Functional redundancy in the control of seedling growth by the karrikin signaling pathway

Abstract

Main conclusion SMAX1 and SMXL2 control seedling growth, demonstrating functional redundancy within a gene family that mediates karrikin and strigolactone responses.

Strigolactones (SLs) are plant hormones with butenolide moieties that control diverse aspects of plant growth, including shoot branching. Karrikins (KARs) are butenolide molecules found in smoke that enhance seed germination and seedling photomorphogenesis. In Arabidopsis thaliana, SLs and KARs signal through the α/β hydrolases D14 and KAI2, respectively. The F-box protein MAX2 is essential for both signaling pathways. SUPPRESSOR OF MAX2 1 (SMAX1) plays a prominent role in KAR-regulated growth downstream of MAX2, and SMAX1-LIKE genes SMXL6, SMXL7, and SMXL8 mediate SL responses. We previously found that smax1 loss-of-function mutants display constitutive KAR response phenotypes, including reduced seed dormancy and hypersensitive growth responses to light in seedlings. However, smax1 seedlings remain slightly responsive to KARs, suggesting that there is functional redundancy in karrikin signaling. SMXL2 is a strong candidate for this redundancy because it is the closest paralog of SMAX1, and because its expression is regulated by KAR signaling. Here, we present evidence that SMXL2 controls hypocotyl growth and expression of the KAR/SL transcriptional markers KUF1, IAA1, and DLK2 redundantly with SMAX1. Hypocotyl growth in the smax1 smxl2 double mutant is insensitive to KAR and SL, and etiolated smax1 smxl2 seedlings have reduced hypocotyl elongation. However, smxl2 has little or no effect on seed germination, leaf shape, or petiole orientation, which appear to be predominantly controlled by SMAX1. Neither SMAX1 nor SMXL2 affect axillary branching or inflorescence height, traits that are under SL control. These data support the model that karrikin and strigolactone responses are mediated by distinct subclades of the SMXL family, and further the case for parallel butenolide signaling pathways that evolved through ancient KAI2 and SMXL duplications.

Introduction

Strigolactones (SLs) are carotenoid-derived phytohormones with butenolide moieties (Alder et al. 2012; Seto et al. 2014; Seto and Yamaguchi 2014). SLs regulate growth at several developmental stages, including shoot branching (Gomez-Roldan et al. 2008; Umehara et al. 2008), senescence (Yamada et al. 2014; Ueda and Kusaba 2015), adventitious root formation (Rasmussen et al. 2012; Sun et al. 2015), root development (Kapulnik et al. 2011; Ruyter-Spira et al. 2011), and leaf shape (Waters et al. 2012; Scaffidi et al. 2013; Lauressergues et al. 2014). SLs also promote symbiotic arbuscular mycorrhizal associations (Akiyama et al. 2005, 2010; Yoneyama et al. 2008; Cardoso et al. 2014), and activate seed germination of root parasitic plants in the Orobanchaceae family (Xie and Yoneyama 2010). Karrikins (KARs), like SLs, are butenolide molecules that elicit growth responses in plants (Flematti et al. 2004; Nelson et al. 2012). Unlike SLs, KARs are not known to be synthesized in plants, but are a byproduct of burning plant material (Flematti et al. 2013). KARs are important signals in post-fire ecology that activate seed germination of many smoke-responsive species (Nelson et al. 2012). KARs also enhance germination and light-responsive growth of Arabidopsis seedlings (Nelson et al. 2009, 2010).

Both SL and KAR responses require the F-box protein MAX2 (Nelson et al. 2011). As an F-box protein, MAX2 functions as an adapter that confers substrate specificity to Skp-Cullin-F-box (SCF) E3 ubiquitin ligase complexes (Stirnberg et al. 2007; Zhao et al. 2014). Although SL and KAR signaling pathways both involve MAX2, the different growth responses to SL and KAR imply they are perceived distinctly (Nelson et al. 2010, 2011; Waters et al. 2012). SLs and KARs are recognized by the paralogous α/β-hydrolases DWARF14 (D14) and KARRIKIN-INSENSI-TIVE2 (KAI2), respectively (Hamiaux et al. 2012; Bythell-Douglas et al. 2013; Guo et al. 2013; Kagiyama et al. 2013; Nakamura et al. 2013; Zhao et al. 2013, 2015; Umehara et al. 2015). D14 and KAI2 control complementary aspects of MAX2-regulated growth; for example, D14 regulates shoot branching, a SL-responsive phenotype, whereas KAI2 regulates seed germination, a KAR-responsive phenotype (Waters et al. 2012). Both D14 and KAI2 contribute to leaf morphology, and both can influence hypocotyl elongation in seedlings.

D14 is an ancient paralog of KAI2 that may have evolved strigolactone specificity following KAI2 gene duplication (Waters et al. 2012; Conn et al. 2015). As demonstrated by promoter-swapping experiments, D14 and KAI2 do not have interchangeable functions (Waters et al. 2015). Parasitic weeds in the Orobanchaceae family have more recently evolved a SL detection mechanism that enables host-responsive seed germination. These SL receptors are KAI2 paralogs that evolved new ligand-specificity following gene duplication, but have still maintained the ability to activate seed germination in Arabidopsis (Conn et al. 2015; Toh et al. 2015; Tsuchiya et al. 2015). This raises questions about the original function of KAI2 in angiosperms and basal land plants. The observation that Arabidopsis kai2 mutants have germination and seedling phenotypes that are opposite to KAR growth responses, which are not shared with d14 or SL-deficient mutants, points to an endogenous, unknown KAI2 ligand (KL) (Flematti et al. 2013; Waters et al. 2014). This hypothesis is supported by recent cross-species complementation studies, which have shown that SmKAI2a from the lycophyte Selaginella moellendorfii and KAI2c genes from the parasitic weeds Phelipanche aegyptiaca and Striga hermonthica can rescue Arabidopsis kai2 phenotypes, but do not transduce KAR or SL signals (Conn et al. 2015; Waters et al. 2015; Conn and Nelson 2015). In parasitic weeds, KAI2c are basal to KAI2 paralogs with KAR or SL-specificity, and are under strong purifying selection (Conn et al. 2015). KAI2c may encode receptors that are specific for KL, which is anticipated to have a butenolide moiety like KARs and SLs (Conn and Nelson 2015). KARs may fortuitously serve as KL analogs in Arabidopsis thaliana. Thus understanding how KARs regulate growth also sheds light on KL signaling mechanisms, and may provide insights into how SL signaling evolved.

To elucidate the signaling mechanism downstream of MAX2, we previously screened for mutations that suppressed germination and seedling growth phenotypes of max2, and identified SUPPRESSOR OF MAX2 1 (SMAX1) (Stanga et al. 2013). While smax1 suppresses max2 phenotypes associated with KAR signaling, SL-regulated phenotypes such as shoot branching are unaffected (Stanga et al. 2013). Shoot branching (or tillering) is controlled, however, by a SMAX1 homolog in rice, DWARF53 (D53) (Jiang et al. 2013; Zhou et al. 2013). D53 is polyubiquitinated and degraded after SL recognition. This aspect of D53 regulation requires D14 and D3, the rice ortholog of MAX2. SL perception by D14 promotes its interaction with D53 and D3, suggesting formation of an SCF^MAX2-SL-D14-D53 complex (Jiang et al. 2013; Zhao et al. 2013, 2015).

SMAX1 and seven SMAX1-LIKE (SMXL) genes compose a family in Arabidopsis that has weak similarity to ClpB/HSP100 proteins (Stanga et al. 2013). Based on a comparison of monocots and dicots (Zhou et al. 2013), the SMXL/D53 family can be divided into four phylogenetic clades: (1) SMAX1 and SMXL2; (2) SMXL3; (3) SMXL4 and SMXL5; and (4) SMXL6, SMXL7, and SMXL8. The closest D53 homologs in Arabidopsis, SMXL6, SMXL7, and SMXL8, have redundant roles in SL-regulated growth (Soundappan et al. 2015; Wang et al. 2015). Loss-of-function mutations in double and triple smxl6,7,8 mutants suppress the shoot branching, auxin transport, PIN1 accumulation, and lateral root density phenotypes of max2, but do not suppress its seed germination or hypocotyl elongation phenotypes (Soundappan et al. 2015; Wang et al. 2015). Thus there is a division of labor among the SMXL family that suggests SMAX1 is a target of KAR/KL-KAI2 signaling, and SMXL6, SMXL7, and SMXL8 are targets of SL-D14 signaling. As max2 phenotypes are a hybrid of d14 and kai2 phenotypes, this parallel signaling mechanism hypothesis is supported by observations that smax1 max2 rosettes mimic d14, and smxl6,7 max2 rosettes mimic kai2 (Soundappan et al. 2015). However, biochemical evidence for preferred receptor-SMXL interactions or KAR-induced SMAX1 degradation have not yet been shown.

The roles of SMXL2, SMXL3, SMXL4, and SMXL5 are currently unknown. As SMAX1 and SMXL6,7,8 are largely sufficient to account for the KAR and SL growth responses tested so far, it is unclear if other SMXL paralogs participate in KAR/KL or SL signaling. However, we previously observed that smax1 seedlings show a weak, but statistically significant response to KAR₂ treatment (Stanga et al. 2013). This suggested that another gene(s) may contribute to KAR/KL signaling. We hypothesized that SMXL2 function may account for this redundancy because it is the closest paralog of SMAX1. Furthermore, there is an increase in SMXL2 expression in seedlings following treatment with KARs, and SMXL2 expression is reduced in max2 (Stanga et al. 2013). If SMXL2 represses responses to KAR/KL, like SMAX1, this transcriptional regulation of SMXL2 might contribute to feedback regulation of KAR/KL signaling. Here we investigate whether SMXL2 is involved in KAR responses, and identify functional redundancy between SMAX1 and SMXL2 in Arabidopsis seedling growth.

Materials and methods

Plant assays and growth conditions

SAIL_596_E08 (smxl2-1) was obtained from the Arabidopsis Biological Resource Center (Sessions et al. 2002). The kai2 mutant allele was htl-3, a gift of Shigeo Toh and Peter McCourt (Toh et al. 2014), and d14-1 was a gift of Mark Waters and Steven Smith (Waters et al. 2012). The smax1-2 and smax1-2 max2-1 mutants were described in Stanga et al.(2013). Genotyping was performed on leaf tissue using the primers listed in Table S1. Unless otherwise indicated, plants were grown as described in Stanga et al. (2013). Plants were grown under fluorescent white light (~80–110 μmol m⁻² s⁻¹) with 16 h light/8 h dark cycles at 21 °C. Plants for the hyponastic growth assay and leaf morphology assays were grown under 8 h light/16 h dark cycles at 21 °C. Pots were randomized within flats to minimize effects of environmental variation. Plants were harvested when most siliques were brown and dried at room temperature in paper bags for 3 days. Seeds were cleaned, equilibrated in a box containing Drie-rite desiccant for 3 days, and stored at −80 °C to preserve primary dormancy. Before assays, seeds were surface-sterilized with a 70 % EtOH, 0.05 % Triton X-100 solution, rinsed with 70 and 95 % EtOH, and air dried. For the germination assay, seeds were plated onto 0.8 % Bacto-agar. Plates were immediately incubated at 24 °C under constant white light (~40–80 μmol m⁻² s⁻¹). Germination was indicated by the emergence of the radicle tip through the endosperm. For all other assays, seedlings were plated on solid 0.5 × Murashige–Skoog (MS, Sigma M0404, 2.17 g/L) medium. If indicated, the media were supplemented with KAR₂, racGR24, GR24^5DS, or GR24^ent−5DS from 1000X stocks dissolved in acetone. Chemicals were prepared and provided by Dr. Adrian Scaffidi (University of Western Australia). Plates were stratified for 3 days in the dark at 4 °C. Seeds were exposed to 3 h white light (~40–80 μmol m⁻² s⁻¹) and returned to darkness for 21 h. Except the etiolation assays, in which seedlings were kept in darkness for 4 days, seedlings were then transferred to red light (~30 μmol m⁻² - s⁻¹) at 24 °C for 4 days before being photographed. Hypocotyls, cotyledons, leaves, and petioles were measured using ImageJ (http://rsb.info.nih.gov/ij/).

Real-time RT-PCR

Real-time RT-PCR was performed as described in Stanga et al. (2013). Total RNA was isolated from 5 d old, red light-grown seedlings using a Spectrum™ Plant Total RNA Kit (Sigma, St. Louis, MO, USA). RNA was DNAse treated with Turbo™ DNA-free (Ambion, Austin, TX, USA), and converted to cDNA with iScript Reverse Transcriptase Supermix, which uses random hexamer and dT primers (Bio-Rad, Hercules, CA, USA). cDNAs were used as templates for real-time PCR in a Roche Light Cycler 480 using LightCycler 480 SYBR Green I Master (Roche, Carlsbad, CA, USA) with the following program: 10 min at 95 °C; 45 cycles of 20 s at 95 °C, 20 s at 60 °C, and 20 s at 72°C; followed by melt curve analysis to analyze product specificity. Primers are listed in Table S1. Crossing point (Cp) values were calculated under high confidence with Light Cycler 480 software. The average Cp of two technical replicates was used to calculate the abundance relative to the CACS reference gene for each sample, with adjustment for primer efficiencies. Relative expression values from four biological samples were used for each data point.

Software and statistical tests

Statistical tests were calculated using either the JMP statistical package (SAS, Cary, NC, USA) or Excel (Microsoft, Redmond, WA, USA). Prior to t tests, data were tested for equal variance by F tests. One-way ANOVA tests were conducted using combined data from experimental replicates. Following ANOVA, post hoc tests with Tukey–Kramer Honest Significant Difference (HSD) or Student’s t tests were conducted to assign statistical groupings. All sample sizes and significance thresholds are indicated in the figure captions.

Results

We obtained a T-DNA insertion allele of SMXL2 (At4g30350) in the Col-0 ecotype from the SAIL collection (Fig. 1a) (Sessions et al. 2002). qRT-PCR performed with primers located upstream, flanking, and downstream of the T-DNA insertion site showed that no full-length SMXL2 transcript is present in smxl2-1 mutants. However, there is residual transcription downstream of the insertion site and increased transcription upstream of the insertion site (Fig. 1b). We generated double and triple mutant combinations between smxl2-1, smax1-2, and max2-1. We hereafter refer to these alleles as smxl2, smax1, and max2.

Fig. 1.

The smxl2-1 allele. a The triangle indicates the position of SAIL_596_E08, the T-DNA insertion in At4g30350. Carats represent introns. Small arrows indicate primer pair loci for the expression analysis shown in panel b. b Abundance of SMXL2 transcript at three loci as determined by real-time RT-PCR using cDNA derived from 5-d old red light-grown seedlings. Expression values are relative to the CACS reference gene and scaled to Col-0 expression. Mean ± SE, n = 4 independent samples, >50 seedlings/sample.*P < 0.05, Student’s t test comparisons to wildtype (Col-0)

SMXL2 redundantly controls seedling growth, but not germination, with SMAX1

We first investigated whether smxl2 influences seedling photomorphogenesis. In contrast to smax1 seedlings, which have shortened hypocotyls, smxl2 were indistinguishable from wildtype under continuous red light (Fig. 2a). The smax1 smxl2 double mutant seedlings, however, showed a substantial reduction in hypocotyl elongation, indicating that smxl2 enhances the smax1 phenotype. SMAX1 is epistatic to MAX2 in hypocotyl growth, and the smax1 smxl2 double mutant also showed clear epistasis to max2.

Fig. 2.

smxl2 enhances smax1 phenotypes in seedlings. a Hypocotyl lengths of 5-d old red light-grown seedlings grown on media containing 1 μM KAR₂ or rac-GR24. Mean ± 99 % CI, n = 100 seedlings. b Hypocotyl lengths of red light-grown seedlings grown on media containing 1 μM KAR₂, GR24^ent−5DS, or GR24^5DS. Mean ± 99 % CI, n = 100 seedlings. c Cotyledon area measured in 5-d old red light-grown seedlings. Mean ± 99 % CI, n = 127–142 cotyledons. d Hypocotyl lengths of 5-d old dark-grown seedlings. Mean ± 99 % CI, n = 45 seedlings. e Hypocotyl lengths of 5-d old dark-grown seedlings treated with or without 50 μM rac-GR24. Mean ± 99 % CI, n = 60 seedlings. f Primary dormant seeds assayed for germination after 5 d at 24 °C. Mean ± SD, n = 4 independent seed batches, 75–100 seeds per sample. *P < 0.05, Student’s t test comparisons to wildtype (Col-0). Statistical groups indicated by letters were determined by ANOVA with post hoc Tukey–Kramer HSD, P < 0.01

KARs enhance seedling growth responses to light, causing reduced hypocotyl elongation and enlarged cotyledons. The commonly used synthetic SL mixture rac-GR24 also inhibits hypocotyl elongation in light-grown seedlings. Both smax1 and smxl2 mutants retain responses to KAR₂ and rac-GR24 treatments. Interestingly, smax1 smxl2 hypocotyl length was equivalent to smax1 seedlings treated with KAR₂ or rac-GR24 (Fig. 2a), and smax1 smxl2 hypocotyls were not further responsive to KAR₂ or rac-GR24 treatments. This suggested that the ability to respond to exogenous KAR or SL had already reached its maximum in smax1 smxl2.

The synthetic SL rac-GR24 is actually a mixture composed of GR24^5DS, a SL analog with a 2′R configuration consistent with all known SLs, and its enantiomer GR24^ent−5DS, which has an non-natural 2′S configuration. GR24^5DS acts through D14, but GR24^ent−5DS primarily acts through KAI2 (Scaffidi et al. 2014). As KAR signaling is KAI2-dependent, KAR₂ and the different GR24 stereoisomers can be used to separate KAI2 and D14 signaling. We found a slight reduction in hypocotyl inhibition response to GR24^5DS in smxl2 (Fig. 2b). However, smax1 smxl2 seedlings were insensitive to all treatments; therefore SMAX1 and SMXL2 appear to be the only downstream targets of both KAI2 and D14-dependent control of hypocotyl growth.

Cotyledon expansion in light-grown seedlings is reduced in max2 (Shen et al. 2007). However, smax1 mutants have larger cotyledons than wildtype, mimicking the effects of KAR treatments (Stanga et al. 2013). We did not observe a difference in smxl2 cotyledon sizes compared to wildtype Col-0 (Fig. 2c). Whereas smax1 strongly suppresses the small cotyledon phenotype of max2, smxl2 did not suppress max2. Unexpectedly, smxl2 reduced the cotyledon expansion effects of smax1, producing an intermediate phenotype. This indicates a complex relationship between SMAX1 and SMXL2 availability and cotyledon growth.

The effects of KAR and SL signaling on hypocotyl elongation are usually observed only under light-grown conditions; in the dark, max2 and kai2 seedlings have wildtype hypocotyl lengths, and 1 μM KAR or rac-GR24 treatments do not reduce hypocotyl elongation (Shen et al. 2007; Nelson et al. 2011; Jia et al. 2014). However, much higher concentrations of 50–75 μM rac-GR24 inhibit dark-grown hypocotyl elongation, and can do so in a MAX2-independent manner (Tsuchiya et al. 2010; Jia et al. 2014). The basis for this MAX2-independent response is unclear, but one possibility is that high rac-GR24 levels may drive formation of receptor-SMAX1/SMXL2 protein complexes that reduce the function of SMAX1/SMXL2 (i.e. by sequestration). If so, dark-grown smax1 and smxl2 mutants might mimic, but be insensitive to, high rac-GR24 treatment. We did not observe a reduction in hypocotyl elongation in dark-grown smax1 or smxl2 seedlings, but smax1 smxl2 and smax1 smxl2 max2 mutants indeed had a significant and equal reduction in hypocotyl length (Fig. 2 d,e). However, dark-grown smax1 smxl2 seedlings retained responses to 50 μM rac-GR24, which substantially reduced dark grown hypocotyl length. Therefore the MAX2-independent response to high rac-GR24 also appears to be SMAX1 and SMXL2-independent.

Finally, we tested whether smxl2 can suppress the enhanced seed dormancy phenotype of max2, similar to smax1. We did not observe any suppression of max2 seed dormancy by smxl2. Thus SMAX1, but not SMXL2 is important for germination control (Fig. 2f). These roles correlate with the high expression of SMAX1, and comparably low level expression of SMXL2, in Arabidopsis seed (Stanga et al. 2013).

SMXL2 redundantly regulates expression of KAR/SL markers in seedlings with SMAX1

To investigate the possibility that smax1 smxl2 seedlings have constitutive activation of the KAR/KL pathway, we used qRT-PCR to measure expression of the transcriptional markers D14-LIKE2 (DLK2), KAR-UP F-BOX1 (KUF1), and INDOLE-3-ACETIC ACID INDUCIBLE1 (IAA1) in light-grown seedlings. DLK2 and KUF1 are induced by KAR and rac-GR24 treatment in seedlings, whereas IAA1 expression is downregulated. In max2 and kai2 seedlings, DLK2 and KUF1 have reduced expression, and IAA1 is upregulated (Nelson et al. 2010, 2011; Waters et al. 2012). We did not observe a significant effect of smxl2 on the expression of these markers in the wildtype or max2 background; however, smax1 smxl2 had an ~25-fold increase in DLK2 expression compared to wildtype, as well as a significant increase in KUF1 transcripts and a decrease in IAA1 transcripts (Fig. 3). As with other seedling phenotypes, smax1 smxl2 was epistatic to max2. The expression phenotypes of these three transcriptional markers correlate with the hypocotyl elongation phenotypes, demonstrating that SMAX1 and to a lesser degree SMXL2 redundantly control seedling growth downstream of MAX2.

Fig. 3.

Expression of KAR/SL-responsive genes. Relative abundance of KUF1, IAA1, and DLK2 transcripts, determined by real-time RT-PCR using cDNA derived from 5-d old red light-grown seedlings. Expression values are relative to the CACS reference gene and normalized to Col-0. Mean ± SE, n = 4 independent samples, >50 seedlings/sample. *P < 0.05, Student’s t test comparisons to wildtype (Col-0)

SMXL2 does not control leaf morphology or shoot branching

We next investigated whether smxl2 influences max2 growth phenotypes at later stages of development. The petioles of max2 plants grown in short day conditions have a hyponastic orientation similar to kai2/htl. The max2 phenotype is qualitatively suppressed by smax1, restoring a flattened rosette, but not by smxl2 (Fig. 4a). Under short day conditions, d14 leaves have shorter petioles than wildtype (Fig. 4b,c). The petioles of max2 leaves are intermediate to d14 and kai2. Petiole length is further reduced in the smax1 max2 double mutant, recapitulating a d14 phenotype (Fig. 4c). SMXL2, however, does not make a significant contribution to petiole length.

Fig. 4.

smxl2 does not suppress max2 rosette phenotypes. a Petiole ▶ orientation of 5-week old plants grown under a short day photoperiod. b 5-week old rosettes grown under a short day photoperiod. c Petiole length, d blade length, blade width, and length:width ratio of 7th and 8th leaves from 35-d-old plants grown under a short day photoperiod. The kai2 mutant allele was htl-3. b and the top of c are composite images in which the background has been removed for clarity. Mean ± 95 % CI, n = 16 (2 leaves each from 8 plants). Statistical groups were determined by ANOVA with post hoc Student’s t test, P < 0.05

SL and KAR insensitive mutants have different leaf morphologies, and max2 leaves are a hybrid of d14 and kai2 phenotypes (Soundappan et al. 2015). The d14 mutant has shorter leaves than wildtype and a low leaf length to width ratio, while kai2 has narrower leaves (Fig. 4d). smax1 smxl2 leaves are slightly smaller than wildtype in terms of leaf length and width (Fig. 4d). Overall smax1 max2 and smax1 smxl2 max2 rosettes seem to show a qualitative shift toward d14 morphology (Fig. 4b), but we did not detect significant differences in the dimensions of the 7th and 8th leaves relative to max2 under these conditions (Fig. 4d).

Finally, we investigated whether smxl2 contributes to shoot growth. Consistent with branching control being regulated by SL and the SMXL6,7,8 clade (Soundappan et al. 2015; Wang et al. 2015), we observed no effect of smxl2 or smax1 smxl2 mutations on the rosette branching or primary inflorescence height phenotypes of max2 (Fig. 5). Taken together, these findings support the model that different SMXL family members have specialized roles in growth and development: SMXL2 contributes to KAR/KL responses alongside the primary regulator, SMAX1, whereas SMXL6, SMXL7, and SMXL8 mediate SL responses.

Fig. 5.

smxl2 does not suppress max2 shoot phenotypes. a Representative images of 46-d-old plants grown under a long day photoperiod. b Axillary branching and c primary inflorescence heights of plants grown as in a. Mean ± SD, n = 12 plants. *P < 0.05, Student’s t test comparisons to wildtype (Col-0)

Discussion

SMAX1 is an important regulator of seed germination, hypocotyl elongation, cotyledon expansion, petiole orientation, and petiole length (under short days) in the SMXL family. We did not observe any mutant phenotypes in the smxl2 single mutant, suggesting that SMXL2 is not as critical as SMAX1 for normal growth. However, we did discover that smxl2 enhances the effects of smax1 on hypocotyl growth and gene expression in seedlings, revealing functional redundancy between SMAX1 and SMXL2 at this developmental stage. Interestingly, light-grown smax1 smxl2 hypocotyls mimicked constitutive growth responses to KAR or rac-GR24, yet were insensitive to these treatments. This implies that SMAX1 and SMXL2 are the only targets of MAX2 in seedlings; i.e. activation of KAI2 or D14 signaling through KAR/SL treatments has no effect because these hypothesized proteolytic targets of MAX2 are already absent.

SMAX1 and SMXL2 are likely to act in the KAR/KL pathway specifically, as they collectively mimic KAR responses, suppress the kai2-associated phenotypes of max2, and do not influence shoot branching, which is SL-regulated. Although rac-GR24 mimics KAR responses in seeds and seedlings, this should not be taken to imply a role for endogenous SLs at these developmental stages, as its enantiomeric components can separately activate KAI2 and D14-dependent signaling (Scaffidi et al. 2014). For example, kai2 mutant seedlings respond to GR24^5DS, due to D14 activity (Scaffidi et al. 2014). However, SL-deficient max seedlings and d14 do not have hypocotyl elongation defects (Nelson et al. 2011; Shen et al. 2012). Therefore, there is a difference between the role of endogenous SLs in hypocotyl elongation and the capacity to respond to exogenous SL treatments. This is supported by the observation that smxl6,7,8 mutations do not suppress the max2 hypocotyl elongation phenotype, although smxl6,7,8 mimics constitutive SL phenotypes in shoot branching, auxin transport, PIN1 accumulation, BRC1 expression, and lateral root density (Soundappan et al. 2015). Our data show that smax1 smxl2 are insensitive to GR24^5DS, suggesting that SMAX1 and SMXL2 are negatively regulated by D14. However, it remains unclear whether signaling between D14 and SMAX1/SMXL2 in seedlings is physiologically relevant, or is crosstalk that is only forced to occur by exogenous SL treatments.

As smxl2 has no discernible phenotype on its own, it raises the question of why SMXL2 is conserved as a functional gene in the Arabidopsis genome. One possibility is that subfunctionalization of the SMXL2 and SMAX1 paralogs occurred after gene duplication, and SMXL2 is the primary signaling agent in a specific developmental or environmental context that we have not yet explored. In some cases the expression of SMXL paralogs correlates with their respective functions; among the SMXL family, SMAX1 is most highly expressed in seeds and controls germination, whereas SMXL7 is predominantly expressed in axillary branches and is the strongest contributor to branching (Stanga et al. 2013, Soundappan et al. 2015). A detailed analysis of SMXL2 expression may point to a specific tissue or process in which to test for its predominant role.

So far the SMXL family can be divided neatly into a clade that mediates KAR/KL responses (SMAX1 and SMXL2) and a clade that mediates SL responses (SMXL6, SMXL7, and SMXL8). Genetic data demonstrating the complete suppression of several KAR and SL-specific max2 phenotypes; as well as biochemical data demonstrating SL-induced physical interactions between D53/SMXL6,7,8 and D14; and D14 and MAX2-dependent proteolysis of D53/SMXL6,7,8 strongly support SMXL proteins as the bona fide targets of MAX2 (Jiang et al. 2013; Stanga et al. 2013; Zhou et al. 2013; Soundappan et al. 2015; Umehara et al. 2015; Wang et al. 2015). Future comparisons of the transcriptomes of smax1 smxl2 and smxl6,7,8 mutants may reveal unique and overlapping features of the downstream transcriptional networks of KAR/KL and SL signaling, and identify candidate genes that carry out MAX2-dependent growth responses. Other max2 phenotypes remain to be tested for links to the SMXL family, including drought tolerance, senescence, secondary growth, adventitious rooting, and root hair growth. Also, the functions of SMXL3, SMXL4, and SMXL5, and whether they too are targets of MAX2 remain important mysteries to be solved. The evolutionary history of the SMXL family will likely prove an exciting area for future investigation, as it may reveal how parallel SL and KAR/KL signaling mechanisms emerged and gained diverse roles in plant growth.

Abbreviations

KAR

Karrikin

SL

Strigolactone

KL

KAI2 ligand

SMAX1

SUPPRESSOR OF MAX2 1

SMXL

SMAX1-LIKE

D3 (14, 53)

DWARF3 (14, 53)

PIN1

PIN-FORMED1

KAI2

KARRIKIN INSENSITIVE2

MAX2

MORE AXILLARY GROWTH2

DLK2

D14-LIKE2

KUF1

KAR-UP F-BOX1

IAA1

INDOLE-3-ACETIC ACID INDUCIBLE1