Development of RNA aptamers targeting Ebola virus VP35

Abstract

Viral protein 35 (VP35), encoded by filoviruses, are multifunctional dsRNA binding proteins that play important roles in viral replication, innate immune evasion and pathogenesis. The multifunctional nature of these proteins also presents opportunities to develop countermeasures that target distinct functional regions. However, functional validation and the establishment of therapeutic approaches toward such multifunctional proteins, particularly for non-enzymatic targets, are often challenging. Our previous work on filoviral VP35 proteins defined conserved basic residues located within its C-terminal dsRNA binding interferon (IFN) inhibitory domain (IID) as important for VP35 mediated IFN antagonism and viral polymerase co-factor functions. In the current study, we used a combination of structural and functional data to determine regions of Ebola virus (EBOV) VP35 (eVP35) to target for aptamer selection using SELEX. Select aptamers, representing two distinct classes, were further characterized based on their interaction properties to eVP35 IID. These results revealed that the aptamers bind to distinct regions of eVP35 IID with high affinity (10–50 nM) and specificity. These aptamers can compete with dsRNA for binding to eVP35 and disrupt the eVP35-nucleoprotein (NP) interaction. Consistent with the ability to antagonize eVP35-NP interaction, select aptamers can inhibit the function of the EBOV polymerase complex reconstituted by expression of select viral proteins. Taken together, our results support the identification of two aptamers that bind filoviral VP35 proteins with high affinity and specificity and have the capacity to potentially target filoviral VP35 proteins as a therapeutic target.

Introduction

Negative stranded RNA viruses possess a limited number of open reading frames (ORFs), such that many of the encoded proteins function at multiple stages of the viral replication cycle. The multifunctional nature of these viral proteins potentially provides several opportunities to develop antivirals. However, target validation of non-enzymatic viral proteins is particularly challenging as inhibitors must be identified de novo. Innovative methods that combine structure/function information can provide critical target validation prior to initiation of more effective small molecule-based methods.

Filoviruses are single-stranded, negative sense RNA viruses that cause severe hemorrhagic fever in humans and non-human primates¹ The family filoviridea includes genera the ebolaviruses and marburgviruses and a proposed genus of Cuevavirus² The genus ebolavirus contains five species, which are Ebola virus (EBOV), Sudan virus (SUDV), Taï Forest virus (TAFV), Reston virus (RESTV), and Bundibugyo virus (BDBV), while the genus Marburg virus has two species, Marburg virus (MARV) and Ravn virus (RAVV)³ Among the five species of Ebolavirus, RESTV does not cause disease in humans, although it is pathogenic in non-human primates^{4, 5} However, RESTV was recently isolated from a swine population in the Philippines, whose animal handlers were seropositive suggesting the potential zoonotic nature of filoviruses^{5, 6} Together, these observations highlight the potential public health risk posed by filoviruses and the need to develop innovative countermeasures.

Filoviral viral protein 35 (VP35) is a multifunctional immune antagonist that functions at several stages in the viral replication cycle^{7– 11} VP35 is known to antagonize numerous components in the IFN induction and signaling pathway^{8– 17}, including RIG-I-like receptors (RLRs)^{11, 18, 19}, IFN kinases IKKε and TBK-1²⁰, and RNA-dependent protein kinase (PKR)^13,21. VP35 also functions as an RNAi silencing suppressor^14,22, a co-factor for the viral polymerase²³, and a structural protein that is part of the viral nucleocapsid^24–27 Some, but not all VP35 mediated functions depend on the ability of VP35 to bind double stranded RNA (dsRNA)^{10, 11, 17, 18, 23, 28} Recent structural and functional studies have identified two conserved basic patches, the first basic patch (FBP) and the central basic patch (CBP), within the C-terminal IFN inhibitory domain (IID) of EBOV VP35 (eVP35) (Figure 1A)^{17, 18, 29, 30} The CBP is critical for dsRNA binding and host immune suppression, likely by preventing RNA recognition by RLRs¹⁸ Viruses containing mutations in the CBP region are replication competent, but are attenuated in viral growth. Moreover, prior infection with a guinea pig adapted CBP-mutant EBOV protected guinea pigs against subsequent infection by wildtype EBOV, further supporting the importance of the VP35 CBP for viral pathogenesis²⁸ In contrast, the VP35 FBP is not required for dsRNA binding or inhibiting host immune responses. The filoviral replication complex minimally consists of the viral polymerase L, nucleoprotein (NP), and VP35, with the addition of EBOV viral protein 30 (VP30) being required for viral transcription in EBOV³¹ Previous studies have shown that mutation of residues within the FBP as well as other basic residues that border the CBP (border basic residues) render VP35 defective in its polymerase co-factor function²³ Mutation of FBP residues leads to loss of critical contacts with the NP protein, an interaction required for replication complex formation^{17, 18, 23, 30} In contrast, mutation of the border basic residues does not result in loss of NP binding²³ Altogether, these studies highlight the importance of VP35 proteins in viral-mediated host immune suppression and viral RNA synthesis, two critical functions which contribute to filoviral pathogenesis.

Figure 1. 1G8-14 and 2F11-14 aptamers bind eVP35 IID WT with high affinity.

(A) eVP35 IID contains two basic patches. The first basic patch (FBP) contains residues K222, R225, K251, and K248 (purple). The central basic patch (CBP) contains residues R312, K319, and R322 (cyan). eVP35 IID WT and eVP35 IID CBP3mut were used as targets for in vitro selection of 1G8-14 and 2F11-14 aptamers, respectively. Filter binding assays were used to assess the ability of (B) eVP35 IID WT and (C) eVP35 IID CBP3mut proteins to bind ssRNA, dsRNA, 1G8-14 aptamer, and 2F11-14 aptamer at protein concentrations of 0.1 µM (gray), 1 µM (green), and 10 µM (blue).

Despite increasing knowledge of filoviral pathogenesis, there are no approved treatments that target filoviral infections^{1, 32–37} Currently, there are several promising vaccine options under investigation^{38, 39}; however, there are relatively few non-vaccine options available for treatment, particularly for post-exposure treatments^{38, 39} The multifunctional nature of filoviral VP35 and its relevance to viral pathogenesis makes VP35 a potential target for therapeutic development. Previous efforts using antisense oligonucleotides and small interfering RNAs have been shown to interfere with EBOV replication by targeting VP35^{12, 40–42} These studies coupled with recent results, which show that EBOVs containing mutant VP35 are highly attenuated and guinea pigs infected with these mutant EBOVs are protected against wildtype EBOV infections²⁸, suggest that VP35 is a viable target for antiviral development.

Development of inhibitors that target key components in pathogens often requires detailed knowledge of structure and function or screening of large compound libraries with highly sensitive reporter(s). Nucleic acid aptamers provide an alternative method for inhibitor development as well as target validation^{43, 44} Aptamer identification and characterization does not require detailed knowledge of the target^45–48 However, in situations where detailed structural and functional information are available, aptamer development can be tailored to enhance the impact of aptamer binding by targeting aptamer interactions to functionally important regions. Using such an approach, here we report the development of aptamers against eVP35 IID. Previous structural and functional results were utilized to identify functionally relevant regions of VP35 to target for aptamer development. Aptamers were identified using SELEX (Systematic Evolution of Ligands by EXponential enrichment^{49, 50}) methods by targeting WT eVP35 IID and a mutant form of the protein. Two aptamers, representing distinct binding modes, were selected for further characterization. These studies revealed that one aptamers (1G8–14) preferentially binds EBOV and RESTV VP35 IID proteins, while the other (2F11-14) binds both EBOV and MARV VP35 IID proteins. Using in vitro studies, we demonstrate the utility of our selected aptamers as potential inhibitors, which can disrupt a critical protein-protein interaction in the filoviral replication complex and the activity of the minigenome system in cells. Taken together, our results suggest that aptamers identified here are valuable reagents for basic research and target validation of filoviral VP35 proteins for antiviral development.

MATERIALS AND METHODS

Protein expression and purification

Mutant eVP35 IID proteins were generated by overlap PCR and the plasmids were sequenced to confirm the mutation(s). Recombinant eVP35 IID WT and mutant proteins were expressed as previously described^{17, 51} The integrity of the proteins was assessed by ¹H/¹⁵N-HSQC NMR spectra.

Crystallization and structure determination

Initial crystallization conditions for eVP35 IID CBP3mut (where CBP residues R312, K319, and R322 are mutated to alanine) and FBP4mut (where FBP residues K222, R225, K248, and K251 were mutated to alanine) were identified using commercially available crystallization screens (Hampton Research), and in-house optimized native crystals were grown at 25 °C using the hanging-drop vapor diffusion method.16.4 mg/mL eVP35 IID CBP3mut was diluted in a 1:1 ratio with reservoir solution containing 2.2 M Na/K phosphate (pH 4.5). 24 mg/mL eVP35 IID FBP4mut was diluted in a 1:1 ratio with reservoir solution containing 0.2 M MgCl₂, 0.1 M HEPES (pH 7.5), 30% PEG 400. Crystals were soaked in reservoir solution containing 25% glycerol and vitrified in liquid nitrogen. Diffraction data was collected at Advanced Photon Source (Sector 19) at 100 K (Table S1). One hundred and eighty frames of data were collected with a frame width of 1.0° and detector-to-crystal distance of 300 mm for eVP35 IID FBP4mut. 450 frames of data were collected with a frame width of 0.2° and detector-to-crystal distance of 250 mm for eVP35 IID CBP3mut. Diffraction data were indexed, scaled and merged using HKL-3000⁵². Intensities were converted to structure factors using CCP4⁵³. The structures were solved using molecular replacement, with chain B from eVP35 IID crystal structure (PDB ID 3FKE) as the search model using MOLREP⁵³ The structures were refined using REFMAC5⁵⁴ and PHENIX⁵⁵ Addition of solvent molecules and manual model building was performed using Coot⁵⁶ TLS parameters were refined using the TLMSD server⁵⁷ Final validation was performed using MOLPROBITY server⁵⁸

In vitro selection of aptamers

In vitro SELEX selection of aptamers was carried out using eVP35 IID wildtype (WT) and eVP35 IID CBP3mut as targets^{59, 60}. Oligonucleotide templates and primers were chemically synthesized with standard desalting by Integrated DNA Technology (Coralville, IA). The dsDNA library was generated from a ssDNA library (5`-GCCTGTTGTGAGCCTCCTGTCGAA(N45)TTGAGCGTTTATTCTTGTCTCCC-3`). Using the dsDNA library as a template, RNAs were in vitro transcribed following established protocols. The transcribed RNA was purified by phenol/chloroform extraction followed by ethanol precipitation and separated out on a 7M urea-8% PAGE. Purified RNAs were incubated with either eVP35 IID WT or eVP35 IID CBP3mut at 37°C for 0.5 to 1 hr in 10 mM HEPES (pH 7.0), 150 mM NaCl, 2mM TCEP (tris(2-carboxyethyl)phosphine), and 1 mM MgCl₂ and filtered through nitrocellulose membrane (Millipore). The protein-bound RNA was eluted with 7M urea, precipitated by ethanol, and reverse transcribed using SuperScript™ reverse transcriptase (Invitrogen). The reverse transcribed cDNAs were amplified by polymerase chain reaction (PCR) using oligo-003 (5`-TAATACGACTCACTATAGGGAGACAAGAATAAACGCTCAA-3`) and oligo-004 (5`-GCCTGTTGTGAGCCTCCTGTCGAA-3`). In vitro transcription, RNA-protein binding, partition, and RT-PCR were repeated for 14 rounds and the resulting binders at rounds 10 and 14 were subcloned into the pCR-XL-TOPO vector (Invitrogen) for sequencing. Negative selection was carried out against the nitrocellulose membrane at rounds 3, 6, 9, and 12 to reduce non-specific interactions. The protein concentration for eVP35 IID WT was decreased from 0.4 µM at round 1 to 0.025 µM by round 14 in order to enhance selection stringency. For eVP35 IID CBP3mut, the protein concentration was decreased from 4 µM at round 1 to 0.05 µM by round 14. The concentration of protein used in each selection was adjusted so that the we can obtain about 1% retention of aptamers during each selection. Thus, the initial values of 0.4 µM for selection against eVP35 IID WT and 4 µM for selection against eVP35 IID CBP3mut were determined through this process.

End labeling of dsRNA and aptamers

³²P 5`-end labeling of RNAs was carried out at 37°C for 1.5 hrs in a 20 µl reaction containing 70 mM Tris (pH 7.6 at 25°C), 10 mM MgCl<sub>2</sub>, 5 mM DTT, 1U/µl T4 polynucleotide kinase (New England Biolabs), 1 µM γ-ATP (Perkin Elmer), 5 µM RNA. <sup>32</sup>P 3`-end labeling of RNA was carried out at 16°C for 18 hr in a 30 µl reaction of 50 mM Tris-HCl (pH 7.8 at 25°C), 10 mM MgCl₂, 10 mM DTT, 1 mM ATP, 10% DMSO, 0.67 U/µl T4 RNA ligase 1 (New England Biolabs), 2 µM ³²P -cytidine-3`,5`-bis-phosphate (³²pCp), 5 µM RNA (dsRNA or aptamer)⁶¹ Labeled RNAs were separated on a 7M urea-8% PAGE and eluted by crush soak method.

Filter binding assays for binding and competition assays for eVP35 IID-aptamer interaction

Filter binding assays were carried out as reported previously⁵¹ For dissociation constant (K_D) measurements, the reaction mixtures contained 1–2 nM ³²P-aptamer or ³²P-dsRNA and eVP35 IID protein ranging from 0–100 µM. For single binding point assays, the reaction mixtures contained 10 nM ³²P-aptamer and 0.5 µM eVP35 IID proteins. In the competition assay, the reaction mixtures included 10 nM ³²P-aptamer, 0.5 µM eVP35 IID or 10 nM ³²P-dsRNA, 10 µM eVP35 IID and unlabeled aptamer or dsRNA ranging from 0–20 µM. All reaction mixtures were incubated at 25°C for 15 min. Binding experiments were carried out in buffer containing 10 mM HEPES (pH 7.0), 150 mM NaCl, 2mM TCEP, 1 mM MgCl₂ in 50 µl reactions in a 96-well vacuum filtration apparatus assembled with a nitrocellulose membrane on top of a positively charged nylon membrane. Membranes were washed and air-dried before exposure for 12–16 hrs along with a calibration standard, scanned on a Typhoon 9410 Variable Mode l Imager in phosphor storage mode, and quantified by ImageQuant (GE Healthcare). The ratio of the intensity on nitrocellulose membrane to that of the sum on nitrocellulose membrane plus nylon membrane was expressed as fractional binding. The binding curves were fit to the equation B=B_max*[M] /(K_D+[M]) using ORIGIN 7.0 software (OriginLab), where B is the fraction bound, B_max is the maximum fractional binding, [M] is the protein concentration, and K_D is the dissociation constant. Competition curves were fitted to the equation B/B₀= 1/(1+[M]/IC₅₀)ⁿ), where B is the fraction of ³²P-labeled aptamers bound to nitrocellulose membrane at different concentrations of competitors (aptamer or dsRNA), B₀ is fraction of ³²P-labeled aptamers bound to nitrocellulose membrane without any competitors, [M] is the competitor concentration, IC₅₀ is the concentration where 50% competition occurs, and n is a constant.

Isothermal titration calorimetry

Quantitative analysis of eVP35 IID protein-aptamer interactions were performed on a VP-isothermal titration calorimeter (VP-ITC) (Microcal). Protein samples were dialyzed against 500 mL of buffer containing 10 mM HEPES (pH 7.0), 150 mM NaCl, 1 mM MgCl₂, and 2 mM TCEP for 12 h at 25 °C. Aptamer stocks were diluted >200-fold with dialysis buffer. The syringe contained eVP35 IID proteins at 100 to 150 µM, and the cell contained aptamers (1G8-14, 2F11-14) at 3 to 5 µM. ITC titrations were carried out using a reference power of 5 µcal/sec. The resulting ITC data were processed and fit to a one-site binding model or a two-site binding model to determine n (number of binding sites) and K_D (dissociation constant) using ORIGIN 7.0 software. Both one and two site binding models were considered for each data set, but selection of the model was based on the quality of the fits. For two site binding, both sequential and independent binding modes were also considered. The two-site binding model used in this study assumes two thermodynamically independent binding events.

Pulldown assays

Pulldown assays were performed in buffer containing 10 mM HEPES (pH 7.0), 150 mM NaCl, 1 mM MgCl₂, and 5 mM 2-mercaptoethanol at 25 °C. MBP-His tagged eVP35 IID WT or mutant proteins were immobilized on amylose resin. Resin bound MBP-His tagged eVP35 IID was incubated with purified His-tagged NP protein and subsequently washed. For aptamer competition assays, either 100µM aptamer or 300 µM dsRNA was incubated with MBP-His tagged eVP35 IID prior to incubation with NP. The level of protein used in the assay reflects our efforts to ensure proper detection of the pulldown experiment. Pulldown samples were separated on SDS-PAGE and analyzed by Western blot using mouse anti-His antibody (Santa Cruz biotechnology), followed by horseradish peroxidase (HRP) conjugated goat anti-mouse antibody (Bio-Rad). Membranes were developed using Millipore Immobilon Western Chemiluminescence HRP substrate and recorded on a ChemiDoc (Bio-Rad).

EBOV transcription/replication assays

The plasmids used in the EBOV transcription/replication assays were described previously^{18, 23, 31} NP, VP35, VP30 and L proteins were cloned into pTM1 with only the cDNA sequence (no virus-derived UTRs were present). Aptamers were cloned into a modified pSilencer plasmid⁶² and the corresponding plasmids were sequenced to verify the accuracy of the aptamers.

Quantitation of total aptamer copy number in transfected cells by qRT-PCR

Genome equivalent (GEq) copy numbers (Total copies,) of aptamers in the minigenome assay were measured using primer/probes targeting the aptamers. Briefly, real-time PCR (RT-PCR) was performed using Superscript II RT-PCR kits (Invitrogen). All RT-PCR mixtures contained 5 µl of RNA eluate and master mixes were set up following the manufacturer’s protocols. Standards and test samples were assayed in triplicate using the CFX96 detection system (Bio-Rad) with the following cycle conditions: 50° C for 10 min, 95°C for 5 min, and 40 cycles of 95°C for 10 sec and 59°C for 30 sec. Threshold cycle (CT) values representing aptamers were used to determine Total copies. To create the GEq standard, DNA from aptamers stocks (100 ng) was used and the number of genomes was calculated using Avogadro’s number and the total molecular weight of the plasmid with aptamer.

Aptamer secondary structure prediction

The computational program RNAfold from Vienna package v.2.0⁶³ was used to predict the secondary structures for RNA aptamers 1G8-14 and 2F11-14. Default parameters on the RNAfold server were used except for the Turner Model 2004⁶⁴, which was used as the RNA parameter source. Results were assessed for the minimum free energy (MFE) structure prediction as well as thermodynamic ensemble predictions, including centroid structure⁶⁵ and base-pairing probabilities using dot plot analysis⁶⁶ as implemented in RNAfold. Resulting secondary structures for the MFE and the centroid were visualized with the RNA graphics program VARNA⁶⁷

Circular dichroism (CD)

CD data were collected for 1G8-14 and 2F11-14 at 10 µM in a quartz cuvette with a 1 mm path-length using a Chirascan™ CD spectrometer. CD spectra were acquired from 200 to 320 nm at 5 and 95 °C at an interval of 1 nm and time-per-point of 2 sec. Temperature melting of aptamers was performed by monitoring the ellipticity at 260 nm from 5 to 95 °C with a 2 °C increment at a 1 °C/min gradient. ORIGIN 7.0 software was used to plot CD spectra and to calculate first derivatives from the temperature melting plots.

RESULTS

Mutation of conserved basic residues does not perturb the overall fold of eVP35 IID

The CBP and FBP contain highly conserved basic residues in VP35 IID that are functionally important as mutations of either the CBP or the FBP impair several VP35 mediated functions⁷ Therefore, a series of structural studies were carried out to characterize the impact of these mutations on the integrity of the structure of eVP35 IID. These studies resulted in crystal structures of eVP35 IID CBP3mut (PDB code 4IJE) and FBP4mut (PDB code 4IJF), which were solved to 1.9 and 2.5 Å, respectively using the eVP35 IID WT (PDB code 3FKE) structure as the search model for molecular replacement (Table S1). These structures revealed minimal changes in the overall protein fold when compared to eVP35 IID (Figure S1). The most notable difference between the structures was observed in a short helix (α5) located within the β-sheet subdomain, in which the CBP3mut and FBP4mut structures make one turn of the helix instead of two. Comparison of the eVP35 IID WT, CBP3mut, and FBP4mut structures revealed that eVP35 IID CBP3mut and eVP35 IID FBP4mut structures display RMSD values of 0.68 and 0.46 Å, respectively, which further suggest minimal structural changes upon mutating out the CBP and FBP. Analysis of surface electrostatics, however, showed significant differences among the three structures (Figure S2). As expected, the CBP3mut resulted in a loss of the basic charge at the CBP, but the FBP was relatively unaffected, while the structure of FBP4mut revealed that the basic charge at the FBP was lost with the CBP unaffected.

SELEX identifies RNA aptamers that binding eVP35 IID

In order to generate different classes of aptamers for eVP35 IID, aptamer selection was carried out against two targets, eVP35 IID WT and eVP35 IID CBP3mut proteins using SELEX (Figure 1A). By using eVP35 IID WT and CBP3mut proteins, we biased the aptamer selection to different conserved regions of eVP35 IID. We expected that the dsRNA binding site formed by CBP residues would dominate aptamer selection using eVP35 IID WT as a target due to the highly basic charge and dsRNA binding properties in this region. In contrast, by mutating the CBP (in eVP35 IID CBP3mut), we reduced the basic nature of the CBP region with the intent that other charged regions, preferably the FBP, could compete during aptamer selection. The initial RNA pool contained RNA species of approximately 90 bases, with 45 nucleotides in the variable region flanked on either side by a 5` and a 3` common region. This initial pool bound eVP35 IID WT with a K_D of 2900 ± 200 nM and bound eVP35 IID CBP3mut with a K_D of 36000 ± 8000 nM. After several rounds of SELEX, each aptamer pool displayed enhanced affinity and specificity for eVP35 IID WT and eVP35 IID CBP3mut proteins. Our progress curves show >10-fold affinity enhancement for the final aptamer pool, with an affinity of 300 ± 100 nM at round 14 for eVP35 IID WT (Figure S3A) and 2800 ± 600 nM at round 10 for eVP35 IID CBP3mut (Figure S3B). We also extended our selection against eVP35 IID CBP3mut up to round 14. Although this pool displayed a high degree of non-specific binding to the nitrocellulose membrane, we were able to identify select individual aptamers from rounds 10 and 14. As discussed below, we were able to further characterize these select aptamers for affinity measurements and ranking.

For our initial SELEX studies, we observed significant enrichment of select sequences from a total of 96 clones. Using ClustalW multiple sequence alignments, sequences with variations less than 5% of the total aptamer length (i.e. 4 nucleotides or less that includes the constant/primer regions and the variable region) were grouped together and represented by a single member. We first selected 9 aptamers from the 1-series and 3 aptamers from the 2-series (Figure S3C). Subsequently, we identified representative members from each group based on their affinity for eVP35 IID WT and their ability to inhibit the interaction between VP35 IID and NP (see below). These selected aptamers were subjected to further characterization and ranking based upon affinity measurements with the top binders from each series selected for further analysis. Aptamer 1G8-14 was chosen to represent the 1-series aptamers, which were selected against eVP35 IID WT, and aptamer 2F11-14 was chosen to represent the 2-series aptamers, which were selected against eVP35 IID CBP3mut. In order to further examine RNA structural determinants, we used ViennaRNA structural prediction software to predict the RNA secondary structure of 1G8-14 and 2F11-14. The resulting secondary structure predictions for the MFE and the centroid were visualized with the RNA graphics program VARNA⁶⁷, and base-pairing probabilities for the MFE and ensemble are shown using dot plot analysis (Figure S4)⁵⁴ Based on the structure predictions, the 1G8-14 MFE and centroid structure displayed 15.12 ensemble diversity and the MFE structure represented 5.12% of the ensemble compared to the 2F11-14 MFE and centroid structure with 13.46 ensemble diversity and 4.23% of the ensemble represented by the MFE. However, the free energy of the ensemble for 1G8-14 was −22.63 kcal/mol compared to −12.75 kcal/mol for 2F11-14 suggesting that the 1G8-14 aptamer may adopt a more stable structure. Our CD data show differences in the corresponding spectra providing some experimental evidence that indicate structural differences between 1G8-14 and 2F11-14 aptamers (Figure S5). However, additional studies are required to appropriately characterize the actual structural differences between 1G8-14 and 2F11-14 aptamers. As an initial step in our efforts to define the optimum aptamer, we carried out boundary determination studies and tested several shorter aptamers based on the results of our boundary determination experiments. These results are described in the section entitled “Primer binding region of 1G8-14 is important for high affinity VP35 IID binding” (see below). These studies supported the use of full length aptamers in our present studies as discussed below.

Aptamers 1G8-14 and 2F11-14 bound eVP35 IID WT with affinities higher than dsRNA. At 1µM protein, the fractional binding for both 1G8-14 and 2F11-14 was >0.5, while the fractional binding for dsRNA was <0.3 (Figure 1B). eVP35 IID CBP3mut bound to 2F11-14 with high affinity and to 1G8-14 only at higher protein concentrations (10 µM protein) (Figure 1C). These binding studies provided an initial indication that we had identified aptamers that bind to eVP35 IID differently, with the CBP residues contributing to 1G8-14 aptamer binding. The ability of eVP35 IID WT and eVP35 IID CBP3mut to bind ssRNA and dsRNA was also assessed. As previously shown, eVP35 IID WT binds to dsRNA, but not ssRNA (Figure 1C). eVP35 IID CBP3mut does not bind to dsRNA due to mutations at key dsRNA binding residues, R312, K319 and R322 (Figure 1C)^{16, 28} Previous studies have shown that eVP35 binds dsRNA containing a 5`ppp moiety with higher affinity than blunt end dsRNA<sup>18, 29</sup> Therefore, we tested if a 5`ppp or 5`OH affected aptamer binding to eVP35 IID. The resulting data suggest that the 5` phosphorylation state of the aptamer does not influence eVP35 IID binding (Figure S6). Additionally, this observation supports the notion that the eVP35-aptamer interactions are distinct from those observed between eVP35 IID and dsRNA^{11, 18}.

Aptamers recognize partially overlapping sites in eVP35 IID

We and others have shown that eVP35-dsRNA interactions are mediated through the CBP region of eVP35 IID^{11, 16, 18, 29, 30} Therefore, a series of basic residues in eVP35 IID were mutated to determine the contribution of these residues, if any, to 1G8-14 and 2F11-14 aptamer binding. Representative single point binding data for 1G8-14 and 2F11-14 aptamers revealed distinct binding patterns. 1G8-14 aptamer binding primarily depends on a subset of residues that are also important for dsRNA binding and IFN inhibition. Single mutation of residues F239, R305, K309, R312, R322, and K339 to alanine had little effect on the 1G8-14 aptamer binding (<30% relative to WT) whereas double/triple mutations (K309A/R312A, K319A/R322A, CBP3mut) significantly diminished binding to 1G8-14 (>70% relative to WT) (Figure 2A). Mutations of the FBP residues (eVP35 IID FBP4mut) also impair binding to 1G8-14; however, further experiments revealed that this decrease in 1G8-14 binding to eVP35 IID FBP4mut is largely due to a decrease in the maximum fractional binding and not due to a substantial change in binding affinity (Figure S7). Moreover, these aptamers readily competed with dsRNA in competition assays (Figure S8). These results are different from those observed for dsRNA binding, where select single point mutation of residues in the CBP abolished dsRNA binding¹⁸ Collectively, these results suggest that aptamer binding involves multiple residues within a given eVP35 IID surface, with the CBP residues providing the largest contribution towards 1G8-14 aptamer binding.

Figure 2. Mutational analysis reveals differences in eVP35 IID-aptamer binding sites.

Fractional binding of aptamers (A) 1G8-14 and (B) 2F11-14 to eVP35 IID WT and mutant proteins measured by filter binding assay. Average fractional binding normalized to eVP35 IID WT are from two independent experiments, each with four repeats for individual eVP35 IID variants.

For the 2F11-14 aptamer, we observe a set of residues different from those important for 1G8-14 binding, including R225, K251, and F239, as mutation of these residues to alanine result in some loss of aptamer binding (<40% relative to WT) (Figure 2B). Most significantly, R305A mutant alone can drastically decrease binding to 2F11-14 (>70% relative to WT). In contrast to 1G8-14, eVP35 IID CBP3mut had limited impact on 2F11-14 binding, consistent with the single point data (Figure 1C). Residues R305 and K309 are important for both aptamers, suggesting that the aptamer binding sites are partially overlapping at or near these residues.

Residues in eVP35 IID CBP and FBP contribute differentially to aptamer binding

To further characterize eVP35-aptamer interactions, we measured the binding affinities of aptamer to eVP35 IID WT and mutant proteins using ITC. eVP35 IID WT bound 1G8-14 aptamer with high affinity compared to dsRNA (Figure 3A,^{18, 23}). One-and two-site binding models were considered, ultimately resulting in the ITC data being fit using a two-site binding model, indicating that there are two independent binding modes for eVP35 IID on 1G8-14eVP35. The first binding mode is of high affinity (K_D,1 = 3.7 ± 0.2 nM, n₁ = 1.1 ± 0.1), which we term the “aptamer binding mode”, whereas the second binding mode showed moderate affinity (K_D,2 = 1400± 500 nM, n₂ = 2.7 ± 0.4) (Figure 3A, Table 1). The affinity and the multiple bindings displayed by the second binding mode suggest that this is the dsRNA binding mode. Mutation of residues in the FBP (e.g. K248A and FBP4mut) had little effect on binding (Figure 3B–C). When we test binding of aptamer to eVP35 IID proteins containing mutations of residues important for dsRNA binding (e.g. R312A and F239A), we find that the second binding mode is eliminated and the data can be fit to a one-site binding model (K_D,1 values of 83 ± 10 nM and 85 ± 50 nM, respectively) (Figure 3D–E). Furthermore, mutation of three CBP residues together (i.e R312A/K319A/R322A triple mutant or CBP3mut) or in combination with mutations of the four FBP residues (i.e. FBP4mut/R312A and FBP4mut/F239A) in eVP35 IID results in complete loss of 1G8-14 aptamer binding (Figure 3F–H). For the second binding mode, these data are consistent with previously reported descriptions of the dsRNA-eVP35 interaction¹⁸ Altogether, our data suggest that the first binding mode is the high affinity aptamer-eVP35 IID interaction (aptamer mode) and the second binding mode is eVP35 IID binding to double stranded regions (dsRNA mode) in the aptamer (Table 2).

Figure 3. The CBP is important for high affinity binding of 1G8-14 aptamer to eVP35 IID.

ITC raw data and corresponding binding isotherms for 1G8-14 binding to eVP35 IID (A) WT, (B) K248A, (C) FBP4mut, (D) R312A, (E) F239A, (F) CBP3mut, (G) FBP4mut/R312A, and (H) FBP4mut/F239A.

Table 1.

Measured binding affinities between aptamers and WT or mutant eVP35 I ID proteins by ITC.

	1G8-14				2F11-14
	n1	KD,1(nM)	n2	KD,2(nM)	n1	KD,1(nM)	n2	KD,2(nM)
WT	1.1±0.1	3.7±0.2	2.7±0.4	1400±500	1.2±0.0a	7.1±0.1	5.1±0.1	380±40
K248A	1.1±0.0b	13±1	3.3±0.4	1600±90	1.7±0.0c	36±7	5.3±0.1	830±300
FBP4mut	1.4±0.2	48±30	3.2±0.6	1200±800	n.d.	n.d.	n.d.	n.d.
R312A	1.2±0.1	83±10	n.d.	n.d.	1.4±0.0d	52±6	n.d.	n.d.
F239A	1.1±0.0e	85±50	n.d.	n.d.	1.4±0.0f	100±8	n.d.	n.d.
CBP3mut	n.d.	n.d.	n.d.	n.d.	1.3±0.1	56±7	n.d.	n.d.
lFBP4mut/ R312A	n.d.	n.d.	n.d.	n.d.
CBP3mut/ K248A					1.6±0.2	180±20	n.d.	n.d.

^a= 1.2 ± 0.04

^b= 1.1 ± 0.03

^c= 1.7 ± 0.03

^d= 1.4 ± 0.01

^e= 1.1 ± 0.02

^f= 1.4 ± 0.04

n.d.: no binding

Table 2.

Summary of ITC results for aptamers binding to WT or mutant eVP35 IID proteins.

	dsRNA	1G8-14		2F11-14
		Kd,1	Kd,2	Kd,1	Kd,2
WT	+ + + +	+ + + +	+ + + +	+ + + +	+ + + +
K248A	+ + + +	+ + + +	+ + + +	+ + + +	+ + + +
FBP4mut	n.a.	+ +	+ + + +	−	−
R312A	−	+	−	+ + +	−
F239A	−	+	−	+ +	−
CBP3mut	−	−	−	+ + +	−
FBP4mut+R312A	n.a.	−	−	n.a.	n.a.
FBP4mut+F239A	n.a.	−	−	n.a.	n.a.
CBP3mut+ K248A	n.a.	n.a.	n.a.	+	−

+ + + + = WT

+++ = 5–9 fold decrease in K_D

++ = 10–19 fold decrease in K_D

+ = >20 fold decrease in K_D

− = no binding

n.a: not available

The ITC binding data for 2F11-14 aptamer to eVP35 IID were fit to a two-site binding model, where K_D,1 = 7.1 ± 0.1 nM (n₁ = 1.2 ± 0.0) and K_D,2 = 380 ± 40 nM (n₂ = 5.1 ± 0.1) (Figure 4A, Table 1). A single point mutation in the FBP, K248A, leads to slightly diminished binding with an approximately 5-fold increase in K_D,1 and a 2.5-fold increase in K_D,2 (Figure 4B). Mutation of four residues in the FBP (eVP35 IID FBP4mut) leads to near complete loss of 2F11-14 binding (Figure 4C), emphasizing a critical role for the FBP in 2F11-14 aptamer binding. Mutation of residues in the CBP eliminates the second binding event with K_D,1 = 56 ± 7 nM, K_D,1 = 100 ± 8 nM, and K_D,1 = 52 ± 6 nM for eVP35 IID CBP3mut, F239A, and R312A, respectively (Figure 4D–F). Interestingly, mutation of K248A in combination with the CBP3mut leads to a ~25-fold increase in K_D,1 compared to eVP35 IID WT (Figure 4G), but does not abolish binding altogether. Thus, in contrast to 1G8-14, residues in the FBP are critical for high affinity 2F11-14 binding. Similar to the results observed for 1G8-14 binding, residues in the CBP contribute to the second binding event, presumably through contacts with double stranded regions of the 2F11-14 aptamer. Our results are summarized in Table 2

Figure 4. The FBP is important for high affinity binding of 2F11-14 aptamer to eVP35 IID.

ITC raw data and corresponding binding isotherms for 2F11-14 binding to eVP35 IID (A) WT, (B) K248A, (C) FBP4mut, (D) CBP3mut, (E) F239A, (F) R312A, and (G) CBP3mut/K248A.

In order to better define the binding interface for eVP35 IID-1G8-14 and eVP35 IID/2F11-14, and how these aptamers are oriented with respect to eVP35 IID, we used a filter binding competition assay to see if 1G8-14 could compete with 2F11-14 for eVP35 IID binding and vice versa. As expected, unlabeled 1G8-14 was able to compete off ³²P-labeled 1G8-14. Unlabeled 1G8-14 can also compete off ³²P-labeled 2F11-14 (Figure S9). However, unlabeled 2F11-14 could only compete off ³²P-labeled 2F11-14, but not ³²P-labeled 1G8-14. Since 1G8-14 was selected against WT eVP35 IID, it is conceivable that 1G8-14 aptamer would make contacts with both basic patches. This notion is supported by our ITC data (described above), showing that the CBP is the primary binding site for 1G8-14, but the FBP also contributes to 1G8-14 binding, and mutations to a CBP residue in combination with mutations to the FBP region (FBP4mut + R312A) results in loss of interaction. In contrast, 2F11-14 aptamer, which was selected against a mutant protein with only the FBP present, may predominantly interact with the FBP surface. Altogether, these data are consistent with the notion that distinct classes of high affinity aptamers can be developed using eVP35 WT and mutant eVP35 IID proteins as SELEX targets.

Primer binding region of 1G8-14 is important for high affinity VP35 IID binding

The structure predictions for the 1G8-14 aptamer (Figure S4) suggested that the primer binding regions may be important for secondary structure. To further evaluate the contributions from secondary structure of the 1G8-14 primer regions, we performed binding site boundary measurements to identify the nucleotides of 1G8-14 important for binding to eVP35 IID (Figure S10). The resulting data suggests that the region between nucleotides 36 and 65 is important. However, given the nature of the experiment, it was not clear if this is due to binding of the structural element shown in Figure S10B or if binding was primarily driven by the sequence between nucleotides 36 to 65 (indicated by arrows in Figure S10C). In this assay, we used partially digested aptamers under native conditions and isolated aptamer fragments that bound eVP35 IID binding. RNA bound eVP35 IID proteins were subsequently subjected to denaturing gel electrophoresis for sequence mapping. In order to differentiate between the two potential aptamer fragments that form the high affinity binding, we generated the minimum RNA sequence (nucleotides 36–65) and subjected that sequence to structure predictions and binding studies. In addition, we also tested RNA fragments corresponding to 1–65 and 36–90 nucleotides and the minimal free energy (MFE) structure predictions are shown in Figure S11 These predictions revealed that bases 36–65 forms a stem-loop (shaded region), which is not a sub-structure predicted for full length 1G8-14 structure (compare Figure S11A with Figure S11B). We performed ITC experiments in order to compare the full-length and truncated 1G8-14 constructs in their ability to bind eVP35 IID as well as mutant eVP35 IID proteins. Truncated 1G8-14 36–65 bound to eVP35 IID WT with a two-site binding mode similar to full-length 1G8-14. However, the truncated 1G8-14 36–65 aptamer was unable to bind eVP35 IID R312A (Figure S12; see Figure S12F vs Figure S12C). The sensitivity to the eVP35 IID R312A mutation was also observed previously for dsRNA binding. We also conducted ITC experiments for truncated 1G8-14 1–65 and 1G8-14 36–90 to assess the importance of the 5’ constant region and 3’ constant region, respectively (Figure S13). Truncated 1G8-14 1–65 bound eVP35 IID WT with a two-site binding mode, but similar to truncated 1G8-14 36–65 was unable to bind eVP35 IID R312A. Additionally, truncated 1G8-14 1–65 binding to eVP35 IID WT appear to be cooperative and this observation will require additional experiments in order to fully explore this behavior. However, it is evident that there are significant differences between WT and the 1–65 truncated aptamer. In contrast, truncated 1G8-14 36–90 bound eVP35 IID WT with a two-site binding mode similar to full length aptamer and more importantly, this truncated aptamer bound eVP35 IID R312A with high affinity (Figure S13I). Together these data suggest that the variable region of 1G8-14 is important for binding to eVP35 IID, but additional contributions from the constant regions are required for the optimal aptamer formation. Based on these observations, we used full length aptamers in our studies described here.

Aptamers disrupt eVP35 IID-NP interactions and inhibit EBOV replication/transcription activity

In order to test the utility of the 1G8-14 and 2F11-14 aptamers, we assessed their ability to inhibit interactions between eVP35 IID and NP proteins. This interaction is important for the function of the filoviral replication complex^{23, 31} Specifically, residues in the FBP of eVP35 IID are important for eVP35 IID-NP interactions and minigenome activity (Figure 5A)²³ Mutation of all four residues in the FBP (eVP35 IID FBP4mut) leads to loss of eVP35 IID-NP binding in a pulldown assay. In contrast, mutation of residues in the CBP (eVP35 IID R305A/K309A and eVP35 IID CBP3mut) does not impact eVP35-NP binding. Since the FBP appears to be important for both 1G8-14 and 2F11-14 binding, we tested the ability of these aptamers to compete with NP for binding to eVP35. Both 1G8-14 and 2F11-14 aptamers were able to inhibit eVP35 IID-NP complex formation, while dsRNA and aptamers 2B3-10 and 2D1-10 are unable to disrupt eVP35 IID-NP interaction (Figure 5B). Aptamers 2B3-10 and 2D1-10, selected against eVP35 IID CBP3mut, are of similar length to 1G8-14 and 2F11-14 aptamers, and they display similar high affinity binding to eVP35 IID (Figure 5C). Together, our data demonstrate the ability of 1G8-14 and 2F11-14 aptamers to specifically disrupt the eVP35-NP interaction (Figure 6). We also tested the ability of aptamers 1G8-14 and 2F11-14 to inhibit RNA synthesis in a replication/transcription assay, where the viral replication complex was reconstituted by expression of select viral proteins³¹ Resulting data show that 1G8-14 and 2F11-14 can antagonize the replication/transcription activity in a dose dependent manner (Figure 6). In contrast to these two specific aptamers, additional aptamers with similar length and affinity show little to no inhibition, despite all six aptamers being present at comparable levels as judged by qRT-PCR (Figure S14). Altogether, these results demonstrate that 1G8-14 and 2F11-14 aptamers can potentially function as competitive inhibitors of the filoviral replication complex by disrupting eVP35 IID-NP interactions. However, it is worth noting that the aptamers can interact with NP at high protein concentrations (>100 µM); thus, it is not clear if additional rounds of aptamer optimization with negative selection against NP will be necessary prior to use of these aptamers in virally infected cells.

Figure 5. 1G8-14 and 2F11-14 aptamers disrupt eVP35 IID-NP interaction.

MBP-His tagged WT or mutant eVP35 IIDs were immobilized on amylose resin and incubated with His-tagged NP proteins in the (A) absence or (B) presence of dsRNA or aptamer. (A) Lanes 1, 3, 5, 7 are input (I) samples for MBP-His eVP35 IID WT, CBP3mut, 305A/309A, and FBP4mut, respectively. Lanes 2, 4, 6, and 8 are final bead (FB) samples for MBP-His eVP35 IID WT, CBP3mut, 305A/309A, and FBP4mut, respectively. (B) Lane 1 shows MBP-His tagged eVP35 IID WT bound to the amylose resin. Lane 2 shows the presence of both His-NP and MBP-His eVP35 IID WT, and lanes 3–8 are final bead samples for pulldowns done in the absence (lane 3) or presence of dsRNA (lanes 4) and aptamers 1G8-14 (lane 5), 2B3-10 (lane 6), 2D1-10 (lane7) and 2F11-14 (lane 8).

Figure 6. 1G8-14 and 2F11-14 aptamers inhibit EBOV replication/transcription activity in a dose dependent manner.

A replication/transcription assay was performed in which plasmids encoding the different aptamers (250, 500 ng, left and right, respectively) were co-transfected with the plasmids required for minigenome replication and transcription (plasmids expressing EBOV L, NP, VP30 and VP35 along with plasmid expressing T7 promoter dependent EBOV minigenome which encodes a Renilla luciferase reporter gene). Also co-transfected was a plasmid that expresses firefly luciferase from an RNA polymerase II promoter. Renilla luciferase activity was normalized firefly luciferase activity. Minigenome reporter activation was expressed as relative activity by setting the negative control (without VP35) to a value of 1. The error bars indicate standard deviation of three independent replicates, *p=0.001, **p=0.0005. The western blot shows expression of NP and eVP35.

Aptamers differentially recognize closely related filoviral VP35 IID proteins

Filoviral VP35 proteins share high sequence similarity in the IID region (97.6% similar, 88.9% identical for RESTV aligned with EBOV; 81.7% similar, 22.9% identical for MARV aligned with EBOV) (Figure S15). Therefore, we tested the ability to the 1G8-14 and 2F11-14 aptamers to differentiate between closely related VP35 proteins. The resulting data showed that 1G8-14 aptamer preferentially binds to eVP35 IID proteins (EBOV and RESTV VP35 IID proteins) (Figure 7A). When binding was tested at protein concentrations <1 µM, EBOV and RESTV showed a fractional binding >0.6, while <0.1 fractional binding was observed for MARV. In contrast, 2F11-14 aptamer bound to all three filoviral VP35 IID proteins with >0.5 fractional binding at protein concentrations > 1µM. (Figure 7B). The ability of 1G8-14 to differentiate between eVP35 IID and MARV VP35 IID proteins is likely due to sequence differences between EBOV and MARV VP35 proteins near the CBP. For example, in EBOV VP35 IID positions 319 and 322 are K319 and R322, respectively, while corresponding positions in MARV VP35 IID are T309 and K312 (Figure 2 and Figure S15). Thus, it appears that the relative sensitivity to different filoviral species is entirely dependent on the relative contributions of the different basic patches.

Figure 7. Aptamers differentially bind to filoviral VP35 IID proteins.

eVP35 IID from EBOV, RESTV, and MARV were tested in their ability to bind (A) 1G8-14 and (B) 2F11-14 aptamers at varying protein concentrations shown as fractional binding.

DISCUSSION

In this study, we have identified and characterized two aptamers against eVP35 IID that bind with low nanomolar affinity using SELEX technology. These aptamers were generated by using two targets, eVP35 IID WT and eVP35 IID CBP3mut, with the goal of identifying different classes of aptamers that interact with distinct sets of residues within filoviral VP35 proteins. Target selection was in large part motivated by available structure-function data and our goal was to drive the selection process to different functionally important regions of this multifunctional non-enzymatic protein. Based on the data described above, our selected aptamers display binding affinities that are 20–100 fold higher than dsRNA and the 92 nucleotide long random RNA pool that was used in the initial SELEX, supporting the specific enrichment of aptamers. From the initial groups of aptamers, 1G8-14 and 2F11-14 aptamers were chosen to represent two aptamer classes as they bound eVP35 IID with high affinity, disrupted the eVP35IID-NP complex, and displayed significant differences at the nucleotide level (~40% difference in the aptamer variable region). Biochemical and functional mapping of select aptamers revealed that 1G8-14 and 2F11-14 aptamers bind eVP35 IID primarily through interacting with residues from the CBP and the FBP, respectively. Although it is apparent that additional interactions between eVP35 IID and aptamers are also formed due to the highly basic nature of the eVP35 IID surface³⁰ Our ITC data revealed two independent binding modes: a high affinity aptamer binding mode and a the dsRNA binding mode (Figure 8). Each aptamer has a single high affinity binding site and multiple dsRNA binding sites. These two binding events are energetically independent, at least in the context of eVP35 IID. Interestingly, mutations to CBP residues, or key residues for dsRNA binding resulted in the presumable loss of the dsRNA binding mode. Therefore, at low eVP35 IID concentrations, we observe the high affinity aptamer binding event with a eVP35 IID:aptamer ratio of 1:1. As we increase eVP35 IID concentrations, eVP35 IID is able to make additional interactions with double stranded regions of the aptamers due to its inherent dsRNA binding function, therefore, allowing multiple eVP35 IID proteins to interact with a single aptamer.

Figure 8. 1G8-14 and 2F11-14 aptamers have multiple binding modes.

VP35 aptamers display two binding modes that include a single high affinity binding site and multiple dsRNA binding sites. At low eVP35 IID WTconcentrations or when binding to key dsRNA binding mutant R312A, only the high affinity site is occupied. At higher concentrations of eVP35 IID WT additional binding events occur via the dsRNA binding mode.

This work demonstrates how integrating available structural and functional data to tailor SELEX selection can result in aptamers that target different functional regions and potentially generate multivalent aptamers. 1G8-14 aptamer was selected against eVP35 IID WT and binds primarily to residues in the CBP with additional contacts to the FBP region. Despite the fact that the CBP is important for both 1G8-14 aptamer and dsRNA binding, there are significant differences in how eVP35 IID binds these two ligands. Most notably, a single point mutation within the CBP completely abolishes dsRNA binding^{16, 18, 28} However, single point mutations have only marginal effects on 1G8-14 aptamer binding (Figure 2). In contrast, double and triple mutation of CBP residues significantly affect the ability of 1G8-14 aptamer to bind eVP35 IID, which supports the notion that the aptamers establish multiple contact sites at the binding interface(s). In addition to the CBP, our data shows that mutations within the FBP affect 1G8-14 aptamer binding (increases K_D,1). Mutation of the FBP along with a single mutation in the CBP region abolishes aptamer binding altogether. These results further support the view that 1G8-14 aptamer recognizes several surfaces across eVP35 IID. In contrast, the main site of interaction on eVP35 IID for the 2F11-14 aptamer appears to be near the FBP, with the CBP providing additional contact points. Our results clearly suggest that both 1G8-14 and 2F11-14 aptamers interact with multiple copies of eVP35 IID and therefore there is a high potential for these aptamers to inhibit multiple VP35-mediated functions in vivo.

In this work, we were able to target the CBP or the FBP of eVP35 IID and show that 1G8-14 and 2F11-14 aptamers have distinct binding footprints through biochemical mapping. Moreover, we are able to distinguish that the CBP is the primary binding site for 1G8-14, while 2F11-14 uses the FBP as its primary binding site. As a multifunctional protein, eVP35 IID inhibits IFN responses to viral infection, in part through dsRNA sequestration, and supports RNA viral synthesis through structural bridging of NP and the viral polymerase L. We have shown that aptamers bind eVP35 IID using multiple points of contact, a mode that is distinct from dsRNA binding. Furthermore, aptamers can compete with dsRNA for eVP35 IID binding, indicating that aptamers have the potential to disrupt the IFN inhibitory function of eVP35. Here we show that aptamers disrupt the eVP35-NP interaction and replication/transcription activity, indicating their potential to inhibit eVP35 polymerase cofactor function. It is interesting to note that selected aptamers can potentially inhibit both the eVP35 immune suppressor function and the polymerase cofactor function.

Aptamers are well known for their high affinity and specificity against their targets. The high specificity of aptamers can be used to distinguish among viral subtypes, which is relevant for the development of aptamers against viral proteins. On the other hand, aptamers with broad specificity can potentially serve as pan-antivirals. One of our selected aptamers, 1G8-14, can distinguish between EBOV and MARV VP35 IIDs, binding to EBOV and RESTV VP35 IIDs with K_D’s 20-fold lower (at 1 µM concentration of protein, see Figure 5A) than MARV VP35 IID. The other selected aptamer, 2F11-14, can bind EBOV, RESTV, and MARV VP35 IIDs with comparable affinities. Thus, we have demonstrated that it is feasible to develop both highly specific and pan-antifiloviral aptamers using VP35 as the SELEX target molecule, and our success in selecting both highly specific and pan-antifiloviral aptamers indicates that there are highly conserved regions of VP35 which are recognized by 2F11-14, as well as species unique regions in VP35, which allows differential binding of 1G8-14 to EBOV and MARV VP35 IIDs.

Filoviruses cause high rates of fatalities during outbreaks and are potential agents of bioterrorism. Although filoviral vaccine options are under investigation, no approved vaccines are available at this time. Furthermore, other therapeutics such as small molecule or drug-like methods to counter disease are largely lacking at present. Work here, using available structure-function data, supports the development of VP35 specific aptamers that can facilitate validation of multifunctional VP35 as an antiviral target. Thus, inhibition of VP35 function using aptamers will not only provides a path to develop antivirals, but also provide novel target sites and mechanisms for antiviral development.

ABBREVIATIONS

EBOV

Ebola virus

RESTV

Reston virus

MARV

Marburg virus

LLOV

Lloviu virus

IFN

interferon

IID

IFN inhibitory domain

VP35

viral protein 35

eVP35 IID

EBOV VP35 IID

RLRs

RIG-I-like receptors

CBP

central basic patch

FBP

first basic patch

SELEX

Systematic Evolution of Ligands by EXponential enrichment

NP

nucleoprotein

ITC

isothermal titration calorimetry

dsRNA

double stranded RNA

IKKε

IkappaB kinase epsilon

TBK-1

TANK-binding kinase 1

PKR

RNA-dependent protein kinase

CD

circular dichroism

TCEP

tris(2-carboxyethyl)phosphine

HRP

horseradish peroxidase

MFE

minimum free energy

RMSD

root-mean-square deviation