SELEX was performed as previously reported with some modifications (18, 19). Dynabeads™ MyOne™ Carboxylic Acid (CA) magnetic beads (Invitrogen, Waltham, MA) were used for NCYM solidification to segregate NCYM-binding DNA molecules from the non-binding molecules. The target beads were prepared by an amine coupling reaction using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC; Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions and washed with the selection buffer [SB; 40 mM HEPES (pH 7.5), 125 mM NaCl, 5 mM KCl, 1 mM MgCl₂, and 0.01% Tween 20]. Briefly, the CA magnetic beads were washed twice with 500 μl of 100 mM 2-Morpholinoethanesulfonic acid mono hydrate (MES) buffer at pH 6.0. After added 50 μl of 100 mM MES buffer and 50 μl of EDC, and the mixture was incubated for 30 minutes at room temperature. The mixture was then mixed with 200 μg of NCYM in 100 mM MES buffer and incubated over night at room temperature to react with the amino group of NCYM and the carboxylic acid of the CA magnetic beads. Next, they were washed twice with 500 μl of PBST (0.1% Tween20 in phosphate buffered saline, PBS) and 500 μl of PBST-BSA (0.1% bovine serum albumin in PBST) was added. Before using the target beads, they were washed twice with 1 ml of SB.

An initial single-stranded DNA (ssDNA) pool, 5′-GGAATGTGGTCCCTCGCAATAAATC-(N30)- GAAATGAGCCCTTTGACCCTGTAC-3′, containing 30 random nucleotides between forward (Fw) and revers (Rv) primer region was purchased from Integrated DNA Technologies (Tokyo, Japan). The selection of aptamers was performed starting from 4.5 nmol of DNAs (~10¹⁵ molecules) in 100 μl of SB. The pool was mixed for 15 minutes with 250 μg of target beads at 25°C. The beads were then washed with SB, and the bound ssDNA was eluted with 7 M urea. After recovery of the eluted ssDNA using Rv primer beads, polymerase chain reaction (PCR) was performed with KOD Dash DNA polymerase (Toyobo, Osaka, Japan), a forward (Fw) primer (5’-GGAATGTGGTCCCTCGCAATAAATC-3’) and a reverse (Rv) primer (5’-GTACAGGGTCAAAGGGCTCATTTC-3’) with modification by the 5’-biotin. Next, the amplified double stranded DNA (dsDNA) was bound to MyOne SA C1 magnetic beads, and the Fw chain (ssDNA) was eluted with 0.02M NaOH. The ssDNA was used for the next round.

After eight rounds of selection, the frequency of ssDNA sequences was determined by next-generation sequencing (NGS) from rounds 3 to 8 of SELEX using a MiniSeq System (Illumina, San Diego, CA, USA). Sequencing data were preprocessed by using the program of PRINSEQ++ (20) and adopted above 99.9% of the base calling accuracy (Q score of 30 and above).

All the Bio-Layer Interferometry (BLI) measurements were performed at 25°C using an Octet® RED96e system (Sartorius AG, Goettingen, Germany). All the samples were placed in a 96 well microplate and the sample volume was 200 μl/well. The microplate was shaked at 1,000 rpm during the measurement. As a ligand, each aptamer with 5’-biotin modification was immobilized on an Octet® SA biosensor chip (Sartorius). For kinetics analysis, different concentrations of NCYM (25-400nM) were used. The dissociation constants between the aptamer and NCYM were calculated using a simple 1:1 biomolecular interaction model according to the manufacturer’s instructions.

To gain insights into the mechanism of interactions between NCYM and the aptamers obtained by SELEX, NCYM-aptamer binding was studied at the single molecule level using AFM. Scaffolds prepared using the DNA origami method were used for this purpose. The DNA origami method allows the creation of structures of any shape and the introduction of functional molecules anywhere in the structure. Therefore, single-molecule observation using DNA origami structures is suitable for evaluating biomolecules and has been used to observe various molecules (21–29). In this study, we used a DNA frame structure (21). The DNA frame contains a space inside, and dsDNA can be introduced into any sequence.

DNA frames were prepared as previously described (21). A solution containing 10 nM M13mp18 ssDNA (tilibit nanosystems GmbH, Germany), 25 nM staples (2.5 eq), 20 mM Tris-HCl (pH 7.4), 10 mM MgCl₂, 1mM EDTA was prepared and annealed at a rate of -1°C per minute from 85°C to 15°C.

Aptamers and DNA oligos for the DNA frame were purchased from Eurofins Genomics K.K. (Tokyo, Japan) and used without further purification. In this study, three types of aptamers, which showed high affinity among the aptamers obtained by SELEX, were employed, and the sequences of which for AFM observation are as follows:

All aptamers had a dsDNA complement sequence (underlined) added via a TTTT sequence (bold) at the 3’ end for introduction into the DNA frame.

The secondary structures of the aptamers were predicted using RNAfold [http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi]. We used the default settings: minimum free energy (MFE) and partition function, yes; avoid isolated base pairs, yes; incorporate G-Quadruplex formation into the structure prediction algorithm, yes; dangling end options, dangling energies on both sides of a helix in any case; energy parameters, DNA parameters (Mathews model, 2004) at 25 or 38°C with a salt concentration of 1.021M.

The resultant dsDNAs were incorporated into the frame structures described above and observed in the presence of NCYM ( Figure S1 ), which was purified according to a previous method (1).

To identify where the aptamers interact with NCYM, we prepared three analogs (short 1, short 2, and short 3) of the No. 1 aptamer based on the secondary structure prediction by RNAfold. The sequences used are as follows:

short 1, short 2, and short 3 lack the 5’ end and 3’ end side (named “5’-3’-end” herein), the 3’ end stem loop, and the central stem loop of the No. 1 aptamer, respectively.

In addition, mutations were introduced into aptamer No.1 without changing its secondary structure. The following sequences are mutants of aptamer No.1, and italic font indicates the introduced mutations.

The AFM images were acquired using an AFM system (NanoWizard UltraSpeed, JPK) equipped with a silicon nitride cantilever (Olympus, BL-AC40TS). In all the measurements, NCYM (100 nM) 10eq was added to the adjusted frame structures having aptamer (10 nM) and incubated at 25°C for 2h. Samples were then double-diluted in annealing buffer (20 mM Tris-HCl (pH 7.4), 10 mM MgCl₂, 1mM EDTA), adsorbed onto fresh mica plates for 10 minutes at room temperature, and washed three times with the annealing buffer. The observations were performed using the same buffer.

20 μl of Dynabeads™ protein G (Thermo Fisher Scientific) slurry was transfered to a clean tube. The tube was placed in a magnetic separation rack for 10-15 seconds, then, the buffer was carefully removed. 2 µg of anti-GSK3β (BD Transduction Laboratories) or mouse IgG (Cell Signaling Technology, Danvers, MA) was dissolved in 200 μl of PBS (0.02% Tween 20) and the solution was added to the beads. The slurry was incubated with rotation at room temperature for 10 minutes. The beads were pelleted using a magnetic separation rack and washed three times with 200 μl of PBS, then resuspended in 100 μl of PBS. 0.17µg (3.7 pmol) of purified GSK3β (Signal Chem, Richmond, Canada), 0.044g (3.7 pmol) of NCYM (1) and DNA aptamers (1eq (3.7 pmol) or 5 eq (18.5 pmol)) were dissolved in 500μl of ice-cold PBS and incubated with rotation at 4°C for 2h. A 100 μl suspension of antibody-conjugated beads was added and incubated for 2h. The beads were pelleted using a magnetic separation rack and washed five times with 1 ml of PBS. The pellet was resuspended in 20 μl of 1X sample buffer and heated at 95°C for 5min.The beads were pelleted using a magnetic separation rack and the supernatant was subjected to the Abby analysis.

The NCYM and GSK3β protein levels were measured using a capillary electrophoretic-based immunoassay (the Abby instrument; ProteinSimple, San Jose, CA, USA), according to the manufacturer’s protocol. Briefly, the samples were combined with 0.1× sample diluent buffer and 5× fluorescent master mix denaturing buffer to acquire 0.8 µg/μl loading concentration. Subsequently, the samples were denatured for 5 min at 95 °C. The primary antibodies used in this study were anti-NCYM (1) and anti-GSK3β (#9315, Cell Signaling Technology, Danvers, MA, USA). The Abby measurements were performed using a 12–230 kDa separation module with 25-min separation at 375 V, 10-min blocking, 30-min primary antibody incubation, and 30-min secondary antibody incubation (DM-001, ProteinSimple, San Jose, CA, USA). At the end of the run, the chemiluminescent signal was displayed as a virtual blot-like image and an electropherogram based on the molecular weight using Compass (ProteinSimple, San Jose, CA, USA).

The expression and purification procedures for NCYM were modified from those described previously (8). Recombinant NCYM with glutathione S-transferase (GST) at the N-terminus was expressed in Escherichia coli BL21 (DE3) cells using the pGEX-6p-1 vector. The cells were cultured in Luria broth medium containing 0.1 mg/ml ampicillin at 30°C. When OD₆₀₀ reached 0.6, protein expression was induced by 0.1 mM of isopropyl-β-D-thiogalactopyranoside, and culture was continued for 6 h at 30°C. The cells were harvested by centrifugation (3,890×g, 15 min, 4°C), and then stored at −30°C until purification.

The frozen cell pellets were thawed and resuspended in phosphate-buffered saline (PBS) supplemented with cOmplete EDTA-free Protease Inhibitor Cocktail (11873580001, Roche, Basel, Switzerland) and lysed by repeated sonication in ice water. The cell lysate was centrifuged at 20,000×g, 4°C for 20 min, then the supernatant was loaded onto GSTrap FF column (17513102, Cytiva, Marlborough, MA, USA) equilibrated with PBS, using a peristaltic pump. After washing out the unbound materials with PBS, GST-tagged NCYM (GST-NCYM) was eluted with an elution buffer (50 mM Tris-HCl and 10 mM reduced glutathione, pH 8.0). The flow-through was reloaded onto the re-equilibrated column, and the eluate was collected once more to increase the final protein yield.

The eluted GST-NCYM solution was dialyzed against a buffer containing 50 mM Tris-HCl, 100 mM NaCl, and 1 mM EDTA at pH 8.0. After adding dithiothreitol to the protein solution at a final concentration of 1 mM, the GST-tag was cleaved with 50 U/L-culture of PreScission Protease (27084301, Cytiva) for over 18 h at 4°C with gentle stirring. To exhaustively degrade the remaining nucleic acids, 6,000 U/L-culture of benzonase (71205-3CN/70746-3CN, Merck, Darmstadt, Germany) was added to the cleaved sample along with MgCl₂ at a final concentration of 2 mM (Mg²⁺ is required for the activation of benzonase), after which the sample was dialyzed against IEX buffer (20 mM MOPS (pH 7.0) and 1 mM DTT) with 2 mM MgCl₂.

The dialyzed sample was loaded onto a HiTrap SP HP column (17115201, Cytiva) equilibrated with IEX buffer, and eluted using a linear gradient of NaCl (0–650 mM). The fractions containing high-purity NCYM, as confirmed by SDS-PAGE, were collected and used for SAXS measurements. The GST-tag did not bind to the column and was detected in the flow-through. The final NCYM yield was 1.25 mg/L-culture, which was estimated with the molecular absorption coefficient of 280 nm ϵ₂₈₀ ^0.1% = 0.558.

The purified NCYM was dialyzed against a buffer containing 10 mM Tris-HCl (pH8.0), 50 mM NaCl, and 5 mM DTT. Powdered DNA aptamers (No. 1) were dissolved directly in the same buffer. Samples of the NCYM-DNA complex were prepared by mixing these two kinds of solutions at the appropriate molar ratios (see below). These solutions were used for the following SAXS measurements.

SAXS measurements were carried out at BL40B2 in SPring-8 (Hyogo, Japan) on solution samples of NCYM (1.4 and 2.8 mg/ml), the DNA aptamer (1.6 and 5.0 mg/ml), and the NCYM-DNA complex. For the measurements of the complex, two kinds of samples, where the molar ratio of NCYM and DNA was 1:1.2 or 1:1.5 ( Table 1 ), were used to extract the scattering curve of the complex by changing the relative contribution of the unbound DNA aptamers. The wavelength (λ) of the incident X-ray was 1.0 Å and the temperature was 293 K with the sample-to-detector distance of 2.2 m. A pixel detector (PILATUS3S 2 M, Dectris) was used to record the scattering patterns.

Data reduction was conducted using the software SAngler (30): The recorded two-dimensional SAXS patterns were circularly averaged to obtain one-dimensional scattering curves, corrected by the incident flux measured with an ion chamber placed upstream of the samples. The net scattering curves of the scattering particles were obtained by subtracting the scattering curves of the buffer from those of the samples with an appropriate scaling factor based on the scattering particle concentration and the partial specific volumes of proteins (0.73 cm³/g) or of DNA (0.53 cm³/g) (31). Finally, the scattering curves were normalized to the absolute scale using H₂O as the standard (32) to estimate the molecular weight of the scattering particles.

Guinier analysis was employed to evaluate the radius of gyration (R_g) of the scattering particle from its scattering curve. A scattering curve I(Q), where Q (=4πsinθ/λ, where 2θ is the scattering angle) denotes the momentum transfer, is represented as follows in a good approximation in the so-called Guinier region (Q･R_g < 1.3) (33):

where I(0) [cm^-1] denotes the forward scattering intensity, from which the molecular weight (MW) of the scattering particle is estimated from the following equation in kDa (32):

where c is the weight concentration [g/l].

For rod-like particles such as the DNA aptamer and the NCYM-DNA complex, cross-sectional Guinier analysis was applied to evaluate their cross-sectional radii of gyration (R_c). In this case, I(Q) is approximated as:

where I_c(0) denotes the forward scattering intensity of the cross-section of the scattering particle. Application of Eq. 2 with I_c(0) instead of I(0) yields the scattering mass per unit length.

The scattering curves taken at the lower particle concentrations were merged with those at the higher concentrations at Q = 0.1 Å⁻¹ and the resultant curves were used for structural modeling. IGOR Pro software (WaveMetrics, Lake Oswego, OR, USA) was used for (cross-sectional) Guinier analyses and for processing the scattering curves.

Human NCYM comprises109 residues. Because NCYM was expressed with a GST-tag at its N-terminus in this study, eight residues were added to the 109 residues even after cutting the tag. For structural modeling, the program GASBOR (34) was employed, where each residue is represented by a sphere with a constant electron density. The obtained model thus consists of 117 spheres. In this study, the GASBOR runs were repeated 10 times (i.e., 10 best-fit models were obtained). For the modeling of the DNA aptamer, DAMMIF (35) was used, where a molecule is represented by an ensemble of spheres called dummy atoms. The input files required for DAMMIF were generated using AUTORG and DATGNOM (36). The maximum Q value (Q_max) used for structural modeling was automatically determined using DATGNOM (Q_maxR_g < 7–8). The DAMMIF runs were repeated 10 times and the resultant 10 models were averaged, followed by filtration using DAMAVER (37). Structural modeling of the NCYM-DNA complex was carried out using the program MONSA (38), where each of the two phases is represented by a dummy atom model while each phase is assigned a designated electron density value. As an input file of MONSA, the following information is required: The values of the electron density of NCYM and the DNA aptamer were set to be 0.09 e/ų and 0.21 e/ų, respectively, which are typical of these types of molecules (39). The volume fractions of NCYM and the DNA aptamer were obtained from the volumes obtained using GASBOR and DAMMIF, respectively. Using these parameters, the MONSA runs were repeated 10 times. In all of the programs above, the best-fit model, the scattering curve of which reproduces the experimental curve well, is determined by simulated annealing.

The scattering curves of the NCYM-DNA complex recorded at 1:1.2 and 1:1.5 molar ratios were found to be superimposable despite the existence of different amounts of unbound DNA aptamers ( Figure S2 ). However, as described in the Results section, NCYM molecules and the DNA aptamers exist as monomers and dimers, respectively, in solution so that the molar ratios of NCYM and the dimeric DNA aptamer in the above samples are 1:0.6 and 1:0.75, respectively, which result in the molar ratios of the complex and the excess amount of NCYM of 0.6:0.4 and 0.75:0.25 assuming that one NCYM monomer binds to a DNA dimer. Because the scattering intensity is proportional to the square of the product of the scattering contrast of a particle and its volume, the scattering contributions of excess NCYM molecules were calculated to be 0.8% and 0.4% for the 0.6:0.4 and 0.75:0.25 samples, respectively. In this estimation, the scattering contrasts of NCYM and DNA were set to be 0.09 e/ų and 0.21 e/ų as described above, and the volumes of a NCYM monomer and a DNA dimer were assumed to be 17000 ų and 68000 ų as obtained by GASBOR and DAMMIF. Thus, it appears that the scattering contribution of free NCYM monomers is negligible, which explains why the scattering curves obtained at NCYM : DNA = 1:1.2 and 1:1.5 are similar to each other within errors. This observation also excluded the possibility that the complex consists of one NCYM monomer and one monomeric DNA aptamer. Based on the above inspection, the curves obtained from the above samples reflect those of the NCYM-DNA complex. The scattering curves taken at NCYM : DNA = 1:1.2 and 1:1.5 were, therefore, averaged to improve the signal-to-noise ratio and the resultant curve was employed as the scattering curve of the NCYM-DNA complex.

To identify DNA aptamers that specifically bind to NCYM, we performed SELEX and employed NGS technology to monitor the progress of the enrichment sequences that bind to the target in the selection pool. We performed the NGS analysis using an Illumina MiniSeq for the NCYM aptamer selection pools from rounds 3 to 8 to identify the aptamer candidates. The ratio of unique DNA sequences in the selection pool per round showed that DNA sequences were enriched in round 6 ( Figure 1A ). No. 1 aptamer candidate, which was the most abundant population in round 8, was quickly enriched from round 6 compared to the other DNA sequences. Guanine contents in the selection pools were increased slightly in round 8 ( Table S1 ). BLI measurements and kinetic analyses showed that top three aptamers strongly bind to NCYM protein at K_D value in the range of 53.9 to 299 nM ( Table S2 ); however, the aptamer No. 2 and 3 were predicted to form different secondary structure at 38°C, showing low structural stability at relatively high temperature compared to aptamer No.1 ( Figure 1B and Figure S3 ).

To clarify the effect of DNA aptamer No.1 on NCYM function, we examined NCYM binding to GSK3β with or without the aptamer. As reported (1), purified NCYM was co-immunoprecipitated with GSK3β ( Figure 4A ). Addition of aptamer No.1 enhanced the interaction between NCYM and GSK3β in dose-dependent manner ( Figure 4B ). This result led us to analyze the NCYM-aptamer No.1 complex because the aptamer may help NCYM to adopt a conformation that facilitates binding to GSK3β.