Thursday, 6 March 2014

Crystallins under the Lens

Written by Jill Faircloth

For generations, anyone who argued against evolutionary theory would point to the human eye and exclaim that nothing so perfectly adapted to its purpose could have evolved in a series of random steps. The well-rehearsed counter argument is that even a very basic recognition of light and shadow via an organic pinhole camera is useful as an aid to survival and that this could provide the first stepping stone towards the sophistication of the vertebrate eye (see references 1 and 2). The theory is supported by a succession of organisms with gradually increasing vision.

On a molecular level, the proof is harder to achieve but Christine Slingsby of Birkbeck's Department of Biological Sciences has used crystallography to do just that. In investigating the structure of the proteins of the vertebrate eye lens, Slingsby has not only greatly increased our understanding of their characteristics and mechanisms but also provided fascinating insights into their evolution.

Professor Slingsby's work is featured in several pages in the PPS course: Greek Key Motif, Beta Sandwiches, Lens Proteins and Cataract and Eukaryotic Genomes. Last year she published a paper (reference 3 here) which summarised the key conclusions of her research during the last ten years. This review is available in the Birkbeck e-library here.

Vertebrate lenses comprise layers of highly elongated fibre cells which give transparency and focus but the refractive power is given by high concentrations of transparent proteins from two superfamilies: the alpha crystallins and the beta-gamma crystallins. These proteins, which are all mainly made up of beta strands, have been co-opted from their original functions to generate a functioning lens.

You don’t need to look far to find the probable origin of α crystallins. They are small heat shock proteins (sHsps), molecular chaperones that are present in most types of cell in most organisms. They are upregulated: that is, produced in greater quantities, by cells under environmental stress as part of the protein homeostasis response.

Despite their name, βγ crystallins are unrelated to α crystallins; all crystallins interact to form a refractive index gradient which can vary as required. Apart from the vertebrate lens where they are very prevalent, and in stark contrast to α crystallins, βγ crystallins are found only in other vertebrate eye tissues (except as a component of a much larger gene/protein known as Aim1) and this makes their origin harder to identify. Beta and gamma crystallins each contain four Greek key motifs organised as two βγ-crystallin domains.

There are several requirements for an eye lens protein. It must be expressed at very high levels, unlike sHsps, so the sHsp gene promoters would have required modification. The proteins must pack tightly and uniformly enough so that there is no irregularity on the scale of the wavelength of light and they must be soluble but must not crystallise or separate into different phases. In addition, lens fibre cells have lost their organelles, which could cause light scattering, and so have no mechanisms for protein repair or disintegration. Accordingly, these proteins need to have a lifespan as long as the vertebrate using them.

One of the main reasons for crystallins having been adopted as lens proteins could be that the two α crystallins are able to dynamically form polymers with highly diverse size and shape. This ability was demonstrated as the first crystal structure of a sHsp revealed a hollow octahedral structure of 24 α crystallin monomers. The next one to be solved showed point group 32 symmetry and was constructed from six dimers arranged in two interlocking rings.

This figure shows the beta-sandwich structure of the alpha-crystallin domain of a monomer, the formation of the dimer with the B6 beta strand exchanging into the partner beta-sandwich, and the oligomer with six dimers forming interconnecting discs. The dimers link using motifs on the C terminal extension which insert into the pocket between the B4 and B8 strands, shown in dark blue, and by interaction of the N terminal helices.

Figure taken from Slingsby, C. et al. (2013. PDB 1GME

In addition to the wide range of alpha-crystallin oligomers, the numerous βγ-crystallin chains can be assembled to create a wide range of polymers which coexist in a polydisperse stable but flexible arrangement of varying density.

Beta-crystallins thus appear to function in a similar way to α-crystallins, forming a diverse range of differently sized hetero-oligomers that adjust the refractive index throughout the lens.

Gamma crystallins are different because they are monomeric and polar. They are present in differing concentrations throughout the lens and their polarity results in distinctive orientations towards other crystallins which may regulate inter-crystallin interactions. There is evidence that disruption of these dipoles results in cataracts.

By examining genomes of organisms which predate the development of the camera eye, Slingsby has shed light on the evolutionary pathway of crystallins as lens proteins. PPS students will bave read about the single-domain βγ-crystallin in the urochordate (invertebrate) sea squirt, Ciona intestinalis, that has exactly the same double Greek key structure as a vertebrate crystallin but includes a calcium binding sequence in each Greek key motif. This ancestral link was further demonstrated by the remarkable discovery that the gene promoter for Ciona-crystallin could successfully target reporter gene expression for proteins associated with vision in vertebrates.

Investigation of the genome of a cephalochordate, which is part of the lineage of both vertebrates and urochordates, revealed a less complex ancestor to βγ-crystallins. Signature sequences from the βγ-crystallins have also been found in bacterial and archaeal proteins. The implication of this is that all of the proteins of the vertebrate lens could well have evolved from proteins present in ancient species with no visual function. An interesting twist is introduced by the knowledge that the nonchordates, or animals without a spine, can use quite similar cellular lenses that involve non-crystallin proteins.

This suggests that lenses evolved independently in different animal kingdoms, relatively late on an evolutionary timescale, utilizing different proteins that were available in the respective phyla, that is proteins which already had an established purpose but which had qualities allowing them to form lenses. Since all species seem to have had access to at least a basic form of βγ-crystallin, it is an impressive demonstration that evolution can not only capitalise on the multiple possibilities presented by one family of proteins to develop a functioning visual system, but also repeat the trick from a different starting point.


  1. Dawkins, R. (1994). The eye in a twinkling. Nature 368, 690-691
  2. Nilsson, D.E., Pelger, S. (1994). A pessimistic estimate of the time required for an eye to evolve. Proc. Biol. Sci. 256(1345): 53-8.
  3. Slingsby, C., Wistow, G.J. and Clark, A.R. (2013). Evolution of crystallins for a role in the vertebrate eye lens. Protein Sci. 22(4):367-80.