Abstract
Molecular imprinting technology offers the unique opportunity to tailor chiral stationary phases with predefined chiral recognition properties by employing the enantiomers of interest as binding-site-forming templates. Added advantages, such as ease of preparation, chemical robustness, low-cost production, and the possibility of shaping molecularly imprinted polymers (MIPs) in various self-supporting formats, render them attractive materials for a broad range of chiral recognition applications. In this review a critical overview on recent developments in the field of MIP-based chiral recognition applications is given, focusing on separation techniques and molecular sensing. Inherent limitations associated with the use of enantioselective MIP materials in high-performance separation techniques are outlined, including binding site heterogeneity and slow mass transfer characteristics. The prospects of MIP materials as versatile recognition elements for the design of enantioselective sensor systems are highlighted.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
Introduction
Molecularly imprinted polymers (MIPs) are increasingly appreciated as highly target-specific polymeric molecular recognition materials with a broad range of potential applications in separation sciences, catalysis, molecular sensing, and drug delivery [1]. One of the most intriguing aspects of molecular imprinting technology is the unique possibility of obtaining MIPs with predetermined enantioselective binding properties for given chiral targets [2, 3]. Compared with conventional chiral selector systems, chirally imprinted polymers have several other advantages, for example ease of preparation, scalability, low material costs, and flexibility to design various self-supporting formats. MIPs have also been shown to be more stable toward thermal and mechanical stress and to tolerate a broad range of solvents, acids, bases, and salts, making them particularly well-suited to operation in challenging environments [4]. MIP-type affinity materials have received much scientific attention in the context of a variety of chiral recognition applications, including enantioselective chromatography [5, 6], capillary electrochromatography [7–11], solid-phase extraction [12–14], binding assays [15], sensing [16–18], catalysis [19–22], and drug delivery [23–25]. Despite the emergence of competing technologies over the two last decades, MIP-mediated chiral recognition phenomena continue to attract interest from the scientific community. An ever-increasing body of data accumulated in numerous studies, however, has provided a basis for clearer understanding of true potential and inherent limitations of MIP materials in chiral recognition applications.
The intent of this review is to provide a critical overview of the current role of MIP-type affinity materials in the multidisciplinary field of chiral recognition, with the focus on applications in enantioselective separation and molecular sensing. The coverage of this review is selective rather than exhaustive, and concentrates on innovative concepts rather than incremental improvements. We offer our sincere apologies to the authors whose work could not be discussed herein.
Molecular imprinting approaches
Molecular imprinting can generally be defined as a process during the course of which polymeric matrices with specific binding sites are generated by template-induced pre-arrangement of complementary interactive functional groups [26]. Depending on the chemical nature of the interactions between the templating molecule and the interactive functional groups, molecular imprinting technology may be categorized into covalent and noncovalent approaches.
The covalent imprinting approach pioneered by Wulff [27, 28] utilizes template species with covalently attached polymerizable functionality (Fig. 1). On co-polymerization with cross-linking agents in porogenic solvents, these template–monomer complexes become integrated into a macroporous network. Template removal is subsequently accomplished by chemical cleavage of the supporting covalent bonds, liberating the corresponding functional groups located within polymer-embedded cavities. These cavities are in size, shape, and functional-group arrangement complementary to the templating molecule, and therefore can act as highly template-specific binding sites. Covalently imprinted polymers are characterized by thermodynamically rather homogeneous binding characteristics, reduced non-specific adsorption, and high yields of stable and functional template binding sites [19, 28]. The template binding of noncovalently imprinted polymers involves the re-formation of the cleaved covalent bonds within the cavities. Because of the high steric requirements imposed by the well-defined binding site geometry the kinetics of the re-binding process may be rather slow. Successful covalent imprinting requires the bonds of the functional groups to the template to be cleavable under relatively mild conditions. This requirement restricts the applicability of the covalent imprinting approach to templates with functional groups that can be converted to readily cleavable derivatives, for example boronic esters, imines, ketals and esters.
Covalent imprinting of 4-nitrophenyl-α-d-mannopyranoside-2,3:4,6-di-O-(4-vinylphenylboronate). Adapted from Ref. [28]
The noncovalent imprinting approach developed by Mosbach [29] circumvents the synthetic efforts associated with covalent imprinting by exploiting weak intermolecular interactions for template-directed pre-arrangement of complementary functional groups. In practice this is achieved by equilibrating the template molecules with an excess of suitable functional monomers in porogenic solvents supportive to intermolecular interactions to form noncovalent template–functional monomer complexes (Fig. 2).
Noncovalent imprinting of a dipeptide derivative by multiple noncovalent interactions with methacrylic acid. Adapted from Ref. [35]
Subsequent polymerization in the presence of crosslinking agents traps some of these pre-polymerization complexes within the growing polymer matrix. Because of the reversible nature of the matrix–template interactions, template removal from the imprinted sites can be accomplished conveniently by extraction with competing solvents. Noncovalent imprinting obviates the need for synthetic manipulations and chemically addressable template structures, thus offering the advantages of operational simplicity and an extended scope of applications. Moreover, taking advantage of the huge repertoire of commercially available chemically diverse monomers, suitable noncovalent interaction motifs can be designed and optimized almost for any template structure of interest. Nevertheless, the noncovalent imprinting concept also suffers from several inherent drawbacks [19, 28]. The excess of functional monomers needed to achieve appreciable levels of template–functional monomer complexation leads to random incorporation of interactive functional groups outside the imprinted cavities, giving rise to nonspecific interactions. As a reflection of the dynamic nature of the reversible pre-polymerization complexes noncovalent imprinted polymers suffer from severe receptor site heterogeneity, imparting unfavorable polyclonal binding characteristics. Another limitation is the relatively poor yield of high-affinity binding sites achievable with noncovalent imprinting approaches. Typically, less than 15% of the loaded template produces functional binding sites, indicating that most of the imprinted cavities are irreversibly lost, by shrinking, on template elution.
Efforts to combine the attractive features of covalent and noncovalent imprinting procedures have led to the development of hybrid technologies, known as the semi-covalent and sacrificial-spacer approaches [26]. These strategies employ covalent template–monomer complexes in the imprinting step but entirely noncovalent interactions for binding. An illustrative example of this appealing concept is the preparation of a cholesterol-imprinted polymer using a carbonate derivative as sacrificial linker motif (Fig. 3 [30]).
Imprinting of cholesterol by the sacrificial spacer (semi-covalent) method. Adapted from Ref. [30]
On copolymerization of the cholesteroyl (4-vinyl)phenyl carbonate adduct with a crosslinking agent the supporting carbonate was cleaved by base-promoted hydrolysis, resulting in a phenol functionality strategically positioned within the imprinted binding sites. Assessment of the molecular recognition properties of the resulting MIP by batch binding revealed thermodynamically homogeneous association behavior with a single binding constant for cholesterol. The acylation of the phenol functionality led to the suppression of cholesterol binding, confirming the crucial role of this functionality in hydrogen bond-mediated template recognition.
Liquid chromatographic applications of MIP-type CSPs
Potential benefits
Analytical and preparative-scale enantiomer separations are routine requirements in many fields of contemporary research, including drug development, the food and agrochemical industries, catalyst technology and enzyme engineering, and the material sciences. A large number of efficient chiral stationary phases (CSPs) for liquid chromatographic applications have been developed to address these needs. Most of the chiral recognition elements incorporated into these CSPs are non-target-specific in nature, however, making reliable prediction of the separability and order of elution of a given pair of enantiomers elusive. Molecular imprinting technology offers the unique opportunity to tailor CSPs with predefined chiral recognition properties by using the enantiomers of interest as binding-site-forming templates. The operational simplicity of this chirality transfer from a templating enantiomer to the polymer network also obviates the need for sophisticated receptor designs, lengthy synthetic routes, and elaborate immobilization procedures.
Consequently, molecular-imprinting technology has been extensively exploited to produce target-specific CSPs for a broad range of chiral compounds [6, 31, 32], for example amino acidic derivatives [33, 34], peptides [35], natural compounds, and a variety of drugs [32]. In general, MIP-type CSPs have excellent chiral recognition properties for the templating chiral species, which are manifested in high enantioselectivity, pronounced substrate-specificity, and predictable order of elution, with the enantiomers employed as templates being the more strongly retained species. A particularly attractive feature of MIP-type CSPs is their capability of discriminating not only between enantiomers but also between structurally closely related stereoisomers. For example, a poly(MAA-co-EDMA) CSP imprinted against N-Ac-l-Phe-l-Trp-OMe could successfully distinguish the template from the corresponding d,d, d,l, and l,d isomers with impressive selectivity factors (α = 17.8, 14.2, and 5.21) [35].
Challenges: engineering of suitable chromatographic MIP formats
The broad chromatographic use of MIP-type CSPs has been hampered by the difficulties associated with the engineering of suitable chromatographic formats and the inherently poor mass-transfer characteristics of imprinted polymers [6]. Most chromatographic chiral recognition studies on MIP-type CSPs have been performed with columns packed with particles derived from bulk polymers by the traditional grinding and sieving procedure. The mechanical processing of bulk polymers, however, inevitably leads to irregular particles with relatively broad size distributions, resulting in packings of irreproducible quality which suffer from poor column efficiency and high back pressure.
Much research effort has been invested in establishing dedicated MIP formats for chromatographic applications, for example porous monoliths [36, 37], spherical beads [38–44] and silica-supported films [45, 46].
The generation of monolithic MIP-type CSPs by in-situ polymerization of pre-polymerization mixtures within the confines of chromatographic columns was pioneered by Matsui and coworkers [36]. Crucial to the success of this approach is the use of dedicated porogenic solvents, supporting the formation of a continuous pore structure. In this regard, cyclohexanol–dodecanol mixtures were found to be particularly effective. Apart from the advantage of being less elaborate to prepare, monolithic MIP-type CSPs have been shown to have chiral recognition properties superior to those of traditionally prepared bulk polymers. Thus, in a recent study Yin et al. [37] compared the chiral recognition performances of a monolithic and a bulk MIP-type CSP imprinted with l-nateglinide, both prepared with identical pre-polymerization mixtures and similar polymerization conditions. Successful enantiomer separation could be achieved on the monolithic CSP whereas a column packed with the bulk MIP particles proved unselective under identical chromatographic conditions. The origin of the superior chiral recognition characteristics of imprinted monolithic compared with bulk polymers has been investigated by Guiochon’s group for Fmoc-l-Trp imprinted MIPs [47]. Detailed analysis of the adsorption isotherms revealed that binding-site distribution was more homogeneous in the monolithic column, mainly because the number of nonspecific affinity sites was reduced. A limiting factor in the preparation of monolithic MIP formats, however, is the relatively polar nature of the porogenic solvents required to achieve pore structures of sufficient permeability. These solvents may compete with crucial template-functional monomer interactions and thus compromise the quality of the imprinted receptor sites.
In the context of chromatographic applications, major research efforts have been focused on establishing efficient procedures for preparation of spherical MIP particles [26]. Established strategies to create polymer beads typically involve suspension polymerization techniques performed in biphasic solvent systems, using an aqueous phase for suspending the monomeric species. The presence of water, however, is usually regarded as incompatible with noncovalent molecular imprinting techniques, because its polar protic nature readily disrupts crucial template–functional monomer interactions.
Nevertheless, MIPs capitalizing on relatively strong intermolecular noncovalent interactions, for example charge-supported hydrogen bonds, have been shown to be compatible with the competing aqueous environment required for suspension–polymerization procedures. Specifically, the two-stage swelling–polymerization procedure developed by Hosoya and Haginaka [42] has been successfully used to prepare enantioselective, uniformly sized spherical MIP particles for a range of pharmaceutically relevant chiral compounds, for example profens [41, 42, 48], calcium-channel blockers [44], and antihistamines [43]. This rather sophisticated procedure uses polystyrene latex as seed particles, which, in a first swelling step, are incubated with emulgators, cross-linking agents, and radical initiators, followed by a second swelling step in the presence of the components forming the template–functional monomer complexes. During these swelling steps the seed particles internalize the organic components and increase in size. On thermal polymerization the particles are transformed into uniformly sized MIP spheres with an size range of 5 to 10 μm. These MIP particles have favorable chromatographic properties, for example improved flow characteristics and excellent pressure resistance, enabling long-term use at enhanced flow rates and temperatures [44].
Mosbach’s group has developed an appealing water-free alternative to conventional suspension polymerization procedures for preparation of MIPs in bead format [39, 49]. In this procedure the potentially interfering water suspension phase is replaced by perfluoroalkane-type solvents. Perfluoroalkanes are essentially insoluble in most organic solvents but are of low polarity and are thus capable of stabilizing the noncovalent template-monomer interactions crucial to successful molecular imprinting. Specific challenges associated with this procedure are the synthesis of dedicated fluorous emulgators for stabilizing the particles formed in the organic-fluorous interphase, the high costs of perfluoroalkanes, and the often substantial optimization efforts. Under optimized conditions, however, this suspension polymerization procedure produces spherical MIP particles with a relatively narrow size distribution (8–25 μm) and favorable mechanical properties. For example, columns packed with poly(MAA-co-TRIM) MIP beads imprinted with Boc-l-Phe could be operated at flow rates up to 5 mL min−1 without suffering significant loss in enantioselectivity [49].
An experimentally greatly simplified suspension polymerization technique for production of MIPs in bead format has recently been reported by Kempe [38, 40]. This method utilizes inexpensive mineral oil as a highly effective dispersion medium. In this environment, pre-polymerization mixtures prepared in polar porogens, for example MeCN and MeOH, were found to form stable emulsions after mechanical dispersion. Exposure of these suspensions to UV-irradiation at low temperature produced fairly monodisperse spherical MIP particles in the size range 10–20 μm (Fig. 4).
Scanning electron micrographs of (S)-propranolol-imprinted MIP beads obtained by dispersion polymerization in mineral oil (top) and MIP particles obtained by conventional bulk polymerization after grinding and sieving (bottom). Reprinted with permission from Ref. [38]
By using this operationally straightforward procedure, the authors successfully created (S)-propranolol-imprinted spherical MIP beads with a binding capacity similar to that of an established MIP material prepared by the traditional bulk polymerization approach. Particularly attractive features of this procedure are simplicity, avoidance of stabilizing agents, low costs, scalability, and reduced time requirements, enabling fully functional MIP beads to be obtained within 4 h. The applicability of this procedure, however, is restricted to polar porogenic solvents, rendering imprinting formulations insoluble in the mineral oil dispersion phase.
A major limitation encountered in the development of new MIP formats for chromatographic applications is the difficulty of controlling polymer morphology (pore volume, pore size, and structure) independently of molecular recognition issues. To decouple the experimental conditions defining the molecular recognition properties of MIPs from pore-forming factors, Sellergren’s group explored strategies to graft MIP films on to the surface of morphologically well-defined silica supports [9, 10, 45, 46, 50, 51]. A key aspect of these approaches is the use of silica materials carrying surface-immobilized free radical azo-initiator species to favor surface-located “grafting-from” polymerization over solution-based polymer growth. These investigations were performed using a well-established MIP model system employing l-phenylalanine anilide (PA) as a template, and MAA and EDMA as functional monomer and crosslinking agent, respectively. In the framework of a comprehensive optimization study [45], the effects of different experimental conditions during the grafting procedure on the chromatographic performance of the resulting surface-modified MIP-silica composites were evaluated. Silica particles with a pore diameter of 10 nm supporting a thin (0.8 nm) MIP film gave the most favorable chromatographic performance with regard to efficiency. The sample-load capacity, however, increased with increasing thickness of the MIP layer, reaching an optimum at 7.0 nm.
Because of the single-point attachment of the grafted azo-initiator, however, solution-phase polymerization leading to undesired particle aggregation could not be efficiently suppressed. In an effort to address this limitation the authors resorted to dithiobenzoate-type iniferters to moderate the inherently high activity of the surface-anchored azoinitiator species [46]. The presence of the iniferters in the reaction mixture favored reversible addition–fragmentation chain transfer (RAFT) polymerization mechanisms, with the benefit of a slower and more readily controllable surface-located polymer growth. Under these special RAFT conditions aggregation phenomena triggered by solution-phase polymerization could be completely suppressed, even over extended reaction times (Fig. 5). The composite MIP particles obtained by use of this imprinting process had highly homogenous MIP films with enhanced mass-transfer characteristics compared with conventional imprinted polymers.
Scanning electron micrographs of surface-grafted silica-MIP composite materials after a reaction time of 120 min (a, c, d) or 240 min (b). (a, b) Composites prepared using azoinitiator-grafted silica particles (pore size 10 nm, azoinitiator surface concentration 1.5 μmol m−2) as support in the presence of RAFT agent. (c, d) Composites prepared using (c) azoinitiator-grafted silica particles (pore size 10 nm, azoinitiator surface concentration 0.33 μmol m−2) or (d) azoinitiator-grafted silica particles (pore size 10 nm, azoinitiator surface concentration 1.5 μmol m−2) in the absence of RAFT agent. The bar represents a distance of 10 μm (a, c, d) or 20 μm (b). Reprinted with permission from Ref. [46]
Drawbacks: inherent thermodynamic limitations
Despite the major advances achieved in the preparation of chromatographically suitable polymer formats, enantioselective MIPs do not compete with established non-target-specific CSPs when challenged with real-world analytical and preparative enantiomer separation problems. Under conventional isocratic mobile phase conditions, MIP-type CSPs suffer from poor chromatographic efficiency and peak tailing, which is especially severe for the more retained (imprinted) enantiomer. Even minor increases in sample loading on MIP-type CSPs cause major losses in enantioselectivity, and complete loss of enantioselectivity may be observed with amounts of sample with which commercial CSP still operate under linear chromatographic conditions.
An in-depth investigation of the underlying thermodynamic and kinetic factors has been performed by Guiochon’s group [52–60]. In a series of studies the enantiomer separation characteristics of MIPs were systematically analyzed using an Fmoc-l-Trp-imprinted polymer as a model CSP. The dynamic adsorption behavior of the imprinted and corresponding non-imprinted polymers were measured for the template and its enantiomer over a large range of concentrations by frontal analysis. The resulting isotherm raw data were used as input to calculate the respective adsorption affinity distributions, which provided valuable information about the relative population of binding sites and their energies. The adsorption affinity distributions obtained for the MIP were indicative of the existence of three types of interaction site with distinct binding energies for both Fmoc-Trp enantiomers. In contrast, only two types of low energy site were found for the corresponding non-imprinted polymer (NIP). Fitting the isotherm data to tri-Langmuir adsorption models enabled quantification of the association energies and densities of the individual binding sites of the MIP. The least populated sites with the highest binding energy were identified as enantioselective sites [60]. This finding is consistent with the poor loading capacity of MIP-type CSPs and explains the dramatic drop in enantioselectivity of MIP-type CSPs even at moderate levels of sample loading. The authors demonstrated that minor variations in the operation conditions of MIP-type CSPs can induce significant shifts in the relative populations and association energy of the MIP binding sites, and thus dramatically alter their chromatographic behavior [55–57]. In an effort to elucidate the factors causing the pronounced peak tailing of MIP-type CSPs, a detailed analysis of the mass-transfer characteristics was conducted. Individual kinetic contributions to the intraparticle diffusion in MIPs were deconvoluted by fitting simulated band profiles to experimentally observed peak shapes [54, 61]. The overall mass transfer kinetics of the MIP-type CSP were found to be dominated by surface diffusion rather than by pore diffusion or external mass-transfer phenomena. In contrast to the NIP, the surface-diffusion coefficient of the template on the MIP-type CSP was directly dependent on loading, with increasing sample loading enhancing the surface diffusion-mediated mass transfer. This behavior was rationalized as a consequence of template binding occurring simultaneously to thermodynamically and kinetically heterogeneous binding sites on the MIP surface. The authors concluded that only at very low concentrations will the template preferentially bind to the enantioselective low-abundance high-affinity sites. Because of multi-point interactions experienced in this environment, template dissociation from these sites, and thus surface diffusion, should be slow. In the presence of higher template concentrations, however, the sparsely populated high-affinity sites become rapidly saturated; further template binding will then occur at the abundant nonselective low-energy sites, the dissociation kinetics of which are much faster, because of the lack of specific interactions. It is this mixed regime of dissociation kinetics that accounts for the observed concentration dependence of the surface diffusion. The heterogeneous mass-transfer kinetics of MIP-type CSPs become chromatographically evident in the extreme asymmetry of the template peak observed with overloading conditions. The authors concluded that, because of the low saturation capacities of the high-affinity sites of MIPs, significant peak distortion may occur for the imprinted enantiomer already at very low sample loads, possibly at concentrations lower than the detection limit.
Different strategies have been suggested to improve the notoriously poor efficiency of enantiomer separations on MIP-type CSPs, including the use of elevated column temperatures [44], gradient elution [34] and the application of supercritical-fluid mobile-phase conditions [62].
For example, Fu et al. demonstrated that the poor efficiency of an (S)-nilvadipine-imprinted MIP CSP can be significantly enhanced by increasing the column temperature from 25 to 70 °C [44]. Thus, the initial plate counts for the (R) and (S) enantiomers of 1,260 and 19,000 could be raised to 3,240 and 27,240 plates m−1, respectively, with the added benefit that the total analysis time was reduced by a factor of three. Under these optimized operation conditions the chromatographic enantiomer separation performance of the MIP-type CSP compared favorably with that achieved on the commercial OMV and α1-AGP protein-type CSPs.
Kempe evaluated the effect of gradient elution on the chromatographic performance of MIP-type CSPs [34]. Using chloroform–AcOH gradients with poly(PETRA-co-MAA) CSP imprinted with Z-l-Tyr, baseline enantiomer separation of the racemic template could be achieved, with separation efficiencies of 980 and 3750 plates m−1 for the l and r enantiomers, respectively (Fig. 6). Similar enhancements in column efficiency were obtained with other MIP-type CSPs imprinted with structurally related N-blocked amino acids and dipeptide derivatives. It should, however, be noted that these improvements in peak shape relate to apparent figures, because the theoretical plate model is strictly defined for isocratic conditions only.
Separation of 100 μg racemic Z-Tyr-OH on a poly(PETRA-co-MAA) MIP-type CSP imprinted with Z-l-Tyr-OH. Gradient elution at 1 mL min−1 with chloroform–AcOH, 96:4 (v/v) (component A), and chloroform–AcOH, 8:2 (v/v) (component B). 0–23 min, 0% B; 23–24 min, 0–100% B; 24–43 min, 100% B; 43–47 min, 0% B. Reprinted with permission from Ref. [34]
SFC mobile phase conditions, employing carbon dioxide in combination with polar modifiers, have been shown to be fully compatible with a large number of commercial CSPs, and to enhance the enantiomer separation performance in terms of speed, efficiency, and preparative throughput [63]. Additional benefits, for example fast column equilibration, ease of product recovery, and reduced safety hazards, have made SFC an attractive alternative to traditional mobile phases in chromatographic enantiomer separation. Recent exploratory efforts to operate MIP-type CSPs with SFC-type mobile phases have, however, led to rather disappointing results. Ellwanger et al. studied the chromatographic behavior of PA on an MIP-type CSP imprinted against the l enantiomer under SFC conditions [62]. A high proportion of polar modifier (40%, MeOH–AcOH 95:5 v/v) was required to achieve analyte elution at a column temperature of 50°C and an outlet pressure of 150 bar. Although some enantiomer separation could be achieved, the separation suffered from poor peak shapes and extremely low efficiency. An increase in outlet pressure to 200 bar resulted in compromised enantioselectivity rather than in improved efficiencies. After few days of operation under SFC conditions, furthermore, the enantiomer-separation capacity of the MIP was almost entirely lost (Fig. 7). This indicated that the enantioselective sites of the MIP-type CSP had largely been destroyed, either by pressure-induced mechanical stress or chemical degradation of important binding site functionality.
Chiral SFC separation of phenylalanine anilide (PA) at 50 °C (a) and 100 °C (c) using a poly(EDMA-co-MAA) MIP-type CSP imprinted with l-PA. (b) Deterioration of the separation performance of the MIP-type CSP after two days of operation in SFC. (d) Retention behavior of PA on the blank polymer under identical conditions. Conditions: flow rate, 2 mL min−1; oven temperature, 50 °C (a, b, d) and 100 °C (c); outlet pressure, 150 bar; mobile phase, CO2 , 40% modifier (MeOH–AcOH, 95:5, v/v) added; detection, 260 nm; injection, 10 mL of a 5 mmol L−1 stock solution. Adapted from Ref. [62]
A very limited amount of data is available on the use of MIP-type CSPs for preparative separation of enantiomers. The few papers addressing this issue claim successful enantiomer separation of 0.1 to 1.0 mg racemic template on analytically-sized MIP columns (Fig. 8) [34, 39, 64].
Elution profiles of d,l-PA injected on to a column (125 mm × 4 mm) packed with a poly(EDMA-co-MAA) MIP-type CSP imprinted with l-PA. Mobile phase MeCN–TFA (0.01%)–water (2.5%). Reprinted with permission from Ref. [6]
The “preparative” loading capacity of MIPs is comparable with that of commercial protein-type CSPs but far inferior to that of enantioselective adsorbents currently used for preparative applications. For example, for π-donor–acceptor-type CSPs baseline enantiomer separation have been reported for sample loads in the range 10 mg racemate g−1 CSP, and even higher loading capacities have been achieved with polysaccharide-derived adsorbents (up to 30 mg racemate g CSP) [65]. General statements on the usefulness of a given CSPs for preparative applications are problematic, however, because preparative loading capacities are known to vary, depending on the chemical nature of the enantiomers and the mobile phase conditions used. Lindholm et al. suggested use of the “true” saturation capacities of the enantioselective sites as a figure of merit for assessment of the preparative potential of different types of CSP [66]. Recent studies by Guiochon’s group have established that MIP-type CSPs have very low saturation capacity for the enantioselective sites (<3 μmol mL−1), with the additional drawback of a highly unfavorable enantioselective to nonselective site ratio [60, 67, 68].
Apart from the prohibitively low loading capacity, the poor chromatographic efficiency of MIP-type CSPs is another serious impediment for their preparative use. Short cycle times between runs are absolutely essential for an economic chromatographic enantiomer separation on the preparative scale. With MIP-type CSPs, at ambient temperature and under isocratic elution conditions, large volumes of mobile phase may be required to achieve complete elution of the template enantiomer, resulting in highly dilute product streams, elaborate recovery procedures, and ultimately poor productivity. As remarked above, operating MIP-type CSPs at elevated temperatures or under gradient conditions may partially alleviate these problems. These measures call for cost-intensive non-standard engineering concepts, however, investment that is difficult to justify because of the low loading-capacity of MIP-type CSPs.
Capillary MIP formats
Capillary electrochromatography (CEC) is the second technique that has been extensively exploited for MIP-based enantiomer separation [69–72]. CEC is a hybrid separation technique using elements of liquid chromatography and electrophoresis, and is often claimed to combine the benefits of both [73]. Specifically, mobile-phase flow through CEC systems is accomplished by endoelectroosmotic flow (EOF) instead by external pumping, giving rise to a stable plug-like, rather than a parabolic, flow profile. This flat flow profile diminishes unfavorable peak dispersion phenomena while promoting pore flow, features that result in dramatically enhanced separation efficiencies compared with pressure-driven chromatographic techniques. The use of stationary phases in CEC enables the design of separation systems with predictable selectivity profiles, and improves sample capacity and circumvents many of the background buffer additive-derived limitations plaguing capillary electrophoresis. An additional attractive feature of CEC separation systems is their inherently small dimensions, which minimize stationary phase requirements and operating costs compared with conventional HPLC formats.
Employing MIP-type CSPs in conjunction with CEC seems to be particularly attractive with regard to possible enhancement of separation efficiency, backpressure-free operation, and reduced consumption expensive functional monomers and templates of limited availability. The preparation of MIP-charged capillary columns suitable for CEC applications was found be more challenging than in HPLC, however, demanding, essentially, the development of dedicated MIP formats.
Capillaries packed with particulate MIPs
Different approaches to MIP-type capillary formats have been proposed, including gel-entrapping of pre-formed particles [74, 75], capillaries packed with silica-grafted MIP films [10], surface-coating procedures [76, 77], in-situ generation of superporous monoliths [7, 78–82], and use of pseudo-stationary phases comprising nanoparticles [83–85]. Early efforts to create appropriate capillary columns involved the packing of MIP particles derived from bulk polymers by grinding and sieving. Stabilization of the particle packings within the capillaries was accomplished by entrapment in supporting matrices of polymer or silicate networks. Following this strategy, Lin et al. produced capillaries in which l-PA-imprinted polymer particles were fixed by incorporation into a polyacrylamide matrix [74]. As a variation on this theme, Vallano and Remcho described a silicate-based sol–gel process with subsequent thermal aging to stabilize packings of nortriptyline-imprinted polymer particles [75]. With both types of column successful enantiomer separation of the racemic templates and related compounds could be achieved. Because of the dense nature of these composites, however, solvent permeability was very limited, making mobile-phase exchange or electroosmotic flow regeneration after breakdown extremely difficult.
Quaglia and coworkers evaluated capillaries packed with silica-grafted MIP composite particles for capillary electrochromatographic enantiomer separation [9, 10]. These composite MIPs for l-PA were prepared by using azoinitiator-modified spherical silica support characterized by a well-defined particle size (10 μm) and pore system (100 nm). The enantioselectivity of fused silica capillaries packed with these supported MIP particles was comparable with that achievable under HPLC conditions, and performance in terms of plate numbers was somewhat better. Under optimized operating conditions and with short capillaries (effective length 8 cm), the enantiomers of PA could be successfully separated within a favorably short analysis time (3 min). In addition, satisfactory inter-intraday and intercapillary reproducibility could be demonstrated. In terms of separation efficiency, however, these phases can hardly compete with established enantioselective CEC media, rendering their utility for analytical applications questionable.
In-situ generation of superporous MIP monoliths
In an effort to address these drawbacks Nilsson’s group has established procedures for the in-situ generation of enantioselective superporous MIP monoliths within the confines of fused silica capillaries [81]. For this purpose, fused silica capillaries were first chemically modified with methacryloxypropyltrimethoxysilane to attach vinyl anchor groups for the polymer rod to be generated. These chemically modified capillaries were charged with MAA–TRIM–(R)-propranolol mixtures in toluene, an apolar porogenic solvent favoring the formation of MAA–template complexes. The filled capillaries were then exposed at subambient temperatures to a UV-source to initiate photochemical polymerization. Precise control over the porous structure of monoliths could be achieved by careful timing of the polymerization reaction. Premature termination of the polymerization reaction was found to be crucial to obtaining highly permeable superporous MIP monoliths (Fig. 9a).
(a) Scanning electron micrograph of the cross section of a superporous monolithic MIP inside a capillary column. (b) Separation of the enantiomers of propranolol on a superporous capillary column. Reprinted with permission from Ref. [81]
The template and the unreacted components of the pre-polymerization mixture could easily be removed from the superporous polymer network formed by flushing with appropriate solvents. It is worth mentioning that the complete procedure, from start of capillary modification to analytical evaluation required less than 3 h. The resulting monolithic MIP phases were shown to be capable of resolving the template enantiomers under optimized CEC conditions in less than 120 s with enantioselectivity α = 1.12 and remarkable resolution R S = 1.29 (Fig. 9b). The authors emphasized the operational simplicity, the speed, the low consumption of material (ca 10 μg template per capillary), and good reproducibility as the major advantages of their approach. In further studies the effect of using different experimental conditions for MIP synthesis on the chiral recognition capabilities of these CEC phases was investigated [86]. The authors found that the enantioselectivity and retention of the imprinted monoliths could be tuned by using additional functional monomers; for example, methyl and butyl methacrylate enhanced both enantioselectivity and resolution. In addition, by performing the imprinting process in presence of two structurally different chiral templates, “multi-specific” enantioselective monolithic CEC phases were obtained. Thus, an MIP-type monolith imprinted with a mixture of single enantiomers of the β-blocker atenolol and the local anesthetic ropivacaine had appreciable enantioselectivity toward both racemic drug compounds. This concept to customize separation media with multi-analyte chiral recognition profiles may offer interesting opportunities in the enantioselective analysis of multi-drug formulations.
Capillaries coated with MIP films
Imprinted polymer films grafted on to the inner surface of capillaries are another attractive format for MIP-based CEC enantiomer separation. The most obvious advantage of this open tubular (OT) configuration over other formats is the lack of significant back pressure, a feature that greatly facilitates column regeneration and mobile phase exchange by hydrodynamic flushing. The use of surface-coated formats also reduces the risks of bubble formation and capillary clogging, phenomena that notoriously perturb the performance of packed and monolithic CEC columns. The major challenge associated with successful realization of OT MIP formats consists in devising efficient strategies for the controlled deposition of functional imprinted polymer layers on the capillary walls. Generally, compromises may have to be made to harmonize the experimental conditions required for film formation with the often different conditions favoring successful molecular imprinting.
This difficulty surfaced in the pioneering study by Brueggemann et al., who reported the generation of OT MIP capillary columns for (S)-2-phenylpropionic acid [87]. As first step fused silica capillaries were treated with methacryloxypropyltrimethoxysilane to provide “primer” vinyl groups for a surface-centered polymerization reaction. These methacrylate-grafted capillaries were then charged with pre-polymerization mixtures of different composition, and subjected to thermally induced polymerization. A large set of different monomer–porogen combinations was systematically evaluated with regard to their film-forming properties. Only a small proportion (ca 20%) of the formulations tested furnished wall-attached MIP films of acceptable quality. The greatest success was achieved with formulations comprising dimethylsulfoxide as porogenic solvent. This solvent is regarded as unfavorable for noncovalent molecular imprinting, however, because it destabilizes functional monomer–template complexes because of its hydrogen-bond-disrupting properties. Accordingly, the CEC performance of the respective MIP-type OT capillaries was rather poor. Although the authors claimed successful separation of the enantiomers of the racemic template molecule, this claim may be questioned as the peak of the template enantiomer could not be detected unambiguously.
Tan et al. used a similar approach to prepare enantioselective MIP-type CEC columns for l-danysl Phe; the method also entailed thermal polymerization of monomer–template–porogen mixtures in the confines of methacrylate-grafted capillaries [77]. To promote the formation of homogenous polymer coatings the resulting capillaries were evacuated to remove residual solvents and then exposed to gas pressure. It was claimed this treatment shrunk the formed MIPs on to the capillary surface. The procedure was, however, of limited efficiency, because permeable OT capillaries were obtained with a 50% success rate only. Successful CEC separations of the enantiomers of dansyl-Phe could, nevertheless, be achieved with these capillaries.
An innovative strategy greatly facilitating the synthesis of enantioselective MIP-type OT CEC capillary columns has been reported by Schweitz [76]. Key to this approach is a “grafting-from” polymerization scheme made possible by covalent attachment of radical initiators on to the inner capillary surface. During MIP formation the thermal–photochemical decomposition of these immobilized initiator species provides a high density of free radicals on the capillary wall, thereby confining the polymerization reaction to the vicinity of the surface rather than to the bulk solution. The feasibility of this concept was demonstrated by the preparation of OT CEC capillaries imprinted with (S)-propranolol. For this purpose an azoinitiator was coupled to aminopropylsilane-modified fused silica capillaries using carbodiimide chemistry. The initiator-grafted capillaries were charged with pre-polymerization mixtures consisting of MAA, TRIM, and the template, dissolved in toluene, DCM, or MeCN as the porogenic solvents. The polymerization was performed photochemcially at ambient temperature for 3 h, followed by removal of the template and unreacted components by simple flushing with MeCN–AcOH. Characterization of the capillaries by electron microscopy revealed that formation of MIP coatings was always successful. The polymer layers differed in thickness, texture, and morphology, depending on the porogenic solvent used (Fig. 10-I).
(I) Scanning electron micrographs of the MIP coatings inside 50-μm fused-silica capillaries showing solvent-dependent differences in appearance. (a) Porogen toluene, (b) porogen DCM, and (c) porogen MeCN. The scale bars indicate 5 μm. (II) Electrochromatograms obtained from separation of the enantiomers of propranolol using MIP coatings synthesized in different solvents: (a) toluene, (b) DCM, (c) MeCN. Order of elution: (R)-propranolol followed by (S)-propranolol. Conditions: capillary, 50 μm inner diameter, 35 cm total length, and 26.5 cm effective length; separation potential, 15 kV; temperature, 60 °C. Reprinted with permission from Ref. [76]
With toluene as porogen a relatively thin coating (0.15 to 0.45 μm) with a rough surface appearance was obtained, whereas DCM and MeCN gave rise to thicker layers (2 to 4 μm) patterned with polymer aggregates. CEC evaluation of these capillaries with racemic propranolol under identical experimental conditions resulted in successful enantiomer separation in all cases (Fig. 10-II). Although the MIP coating prepared in toluene enabled more efficient CEC enantiomer separation, retention and enantioselectivity were higher for coatings formed in the presence of DCM and MeCN.
MIP nanoparticles as pseudo-stationary phases
A recent achievement in CEC is enantiomer separation effected by using chirally imprinted nanoparticles as pseudostationary phases [83–85, 88]. Imprinted nanoparticles in the size range 100–600 nm are readily accessible by precipitation polymerization procedures using highly diluted solution of the imprinting components in porogens with poor solvation power for the resulting polymers [89]. Schweitz et al. synthesized (S)-propranolol-imprinted MAA/TRIM nanoparticles by photopolymerization at low temperatures, and subsequently evaluated these nano-MIPs for CEC enantiomer separation [83]. To avoid interference with analyte detection, the authors resorted to partial-filling CEC techniques, charging the capillary with a plug of a nanoparticle suspension before injection of racemic propranolol. On applying voltage across the capillary, successful enantiomer separation of propranolol could be achieved by migration through the MIP plug. Crucial to the success of the partial-filling mode enantiomer separation was the use of capillaries with suppressed EOF characteristics (by surface silylation) to slow the co-current mobility of the nanoparticle plug. In a further study the authors investigated the effect of the experimental variables of the polymerization process on the CEC performance of the resulting MIP nanoparticles [85]. The use of MAA in combination with weakly interacting functional monomers was found to improve CEC separation efficiency. It was also noted that the template/functional monomer ratio used in the pre-polymerization mixture had a pronounced effect on the size and morphology of the nanoparticles. In a subsequent contribution the authors reported the possibility of preparing enantioselective nanoparticles imprinted with two different, chemically unrelated chiral targets [84]. The synthesis of these “multiply enantioselective” nanoparticles for (S)-propranolol and (S)-ropivacaine, however, proved experimentally challenging, because of the competition of the templates for functional monomers. The concentration of the strongly interactive propranolol had to be reduced substantially relative to that of the weakly interactive ropivacaine to furnish nanoparticles with the desired selectivity. Successful partial filling CEC enantiomer separation of racemic ropivacaine and propranolol could be accomplished with these multiply imprinted nanoparticles.
Although operationally complex, partial filling CEC with imprinted nanoparticles has the advantage of flexibility in the choice of the amount, composition, and type of stationary phase. The non-immobilized nature of the nanoparticle pseudostationary phase also enables easy renewal after each run, avoiding problems arising from column aging and irreversible adsorption of contaminants by packed columns.
Enantioselective MIP membranes
The ever-increasing demand for single-enantiomer compounds has intensified research activity toward the development of industrially viable enantiomer resolution technology. Apart from continuously operating simulated moving bed (SMB) chromatography, enantioselective membrane-based separation technology are regarded as promising [90]. Membrane-based enantiomer separation would have several advantages over chromatographic processes, for example equipment of lower complexity, choice of different process configurations, ease of scale-up, relatively low energy requirements and thus low operating costs [3]. Progress in the field is, unfortunately, seriously hampered by the lack of suitable membrane materials, i.e. which meet the crucial requirements of industrial-scale enantiomer separation. In addition to high target-specific enantioselectivity, ideal membrane materials must provide the features of ready accessibility, scalability, reasonable production costs, and good long-term stability under process relevant conditions. Chirally imprinted MIPs, in principle, fulfill all these requirements and may therefore provide interesting prospects for the development of membrane-based enantiomer separation processes.
Several research papers exploring the development of chirally imprinted MIP-type membranes have focused on different strategies for engineering of formats applicable to enantiomer separation. Dzgoev and Haupt reported the preparation and evaluation of Z-l-Tyr-imprinted composite membranes [91]. Initial attempts at obtaining free-standing membranes directly by casting of a optimized bulk MIP failed because of the poor mechanical stability of the resulting materials. This limitation was successfully addressed by incorporating a specifically adapted MIP formulation mixture into the porous confines of commercial polypropylene membranes, followed by photo-polymerization. The corresponding composite membranes had enhanced mechanical robustness, enabling membrane-diffusion experiments to be conducted. In diffusion experiments performed with single enantiomers enhanced transport of the template enantiomer from 1:1 chloroform–MeOH was observed, providing evidence of facilitated transport assisted by the imprinted enantioselective binding sites. The initial flux rates were found to be 1.17 and 0.34 mmol m−2 h−1 for Z-l-Tyr and Z-d-Tyr, respectively, corresponding to an operational enantioselectivity α = 3.4. When the membrane was challenged under otherwise identical conditions with racemic Z-Tyr, however, the differences between transport rates were less pronounced, indicative of competitive binding phenomena.
Using a similar approach, Donato et al. [92] created enantioselective composite membranes for the anti-inflammatory drug naproxen. The membrane materials were created by impregnation of the polypropylene support with a 4-VPy–EDMA–(S)-naproxen–toluene-based MIP formulation by photopolymerization. The resulting flexible membranes were shown to be dense in nature, with reduced water permeability compared with the parent supporting materials. SEM surface investigations provided evidence of the existence of a 30-μm-thick MIP layer homogeneously covering the surface and pores of the underlying polypropylene carrier membrane. Permeation studies were conducted by pressure-driven (0.4 bar) dead-end filtration with racemic naproxen (7 μg mL−1) in 10 mm PBS at different pH. Enantioselective membrane transport was observed for the template enantiomer under acidic conditions, with the highest enantioselectivity (α = 1.6) achieved at pH 3.7.
In addition to from these “traditional” molecular imprinting approaches, “alternative” molecular imprinting techniques have been introduced by Yoshikawa et al. for preparation of enantioselective membranes [93]. In contrast to established imprinting technology, this alternative strategy utilizes pre-formed non-crosslinked polymers, which serve simultaneously as functional recognition elements for the templates and as supporting matrices for the binding sites to be formed. Imprinted membranes are obtained by simple casting of the template-containing polymer solutions instead by use of cross-linking polymerization reactions, avoiding the unfavorable action of thermal and/or photochemical stress on templates and pre-polymerization complexes. A variety of chiral and achiral polymers have been evaluated for alternative molecular imprinting; examples include blends of peptide-grafted polystyrene–polyacrylonitrile [94], cellulose acetate [95], carboxylated polysulfones [96], and polyamides [97, 98].
Alternative imprinted membranes have been prepared for several amino acid templates, and have been shown to enable enantioselective adsorption and permeation. With membrane materials created from inherently chiral pre-polymers, for example peptide-grafted polystyrenes or cellulose acetate, observation of chiral recognition properties does not come as a surprise. For such membranes it may be difficult to assess to what extent the enantioselective binding properties originate from interactions with the imprinted binding sites or from interactions with the inherently chiral sub-units of the polymeric backbone. Enantioselective membranes have, however, also been obtained by alternative imprinting from inherently achiral pre-polymers. Yoshikawa’s group produced free standing membranes by casting carboxylated polysulfone (degree of substitution 0.88) in presence of Z-d-Glu from THF solutions [96]. Depending on the amount of template used, the resulting membranes had different enantioselective adsorption capacity from 50% aqueous ethanol. Membranes prepared with an excess of template (one or three equivalents relative the total carboxyl group concentration of the polymer) lacked chiral recognition capacity; those prepared with 0.5 equivalents of template were enantioselective, although modestly (α = 1.2). The presence of polymer-embedded binding sites could, however, be confirmed by repeating the imprinting procedure with the enantiomeric template; this also produced a stereodiscriminating membrane but with the reverse enantioselectivity. In subsequent electrodialysis experiments the membranes were found to have enantioselective transport properties, with the template enantiomer permeating more rapidly (Fig. 11).
Time–transport curves for electrodialysis of d and l-Glu through d and l-Glu-imprinted carboxylated polysulfone membranes. Applied potential ΔE = 2.5 V. Adapted from Ref. [96]
The membranes could also be used for chiral recognition of the free Glu enantiomers, with enantioselectivity similar to that for the corresponding Z derivatives. This finding is indicative of imprinted sites with some cross-selectivity, a behavior frequently observed for conventional imprinted polymers. In a subsequent study, performed with a different pre-polymer (poly(hexamethylene terephthalamide–isophthalamide)) and the enantiomers of Z-Glu as templates, the authors obtained imprinted membranes with enhanced chiral recognition properties (α = 2.0) [97]. These findings suggest this approach to imprinting may be an attractive alternative to conventional imprinting techniques for production of selective membrane materials. The molecular memory of these non-crosslinked membrane materials is, however, easily compromised by polymer swelling, which may induce irreversible collapse of the binding domains.
Batch extraction
Over the last decade significant progress has been achieved in enantioselective catalysis and biotechnology, and processes based on these technologies are increasingly being implemented for large-scale production of key chiral intermediates [99–104]. The enantiomeric or diastereomeric purity of the products emerging from these processes is usually high. Achieving complete control over competing stereochemical pathways, however, remains challenging, and simultaneous formation of minor amounts of enantiomeric or diastereomeric impurities is difficult to suppress. The similar physicochemical properties of products and contaminants make the removal of chiral impurities from production batches an elaborate and cost-intensive task. Eventually, lack of efficient purification schemes may disqualify otherwise promising synthetic strategies for production of chiral intermediates for which high purity is required, for example drugs and food additives. Selective MIP-type adsorbents imprinted with these enantiomeric or diastereomeric contaminants may be an attractive and cost-effective solution to this problem. Optimized MIPs with high affinity and target-specificity may enable efficient scavenging of contaminants from production streams, even in presence of a large excess of stereochemically closely related molecules, and thus facilitate product refinement. Ease of synthesis, scalability and durability under chemically demanding conditions are additional properties that make MIPs ideal target-specific adsorbents for industrial use.
The potential of MIPs to serve as “stereoselective polishing agents” has been explored by Ye in an interesting proof-of-principle study [13]. In this study a complex reaction mixture emerging in the course of the synthesis of an aspartam intermediate, N-(benzyloxycarbonyl)aspartylphenylalanine methyl ester (ZAPM), was selected as a challenging model case of a severely contaminated production batch. The reaction mixture, obtained by coupling of Z-l-aspartic anhydride with l-Phe methyl ester, consisted of the required product, α-l,l-ZAPM (59%), the corresponding β-isomeric byproduct, β-l,l-ZAPM (19%), and Z-l-Asp (22%). The authors prepared a range of MIPs for the β-l,l-ZAPM impurity by using different molar ratios of VPy and MAA as functional monomers and MeCN as porogenic solvent. Systematic chromatographic tests and solid-phase extraction experiments were conducted to identify combinations of MIP materials and operating conditions that would be a favorable compromise between byproduct binding and undesired co-binding of the required product. The optimized adsorption conditions were finally used to perform repeated cycles of solid-phase extraction of the crude (diluted) MeCN reaction mixture. The purity of the required product could be enhanced from the initial 59% to 96% by five extraction cycles. In addition, complete removal of Z-l-Asp could be achieved, and the level of contamination with β-Z,Z-l-ZAPM was reduced to 4%.
In a subsequent study the authors applied this promising MIP-based scavenging concept to a real-world purification problem. This challenge entailed the removal of trace amounts of a fermentation contaminant, succinyl l-Tyr, from a production batch of the β-lactamase inhibitor clavulanic acid [14]. For this purpose a highly specific MIP was tailored using mixtures of vinylbenzyl trimethylammonium chloride and methacrylic acid as functional monomers and succinyl l-Tyr as template. The optimized MIP had pronounced affinity (K d = 330 μm) and capacity (B max = 81 mg g−1) for the succinyl l-Tyr contaminant, even in highly polar environments such as MeOH. With this material selective adsorption of less than 2% succinyl l-Tyr could be achieved from MeOH solutions containing up to 15 mg mL−1 clavulanic acid. Clavulanic acid was recovered after this treatment in excellent yield (97.6%) and acceptable chemical purity >99%. The exhausted affinity adsorbent could be easily regenerated by a simple acidic washing procedure without loss of binding performance, and could be reused for further purification cycles.
MIP-based enantioselective sensors
Current progress in the development of asymmetric catalysis and enantioselective biotransformation is driven by large-scale combinatorial discovery efforts [105–107]. In these programs large numbers of catalysts are screened, in parallel, under many sets of conditions, and thousands of samples are accumulated; the product distribution of these must be analyzed, to determine conversion and enantioselectivity, in matter of days. Conventional chromatographic assays for determination of enantiomer purity are instrumentally demanding and designed for serial rather than parallel sample processing; they are, therefore ill-suited to the challenge of high-throughput screening. This situation has created a need for cost-effective and operationally simpler high-throughput techniques for assessment of enantiomer purity [108, 109]. Although still in an early stage of development, enantioselective sensing devices are regarded as interesting alternatives to time-consuming, serially operating separation methods [110–113]. Enantioselective sensors using MIPs as recognition elements may be particularly appealing solutions. The ease and speed with which MIPs with predetermined enantioselectivity can be prepared from inexpensive bulk chemicals obviates the need to screen large numbers of chiral selectors to identify suitable enantioselective recognition elements for a given target analyte. The greater substrate specificity of MIPs compared with conventional chiral selectors may help to suppress cross-reactivity, enabling the design of sensor devices capable of operating in matrix-loaded environments. MIPs may also be shaped in different formats, facilitating the task of interfacing enantioselective recognition elements with different transducer units. MIPs also can be designed as self-reporting materials, combining molecular recognition with signaling functions, a feature that facilitates the monitoring of selective binding. Particularly attractive in this context are secondary stimuli triggered by template binding, for example changes in the MIP swelling state, attenuation of membrane permeability, and quenching and/or amplification of auto-fluorescent properties. These phenomena may provide opportunities for enhancement of insensitive primary transducer responses and the design of complementary readout options.
Gravimetric sensor devices
Piezoelectric thickness-shear-mode acoustic devices, for example quartz micro balances (QCMs) and surface acoustic wave devices (SAWs), have proved particularly versatile transducer elements for the design of MIP-based gravimetric sensing systems [114–118]. The surfaces of these inexpensive and readily available transducers can be easily modified with MIP recognition layers by experimentally simple in-situ polymerization procedures. Mass changes associated with selective analyte binding to the imprinted polymer films induce shifts in the frequency of the acoustic resonators, providing an easily measurable readout for monitoring the underlying molecular recognition events. A particularly appealing feature of these gravimetric sensor devices is their excellent sensitivity, with standard QCM equipment operating at a fundamental frequency of 10 MHz, enabling detection of mass changes in the low-nanogram range [119].
Pioneering efforts to adapt MIP-based QCM sensing technology to chiral recognition of (S)-propranolol have been reported by Haupt et al. [120]. The enantioselective chiral recognition layers were generated on the gold-coated surfaces of 5 MHz quartz crystals employing an established poly(TRIM-co-MAA) MIP formulation. To improve polymer flexibility and film adherence to the gold surface, however, the total concentration of crosslinker (38% of total monomer) had to be reduced relative to the original formulation. To ensure the formation of homogeneous MIP coatings and to minimize inhibition of polymerization by oxygen, the imprinting mixture was sandwiched between the QCM and a quartz cover plate before polymerization. Exposure of this sandwich assembly to UV irradiation produced MIP films 2 μm thick with satisfactory adherence to the supporting resonator surface. For evaluation of the chiral recognition properties the modified QCMs were challenged with increasing concentrations of propranolol enantiomers in MeCN containing AcOH. A pronounced decrease in the resonator frequency was observed for (S)-propranolol over the analyte concentration range investigated (0.05 to 1.5 mmol L−1) whereas (R)-propranolol induced only minor frequency changes. The initial slopes of the binding isotherms were estimated to be −500 and −100 Hz mmol−1 dm3 for (S) and (R)-propranolol, respectively, corresponding to appreciable enantioselectivity of α ≈ 5.0 (Fig. 12).
Resonance frequency change of a 5-MHz Au–quartz electrode coated with a (S)-propranolol-imprinted polymer film exposed to solutions of (R)-propranolol (curve 1) and (S)-propranolol (curve 2) in 33.7 mol L−1 acetic acid in MeCN. Adapted from Ref. [120]
The analytically useful operating range of this QCM sensor, however, was found to be quite narrow (0.1 to 1 mmol L−1 for (S)-propranolol), because of binding site saturation. The QCM was also of limited sensitivity, with a limit of detection of 0.05 mmol L−1 (S)-propranolol. The authors argued that these drawbacks might be addressed by using QCMs with higher resonator frequencies and MIPs with enhanced binding site densities.
Recent work from Percival’s group suggests that enantioselective MIP-type QCM sensors may offer unique opportunities to address “difficult-to-recognize” classes of chiral molecules, for example rather apolar terpenes and zwitterionic amino acids. For example, by employing rather simple poly(TRIM-co-MAA) imprinting formulations these authors could generate efficient enantioselective QCM sensors both for the terpene alcohol l-menthol [121] and the highly polar amino acid l-serine [122]. The l-menthol-imprinted QCM sensor was reported to be capable of detecting the corresponding template in ethanol solution at the 200-ppb level, with an response range of 0–1.0 ppm and promising enantioselectivity of α ≈ 3.4.
An investigation focusing on improving the often rather slow responses of enantioselective MIP-type QCM sensors for (S)-propranolol has been reported recently by Piacham et al. [123]. In this study, an interesting grafting-from polymerization approach was proposed for deposition of ultrathin MIP films on to QCM sensor surfaces to enhance the analyte mass transport characteristics. The procedure involved formation of a carboxy-terminated self-assembled thiol monolayer on the gold-coated surface of a 10 MHz QCM crystal, which then served as a platform for covalent attachment of a free-radical initiator species. The initiator-functionalized QCM were impregnated with dilute MeCN solutions of MAA, TRIM, and (S)-propranolol and exposed at low temperature to UV irradiation to trigger polymerization. Characterization of the resulting QCM surfaces by ellipsometry provided evidence of the successful formation of ultrathin MIP films with a layer thickness <50 nm. Evaluation of these QCM sensors with the individual propranolol enantiomers in aqueous MeCN indeed revealed favorably fast (<1 min) and completely reversible sensor response, enabling multiple measurements without time-consuming surface regeneration. The sensitivity of these sensor layers was rather poor, however, because template concentrations in the mmol L−1 range were required to produce significant frequency shifts. The chiral recognition characteristics of these sensors were, moreover, indicative of somewhat counterintuitive behavior. Thus, the sensor response was essentially nonselective at low template concentrations but enantioselectivity was modest (α ≈ 1.2) when challenged with higher analyte concentrations. This observation suggests that for QCM-type sensors there may be an optimum MIP layer thickness for achieving a favorable compromise between response characteristics, sensitivity, and enantioselectivity.
Optical sensor devices
Nopper et al. investigated enantioselective binding of O,O′-dibenzoyltartaric acid enantiomers to imprinted polymer layers by use of an optical sensor using reflectometric interference spectroscopy (RIfS) as the transducer principle [124]. The enantioselective MIP films serving as recognition elements were created by spin coating optimized pre-polymerization mixtures on to methacroylpropylsilane-functionalized glass slides, followed by photochemical polymerization. The polymerization mixtures used comprised the respective (R,R)- or (S,S)-O,O′-dibenzoyltartaric acid as templates, EDMA as the crosslinking agent, and the strongly interacting ethyl 4-vinylbenzamide as functional monomer. After washing, the sensor layers were integrated in a flow cell and exposed to increasing concentrations of the templates and their respective enantiomers. The corresponding binding-induced changes in the optical thickness were monitored by RIfS. Linear response relationships between optical thickness and analyte concentration were observed over the concentration range studied. The templates were more efficiently bound than their respective enantiomers, but the enantioselectivity achieved (α ≈ 1.2) was modest.
One possibility of reducing the instrumental demands associated with many MIP-based sensing formats is to integrate analyte-responsive chromophoric and fluorophoric functionalities into the polymer matrix. Changes in the optical properties of these MIP materials induced by analyte binding may be exploited as self-reporting signals to monitor the specific molecular recognition event. The feasibility of this approach was evaluated by Liao and coworkers in an effort to prepare an enantioselective MIP-type fluorescence sensor system for l-tryptophan [125]. An appropriate fluorescent functional monomer was synthesized by coupling dansyl chloride via a short bifunctional linker to 3,3-dimethylacrylic acid. For preparation of fluorescent MIPs this monomer was co-polymerized with EDMA and l-tryptophan in the presence of AIBN and MeCN as porogenic solvent under thermal conditions. Binding experiments performed with 10 mmol L−1 l-tryptophan from a two-phase mixture of chloroform and 30 mmol L−1 citric acid (50:50, v/v) produced only minor changes in fluorescence emission. To enhance the sensitivity of the MIP sensor an indirect sensing scheme based on a quencher displacement procedure was adopted. Equilibration of the MIP with 4-nitrobenzaldehyde (3 mmol L−1) as a quencher was found to reduce the inherent fluorescence intensity of the polymer by more than 75%, because of nonspecific binding. Incubation with increasing concentrations of l-tryptophan (1 to 10 mmol L−1) displaced the quenching agent from the imprinted binding sites, leading to concentration-dependent enhancement of the MIP fluorescence. The increase in MIP fluorescence observed on incubation with d-tryptophan was less pronounced (approximately 70% of that obtained with the l enantiomer), and corresponded to enantioselectivity, α, of 1.45.
Another appealing concept for generation of self-reporting enantioselective MIP sensors exploits the unique optical properties of photonic hydrogels featuring a highly ordered 3D structure of spatially well-defined macropores [126]. The optical diffraction wavelength (and other optical properties, for example UV–visible absorbance and fluorescence) of these hydrogels are controlled by the periodicity of their pore structure. Environmental changes affecting the periodic pore structure of the polymer as a result of shrinking and/or swelling phenomena induce shifts in the diffraction wavelength, which can be used as an optical readout to monitor the underlying physical and chemical processes. Changes in the swelling state on analyte binding are frequently seen with MIPs, making them particularly promising materials for the design of self-reporting photonic sensor systems. In a recent study, Hu et al. demonstrated the feasibility of photonic MIP-based materials for sensitive and highly enantioselective sensing of l-dopa [127]. To create these molecularly imprinted hydrogels with a 3D ordered pore structure, supported colloidal crystals assembled from spherical silica nanoparticles were used as sacrificial templates (Fig. 13-I).
(I) Preparation of photonic MIP films: (a) SiO2 colloidal crystals on glass substrate. (b) infiltration of complex solution into colloidal template followed by photopolymerization; (c) photonic MIP film after removal of SiO2 nanoparticles and dopa template molecules; (d) complex of monomer and template molecule; (e) imprinted molecules within the polymer matrix; (f) imprinted cavities with complementary shape and binding sites to the template molecule. (II) UV–visible spectra of blank MIP films exposed to l-dopa (a) and d-dopa (b) in phosphate buffer at different concentrations: 10 mmol L−1 (a); 1 mmol L−1 (b); 0.1 mmol L−1 (c); 0.01 mmol L−1 (d); 1 μmol L−1 (e); 0.1 μmol L−1 (h); 0.01 μmol L−1 (g); and deionized water (h). (c) Three blank MIP films exposed to 0.1 mol L−1 phosphate buffer, 10 mmol L−1 d-dopa, and 10 mmol L−1 l-dopa in phosphate buffer. Reprinted with permission from Ref. [127]
For this purpose, monodisperse silica nanoparticles (186 nm) were first coated on glass substrates to form face-centered-cubic colloidal crystal layers. The void spaces within these layers were infiltrated with a pre-polymerization mixture consisting of MAA as functional monomer, EDMA as cross linking agent, l-dopa as print molecule, AIBN as initiator, and aqueous ethanol as porogenic solvent. After low-temperature photopolymerization the resulting silica–polymer composites were treated with hydrofluoric acid to remove the templating silica nanoparticles, yielding thin (2 μm) films of an l-dopa-imprinted hydrogel comprising a 3D ordered network of interconnected macropores. Highly enantioselective changes in the optical absorbance of these MIP films were observed on immersion in phosphate-buffered solutions of dopa enantiomers. Concentration-dependent shifts in optical absorbance were observed for sensor layers exposed to increasing amounts of l-dopa, with the maximum, blue-shifted by 66 nm relative to the unchallenged polymer, being in the presence of 10 mmol L−1 l-dopa. The optical response of the photonic sensor films was highly sensitive to l-dopa, enabling detection of the analyte down to 10 nmol L−1. In contrast, in the presence of d-dopa the absorbance wavelength of the imprinted films remained essentially unaffected (Fig. 13-II). This highly specific chiral recognition event could be followed even by the naked eye. MIP films immersed in with buffer solutions containing 10 mmol L−1 l-dopa had a blue appearance whereas those in contact with buffer solution containing 10 mmol L−1 d-dopa or pure buffer appeared green. A particularly attractive feature of imprinted hydrogel-type photonic sensors is their fast response times, enabling stable optical readout signals to be achieved after 20 s. These rapid analyte binding kinetics are because the interconnected pore structure of the imprinted hydrogels accelerates analyte mass transfer within the polymeric network.
Electrochemical sensors
Impressive enantioselectivity and substrate specificity for 2-phenylbutyric and 2-phenylpropionic acids have been achieved with ion-selective field-effect transistor (ISFET) devices featuring gate surfaces modified with TiO2-MIP films [128]. These highly efficient sensors were prepared by coating the Al2O3 gate interfaces of the ISFETs with ethanol or toluene solutions of titanium tetrabutoxide and the corresponding 2-phenylalkanoic acid enantiomers, followed by mild thermal curing (Fig. 14).
(a) Schematic configuration of the molecularly imprinted ISFET device. (b) Preparation of enantioselective molecularly imprinted sites for chiral carboxylic acids in a TiO2 film acting as the sensing interface on the ISFET gate. Adapted from Ref. [128]
Subsequent elution of the acidic templates from the imprinted binding sites liberated Ti–OH surface groups, which triggered a change in the gate potential. On re-binding of the acidic templates to the imprinted cavities these free Ti–OH groups became saturated and induced a reverse shift in the gate potential, thus converting the specific association events into electronically readable signals. Contacting the ISFET imprinted with (R)-2-phenylbutanoic acid with solutions containing increasing concentrations (0.0625–1.25 mmol L−1) of the template triggered a pronounced increase of the gate-source voltage. No changes in the gate-source voltage were observed when the ISFET was challenged with comparable concentrations of (S)-2-phenylbutanoic acid and (R)-2-phenylpropionic acid, however. Similar chiral recognition behavior was observed with ISFETs imprinted with the 2-phenylpropionic acid enantiomers. These findings suggest the binding sites in the imprinted TiO2 films have exceptionally high molecular complementarity with the templating molecules and are not only capable of discriminating between enantiomers but also between structurally closely related analogues with similar stereochemical features.
Zhou’s group recently reported use of innovative surface-imprinted technology as a simple yet efficient route to highly enantioselective potentiometric sensors [129]. The surface-imprinted working electrodes of these sensors were prepared by treatment of indium tin oxide (ITO) glass substrates with a mixture of Z-l-aspartic acid (Z-l-Asp) and octadecyltrichlorosilane (ODS) in apolar solvents. The modified ITO–ODS working electrodes produced, in the presence of aqueous solutions of Z-l-Asp, strong, concentration-dependent potential changes relative to an Ag/AgCl-reference electrode, but proved essentially insensitive to the corresponding d- enantiomer. Subsequent control experiments performed with racemic mixtures confirmed the excellent chiral recognition capacity of the surface-imprinted sensors, with potential changes being triggered exclusively by the enantiomer used as template. Potentiometric selectivity coefficients, K POT, were 4–9 × 10−3, corresponding to outstanding enantioselectivity (α = 100–250). The surface-imprinted sensors also proved highly sensitive, covering an analytical concentration range from 5 μmol L−1 to 12 mmol L−1 Z-l-Asp. The long-term stability of the sensors was also impressive, with only 8% loss of signal after 200 measurements. The authors rationalized the exceptional performance characteristics of these sensors in terms of the presence of highly enantioselective imprinted channels within the hydrophobic ODS monolayers which allowed only the template enantiomer access to the underlying proton-sensitive ITO transducer layer.
Conclusions
In principle, molecularly imprinted polymers are a valuable source of enantioselective receptor materials for a broad range of chiral recognition applications. Intelligent exploitation of this potential requires a clear understanding of the inherent limitations of these readily accessible chiral recognition materials. It is obvious that the current generation of noncovalently imprinted polymers suffers from binding site heterogeneity, slow mass transfer kinetics, and relatively low density of high-affinity binding sites, making them a rather poor choice for analytical and preparative chromatographic applications. These limitations, however, seem to be less prohibitive for chiral recognition methods based on equilibrium binding, for example binding assays, batch adsorption, and molecular sensing. In this context, enantioselective imprinted materials have considerable promise as efficient scavengers for refinement of chiral intermediates from industrial production streams. Also, in the near future, intelligently designed sensor devices based on enantioselective molecularly imprinted polymers may be applied as inexpensive and robust tools in high-throughput analysis, e.g. in support of combinatorial discovery of asymmetric (bio-)catalysts. Finally, future advances in molecularly imprinted membranes may provide new economical opportunities for large-scale enantiomer separations.
Abbreviations
- AcOH:
-
acetic acid
- α1-AGP:
-
acidic glycoprotein
- AIBN:
-
α,α-azoisobutyronitrile
- Asp:
-
aspartic acid
- Boc:
-
tert-butylcarboxycarbonyl protecting group
- Bmax:
-
saturation loading capacity
- CEC:
-
capillary electrochromatography
- CSP:
-
chiral stationary phase
- DCM:
-
dichloromethane
- EDMA:
-
ethylene dimethacrylate
- EOF:
-
electroosmotic flow
- Fmoc:
-
9-Fluorenylmethoxycarbonyl protecting group
- Glu:
-
glutamic acid
- HPLC:
-
high-performance liquid chromatography
- ITO:
-
indium tin oxide
- ISFET:
-
ion-selective field-effect transistor
- Kd:
-
dissociation constants
- K POT :
-
potentiometric selectivity coefficient
- MAA:
-
methacrylic acid
- MeCN:
-
acetonitrile
- MIP:
-
molecularly imprinted polymer
- NIP:
-
non-imprinted control polymer
- ODS:
-
octadecylsilane
- OT:
-
open tubular
- OVM:
-
ovomucoid
- PA:
-
phenylalanine anilide
- PBS:
-
phosphate-buffered saline
- PETRA:
-
pentaerythritol triacrylate
- Phe:
-
phenylalanine
- QCM:
-
quartz-crystal microbalance
- RAFT:
-
reversible addition–fragmentation chain transfer
- RIfS:
-
reflectometric interference spectroscopy
- SAW:
-
surface acoustic wave
- SFC:
-
supercritical-fluid chromatography
- SMB:
-
simulated moving bed
- THF:
-
tetrahydrofuran
- Trp:
-
tryptophan
- TRIM:
-
2,2-bis(hydroxymethyl)butanol trimethacrylate
- Tyr:
-
tyrosine
- VPy:
-
4-phenylpyridine
- Z:
-
benzyloxycarbonyl protecting group
References
Alexander CS, Andersson H, Andersson LI, Ansell RJ, Kirsch N, Nicholls IA, O’Mahony J, Whitcombe MJ (2006) J Mol Recognit 19:106
Ramstrom O, Ansell RJ (1998) Chirality 10:195
Maier NM, Franco P, Lindner W (2001) J Chromatogr A 906:3
Svenson J, Nicholls IA (2001) Anal Chim Acta 435:19
Kempe M (2001) Tech Instrum Anal Chem 23:395
Sellergren B (2001) J Chromatogr A 906:227
Schweitz L, Spegel P, Nilsson S (2001) Electrophoresis 22:4053
Spegel P, Nilsson J, Nilsson S (2004) Methods Mol Biol 243:411
Quaglia M, De Lorenzi E, Sulitzky C, Massolini G, Sellergren B, (2001) Analyst 126:1495
Quaglia M, De Lorenzi E, Sulitzky C, Caccialanza G, Sellergren B (2003) Electrophoresis 24:952
Quaglia M, Sellergren B, De Lorenzi E (2004) J Chromatogr A 1044:53
Adbo K, Nicholls IA (2001) Anal Chim Acta 435:115
Ye L, Ramstroem O, Mosbach K (1998) Anal Chem 70:2789
Yu Y, Ye L, De Biasi V, Mosbach K (2002) Biotech Bioeng 79:23
Yoshimatsu K, Reimhult K, Krozer A, Mosbach K, Sode K, Ye L (2007) Anal Chim Acta 584:112
Haupt K (2003) Chem Commun 2003:171
Haupt K, Mosbach K (2000) Chem Rev 100:2495
Haupt K (2001) Analyst 126:747
Wulff G (2002) Chem Rev 102:12002
Emgenbroich M, Wulff G (2003) Chem Eur J 9:4106
Alexander C, Davidson L, Hayes W (2003) Tetrahedron 2025:2003
Whitcombe MJ, Alexander C, Vulfson EN (2000) Synlett 6:911
Suedee R, Srichana T, Rattananont T (2002) Drug Delivery 9:19
Suedee R, Srichana T, Chotivatesin R, Martin GP (2002) Drug Dev Ind Pharm 28:545
Cunliffe D, Kirby A, Alexander C (2005) Ad Drug Del Rev 57:1836
Mayes AG, Whitcombe MJ (2005) Adv Drug Del Rev 57:1742
Wulff G, Sarhan A (1972) Angew Chem Int Ed in English 11:341
Wulff G (1995) Angewandte Chem Int Ed 34:1812
Andersson L, Sellergren B, Mosbach K (1984) Tetrahedron Lett 25:5211
Whitcombe MJ, Rodriguez ME, Villar P, Vulfson EN (1995) J Am Chem Soc 117:7105
Kempe M, Mosbach K (1995) J Chromatogr A 694:3
Ansell RJ (2005) Adv Drug Del Rev 57:1809
Kempe M, Fischer L, Mosbach K (1993) J Mol Recognit 6:25
Kempe M (1996) Anal Chem 68:1948
Ramstroem O, Nicholls IA, Mosbach K (1994) Tetrahedron: Asymmetry 5:649
Matsui J, Kato T, Takeuchi T, Suzuki M, Yokoyama K, Tamiya E, Karube I (1993) Anal Chem 65:2223
Yin J, Yang G, Chen Y (2005) J Chromatogr A 1090:68
Kempe H, Kempe M (2006) Anal Chem 78:3659
Ansell RJ, Mosbach K (1997) J Chromatogr A 787:55
Kempe H, Kempe M (2004) Macromol Rapid Commun 23:315
Haginaka J, Sanbe H, Takehira H (1999) J Chromatogr A 857:117
Haginaka J, Takehira H, Hosoya K, Tanaka N (1997) Chem Lett 555:1997
Haginaka J, Kagawa C (2004) J Chromatogr B 804:19
Fu Q, Sanbe H, Kagawa C, Kunimoto KK, Haginaka J (2003) Anal Chem 75:191
Sulitzky C, Rueckert B, Hall AJ, Lanza F, Unger K, Sellergren B (2002) Macromolecules 35:79
Titirici MM, Sellergren B (2006) Chem Mater 18:1773
Kim H, Guiochon G (2005) Anal Chem 77:93
Haginaka J, Sanbe H (2001) J Chromatogr A 913:141
Mayes AG, Mosbach K (1996) Anal Chem 68:3769
Rueckert B, Hall AJ, Sellergren B (2002) J Mater Chem 12:2275
Sellergren B, Ruckert B, Hall AJ (2002) Adv Mater 14:1204
Kim H, Guiochon G (2005) J Chromatogr A 1097:84
Kim H, Guiochon G (2005) Anal Chem 77:6415
Kim H, Kaczmarski K, Guiochon G (2005) Chem Eng Sci 60F:5425
Kim H, Guiochon G (2005) Anal Chem 77:1718
Kim H, Guiochon G (2005) Anal Chem 77:2496
Kim H, Guiochon G (2005) Anal Chem 77:1708
Kim H, Kaczmarski K, Guiochon G (2006) Chem Eng Sci 61:5249
Kim H, Kaczmarski K, Guiochon G (2006) Chem Eng Sci 61:1122
Kim H, Kaczmarski K, Guiochon G (2006) J Chromatogr A 1101:136
Kim H, Guiochon G (2005) Polym Preprints 46:1178
Ellwanger A, Owens PK, Karlsson L, Bayoudh S, Cormack P, Sherrington D, Sellergren B (2000) J Chromatogr A 897:317
Phinney KW (2000) Anal Chem 72:204
Sellergren B (1989) Chirality 1:63
Francotte ER (2001) J Chromatogr A 906:379
Lindholm J, Fornstedt T (2005) J Chromatogr A 1095:50
Szabelski P, Kaczmarski K, Cavazzini A, Chen YB, Sellergren B, Guiochon G (2002) J Chromatogr A 964:99
Chen Y, Kele M, Sajonz P, Sellergren B, Guiochon G (1999) Anal Chem 71:928
Vallano PT, Remcho VT (2000) J Chromatogr A 887:125
Schweitz L, Nilsson S (2001) Tech Instrum Anal Chem 23:377
Spegel P, Schweitz L, Nilsson S (2003) Electrophoresis 24:3892
Nilsson J, Spegel P, Nilsson S (2004) J Chromatogr B 804:3
Lämmerhofer M, Svec F, Frechet JMJ, Lindner W (2000) Trends Anal Chem 19:676
Lin JM, Nakagama T, Uchiyama K, Hobo T (1996) Chromatographia 43:585
Vallano PT, Remcho VT (2000) J Chromatogr A 888:23
Schweitz L (2002) Anal Chem 74:1192
Tan ZJ, Remcho VT (1998) Electrophoresis 19:2055
Schweitz L, Andersson LI, Nilsson S (2001) Anal Chim Acta 435:43
Schweitz L, Andersson LI, Nilsson S (1998) J Chromatogr A 817:5
Schweitz L, Andersson LI, Nilsson S (1999) Chromatographia 49:S93
Schweitz L, Andersson LI, Nilsson S (1997) Anal Chem 69:1179
Schweitz L, Andersson LI, Nilsson S (1997) J Chromatogr A 792:401
Schweitz L, Spegel P, Nilsson S (2000) Analyst 125:1899
Spegel P, Schweitz L, Nilsson S (2003) Anal Chem 75:6608
Spegel P, Schweitz L, Nilsson S (2001) Electrophoresis 22:3833
Schweitz L, Andersson LI, Nilsson S (2002) Analyst 127:22
Brueggemann O, Freitag R, Whitcombe MJ, Vulfson EN J (1997) Chromatogr A 781:43
Spegel P, Nilsson S (2002) Am Lab 34:29
Ye L, Cormack PAG, Mosbach K (1999) Anal Commun 36:35
Afonso CAM, Crespo JG (2004) Angew Chem Int Ed 43:2
Dzgoev A, Haupt K (1999) Chirality 11:465
Donato L, Figoli A, Drioli E (2005) J Pharm Biomed Anal 37:1003
Yoshikawa M (2002) Bioseparation 10:277
Yoshikawa M, Fujisawa T, Izumi J (1999) Macromol Chem Physic 200:1458
Yoshikawa M, Ooi T, Izumi JI (1999) J Appl Polym Sci 72:493
Yoshikawa M, Izumi JI, Ooi T, Kitao T, Guiver MD, Robertson GP (1998) Polym Bull 40:517
Kondo Y, Yoshikawa M, Okushita H (2000) Polym Bull 44:517
Reddy PS, Kobayashi T, Fujii N (1999) Chem Lett 293
Schoemaker HE, Mink D, Wubbolts MG (2002) Science 299:1694
Breuer M, Ditrich K, Habicher T, Hauer B, Kesseler M, Stuermer R, Zelinski T (2004) Angew Chem Int Ed 43:788
Shaw NM, Robins KT, Kiener A (2004) Asymmetric Catalysis on Industrial Scale 105
Ikunaka M (2004) Catal Today 96:93
Patel RN (2001) Curr Opin Biotechnol 12:587
Hawkins JM, Watson TJN (2004) Angew Chem Int Ed 43:3224
Reetz MT (2001) Angew Chem Int Ed 40:284
Reetz MT (2002) Angew Chem Int Ed 41:1335
Alexeeva M, Carr R, Turner NJ (2003) Org Biomol Chem 1:4133
Reetz MT (2003) Methods Mol Biol 230:259
Reetz MT (2004) Enzyme Functionality 559
Finn MG (2002) Chirality 14:534
Zhu L, Shabbir SH, Anslyn EV (2006) Chem Eur J 13:99
Li ZB, Lin J, Qin YC, Pu L (2005) Org Lett 7:3441
Wolf C, Liu S, Reinhardt BC (2006) Chem Commun 40:4242
Dickert FL, Forth P, Lieberzeit PA, Voigt G (2000) Fresenius J Anal Chem 366:802
Dickert FL, Hayden O, Lieberzeit P, Haderspoeck C, Bindeus R, Palfinger C, Wirl B (2003) Synth Metals 138:65
Dickert FL, Lieberzeit P, Miarecka SG, Mann KJ, Hayden O, Palfinger C (2004) Biosens Bioelectron 20:1040
Dickert FL, Hayden O, Bindeus R, Mann KJ, Blaas D, Waigmann E (2004) Anal Bioanal Chem 378:1929
Dickert FL, Lieberzeit P, Miarecka SG, Mann KJ, Hayden O, Palfinger C (2004) Biosens Bioelectron 20:1040
Marx KA (2003) Biomacromol 4:1099
Haupt K, Noworyta K, Kutner W (1999) Anal Commun 36:391
Percival CJ, Stanley S, Galle M, Braithwaite A, Newton MI, McHale G, Hayes W (2001) Anal Chem 73:4225
Stanley S, Percival CJ, Morel T, Braithwaite A, Newton MI, McHale G, Hayes W (2003) Sens Actuators B 89:103
Piacham T, Josell A, Arwin H, Prachayasittikul V, Ye L (2005) Anal Chim Acta 536:191
Nopper D, Lammershop O, Wulff G, Gauglitz G (2003) Anal Bioanal Chem 377:608
Liao Y, Wang W, Wang B (1999) Bioorg Chem 27:463
Lee YJ, Braun PV (2003) Adv Mater 15:563
Hu X, An Q, Li G, Tao S, Liu J (2006) Angew Chem Int Ed 45:8145
Lahav M, Kharitonov AB, Willner I (2001) Chem Eur J 7:3992
Zhou Y, Yu B, Levon K (2003) Chem Mater 15:2774
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Maier, N.M., Lindner, W. Chiral recognition applications of molecularly imprinted polymers: a critical review. Anal Bioanal Chem 389, 377–397 (2007). https://doi.org/10.1007/s00216-007-1427-4
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00216-007-1427-4