The penetration of sunflower root tissues by the parasitic plant <i>Orobanche cumana</i> is intracellular

Abstract

Sunflower broomrape (Orobanche cumana) is one of the main pests for sunflower crops. This holo-parasitic plant is specific to sunflower crops. Broomrape seeds perceive their host thanks to germination stimulants present in sunflower root exudates (Bouwmeester et al., 2021). Once germinated, the broomrape radicle grows toward the host root (Krupp et al., 2021) and develops papillae, which adhere to the host root and secrete mucilaginous compounds (Joel & Losner-Goshen, 1994). Subsequently, epidermal cells at the tip of the haustorium, a specific parasitic organ, differentiate into intrusive cells that penetrate the host root (Masumoto et al., 2021). This penetration combines physical pressure and degradation of sunflower root cell walls thanks to pectolytic activity enzymes released by the parasitic plant (Shomer-Ilan, 1993; Losner-Goshen et al., 1998). Intrusive cells make their way toward the host root vessels, crossing the successive host root tissues. Transcriptomic analyses showed that, in the case of a susceptible interaction, defense genes were activated only transiently and at a low level (Dos Santos et al., 2003a,b; Letousey et al., 2007). In addition, the expression of the putative defense suppressor gene Par1 of various parasitic plants at the early stages of interaction (Yang et al., 2020; Qiu et al., 2022), suggests manipulation of their host by parasitic plants. Once in contact with the host xylem vessels, intrusive cells differentiate into vessel elements and vascular connections are established (xylem as well as phloem), to insure the nutrient supply of the parasite (Krupp et al., 2019). Although numerous studies have been performed on parasitic seed germination and haustorium development (Yoshida et al., 2016), most of them were focused on the parasitic plants, and the host cellular mechanisms involved during the intrusive cell development were poorly described (Mutuku et al., 2021). How the host cells behave during the massive expansion of the haustorium tissues across the outer root cell layers remained quite unknown. A few studies published in the 1970–90s explored the host cellular reorganization during the early stages of the haustorium penetration of various Orobanchaceae parasitic plant species. It was shown that haustorium development is accompanied by unusual host cell proliferation (Dörr & Kollmann, 1974; Kuijt, 1977). Whether the penetration is intra- or intercellular in the root host was rarely stated. Dörr & Kollmann (1974) and Kuijt (1977) mentioned intercellular growth only of the haustorial cells, with no observations of plasmodesmata interconnecting host and parasite cells for the interactions O. crenata/Vicia faba and O. ramosa/Cannabis sativa. Intercellular penetration between two cortical cells was shown during the interaction between Striga gesnerioides (another Orobanchaceae species) and cowpea (Vigna unguiculata; Reiss & Bailey, 1998). By contrast, the work by Dörr (1969) on the stem parasitic plant species Cuscuta (Convolvulaceae family) on the host Pelargonium zonale revealed intracellular as well as intercellular penetrations preceding the vascular connection between the host and the parasite (Press et al., 1990). In addition, Musselman & Dickison (1975) showed an example of an intrusive cell of the parasitic plant Agalinis aphylla (Orobanchaceae family) penetrating intracellularly a cortical cell through a small opening in the cell wall. Thus, whether sunflower root penetration by the broomrape haustorium is intra-and/or intercellular remained an open question. This knowledge is required in the perspective of subsequently investigate and understand the sunflower cellular mechanisms associated with resistances to O. cumana. In this work, using an efficient selection of the early stages, and combining various microscopy approaches including live-cell imaging of transgenic fluorescent host tissues, we reinvestigated the relationships between host and parasitic tissues at the cell level during the early stages of haustorium penetration. The questions we addressed were as follows: (1) Do intrusive cells penetrate the host root inter or intracellularly? (2) Do the sunflower root cells in the vicinity of the intrusive cells die or stay alive? (3) Are sunflower cell divisions induced at the very early stages of the penetration, and which are the root tissues involved? To answer these questions, we needed to observe attachments at very early stages, that is, haustorium penetration sites sampled before the establishment of vessel connections. To this end, we used a dedicated growth and inoculation device called rhizotron, a plexiglass homemade box, which facilitates the observation of inoculated sunflower roots and selection of attachment sites (Le Ru et al., 2021; Supporting Information Notes S1). For large field and transmission electron microscopy (TEM) observation of stained longitudinal sections of attachments, we used root fragments from young inoculated wild-type sunflower plantlets (i.e. nontransformed). In addition, to get more information on the living status and the subcellular organization of the penetrated cells, we observed attachments using in vivo confocal imaging of living inoculated transgenic composite sunflower plants, that is, obtained by Agrobacterium rhizogenes-mediated transformation. This method generated plants with fluorescent roots, expressing the green fluorescent protein (GFP) targeted to the endoplasmic reticulum (ER) (Figs S1, S2; Table S1). Observation of attachments was performed from 4 to 8 d after inoculation (dai) (Table S2). Broomrape rarely penetrated the host root before 6 dai, while most of the haustoria had reached the inner root tissues (inner cortex to the vessels) at 8 dai. The kinetics were very similar whatever the type of plants and microscopy approach. Interestingly, similarly to our observations, Joel & Losner-Goshen (1994) observed the first stages of attachments at 5–7 dai. Germinated broomrape seeds developed papillae at the tip of the radicle when contacting the host root (Fig. S2i; Joel & Losner-Goshen, 1994). Mechanical pressure of the broomrape in contact with sunflower root epidermal cells led to cell wall deformation (Fig. 1a–c). Differentiated intrusive cells at the broomrape radicle tip were strongly stained by toluidine blue O. They displayed a very dense cytoplasm, a reduced vacuole and a large nucleus containing a darkly stained nucleolus, suggesting a high metabolic activity (Fig. 1a,d,g,j). Imaging early stages of broomrape penetration revealed that intrusive cells penetrated the epidermal layer as well as the successive outer cortical layers intracellularly (Fig. 1d–i). In our culture system, sunflower roots had 4–5 cortical cell layers between the epidermis and the endodermis (Fig. S3). Intracellular penetration of sunflower root cells was observed in all the analyzed penetration sites (21 sites for large field microscopy and 21 sites for confocal microscopy, Table S2). The use of the GFP-ER construct provided information about both the cytoplasmic organization and the nucleus position, thanks to the ER outline labeling the nuclear envelope (Genre et al., 2005). In many cases, the nucleus of the penetrated cell was strikingly positioned close to the intrusive cells (Fig. 1b,c,h,i). The nucleus repositioning close to the intruder is reminiscent of the cellular reorganization of plant cells during bacterial and fungal symbiotic or pathogenic interactions (Genre et al., 2005, 2008, 2009; Fournier et al., 2008). It suggests that the host nucleus perceives the intrusive cell, either through the exerted mechanical pressure (Genre et al., 2009) and/or through unknown chemical signals. However, in contrast to root penetration by symbiotic (Genre et al., 2005, 2008) or pathogenic biotrophic fungi (Koh et al., 2005; Kankanala et al., 2007; Genre et al., 2009), no cytoplasmic aggregation, nor specific ER reorganization were observed ahead of the penetration process. Interestingly, ER was surrounding the broomrape intrusive cells (Fig. 1e,f,h,i,k,l), showing active, though not massive, host intracellular reorganization along with the penetration process. These results suggested active membrane synthesis around intrusive cells requiring nucleus and ER activity in the host cell, and showed that the sunflower penetrated cells remained alive. Deeper root tissues (endodermis and pericycle) were also penetrated intracellularly by intrusive cells (Fig. 1j). In most cases, haustoria penetrated the host root with minimal host cell damage. However, the live-cell imaging approach revealed a few cases of cell death as shown by the absence of fluorescence (2 sites, Fig. S4a,b), or a severe ER disruption (1 site, Fig. S4c,d). Similarly, change of the vacuole structure was observed using large field microscopy for a few sites (7 sites among 21 penetration sites): appearing as a blue smear (Figs 1d, S4e) or light blue material filling the cell (Fig. S4e). One or a few penetrated cells only were affected, adjacent to the intrusive cells in outer root tissues. These results suggested that in some cases, penetration of the intrusive cells got out of control and synchronization of the penetration process and the sunflower cellular reorganization failed, leading to sunflower cell death. This phenomenon remained cell autonomous, without other defense reactions in the surrounding or the deepest root tissues. Furthermore, penetration could result in the separation of the host nucleus from the distal part of the penetrated cell, probably leading to cell death as well. Strikingly, broomrape intrusion was thicker in outer root tissues (Fig. 1) than in inner root tissues, in which only single elongated and separated intrusive cells were detected (Fig. 1e,f,h,i,k,l). Similarly, Dörr (1969) reported intracellular ‘searching hyphae’ for the Cuscuta stem parasite. Sunflower roots were known to swell locally at the site of broomrape attachment by means of cell division (Kuijt, 1977) and as early as 7 dai (Dörr & Kollmann, 1974). In the present study, sunflower root cell divisions were observed as early as 6 dai, close to attachments (9 and 7 sites for sections and live-cell imaging, respectively). Divisions were mostly anticlinal in the cortex and periclinal in the pericycle (Figs 1d,e,h,i, S5). The number of dividing root cell layers and the length of the dividing zone were highly variable (for example, from 1 cortical cell to > 30 cells in a row). These divisions may account for root hypertrophy that was previously observed at the site of 14-dai attachments in rhizotrons (Chabaud et al., 2022). These divisions could be induced indirectly (host hormonal regulation) or directly by the parasitic plant (hormonal release: such as auxin (Ishida et al., 2016) or cytokinin (Spallek et al., 2017)). Whether germinated broomrape seed exudates would be sufficient for the induction of host cell divisions remains an open question. The interface between intrusive cells and the sunflower penetrated root cells at early stages of the interaction was further characterized by TEM (Fig. 2). A 7 dai attachment with the haustorium reaching the 3rd cortical cell layer is illustrated Fig. 2(a,b). Starch grains, a sign of the transition from the autonomous (germination stage) to the parasitical stage (Joel & Losner-Goshen, 1994), were present in the central part of the attachment (Fig. 2a,c). In the outer root cell layers, the interface appears as a thick layer surrounding broomrape (Fig. 2b), as already described for Striga (Reiss & Bailey, 1998; Neumann et al., 1999). The intrusive cells were easily distinguished from sunflower root cortical cells thanks to their dense cytoplasm, containing Golgi stacks, large mitochondria as well as a reduced vacuole (Fig. 2b,e,f) as already reported by Kuijt & Toth (1976) and Kuijt (1977). By contrast, the host cortical cells whether penetrated or not, contained a large vacuole with a thin layer of surrounding cytoplasm (Fig. 2b,d). Mitochondria in the penetrated host cells were present all along the host plasmalemma, suggesting intense activity at the periphery of the host cell such as membrane biosynthesis (Fig. 2d). This dense cytoplasm confirmed that the penetrated host cells were alive at this stage. The parasitic cell wall was present all around the intrusive cells. By contrast, the presence of a host matrix along the anticlinal interface of the host penetrated cell was not always detectable and its appearance varied along the length of the haustorium. On the outermost anticlinal side of the host penetrated cell, the host interface with the haustorium appeared as a dark thick layer, in continuity with the existing periclinal host cell wall (Fig. 2b,d). It could partly result from the invagination of the existing periclinal host cell wall pushed in by the penetrating intrusive cells. The discontinuity of the staining suggests disorganization of this host cell wall/matrix. On the innermost side of the cell, the host matrix was either too thin to be visible (Fig. 2e,f) or appeared as a low-density material (stars in Fig. 2g,h) separating the host cell plasmalemma from the parasitic cell wall, and differing from the existing darker periclinal host cell wall. The use of various fluorescent dyes to distinguish host cell wall from newly made matrix would be interesting, as done for the symbiotic nitrogen-fixing bacterial infection thread (Rae et al., 2021). At the frontline of the haustorium penetration the existing periclinal host cell wall seemed also disorganized (Fig. 2i), suggesting progressive local enzymatic degradation of the host cell wall. This apparent dissolution of the nearby host cell walls (Kuijt, 1977) or a partial digestion of the cell wall at the interface (Kurotani et al., 2020, in the case of the interaction Phtheirospermum japonicum/Arabidopsis thaliana) had been reported previously. While cell wall degrading enzymes which might contribute to this process have been identified from the parasite (Shomer-Ilan, 1993; Losner-Goshen et al., 1998), there is no evidence at the moment of the direct involvement of host enzyme activities involved in host cell wall degradation in this context (Mitsumasu et al., 2015; Yang et al., 2020). Nevertheless, the host cell plasmalemma seemed to remain undisturbed and continuous (Figs 2g, S6b). Both host and parasitic plasma membranes were highly convoluted at the front line of the haustorium (Fig. S6b), suggesting membrane synthesis for the haustorium accommodation (host) and haustorium expansion (parasite). No plasmodesmata were observed on the interface at these early stages, indicating that molecular exchanges between the parasite and the host happened at later stages, or through vessel connections, as interspecific plasmodesmata have been shown in the phloem (Krupp et al., 2019). In some cases, the penetration led to disaggregation of the vacuole of the host cell (Fig. S6c,d), with disruption of the host cell plasmalemma, leading to cell necrosis, as for Striga (Neumann et al., 1999). However, as mentioned above, this was not very common and remained cell autonomous. In addition, no evidence of cell death was observed at the later stages (Chabaud et al., 2022). Altogether, these results showed that the parasitic intrusive cells penetrate the host root cells intracellularly, as a result of degradation of the host cell wall and formation of a new host trans-cellular apoplastic compartment for haustorium accommodation. Most striking among our findings has been the observation of intracellular haustorium penetration of host root tissues, in contrast to most studies on Orobanchaceae. These studies relied mainly on the observation of transverse sections, in contrast to the longitudinal sections used in this work, which made it easier to distinguish between intra- and intercellular processes. Our work showed the intimate broomrape penetration into its host, through the formation of a new apoplastic compartment. It suggested that although host cell wall integrity has been damaged by parasitic cell wall degrading enzymes (Shomer-Ilan, 1993; Losner-Goshen et al., 1998), only minor defense reactions were induced as previously reported for biotrophic pathogenic fungi (Mendgen & Hahn, 2002; Bellincampi et al., 2014). In that respect, genes encoding inhibitors of cell wall degrading enzymes could be good candidates for increasing resistance to broomrape. In addition, as HaOr7 (Duriez et al., 2019) and HaOrDeb2 (Fernandez-Aparicio et al., 2022) encode Leucine-Rich-Repeat Receptors Like Proteins, providing resistance to various O. cumana races, it would be of outstanding interest to characterize the cellular processes involved in these incompatible interactions. Comparing the cellular processes for various O. cumana races could highlight common or different mechanisms. Finally, using these approaches on other major parasitic plant species such as Striga will be of great interest for future resistance development in a larger host range. We thank P. Gresshoff (University of Queensland, Australia) for the A. rhizogenes strain K599. This study was supported by the ‘Laboratoires d'Excellences (LABEX)’ ‘Towards a Unified theory of biotic interactions: roLe of environmental Pertubations’ (TULIP; ANR-10-LABX-41) and/or by the ‘École Universitaire de Recherche (EUR)’ TULIP-GS (ANR-18-EURE-0019). Slide scanning was performed using the Nanozoomer from the Imagery Platform of the Federated Research Institute AgroBiosciences-Interactions-Biodiversité (FRAIB; Castanet-Tolosan, France). This study was performed in the frame of a 2-year project (SunOrCell), funded by ‘Promosol/SeleoPro’ (the association of French Sunflower and Rapeseed Breeders for promoting these crops). The International Consortium of Sunflower Genomics (ICSG) supported the grants for the training students CG and ARS. None declared. M-CA performed cytological experiments (large field microscopy and TEM). CG, AR-S and AL established sunflower transformation experiments. M-CB produced sunflower and broomrape resources. SM coordinated the ICSG, assisted with the construction of the project and the writing of the manuscript. JF gave technical and scientific advice and assisted with the construction and the writing of the manuscript. MC designed the experiments, carried out confocal microscopy and cytology experiments and wrote the manuscript. Fig. S1 Schedule of sunflower transformation experiments, broomrape inoculation of composite plants for confocal microscopy observations. Fig. S2 The various steps of transformation of sunflower plants via Agrobacterium rhizogenes, transfer of composite plants in rhizotrons, broomrape inoculation and observation using confocal microscopy. Fig. S3 Longitudinal section of a sunflower root. Fig. S4 Sunflower cell death or endoplasmic reticulum destructuring associated with haustorium penetration. Fig. S5 Multiple divisions in the host root at the site of broomrape intrusion. Fig. S6 Transmission electron microscopy of a 7 dai attachment. Notes S1 Details of the materials and methods. Table S1 Efficiency of sunflower transformation via Agrobacterium rhizogenes. Table S2 Overview of observed sites. Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts