Abstract
Rooting cells and pollen tubes – key adaptative innovations that evolved during the colonization and subsequent radiation of plants on land – expand by tip-growth. Tip-growth relies on a tight coordination between the protoplast growth and the synthesis/remodeling of the external cell wall. In root hairs and pollen tubes of the seed plant Arabidopsis thaliana, cell wall integrity (CWI) mechanisms monitor this coordination through the Malectin-like receptor kinases (MLRs) such as AtANXUR1 and AtFERONIA that act upstream of the AtMARIS PTI1-like kinase. Here, we show that rhizoid growth in the early diverging plant, Marchantia polymorpha, is also controlled by an MLR and PTI1-like signaling module. Rhizoids, root hairs and pollen tubes respond similarly to disruption of MLR and PTI1-like encoding genes. Thus, the MLR/PTI1-like signaling module that controls CWI during tip-growth is conserved between M. polymorpha and A. thaliana suggesting it was active in the common ancestor of land plants.
Introduction
When plants first colonized terrestrial habitats more than 470 million years ago, they had to cope with a relatively different, water-deprived environment, that represented a major driving force for a series of key adaptations1–3. Among these, the emergence of rooting cells, such as rhizoids and root hairs4, that facilitate substrate anchorage and water/nutrient uptake, and the subsequent evolution of pollen tubes allowing for an efficient water-independent sperm transport for fertilization in seed plants5 use tip-growth for rapid expansion. Tip-growth is an extreme form of polar growth and as such, it relies on a tight coordination in both, time and space, between the turgor pressure-driven deformation/loosening of the preexisting cell wall and the secretion of new plasma membrane and cell wall material to allow for expansion6. It has become apparent that plant cells have developed cell wall integrity (CWI) maintenance mechanisms to control this coordination and avoid growth cessation or rupture. Research on tip-growing cells of the seed plant model Arabidopsis thaliana has revealed CWI signaling pathways that control pollen tube and root hair growth (reviewed in 7–11). CWI maintenance during pollen tube growth is controlled by two pairs of functionally redundant Malectin-like receptor kinases (MLRs) named AtANXUR1/2 (AtANX1/2)12, 13 and AtBUDDHA’S PAPER SEAL1/2 (AtBUPS1/2)14. These pollen-expressed MLRs can interact with each other and are thought to form a receptor complex for the peptide ligands AtRAPID ALKALINIZATION FACTOR 4 (AtRALF4) and AtRALF19 14, 15. Downstream of these MLRs, CWI is regulated by reactive oxygen species (ROS)-producing NADPH-oxidases AtREACTIVE-BURST OXIDASE HOMOLOG H/J (AtRBOHH/J)16, the type-one protein phosphatases AtATUNIS1/2 (AtAUN1/2)17, as well as the receptor-like cytoplasmic kinase (RLCK) AtMARIS (AtMRI)18. Apart from the negative regulators AtAUN1/2, all the above-mentioned proteins positively regulate CWI maintenance. Consistently, their corresponding loss-of-function and knockdown mutants display precocious pollen tube bursting leading to low or no transmission of the mutant alleles by the haploid pollen to the next generation and thus to male sterility. CWI signaling also occurs in Arabidopsis root hairs, where it is governed by the closest homolog of AtANX1/2, the MLR AtFERONIA (AtFER); Atfer mutants develop root hairs that spontaneously burst during growth19. Intriguingly, AtMRI has been shown to also function during root hair CWI maintenance downstream of AtFER18. Concordantly, AtMRI is strongly expressed in both pollen and root tissues and AtMRI protein is located though the plasma membrane of both pollen tubes and root hairs where it is required to maintain CWI18, 20. AtMRI belongs to the Arabidopsis RLCK-VIII subfamily that shares homology with the tomato Pto-interacting protein 1 (PTI1) involved in the Pto-mediated hypersensitive response21. Apart from the bursting root hair and pollen tube phenotypes of Atmri, and abscisic acid (ABA)-insensitivity of Atcark1 mutants22, no phenotype and therefore no function have been reported for any of the other 9 Arabidopsis PTI1-like RLCK proteins.
Recently, a T-DNA insertional mutant screen for defective rhizoid growth phenotypes in the early-diverging land plant Marchantia polymorpha identified mutant alleles of the two unique Marchantia MLR and PTI1-like kinases, MpFERONIA (MpFER) and MpMARIS (MpMRI), respectively23, 24. Here, by focusing on AtMRI and MpMRI and using trans-species complementation assays, we demonstrate that the rhizoid-derived MpMRI and the root hair- and pollen tube-derived AtMRI can function in the CWI pathways of all three tip-growing cell types. Despite their different function, origin and growth environment, we show that all three cell types respond similarly to disruption and overexpression of the PTI1-like encoding genes. Finally, we show that MpMRI functions downstream of MpFER in the rhizoid CWI pathway, revealing that the CWI control of tip-growth in bryophytes and seed plants relies on a common MLR/PTI1-like receptor-like kinases module that has been conserved during land plant evolution.
Results
MpMRI is required for CWI during rhizoid growth
We previously showed that the Arabidopsis PTI1-like AtMRI controls CWI during tip-growth of root hairs and pollen tubes18. A large-scale mutant screen in Marchantia for defects in rhizoid growth revealed that a T-DNA insertion in the genetic locus Mapoly0051s0094.1, renamed MpMRI (see below), coding for a PTI1-like homolog, may cause a morphologically related loss of CWI phenotype in tip-growing rhizoids23. We first reanalyzed the Mpmri-1 (formerly named Mppti-1 or ST13.323) mutant phenotype in young gemmalings and confirmed that the mutant develops significantly shorter rhizoids than the wild-type accessions (Fig. 1a,b). Mpmri-1 also developed fewer intact rhizoids per thallus than Tak-2 (Fig. 2c), because rhizoids frequently burst at their tip, as revealed by brownish spots on the ventral side of thalli grown on solid medium (Fig. 1a, 2d) or cytoplasmic content released by collapsed rhizoids grown in liquid condition (Fig. 1c). To test whether the T-DNA insertion in the MpMRI locus (Supplementary Fig. 1) causes the rhizoid bursting phenotype, we transformed the Mpmri-1 mutant with the native MpMRI coding sequence fused to the red fluorescent protein (RFP) under control of the ubiquitously active promoter of Elongation Factor 1 α (proMpEF1α). Rhizoids were longer in transformed lines than in mutants (Fig. 2a and Supplementary Fig. 2). Bursting of rhizoids and the resulting brown staining observed in Mpmri-1 rhizoids was not observed in any of the transformed lines (Fig. 2d and Supplementary Fig. 2). Thus, disruption of the MpMRI locus indeed causes defective rhizoid growth in Mpmri-1 mutants. The functional MpMRI-RFP protein fusion mainly localizes to the plasma membrane of Mpmri-1 rescued rhizoids, with a weaker fluorescence signal detected in the cytoplasm (Fig. 2e and Supplementary Video 1). This localization pattern resembles the localization of AtMRI in Arabidopsis tip-growing cells18, 20. Therefore, a plasma membrane-localized MpMRI protein is required for CWI in M. polymorpha rhizoids.
AtMRI can maintain CWI during rhizoid growth in Marchantia
Since both AtMRI and the Marchantia PTI1-like homolog MpMRI control CWI in different tip-growing cells, we tested if their function in CWI was conserved. The sequences of the AtMRI and MpMRI proteins are 64 % identical (Supplementary Table 1). We first introduced AtMRI-RFP into the Mpmri-1 mutant. Expression of AtMRI-RFP led to a partial, but significant increase in rhizoid length (Fig. 2b) and apparent cell wall integrity (Fig. 2d). The number of intact rhizoids per thallus was significantly higher in all rescue lines as compared to Mpmri-1 (Fig. 2c). Furthermore, AtMRI-RFP localized to the plasma membrane of rhizoids just like MpMRI-RFP (Fig. 2e and Supplementary Video 2). Taken together, these results demonstrate that AtMRI can partially substitute for MpMRI to maintain CWI in M. polymorpha rhizoids, despite the independent evolution of both genes for approximately 450-470 million years.
MpMRI can maintain CWI during pollen tube and root hair growth in Arabidopsis
Next, we tested the hypothesis that the function of PTI1-like proteins has been universally conserved throughout evolution of rooting cells and/or upon acquisition of new functions (i.e. between rooting cells and pollen tubes). We first introduced MpMRI fused to the yellow fluorescent protein (YFP) into the heterozygous Atmri-1/AtMRI (herein after referred to as mri-1/MRI) background under control of the native AtMRI promoter (proAtMRI), which drives expression in both, root hairs and pollen tubes17, 18. Normally, mri-1 homozygous individuals are never isolated in the self-progeny of heterozygous mri-1/MRI plants because mri-1-dependent pollen tube bursting prevents transmission of the male mri-1 allele and thus, production of homozygous mri-1/mri-1 individuals (Supplementary Table 2)18. By contrast, 7 homozygous mri-1/mri-1 T2 individuals were isolated out of 200 progenies originating from three independent mri-1/MRI T1 lines hemizygous for MpMRI-YFP suggesting that MpMRI expression partially rescues the mri-1 male gametophytic defect.
To confirm that the partial rescue of mri-1 male transmission efficiency resulted from the rescue of CWI during pollen tube growth by MpMRI-YFP expression, we performed in vitro pollen germination assays on homozygous T3 plants (mri-1/mri-1; MpMRI-YFP) originating from three independent T1 lines. Pollen bursting rate was lower in all lines expressing MpMRI-YFP and not significantly different from the bursting rates of either wild type, or mri-1/mri-1 complemented with an AtMRI-YFP construct (Fig. 3a)18. However, the expression of MpMRI-YFP frequently led to abnormal pollen tube morphology in vitro, as indicated by swollen apices and branching (n=56 out of 60; Fig. 3b). These abnormalities were not observed in mri-1 pollen tubes expressing AtMRI-YFP (Fig. 3b) and may explain the incomplete rescue of male mri-1- transmission efficiency by MpMRI-YFP expression. MpMRI-YFP normally localized to the plasma membrane of the pollen tube (Fig. 3b). To exclude that the observed rescue of CWI is not an effect of the artificial in vitro conditions, we also determined the seed set of T3 lines. For all T3 lines, there was a significant increase of seed set (Fig. 3c,d). Taken together, these results demonstrate that MpMRI can partially substitute for AtMRI in mri-1 pollen tubes despite the different functions and growth environment of rhizoids and pollen tubes.
To test if expression of MpMRI-YFP can rescue the loss of CWI in mri-1 mutant root hairs, we analyzed root hair growth in two of the A. thaliana mri-1 T3 lines homozygous for MpMRI-YFP. Not only was root hair length partially restored to wild-type length in all individuals (n > 10; Fig. 4a,c), but the rate of root hair bursting was much lower than in mri-1 mutants (Fig. 4b,c). mri-1 root hairs expressing MpMRI-YFP (Fig. 3b) developed cylindrical root hairs indistinguishable from wild type (Fig. 4c). MpMRI-YFP localized to the plasma membrane of root hairs similarly as it and AtMRI did in other tip-growing cells (Fig. 4c). These results demonstrate that MpMRI can fully replace AtMRI in the CWI maintenance pathway during root hair growth.
Overexpression of PTI1-like proteins inhibits rhizoid, root hair and pollen tube growth
Overexpression of positive or knockdown of negative CWI regulators can lead to growth cessation in Arabidopsis pollen tubes and root hairs16–18. To further confirm that MpMRI function is conserved between A. thaliana and M. polymorpha we tested whether rhizoid growth would be inhibited by overexpression of PTI1-like genes. Constitutively overexpressing MpMRI and AtMRI fused to a ternary Citrine tag (3xCitrine) in wild-type sporelings led to significant rhizoid growth inhibition (Fig. 5a,b and Supplementary Fig. 5a). The shorter rhizoids on lines overexpressing MpMRI- or AtMRI-3xCitrine displayed the expected plasma membrane localization for both MpMRI and AtMRI (Supplementary Fig. 3). Taken together, these results indicate that the repression of growth by PTI1-like kinases has been conserved since A. thaliana and M. polymorpha last shared a common ancestor.
The conserved R240 of PTI1-like proteins exerts an auto-inhibitory effect
A suppressor screen for rescue of male fertility in the double-knockout mutant Atanx1 anx2 helped identify a hyperactive variant of AtMRI, with a R240C substitution in the activation loop of the kinase domain18. Structural similarity analyses of land plant PTI1-like revealed that MpMRI is part of a monophyletic group that includes AtMRI (termed the ‘MRI-like group’ from here on; Supplementary Fig. 4a). Both the invariant lysine residue of the ATP-binding pocket required for kinase activity (K100 in AtMRI) and the STR-motif, which includes arginine R240, are conserved among members of the MRI-like group (Supplementary Fig. 4a). Furthermore, the conserved threonine (T239 in AtMRI) of the conserved STR motif of some PTI1-like kinases can be phosphorylated by other kinases such as the Arabidopsis OXIDATIVE SIGNAL-INDUCIBLE1 (OXI1) and tomato Pto kinases25, 26.
To better understand the mechanistic role of conserved K100, T239 and R240 residues, we designed several structural variants of AtMRI (Supplementary Table 3). First, we designed two kinase-dead variants with a K100N and K100E amino acid substitution, respectively 21, 27, 28 and introduced them in mri-1/MRI under the pollen-specific promoter Lat5229. Expression of both structural AtMRIK100E and AtMRIK100N variants led to the emergence of homozygous mri-1/mri-1 lines in the self-progeny of transformed mri-1/MRI (Supplementary Table 3, orange rows) similarly as for expression of wild-type AtMRI (Supplementary Table 3, green row) or AtMRIR240C (Supplementary Table 3, blue row). Hence, kinase activity does not appear to be required for AtMRI function during pollen tube growth. Second, we tested for the necessity of phosphorylation at residue T239 by engineering a non-phosphorylatable T239A and a phosphomimetic T239E variant, respectively, and introducing them in mri-1/MRI. Again, expression of both AtMRIT239A and AtMRIT239E variants led to the emergence of homozygous mri-1/mri-1 individuals (Supplementary Table 3, yellow rows) indicating that phosphorylation at T239 is not required for AtMRI function during pollen tube growth. These results are in good agreement with a previous study on rice OsPTI1a, where equivalent mutant variants OsPTI1aK96N and OsPTI1aT233A were able to rescue Ospti1a spontaneous lesions and rice blast-enhanced resistance phenotypes30. Moreover, they also suggest that overactivity of AtMRIR240C arises from a substitution of R240 itself rather than from mimicking phosphorylation at T23931. We then produced some variants for R240 itself, namely R240M, R240A and R240S. To assess a potential protein overactivity that may arise from the substitutions, we expressed all our variants in the pollen tube bursting male-sterile AtrbohH rbohJ mutant, whose sterility is rescued by expression of the overactive AtMRIR240C variant, but not the native AtMRI protein18. Similarly, as for wild-type AtMRI, expression of the phosphomimetic AtMRIT239Evariant did not restore AtrbohH rbohJ sterility, with all independent T1 lines producing only a few seeds (Supplementary Fig. 4b). However, expression of any R240 variant led to a clear rescue of AtrbohH rbohJ fertility with seed set increase (Supplementary Fig. 4b). These results confirm that R240C overactivity is not due to phospho-mimicking at residue T239 and indicate that any substitution of residue R240 leads to an overactive version of AtMRI. Thus, residue R240 of AtMRI appear to exert an enigmatic auto-inhibitory effect.
MpMRIR240C inhibits growth in Marchantia and Arabidopsis tip-growing cells
If MpMRI and AtMRI are functionally conserved, we would expect that the hyperactive R240C mutation in each would have similar phenotypic consequences. To test this, we constitutively expressed variant MpMRIR240C-RFP in Marchantia. Rhizoid length was significantly reduced in all independent transformed lines compared to controls (Fig. 5a, b). Also, the growth inhibitory effect derived from MpMRIR240Cexpression was stronger than for MpMRI expression as observed when AtMRIR240C and AtMRI are overexpressed in A. thaliana18, suggesting that the R240C substitution in MpMRIR240C induces a similar effect as for AtMRIR240C.
Next, to test whether MpMRIR240C growth inhibition effect is restricted to rhizoid, we introduced MpMRIR240C-YFP in Arabidopsis wild-type pollen tubes and root hairs under the control of proAtMRI. To directly assess the effect of MpMRIR240C on pollen tube germination and growth, we used several independent lines hemizygously expressing the transgene, i.e. in plants from which half of the pollen is wild type and the other half expresses MpMRIR240C-YFP (Supplementary Table 4), thus allowing direct comparison of transgenic and wild-type pollen tubes within the same line and assays. Pollen tubes were shorter for pollen expressing MpMRIR240C than for wild-type pollen (Fig. 5c and d; Supplementary Fig. 5). We then asked whether these growth inhibitory effects were strong enough to affect pollen germination itself. While MpMRI-YFP did not significantly affect the percentage of transgenic to non-transgenic pollen tubes (expected to be 50% if the transgene does not influence germination), expression of MpMRIR240C-YFP decreased the ratio of pollen tubes carrying the transgene to 37 - 43 % (Supplementary Table 4). Interestingly, plasma membrane invaginations in fluorescent non-germinating pollen grains were observed, a phenotype described upon AtMRIR240C expression, as well (Fig. 5d)18. A more severe effect was observed upon expression of AtMRI-CFP (16 %), while AtMRIR240C-CFP induced a complete inhibition of pollen germination (complete absence of fluorescent pollen tubes), just as reported before (Supplementary Table 4)18. These findings suggest that expression of MpMRIR240C disrupts pollen tube growth more than MpMRI as observed upon expression of AtMRIR240C and AtMRI.
We assessed whether expression of MpMRIR240C has a similar growth inhibitory effect on root hairs. Root hairs were shorter on all lines homozygous for proAtMRI:MpMRIR240C-YFP (originating from the same three independent T1 lines tested for pollen tube growth) than on wild type (Fig. 5e and Supplementary Fig. 6). Again, the growth inhibitory effect observed in root hairs was less pronounced than upon expression of Arabidopsis AtMRIR240C-YFP (two independent lines, Fig. 5e and Supplementary Fig. 6). Altogether, the similar growth inhibitory effects observed on tip-growing Marchantia and Arabidopsis cells upon expression of either MpMRIR240C or AtMRIR240C suggest functional conservation of the auto-inhibitory residue R240 and overactivation of the CWI pathways upon expression of the R240C variants.
MpMRI and MpFER act in a common, evolutionarily conserved signaling pathway
We previously reported that expression of AtMRIR240Cwas sufficient to partially rescue loss of CWI of Atanx1 anx2 pollen tubes as well as Atfer root hairs, thereby positioning MRI downstream of the MLRs AtANX1 and AtFER18. Interestingly, disruption of the unique Marchantia MLR homologue MpFER24 was also reported to trigger loss of CWI in rhizoids23 suggesting that the MLR/PTI1-like signaling module is conserved during rhizoid growth. To further demonstrate the conservation, we expressed a MpMRIR240C-RFP protein fusion in the Mpfer-1 mutant23 (Supplementary Fig. 1). In contrast to Mpfer-1 in which almost every rhizoid bursts at an early stage of growth, rhizoids on 4 independent Mpfer-1 lines transformed with MpMRIR240C-RFP were longer and more frequently intact (Fig. 6 and Supplementary Fig. 7). These data indicate that MpMRI functions downstream of MpFER in the CWI pathway that controls rhizoid growth. They also confirm that the MLR/PTI1-like signaling module controlling CWI in tip-growing cells has been conserved during land plant evolution from the rhizoids of liverworts to the root hairs and pollen tubes of seed plants.
Discussion
During land plant evolution, tip-growth was recruited to participate in several key adaptations to terrestrial habitats that facilitated colonization and radiation of land plants. Among these adaptations are substrate anchorage, water/nutrient uptake and microorganism interactions via rooting cells (rhizoids and root hairs), and water-independent sperm delivery to the egg cell via pollen tubes. While tip-growth in plants performs extremely well for increasing surfaces, exploring environment and transport, it relies on a delicate balance of turgor-driven, anisotropic cell expansion and appropriate secretion of surface membrane and cell wall material that defines cell shape and function. We show that this balance is mediated, at least in part, by the FER/MRI signaling module in tip-growing cells. Failure of this mechanism in plants with defective FERONIA or MARIS function results in loss of cellular integrity because of mechanical failure in the cell wall.
While rhizoids and root hairs represent morphologically similar cell types serving similar biological functions, they develop during different life-phases. Rhizoids are found on the gametophyte of early diverging land plant taxa, while root hairs develop on the roots of vascular plants4, 32. The development of rhizoids or root hairs in all major lineages of land plants suggests that tip-growing cells formed the rooting structures of the first land plants4. Unlike pollen tubes that emerge from pollen grains, rhizoids and root hairs originate from the swelling of some epidermal cells. The development of such is controlled by an ancient conserved genetic mechanism composed of the class I ROOT HAIR DEFECTIVE SIX-LIKE (RSL) basic helix-loop-helix transcription factors33, 34. The control of root hair and rhizoid development by the same transcription factors has been interpreted as an example of deep homology35 since, at the level of land plants, rhizoids and root hairs are analogous structures with similar functions whose development is controlled by a homologous genetic mechanism.
Later during evolution, some vascular plants developed an elaborate system to guarantee an efficient water-independent sperm cell delivery to the female egg cell via pollen tubes, a complex process referred to as siphonogamy5, 36. Siphonogamy may have evolved twice independently within the gymnosperm and angiosperm lineages5, 37. Pollen tubes of seed plants are thought to have evolved from a slow and long-lived haustorial pollen tube involved in nutritional uptake to a fast-growing, short-lived pollen tube whose unique function is the delivery of non-motile sperm in angiosperms5, 38. It is currently unknown whether CWI maintenance mechanisms were already active in the slow haustorial pollen tubes or if they were recruited along the transition to fast growth to avoid growth accidents and infertility. So far, however, there have been no reports of homologous mechanisms conserved between pollen tubes and rhizoids.
In this study, we demonstrate the evolutionarily conservation of the genetic pathway comprised of MLRs and PTI1-like receptor-like kinases that regulates CWI maintenance during tip-growth. We showed that not only disruption but also overexpression of the PTI1-like encoding genes MpMRI and AtMRI led to similar CWI maintenance defects in rhizoids of the early diverging land plant Marchantia as in the root hairs and pollen tubes of the seed plant Arabidopsis. Moreover, trans-species complementation assays showed that AtMRI expression partially rescues the bursting rhizoid phenotype of Mpmri-1 while MpMRI expression rescues partially pollen tube bursting of Atmri-1 and fully its bursting root hair phenotype. These results demonstrate that MpMRI and AtMRI can largely functionally replace each other regardless of the tip-growing cell type or the phylogenetic position of the plant on which they develop. This indicates that the fundamental function of PTI1-like proteins has been conserved since liverworts and seed plants diverged from a common ancestor. Nonetheless, in the Arabidopsis mri-1 mutant, MpMRI-expressing root hairs grew better than MpMRI-expressing pollen tubes. Consequently, the CWI control in sporophytic root hairs is likely to be more similar to that in gametophytic rhizoids than in pollen tubes, the male gametophytes of flowering plants. Thus, our study suggests that a genetic program maintaining CWI in rhizoids was established early during land plant evolution and then adopted by the sporophyte for root hair growth control and much later by the male gametophyte of flowering plants.
One can also wonder whether CWI maintenance during tip-growth represents the original ancestral function of PTI1-like proteins. The fact that several Arabidopsis PTI1-like proteins, close homologues of AtMRI, cannot even partially substitute for AtMRI in pollen tube CWI maintenance39 but MpMRI can, suggests subfunctionalisation of the PTI1-like gene family in the lineage that led to Arabidopsis. Studies in tomato, cucumber, and rice all demonstrate a role for PTI1-like in plant responses to pathogens as either positive (in dicots such as Tomato21, 40 and Cucumber41) or negative (in monocot, Rice28, 42) regulators. Moreover, the Arabidopsis PTI1-like AtCARK1 positively regulates abscisic acid (ABA) signaling by phosphorylating ABA receptors22. Since some ABA responses and signaling components are conserved in Marchantia24, 43–45, and pathosystems are being established in Marchantia46 it would be informative to test if MpMRI regulates ABA signaling and/or immunity in this species. Our previous study showed that the AtMRIR240C structural variant functions as an overactive protein version, whose expression is not only sufficient for tip-growth inhibition in the wild-type background but also rescues loss of CWI in mutants of its upstream MLR regulators AtANX1/2 for pollen tubes and AtFER for root hairs18. Here, we showed that expression of the Marchantia variant MpMRIR240C triggered similar growth inhibitory effects in both, Marchantia and Arabidopsis tip-growing cells suggesting a conserved structural mechanism. Furthermore, expressing the MpMRIR240C variant in the Mpfer mutant partially rescued Mpfer bursting rhizoids, similarly as did AtMRIR240C expression for Atfer bursting root hairs and Atanx1 anx2 bursting pollen tubes. Therefore, the distantly related land plant taxa Arabidopsis and Marchantia share a common MLR/PTI1-like receptor-like kinase signaling module to control CWI in tip-growing cells. Since these two species last shared a common ancestor approximately 450-470 million years ago and given the most likely topologies of land plant phylogeny, we conclude that such a signaling module is an ancient common feature shared by most lineages of extant land plants. Whether such signaling module includes further common regulatory features known from the Arabidopsis CWI maintenance pathways (Fig. 6d), such as extracellular receptor binding of RALF peptides, regulation of ROS-producing NADPH oxidase etc. remains to be investigated.
Materials and Methods
Plant materials and growth conditions
All mutant lines, transgenic lines, oligonucleotide sequences and plasmids used in this study are listed in Supplementary Tables 5-8, respectively. The presence and position of the T-DNA in the respective genetic locus of the insertional mutants Mpmri-1 (ST13.3) and Mpfer-1 (ST14.3)23 was verified via PCR-genotyping and sequencing (Supplementary Fig. 1). Binary vectors were first transformed into Agrobacterium tumefaciens strain GV3101 via electroporation. A pure culture was then used to transform Arabidopsis47 or Marchantia sporelings48 and thalli49.
Arabidopsis seeds (Col-0) were sown on solid half-strength Murashige and Skoog (MS) basal medium (Duchefa Biochemie), stratified (2 d, 4 °C, darkness) and allowed to germinate in a growth chamber set to long day conditions (7-10 d, 21 °C, 60 % humidity, 16 h white light at 80 µmol m-1 s-1/8 h darkness cycle). Seedlings were then transferred and grown on soil in a greenhouse under long day conditions. Vegetative Marchantia stages were grown on solid Johnson’s growth medium50 (0.8 % agar to facilitate growth of rhizoid defective mutants) under long day conditions. To cross male and female Marchantia lineages, 4 weeks old thalli were transferred onto soil and grown for 3 weeks under long day conditions (white light: 40 µmol m-1 s-1; far-red light between 700-880 nm: 15 µmol m-1 s-1) to trigger gametangia formation. Crossings were then carried out manually via spermatozoid transfer from mature male to female gametangia. Emerging sporangia were harvested after 2-3 weeks and sterilized with 1 % sodium dichlorisocyanurate to allow for axenic, vegetative cultivation.
Molecular cloning
To amplify the MpMRI and MpFER coding sequences (CDS), RNA was isolated from Tak-1 and Tak-2 whole-plant material using the Direct-Zol RNA Mini-Preparation Kit (Zymo Research). cDNA synthesis was carried out using the RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) and used as template for the amplification of the respective CDS without stop codon using gateway-compatible primers. The different CDS were then cloned into pDONR207 and transformed into Escherichia coli DH5α cells to generate Gateway-entry clones. The CDS were then remobilized in the respective Gateway destination vectors of choice to generate fluorescent protein fusions (Supplementary Table 7). Amplified sequences and preservation of the reading frames were verified via sequencing of the full-length CDS (entry clones) and the promoter-CDS / CDS-fluorescent protein junctions (final binary constructs).
The new variants MpMRIR240C and AtMRIK100N, AtMRIT239A, AtMRIT239E, AtMRIR240M, AtMRIR240A and AtMRIR240S, were generated via site-directed mutagenesis (SDM). SDM primers were designed using Primer X (http://www.bioinformatics.org/primerx/) to introduce via PCR the different adequate nucleotide substitutions with MpMRI and AtMRI in pDONR207 as template. The PCR-mix were digested with the methylation-sensitive restriction enzyme DpnI to degrade non-mutagenized vector copies. The mutagenized vectors were transformed into E. coli DH5α cells and the respective nucleotide substitution were verified via sequencing before remobilization of the Gateway cassette in the final binary plasmids.
Selection of transformant lines
Putatively transformed Arabidopsis T1 seeds were preselected on solid half-strength MS with Basta (glufosinate ammonium, 10 µg/ mL). Resistant T1 seedlings were transferred and grown on soil. Upon flower emergence, pollen was checked for expression of the respective fluorescent fusion protein. Transgenic T1 lines with clear, homogenous fluorescence in about 50 % of the pollen were propagated. Their seeds were harvested and selected again in the T2 generation. Upon flower emergence, plants were screened for 100 % fluorescent pollen grains, indicating homozygous lines. Plants expressing fluorescent protein fusions under control of proAtMRI were additionally screened for root hair-specific fluorescence at the seedling stage.
Putatively transformed Marchantia sporelings/thalli were preselected on solid Johnson’s growth medium with Chlorsulfuron (100 µg/ mL) or Gentamicin (0.5 µM) and supplemented with Cefotaxime (100 µg/ mL) to avoid contamination with residual Agrobacteria. Resistant plants were then genotyped for presence of the fluorescent protein fusion and additionally checked for Citrine/RFP fluorescence in thalli, gemmae and/or rhizoids. Propagation of the different transgenic lines was done by transplantation and cultivation of gemmae.
Rhizoid growth assays
Gemmae were grown vertically for 7 d under axenic standard conditions (as described above). To determine the mean rhizoid length, the 10 longest rhizoids per gemmaling were measured for n ≥ 10 individuals. To approximately estimate the percentage of bursting rhizoids, the mean number of visibly intact rhizoids per thallus was determined per individual. For time course experiments, data was collected after 7, 14, 21 and 28 d. Images were taken with a Leica MZ 16 F fluorescence binocular (whole-plant), a Leica DM5000 fluorescence microscope (close-up images). Alternatively, fresh gemmaes were grown for 24hr in liquid Johnson’s growth medium sandwiched between a slide and a cover slip separated with a thin layer of parafilm. Young growing rhizoids were then imaged with either a Leica DM5000 fluorescence microscope or a Leica SP8 confocal laser scanning microscope.
In vitro Pollen germination assays and length measurements
Young, freshly opened Arabidopsis flowers were harvested in the morning and pollen grains were allowed to hydrate for 30 min under long day conditions in a humid environment. Stamina were then brushed on a microscopic slide coated with 500 µL of solid pollen germination medium 51. Pollen germination assays and imaging were conducted as described previously17. Bright field microscopy was carried out with a Leica DM5000 microscope and n = 150 – 300 pollen grains per genotype and experiment were imaged and later classified as non-germinated, germinated and intact or germinated and burst.
In order to determine the mean pollen tube length, pollen germination was induced as described above. The length of n = 200 pollen tubes was measured following the grown pollen tube from the origin of germination to the pollen tube tip in three independent experiments.
Seed set assays
Plants were allowed to self-fertilize and n > 10 mature siliques were harvested per individual plant. Siliques were fixed and bleached in 1.5 mL of an acetic acid / ethanol solution (v/v 1:3) for 1 d. Siliques were then opened and their seeds were quantified to determine the mean number of seeds per silique. Images were obtained with a Leica MZ 16 F fluorescence binocular.
Root hair assays
Seeds were allowed to germinate vertically on solid half-strength MS medium supplemented with 2 % sucrose. 4 – 5 d old seedlings were embedded in a microscope slide with liquid 1/10 MS medium. The borders of the glass slide were wrapped with a thin parafilm layer to act as spacer to prevent squashing the young fragile Arabidopsis roots with the coverslip. The most distal part of the root was imaged (Leica DM5000 fluorescence microscope) and digitally divided into sections of 150 µm length starting from the first root hair from the tip. The length of 5 – 10 root hairs per section was determined. The mean of each section was then compared amongst specimens of each genotype. The root hair bursting rate was determined as the percentage of bursting root hairs out of the total root hair number.
Phylogenetic analyses
The AtMRI (AT2G41970) amino acid sequence was used as query to perform a pBLAST homology search (BLOSUM62 matrix) against the protein database on representative species of major land plant taxa. pBLAST was performed on https://phytozome.jgi.doe.gov/ using standard settings for all species. The following species and genome versions were used: AmTr - Amborella trichopoda (v1.0), AT - Arabidopsis thaliana (TAIR10), Mapoly - Marchantia polymorpha (v3.1), Pp - Physcomitrella patens (v3.3), SeMo - Selaginella moellendorffii (v1.2). Only sequences either showing a percent identity >60% and/or annotation as PTO-interacting protein 1-like proteins (KEGG ORTHOLOGY code K13436) were considered. Multiple sequence alignments of all PTI1-like amino acid sequences were built using ClustalX2 52 and illustrated using GeneDoc 2.7.0 and phylogenetic trees were calculated with MEGA753 using the Maximum Likelihood method based on the JTT matrix-based model54 and 1000 repetitions. The tree with the highest log likelihood (-4460.1659) is shown in Supplementary Fig. 4. The percentage of trees in which the associated taxa clustered together is shown next to the branches. The frequency of amino acid usage per residue was illustrated using WebLogo 3 (http://weblogo.threeplusone.com/create.cgi).
Data processing
Image analysis was performed using FIJI/ImageJ 1.50e55. Bar graphs were designed using Excel/Microsoft Office 365. Boxplots were designed using R studio 0.99.892 (RStudio Team, 2015).
Author contributions
J.W., L.D. and A.B.-D. conceived the experiments. J.W., S.S., C.M.F., R.L., and A.B.-D. performed the experiments shown. J.W. and A.B.-D. analyzed the data. J.W., L.D. and A.B.-D. wrote the manuscript with contributions from the other authors.
Competing interests
The authors declare no competing interests.
Supplementary Video 1: Time course of a growing Mpmri-1 rhizoid expressing MpMRI-RFP.
Δt =15 sec, 30 frames. Left side is an overlay of RFP-derived fluorescence in magenta and plastids autofluorescence in green. Right side is bright-field. Scale bar: 10μm.
Supplementary Video 2: Time course of a growing Mpmri-1 rhizoid expressing AtMRI-RFP.
Δt =15 sec, 30 frames. Left side is an overlay of RFP-derived fluorescence in magenta and plastids autofluorescence in green. Right side is bright-field. Scale bar: 10μm.
Acknowledgements
We thank Ueli Grossniklaus and Moritz Rövekamp (University of Zürich, Switzerland) for sharing unpublished data and enriching discussion. We also thank all members of Martin Hülskamp’s, Ute Höcker’s (University of Cologne, Germany) and Liam Dolan’s (University of Oxford, United Kingdom) groups for sharing their facilities. We thank Gabi Meineke (University of Cologne, Germany) for administrative support, Hiroyasu Motose (University of Okayama, Japan) and Benjamin Jaegle (Gregor Mendel Institute, Austria) for valuable intel about live-imaging of Marchantia rhizoids and phylogenetic trees, respectively. This research was partly funded by a short-term stipend of the Deutscher Akademischer Austauschdienst (DAAD) to J.W.; Biotechnology Biology Research Council Doctoral Train Partnerships fellowship in Interdisciplinary Bioscience to S.S.; European Research Council Advanced Grants EVO-500 (250284) and DENOVO-P grant (787613) to L.D; the University of Cologne, the Deutsche Forschungsgemeinschaft Grant BO 4470/1-1, and grant from the University of Cologne Centre of Excellence in Plant Sciences to A.B.-D..