Pandan (Pandanus amaryllifolius Roxb) (family Pandanaceae) is a tropical monocot plant that is widely cultivated in Southeast Asia. Pandan leaves are used for flavoring and as a natural green colorant for several foods. A young pandan plant has a grass-like appearance, but it grows into a palm-like tree with abundant aerial roots when it becomes mature. Pandan is vegetatively propagated using suckers and cuttings [1, 2]. To date, no reports are available regarding viral diseases occurring in P. amaryllifolius, but cucumber mosaic virus has been found infecting the related species P. tectorius in China [3].

In November 2019, virus-like symptoms, including mosaic and chlorosis, were observed on several offshoots of fragrant pandan grass at Mounts Botanical Garden in West Palm Beach, FL (Fig. 1A). Symptomatic leaves from these plants were collected in April 2021 and prepared for RNA sequencing (RNA-Seq) for further investigation.

Fig. 1
figure 1

(A) Pandan leaves exhibiting symptoms of mosaic and chlorosis. (B) Diagram depicting the genome organization of pandanus mosaic associated virus. ORFs are represented by grey boxes, and putative conserved domains are represented by yellow boxes.

Full size image

Total RNA was extracted using an RNeasy Plant Mini Kit (QIAGEN) according to the manufacturer’s instructions. RNA was treated with DNase to remove host genomic DNA, and ribosomal RNA was depleted. Complementary DNA was generated using a Truseq RNA Library Prep Kit and sequenced using a NextSeq 500 Illumina platform (75-bp single-end reads). A total of ~27 million raw single-ended reads were obtained, and reads were processed using CLC Genomics Workbench 11 (QIAGEN). De novo assembly generated a total of 61,968 contigs, and viral contigs were identified by BLAST analysis. Three viral contigs sharing low amino acid (aa) sequence identity (<60%) with members of the genus Badnavirus were identified. Badnavirus contigs comprising a partial genome sequence were investigated further. Additionally, three contigs showed ≥98% aa sequence identity to homologous proteins of orchid fleck virus (OFV). Subsequent analysis generated a nearly complete genomic OFV sequence, including the complete open reading frames of OFV RNA1 and RNA2. The RNA1 and RNA2 fragments were 6,380 and 5,977 nucleotides (nt) long, respectively. A total of 4,795 raw data reads mapped to RNA1 (coverage of 57X per nt), and 20,656 reads mapped to RNA2 (coverage of 262X per nt). OFV is a bisegmented negative-sense RNA virus of the genus Dichorhavirus. The nearly complete genomic sequence including all of the ORFs of the pandan isolate of OFV was submitted to the GenBank database under the accession numbers OK624601 and OK624602 for RNA-1 and RNA-2, respectively. This is the first report of OFV infecting Pandanus sp. No other viral contigs were identified.

DNA sequencing (DNA-Seq) was used to generate data for the new badnavirus genome assembly.

Total DNA from symptomatic pandan leaves was extracted using a DNeasy Plant Kit (QIAGEN) and used for sequencing on a NovaSeq Illumina platform. Sequencing generated approximately 6.75M paired reads, which were assembled into 344,342 contigs by de novo assembly as described previously. A BLAST search resulted in identification of nine contigs with similarity to sequences from members of the genus Badnavirus. Further assembly resulted in a circular DNA sequence of 7,481 bp with 39.8% GC content (Fig. 1B) and an average assembly coverage of 37X per nt position. The full genome sequence obtained from the DNA-Seq analysis completely matched the contigs obtained from the RNA-Seq experiment. The new pandan badnavirus was therefore detected by both RNA- and DNA-based HTS analysis.

The new badnavirus was tentatively named "pandanus mosaic associated virus" (PMaV), and the complete genome sequence was deposited in the NCBI GenBank database under accession number OK624603.

Sequence analysis revealed a motif complementary to the 3’ end of methionine tRNA (5’- TGG TAT CAG AGC GAG GTT-3’), and this sequence was set as the starting point of the genomic sequence. Three putative open reading frames (ORFs) were predicted using ORFfinder (NCBI). The sequence of each putative protein was searched for conserved protein domains using the CD-Search tool (NCBI). ORF 1 (nt position 213–674) codes for a putative protein of 17.7 kDa containing a domain of unknown function (DUF1319), which has been found only in members of the genus Badnavirus. Pairwise alignments showed that the ORF1-encoded protein shared the highest aa sequence identity (44.2%) with taro bacilliform virus-Tz17 (AWK49020). ORF2 (nt position 671–1,039) encodes a putative protein of 13.6 kDa of unknown function. The ORF2-encoded protein shared the highest aa sequence identity (45.3%) with taro bacilliform virus-Ke52 (AWK49017). ORF3 (nt position 1,039–6,837) codes for a putative polyprotein of 220.4 kDa with conserved protein domains including zinc finger, trimeric dUTPase, aspartic protease (AP), reverse transcriptase (RT), and RNase H domains. The most similar ORF3-encoded polyprotein sequences were those of taro bacilliform CH virus (MG017324) and taro bacilliform virus-Aus7 (MG017318), with 59% and 48% identity, respectively. Pairwise alignment of publicly available sequences of the highly conserved RT + RNase H region (approximately 1 kb) from members of the genus Badnavirus showed the highest nt sequence identity (70.7% and 69.4%) to taro bacilliform CH virus-Et17 (MG017324) and jujube mosaic-associated virus-BJ (MN274946), respectively. According to the current ICTV species demarcation criteria for members of the genus Badnavirus (nt differences greater than 20% in the RT + RNase H region) [4], PMaV is proposed to be a new badnavirus.

The phylogenetic relationship of PMaV to other members of the genus Badnavirus was examined using genetic information for 43 badnaviruses corresponding to both ICTV-approved and tentative members of the genus available in the GenBank database [4, 5]. Multiple sequence alignments were performed using Clustal W. Maximum-likelihood (ML) phylogeny was inferred with MEGA 11, and the general time-reversible model [6, 7] was chosen as the best model for our dataset. A phylogenetic tree was constructed based on the nucleotide sequences of the RT-RH1 domain (Fig. 2). In the ML-inferred phylogram, PMaV and taro bacilliform virus (TaBV) were located in sister taxa.

Fig. 2
figure 2

Phylogenetic tree constructed using RT-RH1 nucleotide sequences of pandanus mosaic associated virus (PMaV) and 43 members of the genus Badnavirus. The tree was constructed by the maximum-likelihood method, using the general time-reversible model. Bootstrap values (1000 replicates) are indicated at each node. The tree is drawn to scale, with branch lengths corresponding to the number of substitutions per site. PMaV is indicated by a star, and rice tungro bacilliform virus (genus Tungrovirus) was used as an outgroup.

Full size image

Because members of two different virus species were detected in the pandan sample analyzed in this study, it remains unclear whether the foliar virus-like symptoms observed (Fig. 1A) were associated with PMaV, OFV, or both. Further studies are needed to determine the role of each virus, both as single and mixed infections, in the observed symptomatology.