SQUAMOSA promoter binding protein-like7 regulated microRNA408 is required for vegetative development in Arabidopsis

Authors


For correspondence (e-mail ll4jn@virginia.edu).

Summary

MicroRNAs (miRNAs) are endogenous small RNAs repressing target gene expression post-transcriptionally and are critically involved in various development processes and responses to environmental stresses. MiR408 is highly conserved in land plants and targets several transcripts encoding copper proteins. Although it has been well documented that expression level of miR408 is strongly influenced by a variety of environmental conditions including copper availability, the biological function of this miRNA is still unknown. Here we show that constitutive expression of miR408 results in enhanced growth of seedling and adult plant while knocking down miR408 level by T-DNA insertions or the artificial miRNA technique causes impaired growth. Further, we found that constitutively activated miR408 is able to complement the growth defects of the T-DNA lines. Regarding the molecular mechanism governing miR408 expression, we found that the transcription factors SQUAMOSA PROMOTER BINDING PROTEIN-LIKE7 (SPL7) binds to the GTAC motifs in the MIR408 promoter in response to copper deficiency. Interestingly, constitutive activation of miR408 in the spl7 background could partially rescue the severe growth defects of the mutant. Together these results demonstrate that miR408 is a powerful modulator of vegetative growth. Our finding thus reveals a novel control mechanism for vegetative development based on calculated miR408 expression in response to environmental cues.

Introduction

Adaptation to the environment occurs in all living organisms. Consequently, sophisticated gene networks have evolved to modify development and adjust metabolism in response to changing environmental conditions. We now appreciate that non-coding regulatory RNAs are important constituents of these networks. Among these are the 20–24 nucleotides endogenously encoded miRNAs that are derived from stem loop-structured precursors called pre-miRNAs (Bartel, 2004; Voinnet, 2009). Following the initial discovery in Caenorhabditis elegans (Lee et al., 1993; Wightman et al., 1993), miRNAs are recognized as an important class of regulatory RNAs that modulate gene expression at the post-transcription level in eukaryotes (Bartel, 2004; Voinnet, 2009). After integration into the RNA-induced silencing complex, miRNAs base-pair with complementary sequences within the target genes to direct cleavage (Llave et al., 2002) or translational repression (Brodersen et al., 2008) of the target transcripts.

As gene regulators, miRNAs share several features with transcription factors (Chen and Rajewsky, 2007; Hobert, 2008). For example, miRNAs act in trans and recognize their targets through short sequence motifs. A single miRNA can regulate multiple target genes, and multiple miRNAs may act in combination to regulate the same targets (Chen and Rajewsky, 2007). This, together with the large number of miRNA genes, indicates that miRNAs impact a substantial portion of the transcriptome post-transcriptionally. In plants, hundreds of miRNA target pairs have been predicted, based on high sequence complementarity and dozens characterized in some detail (Jones-Rhoades and Bartel, 2004; Schwab et al., 2005; Fahlgren et al., 2007; Zhang et al., 2007; Alves et al., 2009). It is now well established that many of these gene circuits are critically involved in various developmental processes and responses to environmental stresses (Jones-Rhoades et al., 2006; Garcia and Frampton, 2008; Voinnet, 2009). Equally important to our understanding of the miRNA-mediated biological processes is the expression of MIR genes, which is crucial for temporal and spatial control of the abundance of individual miRNAs. In contrast to the numerous reports that attest to the importance of miRNAs on target gene expression, our knowledge on the regulation of miRNAs is still relatively sparse.

MiR408 has so far been annotated in more than 20 plant species. This places miR408 among the most conserved miRNA families in land plants (Axtell and Bowman, 2008; Kozomara and Griffiths-Jones, 2011), indicating that its function is fundamental to plants. The abundance of miR408 is responsive to copper supply in the environment. It was repeatedly observed in Arabidopsis that the miR408 level is low under standard growth conditions, but increases markedly upon copper starvation (Yamasaki et al., 2007, 2009; Abdel-Ghany and Pilon, 2008). Furthermore, the sequence of miR408 exhibits strong complementarity, with transcripts encoding copper-containing proteins that include three members of the Laccase family (LAC3, LAC12, and LAC13) and Plantacyanin. Cleavage of the four target transcripts mediated by miR408 was reported by Abdel-Ghany and Pilon (2008). These findings clearly support the participation of miR408 in copper homeostasis (Yamasaki et al., 2007; Abdel-Ghany and Pilon, 2008).

It should be noted that the miR408 level is also influenced by a variety of other environmental cues. Induced miR408 accumulation in response to mechanical stress (Lu et al., 2005), dehydration (Kantar et al., 2010; Trindade et al., 2010) and reactive oxygen species (Li et al., 2010) have been reported for various plant species. The diverse and sophisticated expression of the MIR408 gene suggests that multiple transcription factors are involved in its regulation. Based on an analysis of miR398, another copper-responsive miRNA, SPL7 has been proposed to regulate MIR408 in response to low copper conditions (Yamasaki et al., 2009). It has been shown that copper deficiency-induced miR408 accumulation depends on a functional SPL7 in Arabidopsis (Yamasaki et al., 2009); however, it has not been demonstrated whether SPL7 directly regulates miR408 expression in response to changing environments. More importantly, it is not yet clear how the expression dynamics of miR408 impacts plant growth and development.

In the current work, we attempt to elucidate the biological role of miR408 in Arabidopsis through a series of genetic and molecular analyses. We show that the miR408 level is tightly linked to vigor of vegetative growth, with constitutively expressed miR408 resulting in enhanced growth and silenced MIR408 causing impaired growth. Furthermore, we demonstrate that SPL7 regulates miR408 expression through binding to the GTAC motifs in its promoter. Interestingly, constitutive activation of miR408 in the spl7 background could partially rescue the severe growth defects of the mutant. Together, these results reveal that miR408 is a powerful modulator of vegetative growth, and its calculated expression constitutes a control mechanism for vegetative development in response to environmental cues.

Results

MiR408 is expressed throughout the life cycle in Arabidopsis

Using a homology search, we identified miR408 sequences in 51 plant species. Excepting the most 5′ nucleotide, miR408 is deeply conserved in moss, angiosperm and gymnosperm species, suggesting that its function is fundamental to land plants. To begin to elucidate the regulation and function of this conserved miRNA, we first examined the tissue-specific expression profile of miR408 in Arabidopsis. By means of northern blotting and quantitative reverse transcription-coupled PCR (qRT-PCR), we found that mature miR408 is readily detectable in the seven organ types examined, including seedling, root, rosette and cauline leaves, stem, flower and silique (Figure 1a,b). These results indicate that MIR408 is constitutively expressed under standard growth conditions, and therefore has the potential to act throughout development.

Figure 1.

Expression pattern of miR408 in development. (a) RNA blot analysis of mature miR408 accumulation in seedlings and different adult tissues. U6 snRNA was used as a loading control. (b) qRT-PCR analysis of the expression of miR408 and SPL7 in seedlings and different adult tissues. 5S rRNA and Actin7 were used as the endogenous control to normalize the relative expression levels of miR408 and SPL7, respectively. Error bars present SDs (n = 3). (c) The promoter region of MIR408 was fused to the GUS reporter gene to generate the pMIR408:GUS transgenic plants in the wild-type background. Tissues at different development stages were stained for GUS activity and visualized: 1, seedling; 2, rosette leaf; 3, cauline leaf; 4, stem; 5, inflorescence; 6, silique.

MiR408 is encoded by a single gene in Arabidopsis (AT2G47015). Using the annotated pre-miRNA sequence, we identified a match in the available full-length cDNA collection (AF419552). Alignment of the full-length cDNA sequence to the genome allowed us to pinpoint the transcription start site (TSS) and the proximal promoter region. We generated a reporter construct in which the β–glucuronidase (GUS) gene is fused with the native promoter region, approximately 800 base pairs upstream of the TSS (pMIR408:GUS). We then generated transgenic Arabidopsis plants expressing this reporter construct and the vector control (Figure S1). When the transgenic plants were assayed, GUS activity indicates that the MIR408 promoter is active in both the green tissues (cotyledon, rosette and cauline leaf, sepal and silique coat) and vascular tissues (hypocotyl, root, vein and stem) (Figure 1c). Furthermore, we found that the GUS staining pattern is in general consistent with the detected transcript abundance (Figures 1 and S1), although the miR408 levels in stem and silique deduced from the three GUS assays and RNA profiling show a slight difference. Together, these results demonstrate that the cloned MIR408 promoter is functional, and that miR408 is expressed throughout the developmental stages.

Overexpression of miR408 promotes vegetative growth

As a first step to dissect the effects of miR408 on development, we analyzed transgenic plants in which the expression of pre-miR408 is driven by the enhanced cauliflower mosaic virus 35S promoter (referred to hereafter as miR408-OX; Zhang et al., 2011). Compared with the wild type, the accumulation of miR408 is much higher in miR408-OX plants (Figures 2a and S2). To test whether over-accumulated miR408 in the miR408-OX plants results in downregulation of the target genes, we examined LAC13 (AT5G07130) and Plantacyanin (AT2G02850), which are two of the four validated miR408 target genes (Abdel-Ghany and Pilon, 2008). Using the modified 5′ RNA-ligase RACE assay that detects miRNA-guided target cleavage (Llave et al., 2002; Kasschau et al., 2003), we confirmed these to be bona fide miR408 targets. Based on qRT-PCR analysis, we found that expression of both LAC13 and Plantacyanin are significantly repressed in the miR408-OX plants (Figure 2b), indicating that constitutively activated miR408 is functional.

Figure 2.

Molecular and phenotypic analyses of transgenic plants overexpressing miR408. (a) RNA blot analysis of miR408 accumulation in wild-type and miR408-OX seedlings. U6 snRNA was used as a loading control. (b) qRT-PCR analysis of miR408 target genes Plantacyanin and LAC13 in wild-type and miR408-OX seedlings. Error bars present SDs (n = 3). (c) Phenotypic comparison of wild type and miR408-OX plants. The upper left panel shows seedlings grown on MS under continuous light. The upper right panel is the top view of adult plants grown in soil under long-day conditions. The lower left panel shows the morphology of the cotyledons and rosette leaves of adult plants. The lower right panel shows cauline leaves of adult plants. Scale bars: 1 cm. (d) Quantitative analysis of fresh weight, root length, and leaf length and width between wild-type and miR408-OX seedlings. Measurements of fresh weight and root length were performed on seedlings. The fifth rosette leaf of adult plants was used as the representative for leaf length and width. Error bars present SDs (n ≥ 20). *P < 0.01 and **P < 0.001 versus the wild type, by Student's t–test.

Compared with the wild type, miR408-OX plants exhibit elevated growth vigor, even at the seedling stage (Figure 2c), resulting in significantly higher fresh weight and a longer root (Figure 2d). In the adult stage, an enhancement of vegetative development was readily noticeable for the transgenic plants (Figure 2c). All leaves, including the cotyledon of the miR408-OX plants, are morphologically larger than those of wild-type plants, with significantly increased leaf length (between blade and petiole) and leaf width (Figure 2c). As a representative, the fifth rosette leaf was used to quantitatively detect the leaf size difference. We found that the enlarged leaves of the miR408-OX plants increased in size in two directions (40% in length and 24% in width), such that the whole rosette leaf area had significantly increased up to a maximum of 162% compared with the wild type. The size of the cauline leaves and petiole length of the transgenic plants were also significantly increased (Figure 2c). In addition, the rosette and cauline leaves of the miR408-OX plants showed a serrated phenotype compared with the wild type, which became more severe for later rosette leaves (Figure 2c). Thus, an over-accumulation of miR408 leads to the downregulation of its target genes and the promotion of vegetative growth.

Knocking down miR408 results in impaired growth

Our next goal is to assess the impact of reduced miR408 expression on vegetative development by two different approaches. Previously, four Arabidopsis T–DNA lines carrying insertions in the MIR408 locus was examined (Maunoury and Vaucheret, 2011). It was reported that seedlings of these four T–DNA insertion lines exhibit no visible developmental defects when grown on medium supplemented or not with copper (Maunoury and Vaucheret, 2011). We therefore identified two more independent insertion lines, FLAG_545E03 and FLAG_481E04, pertinent to MIR408 (Figure S3). In these two lines, especially in FLAG_481E04, miR408 expression is drastically compromised (Figure 3a). We found both mutants show increased Plantacyanin and LAC13 levels compared with the wild type (Figure 3b). Phenotypic analysis of the two insertion lines is shown in Figure 3c. Morphologically, it can be observed that the insertion lines exhibit reduced vegetative growth as soon as just after germination, with FLAG_481E04 displaying more severe growth defects than FLAG_545E03 (Figure 3c). Quantitatively, the insertion lines have weakened performance in all measured parameters for vegetative development, including fresh weight, root length, leaf length, and leaf width, compared with wild type (Figure 3d).

Figure 3.

Characterization of miR408 loss-of-function lines caused by T–DNA insertion. (a) RNA blot analysis of miR408 accumulation in seedlings of wild type and T–DNA insertion lines FLAG_545E03 (E03) and FLAG_481E04 (E04). U6 snRNA was used as the loading control. (b) qRT-PCR analysis of miR408 target genes Plantacyanin and LAC13 in seedlings of wild type as well as the E03 and E04 mutants. Error bars present SDs (n = 3). (c) Morphology of the wild type and the two FLAG mutants. The left panel shows seedlings and the right panel shows adult plants. Scale bars: 1 cm. (d) Quantitative comparison of the wild type and the FLAG mutants. Fresh weight and root lengths were measured at the seedling stage, whereas leaf lengths and widths were measured at the adult stage. Error bars present SDs (n ≥ 15). *P < 0.01 and **P < 0.001 versus wild type, by Student's t–test.

As an additional approach to knock-down miR408 expression, we used the artificial microRNA method (Schwab et al., 2006), which has been demonstrated as an efficient way to silence endogenous miRNA genes (Eamens et al., 2011). We designed an artificial microRNA, amiR408, driven by the 35S promoter to specifically target the pre-miR408 sequence (Figure S4). By northern blotting and qRT-PCR analyses, we detected drastically reduced miR408 expression in two independent lines transformed with amiR408 (Figures 4a and S4). As an indicator for reduced miR408 activity, both LAC13 and Plantacyanin transcripts were found at much increased levels in the amiR408 plants (Figure 3b). Contrary to the miR408-OX plants but similar to the insertion lines, the growth vigor of the amiR408 plants was severely reduced (Figure 4c), resulting in significantly lower fresh weight and shorter roots in the seedling stage, and reduced leaf length and width in the adult stage (Figure 4d). Together, these results demonstrate that miR408 is critically required for proper vegetative development in Arabidopsis.

Figure 4.

Knocking down miR408 expression by artificial microRNA impairs growth. (a) RNA blot analysis of miR408 accumulation in the seedlings of the wild type and a selected amiR408 transgenic line. U6 snRNA was used as the loading control. (b) qRT-PCR analysis of miR408 target genes Plantacyanin and LAC13 in the wild type and amiR408 transgenic seedlings. Error bars present SDs (n = 3). (c) Phenotypic comparison of wild type and amiR408 transgenic plants. The left panel shows seedlings and the right panel shows adult plants. Scale bars: 1 cm. (d) Quantitative comparison of fresh weight, root length, leaf length and width between wild-type and amiR408 plants. Error bars present SDs (n ≥ 15). *P < 0.01 and **P < 0.001 versus wild type, by Student's t–test.

Considering that At2 g47020, which encodes a peptide chain release factor, overlaps with the miR408 locus, we examined whether this gene is impacted when miR408 is silenced. As shown in Figure S3, using the primer pair flanking the T–DNA insertion sites, we found that the transcript level of At2 g47020 is reduced in three of the T–DNA lines, but increased in the other three when compared with the wild type. Using the primer pair downstream of all insertion sites, we found that the At2 g47020 level is increased in all T–DNA lines except FLAG_545E03. Thus, the T–DNA lines disrupting miR408 does affect the At2 g47020 locus; however, three critical pieces of evidence support the conclusion that the growth defects when knocking down miR408 are independent of the At2 g47020 gene. First, the homozygous T–DNA mutant with At2 g47020 specifically knocked out (SALK_049729 in Figure S3) showed a growth phenotype similar to that of the wild type (Figure S3); therefore, the reduced At2 g47020 transcript level in the two FLAG lines does not contribute to the observed growth defects (Figure 3). Second, we checked four individual amiR408 lines and found comparable expression levels of At2 g47020 with the wild type (Figure S4); thus, the amiR408 lines employed in this study are specific to miR408, and do not downregulate At2 g47020. Finally, we performed a complementation experiment by overexpressing miR408 in the two FLAG lines. As expected, both mutants recovered to normal growth compared with the wild-type plants (Figure S5). Taken together, our results unambiguously demonstrate that the loss of function of miR408 specifically reduces growth.

Transcription of miR408 is induced by copper deficiency in seedling

To gain mechanistic insight into the function of miR408, we studied its regulation in the seedling stage. Searching the proximal promoter regions (from −1 to −600 upstream the TSS) revealed an array of 10 putative GTAC motifs. The tetranucleotide GTAC was found to be the core sequence of copper-response elements in Chlamydomonas, Arabidopsis and Barbula unguiculata (Quinn and Merchant, 1995; Quinn et al., 1999; Kropat et al., 2005; Nagae et al., 2008; Yamasaki et al., 2009). Additionally, the abundance of mature miR408 was found to increase upon copper starvation (Abdel-Ghany and Pilon, 2008; Yamasaki et al., 2009). These observations prompted us to examine whether the induction of miR408 expression by low copper levels occurs at the transcription level in Arabidopsis seedlings. Northern blot (Figure 5a) and qRT-PCR (Figure 5b) analyses confirmed that miR408 is present in young seedlings grown in standard MS media (with a copper concentration of 0.1 μm). Consistently, supplementing MS with 5 μm copper (MS + Cu) effectively prevents the seedlings from accumulating miR408 (Figure 5a,b).

Similar to mature miR408, qRT-PCR analysis showed that pre-miR408 is accumulated to a much higher level under low copper conditions (Figure 5c). To further test whether miR408 accumulation is controlled at the transcription level, we analyzed the pMIR408:GUS transgenic plants in the wild-type background. When grown with copper supplement, GUS staining was visible in cotyledon veins, hypocotyl and the upper part of the root (Figure 5d). By contrast, strong GUS activity throughout the seedling was observed when the seedlings were grown on MS (Figure 5d), indicating that the MIR408 promoter is induced in low copper conditions. Together, our results demonstrate that transcriptional control is important for proper MIR408 expression in response to varying copper levels.

Figure 5.

MIR408 is regulated by SPL7 at the transcription level. (a) RNA blot analysis of miR408 levels in wild-type and spl7 seedlings grown on MS or MS supplemented with 5 μm CuSO4 (MS + Cu). U6 snRNA was used as the loading control. (b, c) qRT-PCR analyses of miR408 and pre-miR408 in wild-type and spl7 seedlings grown under different copper conditions. 5S rRNA and Actin7 were used as the endogenous control, respectively. Error bars present SDs (n = 3). (d) GUS staining of transgenic seedlings expressing the pMIR408:GUS reporter gene in the wild type (pMIR408:GUS/WT) or the spl7 (pMIR408:GUS/spl7) background. Seedlings were grown under different copper regimes, as indicated.

SPL7 directly regulates miR408 expression

To test whether SPL7 binds to the MIR408 promoter via the GTAC motifs, we fused its N–terminus portion (amino acids 1–297) containing the conserved DNA binding domain (SBP) with a His tag. The purified recombinant SPL7-SBP protein from Escherichia coli was then used in a series of electrophoretic mobility shift assays (Figure 6a). We found that digoxigenin-labeled probes I (from −351 to −312 upstream the TSS, including three GTAC motifs), II (from −293 to −266, including two GTAC motifs) and III (from −130 to −91, including two GTAC motifs), but not probe IV (from −77 to −45, including two GTAC motifs), were retarded by the addition of the SPL7-SBP recombinant protein (Figure 6b). In the presence of excessive unlabeled probes, binding between SPL7-SBP and probes II and III was effectively impaired (Figure 6b). By contrast, the competitor probe did not significantly change the shift pattern of probe I (Figure 6b). These results suggest that SPL7 binds to the MIR408 promoter mainly through the GTAC motifs located in the middle of the proximal promoter region.

Figure 6.

SPL7 binds to the GTAC motifs within the MIR408 promoter. (a) Schematic presentation of the MIR408 proximal promoter regions in Arabidopsis. The annotated pre-miRNA is depicted as an open horizontal bar in which the position of the mature miRNA is shaded. TSS (+1) was determined by mapping the corresponding full-length cDNA sequence to the genome. The ovals indicate the positions of the GTAC motifs. Probes used in subsequent EMSA are indicated as short horizontal lines. (b) EMSA analysis of the recombinant SBP domain of SPL7. The ability of SPL7-SBP to retard the migration of four GTAC-containing regions of the MIR408 promoter (probes I–IV) was tested: 1, digoxigenin-labeled probes only; 2, labeled probes incubated with SBP; 3–5, excessive unlabeled probes added at different concentrations as competitors. FP, free probe; SB, shift band. *Indicates non-specific shift band.

To genetically test whether SPL7 is required for miR408 expression, we used an spl7 null mutant in which a T–DNA is inserted in the SPL7 locus (Yamasaki et al., 2009). Northern blotting (Figure 5a) and qRT-PCR analyses (Figure 5b,c) showed that maintaining proper transcript levels as well as the induction of both miR408 and pre-miR408 in response to low copper are blocked by the spl7 mutation. Furthermore, we introduced the pMIR408:GUS reporter gene in the spl7 background to generate the pMIR408:GUS/spl7 plants. Compared with wild-type seedlings expressing pMIR408:GUS (pMIR408:GUS/WT), pMIR408:GUS/spl7 seedlings have much reduced GUS staining with and without copper supplementation (Figure 5d), indicating that the activation of the MIR408 promoter under different copper conditions requires SPL7. Based on these results, we concluded that SPL7 binds to the GTAC motifs within the MIR408 promoter, and is directly required for proper miR408 transcription in response to copper availability. Interestingly, qRT-PCR analysis revealed that SPL7 is expressed at comparable or higher levels in adult organs (Figure 1b; Yamasaki et al., 2009). The parallel expression pattern of miR408 and SPL7 suggests that SPL7 may well play a role for miR408 transcription in other biological contexts.

Finally, we sought to test whether the activation of miR408 could rescue the spl7 mutation, which causes reduced plant size and growth in low copper conditions (Yamasaki et al., 2009). To this end, we introduced the 35S:pre-miR408 transgene into the spl7 background to generate the miR408-OX/spl7 plants. In these plants, an over-accumulation of miR408 (Figure 7a) was observed, as well as downregulation of the target genes LAC13 and Plantacyanin to levels comparable with those in the wild type (Figure 7b). Consistent with the previous report (Yamasaki et al., 2009), both fresh weight and root length of spl7 plants grown on MS are significantly reduced compared with the wild type (Figure 7c,d). These severe growth defects are alleviated when the plants are supplemented with copper (Figure 7c,d). Interestingly, the miR408-OX/spl7 seedlings grown on MS exhibited clearly improved growth vigor compared with spl7 (Figure 7c), with both fresh weight and root length restored to levels significantly higher than the mutant (Figure 7d). Collectively, these results demonstrate that miR408 can partially rescue the spl7 defects, and indicate that SPL7-based regulation of miR408 is an integral part of SPL7 function.

Figure 7.

Overexpression of miR408 partially rescues defects of the spl7 mutant. (a) RNA blot analysis of miR408 accumulation in spl7, wild type and miR408-OX/spl7 seedlings. U6 snRNA was used as the loading control. (b) qRT-PCR analysis of the miR408 target genes Plantacyanin and LAC13 in wild-type, spl7 and miR408-OX/spl7 seedlings. Error bars present SDs (n = 3). (c) Morphology of wild-type, spl7 and miR408-OX/spl7 seedlings grown on MS (left panel) or MS + Cu (right panel). (d) Quantitative measurement of fresh weight and root length of the three genotypes described in (c). Error bars indicate SDs (n ≥ 12). *P < 0.01 and **P < 0.001 versus wild type, by Student's t–test.

Discussion

Proper gene expression is fundamental to the integrity and function of biological processes and developmental programs. We now appreciate that spatial and temporal gene expression is deliberately controlled through sophisticated regulatory networks. High throughput systems-level studies and detailed molecular works have revealed that miRNAs constitute an important part of these regulatory networks (Shalgi et al., 2007; Tsang et al., 2007; Hobert, 2008; Martinez et al., 2008; Yu et al., 2008; Re et al., 2009). It is known that many miRNAs as well as their target genes are deeply conserved across species, indicating that these circuits represent evolutionarily successful mechanisms for regulating gene expression. MiR408 is one of the most conserved miRNA families in land plants. Elucidating its function and regulation thus represents an important research question.

In the current work we show that miR408 is a powerful regulator for vegetative development in Arabidopsis. This conclusion was drawn from two sets of complementary experiments. When over produced by the constitutive 35S promoter, excessive miR408 causes an obvious enhancement of growth vigor. The miR408-OX transgenic plants exhibit enlarged stature compared with wild-type plants, soon after germination (Figure 2). All vegetative organs in the transgenic plants, including cotyledon, rosette leaf, cauline leaf and petiole, are larger than those of wild-type plants (Figure 2). Thus, the whole rosette area of the transgenic plants has significantly increased biomass (Figure 2). Conversely, physiological consequences of the loss of function of miR408 include reduced vegetative organ size and diminished biomass (Figures 3 and 4), phenotypes that are opposite to those of the miR408-OX transgenic plants. Together, these results firmly establish that levels of miR408 are tightly linked to the vigor of vegetative development in Arabidopsis.

The strong influence of miR408 over vegetative growth indicates that its level is subject to strong regulation. Earlier works in Arabidopsis have revealed copper availability as one of the conditions implicated in miR408 regulation (Yamasaki et al., 2007; Abdel-Ghany and Pilon, 2008). Copper is an essential co–factor for numerous proteins because of its ability to cycle between oxidized Cu(II) and reduced Cu(I) states (Burkhead et al., 2009; Palmer and Guerinot, 2009). The most abundant copper-containing protein in plants is plastocyanin, which transfers electrons from the cytochrome b6f complex to photosystem I. Copper is also used as a co–factor by proteins involved in neutralizing reactive oxygen species, lignification of the cell wall, ethylene perception and formation of phenolics in response to pathogen attack (Burkhead et al., 2009). Although diminished cellular copper impedes photosynthesis and other processes, excessive copper leads to the generation of harmful reactive oxygen species and the replacement of other metal co–factors. Therefore, conserved mechanisms have evolved in eukaryotes to tightly control the acquisition, distribution and storage of copper (Festa and Thiele, 2011).

Given the involvement of miR408 in copper homeostasis, it is not surprising that its expression requires SPL7 (Yamasaki et al., 2009). SPL7 is homologous to CRR1 of the green algae Chlamydomonas, which is specifically activated under copper deficiency (Kropat et al., 2005). The GTAC motif found in CRR1 targets (Kropat et al., 2005) has been recognized as the core of the copper-response elements in diverse plants (Quinn and Merchant, 1995; Quinn et al., 1999; Nagae et al., 2008; Yamasaki et al., 2009). It was recently demonstrated that Cu(II) specifically inhibits the DNA binding activity of both CRR1 and SPL7, and prevents transcription activation in vitro (Sommer et al., 2010). SPL7 is thus likely to be the copper sensor in Arabidopsis that regulates gene expression in response to reduced cellular copper. We show that SPL7 transcriptionally regulates miR408 expression through binding to the GTAC motifs in its promoter (Figures 5 and 6). Significantly, constitutively expressed miR408 in the spl7 background partially rescues the severe growth defects of the mutant (Figure 7). Together, these results indicate that the SPL7-miR408 circuit represents a mechanism for modulating growth in response to varying copper conditions.

Interestingly, the miR408 level is influenced by a variety of environmental stresses. In addition to copper deficiency in Arabidopsis, the documented conditions inducing miR408 expression include mechanical stress in poplar (Lu et al., 2005), dehydration or drought in Medicago and Hordeum vulgare (barley; Kantar et al., 2010; Trindade et al., 2010), and reactive oxygen species in Oryza sativa (rice; Li et al., 2010). These observations suggest that miR408 is a general hub for integrating stress signals. Previously, it has been suggested that induced miR408 expression is used to inhibit the accumulation of transcripts encoding copper proteins, thus ensuring the preferential delivery of cellular copper to plastocyanin, which in turn helps to maintain photosynthesis (Burkhead et al., 2009). It can be envisioned that this mechanism may also operate in response to other stresses. Collectively, our results and earlier findings support a model in which various environmental stresses activate the respective signal transduction pathways that entail transcriptional induction of miR408 expression to modulate development and growth (Figure 8). Further analysis of the miR408-mediated gene circuits will thus provide much needed insight into the coordinated and adaptable control of the gene batteries that underpin plant development in response to changing environments.

Figure 8.

A working model illustrating the biological role of miR408 in Arabidopsis. Cellular levels of miR408 are known to be influenced by copper deficiency (Yamasaki et al., 2007; Abdel-Ghany and Pilon, 2008) and a variety of other stress conditions, such as dehydration (Kantar et al., 2010; Trindade et al., 2010) and elevated levels of sucrose (Ren and Tang, 2012). The environmental challenges may activate specific signal transduction pathways, such as SPL7-mediated copper signaling (Yamasaki et al., 2009) to promote the accumulation of miR408, which post-transcriptionally represses target genes such as Plantacyanin and LAC13. Because miR408 levels are tightly associated with growth vigor, the transcriptional regulation of MIR408 thus represents a mechanism for modulating vegetative development based on environmental cues.

Experimental procedures

Plant material and growth conditions

The wild-type plant used in this work was Arabidopsis thaliana ecotype Col–0. The mutant defective in the SPL7 gene was a T–DNA insertion line (SALK_093849), as described previously (Yamasaki et al., 2009). Four MIR408 T–DNA insertion lines (SALK_023586, SALK_038860, SALK_081087 and SALK_082709) and the At2 g47020 T–DNA insertion line (SALK_049729) were obtained from the Arabidopsis Biological Resource Center (ABRC, http://abrc.osu.edu), and another two lines concerning MIR408 were FLAG_481E04 and FLAG_545E03 from French National Institute for Agricultural Research (INRA) collection (WS–4 background; Brunaud et al., 2002). For germination, the seeds were placed on agar-solidified MS media, including 1% sucrose and the indicated concentrations of CuSO4. After incubation for 4 days at 4°C in the dark, seeds were transferred to continuous white-light conditions (with a light intensity of 120 μmol m−2 s−1), and were allowed to grow at 22°C for 7 days unless otherwise mentioned. Measurement for root length was performed on 7–day-old seedlings. Fresh weight was determined for seedlings grown for 12 days. The measurement of leaf size was performed on 3–week-old adult plants.

Generation of transgenic plants

To clone the promoter region of MIR408, two primers (p408–F and p408–R in Table S1) were used to amplify the genomic fragment 820 bp upstream of the TSS. This DNA fragment was then inserted into the pCAMBIA-1381Xa vector (CAMBIA) after HindIII and SpeI double digestion to create the pMIR408:GUS construct, in which the GUS reporter gene with an added start codon is driven by the MIR408 promoter. Plants were transformed with Agrobacterium GV3101 containing the pMIR408:GUS construct by the standard floral-dipping method (Clough and Bent, 1998). The transgenic plants were selected on Hygromycin-containing (25 mg L−1) MS media and allowed to propagate. T2 generation plants were used for GUS activity analysis.

To overexpress miR408 in the spl7 mutant, the construct containing 35S:pre-miR408, as previously described (Zhang et al., 2011), was transformed into the spl7 backgrounds by the standard floral-dipping method (Clough and Bent, 1998). Transgenic plants were selected on MS plates containing 20 μg ml−1 Basta (bioWORLD) and allowed to propagate. The T3 generation plants were selected and analyzed. To overexpress miR408 in the FLAG_481E04 or FLAG_545E03 mutants, the pre-miR408 was amplified from Arabidopsis genomic DNA and then inserted into the pJIM19 binary vector after XhoI and StuI double digestion. Transgenic plants were selected on MS plates containing 50 μg ml−1 hygromycin B (bioWORLD) and allowed to propagate. T2 generation plants were selected and analyzed.

For designing an artificial microRNA targeting pre-miR408, wmd 3 (http://wmd3.weigelworld.org/cgi-bin/webapp.cgi) was used and the sequence 5′-TAGTAAACTCAATCGATGCGC-3′ within pre-miR408 was chosen as amiR408. A 405–bp PCR product containing the pre-miR319A sequence was obtained from Arabidopsis genomic DNA to produce pAth-miR319a, as previously described (Schwab et al., 2010). This vector was used as the DNA template in a series of bridge PCR reactions to substitute miR319 with amiR408. The confirmed final PCR product was cloned into a binary vector pJIM19 (Figure S3a) and transformed into the wild-type background. Transgenic plants were selected on MS plates containing 20 μg ml−1 Basta (bioWORLD, http://www.bio-world.com), and allowed to propagate. T3 generation plants were selected and analyzed. The primers used above are listed in Table S1.

RNA analyses

Total RNA (10 μg), extracted using the TRIzol reagent (Invitrogen, http://www.invitrogen.com), was used for blotting as previously described (Zhang et al., 2011). For qRT-PCR, total RNA was treated with RNase-free DNaseI (Invitrogen), and reverse transcribed using the SuperScript II reverse transcriptase (Invitrogen). The resultant cDNA was analyzed using the SYBR Green master mix (Applied Biosystems, http://www.appliedbiosystems.com) with the ABI 7500 Fast Real-Time PCR System (Applied Biosystems). PCR reactions were performed in triplicate for each biological sample. The Actin7 amplicon was used for normalization. For miRNA quantification, first-strand cDNA was synthesized using the NCode miRNA First-Strand cDNA Kit (Invitrogen). Levels of miRNA were measured by qPCR using an miRNA-specific forward primer and a universal reverse primer supplied by the manufacturer. 5S ribosome RNA was used for normalization. Probe and primer sequences are listed in Table S1.

Histochemical staining for GUS activity

We used T2 transgenic plants expressing the pMIR408:GUS reporter gene in various genetic backgrounds. Seedlings grown on MS with or without copper supplement under continuous light were immersed in GUS staining solution [0.1 m NaPO4 (pH 7.0), 10 mm EDTA, 0.1% Triton X–100, 1 mm K3Fe(CN)6, 2 mm X–Gluc], incubated overnight at 37°C. Following the removal of the staining solution, seedlings were washed with several changes of 50% ethanol until chlorophyll was no longer visible. Images of the GUS staining pattern were taken with a digital camera.

Electrophoretic mobility shift assay

The sequence encoding the SBP domain of SPL7 was amplified by RT-PCR using the high-fidelity Pfusion DNA polymerase (New England Biolabs, http://www.neb.com) and primers, as listed in Table S1. PCR products were cloned into the vector pET-28a (+) (Novagen, now EMD Millipore, http://www.emdmillipore.com) between the NcoI and XhoI restriction sites and sequenced. The resulting plasmid was introduced into E. coli strain BL–21 (Stratagene, now Agilent Technologies, http://www.genomics.agilent.com). Production of SPL7-SBP proteins fused with the His tag was induced by 1 mm isopropyl-β-d-thiogalactopyranoside at 37°C. The recombinant proteins were purified with the Ni-NAT Agarose system (Qiagen, http://www.qiagen.com).

Fragments of the MIR408 promoter containing the GTAC motifs were synthesized as pairs of complementary oligonucleotides and labeled by digoxingenin in the second generation DIG Gel Shift Kit (Roche, http://www.roche.com). Briefly, the SBP domain protein was incubated together with digoxingenein-labeled probes and various concentrations of competitor DNA in 20–μl reaction mixtures containing 20 mm HEPES (pH 7.6), 1 mm EDTA, 10 mm (NH4)2SO4, 1 mm DTT, 0.2% (w/v) Tween 20, 30 mm KCl, 0.05 μg μl−1 poly [d(I–C)] and 0.005 μg μl−1 poly-l-lysine for 15 min at room temperature (22–25°C). The samples were loaded onto 6% (w/v) polyacrylamide gels and run at 4°C at 0.8 V cm−2 in half-strength TBE electrophoresis buffer (44.5 mm Tris, 44.5 mm boric acid and 1 mm EDTA). DNA was blotted onto Hybond+ nylon membrane (GE Healthcare Life Sciences, http://www.gelifesciences.com) and detected with a digoxigenin-specific antibody (Roche). Probes are listed in Table S1.

Acknowledgements

We thank Drs Toshiharu Shikanai and Hervé Vaucheret, respectively, for providing the homozygous SPL7 and MIR408 T–DNA insertion lines. This work was supported by a grant from the National Science Foundation (DBI-0922526).

Ancillary

Advertisement