Embryo growth, testa permeability, and endosperm weakening are major targets for the environmentally regulated inhibition of Lepidium sativum seed germination by myrigalone A

Abstract

Myrigalone A (MyA) is a rare flavonoid in fruit leachates of Myrica gale, a deciduous shrub adapted to flood-prone habitats. As a putative allelochemical it inhibits seed germination and seedling growth. Using Lepidium sativum as a model target species, experiments were conducted to investigate how environmental cues modulate MyA’s interference with key processes of seed germination. Time course analyses of L. sativum testa and endosperm rupture under different light conditions and water potentials were combined with quantifying testa permeability, endosperm weakening, tissue-specific gibberellin (GA) and abscisic acid (ABA) contents, as well as embryo growth and apoplastic superoxide production important for cell expansion growth. Lepidium sativum testa permeability and early water uptake by imbibition is enhanced by MyA. During late germination, MyA inhibits endosperm weakening and embryo growth, both processes required for endosperm rupture. Inhibition of embryo cell expansion by MyA depends on environmental cues, which is evident from the light-modulated severity of the MyA-mediated inhibition of apoplastic superoxide accumulation. Several important key weakening and growth processes during early and late germination are targets for MyA. These effects are modulated by light conditions and ambient water potential. It is speculated that MyA is a soil seed bank-destroying allelochemical that secures the persistence of M. gale in its flood-prone environment.

Introduction

Seed shedding combined with appropriate seed germination responses to environmental cues, including ambient light conditions (e.g. darkness in the soil seed bank), soil water status, or allelochemicals produced by competing plants, control the persistence of a species in the local habitat, as well as long-distance dispersal important for expanding the species range (Donohue et al., 2005; Batlla and Benech-Arnold, 2010). Germination timing is under extremely strong natural selection and influences phenotypic expression of and selection on post-germination traits as it is a major determinant of successful seedling establishment, survival, and colonization of a habitat (Donohue et al., 2010). Phenological adaptation to environmental changes studied with sets of Arabidopsis thaliana recombinant inbred lines demonstrated that natural variation in germination responses to environmental factors is associated with differences in seed-related mechanisms involving hormones and the envelopes covering the embryo (Huang et al., 2010; Barua et al., 2011). While gibberellin (GA) sensitivity was associated with germination responses to multiple environmental factors, abscisic acid (ABA) sensitivity and testa permeability were associated with specific germination responses. Environmental effects on the mechanical and chemical properties of the testa were proposed to constitute major mechanisms for maternal effects on coat dormancy and germination traits (Roach and Wulff, 1987; Debeaujon and Koornneef, 2000; Debeaujon et al., 2000; Donohue, 2009).

In the non-dormant seed state, germination starts with the uptake of water by imbibition of the dry seed, followed by embryo expansion and seed swelling (Holdsworth et al., 2008; Weitbrecht et al., 2011; Linkies and Leubner-Metzger, 2012). The uptake of water is triphasic, with a rapid initial uptake (phase I, i.e. imbibition) followed by a plateau phase (II), and finally phase III water uptake. The completion of germination by radicle protrusion through all seed envelopes depends on the increasing growth potential of the embryo and on the weakening of the restraint of the various seed envelopes. Embryo growth is promoted by GA and inhibited by ABA, and light-promoted A. thaliana seed germination is associated with light-increased GA/ABA ratios (Holdsworth et al., 2008). Arabidopsis thaliana embryo elongation is achieved by cell expansion in specific regions of the radicle–hypocotyl axis (Sliwinska et al., 2009). The envelopes of most angiosperms are made up of the testa (seed coat) and the endosperm, which is retained to a variable degree in mature seeds (Linkies et al., 2010). Arabidopsis thaliana testa mutants show reduced dormancy that is caused by alterations of the testa characteristics including increased permeability (Debeaujon and Koornneef, 2000; Koornneef et al., 2002). The GA requirement for A. thaliana seed germination is determined by testa characteristics, embryonic growth potential, and embryonic ABA content. In many endospermic seeds, germination progresses in two visible steps, with testa rupture and endosperm rupture being two sequential events (Leubner-Metzger, 2002; Petruzzelli et al., 2003; Linkies and Leubner-Metzger, 2012). Mature seeds with a thin endosperm layer include the Brassicaceae species A. thaliana, Lepidium sativum (garden cress), and Sisymbrium officinale (Müller et al., 2006; Holdsworth et al., 2008; Iglesias-Fernandez and Matilla, 2010). Lepidium sativum endosperm rupture is preceded by weakening of the micropylar endosperm (CAP) that covers the radicle. Embryo growth and endosperm weakening depend on hormonally regulated cell wall remodelling mechanisms mediated by hydrolytic enzymes, expansins, and reactive oxygen species (ROS) in the apoplast (Ni and Bradford, 1993; Toorop et al., 2000; Petruzzelli et al., 2003; Da Silva et al., 2008; Müller et al., 2009; Morris et al., 2011; Voegele et al., 2011).

Myrica gale (‘sweet gale’, ‘bog myrtle’, Myricaceae) is a deciduous shrub native to Northern and Western Europe and Canada, adapted to flood-prone habitats (Skene et al., 2000). It grows in acidic peat bogs and at the intertidal zone of lakes and rivers which are often flooded by frequent rises and falls in water level. Myrica gale fruits and leaves secrete resin droplets containing essential oils. Myrica gale fruit exudates contain rare flavonoids, with myrigalone A (MyA) being the major C-methylated dihydrochalcone (Mathiesen et al., 1995; Popovici et al., 2011). Several biological activities have been reported for MyA, including radical scavenging and the inhibition of endoreduplication in radicles ofL. sativum under continuous light conditions (Mathiesen et al., 1997; Oracz et al., 2012). MyA has also been shown to interfere with GA metabolism by inhibiting GA3-oxidase reactions and with GID1-type GA signalling pathways to inhibit endosperm rupture of L. sativum imbibed in the light and to inhibit shoot and root growth of dark-grown eudicot (cress, mustard, and knotweed) and monocot (sorghum) seedlings (Popovici et al., 2011; Voegele et al., 2011; Oracz et al., 2012). The MyA inhibition of seedling growth is therefore evident for a wide range of weed and crop species, as well as for invasive knotweed (Fallopia×bohemia) for which the protective effect of M. gale against invasion is currently being investigated in the ecosystem.

It has been proposed that MyA is an allelochemical with a novel mode of action, but very little is known about the way in which MyA inhibits seed germination. Oracz et al. (2012) found that MyA interferes with GA biosynthesis, apoplastic ROS production, and endosperm cap weakening during the late phase of L. sativum seed germination in continuous light and at optimal conditions for water uptake. Here, in complementary work, investigations are carried out to determine how environmental cues (darkness and restrictive ambient water potential) affect MyA inhibition in order to elucidate its complex and developmentally regulated mode of action. It is shown that MyA interferes with several key mechanisms throughout the whole process of seed germination and that its inhibitory effect is under strong environmental control. It was found that the effects of MyA on seed germination are strongly intensified in darkness compared with light. Furthermore, it is shown that the very early processes of initial embryo imbibition, as well as testa permeability, are a major target for MyA during the early phase of germination. During the late phase of germination, it is demonstrated here by quantitative analysis that embryo growth and molecular mechanisms related to cell expansion growth of the radicle–hypocotyl axis, as well as endosperm cap weakening are affected by MyA. The severe impact of environmental cues such as water deficiency stress and light conditions on the effects of MyA presented here is placed into a speculative model to explain the complex interactions in connection with the ecophysiological roles of MyA as an allelochemical of M. gale important for the adaptation to flood-prone habitats.

Materials and methods

Germination conditions and puncture-force measurements

MyA (purity >99%), extracted and purified as described by Popovici et al. (2011), was generously provided by Professor G. Comte, CESN, University and CNRS, Lyon, France, and used as a methanol stock, corresponding to 0.35% (v/v) methanol in the medium if 5×10^–4 M MyA was used. Germination of after-ripened seeds of L. sativum L. FR14 (‘Keimsprossen’, Juliwa) was analysed using three replicates of 50 seeds per 9cm Petri dish containing two layers of filter paper and 6ml of sterile, distilled water with 0.35% (v/v) methanol (‘CON medium’) (Oracz et al., 2012). Seeds were incubated at 18 °C in a Sanyo Versatile Environmental Test Chamber (MLR-350) in continuous white light (~75 µmol s^–1 m^–2). A binocular stereo microscope was used to score the percentages of testa and endosperm rupture of the seed populations. Puncture-force measurements (Müller et al., 2006) were used for quantifying endosperm cap weakening. For organ- or tissue-specific analyses, whole embryos, RAD (lower one-third of the hypocotyl–radicle axis) or CAP (micropylar endosperm) tissues were used from seeds displaying testa rupture but not endosperm rupture.

Embryo size measurements

Fifty seeds for each data point were imbibed with different concentrations of polyethylene glycol (PEG) 6000 (Carl Roth, Karlsruhe, Germany) and embryos carefully extracted at the times indicated. Embryos were transferred onto a black ceramic plate and images were taken with a binocular stereo microscope (Leica MZ 125) connected to a digital camera (Leica DFC480) at 20-fold magnification using Leica IM1000 software (V4.0 release 132). Twenty-four-bit RGB images at a resolution of 2560×1920 pixel (pixel aspect ratio=1) were saved in TIFF format. Image analysis was performed using the ImageJ distribution Fiji (http://fiji.sc/wiki/index.php/Fiji) with the ImageJ kernel version 1.45r. Following conversion to 8-bit images, thresholding was performed using the Auto Threshold plugin v1.14 (http://fiji.sc/wiki/index.php/Auto_Threshold) implemented in Fiji. The Auto Threshold plugin can binarize 8-bit images using various global (histogram-derived) thresholding algorithms. The ‘Default’ algorithm was selected for thresholding, which is a variation of the IsoData method for image thresholding (Ridler and Calvard, 1978). Options controlling the thresholding algorithm were set to: ignore white=true; ignore black=true; white objects on black ground=true. The resulting binarized images (showing a white embryo on a black background) were used for measuring the actual embryo sizes. Embryo mask areas were determined using ImageJ’s ‘Analyze Particles’ command. Options for measuring the object were set to: size (pixel^2)=100 000–infinity (i.e. no objects smaller than this are going to be measured, thus excluding image artefacts resulting from the thresholding); circularity: 0.00–1.00 (i.e. allowing irregular shapes of the object); include holes=true; show=masks. This operation provides a pixel value for the measured area as well as an image showing the filled outline of the measured area; that is, a black embryo shape on a white background. These embryo mask images were used in the figures to illustrate the area quantification process. Each embryo mask was checked manually to control that the embryo shape was identified properly and without measuring artefacts. Using the described settings for image acquisition and analysis results in each pixel of the measured area representing a real physical embryo area of 7.4 µm².

Testa permeability assay using tetrazolium staining

Entire L. sativum seeds were incubated for 3h or 15h in continuous light at 18 °C with tetrazolium staining solution (Graeber et al., 2010) containing 0.35% methanol (CON) or 5×10^–4 M MyA. The embryos were subsequently extracted, classified according to their staining intensity and patterns, and photographed. Enhanced staining intensity indicates enhanced testa permeability (Debeaujon et al., 2000).

Quantification of hormone and superoxide contents

A total of 200 RADs and 200 CAPs were isolated from L. sativum seeds imbibed for 15h, immediately frozen in liquid nitrogen, and used for quantification of ABA and GA metabolites as previously described (Oracz et al., 2012). To quantify apoplastic superoxide (O₂·^–) contents, the XTT assay was used with 10 RADs or 20 CAPs as described by measuring the absorbance of four biological replicates at 470nm (Müller et al., 2009; Oracz et al., 2012). Apoplastic O₂·^– was localized in situ by nitroblue tetrazolium (NBT) histostaining (dark-blue colour of insoluble formazan compounds) (Oracz et al., 2007).

Results

The MyA-mediated inhibition of endosperm weakening and rupture during L. sativum seed germination is more severe in darkness compared with light

Lepidium sativum two-step seed germination was investigated by scoring testa (seed coat) rupture over time, as well as subsequent endosperm rupture which is connected to visible radicle protrusion through the weakened micropylar endosperm (CAP) and marks the completion of germination. Figure 1 shows that testa and endosperm rupture kinetics of L. sativum seeds imbibed in continuous light or in darkness are highly similar (CON). It was shown earlier (Oracz et al., 2012) that the putative allelochemical MyA delays endosperm weakening and rupture of L. sativum seeds imbibed in continuous light. Interestingly, and in contrast to CON, when seeds were incubated in the presence of MyA, the inhibiting effect of MyA on endosperm rupture was more severe in darkness compared with light. While the time to reach 50% endosperm rupture (ER_50%) was ~16h for light- or dark-imbibed CON seeds, treatment with 5×10^–4 M MyA caused an ER_50% of ~33h in light and ~223h in darkness (Fig. 1b). A concentration of 5×10^–4 M MyA was chosen as a biologically relevant concentration as it results in a close to saturating response and causes the same delay in endosperm rupture as fruit leachates of M. gale on L. sativum seeds imbibed in continuous light (Oracz et al., 2012). As in light, the MyA effect on endosperm rupture was dose dependent in darkness, and testa rupture was not affected (Fig. 1a; Supplementary Fig. S1a available at JXB online). The MyA inhibition in darkness was associated with a high incidence (~60%) of seeds in the population with atypical endosperm rupture, in which the CAP is torn off at its base instead of being penetrated by the radicle (Supplementary Fig. S1c). The stronger effect of MyA in darkness compared with light was mediated, at least in part, by prolonged inhibition of endosperm weakening in darkness (Fig. 1b). Whereas in the light MyA allowed 100% germination and acted phytotoxically on the growing seedling (Popovici et al., 2011; Oracz et al., 2012), in darkness MyA was already phytotoxic prior to the completion of germination. It decreased embryo viability and caused embryo death during the prolonged incubation (Supplementary Fig. S1b). To exclude that MyA is chemically inactivated upon light irradiation, the effect of MyA pre-incubation (light versus darkness) was compared prior to the addition of seeds (Fig. 1c). Light- or dark-pre-incubated MyA inhibited L. sativum endosperm rupture equally, showing that MyA’s effect is persistent under the ambient incubation conditions used here. Endosperm weakening is clearly a major target of MyA, but it has not been investigated so far whether MyA acts in addition on L. sativum embryo growth as another major target.

Fig. 1.

The effect of myrigalone A (MyA) on Lepidium sativum seed germination in continuous light and in darkness. (a) The effect of different concentrations of MyA on the time required for germinating seed populations to reach 50% testa rupture (TR_50%) or endosperm rupture (ER_50%) displayed as dose–responses. (b) Endosperm cap puncture-force (red circles) and endosperm rupture kinetics during germination without (CON) or with 5×10^–4 M MyA added. (c) Endosperm rupture kinetics of germinating seeds, incubated in continuous white light without (CON) and with 5×10^–4 M MyA pre-incubated for 35h in darkness or continuous light. Mean values ±SE of ≥3 replicates of 50 seeds each are shown for testa and endosperm rupture; mean values ±SE of 10–50 seeds are shown for the puncture-force measurements. (This figure is available in colour at JXB online.)

Development of an image analysis method to quantify L. sativum embryo growth

To quantify L. sativum embryo sizes and thereby embryo expansion growth during the process of germination, embryos were extracted from their seed envelopes at the indicated times, examined under a stereo microscope, and photographed thereafter. These raw images were analysed using the ImageJ software distribution Fiji using the workflow presented in Fig. 2a and described in detail in the Materials and methods. The size/shape detection process of this image analysis method worked accurately for different stages such as embryos extracted from fully imbibed seeds (e.g. 3h CON+0mM PEG) with a semi-opened hypocotyl hook, as well as for only partially imbibed seeds in the presence of osmotica (e.g. 3h CON+30mM PEG) with a closed hypocotyl hook (Fig. 2a). The CON+30mM PEG incubation solution corresponded to an ambient water potential (Ψ_medium) of –0.58MPa and caused a 1.6-fold smaller embryo size compared with CON (Ψ_medium=0MPa) at 3h (Fig. 2b).

Fig. 2.

A computer-based image analysis method to quantify L. sativum embryo sizes. (a) Workflow illustrating quantification of embryo sizes. After seed incubation for a specified time and condition, embryos were carefully extracted and photographed. Raw images were thresholded and image particle analysis was used to create embryo masks. The embryo mask areas were quantified using the Fiji distribution of the image analysis software ImageJ as described in detail in the Materials and methods. Using this workflow, 1 pixel of the embryo mask area equals 7.4 µm² real embryo area. Example raw images and mask areas are presented for embryos extracted from 3 h-imbibed seeds with or without 30mM PEG added; pixel and mm² values correspond to the presented embryo mask areas. (b) The effect of decreasing ambient water potential (Ψ_medium) on embryo sizes from seeds incubated for 3h in the light without (CON) and with 5×10^–4 M MyA and the osmoticum PEG, as indicated. Mean values ±SE calculated from ≥50 extracted embryos are presented.

MyA accelerates early water uptake by imbibition ofL. sativum seeds and inhibits embryo growth during the late germination phase

The method developed to quantify L. sativum embryo sizes (Fig. 2a) was applied to embryos extracted at the indicated times from seeds incubated in 0, 10, 20, and 30mM PEG. These PEG concentrations correspond to ambient water potentials (Ψ_medium) of 0, –0.05, –0.25, and –0.58MPa, respectively, for which the embryo sizes in 3 h-imbibed seeds are presented as mean values ± SE (Fig. 2b). In contrast to higher Ψ_medium during the early germination phase, at low Ψ_medium (–0.58MPa) MyA enhances water uptake by imbibition, leading to a MyA-mediated ~1.4-fold larger embryo size at 3h. To verify with an independent method that this MyA-mediated larger embryo size at –0.58MPa is indeed due to increased water uptake, the embryo masses were compared by weighing, and it was found that the 3h MyA-imbibed embryos are heavier compared with the 3h CON embryos, but that their weight is equal after drying (Supplementary Fig. S2 at JXB online). In Fig. 3, the data distribution of the embryo sizes over time is presented as box–whisker plots for the different treatments. Figure 3a shows that without MyA (CON series) the embryo size increased over time at all Ψ_medium, but also that the early (3h) water uptake by imbibition is severely hampered by –0.58MPa (30mM PEG; Figs 2b, 3a; Supplementary Fig. S3a). At 22h without osmoticum (CON+0mM PEG); that is, at ER_50%, a 1.2-fold growth in embryo size (compared with 3h) to 5.4±0.1mm² and an opened hypocotyl hook was evident (Fig. 3a). In contrast to this, the embryo size at low Ψ_medium (CON+30mM PEG) had reached only 4.3±0.1mm² at 22h, no endosperm rupture had occurred, and the hypocotyl hook did not fully open (Fig. 3a). Although an ~1.5-fold increase in embryo size was evident from 3h to 22h in CON+30mM PEG, the absolute embryo size at 22h remained below the absolute embryo size in CON+0mM PEG at 3h (4.6±0.1mm²). The –0.58MPa decrease in Ψ_medium therefore severely inhibited early water uptake by imbibition (3h) and later embryo growth to a size associated with ER_50% (22h).

Fig. 3.

The effect of MyA on the embryo mask areas during seed germination of L. sativum. Time course of mask areas of embryos extracted from seeds at different times during incubation (a) without (CON) and (b) with 5×10^–4 M MyA, and 0, 10, 20, or 30mM PEG in continuous light. Typical examples of embryo masks from each seed population at different times are shown below the graphs. Box–whisker plots each representing data from 50 embryos are shown, with the whiskers representing the minimum and maximum of the data distribution, the bottom and top of the boxes the 25th and 75th percentiles, and the horizontal bar the median.

In contrast to the CON series, for the early phase of water uptake by imbibition (3h, phase I water uptake), treatment of seeds with MyA completely abolished the severe inhibition of the embryo size increase by low Ψ_medium (MyA+30mM PEG) when compared with MyA+0mM PEG (Fig. 3b). After this initial swelling at 3h to 4.0±0.1mm², the embryo size did not increase further over time in MyA+30mM PEG (Fig. 3b; Supplementary Fig. S3a at JXB online). At 35h, the MyA-treated seed population without osmoticum had reached ER_50% and the embryos had grown in size to 6.7±0.2mm², while the embryos of the MyA+30mM PEG series remained at 3.8±0.1mm², without a fully opened hypocotyl hook and with no endosperm rupture (Fig. 3; Supplementary Fig. S3b). By plotting the endosperm rupture responses against the embryo sizes at the various Ψ_medium (Fig. 4a), it was found that the embryo size associated with the onset of endosperm rupture is roughly equal for CON and MyA (~4.6–4.7mm²). In contrast to this, the progression of endosperm rupture required larger embryo sizes in the presence of MyA. Regarding their endosperm rupture responses, the 22h CON and the 35h MyA seed populations are comparable. The ER_50% times (without osmoticum) were associated with embryo sizes of 5.4±0.1mm² (CON) and 6.8±0.2mm² (MyA); that is, 1.3-fold larger embryos upon MyA treatment (Fig. 4). Therefore, and in agreement with the inhibition of endosperm weakening by MyA (Fig. 1a), further embryo growth is required in the presence of MyA to achieve the same endosperm rupture response (Figs 3, 4; Supplementary Fig. S3).

Fig. 4.

The effect of MyA on the relationship between embryo size, growth, and endosperm rupture at different ambient water potentials (Ψ_medium). (a) The percentages of endosperm rupture at different times during the seed germination of L. sativum at various Ψ_medium without (CON) and with 5×10^–4 M MyA (Fig. 3) are plotted against the corresponding embryo mask areas. Note that while the onset of endosperm rupture was associated with an equal embryo size of 4.6–4.7mm², the progression of endosperm rupture required larger embryos in the MyA series compared with CON; the embryo sizes required for 50% endosperm rupture (ER_50%) were 5.4±0.1mm² (CON) and 6.8±0.2mm² (MyA), and therefore increased 1.3-fold upon MyA treatment. (b) Embryo sizes and endosperm rupture (upper panel), and embryo growth rates (lower panel; slopes of the curves in Fig. 3) at different Ψ_medium during late germination. Mean values ±SE are presented from data shown in Fig. 3, and Supplementary Fig. S3b at JXB online. (This figure is available in colour at JXB online.)

To analyse further the effect of MyA on embryo cell expansion growth during late germination [i.e. well separated from the phase I water uptake (imbibition)], the phase II/III water uptake periods were analysed. When all Ψ_medium are taken into consideration, they correspond to incubation times >15h (CON) and >8h (MyA) (Fig. 3). Figure 4b shows that at 22h the embryo sizes in the MyA series are smaller compared with the CON series at any Ψ_medium. MyA therefore inhibits embryos from growing above a critical size required for the onset of endosperm rupture (grey-shaded area, Fig. 4b, upper panel) until 22h. In agreement with this, MyA inhibited the embryo growth rates (growth/time) between 15h and 22h (Fig. 4b, lower panel). This inhibition is, however, transient, as later at 35h in the MyA series when the embryos are larger (Fig 4b, upper panel), the embryo growth rate is higher (Fig. 4b, lower panel) at high Ψ_medium (0 to –0.05MPa) compared with the physiologically similar (in terms of endosperm rupture) 22h CON series. At these physiologically comparable times, namely 22h CON and 35h MyA, the transition zone for passing the critical embryo size required for the onset of endosperm rupture (grey-shaded area, Fig. 4b, upper panel) seems to be about –0.3MPa for both CON and MyA. When the MyA-mediated transient inhibition of embryo growth is overcome during the very late phase of germination, the MyA series embryos need to grow to a larger sizes to yield ER_50% compared with CON. This is obviously due to the inhibitory effect of MyA on endosperm weakening.

Embryo growth is mainly due to radicle–hypocotyl axis (RAD) elongation, and an excess in RAD growth is required to accomplish endosperm rupture in the presence of MyA

To answer the question of which part of the embryo contributes most to the measured embryo growth, the embryo was digitally dissected into two parts: the cotyledons (COT) and the radicle–hypocotyl axis (RAD), with the apical meristem being the cutting site (Fig. 5a). Figure 5b shows that the increases in RAD sizes contribute most to embryo growth. The relative RAD size increases are 4- to 5-fold higher compared with those of the COT, for both the CON and the MyA series. The RAD is therefore the major organ that contributes to embryo growth during L. sativum seed germination. An ~3-fold higher RAD size increase is observed in the presence of MyA to achieve ER_50% at 35h (MyA) in comparison with CON (ER_50% at 22h) (Fig. 5b). Since RAD is the main embryo part contributing to overall growth, the transient inhibition of embryo growth by MyA mentioned in the previous section is due to inhibition of RAD growth. Therefore, in the presence of MyA, the embryos need longer times to grow to larger sizes required for rupture of the less weakened endosperm compared with CON treatment (Figs 3–5).

Fig. 5.

The contribution of different embryo parts to the increase in embryo mask area during L. sativum seed germination.(a) Workflow for quantification of COT and RAD mask areas.The raw images were digitally dissected at the apical meristem with a 90 ° angle towards the stem into a ‘cotyledon’ (COT) part and a ‘radicle–hypocotyl axis’ (RAD) part. COT and RAD masks were created and their areas quantified as shown in Fig. 2a.(b) Increase in COT and RAD mask areas of embryos extracted from seeds incubated without (CON) or with 5×10^–4 M MyA. The relative increase of the COT and RAD sizes from 3h to the time for ER_50% for CON (22h) and MyA (35h) was calculated from 50 embryo raw images taken from seeds incubated in continuous light for each indicated time point. Mean values are presented;SE <1% for CON and MyA-COT, 7% for MyA-RAD.

MyA enhances L. sativum testa permeability during imbibition, but does not affect the kinetics of testa rupture at different water potentials

Figure 6a shows that decreasing water potentials (Ψ_medium) generated by increasing osmoticum concentrations inhibited testa rupture in a dose-dependent manner, with 50% testa rupture (TR_50%) values of ~11h and 40h for 0 and 30mM PEG, respectively. As during early germination at 3h in the case of 30mM PEG the embryo was 1.3-fold larger for MyA compared with CON (Fig. 2b), it was speculated that the MyA-mediated early facilitation of water uptake may be due to promotion of testa rupture. However, it was found that MyA did not affect the kinetics of testa rupture at any PEG concentration tested (Fig. 6a). As there was no significant effect of MyA on testa rupture, it was investigated whether the MyA-mediated early facilitation of water uptake by the embryo was associated with increased testa permeability. To test for testa permeability, tetrazolium assays were conducted with whole seeds, a method that is known from work with A. thaliana testa mutants to detect differences in testa permeability (Debeaujon et al., 2000). Lepidium sativum seeds were imbibed in tetrazolium salt solution in the absence (CON) or the presence of MyA, and the embryos were extracted from the seeds at 3h or 15h (Fig. 6b–d). Figure 6b shows for embryos from 3 h-imbibed seeds that for 0mM PEG there is no obvious difference in the weakly orange embryo staining between CON and MyA. In contrast to this, for seeds imbibed in tetrazolium salt solution with 30mM PEG, the histochemical staining of the 3 h-imbibed embryos of the MyA series appeared to be more intense compared with CON (Fig. 6b). This difference was further enhanced for embryos from seeds imbibed in tetrazolium salt solution for 15h with 30mM PEG (Fig. 6d). Here the CON series seeds delivered 15h embryos with only background staining intensity, suggesting that the testa layers of CON seeds have lower permeability. In contrast to this, the MyA-treated seeds delivered approximately one-third orange-stained and two-thirds red-stained embryos (Fig. 6d), demonstrating that MyA increased testa permeability during early seed germination. This increased testa permeability suggests that MyA causes testa weakening, which in turn may require a lower embryo force for its rupture. The finding that the embryo sizes at the onset of testa rupture and at TR_50% are smaller in MyA-treated seeds compared with CON (Fig. 6e) is in agreement with this hypothesis. This indicates that MyA-enhanced testa permeability is another major target by which MyA exerts its effects on seed germination.

Fig. 6.

The effect of MyA on testa characteristics. (a) Kinetics of testa rupture of L. sativum during seed germination without (CON) and with 5×10^–4 M MyA added, and incubated in continuous light with 0, 10, 20, and 30mM PEG. (b–d) Testa permeability test using the tetrazolium assay. Entire seeds were incubated in tetrazolium staining solution for 3h (b) and 15h (d) without (CON) and with 5×10^–4 M MyA together with 0 and 30mM PEG; embryos were extracted, and staining was scored. (c) Embryos extracted from 3h CON and MyA-treated seeds were stained as positive controls. (e) The relationship between embryo sizes and testa rupture (onset and TR_50%) for CON and MyA at different PEG concentrations. Mean values ±SE of 50 embryos.

MyA inhibits apoplastic superoxide accumulation in the embryo required for cell elongation growth and this effect is more severe in darkness compared with the light

It was shown earlier (Oracz et al., 2012) that MyA inhibits the production of apoplastic O₂·^– in the growth zones of the RAD of embryos extracted from seeds imbibed in continuous light. It is shown here that MyA also inhibits the production of apoplastic O₂·^– in the growth zones of the RAD of embryos extracted from seeds imbibed in darkness (Fig. 7). The NBT histostains demonstrate that MyA inhibits apoplastic O₂·^– production in the RAD of embryos extracted from seeds imbibed for different times in darkness (Fig. 7a). In agreement with a stronger inhibition of endosperm rupture by MyA in darkness (Fig. 1), the quantification of apoplastic O₂·^– in the RAD by the XTT assay showed that an ~2-fold stronger reduction in apoplastic O₂·^– in the RAD was evident in darkness compared with the light (Fig. 7b). In agreement with a stronger inhibition of endosperm weakening by MyA in darkness (Fig. 1a), the XTT quantification showed that a stronger reduction of apoplastic O₂·^– in the CAP was evident in darkness compared with the light (Fig. 7c). The stronger inhibition of apoplastic O₂·^– production at 22h by MyA in darkness compared with the light could be due either to directly lowered MyA contents resulting in a decreased radical scavenger effect or to an indirect effect by altered GA/ABA levels upon light irradiation.

Fig. 7.

Accumulation of apoplastic superoxide (O₂·^–) in L. sativum embryos extracted from seeds incubated without (CON) or with 5×10^–4 M MyA in darkness. (a) In situ localization of the superoxide radical O₂·^– by NBT histostaining of embryos isolated from CON- and MyA-treated seeds incubated in darkness for 15h and 22h. Seeds without endosperm rupture (–ER) are presented for CON and MyA-treated seeds at 15h and 22h of incubation and with completed endosperm rupture (+ER) for CON at 22h. Left images show the lower, and right images show the upper side of the embryos as positioned in the seed during incubation in the Petri dish. Quantification of apoplastic O₂·^– production using the XTT assay in (b) RAD (radicle+lower hypocotyl) and (c) CAP (micropylar endosperm) isolated from CON and MyA seeds incubated for 22h in darkness (this work) and for comparison in continuous light (Oracz et al., 2012). Mean values ±SE of four biological replicates.

MyA alters gibberellin metabolism duringL. sativum seed germination in a tissue-specific and light-dependent manner

MyA severely interfered with GA metabolism and signalling of L. sativum seeds imbibed in continuous light (Oracz et al., 2012). Supplementary Fig. S4 at JXB online shows that MyA did not appreciably affect the ABA contents in the RAD, but severely interfered with GA metabolism in the RAD in a similar manner in darkness and light. In the RAD, MyA caused an ~3-fold decrease in the (bioactive) GA₄ contents, while the inactive precursor GA₉ accumulated ~400- (darkness) and 200-fold (light), demonstrating that MyA inhibits important steps catalysed by GA3-oxidase and thereby may act by decreasing embryo growth (Supplementary Fig. S4b). In contrast to the inhibitory effects of MyA on the GA₄ contents in the RAD in light and darkness, there was no appreciable inhibitory effect on GA₁ in the RAD, but a differential effect of light/darkness on the GA₁₃ contents (Supplementary Fig. S4c, d). MyA also affected the GA metabolite contents of the CAP (Supplementary Table S1), but there was no change that could explain the stronger inhibition of endosperm weakening (Fig. 1a) by MyA in darkness compared with the light.

Discussion

MyA enhances testa permeability and early embryo water uptake by imbibition

Early water uptake by imbibition is driven by the difference in water potential ΔΨ between the ambient incubation medium (Ψ_medium) and the embryo. For the embryo water potential Ψ_embryo, the matric potential is primarily responsible for imbibition, whereas the osmotic potential becomes more important later during germination (Woodstock, 1988; Schopfer, 2006; Weitbrecht et al., 2011). Final values reached for L. sativum embryo water uptake by imbibition at low Ψ_medium (–0.58MPa) were equal without and with MyA treatment, but MyA caused them to be reached much earlier, already at 3h. It is concluded that MyA does not alter Ψ_embryo, but increases the permeability of the seed envelopes to allow faster water uptake. In agreement with this, it was found that MyA increases testa permeability for tetrazolium salts. Several A. thaliana mutants as well as genetic variants of Brassica spp. have increased testa permeability for tetrazolium salts linked to altered testa structure and/or pigmentation (Debeaujon and Koornneef, 2000; Debeaujon et al., 2000; Lepiniec et al., 2006). For other Brassicaceae with heterogeneous seed populations (e.g. Brassica rapa and Sisymbrium officinale), higher water uptake rates and lower dormancy were associated with altered hormone sensitivities in lighter coloured seeds (Matilla et al., 2005; Iglesias-Fernandez et al., 2007). The MyA-mediated increase in the water uptake rate (enhanced water flow-through) and testa tetrazolium salt permeability during L. sativum seed imbibition therefore seem to be caused by MyA-induced alterations of the testa properties. Reduced testa pigmentation is associated with enhanced inhibitor leakage during Sinapis arvensis imbibition (Duran and Retamal, 1989). In Fabaceae species such as soybean, lighter coloured seeds often imbibe too rapidly and this may cause imbibitional damage [i.e. higher solute leakage and decreased vigour and viability (Powell et al., 1986; Chachalis and Smith, 2000; Norikazu and Setsuko, 2008)]. Micro-magnetic resonance imaging of soybean seed imbibition demonstrated that the testa regulates water uptake and minimizes imbibitional injury by ensuring optimal reconstitution of cellular structures (Koizumi et al., 2008). Besides soybean seed testa colour, structural aspects, such as a relatively loose adherence of the testa to the embryo, are known to be associated with an enhanced imbibitional water uptake (Norikazu and Setsuko, 2008). Although MyA did not alter the kinetics of L. sativum testa rupture, the onset and the TR_50% cardinal points of testa rupture were associated with a smaller embryo size when compared with CON (Fig. 6). This finding is in agreement with a MyA-mediated loose adherence of the testa, increased testa permeability, and, in turn, a weaker testa leading to easier rupture already by a less swollen embryo. Increased testa permeability also allows facilitated uptake of inhibitors and phytotoxins from the ambient medium into the embryo, as is known from the genetic weakening for A. thaliana testa mutants (Debeaujon and Koornneef, 2000; Debeaujon et al., 2000) and as can be expected from the biochemical weakening caused by MyA (this work). In agreement with a negative impact of the MyA-induced faster L. sativum embryo imbibition due to increased testa permeability, MyA treatment lowered seed viability in darkness (this work) and prevented seedling establishment by acting as a phytotoxin (Popovici et al., 2011; Oracz et al., 2012).

MyA inhibits endosperm cap weakening and this inhibition is more severe in darkness compared with the light

Prior to radicle protrusion and endosperm rupture of A. thaliana and L. sativum, the embryo RAD elongates due to cell expansion growth (Sliwinska et al., 2009; this work) and the CAP weakens (Müller et al., 2006; Bethke et al., 2007; Linkies et al., 2009; Oracz et al., 2012; this work). An early embryo signal, which can be replaced by GA, initiates endosperm cap weakening which further proceeds in an organ-autonomous manner and is promoted by GA and ethylene, and inhibited by ABA in L. sativum (Müller et al., 2006; Linkies et al., 2009; Graeber et al., 2010; Morris et al., 2011). Molecular mechanisms for endosperm cap weakening depend on environmentally and hormonally regulated expression of cell wall remodelling proteins and/or ROS (e.g. Ni and Bradford, 1993; Bewley, 1997; Petruzzelli et al., 2003; Müller et al., 2009; Voegele et al., 2011). Datura ferox endosperm weakening is promoted by light and inhibited by darkness combined with low ambient water potential (Sanchez et al., 2002). The germination of L. sativum seeds is, however, not inhibited by darkness, but the inhibitory effect of MyA on endosperm rupture was much stronger in darkness compared with continuous light and this was associated with prolonged inhibition of endosperm cap weakening in darkness (Fig. 1b). Light alleviates the inhibitory effects of MyA on endosperm weakening and rupture by a mechanism which does not seem to play a role under CON germination conditions in which light or darkness do not influence germination of L. sativum. Altered ABA or GA contents were not found in the CAP or RAD that might explain the more severe MyA inhibition in darkness. Scarification (removal or pricking of the seed envelopes) experiments with A. thaliana mutant seeds highlight the permeability and constraint weakening of the testa and/or endosperm being major determinants for adaptation to environmental change (Barua et al., 2011). Based on the mechanistic model of two opposing forces during seed germination (Ni and Bradford, 1993; Linkies and Leubner-Metzger, 2012), environmental modulation of the endosperm weakening inhibition combined with increased testa permeability and decreased embryo growth seems to alter the degree of MyA inhibition of seed germination.

The MyA-mediated inhibition of embryo growth and molecular mechanisms for radicle–hypocotyl cell elongation growth are regulated developmentally and by environmental cues

Cell expansion growth required for radicle protrusion and endosperm rupture depends on environmentally and hormonally regulated cell wall-loosening mechanisms mediated by proteins such as xyloglucan endotransglycosylases/hydrolases (XTHs) and expansins (Cosgrove, 2005; Knox, 2008; Voegele et al., 2011) and/or by apoplastic ROS (Renew et al., 2005; Schopfer, 2006; Müller et al., 2009). Specific zones in the A. thaliana and L. sativum embryo hypocotyl–radicle axis exhibit endoreduplication-associated cell expansion growth, and in L. sativum are also associated with apoplastic ROS production as a molecular mechanism for this cell expansion growth (Müller et al., 2009; Sliwinska et al., 2009; Graeber et al., 2010; Oracz et al., 2012). MyA inhibits endoreduplication and apoplastic ROS production in the RAD of L. sativum seeds. In the present study, embryo sizes of L. sativum seeds were quantified and it was demonstrated that MyA inhibits embryo growth in terms of size and its rate during late germination (15–22h). While the embryo sizes at the onset of endosperm rupture were equal between the CON and the MyA series, the embryo size required for ER_50% was 1.4-fold larger for MyA (ER_50% at ~35h) compared with CON (ER_50% at ~22h). A larger embryo due to prolonged RAD growth to achieve this response later in time is in agreement with the MyA-mediated inhibition of endosperm cap weakening.

ABA inhibits embryo cell expansion growth and decreases the embryo growth potential of coffee, tomato, and B. napus (Schopfer and Plachy, 1985; Toorop et al., 2000; Da Silva et al., 2004, 2008). ABA also inhibits the production of apoplastic ROS in the L. sativum RAD, whereas GA and ethylene promote it (Müller et al., 2006; Graeber et al., 2010). Apoplastic ·OH is produced from apoplastic H₂O₂ and O₂·^–, and plasma membrane-localized NADPH oxidase has been proposed to produce the apoplastic O₂·^–; and this apoplastic ROS causes cell wall loosening required for cell expansion growth, as has been demonstrated for seed germination, and embryo and seedling growth (Renew et al., 2005; Schopfer, 2006; Müller et al., 2009). MyA was characterized as an effective C-methylated dihydrochalcone radical scavenger (Mathiesen et al., 1995, 1997) and it was demonstrated here that MyA reduced the apoplastic O₂·^– production in the L. sativum RAD from seeds imbibed in continuous light or darkness between 15h and 22h (Oracz et al., 2012; this work). The ~2-fold lower contents of apoplastic O₂·^– in the RAD in the dark compared with the light (Fig. 7) explains, at least in part, why the MyA-mediated inhibition of L. sativum seed germination is stronger in darkness. As it was shown that light itself does not chemically inactivate MyA, it is speculated that a light-induced accumulation of an unknown MyA detoxification enzyme may be an explanation for the difference between light and darkness; light-induced detoxification is known in other systems (e.g. Kucera et al., 2003; Barrero et al., 2012). In agreement with a requirement for apoplastic O₂·^– production for cell expansion growth, light is known to target NADPH oxidase to the plasma membrane and nucleus of wheat coleoptiles (Schopfer, 2006; Chandrakuntal et al., 2010). It is concluded that MyA, due to its radical-scavenging properties, interferes with the ROS-mediated cell expansion growth required for embryo elongation and seedling growth. Many other allelochemicals, such as sorgoleone and (–)-catechin, inhibit seedling establishment in the ecosystem by causing enhanced ROS production and cell death (Inderjit and Duke, 2003; Weir et al., 2004; Oracz et al., 2007). They therefore act phytotoxically by increasing ROS production, whereas MyA acts phytotoxically by decreasing ROS production in germinating seeds and seedlings.

MyA as a soil seed bank-destroying allelochemical: a speculative ecophysiological working model based on its multiple targets and mechanisms

Ambient temperature and light conditions together with the soil water status are the key abiotic environmental factors that determine seasonal germination pattern and modulate persistence and dormancy of soil seed banks (Walck et al., 2005; Batlla and Benech-Arnold, 2010; Footitt et al., 2011). Light intensity and spectral composition vary for seeds located at different depths in the soil bank and for germinated seeds and growing seedlings depending on shading by their plant neighbours. The soil water status affects the persistence and dormancy of soil seed banks as the fluctuations in the ambient water potential affect the seed hydration levels causing altered hormone and light sensitivities (e.g. Batlla and Benech-Arnold, 2006, 2010; Footitt et al., 2011). Flooding, for example, can cause seed and seedling death of unadapted species, thus opening gaps in the vegetation and facilitating the establishment of flooding-tolerant species (Benech-Arnold et al., 2000). It was found that the phytotoxic effects of the putative allelochemical MyA of M. gale are achieved by targeting several key steps of seed germination and seedling establishment, and that these effects are modulated in a complex manner by the ambient light conditions and water potentials.

Myrica gale is found in environments such as the wet and flooded areas around lakes, along rivers, or in peat bogs, and has a wide distribution in Northern and Western Europe and on the American continent (Skene et al., 2000). It is a shrub, scattered with many plants clustering close together at a growing site, and is described as ‘light loving’ and as a wet site indicator. The fruits of M. gale have fleshy bracts that aid the dispersal of seeds by flotation, keeping them afloat when they fall into the water of their often flooded habitat. Germination of M. gale requires cold stratification and extended exposure to light, which has been suggested to provide a mechanism to ensure that germination only occurs when seeds are located on relatively stable substrate following transport by water. Myrica gale and Myrica rubra dormancy can be released by GA, and for M. rubra it was demonstrated that the endogenous GA and ABA contents in the endocarp and seed coat are regulating coat dormancy and germination (Schwintzer and Ostrofsky, 1989; Chen et al., 2008). Myrica gale fruit exudates are known to contain MyA as the major C-methylated dihydrochalcone and putative allelochemical that inhibits seed germination and seedling growth (Svoboda et al., 1998; Popovici et al., 2011; Oracz et al., 2012). That MyA is indeed an allelochemical of M. gale in its habitat has so far not been proven by ecosystem competition experiments, but such experiments have been conducted for three other (invasive) Myrica species (Vitousek et al., 1987; Walker and Vitousek, 1991). Their leaf litter inhibits target species germination by the release of unknown allelochemicals, and for M. cerifera and M. pensylvanica this inhibition is pronounced under low light conditions (Collins and Quinn, 1982; Tolliver et al., 1995; Shumway, 2000). The complex interaction of shading and allelopathy is in agreement with the finding that darkness and low light conditions cause a stronger MyA-mediated inhibition of seed germination and seedling growth (Popovici et al., 2011; this work). It is speculated that MyA is a soil seed bank-destroying allelochemical that secures the persistence of M. gale in its flood-prone environment.

The speculative ecophysiological working model is based on the finding that MyA acts via multiple inhibitory mechanisms that have important key processes of seed germination and seedling establishment of competitor species as targets. (i) During early germination, MyA enhances testa permeability and thereby facilitates too rapid water uptake during embryo imbibition. This may cause imbibitional injury and untimely loss of embryo desiccation tolerance of competitor species, and may also cause enhanced uptake of phytotoxins like MyA itself, as well as leakage of metabolites including ABA. With their protection mechanisms deactivated by MyA, seeds of these competitor species may be easily prone to decay processes. (iii) During late germination, MyA inhibits endosperm weakening and interferes with embryo growth; and the finding that this is more pronounced in darkness is in agreement with the proposal that MyA is an allelochemical that acts under low light conditions. Seeds of competitor species of the soil bank buried next to M. gale plants or fruits would germinate, but, due to the enhanced phytotoxicity of MyA in darkness, would have reduced viability leading to slow growth and fast decay. (iii) During late germination and subsequent seedling growth, MyA acts as a scavenger of apoplastic ROS required for cell wall loosening and cell expansion growth. MyA thereby targets seed germination and seedling growth mechanisms that are important for a wide range of species.