A peptide that regulates metalation and function of the Arabidopsis ethylene receptor

Ethylene signalling represents one of the classic hormonal pathways in plants, with diverse roles in development and stress responses. The dimeric ethylene receptor localizes to the endoplasmic reticulum (ER) and contains Cu(I) ions essential for ethylene binding and signalling. As for other vesicular cupro-proteins, the final step of Cu(I) maturation at the ER is undefined. We previously discovered that mutants in the Arabidopsis gene POLARIS (PLS), encoding a 36 amino acid peptide, exhibit enhanced ethylene signalling responses. Here we report a 1:2 thiol-dependent Cu(I):PLS2 complex, with an affinity of 3.79 (±1.5) x1019 M-2. We demonstrate PLS interactions with the transmembrane domain of receptor protein ETR1, the Cu(I) chaperones ATX1 and CCH, and Cu(I)-transporting P1B-type ATPase RAN1. Formation of Cu(I)-dependent PLS-cuproprotein interactions at the ER provides a mechanism to modulate the metalation of ETR1, thereby regulating its activity and representing a novel mechanism for plant hormone receptor regulation. Significances statement Ethylene signalling represents one of the classic hormonal pathways in plants, with diverse roles in development and stress responses. The dimeric ethylene receptor localizes to the endoplasmic reticulum (ER) and contains Cu(I) ions essential for ethylene binding and signalling. The polaris (pls) mutant of Arabidopsis has been shown to exhibit defective ethylene responses. Here we show that the POLARIS gene product, a 36 amino acids peptide, binds copper and interacts with the ethylene hormone receptor to regulate its activity in a tissue-specific manner.

Introduction of copper to the ER and ethylene receptor requires the RAN1 (RESPONSIVE TO ANTAGONIST1) protein.This is a predicted copper-transporting P-type ATPase homologous to the yeast Ccc2p and to human Menkes and Wilson disease proteins (15).
Strong loss-of-function mutants of RAN1 in Arabidopsis (e.g.ran1-3, ran1-4) exhibit an enhanced ethylene signalling response (16), consistent with a loss of receptor function, and similar to higher order loss-of-function receptor mutants, which also show an ethylene hypersignalling phenotype (17).Mechanisms of copper homeostasis at ETR1 are unknown, and indeed this is true for other compartmentalized cuproproteins supplied with copper, for example via Ccc2p, Menkes or Wilson ATPases.RAN1 in Arabidopsis has been found to localize to endomembrane systems, including trans-Golgi and ER compartments, and is necessary for both the biogenesis of ethylene receptors and copper homeostasis during plant development (16).
There is also evidence that RAN1 interacts directly with ETR1 and the copper chaperones ANTIOXIDANT1 (ATX1) and COPPER TRANSPORT PROTEIN (CCH), suggestive of a transfer of copper between proteins to deliver it to the ethylene receptor as part of the receptor biogenesis pathway at the ER (18).
Our understanding of the molecular mechanism of copper delivery to the receptor is still incomplete, however.For example, is metalation of the receptor regulated in a tissue-specific way or in response to the hormonal environment in a tissue?Is this process part of the crosstalk mechanism with other hormone signaling pathways?It is well established that ethylene signaling interacts with and is impacted by other hormonal pathways, but does this influence receptor metalation and does this have developmental consequences?In this paper we propose a model in which a new component of the metalation pathway is identified, which provides cell type specificity and a crosstalk regulatory mechanism for ethylene responses, through copper delivery control to the ethylene receptor.

The POLARIS peptide is a negative regulator of ethylene responses
The POLARIS gene of Arabidopsis (AT4G39403) encodes a 36 amino acids peptide (Fig. 1a) translated from a low-abundance transcript that, in light-grown seedlings, is most abundant in the embryo and silique, seedling root tip and leaf vascular tissues, also reflected in promoter-reporter expression patterns (Fig. S1; 19).We have previously shown that the loss-of-function pls mutant has some phenotypic similarities to ran1 loss-of-function alleles and to ctr1, exhibiting a triple response phenotype (short hypocotyl and root, exaggerated apical hook, radial expansion) in the dark in the absence of ethylene (14), and a short root in light-grown seedlings, consistent with its expression in the root meristem in light-grown seedlings (Fig. 1b).Transgenic complementation of the mutant by the PLS gene suppresses the mutant phenotype (19).The pls mutant phenotype is rescued by the gain-of-function ethylene resistant mutation etr1-1 and pharmacological inhibition of ethylene signalling by silver ions (14).Ethylene gas production in the pls mutant is at wildtype levels, indicating the peptide may play a role in ethylene signalling rather than biosynthesis (14).PLS transgenic overexpressers (PLSOx seedlings) in contrast exhibit a suppression of the triple response phenotype when grown in the presence of the ethylene precursor ACC, similar to the gain of function etr1-1 mutant, but this suppression is incomplete (PLSOx seedlings show some response to ACC; 14,19); and they suppress the ctr1 mutant phenotype, indicating that the PLS peptide acts upstream of CTR1 (14).However, this repression of the PLSOx phenotype by ctr1 also is incomplete, with some elongation of the hypocotyl and root in the PLSOx/ctr1 cross compared to the ctr1 mutant in the light, and a slightly longer root for the PLSOx/ctr1 cross grown in air in the dark (14).These observations are indicative of some complexity in the relationship between CTR1 and PLS.
To further investigate the pls molecular phenotype, we carried out an RNA-seq analysis on loss-of-function pls mutant and PLSOx transgenic overexpressor seedlings to identify differentially regulated genes, with mutant and overexpressor each compared with wildtype seedlings of the same ecotype as controls.The pls mutant expresses no full length PLS coding sequence due to a T-DNA insertion in the coding region of PLS and disrupting function (19), while the PLSOx seedlings express significantly higher levels of PLS transcript compared to wild-type (log2 fold 10.8, padj value = 2.02 x 10 -33 ; Tables S1, S2).836 genes were significantly up-regulated and 292 down-regulated in the pls mutant compared to wildtype control seedlings (padj value < 0.05) (Table S1).1487 genes were significantly up-regulated and 1281 downregulated in PLSOx seedlings compared to wildtype (padj value < 0.05; Table S2).Gene Ontology (GO) analysis of genes upregulated in pls mutant seedlings compared to wildtype showed significant enrichment of genes associated with responses to hormone signalling, biotic and abiotic defence responses, and cell death (Table S3; Fig. S2a).Out of 307 GO:0009723 (response to ethylene) genes, 25 were significantly up-regulated and 7 down-regulated in the pls mutant compared with the wildtype, with 40 up-regulated and 18 down-regulated in the PLSOx seedlings (Fig. 1c; Table S3, S4, S5 and S6, 20), indicating a requirement for control over PLS expression levels for correct ethylene responses.While GO:0009723 (response to ethylene) is significantly enriched in genes up-regulated in the pls mutant compared to wildtype (FDR = 0.00062; Table S3), a large number of up-regulated genes in pls are significantly associated with immunity, response to pathogens and hypersensitive response (Table S3 and Fig. S2a).Downregulated genes in pls were significantly enriched in GO categories of hormone biosynthetic processes (GO:0042446, FDR = 0.038), hormone metabolic process (GO:0042445, FDR = 0.0096) and regulation of hormone levels (GO:0010817, FDR = 0.0013) (Table S4, Fig. S2b; 20), consistent with our previous studies that described PLS-dependent crosstalk between ethylene, auxin and cytokinin signalling (21,22,23).In PLSOx overexpressors, both up-and down-regulated genes similarly included those associated with GO terms hormone response, stress response and immune response, and genes associated with photosynthesis are downregulated in PLSOx seedlings, consistent with repressed photosynthetic activity in root tips (Table S5 and S6,Fig. S2c,d;19).These data are consistent with a role for PLS in a number of ethylene responses and other related signalling, stress and developmental processes.
To understand better the relationship between PLS peptide structure and function, and to investigate conservation of PLS function between species, we carried out synthetic peptide feeding experiments using hydroponically grown seedlings.The Arabidopsis relative Camelina sativa contains a gene with partial sequence identity to the Arabidopsis PLS gene, encoding a predicted peptide sequence that is 22 amino acids long and identical to the N-terminal 22 amino acids of the Arabidopsis PLS except for a phenylalanine to serine substitution at position nine (Fig. 1a).We synthesized full-length PLS peptide, PLS(FL) and truncated versions from both Arabidopsis and C. sativa (Fig. 1a) and supplied the peptides to Arabidopsis pls mutant seedlings hydroponically.The full-length peptides from both Arabidopsis and C. sativa and the N-terminal 22 amino acids sequence of the Arabidopsis peptide (N1) were each able to rescue the short primary root length of the Arabidopsis pls mutant (Figs.1d), similar to transgenic overexpression and genetic complementation using the wild type PLS coding sequence (14,19).The mean primary root length of wild type (C24) seedlings did not change across all peptide treatments, while supply of pls seedlings with 50 nM PLS(FL) and N1 peptides showed a rescue of root growth (Fig. 1e), and t-tests showed there was no significant difference between the rescue effects of PLS(FL) and PLS(N1); P > 0.1, n = 25).However, neither a 9 amino acids sequence (N2, Fig. 1e) from the N-terminus, nor C-terminal sequences of 14 (C1) or 24 (C2) amino acids from Arabidopsis PLS were able to rescue the mutant (Fig. 1e); each of these shorter peptides lacks one of the two Cys residues found in the functional longer length peptides PLS(FL) and PLS(N1).
Imaging showed that a fluorescent tagged (5-carboxyfluorescein, 5-FAM) version of the Arabidopsis N-terminal 22 amino acids sequence of the peptide (N1) is taken up by the roots, and also rescues the mutant root phenotype (Fig. S3a,b).

PLS localizes to the cytosolic side of the endoplasmic reticulum
Since genetic studies suggest that PLS acts close to the ethylene receptor (18), we hypothesized that it should localize to the same subcellular compartment.The ethylene receptor in Arabidopsis is localized to the ER (12), and a proPLS::PLS:GFP-generated fusion protein (PLS:GFP) was found to co-localize with both the ER marker dye ER-Tracker TM (Fig. 2a-c) and with an ERtargeted red fluorescent protein RFP-HDEL (24) in transgenic plants (Fig. 2g-i).PLS:GFP also appears to localize to the nucleus and cytoplasm (Fig. 2).As controls, free GFP protein expressed under the control of the PLS promoter is not co-localized to the ER (Fig. 2d-f) and, as expected, the Golgi marker SH:GFP does not co-localize with ER Tracker (Fig. 2m-o).Trans-Golgilocalized SULFOTRANSFERASE1 (ST1) mCherry ( 25) visualization shows PLS:GFP does not localize to the Golgi (Fig. 2j-l).To further clarify the side of the ER membrane on which PLS localizes, transient expression of redox-sensitive GFP (roGFP2) fusions of PLS were carried out.
The different excitation properties of roGFP2 in an oxidizing (ER lumen) or reducing environment (cytosol) allows discrimination of the precise location of PLS fused to RoGFP2.Ratiometric analysis and comparison with proteins of known localization (i.e.cytosolic RoGFP2, ER luminal RoGFP2, cytosolic v-SNARE SEC22-GFP, ER luminal RoGFP2-SEC22; 26) revealed that PLS as either an N-or C-terminal roGFP2 fusion resides at the cytosolic side of the ER and is not localized in the ER lumen (Fig. 2p).

PLS interacts with the ethylene receptor protein ETR1
We hypothesized that PLS plays a role in receptor function and investigated whether this involved direct interaction with the receptor complex.Preliminary experiments using yeast 2hybrid analysis suggested that PLS interacts with ETR1 (Fig. S4).Confirmation of the physical interaction between PLS and ETR1 in plants came from co-immunoprecipitation (Co-IP) analysis.
Agrobacterium containing a plasmid encoding PLS linked to a C-terminal GFP and ETR1 with a C-terminal HA tag was infiltrated into Nicotiana benthamiana leaves for transient expression.
After 3 d, interaction was confirmed by western blotting after Co-IP with either anti-GFP beads (showing PLS pulls down ETR1) or anti-HA beads (showing ETR1 pulls down PLS) (Fig. 3a).
GFP-only controls did not show binding with ETR1, demonstrating the interaction is dependent on the presence of the PLS peptide.The addition of 0.5 μM copper sulphate to the protein extract used for Co-IP experiments stabilized the PLS-ETR1 interaction.The presence of copper ions resulted in almost 3-fold more PLS:GFP detected upon pulldowns with ETR1-HA, or conversely of ETR1-HA pulled down with PLS-GFP, compared to the same assay in the presence of the metal chelator 2 mM EDTA (Fig. 3a, b).
To investigate the specificity of PLS binding, synthetic full length PLS peptide PLS(FL) was introduced into the infiltrated N. benthamiana leaves 30 min before the tissue was harvested.
The addition of 25 nM synthetic PLS caused a ca.80% reduction in PLS-GFP binding to ETR1-HA (Fig. 3c, d), suggesting that the synthetic PLS peptide competes for ETR1 binding, and showing the specificity of PLS for ETR1.Interestingly, the anti-GFP beads bound two sizes of PLS-GFP protein (Fig. 3e), both of which were larger than a GFP-only control, suggesting that the PLS peptide undergoes cleavage, a change in conformation, post-translational modification or incomplete reduction of Cys residues on some PLS.When using ETR1-HA to pull down PLS-GFP, only the larger peptide was present (Fig. 3e, f), suggesting that ETR1 binds the full length PLS peptide, but that PLS may be modified after ETR1 binding.
To pinpoint the interaction site at the receptor in more detail, in vitro binding studies were performed with purified receptor variants and PLS by microscale thermophoresis (MST; Fig. 3g).Binding of PLS was observed only with receptor variants containing the N-terminal transmembrane domain (TMD).In contrast, no binding was detected with ETR1 lacking this domain (ETR1 306-738 ).The TMD harbors the ethylene and copper binding region (27).

PLS binds Cu(I) and forms protein-protein interactions with copper chaperones ATX and CCH and with RAN1
Cysteine residues are common metal-ligand binding residues in low molecular weight copperhandling peptides, and predictions of PLS structure suggests a single a-helix plus unstructured region with two cysteines (CX10C arrangement where X is any amino acid), with some analogy to copper-metallochaperones such as Cox17 or other CX9C twin proteins (28).In view of both structural considerations and the copper-dependency of ETR1 we determined whether the two cysteine residues (C6 and C17) in PLS play a functional role, namely in pls mutant complementation and in copper binding.A mutated Arabidopsis full-length peptide in which both cysteines were replaced with serine, PLS(FL C6S, C17S), was non-functional in hydroponic root feeding assays (Fig. 4a).Furthermore, as indicated above, the 9 amino acids sequence N2 and C-terminal sequences C1 and C2 from Arabidopsis PLS, which each contain only one cysteine residue, were unable to rescue the mutant (Fig. 1e).These results indicate the requirement of the two cysteine residues for biological activity.Interestingly, RNA-seq data showed that nineteen out of 287 genes associated with response to metal ion (GO:0010038) were also down-regulated in the pls mutant (Table S4, Fig. S2b, enrichment FDR = 0.000057; 20).
To determine a possible role for PLS and the cysteine residues in binding Cu(I), synthesised PLS peptide (44.8 μM) was titrated with Cu(I) ions under strictly anaerobic conditions and monitored for copper-dependent spectral features.Both UV-vis (putative metal-to-ligand charge-transfer) and fluorescence (putative tyrosine solvent access) spectra of PLS changed as a function of Cu(I) ion concentration, consistent with complex-formation (Fig. 4b,c).The inflection in the UV-vis data at ca. 20 μM and the absence of further quenching in the fluorescence data above ca.22 μM suggest that PLS binds Cu(I) ions in a 2:1 stoichiometry (Figs. 4b).We further titrated PLS peptide against bicinchoninic acid (BCA), which is a chromophore that binds Cu(I) (29) (Fig. 4c).The rationale was to determine whether PLS could compete Cu(I) from BCA, indicative of binding by PLS.PLS peptide was titrated with Cu(I) in the presence of 87 μM BCA.
In the presence of the PLS peptide (44.8 μM), the concentration of Cu(I) required to saturate BCA increased by ca.20 μM, and the initial gradient is shallower than in the control (Figs.4c,d, S5).
Thus PLS withholds Cu(I) from BCA, implying an affinity for Cu(I) (KCu PLS) within an order of magnitude of KCu of BCA, and these data also provide further support for a 2:1 stoichiometry of binding.Titration experiments also show that synthetic PLS(FL) withholds Cu(I) from BCA (Fig. S6), but the mutant PLS (FL C6S, C17S) does not (Fig. S7), confirming a role for those cysteine residues in Cu(I) binding.As an additional and complementary strategy, to eliminate potential solubility problems of PLS in determining accurately the stoichiometry and affinity for Cu(I) (these studies require complete solubility for precise affinity quantification), we also tested a PLS fusion to maltose binding protein (MBP-PLS) which was found to retain solubility when titrated with Cu(I) ions under strictly anaerobic conditions.MBP-PLS (14 μM) was found to withhold ca.7 μM Cu(I) from BCA, indicative of a 2:1 PLS:Cu(I) stoichiometry, and tight binding is cysteinedependent (Fig. 4e).A β2 affinity of 3.79 (±1.5) x10 19 M -2 was determined by competition against an excess of BCA, and the fit significantly departs from simulations 10x tighter or weaker (Fig. 4f, Fig. S8).
Metal binding in biology is challenging to predict because the formation of metalprotein complexes is a combined function of metal affinity for a given protein and metal availability, which would need to be known for Cu(I) in the Arabidopsis cytosol in the case of PLS (Fig. 2p).Cu(I) occupancy of the cytosolic copper chaperone ANTIOXIDANT PROTEIN 1 (ATX1) tracks with fluctuations in available cytosolic Cu(I) such that its affinity approximates to the mid-point of the range of Cu(I) availabilities within this eukaryotic compartment (30,31).
Arabidopsis ATX1 was therefore expressed and purified to determine a 1:1 ATX1:Cu(I) stoichiometry and affinity KD Cu(I) of 5.47 (±0.6) x10 -18 M (Fig. 4g inset, Fig. S9). Figure 4f reveals that the cytosolic concentration of PLS would need to exceed (improbable) millimolar concentrations for Cu(I)-dependent homodimers to form at this cytosolic Cu(I) availability (mathematical models and equations shown in Supplementary text).It is thus unlikely that the Cu(I):PLS2 complex alone delivers Cu(I) to interacting cuproprotein ETR1.Cu(I)-dependent PLS heterodimeric complexes are the more likely functional species.
To investigate possible interaction between PLS and RAN1, PLS was titrated against fluorescently labelled RAN1 and RAN1 truncation mutants for comparison.RAN1, NterRAN1 and CterRAN1 (DeltaRAN1 in Fig. 4h) were purified and labelled as described (17).NterRAN1 is a truncation containing only the two N-terminal metal-binding-domains, whereas CterRAN1 is a construct lacking this region.Using MST, direct interaction was observed between PLS and RAN1.Dissociation constants indicate that PLS interacts predominantly with the N-terminal metal-binding-domains, but only weakly with the C-terminal region (Fig. 4h).Moreover, the soluble copper chaperones ATX1 and CCH also interact with PLS.Titration experiments were carried out with fluorescently labelled copper chaperones ATX1, CCH and CCHΔ, a mutant lacking the plant specific C-terminal extension; though the C-terminus on its own (DeltaN) is not soluble and can thus not be checked for interaction with PLS.PLS interacts with these soluble copper chaperones at similar affinity as obtained for RAN1 (Fig. 4i).Of relevance here is the observation that ATX1 interacts directly with RAN1 to deliver copper at the ER (32).Cu(I)dependent PLS heterodimeric complexes may therefore modulate multiple cuproproteins in the Cu(I)ETR1 maturation pathway.

Discussion
We present evidence of a role for the PLS peptide as a regulator of ethylene receptor function.
Peptides can act as ligands for receptor kinases to regulate signalling pathways, such as in plant immunity (33), development (34) or in response to abiotic stresses (35).We propose that the PLS peptide regulates the ethylene receptor by a different mechanism, i.e. not as a ligand (which for ETR1 is the hormone ethylene) but as part of the metalation mechanism for Cu(I) delivery to the receptor, a process essential for ethylene binding (10).Given that PLS is expressed in specific cell types (19) it seems likely that PLS adds a level of regulation of receptor function in specific tissues, and most obviously in the root tip.While PLS:GFP localizes to several subcellular compartments including cytoplasm and nucleus, some localizes to the ER where the ethylene receptor is found (Fig. 2).Furthermore we demonstrate that PLS can interact directly with the receptor protein ETR1 (and specifically with the N-terminal copper-binding domains of ETR1), with the Cu-transporting P-type ATPase RAN1, and with two Cu chaperones, ATX1 and CCH.
ATX1 and CCH have also previously been shown to bind RAN1 (32).
Many plant peptides involved in signalling are processed from longer pre-proteins (36).
In contrast, PLS is translated from an open reading frame to produce a functional 36 amino acids peptide; the full-length peptide appears to be cleaved or otherwise modified on binding to ETR1 (Fig. 3e).It is possible that the three arginines (amino acids 10-12) may represent a cleavage site (37,38) which would produce the observed shorter GFP fusion protein detected in co-IP experiments (Fig. 3e).The size of the PLS-GFP cleavage product (ca.30 kD) suggests that this peptide fragment represents GFP (ca.27 kDa) plus a 3 kDa fragment of the C-terminal region of PLS.Such cleavage would likely inactivate PLS function as we demonstrate that shorter N-and C-terminal fragments (each with only one cysteine residue) are not biologically active (Fig. 1e).
Such cleavage might represent a mechanism by which PLS releases Cu(I) to the receptor or to other interacting Cu(I)-transporting proteins RAN1, ATX1 and/or CCH.
Previously there has been proposed a model for copper delivery into the cell via the plasma membrane-located HMA5 transporter to RAN1 via ATX1 and/or CCH (18,32).
Arabidopsis has three known copper chaperones, ATX1, CCH and a Cu chaperone for superoxide dismutase CCS (39,40,41).These are required for copper homeostasis, and both ATX1 and CCH have been shown to interact with RAN1, providing a link with ethylene signalling (41,42).
Chaperones are required for the transport of reactive copper to the correct compartment via these chaperones, avoiding cytotoxicity, following import into the cell via copper transporters such as COPT1 and HMA5; the latter being structurally similar to RAN1 (43,44).Loss of function mutants cch and atx1 have no abnormal phenotype when grown under standard conditions, but the atx1 and the atx1 cch double mutant, though not the cch single mutant, are hypersensitive to exogenous Cu (45).ATX1 and CCH functions have likely diverged, a view supported by their differential transcriptional regulation by copper and different cell type specificities (45).The ran1 mutant is however not copper-hypersensitive, while hma5 is; and the pls mutant is, like ran1, not copper hypersensitive (Fig. S10), indicative of a role that is distinct from that of ATX1 or CCH.The ATX1 interaction with RAN1 is indicative of its role in copper transport from the cell surface to internal targets, including RAN1 and potentially to PLS.
We propose that PLS represents a new component of this pathway, providing some tissue-specificity to the Cu(I) delivery process.Given the low probability that PLS would act as a copper chaperone directly due to its predicted low concentration in the cell (Figs.4f, S8), it more likely partners with an existing chaperone such as ATX1 or CCH close to the ER, which would provide a concentrating effect, facilitating Cu(I) delivery to RAN1 and ETR1.In this scenario, a down-regulation of PLS expression or loss-of-function mutation of the PLS gene would lead to a depletion of correctly coordinated Cu(I) at the receptor through reduced intermediate PLS heterodimer formation with ATX1 and/or CCH and ultimately RAN1, in turn leading to a non-functional receptor that cannot activate CTR1.This would be expected to promote ethylene signaling, consistent with the observed effect of ACC treatment on reducing PLS expression and enhancing ethylene responses (Fig. S11).A functional PLS complex may be removed from ETR1 by the observed cleavage or modification of PLS.Future structural studies should reveal more about the PLS-ETR1-RAN1-chaperone interaction and role of copper ions in ethylene signal transduction, which represents a new paradigm for the regulation of signaling protein function by metals.
While the pls mutant shows evidence for a role of PLS in ethylene signalling, the relationship is not simple.For example, the PLS overexpresser is not completely insensitive to the ethylene precursor 1-aminocyclopropane-1-carboxylate (ACC) and does not completely suppress the ctr1 mutation ( 14), suggesting some distinct function compared with other regulatory components.The pls mutant shares (statistically significant) overlap in DEGs with the ctr1 mutant, whereby 11.9% of the ctr1 DEGs are shared with pls (each compared to the respective wild types; Fig. S12).We have also shown in other previous work that there is crosstalk between PLS, ethylene and other signalling pathways, notably auxin, cytokinin and ABA (e.g.21,22,23), and observed 'leakiness' of the PLS overexpression ethylene phenotype, and distinctiveness of pls compared to other ethylene signaling mutants, may be linked to its restricted expression pattern.The root tip, where PLS is most strongly expressed, is a site of high auxin, whereby an auxin concentration maximum is formed around the stem cell niche to regulate both stem cell identity and function (46).Auxin can induce ethylene biosynthesis by upregulation of the ethylene biosynthetic enzyme ACC synthase (47), and high ethylene concentrations can lead to both reduced meristem function and aberrant division of the quiescent centre and cells within the stem cell niche and columella (48).Therefore ethylene biosynthesis and signalling in the root tip must be tightly regulated to allow growth control.PLS expression is transcriptionally activated by auxin and suppressed by ethylene, together forming a signaling network with cytokinin, also produced in the root tip, to regulate ethylene responses and root growth.This network provides a mechanism to suppress growth-inhibitory ethylene responses in the high auxin environment of the root tip, which is dependent on the tissue-specific expression of PLS (21) (Fig. S11) and is mediated through ethylene receptor metalation.

Plant Material
Seedlings of Arabidopsis thaliana ecotype either C24 or Col-0, or pls mutants or transgenic PLS overexpressers, all from lab stocks, were grown on solid sterile half-strength MS10 medium on 2.5% Phytagel (Sigma-Aldrich) in 90 mm Petri dishes (Sarstedt, Leicester, UK) at 21°C, under a 16 hour photoperiod as described (14,19).Nicotiana benthamiana plants (lab stocks) were grown in a controlled environment (21°C, 16 h photoperiod) for transient expression studies in leaf tissue.For hydroponic feeding studies, Arabidopsis seedlings were cultured in liquid medium (1 ml/well) in sterile 24-well plates (Sarstedt), essentially as described (49).For hormone and peptide assays, one seedling was grown in each well, at 21°C, under a 16 h photoperiod.For peptide feeding experiments, purified freeze-dried peptide was dissolved in DMSO to create a 500 μM stock solution.Peptide stock solution was added to liquid ½ MS10 plant media containing 0.1% DMSO to make a final peptide concentration of 50 or 100 nM (or 10, 25, 50 and 100 nM for dose-dependent assays).For copper treatments, 1 mM CuSO4 solution was filter-sterilized and added to autoclaved liquid ½ MS10 plant media to create final CuSO4 concentrations of 0, 5, 10, 15, 20, 25, 30, 35, 40, 45 and 50 μM.The copper chelator bathocuproine disulphonic acid (BCS) was added to liquid medium to produce final concentrations of 0, 10, 50, 100, 250 and 500 μM.Seedlings were scanned to create a digital image and root lengths of seedlings were measured using ImageJ.Statistical analysis was performed using the Real Statistics Resource Pack software (Release 3.8, www.realstatistics.com) in Excel (Microsoft).

Peptide Synthesis
Peptides were either obtained from Cambridge Research Biochemicals (Billingham, UK) or synthesized in the laboratory by Fmoc solid phase peptide synthesis (SPPS) on a CEM Liberty1 single-channel peptide synthesiser equipped with a Discover microwave unit.Reactions were performed in a 30 ml PTFE reaction vessel with microwave heating and agitation by bubbling nitrogen.Peptide synthesis was carried out using 2-chlorotrityl chloride resin (0.122 mmol g -1 , Novabiochem) using Fmoc-protected amino acids.The first amino acid residue at the C-terminus (histidine) was coupled manually by mixing 76 mg (1 eq.) Fmoc-His(Trt)-OH, 0.09 ml (4 eq.) N,N-diisopropylethylamine (DIPEA), 1 ml dichloromethane (DCM) and 1 ml dimethylformamide (DMF) until the amino acid powder had dissolved.The mixture was added to 0.1 mmol resin and stirred gently for 120 minutes at room temperature.Resin was washed with 3x DCM/MeOH/DIPEA (17:2:1), 3x DCM, 2x DMF and 2x DCM.Amino acid coupling reactions were performed using Fmoc-protected amino acids present in a 5-fold excess (2 M concentration), HOBt (0.5 M HOBt in DMF, used at the activator position) and DIC (0.8 M in DMSO, used at the activator base position).For double and triple couplings the reaction vessel was drained after each coupling cycle and fresh reagents were added.Before each coupling, a room temperature preactivation period of 1 to 2 hours was used.Microwave-assisted couplings were performed for 10 minutes at 75°C at 25W power.Cys and His residues were coupled at low temperature (10 minutes at room temperature followed by 10 minutes at 50°C, 25W).Arg residues were double coupled, firstly by 45 minutes at room temperature plus 5 minutes at 75°C (25W), and second by the standard microwave conditions above.Fmoc group removal was carried out by two piperidine solution treatments (20% piperidine in DMF) in succession: 5 minutes then 10 minutes.Peptide cleavage from resin was carried out in 3 ml 95% TFA in dH - 2O/TIPS (2.85 ml TFA, 0.15 ml dH2O, 0.15 ml triisopropylsilane).Peptide was dissolved in water with a small volume of MeCN and lyophilized to produce a powder using a Christ ALPHA 1-2 LDplus freeze dryer.

Preparative High-Performance Liquid Chromatography (HPLC)
Peptide products were analysed and purified by HPLC at 280 nm.25-50 mg of freeze-dried peptide sample was dissolved in 1 ml 1:1 H2O:MeCN and injected onto a Speck and Burke Analytical C18 Column (5.0 μm, 10.0 x 250 mm) attached to a PerkinElmer (Massachusetts, USA) Series 200 LC Pump and 785A UV/Vis Detector.Separation was achieved by gradient elution of 10-80% solvent B (solvent A = 0.08% TFA in water; solvent B = 0.08% TFA in ACN) over 60 minutes, followed by 80-100% B over 10 minutes, with a flow rate of 2 ml/min.Selected peptide fractions were lyophilized and a mass assigned using MALDI-TOF MS.Peptide sequences were identified using MALDI-TOF MS, using an Autoflex II ToF/ToF mass spectrometer (Bruker Daltonik GmBH, Germany) equipped with a 337 nm nitrogen laser.MS data was processed using FlexAnalysis 2.0 (Bruker Daltonik GmBH).

Imaging
CSLM images were obtained by a Leica SP5 TCS confocal microscope using 40X and 63X oil immersion lenses.For propidium iodide staining, whole Arabidopsis seedlings were incubated in 10 mg/l propidium iodide solution for 90 s. pPLS::PLS:GFP, pPLS::GFP and p35S::GFP seedlings were grown for 7 d on Phytagel ½ MS10 medium before ca. 25 mm of the root tip was removed and mounted in dH2O on a microscope slide, and a 1.5 mm cover slip was placed on top prior to imaging.The ER marker p35S::RFP:HDEL (24) (kindly provided by Dr. Pengwei Wang, Durham University), and the trans-Golgi apparatus marker pFGC-ST:mCherry (obtained from Nottingham Arabidopsis Stock Centre, www.arabidopsis.info)were introduced into pPLS::PLS:GFP plants by the floral dip method of transformation (50) using A. tumefaciens GV3 101.ER was also localized using ER Tracker™ Red (Life Technologies).Seven day-old seedlings were stained for 30 min in the dark in liquid ½ MS10 media containing 1 μM ER Tracker™ Red.

RNA Isolation, RNA Sequencing, RT-qPCR
RNA was extracted from 7 day-old seedlings grown on half strength MS10 medium using the Sigma-Aldrich Plant Total RNA Kit (STRN50) and the On-Column DNase I Digestion Set (DNASE10-1SET), essentially as described (51).Tissue was ground in liquid nitrogen before incubation in lysis solution containing 2-mercaptoethanol at 65°C for 3 min.Debris was removed by centrifugation and column filtration and RNA was captured onto a binding column using the supplied binding solution.DNA was removed by wash solutions and DNase treatment on the column.Purified RNA was eluted using RNAase free water.
The Illumina HiSeq 2500 System was used for RNA sequencing of three biological replicate samples, with libraries prepared using the Illumina TruSeq Stranded Total RNA with Ribo-Zero Plant Sample Preparation kit (RS-122-2401), essentially as described (51).
Ribosomal RNA (rRNA) was removed and purified RNA was quality checked using a TapeStation 2200 (Agilent Technology) with High Sensitivity RNA ScreenTape (5067-5579).
mRNA was fragmented into 120-200 bp sequences with a median size of 150 bp and used as template to synthesize first strand cDNA using reverse transcriptase and random primers, followed by second strand cDNA synthesis with DNA Polymerase I and RNase H. Newly synthesized cDNA had a single adenine base added with ligation of adaptors, before being purified and amplified by PCR to make the final library.Library quality control was carried out again using a TapeStation with D1000 ScreenTape (catalog number 5067-5582).RNA-seq data were aligned to the TAIR10 (EnsemblePlants, release 58) genome sequence with corresponding gtf file using STAR (v 2.7.11a; 52) to obtain read count per gene.Read count per gene was analysed using DESeq2 v 1.40.2 (53) to get p values, adjusted p values, and log2 fold changes.
Differentially expressed genes (DEGs) were identified with the criteria of adjusted p value < 0.05.Gene ontology analysis was performed using AgriGO (54) singular enrichment analysis.
For RT-qPCR, RNA was extracted from 7 day-old seedlings (3 biological replicates, 20 mg of tissue) as described (51).Total mRNA was extracted using Dynabeads ® mRNA DIRECT™ kit with Oligo(dT)25 labelled magnetic beads.cDNA was prepared using a SuperScript ® IV First-Strand synthesis system.Samples were checked for the presence of genomic DNA by PCR with ACTIN2 primers ACT2 forward and reverse.Primer sequences were determined using Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Primers are listed in Table S7.

Protein-protein interaction studies Yeast 2-hybrid
The GAL4 two-hybrid phagemid vector system was used to detect protein-protein interactions in vivo in yeast, using the reporter genes β-galactosidase (lacZ) and histidine (HIS3) in the YRG-2 yeast strain, essentially as described (55).DNA that encodes the target (ETR1) and bait (PLS) was inserted into the pAD-GAL4-2.1 A and pBD-GAL4 Cam phagemid vectors respectively and expressed as hybrid protein.The hybrid proteins were then assayed for protein-protein interaction.
DNA encoding the target and bait proteins were prepared by PCR amplification using primer designed the specifically for the DNA encoding the target (ETR1) and bait (PLS).Each set of primer contained specific endonucleases on the ends of primer corresponding to the endonucleases in the MCS of pAD-GAL4-2.1 A and pBD-GAL4 Cam phagemid vectors.The DNA construct of the target (ETR1) and bait (PLS) with specific restriction sites on the ends was then transformed into the TOPO 2.1 vector to check the sequence of the amplified DNA by sequencing with M13 forward (CTG GCC GTC GTT TTA C) and M13 reverse (CAG GAA ACA GCT ATG AC).The two vectors, pAD-GAL4-2.1 and pBD-GAL4 Cam were digested using specific restriction endonucleases and dephosphorylated prior to ligating the insert DNA.
The DNA encoding the target (ETR1) and bait (PLS) was then ligated into the same reading frame as the GAL4 AD of the pAD-GAL4-2.1 phagemid vector and the GAL4 BD of the pBD-GAL4 Cam phagemid vector.
The following primers were used for PCR amplification: ETR1: Forward primer GAA TCC ATG GAA GTC TGC AAT TGT A (Eco RI on 5' end) Reverse primer GTC GAC TTA CAT GCC CTC GTA CA (Sal I on 5'end) PLS: Forward primer CTG GAG ATG AAA CCC AGA CTT TGT (Xho I on 5' end) Reverse primer GTC GAC ATG GAT TTT AAA AAG TTT (Sal I on 5' end) The pGAL4 control plasmid was used alone to verify that induction of the lacZ and HIS3 genes has occurred and that the gene products are detectable in the assay used.The pLamin C control plasmid was used in pairwise combination with the pAD-WT control plasmid or with the pAD-MUT control plasmid to verify that the lacZ and HIS3 genes are not induced as the protein expressed by each of these pairs do not interact in vivo.
The control plasmids were transformed into the YRG-2 strain prior to the initial transformation of the bait and the target plasmids.The control plasmids were used separately or in pairwise combination in the transformation of the YRG-2 yeast strain.The yeast competent cells were cotransformed with the bait and target plasmids by sequential transformation.First, yeast were transformed with the bait plasmid and assayed for expression of reporter genes.
Second, yeast competent cells containing the bait were prepared and transformed with the target plasmid.

Bimolecular fluorescence complementation analysis (BiFC)
BiFC was carried out essentially as described previously (56).Full-length Arabidopsis ETR1 and PLS cDNA sequences were cloned respectively into the vectors pDH51-GWYFPn (AM779183, to form ETR1-YFPn) and pDH51-GW-YFPc (AM779184, to form PLS-YFPc), and the CTR1 cDNA was cloned into pDH51-GW-YFPc (CTR1c), as a control for ETR1 interactions.Intact YFP plasmid was also used as a positive control and YFPc alone was used as a negative control as described previously (57).Plasmids were kindly provided by Prof. Don Grierson, University of Nottingham.Transient expression studies following microprojectile bombardment on onion cells.Plasmids were adhered to gold particles to make gold-coated cartridges.5-10 of these cartridges were used for bombarding onion peel cells.Agar plates containing bombarded onion peels were incubated for 8 hours in dark and then a 1 cm section of the peel was stained with propidium iodide (10 mg/ml) for 1 min and viewed under confocal microscope.Experiments were repeated at least three times.

Co-immunoprecipitation
To investigate the interaction between the PLS peptide and the ethylene receptor ETR1, two DNA constructs were created by Gateway cloning.The 105-nucleotide PLS gene (without the stop codon) was inserted into the pEarlyGate103 (pEG103) destination vector, containing the p35S promoter and a C-terminal GFP tag, producing a vector containing the p35S::PLS:GFP DNA.The ETR1 cDNA was inserted into the pEarlyGate301 (pEG301) vector to create a p35S::ETR1:HA construct, producing an ETR1 protein with a C-terminal HA tag.

Infiltration into Nicotiana benthamiana
The transient expression of constructs in Nicotiana benthamiana (tobacco) leaves was based on a previously published method (57).Experiments were replicated up to 5 times.Competent Agrobacterium tumefaciens GV3101 cells were transformed with the desired plasmid containing the gene of interest.Individual colonies were used to inoculate liquid LB cultures containing 25 μg/ml gentamicin, 50 μg/ml rifampicin and the specific antibiotic required to select for the desired plasmid.The liquid cultures were grown at 28°C for 14-16 h with shaking at 220 rpm.
Additionally, liquid cultures of GV3101 containing the p19 protein that is encoded by the tomato bushy plant virus were also prepared in order to suppress post-transcriptional gene silencing (57).
The overnight cultures were grown until an OD600 of approximately 0.6 was reached, and then centrifuged at 3000 x g for 5 min.These cells were then twice washed with 10 ml of an infiltration buffer containing 10 mM MgCl2.6H20, resuspended in 1 ml of the same solution and subsequently incubated at room temperature for 3-5 h.Prior to infiltration, each construct was mixed with p19 and infiltration buffer in a 1:1.2:1.8 ratio.
Several small cuts were made with a scalpel on the abaxial surface of the N. benthamiana leaves and were subsequently injected with each of the constructs using a syringe.
The plants were approximately 7 to 10 weeks old; the chosen leaves were healthy and of length 3-6 cm, and 3 to 4 leaves were infiltrated with each construct.

Protein extraction and PLS/ETR1 co-immunoprecipitation
Total protein was extracted from the infiltrated leaves of N. benthamiana plants 3 d after infiltration for co-immunoprecipitation (co-IP/pull-down) experiments to investigate the interaction between PLS and ETR1 proteins.
1.5 g of leaf tissue was harvested from each A. tumefaciens construct infiltration event, frozen with liquid nitrogen and ground gently using a mortar and pestle.For competition assays, ChromoTek (Planegg, Germany) anti-GFP beads were used to immunoprecipitate the PLS-GFP protein, and Sigma-Aldrich (St. Louis, USA) anti-HA beads for the HA-tagged ETR1.
25 μl bead slurry was resuspended in 500 μl ice-cold dilution buffer (20 mM sodium phosphate pH 7.4, 100 mM NaCl, 80 mM KCl, 1% glycerol, 0.1 % Triton, 10 mM DTT, plus 1 mini protease inhibitor cocktail tablet) and centrifuged for 2 minutes at 2500 x g at 4°C.The supernatant was discarded and the beads were washed twice more with 500 μl ice-cold dilution buffer.
The supernatant from the protein sample extraction from N. benthamiana plants was mixed with 50 μl GFP beads or HA beads and incubated for 30 minutes at 4°C, mixing every 2 minutes.The mixture was centrifuged at 2500 x g for 2 min at 4°C, washed twice with 500 μl ice-cold dilution buffer, and the beads were transferred to a new microcentrifuge tube.The target protein was eluted with the addition of 100 μl 2x SDS sample buffer (120 mM Tris pH 6.8, 50 mM 4% (w/v) SDS, 20% (v/v) glycerol) and the sample was boiled for 10 minutes at 95°C to dissociate immunocomplexes from the beads.The mixture was centrifuged for at 2500 x g for 2 minutes at 4°C to separate the beads, and the supernatant was transferred to a new microcentrifuge tube.The supernatant was used in SDS-PAGE analysis.

Sodium Dodecyl Sulphate Polyacrylamide Gel Electrophoresis (SDS-PAGE)
SDS-PAGE was used to separate protein fragments.The complexed proteins from the pull-down assay were analysed on 10-12% acrylamide gels.
Firstly, the resolving gel was prepared by adding the chosen amount of acrylamide (ProtoGel, 30% (w/v) acrylamide, 0.8% (w/v) bisacrylamide solution, National Diagnostics) to the resolving buffer (0.1% (w/v) SDS, 375 mM Tris, polymerized via the addition of 0.1% (v/v) ammonium persulphate solution (APS) and finally set by the addition of 1.4 µl/ml TEMED (NNN'N'-tetramethylethylenediamine).The stacking gel was then prepared again by adding the appropriate amount of acrylamide to the stacking buffer (consisting of 0.1% w/v SDS, 125 mM Tris).Polymerization was activated by adding 0.1% (v/v) APS and set using 4 µl/ml TEMED.

Western Blotting
Following electrophoresis, the SDS gels were first washed in 1x transfer buffer (0.04% (w/v) SDS, 20% (v/v) methanol, 38 mM glycine, 48 mM Tris) for 5 minutes.The proteins were then transferred overnight onto nitrocellulose membranes (Whatman, GE Healthcare Life Sciences, Buckinghamshire, UK) in a 1 litre tank containing transfer buffer at 30 V.
Excess secondary antibody was removed again by washing three times in 1x TBST, as with the primary antibody.In order to visualize the probed blot, the membrane was incubated with ECL Western Blotting Detection Reagent immediately prior to imaging.The horseradish peroxidase quartz cuvettes (Helma), and after addition of probe to the concentration specified, titrated with CuSO4.After each addition, solutions were thoroughly mixed and absorbance spectra recorded using a Lambda 35 UV/Vis spectrophotometer (Perkin Elmer).Titration isotherm data was fitted using simulated affinity curves using DynaFit (58).

Determination of dissociation constants for the PLS-ETR1 interaction
Full-length ETR1 and truncation mutants were purified and labelled as described in (59).94 µM PLS were diluted serially in 50 mM Tris and 300 mM NaCl (pH 7.6).Fluorescently labelled receptor was added at a final concentration of 50 nM.Thermophoretic behaviour was measured in premium capillaries at 50 % LED and 50 % MST power.In case of a binding event, data were fitted using GraphPad Prism 5.

Interaction studies of PLS with copper transporter RAN1 and copper chaperones by microscale thermophoresis
PLS was titrated to fluorescently labelled RAN1 and RAN1 truncation mutants.RAN1, NterRAN1 and CterRAN1 were purified and labelled as described in (18).NterRAN1 is a truncation containing only of two N-terminal metal-binding-domains, whereas CterRAN1 is a construct lacking this region.Direct interaction between PLS and RAN1 was observed using Microscale Thermophoresis.Dissociation constants indicate that PLS interacts predominantly with the N-terminal metal-binding-domains, but only weakly with the C-terminal region.
Dissociation constants were determined using GraphPad Prism 5. Titration experiments were also carried out with fluorescently labelled copper chaperones ATX, CCH and CCHΔ, a mutant lacking the plant specific C-terminal extension.Proteins were purified and labelled as described (18).

Ratiometric analysis of roGFP2 fusion proteins
N-and C-terminal roGFP2 fusion proteins of PLS were generated by Gateway cloning.
Infiltration and transient expression of roGFP2 fusions and control proteins were carried out as described (26).Image acquisition and data analysis were carried out as described in (18).A minimum of 10 leaf optical sections were imaged and used for ratiometric analysis of the redox sensitive excitation properties of roGFP2.

Statistical methods
For gene expression analysis and growth assays, a minimum of three biological replicates was used -see Supplementary Materials, Figure legends, Supplementary Tables (20).For proteinprotein interaction and protein localization studies, assays were carried out independently between 2 and 10 times (Figs. 2, 3 legends).At least three biological replicates were used for plant growth assays, RNA-seq and gene expression (RT-qPCR) experiments, metal binding assays (Fig. 4).Normalised values from at least three biological replicates were then used for one-or two-way analysis of variance (ANOVA) where appropriate and indicated in relevant

5 nM or 25
nM full-length PLS peptide was also infiltrated in the presence of 50 μM MG-132 (a proteasome inhibitor) 30 min prior to tissue freezing.The homogenate was transferred to a precooled microcentrifuge tube. 2 ml of extraction buffer was added (20 mM sodium phosphate pH 7.4, 100 mM NaCl, 80 mM KCl, 1% glycerol, 0.1 % Triton, 10 mM DTT, plus 1 mini protease inhibitor cocktail tablet, Roche, Switzerland) per 20 ml of extraction buffer, and the extra addition of either 2 mM EDTA or 0.5 μM CuSO4 for binding studies), and the solution was ground further and vortexed until the homogenate was smooth.The solution was centrifuged for 12 min at 14000 x g, 4°C.

Figure legends .
Figure legends.Error bars are defined in Figure legends, where relevant.

Figure 1 .
Figure 1.The PLS peptide is required for ethylene control of seedling growth, is structurally and functionally conserved, and complements the Arabidopsis pls mutant.(a).Amino acid sequence of the PLS peptide from Arabidopsis thaliana with Camelina sativa PLS sequence (C.s), and synthetic truncations N1, N2, C1 and C2, indicated by horizontal lines.Two cysteine residues are highlighted in bold.(b) Wildtype (left) and pls mutant (right); bar = 5 mm.(c) Expression levels of 24 ethylene-responsive genes in pls and PLS overexpressing (PLSox) seedlings, compared to wildtype levels.Data are expressed as log2-fold changes in pls mutant and PLSox seedlings compared wildtype with respective significance levels (P-adj value) provided in Table S1.A P < 0.05 and log2-fold of change of ± 0.5 was chosen to identify differentially expressed genes.(d) Effect of Arabidopsis PLS full length peptide, A.t. PLS(FL), and Camelina PLS peptide (C.s.PLS) on Arabidopsis primary root length.Wildtype (C24; blue bars) and pls mutant seedlings (red bars) were grown hydroponically in the presence (100 nM) or absence of peptide for 10 d.(e) Effect of PLS full length and truncated peptides on wild type (blue bars) and pls mutant (red bars) Arabidopsis primary root length.Seedlings were grown hydroponically in the presence of 50 nM peptide for 10 d.C1 = C-terminal 14 amino acids, C2 = C-terminal 24 amino acids, N1 = N-terminal 22 amino acids, N2 = N-terminal 9 amino acids, full length PLS = 36 amino acids.Error bars show ± 1 standard error, n = 25.

Figure 2 .
Figure 2. PLS localizes to the endoplasmic reticulum.(a-o) PLS::PLS:GFP fusion protein (a, g, j) colocalizes with endoplasmic reticulum markers ER Tracker (b, c, c inset) and RFP:HDEL (h, i, i inset), but free GFP does not (d-f, m-o).PLS:GFP staining is seen also in nuclei (n).PLS::PLS:GFP (j) does not co-localize with the trans-Golgi markers ST-mCherry SH:GFP (k, l).Scale bars = 25 μm (c, l), 10 μm (f, i, o).Root epidermal cells in the transition zone were imaged.(p) Ratiometric analysis of roGFP2 fusion constructs transiently expressed in N. benthamiana.Comparison of excitation ratios of PLS-roGFP2 and roGFP2-PLS with control constructs (free roGFP, SEC22 fusions) reveals that PLS localizes to the cytosolic side of the ER.

Figure 3 .
Figure 3. PLS interacts with the ethylene receptor ETR1.(a) Co-immunoprecipitation of PLS:GFP by ETR1:HA (upper panel) in leaves of Nicotiana benthamiana, in the presence and absence of 0.5 μM CuSO4 and EDTA (to remove Cu).Lower panels show presence of ETR1:HA in extracts using anti-HA antibody, and protein input blots using anti-GFP and anti-Rubisco.(b) Densitometric scan of immunoblot.(c) Competition assay showing a reduced binding between PLS:GFP and ETR1:HA in the presence of 0, 5 nM or 25 nM PLS peptide, in the presence of 0.5 μM CuSO4 and 50 μM MG-132, a proteasome inhibitor (upper panel).Lower panel shows ETR1:HA in extracts using anti-HA antibody.(d) Densitometric scan of immunoblot.(e) Co-immunopreciptation of ETR1:HA (upper panel) in leaves of N. benthamiana, showing the effect of EDTA (to remove Cu) on interaction between ERT1:HA and PLS:GFP.Lower panels shows presence of ETR1:HA in extracts using anti-HA antibody, and protein input blots using anti-ETR1-HA and anti-Rubisco.(f) Densitometric scan of immunoblot.a-GFP, anti-GFP antibody; a-HA, anti-HA antibody.(g) Microscale thermophoresis binding curves of different ETR1 truncations with PLS.Binding of PLS was observed with full-length ETR1 and all C-terminal truncations but not with ETR1 306-738 lacking the N-terminal transmembrane part of the receptor.

Figure 1 .
Figure 1.The PLS peptide is required for ethylene control of seedling growth, is structurally and functionally conserved, and complements the Arabidopsis pls mutant.(a).Amino acid sequence of the PLS peptide from Arabidopsis thaliana with Camelina sativa PLS sequence (C.s), and synthetic truncations N1, N2, C1 and C2, indicated by horizontal lines.Two cysteine residues are highlighted in bold.(b) Wildtype (left) and pls mutant (right); bar = 5 mm.(c) Heat maps showing expression levels of 32 ethylene-responsive genes in pls (left panel) and of 58 genes in PLS overexpressing (PLSox) seedlings (right panel), compared to wildtype levels.Data for 3 biological replicates are shown, and are expressed as log 2 -fold changes in pls mutant and PLSox seedlings compared wildtype with respective significance levels (P-adj value) provided in Tables S1 and S2.A P < 0.05 and log 2 -fold of change of ± 0.5 was chosen to identify differentially expressed genes.(d) Effect of Arabidopsis PLS full length peptide, A.t. PLS(FL), and Camelina PLS peptide (C.s.PLS) on Arabidopsis primary root length.Wildtype (C24; blue bars) and pls mutant seedlings (red bars) were grown hydroponically in the presence (100 nM) or absence of peptide for 10 d.(e) Effect of PLS full length and truncated peptides on wild type (blue bars) and pls mutant (red bars) Arabidopsis primary root length.Seedlings were grown hydroponically in the presence of 50 nM peptide for 10 d.C1 = C-terminal 14 amino acids, C2 = Cterminal 24 amino acids, N1 = N-terminal 22 amino acids, N2 = N-terminal 9 amino acids, full length PLS = 36 amino acids.Error bars show ± 1 standard error, n = 25.