DNA methylation plays a role on in vitro culture induced loss of virulence in Botrytis

Pathogenic fungi can lose virulence after protracted periods of culture but little is known of the mechanisms regulating this. Here we assess whether DNA methylation could play a role in this phenomenon by the methylome analysis of virulent and reduced virulence derivative cultures of Botrytis cinerea, and identify the genes/genomic regions affected by these epigenetic modifications. Virulence declined during the eight months culture and recovered after one fungal generation on A. thaliana. Methylation-sensitive amplified polymorphisms show that epi/genetic variation followed virulence changes during culture. Whole genome sequencing showed no significant genetic changes during culture. Conversely, bisulfite sequencing showed significant changes both on global and local methylation patterns. We suggest that virulence is a non-essential plastic character regulated by DNA methylation during protracted in vitro culture. We propose DNA methylation as a regulator of the high virulence/low virulence transition in B. cinerea and as a potential mechanism to control pathogenicity.

in B. cinerea. We anticipate that these results will provide new targets for the control of B. 107 cinerea caused disease. It also provides novel insights into the control of components that 108 dictate saprotrophic and parasitic capability and proposes DNA methylation as a mechanism 109 to control these processes. CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which was this version posted June 17, 2016. ; https://doi.org/10.1101/059477 doi: bioRxiv preprint All isolates obtained from all culture time points produced lesions on A. thaliana to various 114 extents ( Figure 1A). As B. cinerea was successively cultured disease scores decreased over 115 the eight month period (T0-T8) ( Figure 1B). This loss of virulence with time in culture was 116 significant from T3 onwards (T-Test P< 0.05) ( Figure 1B). The disease scores for the T8P 117 challenge did not differ significantly from those for those at T0 culture times suggesting 118 recovered virulence following single passage through a plant ( Figure 1B). Conversely, T8 119 virulence scores were significantly different from those obtained from T0 to T5 and T8P (T-120 Test P< 0.05) ( Figure 1B). Fungal DNA content within the infected areas of the leaf was 121 measured by quantitative PCR to confirm that in planta fungal development increased 122 between T8 and T8P generation as described previously (24). Infected leaves with T0 and 123 T8P cultures did not show significant differences in fungal DNA content ( Figure 1C).

124
However both showed significantly higher levels of fungal DNA (T-Test P< 0.05) when 125 compared to those infected using T8 cultures ( Figure 1C).  Furthermore, a partial recovery of the epigenetic profile was observed on samples cultured 137 for 8 months after one fungal generation on the host plant (T8P) ( Figure 2C). Calculated 138 . CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made contained one or more variants between T1 and T8 (See Table 1 and Supplementary Table 1 176 for a comprehensive list of genes with variants). not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made       Table 3) (28). As a whole, these genes 225 showed the same methylation pattern as described above for the 82 randomly selected genes 226 with an increase in methylation upstream of the TSS ( Figure 5A). More interestingly, these 227 genes showed higher levels of methylation after 8 months in culture than at T1 and aT8P.

228
This observed increase in global methylation on T8 samples seems to be due to an increase  DMRs showing levels of methylation significantly higher than the average ( Figure 6B).

246
When studied individually, changes in methylation within DMRs between samples presented 247 two main pattern types ( was not significantly different between T1 and T8P samples but they were higher or lower 259 . CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made   were duplicated, i.e., present in more than one category). The virulence sensu lato genes 282 included: 12 appressorium-associated genes (26), 17 virulence sensu stricto genes (27) 283 . CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which was this version posted June 17, 2016. ; https://doi.org/10.1101/059477 doi: bioRxiv preprint and 1155 plant cell wall disassembly genes (CAZyme genes) (28). Of these, 478 genes 284 (30.3%) overlapped with one or more detected DMRs (See Table 1 and Supplementary Table   285 6 for a comprehensive list of genes overlapping with DMRs).    To determine the significance of the observed changes in DNA methylation between culture 379 time points we performed an analysis of regional differential methylation identifying 2822 not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which was this version posted June 17, 2016. ; https://doi.org/10.1101/059477 doi: bioRxiv preprint keeping genes did not overlap with any DMRs while genes associated to apoptosis and 406 conidiation showed the higher percentage of overlapping with DMRs (40.0 And 37.5% 407 respectively). Remarkably, both biological processes have been previously shown to be 408 affected by DNA methylation (19,36,37). Interestingly, the 98% of these DMRs overlapped,       Applied Biosystems, USA). Samples were heat-denatured at 95°C for 3-5 min and snap-    . CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which was this version posted June 17, 2016. ; https://doi.org/10.1101/059477 doi: bioRxiv preprint one sample and the other two samples (FDR<0.01)) were determined using a three sample 911 Kruskal-Wallis test. Methylation levels were analysed by sliding window analysis using 912 swDMR. DMRs were determined for all cytosines and for three methylation contexts. DMRs 913 were grouped according to their changing patterns into recovery (T1=T8P) and non-recovery 914 (T1< > T8P). Two subgroups where found for recovery (T1=T8P < T8 (Type 0) and T1=T8P 915 > T8 (Type 2)) and non-recovery (T1>T8=T8P (Type 1a) and T1=T8<T8P (Type 1b)).

916
Percentage of the total DMRs for each pattern type/sequence context is shown in parenthesis. CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which was this version posted June 17, 2016. ; https://doi.org/10.1101/059477 doi: bioRxiv preprint    . CC-BY-NC-ND 4.0 International license available under a not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which was this version posted June 17, 2016. ; https://doi.org/10.1101/059477 doi: bioRxiv preprint