Structural Insights into the Interaction Between Autophagy‐regulating Protein BECN1 and GAPR‐1

Yue Li,Minfei Su,Yuting Zhao,Christopher L. Colbert,Beth Levine,Sangita C. Sinha
DOI: https://doi.org/10.1096/fasebj.31.1_supplement.913.14
2017-01-01
Abstract:Autophagy is a conserved, eukaryotic cell survival pathway, responsible for the degradation and recycling of large and stable macromolecular assemblies, such as protein aggregates, organelles, and intracellular pathogens. BECN1/Beclin 1 is an essential component of the autophagosome nucleation complex that mediates the first committed step in autopahgy. The Golgi‐Associated Plant Pathogenesis Related protein 1 (GAPR‐1) is a mammalian protein that down‐regulates autophagy by binding to BECN1 1 . BECN1 residues 267–284 are required for this interaction 1 . Further, a cell‐permeable peptide, Tat‐BECN1 (residues 267–284) inhibits the BECN1:GAPR‐1 interaction, thereby inducing autophagy. Autophagy induced by treatment with the Tat‐BECN1 (residues 267–284) peptide reduces certain virus infections and protein aggregation. Therefore, understanding the mechanism of this interaction is important for developing therapies for infectious and neurodegenerative diseases. Our goal was to better understand the mechanism of the BECN1:GAPR‐1 interaction using a combination of structural methods, structure modeling and mutagenesis combined with co‐immunoprecipiitation. GAPR‐1 crystallizes as a dimer, although size‐exclusion chromatography indicates that it is a monomer in solution. To date, we have been unsuccessful in crystallizing the complex of wild‐type (WT) GAPR‐1:Tat‐BECN1 (residues 267–284). Therefore, as an alternative strategy, we modeled the BECN1:GAPR‐1 interaction using the CABS‐dock server that searches for the binding sites by utilizing Replica Exchange Monte Carlo dynamics with 10 replicas spread on the temperature scale and final models are reconstructed from the C‐alpha trace to an all‐atom representation. According to this model, Tat‐BECN1 is bound to a groove located at the dimer interface of the WT GAPR‐1. A model of the BECN1 (248–450):GAPR‐1 dimer complex was constructed by aligning the BECN1 (248–450) structure with the GAPR‐1 dimer:Tat‐BECN1 (248–450) model using PyMol. According to these models, dimeric GAPR‐1 is not likely to bind to BECN1 or even the Tat‐BECN1 (248–450) peptide, since BECN1 binding to a GAPR‐1 dimer would result in extensive steric clashes. Several conserved residues in this groove interact with Tat‐BECN1, so we mutated several combinations of these conserved residues, and tested if mutations abrogate the interaction between GAPR‐1 and BECN1. We find that two sets of GAPR‐1 mutations E86A/G102K/H103A/G138N and H54A/E86A/G102K/H103A/G138N abrogate their interactions with BECN1. Furthermore, we determined the crystal structure of the latter mutant, and compared this structure with WT GAPR‐1. We find that the binding groove of the WT GAPR‐1 is more hydrophobic, while the groove of mutant GAPR‐1 is more positively charged, and therefore may not efficiently bind BECN1 residues 267–284, which include many hydrophobic residues. Support or Funding Information NIH grants RO3 NS090939 (S.C.S.) and U19AI109725 (S.C.S) NSF grant MCB‐1413525 (S.C.S.); and NSF and North Dakota EPSCoR grant II‐1355466 doctoral dissertation awards for M.S. (PI: S.C.S.)
What problem does this paper attempt to address?