Cotton leafroll dwarf virus (genus Polerovirus, family Solemoviridae) has been commonly reported affecting cotton plants (Gossypium spp., family Malvaceae) and several weed species (Ramos-Sobrinho et al., 2021; Sedhain et al., 2021). During a recent survey, cacao (Theobroma cacao L.) trees exhibiting virus-like symptoms such as leaf mosaic, vein clearing, and yellow spot were observed in the south part of the state of Bahia, northeastern Brazil, in 2022. Leaf samples were randomly collected from symptomatic cacao plants (n=30) growing in an affected area of approximately 30 ha. Total RNA obtained from pooled cacao samples were subjected to Illumina HiSeq 2500 sequencing as previously described (Keith et al., 2021), and partial sequences of cotton leafroll dwarf virus (CLRDV), and other virus-specific sequence contigs, were de novo assembled according to Ramos-Sobrinho et al. (2021). To further investigate the presence of CLRDV in cacao leaves, total RNA was individually extracted using a modified silica protocol (Rott and Jelkmann, 2001) and used as template for cDNA synthesis with random hexamers using the SuperScript™ IV First-Strand Synthesis System (Invitrogen, CA, USA) following the manufacturer´s protocol. Detection of CLRDV was carried out by reverse transcription-polymerase chain reaction (RT-PCR) with the primers PL4F and o3-R, which amplify the open reading frame 3 (ORF3) encoding the capsid protein (Corrêa et al., 2005). Expected size amplicons (~0.6 kb) were observed from 16 out of 30 symptomatic plants, indicating ~53% of the cacao trees were infected by CLRDV. Considering 14 symptomatic plants tested negative for CLRDV, the symptoms observed here could also be caused by other viral groups or abiotic stress. To confirm the detection of CLRDV, the first half (~3.5kb) of the viral genome was amplified from two representative cacao samples using the primers P20F and P22R (Avelar et al., 2020). The RT-PCR products were gel-purified using the Wizard® SV Gel and PCR Clean-Up System (Promega, WI, USA) and Sanger sequenced. The RNA Illumina sequencing from pooled cacao samples (n=30) yielded 34,610,572 million trimmed reads. Two contigs of 868 and 839 nucleotides (nt) in length, and sharing high nt identity with CLRDV isolates, were assembled from 6,903 and 10,271 reads, at a coverage depth of 795 and 1,224x, respectively. Together, these contigs represent ~29% of the complete viral genome and included part of the 5´-untraslated region, ORF0 and the second half of ORF1-ORF2. Additional CLRDV-like contigs were observed across the viral genome, but they were not considered for further analyses due to the poor sequence quality. The Illumina- and Sanger-derived ORF0 and partial ORF1-ORF2 sequences shared >97% nt identity, suggesting they were congruent. Pairwise sequence comparisons for ORF0, encoding the gene silencing suppressor P0, indicated the cacao-associated isolates shared 99.7 and 99.2% nt and amino acid (aa) identity one with another, respectively. The ORF0 nt sequences showed 91.9-93.8 and 90.7-93.6% identity, while the aa sequences shared 85.8-88.5 and 86.2-90.0% similarity, with CLRDV isolates previously reported in South America and the USA, respectively. Finally, the ~3.5kb nt sequences of cacao-infecting CLRDV isolates shared 92.9-95.8% identity with CLRDV genomes deposited in NCBI-GenBank. The Bayesian phylogenetic tree reconstructed based on ORF0 nt sequences showed the new sequences were more closely related to CLRDV-atypical isolates (GenBank accession nos. KF359946, KF359947, KF906260, and KF906261). Together, these results suggest the new ORF0 sequences belong to CLRDV and were deposited in GenBank under accession nos. ON954058-ON954059. To our knowledge, this is the first report of CLRDV infecting cacao plants, expanding the range of malvaceous hosts of this polerovirus. CLRDV is largely known for causing yield losses in cotton crops, but additional studies are needed to determine if CLRDV infection is detrimental to cacao production.
