The human eye lens has a remarkable, distinct physiology. It is formed of concentric layers of specialised cells, called fibre cells. The majority of these are formed in utero and lose all organelles in order to be transparent enough to correctly refract light onto the retina. This means, however, that the protein content of the lens must last the entirety of the human lifetime and be able to deal with daily exposure to UV radiation. Despite being incredibly stable, eventually lens proteins do begin to aggregate leading to cataracts, a leading cause of blindness worldwide, currently with no pharmaceutical interventions available.

Despite this UV barrage, the lens protein, gamma D crystallin (HGD), shows remarkable stability. This has been attributed to four conserved, highly stabilised tryptophan residues which absorb UV strongly. We wanted to investigate what happens when UV radiation is absorbed by HGD and how this might affect the structure of HGD. To do this we employed a serial crystallography set up. This involves taking a series of diffraction images from tens of thousands of microcrystals (~20 µM in size). The distinct advantage of this approach is it allows us to collect data at room temperature and avoid damage from radiation as each crystal is only shot with the X-ray beam once. For these experiments we used two beamlines i24 and Diamond Light Source, UK and TREXX EMBL@PetraIII, these facilities allow us to quickly and easily collect high quality serial data. It also allows us to conduct time-resolved studies and capture molecular movies in response to a trigger signal – in our case UV light. This study represents an important step towards our goal of using time-resolved crystallography to observe UV induced dynamics in proteins.

Here we took our microcrystals and determined the structure of the protein within a couple of weeks of purification (fresh) or left the crystals to age in the dark for months. Interestingly as the crystals age in the presence of dithiothreitol (DTT - a reducing agent/antioxidant) it accumulates on the surface of the protein through covalent disulfide bonds to surface cysteine residues. This is quite unexpected and has only recently been reported in a similar gamma S crystallin. We hypothesised this might be similar to how the natural antioxidant, glutathione (GSH), behaves in the eye lens and that by binding and blocking key cysteine residues, it prevents HGD aggregating through dimerisation. We also wondered how it would respond to UV light, and remarkably we found that the DTT was cleaved off after exposure. This led us to consider a model where antioxidants may be both reducing aggregation and being recycled by UV.

Initially we compared the different structures and on first glance there is little difference. Initially we expected that UV radiation would eventually damage the tryptophan residues and we hoped to observe this in the structure. However, this was not the case. Instead the DTT is absent in the fresh structure, present in the aged, and removed again by UV light. Despite this, the overall confirmation looks similar. We wondered what would be the best way to determine if the small structural changes accompanying DTT binding and removal were significant.  This is where using RoPE (1, 2) for our analysis was really insightful. We were able to exploit the inherent oversampling involved in taking thousands of serial diffraction stills and compare the conditions in torsional space and by conducting principal component analysis and singular value decomposition (PCA & SVD) we were able to see that the different conditions formed distinct clusters. By comparing each cluster in torsional and coordinate space we were able to reveal subtle global changes that occur upon ageing and are partially reversed with UV radiation – suggesting a return towards a more ‘fresh’ state.

This shows a novel cysteine modification in HGD and points towards a role for tryptophan cysteine electron transfer to contribute to the oxidation balance of the eye lens and to the stability of HGD. It also demonstrates scope for characterising and harnessing exciting photobiology in the UV region in the future using serial crystallographic methods.

1. https://rope.hginn.co.uk/

2. Ginn HM. Torsion angles to map and visualize the conformational space of a protein. Protein Science. 2023; 32(4):e4608. https://doi.org/10.1002/pro.4608