Cloning, characterization and chromosomal location of three genes encoding host-cell-wall-degrading enzymes in Leptosphaeria maculans, a fungal pathogen of Brassica spp.

Gene 248 (2000) 89–97 www.elsevier.com/locate/gene

Cloning, characterization and chromosomal location of three genes encoding host-cell-wall-degrading enzymes in Leptosphaeria maculans, a fungal pathogen of Brassica spp. Adrienne C. Sexton a, *, Martina Paulsen b,1, Johannes Woestemeyer b, Barbara J. Howlett a a School of Botany, The University of Melbourne, Parkville, Vic 3010, Australia b Friedrich-Schiller-Universitat Jena, Institut fur Mikrobiologie, Lehrstuhl fur Allegemeine Mikrobiologie und Mikrobengenetik, FSU Jena, Germany Received 20 November 1999; received in revised form 28 February 2000; accepted 13 March 2000 Received by W. Martin

Abstract The ascomycete, Leptosphaeria maculans, causes blackleg disease of oilseed Brassica spp. such as canola (Brassica napus). We have cloned a gene encoding endopolygalacturonase, pg1, and two genes encoding cellulases, cel1 and cel2, in L. maculans. These genes are not clustered in the genome, as they are located on different chromosomes. The deduced amino acid sequences of all three genes predict an N-terminal signal sequence, as is common for secreted fungal enzymes that degrade plant cell walls. The endopolygalacturonase encoded by pg1 shows the highest similarity (54% amino acid identity) to endopolygalacturonase 4 from Botrytis cinerea. Both cel1 and cel2 appear to encode cellobiohydrolase, and neither gene encodes a recognizable cellulose-binding domain or linker region. Transcription of pg1 is induced in cultures containing 1% polygalacturonic acid or pectin, and cel1 is induced in 1% cellulose or carboxymethylcellulose, as shown by Northern analysis. Glucose represses the induction of cel1 caused by cellulose and carboxymethylcellulose, but does affect transcription of pg1. Transcription of cel2 (but not cel1 or pg1) is detectable during infection of B. napus and B. juncea cotyledons and leaves using reverse transcription-PCR. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Brassica juncea; Brassica napus; Cellulase; Glucose repression; Polygalacturonase

1. Introduction Leptosphaeria maculans (Desm.) Ces. & de Not. (anamorph: Phoma lingam ( Tode ex Fr.) Desm.) causes blackleg disease of oilseed Brassica spp. such as canola (B. napus), and produces serious crop losses world-wide. During infection, spores germinate on cotyledons or leaves, and the emergent hyphae enter the plant tissue via stomata (Chen and Howlett, 1996). Hyphae then Abbreviations: bp, base pair; cDNA, DNA complementary to RNA; dNTP, deoxyribonucleoside triphosphate; kb, kilobase(s); mRNA, messenger RNA; nt, nucleotide(s); PCR, polymerase chain reaction; p.i., post-inoculation; RT-PCR, reverse transcription-polymerase chain reaction. * Corresponding author. Tel.: +61-3-8344-5056; fax: +61-3-9347-1071. E-mail address: [email protected] (A.C. Sexton) 1 Present address: University of Cambridge, Department of Anatomy, Downing St., Cambridge CB2 3DY, UK.

proliferate in the intercellular spaces of the mesophyll layer and then enter the vascular tissue, eventually reaching the base of the stem, where cortical cells are colonized, forming a blackened canker — hence the name ‘blackleg’. Phytopathogenic fungi often produce a variety of enzymes active against plant-wall constituents such as pectin, cellulose and xyloglucan. These enzymes may produce oligosaccharide elicitors/suppressors of hostplant defences (Piel et al., 1997; Boudart et al., 1998; Moerschbacher et al., 1999). In most cases with phytopathogenic fungi, the disruption of one or two genes encoding cell-wall-degrading enzymes has not led to any changes in pathogenicity, as these fungi produce a battery of enzymes, and the loss of one enzyme is likely to be compensated for by other enzymes with similar activities. For example, Sposato et al. (1995) disrupted a cellobiohydrolase gene in Cochliobolus carbonum and found no change in pathogenicity on maize and that the

0378-1119/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 0 0 ) 0 0 14 2 - 6

90

A.C. Sexton et al. / Gene 248 (2000) 89–97

mutant still produced cellulase activities in culture. Disruption of a gene encoding endopolygalacturonase in Cryphonectria parasitica caused no reduction in canker formation on chestnut, and the mutant produced polygalacturonase activities in infected tissues (Gao et al., 1996). Although individual cell-wall-degrading enzymes may not be crucial for pathogenicity, a battery of such enzymes must certainly contribute to the development of many plant diseases. L. maculans produces cellulase and polygalacturonase activities in culture (Hassan et al., 1991; Annis and Goodwin, 1997a) and in stem lesions on B. napus ( Easton and Rossall, 1985). However, none of these studies tested for L. maculans cellulase or polygalacturonase activities during infection of Brassica leaves or cotyledons. In spite of the economic importance of this fungus, the cloning of only three protein-encoding genes (nitrate reductase, nitrite reductase and cyanide hydratase) has previously been published ( Williams et al., 1995; Sexton and Howlett, 2000). In this paper, we describe the cloning and characterization of an endopolygalacturonase gene and two cellulase genes from L. maculans, and report on their transcription in culture and during infection of leaves and cotyledons of two oilseed Brassica spp.

2. Materials and methods 2.1. Culturing of L. maculans isolates Isolates C13 and M1 are derived from single ascospores found on canola trash in Australia. Both isolates can attack most B. napus cultivars, including cv. Midas, but isolate C13 is avirulent on B. juncea cv. Stoke, whilst isolate M1 is virulent (Purwantara et al., 1998). Isolate W4E87 is from Germany and attacks B. napus. Transcriptional induction of L. maculans genes in the presence of various carbon compounds was tested. Pycnidiospores (2×106 spores per 50 ml petri dish) of isolate M1 were cultured in Czapek–Dox broth (bacto saccharose 30 g/l, sodium nitrate 3 g/L, dipotassium sulphate 1 g/l, magnesium sulphate 0.5 g/l, potassium chloride 0.5 g/l, ferrous sulphate 0.01 g/l; Difco, USA) containing yeast extract (1 g/l ) for 5 days at 22°C. Mycelia were washed and transferred for 24 h to minimal medium (Newton and Caten, 1988) containing a test chemical as the carbon source. The nitrogen source in these experiments was 5 mM ammonium tartrate. Initially, a range of carbon sources [glycerol, glucose, sucrose, polygalacturonic acid (Grade III, Sigma, St. Louis, MO), apple pectin (250 Grade, BDH Ltd., UK ), carboxymethylcellulose (sodium salt, Sigma), microcrystalline cellulose ( E. Merck, Germany)] each at 1% w/v, or no carbon source, was used. ( When two test chemicals

were included together, both were at 1% w/v.) For carbon-starvation conditions, mycelia grown for 5 days in Czapek–Dox/yeast extract broth were transferred to minimal medium containing a range of sucrose concentrations (30, 3 or 0.3%) with 17.5 mM ammonium tartrate as the nitrogen source. For nitrogen-starvation conditions, mycelia were transferred to minimal medium containing ammonium tartrate (17.5, 5.0 or 0.5 mM ) with 3% sucrose as the carbon source. Mycelia transferred to media with 30% sucrose (and 17.5 or 0.5 mM ammonium tartrate) served as a control for the effect of a high C:N ratio. After 24 h, mycelia were harvested, rinsed and freeze-dried. 2.2. Isolation of genes encoding polygalacturonase and two cellulases The polygalacturonase gene was isolated from a cDNA library prepared from poly(A)+ RNA from L. maculans isolate M1 grown on 10% Campbells V8 juice, in the lambda phagemid vector Uni-Zap XR (Stratagene, La Jolla, CA). During low-stringency screening of this library with a 300 bp cDNA encoding part of a putative ribosomal protein, a full cDNA clone for endopolygalacturonase ( pg1) was isolated. The pg1 nucleotide sequence had 32% similarity to the probe. This pg1 cDNA was used to isolate the corresponding gene from a genomic library of L. maculans isolate M1 (constructed in Stratagene ZAP Express vector). The genomic clone contained only 100 bp of the upstream sequence, so inverse-PCR was used to obtain a further 400 bp of promoter region. For isolating cellulase genes, a cDNA library was prepared using messenger RNA isolated from mycelia of L. maculans isolate W4E87, which had been growing on complete medium (150 g/l boiled peas, 1% glucose, salts), then transferred to minimal medium containing 1% cellulose (microcrystalline cellulose, Serve, and carboxymethylcellulose, Sigma) and incubated for 4 h at 22°C. The resultant library was screened with a 500 bp SalI fragment of endoglucanase 1 (egl1) from Trichoderma reesei (Penttila et al., 1986), and a hybridizing clone was sequenced and designated cel1. A corresponding genomic DNA fragment for cel1 and an additional hybridizing fragment, cel2, were isolated from digested L. maculans genomic DNA. All DNA sequencing was performed using dye-primer cycle reactions and an automated sequencer (Applied Biosystems Model 373A, USA). 2.3. Genomic and chromosomal Southern analyses of pg1, cel1 and cel2 Southern blots of genomic DNA of isolate M1 were probed with a 600 bp EcoRV/PvuII fragment of the cel1 coding region, an 1100 bp fragment (generated by PCR)

A.C. Sexton et al. / Gene 248 (2000) 89–97

of the cel2 coding region, or a 1.2 kb EcoRI/XhoI fragment of the pg1 coding region, labelled with a[32P]dCTP by random priming using a MegaPrime kit (Amersham, USA). Prehybridization, hybridization and washing were as described by Sexton and Howlett (2000). Chromosomal DNA was prepared from pycnidiospores of L. maculans isolates M1 and C13, and separated using pulsed-field gel electrophoresis (Howlett, 1997). The chromosomal DNA was blotted and probed with the cDNA fragments described above. Prehybridization (2 h) and hybridization (18 h) were in modified Church Buffer (0.36 M NaH PO , 0.14 M 2 4 NaH PO , 1 mM EDTA, 7% SDS) at 65°C. Membranes 2 4 were washed twice in 2× SSC, 0.5% SDS (for cel1 or pg1) or 1× SSC, 0.5% SDS (for cel2) at 65°C, then exposed to Fuji RX film at −70°C for 1–5 days. 2.4. Plant inoculations Ten days after seed germination, cotyledons of B. napus cv. Midas and B. juncea cv. Stoke were inoculated with pycnidiospores of isolate M1 by placing a drop (10 ml containing 5×103 pycnidiospores) over a puncture (1 mm diameter) on each half of each cotyledon (Purwantara et al., 1998). Fourteen days later, cotyledons were harvested in liquid nitrogen and freeze-dried. In separate experiments, plants were sprayed with pycnidiospores (5 ml of 1×107 spores/ml per plant) 24 days after seed germination. Leaves displaying blackleg lesions were harvested in liquid nitrogen at 5, 7 or 14 days after inoculation, and freeze-dried. 2.5. Transcriptional analyses Northern analyses, using cDNA fragments of pg1, cel1 or cel2 as probes, were performed according to Sexton and Howlett (2000). Blots were stripped and re-probed with a a[32P]dCTP-labelled fragment (0.6 kb) of L. maculans cDNA encoding b-tubulin (Sexton, unpublished ) to compare the amounts of RNA loaded. Experiments were repeated at least once. Reverse transcription-PCR (RT-PCR) was used to detect transcription of pg1, cel1, and cel2 genes in total RNA from infected leaves or cotyledons, since Northern analysis did not detect these mRNAs in infected plant tissues (data not shown). The primers used are shown in Table 1, and the PCR conditions were: 5 min at 94°C, followed by 30 rounds of amplification [20 s at 94°C, 30 s at 60°C then 1 min at 72°C ]. As a positive control for cDNA quality, primers for L. maculans b-tubulin (5∞ CCGTATGATGGCCACCTTCTC 3∞ and 5∞ CTCCTGAATGGAGTCGAGTT 3∞) were used for RT-PCR. Bands were amplified from all samples due to tubulin genes being highly conserved between fungi and plants. A double band was sometimes seen — a band

91

with a size of ~640 bp was expected from L. maculans cDNA, and a slightly larger band was also observed, which may correspond to the plant tubulin (there are no reports as yet of any B. napus b-tubulin genes in GenBank, so we cannot predict the exact band size). RT-PCR experiments were duplicated using different batches of infected tissues.

3. Results and discussion 3.1. Nucleotide sequence characterization of a gene encoding endopolygalacturonase A gene with approximately 40–60% predicted amino acid identity to fungal endopolygalacturonases was isolated from L. maculans and is designated pg1. In isolate M1, this gene contains an open reading frame of 1152 bp interrupted by one intron (64 bp), encoding 384 amino acids (GenBank Accession No. AF238081). The intron contains GU 5∞ and AG 3∞ splice sites, which are highly conserved in eukaryotes (Brown and Simpson, 1998). A putative ‘TATA box’ is located 96 bp upstream of the start codon, and the sequence at the translation initiation site differs from the consensus in Neurospora crassa genes by only one base ( Edelmann and Staben, 1994). No other common promoter motifs were identified. The endopolygalacturonase gene is present as a single copy in L. maculans, as Southern analysis using individual enzymes ApaI, BamHI, HindIII, KpnI or XhoI produces a single hybridizing DNA fragment. In accordance with the nucleotide sequence, digestion of genomic DNA with SalI produces two hybridizing fragments (2.0 and 1.4 kb). The gene hybridizes to the largest chromosomal DNA band in both isolates examined (4.4 Mb in isolate M1; 3.5 Mb in isolate C13) (Fig. 1A). This size difference is not unexpected as chromosomal length polymorphisms are common in L. maculans (Howlett, 1997). A conserved motif in the predicted amino acid sequence from amino acid 233 to 246 corresponds to the putative active site [determined using the PSITE program (Solovyev and Kolchanov, 1994)]. A putative hydrophobic signal sequence of 19 amino acids is present at the N-terminal [determined using the SignalP program (Nielsen et al., 1997)]. Fungal endopolygalacturonases are secreted and contain a signal of about 20–40 amino acids. In L. maculans PG1, a putative signal for targetting to the endoplasmic reticulum is located at amino acid 189–192 (Ala–Asn–Glu–Leu), and once at the endoplasmic reticulum, N-linked glycosylation could occur at two sites: amino acid 228–231 (Asn–Ile–Thr– Phe) and 373–376 (Asn–Cys–Thr–Asn). The amino acid sequence encoded by pg1 is most similar (54% identity, 70% similarity) to endopolygalacturonase 4 from Botrytis cinerea, and all six of the B. cinerea endopolyga-

92

A.C. Sexton et al. / Gene 248 (2000) 89–97

Table 1 Primers used to amplify pg1, cel1 and cel2 of L. maculans by reverse transcription-PCR, and expected sizes of amplification products Gene

Primer (5∞ to 3∞)

Band size (bp) expected with a template of: cDNA

Cellulase cel1 Cellulase cel2 Polygalacturonase pg1

GTCATTGATGCCAACTGGCGa GCCCTTATTCGTGGGGTAC GTCATTGATGCCAACTGGCGa GAGACATTCCCTCGCTCATGG CCAGTTTGGCTGATGGCACG

Genomic DNA

400

440

990

1100

847

911

GTCCCATTTCCAGCCGCCAGTGC a The same primer was used for both cel1 and cel2 amplification, as only one base is different in the cel1 gene.

Fig. 1. Chromosomal locations of pg1 (A), cel1 (B) and cel2 (C ) in Leptosphaeria maculans isolates M1 and C13. Chromosomal DNA was separated by pulsed field gel electrophoresis in 0.9% agarose (Sigma Type II ) using a BioRad CHEF-DR II apparatus, with varied switching intervals, voltages and run-times: (B) 2800–1400 s at 50 V for 96 h, 1400–825 s at 60 V for 24 h, 825–600 s at 80 V for 24 h; (B) 180–420 s at 180 V for 24 h, 420–420 s at 100 V for 48h; (C ) 2800–1400 s at 50 V for 96 h, 1400–825 s at 60 V for 24 h, 825–600 s at 80 V for 24 h. Size of standards from Hansenula wingei (HW ) (BioRad) are included. Gels were blotted and probed with a cDNA fragment of each gene.

lacturonases characterized so far contain N-linked glycosylation signals ( Wubben et al., 1999). 3.2. Nucleotide sequence characterization of two genes encoding cellulases Two genes with a high degree of similarity to fungal exocellobiohydrolase I genes were isolated from L. maculans using part of the endoglucanase I (egl1) gene from T. reesei as a probe (Penttila et al., 1986). These genes, designated cel1 and cel2 (GenBank Accession No. AF240000 and AF240001, respectively), contain open reading frames of 1314 and 1341 bp, respectively. Both genes contain three introns, each having the conserved GU 5∞ and AG 3∞ splice sites. A putative ‘TATA box’ is located upstream of the start codon in each gene. No other common promoter motifs were identified. The cel1 gene is present as a single copy in L. maculans, as digestion of genomic DNA by individual enzymes showed single hybridizing fragments (ApaI, 7.0 kb; HindIII, 6.5 kb; KpnI, 12 kb; SalI, 2.8 kb; SpeI,

3.8 kb; XhoI, 9 kb). In accordance with the gene sequence, digestion of genomic DNA with EcoRV or HincII produced hybridizing fragments of 1.6 or 1.2 kb, respectively. Southern analysis using part of cel2 as a probe also produced single hybridizing bands, suggesting that this gene is also present as a single copy (BamHI, 12 kb; HindIII, 5.0 kb; KpnI, 7.5 kb; SalI, 6.0 kb; XhoI, 6.5 kb). As expected from the nucleotide sequence, digestion of genomic DNA with EcoRV or HincII produced hybridizing fragments of 4.0 or 2.0 kb, respectively (data not shown). Cel1 is located on a chromosome with a size of 1.85 Mb in isolate M1 and 1.8 Mb in C13 (Fig. 1B). Cel2 hybridizes to the second largest chromosomal DNA band (3.6 Mb) in isolate M1 and the largest band (3.5 Mb) in isolate C13 ( Fig. 1C ) — this latter band probably comprises two chromosomes of the same size (Howlett, unpublished ). There appears to be low-stringency cross-hybridization of cel1 to other chromosomes (including that containing cel2) in Fig. 1B, suggesting that genes with similar sequences may be present in L. maculans. Although sets

Fig. 2. Alignment of the predicted amino acid sequences of Leptosphaeria maculans CEL1 and CEL2 with other cellulases, including cellobiohydrolase of Cochliobolus carbonum (CcCEL1) (Sposato et al., 1995), exocellobiohydrolase of Cryphonectria parasitica (CpCBH1) ( Wang and Nuss, 1995), endoglucanase of Trichoderma reesei ( TrEG1) (Penttila et al., 1986) and an endoglucanase of basidiomycete Ustilago maydis ( UmEGL1) (Schauwecker et al., 1995). Highly conserved amino acids are highlighted, and the cellulose-binding domain of T. reesei EG1 is underlined.

A.C. Sexton et al. / Gene 248 (2000) 89–97 93

94

A.C. Sexton et al. / Gene 248 (2000) 89–97

of cellulase genes have been characterized from other phytopathogenic fungi, their chromosomal locations have not been reported. Electrophoretic karyotyping offers the opportunity to see whether genes are clustered on particular chromosomes. Cel1 and cel2 are not clustered in L. maculans, unlike genes encoding nitrate reductase and nitrite reductase ( Williams et al., 1995). The predicted amino acid sequences of the two genes exhibit 54% identity ( Fig. 3). The amino acid sequence predicted from cel1 shows the greatest similarity to CBH1 of Cryphonectria parasitica ( Wang and Nuss, 1995), and the amino acid sequence of CEL2 is most similar to that of Cel1 of Cochliobolus carbonum (Sposato et al., 1995) (Fig. 2). Like other fungal cellulases, both predicted proteins have a putative secretion signal sequence at the N-terminal, consisting of 17 and 20 amino acids in CEL1 and CEL2, respectively. Neither protein has a recognizable cellulose-binding domain, such as that near the C-terminal of T. reesei EG1. However, this does not suggest a lack of functionality as other fungal cellulases (such as C. parasitica CBH1 and C. carbonum Cel1) also lack such a domain, and in some cases (e.g. Fusarium oxysporum EG1), enzymatic activity has been demonstrated (Sulzenbacher et al., 1997). There is a codon bias for cytosine in the third position (present in 48% of codons in cel2, 37% in cel1, and 42% in pg1), which has been noted in cellulases and other fungal genes; for example, C. carbonum Cel1 has cytosine in the third position of 58% of codons (Sposato et al., 1995). 3.3. Regulation of pg1, cel1 and cel2 transcription in culture and during infection of oilseed Brassicas Transcription of pg1 is inducible when transferred into medium containing 1 or 3% (w/v) polygalacturonic acid or pectin as the sole carbon source, compared to a much lower level of transcription in sucrose (not shown), glycerol, or glucose, or in the absence of a carbon source (Fig. 3). Glucose does not repress the induction by polygalacturonic acid or pectin. These Northern analyses are consistent with enzyme activity studies by Annis and Goodwin (1997b) which showed induction by polygalacturonic acid (0.1%) with no repression by glucose. These authors found polygalacturonase isozymes with sizes from 45 to 66 kDa, consistent with the predicted protein size of L. maculans PG1 (approximately 42 kDa without glycosylation). Northern analysis showed that mRNA, 1.5 kb in size, hybridizing to cel1 is inducible in liquid culture after incubation for 24 h with 1% microcrystalline cellulose and highly inducible by carboxymethylcellulose, when compared with the much lower level of transcription on glycerol, glucose, or no carbon source (Fig. 4). A smaller band (0.7 kb) of unknown origin also hybridizes to the cel1 probe. The insolubility of microcrystalline cellulose

Fig. 3. Northern analysis of pg1 in Leptosphaeria maculans (isolate M1) mycelia transferred for 24 h to minimal medium containing no carbon source ( lane 1), or as the carbon source: 1% glycerol ( lane 2), 1% glucose ( lane 3), 1% polygalacturonic acid ( lane 4), 1% polygalacturonic acid plus 1% glycerol ( lane 5), 1% polygalacturonic acid plus 1% glucose ( lane 6), 1% apple pectin ( lane 7), 1% apple pectin plus 1% glucose ( lane 8). Total RNA (10 mg) was electrophoresed at 65 V in a 1.2% agarose gel containing formaldehyde and blotted onto nylon membrane, then probed for pg1. The size of hybridizing RNA is marked. The lower panel shows the same blot re-probed for b-tubulin (tub) for comparison of RNA loading.

Fig. 4. Northern analysis of cel1 in Leptosphaeria maculans (isolate M1) mycelia transferred for 24 h to minimal medium containing no carbon source ( lane 1), or as the carbon source: 1% glycerol ( lane 2), 1% glucose ( lane 3), 1% microcrystalline cellulose ( lane 4), 1% microcrystalline cellulose plus 1% glycerol ( lane 5), 1% microcrystalline cellulose plus 1% glucose ( lane 6), 1% carboxymethyl cellulose ( lane 7), 1% carboxymethyl cellulose plus 1% glucose ( lane 8). The lower panel shows the same blot re-probed for b-tubulin (tub) for comparison of RNA loading.

may be the reason for its lower potency as an inducer compared with carboxymethylcellulose, which is more soluble. Glucose represses the induction caused by cellulose or carboxymethylcellulose (Fig. 4). Cellulases of other fungi are subject to glucose repression, such as the cellulases of T. reesei (Ilmen et al., 1997) and of C. carbonum, where sucrose inhibits transcription of CEL1 (Sposato et al., 1995). As with T. reesei, glycerol appears to act as a ‘neutral’ carbon source in terms of transcriptional induction/repression of L. maculans pg1 and cel1. In contrast to cel1, transcription of cel2 was only detected at low levels in culture (on two of four Northern blots), in the presence of carboxymethylcellulose, but this may have been due to cross-hybridization to cel1 mRNA (data not shown). The much higher level of transcription of cel1 compared with cel2 in cultures containing cellulose, and conversely the transcription of cel2 but not cel1 in seedlings (14 days after inoculation, described below),

A.C. Sexton et al. / Gene 248 (2000) 89–97

95

Fig. 5. Detection of transcripts of cel1 and cel2 in L. maculans isolate M1 during infection, using reverse transcription-PCR. Lanes 1 and 17 (MW ): DNA markers (Promega). Lane 2: Infected B. napus cotyledons 14 days post-inoculation (p.i). Lane 3: infected B. juncea cotyledons 14 days p.i. Lane 4: uninfected B. napus cotyledons. Lane 5: uninfected B. juncea cotyledons. Lane 6: infected B. napus leaves 5 days p.i. Lane 7: infected B. juncea leaves 5 days p.i. Lane 8: infected B. napus leaves 7 days p.i. Lane 9: infected B. juncea leaves 7 days p.i. Lane 10: infected B. napus leaves 14 days p.i. Lane 11: infected B. juncea leaves 14 days p.i. Lane 12: uninfected B. napus leaves. Lane 13: uninfected B. juncea leaves. Lane 14: genomic DNA of B. napus. Lane 15: genomic DNA of L. maculans. Lane 16: negative control containing H O instead of template. Primers specific 2 to cel1 and cel2 amplified bands of the expected sizes from L. maculans genomic DNA ( lane 14), but only mRNA of cel2 was detected during infection, in leaves and cotyledons, 14 days after inoculation. Bands corresponding to b-tubulin (tub) are included as a positive control.

may reflect differences in the regulation and perhaps the function of these two genes. The proteins may differ in substrate specificities or access to substrates in the hostcell wall due to the shape of the enzyme. The mature peptide of CEL1 is only six amino acids smaller than CEL2, but amino acids 83–97 of CEL1 do not match the conserved amino acids in other fungal cellobiohydrolases and endoglucanases including CEL2 ( Fig. 2). Northern analysis did not detect induction of pg1, cel1, or cel2 transcription after 24 h of nitrogen- or carbon-starvation conditions (data not shown). The fungus may encounter such conditions during infection of the host plant, and other studies on phytopathogenic fungi suggest a link between N-starvation and expression of disease-related genes ( Talbot et al., 1997; Snoeijers et al., 1999). Genes such as the avirulence gene, Avr9, of tomato pathogen, Cladosporium fulvum, contain multiple copies of a sequence recognized by GATA factors, involved in the regulation of nitrogen utilization (Snoeijers et al., 1999), but these are not present in pg1, cel1, or cel2 of L. maculans. Avr9 is induced after 24 h starvation in culture, and 10 days’ post-inoculation in tomato plants (van den Ackerveken et al., 1994). Transcription of the three L. maculans genes during infection of both B. napus and B. juncea by isolate M1 was studied using RT-PCR (since Northern analysis was not sensitive enough to detect these fungal mRNAs in infected tissues). Transcription of only one gene, cel2, was detected during infection of cotyledons of B. juncea and B. napus and leaves of B. juncea (Fig. 5), and in other experiments, a faint band was also seen in B. napus leaves. The cel2 mRNA is detectable 14 days postinoculation, where lesions are spreading and hyphae are invading vascular cells (Chen and Howlett, 1996), but

not at earlier stages of infection, where the fungus is growing intercellularly ( Fig. 5). During growth through the intercellular spaces of the host tissue, the need for host-cell-wall-degrading enzymes may not be great, and also the fungal biomass at these early stages may be too low (as a proportion of RNA from total infected tissue) for detection of genes that are not highly expressed. However, previously, we have shown that a sufficient amount of infection had occurred (in these same tissue samples) to elicit host defence responses, as Brassica genes encoding chitinase class IV and pathogenesisrelated protein PR-1 were induced (Sexton and Howlett, 2000). Cel2 has the same transcription pattern in infected tissues as cyanide hydratase of L. maculans (Sexton and Howlett, 2000). Like cel2, the cyanide hydratase gene could be expected to be transcribed where degradation of host tissue occurs. Transcription of pg1 and cel1 could not be detected in these tissues after inoculation, at any of the times examined. As positive controls for the reactions, bands of the expected sizes ( Table 1) were amplified from L. maculans genomic DNA. We did not assay for transcription of pg1, cel1 and cel2 in infected stems, where polygalacturonase and cellulase activities have been detected ( Easton and Rossall, 1985). The genes may be transcribed in these later stages of disease, where invasion of the stem cells and lesion formation occurs. Our study is concerned with the infection of leaves and cotyledons as this earlier stage is probably more crucial for establishment or restriction of the fungal infection. 3.4. Conclusions 1. Three L. maculans genes, pg1, cel1 and cel2, encoding endopolygalacturonase and cellobiohydrolases, were

96

A.C. Sexton et al. / Gene 248 (2000) 89–97

cloned and sequenced. Each gene is present as a single copy in the genome. 2. These genes are not clustered in the genome. They are located on three separate chromosomes in L. maculans isolate M1, and in another isolate, C13, pg1 and cel2 hybridize to the same chromosomal band (which probably comprises two chromosomes) and cel1 is on a smaller chromosome. 3. The pg1, cel1 and cel2 genes all encode putative secretion signals (of 19, 17 and 20 amino acids, respectively) at the N-terminal of the proteins. The predicted CEL1 and CEL2 proteins do not have any recognizable cellulose-binding domains. 4. Transcription of pg1 and cel1 is induced when the putative substrate (polygalacturonic acid or cellulose, respectively) is added in culture, whilst only cel2 could be detected (by RT-PCR) in infected plants. The differences in transcription patterns between cel1 and cel2 suggest the presence of specific regulatory mechanisms in L. maculans and suggest that the cellobiohydrolase enzymes could have different functions. Transcription of cel1 is glucose-repressible, whilst induction of pg1 by polygalacturonic acid or pectin is not.

Acknowledgements We are grateful to Prof. M. Penttila for the plasmid containing part of T. reesei egl1, and Prof M. Sacristan for L. maculans isolate W4E87. We thank Dr K. Voight for testing cellulase inducibility on several substrates and Ms K. Popa for assistance with pulsed-field gel electrophoresis experiments. We thank the Australian Grains Research and Development Corporation for awarding a Junior Research Fellowship to A.C.S.

References Annis, S.L., Goodwin, P.H., 1997a. Inhibition of polygalacturonase activity produced by Leptosphaeria maculans with stem extracts of canola and the relationship of inhibition to resistance. J. Phytopathol. 145, 217–223. Annis, S.L., Goodwin, P.H., 1997b. Production and regulation of polygalacturonase isozymes in Canadian isolates of Leptosphaeria maculans differing in virulence. Can. J. Plant Pathol. 19, 358–365. Boudart, G., Lafitte, C., Barthe, J.P., Frasez, D., Esquerre-Tugaye, M.-T., 1998. Differential elicitation of defense responses by pectic fragments in bean seedlings. Planta 206, 86–94. Brown, J.W.S., Simpson, C.G., 1998. Splice site selection in plant premRNA splicing. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49, 77–95. Chen, C.Y., Howlett, B.J., 1996. Rapid necrosis of guard cells is associated with the arrest of fungal growth in leaves of Indian mustard (Brassica juncea) inoculated with avirulent isolates of Leptosphaeria maculans. Physiol. Mol. Plant Pathol. 48, 73–81. Easton, C.J., Rossall, S., 1985. The production of certain cell-wall degrading enzymes by Leptosphaeria maculans in culture and in

stem canker lesions of oilseed rape. Physiol. Plant Pathol. 26, 185–197. Edelmann, S.E., Staben, C., 1994. A statistical analysis of sequence features within genes from Neurospora crassa. Exp. Mycol. 18, 70–81. Gao, S., Choi, G.H., Shain, L., Nuss, D.L., 1996. Cloning and targetted disruption of enpg-1, encoding the major in vitro extracellular endopolygalacturonase of the chestnut blight fungus, Cryphonectria parasitica. Appl. Environ. Microbiol. 62, 1984–1990. Hassan, A.K., Schulz, C., Sacristan, M.D., Woestemeyer, J., 1991. Biochemical and molecular tools for the differentiation of aggressive and non-aggressive isolates of the oilseed rape pathogen, Phoma lingam. J. Phytopathol. 131, 120–136. Howlett, B.J., 1997. Genome analysis of the fungal plant pathogen Leptosphaeria maculans using pulsed-field gel electrophoresis. Electrophoresis 18, 1544–1547. Ilmen, M., Saloheimo, A., Onnela, M.L., Penttila, M.E., 1997. Regulation of cellulase gene expression in the filamentous fungus Trichoderma reesei. Appl. Environ. Microbiol. 63, 1298–1306. Moerschbacher, B.M., Mierau, M., Graebner, B., Noll, U., Mort, A.J., 1999. Small oligomers of galacturonic acid are endogenous suppressors of disease resistance reactions in wheat leaves. J. Exp. Bot. 50, 605–612. Newton, C., Caten, C.E., 1988. Auxotrophic mutants of Septoria nodorum isolated by direct screening and by selection for resistance to chlorate. Trans. Br. Mycol. Soc. 90, 199–207. Nielsen, H., Engelbrecht, J., Brunak, S., von Heijne, G., 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Prot. Eng. 10, 1–6. Penttila, M., Lehtovaara, P., Nevalainen, H., Bhikhabhai, R., Knowles, J., 1986. Homology between cellulase genes of Trichoderma reesei: complete nucleotide sequence of the endoglucanase I gene. Gene 45, 253–263. Piel, J., Atzorn, R., Gabler, R., Kuhnemann, F., Boland, W., 1997. Cellulysin from the plant parasitic fungus Trichoderma viride elicits volatile biosynthesis in higher plants via the octadecanoid signalling cascade. FEBS Lett. 416, 143–148. Purwantara, A., Salisbury, P.A., Burton, W.A., Howlett, B.J., 1998. Reaction of Brassica juncea (Indian mustard ) lines to Australian isolates of Leptosphaeria maculans under glasshouse and field conditions. Eur. J. Plant Pathol. 104, 895–902. Schauwecker, F., Wanner, G., Kahmann, R., 1995. Filament-specific expression of a cellulase gene in the dimorphic fungus Ustilago maydis. Biol. Chem. Hoppe-Seyler 376, 617–625. Sexton, A.C., Howlett, B.J., 2000. Characterisation of a cyanide hydratase gene in the phytopathogenic fungus Leptosphaeria maculans. Mol. Gen. Genet. in press Snoeijers, S.S., Vossen, P., Goosen, T., van den Broek, H.W.J., De Wit, P.J.G.M., 1999. Transcription of the avirulence gene Avr9 of the fungal tomato pathogen Cladosporium fulvum is regulated by a GATA-type transcription factor in Aspergillus nidulans. Mol. Gen. Genet. 261, 653–659. Solovyev, V.V., Kolchanov, N.A., 1994. Search for functional sites using consensus. In: Kolchanov, N.A., Lim, H.A. ( Eds.), Computer Analysis of Genetic Macromolecules. World Scientific, Singapore, pp. 16–21. Sposato, P., Ahn, J.H., Walton, J.D., 1995. Characterization and disruption of a gene in the maize pathogen Cochliobolus carbonum encoding a cellulase lacking a cellulose binding domain and hinge region. Mol. Plant–Microbe Interact. 8, 602–609. Sulzenbacher, G., Schulein, M., Davies, G.J., 1997. Structure of the endoglucanase I from Fusarium oxysporum: Native, cellobiose and ˚ reso3,4-epoxybutyl b--cellobioside-inhibited forms, at 2.3 A lution. Biochemistry 36, 5902–5911. Talbot, N.J., McCafferty, H.R.K., Ma, M., Moore, K., Hamer, J.E., 1997. Nitrogen starvation of the rice blast fungus Magnaporthe

A.C. Sexton et al. / Gene 248 (2000) 89–97 grisea may act as an environmental cue for disease symptom expression. Physiol. Mol. Plant Pathol. 50, 179–195. van den Ackerveken, G.F.J.M., Dunn, R.M., Cozijnsen, A.J., Vossen, J.P.M.J., van den Broek, H.W.J., de Wit, P.J.G.M., 1994. Nitrogen limitation induces expression of the avirulence gene avr9 in the tomato pathogen Cladosporium fulvum. Mol. Gen. Genet. 243, 277–285. Wang, P., Nuss, D.L., 1995. Induction of a Cryphonectria parasitica cellobiohydrolase I gene is suppressed by hypovirus infection and regulated by a GTP-binding-protein-linked signalling pathway

97

involved in fungal pathogenesis. Proc. Natl. Acad. Sci. 92, 11529–11533. Williams, R.S.B., Davis, M.A., Howlett, B.J., 1995. The nitrate and nitrite reductase-encoding genes of Leptosphaeria maculans are closely linked and transcribed in the same direction. Gene 158, 153–154. Wubben, J.P., Mulder, W., ten Have, A., van Kan, J.A., Visser, J., 1999. Cloning and partial characterisation of endopolygalacturonase genes from Botrytis cinerea. Appl. Environ. Microbiol. 65, 1596–1602.