Occurrence of Hop Latent Viroid in Cannabis sativa with Symptoms of Cannabis Stunting Disease in California

In February 2018 we sampled three symptomatic and three asymptomatic Cannabis plants suspected to host a viral agent from a farm in Santa Barbara County, CA. The symptoms included brittle stems, an outwardly horizontal plant structure, and reduced flower mass and trichomes. RNA from each flower sample was extracted using a QIAmp Viral Mini Kit (Qiagen, Hilden, Germany). RNA was purified, concentrated, and ribo-depleted before library construction with a NEBNext Ultra II Directional RNA kit (NEB, Ipswich, MA). Individual libraries were tagged with unique adapters and pooled for sequencing with a NextSeq High Output 300 cycle kit (Illumina, San Diego, CA). Sequence data were assembled with Trinity (Grabherr et al. 2011), and Blobtools (Laetsch and Blaxter 2017) was used to search for viral sequences. Possible viral transcripts were filtered for at least 10× coverage, 95% BLAST identity, and alignment length greater than 80% of query and subject sequences. One hundred twenty-five transcripts showed significant similarity to hop latent viroid (HLVd) sequence EF613183. We used STAR (Dobin et al. 2013) to align reads to the transcripts and found that 2.4% of read pairs from symptomatic libraries (of a total of 153,001,495 read pairs) and 0.0001% of read pairs from asymptomatic libraries (of a total of 159,608,791 read pairs) mapped to the putative HLVd sequences. We also identified a transcript with high homology to C. sativa mitovirus 1 (BK010428). Reads from all libraries mapped to this transcript in similar numbers; thus, the mitovirus is unlikely to be associated with the disease symptoms. To confirm our findings, we used two HLVd-specific primer pairs in reverse transcription PCR (RT-PCR). The Zeigler et al. (2014) primer pair produced the expected HLVd amplicon size from only the symptomatic plants, whereas the primer pair of Eastwell and Nelson (2007) produced the expected HLVd amplification pattern from all samples collected. The amplicons were Sanger sequenced and produced at least 200 bp of data with 95% or more pairwise identity to HLVd sequences in GenBank. One sample, CV1 (MK791751), produced 98% pairwise identity across the whole 256-bp HLVd genome as published by Puchta et al. (AX07397). In February 2019, we sampled seven symptomatic and three asymptomatic plants from another farm in Alameda County, CA. RT-PCR revealed the presence of HLVd in all symptomatic plants and one asymptomatic plant. Amplicons from four of these samples were Sanger sequenced and confirmed to be HLVd (MK791747 to MK791750). Analysis of the sequences obtained from the two locations revealed two single-nucleotide polymorphisms between them at positions 190 and 225, which include the centrally conserved region (Puchta et al. 1988). The consistent detection of HLVd alone in all symptomatic plants from both locations and its occurrence in only a few asymptomatic plants point to an association of the disease with the stunted Cannabis. HLVd can persist in hops without symptoms, and Cannabis cultivators should take measures to minimize mechanical transmission while keeping in mind that HLVd may also be seed transmissible (Pethybridge et al. 2008). Considering the detrimental effects of HLVd to the commercially valuable secondary metabolites of the species, a thorough investigation of susceptibility as well as range and expression of Cannabis stunting disease should be conducted at this important juncture.