ANALYSIS OF STRUCTURE-FUNCTION RELATIONSHIPS IN PELARGONIUM LINE PATTERN VIRUS

    A recent survey made in Spain showed that the Pelargonium line pattern virus (PLPV) is the most frequent viral agent in geranium (Pelargonium spp.), with percentages of incidence ranging from 40 to 90 % depending on the examined geographical area. A similar situation is predictable in neighbouring countries and probably worldwide. PLPV produces isometric particles and has a single stranded monopartite RNA genome of positive polarity. Natural infections of the virus seem to be restricted to species of the genus Pelargonium, though diverse plant species can become experimentally infected. Due to the scarce data on PLPV and to the relevance that the virus is acquiring as pathogen, in this thesis we have intended to deep our knowledge on its biological and molecular characteristics.
    The first objective of this work has been to determine the complete nucleotide sequence of PLPV genomic RNA (gRNA). This molecule comprises 3883 nt and, by in silico analysis, we initially identified six open reading frames (ORFs) potentially encoding proteins of 27 (p27), 13 (p13), 87 (p87), 7 (p7), 6 (p6), and 37 kDa (p37). These ORFs are flanked by an unusually short untranslatable region (UTR) at the 5’ side with just 6 nt, and by a 3’ UTR of 246 nt. The arrangement of the ORFs in the gRNA as well as the majority of its potential products were very similar to those involved in replication (p27 and p87), movement (p7) and encapsidation (p37) of members of the genus Carmovirus (family Tombusviridae). However, some unique features distinguished PLPV from carmoviruses. The latter ones produce two subgenomic RNAs (sgRNAs) for the synthesis of the movement and coat proteins, while a Northern blot analysis revealed that PLPV generates a single sgRNA, of approximately 1.7 kb, presumably involved in the translation of the proteins p7, p6 and p37. In addition, the central region of the carmovirus genome contains two small ORFs usually in different reading frames and with canonical initiation codons which code for two movement proteins. However, only the putative p7 of PLPV showed significant homology with movement proteins, while clear homologous could not be found for p6 (and neither for p13). Moreover, the ORF potentially encoding p6 had a non canonical initiation codon though, alternatively, this protein could be expressed through a -1 frameshift (FS) mechanism that would take place immediately upstream of the stop codon of the ORF (p7), giving rise to a protein of 12 kDa (p7-FS or p12). This possibility was supported by the presence, immediately upstream of the mentioned stop codon, of a heptanucleotide whose sequence fits the canonical frameshifting motif. In addition, the formation of a stem-loop downstream of the putative shifty site was predicted, a type of structure which seems to favour frameshifting. Outstandingly, the arrangement of ORFs as well as the anticipated strategies of gene expression from PLPV sgRNA were very similar to those described for the single sgRNA of the Panicum mosaic virus (PMV), the only member of the genus Panicovirus in the family Tombusviridae. 
   Besides PLPV, a group of small isometric viruses whose natural infections also seem restricted to species from the genus Pelargonium had been described. This group included Pelargonium ringspot virus (PelRSV) and Pelargonium chlorotic ring pattern virus (PCRPV). Remarkably, PCRPV potentially encodes five proteins very similar to those involved in replication (p27 and p87), movement (p7 and p9) and encapsidation (p37) in the genus Carmovirus. However, as predicted for PLPV, PCRPV generates a single sgRNA that is likely involved in the translation of 5’-distal ORFs of the gRNA  and could use a -1 FS mechanism for the expression of an internal small ORF. Some data indicate that these characteristics could be shared by PelRSV and Elderberry latent virus, (ELV) and, on the bases of these common characteristics and of sequence similarities, it was proposed to group the four viral agents in a prospective new genus in the family Tombusviridae, named Pelarspovirus.
   The genomic peculiarities of PLPV (and its possible adscription to a new genus) deserved to be studied in depth, thus the next objective was to obtain an infectious clone of the virus in order to carry out a detailed analysis of the structure-function relationships in the pathogen. To this aim, a full length viral cDNA was fused to the promoter of bacteriophage T7 RNA polymerase and subsequently ligated to a high-copy-number vector. The transcripts generated in vitro from the resulting construct were mechanically inoculated in different experimental hosts. These in vitro synthesized PLPV RNAs gave rise to infections indistinguishable to those established by the parental virus in C. quinoa, N. clevelandii and N. benthamiana plants. The availability of this tool has allowed us to develop strategies aimed at obtaining information on the role of the different ORFs in the biological cycle of the virus and on the mechanisms that mediate their expression.
   Through in vitro translation experiments and bioassays of different mutants, we identified the biologically active ORFs of PLPV and obtained data about the step(s) of the infectious cycle in which they are implicated. The results of this study were consistent with the anticipated functions of ORFs (p27) and (p87), involved in replication, and ORFs (p7) and (p37), involved in movement and encapsidation, respectively, while ORFs (p13) and (p6), initially identified in silico, were not translated in vitro nor required in vivo, so they likely lack biological significance. On the other hand, a new ORF located in the central region of the genome was identified, which bears a weak initiation codon (GUG) and encodes a protein of 9.7 kDa involved in the movement of the virus. Remarkably, this PLPV protein shares notable sequence identity (44.4%) with a 9.4 kDa protein potentially encoded by a centrally located ORF of PCRPV. The initiation codon of such ORF would also correspond to a non-AUG triplet, further extending the resemblances among PLPV and PCRPV, both tentative members of the prospective genus Pelarspovirus. 
   Once the function of the proteins encoded by PLPV was determined, we investigated the strategies that this pathogen employs for translation of all its genes. The layout of in vitro translation experiments confirmed that: i) the gRNA directs efficient translation of proteins p27 and p87 (the latter by read-through of the weak stop codon of gene p27), despite lacking a cap structure at the 5’ end, as most (if not all) members of the family Tombusviridae, and having an extremely short leader sequence (6nt), and, ii) the sgRNA acts as mRNA for the production of the three remaining proteins p7, p9.7 and p37, through leaky-scanning processes. These processes seem to be favoured by the suboptimal context of the starting codon of the gene p7, by the weak codon of initiation (GUG) of the gene p9.7, and by the absence of additional AUG codons in any reading frame in the region preceding the initiation codon of the gen p37 in the sgRNA.
  Finally, in order to get insights into the selective pressures that operate on the PLPV genome, the molecular diversity of the virus was assessed using a collection of isolates from different geographical origins, Pelargonium spp., and collecting times. The genomic region examined was 1817 nt long and comprised the complete movement (p7 and p9.7) and coat protein genes as well as flanking segments including the 3´-proximal region of the RdRp gene and the entire 3’ UTR. The results of sequence analyses have allowed to identify some constraints limiting virus heterogeneity such as the conservation of the aminoacidic sequences of specific domains of the proteins and the preservation of certain foldings of RNA regulatory regions. Covarying amino acids within and between proteins have allowed to infer a potential network of specific interactions probably required to maintain the correct conformation of the viral proteins and to drive the viral RNA from the replication process to its dissemination in the plant. Additionally, through serial passages of a PLPV variant in the experimental host C. quinoa, we have proved that this pathogen is able to evolve in a very fast and reproducible fashion when it is transferred to a new plant species. This result highlights the strong selective pressures imposed by the host on the genetic structure of PLPV populations and suggests that the virus is endowed with an usually elevated mutation rate.