Isocitrate Lyase Is Essential for Pathogenicity of the Fungus Leptosphaeria maculans to Canola (Brassica napus)

A pathogenicity gene has been identified in Leptosphaeria maculans,the ascomycetous fungus that causes blackleg disease of canola(Brassica napus). This gene encodes isocitrate lyase, a componentof the glyoxylate cycle, and is essential for the successfulcolonization of B. napus. It was identified by a reverse geneticsapproach whereby a plasmid conferring hygromycin resistancewas inserted randomly into the L. maculans genome. Twelve of516 transformants tested had reduced pathogenicity on cotyledonsof B. juncea and B. napus, and 1 of these 12 had a deletionof the isocitrate lyase gene, as well as an insertion of thehygromycin resistance gene. This mutant was unable to grow onfatty acids, including monolaurate, and the isocitrate lyasetranscript was not detected. When the wild-type gene was reintroducedinto the mutant, growth on monolaurate was restored and pathogenicitywas partially restored. L. maculans isocitrate lyase is producedduring infection of B. napus cotyledons, while the plant homologueis not. When 2.5% glucose was added to the inoculum of the isocitratelyase mutant, lesions of sizes similar to those caused by wild-typeisolate M1 developed on B. napus cotyledons. These findingssuggest that the glyoxylate pathway is essential for diseasedevelopment by this plant-pathogenic fungus, as has been shownrecently for a fungal and bacterial pathogen of animals anda bacterial pathogen of plants. Involvement of the glyoxylatepathway in pathogenesis in animals and plants presents potentialdrug targets for control of diseases.

INTRODUCTION

Top

Introduction

Fungal plant pathogens employ diverse strategies to infect theirhosts. Mechanisms underlying these strategies are being revealedbecause of the increasing amenity of many fungi to moleculargenetic analysis. In particular, techniques including insertionalmutagenesis have identified fungal pathogenicity genes crucialto disease. Such genes are defined on the basis of mutagenesistechniques; disruption of a pathogenicity gene results in significantor complete loss of disease symptoms when mutants are inoculatedonto hosts that are susceptible to the wild-type pathogen. Thetypes of pathogenicity genes expressed depend on the infectionprocess of a particular fungus, for instance, whether hydrolyticenzymes, specialized infection structures, or toxins are required,as well as the nutritional status of the pathogen (for reviews,see references 8 and 22). Pathogenicity genes may be valuableas fungicide targets for disease control. If such genes arespecific to the fungus, inhibitors may be designed to bind totheir products and block the growth of the fungus without affectingthe plant.

Leptosphaeria maculans (anamorph = Phoma lingam) causes blacklegdisease of canola (Brassica napus) worldwide. Its infectionpathway is well characterized (for a review, see reference 7).Hyphae enter the host through stomata without the aid of specializedinfection structures such as appressoria. The fungus initiallygrows in planta as a biotroph and later becomes necrotrophic.The fungus grows down the stem and produces a dark canker girdlingthe base of the plant, hence the name blackleg. In spite ofthe economic importance of blackleg disease, L. maculans hasnot been studied in great detail at a molecular genetic level.This paper describes an insertional mutagenesis approach toidentifying pathogenicity genes of L. maculans and the characterizationof one such gene encoding isocitrate lyase.

MATERIALS AND METHODS

Top

Materials and Methods

Fungal culturing and transformation. L. maculans isolate M1, which has well-defined host specificity,was mutated. This isolate causes cotyledonary and stem lesionson all of the B. napus and B. juncea (Indian mustard) cultivarstested, and its ability to attack B. juncea is controlled bya single locus (3, 15). Isolate M1 was transformed to hygromycinresistance by using plasmid pUCATPH (11) and restriction enzyme-mediatedintegration in the presence of 10 to 40 U of HindIII, KpnI,or SacI as previously described (17). Transformed protoplastswere plated onto regeneration medium (1.2 M sucrose, 10% Campbell'sV8 juice, 1.5% agar) containing hygromycin (50 µg/ml;Sigma). Hyphae were visible after 2 to 3 weeks and were subjectedto two rounds of hyphal-tip culturing onto 10% V8 juice agarplates containing hygromycin before transfer to nonselectivemedium. For Northern blot analysis, mycelia were grown in completemedium (17.5 g of Czapek-Dox per liter, 0.005% yeast extract)in still culture (50 ml) in petri dishes (15-cm diameter) for15 to 25 days, washed in sterile water, and transferred to completemedium or minimal medium (14) containing no carbon or nitrogensource, ammonium acetate (0.5%) only, or ammonium acetate (0.5%)and glucose (0.25%). Mycelia were harvested 7 h later.

Pathogenicity tests. Transformants were screened for the ability to infect cotyledonsof 14-day-old seedlings of B. juncea cv. Stoke wounded witha 26-gauge needle as previously described (15). B. juncea waschosen for the initial screen so that target pathogenicity genesin L. maculans would include the host specificity gene (3).Lesion development was assessed between 10 and 14 days postinoculationand compared visually to lesions caused by isolate M1. Transformantswith reduced pathogenicity were rescreened on B. juncea cv.Stoke and B. napus cv. Westar, and those with confirmed reducedpathogenicity were characterized further.

Nucleic acid manipulations and analysis of a nonpathogenic mutant. Chromosomal and genomic DNA or total RNA from L. maculans wasisolated, resolved by gel electrophoresis, blotted onto nylonmembranes, and hybridized as previously described (17). Regionsof L. maculans DNA flanking the pUCATPH plasmid were obtainedby inverse PCR, and the resultant products were cloned intopGEM T-Easy (Promega). The inverse PCR product derived froma nonpathogenic mutant, which has a single insertion of pUCATPH,was radiolabeled and used to probe a cosmid library (pWEB vector;EpiCentre Technologies) containing 30- to 40-kb inserts of DNAfrom isolate M1 to obtain the wild-type copy of the mutatedL. maculans gene. Insert DNA from a hybridizing cosmid was subcloned,and resultant inserts were sequenced by using ABI dye-primercycle reactions and an automated sequencer (Applied Biosystemsmodel 373A). The DNA sequence obtained was compared to thosein the GenBank database by using BLAST (1), and genes were predictedby using FGENESH software (http://www.softberry.com) and GENSCAN(www.bionavigator.com). One subclone contained an open readingframe with a high degree of sequence similarity to that of isocitratelyase (icl1). A 6-kb XbaI fragment containing icl1 was ligatedinto the XbaI site of pAN8-1, a plasmid that confers phleomycinresistance (12). Protoplasts of the nonpathogenic mutant ( ${Delta}$ icl1)were transformed with this construct or pAN8-1 as describedabove, and resultant transformants were selected for growthon phleomycin at 20 µg/ml (Cayla).

Putative complemented strains were inoculated onto B. napuscv. Westar cotyledons and pycnidiospores germinated on wateragar plates. Strains were also grown on minimal medium agarsupplemented with monolaurate (0.25% Tween 20, polyoxyethylenesorbitan monolaurate) or glucose (0.25%) as the carbon source.Additionally, growth on this carbon source was tested in thepresence of inhibitors of isocitrate lyase, i.e., 0.5 or 1 mM3-nitropropionate (an analogue of succinate) and 0.5 or 1 mM3-bromopyruvate (an analogue of glyoxylate) (19). After 9 days,colony diameters were compared to those of the wild type (M1)and the ${Delta}$ icl1 mutant. These inhibitors were also added to inoculumand tested for the ability to affect disease development onB. napus cotyledons. The infection patterns of the wild-typeand ${Delta}$ icl1 strains were investigated by light microscopy. Fourteendays after inoculation, cotyledon tissue was placed in boiling95% ethanol for 15 min, cleared in saturated chloral hydratefor 5 days, mounted in Hoyer's medium, and examined with anOlympus light microscope (BH-2) as previously described (18).

To investigate transcription of isocitrate lyase during infection,cotyledons were harvested over the course of infection withisolate M1 and total RNA was examined by Northern blot analysis.The L. maculans icl1 gene and the B. napus isocitrate lyasegene which was cloned by PCR with primers 5'-GAAGGGAGATTCGAGGCG-3'and 5'-GCCACACTTCTTGGTGAC-3' (4) were radiolabeled and hybridizedto the RNA blot. Control RNA samples were purified from germinatedseed, uninfected cotyledons and leaves, and L. maculans mycelia.

RESULTS AND DISCUSSION

Top

Results and Discussion

A total of 516 hygromycin-resistant transformants were screenedon B. juncea and B. napus, and 12 caused either no lesions orlesions less than 30% of the diameter of those caused by isolateM1. The host specificity of these 12 mutants indicates thatnone of them had an insertion of pUCATPH in the locus conferringhost specificity on B. juncea (3). Nine of these mutants hadmultiple insertions of pUCATPH, while three had single copies;one of the latter mutants was characterized further. FlankingDNA obtained by inverse PCR was used to screen a cosmid libraryto obtain the wild-type gene. A sequence including the isocitratelyase (EC 4.1.3.1) gene (icl1) was isolated (GenBank AY118108),and the intron positions in this gene were determined by sequencingfragments of cDNA of L. maculans amplified by PCR with appropriateprimers (Fig. 1). A predicted gene is located downstream oficl1 and transcribed in the opposite direction. This gene hashigh sequence similarity to that which encodes ribose-phosphatepyrophosphokinase (EC 2.7.6.1; prs5) of Neurospora crassa andSaccharomyces cerevisiae and is predicted to be involved innucleotide metabolism (6). Pulsed-field gel electrophoresisshowed that the L. maculans icl1 gene is located on one memberof a chromosome doublet (6/7; size, 2.15 Mb) of isolate M1 (datanot shown). Southern blot analysis of isolates M1 and ${Delta}$ icl1 showedthat deletion of the icl1 gene and an unknown amount of DNAoccurred during integration of pUCATPH (Fig. 1 and data notshown).

View larger version (82K):
[in this window]
[in a new window]
FIG. 1. Alignment of the L. maculans isocitrate lyase gene (icl1) with homologues. (A) Structure of a 5.8-kb region of L. maculans DNA including icl1 (exons are black) and an open reading frame with high sequence similarity to ribose-phosphate pyrophosphokinase (prs5) (exons are grey). Arrows under the genes indicate directions of transcription. The region deleted in the mutant ( ${Delta}$ icl1) is indicated. (B) Alignment of L. maculans icl1 (L.ma) with isocitrate lyase genes from selected organisms: the ascomycetes Penicillium marneffei (P.ma; accession no. AF373018) and C. albicans (C.al; accession no. AF222905), the basidiomycete C. neoformans (C.ne; accession no. AF455253), the plant B. napus (B.na; accession no. L08482), and the bacterium M. tuberculosis (M.tu; accession no. AE007051).

Isocitrate lyase is one of two enzymes in the glyoxylate pathwaythat are involved in the metabolism of two-carbon compoundssuch as acetate and fatty acids. Accordingly, growth of the ${Delta}$ icl1 mutant was tested on such carbon sources. This mutant wasunable to grow on a C₁₂ fatty acid (monolaurate; 0.25% Tween20) or ammonium acetate (0.5%) as the sole carbon source, incontrast to the wild-type isolate (Fig. 2). Both the mutantand the wild type grew on 0.25% glucose. An analogue of glyoxylate,3-bromopyruvate, at a concentration of 0.5 or 1 mM did not inhibitgrowth of isolate M1 specifically on 0.25% Tween 20 (data notshown). However, 3-nitropropionate (an analogue of succinate)at a concentration of 0.5 or 1 mM prevented growth on Tween20 when it was the sole carbon source (Fig. 2). Both moleculeswere toxic to L. maculans at concentrations of greater than5 mM.

View larger version (48K):
[in this window]
[in a new window]
FIG. 2. Growth of L. maculans isolates on complete medium supplemented with hygromycin or phleomycin and on minimal medium supplemented with glucose or monolaurate (Tween 20). A wild-type isolate (M1), an isocitrate lyase mutant ( ${Delta}$ icl1), a ${Delta}$ icl1 isolate transformed with phleomycin resistance gene only (( ${Delta}$ icl1+phl), and a ${Delta}$ icl1 isolate transformed with a wild-type copy of icl1 and the phleomycin resistance gene ( ${Delta}$ icl1+icl1) were tested. Addition of 3-nitropropionate (3-NP; 1 mM) inhibited the growth of all isolates on monolaurate as the sole carbon source.

No transcript of icl1 was observed in the ${Delta}$ icl1 mutant, whilein isolate M1, the gene was induced by starvation conditionsand by 0.5% ammonium acetate. Glucose (0.25%) did not affectthe induction of transcription by acetate (Fig. 3). Glucosecauses catabolite repression of isocitrate lyase in fungal species.However, in Aspergillus nidulans, addition of glucose to acetate-containingmedium at concentrations similar to those used here for L. maculansdid not result in reduced transcription of isocitrate lyase(2). Transcription of icl1 in isolate M1 was detected in cotyledonsof B. napus 14 days after infection (data not shown). Transcriptionof ß-tubulin, a constitutively expressed gene thatis an indicator of fungal biomass during infection, was alsodetected at this stage. A wild-type copy of the icl1 gene wasreintroduced into the ${Delta}$ icl1 mutant, and eight transformants ( ${Delta}$ icl1+icl1)that grew on 0.25% Tween 20 were tested on B. napus (Fig. 3).The resultant lesions for all strains developed more slowlythan those caused by the wild-type isolate; however, they werelarger than those produced by the ${Delta}$ icl1 mutant (Fig. 4).

View larger version (62K):
[in this window]
[in a new window]
FIG. 3. Transcriptional regulation of the isocitrate lyase (icl1) gene of L. maculans. Northern analysis of RNA from isolate M1 or an isocitrate lyase mutant ( ${Delta}$ icl1) where mycelia were grown in complete medium and then transferred to minimal medium with ammonium acetate (lane 1, ${Delta}$ icl1; lane 2, M1), ammonium acetate and glucose (lane 3, M1), or no carbon or nitrogen (lane 4, M1) or to complete medium (lane 5, M1). The blot was probed with icl1. No transcript is observed in the ${Delta}$ icl1 mutant. The gene is induced under starvation conditions and by acetate but is not repressed by glucose. The blot reprobed with a fragment of ß-tubulin (ßtub; bp 8 to 1648 of the nucleotide sequence with GenBank accession no. AF257329) shows equal loading and transfer of RNA.

View larger version (43K):
[in this window]
[in a new window]
FIG. 4. Pathogenicity of L. maculans isolates, including wild-type isolate M1, an isocitrate lyase mutant ( ${Delta}$ icl1), and a strain where ${Delta}$ icl1 was transformed with a wild-type copy of icl1 ( ${Delta}$ icl1+icl1), to cotyledons of B. napus cv. Westar. (A) Reintroduction of the wild-type copy of icl1 restores the ability of ${Delta}$ icl1 to form lesions. (B) Pathogenicity is also partially restored by addition of glucose (2.5%) to inoculum of ${Delta}$ icl1. Pen marks are present on the cotyledons to distinguish inoculations.

Pycnidiospores of the ${Delta}$ icl1 mutant had an extremely low germinationrate, compared to that of the wild type or complemented strains,on water agar plates. Microscopic examination of the woundedinfection site revealed limited hyphal growth of the ${Delta}$ icl1 mutantin planta (Fig. 5). This suggests that, rather than a defectin spore germination leading to reduced pathogenicity, a lackof carbon utilization in planta or on the plant surface (inthe form of fatty acids) may be responsible for limiting growth;this hypothesis is supported by the finding that when 2.5% glucosewas added to the inoculum of the ${Delta}$ icl1 mutant, lesions of sizessimilar to those caused by isolate M1 developed (Fig. 4).

View larger version (116K):
[in this window]
[in a new window]
FIG. 5. Phenotypes of isocitrate lyase mutant ( ${Delta}$ icl1) and wild-type (M1) L. maculans strains in planta. Tissue of B. napus was cleared 14 days after inoculation. The plane of focus is just below the cotyledon surface (s, plant stoma; arrow, fungal hypha). Both isolates colonize the plant, but the ${Delta}$ icl1 mutant shows less hyphal branching and ramification of than does isolate M1. Bar = 10 µm.

The finding that isocitrate lyase is involved in pathogenesisof L. maculans is consistent with recent findings that thisgene is necessary for the full virulence of Mycobacterium tuberculosisand Candida albicans in mice (10, 13). L. maculans icl1 is transcribedduring infection of B. napus, which is consistent with the inductionof transcription of this gene in C. albicans cells grown inthe presence of mouse macrophages (10). Vereecke et al. havereported that, in the plant-pathogenic bacterium Rhodococcusfascians, the second enzyme in the glyoxylate pathway, malatesynthase, is essential for pathogenicity on tobacco (25). Theseauthors proposed that glyoxylate buildup in the bacterium inhibitsfurther growth after colonization of the plant.

Unlike animals, plants utilize the glyoxylate pathway, so inhibitorsof fungal isocitrate lyase and malate synthase would only beeffective for disease control if they did not bind the planthomologues (which is unlikely, given the high degree of aminoacid sequence similarity between isocitrate lyase genes; Fig.1) or if the plant homologues were not induced during infection.We cloned a fragment of the B. napus isocitrate lyase gene toinvestigate its transcription during infection. This gene wastranscribed at high levels in germinating seeds, as seen previously(4). However, transcription was not detected at 3, 6, 9, and12 days postinoculation or in uninfected leaves or cotyledons,although a control plant gene (that which encodes actin) wasdetected (data not shown). Although inhibitors of icl1 mightbe expected to lead to a decrease in lesion size caused by thewild type on B. napus cotyledons, this did not occur when 1mM 3-nitropropionate was added to inoculum (data not shown).Higher concentrations of 3-nitropropionate (>5 mM) causednecrosis, suggesting that this compound is phytotoxic at highconcentrations.

The glyoxylate pathway occurs in the peroxisome, an organelleinvolved in various other functions, including ß-oxidationof fatty acids and metabolism of reactive oxygen species. Therole of peroxisomal proteins in plant-pathogenic fungi for theformation of infection structures such as appressoria has beendemonstrated by disruption of the ClaPEX6 gene in Colletotrichumlagenarium, whereby the mutant does not make appressoria andonly colonizes wounded tissue (9). Appressorial turgor in therice blast fungus Magnaporthe grisea depends on lipid reservesto make glycerol (24), an activity probably reliant on peroxisomalfunction and the glyoxylate pathway. Recently, an isocitratelyase mutant of M. grisea has been shown to have reduced pathogenicity(Nick Talbot, personal communication). An additional pathwayfor utilization of acetyl coenzyme A derived from ß-oxidationof fatty acids involves carnitine acetyltransferases (23). Acarnitine acetyltransferase gene, probably located in the peroxisomesand mitochondria, is required by M. grisea for full pathogenicity(21). Curiously, the transcription of another carnitine acetyltransferasegene, probably located in the cytoplasm, was induced even morehighly than the glyoxylate pathway genes in S. cerevisiae duringphagocytosis by macrophages (10). Unlike M. grisea, L. maculansdoes not make appressoria, so the role of fatty acid utilizationin the latter fungus is not to provide molecules to promoteturgor pressure for infection structures. These results indicatethat the glyoxylate pathway functions in pathogenicity in twopathogenic plant fungi with very different modes of infection.

Although the glyoxylate pathway genes are involved in pathogenicityin several animal and plant pathogens, they are not universalpathogenicity factors. Isocitrate lyase is dispensable for diseasedevelopment in an isolate of S. cerevisiae that is pathogenicto mammals (5), as well as in Cryptococcus neoformans, an opportunisticpathogen of animals (16). Glyoxylate pathway genes are absentfrom the plant-pathogenic bacterium Xylella fastidiosa (20).

ACKNOWLEDGMENTS

We thank Olen Yoder and Gillian Turgeon for plasmid pUCATPH,Peter Punt for plasmid pAN8-1, Richard Oliver for phleomycin,Peter Dunn for protoplasting enzyme, Leanne Wilson for RNA fromB. napus infected with L. maculans, and Nick Talbot for permissionto cite unpublished observations. Also, we thank Alice Churchillfor critical appraisal of the manuscript.

We thank the Australian Grains Research and Development Corporationfor funding.

FOOTNOTES

* Corresponding author. Mailing address: School of Botany, The University of Melbourne, Melbourne, Victoria 3010, Australia. Phone: 61 3 8344 5062. Fax: 61 3 9347 1071. E-mail: bhowlett{at}unimelb.edu.au.

REFERENCES

Top