|
|
(12) EUKARYOTIC PORIN 3-D PREDICTION BY THREADING ON THE CRYSTAL STRUCTURES OF BACTERIAL PORINS I. Jacoboni1, R. Casadio1,2,
R. Ajello3, R. Accardi3 and V. De Pinto3
Eukaryotic porins are a group of membrane proteins whose most known role is to form an aqueous pore channel in the mitochondrial outer membrane. As opposed to the bacterial porin (a large family of protein whose 3-D structure has been determined through X-Ray diffraction), the structure of eukaryotic porins is still a matter of debate. On the other hand, there are growing evidences which confer to this proteins a crucial role both in the connection between cytosolic and mitochondrial metabolism, and in the switch of the life-cycle to cellular apoptosis. Based on evolutionary consideration and on the common finding of the obvious presence of several amphipathic beta strands in both bacterial and eukaryotic porins, we thought it was worthwhile to update our knowledge about the 3D prediction of eukaryotic porins. We used a procedure, not yet applied to porins, referred to as 'Threading' (low homology modelling). This procedure builds a model of protein of unknown structure using an alignment between this sequence and one whose tertiary structure has been determined. Since the sequence identity between bacterial and eukaryotic proteins is very low (below 25%), the alignment between sequence and structure cannot be based on sequence similarity. Therefore secondary structure has been predicted, by the combination of two Neural Network methods, and the alignment has been performed using this prediction. With this method we produced 3D models for N. crassa, yeast and mouse isoform 1 VDAC (Voltage Dependent Anion Channel). These models are essentially made of sixteen b strands which form the walls of the pore. The N-terminal amphipathic a-helix is located on the face of the model where very short segments connect consecutive b strands. On the opposite side of the membrane the model predicts the presence of larger loops among which two very large (about 20 and 30 aa each) ones are present. The position of loops is different from the position predicted in previous models. Potential functional consequences of this model will be discussed. The model has been validated by reviewing experiments of site-specific mutagenesis and residues accessibility by antibodies and proteases. The model shows an excellent agreement with the published results. In addition new experiments of validation of this model are in progress. Acknowledgements: The authors acknowledge grants from MURST (PRIN 98), CNR (1998, 1999) and University of Catania.
For further information contact...Carmen Mannella: carmen@wadsworth.org |
||