Use of ecologically-and evolutionary relevant transcriptomic data to infer functions of fungal pathogen gene orthologues important for limiting fungal stresses caused by interacting host plants and bacteria

278 words Body without references abstract and title pages: 6348 words. Total 8222 21 Figs and 2 tables with 88 XY plots. 42 references Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 31 December 2020 doi:10.20944/preprints202012.0773.v1 © 2020 by the author(s). Distributed under a Creative Commons CC BY license. 2 ABSTRACT (278 words) We identified key genes needed for maintenance and growth and homed in on genes where there could be a competition between maintenance requirements (stress) and growth requirements. Such processes are synthesis of arginine, synthesis of DNA-bases, nitric oxide synthesis needing arginine, autophagy, DNA synthesis and DNA repair. Using procedures previously developed for the use of sets of downloaded transcriptomic data to test hypotheses concerning at what time under the course of infection of plants genes are expressed for the two pathogens Fusarium graminearum and Magnaporthe oryzae, we constructed a simplified regulatory network for these genes for both organisms. Our analysis shows that the transcription effort (cost) to maintain the fungal cells (maintenance) are high before infection and in early infection. During the following biotrophic phase maintenance cost drops for later in the transition to the necrotrophic phase increase dramatically. Finally, in the necrotrophic phase, maintenance is lower again despite the high growth rate that can also cause stress. The expressions of all identified genes behaved almost similar for both fungi except the DNA repair genes PARP/PARG that was not responding or absent in the mainly clonal M. oryzae which might indicate this species is more subject to evolution by point mutations than F. graminearum where sexual reproduction is frequent. The potential consequences of these different roles for PARP/PARG in the development and the accelerated breakage of host species resistance in a Red Queen dynamics scenario is discussed. Our analysis demonstrates the possibility to use large transcriptome datasets and co-regulations between key genes to test hypotheses and discusses the advantages with this technique as complement to molecular techniques employing knockouts and over-expression of target genes to suggest gene roles. Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 31 December 2020 doi:10.20944/preprints202012.0773.v1


INTRODUCTION 55
Growth and maintenance are vital concepts both in cell biology and ecology. Maintenance is all the 56 activities needed to maintain the cell's integrity, including repairing proteins and DNA without making more 57 biomass [1]. Growth is simply the growing of biomass and cells' growth, including all proteins needed for 58 making new cells and maintenance, including the necessary new copies of the genomes [2]. However, the 59 maintenance concept has been questioned since it contains many different processes [3]. 60 Wheat head blight caused by F. graminearum Schwabe (teleomorph stage: Gibberella zeae (Schwein.) 61 Petch) often results in significant crop losses in grains like wheat and barley [4]. Magnaporthe oryzae B.C. 62 Couch (teleomorph Pyricularia oryzae Cavara) cause rice blast resulting in yield and economic losses worldwide 63 [5]. Both fungi are studied by many researchers worldwide, and they are both considered model organisms 64 [6,7]. One interesting difference between them is that F. graminearum often reproduces sexually [8] while M. 65 oryzae is mainly clonal [9]. Genetically, these two Ascomycete pathogens are relatively closely related and 66 differ from yeast and Penicillium/aspergillus species. Both fungi belong to the class Sordariomycetes but in 67 different orders. F. graminearum belongs to Hypocreales and M. oryzae to Magnaporthales. Most genes and 68 gene expression patterns are mirrored in the two species [10]. Both pathogens infect as biotrophs and switch 69 to necrotrophy at a later stage Hours Post Infection (HPI). They are exposed to environmental stresses at the 70 plant surfaces, including possible biotrophic stresses from other organisms. They enter the plant and 71 establishes biotrophic growth inside the plant. The plant defences are low at this time. At about mid-time (HPI), 72 the fungi become detected by the plant's innate immune system that starts attacking the intruders with radical 73 oxygen species (ROS). In response to this the pathogens switch to necrotrophy, killing the host cells and in the 74 case of F. graminearum producing the toxic secondary metabolites deoxynivalenol (DON) [4]. At the end of the 75 necrotrophic stage both fungi switch from biomass growth to conidia production emptying the vegetative 76 mycelium of biomass to form conidia that can spread to other plants and infect them. In a previous study we 77 found that the expression of the key autophagy gene ATG8 increases with HPI in both fungi and can be used 78 as indicator for HPI in downloaded expression data from a large number of experiments [10]. In the same study 79 we identified the His2b gene as an indicator of de novo DNA synthesis and growth since free histones not 80 bound to DNA are cytotoxic [10,11]. Maintenance expression of a specific gene is defined as the relative growth 81 rate normalized transcript expression of the gene, or in other words, gene expression normalized for DNA 82 synthesis [10]. 83 We have previously studied conserved genes involved in fungal maintenance and growth. As the 84 primary gene regulated during autophagy ATG8 [12], the DNA repair gene PARP [13,14], and recently we have 85 worked with genes involved in the synthesis of DNA bases [15] and the amino acid arginine [16]. Arginine was 86 shown to be used together with oxygen to produce nitric oxide (NO), a ROS produced in fungal innate immunity 87 [17]. Plants trigger NO-production during the transition between biotrophy to necrotrophy, and when the 88 fungus is exposed to bacterial MAMPs (microbial-associated molecular patterns) [17]. Together, these genes 89 fit into a conceptual model for how these conserved genes necessary for growth and maintenance are likely 90 to be differentially expressed during different stages of a plant's plant pathogen colonization. During growth-91 dominated stages, the purine synthesis genes are mainly used to make new DNA, while arginine synthesis is 92 primarily needed to make new proteins. ATG8 activity is also crucial for growth since growing fast causes a 93 need to recycle misfolded proteins, protein aggregates, and storage lipid droplets through autophagy [10,12,18] 94 (Fig. 1A).

99
B. When the fungus is exposed to plant-induced stresses, the stress-related network marked red should be most active. There will be 100 a focus on this network if gene expression is normalized for growth using the His2b gene transcriptional expression [

106
This hypothesis that the stress weighted network (Fig. 1B) should be most active just before 107 penetration and especially in the transition between biotrophy to necrotrophy when plant defences are 108 activated was tested. In contrast, during biotrophic growth and later stages of necrotrophy, the growth-related 109 network (Fig. 1A)  production, and the MAMPs responses should mainly be reflected as an increased response of the stress 114 weighted network (Fig. 1B). The stress-weighted network should increase in expression with the expression of 115 the cytochrome p450 gene (CYP(NO,ERG)) that is the gene mainly responsible for the intrinsic NO production 116 with accompanying ROS stress [17]. 117 The data supported the hypothesis, and also, it was found that the PARP gene necessary for DNA repair 118 is expressed very differently in M. oryzae and F. graminearum. An orthologue for PARG necessary for de-119 PARylation of the PARP activity is absent in M. oryzae. It is suggested to interpret the found difference that 120 this reflects the need for the mainly clonal M. oryzae [8,9] to generate variation through mutations to 121 overcome host resistance changes without sexual recombination. Finally, the potential benefits of 122 transcriptomic analyses for suggesting the relative importance of specific gene expressions and roles of genes 123 under relevant natural conditions are discussed 124

MATERIALS AND METHODS 125
The procedures are briefly outlined in the Results and Discussion section and mainly comprise plotting 126 of transcript expression data against transcript expression data (RNAseq or Affymetrix microarray data). All 127 data used for this paper are secondary data and have been described in previous articles and are publicly 128 available (Table 1). Candidate orthologous genes were identified through protein BLAST at NCBI (Table 1). 129  Fig. 2A-E). All five genes show similar temporal patterns of growth-normalized gene expression 151 ( Fig. 2F-J), supporting the presented hypothesis ( Fig. 1). The genes appear essential for maintenance to handle 152 the cellular stresses at low HPIs (LowATG8 expression) and are then upregulated. After that, expression of the 153 genes decreases during biotrophy, followed by a substantial increase during the biotrophy-necrotrophy 154 transition. Finally, all genes' expression decreases again during necrotrophy, tending to a final rise at very late 155 necrotrophy. 156 The shapes of these growth-normalized expression profiles are, in principle, W-shaped.

164
For M. oryzae, as for F. graminearum, an increased transcription was observed for most of the five genes 165 ( Fig. 3A-E). Growth-normalized FgPrp was strongly upregulated, especially at the transition between biotrophy 166 to necrotrophy, while this was not found for MoPrp ( Fig. 3F-J). The lack of regulation in M. oryzae could indicate 167 that very little ROS stress is experienced by the fungus in the transition between biotrophy to necrotrophy. 168 Alternatively, MoPrp is not functionally regulated in response to DNA damages caused by oxidative stress since 169 MoPrp can have lost both function and regulation. 170

176
Oxidative stresses due to hydrogen peroxide or nitric oxide 177 The relative expression for all catalase (CAT) orthologues in both fungi was investigated to investigate 178 ROS stress. Catalase is needed to counteract intrinsically, and plant-made H2O2 nitric oxide dioxygenase (NOD) 179 orthologues are necessary to balance intrinsically and plant-made NO. Five decent candidate orthologues to 180 yeast (Saccharomyces cerevisiae) CatA, genes FgCAT1-5 were found for F. graminearum ( Table 2). In F. 181 graminearum, there are two nitric oxide dioxygenase genes [17], and two orthologues to these were found in 182 M. oryzae ( Table 2). 183

F. graminearum catalase gene responses 184
All five catalase orthologues are activated in the biotrophy to necrotrophy transition (Fig. 4) Table 2). These were also only one of the 202 catalase genes with BLAST similarities to yeast CatA. Of these, only MoCAT1 responds strongly by an 203 upregulation in the transition between biotrophy to necrotrophy (Fig. 6A) and appears to stay high also 204 growth-normalized in the whole necrotrophic stage at high MoATG8 expression (Fig. 6A-iii).

213
However, none of the two catalase orthologues is strongly positively correlated with the MoPrp gene 214 (Fig. 6.B are sharply upregulated in the transition between biotrophy to necrotrophy ( Fig. 7-Ai,ii), most likely because 224 of plant generated NO instead of intrinsically generated NO that seem to dominate at low expression levels of 225 ATG8 (HPI). This pattern is even more pronounced when normalizing for growth (Fig. 7Aiii,iv). The likely 226 response to plant-generated NO in the transition between biotrophy to necrotrophy is made even more 227 probable if, instead, the NOD expression is normalized for the main protein involved in intrinsic fungal NO 228 formation. Now it can be seen that there seems to be a balance between FgNODs and FgCYP(NO,ERG) before 229 the biotrophy-necrotrophy transition (Fig, 7B). Thus, plant NO stress most likely dominates inside the plant, 230 235 while intrinsically produced NO probably dominates at low ATG8 levels before infecting the fungus and 236 biotrophy starts. The ratio of the expression of the two NOD genes per FgCYP(NO,ERG) (FgNOD1 or 237 2/FgCYP(NO,ERG)) as an indicator for intrinsically produced NO versus FgATG8 was plotted to see if intrinsically 238 dominated NO is more likely at low FgATG8 levels (HPI), and it is (Fig. 8A). It was also plotted against intrinsic 239 NO generation indicated by FgCYP(NO,ERG) expression (Fig. 8B). The results indicate that after the biotrophy-240 necrotrophy transition intrinsically produced, NO is negatively correlated with NO defences suggesting that 241 these NO defences are most likely against plant-generated NO. 242

248
NO is highly mutagenic, and the expression of both NODs that indicate needed ROS defences due to 249 NO is strongly correlated with PARP expression, suggesting that more DNA repair is needed at the high NO 250 levels likely caused by the plant defences ( Fig. 9A-B).

257
In M. oryzae, there is only one good BLAST hit for a NOD orthologue, MoNOD ( transition between biotrophy to necrotrophy (Fig. 10Ai). In this case, this is probably also because of plant-260 generated NO instead of intrinsically generated NO but might also be aided by intrinsically formed NO in the 261 necrotrophic stage (at high MoATG8 levels). This pattern is even more pronounced when normalizing for 262 growth (Fig. 10 Aii) where it can be seen that the MoNOD is sharply upregulated in the biotrophy-necrotrophy 263 transition. 264 MoNOD is highly expressed (Fig, 10Aii) at the same time MoNOD/MoCYP(NO,ERG) ratio is low (Fig.  265 10Aiii) when growth adjusted, indicating that NO likely comes from the plant. At the late stages of infection, it 266 is also clear that NO from the plant probably plays a larger role since this expression ratio increases (Fig. 10Aiii). 267 The idea that plant NO stress likely dominates inside the plant while fungal intrinsically produced NO 268 probably dominates at low MoATG8 levels (low HPIs before entering biotrophy) was tested further. The NOD 269 gene expression ratio to MoCYP(NO,ERG) as an indicator for intrinsically produced NO versus MoATG8 was 270 plotted to see if intrinsically dominated NO is more likely at low ATG8 levels and it is (Fig. 10Aiv). To further 271 confirm that this ratio is negatively correlated at high levels, it was also plotted against MoCYP(NO,ERG) ( Fig.  272  10Bi). During plant infection, gene expression of the gene for intrinsically produced NO seems to be negatively 273 correlated with the gene necessary for NO defences (Fig. 10Aiv, Bi), similar to the case for F. graminearum (Fig.  274  8B). 275 F. graminearum (Fig. 9A-B), but no such pattern is visible (Fig. 10C). This lack of regulation supports that in M. 285 oryzae, increased DNA repair by MoPrp is not activated by the plant produced ROS as H2O2 and NO even if the 286 fungus is stressed by these plant produced ROS. 287 Since this difference between FgPrp and MoPrp transcriptional activation was found, it was 288 investigated if there can be differences in the "parylation toolbox" for the two fungi. Part of the PARP signalling 289 pathway is the enzyme poly(ADP-ribosyl) glycohydrolase (PARG), the de-PARylation counterpart to PARP. 290 PARG enzymes have been described in Fusarium oxysporum (FoPARG) [26], and we found an orthologue in F. 291 graminearum ( Table 2). The F. graminearum protein FgPrg has a high similarity and is identical to the FoPARG 292 around the active site. Many orthologues PARGs can be found in fungi, and PARGs appear to be well conserved. 293 However, there were no PARG orthologues to be found in Magnaporthe sp. To test if the PARG orthologue 294 seems to be active in F. graminearum, FgPrg expression, and the FgPrp expression in the in planta data, was 295 plotted together to see if they correlate and they are expressed in a 1/1 ratio at all stages of infection ( Fig.  296  11A-B). For M. oryzae, the lack of a strong correlation of MoPrg with the ROS indicating genes (catalase and 297  NOD) (Fig. 6B and 10C) the ROS indicating genes (catalase and NOD) ( Fig. 6B and 10C)  As can be expected, the genes responsible for producing arginine needed to produce NO are positively 316 regulated with FgCYP(NO,ERG), purine synthesis FgCPA1 FgPrp FgPrg needed for DNA repair (Fig. 12). This 317 regulation supports the notion that intrinsic NO production causes single-nucleotide mutations that need 318 repair. Although the PARG gene FgPrg is upregulated, it does not entirely mirror the FgPrp gene found in the 319 in planta data (Fig. 11). This lack of mirroring can be due to the short time nature of the experiments (1-4h) 320 since protein parylation by PARP is a speedy process

339
The two NODs necessary to regulate NO concentrations specifically are upregulated with increased 340 expression of the NO generating FgCYP(NO,ERG), especially at the higher levels (Fig. 14Ai,ii). NOD1 is localized 341 in the cytoplasm and the nucleus, while NOD2 is not in the nucleus [17]. The FgNOD1/CYP(NO,ERG) ratio can 342 be expected to be negatively correlated with CYP(NO,ERG) to reach high levels of bacterial detection signalling 343 when exposed to MAMPs. Thus it is expected that a negative correlation between FgNOD1/CYP(NO,ERG) ratio 344 and Fg(CYP(NO,ERG) will be found, and that is indeed the case (Fig. 14Bi,ii). There is also a negative correlation 345 between FgNOD2/CYP(NO,ERG), but it seems less tightly co-regulated. over most of the Fg(CYP(NO,ERG) expression range (Fig.14i). 351 genes are expected to be negatively correlated, and they are (Fig. 15).

Potential relative roles of autophagy and apoptosis during the infection stages (HPI) in both fungi 368
In Metarhizium robertsii, belonging to the same order as F. graminearum (Hypocreales), the expression 369 of MrBI-1 has been identified as linked to apoptosis [31]. Apoptosis is needed to empty hyphae (autolysis) for 370 the use of resources in other hyphae (reallocation of resources) or production of conidia [13]. There is only 371 one orthologue to this gene in F. graminearum, FgBI1 (Table 2.). When investigating F. graminearum growth 372 measured as FgHis2b expression versus HPI as measured by FgATG8 expression (Fig. 16Ai)

386
The expression of FgBI1 increases with FgATG8 (HPI) as expected (Fig. 16Bi) since the reallocation of 387 resources from non-productive hyphae to productive hyphae are keys in mycelium development [32,33]. 388 Consequently, growth-normalized FgBI1 expression vs HPI (ATG8) also looks like the ATG8 curve ( Fig. 16Aii and  389 Bii). Apoptosis seems to play a slightly decreasing role compared to autophagy with increasing HPI (ATG8) since 390 the ratio FgBI1/FgATG8 decrease with HPI (ATG8) (Fig. 16iii). 391 Figure 16 and previous figures show that five stages for wheat infection by F. graminearum can be 392 suggested (Table 3). 393 394 The situation for M. oryzae is similar but also different. The growth rate measured as His2b expression 397 does not increase substantially until the transition to necrotrophy and the necrotrophic stage (Fig. 17A). At the 398 switch to necrotrophy and in the necrotrophic stage, autophagy's relative importance to growth is higher, 399 indicating stress and/or reallocation of nutrients through autophagy (Fig. 17A-C). On the other hand, apoptotic 400 emptying of hyphae (autolysis) seems to be much more important at an earlier stage in M. oryzae than in F. 401 graminearum. The emptying of vegetative hyphae can be indicative of considerable simultaneous both growth 402 and sporulation triggered at the transition from biotrophy to necrotrophy as also apoptosis MoBI1 expression 403 increases and stays high at the same level of MoATG8 expression (HPI) growth-normalized (Fig. 17Bii) and 404 ATG8 normalized (Fig. 17Biii). However, for this fungus, there also seems to be a slight decrease before a final 405 spurt in apoptosis (aiding conidiation) at the very high levels of MoATG8 expression (HPI) (Fig. 17Bii-iii).  to non-existent "happy" non-stressed growth period in the necrotrophic stage. 426 427

Conclusion 428
Testing the hypothesis: The analysis supports the hypothesis presented in the introduction and discussed in 429 the Results and Discussion (above). To summarize this, a conceptual model is presented (Fig. 18A). In this 430 Figure 18, it can be seen that the genes responding to stresses and that are needed for maintenance are 431 regulated in a W fashion with HPI (LOG2 ATG8 expression) in F. graminearum ( Fig. 18B and C). For M. oryzae, 432 the stress is similar in the transition between biotrophy to necrotrophy, but the stages before and after are 433 generally not characterized by increased stress except for catalases indicating plant defences induced during 434 necrotrophy (Fig. 18A). The shape of the response profile is not a W but more like a "wizard hat" (Fig. 18B and  435 C). 436  consequences for their respective co-evolution with their hosts: The only one of the genes identified to be 445 important in the network ( Fig. 1 and 2 increased resistance in host……..and so on and so on (Fig. 19). It could be expected that such an automatic host 469 resistance caused increased mutation rate as indicated for M. oryzae is more beneficial to the fungus if fungal 470 host range is narrow and infection cycles are short with ample production of infectious conidia. 471 472 Figure 19. Stress caused by the host causes sexual recombination or increased mutation rate of the fungal pathogens. A non-functioning 473 PARP (PARP/PARG) can potentially increase the mutation rate to benefit a clonally reproducing pathogen. Such overexpression is especially a problem if the gene product is cytotoxic if produced in surplus, like histones 486 [11]. Analyzing correlations between gene expression of genes belonging to physiologically relevant connected 487 processes in transcriptome datasets from natural conditions (not standard lab media) like in the present study 488 can potentially overcome some of these limitations and should be interrogated more frequently by researchers. 489 This insight also calls for developing tools for conditionally up or downregulation of genes and graded 490 regulations. It also calls for in-vitro studies under environmentally more relevant nutrient availabilities and 491 composition than routinely used. Most problematic for understanding the ecological relevant roles for genes 492 is perhaps that no fungi are growing in isolation in nature. They are always surrounded by their characteristic 493 microflora of bacteria [41] and other microorganisms. Lack of these interactions is also why transcriptomic 494 studies of pathogens during natural infection of plants can be better trusted for inferring gene functions than 495 "normal lab media". For the organism in this study's focus, there is a lack of transcriptomic studies for F. 496 graminearum on debris at different ages since incorporating the fungus into the soil on and in debris is part of 497 its life cycle. For this fungus, there is also a lack of transcriptomic studies from the rhizosphere of seedlings at 498 different times since a root tip passed since coming close to an inoculum might trigger fungal rhizosphere 499 activities [42,43]. Such studies will be tricky since the amount of RNA will be tiny, but they are today possible 500 [44]. For M. oryzae, it is evident in the downloaded transcriptomic data [10] that the amount of RNA has not 501 been large since many genes lack expression data under many conditions. Thus, similar techniques for using a 502 small amount of RNA might solve future research problems.