Genetic variability in Gibberella fujikuroi and some related species of the genus Fusarium based on random amplification of polymorphic DNA (RAPD)

Curr Genet (1995) 27:528-535 �9 Springer-Verlag 1995

K. Voigt �9 S. Schleier �9 B. Brtickner

Genetic variability in Gibberella fujikuroi and some related species of the genus Fusarium based on random amplification of polymorphic DNA (RAPD)

Received: 1 September 1994 / 17 November 1994

Abstract One of the most important rice pathogens is Fu- sarium moniliforme (perfect stage: Gibberella fujikuroi), the causal agent of the super-elongation ("bakanae") dis- ease. Thirty-seven strains of this species from different geographical regions were analyzed for their ability to pro- duce gibberellins (GA) and for genetic relatedness by ran- dom amplified polymorphic DNA (RAPD). All GA-pro- ducing isolates showed nearly identical RAPD patterns us- ing 51 oligonucleotide nona- and deca-mers as arbitrary primers. On the other hand, large differences between GA- nonproducing isolates were obtained. Comparison of the RAPD patterns with those of the tester strains of the six known mating populations (A, B, C, D, E, F) of G. fujiku- roi showed that all producer strains belong to mating pop- ulation C and all nonproducer isolates to other mating pop- ulations. Evidence for the usefulness of the RAPD tech- nique to distinguish between mating populations was pro- vided by sexual crossings. Consensus phylogenetic trees based on RAPDs were constructed by the Phylogenetic Analysis Using Parsimony (PAUP) system. In combina- tion with morphological analysis, RAPD can distinguish between different species of the genus Fusarium. These investigations may find an application in the diagnosis of unknown Fusarium spp. and in distinguishing isolates of G. fujikuroi within the section Liseola.

Key words Fusarium �9 Gibberellins �9 RAPD markers �9 PAUP analysis

K. Voigt �9 S. Schleier �9 B. Brtickner ([]) Lehrstuhl ftir Allgemeine Mikrobiologie und Mikrobengenetik, Friedrich-Schiller-Universit~it, Neugasse 24, D-07743 Jena, Germany Present address: 1 Lehrstuhl ftir Allgemeine Botanik und Mikrobiologie, Westf/ili- sche Wilhelms-Universit~t, Schlossgarten 3, D-48149 Mtinster, Germany

Communicated by P. J. G. M. de Wit

Introduction

Fungi belonging to the genus Fusarium are distributed worldwide on many economically important plants includ- ing rice, maize, sorghum, mango, pine, asparagus, pineap- ple, and sugarcane. Some strains produce significant quan- tities of secondary metabolites; (for example, gibberellins (GA) (Brtickner et al. 1989). The production of bioactive metabolites is dependent both on climatic conditions and strain specificity.

Fusarium moniliforme (perfect stage: Gibberetla fuji- kuroi), the causative agent of gibbereltin-induced "baka- nae" disease of rice (Yabuta et al. 1934), stalk rot of maize and sorghum (Klittich and Leslie 1989), and pitch canker disease of pine (Correll et al. 1992), is one of the most widespread species of the genus Fusarium. This fungus has biotechnological importance because of the production of high amounts of gibberellic acid (GA3) and other gibbe- rellins, which are important as virulence factors in the case of"bakanae" and some other fungal plant diseases. The de- gree of virulence depends on the quantity of gibberellins. In mutants of G. fujikuroi selected by restistance to the fun- gicide pefurazoate a decreased gibberellin formation leads to a decreased virulence on rice plants (Takenaka et al. 1992).

E monoliforme is placed in the section Liseola, members of which are identified by the mode of formation of microconidia and colony morphology (Nelson et al. 1983, 1990; Nelson 1992). However, the differences between some species are rather subtle and not very spe- cific.

Formae speciales and races within a species may be characterized by non-morphological parameters, such as pathogenicity tests on a range of hosts, examination of the mycotoxin profile, or the capability to produce specific secondary metabolites. Leslie et al. (1990, 1992) and Leslie (1991) described six different mating populations of G. fujikuroi, called A, B, C, D, E, and F, which appear to represent different biological species. Following the taxonomic system of Nelson (1992), members of the A, C,

F populat ions are F. moniliforme, members of both o f the B and E populations are E subgtutinans, and members o f the D population are E proliferatum (Leslie 1991).

The failure of traditional taxonomic criteria to distin- guish pathogenic and antagonistic isolates of Fusaria means that new rapid methods must be developed for ec- ological and populat ion studies (Coddington et al. 1987).

Several molecular approaches for the monitoring, iden- tification, classification, and determination o f vegetative compatibil i ty groups in Fusarium species have been de- scribed. Fekete et al. (1993) and Migheli et al. (1993) used electrophoretic karyotyping to distinguish several Fusar- ium spp. Other investigations on phytopathogenic fungi are based on restriction fragment length polymorphisms (RFLPs) (Fusarium spp.: Dickman et al. 1989; Sarfatti et al. 1991; Correll et al. 1992; Whitehead et al. 1992. Phy- tophthora isolates: F6rster and Coffey 1992. Colletotri- chum gloeosporioides: Cisar et al. 1994. Geaumanno- myces-Phialophora complex: Ward and Akrofi 1994). However, detection of polymorphisms by techniques in- volving Southern-hybridizat ion analysis is t ime-consum- ing and laborious. An alternative approach, based on the polymerase chain reaction (PCR) (Saiki et al. 1985, 1988), is the random amplification o f polymorphic D N A (RAPD) (Williams et al. 1990) or arbitrarily primed polymerase chain reaction (AP-PCR) fingerprinting (Welsh and McClel land 1990). This PCR-based assay detects D N A polymorphisms using single primers of arbitrary sequence. Many fungi have been investigated by the RAPD technique since its first application in molecular diagnostics: e.g. ar- buscular-mycorrhizal (AM) fungi (Wyss and Bonfante 1993), Aspergilli (Aufauvre-Brown et al. 1992; Megneg- neau et al. 1993), Botrytis cinerea (Btittner et al. 1994), Cladosporium fulvum (Arnau et al. 1994), CochlioboIus carbonum (Jones et al. 1993), Fusaria (Crowhurst et al. 1991; Quellet and Seifert 1993), Heterobasidion annosum (Garbelotto et al. 1993; Stenlid et al. 1994), Penicillium roqueforti (Durand et aI. 1993), Phoma lingam (Goodwin and Annis 1991; Meyer et al. 1992; Schfifer and W6s temeyer 1992), Phoma tracheiphila (Rollo et al. 1990) and Phomopsis subordinaria (Meijer et al. 1994).

In the present report, we assess this molecular technique for its ability to distinguish between GA-produc ing and -nonproducing strains of G. fujikuroi with anamorphs in the Fusarium section Liseola (Kuhlman 1982) isolated f rom different geographical origins. We wished to test the hypothesis that different mating populations might com- prise distinct clades based on R A P D profiles. Furthermore, this paper shows that molecular R A P D markers are rapid and objective criteria for differentiating between isolates of different Fusarium spp.

Materials and methods

Fungal strains. The Fusarium strains used in this study, their sour- ces, host plants, mating populations, mating types, and their ability to produce gibberellins, are listed in Tables 1 and 2.

529

Gibberellin production. For quantitative determination of the gibbe- rellin spectrum we inoculated a production medium [3.5% corn steep liquor, 0.25% (NH4)2504, 0.1% KH2PO4, 6.0% plant oil, estimated at pH 5.0] with 5.0 (v/v)% of mycelial inoculum, cultivated for 48 h in the following medium: 2.5% sucrose, 2.5% corn steep liquor, 0.05% KH2PO4, 0.05% (NH4)2SO 4, 0.7% CaCO 3. After elimination of the mycelium, the culture fluid was diluted and analyzed for gib- berellic acid (GA3) spectrofluorimetrically (Schneider 1988) and by gas chromatography and mass spectrometry (GC/MS) (Takahashi et al. 1986).

Genomic DNA isolation. Isolates were grown in 100 ml of CM liq- uid medium optimized for Fusarium spp. (Pontecorvo et al. 1953) for 3-4 days at 28~ on a rotary shaker set at 200 rpm. The myceli- um was harvested by filtration through sterile Miracloth, washed with distilled water, frozen with liquid nitrogen and lyophilized for 24 h. Lyophilized mycelial tissue was ground into a fine powder with a mortar and pestle and dispersed (in the case of DNA for use in PCR) in extraction buffer as described by Cenis et al. (1993). DNA for Southern hybridization experiments was prepared by growing cul- tures as described above but the preparations followed the protocol of Doyle and Doyle (1990).

Amplification conditions. From 10 to 25 ng of fungal genomic tar- get DNA was used for one RAPD reaction. The amplification assay contained 40 ng of a single primer oligonucleotide, 0.2 mM deoxy- nucleoside triphosphates (Pharmacia, Boehringer) and 1.0 unit of Taq DNA polymerase (HT Biotechnology, Cambridge, UK) in the buf- fer provided by the manufacturer in a total volume of 50 gl. The re- action mixtures were overlayed with one drop of mineral oil (Sig- ma) and incubated in a water bath (Autogene II, CLF Analytische Laborger~ite) for 30 cycles with the following temperature profile: 30 s at 95~ 60 s at 32~ 30 s at 72~ A 72~ incubation for 6 rain was included as the last step for the final primer extension reaction. The amplified DNA fragments were separated by electrophoresis in 2% agarose gels (Gibco, BRL) in lxTAE buffer according to Sam- brook et al. (1989). Gels were stained in a 1.0 gg/ml ethidium bro- mide solution and photographed on a UV-transilluminator. Fifty-one arbitrarily synthesized oligonucleotides were used (Operin Technol- ogies Inc., Alameda Calif., USA; van Kan, Dept. of Phytopatholo- gy, Agricultural University of Wageningen; Klein-Lankhorst et al. 1991; Schfifer and W6stemeyer 1992; W6stemeyer et al. 1992). Prim- ers were 9-10 bases in length and had a GC content of 50-60%.

Southern analysis with a RAPD product as a hybridization probe. One amplification product (C/PS), obtained with template DNA from strains of mating population C and primer P8, was recovered from the gel and purified with a procedure based on adsorption of DNA to glass particles (Gene-Clean/Pharmacia; Vogelstein and Gillespie 1979). This PCR fragment (30-50 ng) was labelled with 32p-dATP (DuPont) using a random primer labelling kit (Gibco, BRL) and used as probe in Southern hybridization experiments with 5 gg of G. fu- jikuroi i1324, m559, N63165 and m567 DNA, which were each re- stricted with 20 units of HindlII according to the recommendations of the supplier (Boehringer, Mannheim). The DNA was blotted to Hybond-N+ nylon membranes (Amersham, Braunschweig) by ca- pillary transfer (Southern 1975). Hybridization was carried out ac- cording to the recommended protocol of Amersham (Feinberg and Vogelstein 1983, 1984). The blot was washed at high stringency (0.lxSSC; 0.1% SDS, final wash at 65~

Data analysis. Similarity and relatedness among isolates were esti- mated by cladistic analysis of RAPD data. Each PCR product was treated as an unordered character with two possible states, presence or absence. PAUP 3.1.1 (Swofford and Maddison 1987) was used for bootstrapping (Felsenstein 1985) to construct phylogenetic trees based on RAPDs generated with 28 isolates of G. fujikuroi (marked in Table 1 with "1-28") and 16 RAPD primers.

Sexual compatibility and crossing procedures. Seven unclassified isolates of G. fujikuroi were crossed with the six tester strains A00102+, A00149-, B00278+, B00281-, C01993+ and C01996-

530

(noted in Table 1 by asterisks). Sexual crosses were performed on carrot agar following the methods described by Klittich and Leslie (1988). All crosses were performed at least twice. Perithecia were collected during 1-3 months after spermatization and directly used for the collection of ascospores.

Results

Gibbere l l i c -ac id assays and G A levels

Thi r ty-seven isolates of G. fujikuroi f rom rice maize and some other host plants have been inves t iga ted for their abi l i ty to produce g ibbere l l ic acid (GA3), the f inal product

of the fungal g ibbere l l in pathway, by spec t rof luor imet ry and a combina t ion of gas chromatography and mass spec- t rometry (GC/MS). Signif icant di f ferences were observed in g ibber i l l in product ion among the strains (Table 1). Those strains in which G A 3 could not be detected spectrof luor i - met r ica l ly were ana lyzed by the h ighly sensi t ive GC/MS technique. Interest ingly, all those strains were able to pro- duce low levels of G A 3. Therefore, we could not strictly di- v ide the strains into absolute producers and non-producers but rather into low- and h igh-produc ing isolates. Gibbere l - l in product ion ranged be tween 0.1 ng/1 ( low producers) , the min imum detectable G A 3 quanti ty using the GC/MS pro- cedure, and approx imate ly 2 g/1 (high producers) .

Table 1 Mating type and mating population, host plant, origin and gibberellic acid production of Gibberellafujikuroi strains used in this study

Isolate Mating Geographic Original Degree of GA 3 GA 3 number popu- origin host GA 3 (spectro- (GC/MS)

lation production fluorimetry) [ng/1] [low/high] ling/l]

Reference

i1324 ? ? m559 ? Japan N631721 ? India N638732 ? Taiwan N631653 ? Nigeria S 14 C Japan m5565 C Japan m5676 C Japan R49 C UV-mutant

of m567 M11517 C Taiwan M65838 C Thailand N636309 C Taiwan pg71~ C China 004011 ? Hungary 82[ 112 ? Germany 7148 C UV-mutant

of m567 B 1-41 a C UV-mutant

of GF-la *A0010213 A+ USA *A0014914 A- USA *K172 A+ California *K173 A- California

*K17415 A- California *K175 A+ California *B0027816 B+ Taiwan *B0028117 B- Taiwan *K142 B- India *K14318 B+ India *C01993 I9 C+ Taiwan *C019942o C+ Taiwan *C0199521 C- Taiwan *C0199622 C- Taiwan *D0050223 D+ USA *D0219324 D- USA *E0099025 E- USA *E0219226 E+ USA *F0137727 F+ USA *F0154028 F- USA

Rice Low 0 ND University of Prague Rice Low 0 ND Gordon, USA Rice Low 0 0,1 Nirenberg, Berlin Rice Low 0 0,66 Nirenberg, Berlin Rice Low 0 0,2 Nirenberg, Berlin Rice High 800 250;<106 Avalos, Seville Rice High 1320 1160;<106 Gordon, USA Rice High 1300 1300x106 Gordon, USA

High 1980 2000;<106 Brtickner, Jena

Rice High 715 560x106 Kuhlman et al. 1982 Rice High 255 240x106 Leslie et al. 1990, 1992 Rice High 220 500;<106 Nirenberg, Berlin Rice High 960 740;<106 Muromtsev, Moskau Maize High 355 415;<106 Ferenczy, Szeged Maize High 430 575;<106 Sembdner, Halle

Low 0 0.1 Brtickner, Jena

Low 0 0.2 McMillian, UK

Maize Low 0 ND Leslie et al. 1990, 1992 Maize Low 0 ND Leslieet al. 1990, 1992 Maize Low 0 ND Kuhlman et al. 1982 Sugarcane, Low 0 ND Kuhlman et al. 1982 Sorghum Maize Low 0 0,74 Kuhlman et al. 1982 Maize Low 0 ND Kuhlman et al. 1982 Sugarcane Low 0 ND Leslie et al. 1990, 1992 Sugarcane Low 0 ND Leslie et al. 1990, 1992 ? Low 0 0,2 Kuhlman et al. 1982 ? Low 0 0,1 Kuhlman et al. 1982 Rice High 670 430x106 Leslie et al. 1990, 1992 Rice High 10 25x106 Leslie et al. 1990, 1992 Rice High 650 500• Leslie et al. 1990, 1992 Rice High 650 645x106 Leslie et al. 1990, 1992 Maize Low 0 ND Leslie et al. 1990, 1992 Maize Low 0 ND Leslie et al. 1990, 1992 Maize Low 0 ND Leslie et al. 1990, 1992 Maize Low 0 ND Leslie et al. 1990, 1992 Sorghum Low 0 ND Leslie et al. 1990, 1992 Sorghum Low 0 ND Leslie et al. 1990, 1992

Letter (A-F) indicates the mating population to which a strain * mating population standard strain 1-28 mark the strains which were used in PAUP analysis ? mating population and mating type unkown ND not determined

belongs: _+ indicates mating type within the mating population.

Table 2 GA 3 production by ,,non-moniliforme" isolates of the ge- nus Fusarium recovered from maize

Species and Geographic Host GA 3 (GC/MS) code no. origin plant [ng/ml]

E culmorum 12551 Hungary maize 1.02 F. crookweIlense 1 Hungary maize 0.78 F. graminearum 1184 Hungary maize 0.41 E graminearum K63 Hungary maize 0.38

531

Until now, the ability to secrete gibberellins has been described only for rice isolates of the species G. fujikuroi. However, we were interested in analyzing isolates of other Fusarium spp. by the highly sensitive GC/MS method. The results in Table 2 show that all of the analyzed isolates are able to produce low amounts of GA 3.

Genetic characterization of G. fu j ikumi isolates by RAPD, Southern hybridization, sexual crosses and PAUP analysis

Of 51 RAPD primers employed, 37 were able to amplify distinct fragments on Fusarium DNA. These RAPD prim- ers have been used to compare a collection of 58 isolates of Fusarium. Depending on the isolate-primer combina- tion, between 3 and 22 DNA segments of various inten- sities were amplified, ranging in size from approximately 0.4 to 3.0 kb. These bands were reproducible in repeated amplifications.

For standardized RAPD analysis, 11 isolates of G. fu- jikuroi with high levels of GA 3 production and seven low- producer strains were used. All amplifications resulted in a completely different RAPD pattern between high- and low-producing strains (Fig. 1). High-producers always formed nearly identical RAPD patterns, whereas the low producers gave very heterologous amplification patterns. An exception is provided by two GA-deficient mutants, 7148 and B1-41a (Fig. 1, lanes 18 and 19, respectively), which were recovered from GA-forming wild-type strains. RAPD assay was not suitable to detect differences in high- producing wild-type (m567) and derivatised GA-deficient mutant strains 7148 and B1-41a. The GA-overproducing mutant strain R49, which was obtained from the wild-type strain m567, could also not be distinguished from its wild- type origin (Fig. 1, lane 5).

Because of the large differences among the low-produc- ing strains and between the low- and high-producer strains taken together, we suggested that these two groups are members of different mating populations. In order to con- firm this, we examined the RAPD patterns of the 14 tester strains listed in Table 1 (marked by asterisks) categorized by one of the six mating populations and compared them with the specific RAPD pattern of high- and low-produc- ing strains under identical primer conditions. Figure 2 shows the evidence of mating-population-specific patterns of the 14 standard strains. As expected, all GA high-pro- ducing strains show the typical pattern for mating popula-

Fig. 1 Random amplified polymorphic DNA from 11 isolates char- acterized by a high level of GA 3 production (lanes 2-12) and seven low-producing isolates of G. fujikuroi (lanes 13-19) using the non- amer primer P8 (5' GGA GCC CAC 3'). Isolate numbers are present above each lane. Selected molecular-size markers (100-bp and 1-kb ladder, Gibco BRL) are indicated in kilobases (kb) to the right

Fig. 2 DNA fragments amplified from genomic DNA samples from members of different mating populations using the decamer primer B15 (5' GGA GGG TGT T 3'). Numbers on the right indicate sizes (in kb) of the components of a 123-bp ladder (Gibco BRL)

tion C, whereas the low-producing strains are members of different mating populations. Isolates K 174, N63172, and N63873 show amplified fragments typical for mating pop- ulation A, whereas strain K143 resembles isolates from mating population B. The isolate N63165 shows a specific

532

amplification pattern different from all other groups (data not shown).

Two maize isolates (0040 and 82/1), of unknown mat- ing population and characterized by high GA 3 production, were analyzed to test the suitability of RAPD for the de- termination of mating populations. Interestingly, the RAPD pattern provided evidence that these strains are not members of mating popultions A, D or E, as expected for maize isolates, but showed the typical C pattern. Primer P8 generated a 0.45-kb fragment (named C/P8) in the mat- ing population-C isolates, illustrating the diagnostic char- acter of this fragment (Figs. 1 and 3). The use of 29 other primers in RAPD assays also gave mating-population-spe- cific amplification products (data not shown).

In order to determine the occurrence of mating popula- tion C-specific sequences being amplified by the RAPD assay we purified genomic DNA from two GA low-pro- ducing strains (m559, N63165), obviously belonging to other than C mating populations, and the high-producing wild-type strain m567 (C). Genomic DNA was digested with HindIII and the Southern blot was hybridized with a mating population C-specific 0.45-kb amplification prod- uct, C/P8, eluted from the agarose gel (Fig. 3, lanes con- taining C01993, C01994, C01995, C01996). Probe C/P8, which does not contain an internal HindIII-restriction site (data not shown), hybridized to two fragments in the HindIII-digested DNA of the three isolates tested (Fig. 4). Interestingly, the RAPD marker C/P8 hybridized to the DNA of m567 (mating population C) (Fig. 4, lane 3) as well as to the DNA of "non-C" isolates (Fig. 4 lanes 1 and 2).

In order to provide evidence for the correct identifica- tion of mating population for the strains N63172, N63873, m559, it324, 0040 and 82/1, each of them was mated in reciprocal crosses with the corresponding A and C tester strains of both mating types, which are known to be sexu- ally compatible isolates. Each cross was performed three times. Five out of six strains (N63873, m559, N63172, i1324, 0040) produced fertile perithecia with viable asco- spores: strains m559, i1324 and N63172 crossed with A00102 +, N63873 crossed with A00149-, and 0040 with C01993 §

For cladistic analysis, all RAPD data obtained from 28 isolates of G. fujikuroi (see Table 1) using 16 nona- and deca-mers were converted to a similarity matrix on the ba- sis of the presence or absence of all RAPD fragments. From these data, consensus phylogenetic trees were constructed by PAUP analysis (Fig. 5). Bootstrapping was applied for determining the statistical significance of cladogram branches. As expected, the cladogram grouped the mem- bers of each mating population together. The C clade also included the two unkown isolates 0040 and 82/1, although these were isolated from maize, whereas the A cluster con- tained N63172 and N63873, isolated from rice plants. N63165 formed a separated outgroup between mating po- pulations A and B. These results show that the host plant alone does not determine the mating population.

Differences in geographic proximity did not reflect re- latedness at all: B00278 (Taiwan), as an example, was most

Fig. 3 DNA fragments amplified from genomic DNA samples from members of different mating populations by using the nonamer pri- mer P8 (5' GGA GCC CAC 3'). Numbers on the right indicate sizes (in kb) of the components of a 100-bp ladder (Gibco BRL)

Fig. 4 Southern hybridization of HindIII-digested genomic DNA of three G. fujikuroi iso- lates probed with the amplifica- tion product C/P8. Lane 1, m559; lane 2, N 63165; lane 3, m567. Approximate sizes are given in kb on the left

closely related to K143 isolated in India. Strains of G. fu- jikuroi from the same location do not necessarily have the same RAPD patterns. We suppose that mating populations are widely dispersed geographically.

In the case of the members belonging to the mating population C the RAPD patterns did not show significant variability among the five different geographical origins.

Fig. 5 Dendrogram using the RAPD polymorphisms of 28 G. fujikuroi strains generated with 16 different random primers and constructed with PAUP 3.1.1 bootstrap analysis from a matrix of similarity coefficients (see text). The numbers on the rectangular cladogram indicate the branch lengths and the per- centages of relativity

8 50%

49 100%

2

'1 70% I 7 54% 3

49

25 92%

12

70 F 1~_2 t O0% 12

16 9 78

73% 1o0% [ s 3

t00% 3

8

4

1

4

1

, 41 7 4 t00% 50% 1

' :3 0 ~,to 34 2 08%

2

7

50 F - L - t00% 5

geographic origin mating population

A00102 CA, USA A A00149 CA, USA A K174 CA, USA A N63873 Taiwan (A) N63172 India (A) N63165 Nigeria ? B00278 Taiwan B B00281 Taiwan B K143 India B E00990 IL, USA E E02192 IL, USA g F01377 KS, USA F F01540 KS, USA F U0040 Hungary (C) 82/1 Germany (C) m556 Japan C m567 Japan C N63630 Taiwan C M1151 Taiwan C M6583 Thailand C pg7 China C C01993 Taiwan C C01994 Taiwan C (301995 Taiwan C C01996 Yaiwan C Sl Japan C

- - D00502 KS-Z, USA D D02193 KS-Z, USA D

533

RAPD analysis of some Fusarium spp.

In order to prove the suitability of RAPD for taxonomic classification, we examined five different species of the genus Fusarium. All 37 primers differentiated these spe-

cies. With each of the 37 nona- or deca-mer primers, spe- cies-specific patterns were obtained. Depending on the primer, from 5 to 15 (average of 10) amplification prod- ucts were generated with lengths ranging from 0.5 to 3.0 kb. The pattern of the unkown Fusarium strains i1329 and i1331 seems to be similar to those of isolates of F. cul- morum (Fig. 6). The RAPD pattern of strain N57, which was originally determined as F. decemcellulare by mor- phological means, resembles F. culmorum. The classifica- tion of this strain may have to be re-examined.

Fig. 6 Differentiation of isolates of five Fusarium spp. DNA was amplified with primer P5 (5' ACG ATC GCG G 3'). The sizes (in kb) of the molecular-weight size marker (1-kb ladder; Gibco BRL) are indicated on the left

Discussion

The study of variation within filamentous fungi has fre- quently been limited by a lack of useful genetic markers (Crowhurst et al. 1991). A powerful tool for studying vari- ation and relatedness between different species is the ran- dom amplification of genomic DNA by means of arbitrary primer sequences. RAPD assays are more convenient than RFLPs because there is no need for radioactive probes. This technique is precise, fast, relatively inexpensive, re- quires only small amounts of DNA obtained from labo- ratory cultures, and can provide markers for population studies. Random primers have been successfully used to analyse isolates of Fusarium spp. collected from all over the world.

In previous investigations, the isolates of our G. fujiku- roi collection shown in Table 1 were typed according to their ability to produce gibberellin GA 3. Two groups have been defined. One group has the potential to produce high levels of GA 3, the other one shows only low amounts of

534

this phytohormone. Non-producers have not been found. Because of GA3-production also by F. graminearum and F. crookwellense (Table 2) we propose gibberellin biosyn- thesis as a general feature in the life cycle of many fila- mentous fungi. Besides G. fujikuroi, gibberellins are im- plicated as pathogenicity factors for Sphaceloma manihot- icola (Pegg 1984), Ustilago maydis (Kahmann, personal communication), and the basidiomycetous head smut fun- gus Sporisorium reilianum of sorghum (Matheussen et al. 1990).

In conclusion, the presence of the GA-biosynthetic pathway is not unique for the rice pathogen G. fujikuroi, but only isolates of the C mating type of this species are able to produce gibberellins in high amounts.

On the basis of sexual compatibility, isolates of G. fu- jikuroi can be grouped into six distinct (inter-sterile) mat- ing populations (A-F) (Leslie et al. 1990, 1992; Leslie 1991). Host specificity and mating populations are highly correlated (Leslie et al. 1992), especially for isolates found exclusively on rice. A grouping into mating populations by sexual crossing is difficult and time consuming. Our results show that RAPD is a suitable technique for the classifica- tion of G. fujikuroi on the basis of mating-population-spe- cific markers. Garbelotto et al. (1993) obtained similar re- sults with the genetic differentiation of intersterility groups (ISG) among isolates of Heterobasidion annosum.

The G. fujikuroi strains representing the mating popu- lation C could not be distinguished by the combined pro- files of all the available RAPD primers which have been tested for general usefulness. Presumably, more RAPD primers will be necessary to reveal genetic variability within the C group. From these results we infer that all C isolates are genetically closely related, whereas there is a high degree of genetic divergence between the isolates belonging to different mating populations. Furthermore, the results suggest that only members of the mating population C are potential GA producers at a high level. In contrast, members of the other mating populations produce minimal amounts of gibberellins (ng-s) and significantly differ in RAPD patterns from members of mating population C.

The need for more primers to detect differences within a mating population demonstrates that genomic variabil- ity within a mating population is substantially lower than variability between mating populations.

The identification of mating populations and variants within mating populations of G. fujikuroi is very useful in population studies and disease management (Goodwin et al. 1991). Based on the RAPD patterns generated by the tester strains, we suggest that two GA-producing maize isolates (0040 and 82/1) are closely related to the members of mating population C. These were the only examples of maize isolates in our strain collection producing GA 3 in high amounts and showed the RAPD patterns typical for mating population C (Fig. 3). Sexual compatibility studies with the standard strains C01993+ and C01996- gave fer- tile ascospores in crosses between the strains 0040 and C01993+, thus confirming the RAPD data. The fact that both isolates have been recovered from maize but not from rice, which is regarded as the characteristic host of C-type

strains, suggests that they have a broad host spectrum and are not specialized on single host plants. Crosses between 82/1 and the two tester strains from type C nevertheless failed. Strains m559, i1324, N63172and N63873 (Fig. 5) belong to type A. Without any exception all isolates crossed with an A standard strain.

Current assays for the taxonomic classification of mem- bers belonging to the Fusarium section Liseola, such as the analytical measurement of bioactive metabolites or the in- oculation of crop seedlings in pathogenicity tests, are time- consuming and can be affected by environmental condi- tions. In contrast, the RAPD assay is rapid, independent of gene expression and is proving to be beneficial for group- ing isolates of any fungus.

Acknowledgements We express our gratitude to Dr. G. B6se (G6ttingen) for technical help in GC/MS analysis, to Dr. A. Mesterhfizy (Szeged, Hungary) for providing us with the "non-mo- niliforme" Fusaria, Dr. H. Nirenberg (Berlin) and Prof. J. L. Leslie for supplying many G. fujikuroi isolates and Prof. E Tudzynski (Miinster) for helpful discussions. We thank U. Jungehtilsing (Mtinster) for initial help and technical advice in PAUP analysis, Dr. Birch-Hirschfeld (Jena) and Dr. D. Blechschmidt (Jena) for careful- ly synthesizing the primer oligonucleotides, and M. Gerlach for ex- cellent technical assistance. We also thank Prof. R.R Oliver (Nor- wich, U.K.) for critical reading of the manuscript. This research was supported by a grant of Deutsche Forschungsgemeinschaft (Br 1245/2-1), which is gratefully acknowledged.

References

Arnau J, Housego AP, Oliver RP (1994) The use of RAPD markers in the genetic analysis of the plant pathogenic fungus Cladospor- iumfuIvum. Curr Genet 25:438-444

Aufauvre-Brown A, Cohen J, Holden DW (1992) Use of randomly amplified polymorphic DNA markers to distinguish isolates of Aspergillusfumigatus. J Clin Microbiol 30:2991-2993

Brtickner B, Blechschmidt D, Schubert B (1989) Fusarium monili- forme Sheld. - a fungus producing a broad spectrum of bioactive metabolites. Zentralbl Mikrobiol 144:3-12

Btittner R Koch F, Voigt K, Quidde T, Risch S Blaich R, Brtickner B, Tudzynski P (1994) Variations in ploidy among isolates of Bot- rytis cinerea: implications for genetic and molecular analyses. Curr Genet 25:445-450

Coddington A, Matthews PM, Cullis C, Smith KH (1987) Restric- tion-digest patterns of total DNA from different races of Fusar- ium oxysporum f. sp. pisi - an improved method for race classifi- cation. J Phytopathol 118:9-20

Cenis JL (1993) Rapid extraction of fungal DNA for PCR amplifi- cation. Nucleic Acids Res 20:2380

Correll JC, Gordon TR, McCain AH (1992) Genetic diversity in Cal- ifornia and Florida populations of the pitch canker fungus Fusar- ium subglutinans f. sp. pini. Phytopathology 82:415-420

Cisar CR, Spiegel FW, TeBeest DO, Trout C (1994) Evidence for mating between isolates of Colletotrichum gloeosporioides with different host specificities. Curr Genet 25:330-335

Crowhurst RN, Hawthorne BT, Rikkerink EHA, Templeton MD (1991) Differentiation of Fusarium solani f. sp. cucurbitae rac- es 1 and 2 by random amplification of polymorphic DNA. Curr Genet 20:391-396

Dickman MB, Partidge JE (1989) Use of molecular markers for mon- itoring fungi involved in stalk rot of corn. Theor Appl Genet 77:535-539

Durand N, Reymond P, Fevre M (1993) Randomly amplified poly- morphic DNAs assess recombination following an induced par- asexual cycle in Penicillium roqueforti. Curr Genet 24:417-420

Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12:13-15

Feinberg AR Vogelstein B (1983) A technique for radiolabelling DNA restriction endonuclease fragments to high specific activ- ity. Anal Biochem 132:6-13

Feinberg AR Vogelstein (1984) Addendum. Anal Biochem 137:266-267

Fekete C, Nagy R, Debets AJM, Hornok L (1993) Electrophoretic karyotypes and gene mapping in eight species of the Fusarium sec- tions Arthrosporiella and Sporotrichiella. Curr Genet 24:500- 504

Felsenstein J (1985) Confidence limits on phylogenies:an approach using the bootstrap. Evolution 39:783-791

F6rster H, Coffey MD (1992) Molecular characterization of Phy- tophthora isolates with non-papillate sporangia causing root rot of raspberry using mtDNA restriction fragment length polymor- phisnas. Mycol Res 96:571-577

Garbelotto M, Bruns TD, Cobb FW (1993) Differentiation of inter- sterility groups and geographic provenances among isolates of Heterobasidion annosum detected by randon amplified polymor- pbic DNA assays. Can J Bot 71:565-569

Goodwin PH, Annis SL (1991) Rapid identification of genetic vari- ation and pathotype of Leptosphaeria maculans by random am- plified polymorphic DNA assays. Appl Environ Microbiol 57: 2482-2486

Jones MJ, Dunkle LD (1993) Analysis of Cochliobolus carbonum races by PCR amplification with arbitrary and gene-specific primers. Mol Plant Pathol 83:366-370

Klein-Lankhorst RM, Vermut A, Weide R, Liharska T, Zabel P (1991) Isolation of molecular markers for tomato (L, esculentum) using random amplified polymorphic DNA (RAPD). Theor Appl Gen- et 83:108-I 14

Klittich C JR, Leslie JF (1988) Nitrate reduction mutants of Fusari- um moniliforme ( GibbereIla fujikuroi). Genetics 118:417-423

Klittich CJR, Leslie JF (1989) Chlorate-resistant, nitrate-utilizing (crn) mutants of Fusarium moniliforme (Gibberella fujikuroi). J Gen Microbiol 135:721-727

Kuhlman EG (1982) Varieties of Gibberella fujikuroi with ana- morphs in Fusarium section Liseola. Mycologia 74:759-768

Kuhlman EG, Dianis SD, Smith DK (1982) Epidemiology of pitch canker disease in a loblolly-pine seed orchard in North Carolina. Phytopathology 80:1160-1166

Leslie JF (1991 ) Mating populations in Gibberellafujikuroi (Fusar- ium section Liseola). Phytopathology 8 t : 1058-1060

Leslie JE Pearson CAS, Nelson PE, Toussoun TA (1990) Fusarium spp. from corn, sorghum, soybean fields in the central and east- ern United States. Phytopathology 80:343-350

Leslie JK Plattner RD, Desj ardins AE, Klittich C JR (1992) Fumon- isin B 1 production by strains from different mating populations of Gibberella fujikuroi (Fusarium section Liseola). Phytopathol- ogy 82:341-345

Matheussen AM, Morgan PW, Frederiksen RA (1990) Implication of gibberellins in head smut (Sporium reilianum) of Sorghum bi- color. Plant Physiol 96:537-544

Megnegneau B, Debets F, Hoekstra RF (1993) Genetic variability and relatedness in the complex group of black Aspergilli based on random amplification of polymorphic DNA. Curt Genet 23:323-329

MeijerG, Megnegneau B, Linders EGA (1994) Variability for iso- zyme, vegetative compatibility and RAPD markers in natural populations of Phomopsis subordinaria. Mycol Res 98:267-276

Meyer W, Lieckfetdt E, W6stemeyer J, B6rner T (1992) DNA fin- gerprintingfor differentiating aggressivity groups of the rape seed pathogen Leptoshaeria maculans. Mycol Res 96:651-657

Migheli Q, Berio T, Gullino ML (1993) Electrophoretic karyotypes of Fusarium spp. Exp Mycol 17:329-337

Nelson PE (1992) Taxonomy and biology of Fusarium moniliforme. Mycopathologia 117:29-36

Nelson PE, Toussoun TA, Marasas WFO (1983) Fusarium species:an illustrated manual for identification. The Pennsylvania State Uni- versity Press, University Park

Nelson PE, Burgess LW, Summerell BA (1990) Some morphologi- cal and physiological characters of Fusarium species in sections Liseola and Elegans and similar species. Mycologia 82:99-106

535

Quellet T, Seifert KA (1993) Genetic characterization of Fusarium graminearum strains using RAPD and PCR amplification. Ge- netics 83:1003-1007

Pegg GF (1984) The role of growth regulators in plant disease. In: Wood RKS, Jellis GJ (eds) Plant diseases:infection, damage and loss. Blackwell Scientific Publications, Oxford, pp 29-48

Pontecorvo G, Roper JA, Hemmons LM, MacDonald KD, Bufton AWJ (1953) The genetics of Aspergillus nidulans. Adv Genet 5:141-238

Rollo F, Salvi R, Torchia P (1990) Highly sensitive and fast detec- tion of Phoma tracheiphila by polymerase chain reaction. Appl Microbiol Biotechnol 32:572-576

Saiki RK, Scharf S, Faloona E Mullis KB, Horn GT, Erlich HA, Arn- heim N (1985) Enzymatic amplification of ~3-globin genomic se- quences and restriction-site analysis for diagnosis of sickle cell anemia. Science 230:1350-1354

Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, Mullis KB, Erlich HA (1988) Primer-directed enzymatic amplifi- cation of DNA with thermostable DNA polymerase. Science 239:487-491

Sambrook J, Fritsch EE Maniatis T (1989) Molecular cloning: a la- boratory manual. Cold Spring Harbor Laboratory. Cold Spring Harbor, New York

Sarfatti M, Abu-Abied M, Katan J, Zamir D (1991) RFLP mapping of I1, a new locus in tomato conferring resistance against Fusar- ium oxysporum f. sp. lycopersici race 1. Theor Appl Genet 82:22- 26

Sch~ifer C, W6stemeyer J (1992) Random-primer dependent PCR differentiates aggressive from non-aggressive isolates of the oil- seed rape pathogen Phoma lingam (Leptosphaeria maculans). J Phytopathol 136:124-136

Schneider G (1988) HPLC yon Gibberellinen. In: Sembdner G, Schneider G, Schreiber K (eds) Methoden zur Pfianzenhormo- nanalyse. Fischer-Verlag, Jena, pp 102-108

Southern EM (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98:503- 517

Stenlid J, Karlsson JO, H6gberg N (1994) Intraspecific genetic vari- ation in Heterobasidion annosum revealed by amplification of minisatellite DNA. Mycol Res 98:57-63

Swofford DL, Maddison WP (1987) Reconstructing ancestral char- acter states under Wagner parsimony. Math Biosci 87:199-229

Takahashi N, Yamaguchi I, Yamane H (1986) Gibberellins. In: Tak- ahashi N (ed) Chemistry of plant homones. CRC Press, Boca Ra- ton, Florida, pp 57-151

Takenaka M, Hayashi K, Ogawa T, Kimura S, Tanaka T (1992) Low- ered virulence to rice plants and decreased biosynthesis of gib- berellins in mutants of Gibberellafujikuroi selected with pefura- zoate. J Pest Sci 17:213-220

Vogelstein B, Gillespie D (1979) Preparative and analytical purifi- cation of DNA from agarose. Proc Natl Acad Sci USA 76:615-619

Ward E, Akrofi AY (1994) Identification of fungi in the Geauman- nomyces-Phialophora complex by RFLPs of PCR-amplified ri- bosomal DNAs. Mycol Res 98:219-224

Welsh J, McClelland M (1990) Fingerprinting genomes using PCR with arbitrary primers. Nucleic Acids Res 18:7213-7218

Williams JGK, Kubelik AR, Livak KJ, Rafalski JA, Tingey SV (1990)DNA polymorpbisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Res 18:6531-6535

Whitehead DS, Coddington A, Lewis BG (1992) Classification of races by DNA polymorphism analysis and vegetative compatibil- ity grouping in Fusarium oxysporum f. sp. pisi. Physiol Mol Plant Pathol 41:295-305

W6stemeyer J, Sch~ifer C, Kellner M, Weisfeld M (1992) DNA poly- morphisms detected by random primer-dependent PCR as a pow- erful tool for molecular diagnostics of plant pathogenic fungi. Adv Mol Gen 5:227-240

Wyss P, Bonfante P (1993) Amplification of genomic DNA of arbus- cular-mycorrhizal (AM) fungi by PCR using short arbitrary prim- ers. Mycol Res 97:1351-1357

Yabuta T, Kambe K, Hayashi T (1934) Biochemistry of "bakanae" fungus of rice. J. Agric Chem Soc Japan 10:1059-1068