Phytophthora methylomes modulated by expanded 6mA methyltransferases are associated with adaptive genome regions

Filamentous plant pathogen genomes often display a bipartite architecture with gene sparse, repeat-rich compartments serving as a cradle for adaptive evolution. However, the extent to which this “two-speed” genome architecture is associated with genome-wide epigenetic modifications is unknown. Here, we show that the oomycete plant pathogens Phytophthora infestans and Phytophthora sojae possess functional adenine N6- methylation (6mA) methyltransferases that modulate patterns of 6mA marks across the genome. In contrast, 5-methylcytosine (5mC) could not be detected in the two Phytophthora species. Methylated DNA IP Sequencing (MeDIP-seq) of each species revealed that 6mA is depleted around the transcriptional starting sites (TSS) and is associated with low expressed genes, particularly transposable elements. Remarkably, genes occupying the gene-sparse regions have higher levels of 6mA compared to the remainder of both genomes, possibly implicating the methylome in adaptive evolution of Phytophthora. Among three putative adenine methyltransferases, DAMT1 and DAMT3 displayed robust enzymatic activities. Surprisingly, single knockouts of each of the 6mA methyltransferases in P. sojae significantly reduced in vivo 6mA levels, indicating that the three enzymes are not fully redundant. MeDIP-seq of the damt3 mutant revealed uneven patterns of 6mA methylation across genes, suggesting that PsDAMT3 may have a preference for gene body methylation after the TSS. Our findings provide evidence that 6mA modification is an epigenetic mark of Phytophthora genomes and that complex patterns of 6mA methylation by the expanded 6mA methyltransferases may be associated with adaptive evolution in these important plant pathogens.

DNA methylation, one of the fundamental epigenetic marks, participates in many biological 43 processes in both eukaryotes and prokaryotes 1-3 . The most studied form of DNA 44 methylation is 5-methylcytosine (5mC), which is a prevalent DNA modification in mammals 45 and plants 4 . The 5mC modification plays a role in many processes, such as transposon 46 silencing, regulation of gene expression and epigenetic memory maintenance 5 . The 47 amount of 5mC present in DNA varies across organisms and is barely detectable or absent 48 in many species, such as the nematode (Caenorhabditis elegans), the fruit fly (Drosophila 49 melanogaster) and brewers yeast (Saccharomyces cerevisiae) 6 . Comparatively, the N6-50 methyladenine (6mA) modification is extensively distributed in prokaryotic genomes. A 51 prominent function of 6mA is in discriminating between host DNA and invading DNA, thus 52 contributing to prokaryote immunity against phages and other invading genetic elements 7 . 53 Besides, 6mA is also involved in DNA replication, repair, virulence, and gene regulation 8-54 11 . 55 In contrast to prokaryotes, the occurrence and biological functions of 6mA methylation 56 in eukaryotic organisms remain largely uncharacterized. There is increasing evidence that 57 6mA is present in eukaryotes, including mammals, nematodes, algae, fruit flies, frogs, and 58 fungi 12-16 . Genome-wide 6mA distribution patterns can be identified by several robust 59 methods such as methylated DNA immunoprecipitation sequencing (MeDIP-seq) 14,15 , 60 6mA-sensitive restriction enzyme digestion coupled with high-throughput sequencing 17 , 61 and single molecule real time sequencing (SMRT sequencing) 12,13 . The 6mA pattern 62 appears to be dynamic during development; for instance, the early embryonic stage of 63 Drosophila has relatively higher 6mA levels compared to later stages 14,18 . Furthermore, the 64 genomic localization of 6mA significantly differs among organisms [13][14][15] . The 6mA 65 modification is widely and evenly distributed in the Caenorhabditis elegans genome. By 66 contrast, 6mA is enriched around transcription start sites (TSS) in early-diverging fungi and 67 Chlamydomonas, and is enriched in transpoable elements in Drosophila. The localization 68 patterns associate with 6mA biological functions. For example, in Chlamydomonas and 69 fungi, 6mA is enriched around the TSSs of actively expressed genes, suggesting that 6mA 70 may be an active mark for gene expression 12,15 , while 6mA appears to suppress 71 transcription on the X chromosome in mouse embryonic stem cell 16 72 Like many other epigenetic marks, 6mA can be reversibly modulated by enzymes such 73 as methyltransferase and demethylase 13,14 . It is known that DAM and M.MunI are classical 74 bacterial 6mA methyltransferases 19 . In eukaryotic cells, enzymes from the MT-A70 protein 75 family that evolved from M.MunI 20 , are considered 6mA methyltransferases. 76 Overexpression of the MT-A70 homolog DAMT-1 from C. elegans in insect cells elevated 77 the 6mA level, whereas knockdown of damt-1 resulted in a decrease in the amount of 6mA, 78 suggesting that DAMT-1 is a potential 6mA methyltransferase in nematodes 13 . However, The Oomycetes are a group of eukaryotic organisms that include a variety of 87 pathogens that infect plants and animals 24 . A notorious example is Phytophthora 88 infestans, the causal agent of potato late blight disease which sparked the Irish famine, 89 resulting in starvation and migration of millions of people in the 1840s 25 . An additional 90 example is Phytophthora sojae, a soybean root pathogen that currently threatens global 91 soybean production. These two species are model organisms among oomycetes 26 . The 92 genomes of these Phytophthora display a bipartite architecture, with gene-sparse and 93 repeat-rich regions (GSR) and gene-dense regions (GDR) 25 . The GSR compartments 94 are associated with accelerated gene evolution, serving as a cradle for adaptive 95 evolution 27-29 . However, the biological roles of DNA modifications and their associations 96 with adaptive genome evolution remain unknown. In this study, we demonstrate that 97 6mA, rather than 5mC, is the major DNA methylation in these two Phytophthora species. 98 We show that P. infestans and P. sojae genomes encode expanded numbers of 6mA 99 methyltransferases (DAMT). Two of the three DAMTs have methyltransferase activity, 100 and the 6mA methylation landscapes are described at the genome-wide level using 101 methylated DNA immuno precipitation sequencing (MeDIP-seq). Although the majority of 102 the methylation sites localized in the intergenic regions, 6mA also prefers to accumulate 103 around TSS regions in a bimodal distribution pattern and may function as a repressive 104 mark of gene expression. The GSR genes show higher a methylation level than the GDR 105 genes. Consistently, most 6mA sites accumulate in repetitive sequences, such as DNA 106 elements and long terminal repeat (LTR) elements. Furthermore, individual knockouts of 107 each of the three DAMT genes results in a reduction of 6mA level in vivo. Moreover, 108 comparative analysis of the MeDIP-seq data of the mutants suggests that the DAMTs 109 may have functional specificity in targeting particular genomic regions. 110 111

Results 112
To determine whether Phytophthora species can accomplish the 5mC modification, we 113 performed a hidden Markov models based sequence similarity search for 5mC 114 methyltransferase homologs in the P. infestans and P. sojae genomes 30,31 . No predicted 115 gene or homologous sequence corresponding to a 5mC methyltransferase was discovered 116 (Supplementary Table 1). To test directly for the presence of 5mC, we analyzed 117 hydrolyzed genomic DNA (gDNA) samples from P. infestans and P. sojae by high-118 performance liquid chromatography (HPLC) and Ultra Performance Liquid 119 Chromatography Electrospray Ionization -Mass Spectrum (UPLC-ESI-MS/MS). We did 120 not detect 5mC in either species at the parts per billion (PPB) level (10 -9 g/mL) 121 ( Supplementary Fig. 1a, b). Furthermore, the endonuclease McrBC that specifically 122 cleaves DNA containing 5mC did not digest Phytophthora gDNA, similarly to gDNA of 123 Drosophila melanogaster, which is known to not carry 5mC (Supplementary Fig. 1c). 124 Therefore, none of the methods we employed could detect 5mC in either P. infestans or P. 125 sojae DNA. 126 Although we did not identify genes encoding for 5mC methyltransferases in P. 127 infestans and P. sojae, we did identify homologs of 6mA methyltransferases and 128 demethylases in the Phytophthora genomes. Initially, we discovered a potential MT-A70 129 homolog in the P. sojae but not P. infestans genome. However, closer examination of the 130 putative P. sojae MT-A70 gene indicated that it is a pseudogene with a premature stop 131 codon. We found that N6-adenineMlase domain-containing (DAMT) proteins are present 132 in all the examined oomycete species, including Phytophthora species, Albugo species, 133 Hyaloperonospora arabidopsidis, Pythium ultimum and Saprolegnia parasitica (Fig. 1a). 134 The Phytophthora and Saprolegnia genomes each encode three predicted DAMT genes, 135 whereas the other species have only one gene. Phylogenetic analyses of the oomycete 136 DAMTs uncovered two distinct gene clades, namely DAMT1/2 and DAMT3 137 ( Supplementary Fig. 2a). In contrast to DAMT1 and DAMT2, DAMT3 is conserved in all 138 the examined oomycete genomes except H. arabidopsidis (Supplementary Fig. 2a, 139 Supplementary Table 2). DAMT3 is located in a genomic region with a high degree of 140 synteny ( Supplementary Fig. 2b), suggesting that it is probably the ancestral gene. DAMT 141 gene expansion in Phytophthora species therefore appears to be due to the emergence of 142 the DAMT1/2 genes. A closer examination of the catalytic motif responsible for binding the 143 methyl group from S-adenosyl-L-methionine 13,20,32 indicates that DAMT1 and DAMT3 144 proteins have functional motifs consisting of the amino acid sequences DPPY and DPPF, 145 respectively. However, this motif was naturally mutated into EPPH in the DAMT2 proteins. 146 A search in the P. infestans and P. sojae online RNA-seq databases revealed that DAMTs 147 are expressed in all the examined growth stages 33,34 (Supplementary Fig. 2d, e). In 148 summary, bioinformatics analyses indicate that Phytophthora species may possess the 149 enzymatic machinery for 6mA DNA methylation. 150 To verify the enzymatic activity of these putative methyltransferases, we measured the 151 in vitro methyltransferase activity of three P. sojae recombinant DAMT proteins. The 152 recombinant proteins, together with 6mA-free lambda DNA and substrate S-adenosyl-L-153 methionine, were incubated together in an in vitro enzymatic assay 35 . These 154 DpnI, which recognizes the 6mA methylated GATC site, in the presence of recombinant 170 PsDAMT1, PsDAMT3, or the bacterial 6mA methyltransferase DAM (Fig. 1b). Notably, 171 PsDAMT3 was the most active methyltransferase in vitro. We did not detect any activity for 172 PsDAMT2 in this assay, even after increasing PsDAMT2 concentration (Fig. 1b). We also 173 performed a complementary methylation assay in the 6mA deficient E. coli strain HST04. 174 In this assay, E. coli gDNA from DH5α and DAM complemented HST04 transformants were 175 digested by DpnI as expected. The E. coli gDNA from PsDAMT1 and PsDAMT3 176 transformants could also be digested by DpnI, whereas those from PsDAMT2 and the 177 transformants of the catalytically dead mutants (PsDAMT1 APPA , PsDAMT3 APPA ) could not 178 be digested (Supplementary Fig. 3a). Overall, these data indicate that the P. sojae 179 DAMT1 and DAMT3 proteins possess methyltransferase activity in a DPPY(F) motif-180 dependent manner. 181 To test for the presence of 6mA in Phytophthora, we used UPLC-ESI-MS/MS to 182 analyze gDNA samples from P. sojae and P. infestans. A peak matching the retention time 183 of standard 6mA was present in the test samples from these two Phythophthora species 184 ( Fig. 2a). Moreover, the same base fragment was detected in the two samples by MS/MS 185 of the 266.12 (mass/charge ratio), which also matched the standard 6mA (Fig. 2b). Thus, 186 the 6mA DNA base modification is present in the gDNA samples. Furthermore, we 187 estimated the abundance of 6mA in P. sojae and P. infestans to be 400 and 500 parts per  Table 2). The 6mA level in these Phytophthora species is approximately 60-fold higher 190 than in Homo sapiens and Mus musculus, but is lower than a few early-diverging fungal 191 species like Hesseltinella vesiculosa and Piromyces finnis 12,36 . To further test for the 192 presence of 6mA in Phytophthora gDNA, we used commercially available antibodies that 193 specifically recognize the 6mA modification; immune blot signals were robustly detected in 194 gDNA samples of P. infestans and P. sojae. (Fig. 2d). Collectively, our results show that 195 6mA is a naturally occurring DNA modification in Phytophthora genomes. 196 We performed methylated DNA immuno precipitation-sequencing (MeDIP-seq) to 197 obtain a genome-wide insight into the Phytophthora 6mA methylome. The MeDIP-seq 198 experiments on gDNA samples from mycelium growth stages included two biological 199 replicates for each of the two Phytophthora species. After assembling sequencing data and 200 seeking 6mA-enriched regions, we mapped 6mA peaks (6mA-enriched regions) at a 201 genome-wide level with FDR < 0.01 by SICER 37 . A total of 12,611 overlapping methylation 202 peaks were captured from the two P. infestans biological replicates. A total of 3,031 203 overlapping peaks were called from two P. sojae replicates (Supplementary Fig. 4a). 204 Genome-wide 6mA methylation profiling data revealed that 86% and 55% of the 6mA 205 peaks were located in the intergenic regions in P. infestans and P. sojae, respectively 206 ( Supplementary Fig. 4b). The higher proportion of 6mA intergenic localization in P. 207 infestans results from the larger overall fraction of intergenic gDNA in the expanded 240 208 Mbp genome of this species compared to P. sojae. In P. sojae, 25% of the 6mA peaks mark 209 gene bodies, whereas 15% and 5% of the methylations occupy positions upstream and 210 downstream of gene bodies, respectively. Comparatively, in P. infestans, these figures 211 correspond to 8%, 4%, and 2% (Supplementary Fig. 4b). Overall, our analyses revealed 212 1,805 and 1,343 genes with 6mA marks in P. infestans and P. sojae, respectively. 213 Profiling of 6mA distribution in methylated genes revealed that 6mA peaks tend to flank 214 the transcriptional start site (TSS) with a clear depletion near the TSS itself (Fig. 3a-c), 215 resembling the bimodal distribution pattern of 6mA detected in other organisms, 216 analyses when we plot relative 6mA levels from all the methylated and non-methylated 232 genes (Fig. 3c). We illustrate normalized 6mA MeDIP-seq reads mapped onto loci from 233 P. sojae Ps_155563, Ps_128235 and P. infestans PITG_02506, PITG_02507, 234 PITG_15808 as typical examples of 6mA localization patterns (Supplementary Fig. 5a,  235 b). To gain further insight into the characteristics of 6mA methylated genes, we 236 conducted a gene ontology (GO) enrichment analysis of methylated genes in both 237 species. Results from the GO analysis suggest that methylated genes are associated 238 with functional categories, such as chromatin binding, enzyme regulator, and hydrolase 239 (Supplementary Fig. 5c). 240 241 Figure. 3 Bimodal distribution pattern of 6mA around TSS in Phytophthora 242 The distribution of 6mA peaks around TSS was profiled by MeDIP-seq. The 6mA 243 occupancy along TSS from -2kb to 2kb is shown. 6mA peaks were enriched around TSS 244 with a bimodal distribution and a local depletion after TSS in P. infestans and P. sojae. 245 (a) 6mA occupancy in P. infestans methylated genes. All P. infestans genes are divided 246 into 6mA methylated genes (n=1805) and non-6mA methylated genes (n=16374). 247 (b) 6mA occupancy in P. sojae methylated genes. All P. sojae genes are divided into 6mA 248 methylated genes (n=1343) and non-6mA methylated genes (n=17853). are more likely to be associated with 6mA as this group of genes tends to have more 266 abundant 6mA levels; in contrast, highly expressed genes tend to have lower 6mA levels 267 (Fig. 3d, e). To further validate these observations, we examined the gene expression 268 levels of methylated and non-methylated genes in both species. We found that methylated 269 genes have significantly lower gene expression compared to non-methylated genes in both 270 species (Supplementary Fig. 6a, b). Thus, the data suggests that 6mA negatively 271 correlates with gene expression levels in the two Phytophthora species. 272 It is well established that genomes of Phytophthora species have experienced repeat-273 driven expansions and are, therefore, rich in repetitive sequences 25-28 . Thus, we examined 274 the association between 6mA peaks and major types of transposable elements (TEs). A 275 total of 37% (P. infestans) and 15% (P. sojae) of the 6mA peaks locate to long terminal 276 repeat (LTR) elements (class I TEs), whereas 8% (P. infestans) and 10% (P. sojae) of the 277 peaks fall within DNA elements (class II TEs), respectively (Fig. 4a). Statistical analyses 278 indicate that 6mA peaks are enriched in TEs at a significant level (Supplementary Fig.  279   7a). Moreover, 6mA levels in TEs are higher than the average genomic level in both species 280 (Supplementary Fig. 7b). We conclude that the 6mA methylome is preferentially 281 associated with TEs in the two Phytophthora species. 282 The genomes of Phytophthora species have a bipartite "two-speed" architecture with 283 between 6mA and genome architecture, we calculated the average 6mA levels for genes 303 located in GDR and GSR. These analyses revealed that genes in the GSR tend to have 304 higher 6mA level than GDR genes (Fig. 4b, c). Similarly, we plotted the 6mA methylation 305 RPKM value of the region corresponding to the 500bp after the TSS according to local 306 gene density (measured as length of 5' and 3' flanking intergenic regions) to generate the 307 genome architecture heatmaps previously described 25,27 . The heatmaps revealed a clear 308 association between the methylome and genome architecture that genes with higher 6mA 309 levels being enriched in the GSR and reduced in GDR (Fig. 4d, Supplementary Fig. 8a).

310
These observations are consistent with our previous finding that 6mA preferentially 311 accumulate in repetitive and TE-rich regions, which fill the intergenic regions in the GSR 312 of Phytophthora genomes. Interestingly, further MeDIP-seq analyses demonstrated that 313 secretome genes, including RxLR effector genes, which are important in Phytophthora-314 host interactions and are primarily localized in the GSR, have significantly higher 6mA 315 levels than core orthologous genes (Supplementary Fig. 8b, c). We conclude that the 316 6mA methylome is preferentially associated with both the genes and intergenic regions 317 that form the gene-sparse compartments of Phytophthora genomes. 318 To further investigate the function of DAMTs in Phytophthora, we individually 319 knocked out DAMT genes in the P. sojae strain P6497 using CRISPR/Cas9 gene editing 320 methodology. We designed two sgRNAs matching two sites in each of the DAMT genes 321 and harvested at least three independent knockout transformants for each gene 322 (Supplementary Table 3, Supplementary Figure 9). We selected homozygous mutants psdamt2-T52 were dramatically reduced to only 8.1% and 5.8% of the wild-type strain, 327 whereas psdamt3-T9 was reduced to 11.0% (Fig. 5a, Supplementary Table 4). Although 328 our biochemistry assays showed that PsDAMTs have different levels of enzymatic 329 Given that DAMT3 encodes a functional methyltransferase that is conserved among 346 all examined oomycete species, we examined the methylome in the psdamt3 T9 mutant in 347 more detail using MeDIP-seq. We observed a significant reduction in 6mA levels around 348 TSS (Fig. 5b). 6mA signals were also weaker in psdamt3 than wild-type at LTR elements 349 and DNA elements regions (Supplementary Fig. 10a). Also, we noted a similar reduction 350 in 6mA levels for both GSR and GDR genes in the psdamt3 mutant compared to wild-type 351 (Supplementary Fig. 10b). We conclude that the PsDAMT3-regulated 6mA methylome is 352 not specifically associated with the bipartite genome architecture. However, close 353 examination of the bimodal methylation pattern around the TSS of 6mA methylated genes 354 uncovered a greater loss in the second peak in the psdamt3 mutant compared to the wild-355 type (Fig. 5c). This unexpected finding indicates that DAMT genes may have some degree 356 of functional specialization, and that PsDAMT3 may have a preference for the methylation 357 of gene bodies after the TSS. Representative genomic segments with typical changes in 358 6mA localization are illustrated for psdamt3 and wild-type P. sojae (Fig. 5d, e). The MeDIP-359 seq data partially explains the significant reduction of total 6mA levels in the psdamt3 360 mutant, but also illustrates the uneven reduction pattern around the TSS, suggesting that 361 there are complex patterns of 6mA methylation by the expanded 6mA methyltransferases 362 of P. sojae. 363 364

Discussion: 365
It recently became evident that 6mA is not only an important epigenetic mark in 366 prokaryotes but is also a feature of eukaryotic genomes. Here, we demonstrate that 6mA 367 methylation occurs in the oomycete plant pathogens P. infestans and P. sojae, and 368 document the methylome of these species. Remarkably, the 6mA methylomes are 369 preferentially associated with genes and transposable elements that form the gene-sparse 370 compartments of Phytophthora genomes and have been implicated in adaptive evolution 371 of these pathogens. We discovered that 6mA methyltransferases have expanded into three 372 enzymes in Phytophthora which do not appear to be fully redundant given that each of the 373 single knock-out mutants had a significantly reduced methylome. Based on mutant 374 analyses, we noted that PsDAMT3 may have a preference for methylation of gene bodies 375 after their TSS. Overall, the observed 6mA patterns around the TSS in the PsDAMT3 376 mutant suggest complex patterns of 6mA methylation by the expanded 6mA 377 methyltransferases of P. sojae. 378 Although 6mA appears to be prevalent in eukaryote genomes, most studies report low 379 levels of abundance. Previous studies documented significant variation of 6mA abundance 380 (6mA/A) ranging from 0.00019% to 2.8% among different eukaryotes 12 . Here, we 381 determined the abundance of 6mA to be 0.05% and 0.04% in the mycelium stage of P. 382 infestans and P. sojae, respectively. Beside modification abundance, genome distribution 383 pattern is another way to value 6mA biological significance. Unlike reports from some other 384 organisms, 6mA is not evenly distributed across Phytophthora genome from our research.  (Supplementary Fig. 8b, c). This data is also consistent with the 428 observation that 6mA marks are associated with low gene expression levels and may 429 therefore contribute to the global down-regulation of virulence effector genes during 430 vegetative stages, as proposed for the fungal pathogen Leptosphaeria maculans 43 . 431 Alternatively, 6mA marks may contribute to the stochastic gene silencing of effector genes, 432 thus enabling the emergence of pathogen races that evade plant immunity. Indeed, effector 433 gene silencing has been linked to rapid evolution in both P. sojae and P. infestans 44,45 . In 434 addition, Phytophthora species tend to exhibit high levels of expression polymorphisms in 435 genes located in the GSRs 46 . In summary, we hypothesize that 6mA is involved in virulence 436 gene expression, thus shaping host adaptation and enhancing evolvability in the plant 437 pathogen Phytophthora. 438 In this study, we identified genes predicted to encode 6mA methyltransferases and 439 demethylases in P. infestans and P. sojae. We initially focused on studying the functionality 440 of the predicted 6mA methyltransferases to provide evidence that Phytophthora species 441 have the inherent capability to perform this DNA modification. Our present work shows that 442 DAMT homologs are the major N6-adenine methyltransferases in Phytophthora species. from the PFAM database were used to BLAST search homologous enzymes with an E-506 value cut-off 10 -5 29,30 . 507

Dot blot assay 508
Genomic DNA of P. sojae and P. infestans were extracted using TIANGEN DNAsecure 509 Plant kit. Different amounts of gDNA were denatured at 95 ⁰ C for 5 mins and chilled in ice 510 for 10 mins. DNA were spotted on HybondTM-N+ membranes. The membrane was allowed 511 to dry at 37 ⁰ C for 20 mins and then crosslinked using HL-2000 HybriLinker for 5 mins. 512 The membrane was blocked in 5% milk PBST for 1h at room temperature, and then 513 incubated with 6mA antibody (sysy202003) in 5% milk PBST overnight at 4 ⁰ C. After 3x 514 10 min washes with PBST, DNA and membrane were incubated with secondary antibody 515 (ab6721) for 30 mins at room temperature. After 3x 10 min washes with PBST, the 516 membrane was treated with Pierce ECL Western Blotting Substrate (Prod#32106) and 517 detected by Tanon-5200Mutil. 100 ng input DNA of every samples were loaded on 1% 518 agarose gels, followed by air drying for 5 mins and photographed using Clinx GenoSens. 519

HPLC analysis for 5mC 520
The HPLC separation was performed on a Zorbax SB-C18 column (2.1 mm x 150 mm, 521

DpnI-dependent methylation assay 549
DpnI-dependent methylation assay was performed as previously described 34  system were incubated at 37 ⁰ C for 1 hour, and then 65 ⁰ C for 15 min to stop the reaction. 554 The methylated DNA was digested by 5U DpnI at 37 ⁰ C for 1 hour. Digestion was stopped 555 by heat inactivation by incubating at 80 ⁰ C for 20 mins. 1% agarose gel electrophoresis 556 was used to check digestion. PsDAMT1, PsDAMT2, PsDAMT3, DAM, PsAvr3c were 557 cloned into pET32a-c (+). The recombinant plasmids were transformed into E. coli HST04 558 strain (dam-, dcm-). Bacteria were grown overnight at 37 ⁰ C. E. coli gDNA were extracted 559 using TIANamp Bacteria DNA Kit. DpnI was bought from NEB and used as protocol 560 described. All the samples and standards were loaded at 1 µg. 1% agarose gel 561 electrophoresis was used to check digestion results. 562

MeDIP-seq (6mA-IP-seq) 563
MeDIP-seq used in this paper was optimized from several protocols 13-15 . gDNA was 564 extracted using TIANGEN DNASeure Plant Kit and then treated with RNase A overnight. 565 Then the gDNA was diluted to 100 ng/µl with TE buffer, 100µl diluted gDNA was put in each 566 tube and sonicated to 200-400 bp using Biorupter UCD-600. The 200-400 bp sized DNA 567 was extracted using Takara Gel DNA Extraction Kit ver.4.0. DNA was denatured at 95 ⁰ C 568 for 10 mins and chilled in ice immediately for 5 mins. 20 µl of denatured DNA was stored 569 as input. The rest of the DNA was incubated with 3 µg 6mA antibody at 4 ⁰ C for more than 570 6 hours. Dyna beads (Thermofisher 10001D) were washed twice using 1×IP buffer and 571 pre-blocked in 0.8 mL 1×IP buffer with 20 µg/µl BSA. Pre-blocked beads were washed 572 twice using 1×IP buffer (5×IP buffer: 50mM Tris-HCl, 750mM NaCl and 0.5% vol/vol 573 IGEPAL CA-630), and then incubated DNA-antibody was added to pre-blocked beads, and 574 rotated overnight at 4 ⁰ C .The beads were washed 4 times for 10 mins with 1×IP buffer. 575 IP products were suspended in 400 µL preheated elution buffer at 65 ⁰ C for 15 mins to 576 yield the 6mA-IPed library; repeat this step using 300 L preheated elution buffer (Elution 577 buffer: 50 mM NaCl, 20 mM Tris-HCl, 5 mM EDTA, 1% SDS). Eluted DNA was combined 578 and then added to an equal volume of phenol-chloroform-isopentanol, vortexed and 579 centrifuged at 13000 rpm for 5 mins at room temperature. The aqueous phase was 580 transferred into a new tube and mixed with an equal volume of ethanol to precipitate the 581 eluted DNA. The library was prepared using VAHTS TM Turbo DNA Library Prep Kit for 582 Illumina and AHTS TM Multiplex Oligos set 1 for Illumina. Sequencing was done by BGI 583 (Shenzhen) and GENEWIZ (Suzhou). 584

RNA extraction, RNA-seq and qRT-PCR 589
Total RNA of 3-day-old P. sojae hyphae were isolated using Omega Total RNA Kit I 590 according to the manufacturer's manual. RNA quality was measured using Nanodrop ND-591 1000 and 1% agarose gel electrophoresis. RNA-seq service was provided by BGI and 592 1gene. RNA reverse transcription was conductd using Takara PrimerScript TM RT reagent 593 Kit with gDNA eraser. Quantitative RT-PCR was performed using the ABI PRISM 7500 Fast 594

Real-Time PCR System. 595
High-throughput sequence data analysis 596 RNA-seq data was mapped to P. sojae v1.1 using Tophat2, and MeDIP-seq data was 597 mapped to P. sojae v1.1 using bowtie2. Gene expression data was generated by Cufflinks. 598 MeDIP-seq data was normalized and visualized using deepTools 51 and IGV 52 . 6mA 599 methylation peaks were called using SICER. Figure of two-speed genome was produced 600 as describe before 53 , 6mA distribution was calculated as log2(IPRPKM/inputRPKM+1), RPKM 601 value from the regions before TSS 500bp (real length and reads number will be calculated 602 if the length of flanking intergenic regions<500bp). Phytophthora repeat sequences were 603 referenced in previous publications 24 and re-annotated here by RepeatMasker 54 . 604 605