Comparative studies for amyloid beta degradation: “Neprilysin vs insulysin”, “monomeric vs aggregate”, and “whole Aβ₄₀ vs its peptide fragments”

Abstract

Amyloid beta (Aβ) proteins are produced from amyloid precursor protein cleaved by β- and γ-secretases, and are the main components of senile plaques pathologically found in Alzheimer's disease (AD) patient brains. Therefore, the relationship between AD and Aβs has been well studied for both therapeutic and diagnostic purposes. Several enzymes have been reported to degrade Aβs in vivo, with neprilysin (NEP) and insulysin (insulin-degrading enzyme, IDE) being the most prominent. In this article, we describe the mass spectrometric characterization of peptide fragments generated using NEP and IDE, and clarify the differences in digestion specificities between these two enzymes for non-aggregated Aβ₄₀, aggregated Aβ₄₀, and Aβ₄₀ peptide fragments, including Aβ₁₆. Our results allowed identification of all the peptide fragments from non-aggregated Aβ₄₀: NEP, 23 peptide fragments consisting of 2–11 amino-acid residues, 17 cleavage sites; IDE, 23 peptide fragments consisting of 6–33 amino-acid residues, 15 cleavage sites. Also, we confirmed that IDE can digest only whole Aβ₄₀, whereas NEP can digest both Aβ₄₀ and partial structures such as Aβ₁₆ and peptide fragments generated by the digestion of Aβ₄₀ by IDE. Furthermore, we confirmed that IDE and NEP are unable to digest aggregated Aβ₄₀.

Graphical abstract

Highlights

• •Two Aβ degrading enzymes (neprilysin, NEP and insulysin, IDE) were examined.
• •NEP digested Aβ₄₀ into 23 peptides consisting of 2–11 amino-acid residues.
• •IDE digested Aβ₄₀ into 23 peptides consisting of 6–33 amino-acid residues.
• •Neither IDE nor NEP can digest aggregated Aβ₄₀.
• •NEP can digest Aβ fragment peptides including Aβ₁₆, although IDE cannot.

Abbreviations

Aβ

amyloid beta

AD

Alzheimer's disease

APP

amyloid precursor protein

CSF

cerebrospinal fluid

EIC

extracted ion chromatogram

ESI

electrospray ionization

FA

formic acid

HFIP

1,1,1,3,3,3-hexafluoro-2-propanol

IDE

insulysin, insulin-degrading enzyme

LC

liquid chromatography

MeCN

acetonitrile

MS

mass spectrometry

MS/MS

tandem mass spectrometry

MSⁿ

multiple-stage mass spectrometry

NEP

neprilysin

ODS

octadecylsilyl

PBS

phosphate buffered saline;

PGC

porous graphitic carbon

r.t.

room temperature

THF

tetrahydrofuran

ThT

thioflavin T

TICC

total ion current chromatogram

t_R

retention time.

1. Introduction

Alzheimer's disease (AD) is a progressive neurodegenerative disease that causes memory and cognitive impairment and accounts for about 50–56% of dementia cases worldwide [1,2]. The pathological features of AD include senile plaques and neurofibrillary tangles consisting of aggregated amyloid β (Aβ) proteins [1,2] and hyperphosphorylated tau proteins [3], respectively. Of these, the deposition of Aβ is recognized as the central event in the etiology of AD because it is the primary event in the disease process [4].

Aβs are produced by cleavage of an integral membrane amyloid precursor protein (APP) at the N-terminus by β-secretase, followed by cleavage at the C-terminus by γ-secretase [5]. Variations in the position cleaved by γ-secretase generates several homologs with different lengths. Aβ₄₀ and Aβ₄₂ are the major homologs generated and are 40 and 42 amino-acid residues long, respectively. Aβ monomers can easily aggregate under physiological conditions to form fibrils via soluble oligomers [6]. Neurotoxicity of the soluble oligomers [7] and synaptic plasticity/memory impairment caused by the dimer [8] have also been reported.

Almost twenty enzymes are known to contribute to Aβ degradation, including neprilysin (NEP), insulysin (insulin-degrading enzyme, IDE), angiotensin-converting enzyme, and cathepsin B. The therapeutic utility of these enzymes in AD has been studied [[9], [10], [11], [12]], including extensive studies of NEP and IDE in AD patients [13,14]. Furthermore, fluorescence assays for quantifying the enzyme activities of NEP and IDE have also been reported [15].

NEP is a zinc-dependent metalloprotease present in, for example, the brain, heart, peripheral vasculature, adrenal gland and lungs. NEP degrades not only Aβ but also a variety of bioactive peptides such as angiotensins and enkephalins [16]. In control mice, infusion of an NEP inhibitor (thiorphan) has been reported to induce Aβ₄₀ and Aβ₄₂ deposition and fibrillization [17], and in APP transgenic mice, the overexpression of NEP has been reported to reduce senile plaques [18].

IDE is also a zinc-dependent metalloprotease that is present in the liver, kidney, and brain [19] and degrades not only insulin and glucagon, but also Aβs to regulate Aβs in vivo [20]. In IDE-deficient mice, the degradation of Aβ was reduced by more than 50% concomitantly with the increase in the amount of Aβ in the brain [20], suggesting that IDE may be a link between AD and type 2 diabetes mellitus (which is a risk factor for AD) [21]. Therefore, Aβ-degrading enzymes such as NEP and IDE are recognized as key enzymes for AD therapy and as links between AD and its risk factors.

Generally, there are two major diagnostic strategies for AD: PET imaging using Pittsburgh Compound B (2-(4'-[¹¹C]methylaminophenyl)-6-hydroxybenzothiazole) for Aβ plaques and neurofibrillary tangles in brain [22], and the quantification of Aβ₄₀, Aβ₄₂, and tau in cerebrospinal fluid (CSF) using an immunosorbent assay or liquid chromatography (LC)-mass spectrometry (MS) [23]. However, there are few reported attempts to analyze specific Aβ peptide fragments generated by degrading enzymes, such as the N-terminal [24] and C-terminal [25] Aβ peptide fragments in CSF. Therefore, the potential use of specific Aβ peptide fragments generated by Aβ-degrading enzymes as clearance and diagnostic markers requires clarification of their detailed specificities, including details of the cleavage sites, the effect of combining several degrading enzymes, and the activities of the various components involved in the aggregation process (monomer, oligomer, and fibril).

In this article, we describe comparative studies on Aβ degradation conducted using two degrading enzymes (NEP vs. IDE), various sizes of Aβ (whole Aβ₄₀ vs. its peptide fragments), and two aggregation states of Aβ (non-aggregated Aβ₄₀ vs. aggregated Aβ₄₀).

2. Materials and methods

2.1. Reagents and materials

Specific reagents and materials were purchased as follows: Aβ₄₀ (human, DAEFRHDSGY EVHHQKLVFF AEDVGSNKGA IIGLMVGGVV) and Aβ₁₆ (human, DAEFRHDSGY EVHHQK) (Anaspec, Inc., Fremont, CA, USA); NEP (human recombinant, solution in Tris, NaCl and ZnCl₂) (R&D Systems, Inc., Minneapolis, MN, USA); IDE (human recombinant, solution in Tris and NaCl) (Bon Opus Biosciences, LLC., Millburn, NJ, USA); sequencing grade modified trypsin (Promega Co., Madison, WI, USA); acetonitrile (MeCN), formic acid (FA), and tetrahydrofuran (THF) (Nacalai Tesque, Inc., Kyoto, Japan); thioflavin T (2-[4-(dimethylamino)phenyl]-3,6-dimethyl-1,3-benzothiazol-3-ium chloride, ThT) (Sigma-Aldrich, Inc., St. Louis, MO, USA); 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan); OptiPlate-384, White Opaque 384-well MicroPlate (384 well plate) (PerkinElmer, Waltham, MA, USA); AMPLIseal (plate seal) (Greiner Bio-One, Baden-Württemberg, Germany); and Protein LoBind® tubes, 0.5 mL (Eppendorf, Hamburg, Germany). All other general chemicals, vials, and gases were of the highest grade available and were obtained from local providers.

2.2. Reagent solutions

All stock solutions were stored in Protein LoBind® tubes to minimize the adsorption of Aβs.

Preparation and storage of Aβ₄₀ (0.25 mM) solution: Lyophilized Aβ₄₀ in its original vial was dissolved and monomerized in a corresponding amount of HFIP [26] to give a 1 mg/mL solution. This solution was sonicated for 8 min, incubated for 1 h, re-sonicated for 8 min at room temperature (r.t.), and evaporated to dryness under a N₂ stream. The residue was dissolved in a corresponding amount of 0.1% (v/v) aq. NH₄OH/MeCN (4:1, v/v) [27] to give a 0.25 mM solution. After 12 min sonication, the solution was divided into aliquots (30 μL) and stored at −80 °C. Prior to use, the solutions were thawed, mixed by pipetting repeatedly (10 μL, × 10), then an aliquot was transferred into a tube (0.5 mL) and evaporated under a N₂ stream. The residue was re-dissolved in H₂O/MeCN (4:1, v/v) by sonicating for 3 min to give a 0.25 mM solution.

Preparation and storage of Aβ₁₆ (0.55 mM) solution: Lyophilized Aβ₁₆ in its original vial was dissolved in a corresponding amount of H₂O/MeCN (4:1, v/v) and sonicated for 12 min to give a 0.55 mM solution. The solution was divided into aliquots (30 μL) and stored at −80 °C. Prior to use, the solution was thawed and used as-is.

Preparation and storage of ThT (0.125 mM) solution: ThT was dissolved in a corresponding amount of H₂O to give a 1.25 mM solution. The solution was divided into aliquots (50 μL) and stored at −30 °C. Prior to use, the solution was thawed and diluted with PBS to give a 0.125 mM solution.

Preparation and storage of NEP solution: A commercially provided solution of NEP (0.44 μg/μL, 22.7 μL) was divided into aliquots (2.27 μL) and stored as stock solutions at −30 °C. Prior to use, the solution was thawed and diluted with 25 mM Tris-HCl buffer (containing 200 mM NaCl and 5 μM ZnCl₂) to give a 0.1 μg/μL solution. This was further diluted with 10 mM PBS to give a 14 ng/μL solution.

Preparation and storage of IDE solution: A commercially provided solution of IDE (0.6 μg/μL, 16.7 μL) was diluted with 20 mM Tris-HCl buffer (containing 150 mM NaCl) to give a 0.1 μg/μL solution. The solution was divided into aliquots (10 μL) and stored as stock solutions at −80 °C. Prior to use, the solution was thawed and diluted with 10 mM PBS to give a 14 ng/μL solution.

Preparation and storage of trypsin solution: Lyophilized trypsin (20 μg) was dissolved in the supplied trypsin resuspension buffer (50 mM aq. acetic acid) to give a 0.1 μg/μL solution. The solution was divided into aliquots (10 μL) and stored as stock solutions at −80 °C. Prior to use, the solution was thawed and used as-is.

Control solution to monitor Aβ₄₀ aggregation: A mixture of Aβ₄₀ (0.25 mM, 10 μL), PBS (200 μL) and ThT (0.125 mM, 40 μL) was vortex-mixed for 30 s. The solution was mixed by repeated pipetting (75 μL × 5), transferred to 3 wells (75 μL each) of a 384 well plate, and covered with a plate seal for use in ThT assays [28]. The blank sample was prepared with H₂O/MeCN (4:1, v/v, 10 μL) instead of Aβ₄₀ (0.25 mM, 10 μL) and transferred to 3 wells (75 μL each).

2.3. Proteolysis of Aβ₄₀ or Aβ₁₆

All the reactions were performed in Protein LoBind® tubes except for the ThT assay. Digestion studies were performed using a ratio of Aβ (substrate):enzyme = 40:1 (w/w).

Digestion of non-aggregated Aβ₄₀ by NEP or IDE: A mixture of Aβ₄₀ (0.25 mM, 15 μL), PBS (105 μL), and NEP or IDE (14 ng/μL 30 μL) was incubated at 37 °C for 1, 3, and 7 days.

Digestion of non-aggregated Aβ₄₀ by both NEP and IDE: A mixture of Aβ₄₀ (0.25 mM, 15 μL), PBS (105 μL), NEP (14 ng/μL, 15 μL) and IDE (14 ng/μL, 15 μL) was incubated at 37 °C for 3 and 7 days.

Digestion of non-aggregated Aβ₄₀ by IDE followed by NEP: A mixture of Aβ₄₀ (0.25 mM, 30 μL), PBS (210 μL) and IDE (14 ng/μL, 60 μL) was incubated at 37 °C for 7 days. An aliquot of the solution (100 μL) and NEP (14 ng/μL, 20 μL) was mixed and incubated at 37 °C for 7 days. As a control, an aliquot of the solution (100 μL) and PBS (20 μL) was incubated at 37 °C for 3 days.

Digestion of Aβ₁₆ by NEP or IDE: A mixture of Aβ₁₆ (0.55 mM, 10 μL), PBS (70 μL), and NEP or IDE (14 ng/μL, 20 μL) was incubated at 37 °C for 3 days.

Digestion of aggregated Aβ₄₀ by IDE, NEP, or trypsin: A mixture of Aβ₄₀ (0.25 mM, 20 μL) and PBS (480 μL) was vortex-mixed for 30 s. The solution was mixed by repeated pipetting (75 μL × 5), transferred to wells (75 μL each) of a 384 well plate, and covered with a plate seal. After incubation at 37 °C for 24 h in a Gemini XPS microplate spectrofluorometer (Molecular Devices, LLC, San Jose, CA, USA), IDE, NEP, or trypsin (0.1 μg/μL, 0.81 μL) was added and the mixture was incubated at 37 °C for 3 and 7 days. Aggregation was confirmed by the ThT assay using the above control solution to monitor Aβ₄₀ aggregation.

2.4. Conditions for the ThT assay

A Gemini XPS microplate spectrofluorometer was used in kinetic mode with the following parameters: temperature, 37 °C; read mode, top read; wavelength, ex. 456, em. 489 nm; sensitivity, 12 readings PMT sensitivity Medium; run time, 24 h; interval, 15 min; automix, before first read, 15 s; between reads, 300 s; autocalibrate, on; assay plate type, 384 well standard clrbtm; column priority; carriage speed, normal; and auto read, off. SoftMax Pro Software (version 5.4.1, Molecular Devices) was used for data analysis. The data were organized using Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA, USA) to calculate the average, relative error. The fluorescence intensities were used after subtracting the corresponding controls.

2.5. LC conditions

An Agilent 1100 HPLC system (Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with a 1100 G1312A binary pump, 1100 G1379A degasser, 1100 G1367A autosampler, 1100 G1316A column heater, and a 1100 G1315B photodiode array was used for LC systems 1–4 with the following common chromatographic conditions: Mobile phases (A) 0.1% (v/v) FA in H₂O, (B) 0.1% (v/v) FA in MeCN; flow rate 0.2 mL/min; and column temperature, 40 °C.

LC system 1 (for hydrophobic peptide fragments from non-aggregated Aβ₄₀): A Cosmosil® 5C₁₈-AR-II (octadecylsilyl, ODS) column (150 × 2.0 mm i.d., 5 μm, 120 Å; Nacalai Tesque, Inc.) was used with the following linear gradient: 0 min, 0% B; 140 min, 42% B; 141 min, 65% B; 155 min, 65% B; 156 min, 20% B; 165 min, 45% B; 166 min, 20% B; 175 min, 45% B; 176 min, 20% B; 185 min, 45% B; 186 min, 20% B; 195 min, 45% B; 196 min, 20% B; 205 min, 45% B; 206 min, 20% B; 215 min, 45% B; 216 min, 0% B; and 240 min, 0% B. An aliquot of the solution (20 μL) was injected into the system. The eluate obtained between 3–140 min was introduced into the MS system.

LC system 2 (for polar peptide fragments from non-aggregated Aβ₄₀ and Aβ₁₆): A Hypercarb™ porous graphitic carbon (PGC) column [29] (100 × 2.1 mm i.d., 5 μm; Thermo Fischer Scientific Inc., Waltham, MA, USA) was used with the following linear gradient: 0 min, 0% B; 100 min, 50% B; 101 min, 65% B; 115 min, 65% B; 116 min, 0% B; and 140 min, 0% B. An aliquot of the solution (20 μL) was injected into the system. The eluate obtained between 3–100 min was introduced into the MS system.

LC system 3 (for hydrophobic peptide fragments from Aβ₁₆): A Cosmosil® 5C₁₈-AR-II column (150 × 2.0 mm i.d., 5 μm, 120 Å; Nacalai Tesque, Inc.) was used with the following linear gradient: 0 min, 0% B; 140 min, 42% B; 141 min, 65% B; 155 min, 65% B; 156 min, 0% B; and 180 min, 0% B. An aliquot of the solution (15 μL) was injected into the system. The eluate obtained between 3–140 min was introduced into the MS system.

LC system 4 (for hydrophobic peptide fragments from aggregated Aβ₄₀): A Cosmosil® 5C₁₈-AR-II column (150 × 2.0 mm i.d., 5 μm, 120 Å; Nacalai Tesque, Inc.) was used with the following linear gradient: 0 min, 0% B; 140 min, 42% B; 141 min, 65% B; 155 min, 65% B; 156 min, 20% B; 165 min, 45% B; 166 min, 20% B; 175 min, 45% B; 176 min, 20% B; 185 min, 45% B; 186 min, 20% B; 195 min, 45% B; 196 min, 20% B; 205 min, 45% B; 206 min, 20% B; 215 min, 45% B; 216 min, 0% B; and 240 min, 0% B. THF (50 μL) was injected tenth at 4-min intervals from 160 min to prevent carryover. An aliquot of the solution (30 μL) was injected into the system. The eluate obtained between 3–140 min was introduced into the MS system.

2.6. MS conditions

An LCQ-DECA ion-trap mass spectrometer (Thermo Fischer Scientific Inc.) equipped with an electrospray ionization (ESI) source was used in positive ion mode with the following parameters: spray voltage, 4.5 kV; capillary temperature, 300 °C; sheath gas flow rate, 85.0 arb; and auxiliary gas flow rate, 15.0 arb. The parameters for data-dependent MS/MS were as follows: Full scan range, m/z 100–2000 (for LC systems 1, 3, and 4) or 100–1000 (for LC system 2); precursor, top 3 ions; default charge state, 2; default isolation width, 2; normalized collision energy (CE), 45%; activation Q, 0.25; activation time, 30 ms; minimum MS signal required, 100,000; and minimum MSⁿ signal required, 5000. Xcalibur™ (version 2.0 SR2) was used for the data analyses.

2.7. Criteria for identification of peptide fragments

Proteome Discoverer (version 1. 3) (Thermo Fischer Scientific Inc.) was used to identify peptide fragments using the following parameters: Minimum precursor mass, 100 Da; maximum precursor mass, 5000 Da; enzyme, no-enzyme (unspecific); precursor mass tolerance, 2 Da; fragment mass tolerance, 0.8 Da; dynamic modification, oxidation (methionine); target false discovery rate (strict), 0.01; and target false discovery rate (relaxed), 0.05. All the identified peptide fragments were confirmed by checking the MS/MS spectra.

3. Results and discussion

3.1. General experimental setting

Aβs are challenging proteins to handle and analyze because they tend to adsorb and aggregate, and their concentration often changes because they adsorb on tubes and proteins [30]. Furthermore, significant carryover from previous injections occurs during HPLC analyses [31]. Therefore, Aβ₄₀ was initially monomerized using HFIP [26] and stored as a 0.1% (v/v) aq. NH₄OH/MeCN (4:1, v/v) solution [27]. The ODS column was washed using a quick zigzag gradient alternating between 20% B and 45% B before column equilibration (LC systems 1 and 4). Multiple injections of THF were made for LC system 4 (for aggregated Aβ₄₀). Short peptides are generally too polar (hydrophilic) to be retained on versatile ODS columns. Therefore, we concomitantly used a PGC column, which is good for hydrophilic peptides [29], to cover all the peptide fragments.

3.2. Non-aggregated Aβ₄₀ incubated with NEP

Aβ₄₀ was digested into 23 peptide fragments (2–11 amino-acid residues, 17 cleavage sites) (Fig. 1) without residual intact Aβ₄₀ after 3 days. Interestingly, the intensity patterns of several peptide peaks changed during prolonged incubation up to 7 days (Fig. 1): the peak corresponding to F²⁰AEDVGSNKGA³⁰ (m/z 1094.43, retention time (t_R) = 28.30 min on ODS) decreased as the intensities of the peaks corresponding to F²⁰AEDVGSNK²⁸ (m/z 966.33, t_R = 25.74 min on ODS) and F²⁰AEDVG²⁵ (m/z 637.12, t_R = 33.65 min on ODS) increased, suggesting that long peptide fragments can be further digested by NEP. Each amino acid sequence and its t_R identified by LC-MS with LC systems 1 and 2 is summarized in Table 1.

Fig. 1.

TICCs of non-aggregated Aβ₄₀ incubated with NEP for 1 day (top), 3 days (middle), and 7 days (bottom). (A) LC system 1 (ODS for hydrophobic peptides), (B) LC system 2 (PGC for polar peptides). Identified peptides are shown in Table 1, together with the charge state, precursor MS (m/z), and t_R (min). Arrows in the chromatogram (ODS/1 day) indicate that the amount of the peptide increased (↑) or decreased (↓) during prolonged incubation.

Table 1.

Peptide fragments found from non-aggregated Aβ₄₀ incubated with NEP.

3.3. Non-aggregated Aβ₄₀ incubated with IDE

Aβ₄₀ was digested into 23 peptide fragments. The intensity of each peak gradually increased in a time-dependent manner on the decrease in residual Aβ₄₀. However, some Aβ₄₀ remained even after 7 days, suggesting that IDE is less active than NEP in degrading Aβ₄₀ in our experimental setting (Fig. 2). The peptide fragments tended to be longer (5–33 amino-acid residues, 15 cleavage sites), and most of the peptide fragments were longer than 12 amino-acid residues, except for G²⁹AIIG³³, M³⁵VGGVV⁴⁰, and L³⁴MVGGVV⁴⁰. In contrast, the longest peptide fragment obtained using NEP consisted of 11 amino-acid residues. Each amino acid sequence and its t_R identified by LC-MS with LC systems 1 and 2 is summarized in Table 2.

Fig. 2.

TICCs of non-aggregated Aβ₄₀ incubated with IDE for 1 day (top), 3 days (middle), and 7 days (bottom). (A) LC system 1 (ODS for hydrophobic peptides), (B) LC system 2 (PGC for polar peptides). Identified peptides are shown in Table 2, together with the charge state, precursor MS (m/z), and t_R (min).

Table 2.

Peptide fragments found from non-aggregated Aβ₄₀ incubated with IDE.

3.4. Non-aggregated Aβ₄₀ incubated with both NEP and IDE

Since multiple proteases are involved in Aβ elimination in vivo, Aβ₄₀ was incubated with both NEP and IDE (Fig. 3). The total ion current chromatogram (TICC) and the identified peptide fragments were almost identical to those obtained using NEP only (Fig. 1). In addition, the long peptide fragments listed in Table 2 (12∼ amino-acid residues, obtained using IDE) were not detected, suggesting that the long fragment peptide generated by IDE was further digested by NEP. Each amino acid sequence and its t_R identified by LC-MS with LC systems 1 and 2 is summarized in Table 3.

Fig. 3.

TICCs of non-aggregated Aβ₄₀ incubated with both NEP and IDE for 3 days (top) and 7 days (bottom). (A) LC system 1 (ODS for hydrophobic peptides), (B) LC system 2 (PGC for polar peptides). Identified peptides are shown in Table 3, together with the charge state, precursor MS (m/z), and t_R (min).

Table 3.

Peptide fragments found from non-aggregated Aβ₄₀ incubated with both NEP and IDE.

3.5. Non-aggregated Aβ₄₀ incubated with IDE followed by NEP

Since the experiment above suggested that the long fragment peptide obtained using IDE could be further digested by NEP, Aβ₄₀ was incubated with IDE, followed by digestion using NEP. The TICC pattern of the peptide fragments (Fig. 4) was identical to the one obtained using NEP only (Fig. 1). In addition, the long peptide fragments obtained by digestion with IDE (12∼ amino acid residues) were not detected, similar to the results obtained by digestion with both NEP and IDE (Fig. 3), confirming that the long fragment peptides obtained using IDE were further digested by NEP. Each amino acid sequence and its t_R identified by LC-MS with LC system 1 is summarized in Table 4.

Fig. 4.

TICCs of non-aggregated Aβ₄₀ incubated with IDE, followed by NEP. (A) Control (incubated with IDE only), (B) experiment (incubated with IDE, followed by NEP). Identified peptides are shown in Table 4, together with the charge state, precursor MS (m/z), and t_R (min).

Table 4.

Peptide fragments found from non-aggregated Aβ₄₀ incubated with IDE followed by NEP.

3.6. Aβ₁₆ incubated with NEP or IDE

To further compare the digestion patterns between NEP and IDE, a shorter Aβ₄₀ homologue (Aβ₁₆) was incubated with IDE or NEP. The amino acid sequence of Aβ₁₆ contains 4 and 5 cleavage sites identified in experiments using Aβ₄₀ digested by NEP (Table 1) and IDE (Table 2), respectively. The cleavage sites, shown as slashes, are: NEP, DAE/FR/HDSG/YE/VHHQK; IDE, D/AEFRHD/SGYEVH/H/Q/K. Thus, Aβ₁₆ was digested by NEP into the five peptides D¹AE³, F⁴R⁵, H⁶DSG⁹, Y¹⁰E¹¹, and V¹²HHQK¹⁶ (Fig. 5 AB), and the cleavage sites were identical to those in Aβ₄₀. In contrast, Aβ₁₆ was not digested by IDE (Fig. 5CD). These results suggest that IDE digests Aβ₄₀ by recognizing the whole structure, whereas NEP digests Aβ₄₀ by recognizing the partial structure.

Fig. 5.

TICCs of non-aggregated Aβ₁₆ incubated with NEP (A and B) or IDE (C and D). LC system 3 (ODS for hydrophobic peptides) was used for (A) and (C). LC system 2 (PGC for polar peptides) was used for B and D.

3.7. Aggregated Aβ₄₀ incubated with NEP, IDE or trypsin

The Aβ sequence has an amphipathic character because the N-terminal segment is hydrophilic whereas the C-terminal segment is hydrophobic. Monomeric Aβ prefers to adopt random coil or α-helix structures, but gradually changes to a β-sheet structure during the aggregation process [6]. Aggregated Aβ₄₀ was prepared by monitoring using the ThT assay, which is a β-sheet-specific fluorescence assay used as the “gold standard” for selectively identifying amyloid fibrils [28]. The fluorescence intensity was maximum after 24 h incubation and then remained essentially unchanged. After confirming aggregation, a control solution prepared without ThT (no ThT in the digestion samples) was mixed with NEP, IDE, or trypsin (positive control, cleavage at the C-terminus of R and K), and incubated at 37 °C. Aggregated Aβ₄₀ was not digested by either NEP or IDE by prolonged incubation up to 7 days (Fig. 6AB), suggesting that Aβ₄₀ gained resistance against digestion by both NEP and IDE. In contrast, aggregated Aβ₄₀ was digested by trypsin to form four peptide fragments: D¹AEFR⁵, H⁶DSGYEVHHQK¹⁶, L¹⁷VFFAEDVGSNK²⁸, and G²⁹AIIGLMVGGVV⁴⁰ (Fig. 6C).

Fig. 6.

TICCs of aggregated Aβ₄₀ incubated with NEP (A), IDE (B), or trypsin (C). Incubated for 3 days (top) and 7 days (bottom).

4. Conclusion

In this article, we described the mass spectrometric characterization of the digestion specificities of NEP and IDE for non-aggregated Aβ₄₀, aggregated Aβ₄₀, and Aβ₄₀ peptide fragments, including Aβ₁₆. Howell et al. [32] and Leissring et al. [33] reported similar degradation studies but overlooked several polar peptide fragments and did not examine the relationship between NEP and IDE. We identified all the peptide fragments from non-aggregated Aβ₄₀, as follows (Fig. 7): NEP, 23 peptide fragments consisting of 2–11 amino-acid residues (17 cleavage sites), and for IDE, 23 peptide fragments consisting of 6–33 amino-acid residues (15 cleavage sites). Our use of a PGC column [29] made it possible to retain polar peptide fragments and to identify novel cleavage sites: for example, a recent review suggested that Aβ₄₀ is cleaved at only 10 and 5 sites by NEP and IDE, respectively [12]. Also, we confirmed that IDE can digest only whole Aβ₄₀, in contrast with NEP, which can digest whole Aβ₄₀, partial structures such as Aβ₁₆, and peptide fragments generated from Aβ₄₀ by IDE digestion. Furthermore, we confirmed that IDE and NEP cannot digest aggregated Aβ₄₀, although trypsin can.

Fig. 7.

Identified peptide fragments from non-aggregated Aβ₄₀ generated by NEP, 20 peptides consisting of 2–11 amino-acid residues (16 cleavage sites) and IDE, 23 peptides consisting of 6–33 amino-acid residues (15 cleavage sites). Amino acids written in red (E²²DVGS²⁶) indicate the turn region. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Several previously reported articles support our results. Shen et al. reported the detailed substrate recognition and catalytic mechanism of IDE by digesting insulin, Aβ₄₀, amylin, and glucagon with IDE [34]. The N-terminal and C-terminal domains of IDE form a pocket to encapsulate substrates [34], and the change from the open conformation to the closed conformation results in digestion [35]. Therefore, large oligomers and fibril Aβ₄₀ are too big to enter the catalytic site, in contrast to small monomeric Aβ₄₀ in a random coil or α-helix conformation. This can also explain why APP is not digested by IDE [32]. It is interesting that several peptides containing the turn region (E²²DVGS²⁶) were found in non-aggregated Aβ₄₀ digested by IDE: insulin contains a turn region that is the degradation target of IDE [34]. NEP utilizes a different catalytic mechanism because it is more flexible regarding substrate size. We found that NEP digested smaller peptide fragments generated from Aβ₄₀, including Aβ₁₆. Kanemitsu et al. [36] reported that NEP can degrade the dimer of Aβ₄₀, and the dimer, trimer and tetramer of Aβ₄₂. They did not examine the degradation of aggregated Aβ. However, these findings suggest that the flexibility of NEP regarding substrate size can cover not only Aβ peptide fragments but also small Aβ oligomers.

The therapeutic utility of Aβ-degrading enzymes in AD has been studied [[9], [10], [11], [12]], and thus the resulting peptide fragments in plasma and cerebrospinal fluid should be analyzed to estimate their usefulness as Aβ clearance markers. Moreover, Kallikrein-related peptidase 7 [37] and synthetic peptides derived from the Box A region of Tob 1 protein [38] were recently reported to degrade even aggregated Aβ. Therefore, if there are specific Aβ peptide fragments derived from aggregated Aβ in senile plaques, they could be used as therapeutic markers.

CRediT author contribution statement

Dai Kato: Investigation, Data Curation, Visualization. Yoshiaki Takahashi: Investigation, Data Curation. Haruto Iwata: Investigation, Data Curation. Yusuke Hatakawa: Project administration. Seon Hwa Lee: Project administration, Writing - Review & Editing. Tomoyuki Oe: Supervision, Conceptualization, Methodology, Writing - Original Draft, Writing - Review & Editing, Funding acquisition.

Declaration of competing interest

The authors declare that there is no conflict of interest.

Data availability

Data will be made available on request.