The BME3 (Blue Micropylar End 3) GATA zinc finger transcription factor is a positive regulator of Arabidopsis seed germination


(fax +1 541 737 3479; e-mail


In many plant species, seed dormancy is broken by cold stratification, a pre-chilling treatment of fully imbibed seeds. Although the ecological importance of seed response to cold temperature is well appreciated, the mechanisms underlying the physiological changes during cold stratification is unknown. Here we show that the GATA zinc finger protein expressed in Arabidopsis seeds during cold stratification plays a critical role in germination. Characterization of an enhancer-trap population identified multiple lines that exhibited β-glucuronidase (GUS) expression in the micropylar end of the seed (named Blue Micropylar End, BME lines). One of these lines, BME3, had a T-DNA insertion site in the 5′ upstream region of a GATA-type zinc finger transcription factor gene (termed BME3-ZF). The BME3-ZF mRNA accumulated in seeds during cold stratification. Characterization of the BME3-ZF promoter indicated that this gene was activated specifically in the embryonic axis, which was still enclosed by the endosperm. The zinc finger gene knockout plants produced seeds exhibiting deeper dormancy, which showed reduced response to cold stratification. The ungerminated knockout seeds exhibited testa rupture, but failed to penetrate the endosperm layer. Application of gibberellic acid (GA3) rescued impaired germination of knockout seeds without cold stratification, indicating that the normal GA signal transduction pathway is present in the knockout mutants. Expression of GA20-oxidase and GA3-oxidase genes was greatly reduced in the knockout seeds, suggesting the potential involvement of the zinc finger protein in GA biosynthesis. These results suggest that the GATA zinc finger protein is a positive regulator of seed germination.


The biological functions of GATA zinc finger proteins are well characterized in fungi and animals. Recent studies have provided emerging evidence that plant GATA zinc finger transcription factors also play significant roles in developmental control and responses to the environment. HANABA TARANU (HAN) encodes a GATA-3-like zinc finger protein containing the typical 18-aa-residue zinc finger loop (CX2CX18CX2C) found in plants, and is important in determining the organ boundaries in the flower, shoot apical meristem and developing embryos. HAN controls the number and position of cells expressing WUSCHEL, and affects cell proliferation and differentiation (Zhao et al., 2004). Zinc finger protein expressed in Inflorescence Meristem (ZIM), a novel Arabidopsis GATA zinc finger transcription factor, has a zinc finger loop with 20 aa residues (CX2CX20CX2C). ZIM is involved in hypocotyl and petiole elongation. The domain structure of the ZIM protein is found exclusively in plants (Shikata et al., 2004).

GATA zinc finger transcription factors are also associated with plant responses to environmental stimuli and defense mechanisms. Many light-responsive promoters found in plants contain GATA motifs (Teakle et al., 2002), suggesting that GATA zinc finger proteins are general regulators in light signal transduction. It is known that Arabidopsis GATA-1, GATA-2, GATA-3 and GATA-4 zinc finger genes are not developmentally regulated, but are responsive to light signals (Teakle et al., 2002). The Arabidopsis GATA zinc finger protein CONSTANS (CO) (Putterill et al., 1995) and its homologues in rice (Oryza sativa; Song et al., 1998), wheat (Triticum aestirum; Nemoto et al., 2003), and perennial ryegrass (Lolium perenne; Martin et al., 2004) are associated with photoperiodic control of flowering. Expression of the moss (Physcomitrella patens) CO/COL homologue PpCOL1 is controlled by the circadian clock, and is regulated photoperiodically at the gametophore stage when the rate of sporophyte formation is affected by day length. This suggests that the GATA zinc finger protein is also involved in the photoperiodic regulation of reproduction in moss (Shimizu et al., 2004).

In tobacco (Nicotiana tabacum), the AGP1 GATA zinc finger protein binds to the specific motif AGATCCAA in the promoter region of NtMyb2, which is a regulator of the phenylalanine ammonia lyase (PAL) gene. Both PAL and NtMyb2 are induced by various stresses such as wounding and elicitor treatments (Sugimoto et al., 2003). These results suggest that GATA zinc finger proteins also play critical roles in the defense response of plants. Thus zinc finger proteins are required for both developmental and environmental controls in plants.

Seed germination, which is a highly specialized phase of plant development, is also controlled by environmental signals such as cold temperature, light and soil nitrate concentration. Little is known about the involvement of GATA zinc finger proteins in seed germination. We found a GATA-type zinc finger transcription factor which was expressed in the micropylar end of germinating Arabidopsis seeds. Activation of genes in the micropylar end in both embryo and endosperm plays a significant role in seed germination of other plant species. Hydrolases and cell-wall proteins, which are involved in endosperm weakening or generation of embryo growth potential, are induced in a micropylar-specific manner in tomato (Lycopersicon esculentum; Chen and Bradford, 2000; Chen et al., 2002; Nonogaki et al., 2000; Wu et al., 2001), tobacco (Leubner-Metzger et al., 1996), and coffee (Coffea arabica; da Silva, 2002). We characterized a micropylar-specific GATA zinc finger protein in Arabidopsis seeds. The biological function of this protein in breaking seed dormancy and positively regulating seed germination in Arabidopsis is discussed.


Isolation of the BME enhancer-trap lines

Arabidopsis enhancer-trap lines [Thomas Jack lines, CS31086, Arabidopsis Biological Resource Center (ABRC), Ohio State University] were screened for seed germination-associated genes, and 121 lines showing tissue-specific β-glucuronidase (GUS) expression in seeds were isolated and donated to ABRC (Liu et al., 2005). In this screen we found multiple independent lines exhibiting strong GUS signals at the micropylar end of the seed (Figure 1a). These were designated Blue Micropylar End (BME) lines. GUS expression in these lines was detected when the radicle tip was still enclosed by the thin endosperm (stages I–III), indicating that the trapped genes are associated with sensu stricto germination, which is defined as physiological events before radicle emergence.

Figure 1.

GUS expression in BME (Blue Micropylar End) Arabidopsis enhancer-trap seeds.
(a) Representative images of GUS expression observed in BME lines. Three different views of seeds of a BME line (Arabidopsis Biological Resource Center, stock number CS24447) are shown. In most BME lines GUS activity was detected in both embryo and endosperm before (stages I–III, germination) and after radicle emergence (stage IV, post-germination). Note that the radicle tip has penetrated the thin endosperm layer at stage IV.
(b) GUS expression detected in BME3 seeds. Relatively weak expression was observed in this line.
(c) Left, GUS activity detected at top and bottom of developing siliques (inset: close-up view of receptacle); middle, GUS activity in mature silique; right, close-up views of top and bottom of mature silique.

We further characterized one of these BME3 lines which exhibited a relatively weak GUS signal in the embryo and endosperm (Figure 1b). GUS expression in BME3 started before radicle emergence. GUS expression was not detected in developing seeds, but the top and bottom of developing and mature siliques in BME3 exhibited GUS expression (Figure 1c). The progeny of the BME3 line also showed the same expression patterns, indicating that the putative T-DNA insertion was transmitted genetically.

Identification of the T-DNA insertion site in the BME3 line

To identify the BME3 gene in the enhancer-trap line, genome-walking PCR was conducted using T-DNA right-border (RB)-specific primers (RB1 and RB2) and adapter primers (see Experimental procedures). A DNA fragment of approximately 0.3 kb was amplified in the genome-walking PCR reaction (data not shown) and sequenced. Analysis of sequence data using the SIGnAL T-DNA Express Arabidopsis Gene Mapping Tool ( indicated that the T-DNA insertion was located in Arabidopsis chromosome III. The T-DNA was inserted in the 5′ upstream region of At3g54810 (GATA-type zinc finger family protein) (Figure 2a). The nearest gene, At3g54800 (lipid-binding START domain-containing protein), was located approximately 5 kb distant from the insertion site (data not shown). To verify the predicted T-DNA insertion site, PCR was conducted using the T-DNA right border-specific primer (RB2) and a gene-specific primer (GSP) (Figure 2a). The predicted size of DNA fragment (0.5 kb) was amplified from the genomic DNA of the BME3 line, but not from wild-type Arabidopsis genomic DNA (data not shown). The sequence of this PCR product contained the junction sequence between genomic DNA and the T-DNA right border, confirming the position of the predicted insertion site.

Figure 2.

Identification and verification of the T-DNA insertion site and trapped gene in the BME3 enhancer-trap line.
(a) Schematic representation of the predicted T-DNA insertion site in BME3 line. The T-DNA flanking region was amplified by genome-walking PCR using T-DNA right border-specific primers RB1 and RB2 and adapter-specific primers (see Experimental procedures). The putative insertion site was identified by analyzing the sequence of the amplified DNA fragment with SIGnAL T-DNA Express Arabidopsis Gene Mapping Tool ( The positions of the two primers, GSP and RB2, used for verification of the T-DNA insertion site and the T-DNA insertion sites of two putative SALK knockout lines [SALK_131396 (ZFKO1)and SALK_148073 (ZFKO2)], used for gene-function analysis (Figures 6 and 7), are also shown in this scheme.
(b) PCR products amplified using At3g54800 and At3g54810 gene-specific primers with wild-type Arabidopsis genomic DNA (gDNA) and with reverse transcription products of total RNA extracted from germinating seeds (12 h imbibition following 3 days’ cold stratification). The predicted size (0.8 kb) of RT-PCR product was obtained from At3g54810 but not from At3g54800, while PCR using gDNA with the same sets of primers amplified the predicted size (1.9 and 1.1 kb for At3g54800 and At3g54810, respectively) of DNA fragments. Sizes (kb) of amplified DNA fragments are shown at both sides of the panel.

To examine potential expression of the candidate gene in seeds, RT-PCR was performed using RNA extracted from wild-type Arabidopsis seeds with gene-specific primers for At3g54800 and At3g54810. An RT-PCR product of the predicted size (0.8 kb) was obtained for the At3g54810 gene, while no PCR product was obtained for the At3g54800 gene (Figure 2b, RT). This RT-PCR product was sequenced and found to match the coding region of At3g54810. DNA fragments of the predicted size (1.9 and 1.1 kb for At3g54800 and At3g54810, respectively) were amplified for both genes using wild-type Arabidopsis genomic DNA (Figure 2b, gDNA). These results indicate that the At3g54810 gene is expressed in germinating Arabidopsis seeds.

BME3 encodes a GATA zinc finger transcription factor

The BME3 (At3g54810) gene encodes a transcription factor with a GATA-type zinc finger (termed BME3-ZF). The zinc finger domain was found in the second exon of the BME3-ZF gene (Figure 3a). The amino-terminal region of BME3-ZF contains an acidic domain which is enriched with Asp (D) and Glu (E) (Figure 3a,b). The predicted isoelectric point for the amino-terminal 100 aa of BME3-ZF was calculated to be approximately 3.5. The characteristic Ser and Lys repeats were found in the amino acid sequence. The BME3-ZF protein contained the motif typical of the CX2CX18CX2C-type zinc finger (Figure 3b). Amino-acid sequence alignment of BME3-ZF and other related GATA zinc finger proteins in plants and fungi showed identical amino acids in the highly conserved regions of the C2C2 zinc finger domain. The BME3-ZF homologues in N. tabacum (NTAGBP-5) and Oryza sativa (OsGATA1) have extensive regions of amino acids identical to those found in BME3-ZF (Figure 3c).

Figure 3.

Gene structure and amino acid sequence of BME3 GATA zinc finger transcription factor.
(a) Schematic representation of gene structures of four different subfamilies of plant GATA zinc finger proteins (upper panel). Representative exon–intron organization in each subfamily shown according to Reyes et al. (2004). Black boxes, position(s) of zinc finger domains. The gene structure of the BME3 GATA zinc finger transcription factor gene (BME3-ZF) is shown (lower). Yellow and black, positions of the acidic domain (AD) encoded in the first exon and the zinc finger domain encoded in the second exon of BME3-ZF, respectively.
(b) Amino acid sequence of BME3-ZF. Asp (D) and Glu (E) in the amino-terminal acidic region; yellow and green, Ser- (S) and Lys- (K) repeats, respectively. Red, Cys (C); gray, other amino acids in the zinc finger domain.
(c) Amino acid sequence alignment of zinc finger domains with flanking regions from BME3-ZF (At3g54810) and four other Arabidopsis GATA zinc finger proteins representing class A (At5g66320), B (At4g36620), C (At3g21175) and D (At4g17570) zinc finger proteins. BME3-ZF homologues in tobacco NtAGBP-5 and rice OsGATA-1, and a fungal GATA zinc finger protein Gibberella fujikuroi GfAreA, are also compared. Black, identical amino acids in the zinc finger domain of all proteins. Blue, extended identical amino acids found in the flanking region of the zinc finger domain in class A proteins. Red dots, Cys (C); gray dots, other amino acids in the zinc finger domain.

The BME3 zinc finger transcription factor gene is expressed during cold stratification

To examine the temporal expression of BME3-ZF in Arabidopsis seeds, semi-quantitative RT-PCR was conducted. BME3-ZF mRNA accumulation was detected in dry seeds at a very low level, increased during imbibition at 4°C (cold stratification) for the first 18 h, and decreased thereafter (Figure 4a). Relatively low, but constant, levels of mRNA accumulation were detected in seeds transferred to 22°C after 24 h cold stratification (data not shown). These results suggested that the biological function of BME3-ZF might be associated with physiological changes during cold stratification. The timing of BME3-ZF expression was a little earlier than that of the GA biosynthesis-enzyme genes AtGA20ox3 and AtGA3ox1, which are also induced during cold stratification (Yamauchi et al., 2004).

Figure 4.

Expression of BME3-ZF gene and germination in dormant Arabidopsis seeds.
(a) Semi-quantitative RT-PCR of BME3-ZF. RNA was extracted from seeds during cold stratification (0, 6, 12, 18 and 24 h) and used for reverse transcription. The expression of GA20-oxidase (GA20ox3, At5g07200) and GA3-oxidase (GA3ox1, At1g15550) genes, expressed in Arabidopsis seeds during cold stratification (Yamauchi et al., 2004), was also examined. Actin2 (At3g18780) was used as a control.
(b) Germination of dormant Arabidopsis seeds with (+CS) or without (−CS) 3 days’ cold stratification. These seeds were used for experiments in (c). ‘+CS (cold stratification)’ indicates exposure to 4°C for 3 days and an additional imbibition period.
(c) Semi-quantitative RT-PCR of BME3-ZF. RNA was extracted from dormant Arabidopsis seeds imbibed for 24, 36 and 48 h without cold stratification (−CS) and used for RT-PCR. Actin2 (ACT2) was used as a control.

It is possible that BME3-ZF is a cold-inducible gene and is not associated with germination per se. To examine whether gene expression is correlated with seed germination, we characterized BME3-ZF expression in dormant Arabidopsis seeds kept in our laboratory. The dormant seeds had been harvested, dried and immediately stored at −80°C to maintain deep dormancy. The seeds were still capable of germinating without cold stratification, but at slower rates (Figure 4b). To determine if cold treatment is essential for the induction of BME3-ZF, we analyzed gene expression in dormant seeds incubated at 22°C without cold stratification. As shown in Figure 4(c), BME3-ZF expression was induced before radicle emergence. These results indicate that induction of BME3-ZF is not absolutely dependent on cold induction, and suggest that the gene is probably associated with, and a prerequisite for, the induction of seed germination.

BME3-ZF transcription factor gene is activated in the embryonic axis

Although characterization of the BME3 enhancer-trap line allowed us to identify expression of BME3-ZF in Arabidopsis seeds, Southern blot analysis using a GUS-specific probe indicated that this enhancer-trap line might have multiple T-DNA insertions (data not shown). There was a possibility that GUS expression detected in BME3 enhancer-trap seeds reflected the promoter activity of other genes. To examine the spatial pattern of BME3-ZF expression, we amplified the 5′ upstream region (−1075 to +19) of BME3-ZF by PCR, fused the promoter sequence to the GUS gene, and transformed Arabidopsis with this promoter::reporter construct. Seeds were produced from the transgenic plants and tested for GUS expression.

Germinating seeds showed very weak GUS activity after testa rupture, reproducing the characteristic expression of BME3 seeds (Figure 5a). The GUS signal was detected in the embryo, which was enclosed by the thin, transparent endosperm layer (Figure 5b). The presence of GUS activity in the embryonic axis was confirmed by inspecting the excised embryo; little or no activity was observed in the endosperm (Figure 5c). These results indicate that the BME3-ZF gene is activated mainly in the embryo. The low-intensity GUS signal detected in the endosperm of BME3 enhancer-trap seeds (Figure 1b) might have diffused from embryonic tissues, or could reflect the expression of another trapped gene in the BME3 line. GUS activity was also detected at the top and bottom (receptacle) of siliques (data not shown) of the transformants, which was consistent with the GUS-expression patterns observed in the original enhancer-trap line. The GUS gene of the inserted T-DNA in the enhancer-trap line was in opposite orientation to the BME3-ZF gene. Probably, the trapped enhancer can function in both directions.

Figure 5.

GUS expression driven by the 5′ upstream region of BME3-ZF gene. The promoter region (−1075 to +19) was fused with the GUS reporter gene and used to transform wild-type Arabidopsis.
(a) GUS activity in germinating seed with testa rupture. A very weak signal is seen inside seed.
(b) The same seed with testa removed. The embryo is still enclosed by the thin, transparent endosperm. Note that a GUS signal is detected in the axis inside the de-coated seed.
(c) Embryo (left) and endosperm (right) excised from the same seed.

Disruption of BME3 zinc finger transcription factor gene causes impaired seed germination

Two putative knockout lines of the At3g54810 GATA zinc finger gene, Salk_131396 (termed ZFKO1) and Salk_148073 (ZFKO2), which had T-DNA insertion sites in the second exon and promoter region, respectively, were obtained to investigate gene function (see Figure 2a for insertion sites). Homozygous plants were identified by PCR screening. No apparent phenotype was observed in seedlings, in developing siliques, or at later stages of plant growth in the homozygous lines. Seeds were harvested from homozygous plants of ZFKO1 and ZFKO2. Expression of BME3-ZF in these putative knockout lines was analyzed by semi-quantitative RT-PCR using RNA extracted from seeds treated at 4°C for 18 h. The expression levels of BME3-ZF were drastically reduced in both ZFKO1 and ZFKO2 compared with expression in wild-type Arabidopsis seeds (Figure 6a).

Figure 6.

Gene expression and germination phenotype of knockout plants.
(a) Semi-quantitative RT-PCR of BME3-ZF gene for wild-type Arabidopsis (WT) and the two putative SALK knockout lines [SALK_131396 (ZFKO1) and SALK_148073 (ZFKO2)]. RNA was extracted from seeds treated at 4°C for 18 h. Actin2 (ACT2) was used as a control.
(b) Germination of wild-type Arabidopsis (WT) and the ZFKO1 and ZFKO2 knockout seeds without cold stratification (−CS).
(c) Germination of wild-type Arabidopsis (WT) and the ZFKO1 and ZFKO2 knockout seeds with cold stratification (+CS). ‘+CS’ indicates exposure to 4°C for 3 days and an additional imbibition period.
(d) Germination of wild-type Arabidopsis (WT) and the ZFKO1 and ZFKO2 knockout seeds after 120 h in water, 0.2% (w/v) potassium nitrate (KNO3) and 10 μm gibberellic acid (GA3) without cold stratification (−CS).
Each data point in panels (a–c) represents the average of three replicates; vertical bars, SD.

Germination of wild-type and knockout Arabidopsis seeds was initially compared without cold stratification. The final germination percentage was less in ZFKO1 and ZFKO2 seeds compared with that of wild-type Arabidopsis seeds (Figure 6b), suggesting that disruption of the zinc finger gene caused a deeper dormancy. Cold stratification induced full germination of wild-type Arabidopsis seeds, but the final germination percentage of ZFKO1 and ZFKO2 seeds reached <70% even after 5 days’ incubation at 22°C following 3 days’ cold stratification (Figure 6c). Further incubation did not increase the germination percentage. The ungerminated ZFKO1 and ZFKO2 seeds were still viable, as they showed positive staining by a tetrazolium test (data not shown). These results demonstrated that ZFKO1 and ZFKO2 seeds had a reduced response to cold stratification.

As >30% of ZFKO1 and ZFKO2 seeds still failed to germinate, even with cold stratification, we examined the effect of potassium nitrate (0.2% w/v), which is typically used instead of cold stratification to break seed dormancy in seed-testing laboratories. Although germination was slightly enhanced by potassium nitrate, seeds of the two knockout lines still showed reduced germination levels compared with wild-type seeds after this treatment (Figure 6d). Gibberellic acid (GA3, 10 μm) stimulated germination of the knockout seeds, with over 90% of seeds germinating (Figure 6d). It is interesting that the deep dormancy of ZFKO1 and ZFKO2 seeds, which was not even overcome by 3 days’ cold stratification, was broken by exogenous GA without cold stratification. This result suggests that the normal GA signal transduction pathway is present in these two mutant lines.

The impaired germination phenotype was observed consistently in both ZFKO1 and ZFKO2 knockout seeds. The degree of seed dormancy, however, can also be affected by the physiological state of the maternal plants. That is, the environment that the maternal plants experience during seed development could affect germinability of the progeny seeds. It was possible that the difference in wild-type and knockout seed-germination performance was due to small differences experienced by their maternal plants during seed production, although the wild-type, ZFKO1 and ZFKO2 seeds were produced side-by-side under the same conditions. To confirm the low germination phenotype of ZFKO1 and ZFKO2 lines, the next generation of seeds from these two mutant lines and wild-type Arabidopsis were produced again under the same conditions and tested for the germination phenotype. The impaired germination phenotype was reproducible in the second generation of ZFKO1 and ZFKO2 lines (termed ZFKO1-2nd and ZFKO2-2nd, respectively). More than 35% of seeds in both lines failed to germinate after 5 days incubation at 22°C following 3 days cold stratification (data not shown). Most of the seeds that failed to germinate exhibited testa rupture (Figure 7a), although there was a delay in the timing of testa rupture in the mutant seeds. This result indicates that the mutant embryos had at least a small increase in the growth potential to cause testa rupture, but could not penetrate the endosperm layer. When the remaining ungerminated seeds were transferred to GA3 solution (10 μm) they germinated within 1 day (data not shown), which was consistent with experiments on the previous generation of seeds. These results led to the hypothesis that impaired germination of the knockout seeds may be due to reduced GA synthesis rather than decreased sensitivity to GA. To test this hypothesis, we examined the expression of GA20ox3 and GA3ox1 genes, which are the major GA-biosynthesis enzymes induced by cold stratification (Yamauchi et al., 2004). Strikingly, we found that both genes are downregulated in the progeny of both ZFKO1 and ZFKO2 lines (Figure 7b), indicating that BME3-ZF is probably associated with the upstream event(s) of GA biosynthesis.

Figure 7.

Germination phenotype and expression of BME3-ZF and GA biosynthesis enzyme genes in the second generation of knockout seeds.
(a) Photograph of knockout seeds that failed to germinate after 5 days’ incubation at 22°C following 3 days’ cold stratification. Note that testa rupture and slight elongation of the endosperm occur in the ungerminated seeds.
(b) Semi-quantitative RT-PCR of BME3-ZF, GA20ox3 and GA3ox1 genes for wild-type (WT) and knockout (ZFKO1-2nd and ZFKO2-2nd) seeds. RNA was extracted from seeds treated at 4°C for 24 h. The same sets of reverse transcription products were used for four different genes, including Actin2 (ACT2) control.
(c) Photograph of wild-type (WT) and knockout (ZFKO1-2nd and ZFKO2-2nd) seeds at the early post-germinative stage (3 days) on an agarose plate without cold stratification.

We also found that, when the ungerminated knockout seeds were plated on agarose medium with Murashige–Skoog (MS) salt (Murashige and Skoog, 1962) and 1% (w/v) sucrose, they were able to germinate. The promotion of knockout seed germination on agarose medium was also observed in the absence of MS salts and sucrose, suggesting that higher retention of water around seeds, rather than nutrition of the medium, enhanced radicle emergence. However, comparison of wild-type and knockout seed germination on agarose medium also showed large differences in germination speed (Figure 7c).


Potential of the ‘Seed-GUS-Expression’ enhancer-trap library for seed germination research

We previously reported the generation of a library of ‘Seed-GUS-Expression’ enhancer-trap lines (Liu et al., 2005;, one of which (BME3) was used in the present study. The seeds, which have been donated to the Arabidopsis Biological Resource Center at Ohio State University, USA, are now available (CS24362–CS24480) for the international research community to identify seed germination-associated genes. We previously reported the identification of insertion sites in the vicinity of genes encoding signal transduction proteins and several types of transcription factors in these enhancer-trap lines (Liu et al., 2005); however, examples of the successful application of the enhancer-trap lines for functional analysis of seed germination-associated genes were yet to be demonstrated. The present work provides the proof-of-concept study for the ‘Seed-GUS-Expression’ enhancer-trap lines. Although the BME3 line had several T-DNA insertions (data not shown), characterization of the 5′ upstream sequence of the GATA zinc finger gene verified the tissue-specific GUS expression detected in the BME3 seeds. This study also shows that a combination of gene-expression analysis using wild-type seeds and functional analysis using knockout plants allows us to identify genes crucial for seed germination.

BME3-ZF is a subfamily I GATA-type zinc finger protein

In eukaryotes, zinc finger proteins are an abundant group of transcription factors which have a wide range of biological functions (Laity et al., 2001). The zinc finger genes are by far the largest family of transcription factors (762 members), followed by homeobox genes (199 members) and basic helix–loop–helix genes (117 members) (Messina et al., 2004). Analysis using the SIGnAL ‘T-DNA Express’ Arabidopsis Gene Mapping Tool ( indicates that >800 putative zinc finger genes are present over five chromosomes of Arabidopsis. Zinc finger proteins can be classified into several types, such as C2H2; C2HC; C2C2; C2HCC2C2; and C2C2C2C2 types, based on the number and position of Cys and His residues in the zinc finger domains (Huang et al., 2004). BME3-ZF is classified as a C2C2 GATA zinc finger protein which recognizes the consensus sequence motif (A/T)GATA(A/G). The GATA-type zinc finger proteins were originally identified in vertebrates (Evans and Felsenfeld, 1989; Evans et al., 1988) and are also present in fungi, animals and plants (Reyes et al., 2004). The DNA-binding domains were typically defined as CX2CX17−20CX2C, where X is any amino acid. The CX2CX17CX2C and CX2CX18CX2C configurations are found in fungal GATA-type zinc finger proteins (Teakle and Gilmartin, 1998). Most animal zinc finger proteins contain CX2CX17CX2C and have two zinc finger domains (Patient and McGhee, 2002; Reyes et al., 2004). Zinc finger domains with 18–20 residue loops (CX2CX18–20CX2C) can be found in plants (Nishii et al., 2000; Reyes et al., 2004). In Arabidopsis, 29 GATA zinc finger protein genes have been identified (Reyes et al., 2004). The coding region of the BME3-ZF gene contains two exons, with the last exon encoding the complete zinc finger motif (Figure 3a). The exon–intron organization seen in the BME3-ZF gene structure is also found in 13 Arabidopsis genes which are categorized as subfamily I GATA zinc finger proteins. Another characteristic of subfamily I GATA zinc finger proteins is the presence of an acidic amino-terminal domain with pI < 4 (Reyes et al., 2004). Although the function of an acidic domain is unknown, the presence of this characteristic in BME3-ZF also indicates that BME3-ZF falls into subfamily I (Figure 3a,b). Reyes et al. (2004) conducted comprehensive alignments of CX2CX18CX2C (and CX2CX17CX2C) GATA zinc finger domains and their flanking amino acid sequences of Arabidopsis and rice. They grouped the plant CX2CX18CX2C zinc finger domains into four classes (A–D) based on the conservation of sequences in the flanking region. According to their classification, BME3-ZF is a class A GATA zinc finger which also includes NtAGBP-5 in tobacco and OsGATA-1 in rice (Figure 3c).

BME3-ZF is a positive regulator of seed germination

Seeds exhibit no apparent morphological changes during imbibition, except for seed-size change following water uptake, and testa rupture which occurs only during the later stages of seed imbibition. Apparent embryo growth is a post-germinative event and occurs only after rupture of the endosperm layer by the radicle. However, generation of embryo growth potential, which is a prerequisite for the completion of germination, takes place while the embryo is enclosed by the endosperm during imbibition. The molecular mechanisms controlling this early event in seed germination are not comprehensively understood.

Gibberellic acid releases seed dormancy in many different species, including Arabidopsis. The mechanisms of the GA-signaling pathway in seed-dormancy release are emerging. The DELLA transcription factors play significant roles in GA signal transduction during seed germination (Lee et al., 2002; Peng and Harberd, 2002; Tyler et al., 2004; Wen and Chang, 2002). The DELLA proteins are repressors of seed germination and are probably degraded by the ubiquitin-26S-proteosome pathway (Dill et al., 2004; Itoh et al., 2003; McGinnis et al., 2003).

Only a few factors involved in seed dormancy and sensu stricto germination have been characterized, other than DELLA transcription factors. Dof zinc finger proteins DAG1 and DAG2 are examples of well-characterized transcription factors associated with cold stratification and light response for seed germination. These two Dof zinc finger proteins share high sequence homology and an identical zinc finger domain, but they exhibit completely opposite functions in terms of response to cold stratification, light and GA. DAG1 and DAG2 play negative and positive roles, respectively, in Arabidopsis seed germination. Both genes are expressed in the maternal tissues, such as funiculus, during seed development (Gualberti et al., 2002; Papi et al., 2000). A novel-type zinc finger protein MEDIATOR OF ABA-REGULATED DORMANCY (MARD1), which is a negative regulator of seed germination, is associated with maintenance of seed dormancy imposed by abscisic acid (He and Gan, 2004).

The GATA zinc finger protein was identified as another member of sensu stricto germination-associated genes in the present study. Disruption of the BME3-ZF reduced the response of mutant seeds to cold stratification and caused a deeper dormancy. This suggests that BME3-ZF positively mediates the developmental shift from dormancy to germination. The mutant seeds showed very slow germination, and >30% of seeds failed to germinate even when cold-stratified (Figure 6c). Ungerminated knockout seeds exhibited testa rupture and slight elongation of the endosperm out of the testa (Figure 7a). The rupture of testa is not simply due to physical changes following water uptake, because seeds imbibed during cold stratification or completely dormant seeds imbibed at 22°C do not show testa rupture. Generation of growth potential of the embryo is probably required for testa rupture. Therefore the embryos of the mutant ungerminated seeds generated a minimal growth potential, enough for testa rupture but insufficient for the embryo to penetrate the endosperm layer. It is not known (and technically difficult to examine) whether a single cell layer of the Arabidopsis endosperm provides significant mechanical resistance to prevent radicle protrusion. The thin endosperm seems to have some resistance that could not be overcome by the radicle of the mutant embryo. The BME3-ZF promoter was activated in the axis in terms of seed expression (Figure 5). The function of BME3-ZF is probably associated with a change in the embryonic axis, providing further growth potential increase which is necessary for the radicle to overcome the residual mechanical resistance of the endosperm after testa rupture.

The impaired germination phenotype of the knockout seeds was rescued by exogenous GA (Figure 6d), which is associated with cell enlargement and elongation of the embryo in Arabidopsis seeds (Yamaguchi et al., 2001). The conversion of an inactive form of GA9 to bioactive GA4, which is catalyzed by GA3-oxidase, is a rate-limiting step for the synthesis of bioactive GA in Arabidopsis seeds (Yamaguchi and Kamiya, 2001; Yamaguchi et al., 1998, 2001). The GA3-oxidase genes GA4 (GA3ox1) and GA4H (GA3ox2), which are regulated by a plant photoreceptor phytochrome (Yamaguchi et al., 1998), play a crucial role in stimulating Arabidopsis seed germination. Yamauchi et al. (2004) discovered that cold stratification also induces GA-biosynthesis genes GA20ox3 and GA3ox1, suggesting that the promotional effect of cold stratification on seed germination is mediated by GA synthesized in seeds.

The knockout seeds showed sensitivity to exogenous GA3 (Figure 6d). Impaired germination of ZFKO mutant seeds is probably due to a reduced level of GA biosynthesis. Expression of the GA-biosynthesis genes GA20ox3 and GA3ox1 was downregulated in the knockout seeds (Figure 7b). BME3-ZF probably acts upstream of GA biosynthesis. Expression of BME3-ZF starts just prior to expression of the GA-biosynthesis genes during cold stratification (Figure 4a). GA3ox1 is also expressed in the axis of Arabidopsis seeds (Yamaguchi et al., 2001). These findings also support the potential function of BME3-ZF in GA biosynthesis during and following cold stratification. The knockout plants did not show phenotypes typical of GA-deficient mutants such as dwarfism. The regulation of GA biosynthesis by BME3-ZF could be seed-specific.

More than 60% of the knockout seeds were still capable of germinating despite the drastic reduction in BME3-ZF expression (Figure 6). This can be explained by a partial reduction in GA biosynthesis. Alternatively, it is also possible that there is redundancy in BME3-ZF gene function. The possibility that BME3-ZF also modulates germination by regulating other plant hormones cannot be excluded. Studies using the GA signal transduction mutant sly and GA-biosynthesis mutant ga3 (Steber and McCourt, 2001) and the ethylene-insensitive mutant etr-1–2 (Chiwocha et al., 2005) clearly showed cross-talk between GA and other plant hormones such as brassinosteroids, ethylene and cytokinins. Identification of potential interacting partner(s) of BME3-ZF and other germination-associated transcription factors will provide a clear picture of the biology of seed germination.

Experimental procedures

Screening of the Arabidopsis enhancer-trap lines

The Arabidopsis enhancer-trap lines (Thomas Jack lines, 1130 pools of 10 lines, CS31086, ABRC) (Campisi et al., 1999) were used for screening. Approximately 50 seeds from each sub-pool were placed on a small, round filter paper (approximately 7 mm diameter), moistened with water, and placed on a metal stage for SEM specimens. The metal stage was cleaned thoroughly each time before placing another sub-pool of seeds, to avoid contamination. One hundred sub-pools of 50 seeds placed on small filter paper were incubated on two layers of larger (15-cm diameter) filter paper placed in a plastic Petri dish, and incubated as described above. After 3 days’ pre-chilling at 4°C and 22 h incubation at 22°C, the small filter papers holding seeds were briefly blotted on filter papers to remove excessive water, then examined for GUS expression.

GUS staining

GUS staining of seeds and other tissues was performed as described previously (Weigel and Glazebrook, 2002) using 100 mm sodium phosphate buffer pH 7.0 containing 0.1% (v/v) Triton X-100 and 2 mm X-Gluc (RPI, Mount Prospect, IL, USA). Staining was examined after overnight incubation at room temperature (approximately 22°C).

Identification and verification of T-DNA insertion site and trapped gene

Genomic DNA was extracted from Arabidopsis leaves using phenol extraction according to the QUICK-PREP method described at and used for genome-walking PCR experiments to identify the T-DNA insertion site. Genome-walking PCR was performed using Genome Walker Kit (Clontech, Palo Alto, CA, USA) according to the manufacturer's manual. Briefly, genomic DNA of the BME3 enhancer-trap line was digested with DraI. Adapter DNA provided with the kit was ligated to the DraI-digested genomic DNA fragments. The first PCR was conducted using Ex Taq DNA polymerase (Takara, Madison, WI, USA), an adapter primer (5′-GTAATACGACTCACTATAGGGC-3′) provided with the kit, and a T-DNA right border-specific primer (RB1: 5′-TCTAGAGTCGACCTGCAGGCATGCAAGCTT-3′). Second-round PCR was done using the nested adapter primer (5′-ACTATAGGGCACGCGTGGT-3′) and the nested T-DNA right border-specific primer (RB2: 5′-TCCCAACAGTTGCGCACCTGAATGGCGAAT-3′). The conditions used for the first PCR were: one cycle at 94°C (4 min); one cycle at 80°C (2 min); seven cycles at 94°C (25 sec), 72°C (3 min); 32 cycles at 94°C (25 sec), 67°C (3 min); followed by one cycle at 67°C (7 min). The conditions used for the second PCR were: one cycle at 94°C (4 min); one cycle at 80°C (2 min); five cycles at 94°C (25 sec), 72°C (3 min); 20 cycles at 94°C (25 sec), 67°C (3 min); followed by one cycle at 67°C (7 min). The amplified DNA fragment was sequenced at the Center for Gene Research and Biotechnology, Oregon State University.

To verify the predicted T-DNA insertion site in the BME3 line, a gene-specific primer (GSP: 5′-ACGACGTTTACGCGTACATG-3′) was designed for the 5′ upstream sequence of the At3g54810 gene (located downstream of the putative T-DNA insertion site) and used in PCR with RB2 primer and genomic DNA from the BME3 line. The following conditions were used for PCR: initial denaturation at 94°C (4 min); touchdown cycles [94°C (15 sec), 69 → 63°C (15 sec), 72°C (30 sec)] (one cycle for each temperature) and 25 cycles at 94°C (15 sec), 62°C (15 sec) and 72°C (30 sec) followed by extension at 72°C (7 min).

To examine the expression of the candidate trapped genes identified by genome-walking PCR in imbibed seeds, specific primers were designed for At3g54800 (5′-TTATTGCTGCGACGCTACTG-3′ and 5′-TTGAGCCGACAAGTTCCTAG-3′) and At3g54810 (5′-CTACCATCTGGACCACTCAT-3′ and 5′-TACCATCACTGCATCTCTTG-3′), and RT-PCR was conducted using RNA extracted from 12-h imbibed seeds. The following conditions were used for PCR: initial denaturation at 94°C (4 min); touchdown cycles [94°C (15 sec), 67 → 61°C (15 sec), and 72°C (30 sec)] (one cycle for each temperature) and 25 cycles at 94°C (15 sec), 60°C (15 sec) and 72°C (30 sec), followed by extension at 72°C (7 min).

Semi-quantitative RT-PCR

Total RNA was extracted from Arabidopsis seeds imbibed for 0, 6, 12, 18 and 24 h at 4°C, and for an additional 6, 12 and 18 h at 22°C, using standard phenol–sodium dodecyl sulphate extraction (Sambrook et al., 1989). Two μg DNase-treated total RNA were used for reverse transcription (RT) with a RETROscript Kit (Ambion, Austin, TX, USA). The RT product was subjected to semi-quantitative PCR using the primers for the At3g54810 gene (described above). The conditions for semi-quantitative PCR were essentially as described above, except that 20 cycles were used at 94°C (15 sec), 60°C (15 sec) and 72°C (30 sec). An actin gene ACT2 (An et al., 1996) was used as a control in the semi-quantitative PCR with specific primers (5′-GCCATCCAAGCTGTTCTCTC-3′ and 5′-GAACCACCGATCCAGACACT-3′).

Construction of BME3-ZF promoter::GUS reporter cassette

The 5′ upstream region of BME3-ZF (−1075 to +19) was amplified using forward (5′-CTGAGCTCCGTCCTAGGGGAAAAACTCA-3′) and reverse (5′-CAGTCGACGGAGTGGGGAGAAGTGAAGA-3′) primers, which contained SacI and SalI restriction enzyme sites at their 5′ and 3′ ends, respectively. The PCR products were digested with SacI and SalI and cloned into the SacI and SalI sites of the shuttle vector pRJG23 (Grebenok et al., 1997) that contained the uidA (GUS) gene. The promoter-GUS construct from pRJG23 was removed with SacI and SpeI and subcloned into SacI and XbaI sites in pGPTV-KAN binary vector (Becker et al., 1992) to produce the BME3-ZF promoter::GUS binary vector.

Arabidopsis transformation and screening

For transformation, 200 ml YEP medium [1% (w/v) yeast extract, 1% (w/v) peptone, 0.5% (w/v) NaCl] (Weigel and Glazebrook, 2002) containing 50 μg ml−1 kanamycin was inoculated with 5 ml overnight culture of Agrobacterium tumefaciens EHA 105 strain (Hood et al., 1993) harboring the BME3-ZF promoter::GUS binary vector and grown for an additional 16 h at 28°C with vigorous shaking. Cells were harvested by centrifugation at 6000 g with a Sorvall RC-5B centrifuge (DuPont Instruments, Wilmington, DE, USA) at ambient temperature, resuspended in 400 ml 5% (w/v) sucrose solution containing 0.02% (v/v) Silwet L-77 detergent (Lehle Seeds, Round Rock, TX, USA), and used for transformation with the floral dip method as described previously (Clough and Bent, 1998). Seeds were harvested and stored at room temperature (approximately 22°C). For screening, seeds were sterilized in 70% (v/v) ethanol for 1 min, and in 50% bleach solution containing 0.1% (v/v) Tween 20 for 10 min, followed by several washes with sterile water. Kanamycin-resistant plants were selected by growing plants for 14 days on 0.7% (w/v) agarose plates containing 4.3 mg ml−1 MS salts, 1% (w/v) sucrose and 25 μg ml−1 kanamycin.

Germination test

Fifty seeds were germinated in a Petri dish containing two layers of filter paper moistened with 4 ml water or test solution (three replicates per treatment). Germination and seedling establishment were also tested on 0.7% (w/v) agarose plates with or without 4.3 mg ml−1 MS salts and 1% (w/v) sucrose. Cold stratification was conducted at 4°C in the dark. Germination was performed at 22°C with 16-h light and 8-h dark periods.


We are grateful to the Arabidopsis Biological Resource Center, Ohio State University, USA for propagating the ‘Seed-GUS-Expression’ enhancer-trap lines for the international seed research community and providing the knockout lines for our research. This work was supported by National Science Foundation grant IBN-0237562 to H.N.