The higher order of organization through dimerization, i.e., the formation of bonded dimers (double-hammerhead structures), is a particular feature of ASBVd to do more under tight evolutionary constraints for such small genomes that should preserve functional HHR [74
]. The simultaneous presence of viroids or viroid-like RNAs with different catalytic profiles: from inactive or bad catalysts as single-hammerhead structures to intermediate and good catalysts can be the result of the adaptation process to specific hosts and of the survival of properties which originate from the RNA World. On the other hand, the host range of viroids that is restricted to plants has often been presented as an argument against the hypothesis to consider viroids as relics from the RNA World. However, they are also functional in eukaryotic cells [71
]. So, we cannot rule out that they have not been found yet in other organisms until they have been searched using the latest technologies applied in RNomics.
ASBVd has adapted to different hosts: it normally replicates in the chloroplast of avocado, but it can also replicate in the cyanobacterium Anabaena
. We have further showed that ASBVd replicates and persists in S. cerevisiae
, a non-photosynthetic eukaryote. It is legitimate to speculate on extant viroids as descendants of “free-living” proviroids that invaded ancient cyanobacteria, which would later become endosymbionts, evolving in chloroplasts by usurping the biochemistry of their hosts. Finally, in a “viroids as living fossils” scenario [36
], we must envisage that polymerase and ligase activities might have been lost by the viroid ancestors since it has been demonstrated that these two activities are possible in vitro. They were indeed discovered in artificially evolved RNAs by SELEX, for instance, in the RNA-polymerase ribozyme (PDB ID: 3IVK) [76
] and RNA ligase ribozyme (PDB ID:2OIU) [77
]. Viroid-like ancestors from the RNA World would need to carry three catalytic activities: RNA polymerase, RNA ligase, and RNase which is the sole remaining ribozyme activity in the Avsunviroidae family [78
Assuming that viroid-like RNAs and viroids have lost two of their initial catalytic activities, we may propose an evolutionary path from the RNA World to the DNA World (Figure 7
). The viroids of the Avsunviroidae family still carry an RNAse activity which is necessary for their life cycle. During the transition, the viroids could override the constraints imposed initially by the “biocatalyst/template” paradox. In the RNA World, the proviroids with a “bad catalyst” profile are good templates for replication and preserve the necessary RNase activity through the dimerization into double-hammerhead structures which are catalytically active. In the RNA World, the dimerization would also confer a selective advantage against RNA degradation [61
]. Then, slight changes in the sequence of the HHR motif could lead to some intermediate configuration where the catalytic activity is improved in a strand-specific way: the HHR fold is stabilized only on one strand (e.g., by lengthening the stem III), making the viroid more efficient as a catalyst. The opposite strand would keep being a good template with a catalytic activity relying on the formation of bonded dimers. Additional concerted changes in the HHR sequence could result in more stable HHR folds on both strands carrying a highly efficient catalyst. In a cellular environment, the viroids can evolve by a selection for better catalysts, where the ribozyme-dependent RNase activity is optimal while the template can become suboptimal. By using host enzymes for the other catalytic activities, there is no more selective pressure for good templates (Figure 7
In this hypothetical evolutionary path, ASBVd would correspond to some half-way state where the catalytic efficiency has been partially optimized. The reason it did not evolve towards a fully optimal catalyst found in other viroids may seem puzzling. However, it might be explained eventually by different sources of evolutionary constraints creating antagonistic epistases [75
] in order to escape from the plant’s immune response, and to preserve at the same time the HHR folds on both strands and the associated long-distance interactions necessary for catalytic efficiency (Figure 4
). Only a few nucleotide substitutions and insertions are sufficient to alter the virulence/latency of ASBVd variants [79
]. One may invoke that ASBVd has an extremely narrow host range and thus has co-evolved under host-specific evolutionary constraints. Two other members of the Avsunviroidae family, ChCMVd and PLMVd, also have a narrow host range but they do harbor some stable HHR folds on both strands with high catalytic efficiency. However, these two viroids belong to genus Pelamoviroid
and adopt complex multi-branched secondary structures which are very different from the simple rod-like structure of ASBVd [80
]. Thus, the peculiar secondary structure of ASBVd (only member of the genus Avsunviroid) might be tightly bound to the evolutionary trajectory of this viroid as a special case of “intermediate” catalyst. ASBVd would be the only example of true viroid with a sub-optimal catalyst, since all the other listed “bad catalysts” correspond to viroid-like RNAs. So, it might correspond to a trace for direct inheritance from the RNA World.
The existence of another mode of dimerization in viroids through the self-assembly of monomeric RNAs is also a particularity of ASBVd. As mentioned previously, ASBVd exists in two non-symmetrical polarities (+) and (−) with different structures and catalytic activities. A joined thermodynamic analysis of structural and catalytic data indicates that the rate-determining step corresponds to a dimer/monomer transition. Models suggest that the intermolecular contacts stabilizing the dimer (between HI and HII domains) in HHR (−) compete with the intramolecular ones, stabilizing the active conformation of the full-length HHR required for efficient self-cleavage (Figure 4
). Similar competing intra- and inter-molecular contacts are proposed in ASBVd (−), though with a remoter region from an extension of the HI domain. In vivo, ASBVd (+) can be found in different concatenated multimeric forms up to octamers, while ASBVd (−) is just present as a monomer or dimer with a more efficient cleavage activity. Each polarity could play a distinct role during the viroid life cycle. Lowering the temperature-dependent catalytic activity of ASBVd (−) might play a regulatory role associated with the day/night cycle. Alternatively, it could be a mechanism aimed at synchronizing the transcription of both (+) and (−) strands. The dimerization through weak non-bonded interactions could attenuate the virulence and allow the viroid to co-evolve within its hosts by switching between phases of virulence or latency (Figure 7
). Such a mechanism is already in action with the apparition of ASBVd variants responsible for a transition to a milder form of infection [79
]. This ability to form non-bonded dimers could result from some recent event since it is only reported for ASBVd to date.
The Avocado Sun Blotch Viroid (ASBVd) is unique since it combines “modern” and “ancient” features that may be attributed to either the RNA or DNA World. Proviroids would have been good templates for replication but poor catalysts (biocatalyst/template paradox). As a sub-optimal catalyst, ASBVd is reminiscent of those proviroids whose dimerization through double-hammerhead structures would have protected them from degradation in a clay mineral environment. ASBVd still has this ability to form such dimers, a trait which is also found in many mobile elements or viroid-like RNAs harboring a HHR motif but absent from the other known viroids of the Avsunviroidae family. On the other hand, the formation of non-bonded dimers, which is unique to ASBVd, tends to decrease the catalytic activity; it could be a recent adaptation and the sign of co-evolution in the DNA World.