A model for the complex between the hypoxia-inducible factor-1 (HIF-1) and its consensus DNA sequence

Hypoxia-inducible factor-1 (HIF-1) is a heterodimeric transcription factor activated by hypoxia. When activated, HIF-1 mediates the differential expression of genes such as erythropoietin and Vascular Endothelial Growth Factor (VEGF) during hypoxia. It is composed of two different subunits, HIF-1 alpha and ARNT (Aryl Receptor Nuclear Translocator). These two subunits belong to the bHLH (basic Helix-Loop-Helix) PAS (Per, Ahr/ARNT, Sim) family. The bHLH domain of these factors is responsible for dimerization through the two helices and for DNA binding through their basic domain. In this work, we used various methods of molecular modeling in order to develop a 3D structure for the HIF-1 bHLH domain bound to its DNA consensus sequence. Firstly, the 3D structure of the bHLH domain of both subunits based on their amino acid sequence was defined. Secondly, we compared this model with data from known crystal structures of basic leucine zipper-DNA and bHLH-DNA complexes in order to determine a potential canvas for HIF-1. Thirdly, we performed a manual approach of the HF-1 bHLH domain onto the DNA recognition site using this canvas. Finally, the protein-DNA complex 3D structure was optimized using a Monte Carlo program called MONTY. The model predicted a pattern of interactions between amino acids and DNA bases which reflect for ARNT what is experimentally observed among different X-ray structures of other bHLH transcription factors possessing the H (His), E (Glu), R (Arg) triad, as ARNT does. On the other hand, only the Arg residue is conserved in HIF-1 alpha. We propose from this model that a serine replaces the histidine while an alanine and a lysine also make contacts with DNA. From these results, we postulate that the specificity of HIF-1 toward its DNA sequence could be driven by the HIF-1 alpha subunit. The predicted model will be verified by X-Ray currently ongoing.