Post-Translational Chemical Modification of Amyloid-β Peptides by 4-Hydroxy-2-Nonenal

Abstract

Background:

The extraction and quantification of amyloid-β (Aβ) peptides in brain tissue commonly uses formic acid (FA) to disaggregate Aβ fibrils. However, it is not clear whether FA can disaggregate post-translationally modified Aβ peptides, or whether it induces artifact by covalent modification during disaggregation. Of particular interest are Aβ peptides that have been covalently modified by 4-hydroxy-2-nonenal (HNE), an oxidative lipid degradation product produced in the vicinity of amyloid plaques that dramatically accelerates the aggregation of Aβ peptides.

Objective:

Test the ability of FA to disaggregate Aβ peptides modified by HNE and to induce covalent artifacts.

Methods:

Quantitative liquid-chromatography-tandem-mass spectrometry of monomeric Aβ peptides and identify covalently modified forms.

Results:

FA disaggregated ordinary Aβ fibrils but also induced the time-dependent formylation of at least 2 residue side chains in Aβ peptides, as well as oxidation of its methionine side chain. FA was unable to disaggregate Aβ peptides that had been covalently modified by HNE.

Conclusion:

The inability of FA to disaggregate Aβ peptides modified by HNE prevents FA-based approaches from quantifying a pool of HNE-modified Aβ peptides in brain tissue that may have pathological significance.

Keywords

Ascorbate copper deuterium-stabilized fatty acids liposomes mass spectrometry mouse oxidative stress polyunsaturated

INTRODUCTION

The accumulation of amyloid-β (Aβ) peptides as aggregated fibrils in cortical brain tissue is one of the pathological hallmarks of Alzheimer’s disease (AD), but quantifying the extent of this accumulation is challenging. Agents that selectively bind to fibrils (e.g., Congo red, thioflavin dyes) are imprecise due to their variable affinity for amyloid fibrils and non-stoichiometric binding, while others (e.g., Pittsburgh compound B, florbetapir) are intended for imaging the distribution and extent of accumulation rather than for quantifying a precise amount. Monomeric Aβ peptides may be quantified with sandwich ELISA assays, but they can quantify fibrillar Aβ only to the extent that fibrils from brain tissue may be extracted and disaggregated into monomeric Aβ peptides. Consequently, their accuracy is compromised to the extent that extraction and disaggregation of fibrils is incomplete, that antibody recognition sites are altered by chemical modification occurring during extraction and disaggregation, and that sequence variations at both the amino and carboxy termini occur outside the antibody recognition sites. Some of these problems may be solved by liquid-chromatography-tandem-mass spectrometry (LC-MS/MS) and the use of internal standards, but LC-MS/MS remains susceptible to inaccuracy if extracted fibrils are not completely disaggregated, or if sample processing induces unanticipated artifactual chemical modifications.

The present study focuses on whether formic acid (FA) can disaggregate Aβ peptides that have been covalently modified by 4-hydroxy-2-nonenal (HNE). This focus on HNE and HNE-modified Aβ peptides is prompted by several lines of evidence suggesting a possible role for HNE in the pathogenesis of AD. First, a role for oxidative stress in AD pathogenesis has broad evidentiary support [1], HNE is a natural product of oxidative stress acting on arachidonic acid (ARA) [2]— a polyunsaturated fatty acyl chain that is concentrated in the brain, and HNE concentrations are elevated in AD [3 –6]. Second, HNE spontaneously reacts with unprotonated His side chains via Michael addition [7], and elevated levels of HNE-modified His side chains are present in the hippocampus of patients with AD [8]. When the His side chains of Aβ peptides are modified in this way, their tendency to aggregate increases dramatically [9], especially in the presence of lipid membranes [10 –13]. Third, and most importantly, proteins in the immediate vicinity of amyloid plaques are heavily modified by HNE, as demonstrated in mouse [14, 15], canine [16, 17], and human [18 –20] brain, with a distribution having been likened to a “ring of fire” around amyloid plaques. Thus, the modification of proteins in the brain by HNE features prominently in a plausible mechanism linking oxidative stress to the formation of amyloid plaques [1]. The link is especially provocative when the ability of Aβ peptides to bind copper ions, aggravate oxidative stress, and produce HNE is considered [21, 22].

The experiments described below have examined the ability of 70% FA to disaggregate Aβ fibrils that have been covalently modified by HNE. The question of whether FA can disaggregate Aβ fibrils is important because the amount of HNE-modified Aβ peptide in the brain is unknown, and quantifying it requires disaggregation into monomeric forms. The reactions involved are illustrated schematically in Fig. 1. Unmodified 40-residue monomeric Aβ peptides (Aβ₄₀) form amyloid fibrils (Aβ_40F) spontaneously at concentrations above 10μM (reaction ‘a’) [23]. 70% FA disaggregates fibrils back into monomers (Aβ_40M, reaction ‘b’). Oxidative lipid degradation products such as HNE are known to be produced in the immediate vicinity of amyloid plaques [15 , 20] and react spontaneously with the 3 histidine side chains in Aβ₄₀ [7 , 25]. When HNE reacts with Aβ_40M rapid aggregation ensues (reaction ‘e’). Alternatively, HNE may react with preformed Aβ_40F (reaction ‘c’). Upon treatment of HNE-treated fibrils with FA, they may disaggregate, lose any HNE modifications that have occurred, or both (reactions ‘d’ and ‘f’). FA treatment may also create artifactual species by chemically modifying Aβ₄₀ to create formate esters with serine side chains (reaction ‘g’), or sulfoxides/sulfones in place of methionine side chains (reaction ‘h’).

Fig. 1

Scheme outlining the chemical reactions considered in the present study. Reaction (a) represents the spontaneous aggregation of Aβ_40M into fibrils over time, whereas (b) represents the disaggregation of Aβ_40F upon treatment with 70% formic acid (FA). Reaction (c) represents the spontaneous reaction of HNE with pre-formed fibrils, while reactions (d) represent the disaggregation of HNE-treated pre-formed fibrils into Aβ_40M and Aβ₄₀-HNE adducts. Upon disaggregation, both products must be considered because HNE modification of fibrils may not be complete. Reaction (d) also allows for the possibility that FA may release Aβ_40M from an HNE-Aβ_40M adduct. Reaction (e) represents the spontaneous (and rapid) reaction between Aβ_40M and HNE to form fibrils, while the three reactions labeled (f) represent the release of Aβ_40M from such fibrils upon treatment with 70% FA, either directly or by release Aβ_40M from the HNE adduct. Reactions labeled (g) represent the formation of formic acid esters on Aβ_40M, while reactions labeled (h) represent the oxidation of Met35 (+O) upon treatment with FA. Reactions (g) and (h) may combine to yield analogs of Aβ_40M that have both FA esters and oxidized Met35 residues. In the present study, Aβ_40M (indicated by a solid green circle) was quantified by LC-MS/MS using an isotope-labeled internal standard. Fibrillar forms of Aβ₄₀ (indicated by dashed orange circles) were morphologically characterized by EM. Analogs of Aβ₄₀ bearing FA esters or oxidized Met35 residues (indicated by a dotted light blue circles) were semi-quantitatively characterized by LC-MS/MS. Analogs of Aβ₄₀ bearing both FA esters and oxidized Met35 residues were not characterized.

These reactions have been studied by means of quantitative LC-MS/MS with a stable-isotope-labeled form of Aβ₄₀ as an internal standard. The use of LC-MS/MS provides for the precise and unambiguous quantitation of Aβ_40M and insight into the creation of chemically modified forms of Aβ₄₀ under conditions used to disaggregate fibrils.

METHODS

Chemicals and reagents

1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) was obtained from Sigma-Aldrich (St. Louis, MO) and redistilled before use. Amyloid-β (1–40) prepared from HFIP (Aβ₄₀) was purchased from rPeptide (Bogart, GA). ¹³C₆-Leu was obtained from Cambridge Stable Isotopes (Tewksbury, MA). ¹³C₁₂-Aβ₄₀ for use as an internal standard was synthesized with ¹³C₆-Leu at positions 17 and 34 by Life Tein (Somerset, NJ). Its concentration was determined by integrated absorbance at 280 nm, using an extinction coefficient of 1137 AU M^–1 cm^–1 and an in-line absorbance detector. HNE was obtained from Cayman Chemical (Ann Arbor, MI). FA and trifluoroacetic acid (TFA) were obtained from Fluka. LC/MS grade acetonitrile and 0.1 N hydrochloric acid (HCl) were obtained from Fisher Scientific (Waltham, MA).

Aβ₄₀ preparation, aggregation, and disaggregation

Monomeric Aβ₄₀ (Aβ_40M) was prepared by dissolving the contents of a tube of commercially-prepared powder, labeled “1 mg”, in 1 mL of a mixture of HFIP and 8.3 mM HCl solution (2:3, v/v). This solvent mixture helps disaggregate preformed fibrils, removes trace amounts of any trifluoroacetic acid added during synthesis or purification, and is easily removed by lyophilization.

This solution was lyophilized overnight, then redissolved in the same solution and re-lyophilized overnight. The resulting material was dissolved in 2.3 ml of PBS to yield a solution that was nominally 100μM, and fibril formation was slowed for up to 8 h by keeping the sample at 4°C. Fibrillar Aβ₄₀ (Aβ_40F) was then formed by warming the nominally 100μM PBS solution to 37°C and incubating without agitation for 5 days. Disaggregation was accomplished by adding sufficient FA to make the solution 70% FA.

The effects of HNE on Aβ₄₀ aggregation and disaggregation

The effects of HNE on Aβ_40M aggregation was examined by adding 10μL of a 10 mg/mL HNE in ethanol to 600μL of nominally 100μM Aβ₄₀ in PBS, which yielded a molar ratio of HNE:Aβ₄₀ slightly greater than 10:1. The effects of HNE on Aβ_40F disaggregation was examined by adding 10μL of a 10 mg/mL HNE in ethanol to a 600μL portion of the 100μM Aβ_40M solution that had been incubated at 37°C without agitation for 5 days to form Aβ_40F.

Quantitation of Aβ_40M and modified forms

The concentration of Aβ_40M was quantified by vortex mixing for 1 min, transferring 40μL into a 7×20 mm polypropylene tube, and centrifuging at 33,000× g for 1 h at 4°C to pellet any formed fibrils. An aliquot of the supernatant was mixed with a calibrated amount of ¹³C₁₂-Aβ₄₀ and sufficient neat FA to make the solution 70% FA, and 15μL of the resulting mixture was analyzed by LC-MS/MS. In all cases, therefore, the concentration of Aβ_40M was derived from the ratio of Aβ₄₀ and ¹³C₁₂-Aβ₄₀ integrated areas and the known concentration of the internal ¹³C₁₂-Aβ₄₀ standard.

Samples were injected into a 50 mm×1 mm 3.5μm Eclipse XDB-C8 column (Agilent, Santa Clara, CA) maintained at 60°C in a column oven. All tubing ahead of the column was steel, to avoid nonspecific adsorption. The mobile phase consisted of 0.01% TFA in water-acetonitrile (95:5, v/v) (solvent A) and 0.01% TFA in acetonitrile (solvent B). The TFA concentration was reduced from the more commonly used 0.1% to improve signal strength. The solvent gradient program was: 0–1 min 20% B, 1–5 min 20⟶100% B, 5–7 min 100% B, 7–9 min 100⟶20% B with a flow rate of 0.1 mL/min.

Aβ₄₀, ¹³C₁₂-Aβ₄₀, formic acid esters of Aβ₄₀ (Aβ₄₀-FA and Aβ₄₀-FA₂), Aβ₄₀ with one, two, or three HNE Michael adducts (Aβ₄₀-HNE, Aβ₄₀-HNE₂, Aβ₄₀-HNE₃), Aβ₄₀(MetO), and Aβ₄₀(MetO₂) were detected by LC-MS/MS in positive MRM mode (Table 1) on a Sciex API 4000 mass spectrometer. Instrument parameters for all monitored transitions were curtain gas 10 psi, nebulizer gas (GS1) 50 psi, drying gas (GS2) 20 psi and 450°C, nitrogen collision gas 12 psi, and entrance potential 10 V. Other instrumental parameters vary with the transitions monitored and are listed in Table 1. In each case, the b39 ion was monitored in the +4 charge state.

Table 1

Parameters for MS/MS analysis of Aβ₄₀, Aβ₄₀-FA esters, and Aβ₄₀-HNE adducts

Compounds	Formula	Precursor ion* (m/z)	Product ion** (m/z)	DP (V)	CE (eV)	CXP (V)
¹²C-Aβ40	C₁₉₄H₂₉₅N₅₃O₅₈S	1083.3	1054.0	121	33	34
¹³C-Aβ40 (Int.Std.)	C₁₉₄H₂₉₅N₅₃O₅₈S	1086.3	1057.0	121	33	34
¹²C-Aβ40-FA	C₁₉₅H₂₉₅N₅₃O₅₉S	1090.3	1061.5	85	34	35
¹²C-Aβ40-FA₂	C₁₉₆H₂₉₅N₅₃O₆₀S	1097.3	1068.5	75	34	37
¹²C-Aβ40-HNE	C₂₀₃H₃₁₁N₅₃O₆₀S	1122.3	1093.6	75	36	37
¹²C-Aβ40-HNE₂	C₂₁₂H₃₂₇N₅₃O₆₂S	1161.4	1032.6	80	36	38
¹²C-Aβ40-HNE₃	C₂₂₁H₃₄₃N₅₃O₆₄S	1200.4	1171.6	100	38	40
¹²C-Aβ40-MetO	C₁₉₄H₂₉₅N₅₃O₅₉S	1087.3	1058.0	121	33	34
¹²C-Aβ40-MetO₂	C₁₉₄H₂₉₅N₅₃O₆₀S	1091.3	1062.0	121	33	34
¹²C-Aβ40-HNE-H₂O	C₂₀₃H₃₀₉N₅₃O₅₈S	1117.8	1089.1	75	36	37
¹²C-Aβ40-(HNE-H₂O)₂	C₂₁₂H₃₂₃N₅₃O₆₀S	1152.3	1123.1	80	36	38
¹²C-Aβ40-(HNE-H₂O)₃	C₂₂₁H₃₃₇N₅₃O₆₁S	1186.4	1158.1	100	38	40

^*+4 charge state. ^**b39 ion in the +4 charge state.

Internal standards for Aβ₄₀ FA esters, Aβ₄₀ with HNE adducts, and Aβ₄₀ with oxidized Met35 residues were not available. Therefore, variable factors such as ionization and detection efficiencies can affect signal magnitudes and could not be corrected. However, injection volumes were held constant, and the signals arising from calibrated amounts of ¹³C₁₂-Aβ₄₀ were consistent with a coefficient of variation of 5.0%. Therefore, quantitative information about Aβ₄₀ FA esters, Aβ₄₀ with HNE adducts, and Aβ₄₀ with oxidized Met35 residues are presented as relative amounts of each species at the times indicated, but as individual graphs so that a quantitative comparison of one species to another is not implied.

Fibril morphology

Aβ₄₀ samples (∼100 ng) in 5μL of buffer were placed onto freshly glow-discharged carbon films on 300 mesh nickel grids (Electron Microscopy Sciences, Hatfield, PA, USA) for 2 min and blotted with filter paper. The blotted samples were then negatively stained using 1% (w/v) ammonium molybdate (Sigma-Aldrich), adjusted to pH 7.4 with ammonium hydroxide, applied for 2 min, blotted, and air-dried. Images were recorded using a JEM-1010 transmission electron microscope (JEOL, Tokyo, Japan), operating at 80 kV, and equipped with a side-mounted CCD digital camera.

RESULTS

Aβ₄₀ aggregation and disaggregation

The time course of Aβ_40M aggregation (Fig. 1, reaction a) was characterized by measuring the amount of Aβ_40M remaining in solution over 5 days. The concentration of Aβ_40M with a nominal concentration of 100μM (i.e., based on the supplier’s package label) was verified to be 110±17μM on day 0, and it declined to <1μM on day 3 (Fig. 2A). Fibrils with a typical morphology were observed by EM (Fig. 3A). Treatment with 70% FA (Fig. 1, reaction b) on day 5 restored the concentration of Aβ_40M to 113±14.0μM within 10 min (Fig. 2A). Thereafter, the concentration of Aβ_40M decreased to 36±2.4μM within 24 h. On day 7 (48 h after FA treatment), no fibrils were observed, but many small globular structures reminiscent of objects previously designated soluble oligomers were present (Fig. 3B).

Fig. 2

Aβ_40M recovery after fibril formation. A) Aβ_40M concentrations as a nominally 100μM solution was allowed to fibrillize, and then treated with 70% FA on day 5. Absolute concentrations in this panel were determined with an isotope-labeled internal standard. The symbols each represent the result of a single determination, and three independent results are plotted on each day although they are often obscured by overlap. A line is drawn connecting the mean values. B) Signals arising from Aβ₄₀-FA esters 10 min (=day 5), 24 h (=day 6), and 48 h (=day 7) after FA addition. C) Signals from Aβ₄₀-FA₂ ester after FA addition. The signals in B and C are only semi-quantitative and are not necessarily comparable between panels. In all panels, the results of 3 determinations are plotted on each day, and a line is drawn connecting the mean values.

Fig. 3

Aβ_40F morphologies by EM. A) 100μM Aβ_40M after incubation for 5 days at 37°C, without agitation. B) Same as (A) 2 days after treatment with 70% FA. C) 100μM Aβ_40M treated with ∼1 mM HNE after incubation for 5 days at 37°C, without agitation. D) Same as (C) 2 days after treatment with 70% FA. E) Aβ_40F (formed after incubating 100μM Aβ_40M for 5 days at 37°C, without agitation) treated with ethanol. F) Aβ_40F (formed after incubating 100μM Aβ_40M for 5 days at 37°C, without agitation) treated with ∼1 mM HNE in ethanol.

To determine whether the decrease of Aβ_40M after day 5 was due to the artifactual formation of FA esters, the time courses for the formation of Aβ₄₀-FA and Aβ₄₀-FA₂ are shown in Fig. 2B and 2C (Fig. 1, reactions g). Following FA treatment, signals corresponding to Aβ₄₀-FA and Aβ₄₀-FA₂ increased significantly after 24 h, but then decreased (Fig. 2B, C). These determinations were not quantitative, however, because internal standards were not prepared for the multiple chemical species and isomeric forms that are produced. Therefore, the extent to which the formation of FA esters accounted for the decrease of Aβ_40M after day 5 is not known.

The methionine side chain is susceptible to spontaneous oxidation in the presence of oxygen, forming both the sulfoxide (Aβ₄₀-MetO) and the sulfone (Aβ₄₀-MetO₂). These reactions may be accelerated under acidic conditions and hence, by FA treatment. To determine whether the decrease of Aβ_40M after day 5 was due to either spontaneous or acid-catalyzed methionine oxidation (Fig. 1, reactions h), reference samples of Aβ₄₀-MetO and Aβ₄₀-MetO₂ were prepared by mixing 10μL of 3% H₂O₂ and 90μL of 10μM Aβ_40M (molar ratio of H₂O₂ to Aβ_40M of approximately 9800:1). After 20 min at room temperature, 12μL of this mixture and 28μL of FA were mixed to yield 70% FA, and 15μL of this solution was analyzed by LC-MS/MS. Under conditions where unmodified Aβ₄₀ eluted at 4.1 min (Fig. 4A,B), H₂O₂ treatment yielded a parent ion with m/z = 1091.5 eluting at 3.8 min, and disappearance of the ion corresponding to Aβ₄₀ (Fig. 4C,D). Product ion analysis of the 1091.5 peak yielded a peak at 1062.3, indicating that it was Aβ₄₀-MetO₂ (Fig. 4E,F).

Fig. 4

LC-MS/MS analysis of Aβ₄₀-MetO and Aβ₄₀-MetO₂ produced by H₂O₂ treatment of Aβ₄₀. A) TIC of Aβ₄₀. B) Q1 of the peak in (A) at 4.0 min. C) TIC of Aβ₄₀ with H₂O₂. D) Q1 of the peak in (C) at 3.8 min. E) TIC of a product ion at m/z = 1091.5 of Aβ₄₀ with H₂O₂. F) Product ion spectrum of the peak in (F) at 3.9 min.

To determine whether FA treatment alone could contribute to the decrease of Aβ_40M after day 5 by causing Met35 oxidation, Aβ_40M was treated with 70% FA and immediately analyzed by LC-MS/MS. A large peak corresponding to Aβ₄₀-MetO (Fig. 5) was detected within 1 min of FA treatment, suggesting that the decrease of Aβ_40M after day 5 was due to both the formation of FA esters and the oxidation of methionine. A much smaller signal arising from Aβ₄₀-MetO₂ was detected.

Fig. 5

LC-MS/MS analysis of Aβ₄₀-MetO and Aβ₄₀-MetO₂ produced by FA treatment. A) Single reaction monitoring of the m/z 1083.3⟶1054.0 transition for Aβ₄₀. B) Single reaction monitoring of the m/z 1087.3⟶1058.0 for Aβ₄₀-MetO. C) Single reaction monitoring of the m/z 1091.3⟶1062.0 for Aβ₄₀-MetO₂ (note that the split signal for this transition is three orders of magnitude smaller than the signal for Aβ₄₀-MetO).

The effects of HNE on Aβ_40M aggregation

HNE added to Aβ_40M on day 0 increased the rate of Aβ_40M aggregation (Fig. 1, reaction e), with the loss of Aβ_40M ∼99% complete by day 1 (Fig. 6A). The initial concentration of Aβ_40M was nominally 100μM, but when measured 70 min after adding HNE, it was only 64.0±9.6μM. The loss of Aβ_40M may be due to the formation of Aβ₄₀-HNE adducts or aggregation into fibrils during the 60 min required for centrifugation and 10 min required for LC-MS/MS. The addition of FA on day 5 (Fig. 1, reactions f) only restored the concentration of Aβ_40M to 0.6±0.1μM (Fig. 6A).

Fig. 6

A) Aβ_40M concentrations as a nominally 100μM solution is allowed to fibrillize in the presence of ∼1 mM HNE at 37°C with HNE without agitation and is then treated with 70% FA on day 5. Absolute concentrations in this panel were determined with an isotope-labeled internal standard. The symbols each represent the result of a single determination, and three independent results are plotted on each day although they are often obscured by overlap. A line is drawn connecting the mean values. B) Signals from Aβ₄₀-HNE 10 min (=day 5), 24 h (=day 6), and 48 h (=day 7) after FA treatment. C) Signals from Aβ₄₀-HNE₂ after FA treatment. D) Signals from Aβ₄₀-HNE₃ after FA treatment. The signals in B-D are only semi-quantitative and are not necessarily comparable between panels. In all panels, the results of three determinations are plotted on each day, and a line is drawn connecting the mean values.

After HNE addition, 5 days of incubation, and FA addition, Aβ₄₀-HNE, Aβ₄₀-HNE₂, and Aβ₄₀-HNE₃ were readily detected on day 5, but their signals declined after day 5, due either to acid-catalyzed release of the HNE adduct, or by dehydration of the hemiacetal (Fig. 6B-D). Again, however, these determinations only reflect signal strength, and are not quantitative. Acid-catalyzed release of the HNE adduct seems unlikely, however, because unmodified Aβ₄₀ did not appear as the HNE-modified forms declined (Fig. 6A). Each HNE adduct adds 156 amu to the mass of Aβ₄₀ when forming a hemiacetal with the His side chains. Each hemiacetal may subsequently lose H₂O to form an adduct that adds only 138 amu to the mass of Aβ₄₀, and the neutral loss of H₂O may occur in the electrospray source. HNE may also add 138 amu to the mass of Aβ₄₀ by forming a Schiff base with Lys side chains, although these adducts generally require reduction with borohydride to stabilize them for detection. In either case, these adducts cannot be distinguished under the conditions of our measurements. However, HNE adducts corresponding to +138 atomic mass units yielded much larger signals than adducts corresponding to +156 amu (Fig. 7).

Fig. 7

LC-MS/MS signals arising after Aβ_40M was incubated with HNE for 9 days at 37°C without agitation. On day 9, the samples were treated with 70% FA and immediately analyzed by LC-MS/MS for Aβ_40M with one, two, or three +156 amu adducts, and one, two, or three +138 amu adducts.

HNE causes Aβ_40M to form clumps of much thinner “winding fibrils” (Fig. 3C), while FA treatment of these winding fibrils yields many small globules, similar to Fig. 3B, and reminiscent of objects previously designated soluble oligomers (Fig. 3D).

The effects of HNE on Aβ_40F disaggregation

HNE in ethanol was added to Aβ_40F, which had formed after a 5-day incubation of Aβ_40M at 37°C without agitation (Fig. 1, reaction c). After 2 additional days of incubation at 37°C without agitation (i.e., on day 7), the concentration of Aβ_40M was 0.04±0.01μM (Fig. 8A). These HNE-treated fibrils were then treated with 70% FA for 10 min (Fig. 1, reactions d) and the Aβ_40M concentration was found to be 5.1±0.1μM. In contrast, initial treatment with ethanol instead of HNE yielded 76.0±4.9μM (Fig. 8B). As demonstrated previously, HNE spontaneously forms Michael adducts with the three His side chains of Aβ₄₀, forming Aβ₄₀-HNE, Aβ₄₀-HNE₂, and Aβ₄₀-HNE₃ [9]. All three of these species were detected after FA treatment on day 7 by LC-MS/MS, but their signals subsequently declined, due either to the release of the HNE adduct, or by dehydration of the hemiacetal (Fig. 9A–C). Concurrently, signals corresponding to Aβ₄₀-FA and Aβ₄₀-FA₂ increased for one day then appeared to plateau (Fig. 10). The morphology of Aβ_40F was unaffected by HNE as well as FA treatment (Fig. 3E, F), indicating that FA was unable to disaggregate Aβ_40F after HNE treatment.

Fig. 8

The effect of HNE on the disaggregation of Aβ_40F. Fibrils formed by a 5-day incubation of Aβ_40M were treated with HNE on day 5, and with FA on day 7. A) The treatment of preformed fibrils with HNE in ethanol precluded disaggregation by FA. B) Treatment with ethanol alone did not preclude disaggregation by FA. Absolute concentrations in this panel were determined with an isotope-labeled internal standard. The symbols each represent the result of a single determination, and three independent results are plotted on each day although they are often obscured by overlap. A line is drawn connecting the mean values.

Fig. 9

Fibrils formed by a 5-day incubation of Aβ_40M were treated with HNE on day 5, and 70% FA on day 7. A) Signals corresponding to Aβ₄₀-HNE. B) Signals from Aβ₄₀-HNE₂. C) Signals from Aβ₄₀-HNE₃. The symbols each represent the result of a single determination, and three independent results are plotted on each day although they are often obscured by overlap. A line is drawn connecting the mean values.

Fig. 10

Fibrils formed by a 5-day incubation of Aβ_40M were treated with HNE on day 5, and 70% FA on day 7. A) Signals from Aβ₄₀-FA ester. B) Signals from Aβ₄₀-FA₂ ester. The symbols each represent the result of a single determination, and three independent results are plotted on each day although they are often obscured by overlap. A line is drawn connecting the mean values.

DISCUSSION

The foregoing results are novel because they highlight the potential of HNE, a compound known to be produced in the vicinity of amyloid plaques, to confound the quantitative analysis of the Aβ peptides comprising those plaques. Aβ fibrils must be disaggregated into monomeric Aβ peptides for accurate quantitation by LC-MS/MS. Disaggregation is most commonly accomplished with 70% FA, but the duration of treatment and conditions such as temperature have not been heretofore evaluated with respect to the quantitative recovery of monomeric Aβ peptides. The use of mass spectrometry with isotope-labeled internal standards in this work assures us that the peptide is positively identified and monomeric. Our results demonstrate the quantitative recovery of Aβ_40M from fully-formed fibrils after only 10 min in 70% formic acid at room temperature (Fig. 2A).

Although formic acid treatment rapidly disaggregates Aβ fibrils, formic acid can also induce the formation of artifactual species, in particular formate esters with the side chain –OH groups of Ser10 and Ser26 in Aβ₄₀, and possibly also of Tyr10. Our results demonstrate that monomeric Aβ₄₀ declines dramatically within 24 h of beginning treatment with 70% formic acid, and that peptide species consistent with the formation of single and double formate esters are formed concomitantly (Fig. 2B, C). Formic acid may also cause modification of the Met35 side chain. Dimethyl sulfoxide (DMSO) is commonly used to aid in the solubilization of Aβ peptides, and the acid-catalyzed conversion of the methionine side chain to the sulfoxide and sulfone by DMSO is a known process [26, 27]. In this study, we found that significant amounts of Aβ₄₀-MetO (the sulfoxide) and Aβ₄₀-MetO₂ (the sulfone) were produced by 70% formic acid treatment even in the absence of DMSO (Fig. 5).

Another possible explanation for an apparent inability to disaggregate fibrils is crosslink formation. HNE is bifunctional and capable of forming both a Michael adduct and a Schiff base. We have previously examined HNE-modified Aβ peptides that were solubilized by limited proteolysis [9]. Under conditions of 10-fold excess HNE, Aβ peptides with one, two, or three Michael adducts on His side chains are formed. However, there were no Aβ peptides with Schiff base or Michael adducts on the Lys side chains, consistent with an earlier study of ubiquitin [28] which does not aggregate upon HNE modification. Another reason to discount crosslink formation is that Schiff bases form reversibly, and are difficult to detect by mass spectrometry without reduction by borohydride [29].

Therefore, it is not clear whether formate ester formation is solely responsible for the disappearance of monomeric Aβ₄₀ in 70% formic acid over 24 h, or additional processes such as proteolysis or methionine oxidation contribute, but it is clear that the quantitative analysis of Aβ₄₀ monomer is best done shortly after treatment with 70% formic acid.

It has been shown that oxidative lipid degradation products such as HNE are produced in the immediate vicinity of amyloid plaques [15 , 20]. Therefore, the effects of histidine modification by HNE on our ability to disaggregate amyloid fibrils must be considered. Earlier studies have shown that HNE treatment strongly promotes Aβ₄₀ aggregation, even when HNE is only present in substoichiometric amounts [1 , 30]. In the present study, we confirmed that HNE causes rapid aggregation of Aβ₄₀, but also that HNE modification of monomeric Aβ₄₀ prevents the disaggregation of amyloid fibrils by 70% formic acid (path ‘e’ in Figs. 1, 6 and 7). It is significant that HNE modification of preformed fibrils (path ‘c’ in Fig. 1) also precludes disaggregation by 70% formic acid (Figs. 8–10). The mechanism of this interference may involve the loss of His side chains upon HNE modification so that they can no longer be protonated by 70% formic acid. In any case, HNE modification interferes with disaggregation by formic acid regardless of whether the modification occurred prior to fibril formation or after fibril formation. We have been unable to find conditions that disaggregate Aβ_40F after modification by HNE [31, 32], although disaggregation via proteolytic cleavage with preservation of the HNE adducts is clearly possible [9]. Proteolysis would obscure the nature of the full-length polypeptide, however, so that quantitation of post-translationally modified Aβ peptides in vivo remains a challenge.

This study focused on Aβ₄₀, whereas the amyloid that accumulates in AD pathology is a mixture of various N-terminal and C-terminal variants. Aβ₄₂, for example, appears to predominate in early amyloid plaques and some animal models of AD [33]. Aβ₄₂ is much less soluble than Aβ₄₀, and it is even more aggregation-prone than Aβ₄₀ when modified by HNE [11]. Therefore, it is likely that HNE-modified Aβ₄₂ fibrils will also resist disaggregation by FA. We have only succeeded in analyzing HNE-modified Aβ₄₂ after limited proteolysis [9]. However, proteolytic cleavage is unsuitable for quantitative analysis because of incomplete and variable cleavage efficiency, as well as an inability to place an internal standard in the same physicochemical state as the analyte.

The most common approach to Aβ peptide quantitation is the sandwich ELISA assay. Many commercial kits are available, and they are relatively easy to perform on batched samples. However, these assays cannot gauge the extent to which Aβ peptides are lost during extraction and disaggregation. Moreover, the various assay kits rely on antibodies with differing specificities, and the quantitation standards provided in the kits are typically proprietary. Users must rely on statements by the kit manufacturer about potential interference by modified forms of Aβ, if such statements are even made, or perform their own tests for interference on a particular kit.

Quantitation by LC-MS/MS is more time-consuming, instrumentation-intensive, and labor-intensive than ELISA assays. However, LC-MS/MS circumvents some of the most serious problems inherent in ELISA assays. For example, the use of an isotope-labeled internal standard controls for peptide losses during sample processing, while the reliance on retention time and the precise masses of parent/daughter ion m/z ratios renders peptide identification unambiguous. The chief limitations of mass spectrometry for quantitation are that species for which an isotope-labeled internal standard is not available cannot be quantified reliably, and that species not anticipated in advance are not detected.

In summary, the quantitative mass spectrometry analyses described herein provide insight that helps avoid pitfalls in the quantitative analysis of Aβ peptides, whether performed by mass spectrometry, ELISA, or another method. Peptide modification by formic acid must be anticipated and minimized for analysis, while problems caused by the in vivo modification of Aβ peptide by reactive substances such as HNE must be acknowledged.

DATA AVAILABILITY

The data supporting the findings of this study are available on request from the corresponding author.

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

This work was supported by a grant from the NIH (AG057197).

CONFLICT OF INTEREST

The authors have no conflicts of interest to report.

References

Axelsen

, Komatsu

, Murray IV

(2011) Oxidative stress and cell membranes in the pathogenesis of Alzheimer’s disease. Physiology (Bethesda) 26, 54–69.

Schneider

, Tallman

, Porter

, Brash

(2001) Two distinct pathways of formation of 4-hydroxynonenal. Mechanisms of nonenzymatic transformation of the 9-and 13- hydroperoxides of linoleic acid to 4-hydroxyalkenals. J Biol Chem 276, 20831–20838.

Lovell

, Ehmann

, Mattson

, Markesbery

(1997) Elevated 4-hydroxynonenal in ventricular fluid in Alzheimer’s disease. Neurobiol Aging 18, 457–461.

Markesbery

, Lovell

(1998) Four-hydroxynonenal, a product of lipid peroxidation, is increased in the brain in Alzheimer’s disease. Neurobiol Aging 19, 33–36.

McGrath

, McGleenon

, Brennan

, McColl

, McIlroy

, Passmore

(2001) Increased oxidative stress in Alzheimer’s disease as assessed with 4-hydroxynonenal but not malondialdehyde. QJM 94, 485–490.

Sayre

, Zelasko

, Harris

PLR

, Perry

, Salomon

, Smith

(1997) 4-hydroxynonenal-derived advanced lipid peroxidation end products are increased in Alzheimer’s disease. J Neurochem 68, 2092–2097.

Uchida

, Stadtman

(1992) Modification of histidine-residues in proteins by reaction with 4-hydroxynonenal. Proc Natl Acad Sci U S A 89, 4544–4548.

Fukuda

, Kanou

, Shimada

, Sawabe

, Saito

, Murayama

, Hashimoto

, Maruyama

, Ishigami

(2009) Elevated levels of 4-hydroxynonenal-histidine Michael adduct in the hippocampi of patients with Alzheimer’s disease. Biomed Res 30, 227–233.

Murray IV

, Liu

, Komatsu

, Uryu

, Xiao

, Lawson

, Axelsen

(2007) Membrane mediated amyloidogenesis and the promotion of oxidative lipid damage by amyloid beta proteins. J Biol Chem 282, 9335–9345.

10.

Koppaka

, Paul

, Murray IV

, Axelsen

(2003) Early synergy between Ab42 and oxidatively damaged membranes in promoting amyloid fibril formation by Ab40. J Biol Chem 278, 36277–36284.

11.

Murray IV

, Sindoni

, Axelsen

(2005) Promotion of oxidative lipid membrane damage by amyloid beta proteins. Biochem 44, 12606–12613.

12.

Komatsu

, Liu

, Murray IV

, Axelsen

(2007) A mechanistic link between oxidative stress and membrane mediated amyloidogenesis revealed by infrared spectroscopy. Biochim Biophys Acta 1768, 1913–1922.

13.

Liu

, Komatsu

, Murray IV

, Axelsen

(2008) Promotion of amyloid beta protein misfolding and fibrillogenesis by a lipid oxidation product. J Mol Biol 377, 1236–1250.

14.

Smith

, Hirai

, Hsiao

, Pappolla

, Harris

PLR

, Siedlak

, Tabaton

, Perry

(1998) Amyloid-beta deposition in Alzheimer transgenic mice is associated with oxidative stress. J Neurochem 70, 2212–2215.

15.

Ellis

, Fang

, Maheshwari

, Roltsch

, Holcomb

, Zimmer

, Martinez

, Murray IV

(2010) Lipid oxidation and modification of Amyloid-beta (A beta)and. J Alzheimers Dis 22, 593–607.

16.

Papaioannou

, Tooten

PCJ

, van Ederen

, Bohl

JRE

, Rofina

, Tsangaris

, Gruys

(2001) Immunohistochemical investigation of the brain of aged dogs. I. Detection of neurofibrillary tangles and of 4-hydroxynonenal protein, an oxidative damage product, in senile plaques. Amyloid 8, 11–21.

17.

Rofina

, Singh

, Skoumalova-Vesela

, van Ederen

, van Asten

, Wilhelm

, Gruys

(2004) Histochemical accumulation of oxidative damage products is associated with Alzheimer-like pathology in the canine. Amyloid 11, 90–100.

18.

Ando

, Brannstrom

, Uchida

, Nyhlin

, Nasman

, Suhr

, Yamashita

, Olsson

, El Salhy

, Uchino

, Ando

(1998) Histochemical detection of 4-hydroxynonenal protein in Alzheimer amyloid. J Neurol Sci 156, 172–176.

19.

Ghanbari

, Ghanbari

, Harris

PLR

, Jones

, Kubat

, Castellani

, Wolozin

, Smith

, Perry

(2004) Oxidative damage in cultured human olfactory neurons from Alzheimer’s disease patients. Aging Cell 3, 41–44.

20.

Furman

, Murray IV

, Schall

, Liu

, Ghiwot

, Axelsen

(2016) Amyloid plaque-associated oxidative degradation of uniformly radiolabeled arachidonic acid. ACS Chem Neurosci 7, 367–377.

21.

Hayashi

, Shishido

, Nakayama

, Nunomura

, Smith

, Perry

, Nakamura

(2007) Lipid peroxidation and 4-hydroxy-2-nonenal formation by copper ion bound to amyloid-beta peptide. Free Rad Biol Med 43, 1552–1559.

22.

Eskici

, Axelsen

(2012) Copper and oxidative stress in the pathogenesis of Alzheimer’s disease. Biochem 51, 6289–6311.

23.

Jarrett

, Berger

, Lansbury

(1993) The carboxy terminus of the beta amyloid protein is critical for the seeding of amyloid formation: Implications for the pathogenesis of Alzheimer’s disease. Biochem 32, 4693–4697.

24.

Esterbauer

, Schaur

, Zollner

(1991) Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Rad Biol Med 11, 81–128.

25.

Bolgar

, Gaskell

(1996) Determination of the sites of 4-hydroxy-2-nonenal adduction to protein by electrospray tandem mass spectrometry. Anal Chem 68, 2325–2330.

26.

Lipton

, Bodwell

(1976) Specific oxidation of methionine to methionine sulfoxide by dimethyl sulfoxide. J Agric Food Chem 24, 26–31.

27.

Schechter

, Berger

(1967) On the size of the active site in proteases. I. Papain. Biochem Biophys Res Commun 27, 157–162.

28.

Grasso

, Axelsen

(2017) Effects of covalentmodification by 4-hydroxy-2-nonenal on the noncovalent oligomerization of ubiquitin. J Mass Spec 52, 36–42.

29.

Sayre

, Lin

, Yuan

, Zhu

, Tang

(2006) Protein adducts generated from products of lipid oxidation: Focus on HNE and ONE. Drug Metab Rev 38, 651–675.

30.

Koppaka

, Axelsen

(2000) Accelerated accumulation of amyloid beta proteins on oxidatively damaged lipid membranes. Biochemistry 39, 10011–10016.

31.

Komatsu

, Feingold-Link

, Sharp

, Rastogi

, Axelsen

(2010) Intrinsic linear heterogeneity of amyloid beta protein fibrils revealed by higher resolution mass-per-length determinations. J Biol Chem 285, 41843–41851.

32.

Grasso

, Komatsu

, Axelsen

(2017) Covalent modifications of the amyloid beta peptide by hydroxynonenal: Effects on metal ion binding by monomers and insights into the fibril topology. J Inorg Biochem 174, 130–136.

33.

Oakley

, Cole

, Logan

, Maus

, Shao

, Craft

, Guillozet-Bongaarts

, Ohno

, Disterhoft

, Van Eldik

, Berry

, Vassar

(2006) Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: Potential factors in amyloid plaque formation. J Neurosci 26, 10129–10140.

Post-Translational Chemical Modification of Amyloid-β Peptides by 4-Hydroxy-2-Nonenal

Abstract

Background:

Objective:

Methods:

Results:

Conclusion:

Keywords

INTRODUCTION

METHODS

Chemicals and reagents

Aβ40 preparation, aggregation, and disaggregation

The effects of HNE on Aβ40 aggregation and disaggregation

Quantitation of Aβ40M and modified forms

Fibril morphology

RESULTS

Aβ40 aggregation and disaggregation

The effects of HNE on Aβ40M aggregation

The effects of HNE on Aβ40F disaggregation

DISCUSSION

DATA AVAILABILITY

Footnotes

ACKNOWLEDGMENTS

FUNDING

CONFLICT OF INTEREST

References

Aβ₄₀ preparation, aggregation, and disaggregation

The effects of HNE on Aβ₄₀ aggregation and disaggregation

Quantitation of Aβ_40M and modified forms

Aβ₄₀ aggregation and disaggregation

The effects of HNE on Aβ_40M aggregation

The effects of HNE on Aβ_40F disaggregation