Click-Particle Display for Base-Modified Aptamer Discovery

Base-modified aptamers that incorporate non-natural chemical moieties can achieve greatly improved affinity and specificity relative to natural DNA or RNA aptamers. However, conventional methods for generating base-modified aptamers require considerable expertise and resources. In this work, we have accelerated and generalized the process of generating base-modified aptamers by combining a click-chemistry strategy with a fluorescence-activated cell sorting (FACS)-based screening methodology that measures the affinity and specificity of individual aptamers at a throughput of ∼107 per hour. Our “click-particle display (PD)” strategy offers many advantages. First, almost any chemical modification can be introduced with a commercially available polymerase. Second, click-PD can screen vast numbers of individual aptamers on the basis of quantitative on- and off-target binding measurements to simultaneously achieve high affinity and specificity. Finally, the increasing availability of FACS instrumentation in academia and industry allows for easy adoption of click-PD in a broader scientific community. Using click-PD, we generated a boronic acid-modified aptamer with ∼1 μM affinity for epinephrine, a target for which no aptamer has been reported to date. We subsequently generated a mannose-modified aptamer with nanomolar affinity for the lectin concanavalin A (Con A). The strong affinity of both aptamers is fundamentally dependent upon the presence of chemical modifications, and we show that their removal essentially eliminates aptamer binding. Importantly, our Con A aptamer exhibited exceptional specificity, with minimal binding to other structurally similar lectins. Finally, we show that our aptamer has remarkable biological activity. Indeed, this aptamer is the most potent inhibitor of Con A-mediated hemagglutination reported to date.

The same experiments was performed to determine the appropriate polymerase to incorporate DBCO-PEG4-dUTP, and our analysis of the final reaction (gel not shown) likewise indicated that KOD-XL provides the highest yield and purity.
The cap of the tube was then removed, and the de-capped tube was immediately placed in a 20 mL vial equipped with a rubber septum, followed by Ar flushing for 5 min. We incubated the sealed vial in the dark for two hours. The reaction product was purified with a Centri-Spin 10 column (Princeton Separations). 200 L of concentrated ammonium hydroxide (18 M) was added to the purified product, and the solution was incubated at room temperature for 3 hours. 400 L nbutanol was then added, vortex mixed, and centrifuged at 16,000 ×g at 4 °C for 2 min. The top organic layer was removed and discarded. The bottom aqueous layer was purified by an Oligo Clean and Concentrator spin column (Zymo Research), followed by HPLC analysis.

Synthesis of N-methyliminodiacetic acid (MIDA)-protected p-azidomethyl
phenylboronic acid (p-AMPBA). Following a modified procedure based on literature 1 ( Figure   S4), a 50 mL round bottom flask was charged with p-azidomethylphenyl boronic acid pinacol ester (0.26 g, 1 mmol), 3 mL 1 M HCl aqueous solution, polymer-bound boronic acid (3.5 g, 2.6 mmol/g, 9 mmol), and 18 mL acetonitrile. This mixture was stirred at room temperature for 36 h, with reaction progress monitored by thin layer chromatography (1:1 hexane:ethyl acetate as eluent). After the reaction was complete, the solvent was evaporated in vacuo, and 5 mL water was added. The crude sample was frozen by liquid nitrogen and lyophilized overnight. The residue was added into a 500 mL flask, along with MIDA (0.15 g, 1 mmol), 18 mL benzene, and 2 mL DMSO.
The flask was then fitted with a Dean-Stark trap and a reflux condenser, and the mixture was refluxed with stirring for 16 h followed by concentration in vacuo. The resulting crude product was adsorbed onto Florisil gel from an acetonitrile solution. The resulting powder was dry-loaded on top of a silica gel column slurry-packed with ethyl acetate. The product was eluted using a gradient Optimization of SPAAC click conjugation reaction. We optimized the reaction conditions for coupling boronic acid to DBCO-PEG4-dUTP via SPAAC using sequence T-29-DBCO (Table S1). 1 l 1 mM T-29-DBCO was added to 10 l 2.78 mM p-AMPBA in 1X PBS, 100 mM NaCl and incubated overnight. Excess p-AMPBA was removed using a 3k Amicon filter and deionized water. Half of the reacted product was deprotected with 0.5M NaOH for 10 minutes at room temperature. Both deprotected and protected products were washed with deionized water.

Optimization of PCR amplification, click conjugation of 2-azidoethyl 2,3,4,6-tetra-O-
in DMSO (100 eq) and 40 L of 20 mM sodium phosphate buffer, pH 8 (pre-degassed by bubbling N2 through) with 20 L of base-modified DNA solution in a 1.5 mL Eppendorf tube. Click chemistry was initiated by adding a 20 L premixed solution of 1:1 Cu:TBTA (10 mM, prepared from 1 mg CuBr + 0.7 mL 10 mM TBTA in 4:3:1 water:DMSO:t-BuOH). The cap of the tube was removed, and the tube was immediately placed in a 20 mL vial equipped with a rubber septum, followed by Ar flushing for 5 min. We incubated the sealed vial in the dark for two hours. To this DNA solution, we added 10 L of 3 M sodium acetate (pH 5.2) and 330 L of 100% ethanol, followed by freezing at -80 °C for 30 min. The frozen stock was then centrifuged for 30 min at 21,000 × g at 4 °C to precipitate the DNA. We resuspended the material in 350 L of 1X bind and wash buffer (B&W; 5 mM Tris, 0.5 mM EDTA, 1 M NaCl, pH 7.5).
We then added 350 L MyOne C1 streptavidin beads to a 1.5 mL Eppendorf tube. We captured the beads on the side of the tube with a magnet and removed the supernatant. The beads were washed three times with 350 L 1X B&W. The click product was added to the beads and mixed on a rotator for 30 min. The beads were then captured on a magnetic rack and the supernatant was discarded. The beads were washed three times with 350 L 1X B&W, and then treated with 100 L freshly-prepared 0.25 M NaOH solution to generate single-stranded DNA (ssDNA). The beads were captured by magnet, and the supernatant was collected and desalted using a Centri-Sep column (Princeton Separations).
We deprotected the acetyl groups by adding 200 L concentrated ammonium hydroxide (18 M) to the collected oligos and incubating for 4 hours at room temperature. 450 L n-butanol was then added to the solution, followed by vortexing, and centrifuging at 21,000 × g at 4 °C for 1 min. The top organic layer was removed and discarded. The resulting base-modified aptamer solution was then desalted by a Centri-Spin-10 column (Princeton Separations).
General procedure for generating particle-displayed base-modified aptamers.
Monoclonal, particle-displayed base-modified aptamers were generated by emulsion PCR. The oil phase was made up of 4.5% Span 80, 0.45% Tween 80, and 0.05% Triton X-100 in mineral oil.
The aqueous phase consisted of 1x KOD XL DNA polymerase buffer, 50 U KOD XL DNA polymerase, 0.2 mM dATP, 0.2 mM dGTP, 0.2 mM dCTP (or 0.2 mM 5-formyl-deoxycytidine for the Con A screen), 0.2 mM C8-alkyne-dUTP or DBCO dUTP, 10 nM FP (C-FP or E-FP), 1 M fluorescently-labeled RP (C-RP or E-RP), ~1 pM template DNA, and ~10 8 1 m FPconjugated magnetic beads. For each reaction, 1 mL of aqueous phase was added to 7 mL of oil phase and emulsified at 620 rpm for 5 min in an IKA DT-20 tube using the IKA Ultra-Turrax After PCR, the emulsions were collected into an emulsion collection tray (Life Technologies) by centrifuging at 300 x g for 2 min. The emulsion was broken by adding 10 mL 2-butanol to the tray, and the sample was transferred to a 50 mL tube. The tube was vortexed for 30s, and the particles were pelleted by centrifugation at 3,000 x g for 5 min. The oil phase was carefully removed, and the particles were resuspended in 1 mL of emulsion breaking buffer (100 mM NaCl, 1% Triton X-100, 10 mM Tris-HCl, pH 7.5, and 1 mM EDTA) and transferred to a new 1.5 mL tube. After vortexing for 30s and 90s of centrifugation at 15,000 × g, the supernatant was removed. The tube was placed on a magnetic separator (MPC-S, Life Technologies), and the remaining supernatant was removed. The particles were washed three times with 1x PBS buffer using magnetic separation, then stored in 200 L 1x PBS, 0.1% Tween 20 at 4 °C before click modification.

SPAAC:
This click chemistry reaction was employed when p-AMPBA was used as the click modification for the epinephrine selection. ~10 7 particles displaying DNA incorporating DBCO-PEG4-dUTP were incubated overnight at room temperature with an excess of p-AMPBA (>1.5 mg) in 1 ml of storage buffer (10 mM Tris, pH 7.4, 0.025% Tween 20). Before use in the particle display binding assays, particles with the p-AMPBA modification were washed two times with 0.5M NaOH for 10 minutes to deprotect the functional group.
CuAAC: This click chemistry reaction was employed when mannose was used as the click modification for the Con A selection. ~10 7 particles displaying the DNA library incorporating C8alkyne-dUTP were resuspended in 10 L 1x PBS and combined with 25 L of 20 mM Na2HPO4 ,  We confirmed that base-modified aptamers are efficiently displayed on the particle surface by fluorescently labeling the 3'-end of the base-modified aptamers to allow for FACS characterization after emulsion PCR ( Figure S5). A cleavable disulfide linker was also incorporated between the aptamer and the particle to allow cleavage of the modified DNA for electrophoretic analysis (Figure S6). The slightly lower mobility of the cleaved base-modified aptamer was attributed to the extra mass from the "scar" of the disulfide linker.  (Figure S7a), we found that Taq efficiently generated DNA of the correct length from the base-modified aptamers ( Figure S7b). Sanger sequencing showed that the product generated by Taq polymerase was identical to the starting template, confirming the fidelity of the reverse transcription process (Figure S8).
The same experimental procedure was performed for base-modified aptamer M1displaying particles with the boronic acid modification, and we determined that Taq polymerase efficiently reverse-transcribed our sequence. ssDNA was generated by using MyOne SA C1 beads to capture the double-stranded DNA, after which the sense strand was eluted by incubating the DNA-coated beads with 0.1 M NaOH for 10 minutes. DNA was recovered by adding 1/10 vol 3M NaOAc to neutralize pH followed by purification with a Qiagen MiniElute cleanup kit.
SELEX was repeated with the isolated sense strands under the same conditions (l epinephrine beads in 200ul selection buffer, rotating 30 min RT) and recovered in the same way.
The output DNA of the second round of SELEX was used as the input DNA for particle display Dynabeads (Thermo Fisher) for 3 hours at room temperature. Supernatant was saved for particle display selection by magnetic separation.
Click-PD screening. For each round of screening for epinephrine, ~10 7 boronic acidmodified aptamer particles were incubated in 100 ul binding buffer at 95 °C for 2 min, and then cooled at room temperature for 30 min. The beads were incubated with the appropriate concentration of FITC-epinephrine conjugates in 1mL on the rotator in the dark for 1 h. After removing the supernatant, the particles were resuspended in 1 mL selection buffer, and then analyzed using a BD FACS Aria III. In each round, the sort gate was set to collect aptamer particles that show high binding affinity towards epinephrine (i.e., high FITC fluorescence). As the rounds progressed, we applied higher stringency for the sort gates. The collected population ranged from 0.1-0.3%. After sorting, the collected base-modified aptamer particles were resuspended in 20 L PBS and reverse transcribed into canonical DNA by Taq polymerase.
For each round of screening for Con A, we incubated ~10 8 mannose-modified aptamer particles with 1 nM biotinylated Con A (Sigma) and 250 nM FITC-conjugated PSA (Sigma) in selection buffer (SB; 1 x PBS, 2.5 mM MgCl2, 1 mM CaCl2, 0.1 mM MnCl2, 0.01% Tween 20) for 1 hour in the dark on a rotator. After incubation, 2 L of 2 mg/mL streptavidin-conjugated Alexa Fluor 647 was diluted in 998 L of selection buffer. The aptamer particles were resuspended in the 1 mL mixture and incubated for 10 min in the dark on a rotator to fluorescently label the biotinylated Con A. The particles were washed once and resuspended in SB. The sample was then analyzed with the BD FACS Aria III, with the sort gate set to collect base-modified aptamer particles that exhibit high binding to Con A and low binding to PSA. 0. Sequences with low quality were filtered out using "Filter by quality", accepting only sequences with more than 90% of the bases having a quality score of 20 or above. The FASTAptamer toolkit was used to identify sequence clusters (sequences varying by 2 or fewer bases) and calculate the degree enrichment of each sequence from round to round.
General procedure for particle-based binding assay for fluorescently labeled targets. ~10 6 particles were incubated with varying concentrations of fluorescently labeled protein in SB for 1 hour on a rotator at RT. After incubation, the particles were washed once in 100 uL cold SB and resuspended in 100uL cold SB. The particles were analyzed using the BD Accuri C6 flow cytometer, and the mean fluorescence and/or percentage of bound particles were measured in the relevant fluorescence channel(s). This tube was then sealed tightly before heating on a thermal block at 70 °C for 10 min. The sample was cooled in an ice bath before opening the cap. The tube was placed on the magnet, and the supernatant was transferred to a separate tube. 100 L of 18 M ammonium hydroxide was again added to the beads, and the heating procedure was repeated.

Generation of solution-phase base-modified aptamers with 5'-biotinylation for
The supernatants from the two ammonium hydroxide treatment steps were combined and then mixed with 4.5 mL n-butanol before vortexing and centrifuging at 16,000 ×g at 4 °C for 10 min. The supernatant was removed and discarded. The sample was dried over vacuum centrifugation, and then resuspended in 100 L water. To this solution, we added 50 L of 5 M NH4OAc and 415 L of cold 100% ethanol before freezing at -80 °C for 30 min. We centrifuged for 30 min at 21,000 ×g at 4 °C to precipitate the base-modified aptamer. The pellet was washed once with 70% v/v cold ethanol in water, then dissolved in 100 L water.
Bio-layer interferometry measurement of selected base-modified aptamers. ConA-3-1 and ConA-3-1m were diluted to 50 nM in SB. Solutions of 0, 1, 2, 4, 8, 16, 32, and 64 nM Con A were prepared in SB. The solutions were loaded into a 384-well plate, with 100 L of SB, 80 L of biotinylated aptamer, and 100 L of Con A solution for each reaction. The following steps were run on the ForteBIO Octet RED384 with Super Streptavidin biosensors: 60s in buffer for equilibration, 5 min in aptamer solution to load the aptamer onto the biosensors, 60s in buffer for a baseline measurement, 10 min in Con A solution to measure association, and 10 minutes in buffer to measure dissociation. Analysis was performed using Octet Data Analysis software, including the alignment of the different measurements and global fitting of the experimental data to a binding model to extract Kd, kon, and koff.
Lectin array assay to probe base-modified aptamer specificity. The following procedure was adapted from RayBiotech's product manual for the Lectin Array 70. First, we dried the glass slide. The slide with the pre-printed lectin array was equilibrated to room temperature inside the sealed plastic bag for 20-30 minutes. We then annealed 30 L of 0.5 M ConA-3-1m in 1X PBS by incubating the solution at 95 °C and slowly cooling down to 4 °C at a ramp rate of 0.1 °C/second. We incubated at 4 °C for 5 min. We then added 100 L sample diluent (included in the lectin array package) into each well of the array and incubated at room temperature for 30 min to block the slides. We removed the buffer from each well. After diluting ConA-3-1m to the desired concentration with SB, we added 100 L of diluted ConA-3-1m to each well and incubated the arrays at room temperature for 3 hours. We then removed the samples from each well, and washed each well five times (5 min each) with 150 L of 1X wash buffer I (included in the lectin array package, supplemented with 2.5 mM MgCl2, 1 mM CaCl2, and 0.1 mM MnCl2) at room temperature with gentle shaking. We completely removed the buffer between each wash step. We then washed two times (5 min each) with 150 L of 1X wash buffer II (included in the lectin array package, supplemented with 2.5 mM MgCl2, 1 mM CaCl2, and 0.1 mM MnCl2) at room temperature with gentle shaking. We completely removed the wash buffer between each wash step.
We then briefly spun down the Cy3 equivalent dye-conjugated streptavidin tube (included in the lectin array package) and added 1.4 mL of sample diluent to the tube, mixing gently. We added 80 L of Cy3 equivalent dye-conjugated streptavidin to each well and incubated in the dark at room temperature for 1 hour. We decanted the samples from each well, and washed five times with 150 L of 1X wash buffer I at room temperature with gentle shaking, completely removing the wash buffer after each wash step. We disassembled the slide assembly by pushing the clips outward from the slide side and carefully removing the slide from the gasket. We placed the slide in the slide washer/dryer (a four-slide holder/centrifuge tube included in the lectin array package), adding enough 1x wash buffer I to cover the whole slide (~30 mL), and then gently agitated at room temperature for 15 minutes. After decanting wash buffer I, we washed with 1x wash buffer II (about 30 mL) with gentle shaking at room temperature for 5 minutes. Finally, we dried the slide by centrifugation at 200 × g on a microscope slide spinner and scanned the slide on a GenePix Array scanner, monitoring the Cy3 dye channel at PMT 500.
Determining Con A concentration to induce complete hemagglutination. Human erythrocytes were washed and resuspended in 1X PBS in a 96-well U-shaped well plate at 1% hematocrit, with Con A concentrations ranging from 2-250 g/mL. We let the plate stand at room temperature for 1 hour before visualizing the deposition of erythrocytes at the bottom of the well.
The optical densities at 655 nm were then measured on a Tecan M220 plate reader.
Hemagglutination inhibition assay. We annealed 30 L of 0.5 M ConA-3-1m in 1X PBS by heating the solution to 95 °C and slowly cooling down to 4 °C at a ramp rate of 0.1 °C/second, followed by incubation at 4 °C for 5 min. We incubated the annealed aptamer at a range of concentrations from 9.6-300 nM with 150 nM Con A in 1X PBS for 30 min in a 96-well Ushaped well plate. Human erythrocytes were added to produce a cell suspension of 1% hematocrit in a total volume of 50 L per well. After 1 hour of incubation at RT, the hemagglutination status of the samples was visualized, and the optical densities of the cell suspensions at 655 nm were monitored by a plate reader.
Microscopic characterization of human erythrocyte agglutination. We annealed 2 L of 4 M ConA-3-1m in 1X PBS by incubating the solution at 95 °C and slowly cooling down to 4 °C at a ramp rate of 0.1 °C/second, and then incubated at 4 °C for 5 min. We then added 1 L of 6.5 M Con A and incubated for 30 min at room temperature. We prepared an erythrocyte suspension to a final hematocrit of 20% in PBS, and 7 L of erythrocyte suspension was combined either with the Con A-aptamer complex or 3 L of 1X PBS. 10 L of this mixture was loaded onto glass slides, covered with coverslips, and immediately visualized using 10X and 40X objective lenses on a microscope. Optical microscopy imaging was performed on an Olympus CKX-41 inverted microscope with color digital camera. The images were processed with ImageJ software.    Left, structure of the "scar" of the disulfide linker after base-modified aptamer cleavage. The disulfide linker between the forward primer and the particle is cleaved by TCEP treatment followed by alkylation using iodoacetamide. Right, our click chemistry reaction conditions efficiently modified particle-coupled base-modified aptamers. Lane 1 contains the reaction product M1 formed in solution (see Figure S3a), and lane 2 contains base-modified aptamer cleaved from beads after emulsion PCR and on-bead click reaction.      (Table S2) were tested by flow cytometry assay for binding from 0 to 10 M FITC epinephrine. Kd was determined to be 0.3 M ± 0.18 for 4-1, and could not be determined for the other sequences. Table S2. Selected sequences from high-throughput DNA sequencing of epinephrine binders.

Sequence Random region only (5'-3')
4-1 GTACGTGAATCCATGGGGACGGAGAGACGGCACGAACCAA 4-2 GGAGAGGGCTCAGACCAACAGAGGGGAAATGGGGCGGTGG 4-3 GGAGAGGGCTCAGACCAACAGAGGGGAAATGGGGCGGTGA 4-7 GTACGCGAATCCATGGGGACGGAGAGACGGCACGAACCAA Figure S12. Binding of selected sequences to fluorescently-labeled Con A in a particle-based assay. Two criteria were considered to identify the top-performing base-modified aptamer. First, >60% of base-modified aptamer-displaying particles should bind Con A in a particle-based fluorescent assay. Second, the base-modified aptamer should originate from a cluster that has undergone >2,000-fold enrichment. ConA-3-1 met both of these criteria.   predicted by mFold. Note that modified nucleotide 4 has been substituted with dT in the simulation. The circled nucleotide positions were mutated to dA individually or in pairs, and the binding of the mutant base-modified aptamers was characterized in a particle-based fluorescent assay. B) The relative fluorescence signals of the mutant sequences, which are shown in C). The error bars were derived from three experimental replicates. The fluorescence signals were first normalized to particle coating, and then to the relative signal of ConA-3-1m.
Figure S16. ConA-3-1m exhibits little binding to NPA, LcH, and VFA lectins. We incubated particles coated with ConA-3-1m with fluorescently-labeled mannose-binding lectins. These were then washed and analyzed by FACS based on mean fluorescence of the population. Error bars were derived from three experimental replicates. Figure S17. Aptamer array binding for ConA-3-1m and previously described aptamer Seq 1. A) Raw binding signal of ConA-3-1m aptamer binding to Con A (red) and other lectins. We did not test concentrations high enough to determine a Kd value for off-target binding. The normalized plot of this data is shown as Fig 6B. B) Raw binding signal of the previously published Seq 1 aptamer (red) to Con A and other lectins on an array, scaled to match the axes for the raw binding signal of ConA-3-1m. The signal produced by Seq 1 is negligible compared to our aptamer, and only VVA (yellow) continues to show non-specific binding. C) Normalized signal of Seq 1 binding to Con A (red) and other lectins. The signal from each lectin was normalized to the maximum Con A signal measured in each assay.

Table S4. Additional information on lectins spotted on the lectin array.
The following information is replicated from Lectin Array 70 product manual.