Four filamentous ascomycetes, Leptosphaeria maculans, L. biglobosa, Oculimacula yallundae and O. acuformis, were transformed via Agrobacterium tumefaciens-mediated transformation with the genes encoding DsRed and GFP. Using vectors pCAMDsRed and pCAMBgfp, either germinated conidia of Leptosphaeria spp. and O. yallundae or physically fragmented cultures of Oculimacula spp. were transformed. In vitro, the expression of the two reporter proteins in mycelium of both Oculimacula and both Leptosphaeria species was sufficient to distinguish each species in co-inoculated cultures. In planta, transformants of L. maculans or L. biglobosa expressing DsRed or GFP could be observed together in leaves of Brassica napus. Either reporter protein could be used to view the colonization of leaf petioles by both Leptosphaeria spp. and growth in the xylem vessels could be clearly observed. With the generation of these transformants, further studies on interactions between pathogen species involved in disease complexes on various host species and between opposite mating types of the same species are now possible.
Plant diseases are often caused by pathogen species complexes . When several closely related pathogens interact to produce symptoms, it is difficult to determine the contributions of each species. Two examples of global economic importance are phoma stem canker (blackleg) of oilseed rape  and eyespot of cereals . In Europe, phoma stem canker is associated with a complex of two co-existing species, Leptosphaeria maculans and L. biglobosa, although up to seven sub-species have been identified . Eyespot is associated with a complex of at least two species, recently re-named as Oculimacula yallundae (formerly Tapesia yallundae) and O. acuformis (T. acuformis) . For both pairs of related species, slight differences in epidemiology between the two species have been noted [5–7]. However, the details of the interactions between the species, in relation to environmental or agronomic factors, are difficult to study with existing methods, such as isolation onto selective culture media or species-specific PCR [6–9].
The use of the Aequorea victoria gfp gene encoding green fluorescent protein (GFP) to create transgenic strains constitutively expressing this reporter protein has, since 1996, provided an excellent tool for the in vitro and in planta observation of several plant pathogenic fungi, including O. yallundae and L. maculans[10–12]. Recently, three filamentous ascomycete species have been transformed with the Discosoma reef coral gene DsRed[13,14]. Despite a small overlap in their emission spectra, DsRed offers a true alternative to GFP and discrimination between GFP and DsRed transformants is possible using fluorescence microscopy, without the need for confocal microscopy .
To our knowledge, L. biglobosa and O. acuformis have not been transformed in the laboratory. L. maculans has been transformed using polyethylene glycol (PEG)-mediated transformation [11,15,16], electroporation  or Agrobacterium tumefaciens-mediated transformation . O. yallundae has previously been transformed only by PEG-mediated transformation [10,19–21]. GFP-expressing transformants of O. yallundae and L. maculans have previously been generated by PEG-mediated transformation [10,11] and vectors currently available to modify ascomycetes with the DsRed gene require the use of PEG-mediated transformation. However, since the enzyme preparation Novozyme 234, used for the generation of protoplasts during PEG-mediated transformation, became commercially unavailable in the late 1990s, problems with preparation of protoplasts from L. maculans and O. yallundae have been encountered . In our laboratory, it was not possible to modify L. maculans or O. yallundae using PEG-mediated transformation.
In this paper, we report the construction of a vector for Agrobacterium-mediated transformation of filamentous ascomycetes with the reef coral gene DsRed, the genetic modification of Leptosphaeria spp. and Oculimacula spp. with the genes encoding for GFP and DsRed and an evaluation of the potential use of these two reporter proteins to observe co-existing fungal pathogens in vitro and in planta.
2Materials and methods
2.1Bacterial strains and fungal isolates
Escherichia coli strain NovaBlue™ (Novagen, UK) was used during the construction and maintenance of plasmids. Electro-competent A. tumefaciens cells of strain Agl1 were prepared according to a standard protocol . The fungal isolates used in this work are maintained within the Rothamsted culture collection (curated by Paul Hornby and Martin Urban). L. maculans‘brassicae’ isolate ME24 and L. biglobosa‘brassicae’ isolate B2003-2-8 were obtained from winter oilseed rape stubble collected shortly before harvest in 2002 at Darrington, Yorkshire (ME24) or in 2003 at Orston, Nottinghamshire (B2003-2-8), using methods described ; species identification was confirmed by ITS-RFLP . Isolates were maintained on 20% V8 medium and conidia harvested from 2-week-old cultures by flooding with sterile distilled water and filtration through two layers of sterile Miracloth (Calbiochem, USA). Opposite mating-type O. yallundae isolates 22–432 (MAT1-1) and 22–433 (MAT1-2)  were obtained from P.S. Dyer (University of Nottingham, UK); O. acuformis isolate 185 was provided by G.L. Bateman (Rothamsted Research) and had been obtained from winter wheat collected in 1988 at Rothamsted. Oculimacula isolates were maintained on 20% PDA (Oxoid) medium and conidia of O. yallundae MAT1-2 were produced using methods previously described . Stock cultures of fragmented O. yallundae MAT1-2 hyphae were prepared by suspending aerial hyphae from one 9 cm culture plate in 10 ml of Czapek Dox Broth (CDB, Oxoid), physically fragmenting the mycelium for 1 min using a Polytron (Ultra-Turrax® T25 Basic, Werke GmbH) at maximum speed and incubating shaking at 200 rpm and 17 °C for 3 days.
2.2Plasmids and nucleic acid manipulations
Plasmids pGpdGFP  and pPgpd-DsRed  were obtained from B. Howlett (University of Melbourne, Victoria, Australia) and L. Mikkelsen (Royal Veterinary and Agricultural University, Frederiksberg, Denmark), respectively. Plasmid pCAMBIA1301 was purchased from CAMBIA (Canberra, Australia) and pCAMBgfp  was obtained from A. Sesma-Galarraga (John Innes Centre, Norwich, UK).
Standard molecular biology procedures were performed as previously described . For the construction of pCAMDsRed, the T-DNA of pCAMBIA1301 was modified to include the PtrpC-hph expression cassette from pGpdGFP and the PgpdA-DsRed-Express-TtrpC expression cassette from pPgpd-DsRed. Primers PtrpC-BstXI (5′ CCACCATGTTGGAGCATTTTTGGGCTTGGCTGG 3′) and hph-XhoI (5′CTCGAGCTATTCCTTTGCCCTCGGACGAG 3′) were used to amplify the PTrpC-hph coding sequence from pGpdGFP using a proof-reading polymerase (Platinum® Pfx, Invitrogen). The PCR product was blunt-end ligated into pSTBlue-1 (Novagen) and sequenced to ensure that no errors had been incorporated during PCR. The PtrpC-hph cassette was released from pSTBlue-1 by digestion with Bst XI and Xho I. pCAMBIA1301 was digested with Bst XI and Xho I to remove the existing CaMV35S promoter and hygromycin (hph) resistance gene; the PtrpC-hph cassette was then ligated into the remaining backbone of the vector to create pCAMhph. Next, the PgpdA-DsRed-Express-TtrpC expression cassette was released from pPgpd-DsRed by digestion with Eco RI and Hin dIII and blunt-ended using T4 DNA Polymerase (Promega). pCAMhph was digested with Bst EII and Bst XI (removing the other CaMV35S promoter and the GUSA gene present in pCAMBIA1301), blunt ended using T4 DNA polymerase and de-phosphorylated using Calf Intestinal Alkaline Phosphatase (MBI Fermentas). The PgpdA-DsRed-Express-TtrpC expression cassette was then ligated into the pCAMhph backbone to create pCAMDsRed (Fig. 1). Configuration of the construct was determined by restriction digestion with enzymes Eco RI, Nco I and Xho I.
Fungal genomic DNA was extracted using the CTAB protocol . To determine copy number of GFP- and DsRed-reporter genes a 543-bp GFP and a 579-bp DsRed fragment was amplified using primer pairs U61 (5′ TAAACGGCCACAAGTTCA 3′)/U62 (5′ TGCTCAGGTAGTGGTTGT 3′) and U32 (5′ CTCCTCCGAGGACGTCATCAA 3′)/U33 (5′ ACGTAGTAGTAGCCGGGCAGCT 3′). Both fragments were α32P-dCTP-labelled using the Rediprime II Random Prime Labelling System (Amersham Biosciences) following the instructions of the manufacturer. Twenty microgram aliquots of DNA of each transformant were restriction digested with EcoR I and then separated on a 1% agarose gel, blotted onto Hybond™ N+ membrane (Amersham Biosciences). Blots were incubated for 8 h in Church buffer at 65 °C  and washed twice in 1% SDS/100 mM NaHPO4 (pH 7.0) for 30 min at the same temperature. The radioactive image was visualized using a Typhoon 8600 Variable Mode Imager (Molecular Dynamics).
Agrobacterium-mediated transformation was done by a slightly modified version of an existing method . L. maculans or L. biglobosa conidia (108) were incubated with shaking at 200 rpm in 40 ml of PDB (Sigma) at 28 °C until >50% of conidia had germinated (approximately 28 h for L. maculans and 18 h for L. biglobosa). Conidial concentrations were then adjusted to 107 conidia ml−1 (including un-germinated conidia) in the induction medium (IM) , adjusted to pH 6. A 1:30 dilution of O. yallundae MAT1-2 physically fragmented mycelial stock culture or 104 conidia of O. yallundae MAT1-2 were inoculated into 10 ml of CDB. Approximately 50 or 30 cm2 of PDA colonized by 4-week-old cultures of O. yallundae MAT1-1 or O. acuformis, respectively, were transferred to 10 ml of CDB and physically fragmented. All Oculimacula cultures were incubated shaking at 200 rpm at 17 °C for 3 days, followed by re-suspension of the germinated conidia in 1 ml of IM. Alternatively, 500 μl of O. yallundae MAT1-2 mycelial culture or 100 μl of O. yallundae MAT1-1 or O. acuformis mycelial cultures were re-suspended in 1 ml of IM.
To initiate transformation, 200 μl aliquots of prepared fungal material were mixed with 200 μl of log-phase Agrobacterium Agl1 cells carrying the transforming plasmid . The cells were pelleted by centrifugation immediately after mixing, re-suspended in 50 μl of IM and applied to 3 × 3 cm pieces of sterile neutral nylon membrane (Hybond™-N, Amersham Bioscience, UK) placed on solid IM (1.5% agar). Plates were sealed, incubated at 22 °C in darkness for 3 days and the nylon membrane pieces were then transferred to selection medium (PDA containing 50 μg ml−1 hygromycin and 250 μg ml−1 cefotaxime). Alternatively, an overlay selection method  was used and transformants were selected by overlaying with PDA (or 20% V8 juice agar) containing 100 μg ml−1 hygromycin and 500 μg ml−1 cefotaxime. Plates were sealed and incubated at 22 °C in darkness until fungal colonies were visible.
A Leica MZ FLIII stereo-microscope, equipped with filters GFP1 (excitation 425/60 nm, emission 480 nm) and GFP3 (470/40 nm, 525/50 nm) from Leica Microsystems (Milton Keynes, UK) and the red shifted TRITC filter (545/30 nm, 620/60 nm) from Chroma (Rockingham, USA), was used both to screen colonies of transformants growing on nylon membranes for fluorescent protein expression and for observation of fungal colonization in plant tissues. A Zeiss Axiophot fluorescent microscope equipped with GFP fluorescence filters GFP1 (450–490 nm, LP 515 nm) and GFP3 (470/20 nm, BP 505–530 nm) from Zeiss Ltd. (Welwyn Garden City, UK) and the red shifted TRITC filter was used for observation of GFP and DsRed fluorescence in spores and hyphae in vitro and in planta. Both microscopes are equipped with a mercury lamp and photomicrographs were taken by digital camera (Leica DC 300FX) operated with IM50 software (Leica DC Twain, v. 18.104.22.168).
3Results and discussion
Four filamentous ascomycetes, Leptosphaeria maculans, L. biglobosa, Oculimacula yallundae and O. acuformis, were successfully transformed via Agrobacterium tumefaciens-mediated transformation with both GFP and DsRed, using vectors pCAMDsRed (Fig. 1) and pCAMBgfp . This is the first report that L. biglobosa and O. acuformis have been genetically modified in the laboratory. Furthermore, this is the first report of Agrobacterium-mediated transformation of filamentous ascomycetes with the DsRed gene, whereas PEG-mediated transformation has been used before [13,14]. The successful use of one Agrobacterium-mediated transformation protocol for the modification of all four species circumvents the need for PEG-mediated transformation, which often requires species-specific protocols and is also currently problematic .
Hygromycin-resistant transformants expressing either GFP or DsRed from the ToxA or PgpdA promoters, respectively, were obtained for all four species within 2–4 weeks after transformation. GFP or DsRed expression of colonies growing on nylon membranes was observed with a stereo-microscope (Fig. 1(a)–(d)). Strongly expressing colonies were removed from the membrane with a sterile tooth-pick and transferred to fresh solid media plates. The efficiency of transformation was 30–40 and 20–25 transformants per 107 spores for L. maculans and L. biglobosa, respectively. Around 10 times more L. biglobosa transformants were obtained when selection was done using a 20% V8 juice agar + hygromycin overlay, whereas only 1–2 L. biglobosa transformants were obtained per 107 spores when a PDA + hygromycin overlay was used. On the contrary, more L. maculans transformants were obtained with a PDA + hygromycin overlay. This could indicate a greater requirement of L. biglobosa for nutrients during the recovery period after transformation. Individual L. maculans transformants expressed either reporter protein moderately or strongly, whereas individual L. biglobosa transformants expressed GFP weakly and DsRed either weakly or strongly. Curiously, few L. biglobosa transformants expressed either reporter protein moderately.
Transformation of both Oculimacula spp. and both mating-types of O. yallundae was successfully achieved using either germinated conidia or cultures of physically fragmented mycelium. No preference for either selection method was observed. However, in each transformation more O. acuformis than O. yallundae transformants were obtained (Fig. 2) and the two species differed in vector efficiency; i.e, numbers of transformants obtained with vectors pCAMBgfp or pCAMDsRed. For O. yallundae, more GFP transformants were generated (GFP n=30, DsRed n=6), whereas for O. acuformis more DsRed transformants were generated (DsRed n=40, GFP n=27). The vector efficiency coincided with the strength of expression of the reporter protein; O. yallundae transformants expressed GFP strongly whereas O. acuformis transformants expressed DsRed strongly. The preference of both L. biglobosa and O. acuformis for DsRed expression could be due to greater compatibility of the PgpdA promoter, rather than preferential reporter protein synthesis. Similar strengths of GFP and DsRed expression had been observed previously in Trichoderma when the PgpdA promoter expressed either gene . To determine whether there is a species-specific preference for one of the two reporter proteins, another plasmid could be created, in which the ToxA promoter regulates the expression of the DsRed gene.
In vitro, the reporter proteins GFP and DsRed were expressed in mycelium of both Oculimacula (Fig. 3(a)–(d) and Fig. 1(u) and (v)) and both Leptosphaeria (Fig. 3(e)–(n)) species. This made it possible to distinguish each species when they were co-inoculated onto the same piece of agar (Fig. 3(a)–(d)). The GFP and DsRed reporter proteins were also expressed in conidia of all four species (Fig. 1(e)–(t)). Older mycelium occasionally included segments of hyphae with reduced expression (Fig. 3(e)) or without expression (Fig. 3(h)). Light micrographs of L. maculans hyphae (Fig. 3(g)) indicated that non-expressing hyphal sections included collapsed cells. Growing hyphal tips always expressed GFP or DsRed strongly (Fig. 3(f)) and the GFP or DsRed expression of the transformants was continuous; i.e, no spontaneous loss of expression was observed for any of the transformants of the four species, even after continuous sub-culture over a period of one year.
GFP expression was visualized using GFP1 or GFP3 filters and some DsRed expression was observed with the GFP1 filter (Fig. 3(b) and Fig. 1(f) and (j)). Although DsRed was best viewed with the TRITC filter (Fig. 3(d) and Fig. 1(h) and (l)), the possibility of using one filter for in vitro co-visualization of two pathogen species (Fig. 3(b)) or two mating types of one species would allow faster analysis of samples without the need for computer-aided overlay of images taken with different filters. The transformation of both mating-types of a fungal species with different reporter genes and the ease of co-visualization of both species in one sample creates new opportunities for studying in vitro interactions between opposite mating types of filamentous fungi, such as O. yallundae.
Transformants expressing GFP or DsRed were observed together in planta. Using either GFP1 or TRITC filters, L. maculans and L. biglobosa were observed as germinating conidia on the surface of B. napus leaves (Fig. S2(a) and (b)), during different stages in stomatal penetration (Fig. 2(b)) and colonizing leaf tissues (Fig. 3(i)–(n) and Fig. S2(c) and (d)). Pycnidia were clearly visible in heavily colonized leaf tissue with either marker protein and their formation could be followed from initiation in the sub-stomatal cavity (Fig. S2(e)–(g)). Use of the GFP1 filter also gave information on the physiological state of plant tissues near to the fungal hyphae. Chloroplasts of healthy tissue autofluoresced red, whereas senescent plant cells either autofluoresced yellow-green or did not autofluoresce (Fig. 3(i) and (l)). A co-inoculation experiment on a B. napus leaf showed that extensive cell death occurred rapidly in regions colonized by L. biglobosa, whilst cells in areas colonized by L. maculans initially remained alive (Fig. 3(i)–(k)) and only subsequently died (Fig. 3(l)–(n)). These results indicate an initial hemi-biotrophic phase of leaf tissue colonization by L. maculans. The use of GFP and DsRed labelled isolates thus provides an opportunity to investigate differences in spore germination, penetration, colonization and nutrient acquisition between closely related fungal species, such as L. maculans and L. biglobosa. Differences in any of these parameters could generate species-specific epidemiological changes under field conditions.
A key stage in the pathogenesis of phoma stem canker is the spread of the pathogens from the leaf to the stem . Colonization of the vascular bundle in oilseed rape petioles by Leptosphaeria spp. was clearly observed using transformants expressing either reporter protein (Fig. 3(o)–(s)). In this case, with a strongly expressing DsRed transformant, more informative images were obtained using the GFP1 filter (Fig. 3(r)) than the TRITC filter (Fig. 3(s)). These images confirm that hyphae of both Leptosphaeria spp. can spread from leaves to stems down the xylem vessels of leaf petioles . They offer new possibilities for studying interactions between L. maculans, L. biglobosa and B. napus during this symptomless phase of pathogen growth, which may be important in development of strategies for controlling the pathogens through host resistance or fungicides .
To determine GFP- and DsRed-reporter gene integration and copy number in the four filamentous ascomycetes, genomic DNA was isolated from 18 purified Leptosphaeria and 15 Oculimacula species transformants for Southern blot analysis (Fig. 4). The genomic DNA was digested with EcoR I, which cuts once the T-DNA of GFP transformants and twice within close vicinity the T-DNA of the DsRed transformants. The presence of different sized bands shows random integration, single bands confirm single integration, while two or more bands are indicative of multi-copy integration events. Hybridization patterns show that the majority of transformants have integrated only a single copy of the reporter gene similar to A. tumefaciens transformation systems of other filamentous fungi [30,32]. The expression strength of the GFP- and DsRed-reporter genes in the transformants was not found to be linked to the copy number of these genes, but may rather be dependent on the integration locus.
We have successfully used GFP and DsRed expressing transformants for the visualization, both in vitro and in planta, of two pairs of co-existing related pathogen species involved, respectively, in the oilseed rape phoma stem canker  and cereal eyespot  disease complexes. The development of L. maculans, L. biglobosa, O. yallundae and O. acuformis transformants strongly expressing these fluorescent reporter proteins will allow further studies on interactions between these pairs of species during host colonization and suggests that other filamentous fungi may be transformed with GFP and DsRed to investigate interactions between different pathogens involved in disease complexes.
Rothamsted Research receives grant-aided support from the UK Biotechnology and Biological Sciences Research Council (BBSRC); ME received additional funds from DuPont. Barbara Howlett, Ane Sesma-Galarraga and Lisbeth Mikkelsen are thanked for the generous donation of plasmids pGpdGFP, pCAMgfp and pPgdp-DsRed. Raffaella Carzaniga and Julia Goodwin advised and helped with fluorescence microscopy. We thank Yong-Ju Huang for her help with the culturing of transgenic fungi. All work involving the construction, use and storage of genetically modified plant pathogens was done under Defra licence No. PHL 174B/4650(11/2003).
Appendix A Supplementary data
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.femsle.2005.09.041.