Methylation of cytosines within CpG dinucleotides is a feature of vertebrate DNA. The precise role of DNA methylation is unknown to date, although it has been implicated in several processes relating to transcriptional regulation. One approach to study DNA methylation is the characterization of proteins that bind specifically to methylated DNA. One such family of proteins is the methyl-CpG binding domain (MBD) containing family and MBD1 is a member of this family. MBD1 is implicated in transcriptional repression and various mechanisms by which it might bring about gene silencing have been proposed. These are mainly based on studies reporting interactions between MBD1 and various proteins that regulate chromatin structure. Also MBD1 function can be modified by PIAS proteins, which stimulate its conjugation to SUMO (small ubiquitinlike modifier).The original aim of this work was to address two questions about MBD1: (1) Does MBD1 form part of a stable complex with other factors, and if so, what are the identities of the other components? Purification of MBD1 revealed the presence of no stably bound interacting proteins. However, some evidence indicates MBD1 may interact with itself and form dimers, a finding which impacts on many aspects of the function of MBD1. Also a proteomics screen for transient interaction partners identified candidate binding partners for MBD1 and the related protein MeCP2, which may throw light on the function of these proteins. (2) Are there any activities which regulate MBD1 function by the removal of SUMO from this protein? No activities capable of removing SUMO from native MBD1 were found but it was demonstrated that this modification leads to the destabilization of MBD1 in vitro. The relevance of this finding in vivo is yet to be determined.

Post-translational protein modifications (PTM) vastly increase the complexity and functional diversity of the proteome, to precisely regulate crucial cellular processes. The plant immune system is composed of complex signalling networks that are influenced by various PTMs. Activation of plant immunity is associated with a rapid burst of nitric oxide (NO), which can covalently modify cysteine thiols within target proteins by a process termed S-nitrosylation to form S-nitrosothiols (SNOs), constituting a redox-based PTM. Another key PTM involved in plant immunity is SUMOylation, an essential mechanism involving the conjugation of the small ubiquitin-like modifier (SUMO) peptide to lysine residues within target proteins. Although the targets and mechanisms of S-nitrosylation and SUMOylation are becoming evident, how these key PTMs are themselves regulated remains obscure. Work presented in this thesis reveals that during plant immune signalling, the sole Arabidopsis thaliana SUMO conjugating enzyme, SUMO CONJUGATING ENZYME 1 (SCE1), is S-nitrosylated at a highly conserved, but previously uncharacterized cysteine. S-nitrosylation of SCE1 was shown to inhibit its SUMO conjugating activity in vitro and mutational analysis revealed that the site of this modification, Cys139, is not required for enzyme activity but rather constitutes a redox-sensitive inhibitory switch. Generation and characterization of transgenic Arabidopsis plants overexpressing both wild-type and mutant forms of SCE1 revealed that Cys139 is required for efficient immunity against bacterial pathogens. Furthermore, after immune activation, S-nitrosylation of this residue inhibits global SUMOylation of proteins. These results provide evidence of a novel means of crosstalk between S-nitrosylation and SUMOylation in the context of plant immunity. The abundant cellular antioxidant, glutathione (GSH), is S-nitrosylated to form S-nitrosoglutathione (GSNO), which is thought to constitute a stable reservoir of NO bioactivity. In Arabidopsis, GSNO levels are controlled by the enzyme S-NITROSOGLUTATHIONE REDUCTASE 1 (GSNOR1), which indirectly influences the levels of protein SNOs. In this study, transgenic plants overexpressing FLAG-epitope tagged GSNOR1 were generated in various mutant backgrounds, including nitric oxide overproducer 1 (nox1), to further investigate the roles of GSNOR1 and NO in plant immunity. It was shown that ectopic GSNOR1 expression completely recovers developmental and disease susceptibility phenotypes of gsnor1, but not nox1 mutant plants, highlighting in vivo differences between accumulation of GSNO and free NO. Surprisingly, elevated NO levels in nox1 plants promote S-nitrosylation of GSNOR1, inhibiting its enzymatic activity. This suggests a previously unreported means by which NO might regulate its own bioavailability. Further work in this study revealed that recombinant GSNOR1 can be SUMOylated in vitro, which appeared to increase its enzymatic activity. Several potential SUMO modification sites were identified within GSNOR1 and mutational analysis revealed that at least one of these, Lys191, is SUMOylated. Co-immunoprecipitation experiments revealed that transgenic GSNOR1 might be SUMOylated in vivo, although the site(s) and biological function of SUMOylation were not identified. Nonetheless, these results reveal another possible layer of interplay between S-nitrosylation and SUMOylation.

Small ubiquitin-like modifier (SUMO) regulates numerous biological functions. In a previous study we found that sumoylation of HSF2 is involved in regulating HSF2 bookmarking function, but the mechanism that mediates this regulation was unknown. The results in my work support the intriguing hypothesis that polycomb protein, Mel-18, actually functions as an anti-SUMO E3 protein, interacting both with HSF2 and the SUMO E2 Ubc9, but acting to inhibit Ubc9 activity and thereby decrease sumoylation of the HSF2. This study also suggested that Mel-18 negatively regulates the sumoylation of other cellular proteins, and we extend its targets to RanGAP1 protein. The results also show that RanGAP1 sumoylation is decreased during mitosis, and that this is associated with increased interaction between RanGAP1 and Mel-18. Previous studies showed little evidence of anti-SUMO E3 proteins, however, my study, taken together, found Mel-18 actually functions as a novel anti-SUMO E3 protein, interacting both with substrates and the SUMO E2 Ubc9 but acting to inhibit Ubc9 activity to decrease sumoylation of target proteins and also provide an explanation for how mitotic HSF2/RanGAP1 sumoylation is regulated. This finding also gives a clue for a future study direction in Mel-18 as a tumor suppressor: the anti-SUMO E3 function. Additionally, we identify a single-nucleotide polymorphism in another human PcG protein, Bmi-1, that changes a cysteine residue within its RING domain, cysteine 18, to a tyrosine. This C18Y polymorphism is associated with a significant decrease in levels of the Bmi-1 protein. Furthermore, the C18Y Bmi-1 protein exhibits a very high level of ubiquitination compared to wild-type Bmi-1, suggesting that that the low levels of this form of Bmi-1 are due to its destruction by the ubiquitin-proteasome system. Consistent with this hypothesis, treatment of cells with the proteasome inhibitor MG-132 results in a significant increase in levels of C18Y Bmi-1. This is the first example of a polymorphism in human Bmi- 1 that reduces levels of this important protein.

- / Astrocyte reactivity is implicated in the neuronal loss underlying Alzheimer's disease. Curcumin has been shown to reduce astrocyte reactivity, though the exact pathways underlying these effects are incompletely understood. Here we investigated the role of the small ubiquitin-like modifier (SUMO) conjugation in mediating this effect of curcumin. In beta-amyloid (Aβ)-treated astrocytes, morphological changes and increased glial fibrillary acidic protein (GFAP) confirmed reactivity, which was accompanied by c-jun N-terminal kinase activation. Moreover, the levels of SUMO-1 conjugated proteins, as well as the conjugating enzyme, Ubc9, were decreased, with concomitant treatment with curcumin preventing these effects. Increasing SUMOylation in astrocytes, by over-expression of constitutively active SUMO-1, but not its inactive mutant, abrogated Aβ-induced increase in GFAP, suggesting astrocytes require SUMO-1 conjugation to remain non-reactive.