Roselab

Small angle X-ray (SAXS) and neutron (SANS) scattering provide low resolution (> 10 Å) structural information from samples in solution. SAS data can be used to develop de novo molecular envelope models, or to combine the crystal structures of individual components of a larger complex.

We are using SAS data to model cooperative interactions between transcription factors on the insulin promoter. We have collected scattering data of the DNA binding domains of E47/NeuroD1 and Pdx1 bound to a rat insulin I promoter sequence. We have shown that we can use scattering data to distinguish between structural models of these insulin promoter complex (Figure 1). We are also developing de novo molecular envelope models of the complex for docking crystal structures of E47 and NeuroD1 (Figure 2). We are in the process of developing a more detailed model using contrast variation with neutron scattering. This technique takes advantage of the different scattering length densities of protein and DNA, and of per-deuterated and hydrogenated proteins. The combined use of SAS and crystallography promises to allow the modeling of large multi-component transcriptional complexes in the future.

This project is in collaboration with Oak Ridge National Labs, Center for Structural Molecular Biology.


	Figure 1: To demonstrate the feasiblitly of developing models of transcription factor/DNA complexes using SAS data, we collected scattering data from a) DNA alone, b) DNA bound to two transcription factors with 9 basepairs between binding sites. Tthe basic-helix-loop-helix domain of E47/NeuroD1 is in green and the homeodomain of Pdx1 is in black. c) Same as b) but with binding sites spaced 7 basepairs apart. The pair distribution function, P(r) (left) represents the distribution of distances in the scattering molecules. a) shows a single peak, similar to scattering of a simple ellipsoid shape. b) shows 2 peaks, indicating E47/NeuroD1 and pdx1 are separated in space and not interacting. c) shows a single peak, indicating that the two factors are situated so they can interact on the DNA.		Figure 2: Using scattering data to cacluate a de novo molecular envelope An averaged molecular envelope was calculated from the scattering of the E47/NeuroD1 and Pdx1 DNA binding domains bound to DNA. The crystal structures of the proteins have been built into bulges in the envelope by hand. The remainder of the envelope is occupied by DNA. The envelope was calculated using the programs Dammin and Damaver by the Biological Small Angle Scattering Group.

Figure 1: To demonstrate the feasiblitly of developing models of transcription factor/DNA complexes using SAS data, we collected scattering data from a) DNA alone, b) DNA bound to two transcription factors with 9 basepairs between binding sites. Tthe basic-helix-loop-helix domain of E47/NeuroD1 is in green and the homeodomain of Pdx1 is in black. c) Same as b) but with binding sites spaced 7 basepairs apart. The pair distribution function, P(r) (left) represents the distribution of distances in the scattering molecules. a) shows a single peak, similar to scattering of a simple ellipsoid shape. b) shows 2 peaks, indicating E47/NeuroD1 and pdx1 are separated in space and not interacting. c) shows a single peak, indicating that the two factors are situated so they can interact on the DNA.

Figure 2: Using scattering data to cacluate a de novo molecular envelope An averaged molecular envelope was calculated from the scattering of the E47/NeuroD1 and Pdx1 DNA binding domains bound to DNA. The crystal structures of the proteins have been built into bulges in the envelope by hand. The remainder of the envelope is occupied by DNA. The envelope was calculated using the programs Dammin and Damaver by the Biological Small Angle Scattering Group.