Role of Counterions in Molecularly Imprinted Polymers for Anionic

Subscriber access provided by University of Winnipeg Library

New Concepts at the Interface: Novel Viewpoints and Interpretations, Theory and Computations

On the Role of Counterions in Molecularly Imprinted Polymers for Anionic Species Sabine Wagner, Carlos Zapata, Wei Wan, Kornelia Gawlitza, Marcus Weber, and Knut Rurack Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.8b00500 • Publication Date (Web): 23 May 2018 Downloaded from http://pubs.acs.org on May 24, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

On the Role of Counterions in Molecularly Imprinted Polymers for Anionic Species Sabine Wagner,†,§ Carlos Zapata,†,‡,#,§ Wei Wan,† Kornelia Gawlitza,† Marcus Weber,‡ and Knut Rurack†,* † Chemical and Optical Sensing Division (1.9), Bundesanstalt für Materialforschung und – prüfung (BAM), Richard-Willstätter-Str. 11, D-12489 Berlin, Germany ‡ Computational Molecular Design Group, Department of Numerical Mathematics, Zuse Institute Berlin, Takustrasse 7, D-14195 Berlin, Germany # School of Analytical Sciences Adlershof (SALSA), Humboldt-Universität zu Berlin, Unter den Linden 6, D-10099 Berlin, Germany

KEYWORDS. Anion receptors; Fluorescence sensing; Molecular dynamics simulations; Molecularly imprinted polymers; Rational design

ACS Paragon Plus Environment

1

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 45

ABSTRACT. Small-molecule oxoanions are often imprinted noncovalently as carboxylates into molecularly imprinted polymers (MIPs), requiring the use of an organic counterion. Popular species are either pentamethylpiperidine (PMP) as a protonatable cation or tetraalkylammonium (TXA) ions as permanent cations. The present work explores the influence of the TXA as a function of their alkyl chain length, from methyl to octyl, using UV/vis absorption, fluorescence titrations and HPLC, as well as MD simulations. Protected phenylalanines (Z-L/D-Phe) served as templates/analytes. While the influence of the counterion on the complex stability constants and anion-induced spectral changes shows a monotonous trend with increasing alkyl chain length, already at the prepolymerization stage, the cross-imprinting/rebinding studies showed a unique pattern that suggested the presence of adaptive cavities in the MIP matrix, related to the concept of induced-fit of enzyme-substrate interaction. Larger cavities formed in the presence of larger counterions can take up pairs of Z-X-Phe and smaller TXA, eventually escaping spectroscopic detection. Correlation of the experimental data with the MD simulations revealed that counterion mobility, the relative distances between the three partners, and the hydrogen bond lifetimes are more decisive for the response features observed than actual distances between interacting atoms in a complex or the orientation of binding moieties. TBA has been found to yield the highest imprinting factor, also showing a unique dual behavior regarding the interaction with template and fluorescent monomer. Finally, interesting differences between both enantiomers have been observed in both theory and experiment, suggesting true control of enantioselectivity. The contribution concludes with suggestions for translating the findings into actual MIP development.


2

Page 3 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Introduction Anion sensing has received strongly increasing attention over the past three decades, because anions are of paramount importance in all biological systems, most of all in the form of nucleotides and amino acids.1-3 Furthermore, important ingredients of industrial products such as pesticides and fertilizers are also often potentially anionic in nature.4 The detection and quantification of anionic species is thus an essential task in many areas ranging from medical diagnostics to environmental monitoring.5, 6 However, the detection of anions is comparatively difficult, because of their significantly lower charge density compared to isoelectronic cations and their rather complex shapes, often precluding strong electrostatic interactions.7 The structural variety of anions also greatly increases the difficulty in receptor design. Especially for synthetic molecular receptors, advanced synthesis knowledge and skills are often required to obtain binders with appropriate selectivity. Biological receptors such as antibodies on the other hand generally possess a high affinity to the respective anionic species, yet the high costs and chemical/thermal instabilities of these materials limit their application range.8, 9 Recently, molecularly imprinted polymers (MIPs) have attracted considerable attention in sensing applications as artificial receptors.10-13 MIPs are prepared in a target-assisted approach in which the monomer and crosslinker are first preassembled with the analyte molecule of interest, the so-called template, to form a stable complex in solution. A solvent or porogen appropriate for the polymerization reaction is used to dissolve or disperse the components and constitutes the pore-forming element during the polymerization step. After polymerization and template removal, for instance by extraction, a comparatively rigid and chemically robust porous material is obtained which retains cavities that can rebind the target molecule in the analytical reaction.14 The abundance of commercially available and custom-designed monomers enables the


3

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 45

preparation of a vast number of MIPs for targeting a wide range of species. In addition, MIPs can also be synthesized directly on a support material so that no further immobilization step for a receptor becomes necessary. Especially for applications in which MIPs do not only have to retain a certain species for separation or enrichment purposes but are endowed with a second function to for instance signal its presence, commercially available monomers are often not sufficiently strong binders and design of new monomers is required. For example, functional urea-based monomers are promising candidates for MIPs targeting oxoanionic species.15-19 Their high selectivity stems from the fact that urea can provide two directional hydrogen bonds through its NH groups toward the two oxygen atoms of the carboxylate group. Ghosh and Boiocchi recently reported that the binding affinity between a urea and an oxoanion depends on the charge density on the oxygen of the anion and its basicity as well as the acidity of the urea NHs.20, 21 However, MIPs are often prepared in organic solvents in which ionic species are not well solvated. This in turn means that also the type, electronic and chemical nature of the counterion might become decisive for successful receptor preparation. Surprisingly, very little research has been undertaken to elucidate the role of counterions in the preparation of MIPs toward anions. Here, we will explore the effects of counterions in MIP synthesis by both, experimental and theoretical means, and correlate these findings to the results obtained in analytical rebinding studies.

Results and discussion Scheme 1 gives an overview of the key components and synthetic steps of the system studied here. First, we opted for MIPs that contain a fluorescent probe monomer as the designated binding partner. Such fluorescent MIPs are not only very attractive sensor materials, but the


4

Page 5 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

functional monomer reports on its status (bound or not bound, microenvironmental polarity as well as acidity/basicity) at any time during the prepolymerization, polymerization and rebinding steps simply through the band maxima in absorption and/or fluorescence and its fluorescence yield.22,

23

Second, we investigated a format that recently proved to possess superior sensing

performance in the enantioselective recognition of amino acids, that is, a few-nanometer thin MIP layer grafted from silica core particles. The grafting is achieved by RAFT (reversible addition-fragmentation chain transfer) polymerization of a fluorescent monomer (Nnitrobenzoxadiazole-N′-ethylmethacrylate-functionalized

urea

1),

a

skeletal

comonomer

(benzylmethacrylate, BMA) and a crosslinker (ethylene glycol dimethacrylate, EDMA) from the surface of aminopropyltriethoxysilane (APTES)- and RAFT agent-functionalized silica microparticles.18 Z-D-Phenylalanine (Z-D-Phe) was chosen as the template/analyte here because it is a suitable model amino acid and because these investigations allowed to countercheck and expand previous results obtained for the L-analogue.18, 24, 25 The cations investigated in this study were tetramethyl- (TMA), tetraethyl- (TEA), tetrabutyl- (TBA), tetrahexyl- (THA) and tetraoctylammonium (TOA; the latter three as linear n-alkyl variants) as permanent cations and pentamethylpiperidine (PMP) as protonatable cation. These cations are by far the most frequently used counterions when MIPs are developed for carboxylate-containing smallmolecule analytes. Binding studies of 1 and Z-D-Phe with different counterions. When carefully screening the literature, the use of counterions during the characterization of the (isolated) host–guest pairs, as well as during the stages of choosing components for a MIP, MIP preparation and rebinding studies, is not always consistent. For instance, PMP was used frequently for the in situ deprotonation of acidic templates to imprint the corresponding anions, yet the host–guest


5

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 45

features, that is, the complex stability constant of template and functional urea monomer, were reported for the TBA salt, respectively.26, 27 Memory effects for the counterion have also been observed accordingly.27 Thus, the counterion seems to have a more important role than simply acting as a cation that guarantees solubility of the carboxylate salt and maintains charge neutrality during polymer preparation. Moreover, when developing MIPs for optical sensing applications one has to rely on as strong as possible interactions between template and functional monomer, not only for efficient binding but also for achieving the best possible spectroscopic response. The detailed study reported here thus lay the focus on both, the influence of the counterion on the formation of the preassembled complex in the prepolymerization mixture, and the rebinding performance of the resulting crosslinked polymer matrix.

Scheme 1. Synthesis of RAFT agent-coated SiO2 particles (350 nm) and silica core/MIP shell nanoparticles using TMA, TEA, TBA, THA and TOA as counterions (TXA). Monomer 1 was synthesized as described previously.18 The urea moiety endows the monomer with a considerably high affinity for carboxylates. In addition, the directly fused


6

Page 7 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

nitrobenzoxadiazole (NBD) chromophore is sensitive to the change of the electron density on the urea’s NH group attached to the NBD’s 4-position and can thus respond to the binding of a carboxylate through two directional H-bonds by a modulation of the NBD’s absorption and fluorescence. The methacrylate group helps to embed the complex covalently into the polymer matrix. UV/vis absorption and fluorescence titrations were used to assess the binding affinity between anionic guest and fluorescent host in dilute solution. Due to the rather acidic character of the NBD urea, anions with a high charge density or high partial charges on the oxygens can deprotonate the monomer, especially in organic solvents. Usually, only weak bases form Hbonds with the urea group of 1 and induce a bathochromic shift in absorption and an increase in fluorescence as a result of the reinforcement of an intramolecular charge transfer (ICT) process that is operative within the chromophore upon optical excitation. The influence of the different counterions on these features is discussed in the following. Interactions in dilute solution. The use of PMP to deprotonate Z-D-Phe, and hence HPMP+ to act as the counterion, led to a red-shift of the absorption spectra of 1 by ca. 20 nm in CHCl3 (Figure 1). The fluorescence increased accordingly, as expected for NBD dyes with reinforced ICT character.28 Since PMP itself does not have an influence on the spectroscopic properties of 1, the spectral changes are attributed to the formation of the 1:Z-D-Phe/HPMP ternary complex or ensemble. A complex stability constant of logK = 2.55±0.01 was determined through fitting the titration results with the software HypSpec, and only two UV/vis-spectroscopically active species were found to be involved in the whole titration process, free 1 and 1:Z-D-Phe/HPMP (Figure 1).


7

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 45

In a second step, monomer 1 was titrated with the different tetraalkylammonium (TXA) salts of Z-D-Phe, varying from TMA to TOA. Also for these counterions, strong fluorescence enhancement was observed upon addition of the templates as shown exemplarily in Figure 2.

Figure 1. Species spectra (top) and species distribution during the titration of 1 (5 µM) with Z-DPhe/HPMP in CHCl3 (bottom); free 1 = black/squares, H-bonded complex = red/circles.


8

Page 9 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 2. Fluorescence spectra of 1 (5 µM) in CHCl3 in the absence (red spectrum) and presence of 0–250 µM Z-D-Phe/TBA (increasing concentrations in grey, endpoint spectrum in blue); note that the decrease in fluorescence in the high concentration range is due to the deprotonation effect described in the text and illustrated in Figure 3. Figure 3 further reveals that addition of Z-D-Phe/TXA leads to a bathochromic shift of the absorption maximum of 1 from 405 to ca. 440 nm with a concomitant appearance of a new, redshifted band at 510 nm. Based on work from others and ours, the red-shifted species is ascribed to the deprotonated form 1–.17, 18, 20, 21 Since the urea unit is electronically conjugated to the NBD chromophore and forms the electron donating part of a donor–acceptor π system, an increase of the electron richness at the urea by deprotonation of one NH leads to an increase in donor strength and hence to a strong bathochromic shift. The absence of well-defined isosbestic points on the other hand suggests that not only a single acid-base reaction is taking place but that most likely two competing reactions are involved: hydrogen bond-mediated complexation and deprotonation. Comparing all the different TXA counterions, the deprotonation tendency increases from Z-D-Phe/TMA to Z-D-Phe/TOA. Especially in case of Z-D-Phe/THA and Z-DPhe/TOA, deprotonation starts already at considerably low analyte concentrations. Data analysis with the HypSpec software supports our phenomenological interpretation, that is, three species are required to obtain a good fit. Thus, during titration of 1 with various salts, the free monomer is first converted into the H-bonded complex with an absorption maximum at 440 nm and then into the anion 1– (λabs = 510 nm). The logK data derived from the fit for 1:Z-D-Phe complex formation are given in Table 1.


9

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 45

Figure 3. Absorption titration spectra of 1 (5 µM) with Z-D-Phe/TBA in CHCl3 (top, c = 2–500 µM), species spectra (middle) and species distribution during the titration (bottom; free 1 in red, H-bonded complex in black and 1– in blue).


10

Page 11 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Table 1. Complex stability constants and absorption spectral shifts of 1:Z-D-Phe complexes with HPMP+ and different TXA+ counterions in CHCl3 as derived from HypSpec fits. Counterion

logK a

b ൫∆ߣ௔௕௦ ૚:௉௛௘ି૚ ൯ௗ /nm

b ൫∆ߣ௔௕௦ ૚:௉௛௘ି૚ ൯௣௣ /nm

HPMP+

2.55 ± 0.01

21

–

TMA+

4.32 ± 0.03

29

12

TEA+

4.43 ± 0.04

34

14

TBA+

4.75 ± 0.02

35

16

THA+

4.92 ± 0.01

38

18

TOA+

5.06 ± 0.03

40

20

a: Obtained at low (µM) concentrations. b: Difference between absorption maxima of 1:Z-DPhe and 1, at low (d) and prepolymerization (pp) concentrations.

Comparison of these data with those of PMP/Z-D-Phe mentioned above suggests that in contrast to several published works, PMP is not as effective as the TXA counterions for Z-D-Phe imprinting because of the lower affinity (that is, smaller logK) of the salt Z-D-Phe/HPMP to 1 in CHCl3. This is also reflected by the distinctly smaller template-induced absorption shift in the case of PMP. Apparently, HPMP+ resides in closer vicinity to the anion center, possibly due to stronger electrostatic interactions. Such behavior would result in competition between the two hydrogen bonds pointing from the urea NH to the carboxylate oxygens and the single hydrogen bond that can be formed between the proton on the HPMP+ and the carboxylate, leading to an overall decreased basicity and a weaker fluorescence response. Interactions under prepolymerization conditions. After the binding behavior of the key partners, the fluorescent monomer and the template, has been established in a suitable solvent, the next step is to verify that these favorable signaling features are retained under conditions that


11

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 45

are required for polymerization. For successful imprinting, high affinity between the functional monomer and the template is necessary. In addition, complex stability constants logK higher than 3 are preferred to reach stoichiometric noncovalent imprinting whereby over 95% of template is complexed in the prepolymerization mixture.16, 29, 30 Stronger ion-pairing and enhanced ternary complex formation in the prepolymerization mixture might in turn have an adverse effect on designated cavity formation during the polymerization step. Moreover, since the features reported for low concentrations in Table 1 are retained at higher concentrations under prepolymerization conditions (Table 1; see also Figure 4), though to a lesser extent, we set out to prepare a series of MIPs and studied in detail their rebinding behavior as a function of the chain length on the ammonium-N of the TXA cations. It should be kept in mind that prepolymerization conditions do not only mean higher concentrations but also the presence of an excess of comonomer and crosslinker. In our case, the latter are more polar than CHCl3 and can thus also influence the interaction between fluorescent monomer and template and hence the shift in absorption. With respect to optimum imprinting it is additionally important to avoid any nondirectional electrostatic interactions due to species’ deprotonation, because non-directional interactions can lead to a higher heterogeneity of the cavities.


12

Page 13 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 4. Absorption spectra of 1 (1 mM) in CHCl3 (black), 1 with various Z-L-Phe/TXA in CHCl3 and in the presence of BMA and EDMA at prepolymerization conditions; with TXA = TMA (red), TEA (blue), TBA (orange), THA (magenta) and TOA (green).

Preparation and performance of the MIP particles. MIP shells on silica particles were prepared by RAFT polymerization employing the prepolymerization mixtures containing 1 and the various Z-D-Phe/TXA salts as mentioned above. The coating of the surface with the RAFT agent ensures the formation of a ca. 5 nm thin and homogeneous MIP film on the core particles (Figure S1). Noncovalent complex imprinting is mediated by EDMA as the crosslinker. Finally, Z-D-Phe/TXA is removed from the MIP network by washing with acidic methanol, leaving specific cavities with binding sites in the shell that are complementary to the template molecule. Before analyzing the effect of the different TXA counterions in detail, it is important to assess the overall performance of the system. In the present case of imprinting amino acids, the most important figures of merit are the (enantio)selectivity, that is, the discrimination of the core-shell (CS) sensor particles MIP-Z-D-Phe/TXA@SiO2 between Z-D-Phe and its enantiomer Z-L-Phe, and the selectivity regarding chemically competing amino acids such as Z-D-Glu, possessing two carboxylate moieties, or Z-D-Tyr, able to undergo π-stacking interactions with the comonomer


13

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 45

BMA. Rebinding titrations were thus performed with Z-D-Phe/TXA, Z-L-Phe/TXA, Z-DGlu/TXA and Z-D-Tyr/TXA. Figure 5 shows a representative binding behavior of Z-D-Phe to MIP-Z-D-Phe/TBA@SiO2 and NIP@SiO2 nanoparticles using TBA as counterion, and the corresponding cross-sensitivity studies with Z-L-Phe/TBA, Z-D-Glu/TBA and Z-D-Tyr/TBA. NIP@SiO2 are the nonimprinted control particles prepared according to an identical protocol only in the absence of the template, that is, Z-D-Phe and its corresponding counterion TXA. According to Figure 5, MIP-Z-D-Phe/TBA@SiO2 nanoparticles obviously show a higher sensing response to Z-D-Phe/TBA compared to the enantiomer and the other amino acids. This favorable selectivity is attributed to the formation of rather well-matching cavities during the imprinting process and the retention of the directed hydrogen bonding between the carboxylate group of the analyte and the NH groups of the urea. Thus, the enantioselectivity factor EF = dFZ/dFZ-L-Phe was determined to 5.6. Apparently, Z-L-Phe does not fit well into the specific

D-Phe

cavity due to its different stereoisomeric structure compared to Z-D-Phe. The MIP showed also a good discrimination against Z-D-Glu/TBA and Z-D-Tyr/TBA with a discrimination factor of 3.7 for both analytes. The NIP@SiO2 particles show only a very weak response, stressing the fact that non-specific binding is negligible in this case. The difference in selectivity between MIPs and NIPs results in a high imprinting factor of 9.6. Compared to our previous study with this system,18 the imprinting effect could be further enhanced due to the thinner MIP shell and a presumable reduction of heterogeneity among the cavities.


14

Page 15 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 5. Sensing response of MIP-Z-D-Phe/TBA@SiO2 (1 mg mL–1) in CHCl3, toward template (black squares), corresponding enantiomer Z-L-Phe (red circles) and two other selected amino acids Z-D-Tyr (blue down-triangle) and Z-D-Glu (orange diamonds) and response of NIP@SiO2 (1 mg mL–1) toward template (green up-triangles); the response of NIP@SiO2 to all four species is identical within uncertainty; measurement uncertainties as indicated for selected data points. dF/F0 = (F–F0)/F0 is the normalized fluorescence enhancement. Influence of the counter-ion during imprinting. Having established the favorable performance and repeatability of our approach, we proceeded to investigate whether the counterion effect seen in the prepolymerization mixture can also influence the imprinting process during the MIP synthesis. We studied the impact of the counterion on carboxylate imprinting, that is, in how far the use of the different TXA salts governs the MIP’s selectivity in the analytical reaction. Figure 6 reveals that the MIPs which were prepared with smaller counterions like TMA showed lower sensing responses after titration with the corresponding Z-D-Phe/TXA stock solutions.i The sensing response increases from TMA to TOA, reflecting well the order of logK

i

To avoid confusion, MIP-Z-D-Phe/TMA@SiO2 was titrated with Z-D-Phe/TMA, MIP-Z-DPhe/TEA@SiO2 was titrated with Z-D-Phe/TEA, etc.


15

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 45

values in Table 1. The weaker fluorescence enhancement found in the presence of TMA suggests that also in the MIP TMA approaches closer to the carboxylate center, leading to weaker interaction with the NBD’s urea binding site as was discussed before. In contrast, the larger counterions seem to stay farther away from the 1:Z-D-Phe complexes also in the cavities (vide infra), entailing a more pronounced spectroscopic response due to stronger H-bonding between 1 and Z-D-Phe (cf. Table 1). Analysis of the curves in Figure 6 yielded complex stability constants for the MIP particles as collected in Table 2. The data show that the logK of the MIP particles increases from TMA to TOA, consistent with the trend of the logK in Table 1. The data in Figure 6 immediately raise the question whether only the TXA-modulated interaction between carboxylate and urea is decisive or whether the cavities as such have a major influence. We asked ourselves whether in all cases Z-D-Phe/TXA is imprinted or whether for the more lipophilic THA and TOA, mainly Z-D-Phe is imprinted yet the counterions are sitting somewhere close in the polymer network (for electroneutrality), but not necessarily directly in the cavity.

Figure 6. Sensing response of the various MIP-Z-D-Phe/TXA@SiO2 CS particles (1 mg mL–1) in CHCl3 with X = M (red squares), E (orange circles), B (black up-triangles), H (blue downtriangles) and O (green diamonds) toward the respective template Z-D-Phe/TXA used for imprinting, and representative NIP@SiO2 with Z-D-Phe/TBA (magenta left-triangle);


16

Page 17 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

measurement uncertainties as indicated for selected data points. dF/F0 = (F–F0)/F0 is the normalized fluorescence enhancement. Table 2. Complex stability constants for MIP-Z-D-Phe/TXA@SiO2 and corresponding Z-DPhe/TXA in CHCl3; the counterion used in imprinting and for rebinding was always identical. Counteriona

logK

TMA+

4.10 ± 0.01

TEA+

4.98 ± 0.02

TBA+

5.16 ± 0.02

THA+

5.39 ± 0.01

TOA+

5.71 ± 0.01

We thus performed another series of experiments by titrating each MIP-Z-D-Phe/TXiA@SiO2 CS particle solution with the different tetraalkylammonium salts TXrA of Z-D-Phe to screen for possible recognition patterns among the MIP sensor particles. The results of these titration experiments are collected in Figure 7. The different subscripts of TXiA and TXrA refer to the use of the different counterions during the imprinting (TXiA) and the rebinding titration (TXrA).


17

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 45

Figure 7. Sensing response of each MIP-Z-D-Phe/TXiA@SiO2 (1 mg mL–1) in CHCl3 toward each tetraalkylammonium salt of Z-D-Phe/TXrA (at 0.048 mM).

Figure 7 allows to draw the following conclusions. - The nature of the counterion of the analyte Xr is least decisive when the smallest Xi = M is used. - The nature of the counterion of the analyte Xr is most decisive when the largest Xi = O is used. - In all cases of TXiA, the highest response is found when a TOA salt is used as TXrA. - Counterion selectivity is only found for the pair MIP-Z-D-Phe/TOiA@SiO2 and Z-DPhe/TOrA. - In all cases, a good response is found for all TXrA with a size that is larger than or equal to that of the TXiA used. Considering the system of moderately polar template/analyte, fairly polar fluorescent monomer, slightly polar polymer matrix and counterions ranging from polar (TMA) to largely nonpolar (TOA), the following two embedding scenarios seem intuitionally possible (Scheme 2). First, regardless of the polarity of the counterion and because of electrostatic attraction, the cavity is always formed around both, template and counterion (scenario A). Second, because TBA, THA and especially TOA are rather lipophilic and possess a low charge density, these counterions are not necessarily co-embedded into the cavity but can reside in the porous polymer network close by (scenario B). In case A, the size of the cavity should increase monotonically as Xi is changed from M over E, B and H to O. For scenario B, the cavity size first increases for the more tightly bound and polar M and E but then levels off for the less polar counterions when


18

Page 19 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

these are located in the network nearby. For B, H and O, the size of the cavity should then match rather closely that of Z-D-Phe alone. In the next section, we discuss in detail the responses to be expected in these two cases.

Scheme 2. Graphical representation of the two imprinting scenarios A (top) and B (bottom) described in the text with increasing size of counterion from left (TMA) to right (TOA). Crosssection of the cavities with the size of all the relevant partners being roughly proportional. Grey balls denote the TXA counterions, template in blue and fluorescent monomer in green. Response expected for scenario A. For Z-D-Phe/TMiA@SiO2, Z-D-Phe/TMrA should reoccupy the cavity best. All the other Z-D-Phe/TXrA should bind to a (progressively) lesser extent or not at all. For Z-D-Phe/TEiA@SiO2, Z-D-Phe/TErA should bind best, Z-D-Phe/TMrA to a lesser degree and Z-D-Phe/TBrA, Z-D-Phe/THrA and Z-D-Phe/TOrA even less or not at all, etc. A response pattern such as in Figure S2 would be expected. Response expected for scenario B. In this scenario, the cavity size would only increase from TMA to TEA because these counterions bind tightly enough to 1:Z-D-Phe so that the ternary complexes are only fully imprinted for Xi = M, E. For TXiA with Xi = B, H and O, the more lipophilic counterions retreat into the porous polymer network already during MIP formation and


19

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 45

the cavities formed in all these cases largely reflect the size of Z-D-Phe alone. Upon rebinding to Z-D-Phe/TMiA@SiO2 and Z-D-Phe/TEiA@SiO2, Z-D-Phe/TMA and Z-D-Phe/TEA should bind best but for the larger TXiA, the influence of the counterion should be much less prominent once the ability of Xr for ternary complex formation is low; a response as in Figure S3 would be expected. Expected vs. observed responses. According to Figure 7, all the data on the right side of the diagonal described by the data points TXiA = TXrA indicate progressively stronger binding when Xr > Xi. A pure size effect of a (considerably rigid) cavity as expected for scenario A thus does not seem to be the key decisive factor. This part of Figure 7 suggests that scenario B might be effective, that is, that the larger TXrA could penetrate the network and responses would be generally high. However, scenario B does not account for the left side of the line TXiA = TXrA in Figure 7, i.e., the weak response for essentially all cases of Xr < Xi (Figure 7 vs. Figure S3). Ternary complex imprinting into a flexible cavity in a porous network (alternative scenario C). If we assume that the cavity is not as rigid as it might intuitively seem from the point of view of imprinting, enantioselectivity or discrimination factors, but rather flexible, the Xr > Xi part of Figure 7 can be understood by analogy with a very well-known biological behavior, the induced fit of enzyme–substrate complementarity.31 If we assume that cavities in MIPs are embedded into a porous network—like here, guaranteeing macroscopic equilibration and response times on the order of seconds—and are flexible, the increasing size of the counterion might progressively be accommodated by an adaptive cavity. Since the counterion’s increase in size correlates with an increase in spectroscopic response (Figure 6), the gradually and moderately increasing data bars on the right side of the diagonal in Figure 7 seem to match well with this model. However, the distinctly weaker response on the left side can still not fully be explained by fit-induced cavities


20

Page 21 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

because an adaptive cavity should also accommodate ternary complexes with smaller counterions (at least in a progressive manner, see left side of Figure S2). The situation yet changes if we consider that Figure 7 shows the spectroscopic response of the system, that is, the fit into the cavity plus the designated hydrogen bonding interaction between monomer and template. A weak signal in case of a smaller counterion can have two reasons. First, Z-D-Phe and counterion are virtually not bound. Second, the two species reside in the cavity, yet the cavity’s rather large size and flexibility allows for more efficient ion pair formation than interaction between template and fluorescent monomer or ternary complex formation (Scheme 3B). Uptake would thus happen in all cases, yet would largely escape spectroscopic detection for TXrA < TXiA. This scenario would also be supported by the previously determined complex stability constants of the MIP particles.

Scheme 3. A) Case of only Z-D-Phe fitting tightly into a cavity with counterion residing in the network (identical to last scenario of Scheme 2B). B) Case of cavity being too large for Z-DPhe/counterion pair so that the attractive forces exerted on Z-D-Phe/TXrA are not strong enough to dissociate the ion pair; Z-D-Phe is thus largely unable to H-bond to the urea moiety. Control experiments by HPLC. While fluorescence spectroscopy can report only on that fraction of templates that is interacting in a designated fashion with fluorescent monomer, retention experiments by HPLC provide us with the total amount of template that is bound in a MIP, the overall binding capacity of the cavities. If uptake happens also for ternary complexes


21

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 45

with TXrA < TXiA, but remains spectroscopically silent, this should be seen in the HPLC experiments. Applying the same conditions as for the spectroscopic titration experiments, a solution of 1 mg mL–1 MIP sensor particles containing 0.048 mM of Z-D-Phe/TXrA was prepared and equilibrated for 14 min. After centrifugation, the concentration of Z-D-Phe/TXrA in the supernatant was quantified by HPLC measurements. Back calculation allowed the determination of the adsorbed amount of analyte independent of the analyte’s influence on the spectroscopic properties. The calculated concentrations for the MIPs prepared with TMiA, TBiA and TOiA are presented in a 3D diagram (Figure 8). It is obvious that for Z-D-Phe/TMiA@SiO2 a somewhat higher amount of Z-D-Phe/TBrA compared to Z-D-Phe/TMrA is bound to the MIP, and even a slightly higher amount of Z-D-Phe/TOrA is retained. The mechanism of fit-induced cavity formation seems to allow the ternary complex with TBrA to be efficiently retained. Even the much bigger TOrA does not impede retention when the smallest counterion is used during imprinting. In addition, considering the magnitude of the spectroscopic responses (e.g., Figure 6 for all the optimal pairs), less retained Z-D-Phe/TOrA can induce a higher spectroscopic response than better retained Z-D-Phe/TBrA, and especially than strongly interacting TMrA, arriving at an order as in Figure 7. For Z-D-Phe/TBiA@SiO2, the HPLC data show that Z-D-Phe/TOrA can also be accommodated to a sizeable degree, though Z-D-Phe/TMrA in particular is efficiently retained. Comparison with the fluorescence response in Figure 7 then immediately suggests that indeed the binding scenario sketched in Scheme 3B seems very plausible. Finally, the data shown for Z-D-Phe/TOiA@SiO2 in Figure 8 nicely complement this mechanistic explanation. An overview of all the response patterns is collected in Figure S4 for better comparison.


22

Page 23 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 8. Quantitative analysis of Z-D-Phe/TXrA retention by MIP-Z-D-Phe/TXiA@SiO2 sensor particles (1 mg mL–1) in CHCl3 by HPLC. For better illustration, the Z axis (rebinding) was reversed, as compared to Figure 7. Enantioselectivity control studies. Returning to the fluorometric characterization of the various materials, fit-induced cavity formation is also supported by the enantioselectivity patterns of the MIP sensor particles as determined from titrations with the TXrA salts of Z-D-Phe and the corresponding

L-isomer

as the enantiomeric twin. Table 3 shows representative

enantioselectivity factors of MIP-Z-D-Phe/TEiA@SiO2, MIP-Z-D-Phe/THiA@SiO2 as well as MIP-Z-D-Phe/TOiA@SiO2, which were calculated according to EF = dFZ-D-Phe/TXiA/dFZ-LPhe/TXiA.

If a considerably small cavity is imprinted like in case of MIP-Z-D-Phe/TEiA@SiO2, the enantiomeric selectivity is largely similar regardless of TXrA when Xr ≥ E. On the contrary, in presence of the smaller TMrA, no enantioselectivity is found. The ternary complex is obviously too mobile and TMrA interacting too strongly with Z-D/L-Phe so that no preference is found for Z-D-Phe compared to Z-L-Phe. On the other hand, MIP-Z-D-Phe/THiA@SiO2 as well as MIPZ-D-Phe/TOiA@SiO2 can receive all the Z-D/L-Phe/TXrA (TXrA < THiA or TOiA) into their


23

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

considerably

large

cavities,

revealing

absence

Page 24 of 45

of

pronounced

enantioselectivity.

Enantioselectivity is only found when the counterions are big enough to closely fill the space in the adaptive cavity. In addition, this enantioselectivity is distinctly higher than the one found for the smaller counterions, suggesting that although the cavity is adaptive, it is still sufficiently small to prevent accommodation of species that have a stereoisomerically different preferential orientation.

Table 3. Enantioselectivity factors of MIP-Z-D-Phe/TXiA@SiO2 with X = E, H and O. Salts used for rebinding

EFMIP-Z-D-Phe/TEiA@SiO2

EFMIP-Z-D-Phe/THiA@SiO2

EFMIP-Z-D-Phe/TOiA@SiO2

Z-D/L-Phe/TMrA

1.1

1.4

1.1

Z-D/L-Phe/TErA

2.3

1.5

1.2

Z-D/L-Phe/TBrA

2.4

1.7

1.3

Z-D/L-Phe/THrA

2.8

5.8

1.7

Z-D/L-Phe/TOrA

2.5

5.4

5.6

Control studies with NIP@SiO2. Until now, we have only discussed the various scenarios pertaining to MIPs and TXiA/TXrA pairs. Table 4 collects the imprinting factors for all TXrA, that is, their responses when bound by the corresponding MIP-Z-D-Phe/TXiA@SiO2 compared to the responses found by binding to NIP@SiO2 (see also Figure 5 above). The IFs were calculated as IF = dFMIP/dFNIP. It should be mentioned that because the application-oriented background is sensing, we only considered here spectroscopic imprinting effects. The imprinting effect has been already introduced and discussed above for MIP-Z-DPhe/TBiA@SiO2 (Figure 5), which showed a favorable IF = 9.6. Table 4 reveals that MIPs


24

Page 25 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

imprinted with smaller TXiA gave lower IF values, most likely due to the stronger interaction with the anion and, therefore, reduced response. The decrease of the imprinting effect when considering the more lipophilic counterions from TBiA to TOiA is then tentatively attributed to a higher nonspecific binding in the moderately low-polar organic NIP network or on its surface.

Table 4. Imprinting factors for Z-D-Phe/TXrA and MIP-Z-D-Phe/TXiA@SiO2 (Xr = Xi) referenced against NIP@SiO2. TXi/rA

IF

TMA

4.1

TEA

4.8

TBA

9.6

THA

8.9

TOA

6.2

Molecular dynamics simulations of the prepolymerization mixture. To complement the analysis of the counterion effects on the binding of a model amino acid template, we carried out molecular dynamics simulations. The search for rational design approaches to synthesize MIPs with higher performance has led to computational strategies to complement the usual chromatographic and spectroscopic techniques. Nicholls et al. and Cowen et al. presented comprehensive reviews of in silico methods in the field.32, 33 In this work, we focus on the prepolymerization state and assume as a simplification that the fluorescent monomer–template complex has freedom of movement, i.e. that it is not embedded in chains anchored to the silica core. Although doing so we exclude matrix-related effects, this analysis serves as a versatile tool to gain quantitative insights about the binding stability as well


25

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 45

as the molecular interactions and chemical affinity between different components prior to polymerization. Moreover, in line with findings by others34-36 as well as the molecular structures of our components, we did not specifically consider template–template or monomer–monomer association (see Supporting Information for more detailed considerations). The starting arrangement prior to energy minimization was set so that 1 and template were positioned by aligning the urea group in the monomer and the carboxylate in the template in a coplanar configuration (Figure S5). Next, the counterion was located manually as close to the urea-carboxylate H-bonded site as the steric restrictions allowed (in all cases no more than 5 Å away from the ammonium-N). This configuration was adopted for all simulations, allowing the hydrogen bonds between the urea and the carboxylate to be present from start. Simulations carried out in vacuo showed that when initially separated by 20 Å, the complex reaches this conformation in about 0.4 ns. The first step in the analysis was to examine the effect of the counterions on the stability of the complex. The closer the cation’s ammonium nitrogen is to the COO– group, the stronger its electrostatic attraction to the O atoms, pulling them away from the urea-Hs in a ternary complex equilibrium. With this in mind, we set out to examine a possible correlation between the distance from the center of mass of the counterion to the center of mass of the carboxylate oxygens, dTXA– O(Phe),

and the distance between the center of mass of the urea hydrogens in 1 and the center of

mass of the carboxylate oxygens, dO(Phe)–H(1). The two distances were extracted every picosecond for 40 trajectories of each template enantiomer to create the bivariate histograms in Figure 9.


26

Page 27 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 9. Correlation between distances template–counterion and monomer–template, values for runs with Z-D-Phe. The counterion mobility is governed by an interplay between electrostatic and van der Waals forces. Plot for Z-L-Phe in Figure S6. During the entire simulation time, and for all TXA, the complex remained stable with monomer and template close to each other, maintaining the H bonds between urea and carboxylate. The correlation between both distances is not direct, but a detailed inspection of Figure 9 reveals interesting features regarding the dynamics of the complex. The red areas represent regions with the highest density of observations. In all five cases, the state with the highest density corresponds to a distance template–counterion of 5 Å and a distance monomer– template of 1.8–2.0 Å. The last subplot contains the values for all the 250 runs per enantiomer system considered. In general, dO(Phe)–H(1) oscillates between 1.4 and 3.3 Å independently of how far away TXA is located.


27

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 45

The counterion itself moves near the complex and its mobility increases with increasing size of the alkyl branches, reaching a maximum with TBA and then decreasing toward TOA. The highest number of states with the center of mass of TXA away from the carboxylate occurs in the TBA case. As expected from its highly localized charge and its minimal steric hindrance, the counterion with the shortest branches, TMA, experiences the highest Coulomb attraction and thus remains anchored more strongly by the interaction with the carboxylate-Os. With bigger branches, the van der Waals interactions increase, charges on the counterion delocalize, and the electrostatic attraction to the template decreases. As the largest counterion, the mobility of TOA is decreased compared to THA. As for the template enantiomer type, systems with the

L

enantiomer exhibit, on average, states with a greater separation between carboxylate-Os and counterion-N. This separation reaches as much as 16 Å for the system Z-L-Phe/TOA, which is 7 Å farther than the maximum distance carboxylate/counterion-N for Z-D-Phe/TOA (Figure S6e vs. Figure 9e). In the TOA case, this difference is confirmed by a 5 kJ mol–1 stronger interaction energy including electrostatic and van der Waals interactions (see Figure 11 below and Table S1). Comparison of the findings with the complexation data of the TOA cases in Tables 1 and S2 suggests that indeed stereochemistry might exert subtle but distinct effects: logK and spectral shifts are higher and more pronounced for Z-L-Phe/TOA than for Z-D-Phe/TOA, confirming the greater separation of Z-L-Phe and TOA observed in simulations. Moreover, the theoretical data support well the experimental findings and their interpretation reported above in the section on the enantioselectivity control studies. A bond lifetime analysis was carried out to investigate the kinetic aspect of the H bonds formed. Each curve in Figure S7 and Figure S8 corresponds to one of the 50 trajectories run for each TXA per template enantiomer. Contacts between carboxylate and urea moieties are plotted


28

Page 29 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

as a function of the time that the contact was continuously present. A contact is counted when a carboxylate-O resides inside a cutoff of 0.38 nm around the center of mass of the urea group. It is important to highlight here that by ‘contacts’ in this analysis we are not referring to the H bonds themselves but to the extent in which the oxygens’ van der Waals volumes trespass the cutoff sphere. This contact variable is therefore continuous, so when it takes the value of 1.5, e.g., it means one oxygen is completely inside the cutoff while the other one is half inside. The 0.38 nm cutoff value was set so that it would be able to include the volume of both oxygens completely inside the cutoff during some instants of the trajectories. Longer contact times for a given degree of contact (or extent of cutoff trespassing) imply that the given H bond arrangement, in terms of distance, occurs more frequently or lasts longer. The mean value of the lifetime functions was calculated by numerical integration of each of the curves. Larger mean values indicate that oxygens remain close to the H atoms longer. Average values of the number of contacts for each counterion and enantiomer are presented in Figure 10.

Figure 10. Mean number of oxygen contacts inside 0.38 nm cutoff around the center of mass of urea hydrogens, averaged over 50 trajectories for each counterion and template enantiomer. Error bars = standard error of the mean.


29

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 45

The mean value of the contact functions increases with counterion size, more monotonically in the Z-D-Phe mixtures, indicating that oxygen contacts last longer with larger TXA. In contrast, systems with TMA display a weaker binding frequency, which makes them the least attractive counterion alternative when aiming at the formation of cavities with a narrow shape distribution and thus a higher MIP selectivity. It is worth mentioning here that although the smallest counterions exhibit the lowest mean dO(Phe)–H(1) with the

L-template,

the contact lifetime

represents a more comprehensive descriptor for the H bonds binding dynamics, since it includes the contact times and the percentage of cutoff penetration of both carboxylate oxygens. When comparing these data with the complex stability constants and complexation-induced shifts it is obvious that the simulations support the experimental findings very well in a qualitative fashion (Figure S9). The higher number of average contacts suggests that the bigger branches of THA and TOA may act like a hinge where the counterion is attracted to the template from one side of its branches and to the fluorescent monomer 1 from the other. Because THA and TOA can extend their branches fully and thereby reach other species, in particular the solvent, it is less likely that their free branches approach and disturb the H bonds, leading to more stable fluorescent complexes. To look at this more closely, we quantified the interaction energies of the counterions. Table S1 contains the mean values for the Coulomb and Lennard-Jones short-range interaction energies of the five alkyl-branched counterions with all the other species in the mixture. The strongest interaction TMA experiences takes place with the template (47% of TMA’s overall interactions), given by the highly focused electrostatic attraction. For TBA, THA and TOA on the other hand the stronger interaction occurs with the solvent (ranging from 49 to


30

Page 31 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

62%), which explains in part the higher mobility of the long-branched species. The interaction of TOA with the template accounts for only 20% of its interactions with all species; nevertheless, its value remains the highest among the five counterions (~70 kJ mol–1), considering both van der Waals and electrostatic forces (Figure 11). Additionally, the interaction counterion/monomer 1 is governed almost exclusively by van der Waals forces, increasing with the branch size.

Figure 11. Short range Coulomb plus van der Waals mean interaction energies between template and counterion. Averaged over 50 trajectories of 15 ns per counterion and per templateenantiomer. The most convenient scenario for a successful imprinting is to have monomer and template as close as possible, and counterion as far, and thus we quantified an observable d1 defined as the distance template–counterion minus the distance template–urea, taking as reference points, as before, the center of mass of the two carboxylate oxygens, the center of mass of the counterion and the center of mass of the urea hydrogens. According to the probability densities for d1 plotted in Figure 12, excluding TBA, the bigger the counterion the more likely to have states with higher values of d1 and therefore more ideal imprinting conditions.


31

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 45

Figure 12. Probability densities for d1 = dTXA–O(Phe) – dO(Phe)–H(1). Z-D-Phe (top) and Z-L-Phe (bottom). While other counterions reveal one main peak for d1, TBA shows two main peaks and a smaller one at around 0.12–0.15 nm. These three peaks correspond to metastable states and appear as three distinctive clusters with the higher population of states in Figure 9c. Being the middle-sized counterion, the length of its branches confers TBA a dual behavior. On one hand, TBA stays close without disturbing the urea H bonds (lower cluster left). On the other hand, it moves away from the complex without interfering, attracted by van der Waals interactions with other species (lower cluster right). Furthermore, since its branches are long enough, they can also more closely approach the urea and the carboxylate, disturbing to some


32

Page 33 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

extent the H bonds without moving away its center of mass (upper cluster: higher distance monomer–template and lower template–counterion). The attraction between TBA and other species is not enough to move the counterion too far. In contrast, THA and TOA have larger branches and it is sterically and energetically more difficult for them to come closer and disturb the H bonds. Besides, being able to reach other species, THA and TOA remain farther, favoring a small distance monomer–template and a high distance template–counterion (best scenario for imprinting). Concerning the smaller cations, TEA and TMA, the disturbance of their shorter branches is minimal but then the interference is dominated by the concentrated electrostatic charge which attracts strongly the template oxygens. This analysis was corroborated by a radial distribution function (RDF) analysis considering only the atoms in the counterion around the carboxylate group (Figure 13). While the average distance between the TBA nitrogen and the template oxygens is approx. 4.65 Å, a fully extended TBA branch is approx. 5.18 Å. Right after the 3.0 Å cutoff in the RDFs, the repulsive forces decay and the first counterion atoms appear. At 3.9 Å a first peak is observed, reflecting the atoms in the extremes of the branches that are closer to the 3.0 Å barrier. For TBA, this peak is considerably higher, confirming that more of the atoms in its branches tend to remain, on average, longer near the carboxylate. This peak also suggests that TBA might tend to adopt a conformation with two of its branches surrounding tangentially the 3.0 Å cutoff, which would explain well the peak at 3.9 Å. For THA, the first RDF peak at 3.9 Å loses strength and disappears completely in TOA, indicating that the branches point in different directions and adopt other conformations induced by stronger interactions with other species. The homogeneous RDFs of the TOA trajectories suggest the prevalence of fully extended conformations. Taking this into consideration, it is


33

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 45

evidenced that the bigger counterions, THA and TOA, indeed exhibit the previously mentioned hinge effect, yielding more stable hydrogen bonds between urea and carboxylate.

Figure 13. Radial distribution functions for counterion atoms around the center of mass of the carboxylate oxygens. RDF plotted for 25 trajectories of each counterion (superimposed). Z-DPhe mixtures. Finally, we wanted to examine a possible preference of the system concerning the directionality of the H bonds depending on the type of counterion. To carry this out, we tracked the alignment of the urea and the carboxylate oxygens through the RMSD of the two groups as a whole, taking as a reference the coordinates of the pre-energy-minimization state in which the two groups were positioned in a coplanar arrangement. For neither the Z-D-Phe nor for the Z-LPhe systems was any trend related to the TXA sizes observed in the RMSD (Table S3), meaning there are no bond-orientation effects due to the counterion. According to the presented analysis, MD simulation results indicate that larger counterions lead to a stronger and more stable pair of urea–carboxylate H-bonds. The van der Waals interactions exerted by TOA’s branches on the template and the monomer favor a longer-lasting proximity between urea and carboxylate. This finding corroborates the results of the spectroscopic studies and confirms that larger counterions are the best candidates for generating


34

Page 35 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

powerful MIPs among the five tetraalkylammonium species considered (as well as PMP). Contrary to this, a smaller counterion like TMA yields a weaker binding strength, a fact that is supported by the experimental spectroscopic findings of smallest shifts and least fluorescence enhancement (Table 1, Figure S10). Between the tetraethyl- and the tetrahexylammonium, binding features predicted by simulation are similar, and experimental data should be relied upon more for an optimal MIP system selection. In conclusion, the molecular effects derived from the changes in counterion size are not strong enough to alter the stability of the complex studied, but the size tuning can definitively improve the template binding with the consequent improvement of the MIP sensing response. Further attempts to step on from these molecular correlations to predicting spectroscopic properties largely failed at TXA level and was only successful for discriminating TXA against PMP (see Section III in Supporting Information).

Conclusions Cavities are significantly determining the performance of molecularly imprinted polymers. When imprinting ionic templates, a counterion is always present in the polymerization mixture, potentially influencing the final performance of a system. We have investigated here in experimental and theoretical detail the influence of the popular tetraalkylammonium permanent cations, often employed in the imprinting of small-molecule oxoanionic species. By using a fluorescent urea monomer as the primary binding partner for Z-X-Phe as template anions, studies at dilute and at prepolymerization conditions allowed us to extract two trends, namely, that complex stability constants and spectroscopic responses correlate with the chain length of the counterion on the order of methyl, ethyl, butyl, hexyl and octyl. As the corresponding MD


35

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 45

simulations have shown, simple descriptors such as H bond lengths, atomic distances within the binding moieties, or tracking whether the host–guest pair resides in an optimal coplanar conformation are not sufficient to predict these chain length effects. A deeper insight was gained by using extra parameters such as the mobility of the counterion, the relative distances between the three partners, and the H bond lifetimes. In addition, the consideration of Coulomb and van der Waals forces also reinforced the trends observed. Up to the point of the prepolymerization mixture, both theoretical descriptors and experimental complexation data would hint at the fact that the larger the counterion, the stronger and potentially more powerful the resulting MIP to be expected. However, experimental investigation of the actual MIPs revealed that the picture is not so simple. Only if an optimal ionpair imprinting is aimed at for a sensory MIP like the fluorescent MIPs currently becoming very popular, the choice of TOA would be the best. This is however not because of the best possible discrimination in analyte uptake—the analyte is also taken up in the presence of the other, smaller counterions, but because the TOA prevents least the desired directional imprinting so that a strong spectroscopic response is generated. From the point of view of a sensory application, it is thus necessary to consider the best compromise between uptake, response, suppression of nonspecific binding and the fact that an unknown sample contains counterions of unknown size and type already when choosing the salt for imprinting. Based on our present results, TBA is the counterion of choice because it combines favorable analytical figures of merit while MIPs produced with TBA tolerate larger counterions. The latter behavior was tentatively ascribed by us to the presence of adaptive cavities. The differences in enantioselective imprinting found by us both experimentally and theoretically are a very interesting outcome of this work which, however, requires more


36

Page 37 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

fundamental studies before general conclusions can be drawn. Apparently, the stereochemistry is affecting the interaction between the three partners and, as a general trend, an increasing size of the alkyl branches is reinforcing these effects. Although the average number of contacts is not a direct quantitative predictor of the behavior observed experimentally, the overall trend of the curves showing the absorption band shift and the number of contacts is relatively similar for each enantiomer. Whether this feature can in practice be used for enhanced enantioseparation in the future remains to be elucidated. When aiming at actual sensor design, the rich counterion effects found here should at best be circumvented by imprinting for instance a TBA/oxoanion template salt and, during the analytical workflow, adding an excess of a TBA salt that is not interfering with oxoanion binding so that enough counterion molecules are available for achieving the desired rebinding and signaling behavior. We have recently employed this strategy successfully.37 With respect to prediction of MIP behavior, the present simulations have added several new parameters to the existing toolbox for modeling prepolymerization mixtures. They have also shown that the prediction of spectral responses is to a certain degree possible between different classes of counterions. Given the complexity of MIP systems, more detailed solvent-explicit approaches using quantum mechanical calculations are still highly demanding, computationally speaking. Multiscale approaches appear as a promising alternative for reproducing accurately the local binding behavior at the mesoscopic scale of a polymer network/matrix. Further steps in this direction will expand the toolbox for better rational design of MIPs.


37

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 45

ASSOCIATED CONTENT Supporting Information. Experimental and computational details. AUTHOR INFORMATION Corresponding Author *[email protected] Author Contributions § These authors contributed equally.

ACKNOWLEDGMENT The work was supported by the Excellence Initiative of the German Research Foundation (DFG) within the framework of the Graduate School SALSA (School of Analytical Sciences Adlershof), BAM/BMWi’s “Menschen, Ideen, Strukturen”-Program (Ideen_2013_69) and an individual research grant of DFG (RU 1622/1-1).


38

Page 39 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

REFERENCES 1.

Bowman-James, K.; Bianchi, A.; García-España, E., Anion Coordination Chemistry. WileyVCH: Weinheim, 2011.

2.

Leitner, A.; Sturm, M.; Lindner, W., Tools for analyzing the phosphoproteome and other phosphorylated biomolecules: A review. Analytica Chimica Acta 2011, 703, 19-30.

3.

Gale, P. A.; Busschaert, N.; Haynes, C. J. E.; Karagiannidis, L. E.; Kirby, I. L., Anion receptor chemistry: highlights from 2011 and 2012. Chemical Society Reviews 2014, 43, 205-241.

4.

Mason, C. F., Biology of Freshwater Pollution. Longman Scientific & Technical: New York, 2001; p 389.

5.

Frawley Cass, S. M.; Tepe, J. J., Identification of phosphoproteins and their impact as biomarkers in cancer therapeutics. Current Signal Transduction Therapy 2011, 6, 113-140.

6.

Albelda, M. T.; Frias, J. C.; Garcia-Espana, E.; Schneider, H. J., Supramolecular complexation for environmental control. Chemical Society Reviews 2012, 41, 3859-3877.

7.

Schmidtchen, F. P.; Berger, M., Artificial organic host molecules for anions. Chemical Reviews 1997, 97, 1609-1646.

8.

Haupt, K.; Mosbach, K., Molecularly imprinted polymers and their use in biomimetic sensors. Chemical Reviews 2000, 100, 2495-2504.

9.

Baggiani, C.; Anfossi, L.; Giovannoli, C., MIP-based immunoassays: State of the Art, limitations and Perspectives. In Molecular Imprinting, 2013; Vol. 1, p 41.


39

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 45

10. Ye, L.; Haupt, K., Molecularly imprinted polymers as antibody and receptor mimics for assays, sensors and drug discovery. Analytical and Bioanalytical Chemistry 2004, 378, 1887-1897. 11. Reimhult, K.; Yoshimatsu, K.; Risveden, K.; Chen, S.; Ye, L.; Krozer, A., Characterization of QCM sensor surfaces coated with molecularly imprinted nanoparticles. Biosensors & Bioelectronics 2008, 23, 1908-1914. 12. Moreno-Bondi, M. C.; Navarro-Villoslada, F.; Benito-Pena, E.; Urraca, J. L., Molecularly imprinted polymers as selective recognition elements in optical sensing. Current Analytical Chemistry 2008, 4, 316-340. 13. Lattach, Y.; Fourati, N.; Zerrouki, C.; Fougnion, J. M.; Garnier, F.; Pernelle, C.; Remita, S., Molecularly imprinted surface acoustic wave sensors: The synergy of electrochemical and gravimetric transductions in chemical recognition processes. Electrochimica Acta 2012, 73, 36-44. 14. Whitcombe, M. J.; Vulfson, E. N., Imprinted polymers. Advanced Materials 2001, 13, 467+. 15. Manesiotis, P.; Hall, A. J.; Emgenbroich, M.; Quaglia, M.; De Lorenzi, E.; Sellergren, B., An enantioselective imprinted receptor for Z-glutamate exhibiting a binding induced color change. Chemical Communications 2004, 2278-2279. 16. Hall, A. J.; Manesiotis, P.; Emgenbroich, M.; Quaglia, M.; De Lorenzi, E.; Sellergren, B., Urea host monomers for stoichiometric molecular imprinting of oxyanions. Journal of Organic Chemistry 2005, 70, 1732-1736.


40

Page 41 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

17. Wagner, R.; Wan, W.; Biyikal, M.; Benito-Pena, E.; Moreno-Bondi, M. C.; Lazraq, I.; Rurack, K.; Sellergren, B., Synthesis, spectroscopic, and analyte-responsive behavior of a polymerizable naphthalimide-based carboxylate probe and molecularly imprinted polymers prepared thereof. Journal of Organic Chemistry 2013, 78, 1377-1389. 18. Wan, W.; Biyikal, M.; Wagner, R.; Sellergren, B.; Rurack, K., Fluorescent sensory microparticles that "light-up" consisting of a silica core and a molecularly imprinted polymer (MIP) shell. Angewandte Chemie-International Edition 2013, 52, 7023-7027. 19. Wan, W.; Descalzo, A. B.; Shinde, S.; Weisshoff, H.; Orellana, G.; Sellergren, B.; Rurack, K., Ratiometric fluorescence detection of phosphorylated amino acids through excited-state proton transfer by using molecularly imprinted polymer (MIP) recognition nanolayers. Chemistry-a European Journal 2017, 23, 15974-15983. 20. Ghosh, A.; Verma, S.; Ganguly, B.; Ghosh, H. N.; Das, A., Influence of urea N-H acidity on receptor-anionic and neutral analyte binding in a ruthenium(II)-polypyridyl-based colorimetric sensor. European Journal of Inorganic Chemistry 2009, 2496-2507. 21. Boiocchi, M.; Del Boca, L.; Gomez, D. E.; Fabbrizzi, L.; Licchelli, M.; Monzani, E., Nature of urea-fluoride interaction: Incipient and definitive proton transfer. Journal of the American Chemical Society 2004, 126, 16507-16514. 22. Wan, W.; Wagner, S.; Rurack, K., Fluorescent monomers: "bricks" that make a molecularly imprinted polymer "bright". Analytical and Bioanalytical Chemistry 2016, 408, 1753-1771. 23. Gawlitza, K.; Wan, W.; Wagner, S.; Rurack, K., Fluorescent molecularly imprinted polymers. In Advanced Molecularly Imprinting Materials, John Wiley & Sons, Inc.: 2016; pp 89-128.


41

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 45

24. Hamase, K.; Iwashita, K.; Zaitsu, K., Enantio-selective derivatization of amino compounds in the presence of a molecular imprint polymer. Analytical Sciences 1999, 15, 411-412. 25. Peng, H.; Zhang, Y.; Zhang, J.; Xie, Q.; Nie, L.; Yao, S., Development of a thickness shear mode acoustic sensor based on an electrosynthesized molecularly imprinted polymer using an underivatized amino acid as the template. Analyst 2001, 126, 189-194. 26. Shinde, S.; Bunschoten, A.; Kruijtzer, J. A. W.; Liskamp, R. M. J.; Sellergren, B., Imprinted polymers displaying high affinity for sulfated protein fragments. Angewandte Chemie International Edition 2012, 51, 8326-8329. 27. Emgenbroich, M.; Borrelli, C.; Shinde, S.; Lazraq, I.; Vilela, F.; Hall, A. J.; Oxelbark, J.; De Lorenzi, E.; Courtois, J.; Simanova, A.; Verhage, J.; Irgum, K.; Karim, K.; Sellergren, B., A phosphotyrosine-imprinted polymer receptor for the recognition of tyrosine phosphorylated peptides. Chemistry-a European Journal 2008, 14, 9516-9529. 28. Fery-Forgues, S.; Fayet, J.-P.; Lopez, A., Drastic changes in the fluorescence properties of NBD probes with the polarity of the medium: involvement of a TICT state? Journal of Photochemistry and Photobiology A: Chemistry 1993, 70, 229-243. 29. Wulff, G.; Gross, T.; Schönfeld, R.; Schrader, T.; Kirsten, C., Molecular imprinting for the preparation of enzyme-analogous polymers. In Molecular and Ionic Recognition with Imprinted Polymers, American Chemical Society: 1998; Vol. 703, pp 10-28. 30. Wulff, G.; Knorr, K., Stoichiometric noncovalent interaction in molecular imprinting. Bioseparation 2001, 10, 257-276.


42

Page 43 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

31. Koshland, D. E., Application of a theory of enzyme specificity to protein synthesis. Proceedings of the National Academy of Sciences of the United States of America 1958, 44, 98-104. 32. Nicholls, I. A.; Andersson, H. S.; Charlton, C.; Henschel, H.; Karlsson, B. C. G.; Karlsson, J. G.; O'Mahony, J.; Rosengren, A. M.; Rosengren, K. J.; Wikman, S., Theoretical and computational strategies for rational molecularly imprinted polymer design. Biosensors & Bioelectronics 2009, 25, 543-552. 33. Cowen, T.; Karim, K.; Piletsky, S., Computational approaches in the design of synthetic receptors - A review. Analytica Chimica Acta 2016, 936, 62-74. 34. O'Mahony, J.; Wei, S.; Molinelli, A.; Mizaikoff, B., Imprinted polymeric materials. Insight into the nature of prepolymerization complexes of quercetin imprinted polymers. Analytical Chemistry 2006, 78, 6187-6190. 35. Karlsson, B. C. G.; O'Mahony, J.; Karlsson, J. G.; Bengtsson, H.; Eriksson, L. A.; Nicholls, I. A., Structure and dynamics of monomer-template complexation: An explanation for molecularly imprinted polymer recognition site heterogeneity. Journal of the American Chemical Society 2009, 131, 13297-13304. 36. Olsson, G. D.; Karlsson, B. C. G.; Shoravi, S.; Wiklander, J. G.; Nicholls, I. A., Mechanisms underlying molecularly imprinted polymer molecular memory and the role of crosslinker: resolving debate on the nature of template recognition in phenylalanine anilide imprinted polymers. Journal of Molecular Recognition 2012, 25, 69-73. 37. Wagner, S.; Bell, J.; Biyikal, M.; Gawlitza, K.; Rurack, K., Integrating fluorescent molecularly imprinted polymer (MIP) sensor particles with a modular microfluidic platform


43

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 44 of 45

for nanomolar small-molecule detection directly in aqueous samples. Biosensors & Bioelectronics 2018, 99, 244-250.


44

Page 45 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

ToC Graphic


45

Role of Counterions in Molecularly Imprinted Polymers for Anionic

Recommend Documents