HIF-3alpha: structural studies, identification and characterization of a high-affinity ligand of hypoxia inducible factor (HIF) and structural studies of prolil-hydroxilases (PHDs) from non usual organisms

Author(s)

Angela Maria Fala

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Biologia
• Defense date: 2017-07-20

Examining board members

Andre Luís Berteli Ambrosio; Daniela Barretto Barbosa Trivella; Daniel Maragno Trindade; Adriana Franco Paes Leme; Cristiane Rodrigues Guzzo Carvalho

• Advisor: Andre Luís Berteli Ambrosio

Abstract

HIFs (Hypoxia-inducible-factors) are heterodimers formed by subunits alpha and beta that regulate the transcription of more than 100 genes, of which are involved in the development and progression of cancer. Three isoforms for the alpha-subunit are known, and under normoxia conditions, they are regulated by PHDs (Prolyl hydroxylases) which lead to the hydroxylation of two residues of proline and subsequent assignment for degradation via the proteosome. PHDs are dioxygenases and the master regulator for HIFs. There are known four isoforms and all of them can hydroxylate the alpha subunits of HIF under normoxia. The complex of HIF-alpha:HIF-beta can regulate the transcription of more than 100 genes, many of them are involved in the development and progression of cancer, such as the regulation of genes involved in angiogenesis, erythropoiesis, cell proliferation and apoptosis. The subunit -3alpha is the last addition to the HIF-alpha class subunits (HIF-3alpha) and a distinctive one in a number of ways. Firstly, the gene encoding for HIF-3alpha (HIF3A) is itself under regulation of HIF-1alpha and exclusively subject to alternative splicing, with six isoforms confirmed by protein levels so far. Besides the complete absence of C-terminal (CTAD), HIF-3alpha is considered the short isoform in comparison with HIF-1alpha and -2alpha. Secondly, the only common domain between all -3alpha isoforms is the C-terminal Per-ARNT-Sim sensor domain (PASb), which is responsible for heterodimerization with HIF-beta. The aims of this project were to study the PASb domain of human HIF-3alpha and the prolyl hydroxylases of unusual organisms. We also intend to produce and perform structural assays with the PASb-3alpha4 isoform of HIF-3alpha, but after several attempts at expression and purification, the protein continued in the insoluble form, which made it impossible to continue the studies. Here we report the crystal structure of the PASb domain of HIF-3alpha at 1.15 Å maximum resolution and we identify an exclusive bulky hydrophobic cavity with 510Å3, bound to the unsaturated 18-carbon-long fatty acids as an unexpected and novel high-affinity intracellular cofactors of HIF-3alpha. The lipids were extracted from the samples and two phospholipids were identified by mass spectrometry as a Lysophosphoethanolamine 1-(11Z-octadecenoul)-2-hydroxil-sn-glycero-3-phosphoethanolamine (18:1-LPE) and Lysophosphoglycerol 1-(11Z-octadecenoyl)-2-hydroxil-sn-glycerol-3-phospho-(1'-sn-glycerol) (18:1 LPG). To explore the hydrophobic cavity and identify other lipids that could be interacting with the protein, we performed delipidation assays and titration of free fatty acids with different sizes and insaturation that could be present in the human organism. The results showed a higher affinity with oleic and linoleic fatty acids (18:1 and 18:2 carbons) with kd of 39±1nM and 0,8±0,2 nM respectively. Overall, this is the first evidence that an eukaryotic PAS domain may serve as a sensor for the lipid levels of the cell, possibly dictating HIF-3alpha-dependent gene expression in response to biosynthetic and bioenergetic outcomes, such as lipid accumulation or depletion, upon exposure to hypoxia. We also planned to study the three prolyl hydroxylases (PHD1, 2 and 3) as a collaboration with Dr. Christopher Schofield and their group (Chemistry Department-Oxford University). The genes of organisms that contained the predicted sequences for these enzymes, such as walrus, platypus and python. Were selected eight organisms, from which we obtained soluble proteins for six of them. Two isoforms for PHD3 (platypus and python genes) were shown to be more stable and purified with high purity. As result of the enzymatic activity of PHD3 for python and platypus organisms, we observed that the enzymes are active and they preferentially hydroxylate the proline contained in the CODD domain of HIF. Upon crystallization assay of PHD3 from python, we obtained small malformed crystals and after some refinement cycles we did not obtain single crystals that could be taken to diffraction. For the remaining PHDs that were obtained in a soluble form, several optimizations on the purification protocols had to be made to obtain more stable and promising samples, so that crystallization and enzymatic activity assays can be performed (AU)

• FAPESP's process: 14/04927-4 - Identification of a novel cellular cofactor for the hypoxia-inducible transcription (HIF) and structural studies of the catalytic domains of isoforms 1 and 3 of the human prolyl-hydroxylases (PHD1 and PHD3) in complex with inhibitors
• Grantee: Angela Maria Fala
• Support Opportunities: Scholarships in Brazil - Doctorate