Matriptase processing of APLP1 ectodomain alters its homodimerization

Abstract

The amyloid beta peptide (Aβ) is derived from the amyloid precursor protein (APP) by secretase processing. APP is also cleaved by numerous other proteases, such as the type II transmembrane serine protease matriptase, with consequences on the production of Aβ. Because the APP homolog protein amyloid-like protein 1 (APLP1) shares similarities with APP, we sought to determine if matriptase also plays a role in its processing. Here, we demonstrate that matriptase directly interacts with APLP1 and that APLP1 is cleaved in cellulo by matriptase in its E1 ectodomains at arginine 124. Replacing Arg124 with Ala abolished APLP1 processing by matriptase. Using a bioluminescence resonance energy transfer (BRET) assay we found that matriptase reduces APLP1 homodimeric interactions. This study identifies matriptase as the first protease cleaving APLP1 in its dimerization domain, potentially altering the multiple functions associated with dimer formation. Alzheimer's disease (AD) is a neurodegenerative disease characterized by a progressive and accelerated loss of neurons, leading to cognitive disorders and is currently the most common dementia 1. Accumulation of extracellular amyloid beta (Aβ) whether in the form of plaques, oligomers or soluble monomers is a fundamental hallmark of AD 2,3. In the pathogenic amyloidogenic pathway, successive APP cleavages by βand γ-secretase results in the production of Aβ 4. Recent treatment strategies targeting elements of the amyloidogenic pathway have failed to slow the progression of symptoms. Therefore, a better understanding of the mechanisms involved in AD is needed and several teams have focused on the physiological role and biosynthesis/processing of APP family members. Amyloid-like protein 1 (APLP1) is part of the same family and is homologous to APP 5. According to the Human Protein Atlas 6,7 , APLP1 is enriched in the human brain while APP is ubiquitously expressed, consistent with data obtained in mice 8. APP and APLP1 are type I transmembrane proteins sharing conserved luminal E1 and E2 domains 5. The E1 domain, rich in cysteines, is comprised of two subdomains, a growth factor-like subdomain (GFLD) that binds heparin and that stimulates neurite growth, as well as a CuBD subdomain that binds Cu and Zn ions 9. The E2 domain forms an antiparallel dimer and binds heparin in its dimeric form 10. Finally, the C-terminal domains of these proteins contain a YENPTY motif that serves as an endocytosis signal 11. Although both proteins can be cleaved by secretases, the Aβ sequence is only found in APP 12,13. APP and APLP1 are both involved in neuronal differentiation, synaptogenesis, neurite growth, and synaptic plasticity 14-16. APP and APLP1 are known to form homo-and heterodimers 17 , which are in part dependent on the conserved E1 domain 18. These dimeric interactions occur at the plasma membrane on a single cell (cis interaction) but also occur between transmembrane proteins of adjacent cells (trans interaction) 19-21. APP/APLP1 interactions promote cell adhesion in a homo-and heter...

APP family member dimeric complexes are formed predominantly in synaptic compartments

“…In addition, processing by e.g. the transmembrane serine protease Matriptase, which cleaves APLP1 within the E1 domain and thereby negatively impacts APLP1 homodimerization may contribute to the complex pattern of APLP1 on BN gels [ 55 ]. We already demonstrated that all APP family members can heterodimerize with each other in all possible combinations, which has also been shown endogenously in mouse brains via co-immunoprecipitation of APP and APLP1 in synaptosomes [ 26 ] (Additional file 1 : Fig.…”

„APP family member dimeric complexes are formed predominantly in synaptic compartments“

“…In contrast, APLP1 revealed a pattern of steady complex formation between 200 and 800 kDa, involving also dimerized complexes (Figure 1, 2) . The reason for the pattern of diverse high molecular weight complexes might be based on heparin or zinc induced APLP1 dimerization (August et al, 2019; Dahms et al, 2015; Dunsing et al, 2017; Mayer et al, 2014; Xue et al, 2011), differently glycosylated complexes (Paliga et al, 1997; Schilling et al, 2017), or processing by the transmembrane serine protease Matriptase, which cleaves APLP1 within the E1 domain and thereby negatively impacts APLP1 homodimerization (Lanchec et al, 2020).…”