Cell death is a conserved innate disease resistance response of brown algae against the oomycete Eurychasma dicksonii

Summary

Brown algae (Phaeophyta) encompass key primary producers of temperate and cold coastal seas, such as kelps, the cultivation of which is rapidly expanding worldwide. Here, we show that across ten brown algal species, innate resistance against the intracellular oomycete pathogen Eurychasma dicksonii is mediated by local cell death and accompanied by cell-wide deposition of β1-3 glucans and fluorescent metabolites, the accumulation of reactive oxygen species, and the expression of programmed cell death (PCD) markers. This response also occurs in compatible strains for a fraction of the infected algal cells, which makes it a quantitative trait. It is followed by the induction of two inducible autophagy-mediated defense responses already known for the oomycete pathogen Anisolpidium ectocarpii. Overall, we unveil hypersensitive-like cell death as a conserved, core part of brown algal innate defenses, that is intertwined with a second line of inducible, autophagy-mediated defenses. These results open unprecedented avenues to elucidate the molecular mechanisms of pathogen recognition and disease resistance in brown algae.

Graphical abstract

Highlights

• •We describe the cellular mechanisms underpinning brown algal defenses against a pathogen
• •A local cell death, hypersensitive-like, response provides a first line of defense
• •This response also occurs quantitatively in brown algae susceptible to this pathogen
• •Inducible autophagy, both in the pathogen and its host, provides a second line of defense

Biological sciences; Molecular biology; Botany

Introduction

Brown algae (Phaeophyta) are a monophyletic group of multicellular heterokonts (Straminipila), one of many phototrophic, plastid-bearing eukaryotic lineages collectively known as algae.^1,2 Brown algae, particularly kelps and fucoids, form highly productive underwater canopies in cold and temperate marine coastal ecosystems, thus providing habitat for numerous organisms. Additionally, brown algal cultivation has been expanding rapidly worldwide for a wealth of applications, mainly as food and feed for aquaculture.^3,4 Like any other organism, brown algae are plagued by pathogens, which are thought to contribute to population structuring and to act as pleiotropic environmental and evolutionary drivers.⁵ Brown algal canopies are declining globally, but the possible role of pathogens in this process is largely unexplored.^6,7 On the other hand, the expansion of seaweed aquaculture is accompanied by an increasingly severe impact of diseases in farms, with hardly any control method in place.⁸ In fact, knowledge about algal defenses is mostly restricted on early signaling events, notably lipid signaling and reactive oxygen species production triggered by chemical elicitors.⁹ There are hardly any data as to how these signaling mechanisms actually lead to defense or resistance against pathogens, mostly due to the limited availability of pathosystems amenable to physiological, molecular, and genomic investigation. Developing this knowledge is key not only to understand the impact of pathogens on natural algal populations but also to design seaweed-breeding programs.

Historically, the oomycete pathogen Eurychasma dicksonii was described as “the most common and widespread of the marine fungi” in coastal environments.¹⁰ It is an obligate intracellular biotroph that parasitizes individual cells of its brown algal hosts (Figure 1). While in the field, Eu. dicksonii is most commonly reported causing epidemic outbreaks in filamentous Ectocarpales,¹¹ it can infect at least brown algal 45 species in laboratory cultures,¹² including the gametophytes of commercially relevant kelp species. We previously reported that clonal strains of the genome model Ectocarpus exhibit differential susceptibility to Eu. dicksonii isolates,¹³ thus evidencing an innate component to the resistance of some algal genotypes against this pathogen. Separately, time course inoculations performed on a disease-susceptible, gametophytic strain of the giant kelp Macrocystis pyrifera led us to identify inducible, local, and systemic defenses against another intracellular, obligate biotroph oomycete, Anisolpidium ectocarpii.¹⁴

Figure 1.

Infection cycle of the oomycete Eurychasma dicksonii in brown algae

(A) The infection cycle starts when a zoospore encysts at the surface of an algal cell (A, arrowhead), and injects its content into the host cytoplasm (inset in A).

(B and C) Eu. dicksonii then grows intracellularly (B, double arrowhead), progressively filling its host cell and causing hypertrophic growth (C, double arrowhead).

(D) Each Eu. dicksonii thallus ultimately differentiates into a sporangium (D, double arrowheads and inset), which releases new zoospores (arrows).

(E and F) In this study, we show that resistance against infection is mediated by the early (E) or late (F) death of the attacked host cell (asterisks), preventing the completion of the Eu. dicksonii development cycle.

Arrowheads in (A), (C), and (E): attacking Eu. dicksonii spore.

Scale bars, A–C, E, and F, inset in D: 10 μm; D: 100 μm; inset in A: 1 μm.

In many eukaryotes, programmed cell death (PCD) and autophagy often interlink in mediating innate and inducible (or acquired) immune reactions. Based on the available evidence, the molecular machinery in PCD is highly conserved in a trans-kingdom way,¹⁵ and involves tightly structured signaling cascades (e.g., mediated by oxidative stress) and effector dynamics, with the orchestrated activity of several caspases.¹⁶ Similarly, autophagy involves the performance of autophagy-related genes (ATGs), that are not only present but also are very conserved between eukaryotes.¹⁷ In land plants, the hypersensitive response (HR) was originally defined phenotypically as a rapid, localized death of the first infected cell(s), resulting in the efficient, though not always total restriction of the pathogen at the point of ingress. Subsequently, and despite some mechanistic variations between species, this phenotypic cell death response was tightly associated with the execution of PCD, the deployment of defenses such as cell wall reinforcement and the accumulation of phytoalexins, together with the plant’s ability to specifically recognize the pathogen with resistance genes.¹⁸ However, aside from a few close relatives of land plants (i.e., the Charophyceae), algal lineages—including brown algae—lack clear orthologues of plant disease resistance genes¹⁹ or other known immune receptors, although two fast-evolving families of LRR GTPases and NB-ARC kinases have been speculated to mediate pathogen detection in Ectocarpus.²⁰ Yet, cell death responses phenotypically resembling the plant HR have been described in the diatom Asterionella formosa against the chytrid Rhizophydium planktonicum²¹ and in the red seaweeds Ceramiales and Gracilariales against red algal parasites, oligosaccharide and protein elicitors of bacterial origin.^22,23 Mechanistic investigations unveiled the induction of PCD in diatoms,²⁴ chlorophytes,²⁵ and haptophytes²⁶ in response to abiotic stress. To date, the unicellular haptophyte Emiliania huxleyi is the only algal lineage in which PCD, and to some extent autophagy, have been implicated in the resistance against bacterial and viral pathogens.²⁷

Here, we investigated the mechanism underpinning the innate resistance of brown algae against Eu. dicksonii, both at macroscopic and cellular levels, in several species of Ectocarpus. Having identified markers of defenses and of PCD, we then took advantage of the pathogen’s broad host spectrum to investigate their conservation across ten species, spanning four different brown algal orders. Finally, we investigated the link between PCD-mediated innate resistance and the onset of inducible autophagy-mediated defenses.

Results

Innate differential susceptibility to Eurychasma dicksonii across brown algae

Building on earlier observations of differential susceptibility of the model alga Ectocarpus to the oomycete pathogen Eu. dicksonii,^12,13 we expanded the scope of our host range study to three strains of Eu. dicksonii and ten algal species that span four different orders (Ectocarpales, Laminariales, Tilopteridales, and Discosporangiales, see key resources table). For the sake of consistency with the terminology used for land plants and red seaweeds, we distinguished incompatible from compatible interactions. We define the former as interactions where complete phenotypic resistance was observed, translating into the absence of any mature pathogen sporangium under optic microscopy; in contrast, the latter are interactions where the completion of the pathogen development cycle was observed, leading to the release of new infectious spores. In this manner, five algal strains incompatible with one or more strains of Eu. dicksonii were identified, corresponding to four algal species (Ectocarpus siliculosus, Ec. crouaniorum, Ec. fasciculatus, and Pylaiella littoralis; Table 1).

Table 1.

Compatibility and incompatibility between the algal and pathogen strains used in this study

	Eu. dicksonii 05	Eu. dicksonii 96	Eu. dicksonii 06
Ec. sp. 022-10	I (a)	I (a)	I (a,b)
Ec. siliculosus 32m	C (N/D)	C (a)	C (N/D)
Ec. fasciculatus 007-04	C (a)	I (a)	C (a)
Ec. crouaniorum	I (a)	C (N/D)	C (a)
P. littoralis IRg	I (a)	C (a)	I (a)
P. littoralis Ros	N/D	N/D	I (a)
P. littoralis CH	I (a)	N/D	I (a)
P. littoralis BHI	I (a)	N/D	I (a)
A. crinita	C (N/D)	C (N/D)	C (a)
L. digitata	C (a)	C (a)	C (a)
M. pyrifera PU	C (a)	C (a)	C (a)
M. pyrifera FB	C (a)	C (a)	C (a)
M. pyrifera Mau	C (a)	C (a)	C (N/D)
T. mertensii	C (a)	C (a)	C (a)
C. tenellus	C (N/D)	C (N/D)	C (a)
D. mesarthrocarpum	C (a)	C (a)	C (a)

I, incompatibility (full resistance); C, compatibility (partial or complete susceptibility to infection); N/D, not determined.

^aOccurrence of hypersensitive cell death. See strain details in Table S1.

^bIn one occasion (one independent experiment), the completion of the infection cycle was observed in one experiment, in a few challenged cells. In all other experiments performed by several experimenters over a decade, this interaction was incompatible.

Incompatibility is mediated by a localized “hypersensitive” cell death response

We observed that the inoculation of Eu. dicksonii 96 on the incompatible strain Ec. fasciculatus 007-04 resulted in the algal filaments becoming interspersed with dead cells (Figures 2A and 2B), recognizable by their loss of chlorophyll fluorescence, their blue color under UV light, and by their green color under differential interference contrast (DIC) or bright field. These cells can further be stained with the death marker Evans blue (Figure 2C), a stain widely used in plant biology to evidence heightened permeability of the plasmalemma, and by extension, cell death.²⁸ Spore counts showed that a much higher proportion of dead algal cells have an Eu. dicksonii spore attached to their surface compared to living algal cells (Figure 2D, α < 0.001). Since a mock-inoculated control does not accumulate dead cells, we conclude that the attachment of Eu. dicksonii spores to the surface of Ec. fasciculatus 007-04 (or possibly the subsequent infection attempt) triggers this cell death reaction. As this reaction restricts the development of the pathogen locally, we conclude that it phenotypically compares to a HR. We were able to generalize these results to all five incompatible interactions identified and therefore conclude that all incompatible algal strains investigated withstand infection by mounting a “hypersensitive” cell death reaction against Eu. dicksonii (Figures 2H and S1; Table 1).

Figure 2.

The hypersensitive response of Ectocarpus against Eu. dicksonii is mediated by programmed cell death

Exemplary data acquired on the incompatible interaction between Ectocarpus fasciculatus 007-04 and Eu. dicksonii 96, except (G) that correspond to Ec. siliculosus 022-10 and Eu. dicksonii 06. See information for other incompatible strain combinations. The white arrows point to Eu. dicksonii spores.

(A) Dead algal cells—recognizable under UV by the loss of chlorophyll autofluorescence—accumulate in a resistant Ectocarpus culture co-incubated for 30 days with Eu. dicksonii (inset: mock control). Eu. dicksonii spores are stained in green with CFSE.

(B) The dead algal cells shown in (A) appear greenish under DIC and typically carry an empty Eu. dicksonii spore at their surface. Filament regeneration is also visible 30 dai (asterisks, see Figure S15 for details).

(C) Evans blue staining of an Ectocarpus cell challenged by Eu. dicksonii (13 dai).

(D) Relationship between host cell death and Eu. dicksonii spore encystment (30 dai, one-tailed Z test, the asterisk highlights α < 0.001, 758 cells examined in total).

(E) β1-3 glucan deposition across the cell wall of an Ectocarpus cell reacting to Eu. dicksonii (aniline blue staining). The inset shows an overlay of the blue epifluorescence of aniline blue with the same field of view under DIC. An empty Eu. dicksonii spore is pointed at the surface of the blue-stained algal cell in the center.

(F) Labeling of an Ectocarpus cell challenged with Eu. dicksonii using a heterologous anti-metacaspase antibody (MCP, red channel). Algal nuclei and Eu. dicksonii structures appear in the blue (TOPRO-3) and green (aniline blue) channel, respectively (see Figure S2 for details).

(G) TUNEL labeling an Ectocarpus nucleus reacting to Eu. dicksonii (overlay of DAPI, TUNEL, and DIC (see Figure S3 for details).

(H and I) Accumulation of reactive oxygen species in Ectocarpus reacting to Eu. dicksonii. (H) Superoxide, 4 dai. (I) Hydrogen peroxide, 8 dai.

(J) TEM evidence of plastid degeneration (Cp; magnified in the upper inset) and swollen mitochondria (M; lower inset) in an Ectocarpus cell reacting to Eu. dicksonii.

(K) Magnification from (J) showing the obstruction of plasmodesmata (pl) by the deposition of new fibrillar cell wall layers (black arrows; see Figure S4 for a mock control).

Scale bars, A: 200 μm; B, C, E, H, and I: 10 μm; J: 4 μm; insets in J: 500 nm; K: 1 μm.

DIC, Differential Interference Contrast; dai, days after inoculation.

Hypersensitively reacting cells express defense and PCD markers

The induction of defense hallmarks or PCD markers was further investigated in incompatible interactions; in the light of the correlation identified previously, the cells of incompatible algal strains that carried an Eu. dicksonii spore at their surface were considered challenged by the pathogen and therefore assumed to react hypersensitively. Such cells deposited β1-3 glucans in their wall (Figure 2E) and were immunolabeled with a polyclonal metacaspase (MCP)-specific antibody previously used as a PCD marker in the haptophyte Emiliania huxleyii²⁹ (Figures 2F and S2). Note that whereas MCPs are present in the Ectocarpus genome, we were unable to confirm the specificity of the antibody with western blot in our algal cultures, probably due to the low overall expression of the protein(s) and the dilution of hypersensitively reacting cells in our algal cultures. Additionally, algal cells could be labeled in a TUNEL assay following inoculation with Eu. dicksonii, a histochemical assay widely used across eukaryotes to detect fragmented nuclear DNA²⁸; this observation is indicative of a DNA fragmentation that was clearly absent in mock-inoculated algal material (Figures 2G and S3). Note that due to the necessary sectioning of algal filaments in this protocol, it was not always possible to correlate TUNEL labeling of an algal nucleus with the attachment of a pathogen spore on the surface of the corresponding host cell. However, the TUNEL assay also revealed that cell-death-mediated resistance can occur at various stages during the infection attempt, because algal nuclei were seen degenerating either before or after the pathogen nucleus had penetrated in the host algal cell, and sometimes after it had divided in a multinucleated syncytium (Figures S3 and S5). Overall, all our observations are consistent with TUNEL-positive algal nuclei strictly correlating with an infection attempt of the cell by Eu. dicksonii. Finally, hypersensitively reacting algal cells also produced reactive oxygen species, namely superoxide and hydrogen peroxide (Figures 2H, 2I, and S6). Under TEM, they exhibited digested plastids and swollen mitochondria (Figure 2J), and tended to obstruct their plasmodesmata through the deposition of fibrillary cell wall layers (Figure 2K; compare with a mock control in Figure S4).

In compatible algal strains, a fraction of infected cells also mounts a hypersensitive response

Aside from incompatibility, the most frequent outcome of our host range study was in fact disease compatibility (Table 1). In compatible interactions, we soon realized that successfully infected host cells co-occur with dead algal cells devoid of visible intracellular pathogen structures under bright field microscopy, similar to the ones observed in incompatible interactions (Figures 3A and S7). Likewise, spore counts showed that such dead algal cells in compatible strains are disproportionately likely to carry Eu. dicksonii spores at their surface (Figure 3B). We conclude that in compatible interactions, only a fraction of host cells attacked by the pathogen is successfully infected by Eu. dicksonii; the remainder is capable of mounting a hypersensitive reaction and expresses the same inducible markers as observed in incompatible interactions, namely β-1,3-glucan deposition (Figure 3C), MCP (Figure 3D), DNA degradation (Figure 3E), and accumulation of superoxide and hydrogen peroxide (Figures 3F–3H). Additionally, we observed that hypersensitively reacting cells of compatible strains tend to specifically accumulate blue fluorescent compounds, the chemical nature of which we did not attempt to ascertain (Figure 3I).

Figure 3.

In compatible brown algal strains, resistant and susceptible host cells co-exist

(A) In most strain combinations tested, successful infections (arrowhead) coexist with hypersentively reacting algal cells (double arrowhead), leading to partial disease resistance.

(B) Proportion of dead (black bars) and healthy (gray bars) host cells from brown algae carrying Eu. dicksonii spores. Spore counts were carried out as in Figure 2D, using representative strains across the class Phaeophyceae available in our laboratory. In these partially resistant strain combinations dead algal cells and phenotypically healthy cells devoid of an intracellular Eu. dicksonii thallus were scored for the presence of a spore at their surface. ∗ indicates that the recorded proportions are significantly different with α < 0.001 (one tailed Z test); #: idem with α < 0.002. For host codes, see key resources table.

(C) Whole-cell β1-3 glucan deposition in a hypersensitively reacting host cell (aniline blue staining) of Ec. crouaniorum 06-29-7 vs. Eu. dicksonii 06 (partial resistance).

(D) Metacaspase expression in the compatible interaction between M. pyrifera Mau and Eu. dicksonii 96. The cell on the right expresses metacaspase indicative of undergoing cell death whereas in the host cell on the left a Eurychasma sporangium has developed (see Figure S2 for details).

(E) Three phenotypes in one filament of the compatible interaction between M. pyrifera Mau and Eu. dicksonii 05. The first cell on the left bears a mature dehiscent Eurychasma sporangium, the middle cell early cell death with a TUNEL-positive labeled host nucleus (arrow), and in the cell on the right Eurychasma has initiated its intracellular development (nuclei visible, arrowheads) but the host cell underwent cell death (TUNEL-positive algal nucleus [arrow]). Individual channels are shown in Figures S9D–S9F.

(F–H) ROS production in challenged brown algae. (F) Superoxide production in M. pyrifera Mau challenged by Eu. dicksonii 05. (G) Hydrogen peroxide production in M. pyrifera Mau against Eu. dicksonii 96 and (H) Ec. siliculosus 32m against Eu. dicksonii 96.

(I) Differential accumulation of blue fluorescent metabolites in a swollen hypersensitively reacting (arrowhead) and a susceptible (double arrowhead) Ectocarpus sp. 32m cell inoculated with Eu. dicksonii 96 (10 dai).

Arrow: Eu. dicksonii spore.

Completion of the pathogen development in host cells correlates with the late induction of defense and PCD markers

Blue fluorescent compounds were occasionally visible in cells bearing mature Eu. dicksonii sporangia (Figure S8A), suggesting that susceptibility results from the delayed onset of defenses. Likewise, focal callose deposition papillae were present even in successfully infected cells (Figures S8B–S8D) as was ROS production (Figure S8E; see also studies by Tsirigoti et al.³⁰ and Strittmatter er al.³¹). In some cases, TUNEL-positive algal nuclei could be observed in cells with dehiscent Eurychasma sporangia (Figures S8F and S9). Those reactions probably correspond to the onset of a local defense response that may contribute but is not always sufficient to arrest pathogen penetration.

Inducible host and pathogen autophagy is a second line of defense in compatible interactions

A study by Murúa et al.¹⁴ showed that in disease-susceptible gametophytes of the kelp M. pyrifera, two local, inducible, and autophagy-mediated responses contribute to the restriction of another intracellular oomycete, A. ectocarpii: firstly, pathogen thalli progressively become fully abortive over time, a phenomenon presumed to result from the hijacking of their autophagy machinery by the algal host; secondly, infected algal cells become able to digest and eliminate the pathogen via xenophagy. Both responses are tractable with TEM and quantifiable using acidotropic dyes such as monodansylcadaverine (MDC) and LysoTracker red. This inducible response is only observed in compatible interactions, as in incompatible interactions MDC labeling has not been detected (except of plurilocular sporangia in Ectocarpalean seaweeds).

Firstly, and as observed with A. ectocarpii, Eu. dicksonii thalli turned increasingly MDC-positive over an infection time course (Figures 4A and S10). The first MDC-positive Eu. dicksonii thalli were observed from 12 days after inoculation (dai) onwards (Figures 4B and S11) in M. pyrifera and Ec. siliculosus. The proportion of MDC-positive thalli in the Eu. dicksonii population increased significantly reaching values ranging from 40% to close to 80% at 35 dai (Figure S11, p < 0.05). A similar increase was also observed with LysoTracker over a 35-day time course, although a quantification was not attempted (Figures 4C and S12). We were able to correlate the onset of MDC and LysoTracker labeling in vivo with the appearance of in TEM of highly vacuolated, abortive Eu. dicksonii syncytia with no recognizable organelles (Figure 4D). Note that such abortive Eu. dicksonii thalli occur in dead host cells with degraded organelles.

Figure 4.

Abortive autophagy of Eurychasma and xenophagy in susceptible brown algal strains

(A) MDC+ (arrow) E. dicksonii 96 in Ec. siliculosus 32m. ∗: empty Eu. dicksonii sporangium thallus.

(B) Time course accumulation of autophagic thalli of Eu. dicksonii 96 population in its host Ec. siliculosus 32m, as scored with monodansylcadaverine (MDC). Individual time courses are shown for six replicates, the y axis represents the ratio of MDC-positive thalli over the total of pathogen thalli examined. Time points where MDC-positive Eu. dicksonii were not found are designated as red dots. Letters on every sampling day designate the statistically significant differences between time points (multiple comparisons, Friedman tests), where a < b and p < 0.05. MDC ratio time courses for Ectocarpus and Macrocystis against Eu. dicksonii are shown in Figure S11.

(C) M. pyrifera filament with LysoTracker-positive (right) Eu. dicksonii thalli in 35 dai. Arrow: Eurychasma thallus (DIC in inset). Scale bars, 30 μm.

(D) Degenerated Eu. dicksonii syncytium (Sy) inside an already infected M. pyrifera cell. Arrow: encysted Eu. dicksonii spore. Scale bars, 10 μm.

(E) MDC+ host cell with a developing Eu. dicksonii thallus inside (arrow). Scale bars, 20 μm.

(F) M. pyrifera cell infected by multiple unwalled Eu. dicksonii thalli (6-week-old culture, Eury 06). Small (e.g., dividing plastids) (double arrowheads) are visible all around the cell. Note also localized cell wall thickenings (black arrows). Sp, encysted Eu. dicksonii spore at the host cell surface; Sy, Eu. dicksonii syncytium; HN, host nucleus, closely associated to parasitic thalli and a very active Golgi apparatus; HCy, host cytosol. Scale bars, 5 μm. Inset: comparable M. pyrifera cell with multiple infections, as seen under DIC microscopy. White arrowheads: Eu. dicksonii syncytia. White arrow: encysted spore. Scale bars, 7 μm.

(G and H) Digestion of Eu. dicksonii thalli after incorporation into the host digestive vacuole. (G) From image (F), a magnification shows that the Eu. dicksonii plasma membrane (arrows) and vacuolar membranes (arrowheads) are locally digested (double arrowheads), leading to the loss of integrity of the thallus. Scale bars, 1 μm. (H) Advanced degradation of Eu. dicksonii thallus. Note that the host organelles look intact, but that plastids are dividing (double arrowhead). HM, host mitochondria; Cp, chloroplasts; HN, host nucleus; Sy, Eu. dicksonii syncytial debris. Scale bars, 2 μm.

Secondly, another population of infected host cells (i.e., containing intracellular Eu. dicksonii thalli) appeared over time, that were vacuolized and showed strong MDC labeling (Figures 4E and S13). In this case, the host cell appeared to survive the infection, as judged by the normal appearance of their organelles. TEM imaging was conducted on a 6-week-old M. pyrifera-Eu. dicksonii co-incubation, in order to allow for several cycles of re-infection and the full induction of the autophagic response; at that point, M. pyrifera cells were often infected with several unwalled Eu. dicksonii thalli (Figure 4F). Such host cells contained unwalled and vacuolated Eu. dicksonii thalli, some of them with recognizable well-shaped organelles such as nuclei and mitochondria (Figures 4F and S14A). Some Eu. dicksonii thalli also show vesicles being assimilated into vacuoles (Figure S14B), which we interpret as evidence of autophagy-mediated abortion. However, in other Eu. dicksonii thalli, some vacuoles and well-delimited host pathogen interfaces are lost, suggesting the pathogen is being digested by Macrocystis (Figure 4G). This observation coincides with some Eu. dicksonii thalli being labeled with MDC only in the interface with M. pyrifera and Ectocarpus vacuoles (Figure S13; Table S1). In a later stage of this M. pyrifera-mediated autophagy process, Eu. dicksonii thalli are barely recognizable whereas the central host vacuole is filled with debris (Figure 4H). In conclusion, our observations on this pathosystem suggest that autophagy is also induced both in M. pyrifera and Eu. dicksonii, consistent with a conservation of the elimination of the pathogen through xenophagy and the subversion of Eu. dicksonii’s autophagy’s machinery toward abortive development. After applying the autophagy inhibitor chloroquine, abortive autophagy of pathogen and xenophagy phenotypes are lost in Ectocarpus Ec32m but Eurychasma tended to accumulate debris in their vacuoles (Figure S15), suggestive of the ATG machinery is involved in these processes.

Regeneration may follow host cell death

In most, though not all, interactions examined, host cell death is followed by regeneration according to a pattern that depends on the algal host and the cell type considered (Table S1). For example, Figure S16 illustrates how, in the erect filaments of Ec. fasciculatus 007-04, the cells flanking a dead host cell first divide asymmetrically, producing rhizoid-like filaments that grow indefinitely toward each other. This process is fundamentally symmetrical although lack of growth synchronization between both converging rhizoids might lead to seemingly asymmetrical filament regeneration (Figure S15B).

Discussion

Hypersensitive cell death is a conserved, innate disease resistance mechanism of brown algae against Eurychasma dicksonii

Here, we identify incompatible interactions between three strains of Eu. dicksonii and four species of Ectocarpales. In all of them, disease resistance stems from the death of cells challenged by Eu. dicksonii (Figure 5A). In all strain combinations investigated, this cell death reaction is localized at the point of pathogen ingress, accompanied by the accumulation of reactive oxygen species, the deposition of blue fluorescent metabolites in the cell and of β1-3 glucans in the cell wall, and the localized degradation of nuclear DNA, a frequent hallmark of PCD.³² Ultrastructural changes also occur, such as swelling of mitochondria and the obturation of plasmodesmata, which are also associated to PCD and senescence in plants.³³ Though we acknowledge that the molecular mechanisms underpinning pathogen recognition and the host response might differ from plants, we thus refer to this cell-death-mediated resistance as a HR due to its phenotypical resemblance with the cell-death-mediated innate defense responses of land plants, diatoms, and red algae that are already named as such (e.g., Weinberger et al., 2000, and Canter and Jaworski, 1979; see last section of the introduction for details).

Figure 5.

Sequential contribution of the cell-death-mediated (hypersensitive-like) response and of autophagy-mediated, inducible defenses in the resistance to infection by Eurychasma dicksonii: a working model

(A) Innate defense reaction. The pathogen development leading to susceptibility (purple box on the left, thick vertical arrows). At each stage, the horizontal thin arrows highlight the possible induction of a hypersensitive response and defense markers. The interaction is incompatible (yellow box) when all Eu. dicksonii cells undergo cell death before the pathogen sporulates; compatibility arises when only a fraction of infected algal cells undergo cell death before sporulation occurs, so that macroscopically, innate resistance to Eu. dicksonii is a quantitative trait.

(B) Inducible defenses. In compatible interactions, newly released pathogen spores amplify the infection, which progressively induces autophagy in the pathogen and the host (enlarging black arrows). As both mechanisms arrest the pathogen’s development, increased disease resistance might ensue (enlarging gray arrows), compared to the innate level shown in (A).

Taking advantage of the broad host spectrum of Eu. dicksonii, we investigated compatible interactions across ten algal species spanning four orders of brown algae: the Ectocarpales and the Laminariales are closely related, whereas the Tilopteridales are more distant and the Discosporangiales is the earliest diverging order in the class Phaeophyceae.³⁴ Macroscopically, the balance between successful cell infection and HR results in a continuum between full resistance and full disease susceptibility.¹³ Across all ten species investigated, compatible interactions with Eu. dicksonii always result from some host cells being successfully overcome by the pathogen, while others resist infection through hypersensitive death before the intracellular development of the parasite is completed (Figure 5A). In other words, every algal species investigated was to some extent capable of resisting infection via an HR (Table 1), and the conservation of defense markers across the phylogenetic breadth of brown algae therefore strongly suggests that innate defense mechanisms against Eu. dicksonii are conserved in this lineage. Our findings strongly resonate with an early report on the diatom Asterionella formosa,²¹ implying that cell-death-mediated defense responses against pathogens might be more broadly conserved across the Stramenopiles.

Cell-death-mediated resistance against Eurychasma dicksonii as a quantitative trait: Implications for improvement of brown algal crops

The co-existence of infected and HR-undergoing host cells in compatible interactions strongly suggests that all algal strains and species under study have the potential to mount an HR response. In addition to suggesting that this program is genetically determined and conserved at least to some extent, our findings strongly imply that the proportion of resistant host cells against Eu. dicksonii correlates with the severity of infection observed macroscopically. This is similar to the continuum of partially resistant cultivars that are often observed in agronomy.³⁵ Likewise, our results suggest that brown algal HR is a quantitative trait, whose expression is linked with the timely recognition of the pathogen. Such recognition systems are virtually unknown in this phylum, and their characterization will ultimately be key to inform breeding programs in seaweeds. In the more immediate future, however, the new cell death and resistance-associated defense markers described here, combined to the quantitative, non-invasive fluorometry and nephelometric assays that we have developed earlier,³⁶ have immediate practical implications for crop improvement and breeding of the most economically valuable kelp species, and should enable the identification of quantitative trait loci associated to resistance.

A novel working model integrating innate cell-death and inducible autophagy-mediated defenses of brown algae against Eurychasma dicksonii

In an earlier report, we demonstrated that autophagy is an inducible defense response of some Ectocarpales and Laminariales against another intracellular oomycete pathogen, Anisolpidium ectocarpii.¹⁴ These autophagy-mediated defenses were observed in susceptible interactions and expressed both through the induction of an abortive program in the pathogen thallus and xenophagy, i.e., the digestion of the pathogen within the host’s vacuole. The final outcome was the elimination of the pathogen. Here, we show that both responses are also induced in response to infection by Eu. dicksonii. Thus, we propose a working model where, hypersensitive cell-death-mediated resistance at the point of initial pathogen ingress is a first barrier against infection, followed by the induction of autophagic responses in compatible strains (Figure 5B). PCD is concomitant with the accumulation of ROS and localized cell wall appositions. Thus, a local PCD program is expressed from the beginning of the infection, aiding to stop/delay Eurychasma dissemination by sacrificing the challenged host cell(s). In compatible interactions, it also occurs in infected host cells during different stages of the pathogen development. Cell death can be induced in the pathogen as well, interrupting the pathogen’s life cycle. The pathogen completes its life cycle and successfully sporulates when PCD fails to occur or only happens after the pathogen zoosporangium has matured. In such compatible interactions, additional barriers of inducible defenses are then induced, such as xenophagy or the inducibility of abortive autophagy of these pathogens.

The molecular mechanisms underpinning the brown algal HR and inducible autophagy-mediated defenses remain to be elucidated. Both in partially and fully resistant interactions, the HR is strictly correlated to the deposition of blue fluorescent metabolites; we hypothesize that these are phlorotannins, a structurally complex family of stress-inducible phenolic metabolites³⁷ with demonstrated antimicrobial properties in vitro.³⁸ Beta-1,3-glucan deposition, identified by aniline blue staining, was common in papillae from hypersensitive host cells. Remarkably, these markers are reminiscent of the localized phytoalexin deposition and cell wall reinforcement characteristic of the plant HR.³⁹ It is tantalizing to speculate that the link between PCD and innate immunity, secondary metabolism induction and cell wall reinforcement might be ancestral traits that predate the independent evolution of multicellularity in plants, animals, and heterokonts.^40,41 PCD-related genes characterized in other eukaryotes are essentially all present in the Ectocarpus genome, including a type III MCP and Bax1-like protein.⁴² Similar to plants, members of BCl2 family and of the NFκB transduction pathway are seemingly absent. The novel bioassays and markers described here, combined with the recent development of gene knock-out protocols on brown algae,⁴³ open unprecedented avenues to tackle these fundamental questions.

Limitations of the study

One experimental limitation relates to the use of the heterologous, polyclonal MCP antibody: we only ever observed staining in algal cells challenged by the pathogen, to the exception of any other cell in Eurychasma-inoculated cultures. However, we did not manage to detect the protein(s) with western blot, which might just be related to the low density of hypersensitively reacting cells in our experimental set-up. As only restricted amounts of this polyclonal antibody were available, it was not possible to perform additional controls to check its specificity, especially as there is, to the best of our knowledge, no established protocol to trigger PCD in a brown alga. Yet, we believe that it is worthwhile including our MCP observation here, especially to encourage further investigation in this direction.

More generally, how pathogen recognition takes place and the molecular mechanisms of the defense responses triggered remain undetermined. We note however, that the plant HR was first characterized cytologically in the first half of the 20th century, and the concept expanded later to diatoms and red algae without any knowledge of the receptors or molecular actors involved.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Claire M. M. Gachon (claire.gachon@mnhn.fr).

Materials availability

This study did not generate new unique reagents. The accession numbers for the algae and pathogens held by the Culture Collection for Algae and Protozoa (www.ccap.ac.uk) are available in the key resources table.

Data and code availability

• •Access to laboratory notebooks can be requested from the lead contact.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We gratefully acknowledge funding from the UK Natural Environment Research Council: grants NE/D521522/1 (F.C.K.); NE/F012705/1 (P.v.W., F.C.K., C.M.M.G., and S.W.), and NE/J00460X/1 (C.M.M.G.). C.M.M.G. also received support from the EU FP7 Marie Curie actions (MIEF-CT-2006-022837 and PERG03-GA-2008-230865). M.S. was supported by an FP7 ECOSUMMER Marie Curie PhD fellowship (MEST-CT-2005-20501). F.C.K., C.K., and A.T. were supported by the Total Foundation. C.K. and A.T. received awards from the FP7 program ASSEMBLE. P.M. was funded by Conicyt-ANID for his PhD studies at the University of Aberdeen (Becas Chile Nº 72130422) and Nucleo Milenio MASH (NCN2024_037). We gratefully acknowledge Kay Bidle (Rutgers University) for providing the E. huxleyii anti-metacaspase antibody, and Emmanuelle Evariste (SAMS) and Gillian Milne (Aberdeen Microscopy Facility) for technical assistance. Emmanuelle Evariste received an undergraduate summer studentship from the British Phycological Society.

Author contributions

Conceptualization, all co-authors; data curation, C.M.M.G., M.S., P.M., M.B., A.T., and S.W.; formal analysis, all co-authors; funding acquisition, project administration, resources, and supervision, C.M.M.G., C.K., P.v.W., and F.C.K.; investigation and methodology, C.M.M.G., D.G.M., M.S., P.M., M.B., A.T., and S.W.; writing – original draft, C.M.M.G., M.S., P.M., M.B., A.T., and S.W.; writing – review & editing, all co-authors.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Polyclonal antibody targeted against a metacaspase of Emiliania husleyi	Kay Bidle29	not available
Experimental models: Organisms/strains
Ectocarpus siliculosus	A. Peters, collected in 2002 in San Juan, Peru. Fully sequenced genome strain. Clonal haploid sporophyte derived parthenogenetically from a male gametophyte.	32m (CCAP 1310/4)
Ectocarpus siliculosus	D.G. Müller, collected Jul. 1973, in Kapisigdlit, Godhåbsfjorden, West Greenland. Original name Ec Gro G30. Clonal unialgal field isolate of unknown ploidy.	022-10 (CCAP 1310/214)
Ectocarpus fasciculatus	D.G. Müller, Roscoff, France. Original name Ec fas Ros 007-04, synonym: Ec fas Ros 93-31Z7 G12f. Clonal haploid sporophyte derived in the laboratory from a female gametophyte.	007-04 (CCAP 1310/13)
Ectocarpus crouaniorum	C. Gachon, 2006, Oban, Scotland. Original name Ec Oban 06-29-7. Monoeukaryotic clonal field isolate, original field host of Eu. dicksonii CCAP 4018/3	06-29-7 (CCAP 1310/300)
Pylaiella littoralis	D.G. Müller, Isla Diego Ramirez, Chile. Unialgal clonal strain.	IRg (CCAP 1330/3)
Pylaiella littoralis	D.G. Müller, collected in 1993 in Concarneau Harbour, France. Original name Pyl Ros 93-24-1. Epiphytic on drift Fucus vesiculosus. Unialgal clonal strain.	Ros
Pylaiella littoralis	D.G. Müller, collected Feb. 13, 1985. Puyuhuapi, Puente La Union, Chile XI Region. Orignal name Pyl CH 85-21. Plantlet collected in the intertidal zone, bearing unilocular sporangia. Unialgal clonal strain.	CH
Pylaiella littoralis	culture obtained from Dr. Rainer Schmid, Queen's Univ. Belfast Northern Ireland in May 1994. Origin: Ballyhenry Island, near Portaferry, Co. Down. Original name Pyl BHI 293. Unialgal clonal strain.	BHI
Acinetospora crinita	D.G. Müller Aug 23, 1988. Kaikoura, New Zealand South Island, Original name Acinet NZ 88-14-1 Point Kean, intertidal, sea grass bed, on mud. Unialgal clonal strain.	name Acinet NZ 88-14-1
Laminaria digitata	D.G. Müller. Clonal haploid female gametophyte.	Ros (CCAP 1321/4)
Macrocystis pyrifera	D.G. Müller, Feb 13, 1985. Puyuhuapi, Puente La Union, Chile XI Region. Epilithic, upper sublittoral. Original name : Mac pyr Puy 85-20-3 f. Clonal haploid female gametophyte	PU
Macrocystis pyrifera	D.G. Müller, Fuerte Bulnes, Chile. Original name SAm 70-2 f. Clonal haploid female gametophyte.	FB (CCAP 1323/3)
Macrocystis pyrifera	D.G. Müller, Maullin estuary, Chile. Original name Mau 99-27f. Clonal haploid female gametophyte.	Mau (CCAP 1323/1)
Tilopteris mertensii	Collected by P. Kornmann and D. G. Müller, May 22, 1984, Helgoland (Germany). Drift between main Island and Duene. Unialgal clonal isolate.	Hel1
Choristocarpus tenellus	D. G. Müller, collected with Scuba diver, 18 m deep, near the Marine Station Villefranche-sur-mer, France (Mediterranean Sea) Sept. 1. 1980. Axenic clonal isolate.	6 ax
Discosporangium mesarthrocarpum	D. G. Müller, collected Oct 15, 2004, Hydra Island (Greece). Entangled in a trawling net, creeping on Sargassum linearis. Unialgal clonal isolate.	04-14
Eurychasma dicksonii	F.C. Küpper and D.G. Müller, collected in 1996, at Aith Voe, Bressay, Shetland, UK. Found infecting epiphytic Pylaiella littoralis on Ascophyllum nodosum. Isolate derived from a unique Eu. dicksonii thallus, ploidy unknown.	05 (CCAP 4018/1)
Eurychasma dicksonii	D.G. Müller, Collected in 2005, at Le Caro, Rade de Brest, Brittany, France. Found infecting E.siliculosus (CCAP 1310/299). Isolate derived from a unique Eu. dicksonii thallus, ploidy unknown.	96 (CCAP 4018/2)
Eurychasma dicksonii	C. Gachon, collected in 2006 at Dunstaffnage Bay, Oban, Scotland. Found infecting E. crouaniorum (CCAP 1310/300). Isolate derived from a unique Eu. dicksonii thallus, ploidy unknown. Original name Eu. dicksonii Oban 06-29-7	06 (CCAP 4018/3)
Software and algorithms
Axiovision 4.5 and 4.7	Zeiss Ltd	Not applicable

Experimental model and study participant details

Algal and pathogen strains are described in the key resources table. Cultures were grown grown at 15°C (20°C for the Mediterranean strain D. mesarthrocarpum) in half-strength Provasoli medium, under a 12:12 photoperiod and a light intensity of 2-10 μmol photon m^-2 s^-1. Algal cultures (except C. tenellus) contained naturally-occurring epiphytic bacteria. Since Eu. dicksonii is an obligate biotroph, inoculations were performed by co-incubating the target alga with an infected donor as previously described.¹³ A mock-inoculation was systematically prepared as a negative control. Depending on the strain combination used, results were scored between 5-8 dai for incompatible interactions and 10-21 dai for compatible interactions. For spore counts, Eu. dicksonii structures were stained with Congo red between 15 and 23 dai, depending on the rapidity of symptom development for each algal host. In each sample, an average of 150 algal cells was examined and categorized as either dead or healthy. Whether or not an Eu. dicksonii spore was encysted at the surface of those cells was recorded. In partially compatible interactions, visibly infected algal cells that contained an intracellular Eu. dicksonii thallus were disregarded.

Method details

Staining techniques

Live Eu. dicksonii structures cultures were stained for 15 min in Provasoli medium containing 0.01 mg mL^-1 of the cellulosic stain Congo red (Sigma-Aldrich), and briefly washed with sterile seawater. For epifluorescence observation, Eu. dicksonii spores were stained with the vital stain carboxy-fluorescein succinimidyl-diester (CFSE), at a 2 μL / mL final concentration for 5-15 min and washed with sterile seawater. Cell death was monitored with Evans blue staining.⁴⁴ Cultures were stained with a 0.04% Evans blue solution for 15 min, followed by one washing step in sterile seawater. Aniline blue is a fluochrome commonly used to detect β1-3 glucans (callose, Tsirigoti et al., 2015). Algal cultures were incubated for 5 min at 95°C in KOH (1 mol L^-1). Following 3 washes with sterile seawater, they were incubated in 0.025% aniline blue dissolved in KOH/EtOH. The samples were mounted on a slide in antifading agent (Slowfade, Invitrogen).

The histochemical detection of the reactive oxygen species hydrogen peroxide (H₂O₂) and superoxide (O₂^-) was performed using 3,3'-Diaminobenzidine (DAB) and Nitroblue Tetrazolium Chloride (NBT) respectively, as previously described.³¹ Monodansylcadaverin (MDC) and Lysotracker red® staining were carried out after Murua et al.,¹⁴ Bright field observations were made under differential interference contrast (DIC) on a Zeiss Axioscope 2 upright microscope fitted with x10, x20, x40, and x100 objectives. Epifluorescence observations were made using DAPI (G 365 nm, FT 395 nm, LP 420 nm) and FITC (BP 450-490 nm, FT 510 nm, LP 515 nm) filter sets. Pictures were recorded with a digital AxioCam camera coupled to the software Axiovision 4.5 or 4.7 (Zeiss, Germany).

Immunofluorescence

Infected thalli were fixed in 4% paraformaldehyde for 20 min on ice, and washed three times in in phosphate buffered saline (PBS) containing 0.05% NaN₃. Paraffin-embedded samples were cut with a Leica Ultramicrotome (RM 2125RT) in 5 μm-thick slices. After rehydration, the samples were incubated with a polyclonal anti-Emiliania metacaspase antibody (1:250) in PBS containing 10 % BSA for 1 h at 37°C {Bidle, 2007 #3125}. The samples were washed 3× for 5 min with PBS before incubation with the secondary antibody (Alexa Fluor 555 Goat anti rabbit; Invitrogen; 5 μg/mL, 1 h, 37°C). The slides were washed 2× for 5 min with PBS and incubated for 10 min with TO-Pro3 iodine (1:1000 Invitrogen) and aniline blue (0.05% in PBS). An additional wash (2× for 5 min in PBS) was performed before mounting in Vectashield (Vectorlabs).

TUNEL assay

Infected thalli were fixed, embedded and cut in 5 μm slices in as described in the previous section. The embedding material was subsequently removed by incubation at 60°C for 25 min, followed by 2 washes in 100% xylene for 15 min each, gradual washes in 100%, 70% and 50% ethanol for 5 min each. Subsequently the algal samples were treated with 100 μL of proteinase K in PBS at a final concentration of 20 μg/mL for 10 min at room temperature. Then samples were washed in PBS for 10 min, followed by one wash in 50% methanol for 10 min, gradual washes in 100%, 70% and 50% ethanol for 5 min each and one final wash in PBS. Subsequently, dUTP labelling was performed (Promega DeadEnd™ Fluorometic TUNEL system). The samples were incubated in 100 μL equilibration buffer for 10 min, which was subsequently removed from the samples. Then 60 μL of incubation buffer containing nucleotide mix and rTdT enzyme were added and incubated at 37 for 1 hour. Afterwards the samples were washed twice in 2× SSC buffer for 15 min and once in PBS for 5 min. Nuclei were stained with 0.1% v/v DAPI in PBS for 10 min, followed by two washes in PBS for 5 and 10 min respectively. Finally, the samples were mounted in Slowfade and observed via fluorescence microscopy. Negative controls omitting the dUTP labelling and positive controls treated with DNase were also included.

Transmission electron microscopy

Eurychasma-challenged infected cultures were fixed in 2.5% glutaraldehyde, 0.5% caffeine, 3% NaCl, 0.1% CaCl₂ 0.1 M SodiumCacodylate prepared with sterile filtered seawater(pH 7.4) for a couple of days, at room temperature, washed (3x 15 min) in 0.1 M Sodium Cacodylate, 3% NaCl, 0.5% CaCl₂. The material was postfixed in 1% OsO₄ in the same buffer for 2 hours at 4^oC and washed. We applied Uranyl acetate (2% in distilled water) fort 1 hr, dehydrated the biomass with ascending acetones series (15 min each) and infiltrated with Spurr’s resin. Polymerized blocks were sectioned (90 nm) with an ultramicrotome (Leica UC6), placed on copper grids and contrast-stained with lead citrate (3%). Sections were imaged using a JEM- 1400 Plus (JEOL) TEM with an AMT UltraVue camera, available at the Aberdeen microscopy facility.⁴⁵ Autophagy inhibitor treatment was employed with 200 μM chloroquine (ATG15 inhibitor) in Macrocystis-challenged Eurychasma 05 (6-weeks old), for 10 days incubation.¹⁴

Quantification and statistical analysis

The statistical details are indicated in the figure legends, including the tests used and significance thresholds. Proportions were compared using one-tailed Z-test ,with the significance expressed as α, and the significance threshold set at 0.001. For time-courses, multiple comparisons were made using a one-tailed Friedman test, and p value were considered significant when <0.05.

Published: July 24, 2025