Colletotrichum higginsianum effectors exhibit cell to cell hypermobility in plant tissues and modulate intercellular connectivity amongst a variety of cellular processes

Multicellular organisms exchange information and resources between cells to co-ordinate growth and responses. In plants, plasmodesmata establish cytoplasmic continuity between cells to allow for communication and resource exchange across the cell wall. Some plant pathogens use plasmodesmata as a pathway for both molecular and physical invasion. However, the benefits of molecular invasion (cell-to-cell movement of pathogen effectors) are poorly understood. To begin to investigate this and identify which effectors are cell-to-cell mobile, we performed a live imaging-based screen and identified 15 cell-to-cell mobile effectors of the fungal pathogen Colletotrichum higginsianum. Of these, 6 are “hypermobile”, showing cell-to-cell mobility greater than expected for a protein of its size. We further identified 3 effectors that can indirectly modify plasmodesmal aperture. Transcriptional profiling of plants expressing hypermobile effectors implicate them in a variety of processes including senescence, glucosinolate production, cell wall integrity, growth and iron metabolism. However, not all effectors had an independent effect on virulence. This suggests a wide range of benefits to infection gained by the mobility of C. higginsianum effectors that likely interact in a complex way during infection.


184
Identification of candidate cell-to-cell mobile Colletotrichum higginsianum effectors 185 To establish a candidate list of putative C. higginsianum effectors (hereafter referred to as 186 effectors) exhibiting cell-to-cell mobility, we mined published transcriptome data covering different infection stages (O'Connell et al., 2012;Dallery et al., 2017). Many effectors are 188 secreted from pathogens into host plant tissues (Lo Presti et al., 2015), and therefore we limited 189 our candidate cell-to-cell mobile effectors as those that encode conventionally secreted proteins is alive) and thus defined candidate cell-to-cell mobile effectors as those that have enhanced 196 expression in PA and BP phases relative to the other 2 stages (i.e. PA/VA, BP/VA, PA/NP, and 197 BP/NP >2). We further limited candidates to those for whichPA reads > 50 and BP reads > 20.

198
These criteria produced a list of 46 candidate effectors.

199
Plant proteins that are known to be cell-to-cell mobile are typically soluble within the 200 cytoplasm or the nucleus (Kim et al., 2002;Gallagher et al., 2004;Gallagher and Benfey, 2009;201 Chen et al., 2013) and we made the assumption that C. higginsianum cell-to-cell mobile 202 proteins would have similar properties. Thus, to further refine our list of candidate cell-to-cell 203 mobile effectors, we cloned these 46 candidate effectors as fusions to GFP and screened for 204 nucleocytoplasmic and nuclear subcellular localisations by transient transformation of N. 205 benthamiana. Of these 46 effector-GFP fusions, 25 showed nucleocytoplasmic localisation but 206 none showed a specific nuclear localisation.

207
Live cell screening for cell-to-cell mobility 208 To assay the cell-to-cell mobility of candidate effectors, we performed a live cell imaging-209 based screen using transient transformation of single epidermal cells in N. benthamiana. For 210 this we used Golden Gate modular cloning (Engler et al., 2008) to assemble effector-GFP 211 fusions and a cell transformation marker in a single vector as a dual expression cassette vector 212 (Fig. 1a). For a cell transformation marker, we used dTomato fused to a nuclear localisation 213 sequence (NLS-dTomato), reasoning that the dimerization properties of dTomato would make 214 a protein complex too large to move from cell to cell.

215
To confirm the utility of NLS-dTomato as a cell transformation marker, we generated 216 constructs that express GFP or 2xGFP with NLS-dTomato (pICH4723.GFP.NLS-dTomato and 217 pICH4723.2xGFP.NLS-dTomato respectively). Agrobacterium infiltration of N. benthamiana 218 leaves demonstrated that both fluorophores were expressed in the transformed cell. While GFP was frequently seen to move freely from the transformed cell, both NLS-dTomato and 2xGFP were mostly retained within the transformed cell ( Fig. 1 and S1).

264
To determine if hypermobility is associated with a general increase in mobility that would of mRFP relative to the Col-0 control (Fig. 3b). Thus, this data indicates that Ch132 also 288 modifies plasmodesmal function like ChEC127 and ChEC8, and that ChEC123 mediates its 289 translocation by a possible active mechanism that does not involve plasmodesmal modification.  and iron starvation responses and transport (Fig. 5b). Constitutive expression of ChEC132 also 315 induced down regulation of genes associated with iron transport and responses, in addition to cell wall modification, growth, syncytium formation (5 EXPANSIN genes also represented in the cell wall loosening GO term), and lipid metabolism (Fig. 5a).  Table S2).

333
In addition to identifying that cell-to-cell mobility is possible for a range of effectors, our 334 data suggests that effectors from C. higginsianum can move through plasmodesmata by 335 different mechanisms. Firstly, we identified a subset of effectors that move 'passively' from 336 cell to cell, i.e. they move through plasmodesmata at a rate expected for soluble molecules of 337 the same size (Fig. 2). Like the endogenous plant transcription factor LEAFY (Wu et al., , these molecules can be considered to have no mechanism for active translocation and 339 simply move from cell to cell as soluble, freely diffusive molecules. Secondly, we identified 340 hypermobile effectors that modify plasmodesmal function such as ChEC127 and ChEC132.

341
These effectors trigger plasmodesmata opening to a degree such that the effector itself ( Fig.   342 2), as well as other soluble molecules (as observed for mRFP; Fig. 3b), can move faster and 343 further to neighbouring cells. Thirdly, we identified one hypermobile effector, ChEC123, that 344 has no general effect on plasmodesmal function ( Fig. 2; Fig. 3). This effector could therefore  pathogen virulence, ChEC127 induced differential expression of only 13 genes (Table S4), 366 suggesting the mechanism by which it promotes virulence does not involve significant 367 perturbation of gene expression. By contrast, ChEC123 and ChEC132, which did not 368 independently promote virulence, did induce significant changes in gene expression (Table S3; 369 Table S5). ChEC123 downregulated genes associated with glucosinolate production (glycosyl 370 compound catabolism) and iron starvation and transport, indicating it may regulate defence and 371 nutritional processes (Fig. 5b). The same effector up-regulated genes associated with leaf 372 senescence which might contribute to the transition of C. higginsianum to the necrotrophic 373 lifestyle (Fig. 5c). ChEC132 down regulated genes associated with iron metabolism (Fig. 5a) 374 but most significantly perturbed processes associated with plant cell wall modification and 375 growth. This raises the possibility that ChEC132 perturbs growth processes, possibly to limit    Bonferroni corrected confidence interval of the mean (p < 1x10 -5 , red ribbon) was calculated.

592
The point density shows the number of replicates at that value.           The standard curve is a quasi-Poisson generalised linear model with a log link function and the Bonferroni corrected confidence interval of the mean (p < 1x10 -5 ) (red ribbon). Effectors were determined to be significantly mobile (purple squares) by an exact Poisson test indicating the rate of movement is significantly different to the standard curve (p < 1x10 -5 ). Data are means ± standard error (n >30, p < 1 x 10 -5 )  (a) Mobility of NLS-dTomato plotted against the relative mobility (Mob r ) of a co-expressed effector in N. benthamiana leaf epidermal cells. The standard curve represents NLS-dTomato mobility in the presence of GFP variants of different sizes with Bonferroni corrected confidence interval of the mean (p < 1x10 -5 red ribbon). NLS-dTomato movement was analysed by an exact Poisson test, identifying that ChEC8 and ChEC127 increase NLS-dTomato mobility (b) Mobility of mRFP (number of cells) in Arabidopsis leaf tissue assayed by microprojectile bombardment assays. Independent transgenic lines expressing ChEC127-GFP lines and ChEC132-GFP showed increased mobility of mRFP relative to Col-0. Boxes signifies the upper and lower quartiles, and the whiskers represent the minimum and maximum within 1.5 × interquartile range. The number of bombardment sites (n) counted is ≥ 92. Data was analysed by bootstrap analyses and asterisks indicate statistical significance compared with Col-0 plants (**p < 0.01 and *p < 0.05) Detached leaves from 4-5-week-old Arabidopsis were inoculated with C. higginisianum spores and lesion areas were measured 6 dpi (n > 60) and bootstrap analysis of the lesion area means identified that lesions are larger in ChEC127 expressing plants. CirGO visualisation of GO Terms enriched amongst (a) genes down-regulated by ChEC132, (b) genes down-regulated by ChEC123, and (c) genes upregulated by ChEC123. Slice size represents the proportion of DEGs attributed to this GO Term. The inner ring slices represent 'parent' GO terms and the labelled outer ring slices represent 'child' GO terms.