Introduction

MIR1444 genes were first reported in 2008 in Populus trichocarpa1, a model tree species with the whole genome sequence available2. Like other microRNA genes (MIRs), after transcription from the MIR1444 loci by RNA polymerase II, long MIR1444 transcripts are processed into 21-nucleotide mature microRNAs (miRNAs), termed miR1444s, under the catalysis of Dicer-like 1 (DCL1) in interaction with several other proteins3,4,5,6. Mature miR1444s target a subset of polyphenol oxidase genes (PPOs) for cleavage in Populus trichocarpa, playing significant regulatory roles in copper homeostasis and stress responses in plants1,7,8,9.

Compared with those deeply conserved old miRNAs, such as miR156, miR159, miR171, miR390 and miR408 inherited from the ancestral embryophyte and miR162, miR164, miR397, miR398, miR399 and miR482 evolved in the ancestral spermatophyte10, miR1444s are evolutionarily young. MIR1444 genes were reported only in P. trichocarpa1,36. Based on the fossil record and comparative genomic analysis, the two sister genera of Salicaceae were estimated to diverge from each other approximately 45 to 52 Ma23,24,25. The whole genomes of two Salix species, including S. purpurea native to Europe and western Asia and S. suchowensis native to the north of China, have been sequenced recently23 (https://phytozome.jgi.doe.gov/pz/portal.html). Since no Salix MIR1444 gene had been reported previously, we first searched current assemblies of the S. purpurea and the S. suchowensis genomes. A MIR1444 gene was identified from each of the two Salix genomes (Fig. 1). We next carried out blast analysis of the identified spu- and ssu-MIR1444 sequences against the high-throughput illumina and 454 sequencing data of S. purpurea and S. suchowensis (Table S1). The results showed that spu-MIR1444 and ssu-MIR1444 were the expressed MIR1444 genes in S. purpurea and S. suchowensis, respectively.

S. matsudana, also known as Chinese willow, is a species of willow native to northwest of China. Although its genome has not been sequenced, there are 228.7 million of RNA-seq reads from five illumina runs available (Table S1)12,37. Blast analysis of MIR1444 sequences against RNA-seq data of S. matsudana showed the existence of an expressed MIR1444 gene, termed sma-MIR1444, in S. matsudana (Fig. 1). Taken together, the results suggest that there is only one MIR1444 gene in a genome of Salix species, such as S. suchowensis, S. purpurea, and S. matsudana.

Molecular cloning of Idesia Polycarpa MIR1444 precursors and phylogenetic analysis of Salicaceae MIR1444s

Idesia is a genus formerly placed in the family Flacourtiaceae, but now included in the family Salicaceae21. Idesia comprises the single species I. polycarpa Maximowicz, which is native to eastern Asia, including China, Japan and Korea. Molecular and morphological evidence has shown that Idesia is the most closely related taxa to Populus and Salix18,19,21. Although the genome of I. polycarpa has not been decoded, transcriptome of male and female flower buds has been sequenced using high-throughput Illumina HiSeq 2000 (SRX1421098). Blast analysis of Populus and Salix MIR1444 genes against RNA-seq data of I. polycarpa flower buds identified partial sequences of two ipo-MIR1444 genes. We then designed primers based on the partial sequences and carried out PCR of ipo-MIR1444 precursors. The results suggest that I. polycarpa contains two MIR1444 genes, termed ipo-MIR1444a and ipo-MIR1444b, respectively (Fig. 1).

Through genome- and transcriptome-wide analysis and molecular cloning, we identified a total of 23 MIR1444 genes in Populus, Salix and Idesia (Fig. 1). A neighbor-joining (NJ) phylogenetic tree for precursor sequences of the identified MIR1444s was constructed using MEGA5.038 (Fig. 2a). MIR1444s could be divided into two groups, 1444a and 1444b. Salix MIR1444s cluster with Populus and Idesia MIR1444b precursors in the 1444b group, while Populus and Idesia MIR1444a precursors cluster together in the 1444a group (Fig. 2a). It suggests that Salix MIR1444s are homologs of Populus and Idesia MIR1444b genes. The homolog of Populus and Idesia MIR1444a genes was lost in Salix plants.

Figure 2: Conservation and divergence of MIR1444s.
figure 2

(a) Phylogenetic relationship of MIR1444 precursors in various Populus, Salix and Idesia species. It includes P. euphratica (peu), P. tomentosa (pto), P. deltoids (pde), P. balsamifera (pba), P. tremula (ptra), P. tremuloides (ptro), P. simonii (psi), P. pruinosa (ppr), S. suchowensis (ssu), S. matsudana (sma), S. purpurea (spu), and I. polycarpa (ipo). Groups 1444a and 1444b indicate two groups identified. (b) and (c) Hairpin structures of ptr-MIR1444a (b) and ptr-MIR1444b (c). miR1444a and miR1444b are shown in red. miR1444a* and miR1444b* are shown in blue. The 5′ and 3′ ends of miRNA/miRNA* pairs identified in the small RNA libraries are shown by broken lines with same color. The number of reads per million is shown at the 5′ end of each 21 nt small RNA identified. (d) and (e) High throughput sequencing analysis of small RNAs from ptr-MIR1444a (d) and ptr-MIR1444b (e) precursors.

Full size image

Divergence of mature miR1444a and miR1444b in P. trichocarpa

Comparison of ptr-MIR1444a and ptr-MIR1444b precursors showed sequence divergence, which led to varied stem-loop structures with different stabilities (ΔG) (Fig. 2b,c). ptr-miR1444s and their corresponding miRNA* sequences are relatively conserved, while the nucleotides outside of this region show greater variation. Comparison of ptr-miR1444a and ptr-miR1444b showed two nucleotide changes (U-to-C, C-to-G) near the 3′ end. Variation at three locations were found between ptr-miR1444a* and ptr-miR1444b*.

In order to further analyze the mature miR1444s in P. trichocarpa, we sequenced six small RNA libraries using high-throughput Illumina sequencing technology. A total of 323,577,100 raw reads were obtained. After removing adaptors and low quality sequences, 211,425,411 clean reads represented by 30,938,856 unique small RNA sequences were obtained. We mapped the unique small RNA sequences to the precursors of ptr-miR1444a and ptr-miR1444b using SOAP2 with no mismatch allowed, respectively39. Of the 244 small RNA hits on ptr-miR1444a precursor, six different miRNA/miRNA* duplexes were found (Fig. 2b). Among them, ptr-miR1444a showed an extremely high abundance of 1,376 RPM (reads per million), accounting for 74.08% of the reads map** to that site. Similarly, five distinct miRNA/miRNA* pairs were obtained from the 171 small RNA hits on ptr-miR1444b precursor (Fig. 2c). Ptr-miR1444b had 796 RPM, accounting for 25.38% of the mapped reads. Compared to ptr-miR1444a, the expression of ptr-miR1444b is lower in plantlets. Interestingly, we found that ptr-miR1444a started 2 nt upstream relative to ptr-miR1444b (Fig. 2b,c). Details on the location of ptr-miR1444a/b and their corresponding miRNAs* are shown in Fig. 2d and e, respectively. Sequence divergence between ptr-miR1444a and ptr-miR1444b was confirmed by map** sequence reads from three previously reported small RNA libraries40 (GSM717875, GSM717876 and GSM717877) to ptr-MIR1444 precursors (Fig. S2). It should be noticed that, in addition to ptr-miR1444b, the ptr-MIR1444b precursor generates another 21 nt small RNA with high sequence reads (1,665 RPM) in our small RNA libraries (Fig. 2c). This small RNA was produced from the 5′ arm of ptr-MIR1444b. The function of this small RNA remains to be elucidated. Taken together, the results suggest that the natural variation between ptr-MIR1444a and ptr-MIR1444b may affect foldback structures and miRNA biogenesis.

Identification and characterization of PPO genes in Populus and Salix

It has been shown that mature miR1444s target PPO genes for cleavage in P. trichocarpa1,7,8. PPOs are copper-binding enzymes catalyzing the dehydrogenation of catechols to the corresponding o-quinones41. It also act as bifunctional enzymes to oxidize monophenols first to o-diphenol intermediates and then to the corresponding o-quinones41. Plant PPO proteins typically consist of N-terminal targeting signal, dicopper centre and C-terminal region42,43. The dicopper centre contains two copper-binding domains, CuA and CuB, each of which is approximately 50 amino acids in length. Analysis of the PPO gene families in 25 sequenced genomes from chlorophytes, bryophytes and lycophytes showed that the number of PPO genes in a genome varied among 0 to 1343. No PPO genes were found in Arabidopsis. In order to identify miR1444-targeted PPO genes, we searched current assemblies of the P. trichocarpa (v3.0), P. euphratica (v1.0), S. purpurea (v1.0) and S. suchowensis (v1.0) genomes2,22,23 (https://phytozome.jgi.doe.gov/pz/portal.html). We identified a total of 34 full-length PPO genes, including 15 from P. trichocarpa, 6 from P. euphratica, 9 from S. purpurea and 4 from S. suchowensis (Table 1). Compared with P. trichocarpa and S. purpurea, the number of full-length PPO genes in P. euphratica and S. suchowensis were significantly less. It could be due to the incompleteness of the current assemblies of the P. euphratica and S. suchowensis genomes. In addition to the full-length genes, a total of 26 partial PPO sequences were identified in the P. trichocarpa, P. euphratica, S. purpurea and S. suchowensis genomes (Table S2). Some of the identified partial sequences could result from the incompleteness of genome assemblies, while a significant proportion appeared to be pseudogenes.

Table 1 Sequence features of PPO genes in P. trichocarpa, P. euphratica, S. purpurea and S. suchowensis.
Full size table

Among the 34 PPO genes, PtrPPO10 and SpuPPO7 contain three introns, PtrPPO7 and SpuPPO4 have two introns, PtrPPO4, PtrPPO9, PtrPPO13, PeuPPO6 and SpuPPO9 include an intron, while the other 25 PPOs are single exon genes (Fig. 3a). We constructed a neighbor-joining (NJ) phylogenetic tree for the deduced PPO proteins using MEGA5.038 (Fig. 3b). As a result, the identified 34 Populus and Salix PPOs could be divided into 5 groups. Group 1 is the largest one. It contains 14 PPOs, including 10 from P. trichocarpa, two from P. euphratica, and two from S. purpurea (Fig. 3b). Groups 2 and 3 contain five and four PPOs, respectively. These PPOs are from four different plant species. Group 4 contain eight PPOs, four of which are from S. purpurea. The three PPOs containing a secretory pathway signal peptide, including PtrPPO13, PeuPPO6 and SpuPPO9, are clustered in group 5 (Fig. 3b, Table 1).

Figure 3: miR1444-mediated cleavage of PPOs in Populus and Salix.
figure 3

(a) Gene structures of PPOs. (b) Unrooted neighbor-joining tree of PPOs. Five groups identified are shown. (c) Validation of miR1444-mediated cleavage of PPOs in P. trichocarpa. The mRNA cleavage sites were determined by the modified 5′ RLM-RACE method. The mRNA sequence of each complementary site from 5′ to 3′ and the ptr-miR1444a and ptr-miR1444b sequences from 3′ to 5′ are shown. Watson-Crick pairing (vertical dashes) and G:U wobble pairing (circles) are indicated. Vertical arrows indicate the 5′ termini of miRNA-guided cleavage products with the frequency of clones shown.

Full size image

miR1444-mediated cleavage of PPO transcripts

Plant miRNAs have perfect or near-perfect complementarities to their targets, allowing an effective prediction of target sequence through computational approaches like psRNATarget44,45. Among the 34 identified PPO genes, complementary sequences of miR1444s were found in 31 genes (Fig. 4). It indicates that 31 of the 34 PPO genes are potential targets of miR1444s. The target sites locate in regions encoding CuB, a conserved domain of PPO proteins (Fig. 4). Analysis of the subcellular localization of PPO proteins using TargetP 1.146 showed that the majority of the predicted targets of miR1444 contained a chloroplast transit peptide at the N-terminus, while the three non-targets of miR1444s included a secretory pathway signal peptide (Table 1).

Figure 4: Multiple sequence alignment of MIR1444 precursors and miR1444-targeted PPOs in Populus and Salix.
figure 4

The schematic diagram of PPOs is shown. cDNA regions coding for conserved domains are shown and indicated by heavy lines with different colors. Mature miRNAs of the MIR1444 gene family target to the cDNA region encoding CuB domain. Sequence alignments of the sense orientation of PPOs and the miRNA* foldback arms and the antisense sequences of miRNA foldback arms are shown in the expanded region. Blue boxes indicate miRNA complementary sites.

Full size image

Plant miRNAs regulate target gene expression mainly through direct cleavage of target mRNAs at the 10th miRNA nucleotide from the 5′ end44,47. To validate miRNA-mediated cleavage of predicted targets, we carried out rapid amplification of 5′ complementary DNA ends (5′-RACE) for fourteen PPO genes on mRNAs isolated from leaves and xylem of P. trichocarpa as described previously48. Six of the PPO genes tested were indeed cleaved by miR1444s in vivo (Fig. 3c), verifying the results from computational prediction (Fig. 4). Examination of the cleavage sites showed that PtrPPO11 was cleaved by ptr-miR1444a, PtrPPO2, PtrPPO6, PtrPPO9 and PtrPPO15 were regulated by ptr-miR1444b, while PtrPPO3 was targeted by both ptr-miR1444a and ptr-miR1444b. It should be noted that the frequency of ptr-miR1444a-guided cleavage products of PtrPPO3 (12/13) is much higher than that of ptr-miR1444b (1/13). Similar results have been obtained previously1. It indicates that PtrPPO3 transcripts are predominantly cleaved by ptr-miR1444a in the plant tissues analyzed, although they are also targets of ptr-miR1444b. Taken together, the results confirm the divergence of mature ptr-miR1444a and ptr-miR1444b, and suggest that both of them are functional.

MIR1444 genes show extensive similarity to PPO targets

In Arabidopsis, several MIRs, such as MIR161 and MIR163, were generated from inverted gene duplication events of target genes49,50. To elucidate the origin of MIR1444 genes, we aligned foldback arms and their flanking sequences of MIR1444 precursors with PPO targets. Similarity was detected among miRNA*-containing 5′ arms, complementary sequences of miRNA-containing 3′ arm, and 78 nt PPO sequences containing target sites and partial of CuB conserved domain-encoding sequences (Fig. 4). The complementarity or similarity to PPO targets was relatively high in the miRNA and miRNA* regions (Fig. 4), suggesting they were under evolutionary constraints and indicating the significance of miR1444s. Compared with miRNA*-containing 5′ arms, complementary sequences of miRNA-containing 3′ arms showed higher similarity to PPO targets. P. trichocarpa ptr-MIR1444a and ptr-MIR1444b showed the greatest similarity (~70%) with the three PPO genes located in chromosome 11, including PtrPPO1, PtrPPO3 and PtrPPO10 (Fig. 4). spu-MIR1444 showed the greatest similarity with SpuPPO1 and SpuPPO2 in S. purpurea. ssu-MIR1444 showed the greatest similarity with SsuPPO2 and SsuPPO3 in S. suchowensis. All of the PPO genes with the greatest similarity with the corresponding MIR1444s are members of groups 2 and 3 (Fig. 3b). It indicates that MIR1444s and PPO genes in these groups probably have a common ancestor. Among the four P. trichocarpa and P. euphratica MIR1444 genes, MIR1444b showed greater similarity with PPO genes than MIR1444a in P. trichocarpa and P. euphratica, respectively. It indicates that MIR1444b is more conserved than MIR1444a.

Expansion of MIR1444 genes through the salicoid genome duplication event

Analysis of the fourfold synonymous third-codon transversion position (4DTV) values in P. trichocarpa and P. euphratica has shown that two ancient whole-genome duplication (WGD) events occurred in the Populus lineage2,22. The recent WGD event, known as the Salicoid duplication event, was estimated to occur around 60 to 65 Ma, and both genera Salix and Populus shared this event2,22,23. P. trichocarpa MIR1444a in chromosome 8 and MIR1444b in chromosome 10 are located in two homologous genome blocks arising from the Salicoid duplication event2 (Fig. 5a). Sequence alignments using zPicture (http://zpicture.dcode.org/) and Generic Synteny Browser (GBrowse-syn) (http://me.lzu.edu.cn/stpd/#main_tabs=2) showed that the genomic DNA segments of peu-MIR1444a in scaffold 1.1 and peu-MIR1444b in scaffold 9.1 shared high level of sequence homology with the segments of ptr-MIR1444a in chromosome 8 and ptr-MIR1444b in chromosome 10, respectively (Fig. 5b,c). The results indicate the expansion of MIR1444 genes through the Salicoid duplication event.

Figure 5: Duplication and divergence of MIR1444 genes in Populus.
figure 5

(a) Locations of P. trichocarpa (ptr) MIR1444a in chromosome 8 (chr8) and MIR1444b in chr10. The schematic diagram of two homologous genome blocks (green) arising from the Salicoid duplication event was adapted from Tuskan et al.9. (b–e) Dot-plots of genomic DNA segments in P. trichocarpa and P. euphratica (peu). The alignment and visualization of two MIR1444s and/or surrounding genomic DNA sequences was performed using zPicture (http://zpicture.dcode.org/). Arrows indicate the corresponding locations of miR1444a and miR1444b.

Full size image

Although MIR1444a and MIR1444b are located in genome blocks arising from the Salicoid duplication event in P. trichocarpa and P. euphratica, significant divergence has occurred in the promoter and downstream regions of MIR1444a and MIR1444b genes (Fig. 5d,e). In contrast, the homologies are higher between genomic DNA regions surrounding ptr-MIR1444a and peu-MIR1444a and between the regions surrounding ptr-MIR1444b and peu-MIR1444b (Fig. 5b,c). The difference of sequence similarities is consistent with the difference of estimated time of the Salicoid duplication event (around 60 to 65 Ma) and P. trichocarpa and P. euphratica divergence (about 14 Ma)2,22,23.

Loss of MIR1444a homologs in Salix through DNA segment deletion

Although the Salicoid duplication event took place before the divergence of the Populus and Salix lineages, only one MIR144 gene was identified in the genomes of S. purpurea and S. suchowensis (Fig. 1). Sequence comparison analysis showed that the genomic regions of MIR1444s in S. purpurea chromosome 10 and S. suchowensis scaffold 90 were highly similar to MIR1444b in P. trichocarpa chromosome 10 and P. euphratica scaffold 9.1 (Fig. 6a–d). It suggests that Salix MIR1444s are homologs of P. trichocarpa and P. euphratica MIR1444b. zPicture (http://zpicture.dcode.org/) analysis showed that S. purpurea chromosome 8 and S. suchowensis scaffold 69 were aligned with P. trichocarpa chromosome 8 and P. euphratica scaffold 1.1 (Fig. 6e–h)2,22,23. However, no homologs of P. trichocarpa and P. euphratica MIR1444a were found in S. purpurea chromosome 8 and S. suchowensis scaffold 69. It was resulted from deletion of a DNA segment with length about 8 kb (Fig. 6e–h). Thus, the loss of P. trichocarpa and P. euphratica MIR1444a homologs in S. purpurea and S. suchowensis occurred after the divergence of Populus and Salix lineages.

Figure 6: Dot-plots of genomic DNA segments in Salix and Populus.
figure 6

The alignment and visualization of two MIR1444s and/or surrounding genomic DNA sequences was performed using zPicture (http://zpicture.dcode.org/). Arrows indicate the corresponding locations of mature miR1444s in Salix and miR1444bs in Populus. Triangles show the positions of P. trichocarpa (ptr) and P. euphratica (peu) miR1444as that are absent from S. purpurea (spu) chromosome 8 (chr8) and S. suchowensis (ssu) assembled genomic DNA scaffold69.

Full size image

Discussion

Plant miR1444s are functionally significant miRNAs. They regulate copper homeostasis and stress responses through cleaving the transcripts of PPO genes in P. trichocarpa1,7,8,9. Previously, MIR1444 genes were reported only in P. trichocarpa, although the mature sequences of miR1444 had been identified from S. matsudana and various Populus species1,7,11,12,14,15,16,17. In this study, we found two MIR1444 genes in various Populus species and one in Salix. Also, we identified two MIR1444 genes from I. polycarpa, a Salicaceae species most closely related to Populus and Salix18,19,21. This brings the number of authentic MIR1444 precursors to 23, including 18 from Populus, 3 from Salix and 2 from Idesia. The 23 MIR1444s can be divided into two groups. Populus and Idesia MIR1444b precursors and Salix MIR1444s were separated from Populus and Idesia MIR1444a precursors in the phylogenetic NJ tree. It shows the divergence between MIR1444a and MIR1444b precursors in Populus and Idesia and suggests that Salix MIR1444s have higher level of homology with MIR1444b precursors than MIR1444a precursors in Populus and Idesia.

In P. trichocarpa, in addition to the precursors, divergence was also observed in mature sequences of miR1444a and miR1444b. Ptr-miR1444a and ptr-miR1444b contain two nucleotide changes (U-to-C and C-to-G) at positions close to the 3′ end. Compared with ptr-miR1444b, ptr-miR1444a started 2 nt upstream. Examination of the precursor sequences and secondary structures showed difference between ptr-MIR1444a and ptr-MIR1444b. It indicates that DCL1-mediated processing of primary miRNAs is probably affected by the precursor sequences and mismatch patterns between miRNA and miRNA*. The results are consistent with previous findings51,52,53,54,55,56. Although the underlying mechanism positioning the cleavage sites in ptr-MIR1444a and ptr-MIR1444b remains to be elucidated, both of the mature ptr-miR1444a and ptr-miR1444b are functional. They were validated to target a subset of PPO genes for cleavage in P. trichocarpa (Fig. 3c). In consistent with the divergence of miRNA sequences, ptr-miR1444a and ptr-miR1444b directed the cleavage of PPO transcripts at two positions with 2 nt distance.

PPOs exist widely in land plants, fungi and some bacteria57, although Arabidopsis thaliana, A. lyrata and various species of green algae do not contain PPO genes43,58. In plants, the number of PPOs varied significantly among species43,59. Through genome-wide analysis, we identified a total of 34 full-length PPO genes in Salicaceae plants, including 15 PtrPPOs, 6 PeuPPOs, 9 SpuPPOs and 4 SsuPPOs. Land plant PPO genes were originally transferred from a bacterium about 450 to 500 Ma through an ancient horizontal gene transfer event during the transition of an early common ancestor of land plants to live on land57. The dynamic of the PPO gene family reflects that PPO genes have undergone frequent gain and lost during plant evolution. Expansion of the PPO gene family in some plant species appears a consequence of gene duplication events, such as the WGD events and the tandem duplication events. The lack of PPO genes in Arabidopsis and the identification of many partial PPO sequences in the genomes of P. trichocarpa, P. euphratica, S. purpurea and S. suchowensis suggest that PPO genes may be lost through deletion and mutation during chromosome rearrangement.

Based on the analysis of Arabidopsis and rice MIR loci, three possible models of MIR origin have been proposed6. In the first model, MIR genes were generated from inverted gene duplication events of target genes49,50. The second model termed spontaneous evolution6,60. According to this model, evolutionarily young MIRs were originated from high density of small-to-medium sized fold-back sequences scattered throughout the plant genome. The last model indicated the derivation of some evolutionarily young plant MIRs from transposable elements6,61. Through the analysis of MIR1444 genes in Populus, Salix and Idesia, we proposed that MIR1444 genes were originated from PPOs through an inverted gene duplication event (Fig. 7). This hypothesis is supported by the fact that MIR1444 genes show extensive similarity to their PPO targets, and the existence of many partial PPO sequences in plant genomes. The inverted duplication event resulted in tail-to-tail orientations of complete or partial PPO gene sequences as proposed for Arabidopsis MIR16349. The tail-to-tail-orientated PPO sequences were diversified through sequence mutation to shorten and gain of bulges in the foldback structure. Continuous mutation generates evolutionarily young MIR1444 genes with sequences and mismatch patterns in and surrounding the miR1444:miR1444* region to be compatible with DCL1-mediated processing6. It has been shown that Salix and Populus share the Salicoid duplication event estimated to be occurred around 60 to 65 Ma2,22,23. Through the comparative analysis of MIR1444 genes in Populus and Idesia, we found that Idesia could also share the Salicoid genome duplication event and Populus and Idesia MIR1444 genes were all expanded through the event (Fig. 7). After duplication, the sequence of MIR1444 genes was further diversified through mutation as evidenced by precursors and mature miR1444 sequences in Populus and Idesia or lost through DNA segment deletion during chromosome rearrangement in Salix (Fig. 7).

Figure 7: Proposed model for MIR1444 origin and evolution.
figure 7

MIR1444 genes were originated from PPOs through an inverted gene duplication event. It resulted in tail-to-tail orientations of complete or partial PPO gene sequences, which were diversified through sequence mutation to shorten and gain of bulges in the foldback structure. Continuous mutation generates MIR1444 genes with sequences and mismatch patterns in and surrounding the miR1444:miR1444* region. MIR1444 genes were expanded through the Salicoid genome duplication event and then further diversified through sequence mutation. A copy of the duplicated MIR1444 genes was lost through DNA segment deletion during chromosome rearrangement in Salix. Pink arrows indicate transcription direction of PPO genes.

Full size image

Evolutionarily young miR1444s regulate PPOs in Populus, Salix and Idesia. The target site locates in a region encoding the conserved CuB domain. Duplication events yielding foldbacks from highly conserved domains are considered to be under strong negative selection, since miRNAs usually target to the regions outside of family-defining domains49. Thus, the origination of conserved CuB domain-targeted MIR1444 genes could be under strong negative selection. It indicates that miR1444s are particularly important for Salicaceae plants. The generation of a miRNA targeting conserved domains is beneficial to regulate the whole gene family or the majority of family members. This type of regulation seems to be vital for a family of genes with similar and significant functions. It has been shown that cytosolic PPOs can trigger a stress response and cause the accumulation of anthocyanins, which eventually lead to reduce the growth rates of plants57. A mechanism to alleviate the effects of cytosolic PPOs is targeting them to plastids57. miR1444s may play a role of fine adjustment during this process. They cleave those PPOs without the intact chloroplast transit peptide-encoding sequence and control the level of plastid-targeted PPOs. The regulation of PPOs is more sophisticated in Populus plants, which contain sequence divergent ptc-miR1444a and ptc-miR1444b. Taken together, evolutionary force-driven origination of MIR1444s may be important for long-term growth and survival of Salicaceae trees in stressful environments. In view that PPOs are dynamic and flexible enzymes evolved to play a variety of specific functions in different plants43, miRNA-mediated and lineage-specific regulation of PPO genes is likely to exist in other PPO-containing plants. Consistently, a grapevine PPO has recently been found to be regulated by a novel miRNA, termed Vv-miR05862. Further investigating miRNAs in other plant species will help to test this hypothesis.

Methods

Plant materials

Seeds of Idesia polycarpa Maxim. var. vestita Diels were collected in a field nursery at Shuyang, Jiangsu Province, China. Populus trichocarpa (genotype Nisqually-1) plants were grown in a greenhouse for about seven months63. To induce shoots, stem segments of vigorously growing plants were cultivated on WPM agar media for three weeks in a controlled growth chamber under the following conditions: temperature 24–26 °C, humidity 60–70%, 16-h light/8-h darkness7,64. Shoots with 1–2 cm height were then excised and cultivated on WPM media with different zinc ion concentration for one month. Stems, leaves and roots were harvested separately, and stored immediately in liquid nitrogen.

Populus and Salix MIR1444 gene identification

P. trichocarpa MIR1444 precursors were downloaded from miRBase23. Genomic loci of MIR1444s were identified through BLAST analysis of ptr-MIR1444 precursors against the current genome assemblies of P. trichocarpa (v3.0), P. euphratica (v1.0), S. purpurea (v1.0) and S. suchowensis (v1.0) using BLASTn2,22,23,29,65. Transcriptome-wide identification of MIR1444 genes was carried out through BLAST analysis of ptr-MIR1444s against the Nt and EST databases and/or RNA-seq reads from illumina and 454 runs using BLASTn65. The SRA accession numbers for illumina and 454 RNA-seq data are listed in Supplementary Table S1.

Prediction of stem-Loop structures

Secondary structures were predicted by the mfold program using the default parameters66. In each case, only the lowest energy structure was selected as described previously48.

Molecular cloning of ipo-MIR1444a and ipo-MIR1444b genes

Blast analysis of Populus and Salix MIR1444 genes against RNA-seq data of Idesia polycarpa flower buds was carried out using BLASTn65. Genomic DNA was isolated from I. polycarpa seeds treated with distilled water for 12 h using the Plant Genomic DNA kit (Tiangen, Bei**g, China). PCR amplification was performed using 100 ng genomic DNA as templates under the following conditions: pre-denaturation at 95 °C for 10 min, then 30 cycles of amplification at 94 °C for 30 s, 55 °C for 30 s and 72 °C for 1 min, followed by a final extension at 72 °C for 10 min. PCR products were gel-purified and cloned into pMD18-T vector (Takara, Shiga, Japan) for sequencing at Sangon Biotech (Shanghai, China). Primers used were listed in Supplementary Table S3.

Small RNA library construction and high throughput sequencing

Total RNA was extracted from leaves, stems and roots of one-month old P. trichocarpa cultivated in vitro using the total RNA purification kit (LC Sciences, Houston, TX, USA). The quality and quantity of RNA were examined using gel analysis and Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA, USA). Equal amount of RNA from each tissue was mixed. A total of 1 μg high quality RNA (RNA integrity number or RIN > 8.5) was used for small RNA library construction as described67. Briefly, 15–30 nt small RNAs were purified from a 15% denaturing polyacrylamide gel and then ligated sequentially to 5′ and 3′ RNA/DNA chimeric oligonucleotide adaptors by T4 RNA ligase 2. RT-PCR amplification was performed. The resulting small RNA libraries were sequenced using the Genome Analyzer GA IIx (Illumina, San Diego, CA, USA) at LC Sciences (Hangzhou, China). Raw reads were firstly processed to filter out the adapter and low quality and low-copy sequences. The obtained clean small RNA sequences were mapped to the precursors of ptr-miR1444a and ptr-miR1444b using SOAP2 with no mismatch allowed39.

Genome-wide identification of PPO genes

To identify full-length P. trichocarpa PPO genes, we first searched P. trichocarpa genome (v3.0) for the conserved pfam polyphenol oxidase middle domain (PF12142)2. The retrieved proteins were then analyzed for conserved domains (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). Those with three conserved domains, including N-terminal targeting signal, dicopper centre and C-terminal region, were considered as full-length PtPPO proteins. Partial sequences of P. trichocarpa PPO genes were obtained by BLAST analysis of full-length PtPPO proteins against P. trichocarpa (v3.0) using tBLASTn2,65.

To identify PPO gene sequences in P. euphratica, S. purpurea and S. suchowensis, we carried out BLAST analysis of PtPPO proteins against the current genome assemblies of P. euphratica (v1.0), S. purpurea (v1.0) and S. suchowensis (v1.0) using tBLASTn22,23,65. An e-value cut-off of 1e–10 was applied to the homologue recognition. Gene models of PPOs were predicted based on the alignments between the retrieved DNA sequences and PPOs from other plant species using BLASTx65. The retrieved DNA sequences were also BLAST-analyzed against the gene models available in the databases of P. euphratica (v1.0), S. purpurea (v1.0) and S. suchowensis (v1.0) using BLASTx65.

Bioinformatics analysis and phylogenetic tree construction

The molecular weight (MW) and theoretical isoelectric point (pI) were predicted using the Compute pI/MW tool on the ExPASy server (http://web.expasy.org/compute_pi/). Intron/exon structures were analyzed using GSDS 2.068. Subcellular localization of PPO proteins was predicted using TargetP 1.146. Conserved domains were analyzed by searching the deduced amino acid sequence of PPOs against the NCBI conserved domain (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). Multiple sequence alignment was carried out using DNAMAN. Phylogenetic tree was constructed using MEGA5.0 by the neighbor-joining (NJ) method with 1000 bootstrap replicates38. Alignment of large genomic DNA segments was performed using zPicture (http://zpicture.dcode.org/) and Generic Synteny Browser (GBrowse-syn)29.

psRNATarget analysis of miR1444 targets

Targets of ptr-miR1444a, ptr-miR1444b, peu-miR1444a, peu-miR1444b, spu-miR1444 and ssu-miR1444 were predicted using psRNATarget with the default parameters45. PPO genes identified from P. trhichcarpa, P. euphratica, S. purpurea and S. suchowensis were used as target transcript candidates. The maximum expectations of 3.0 and the target accessibility-allowed maximum energy to unpair the target site of 25.0 were applied.

5′ RLM-RACE validation of miR1444-directed cleavage

miR1444-directed cleavage of PPO genes were validated using the modified RNA ligase-mediated rapid amplification of 5′ cDNA ends method (5′ RLM-RACE) as described previously48. The SMARTer™ RACE cDNA Amplification kit (Clontech, Mountain View, CA, USA) was used. Total RNA was isolated from pooled tissues containing leaves and xylem of P. trichocarpa. The nesting and the nested PCR amplification of cleaved transcripts was performed using primers listed in Supplementary Table S3.

Additional Information

How to cite this article: Wang, M. et al. Origin and evolution of MIR1444 genes in Salicaceae. Sci. Rep. 7, 39740; doi: 10.1038/srep39740 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.