Aptamer-Antibody Hybrid ELONA That Uses Hybridization Chain Reaction to Detect EN2, A Urinary Biomarker for Bladder and Prostate Cancer

Development of ssDNA aptamer specific to EN2

N-terminal His-tagged recombinant EN2 protein was cloned and purified using a bacterial expression system and immobilized metal affinity chromatography (Fig. S1). The result of SDS-PAGE indicated the purity of the recombinant EN2 to be > 90% (Fig. S1B). DNA aptamers for the recombinant EN2 were selected by magnetic-bead SELEX using a single-stranded DNA library containing approximately 3 x 10¹⁴ molecules. After 12 rounds of selection, an H90 aptamer was obtained, then trimmed to be EBA of 50 mer to ensure that random sequences remained and secondary structure was preserved; EBA had K_d = 8.26 nM (R² = 0.971) (Table S1; Fig. 2), which indicates high binding affinity to EN2.

Design and signal amplification of aptamer-mediated HCR

Trigger, H1, and H2 sequences for HCR were designed by reference to a previous study (Table S2)¹⁸. To evaluate the HCR feasibility of the designed trigger, H1, and H2 sequences, theoretical calculation was first performed using NUPACK software (www.nupack.org)¹⁹. The results showed that the Gibbs free energies between trigger and H1 were − 34.00 kcal/mol, and between H1 and H2 were − 27.44 and − 27.08 kcal/mol, respectively, suggesting that HCR is attainable (Fig. 3A). We also investigated the binding affinity of detector, the connected form of EBA and the trigger, to generate an aptamer-mediated HCR; the EMSA result of the EBA and detector indicated that the binding affinity of the EBA to EN2 was sustained even when it was linked to the trigger sequence (Fig. S2).

To test the designed sequences for HCR in experiments, HCR products for the reaction of detector, H1, and H2 were analyzed using gel electrophoresis. Results (Fig. 3B) showed that HCR was performed when all of the reagents existed. Taken together, the results suggest that the HCR products could interact with EN2 and consequently increase the signal-to-background ratio (S/B), which is defined as A_EN2/A_blank, where A_EN2 represents the absorbance at 450 nm in the presence of EN2, and A_blank represents the absorbance of buffer solution at 450 nm. To compare the S/B, the absorbance intensities of 10 nM EN2 were measured in the ELONA without and with HCR. The S/B without HCR was 1.026 ± 0.054, whereas the S/B with HCR was 3.995 ± 0.679, a 3.89-fold increase in S/B with HCR (Fig. 3C). We also measured signals without the detector to ensure the signal amplification by HCR rather than hairpins H1 and H2 themselves (Fig. S4). Consequently, the HCR increased S/B, and accordingly the sensitivity of the sensor.

Optimization of the key parameters

To achieve the highest sensitivity of EN2 detection, we optimized the conditions, including antibody concentration, parameters affecting HCR, dilution degree of poly-HRP, and binding time of poly-HRP. Each component was determined using a maximum ΔA (A_EN2 - A_blank). First, the optimal concentration of antibody as a capture agent was determined to be 5 µg/mL (Fig. S5). Of parameters that affect HCR, the reaction buffer, spacer length between EBA and trigger, and particular order in which related reactions occur were considered, to facilitate both EN2 binding of EBA and HCR of trigger. Signal intensity was highest when binding buffer was used and 10 consecutive adenosine monophosphates (10A) were inserted as a spacer. No difference occurred between adding the HCR product after HCR in advance (pre-HCR) and adding H1 and H2 after reacting detector only with the EN2 (pro-HCR) (Fig. S6); we selected pre-HCR for further applications because of time efficiency. Among the parameters that affect HCR, reaction buffer had a significant effect on ΔA. The binding buffer contained both MgCl₂ and tween-20, whereas tris buffer previously reported in the literature¹⁵ contained only MgCl₂. MgCl₂ and tween-20 participate in stabilization of DNAs and proteins, respectively^20,21, and these effects may have improved the capabilities of the detector. In addition, incorporation of 1:500 diluted poly-HRP for 5 min was sufficient to generate signals; 1:250 diluted poly-HRP was not selected because it had a high background signal (Fig. S7). These processes yielded an optimal ELONA for detection of EN2.

Analytical characteristics of the ELONA

Once the reaction conditions were established, we assessed the dynamic range, sensitivity, specificity, and stability of ELONA. We evaluated the dynamic range and sensitivity by measuring absorbance intensities from buffers containing EN2 in the range of 0 nM to 200 nM (Fig. 4A). The ELONA showed a good linear relationship between the concentrations of EN2 and signal intensities (R² = 0.9906), so the dynamic range was 0.39 nM to 25 nM and the limit of detection (LOD) was 0.34 nM (11.6 ng/mL). The average coefficient of variation (CV) was 7.08%, which confirms great reproducibility (Fig. 4B). The data are meaningful in that the mean urine EN2 level is 5.76 nM (197 ng/mL) for bladder cancer patients², and is 23.5 nM (804 ng/mL) with the cut-off value of 1.24 nM (42.4 ng/mL) for prostate cancer patients^7,22.

We expanded the capability of the ELONA to enable the detection of EN2 in clinical conditions by applying it to artificial urine medium (AUM). Prior to the urine-spiking test, the AUM percentage of the solution for the ELONA was optimized to be 50% (v/v) (Fig. S8); the signal intensities of the EN2-spiked 5% AUM solutions were higher than those of the 50%, but 50% was selected because of the low background signal and low dilution of EN2 concentrations in real clinical samples. In the EN2-spiked AUM samples, the absorbance curves from 0 nM to 800 nM were obtained (Fig. S9). The dynamic range was determined from 3.12 nM to 50 nM with a LOD of 2.69 nM and R² = 0.9966 (Fig. 4C); the real minimal detectable content of EN2 in urine samples is 5.38 nM (184 ng/mL). The average CV was 13.40%, and the recoveries, calculated as the percentage of the practical concentration divided by the theoretical concentration, varied from 88.84–104.67% (Table S3).

The specificity of the ELONA was evaluated in AUM using prostate specific antigen (PSA), cancer antigen 125 (CA125), IgG, and bovine serum albumin (BSA) (Fig. 5); PSA and CA125 are known as genitourinary-related tumor markers and can coexist in clinical samples^23,24, which may interfere with accurate detection of EN2. Signals for proteins other than EN2 were negligible. These results suggested that the developed ELONA is reproducible and reliable to detect EN2 in urine-added conditions.

The ELONA was stable enough to be stored for more 3 months at 4℃ (Fig. S10); this stability permits use of pre-coated plates, and can thereby reduce the detection time to < 2.5 h. Thus, the developed ELONA can be practically applied to the diagnosis of bladder and prostate cancer by testing urine. In order to further highlight the advantages of the designed ELONA in detecting EN2, the analytical properties were compared with those of other methods (Table 1).

Materials

Genomic DNA of EN2 was purchased from Thermo Fisher Scientific (USA). The pET28a plasmid was purchased from Novagen (Germany). Nucleospin Gel and PCR Clean-up/Plasmid EasyPure kit were acquired from Macherey-Nagel (Germany). E. coli strain Rosetta (DEC), Dynabeads his-tag Isolation and Pulldown, MyOne Streptavidin C1, and TOPO TA cloning kit were purchased from Invitrogen (USA). HisTrap Ni-NTA affinity column and Sephadex G-25 Superfine resin were purchased from GE Healthcare (USA). All DNAs were synthesized from Bionics Inc. (South Korea). For ELONA, EN2 monoclonal antibody was purchased from Abnova (Taiwan). Streptavidin Poly-HRP (poly-HRP), 3,3′,5,5′-tetramethylbenzidine (TMB) substrate kit, 96-well polystyrene plates, and sulfuric acid (H₂SO₄) were purchased form Thermo Fisher Scientific (USA). Bovine serum albumin (BSA) and IgG were purchased from Sigma-Aldrich (USA). Prostate specific antigen (PSA) and cancer antigen 125 (CA125) were purchased from Sino Biological Inc. (USA) and Abnova (Taiwan). Artificial urine medium (AUM) was purchased form Pickering Laboratories Inc. (USA). All materials were of analytical grade. All aqueous solutions were prepared in deionized water (> 18 MΩ) obtained using a Direct-Q system from Merck Millipore (USA).

Preparation and purification of EN2

The EN2 genes, which corresponds to protein P19622 (UniProtKB database), were amplified using PCR with sequence specific-forward primers including BamHⅠ linker (EFP, 5’-CCC GGA TCC ATG GAG GAG AAT GAC CCC AAG C-3’) and reverse primers including XhoⅠ linker (ERP, 5’-CCC CTC GAG CTA CTC GCT GTC CGA CTT GC-3’). PCR products were purified, and subsequently digested using BamHⅠ and XhoⅠ. They were cloned into pET28a vectors that had been pre-digested under the same conditions, which leads to the addition of the His-tag at the N-terminal of EN2. The plasmids were then transformed into E. coli Rosetta (DEC) to allow over-expression of EN2.

Bacteria were grown at 37℃ in lysogeny broth (LB) supplemented with 50 µg/mL kanamycin and 25 µg/mL chloramphenicol until the optical density (OD) at 600 nm reached 0.6. Point expression was induced by addition of 200 µM isopropyl β-d-1-thiogalactopyranoside (IPTG) and incubation at 37℃ for an additional 6 h.

Bacterial cells were harvested and sonicated, then the lysate was cleared by centrifugation at 18,000 rpm for 40 min and applied to a HisTrap Ni-NTA column. The fusion proteins were eluted with an imidazole gradient, then the eluates were added to a desalting column with storage buffer (50 mM Tris-HCl, 100 mM NaCl, 0.5 mM β-mercaptoethanol, and 5% (v/v) glycerol, pH 8.0). The products were stored at -80℃ in aliquots containing 20% (v/v) glycerol, then analyzed using 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

in vitro selection for EN2-specific ssDNA aptamer

The overall in vitro selection was conducted using magnetic bead SELEX and performed in 100 µL of binding buffer (20 mM Tris–HCl, 50 mM NaCl, 5 mM KCl, 5 mM MgCl₂, pH 8.0)^27,28. A library template was synthesized as ssDNA containing a central random region of 40 nucleotides (DNA library: 5’-CAC CTA ATA CGA CTC ACT ATA GCG GAT CCG A-N₄₀-CTG GCT CGA ACA AGC TTG C-3’). For the selection, 20 µL of pre-washed NTA magnetic beads (Dynabeads His-Tag Isolation & Pulldown) were incubated with 500 pmol of His-tagged EN2 for 1 h at room temperature (RT), then washed to remove unbound proteins using an external magnetic separator. Then 500 pmol of ssDNA library was heated at 95℃ for 5 min then cooled on ice for 1 h to stabilize the naturally-occurring secondary structures, then incubated with EN2-immobilized magnetic beads for 1 h at RT. Unbound ssDNAs were collected and measured by UV absorbance at 260 nm to calculate the amount of bound ssDNAs.

Then the EN2-ssDNA complexes were eluted using binding buffer supplemented with 300 mM imidazole, and the eluted ssDNAs were precipitated using 70% (v/v) ethanol and amplified by PCR with pfu polymerase, using forward primers (5’-CAC CTA ATA CGA CTC ACT ATA GCG GA-3’) and biotinylated reverse primers (5’-biotin-GCA AGC TTG TTC GAG CCA G-3’). The resulting dsDNAs were added to 70 µL of streptavidin-coated magnetic beads (Dynabeads MyOne Streptavidin C1) in coupling buffer (5 mM Tris-HCl, 1 M NaCl, 0.5 mM EDTA, 0.0025% (v/v) tween-20, pH 7.5) for 1 h at RT, then washed in coupling buffer. Then non-biotinylated ssDNAs were eluted using 200 mM NaOH and sequentially precipitated using 70% (v/v) ethanol. The generated ssDNAs were used as the library for the next round of SELEX. After the 12th round of SELEX, the eluted ssDNAs from the EN2-immobilized magnetic beads were amplified by PCR using unmodified primers.

Finally, the amplified dsDNAs were cloned into pCR 2.1-TOPO TA vectors, and the constructs were transformed to E. coli TOP10 cells (TOPO TA Cloning Kit). The plasmids were purified using a Nucleospin Plasmid EasyPure kit, and the inserts were sequenced. The secondary structures of the aptamer candidates were predicted using the Mfold program (http://www.unafold.org/mfold/applications/dna-folding-form.php)²⁹.

Measurement of dissociation constant ( K _d )

The dissociation constants K_d were determined using a magnetic-bead fluorescent assay. 5’ FAM modified aptamers were heated at 95℃ for 5 min then cooled on ice for 1 h in binding buffer prior to use. First, 100 pmol of His-tagged EN2 was shake-incubated with 10 µL of pre-washed NTA magnetic beads for 1 h at RT, then the EN2-immobilized beads were separated using an external magnetic separator. Next, the EN2-immobilized beads were reacted with several concentrations (0–100 nM) of 5’-FAM modified aptamers for 1 h at RT in darkness. The suspensions were washed twice to remove unbound aptamers, then the EN2-aptamer complexes were obtained using binding buffer supplemented with 300 mM imidazole. Finally, the fluorescence intensities of the eluates were measured by 485/20 nm excitation and 528/20 nm emission filter on a Synergy™ HT microplate reader (BioTek, USA). K_d values were calculated by fitting the data to the one-site-binding (total and nonspecific binding) model of GraphPad Prism 5.0 software (GraphPad, USA)³⁰.

Aptamer-antibody hybrid ELONA that uses HCR

In preparation, 5’biotinylated detector, hairpin 1 (H1), and hairpin 2 (H2) were heated at 95°C for 5 min, then cooled on ice for 30 min in binding buffer. First, 96-well polystyrene plates were coated with 5 µg/mL of EN2 monoclonal antibody (100 µL) at 4℃ overnight. The plates were blocked using 200 µL of blocking buffer (10 mg/mL BSA in PBS) for 1 h at RT. Subsequently, the plates were washed with wash buffer (0.05% (v/v) tween-20 in PBS), then a two-fold serial dilution of EN2 (0, 0.20, 0.39, 0.78, 1.56, 3.13, 6.25, 12.5, 25, 50, 100 and 200 nM) in PBS was added individually to wells of the plates for 1 h at RT. At the same time, detector, H1, and H2 were mixed to 100 nM each, then reacted for 1 h at RT to complete the HCR process. Then the plates were washed three times with wash buffer, and the HCR products were treated for 1 h at RT. The plates were washed three times, then incubated with 100 µL of poly-HRP diluted 1:500 in blocking buffer for 5 min at RT. The plates were washed five times, then reacted with TMB solution for 20 min at RT in darkness. The reactions were stopped by 1 M sulfuric acid, and the absorbance at 450 nm (A₄₅₀) was measured using a Synergy™ HT microplate reader (BioTek, USA). For EN2-spiked urine samples, a two-fold serial dilution of EN2 from 800 nM was prepared in PBS including 50% (v/v) AUM, and other procedures were performed as above. All data were obtained in triplicate at each EN2 concentration. To analyze the sensitivity of the assay, the limit of detection (LOD) was calculated as 3σ/s, where σ is the standard deviation (s.d.) of the sample absorbance, and s is the slope of the linear relationship between A₄₅₀ and EN2 concentration. The coefficient of variation (CV) was calculated as the ratio of s.d. to the mean.