STRUCTURAL STUDIES OF CHICKEN BRAIN ALPHA-SPECTRIN
Valerie Grum, Dongning Li, Ruby MacDonald and Alfonso Mondragón* Department of Biochemistry, Molecular Biology and Cell Biology, Northwestern University, Evanston, Il *Corresponding author: a-mondragon@nwu.edu
Proteins of the spectrin superfamily are uniquely designed for the vital task of providing cells with a deformable skeleton and a flexible matrix for organizing the distribution of regulatory proteins. The members of this ubiquitous family, alpha-spectrin, dystrophin, and a membrane-bound tyrosine phosphatase, possess an elongated, flexible domain of repeating units of 106-109 amino acids, each folded into a triple-helical bundle. Although intact spectrin had been widely studied, it is not known how the molecular structure of its repeating unit domain codes for the flexibility and deformability crucial to this family of proteins. The structure of one repeat is now known and it shows that it is indeed a three-helix bundle. Nevertheless, the relative arrangement of the repeats, the linker region between them, and the general disposition of the repeats in the molecule are still largely unknown. To address these questions, we have been studying the structure of two linked repeats from chicken brain alpha-spectrin.
We have grown crystals of four different constructs of a two-repeat fragment of chicken brain alpha-spectrin. The constructs differ by the position of the amino and carboxy termini. Two of the four crystals diffract beyond 2.0Å, one diffracts to ~3.1Å, and the last one to only 4.0Å. We have solved all four crystal structures by a variety of methods. The first structure solved was the one that diffracts to lower resolution. The two high resolution forms were solved by heavy atom methods, and the fourth form by Molecular Replacement.
Results:
We collected data on the MAR-CCD detector at DND CAT on three of these crystals forms. In one case the structure had been solved previously by Multiple Isomorphous Replacement but we only had data to medium resolution. We took advantage of the intensity of the beam at the APS to be able to collect data to 2.0Å resolution. The data are of excellent quality and we are in the process of refining the structure to the resolution limit. A picture of the current electron density map is shown below.
The other two crystal forms were not solved previously and we collected the data that allowed us to solve the structures. In the first case the crystals diffract to beyond 1.8Å even though they are fairly thin plates. We collected data on native crystals and also on a mercury derivative. We also collected a complete 4 wavelength MAD experiment on selenium-methionine protein. The structure was solved by SIRAS using the mercury data. The map is of extremely good quality and the entire molecule has now been traced and is being refined to 1.7Å. A helix in the SIRAS map is shown below.
The third structure solved was of a longer construct. The crystals diffract just beyond 3.1Å resolution and have a fairly large unit cell, with four monomers in the asymmetric unit. The crystals are small and decay rapidly in the x-ray beam so we had to collect native data from several frozen crystals. To improve the signal to noise ratio, we collected the data in 0.2° steps. It is only possible to do an experiment of this type by using a CCD detector as otherwise it would be prohibitively slow. The data is of extremely good quality and has allowed us to solve the structure by Molecular Replacement techniques. Refinement of this structure is in progress
Acknowledgments. We thank H. Feinberg, K. Perry, and X. Yang for help with data collection and D. Keane, J. Quintana and other members of DND-CAT for all their help.