Enhanced Binding and Biosensing of Carbohydrate-Functionalized

Key Laboratory of Mesoscopic Chemistry (Ministry of Education), School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210093, Peop...
7 downloads 16 Views 2MB Size
Download PDF
11330

J. Phys. Chem. B 2009, 113, 11330–11337

Enhanced Binding and Biosensing of Carbohydrate-Functionalized Monolayers to Target Proteins by Surface Molecular Imprinting Haifu Zheng and Xuezhong Du* Key Laboratory of Mesoscopic Chemistry (Ministry of Education), School of Chemistry and Chemical Engineering, Nanjing UniVersity, Nanjing 210093, People’s Republic of China ReceiVed: June 27, 2009

Concanavalin A (Con A) binding to the surfaces of mannose-functionalized binary monolayers was enhanced by surface molecular imprinting technique. The protein surface imprinting was prepared from binary Langmuir monolayers at the air-water interface through lateral reorganization of glycolipids directed by Con A in the subphase solution to form more specific bivalent binding sites, followed by horizontal immobilization of the binary monolayers and preservation of the enhanced affinity. The favorable spatial arrangement of the mannose ligands through lateral delivery matched well with protein binding pockets, and the steric crowding/ hindrance of neighboring ligands was minimized. The amounts of specifically bound proteins on the imprint surfaces are almost independent of surface density of the ligands, in contrast to the dependence of the bound amounts on surface density of the ligands for the control surfaces. The benefits of the protein surface imprinting included excellent mass transfer, ease of integration into sensor systems, directed creation of imprint sites, and biologically friendly aqueous media. This strategy generated tailor-made surfaces with high protein affinity and opens the possibility of surface design of intellectual materials and preparation of biosensors. Introduction Synthetic materials capable of selectively recognizing proteins are important in separations, biosensors, and the development of biomedical materials.1,2 The technique of molecular imprinting creates specific recognition sites in polymers by using template molecules.3,4 While successful for small template molecules, applying traditional bulk imprinting to macromolecules such as proteins has proven difficult due to inherent technical problems.2,5,6 These primarily include reduced mass transfer/permanent entrapment of macromolecule templates in polymer matrix and restricted selection of aqueous media.2,6 Imprinting of proteins represents one of the most challenging tasks.2 Surface imprints in polymeric thin films have improved binding kinetics by increasing mass transfer and reducing entrapment of the protein templates.7 Langmuir monolayers have half a framework structure of cellular membranes and have been demonstrated to mimic cell surface rearrangement due to the lateral mobility of the lipid components at the air-water interface.8-11 Lateral mobility is expected to play a key role in free rearrangement of ligand molecules in membranes,12-14 because a ligand for the second and subsequent binding events can be delivered through the lateral rearrangement of lipid components.15 The rearranged multicomponent monolayers directed by proteins in the subphase with cooperative and multivalent interactions can be immobilized by horizontal transfer onto sensor surfaces. Created specific binding sites can be preserved for protein recognition after bound proteins are removed. This process is simply a two-dimensional version of the solution-phase interactions between templates and functional monomers for the preparation of molecularly imprinted materials. The benefits of the lipid monolayer imprinting include excellent mass transfer, directed formation of imprint sites, and biologically friendly aqueous media, which are highly desirable * To whom correspondence should be addressed. E-mail: xzdu@ nju.edu.cn. Fax: 86-25-83317761.

for protein molecular imprinting.5,6 Few attempts to imprint proteins in the monolayers through electrostatic interactions were limited to specificity and regeneration.16-18 Protein-carbohydrate interactions play a crucial role in many cellular processes including cell adhesion, trafficking, metastasis, and immune response.19-22 These specific interactions occur through glycoproteins, glycolipids, and polysaccharides on cell surfaces and lectins.23 Lectins typically possess shallow binding pockets that are solvent-exposed, and therefore the proteincarbohydrate interactions are generally weak.20,24,25 It is possible for the lectins to participate in multivalent binding or the formation of several simultaneous binding events.20 The interaction strength and specificity can be compensated for by the presentation of multivalent protein binding. Multivalent interactions may achieve higher binding affinity, allow signaling through oligomerization, and induce changes in the distribution of molecules at the membrane interface.15,26 On the other hand, multivalent interactions are crucial for discovering a broad range of biosensors with various functions for use in medicine, environment, and food processing.15,25 Surface-based carbohydrate systems are usually used to facilitate the study of lectin recognition in comparison with the solution-based ones.27,28 Concanavalin A (Con A) is usually selected as a target protein because its multivalent binding to biological ligands such as mannose moieties has been extensively investigated and a number of methods have been developed to evaluate this binding.29,30 Con A (pI 4.5-5.6)31 is a multivalent binding protein found in jack bean and at pH < 6.0 exists as a dimer and at pH > 7.0 as a tetramer (molecular weight, 104 kDa for tetramer) with the unit cell dimensions of 6.32, 8.69, and 8.93 nm.32,33 Con A tetramer has four carbohydrate binding sites and presents two binding sites on each face. The orientation of these two binding sites allows Con A to engage in bivalent interactions in the planar lipid monolayers.34 A few model systems, such as self-assembled monolayers (SAMs),35-39 vesicles,26,29,40 and supported lipid bilayers

10.1021/jp9060279 CCC: $40.75  2009 American Chemical Society Published on Web 07/20/2009

Biosensing of Carbohydrate-Functionalized Monolayers

J. Phys. Chem. B, Vol. 113, No. 32, 2009 11331

Figure 1. Chemical structures of synthetic glycolipids and protein-repelling lipids.

(SLBs),41,42 offer platforms for studying the protein-carbohydrate interactions. It has been shown that the surface density and spatial arrangement of the ligands play an important role in protein binding.20,35,39 However, the surface densities of ligand molecules in these model systems cannot be precisely controlled. Although Con A binding to mannose-functionalized SAMs has been extensively studied,35-39 the ligand densities in the SAMs are generally different from their original mole fractions in the solution from which the SAMs are formed;39,43 on the other hand, the bivalent protein binding is obviously depressed due to the limitation of lateral rearrangement of the ligands through covalent attachment onto the surfaces. However, the bivalent protein binding can be realized through the interactions with unilamellar vesicles or SLBs rearranged from vesicles by means of lateral mobility of glycolipids,26,40,42 but the surface densities of the ligands on the membranes/bilayers cannot be precisely controlled either, because multicomponent lipids with different headgroups are known to be heterogeneously distributed on the two leaflets of vesicles. In comparison with the model systems formed from bulk solutions, the surface densities of different components in the Langmuir monolayers can be precisely controlled. Moreover, the influence of spatial arrangement of the ligands on protein binding is still largely unclear. The glycolipids in the Langmuir monolayers are capable of lateral mobility to form specific interactions with Con A in the subphase, so that the favorable spatial arrangement of the glycolipids for enhanced protein binding can be reached and the binding affinity can be preserved by the lipid monolayer imprinting. Results and Discussion Surface Behaviors of Binary Monolayers at the Air-Water Interface. Double-chained glycolipids with the mannose moieties and lipids with the protein resistant headgroups were synthesized (see Supporting Information), and their chemical structures are shown in Figure 1. The four synthetic glycolipids/ lipids have identical hydrocarbon chains and glyceryl skeletons except for the nonionic hydrophilic moieties. The corresponding surface pressure (π)-area (A) isotherms of the four individual monolayers at the air-water interface are shown in Figure 2. DPEM and DPE with an oligo(ethylene glycol) (OEG) spacer/ moiety displayed typical liquid expanded-liquid condensed

Figure 2. Surface pressure-area isotherms of the monolayers of DPEM, DPE, DPM, and DPG on the phosphate-buffered saline (PBS, 10 mM, 150 mM NaCl, pH 7.4) containing Ca2+ (0.1 mM) and Mn2+ (0.1 mM) at 22 °C.

phase transition with a plateau around at 9.5 and 2.5 mN/m, respectively, primarily due to the large sizes of the hydrophilic headgroups. Their limiting areas by extrapolating linear regions of the liquid-condensed phases to zero surface pressure were 0.421 and 0.40 nm2/molecule, respectively. The surface pressures and plateau ranges of the phase transitions were accordingly reduced for their counterparts DPM and DPG without OEG spacer/moiety, so that a liquid-condensed monolayer was only observed for DPG with a limiting area of 0.383 nm2/ molecule. The limiting area of DPM estimated from the liquidcondensed phase was 0.484 nm2/molecule. According to the surface phase rule, if the two components are miscible, then the collapse pressure of their monolayers will change depending on the composition.44 By contrast, in the case of two completely immiscible components, their monolayers will exhibit two collapse points, each occurring at the same surface pressure regardless of monolayer composition.44 The binary monolayers of DPEM and DPE at the air-water interface obviously displayed a single collapse pressure varying with mole fraction of DPEM (XDPEM) (Figure S1a). The miscibility of the binary monolayers can be further investigated from the variation of

11332

J. Phys. Chem. B, Vol. 113, No. 32, 2009

Zheng and Du

mean molecular areas. At a given surface pressure, the excess area can be represented by comparing the mean molecular area of a mixed monolayer consisting of components 1 and 2 with that of an ideally mixed monolayer44-46

Aex(π) ) A12(π) - Aid(π) ) A12(π) - [X1A1(π) + X2A2(π)] where A12 and Aid are the experimental and ideal molecular areas of the mixed monolayer at a given surface pressure, respectively. A1 and A2 stand for the molecular areas of pure components 1 and 2 at the same surface pressure, and X1 and X2 denote the mole fractions in the mixture. If two components are ideally mixed or totally immiscible, the excess molecular area, Aex, will be zero. Any derivation from ideality would indicate that the two components are miscible. In general, negative derivation implies an attractive interaction and/or better miscibility between them, while a repulsive interaction and/or poor miscibility will lead to positive derivation from an ideal behavior.46,47 Seen from the excess molecular area as a function of XDPEM in the binary mixtures (Figure S1b), the excess molecular areas fluctuated around the ideal behaviors with small positive and negative deviations above the phase transition pressures. These results suggest that DPEM and DPE were miscible at the molar ratios investigated and their surface behaviors were close to the ideal mixtures to some extents due to their compatible chemical structures. For the binary monolayers of DPM and DPG, a similar case occurred with a small positive deviation (Figure S2), whereas the binary mixtures of DPM and DPE exhibited a different surface behavior (Figure S3). Positive deviations of the excess molecular areas for various molar ratios at different surface pressures were observed because of the large difference in the hydrophilic headgroup between DPM and DPE. It is obvious that poor miscibility occurred between them. Protein Binding to Immobilized Binary Monolayers. Surface plasmon resonance (SPR) is an excellent method for the investigation of protein binding at the solid-water interface in real time without labeling analytes and gives rich information on protein binding kinetics and protein-bound amount. It has been shown that the monolayers at a surface pressure of 30 mN/m are enough to resist protein penetration into the hydrophobic chain regions besides being capable of lateral mobility.48 The surface pressure of 30 mN/m is equivalent to the lateral pressure of cellular membranes.49-51 During the SPR monitoring, the monolayers at the air-water interface at 30 mN/m were first horizontally immobilized with the octadecanethiol (ODT)modified SPR gold surfaces followed by protein injection underneath the monolayers, which were equivalent to SAMs to some degree but different from them because the binary Langmuir monolayers might be already miscible prior to immobilization. Negligible Con A specific binding to the individual monolayers of DPE and DPG was observed (Figure 3). It is well-known that the OEG moieties with the numbers of EG units g3 possess excellent biocompatibility to be protein resistant.52,53 Generally, the hydroxyls (OH) are not good proteinrepelling groups; however, the DPG monolayer here was capable of preventing nonspecific Con A binding. In contrast, a significant Con A binding to the glycolipid monolayers of DPEM and DPM was observed (Figure 3) due to the specific recognition between the mannose ligands and Con A, but the amount of specifically bound proteins on the surface of the DPEM monolayer at saturation was greater than that on the DPM surface. Although significant amounts of mannose ligands in the glycolipid monolayers were available for Con A, no much room could accommodate interrogating Con A for the access

Figure 3. SPR sensorgrams of Con A binding to the immobilized monolayers of DPEM, DPE, DPM, and DPG on the PBS solution (pH 7.4) containing Ca2+ and Mn2+ at 22 °C at the surface pressure 30 mN/m, respectively: arrow a, injecting Con A with a final concentration of 100 µg/mL; arrow b, rinsing with PBS solution (pH 7.4).

of the ligands to protein binding pockets due to the steric crowding of neighboring ligands. The flexible OEG spacers in the DPEM monolayer (one OEG spacer per double chain) facilitated the mannose ligands to access readily to Con A to a certain extent in comparison with the DPM monolayer without OEG spacer. Con A binding to various kinds of immobilized binary monolayers is shown in Figure 4. For the binary mixtures of DPEM and DPE (Figure 4a), the amount of specifically bound protein on the surfaces increased rapidly with increasing XDPEM from 0.01 to 0.1 followed by a drop at XDPEM ) 0.2, then an increase in the amount continued upon further increase of XDPEM. The amount of bound proteins at XDPEM ) 0.1 was even higher than that at XDPEM ) 0.5. The binary surface at XDPEM ) 0.1 afforded the best affinity for Con A. The amounts of specifically bound proteins on the various binary surfaces at saturation, estimated on the basis of the relationship (an angle shift of 0.1° ∼ a protein surface density of 1 ng/mm2),54 against mole fraction of the ligands are presented in Figure 4d. The phenomena of the obvious increase in the amount of specifically bound proteins for the Con A-mannose interactions at small surface densities of ligands were observed previously.10,35,39,41 These observations were attributed to the tight packing of hydrocarbon chains in the binary monolayers resulting in limited access of the ligands to the protein binding pockets.36,39 Con A is capable of forming two attachment points to the surface, and the distance between the two points on the proteins is approximately 6.5 nm, as determined from X-ray structural analysis.55 It is obvious that spatial arrangement of the ligands in the binary monolayers had a large impact on protein binding. This suggests that the ligands with favorable distribution in the binary monolayers could match well with the protein binding pockets to improve protein affinity even at a given XDPEM. The increase in the amount of specifically bound proteins upon further increase in XDPEM indicated that the reduced protein binding due to the steric crowding of the ligands could be compensated for by an increase in surface density of the ligands to some degree considering the formation of stable bivalent binding. It is clear that the amount of specifically bound Con A was related not only to surface density of the ligands but also to steric crowding of neighboring ligands and finally determined by the balance

Biosensing of Carbohydrate-Functionalized Monolayers

J. Phys. Chem. B, Vol. 113, No. 32, 2009 11333

Figure 4. SPR sensorgrams of Con A binding to various kinds of immobilized (control) binary monolayers on the PBS solution (pH 7.4) containing Ca2+ and Mn2+ at 22 °C with XDPEM or XDPM at the surface pressure 30 mN/m: (a) DPEM and DPE; (b) DPM and DPG; (c) DPM and DPE. (d) Amount of specifically bound Con A at saturation as a function of mole fraction of glycolipids. Arrow a, injecting Con A with a final concentration of 100 µg/mL; arrow b, rinsing with PBS solution (pH 7.4).

TABLE 1: Association Constants (Ka) of Con A Binding to the Binary Monolayers of DPEM and DPE with Different XDPEM Ka (M-1) monolayers

XDPEM ) 0.1

control (immobilized) imprint

2.63 × 10 1.42 × 107 6

0.2

0.3

0.5

4.96 × 10 9.01 × 10 3.01 × 106 2.39 × 106 5

5

1.97 × 106 2.60 × 106

between them. The association constants of Con A binding to the immobilized binary monolayers with different XDPEM calculated on the basis of the methods9 are listed in Table 1. The values are consistent with those reported previously.9,20,25,39,56-58 For the binary monolayers of DPM and DPG (Figure 4b), the case was very similar to that for the binary monolayers of DPEM and DPE. This further demonstrates that surface density and spatial arrangement of the ligands play an important role in protein binding. But for XDPM g 0.3, the amounts of specifically bound proteins were diminished in comparison with the binary monolayers of DPEM and DPE. It is obvious that the steric crowding effect of the ligands was dominant when XDPM was large. Similarly, the flexible OEG spacers in the binary monolayers of DPEM and DPE were favorable to the access of the ligands to the protein binding pockets in comparison with the binary monolayers of DPM and DPG. In order to give a deeper insight into the influence of steric crowding/hindrance of the ligands on protein binding, the binary

monolayers of DPM and DPE were investigated (Figure 4c). The amounts of specifically bound proteins at different XDPM were considerably reduced but increased gradually with XDPM in the absence of a maximum value at XDPM ) 0.1 as in the binary monolayers of DPEM and DPE, DPM and DPG. It is obvious that in the binary surfaces of DPM and DPE the mannose ligands were embedded in the OEG moieties of DPE. The steric hindrance from the OEG moieties inhibited Con A from binding to the mannose ligands. It is most likely that the protein-carbohydrate interactions occurred in the regions of local DPM domains or with the sparse distribution of DPE for the accommodation of Con A. Although the poor miscibility occurred between DPM and DPE inferred from the isotherms of their binary monolayers (Figure S3), the probability of the formation of this kind of local DPM domains should be small for the accommodation of relatively large proteins; moreover, in the local DPM domains the steric crowding of neighboring ligands at high XDPM as discussed above led to a further reduce in protein binding. It is not difficult to understand that the amount of specifically bound proteins increased gradually with XDPM because the probability of the formation of the local DPM domains became relatively large. It is clear that the steric hindrance effect from the OEG moieties for protein binding was more than that from the ligand crowding. Scheme 1 illustrates Con A binding to the three kinds of immobilized binary

11334

J. Phys. Chem. B, Vol. 113, No. 32, 2009

Zheng and Du

SCHEME 1: Schematic Illustration of Con A Binding to the Immobilized (Control) Binary Monolayers of DPEM and DPE, DPM and DPG, and DPM and DPE

Figure 5. p-Polarized difference IRRAS spectra of the binary monolayers of DPEM and DPE at the air-water interface at the surface pressure 30 mN/m at 22 °C after (binding saturation) and before protein binding at XDPEM: (a) 0.1; (b) 0.2; (c) 0.3; (d) 0.4.

monolayers. From these binary surfaces, it can be concluded that the enhanced Con A binding could be achieved when the mannose ligands were extended from the inert surfaces in a suitable spatial arrangement to match well with the protein binding pockets on demand. Enhanced Protein Binding to Binary Monolayers by Surface Molecular Imprinting. In situ infrared reflection absorption spectroscopy (IRRAS) technique is one of the leading methods for structural analyses of monolayers at the air-water interface59,60 and currently the only physical method able to directly monitor protein secondary structures in situ in Langmuir monolayers.61 This technique was used to investigate Con A binding to the binary monolayers of DPEM and DPE at the air-water interface at 30 mN/m. The difference spectra after and before protein binding to the monolayers with different XDPEM are presented in Figure 5. Protein injection resulted in the appearance of amide I and amide II bands, and the position of the amide I band centered at 1640 cm-1 suggested the presence of a great amount of antiparallel β-sheet conformation, which is consistent with the secondary structure of native Con

A.33 The spectral features indicate that the native secondary structures were basically maintained for the bound proteins on the surfaces of the hydrophilic headgroups of the binary monolayers. Since no amide I or amide II band was observed from the DPE monolayer, it is suggested that the amide I and II bands observed from the binary monolayers reflected specific protein binding. More interestingly, the intensities of the amide I bands for the binary monolayers at the air-water interface with different XDPEM were very comparable, which was different from the amounts of specifically bound proteins on the corresponding immobilized monolayers at the solid-water interface varying with XDPEM (Figure 4a). One plausible explanation is that the glycolipids underwent lateral reorganization to form a new spatial arrangement in the binary monolayers at the air-water interface directed by Con A in the subphase. The optimal spatial arrangement of the mannose ligands at the interface could match well with the protein binding pockets and minimize the steric crowding of neighboring ligands. In order to further verify the formation of the new spatial arrangement of the ligands and preserve the enhanced Con A binding to the binary monolayers, the SPR technique was used to investigate protein binding to the surfaces of initially fluid and immobilized (control) binary monolayers of DPEM and DPE (Figure 6). For the initial protein binding to the Langmuir monolayers at the air-water interface, the binding kinetics were not attained as the SPR sensors could not be in contact with the monolayers if they were to remain laterally mobile. However, the final adsorption amounts were obtained by horizontally placing the ODT-modified SPR sensors in contact with the monolayers after identical binding times. The whole process was equivalent to the preparation of protein surface imprinting. At a given XDPEM, the initially fluid and control monolayers had the same surface densities because the identical amounts of the lipid samples were spread at the fixed microtrough area to reach the surface pressure of 30 mN/m (see Supporting Information). This means that the SPR angle shifts, which were contributed from the binary lipid monolayers and premodified ODT monolayers on the SPR sensors, were practically identical in the two cases. The SPR angle shifts due to the protein binding were only presented after the subtraction of the contributions from the lipid and ODT monolayers. For the first binding stages, the amounts of specifically bound Con A on the surfaces of the imprint monolayers were increased

Biosensing of Carbohydrate-Functionalized Monolayers

J. Phys. Chem. B, Vol. 113, No. 32, 2009 11335

Figure 6. SPR sensorgrams of Con A binding to and desorption from the surfaces of imprint and control monolayers of DPEM and DPE at 22 °C at the surface pressure 30 mN/m with XDPEM: (a) 0.1; (b) 0.2; (c) 0.3; (d) 0.5. Arrow a, injecting Con A with a final concentration of 100 µg/mL; arrow b, rinsing with PBS solution (pH 7.4); arrow c, rinsing with acetate-acetic acid buffer solution (pH 1.5).

varying with XDPEM in comparison with the control ones. In the case of XDPEM ) 0.1, the amount on the imprint monolayer increased by about 30% in comparison with the control one. It is most likely that the enhanced protein binding resulted from the spatial rearrangement of DPEM in the monolayer at the air-water interface directed by Con A in the subphase. The optimal spatial arrangement of the ligands means that significant amounts of bivalent binding sites were created to meet the requirement for the separation distance of about 6.5 nm54 for the two binding sites of Con A on demand and the steric crowding of neighboring ligands were minimized as possible as they could. The association constant in the case of the imprint monolayer was increased to be 1.42 × 107 M-1 (Table 1). The protein-directed assembly of the binary monolayer gave rise to a new spatial arrangement of the ligands in the monolayer complementary to the protein binding pockets, which was different from the original ligand arrangement in the monolayer determined only by the interaction between the two lipid components in the absence of protein. It is clear that the protein binding was related not only to the surface density of the functional ligands but also to the spatial arrangement of the ligands in the monolayers. In the case of XDPEM ) 0.2, the amount of specifically bound proteins on the imprint monolayer increased by as much as 130% in comparison with the control one. It is easy to understand the significant improvement in protein binding in

this case since the corresponding control monolayer had a low affinity for Con A. The favorable spatial arrangement of the ligands on the surface of the imprint monolayer minimized the steric crowding of neighboring ligands and created more bivalent binding sites for the proteins. The binding affinity was enhanced

Figure 7. Amounts of specifically bound Con A on the surfaces of the imprint and control monolayers of DPEM and DPE as a function of XDPEM, together with the IRRAS intensities of amide I bands after protein binding to the binary monolayers at the air-water interface for comparison.

11336

J. Phys. Chem. B, Vol. 113, No. 32, 2009

Zheng and Du

SCHEME 2: Schematic Illustration of the Formation of a Protein-Imprinted Binary Monolayer in Comparison with a Control Monolayer

from the association constant of 4.96 × 105 M-1 for the control monolayer to 3.01 × 106 M-1 for the imprint one (Table 1). Upon an increase in XDPEM to 0.3, the case was similar to that at XDPEM ) 0.2. The amount of specifically bound Con A for the imprint monolayer was 50% greater than that for the control monolayer. Note that the steric crowding of neighboring ligands in the imprint monolayer could not be ignored in this case due to relatively large XDPEM. Upon further increase to XDPEM ) 0.5, the amount on the imprint surface was slightly improved. This was because the lateral reorganization of DPEM could not cause a significant change in the spatial arrangement of the ligands due to the excess DPEM. It was difficult to reduce considerably the steric crowding of the neighboring ligands at high XDPEM. The amounts of specifically bound Con A on the surfaces of the imprint and control monolayers against XDPEM are presented in Figure 7, together with the IRRAS intensities of the amide I bands for the binary monolayers at the air-water interface for comparison. The amounts on the imprint surfaces were almost independent of XDPEM due to the preferential spatial arrangement of the ligands, in contrast to the amounts on the control surfaces dependent on XDPEM, which was consistent with the amounts of bound proteins (from the intensities of the amide I bands) in the monolayers at the air-water interface almost independent of XDPEM. This demonstrates that the enhanced protein binding was attributed to the optimal spatial arrangement of the ligands through the lateral delivery of the glycolipids at the air-water interface and this surface imprinting technique was suited for the preparation of protein imprinting from aqueous media. After the specifically bound proteins were washed with the acidic buffer solution of pH 1.5 followed by the initial buffer solution (pH 7.4), both the imprint monolayers and the control monolayers were completely regenerated (Figure 6). During the rebinding stages, the protein binding/desorption kinetics on the control surfaces were nearly the same as those during the first binding stages, and the amounts of specifically bound proteins at saturation and protein desorption kinetics on the imprint surfaces were almost identical to those during the first binding stages. These mean that the optimal spatial arrangement of the glycolipids formed in the monolayers at the air-water interface

were imprinted and preserved for the subsequent protein binding events. The creation of an imprint binary monolayer with enhanced protein binding is schematically represented in Scheme 2 in comparison with a control monolayer. This strategy generated tailor-made surfaces with enhanced protein binding in aqueous media, allowed the integrity of the imprint surface into sensor systems for biosensing, and eliminated problems associated with mass transfer/entrapment of proteins in a matrix and organic solvents. The surface imprinting technique can apply to other functional lipids and target proteins in aqueous media. On the other hand, the SPR and IRRAS techniques used here were well suited for the in situ studies of interfacial systems relevant to aqueous solutions, such as the solid-water interface and air-water interface, so that physiological environments of the proteins could be maintained not only during the preparation of the imprint monolayers but also in the course of protein recognition and biosensing. A combination of the two techniques could provide abundant information on protein binding to the monolayers at the interface both in real time and on the molecular level. Conclusions Con A binding to three kinds of immobilized binary monolayers functionalized with mannose ligands was investigated in detail. The amount of specifically bound proteins was found to be related to surface density of the ligands, steric crowding of neighboring ligands, and steric hindrance of protein-repelling matrix components. It is concluded that enhanced Con A binding to the binary monolayers could be achieved when the ligands were extended from the protein-resistant surfaces in a suitable spatial arrangement to match well with protein binding pockets on demand. The affinity of the binary surfaces to Con A was enhanced by surface molecular imprinting, which was formed by self-assembling glycolipids in the protein-resistant matrix monolayers at the air-water interface through lateral reorganization directed by Con A in the subphase followed by horizontal immobilization of the binary monolayers. The favorable spatial arrangement of the mannose ligands through lateral delivery matched well with the protein binding pockets, and

Biosensing of Carbohydrate-Functionalized Monolayers the steric crowding/hindrance of neighboring ligands was minimized. The amounts of specifically bound proteins on the imprint surfaces were almost independent of surface density of the ligands, in contrast to the dependence of the bound amounts on surface density of the ligands for the control surfaces. The benefits of the protein surface imprinting included excellent mass transfer, directed formation of imprint sites, biologically friendly aqueous media, and ease of integration into sensor systems for information storage and biosensing. This strategy generated tailor-made surfaces with enhanced protein binding and opens the possibility of surface design of intellectual materials and preparation of biosensors. Acknowledgment. This work was supported by the National Natural Science Foundation of China (Grants 20673051, 20635020, and 20873062), the Natural Science Foundation of Jiangsu Province (Grant BK2007519), and the program for New Century Excellent Talents in University (NCET-07-0412). Supporting Information Available: Experimental section including glycolipid/lipid syntheses, monolayer preparation and isotherms, SPR and IRRAS measurements, and miscibility of the binary monolayers of DPEM and DPE, DPM and DPG, and DPEM and DPE at the air-water interface. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Vidyasankar, S.; Arnold, F. H. Curr. Opin. Biotechnol. 1995, 6, 218. (2) Bossi, A.; Bonini, F.; Turner, A. P. F.; Piletsky, S. A. Biosens. Bioelectron. 2007, 22, 1131. (3) Wulff, G. Angew. Chem., Int. Ed. 1995, 34, 1812. (4) Mosbach, K. Trends Biochem. Sci. 1994, 19, 9. (5) Hansen, D. E. Biomaterials 2007, 28, 4178. (6) Dhruv, H.; Pepalla, R.; Taveras, M.; Britt, D. W. Biotechnol. Prog. 2006, 22, 150. (7) Shi, H.; Tsai, W.-B.; Garrison, M. D.; Ferrari, S.; Ratner, B. D. Nature 1999, 398, 593. (8) Leblanc, R. M. Curr. Opin. Chem. Biol. 2006, 10, 529. (9) Ebara, Y.; Okahata, Y. J. Am. Chem. Soc. 1994, 116, 11209. (10) Wang, S.; Leblanc, R. M. Biochim. Biophys. Acta 1999, 1419, 307. (11) Glomm, W. R.; Voden, S.; Halskau, Ø., Jr.; Ese, M.-H. Anal. Chem. 2009, 81, 3042. (12) Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651. (13) Spencelayh, M. J.; Cheng, Y.; Bushby, R. J.; Bugg, T. D. H.; Li, J.-J.; Henderson, P. J. F.; O’Reilly, J.; Evans, S. D. Angew. Chem., Int. Ed. 2006, 45, 2111. (14) Philips, K. S.; Wilkop, T.; Wu, J.-J.; Al-Kaysi, R. O.; Cheng, Q. J. Am. Chem. Soc. 2006, 128, 9590. (15) Yang, T.; Baryshnikova, O. K.; Mao, H.; Holden, M. A.; Cremer, P. S. J. Am. Chem. Soc. 2003, 125, 4779. (16) Du, X.; Hlady, V.; Britt, D. Biosens. Bioelectron. 2005, 20, 2053. (17) Du, X.; Wang, Y. J. Phys. Chem. B 2007, 111, 2347. (18) Turner, N. W.; Wright, B. E.; Hlady, V.; Britt, D. W. J. Colloid Interface Sci. 2007, 308, 71. (19) Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357. (20) Smith, E. A.; Thomas, W. D.; Kiessling, L. L.; Corn, R. M. J. Am. Chem. Soc. 2003, 125, 6140. (21) Chen, I. J.; Chen, H. L.; Demetriou, M. J. Biol. Chem. 2007, 282, 35361. (22) Philobello, K. T.; Mahal, L. K. Curr. Opin. Chem. Biol. 2007, 11, 1.

J. Phys. Chem. B, Vol. 113, No. 32, 2009 11337 (23) Sharon, N.; Lis, H. Science 1989, 246, 227. (24) Kiessling, L. L.; Pohl, N. L. Chem. Biol. 1996, 3, 71. (25) Liang, P.-H.; Wang, S.-K.; Wong, C.-H. J. Am. Chem. Soc. 2007, 129, 11177. (26) Thomas, G. B.; Rader, L. H.; Park, J.; Abezgauz, L.; Danino, D.; DeShong, P.; English, D. S. J. Am. Chem. Soc. 2009, 131, 5471. (27) Kiessling, L. L.; Cairo, C. W. Nat. Biotechnol. 2002, 20, 234. (28) Wang, D.; Liu, S.; Trummer, B. J.; Deng, C.; Wang, A. Nat. Biotechnol. 2002, 20, 275. (29) Martin, A. L.; Li, B.; Gillies, E. R. J. Am. Chem. Soc. 2009, 131, 734. (30) Gestwicki, J. E.; Cairo, C. W.; Strong, L. E.; Oetjen, K. A.; Kiessling, L. L. J. Am. Chem. Soc. 2002, 124, 14922. (31) Barbucci, R.; Magnani, A.; Chiumiento, A.; Pasqui, D.; Cangioli, I.; Lamponi, S. Biomacromolecules 2005, 6, 638. (32) Becker, J. W.; Reeke, G. N., Jr.; Cunningham, B. A.; Edelman, G. M. Nature 1976, 259, 406. (33) Hardman, K. D.; Ainsworth, C. F. Biochemistry 1972, 11, 4910. (34) Page, M. I.; Jencks, W. P. Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 1678. (35) Horan, N.; Yan, L.; Isobe, H.; Whitesides, G. W.; Kahne, D. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 11782. (36) Houseman, B. T.; Mrksich, M. Angew. Chem., Int. Ed. 1999, 38, 782. (37) Yonzon, C. R.; Jeoung, E.; Zou, S.; Schatz, G. C.; Mrksich, M.; Van Duyne, R. P. J. Am. Chem. Soc. 2004, 126, 12669. (38) Ding, L.; Cheng, W.; Wang, X.; Ding, S.; Ju, H. J. Am. Chem. Soc. 2008, 130, 7224. (39) Grant, C. F.; Kanda, V.; Yu, H.; Bundle, D. R.; McDermott, M. T. Langmuir 2008, 24, 14125. (40) Bondurant, B.; Last, J. A.; Waggoner, T. A.; Slade, A.; Sasaki, D. Y. Langmuir 2003, 19, 1829. (41) Mann, D. A.; Kanai, M.; Maly, D. J.; Kiessling, L. L. J. Am. Chem. Soc. 1998, 120, 10575. (42) Taylor, J. D.; Linman, M. J.; Wilkop, T.; Cheng, Q. Anal. Chem. 2009, 81, 1146. (43) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1988, 110, 3665. (44) Gaines, G. L., Jr. Insoluble Monolayers at Liquid-Gas Interfaces; Wiley-Interscience: New York, 1966. (45) Chou, T.-H.; Chang, C.-H. Langmuir 2000, 16, 3385. (46) Rey Go´mez-Serranillos, I.; Min˜ones, J., Jr.; Dynarowicz-£1tka, P.; Minones, J.; Conde, O. Langmuir 2004, 20, 11414. (47) Tanwir, K.; Tsoukanova, V. Langmuir 2008, 24, 14078. (48) Wang, Y.; Du, X. Langmuir 2006, 22, 6195. (49) Zwaal, R. F. A.; Demel, R. A.; Roelofsen, B.; Deenen, L. L. M. Trends Biochem. Sci. 1976, 1, 112. (50) Serfis, A. B.; Katzenberger, R.; William, K.; Tran, N. J. Colloid Interface Sci. 1999, 215, 356. (51) Peetla, C.; Labhasetwar, V. Mol. Pharm. 2008, 5, 418. (52) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164. (53) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10714. (54) Stengberg, E.; Persson, B.; Roos, H.; Urbaniczky, C. J. Colloid Interface Sci. 1991, 143, 513. (55) Derewenda, Z.; Yariv, J.; Helliwell, J. R.; Kalb, A. J.; Dodson, E. J.; Papiz, M. Z.; Wan, T.; Champbell, J. EMBO J. 1989, 8, 2189. (56) Linman, M. J.; Taylor, J. D.; Yu, H.; Chen, X.; Cheng, Q. Anal. Chem. 2008, 80, 4007. (57) Pei, Z.; Anderson, H.; Aastrup, T.; Ramstro¨m, O. Biosens. Bioelectron. 2005, 21, 60. (58) Critchley, P.; Clarkson, G. J. Org. Biomol. Chem. 2003, 1, 4148. (59) Mendelsohn, R.; Brauner, J. W.; Gericke, A. Annu. ReV. Phys. Chem. 1995, 46, 305. (60) Zheng, J.; Leblanc, R. M. Infrared Reflection Absorption Spectroscopy of Monolayers at the Air-Water Interface. In AdVanced Chemistry of Monolayers at Interfaces; Imae, T., Ed.; Elesevier BV: Amsterdam, 2007; Chapter 10. (61) Cai, P.; Flach, C. R.; Mendelsohn, R. Biochemistry 2003, 42, 9446.

JP9060279