Transcriptome and small RNAome profiling uncovers how a recombinant begomovirus evades RDRγ-mediated silencing of viral genes and outcompetes its parental virus in mixed infection.

"plos publication stage vor-update-to-uncorrected-proof publication update 2024-01-25 data availability the srna-seq and mrna-seq data produced and analysed in this study has been deposited in the ncbi sra (short read archive) database as bioprojects prjna1016149 (srr26047678-srr26047725) and prjna1018203 (srr26081500-srr26081547) respectively.; data availability the srna-seq and mrna-seq data produced and analysed in this study has been deposited in the ncbi sra (short read archive) database as bioprojects prjna1016149 (srr26047678-srr26047725) and prjna1018203 (srr26081500-srr26081547) respectively. using the illumina sequencing data ( s1 and s2 datasets) we first measured the collective loads of viral mrnas and viral srnas in reads per million (rpm) of total (plant+viral) mrna and total (plant+viral) srna reads respectively.; mapping of illumina mrna-seq 100 nt paired-end reads on the reference genomes of il and is76 revealed the three pol ii transcription units previously reported for begomoviruses one rightward (virion strand) unit for the v2-v1 mrna and two leftward (complementary strand) units for 3'-coterminal c1-c4 and c2-c3 mrnas (figs2 and 3 and s3 dataset ).; for each condition illumina mrna-seq 100 nt paired-end reads were mapped onto the reference sequences of il and is76 genomes with zero mismatches (see s3 dataset for more details of mapping).; for each condition illumina 100 nt paired-end reads were mapped onto the reference sequences of il and is76 genomes with zero mismatches (see s3 dataset for more details of mapping).; mapping illumina srna reads on the reference genomes of il and is76 revealed that viral srnas are derived from both strands of the entire virus genomes in both s and r plants and at both time-points (figs6 7 8 and s3 and s4 dataset ) and are dominated by the three major size-classes ( fig 9 and s4 dataset ; see below).; for each condition illumina srna-seq reads in the size range from 20 to 25 nts were mapped onto the reference sequences of il and is76 genomes with zero mismatches (see s4 dataset for more details of mapping and maps of each size class of viral srnas).; for each condition illumina srna-seq reads in the size range from 20 to 25 nts were mapped onto the reference sequences of il and is76 genomes with zero mismatches (see s4 dataset for details of read mapping and counting in mixed infection and for maps of each size class of viral srnas).; in support of this hypothesis our illumina sequencing analysis of both il- and is76-infected plants revealed low-abundance long rna reads covering the antisense strands of the rightward and leftward transcription units and both strands of the ir which likely represent remnants of the presumptive readthrough transcripts ( s1 fig and s3 dataset ).; in addition to il genome-derived srnas the highly abundant srnas derived from all regions of the is76 genome (figs7 and 8 ) could further repress il gene expression because most of them share 100% identity with the il genome within the transcription units and the ir outside of the recombination region (see s5 dataset ).; consistent with previous studies of tylcv-il [ - ] and other begomoviruses (e g [ ]) the three major (and functional) size-classes of viral sirnas (21 22 and 24 nt) derived from both strands of the il and is76 genomes were observed in all conditions ( fig 9 and s4 dataset ).; in contrast the hotspots of the less abundant 22 nt class and much more pronouncedly those of the low-abundance 24 nt class were also spread to the ir and the terminator region especially at 30 dpi ( s4 dataset ).; this is despite the hotspots of both 22 nt and 21 nt sirnas became more evenly distributed along both strands of the entire il and is76 genomes ( s4 dataset ).; interestingly the higher abundance of complementary strand-derived srnas observed within the ir and the v2-v1 unit in r plants is largely due to increased proportion of the 24 nt sirnas that exhibit the complementary strand bias in both r and s plants in most conditions ( s4 dataset ).; interestingly a tata-associated composite element (tace) conserved in many genera of geminiviridae which often contains a trap-responsive conserved late element (cle) or its variants with gc-rich sequences [ ] is not affected by is76 recombination whereas an additional cle located at the upstream position of il is mutated in is76 ( s5 dataset ).; in fact a total number of cytosines on both strands of the recombination region is higher in il (s5 dataset).; mapped viral reads were sorted by polarity (forward reverse) and in the case of srnas also by size (from 15 to 34 nts) and 5'-terminal nucleotide identity (5'a 5'u 5'g 5'c) and then counted ( s1 dataset for mrna counts and s2 dataset for srna counts).; single-nucleotide resolution maps of viral mrna and srna reads ( s3 and s4 datasets respectively) were generated using misis-2 [ ].; unlike tylcv-is76' the wild type recombinant can be distinguished from tylcv-il not only by 19 snps and three indels of 2 3 and 9 nts in the recombination region (between the replication origin at position 1 and the recombination breakpoint at position 84 of il or position 76 of is76) but also by other snps scattered along the viral genome (17 snps in the v2-v1 transcription unit 3 snps in the c2-c3 transcription unit 7 snps in the c1-c4 transcription unit 2 snps in the intergenic region upstream of the replication origin and 1 snp in the intergenic region downstream the recombination breakpoint ( s5 dataset ).; in all panels the percentages are for two biological replicates per each condition with the standard error shown with a capped vertical line and the mean value indicated above (pdf) click here for additional data file s1 dataset the illumina 100 nt paired-end reads from each library (two biological replicates per condition: pool 1 and pool 2) were mapped without ( a ) or with ( b ) mismatches to the reference sequences of the solanum lycopersicum genome (nuclear chloroplast and mitochondrion) and the viral genomes (il and is76) sorted by polarity (forward reverse total) and counted.; the counts of plant and viral reads mapped without mismatches were then normalized per million of total reads (rpm) in each library ( c ) (xlsx) click here for additional data file s2 dataset the illumina 15-34 nt reads from each library (two biological replicates per condition: pool 1 and pool 2) were mapped without ( a ) or with ( b ) mismatches to the reference sequences of the solanum lycopersicum genome (nuclear chloroplast and mitochondrion) and the viral genomes (il and is76) sorted by size (15 nt through 34 nt) and polarity (forward reverse total) and then counted.; the counts of plant and viral reads mapped without mismatches were then normalized per million of total reads (rpm) in each library ( c ) and were also sorted by 5'-terminal nucleotide identity (5'a 5'c 5'g 5'u) and then counted in percentage of total ( d ) (xlsx) click here for additional data file s3 dataset for each condition illumina 100 nt paired-end reads of the two biological replicates (1 and 2) were mapped onto the reference sequences of il and is76 using bwa and the resulting bam files were analysed by misis-2 (seguin et al 2016 [ ]) to generate tables of reads mapped to each reference sequence with zero mismatches.; the average percentage at all snps of the viral genome (or its selected region) was applied on all parts of the genome (or its selected region) that contain no snps to estimate the number of reads derived from the entire genome of each virus (or its selected region) or each strand of the viral genome (or its selected region) (xlsx) click here for additional data file s4 dataset for each condition illumina 20-25 nt reads of the two biological replicates were combined and mapped onto the reference sequences of il and is76 using bwa and the resulting bam files were analysed by misis-2 (seguin et al 2016 [ ]) to generate tables of reads mapped to each reference sequence with zero mismatches and sorted by size and polarity.; the average percentage at all snps of the viral genome (or its selected region) was applied on all parts of the genome (or its selected region) that contain no snps to estimate the number of reads derived from the entire genome of each virus (or its selected region) or each strand of the viral genome (or its selected region) (xlsx) click here for additional data file s5 dataset the start and stop codons of viral orfs are coloured in red and underlined the caat and tata-boxes of the promoters coloured in brick red the tata-associated composite element (tace) and conserved late elements (cle) highlighted in green and cyan respectively the iterons highlighted in grey and snps and indels highlighted in yellow (pdf) click here for additional data file. xlsx) click here for additional data file s2 data availability the srna-seq and mrna-seq data produced and analysed in this study has been deposited in the ncbi sra (short read archive) database as bioprojects prjna1016149"

"The authors declare no competing interests."

"The study was supported by ANR through the PRIMA project “Prevention and control of new and invasive geminiviruses infecting vegetables in the Mediterranean” coordinated by M.P. in partnership with M.M.P., C.U. and C.P. (https://anr.fr/Project-ANR-18-PRIM-0003). The PhD salary of M. J. was funded by the ANR project and extended for 4 months by CIRAD (https://www.cirad.fr/). A.F. was supported by a CIRAD South action travel grant. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."