Biochemical and structural studies of NRBF2 : a critical autophagy modulator that targets the Beclin1-Vps34 complex

Abstract

Nuclear receptor binding factor 2 (NRBF2) is a critical modulator of the mammalian class III phosphatidyl-inositol-3 kinase (PI3KC3) complex I. Core members of this complex include the phosphatidyl-inositol-3 kinase Vps34, the serine/threonine kinase Vps15, the scaffolding protein Beclin1 and the Beclin1-binding autophagy enhancer Atg14L. Studies have shown that NRBF2 binds to complex I and promotes cellular autophagic response by enhancing the lipid kinase activity of Vps34. How NRBF2 specifically interacts with complex I but not the UVRAG-containing complex II to promote Vps34-mediated autophagy process is not clear. Here we have conducted biochemical and structural studies of the NRBF2 coiled-coil (CC) domain to elucidate the molecular mechanism of NRBF2-mediated autophagy modulation. We have determined the crystal structure of NRBF2 CC domain. The structure reveals two helices wrapped around each other in parallel fashion, conforming to the architecture of a canonical coiled-coil dimer. The dimer interface contains multiple leucine-zipper pairings, rendering the dimeric structure highly stable. This structure is in stark contrast to that observed in Atg38, the yeast homolog of NRBF2. The CC domain of Atg38 is an asymmetric dimer, with only one helix being straight and the other bent in the middle. The dimer interface of Atg38 CC domain also contains multiple electrostatically repulsive pairings, likely rendering this structure less stable. It has been reported that Atg38 dimer is only associated with one copy of complex I while NRBF2 can link two copies to form a large dimeric complex. This difference in stoichiometry for NRBF2- or Atg38-associated complex I may be due to the differential stability of their respective CC domain. We also conducted a series of cell-based experiments to delineate how the oligomeric state of NRBF2 as determined by its CC domain affects its function in autophagy modulation. Our competitive Co-IP experiments confirm that the CC domain of NRBF2 is not responsible for its specific association with complex I. Instead, the MIT domain of NRBF2 and the C2 domain of UVRAG bind to Vps15 in competitive manner. As a result, NRBF2 can only associate with Atg14L-containing complex I, but not UVRAG-containing complex II. Additionally, by making mutations and substitutions within the CC domain, we engineered NRBF2 constructs that would adopt monomeric, dimeric and tetrameric state respectively. Co-IP experiments confirm that monomeric NRBF2 is the least competitive against UVRAG in terms of binding to Vps15. Dimeric and tetrameric constructs are noticeably more competitive and help to promote the formation of complex I. Furthermore, in terms of rescuing autophagy activity in NRBF2 knockout cells, monomeric construct is also the least effective while dimeric and tetrameric constructs lead to full recovery and even enhancement of the autophagy activity. In summary, my thesis work has provided biochemical and structural information to help understand the functional role of NRBF2 in autophagy regulation. The CC domain of NRBF2 exerts positive influence on complex I by mediating its oligomeric state and promoting its competitiveness against UVRAG-containing complex II. Collectively, these two effects may lead to enhanced activity of complex I to up-regulate the autophagy process.