Dec 2002 - Jan 2003
Supervisors: Augusto Becerra & Bo Wang
Fusarium oxysporum f. sp. vasinfectum (Fov) is a fungal disease, which causes cotton (Gossypium) death through occlusion of the plants vascular system. The Fungal spores may be transported easily by soil and water enabling fungi to spread quickly via irrigation methods and vehicles. Whilst existing in major overseas cotton producing countries for many years, Fov was only recently detected in Australian cotton fields in the early 1990s. When sequenced it was found that the Fov in Australia was genetically quite diverse from its overseas counterparts, thus it is most likely of indigenous origin. It was hypothesised the pathogen may have originated in native Gossypium.
In order to test the hypothesis Bo Wang collected stems of native cottons (Gossypium australe, G. bickii, G. nelsonii and G. sturtianum). Fusarium species were found in 31 stems (3%), these were isolated and identified using morphological features. Only one of the Fusarium strains was identified as a F. oxysporum (Fo), the vast majority (25) being F. semitectum (Fs) with two F. chlamydosporum (Fc), and one each of F.dimerum (Fd), F. equiseti (Fe), and F. scirpi (Fsc).
The summer project was designed to test the 31 Fusarium isolates for their pathogenicity towards cultivated cottons (G. hirsutum). The main questions the project wanted to answer were: Can Fusarium strains from native Gossypium infect cultivated cottons? How do the fungi enter the vascular system of cultivated cottons? Once the fungi enter the vascular tissue can they spread within the plants? Where do the Fusarium isolates fit in an evolutionary sense in relation to other fungi including Fov? Are there any correlations between the isolates genetics and their pathogenicity?
There were two parts to the project these being a morphological analysis: to test for pathogenicity and a genetic analysis: to access isolates relationship to Fov and other Fusarium species.
The fungi was grown and inoculated into a Fusarium susceptible cultivar of G. hirsutum along with one set of water controls. Four inoculation methods were adopted:
After six weeks no plants displayed signs of leaf wilt or other external disease symptoms as occurs for Fov. The stem prick method was the most effective with 34% plants displaying some vascular infection. Only 0.01% plants from root dip revealed any vascular fungal activity and most of these displayed obvious wound scars around root area suggesting the fungi did not in fact enter via the roots. There was no significant difference between the stem prick and the injection methods, implying age of host plant (at least between six and ten weeks) is not important for pathogenicity of endovascular fungi. The genetic groupings as defined in the genetic analysis did not have any significant correlation with pathogenicity. The individual isolates were significantly different to one another in their pathogenicity and interestingly enough responded differently to the four inoculation methods (the probability values in a two-way ANOVA for isolates and isolate*treatment were both P<0.0001).
Seven strains were chosen for re-isolation of the endovascular fungi, representing the various genetic groupings. Whilst there was some evidence of contamination, (Penicillin was found growing in two plates!) this was minimal and overall the fungi that grew out of the stems was that with which the plants were inoculated.
Genetic analyses were run on DNA extracted from the 31 strains. Several primers were tried however, only the Internal Spacer (ITS) proved successful. Restriction Fragment Length Polymorphisms (RFLP) were run using the ITS PCR product to check for clones and group isolates. Three isolates showed distinct cut sites with the enzyme HpaII (Fd, Fo and Fsc). Two large Fs groups were apparent with both HpaII and MboI and isolate 24 was further distinguished from the main Fs group (FsGp1) with MboI. The Fc were also distinct although very similar to the smaller Fs group (FsGp2). 18 sample strains were chosen to represent the various groups apparent in the RFLPs for sequencing.
The sequences were entered into the computer and a search (BLAST) run for the best matches. The Fo isolate did not match closely with other Fo including Fov. The closest match for the Fo was F. udum suggesting it was sister to Lineage B (soil isolates from an analysis being conducted by Curt Brubaker and Bo Wang), (this was later confirmed). The best BLAST matches for the FsGp1 were samples isolated by Augusto Becerra from cultivated cotton in Narrabri.
Two trees were produced from the alignment of the 18 isolates sequenced + 38 additional sequences selected from the BLAST search results. One analysis coded gaps as a "new-state" and the other as missing data; the results were extremely congruent. The phylogenetic trees were also fairly concordant with RFLP results. The two Fs groups sequenced out on quite separate branches of both phylogenetic trees. FsGp2 was more closely related to the Fc isolates, thus some revision may be needed regarding Fusarium classification, which at present is based on morphological characteristics.
From the data gathered in the project it can be concluded that the Fusarium strains isolated from native cotton plants are able to infect cultivated cotton, indeed strains similar to FsGp1 have already been isolated from cultivated cotton by A. Becerra. None of the strains isolated appear to be harmful to cultivated cotton at present and none sequenced were closely related to Fov. The Fusarium strains tested need some form of wounding in order to infect plants as they are unable to enter via roots. Pathogenicity was not found to be related to sequence patterns within the ITS region. Overall isolates could be distinguished by their significantly different responses in pathogenicity to different treatments.