Based on expression analysis for the miRNA families of wheat deposited in miRNA database¹, a suite of the miRNA members were shown to be responses to drought, salt, and deprivations of nitrogen (N), and phosphate (Pi) (our unpublished data). Among these, TaemiR408 (accession number MI0006177) showed induced expression upon both Pi starvation and salt stress, suggesting the miRNA-mediated plant responses to these stressors. To understand whether miR408 is evolutionarily conserved across plant species, we amplified the TaemiR408 homolog of tobacco, a model eudicot plant species frequently used in genetic analysis, based on reverse transcriptase-polymerase chain reaction (RT-PCR) using the TaemiR408 specific primers (Supplementary Table S1). Root cDNA of tobacco (Nicotiana tabacum cv. Wisconsin 35) was acted as templates.

miRNA-mediated biological processes closely associate the miRNA regulation on target genes at posttranscriptional or translation level. To define the target genes of TaemiR408, an online tool referred to as psRNATarget (Plant microRNA Potential Target Finder²) was adopted to predict the genes putatively interacted by TaemiR408. As suggested, the mature TaemiR408 sequence was subjected to scanning against two of wheat cDNA databases, including (i) Triticum aestivum (bread wheat), transcript, cDNA library, TGACv1 and (ii) Triticum aestivum (wheat), DFCI Gene Index (TAG), version 12, released by 2010-04-18. Biological functions of the six target genes identified were defined based on gene BLASTn search analysis in NCBI.

The expression patterns of TaemiR408, miR408 tobacco homolog (referred to as NtMIR408 hereafter), and the target genes of TaemiR408 upon Pi starvation and salt stress were evaluated. With this aim, wheat (cv. Shiyou 20) and tobacco (cv. Wisconsin 35) seedlings were cultured normally in standard Murashige and Skoog (MS) solution to the fourth leaf stage in a growth chamber under follow condition: photoperiod of 12 h/12 h (day/night) with 300 μmolE⋅m^-2⋅s^-1 light intensity during light phase, temperature of 28°C/23°C (day/night), and a relative humidity from 65 to 70%. They were transferred in the modified MS solutions either containing reduced Pi (0.012 mM Pi) or supplemented with NaCl (200 mM) for stress treatments and followed recovery treatments, which were established by transferring the 48 h-stressed seedlings again to standard MS solution. At time points of 0 h (before treatment), 6, 12, 24, and 48 h after stress treatment, and 6, 12, 24, and 48 h after recovery treatment, root tissues were collected and subjected to expression evaluation for TaemiR408, NtMIR408, and the target genes based on qRT-PCR performed as previously described (Guo et al., 2013). Briefly, total RNA from roots was extracted by TRIzol reagents (Invitrogen, United States). After treatment with RNase-free DNase (TaKaRA, Dalian, China) to avoid genomic DNA contamination, the total RNA (2 μg) was subjected to synthesis for the first-strand cDNA with RT-AMV transcriptase (TaKaRa, Dalian, China) in 20 μL reaction volume using oligo (dT)18 at 42°C for 30 min, as suggested. qRT-PCR analysis was performed in a total volume of 25 μL containing 12.5 μL of SYBR Premix ExTaqTM (TaKaRa, Dalian, China), 0.5 μL of forward and reverse primers, 1 μL cDNA and 10.5 μL nuclease-free water. Transcripts of the miRNA members and the target genes were calculated based on the by 2^-ΔΔCT method using wheat Tatubulin and tobacco Nttubulin as an internal control. The gene-specific primers used for qRT-PCR analysis are shown in Supplementary Table S1.

To verify the target genes acted by TaemiR408, we analyzed the cleavage products of them based on Poly (A) polymerase -mediated 3′ rapid amplification of cDNA ends (PPM-RACE) and RNA ligase-mediated 5′ rapid amplification of cDNA ends (RLM-RACE) as previously reported (Song et al., 2012). Briefly, total RNA of wheat (cv. Shiyou 20) roots at various time points including at normal growth (0 h, CTR), 48 h after Pi starvation, and 48 h after Pi normal recovery was polyadenylated at 37°C for 60 min in a 50 μl reaction mixture (containing 5 μg of total RNA, 1 mM ATP, 2.5 mM MgCl₂, and 8 U poly (A) polymerase) (Ambion, United States). They were then subjected to ligation to a 5′ adapter (5′-CGACUGGAGCACGAGGACACUGACAUGGACUGAAGGAGUAGAAA-3′) using T4 RNA ligase (TaKaRa, Japan). PPM-RACE and RLM-RACE were performed using the GeneRacer kit (Invitrogen). The GeneRacer Nested Primer and gene-specific primers for PPM-RACE and RLM-RACE reactions are shown in Supplementary Table S1. mRNA cleavage products were detected based on qRT-PCR performed similarly as previously described (Guo et al., 2013).

A transgene analysis was performed to characterize whether TaemiR408 is involved in the mediation of plant responses to Pi starvation and salt stresses. Given the conserved nature of miR408 across wheat and tobacco as described above, we transformed TaemiR408 into tobacco, due to its genetic transformation conveniently relative to that of wheat. To this end, RT-PCR was performed to amplify the TaemiR408 precursor sequence using specific primers (Supplementary Table S1), the amplified product was then inserted into the NcoI/BstEII restriction sites in binary vector pCAMBIA3301 under the control of CaMV35S promoter. Then the expression cassette was integrated into Agrobacterium tumefaciens strain EHA105 using conventional heat shock approach. Genetic transformation of tobacco (cv. Wisconsin 35) using EHA105 transformants and further generation of tobacco lines with TaemiR408 overexpression were conducted as described previously (Sun et al., 2012). The TaemiR408 transcripts in transgenic lines were evaluated based on qRT-PCR.

Two T3 lines showing more TaemiR408 transcripts (Supplementary Figure S2A) and one copy insertion of target in genome (Supplementary Figure S2B), OE2 and OE3, were selected to characterize the TaemiR408 function in mediating responses to Pi starvation and salt stress. With this aim, 10-day-old evenly seedlings of the transgenic lines and wild type (WT) were cultured in plastic pots filled by vermiculite and supplied by standard MS solution (sufficient-Pi, 1.2 mM Pi) or modified MS solution containing reduced Pi (deficient-Pi, 0.05 mM Pi) for Pi starvation treatment, and with standard MS solution (control) or modified MS solution supplemented with salt (200 mM NaCl) for salt treatment. Three weeks later, the transgenic and WT plants were subjected to evaluation of phenotypes, biomass, P-associated traits, and osmolyte contents. Of these, the phenotypes were recorded based on images taken by a digital camera; biomass were obtained after oven drying; P concentrations and P amounts were assessed as described previously (Guo et al., 2011); contents of osmolytes (i.e., proline and soluble sugar) were analyzed as reported (Du et al., 2013). In addition, to understand expression patterns of the target genes in transgenics, we searched the N. tabucum cDNA database [DFCI Gene Index (NTGI), version 7] and identified three TaemiR408 target homologs, including NtBCP (TC142445), NtKRP (FS391808), and NtAMP (TC145853), using the online tool psRNATarget (Plant microRNA Potential Target Finder³). The expression patterns of these target genes in Pi starvation- and salt stress-challenged transgenic lines (OE2 and OE3) and wild type (WT) were assessed based on qRT-PCR using gene specific primers (Supplementary Table S1).

Photosynthesis modified by internal and environmental cues impacts greatly on plant growth, development, and the response of plants to abiotic stresses, which causes reduction on dry mass production and yield potential (Gu et al., 2017). To address if the TaemiR408-modified plant biomass under Pi starvation and salt stress associate with the modulation of miRNA on photosynthesis behaviors, three parameters effectively reflecting photosynthetic function of leaves, including photosynthetic rate (Pn), PSII efficiency (ΦPSII), and non-photochemical quenching (NPQ), were investigated in the upper leaves of transgenic and WT plants after the Pi starvation and salt treatments. The parameters were assessed as described previously (Guo et al., 2013).

The Pi taken up of plants from growth media as well as internal P translocation across tissues are mediated by the membrane-bound phosphate transporters (PTs), during which uses ATP as the driving power (Mudge et al., 2002; Shin et al., 2004). The improved Pi accumulation in TaemiR408 overexpressors under Pi starvation prompted us to characterize the expression patterns of the tobacco PT genes. To this end, a suite of tobacco PT genes, including NtPT (DI040486), and NtPT1 to NtPT5 (AB020061, AF156696, AB042950, AB042951 and AB042956, respectively) was subjected to expression evaluation in the Pi-deprived transgenic lines and WT based on qRT-PCR using gene specific primers (Supplementary Table S1), with internal reference Nttubulin for expression normalization.

PT gene expression analysis as above revealed that one of them, NtPT2, exhibited significantly upregulated expression in transgenic lines relative to WT, suggesting its potential in the regulation of Pi acquisition in Pi-deprived transgenic plants. To functionally characterize its role in mediating Pi taken up, transgenic lines with knockdown of this PT gene were generated. With this aim, RT-PCR was performed to amplify the open reading frame (ORF) of NtPT2 in antisense orientation using specific primers (Supplementary Table S1), the amplified product was inserted into NcoI/BstEII restriction sites of binary vector pCAMBIA3301 downstream the CaMV35S promoter. Integration of the expression cassette into EHA105 and transformation of NtPT2 into tobacco were performed to be similar in establishment of the TaemiR408 overexpressors. NtPT2 transcripts in lines with knockdown of this PT gene were evaluated based on qRT-PCR. Two lines with much less target transcripts at T3 generation (Anti-PT2-2 and Anti-PT2-5) were selected for further functional analysis. For this, 10-day-old transgenic and WT seedlings were vertically cultured on agar MS medium (sufficient-Pi, 1.2 mM) and modified MS nutrients with reduced Pi (deficient-Pi, 0.05 mM) for Pi normal and starvation treatments, respectively. Two weeks later, phenotypes, biomass, and P-associated traits in transgenic lines and WT were evaluated performed to be similar in assessing these traits in OE2 and OE3.

ABA-dependent signaling pathways are closely associated with plant response to the abiotic stresses, including water deficit and salt stress (Fujii and Zhu, 2009). Of which, the ABA receptors and SnRK2 family proteins act as critical mediators in transduction the ABA signaling (Ma et al., 2009; Park et al., 2009). To characterize if the TaemiR408-mediated salt response is related to the ABA signaling, a suite of ABA receptor and SnRK2 encoding genes released in the NCBI GenBank database, including seven ABA receptor genes (i.e., NtPYR1, NtPYL2, NtPYL4, NtPYL8, NtPYL9, NtPYL11, and NtPYL12) and ten SnRK2 kinase genes (i.e., NtSAPK1, NtSAPK2, NtSAPK2;1, NtSAPK2A, NtSAPK2A;1, NtSAPK2E, NtSAPK2E;1, NtSAPK2I, NtSAPK3, and NtSAPK7), was subjected to expression evaluation in the salt-challenged TaemiR408 overexpressors and WT. qRT-PCR was performed to understand the transcripts abundance of the ABA receptor and SnRK2 family genes, in which, Nttubulin was used as internal reference. Accession numbers and specific primers for these ABA signaling genes are listed in Supplementary Table S1.

Expression analysis revealed that NtABR2 and NtSAPK3 are significantly upregulated in the TaemiR408 overexpressors after salt stress, suggesting their potential in mediating salt stress response. Therefore, transgenic lines with knockdown of them were generated to define their function in mediating salt tolerance. With this purpose, RT-PCR was performed to amplify the ORFs of NtABR2 and NtSAPK3 in antisense orientation using specific primers (Supplementary Table S1), the ORFs were then inserted into NcoI/BstEII restriction sites of binary vector pCAMBIA3301 under the control of CaMV35S promoter. Transgenic lines were generated similarly in establishment of the TaemiR408 overexpressors as aforementioned. The T3 lines Anti-PYL2-2 and Anti-PYL2-3 for NtABR2 and Anti-SAPK3-3 and NtSAPK3-4 for NtSAPK3 were subjected to salt stress treatment by vertically culturing the evenly 7-day-old seedlings on agar media supplemented without or with 200 mM NaCl. Two weeks after treatments, phenotypes, biomass, and contents of proline and soluble sugar in these lines were assayed to be similar in assessment of those in TaemiR408 overexpression lines.

Averages of gene expression levels in qRT-PCR analysis, plant biomass, photosynthesis parameters, P concentrations and P accumulative amounts, proline contents, and soluble sugar contents in the transgenic lines and WT were derived from the results of four replicates. Standard errors of averages and significant differences were analyzed using the Statistical Analysis System software (SAS Corporation, Cory, NC, United States).

The mature sequence of TaemiR408 is 21 nt-long in length (5′-cugcacugccucuucccuggc-3′), which is situated in a 187 nt-long precursor. Supplementary Figure S1A shows the secondary stem-loop structure established by the TaemiR408 precursor. Given that most miRNAs share conserved nature across eukaryotes, we addressed if miR408 is also present in both monocot and eudicot by identifying the miRNA paralog in tobacco. Sequencing analysis on the RT-PCR product derived from tobacco (cv. Wisconsin 35) revealed an identical precursor of TaemiR408 in N. tabacum (designated as NtMIR408) (Supplementary Figure S1B). Our amplified NtMIR408 is not same as the miR408 member deposited in N. tabacum miRNA database (Accession No. MI0021410), suggesting that miR408 is conserved and the tobacco miR408 family consists of a set of members.

Based on running an online tool, the genes putatively targeted by TaemiR408 were predicted. Results indicated that totally six genes were interacted by this wheat miRNA, including five from wheat cDNA database and one from expressed sequence tag (EST) database. The target genes from the cDNA database were as follows: TRIAE_CS42_1BTRIAE_CS42_5DL_TGACv1_433117_AA1402600, a gene encoding chemocyanin putatively functional in biochemical metabolism (TaCP, KC852069); TRIAE_CS42_5AL_TGACv1_379889_AA1257020, a gene coding for mavicyanin functioning in primary biochemical process (TaMP, XM_020292436); TRIAE_CS42_7AS_TGACv1_571397_AA1847350, a gene encoding blue copper protein involving secondary biochemical metabolism (TaBCP, XM_020294339); TRIAE_CS42_U_TGACv1_641170_AA2087300, a gene for filament involving the cellular microtubule organization (TaFP, XM_020310872); and TRIAE_CS42_6BS_TGACv1_513906_AA1651810, a gene for F-box/kelch-repeat protein functioning in protein-protein interaction (TaKRP, XM_020290582). The target gene from the EST database encodes an AMP-binding protein involving the binding of cyclic AMP (TaABP, BJ225979). Figure 1 shows the base pairing characterization between TaemiR408 and the target genes. BLASTn analysis for the TaemiR408 target genes suggested that they are categorized into diverse function families, including three involved in biochemical metabolism (i.e., TaCP TaMP, and TaBCP), two in signaling transduction (TaKRP and TaABP), and one in microtubule organization (TaFP). Thus, TaemiR408 establishes putative action module(s) with the target genes and exerts distinct biological functions in plants.

Expression patterns of TaemiR408, NtMIR408, and the target genes of TaemiR408 upon Pi starvation and salt stress were investigated in more detail. Results indicated a similar expression pattern for TaemiR408 and NtMIR408 under the stresses as well as upon recovery treatments; the two miRNAs both showed upregulated transcripts abundance over 48 h regimen of the stressors and whose elevated expression was downregulated along with 48 h regimen of the normal recoveries (Figures 2A, 3A). All of the target genes displayed reverse expression patterns to miRNAs upon Pi starvation and Pi recovery treatments as well as upon salt stress and salt recovery treatments; they showed a downregulated expression pattern over 48 h regimen of the stressors and whose repressed transcripts under stresses were gradually restored along with 48 h regimen of the recoveries, albeit that the expression amplitudes of them in response to these stressors are different (Figures 2B–G, 3B–G). Therefore, TaemiR408 regulates the target genes post-transcriptionally and establishes putative miRNA/target modules that mediate plant Pi starvation and salt responses.

To experimentally verify the target genes, PPM-RACE and RLM- RACE that effectively detect the 3′ and 5′ end fragments of the cleaved products after miRNA mediation, respectively, were performed. Among the six target genes (i.e., TaCP, TaMP, TaBCP, TaFP, TaKRP, and TaAMP) that showed converse expression patterns to TaemiR408 upon Pi starvation and salt stress, we detected the 3′ end products of TaCP, TaMP, TaBCP, and TaFP based on PPM-RACE and the 5′ end products of TaKRP and TaAMP by RLM- RACE in the roots of normal growth (0 h, CTR), 48 h after Pi starvation, and 48 h after Pi recovery. At the assayed time points, the cleavage products of the target genes are all consistent with the miRNA transcripts abundance (Figures 2A, 4A). These results confirm the regulation of these target genes under TaemiR408 at the posttranscriptional level.

Two TaemiR408 overexpressors (OE2 and OE3) were cultured under contrasting Pi levels for miRNA functional analysis in mediating Pi starvation response. Under the sufficient-Pi condition, OE2 and OE3 exhibited comparable phenotypes (Figure 4B), biomass (Figure 5A), photosynthesis parameters (i.e., Pn, ΦPSII, and NPQ) (Figures 5B–D), P concentrations (Figure 6A), and P accumulative amounts (Figure 6B) with WT. In contrast, under the Pi-starvation stress, the transgenic lines displayed modified growth, P-associated traits, and photosynthetic parameters, showing improved phenotypes (Figure 4B), elevated biomass (Figure 5A), P accumulation (Figure 6A), Pn and ΦPSII, and decreased NPQ (Figures 5B–D) relative to WT, although similar Pi concentrations were observed in the Pi-deprived transgenic and WT plants (Figure 6A). Additionally, the miR408 target genes including NtBCP, NtKRP, and NtAMP showed lowered expression levels in the transgenic lines (OE2 and OE3) than in wild type under low-Pi stress (Supplementary Figures S3A–C). These results suggest the critical role of TaemiR408 in mediating Pi starvation tolerance through regulation of target genes at the posttranscriptional level.

The more P amounts in Pi-deprived TaemiR408 overexpression lines suggested the contribution of an increased Pi acquisition capacity of root system to the miRNA-mediated low-Pi tolerance. To characterize the PT genes involving the elevated root Pi taken up in transgenic lines, six tobacco PT genes (i.e., NtPT, and NtPT1 to NtPT5) were subjected to expression evaluation in the Pi-deprived transgenic and WT plants. Among them, NtPT2 showed significantly upregulated expression in lines OE2 and OE3 compared with that in WT, which was contrast to other genes displaying unaltered expression among the Pi-deprived transgenic and WT plants (Figure 7A). This finding suggests the transcriptional regulation of NtPT2 under TaemiR408.

Two lines with NtPT2 knockdown (Anti-PT2-2 and Anti-PT2-5, Supplementary Figure S4) were cultured under two contrasting Pi conditions to characterize the PT role in mediating Pi acquisition. Under sufficient-Pi condition, no obvious variations were observed on phenotype, biomass, P concentration, and P accumulation among the transgenic lines and WT (Figures 7B–F). Under Pi starvation treatment, however, the lines with NtPT2 knockdown exhibited deteriorated phenotypes (Figure 7C), and decreased biomass (Figure 7D) and P amounts (Figure 7E) relative to WT. NtPT2 thus acts as a high-affinity PT gene and contributes to the TaemiR408-improved Pi accumulation.

Similar to the investigation for TaemiR408-mediated plant Pi starvation response as described above, OE 2 and OE 3, two TaemiR408 overexpressors, were cultured in vermiculite supplemented with NaCl (200 mM) to evaluate the miRNA-mediated salt tolerance. Under normal condition, the transgenic lines exhibited comparable phenotypes (Figure 4) and biomass (Figure 5A) with WT, which were comparable to those cultured under the sufficient-Pi condition as aforementioned. Additionally, no obvious variations were observed on photosynthesis parameters (i.e., Pn, ΦPSII, and NPQ) (Figures 5B–D) and contents of proline (Figure 6C) and soluble sugar (Figure 6D) in transgenic and WT plants. Under salt stress treatment, the TaemiR408 overexpression lines displayed improved phenotypes, biomass, elevated proline and soluble sugar contents, and modified photosynthesis parameters (i.e., increased Pn and ΦPSII and decreased NPQ) with respect to WT (Figures 4, 5A–D, 6C,D). Moreover, expression analysis revealed that NtBCP, NtKRP, and NtAMP, three target genes of miR408, exhibited reduced transcripts abundance in the transgenic lines (OE2 and OE3) than in the wild type plants under salt stress treatment (Supplementary Figures S3A–C). Therefore, TaemiR408 regulates the target genes at posttranscriptional level and plays important roles in mediating salt tolerance, which is associated with the miRNA-improved osmoregulatory capacity and photosynthesis function of plants upon salt stress.

Abscisic acid is swiftly induced in the tissues of plants upon osmotic stresses, such as high salinity and drought, which further involves the plant response to above stressors through an ABA-dependent pathway (Fujii and Zhu, 2009). That the TaemiR408-mediated salt tolerance associates elevated osmolytes prompted us to investigate whether the ABA signaling pathway is modulated by this miRNA member. To address this, a large set of the ABA signaling component genes, including seven encoding ABA receptors and ten coding for SnRK2 family proteins, were subjected to expression evaluation in the salt-challenged TaemiR408 overexpression lines. Among the genes examined, an ABA receptor gene NtPYL2 and a SnRK2 gene NtSAPK3 displayed significantly upregulated expression in the transgenic lines relative to WT (Figures 8A,B). Other genes aside from these two behaved unaltered transcripts in the transgenic and WT plants (Figures 8A,B). These expression results suggest the transcriptional regulation of NtPYL2 and NtSAPK3 under TaemiR408 and the involvement of them in modulating ABA signaling that possibly impacts on osmolytes accumulation, photosynthesis, and salt tolerance of plants upon salt stress.

The function of NtPYL2 and NtSAPK3 in mediating salt response was investigated using transgenic lines with knockdown of these genes. Expression analysis revealed that the lines Anti-PYL2-2 and Anti-PYL2-3 showed drastic downregulation of NtPYL2 whereas Anti-SAPK3-3 and Anti-SAPK3-4 repressed expression of NtSAPK3 (Supplementary Figures S5, S6). Therefore, these lines together with WT cultured under normal condition or salt stress were subjected to evaluation of the gene function in mediating plant salt response. Under normal growth condition, all these lines exhibited comparable phenotypes, biomass, and proline and soluble sugar contents with WT seedlings (Figures 8C,E, 9A–F). Under salt stress treatment, however, the lines displayed dramatically modified growth and osmolytes contents, showing smaller stature (Figures 8D,F), lower biomass (Figures 9A,B), and reduced contents of proline (Figures 9C,D) and soluble sugar (Figures 9E,F) than those of WT. These results suggested that the TaemiR408-mediated salt tolerance closely associates the miRNA role in modulating ABA signaling pathway via transcriptionally regulating distinct ABA signaling genes.