SMAX1 and SMXL2 inhibits Arabidopsis germination in the dark

The conditions used to germinate Arabidopsis vary widely in the literature. Typically, after-ripened seeds are exposed to various sterilizing agents and then plated onto rich nutrient agar plates with or without a sugar source. Moreover, many dark-based experiments involve pre-exposing seeds to a complicated light regime involving red and far-red light exposure before germination to ensure the ratio of active (P_fr) to inactive phytochrome (P_r) is experimentally correct. To analyze the germination of various genotypes and perform large-scale mutant screens under light-limited conditions, we simplified dark germination conditions of treated seeds in the following way.

First, we removed the confounding effects of seed age on levels of dormancy by only using seeds that had ripened for at least six months; a timeframe based on observations that at least three months of seed ripening is required to maximally decay PHYB to its P_r form [41]. We tested this by giving seeds a five-minute pulse of far-red light and found similar germination responses to our dark conditions (S1 Fig). Second, seeds were plated on water and agar media devoid of nutrients and a carbon source, which removes the promotive effects of nitrogen and carbon on germination. Finally, we were concerned that seed sterilization may influence seed germinability in the dark and tested the effects of no sterilization, ethanol sterilization, and bleach sterilization on the germination of seeds in the dark (S2A Fig). We found unsterilized or ethanol-sterilized seeds of wild-type (Col-0) reproducibly did not germinate under our dark conditions (S2A Fig). Bleaching seed, however, markedly increased the germination of these genotypes in the dark, which most likely reflects a weakening of the seed coat due to the bleach (S2A Fig). Because experiments would now require unsterilized seeds, we added the fungicide benomyl to the media and found this has no obvious effect on germination or seedling viability (S2B Fig). Based on these results, our dark germination experiments of plating six-month-old dry unsterilized seeds directly onto water agar plates followed by immediate placement in the dark is a useful and simple method of testing dark germination response in Arabidopsis.

To further characterize wild-type seed germination under light-limiting conditions, we exposed imbibed seeds to a low light fluence (~0.3 μmol m^-2 s^-1) for increasing time periods to determine a threshold of light duration required for germination. Wild-type seeds showed a slow, steady rise in germination as low-light exposure progressed. In contrast, htl-3 (kai2) seed germinated poorly over the same timeframe, suggesting that the AtKAI2 pathway positively contributed to germination under limiting light conditions, which is consistent with the original identification of KAI2 as ‘hyposensitive to light’ at the level of inhibition of hypocotyl growth (S2C Fig) [13]. Two additional genetic experiments further supported this premise. First, two well-characterized htl-3 lines overexpressing AtKAI2 (AtKAI2OX-A, AtKAI2OX-B) germinated in the absence of light at effective KAR₂ concentrations (EC₅₀) in the low nanomolar range (Fig 1A and 1B) [8]. Second, genetic inactivation of the downstream negative regulator of AtKAI2 signaling, SUPPRESSOR OF MAX2 1 (SMAX1) resulted in partial germination in the dark (Fig 1C). When a loss-of-function mutation in SUPPRESSOR OF MAX2 LIKE 2 (SMXL2), a partially redundant ortholog of SMAX1, was introduced into this line (smax1-2; smxl2-1), almost all seeds germinated constitutively in the dark (Fig 1C). Consistent with our genetic results, the expression of genes known to be induced by AtKAI2 activation (DLK2, KUF1, BBX20) increased in dark-germinated AtKAI2OX-A seeds treated with KAR₂ (Fig 1D) [42,43]. As predicted, these genes were constitutively expressed in smax1-2 single and smax1-2; smxl2-1 double mutants (Fig 1D). Our results indicate that SMAX1 and SMXL2 contribute to the repression of germination of Arabidopsis seeds in the dark.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Activation of AtKAI2 signaling promotes Arabidopsis germination in the dark.

(A) Representative images of Arabidopsis genotypes germinating on minimal mock media (DMSO) or media containing 0.5 μM KAR₂ under continuous dark conditions. (B) Effective concentrations (EC₅₀) of KAR₂ required to germinate 50% of two AtKAI2 overexpression lines (AtKAI2OX-A, AtKAI2OX-B) [8], wild-type Col-0, and htl-3 seed. Each point represents the mean germination percentage of three biological replicates. Bar = SD. (C) Dark germination of Col-0, smax1-2, and smax1-2; smxl2-1 double homozygous mutant. Each open circle represents the germination percentage of a biological replicate. Bar = SD (D) Expression (RT–qPCR analysis) of KAR-induced genes (DLK2, KUF1, and BBX20) [42,43] in dark-germinated AtKAI2OX-A seeds on 0.5 μM KAR₂. AtACTIN8 gene was used as an internal control. Lower case letters indicated a significant difference compared to untreated (DMSO) control seeds (P < 0.05, ANOVA with post-hoc one-sided Fisher’s least significant difference). P values, sample sizes and plot elements are provided in S1 Table.

https://doi.org/10.1371/journal.pgen.1011447.g001

Light reverses SMAX1 accumulation in the dark

In Arabidopsis, SMAX1 and SMXL2 are degraded by the F-box protein, MAX2, when AtKAI2 is activated by either the (-)-(2’S) isomer of rac-GR24 or by KARs [44–46]. Evidence for this model, however, often requires special experimental conditions because of difficulties in detecting SMAX1 protein in Arabidopsis [44]. Our result that SMAX1 is a negative regulator and represses germination in the dark led us to hypothesize that the SMAX1 protein in Arabidopsis should accumulate in the dark. To test this, we fused eYFP (enhanced yellow fluorescent protein) to SMAX1 and placed it under the control of the ubiquitously expressed 35S promoter (35S::eYFP-SMAX1). We generated stable transgenic lines in a smax1-2 loss-of-function mutant background and found that the eYFP-SMAX1 transgene complemented smax1-2 phenotypes, such as hypocotyl length and germination in the presence of the GA biosynthesis inhibitor, paclobutrazol (PAC) (S3 Fig).

In contrast to seeds germinated in the light, dark-grown seedlings showed a strong nuclear-localized eYFP signal throughout the root (Fig 2A). Importantly, the dark-dependent SMAX1 accumulation disappeared within six hours of light exposure (Fig 2A). When we performed the reciprocal experiment, where light-grown seedlings were moved to the dark, we observed nuclear-localized eYFP signal within six hours (Fig 2B). This signal accumulated to higher levels after twelve hours in the dark (Fig 2B). To assess if the eYFP-SMAX1 line was responding correctly, we reperformed our light shift experiments in the presence of rac-GR24. Exposure to rac-GR24 activates AtKAI2 signaling through the (-)-(2’S) isomer and should result in SMAX1 degradation. Indeed, only weak YFP signal was detected in rac-GR24 treated roots under all conditions (Fig 2A and 2B). In summary, SMAX1 accumulates in the dark. Genetic analysis indicates that SMAX1 accumulation represses germination in the absence of light stimulus. Upon exposure to light, the SMAX1 signal disappears. The ability of rac-GR24 to prevent SMAX1 accumulation in the dark suggested that SMAX1 stability depends on activation of AtKAI2 signaling.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. SMAX1 accumulates in the dark.

(A) Representative z-stack projection images of an overexpression eYFP-SMAX1 line (smax1-2; 35S::eYFP-SMAX1). Seeds were exposed to light for six hours to initiate germination and then transferred to total darkness for 48 hours. After 48 hours in the dark, seedlings were exposed to light for 6 or 12 hours. The red outline is propidium iodide staining of the cell walls. The duration of light and dark treatments is drawn at the top of the figure. (B) Representative z-stack projection images of an overexpression eYFP-SMAX1 line (smax1-2; 35S::eYFP-SMAX1) after 48 hours of continuous light germination with or without 1μM rac-GR24. Seeds were exposed to light for 48 hours. After 48 hours in the light, seedlings were placed in the dark for 6 or 12 hours. The red outline is propidium iodide staining of the cell walls. The duration of light and dark treatments is drawn at the top of the figure. Scale bar = 25μm.

https://doi.org/10.1371/journal.pgen.1011447.g002

SMAX1 regulates PIF1 to mediate hormone action

Although SMAX1 and SMXL2 repress seed germination under dark conditions, it is not clear how these proteins regulate this process. Activation of AtKAI2 is thought to primarily target SMAX1 and SMXL2 proteins for degradation [44–46]. Taking advantage of this specificity, we compared global gene expression changes between untreated and KAR₂-treated AtKAI2OX seeds imbibed in the dark. Despite the absence of light stimulus, KAR₂-treatment of AtKAI2OX seed upregulated the expression of many genes typically only induced by light (Fig 3A and S2 Table) [47]. Seedling greening after germination is a highly regulated process that balances the establishment of photosynthesis whilst minimizing photooxidative damage and many of the involved genes are controlled by phytochrome through inhibition of PIF1 [48]. Consistent with this, while cataloguing components involved in phytochrome signaling, we noticed that expression of several basic helix-loop-helix (bHLH) family members including PACLOBUTRAZOL RESISTANCES (PREs), PHYTOCHROME RAPIDLY REGULATED (PARs) and LONG HYPOCOTYL FAR RED1 (HFR1) that contribute to seedling greening, were upregulated in KAR₂-treated AtKAI2OX seed (Fig 3B and S2 Table). These bHLHs are thought to fine-tune phytochrome responses by forming feedback loops where the transcription factor, PIF1, promotes HFR1, HECATES (HECs), and PAR expression [49–51]. In turn, the HFR1, HECs, and PAR proteins sequester PIF1 to prevent transcription of downstream targets [49–51]. Notably, 94 of the 166 known direct targets of PIF1 were activated two-fold by KAR₂ (Fig 3C and S3 Table). [29]. This large overlap of gene expression together with the shared dark germination phenotype of pif1-1 and smax1-2 mutants, suggest that a genetic relationship exists between these pathways. Indeed, epistatic analysis between htl-3 and pif1-1, which germinates poorly and constitutively under light-limiting conditions, respectively, places PIF1 at or downstream of AtKAI2 in a genetically dependent pathway (Fig 3D).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Activated AtKAI2 signaling perturbs light signaling.

(A) Heatmap representation of the expression of genes annotated as components of the photosynthetic apparatus in dark germinated AtKAI2OX seed treated with KAR₂. All heatmaps depict log₂-transformed fold-change in gene expression between AtKAI2OX+KAR₂/AtKAI2OX at 24h. See S2 Table for gene lists and values. (B) Heatmap representation of the expression of genes annotated as regulatory components of phytochrome signaling in dark-germinated AtKAI2OX seed treated with KAR₂. (C) Heatmap representation of the expression of annotated PIF1-regulated genes [29] in dark-germinated AtKAI2OX seed with KAR₂ at 24 hours. See S3 Table for gene lists and values. (D) Representative pictures of htl-3, pif1-1, and htl-3; pif1-1 seed placed under dark conditions for five days.

https://doi.org/10.1371/journal.pgen.1011447.g003

Since many of the genes downstream of PIF1 are annotated as responsive to hormones, PIF1 appears to contribute to establishing the correct combination of hormonal responses required for germination [29]. We observed that the expression of genes involved in hormone response affected by the loss of PIF1 were also regulated by KAR₂ activation of AtKAI2 in dark germinated AtKAI2OX seed (Fig 4A and S4 Table). The ratio of the hormones, ABA to GA is crucial for tuning the seed for germination [52]. For example, PIF1 partially represses germination by increasing ABA and decreasing GA levels through SOM [31]. SOM increases ABA levels by activating ABA anabolic genes, ABA1, NCED6, and NCED9 as well as repressing the ABA catabolic gene, CYP707A2 [31]. SOM also decreases the expression of GA anabolic genes, GA3ox1 and GA3ox2 and increases the transcript levels of the GA catabolic gene, GA2ox2 [31]. We found that KAR₂ activation of AtKAI2 in dark-germinated AtKAI2OX seed decreased transcript levels of the SOM targets ABA1, NCED6 and NCED9 (Fig 4B and S2 Table). AtKAI2 activation also increases the levels of GA3ox1 and GA3ox2 transcripts whilst decreasing GA20ox3 (Fig 4B). Unlike SOM, however, AtKAI2 activation induced expression of a different CYP707 homolog (CYP707A3 versus CYP707A2) and decreased expression of GA2ox2 (Fig 4B). Even considering this nuance, the transcript profile of KAR₂ activation of AtKAI2, demonstrated an overall similarity to the transcript profile of pif1 loss-of-function seeds. Consistent with a decrease in ABA and increase in GA levels, the expression of genes annotated as ABA-regulated, decreased upon AtKAI2 activation while transcripts of GA-responsive genes increased (Fig 4A and 4B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Activating AtKAI2 signaling perturbs hormone signaling.

(A) Heatmap representation of the expression of genes annotated as PIF1-regulated hormonal response genes [29] in dark-germinated AtKAI2OX seed with KAR₂ relative to mock treated (AtKAI2OX+KAR₂/AtKAI2OX). All heatmaps depict log₂-transformed fold-change at 24h. See S4 Table for gene lists and values. (B) Heatmap representing select ABA-and GA-regulated genes in dark germinated AtKAI2OX seed with KAR₂. Dark red pathways are ABA-regulated genes. Purple pathways are GA-regulated genes. Heat map values are the ratio of AtKAI2OX+KAR₂/AtKAI2OX gene expression after 24 hrs. See S2 Table for gene lists and values. (C) Germination of smax1-2; smxl2-1 seed in the dark on 3 μM ABA. Each open circle represents the germination percentage of a biological replicate. Bar = SD. (D) Germination of smax1-2; smxl2-1 seed in the dark on 20 μM PAC. Each open circle represents the germination percentage of a biological replicate. Bar = SD. Asterisks indicated a significant difference compared to untreated (mock) smax1-2; smxl2-1 seeds (P < 0.05, one-way ANOVA with post-hoc Tukey honest significant difference test). P values, sample sizes and bar plot elements are provided in S1 Table.

https://doi.org/10.1371/journal.pgen.1011447.g004

To functionally test the contributions of ABA and GA to AtKAI2 signaling, we next manipulated the levels of these hormones in the smax1-2; smxl2-1 double mutant and monitored dark germination. Adding ABA to smax1-2; smxl2-1 seeds partially inhibited germination in the dark (Fig 4C) as measured by radicle emergence, but ABA did fully inhibit seedling emergence (S4 Fig). These phenotypes are consistent with the previous finding that ABA does not increase the DELLA repressors of GA signaling to levels needed to inhibit seed coat rupture [53]. The inability of smax1-2; smxl2-1 seed to overcome the capacity of ABA to stop seedling greening also suggests that ABA signaling is functional after testa rupture. In contrast to increasing ABA, we found that reducing GA levels by exposing smax1-2; smxl2-1 seed to paclobutrazol (PAC), a GA biosynthesis inhibitor, fully inhibited germination (Fig 4D).

Identification of dark germination mutants

The ability of the smax1-2 loss-of-function mutant to germinate in the dark encouraged us to further probe this process by performing an unbiased forward genetic screen to identify new mutations in AtKAI2 signaling. We screened 48,000 EMS mutagenized Col-0 seeds (48 M₂ pools) for mutants capable of germinating under our dark conditions and identified 122 potential mutants (Fig 5 and S5 Table). To speed up the mutant classification, we used a chemical genetics approach where various compounds that are known to differentiate different hormone- and light-related mutants were tested against the mutant collection. We exposed mutant seed to ABA and PAC in the light as these two compounds are commonly used to probe the ABA and GA signaling pathways as they pertain to Arabidopsis wild type seed germination. We also included a condition of PAC plus rac-GR24 to see if any PAC sensitive mutants could be rescued by addition of this artificial SL/KL. This information was then used to classify groups using two-dimensional hierarchical clustering. The response of a reference set of known hormonal or light signaling mutants to our compounds were also queried through this system to further catalog our dark germination lines. This reference set included an ABA-auxotroph, aba deficient2 (aba2-2), two ABA-insensitive mutants, aba insensitive1 (abi1-1), aba insensitive4 (abi4-5), three constitutive light signaling mutants, phytochrome interacting factor1 (pif1-1), de-etiolated1 (det1-1), constitutive photomorphogenic1 (cop1-4) and the constitutive AtKAI2 signaling mutant, smax1-2.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. A genetic screen for dark germination mutants.

Heatmap representation of mutant germination phenotypes on 20 μM PAC, 20 μM PAC, and 1 μM rac-GR24, 2 μM ABA, or in the absence of light. The germination responses were hierarchically clustered using the complete linkage method with Euclidean Distance. The mutants fall into three groups: det1-1 like (Group 1), abi1-1 like (Group 2), and smax1-2 like (Group 3). The photographs show the bleached seedling phenotype of various mutants grown on 2.5 μM CTL-VI. Bolded cells indicate mutants which showed wild-type sensitivity to CTL-VI. smax1-4 is a novel SMAX1 allele. See S5 Table for putative mutant list and values.

https://doi.org/10.1371/journal.pgen.1011447.g005

Using these criteria, we identified three major groupings. Group 1 mutants did not germinate well on either PAC or ABA, indicating they are not perturbed in GA or ABA signaling (Fig 5). Within this group, the light signaling mutants pif1-1 and det1-1, which also germinate in the dark, did not germinate well on PAC or ABA suggesting this group may be enriched for mutants defective in light signaling (Fig 5). Group 2 mutants showed moderate germination in the dark, high germination on ABA, and varying germination on PAC (Fig 5). As expected, this group included abi1-1 and abi4-5 and therefore most likely is enriched for ABA signaling mutants (Fig 5). Finally, Group 3 mutants, which included smax1-2 and aba2-2 seed, generally germinated well in the dark and on PAC in the presence or absence of rac-GR24 but did not germinate on ABA (Figs 5 and S4). Mutants perturbed in KL signaling can be identified using a compound called cotylimide-VI (CTL-VI) [54]. Addition of CTL-VI to germinating Arabidopsis seed bleaches their cotyledons and max2 mutants, which are insensitive to SL or KL signaling are insensitive to this bleaching effect [54]. We, therefore, thought mutants constitutive for KL signaling, like smax1-2, may show increased sensitivity to CTL-VI application versus ABA auxotrophs with respect to cotyledon bleaching. Indeed, CTL-VI preferentially bleaches smax1-2 cotyledons compared to aba2-2 and abi4-5 cotyledons (Figs 5 and S5). Interestingly, five mutants from Group 3 showed a similar bleaching phenotype on CTL-VI to that of smax1-2 and subsequent sequencing for the SMAX1 locus identified a new smax1 allele (smax1-4) carrying a nonsense mutation in codon 118 (CAA→TAA). The smax1-4 allele is predicted to produce the smallest truncated product among all the known SMAX1 alleles [37] and therefore could represent a good null mutant. Importantly, our isolation of a novel allele in SMAX1, smax1-4, indicated that our screening conditions for constitutive dark germinating mutants can identify constitutive AtKAI2 signaling mutants and directly connects the dark germinating phenotype to the AtKAI2 signaling pathway.

SMAX1 and SMXL2 inhibits Arabidopsis germination in the dark

Generally, for plants like Arabidopsis, GA signaling is the major germination promotive pathway while the KL pathway appears to only have a minor role. By contrast, fire-following ephemeral weeds and some parasitic plant species have an absolute requirement for HTL/KAI2 for germination and are not responsive to GA [3,7,8,11]. We believe we have reconciled these species differences by showing that Arabidopsis germination like parasitic and ephemeral weeds can also become completely dependent on AtKAI2 signaling under appropriate environmental conditions. Consistent with this, smax1; smxl2 double mutants show efficient germination under dark conditions. Loss of SMAX1 and SMXL2 function by the activation of AtKAI2 produced a similar transcriptional signature as loss of the light-dependent repressor, PIF1. Epistatic analysis suggested that these genes act in a common signaling pathway. On this note, SMAX1 reduces the inhibitory effect of PHYB on PIF4, which indicates that phytochrome requires activated AtKAI2 signaling to repress PIF function in the seedling [55]. Since our genetic epistasis and transcriptome comparisons demonstrated a functional and mechanistic connection between SMAX1 and PIF1, we predict that SMAX1 and SMXL2 are also necessary for PIF1 function in the dark.

The ability of light to inhibit SMAX1 accumulation could explain some of the previous experimental difficulties involving SMAX1 cell biology. Fluorescence is not detected in transgenic Arabidopsis containing SMAX1-GFP fusions even when the F-box MAX2, which is thought to control SMAX1 degradation, is genetically removed [44]. By contrast, under our dark growth conditions, we detected accumulation of eYFP-SMAX1 in planta with the signal diminishing over a few hours in the presence of rac-GR24 or light. The ability of light to inhibit SMAX1 accumulation is similar to a previous study that did not detect SMAX1-GFP in light-grown seedlings unless the SMAX1 degron motif was deleted [44]. However, our results differed from SMAX1 accumulation studies in the hypocotyl, which conclude that light encourages SMAX1 accumulation [55]. Perhaps these results reflect differences in the light regime as well as the tissue and developmental stage used in the studies. In the hypocotyl study, older seedlings were exposed to a short-day light exposure of 8 hours light and 16 hours dark, in contrast to our experiments which were conducted on germinating seeds under constant light or dark. The difference in light exposures could affect the dynamics of SMAX1 accumulation. At different developmental stages, SMAX1 stability could differ in the dark compared to light. Unfortunately, the absence of a KL ligand candidate limits speculation on how light conditions may affect this ligand composition, concentration or sensitivity. However, light and the addition of KARs or rac-GR24 stimulate AtKAI2 transcription, which is consistent with our findings that SMAX1 is degraded in the light during early stages of germination [12,40]. Whatever the case, special consideration regarding the development stage of the plant and light regime should be given when studying SMAX1 stability.

Low nanomolar KAR₂ concentrations were sufficient to stimulate high germination rates of AtKAI2OX seed in the dark, which was unexpected given that KAR₂ addition to AtKAI2OX seed in the light only partially germinates seeds depleted of GA [40]. There are numerous reports of GA and SL interactions at the level of synthesis; for example, GA addition negatively regulates the levels of SL in rice [56]. As light is known to increase GA synthesis this would suggest SL levels are lower in the light perhaps sensitizing the AtKAI2 signaling pathway to exogenous KL [30]. Of course, because KL is not identified, we are not sure this compound is regulated in a similar manner to SL. More directly, SMAX1 interacts with DELLA domain proteins and this interaction appears to enhance SMAX1 transcription and inhibit expression of a gene involved in GA biosynthesis [38]. These relationships support that DELLA and SMAX1 positively enhance one another. Whatever the case, our genetic studies suggest that the SMAX1-DELLA module act together on Arabidopsis germination under low light conditions [38].

An alternative explanation for light-dependent SL sensitivities could revolve around the role of PIF proteins [57,58]. Although this interaction appears antagonistic in the hypocotyl, the PIF1-DELLA module appears to be positive in the seed. PIF1 activation reduces GA levels, which in principle would encourage DELLA protein stabilization (Fig 6). In this context, PIF1 function also requires SMAX1 and SMXL2, thus inactivation of these proteins through SL signaling increases GA resulting in DELLA protein degradation (Fig 6). Accordingly, seed germination in the dark will be PAC dependent. In the light, all PIF proteins are inactivated by phytochromes, consequently DELLA repressors are inactive (Fig 6). This light-dependent loss of DELLA function removes the requirement of GA for germination thereby making seed germination insensitive to PAC.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. A working model on the contributions of SMAX1 and SMXL2 to germination under dark versus light conditions.

Far left panel, in the dark, PIF proteins actively repress light response genes. Under low KL conditions, SMAX1 and SMXL2 activate PIF1 to reduce the GA to ABA ratio discouraging germination. SMAX1 and SMXL2 also could act to directly dampen germination as SMAX1 interacts with the EAR motif repressors, TOPLESS1 (TPL1) and TOPLESS RELATED1, 2, 4 (TPR1, 2, 4) [59–62]. Center left panel, increasing KL inhibits SMAX1 and SMXL2 function to both alleviate the repressive effects of these proteins and increase the GA/ABA ratio. Although this allows germination in the dark, inhibition of GA synthesis by addition of the GA synthesis inhibitor, PAC, will decrease the germination response due to the repressive effects of DELLA proteins. Right central panel, in the light, phytochromes inactivate PIF proteins resulting in DELLA protein inactivation. Although seeds no longer require GA to germinate, SMAX1 and SMXL2 proteins still maintain their function to attract TPL and TPR. Also, SMAX1 is thought to inhibit PHYB inactivation of PIF4 in the light. This may be true of other PIF proteins. These events would suggest a weaker germination response in low KL conditions. Far right panel, increased KL levels inhibit SMAX1 and SMXL2 function and this in combination with PIF inactivation of DELLA repressors will lead to a stronger germination response in the presence of PAC.

https://doi.org/10.1371/journal.pgen.1011447.g006

Like all working models there are complexities. For example, differences in light versus dark germination in response to karrikins may also reflect metabolic differences under these environmental conditions. Possibly, KARs are converted to more potent ligands in the dark or perhaps are degraded more effectively in the light. It is important to note that neither the addition of KAR₂ nor the overexpression of AtKAI2 was solely sufficient to trigger germination under our dark conditions, which suggests that the expression of AtKAI2 and production of KL are tightly controlled. Whatever the case, our results demonstrated a major role for SMAX1 and SMXL2 signaling in regulating Arabidopsis germination. Moreover, we suggest darkness is an optimal environmental condition to study the roles of AtKAI2 signaling in Arabidopsis germination.

Conservation of the roles of HTL/KAI2 signaling in germination

Our analysis gives some insight into the evolution of germination signaling pathways in ephemeral weeds that are fire followers. Adapting a signaling pathway for a particular process broadly follows scenarios where either the pathway is established for its function by natural selection (historical genesis) or the pathway, which has other functions, is co-opted for new uses (current utility) [63]. Fire-following behaviour is a derived trait, thus our findings in Arabidopsis suggest the role of HTL/KAI2 signaling in these species is the product of historical genesis rather than current utility. Our results suggest in generalist plants like Arabidopsis, SMAX1 and SMXL2 inhibit germination in the absence of light like PIF1. Unlike PIF1, however, which is modulated by a ubiquitous environmental cue, light, SMAX1 and SMXL2 levels depend on light and specific small molecules generated by distinct environmental conditions. Because SMAX1 and SMXL2 function is necessary to repress dark germination, any selective pressures affecting their levels will dominate germination behaviours.

These findings also suggest that smoke-derived karrikins substitute for light in germinating the positive photoblastic lettuce cultivar, L. sativa cv. Grand Rapids, through activation of the KAI2 pathway leading to inactivation of SMAX1 [20–22]. Finally, our results emphasized that defining signaling pathways by their ligands and receptors can be confounding when these signals converge on conserved downstream effectors which are also modulated by other signaling pathways. We suggest when looking at fundamental processes such as germination, it is more informative to reference pathways by their bottom effectors like SMAX1, rather than by specific receptors [64].

The genetic space of dark germination

The establishment of simple dark conditions creates an experimentally tractable environment to identify genes involved in the germination process. Furthermore, these conditions can be combined with small molecule compounds that perturb germination to quickly classify mutants into categories for subsequent genetic analysis. Traditionally, mutants are classified through genetic complementation, which is tedious in Arabidopsis. Using a chemical biology approach to classify mutant lines is based on the concept that each chemical represents a unique environmental condition [65]. With enough compounds and two-dimensional hierarchical clustering, in principle, different classes of mutations can be grouped without performing complementation testing. By focusing on compounds that perturb germination and early seedling growth, we showed the utility of using only four compounds to classify dark germination mutants into three categories, one of which included a new SMAX1 allele.

In this study, we limited ourselves to chemical mutagenesis which will mostly enrich for loss-of-function mutations in regulators that impede germination under dark conditions but similar experiments with mis-expression collections would identify positive components [66,67]. In our screen, we identified three classes of mutations. DET1, for example, is known to directly interact with and promote the protein stability of PIF1 and our screen placed loss-of-function mutations in DET1 and PIF1 into one class (Group 1) [68]. Mutations in another key light component, COP1 resulted in low germination rates in the dark but this weak phenotype most likely reflects the leakiness of the cop1-4 allele we used as a reference [69]. Hence, our dark germination screens should enrich for mutants that influence light signaling.

Our screen should also identify new mutations in ABA synthesis and signaling. This may be useful as the identification of mutants in the ABA pathway action often requires screening in the presence of exogenous hormones. Because our dark screen does not require any chemical additions, new alleles may be more relevant to seed physiology. Finally, we expected to identify new negative regulators in AtKAI2 signaling and to this end, we identified a new SMAX1 nonsense allele. Presently, all the current SMAX1 alleles either contain nonsense mutations or T-DNA insertions [37].

Our study presents new evidence that activation of the AtKAI2 signaling pathway can modulate hormone action to germinate in the absence of light. Loss-of-function of the downstream negative regulator, SMAX1, can constitutively germinate in the dark and through cell biology experiments, we show that SMAX1 is stable in the dark. Interestingly, the gene expression profile of activated AtKAI2 signaling reflects that of a light signaling mutant, pif1, and we genetically place PIF1 at or downstream of the AtKAI2 signaling pathway. Lastly, we utilize the dark germinating phenotype of constitutive AtKAI2 signaling mutants to screen for mutants in this pathway and display the advantages of screening during light-limiting conditions. Therefore, we present novel insights into the AtKAI2 signaling pathway with respect to the light signaling pathway.

Plant material and plant growth

Arabidopsis thaliana Col-0 was the default wild-type (WT) for this study, except for abi1-1 (Ler) [70]. htl-3 (kai2) [71], smax1-2 (SALK_128579) [37], smax1-2/smxl2-1 [37], pif1-1 [26], abi4-5 [72], cop1-4 [72], aba2-2 [73], abi1-1 [70], det1-1 [74], and Arabidopsis kai2 mutant overexpressing Arabidopsis AtKAI2 (35S::AtKAI2) or Striga ShHTL7 (35S::ShHTL7) [8] were described previously. Arabidopsis thaliana plants were grown in a continuous 24-hour light chamber at 22°C. Seeds were harvested and stored in constant conditions (24°C in the dark) for after ripening. All genotypes used in this study are listed below and in S7 Table.

Plasmid construction and plant transformation

The 35S::eYFP cassette was amplified from pGWB542 and cloned into pGREENII-0179 vector using Fast Digest KpnI and ApaI restriction enzymes (ThermoScientific, cat no. FD0524, cat no. FD1414) and primers 35S-YFP_F_KpnI:5’ CGGGGTACCTGAGACTTTTCAACAAAGGGT 3‘ and 35S-YFP_R_ApaI:5’ GCGGGCCCCTTGTACAGCTCGTCCATG 3’. Arabidopsis SMAX1 was amplified from Col-0 cDNA using primers AtSMAX1_F_ApaI: 5’ GCGGGCCCATGAGAGCTGGTTTAAGTACGA 3’ and AtSMAX1_R_HindIII: 5’ CGGAAGCTTTCATACTGCCAAAGTAATAGTTGT 3’ and cloned into pGREENII-0179-eYFP using Fast Digest ApaI and HindIII (Thermofisher, cat no. FD1414, cat no. FD0504), resulting in the pGREENII0179-eYFP-AtSMAX1 construct. All primers used in this study are listed in S6 Table. The construct was transformed into the smax1-2 mutant by Agrobacterium tumefaciens GV3101- mediated floral dip transformation [75].

Double mutant construction

Putative htl-3; pif1 double mutants were identified from F₂ offspring of htl-3; ShHTL7OX x pif1 cross by first screening for the pif1 dark germination phenotype. Plants were then screened for the htl-3 round rosette leaves and long hypocotyl phenotype. These lines were also screened on glufosinate ammonium (BASTA) to look for lines that were 100% sensitive on BASTA to ensure that the ShHTL7OX transgene was absent. Potential htl-3; pif1 homozygous lines were confirmed by genotyping of individual plants by PCR for the htl-3 and pif1 allele. Primers used for genotyping are listed in S6 Table. See S6 Fig for workflow and PCR gels confirming htl-3; pif1 double mutants.

Dark germination assays

Seeds for dark germination assays were left to dry for at least six months in a glass tube and closed cardboard storage box at room temperature (24°C) in the laboratory before use in germination assays. Approximately 25 to 35 of unsterilized dry seeds per replicate were sprinkled directly onto 10 mm petri plates in the dark containing water agar (0.8%) and the fungicide benomyl (10 μg/ml) (Toronto Research Chemicals, B161380). Seeds were exposed to low light (~0.1 μmol m^-2 s^-1) for less than 30 seconds while plating and then plates were wrapped with three layers of aluminum foil. Plates were placed at 4°C for four days and then incubated at 22°C for five days in the dark. Germination was quantified using radicle emergence. Seedling emergence was quantified by counting fully elongated hypocotyls.

For all chemical treatments, rac-GR24 was used which is a racemic mixture of (+)-GR24 and (-)-GR24. A stock solution of rac-GR24, PAC (Cayman Chemical Company, 18864), and ABA (Sigma-Aldrich, A1049-100MG) was diluted in DMSO, 100% ethanol, and 100% methanol, respectively. For dark germination assays on compounds, the fungicide benomyl (10 μg/ml) and the appropriate compound at specific concentrations (3 μM ABA, 20 μM PAC) or the solvent were added to the water agar media (0.8%) at 1/1000 concentration.

Light germination assays

Germination assays on paclobutrazol, cotylimide, and ABA in the light were performed as previously described [40, 54]. Seeds were surfaced sterilized with 70% (v/v) ethanol and gently mixed for five minutes. After the 70% ethanol was removed, 100% ethanol was added, gently mixed for 30 seconds, and then removed. Once dry, approximately 25 to 35 seeds were sprinkled directly onto 10mm petri plates of ½ MS (Murashige-Skoog) agar (0.8%) containing the appropriate compound at specific concentrations or the solvent at 1/1000 concentration. Plates were placed at 4°C for four days and then incubated at 24°C for five days in the light. Germination was quantified using radicle emergence. Seedling emergence was quantified by counting fully elongated hypocotyls.

Far-red germination assays

Approximately 25 to 35 of unsterilized seeds per technical replicate were sprinkled directly onto 10 mm petri plates in the dark containing water agar (0.8%) and the fungicide benomyl (10 μg/ml) (Toronto Research Chemicals, B161380). Plates were wrapped with three layers of aluminum foil and placed at 4°C for four days. Plates were treated with far-red light at ~5 μmol m^-2 s^-1 for 5 minutes and then wrapped in three layers of aluminum foil and incubated at 22°C for five days in the dark. Germination was quantified using radicle emergence.

Transcript profiling and data analysis

Total RNA from three biological replicates of KAI2OX seeds that germinate in the dark for 24 hours in the presence of 0 μM or 0.5 μM KAR₂ was isolated using the mini hot phenol method [76], following the clean-up using RNase-free DNase I and Monarch RNA Cleanup Kit (New England Biolabs) and was used for single-end 1X75bp RNA-Seq on a NextSeq500 sequencing platform (Illumina). After examining the data quality using FastQC (v0.11.5) [77], the fastq data containing the raw reads for each sample was trimmed and filtered using htStream (v1.3.0) (UC, Davis) [78] to generate clean reads. For each read, trimming and filtering processes include removing the last base, removing the low-quality fragments, removing ‘N’s and returning the longest fragments, and discarding reads that are less than 25 bp. Clean reads from each sample were subjected to HISAT2 [79] (v2.1.0) for genome mapping using Arabidopsis thaliana TAIR10 genome as a reference. Samtools (v1.10) [80] was used to remove the unmapped and multi-loci mapped reads and convert the uniquely mapped reads into a BAM file. Transcript assembles for each sample were performed using StringTie (v2.1.3) [81] against Arabidopsis thaliana Araport11 [82] annotation and transcriptome assembles from each sample were merged using the merge function in StringTie to generate a combined transcriptome across all sequencing samples to serve as a new reference transcriptome. A customized Python script, stringtieGeneIdReplace2.py was written to replace the StringTie default ’MSTRG’-tagged gene_id with the TAIR10 gene locus name in order to facilitate the comparisons of genes and transcripts between different samples. A StringTie provided Python script, prepDE.py, was used to extract the read counts matrix from each transcriptome assemble and this counts matrix served as the input data for DESeq2 (v1.18.1) [83], an R [84] package for identifying the differentially expressed genes between various comparisons. The transcriptome data is stored under GEO accession number GSE161704. Python script, stringtieGeneIdReplace2.py can be accessed from https://github.com/maplexuci/stringtie_gene_id_replacement.

Gene expression analysis by quantitative RT-PCR

Seed total RNA from three biological replicates of each treatment was isolated using mini hot phenol method [75], following the clean-up using RNase-free DNase I and Monarch RNA Cleanup Kit (New England Biolabs). One μg of DNA-free total RNA was reverse transcribed using LunaScript RT SuperMix Kit (New England Biolabs) according to the manual. cDNA product was diluted by 10 times and 1 μL was used in the quantitative RT-PCR reaction. The primers used are listed in S6 Table. Real-time PCR was performed using Luna Universal qPCR Master Mix (New England Biolabs) on a Bio-rad CFX96 Real-time Detection System (Bio-rad). The PCR reaction was performed in triplicate for each biological replication. Expression levels were normalized against the expression of the endogenous control gene, AtACTIN8, and were calculated using 2^-ΔΔCt method [85]. All the expression data was presented as fold change value and was transformed into log2 value for statistical analysis.

Confocal microscopy

The floral spray technique was used to construct the smax1-2; 35S::eYFP-SMAX1 line [75]. T1 transgenic lines were selected for resistance on hygromycin and propagated to the T2 generation. Line 1 in the T2 generation was phenotyped for PAC and hygromycin resistance and used for confocal imaging. For imaging of smax1-2; 35S::eYFP-SMAX1, seeds were surfaced sterilized with 70% (v/v) ethanol and gently mixed for five minutes. After the 70% ethanol was removed, 100% ethanol was added, gently mixed for 30 seconds, and then removed. Once dry, seeds were sprinkled directly onto 10mm petri plates of ½ MS (Murashige-Skoog) agar (0.8%) in the dark and then wrapped three times in aluminum foil. Plates were placed at 4°C for four days and then exposed to white light (~40 μmol m^-2 s^-1) for six hours. Plates were wrapped in three layers of aluminum foil and placed in a room-temperature dark cabinet for 48 hours depending on the light switch experiment and imaged.

Seedlings with specific light treatments were transferred to glass slides with staining solution containing 10μg/ml propidium iodide (PI) and visualized for nuclear localized eYFP fluorescence. The Leica TCS SP5 confocal laser microscope was used, and all images are z-stack projections with 3 μm per step overlayed in ImageJ. All images are imaged with the 40x oil immersion objective lens.

Genetic screen for new smax1 alleles

To generate an M₁ EMS population, approximately 20,000 seeds were mutagenized in 0.2% for 16 hours and then washed in distilled water 10 times before sowing onto soil. For the primary screen, approximately 48,000 M₂ seeds from 48 pools of EMS mutagenized Col-0 were plated onto water-agar plates. The plates were immediately wrapped in three layers of aluminum foil and stratified for four days at 4°C. After stratification, the plates were incubated in the dark at room temperature for five days. Any dark germinating seedlings were then grown to produce M₃ seeds. For the phenotypic screening, M₃ seeds were plated onto plates containing minimal ½ MS-agar or ½ MS-agar supplemented with either 20 μM PAC, 20 μM PAC + 1 μM GR24, 2 μM ABA, or 2.5 μM CTL-VI. The plates were stratified for four days at 4°C and then incubated under constant white light (~40 μmol m^-2 s^-1) at room temperature for seven days. After seven days, seed germination counted by radicle emergence and ABA resistance counted by cotyledon greening was quantified. A heatmap was generated from this data and hierarchically clustered using the Euclidean distance metric in RStudio.

Statistical analysis

One-way ANOVA analysis was performed using SPSS software (IBM) and Tukey HSD method was used for multiple comparison. Descriptive statistics was perfomed using an online tool CalculatorSoup (https://www.calculatorsoup.com/calculators/statistics/descriptivestatistics.php)

Python libraries scipy, pandas and numpy were used in calculating Pearson’s correlation coefficience R², P-value and 99% confidence interval.