Concerns regarding carbon emissions, increasing demands on water supplies and environmental pollution have meant that the European protected horticulture industry is being challenged to develop more sustainable greenhouse climate management systems. These new systems can however potentially impact on pest and disease (P & D) pressures and the efficacy of biological control agents (BCAs). This thesis aimed to use a combination of experimental work and simulation models to compare novel and traditional greenhouse climate management scenarios in Spain and the Netherlands using two model P & D systems. These were Oidium neolycopersici (powdery mildew) and its BCA, Bacillus subtilis, on tomato, and Tetranychus urticae (the two-spotted spider mite) and its BCA, Phytoseiulus persimilis, on ornamentals. Experiments showed that latent period, disease development and sporulation of Oidium neolycopersici were strongly influenced by temperatures between 10-33°C and that the control efficacy of B. subtilis was significantly influenced by temperature and humidity in the ranges 10-33°C and 50-95% RH. The functional response of P. persimilis was found to be significantly affected by ambient humidities of 57-99% RH, with predation highest at 85% RH and lowest below 76% RH. These results, in combination with existing data, were used to construct dynamic P & D models. A greenhouse climate model, based on observed temperatures in European greenhouses, was constructed to provide data on the diurnal and seasonal variation in temperature and humidity for different climate management scenarios. The predictions from the P & D models allowed climate control regimes in different greenhouses in Spain and the Netherlands to be identified, which minimised P & D pressures and maximised the efficacy of the BCAs. The implications of these findings for greenhouse climate management are discussed.

Bulb initiation is a process which is photoperiodically driven, drawing parallels with flowering. Photoperiodic flowering is well characterised at molecular and genetic levels and occurs when photoreceptors interact with the circadian clock, regulating the expression of CONSTANS (CO), which itself regulates the expression of floral pathway integrating genes such as FLOWERING LOCUS T (FT), leading to floral initiation. Two genes which regulate CO transcription are FLAVIN-BINDING, KELCH REPEAT, F-BOX (FKF1) and GIGANTEA (GI). The onion genome is very large with a high level of duplication, presenting challenges for any molecular-based study. The aim of this study was to test the hypothesis that genes controlling daylength response are conserved between the model plant Arabidopsis and onion and hence between the different end-processes bulbing and flowering. Bulbing ratios were used to measure the response of onion plants to long day (LD) and short day (SD) conditions and the reversibility of the bulbing process. It was shown that bulbing is reversible, with a delay when plants are transferred from SDs to LDs, suggesting the accumulation of an inhibitor. Diurnal expression patterns of onion genes homologous to Arabidopsis flowering time genes were examined using quantitative RT-PCR. Phylogenetic analyses were conducted to validate the identity of the homologues. Molecular and phylogenetic data suggests that an onion GIGANTEA (GI) homologue was isolated. Peaks of expression of ZT10 in LDs and ZT7 in SDs mirror the expression of Arabidopsis GI. Homologues of FKF1 and the circadian clock gene ZEITLUPE (ZTL) were also characterised. The putative FKF1 homologue showed different expression patterns in varieties exhibiting different daylength responses. These differences may contribute to the different daylength responses. A CO-like gene, which is closely related to Arabidopsis COL4, and three FT-like genes were also characterised. It appears that many of the genes controlling daylength response are conserved in onion.

Broccoli (Brassica oleracea var. italica) is a popular vegetable, known for its nutritional benefits. However, the marketability of broccoli is limited by a short shelf life. Broccoli is susceptible to rapid postharvest senescence, which causes visible head yellowing and wilting from dehydration. Visible quality loss is also accompanied by a decline in nutrients, resulting in a product with reduced postharvest nutritional value. These factors combined cause broccoli to become unmarketable, leading to severe wastage in the retail chain. Postharvest yellowing in broccoli has been shown to be controlled by genotype, as a doubled haploid population (MGDH) created from an F1 cross between GD33, a poor performing DH line (yellow in 2 days) and Mar34, a good performing DH line (staying green > 4 days) exhibited natural variation for shelf life. Therefore, to investigate the genetic control of quality in broccoli, the fixed mapping population was assessed for shelf life, morphological traits and vitamin C content and stability in replicated field trials. Visual inspections identified head yellowing, stem turgor and bud compactness as the main traits affecting the marketability of broccoli. Two methods to quantify head yellowing were also evaluated. Spectrophotometer readings were found to be more sensitive than Image J in detecting colour change, but Image J data was more reproducible. Vitamin C quantification using HPLC, confirmed that natural variation was present in the MGDH population at harvest. Vitamin C content during postharvest storage, detected by plate assays, found vitamin C to be unstable, degrading quickly after harvest. A unique broccoli x broccoli linkage map, covering ~72.9% of the B.oleracea genome, was also constructed by genotyping the MGDH population with SSR and AFLP markers. QTL analysis of the trait data positioned 48 significant QTL in the linkage map for head yellowing (4), colour co-ordinates (17), morphological traits (17), bud quality (2) and postharvest vitamin C content (3) and stability (5). The identification of QTLs associated with the above traits has provided useful information for breeders to breed for improved nutritional and quality in broccoli using marker-assisted selection (MAS). The location of QTLs has also provided targets for fine-scale mapping and for the identification of candidate genes underlying traits.

The 9-cis-epoxycarotenoid-dioxygenase (NCED) is a key regulatory enzyme in ABA biosynthesis in plants and its overexpression increases ABA levels that can increase water use efficiency (WUE). The use of the „super promoter‟ (sp) to drive an LeNCED1 transgene in tomato greatly improved WUE without affecting long-term plant growth but caused delayed seed germination and reduced rates of seedling establishment. The first aim of the present study was to generate useful, novel variation in NCED transgene expression by allowing an LeNCED1 transgene, driven by a histone H2A promoter (pH2A), to transpose to new positions in the genome. A stabilized activator element linked to a GUS marker gene (sAc) was used to allow transposition of an engineered dissociation element (Ds) containing the LeNCED1 transgene of format Ds1::pH2A::LeNCED1::Ds2 (T-DNA-Ds). In F1 (sAc × Ds) plants transposition was demonstrated by detection of empty donor site through PCR. The F2 generation was screened for stable integration of the Ds element and reduced gs. Three F3 families, namely 59F3, 102F3 and 116F3, were investigated: they showed multiple stable Ds transposition events, had low gs, and a range of growth rates. Genotype 102F5 had wild-type seed germination, higher bulk leaf ABA and xylem sap ABA and 60% higher gravimetric WUE (WUEp). However, 102F5 accumulated 32% less above ground dry weight than wild type under well-watered conditions 56 days after germination. The sAc and Ds approach allowed production of T-DNA-Ds and nptII kanamycin free 102F5 line which retained transposed Ds (Tr-Ds) elements. These progeny were cross pollinated with sAc to reactivate the transposition of the LeNCED1 transgene and should be an excellent material for generating further variation in ABA content. The second aim was to overcome the seed dormancy in high ABA, genotypes such as sp5 and sp12, by testing novel hydroxamic acid NCED inhibitors for their ability to stimulate germination and seedling establishment. Improvement in seed germination in sp12 was observed with two hydroxamic acid compounds, and an increase in seedling growth rate was also observed, although this was not statistically significant.

The phase change from vegetative to reproductive growth in Agaricus bisporus is a complex process involving changes in morphology at least in part regulated genetically and influenced by various environmental signals. This work was aimed at understanding how the morphology changes, and the specific environmental parameters are involved, and which genes show changes in transcription during the phase change process in A. bisporus. Different resources and methodologies were developed and applied to investigate this process including digital time-lapse photography, genome database assembly, design, validation and normalisation of a custom oligonucleotide gene expression microarray and analysis of microarray-generated gene expression profiles showing the response of this fungus under stimulatory and non-stimulatory environmental conditions. Key stages that occur during reproductive differentiation and development were identified and defined. It was found that temperature and the mushroom volatile, 1-octen-3-ol, act as an on/off switches as they block specific stages of the phase change while carbon dioxide acts as a quantitative regulator as high amounts of this molecule reduce the number of primordia and fruit bodies that develop. Gene expression profiles were constructed showing the changes in gene expression in peat-based A. bisporus samples, grown under commercial cultivation conditions which were designed to stimulate reproductive growth, and experimental cultivation conditions which were designed to separate out the effects of the three environmental parameters mentioned previously. It was found that 52 genes were differentially expressed in A. bisporus during the phase change from vegetative mycelium into fruit body primordia. A comparison with the gene expression profiles constructed for the experimental growth conditions, in correlation with morphological observations enabled the separation of these 52 genes into 3 clusters. One cluster contained 4 genes that are likely to be involved in the regulation of the “early” phase change, a second cluster contained 11 genes that are likely to be involved in the regulation of the “late” phase change and the third cluster contained 37 genes that are likely to be involved with physiological processes supporting the phase change.

For many plants day length is critical for the control of flowering time, as the ability to respond to environmental signals is fundamental for the induction of flowering in optimal conditions. Arabidopsis is a facultative long day plant and as such flowers under both long and short days but sooner under long day conditions. To identify novel flowering mutants of the photoperiodic pathway a T-DNA tagged population of Arabidopsis was obtained from NASC (INRA Versailles lines) and screened for altered flowering times in long and short days. A novel flowering time mutant, termed day neutral flowering (dnf) was isolated as a result of this screen which displayed no short-day inhibition of flowering. The dnf mutant, therefore, flowers at the same time under both long and short day conditions. Complementation of the dnf mutant was carried out by re-introducing the wild-type DNF gene under the control of its endogenous promoter. Further complementation was also achieved using N-terminal TAP-tag fusion proteins. This complementation confirmed that the dnf mutation is responsible for the observed phenotype. The ability to perceive daylength in Arabidopsis is achieved by the coincidence of light with (CONSTANS) CO expression. This promotes the expression of FLOWERING LOCUS T (FT) which in turn leads to the promotion of flowering. Quantitative real-time PCR (qRT-PCR) has shown that the expression of both CO and FT is altered in the dnf mutant, which is indicative of its early flowering phenotype. Furthermore the expression of GIGANTEA (GI) was unaltered in dnf. These results position DNF upstream of CO and downstream of GI within this photoperiodic flowering pathway. Crosses between dnf and the co-2 mutant flowered at the same time as the co-2 mutant and as such confirms that DNF lies on the photoperiodic flowering pathway. The fact that dnf flowers early in SD combined with the qRT-PCR analysis of GI, CO and FT suggest that DNF plays a role in the repression of CO in such photoperiods. The DNF protein is 141 amino acids in length and contains a predicted membrane spanning domain between residues 13 and 33 and a putative RING finger domain between residue 79 and 121. This putative ring finger domain within DNF shows sequence similarity to both an E3 ligase and a PHD domain. Many E3 ubiquitin ligases have been shown to target specific proteins for degradation by the 26S proteosome whereas PHD domains have been implicated in the regulation of gene transcription through chromatin remodelling. The biochemical function of DNF is unknown, however one method to elucidate its role is to look at possible protein interactors. With this aim in mind we have decided to use an in vivo pull-down approach and have produced plants expressing a TAP tagged DNF protein for this purpose. However, due to time constraints pull down and MASS SPEC analysis was unable to be performed. The results of EGFP expression assays revealed that DNF protein is localised in the cytoplasm when driven by its native promoter and in the cytoplasm and nucleus of plant cells when over-expressed from the strong P35S promoter. The hypothesised role of DNF in the control of flowering will be discussed.

The phytopathogen Botrytis cinerea is responsible for the devastating grey mould disease that affects hundreds of economically important crop species. B. cinerea represents a necrotrophic pathogen that must kill host tissue before if can consume the nutrients, this distinguishes it from other biotrophic pathogens that exist parasitically. Importantly, B. cinerea is capable of infecting the model plant organism Arabidopsis thaliana. Together with the availability of the sequenced B. cinerea genome and the available molecular tools that now allow fungal genome manipulation makes the pathosystem ideal for studying necrotrophic pathogen life style from a systems biology perspective. This thesis focuses on the transcriptional responses of the A. thaliana host to B. cinerea infection. A high resolution transcriptome time series experiment was conducted to compare transcriptional variation between infected and mock infected A. thaliana leaves over 48 hours. This identified 9838 unique host genes differentially expressed over the course of infection. High resolution temporal expression profiles of genes were used to build transcriptional gene regulatory networks using a Variational Bayesian State Space Modeling technique. Approximately 56% of principle network components identified by this method and tested using a reverse genetics approaches showed an effect on defence against B. cinerea. This represents a significant increase in the predictive power (of gene essentiality) when using this method compared to classical forward genetics approaches and simple reverse genetic approaches following on from expression profiling studies. Attempts were made to resolve the local networks surrounding two of these previously uncharacterised principle network components involved in defence against B. cinerea using further transcriptome expression profiling and Yeast-1-Hybrid analysis. Subsequent re-modeling and experimental studies identified a number of high probability targets and several potential regulators of these principle network components. Overall the A. thaliana-B. cinerea interaction presents a experimentally tractable pathosystem for studying necrotrophic plant defence from a systems biology perspective.

The impact of climate change on plant disease is an important concern for agriculture. Tools from natural and social science are used in this interdisciplinary study in an innovative way to assess potential impacts on the UK strawberry sector. Records of agricultural statistics and disease incidence covering a 90-year period were analysed to study agricultural change and the past influence of disease on the sector. Future change in potential disease incidence was then modelled for three of the most common diseases by building probabilistic projections for 2020 to 2080 using the UKCP09 scenarios. Using these disease scenarios, data were collected from strawberry growers, through a national survey and case study work in two contrasting areas of strawberry production, covering around 40% of the UK sector. A number of major outcomes were obtained. The introduction of polytunnels was seen as the most important influence on change in the sector, tripling crop yields and reducing the climatic impact on yield variability. Disease was found to vary spatially and temporally throughout the country, emphasizing the need for development of resistant cultivars, use of pesticides and a change in cultivation methods. Changes in future disease incidence were predicted for all three pathogens, with a high degree of spatial variation. The outcome suggests that the UK Strawberry sector may be vulnerable, not only to the impacts of plant disease, which has affected both the distribution of the sector throughout the UK and profitability of some of the businesses, but especially to pressures arising from other factors such as labour and decreasing profit margins. When coupling these with important policy changes such as the change in the EU Pesticides Directive, a challenging picture emerges for the future of the sector in the UK. Lesson learned from this sector may be applicable to other sectors.

Basal Stem Rot (BSR) of oil palm (EIaeis guineensis) is caused by Ganoderma boninense, and is one of the most commercially devastating diseases in South East Asia. It can kill more than 80 percent of stands by the time they are halfway through their normal economic life. High incidence of BSR results in economic losses, due to zero yield from dead palms and significantly reduced weight and number of fruit bunches in infected but living palms. Many methods have been attempted to control BSR. These include clean clearing, fallow period before re-planting, burning of waste or dead material, windrowing, use of fungicides as soil drenching or by tree injection, surgical removal of infected material, biological control and development of varieties with enhanced resistance. To date no method gives good control of Ganoderma infection in established plantations and some have technical limitations in application. The aim of this study was to investigate a possible source of disease resistance associated with the accumulation of phenolics, after inoculation with G. boninense and elicitation with the biopolymer chitosan. The identity of isolates of Ganoderma from Langkon Oil Palm Estate in Sabah, Malaysia, which were used in this project, was confirmed using Ganoderma Selective Medium and DNA sequence analysis after PCR amplification. That latter method showed that the Sabah isolates were very similar to virulent G. boninense strains FA5017 or FA5035 from West Malaysia, with a maximum similarity of 98%. A method to inoculate oil palm seedlings with Ganoderma, based on root treatment with macerated mycelium, was developed which facilitated experiments with consistent and rapid disease development. This proved an improvement compared to the traditional technique of inoculation with colonised rubber wood blocks. Ekona and Calabar varieties of oil palm were found to be more susceptible to G. boninense infection compared to the variety AVROS, based on a higher accumulation of ergosterol (a fungal sterol) and infection scores based on the Modified Disease Severity Score. Results from both parameters showed a strong correlation between each other and provided a quick assessment on the progress of pathogen infection for use in this project. It should be noted, however, that none of the varieties tested exhibited a high level of disease resistance. Following preliminary studies to quantify the accumulation of total phenolics, using Folin-Ciocalteu reagent, High Performance Liquid Chromatography was used to monitor changes in the concentration of three specific phenolics, which were considered to be of primary importance. These were syringic acid, caffeic acid and 4-hydroxybenzoic acid (4-HBA). Their identification was confirmed using co-injection with pure standards and by LCMS/Q-tof. The variety AVROS showed a slightly higher content of all these three phenolics compared to Ekona and Calabar. Stimulation of their accumulation was promoted by the addition of chitosan to the plant-growing medium. By the end of the time course, the concentration of these phenolics decreased in the oil palm tissues inoculated with G. boninense suggesting possible metabolism by this pathogen. This loss was, however, not detected in tissues elicited with chitosan alone and was greatly reduced when G. boninense was combined with chitosan. In vitro studies on antifungal activity with phenolics incorporated either in 10% Potato Dextose Agar, Oil Palm Root Agar or Oil Palm Root Broth were done, using concentrations (50-110 pg mL 1) of phenolic acids typically detected in oil palm root tissues. Syringic acid was found to be the most antifungal of the phenolics tested. Caffeic acid had lower activity and 4-HBA was virtually inactive. Concentrations of syringic acid detected in root tissues, especially in the presence of chitosan, that might be expected to inhibit growth of G. boninense. The pathogen was shown to be able to degrade all of the phenolic acids tested. However, at the highest concentration tested, metabolism was greatly delayed, only occurring after a lag phase in the growth of pathogen. TLC bioassay showed the degradation products of all these phenolics were no longer antifungal. Thus, accumulation of phenolic acids, especially syringic acid, may provide a useful marker in future breeding of resistant varieties.

Lethal diseases (LD) of coconut caused by phytoplasmas have destroyed millions of palms globally and pose a serious threat to the coconut industry in Ghana. This study investigated the genetic basis of resistance/tolerance of coconut varieties and hybrids to LD. The study was also aimed at developing molecular markers for cultivar verification to be used for the sustainable breeding of high value varieties and hybrids. Using PCR diagnosis and monitoring of symptoms of LD over a three year period, the study determined that escapee palms of the West African Tall ecotype (WAT) found in LD-devastated fields die once infected and do not represent resistant/tolerant sub-populations. Six monthly observations of symptoms and PCR diagnosis showed that LD-infection and symptom development occurred all year round. Quantification of phytoplasma amounts using novel quantitative real-time PCR methods did not reveal a pattern in pathogen disease titres between the rainy and dry seasons or between plant parts. Due to a lack of LD-infection in the sampled SGDxVrT and MYDxVrT palms, the basis of their resistance/tolerance could not be determined. Out of 44 microsatellites markers assessed for their usefulness in differentiating between the Ghanaian breeding materials, only two of the markers, CnCirC12 and CAC65 initially appeared to be associated with alleles specific for the susceptible 'West Africa Tall' variety but screening with further samples showed these two markers also to not be specific. The study found that palms of each variety did not show consistent genotypes for variety-specific SSR markers to be identified. Diagnostic assays based on the LAMP technique (DNA amplification at a single temperature using Bst polymerase) were assessed for their potential for in-field use. The simplicity of the technique and the rapidity with which results are obtained <30 min) demonstrated that this non-PCR technique could be a future method of choice for field diagnostics in Ghana and in Africa at large. Seeds from LD-infected palms were assessed for their ability to transmit the LD phytoplasma to progeny plants. Out of 105 coconut seedlings derived from LD-infected palms, none was found to be infected or developed LD after six monthly PCR diagnosis for 18 months. This study concludes that LD is unlikely to be seed transmitted and that the DNA fragments detected in coconut embryos may not represent a viable organism