A transcriptome-wide, organ-specific regulatory map of Dendrobium officinale, an important traditional Chinese orchid herb

Abstract

Dendrobium officinale is an important traditional Chinese herb. Here, we did a transcriptome-wide, organ-specific study on this valuable plant by combining RNA, small RNA (sRNA) and degradome sequencing. RNA sequencing of four organs (flower, root, leaf and stem) of Dendrobium officinale enabled us to obtain 536,558 assembled transcripts, from which 2,645, 256, 42 and 54 were identified to be highly expressed in the four organs respectively. Based on sRNA sequencing, 2,038, 2, 21 and 24 sRNAs were identified to be specifically accumulated in the four organs respectively. A total of 1,047 mature microRNA (miRNA) candidates were detected. Based on secondary structure predictions and sequencing, tens of potential miRNA precursors were identified from the assembled transcripts. Interestingly, phase-distributed sRNAs with degradome-based processing evidences were discovered on the long-stem structures of two precursors. Target identification was performed for the 1,047 miRNA candidates, resulting in the discovery of 1,257 miRNA--target pairs. Finally, some biological meaningful subnetworks involving hormone signaling, development, secondary metabolism and Argonaute 1-related regulation were established. All of the sequencing data sets are available at NCBI Sequence Read Archive (http://www.ncbi.nlm.nih.gov/sra/). Summarily, our study provides a valuable resource for the in-depth molecular and functional studies on this important Chinese orchid herb.

Figure 1. Heatmaps showing the expression patterns of the transcripts highly accumulated in specific organs of Dendrobium officinale.

(A) The expression patterns of the transcripts highly expressed in the flowers. (B) The expression patterns of the transcripts highly expressed in the leaves. (C) The expression patterns of the transcripts highly expressed in the roots. (D) The expression patterns of the transcripts highly expressed in the stems. The expression levels (normalized in RPKM, reads per kilobase of exon model per million mapped reads; please refer to Materials and Methods for RPKM calculation) of all the transcripts were rescaled by log10. For each organ, there are two biological replicates (“Root1” and “Root2”, for example). The heatmaps were drawn by using Treeview. Please refer to Table S2 for the detailed list of all the transcripts.

Figure 2. Examples of highly structured microRNA (miRNA) precursor candidates identified from Dendrobium officinale.

(A) The transcript comp108689_c0_seq1 assembled by RNA-seq reads could form an internal hairpin structure with a long-stem region (partially delineated by a pink box) encoding dof-miR-109 (red line) and dof-miR-109* (blue line). The accumulation levels (normalized in RPM, reads per million; please refer to Materials and Methods for RPM calculation) of dof-miR-109 and dof-miR-109* in the eight small RNA (sRNA) sequencing libraries (roots, flowers, stems and leaves, two biological replicates for each organ) are shown in the two diagrams respectively. The expression levels of dof-miR-109* in the two libraries from the stems were highlighted in pink background color since degradome signatures from the stems of Dendrobium officinale were detected at the 3’ end of dof-miR-109*. (B) The transcript comp168357_c1_seq16 assembled by RNA-seq reads could form an internal hairpin structure with a long-stem region (partially marked by a pink box) encoding dof-miR-1023 (red line) and dof-miR-1023* (blue line). The accumulation levels (in RPM) of dof-miR-1023 and dof-miR-1023* in the eight sRNA libraries are shown in the two diagrams respectively. The expression levels of dof-miR-1023 in the eight libraries were highlighted in pink background color since degradome signatures from the four organs of Dendrobium officinale were detected at the 5’ end of dof-miR-1023. The expression levels of dof-miR-1023* in flowers and leaves were also highlighted because degradome signatures from the two organs were detected at the 3’ end of dof-miR-1023*. (C) The transcript comp170707_c6_seq1 assembled by RNA-seq reads could form an internal hairpin structure with a long-stem region (partially delineated by a pink box) encoding two miRNA dof-miR-154 and dof-miR-988 (both denoted by red lines). The accumulation levels (in RPM) of the two miRNAs in the eight sRNA libraries are shown in the diagram together. And, their expression levels in the stems and leaves were highlighted in pink background color because degradome signatures from the two organs were detected at the 3’ end of dof-miR-154 and at the 5’ end of dof-miR-988. The secondary structures of the three transcripts were predicted by using RNAfold (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi).

Figure 3. Tandemly distributed small RNAs (sRNAs) identified on the highly structured microRNA (miRNA) precursor candidates in Dendrobium officinale.

(A) The transcript comp124801_c0_seq1 assembled by RNA-seq reads could form an internal hairpin structure with a long-stem region (partially delineated by a pink box). In addition to generating miRNAs (dof-miR340, dof-miR341, dof-miR1002 and dof-miR1004) and miRNA*s (dof-miR1002* and dof-miR1004*), the long-stem region potentially encodes three pairs of tandemly distributed sRNAs (124801_sRNA1 and 124801_sRNA6, 124801_sRNA2 and 124801_sRNA5, and 124801_sRNA3 and 124801_sRNA4). Each pair possesses 2-nt 3’ overhangs. Five degradome signatures (124801_degr1 to 124801_degr5) were detected at the ends of certain tandemly distributed sRNAs. And, 124801_degr3 also appeared at the 5’ ends of dof-miR-340 and dof-miR-341, and 124801_degr4 and 124801_degr5 are present at the 3’ ends of dof-miR-1004. The accumulation levels (normalized in RPM, reads per million; please refer to Materials and Methods for RPM calculation) of the degradome signatures, the miRNAs, the miRNA*s and the tandemly distributed sRNAs are shown in the diagrams on the right of the panel. Their accumulation levels in the stems of Dendrobium officinale were highlighted in pink background color. (B) The transcript comp168357_c1_seq6 assembled by RNA-seq reads could form an internal hairpin structure with a long-stem region (partially included in a pink box). Within this region, three pairs of sRNAs (including 168357_sRNA2 and 168357_sRNA6, 168357_sRNA3 and 168357_sRNA5, and the dof-miR-1023/dof-miR-1023* duplex) along with two unpaired sRNAs (168357_sRNA1 and 168357_sRNA4) were identified to be distributed tandemly. Each pair possesses 2-nt 3’ overhangs. Eleven degradome signatures (168357_degr1 to 168357_degr11) were detected at the ends of certain tandemly distributed sRNAs. The accumulation levels (in RPM) of the degradome signatures, the miRNAs, the miRNA*s and the tandemly distributed sRNAs are shown in the diagrams on the right of the panel. The secondary structures of the two transcripts were predicted by using RNAfold (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi).

Figure 4. MicroRNA--target pairs supported by degradome signatures, and comparison of expression patterns between the microRNAs and their targets.

(A) Comp155533_c0_seq1 and comp155533_c0_seq3 were predicted to be targeted by seven microRNAs (dof-miR-1017/1020/1021/1022/1023/1024/1034). The scatter diagram shows the degradome-seq data supporting the predicted regulation. Y axis measures the levels (in RPM, reads per million) of the degradome signatures, and X axis represents the region surrounding the microRNA binding site (grey horizontal line). Compared with the signatures from the other organs, the most abundant one was identified from the flowers (arrowhead). The first line chart shows the levels (in RPKM, reads per kilobase of exon model per million mapped reads) of the two targets, and the other two charts show the levels (in RPM) of three microRNAs (dof-miR-1017/1021/1034). Consistent with the degradome intensity, both the targets and the microRNAs are highly accumulated in the flowers (grey background color). (B) Comp156499_c0_seq2 was predicted to be targeted by seven microRNAs (dof-miR-985/986/987/988/989/990/991). The scatter diagram shows the degradome-seq data supporting the predicted regulation. Y axis measures the levels of the degradome signatures, and X axis represents the region surrounding the predicted microRNA binding site (grey horizontal line). Compared with the signatures from the other organs, the most abundant one was identified from the leaves (arrowhead). The first line chart shows the level of the target, and the other two show the levels of two microRNAs (dof-miR-988/989). Consistent with the degradome intensity, the two microRNAs are highly accumulated in the leaves (grey background color). (C) Comp126829_c0_seq1 was predicted to be targeted by dof-miR-1047. The scatter diagram shows the degradome-seq data supporting the predicted regulation. Y axis measures the levels of the degradome signatures. X axis represents the region surrounding the predicted microRNA binding site (grey horizontal line). Within this region, the degradome signals were only detectable in the flowers, and the most abundant cleavage signal was denoted by an arrowhead. The first line chart shows the level of the target, and the other chart shows the level of dof-miR-1047. Consistent with the degradome intensity, the target is highly expressed in the flowers (grey background color).

Figure 5. Degradome-seq data-supported, microRNA-mediated regulatory networks in Dendrobium officinale.

(A) MicroRNA-mediated network implicated in plant development. (B) MicroRNA-mediated network involved in hormone signaling. (C) Transcripts encoding Argonaute 1 are targeted by dof-miR-642, dof-miR-644, dof-miR-645, dof-miR-646 and dof-miR-647. (D) MicroRNA-mediated network involved in secondary metabolism. For the above networks, all of the microRNA–target regulatory relationships are supported by degradome signatures (please refer to Figure S5 and Table S6). The functional involvement of the networks was deduced based on the annotations of the target transcripts. The networks were drawn by using Cytoscape.