Oriented Immobilization of Fab Fragments by Site-Specific

Aug 14, 2015 - Fab fragments exhibit numerous advantages over intact antibodies such as smaller size, reduced immunogenicity, and ease of production w...
3 downloads 10 Views 2MB Size
Article pubs.acs.org/Langmuir

Oriented Immobilization of Fab Fragments by Site-Specific Biotinylation at the Conserved Nucleotide Binding Site for Enhanced Antigen Detection Nur Mustafaoglu,† Nathan J. Alves,† and Basar Bilgicer*,†,‡,§,∥,⊥ †

Department of Chemical and Biomolecular Engineering, ‡Department of Chemistry and Biochemistry, §Advanced Diagnostics and Therapeutics, ∥Mike and Josie Harper Cancer Research Institute, and ⊥Center for Rare and Neglected Diseases, University of Notre Dame, Notre Dame, Indiana 46556, United States

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

S Supporting Information *

ABSTRACT: Oriented immobilization of antibodies and antibody fragments has become increasingly important as a result of the efforts to reduce the size of diagnostic and sensor devices to miniaturized dimensions for improved accessibility to the enduser. Reduced dimensions of sensor devices necessitate the immobilized antibodies to conserve their antigen binding activity for proper operation. Fab fragments are becoming more commonly used in small-scaled diagnostic devices due to their small size and ease of manufacture. In this study, we used the previously described UVNBSBiotin method to functionalize Fab fragments with IBA-EG11-Biotin linker utilizing UV energy to initiate a photo-cross-linking reaction between the nucleotide binding site (NBS) on the Fab fragment and IBA-Biotin molecule. Our results demonstrate that immobilization of biotinylated Fab fragments via UV-NBSBiotin method generated the highest level of immobilized Fab on surfaces when compared to other typical immobilization methods while preserving antigen binding activity. UV-NBSBiotin method provided 432-fold, 114-fold, and 29-fold improved antigen detection sensitivity than physical adsorption, NHS-Biotin, and ε-NH3+, methods, respectively. Additionally, the limit of detection (LOD) for PSA utilizing Fab fragments immobilized via UV-NBSBiotin method was significantly lower than that of the other immobilization methods, with an LOD of 0.4 pM PSA. In summary, site-specific biotinylation of Fab fragments without structural damage or loss in antigen binding activity provides a wide range of application potential for UV-NBS immobilization technique across numerous diagnostic devices and nanotechnologies.



INTRODUCTION Technological advancements in small-scaled device fabrication improve the sensitivity and selectivity of measurements while requiring smaller sample sizes and less energy. Such devices are now commonly used for immunosensor-based applications including screening of pathogens, disease identification, and point-of-care (POC) analysis; subsequently, their utility in medical diagnostics have been gaining favorable interest.1,2 Antibodies, antibody fragments and their derivatives have been widely used in such systems due to their high specificity for antigen recognition.3−5 Antibody Fab fragments, which are essentially functional domains of an antibody’s structure that maintain antigen binding activity, are becoming the preferred components of small-scaled medical diagnostic devices.3,6,7 As devices decrease in size, the number of detection components that can physically fit within the detection area is greatly reduced. For this reason, the percentage of detection component that remains active post immobilization is critical for the proper functioning of such devices. Commonly used antibody immobilization methods often result in partial or total loss of antigen binding activity. Therefore, various small-scaled devices including biosensors, affinity chromatography, immunoassays, and biomedical screening systems rely on oriented © XXXX American Chemical Society

antibody immobilization techniques to improve antigen binding activity of surface immobilized antibodies.8−12 Developing reliable and reproducible oriented immobilization techniques has been one of the most formidable challenges during biosensor device design. The most commonly utilized method for antibody immobilization on detection surfaces is physical adsorption, which results in randomly oriented antibody molecules on the surface due to nonspecific multipoint attachment between the antibody and the surface.13−15 This multipoint attachment is driven by a combination of electrostatic, hydrophobic/hydrophilic, and van der Waals interactions. Random and unstable binding of proteins to the surface, through physical adsorption, causes irreproducible results yielding significant decreases in antigen binding activity and increased inconsistency between devices.9,16,17 Additionally, dramatic variations in the topology of the antibody coated surfaces due to diverse random orientations can also negatively affect antigen binding specificity.18,19 Received: May 11, 2015 Revised: August 5, 2015

A

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

Langmuir

number of these strategies are based on diverse chemical attachment methods and fail to yield reproducible and reliable results.36−38 Commonly used modification methods for intact antibody functionalization often cause even more dramatic damage to the antigen binding site of Fab fragments due to the Fab fragment’s smaller modifiable surface area compared to intact antibodies. Hence, modification of Fab fragments without impacting antigen binding activity remains a central challenge that needs to be overcome. In this study, we demonstrate an application for oriented immobilization of UV-NBS biotinylated Fab fragments onto streptavidin-coated surfaces for enhanced antigen detection systems. We selected a prostate specific antigen (PSA) detection system to demonstrate the utility of the UV-NBSBiotin method for Fab fragment immobilization without impacting antigen binding activity or overall Fab fragment stability. PSA, also known as gamma-seminoprotein or kallikrein-3 (KLK3), is a 30−34 kDa glycoprotein enzyme.39 The serum level of PSA has been used as an indicator of prostate cancer, prostatitis and benign prostatic hyperplasia.40 Here, we use a mouse IgG PSAspecific capture antibody (IgGPSA, clone B731M), free PSA, and HRP conjugated PSA detection antibody (det-IgGPSA, clone 906) to demonstrate the advantages of the UV-NBSBiotin method compared to physical adsorption, ε-NH3+, and NHSBiotin methods. We repeated the experiments using another clone of PSA-specific capture antibody (IgGPSA, clone 5A6) in order to show usability of the UV-NBSBiotin method across different antibodies and to demonstrate its reliability.

Chemical conjugation is another commonly used method for immobilization of antibodies onto detection surfaces. Generating covalent bonds between the protein and the surface often requires harsh conditions that can cause irreversible damage to the antibody structure leading to significant loss of antigen binding activity and selectivity.1,19−21 Chemical conjugation methods of antibodies generally utilize existing amino acid side chains within the antibody as points of attachment such as εamino groups of lysines, β-carboxyl groups of aspartic acids, γcarboxyl groups of glutamic acids, and thiol groups generated by reduction of cysteine disulfide bonds are the most commonly used moieties for chemical conjugation reactions.15,22−24 Unfortunately, most of these moieties exist as multiple copies on a variety of locations on an antibody, and conjugation at these sites often still results in loss of antigen binding activity due to the inaccessibility of binding sites due to physical hindrance or random immobilization.21,25 It is noteworthy that, although targeting the cysteine residues provides specific sites for oriented conjugation, high levels of antigen binding activity post immobilization is typically not observed due to the harsh procedures used in disulfide bond reduction that impact the antibody tertiary structure.23,26,27 We have previously reported UV-NBS, UV-NBSBiotin, and UV-NBSThiol methods that have been developed in our laboratory as universal methods for antibody28 and Fab29 functionalization, as well as oriented surface immobilization.26,27,30 These methods utilize indole 3-butyric acid’s (IBA) affinity (Kd = 1−8 μM) for the nucleotide binding site (NBS) which is a highly conserved binding pocket located in the variable domain of the Fab fragment.30,31 The conservation of the NBS was characterized in our previous in silico studies by overlaying >260 antibody crystal structures available in the RCSB Protein Data Bank.30,31 The NBS consists of two tyrosine residues (Y42 and Y103) on the light chain and one tyrosine (Y103) and one tryptophan residue (W118) on the heavy chain.30 In previous studies, the mechanism of photocross-linking between IBA and NBS was analyzed by mass spectroscopy MS/MS fragmentation, in silico docking minimization, and Western blot.30 The results indicated that a reactive radical driven cross-linking event through exposure to UV light is predominantly occurring between IBA and the Y42 NBS residue present on the antibody light chain.30 Utilizing the UV-NBS method we have successfully functionalized antibodies and Fab fragments with various functional groups including: affinity tags (biotin and thiol), imaging molecules (fluorescein), chemotherapeutics (paclitaxel), and cell penetrating peptides (cyclic iRGD peptide).27−30 Thus, the UV-NBS cross-linking technique provides a universal functionalization method of antibodies that maintains antibody structure and activity. Utilization of UV-NBS cross-linked antibodies for oriented immobilization to detection surfaces also yields higher levels of antibody immobilization and antigen binding sensitivity compared to other commonly used immobilization methods such as physical adsorption, NHS-Biotin, and ε-amine.26,27,30 Fab fragments exhibit numerous advantages over intact antibodies such as smaller size, reduced immunogenicity, and ease of production while still retaining all of the necessary functional components of intact antibodies to allow for specific antigen targeting. These benefits have expanded their utilization in nanoscaled immunosensors, drug delivery systems, detection platforms, and purification systems.6,32,33 For these reasons, numerous Fab fragment modification strategies have been developed, with varying degrees of success.6,34,35 A large



EXPERIMENTAL SECTION

Materials. IBA, biotin N-hydroxysuccinimide ester (Biotin-NHS), N,N-diisopropylethylamine (DIEA), and carbonate-bicarbonate buffers were purchased from Sigma-Aldrich (St. Louis, MO). NovaPEG Rink amide resin was purchased from Novabiochem (Billerica, MA). MonoN-t-boc-amido-dPEG11-amide was purchased from Quanta Biodesign (Powell, OH). Mouse anti-PSA (IgGPSA capture antibody, clone B731M), mouse anti-PSA (IgGPSA capture antibody, clone 5A6), HRP conjugated mouse anti-PSA (det-IgGPSA detection antibody, clone 906), and purified free prostate specific antigen (PSA) were purchased from Meridian Life Science, Inc. (Memphis, TN). Peroxidaseconjugated affiniPure goat anti-mouse IgG and streptavidin-HRP were purchased from Jackson ImmunoResearch (West Grove, PA). Heat shock isolated bovine serum albumin (BSA), Amicon Ultra centrifugal filters (0.5 mL, 10K), and Coomassie R-250 were purchased from EMD Millipore (Billerica, MA). Amplex Red Assay Kit and SeeBlue Plus2 Prestained Standard were purchased from Life Technologies (Grand Island, NY). TEMED, Fab Preparation Kit, maleic anhydride amine reactive 96-well plates, and streptavidin coated 96-well plates were purchased from Thermo Scientific (Rockford, IL). High binding plates were purchased from Corning (Tewksbury, MA). Tris-Gly running buffer, transfer buffer, and tris buffered saline (TBS) were purchased from Boston Bioproducts (Ashland, MA). Synthesis of IBA-EG11-Biotin. The IBA-EG11-amine was synthesized by coupling of IBA to mono-N-t-boc-amido-dPEG11-amine in solution following HBTU activation in DMF and DIEA at room temperature (RT) for 3.5 h while agitating. DMF was evaporated using rotate evaporator. t-boc protecting group was cleaved in a solution of 4% triisopropylsilane, 4% Millipore water, and 92% trifluoroacetic acid (TFA) for 45 min at RT. RP-HPLC system was used for purification of IBA-EG11-amine on a Zorbax C18 column, and MALDI-TOF MS was used for characterization. The calculated exact mass for the IBAEG11-amine (C36H63N3O12) was 729.44 Da; found 730.49 Da. To synthesize IBA-EG11-Biotin, biotin was conjugated to the purified IBAEG11-amine. It was also purified via RP-HPLC on the Zorbax C18 column and characterized with MALDI-TOF MS (calculated exact mass is 955.52 Da; found 956.765) (Figure S1). The yield was 70%, B

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

Langmuir

Figure 1. (A) Schematic representation of Fab fragment cleavage of antibody using Papain enzyme resin. After cleavage, Fab fragments were purified via a Protein A column and their purity was checked with SDS-PAGE. Nucleotide binding site (NBS) is between the heavy and the light chain of the variable region of the Fab fragment. (B) Schematic representation of the UV photo-cross-linking of IBA-EG11-Biotin to Fab fragments. A covalent bond forms between the IBA-EG11-Biotin and Fab fragment at the NBS site upon UV exposure while preserving the Fab fragment’s antigen binding activity. The biotinylated Fab fragments were then incubated on streptavidin coated plate surfaces to facilitate oriented immobilization of the Fab fragments. and the product purity was confirmed using RP-HPLC on the analytical Zorbax C18 column to be >95%. Cleavage of Fab Fragments of Antibody. The buffers of the PSA specific B731M and 5A6 antibodies were exchanged using membrane spin concentrator (10 kDa) in order to remove the preservatives such as sodium azide. The antibodies were further desalted using desalting columns. The amount of 500 μL of antibody (1 mg/mL) was cleaved using Papain enzyme resin for 2 h at 37 °C. The cleaved Fab fragments were purified using a Protein A column. The purity of the Fab fragments was determined by running on a 10% SDS-PAGE gel under reducing conditions, and then the percentage of cleaved product’s purity was analyzed using ImageJ software. Photo-Cross-Linking of IBA-Conjugated Biotin Ligand (IBAEG11-Biotin) to B731M Fab and 5A6 Fab Fragments. Fab fragments were incubated with the IBA-ligand (IBA-EG11-Biotin) for 1 h prior to UV exposure at RT. The UV energy is exposed to the samples in controlled amounts at a wavelength of 254 nm using a Spectroline UV Select Series Cross-linker from Spectronics. Assessing Biotinylation and Antigen Binding Activity of the Fab via ELISA. Fab Biotinylation. The Fab fragments were exposed to UV energy in various amounts in the presence of 300 μM IBAEG11-Biotin. To determine the photo-cross-linking efficiency and the stability of Fab fragments under different amounts of UV energy, Fab fragments were physically absorbed on high binding plates. The plates were washed to remove any unbound components using an automated plate washer (MDS Aquamax 2000) three times with 200 μL of PBS including 0.05% Tween20 at pH 7.4. All plate surfaces were then blocked with BSA blocking buffer (200 μL of 5% BSA in PBS pH 7.4 with 0.1% Tween20) for 30 min at RT. The adsorption of the biotinylated Fab fragments were detected by using a 1:5000 dilution of HRP conjugated streptavidin (1 mg/mL stock concentration) in blocking buffer for 1 h. Antigen Binding Activity. PSA (3 nM) was incubated to high binding 96-well ELISA plates in PBS buffer at pH 7.4 for 1.5 h at RT in order to generate antigen coated ELISA plates. Following the washing and blocking steps, the biotinylated Fab fragments exposed to increasing amounts of UV energies were incubated on the antigen-

coated plates. The wells were then incubated with a 1:5000 dilution of HRP-conjugated streptavidin (1 mg/mL stock concentration) in BSA blocking buffer for 1 h to quantify the total amount of antigen bound Fab fragments. Amplex Red was used as an HRP substrate to form fluorescent product. Fluorescent intensity was measured using a Molecular Devices SpectraMax M5 plate reader (ex. 570 nm, em. 592 nm) for all ELISA assays. Control experiments performed without IBA-EG11Biotin were used as background for biotinylation detection measurements. All experiments were performed in triplicate and the data represents means (±Standard Deviation). UV-NBSBiotin Fab Immobilization Method. Fab fragments (B731M or 5A6) incubated with 200 μM IBA-EG11-Biotin in PBS pH 7.4 for 1 h. The incubated Fab and ligand solution were then exposed to 0.5 J/cm2 of UV energy. The unbound IBA-EG11-Biotin was removed via 4−5 cycles of membrane filtration (0.5 mL, 3 kDa cutoff). After purification, biotinylated Fab was incubated on prewashed streptavidin coated ELISA plates in blocking buffer for 1.5 h at RT. In all cases, unbound Fab was then washed using the automated plate washer. Fab immobilized wells were then blocked with BSA blocking buffer for 1 h to prevent nonspecific adhesion/ interactions. Non-Site-Specific Immobilization Methods. Physical Adsorption Immobilization Method. Fab fragments were incubated directly on high binding ELISA plates in PBS (pH 7.4) for 1.5 h at RT. ε-NH3+ Immobilization Method. Lysine side-chains present on the Fab surface were directly reacted to amine reactive maleic anhydride 96-well plates in PBS buffer at pH 8.0 for 1.5 h at RT. Any remaining reactive sites were then quenched with 50 mM Tris buffer with 100 mM NaCl at pH 8.0 for 30 min at RT. NHS-Biotin Immobilization Method. Fab fragments were biotinylated with NHS-biotin following the manufacturer suggested protocol: Fab fragments were buffer exchanged to sodium carbonate at pH 9.5. Next, 2.2 mg of NHS-D-Biotin was dissolved in 100 μL of DMSO. NHS-D-Biotin was then added into the Fab Fragment solution in a 1:10 ratio with gentle stirring and incubated for 4 h at RT. The unreacted NHS-Biotin was removed via membrane filtration (0.5 mL, C

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

Langmuir

Figure 2. (A) Photo-cross-linking efficiency of IBA-EG11-Biotin to the Fab fragment at the NBS was determined by an ELISA assay. Fab fragments were incubated with IBA-EG11-Biotin linker for 1 h to allow association, and then exposed to increasing UV energies. The biotinylation levels of the Fab fragments were detected by HRP conjugated streptavidin after direct adsorption to high binding plate surfaces (red squares). Further, PSA immobilized surfaces were used to assess antigen binding activity (blue diamonds). (B) Schematic representation of the ELISA set up used for determining the photo-cross-linking efficiency of UV-NBSBiotin method with direct adsorption (biotinylation) and binding to surface immobilized PSA (antigen binding). (C) Effects of UV energy on antigen binding activity and Fab fragment structural stability determined by ELISA utilizing an HRP conjugated anti-Fab detection antibody. The Fab structural stability upon UV exposure was determined with a direct surface adsorption ELISA assay (red squares) and an antigen binding indirect ELISA assay (blue diamonds). (D) Schematic representation of the ELISA set up used for determining the effects of UV energy on Fab fragment structure stability and antigen binding activity. All data represent means (±SD) of triplicate experiments. Determination of Activity Ratio of Fab Immobilized Surfaces. Activity ratio of immobilized Fab surfaces was calculated by taking the ratio of antigen detection efficiency and the initial incubated Fab fragment concentration. A constant Fab fragment concentration (1 nM) was used in all experiments to determine the antigen detection efficiency. To calculate the activity ratio, the fluorescence signals from the antigen detection efficiency experiments at each antigen concentration were divided by the signal of Fab fragment immobilization at 1 nM. Determination of Lower Limit of Detection. The lower limit of detection (LOD) was calculated utilizing the linear regression curve of the antigen detection signal vs antigen concentration plots. The LOD was determined as the PSA concentration for each immobilization method at 3 standard deviations from the mean of the zero PSA standard.41 While theoretically the assay can detect PSA at lower concentrations, this estimate of the LOD is used to ensure a reproducible value for the LOD at a PSA concentration that can be reliably determined taking into consideration the intrinsic variability of each assay by utilizing three standard deviations.

3 kDa). NHS-biotinylated Fab was then incubated on streptavidin coated plate surfaces in BSA blocking buffer for 1.5 h at RT. All surfaces were then washed and blocked using BSA blocking buffer for 30 min. Determination of Total Fab Content of Fab immobilized surfaces. The total amount of surface immobilized Fab for each of the four immobilization techniques was quantified using a 1:2000 dilution of HRP conjugated anti-mouse secondary antibody from goat (1 mg/ mL stock concentration). The initial Fab amounts ranged from 0 to 5 nM (or 0−500 fmol total incubated Fab present in the solution of each well). Amplex Red, a HRP substrate, was used to quantify the amount of Fab via monitoring formation of the fluorescent product with the results reported in RFU (relative fluorescence units). The results represented in the graph of Fab immobilization signal vs initial Fab concentration were fit by linear regression. The slope of the linear regression line was determined to be the antibody immobilization efficiency for each of the immobilization methods. Control experiments performed without immobilized antibody were used as background. Determination of Antigen Detection Efficiency and Assay Sensitivity. The antigen detection efficiencies of Fab fragments were determined by ELISA for all four immobilization methods. Biotinylated Fab immobilized surfaces at 1 nM (100 fmol) of initial Fab were incubated with PSA (0−5 nM) in 100 μL of BSA blocking buffer for 30 min. Unbound PSA was washed and then the wells were incubated with 1 nM of HRP conjugated det-IgGPSA antibody in blocking buffer for 1 h at RT. After Amplex Red addition, the fluorescence product formation was observed at 570 nm excitation and 592 nm emission wavelengths. The resulting antigen detection signal vs antigen concentration plots were fit by linear regression. The slope of the regression line was used to determine the assay sensitivity for each of the immobilization methods. Control experiments performed without antigen and without detection antibody were used as background for antigen detection measurements.



RESULTS AND DISCUSSION The biotinylated Fab fragments were generated by covalent conjugation of IBA-EG11-Biotin ligand to the NBS via UV energy exposure at 254 nm (Figure 1). The amount of UV energy that is needed to achieve photo-cross-linking between the NBS and IBA-Biotin ligand without causing a reduction in antigen activity was determined using an ELISA assay. For this purpose, Fab fragments (3 nM) were incubated with IBA-EG11Biotin (300 μM) for 1 h at RT and were then exposed to increasing amounts of UV energy from 0 to 5 J/cm2 in order to form a covalent cross-linking between IBA and the NBS of Fab fragment. The samples were then incubated on high binding D

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

Langmuir

necessary to allow for photo-cross-linking at the NBS while having minimal impact on antigen binding activity and Fab structural stability. The site-specific photo-cross-linking of IBA to the NBS on the Fab fragment was verified with a Western blot assay. Western blotting was performed with IBA-EG 11-Biotin conjugated Fab fragments under reducing conditions. Streptavidin-HRP was used as a probe to detect which chain the biotinylation occurred at since the NBS is located between the heavy and light chain of the Fab fragment. After Protein A column purification of the cleaved Fab fragment, an SDS-PAGE gel was performed under reducing conditions both to verify the purity of Fab and to compare the molecular weights of the intact antibody fragments with the Fab fragments. The purity of the Fab fragments was calculated by ImageJ as >85% by band intensity measurements (Figure S3). On the blotted film, biotinylation was predominantly observed on the light chain of the Fab fragment, with the yield of conjugation being dependent on the amount of UV exposure (Figure 3). The

ELISA plates in coating buffer (pH 9.6) for 1 h. StreptavidinHRP was used to probe for biotinylation of the Fab fragments utilizing fluorescent product formation from Amplex Red for quantification purposes (Figure 2B). IBA-Biotin photo-crosslinking efficiency increased with increasing UV energy and reached a plateau at 0.5 J/cm2 (Figure 2A). This plateau illustrates that UV energies >0.5 J/cm2 were sufficient to achieve a covalent biotinylation of nearly all Fab fragment nucleotide-binding sites. Additionally, antigen recognition capabilities of the biotinylated Fab fragments were also assessed via an ELISA assay utilizing PSA coated plate surfaces (Figure 2B). Similar results were observed compared to the biotinylation assay demonstrating increased photo-cross-linking efficiency from 0−0.5 J/cm2 and a plateau occurring at 0.5−2 J/ cm2 validating covalent conjugation between the IBA-Biotin and NBS as well as antigen binding activity. A reduced signal intensity at high UV energies (2−5 J/cm2) was observed and is indicative of damage to the antigen binding site preventing the Fab fragment from binding the surface immobilized PSA (Figure 2A). Consequently, we determined 0.5−2 J/cm2 to be an effective UV exposure range that can be used for efficient photo-cross-linking of IBA-Biotin to Fab fragments, without reducing antigen binding activity. The IBA-EG11-Biotin concentration is another critical factor that affects the photo-cross-linking efficiency as the percent of NBS pockets that is bound to IBA prior to UV exposure governs the yield of the reaction. For this reason, we sought to determine the saturating concentration of the IBA-ligand with an ELISA assay by incubating varying concentrations of IBABiotin against a constant concentration of Fab (3 nM) and then exposing the mixture to 0.5 J/cm2 of UV energy. Following the UV exposure, the IBA-EG11-Biotin conjugated Fab fragments were directly adsorbed on high binding plates in coating buffer for 1 h. The biotinylation of the Fab fragments were determined utilizing Streptavidin-HRP. A higher photo-crosslinking efficiency was achieved at higher ligand concentrations reaching a plateau at ∼300 μM of IBA-EG11-Biotin (Figure S2). The UV energies used to generate site-specific photo-crosslinking of IBA-ligand to the NBS in the Fab fragment may potentially have damaging effects on the antigen binding site as well as disrupt the Fab structure. To ensure that the antigen binding activity and structure of the Fab are preserved during the photo-cross-linking reaction, the effect of UV energy on the Fab fragment was explored with ELISA assays. In these assays, Fab fragments (3 nM) were incubated with a saturating amount of IBA-EG11-Biotin (300 μM) for 1 h prior to UV exposure. Increasing amounts of UV energy (0−5 J/cm2) were used to initiate the photo-cross-linking reaction. The cross-linking products, biotinylated Fab samples, were then incubated and adsorbed on high binding ELISA plates for 1 h. The Fab fragments were then detected using an HRP conjugated antimouse IgG secondary antibody to probe for conservation of the general Fab structure post UV exposure (Figure 2C and D). In a separate experiment, we evaluated the structural stability of the same biotinylated Fab samples by measuring their antigen binding activity using PSA coated ELISA surfaces (Figure 2D). According to either method, there was no observable reduction in antigen binding activity or disruption to the Fab structure up to UV energies of 0.5 J/cm2 (Figure 2C). In both assays, a reduction in the signal intensity at high UV energies (>0.5 J/ cm2) was observed that is indicative of damage to the Fab fragment structure and antigen binding activity. Consequently, these results demonstrate that UV energies ≤0.5 J/cm2 are

Figure 3. Determination of site-specific conjugation of IBA-EG11Biotin to the Fab fragment NBS via SDS-PAGE and Western blot. The full-length IgGPSA antibody (B731M) and Fab fragment were run on 10% SDS-PAGE gels under reducing conditions. The biotinylated Fab fragments exposed to increasing amounts of UV energy were also run on 10% SDS-PAGE gel and the results were compared to the fulllength antibody. The light chain of the B731M matched to the upper band of the Fab fragments indicating that the upper band is the light chain and the lower band is the heavy chain of the Fab fragments. To show the specific location of biotinylation on the Fab fragment, a Western blot assay was performed using HRP-streptavidin to probe for biotinylation. The blotted film reveals that the biotinylation of Fab fragments at the NBS is occurring dominantly at the light chain.

dominant biotinylation band precisely matched the light chain of the Fab as indicated by the apparent molecular weight marker on the SDS-PAGE compared to the intact B731M antibody light chain. After identifying the optimal conditions for photo-crosslinking, we examined the immobilization and antigen recognition efficiency of Fab fragments biotinylated using the UV-NBSBiotin method. The performance of UV-NBSBiotin sitespecific immobilization method was compared to three other commonly used antibody immobilization methods: lysine side chain immobilization (ε-NH3+), NHS-Biotin, and physical adsorption. Essentially, these immobilization techniques rely on adhesion to surface through reactive amine chemistry at arbitrary lysine residues, biotinylation using NHS chemistry at arbitrary lysine residues, or nonspecific hydrophobic interactions, respectively, leading to highly disordered immobilizaE

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

Langmuir

Figure 4. Comparison of the UV-NBSBiotin method to ε-NH3+, NHS-Biotin, and physical adsorption immobilization methods via ELISA assay. Fab fragments were immobilized to 96-well plates using all four methods. (A) Level of immobilization was determined using increasing concentrations of Fab fragments, which were detected by a HRP conjugated secondary antibody. (B) Schematic representation of the ELISA used for each method determining the immobilization of biotinylated Fab fragments. (C) Antigen detection sensitivity was determined using increasing concentrations of PSA (0−1 nM), and then a detection antibody (det-IgGPSA) was added and quantified by an HRP conjugated anti-Fc antibody as a reporter. (D) Schematic representation of the ELISA used for each method determining antigen binding efficiency of biotinylated Fab fragments. All data represent means (±SD) of triplicate experiments.

tion on surfaces. We compared the UV-NBSBiotin method to these three commonly used immobilization methods in terms of Fab fragment immobilization efficiency, antigen recognition sensitivity, and activity ratio. To investigate the immobilization efficiency of each technique, varying concentrations of Fab fragments (0−5 nM) were incubated in 96-well ELISA plates. Following the blocking and washing steps, the immobilization efficiency was determined by using HRP conjugated secondary antibody (Figure 4B). The slope of the linear regression fit was used as a metric to assess the Fab fragment immobilization efficiency. The UV-NBSBiotin method demonstrated 2-, 3-, and 71-fold higher immobilization efficiency compared to ε-NH3+, physical adsorption, and NHS-Biotin methods, respectively (Figure 4A). The results demonstrate that the UV-NBSBiotin method was the most efficient immobilization technique that generated the highest level of Fab fragment immobilization on the detection surface. Next, we investigated the antigen binding sensitivity of the immobilized Fab fragments using a different ELISA setup. For this purpose, we incubated varying concentrations of PSA over detection surfaces that were prepared using the four different immobilization techniques. An HRP conjugated det-IgGPSA antibody was then used to detect the PSA in a sandwich type

ELISA assay (Figure 4D). The antigen binding sensitivity of the Fab fragments were determined with the coefficient of the natural log from the regression line. The UV-NBSBiotin method demonstrated the highest sensitivity with a coefficient of 4064.6 (R2 = 0.97), which indicates 29-, 114-, and 432-fold higher sensitivity than ε-NH3+, NHS-Biotin, and physical adsorption methods, respectively (Figure 4C). These results demonstrate that the UV-NBSBiotin method delivers not only the highest level of Fab fragment immobilization to surfaces but also the highest retained activity of immobilized Fab fragments on the surface. To further emphasize the antigen binding activity of Fab fragments after surface immobilization, we calculated the relative activity ratio for all four methods by using the ratio of the signal from the antigen binding intensity to the signal from the surface immobilized Fab fragment (Figure S4). Predictably, the UV-NBSBiotin method showed the highest relative activity ratio of all four methods, representing 6-, 2-, and 193-fold enhancement in relative Fab fragment activity compared to ε-NH3+, NHS-Biotin, and physical adsorption methods, respectively at 0.5 nM (50 fmol) antigen concentrations (Table S1). It is worth mentioning that, while the NHS-Biotin method appears to achieve high activity, the coefficient of determination (R2) values of the immobilization F

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

Langmuir

demonstrate that the UV-NBSBiotin method can be utilized to significantly enhance the level of Fab fragment immobilization providing for higher antigen detection efficiency and improved antigen binding sensitivity compared to three other commonly used immobilization methods: ε-NH3+, physical adsorption, and NHS-Biotin. Comparing the UV-NBSBiotin immobilization of intact IgGPSA to immobilized Fab fragments, as presented here, demonstrates a very similar result with an improved antigen detection efficiency and LOD utilizing Fab fragments in place of intact antibody. This comparison can also be used to demonstrate that the antigen binding activity of the Fab fragments are preserved post removal of the Fc domain as well as post UV exposure and immobilization processes. This study demonstrates an optimized method to sitespecifically photo-cross-link IBA-Biotin to the conserved NBS of Fab fragments for oriented Fab immobilization to diagnostic surfaces. This covalent conjugation technique preserves the Fab structure, antigen binding activity, affinity and sensitivity. The UV-NBS method can also be utilized for site-specific functionalization of Fab fragments with various other IBAlinked functional ligands such as fluorescent probes, affinity tags, chemotherapeutics, and peptides. While the Fab fragments utilized in this study were derived from full length antibodies, we anticipate that other antibody fragments such as Fab2, Fab′ and ScFv that contain an NBS site can also be functionalized utilizing UV-NBS method with a variety of molecules carrying an IBA-moiety. Hence, site-specific modification of Fab fragments via the UV-NBSBiotin method has the potential to be applied across numerous biosensor and immunosensor platforms as well as in a vast array of therapeutic applications. Immobilization of functionalized Fab fragments via UV-NBS method can be applied to other platforms such as functionalizing nanoparticles, which is currently an ongoing project in our laboratories, to be used in enhanced biosensing applications. The results presented in this study establish the UV-NBSBiotin method as a universal, practical, gentle, and reproducible method for site-specific functionalization of Fab fragments as well as intact antibodies for use in oriented immobilization to detection surfaces.

efficiency and antigen detection efficiency indicate a very inconsistent detection system. Taken together, the biotinylated Fab fragments produced via the UV-NBSBiotin method preserves antigen binding activity through site-specific functionalization and subsequent oriented immobilization of the Fab fragments. Antigen binding activity was significantly reduced when utilizing the other three immobilization methods due to either damage to the Fab fragments or as a result of nonorientated immobilization. Determination of the optimum conditions for Fab fragment photo-cross-linking with IBA-EG11-Biotin ligand via the UVNBSBiotin method was repeated using Fab fragments produced from another mouse IgGPSA (clone: 5A6) capture antibody (Figure S5). Optimum conditions for the 5A6 Fab fragments photo-cross-linking reaction show similar trends compared to photo-cross-linking of the B731M Fab fragment. It is important to note that the antigen binding site of the 5A6 Fab fragment were slightly more sensitive to UV energy than the B731M Fab fragment. Furthermore, comparison experiments of UVNBSBiotin method with the three commonly used immobilization techniques were also performed for the 5A6 Fab fragment (Figure S6). While the B731M clone was superior to the 5A6 clone in PSA antigen detection, the results compared similarly as the UV-NBSBiotin method still demonstrated the highest level of Fab fragment immobilization, antigen binding sensitivity and activity ratio compared to the ε-NH3+, physical adsorption and NHS-Biotin methods. The UV-NBSBiotin method represents 2.5-, 1.3-, and 92-fold enhancement in the relative Fab fragment activity compared to ε-NH3+, NHS-Biotin, and physical adsorption methods, respectively, at 0.5 nM (50 fmol) antigen concentrations (Table S1). The consistency across these results for the two different Fab fragments is indicative of the reliability of UV-NBSBiotin method. The limit of detection (LOD) for PSA utilizing the immobilized Fab fragments via the UV-NBSBiotin method was significantly lower at ∼0.4 pM (∼0.04 fmol) when using the B731M clone of IgGPSA Fab fragments (Table S2). This is >3 orders of magnitude lower than the LOD of ε-NH3+, NHSBiotin, and physical adsorption methods and is a direct result of achieving more active immobilized Fab on the detection surface than any other immobilization technique tested (Table S2). While the overall assay conditions for utilizing the UV-NBSBiotin method were optimized independently for the intact IgGPSA and the Fab fragment IgGPSA, the LOD of the Fab fragments is significantly improved when compared to that of the intact antibody of ∼0.02 nM (∼2 fmol), that was reported in our previous publication.27 Immobilization of Fab fragments via the UV-NBSBiotin method also demonstrates a comparable antigen detection sensitivity at a 1.3-fold improvement compared to previous results with intact antibody (regression line equation: y = 3115ln(x) + 3508.2).27 This result indicates that oriented immobilization of Fab fragment utilizing the UV-NBSBiotin method can provide higher antigen binding capabilities and enhanced antigen detection sensitivity compared to intact antibody.



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b01734. Molecular structure and characterization of IBA-EG11Biotin molecule; impurity analysis of Fab fragments; impurity analysis of Fab fragments; activity ratio of Fab fragments after immobilization; characterization of 5A6 antibody Fab fragment photo-cross-linking reaction; comparison of UV-NBS Fab fragment immobilization method; activity ratio of Fab immobilized surfaces; LOD values (PDF)





CONCLUSIONS Fab fragments, with their smaller size and simpler production methods relative to intact antibodies, are well suited for use in various diagnostic and therapeutic applications. Oriented immobilization of Fab fragments on detection surfaces is, however, a challenging problem that has limited their implementation in many applications. In this study, we

AUTHOR INFORMATION

Corresponding Author

*Mailing address: Department of Chemical and Biomolecular Engineering, Department of Chemistry and Biochemistry, Mike and Josie Harper Cancer Research Institute, Center for Rare & Neglected Diseases, Advanced Diagnostics & Therapeutics Initiative, 171 Fitzpatrick Hall, Notre Dame, IN 46556-5637, G

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(17) Vijayendran, R. A.; Leckband, D. E. A Quantitative Assessment of Heterogeneity for Surface-Immobilized Proteins. Anal. Chem. 2001, 73, 471−480. (18) Parolo, C.; de la Escosura-Muñiz, A.; Polo, E.; Grazú, V.; de la Fuente; Jesús, M.; Merkoçi, A. Design, Preparation, and Evaluation of a Fixed-Orientation Antibody/Gold-Nanoparticle Conjugate as an Immunosensing Label. ACS Appl. Mater. Interfaces 2013, 5, 10753− 10759. (19) Firestone, M. A.; Shank, M. L.; Sligar, S. G.; Bohn, P. W. Film Architecture in Biomolecular Assemblies. Effect of Linker on the Orientation of Genetically Engineered Surface-Bound Proteins. J. Am. Chem. Soc. 1996, 118, 9033−9041. (20) Rao, S. V.; Anderson, K. W.; Bachas, L. G. Oriented Immobilization of Proteins. Microchim. Acta 1998, 128, 127−143. (21) Moreau, T.; Faye, C.; Baque, M.; Desvignes, I.; Coussot, G.; Pascal, R.; Vandenabeele-Trambouze, O. Antibody-Based Surfaces: Rapid Characterization using Two Complementary Colorimetric Assays. Anal. Chim. Acta 2011, 706, 354−360. (22) Patel, N.; Davies, M. C.; Hartshorne, M.; Heaton, R. J.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M. Immobilization of Protein Molecules Onto Homogeneous and Mixed Carboxylate-Terminated Self-Assembled Monolayers. Langmuir 1997, 13, 6485−6490. (23) Vikholm-Lundin, I.; Albers, W. M. Site-Directed Immobilisation of Antibody Fragments for Detection of C-Reactive Protein. Biosens. Bioelectron. 2006, 21, 1141−1148. (24) Wimalasena, R. L.; Wilson, G. S. Factors Affecting the Specific Activity of Immobilized Antibodies and their Biologically Active Fragments. J. Chromatogr., Biomed. Appl. 1991, 572, 85−102. (25) Orr, V.; Zhong, L.; Moo-Young, M.; Chou, C. P. Recent Advances in Bioprocessing Application of Membrane Chromatography. Biotechnol. Adv. 2013, 31, 450−465. (26) Alves, N. J.; Kiziltepe, T.; Bilgicer, B. Oriented Surface Immobilization of Antibodies at the Conserved Nucleotide Binding Site for Enhanced Antigen Detection. Langmuir 2012, 28, 9640−9648. (27) Alves, N. J.; Mustafaoglu, N.; Bilgicer, B. Oriented Antibody Immobilization by Site-Specific UV Photocrosslinking of Biotin at the Conserved Nucleotide Binding Site for Enhanced Antigen Detection. Biosens. Bioelectron. 2013, 49, 387−393. (28) Alves, N. J.; Mustafaoglu, N.; Bilgicer, B. Conjugation of a Reactive Thiol at the Nucleotide Binding Site for Site-Specific Antibody Functionalization. Bioconjugate Chem. 2014, 25, 1198−1202. (29) Mustafaoglu, N.; Alves, N. J.; Bilgicer, B. Site-specific Fab Fragment Biotinylation at the Conserved Nucleotide Binding Site for Enhanced Ebola Detection. Biotechnol. Bioeng. 2015, 112, 1327−1334. (30) Alves, N. J.; Champion, M. M.; Stefanick, J. F.; Handlogten, M. W.; Moustakas, D. T.; Shi, Y.; Shaw, B. F.; Navari, R. M.; Kiziltepe, T.; Bilgicer, B. Selective Photocrosslinking of Functional Ligands to Antibodies Via the Conserved Nucleotide Binding Site. Biomaterials 2013, 34, 5700−5710. (31) Handlogten, M. W.; Kiziltepe, T.; Moustakas, D. T.; Bilgicer, B. Design of a Heterobivalent Ligand to Inhibit IgE Clustering on Mast Cells. Chem. Biol. 2011, 18, 1179−1188. (32) Sharma, H.; Mutharasan, R. Half Antibody Fragments Improve Biosensor Sensitivity without Loss of Selectivity. Anal. Chem. 2013, 85, 2472−2477. (33) Holliger, P.; Hudson, P. Engineered Antibody Fragments and the Rise of Single Domains. Nat. Biotechnol. 2005, 23, 1126−1136. (34) Humphreys, D. P.; Heywood, S. P.; Henry, A.; Ait-Lhadj, L.; Antoniw, P.; Palframan, R.; Greenslade, K. J.; Carrington, B.; Reeks, D. G.; Bowering, L. C.; West, S.; Brand, H. a. Alternative Antibody Fab′ Fragment PEGylation Strategies: Combination of Strong Reducing Agents, Disruption of the Interchain Disulphide Bond and Disulphide Engineering. Protein Eng., Des. Sel. 2007, 20, 227−234. (35) Lee, W.; Oh, B.; Lee, W.; Choi, J. Immobilization of Antibody Fragment for Immunosensor Application Based on Surface Plasmon Resonance. Colloids Surf., B 2005, 40, 143−148. (36) Vikholm, I. Self-Assembly of Antibody Fragments and Polymers Onto Gold for Immunosensing. Sens. Actuators, B 2005, 106, 311−316.

USA. E-mail: [email protected]. Fax: 574-631-8366. Tel: 574631-1429. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■ Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

ACKNOWLEDGMENTS This work was supported by the NSF (Grant Award Number CBET-1263713). The authors thank the Notre Dame Mass Spectrometry and Proteomics Facility for usage of mass analysis instrumentation.



REFERENCES

(1) Lu, B.; Smyth, M. R.; O'Kennedy, R. Tutorial Review. Oriented Immobilization of Antibodies and its Applications in Immunoassays and Immunosensors. Analyst 1996, 121, 29R−32R. (2) Gray, J. J. The Interaction of Proteins with Solid Surfaces. Curr. Opin. Struct. Biol. 2004, 14, 110−115. (3) Makaraviciute, A.; Ramanaviciene, A. Site-Directed Antibody Immobilization Techniques for Immunosensors. Biosens. Bioelectron. 2013, 50, 460−471. (4) Ferreira, N. S.; Sales, M. G. F. Disposable Immunosensor using a Simple Method for Oriented Antibody Immobilization for Label-Free Real-Time Detection of an Oxidative Stress Biomarker Implicated in Cancer Diseases. Biosens. Bioelectron. 2014, 53, 193−199. (5) Zhang, W.; Patel, K.; Schexnider, A.; Banu, S.; Radadia, A. D. Nanostructuring of Biosensing Electrodes with Nanodiamonds for Antibody Immobilization. ACS Nano 2014, 8, 1419−1428. (6) Saerens, D.; Huang, L.; Bonroy, K.; Muyldermans, S. Antibody Fragments as Probe in Biosensor Development. Sensors 2008, 8, 4669−4686. (7) Lu, B.; Xie, J.; Lu, C.; Wu, C.; Wei, Y. Oriented Immobilization of Fab′ Fragments on Silica Surfaces. Anal. Chem. 1995, 67, 83−87. (8) Wan, Y.; Su, Y.; Zhu, X.; Liu, G.; Fan, C. Development of Electrochemical Immunosensors Towards Point of Care Diagnostics. Biosens. Bioelectron. 2013, 47, 1−11. (9) Trilling, A. K.; Beekwilder, J.; Zuilhof, H. Antibody Orientation on Biosensor Surfaces: A Minireview. Analyst 2013, 138, 1619−1627. (10) Baniukevic, J.; Kirlyte, J.; Ramanavicius, A.; Ramanaviciene, A. Application of Oriented and Random Antibody Immobilization Methods in Immunosensor Design. Sens. Actuators, B 2013, 189, 217−223. (11) Buijs, J.; Norde, W.; Lichtenbelt, J. W. T. Changes in the Secondary Structure of Adsorbed IgG and F (Ab′) 2 Studied by FTIR Spectroscopy. Langmuir 1996, 12, 1605−1613. (12) Rao, J. H.; Lahiri, J.; Isaacs, L.; Weis, R. M.; Whitesides, G. M. A Trivalent System from Vancomycin Center Dot D-Ala-D-Ala with Higher Affinity than Avidin Center Dot Biotin. Science 1998, 280, 708−711. (13) Lee, J. E.; Seo, J. H.; Kim, C. S.; Kwon, Y.; Ha, J. H.; Choi, S. S.; Cha, H. J. A Comparative Study on Antibody Immobilization Strategies Onto Solid Surface. Korean J. Chem. Eng. 2013, 30, 1934− 1938. (14) Bonroy, K.; Frederix, F.; Reekmans, G.; Dewolf, E.; De Palma, R.; Borghs, G.; Declerck, P.; Goddeeris, B. Comparison of Random and Oriented Immobilisation of Antibody Fragments on Mixed SelfAssembled Monolayers. J. Immunol. Methods 2006, 312, 167−181. (15) Kausaite-Minkstimiene, A.; Ramanaviciene, A.; Kirlyte, J.; Ramanavicius, A. Comparative Study of Random and Oriented Antibody Immobilization Techniques on the Binding Capacity of Immunosensor. Anal. Chem. 2010, 82, 6401−6408. (16) Spitznagel, T. M.; Jacobs, J. W.; Clark, D. S. Random and SiteSpecific Immobilization of Catalytic Antibodies. Enzyme Microb. Technol. 1993, 15, 916−921. H

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.langmuir.5b01734

(37) El Khoury, G.; Lowe, C. R. A Biomimetic Protein G Affinity Adsorbent: An Ugi Ligand for Immunoglobulins and Fab Fragments Based on the Third IgG-binding Domain of Protein G. J. Mol. Recognit. 2013, 26, 190−200. (38) Billah, M. M.; Hodges, C. S.; Hays, H. C.; Millner, P. Directed Immobilization of Reduced Antibody Fragments Onto a Novel SAM on Gold for Myoglobin Impedance Immunosensing. Bioelectrochemistry 2010, 80, 49−54. (39) Lilja, H. Biology of Prostate-Specific Antigen. Urology 2003, 62, 27−33. (40) Velonas, V. M.; Woo, H. H.; Remedios, C. G. d.; Assinder, S. J. Current Status of Biomarkers for Prostate Cancer. Int. J. Mol. Sci. 2013, 14, 11034−11060. (41) Han, H. J.; Kannan, R. M.; Wang, S.; Mao, G.; Kusanovic, J. P.; Romero, R. Multifunctional Dendrimer-Templated Antibody Presentation on Biosensor Surfaces for Improved Biomarker Detection. Adv. Funct. Mater. 2010, 20, 409−421.

I

DOI: 10.1021/acs.langmuir.5b01734 Langmuir XXXX, XXX, XXX−XXX