TEMPRANILLO is a direct repressor of the microRNA miR172.

In the age-dependent pathway, the microRNA 156 (miR156) is essential for the correct timing of developmental transitions. MiR156 negatively regulates several SPL genes, which promote the juvenile-to-adult and floral transitions in part through up-regulation of miR172. The transcriptional repressors TEMPRANILLO1 (TEM1) and TEM2 delay flowering in Arabidopsis thaliana at least through direct repression of FLOWERING LOCUS T (FT) and gibberellins biosynthetic genes, and have also been reported to participate in the length of the juvenile phase. TEM mRNA and miR156 levels decrease gradually, allowing progression through developmental phases. Given these similarities, we hypothesized that TEMs and the miR156/SPL/miR172 module could act through a common genetic pathway. We analyzed the effect of TEMs on miR156, SPL and miR172 levels, tested binding of TEMs to these genes using chromatin immunoprecipitation and analyzed the genetic interaction between TEMs and miR172. TEMs played a stronger role in the floral transition than in the juvenile-to-adult transition. TEM1 repressed MIR172A, MIR172B and MIR172C expressions and bound in vivo to at least MIR172C sequences. Genetic analyses indicated that TEMs affect the regulation of developmental timing through miR172. This article is protected by copyright. All rights reserved.


Accepted Article
This article is protected by copyright. All rights reserved.

Comparison of phenotypes caused by alterations in TEM1/TEM2 and the agedependent pathway
In order to uncover the mechanism of TEM1/TEM2 action on phase change, we compared the phenotypes of plants with altered TEM levels with those of plants affected in the agedependent pathway under SDs and LDs. It has been shown that miR156 maintains the juvenile traits by repressing SPLs, that in turn activate miR172 which will promote adult epidermal identity, such as the appearance of trichomes on the abaxial surface of rosette leaves (Poethig 2003, 2013, Huijser and Schmid, 2011, Yu et al., 2015Wang et al., 2019).
The timing of abaxial trichome formation is correlated with flowering time, consistent with the fact that juvenile-to-adult vegetative phase change contributes to the acquisition of the competence to flower. Thus, miR172 is involved in the timing of trichome formation, and its AP2 target genes, together with KANADI, have been shown to mediate the temporal and spatial integration for abaxial trichome formation (Wang et al., 2019). Interestingly, alterations of TARGET OF EAT 1 (TOE1) or miR172 levels affect the timing of abaxial, but not adaxial, trichome formation (Wang et al., 2019). Consequently, the length of juvenile or adult phases can be determined by counting the leaves without and with abaxial trichomes, respectively.
Under SDs, we confirmed previously reported results and used them as control for our tem mutant plants growing in parallel. Thus, 35S::miR156 plants showed a dramatically extended juvenile phase, consistent with the results of Wu et al. (2009), and a shortened adult phase (Figures 1a,b,S1a,b;Table S2a,b), resulting in late flowering compared with wild-type plants (Figures 2a,b,S2a,b;Table S3a,b). Plants in which miR156 activity was inhibited (35S::MIM156) and plants expressing a miR156-resistant form of SPL9 (pSLP9::rSPL9) lacked the juvenile phase, in agreement with a previous report (Wu et al., 2009), and had an adult phase similar to that of wild-type plants under SDs (Figures 1a,b, S1a-d; Table S2a,b). Therefore, 35S::MIM156 and pSLP9::rSPL9 plants flowered with fewer leaves than wildtype plants (Figures 2a,S2a,c; Table S3a) due to the absence of the juvenile phase.

Accepted Article
This article is protected by copyright. All rights reserved.
As miR156 and TEM1/TEM2 follow similar temporal expression patterns and have partially overlapping phenotypic effects, we tested whether miR156 regulates TEM1/TEM2 levels or vice versa. 35S::miR156 and 35S::MIM156 plants did not show significant alterations in TEM1 and TEM2 mRNA levels that could correlate with their flowering phenotypes under LD and SD conditions ( Figure S5). For the analysis of the effect of TEMs on miR156, we collected samples at two time points under LD, one during the light period (ZT8), when TEM mRNA levels are low (Castillejo & Pelaz, 2008;Osnato et al., 2012) and the miR156targeted SPL9 mRNA peaks ( Figure S6), and one during the night (ZT18), when TEM mRNAs peak (Castillejo & Pelaz, 2008;Osnato et al., 2012). Although RNA blots with samples collected at one time point, ZT8 in 10-day-old plants, showed slight differences ( Figure S7a,b), when we examined miR156 levels in tem1 tem2 mutants at different ages, we did not observe substantial differences relative to wild-type plants ( Figure S7c). Even the miR156 level reduction happened at the same time, after day 12, in both genotypes and, therefore, we could not confirm that TEMs have an effect on miR156 ( Figure S7c).
Then we tested the effect of TEMs on the expression of SPL genes under LDs. As previously, when we studied SPL3, SPL9 and SPL15 expressions at one time point we found that SPL3 transcript abundance was slightly increased in 10-day-old tem1 tem2 plants (Figure 5a-c) and that the expression of all three genes was reduced in 35S::TEM1 plants (Figure 5a-c), but an analysis of SPL mRNA levels at different ages confirmed that only SPL3 and SPL9 were down-regulated in 35S::TEM1 plants (Figure 5d-i) and none was affected in tem1 tem2 mutant plants (Figure 5d-i).

TEMs down-regulate miR172 and bind to MIR172C chromatin
Since TEMs were reported to have a role in the length of the juvenile phase (Sgamma et al., 2014) and we have not observed expression changes of miR156 or SPL genes in tem1 tem2 mutant plants, we wondered if the effect on the phase transition could be through miR172 whose genes transcription is activated by SPLs (Wu et al., 2009). Consistent with this hypothesis, we found increased abundance of mature miR172 in tem1 tem2 and reduced abundance in 35S::TEM1 when we analyzed 10-day-old plants (Figure 6a,b). Similar results

Accepted Article
This article is protected by copyright. All rights reserved.
were obtained by RNA blot (Figure 6a,b) and RT-qPCR ( Figure S8). 35S::miR156 and pSPL9::rSPL9 plants used as controls showed slightly reduced and increased miR172 levels, respectively (Figure 6a), as expected (Wu et al., 2009). Then we examined mature miR172 levels in tem1 tem2 mutants at different ages. Again, we found that tem1 tem2 mutants show increased miR172 levels up to day 12 (Figure 6c), indicating that TEMs down-regulate miR172 at early developmental stages, consistent with their effect on the juvenile-to-adult and floral transitions.
We found putative RAV binding sites in four MIR172 genes (Table S6). We chose MIR172C gene because it has several putative RAV binding sites and it was possible to design specific oligos to test binding of TEM1 and TEM2 by ChIP. TEM1 clearly bound to a fragment containing two putative RAV binding sites and TEM2 also bound to some extent ( Figure 6d). This suggests that TEMs repress miR172 expression through direct binding to, at least, MIR172C chromatin. We used transient expression assays in Arabidopsis protoplasts to test whether this is the case. Indeed, we found that TEM1 represses expression of a reporter construct that carries a fragment of the MIR172C gene containing the two RAV binding sites bound in ChIP experiments ( Figure 6e). In addition, MIR172C transcript levels are increased in tem1 tem2 and RNAi-TEM1/2 plants, and are decreased in 35S::TEM1 plants (Figure 6f).
Similarly, we found that MIR172A and MIR172B transcript levels are increased when TEM genes are down-regulated and that MIR172A levels are decreased when TEM1 is overexpressed ( Figure 6f). Upregulation of the three MIR172A, MIR172B and MIR172C genes resulted in the high mature miR172 levels observed in tem1 tem2 mutant plants ( Figure   6a,c and S8). TEMs, therefore, down-regulate miR172 levels by binding to at least the MIR172C gene and repressing its transcription.

TEMs act partially through miR172
If the early vegetative phase change and early flowering of tem1 tem2 and RNAi-TEM1/2 plants is due to increased miR172 abundance, silencing of miR172 should suppress the vegetative phase change and flowering phenotypes of tem1 tem2 and RNAi-TEM1/2.
Unfortunately, we found that several transgenes were silenced when introduced in tem1 and 35S:TEM1 mutant backgrounds ( Figure S9a,b,c). Silencing of transgenes by T-DNA insertion mutations has already been reported (e.g. Daxinger et al., 2008;Wu et al., 2009). Therefore, to analyze the genetic interaction between TEMs and miR172 we crossed 35S::MIM172 with RNAi-TEM1/2 and checked that the 35S::MIM172 transgene was not

Accepted Article
This article is protected by copyright. All rights reserved.  Table S7a,b), and they flowered with more leaves than RNAi-TEM1/2, but with fewer leaves than 35S::MIM172 plants (Figures 8, S12; Table S8a). These results indicate that 35S::MIM172 suppresses partially the early juvenile-to-adult transition and the early flowering of RNAi-TEM1/2. Taking together these results and the effect of TEMs on miR172 levels, we can conclude that TEMs delay flowering, and to a lesser extent the vegetative phase change, partially through miR172.

TEMs regulate miR172 levels
Several flowering-time regulators, such as FCA, GIGANTEA, SHORT VEGETATIVE PHASE, SUPPRESSOR OF OVEREXPRESSION OF CO 1, SPL9, and SPL15, affect miR172 levels. Several of these genes regulate transcription of MIR172 genes, whereas others affect processing of miR172 primary transcripts (Jung et al., 2007;Wu et al., 2009;Lee et al., 2010;Cho et al., 2012;Jung et al., 2012b;Tao et al., 2012;Hyun et al., 2016). In addition, proteins of the POLYCOMB REPRESSIVE COMPLEX 1 (PRC1) and PRC2, which establish and maintain transcriptional repression through histone modifications, repress MIR172B expression (Picó et al., 2015). Our work establishes that TEMs are new miR172 regulators ( Figure 6). Whereas all the previously reported miR172 regulators have been shown to affect expression or processing of MIR172A and/or MIR172B, we show here that TEMs regulate MIR172A, MIR172B and MIR172C expressions. Therefore, multiple flowering-time regulators fine-tune mature miR172 levels by regulating several MIR172 genes. Further work will be required to understand whether there are interactions among all these factors and how they contribute to establish the temporal and spatial pattern of miR172 expression.

Accepted Article
This article is protected by copyright. All rights reserved.

TEMs play a more modest role in the juvenile-to-adult transition
We had previously shown that TEM1 and TEM2 play a role in the timing of the floral transition in Arabidopsis (Castillejo & Pelaz, 2008;Osnato et al., 2012). We show here that they also somewhat affect the timing of vegetative phase change. In a previous work, it was showed a more dramatic effect of TEMs on the juvenile phase (Sgamma et al., 2014).
However, they determined the length of the juvenile phase by the response of flowering to inductive photoperiods and measured number of days from germination, whereas we measured the number of leaves without abaxial trichomes (Telfer et al., 1997;Huijser & Schmid, 2011;Wang et al., 2019). In addition, we cannot rule out that the difference can be due to the use of different growth conditions. Taking into account that determining the juvenile phase length by measuring the competence to flower after LD exposure may not be the same as determining it by counting the number juvenile leaves, our results (Figure 1 and 3), although with a much weaker effect, seem consistent with those of Sgamma et al. (2014) as we both observed that TEM1 and TEM2 redundantly delay vegetative phase change.
Unexpectedly, when TEM1 is overexpressed, instead of an extension of the adult phase, we observed a dramatic lengthening of the juvenile phase under both LDs and SDs, together with an extreme shortening or even suppression of the adult phase under LDs (Figure 1i,j).

Accepted Article
This article is protected by copyright. All rights reserved.

Analyses of miRNA and transcript levels
For miRNA analyses, total RNA was extracted from pools of at least 10 plants using the Real ARNzol Spin kit (+PVP; Durviz) or using Trizol (Ambion) following manufacturer's instructions. RNAs were treated with DNase I using the DNA-free kit (Ambion) and were precipitated with sodium acetate. Stem-loop reverse transcription followed by quantitative real time PCR (stem-loop RT-qPCR) and RNA blots were performed as previously described (Martin et al., 2009). Primer and probe sequences are shown in Table S1.  Table  S1.
The analysis of MIR172C RNA levels was performed using TaqMan assays provided by Applied Biosystems (assay IDs: ARKA3K9 for MIR172C and At02163341_gH for IPP2).
The qPCR reactions, performed in triplicate in a volume of 14 µl, contained 2 μl of cDNA, 1X TaqMan Multiplex Master mix (Applied Biosystems) and 1X TaqMan assay, and were incubated as indicated above. Quantification of MIR172C RNA was standardized to IPP2 mRNA levels and data were analyzed using the 2 -ΔΔC T method (Livak & Schmittgen, 2001).

GUS staining
Histochemical analyses of GUS expression were performed as previously described (Blázquez et al., 1997).

Accepted Article
This article is protected by copyright. All rights reserved.

Identification of putative RAV binding sites
To find putative RAV binding sites, sequences of interest were searched for the pattern

Chromatin immunoprecipitation analyses
Chromatin immunoprecipitation (ChIP) experiments were performed using a modified version of a previously reported protocol (Matias-Hernandez et al., 2010). Direct binding of TEM1 and TEM2 to the regulatory regions of putative targets was assayed using the 35S:TEM1-HA and 35S:TEM2-HA lines previously described (Castillejo & Pelaz, 2008).
Wild-type plants were used as negative controls. The crosslinked DNA was immunoprecipitated with an anti-HA antibody (Sigma) and purified using Protein A-Agarose resin (Millipore). Enrichment of the target regions was determined by qPCR using different primer sets specific for putative direct targets, as listed in Table S1. The qPCR assay was conducted in triplicate using a SYBR Green Assay (SYBR Green Supermix, Roche) and was performed in a Roche LightCycler ® 480 System. For the binding of TEM1 and TEM2 to the selected genomic regions, the affinity of the purified sample obtained in the 35S:TEM1-HA and 35S:TEM2-HA lines was compared with the affinity-purified sample obtained in the wild-type background, which was used as negative control. Fold enrichment, relative to wildtype input DNA immunoprecipitated with no antibody, was calculated using the 2 -ΔΔC T method (Livak & Schmittgen, 2001).

Transient expression assays
To generate the set of reporter vectors, different regions of MIR172C were cloned as SalI-PstI fragments in a modified pGreenII 0800-LUC vector carrying p35S::LUC (as reporter) and

Accepted Article
This article is protected by copyright. All rights reserved.
manufacturer's protocol. Transcriptional repression activity of TEM1, based on the relative ratio of Luciferase (LUC) and Renilla (REN) mRNA abundance, was assessed by RT-qPCR.
Primers used to generate the reporter vectors and for qPCR are listed in Table S1.

Genetic crosses
RNAi-TEM1/2 and 35S::MIM172 plants were crossed and the F 1 generation was allowed to self-pollinate. F 2 plants were selected based on resistance to kanamycin and Basta and were genotyped by checking resistance of the F 3 generation to kanamycin and Basta. F 3 homozygous plants were used for all the experiments.

Statistical analyses
Statistical analyses were performed with GraphPad Prism 6 or 7 software (GraphPad comments from all other authors.
The authors declare that there is NO conflict of interest.

Figure S1
Effect of TEMs, miR156, SPL9 and miR172 on the juvenile-to-adult transition.

Accepted Article
This article is protected by copyright. All rights reserved.

Table S1 Primer sequences
Tables S2a-S2j Statistical analyses of data from Figure 1 Tables S3a-S3j Statistical analyses of data from Figure 2 Tables S4a-S4b Statistical analyses of data from Figure 3 Tables S5a-S5b Statistical analyses of data from Figure 4 Table S6    Asterisks indicate statistically significant differences relative to the WT (*, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001; ns, not significant), using one-way ANOVA followed by Dunnett's multiple comparison test (a-h) or using the Student's t test (i,j). Histograms and statistical analyses of these data are shown in Figure S1 and Tables S2a-S2j, respectively.

Accepted Article
This article is protected by copyright. All rights reserved. independent experiments, n=15-68. Asterisks indicate statistically significant differences relative to the WT (*, P ≤ 0.05, **, P ≤ 0.01, ***, P ≤ 0.001, ****, P ≤ 0.0001; ns, not significant), using one-way ANOVA followed by Dunnett's multiple comparison test (a-h) or using the Student`s t test (i,j). Histograms and statistical analyses of these data are shown in Figure S2 and Tables S3a-S3j, respectively. Asterisks indicate statistically significant differences relative to the WT (***, P ≤ 0.001), using the Student`s t test. Histograms and statistical analyses of these data are shown in Figure S3 and Tables S4a-S4b, respectively.

Accepted Article
This article is protected by copyright. All rights reserved.
as normalization control and normalized levels in the wild type (a-c) or in the wild type at day 6 (d-i) were set to 1. Each symbol represents the mean of the combination of 2-5 independent experiments, n=2-5, and lines represent the standard error of the mean (SEM).

Accepted Article
This article is protected by copyright. All rights reserved. conditions. The number of juvenile and adult leaves was determined by counting the rosette leaves without and with abaxial trichomes, respectively. The lines within the boxes represent the median and the whiskers represent the minimum and maximum values. Data represent the combination of 3 independent experiments, n=59-60. Asterisks indicate statistically significant differences (**, P ≤ 0.01, ****, P ≤ 0.0001; ns, not significant), using one-way ANOVA followed by Tukey's multiple comparison test. Although the WT is shown as reference, only the relevant statistical analyses for the transgenic lines are shown. Histograms and statistical analyses of these data are shown in Figure S11 and Tables S7a-S7b, respectively. conditions. The lines within the boxes represent the median and the whiskers represent the minimum and maximum values. Data represent the combination of 3 independent experiments, n=60. Asterisks indicate statistically significant differences (***, P ≤ 0.001; ns, not significant), using one-way ANOVA followed by Tukey's multiple comparison test.
Although the WT is shown as reference, only the relevant statistical analyses for the transgenic lines are shown. Histograms and statistical analyses of these data are shown in Figure S12 and Tables S8a-S8b, respectively. (c) WT, RNAi-TEM1/2, 35S::MIM172 and