Common Components of the Strigolactone and Karrikin Signaling Pathways Suppress Root Branching in Arabidopsis1[OPEN]

Dear Editor, Elaborate root branching enables plants to efficiently access resources in a heterogeneous soil matrix. In the eudicot model species Arabidopsis (Arabidopsis thaliana), root branching occurs on the primary tap root where lateral roots (LRs) emerge to provide additional surface area for nutrient uptake and anchorage to the soil. Adventitious roots (ARs) can emerge from the root–shoot junction under normal growth conditions or from the hypocotyl under stress conditions (Steffens and Rasmussen, 2016), thus providing additional root-branching capacity. Phytohormones including strigolactones (SLs) control the development of both LRs and ARs (for review, Rasmussen et al., 2013). Higher levels of SLs, achieved through application of exogenous synthetic SL (GR24rac), tend to arrest the development of LR primordia at stage V, depending on the plant’s auxin status, and lead to overall lower LR density (LRD; Ruyter-Spira et al., 2011). GR24rac also suppresses LR priming and emergence especially near the root–shoot junction (Jiang et al., 2016). Plants insensitive to SL, such as the Arabidopsis mutant more axillary growth2 (max2) that lacks an F-box Leu-rich repeat protein (which is part of a SKP1-CULLIN-F-box class of E3 ubiquitin-ligase complex and necessary for SL perception), show an increased LRD (Kapulnik et al., 2011; Ruyter-Spira et al., 2011; Soundappan et al., 2015; Villaécija-Aguilar et al., 2019). SLs were also shown to repress AR development (Rasmussen et al., 2012), because Arabidopsis max2 as well as mutants impaired in SL synthesis (max3 and max4) develop more ARs per plant than wild type. Themax2mutant has significantly more ARs compared to the two synthesis mutants (Rasmussen et al., 2012). Overall, the characterization of synthesis and perception mutants, as well as the use of GR24rac, has helped demonstrate a role for SL in suppressing root branching. Interestingly, the SL perception pathway shares elements (e.g. MAX2) with the karrikin/KAI2-Ligand (KAR/KL) pathway (Nelson et al., 2011; Waters et al., 2012, 2017). Moreover, aspects of root architecture regulation previously assigned solely to the SL pathway are now also partly assigned to the KAR/KL pathway, such as the suppression of LR development (Villaécija-Aguilar et al., 2019). Here we report that the


Dear Editor,
Elaborate root branching enables plants to efficiently access resources in a heterogeneous soil matrix. In the eudicot model species Arabidopsis (Arabidopsis thaliana), root branching occurs on the primary tap root where lateral roots (LRs) emerge to provide additional surface area for nutrient uptake and anchorage to the soil. Adventitious roots (ARs) can emerge from the root-shoot junction under normal growth conditions or from the hypocotyl under stress conditions (Steffens and Rasmussen, 2016), thus providing additional root-branching capacity. Phytohormones including strigolactones (SLs) control the development of both LRs and ARs (for review, Rasmussen et al., 2013). Higher levels of SLs, achieved through application of exogenous synthetic SL (GR24 rac ), tend to arrest the development of LR primordia at stage V, depending on the plant's auxin status, and lead to overall lower LR density (LRD; Ruyter-Spira et al., 2011). GR24 rac also suppresses LR priming and emergence especially near the root-shoot junction (Jiang et al., 2016). Plants insensitive to SL, such as the Arabidopsis mutant more axillary growth2 (max2) that lacks an F-box Leu-rich repeat protein (which is part of a SKP1-CULLIN-F-box class of E3 ubiquitin-ligase complex and necessary for SL perception), show an increased LRD (Kapulnik et al., 2011;Ruyter-Spira et al., 2011;Soundappan et al., 2015;Villaécija-Aguilar et al., 2019). SLs were also shown to repress AR development (Rasmussen et al., 2012), because Arabidopsis max2 as well as mutants impaired in SL synthesis (max3 and max4) develop more ARs per plant than wild type. The max2 mutant has significantly more ARs compared to the two synthesis mutants (Rasmussen et al., 2012). Overall, the characterization of synthesis and perception mutants, as well as the use of GR24 rac , has helped demonstrate a role for SL in suppressing root branching.
Interestingly, the SL perception pathway shares elements (e.g. MAX2) with the karrikin/KAI2-Ligand (KAR/KL) pathway (Nelson et al., 2011;Waters et al., 2012Waters et al., , 2017. Moreover, aspects of root architecture regulation previously assigned solely to the SL pathway are now also partly assigned to the KAR/KL pathway, such as the suppression of LR development (Villaécija-Aguilar et al., 2019). Here we report that the Figure 1. KAI2 and MAX2 prevent JR development under normal growth conditions. A to D, Data are shown as the percentage of 9-d-old plants showing 0, 1, 2, or 3 JRs, for n . 47 plants per genotype combined from at least three experiments. In C, lines 12H and 10G are independent lines of the complemented kai2-2 mutant (Waters et al., 2015b). Plants were grown on 0.8% (w/v) agar, supplemented with onehalf strength Murashige and Skoog salt, as reported in Swarbreck et al. (2019). Symbols indicate statistical significance in the distribution was tested using the Fisher's exact test ($P , 0.1, *P , 0.05, and ***P , 0.001). n.s., No significance.
suppression of the development of junction roots (JRs), a type of AR (Rasmussen et al., 2012), can also be partly attributed to the KAR/KL pathway, supporting the idea that there is some overlap in SL and KAR/KL pathways at least in terms of the regulation of root architecture.
KARs are small butenolide compounds found in smoke that can trigger germination in fire-follower as well as non-fire follower plant species such as Arabidopsis, while KLs are endogenous compounds yet to be identified that are hypothesized to trigger the KAR perception pathway (Conn and Nelson, 2016). Besides germination, KAR/KLs have been implicated in the regulation of root development, especially LR formation (Villaécija-Aguilar et al., 2019) and root skewing (Swarbreck et al., 2019). KAI2 was identified as a receptor for KAR/KL (Waters et al., 2012); it is an a/b hydrolase that belongs to the same family as D14, the SL receptor (Hamiaux et al., 2012;Chevalier et al., 2014;de Saint Germain et al., 2016;Yao et al., 2016). In the presence of SL, D14 forms a complex with MAX2 (Zhao et al., 2015;Yao et al., 2016), which is associated with suppressor proteins of the SUPPRESSOR OF MAX2 1 (SMAX1)/SUPPRESSOR OF MAX2 1-LIKE (SMXL) family (Stanga et al., 2013;Soundappan et al., 2015;Wang et al., 2015). The receptor and suppressor complex can be ubiquitinated then degraded through the 26S proteasome (Chevalier et al., 2014;Hu et al., 2017). Similarly, the presence of a KAR surrogate induces the formation of a complex among KAI2, MAX2, and a member of the SMXL family (SMXL2; Wang et al., 2020). KAI2 degradation was previously shown to happen independently from MAX2 (Waters et al., 2015a). max2 mutants are insensitive to both SL and KAR/KL and display phenotypes that can be attributed to both, while d14 mutants are specifically insensitive to SL and kai2 mutants to KAR/KL. Arabidopsis plants grown for 9 d on the surface of agar (0.8% [w/v], supplemented with one-half strength Murashige and Skoog salt as reported in Swarbreck et al., 2019), develop JRs at the root-shoot junction. These JRs were counted on Arabidopsis mutants lacking specific proteins involved in the SL and/or KARs perception pathways (growth conditions and imaging methods were as reported in Swarbreck et al., 2019). Wild-type and mutant Arabidopsis plants (lines described in Swarbreck et al., 2019, except for the kai2d14 double mutant, a gift from Dr. Mark Waters, The University of Western Australia; Waters et al., 2015b) tended to have only a few JRs per plant, and many plants had no JRs. In the Columbia (Col) ecotype background, a maximum of one JR per plant was observed (Fig. 1). Accordingly, we report the percentage of plants that have developed a JR rather than calculating an average JR number per plant. A high percentage of plants showing no JR would be consistent with a low average AR in other reports.
A higher proportion of max2 plants had a JR (9.4%, Fisher's exact test two-sided, P , 0.001) compared to Col (1.3%). A slightly higher proportion of d14 plants also had a JR (5.4%, Fisher's exact test two-sided, P 5 0.054) compared to Col (1.3%). However, a significantly higher proportion of max2 plants had a JR compared to d14 (Fisher's exact test two-sided, P , 0.001). Our data demonstrate that MAX2, and to some extent D14, suppresses the development of JR in Arabidopsis. These results are consistent with the significant increase in average number of etiolation-induced hypocotyl ARs in max2, as reported in Rasmussen et al. (2012).
To test whether the MAX2-dependent suppression of JR development was solely due to SL perception, we also characterized the phenotype of mutants impaired in the KAR/KL perception pathway. In the Landsburg erecta (Ler) background, a higher proportion of both max2-8 and kai2-2 plants have one or more JRs compared to the wild type (Fisher's exact test, P , 0.001; Fig. 1B). Given that the phenotype of the double mutant kai2-2 max2-8 matches that of both  Figure 1 and as reported in Swarbreck et al. (2019), except for C, where agar was supplied from Melford. Statistical significance was assessed using ANOVA and Tukey Test as post hoc test, and differences between mutants and wild type are indicated by asterisks (*P , 0.05 and ***P , 0.001). n.s., No significance. single mutants (Fig. 1B), we can conclude that both KAI2 and MAX2 function in the same genetic pathway to suppress JR development in Arabidopsis. Using transformed kai2-2 mutant lines, we could also demonstrate the specific effect of KAI2 on JR development ( Fig. 1C; we discuss below the JR phenotype of kai2-1, which is different from kai2-2). In a separate experiment, we counted the JR in the kai2d14 double mutant to estimate whether the phenotype of max2 could simply be attributed to signaling through the SL and KL pathway or whether additional pathways were implicated. We found no significant difference in the proportion of kai2d14 plants presenting no JR (64.3%) compared to max2-8 (52.3%, Fisher's exact test two-sided, P 5 0.7, with n . 23 plants in three independent experiments).
The degradation targets for the complex involving MAX2 also provide an additional point where the SL and KAR/KL pathways share elements. In Arabidopsis, the SMXL family includes eight members (Soundappan et al., 2015). While SMAX1/SMXL2 have been implicated in the KAR/KL perception, SMXL6, SMXL7, and SMXL8 tend to be degraded in response to SL perception (Soundappan et al., 2015). However, there are some elements of overlap, because the root-skewing phenotype can be suppressed by both SMAX1/SMXL2 as well as SMXL6, SMXL7, and SMXL8 (Swarbreck et al., 2019;Villaécija-Aguilar et al., 2019). More recently, SMXL2 ubiquitination and degradation was shown to be induced by both SL and KL perception (Wang et al., 2020). Here we show that SMAX1, as well as SMXL6, SMXL7, and SMXL8, suppresses the JR phenotype of max2 mutant plants, because smax1 max2 and smxl6, smxl7, and smxl8 max2 mutants have a similar proportion of JRs as wild type (Fig. 1D). Overall, our data show that under normal growth conditions, the KAI2/MAX2/SMXL signaling pathway suppresses the development of JRs in Arabidopsis.
SLs were shown to halt LR development in Arabidopsis (Ruyter-Spira et al., 2011), although recently Villaécija-Aguilar et al. (2019 suggested that the increased LRD phenotype of max2 mutants could be attributed to the impairment in the KAR/KL perception pathway. In the Col background, we observe an increased LRD in the max2 mutant ( Fig. 2A; Tukey Honestly Significant Difference [HSD] test, P , 0.001), in agreement with reports from the literature. We also report an increased LRD in the d14 mutant compared to Col ( Fig. 2A; Tukey HSD test, P , 0.001). However, similarly to the JR phenotype, max2 mutants show higher LRD compared to d14 ( Fig. 2A; Tukey HSD test, P , 0.001). We also measured an increased LRD in kai2-2 in Col ( Fig. 2A; Tukey HSD test, P , 0.001). In the Ler background, LRD is increased for max2 and kai2-2, with no further increased LRD in the kai2 max2 double mutant (Fig. 2B). These data suggest that KAI2 and MAX2 function in the same genetic pathway to regulate LRD.
Notably, in the Ler background two alleles of kai2 had initially been characterized; kai2-1 carries a mutation leading to a modification of a conserved Gly residue to Glu, while kai2-2 carries a Ds transposon insertion in an intron (Waters et al., 2012). The mutations were allelic, and no KAI2 protein could be detected in either mutant (Waters et al., 2015b). We report here that while kai2-2 shows an increased LRD compared to wild type (Fig. 2, B and C), this is not the case for kai2-1 (Fig. 2C). There was no significant difference in primary root length (Ler, 27.56 6 0.35 mm; kai2-1, 21.98 6 0.4 mm; kai2-2, 22.79 6 0.52 mm). We note that kai2-1, contrary to kai2-2, also shows no differences in JR development compared to the wild-type Ler (proportion of plants with no Figure 3. Hypothetical model of the complexes involved in SL and KAR/KL signaling and regulation of root system architecture, particularly suppression of JRs and LR formation. The perception of SL through D14/MAX2/SMXL6, SMXL7, and SMXL8 can lead to modification of the shoot architecture, secondary growth, and leaf morphogenesis. The perception of K (i.e. KAR or KL) regulates germination, photomorphogenesis, leaf morphogenesis, arbuscular mycorrhiza (AM) symbiosis, and root hair development. In addition, the careful analysis of root skewing (Swarbreck et al., 2019;Villaécija-Aguilar et al., 2019) suggest that an alternative complex may form that involves MAX2/KAI2 and SMXL6, SMXL7, and SMXL8, which could also regulate LR development (see Fig. 2; Villaécija-Aguilar et al., 2019) and JR (see Fig. 1). Further experiments are necessary to confirm that these interactions do occur at the protein level. JR, 90% for Ler, 88.7% for kai2-1; proportion of plants with 1 JR 10% for Ler, 11.3% for kai2-1, Ler, n 5 150 plants, kai2-1, n 5 133 plants, Fisher's exact test, P 5 0.85). We measured LRD in complemented kai2-2 mutant lines expressing KAI2 under the native promoter and showed that LRD tended to be lower in the transformed lines compared to the mutant (Fig. 2D). However, the LRD of transformed lines was also significantly higher compared to the wild type (Tukey HSD test, P , 0.001). Our data showing an increased LRD in kai2-2 in two different ecotypes support the findings of Villaécija-Aguilar et al. (2019), who reported an increased LRD in two kai2 mutants in the Col background. At this point, it is not clear why kai2-2 but not kai2-1 showed an increased LRD, given that the KAI2 protein is absent in both mutants (Waters et al., 2015b). Waters et al. (2012) also showed that kai2-1 had a slightly greater germination rate compared to kai2-2, but for other root phenotypes, kai2-1 and kai2-2 are similar (Swarbreck et al., 2019). Perhaps kai2-1 is not a null allele and can produce sufficient KAI2 protein to show a different phenotype compared to kai2-2, albeit this is not detected in western blot.
Clearly there is more overlap between the SL and KAR/KL signaling pathways than previously thought, and these mostly relate to some of the root phenotypes observed. To add to the complexity, roots are highly responsive to small changes in environmental conditions and establishing a clear root system architecture phenotype for a given mutant, even when grown under controlled conditions, often requires characterizing a high number of plants. The root skewing phenotype is even more sensitive to small changes, with skew values reported for Col that vary greatly (from 5°to .15°toward the right, when measured from the back of the plate). In trying to determine which of the SMAX1/ SMXL suppressors was also involved in root skewing regulation together with KAI2 and MAX2, we previously showed that mutants deficient in SMAX1 and SMXL2, as well as mutants in SMXL6, SMXL7, and SMXL8 had a reduced rightward skewing angle compared to the wild type (Swarbreck et al., 2019). Interestingly, Villaécija-Aguilar et al. (2019) could find a similar phenotype but only in one laboratory setting, not in the two that were tested. Together, these data ( Fig. 1; Swarbreck et al., 2019;Villaécija-Aguilar et al., 2019) do not support the strict dichotomy that has been proposed previously, i.e. that only the MAX2/KAI2/ SMAX1,SMXL2 complex is relevant to KAR/KL signaling while the MAX2/D14/SMXL6, SMXL7, and SMXL8 complex is relevant to the SL perception pathway (Fig. 3). Our data support a third hypothetical complex MAX2/KAI2/SMXL6, SMXL7, and SMXL8 that is involved in the regulation of root skewing and JR development (Fig. 3). A more varied profile of complexes with the SMAX1/SMXL target proteins would help ensure a more appropriate response in Arabidopsis root system architecture to a highly heterogeneous soil environment.