Karrikins Discovered in Smoke Trigger Arabidopsis Seed Germination by a Mechanism Requiring Gibberellic Acid Synthesis and Light^,,

Abstract

Discovery of the primary seed germination stimulant in smoke, 3-methyl-2H-furo[2,3-c]pyran-2-one (KAR₁), has resulted in identification of a family of structurally related plant growth regulators, karrikins. KAR₁ acts as a key germination trigger for many species from fire-prone, Mediterranean climates, but a molecular mechanism for this response remains unknown. We demonstrate that Arabidopsis (Arabidopsis thaliana), an ephemeral of the temperate northern hemisphere that has never, to our knowledge, been reported to be responsive to fire or smoke, rapidly and sensitively perceives karrikins. Thus, these signaling molecules may have greater significance among angiosperms than previously realized. Karrikins can trigger germination of primary dormant Arabidopsis seeds far more effectively than known phytohormones or the structurally related strigolactone GR-24. Natural variation and depth of seed dormancy affect the degree of KAR₁ stimulation. Analysis of phytohormone mutant germination reveals suppression of KAR₁ responses by abscisic acid and a requirement for gibberellin (GA) synthesis. The reduced germination of sleepy1 mutants is partially recovered by KAR₁, which suggests that germination enhancement by karrikin is only partly DELLA dependent. While KAR₁ has little effect on sensitivity to exogenous GA, it enhances expression of the GA biosynthetic genes GA3ox1 and GA3ox2 during seed imbibition. Neither abscisic acid nor GA levels in seed are appreciably affected by KAR₁ treatment prior to radicle emergence, despite marked differences in germination outcome. KAR₁ stimulation of Arabidopsis germination is light-dependent and reversible by far-red exposure, although limited induction of GA3ox1 still occurs in the dark. The observed requirements for light and GA biosynthesis provide the first insights into the karrikin mode of action.

Germination is a critical event in the plant life cycle, as the timing of emergence from the protective seed coat is crucial for survival and reproductive success. A variety of abiotic stimuli, including light, temperature, and nitrates, provide information about the external environment that affects germination. Seed dormancy gates responses to these factors. Upon maturation, physiologically dormant seeds are in a primary dormant (PD) state, which is lost during afterripening. The transition between a PD and nondormant state is both gradual and reversible and results in relaxation of the set of environmental conditions under which a seed will germinate (Baskin and Baskin, 2004; Finch-Savage and Leubner-Metzger, 2006).

Despite decades of research, seed dormancy remains a complex physiological state that is not well understood. The plant hormones abscisic acid (ABA) and GA are mutually antagonistic central players in the germination decision (Finch-Savage and Leubner-Metzger, 2006; Finkelstein et al., 2008). The role of dormancy establishment and maintenance has been attributed to ABA, while GA has been implicated in the initiation and completion of germination. The ratio of ABA to GA signaling, rather than absolute amounts of the hormones, appears to be critical to dormancy breaking (Finch-Savage and Leubner-Metzger, 2006). Environmental stimuli and phytohormones influence the ABA/GA balance, although the mechanisms of signal integration and hormone cross talk are still largely unknown.

In many biodiverse regions, fire events provide an irregular but important opportunity for seedling establishment by freeing up key resources such as light, space, and nutrients (Van Staden et al., 2000; Dixon et al., 2009). A clear example of this is seen in the flush of new growth in the immediate postfire environment, indicating potent activation of the soil seed bank. Heat is not required for the germination response, as cold smoke application induced an up to 48-fold increase in the number of germinating seedlings and approximately 3-fold enrichment in species abundance in field trials (Roche et al., 1997; Rokich et al., 2002). It has now been well established that smoke is a broadly effective stimulant that enhances germination of approximately 1,200 species in more than 80 genera worldwide (Dixon et al., 2009). Attempts to study smoke effects on plant physiology have been confounded by the complex mixture of components within smoke, some of which confer toxicity at high concentrations. Bioassay-guided fractionation of smoke water culminated in the discovery and synthesis of the primary germination stimulant 3-methyl-2H-furo[2,3-c]pyran-2-one (KAR₁; Flematti et al., 2004). With the recent identification of three analogous active compounds in smoke water fractions (Fig. 1A; G. Flematti, unpublished data), this family of butenolide molecules have been designated karrikins, after “karrik,” the first recorded Aboriginal Nyungar word for smoke (Dixon et al., 2009).

Figure 1.

Partial structural similarity between the karrikin family of plant growth regulators and strigolactones. A, Molecular structures of KAR₁ (3-methyl-2H-furo[2,3-c]pyran-2-one), KAR₂ (2H-furo[2,3-c]pyran-2-one), KAR₃ (3,5-dimethyl-2H-furo[2,3-c]pyran-2-one), and KAR₄ (3,7-dimethyl-2H-furo[2,3-c]pyran-2-one). B, Molecular structures of a natural (strigol) and a synthetic (GR-24) strigolactone.

The parent molecule, KAR_1, is a potent stimulant that enhances germination in some species at subnanomolar concentrations (Flematti et al., 2004; Stevens et al., 2007). In field trials, KAR₁ is effective at less than 5 g ha⁻¹ compared with 10 ton ha⁻¹ smoke water and thus may have practical value in agriculture, conservation, and restoration (Stevens et al., 2007). Smoke water fractions containing KAR₁ have been reported to enhance seedling vigor of several weed and crop species, indicating potential use for KAR₁ as a seed priming agent to improve germination and seedling establishment (Jain et al., 2006; Jain and Van Staden, 2006; Kulkarni et al., 2006; van Staden et al., 2006; Daws et al., 2007a). Since its discovery, a widespread capacity for KAR₁ germination response among angiosperms has been demonstrated (Flematti et al., 2004; van Staden et al., 2004, 2006; Merritt et al., 2006; Daws et al., 2007a; Stevens et al., 2007). Thus, karrikins may be considered a novel class of plant growth regulators with broad impact. To gain a better understanding of the mechanism by which karrikins trigger seed germination and explore their interaction with ABA and GA, we examined KAR₁ responses in Arabidopsis (Arabidopsis thaliana).

RESULTS

Karrikins Enhance Germination of Primary Dormant Arabidopsis Seeds

To determine if Arabidopsis is a suitable model system for studies of karrikin action, PD seeds of the Landsberg erecta ecotype (Ler) were tested for germination enhancement by KAR₁. While 1 μm KAR₁ strongly promoted germination, the phytohormones GA and epibrassinolide (EBR) and the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC) had little or no effect on PD seeds at similar concentrations (Fig. 2, A and B). The KAR₁ effect on germination rates was comparable to that of 10 mm KNO₃, an effective dormancy-breaking treatment (Alboresi et al., 2005). The combination of these two treatments resulted in a synergistic response.

Figure 2.

Karrikins enhance germination of Arabidopsis. A, PD Ler seeds after 5 d of imbibition on water (left) or 1 μm KAR₁ (right). B, Germination of PD Ler seeds imbibed on 10 mm KNO₃, 1 μm KAR₁, GA₄, or EBR, or 10 μm ACC. C, Germination of PD Ler seeds after 7 d of imbibition on 1 nm to 10 μm KAR₁, KAR₂, KAR₃, or KAR₄, GR-24, or GA₄. D, Germination of B. tournefortii and the parasitic weed O. minor in the presence of GR-24 or KAR₁. E, Germination of PD Arabidopsis ecotypes after 7 d of imbibition (except Cvi-0, which is shown at 14 d). Time courses of germination are shown in Supplemental Figure S1. F, Germination of progressively AR PD Ler seeds after 4 d of imbibition. G, Germination of PD Ler stratified (ST) for 3 d in the dark at 4°C or 10-month AR Ler seeds (AR) ± 1 μm KAR₁. The x axis indicates the time after transfer to continuous light at 20°C or the time of imbibition.

Karrikins have no distinct structural similarity to known plant hormones, although the A ring of the KAR₁ molecule is analogous to the D ring of strigolactones (Fig. 1B; Flematti et al., 2004). Strigolactones are exuded by the roots of host plants and are highly active germination stimulants for a number of parasitic weed species in the Striga, Orobanche, and Alectra genera (Humphrey et al., 2006). Smoke water extracts, in which KAR₁ is present, have also been reported to stimulate parasitic weed germination (Bar Nun and Mayer, 2005; Daws et al., 2007b). To test whether karrikins and strigolactones share a similar mechanism of action, we compared the germination response of PD Ler seeds to the four karrikins and the synthetic strigolactone GR-24 (Fig. 2C). KAR₂ was the most active analog and clearly enhanced germination at concentrations as low as 10 nm, while KAR₁ and KAR₃ were slightly less effective. Although KAR₄ stimulates germination of Lactuca sativa and Solanum orbiculatum (Flematti et al., 2007), KAR₄ was either completely inactive or inhibitory at similar concentrations in Arabidopsis. GR-24 was capable of stimulating PD Ler germination but required approximately 100-fold greater concentrations than the weakest active karrikin, KAR₃, to achieve the same effect. Brassica tournefortii, a strongly karrikin-responsive species, also had minimal response to GR-24 (Fig. 2D). Conversely, KAR₁ was unable to trigger germination of Orobanche minor, while GR-24 was active at concentrations as low as 1 nm (Fig. 2D). The differences in karrikin and strigolactone efficacy among these species suggest distinct modes of action or a species-specific capacity for response.

KAR₁ Effects on Germination Vary with Ecotype and Dormancy Depth

There is substantial natural variation in primary seed dormancy among Arabidopsis ecotypes (van Der Schaar et al., 1997; Alonso-Blanco et al., 2003). PD seeds from seven other ecotypes were tested for germination enhancement by KAR₁. A range of responses were observed in which KAR₁ had moderate to negligible effects on germination (Fig. 2E). KAR₁ typically caused an early, limited enhancement of seed germination rates, suggesting that a subset of the seed population had been pushed over a dormancy threshold within an imbibition time window (Supplemental Fig. S1). Notably, the control levels of germination and responses to KNO₃ were also variable across ecotypes. These results may be attributed to natural variation either in the capacity for KAR₁ perception and response or in seed dormancy depth. To address the latter hypothesis, we examined the germination of Ler seeds with different depths of dormancy. Although after 7 d of imbibition the germination enhancement by KAR₁ was apparent and equivalent to 6 weeks of afterripening (Supplemental Fig. S2), KAR₁ was unable to fully stimulate germination of PD seeds within a 4-d period (Fig. 2F). With afterripening, KAR₁ enhanced germination within this time frame at a rate outpacing the rise in control germinability. However, as dormancy was further lost, there was a corresponding reduction in the apparent effects of KAR₁ on germination relative to the control. Thus, in either highly dormant or nondormant states seeds may not show an obvious response to KAR₁ treatment, particularly when assaying germination at a single time point. We found that KAR₁ had a positive effect on seed germination rates even after removal of PD Ler seed dormancy by cold stratification or extended afterripening (Fig. 2G). Karrikins, therefore, display characteristics of dormancy-breaking and germination-stimulating compounds.

GA Biosynthesis Is Required for KAR₁ Promotion of Germination

To investigate the interaction of karrikins with ABA and GA pathways, we tested the germination responses of several phytohormone mutants (Table I). We first examined mutants with reduced ABA biosynthetic capacity. aba3-2 had a small response to KAR₁ after 48 h but otherwise exhibited no difference in germination relative to the control (Fig. 3A). We reasoned that if karrikins act through reducing ABA levels or sensitivity, the lack of ABA in this mutant may prevent the detection of a KAR₁ effect. In the presence of exogenous ABA, germination rates were inhibited, although aba3-2 seeds still showed no response to KAR₁. Similar results were obtained with aba2-1 and aba3-1 alleles, and afterripened (AR) Ler and aba1-3 germination was not further enhanced by KAR₁ in the presence of ABA (Supplemental Fig. S3, A–D). These results could indicate that KAR₁ acts primarily by reducing ABA synthesis during imbibition or that high levels of ABA block the karrikin response. Alternatively, as it has been established that exogenous ABA application does not induce the same transcriptional state in seeds as natural dormancy (Carrera et al., 2008), it may be that ABA is required during seed maturation to form a seed with KAR₁-responsive dormancy.

Table I.

Summary of Arabidopsis phytohormone mutant responses to KAR₁

The ecotype for each mutant is indicated. Special conditions (Treatment) are noted (i.e. in the presence [ABA/GA] or absence [−] of exogenous hormone or dormancy state [PD/AR] of the seed). Where not noted, all seeds were of the primary dormancy level within a few days of harvest. KAR₁ response in germination is classified as − (0%–10%), + (10%–40%), or ++ (>40%). Col, Ecotype Columbia.

Allele	Ecotype	Treatment	KAR1 Response	Figure
ABA synthesis
aba1-3	Ler	−	+	S3B
		ABA	+	S3B
aba2-1	Col	−	−	S3C
		ABA	−	S3C
aba3-1	Col	−	−	S3D
		ABA	−	S3D
aba3-2	Ler	−	?	3A
		ABA	−	3A
ABA catabolism
cyp707a1 a2	Col	PD	−	S3E
		AR	+	S3F
cyp707a1 a3	Col	PD	−	S3E
		AR	+	S3F
cyp707a2 a3	Col	PD	+	S3E
		AR	+	S3F
ABA response
era1-2	Col		−	S3G
GA synthesis
ga1-3	Ler	−	−	3B
		GA	−	3C
ga4-1	Ler		++	S4A
ga3ox1	Col		−	S4C
ga3ox2	Col		++	S4D
ga3ox1 ox2	Col		−	S4E
ga5-1	Ler		++	S4B
GA response
sly1-2	Ler		+	3D
sly1-10	Ler		+	3D

Figure 3.

Responses of phytohormone mutants to KAR₁. A, Germination of aba3-2 on ABA with (black symbols) or without (white symbols) 1 μm KAR₁. B, Germination of ga1-3 after a 3-d dark stratification at 4°C followed by 10 d in continuous light at 20°C. The medium was 0.5× Murashige and Skoog salts (MS; contains nitrates) with 10 μm GA₄, 1 μm KAR₁, 10 μm ACC, or 1 μm EBR. C, Germination of ga1-3 in the presence of 1 or 10 μm GA₃ ± 1 μm KAR₁. D, Germination of sly1-10 and sly1-2 alleles.

ABA catabolism occurs primarily through an oxidation pathway resulting in phaseic acid via an 8′-hydroxy ABA intermediate (Nambara and Marion-Poll, 2005). The CYP707A family of cytochrome P450s have been characterized as ABA 8′-hydroxylases. Mutations in these genes lead to increased ABA levels in mature seeds and a reduction in germinability (Okamoto et al., 2006). PD cyp707a2 cyp707a3 seeds were responsive to KAR₁, while the highly reduced germination of cyp707a1 cyp707a3 and cyp707a1 cyp707a2 showed no enhancement (Supplemental Fig. S3E). With afterripening, however, germination of all three double mutant lines was mildly enhanced by KAR₁ (Supplemental Fig. S3F). KAR₁ was also unable to overcome the enhanced dormancy of the ABA-hypersensitive era1-2 mutant (Cutler et al., 1996), although KNO₃ could eventually induce germination (Supplemental Fig. S3G). Thus, high endogenous levels of ABA signaling can block induction of germination by KAR₁.

GA is typically required for germination and can overcome dormancy or promote germination in restrictive conditions. Alleles of GA3ox1 (ga4-1) and GA20ox1 (ga5-1) have reduced GA biosynthetic capacity but were both strongly responsive to KAR₁ (Supplemental Fig. S4, A and B). Germination is blocked in the GA-deficient mutant ga1-3, but this can be overcome by other hormones, such as brassinosteroid and ethylene, or by reduced ABA synthesis during seed maturation (Karssen et al., 1989; Debeaujon and Koornneef, 2000; Steber and McCourt, 2001). In our hands, very limited ga1-3 germination was achieved with either ACC or EBR, even with the inclusion of nitrates in the medium and a stratification treatment (Fig. 3B). However the ga1-3 mutant had no germination in the presence of KAR₁. Similarly, while a ga3ox2 allele was KAR₁ responsive, the highly reduced germination of a ga3ox1 ga3ox2 double mutant was not recovered by KAR₁ (Supplemental Fig. S4, D and E). If KAR₁ acts primarily through enhancement of GA biosynthesis and not sensitivity to GA, a lack of germination promotion in GA-supplemented ga1-3 seeds would be expected. Indeed, KAR₁ produced little to no change in the germination response of ga1-3 to different concentrations of GA₃ (Fig. 3C).

GA signaling is mediated by SLEEPY1 (SLY1), an F-box protein that targets the DELLA repressor proteins for degradation in a GA-dependent manner (Dill et al., 2004; Finkelstein et al., 2008). Interestingly, KAR₁ treatment partially restored germination of the GA signaling-deficient sly1-2 and sly1-10 alleles, while KNO₃ was less effective (Fig. 3D). This suggests that nitrate and karrikin have distinct modes of action and that KAR₁ may partially influence germination in a DELLA-independent manner.

Transcription of GA 3-Oxidases and CP1 Is Induced by Active Karrikins

To gain further insight into KAR₁ effects on hormone metabolism during PD Ler seed imbibition, we performed quantitative reverse transcription (qRT)-PCR analysis for a set of genes involved in the synthesis, catabolism, or response of ABA and GA. Transcriptional changes were assessed during the first 48 h of imbibition, before KAR₁-treated seeds begin to germinate (Fig. 2B), in order to identify early regulatory events that may result in a germination decision.

No transcript differences were observed for the major ABA biosynthetic enzymes ABA1, ABA2, ABA3, NCED6, or AAO3 in response to KAR₁ (Supplemental Fig. S5A). AAO4, which can have a role in ABA biosynthesis in the absence of the major isozyme AAO3 (Seo et al., 2004), was up-regulated after 24 h by KAR₁. In a comparison of dormant and nondormant Arabidopsis seeds, expression of the ABA 8′-hydroxylase CYP707A2 was 4-fold up-regulated in nondormant seeds at 6 h of imbibition (Millar et al., 2006). However, CYP707A2 and CYP707A3 showed no change in expression in response to KAR₁ treatment during imbibition (Supplemental Fig. S5B). ABA can also be inactivated through the formation of an ABA-Glc ester conjugate by the glucosyltransferase UGT71B6 (Priest et al., 2005). UGT71B6 transcript abundance was unaffected by KAR₁ (Supplemental Fig. S5B).

ABI genes were identified as components of ABA signal transduction pathways through genetic screens for ABA-insensitive mutants (Finkelstein et al., 2002). Of the five ABI genes tested, only ABI4 transcripts were affected by KAR₁ (Supplemental Fig. S5C). At 48 h, ABI4 expression was up-regulated by KAR₁. ABI4 encodes an AP2 domain transcription factor (Finkelstein et al., 1998) and confers sensitivity to sugar and salt stress (Arenas-Huertero et al., 2000; Huijser et al., 2000; Laby et al., 2000; Quesada et al., 2000). The implications of ABI4 induction by KAR₁ are currently unclear, but they may reflect preparation for seedling emergence. FRY1/SAL1 encodes an inositol 1,4,5-triphosphate phosphatase and acts as a negative regulator of ABA signaling (Xiong et al., 2001). We observed up-regulation of FRY1 in KAR₁-treated seeds after 48 h of imbibition (Supplemental Fig. S5C).

GAs are derived from geranylgeranyl diphosphate through the action of a series of terpene cyclases, cytochrome P450s, and 2OG-dependent dioxygenases (Olszewski et al., 2002). No transcript changes were observed for GA2/KS1, GA3/KO1, KAO1, or KAO2 (Supplemental Fig. S6A). Among the GA 20-oxidases, which catalyze the penultimate step in GA biosynthesis, expression of GA20ox2, GA20ox3, and GA20ox4 was relatively unaffected by KAR₁, but GA20ox1 was initially suppressed by KAR1 and then induced at 48 h (Supplemental Fig. S6A). GA 3-oxidases catalyze the final step of GA biosynthesis to produce active GA₄ and GA₁. GA3ox1/GA4 and GA3ox2 are the primary genes encoding GA 3-oxidases (Yamaguchi et al., 2001). GA3ox1 and GA3ox2 are light responsive, and GA3ox1 is also regulated by stratification and GA feedback inhibition (Yamaguchi et al., 1998; Yamauchi et al., 2004). KAR₁ induced significant up-regulation of both GA3ox1 and GA3ox2. GA3ox1 transcripts showed 4-fold up-regulation by KAR₁ as early as 6 h of imbibition, while GA3ox2 was differentially expressed after 12 h (Fig. 4A). The GA 2-oxidases are responsible for GA catabolism (Thomas et al., 1999). No KAR₁-induced expression changes were observed for GA2ox2 or GA2ox3 (Supplemental Fig. S6B).

Figure 4.

Karrikins induce GA 3-oxidase and CP1 transcripts. A, Relative expression of GA3ox1 and GA3ox2 during a 48-h time course of PD Ler seed imbibition on water (white symbols) or 1 μm KAR₁ (black symbols) in the light at 20°C. Inset bar graph indicates fold change in expression (KAR₁/water) at the indicated time points. B, Quantitation of ABA and GA₄ in dry, 24-h, and 48-h KAR₁-imbibed (±1 μm) seeds. The y axis indicates ng g⁻¹ preimbibition seed weight. C, Expression of GA3ox1, GA3ox2, and CP1 in PD Ler seeds after 24 h of imbibition in the light at 20°C on water, 1 μm KAR₁, KAR₂, KAR₃, or KAR₄, 10 μm GR-24, 10 μm GA₄, or 10 mm KNO₃. Relative expression values for each gene in seeds imbibed in water were set to 1, and other expression values were scaled accordingly.

Among the five DELLA proteins, RGL2 is considered the main repressor in seed germination (Lee et al., 2002; Tyler et al., 2004; Cao et al., 2005). We detected no KAR₁-induced changes in transcript abundance during imbibition (Supplemental Fig. S6C). The putative Cys proteinase CP1 is a useful marker for GA signaling, as it is up-regulated by GA₄ in a concentration-dependent manner and also induced by GA-sensitizing treatments such as stratification (Yamauchi et al., 2004). CP1 exhibited a very similar pattern of expression as GA3ox2 and was up-regulated by KAR₁ after 24 h (Supplemental Fig. S6C). This suggested that a rise in GA levels consistent with GA3ox expression was occurring in KAR₁-treated seeds.

The relatively few changes in gene expression we observed implied that KAR₁ enhances GA synthesis via GA3ox1 and GA3ox2 but does not directly affect ABA pathways. To assess these predictions, we quantitated changes in the levels of ABA and GAs during imbibition of PD seeds in response to KAR₁ (Fig. 4B). ABA levels declined substantially during the first 48 h of imbibition but were unaffected by KAR₁. Interestingly, we did not detect a dramatic rise in GA₄ levels predicted by GA3ox and CP1 expression levels at 48 h. At 24 h, GA₄ levels were the same in control and KAR₁-treated seeds, and at 48 h, there was a statistically significant (Student's paired t test, P < 0.02), although small (10%), increase in GA₄ induced by KAR₁.

We examined the transcript levels of GA3ox1, GA3ox2, and CP1 in PD seeds imbibed for 24 h to compare the expression changes observed with KAR₁ and other germination stimulants. KAR₁, KAR₂, and KAR₃, but not KAR₄, induced expression of GA3ox1, GA3ox2, and CP1. The degree of up-regulation of these genes corresponded to the effectiveness of each treatment on stimulating germination (Fig. 4C). GR-24 produced a slight enhancement of GA3ox expression but did not result in CP1 up-regulation, at least at this time point. Exogenous GA₄ treatments produced the expected transcriptional effects: GA3ox1, which is feedback inhibited, was down-regulated, GA3ox2 was unaffected, and GA-responsive CP1 was strongly induced. KNO₃ was particularly effective at inducing GA3ox and CP1 transcripts, which may explain its broad effectiveness in Arabidopsis as a dormancy-breaking treatment.

KAR₁ Requires Light to Enhance Arabidopsis Germination

Moisture and light are minimal requirements for normal Arabidopsis seed germination. We found that karrikin could not replace the light requirement (Fig. 5A). Under continuous light, KAR₁ induced nearly complete germination of PD Ler seeds within 7 d (Fig. 2A) but caused minimal germination (2%) of PD seeds incubated in the dark after a 2-h initial white light exposure. AR seeds were much more responsive than PD seeds to the early light exposure but did not achieve maximal germination without KAR₁. When the early light treatment was reversed by a far-red (FR) light pulse prior to dark incubation, no germination was observed for KAR₁-treated seeds regardless of seed dormancy state. Light induces GA synthesis in seeds, and exogenous GA is sufficient to induce germination of Arabidopsis seeds in the dark. To determine whether KAR₁ action requires light because GA synthesis is not triggered, we tested the germination of GA-treated seeds in the presence and absence of KAR₁ during dark incubation. GA supplements were not sufficient to restore KAR₁ promotive effects, and KAR₁ had little or no effect on seed sensitivity to exogenous GA (Fig. 5A). It is also interesting that PD and AR seeds had nearly equivalent responses to GA after FR reversal of the early light exposure. Thus, light, but not dormancy loss through afterripening, enhances seed sensitivity to GA.

Figure 5.

KAR₁ requires light to induce germination and does not enhance GA perception. A, PD or 8-month AR Ler seeds were imbibed for 2 h under white light (W) on medium supplemented with 1 μm KAR₁, 1 μm GA₄, or 10 μm GA₄ and then placed into continuous darkness. A subset of seeds (right) were treated with 5 min of FR light (W:FR) immediately prior to transfer to dark conditions. Germination was assessed after 7 d. ^#, No germination. B, Expression of GA3ox1 transcripts in PD and AR seeds after 22 h of dark imbibition following the same light treatments as described for A. Level of expression in AR white light KAR₁-treated seeds was set to 100, and other values were scaled accordingly. ^#, Undetectable transcript.

To further examine the relationship of light to KAR₁ effects on germination, we tested the expression of the KAR₁-responsive genes GA3ox1 and CP1 under the same conditions. Interestingly, KAR₁ stimulated GA3ox1 expression in AR seeds in the dark even after FR exposure (Fig. 5B). However, the levels of GA3ox1 expression were dramatically enhanced by the combination of light and KAR₁ in AR seeds. CP1 expression showed a similar trend (data not shown). Notably, the strong induction of GA3ox1 occurred only in the treatment producing maximum dark germination. Thus, while KAR₁ can up-regulate GA3ox1 expression independently of light, its effects are insufficient to induce germination in the absence of light.

DISCUSSION

Karrikins are a novel family of plant growth regulators that affect key processes for a broad range of angiosperms. However, very little is known concerning the potential mechanisms of karrikin action in germination or seedling establishment. We provide, to our knowledge, the first demonstration that three karrikins, KAR₁, KAR₂, and KAR₃, promote germination of Arabidopsis seeds. At high concentrations, the structurally related synthetic strigolactone GR-24 enhanced germination of two karrikin-responsive species. However, KAR₁ was completely inactive on the smoke water and the strigolactone-responsive parasitic weed O. minor, suggesting that another component of smoke water is responsible for triggering Orobanche germination. These results indicate that karrikins and strigolactones are not interchangeable and may act via different mechanisms. The level of GA3ox induction corresponded with the efficacy of these germination stimulants (KAR₂ > KAR₁ > KAR₃ > GR-24). However, it cannot be assessed from GA3ox and CP1 expression alone whether karrikin and strigolactone signals are perceived and transduced via common pathways, as both stimulants may ultimately enhance GA synthesis as part of the germination process.

KAR₁ was not equally effective across all tested Arabidopsis ecotypes. Different depths of primary seed dormancy in these lines may provide at least a partial explanation for this observation. A progressive receptiveness to germination-promoting factors (nitrates, stratification) has been observed during afterripening of Arabidopsis Cvi seeds, indicating dynamic capacities for seed response to stimuli (Finch-Savage et al., 2007). Similarly, KAR₁ may be an ineffective stimulant under certain dormancy states. As an example, B. tournefortii germination is particularly sensitive to KAR_1, but some seed collections require several months of afterripening to become responsive (Stevens et al., 2007). As seeds in the soil seed bank cycle between dormancy states in response to afterripening or environmental stimuli, limited windows of opportunity for karrikin action may be created. In support of this, smoke treatments to enhance seed bank germination have marked differences in effectiveness at specific times of the year (Rokich and Dixon, 2007). Natural variation provides another likely explanation for variable KAR₁ responses. Here, we do not simply refer to variation in the depth of established seed dormancy but in the mechanisms by which dormancy is broken. Light, cold, and nitrate commonly promote germination across Arabidopsis ecotypes, but the contribution of each signal toward germination commitment can vary. For instance, polymorphisms in the photoreceptor phytochrome B have been identified as a source of natural variation in light responses (Filiault et al., 2008). In consideration of the range of KAR₁ effectiveness even within a single species, it would not be surprising to discover that karrikins activate germination under specific conditions for many more taxa than are currently known to be responsive.

Uncovering cross talk between karrikins and phytohormones is important for determining a mechanism for karrikin promotion of germination. KAR₁ was ineffective at overcoming inhibition of germination by exogenous ABA or high endogenous ABA signaling. While high concentrations of GA₃ are often effective at stimulating germination of KAR₁-responsive species, there has been no direct evidence for GA-mediated KAR₁ signaling (Merritt et al., 2006; Daws et al., 2007a; Stevens et al., 2007). We demonstrate that KAR₁ requires GA synthesis to induce germination, as both ga1-3 and ga3ox1 ga3ox2 mutants were unresponsive. KAR₁ did not affect seed sensitivity to GA, either in ga1-3 or in dark-incubated wild-type plants.

In support of the germination phenotypes, KAR₁ did not affect transcript abundance of the majority of genes involved in ABA and GA biosynthesis and catabolism but did induce GA3ox1 and GA3ox2. It is difficult to assess whether GA3ox induction by KAR₁ is a cause or a result of the seed's commitment to germination. As GA3ox1 expression is influenced by cold, light, and GA levels, it may serve as a signal integration point with a direct effect on germination. While KAR₁ enhancement of germination requires both light and GA biosynthetic capacity, it cannot be concluded that KAR₁ acts solely or directly through enhancement of light-induced GA3ox transcription.

The abundance of ABA during imbibition was unaffected by KAR₁, while GA₄ levels were only slightly increased. As dormancy loss has been reported to result in a faster decline of ABA levels during seed imbibition (Ali-Rachedi et al., 2004; Millar et al., 2006) and afterripening leads to enhanced GA sensitivity (Karssen et al., 1989), KAR₁ does not overcome seed dormancy in a similar manner. Although we had anticipated changes in hormone levels prior to initiation of germination, these results correspond with a similar experiment in which strong induction of GA3ox transcripts preceded germination but a rise in GA₄ levels was concomitant only with radicle emergence (Ogawa et al., 2003). Therefore, a substantial rise in total seed GA₄ levels may be a result of a seed's commitment to germination, rather than a cause. As there is evidence for spatial separation of transcripts for early and late phases of the GA biosynthetic pathway in embryonic tissues (Yamaguchi et al., 2001), it would be interesting if production of active GAs is restricted to a small “decision” sector of the seed prior to the initiation of germination and thus initially contributes little to the overall seed GA₄ abundance. Alternatively, the initial suppression of GA20ox1 by KAR₁ may restrict the GA precursor supply during the first 24 h of imbibition.

The broad conservation of karrikin perception, even among taxa not obviously subject to selective pressures from a “fire-prone” environment, suggests several intriguing hypotheses. First, karrikins may be generated via other mechanisms than fire. A biotic source such as bacteria or fungi, or the slow chemical breakdown of organic matter at the soil surface, could provide alternative sources of karrikin. Second, there may be a strong selective advantage for species that have maintained a karrikin signaling system even in ecosytems with rare fire events. Third, karrikins may be endogenous plant hormones that await identification. Our demonstration of a karrikin response in Arabidopsis seed germination opens the door for a genetic approach to addressing these possibilities.

MATERIALS AND METHODS

Plant Growth and Germination Assays

Arabidopsis (Arabidopsis thaliana) plants were grown in soil under continuous white light, 22°C, 60% relative humidity conditions. Harvested plants were dried for 4 to 7 d under ambient conditions in paper bags. Seeds pooled from multiple parent plants (>10) were then cleaned and stored at −80°C to preserve primary dormancy. Cryostorage had no obvious effect on seed viability or germination. AR seeds were maintained in the dark at room temperature. For germination assays, seeds were surface sterilized for 5 min with inversion in 70% ethanol and 0.05% Triton X-100, rinsed with 70% ethanol, rinsed again with 95% ethanol, and rapidly dried on filter papers in a sterile laminar flow cabinet. Sterilization treatment had no effect on germination. Sterilized seeds were sprinkled onto 0.8% Bacto-agar plates supplemented as indicated. Experiments comparing the four karrikins and GR-24 used methanolic 1,000× stocks, except for brassinolide, which was dissolved in dimethyl sulfoxide. In all other experiments, an aqueous 1,000× KAR₁ stock was used. Hormone stocks were prepared for (+)-ABA (A.G. Scientific; ethanol), GA₃ (Sigma; ethanol), GA₄ (from L.N. Mander, Australian National University; ethanol or methanol), epibrassinolide (Sigma; methanol), and ACC (Sigma; water). Karrikins were synthesized as described previously (Flematti et al., 2007). Immediately after plating, plates were sealed with gas-permeable Leukopor tape and transferred to continuous light (approximately 100 μE), 20°C conditions, and 25 seed sectors were outlined prior to germination. Germination, defined as emergence of the radicle, was assessed for three replicates each of 100 or 150 seeds (300 or 450 seeds total per treatment per time point) from one large seed batch per genotype. Average germination and se across the three replicates are shown. Similar trends from multiple independent seed batches were consistently observed for Ler and ga1-3. FR light-emitting diodes (L735-04AU; Epitex) were used to supply FR light treatment where indicated. FR light-emitting diode light intensity (6 μE) was determined with a Warsash Scientific EPP2000 spectrometer with irradiance calibration. The ga1-3 mutant required two to three spray treatments of 10 μm GA₃ to recover fertility, but all other mutants were grown without additional treatments.

Orobanche minor seeds (collected from King's Park, Perth, Western Australia) were surface sterilized and preconditioned on filter paper dampened with 0.5 mL of water for 2 weeks at 20°C in the dark. A second filter paper was then added to cover the seeds and dampened with 0.5 mL of 2× aqueous germination stimulant. Germination was scored for three replicates each of 75 seeds after 7 d of dark imbibition at 20°C.

Brassica tournefortii (collected from Meckering, Western Australia) was sown on 0.8% agar supplemented with 1,000× methanolic stocks of KAR₁ or GR-24. Germination for three replicates each of 50 seeds was assayed after 7 d of imbibition in the dark at 20°C.

qRT-PCR Analysis

Imbibed seed samples (50 or 80 mg preimbibition weight) were frozen in liquid nitrogen and stored at −80°C until processing. RNA was isolated using the RNAqueous Kit with Plant RNA Isolation Aid (Ambion) and LiCl precipitation to aid in the removal of contaminating polysaccharides. RNA integrity was assessed by BioAnalyzer (Agilent). RNA was treated with Turbo DNA-free (Ambion) and subsequently converted to cDNA using the iScript cDNA Synthesis Kit (Bio-Rad). qRT-PCR was performed on a Roche LC480 using LightCycler 480 SYBR Green I Master (Roche). Cycle conditions were 95°C for 10 min; 45 cycles of 95°C for 20 s, 60°C for 20 s, and 72°C for 20s; followed by melt curve analysis. When possible, primer pairs were designed across introns. Most primer pairs were designed by AtRTPrimer (Han and Kim, 2006), except NCED6 (Millar et al., 2006). Primer sequences are listed in Supplemental Table S1. Melt curve analysis of real-time PCR products was always performed to ensure clean amplification. Crossing point values were calculated under high confidence. Two independently prepared biological replicates per treatment/time point were examined in all studies, with at least two technical replicates of each real-time PCR. The average crossing point value of two technical replicates was used to calculate expression relative to an internal reference gene adjusted by primer efficiencies. We tested several potential reference genes identified by Czechowski et al. (2005), and for our analyses we chose At5g46630, a clathrin adaptor complex subunit that we term CACS. The average relative expression and se for the two biological replicates are shown. Levels of relative expression in dry seeds or the water control were typically adjusted to 1, and all other values were scaled accordingly. For the PD Ler imbibition time course, two of the three independent seed batches used for ABA/GA quantitation were used. All three seed batches had germination rates similar to that shown in Figure 2B, and no radicle emergence occurred by 48 h of imbibition.

Quantitation of ABA and GA

Per time point, three replicates of 300 mg of seeds (preimbibition weight) from independent PD Ler seed batches were pulverized in liquid nitrogen using a ball mill. Powder was extracted with 5 mL of 80% (v/v) methanol-water (with 0.1% acetic acid [AA]) with the following deuterated internal standards added: 10 ng g⁻¹ [17,17-²H₂]GA₁ and [17,17-²H₂]GA₄ (from L.N. Mander) and 20 ng g⁻¹ [²H₆]ABA (from L.I. Zaharia, National Research Council, Canada). The suspension was stored at 4°C for 24 h and filtered. The residue was rinsed with a further 50 mL of 80% methanol/0.1% AA water and combined with the original extract. The extract was reduced to an aqueous solution under reduced pressure at 35°C to which brine (10 mL) and 0.1% AA water (20 mL) were added. The aqueous solution was partitioned three times with equal volumes of ethyl acetate, and the combined organic phase was evaporated to dryness under reduced pressure at 35°C. The ethyl acetate extract was dissolved in 15% methanol/0.1% AA water (10 mL) and was applied directly to a preconditioned C₁₈ Sep-Pak (Waters; 1 g) cartridge. The retained material was rinsed with 15% methanol/0.1% AA water (10 mL), and the GAs and ABA were eluted with 80% methanol/0.1% AA water (10 mL). The 80% methanol fraction was evaporated to dryness under reduced pressure at 35°C, and the sample was dissolved in methanol (2 mL) and methylated with excess ethereal diazomethane. The sample was dried and dissolved in 20% methanol/0.1% AA water (1 mL) and separated by HPLC (Hewlett-Packard 1050 apparatus). The sample was injected (1 mL) onto a C₁₈ reversed-phase column (Grace-Davison Apollo; 250 × 10 mm, 5 μm) and eluted with 20% acetonitrile/0.1% AA water, which increased to 100% acetonitrile (+0.1% AA) over 40 min. Fractions were collected every minute for 40 min. After all of the extracts had been separated, retention times of the GAs and ABA were established by running authentic samples and monitoring for UV A₂₀₆. GA₁ and GA₄ eluted at 11.24 and 25.35 min, respectively, and ABA eluted at 20.06 min. Individual fractions containing these hormones plus one fraction on either side were combined and dried under reduced pressure. The ABA fraction was dissolved in dry acetonitrile (12 μL) and analyzed by gas chromatography-mass spectrometry (GC-MS; Shimadzu QP2010). The GA fractions were treated with dry pyridine (50 μL) and BSTFA (Sigma-Aldrich; 50 μL) at 60°C for 20 min before drying under a stream of nitrogen. The GA fractions were dissolved in dry acetonitrile (12 μL) and analyzed by GC-MS. For GC-MS, the samples (1 μL) were injected onto a BPX-5 capillary column (SGE 30 m × 0.25 mm, 0.25 μm) using helium as the carrier gas (1 mL min⁻¹), and the inlet temperature was 280°C. The initial oven temperature was set at 50°C and held for 1 min before increasing to 200°C at 15°C min⁻¹. The temperature was increased at 3°C min⁻¹ up to 270°C followed by 15°C min⁻¹ to 320°C and held for 5 min. The transfer line was set at 280°C, the ion source was 200°C, and the ionization potential was 70 eV. The analyses were performed in selected ion monitoring mode, monitoring for the following sets of ions: 190 and 194 for ABA/d₆ABA, 284 and 286 for GA₄/d₂GA₄, 506 and 508 for GA₁/d₂GA₁. Amounts were calculated based on the ratio of these ions with corrections made to account for contributions to the area of the deuterated peak from endogenous ions. Deuterated standards were calibrated with authentic samples to provide an accurate measurement of the endogenous hormone levels.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Time course of germination for PD Arabidopsis ecotypes.

• Supplemental Figure S2. Seven-day germination of afterripening Ler seed.

• Supplemental Figure S3. KAR₁ effects on ABA mutant seed germination.

• Supplemental Figure S4. KAR₁ effects on GA mutant seed germination.

• Supplemental Figure S5. Transcriptional changes of ABA-related genes during imbibition.

• Supplemental Figure S6. Transcriptional changes of GA-related genes during imbibition.

• Supplemental Table S1. Primer sequences for qRT-PCR analysis.