Strigolactones inhibit auxin feedback on PIN-dependent auxin transport canalization

Directional transport of the phytohormone auxin is a versatile, plant-specific mechanism regulating many aspects of plant development. The recently identified plant hormones, strigolactones (SLs), are implicated in many plant traits; among others, they modify the phenotypic output of PIN-FORMED (PIN) auxin transporters for fine-tuning of growth and developmental responses. Here, we show in pea and Arabidopsis that SLs target processes dependent on the canalization of auxin flow, which involves auxin feedback on PIN subcellular distribution. D14 receptor- and MAX2 F-box-mediated SL signaling inhibits the formation of auxin-conducting channels after wounding or from artificial auxin sources, during vasculature de novo formation and regeneration. At the cellular level, SLs interfere with auxin effects on PIN polar targeting, constitutive PIN trafficking as well as clathrin-mediated endocytosis. Our results identify a non-transcriptional mechanism of SL action, uncoupling auxin feedback on PIN polarity and trafficking, thereby regulating vascular tissue formation and regeneration.

P lant development is characterized by self-organizing processes, such as the regular patterns of organ initiation at the shoot apical meristem, branching of roots and shoots, the connection of newly formed organs with pre-existing vasculature, or the spontaneous occurrence of vasculature veins in developing leaves. The plant hormone auxin and its directional transport through tissues have been implicated in all these traits 1 . The process that is called auxin canalization establishes narrow auxin transport routes between cells and tissues of relatively high auxin concentration (source), to locations where auxin is being depleted (sink) [2][3][4] . A self-reinforcing system has been proposed to drive canalization. In this system, auxin feeds back on PIN-FORMED (PIN) auxin transporters by promoting the expression of PIN genes specifically in auxin transport routes and by localizing PINs to plasma membrane (PM) domains facing the auxin sink 3,5 . Auxin is typically transported basipetally from source to sink and canalization seems to be driven by auxin sink rather than source [6][7][8] . For example, developing vasculature tissue is characterized by relatively high auxin contents; therefore, sink strength in such a system primarily depends on PIN-dependent auxin flux rates depleting auxin from sources. New vein patterns in leaves develop away from a localized sink at the base, opposite to the direction of auxin flow [6][7][8] . Vasculature formation and its connection to already existing vascular strands hence is intimately linked to the effects of auxin flux on the subcellular positioning of PINs, which in turn defines auxin flux rates and directionality [6][7][8] . Related feedback mechanisms controlling PIN polarity have been described for additional developmental processes, such as embryonic axis formation 9,10 , shoot and root organogenesis 11,12 as well as the control of the directional growth of organs 13 .
In spite of the biological significance of PIN proteins function, the mechanisms by which auxin controls polarization of PINs have remained conceptually unclear. Modeling of auxin-mediated polarization 14 has linked auxin feedback on PIN polarity to the auxin effect on PIN subcellular trafficking [15][16][17] . PM-associated PINs are internalized in clathrin-coated vesicles in a process called endocytosis [18][19][20] that might precede PIN relocation to different (plasma) membrane domains 21,22 . Numerous pharmacological and genetic determinants that impact on specific cellular events in the control of auxin transport have shaped our current picture of cellular mechanisms and crosstalk therein 1 . For example, treatment with the fungal toxin Brefeldin A (BFA) revealed the constitutive endocytic recycling of PIN proteins. This requires BFA-mediated interference of GNOM ARF-GEF activity, causing PM proteins, including PINs, to aggregate in cells 18,20 . Notably, auxin itself appears to inhibit the process of endocytosis and antagonize the BFA effect on PIN recycling 15 . In addition to auxin effects on PINs, various other plant hormones can influence PIN-dependent auxin transport, such as cytokinins 23 , brassinosteroids 24,25 , gibberellins 26 , salicylic acid 27 , abscisic acid 28 , and strigolactones (SLs) 29 . However, much remains to be uncovered about the modes of action, by which these plant hormones regulate PINs, thereby ultimately defining plant development.
SLs represent a recently discovered class of plant growth regulators and their developmental roles and signaling mechanisms are not yet fully characterized. SLs have been shown to influence a range of plant traits including shoot branching 30,31 , shoot gravitropism 32 , secondary growth 33 , adventitious rooting 34 as well as lateral rooting and root hair elongation 35,36 . Many of the processes targeted by SLs also require auxin transport, specifically its canalization as proposed for the classical SL effect on shoot branching 37 . This response has been linked to SL-mediated interference of PIN PM targeting and the modulation of auxin flux 29,37 , but we still lack insight into the mechanisms by which SLs might impact the sorting of PINs.
Our observations in this study extend the spectrum of physiological SL effects on processes associated with auxin transport canalization, namely leaf venation and vascular tissue regeneration and formation induced by wounding or external auxin sources. At the cellular level, we show that SLs specifically interfere with the feedback of auxin on PIN polarization and clathrin-mediated internalization, providing a mechanistic framework for the molecular action of SLs in many developmental processes.

Results
SLs interfere with auxin canalization in pea. Inhibition of shoot branching is among the best-understood responses of SLs in flowering plants. This is a process that involves auxin canalization, because when buds are released from dormancy, they initiate the formation of PIN1-expressing channels to increase vascular connections with the main vasculature 5 . These channels appear similar to those that form after adding exogenous auxin to the side of the stem 3,5 . These canalization events can be inhibited by auxin produced in shoot apices, hinting at a possible mechanism, by which dominant shoot tips might control branching 5 . When SLs are applied directly to buds after decapitation, they inhibit bud outgrowth 38 and reduce the transport of auxin (indole-3acetic acid, [ 3 H]-IAA) from buds into the stem (Supplementary Fig. 1a-d). However, the precise action of SLs in controlling auxin canalization and vascularization is less obvious 39 . Therefore, we explored the effect of synthetic SL, rac-GR24 (hereafter called GR24) using intact or fully decapitated pea (Pisum sativum) plants that had been treated with auxin (indole-3-acetic acid; IAA).
First, we analyzed PIN1 channel and subsequent vasculature formation originating from an artificial lateral auxin source. Lateral, local auxin application in lanolin paste onto pea stems just below a cut (Fig. 1a) was sufficient to induce the formation of PIN1-expressing auxin channels and subsequent vascular connections to the stem vasculature 4,40 . In our control situation, strong PIN1 expression in the vicinity of the local IAA application was observed with a predominantly lateral PIN1 localization, pointing away from the auxin source (Fig. 1b). An initially large area of PIN1-expressing cells narrowed down about 5 days after auxin application, resulting in the establishment of fully defined and strongly polarized narrow PIN1 channels, often accompanied by differentiated xylem vessels (Fig. 1b). This is in agreement with the classical canalization hypothesis in the absence of competing auxin sources 2 . In contrast, coapplication of GR24 interfered with strongly polarized PIN1 expression as well as with the formation of PIN1 channels and continuous de novo vasculature; only occasional fragmented xylem cells appeared instead (Fig. 1b).
Related observations were made when we analyzed PIN1 expression in fully decapitated pea stems. IAA application to the stump (Fig. 1c) led to a massive increase of PIN1 expression in the polarized field below the application site within the first 3 days, while the formation of narrow PIN1-expressing channels accompanied with differentiated xylem strands became visible after 5 days (Fig. 1d). The simultaneous application of GR24 strongly inhibited this process, as we failed to observe a pronounced increase in polarized PIN1 expression and channel formation under these conditions (Fig. 1d).
Together, our findings suggest an inhibitory role of SLs in the formation of new auxin-conducting, PIN-expressing channels induced from auxin sources. This effect of SLs on auxin canalization would offer a plausible explanation for how SLs regulate auxin transport, vascularization and branching.
SLs inhibit vasculature formation and regeneration. To further explore the role of SLs in other processes that have been mechanistically linked to canalization, we examined canalizationdependent vasculature regeneration following wounding 2 , which has recently been established in Arabidopsis (Arabidopsis thaliana) stems 41 (Fig. 2a). This allowed us to use the extensive genetic toolkit in the model species, and also allowed us to test plantproduced, endogenous SLs rather than relying on synthetic SLs.
In control experiments, we observed vascular regeneration initiated with a broad PIN1 expression field and auxin accumulation above the wound about 2 days after wounding. This was subsequently followed by the establishment of narrow, polarized PIN1-expressing, auxin-conducting channels circumventing the wounded site 41 during the next days ( Supplementary  Fig. 2a). Strikingly, vasculature regeneration in the SL biosynthesis mutants more axillary growth (max)1-1 and max4-1 occurred as fast as in control ( Supplementary Fig. 2b), but the regenerated vasculature in the mutants was more abundant than that in the wild-type control (Fig. 2b). To test the effect of increased endogenous SLs, we employed conditional overexpression of the cytochrome P450 monooxygenase MAX1 in the max1 mutant background (DEX≫MAX1 max1-1; hereafter named DEX≫MAX1). In un-induced control plants, the first vessels around the wound appeared after 4 days and fully regenerated vasculature was observed after 6 days (Fig. 2c, Supplementary  Fig. 2c). In contrast, although we detected clusters of isolated vessel-like cells that developed from callus in both untreated and dexamethasone (Dex)-treated plants (Fig. 2c), there was no regeneration of vasculature around the wound observed after Dex induction (Fig. 2c, Supplementary Fig. 2c). These results are in line with the observations that we made in pea, substantiating an inhibitory role for SLs in the regulation of canalization-mediated vasculature regeneration.
We also determined the efficiency of vasculature regeneration in mutants affected in SL/karrikin-related signaling, including a mutant allele affected in the MAX2 F-box protein (max2-3) and a double mutant affected in SL/karrikin receptors dwarf14-1 hyposensitive to light-3 (d14-1 htl-3) 42 . In both genotypes, regeneration occurred faster (Fig. 2d) and the regenerated strands were more abundant compared with wild-type control (Supplementary Fig. 2d), suggesting that SL/karrikin signaling normally restricts vasculature regeneration.
To directly assess whether SL/karrikin signaling is involved in auxin channel formation, we analyzed the expression of the DR5 auxin response reporter (DR5rev::GFP) during regeneration. Comparison between wild type and max2-3 revealed that the DR5-positive channels formed faster and more abundantly when SL/karrikin signaling was compromised (Fig. 2e, f). Consistently, the layer of regenerated vasculature was also formed earlier and thicker in the max2-3 mutant (Fig. 2e, f).
Together, these results identify SLs as crucial regulators of vasculature regeneration after wounding, and that increased SL levels inhibit, whereas decreased SL biosynthesis or compromised SL signaling promotes, canalization-mediated vasculature regeneration. Another presumably auxin canalization-dependent process that involves vasculature patterning along auxin channels is de novo leaf venation formation 2,6,8,43 . We questioned whether SLs might participate in this process as well, and thus examined leaf vascular development in presence of GR24 or upon induction of endogenous SL biosynthesis. After growth on GR24, simplified leaf vascular network patterns with occasional discontinuities were detected ( Supplementary Fig. 3a, b). Dex-treated DEX≫MAX1 plants also caused more simplified leaf veins with more free ends ( Supplementary Fig. 3c, d). This two-component glucocorticoid system can occasionally cause non-specific growth defects 44 . However, our Dex-treated transgenic plants expressing only the chimeric GAL4-VP16-GR (GVG) transcription factor grew normally and leaf vasculature was unaffected (Dex: 11.1 freeending veins per leaf, n = 20 leaves; Control: 10.7 free-ending veins per leaf, n = 20 leaves). Therefore, these data support the Red arrows crossed with black X represent inhibited auxin flow. Lanolin pastes containing IAA or IAA/GR24 (marked in yellow) were applied to the side of the pea stem below the wound. Dashed red arrow crossed with dashed black X represents formation of a new auxin flow route from the auxin source that can be inhibited by GR24 application. b Immunolocalization of PIN1 in the primary stem. The concentrations of IAA and GR24 applied locally in lanolin pastes were 0.16 and 0.09 µM, respectively. In total, 10 plants were analyzed for each treatment. White rectangles indicate the places of IAA or IAA/GR24 application. Arrowheads indicate polarity of the PIN localization. Arrows indicate newly formed auxin channels. The fluorescence signals were evaluated on Olympus Fluoview 200 confocal scanning microscope with UPlanFI 20×/0.5 and/or UPlanApo 10×/ 0.40 objectives. PIN1 immunolocalization signals (red) are overlaid with the transmitted light images. Scale bar, 100 μm. c Scheme of decapitated plants treated with IAA or IAA/GR24 paste on the stump. Red arrows represent auxin (IAA) flow. Red arrow crossed with dashed black X represents formation of a new auxin flow route from the auxin application to the stump that is inhibited by GR24 application. Dashed red arrow crossed with dashed black X represents intermitted auxin flow after temporary bud activation. Dashed green arrow represents temporary bud outgrowth. Red arrows crossed with black X represent inhibited auxin flow. d Immunolocalization of PIN1 in the stem of decapitated plants. In total, 10 plants were analyzed for each treatment. Arrowheads indicate polarity of the PIN localization. Arrows indicate newly formed auxin channels. The fluorescence signals were evaluated on Olympus Fluoview 200 confocal scanning microscope with UPlanFI 20×/0.5 objective. PIN1 immunolocalization signals (red) are overlaid with the transmitted light images. Scale bar, 100 μm. The above experiments were repeated three times with similar results. Images shown are representative of each treatment. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-17252-y ARTICLE notion that SLs regulate vasculature regeneration as well as de novo formation during venation patterning in leaves.
SLs interfere with auxin-mediated PIN polarization. The mechanism by which a local auxin source promotes the formation of auxin channels and vascularization is largely unknown. The classical canalization hypothesis proposes positive auxin feedback on auxin transport directionality 14 , which can be realized at the cellular level by the effect of auxin on PIN polar distribution. This can be visualized by auxin-mediated PIN polarity rearrangements in Arabidopsis roots 4,45 . In primary roots, PIN2 localizes to the apical side of epidermal cells, and preferentially to the basal cell side in the young cortex cells 22 . Auxin (synthetic 1naphthaleneacetic acid; NAA or natural; IAA) treatments led to rearrangement in PIN2 distribution to the outer lateral sides of cortex cells 4 (Fig. 3a-c). This PIN lateralization effect of auxin was consistently attenuated by induction of SL biosynthesis in the DEX≫MAX1 line (Fig. 3a, b), and by NAA/GR24 co-treatment in wild-type seedlings (Fig. 3c). In addition, we tested the effect of a natural SL, (+)-5-deoxystrigol (5DS), on NAA-mediated PIN2 lateralization and observed an inhibitory response similar to that of GR24 ( Supplementary Fig. 6c). In contrast, the GR24 effect on auxin-mediated PIN2 lateralization was diminished in max2-3 (Fig. 3d).
These observations suggest that SLs, acting via a MAX2dependent signaling pathway, not only negatively regulate the canalization processes at the organ and tissue levels, but also auxin-mediated polarization of PIN transporters at the level of individual cells.
SLs interfere with auxin effect on PIN endocytic recycling. How auxin can regulate PIN polarity and, in particular, how a localized auxin source can lead to the coordinated polarity changes in a whole field of cells, is conceptually unclear. Modeling of canalization and PIN polarization suggests that auxin feedback on PIN polarity can be related to the known inhibitory auxin effect on PIN internalization in individual cells 14 . PIN proteins are known to constitutively cycle between the PM and the endosomes 18,46 .   [48][49][50] , and intracellular PIN accumulation is rapidly and transiently inhibited by auxin itself 15 .
We investigated the SL effect on auxin-mediated inhibition of PIN endocytic recycling. As shown previously 15 , PIN proteins accumulated intracellularly after BFA treatment and such internalization was inhibited by NAA (Fig. 3e). GR24 treatment showed no effect on BFA-induced PIN intracellular accumulation ( Supplementary Fig. 4a, b)  increased accumulation of PIN1 and PIN2 in BFA-induced compartments, upon co-treatment with NAA/GR24 (Fig. 3e, f,  Supplementary Fig. 4c-e). Similarly, 5DS also interfered with the auxin effect on the BFA-induced PIN intracellular accumulation (Fig. 3h, Supplementary Fig. 4f).
Note that some of these short-term pharmacological experiments required high concentrations of rac-GR24. High GR24 concentrations may impact on photoreceptor pathways 51 and the use of rac-GR24 may lead to non-SL responses due to stereoisomer specificity 52 . However, 5 μM GR24 or higher can be required to trigger responses, particularly in roots 37,53-55 . We aimed to resolve these issues by testing the transgenic line DEX≫MAX1 that stimulates endogenous SL biosynthesis 29 , and also comparing GR24 treatment responses with SL mutants. The same antagonistic SL effect on auxin-mediated inhibition of PIN internalization was observed in Dex-treated DEX≫MAX1 line ( Supplementary Fig. 4g-k). Furthermore, we tested whether the effect of SLs on PIN trafficking depends on SL signaling components. In the absence of GR24, BFA-induced PIN internalization or NAA-mediated inhibition was similar in the max2 SL/karrikin signaling mutant or the d14 SL-specific signaling mutant as that of the wild type ( Fig. 3e-g, Supplementary Fig. 4d, e). Importantly, these mutants showed less sensitivity to GR24 in counteracting the NAA action on PIN endocytic trafficking ( Fig. 3e-g, Supplementary Fig. 4d, e), which appeared consistent with the results from other pharmacological studies.
In summary, our findings imply that synthetic or endogenous SLs interfere with the antagonistic auxin effect on BFA-induced intracellular accumulation of PINs, by acting via D14-and MAX2-dependent SL signaling.
SLs interfere with auxin effect on endocytosis. PIN proteins are internalized by clathrin-mediated endocytosis (CME) 19 and this endocytic pathway is inhibited by auxin through a TIR1independent mechanism 16 . Notably, in shoots, SL action has been linked to clathrin-mediated PIN internalization, acting independently of de novo protein synthesis 37 . To gain further insights into the mode of SL action in uncoupling auxin feedback on PIN internalization, we asked whether SLs specifically interfere with the auxin effect on CME in roots. Quantitative evaluation of uptake of the fluorescent endocytic tracer FM4-64 56 revealed that, as for BFA-induced PIN internalization ( Supplementary Fig. 4a,  b), FM4-64 uptake itself was not influenced by GR24 treatment (Fig. 4c). However, in contrast, NAA-mediated inhibition of FM4-64 uptake was clearly suppressed in response to GR24 (Fig. 4a, c).
Auxin inhibition of endocytosis coincides with auxin depleting clathrin from the PM. A clathrin light chain fused to GFP reporter protein (CLC-GFP) was found associated with intracellular endomembranes as well as with the PM 57 , and NAA treatment caused a decrease of clathrin-associated fluorescence preferentially at the PM 16 . We then tested the consequence of GR24 treatment, which revealed no visible effect on CLC-GFP localization (Fig. 4d). However, upon co-incubation with NAA, GR24 counteracted NAA-mediated depletion of the PM-localized CLC-GFP, reflected in an efficient recovery of CLC-GFP signal at the PM (Fig. 4b, d). Thus, while SLs alone do not have an obvious effect on PIN internalization, endocytosis or clathrin association with the PM, they specifically interfere with auxin effects on these processes.
Notably, the GR24 effect on auxin-inhibited BFA-induced PIN internalization and FM4-64 uptake could also be observed upon inhibition of de novo protein synthesis (Fig. 4e, Supplementary  Fig. 5a-d). We further obtained comparable results that affecting transcription by cordycepin treatments did not influence the ability of GR24, as it was still effective enough to counteract the NAA effect on FM4-64 uptake (Supplementary Fig. 5e) and CLC-GFP depletion from the PM (Supplementary Fig. 5f). Given the involvement of protein ubiquitination in the control of PIN2 endocytic sorting 58 , we determined the 5DS effect on NAAregulated PIN2 recycling in presence of avadomide and carfilzomib, targeting ubiquitination and proteasome activity, respectively. Both drugs antagonized 5DS effects on NAAinduced inhibition of PIN2 internalization ( Supplementary  Fig. 5g), highlighting an involvement of protein ubiquitination and proteasome activity in the transmission of the SL effect on the control of auxin feedback on protein endocytosis.
Together, these results are in agreement with, and further extend, previous findings for SL effects on PIN internalization in shoots 37 , as they establish a non-transcriptional SL effect modulating auxin-dependent control of PIN internalization by CME in roots.
Protein trafficking is linked with auxin canalization. We then asked if the observed SL cellular effects are functionally related with SL regulation of auxin canalization. For this purpose, we tested various mutants defective in CME (chc2-1 and chc2-2 alleles) or the auxin effect on PIN trafficking (doc1 and big). Indeed, all these mutants interfered with developmental processes involving canalization, such as leaf venation patterning (Supplementary Fig. 5h), vascular tissue regeneration and formation induced by wounding or external auxin sources 59 . These data strengthen the previously established connection 59 between subcellular trafficking and their dynamic auxin regulations at the level of individual cells and long-term auxin-induced canalization processes across tissues. Immunolocalization of PIN2 is shown (a). Arrowheads indicate the polarity of PIN localization. Scale bars: 5 µm. Ratio between mean fluorescence intensity of the lateral and basal membrane in young cortex cells was scored (b n ≥ 34 cells). c Exogenous SL effect on auxin-mediated PIN2 polarity changes in young cortex cells. Ratio between mean fluorescence intensity of the lateral and basal membrane in young cortex cells was scored (c n ≥ 70 cells). Data are expressed as mean ± s.e.m. Means with different letters are significantly different at P < 0.05 (one-way ANOVA with Fisher LSD test) (b, c). d Less pronounced effect in max2-3 SL/karrikin signaling mutant in terms of GR24 inhibition of NAA action on PIN2 lateralization (d n ≥ 30 cells). Data are expressed as mean ± s.e.m. P value was calculated using Welch's two-tailed t-test. e, f GR24 effect on NAA inhibition of BFA-induced PIN1 internalization. Immunolocalization of PIN1 in root cells is shown (e). Arrowheads indicate PIN1 proteins internalized into BFA compartments. Scale bars: 5 µm. The number of BFA bodies per root cell in NAA/BFA-or NAA/GR24/BFA-treated wild-type and max2 seedlings was scored (f n ≥ 425 cells). Data are expressed as mean ± s.e.m. Means with different letters are significantly different at P < 0.05 (one-way ANOVA with Fisher LSD test). g, h Quantification of PIN1 internalization in roots. Both synthetic SL GR24 (25 μM, g; 50 μM, h) and natural SL 5DS (50 μM, h) were applied. The same position of root tip was imaged with the same microscope settings for each independent experiment. The roots (n ≥ 5) were then scored blind and the percentage of roots displaying almost undetectable (Score 0), weak (Score 1), stronger (Score 2), or very severe (Score 3) PIN1 internalization was determined. The above experiments were repeated three times with similar results. Images shown are representative of each treatment. Source data of b-d and f-h are provided in the Source data file.
SL/karrikin signaling interferes with auxin feedback. SLs and karrikins act via MAX2-dependent signaling, and we therefore questioned whether both signals exert long-term effects on vasculature patterning in response to external auxin sources (droplets of auxin in lanolin). IAA application promoted the formation of PIN1-mediated auxin channels and regeneration of vascular strands in Arabidopsis, whereas GR24 treatment alone had no influence neither on auxin channel nor on vascular tissue formation ( Supplementary Fig. 6a, b). Although SL signaling is essential for lateral bud development, in this experimental set-up, regardless of the presence or absence of lateral buds, GR24 effectively inhibited IAA-induced formation of PIN1-GFP channels and vascular strands from the position where IAA and GR24 were co-applied ( Supplementary Fig. 6a, b). Notably, antagonistic effects on auxin channel and vasculature formation were also observed in response to karrikinolide (KAR 1 ; Supplementary  Fig. 6a, b).
Moreover, we also examined short-term effects of KAR 1 on auxin feedback at the subcellular level. As in case with SLs, KAR 1 treatment antagonized NAA-mediated PIN2 lateralization (Supplementary Fig. 6c), and attenuated the NAA effect on BFA-induced PIN2 intracellular accumulation as well (Supplementary Fig. 6d).
Taken together, these data suggest that MAX2-dependent SL and karrikin signaling interfere with auxin feedback control of canalization at the tissue level as well as PIN polarity and trafficking at the cellular level.

Discussion
Our observations extend the current knowledge about developmental processes regulated by SLs and provide insights into the NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-17252-y ARTICLE NATURE COMMUNICATIONS | (2020) 11:3508 | https://doi.org/10.1038/s41467-020-17252-y | www.nature.com/naturecommunications cellular mechanism of SL action. We show that SLs negatively regulate vascularization of leaves, vasculature regeneration after wounding as well as de novo formation of vasculature from artificial exogenous auxin sources. These processes, together with well-documented effects of SLs on shoot branching, are thought to at least partially depend on the canalization of auxin flow through narrowed auxin-conducting channels that demarcate future vasculature.
Prerequisites for canalization involve the feedback regulation of directional auxin transport, as manifested at the cellular level by the auxin effect on polar, subcellular localization of PIN auxin transporters 4,40 . Our results show that endogenous as well as exogenous SLs interfere with canalization-dependent developmental processes, and specifically interfere with auxin feedback on PIN polarity and clathrin-mediated endocytosis of PIN proteins. This SL action does not require the regulation of transcription and occurs through the known D14-and MAX2mediated signaling pathways. Thus, SLs may repress a mechanism that enables auxin to inhibit PIN internalization and polarization or SLs may inhibit auxin bioactivity in this cellular context. This is indicated by a proposed role for SLs in the regulation of auxin biosynthesis in context of shoot gravitropism 32 . However, auxin biosynthesis and auxin levels could also be repressed as a consequence of inhibition of auxin transport 60 . Moreover, in our tests, SLs also inhibit the action of exogenously applied auxin, suggesting SLs act downstream of auxin biosynthesis.
It was suggested previously that, in context of shoot branching, SLs destabilize PINs via promoting their internalization from the PM 29,37,61 . However, our observations in roots suggest that SLs do not affect endocytosis or PIN internalization per se, but specifically uncouple the effect of auxin on endocytosis and trafficking processes. Alternatively, SLs could divert endocytic PIN trafficking into an auxin-insensitive pathway, thus making PIN retrieval from membranes more efficient and possibly auxininsensitive. In any case, given that SLs also interfere with canalization-mediated processes in context of branching and vascular tissue formation and regeneration in shoots, it is likely that the above-mentioned PIN1-GFP-based observations in shoots 29,37,61 are in fact results of the here-identified SL effects on auxin feedback on PIN internalization.
Our findings identify a cellular mechanism, acting downstream of D14 and MAX2-dependent SL and karrikin signaling, and provide a mechanistic framework for the important part of developmental roles of the pathways, including vascularization and the regulation of root and shoot architecture. Further work should identify the precise molecular links between the SL/karrikin-related pathways and auxin feedback on PIN polarity.
Arabidopsis was stably transformed with pTA7002 67 to only express the empty Dex-inducible GVG cassette. This cassette can occasionally cause unspecific growth and defense defects 44 . However, we observed normal plant growth upon Dex treatment. Auxin transport assays in pea. For auxin transport assay on the axillary buds, the upper axillary buds were treated with water lanolin pastes or pastes with GR24 (0.03 μM). After 4 h, the treated and untreated plants were decapitated 10 mm above the upper bud. 0.5 μl of [5-3 H]-IAA (American Radiolabeled Chemicals, 925 Gbq mmol −1 , 6666 Bq μl −1 ) diluted in 50% ethanol was then applied to the tip of the axillary buds after decapitation in 6 h. Following 1.5 h treatment, the stems at a distance of 0-4 and 4-8 mm below the upper axillary buds were cut into 4 mm segments, respectively. All samples were incubated in a dioxane-based liquid scintillator cocktail overnight. The [ 3 H] activity was then measured with a scintillation spectrophotometer Packard TRI/Carb 2000 (Packard).
Gene expression analyses. For gene expression of PsDRM1, GR24 (0.03 μM) in water lanolin paste was applied on the upper axillary bud of decapitated plants as a ring. PsDRM1 expression was then followed in the untreated lower and treated upper axillary buds.
Total RNA was extracted from buds of pea plants using the RNeasy Plant Mini Kit (Qiagen). RNase-free DNase I (Qiagen) was used to remove genomic DNA. RNA was then reverse transcribed using the Superscript III cDNA kit (Invitrogen). Resulting cDNAs were used to detect PsDRM1 gene expression by quantitative Real-Time PCR (qRT-PCR) using LC 480 SYBR Green I Master Mix (Roche Diagnostics) with the specific primers 40 (Supplementary Table 1). The gene expression normalization was performed by using the combination of three reference genes (Psβ-tubulin, PsActin, and PsEF1-α).
Vascular tissue formation analyses. Young Arabidopsis plants with inflorescence stems having primary tissue architecture (vascular bundles separated by interfascicular parenchyma sectors) were used for analyzing vasculature regeneration and formation after wounding or from local application of compounds according to method described in previous study 41,68 . Briefly, first step was aimed to obtain a closed ring of active vascular cambium and secondary tissue architecture in immature inflorescence stems; second step was to analyze regeneration of incised vascular cambium and formation of new vessels in wounded stems.
In situ expression and localization analyses. In Arabidopsis, whole-mount immunolocalization was performed following the published protocol 69 . Antibodies were diluted as follows: 1:1000 for rabbit anti-PIN1 15 (produced and processed in lab); 1:1000 for rabbit anti-PIN2 70 (produced and processed in lab); and 1:600 for CY3-conjugated anti-rabbit secondary antibody (Sigma, C2306). In pea, water lanolin pastes containing IAA (0.16 μM), or IAA/GR24 (0.16 µM/0.09 µM) were applied on the stem stump or on the stem 2 mm below lateral incision. Immunolocalization was performed on longitudinal pea stem segments as described for Arabidopsis stem 69 . The Arabidopsis anti-PIN1 antibody can also recognize the homologous PIN protein in pea 4 . Antibodies were diluted as follows: 1:1000 for rabbit anti-PIN1 15 (produced and processed in lab); and 1:500 for CY3-conjugated anti-rabbit secondary antibody (Sigma, C2306). All the fluorescence signals were evaluated on Zeiss LSM 700, Zeiss LSM 710, Zeiss Observer. Z1, Leica TCS SP2, Olympus Fluoview FV1000, or Olympus Fluoview 200 confocal scanning microscopes. Unless otherwise noted, the same microscope settings were usually used for each independent experiment and pixel intensities were taken into account when comparing the images between different samples. Images were finally assembled in Adobe Photoshop CC 2015 and Adobe Illustrator CS6.

Data availability
The data supporting the findings of this study are available within the paper and its supplementary information files, or from the corresponding authors upon reasonable request. The source data underlying Figs. 2c-e, 3b-d, f-h, 4c-e, and Supplementary Figs. 1b-d, 2b, 3b, d, 4a, b, e, f, i-k, 5c-h, 6b-d are provided as a Source data file.