Entanglement of plant immunity and endomembrane trafficking revealed by plant–powdery mildew fungal interactions

Summary

Plant immunity is complex, and studies of leaf epidermal cells attacked by powdery mildew fungi have been instrumental in revealing how it relies on plant endomembrane trafficking. Immunity against these biotrophic fungi is manifested as cell wall deposits (‘papillae’) and the hypersensitive reaction (HR), both involving plant endomembrane traffic. Papillae contain extracellular vesicles (EVs), some of which depend on the ‘endosomal sorting complexes required for transport’ machinery, while others depend on mechanisms yet to be uncovered. Existence of two such EV pathways agrees with several old electron microscopy observations of papillae. Interestingly, resistance protein‐activated HR has been shown to be a form of immunity also depending on membrane trafficking, namely the pathway to the vacuole.

	Contents
	Summary	799
I.	Introduction	799
II.	Two‐layered cell wall depositions in response to powdery mildew fungi formed by separate membrane trafficking pathways	800
III.	Second layer cell wall deposition is late endosome/ESCRT‐dependent	801
IV.	First layer cell wall deposition is ESCRT‐independent	802
V.	PM fusion of both ESCRT‐dependent and independent endosomes is mediated by the same SNARE proteins	802
VI.	Potential role of RabA GTPase and the exocyst tethering complex in papilla formation	803
VII.	The late endosome pathway is required for ETI	803
VIII.	Conclusions	803
	Acknowledgements	803
	References	803

I. Introduction

Plant diseases severely affect crops, and introduced resistances are generally only efficient for a few years as they are overcome by new pathogen genotypes. Yield reductions impact farmers' economy and inflict demand for more agricultural land to secure food availability, which in turn has global consequence by negatively affecting biodiversity and carbon sequestration. Furthermore, fungicides for disease control may negatively affect environmental and human health. Altogether, this underlines the need for deeper insight into the plant immune system, allowing it to be managed ingeniously to provide efficient and durable disease resistance.

The plant immune system is activated in two steps (Jones et al., ²⁰²⁴). Pathogen‐associated molecular patterns bind to plasma membrane (PM)‐localized pattern recognition receptor (PRR)‐kinase complexes and activate pattern‐triggered immunity (PTI). PTI responses generally involve transcriptional reprograming, deposition of callose and other cell wall materials at sites of attack, and secretion of antimicrobial proteins and metabolites (DeFalco & Zipfel, 2021). These deposits can be polarized at attack sites in what is referred to as ‘papillae’, which often stop invasions by filamentous pathogens (e.g. Chowdhury et al., ²⁰¹⁴). Pathogens transfer effector proteins to the host tissue to bind and suppress proteins essential for immunity (Yuen et al., ²⁰²³; Bhandari & Brandizzi, ²⁰²⁴). Meanwhile, plants can detect such effectors and activate effector‐triggered immunity (ETI). This involves nucleotide‐binding leucine‐rich repeat (NLR) receptors, classified as coiled‐coil NLR (CNL), toll‐interleukin NLR (TNL) and RPW8‐like NLR, according to their N‐terminal domains. Upon effector recognition, NLRs oligomerize to form ‘resistosomes’, and usually an associated prolonged Ca²⁺‐influx into the cytosol activates ETI responses, often including hypersensitive reaction (HR) (Jones et al., ²⁰²⁴).

Powdery mildew fungi attack individual leaf epidermal cells, which makes the cellular immune responses highly amenable for microscopy. Thus, studies of cellular delivery of material to papillae have revealed interesting details on how membrane trafficking is crucial for immunity. We will discuss these findings and how some of them challenge current views on membrane trafficking in plants.

Membrane trafficking transports proteins, lipids and metabolites in eukaryotic cells by help of numerous proteins, as described in some excellent reviews (e.g. Vukašinović & Žárský, ²⁰¹⁶; Singh & Jürgens, ²⁰¹⁸; Rodriguez‐Furlan et al., ²⁰¹⁹; Ito & Uemura, ²⁰²²; Shimizu & Uemura, ²⁰²²). In short, transport of vesicles carrying cargo molecules involves vesicle budding, transport and fusion to their right destination, all processes tightly regulated by specific proteins. These include small GTPases, switching between an inactive GDP‐bound cytosolic state, and an active GTP‐ and membrane‐bound state. GTPases are activated by guanine‐nucleotide exchange factors (GEFs), exchanging GDP for GTP. Vesicle budding is regulated by ARF/SAR1‐GTPases, while fusions of vesicles, and of endosomes, to target membranes are regulated by Rab‐GTPases. Typically, proteins destined for exocytosis or organelles are synthesized at the endoplasmic reticulum (ER) and, via the Golgi, transported to the trans‐Golgi network (TGN). Interestingly, plants appear to lack an equivalent to the early endosome in mammals. Instead, the TGN functions as sorting hub sending proteins to the PM by secretory vesicles in the exocytic pathway or to the vacuole in a process involving multivesicular bodies (MVBs). Membrane proteins destined for vacuolar degradation become incorporated into MVB intraluminal vesicles (ILVs) by the endosomal sorting complexes required for transport (ESCRT) complex. The subsequent MVB‐to‐vacuole fusion involves Rab5‐to‐Rab7 transition mediated by MON1/CCZ1, and binding of GTP‐bound Rabs to the CORVET and HOPS tethering complexes. Meanwhile, proteins and lipids retrieved from the PM are transported in endocytic vesicles to the TGN, either to be recycled back to the PM or to be sent for degradation through the vacuolar pathway. Fusion to the PM involves Rab‐GTPases and the exocyst tethering complex, including Exo70. Finally, membrane fusions are mediated by SNARE proteins (e.g. syntaxins). When they intertwine and form the SNARE complex, the two membranes are forced together.

These two plant cellular phenomena, immunity and membrane trafficking, are not only exceedingly complex, but are also highly entangled (Nielsen, 2024). This became evident with the discovery of two orthologous Syntaxins of Plants (SYPs) named PENETRATION 1 (PEN1/SYP121/SYR1) and required for mlo‐specified Resistance 2 (ROR2) in Arabidopsis and barley (Collins et al., ²⁰⁰³). Recent excellent reviews have described pathogen effectors targeting membrane trafficking components in different patho‐systems (Yuen et al., ²⁰²³; Bhandari & Brandizzi, ²⁰²⁴), highlighting the importance of these pathways for the regulation of immunity. Many effectors, for instance, inhibit conventional vesicle trafficking to the PM and thereby affect secretion of antimicrobial proteins to the apoplast and insertion of PRRs in the PM. An example is the effector RxLR24 from the potato late blight pathogen, Phytophthora infestans, that targets members of the RabA subfamily mediating vesicle–PM fusion (Tomczynska et al., ²⁰¹⁸). RabAs interact with the exocyst complex at the PM, and, for example, the Xanthomonas core effector, XopP, targets the exocyst Exo70B1 to inhibit exocytosis (Michalopoulou et al., ²⁰²²). An example from the barley powdery mildew fungus (Blumeria hordei, Bh) is the effector BEC4, targeting an ARF‐GTPase‐activation protein. Knockout of its Arabidopsis orthologue suggested its importance for preinvasive immunity to powdery mildew fungi (Schmidt et al., ²⁰¹⁴). Besides advancing the understanding of plant–pathogen interactions, these studies have also helped reveal important unknowns in plant membrane trafficking.

II. Two‐layered cell wall depositions in response to powdery mildew fungi formed by separate membrane trafficking pathways

The PM syntaxins, PEN1 and ROR2, are important for penetration resistance to Bh, and this polar deposition of defence barriers at the cell interface in response to filamentous pathogens has been well studied using microscopy techniques. Arabidopsis pen1 and barley ror2 mutants apparently form normal papilla at attack sites, but at a 2‐h delay, explaining the reduced penetration resistance (Assaad et al., ²⁰⁰⁴; Böhlenius et al., ²⁰¹⁰). These studies and Bhat et al. (2005) remarkably observed the PM transmembrane proteins, PEN1 and ROR2, to be deposited inside the papilla structure, outside the PM. This was later found to be due to these proteins labelling extracellular vesicles (EVs), which are typically derived from the release of ILVs into the apoplast upon MVB–PM fusion (Meyer et al., ²⁰⁰⁹; Nielsen et al., ²⁰¹²; Rutter & Innes, ²⁰¹⁷). The localization of PEN1 and ROR2 at EVs is likely due to a ‘cellular mistake’, reflecting insufficient sorting at the limiting membrane of the MVB when redirecting cargo to the ILVs. This makes syntaxins nonfunctional in EVs, having their active SNARE domain inside the EV lumen. Nonetheless, PEN1/ROR2‐labelled EVs being in papillae indicates polarized MVB fusion to the PM, as indicated by transmission electron microscopy (TEM; An et al., ^2006a,b).

TEM by Heitefuss & Ebrahim‐Nesbat (1986), Assaad et al. (2004), An et al. (2006a,b), Chowdhury et al. (2014) and others clearly show that papillae deposited in barley and Arabidopsis at Bh attack sites consist of two distinct layers. In barley, the first deposited (inner) layer is highly electron dense, while the second layer deposited on top is electron light. Heitefuss & Ebrahim‐Nesbat (1986) also found that this second layer extends along the haustorial neck when the fungus successfully penetrates and forms the haustorial feeding structure in the plant cell. In certain plant–powdery mildew fungal interactions, such as Arabidopsis attacked by Golovinomyces orontii, haustoria become more or less fully encased in callose deposits, and from TEM studies these ‘encasements’ appear continuous with the second layer of the papillae (Heitefuss & Ebrahim‐Nesbat, 1986; Micali et al., ²⁰¹¹). Their different compositions suggest that the first layer of the papillae is formed by a membrane trafficking pathway different from the one forming the second papilla/encasement layer. The revealing of these pathways is described below. See also Nielsen (2024).

III. Second layer cell wall deposition is late endosome/ESCRT‐dependent

The haustorial encasement, which appears to be an extension of the second layer of the papilla, was strongly inhibited in Arabidopsis vps9a‐2, whereas papillae are barely affected (Nielsen et al., ²⁰¹⁷). VPS9a is the GEF of Rab5's, which in turn are essential for activation of MON1/CCZ1, the GEF of Rab7's. This indicated the late endosome/ESCRT‐based MVBs, normally fusing with the tonoplast, to be essential for encasement formation. Additionally, both papillae and encasements contain PEN1, which was clearly reduced in encasements of the vps9a‐2 mutant. Besides, the constitutively active Rab5^Q69L ended up in encasements together with PEN1, but not in papillae (Nielsen et al., ²⁰¹⁷). The involvement of ESCRT‐based MVBs was later substantiated in barley, when MON1 and VPS18 were found to be targeted by Bh effectors CSEP0162 and CSEP0214, respectively. Over‐expressing CSEP0214 and CSEP0162, as well as silencing MON1 in barley, inhibited encasement formation, as did mutating Arabidopsis MON1. Meanwhile, these manipulations of the latter hardly affected papilla formation and penetration resistance (Nielsen et al., ²⁰¹⁷; Liao et al., ²⁰²³; Deb et al., ²⁰²⁵; Sabelleck et al., ²⁰²⁵). In addition to PEN1, the tetraspanin TET8, known to label late endosome/ESCRT‐based EVs in HeLa cells (Larios et al., ²⁰²⁰), is also found in encasements (Rubiato et al., ²⁰²²). In summary, these observations point at the second layer of the papilla and the encasement to be deposited by the involvement of ESCRT‐based MVBs. However, a constitutively expressed vacuolar marker neither accumulated in papillae nor in encasements, arguing against direct involvement of the vacuolar trafficking pathway (Munch et al., ²⁰²⁶). It therefore remains unresolved which pathway transports cargo for the encasement, and whether this pathway also plays a role in other situations. Interestingly, the target of CSEP0214, VPS18, which is a component of the CORVET and HOPS tethering complexes, is required for encasement formation (e.g. Sabelleck et al., ²⁰²⁵). This observation is striking, as we would expect the exocyst to be mediating fusions to the PM, as discussed below.

IV. First layer cell wall deposition is ESCRT‐independent

Inhibition of the Arabidopsis ARF‐GEF, GNOM, suppressed initial secretion of membrane‐bound PEN1 at Bh attack sites. As GNOM plays a nonredundant role in recycling of PM proteins, this indicated that PEN1 is recycled from the PM and secreted to the papilla on MVB‐derived EVs (Nielsen et al., ²⁰¹²). However, studies by Nielsen et al. (2017) and Liao et al. (2023) indicated that papilla formation and penetration resistance are poorly affected by inhibition of the vacuolar pathway, which typically involves ESCRT‐dependent MVBs. This suggests that the first layer of the papillae, containing PEN1/ROR2‐labelled EVs, is deposited by a membrane recycling pathway, involving ESCRT‐independent MVBs. This was confirmed in barley epidermal cells where mCherry‐ROR2‐containing papillae still formed despite co‐expression of the dominant‐negative mYFP‐VPS4^E232Q, arresting ESCRT‐III and stimulating formation of incomplete MVBs (Doukoudaki, 2023). This result was repeated in Fig. 1, also showing co‐localization of mYFP‐VPS4^E232Q and mCherry‐ROR2 in endosomes, likely to be incomplete MVBs carrying material aimed for the ESCRT‐dependent EVs in the second papilla/encasement layer.

Fig. 1.

Endosomal sorting complexes required for transport (ESCRT)‐independent extracellular vesicle (EV) secretion at powdery mildew fungal attack site. Blumeria hordei (isolate C15)‐attacked barley cv Ingrid epidermal cells transiently transformed by particle bombardment to express the extracellular vesicle marker, mCherry‐ROR2 and the dominant‐negative ESCRT‐III component, mYFP‐VPS4^E232Q, imaged 18 h after inoculation. mYFP was used as a negative control. Closed arrow, extracellular accumulation of mCherry‐ROR2. Arrowheads, co‐localized mCherry‐ROR2 and mYFP‐VPS4^E232Q in cellular aggregates, likely to be incomplete MVBs. Open arrow, fungal appressorium. Bars, 10 μm. ROR2, required for mlo‐specified Resistance 2; YFP, yellow fluorescent protein; MVB, multivesicular body.

V. PM fusion of both ESCRT‐dependent and independent endosomes is mediated by the same SNARE proteins

While pen1 single mutants have distinct penetration resistance phenotypes (Collins et al., ²⁰⁰³; Assaad et al., ²⁰⁰⁴), indicating a PEN1‐specific function, the pen1 syp122 double mutants, also mutated in the closest homologue of PEN1, have a strong autoimmune phenotype (Schultz‐Larsen et al., ²⁰¹⁸). This shows existence of shared PEN1/SYP122 function(s). Such shared function(s) were demonstrated by Rubiato et al. (2022). They showed that the PEN1 and SYP122 PM syntaxins together are required for deposition of material for both the primary papilla and the secondary papilla/encasement, not only at Bh attack sites, but also at the attack site of Colletotrichum destructivum and Phytophthora infestans. This suggests that PEN1 and SYP122 mediate fusion of both ESCRT‐dependent and independent MVBs to the PM.

VI. Potential role of RabA GTPase and the exocyst tethering complex in papilla formation

Powdery mildew resistance, resulting from knockout of Mlo genes, is known in many plant species. The mechanism is not well understood but involves enlarged papillae and loss of fungal penetration (Kusch & Panstruga, 2017). Recently, Arabidopsis MLO6 has been found to interact with the exocyst component, EXO70H4, in leaf trichomes, while being important for normal localization of the immunity‐associated callose synthase, PMR4 (Huebbers et al., ²⁰²⁴). EXO70H4, homologous to EXO70B2 and previously found to interact with PEN1, contributes to secretion of PEN1‐labelled EVs into papillae (Ortmannová et al., ²⁰²²). This leads Huebbers et al. (2024) to suggest that MLO contributes to the exocyst‐mediated tethering of endosomes carrying callose synthase while interacting with the PEN1 SNARE complex. In mammals, the exocyst tethers Rab11 GTPase during endosome/PM fusion (Takahashi et al., ²⁰¹²). Thus, it is potentially relevant that the Arabidopsis Rab11 subfamily member, RabA2a, interacts with PEN1 and SYP122 (Pang et al., ²⁰²²). Future studies will hopefully unravel the functional links between these elements and how they contribute to formation of the two‐layered papillae.

VII. The late endosome pathway is required for ETI

It is noteworthy that inhibition of vacuolar traffic prevents execution of HR mediated by certain CNLs and TNLs (Munch et al., ²⁰¹⁵; Hatsugai et al., ²⁰¹⁸; Schultz‐Larsen et al., ²⁰¹⁸). More recently, silencing of MON1 and introduction of the Bh effector CSEP0214, targeting VPS18, inhibited CNL‐mediated HR in barley (Liao et al., ²⁰²³; Sabelleck et al., ²⁰²⁵). The precise mechanism behind this remains unclear, although analyses of Arabidopsis vacuolar sorting receptor mutants have revealed that blocking vacuolar traffic during immunity inhibits transport of hydrolytic enzymes associated with cell death, as well as tonoplast‐PM fusion previously associated with the execution of HR (Hatsugai et al., ²⁰⁰⁹; Zhu et al., ²⁰²⁵).

VIII. Conclusions

In summary, studies of microscopy‐friendly polarized depositions of EVs, callose, etc. into papillae in powdery mildew fungal attacked epidermal cells have revealed that the first papilla layer is ESCRT‐independent, while the second layer, sometimes including an encasement, is ESCRT‐dependent and deposited using vacuolar trafficking proteins. Fusion of both ESCRT‐dependent and independent MVBs to the PM, releasing ILVs as EVs, is mediated by PEN1 and SYP122 syntaxins in Arabidopsis. Besides these PTI responses, NLR‐mediated ETI also appears to depend on vacuolar traffic. See model in Fig. 2. These observations leave us with the remaining open questions:

1. How are ESCRT‐dependent MVBs redirected to the PM? We observe involvement of Rab5/Rab7 and CORVET/HOPS tethering complex in encasement formation, normally associated with MVB‐tonoplast fusion. Yet, encasements lack accumulation of a vacuolar marker. The PM fusion depends on PEN1 and SYP122, but how RabA's and the exocyst are involved awaits future studies.

2. How are ILVs formed independently of the ESCRT machinery?

3. Do these unconventional protein secretion pathways play a role in other contexts of plant development or plant immunity?

4. What are the precise molecular mechanisms behind the dependency of NLR‐mediated HR on vacuolar traffic?

Fig. 2.

Model for membrane trafficking in the immune response to powdery mildew fungi. The papilla is made up of two layers, where the second extends into the encasement around the haustorium. The first and second layers are deposited by endosomal sorting complexes required for transport (ESCRT)‐independent and ESCRT‐dependent multivesicular body (MVB) pathways, respectively. Both layers contain extracellular vesicles (EVs) labelled by the syntaxin PEN1 in Arabidopsis and ROR2 in barley. In Arabidopsis, the second layer also contains SYP122‐labelled EVs. The ESCRT‐dependent pathway is in addition required for nucleotide‐binding leucine‐rich repeat (NLR)‐mediated hypersensitive reaction (HR) and targeted by Blumeria hordei effectors, CSEP0162 and CSEP0214.

Competing interests

None declared.

Disclaimer

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.