Small RNA and methylation responses in susceptible and tolerant landraces of cassava infected with South African cassava mosaic virus.

Endogenous small RNAs (sRNAs) associated with gene regulatory mechanisms respond to virus infection, and virus-derived small RNAs (vsRNAs) have been implicated in recovery or symptom remission in some geminivirus-host interactions. Transcriptional gene silencing (TGS) (24 nt vsRNAs) and post transcriptional gene silencing (PTGS) (21-23 nt vsRNAs) have been associated with geminivirus intergenic (IR) and coding regions, respectively. In this Illumina deep sequencing study, we compared for the first time, the small RNA response to South African cassava mosaic virus (SACMV) of cassava landrace TME3 which shows a recovery and tolerant phenotype, and T200, a highly susceptible landrace. Interestingly, different patterns in the percentage of SACMV-induced normalized total endogenous sRNA reads were observed between T200 and TME3. Notably in virus-infected T200 there was an increase in 21 nt sRNAs during the early pre-symptomatic response (12dpi) compared to mock, while in TME3, the 22 nt sRNA size class was predominant at 32days post infection with SACMV. While vsRNAs of 21-24 nt size classes mapped to the entire SACMV DNA-A and DNA-B genome components in T200 and TME3, vsRNA population counts were lower at 32 (symptomatic stage) and 67 dpi (recovery stage) in tolerant TME3 compared with T200 (non-recovery). It is suggested that the high accumulation of primary vsRNAs, which correlated with high virus titers and severe symptoms in susceptible T200, may be due to failure to target SACMV-derived mRNA. Likewise, in contrast, in TME3 low vsRNA counts may represent efficient PTGS of viral mRNA, leading to a depletion/sequestration of vsRNA populations, supporting a role for PTGS in tolerance/recovery in TME3. Notably, in TME3 at recovery (67 dpi) the percentage (expressed as a percentage of total vsRNA counts) of redundant and non-redundant (unique) 24 nt vsRNAs increased dramatically. Since methylation of the SACMV genome was not detected by bisulfite sequencing, and vsRNA counts targeting the intergenic region (where the promoters reside) were very low in both the tolerant or susceptible landraces, we could not provide conclusive evidence that 24 nt vsRNA-mediated RNA directed genome methylation plays a central role in disease phenotype in these landraces, notwithstanding recognition for a possible role in histone modification in TME3. This work represents an important step toward understanding variable roles of sRNAs in different cassava genotype-geminivirus interactions.