Col-0 was used as wild type. T-DNA mutants were all in Col-0 background and provided either by the SALK collection (Alonso et al., 2003), herk1 (SALK_008043), herk2.1 (SALK_105055), herk2.2 (SALK_107146), the SAIL collection (Sessions et al., 2002), the1-4 (SAIL_683_H03), and the GABI-Kat collection (Kleinboelting et al., 2012), fer-4 (GABI_GK106A06). The loss-of-function allele, the1-6, was isolated in a suppressor screen of ctl1-1/pom and described in Merz et al. (2017). The THE1:GFP overexpression line was described in Hématy et al. (2007).

Seeds were surface-sterilized in 5% sodium hypochlorite and rinsed three times with sterile deionized water and then transferred to nutrient agar medium plates containing 1/10 strength Hoagland salts, 1% (w/v) sucrose and 1% (w/v) agar (Duchefa). For metal ion treatments metal salts were added after autoclaving to final concentrations of 10 μM CdCl₂, 5 μM CuSO₄, 15 μM NiSO₄, 100 μM Pb(NO₃)₂, and 100 μM ZnSO₄. After 2 days of imbibition at 4°C in the dark, plates were vertically incubated in a growth chamber at 22°C with constant light (80 μmol m² s^-1). For measurements of etiolated hypocotyls, plates were wrapped in aluminum foil after exposure to light for 5 h. Isoxaben treatment for gene expression analyses was essentially done as described in Merz et al. (2017). Ni treatment for gene expression analyses was done for 6 h on 4 days after germination (dag) etiolated seedlings.

For pH assays sterilized seeds were placed on nutrient agar plates as described above supplemented with 150 μM BCP and metal ions in indicated concentrations. Plates were scanned on day 5 after germination.

The plates were scanned on days 3, 4, and 5 after germination for root growth and on day 5 after germination for hypocotyl measurements. The lengths were evaluated with the ImageJ software by freehand tracking.

Total RNA of 10–12 day old seedlings germinated and grown on control and metal ion supplemented medium were snap frozen and isolated using a LiCl/CTAB method. After grinding roughly 100 mg frozen seedlings 1 mL of pre-heated RNA extraction buffer (2% [w/v] hexadecyltrimethylammonium bromide, CTAB; 2% [w/v] polyvinylpyrrolidone, PVP; 100 mM Tris/HCl pH 8.0; 25 mM EDTA; 2 M NaCl; 0.5 g/L spermidine and 2.7% [v/v] 2-mercaptoethanol) was added, mixed and incubated at 65°C for 5 min. CTAB was removed through two times separation with 1 mL of ice-cold chloroform: isoamylalcohol (24:1) and centrifugation at 4°C. RNA in the supernatant was precipitated with 250 μL 10 M LiCl at 4°C for more than 1.5 h. After centrifugation and EtOH washes the pellet was dissolved in 20 μL RNase free water and stored at -80°C. RNA was quantified with the Qubit (Invitrogen) and the NanoDrop systems. RNA integrity was controlled on Agilent 2100 Bioanalyzer using RNA Nanochip. RNA samples were treated with DNaseI using the TURBO DNA-free^TM kit (Ambion). cDNAs were synthesized from 4 μg total RNA using the Transcriptor reverse transcriptase (Roche) with 500 pmol of oligo(dT)₁₈ primer. The reaction was stopped at 85°C for 5 min without further treatment according to the manufacturer’s instructions. cDNA was diluted 20 times with distilled water and tested by PCR using specific primers flanking an intron sequence to confirm the absence of genomic DNA contamination. cDNA synthesis for the THE1 downstream gene expression was done as described in Merz et al. (2017).

AP2M (ADAPTOR PROTEIN-2 MU-ADAPTIN) has been validated to be the most stably expressed gene among eight tested and was used to normalize the RT-qPCR data (Gutierrez et al., 2012). CT and PCR efficiency (E) values were used to calculate expression using the formula E_T(^CT_ctr^-CT_m)/E_R(^CT_ctr^-CT_m), where T is the target gene and R is the reference gene, CT is the crossing threshold value, m refers to cDNA from the metal ion treated seedlings, and ctr refers to cDNA from the control medium. All RT-qPCR results presented are means from three independent biological replicates and for each independent biological replicate, the relative transcript amount was calculated as the mean of three technical replicates, using the method for calculation of SE values in relative quantification recommended by Rieu and Powers (2009). RT-qPCR for the THE1 downstream gene expression was done as described in Merz et al. (2017). Primers used for the RT-qPCRs are summarized in Supplementary Table 2.

In an expression survey for genes responsive to the metal ions Cd, Ni, and Pb in seedlings, the CrRLK1L members THE1, FER, and HERK1 exhibited a remarkable expression pattern. All three were induced by Ni, only slightly for FER but strongly for HERK1 and THE1 (Figure 1). Pb did not induce any transcriptional regulation whereas Cd triggered upregulation of THE1 and HERK1.

Growth and cell expansion responses of germinating Arabidopsis seedlings to elevated metal ion concentrations depend strongly on the composition of the growth medium and on the cation exchange capacity of the agar. The high cation concentration such as provided by the frequently used 1x Murashige and Skoog medium (Murashige and Skoog, 1962) interferes with the uptake of additional metal ions. On 1x MS medium 400 μM Zn and 40 μM Cd were needed to observe a slight root length reduction for Zn and a 30% reduction for Cd (Kobae et al., 2004). A comparable experiment with germinating seedlings on 1/10 Hoagland medium supplemented with metal ions needed only 50 μM Zn for a 43.2% and 2 μM Cd for a 57.6% root growth inhibition (Tennstedt et al., 2009). Since the MS medium is also Fe- satiated and Cu-deficient all metal ion experiments were performed on 1/10 Hoagland medium (Supplementary Table 1).

Root growth depends on two parameters, the rate of cell division and cell expansion, while dark stimulated hypocotyl growth depends solely on cell expansion. To separate the cell division from cell expansion effects, root growth was determined on seedlings germinating and grown for 5 days in the light and hypocotyl elongation was assessed on 5 dag dark-grown seedlings. In both sets of experiments identical 1/10 Hoagland media supplemented with either 10 μM Cd, 5 μM Cu, 15 μM Ni, 100 μM Pb, or 100 μM Zn were used.

Generally, hypocotyl elongation is less sensitive to metal ions (Figures 2A–C). While on 10 μM Cd root growth was reduced by 70% in relation to control medium the growth inhibition of hypocotyls was only 43%. Similar results were obtained for 5 μM Cu with 46% root and 30% hypocotyl inhibition, 100 μM Zn with 47% root and 41% hypocotyl inhibition, 15 μM Ni with 74% root and 25% hypocotyl inhibition and 100 μM Pb with 78% root and 49% hypocotyl inhibition (Figure 2C). The most dramatic difference between root and hypocotyl growth inhibition was seen for Ni followed by Pb and Cd (Figure 2C) indicating that these metal ions might influence cell division more severely than cell expansion. It might also be possible that cell expansion in the root is differently affected by these metal ions than in etiolated hypocotyls. The difference between root and hypocotyl growth is particularly true for Ni where root growth rate was continuously declining and nearly stopped 5 days after germination (dag) (Figure 2D). Pb shows the opposite effect and root growth recovered starting at 5 dag and roots reached nearly the same length as wild type at 11 dag (Figure 2E).

Apart of the dissimilar effects of metal ions on cell division and cell expansion another explanation for the differing sensitivities between root and hypocotyl growth would be that metal ion transport to hypocotyls is less effective. Data on metal ion concentrations in separated roots and hypocotyls and quantifications of the cell division rate and cell sizes would resolve these possibilities. Although the results of these analyses would contribute to the understanding of the root/hypocotyl response differences, the focus of this study was to determine if members of the CrRLK1L family are involved in mediating growth responses to metal ions.

Previous growth assays with single mutants impaired in HERK1 and HERK2 expression did not show any obvious growth phenotype. But the mutants developed strong cell elongation defects in combination with the the1-4 allele (Guo et al., 2009a,b). However the1-4 turned out to be a hypermorphic allele. Thus the growth defects of the double and triple mutants indicate that HERK1 and HERK2 act antagonistically and not redundantly to THE1 (Merz et al., 2017). On 1/10 Hoagland control medium, roots of herk1 seedlings grow similarly to wild type in contrast to both herk2 alleles, which exhibited a significantly faster root growth (Figure 3A and Table 1).

On metal ion supplemented medium root growth of herk1 seedlings is stronger inhibited upon Cd, Ni, Cu, and Zn than wild type. Since the two herk2 alleles already grow faster on control medium, analyzing the effect of metal ions on these mutants by only looking at the absolute root length data might be misleading. Therefore, their root growth on metal ions was normalized to control conditions (Figure 3B and Table 1). These analyses revealed that the apparent higher tolerance to Ni, and Zn diminished and only the reduced lead response remained while an enhanced Cd, Cu, and Ni sensitivity was revealed. In summary, both HERK genes are likely to be important to support root growth upon higher Cd, Cu, and Ni concentrations while the Zn response is HERK1 specific and the Pb response is HERK2 specific.

As with wild type the hypocotyl elongation in the mutants differs from root growth responses. Already on control medium herk1 elongates significantly less than wild type and the two herk2 alleles (Figure 3C and Table 1). Therefore, normalization to control conditions was necessary to reveal that hypocotyl elongation of herk1 was less disturbed by Cd, Ni, Zn, and Pb than wild type (Figure 3D and Table 1). Identical responses were found in the two herk2 mutants (Figures 3C,D and Table 1). In conclusion, both HERK genes appear to mediate the inhibition of hypocotyl elongation upon Cd, Ni, Zn, and Pb while they support root growth on Cd, Cu, and Ni.

THE1 signaling negatively affects cell expansion upon inhibition of cellulose synthesis (Hématy et al., 2007; Merz et al., 2017). Therefore THE1 is the strongest candidate to uncover metal ion induced cell wall damage responses. Well-characterized loss-of-function and gain-of-function alleles were available to compare their responses to metal ions.

Already on control medium root growth of THE1-related knock-out and overexpression lines behaved inversely (Figure 4A and Table 1). While the loss-of-function mutant the1-6 developed shorter roots, roots of the THE1:GFP overexpressor grew faster than wild type. Similar contrasting root growth responses were significant on Cd and Ni. On Cu only THE1:GFP and the hypermorphic allele the1-4 developed shorter roots as did the1-6 on Zn (Figure 4B and Table 1).

We previously showed that the hypermorphic the1-4 expresses a transcript encoding a predicted membrane-associated truncated protein lacking the kinase domain. Differences between THE1:GFP and the hypermorphic the1-4 allele, as seen on control, Ni and Zn, possibly relate to the missing cytoplasmic domain in the1-4 which might be important for feedback regulation (Merz et al., 2017).

Hypocotyl elongation in the dark behaved different to roots for the1-6 on control Cd, and Pb, for THE1:GFP on control, Cu and Zn and for the1-4 on all conditions (Figures 4C,D and Table 1). While in roots the strongest effect was shown by the1-6 on Cd, for hypocotyls it was on Ni by the1-4 and THE1:GFP (Figure 5A). The opposing effects of the loss- and gain-of function alleles on hypocotyl elongation were even more distinct on 20 and 30 μM Ni (Figures 5B,C). These results indicate that THE1 mediates Ni specific inhibition of hypocotyl elongation in the dark and promotes Cd specific growth tolerance in roots.

As mentioned above THE1 is known to inhibit cell elongation in hypocotyls upon cell wall damage. Therefore, we examined if Ni induces cell wall damage specific responses such as the expression of the downstream gene in the THE1 signaling pathway, EDGP/At5g19110. The expression of At5g19110 was strongly induced in the overexpressor and the hypermorphic alleles and did not respond in the loss-of-function allele, the1-6 in chemically induced cell wall damage by treatment with the cellulose biosynthesis inhibitor isoxaben. Upon Ni treatment a very low expression in the different THE1 alleles was quantified with no significant alterations to wild type (Figure 5D). These data suggest that Ni perception by THE1 is independent of cell wall damage specific gene expression responses.

FER is the best studied member of the CrRLK1L gene family. FER has been identified due to its crucial function during pollen-tube ovule interaction (Escobar-Restrepo et al., 2007) but is also important for bacterial pathogen interaction (Keinath et al., 2010; Stegmann et al., 2017), mechanosensing (Shih et al., 2014) and in the antagonistic cross-talk between ABA and auxin in vegetative tissues (Yu et al., 2012). Mutants of FER exhibit diverse pleotropic phenotypes. They are stunted, develop smaller leaves (Guo et al., 2009b) and shorter etiolated hypocotyls which are hypersensitive to ethylene and insensitive to brassinosteroids (Deslauriers and Larsen, 2010). Moreover fer mutants develop shorter root hairs (Duan et al., 2010; Du et al., 2016) and larger seeds (Yu et al., 2014) and exhibit a reduced gravitropic response which correlates with less apoplastic alkalization at the lower side of gravistimulated roots (Barbez et al., 2017).

In our growth analyses root lengths of 5 days old fer-4 seedlings were significantly shorter on control medium and in most metal ion treatments except on Pb (Figure 6A and Table 1). However when root growth was normalized to control conditions fer-4 was hypersensitive to Zn and Cu while it was more tolerant to Pb (Figure 6B and Table 1). In concordance with published data etiolated fer-4 hypocotyl reached only 25% of the size of wild type (Figure 6C and Table 1). Thus the different effects of the metal ions were revealed after normalization to control conditions (Figure 6D and Table 1) and showed that fer-4 is hypersensitive to Ni and has a higher tolerance to all other tested metal ions.

FER binds and is activated by secreted peptides of the RALF family (Haruta et al., 2014) triggering alkalinization of the apoplast. This alkalinization is thought to counteract the acidification crucial for cell expansion. Light grown fer-4 seedlings acidified a bath medium faster than wild type which is indicative for a higher plasma membrane H⁺-ATPase activity (Haruta et al., 2014).

Since growth of etiolated hypocotyl depends on rapid cell expansion we tested the acidification capacity of the CrRLK1L mutants on control medium and upon metal ion treatments (Figures 7A–C). All genotypes acidified the medium supplemented with the pH indicator bromocresolpurple (BCP) (Figure 7D). Only fer-4 exhibited a remarkable alkalinization around the hypocotyl while the root was still acidifying the medium (Figure 7B). This alkalinization might be one of the reasons why etiolated hypocotyls of fer-4 are short and indicate that the plasma membrane H⁺-ATPase activity is differently regulated from roots. The phenotype further suggests that FER is necessary for the acidification of etiolated hypocotyls. It is conceivable that in the absence of FER RALF expression is triggered and another receptor in the hypocotyl mediates the alkalinization. The alkalinization surrounding the fer-4 etiolated hypocotyl was still visible on medium supplemented with metal ions. Hypoctoyl mediated alkalinization of the medium is strongest on control medium, followed by Cu, Ni, and only slightly on Cd and Zn. Since, the medium supplemented with Pb has a lower pH itself the alkalinization takes longer. The only other genotype with reduced acidification capacity was the overexpressor THE1:GFP on Ni (Figure 7C and Supplementary Figure 1). Similar to fer-4, THE1:GFP developed shorter hypocotyls on Ni. It might be possible that overexpression of THE1:GFP increases the sensitivity to Ni induced hypocotyl specific RALFs and that THE1 might be involved in their signaling pathway.