Selection of optimised ligands by fluorescence-activated bead sorting

The chemistry of aptamers is largely limited to natural nucleotides, and although modifications of nucleic acids can enhance target aptamer affinity, there has not yet been a technology for selecting the right modifications in the right locations out of the vast number of possibilities, because enzymatic amplification does not transmit sequence-specific modification information. Here we show the first method for the selection of specific nucleoside modifications that increase aptamer binding efficacy, using the oncoprotein EGFR as a model target. Using fluorescence-activated bead sorting (FABS), we have successfully selected optimized aptamers from a library of >65 000 variations. Hits were identified by tandem mass spectrometry and validated by using an EGFR binding assay and computational docking studies. Our results provide proof of concept for this novel strategy for the selection of chemically optimised aptamers and offer a new method for rapidly synthesising and screening large aptamer libraries to accelerate diagnostic and drug discovery.

Oligonucleotide Synthesiser: All library components were synthesised on an Expedite™ 8909 Nucleic Acid Synthesiser system provided by Biolytic. Phosphoramidites were dissolved in dry acetonitrile to the concentrations as suggested by the supplier, solvents were used as provided. Oligomers were synthesised on a 1 µM scale.
Mass Spectrometer (Small Molecule): Electrospray mass spectra were recorded on a Bruker micrOTOF-Q II mass spectrometer. The samples were analysed by injecting 2 µL of 0.1 mg/mL solutions into a flowing stream of 95% methanol, 10 mM ammonium acetate at a flow rate of 20 μL/min. The samples were analysed in both positive and negative ion modes. All the masses are mono isotopic and lock mass corrected and therefore should be within 5 ppm of the calculated mass.
Mass Spectrometer (Oligonucleotides): Mass spectra were recorded on Waters QToF ESI-LC/MS/MS. Using columns nanoE MZ Sym C18 Trap Column 5 μm and nanoE MZ HSS T3 Column 1.8 μm 75 μm x 150 mm. The samples were analysed by injecting 2 µL of 0.1 mg/mL solutions into a flowing stream of 95% methanol, 10 mM ammonium acetate at a flow rate of 20 μL/min. The samples were analysed in both positive and negative ion modes. Mobile phase A was 8 mM tetraethyl ammonium bromide (TEAB) in LC-MS grade water adjusted to the pH 7.5-7.8. Mobile phase B was 8 mM TEAB in a 1:1 ratio of LC-MS grade methanol and water. Initially A was held at 92% and B at 8%. At 30 mins B was increased to 65% and up to 95% at 31 mins. At 33 mins B was dropped back to 8% and then held until 60 mins. The column is maintained at 50˚C. 2,3 NMR: NMR spectra were obtained on a Bruker AVII 400 MHz spectrometer and were calibrated to the centre of the set solvent peak and chemical shifts were reported in parts per million (ppm). All NMR samples were run in DMSO-d6.
Scanning Election Microscope: SEM images were obtained on a Hitachi S-3400N, using a 10 kV electron beam and secondary electron detector.

S2 Bead suitability
Imaging of TentaGel® M NH₂ Monosized Amino TentaGel Microspheres using SEM. SEM images were obtained on a Hitachi S3400N, using a 10 kV electron beam and secondary electron detector. The sample was deposited on carbon tabs placed on aluminium sample holders. The images were visualised on ImageJ and the data analysis was done using Origin Lab software.
Amidation of 10-hydroxydecanoic acid with TentaGel® M NH₂ Monosized Amino TentaGel Microspheres. N,N'-Dissopropylcarbodiimide (0.2 g) was dissolved in dimethylformamide (10 mL). 10-Hydroxydecanoic Acid (0.4 g) was added to the solution. TG-beads (0.1 g) were swelled in dimethylformamide (10 mL) and then added to the solution. This solution was stirred for three hours at room temperature. It was spun down and cleaned with dimethylformamide (10 mL x 3) and then with ethanol (10 mL x 2). A white power was produced (0.95 g, 95%). A Kaiser test was conducted producing a yellow colour (successful).
Fluorescent tagging of TentaGel® M NH₂ Monosized Amino TentaGel Microspheres. 6-Carboxyfluorescein. N,N'-Disopropylcarbodiimide (0.32 g) was dissolved in dimethylformamide (10 mL). The amine-modified TGbeads (0.05 g) were swelled in dimethylformamide (10 mL), this was added to the solution along with 6carboxyfluroescein (0.05 g). The solution was stirred for 3 hours at room temperature. It was spun down and cleaned with dimethylformamide (10 mL x 3) and then with ethanol (10 mL x 2). The product was then air dried and an orange powered was collected (TGCFluor100) (0.074 g, 74%). Samples TGCFluor100, TGCFluor73 and TGCFluor36 are made per amounts in table. Rhodamine B. N,N'-Dissopropylcarbodiimide (0.15 g) was dissolved in dimethylformamide (10 mL). The amine modified TG-beads (0.05 g) were swelled in dimethylformamide (10 mL), this was added to the solution along with Rhodamine B (0.05 g). The solution was stirred for 3 hours at room temperature. It was spun down and cleaned with dimethylformamide (10 mL x 3) and then with ethanol (10 mL x 2). The product was then air dried and a pink powered was collected (TGRhodB100) (0.03 g, 60%). Samples TGRhodB100, TGRhodB73, TGRhodB36, TGRhodB17 an TGRhodB1.7 are made per amounts in table. Leica DMR microscope equipped with a Leica DFC9000 GT digital camera. The illumination source was a CoolLED pE-300 ultra fluorescence microscopy Illumination System. Images were acquired using Leica Application Suite X software. Excitation 495nm and emission 525/50 nm. Excitation 515-560 nm and emission long pass 560-590 nm. A small sample of 6-carboxyfluorescein tagged TG-beads in water were smeared onto a glass slide. This was placed on the microscope platform and visualised using a fluorescent lamp.
Flow Cytometry analysis of tagged TentaGel® M NH₂ Monosized Amino TentaGel Microspheres. 6-Carboxyfluorescein. Flow cytometry was performed on the tagged TentaGel® beads (TG-beads) using a BD FACSJazz TM Cell Sorter. The FACSJazz lasers were calibrated before the samples were run with the Rainbow Calibration Particles, 8 peaks (3.0-3.4 µm). Non-labelled TG-beads were dispersed in 5 mL sheath fluid and put through the machine to check the calibration. Then samples TGCFluor100, TGCFluor73 and TGCFluor36 were run through the FACS dispersed in 5 mL sheath fluid, being observed on laser 488 nm with filter 513/17 nm. All flow cytometry data was processed using BD FACS™ Sortware program.
Rhodamine B. The FACSJazz lasers were calibrated before the samples were run with the BD Rainbow Calibration Particles (8 peaks). Non-labelled TG-beads were dispersed in 5 mL sheath fluid and put through the machine to check the calibration. Then samples TGRhodB100, TGRhodB73, TGRhodB36, TGRhodB17 and TGRhodB1.7 were run through the FACS dispersed in 5 mL sheath fluid, being observed on laser 488 nm with filter 585/29 nm.
Addition of 2-cyanoethyl diisopropyl chlorophosphoramidite to DMT protected 2'-desoxy-2'-fluoro-5vinyl-uridine. 10 (0.88 g), and dimethylaminopyridine (0.14 g, 1.14 mmol) were dissolved in 50 mL of dry DCM, diisopropylethylamine (0.84 mL, 6.51 mmol) and 2-cyanoethyl-diisopropyl-chlorophosphoramidite (0.83 mL, 3.50 mmol) were added. Under an atmosphere of nitrogen, the mixture was then stirred at room temperature for 2 hours. After the DCM was removed in vacuo a thick yellow oil was obtained (13) (1.63 g). 31 P NMR was run on the product, to prove the addition of the phosphoramidite group. No other analysis was run on it because of the air sensitive nature of the product. 31 P NMR (DMSO-d6 capillary in DCM): δ 139.00 (s, 1P).

S4. Aptamer Library Synthesis
General Expedite™ 8909 DNA Synthesiser set up for the synthesis of aptamers and aptamer library. The phosphoramidite samples were dissolved in 20 mL (20 mL for 1 g of sample) acetonitrile (DCM for Tentagel beads), put in bottles and screwed into the synthesiser lines. The other reagents put on to the machine: oxidizer (0.02M iodine, 20% pyridine), Cap A Mix (THF/Pyridine/acetic anhydride 8:1:1), Cap B Mix (10% methylimidazole in THF), deblock (3% trimethylamine in DCM) and ETT activator solution (0.25 M, 5ethylthio-1H-tetrazole in acetonitrile). A leak test is run to check nitrogen is not leaking from the lines. If passed, the lines are then flushed with the new reagents added to them. The beads are added to the column, which is then fitted onto the synthesiser, which is then flushed with acetonitrile several times. The sequence and protocol are then selected using Validate XP connected to the Expedite™ 8909 DNA Synthesiser. The aptamer sequences are then run and monitored using the trityl monitor.
Synthesis of MinE07 Aptamers. All library components were synthesised on an Expedite™ 8909 Nucleic Acid Synthesiser system provided by Biolytic. Phosphoramidites were dissolved in dry acetonitrile to the concentrations as suggested by the supplier, solvents were used as provided. Oligomers were synthesised on a 1 µM scale. Universal UnyLinker support (0.021g) was added to a synthesiser column. See General Expedite™ 8909 DNA Synthesiser set up for the synthesis of aptamers and aptamer library for the setup of the synthesiser. The sequences uploaded on to the synthesiser: Synthesis of MinE07Lib Aptamer Library. -OH modified TG-beads (0.0087 g) were added to a synthesiser column. This was enough to synthesis 200 copies of every possible sequence. Round 1: 3'-PCLinker-fCfC-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. Round 2: Uridine 1: 3'-fU-5' -The TG-beads were taken out of the main column and split in 4 columns (U-Ph, U-Vi, U, U-I). A synthesis circle using compound 12 was performed on column U-Ph. A synthesis circle using compound 13 was run through column U-Vi. A synthesis circle using compound 14 was run through column U. A synthesis circle using unmodified uridine phosphoramidite was run through column U-I. Round 3: 3'-rGfCfCrArArGrGfCfCrGrArArAfC-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. Round 4: Uridine 2: 3'-fU-5' -The same synthesis procedure was performed as for round 2. Round 5: 3'-rG-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. Round 6: Uridine 3: 3'-fU-5' -The same synthesis procedure was performed as round 2. Round 7: 3'-rAfCrGrArArArArGrA-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. Round 8: Uridine 4: 3'-fU-5' -The same synthesis procedure was performed as round 2. Round 9: 3'-rGfCfCrGfC-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. Round 10: Uridine 5: 3'-fU-5' -The same synthesis procedure was performed as round 2. Round 11: 3'-rArA-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. Round 12: Uridine 6: 3'-fU-5' -The same synthesis procedure was performed as round 2. Round 13: Uridine 7: 3'-fU-5' -The same synthesis procedure was performed as round 2. Round 14: Uridine 8: 3'-fU-5' -The same synthesis procedure was performed as round 2. Round 15: 3'-rArGrGfCrArGrG-5'. This sequence was loaded onto the synthesiser and run on all TG-beads. The weight of the final MinE07Library TG-beads was 0.0029 g. Table 1. Response values from the trityl monitors for each round of synthesis. Note that the trityl yields in the aptamer library synthesis go up and down during the split and mix depending on the exact portion of TG-beads taken from the total batch. S5. Fluorescence-activated bead sorting Two-way sort of Accudrop Beads to confirm photo-bleaching theory. Accudrop beads (2 mL) were put into a FACS sized falcon tube. Standard calibration was performed. The drop-delay value was adjusted while viewing BD FACS Accudrop beads in the centre and side sort streams that are illuminated by a red laser. Sort monitoring was undertaken with live video feed of breakoff point, waste collection, and side streams. The accudrop beads were sorted into two gates. The beads that were sorted into positive gate were then put back through the FACS and sorted again in gates. This was repeated again. The 488 nm laser was used with filter 513/17 nm. All data was collected and analysed using the BD FACS Sortware sorter software program.

Round Nucleotides
Two-way sort of 100% 6-carboxyfluorescein tagged microspheres from plain microspheres to confirm photo-bleaching theory. Standard calibration was performed. 6-carboxyfluorescein tagged TG-beads (2 mL) were put into a FACS sized falcon tube. They were sorted into two gates. The microspheres that were sorted into positive were then put back through the FACS and sorted again in gates P7 and P8. This was repeated again. The 488 nm laser was used with filter 513/17 nm.
Sorting MinE07Lib to extract the top binding MinE07 modified aptamers. All of the MinE07Library (1 µM, 0.0029 g) was added to an Eppendorf tube. Binding buffer (100 µL) was added and the tube was incubated at room temperature for 15 minutes. This was centrifuged at 300rpm for 1 minute; the binding buffer was removed and the microspheres were then washed with binding buffer (2 x 100 µL). EGFR-Fc (0.03357 mg/mL, 100 µL) was then added to the tube and this was incubated at room temperature for 60 minutes. This was spun down at 300 rpm for 1 minute, the EGFR-Fc was removed, and the microspheres were then washed with wash buffer (2 x 100 µL). Protein A-FITC (0.03357 mg/mL, 100 µL) was then added to the tube and this was incubated at room temperature for 60 minutes. This was spun down at 300 rpm for 1 minute, the Protein A-FITC was removed, and the microspheres were then washed with wash buffer (2 x 100 µL). Sheath fluid (4 mL) was added, and the microspheres sonicated. The sample was then added to a FACS sized falcon tube.
Standard calibration was performed. The MinE07Lib-EGFR-Fc-Protein-A-FITC microspheres sample was twoway sorted three times, changing the gates each time to match the photobleaching and to select less each time. The FACS was calibrated using the 96 well plate set up program along with the accudrop beads. The top microspheres selected from the first three rounds of sorting the MinE07Lib-EGFR-Fc-Protein-A-FITC microsphere were then sorted across two 96 well plates with 100 µL of RO water. The 488 nm laser was used with filter 513/17 nm. All data was collected and analysed using the BD FACS Sortware sorter software program. Analysis of top Aptamers LC-MS/MS data. The data generated by LC-MS/MS of the top hit aptamers MinE07UA, MinE07UB and MinE07UC was analysed using the program RoboOligo. 8 Unlabelled peaks in the MS/MS graphs correspond to fragments that are less than 4 nucleotides long, and so could correspond to multiple places along the chain, or are too close to the limit of detection to be considered as accurate                                                               Desalt of MinE07-Biotin, MinE07-U-Ph-Biotin, MinE07-U-Vi-Biotin and MinE07-U-I-Biotin Aptamers using Zetadex. Purification of the RNA aptamers was carried out by size exclusion gel chromatography using Zetadex resin (emp Biotech). A 20 mL column was prepared with 1 mL of cotton wool at the bottom of the column. A slurry of zetadex resin and autoclaved water was prepared for use in the column. 1 mL of aptamer was loaded onto the column and the column was flushed with deionised water. 1 mL fractions were collected. This was repeated for each aptamer, with a new column each time. The fractions were run on the Nanodrop UV-Vis spectrophotometer. The fractions containing an RNA signal were pooled. dried down and resuspended in reverse-osmosis purified water and stored in the freezer (-20°C).
PAGE-based analysis of Aptamers. TBE denaturing polyacrylamide gels were prepared from a denaturing 20% polyacrylamide stock solution and TBE buffer, with the final concentrations of polyacrylamide at 15%. The gels were polymerised by addition of 5 μL TEMED (1 μL/mL of gel), followed by APS 5 μL (40% stock concentration). The solution was mixed thoroughly and poured between glass plates with 0.75 mm spacers, followed by insertion of a 0.75 mm comb. The gel was then left to fully polymerise. Approximately 80% of the cassette was filled with TBE buffer. The comb was removed after polymerisation and the wells were flushed with deionised water, and then buffer. 10 μL at 20 nM of each aptamer sample in water was prepared along with 10 μL urea (8 M). 20 µL of each sample was loaded onto the polyacrylamide gel. All gels were run at 300 V, 15 mA, for 60 minutes. Gels were stained in Stains-All prepared in isopropanol-tris buffer, for over an hour. The gels were then rinsed to remove excess stain using water before being imaged on an Epsom scanner.
PAGE-based Purification of Aptamers. MinE07-Biotin, MinE07-U-Ph-Biotin, MinE07-U-Vi-Biotin and MinE07-U-I-Biotin aptamers required purification by Poly Acrylamide Gel Electrophoresis (PAGE). A 1.5 mm 15% denaturing PAGE gel was produced by diluting 37.5 mL of 20% acrylamide denaturing stock solution with 12.5 mL of 1 x TBE buffer. To this 50 µL of TEMED and 130 µL of 40% APS stock solution were added to induce polymerisation. The gel was poured between two glass plates and a 1.5 mm comb was inserted. The comb was removed after casting and the well washed with TBE buffer. The gel was pre-run at 300 V for 1 hour before the sample was loaded. The aptamers with 8 M urea were loaded onto the gel and run at 250 V for half an hour before being run for a further 1.5 hours at 300 V. The gel was removed from the glass plates and placed onto cling film, which was then placed onto a silica TLC plate; the gel was illuminated with UV light to visualise each aptamer. The aptamer band was cut out of the bulk gel and placed into a 15 mL falcon tube. The gel was homogenised, and 10 mL of autoclaved water was added. The solution was mixed and rapidly frozen in liquid nitrogen before being incubated overnight at 60 °C in a water bath. The solution was split into two equal portions and centrifuged for five minutes, the supernatants were collected, and the pellet extracted using a pipette with 2 mL autoclaved water. This was repeated three times. The supernatants were dried by centrifugal vacuum concentration at 60 °C for five hours and the pellets were re-suspended in autoclaved water.
Ethanol Precipitation of Aptamers after PAGE Purification. Aptamers were desalted by ethanol precipitation; the samples were incubated with 3 M sodium acetate and 100% ethanol over night before being centrifuged for 20 minutes. The supernatant was removed, and the pellet washed with 1 mL of ethanol and centrifuged again; this was repeated three times for each sample. The samples were then left to air dry before being re-suspended in autoclaved water.
Determination of Aptamer concentration by UV-Visible spectrophotometry. Aptamer concentrations were analysed by UV-Visible spectrometry using a Nanodrop spectrophotometer.UV-visible absorption spectra were recorded on a NanoDrop One UV-Vis spectrophotometer by Thermo Fisher Scientific. 2 µL of deionised water was placed onto the stage and this blank was run through the machine. The stage was cleaned with deionised water and a Kimtech wipe. 2 µL of sample was then placed onto the stage and a spectrum was run between the regions of 200 -360 nm. Each sample was repeated until a minimum of three concordant results were achieved. Using the calculated A260 from Intregrated DNA Technologies (IDT) (2.11) and the measured A260 from the UV-Vis data, a concentration was calculated by multiplying them together.                   Streptavidin-AP at a 1:1000 dilution in binding buffer was prepared, 20 µL was added to each well and it was incubated for 30 minutes at room temperature shaking at 450 rpm with 20 µl of Tropix CDP-Star ready to use substrate. The plate was then read using the chemiluminescence protocol on the Victor X4 plate reader to check it was giving a signal. 20 µL of HABA (5 µM) was added to the wells and it was incubated for 30 minutes at room temperature shaking at 450rpm. MinE07UA-Biotin, MinE07UA'-Biotin, MinE07UB-Biotin, MinE07UC-Biotin and MinE07UC'-Biotin were added to relevant wells and incubated for 30 minutes at room temperature shaking at 450rpm. The plate was then read using the chemiluminescence protocol on the Victor X4 plate reader.

S8. Aptamer Validation
EGFR Binding Assay with MinE07UA-Biotin, MinE07UA'-Biotin, MinE07UB-Biotin, MinE07UC-Biotin and MinE07UC'-Biotin. The protein A plates were washed 3 times with 150 µL of wash buffer per well. EGFR-Fc protein was prepared in wash buffer at 1 µg/mL. 100 µL EGFR-Fc coating buffer per well was added and incubated for 30 minutes at 450 rpm at room temperature. The liquid was flicked out of the well plate. The plate was washed 3 times with wash buffer and then 50 µL 1x binding buffer to each well was added, and then incubate at 450 rpm for 10 minutes. MinE07UA-Biotin, MinE07UA'-Biotin, MinE07UB-Biotin, MinE07UC-Biotin and MinE07UC'-Biotin aptamers were prepared at concentrations 500 nM, 158 nM, 50 nM, 15.8 nM, 5 nM, 1.58 nM, 1 nM and 0.5 nM in binding buffer. The diluted aptamer was then denatured in the Eppendorf master cycler PCR machine (85°C for 5 minutes, then cooled to 25°C at 0.1°C per second and held at 25°C). The binding buffer was removed from the plate and washed 3 times with wash buffer. 100 µL of aptamer was added to the relevant wells. This was then incubated for 60 minutes at room temperature at 450 rpm.
The liquid was flicked out of the well plate, and it was washed 3 times with wash buffer. Streptavidin-AP at a 1:1000 dilution in binding buffer was prepared, 100 µL was added to each well and it was incubated for 30 minutes at room temperature shaking at 450 rpm. The plate was then washed with wash buffer and then 95 µL of Tropix CDP-Star ready to use substrate was added and incubated for 10 minutes with shaking at 450rpm. The plate was then read again using the chemiluminescence protocol was then read on the Victor X4 plate reader. This was repeated until 3 biological repeats were achieved. Data analysis was done on Prism GraphPad Version 9.0.0 using the one site -specific binding model to calculated Bmax and Kd.

S9. Computational Docking Studies
ViennaRNA 2.0 was used to model a 2D predicted secondary structure of the MinE07 aptamers, using the thermodynamic structure prediction and the RNAfold web server. 9 The dot and bracket sequence was used to generate a PDB file to create 3D structure using RNA composer. 10,11 The PDB file was opened in text editor and all the 2'OH atoms are deleted from uridine and cystine. This PDB can be visualised using the cross platform molecular builder and editor program, Avogadro (avogadro.cc). 12 The geometry was then optimised using the optimise geometry option in Avogadro with force field MMFF94, which is in a vacuum. EGFR PBD file (1IVO) was taken from the Protein Data Bank. 13 For docking studies the following programs were used Autodock Tools (MGLTools 1. 5    The results for the optimal secondary structure gave a minimum free energy of -9.30kcal/mol. ViennaRNA was used to model a 2D predicted secondary structure of the control MinE07 aptamers (rna.tbi.univie.ac.at/). This was done by using the thermodynamic structure prediction and the RNAfold web server.  .)))))..)))))) Figure 162. The MinE07 dot and bracket sequence used to generate a PDB file. The dot and bracket sequence was used to generate a PDB file to create 3D structure using RNA composer (rnacomposer.cs.put.poznan.pl/).