Heterologous expression of a rice miR395 gene in Nicotiana tabacum impairs sulfate homeostasis

Sulfur participates in many important mechanisms and pathways of plant development. The most common source of sulfur in soil –SO42−– is absorbed into root tissue and distributed into aerial part through vasculature system, where it is reduced into sulfite and finally sulfide within the subcellular organs such as chloroplasts and mitochondria and used for cysteine and methionine biosynthesis. MicroRNAs are involved in many regulation pathways by repressing the expression of their target genes. MiR395 family in Arabidopsis thaliana has been reported to be an important regulator involved in sulfate transport and assimilation, and a high-affinity sulphate transporter and three ATP sulfurylases (ATPS) were the target genes of AthmiR395 (Arabidopsis thaliana miR395). We have cloned a miR395 gene from rice (Oryza sativa) and studied its function in plant nutritional response. Our results indicated that in rice, transcript level of OsamiR395 (Oryza sativa miR395) increased under sulfate deficiency conditions, and the two predicted target genes of miR395 were down-regulated under the same conditions. Overexpression of OsamiR395h in tobacco impaired its sulfate homeostasis, and sulfate distribution was also slightly impacted among leaves of different ages. One sulfate transporter (SULTR) gene NtaSULTR2 was identified to be the target of miR395 in Nicotiana tobacum, which belongs to low affinity sulfate transporter group. Both miR395 and NtaSULTR2 respond to sulfate starvation in tobacco.


Results
Sulfate regulates the expression of OsamiR395 and its target genes. According to previous research and miRNA database (http://mirbase.org), 24 family members belonging to four clusters comprise OsamiR395 family 26 . The sequence of mature OsamiR395 is highly conserved while the pre-microRNA sequences are divergent. It has previously been demonstrated in Arabidopsis that mature AthmiR395 transcript accumulates under sulfur-limited conditions 18,19,27 . To investigate whether OsamiR395 also responds to low sulfate conditions as its counterpart in Arabidopsis, transcript level of OsamiR395 in two weeks old rice plants grown in N6 solid medium supplemented with different concentrations of sulfate was analyzed. Both northern blotting and stemloop RT-PCR results showed that the transcripts of mature  Fig. 1a,b).
In plant nucleus, miRNA gene is first transcribed into a long pri-miRNA, which is then processed into pre-miRNA and finally mature miRNA that is later translocated by HASTY into cytoplasm and induces the degradation of its target gene(s). To further understand whether OsamiR395 is regulated at the transcription level or post-transcription level, real-time PCR experiment was conducted to investigate the transcript level of pri-OsamiR395h in two weeks old rice plants grown in N6 solid medium supplemented with 0, 20, 1500 or 2000 μ M SO 4 2− . Real-time PCR results showed that excess sulfate could repress the accumulation of pri-OsamiR395h transcript (Fig. 1c). Conversely, the transcription level of pri-OsamiR395h increased significantly under sulfate deficient conditions (0 and 20 μ M SO 4 2− , Fig. 1c). Transcript levels of pri-and mature OsamiR395 exhibit the same trend under sulfate starvation stress, indicating that OsamiR395 expression is transcriptionally regulated by sulfate. Sulfate starvation stress induces the expression of pri-OsamiR395h, leading to the production of more mature OsamiR395 transcripts.
Computational analysis of the rice genome sequences leads to the identification of four putative targets of OsamiR395, including one ATPS and three sulfate transporter genes, OsaSULTR2;1, OsaSULTR2 and OsaSULTR3;4 ( Fig. 2a) 17,27 . RT-PCR results indicated that OsaATPS did not exhibit any responses in both roots and leaves under − S stress. OsaSULTR3;4 did not respond to sulfate treatment in leaves either, but was down-regulated in roots with the increasing sulfate concentrations, exhibiting similar expression pattern as OsamiR395 (Fig. 2b). OsaSULTR2;1 and OsaSULTR2 genes were both down-regulated in leaves with the increasing sulfate concentrations (Fig. 2b), similar to the expression pattern of OsmiR395 in response to sulfate treatment (Fig. 1). On the contrary, they were both up-regulated in roots in response to increasing sulfate concentrations Scientific RepoRts | 6:28791 | DOI: 10.1038/srep28791 (Fig. 2b). It should be noted that OsaSULTR2 exhibited the highest induction under 20 μ M sulfate, suggesting that other regulation machineries may also participate in the regulation of the OsaSULTR2 gene under this particular condition. These results support the hypothesis that OsaSULTR2;1 and OsaSULTR2 are the putative target genes of, and regulated by OsamiR395 in rice roots. In rice leaves, however, OsamiR395-mediated transcript cleavage of the OsaSULTR2;1 and OsaSULTR2 genes may not be able to take place due to their non-overlapping tissue-specific expression. Instead, there may exist some other mechanisms regulating the expression of OsaSULTR2;1 and OsaSULTR2. This is also likely the case for OsaSULTR3;4 in roots. Similar phenomena have previously been observed in Arabidopsis 18,19 . It should be noted that there are multiple mismatches in the OsamiR395 target sequence of the OsaSULTR3;4 (Fig. 2a). This raises the question of whether or not OsaSULTR3;4 is indeed the true target of OsamiR395.
To confirm the results of semi-quantitative RT-PCR, real-time RT-PCR was conducted to determine the expression levels of OsamiR395 and its putative targets in rice under −S condition (N6 medium without sulfate) and + S condition (regular N6 medium). Real-time PCR results were consistent with that of the semi-quantitative RT-PCR. In both leaves and roots, pri-and mature OsamiR395 were up-regulated under −S condition (Fig. 2c). Among the four putative target genes, only OsaSULTR2;1 and OsaSULTR2 were significantly down-regulated in rice roots under −S condition, exhibiting opposite trend of expression to OsamiR395 (Fig. 2d), in agreement with the results obtained by semi-quantitative RT-PCR and supporting the notion that OsaSULTR2;1 and OsaSULTR2 are the putative targets of OsamiR395 in rice roots.
Expression of the OsamiR395 and its target genes is spatiotemporally regulated. Besides the response of OsamiR395 and its targets to sulfate starvation stress, we also investigated the expression patterns of Total RNA samples were prepared as in (a) and used for stem-loop real-time PCR analysis. OsaSIZ1 was used as a reference gene. Data are presented as means of three technique replicates, error bars represent SD (n = 3). (c) Real-time PCR analysis of rice pri-OsamiR395h under different sulfate concentrations. Total RNA samples were prepared as in (a) and used for real-time PCR analysis. OsaSIZ1 was used as a reference gene. Data are presented as means of three technique replicates, error bars represent SD (n = 3). The statistically significant difference between groups was determined by one-way ANOVA (F(df between , df within ) = F ration, p = p-value, where df = degrees of freedom). Means not sharing the same letter are statistically significantly different (P < 0.05).
Scientific RepoRts | 6:28791 | DOI: 10.1038/srep28791 OsamiR395 and its target genes in different developmental stages and tissues. To this end, we particularly focused on the primary miRNA level for one of the rice OsamiR395 genes, OsamiR395h and the expression of its putative target genes in both roots and leaves at different developmental stages under normal growth conditions. The RT-PCR results showed that the expression of pri-OsamiR395h was strongly induced only in the roots of the four weeks old plants, but otherwise remained very low in both roots and leaves in any other developmental stages (Fig. 3).
The expression of the ATPS again was quite stable in both tissues throughout the rice development, but an elevated expression level in roots was observed compared to that in leaves (Fig. 3). The expression levels of the three sulfate transporter genes were variable, but none of them was inversely correlated with that of the OsamiR395h (Fig. 3).
Heterologous expression of pri-OsamiR395h in Nicotiana tabacum. To further study the role OsamiR395 plays in sulfate transportation and distribution, we generated a chimeric DNA construct containing RT-PCR analysis of expression levels of the OsamiR395 putative targets. Total RNA samples used for RT-PCR were extracted from leaf and root tissues of two weeks old rice grown in N6 medium with 0, 20, 1500 or 2000 μ M (NH 4+ ) 2 SO 4 and used for RT-PCR analysis. OsaSIZ1 was used as a reference gene. Experiment was repeated three times. (c) Stem-loop real-time RT-PCR analysis of mature OsamiR395 and real-time RT-PCR analysis of pri-OsamiR395h. Total RNA samples were prepared from leaf and root tissues of two weeks old rice grown in regular N6 medium (+ S) or N6 medium without SO 4 + (− S) and used for RT-PCR analysis. OsaSIZ1 was used as a reference gene. (d) Real-time RT-PCR analysis was also conducted to determine the expression levels of the OsamiR395 putative targets in rice leaves and roots. Total RNA samples were prepared as in (c) and used for real-time RT-PCR analysis. OsaSIZ1 was used as a reference gene. For (c,d), data are presented as means of two independent biological replicates and three technical replicates, error bars represent SD (n = 6). Asterisks indicate the significant differences between expression levels under − S and + S conditions. P < 0.05 is marked as*, P < 0.01 is marked as* * .
Scientific RepoRts | 6:28791 | DOI: 10.1038/srep28791 the pri-OsamiR395h sequence driven by the CaMV35S promoter (Fig. 4a). This construct was then introduced into tobacco (Nicotiana tabacum) to produce a total of 10 independent transgenic events. RT-PCR analysis suggested rice pri-OsamiR395h was successfully expressed in tobacco (Fig. 4b), and small RNA northern blotting result suggested rice pri-OsamiR395h was successfully processed into mature miRNA (Fig. 4c). The detection of tobacco endogenous mature NtamiR395 in northern blotting indicated that mature NtamiR395 shares a highly conserved sequence with its rice homolog. Three independent transgenic events were selected for further analysis.  Overexpression of the rice pri-OsamiR395h impairs sulfate homeostasis and leads to retarded plant growth in transgenic tobacco. It has previously been shown that overexpression of AthmiR395 in Arabidopsis impairs its sulfate distribution and assimilation 19 . To evaluate the impact of the OsamiR395 in tobacco sulfate metabolism and plant development, we first measured the total sulfur contents in transgenic tobacco plants and wild type (WT) controls. Not surprisingly, the total leaf sulfur content of all the transgenic lines was 2.16 to 2.50 times higher than that in WT controls. On the contrary, the root sulfur content in transgenic lines was 32% to 42% less than that in WT controls (Fig. 5a).
Next, we determined the sulfate-S (sulfate-sulfur) concentration in WT and transgenic plants. Again, the difference in sulfate-S concentrations between transgenics and WT controls was similar to that of the total sulfur contents. In transgenic leaf tissues, the sulfate-S concentration was 1.35 to 1.96 times higher than that in WT leaves, whereas in roots, transgenics had 38% to 57% less sulfate than WT controls (Fig. 5b). This result indicated that the high-level of miR395 accumulation in transgenic plants impacts the uptake and transportation of sulfur and sulfate.
Similar to a previous report in Arabidopsis that overexpression of AthmiR395 represses the expression of sulfate transporter gene AthSULTR2;1 and causes impaired sulfate distributions between leaves of different ages 19 , we also observed that leaf sulfate distribution patterns are different between transgenic tobacco plants and WT controls (Fig. 5c). Because sulfate or sulfur compounds could be transported from old to young leaves under normal or sulfate-adequate conditions 28 , sulfate accumulation in young leaves should be higher than that in old ones as observed in WT control plants (Fig. 5c). Contrary to this, transgenic tobacco plants accumulate fewer sulfates in younger leaves than in older ones (Fig. 5c), indicating that sulfate delivery pathway is impaired in transgenics, which is most likely one of the consequences caused by repressed expression of sulfate transporter genes. Furthermore, compared with WT controls, transgenic tobacco exhibited retarded growth (Fig. 6a,d). As shown in Fig. 6b,c, one-month-old transgenic plants displayed shorter root and less fresh weight than wild type controls, a similar phenotype observed in transgenic Arabidopsis overexpressing AthmiR395 19 . The slow-growth phenotype of transgenic plants suggests that the expression of ATPS may also have been strongly repressed in transgenics, Identification of miR395 target gene in tobacco. To understand how the excess miR395 impacts tobacco sulfate homeostasis at the molecular level, we sought to identify putative new target genes of miR395 using two approaches 29 . We first used the DNA sequences of the Arabidopsis SULTR2;1 and ATPS genes to blastn against the Nicotiana tabacum EST sequences. All the DNA sequences with high similarity (identity of more than 70%) were used to do alignment with complementary sequence of the mature OsamiR395h. The following criteria were used to determine the predicted target sequences with minor modifications: (1) No more than four mismatches between OsamiR395h and its predicted target genes; (2) No more than two constitutive mismatches between OsamiR395h and its predicted target genes; (3) No mismatches between position 10 and 11; (4) No gaps between OsamiR395h and its predicted target genes 29 . Besides, we also designed primers based on the AthmiR395 target genes (AthSULTR2;1 and AthATPS1, 3,4) to amplify and identify the putative homologous genes in tobacco.
Using these approaches, we identified a novel gene named NtaSUTLR2 to be a putative target of OsamiR395h (Fig. 7). Semi-quantitative RT-PCR analysis revealed that NtaSULTR2 was significantly down-regulated in transgenic tobacco (Fig. 7a). We cloned the full-length cDNA sequence of NtaSULTR2 using RACE (Rapid Amplification of cDNA Ends) method, and identified the target site of miR395 that is located between 135 bp and 156 bp of its coding region. There are four mismatches and three mismatches between NtaSULTR2 target sequence and mature OsamiR395 and NtamiR395, separately (Fig. 7b), indicating that NtaSUTLR2 should be efficiently regulated by miR395 because of their near perfect complementary sequence.
We further characterized NtaSULTR2 by generating a phylogenetic tree using protein sequence of NtaSULTR2 and other sixteen well-studied sulfate transporters from rice and Arabidopsis using MEGA6. In this phylogenetic tree, NtaSULTR2 protein is classified into the second group of sulfate transporter subfamily together with AthSULTR2;1, AthSULTR2;2 and OsaSULTR2;1 proteins (Fig. 7c). The three sulfate transporters from Arabidopsis and rice are low-affinity sulfate transporters and involved in the inter-organ delivery of sulfate in vascular to transport sulfate from root to leaf, and distribution of sulfate between leaves 4,7,8 .
Taken together, our results indicate that overexpression of OsamiR395h in tobacco represses sulfate transporter NtaSULTR2, which may play an important role in sulfate transportation and distribution, thus interrupting sulfate homeostasis and distribution in transgenics.
Sulfate regulates tobacco NtamiR395 and NtaSULTR2. To confirm that NtaSULTR2 is the target of miR395 in tobacco, we investigated the expression level of both NtaSULTR2 and mature NtamiR395 under different sulfate concentrations.
In leaf tissues, the transcription of the mature NtamiR395 was gradually up-regulated, contrary to the gradually reduced sulfate concentration. However, NtaSULTR2 did not exhibit an opposite, but a similar expression pattern to NtamiR395 with its lowest transcript level being under 1500 μ M, but not 2000 μ M (NH 4+ ) 2 SO 4 (Fig. 8a).
In root tissues, the situation was different. The transcript level of the mature NtamiR395 increased in response to sulfate depletion, similar to that observed in leaves, whereas NtaSULTR2 exhibited a roughly opposite, but more complex expression pattern (Fig. 8b). Compared to sulfate depletion conditions with 0 μ M (NH 4+ ) 2   that the spatial expression pattern of AthmiR395 does not overlap with that of AthSULTR2;1 18,19,30 , which could probably also explain the similar observation in tobacco from this study.
MiR395 mediates the cleavage of NtaSULTR2 mRNA. To further confirm that NtaSULTR2 is the true target of miR395, we conducted RLM-RACE (T4 RNA Ligase Mediated Rapid Amplification of cDNA Ends) to verify that NtaSULTR2 transcripts are cleaved by miR395. We used RNA from the miR395-overexpressing transgenic tobacco plants to facilitate the detection of cleaved NtaSULTR2 mRNA.
We used the forward primer ASP (Adapter Specific Primer) and the reverse primer GSP (Gene Specific Primer) to conduct the first round PCR after the adapter-linked first strand cDNA ends were generated. The RNA adapter has a length of 44 bp, and the reverse GSP is localized 545 bp downstream of the predicted miR395 target site in the NtaSULTR2 mRNA, so the product of the first round PCR should have a length of about 589 bp. As shown in Fig. 9, the first round PCR with transgenic tobacco cDNA indeed generated a clear band of about 600 bp.
A second round PCR was then conducted using the first round PCR product as template and a new set of primers to confirm the authenticity of the PCR product. The forward primer NASP (Nest Adapter Specific Primer) is localized on the adapter from 14 bp to 44 bp, and the reverse primer NGSP (Nest Gene Specific Primer) is localized 463 bp downstream of the predicted miR395 target site in the NtaSULTR2 mRNA, so the product of the second round PCR should be about 493 bp. As shown in Fig. 9, the second round PCR indeed generated a clear main band of about 500 bp as expected. Cloning and sequencing of the PCR product further confirmed the predicted miR395 cleavage site in the NtaSULTR2 mRNA.

Discussion
Previous studies on Arabidopsis miR395 have indicated its involvement in sulfate starvation response by repressing the expression of genes in sulfate transportation and assimilation pathways.
Under − S condition, the accumulation of AthmiR395 is enhanced under low internal sulfate levels, and correlated to GSH pool, indicating that the regulation of AthmiR395 is mediated by internal sulfate level and redox NtaL25 was used as a reference gene. Data are presented as means of three technical replicates and two biological replicates, error bars represent SD (n = 6). The statistically significant difference between groups was determined by one-way ANOVA (F(df between , df within ) = F ration, p = p-value, where df = degrees of freedom). Means not sharing the same letter are statistically significantly different (P < 0.05). signaling in Arabidopsis 22,31 . The increased AthmiR395 then represses the expression of AthATPS1, AthATPS3, AthATPS4 and AthSULTR2;1 18,22 . Further study in Arabidopsis revealed a whole picture of how AthmiR395 is involved in plant response to sulfate starvation. When sulfate supply is limited, the induced AthmiR395 mediates the degradation of ATPS mRNA leading to the accumulation of sulfate in leaf tissues as a result of decelerated sulfate assimilation 19 . At the same time, the cleavage of AthSULTR2;1 mRNA in shoots by AthmiR395 results in blocked sulfate transport into new leaves from old ones 19 . Furthermore, the impaired sulfate homeostasis and reduced sulfate assimilation impact seed germination under ABA-treated condition 32 .
MiR395 is highly conserved across species, which strongly suggests that its function in regulating plant response to nutrition, particularly sulfate supply could also be conserved during evolution. Our results in rice indicate that indeed, the transcript of mature OsamiR395 increases under − S condition, and this change in expression might be regulated at the transcription level (Fig. 1). Computational prediction led to the identification of four putative target genes of OsamiR395 in rice. We confirmed that OsaSULTR2;1 and OsaSULTR2 are regulated by OsamiR395 in roots suggesting that they may be the OsamiR395 target genes.
Knowledge about the functions of rice sulfate transporters is limited. Phylogenetic analysis grouped the fourteen rice sulfate transporters together with their Arabidopsis counterparts 11 , suggesting that they may share similar function. OsaSULTR2;1 and OsaSULTR2 may be responsible for the root-to-shoot sulfate transportation and distribution of sulfate between leaves of different ages. Our results (Fig. 2b-d) showed that the expression patterns of rice sulfate transporter genes were different from their Arabidopsis homologs, both OsaSULTR2;1 and OsaSULTR2 were reduced in leaves with the increasing sulfate concentrations. We speculate that the two sulfate transporter genes and miR395 may be differentially expressed in different leaf tissues and thus, OsaSULTR2;1 and OsaSULTR2 may not be subjected to miR395 regulation. Instead, other regulatory machineries may participate in the control of their expression in response to sulfate levels. It is likely that when rice plants are subjected to sulfate starvation, there is a need for the two sulfate transporters to be active, driving the transportation of sulfate from old leaves to younger ones to ensure plant growth and development. However, with abundant sulfate supply in the environment, there is no need for sulfate distribution to young leaves, and therefore the expression of both OsaSULTR2;1 and OsaSULTR2 declines.
The miRNA-mediated gene regulation mechanism emerged about 425 million years ago, which is at a very early stage of plant phylogeny prior to the divergence of monocot and dicot plants 33 . This suggests that monocot Figure 9. Confirmation of miR395-mediated cleavage of NtSULTR2 mRNA. RLM-RACE (T4-RNA ligase mediated amplification of 5′ cDNA ends) was conducted to confirm the cleavage of NtSULTR2 mRNA. Total RNA samples were isolated from two weeks old transgenic tobacco. 44 bp RNA adapter was ligated to the purified RNA by using T4 RNA ligase. Adapter-linked RNA was then used to synthesize first strand cDNA, followed by amplification of 5′ ends using the forward primer ASP and the reverse primer GSP. The 589 bp product from the first round PCR was then used as template for the second round PCR using the forward nest primer NASP and the reverse nest primer NGSP, producing a 493 bp second round PCR product. M: DNA molecular weight marker. OE: overexpression line. Red lines indicate miR395 cutting site. and dicot plants should have a similar miRNA-mediated gene regulation mechanism, and some highly conserved miRNA families regulating the same biological process have evolved from the same gene ancestors. Indeed, research data in the past twenty years indicate that 21 miRNA families, such as miR156 and miR399, are conserved in sequence across monocots and dicots 25 . More specifically, Zhang et al. found that 9 miRNA families are highly conserved 33 , 10 miRNA families are moderately conserved and 16 miRNA families including miR395 are lowly conserved across plant species. In a later work, miR395 family was identified in the common ancestor of all embryophytes 25 . Besides the miRNA sequences, the genes involved in miRNA and siRNA biogenesis pathways are also conserved across species. In plants, Dicer-like (DCL) is a key protein in the miRNA genesis pathway. DCL interacting with HYPONASTIC LEAVES1 (HYL1) and C2H2-zinc finger protein SERRATE (SE) in D-bodies cleaves the pri-miRNA from the base to yield a pre-miRNA with stem-loop structure, and this pre-miRNA is sliced again to yield mature miRNA [34][35][36][37] . Phylogenetic analysis indicated that divergence of DCL1 gene associated with miRNA production from other DCLs could be traced to the time before the emergence of moss Physcomitrella patens 36 , indicating that DCLs may have the same origin and are conserved across vascular plants.
Based on previous findings, we hypothesize that miRNA biogenesis pathway in dicots could accept pri-miRNAs from monocots, and process it into mature miRNA with function. To verify our hypothesis, the full-length DNA sequence of pri-OsamiR395h was cloned from rice genome. The expression cassette of the CaMV35S-controlled rice pri-OsamiR395h was then prepared and introduced into tobacco genome. By performing small molecule northern blotting, we observed high transcript level of miR395 in transgenic tobacco under normal condition, indicating that rice pri-OsamiR395h could be successfully expressed and processed into mature miR395h in tobacco (Fig. 4). At the same time, we also observed low level of endogenous mature miR395 in WT tobacco, confirming that tobacco mature miR395 is highly conserved with its rice homolog. All of the three transgenic tobacco lines exhibited impaired sulfate homeostasis and distribution (Fig. 5). Furthermore, transgenic plant had retarded growth phonotype (Fig. 6). All the facts suggest that mature OsamiR395 functions in transgenic tobacco.
Data obtained from this research revealed that the sulfate-S contents in transgenic tobacco are higher in leaf tissue, but lower in root tissue than those in WT controls. An even more significant difference in total sulfur content was observed between WT controls and OsamiR395h overexpression plants (Fig. 5a,b). Besides, we also observed that sulfate distribution between leaves of different ages is impaired in transgenic tobacco plants (Fig. 5c).
To reveal the molecular mechanism underlying miR395-mediated plant sulfate metabolism, we studied genes impacted by excessive dose of miR395 in transgenic tobacco, and identified a novel sulfate transporter gene NtaSULTR2 belonging to the second group of sulfate transporter genes (Fig. 7). Based on the results of real-time PCR and RML-RACE, we verified that NtaSULTR2 is the target gene of miR395 (Figs 8 and 9). We believe that the repression of NtaSULTR2 gene in transgenic tobacco plants partially impaired the sulfate homeostasis. In Arabidopsis shoot tissue, sulfate transporter AthSULTR2;1 is localized in both xylem and phloem, particularly in phloem parenchyma cells surrounding sieve and companion cells, and involved in distribution of sulfur between leaves of different ages 7,28 . We conjecture that in tobacco shoot tissue, NtaSULTR2, likes its homologs in Arabidopsis, retrieves sulfate from mesophyll cells to xylem and phloem cells, and sulfate is transported from old leaves to young leaves. But in transgenic plants, the delivery of sulfate from old leaves to young leaves is impaired because of significantly repressed NtaSULTR2 gene (Fig. 5c).
Although no ATPS gene have been identified and cloned in tobacco, we believe that there must be one or more ATPS gene(s) repressed in transgenic tobacco, causing interrupted sulfate assimilation. The interruption of the sulfate assimilation pathway would cause a shortage in cysteine and other sulfate metabolic products, resulting in retarded plant growth and triggering plant sulfate starvation signaling, which would promote sulfate absorption and transport into leaf tissue, and consequently a much more sulfur accumulation in leaves of transgenics than in that of WT controls (Fig. 5a,b).

Materials and Methods
Plant materials and growth conditions. To investigate the expression levels of OsamiR395 and its targets in rice under different sulfate concentrations, rice seeds were surface sterilized and grown in N6 medium under 16h light/8h dark at 28 °C 38 . Sulfate salts of the N6 medium were replaced with chloride salts and supplemented with 0, 20, 1500 or 2000 μ M (NH 4+ ) 2 SO 4 . Sterilized rice seeds were also grown in regular N6 medium (+ S) and N6 medium without SO 4 + (− S) under 16 h light/8 h dark at 28 °C. Two weeks old plants were harvested for RNA isolation.
To investigate the expression patterns of OsamiR395 and its targets in different developmental stages and tissues of rice, rice seeds were grown in soil in a greenhouse. Root and leaf samples were collected two, four and eight weeks after germination.
To investigate the expression levels of pri-OsamiR395h, mature miR395 and NtaSULTR2 in tobacco, tobacco seeds were surface sterilized and grown in MS medium under 16h light/8h dark at 22 °C 39 . To prepare MS mediums with different sulfate concentrations, sulfate salts of the MS medium were replaced with chloride salts and supplemented with 0, 20, 1500 or 2000 μ M (NH 4+ ) 2 SO 4 . Two weeks old and four weeks old plants were harvested for RNA isolation.
To measure total sulfate content and sulfate-S concentration in tobacco, and to determine the growth rate of tobacco, tobacco were grown in soil in a greenhouse. Four weeks old and 12 weeks old plants were collected for analysis.
Genomic DNA and total RNA isolation, and cDNA synthesis. Plant genomic DNA was isolated following previously described method 40 .
Scientific RepoRts | 6:28791 | DOI: 10.1038/srep28791 Total RNA was isolated from 100 mg plant samples with Trizol reagent (Ambion, USA), and the genomic DNA is removed by using RNase-free DNase I (Invitrogen, USA). 2 μ g total RNA was used to synthesize first strand cDNA with SuperScript III Reverse Transcriptase (Invitrogen, USA) according to manufacturer's instructions. The first strand cDNA was used for semi quantitative RT-PCR and regular real-time PCR.
To determine the transcript level of mature miR395, the first-strand cDNA used for stem-loop real-time PCR was synthesized following the regular SuperScript III Reverse Transcriptase (Invitrogen, USA) mediated method, except that the oligo (dT) 20 was replaced with miR395 specific reverse transcription primer. Primers were all listed in Supplementary Table S1.
Semi-quantitative RT-PCR, stem-loop and regular real-time PCR. To conduct semi-quantitative RT-PCR, first-strand cDNA samples were diluted to 0.25 times based on the concentration of the first-strand cDNA samples. The loading volume of the cDNA samples was adjusted basing on the transcript level of a reference gene.
To conduct stem-loop and regular real-time PCR, first-strand cDNA samples were diluted to 0.025 to 0.005 times based on the concentration of the first-strand cDNA samples. Both stem-loop and regular real-time PCR were performed using SYBR Green Supermix (Bio-Rad, USA) following manufacturer's instructions, and iQ5 real-time detection system (Bio-Rad USA) was used to detect and analyze the real-time PCR result.
Stem-loop and regular real-time PCR results were determined by using Δ Δ Ct method. Δ Ct was defined as Ct test − Ct 0h , in which Ct test stands for threshold cycle of one gene after treatment, and Ct 0h stands for threshold cycle of one gene before treatment. Δ Δ Ct was defined as Δ Ct reference − Δ Ct target , in which Δ Ct reference stands for Δ Ct of the endogenous gene used as a reference, and Δ Ct target stands for Δ Ct of target gene. Finally, related expression ratio was calculated as 2 ΔΔCt .
Primers used for semi-quantitative RT-PCR, stem-loop real-time PCR and regular real-time PCR were all listed in Supplementary Table 1.
Small molecule Northern blotting. Small molecule northern blotting was performed following the method previously described with minor modification 41 . 10 μ g total RNA denatured at 95 °C was separated in 12.5% urea-polyacrylamide gel and transferred to Hybond-N+ nylon membrane (Amersham, USA) in a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad, USA). To prepare radiolabeled probe for detecting mature miR395, DNA oligonucleotide GAGTTCCCCCAAACACTTCAC was synthesized (http://www.idtdna.com/site) and labeled with γ -[32P]-ATP by using T4 polynucleotide kinase. RNA membrane was then hybridized with radiolabeled probe and detected on a phosphorimaging screen.

Plasmid construction, bacterial strains and plant transformation. The predicted pri-OsamiR395h
was amplified from rice genomic DNA and cloned at downstream of CaMV35S (Cauliflower Mosaic Virus 35S) promoter of binary vector pZH01, resulting in CaMV35S/OsamiR395h-CaMV35S/hygromycin 42 . This chimeric gene expression construct was then mobilized into Agrobacterium tumefaciens strain LBA4404 by electroporation for tobacco transformation. The Escherichia coli strain used in this experiment was DH5α .
The primers used for plasmid construction were all listed in Supplementary Table 1.
Determination of total sulfur content and sulfate-sulfur concentration. For determination of total sulfur, plant samples were collected and dried for 48 h at 80 °C. Total sulfur contents in dry samples were determined as previously described 43 sulfate-S concentration was determined following a previous method with minor modification 44 . 10 mg dry plant sample or 200 mg fresh plant sample was immersed in 1 ml 0.1 M HCl for 2 h at room temperature, followed by 20 min centrifugation at 12000 g. Clear supernatant liquid was then transferred to a 50 ml Erlenmeyer flask and made to 20 ml by water. One ml of barium chloride-gelatin reagent was added to the liquid. After 40 min (no more than 120 min), absorbance of the resulting cloudy liquid was determined at 454 nm by using a spectrometer.
Rapid amplification of cDNA ends. To obtain 5′ cDNA end and 3′ cDNA end of NtaSULTR2, total RNA was extracted from 100 mg two weeks old WT tobacco with Trizol reagent (Ambion, USA) and treated with RNase-free DNase I (Invitrogen, USA) to remove genomic DNA. 1 μ g total RNA was then used to amplify 5′ end and 3′ end cDNA of NtaSULTR2 with SMARTer RACE 5′ /3′ commercial kit (Clontech, USA) following the manufacture's instruction. Then, the 5′ end and 3′ end cDNA fragments were sequenced. Sequence information was used to design primers for cloning of full-length NtaSULTR2 cDNA. The primers used for RACE and for cloning of full length NtaSULTR2 cDNA were all listed in Supplementary  Table S1. T4-RNA ligase mediated amplification of 5′ cDNA ends. To verify miR395 cleavage site within NtaSULTR2, T4-RNA ligase mediated amplification of 5′ cDNA ends was conducted following a previously described method 45 . Briefly, total RNA was isolated from 100 mg plant sample using Trizol reagent (Ambion, USA), followed by purification of RNA with RNeasy mini kit (Qiagen, Germany). RNA adapter was ligated to the purified RNA by using T4 RNA ligase (New England Biolabs, USA). Based on the fact that miRNA-mediated mRNA cleavage will generate 5′ -monophosphate ends on the 3′ end cleavage product of the target mRNAs, it is possible to ligate RNA oligonucleotide adapter to the 5′ terminus of the 3′ end cleavage product by using T4 RNA ligase, whereas such RNA oligonucleotide adapter would not be ligated to mRNAs with conventional 5′ cap 45 . Adapter-linked RNA was then used to synthesize first strand cDNA with SuperScript II Reverse Transcriptase (Invitrogen, USA), followed by amplification of 5′ ends using the forward primer ASP and the reverse primer GSP. The product from the first round PCR was then used as template for the second round PCR with the forward nest primer NASP and the reverse nest primer NGSP. PCR product was cloned for sequencing.
The primer sequences used for RML-RACE were all listed in Supplementary Table 1.
Scientific RepoRts | 6:28791 | DOI: 10.1038/srep28791 Phylogenetic analysis of sulfate transporters. Phylogenetic tree of NtaSULTR2 and other sulfate transporter genes in rice and Arabidopsis inferred using the Neighbor-Joining method 46 . The optimal tree with the sum of branch length = 3.89795523 is shown. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Poisson correction method 47 and are in the units of the number of amino acid substitutions per site. The analysis involved 17 amino acid sequences. All positions containing gaps and missing data were eliminated. There was a total of 347 positions in the final dataset. Evolutionary analyses were conducted in MEGA6 48 . WT: wild type plant. OE: overexpression line.
Statistical analysis. Student's t test was used to test the difference between the means from two groups. P < 0.05 was considered to be statistically significant and marked as *P < 0.01 was considered to be statistically highly significant and marked as* * . One-way ANOVA (F(df between , df within ) = F ration, p = p-value, where df = degrees of freedom) with post hoc comparisons using the Tukey HSD test was used to determine the statistically significant difference between the means from three or more groups. Means not sharing the same letter are statistically significantly different (P < 0.05).