Unmasking the information encoded as structural motifs of viral RNA genomes: a potential antiviral target.

RNA viruses show enormous capacity to evolve and adapt to new cellular and molecular contexts, a consequence of mutations arising from errors made by viral RNA-dependent RNA polymerase during replication. Sequence variation must occur, however, without compromising functions essential for the completion of the viral cycle. RNA viruses are safeguarded in this respect by their genome carrying conserved information that does not code only for proteins but also for the formation of structurally conserved RNA domains that directly perform these critical functions. Functional RNA domains can interact with other regions of the viral genome and/or proteins to direct viral translation, replication and encapsidation. They are therefore potential targets for novel therapeutic strategies. This review summarises our knowledge of the functional RNA domains of human RNA viruses and examines the achievements made in the design of antiviral compounds that interfere with their folding and therefore their function.

atomic force microscopy

cis‐acting replicating element

dengue virus

encephalomyocarditis virus

foot and mouth disease virus

functional RNA domain

genome‐scale ordered RNA structures

initiation factor

human coronavirus

highly variable region

hypervariable region

infectious bronchitis virus

intergenic region

internal ribosome entry site

initiator trans‐acting factor

Japanese encephalitis virus

polyA binding protein

poly‐rC binding protein

pseudoknot

polypyrimidine tract binding protein

systematic evolution of ligands by exponential enrichment

species

untranslatable region

yellow fever virus

INTRODUCTION

The appearance of viral mutants resistant to antiviral compounds is a major public health concern. The continuous generation of such mutants is a characteristic of RNA viruses, many of which, for example, hepatitis C virus (HCV), human immunodeficiency virus (HIV) and DENV (all with a single‐stranded, positive RNA genome), cause serious diseases. The genome of RNA viruses acts both as a template for replication and as an mRNA. During replication, mutations are introduced into the genome via errors in the action of viral RNA‐dependent RNA polymerase; this can generate mutants that may greatly contribute to viral proliferation—although it can also lead to the disappearance of the virus if such changes leave it non‐viable or induce acute pathogenesis in the host, the so‐called error catastrophe (for reviews, see 1, 2). Viruses must, therefore, strive to reach an equilibrium that allows for adaptive fitness without compromising their persistence. To accomplish this, RNA viruses have developed an information storage system that does not code for proteins, but for specific, complex and highly conserved, folded genomic regions that play direct, functional roles in the viral cycle. These viral regions present an adaptive behaviour that has evolved with the selective pressure to achieve high genetic robustness 3, 4, 5. As a result, mutations in viral RNA genomes present minor incidence in the functional elements 4.

RNA molecules show two levels of compaction: their secondary and their tertiary structures. The secondary structure is the result of double‐stranded and single‐stranded regions arrangements, for example, helices closed by loops. The tertiary structure is determined by the relationships established between secondary structure elements. The combination of these two folding levels establishes the final shape of the RNA molecules, the so‐called RNA structurome 6. This conformational view has prompted the search of structured RNA motifs as FRDs, not only in the viral genomes 7 but also in the transcriptome of all living organisms (for a review, see 6).

A main feature of FRDs included in viral genomes is their high genetic robustness, likely derived from their essentiality for the consecution of the infective cycle 3, 4, 5. FRDs are cis‐acting elements typically involved in RNA–RNA and RNA–protein interactions, required for the correct execution and regulation of essential viral processes 8, 9, 10. One of the FRDs that have attracted the most attention is the IRES. IRES regions direct the recruitment of the cellular translation machinery in a cap‐independent manner 11. Similarly, viral replication and maturation are regulated by the direct binding of viral polymerases and related cofactors to FRDs in the 3′ end of the genomic RNA, as well as by direct RNA–RNA interactions between distant regions of the virus genome 12.

The functional importance of structural elements in the viral cycle and their high conservation rate 8, 9, 13, 14, 15 suggest that they might make good therapeutic targets, and the idea of developing novel antiviral tools (mostly RNA based) to attack them has gained currency in recent years. This review briefly summarises what is known about the FRDs of a number of viruses, paying special attention to those infecting humans, and recapitulates the state of the art in the development of therapeutic compounds that might interfere with FRD folding and, therefore, function.

VIRAL IRESs: RNA TAGS FOR EFFICIENT TRANSLATION

Many viral genomes achieve efficient translation initiation by the use of IRESs, which replace IFs by the combined use of several FRDs. These regions present high sequence and structural diversity across different viral groups, which define the mechanisms by which they promote translation. IRESs are therefore divided into four groups 16 according to common IF requirements (Figure 1):

Figure 1

Diagram showing the secondary structure of different types of viral internal ribosome entry sites (IRESs). (a) Dicistroviridae intergenic region (IGR) IRES. (b) Hepatitis C virus (HCV), with the minimum IRES region boxed. (c) Foot and mouth disease virus (FMDV) IRES. (d) HIV‐1 IRES elements. Functional domains are shadowed or indicated as noted. Start translation codons are marked by arrows. PK, pseudoknots

Packed IRESs. This group encompasses the IRES elements located in the IGR of the genome of the Dicistroviridae family (whose members infect insects). The folding of these elements represents a highly conserved, compact structure that, via the so‐called all‐RNA‐based mechanism, mediates the recruitment of the ribosome with no need of additional protein factors 17. The IRES structure can be split into three domains (Figure 1(a)), each defined by a PK motif (PKI–PKIII). This conformation is considered the ‘prefolded’ isoform of the IRES. Domain PKIII mimics the folding of the initiator transfer RNA (tRNA), allowing it to occupy the P and E sites of the 40S ribosomal subunit. This promotes structural changes in the IRES that finally allow the binding of the 60S particle and the efficient initiation of translation from a non‐AUG codon 17, 18, 19, 20, 21, 22. Thus, the conformation of PKIII replaces the need for an initiator tRNA 23.

Extended IRES with compact regions. The prototype of this group, the HCV IRES, folds as an extended region with tightly compact domains. It is mostly contained in the 5′UTR of the RNA genome but also spans a short stretch of the coding sequence (Figure 1(b)) 24, 25. It can be divided into two major domains (II and III) 26 and a short stem‐loop containing the start codon (domain IV) 27. The extended folding of the IRES allows for the proper coaxial alignment of domains II and III on either side of a complex PK structure 28. This allows easy access for the 40S ribosomal particle; subsequent IRES conformational changes directly position the appropriate start codon in the P site 29. Then, binding of eIF3 at the IIIabc junction aids the incorporation of the ternary complex eIF2‐tRNAi Met and the joining of the 60S subunit (for a review, see 30). Interestingly, the requirement for eIF2‐tRNAi Met can be minimised under certain stress conditions 31, providing a selective advantage for efficient HCV propagation even in unfavourable physiological environments. This suggests that an IRES may change its mode of function depending on environmental conditions.

Flexible IRESs. These are mainly represented by picornavirus type II IRESs (FMDV and EMCV among others; Figure 1(c)) 9. Their overall 3D structure reflects a greater degree of flexibility than that shown by IGR or HCV‐like IRESs before ribosomal recruitment 32, 33, 34, 35. Nonetheless, they retain compact self‐folded stem‐loops and a GNRA sequence motif included within a tetraloop that directs the structural organisation of central domain 3 32. Flexible IRESs cannot directly recruit the 40S ribosomal subunit; rather, they first bind eIF4G, eIF4A, eIF4B and eIF3 at domains 4 and 5 36, 37 to finally induce conformational rearrangements that promote the recruitment of the 43S complex. The presence of domains 2 and 3 is crucial because they bind additional ITAFs, such as the PTB, PABP and ITAF45 38, 39, 40, 41. ITAFs are thought to provide chaperone activity but may also help overcome cell restrictions under stress conditions. Hence, the flexible folding provides specific biophysical properties that define a characteristic mode of action.

Inducible IRES. In lentiviruses, the use of IRES elements is restricted to certain environmental conditions, the initiation of protein synthesis normally being accomplished by a cap‐dependent mechanism. Two different IRES elements lie within the HIV genome: one in the 5′UTR, which only operates in the G2/M phase of the cell cycle 42, 43, and one embedded within the coding sequence (Figure 1(d)) 42, 44. Enormous efforts have gone into fathoming the induction mechanism(s) of these elements, but without much success. It is well known that the 5′UTR of HIV‐1 RNA can adopt two different conformations, but a direct translational regulation mediated by conformational exchanges in the genomic RNA has been ruled out 45. Furthermore, it seems likely that unknown ITAFs are differentially loaded by each conformer, providing a sophisticated and opportunistic viral strategy aimed at overcoming environmental difficulties. It is tempting to propose that the induction of the IRES within the coding sequence is further modulated by long‐distance RNA–RNA interactions with the 5′UTR, thus promoting alternative conformations that might regulate its functioning 46.

FUNCTIONAL DOMAINS LOCATED AT THE 3′UTR

FRDs are needed in the 3′ end of viral genomes to promote the efficient binding of the polymerase and the regulation of the elongation step during RNA synthesis 7, 12. They are also important partners in the preservation of viral genomic stability and the control of processes mediated by the 5′ end 47. The latter is mediated by the establishment of a dynamic crosstalk between both termini of the genomic RNA. The sequence and 3D structure of the 3′ cis‐acting domains are quite diverse, even among members of the same taxon (Figure 2), which must entail substantial differences in the mechanism of action if the same function is to be achieved. Such great complexity and variability preclude a detailed discussion of all the 3′ RNA structures present in different RNA viruses; only those most extensively studied are therefore mentioned in the following.

Figure 2

Diagram showing the structural diversity of the 3′ untranslatable regions (UTRs) of different viruses. (a) Several examples of 3′UTRs of enterovirus, (b) aphthoviruses, (c) hepatitis C virus (HCV), (d) Flavivirus sp. and (e) the Coronaviridae family. Proposed pseudoknot (PK) structures are indicated with a line. Structural motifs defining functional RNA elements are noted. Stop codon is marked by an arrow. BSL, bulged stem‐loop; HV, highly variable region; HVR, hypervariable region; N, any nucleotide; R, purine

3′UTRs in the Picornaviridae

Enterovirus sp.:

The genus Enterovirus includes human polioviruses, rhinoviruses and the human enteroviruses of groups B, C and D (which include different human coxsackieviruses). Their genomic RNA ends with the so‐called oriR region, which is composed of a short, highly folded 3′UTR (75–100 nucleotides) capped by a polyA tail 48, 49. The oriR has been shown to play an important role in viral replication via its direct recruitment of viral RNA‐dependent RNA polymerase and host cell factors such as PCBP, the La autoantigen and eEF1A 15. A common architecture at the 3′UTR can be inferred for all the members of the same viral group (Figure 2(a)). Thus, rhinoviruses have a single stem‐loop (named Y), polioviruses contain two stem‐loops (X and Y) and coxsackieviruses have three consecutive stem‐loops (X, Y and Z) 48. The presence of two or more stem‐loops defines a typical PK motif involving the apical loops of domains X and Y 50, 51. This PK may regulate dynamic spatial rearrangements in domains X and Y that fine‐tune viral RNA synthesis 52, 53. Despite sequence diversity, enterovirus 3′UTR represents a good example of how RNA genome structure has diversified yet retains common essential functions.

Aphthoviruses:

FMDV is the prototype of the aphthovirus group. Its genome contains an essential oriR region at the 3′ end consisting of a well‐defined 3′UTR of around 90 nucleotides, plus a genetically encoded polyA tract (Figure 2(b)). The 3′UTR adopts a quasi‐globular conformation defined by two stem‐loops, SL1 and SL2. It was recently shown that, whereas the deletion of SL2 completely interferes with virus viability, mutants defective for SL1 only show reduced replication capacity 54. The oriR region in aphthoviruses is also involved in translational control by the establishment of molecular bridges with the IRES element at the 5′ end of the viral genome 55.

cis‐Acting elements within the 3′ end of genomic Flaviviridae RNA

Hepacivirus sp.:

The 3′UTR of the HCV genome is an evolutionarily conserved structural element of around 200 nucleotides with three different domains (Figure 2(c)): the HV, a polyU/UC tract of variable length and composition and the 3′X tail consisting of three stem‐loops (3′SLI–3′SLIII) 56, 57. Both the polyU/UC tract and the 3′X tail contain elements important in replication and infectivity in cell culture and a chimpanzee model 58, 59, 60, 61, 62, 63, 64. The 3′UTR in hepaciviruses operates by the recruitment of viral proteins, such as RNA‐dependent RNA polymerase and helicase 65, as well as numerous host factors 66, 67, 68, 69, 70. Moreover, the formation of high‐order structures involving RNA motifs in the 3′ end of the coding sequence provides additional regulation of 3′UTR folding and activity 71, 72. Together, the aforementioned elements constitute the complete functional scaffold required to achieve precise RNA replication.

Flavivirus sp.:

This genus includes important human pathogens, such as DENV, YFV and JEV. Their positive RNA genome contains a 3′UTR ranging from 350 to 700 nucleotides in length. Despite this heterogeneity in size, common secondary structure elements are found at the very 3′ end of the genomic RNA, defining the so‐called domain III, which is indispensable for replication 73 (Figure 2(d)). Domain III is composed of a short hairpin followed by a long and stable stem‐loop (3′SL). It has been shown that the formation of a PK motif involving the apical loop of the short hairpin and an internal loop of the 3′SL is critical for the recognition of viral and cellular proteins during the initiation of RNA synthesis 74.

3′ terminal domains in coronaviruses

The genome of the Coronaviridae family encodes for five to seven overlapping, capped and polyadenylated subgenomic mRNAs 75. The 3′UTR contains two highly conserved structural motifs: an essential bulged stem‐loop (BSL) immediately followed by a hairpin (Figure 2(e)) 76, 77. Under certain conditions, this folding may be displaced towards the formation of a PK motif between the base of the first domain and the apical loop of the hairpin 78. This leads to a conformational equilibrium that must be preserved for RNA replication to occur. At the 3′ position of the PK, an HVR is found. It is highly divergent in terms of both primary and secondary structures, with the exception of the 5′‐GGAAGAGG‐3′ motif. This octanucleotide is situated 70–80 nucleotides from the 3′ end of all coronavirus genomes, thus being considered a characteristic signature of these pathogenic agents 75. Such stringent conservation implies a critical yet unknown function in the coronavirus biology 79.

There is a long list of viral and cellular proteins that also bind the 3′UTR of different coronavirus and affect replication 75. These include viral RNA polymerase, non‐structural virus proteins, nuclear factors and proteins involved in the translation process. Thus, once again, RNA–RNA and RNA–protein interactions provide a complex regulatory network that controls viral infection.

CIS‐ACTING RNA ELEMENTS WITHIN THE CODING SEQUENCE

Recent work has shown a great variety of cis‐acting RNA motifs embedded throughout the coding sequence 80, 81, 82, 83, 84, 85, 86. These operate in a coordinated fashion with other functional regions and with viral and host proteins, providing complex control systems for replication.

Some of the best‐characterised CREs are found in the picornavirus genome 87, 88, 89. All of them contain one or more short hairpins responsible for efficient priming during RNA synthesis 90. Two adenine residues placed in an apical loop of one of these hairpins function as a template for the uridylylation of the essential viral protein VPg, involved in picornaviral replication 90. This mechanism has been shown to contribute to the overcoming of the rate‐limiting effects of low nucleotide concentrations 91, thus providing a robust system for the initiation of replication.

A CRE was identified within the 3′ end of the coding sequence of the HCV genome 81. It is a highly conserved structure composed of three stem‐loops (5BSL3.1–5BSL3.3) (Figure 3(a)) 92, which were shown to be required for viral replication 64, 93. The establishment of long‐range RNA–RNA interactions between 5BSL3.2 and the 3′SLII within the 3′UTR 71, 94, 95, 96, and with the SL9110 hairpin motif in the coding sequence 72, 95, 96, is critical for the proper functioning of this region during viral RNA synthesis (Figure 3(a)) 71, 72. Importantly, the existence of an additional, direct, long‐range RNA–RNA interaction involving the subdomain IIId of the IRES and the essential stem‐loop 5BSL3.2 of the CRE region has important consequences for the regulation of the HCV IRES function (Figure 3(a)) 96, 97, 98.

Figure 3

Circularisation mechanisms in different viral genomic RNA molecules. Detailed diagrams of the secondary structure proposed for the 5′ and 3′ ends, as well as for the described cis‐acting replicating elements (CREs), are shown. (a) Hepatitis C virus. The region required for internal ribosome entry site (IRES) activity is noted. The 3′ end of the viral genomic RNA is organised into two structural elements: the CRE region and the 3′ untranslatable region (UTR). (b) Foot and mouth disease virus RNA genome with the VPg viral protein bound to its 5′ end. The IRES region is indicated. Binding sites for eIF4G and polyA binding protein (PABP) are shadowed. The circular form is stabilised by protein–protein (eIF4G and PABP, among others) and RNA–RNA interactions. (c) Flavivirus genome is capped at its 5′ end (indicated by m7G) to ensure viral protein synthesis. The locations of the complementary sequences, cyclisation sequence (CS), upstream AUG region (UAR) and downstream AUG region (DAR), are marked by grey solid lines. The black solid lines indicate long‐distance interactions (direct RNA–RNA and protein–protein contacts). Arrows denote the position of the start and stop translation codons. Pseudoknots (PK) are noted as described in Figure 1

5′–3′ COMMUNICATION IN VIRAL GENOMES

End‐to‐end communication in viral genomes is directly responsible for promoting and regulating different steps of the infectious cycle. Via the establishment of long‐distance contacts, differentially evolved functional regions at both ends of viral genomes can act in a coordinated fashion to generate molecular switch control systems between different steps of the viral cycle. Such crosstalk seems to provide important benefits, such as an increase in the local concentration of essential proteins and cofactors and additional protection against the action of exonucleases. Genome circularisation provides new proof of the great versatility shown by compact viral RNA genomes.

The formation of the circular topology in picornaviruses is dependent on direct RNA contacts between the IRES and the 3′UTR (Figure 3(b)) 55. Furthermore, the RNA binding capacity of different factors, such as eIF4G to the IRES region or PABP to the 3′ polyA tail, ensures the stability of the closed‐loop conformation.

In hepacivirus genomes, circularisation was initially proposed to be mediated by protein factors alone 47, which would simultaneously bind to both ends of the viral genome. More recently, a long‐range contact between the IRES and the CRE was described to induce significant conformational changes in precise domains of the IRES (Figure 3(a)) 97, 99, resulting in the regulation of the IRES function 98. Undoubtedly, these results do not exclude the participation of viral and cellular proteins in the stabilisation of circular topology, nor the performance of important functional roles 100, 101, 102, 103.

In other viruses, such as Flavivirus sp., RNA genome circularisation does not require the presence of proteins but the use of canonical and complementary sequences (cyclisation sequence, upstream AUG region and downstream AUG region) at either end of the viral genome 104, 105, 106 (Figure 3(c)). These motifs achieve the association of the genomic termini by direct base pairing. The closed‐loop topology is essential for viral propagation and could be very advantageous for the regulation of viral translation and RNA synthesis, as well for the control of the switch between different steps of the viral cycle.

VIRAL RNA STRUCTURES AS NOVEL TARGETS FOR THERAPEUTIC AGENTS

Complex viral evolutionary dynamics are a great obstacle in the development of efficient therapeutic tools. The constant appearance of sequence mutants resistant to current treatments has prompted the search for alternative targets. Structurally conserved regions involved in the progression of the viral cycle are excellent candidates. The combination of different compounds targeting multiple viral regions and processes might contribute greatly to reducing the appearance of resistant variants. The following section summarises the main advances recently made in the development of molecules directed against different functional domains of viral RNA genomes.

Nucleic‐acid‐based drugs

Nucleic acids, particularly RNA, are excellent candidates for the development of antiviral agents targeting functional regions of viral genomes. The success of these strategies relies on sequence complementarity and/or structural recognition between the inhibitor and the target. Therefore, highly conserved genomic domains are the most promising regions for targeting new nucleic‐acid‐based antiviral compounds. Importantly, a number of aspects need to be taken into account in pursuing this goal 107, such as the cellular uptake of the inhibitor, target and inhibitor colocalisation, the active inhibitor concentration inside the cell, the intracellular inhibitor stability and the access and binding to the target region. In recent years, great progress has been made in the field of nucleic acid treatment, and some of the aforementioned problems have been now successfully solved. This has led to the design of new generations of antiviral RNA and DNA drugs that are currently being extensively developed 108.

Antisense oligonucleotides:

Antisense oligonucleotides are short nucleic acid molecules whose sequence is complementary to an existing motif in a target RNA. The molecular mechanisms of action of antisense oligonucleotides are diverse (Figure 4(a)) 109. They act by hiding essential sites in the viral genomic RNA, such as those required for protein synthesis, replication and packaging, or if DNA based, by inducing target degradation mediated by RNase H 108.

Figure 4

Nucleic‐acid‐based antiviral compounds. (a) Mechanism of action for antisense oligonucleotides and siRNAs. RNase H mediates the cleavage of the heteroduplex antisense deoxyoligonucleotide : target RNA. For non‐DNA oligomers, inhibition is achieved by directing the oligonucleotides against functional regions of the RNA, such as those involved in the recruitment of the translational or replicative machinery. Small interfering RNAs (siRNAs) bind to the macromolecular complex RNA‐induced silencing complex (RISC), which selects the guide strand that recognises the target sequence in the RNA of interest. This promotes the RNA degradation by specific endonucleases. (b) Diagram of a systematic evolution of ligands by exponential enrichment process to identify RNA aptamers targeting a specific RNA domain. Suppression of transcription and reverse transcription allows for the isolation of active DNA aptamers. RT‐PCR, reverse‐transcription polymerase chain reaction

Antisense oligonucleotides can be administered either directly, by subcutaneous injection or as pharmacological compositions. In this context, great advances in the combinatorial chemistry have allowed for the development of oligonucleotides with chemical substitutions at precise positions 108. These modifications confer important features, such as significant half‐life increase in serum and higher thermal stability of the duplex 110. This strategy was followed by McHutchinson et al. for the design of an anti‐HCV compound 111, but with low success in phase I clinical trials.

A different administration strategy is based on the intracellular production of unmodified antisense oligonucleotides. It requires gene therapy protocols to deliver the construct inside the cell. Its main advantage is that chimeric constructs encompassing different antisense sequences (or even different nucleic‐acid‐based inhibitors) can be produced, thus targeting multiple sites simultaneously. The use of this strategy may greatly contribute in diminishing the appearance of resistant viral variants. This method has been successfully used for the inhibition of the HIV‐1 replication 112, 113, 114.

Small interfering RNAs (siRNAs):

The cellular RNAi pathway can be triggered by synthetic double‐stranded siRNA molecules. These compounds are usually 21 nucleotides long, with a two‐nucleotide 3′ overhang. The siRNAs are loaded into the RNA‐induced silencing complex, where the sense or guide strand is selected to target the complementary sequence within the genomic viral RNA for further degradation (Figure 4(a)) 115.

To date, numerous studies have been conducted using siRNAs as antiviral drugs 115. Importantly, previous advances made for the therapeutic use of antisense oligonucleotides have served as the starting point for the development of this new strategy. Thus, the incorporation of stabilising chemical modifications in the passenger (antisense) strand may greatly increase the potency of the siRNA without toxicity 116.

The application of gene therapy strategies to promote a long‐term decrease of viral loads has been also studied for the siRNA technology. One of the main drawbacks of this alternative is the appearance of escape mutants 117. To overcome this problem, several strategies have been developed from the generation of constructs combining multiple siRNAs targeting different regions of the viral genomes 118.

The great potential of siRNA‐based therapeutics has prompted the development of novel internalisation strategies that should allow for a better‐sustained antiviral response. Important contributions have been made in the field for the HCV infection. For example, the use of nanosomes (lipidic nanoparticles) achieves the encapsulation of multiple siRNAs with a 100% delivery yield in a liver tumour xenotransplant mouse model of HCV 119. Alternatively, multilayered polyelectrolyte films have been also shown to be an efficient way of nucleic acid intracellular delivery, providing dose‐dependent, specific and long‐term inhibition of HCV replication in hepatocyte‐derived cells 120.

Aptamers:

Aptamers are nucleic acids that bind to a specific ligand with high affinity. They are isolated by SELEX, an in vitro selection strategy 121, 122. This process consists of iterative series of the synthesis, binding, positive selection and amplification of a randomised oligonucleotide pool to yield a population enriched in those molecules that can bind to a desired target molecule (Figure 4(b)). This strategy has identified RNA and DNA molecules of potential therapeutic use in a great variety of diseases, including many viral infections 123. In addition, aptamers offer useful information about the 3D architecture of the target molecule. They are therefore a valuable alternative to complementary oligonucleotides (antisense and siRNAs). An aptamer's mode of action is dependent on its folding and mediated by direct loop–loop interactions (Figure 4(b)), which block the function of the RNA target domain, either by interfering with the recruitment of essential factors or by affecting the folding of the surrounding area and thus impeding RNA–RNA interactions, or both. This may contribute to the diminished chance of appearance of resistant viral mutants.

Among the preferred viral targets, the most extensively investigated has been the 5′UTR, including that of the genomic RNA of HIV 124, 125, 126, 127, 128, 129 and HCV 130, 131, 132, 133. Other functional RNA regions have been used as targets for aptamers, such as the 3′UTR 130 and the CRE region of HCV 134. These inhibitors have returned promising results both in cell culture and in vitro and have been used to produce antiviral compounds efficient in the nanomolar range 135, 136. Furthermore, this research has served as a starting point for the development of novel strategies combining two or more inhibitor RNA molecules 137, 138, 139, 140, 141, 142.

The emergence of SELEX methodologies has provided novel therapeutic and diagnostic agents with improved pharmacokinetic and pharmacodynamic properties and low immunogenicity. These compounds are already being investigated by several companies for possible clinical use (reviewed in 143). The conjugation of the 5′ end of the aptamer sequence with stabilising agents such as polyethylene glycol has been shown to be very effective in pharmacological formulations 144. The additional incorporation of ligands for specific cell surface receptors helps in the efficient and precise delivery of agents to target cell types or tissues. Given the great progress being made in this field, aptamers may be in widespread use as antiviral agents in the future.

Antibiotics against RNA functional domains

The use of antibiotics for targeting RNA structural elements was firstly described for the aminoglycoside paromomycin. This compound recognises the aminoacyl A site of the bacterial ribosome by its interaction with an enlarged deep grove defined by two bulging and one unpaired adenine residues 145, 146. This 3D conformation was shown to be strongly similar to that acquired by the HIV‐1 dimerisation initiation site (DIS) kissing–loop complex 147. Further modelling and crystal structure resolution provided a strong basis for the development of modified, high‐affinity antivirals derived from aminoglycosides 148 and a more accurate view of the folding of the HIV DIS region 149. These results will greatly prompt the design of small, improved ligands able to interfere with the folding of the essential DIS domain.

Branched peptides targeting the trans‐activation response element of HIV

Branched peptides are a common therapeutic tool used in the development of synthetic peptide vaccines and drug delivery vehicles 150. Their usefulness as efficient ligands of RNA molecules is associated with the so‐called multivalent targets, that is, FRDs with multiple interacting sites. Additionally, amino acids may carry a radical of specific molecular architecture that increases the target selectivity. These molecules have been isolated from random peptide libraries to generate efficient, non‐toxic and cell‐permeable antiviral compounds that target the HIV trans‐activation response element 151.

Benzimidazole derivatives

The benzimidazole nucleus is the main constituent of many therapeutic drugs, all of them based on the combination of heterocyclic compounds to yield isosters of naturally occurring nucleotides. The chemical composition of these isosters makes benzimidazole derivatives excellent candidates for interaction with natural biopolymers. One of these derivatives, ISIS‐11, has attracted much attention in recent years as a potential anti‐HCV molecule. This compound specifically interacts with the basal portion of the IRES domain II, the so‐called domain IIa 152, 153. During the recruitment of the 40S ribosomal particle, domain II adopts an L shape that directs the apical hairpin loop IIb towards the 40S subunit E site, thus inducing a conformational shift that promotes the positioning of the start codon close to the active site. In the presence of ISIS‐11, domain IIa acquires the shape of an open helix that alters the proper folding of domain II, thus impeding the correct positioning of the viral RNA in the catalytic binding cleft. These findings confirm the important principle of interfering with RNA folding to achieve an inhibitory effect.

CONCLUDING REMARKS

The compactness of RNA viral genomes provides important advantages in terms of propagation efficiency. Such compactness, however, requires the intervention of a supracoding system (i.e. information in the genome beyond that of the nucleotide sequence) to direct the folding required; without such a supracoding system, no such compactness would exist, and essential functions of the genome in transcription, replication and encapsidation associated with this folding would be lost. Structurally conserved functional RNA elements interact with protein factors and other RNA domains to direct and regulate these functions as well as switch between different steps of the viral cycle. Interfering with the functioning of these structural domains offers a potential means of treating viral infections. The great advances made in the field of chemical synthesis and nucleic acid production have prompted novel initiatives involving small molecules that can alter the 3D conformation of viral genomic RNA domains. This might be used to complement classical antiviral therapies, such as the use of neutralising antibodies or interferon treatment. The combination of clinical strategies is currently under extensive investigation with the aim of improving therapeutic responses and long‐term outcomes while minimising toxicity and secondary effects.

CONFLICT OF INTEREST

The authors have no competing interest.