Stabilization of Surface-bound Antibodies for ELISA based on a Reversable Zeolitic Imidazolate Framework-8 Coating

Abstract

Immunoassays typically must be stored under refrigerated conditions because antibodies, after being immobilized to solid surfaces, tend to lose their recognition capabilities to target antigens under non-refrigerated conditions. This requirement hinders application of immunoassays in resource-limited settings including rural clinics in tropical regions, disaster struck areas, and low-income countries, where refrigeration may not be feasible. In this work, a facile approach based on a reversable zeolitic imidazolate framework-8 (ZIF-8) coating is introduced to stabilize surface-bound antibodies on enzyme-linked immunosorbent assay (ELISA) plates under non-refrigerated conditions. Using a sandwich ELISA for the detection of neutrophil gelatinase-associated lipocalin (NGAL), a urine biomarker for acute kidney injury, as a model system, ZIF-8 is demonstrated to be able to uniformly coat the surface-bound anti-NGAL IgG, and stabilize the dynamic range and detection sensitivity of the assay after storage at an elevated temperature (50°C) for at least 4 weeks. The stabilization efficacy of the ZIF-8 coating is comparable to the current “gold standard” refrigeration approach, and superior to the commonly used sucrose coating method. This approach will greatly improve the shelf-life and stability of antibody-coated ELISAs and other types of assays which utilize surface-bound antibodies, thus extending biomedical research and medical diagnostics to resource-limited settings.

Introduction

Biomedical research and clinical diagnostics rely on the availability of high-quality immunoassays such as enzyme-linked immunosorbent assays (ELISAs), chemiluminescent microparticle immunoassays (CLIAs), and lateral flow assays (LFAs). The structural integrity and functionality of biomolecule reagents used to construct these assays are critical to the assay performance including sensitivity and specificity. However, due to the poor stability of these biomolecules (including antibodies, antigens, and enzymes) at ambient and elevated temperatures, they are prone to lose their structure and functionality before use.[1–3] Antibodies, after being immobilized to solid surfaces (e.g., on ELISA plates, micro- or nano- particle surfaces), are especially prone to lose their affinity and specificity to antigens under non-refrigerated conditions.[4–8] Hence, immunoassays should be maintained under refrigerated conditions to preserve their biofunctionality (i.e., analyte detection capability). In the case of ELISA, an extensively used immunoassay, manufacturers often provide “capture antibody” pre-coated microplates to shorten the total assay time (i.e., saving 12 h for antibody coating), and require these pre-coated plates to be stored under refrigeration (2–8°C).[9, 10] A similar requirement exists for CLIAs, in which antibody-coated microparticles must be stored at 2–8°C.[11] Thus a temperature-controlled supply chain, termed the “cold chain”, is required during storage, transport and handling of these immunoassays. Apart from causing a financial and environmental burden, such “cold chain” systems may not be feasible in resource-limited settings such as rural clinics in tropical regions, disaster struck areas, and low-income countries with a high burden of disease and poor infrastructure, leading to higher mortality and morbidity rates.[12] To improve the accessibility to such assays in resource-limited settings, it is imperative to develop a refrigeration-free technology to provide stable and reliable bioassays.

Zeolitic imidazolate framework-8 (ZIF-8), a typical type of metal-organic framework (MOF) material, has received increased scientific and technological interest owing to its unique capacity to encapsulate and stabilize proteins.[13–18] When adding ZIF-8 precursors (zinc ion and 2-methyl imidazole) into an aqueous solution containing protein molecules, ZIF-8 crystals tend to form around the proteins and encapsulate them within rigid crystal frameworks, a process analogous to biomineralization.[15, 19] As a result, the proteins are isolated from their ambient environment, and the protein structure and function are preserved and stabilized under normally denaturing conditions.[20, 21] As a novel refrigeration-free approach, ZIF-8 has been shown to encapsulate and stabilize a variety of free (unbound) proteins including protein biomarkers in urine and blood,[22] and protein therapeutics such as insulin and monoclonal antibodies,[23, 24] therefore enabling the handling, transport, and storage of biospecimens and biologics under ambient and elevated temperatures. For surface-bound proteins, it was recently demonstrated that ZIF-8 can form a protective coating on antibody-nanoparticle conjugates and stabilize them at room and elevated temperatures.[25, 26] For example, a plasmonic biosensor was stabilized under 40°C for 1 week, in which antibodies were conjugated onto gold nanoparticle surfaces,[25] similar to a lateral flow assay.[27, 28] The protective ZIF-8 coating can be completely removed within 5 min before use of the biosensor. This approach demonstrated stabilization of antibodies bound to nanoparticle surfaces, but the applicability of this approach to the stabilization of surface-bound antibodies on ELISA plates has yet to be reported. Considering the broader application of ELISAs and the increased need for antibody pre-coated ELISA plates, as well as the large number of antibodies on the plate surface versus the small number of the antibodies on a nanoparticle surface, it would be extremely valuable to investigate if and how the ZIF-8 approach could be used for stabilization of surface-bound antibodies on ELISA plates.

In this work, we report a facile approach based on a reversable ZIF-8 coating to stabilize surface-bound antibodies on ELISA plates under non-refrigerated conditions. We show that a ZIF-8 coating can preserve recognition capabilities of antibodies that have been coated onto ELISA plates after being subjected to elevated temperatures. Just before use, a quick washing step completely removes the protective ZIF-8 coating without disrupting the ELISA workflow (Figure 1). To demonstrate this approach, we use a sandwich ELISA for the detection of neutrophil gelatinase-associated lipocalin (NGAL), a urine biomarker for acute kidney injury, as our model system. The anti-NGAL antibody pre-coated plates with and without ZIF-8 coatings are stored at 50°C for 4 weeks (i.e., the “stress test”), and are compared with the antibody-coated plates stored at 4°C for 4 weeks in terms of their dynamic range and detection sensitivity. We also compare the ZIF-8 coating approach with another commonly used method, a sucrose coating. We demonstrate that the ZIF-8 coating can preserve the recognition capabilities of surface-bound antibodies on the ELISA plates after storage at 50°C for 4 weeks (i.e., maintaining their dynamic range and detection sensitivity in detection of urine NGAL), and the stabilization efficacy of the ZIF-8 approach is superior to the sucrose method. Furthermore, we employ atomic force microscopy (AFM) to not only provide nanoscale morphologies of ELISA plate surfaces after each coating step, but more importantly, the use of AFM-based force spectroscopy offers a direct means to assess antibody recognition capabilities on ELISA plates based on recognition binding force and frequency to their target antigens. Overall, we believe this facile approach will greatly improve the shelf-life and stability of antibody-coated ELISAs, thus extending biomedical research and medical diagnostics to resource-limited settings and underserved populations by enabling reliable bioassays in regions where they are currently inaccessible.

Figure 1. Schematic illustration.

The concept of using a ZIF-8 coating to stabilize the surface-bound antibodies on ELISA plates under non-refrigerated conditions. Just before using the plate for sample analysis, a quick washing step (1 min) completely removes the protective coating without disrupting the ELISA workflow. For stabilization of surface-bound antibodies, the ZIF-8 approach (middle line in the figure) offers comparable efficacy to the current “gold standard” refrigeration storage (upper line). In contrast, without a coating after storage under non-refrigerated conditions, the surface-bound antibodies lose their recognition capabilities to antigens (lower line). This ZIF-8 approach will greatly improve the shelf-life and stability of antibody pre-coated ELISA plates under non-refrigerated conditions, thus extending biomedical research and medical diagnostics to resource-limited settings and underserved populations where refrigeration or the “cold chain” may not be feasible.

Results and Discussion

As a proof-of-concept, a sandwich ELISA for the detection of neutrophil gelatinase-associated lipocalin (NGAL) was employed as a model system. Urine NGAL is a reliable non-invasive biomarker for acute kidney injury as its level increases by two to three orders of magnitude (100–1000 times) during acute kidney injury.[29] Considering that acute kidney injury is a common condition with a high risk of death that often occurs in disaster struck areas,[30] developing tropical countries,[31] and battle fields,[32] it is of great clinical importance to enable the use of stable and reliable NGAL assays to accurately measure the urine NGAL levels in these resource-limited settings. As with most sandwich ELISAs, a monoclonal “capture antibody”, anti-NGAL IgG, is first coated onto the 96-well plate and then blocked with bovine serum albumin (BSA). Since this initial step (the “capture antibody” coating) requires 12 h (i.e., overnight), manufacturers often provide “capture antibody” pre-coated ELISA plates (blocked with BSA) to shorten the total assay time to 3–4 h, and require these antibody pre-coated plates to be stored at 2–8°C. In this work, we show that a protective and removable ZIF-8 coating on the top of the anti-NGAL IgG pre-coated plate can preserve the recognition capabilities of these surface-bound antibodies under ambient and elevated temperatures.

To demonstrate this approach, we first investigated the feasibility and uniformity of the ZIF-8 coating on the antibody pre-coated plate, and verified if the coating could be removed just before using the assay for sample analysis. Herein, we carefully characterized the surface of the plate after each step. Atomic force microscopy (AFM) was used to examine the surface morphologies of the plate well after each step (Figure 2A), including: I. Bare well; II. The well after anti-NGAL IgG coating; III. The antibody coated well after BSA blocking; IV. The well after ZIF-8 coating; and V. The well after removing the ZIF-8 coating. To mimic clinical practice, a 96-well plate was coated and washed as usual, while the well bottom after each step was cut and immobilized to a glass slide for AFM imaging in air (Figure S1). To our knowledge, there is no study to date attempting to characterize the 96-well bottom using AFM in this stepwise manner, probably due to the difficulty in reaching the bottom of the plate well. We believe the simple characterization strategy developed here (i.e., coating plates as usual, whereas cutting the bottom of the well after each step for nanoscale characterization) will be very helpful for understanding the surface coating of 96-well plates, especially when developing new ELISAs. The AFM images after each step are shown in Figure 2B (2×2 μm²) and 2C (500×500 nm²). After coating the bare polystyrene well (column I) with anti-NGAL IgG, the well surface exhibited isolated but well-distributed particulate features (column II), indicating the successful coating of the “capture antibodies” on the plate surface. The height profile analysis of the particles (Figure S2) suggests that the antibodies formed a monolayer on the polystyrene well surface, since the heights (~4–8 nm) are consistent with the dimension of the single IgG (14.5 nm×8.5 nm×4 nm).[33] After blocking the antibody-coated plate with BSA (column III), the well surface showed a decrease in empty areas (i.e., the bare well), confirming that BSA blocking minimizes nonspecific binding. Following the BSA blocking, the precursor solution of ZIF-8 (a mixture of 2-methylimidazole and zinc acetate) was added into the well, incubated for 1 h, and pipetted out. The AFM images of the well surface then displayed dense grainy features (column IV), distinct from the surface morphology before coating (column III), indicating the formation of a uniform coating with a complete coverage of the surface-bound antibodies. Apart from the facile coating procedure, it is also important to ensure the coating can be removed prior to use of the plate for sample analysis. The AFM images indicated that the coating can be completely removed by washing the well with pure water (pH=6) for 1 min (column V). This is because the coordination between zinc ion and 2-methylimidazole can be broken by the slightly acidic solution.[15] To further ascertain that the coating was completely removed, a zinc assay kit was used to determine the zinc concentration in the washing solution. The wells with ZIF-8 coatings were washed twice by pH=6 water for 1 min, and the washing solutions were collected and analyzed for zinc concentrations, and compared to a control solution (pH=6 water after 1 min incubation with wells without ZIF-8 coating). The results indicated that the second-time washing solution only contained less than 1% of zinc, compared to zinc in the first-time washing solution, suggesting that almost no zinc residues would be present in the subsequent sample analysis (Figure S3). To confirm the formation of ZIF-8 coating on the plate well surface, Raman spectroscopy and X-ray diffraction (XRD) were used to characterize the well surface before and after the coating. The Raman spectrum obtained from the well surface before coating showed a typical broad amide I band between 1630 and 1690 cm⁻¹,[34] whereas the well surface after coating exhibited not only the amide I band of the surface-bound proteins (antibodies and BSA), but also the characteristic bands of 2-methylimidazole at 1147, 1184, 1384, 1466 and 1509 cm⁻¹, which can be attributed to C−N stretching, C−N stretching plus N−H wagging, CH₃ bending, C−H wagging, and C-C stretching plus N−H wagging, respectively (Figure 2D).[35] XRD measurements also demonstrated the formation of a ZIF-8 coating on the well surface (Figure 2E). The XRD patterns and related peak positions are identical with the typical structure of pure ZIF-8.[36] In comparison, the well surface before coating did not present any XRD patterns due to the amorphous protein monolayer. Taken together, our AFM imaging of the well surface after each step, along with Raman spectroscopy and XRD, demonstrate the successful formation of a uniform ZIF-8 coating on the antibody-precoated 96-well plate, and that the ZIF-8 coating can be quickly and completely removed just before plate use.

Figure 2. Surface characterization of a 96-well plate.

(A) Schematic showing coating and washing of a single well in a 96-well plate in a stepwise manner. Column I: Bare well; II: The well after anti-NGAL IgG coating; III. The antibody coated well after BSA blocking; IV. The well after ZIF-8 coating; and V. The well after removing the ZIF-8 coating. (B) 2×2 μm² AFM images of the well surface after each step (corresponding to column I to V). (C) 500×500 nm² AFM images of the well surface after each step (corresponding to column I to V). (D) Raman spectra of the well surface before and after ZIF-8 coating (corresponding to column III to IV). (E) XRD spectra of the well surface before and after ZIF-8 coating (corresponding to column III to IV). Simulated ZIF-8 PXRD pattern is shown as blue (CCDC ID number: 864309).

Next, we set out to investigate if the ZIF-8 coating can stabilize the surface-bound “capture antibodies” under non-refrigerated temperatures. The anti-NGAL IgG pre-coated plates (blocked with BSA) with and without ZIF-8 coatings were first stored at 50°C for 4 weeks, and then evaluated for their dynamic range and detection sensitivity (limit of detection, LOD) by analyzing a series of concentrations of NGAL standards. The freshly prepared antibody-coated plates, and antibody-coated plates after storage at 4°C for 4 weeks were used as controls. In addition, we included another commonly used method, a sucrose coating,[37, 38] as a comparison to our ZIF-8 approach. Therefore, the anti-NGAL IgG-coated plates in five different conditions were evaluated to establish the standard curves, including: freshly prepared antibody-coated plates (Figure 3A); antibody-coated plates after storage at 4°C for 4 weeks (Figure 3B); antibody-coated plates after storage at 50°C for 4 weeks (Figure 3C); antibody-coated plates with ZIF-8 coatings after storage at 50°C for 4 weeks (Figure 3D); and antibody-coated plates with sucrose coatings after storage at 50°C for 4 weeks (Figure 3E). The results shown in Figure 3A–3C, and 3F reveal that the antibody-coated plates after storage at 4°C for 4 weeks had similar dynamic range and LOD with freshly prepared antibody-coated plates, whereas the antibody-coated plates after storage at 50°C for 4 weeks exhibited inferior dynamic range and LOD. These results also justify the fact that manufacturers always require antibody-coated plates to be stored at 2–8°C. Here, we selected 50°C for 4 weeks as our “stress test” because it mimics a harsh storage condition in tropical areas and a surrogate for long-term storage stability at ambient temperatures. Figure 3D and 3F represent our most important results, which indicate that the antibody-coated plates with ZIF-8 coatings after storage at 50°C possessed almost unchanged dynamic range and LOD with freshly prepared and refrigerator-stored antibody-coated plates for at least 4 weeks. However, the antibody-coated plates with sucrose coatings still showed compromised dynamic range and LOD (Figure 3E). This imperfect stabilization could be due to the incomplete coating of sucrose on the plate well surface (Figure S4). Overall, we can conclude that the ZIF-8 coating is able to stabilize the antibody-coated plates at 50°C for at least 4 weeks, and the stabilization efficacy of ZIF-8 approach is comparable to the current “gold standard” refrigeration approach, and superior to the commonly used method based on sucrose coatings. Apart from antibody, we believe this technology could also be used for preservation of antigen-based plates as well as nucleic acid-based biosensors, considering that: (i) These biomolecules are also susceptible to harsh environment; and (ii) ZIF-8 has shown versatile encapsulation capacity for proteins and nucleic acids.[22, 39–42]

Figure 3. ELISA standard curves.

Plots showing the standard curves of the human NGAL ELISA plates after storage at different conditions: (A) Freshly prepared antibody-coated plates; (B) Antibody-coated plates after storage at 4°C for 4 weeks; (C) Antibody-coated plates after storage at 50°C for 4 weeks; (D) Antibody-coated plate with ZIF-8 coatings after storage at 50°C for 4 weeks; and (E) Antibody-coated plates with sucrose coatings after storage at 50°C for 4 weeks. Each data point in standard curves represents the average value from duplicates of three plates (Error bars: N=6). The LOD is defined as the analyte concentration corresponding to the mean optical density of blank plus three times of its standard deviation (3σ method). (F) The comparison of LOD of the ELISA plates after storage at different conditions. Error bars in LOD represent the standard deviation across three plates. The antibody-coated plates with ZIF-8 coatings showed similar LOD and dynamic range with freshly prepared and refrigeration stored plates.

Apart from dynamic range and LOD of ELISA plates, it is also critical to ascertain the reproducibility of the ZIF-8 approach across different wells and plates. In other words, we need to make sure that the ZIF-8 approach can stabilize surface-bound antibodies on different wells and plates to the same extent, and does not create significant variations across different wells and plates when analyzing the same samples. Therefore, we turned our attention to two important quality control parameters of ELISA plates: intra-assay coefficients of variability (intra CV%) and inter-assay coefficients of variability (inter CV%). The CV is a dimensionless number defined as the standard deviation of a set of measurements divided by the mean of the set, which is to describe the well-to-well and plate-to-plate consistency of the ELISA plates.[43] As suggested by ELISA manufacturers, the ideal intra CV% should be less than 10% and inter CV% should be less than 15%. Herein, we spiked known concentrations of NGAL into artificial urine to mimic patient samples and analyzed them using ZIF-8 coated plates after storage at 50°C for 4 weeks. The intra CV% was determined by coefficient of variability of sample duplicates within the same plate while three sample concentrations were analyzed. The inter CV% was determined by coefficient of variability of sample quadruplicates on five different plates while two sample concentrations were analyzed. Our results (intra CV% <3% and inter CV% <7%, as detailed in Figure S5, S6 and S7) show that the ZIF-8 coating approach exhibits excellent reproducibility across different wells and plates, which will be extremely valuable for manufacturing scale-up and high throughput testing.

Finally, to further verify the stabilization efficacy of our ZIF-8 approach on surface-bound antibodies on the 96-well plates, we employed single molecule AFM-based force spectroscopy, to quantify the recognition capability of anti-NGAL IgG coated on the plate well surface. The AFM-based force spectroscopy is a unique quantitative tool for studying molecular interactions, especially for antibody-antigen interactions.[44–46] The questions we expect to answer here are: How does the elevated temperature affect the assay performance? Is it through decreasing average interaction force of antibodies to antigens or reducing the number of available surface-bound antibodies that can bind antigens?

Perhaps more importantly, can the ZIF-8 coating counteract these deleterious effects resulting from elevated temperatures? The principle of AFM-based force spectroscopy involves the functionalization of an AFM tip with antigens (i.e., NGAL here), followed by the use of an NGAL functionalized AFM tip to probe the anti-NGAL IgG pre-coated well surface in a liquid environment. Based on the previous studies,[47–49] the molar ratio of silane-PEG-NHS (for covalently linking the NGAL) to silane-PEG-OH (for minimizing nonspecific binding) was optimized at 1:10,000 to ensure that only single or few NGAL molecule(s) were immobilized on the silicon AFM tip (Figure 4A). In the process of a typical force measurement, the NGAL functionalized AFM tip is slowly brought close to the anti-NGAL IgG pre-coated well surface, contacting with the surface to allow for interaction, and pulled away from the surface. The interaction force between antigen and antibody is then derived from the bending of the AFM cantilever from its equilibrium (unbent) position during the pulling (Figure 4B). In the first set of experiments, an NGAL functionalized AFM tip was applied to collect an array of force curves (24×24) over a 2×2 μm² area on a freshly prepared anti-NGAL IgG coated well surface (Figure 4D). This generated an interaction force map in the same area when displaying the interaction force value for each force curve corresponding to its surface coordinates (Figure 4D, column II). Five independent experiments were conducted, meaning that a total of five arrays of force curves (i.e., 24×24×5=2880 curves) were collected on the surfaces of five wells using five different NGAL functionalized AFM tips. Of these force curves, over 50% of the curves manifested specific interactions and were used to build the force distribution histogram (Figure 4D, column III). As in our previous works,[26, 44] we use “rupture distance >50 nm” as a threshold (meaning the rupture event occurs after 50 nm of pulling from the surface) to screen force curves showing specific interactions according to the contour length of the polyethylene glycol linker (47.9 nm for 5 kDa PEG)[50], whereas the force curves with rupture distance <50 nm are considered as nonspecific interactions and excluded (Figure 4C). The distribution of the interaction force between NGAL on the AFM tip and anti-NGAL IgG on the well surface show two peaks (representing most probable interaction forces at 113 and 222 pN, Figure 4D, column III), and the force value of the second peak is very nearly twice the first, suggesting that, for the second peak, two NGALs on the AFM tip interact with two Fab domains of surface-bound anti-NGAL IgG simultaneously. In other words, the interaction force of the single molecule pair of NGAL and anti-NGAL IgG should be 113 pN (the first peak). To ensure the measured forces were indeed from specific binding of NGAL and anti-NGAL IgG, we conducted two control experiments: (i) An AFM tip only functionalized with silane-PEG-OH was used to probe the anti-NGAL IgG coated plate surface; and (ii) The anti-NGAL IgG coated plate well was first incubated with NGAL solution and then probed with an NGAL functionalized AFM tip. In both experiments, over 95% force curves show no interactions (Figure S8), which suggests the measured forces were from the specific binding between NGAL and anti-NGAL IgG. Following the first set of experiment on freshly prepared antibody-coated plates, we used the NGAL functionalized AFM tips to probe the anti-NGAL IgG-coated plates after storage at four different conditions, including: antibody-coated plates after storage at 4°C for 4 weeks (Figure 4E); antibody-coated plates after storage at 50°C for 4 weeks (Figure 4F); antibody-coated plates with ZIF-8 coatings after storage at 50°C for 4 weeks (Figure 4G); and antibody-coated plates with sucrose coatings after storage at 50°C for 4 weeks (Figure 4H). The force maps and force distribution histograms indicate that antibody-coated plates with ZIF-8 coatings after storage at 50°C for 4 weeks still possess similar binding frequency (~50%) and binding force (107 pN) to NGAL, comparable to those of freshly prepared and refrigeration-stored antibody-coated plates. In contrast, the plates without coating or with sucrose coatings after storage at 50°C for 4 weeks display significantly decreased binding frequency and binding force to NGAL. This is an interesting finding implying that the compromised dynamic range and detection sensitivity of antibody pre-coated ELISA plates after storage under non-refrigerated conditions are due to the decrease of both average binding force of surface-bound antibodies to antigens and binding frequency (i.e., the number of available antibodies that can bind antigen). For the plates without coating, it is likely that the complete denaturing of the antibodies results in defective binding (i.e., significant lower average binding force) to their target antigens, further leading to very low binding frequency. In the case of plates with sucrose coatings, the “patchy” sucrose coatings as shown in Figure S4 could still partially preserve the Fab domains of antibodies underneath through limiting the protein chain mobility. More importantly, the ZIF-8 coating here is capable of maintaining both binding force and binding frequency of surface-bound antibodies to target antigens, leading to stabilized assay performance including dynamic range and detection sensitivity. These results highlight the importance of understanding the single molecule interaction in design and evaluation of highly sensitive bioassays since the fine differences at the single molecule level manifest significantly different assay performance. Therefore, AFM-based force spectroscopy provides a fundamental means to understand assay performance, as well as a quantifiable method to predict the stabilization efficacy of different approaches, from a single molecule perspective.

Figure 4. AFM-based force spectroscopy investigating recognition capabilities of surface-bound antibodies.

(A) Schematic illustrating the functionalization of an AFM tip with single antigen (NGAL here) using a mixed PEG silane approach. The molar ratio of silane-PEG-NHS (for covalently linking the NGAL) to silane-PEG-OH (for minimizing nonspecific binding) was optimized at 1:10,000 to ensure that only single or few NGAL molecules were immobilized on the silicon AFM tip. (B) Schematic illustrating the process of a typical force measurement. The NGAL functionalized AFM tip was slowly brought close to the anti-NGAL IgG pre-coated well surface, contacting with the surface to allow for interaction, and pulled away from the surface. The interaction force between antigen and antibody was then derived from the bending of the AFM cantilever from its equilibrium (unbent) position during the pulling. (C) Typical AFM force-distance curves collected in the experiments. Curves with a rupture distance >50 nm were classified as specific interactions, whereas curves with a rupture distance <50 nm were considered as nonspecific interactions. The asterisks represent the rupture events during the pulling of the AFM tips. (D-F) Column I: Schematic illustrating the antibody-coated plates stored at different conditions. Column II: Representative force maps (24×24 force curves) collected over a 2×2 μm² area on plate well after storage at different conditions. Column III: Force distribution histograms established from force curves from five independent experiments. At each condition, a total of five force maps were collected on the surfaces of five wells using five different NGAL functionalized AFM tips. Of these force curves, curves manifesting specific interactions were used to build the force distribution histogram, followed by Gaussian fits of the data.

Conclusions

We have introduced a facile and reversable coating based on ZIF-8 for stabilization of surface-bound antibodies of ELISA plates under non-refrigerated conditions. By using a sandwich ELISA for the detection of urine NGAL (a biomarker for acute kidney injury) as a model system, we demonstrated that a ZIF-8 coating formed on anti-NGAL IgG coated 96-well plate surface is able to stabilize the recognition capability of the surface-bound anti-NGAL IgG, and maintain the assay dynamic range and detection sensitivity after storage at 50°C for at least 4 weeks. Nanoscale AFM imaging showed that ZIF-8 can form a uniform coating on the surface-bound antibodies within 1h, and the coating can be completely washed away within 1 min just before use of the assay, without disrupting the assay workflow. The stabilization efficacy of the ZIF-8 coating is comparable to the current “gold standard” refrigeration approach, and superior to the commonly used sucrose coating method.[37, 38] Moreover, our ZIF-8 coating approach exhibits excellent reproducibility across different wells and plates in terms of antibody stabilization, which is extremely valuable for manufacturing scale-up and high throughput testing. Furthermore, we demonstrated single molecule AFM-based force spectroscopy can quantify the recognition capability of single surface-bound anti-NGAL IgG antibodies to NGAL antigens. The results indicate that the stabilized assay performance, including dynamic range and detection sensitivity, offered by the ZIF-8 coating is because the ZIF-8 coating maintains both average binding force and binding frequency of surface-bound antibodies to antigens. Overall, we believe this facile approach will greatly improve the shelf-life and stability of antibody-coated ELISAs and could be applicable to other types of immunoassays which utilize surface-bound antibodies. Therefore, this approach will enable implementing reliable bioassays in regions where they are currently inaccessible, thus extending biomedical research and medical diagnostics to resource-limited settings and underserved populations. Future work is under the way to determine the applicability of this approach in preservation of antigen and nucleic acid-based assays.

Experimental Section

Materials.

Human Lipocalin-2/NGAL ELISA kit (DY1757, including rat anti-human NGAL capture antibody, biotinylated goat anti-human NGAL detection antibody, streptavidin-HRP, and recombinant human NGAL standard) was purchased from R&D Systems. The flat-bottom 96-well plates were purchased from Corning (Product number: 3601). Other materials for running the ELISA including PBS buffer, wash buffer with Tween 20, reagent diluent with BSA, substrate solution and stop solution are also purchased from R&D Systems. 2-Methylimidazole, zinc acetate dihydrate, sucrose, and artificial urine were purchased from Sigma-Aldrich. Silane-PEG-OH (2 kDa) and silane-PEG-NHS (5 kDa) were purchased from NANOCS.

ELISA plate preparation and assay procedure.

To prepare an antibody-coated ELISA plate, first, coat the 96-well plate with capture antibody by adding 100 μl of capture antibody solution to each well and incubate the plate overnight (12 h) at room temperature. Then each well was washed with wash buffer for three times to remove the unbound capture antibody. Then each well was blocked with 300 μl of reagent diluent (with BSA) for 1 h at room temperature. After blocking, each well was washed again with wash buffer for three times and blotted against clean paper towels to remove the remaining liquid. Then the plates were ready for establishing the standard curve and sample analysis. If the plates were not used for analysis immediately, the plates were first dried in fume-hood for 1 h, sealed, and stored in a refrigerator (at 4°C) for 4 weeks. For the assay procedure, 100 μl of NGAL standards (a series concentrations of NGAL in reagent diluent) or samples (NGAL spiked in artificial urine) were added to each well and incubated for 1 h at room temperature. After three times of washing to remove the unbound NGAL, 100 μl of detection antibody solution was added into each well and incubated for 1 h. Then each well was washed for three times, and 100 μl of streptavidin-HRP was added into each well and incubated for 20 min. Again, each well was washed for three times, and 100 μl of substrate solution was added into each well and incubated for 20 min. Finally, 50 μL of stop solution was added to each well and the plate was then read by a microplate reader at 450 nm for quantification.

ZIF-8 and sucrose coating, and “stress test”.

For ZIF-8 coating, after the plates were coated with capture antibody, each well was added with 300 μl of ZIF-8 precursor solution and incubated for 1 h. Right before adding into each well, the ZIF-8 precursor solution was freshly prepared by mixing 150 μl of 2-methylimidazole solution (160 mM, in pure water) and 150 μl of zinc acetate dihydrate solution (40 mM, in pure water) and agitating for 10 s. After 1 h of incubation, the solution was pipetted out from each well and the plate was blotted against clear paper towels to remove the remaining liquid. Then after dried in fume hood for another 1 h, the plate was sealed, and stored in an oven (50°C) for 4 weeks as the “stress test”. To remove the ZIF-8 coating, 300 μl of slightly acidic pure water (pH=6) was added into each well and incubated for 1 min. Then the water was pipetted out and the plate was ready for subsequent assay procedure. Similar procedure was employed for sucrose coating, in which 3% (weight/volume) of sucrose solution was used for sucrose coating.

Characterization of well surface.

After each step, the bottom of the plate well was cut and glued onto a glass slide for AFM imaging in air. AFM images were collected using a BioScope Resolve AFM (Bruker) in a ScanAsyst mode by using ScanAsyst-Air cantilevers (Bruker). The Raman spectra were acquired on plate well bottom before and after ZIF-8 coating by a Foram X3 Raman spectrometer with a 40× objective and a 532 nm diode laser as an illumination source. The X-ray diffraction (XRD) analysis of plate well bottom before and after ZIF-8 coating were conducted with a Bruker D8-Advance X-ray powder diffractometer using Cu Kα radiation (λ = 1.5406 Å) with scattering angles (2θ) of 5−30°.

AFM-based force spectroscopy.

DNP cantilever D (Bruker, with nominal tip radius: 20 nm and nominal spring constant: 0.06 N/m) was used for force measurement. Before functionalizing the tips with NGAL antigen, the tips were first immersed in a mixture of ethanol and isopropyl alcohol (1:1 v/v) for 10 min to remove the organic contaminants and then treated with oxygen plasma to induce hydroxyl groups on the tip surface. Then the tips were immersed in a mixed silane PEG solution (0.5 μM silane-PEG-NHS and 5 mM silane-PEG-OH dissolved in anhydrous DMSO, molar ratio of 1:10,000) for 1 h at room temperature. Then the tips were washed with anhydrous DMSO and pure water to remove unbound silane-PEGs. After washing, the tips were incubated with NGAL solution (100 ng/mL in PBS) for 3 h at 4 °C. Then the tips were rinsed with PBS to remove unbound NGAL antigens. Before collecting force curves, the NGAL-functionalized tips were calibrated on clean silicon using the thermal noise method, yielding spring constant values ranging from 0.05 to 0.1 N/m. Then the functionalized tips were used to probe antibody-coated plate well surfaces by collecting force-distance curves in PBS. A force-volume mode was used to collect 24×24 force distance curves over areas of 2×2 μm². All force curves were recorded by applying a loading force of 300 pN, with a constant retraction speed of 0.5 μm s⁻¹, a ramp size of 0.5 μm, and 0.1 s of surface delay. AFM force-distance curves and interaction force maps were analyzed using the NanoScope analysis software (v1.9, Bruker). Individual force-distance curves detecting specific interactions were analyzed using the NanoScope analysis software.