Published online: 23 September 2002, doi:10.1038/nbt743 October 2002 Volume 20 Number 10 pp 1035 - 1039

 
 
Engineering hypervirulence in a mycoherbicidal fungus for efficient weed control

Ziva Amsellem¹, Barry A. Cohen^{1, 2} & Jonathan Gressel¹
 
 

1. Plant Sciences, Weizmann Institute of Science, Rehovot, Israel 76100.
2. current address: CBD Technologies, Ltd., Park Tamar, Rehovot, Israel.
Correspondence should be addressed to J Gressel. e-mail: jonathan.gressel@weizmann.ac.il
 

Most successful strategies have involved the biotechnological control of alien weeds by introducing the pathogens or insects that kept the wild species in balance in its center of origin ("classical biocontrol"). In contrast, most of the important weeds of major row crops have become globally distributed, considerably evolved from their wild cohorts at the centers of origin, and cannot be controlled by classical methods. Pathogens have been isolated that prey on many of these major global weeds. They provide a modicum of control when very high levels of inoculum are applied (inundative biocontrol), usually >10⁴ spores/cm², which is more than four orders of magnitude greater than required for 100% efficiency of the initial application. Thus, most proposed organisms have not been sufficiently cost effective, despite efforts in efficient production and formulation of inocula^{1, 2} and in designing application technologies.

Pathogen effectiveness depends on a 100% humidity level (near the dew point) for 6–18 h after application, an environmental requirement that is difficult to meet. The major marketed organism, a strain of Colletotrichum gleosporioides (COLLEGO, Encore Technologies, Plymouth, MN), is used in rice paddies (where the humidity for establishment of the mycoherbicide is adequate) to control Aeschynomene virginica (northern jointvetch), a weed that is difficult to chemically control.

A host-specific hypervirulent pathogen that controls large and dense populations of a high-density major global weed to the extent that farmers require in row crops (i.e., similar to control achieved with a chemical herbicide) would probably have become extinct soon after evolving (unless its virulence became attenuated), as would the target weed (unless it evolved resistance); dead organisms cannot reproduce. Considering the lack of success attained with potential inundative mycoherbicides, despite years of research, it has been suggested that a solution could come from genetically engineering hypervirulence into weed-specific pathogens^{3, 4}. Requisite fail-safe mechanisms have been proposed to prevent the spread of the transformed hypervirulent organisms as well as to mitigate introgression of transgenes into related organisms.⁵.

We have recently engineered two genes, whose products are known to be safe to humans, that sequentially convert tryptophan into the plant hormone indoleacetic acid (IAA), into two Fusarium species that attack the parasitic Orobanche spp.⁶ Mutant or transgenic suppression of IAA production by pathogens leads in many cases to a loss of virulence^7-9. Our report was the first demonstration that overproduction of IAA provides a slight but statistically significant modicum of increase in virulence⁶, but not the huge magnitude of hypervirulence necessary in row-crop agriculture.

Thus, we chose to engineer NEP1 from Fusarium^{10, 11}, a gene that is largely responsible for the natural virulence of the species expressing it. The protein product Nep1 is a potent phytotoxin when injected into leaves or into the translocation stream of plants¹², or when applied to leaves along with a potent detergent to facilitate penetration¹³, or when derived from a fermenter and applied along with a non-Nep1-producing pathogen^{14, 15}. Re-engineering NEP1 into Fusarium with an enhanced promoter led to overproduction of Nep1 in the fermenter¹⁵ but did not result in hypervirulence of Fusarium¹⁶. We hypothesized that NEP1 might act differently in a heterologous organism; many Fusarium strains possess NEP1 genes that are not expressed, suggesting stringent controls on overexpression in this genus, controls that may not exist in unrelated fungi.

We demonstrate here that NEP1 increases the virulence of a strain of Colletotrichum coccodes that is pathogenic on Abutilon theophrasti, a weed that resists chemical control and is almost impossible to control in cotton, a member of the same botanical family. Cotton and Abutilon are naturally resistant to the same selective herbicides because they degrade herbicides using the same metabolic pathways. The proposed Colletotrichum wild-type mycoherbicide can sporadically kill Abutilon when applied to young seedlings bearing only cotyledons or, at most, one true leaf. It requires a week to kill these young plants¹⁷.

Verification of transformation. Separate plasmid DNA preparations containing the NEP1 gene and the hph hygromycin B–resistance genes were used to co-transform plasmid DNA into the Colletotrichum. We generated 77 hygromycin-resistant transformant colonies, of which 68 were also GUS⁺, and 28 very strongly GUS⁺. The presence of the NEP1 gene was confirmed by PCR amplification of genomic DNA fragments of three putative hygromycin B–resistant transformants having the strongest GUS expression (Fig. 1A). Western blot analyses of culture filtrates indicated that the transformed strains produced large amounts of Nep1 protein (more than 10 mg/ml of culture filtrate), while none was produced by the wild type (data not shown).

Hypervirulence of transformed strains. We tested the virulence of the transformed strains by spraying Abutilon plants at various ages with different levels of wild-type and transgenic fungus. The transgenic lines shown in Figure 1A were far more virulent than the wild type, but strain T-20 was the most potent (T-20 > T-31 > T-29 >>> wild type). All additional work reported here was with transgenic strain T-20. The Nep1 protein was identified in extracts of Abutilon cotyledons infected by T-20 NEP1 Colletotrichum, but not in the cotyledons infected by the wild type (Fig 1B). Smaller amounts of Nep1 were found at 40 h, when the diseased tissue was fully wilted.

The wild-type strain was considered for field use when Abutilon plants are in the cotyledonary stage, or at most with a single true leaf¹⁷. However, even an extremely virulent strain would not provide the farmer with an acceptable time window for application, as fields are not always accessible and weed germination occurs over a period of weeks. Farmers need a single application of a mycoherbicide or herbicide when the first weed plants to germinate reach the stage of two or three true leaves and younger plants are also present. We thus compared the NEP1 Colletotrichum with the wild type on Abutilon plants of different ages. Three days after treatment with NEP1 Colletotrichum, the treated plants through the three-leaf stage were dead and the wild type caused only severe damage to plants at the cotyledonary stage, with progressively less damage to plants having increasing numbers of leaves (Fig. 3A). The wild type succeeded in killing the seedlings in the cotyledonary stage by eight days, with some plants killed through the two-leaf stage and plants at the three-leaf stage showing some symptoms of damage (Fig. 3B). The plants with three true leaves treated with the wild type all had a remaining live leaf, and the plants recovered (Fig. 3C), demonstrating that the numerical damage assessment (Fig. 3A, B) overestimates the utility of the wild-type mycoherbicide, but not the transgenic strain. In another experiment (not shown), half of the plants treated with the wild-type strain at the two-leaf stage and 20% at the one-leaf stage recovered, yet no plants treated with the transgenic strain recovered through the three-leaf stage.

Shortening the duration of the high-humidity requirement for disease establishment is an important strategy for successful use of mycoherbicides, even though novel formulations have been helpful in allowing survival at shorter dew periods¹⁹. Experiments were done to ascertain whether the NEP1 Colletotrichum could establish itself during a shorter dew period. The results (Fig. 4) demonstrate that the hypervirulence of NEP1 Colletotrichum allows it to establish and damage the weed more quickly than the wild type. Thus, formulation of the organism for field use should be easier than with the wild type.

Transgenic NEP1 Colletotrichum quickly kills Abutilon through the stage of three true leaves at far lower doses than used previously with the wild type to control plants partially in the cotyledonary stage. Even though the transgenic Colletotrichum has the gene for the Nep1 toxin, its host specificity was not different from that of the wild-type strain on maize and cotton, but it did differentially affect other crops (Table 1). In contrast to other Colletotrichum spp.²⁰, this Colletotrichum strain does not require spores for pathogenicity; it can be formulated as mycelia. Mycelia offer the advantages that mycelial propagules are more efficiently produced than spores in commercial fermenter production of inoculum, and mycelial formulations establish infections faster than spores²¹. Mycelia use has a biosafety advantage over spores as a fail-safe mechanism: deletion mutations in sporulation can be isolated and used; such mutants cannot easily spread off target and have poor environmental persistence^{5, 22, 23}. Such deletion mutations will preclude the production of both sclerotia and conidia, the two resting forms of this species, precluding their persistence in the environment. Other genes could be added to the construct to mitigate sexual or asexual transmission of hypervirulence to related Colletotrichum crop pathogens⁵. We consider it unwise to conduct uncontained experiments with such hypervirulent organisms before these or other fail-safe mechanisms are installed and tested in containment.

Organisms and culture. The previously used^24-26 Colletotrichum coccodes (Wallr.) Hughes strain AG-90 (accession no. DAOM 183088 in the National Mycological Herbarium, Ottawa, Canada) specific to Abutilon was provided by Alan K. Watson (McGill University, Montreal, QC, Canada) and cultured on potato glucose agar (PDA) or broth (PDB).

Abutilon theophrasti Medic. seeds were collected south of Rehovot, Israel or purchased from HerbiSeed (Wokingham, UK). The germination of older seed batches was enhanced by bleaching for 1 h in 2% sodium hypochlorite on a shaker, rinsing, and soaking in water at 37°C overnight. Seeds were sown in a peat vermiculite mixture in 13 8 6 cm plastic trays and cultivated in a 25–30°C day/15–25°C night greenhouse. Except when indicated, cotyledons were inoculated before emergence of the first true leaf.

Plasmids. Two plasmids were co-transformed into the fungus: plasmid GPD1:GUS (ref. 27) expressing hygromycin resistance and gus genes was provided by Dr. D.C. Straney, University of Maryland. Plasmid pPB-FO 11-45, provided by Bryan Bailey (USDA, Beltsville MD), was previously used to express the NEP1 (necrosis and ethylene-inducing peptide) gene in fungi²⁸. Plasmid 11-45 (5,018 bp) was constructed by inserting the NEP1 gene (GenBank accession number AF036580) between two XhoI sites (-505 to +1096) cloned in to the ClaI (5') site after the trpC promoter, and the BamHI (3') site before the trpC terminator of a pTHT/Bluescript SK⁺ plasmid¹⁵.

PCR analysis of gene introgression in chromosomal DNA. PCR was used to amplify the NEP1 gene fragments in the genomic DNAs of putatively transformed strains to assay for integration into the fungal genome. Genomic DNA was extracted from mycelia according to Daboussi et al³⁰. The unique primer pairs chosen from the NEP1 gene sequence were as follows: lower, 5'-CGGCAGCAGCGTAGAGGGTAG-3'; upper, 5'-CCGACGGTTGTCAGCCATACAC-3'. The primers were prepared by the MBC Molecular Biology Center, Ltd. (Rehovot, Israel). The optimal annealing temperature for the primers of 58°C was determined using the Oligo primer analysis version 5.1 computer software program. The PCR program was as follows: double-stranded DNA denaturation at 94°C for 1 min, primer extension at 72°C for 1 min, and annealing for 1 min at 58°C for 30 cycles, in a Peltier Cycler PTC-200 (MJ Research) PCR apparatus. The PCR products were analyzed by gel electrophoresis by ethidium bromide staining. The 1.5% agarose gels were prepared with TEA buffer (40 mM Tris, 1 mM EDTA, 20 mM acetic acid—adjusted with NaOH to pH 8.0), and a 1 kb DNA ladder GeneRuler #smo331 (MBI Fermentas) was used as the size markers.

Inoculation and fungal virulence. The virulence of the transgenic fungi versus the wild type was tested as follows. Fungal inoculum was scraped from one 2 cm disk, washed free of spores, and cultured in 250 ml flasks containing 100 ml PDB. Mycelia were used to inoculate the plants because they were much more efficient in infection than spores (data not shown). The washed hyphae were suspended in 100 ml distilled water, chopped for 1 min at medium speed using an Ultra-Turrax homogenizer, collected in vacuo on Miracloth (Calbiochem, San Diego, CA), and washed with 100 ml of glass-distilled water to force any remaining spores through the filter. A 1.8 g mycelial pellet of each strain was homogenized in 60 ml water again for half a minute as above to resuspend the chopped hyphal pieces used as propagules. Tween-20 was added to attain 0.01% wt/vol, and suspensions were used as such or dilutions prepared for dose/response analyses. The propagule concentration was determined by 1:10,000 serial dilution, and 100 l aliquots were spread on PDA plates. Colonies were counted after two days of incubation at 25°C, routinely giving 8–10 10⁶ propagules/ml before dilution. A 20 ml sample of homogenate in 0.01% Tween-20 was sprayed to duplicate trays with an atomizer to runoff, each containing 13 Abutilon seedlings at the cotyledonary growth stage (unless otherwise noted). The trays with the seedlings were incubated above wet filter paper in sealed plastic boxes in the dark for a 24 h dew period, unless otherwise noted. After the period of high humidity, plants were cultured in a growth room with a 14 h photoperiod at 40 E/m²/s light, at 25°C.

The severity of damage to the cotyledons and leaves was estimated visually as follows: 0 = no infection or a hypersensitive reaction; 1 = moderate infection, plants still alive; 2 = severe infection, plants hardly alive, but one leaf may be alive and the plants recover; 3 = complete desiccation and abscission of all leaves, dead plants.

Determination of Nep1 protein from healthy and diseased Abutilon cotyledons. Proteins were extracted 24 h after inoculation (as when the sealed dew period boxes were opened) and 40 h after inoculation. Half-gram samples of infected and noninfected cotyledons were extracted with a mortar and pestle with 1 ml of 20 mM MES, 300 mM KCl, pH 5.0 buffer containing protease inhibitor cocktail (Sigma, P-9,599) and 1 mM phenyl methyl sulfonyl fluoride (PMSF). The slurries were centrifuged for 10 min at 12,000 g, and the supernatants were collected. The protein contents were determined by the Bio-Rad Bradford protein assay with BSA as a standard.

Protein samples (20 g) were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), followed by western blotting using the alkaline phosphatase conjugate (Sigma) reaction, basically as outlined by Bailey et al.¹⁰, except that TBST was used without milk, when the Nep1 antibody was used.

All experiments were repeated at least two times with similar results.