The LmSNF1 Gene Is Required for Pathogenicity in the Canola Blackleg Pathogen Leptosphaeria maculans

Leptosphaeria maculans is a fungal pathogen causing blackleg in canola. Its virulence has been attributed, among other factors, to the activity of hydrolytic cell wall degrading enzymes (CWDEs). Studies on the pathogenicity function of CWDEs in plant pathogenic fungi have been difficult due to gene redundancy. In microorganisms many CWDE genes are repressed by glucose and derepressed by the function of the sucrose non-fermenting protein kinase 1 gene (SNF1). To address the molecular function of SNF1 in L. maculans, the ortholog of SNF1 (LmSNF1) was cloned and functionally characterized using a gene knockout strategy. Growth of the LmSNF1 knockout strains was severely disrupted, as was sporulation, spore germination and the ability to attach on the plant surface. When inoculated on canola cotyledons, the LmSNF1 knockout strains could not cause any symptoms, indicating the loss of pathogenicity. The expression of 11 selected CWDE genes and a pathogenicity gene (LopB) was significantly down-regulated in the LmSNF1 knockout strains. In conclusion, knockout of LmSNF1 prevents L. maculans from properly derepressing the production of CWDEs, compromises the utilization of certain carbon sources, and impairs fungal pathogenicity on canola.


Introduction
Plant pathogenic fungi secrete an array of cell wall degrading enzymes (CWDEs) capable of depolymerizing the polysaccharides of primary cell walls [1][2][3].These enzymes received special attention from researchers and many of them have been demonstrated to be important for the pathogenicity of various fungi.However, as a result of the redundancy of genes and enzymatic activity, reverse genetic approaches that rely on single gene deletion or silencing have not effectively addressed the importance and function of these enzymes in pathogenicity [4][5][6][7].
In many microorganisms including fungi, CWDEs are subject to catabolite repression, a mechanism that controls the preferential use of easily fermentable carbon sources, such as glucose, by repressing genes that are used to metabolize other carbon sources, such as sucrose, galactose, pectin, and xylose [8,9].To test the role of CWDEs as a whole in fungal pathogenicity, disruption of the elements that control the derepression mechanism could be a more reliable approach than working on single CWDE genes [10].In Saccharomyces cerevisiae, the SNF1 (sucrose non-fermenting 1) gene has been shown to play a central role in carbon catabolite repression [11].SNF1 encodes a protein kinase (Snf1) that phosphorylates Mig1, a DNA-binding transcriptional repressor responsible for catabolite repression [12], leading to the derepression of all genes under Mig1 control.The ortholog of Mig1 in filamentous fungi is called creA [8].Repression of CWDE genes by creA has been studied in a number of plant pathogenic fungi [13][14][15][16].
SNF1 orthologs have been studied in plant pathogenic fungi, including Cochliobolus carbonum [10], Fusarium oxysporum [17], Sclerotinia sclerotiorum [18], Magnaporthe oryzae [19], F. graminearum [20], Alternaria brassicicola [13], Verticillium dahliae [21] and Penicillium digitatum [22].In most of these studies, disruption of SNF1 resulted in a reduction in the ability of the corresponding species to grow on certain carbon sources and in transcription of many CWDE genes.Components of fungal pathogenicity such as sporulation, spore germination and appressorium formation were impaired in the gene knockout strains and, consequently, the virulence on the host plants was reduced.
Blackleg of canola (Brassica napus), caused by the Dothideomycete Leptosphaeria maculans and L. biglobosa, is an important disease in many countries [23,24] including Canada, Australia and most of Europe.Compared to L. biglobosa, L. maculans is more aggressive [25].Worldwide, the two species coexist in many regions, with L. maculans typically causing cortical infection near the base of the stem and L. biglobosa resulting in more superficial stem lesions or pith damage [26].In China, although blackleg is prevalent, it is caused only by L. biglobosa [25].In western Canada, both species may be found in association with basal-stem canker symptoms, with L. biglobosa identified at lower frequencies [27].
In spite of the economic importance of blackleg disease, the pathogens involved have not been studied in great detail at the molecular level until recently.Four avirulence genes, AvrLm1 [28], AvrLm6 [29], AvrLm4-7 [30] and AvrLmJ1 [31] have been cloned and characterized.A few protein-encoding genes playing a role during pathogenesis also have been functionally studied, including an isocitrate lyase gene [32], the lopB gene with unknown function [33], the THIOL gene encoding a 3-Ketoacyl-CoA thiolase [34], the Ipa gene [35], the Lmpma1 gene encoding a plasma membrane H + -ATPase isoform [36], the Lmgpi15 gene encoding a component of the glycosylphosphatidylinositol anchor biosynthesis pathway [37], the LmIFRD gene [38], and the Lmepi gene encoding a key enzyme of the Leloir pathway [39].In most of these reports, gene disruption mutants were studied and the disruption was achieved exclusively by the method of Agrobacterium tumefaciens-mediated random insertional mutagenesis.
The recent release of the whole genome sequence of L. maculans [40] laid a foundation for genomic studies of the blackleg pathosystem.While global studies of genes involved in this hostpathogen interaction have become easier, functional studies of pathogenicity genes by targeted gene knockout also are much facilitated.However, it is has been reported that in L. maculans at least 7 kb of flanking DNA is required for homologous recombination, an innate mechanism on which targeted gene knockout relies, and up to several hundred transformants need to be screened to achieve gene replacement [41].To date, characterization of pathogenicity genes by targeted gene knockout has only been reported once in this fungal species [42].Thus in the present study, we report results from a targeted gene knockout of the L. maculans SNF1 gene and the subsequent characterization of the gene knockout stains.We demonstrate that SNF1 is required for expression of genes involved in cell wall degradation and for spore germination, spore attachment and pathogenicity on canola plants.

Ethics statement
No specific permission was required for the field from which the L. maculans wild-type was derived.All of the field studies were carried out in a closed and protected green house or a growth chamber in Crop Diversification Centre North.This study did not involve endangered or protected species.

Chemicals and standard techniques
All chemicals were purchased from Fisher Scientific (Ottawa, ON, Canada) unless otherwise specified.Polymerase chain reactions (PCR) were conducted in an Eppendorf Mastercycler thermal cycler (Eppendorf, Hamburg, Germany) using the PCR master mix from Promega (Madison, WI, USA).Quantitative PCR (qPCR) assays were conducted in a StepOnePlus Real-Time PCR System (Life Technologies, Carlsbad, CA, USA) using the GoTaq 2-Step RT-qPCR System (Promega) or the Power SYBR Green PCR Master Mix (Life Technologies).qPCR primers were designed with the online program Primer3web (http://primer3.ut.ee).DNA purification of amplicons from the reaction mixture or agarose gel was performed using a Wizard SV Gel and PCR Clean-Up System (Promega).Similarity searches of nucleic acid and protein sequences were performed using the Basic Local Alignment Search Tool (BLAST) against the National Center for Biotechnology Information (NCBI) database unless otherwise specified.Sequencing was done by the Department of Biological Sciences, University of Alberta (Edmonton, AB, Canada).Other molecular techniques, if not specified, were performed according to the protocols described by Sambrook and Russell [43].

Sequence retrieving and analyses
Using the sequence of the open reading frame (ORF) of the SNF1 gene from C. carbonum ( [10]; GenBank accession number: AAD43341) as a query, a BLAST search was conducted against the L. maculans whole genome database (http://genome.jgi-psf.org/Lepmu1).This search produced a single hit which are referred to as LmSNF1 hereafter in this report.The DNA sequence, including the ORF of LmSNF1 and 3,500 bp of its 59 upstream and 39 downstream regions, was retrieved and served as the basis for the design of primers for PCR and qPCR analyses (Figure 1, Table S1).Three PCR fragments covering the ORF were amplified from fungal genomic DNA using the primer pairs g/p, i/q and j/r (Figure 1, Table S1) and sequenced using the same primers.
Orthologs of Snf1 from other fungi were identified by searching each of the accessible fungal genomes hosted at the Broad Institute (http://www.broadinstitute.org)Fungal Genome Database by BLAST with the deduced amino acid sequence of LmSNF1 (referred to as LmSnf1) as the query.The amino acid sequence of the best hit from each fungal species was retrieved.
Twelve CWDE genes were selected from the L. maculans carbohydrate-active enzymes (CAZy) database (http://www.cazy.org/e1438.html).Three other protein-encoding genes that have been shown to be pathogenicity-related were selected from the pathogen host interactions (PHI) database (http://www.phi-base.org).Based on the mRNA sequence for each of these genes, a pair of qPCR primers was designed (Table S2).
Molecular weight and the isoelectric point (pI) of LmSnf1 were calculated using the online compute pI/Mw tool (http://web.expasy.org/compute_pi).Multiple alignments between LmSnf1 and its orthologs and the subsequent construction of a phylogenetic tree were performed with ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2).

Fungal and plant materials
A blackleg susceptible canola cultivar Westar was used as the host for L. maculans.The L. maculans strain S4, isolated from an infected canola plant in a field (GPS coordinates: 53.578, 2 113.222) in Strathcona County, Alberta, Canada, in 2012 was used as the wild-type and is referred to as WT hereafter.The WT and transformants were maintained on 4% potato dextrose agar (PDA) or V8 agar (150 ml V8 juice, 1.5 g CaCO 3 and 15 g agar for 1 L) in Petri dishes.For cultures to be used in quantitative assays, 25 ml of medium was placed in each 10-cm diameter Petri dish with a peristaltic pump (Integra Biosciences, Hudson, NH, USA).

Plant inoculation
Seeds of 'Westar' were sown in 50-ml plastic cups filled with Sunshine mix #4 soil (Sun Gro Horticulture, Vancouver, BC, Canada).The seedlings were maintained in a growth chamber with a 16 h photoperiod at 22uC day/18uC night.To prepare the inoculum, 14-day-old sporulating cultures of the fungus on V8 agar were flooded with sterile distilled water and the resultant conidial suspension was passed through eight layers of cheesecloth.Before inoculation, the paraxial surface of the cotyledons of 10-day old seedlings was wounded with a pair of forceps.The wounded sites were inoculated with a 5-ml droplet of the conidial suspension.After the inoculum was allowed to air dry, the plants were incubated in transparent plastic bags to maintain high relative humidity.After 48 hours, bags were removed and the plants were maintained in the growth chamber.

Nucleic acid manipulations
Genomic DNA was extracted from fungal mycelia or infected plant tissue using a Qiagen DNeasy Plant Mini Kit (Qiagen, Toronto, ON, Canada) unless otherwise specified.Total RNA was isolated from frozen mycelium or inoculated plant tissue with an RNeasy Plant Mini Kit (Qiagen).The RNA samples were further treated with RNase-Free DNase (Qiagen) to remove DNA contaminants.The purified RNA was used for synthesis of the first stranded cDNA using an iScrip cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA) or a Promega GoTaq 2-Step RT-qPCR System (Promega).

Transcript induction
For in vitro induction of transcripts, mycelium of WT generated after 48 h incubation in YPG broth (3 g yeast extract, 10 g peptone and 20 g glucose for 1 L) was transferred to fresh YPG broth, Czapek-Dox minimal medium (referred to as CD hereafter; 2 g NaNO3, 0.5 g KCl, 1 g K 2 HPO 4 , 0.5 g MgSO 4 ?7H 2 O and 0.01 g FeSO 4 ?7H 2 O for 1 L) or CD supplemented with 15% V8 juice or 1% glucose or pectin as the sole carbon source and incubated with agitation for another 24 h.The mycelia were harvested and subjected to RNA extraction as above.

Quantitative PCR
For qPCR analysis, the L. maculans housekeeping gene encoding ACT1 (GenBank accession number: FJ172241) was used as an endogenous control and amplified with the primers ActinF: AGTGCGATGTCGATGTCAG and ActinR: AAGAGCGGT-GATTTCCTTCT [38].Primers for LmSNF1 were the primer pair i/p (Figure 1, Table S1) and those for the selected CWDE and other pathogenicity-related genes are listed in Table S2.
The qPCR was conducted in a 15-ml reaction mixture containing 1.25 ml each of 5 mM forward and reverse primer and 200 ng of the cDNA template.The amplification conditions consisted of an initial denaturation step of 95uC for 10 min, followed by 40 cycles of 15 s at 95uC and 1 min at 60uC.After completion of the PCR amplification, a melting curve analysis was run to evaluate the amplification specificity.Primer efficiency was verified using standard curves generated from the wild-type genomic DNA.Expression of genes was calculated via the deltadelta method, with relative expression = 2 2 (DCt sample-DCt control) [44].All qPCR amplifications were performed in triplicate with each of three cDNA replicates derived from different preparations of mycelium or plant tissue.

Construction of transforming DNA for gene knockout
Two sets of transforming DNA were constructed by fusion PCR [45] and each set was used in fungal transformation.To generate the first construct, a 3.2 kb cassette containing the Escherichia coli hygromycin B phosphotransferase gene (HPH) that has been rendered suitable for fungal expression was amplified from the vector pGC1-1 [46] using the primer pair y1/z3 (Figure 1, Table S1).Two genomic fragments at 1.1 kb and 0.3 kb were amplified from the 59 and 39 flanking regions of LmSNF1 using the primer pairs d/o and k/t, respectively.The 59 ends of the primers o and k were tailed with the reverse-complemented sequences of primers y1 and z3, respectively.The three amplified fragments were mixed at a molar ratio of 1:1:1 and used as a template to amplify a 4.6 kb fusion fragment with the primer pair e/s.The resulting fragment was cloned into the pGEM T-Easy vector (Promega) by TA cloning.
For the second construct, two fusion fragments were generated by a split-marker strategy [47].To generate the first fusion fragment, a 2.3 kb fragment was amplified from pGC1-1 using the primer pair y1/z2 and a 3.4 kb fragment was amplified from the 59 flanking region of LmSNF1 with the primer pair a/o.These two fragments were mixed at a molar ratio of 1:1 and used as template to amplify a 5.7 kb fusion fragment with the primer pair b/z1.For the second fusion fragment, a 1.7 kb fragment was amplified from pGC1-1 using the primer pair y2/z3 and a 3.3 kb fragment was amplified from the 39 region of LmSNF1 using the primer pair k/x.Primer k was tailed at the 59 end with the reverse-complemented sequence of primer z3.These two fragments were mixed at a molar ratio of 1:1 and used as a template to amplify a 5.0 kb fusion fragment using the primer pair y3/w.The 5.7 and 5.0-kb fusion fragments were cloned into the pGEM T-Easy vector by TA cloning.

Fungal transformation
PEG-mediated protoplast transformation followed the protocol of Feng et al. [48].Each transformation experiment was conducted in triplicate in 1.5-ml tubes each containing 3610 7  protoplasts in a volume of 300 ml and supplemented with 30 ml of transforming DNA.The DNA consisted of 3 mg of the first construct or 2 mg of each of the two fragments of the second construct, all of which were generated by NotI digestion of the corresponding plasmid DNA followed by gel purification.The hygromycin B-resistant colonies that grew on the selective media were transferred to PDA plates containing 100 mg/ml of hygromycin B for a second round of screening.Colonies identified in the second round of screening were regarded as successful transformants.Unless otherwise specified, growth medium for the transformants was always supplemented with 100 mg/ml of hygromycin B.

Confirmation of gene knockout
Genomic DNA of WT and all transformants was extracted from mycelium generated on PDA plates according to the method described by Feng et al. [49].Based on these DNA samples, transformants were screened for a gene replacement event by PCR analysis using primers specific to HPH, the LmSNF1 ORF or the genes located on the 59 and 39 regions of LmSNF1.
Based on the PCR results, all of the putative gene knockout (Ko) strains, one ectopic insertion strain (Ect) and WT were subjected to Southern analysis using a DIG DNA Labeling and Detection kit (Roche Diagnostics, Laval, QC, Canada).All strains were grown in YPG broth for two days and genomic DNA was extracted as per the phenol protocol [43].Twenty mg of genomic DNA from each strain was digested with EcoRI and fractionated on a 1% agarose gel.A 1.7-kb fragment amplified from WT using the primer pair j/ u was used as the probe.
Gene knockout events were further confirmed by qPCR analysis.Total RNA was extracted from healthy canola, WT, Ect and the six Ko strains after transcript induction in CD medium for 24 hours.Transcript accumulation of LmSNF1 and its 59 and 39 genes was assessed.

Growth assay
Growth of WT and the transformants was measured on Petri dishes containing CD medium supplemented with 15% V8 juice or 1% glucose or pectin as the sole carbon source (all media without hygromycin B).Briefly, culture plugs (0.5 cm in diameter) were excised from the margins of a colony growing on PDA and were transferred to the center of freshly-prepared plates.The plates were incubated under continuous light at 25uC and colony diameter was measured in two directions at right angles to each other after 6 days.The average colony radius was calculated and used to represent fungal growth.This experiment was conducted as a randomized complete block design with three replicates and repeated with similar results obtained.
To measure mycelial growth in liquid media, three plugs (0.5 cm in diameter) were inoculated into 50 ml of medium in a 125-ml flask.The media used were same as the plate assay but without agar.The flasks were shaken at 150 rpm at 25uC for 6 days.The mycelial mats were vacuum-filtered through two layers of filter paper and lyophilized.The mass of the lyophilized mycelium was recorded.This experiment was conducted as a randomized complete block design with three replicates and repeated with similar results obtained.

Sporulation assay
The ability of the transformants to produce conidia was assessed by induction of sporulation on plates containing 25 ml V8-agar without hygromycin B under continuous light at 25uC.After 10 days, conidia formed on the plate were harvested in 100 ml H 2 O and filtered through eight layers of cheesecloth.The conidial concentration in the suspension was measured with a haemocytometer and expressed as the number of conidia per plate or per square millimeter of the V8-agar culture.This experiment was conducted as a randomized complete block design with three replicates and repeated with similar results obtained.

Germination assay
Ten ml of 0.8% PDA at 45uC was mixed with 100 ml of conidial suspension at a concentration of 2610 7 conidia/ml.The mixture was poured into a Petri dish and kept at 25uC in the dark.After 24 hours, a small piece was excised and transferred to a haemocytometer and covered with a cover slide.The cover slide was pressed with one click of a retractable pilot pen to distribute the mixture evenly in the counting chamber.Under a Zeiss AXIO microscope (Carl Zeiss, Thornwood, NY, USA), the total number of conidia and number of germinated conidia were counted within an area of 0.36 mm 2 (nine 0.260.2mm cells) in the center of the counting chamber.For each strain, three plates (repeats) from different conidial suspensions were prepared and from each plate 10 samples were counted.This experiment was repeated with similar results obtained.

Attachment assay
A conidial suspension from each strain was prepared and adjusted to a concentration of 1610 7 conidia/ml.Two 10-ml droplets of the suspension were applied to the surface of a cotyledon excised from a 10-day old 'Westar' seedling.The cotyledons were placed on water-soaked filter paper in Petri dishes for two hours to let the droplets dry.The Petri dishes were sealed and kept in darkness at room temperature.At 6 and 20 hai, the attachment of conidia to the surface of the cotyledon was assessed by counting the number of washed-off conidia after washing.Briefly, one inoculated cotyledon was transferred into a 2-ml tube containing 1 ml washing buffer (0.02% Tween 20).The tubes were shaken at 1,400 rpm for 30 seconds with an Eppendorf Thermomixer.An 800 ml aliquot of the suspension was transferred into a 1.5 ml tube.After centrifugation at 14,000 rpm for 2 min, 700 ml of the supernatant was removed and the remaining 100 ml was used to quantify the spores.For each sample, two counts were made in a 0.04 mm 2 area (one 0.260.2mm cell at the center of the counting chamber) and 10 samples were examined for each tube.The average of the 20 counts was used to calculate the attachment ratio, which was regarded as the data from one experiment unit.For each strain, the attachment was assessed on three cotyledons (repeats).

Pathogenicity test
Inoculation of 'Westar' was conducted as described above.On each cotyledon, inoculation of a Ko or the Ect strain was accompanied by inoculation with WT side by side.For each strain, 50 cotyledons on approximately 30 plants in five pots were inoculated with conidial suspension at a concentration of 1610 6 conidia/ml or 1610 7 conidia/ml.Plants were examined daily from 4 to 10 days after inoculation (dai) for the development of blackleg symptoms.

Statistical analysis
Data were analyzed for statistical significance using the general linear model (GLM) procedure of the Statistical Analysis System (SAS Institute, Cary, NC, USA).All of the data were subjected to one-way analysis of variance and, when appropriate, a multiplerange test was conducted by Fisher's least significant difference (LSD) at P#0.01.

LmSNF1 shares high sequence similarity with its orthologs
Sequencing of the LmSNF1 PCR fragments revealed an ORF of 2,782 bp, which is identical to the SNF1 gene (XM_003844721) in the whole-genome sequenced L. maculans strain JN3.The ORF is interrupted by two introns of 59 and 53 bp.It encodes a polypeptide of 889 amino acids with a calculated molecular weight of 99 kDa and a pI of 8.8.
LmSnf1 shares 82%, 80% and 79% overall identities with its orthologs in Stagonospora nodorum (AAR02440), C. carbonum (AAD43341) and Pyrenophora tritici-repentis (XP_001932670), respectively.Multiple alignment of LmSnf1 with its orthologs in other fungi indicated that their sequence homology is higher in the Nterminal kinase domain and lower in the C-terminal regulatory domain (data not shown).A phylogenetic tree was constructed based on the alignment.Comparison of this tree with the phylogenomic tree illustrated at the Broad Institute Fungal Genome database, which was constructed using more than 70 genes that are conserved among all fungal species (Dr.Li-Jun Ma, Broad Institute, personal communication), revealed a high consistence between the two trees, especially among the species in Ascomycota (Figure S1).

Expression of LmSNF1 is induced by pectin and during pathogenesis
In YPG, CD and CD supplemented with glucose, expression of LmSNF1 was at lower levels than the housekeeping gene ACT1.Expression of LmSNF1 increased in V8 and pectin media, in which it was similar to ACT1.Compared to the glucose medium, expression in V8 and pectin media was significantly higher (Figure 2A).
In planta expression of LmSNF1 was investigated over a 0-144 hai time course.No transcript of either LmSNF1 or ACT1 could be detected at 0 and 12 hai.From 24 to 96 hai, expression of LmSNF1 increased and then decreased, with an expression peak at 48 hai.From 96 to 144 hai, expression remained constant but at a relative low level (Figure 2B).This result indicates that the activation of LmSNF1 expression is concurrent with an early stage of infection by the fungus.

LmSNF1 was knocked out from the fungal genome
Transformations using the first and second constructs produced 21 and 5 transformants, respectively.PCR screening for the absence of LmSNF1 revealed that 2 out of 21 strains transformed with the first construct and four out of five strains transformed with the second construct were putative gene knockout strains (Figure S2).These strains were named Ko1 and Ko2 (with the first construct) and Ko3 to Ko6 (with the second construct).The ectopic insertion strain (Ect) transformed with the second construct, in which HPH was integrated into the genome but LmSNF1 remained unchanged, was used as a control in addition to WT.
The gene knockout event was confirmed by Southern (Figure 3A) and qPCR analyses.A hybridization signal was observed in WT at 4.0 kb, which is the position of the band produced by digestion with EcoRI at two sites, one within the LmSNF1 ORF and another at the 39 region (Figure 1).In Ko1-Ko6, signals were detected at 3.8 kb, which would have resulted from the same cutting site in the 39 region and another cutting site within the HPH ORF (Figure 1).From Ect, hybridization signals were observed at both 4.0 and 3.8 kb, indicating the presence of both HPH construct and LmSNF1.Against the cDNA, amplification of LmSNF1 could be detected by qPCR analysis from WT and Ect, but not from the Ko strains.These results indicate that the six strains (Ko1-6) are true gene knockout strains.
To confirm that knockout of LmSNF1 did not affect its 59 and 39 genes, a qPCR analysis was conducted with cDNA from YPG cultures of WT, Ect, Ko2 and Ko4.No significant differences in the expression of either the 59 or the 39 gene could be detected among these four strains (Figure 3B).

Knockout of LmSNF1 affects vegetative growth and sporulation
After 6 days of incubation on CD supplemented with V8 juice, glucose or pectin, all of the six gene knockout strains exhibited slower growth relative to WT and Ect (Table 1).This became even more evident after 11 days of incubation (Figure S3).Similarly, in all of the three liquid media, the Ko strains produced less dry  weight than WT and Ect after 6 days of incubation (Table 1).These results indicate that the vegetative growth of L. maculans is impaired as a result of the loss of LmSNF1.
After 10 days of growth on V8-agar, abundant sporulation was evident for all strains.Assessment of the number of conidia per plate indicated that all Ko strains produced less conidia than WT and Ect.Among the Ko strains, Ko2 and Ko4 produced the most conidia (Figure 4A).When sporulation was assessed as the number of conidia per area of colony, Ko1, Ko3, Ko5 and Ko6 produced fewer conidia than WT, while Ko2 and Ko4 did not exhibit a significant decrease in sporulation (Figure 4B).These data suggest that, at least in the strains Ko1, Ko3, Ko5 and Ko6, knockout of LmSNF1 impairs fungal sporulation.

Knockout of LmSNF1 affects conidial germination and attachment on leaf surface
Conidial germination of all strains was investigated in 0.8% PDA (Figure S4).After 24 hours, 87.7% and 88.2% of conidia from WT and Ect, respectively, had germinated, percentages which were significantly higher than for conidia of the six Ko stains.Within the Ko strains, Ko4 and Ko6 showed a higher germination rate than the others (Figure 5).
Conidial attachment to the plant surface was assessed at 6 and 20 hai.Compared to WT and Ect, the strains Ko1, Ko2, Ko5 and Ko6 showed significantly impaired conidial attachment rates at 6 hai (Figure 6A).At 20 hai, Ko1, Ko3 and Ko6 showed reduced attachment (Figure 6B).Although the reduction in conidial attachment was not significant for some Ko strains at either of the time points, a trend could be observed in the two experiments, in which knockout of LmSNF1 resulted in an impairment in the percentage of conidia that could attach to the plant surface.

Knockout of LmSNF1 causes down-regulation of cell wall degrading enzyme genes
Expression of 12 CWDE genes and three pathogenicity genes was investigated in WT after incubation in CD supplemented with 1% glucose for 48 hours or in canola tissue at 4 dai.For eight of the 15 genes, expression could not be detected in glucose medium.By contrast, expression could be detected for all the 15 genes in Table 1.Growth of the Leptosphaeria maculans wild-type, one ectopic insertion strain and six LmSNF1 knockout strains on solid and in liquid media.planta at 4 dai.In planta up-regulation was evident for 14 genes (Table S3).In CD medium containing pectin, 11 CWDE genes were shown to be down-regulated in the two Ko strains compared to in WT (Table 2).Expression of one gene encoding a chitin deacetylase was at similar level in WT and Ko strains.Among the three selected pathogenicity genes, two were expressed at similar levels across the strains.The expression of the gene encoding the pathogenicity protein LopB was diminished in the Ko strains, indicating that this gene is under control of LmSNF1.

Knockout of LmSNF1 causes loss of fungal pathogenicity
The pathogenicity of all Ko strains and Ect were assessed on cotyledons of canola.After inoculation, blackleg symptom development was monitored from 4 to 7 dai.For WT and Ect, blackleg lesions started to become evident at 4 dai and enlarged from that point on (Figure S5).At 7 dai, all cotyledons inoculated with WT and Ect strains developed typical blackleg lesions (Figure 7A).By contrast, none of the six Ko strains caused any lesions (Figure 7B).Similar results were observed after inoculations with two concentrations of conidia, 1610 5 or 1610 6 conidia/ml.This indicates that knockout of LmSNF1 renders the fungus nonpathogenic on canola.

Discussion
Targeted gene knockout in L. maculans Successful introduction of foreign DNA into a fungal genome is dependent on the DNA double-strand break (DSB) repair mechanisms.Eukaryotes have two main DSB repair pathways: homologous recombination and nonhomologous end-joining (NHEJ) [50].A gene replacement event requires the occurrence of DSB in the vicinity of the targeted gene and the utilization of homologous recombination to repair the DSB.However, although some species such as S. cerevisiae use the homologous recombination system more frequently, many filamentous fungi seem to prefer NHEJ [51].
Targeted gene knockout has been reported to be difficult in L. maculans: flanking sequences greater than 7 kb are required to allow the occurrence of homologous recombination and hundreds of transformants need to be screened [41,42].In the present study, the first construct contained 59 (1.1-kb) and 39 (0.3-kb) geneflanking sequences.Transformation using this construct produced 21 transformants and two of which were shown to be gene replacement strains.The gene replacement efficiency (2/ 21 = 9.5%) is comparable with that observed in other fungi when ,1 kb flanking sequences from both the 59 and 39 regions were used, for example 13.0% in F. graminearum [6] and 14.8% in S. nodorum [48].Wilson et al. [42] obtained two L. maculans gene replacement strains from 23 transformants by using a construct carrying 3.7-kb 59 and 3.2-kb 39 homologous sequences.These authors also used a construct carrying 1.4-kb 59 and 1.5-kb 39 homologous sequences but failed to generate any gene replacement strains, albeit 89 transformants were obtained.Based on the gene replacement efficiencies observed in the present study and by Wilson et al. [42], it can be inferred that the NHEJ DSB repair mechanism is preferentially used in L. maculans relative to other filamentous fungi, and that gene replacement in this species maybe not as difficult as previously reported.
In the present study, we also used a split-marker gene replacement strategy to construct the transforming DNA.Previously, it was reported that the efficiency of gene replacement in other fungi could be enhanced by employing a split-marker approach where regions flanking the gene of interest are fused to overlapping partial segments of a selectable marker [47,52].This was supported by the present study: five transformants were obtained and four of them were LmSNF1 replacement strains.

Expression of LmSNF1 in L. maculans
The expression of LmSNF1 was similar in CD medium with or without glucose.This is consistent with earlier reports on the expression of SNF1 in S. sclerotiorum [18] and F. graminearum [20], in which the expression was independent of the presence of glucose.By contrast, the expression of LmSNF1 was higher in pectin and V8 media than in CD and CD with glucose.This also is consistent with what was observed with respect to the expression of SNF1 in Colletotrichum gloeosporoides f. sp.malvae [53], in which expression was consistently higher in glycerol than in glucose cultures, but differed from a report on S. sclerotiorum [18], in which SNF1 expression was apparently constant in media containing either glucose or other carbon sources, including pectin.Based on the expression data obtained in the present study, it seems that some of the components found in V8 juice, including pectin, are responsible for the LmSNF1 up-regulation.
The up-regulation of LmSNF1 observed at 48 hai is consistent with reports of C. gloeosporoides f. sp.malvae [53] and S. sclerotiorum [18].In both species, expression of SNF1 peaked during the early stages of infection and then remained constant at much lower levels.At 48 hai, most of the L. maculans spores had germinated and the fungus started to invade the plant by extensive hyphal growth.Thus, it appears that LmSNF1 expression is related to the phase of growth rather than to the physiological conditions encountered during plant invasion.

Expression of CWDE and pathogenicity genes in the LmSNF1 Ko strains
The 12 CWDE genes were chosen from the CAZy database, which hosts four groups of enzymes: glycoside hydrolases (GHs), Figure 6.Conidial attachment on canola cotyledons of Leptosphaeria maculans strains.WT, the wild-type; Ect, the ectopic insertion strain; Ko, LmSNF1 knockout strains.Two 10-ml droplets of a spore suspension (1610 7 conidia/ml) were applied on the surface of one cotyledon.At 6 (A) and 20 (B) hours after inoculation, the cotyledon was washed with water containing 0.02% Tween 20.The washed out spores were counted and the ratio of attachment was calculated.Means in the plot topped by the same letter do not differ based on Fisher's LSD test at P#0.01 (n = 3).doi:10.1371/journal.pone.0092503.g006glycosyltransferases (GTs), polysaccharide lyases (PLs), and carbohydrate esterases (CEs).Among these four groups, CE, GH, and PL are often considered as CWDEs due to their important roles in plant biomass decomposition by fungi and bacteria [54].In the present study we chose four genes from each of these three groups to investigate their expression in the LmSNF1 Ko strains.
Eleven out of these 12 genes were up-regulated during pathogenesis at 4 dai.One gene encoding a glycogen debranching enzyme had a similar expression profile in the glucose media and in planta.In Blumeria graminis f. sp.hordei, glycogen debranching enzymes are abundant during spore germination but become sparse in the later stages of infection [55].While the 4 dai time point examined in this study occurred after the up-regulation stage of this gene, the glycogen debranching enzyme gene was, nonetheless, among the 11 CWDE genes that were downregulated in the LmSNF1 Ko strains (Table 2).The expression of one gene encoding a chitin deacetylase was at similar levels in WT and LmSNF1 Ko strains, indicating that this gene is not under regulation of LmSNF1.That some CWDE genes were not downregulated was also observed in the SNF1 knockout strains of other plant pathogenic fungi [19,56].In Ustilago maydis in particular, knockout of SNF1 reduced the expression of several CWDE genes but also increased the expression of two xylanase genes [56].
The three pathogenicity-related genes (isocitrate lyase, LopB and the plasma membrane ATPase 1) had been studied previously [32,33,36].All of these genes are under control of glucose repression (Table S3) and in general, the in vitro and in planta expression of these genes was consistent with previously described expression profiles.In the LmSNF1 Ko strains, only the LopB gene was down-regulated and the expression of the isocitrate lyase and the plasma membrane ATPase 1 was similar to the WT, indicating that the latter two are not under control of the LmSNF1 derepression.

Function of LmSNF1 on vegetative growth, sporulation, conidial germination and attachment
The radial growth of LmSNF1 Ko strains was similarly impaired when they were grown on media with or without glucose.This  Cotyledons of canola cultivar Westar were inoculated with 5-ml droplets of conidial suspension from each strain at a concentration of 1610 6 conidia/ml.The pictures were taken at 6 day after inoculation.WT, the wild-type; Ect, the ectopic insertion strain; Ko1, LmSNF1 knockout strain.doi:10.1371/journal.pone.0092503.g007 observation is consistent with reports in M. oryzae [19] and P. digitatum [22], but conflicts with all other reports, in which the SNF1 mutants grew similarly with WT in glucose containing media.The SNF1 genes have been reported to control sucrose and pectin utilization in differentially in various plant pathogenic fungi.For example, the growth of SNF1 mutants of C. carbonum [10] and V. dahliae [21] was similar to the wild-types when they were grown on media with sucrose as the only carbon source, whereas the growth of SNF1 mutants of F. graminearum [20] and P. digitatum [22] was significantly impaired.The growth of C. carbonum and P. digitatum SNF1 mutants was seriously affected when these fungi were grown on media with pectin.The growth of V. dahliae SNF1 mutants [21] and LmSNF1 Ko strains was partially affected, whereas the growth of a F. oxysporum SNF1 mutant [17] was not affected when it was grown on pectin-supplemented media.The differences in carbon utilization by SNF1 mutants indicate the potential existence of distinct carbon utilization mechanisms during infection of plants by pathogenic fungi.Furthermore, in S. cerevisiae, SNF1 also has been shown to be involved in the regulation of nitrogen metabolism, gluconeogenesis, and the glyoxylate and tricarboxylic acid cycles as well as in respiration and b-oxidation [57,58].It is possible that LmSNF1 regulates these processes in L. maculans and thus affects its growth on different media.Sporulation and spore germination were also affected by the loss of LmSNF1, which agreed with a report on the F. graminearum SNF1 mutant [20].In a SNF1 mutant of C. carbonum, both sporulation and spore germination were similar to the wild-type [10].In a SNF1 mutant of M. oryzae, the sporulation ability was abolished [19], while in a SNF1 mutant of P. digitatum sporulation was reduced but the spores germinated normally [22].Sporulation and spore germination in filamentous fungi are controlled by multiple pathways and factors including nutrients, light, salt, and oxygen [59,60].The lower production of conidia of the LmSNF1 Ko strains could result from inadequate or unbalanced nutrient accumulation during spore formation.LmSNF1 also could trigger the expression of genes necessary for conidial germination.Moreover, an inability to accumulate the necessary nutrients or to trigger the genes involved in weakening the conidial cell wall architecture could delay or prevent conidial germination.
The attachment of fungal spores to the plant surface is the first step in committing a pathogen to the establishment of disease.Knockout of LmSNF1 affected the conidial attachment to the plant surface.In plant pathogenic fungi, two basic strategies for spore attachment to the host surface have been postulated: ''passive attachment'' resulting from the presence of preformed materials on the fungal surface and ''active attachment'' requiring the de novo production of fungal adhesives [61].CWDEs capable of degrading components of the plant cuticle, such as cutinase [62] and lipase [48,63], have been demonstrated to be important for fungal attachment after a brief period in contact with the plant.Esterase, including lipase and cutinase, produced by conidia of L. maculans during the stages of germination might be important for active attachment.Impairment of conidial attachment in the LmSNF1 Ko strains is likely a result of the down-regulation of esterases or other enzymes required for attachment.

LmSNF1 as a pathogenicity factor
In plant pathogenic fungi, SNF1 orthologs are widely known to be required for fungal virulence, mostly by activating the expression of CWDEs.Two groups can be identified among the plant pathogenic fungal species in which the SNF1 gene has been knocked out.The first group includes M. oryzae and U. maydis.Knockout of M. oryzae SNF1 did not affect the expression of the tested CWDE genes but abolished the fungal pathogenicity [19], whereas knockout of U. maydis SNF1 caused down-regulation of some but up-regulation of other CWDE genes and slightly decreased fungal pathogenicity [56].The second group includes P. digitatum, F. graminearum, F. oxysporum, V. dahlia, C. carbonum and L. maculans, in which knockout of SNF1 abolished the expression of most or all of the tested CWDE genes and decreased the fungal pathogenicity.Interestingly, both species in the first group are known to produce well-developed appressoria during penetration.In the second group, only C. carbonum and V. dahliae produce appressoria, but both species do not require the formation of appressoria to cause disease [64,65].Activation of CWDEs during the early stages of infection maybe more important in species of the second group than in those of the first group, and the link between SNF1, CWDEs and pathogenicity is more strictly complied.
L. maculans invades the plant via the stomatal apertures [41].The hyphae need to grow on surface of the plant epidermis until they encounter a stoma.During this short period, attachment to the plant surface is critical because water can wash away deposited conidia and rain drops can cause strong air movement over short distances.It can be inferred that impairment of conidial attachment in the LmSNF1 Ko strains would contribute to the loss of pathogenicity under natural conditions.However, the wound inoculation protocol employed in the current study bypassed the attachment and penetration phases of infection.Thus, the observed loss-of-pathogenicity most likely reflects impaired spore germination and reduced vitality of the germinated spores.Biochemically the reduced virulence of SNF1 mutants in C. carbonum and F. oxysporum was explained by their reduced ability to utilize complex carbon sources and the reduced expression of CWDEs [10,17].Similar mechanisms also may explain the loss-ofpathogenicity of the LmSNF1 Ko strains.During the initial interaction of L. maculans with canola, germination of the spores or the presence of chemicals on plant surface activate LmSNF1 and result in the subsequent induction of CWDEs, which function as virulence factors that contribute to pathogenesis.
In summary, the present study demonstrated that the LmSNF1 gene is required by L. maculans to cause blackleg disease on canola.The down-regulation of CWDE genes in the LmSNF1 Ko strains provides additional evidence for the importance of CWDEs in fungal pathogenicity.In addition, the study also provided data to support the suitability of L. maculans for targeted gene knockout.Table S1 Sequences of primers indicated in Figure 1.

(PDF)
Table S2 Primers for CWDEs and pathogenicity genes.(PDF) Table S3 Expression of selected CWDE and pathogenicity genes in Leptosphaeria maculans after cultured in minimal medium supplemented with 1% glucose or in infected canola cotyledons at 4 day after inoculation.(PDF)

Figure 2 .
Figure 2. Expression analysis of LmSNF1 in the wild type of Leptosphaeria maculans by quantitative PCR.A, Expression of LmSNF1 24 h after culturing in YPG or minimal media (CD) supplemented with various carbon sources.B, Expression of LmSNF1 in infected canola tissue.Canola cultivar Westar was inoculated with the fungus and the infected plant samples were collected at 0-144 hours after inoculation.Means in the plot topped by the same letter do not differ based on Fisher's LSD test at P#0.01 (n = 3).doi:10.1371/journal.pone.0092503.g002

Figure 3 .
Figure 3. Analysis of LmSNF1 gene knockout mutants of Leptosphaeria maculans.WT, the wild-type; Ect, the ectopic insertion strain; Ko, LmSNF1 knockout strains.A, Southern blot.Genomic DNA from the fungal strains was digested with EcoRI and hybridized with a PCR fragment that encompasses a part of LmSNF1 open reading frame and its 39 region.B, Assessment of expression of the 59 and 39 genes by quantitative PCR.Means in the plot topped by the same letter do not differ based on Fisher's LSD test at P#0.01 (n = 3).doi:10.1371/journal.pone.0092503.g003

Figure 4 .Figure 5 .
Figure 4. Sporulation of Leptosphaeria maculans strains on V8 agar plates assessed by number of conidia per plate (A) and number of conidia per mm 2 of culture (B).WT, the wild-type; Ect, the ectopic insertion strain; Ko, LmSNF1 knockout strains.Means in the plot topped by the same letter do not differ based on Fisher's LSD test at P#0.01 (n = 3).doi:10.1371/journal.pone.0092503.g004

Figure
Figure S1 Comparison of the fungal phylogenetic tree based on sequences of 70 protein encoding genes (left) and the tree based on SNF1 sequences (right).Clades identical in the two trees are connected by dotted lines.(PDF) Figure S2 PCR analysis of LmSNF1 gene knockout mutants of Leptosphaeria maculans.WT, the wild-type; Ect, the ectopic insertion strain; Ko, LmSNF1 knockout strains.Labels on top indicate the target genes (forward/reverse primers).(PDF) Figure S3 Growth of Leptosphaeria maculans strains in minimal media supplemented with 15% V8 juice (A), 1% glucose (B) or 1% pectin (C).(PDF) Figure S4 Germinated (arrow) and ungerminated (arrow head) conidia of Leptosphaeria maculans under microscope (10640).

Figure
Figure S5 Lesion development on canola cotyledons after inoculated with Leptosphaeria maculans strains.(PDF)

Table 2 .
Expression of selected cell wall degrading enzyme genes in the wild-type and two LmSNF1 knockout strains after culturing in minimal medium supplemented with 1% pectin.