Cells tolerate exposure to cytotoxic compounds through the action of ATP‐driven efflux pumps belonging to the ATP‐binding cassette (ABC) superfamily of membrane transporters. Phytopathogenic fungi encounter toxic environments during plant invasion as a result of the plant defense response. Here we demonstrate the requirement for an ABC transporter during host infection by the fungal plant pathogen Magnaporthe grisea. The ABC1 gene was identified in an insertional mutagenesis screen for pathogenicity mutants. The ABC1 insertional mutant and a gene‐replacement mutant arrest growth and die shortly after penetrating either rice or barley epidermal cells. The ABC1‐encoded protein is similar to yeast ABC transporters implicated in multidrug resistance, and ABC1 gene transcripts are inducible by toxic drugs and a rice phytoalexin. However, abc1 mutants are not hypersensitive to antifungal compounds. The non‐pathogenic, insertional mutation in ABC1 occurs in the promoter region and dramatically reduces transcript induction by metabolic poisons. These data strongly suggest that M.grisea requires the up‐regulation of specific ABC transporters for pathogenesis; most likely to protect itself against plant defense mechanisms.
The ability to withstand toxic compounds in the natural environment is an important adaptation for all microorganisms. In particular, microorganisms often encounter an array of pre‐formed and inducible chemical barriers during their attempts to infect plants. To defend themselves against fungal pathogens, plants produce an array of antifungal proteins (Dixon and Lamb, 1990; Dixon et al., 1994) as well as specialized antibiotics called phytoalexins (for a review see Osbourn, 1996). Only pathogens which can evade these plant defense responses during early infection stages are able to survive and cause disease. Although some fungi are capable of producing phytoalexin‐detoxifying enzymes, mutants defective in these enzymes show only modest reductions in virulence (VanEtten et al., 1989, 1994). Thus, other fungal mechanisms must exist to ensure survival in plant cells. An understanding of these pathogenic mechanisms can lead to improved and informed design of antifungal compounds and genetically engineered crops.
The rice blast fungus Magnaporthe grisea causes the most economically devastating disease of cultivated rice (Ou, 1985). Genetic and physiological studies have provided an ever increasing understanding of the mechanisms of fungal attachment and pathogenicity (Talbot, 1995; Hamer et al., 1997). Penetration is mediated by the differentiation of a specialized cell called the appressorium. Once inside the plant cell, the fungus forms bulbous infection hyphae that rapidly spread to adjacent cells. Eventually, lesions develop at the infection site which yield mycelia that sporulate and release more conidia to reinitiate the disease cycle. Few genes are known to be required for invasive growth, although numerous fungal metabolites have been proposed to act as toxins (Umetsu et al., 1972; Nukina, 1988) and many rice metabolites have been proposed as phytoalexins (Akatsuka et al., 1985; Kodama et al., 1992).
To identify genes required for invasive growth and survival in plant cells, we performed an insertional mutagenesis screen using DNA‐mediated transformation and screened transformants for defects in pathogenicity by leaf sheath injection (Xu and Hamer, 1996; Lau and Hamer, 1998). Insertional mutagenesis has been used to identify genes involved in fungal morphogenesis and pathogenesis (Lu et al., 1994; Bölker et al., 1995; Dufresne et al., 1998; Sweigard et al., 1998). Leaf sheath injection obviates the need for M.grisea to infect from appressoria, and thus would be predicted to identify genes involved in plant cell invasion.
Here we describe the characterization of a novel insertional mutant of M.grisea dramatically reduced in pathogenicity. The affected gene, ABC1, encodes a protein with a high degree of relatedness to proteins of the ATP‐binding cassette (ABC) transporter family. We show that mutations in ABC1 arrest fungal growth early in pathogenicity. ABC1 mutants do not display drug sensitivity phenotypes characteristic of other mutated fungal ABC transporter genes. An analysis of the regulation of ABC1 mRNA levels in wild‐type and mutant strains shows that up‐regulation of ABC1 is essential for pathogenicity in M.grisea. Our findings provide the first evidence that specific ABC transporters may play a role in fungal pathogenicity, most likely as a defense against antimicrobial compounds produced by the host.
Isolation of ABC1
In an insertional mutagenesis screen for non‐pathogenic mutants of M.grisea wild‐type strain Guy11 (Lau and Hamer, 1998), the mutant strain TF7‐3131 was isolated as a strain with dramatically reduced pathogenicity on rice cultivar CO‐39. Random ascospore analysis indicated that the reduced pathogenicity phenotype co‐segregated with the hygromycin resistance marker conferred by the integrating plasmid (data not shown). The plasmid insertion in strain TF7‐3131 was rescued and transformed in Escherichia coli. The recovered plasmid, pRP27, contained a 3.9 kb BglII DNA fragment, which detected a restriction polymorphism between Guy11 DNA (wild‐type) and the TF7‐3131 genomic DNA digested with the enzymes BglII, KpnI, MluI, EcoRI and XhoI (Figure 1A). A subcloned fragment contained in pRP27 revealed high similarity to ATP‐binding cassettes (ABCs) from genes of the ABC transporter superfamily. This fragment was used to isolate clones from a genomic λ library and a cDNA library. DNA sequence analysis from the genomic λ clones and cDNA clones revealed a large open reading frame (ORF) for an ABC transporter in M.grisea, hereafter designated ABC1. Low stringency Southern blot hybridization demonstrated that ABC1 is a single‐copy gene (data not shown). The plasmid insertion of strain TF7‐3131 is located in the putative promoter region 718 bp upstream of the start codon of ABC1 (Figure 1B). The insertional mutation in TF7‐3131 was designated abc1‐1.
Features of the ABC1 gene
An ORF for a protein of 1619 amino acids (Figure 2) was deduced by sequence analysis. This required proposing the existence of five introns of 231, 73, 68, 72 and 85 bp in length, respectively. The presence of these five introns was verified by sequencing three independent cDNA clones. The ABC1 genomic sequence is available in the DDBJ/EMBL/GenBank database under the accession No. AF032443. The ORF is preceded by putative promoter elements, including CAAT consensus sequences at positions −524 and −477, and a TATA‐like consensus (TATAAC) at position −114.
Comparison of the deduced protein with sequences in the SWISS‐PROT protein sequence database revealed significant similarity with members of the ABC transporter superfamily. The highest degree of identity was obtained with Cdr1 from Candida albicans (Prasad et al., 1995) and Pdr5 from Saccharomyces cerevisiae (Balzi et al., 1994) using the FASTA program (Pearson and Lipman, 1988). Both proteins show 47% identity to Abc1. Hydropathy analysis (Kyte and Doolittle, 1982) and protein alignments (Figure 2A) show that Abc1 displays a structure and domain organization typical of membrane proteins of the ABC transporter superfamily. It is composed of two homologous halves, each comprising one N‐terminal hydrophilic domain followed by a C‐terminal hydrophobic domain (Figure 2B). Each hydrophilic domain contains an ABC typical of ABC transporters: the N‐terminal hydrophilic domain consists of a Walker A motif (GPPGSGCST), an ABC signature sequence (VSGGERKRVTIA) and a Walker B motif (QCWDNSTRGLD) (Walker et al., 1982). The C‐terminal ABC shows a Walker A motif (GVSGAGTTL), a Walker B motif (LFVDEPTSGLD) and a modified ABC signature sequence (NVEQRKRLTIGV) identified in fungal ABC transporters from S.cerevisiae and C.albicans (Decottignies and Goffeau, 1997). For each of the two hydrophobic domains, six transmembrane‐spanning domains were predicted using the program PREDICT PROTEIN (Rost et al., 1995) (Figure 2B).
Disruption of ABC1
The insertional mutation in strain TF7‐3131 (abc1‐1) results in a dramatic reduction in disease symptoms (see below). However, the insertion may not create a null phenotype because it occurred 718 bp upstream of the ORF. Therefore, we created an ABC1 deletion allele (abc1‐2Δ) by one‐step gene replacement (Figure 3A). We replaced 1.9 kb of the ORF of ABC1 with a gene encoding hygromycin phosphotransferase (HPH) (Carroll et al., 1994) in gene replacement vector pΔABC1. The deleted region contains almost all of the N‐terminal half of the ABC transporter (Figure 2B) including the N‐terminal ABC and the N‐terminal six transmembrane domains. pΔABC1 was linearized with NotI and transformed into strain Guy11. Transformants were selected on hygromycin‐containing medium. Ten out of 48 hygromycin‐resistant transformants that had the ABC1 gene replaced with the abc1‐2Δ allele (data not shown) were identified in a preliminary screen. ABC1 deletion transformants were confirmed by Southern blot analysis. Hybridization analysis with probe BA1.6 (Figure 3A) showed that the strain Guy11 transformants AM30 and AM36 contained the ABC1 3.0 kb EcoRI fragment (Figure 3B). In addition, transformants AM30 and AM36 contained a high molecular weight fragment of >14 kb, indicating an integration of the gene replacement vector elsewhere in the genome. Transformants AM25, AM27, AM31 and AM35 show the 1.9 kb EcoRI fragment as expected for a correct gene replacement event.
Infection assays on rice cultivar Sariceltic were performed to test the pathogenicity of ABC1 deletion mutants. Mutants TF7‐3131 (abc1‐1) and AM25 (abc1‐2Δ) are reduced in pathogenicity to the same extent when compared with control inoculations of strain Guy11 or transformed strains with an intact ABC1 gene (AM30) (Figure 4 and Table I). Pathogenicity assays of additional independently derived abc1‐2Δ strains gave identical results (data not shown).
Magnaporthe grisea strain Guy11 can cause lesions on other rice cultivars and on barley. Thus we inoculated two other host plants, rice cultivar CO‐39 and barley cultivar Golden Promise, with strains AM25 (abc1‐2Δ) and TF7‐3131 (abc1‐1). The pathogenicity defect of abc1 mutants was similar on all tested host plants (data not shown).
Resistance to rice blast is correlated with visible necrotic pinpoint lesions, called hypersensitive (Type 1) lesions (Valent et al., 1991). Microscopic examination of infected leaves did not show a high induction of Type 1 lesions (data not shown), suggesting that abc1 mutants arrest growth and die early in the pathogenicity process. However, rare large lesions were observed following inoculation with the insertional mutant (abc1‐1) and the deletion mutant (abc1‐2Δ) (see Figure 4, leaves D and E). The formation of rare large lesions is not due to rare suppressor mutations because strains obtained from these rare lesions produced phenotypes identical to the original mutants when re‐inoculated onto susceptible hosts (data not shown).
A detached leaf assay (see Materials and methods, and Figure 5) was used to see if abc1 mutants were weakly pathogenic. In this assay, more concentrated inoculations can be used to test if strains induce a hypersensitive response, or if they are weakly pathogenic. Drops of conidial spore suspensions were placed on 4×12 mm pieces of rice leaves from cultivars CO‐39 and Sariceltic. After 6 days incubation, only rice leaves inoculated with strains Guy11 and AM30 showed visible fungal growth and sporulation. Strain nn78, which is defective in appressorium formation, and mutants AM25 (abc1‐2Δ) and TF7‐3131 (abc1‐1) did not show detectable fungal growth. Furthermore, viable abc1 cultures could not be recovered from these assays (data not shown). These results demonstrate that the block in pathogenicity in abc1 mutants is caused during early steps in pathogenesis.
Table I and Figure 6 show the results of a quantitative penetration assay on detached barley leaves (see Materials and methods). The isogenic wild‐type strain penetrated barley epidermal cells and differentiated dense spreading infection hyphae within 60 h (Figure 6A and B). In contrast, the majority of abc1 mutants formed appressoria that failed to elaborate extensive infection hyphae (Figure 6C). Only in rare instances were small infection hyphae observed (Figure 6D).
The ability of M.grisea to infect its host and cause disease symptoms is a complex phenotype. Fungal conidia have to germinate, form appressoria and produce lesions. We compared growth rates and conidiation on oatmeal plates, spore attachment and the rate of appressorium formation on plastic coverslips. However, we could not detect any significant difference between Guy11 and abc1 mutants (Table I). We conclude that the majority of abc1 mutant conidia fail early in pathogenesis, most likely just after penetration.
ABC1 and drug sensitivity
Mutations in the yeast multidrug resistance (MDR) transporter proteins Pdr5 and Cdr1 result in acute sensitivity to a range of metabolic poisons and antifungal drugs (Balzi et al., 1994; Hirata et al., 1994; Sanglard et al., 1996). To determine whether ABC1 plays a role in multidrug resistance, we tested the effects of metabolic poisons and antifungal drugs on wild‐type strains and mutant strains AM25 and TF7‐3131. In Table II, sensitivity concentrations for each drug were determined using the wild‐type strain Guy11, so as to reduce appressorium formation to 50% in spore germination assays. In growth assays on agar plates, minimal inhibitory drug concentrations (MICs), sufficient to suppress fungal mycelial growth, were determined. The tested drugs included protein synthesis inhibitors (cycloheximide and chloramphenicol), sterol biosynthesis inhibitors (fenarimol, fenpropimorph, metconazole, miconazol, propiconazole tebuconazole and terbinafine), an inhibitor of protein secretion (brefeldin A), a microtubule poison (benomyl), the mutagenic agent 4‐nitroquinoline‐1‐oxide (4‐NQO) and the rice phytoalexin sakuranetin (Kodama et al., 1992). Our assays did not detect any difference in drug sensitivities between abc1 mutants and Guy11 (Table II).
Transcriptional regulation of ABC1
MDR‐type ABC transporters, such as Pdr5 or the homologous transporter Snq2 in S.cerevisiae which are involved in the efflux of xenobiotic compounds, are transcriptionally up‐regulated by exposure to metabolic poisons and by heat shock (Miyahara et al., 1996). To investigate whether ABC1 is subject to MDR regulation, we examined the effects of a variety of antifungal compounds and metabolic poisons on ABC1 mRNA levels by RNA blotting.
For these experiments, a genomic 1.8 kb SalI–KpnI DNA fragment containing the N‐terminal hydrophilic ABC domain and parts of the N‐terminal transmembrane‐spanning domains was used as a probe. The specificity of this probe was tested in a blot analysis of RNA prepared from strain AM25 to exclude the possibility of cross‐hybridization with other ABC transporters. In AM25, the corresponding 1.8 kb SalI–KpnI DNA fragment is deleted and no signals were detected in RNA prepared from this strain (data not shown).
Figure 7A shows that treatment of wild‐type cultures with the antifungal compounds miconazole and metconazole, as well as the rice phytoalexin sakuranetin and the protein synthesis inhibitor hygromycin dramatically induced ABC1 transcript levels. Stationary phase growth, dimethylsulfoxide (DMSO) treatment, heat shock or nitrogen starvation had no effect on ABC1 transcript levels (data not shown).
Interestingly, the insertion of the transforming plasmid into the 5′ region of ABC1 results in the same pathogenicity phenotype as the deletion allele. Thus loss of pathogenicity in abc1‐1 strains may be due to an inability to express or up‐regulate ABC1. To investigate this possibility, we monitored ABC1 transcript levels in TF7‐3131 and Guy11 before and following exposure to cycloheximide.
Figure 7B shows that ABC1 is transcribed at a very low level in strains Guy11 and TF7‐3131. The addition of cycloheximide results in a 200‐fold induction of ABC1 transcript levels in the wild‐type strain but only a slight (<4‐fold induction) in strain TF7‐3131. We conclude that the plasmid insertion in the abc1‐1 allele does not abolish ABC1 expression but significantly reduces ABC1 induction by metabolic poisons. Because strain TF7‐3131 has a pathogenicity phenotype indistinguishable from that of an abc1 deletion allele, we conclude that up‐regulation of ABC1 mRNA is required for pathogenicity.
Molecular analysis of an insertional mutant of the rice blast pathogen M.grisea identified the ABC transporter‐encoding gene, ABC1, as an important pathogenicity determinant. Fungal ABC transporters play important roles as efflux pumps, providing resistance to a variety of metabolic poisons (Balzi et al., 1994; Hirata et al., 1994; Sanglard et al., 1996). Studies of an insertional mutation and a deletion mutation show that ABC1 is required as an efflux pump during early stages of M.grisea pathogenesis of rice.
Abc1 has high levels of similarity to the yeast ABC transporters Pdr5 and Cdr1. Together with Abc1, these proteins and Pdr5 homologs in the filamentous fungus Aspergillus nidulans (Del Sorbo et al., 1997) share an N‐terminal modified ABC signature sequence found in fungal ABC transporters which act as efflux pumps (Decottignies and Goffeau, 1997). Like PDR5, ABC1 transcript levels are dramatically elevated in response to metabolic poisons. However, unlike the yeast transporter genes PDR5 and CDR1, promoter or deletion mutations in ABC1 do not result in acute sensitivity to metabolic poisons or antifungal compounds. This finding suggests that ABC1 may not play a general role in MDR but may have a more specialized role in pathogenicity.
Both C.albicans and S.cerevisiae contain other ABC transporters involved in MDR. The S.cerevisiae transporter, Snq2, has distinct but overlapping drug specificities with Pdr5 (Hirata et al., 1994). Mutations in the C.albicans transporter Cdr2 do not result in acute drug sensitivity; however, Δcdr1Δcdr2 strains appear to be more sensitive than Δcdr1 mutants (Sanglard et al., 1997). By analogy with C.albicans, M.grisea may contain a second ABC transporter. A partial overlap in function between this transporter and Abc1 may account for the rare lesion phenotype observed in whole plant pathogenicity assays. Experiments to test this hypothesis are in progress.
A role for ABC1 in pathogenicity
A surprising result from this study is that an insertional mutation 718 bp upstream of the start codon of ABC1 dramatically reduces the up‐regulation of ABC1 transcription after exposure to metabolic poisons and results in a pathogenicity phenotype indistinguishable from a deletion mutation. This finding argues strongly for the up‐regulation of ABC1 as an important consequence and requirement for M.grisea pathogenicity of rice.
At least two hypotheses can be proposed to account for the role of ABC1 in pathogenesis. First, Abc1 may be required to export a fungal toxin during an early stage of pathogenesis. This hypothesis seems unlikely for several reasons. Despite a variety of studies identifying low molecular weight compounds produced by M.grisea, there is no definitive evidence for a toxin in rice blast disease (Ou, 1985), and early stages of blast fungus pathogenesis are characterized by a lack of plant cell damage or necrosis. Furthermore, abc1 mutants show an abrupt and dramatic growth arrest immediately after penetration. Tox− strains of other fungal pathogens do not demonstrate this type of growth arrest. Finally, abc1 mutants are fully viable whereas transporter mutants of the toxigenic maize pathogen Cochliobulus carbonum are predicted to be lethal due to toxin accumulation (Pitkin et al., 1996). We recognize that despite these and other arguments, until the substrate for ABC1 is identified, we cannot rule out this hypthesis.
The alternative hypothesis is that Abc1 provides a defense function during early stages of pathogenesis, by acting as an efflux pump to provide resistance to antimicrobial compounds present in rice and/or barley cells. Support for this role is found in the extensive similarity of Abc1 to other fungal ABC transporters which act as MDR‐type efflux pumps, and the observation that ABC1 transcripts are up‐regulated in the presence of metabolic poisons, antifungal drugs and the rice phytoalexin, sakuranetin. Interestingly this up‐regulation is required for pathogenicity because an insertional mutation which disrupts this regulation is non‐pathogenic. Although we have yet to identify the substrates of Abc1, numerous molecules isolated from rice cells have been proposed to act as phytoalexins during rice blast disease (Akatsuka et al., 1985; Sekido et al., 1987; Kodama et al., 1992). The availability of isogenic M.grisea strains containing or lacking ABC1 may facilitate the identification and characterization of biologically relevant molecules.
ABC‐type efflux pumps have been shown to play important roles as genetic determinants of MDR in humans (Decottignies and Goffeau, 1997) and fungi (Nishi et al., 1992; Balzi et al., 1994; Sanglard et al., 1997). It has long been speculated that fungi possess efflux pumps to resist metabolic poisons present in the environment. We propose that a natural role for these fungal transporters is to permit initimate associations between plant and fungal cells. If this hypothesis is correct, then the role we have proposed here for the Abc1 transporter of M.grisea is likely to be generalized for other invasive fungal phytopathogens.
Materials and methods
Growth and manipulation of M.grisea
Strains of M.grisea used in this study are listed in Table III. Methods for propagating, transforming and manipulating M.grisea have been described previously (Crawford et al., 1986; Talbot et al., 1993; Xu and Hamer, 1996). Minimal agar was prepared as described (Harris et al., 1994).
Low stringency Southern blots were performed at 55°C. For RNA expression experiments, rapidly growing liquid cultures in CM medium were filtered through Miracloth (Calbiochem), washed with sterile water and resuspended in half the volume of fresh medium containing the respective compound. The cultures were then incubated for another 5 h, harvested again by filtration and immediately frozen in liquid nitrogen. Heat shock experiments were performed by shifting rapidly growing cultures to 37°C. Aliquots were taken after 0, 10 and 30 min and 1 h. Northern analyses were done as described by Virca et al. (1990).
For the detached leaf assay, second leaves from rice or barley were cut into 4×12 mm pieces and floated on sterile water in a 24‐well tissue culture plate (Sarstedt). Conidial suspensions were prepared in water and filtered through two layers of Miracloth. Small droplets of conidial suspensions (1×105 conidia/ml) were applied onto floated leaf pieces using a 26G1/2 needle attached to a 1 ml tuberculin syringe. Samples were stored in the dark at 25°C. After 6 days, rice leaf pieces were dried and photographed. For the penetration assay on cut barley leaves, leave samples were incubated for 60 h with fungal conidia (5×104 conidia/ml) as described above. Samples were then fixed with 1 M KOH, heated to 70°C for 30 min, stained with aniline blue, observed and photographed by epifluorescence microscopy (Hood and Shew, 1996).
Vegetative growth and conidiation rates were measured as described (Liu and Dean, 1997) using oatmeal agar plates. Spore attachment and appressorium formation rates were determined using plastic microscope coverslips (Liu and Dean, 1997; Xu et al., 1997).
Cloning and sequence analysis of ABC1
Plasmid rescue was performed as described earlier by Kuspa and Loomis (1992). The recovered plasmid, pRP27, contained a 3.9 kb genomic DNA fragment, which was used to screen genomic clones from a λGEM‐11 genomic Guy11 library and cDNA clones from a λZAPII (Stratagene) cDNA library constructed with RNA isolated from nitrogen‐starved Guy11 mycelial culture (Xu and Hamer, 1996). Because the genomic clones did not include 2.4 kb of the 3′ end of the ORF, two sets of primers AB3 (GGAATTCCATCT TGGCCGACCAGGTCA) and AB4 (GGAATTCCACTTTTCAGGAAACATACG); AB5 (CCGGTTTCATCCAGAGGC) and AB6 (GTTCGCAATGTTGCCACC), derived from the cDNA sequence, were used to amplify genomic DNA of the corresponding region. The sequence is available in the DDBJ/EMBL/GenBank database under accession No. AF032443.
Plasmids and plasmid constructions
pBA1.6 contains a 1.6 kb fragment of ABC1 (nucleotides 3026–4691 in AF032443) amplified by primers AB5 and AB6 in pGEM‐T (Stratagene). pBL10K contains a 10 kb KpnI fragment of genomic DNA comprising the upstream region and 2.5 kb of the 5′ end of the ABC1 gene in pBluescript KS+ (Stratagene). pSK1.8 contains a 5.7 kb BamHI fragment of pBL10K containing the 5′ end of the ABC1 gene in pBC KS−.
Disruption of the ABC1 gene
To generate pBC1, the 1.6 kb BamHI in pBA1.6 was cloned into pBC KS− (Stratagene). The 5′ region of ABC1 was introduced into pBC1 as a 5.5 kb SalI fragment from pBL10K to generate pBC2. The modified HPH gene (Carroll et al., 1994) was inserted between both ABC1 fragments in pBC2 as a 1.5 kb EcoRI fragment to generate the replacement vector pΔABC1. For M.grisea transformation, pΔABC1 was linearized with NotI and transformed into Guy11. Putative transformants were pre‐screened by PCR analysis of crude mini DNA extracts (Wu et al., 1997) with ABC1 primers AB1 (GCAGAAGTTGGTAAGTAGTTGCAT) and AB2 (AAGATCTGGATGGACCGAGTTAGC) for the absence of a 1.1 kb ABC1 fragment.
Sterol biosynthesis inhibitors were a gift from Dr Mark Bossie (American Cyanamid Company). The rice phytoalexin, sakuranetin, was a gift from Drs Shigeyuki Mayama (Kobe University) and Osamu Kodama (Ibaraki University). Brefeldin A was dissolved in ethanol. All other drugs were dissolved in DMSO.
Spore germination tests were performed as published (Akatsuka et al., 1985) with slight modifications. Briefly, conidial suspensions of 2×104 conidia/ml were prepared in water and filtered through Miracloth. Then, 99 μl drops of spore suspensions were placed on plastic coverslips, and different amounts of drugs were added in a 1 μl volume. Drug concentrations were determined which suppress appressorium formation rates of fungal conidia down to 50% (EC50) (Table II). For tests on agar plates, increasing amounts of drug, eventually sufficient to stop growth, were incorporated into minimal agar. Plugs (6 mm) of 10‐day‐old mycelium were placed on the agar, and radial growth of colonies was evaluated after 8–12 days. MICs of the tested drugs are given (Table II). Experiments were done in duplicate.
We thank Drs Kiichi Adachi and Jin‐Rong Xu for useful suggestions regarding experiments, and Jane Einstein for help with sequencing. We also thank all members of the Hamer laboratory for their daily support and stimulating discussions. Comments on the manuscript by several reviewers are greatly appreciated. M.U. was a recipient of a postdoctoral fellowship from the Deutsche Forschungsgemeinschaft. This work was supported in part by a National Science Foundation Grant and a Presidential Faculty Fellow Award to J.E.H.
- Copyright © 1999 European Molecular Biology Organization