Sequential NMR resonance assignment and structure determination of the Kunitz-type inhibitor domain of the Alzheimer's .beta.-amyloid precursor protein

Abstract

Certain precursor proteins (APP751 and APP770) of the amyloid beta-protein (AP) present in Alzheimer's disease contain a Kunitz-type serine protease inhibitor domain (APPI). In this study, the domain is obtained as a functional inhibitor through both recombinant (APPIr) and synthetic (APPIs) methodologies, and the solution structure of APPI is determined by 1H 2D NMR techniques. Complete sequence-specific resonance assignments (except for P13 and G37 NH) for both APPIr and APPIs are achieved using standard procedures. Ambiguities arising from degeneracies in the NMR resonances are resolved by varying sample conditions. Qualitative interpretation of short- and long-range NOEs reveals secondary structural features similar to those extensively documented by NMR for bovine pancreatic trypsin inhibitor (BPTI). A more rigorous interpretation of the NOESY spectra yields NOE-derived interresidue distance restraints which are used in conjunction with dynamic simulated annealing to generate a family of APPI structures. Within this family, the beta-sheet and helical regions are in good agreement with the crystal structure of BPTI, whereas portions of the protease-binding loops deviate from those in BPTI. These deviations are consistent with those recently described in the crystal structure of APPI (Hynes et al., 1990). Also supported in the NMR study is the hydrophobic patch in the protease-binding domain created by side chain-side chain NOE contacts between M17 and F34. In addition, the NMR spectra indicate that the rotation of the W21 ring in APPI is hindered, unlike Y21 in BPTI, showing a greater than 90% preference for one orientation in the hydrophobic groove.

The relationship between amide proton chemical shifts and secondary structure in proteins

“…The criteria used to select the proteins that were used in the database were that the protein should be almost completely assigned, should have been shown to have very similar structures in solution and in the crystal, and its crystal structure should have been determined at high resolution. The 15 proteins used which satisfy these conditions are: human lysozyme (Redfield and Dobson, 1990), hen egg white lysozyme (Redfield and Dobson, 1988), bacteriophage T4 lysozyme , bovine pancreatic ribonuclease A (Robertson et al, 1989), ribonuclease H (Yamazaki et al, 1993), human ubiquitin (DiStefano and Wand, 1987), streptococcal protein G B2-domain (Orban et al, 1992), bovine pancreatic trypsin inhibitor (BPTI) (Wagner et al, 1987), FKS06-binding protein (Xu et al, 1993), glucose permease IIA (Fairbrother et al, 1992), III G~ (Pelton et al, 1991), tendamistat (Kline and Wiithrich, 1986), Alzheimer's [3-amyloid precursor protein (Heald et al, 1991), apo-neocarzinostatin (Adjadj et al, 1990), and ColE1 top protein (Eberte et at., 1990). The Brookhaven Databank names are given in Table 1 (Bernstein et al, 1977;Abola et al, 1987).…”

Processing of the Pre‐β‐amyloid Protein by Cathepsin d is Enhanced by a Familial Alzheimer's Disease Mutation

“…Rabbit antiserum to human cathepsin D was from Dako Corp. IgG was purified from rabbit sera using Avid AL [23] and coupled to CnBr-activated Sepharose 4B [24]. The methods and instrumentation for peptide synthesis [25], amino acid analysis [26] and mass spectrocopy [26] were as cited previously. The protein content of crude extracts was determined by the Bradford assay [27].…”

“…Sequencing and immunoblots were developed with Coomassie-brilliant blue (R250) and C286.8a respectively. Segments of the sequencing blot containing Coomassie-brilliant blue or C286.8a-reactive fragments were subject to N-terminal sequencing as described [25]. Data are reported for a digest in which the amount of APP,,, sequence was reduced by 80%.…”

Analysis of the Overall Structure of the Multi-Domain Amyloid Precursor Protein (APP)

“…First insights into the domain architecture of APP could be obtained from homology considerations resulting in the initial definition of the E1- and the E2-domains within the large ectodomain [6,7,8]. Highly resolved structures of a number of individual domains of APP have been determined in the last ~20 years, such as those of its growth-factor-like domain (GFLD) [9], its copper-binding domain (CuBD) [10,11], its Kunitz-type protease inhibitor domain (KPI) (not present in the neuronal APP 695 splice form) [12,13], the central APP domain (called CAPPD or E2-domain) [14,15,16] the structure of its membrane-proximal region [17] and that of its intracellular domain AICD [18,19]. In addition, the crystal structure of the entire E1-domain of APP shows that its constituting CuBD and GFLD interact tightly with one another, forming one rigid entity [20].…”