The Hidden Truths of Fungal Virulence and Adaptation on Hosts: Unraveling the Conditional Dispensability of Minichromosomes in the Hemibiotrophic Colletotrichum Pathogens

Colletotrichum spp. are ascomycete fungi and cause anthracnose disease in numerous crops of economic significance. The genomes of these fungi are distributed among ten core chromosomes and two to three minichromosomes. While the core chromosomes regulate fungal growth, development and virulence, the extent to which the minichromosomes are involved in these processes is still uncertain. Here, we discuss the minichromosomes of three hemibiotrophic Colletotrichum pathogens, i.e., C. graminicola, C. higginsianum and C. lentis. These minichromosomes are typically less than one megabase in length, characterized by containing higher repetitive DNA elements, lower GC content, higher frequency of repeat-induced point mutations (RIPMs) and sparse gene distribution. Molecular genetics and functional analyses have revealed that these pathogens harbor one conditionally dispensable minichromosome, which is dispensable for fungal growth and development but indispensable for fungal virulence on hosts. They appear to be strain-specific innovations and are highly compartmentalized into AT-rich and GC-rich blocks, resulting from RIPMs, which may help protect the conditionally dispensable minichromosomes from erosion of already scarce genes, thereby helping the Colletotrichum pathogens maintain adaptability on hosts. Overall, understanding the mechanisms underlying the conditional dispensability of these minichromosomes could lead to new strategies for controlling anthracnose disease in crops.


Introduction
Colletotrichum spp.are ascomycete fungi and represent an economically significant group of pathogens of numerous crops worldwide, resulting in severe yield losses.The species include, but are not limited to, C. graminicola, C. lentis and C. higginsianum.C. graminicola (Ces.)G.W. Wilson causes the anthracnose leaf blight and stalk rot disease in maize (Zea mays L.), which is one of the top diseases affecting grain/silage production in maize-growing regions of the world, especially in the USA and Canada.From 2016 to 2019, the USA and Canada experienced a 0.21% annual yield reduction (3.03 million metric tons) due to anthracnose, resulting in a direct financial loss of USD 579.3 million [1].C. lentis Damm causes anthracnose disease in lentils, which is the most destructive disease of lentils in Canada.In epidemic years, grain yield loss can reach up to 70% [2], which is equivalent to a direct financial loss of around USD 300 per ha for a lentil producer.C. higginsianum Sacc.infects cruciferous crops, such as B. rapa ssp.pekinensis (Chinese cabbage), Brassica napus (rapeseed), B. rapa ssp.parachinensis (Chinese flowering cabbage) and B. rapa ssp.chinensis (bok-choi), and the model plant species Arabidopsis thaliana [3][4][5].
These Colletotrichum spp.employ a hemibiotrophic infection strategy (biotrophy-tonecrotrophy or sequential biotrophy) to infect and subsequently colonize the invaded tissues.Fungal infection is initiated by single-celled asexual spores, called conidia, which, upon landing on the plant surface (e.g., leaves, leaf sheathes, leaf whorls and/or stalk rinds), germinate to form regular or irregular dome-shaped appressoria.A narrow penetration peg emanating from the contact point of the appressorium with the plant surface pierces the epidermal cuticle and cell wall and elaborates the infection vesicle, which then expands to form thick biotrophic hyphae.In the case of C. lentis and C. higginsianum, the biotrophic hyphae (primary hyphae) are confined to the first invaded epidermal cells, and a complete switch to necrotrophy ensues thereafter, where thin necrotrophic hyphae differentiate from the biotrophic hyphae that kill and macerate the colonized tissues.However, C. graminicola exploits a unique hemibiotrophic infection strategy called sequential biotrophy [6], where the biotrophic hyphae of C. graminicola continuously proliferate well beyond the first infected epidermal cells and, therefore, the edges of infection courts remain biotrophic while the centers thereof become necrotrophic characterized secondary hyphae.Watersoaked lesions appear at the infection court within a few days of infection, and acervuli with black setae and conidia embedded in the salmon-pink-colored gelatinous matrix are produced in the anthracnose lesions within a week.These conidia are dispersed by splashing and blowing raindrops to infect neighboring healthy plants.
How these Colletotrichum spp.modulate their virulence/pathogenicity on their hosts have been the subject of numerous studies, which were primarily focused on pinpointing single genes located on core chromosomes through phenotyping of their knockout mutants on hosts.Until recently, delimiting the role of chromosomes was an arduous task, primarily concomitant with the lack of chromosome-level genome assemblies and functional analysis approaches.However, with the advent of second-and third-generation sequencing, and highly efficient genome assemblers, fungal genomes are being sequenced at the chromosome level.More recently, a gln-tRNA-based CRISPR/Cas9 minichromosome deletion system was developed, enabling the determination of the biological role of large genomic regions, including minichromosomes, by generating their knockout mutants and phenotyping them for fungal growth and development and virulence/pathogenicity on hosts [7].
Whole-genome sequencing of the three Colletotrichum spp.(C.graminicola, C. lentis and C. higginsianum) reveals that they carry ten core chromosomes (>3 Mb in length) and two to three minichromosomes (<1 Mb in length) [7][8][9][10][11].Molecular genetics and functional analyses reveal that one of the minichromosomes shows conditional dispensability, i.e., this minichromosome is dispensable for fungal growth and development but modulates fungal virulence on hosts [7,9,11].Here, we review the minichromosomes of three Colletotrichum spp.(C.graminicola, C. lentis and C. higginsianum), highlighting the genic and genomic diversity underlying the virulence/pathogenicity and adaptation of these pathogens on hosts.
T1-3-3-Chr12 (0.62 Mb) and M1.001-Chr12 (0.56 Mb) carry 31 and 43 genes, respectively; none are homologous genes.Eight and six, respectively, of the T1-3-3-Chr12 and M1.001-Chr12 are strain-specific (Supplementary Tables S3 and S4).The remaining genes in T1-3-3-Chr12 encode proteins of unknown functions except for GME10862, which codes for LPXTG domain-containing protein.The LPXTG domain, in conjunction with a C-terminal sorting signal, enables the anchoring of surface proteins in the bacterial cell walls [24]; however, the presence of LPXTG domain-containing proteins in fungal pathogens has not been reported, let alone their role in these organisms.GME10862 does not, however, express during vegetative growth and maize infection.The remaining genes in M1.001-Chr12 also code for proteins of unknown functions except for CGRA01v4_15056, CGRA01v4_15083, CGRA01v4_15084, CGRA01v4_15085, CGRA01v4_15095, CGRA01v4_15096 and CGRA01v4_ 15097, which encode sentrin/sumo-specific protease, linoleate diol synthase, chloroperoxidase, VID27 cytoplasmic protein, BRO1-like domain-containing protein and methyl-transferase type 11, respectively.Similar to Ulp1, sentrin/sumo-specific protease is a SUMO protease that deSUMOylates the substrate proteins by cleaving off SUMOs.Plant pathogenic ascomycete fungi secrete chloroperoxidases into hosts that catalyze the H 2 O 2dependent oxidation of Cl − to hypochlorous acid [25][26][27][28].A vanadium chloroperoxidase (MoVcpo) from M. oryzae induces reactive oxygen species (ROS; e.g., H 2 O 2 ) in rice cells and appears to act as a pathogen-associated molecular pattern.In addition, MoVcpo is required for conidiogenesis, conidial germination, cell wall integrity, osmotic stress tolerance and M. oryzae virulence on rice [29].BRO1-like domain-containing proteins are known to be involved in the fungal pH signaling pathway [30] and may be essential for the hemibiotrophic lifestyles of pathogens like M. oryzae and Colletotrichum spp., which require an alkaline host apoplastic environment during the biotrophic phase of infection and an acidic host environment during the necrotrophic phase of infection [31].Interestingly, only two genes (GME10839 and GME10865) encoding proteins of unknown functions in T1-3-3-Chr12 are induced during maize infection (log2fc ≥ 1.5, p < 0.01) [7].However, these genes do not regulate colony growth, conidiogenesis or C. graminicola virulence on maize.Six of the eight T1-3-3-specific genes are not expressed, whereas the remaining GME10852 and GME10863 are repressed during maize infection.Transcriptional repression of the genes in T1-3-3-Chr12 might be concomitant with the presence of four miRNAs (mi-598) (Supplementary Table S2).
A lack of effective molecular techniques to knock out fungal pathogen minichromosomes has hindered the delimitation of their biological functions.Until recently, targeted deletion of minichromosomes has not been attempted in any fungal species, including Colletotrichum spp., which are recalcitrant to genetic manipulations; this would have provided direct evidence of their role in fungal growth and development, and virulence/pathogenicity.Recently, we have developed a glutaminyl (gln)-tRNA-based CRISPR/Cas9 genome deletion system to functionally characterize the minichromosomes of T1-3-3 [7].The system involves two types of vectors: pCas9-Cg_tRp-sgRNA carrying 20 bp minichromosome-specific protospacers and pCE-Zero-Hpt carrying the selectable marker gene hygromycin phosphotransferase (Hpt) flanked by the minichromosome-specific homology arms of 1000 to 2000 bp.In pCas9-Cg_tRp-sgRNA, a T1-3-3-specific gln-tRNA promoter efficiently drives the expression of sgRNAs.Upon the PEG/CaCl 2 -mediated co-transformation in the T1-3-3 protoplasts, the pCas9-Cg_tRp-sgRNA vectors generate multiple simultaneous DNA double-strand breaks (DSBs) across a targeted genomic region within a minichromosome, followed by homology-directed repair of DSBs with the homology arms.Using this system, functionally nullisomic mutants of Chr11, Chr12 and Chr3 of T1-3-3 were obtained.The ∆Chr11, ∆Chr12 and ∆Chr13 mutants lack 618.61, 498.69 and 393.98 Kb genomic regions of their respective minichromosomes, which carry 32, 31 and 26 genes, respectively.Phenotypic analysis of these mutants showed that ∆Chr11, ∆Chr12 and ∆Chr13 were indistinguishable from T1-3-3 in colony growth and conidiation and that only ∆Chr12 was attenuated in virulence on the maize inbred line B73.This suggests that Chr12 is a conditionally dispensable minichromosome, which is dispensable for fungal growth and development but indispensable for fungal virulence on maize [7].
The reference genome of the C. lentis race 0 isolate CT-30 was sequenced using the second-generation Illumina sequencers (Illumina HiSeq 2000 and MiSeq), resulting in an assembly of 56.10 Mb distributed among 50 scaffolds.To assemble the genome at the chromosome level, an ascospore-derived population of 94 progeny isolates originating from a cross between CT-30, the race 1 isolate CT-21 and the parental isolates were resequenced on Illumina HiSeq 2000 at 19.08-fold genomic coverage.Mapping the resulting sequence reads onto the CT-30 genome yielded 14,132 high-quality single-nucleotide polymorphisms (SNPs), the most frequent genetic variations across the genomes.Genetic linkage mapping of these SNPs based on their genetic recombination resulted in 12 linkage groups, which were utilized to order and orientate 50 scaffolds into 12 chromosomes.Ten of the chromosomes are core chromosomes and >3 Mb in length, whereas two are minichromosomes (Chr11 and Chr12) [9].Unlike minichromosomes, Chr11 is unusually large (1.52 Mb); Chr12 is 0.39 Mb in length.Unlike C. graminicola, the minichromosomes in C. lentis are less affected by RIPMS, e.g., Chr11 (9.21%) and Chr12 (8.02%); therefore, they carry higher GC contents, e.g., Chr11 (37.57%) and Chr12 (37.26%) (Table 1; Supplementary Tables S7 and S8).
Unlike C. graminicola, the C. lentis minichromosomes carry virulence-/pathogenicityrelated genes, such as those coding for effectors, CAZymes and secondary metabolite enzymes.Chr11 carries 164 genes, 39 of which are unique to CT-30.Thirty-eight of the Chr11 genes code for proteins of unknown functions, whereas the remaining eighty-seven genes encode proteins, including nine effector candidates and four secondary metabolite backbone synthesis enzymes (SMBSEs).The C. lentis genome carries four genes, which code for NUDIX hydrolase domain-containing proteins.Of these, three were located on core chromosomes (scaffold1-552 (Chr9), scaffold12-9 and scaffold12-65 (Chr7)) and one (scaffold14-4) on minichromosome Chr11.scaffold12-65 is induced exclusively at the hemibiotrophic switch, where the necrotrophic hyphae start emanating from the biotrophic.Scaffold12-65 (aka ClNUDIX) induces a hypersensitive cell death response (HR) in Nicotiana tabacum.The C. lentis and M. oryzae strains overexpressing ClNUDIX trigger HR in the lentil cultivar Eston and the barley cultivar CDC silky, respectively.ClNUDIX initially accumulates along the cell periphery and later is taken up in endocytosis pits or early endosomes.The effector may hydrolyze the pyrophosphate bonds of inositol pyrophosphates (IP6 and IP7).The NUDIX motif within the NUDIX hydrolase domain specifically binds to IP6 and IP7, which in turn tether to clathrin-associated proteins, such as adaptor protein 2, a constituent of clathrin protein-coated endocytic pits and early endosomes that originate from the endocytic pits.The hydrolysis of inositol pyrophosphates by NUDIX hydrolase may dismantle the endocytosis pit or the early endosomes, resulting in the loss of plasma membrane integrity and causing non-specific HR due to the influx of extracellular proteins into cytoplasm.Since HR cell death coincides with the hemibiotrophic switch phase, it signals the pathogen to initiate the anthracnose-causing necrotrophic phase [32].Like scaffold12-65, scaffold14-4 is also exclusively induced during 48 hpi (log2fc 5.0, p < 0.01), corresponding to the hemibiotrophic switch phase of infection.These NUDIX hydrolase genes might have redundant functions in C. lentis.Silencing the NUDIX domains of these effectors may provide evidence for their direct involvement in the hemibiotrophic switch.
Chr11 contains 133 genes, 98 of which code for hypothetical proteins, whereas the remaining 35 encode proteins of known functions, including 7 effector candidates and 1 each of CAZyme (glycosyl hydrolase family 92) and SMBSE (PKS-NRPS hybrid enzyme).One of the effector candidate genes, CH63R_14535, encodes a LysM domain-containing protein.Fungal LysM domain-containing proteins are small secreted proteins that bind to chitin oligosaccharides during plant-pathogen interactions, thereby preventing chitintriggered immunity [35,36].
In a forward genetic screen for C. higginsianum virulence on A. thaliana via Agrobacterium tumefaciens-mediated transformation (ATMT)-based insertional mutagenesis, Plaumann and colleagues [11] identified two T-DNA insertion mutants, vir-49 and vir-51, that lacked Chr11.Both mutants exhibited a decreased level of virulence on A. thaliana Col-0 due to an inefficient shift from biotrophy to necrotrophy, evidenced by less than 3% of the primary hyphae generating secondary hyphae in the infected cells.However, the lack of the chromosome did not impact vegetative fungal growth or conidiation, indicating that Chr11 is a conditionally dispensable minichromosome.

Perspectives
With the inception of third-generation sequencing and the subsequent correction of the resulting long reads using second-generation sequencing, it has become possible to generate gapless fungal genome assemblies at the chromosome level, enabling the identification of minichromosomes.Minichromosomes in the hemibiotrophic Colletotrichum pathogens appear to be strain-specific innovations as the strains within species, let alone between species, do not share minichromosomes.In addition, they are the gene-sparse regions of the genomes and are not enriched with genes known to be implicated in fungal virulence/pathogenicity on hosts, including but not limited to effectors, CAZymes and secondary metabolism genes.The minichromosomes, especially in the C. graminicola strains, are highly compartmentalized into AT-rich and GC-rich blocks, with the former occupying the majority of the chromosomes, and hence contain a higher content of repetitive DNA elements and lower gene contents.The three Colletotrichum pathogens that we discussed carry a single conditionally dispensable minichromosome, which, especially in C. graminicola and C. lentis, are transcriptionally less active than the core chromosomes.The limited number of transcriptionally active genes in the conditionally dispensable minichromosomes Cg-Chr12 and Cl-Chr12 located in the GC-rich blocks that are more protected from erosion by RIPMs than Ch-Chr11 might be an adaptation, restricting C. graminicola and C. lentis to infect a single host species, as opposed C. higginsianum, which can invade multiple species of crucifers.It would be interesting to assess the role of minichromosomes in the adaptation to various stresses, e.g., osmotic, oxidative and salt stresses, and cell wall damage, enabling fungal pathogens to invade and successfully colonize host species.

Author
Contributions: V.B. designed the research; V.B., M.Z. and W.M. analyzed data; and V.B., Y.-L.P., W.Z. and J.Y. wrote the paper.All authors have read and agreed to the published version of the manuscript.Funding: This research was made possible through the generous support of the National Natural Science Foundation of China (Grant No. 32172363) and the Chinese Universities Scientific Fund (Grant No. 10092004).

Table 1 .
General features of minichromosomes in four strains of three hemibiotrophic Colletotrichum pathogens and features thereof.