Brassica napus is a commercially important crop worldwide and its use is quickly increasing due to its beneficial oil products and biofuel demands. Yield can be lost through infection by a fungal pathogen, Leptosphaeria maculans, the causal agent of stem canker (blackleg). An early indication of the presence of stem canker is a lesion (leaf spot) on the cotyledons or early leaves. The leaf spot stage of the disease was used in this work to ascertain if biological control agents applied both individually and in combination decreased the lesion area and also to quantify the amount of L. maculans DNA present using quantitative polymerase chain reaction (qPCR). The natural production of antibiotics by some bacteria is a commonly found form of antagonistic biological control. Bacillus amyloliquefaciens and Pseudomonas chlororaphis spp. aureofaciens 30.84 evaluated in this work both produce antibiotics and were assayed for their ability to provide control of Leptosphaeria maculans. Known active strains and field isolates of Bacillus and Pseudomonas were tested as potential biocontrol agents in vitro and then used in in planta assays. The in planta assays using bacterial isolates applied individually indicated that all the bacteria gave statistically significant control of L. maculans at the leaf spot stage. Those isolates with highest activity were further evaluated in combination, to determine if improved control of leaf spot occurred. Firstly, however, it was important to confirm the two bacteria would be compatible and antibiotics would be produced. To this aim, an in vitro assay using mutant Chromobacterium violaceum confirmed Pseudomonas chlororaphis spp. aureofaciens upregulated antibiotic production using acyl-homoserine lactones, signalling molecules. Consequently, it was vital that the Bacillus applied with it did not produce lactonase which would denature these molecules. PCR was used to confirm the enzyme was not present. It was, however, shown using in planta assays that combinations of Bacillus and Pseudomonas did not halt the infection or growth of L. maculans, but appeared to lead to increased lesion size. Colonisation of the cotyledons by the bacterial biological control agents applied onto the cotyledons was monitored by washing recovery, serial dilution, plating and colony counting along with qPCR of the DNA. All bacteria colonised successfully when applied individually. However, the populations decreased from the quantity at time zero when they were applied in combination, indicating they were unable to colonise the cotyledons successfully under those circumstances. To quantify Leptosphaeria infection, the concentration of ergosterol, a fungal sterol, was quantified to measure the colonisation of cotyledons. Concentrations were assessed using high performance liquid chromatography (HPLC). This assay was not successful no free ergosterol could be detected. This was probably due to L. maculans either having small amounts of ergosterol in its cell membranes, or having most of the ergosterol esterfied and unsuitable for quantification using this method. Polymerase chain reaction (PCR) was used to ascertain the presence of fungal hyphae within asymptomatic regions of cotyledons. It was found that the fungal DNA was detected within all areas of the cotyledon irrespective of whether the leaf spot could be seen. This result highlights the unreliability of the common method of visually assessing the presence and/or severity of L. maculans infection using leaf spot area. To monitor the populations of bacteria and the fungus in real time, DNA was extracted from the cotyledons and quantified using quantitative PCR (qPCR). The amount of L. maculans DNA isolated decreased when the BCAs were applied individually, and increased when the BCAs were applied in combination (when compared with the amount isolated from the control cotyledons). These results confirmed earlier, non-molecular assessments. To provide a benchmark for biocontrol activity, fungicides used in the control of leaf spot on oilseed rape were tested under the standard experimental conditions. Whilst control was obtained, it was not as effective as when used in the field, probably due to the formulations being optimised for field conditions. Fungicides targeted at wheat pathogens were also tested for control against L. maculans. Field application rates of these fungicides were not successful, as all damaged the epidermis of the cotyledon, resulting in death of the plant. Application of ¼ field rate still induced epidermal damage in all cotyledons except those sprayed with Q8Y78 (now called Refinzar®), where a necrotic lesion could be seen without pycnidia, at day 15 after inoculation.

In the context of increasing food insecurity, new integrated and more sustainable crop protection methods need to be developed. The phyllosphere, i.e. the leaf habitat, hosts a considerable number of microorganisms. However, only a limited number of these are pathogenic and the roles of the vast majority still remain unknown. Managing the leaf-associated microbial communities is emerging as a potential integrated crop protection strategy. This thesis reports the characterisation of the phyllosphere of barley, an economically important crop in Scotland, with the purpose of developing tools to manipulate it. Field experiments were carried out to determine the composition of the culturable bacterial phyllosphere. The leaf-associated populations were demonstrated to be dominated by bacteria belonging to the Pseudomonas genus. Two bacterial isolates, Pseudomonas syringae and Pectobacterium atrosepticum, hindered the growth of Rhynchosporium commune, the causal agent of the leaf scald, but promoted the development of powdery mildew symptoms, caused by Blumeria graminis f.sp. hordei. However, using a molecular fingerprinting technique, namely T-RFLP, the global community was shown to be significantly richer and more diverse than indicated by the culture-based methods, thus increasing the complexity of interactions taking place in the phyllosphere. Various factors were found to affect the structure of the phyllobacteria significantly. Under controlled conditions, a root-associated symbiont, Piriformospora indica, was shown to increase the plant fitness and shift the abundance of the most common bacteria. In the field, both agro-chemical treatments tested, conventional fungicide and an elicitor mixture, increased the bacterial diversity, but fungicide application resulted in a higher yield and better protection against diseases. Finally, the plant genotype also affected the phyllosphere structure. Mutations in the leaf epicuticular waxes led to significant changes in the bacterial diversity and differentially affected disease in the field. A negative correlation between bacterial diversity and scald infection was also observed.

Verticillium chlamydosporium is a fungal pathogen of eggs and females of plantparasitic nematodes. The fungus produced an alkaline serine protease in submerged culture. This enzyme, VCPI, was characterized as a class II subtilisin, based on amino acid sequenceh omology. Several of its characteristics, e.g. molecular mass (33 kDa), pI (ca 10) and broad substrate utilisation, are typical of fungal subtilisins. Although some immunological cross-reactivity existed with other enzymes of this class, an antigenic fingerprint was obtained that was distinct, even from the subtilisin that was its closest homologue based on amino acid sequence, PrI from the entomogenous fungus Metarhizium anisopliae. There was circumstantial evidence, suggesting that this fungal protease was involved in the infection of nematode eggs, which have a largely proteinaceous eggshell. First of all, the enzyme was able to remove the outer protein layer from eggs of the susceptible root-knot nematode, Meloidogyne incognita, exposing the underlying chitin layer. Scanning electron microscopy revealed that fungal hyphae on the egg surface left an imprint, presumably through enzymatic action. There was also evidence of the protease weakening the eggshell, as enzyme-treated nematode eggs were more easily lysed and infected by the fungus than those not pre-incubated in the enzyme. A polyclonal antibody against VCPI demonstrated protease production by the fungus, prior to, or concurrent with, penetration. The enzyme was associated with appressoria, i.e. fungal infection structures. In contrast to the susceptible root-knot nematode, VCPI had little impact on the egg shell of the potato cyst nematode Globodera rostochiensis. It is suggested that the limited in situ hydrolysis of G. rostochiensis egg shell proteins is a factor contributing to its relative resistance to the fungus. Regulation studies in batch culture showed that production of the protease VCPI was repressed by high carbon and nitrogen levels. Its basic regulatory mechanism was that of repression/derepression. However, the highest protease titre was obtained when M incognita eggs were present in the medium, suggesting induction by the host. Collagen and chitin were possibly responsible for this inductive effect. In conclusion, it is believed that VCPI is a protease with a dual role for V chlamydosporium. During saprotrophic growth, VCP1 would allow the fungus to scavenge nutrients from a wide range of protein sources. However, the enzyme also has a designated function in penetration of the host, which makes it a versatile tool for a fungus that can switch trophic modes during its life-cycle. The achievements of this research include the first demonstration in a nematode-attacking fungus of: -a well-characterized protease, including data on stability, kinetics and isoforms; -a subtilisin-like protease in an egg-parasitic nematophagous fungus; -a pathogenicity-related enzyme in V chlamydosporium; -a determinant of host specificity; - enzyme regulation in general, and induction by the host, in particular.