A deeply conserved amino acid required for VAPYRIN localization and function during legume–rhizobial symbiosis

Most legume plants have the ability to form an endosymbiosis with rhizobia, which fix nitrogen in special root outgrowths called nodules (Sprent, 2009; Downie, 2014; Martin et al., 2017). To enter host roots and eventually colonize the nodules, rhizobia need to travel through several root cell layers, a process known as rhizobial infection (Gage, 2004; Murray, 2011). Depending on host–rhizobia combinations and symbiotic conditions, rhizobia can enter legumes by either intercellular or intracellular infection, with the latter strategy exploited by most legumes such as Medicago truncatula and Lotus japonicus (Quilbé et al., 2022). Intracellular infection starts in root hairs following reciprocal exchange and perception of chemical signals between rhizobia and legume host leading to the activation of Nod factor signaling pathway, which triggers the expression of structural genes required for accommodating the rhizobia (Oldroyd et al., 2011; Yang et al., 2022; Geurts & Huisman, 2023; Rübsam et al., 2023). Rhizobial infection requires the formation of an intracellular structure, the infection thread (IT), which serves as a passageway for the entry of rhizobia into the root (Parniske, 2018). Following inoculation with compatible rhizobia, some root hairs from the susceptible zone redirect their growth to curl around and entrap the attached rhizobia within an infection chamber, where the rhizobia propagate forming a micro-colony (MC). From the infection chamber, an inward protrusion of the cell wall and membrane occurs, giving rise to a tubular IT that extends through the root hair. The process repeats itself in the underlying cortical cells, allowing rhizobial colonization of the emerging nodule (Fournier et al., 2008, 2015). The ITs then ramify in cells of the developing nodule, and as the nodule matures, the rhizobia are exocytotically released to form organelle-like structures called symbiosomes where they fix nitrogen (Ivanov et al., 2012). Although the molecular mechanism underlying IT development remains obscure, several host players have been characterized, including VAPYRIN (VPY). The vpy mutant shows delayed initiation of ITs, often resulting in oversized MCs (Murray et al., 2011; Liu et al., 2019). VPY is also required for arbuscule development in arbuscular mycorrhiza (AM) symbiosis (Feddermann et al., 2010; Pumplin et al., 2010; Murray et al., 2011; Zhang et al., 2015; Bapaume et al., 2019; Chen et al., 2021). VPY encodes a plant-specific protein with a major sperm protein (MSP) domain and several ankyrin repeats, both of which are protein–protein interaction domains (Tarr & Scott, 2005; Li et al., 2006). In the context of nodulation, VPY interacts with LIN, a protein also essential for IT development, via its ankyrin repeats region, and is positively regulated by LIN, likely by stabilizing VPY protein by so far unknown mechanisms, in both M. truncatula and L. japonicus (Kuppusamy et al., 2004; Kiss et al., 2009; Yano et al., 2009; Liu et al., 2019; Liu et al., 2021). VPY and LIN form an ‘infectosome’ protein complex with RHIZOBIUM-DIRECTED POLAR GROWTH (RPG) and an exocyst subunit EXO70H4, regulating IT development, likely via polarized exocytosis (Arrighi et al., 2008; Liu et al., 2019; Roy et al., 2020; Wang et al., 2022; Jhu & Oldroyd, 2023; Lace et al., 2023; Li et al., 2023). The core infectosome components, VPY, LIN and RPG accumulate at the extending IT tip and form one or a few cytoplasmic puncta near the nucleus (Liu et al., 2019; Lace et al., 2023; Li et al., 2023). The functional significance and mechanism underlying the interaction of these infectosome components remain to be elucidated. Here, we identify a single amino acid in VPY, which is essential for its role in rhizobial infection and its interaction with LIN, and determines its subcellular localization pattern. Medicago truncatula ecotype R108 and Jemalong A17 were used as wild-type (WT) plants. vpy-2 mutant is a Tnt1 retrotransposon insertional mutant (NF6898) generated in R108 background (Murray et al., 2011). For germination, Medicago seeds were treated with sulfuric acid for 3–5 min and then washed with sterilized deionized water several times. The seeds were then treated with bleach for 4 min, followed by washing with sterilized deionized water. The seeds were left in water for imbibition before being transferred to water agar plates, covered with tin foil and left in a 4°C fridge for 2 d. The seed plates were incubated at 22°C overnight, and for phenotype analysis, seedlings were transferred to a mixture of vermiculite and perlite (1 : 1) and grown in controlled environment chambers (16 h : 8 h, light : dark) before inoculation with rhizobia. Gateway cloning was used to make the constructs for yeast two-hybrid (Y2H). The full-length and different truncated fragments of VPY containing sequential combinations of ankyrin repeats (ANK1 to ANK9) were amplified (for primers see Supporting Information Table S1) using Phanta Max Super-Fidelity DNA Polymerase (Vazyme, Nanjing, China) and ligated into pDONR207 using Gateway BP Clonase (Invitrogen) to make pDONR-ANK1-9, ANK2-9, etc. These constructs were then cloned into destination vector pDEST-GADT7 using Gateway LR Clonase (Invitrogen) to make yeast AD constructs. Constructs used for amino acid (GRDQ) screening were first made using Mut Express II Fast Mutagenesis Kit (Vazyme) based on pDONR207-ANK6*-9 or pDONR207-VPY and then ligated into pGADT7 using Gateway LR Clonase (Invitrogen). These AD constructs were transformed into yeast strain AH109. pDEST-GBKT7-LIN (BD-LIN) was transformed into yeast strain Y187. The PEG/LiAC method was used for yeast transformation (Clontech, Beijing, China). Protein–protein interactions were detected using a yeast mating experiment followed by drop tests. For yeast mating experiments, yeast transformed with empty pDEST-GADT7 (AD-) were used as a negative control for the protein interaction test. Combinations of BD-p53 + AD-T and BD-Lam + AD-T were used as a positive control and a negative control, respectively, for the Y2H system. Yeast was allowed to grow for 3 d on SD-Leu-Trp (growth control) and SD-Ade-His-Leu-Trp before the results were recorded. Agrobacterium rhizogenes ARqua1 strain was transformed with the binary vectors made in this study by electroporation and used to generate composite plants comprising a transgenic hairy root system with untransformed aboveground tissues (Boisson-Dernier et al., 2001). Constructs for the localization of different VPY variants were made by Golden Gate cloning (Patron et al., 2015). DNA fragments of VPY and a 2.7-kb promoter of VPY (pVPY) were synthesized by Life Technologies (Beijing, China) and used as Level 0 modules, and a Level 1 vector was then assembled to make pVPY:VPY-GFP (Liu et al., 2019). Level 1 vectors of pVPY:VPYD394A-GFP, pVPY:VPYD394R-GFP and pVPY:VPYD394E-GFP were made by using Mut Express II Fast Mutagenesis Kit (Vazyme) based on pVPY:VPY-GFP. Level 1 vectors were then assembled into a Level 2 backbone vector EC50507 to make pVPY: VPY-GFP pAtUBQ10:DsRED, pVPY:VPYD394A-GFP pAtUBQ10:DsRED, pVPY:VPYD394E-GFP pAtUBQ10:DsRED and pVPY:VPYD394R-GFP pAtUBQ10:DsRED constructs. The four Level 2 constructs described above were used to study the subcellular localization of VPY and VPY D394 mutants. A confocal laser microscopy-based live-cell imaging system was used in this study (Fournier et al., 2008). Two weeks after hairy root transformation, composite plants transformed with the above-mentioned constructs were transferred to plates containing modified Fåhraeus medium with MgSO4 (3 mM), 0.5% Phytagel and 2-amino ethoxyvinyl glycine (AVG, 100 nM). The roots were covered with Lumox film (Sarstedt, Leicester, Leicestershire, UK). Plants were then grown vertically in a growth room with a 16 h : 8 h, 22°C, light : dark photoperiod. The transgenic plants were inoculated with Sinorhizobium meliloti strain Rm1021 or mCherry-tagged S. meliloti strain Rm1021 (OD600 = 0.001) and from 5 d postinoculation (dpi) onward root hairs at different infection stages were imaged under a Leica STELLARIS 5 or Olympus FV3000 confocal laser scanning microscope system. The excitation wavelengths for GFP and DsRED/mCherry were 488 and 561 nm, respectively. Fluorescent signals were collected at 500–530 nm (GFP) and 570–610 nm (DsRED/mCherry). Pseudo-colors presented in figures for GFP and DsRED/mCherry are green and red, respectively. Confocal images were processed using LAS X Office (Leica) or Olympus FV31S-SW software to give maximal z-projections of stacks and merged pictures. Constructs used for vpy-2 complementation experiments were pVPY:VPY-GFP pAtUBQ10:DsRED, pVPY:VPYD394A-GFP pAtUBQ10:DsRED, pVPY:VPYD394E-GFP pAtUBQ10:DsRED and pVPY:VPYD394R-GFP pAtUBQ10:DsRED, and empty vector (EV) was used as a negative control. The EV is a modified gateway destination vector pK7WGF2-R (in which the CaMV 35S promoter was replaced by the LjUBQ1 promoter). All the above constructs were transformed into vpy-2, and EV was also transformed into WT as a positive control for nodulation and rhizobial infection assays. Composite plants generated via A. rhizogenes-mediated hairy root transformation were transferred into pots containing vermiculite : perlite (1 : 1) in a growth room and were inoculated with S. rhizobium 1021 harboring pXLGD4 (hemA:lacZ; Rm1021-LacZ) at a concentration of OD600 = 0.05. Rhizobial infection phenotypes were investigated at 7, 14 and 21 dpi and nodulation quantification was done at 21 dpi. For rhizobial infection assay, roots from composite plants were fixed for 1 h in 2.5% glutaraldehyde and then stained in X-GAL (5-Bromo-4-chloro-3-indolyl-ß-d-galactoside; Sangon Biotech, Shanghai, China) solution (0.8 mg ml−1) with 100 mM sodium phosphate (pH 7.4), 10 mM KCl and 1 mM MgSO4 at 28°C overnight in the dark. Infection events as indicated by blue staining of rhizobia from different stages, including MCs, elongating infection threads (eIT), fully elongated ITs in root hairs (IT) and ITs ramified beyond epidermis (rIT) were quantified separately. Abnormal infections, for example enlarged MCs, blocked ITs, and so on, of each stages, were also quantified in composite plants transformed with each constructs. VPY homologs were retrieved from Phytozome (https://phytozome-next.jgi.doe.gov/blast-search) by Blast of proteome data using VPY sequence. Representative species were selected from different plant clades. The phylogenetic tree (maximum likelihood) was made using Phylogeny.fr (Dereeper et al., 2008) and was viewed using iTOL (Letunic & Bork, 2016). Protein alignment was done using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). All protein models were generated with default parameters using AlphaFold2 (Jumper et al., 2021) and implemented in the ColabFold interface available on the Google Colab platform (Mirdita et al., 2022). Protein sequences of MtVPY, MtVPYL, MtVPYL2 from M. truncatula, PhPAM1 from Petunia hybrid, and PpVPYL from Physcomitrella patens were all downloaded from Phytozome v.13 or NCBI. VPYD394A, VPYD394E and VPYD394R are designed VPY variants with mutated D394. The PyMol (https://pymol.org/2/) and UCSF ChimeraX (https://www.cgl.ucsf.edu/chimerax/) visualization tool were used to analyze the structure. Previous studies in M. truncatula and L. japonicus show that the interaction between VPY and LIN is mediated by VPY C-terminal half, which contains nine ankyrin repeats (Liu et al., 2019, 2021; Lindsay et al., 2022). The Y2H assay has been shown to be a useful and reliable system to test interactions between VPY, LIN and RPG (Liu et al., 2019, 2021; Li et al., 2023). To further explore whether any of the ankyrin repeats are essential for VPY's interaction with LIN, Y2H assays were performed using a series of truncated VPY-C fragments, ranging from the whole region containing all nine ankyrin repeats (ANK 1–9) to a small C-terminal region containing only the last two ankyrin repeats (ANK 8–9; Figs 1a, S1). The initial screen showed that all tested fragments could interact with LIN, except ANK 7–9 and ANK 8–9, suggesting that the region between ANK6 and ANK7, which could be a putative linker, is indispensable for the interaction (Figs 1a,b, S1). Three additional fragments, one including ANK7-9 plus the putative linker region between ANK6 and ANK7 (L-ANK7-9), a longer one starting from within ANK6 (ANK6*-9) and a shorter one starting from within the putative linker region (L*-ANK7-9), were tested, and the region essential for the interaction was further narrowed down to 8 aa, which distinguishes L*-ANK7-9 from ANK7-9 (Figs 1b,c, S1). A comparison with VPY orthologs from L. japonicus, Trifolium pratense, Vigna unguiculata, Phaseolus vulgaris, Glycine max, Lupinus albus and Arachis hypogaea revealed that the last four amino acids, that is GRDQ, are conserved (Fig. 1b). We then generated a series of ANK6*-9 versions by mutating either G392, R393, D394 or Q395 to A and found that while mutation G392, R393 or Q395 to A had no effect on the interaction, D394A disrupted the interaction between ANK6*-9 and LIN (Fig. 1d). We then tested the interaction between full-length VPY or VPY harboring mutated Asp394 (VPYD394A) and LIN, and found that Asp394 was indeed essential for the interaction between VPY and LIN (Fig. 1d). Thus, we have found a single amino acid from the putative linker region between ANK6 and ANK7 of VPY, which is essential for its interaction with LIN. VPY shows a punctate subcellular localization pattern in cells undergoing infection by rhizobia and the ankyrin repeat domain was previously shown to be important for this phenomenon (Liu et al., 2019). To test whether Asp394 is essential for this unique localization, separate constructs were made with GFP fused to the C-terminus of VPY and of VPYD394A driven by the native VPY promoter. Live-cell confocal imaging was used for the observation in roots of pVPY:VPY-GFP and pVPY:VPYD394A-GFP composite transgenic plants generated via A. rhizogenes-mediated hairy root transformation. As reported previously, VPY-GFP was present throughout the cytoplasm, sometimes in the nucleus as a diffuse signal, and showed a strong fluorescent signal in cytoplasmic puncta in most root hairs following rhizobial inoculation (Fig. 2a). Instead, although VPYD394A-GFP was detected in the cytoplasm and within the nucleus at a seemingly comparable level to that of VPY-GFP in most root hairs, it did not show the typical punctate distribution of VPY-GFP (Fig. 2b). Occasionally, VPYD394A-GFP showed an aggregate pattern, with an irregular shape, unlike the puncta seen with VPY-GFP, which were smooth and round (Fig. 2e). Sometimes, puncta were also seen in pVPY:VPYD394A-GFP root hairs, but these were usually smaller and weaker than that seen in pVPY:VPY-GFP root hairs (Fig. 2b). Similar observations were made in both straight and curled root hairs after rhizobial inoculation (Fig. 2f,g). In root hairs harboring ITs, VPY-GFP puncta were at the IT tip and near the nucleus as previously reported (Figs 2c, S2). By contrast, no VPYD394A-GFP signal was found at the IT tip, even in root hairs showing puncta near the nucleus, indicating D394 is required for VPY punctate accumulation at the IT tip (Fig. 2c,d). To further investigate whether the weak puncta found in pVPY:VPY D394A-GFP root hairs were different from VPY-GFP, we made a construct containing both pVPY:VPY-GFP and pVPY:VPYD394A-tdTomato, which we used for hairy root transformation. Observations in composite plants showed that while both VPY-GFP and VPYD394A-tdTomato often localized to cytoplasm and nucleus, in 78% of root hairs with VPY-GFP puncta, the puncta showed no tdTomato and in most of the root hairs that showed the tdTomato fluorescent signal was weak (Fig. shows that the VPYD394A puncta to that of we investigated whether VPYD394A could IT development in the vpy experiments were performed by pVPY:VPY-GFP and pVPY:VPYD394A-GFP into and infection events were quantified 21 dpi with S. meliloti The results show that had a of nodules as while were no nodules in and pVPY:VPYD394A-GFP which (Figs and nodule including some outgrowths with that are typical of vpy (Figs et al., of rhizobial infection events revealed MCs and no elongated ITs in and (Figs ramified ITs were in the of with all being blocked in as seen in vpy (Figs These results show that being required for VPY's localization at IT and cytoplasmic puncta, Asp394 is essential for the of VPY in the growth of To the level of Asp394 we retrieved plant VPY from and plants such as and patens S1). These homologs into which we VPY, and (Figs The alignment showed that the Asp394 was present in all all including which is not an The nine that the either than four ankyrin repeats or the region around The conserved of Asp394 that it not be by a amino and to test this we mutated to amino and to a amino acid We first tested the interaction between LIN and these two VPY and found that LIN with but not with VPYD394R in yeast (Fig. Two pVPY:VPYD394E-GFP and were made with pVPY:VPY-GFP and pVPY:VPYD394A-GFP, were used to generate composite plants. Confocal revealed that and showed a pattern to with no puncta or only weak puncta in most root hairs (Figs puncta were found at the IT tip in pVPY:VPYD394E-GFP or pVPY:VPYD394R-GFP although a weak GFP signal could be seen in pVPY:VPYD394R-GFP plants (Figs These constructs were then transferred into vpy-2 for complementation assays at 7, 14 and 21 dpi with S. and only pVPY:VPY-GFP could vpy-2 (Fig. These results show that and could not for suggesting its role is Although Asp394 was the interaction in nodulation, its in VPY homologs of plants it be also required for interactions and development of plants. In the the punctate pattern of mutated VPY variants is of VPY localization in the mutant (Figs indicating that Asp394 the and of VPY by LIN rhizobial infection (Liu et al., 2019, 2021). To whether D394 is for the of the ankyrin repeats or the protein of VPY we the structures of VPY and its mutant variants using with et al., the ankyrin repeats of VPY are into two and with ANK6 forming a in between (Fig. D394 is in the region between ANK6 and ANK7, and a by with and (Fig. the of D394 also interacts with in the of ANK7, and to (Fig. In the mutant variants although the first of the ankyrin repeats with that of VPY, the region of all mutant variants shows structural with the including the region around D394 (Fig. that D394 is important to the structural of this region, to the between and this to the interaction between VPY and LIN further The of VPY variants is with results for VPY protein in yeast by which showed seemingly expression of different VPY variants (Fig. that mutation of D394 did not protein & also a conserved from the of D394 from We made further on and some VPY including M. truncatula and genes are not and PpVPYL from a host (Figs From the only shows to VPY for the ankyrin repeats, the of and ANK7-9 (Fig. suggesting that the between the two separate ankyrin domains be essential for protein interactions one is that D394 to a protein with an to LIN, for example a LIN with a in the interaction with that was by LIN to the complex to mutation the interaction between VPY and LIN in yeast (Fig. but disrupted VPY localization pattern and rhizobial infection (Figs could be D394 the interaction with not only LIN but also essential components of the infectosome complex in has been shown that important of the could also VPY punctate accumulation et al., 2023). RPG VPY and whether D394 is required for this is further be that although D394 was required for the punctate localization of VPY, we found that its mutation also in the localization pattern (Fig. These in localization also be for its rhizobial infection, as reported for VPY formation symbiosis et al., 2022). its study on the role of Asp394 for VPY in the and VPY homologs in species such as to light on the of symbiotic intracellular infection in plants et al., 2020; et al., 2020; & 2021). We Liu and from the for Life of and of for their with We all of the for their was by for in the of and by and of and and the for the of and designed the and performed the and made the protein and the and the with from all All data generated in this study are in the and Supporting Information of this of VPY localization of VPY at the tip of an infection of VPY and VPYD394A localization in the root hair. VPYD394A to vpy-2 of vpy-2 transformed with empty of vpy-2 transformed with a construct containing tree of plant VPY test between VPY variants and LIN. of Asp394 to or VPY localization pattern. localization in root containing an infection localization in root containing an infection Protein of VPY and for VPY variants in yeast cells used for yeast of VPY, and of cytoplasmic and localization pattern for VPY VPY homologs protein sequences and Table used in this Table species and VPY homologs used for and is not for the or of any Supporting Information by the than be to the The is not for the or of any by the than be to the for the