Regulation of plant phototropic growth by NPH3/RPT2-like substrate phosphorylation and 14-3-3 binding

Polarity underlies all directional growth responses in plants including growth towards the light (phototropism). The plasma-membrane associated protein, NON-PHOTOTROPIC HYPOCOTYL 3 (NPH3) is a key determinant of phototropic growth which is regulated by phototropin (phot) AGC kinases. Here we demonstrate that NPH3 is directly phosphorylated by phot1 within a conserved C-terminal consensus sequence (RxS) that is necessary to promote phototropism and petiole positioning in Arabidopsis. RxS phosphorylation also triggers 14-3-3 binding combined with changes in NPH3 phosphorylation and localisation status. Mutants of NPH3 that are unable to bind or constitutively bind 14-3-3 s show compromised functionality consistent with a model where phototropic curvature is established by signalling outputs arising from a gradient of NPH3 RxS phosphorylation across the stem. Our findings therefore establish that NPH3/RPT2-Like (NRL) proteins are phosphorylation targets for plant AGC kinases. Moreover, RxS phosphorylation is conserved in other members of the NRL family, suggesting a common mechanism of regulating plant growth to the prevailing light environment.

T he ability to sense and respond to the prevailing light conditions is instrumental for plants to adapt their growth and development to the external environment. Phototropism allows plants to re-orientate shoot growth towards a directional light source, which promotes light capture and early seedling growth 1 . Phototropism is induced by UV/blue light and is mediated by two phototropin (phot) light-activated kinases, phot1 and phot2 2 . Phot1 is the primary phototropic receptor and functions over a wide range of fluence rates, whereas phot2 activity requires higher light intensities 3 . Phots also control physiological responses such as chloroplast movement, leaf positioning, leaf expansion and stomatal opening 4 , which together serve to optimise photosynthetic efficiency and growth [5][6][7] .
Phototropins are plasma membrane-associated kinases containing two light, oxygen, or voltage-sensing domains (LOV1 and LOV2) at their N-terminus, which bind oxidised flavin mononucleotide (FMN) as a UV/blue light-absorbing cofactor 8,9 . Light perception, primarily by LOV2, results in activation of phototropin kinase activity and receptor autophosphorylation 10,11 . Although multiple phosphorylation sites have been identified within phot1 and phot2 12 , only sites within the kinase activation loop have been shown to be important for signalling, and kinaseinactive variants of phot1 and phot2 are non-functional 13,14 .
Despite the importance of phot kinase activity for downstream signalling, only a limited number of substrates have been identified to date. BLUE LIGHT SIGNALLING 1 (BLUS1) and CONVERGENCE OF BLUE LIGHT AND CO 2 1 (CBC1) are phot1-kinase substrates involved in blue light-induced stomatal opening 15,16 , while phosphorylation of ATP-BINDING CAS-SETTE B19 (ABCB19) and PHYTOCHROME KINASE SUBSTRATE 4 (PKS4) by phot1 modulates hypocotyl phototropism [17][18][19] . Given the variety of physiological responses mediated by phot signalling, further phot kinase substrates likely await identification 20 .
Phototropism results from the establishment of lateral gradients of the phytohormone auxin, which leads to increased cell expansion on the shaded side of the hypocotyl 1 . NON-PHOTOTROPIC HYPOCOTYL 3 (NPH3) is an essential signalling component for phototropism and is required for the formation of the lateral auxin gradients 21,22 . NPH3, together with ROOT PHOTOTROPISM 2 (RPT2), are the founding members of the NPH3/RPT2-Like (NRL) protein family, which contains 33 members in Arabidopsis 23,24 . The primary amino-acid structure of NPH3 can be separated into three regions based on sequence conservation with other NRL proteins: an N-terminal BTB (brica-brac, tramtrack and broad complex) domain, a central NPH3 domain and a C-terminal coiled-coil domain 24 . The C-terminal portion of NPH3, including the coiled-coil domain, is proposed to facilitate localisation of NPH3 to the plasma membrane 25 as well as mediating direct interaction with phot1 21 . NPH3 is reported to function as a substrate adaptor in a CULLIN3-based E3 ubiquitin ligase complex targeting phot1 for ubiquitination 26 . Ubiquitination of phot1 may be involved in receptor desensitisation, particularly under high-light irradiation 26 , but its importance in phot1 signalling is currently unknown.
Although the biochemical function of NPH3 remains unresolved, activation of phot1 by blue light results in dynamic changes to NPH3 phosphorylation status and subcellular localisation 27,28 . NPH3 is phosphorylated on multiple sites in darkness, including sites located towards the N-terminus within the NPH3 domain 29 , and localises to the plasma membrane 27 . Upon blue light perception, NPH3 is rapidly dephosphorylated 30 and becomes internalised into aggregates, which transiently attenuates its interaction with phot1 27,28 . These effects are reversible in darkness, with the kinetics of NPH3 rephosphorylation matching the photoactive lifetime of phot1 7 . The kinases and phosphatases that modulate NPH3 phosphorylation status are unknown, however, reduced levels of dephosphorylation, and relocalisation into aggregates, correlates with enhanced phototropic responsiveness observed in de-etiolated (green) seedlings 28 .
Along with NPH3, two other NRL family members also have known roles in phot signalling pathways. RPT2 interacts with both phot1 and NPH3 31,32 , it is proposed to influence NPH3 phosphorylation status and promote the reconstitution of the phot1-NPH3 complex to sustain signalling under higher light intensities 27 . In line with this, phototropic responsiveness in mutant seedlings lacking RPT2 decreases as light intensity is increased 33 . Similarly, RPT2 expression levels are low in darkness, but increase with irradiation in a fluence-dependent manner 33 . RPT2, together with NPH3, is also involved in phot-mediated leaf positioning and leaf expansion responses 25,34 . NRL PROTEIN FOR CHLOROPLAST MOVEMENT 1 (NCH1) is positioned within the same clade as RPT2 in the Arabidopsis NRL phylogenetic tree 24 . NCH1 and RPT2 redundantly mediate chloroplast accumulation movements in response to low-intensity light 35 .
Phot signalling is dependent upon reversible changes in phosphorylation 12 . 14-3-3 proteins are present in all eukaryotic organisms and bind to target proteins through the identification of phospho-serine/threonine motifs 36,37 . 14-3-3 binding can produce a variety of consequences, such as regulation of enzymatic activity, changes in subcellular localisation, protein stability or alteration of protein-protein interactions 38 . 14-3-3 proteins are known to bind to phot1 and phot2 following receptor autophosphorylation 13,32,39,40 , whereas NPH3 and RPT2 have both been identified as components of the 14-3-3 interactome 41,42 . However, the functional relevance of these interactions and the roles of 14-3-3 proteins in phot signalling remains unclear.
Despite the importance of NRL proteins in blue light-mediated responses, how signalling is initiated upon phot activation is still not known. In the present study, we identify NPH3 as a substrate for phot1-kinase activity. Phosphorylation of NPH3 at the C-terminus by phot1 results in 14-3-3 binding, which is required for early signalling events and promotes NPH3 functionality. The C-terminal phosphorylation site of NPH3 is conserved in several NRL family members, including RPT2, suggesting phot-mediated phosphorylation and 14-3-3 binding may represent a conserved mechanism of regulation.

Results
Light-dependent 14-3-3 binding to NPH3. In order to identify additional components involved in blue light signalling, GFP-NPH3 was immunoprecipitated from etiolated nph3 mutant seedlings expressing functional NPH3::GFP-NPH3 28 . Anti-GFP immunoprecipitations (IPs) were performed on total protein extracts from seedlings maintained in darkness or after a brief blue light treatment (20 μmol m −2 s −1 for 15 min) to capture early signalling events. Co-purifying proteins were analysed by label-free quantitative tandem mass spectrometry (MS; Supplementary Data 1) to allow the identification of proteins whose abundance changed following blue light irradiation. As expected, phot1 was recovered in the IPs from both dark-and light-treated seedlings, but at a higher abundance in the dark (Supplementary Data 2). This is in agreement with previous results showing NPH3-phot1 interactions are attenuated by blue light 27 . Conversely, several 14-3-3 isoforms were detected at greater abundance following blue light irradiation (Fig. 1a). 14-3-3 proteins bind to target proteins through recognition of phospho-serine/threonine-containing motifs. Arabidopsis expresses 13 different 14-3-3 isoforms that can be phylogenetically divided into the epsilon and non-epsilon groups 43 . Farwestern blotting was performed to assess direct 14-3-3 binding to GFP-NPH3. Binding of recombinant 14-3-3 Lambda (nonepsilon group member) and 14-3-3 Epsilon (epsilon group member) fused to glutathione-S-transferase (GST) was not detected for GFP-NPH3 IPs from etiolated seedlings maintained in darkness (Fig. 1b). Blue light irradiation results in enhanced electrophoretic mobility of GFP-NPH3 owing to its rapid dephosphorylation 30 . Concurrently, binding of 14-3-3 Lambda and Epsilon was observed following irradiation, whereas no binding was observed when GST alone was used as the probe. In line with the results from IP-MS analysis, no specificity in binding of 14-3-3 proteins from epsilon and non-epsilon groups was detected. These results suggest that blue light irradiation triggers both phosphorylation, and concomitant 14-3-3 binding, as well as dephosphorylation events on NPH3.
Analysis of phosphorylation sites within NPH3. Activation of phot1 by blue light results not only in rapid changes in the phosphorylation status of NPH3 but also its subcellular localisation 27,28 . In the darkness, NPH3 localises predominantly to the plasma membrane but is rapidly internalised into aggregates upon blue light treatment. Based on data from global phosphoproteomics experiments 44,45 three regions of NPH3 (M1, M2 and M3) containing the majority of experimentally identified phosphopeptides were selected for mutational analysis (Fig. 2a). Within each of the regions, all of the serine and threonine residues were replaced with alanine to mimic the dephosphorylated state. The mutations were introduced into the NPH3::GFP-NPH3 construct, transiently expressed in the leaves of Nicotiana benthamiana and compared with the expression of the non-mutated GFP-NPH3 control. Transfected N. benthamiana plants were dark-adapted before confocal observation. The localisation of transiently expressed GFP-NPH3 was similar to that of functionally active GFP-NPH3 in Arabidopsis 28 described above, and repeated scanning with the 488-nm laser used to concomitantly excite GFP along with endogenous phot1, induced relocalisation of GFP-NPH3 into aggregates (Fig. 2b). The localisation of each of the transiently expressed NPH3 phospho-mutants was the same as GFP-NPH3 when imaged immediately (scan 1). Repeated laser scanning was effective in inducing relocalisation for both NPH3-M1 and M2 constructs, whereas the NPH3-M3 mutant failed to show any light-induced changes in subcellular localisation.
Phot1-induced changes in NPH3 localisation are correlated with changes in NPH3 phosphorylation status in transgenic Arabidopsis seedlings 27,28 . Immunoblot analysis of protein extracts from darkadapted leaves of N. benthamiana transiently expressing GFP-NPH3 irradiated with blue light also showed enhanced electrophoretic mobility compared with leaves maintained in darkness (Fig. 2c), although to a lesser degree than observed in etiolated Arabidopsis seedlings expressing GFP-NPH3 when equivalent light treatments were used (Fig. 1b). Both the NPH3-M1 and M3 mutants were affected for this response, whereas the NPH3-M2 mutant response was similar to GFP-NPH3 (Fig. 2c). The NPH3-M1 mutant showed enhanced electrophoretic mobility in the dark compared with the GFP-NPH3 construct, with a further slight enhancement following blue light treatment. The NPH3-M1 mutant contains mutations of serine residues S213, S223, S233 and S237, mutation of which was previously shown to contribute to reducing the electrophoretic mobility of NPH3 in darkness 29 . Conversely, the NPH3-M3 mutant migrated at the same position as GFP-NPH3 in the dark, even following blue light irradiation. Therefore, mutation of phosphorylation sites within the M3 region at the C-terminus of NPH3 prevents both blue light-induced dephosphorylation of sites within the NPH3 domain, which contribute to reducing the electrophoretic mobility, as well as subcellular relocalisation into aggregates.
The C-terminal amino-acid sequence of NPH3 is highly conserved in angiosperms ( Supplementary Fig. 1a) and contains two serine residues, S744 and S746 in Arabidopsis NPH3. Mutation of either serine residue to alanine, singularly or together, prevented (for S744A and S744A S746A) or greatly reduced (for S746A) the light-induced relocalisation response when transiently expressed in N. benthamiana (Fig. 2d). Similarly, these mutations also prevented dephosphorylation of NPH3 following blue light irradiation (Fig. 2e). Therefore, mutation of S744 and/or S746 can reproduce the results obtained with the NPH3-M3 mutant. Although serine to alanine mutations effectively blocks phosphorylation of the respective residue, phosphomimetic substitutions aim to mimic the phosphorylated state by replacement with a negatively charged amino acid. However, mutation of S744 and S746 to aspartate produced similar results to the alanine mutations; loss of light-induced relocalisation and dephosphorylation ( Supplementary Fig. 1b, Supplementary Fig. 1c).

S744 is required for 14-3-3 binding and early signalling events.
To examine the effects of the C-terminal serine residues S744 and S746 in NPH3 signalling, we generated transgenic Arabidopsis expressing NPH3::GFP-NPH3 containing S744A S746A, S744D S746D, S744A or S746A mutations in the nph3 mutant background. Confocal imaging of hypocotyl cells of etiolated seedlings expressing NPH3 S744A S746A or NPH3 S744D S746D showed that both mutants did not relocalise into aggregates following irradiation with the 488 nm laser, in contrast to the GFP-NPH3 control (Fig. 3a). The single NPH3 S744A mutant also lacked this response, whereas the NPH3 S746A mutant was unaffected. Furthermore, analysis of NPH3 dephosphorylation showed that seedlings expressing NPH3 S744A S746A, S744D S746D or S744A exhibited no change in electrophoretic mobility with blue light treatment, in contrast to NPH3 S746A mutant and GFP-NPH3 expressing lines, which both displayed enhanced mobility with blue light treatment (Fig. 3b). Whereas results from transient expression analysis in N. benthamiana showed both S744 and S746 were involved in these early signalling responses (Fig. 2d, e), analysis of transgenic Arabidopsis identifies only S744 as being required.
To determine whether S744 was also required to mediate interactions between NPH3 and 14-3-3 proteins, far-western blotting was performed on anti-GFP IPs from nph3 seedlings expressing GFP-NPH3 or GFP-NPH3 containing S744A or S746A mutations (Fig. 3c). The binding of recombinant 14-3-3 Epsilon was evident for both GFP-NPH3 and NPH3 S746A mutant in a light-dependent manner, with the signal for S746A being substantially lower. However, no binding could be detected for the NPH3 S744A mutant. Phosphorylation of S744 is, therefore, necessary for 14-3-3 binding, subcellular relocalisation and dephosphorylation of N-terminal sites (including S213, S223, S233 and S237) in response to blue light perception.
Phot1 phosphorylates NPH3 at position S744 in a lightdependent manner. Given the evidence for light-induced phosphorylation of NPH3, we examined whether NPH3 was a direct substrate for phot1-kinase activity using a gatekeeper engineered phot1 (phot1 GK ), which can accommodate the bulky ATP analogue N 6 -benzyl-ATPγS as a thiophospho-donor 20 . NPH3, or the NPH3 S744A mutant, were co-expressed in a cell-free expression system with phot1 GK and used for in vitro kinase assays in the presence of N 6 -benzyl-ATPγS. Light-induced thiophosphorylation, which can be detected by immunoblotting with anti-thiophosphoester antibody following chemical alkylation of the incorporated thiophosphates, was detected for NPH3 but not for the NPH3 S744A mutant (Fig. 4a), showing phot1 can specifically phosphorylate residue S744 of NPH3 in vitro. To detect the phosphorylation status of S744 in vivo, we raised a phospho-specific antibody (pS744).
Phosphorylation of S744 was observed in WT and GFP-NPH3expressing seedlings in a light-dependent manner and mutation of S774 resulted in a loss of signal, demonstrating the specificity of the pS744 phospho-specific antibody (Fig. 4b). Phosphorylation of S744 was also detectable for NPH3 S746A mutant-expressing seedlings at a reduced level, similar to the results observed for 14-3-3 binding (Fig. 3c). Phot1 is the main photoreceptor mediating phototropism to low (<1 μmol m −2 s −1 ) and high (>1 μmol m −2 s −1 ) fluence rates of blue light, whereas phot2 functions predominantly at higher light intensities 3 (>10 μmol m −2 s −1 ). Phosphorylation of S744 occurred in WT seedlings in response to both low blue (0.5 µmol m −2 s −1 ) and high blue (50 µmol m −2 s −1 ) light treatments concomitantly with dephosphorylation of sites within the NPH3 domain, detected via changes in electrophoretic mobility when probed with anti-NPH3 antibody (Fig. 4c). These responses were absent in phot1 phot2 double mutant and phot1 single mutant seedlings, but unchanged in the phot2 single mutant, demonstrating that phosphorylation of S744 and dephosphorylation of sites within the NPH3 domain, which alter electrophoretic mobility, are phot1-specific responses in etiolated seedlings.
To assess the kinetics of changes in NPH3 phosphorylation status, we performed time-course experiments. Dephosphorylation of sites that alter NPH3 electrophoretic mobility required 15 min of blue light irradiation (Fig. 5a), whereas phosphorylation of S744 was detected within 30 s and maintained over the 2 h irradiation period. When etiolated seedlings were returned to darkness following blue light exposure, S744 was dephosphorylated within 15 min, matching the time required for rephosphorylation of sites responsible for the electrophoretic mobility shift (Fig. 5b). Therefore, phot1 phosphorylation of S744 is rapid, occurring before light-induced dephosphorylation of sites within the NPH3 domain, and reversible in darkness.
Phot1 phosphorylation of NPH3 promotes functionality. Arabidopsis mutants lacking NPH3 fail to exhibit hypocotyl phototropism under a variety of different light conditions 33,46 . Phototropism in two independent homozygous transgenic nph3 mutants expressing NPH3::GFP-NPH3 is restored to levels comparable to non-transgenic WT seedlings when irradiated with 0.5 µmol m −2 s −1 of unilateral blue light (Fig. 6a). In contrast, the magnitude and kinetics of phototropic curvature were reduced in seedlings expressing GFP-NPH3 with both S744 and S746 residues mutated to alanine or aspartate (Fig. 6a, Supplementary Fig. 2a). Similarly, phototropism was reduced in seedlings expressing GFP-NPH3 containing the S744A mutation, while the NPH3 S746A mutant-expressing seedlings were fully functional (Fig. 6b, Supplementary Fig. 2b). To determine whether the reduced phototropic responsiveness of the S744A mutant is due to altered photosensitivity, phototropism was further assessed under lower (0.05 µmol m −2 s −1 ; Fig. 6c, Supplementary Fig. 2c) and higher (20 µmol m −2 s −1 ; Fig. 6d, Supplementary Fig. 2d) intensity blue light irradiation. Under both fluence rates, transgenic lines expressing the NPH3 S744A mutant were less responsive than the GFP-NPH3 or NPH3 S746A mutantexpressing lines. NPH3 also functions in phototropin-mediated leaf positioning, particularly in low light environments 25 . In WT seedlings transferred to low-intensity white light (10 µmol m −2 s −1 ) the petioles of the first true leaves were positioned obliquely upwards in order to maximise light capture, while the petioles of nph3 mutant seedlings were positioned horizontally (Fig. 6e). Seedlings expressing GFP-NPH3 or the NPH3 S746A mutant were complemented for petiole positioning, while the response of seedlings expressing the NPH3 S744A mutant was significantly reduced (Fig. 6e), which was also observed for the NPH3 S744A S746A and S744D S746D transgenic lines (Supplementary Fig. 3). These results demonstrate that phot1 phosphorylation of S744 positively regulates NPH3 function.
To confirm these results in stable transgenic lines, Arabidopsis nph3 mutants were transformed with NPH3::GFP-NPH3 containing the R18 or mR18 sequences. Confocal imaging of hypocotyl cells of etiolated seedlings revealed similar patterns of localisation observed in N. benthamiana, with GFP-NPH3-R18 forming aggregates in darkness, whereas GFP-NPH3-mR18 failed to relocalise following repeated laser scanning (Fig. 7c). Consistent with the subcellular localisation patterns, analysis of NPH3 dephosphorylation showed that lines expressing GFP-NPH3-mR18 display no change in electrophoretic mobility following blue light treatment, whereas a portion of GFP-NPH3-R18 exhibited enhanced electrophoretic mobility both in darkness and after irradiation (Fig. 7d). Farwestern blotting was used to confirm the constitutive binding of recombinant 14-3-3 Epsilon to GFP-NPH3-R18 immunoprecipitated from seedlings maintained in darkness and following blue light irradiation, as well as the absence of 14-3-3 binding to GFP-NPH3-mR18 (Fig. 7e). Together, these results show that engineered 14-3-3 binding, independent from phot1-mediated S744 phosphorylation, is partially sufficient to induce changes in NPH3 dephosphorylation and localisation status.

Discussion
In this study, we used MS to identify proteins coimmunoprecipitating with GFP-NPH3. This revealed 14-3-3 proteins as NPH3 interactors specifically following a blue light treatment (Fig. 1). Using a chemical-genetic approach, we have found that NPH3 is phosphorylated by phot1 on the C-terminally positioned S744 in a light-dependent manner (Fig. 4a). Moreover, the generation of anti-pS744 antibodies confirmed light-induced phosphorylation of S744 in vivo (Fig. 4c). Phototropins are members of the AGCVIII (protein kinase A, cyclic GMP-dependent protein kinase and protein kinase C) subfamily of protein kinases 50 and S744 is part of a PKAlike phosphorylation consensus sequence (RxS), as are the previously identified phot1-kinase substrates BLUS1 15 , CBC1 16 and PKS4 19 ( Supplementary Fig. 6a). Phot1-mediated phosphorylation of S744 is required to elicit the previously documented early cellular events associated with NPH3 activation such as dephosphorylation 30 and subcellular relocalisation 27,28 . This is consistent with previous observations of changes in NPH3 electrophoretic mobility correlating with the lifetime duration of phot1 activation in planta 7 and occurring locally only in cells/tissues where both proteins are present 51 . Furthermore, a constitutively active phot1-variant can induce NPH3 dephosphorylation in darkness 52 . The phosphorylation status of residues S213, S223, S233 and S237 contribute to reducing the electrophoretic mobility of NPH3 in darkness 29 , however other unidentified sites are also involved ( Fig. 2c; 27 ). Recently, phosphopeptide mapping of YFP tagged NPH3 immunoprecipitated from etiolated seedlings maintained in darkness or irradiated with blue light identified seven phosphorylation sites, including S213, S223 and S237 53 . However, differential phosphorylation was evident for only two phosphorylation sites, phosphorylated S213 and S237 was detected in samples from seedlings maintained in darkness and absent following blue light irradiation 53 . The kinase(s) and phosphatase(s) regulating the phosphorylation status of these sites are currently unknown and their role in regulating NPH3 signalling is unclear. However, mutation of S213, S223, S233 and S237 to alanine, or deletion of amino-acid residues S213-S239, did not impact their ability to restore phototropism in nph3 mutant seedlings 29 , or form aggregates when transiently expressed in N. benthamiana (Fig. 2b). Phosphorylation of S744 creates a 14-3-3 binding site (Fig. 3c), which conforms to the C-terminal mode III 14-3-3-binding motif pS/ pTX 1-2 -COOH 38 . We created a translational fusion between NPH3 and the synthetic R18 peptide to study the role of 14-3-3 binding in the absence of phot1 phosphorylation (Fig. 7e). 14-3-3 binding alone was able to induce NPH3 relocalisation into aggregates (Fig. 7b, c) and partially reduce the electrophoretic mobility of NPH3 (Fig. 7d), in the absence or presence of light. Light-dependent 14-3-3 binding has also been shown for phot1; non-epsilon 14-3-3 s bind to 3 phosphorylation sites located between the LOV1 and LOV2 photosensory domains 32 , but the functional relevance of this interaction is unknown as mutation of 2 of the phosphorylation sites did not impair functionality 25 . In contrast, no isoform specificity was observed for 14-3-3 binding to NPH3, with both epsilon and nonepsilon isoforms shown to interact (Fig. 1). Functional redundancy between 14-3-3 isoforms means loss-of-function mutants often show few, if any, phenotypes, with even quadruple non-epsilon 14-3-3 mutants displaying mild growth phenotypes under non-stress growth conditions 54 , with no obvious differences in phototropism or NPH3 dephosphorylation kinetics observed compared with WT seedlings (Supplementary Fig. 7). However, conditional RNA interference lines targeting three 14-3-3 epsilon members (epsilon, mu and omicron) displayed several auxin-related phenotypes, including reduced hypocotyl elongation and defects in root and hypocotyl gravitropism, due to altered polarity of the PIN-FORMED (PIN) auxin transporters as a consequence of 14-3-3 regulation of cellular trafficking 42 . NPH3 is also reported to be required for phot1-driven changes in PIN2 trafficking during negative phototropic bending of roots 55 .
The biochemical basis underpinning phototropism is the formation of a gradient of phot1 activation across the stem 56 , which results in an asymmetric accumulation of auxin on the shaded side through an unidentified mechanism 2 . We previously demonstrated that a gradient of GFP-NPH3 relocalisation occurs across the hypocotyl of Arabidopsis seedlings during unilateral irradiation with blue light 28 . Here, we report that seedlings expressing mutants of GFP-NPH3 unable to form such a gradient, either through mutation of the phosphorylation site required for 14-3-3 binding (S744) or owing to constitutive 14-3-3 binding via the R18 peptide, have a severely compromised phototropic response. Thus, phototropic curvature likely involves signalling outputs mediated by a gradient in NPH3 localisation across the stem. Our current findings are therefore consistent with 14-3-3 proteins being instrumental components regulating auxin-dependent growth 42 .
The phot1 phosphorylation consensus sequence of NPH3 is also conserved in several other NRL proteins including RPT2, NCH1, and members of the NAKED PINS IN YUCCA (NPY) clade (Supplementary Fig. 6b). Notably, RPT2 was identified in immunoprecipitants of seedlings expressing 14-3-3 epsilon-GFP 42 . We could also detect phosphorylation of RPT2 on the corresponding serine residue (S591) when co-expressed with phot1 GK in vitro kinase assays ( Supplementary Fig. 6c), as well as light-dependent phosphorylation of DEFECTIVELY ORGANISED TRIBUTARIES 3 (DOT3), NPY1 and NRL1 (Supplementary Fig. 6d). It is therefore possible the residual functionality seen in GFP-NPH3 S744A seedlings (Fig. 6) arises from co-action with other NRL family members. The NPY clade of NRL proteins function redundantly to mediate organogenesis and root gravitropism [57][58][59] . These responses are not reported to involve phototropin signalling, but rather the related AGCVIII kinases PINOID (PID) and its close homologues WAG1 and WAG2, and the D6 PROTEIN KINASE (D6PK) family 60 . PID/WAGs and D6PKs phosphorylate PIN transporters on RxS phosphorylation site motifs 50 and physically interact with NPY proteins to maintain PIN polar localisation and therefore directional auxin transport 60 . Furthermore, aggregate formation is not limited to NPH3 but has also been documented for NPY1 when expressed in Arabidopsis protoplasts 57 . Therefore, phosphorylation and concomitant 14-3-3 binding to the C-terminus may represent a conserved mechanism of regulation for NRL proteins.
Determining the biochemical function of NPH3, and related family members, is now required to understand how phots signal via NRL proteins to coordinate different light-capturing processes in plants that will ultimately offer new opportunities to manipulate plant growth through alterations in photosynthetic capacity.

Methods
Plant material and growth. Wild-type Arabidopsis (gl-1, ecotype Columbia), nph3-6 21 , 14-3-3 quadruple mutants 54 and the GFP-NPH3 transgenic line 28  Transient expression in Nicotiana benthamiana. To create transformation vectors encoding NPH3 with multiple serine and threonine residues mutated to alanine, fragments of NPH3 were synthesised (ThermoFisher Scientific) encoding the 13 alanine substitutions for NPH3-M1, 8 substitutions for NPH3-M2 and 15 substitutions for NPH3-M3. The synthesised fragments were introduced into NPH3::GFP-NPH3 using KpnI and MluI restriction sites for GFP-NPH3-M1, MluI and PstI restriction sites for GFP-NPH3-M2 and PstI and BamHI restriction sites for GFP-NPH3-M3. Agrobacterium-mediated transient expression in Nicotiana benthamiana was performed as reported previously 61 . Agrobacterium tumefaciens strain GV3101, transformed with the plasmid of interest, was resuspended in infiltration buffer (10 mM MgCl 2 , 10 mM MES-KOH [pH 5.6] and 200 mM acetosyringone) at an OD 600 of 0.4 and syringe-infiltrated into leaves of 3-4-weekold N. benthamiana plants. Plants were dark-adapted for 16 h before 1 cm leaf discs for confocal observation or protein extraction were taken 2 d post infiltration. For blue light irradiation, leaf discs were placed abaxial-side upwards on the surface of MS medium agar plates for the duration of the treatment.
Transformation of Arabidopsis. Amino-acid substitutions of S744 and/or S746 were introduced into the pUC-SP vector containing the NPH3-coding sequence by site-directed mutagenesis and verified by DNA sequencing. The coding sequence of NPH3 in the NPH3::GFP-NPH3 pEZR(K)-LC binary vector 28 was replaced by the coding sequences containing the phosphosite mutations using Gibson Assembly (New England Biolabs). To create transformation vectors NPH3::GFP-NPH3-R18 and NPH3::GFP-NPH3-mR18, a fragment encoding amino-acid residues 419-743 was PCR amplified from NPH3 pUC-SP with primers containing the R18 or mR18coding sequence and inserted into NPH3::GFP-NPH3 using MluI and BamHI restriction sites. All primer sequences are available in Supplementary Table 1. The nph3-6 mutant was transformed with Agrobacterium tumefaciens strain GV3101 using a streamlined floral dipping protocol 62 . A 500 ml saturating A. tumefaciens culture, transformed with the plasmid of interest, was grown in YEBS medium at 28°C in a shaking incubator, diluted with 500 ml of 5% (w/v) sucrose and Silwet® L-77 added to a final concentration of 0.01% (v/v). Flowering nph3-6 mutant plants were briefly dipped into the solution and sealed in a plastic bag overnight. Plants were dipped for a second time 3-5 days later. Based on the segregation of kanamycin resistance independent homozygous T3 lines, or for GFP-NPH3-mR18 transgenics single-insertion T2 lines, were selected for analysis.
Phototropism. Phototropism was performed using free-standing etiolated seedlings 51 . Seeds were sown in transparent plastic entomology boxes (Watkins and Doncaster) on a layer of silicon dioxide (Honeywell, Fluka), watered with quarter-strength MS medium and grown in darkness for 64-68 h. Seedlings were placed into unilateral blue light and images were recorded every 10 min for 4 h with a Retiga 6000 CCD camera (QImaging) connected to a personal computer running QCapture Pro 7 software (QImaging) with supplemental infra-red illumination. Hypocotyl curvature was measured from two biological replicates, with~10 seedlings measured from each replicate, using Fiji software 63 . Circular histograms were produced using Oriana software (Kovach Computing Services). Box and whisker plots, and one-way analysis of variance (ANOVA) with Tukey post test, were performed using Graphpad Prism.
Leaf positioning. Seedlings were grown on soil for 9 d under 80 µmol m −2 s −1 white light before transfer to 10 µmol m −2 s −1 white light for 4 d. One cotyledon was removed, seedlings were placed flat on an agar plate, and plates were placed on a white light transilluminator and photographed. Petiole angles from the horizontal were measured from three biological replicates, with 20 seedlings for each replicate, using Fiji software. Box and whisker plots, and one-way ANOVA with Tukey post test, were performed using Graphpad Prism.