Atypical E2F activity coordinates PHR1 photolyase gene transcription with endoreduplication onset

In a microarray experiment that compared the transcriptome of wild‐type Arabidopsis plants with that of transgenic plants either overexpressing or knocked out for E2Fe/DEL1, the photolyase gene PHR1 was identified as one of the genes of which the expression level was significantly modulated (Lammens et al, 2008). To confirm this result, we investigated the PHR1 transcript level in control versus E2Fe/DEL1^OE and E2Fe/DEL1^KO leaves by quantitative real‐time PCR analysis. The transcript levels of PHR1 were ∼2.9‐fold higher and 4.0‐fold lower in the E2Fe/DEL1^KO and E2Fe/DEL1^OE lines than those of the wild‐type plants (Figure 1A). These data show that E2Fe/DEL1 operates as a transcriptional repressor of the PHR1 gene. Consistently, the PHR1 promoter holds a putative E2F‐binding site (AAACCGGC) located 459 bp upstream of its translation start site. To assess whether E2Fe/DEL1 controls the PHR1 transcription activity through direct binding to its promoter, a chromatin immunoprecipitation (ChIP) experiment was performed using an E2Fe/DEL1‐specific antibody (Lammens et al, 2008). E2Fe/DEL1^KO plants were used as a negative control. ChIP with primers spanning the E2F‐binding site clearly resulted in an enriched precipitation of the PHR1 promoter in wild‐type versus E2Fe/DEL1^KO samples (Figure 1B). Locus scanning by quantitative real‐time PCR with eight primer pairs covering the first 1366 base pairs of the PHR1 promoter showed a preferential association with the region comprised between −304 and −506 base pairs upstream of the start codon (Figure 1C), matching the position of the predicted E2F‐binding site. According to these results, E2Fe/DEL1 probably operates as a direct transcriptional regulator of the PHR1 gene.

As transcript abundance does not always reflect enzyme activity, photolyase activity levels were measured in the different transgenic lines by irradiating plants with 116 kJ/m² UV‐B (5 h treatment). After treatment, plants were either harvested immediately or after an additional 5‐h white light treatment, allowing CPD repair. The analysis focused on leaf 5, because it expressed E2Fe/DEL1 at the time the treatment was applied. In all three lines, the accumulation of CPDs was equal in plants harvested directly after the UV‐B treatment, illustrating that modified E2Fe/DEL1 levels do not provide any protective benefit (Figure 2). Five hours of recovery in white light allowed control plants to repair ∼34% of the CPDs, in contrast, to 51% and only 16% in E2Fe/DEL1^KO and E2Fe/DEL1^OE, respectively. These data demonstrate that the changes in PHR1 transcript due to knockout or overexpression of E2Fe/DEL1 are accompanied by changes in the CPD repair rate. The amount of CPDs of non‐treated plants did not differ significantly (Supplementary Figure S1), probably because under these conditions the PHR1 levels are not limiting to repair the low level of CPDs that occur in the absence of UV‐B radiation.

To test whether the observed changes in PHR1 transcript levels and CPD repair correlated with a differential response of the transgenic plants towards the UV‐B treatment, 18‐day‐old plants were transferred to a sun simulator where they were submitted to four different doses of UV‐B, corresponding to 35 kJ/m² (3 h treatment), 59 kJ/m² (5 h), 82 kJ/m² (7 h), and 106 kJ/m² (9 h), and photographed 9 days after the UV‐B irradiation. No phenotypic differences could be observed for the non‐treated plants (Figure 3A). Increasing doses of UV‐B caused macroscopic damage. Additionally, plants clearly showed a dose‐dependent decrease in size and biomass (Figure 3A and B). However, the dose response depended clearly on the genotype. Whereas the control displayed a reduced size and biomass at 35 kJ/m² and clear damage at 59 kJ/m², E2Fe/DEL1^KO plants performed clearly better at these doses. By contrast, the E2Fe/DEL1^OE plants showed a stronger growth reduction than control plants at 35 kJ/m². These results illustrate an increased tolerance and sensitivity towards UV‐B of E2Fe/DEL1^KO and E2Fe/DEL1^OE plants, respectively. A typical daily UV‐B dose on a mid‐latitude (48°N) summer day (at 500 m altitude) is 45 kJ/m² (own measurements), indicating that the E2Fe/DEL1 plants display a differential response in the range of naturally occurring UV‐B doses.

Effects on organ size can be due either to a reduced cell number or cell size, resulting from an inhibition of cell division or cell expansion, respectively. To distinguish between these possibilities, we measured the leaf size, abaxial epidermal cell size, and cell number of control and E2Fe/DEL1^KO plants. Measurements were done on the fifth leaf that still divided at the moment of the UV‐B treatment and, thus, expressed E2Fe/DEL1 (Lammens et al, 2008). Control and E2Fe/DEL1^KO plants were treated with 35 kJ/m² (3‐h treatment) of UV‐B and harvested after 9 days. Leaf size, cell number, and cell size were comparable under untreated conditions in E2Fe/DEL1^KO and wild type. The leaf size decreased after UV‐B treatment in both lines, but the effect was significantly (P‐value Student's t‐test <0.05) less for E2Fe/DEL1^KO, reaching 65.9% of the area of the non‐irradiated plants versus only 39.5% for wild‐type leaves (Figure 4A). The abaxial epidermal cell number was reduced by 42.3% in the UV‐B‐treated E2Fe/DEL1^KO leaves versus 55.4% in the control plants (Figure 4B), illustrating that the cell division in the E2Fe/DEL1^KO plants was less affected. Average cell size was slightly, but significantly (P‐value Student's t‐test <0.01) reduced in the control, but not in E2Fe/DEL1^KO (Figure 4C). Interestingly, after UV‐B treatment, the cell size of E2Fe/DEL1^KO plants was slightly, but significantly, larger than that of wild‐type plants (P‐value Student's t‐test <0.01). As cell size is in part determined by the DNA content (Melaragno et al, 1993; Sugimoto‐Shirasu and Roberts, 2003), DNA ploidy levels were measured through flow cytometry at a 3‐day interval after irradiation. As reported previously, under control conditions, the endoreduplication index (EI) of E2Fe/DEL1^KO plants increased faster than that of wild‐type plants as the leaf matured (Vlieghe et al, 2005; Lammens et al, 2008). UV‐B treatment delayed the increase in EI in both lines, but 12 days after irradiation, the EI level of the E2Fe/DEL1^KO plants regained control levels, whereas it remained reduced in wild‐type plants (Figure 4D). These data indicate that the absence of E2Fe/DEL1 helps in overcoming the inhibitory effect of UV‐B irradiation on the endocycle progression.

To test whether the sustained endoreduplication of the E2Fe/DEL1^KO plants contributes to their observed UV‐B tolerance, we measured the DNA ploidy level and size of individual pavement cells, both under control and UV‐B treated (35 kJ/m²) conditions. As observed for the complete leaf, UV‐B treatment resulted into a clear decrease in the relative abundance of pavement cells with 8C and 16C DNA content in control plants (Table I). By contrast, endoreduplication was stimulated in the E2Fe/DEL1^KO background, as exemplified by the increase in the 8C population. As reported before (Melaragno et al, 1993), within every DNA ploidy class a broad distribution in cell sizes was observed, reflecting ploidy‐independent growth (Table I). However, for both genotypes, the DNA content of a cell and its average size correlated clearly, with 16C cells ∼5–10‐fold larger than those with a 2C DNA content, indicating ploidy‐dependent growth (Table I). To estimate the relative contribution of each DNA ploidy class to the final size of the leaf, the proportion of cells within a given DNA ploidy class was multiplied with the average cell size. In wild‐type plants, the 8C and 16C cells accounted for ∼70% of the growth under non‐stress conditions, but their contribution dropped to 50% after UV‐B treatment (Table I). In contrast, the two highest DNA ploidy classes contributed for 64–70% to the E2Fe/DEL1^KO leaf size under both conditions, which can be attributed mainly to the high proportion of 8–16C cells. Thus, E2Fe/DEL1^KO plants apparently triggered a better growth performance after UV‐B treatment because of the presence of high‐ploidy cells, which on average are larger than the cells with a low‐ploidy level.

As the E2Fe/DEL1 gene is predominantly active in dividing tissues (Lammens et al, 2008), mainly leaves that are mitotically active would be expected to benefit from the absence of E2Fe/DEL1 at the moment of the UV‐B treatment. To test this hypothesis, we compared the recovery of a young leaf (leaf 5) with that of an older one (leaf 3). When the UV‐B treatment was applied, the fifth leaf was still predominantly dividing, whereas the third became less mitotically active, as quantified by their difference in E2Fe/DEL1 transcript levels (Figure 5A). When the reduction in leaf size was measured of irradiated versus untreated plants 9 days after the UV‐B treatment, both the young and old E2Fe/DEL1^KO leaves performed better than those of the wild‐type plants (Figure 5B), but the recovery was clearly more pronounced in the youngest leaf. Flow cytometric analysis indicated that the endocycle progression was clearly affected in both the young and old E2Fe/DEL1 leaves (Figure 5C and D). However, whereas the EI levels remained relatively low in the oldest leaf in the treated versus untreated plants during the first 9 days after irradiation, the EI increased steadily in the young leaf. These data indicate that the young leaves are more proficient in overcoming the inhibition of endoreduplication by UV‐B irradiation.

The recovery of the endoreduplication process in E2Fe/DEL1^KO leaves suggested that the PHR1 activity might aid endocycle progression. To test this hypothesis, the DNA ploidy levels of a control and a PHR1 knockout (PHR1^KO) line were compared. Eighteen‐day‐old plants were irradiated with 34 kJ/m² UV‐B and 9 days later, the fifth leaf was harvested for DNA ploidy analysis. In the absence of UV‐B, the EI did not differ significantly in the wild‐type and PHR1^KO lines (Figure 6). The UV‐B treatment affected the DNA ploidy distribution only marginally in the control background (Figure 6A and B). By contrast, endoreduplication was severely inhibited in the PHR1^KO plants (Figure 6C–E), indicating that the PHR1 expression is required to overcome the inhibitory effect of UV‐B on the endocycle progression.

Because of its observed link with PHR1, we tested whether E2Fe/DEL1 transcription is controlled by UV‐B radiation. Eighteen‐day‐old seedlings were irradiated with 35 kJ/m² UV‐B and harvested immediately after the treatment (5 h after onset) for RNA extraction. Within this short treatment, PHR1 levels increased approximately six‐fold (Figure 7A). PHR1 levels were higher in the E2Fe/DEL1^KO plants under both control and UV‐B conditions, but its level of induction was identical to that observed in control plants, indicating that E2Fe/DEL1 did not interfere with induction of PHR1 by UV‐B. In contrast, E2Fe/DEL1 levels were strongly downregulated by UV‐B treatment (Figure 7B).

UV‐B is a component of the sunlight in the biosphere. It can potentially damage plant membranes, proteins, and DNA. A recent sequence comparison between the genome of a reference strain with that of its descendants indicated that UV‐B radiation could be in part responsible for the overall mutation rate of 7 × 10⁻⁹ base substitutions per site per generation (Ossowski et al, 2010). Plants try to overcome the UV‐B‐induced DNA damage through photoreactivation by photolyases (Britt, 1999). Here, we identified E2Fe/DEL1 as a transcriptional repressor of PHR1 that controls the photolyase activity level of cells. E2Fe/DEL1 appears not to control the induction of PHR1 by UV‐B, but rather its basal expression level. Previously, we have found that E2Fe/DEL1 operates as a repressor of the endocycle onset through transcriptional repression of the APC activator gene CCS52A2 (Vlieghe et al, 2005; Lammens et al, 2008). As demonstrated, UV‐B inhibited both the mitotic cell cycle and endocycle, indicating that CPDs interfere with the polymerases that duplicate the genome in mitotically dividing and endocycling cells. When compared with those of control plants, both processes were significantly less inhibited in the E2Fe/DEL1^KO plants. Likely, the higher PHR1 levels under UV‐B conditions contributed to an accelerated removal of CPDs, probably contributing to a more efficient untangling of damaged DNA, and thus resulting in a fast resumption of DNA replication. Interestingly, E2Fe/DEL1 levels were quickly downregulated upon UV‐B treatment, suggesting a role in the DNA damage response (DDR). Mammalian atypical E2Fs (E2F7 and E2F8) have also been linked with DDR, steering the decision between cell survival and apoptosis of a DNA‐damaged cell. Low E2F1 levels stimulate DNA repair, whereas high E2F1 activity triggers the transcription of genes driving the apoptotic program (Lin et al, 2001; Stevens and La Thangue, 2003; Stevens et al, 2003). E2F1 appears to be a direct E2F7/E2F8 target gene and in the absence of both atypical E2Fs, cells massively undergo apoptosis (Li et al, 2008; Zalmas et al, 2008). As discussed previously (Cools and De Veylder, 2009), plants have no need for an apoptotic program to eliminate severely damaged cells due to the lack of cell migration. As plant germ cells originate from somatic lineages, transmission of damaged DNA might be prevented by stimulating cells to enter the differentiation pathway, safeguarding the progeny from DNA mutations. Recently, endoreduplication has been demonstrated to be an essential step in the differentiation of cells (Bramsiepe et al, 2010). Therefore, downregulation of E2Fe/DEL1 might be an initial trigger to eliminate potentially harmful dividing cells by promoting their differentiation through upregulation of CCS52A2 transcription, whereas the accompanying increase in PHR1 levels might ensure that endocycle progression can occur, contributing to plant growth (see below).

Endoreduplication has been suggested to aid in the protection against UV‐B radiation, as indicated by the increased UV‐B tolerance of the uvi4 mutant that displays an increased DNA ploidy level (Hase et al, 2006). The presence of multiple copies of crucial genes could reduce the risk of losing essential genes by gene mutation or help repair of damaged copies by homologous recombination with an intact copy as template (Kondorosi and Kondorosi, 2004). However, the latter seems to be unlikely, because alignment and cohesion of the replicated sister chromatids are rapidly lost in endoreduplicating cells (Schubert et al, 2006). Moreover, double‐stranded repair by homologous recombination is strongly inhibited in differentiated tissues (Boyko et al, 2006). Alternatively, endoreduplication might sustain growth of tissues by supporting cell expansion under conditions that prevent cell proliferation (Barow and Meister, 2003). The size of a leaf is determined by both cell proliferation and cell expansion. Under our experimental conditions, UV‐B treatment resulted in a decrease of both cell number and the average cell size. The small reduction in leaf size of the mutant E2Fe/DEL1^KO plants upon radiation correlated with an elevated percentage of high‐ploidy cells (8C and 16C), which can be explained by a fast resumption of the endoreduplication process after UV‐B treatment due to a high basal PHR1 level. Correspondingly, the endocycle progression was severely inhibited in PHR1‐deficient plants under UV‐B. As high‐ploidy cells are on average larger than low‐ploidy cells (2C and 4C), we speculate that E2Fe/DEL1^KO plants use the growth potential stored in their polyploid cells to compensate for the reduction in cell number. Efficient photorepair for the endocycle progression to aid growth might explain the need to coordinate the transcriptional control of PHR1 and CCS52A2 by E2Fe/DEL1, because coupling of DNA repair with replication gives the possibility of ploidy‐driven cell growth to compensate for the UV‐B‐caused reduction in cell number. In this manner, a maximum leaf surface can be obtained after UV‐B stress, allowing maximum photosynthesis to ensure sufficient energy to produce offspring.

All A. thaliana (L.) Heyhn. plants were of the Columbia‐0 (Col‐0) or Landsberg (Ler) ecotype. The transgenic plants E2Fe/DEL1^KO (Col‐0) and E2Fe/DEL1^OE (Col‐0) had been described previously (Vlieghe et al, 2005). The phr1 line (Ler) was obtained from the European Arabidopsis Stock Center (catalogue number N8325). Plants were grown under control conditions of a 16‐h photoperiod under white light exposure (TL‐D 36W840 tubes; Philips, Eindhoven, The Netherlands). For UV‐B treatment, Q‐Panel UV‐B‐313EL tubes were added. A 95‐μm thick cellulose acetate film (Kunststoff‐Folien‐Vertrieb GmbH, Hamburg, Germany) placed between the light source and the plants prevented any UV‐C radiation to reach the plants. Light and UV‐B intensities were measured with a SpectroSense2+ device (SKYE, Powys, UK) and equaled 70 μmol/m²/s and 1.9 W/m², respectively. For sun simulator treatments, the photosynthetically active radiation (PAR; 400–700 nm) was 172 μmol/m²/s (±5%) and the UV‐B irradiance 3.27 W/m² (±5%) under soda lime 4 mm Sanalux glass sheets (SCHOTT DESAG AG, Grünenplan, Germany). The climate parameters were adjusted to 75% relative humidity and 18°C temperature. Spectral irradiance was measured from 250 to 850 nm, with a double‐monochromator TDM300 (Bentham, Reading, UK) in steps of 1 and 2 nm in the UV‐B and visible range, respectively. During exposure, PAR, UV‐A, and UV‐B irradiances were continuously monitored. For more detailed information on the sun simulator treatments at the Helmholtz Zentrum München, see Döhring et al (1996) and Thiel et al (1996).

Plant material was chopped in 200 μl of Cystain UV Precise P Nuclei extraction buffer (Partec, Münster, Germany), supplemented with 800 μl of staining buffer. The mix was filtered through a 50‐μm green filter and read through the Cyflow MB flow cytometer (Partec). The nuclei were analysed with the FloMax software. The EI was calculated from the percentage values of each ploidy class with the formula: EI=[(0 × %2C)+(1 × %4C)+(2 × %8C)+(3 × %16C)+(4 × %32C)]/100.

Two‐week‐old plants were submitted to UV‐B radiation, then left to recover for 5 h in white light, and harvested before or after the recovery. DNA was extracted with the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany), titrated, and processed with the protocol for MBL D194‐1 anti‐CPD monoclonal antibody and ECL anti‐mouse IgG horseradish peroxidase‐linked whole antibody from sheep.

Plants were grown on half‐strength Murashige and Skoog medium, 0.8% agar at 20°C and 16 h photoperiod. RNA was extracted with the RNeasy Plant Mini Kit (Qiagen) and cDNA was synthetized with the SuperScript Reverse Transcriptase kit (Invitrogen, Carlsbad, CA). Quantitative PCR ran with a LightCycler (Roche) with the SYBR Safe enzyme reaction mix (Roche). The primers were forward PHR1, CTCGTTGGAGCAGTTAGAGAAGG; reverse PHR1, TGAACGCCACATATAGACCACATG; forward DEL1, TCTTCTCAGCCTCACTCTAGTAGC; and reverse DEL1, GCCGCCTCACTTTAGTTCGC. Actin was used as a control with forward and reverse primers GGCTCCTCTTAACCCAAAGGC and CACACCATCACCAGAATCCAGC, respectively.

Eight days after sowing, whole seedlings were grown on half‐strength Murashige and Skoog medium and 0.8% agar for ChIP as described (Bowler et al, 2004). Actin was used as a normalization gene with the following primers: forward, GGCTCCTCTTAACCCAAAGGC; reverse, CACACCATCACCAGAATCCAGC. The primers used to scan the PHR1 promoter were CATTACGAAGGAGGAGAGACGG and GCAATCACAGTCTGACACGTGTTG (0–215), CAACACGTGTCAGACTGTGATTGC and CCCGAACATCTATACACATTACAC (191–324), GTGTAATGTGTATAGATGTTCGGG and GTTAGTTACGCTGTTAGGAGAGC (304–506), GCTCTCCTAACAG‐CGTAACTAAC and CTCTTGAAATGGTCGTTTTATACCTCAAC (487–746), GTTGAGGTATAAAACGACCATTTCAAGAG and CTGTTAGGACATTTTGGACCATAATGC (682–836), GCATTATGGTCCAAAATGTCCTAACAG and GAGTAAACCCTAGATAAGCACGATC (809–1049), GATCGTGCTTATCTAGGGTTTACTC and GAAAGAGTGATTATCATTCCTCATGATATG (1025–1164), CATATCATGAGGAATGATAATCACTCTTTC and CATACGCTCACGCTAGTCGTTTG (1135–1366).

The protocol for epidermal peels was adapted from Melaragno et al (1993). Leaf 5 of 5‐week‐old plants were fixed in a solution containing ¾ of 95% ethanol and ¼ glacial acetic acid for 2 h at room temperature, stored in 70% ethanol at 4°C. Fixed tissue was soaked in water. Leaves were placed with the abaxial epidermis side down on a glass slide and held in place with forceps, while the tissue was removed from the adaxial epidermis with another pair of forceps. A drop of 4′,6‐diamino‐2‐phenylindole at a concentration of 0.005 mg/ml in McIllvaines's buffer, pH 4.1 (60 ml 0.1 M citric acid+40 ml 0.2 M Na₂HPO₄), was placed on the epidermal peel. Peels were observed under a × 20 objective on an Axioskop equipped with an Axiocam CCD camera (Zeiss). Images were obtained with the Axiovision software and were analysed with the public domain image analysis program ImageJ (version 1,40; http://rsb.info.nih.gov/ij/). Relative fluorescence units were measured as integrated density, which are the product of the area and the average fluorescence of the selected nucleus. Ten guard cell nuclei were used as a 2C control for each image. The growth contribution of each ploidy level to the final leaf surface (or LGC) was calculated according to the following formula:

where n is the maximum observed DNA ploidy classes in which i can take the value 2, 4, 8, 16, …, p is the proportion of nuclei in a given ploidy class, and a is the average area associated to a given ploidy level.

Supplementary data are available at The EMBO Journal Online (http://www.embojournal.org).