Primers and plasmids
Amplification of the hygromycin cassette required primers YgF (5′-GTTGGCGACCTCGTATTGG) and HyR (3′-CTTACCACCTGCTCATCACCT). The primers CTB2-1F (5′-CGCTAGATTTAGGTGTTGGA), CTB2-2R (5′-agatgccgaccgaacaagagctgtcccccGCAATCTTTCTTCCTATGCT), CTB2-3F (5′-caatgctacatcacccacctcgctcccccCGTTTCCAAGTCCAAGATCTG), CTB2-4R (5′-CTCTTTCGTCCCTCGTATCTC), CTB2-5F (5′-AACCTCCTTTGCGTATTCTC), and CTB2-6R (5′-ATGTTTCCGAGTTCTTGATGTG) were used for the CTB2 deletion (hygromycin-overlapping sequences underlined). Finally, primers CTB2-int-F (5′-AGCATAGGAAGAAAGATTGC) and CTB2-int-R (5′-CAGATCTTGGACTTGGAAACG) served as control primers for the CTB2 gene. Primers were designed using the homologous sequence from C. nicotianae (accession number DQ991505). Plasmids used in this study were pGEMT (Promega, Mannheim, Germany) for TA cloning of PCR fragments, pII99DsRed , and pGEMThyg (Le Thi Thu Giang, University of Hamburg, Germany).
Fungal and bacterial strains
Escherichia coli strain XL1-blue (Stratagene, La Jolla, CA, USA) was used for molecular cloning. The C. beticola isolates Ahlburg and Ferrara were kindly provided by Planta GmbH, Einbeck, Germany. Ahlburg is known to be the more virulent isolate (unpublished results).
C. beticola was cultured in complete liquid media  for DNA extraction, and on complete media agar plates for maintenance. Transformants were grown under selective pressure with 50 μg/ml hygromycin B (Duchefa, Netherlands) and 100 μg/ml geneticin (Invitrogen, Germany), respectively. Toxin production was carried out on potato dextrose agar (PDA) plates (pH 5.6) under ambient light. Bacteria were cultured in Luria-Bertani liquid media or on Luria-Bertani plates with 200 mg/ml ampicillin (Sigma Aldrich, Munich, Germany). MS medium for the in vitro culture of sugar beets was prepared as described .
CTB2 was disrupted by double homologous recombination using fusion PCR to generate a construct consisting of a 1219-bp 3′ part of the gene, the hygromycin resistance cassette, and a 796-bp 5′ part of the gene (Figure 1A). The hygromycin cassette was released from pGEMThyg with SmaI, while the 5´- and 3´-parts of CTB2 were amplified from genomic DNA with primers CTB2-1F, CTB2-2R, CTB2-3F, and CTB2-4R. In the fusion PCR, equal amounts of the upper fragment, hygromycin cassette, and lower fragment were fused, and the resulting 3.7-kb fragment was amplified with primers CTB2-nestF and CTB2-nestR. The final PCR product was cloned into pGEMT and released with NotI and ApaI.
Protoplasts were prepared as described previously , and polyethylene glycol-mediated transformation was carried out as previously described . Buffers were prepared as described previously . Selection with the antibiotic hygromycin B (Duchefa, Harlem, Netherlands) was carried out with 50 μg/ml. Retransformation of hygromycin resistant transformants was performed using geneticin (Duchefa, Harlem, Netherlands) selection at a final concentration of 100 μg/ml.
Genomic DNA from wild type and gene-disruption strains was digested overnight with BspHI (New England Biolabs, Frankfurt am Main, Germany). The digested DNA was separated on a 0.8% agarose gel at 80-100 V. DNA was transferred by capillary blotting to a Hybond NX membrane (Amersham Biosciences, Little Chalfont, UK) and then hybridized with a digoxygenin-labeled (Roche, Mannheim, Germany) DNA hygromycin probe. Detection and visualization procedures were followed according to the manufacturer’s manual (Roche).
Cercosporin was extracted with 5 N KOH as described by Chung . Plates were prepared from PDA at pH 5.6, evenly inoculated with C. beticola conidia, and maintained under daylight conditions and room temperature for two weeks. The agar was cut into small squares and covered with KOH overnight in a glass beaker. Agar pieces were removed with a small household sieve. The cercosporin contents of the ΔCTB2 strains of Ahlburg and Ferrara, as well as the wild type strains and the ectopic transformants were measured after extracting toxin from PDA plates cultured for two weeks in ambient light. Each measurement was repeated three times from three independent extractions. A 5 mM cercosporin standard was kindly obtained from Planta GmbH. Absorbance was measured in a spectrometer (Ultrospec 3000, Pharmacia Biotech) at 480 nm and the cercosporin content was calculated from the molecular weight of 534.51 g/mol (Sigma). For testing on plants, cercosporin was extracted from PDA plates directly with water, which worked as well as extraction with KOH. Plants were dipped in plate extract containing approximately 750 μM cercosporin (isolate Ahlburg) and 320 μM (isolate Ferrara).
Sugar beet plants were grown in a growth chamber at 18°C with 16 hours of light. To produce conidia for the plant inoculation, PDA plates (pH 5.6) were equally seeded with mycelium mashed in a Waring blender and incubated under daylight for two weeks. Roughly one plate was inoculated for each plant to be infected. The conidia were then carefully scraped off the surface with sterile water and a spatula. The scraped material was filtered through a small household sieve, one layer of miracloth, and a 200 μm Wilson sieve to completely remove all agar plucks. The conidia were then counted in a Fuchs-Rosenthal hemocytometer. Twelve-week-old sugar beet plants were inoculated by thoroughly applying 50 ml of a conidia suspension of 2.0*104 conidia/ml onto the adaxial and abaxial leaf surfaces with a spray bottle. Using 50 ml the leaves of one plant were sufficiently covered. Twenty-five plants were inoculated with each fungal strain. The plants were covered with a foil tent for ten days and exposed to 18 hours of light (2.0*104 lux, 400-600 nm = 4.0-6.0*103 K) at temperatures of 24°C for day and 18°C for night.
For testing the toxin on plants grown in vitro, one-week-old in vitro-cultured plants were dipped into cell-free plate extract or the water control.
Leaves inoculated with the wild type and Δctb2 strains and strains expressing DsRed were investigated with MZ FL III microscope (Leica Microsystems, Heerbrugg, Switzerland). The microscope was equipped with a Leica 1.0 × objective and Leica DFC 500 fluorescence camera. To visualize plant necroses under white light conditions reflected light of an external halogen lamp KL 1500 Electronic (Schott, Mainz, Germany) was used. The DsRed fluorescence was detected with the Leica DsRed excitation filter at 546/12 nm and a long pass filter at 560 nm. The LAS Leica software (version 2.7.1) was used for image acquisition and procession.
High resolution fluorescence microscopy was performed with Zeiss Axio Imager.Z1 microscope equipped with a Zeiss Apotome and AxioCamMRm CCD camera. A UV (ultra violet) lamp HAL 100 served as light source. DsRed was excited in the range of 538 to 562 nm and detected in the 570 to 640 nm range. The plant apoplast was excited in the range of 335 to 383 nm and its blue autofluorescence was detected at 420 to 470 nm. Image processing, including overlay of independently detected DsRed and plant autofluorescence as well as generation of maximum intensity projections (MIP) of z-stacks were done with Zeiss AxioVision software (version 4.8.1). All presented images are MIPs of the respective z-stack.
Disease ratings were performed according to a method that relates the amount of leaf spots with disease severity . Single leaves were rated according to the following disease index: absence of necrotic areas (leaf spots) = 0, necrotic area < 1% = 1, necrotic area 2-5% = 2, necrotic area 6-10% = 3, necrotic area 11-20% = 4, necrotic area 21-40% = 5, necrotic area 41-60% = 6, necrotic area 61-80% = 7, necrotic area 81-100% = 8, and leaf dead = 9.