Plants with visible crown rot or leather rot symptoms were collected from different strawberry and apple plantations in Norway. Infected rhizomes (crowns) or fruits were rinsed in tap water for 20 min to remove superficial soil and debris. The samples were surface sterilized with 70% ethanol for 10 s, followed by treatment with 0.5% sodium hypochlorite for 90 s, and finally rinsed three times with distilled water. Infected plant parts were sliced into rectangular cubes using a sterilized scalpel and were placed on solid PARP medium (Jeffers and Martin, 1986). The plates were incubated in the dark at room temperature (20°C) for 5–7 days and were examined for mycelium growth and typical sporulation patterns (sporangia and oospores) of P. cactorum under light microscopy.

Eleven P. cactorum strains were tested for their ability to cause (1) crown rot in the rhizome of the susceptible strawberry cultivars Polka and Korona; and (2) leather rot on unripe fruits of Korona. Five of the P. cactorum strains from naturally infected strawberry fruits (252360, 252362, 252363, 252364, 252365), and one strain from apple fruit (251539) were tested for their ability to cause crown rot in the highly susceptible cultivar Polka, to identify the leather rot pathotype. The four strains from naturally infected strawberry rhizomes (251171, 251615, 251616, 251683) were tested for their ability to cause crown rot and leather rot in “Korona” (moderately resistant to P. cactorum; Eikemo and Stensvand, 2015), together with previously isolated crown rot strains 10300 (Armitage et al., 2018) and P414 (Nellist et al., 2021). A moderately resistant cultivar was used in this case to study the aggressiveness of different P. cactorum strains based on the degree of necrosis in the rhizome.

Strawberry plants were grown in a greenhouse at 18°C with a 16/8 h light/dark regime. For pathogen inoculation, each P. cactorum strain was grown on 10% V8 (vegetable juice) agar plates (Jeffers, 2015) and incubated at room temperature (20°C) for 2 weeks to promote sporangia formation. Zoospore suspensions were prepared using autoclaved pond water as previously described by Eikemo et al. (2000). Plant tissue was gently wounded using a sterilized scalpel (for rhizome inoculation) or a pipette tip (for fruit inoculation) and inoculated with 2 × 10⁵ zoospores of P. cactorum. Subsequently, the rhizome infection was followed for 4 weeks, and disease symptoms were scored 8-1, where scores 8, 7, 6, 5 represent plants that died due to wilting during the first, second, third, and fourth week after inoculation with P. cactorum, respectively. Plants that survived 4 weeks after inoculation were bisected longitudinally using a sterile scalpel and scored based on the degree of necrosis in the rhizome (4 = clear necrosis covering at least 50% of the rhizome area, 3 = small patches of necrosis, 2 = minor dark brown speckles, 1 = no symptoms) as described by Bell et al. (1997). Inoculated fruits were given a binary score (0 = no infection and 1 = infection).

Genomic DNA was isolated from the mycelium of five P. cactorum strains 251539, 251616, 251683, 252360, and 252365 using a modified phenol-chloroform extraction protocol for Phytophthora (Joint Genome Institute, 2021). Briefly, 5 g of mycelial mat grown in 100 mL V8 broth was collected using a Miracloth (EMD Millipore, Merck, Germany) and ground using a pre-cooled mortar and pestle, plus liquid N_2. Ten mL of extraction buffer (0.2 M of tris(hydroxymethyl)aminomethane, pH 8.5, 0.25 M sodium chloride, 25 mM ethylenediaminetetraacetic acid, and 0.5% sodium dodecyl sulfate) was added to the mycelial powder and mixed well using a pipette. Subsequently, 20 μL of 100 μg/mL RNase A (Qiagen, Germany) was added and the lysate was incubated at 65°C for 10 min. Following the incubation, 7 mL of tris-equilibrated phenol, pH 8.0 and 3 mL of chloroform was added, gently mixed, and incubated at room temperature (20°C) for 1 h. Samples were centrifuged at 6,000 g for 15 min at 4°C. The aqueous phase was added to an equal volume of chloroform, gently mixed, and centrifuged at 6,000 g for 5 min at 4°C. After centrifugation, the aqueous phase was transferred to a pre-cooled 50 mL centrifuge tube (VWR, United States), gently mixed with 0.6 volume of chilled isopropanol, incubated on ice for 30 min, and centrifuged at 6,000 g for 30 min at 4°C to recover the precipitated DNA. The DNA pellet was washed three times with 70% ethanol, air dried and dissolved in 300 μL of nuclease free water. The DNA concentration was measured using the NanoDrop^™ 2000 Spectrophotometer (ThermoFisher Scientific, United States). An aliquot of the isolated DNA was analyzed using agarose gel electrophoresis run at 30 V for 16 h to assess the quality of the intact genomic DNA for sequencing.

Twenty micrograms (50 ng/μL) of the isolated DNA for each sample was sent to the Norwegian Sequencing Centre, Oslo, for library preparation and sequencing. All five P. cactorum genomic libraries were prepared according to the manufacturer’s protocol for Multiplexing SMRTbell libraries using the SMRTbell Express Template Prep Kit 2.0 (Pacific Biosciences, United States). DNA was fragmented to 20 kb fragments using Megaruptor^® 3 (Diagenode s.a., Belgium) and size selected using Bluepippin with 15 kb cut-off. The libraries were sequenced on one 8 M SMRT cell using the Sequel II Binding Kit 2.0 with the sequencing chemistry v2.0 from the PacBio Sequel II (Pacific Biosciences, United States). Reads were demultiplexed using Demultiplex Barcodes pipeline on SMRT Link v 8.0.0.80529 (SMRT Tools v 8.0.0.80502).

Raw subreads were trimmed for adapters and low-quality regions and processed to obtain error corrected CCS (circular consensus sequencing) reads. The high-quality CCS (HiFi) reads (Phred quality score ≥ Q20) with a minimum predicted accuracy 0.9 were used for de novo genome assembly using the automated Flye v2.8.2 pipeline (Kolmogorov et al., 2019) in OmicsBox (2019). Because of the high-quality of the sequenced reads and to increase the possibility of detecting true sequence diversity among the five strains of P. cactorum, de novo genome assembly was performed. Assembly statistics were generated using QUAST v5.0.2 (Gurevich et al., 2013) and BUSCO statistics were obtained using BUSCO v4.1.2 (Simão et al., 2015) with the eukaryota_odb10 database (updated on 2020-03-06). The repeat contents of the assembled genomes were obtained using RepeatMasker Open-4.0 (Smit et al., 2015) and RepeatMasker combined library of Dfam_Consensus and RepBase v20170127 (Bao et al., 2015; Hubley et al., 2016). Genes were predicted using the assembled contigs of each P. cactorum genomes in the ab initio mode of Augustus v3.3.3 (Stanke et al., 2008). Augustus was trained using predicted protein sequences from P. cactorum strain 10300 (Armitage et al., 2018).

The predicted proteome of the P. cactorum strains were annotated using a combination of BLASTP (Altschul et al., 1990) against the NCBI non-redundant (Nr) database and InterProScan (Zdobnov and Apweiler, 2001) to assign individual sequences with gene ontology (GO) terms (Ashburner et al., 2000) using Blast2GO (Götz et al., 2008).

Secreted proteins of P. cactorum (secretome) were predicted using a combination of tools that included SignalP v5.0 (Armenteros et al., 2019b) to detect signal peptide cleavage sites, Phobius (Käll et al., 2007) to predict transmembrane domains and signal peptide cleavage sites, TargetP v2.0 (Armenteros et al., 2019a) to predict the presence of mitochondrial transit peptides and signal peptide cleavage sites. The presence of endoplasmic reticulum retention signals (KDEL or HDEL motifs) were predicted as previously described by Evangelisti et al. (2017) using PROSITE-Scan (de Castro et al., 2006). No subcellular localization of “extracellular” as criterion for predicting secretion was considered, as recommended for oomycetes (Sperschneider et al., 2015). A protein was defined as “secreted” if it contained a signal peptide cleavage site within the first 70 amino acids in the N-terminus, had no more than one transmembrane domain, as recommended for oomycetes (Sperschneider et al., 2015), and had no mitochondrial localization or endoplasmic reticulum retention signals. The predicted secretome was used to predict the apoplastic proteins using ApoplastP (Sperschneider et al., 2018).

Genes encoding RxLR and CRN effectors were predicted using the six-frame translation of open reading frames (ORFs) with a minimum cutoff of 100 aa using getorf from EMBOSS (Rice et al., 2000). The translated ORFs were used to search for canonical RxLR-EER motifs (for RxLR effectors) and the LFLAK-HVLV motifs (for CRN effectors) in R using the effectR package (Tabima and Grünwald, 2019). Predicted RxLR and CRN effector proteins were searched for signal peptides in the N-termini using SignalP v5 and annotated using BLASTP (Altschul et al., 1990) against the NCBI non-redundant (Nr) database. Candidate effectors with a predicted signal peptide in the N-terminus and no more than one transmembrane domain were considered high-confidence effectors.

Orthologues of the predicted proteome of the P. cactorum strains sequenced in this study (251539, 251616, 251683, 252360, and 252365), the P. cactorum 10300 strain (Armitage et al., 2018), the Yunnan strain (Yang et al., 2018a), and predicted proteomes of publicly available P. capsici LT1534 v11.0 (Lamour et al., 2012), P. infestans T30-4 (Haas et al., 2009), P. palmivora ZC01 v1 (Gil et al., 2020), P. parasitica INRA-310 (NCBI dataset, GCA_000247585.2 PP_INRA-310_V2, unpublished), and P. sojae P6497 v3.0 (Tyler et al., 2006) were identified using OrthoFinder v2.4.0 (Emms and Kelly, 2019). The concatenated orthologues (one ortholog per orthogroup per genome) shared by the P. cactorum strains and other Phytophthora spp. were used to construct a phylogenomic tree using the maximum-likelihood method implemented in the phangorn package (Schliep, 2011), with the Le Gascuel (LG) model for amino acid substitutions (Le and Gascuel, 2008). Predicted secreted proteins of the sequenced P. cactorum strains were compared using Orthovenn2 with expectation (E) value equal to 1e-10 and inflation value equal to 1.5 (default) to obtain higher resolution between groups (Xu et al., 2019). Predicted RxLR and CRN proteins were clustered using the OrthoFinder v2.4.0 and visualized in R using the Tidyverse package collection (Wickham et al., 2019).

The HiFi reads were aligned to the de novo assembled genome of P. cactorum crown rot strain 251616 (used as reference) using pbmm2 v1.3.0¹ wrapper for minimap2 aligner (Li, 2018). The output bam files were indexed using samtools v1.10 (Danecek et al., 2021), and variants were called in the HiFi reads using DeepVariant v1.0.0 with a pre-defined model (--model_type = PACBIO; Poplin et al., 2018). Phasing of DNA variants was performed using WhatsHap v1.0 (Patterson et al., 2015). The phased variants (SNPs) were filtered and sorted using VCFtools v0.1.15 (Danecek et al., 2011). The resulting phased SNPs and indels (unphased) were annotated, and variant effects were predicted using SnpEff (Cingolani et al., 2012). To perform variant annotations, a manual database was built using the general feature format (GFF3) file of the predicted genes of strain 251616. High-quality SNPs and indels with Phred quality score ≥ 20 were used for downstream analysis. Pairwise alignment of selected protein-coding genes from strain 251616 and a representative leather rot strain 251539 was done using LASTZ v1.04.03² (Harris, 2007). The rate of nonsynonymous and synonymous substitutions (dN/dS or ω) was calculated using KaKs_Calculator 2.0 (Wang et al., 2010). Genes with dN/dS ratio greater than one and p < 0.05 (calculated by the Fisher’s extract test) were considered under positive selection.

Seven P. cactorum strains that were isolated from strawberry fruits and apple fruit were tested for virulence in the rhizome (crown) by inoculation in the highly susceptible cultivar Polka. Four of these (252361, 252362, 252363, 252364) developed brown necrotic lesions in the rhizome, and were therefore classified as crown rot (CR) pathotype. In contrast, the strain from apple fruit 251539, and two other strawberry fruit strains 252360 and 252365, did not develop crown rot symptoms (Table 1; Figure 1A). Consequently, the latter three P. cactorum strains were classified as leather rot (LR) pathotype.

The four P. cactorum strains isolated from rhizomes with symptoms of crown rot, 251171, 251615, 251616, and 251683, were tested for virulence by inoculation into the rhizome of the moderately resistant cultivar Korona, together with two control crown rot strains (10300 and P414). As expected, all strains developed necrosis in the rhizome, but symptoms varied in severity (Table 1; Figure 1A). Strain 251616 developed a severe necrosis in the strawberry rhizome and is hereafter referred to as high virulence crown rot strain, while the strain 251683 developed few necrotic lesions and is therefore referred to as low virulence crown rot strain (Table 1; Figure 1A). All these strains also developed leather rot on strawberry fruit 4 days after inoculation of berries (Figure 1B).

The genomes of two CR strains and three LR strains were sequenced. A total of approximately 3.9 million raw reads were obtained with an average read length of 73 kilobases (kb). High-quality CCS (HiFi) reads used for de novo genome assemblies of the five P. cactorum strains resulted in genome sizes of 66.4 to 67.6 Mb, each in 178 to 203 contigs with N50 values ranging from 892.1 to 1036.2 kb. The repeat content of the de novo assembled P. cactorum genomes ranged from 14.5 to 16.1 Mb (mean 15.7 Mb), which constituted a mean of 23.3% of the total genome size. The CR strains of P. cactorum, 251616 and 251683, contained slightly more repetitive elements than the LR strains, 251539, 252360, and 252365 (Table 2).

The total number of predicted complete genes in the assembled genomes of the five P. cactorum strains ranged from 17286 to 17398 (Supplementary Table S1). All five P. cactorum assemblies contained 239 of 255 core eukaryotic genes (CEGs), representing a genome completeness of 94%, which was comparable to the percentage of CEGs (96%) in the predicted proteomes. The near complete set of CEGs suggested overall accurate gene prediction and support the validity of the used model. The percentage of CEGs was similar in the assembly (94–96%) to those of the previously published CR strains 10300 and P414 assembly (Table 2).

Highly contiguous genome assemblies provide a better framework to understand the biology of the studied organisms by offering better information on the genomic context of traits of interest. In this work, long HiFi reads facilitated high-quality genome assemblies of the five sequenced P. cactorum strains. The total number of contigs ranged from 178 to 203, which was a major improvement from the 4,623 contigs of our previously sequenced P. cactorum 10300 genome assembly and 3644-20136 contigs of other publicly available P. cactorum genome assemblies (Grenville-Briggs et al., 2017; Armitage et al., 2018; Nellist et al., 2021). The genome sizes of the P. cactorum strains sequenced here were similar to the 10300 assembly and several other P. cactorum assemblies recently reported (Nellist et al., 2021), and therefore, according to expectations. The high number of contigs and low N50 value in the Yunnan assembly might help to explain the unexpectedly large genome size of 121.5 Mb and a distinct clade in the phylogenomic analysis in this work (Figure 2A). The previously sequenced genome of strain P414 (deposited as a reference genome in NCBI) assembled into 194 contigs (Nellist et al., 2021), which is in the range of our five newly sequenced genomes reported here. However, the new genome assemblies have 38–60% higher (and better) N50 value than the P414 assembly, which likely improves genome quality, particularly in regions of repetitive sequences (Table 2).

The whole genome phylogenomic study showed that P. cactorum strains from the strawberry rhizome or fruit tissues were genetically more similar than the apple strain 251539. The result was consistent with the variant analysis that identified a higher number of SNPs and small indels in the apple strain than the P. cactorum strains from strawberry host. A similar study showed that P. cactorum from strawberry and apple hosts were genetically two distinct population that formed two non-recombining clades for strawberry CR and apple strains (Nellist et al., 2021). The present study supports the previous findings that describe P. cactorum as species complex and should not be considered as generalist pathogen.

Understanding the mechanisms of plant recognition of pathogen molecules (PAMPs/MAMPs) and identifying virulence determinants of the pathogen is essential to develop effective strategies to control strawberry diseases. The comparative genomic analyses of the crown- and leather-rot strains revealed genes encoding proteins potentially involved in pathogenesis, with sequence variations and presence or absence polymorphisms. The presence of two unique pectate lyase genes (251616_g6270 and 251616_g6271) in the CR strain 251616 might contribute to the virulence in the strawberry crown. Interestingly, the homologous gene Pcac1_g14072 of the previously sequenced CR strain P414 was shown to be upregulated during early infection in strawberry (Nellist et al., 2021). Pectate lyases secreted by plant pathogens are known to degrade networks of pectin present in the plant cell wall. They cleave the glycosidic bonds of two saccharide units via a β-elimination mechanism (Collmer and Keen, 1986; Herron et al., 2000). Studies have shown that pectate lyases contribute to virulence of bacteria, fungi, nematodes, and oomycetes (Wegener, 2002; Fu et al., 2013; Cho et al., 2015; Chen et al., 2021). A study of pectate lyase genes PL1, PL15, PL16, and PL20 from P. capsici showed that overexpression of these genes in a mildly virulent strain transformed it to a highly aggressive strain (Fu et al., 2015). Another study demonstrated that deletion of a pectate lyase gene PEL1 in the fungal pathogen Verticillium dahliae compromised virulence in Nicotiana benthamiana and cotton plants (Yang et al., 2018b). The two pectate lyases genes in the high virulence CR strain (251616_g6270 and 251616_g6271) are promising candidates to further study P. cactorum virulence in strawberry.

A unique cluster of putative CBD proteins was observed in all the LR strains. Previous studies suggested that CBD proteins are involved in oomycetes cell wall development and interacts with plant cell wall components (Gaulin et al., 2006; Jones and Ospina-Giraldo, 2011). The presence of CBD proteins at the cellular surface of Phytophthora facilitates binding to cellulosic substrates (Gaulin et al., 2002). In plants, alteration of cellulose in the plant cell wall is a warning signal to activate defense responses (Dumas et al., 2008). Although, an increased accumulation of cytosolic calcium was reported in response to CBEL in tobacco cells (Dumas et al., 2008), the exact mechanism of plant defense is unknown. Studies have shown that Phytophthora CBD proteins with lectin-like activities such as CBELs act as potent elicitors of plant defense (Mateos et al., 1997; Gaulin et al., 2006). In this study, the CBD protein-coding genes (251539_g13496, 252360_g4306, 252365_g5402, and 251683_g8771) was of particular interest as these were identified in a cluster unique to the LR strains and the low virulence CR strain. A similar CBD gene (251616_g13858) with four nonsynonymous but no synonymous SNPs was detected in the high virulence CR strain, reflecting signs of positive selection or result of functionally neutral mutation fixed due to drift.

In addition to the variations in the CBD proteins between the high virulence CR strain and the LR strains (including the low virulence CR strain), two single amino acid polymorphisms (T154M and T169I) were observed in the C-terminus of a beta elicitin protein. Interestingly, threonine residues were conserved in both the alleles of the beta elicitin in the high virulence CR strain 251616 and in the previously sequenced CR strains 10300 and P414, except in the strains 15–13 (M¹⁵⁴, I¹⁶⁹) and 17–21 (M¹⁵⁴, T¹⁶⁹; Supplementary Table S3). All LR strains from strawberry fruit and apple including the low virulence CR strain 251683 contained one or both alleles with methionine and isoleucine at these positions. Many Phytophthora elicitins are rich in threonine, serine and proline residues at their C-terminal domain and these residues are suggested to be involved in association with the cell wall through extensive glycosylation (Jiang et al., 2006). Elicitins are often recognized as PAMPs by the host defense machinery, possibly by immune receptors such as BRI1-associated receptor kinase 1 (BAK1) and cell wall-associated receptor-like kinase 1 (WAK1) (Chaparro-Garcia et al., 2011; Raaymakers and Van den Ackerveken, 2016). Pernollet (1993) studied the role of threonine residues in α-elicitin from P. cactorum and proposed that the replacement of a lysine residue by threonine is partly responsible for reduced necrotic activity of the α-elicitin, possibly by avoiding the host defense machinery (Pernollet, 1993). Similar peculiarities were observed in a beta elicitin protein sequence from P. megasperma that was less toxic than related proteins from other Phytophthora spp. (Huet and Pernollet, 1993). Furthermore, cysteine-rich secretory proteins encoded by the SCR-108-like gene cluster detected in the LR strains are interesting candidates for host recognition as similar SCR (e.g., SCR96) of P. cactorum was reported to induce cell death responses in other host plants (Chen et al., 2016). Further study is needed to examine if the putative elicitor genes such as CBD, beta elicitin and SCR-108-like from the LR strains are expressed during interaction with strawberry. It is worth mentioning that the expression of the beta elicitin genes (Pcac1_g22870, Pcac1_g22871) from the CR strain P414 were upregulated during strawberry infection in the transcriptome dataset previously reported by Nellist et al. (2021).

Several genes encoding annexin family members, Crinklers, cysteine-rich secretory protein, a mucin protein, a putative transmembrane protein had sequence variation in the LR strains compared to the high virulence CR strain. Some of the gene products from these, such as an annexin family protein and mucin have been reported to be associated with the cell surface in the close relative P. infestans (Grenville-Briggs et al., 2010). Larousse et al. (2014) studied the role of mucin-like genes in P. parasitica and found that their gene products accumulated on the surface of biofilms, which favor attachment and promote virulence through aggregation. Thus, sequence variations in these genes might affect P. cactorum tissue specific interaction or pathotype.

The high virulence CR strain may possess altered recognition domains or have lost potential elicitors as a mechanism to escape host defense. Since the LR strains were able to cause disease in strawberry fruits after inoculation, but not in the rhizome of the same cultivar, led us to question whether activation of host defense machinery was tissue specific. Eikemo and Stensvand (2015) studied the resistance level of both the tissues (rhizome and fruit) of the same strawberry cultivars against crown rot and leather rot. They reported that cultivars most robust to leather rot was susceptible to crown rot and vice versa. Thus, the mechanisms for resistance against the two diseases seem to be different on the two tissues types. Casado-Díaz et al. (2006) studied the spatial expression pattern of defense-related genes in strawberry during infection with Colletotrichum acutatum, a hemibiotrophic fungal pathogen. They reported that expression of a leucine rich repeat receptor-like protein-coding gene Falrrk-1 and a chitinase gene Fachit-1 was strongly downregulated in infected fruit tissue, while their expression increased several folds higher in the infected rhizome (crown) tissue. They also found that strawberry fruit was more susceptible to C. acutatum infection than crown tissues. Thus, host-pathogen interactions may be influence by tissue architecture and/or differential expression of defense-related genes. The identification of potential elicitor genes of P. cactorum warrants further investigation to understand the virulence mechanisms of the CR and LR pathotypes in strawberry. In addition, several secreted protein-coding genes were detected only in the genome of P. cactorum strains from strawberry host and not present in the genome of apple strain. Most of these genes encode proteins that have a putative role in pathogenesis, suggesting co-evolution and acquisition of pathogenesis-related genes during host adaptation as observed in other pathosystems (Gaulin et al., 2018; Jangir et al., 2021).

The new genome assemblies of P. cactorum identified a greater number of predicted RxLRs via the REGEX and HMM models than previously reported for other strains of the species (Armitage et al., 2018; Yang et al., 2018a; Nellist et al., 2021). Remarkably, most of the predicted CRNs lacked a conventional signal peptide in the N-terminus. It has been proposed that CRN effectors that lack a conventional signal peptide might translocate into host cells via an unconventional secretion pathway (Meijer et al., 2014; Amaro et al., 2017; Voß et al., 2018). The number of predicted CRN effectors was in agreement with the previously published genomes from P. cactorum strains (Armitage et al., 2018; Nellist et al., 2021) and P. palmivora P16830 (Evangelisti et al., 2017).

The analysis of the predicted RxLR and CRN effectors provided evidence for expansion and contraction of certain gene clusters in the investigated P. cactorum strains (Figure 6). In the high virulence CR strain, the unique CRN gene 251616_g2371 and the expansion of CRNs genes in the orthogroups OG0000310 and OG0005694 need further investigation for their virulence function in the P. cactorum-strawberry interaction. In addition, the loss of RxLR genes in orthogroup OG0000414 in this strain should be considered for future functional studies, if they are expressed during infection of strawberry. Among the reduced effector genes in this strain was an RxLR gene with sequence similarity to Avr1b-1 from P. palmivora. Gene loss has been recognized as an important mechanism for emergence of virulence in bacterial, fungal, oomycete pathogens (Maurelli, 2007; Qutob et al., 2009; Bliven and Maurelli, 2012; Rouxel and Balesdent, 2017; Siscar-Lewin et al., 2019). For instance, loss of Avr-Co39 in rice-infecting haplotypes of Magnaporthe oryzae has been proposed as an adaptive mechanism to escape recognition by the R protein Pi-Co39 (Couch et al., 2005). However, loss or mutated effectors may impose a fitness penalty on the pathogen (Leach et al., 2001; Montarry et al., 2010). The functional redundancy of paralogous genes of RxLRs of P. cactorum could play a potential role in conserving virulence function and fitness, as addressed for bacterial and oomycete effectors during pathogenesis (Birch et al., 2008; Ghosh and O’Connor, 2017). Nellist et al. reported that over half of the candidate effectors were not expressed or showed low expression in planta (Nellist et al., 2021). Thus, further expression analysis is required to pinpoint candidate effectors involved in tissue-specific interactions of P. cactorum in strawberry.