**3. Aptamer structures**

The 10G-1 RNA aptamer was selected against NS3 protease domain using a 12–18 nt random‐ ized library and can reduce protease activity by 20% compared with serine protease inhibitors [74]. However, a new SELEX protocol was used to select anti-NS3 aptamers with improved binding and inhibition activities increasing the structural pool complexity by using larger randomized domains (120 nt) and competition against 10G-1. As a result, two new RNA aptamers were selected (G6–16 and G6–19) showing efficient NS3 binding. The G6–16 concentration needed to inhibit 50% of the NS3 activity was 3 µM, and although both aptamers inhibited the protease and helicase activity, they showed lower efficacy compared with known serine protease inhibitors [75]. To further improve aptamer efficacy and inhibit the NS3 RNA binding helicase, a truncated form (∆NS3) only including the protease domain was used as a target, and the random sequence of the RNA pool was reduced to 30 nt to ease the SELEX process. Three highly specific aptamers (G9-I, G9-II and G9-III) were obtained against ∆NS3 containing the conserved sequence GA(A/U)UGGGAC that was present inside an identical loop in all aptamer structures. These aptamers showed *K*<sup>D</sup> values of 11.6 nM, 6.3 nM and 8.9 nm, respectively. The G9 aptamers produced 90% of NS3 protease activity inhibition alone and 70% in the presence of NS4A used to simulate physiological conditions [76]. Structure analyses suggested that interaction of stem I and stem-loop II is essential to G9-I aptamer NS3 binding. To achieve *in vivo* applications, the G9-II aptamer was conjugated with cis-acting genomic human hepatitis delta virus (HDV) ribozymes. The aptamer was inserted into the nonfunctional stem IV region of the HDV ribozyme promoting *in vivo* stable structure that lasted up to 4 days after transfection. The HDV ribozyme–G9-II aptamer (HA) was attached to nuclear export signal CTEM45 (HAC) and ligated in tandem to increase the aptamers dosage in cells. These new constructs showed efficient NS3 protease inhibition *in vivo* and *in vitro* [77]. To also inhibit NS3 helicase activity, a poly U tail (14U) was added to the minimum functional sequence of the G9-I aptamer (∆NEOIII) to mask and inhibit the helicase substrate-binding region [78]. NEOIII-14U displayed dual functions by inhibiting NS3 protease activity *in vivo*

108 Nucleic Acids - From Basic Aspects to Laboratory Tools

and *in vitro* and inhibiting the NS3 unwinding helicase reaction (IC50 1 µM) [79].

petitive mechanism [9].

More RNA aptamers were selected against NS3 helicase domain, including the conserved sequence GGA(U/C)GGAGCC at stem-loop regions. Further deletion and mutagenesis analyses demonstrated that the whole structure of the conserved stem-loop is needed for helicase inhibition. Aptamer #5 presented the best inhibition of helicase *in vitro* activity with an IC50 of 50 nM [77]. Bifunctional aptamers constructed conjugating RNA aptamers ∆NEOIII and G9-II with aptamer #5 through an oligo U spacer. The spacer length was optimized by protease and helicase inhibition assays [80]. The resulting advanced dual-functional (ADD) aptamers (NEO-34-s41 and G925-s50) showed superior inhibitory activities of NS3 [81].

NS5B is an RNA-dependent RNA polymerase that synthesizes the HCV-negative strand RNA using genomic positive RNA strand as a template. NS5B has an essential role in the HCV's life cycle and its variability has been associated with worse disease prognosis [82]. The highly specific B.2 RNA aptamer selected against a truncated NS5B target (NS5B∆C55) presented a conserved sequence that was folded on stem loop structure associated with a tight interaction to NS5B (*K*D = 1.5 ± 0.2 nM). Also, B.2 demonstrated inhibition of NS5B activity by a noncom‐

Nucleic acid aptamers have a diverse range of secondary structures such as stems, loops, symmetric or asymmetric internal loops, bulge, single-base bulges and junctions. Aptamer internal loops and bulges generally present different conformations in solution and adopt defined secondary and tertiary structures on ligand–aptamer complex [92]. This effect was observed on aptamer Sc5-c3 selected against HPV-16 VLPs. Sc5-c3 showed a hairpin structure with an internal loop, where the main loop (ML) presented two different structures in the absence of a target (Table 2). Sc5-c3 transition structure was demonstrated by ribonuclease mapping. Further experiments using Sc5-c3 mutants generated both stable stem and stable loop conformations, demonstrating that the loop structure binds better to the VLPs [58]. Thus, as observed in several aptamers, the binding region remains as a flexible single strand as bulges or loops stabilize conformation arrangements in the presence of a target, producing a very specific binding.

Although bulges and loops are quite common target-binding motifs in aptamer RNAs, they are not the only structures present in aptamer–target complexes. Pseudoknots and G-quad‐ ruplexes have also been reported as functional components of aptamers [93]. For example, some of the aptamers isolated against HIV integrase (93 del and 112 del) presented a G-rich nucleic acid sequence that was stabilized in the presence of K+ as G-tetrad, increasing their inhibitory effect [25]. Later reports showed that 93 del adopts an unusually stable dimeric quadruplex structure [94].

The binding properties of an aptamer are dictated by its sequence and subsequent folding into secondary and tertiary structures. Recently, functional RNA structures were classified as critical, connecting, neutral and forbidden structures regarding their particular roles within a structure [95]. This classification is also applicable to nucleic acid aptamers and is an important clue to design novel and functional variants for viral detection or therapy.
