Aptamer-targeting of Aleutian mink disease virus (AMDV) can be an effective strategy to inhibit virus replication

Aleutian mink disease (AMD), which is caused by Aleutian mink disease virus (AMDV), is an important contagious disease for which no effective vaccine is yet available. AMD causes major economic losses for mink farmers globally and threatens some carnivores such as skunks, genets, foxes and raccoons. Aptamers have exciting potential for the diagnosis and/or treatment of infectious viral diseases, including AMD. Using a magnetic beads-based systemic evolution of ligands by exponential enrichment (SELEX) approach, we have developed aptamers with activity against AMDV after 10 rounds of selection. After incubation with the ADVa012 aptamer (4 μM) for 48 h, the concentration of AMDV in the supernatant of infected cells was 47% lower than in the supernatant of untreated cells, whereas a random library of aptamers has no effect. The half-life of ADVa012 was ~ 32 h, which is significantly longer than that of other aptamers. Sequences and three dimensions structural modeling of selected aptamers indicated that they fold into similar stem-loop structures, which may be a preferred structure for binding to the target protein. The ADVa012 aptamer was shown to have an effective and long-lasting inhibitory effect on viral production in vitro.

virus-1 27,28 , hepatitis C virus 24,29 , influenza virus 30,31 , hepatitis B virus 18 , dengue virus 2 32 and rabies virus 33 . We believe that aptamers can be used to supplement antibody treatments 34 . In this study, we obtained aptamers for AMDV VP2 protein using in vitro magnetic beads-based SELEX, and evaluated the antiviral activity of the selected aptamers in an AMDV infection assay. The identification of antiviral aptamers would represent an important development in the search for treatments for AMDV.

Results
Generation of aptamers against AMDV VP2 protein. Recombinant AMDV VP2 protein was expressed and purified by Ni-chelating affinity chromatography. The purified VP2 protein (aa: 200-588, ~ 55 kDa) was detected as a single band, with the expected molecular weight of ~ 55 kDa (Fig. 1a). The purified VP2 protein was also recognized by VP2-specific monoclonal antibodies (mAbs) and rabbit polyclonal antibody (pAbs) using a western blot assay (Fig. 1b). The full length gels and blots images was included in the Supplementary Information file.
The purified VP2 protein was coupled to Dynabeads MyOne Carboxylic Acid magnetic beads in sodium acetate buffer (pH 4.0). Using flow cytometry, the peak for magnetic beads coupled to VP2 protein was significantly right-shifted compared with that of blank magnetic beads (Fig. 1c). The magnetic beads coupled to recombinant AMDV VP2 protein can be used for target selection.
SELEX was used to identify AMDV VP2-binding ssDNA aptamers from a random 40-nt ssDNA aptamer library consisting of ~ 10 14 sequences. At every SELEX cycle, negative selection against canine parvovirus (CPV) VP2 protein-coupled magnetic beads was carried out to remove nonspecific bead-bound binders. The SELEX process leads to the enrichment of combined oligonucleotides and the diversity of the oligonucleotides is exponentially reduced during the process. The pools from the 1st, 4th, 6th, 8th and 10th selection cycles were used to monitor the SELEX process by quantitative real-time PCR (qPCR). We noticed that the plateau phase of the qPCR curves tended to be stable during the SELEX process (Fig. 2a). The sharply decreased amplification curves in the first few rounds indicate the high diversity of the pool, and the plateau phase amplification curves in the last few rounds indicate the reduced diversity of the pool. If the plateau phase of the qPCR curves from last few rounds tends to be stable, we can thus assume that the diversity of oligonucleotides was exponentially reducing during the SELEX process and that the screening process was successful. Our results are consistent with those of Luo et al. 35 .
After 10 rounds of selection, the aptamers from the 1st, 4th, 6th, 8th and 10th pool were sequenced by deep sequencing, using the Illumina HiSeq2000 platform, to identify specific aptamers for AMDV VP2. The top 100 enriched aptamers provided 40,725 sequences and sequences with > 1% frequency of occurrence are listed in Table 1. The enrichment reads number from the five pools was normalized and the enrichment tendency is also shown in Table 1. The enrichment data indicate that some aptamers, such as ADVa001, ADVa002, ADVa004, ADVa006, ADVa008, ADVa011, ADVa012 and ADVa014, were growing exponentially during the 10 selection rounds. The top 100 enriched aptamers were clustered using MEGA version 6.0 software, and the clustering tree was plotted using FigTree version 1.4.3 software (http://tree.bio.ed.ac.uk/) (Fig. 2b). Four major families were identified, and four aptamers (ADVa001, ADVa011, ADVa012 and ADVa014, drawn as red lines in Fig. 2b) from each clustering family were selected for further characterization.

Characterization of aptamers.
The Kd values of the four selected aptamers (ADVa001, ADVa011, ADVa012 and ADVa014) were measured using an enzyme-linked oligonucleotide assay (ELONA). The resulting mean values were normalized and plotted against the aptamer concentration using Sigma Blot software. The  To further demonstrate the specificity of the selected aptamers, ELONA was performed using the three counter targets, CPV VP2 protein, CAV and BSA. ADVa001, ADVa012 and ADVa014 were found to have good specificity (P < 0.01) for AMDV VP2 protein over CPV VP2 protein, CAV and BSA (Fig. 3), but ADVa011 had poor specificity. ADVa014 thus has good affinity and specificity, and can be used to distinguish between AMDV and CPV using ELONA.
The aptamer sequences, excluding the constant regions, were used to predict secondary structure. The sequences of ADVa001and ADVa011 show that the molecules have a 2-bp stem/4-base loop at the 5′ end.
The secondary structure of ADVa012 shows a molecule with a 2-bp stem/4-base loop at the 5′ end and a 3-bp stem/15-base loop at the 3′ end, connected by two unpaired bases. The secondary structure of ADVa014 shows a molecule with a 2-bp stem/5-base loop at the 5′ end and a 3-bp stem/9-base loop at the 3′ end, also connected by two unpaired bases (Fig. 3).
We predicted the 3D structures of the four aptamers using the DNA model on the RNA structure 6.0 and 3dRNA-V2.0 web servers (http://bioph y.hust.edu.cn/home.html). We found that all four aptamers can form a 5′ end stem-loop structure (red circles). ADVa012 and ADVa014 can additionally form a 3′ end stem-loop structure (blue circles) (Fig. 3). The two stem-loop motifs may be associated with the binding affinity of the aptamers.

Inhibition of virus replication by ADVa012.
After incubation for 1 h at 32 °C, AMDV-HLJ strain (MOI (Multiplicity of Infection) = 1,) and different aptamers (2 μM) were added simultaneously to Crandell-Rees Feline Kidney (CRFK) cells. At 48 h after addition of the ADVa012 aptamer, the amount of secreted virions corresponded to 5.9 × 10 5 viral DNA genome copies/μL culture supernatant, as measured by qPCR. ADVa012 resulted in 32% lower virus concentration in the culture supernatant, compared with the blank control, whereas the virus concentration in the culture supernatant of cells treated with random library aptamers was have no significantly different to the blank control. ADVa014 showed weaker effects (12% inhibition) than ADVa012, ADVa001 and ADVa011 showed no significant inhibitory effect (Fig. 4a). ADVa012 dose-dependently reduced levels of AMDV over the concentration range 0.5-4 μM, and reduced levels of the virus by 47% at a concentration of 4 μM (Fig. 4b).
The half-life of ADVa012 in cell culture was determined by measuring the residual amount of the aptamer between 0 and 60 h after transfection. The aptamer was found to be exponentially degraded, with a half-life (t 1/2 ) of log 10 (2)/0.0095 h≈32 h (Fig. 4c). A good inhibitory effect of ADVa012 on virus production can thus be expected in vitro. A CCK-8 assay showed no apparent cytotoxic effect (P > 0.05) in cells treated with different concentrations of ADVa012 (Fig. 4d).

Identification of domain on VP2 protein recognized by aptamers.
To identify the domain on the VP2 protein that is recognized by aptamers ADVa012 and ADVa014, six overlapping His-fused fragments spanning the VP2 protein were expressed and subjected to ELONA. Six recombinant truncated His-fused proteins were successfully expressed in a prokaryotic expression system (Fig. 5a,b). ELONA was used to determine the truncated fragments to which the aptamers bind.
The P4 fragment is involved in binding to ADVa012 (P < 0.01) and other fragments showed weaker binding, compared with AMDV antigen (P = 0.0025, P < 0.01, Fig. 5c). The P4 fragment is also involved in binding to ADVa014 in comparison with the AMDV antigen (P = 0.0041, P < 0.01), the P3, P5 and P6 fragments showed weaker binding effect (P < 0.01), and the P1 and P2 fragments showed the similar binding effect in comparison  TGGTG TGT GTG TGT GTG TGT GTG CCT ATG CGT GCT ACC GTGA  CCT ATG CGT GCT ACC GTG AA  10.04   ADVa002  AGC AGC ACA GAG GTC AGA TGGTG TGT GTG TGT GTG TGT GCC TAT GCG TGC TAC CGT GAAT  CCT ATG CGT GCT ACC GTG AA  7.16   ADVa003  AGC AGC ACA GAG GTC AGA TGGTG TGT GTG TGT GTG TGT GTG TGC CTA TGC GTG CTA ADVa020  AGC AGC ACA GAG GTC AGA TGTGT TGT GCG GTT TGA GGG TGC CCT ATG CGT GCT ACC GTGA  CCT ATG CGT GCT ACC GTG AA  1.05   ADVa021  AGC AGC ACA GAG GTC AGA TGGGG TTG AGT TGG TCG TGT GGC CTA TGC GTG CTA CCG TGAA  CCT ATG CGT GCT ACC GTG  www.nature.com/scientificreports/ with the AMDV antigen (P > 0.05, Fig. 5c). These data suggest that the binding region for both ADVa012 and ADVa014 may be principally localized within the P4 fragment of the VP2 protein. We speculate, therefore, that the binding region recognized by ADVa012 and ADVa014 may be mainly located between aa residues 397 and 474 of the VP2 protein, and that residues within the P1 and P2 fragments may also contribute to binding to ADVa014.

Discussion
Amdoparvoviruses cause severe diseases in a wide range of animals, resulting in major economic losses for farmers and threatening wild animal populations 2 . In adult mink, AMDV infects macrophages. Viral entry is probably mediated by cellular Fc receptors, which recognize antigen-antibody complexes and enhance viral entry into cells. This phenomenon, known as antibody-dependent enhancement 36 , means that immunosuppressed mink do not develop lesions when infected and is the reason for the failure of many vaccination strategies, although promising results have been obtained with an NS1 DNA vaccine 4 and truncated mutant DNA vaccines 13 . There is, therefore, a pressing need to develop therapeutic agents for the prevention and control of AMD. Efficient monitoring of the progress of SELEX is essential for successful selection of aptamers with good binding affinities 37 . We used qPCR to monitor the SELEX process and noticed that the diversity of oligonucleotides was exponentially reducing during the SELEX process. The sharply decreased qPCR curves indicate high diversity of the pool and the stable curves in the plateau phase indicate low diversity of the pool. We found that the plateau phase of the qPCR curves tended to be stable during the SELEX process and, if the plateau phase from the final ssDNA pool is stable, we can assume that the screening process has been successful. This result is consistent with that of Luo et al. 35 . We also noticed that PCR bias could be avoided by using the cycle threshold (Ct) values from the last pool sample to estimate the number of PCR cycles for the following round. In our experience, the use of C t + 8 as the number of PCR cycles for the following round is suitable for SELEX experiments. www.nature.com/scientificreports/ Although high-throughput sequencing (HTS) could not be performed in real-time, the purified DNA products from the 1st, 4th, 6th, 8th and 10th pools were analyzed by HTS. Bioinformatics analysis of the HTS data provided information about the number of sequences in the analyzed pools, their sequences, and enrichment of sequences during the SELEX process 38 . A comprehensive study of the data for five pools during the SELEX process was in agreement with the results obtained by qPCR. The HTS data on the enrichment tendency from the five pools indicated that many aptamers were growing exponentially during 10 rounds of selection, and some of these aptamers were shown to have high affinity and good specificity. The combination of HTS and bioinformatics analyses thus offered a more efficient tool for ensuring a high success rate in the selection of aptamers 39 .
The selected aptamers (ADVa012 and ADVa014) were found to specifically inhibit infection of CRFK cells by AMDV and to effectively reduce AMDV genome copy numbers in the infection system. At 48 h post-transfection, treatment with 1 μM ADVa012 led to a 32% lower concentration of virus in the culture supernatant and treatment with 4 μM ADVa012 led to a 47% lower concentration of virus. In contrast, after treatment with a random library of aptamers, the concentration of virus in the supernatant was not significantly different from that in the control group. ADV a012 dose-dependently reduced AMDV secretion in the infection system and had a half-life in CRFK cells of ~ 32 h in DMEM supplemented with 5% FBS at 37 °C under 5% CO 2. In this study, we did not use full physiological conditions, just to improve the transfection efficiency using liposome transfection method, so we did not make direct comparison with other study. To validate our results, the experiments were performed three times, and the means of three duplicates were averaged and plotted. A strong and long lasting inhibitory effect of ADVa012 on virus production can thus be expected in vitro. It is also likely that ADVa012 could provide The six recombinant truncated His-fused proteins (P1 to P6) were expressed in a prokaryotic expression system; (c) Binding affinities of ADVa012 and ADVa014 for different domains of the VP2 protein were measured by ELONA. The six recombinant truncated His-fused proteins (P1 to P6) were coated onto the plates and biotin-labeled aptamers were added. The AMDV antigen and BSA were used as positive and negative controls. The P4 fragment is involved in binding to ADVa012 compared with AMDV antigen (P = 0.0025, P < 0.01), and is also involved in binding to ADVa014 in comparison with the AMDV antigen (P = 0.0041, P < 0.01). Results are the averages of three independent experiments performed in triplicate.*P < 0.05; **P < 0.01. www.nature.com/scientificreports/ a treatment for AMDV. In this study, we just to explore an antiviral approach, without much consideration of practical application. After careful accounting, it costs about ¥300 to synthesize 10 nmol ADVa012-80 bp aptamer, so the cost of each treatment is between ¥20 and ¥30. That's really high cost. But, if an industrial-grade synthetic process is used, the cost will reduce largely, and will be advantageous to extend and use the technology. . Solutions were prepared using ultrapure water (resistivity 18.2 Ω), purified using a Milli-Q system (Millipore, Madrid, Spain). CRFK cells were grown in Dulbecco's modified Eagle's medium (HyClone, Logan, UT, USA), supplemented with 10% fetal bovine serum (FBS, HyClone) and 1% penicillin/streptomycin at 37 °C under 5% CO 2 . The AMDV-HLJ (GenBank accession# MF285607) strain was isolated, cultured and preserved in our laboratory 40 . The AMDV antigen (AMDV-G strain) was purchased from Jilin Teyan Biotechnology Co., Ltd. (Changchun, China). The CAVwas maintained in our laboratory. The recombinant AMDV VP2 and CPV VP2 proteins were expressed using the Escherichia coli expression vector pET30a (Qiagen, Hilden, Germany). Manipulation of microorganisms was carried out in a biosafety level II cabinet (BIO II A; Telstar Industrials, Terrassa, Spain).

Expression and purification of recombinant AMDV VP2 protein.
Recombinant AMDV VP2 protein was expressed using the Escherichia coli expression vector pET30a and the His-tagged protein was purified by Ni-chelating affinity chromatography, as described in detail in our previous study 40 . The production of purified recombinant VP2 protein was used to select aptamers in vitro. The mAbs and pAbs used in this study were maintained in our laboratory 41 . Oligonucleotide library and primers. An 80-base random ssDNA library (about 10 14 oligonucleotides) containing 40 bases in random sequence flanked by defined primer-binding sites was synthesized by Comate Bioscience Co., Ltd. (Jilin, China) and purified by High Performance Liquid Chromatography (HPLC). The modified primers P1-F and P2-polyA-S, where F is 6-FAM, polyA is twenty poly (A) at the 5′ end and S is the spacer phosphoramidite 18, were synthesized by Comate Bioscience Co., Ltd. The forward primer P1-F and reverse primer P2-polyA-S were used to synthesize FAM-labeled double-stranded DNA (dsDNA) with a polyA tail in the antisense strand. The FAM-labeled dsDNA was heat-denatured and induced to produce ssDNA by 7 M urea-polyacrylamide gel electrophoresis (PAGE). The primers, P1 and P2, used to monitor the SELEX process by real-time-PCR, were synthesized by Comate Bioscience Co., Ltd. P1 and P2 were used with five forward primers with different barcodes (Lib3S1-01, Lib3S1-04, Lib3S1-06, Lib3S1-08 and Lib3S1-10) to produce the sequencing library. All random libraries and primers are listed in Table 2. The folding procedure for ssDNA was shown as below, 95 °C, 60 s; 95 °C, 90 s, − 1 °C per cycle, 70 cycles.
To test the coupling efficiency, aliquots (2 μL) of protein-coupled beads were suspended in HEPES buffer (20 mM HEPES, 150 mM NaCl, 1 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , pH 7.4, 500 μL) containing 1 × Sypro orange protein gel stain. After heating for 5 min at 85 °C and supersonic treatment for 5 min, the extent of coupling was determined using an FC-500 flow cytometer (Beckman Coulter, Fullerton, CA, USA), with uncoupled blank magnetic beads as the negative control. SELEX protocol and process monitoring. At the beginning of the in vitro selection process, proteincoupled beads (50 μL, ~ 2.5 × 10 8 beads) and ssDNA pool (500 pmoL) were incubated in SELEX binding buffer for 1 h at room temperature. After magnetic adsorption for 5 min and washing three times with binding buffer, the concentrated mixture was diluted to 100 μL using sterile both deionized and distilled water, and heated for 5 min at 100 °C. The collected elution mixture was used as a template for amplification of bound aptamers by PCR (15-25 cycles of 30 s at 95 °C, 30 s at 56 °C, 30 s at 72 °C, followed by 2 min at 72 °C) to obtain the dsDNA pool. After heat denaturation, the FAM-labeled ssDNA band was separated immediately from the 7 M urea-PAGE without nucleic acid dye. Following ethanol precipitation, the ssDNA was recycled and resuspended in sterile deionized distilled water. The selected ssDNA pool was used for the next round of selection. The target amount was reduced in successive cycles (2.5 × 10 8 beads for the first three cycles, 1 × 10 8 beads for the other cycles) to enhance the selection strength. The washing strength was also enhanced gradually by increasing the frequencies of washes (from three to five) to acquire aptamers with high affinity and specificity.
The SELEX process is the enrichment of combined oligonucleotides, and the diversity of oligonucleotides is an exponentially reducing process. Luo's qPCR method 35 was used to monitor the SELEX process. Ct values were measured for the samples and blank and used to evaluate amplification efficiency.
The raw sequence data were then processed in FASTQ format for bioinformatics analysis using a Linux Cen-tOS 6.3 cluster (4 cores, 64 GB). Among the obtained sequences, ssDNA aptamers with > 1% occurrence were clustered and subjected to further assessment. The top 100 enriched aptamers were clustered, and the phylogenetic tree was constructed using the maximum likelihood method in the program MEGA version 6.0. The clustering tree was plotted using FigTree version 1.4.3 software (http://tree.bio.ed.ac.uk/). The top enriched aptamers from each clustering family were selected as candidate aptamers for binding affinity and specificity studies.
Characterization of selected aptamers. The dissociation constants (Kd) of selected aptamers were measured by a modified ELONA 46,47 . After antigen coating and blocking steps, the biotinylated ssDNA aptamers were added at different concentrations (250-4000 nM) in binding buffer and incubated for 1 h at room temperature. HRP-conjugated streptavidin (1:2000; BD Company, San Diego, CA, USA) was then added and incubated for 30 min at 37 °C. After adding substrate and stop buffer, the optical density of 450 nm (OD 450 )was measured using a microplate reader (Bio-Tek Instruments, Inc., Winooski, VT, USA). The resulting absorbance values (OD 450 ) were normalized and plotted against the aptamer concentration using Sigma Plot version 12.0 software. The K d values of the different selected aptamers were estimated using the equation Y = B max X/(K d + X).
To determine the specificity of the selected aptamers, recombinant AMDV VP2, CPV VP2, CAV-1 and BSA were analyzed by ELONA. The mean of absorbance value (OD450) were calculated and analyzed by the nonparametric Student t-tests.

Inhibition of AMDV production by selected aptamers.
To investigate whether the aptamers could specifically inhibit AMDV production, the antiviral effect of four selected aptamers was measured. CRFK cells were infected with AMDV-HLJ strain, and viral genome copy numbers were used to represent viral titer. Briefly, AMDV-HLJ strain (MOI = 1) and different aptamers (2 μM) were added simultaneously to CRFK cells. The cells were incubated for 48 h at 32 °C after aptamer addition. The genomes of secreted virions were recovered by proteinase K digestion, followed by phenol-chloroform extraction. Viral genome copy numbers were measured using qPCR. The primers (AMDV-F and AMDV-R) are listed in Table 2. The antiviral effect of different concentrations (0.5-4 μM) of ADVa012 was measured to investigate dose-dependency. The ssDNA library was used as the negative control. The CRFK cell culture without aptamers or ssDNA library was used as blank control. Results from at least three independent assays for each aptamer were averaged. The mean of the viral genome copy numbers were calculated and analyzed by the one-way ANOVA with Tukey test for multiple comparisons.
Cytotoxicity assays. The cytotoxicity of the aptamers was determined using a CCK-8 assay to measure cell viability. Before aptamer treatment, CRFK cells (1.0 × 10 5 cells/well) were seeded in triplicate in a 96-well www.nature.com/scientificreports/ plates. The cells were then incubated with the aptamer (1-10 μM) for 48 h at 37 °C. Cell viability was determined using the CCK8 assay reagent (Beyotime Biotech, Beijing, China), according to the manufacturer's instructions. After incubation for 3 h at 37 °C, the absorbance at 490 nm was measured using an ELISA plate reader. Three independent experiments were performed, and the result was calculated with respect to the control samples. The mean of absorbance value (OD 490 ) were calculated and analyzed by the one-way ANOVA with Tukey test for multiple comparisons.
Determination of half-life of ADVa012 in cell culture. The construct ADVa012-80 bp, consisting of the constant regions (Table 2), was used to determine the half-life of ADVa012 in cell culture. The aptamer was added to CRFK cells with 5% FBS at 37 °C in 24-well plates with 2 μg (5.9 × 10 13 molecules) of ADVa012 per well using liposome transfection method. After 6 h, the cells were washed five times with pre-warmed PBS (500 μL) to remove non-internalized aptamers. The aptamers were harvested 1, 12, 24, 36, 48 and 60 h post-transfection by lysing the cells with lysis buffer (50 mM Tris-HCl, 100 mM NaCl, 20 mM EDTA, 0.5% v/v Nonidet P-40, pH 7.5, 200 μL per well). The aptamers were recovered by the phenol-chloroform method, and the recovered aptamers were quantified by qPCR, as described in the "SELEX protocol and process monitoring" section. The aptamer copy numbers were used to plot the half-life curve.
Domain on VP2 protein recognized by aptamers. To identify the domain on the VP2 protein that is recognized by ADVa012 and ADVa014, six overlapping His-fused fragments spanning the VP2 protein were expressed and purified by Ni-chelating affinity chromatography. Information for the six recombinant truncated His-fused proteins is provided in Fig. 5a. ELONA was used to determine the truncated fragments to which the aptamers bind. Results are the averages of three independent experiments performed in triplicate. The mean of absorbance value (OD 450 ) were calculated and analyzed by the one-way ANOVA with Tukey test for multiple comparisons.
Statistics. Statistical analysis was carried out using GraphPad Prism version 5.01 (GraphPad Software, Inc. http://www.graph pad.com), and conducted by one-way ANOVA with Tukey test for multiple comparisons. P values < 0.05 were considered to be statistically significant.