INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The Dof proteins are a wide family of transcription factors, recently discovered and present only in plants. These proteins are characterized by a strongly conserved 52-amino acid domain encompassing a single CX₂CX₂₁CX₂C zinc finger (Kisu, et al., 1995; Yanagisawa, 1995; Zhang et al., 1995; De Paolis, et al., 1996; Vicente-Carbajosa et al., 1997; Mena et al., 1998). The strong conservation of the zinc finger domain is reflected in a very similar DNA binding site for all the Dof proteins, which includes a core CTTT sequence (Yanagisawa and Schmidt, 1999). In contrast, outside the conserved Dof domain, these proteins diverge widely and are involved in different regulatory circuits all typical of and of general relevance for plants. A number of Dof proteins are being characterized in several plants. By means of transient expression assays in protoplasts, it has been shown that the maize (Zea mays) Dof1 and Dof2 proteins control the expression of genes involved in carbon metabolism (Yanagisawa and Sheen, 1998; Yanagisawa, 2000). Another maize protein, prolamine-box binding factor, binds to the prolamine box in zein gene promoters and interacts with the transcriptional activator Opaque2 (Vicente-Carbajosa et al., 1997). Its barley (Hordeum vulgare) counterpart barley prolamine-box binding factor has been shown by means of transient expression experiments in developing barley endosperms to be capable of trans-activating the B-hordein promoter (Mena et al., 1998). In pumpkin (Cucurbita pepo), the Dof protein AOBP binds to the promoter of the abscorbate oxidase gene (Kisu et al., 1998; Shimofurutani et al., 1998), whereas we have shown that the tobacco (Nicotiana tabacum) Dof protein NtBBF1 controls the tissue-specific and auxin-inducible expression of the oncogene rolB in transformed plants (Baumann et al., 1999). In Arabidopsis, analysis of the now almost complete genomic sequence indicates the presence of some 40 members of the Dof gene family. Three of them, OBP1 (Chen et al., 1996), OBP2, and OBP3, have similar properties in vitro. They are all capable of interaction with OBF4, a transcriptional regulator of a stress-induced gene encoding glutathione S-transferase, but show distinctive expression patterns, suggesting distinct functions in different plant organs (Kang and Singh, 2000).

The only Dof gene for which an effect in plants has been so far convincingly demonstrated is DAG1, which we have recently shown to be involved in seed germination in Arabidopsis. The knockout mutant in DAG1, isolated from a T-DNA insertion collection, produces seeds that do not develop dormancy and are capable of germinating in the dark (Papi et al., 2000). Seeds of several annuals, including Arabidopsis, develop dormancy during the late stages of their development: Although mature, they are not capable of germinating under favorable conditions when freshly harvested or naturally detached from the mother plant. Seed dormancy can be relieved by a period of dry storage referred to as "after ripening"; in Arabidopsis, stored nondormant seeds need illumination with (red) light to germinate (Koornneef and Karssen, 1994; for review, see Bewley, 1997). Disruption of the Dof gene DAG1 causes mutant seed to loose dormancy and dependence upon red light for germination. In addition, we showed that the gene DAG1 is expressed only in the mother plant and not in the seed at any stage of development. In accordance, the segregation pattern of the dag1 mutant seed phenotype in the progeny indicates that the effect of the mutation is maternal (Papi et al., 2000).

In this work, we compared dag1 mutant seeds with the corresponding Wassilewskija (Ws) wild type for sensitivity to light and (structural) characteristics of the testa and we show that both are altered. We also analyzed in detail the expression pattern of DAG1 in plants, and we show that the gene is specifically expressed in the phloem of all organs of the plant but not in the seed or in the embryo at any stage of development.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Germination of dag1 Mutant Seeds Has a Higher Sensitivity to Red Light

We have previously shown that, although the seeds from dag1 mutant plants germinate in substantially higher proportions than wild-type Ws seeds in the absence of a pulse of red light, irradiation with far-red light inhibits germination of both types of seeds (Papi et al., 2000).

To assess potential differences in the sensitivity of dag1 and Ws seeds to light, we scored germination in total darkness after exposure of the seeds to different conditions of irradiation, as reported in Figure 1. Before any other irradiation, seeds were exposed for 30 min to long-wavelength far-red light to minimize the levels of active Pfr phytochome (Heim and Schäfer, 1982).

View larger version (17K): [in this window] [in a new window]   Figure 1.   Germination of dag1 mutant seeds is more sensitive to red light. Ws (white squares) and dag1 (black circles) seeds were irradiated with far red light for 30 min and germination in the dark was scored after a pulse (5 min) of red light (660 nm) of different fluence rates (A); a pulse (5 min) of monochromatic light of different wavelengths (4 µmol m2 s1; B); a pulse (5 min) of red light (660 nm, 20 µmol m2 s1) and a subsequent pulse of saturating long-wavelength far-red light after different intervals of time (the dotted and dashed line indicate the germination of, respectively, Ws and dag1 seeds in the absence of the long-wavelength far-red pulse; C). Means of at least 50 seeds sown in triplicate are shown. Error bars represent SEs.

In Figure 1A, we report the germination scores obtained upon irradiation of the seeds for 5 min with different red light (660 nm) fluence rates, which generate different amounts of active (Pfr) phytochrome (Mancinelli, 1994). As can be seen from the respective response curves, germination of the mutant seeds is induced with fluence rates substantially lower than those needed for Ws seeds. In particular, 50% germination was obtained at a fluence rate of approximately 2 µmol m² s¹ in the case of Ws seeds, whereas a fluence rate six times lower (0.33 µmol m² s¹) is sufficient for dag1 seeds. In Figure 1B, the effects on germination of a 5-min pulse with light of different wavelengths are shown. The capability of wild-type Arabidopsis seeds to germinate, which is strongly induced by red light at 660 nm, drops rapidly with the increase of the wavelength, zeroing at and above 690 nm. In contrast, the curve of dag1 seeds shows that germination is still induced to some extent even by wavelengths around and above 750 nm.

Finally, seeds were irradiated for 5 min with red light and subsequently exposed to a long-wavelength far-red light pulse (to inactivate Pfr) after different intervals of time in the dark. In Figure 1C, curves for the germination of dag1 and Ws seeds under these conditions are compared. At least 12 h are needed before the far-red pulse becomes noticeably ineffective (i.e. before germination of Ws seeds reach some degree of independence of Pfr). In contrast, already 4 h after being induced by red light, the far-red pulse can no longer reverse the germination of a significant percentage of dag1 seeds.

The Testa of dag1 Seeds Appears Normal But Shows Histochemical Alterations

In Arabidopsis, alterations in the physicochemical and/or structural characteristics of the testa generally result in altered seed germination properties (Koornneef and Karssen, 1994; Bewley, 1997). Because the testa is of maternal origin as is the effect of the dag1 mutation on seed germination (Papi et al., 2000), we assessed whether dag1 seeds had altered seed coats by a number of different histochemical procedures, as reported in Figure 2.

View larger version (83K): [in this window] [in a new window]   Figure 2.   The seed coat of dag1 seeds shows histochemical alterations. A and B, Toluidine blue-stained sections of a Ws and, respectively, dag1 seed coat (ml, mucilage layer; c, columella; ep, epidermis; p, palisade layer; cp, crushed parenchymatic layers; e, endothelium layer; a, aleurone layer; bar = 40 µm); C and D, epidermis of Ws and, respectively, dag1 seeds as viewed with Nomarski differential interference optics (c, columella; bar = 40 µm); E and F, Ws and, respectively, dag1 seed stained for 15 min with ruthenium red after a 3-min imbibition (bars = 200 µm); G and H, Ws and, respectively, dag1 seeds stained with ruthenium red for 15 s without prior imbibition (bar = 1,000 µm); I and J, Ws and, respectively, dag1 seeds stained with tetrazolium salts (bar = 1000 µm); K and L, immature Ws and, respectively, dag1 seeds stained with vanillin (bar = 500 µm).

Comparative microscopic observation of Ws and dag1 mutant seeds revealed no substantial difference in the structure of their seed coats, as shown in Figure 2, A and B. In both types of seeds, the four layers of the mature testa are clearly visible. These are the epidermis and palisade, which form the outer integument, and the two layers (a crushed layer generated by the collapse of the two parenchymatic layers of the immature testa and the endothelium) of the inner integument (Debeaujon et al., 2000). Epidermal cells of Arabidopsis seeds produce mucilage, which is released to surround the seed upon imbibition. The epidermis of the dag1 seed coat appears normal, and its cells exhibit a normal central volcano-shaped structure (columella; Western et al., 2000), as also shown in Figure 2, C and D. When dag1 and Ws seeds were imbibed (for 3 min) and subsequently stained for 15 min with ruthenium red, a dye that stains acidic polysaccharides (Frey-Wyssling, 1976), they showed identical mucilage layers. In both types of seeds, an outer more diffuse and an inner more compact mucilage layer (Western et al., 2000) are clearly visible (Figs. 2, E and F). However, when seeds were stained for only 15 s without prior imbibition, dag1 seeds stained much more than Ws seeds, indicating that their mucilage was more quickly extruded (Fig. 2, G and H).

Staining of the seeds with tetrazolium saltswhose uptake reflects the permeability of the testa, in that the embryo can be stained by tetrazolium only if the dye can reach it through the seed coat (Debeaujon et al., 2000)is shown in Figure 2, I and J. A much larger proportion of dag1 seeds (46%) than Ws seeds (14%) was colored by tetrazolium.

Condensed tannins (proanthocyanidins) accumulated in the endothelium of the testa and derived from progressive oxidation of uncolored proanthocyanidins are responsible for the brownish color of Arabidopsis wild-type seeds (Debeaujon et al., 2000). In this respect, no difference can be observed between mature Ws and dag1 seeds, which are equally brownish (not shown). When the vanillin staining, which stains uncolored proanthocyanidins (Debeaujon et al., 2000), was performed on immature Ws and dag1 seeds, the latter resulted manifestly more colored than the former, as shown in Figure 2, K and L. However, if the staining reaction was protracted, eventually Ws seeds resulted as intensely colored as dag1 seeds (not shown).

DAG1 Is Expressed in the Phloem of All Organs of Arabidopsis Plants

We had previously shown by in situ RNA hybridizations on Arabidopsis flowers, siliques, ovules, and immature embryos that the DAG1 gene is expressed in the vasculature of flowers and of fertilized immature siliques but not in fully mature siliques nor in ovules or embryos (Papi et al., 2000). We now extended the analysis of the expression pattern of DAG1 to the whole plant. The results of the in situ RNA hybridizations on different organs of Arabidopsis plants are reported in Figure 3. As shown in Figure 3A, in addition to being inactive in the developing embryo (Papi et al., 2000), DAG1 is not expressed in any tissue of fully developed embryos in mature seeds. In contrast, as seen in Figure 3B, a very clear DAG1 mRNA signal is visible in longitudinal sections of roots of plantlets and of adult plants. Expression of the gene is localized in the central cylinder of the root and sharply begins (see arrow) approximately 250 µm from the root tip, where differentiation of the protophloem begins (M.M. Altamura, unpublished data; Dolan et al., 1993). No signal is detected in the meristematic region of the root apex, suggesting that activation of DAG1 is associated to the functioning of the phloem rather than being related to its differentiation. In this latter case, expression in the procambium in the root apical meristem and in the procambium in the mature embryo would be observed.

View larger version (141K): [in this window] [in a new window]   Figure 3.   DAG1 is expressed in the phloem of all organs of Arabidopsis plants. Dark-field images of in situ mRNA hybridizations with tritium-labeled antisense DAG1 riboprobe on sections of: mature seed (A; bar = 100 µm); primary root including the apex (B; longitudinal section, bar = 100 µm), arrow points to where differentiation of the protophloem begins and the autoradiographic signal becomes visible; primary root (C; transverse section, bar = 50 µm), arrow points to the phloem region, where the autoradiographic signal is localized; primary root at the site where a secondary root is formed (D; transverse section, bar = 100 µm); upper part of the floral stem including the main apex (E; longitudinal section, bar = 200 µm); floral stem (F; longitudinal section, bar = 100 µm); floral stem (G; transverse section, bar = 200 µm); floral stem and attached leaf (H; transverse section, bar = 200 µm), arrows point to leaf midrib bundle and to secondary veins; floral bud (I; longitudinal section, bar = 200 µm), arrow points to the autoradiographic signal in the immature carpel.

Expression of DAG1 associated with the phloem bundles is confirmed by the signal observed in transverse sections of the root, as shown in Figure 3C (see arrow). To complete the analysis in roots, in Figure 3D is shown a transverse section of a primary root where a secondary root is already emerged from the pericycle of the former. As can be seen, in the primary root, the DAG1 mRNA signal is localized in the phloem; and in the secondary root, the signal begins where protophloem differentiates but is absent (as in Fig. 3B) from the apical meristem and adjacent procambium. Expression of DAG1 associated with the protophloem is also clearly visible in Figure 3E, where a longitudinal section of the upper part of the floral stem is shown. The DAG1 signal is not present in the apical dome and is concentrated in the phloic part of the procambial traces of the stem and of the flower pedicels. In more basal parts of the stem, high levels of DAG1 transcript are clearly visible in the phloem, as can be seen in Figure 3F, which shows (at greater magnification) the continuation of the lower part of Figure 3E. The transverse sections of the stem shown in Figure 3, G and H, confirm the specificity of expression of DAG1 in the phloem. The transverse section in Figure 3H is cut in correspondence of a cauline leaf and shows that in this latter the DAG1 mRNA signal is associated to the midrib bundle and to secondary veins (see arrows), indicating phloem-specific expression of the gene.

Finally, in Figure 3I, we report a longitudinal section of a floral bud, which shows the presence of DAG1 mRNA in the pedicel bundles and in the procambial traces connecting the receptacle to the stamen filaments and to the ovary. Also partially visible (see arrow) is the signal in the vasculature of the immature carpel, where the presence of DAG1 mRNA at later stages of development was documented by a previous series of in situ hybridizations (Papi et al., 2000). No DAG1-specific signal was detectable in any of the above sections with the DAG1 sense riboprobe (not shown). Altogether, these data indicate that the DAG1 gene is specifically expressed in the phloem of all organs of Arabidopsis plants.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

In Arabidopsis, seed germination requires light to convert the inactive Pr form of phytochrome into the active Pfr that triggers germination via a largely unknown pathway involving induction of GA biosynthesis (Yamaguchi et al., 1998) and/or increase in GA sensitivity in the embryo (Karssen and Laçka, 1986). Gibberellins stimulate growth of the embryo and elongation of the embryo radicle and/or induce the expression of genes encoding enzymes that degrade the cell walls of endosperm and seed coat (Groot and Karssen, 1987; Debeaujon and Koornneef, 2000). As a result, the embryo radicle protrudes from the seed bringing the germination process to completion.

Mutant dag1 seeds are more readily induced to germinate than Ws wild-type seeds by irradiation with red light. Comparison of photon fluence response curves clearly indicates that substantially lower red light fluence rates are needed to trigger germination of dag1 seeds than what is required for Ws seeds. Lower red-light photon fluences correspond to lower Pfr to Ptot ratios (Mancinelli, 1994), thus dag1 seeds require substantially less phytochrome signaling to germinate. This is confirmed by the effects of pulse-irradiation of the seeds with light of different wavelengths. Similar to observations by others (Shinomura et al., 1996), only red light below 690 nm had any inductive effect on phytochrome-induced germination of Ws seeds. In contrast, clear effects of much higher wavelengths could be observed on dag1 seeds. Irradiation with progressively higher wavelengths generates progressively lower Pfr to Ptot ratios (Mancinelli, 1994). Our results indicate that induction of germination of Ws seeds requires about 50% Pfr, whereas less than 4% Pfr is still effective for dag1 seeds. Induction of dag1 seed germination also requires signaling for a much shorter time. The experiment reported in Figure 1C indicates that germination of WS seeds depends on continuos signaling by Pfr for more than 12 h after the initial triggering. In contrast, far-red reversion of germination of dag1 seeds can only be achieved within 4 to 8 h from the initial inductive red pulse, indicating that inactivation of the DAG1 gene causes germination to become much sooner independent of Pfr signaling.

Mutant dag1 seeds show alterations in the seed coat. Although major structural alterations in the testa could not be detected, the staining assays point to some aberrant characteristics. Tetrazolium salts hardly penetrate the intact testa of wild-type seeds to stain the embryo. Mutants with major structural alterations in the seed coat (e.g. ats and ap2; Debeaujon et al., 2000) and mutant seeds lacking the protective layer of condensed tannins in the testa endothelium (e.g. ttg and tt; Debeaujon et al., 2000) are permeable to tetrazolium. Unstained dag1 seeds are of the same brownish color as Ws seeds, indicating that condensed tannins are present, as also shown by the coloration of immature seeds by vanillin that stains proanthocyanidins. Thus, rather than a higher permeability to tetrazolium, the much higher percentage of tetrazolium-stained dag1 seeds may reflect a greater mechanical fragility of their seed coat. The faster staining by vanillin of immature dag1 seeds rather than by hyperaccumulation of proanthocyanidins may be accounted for by a greater degradability of the dag1 testa under the staining conditions utilized. In fact, when the staining procedure was carried out for longer times, Ws immature seeds resulted as intensely colored as dag1 seeds. Staining the seed mucilage with ruthenium red provides further indication that the testa of dag1 seeds is in some way altered. Extrusion of mucilage (Western et al., 2000; Windsor et al., 2000) rapidly follows imbibition of Arabidopsis seeds. Mucilage is released from epidermal cells of the seed coat after breakage of their outer tangential cell wall (Western et al., 2000; Windsor et al., 2000) because of the rapid expansion of the mucilage (Goto, 1985) upon hydration. Ruthenium red visualizes identical mucilage capsules surrounding dag1 and Ws seeds, but staining of the former is quicker as mucilage is more readily released. A quicker release of mucilage suggests that mutation of DAG1 results in a weakened cell wall of the epidermal cells of the testa.

A possible rationalization of the data presented here, including expression of the gene in the phloem of the mother plant but not in the seed, could be that inactivation of DAG1 enhances phytochrome signaling by controlling the transport of signal transduction component(s). It is important that this effect is limited to seed germination, because fluence rate response curves of hypocotyl elongation in red light do not differ between Ws and dag1 (data not shown). The observed weakening of the testa in dag1 seeds could be a consequence of a process of degradation triggered by low light fluence rates during maturation of the seeds. DAG1 or a secondary signal that depends on an intact DAG1 gene could alternatively control phytochrome amounts in mutant embryos. Arabidopsis seeds with enhanced levels of phytochrome B require significantly less photon fluences for induction of germination and can be induced to germinate to some extent even by light of 715 nm (Shinomura et al., 1998).

DAG1 may also control the production or transport of component(s) needed for seed coat integrity. A weaker seed coat would affect positively seed germination. Besides the dag1 mutant, the only other maternal mutants affected in seed germination identified so far have altered seed coats (Koornneef, 1981, 1990; Léon-Kloosterziel, 1997; Debeaujon et al., 2000). Inactivation of DAG1 may cause alterations resulting in a more fragile seed coat.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

Arabidopsis, ecotype Ws and dag1 mutant plants were grown in a growth chamber (24°C/21°C, 16/8 h day/night, 300 µEinstein/m²) as previously described (Papi et al., 2000).

Light Treatments and Germination Assays

Stored Ws and dag1 seeds, were sown in triplicate in petri dishes on four layers of filter paper 2043BMGL (Schleicher & Schüll, Dassel, Germany), soaked with 5 mL of distilled water, in a green-safelight chamber. Before any light treatment, seeds were irradiated for 30 min with a far-red light (Osram Linestra fluorescent tubes, combined with filters KG 3/2501/3 and PG 627/3, Osram, Munich; Schott, Mainz, Germany; Röhm und Haas, Darmstadt, Germany; fluence rate 20 µmol m² s¹). A 24-h dark treatment at 4°C was followed by the different light treatments. After the light treatment, seeds were kept in darkness at 25°C, and germination rates were scored after 6 d. Monochromatic light of different wavelengths (660, 685, 690, 694, 703, 718, 743, 758, 776, and 794 nm) was applied using Leitz Prado light projectors with appropriate interference filters (Schott). To obtain long-wavelength far-red light, an 8-mm-thick RG9 cut-off filter (Schott) was used (maximal transmission 775 nm).

Seed Coat Analysis and Seed Staining

For mucilage detection, seeds were incubated for 15 min (after imbibition for 3 min in water) or for 15 s (without prior imbibition) in an aqueous solution of 0.03% (w/v) ruthenium red at room temperature and rinsed with water before observation under a stereomicroscope (Leica MZ12, Weitzlar, Germany). For the tetrazolium staining assay, intact seeds were incubated in a 1% (w/v) aqueous solution of 2,3,5-triphenyltetrazolium chloride (Merck, Darmstadt, Germany) at 30°C in darkness for 2 d according to Wharton (1955). For the vanillin staining assay, intact immature seeds were incubated for 10 min or longer in a solution of 1% (w/v) vanillin in 6 N HCl at room temperature, as described by Aastrup et al. (1984).

Microscopic Analysis of Seeds

Mature seeds imbibed for 30 min in water were fixed for 24 h at 4°C in 5% (v/v) glutaraldehyde before embedding in Technovit 7100 historesin (Kulzer, Hereaus, Germany). Sections (5 µm thick) obtained with a microtome (Leica RM 2145) were stained for 1 min with 1% (w/v) toluidine blue O in 0.1 M phosphate buffer at pH 7.2. Observations and photographs were taken under a DAS Leica DMRB microscope (Leica, Heerbrugg, Switzerland). Seed epidermis was also analyzed with Normaski differential interference optics applied to the same microscope.

In Situ mRNA Hybridizations

A 400-bp PCR fragment containing the C terminus of the DAG1 coding sequence up to the 3'-untranslated region was cloned in the pCR2.1 vector (Stratagene, La Jolla, CA) as described previously (Papi et al., 2000). For the antisense riboprobe, the construct was linearized and in vitro transcribed with T7 RNA polymerase in the presence of [³H]UTP. The sense probe was derived from a construct containing the same PCR-amplified fragment cloned in the opposite orientation. Conditions for tissue fixation, paraffin embedding, hybridization (on 8-µm sections), and washings were as described by Drews et al. (1991). The sections were observed and photographed under dark field with a DAS Leica DMRB microscope.