Auxin and epigenetic regulation of SKP2B , an F-box that represses lateral root formation

In plants, lateral roots are originated from pericycle founder cells that are specified at regular intervals along the main root. Here, we show that Arabidopsis thaliana SKP2B, an F-box protein, negatively regulates cell cycle and lateral root formation as it represses meristematic and founder cell divisions. According with its function, SKP2B is expressed in founder cells, lateral root primordia (LRP) and in the root apical meristem (RAM). We identified a novel motif in SKP2B promoter that is required for its specific root expression and auxin-dependent induction in the pericycle cells. Next to a transcriptional control by auxin, SKP2B expression is regulated by histone H3.1/H3.3 deposition in a CAF-depended manner. SKP2B promoter and the 5’ of the transcribed region are enriched in H3.3, which is associated with active chromatin states, over H3.1. Furthermore, the SKP2B promoter is also regulated by H3-acetylation in an auxin- and IAA14-dependent manner, reinforcing the idea that epigenetics represents an important regulatory mechanism during lateral root formation. promoter fragments were amplified by PCR and resolved in agarose gels. We also amplified promoter regions of root-expressed genes. We used a near-localized SKP2B gene (At1g77100, PIN6), the cell cycle and auxin up-regulated gene CYCB1;1 , (At4g37490), an auxin down-regulated gene ( GRP , At4g30450) and the ACTIN2 ( ACT2 , At3g18780). The H3K9K14 ChIP assays and data analysis were carried out basically as previously described (Ramirez-Parra et al., 2007), using chromatin isolated form roots cell of 7 day old plants. We fixed the roots in presence of 3 mM Sodium Butyrate (Sigma) and immunoprecipitation was carried out with anti-H3ac antibody (Usptate-Millipore #06-599). FastStart DNA Master SYBR Green I (Roche) was used for quantitative real-time PCR. Data correspond to the average of two independent biological experiments and three independent qPCR analyses per experiment. Primer sequences and conditions are available upon request.

marks associated with active transcription, occurs in the SKP2B promoter and that such modifications are auxin-and IAA14/SLR-dependent.

SKP2B is expressed in dividing tissues and during early stages of the lateral root initiation
As SKP2B functions in cell cycle, we studied its transcriptional regulation during the cell cycle. SKP2B showed two expression peaks that correlate with S and G2/M phases (Supplemental Fig. S1). To analyze its spatio-temporal expression pattern we constructed a transgenic line expressing the GUS reporter under the SKP2B promoter (named SKP2Bp:GUS).
Histochemical GUS staining showed that SKP2B was expressed in dividing areas (shoot and root meristems), in the leaf vasculature and in flowers . In roots, SKP2B is expressed in the RAM and in patches along the main root that correlates with LRP in all developmental stages, from stage 0 to VIII (Fig. 1A, 1F-O). Microscopic analyses revealed that SKP2B was also expressed in undivided cells close to the root tip that was restricted to pericycle cells at the xylem pole ( Fig. 1E-G), likely corresponding to founder cells.
In Arabidopsis, LR formation follows an acropetal sequence of development, with the earliest stages localized close to the root tip. The marker lines DR5p: GUS and GATA23 expression are considered to report the earliest events associated with LR initiation (Benková et al., 2003;Dubrovsky et al., 2008;De Rybel et al., 2010). Comparisons between DR5p:GUS and roots (Fig. 2D). Next, we analyzed in detail LR formation in the skp2b mutant, finding that skp2b mutants developed more LR (primordia plus emerged LRs) per millimeter than the control (Fig. 2E). When we analyzed the developmental stages of LRP (according to Malamy and Benfey, 1997), we found that 8 days old skp2b roots significantly contained more LRP in stages I and II than control, but we did not observe differences in the number of emerged LR (Fig. 2F). However, when we analyzed 13 days old seedlings, the number of emerged LRs was significantly higher in skp2b than control plants (Fig. 2G). Taken together, these data indicate that SKP2B acts as a repressor of cell division and LR formation.

SKP2B expression in the root is regulated by auxin
Auxin signaling plays a central role in the specification of founder cells (De Rybel et al., 2010) and during LRP development (reviewed in Perét et al., 2009). Since SKP2B functions in the LR formation we decided to analyze whether auxin controls the expression of SKP2B in the root. After 3 hours of auxin treatment, SKP2B was initially induced in the pericycle (Fig.   3A), but after 5 or 7 hours, GUS staining was also localized in the surrounding cortex and epidermis, although staining was always stronger in the pericycle layer (Fig. 3A). This data is consistent with the finding that SKP2B expression increases in the pericycle cells after 2 and 6 hours auxin treatment (Parizot et al., 2010). In addition, treatment of SKP2Bp:GUS with 1-Nnaphthylphthalamic acid (NPA), which inhibits auxin efflux and blocks LR development, eliminated SKP2B expression in the root, except from the root tip (Fig. 3B). It is possible that NPA impides founder cell specification and LR formation, and consequently SKP2B expression.
To answer this, we grew Arabidopsis seedlings in medium containing 0 or 5 µM of NPA for 7 days. Afterwards, seedlings were transferred to a fresh medium without NPA for 3 extra days and LR were counted only in the root portions that were grown the first 7 days. As shown in Figure 3C, NPA severely compromised, but did not eliminate the pericycle cell competence to further form LRP, suggesting that NPA does not completely block founder cell specification.
of IAA14/SLR. A gain-of-function mutation in IAA14 (slr-1) leads to plants without LR (Fukaki et al., 2002). Histochemical analyses of slr-1/SKP2Bp:GUS showed GUS staining only in the root meristem (Fig. 3G). Auxin treatment of slr-1 did not induce SKP2B expression (Fig. 3G), except for a reproducible expression in few pericycle cells in the differentiation zone ( Fig. 3H) that could represent specified founder cells.
Mutations affecting the auxin signaling reduce the number of LR (reviewed in Mockaitis and Estelle, 2008). We crossed SKP2Bp:GUS with auxin signaling mutants (tir1-1, axr1-12 and ibr5-1) reported to develop fewer LR than wild type. We found that auxindependent SKP2B induction was impaired in the axr1-12, a strong auxin signaling mutant (Hobbie and Estelle, 1995) (Fig. 4A), while mutations in TIR1 or IBR5 slightly reduced SKP2B induction. Next we studied the number of LR specified in these mutants, finding that tir1-1 and axr1-12 had a fewer number of GUS-stained LRP (Fig. 4B), while ibr5-1 developed similar number than control roots, suggesting that IBR5 activity is needed for the emergence of LR rather than for LR specification, likely due to the function of the IAA28-module is not affected in this ibr5-1 mutant (Strader et al., 2008).

Identification of a novel root-specific expression motif
In order to identify domains responsible for SKP2B expression in founder cells and LRP, we generated different constructs containing deleted versions of the SKP2B promoter. adenine. Next, by directed mutagenesis we generated de novo this mutant construction (SKP2B[1Kb-mut]p:GUS), replacing the same cytosine -397 by adenine. SKP2B [1Kbmut]p:GUS roots showed GUS staining in the root meristem but not in LRP ( Fig. 5D-E), demonstrating the relevance of this residue for its expression in LRP. In addition, this mutation also compromised the SKP2B auxin-dependent induction (Fig. 5F).
With this information, we analyzed the DNA sequence surrounding this cytosine using the PLACE motif search program (http://www.dna.affrc.go.jp/PLACE/index.html) to look for cis-elements. We found a plant motif denominated as "root specific motif" located between nucleotides -387 and -409 from the ATG (Supplemental Fig. S3). We have conducted an in silico analysis to search for promoters that contain at least 1 copy of this root motif, allowing only 1 mismatch in the sequence. We have identified more than 500 genes that contain this motif (Supplemental table I). When comparing these genes with those found in the pericycle cells (De Smet et al., 2008;http://www.ncbi.nlm.nih.gov/geo/query/acc. cgi?acc=GSE6349), we found that about 60% of these genes are expressed in the pericycle cells, while only a percentage of 36.5±3.8 % was obtained when random sampling (3 different random samples of 600 genes each). Moreover, 4% of these genes that contain this motif are induced in response to auxin in the pericycle (Supplemental table I), while random sampling only retrieved 0.15±0.028 %. These data indicate a positive correlation between the presence of this motif and pericycle expression.

SKP2B promoter is enriched in histone H3.3
Next, we wanted to get insight in the upstream molecular signaling that controls SKP2B expression in LRP. To do this, we conducted a yeast one-hybrid screening using the SKP2B[0.41Kb]promoter. We chose this promoter region due to the root specific motif identified, fuse to a minimal 35S, was not sufficient to drive SKP2B expression in planta (data not shown). We isolated 10 clones that after re-testing them in a medium containing 5, 10, 15 or to grow only in the presence of low 3-AT level (5 mM), while those transformed with H3.3 grew up to 20 mM of 3-AT (Supplemental Fig. S4), suggesting that H3.3 has higher affinity for SKP2B promoter than H3.1. This appealing result led us to investigate the H3 status across the SKP2B gene by mapping H3.1 and H3.3 occupancy in DNA extracted from roots. ChIP analyses using plants expressing Myc-tagged H3.1 or H3.3 (Stroud et al., 2012;see methods) followed by PCR amplification of different regions (Fig. 6A) revealed that the SKP2B promoter contained mostly H3.3, while H3.1 was untraceable (Fig. 6B). When the coding region was analyzed, the H3.3 amount increased significantly, but now the H3.1 was detectable, consistent with active transcription of this gene (Fig. 6B). We found that root expressed genes (see methods; Supplemental Fig. S5) also showed H3.3/H3.1 enrichment, but in these cases the levels of H3.1 detected in these promoters were much higher than in SKP2B promoter (Fig. 6B).
Since H3.3 is associated with actively transcribed chromatin and SKP2B is highly transcribed in roots upon auxin treatment, we assessed the effect of this hormone on H3.3 deposition. We found that H3.3 deposition did not increase by auxin treatment (Supplemental Fig. S6), suggesting that H3.3 deposition might be more related to specific cell type expression than to auxin response. HIRA1 chaperone replaces H3.3 for H3.1 in differentiating cells after they exit the cell cycle (Lennox and Cohen, 1988). A HIRA1 homolog was identified in Arabidopsis (Phelps-Durr et al. 2005), but its function is still poorly known and mutations in this gene result in an embryonic lethal phenotype, complicating genetic studies. On the other hand, CAF-1 complex is dedicated to the replication-coupled deposition of H3.1/H4 dimers (Polo and Almouzni, 2006) and viable mutants in Arabidopsis for CAF-1 complex subunits, fas1 and fas2, have been described (Kaya et al., 2001;Serrano-Cartagena et al., 1999). Thus, using this mutants, we decided to study whether alterations in H3.1 deposition influences on SKP2B expression.
Histochemical analyses of fas1-4/SKP2Bp:GUS or fas2-1/SKP2Bp:GUS eliminated SKP2B expression in LRP, but not in the main root meristem nor in the LRP surrounding cortex and epidermis ( Fig. 6C and Supplemental Fig. S7). We also found that SKP2B auxin induction was compromised in the fas1-4 mutant (Fig. 6D), suggesting that the correct H3.1 incorporation is needed for both its cell specific expression and auxin induction. Unexpectedly, when we analyzed in fas1-4 the expression of SKP2B[0.5Kb]p, a promoter region that specifically drives the expression in LRP, we found a correct GUS staining in LRP (Fig. 6C) and as well as auxin induction in the vascular tissue (Fig. 6D). These data indicate that maintenance of SKP2B expression in founder cells and LRP relays on SKP2B[0.5Kb]p region, but it is influenced by the CAF-1 function in the proximal upstream region.
Next we wondered whether the lack of SKP2B expression in the LRP in the fas1-4 mutant is a general effect on LRP-expressed genes or it is locus specific. To study this, we 1 0 generated fas1-4/GATA23p:GUS plants. GATA23 is only expressed in early LRP (De Rybel et al., 2010) and mutation in FAS1 did not affect its expression in early LRP (Supplemental Fig.   S7), suggesting that the regulation of SKP2B expression by CAF-1 activity is locus specific and it might represent a good example of how H3.1/H3.3 deposition regulates gene expression.
Finally, using the SKP2Bp:GUS reporter we studied the LR formation in the fas1-4 mutant. We found that fas1-4/SKP2Bp:GUS developed lower number of LRP and emerged LR per root length than SKP2Bp:GUS plants (Fig. 6F), indicating that FAS1 is needed for LR specification and emergence.

SKP2B promoter is regulated by auxin-dependent histone acetylation
In addition to histone H3 exchange, H3 acetylation on promoters plays an important role in regulating gene transcription. We carried out ChIP analyses using an antibody that recognizes H3K9ac and H3K14ac and PCR amplification of different regions of the SKP2B promoter. We found that SKP2B promoter was labeled by H3K9/K14ac (Fig. 7A). Next, we analyzed its H3 acetylation level in response to auxin in both wt and slr-1 roots. Interestingly, we found that auxin significantly promotes acetylation in the SKP2B promoter and to a lesser extend in the coding region, and such acetylation was significantly reduced in the slr-1 background. In this mutant, auxin treatment slightly increased the acetylation level in the SKP2B promoter, but never to the control ones (Fig. 7B). Similarly, we detected that root-expressed promoters of CYCB1;1, GRP, PIN6 and ACT2 also contained acetylated H3 (Fig. 7C). The slr-1 dominant mutation also seems to reduce the H3 acetylation level in the root-expressed promoters, but the reduction was significantly lower than in SKP2Bp. The bigger changes were found in the CYCB1;1 promoter, what was expected since the expression of this locus is induced by auxin in the root (Himanen et al., 2002), and in the PIN6 promoter, in which acetylation level was reduced by auxin treatment (Fig. 7C).
Next, we analyzed the effect of trichostatin A (TSA), a histone deacetylase inhibitor, on SKP2Bp:GUS and on the auxin signaling marker DR5p:GUS. Short TSA treatment (12 hours) led to higher and delocalized GUS staining in the basal meristem and transition zone in SKP2Bp:GUS roots (Fig. 7D). TSA-treated DR5p:GUS seedlings also showed a significantly increased GUS staining in the vasculature of the basal meristem and transition/differentiation zone ( Fig. 7D), similar to what was found in auxin treated seedlings. Conversely, we found lower levels of GUS staining in the most basal LRP in TSA-treated roots ( Figure 7E).
Remarkably, when seedlings were grown for 3 days in the presence of TSA, instead of 12 hours, the root growth was significantly delayed (Supplemental Fig. S8A-B). Moreover, we found that 1 1 promote LR formation in the slr-1 mutant (Fukaki et al., 2006). When slr-1 was treated with TSA for 3 days, we found SKP2B expression in specific cells in the pericycle (Supplemental Fig. S8E), likely corresponding to founder cells. Remarkably, all these SKP2B expression points appeared only on the root sections grown in the presence of TSA. Taken together our data suggest that H3 acetylation regulates auxin responsiveness in the basal meristem and SKP2B expression in the root.

SKP2B is a negative cell cycle regulator in the root system
Both, SKP2A and SKP2B were identified by its homology with the human Skp2, which is a key regulator of cell division (del Pozo et al., 2002). SKP2A is an auxin binding F-box protein that functions as a positive regulator of cell division (Jurado et al., 2008;Jurado et al., 2010). Despite the high homology of both F-box proteins, SKP2B functions as a negative regulator of cell division in the root meristem and in the founder cells.
Here, we show that skp2b mutant develops higher number of LRP in stages I and II than wild type roots, suggesting that SKP2B participates in the first anticlinal division of founder cells, idea that is supported by the SKP2B expression pattern. Based on this, we think that SKP2B might contribute to maintain founder cells undivided until the correct developmental time. However, although statistically significant, the increase in the number of LRP is not stunning. This could be explained by the redundant mechanisms that govern the cell division process, and that deprivation of SKP2B function is partially compensated by the function of other proteins, attenuating the skp2b root phenotype. This partial compensation has been shown for other cell cycle proteins such as Cdt1 (Nishitani et al. 2001) or p27/Kip1 (Carrano et al., 1999Amador et al., 2007;Müller et al., 1997). In addition, this idea is supported by the fact that the double mutant for SKP2B and RKP1 (KPC1-related RING finger protein), another E3 ligase that collaborates in the KRP1 proteolysis (Ren et al., 2008), develops more LRP in early stages than either singles mutants or wt plants (Supplemental Fig. S9). In view of these results, it is reasonable to think that additional and redundant mechanisms govern founder cells division, and that SKP2B function is just one of them.
This role as negative cell division regulator might conflict with the proposed role of SKP2B in the degradation of the cell division repressor KRP1 (Ren et al., 2008). However, it is possible that SKP2B degrades other targets in addition to KRP1, as has been shown for many E3 ligases including HsSkp2, which targets cell cycle repressors such p27 (Kossatz et al., 2004) and cell cycle activators such as E2F1 (Marti et al., 1999)

Identification of a novel and specific root motif
Until now, few root-specific motifs have been described. In this work we have identified a promoter domain and a motif that are needed for specific root expression of SKP2B.
An in silico search using this motif led us to identify more than 500 genes that contain it in their promoters (Supplemental table I). It is remarkable that more than 60 % of these genes are expressed in pericycle, suggesting that this motif might be needed for pericycle expression. A mutation in the cytosine in position -397 of this motif blocked SKP2B expression in LRP and almost its auxin responsiveness. An in silico search revealed the existence of an Aux-RE downstream of this motif (Supplemental Figure S3), suggesting that this cytosine might influence on this Aux-RE. One possibility is that this cytosine is regulated by methylation. However we do not think that methylation is important for SKP2B regulation, since SKP2Bp:GUS plants treated with 5-aza-2'-deoxycytidine, a inhibitor of DNA methylation, did not show differences in GUS staining (data not shown). Recently, the screening of a Ds-element enhancer trap lines in Arabidopsis led to the identification of a root-specific promoter in the At1g73160 gene (Vijaybhaskar et al., 2008), which contains a copy of the here identified LR specific motif.
Based on these observations, we can conclude that the motif identified in this work is important to confer expression in LRP as well as to auxin response.

3
to study LR development for being the earliest reporter associated with this process (Benková et al., 2003;Dubrovsky et al., 2008), suggesting that some of these LRP are arrested (Zolla et al., 2010;this work). Taken together, we think that SKP2Bp:GUS is an excellent and trustworthy maker to study LR development in different conditions or mutant backgrounds. For example, the use of this marker has easily showed that mutations in AXR1 affects to LR specification while a mutation in IBR5 affects more to LR emergence than to specification (Fig 4B).

Epigenetic regulation of SKP2B
Here, we present evidences that SKP2B is regulated by novel mechanisms involving histone exchange and auxin-dependent acetylation. Supporting this role of histone acetylation in auxin response and root development, it has been shown that TSA, a histone deacetylase inhibitor, treatment partially rescues the LR formation defect in slr-1 (Fukaki et al., 2006). We have shown that TSA activated SKP2B expression in founder cells/LRP in the slr-1. Since SKP2B was initially expressed at the position of transference to TSA-containing medium and in the most apical part of the roots, we think that TSA is promoting founder cell specification and division of these cells in the slr-1 roots rather than activating the division of pre-specified founder cells. In addition, we show that a short-time TSA treatment induces the auxin response marker (DR5p:GUS) and SKP2B expression in the 1 6

Yeast One hybrid
The promoter region of SKP2B containing 410 bp upstream from the ATG was cloned into the gateway adapted pHISi-1 vector to prepare reporter yeast harboring HIS3. The construct was linearized with XhoI and 1 µg was used for yeast transformation using the strain Y187. The transformation was carried out as described in the Matchmaker protocol (Clontech, PT3529-1; http://www.clontech.com/images/pt/PT3529-1.pdf), using a cDNA library generated from mRNA isolated from auxin treated 5 days old Arabidopsis seedlings (kindly provided by W. Gray). The screening was carried out in a Dropout base medium without leucine and histidine and containing 5 mM of 3AT. Approximately 1.2 million yeast transformants were screened, and 10 positive clones were isolated, which were re-grown in a minimum medium containing 5, 10, and 20 mM of 3-AT. The DNA of these positive clones were PCR-amplified and sequenced.
The full length cDNA of HISTONE H3.1 and H3.3 were cloned in the pGAD42 to transform into the yeast strain containing the SKP2B[0.41]p version to test the activation potential of both proteins.

Chromatin immunoprecipitation (ChIP) assays
To determine the H3.3 and H3.1 enrichment on promoter, we used Arabidopsis transgenic plant expressing promoter and coding sequence of H3.1 (HTR13; At5g10390) or H3.3 (HTR5; At4g40040) fused in frame to Myc-tag and expressed under their own promoters (Stroud et al., 2012). For the ChIP assays, we used chromatin isolated roots of 7 day-old plants grown in MS agar plates in a 16-hour light 8-hour dark at 22ºC. The ChIP experiment was carried out as in 1 8

Root growth assays and microscopic analysis
Primary root length was determined as described previously by (Lucas et al., 2011). All data are the mean value of at least 50 plants, and these experiments were repeated twice, obtaining similar values in each experiment. Data values were statistically analyzed using the t-Student function. Total number and stages of LRP were counted according to methods used previously (Malamy and Benfey, 1997) and root meristem size was calculated based on the number of meristematic cortex cells (Casamitjana-Martinez et al. 2003).        (a-c) and one in the coding (d) were PCR amplified and separated in an agarose gel. As a control, the ChIP assays were carried out using an anti-IgG. B) Relative acetylation levels on SKP2B locus. ChIP assays of 7 day-old wild type (WT) or slr-1 mutant Arabidopsis roots treated with or without auxin (aux) using antibodies specific for diacetylated-H3. As a control, the ChIP assays were carried out using an anti-IgG. Quantitative PCR was used for relative quantification. The data was normalize to the levels in WT. a: p < 0,001; b: p<0,02; c: p<0,05 by two-sided t-test; n = 6. Error bars represent means ± standard error of the mean (SEM). C)

Supplemental Data Supplemental
Relative acetylation levels on promoters of root expressed genes. The data was normalize to the levels in WT. a: p < 0,001; b: p<0,02; c: p<0,05 by two-sided t-test; n = 6. Error bars represent means ± standard error of the mean (SEM