappe,C. O. Tetrahedron 1993,49,6937–6963.

[92]Wu,Y. Q.; Wang,Y.; Yang,A. L.; Li,T. Anal. Chem. 1999,71,1688–1691.

[93]Murer,P.; Lewandowski,K.; Svec,F.; Fréchet,J. M. J. Anal. Chem. 1999,71,1278–1284.[94]Murer,P.; Lewandowski,K.; Svec,F.; Fréchet,J. M. J.Chem. Commun. 1998,2559–2560.[95]Pirkle,W. H.; Welch,C. J. J. Chromatogr. A 1996,731,322–326.[96]Lanza,F.; Sellergren,B. Anal. Chem. 1999,71,2092–2096.

[97]Takeuchi,T.; Fukuma,D.; Matsui,J. Anal. Chem. 1999,71,285–290.

[98]Sabourin,L.; Ansell,R. J.; Mosbach,K.; Nicholls,I. A. Anal. Commun. 1998,35,285–287.[99]Stinson,S. C. Chem. Eng. News 1999,October 11,p.110.

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文