Inactivation of the ELIP1 and ELIP2 genes affects Arabidopsis seed germination


Author for correspondence:
Paola Vittorioso
Tel: +39 06 49912922


  • Light regulates Arabidopsis seed germination through the phyB/PIL5 (PHYTOCHROME INTERACTING FACTOR 3-LIKE 5) transduction pathway, and we have previously shown that the Dof transcription factor DOF AFFECTING GERMINATION1 (DAG1) is a component of this pathway.
  • By means of microarray analysis of dag1 and wild type developing siliques, we identified the EARLY LIGHT-INDUCED PROTEIN1 and 2 (ELIP1 and ELIP2) genes among those deregulated in the loss-of-function dag1 mutant.
  • We analysed seed germination of elip single and double mutants, of elip dag1 double mutants as well as of elip1 elip2 dag1 triple mutant under different environmental conditions.
  • We show that ELIP1 and ELIP2 are involved in opposite ways in the control of this developmental process, in particular under abiotic (light, temperature, salt) stress conditions.


The Dof family of transcription factors is present throughout the plant kingdom, including the algae, mosses and ferns (Lijavetzky et al., 2003; Moreno-Risueno et al., 2007) but absent from yeast and animals (for a review, see Yanagisawa, 2002).

The DOF transcription factors have been shown to participate in the regulation of vital processes exclusive to plants. In particular, several Dof proteins are involved in light-responses and light-regulated processes (Yanagisawa & Sheen, 1998; Park et al., 2003; Imaizumi et al., 2005; Ward et al., 2005). We have shown that the DOF AFFECTING GERMINATION 1 (DAG1) protein is involved in light-dependent seed germination in Arabidopsis (Papi et al., 2000; Gualberti et al., 2002). It is known that dag1 knock-out mutant seeds require lower R light fluence rates than wild-type (WT) seeds to germinate (Papi et al., 2002). More recently, we have shown that in the light-dependent phyB-mediated pathway leading to seed germination, DAG1 acts downstream of PHYTOCHROME INTERACTING FACTOR 3-LIKE 5 (PIL5), and negatively regulates gibberellin (GA) biosynthesis by directly repressing the AtGA3ox1 GA biosynthetic gene (Gabriele et al., 2010).

The EARLY LIGHT-INDUCED PROTEIN (ELIP) family consists of >100 stress-related proteins in prokaryotic and eukaryotic photosynthetic organisms. The ELIPs are nuclear-encoded, CAB (CHLOROPHILL a/b BINDING)-related proteins, localized in the chloroplast thylakoid membranes (Grimm et al., 1989). Unlike CAB proteins, which are constitutively expressed in thylakoids, ELIP proteins show a transient accumulation in response to environmental stress (Adamska, 2001). The ELIPs were first discovered to be transiently expressed during the greening of etiolated plants (Grimm et al., 1989). Both ELIP mRNAs and proteins appear considerably faster than other light-induced genes during the early stage of de-etiolation, and disappear before chloroplast development is completed (Meyer & Kloppstech, 1984; Grimm & Kloppstech, 1987). The ELIP proteins are also present in plant tissues that undergo desiccation, such as fern spores (Adamska, 2001) and Craterostigma plantagineum leaves (Alamillo & Bartels, 2001). In Tortula ruralis, a model plant for the study of response to water deficit-stress, two ELIP genes (ELIPa and ELIPb) have been shown to be regulated by a variety of stress, such as slow desiccation, rapid-dessication/rehydration, salinity, ABA and rehydration under high light levels (Zeng et al., 2002).

Two closely-related ELIP proteins (ELIP1 and ELIP2, with 81% aminoacid identity) are present in Arabidopsis. Both proteins are differentially expressed in response to light stress in green and senescing leaves (Heddad et al., 2006). It has been suggested that ELIP proteins play a photoprotective role under light-stress conditions, either by transiently binding to free chlorophyll, thus preventing photo-oxidative stress (Hutin et al., 2003), and/or by participating in energy dissipation to protect the photosystem II (PSII) reaction centre from photoinhibition (Montané & Kloppstech, 2000). However, elip1 elip2 double mutants show only a slight reduction in chlorophyll content, indicating that inactivation of both ELIP genes does not affect substantially the response to photoinhibition and photo-oxidative stress (Casazza et al., 2005; Rossini et al., 2006). By contrast, analysis of Arabidopsis lines overexpressing ELIP genes suggests that these proteins reduce the chlorophyll content by inhibiting chlorophyll biosynthesis rather than increasing chlorophyll degradation (Tzvetkova-Chevolleau et al., 2007). These data on the overexpressing lines are not in agreement with the behaviour of the double mutant line and it has been hypothesized that some ELIP-like proteins such as HLIPs (HIGH LIGHT-INDUCED PROTEINS) and SEPs (STRESS-ENHANCED PROTEINS) may compensate the loss of ELIP functions in the double mutant (Tzvetkova-Chevolleau et al., 2007).

Recently, the ELIP genes have been suggested to participate in the phytochrome signalling pathway leading to seed germination in tomato. Tomato seeds can germinate in darkness as a result of Pfr present in the dry seeds. Continuous irradiation with far-red light (FRc) can inhibit this process by converting Pfr in Pr through a high irradiance response (HIR) (Schichjo et al., 2001). Transcriptome analysis of seeds exposed to a FRc inhibitory treatment, with or without a subsequent red light pulse, revealed that the single tomato ELIP gene is upregulated by FRc and downregulated by a red light pulse (Auge et al., 2009).

In this work, we provide evidence that the ELIP genes of Arabidopsis are involved in seed germination and that they act downstream of the Dof protein DAG1, although they are not its direct regulatory targets. Our data also suggest that ELIP1 and ELIP2 may be involved in regulating seed germination under abiotic stress conditions. Moreover, we present molecular data showing that the ELIP genes are differentially regulated by light in seeds.

Materials and Methods

Plant material

The dag1 allele shown in this work is the dag1-1 allele described in Papi et al. (2000) in the Wassjilieskja (Ws-4) ecotype. The elip1, elip2 and elip1elip2 mutants (Col-0) were kindly provided by C. Soave (Casazza et al., 2005). The double mutant elip1 dag1 and elip2 dag1 and the triple mutant elip1 elip2 dag1 were obtained by crossing the single mutants or the double mutant elip1 elip2, growing seeds on kanamycin-containing medium, and identified in the F3 generation by PCR analysis using specific primer pairs for the mutations, listed in the Supporting Information Table S1. As the single mutants were in different ecotypes, the parental lines (dag1, elip1, elip2 and elip1elip2) and the wild type were also selected from the crosses. Several lines for each genotype were selected and analysed in order to minimize the effect of the two different ecotypes on the phenotypes of interest.

Growth conditions and light sources

All Arabidopsis thaliana lines used in this work were grown in a growth chamber at 24°C : 21°C with 16 h : 8 h, day : night cycles and a light intensity of 300 μmol m−2 s−1 as previously described (Papi et al., 2000). For all monochromatic experiments, seeds were handled under a dim-green safelight. For the seed germination assays under high–low light, seeds were exposed to four lamps Osram Xenophot 64653 HLX ELC 24 V 250 W regulated (Arbeitsgruppe Technik, Institut für Biologie II, Unirversitaet Freiburg, Germany), respectively, at 1000 or 22 μmol m−2 s−1. Red light was applied using eight fluorescent tubes (Philips F32T8/TL741, CLF Plant Climatics GmbH, Emersacker, Germany) and a ‘red fire’ filter (LEE Filters, R.E.C., Rome, Italy) to select the appropriate wavelength (670–690 nm). Far-red light was applied using six incandescent bulbs (40 W, Radium Lampenwerk GmbH, Wipperfurth, Germany), adding a ‘blood red’ and a ‘blue medium’ filters (LEE Filters) to select the appropriate wavelength (700–775 nm).

Seed germination assay

All seeds used for germination tests were harvested from mature plants grown at the same time, in the same conditions, and stored for 4–5 wk in the dark under dry conditions at room temperature. For seed germination assays, triplicate sets of 60–100 nonsterilized seeds for each genotype were sown on five layers of filter paper 595 (Schleicher & Schüll, Dassel, Germany), soaked with 5 ml water, under dim-green safe light. All germination assays have been performed with different seed batches. For the NaCl germination assay, sterilized seeds were sown in Petri dishes on 0.8% agar dissolved in water containing increasing concentrations of NaCl (Sigma).

Synthesis of biotinylated cRNA

Each RNA preparation was tested for degradation using the Agilent 2100 Bioanalyzer (Agilent technologies, Palo Alto, CA, USA). cDNA was synthesized from 4 μg of total RNA using one-cycle target labelling and control reagents (Affymetrix, Santa Clara, CA, USA) to produce biotin-labelled cRNA. The cRNA preparation (15 μg) were fragmented at 94°C for 35 min into 35–200 bases in length.

Hybridization, washing and scanning of microarrays

Three biological replicates for each condition were independently hybridized. If the quality control was correct, then 10 μg of fragmented cRNA were hybridized to the Arabidopsis ATH1 Genome array (Affymetrix), containing 22 500 transcript variants from 24 000 well-characterized Arabidopsis thaliana genes. Each sample was added to a hybridization solution containing 100 mM 2-(N-morpholino) ethanesulphonic acid, 1 M Na+, and 20 mM of EDTA in the presence of 0.01% of Tween-20 to a final cRNA concentration of 0.05 μg ml−1. Hybridization was performed for 16 h at 45°C. Each microarray was washed and stained with streptavidin-phycoerythrin in a Fluidics station 450 (Affymetrix) and scanned at 2.5 μm resolution in a GeneChip Scanner 3000 7G System (Affymetrix) Data analyses were performed using GeneChip Operating Software (GCOS).

Data analysis

Analysis was performed using the affylmaGUIR package (Wettenhall et al., 2006). Robust multi-array analysis (RMA) algorithm was used for background correction, normalization and expression levels summarization (Irizarry et al., 2003). Next, differential expression analysis was performed with the Bayes t-statistics from the linear models for microarray data (limma), included in the affylmGUI package. P-values were corrected for multiple-testing using the Benjamini–Hochberg’s method (False Discovery Rate; Benjamini & Hochberg, 1995; Reiner et al., 2003). Genes were considered to be differentially expressed if the corrected P-values were < 0.05. In addition, only genes with a signal log ratio > 1 or < −1 were considered for further analysis.

Transcript analyses

For real-time reverse-transcription polymerase chain reaction (RT-PCR) analyses, total RNA was extracted, purified and the cDNA synthesized from imbibed seeds as described in Penfield et al. (2005). Real-time PCR was performed with SYBR-green I master using the LightCycler 480 instrument (Roche). A 1 μl sample of the diluted cDNA was used, along with the specific primers, listed in Table S1. Relative expression levels were normalized with that of the At5g60390 gene. Each point represents the mean and SD of two independent biological repeats.

DAG1-HA construct and chromatin immunoprecipitation (ChIP) analysis

The DAG1-HA recombinant clone was obtained as described in Gabriele et al. (2010). Chromatin immunoprecipitation was performed according to the protocols described by Saleh et al. (2008) and Bowler et al. (2004), with 12 h DAG1-HA imbibed seeds. Antibodies against HA tag (for DAG1-HA) and myc tag (negative control) were used for immunoprecipitation. Equal amounts of starting plant material and ChIP products were used for PCR reaction. The primers used are listed in Table S1. The ChIP experiments were performed two to four times with biological replicates.


DAG1 inactivation affects the expression of ELIP1 and ELIP2

In pursuing the identification of targets of the transcription factor DAG1, a microarray study was conducted using an Arabidopsis ATH1 Affymetrix array and analysing global gene expression changes in developing siliques of wild type and of the dag1 mutant (Papi et al., 2000). Siliques from stages F12 to F16 (Smyth et al., 1990), where we had previously shown that the gene DAG1 is strongly expressed (Papi et al., 2000), were sampled for microarray hybridizations. Consistency of the microarray data was assessed by comparing the mean normalized signal intensities obtained from three independent biological replicates. Among the several hits identified, we focused our attention on the ELIP1 and ELIP2 genes as microarray data available at the Array Gene Chips collection ( indicated that the ELIP genes are highly expressed in seeds. In particular, ELIP1 is strongly expressed in dry seeds and during the first hours of imbibition (NASCARRAYS-195;, whereas both ELIP1 and ELIP2 are present in germinated embryo (NASCARRAYS-386;

Mutations in the ELIP genes impairs seed germination

As DAG1 acts as a negative regulator of seed germination (Papi et al., 2000; Gabriele et al., 2010), we set out to assess whether its putative targets ELIP1 and ELIP2 would also be involved in this developmental process. We performed seed germination assays on the elip1 and elip2 loss-of-function single mutants (Casazza et al., 2005), and the elip1 elip2 double mutant (Rossini et al., 2006).

Seed germination assays were performed under standard illumination conditions, on seeds grown, harvested and stored at the same time and under the same conditions. As shown in Fig. 1(a), germination of both elip1 and elip2 single mutant seeds was slightly slower than germination of the wild type (Col-0) seeds at 24 h (16% and 17% for elip1 and elip2, respectively, compared with 47% for the wild type). Interestingly, elip1 elip2 double mutant seeds show an even lower germination rate, as 50% germination was attained after 48 h.

Figure 1.

Germination analysis of the elip1 and elip2 mutant seeds. (a) Germination rates of elip1(open squares), elip2 (closed circles), and elip1elip2 double mutant (open triangles) seeds compared with the wild type (open circles). Arabidopsis seeds were grown for 5 d under white light, in standard conditions. (b) Germination rates of elip1, elip2 and elip1elip2 mutant seeds grown for 5 d in total darkness. Error bars represent the SD of three biological repeats. *,  0.05; **,  0.005. P-values were obtained from a Student’s unpaired two-tail t-test comparing the mutant/transformant with its control.

Seed germination assays in total darkness without a light pulse, confirmed the lower germination rate of elip mutant seeds: only up to 5%elip1 seeds germinated, while elip2 and elip1 elip2 double mutant seeds did not germinate at all, compared with c. 35% germination of wild-type Col-0 seeds (Fig. 1b).

As both ELIP genes are deregulated in the dag1 mutant, we produced the elip1 dag1 and elip2 dag1 double mutants and analysed seed germination to assess whether inactivation of DAG1 would (at least in part) compensate for the seed germination impairment caused by loss of elip1 and elip2. As the single mutants were in different ecotypes (Col-0 and Ws), the parental lines (dag1, elip1, elip2 and elip1 elip2) and the wild type were also selected from the crosses. Moreover, as it has been proposed that the ELIP proteins might be functionally redundant (Rossini et al., 2006), we isolated also the elip1 elip2 dag1 triple mutant. Three independent lines of the double and triple mutants were isolated and analysed for seed germination under normal light conditions and in total darkness. As the three lines of each multiple mutant showed substantially the same behaviour, only the results of one representative line for both elip1 dag1, elip2 dag1 and elip1 elip2 dag1 are shown in Fig. 2. Germination assays under standard illumination conditions revealed that the germination rate of elip1 dag1 double mutant seeds was similar to that of wild-type (Col-0/Ws) seeds, indicating that inactivation of DAG1 in the elip1 background compensates for the slight delay of elip1 mutant seeds. By contrast, the kinetics of germination of elip2 dag1 double mutant seeds was comparable to that of elip2 seeds, indicating that inactivation of DAG1 has different effects in the elip1 and elip2 backgrounds. Interestingly, germination of the elip1 elip2 dag1 triple mutant was almost complete in the first 48 h, compared with c. 55% of wild type Col-0/Ws seeds (Fig. 2a). We had shown that in total darkness dag1 seeds are still able to germinate to some extent, while wild-type Ws seeds are not (Papi et al., 2000). Seed germination assays in total darkness revealed that the elip1 elip2 dag1 triple mutant seeds germinated up to 35%, elip1 dag1 double mutant seeds up to 12% while, as expected, wild-type Col-0/Ws seeds did not germinate at all (Fig. 2b). The germination rate of elip1 dag1 seeds was close to that of dag1 seeds (18%), indicating that loss of DAG1 compensates the loss of ELIP1.

Figure 2.

Germination analysis of the double mutants elip1 dag1 and elip2 dag1 and of the triple mutant elip1 elip2 dag1. (a) Germination rates of elip1 dag1 (closed diamonds), elip2 dag1 (open diamonds) and of elip1 elip2 dag1 triple mutant (closed triangles) seeds compared with dag1 (closed squares) and wild-type (open circles) seeds. Arabidopsis seeds were grown for 5 d under white light, in standard conditions. (b) Germination rates of elip1 dag1 and elip2 dag1 double mutants and of elip1 elip2 dag1 triple mutant seeds grown for 5 d in total darkness. Error bars represent the SD of three biological repeats. *,  0.05; **,  0.005. P-values were obtained from a Student’s unpaired two-tail t-test comparing the mutant/transformant with its control.

Expression of ELIP1 and ELIP2 is controlled by the same environmental (stress) factors – light, temperature and dehydration (Bartels et al., 1992; Adamska & Kloppstech, 1994; Montanéet al., 1997; Zeng et al., 2002) – that trigger seed germination (Bewley, 1997). To assess the effect of loss of ELIP function on seeds germinating under stress conditions, we performed seed germination assays on the elip1 and elip2 single mutants, on the elip1 elip2 double mutant and on the elip1 elip2 dag1 triple mutant, under different stress combinations: high/low light intensity and high/low temperature. Seeds were pretreated for 5 min with far-red light (25 μmol m−2 s−1), then exposed to a 5 min pulse of high- or low-intensity white light (1000 μmol m−2 s−1 and 22 μmol m−2 s−1, respectively). Subsequently, seeds were incubated at high (30°C), low (11°C) or normal growth temperature (22°C) in the dark for 5 d. As shown in Fig. 3, low temperature did not affect seed germination, either coupled with low- or high-intensity light pulse, as the germination rate of wild type Col-0/Ws seeds incubated at 11°C or 22°C was similar. Conversely, high temperature, coupled with both a low- or a high-intensity light pulse, greatly affected seed germination, as germination of wild-type seeds was from 0 (low-intensity pulse) to 4% (high-intensity pulse). The elip1 and elip2 single mutants showed a different behaviour under both light treatments, at 22°C, with elip2 seeds germinating up to twofold more than elip1 seeds (65.3% vs 33% under low light, 78,6% vs 40% under high light). Interestingly, the triple mutant elip1 elip2 dag1 was the only line capable of germinating after incubation at 30°C to 40% and 60% with a low- or high-intensity light pulse, respectively. Moreover, triple-mutant seeds showed a higher germination rate than wild-type seeds under both high light (96% vs 60%) and low light (88% vs 24%) at 22°C.

Figure 3.

Germination analysis of the elip mutants under light stress conditions. (a) Germination rates of elip1 (dark tinted bars), elip2 (light tinted bars), elip1 elip2 (open bars), double mutant and of elip1 elip2 dag1 (graded tinted bars), triple mutant seeds, compared with wild-type seeds (closed bars), under low light (22 μmol m−2 s−1). (b) Germination rates of elip1, elip2, elip1 elip2 double mutant and of elip1 elip2 dag1 triple mutant seeds compared with wild-type seeds, under high light (1000 μmol m−2 s−1). Arabidopsis seeds were treated for 5 min with low or high light, then grown for 5 d at three different temperatures (30°C, 22°C and 11°C) in darkness. Error bars represent the SD of three biological repeats.

It has been reported that the ELIP genes can be induced also by high salinity (Zeng et al., 2002), so we performed germination assays under different NaCl concentrations (0, 50, 75 and 100 mM), incubating seeds of the elip1 and elip2 single, elip1 elip2 double and elip1 elip2 dag1 triple mutant lines in the dark for 5 d at high (30°C) and low (11°C) temperatures or at 22°C, after a pulse of red light to trigger germination. At 22°C, increasing concentrations of NaCl reduced the germination rate of wild-type seeds and of all mutant seeds (Fig. 4a). The elip2 single mutant seeds were more tolerant to salt stress than the Col-0/Ws wild type at the three different temperatures. Conversely, the germination rate of the elip1 mutant seeds was substantially lower than the wild type even in the absence of NaCl, as a higher fluence rate (or a longer light treatment) was needed to induce seed germination.

Figure 4.

Germination analysis of the elip mutants under salt stress conditions. Germination rates of elip1 (dark tinted bars), elip2 (light tinted bars), elip1 elip2 double mutant (open bars), and of elip1 elip2 dag1 triple mutant seeds (graded tinted bars), compared with wild type seeds (closed bars), on increasing NaCl concentrations (0, 50, 75, 100 mM). Arabidopsis seeds were pulsed with red light to induce germination, then grown for 5 d in darkness at three different temperatures: (a) 22°C, (b) 30°C and (c) 11°C. Error bars represent the SD of three biological repeats.

Seeds of the elip1 elip2 dag1 triple mutant were still able to germinate up to 30% even at 100 mM NaCl. At 30°C, elip1 elip 2dag1 seeds had a germination rate of 81% at 75 mM NaCl and of 22% at 100 mM NaCl, compared with the null germination rate of wild-type seeds already at 75 mM NaCl (Fig. 4).

These results indicate that the elip mutations affect tolerance to light and salt stress, with elip2 causing an increase of seed germination and elip1 resulting in the opposite phenotype. Moreover, inactivation of DAG1 in the elip1 elip2 background further increases the germination rate under abiotic stress conditions.

Expression of ELIP1 and ELIP2 in Arabidopsis seeds

Expression of the ELIP genes has been extensively analysed only in seedlings and in adult plants (Harari-Steinberg et al., 2001; Casazza et al., 2005; Heddad et al., 2006), thus we analysed their expression in seeds. As DAG1 is involved in the phyB-mediated seed germination process, analysis of ELIP expression was performed by quantitative RT-PCR on RNA extracted from wild-type and dag1 seeds imbibed 24 h under white light, red light or in darkness, to assess whether the ELIP genes are under the control of phyB. In addition, we analysed the amount of the ELIP1 transcript in seeds from the elip2 dag1 and of the ELIP2 transcript in seeds of elip1 dag1.

The results shown in Fig. 5 indicate that in wild-type seeds the level of expression of the ELIP genes is about the same under white and red light and in darkness (Fig. 5). Interestingly, the loss of DAG1 results in a fourfold upregulation of ELIP1 under red light, while simultaneous loss of DAG1 and ELIP2 enhances ELIP1 transcription by c. 50% (Fig. 5a). ELIP2 transcription is only substantially affected in elip1 dag1 double mutant seeds under white light (Fig. 5b), suggesting that ELIP2 may respond to phytochromes other than phyB. In order to verify whether the ELIP2 gene could be regulated through phyA, we also performed the expression analysis on seeds imbibed 24 h under far-red light (Fig. 5a,b). The results indicate that ELIP2 is likely to be regulated through phyA, and that inactivation of DAG1 and of both DAG1 and ELIP1, dramatically decreases its expression level.

Figure 5.

Expression analysis of the ELIP1 and ELIP2 genes in Arabidopsis seeds. Relative expression level of the ELIP1 (a) and ELIP2 (b) genes in wild type, dag1 and, respectively, elip2 dag1 and elip1 dag1 seeds, imbibed for 24 h, exposed to white light (WL) red light (RL), far-red light (FR) or kept in darkness (D). Relative mRNA levels of each gene are presented by the ratio of the corresponding mRNA level of wild type WL, RL, FRL or D which was set to 1. Error bars represent the SD of two biological repeats.

DAG1 does not bind the ELIP1 and 2 promoters

In silico analysis by means of Promomer ( (Toufighi et al., 2005) suggested that the promoters of the ELIP genes are poor candidates as Dof transcription factor targets compared with the positive control AtGA3ox1, which we have recently shown to be a target of, and to be bound by, DAG1 (Gabriele et al., 2010) (Table 1). Chromatin immunoprecipitation assays using the dag1 mutant lines overexpressing the DAG1-HA chimeric protein (Gabriele et al., 2010) shown in Fig. 6 did not lead to amplification of specific bands. These results indicate that DAG1 does not bind directly to the ELIP1 and ELIP2 promoters.

Table 1.   The ELIP1 and ELIP2 promoter sequences do not show a significant high number of Dof CTTT binding sites
GeneAGI ID codeSet AveaSignificanceb
  1. aThe number is the CTTT occurrence in the gene promoter.

  2. bThe significance value is 1 – the percentile occurrence of the promoter’s CTTT count relative to all promoters in the genome. Values ≤ 0.050 are statistically significant.

Figure 6.

Chromatin immunoprecipitation (ChIP)-PCR analysis of the ELIP1 and ELIP2 promoters. (a) Graphic representation of the ELIP1 and ELIP2 promoters. Underlying thick lines marked by uppercase letters indicate various promoter fragments used for PCR. (b) In vivo formaldehyde crosslinking followed by immunoprecipitation and PCR amplification. The primer sequences used for ChIP PCR are shown in the Supporting Information, Table S1.


In Arabidopsis, seed germination is mediated mainly by phyB (Shinomura et al., 1994), and the transcription factor PIL5 is the key regulator of this developmental process (Oh et al., 2004, 2007). The Dof protein DAG1, previously characterized as a transcription factor involved in the maternal control of seed germination (Papi et al., 2000, 2002; Gualberti et al., 2002) was recently shown to be a component of the phyB/PIL5 signalling, acting as a negative regulator of the GA biosynthetic gene AtGA3ox1 (Gabriele et al., 2010).

In this work, transcriptome analysis of the dag1 mutant identifies ELIP1 and ELIP2 among the genes acting downstream of DAG1. The ELIP genes are strongly expressed in imbibed seeds (ELIP1) and in germinating embryos (ELIP1 and ELIP2), as revealed by analysis of available seed microarray data on NASC ( However, the ELIP genes are unlikely to be DAG1 direct targets because ChIP experiments indicated that the latter does not bind to ELIP promoters. However, in silico analysis revealed the presence of one G-box element (CACGTG) in the ELIP1 promoter (−290 from the ATG), as well as multiple copies of E-box elements (CANNTG) in both ELIP1 (−890, −839, −832 and −394 from the ATG) and ELIP2 (−809, −730 and −432 from the ATG) promoters, which serve as binding sites for many bHLH transcription factors (Menkens et al., 1995), such as the PIF and PIL proteins, which are known to be involved in several phytochrome-mediated pathways (Castillon et al., 2007). In addition, several WRKY binding sites are present in both promoters (−635, −627, minus strand, −480 and −275 from the ATG in ELIP1; −912, −720 and −436 from the ATG in ELIP2). The WRKY plant transcription factors, such as the Dof proteins, appear to be involved in the regulation of various physiological programs unique to plants, such as seed germination (Eulgem et al., 2000). An example is the WRKY2 transcription factor that mediates seed germination and post-germination developmental arrest by ABA (Jiang & Yu, 2009). Interestingly, in barley aleurone cells, the WRKY protein HvWRKY38 works in a complex with the Dof protein BPBF in repressing gibberellin signalling (Zou et al., 2008). At the moment we cannot rule out the possibility that DAG1 may regulate the ELIP genes by acting in a complex with other transcription factors, possibly only under particular abiotic stress conditions. In any case, although we do not know yet how DAG1 regulates/affects expression of the ELIP genes, the seed germination phenotype of the elip mutants establishes de facto a relationship between DAG1 and the ELIP genes. It will be of interest in the future to uncover the molecular basis of DAG1-mediated ELIP expression.

The nuclear-encoded ELIP genes encode chloroplast thylakoid membrane proteins and are regulated at the transcriptional and post-transcriptional level by light and abiotic stress (Harari-Steinberg et al., 2001). However, the function of the ELIP proteins is still unknown (Casazza et al., 2005; Rossini et al., 2006).

We show here that loss-of-function mutations of the ELIP genes affects the seed germination process. Under standard conditions, seed germination of the elip1 and elip2 mutants, as well as of the elip1 elip2 double mutant is only slightly delayed for the first 48 h, but the percentage of germinating seeds is unchanged compared with wild-type Arabidopsis.

By contrast, germination of elip mutant seeds under light, temperature and salinity stress is substantially altered. Interestingly, mutations of ELIP1 and ELIP2 influence seed germination with opposite effects: elip1 seeds germinate less, while elip2 seeds germinate (up to threefold) more than the wild type.

Inactivation of DAG1 has different effects in the elip1 and elip2 background, as far as seed germination, as only the elip1 dag1 double mutant shows the same behaviour of dag1 mutant seeds, under normal germination conditions. Moreover, loss of DAG1 in the elip1 elip2 background further increased the germination potential compared with elip1 elip2 double mutant seeds under normal and stress conditions, suggesting that the dag1 mutation is epistatic over the elip mutations.

Our data also suggest that expression of the ELIP1 and ELIP2 genes in seeds is regulated by different phytochromes, as expression of ELIP1 is deregulated only under red light whereas ELIP2 is strongly reduced under far-red light. It has been shown that the contribution of phyB to germination decreases with prolonged exposure to cold, whereas phyA may have a role in promoting germination at high temperature and phyE at low temperature (Heschel et al., 2007). In this perspective, the fact shown here that germination of elip1 mutant seeds is heavily affected at high and low temperatures suggests that ELIP1 may be mostly responding to phyB, whereas ELIP2 may respond also to phyA and phyE, as elip2 mutant seeds germinate up to 70–80% at 30°C and 11°C, respectively. In agreement with the germination data, the expression analysis also suggest that loss of DAG1 has opposite effects on the expression of the ELIP genes: while ELIP1 is upregulated in dag1 mutant seeds, ELIP2 expression is strongly reduced both in dag1 and elip1 dag1 mutant seeds.

The alterations in seed germination observed in elip mutants may be accounted for by altered chloroplast function. Chloroplasts seem necessary for seed germination: while immature Arabidopsis embryos contain photosynthetically active chloroplasts, chlorophyll and thylakoid membranes are lost during seed dehydration, and exposure to light induces regreening during seed germination (Waters & Langdale, 2009). Accordingly, the snowy cotyledon (sco) mutant, lacking a plastid-targeted elongation factor G (SCO1) essential for chloroplast development, shows delayed germination (Ruppel & Hangarter, 2007). In addition, PIL5, the master regulator of phyB-mediated seed germination, has been shown to control chlorophyll biosynthesis (Moon et al., 2008).

As chloroplasts are necessary during seed germination, their function may be even more critical when this process takes place under abiotic stress. The photosynthetic function of chloroplasts is an important sensor in the integration of environmental signals into molecular signals (Taylor et al., 2009), and chloroplasts are primarily involved in the response to environmental stress, as recently demonstrated by analysis of the steady-state changes in chloroplast proteome under a number of environmental stress (Taylor et al., 2009). It should be pointed out that expression of the ELIP1 and ELIP2 genes is controlled by the same environmental signals – light, temperature and dehydration (Bartels et al., 1992; Otto et al., 1992; Adamska & Kloppstech, 1994; Montanéet al., 1997; Zeng et al., 2002) – that trigger seed germination (Bewley, 1997).

Our data suggest the possibility that the transcription factor DAG1 may influence chloroplast development during seed maturation, with loss of DAG1 therefore resulting in deregulation of chloroplast proteins such as ELIP1 and ELIP2. In turn, inactivation of ELIP1 and ELIP2 affects seed germination, particularly under abiotic stress conditions.


We thank Prof. Carlo Soave who kindly provided the elip1, elip2 and elip1 elip2 insertional mutants, Roberto Solano for helpful discussion and Samanta Salvucci for technical assistance. The seed germination assays under high/low light intensity were performed in the laboratories of Prof. G. Nehuaus and Dr Marta Rodriguez Franco (University of Freiburg, Germany). This work was partially supported by research grants from the Ministero dell’Istruzione, dell’Università e della Ricerca, Fondo per gli Investimenti della Ricerca di Base, Networking the European Research Area, Plant Genomics and by Istituto Pasteur-Fondazione Cenci Bolognetti and the Ministero dell’Istruzione, dell’Università e della Ricerca Progetti di Ricerca di Interesse Nazionale to P.C.