Light Protection in Plants: Characterisation of Violaxanthin de-epoxidase

Abstract: Plants and algae need light to drive the photosynthetic machinery. An excess of light will, however, result in damage to the photosynthetic machinery and to the rest of the organism. The surplus of energy, when exposed to light stress is converted into less harmful heat energy through a process called non-photochemical quenching. This quenching is partly dependent on the xanthophyll pool located inside the thylakoid membrane. During light stress the xanthophyll violaxanthin is converted to zeaxanthin by violaxanthin de-epoxidase (VDE), triggering the zeaxanthin dependent light quenching. VDE is located on the lumen side of the thylakoid and is activated by the reduction of pH caused by photosynthesis. The sequence of VDE has been divided into three domains, a conserved N-terminal domain rich in cysteines, a lipocalin-like domain expected to bind the substrates and a negatively charged C-terminal domain rich in glutamic acid. In this work we have constructed cysteine mutants that revealed the importance of 12 of the 13 cysteines of VDE to the activity. These 12 cysteines were found to form disulphides in a pattern giving two hairpin structures and also increased the thermal stability of VDE. The active site of VDE does not appear to be exclusively located in the cysteine rich N-terminal domain. The expression of the N-terminal domain without the rest of VDE did not show catalytic activity. The rest of VDE without the N-terminal domain was also not able to catalyse the reaction, but after mixing of these two constructs the activity returned. This shows that the N-terminal domain and the rest of VDE can fold independently and also indicates that the active site is localised in the interface between the N-terminal domain and the lipocalin-like domain. Crosslinking of monomeric VDE could localise the N-terminal domain near the opening of the lipocalin barrel, where violaxanthin is predicted to bind. The glutamic rich C-terminal domain could be truncated from VDE while the rest of VDE remained active. This showed that the C-terminal domain was not required for the catalytic activity of VDE. The truncation did, however, cause a great loss of activity and a shift in how the VDE activity depends on pH. The C-terminal domain could be linked with the ability of VDE to oligomerise at the pH required for activity. The pH dependent oligomerisation of VDE was lost after truncation of the C-terminal domain. A reduction in pH towards the pH required for optimal activity also causes a strong formation of α-helical structures involved in coiled coils. This formation of secondary structure was also lost after truncation of the C-terminal domain, which is predicted to contain coiled coils. A likely scenario is that the pH activation of VDE involves an oligomerisation event caused by coiled coils at the C-terminal domain. The oligomerisation of VDE could also be seen using chemical crosslinking at different pH, showing a monomeric state at neutral pH while oligomeric interactions occur at lower pH. Small angle x-ray scattering gave indications of a dimeric symmetry of the oligomeric state, while also revealing an elongated shape of monomeric VDE with the lipocalin-like domain localised in the centre. We also show that the previously proposed symmetric docking of violaxanthin into the dimeric state of the lipocalin-like domain appears less likely compared to a non-symmetric binding, based on the observation that only one side of violaxanthin is converted per substrate binding of VDE.