Spatially distinct phytohormone responses of individual Arabidopsis thaliana 1 root cells to infection and colonization by Fusarium oxysporum

31 Jasmonic acid (JA), ethylene (ET) and salicylic acid (SA) are the three major phytohormones 32 coordinating a plant’s defense response to pathogenic attack. While JA and ET are assumed to 33 primarily control the defense against necrotrophic pathogens, SA-induced defense responses target 34 mainly biotrophic microbes, and can include drastic measures such as programmed cell death as 35 part of the plant’s hypersensitive response (HR). Fusarium oxysporum is a hemibiotrophic fungal 36 pathogen of several plant species, including many important food crops, and the model plant 37 species Arabidopsis thaliana . Colonization of the plant’s root vascular tissue by the fungus 38 eventually results in wilting and plant death. A general role for JA, ET and SA in combating 39 infection and colonization of the plant by F. oxysporum has been demonstrated, but their distinct 40 roles and modes of action have so far not been described. Here, using high resolution microscopy 41 with fluorescent marker lines of A. thaliana roots infected with F. oxysporum we show that SA 42 acts spatially separate from JA, in a distinct set of root cells immediately neighboring the fungal 43 colonization site. There, SA induces HR to stop the spread of colonization. JA acts in a different, 44 but also clearly defined set of cells, slightly removed from the colonization site, where it initiates 45 a defense response to actively resist the invader. ET is activated in a stretch of cells that covers 46 both, the cells with activated SA and JA signaling, and may be involved in creating these two 47 distinct zones. These results show how the three phytohormones act together, but spatially and 48 functionally separate from each other, to fight this hemibiotrophic pathogen. Such a high- 49 resolution analysis to resolve the plant’s immune response to pathogenic infection on an individual 50 cell level and in intact tissue has so far been lacking.

inconclusiveness in relation to known pathways, such as the phytohormones, may lie in the tissue 116 and timing used in the experiments. In contrast to other plants, such as tomato, where F. oxysporum 117 can infect the plant via several openings simultaneously, the fungus generally appears to infect A. 118 thaliana solely via the root tip, thereby limiting the number of infection sites and thus infected 119 tissue available to be subjected to transcriptomic analyses (Czymmek et al., 2007;Wang et al., 120 2022a; Martínez-Soto et al., 2022). Thus, by using either whole seedlings or the whole root system 121 to extract RNA, the transcriptional changes in these few infected cells may be diluted too much by 122 the other uninfected cells. Furthermore, as the plant also senses the fungal hyphae in its rhizosphere 123 via their MAMPs, a majority of root cells will launch a general MAMP-response, which further 124 dilutes the response of the cells undergoing infection. 125 For our work, we have set up a microscopy-based live-imaging approach monitoring 126 transcriptional activity of different pathways in planta, over time and on an individual cell level. 127 We have previously published our pGG-PIP GreenGate entry vector collection with regulatory 128 sequences for plant immunity genes, including reporters for the different phytohormone pathways 129 (Calabria et al., 2022). Using this approach, we detected local responses of genes involved in SA, 130 JA and ET biosynthesis, signaling and regulation, targeted to specific tissues and to a very limited 131 number of root cells upon infection with Fo5176. Interestingly, we found that SA acts spatially 132 separate from JA in a distinct set of root cells in direct contact with the fungus, where it appears 133 to induce HR. JA in turn also acts in a clearly defined set of cells which are slightly removed from 134 the colonization site and SA-induced HR zone, where it appears to induce the plant's active defense 135 against the pathogen.

137
Fo5176 colonization of the root leads to cell death of the infected tissue 138 To monitor the infection and colonization of A. thaliana roots by Fo5176, we cultivated 10-day 139 old seedlings in square petri dishes, added fungal spores and then imaged hyphal growth, infection 140 and colonization over the next 14 days. We used a transgenic Fo5176 line expressing a cytoplasmic 141 tdTomato fluorophore to visualize the fungus (Calabria et al., 2022). One of the first things 142 noticeable when imaging infected plants was that the tissue surrounding the colonization site 143 appeared to undergo immediate cell death. When cells die, the plasma membrane and vacuole collapse, leaving behind the cell wall material. As this material scatters light, the dead tissue 145 appears as bright white in the image (Fig. 1A). In fact, when the vasculature of a root tip was 146 colonized by Fo5176, the entire tip underwent cell death (Fig. 1A, B). By imaging a plant line 147 expressing a fluorescent marker for the plant plasma membrane (PM) (mNeonGreen(mNG)-148 LTI6b), this became even more apparent, as the fluorescence signal from the PM outlined all cells 149 of the plant but disappeared completely from all cells around the colonized tissue (Fig. 1C, D). As 150 fungal colonization progressed through the vasculature as a 'colonization front', a 'cell death front' 151 was also observed that moved in line with it. In contrast though, the cell death front spanned the Fluorescent fungal hyphae (magenta) can be seen in the tip and the vasculature. The colonization front (magenta bars to the left and right of the root) aligns with the cell death front. C, D: Overlay of bright field and fluorescence image (C) and fluorescence image only (D) of a root expressing the PM-marker mNG-LTI6b (yellow) and colonized by Fo5176 (magenta). Fluorescent hyphae (magenta) can be seen in the tip and the colonized vasculature. The fluorescence coming from the PM-marker (yellow) is lost at the colonization front (magenta bars). entire circumference of the root, including the outer tissues, epidermis, cortex, and endodermis, 153 which are not colonized by the fungus. Because cell death was not limited to the colonized tissue, 154 we reasoned that cell death may be the result of a hypersensitive response (HR) initiated by the 155 plant, rather than caused directly by Fo5176.

156
Salicylic acid signaling is activated in the hypersensitive response zone 157 As HR is often activated via an SA-and PATHOGENESIS-RELATED PROTEIN 1 (PR1)-158 dependent pathway, we next investigated if SA biosynthesis and signaling via the antimicrobial 159 protein PR1 are indeed activated in the cells bordering the colonized tissue. We utilized reporters 160 for the key SA biosynthesis enzyme ENHANCED DISEASE SUSCEPTIBILITY 16 (EDS16; also 161 named SALICYLIC ACID DEFICIENT 2 (SID2)) and PR1. In an uninfected control, EDS16 was   1-CARBOXYLIC ACID SYNTHASE 2 (ACS2), a key enzyme in the ET biosynthesis pathway. 179 We found that AOS expression was mostly absent in the uninfected control roots, with very weak 180 expression limited to sites of natural stress occurrence, such as lateral root emergence sites (Fig.  S2A, B). Similarly, PDF1.2 was not expressed in the root; however, it was strongly expressed in 182 above ground tissue. ERF1 was expressed in all differentiated tissues, but was absent from the root 183 tip, including the MZ and EZ (Fig. S2C-F). This contrasting expression pattern between AOS and 184 ERF1 most likely represents activation via the ET-, rather than the JA-pathway, since ET also acts 185 in several developmental pathways, while JA-signaling is mostly restricted to stress conditions.   (Fig. 6A, B). Following colonization by Fo5176, we observed strong upregulation in in this zone, expression still appeared to be stronger in the cells surrounding the central cylinder, 250 and therefore not directly overlapping WRKY11 expression (Fig. 6C, D). Thus, WRKY70 may 251 indeed be responsible for promoting SA biosynthesis and signaling in the HR zone, while the 252 mutual antagonism of WRKY11 and WRKY70 could contribute to establishing these two distinct 253 response zones.

255
With the work described here we aimed at clarifying some of the inconclusive data concerning the 256 roles of JA, ET and SA in regulating the plant's defense against the fungal pathogen Fo5176. 257 Consistent with our own microscopic observations, and previously described by Cyzmmek et al. reasoned that previous bulk transcriptomic analyses using whole seedlings, or the entire root 260 system were insufficient to detect highly specific events occurring within the limited amount of 261 infected tissue. This hypothesis was strengthened by our initial observation, that the infected tissue 262 undergoes immediate cell death (Fig. 1), thereby limiting the amount of infected tissue available 263 for transcriptomic analyses even further. By employing a microscopy-based approach to live-264 image the plant's response to infection in real time and at individual cell resolution, we were then 265 indeed able to show that JA, ET and SA all contribute to the plant's defense against Fo5176, albeit 266 in clearly distinct groups of cells (Fig. 7).

267
An SA signaling pathway establishes the HR zone next to the colonization site 268 We first noticed that colonization of the vasculature was accompanied by local cell death of the 269 entire infected root tip. As the fungus continued to move through the root, colonizing more cells, 270 it formed a clearly visible 'colonization front' in the vasculature. As this colonization front moved 271 forward, a 'cell death front' moved in line with the colonization front. The only difference between 272 these two fronts being that the cell death front ran along all tissues of the root, from epidermis to 273 epidermis, and not just in the vasculature, as is seen with the colonization front (Fig. 7). We 274 hypothesized that this cell death front was the result of the plant's own HR, rather than 275 necrotrophic behavior by the fungus, and therefore imaged our reporters for SA biosynthesis and 276 signaling, since SA is a known inducer of HR (Pitsili et al., 2020). Indeed, both SA biosynthesis 277 and signaling, were upregulated precisely in a small group of cells directly in contact with the 278 already colonized and dead tissue. Our interpretation of this observation is that SA induces HR in 279 the cells that will otherwise be colonized next by the pathogen, thereby trying to prevent the fungus 280 from spreading further. In our model we designated this group of cells the 'HR-zone' (Fig. 7). JA, 281 on the other hand, was upregulated in a distinct group of vascular cells beginning slightly (4-7 282 cells) removed from the HR-zone. In this region, that we named the 'defense response zone', we 283 think the plant launches an active defense response to combat the invader (Fig. 7). Such defense 284 can include the activation of NADPH oxidases to produce ROS in the apoplast that may damage 285 the fungus, an acidification of the apoplast via proton pumps, or the production of antimicrobial compounds (Wang et al., 2022a). ET activity bridges the boundary between these two zones, may 287 facilitate boundary formation between the two zones, and also contributes to the defense response 288 zone together with JA.

289
WRKY70 acts via SA to establish the HR zone, and WRKY11 acts via JA to establish a 290 defense response zone'. 291 We then asked how this distinct spatial separation of the two hormonal responses can be achieved 292 and investigated the role of a WRKY transcription factor module, consisting of WRKY11 and 293 WRKY70 that have previously been described as controlling JA and SA signaling respectively 294 (Journot-Catalino et al., 2006). In this regard, we not only found that WRKY11, a positive regulator 295 of JA, was indeed upregulated in the defense response zone, while WRKY70, a positive regulator 296 of SA, was upregulated in HR-zone, but we also observed that these two transcription factors 297 altered their expression patterns. In the uninfected state, WRKY11 was expressed in all cells and 298 tissues of the root, in accordance with a proposed role as a negative regulator of basal defense gene 299 expression (Journot-Catalino et al., 2006). This is consistent with the hypothesis that under regular 300 growth conditions, defense genes need to be downregulated/repressed, to allow for normal 301

Figure 7: Model showing the distinct activity zones of the phytohormones and transcription factors
Fo5176 infects an A. thaliana root tip via the meristem in the tip, where it colonizes the vasculature (purple cells = colonized zone). The tissue around this zone undergoes cell death (dark grey cells). As the colonization front (purple line) progresses in the vasculature, so does a cell death front (black line) across all tissues. In an apparent attempt to prevent the fungus from infecting more tissue, the plant triggers a cell death response in a group of cells immediately adjacent to the colonization front (yellow cells), the HR zone. HR is activated by WRKY70 via SA. Slightly removed from the HR and colonized zones, the plant launches an active defense response to combat the pathogen in the defense response zone (orange cells). This response is dependent on WRKY11 and JA/ET-biosynthesis and signaling. WRKY11/JA and WRKY70/SA are mutually antagonistic, thereby establishing the two spatially separate zones of distinct action. development of the plant, as part of the defense-development trade-off. Following colonization of 302 the plant by Fo5176, WRKY11 was robustly upregulated in the defense response zone, likely 303 activating JA biosynthesis and signaling. In contrast, WRKY11 expression was no longer observed 304 in all other tissues. This would enable the release of basal defense genes from repression and prime 305 such tissue for a possible pathogen attack. Furthermore, WRKY11 would also contribute to the 306 spatial separation of the defense response and HR zones, since WRKY11 and JA jointly repress 307 SA via WRKY70. Consequently, WRKY70 expression is excluded from the defense response zone 308 following colonization by Fo5176 and most likely excludes WRKY11 and JA from the HR zone.

309
Thus, the mutual exclusion of WRKY11 and WRKY70 appears to establish the separate defense 310 response and HR zones (Fig. 7).

324
Sensing of a pathogen results in the upregulation of SA, ET and ROS. All three are then involved 325 in priming the plant for impeding attack, but ROS also acts to repress SA/ET-induced HR.

326
However, ROS can only block the SA/ET->HR pathway as long as SA and ET concentrations are 327 at moderate levels. Acute pathogenic attack that results in cell damage elevates local SA and ET 328 concentrations to a point at which ROS can no longer counteract it (Fig. 8). In this situation of 329 high local concentrations of SA, ET and ROS, this co-incidence results in HR. In the case of 330 vascular colonization by Fo5176, this situation would affect the cells of the HR-zone that are in 331 immediate contact with the already colonized, and thus damaged cells. Slightly removed from the colonization front, in the cells that we call the boundary between HR and defense response zone, 333 local SA and ET concentrations are lower, so ROS can still block the induction of HR to prevent 334 RCD (Fig. 8). Local signaling by residual SA, JA, ET and ROS also primes these cells for 335 impeding attack. Finally, in the defense response zone, SA is excluded, and now JA and ET jointly 336 induce the active defense response of the plant, with ROS contributing to this defense system (Fig.   337 8). This model is purely hypothetical at this point, but in our previous work we could show that a

345
In this pathway, PEP1 is produced in response to pathogen-induced cell damage. This peptide is 346 bound by the PEPR2, which triggers a phosphorylation chain from PEPR2's intracellular kinase In the cells of the HR zone, PEP1, via PEPR2 and BIK1, activates RBOHD to produce ROS. In combination with an over-accumulation of SA and ET via WRKY70-and cell damage-signaling, these factors induce HR. In the boundary, SA and ET accumulate at lower concentrations because the cell damage signaling is lacking and WRKY11 signaling partially suppresses WRKY70 signaling. Thus, ROS can now suppress HR and thus RCD. In the cells of the defense response zone, WRKY11 activates JA biosynthesis, while ET is still active, potentially via PEP1 signaling. JA and ET jointly activate active defense via ERF1. domain to the cytoplasmic BIK1 and the transmembrane RBOHD, which eventually produces a 348 ROS-burst (Fig. 8) (Liu et al., 2013;Tintor et al., 2013;Holmes et al., 2018). 350 Whilst the inclusion of the WRKY-module and the dual role of RBOHD in contributing to defense 351 and controlling RCD led to an elegant hypothesis of how the separate zones can be established, 352 further work is needed to validate these observations. First, it will be important to image the 353 phytohormone markers in wrky11 and wrky70 mutant plants, to definitively establish that they act plates were placed into a growth cabinet and grown for 10 to 14 days. After this, seedlings that 415 survived selection were transferred to fresh plates without antibiotics and grown until they were 416 too big for the plates, at which point they were transferred to soil until maturation of siliques.

417
Fungal growth and transformation 418 The Fo5176 line expressing the cytoplasmic tdTomato (tdT) was described in our previous paper,