Vikas Bhandawat

Delta – Notch pathway is an important signal transduction pathway in Drosophila development (1). The two main players in this pathway are Delta and Notch proteins.

Both Delta and Notch are large cell surface proteins (100 kD and 300 kD) respectively.

The pathway is activated by the interaction of Delta with Notch. Two of the 36 EGF repeats on Notch are implicated in this interaction. The Delta interacts through a DSL (Delta –Serrate-Lag) domain found on all Delta like protein. Due to the large size of these proteins structure determination by NMR or X-ray crystallography is difficult. The purpose of this modeling project is to get some structural insight into the mode of interaction of Delta with Notch. A model for Drosophila Notch repeats 11 and 12 (DNEGF 11, DNEGF 12) and Delta DSL domain was made.

The project has been carried out using InsightII on the SGI system. The methods used were sequence alignment, homology modeling, energy minimization, electrostatic computation. The sequence alignment was done using the BLOSSUM algorithm.

Homology modeling was done using the Homology module in the InsightII. The homologous structures used were coagulation factor Xa (PDB code is 1FAX) for DNEGF11 and coagulation factor X (PDB code1CCF) for DNEGF 12. Energy minimization was performed using cff91 forcefield. The models were evaluated using standard evaluation methods. Electrostatic computation was performed using Delphi package inside InsightII

Model for Notch 11 and 12. A homology model for Notch repeat 11 and 12 were constructed using the coordinates of EGF repeats whose structure has already been solved. The structure was fine tuned by minimizing using the cff91 forcefield and conjugate gradients algorithm to a maximum derivative value of energy as 0.1. The positions of the side chains were fixed so as to minimize bumps. The models for the two domains are shown in Fig 1a. and 1b. respectively. The Ramachandran Plots are shown for the two domains. (Fig. 2a and 2b)

DNEGF11 and DNEGF12 Interface. It has been reported that calcium binding EGF repeats alternate with non calcium binding repeats (2). The interface between the two repeats take different forms depending upon the number of residues in the interface (2). When there is only one residue separating the two EGF repeats it takes a more elongated form. For proteins like Notch where two residues separate the EGF repeats the structure is more compact with the two EGF repeats forming an angle of 110^o (3). The structure of clotting factor IX that takes such a structure was used to model the interface (Fig 3). It was found that the consensus Ca²⁺ binding residues were positioned appropriately to bind the calcium.(Fig.4)

Modeling of the DSL domain. A sequence alignment of DSL domains of different proteins that interact with Notch shows very little sequence homology except for the positions of the cystienes. In selecting the protein to model the DSL domain, the conservation of the position of cystienes was given the highest priority. The model was fine tuned as described for DNEGF11. The model is presented in fig 6. The DSL domain has very little tendency to collapse into a particular secondary structure.

Hydrophobicity A surface for both DSL and DNEGF11 and 12 was made and colored according to hydrophobicity ( Fig 7a, 7b). A large hydrophobic patch would indicate towards interaction through hydrophobic patches. The hydrophobic surface failed to reveal any such patches indicating some other forms of interaction.

Electrostatic surfaces. Since hydrophobicity failed to provide any clue to the mode of interaction between the two proteins a electrostatic surface of the domains were constructed. The DSL surface showed a positively charged surface and a negatively charged surface and it is possible that interaction takes place through charge interaction.(Fig 8a, 8b).

Conclusions

Models for Delta DSL domain and Notch EGF repeats 11 and 12 was made. Their surface properties indicate that the interaction may be through the charges on the surface. Interaction mediated by Ca²⁺ is also possible.

1. Muskavitch, M., Dev. Biol., 166, 415-430 (1994).
2. Downing, A. K., et. al.,Cell, 85, 597-605 (1996).
3. Rao, Z., et. al, Cell, 82, 131-141(1995).