Zinc-Mediated Binding of Nucleic Acids to Amyloid-β Aggregates: Role of Histidine Residues

Abstract

Amyloid-β peptide (Aβ) plays a central role in Alzheimer’s disease (AD) pathogenesis. Besides extracellular Aβ, intraneuronal Aβ (iAβ) has been suggested to contribute to AD onset and development. Based on reported in vitro Aβ-DNA interactions and nuclear localization of iAβ, the interference of iAβ with the normal DNA expression has recently been proposed as a plausible pathway by which Aβ can exert neurotoxicity. Employing the sedimentation assay, thioflavin T fluorescence, and dynamic light scattering we have studied effects of zinc ions on binding of RNA and single- and double-stranded DNA molecules to Aβ₄₂ aggregates. It has been found that zinc ions significantly enhance the binding of RNA and DNA molecules to pre-formed β-sheet rich Aβ₄₂ aggregates. Another type of Aβ₄₂ aggregates, the zinc-induced amorphous aggregates, was demonstrated to also bind all types of nucleic acids tested. To evaluate the role of the Aβ metal-binding domain’s histidine residues in Aβ-nucleic acid interactions mediated by zinc, Aβ₁₆ mutants with substitutions H6R and H6A-H13A and rat Aβ₁₆ lacking histidine residue 13 were used. The zinc-induced interaction of Aβ₁₆ with DNA was shown to critically depend on histidine residues 6 and 13. However, the inclusion of H6R mutation in Aβ₄₂ peptide did not affect DNA binding to Aβ₄₂ aggregates. Since oxidative and/or nitrosative stresses implicated in AD pathogenesis are known to release zinc ions from metallothioneins in cytoplasm and cell nuclei, our findings suggest that intracellular zinc can be an important player in iAβ-nucleic acid interactions.

Keywords

Alzheimer’s disease amyloid-β peptide DNA histidine residues RNA zinc

INTRODUCTION

The aggregation of amyloid-β (Aβ), a 39/42-amino-acid-long peptide, plays a pivotal role in Alzheimer’s disease (AD) pathogenesis though molecular events by which Aβ aggregation triggers the AD onset are still obscure [1]. According to the amyloid cascade hypothesis, extracellular Aβdeposition is crucial in the disease process. However, a growing body of evidence suggests that, besides the extracellular Aβ, the accumulation of intraneuronal Aβ (iAβ) can contribute to pathological alterations leading to AD [2]. The first study showing that the occurrence of iAβ precedes the extracellular Aβ deposition in the form of plaques was reported more than two decades ago [3]. It has been later suggested [4] that the excessive accumulation of Aβ₄₂, 42-amino-acid-long Aβ isoform, within neurons might result in cell lysis, thereby enhancing extracellular amyloid deposition and neuron loss. Today there are numerous reports based on different transgenic mouse and rat models which demonstrate the direct link between iAβ accumulation and neuron loss (e.g., [5, 6] and references therein).

One plausible hypothesis suggests that iAβ can contribute to neuron death through interference with the normal DNA expression by interaction with genomic DNA [7]. Indeed, the localization of iAβ was found in cell nuclei [8 –10] and the binding of Aβ to DNA was shown in vitro [10 –19]. It has been reported that aggregates of Aβ but not its monomeric form are able to bind DNA [11, 14]. Aβ-DNA interactions are known to produce conformational changes in DNA [12 , 15] and modulate DNA stability [19]. The potential activity of Aβ as a transcription factor per se has also been suggested [20]. It is noteworthy that amyloid-like protein aggregates when localized in cytoplasm interfere with the nucleus-to-cytoplasm transport of mRNA [21]. In perikaryal cytoplasm, aggregated iAβ₄₂ appears as densely packed granules [22] and can thus potentially interfere with DNA expression by affecting mRNA transport. The in vitro interaction of RNA with aggregated Aβ has been demonstrated [11].

Zinc plays an important role in AD pathogenesis [23]. Zinc ions form stable complexes with Aβ, mediated mostly by histidine residues located in the N-terminal region comprised of amino acid residues 1–16 and known as the metal-binding domain (MBD) [24]. Upon binding to MBD, zinc can induce rapid Aβ aggregation [24]. In neurons, the total zinc concentration is around 250 μM, but zinc ions are tightly bound mostly by metallothioneins (MTs) [25]. However, under conditions of oxidative stress a substantial release of intraneuronal zinc (iZn^2 +) from MTs has been shown (e.g., [26] and references therein). The oxidative and nitrosative stresses liberate iZn^2 + from metalloproteins, leading to accumulation of chelatable zinc in hippocampal neuronal perikarya in vivo [27]. Moreover, the nitrosative stress has been demonstrated to induce translocation of MTs to a cell nucleus and intranuclear iZn^2 + release [28]. Thus, under conditions of oxidative and/or nitrosative stresses (which are implicated in AD pathogenesis [29, 30]) iZn^2 + can be available for complexing with iAβ. Whether Zn^2 + can influence Aβ interactions with DNA or RNA is open to question.

Earlier we have shown that synthetic peptide Aβ₁₆ representing MBD of Aβ binds to duplex DNA in the presence of Zn^2 + [31], and thus MBD can potentially serve as a zinc-dependent DNA-binding site of Aβ. In the present report, we are extending the study of zinc-induced Aβ-DNA interactions to the full-length Aβ peptide, Aβ₄₂. Since iAβ aggregate structure is still mysterious, we have tested two types of Aβ₄₂ aggregates: fibrillar (β-sheet rich) aggregates (further referred to as fAβ₄₂ aggregates) and zinc-induced amorphous (lacking an ordered structure) aggregates. It has been found that Zn^2 + significantly enhances binding of nucleic acids (NAs) to fAβ₄₂ aggregates. The zinc-induced Aβ₄₂ aggregates were shown to also bind DNA and RNA molecules. The role of Aβ histidine residues in the zinc-induced Aβ-NA interactions was evaluated using H6R-Aβ₄₂, H6R-Aβ₁₆, and H6A-H13A-Aβ₁₆ mutants with substitutions H6R and H6A-H13A, and rat Aβ₁₆ deficient in H13. It has been found that zinc-induced Aβ₁₆-DNA interactions critically depend on H6 and H13 residues. Yet, H6R mutation introduced into the full-length Aβ peptide had no effect on ability of Aβ₄₂ aggregates to bind DNA molecules.

MATERIALS AND METHODS

2.1 Materials

Synthetic peptides Aβ₄₂, H6R-Aβ₄₂, Aβ₁₆, rAβ₁₆, H6R-Aβ₁₆, and H6A-H13A-Aβ₁₆ (purity > 98%) were purchased as a lyophilized solid from Biopeptide Co., LLC (USA). Aβ₁₆, rAβ₁₆, H6R-Aβ₁₆, and H6A-H13A-Aβ₁₆ peptides were acetylated and amidated on N- and C-termini, respectively. The peptides’ amino acid sequences were confirmed with Fourier transform ion cyclotron resonance mass-spectrometry as described elsewhere [32]. Fluorescently labeled and biotinylated oligonucleotides were synthesized and purified in-house, using a DNA/RNA synthesizer ASM-800 (Biosset, Russia). Peptide and oligonucleotide sequences are shown in Table 1. Total RNA from cultured HepG2 cells was isolated using an RNeasy RNA isolation kit (Qiagen, the Netherlands) following manufacturer’s instructions. The RIN (RNA Integrity Number) of the RNA preparation was equal to 9.1 as assessed with 2100 Bioanalyzer (Agilent, USA). Deoxyribonucleic acid sodium salt from salmon testes (double-stranded DNA, dsDNA) was purchased from Sigma-Aldrich (USA). Other chemicals used were obtained from Sigma-Aldrich and were of an analytical grade or higher. Chlorides of zinc, calcium, and magnesium served as a source of divalent ions. The Milli-Q quality water (mqH₂O) was used to prepare all solutions; mqH₂O and stock buffer solutions were filtered through a 0.22 micron syringe filter (Millipore, USA) before use.

Aβ peptide and nucleic acid solutions

Solutions of Aβ₁₆, rAβ₁₆, H6R-Aβ₁₆, and H6A-H13A-Aβ₁₆ in mqH₂O were prepared as described earlier [33]. The peptide solutions were kept on ice and diluted with appropriate buffers just prior to use to provide peptide solutions in buffer composed of 10 mM HEPES (pH 7.0) and 50 mM NaCl (further referred to as ‘buffer H’). Solutions of Aβ₄₂ peptides were prepared closely following the protocol of Stine et al. [34]. Aβ₄₂ peptides were treated with hexafluoroisopropanol (HFIP), dried, and dissolved in dimethyl sulfoxide (DMSO). The Aβ₄₂ solutions in DMSO were diluted 10-fold with cold mqH₂O and used to prepare either fAβ₄₂ aggregates or amorphous (zinc-induced) Aβ₄₂ aggregates. For the latter, the Aβ₄₂ solutions in mqH₂O were diluted with the appropriate buffer to provide peptide solutions in buffer H mostly in a monomeric state [34] (further referred to as ‘mAβ₄₂ preparations’) and used immediately to produce zinc-induced aggregates. H6R-Aβ₄₂ peptide solutions were prepared in the identical manner.

DNA solutions were prepared by dissolving lyophilized oligonucleotides and dsDNA in buffer H. Solutions of RNA in buffer H were obtained by diluting the HepG2 total RNA preparation in mqH₂O with the appropriate buffer. Concentrations of NAs were measured spectrophotometrically. The NA solutions were adjusted to desired concentrations with buffer H supplemented where necessary with Zn^2 +, Ca^2 +, or Mg^2 + ions.

Preparation of Aβ₄₂ aggregates

The fAβ₄₂ aggregates were prepared by incubating Aβ₄₂ peptides under acidic conditions [34] and adjusting with the appropriate buffer to provide fAβ₄₂ aggregate preparations in buffer H. The preparations were sonicated and kept on ice until further use. To prepare zinc-induced Aβ₄₂ aggregates, aliquots of mAβ₄₂ solution were mixed with buffer H supplemented with ZnCl₂ at various concentrations. The solutions were left at room temperature for 30 min and used for DNA and RNA binding experiments. Fibrillar and zinc-induced H6R-Aβ₄₂ aggregates were prepared in the same ways.

2.3.1 Nucleic acids binding to Aβ₄₂ aggregates

Aliquots of the preparation of zinc-induced Aβ₄₂ aggregates were mixed with aliquots of NA solutions (containing Zn^2 + concentration matching that in the zinc-induced Aβ₄₂ aggregate preparation), incubated for 15 min at room temperature, and subjected to centrifugation (16,000 g, 20°C, 30 min). Aliquots of fAβ₄₂ preparations and NA solutions were pre-incubated with Zn^2 + ions, mixed, and incubated for 15 min at room temperature, followed by centrifugation as described above. In control experiments, Ca^2 + and Mg^2 + ions known not to bind to Aβ were used instead of Zn^2 +. Alternatively, the peptide/NA mixtures were placed in VectaSpin Micro centrifugal filter units (GE Healthcare, USA) with an Anapore membrane (pore size - 0.02 μm) and subjected to ultrafiltration (10,000 g, 20°C, 30 min).

The amount of NA in the supernatant (filtrate) was estimated by measuring the fluorescence of FAM-labeled DNA-oligonucleotides or by using a specific dye, SYBR Green I, which fluoresces upon binding to NAs. Fluorescence was measured using an Infinite M200 PRO microplate reader (TECAN, Switzerland). The emission and excitation wavelengths were set at 494 nm and 520 nm, respectively. The fraction of NAs bound to Aβ₄₂ aggregates was calculated as ((F_Aβ = 0–F_Aβ)/F_Aβ = 0) 100%, where F_Aβ and F_Aβ = 0 are fluorescence intensities for NA solutions incubated or not incubated with Aβ₄₂, respectively, and subjected to centrifugation (filtration). The concentration of rDNA bound to fAβ₄₂ aggregates ([rDNA]_bound) was calculated as [rDNA] (F_Aβ = 0–F_Aβ)/F_Aβ = 0), where [rDNA] is the total rDNA concentration in the rDNA/fAβ₄₂ mixture.

More details on the preparation of Aβ42 aggregates, NA, and peptide solutions as well as on binding experiments can be found in the Supplementary Material.

Turbidimetry and biosensor analysis

Turbidimetry and biosensor analysis were employed to study zinc-induced interactions of NAs with Aβ₁₆ peptides. An aliquot of the peptide solution was pre-incubated with Zn^2 + for 15 min and mixed with an equal volume of the NA solution containing the matching Zn^2 + concentration. After 15-min incubation at room temperature, turbidity of the mixture was measured as optical density at 405 nm (A₄₀₅) using the NanoDrop-1000 spectrophotometer. Ca^2 + and Mg^2 + were used instead of Zn^2 + in control experiments. The biosensor analysis was carried out on a Biacore-3000 (GE, USA) optical biosensor utilizing the surface plasmon resonance (SPR) effect. The immobilization of biotinylated rDNA (B-rDNA, Table 1) onto the surface of biosensor optical chip and real-time monitoring of zinc-induced rDNA/peptide interactions were performed as described earlier [31].

Dynamic light scattering

Dynamic light scattering (DLS) measurements were carried out on a Zetasizer Nano ZS apparatus (Malvern Instruments Ltd., UK) which is able to measure particles sizes in the range of 0.6 nm to 10 μm. The instrument software provides particle size distribution and the average particle diameter, approximating a heterogeneous population of Aβ aggregates by a population of spherical particles with the identical distributions of diffusion coefficient. Consequently, the characteristic size of Aβ aggregates is expressed in terms of average “diameter”.

Thioflavin T fluorescence assay

Thioflavin T (ThT) fluorescence was measured on the Infinite M200 PRO microplate reader with the emission and excitation wavelengths set at 482 nm and 450 nm, respectively. The final Aβ₄₂ and ThT concentrations were correspondingly 4 and 9 μM. All measurements were made in triplicates. The relative ThT fluorescence was calculated as F/F₀, where F and F₀ are ThT fluorescence in the presence and absence of Aβ₄₂, respectively.

Statistical analysis

Based on the results of three independent measurements (Table 2), the statistical differences in binding of NAs to fAβ₄₂ aggregates were assessed by comparing the mean values of bound fractions. The assessment was carried out using the 2-tailed Student’s t-test with the following parameters: the degree of freedom = 4, p = 0.05, and t = 2.78.

RESULTS

The binding of nucleic acids to fibrillar Aβ₄₂ aggregates

To ensure that Aβ₄₂ aggregates formed under acidic conditions do possess the β-sheet rich structure, fAβ₄₂ preparations were tested with the ThT fluorescence assay prior to use. All fAβ₄₂ preparations have exhibited the mean values of ThT relative fluorescence, F/F₀, ranging from 18 to 23 (Supplementary Table 1) while the mAβ₄₂ preparations had the F/F₀ values within the interval of 1.5 to 2.7. The high values of F/F₀ indicate the significant β-sheet content and suggest that fAβ₄₂ aggregates indeed have the fibrillar structure. The characteristic size of the sonicated fAβ₄₂ aggregates measured by DLS was found to lie within the range of about 660 to 800 nm in diameter (Supplementary Table 1). No appreciable changes in the aggregate size were observed upon addition of Zn^2 +, Mg^2 +, or Ca^2 + ions at the concentration of 150 μM (Supplementary Figure 1).

As seen from Fig. 1, Zn^2 + enhances rDNA-fAβ₄₂ interaction. In the absence of Zn^2 +, about 35% of rDNA molecules are bound to fAβ₄₂ aggregates. At a 2-fold molar excess of Zn^2 + over Aβ₄₂, the fraction of rDNA bound to fAβ₄₂ grows to ∼90% (Fig. 1A). Table 2 presents results on the NA binding to fAβ₄₂ aggregates for various types of NAs. All NAs tested are able to interact with fAβ₄₂ aggregates in the absence of Zn^2 + though with different efficiencies. When tested NAs were compared with each other, the statistical analysis has revealed statistically significant differences (p = 0.05) in mean values for the bound fraction except for the rDNA/dsDNA pair where no statistically significant differences (p = 0.05) were found in the “no ions” and “Ca^2 +” cases (Table 2). Ca^2 + and Mg^2 + at the concentration of 150 μM had no impact on NA binding to fAβ₄₂. In contrast, Zn^2 + enhanced the NAs binding: more than 90% of RNA and DNA molecules were in a complex with fAβ₄₂ at 150- μM Zn^2 + concentration (a 3-fold molar excess of Zn^2 + over fAβ₄₂, Table 2). The fibrillar aggregates of H6R-Aβ₄₂ mutant (50 μM peptide) were found to bind (94±2)% of rDNA molecules in the presence of 150 μM Zn^2 +. When rDNA-fAβ₄₂ complexes were removed from rDNA/fAβ₄₂ mixtures by filtration instead of sedimentation, the estimates for the fraction of rDNA bound to fAβ₄₂ aggregates were quite close to those obtained with the sedimentation method(Table 2).

The binding of nucleic acids to zinc-induced Aβ₄₂ aggregates

To study whether NAs can interact with zinc-induced Aβ aggregates, as a first step the Aβ aggregation process has been initiated in mAβ₄₂ solutions by adding various amounts of Zn^2 +. It is known that Zn^2 + induces a rapid formation of amorphous Aβ aggregates which lack the β-sheet organization (e.g., [24]). Using DLS, Aβ₄₂ aggregates have been detected in 50- μM mAβ₄₂ solutions after addition of 1 to 3 molar equivalents of Zn^2 +. In 30 min after Zn^2 + addition, the aggregates had the characteristic size around 2.3 μm, regardless of amount of Zn^2 + added (Fig. 1B). No appreciable changes of ThT fluorescence in the mAβ₄₂ solutions were observed upon addition of Zn^2 + as tested with the ThT assay (data not shown) thus confirming that the formed aggregates are amorphous.

Zinc-induced Aβ₄₂ aggregates are able to bind NAs (Fig. 1C). Though the size of Aβ₄₂ aggregates was independent of Zn^2 + concentration within the concentration range studied (Zn^2 +/mAβ₄₂ molar ratio of 1 to 3), only approximately 60% of RNA molecules and 40% of DNA molecules were bound at the equimolar ratio of Zn^2 +/mAβ₄₂. The fraction of NAs bound to Aβ₄₂ aggregates rises to ≈90–100% at Zn^2 +/mAβ₄₂ = 3 (Fig. 1C). As seen from Fig. 1B, a fraction of mAβ₄₂ peptides (as measured with the BCA assay) is not engaged in the formation of large zinc-induced aggregates at the equimolar ratio of Zn^2 +/mAβ₄₂ – about 20% of peptides were found in the supernatant after centrifugation (compared to almost zero level at Zn^2 +/peptide molar ratios of 2 and 3). Thus, the smaller number of zinc-induced Aβ₄₂ aggregates formed at Zn^2 +/mAβ₄₂ = 1 compared to that at Zn^2 +/mAβ₄₂ = 3 can account for the differences in NA binding observed at these Zn^2 +/mAβ₄₂ molar ratios (Fig. 1C).

The fraction of H6R-Aβ₄₂ peptides engaged into aggregate formation was found to be considerably lesser than that for Aβ₄₂ peptides under identical conditions while sizes of zinc-induced Aβ₄₂ and H6R-Aβ₄₂ aggregates were practically undistinguishable at Zn^2 +/mAβ₄₂ molar ratios of 1 and above (Fig. 1B). One may conclude that the smaller number of H6R-Aβ₄₂ aggregates are formed at the same Zn^2 + concentrations (compared to Aβ₄₂) that can account for appreciably less amount of rDNA bound to the aggregates at Zn^2 +/mAβ₄₂ molar ratios studied in the case of H6R-Aβ₄₂ peptides (Fig. 1C).

It should be noted that a limited fraction of NAs (≈5%) sediment in mAβ₄₂ solutions in the absence of Zn^2 + (i.e., in the absence of zinc-induced aggregates, Fig. 1C). This can be attributed to the binding of NAs to pre-existing Aβ₄₂ aggregates (similar to the binding of NAs to fAβ₄₂ aggregates) not disrupted by the treatment with HFIP and DMSO and present in mAβ₄₂ solutions in the small number. Indeed, a particulate material has been detected in mAβ₄₂ solutions (illustrated by Supplementary Figure 2A). Most of this material can be removed by centrifugation (Supplementary Figure 2B) that decreases the peptide concentration in mAβ₄₂ solution by about 15% (estimated with the BCA assay). Moreover, no decrease of rDNA concentration in the supernatant of mAβ₄₂/rDNA mixture was observed if the mAβ₄₂ solution was subjected to centrifugation (16,000 g, 30 min, 4°C) prior to mixing with rDNA.

Affinity of rDNA to fAβ₄₂ aggregates in the presence and absence of Zn^2 +

Figure 2 shows the dependence of [rDNA]_bound on the molar concentration of free (unbound) rDNA ([rDNA]_free) in the absence and presence of 150 μM Zn^2 +. In the former case, the best fitting to experimental points has been provided by the one-site binding model while in the latter case – by the two-site binding model. The following binding equations were used [35]:

[rDNA]_bound=(B_max·[rDNA]_free/([rDNA]_free+ K_D)), (1)

[rDNA]_bound=(B_max1·[rDNA]_free/([rDNA]_free+ K_D1))+(B_max2·[rDNA]_free/([rDNA]_free+K_D2)), (2)

where B_max and K_D are the maximal binding capacity and the equilibrium dissociation constant, respectively, for the one-site binding model, and B_max1, B_max2, K_D1, and K_D2 – the maximal binding capacities and equilibrium dissociation constants for the first and second binding sites in the two-site binding model. The solid lines (curves 1 and 2, Fig. 2) are the best-fits of Equations 1 and 2 to the experimental points, made with the SigmaPlot software package (Systat Software, USA). The values of maximal binding capacities and equilibrium dissociation constants are presented in Table 3. As seen, the binding of rDNA to fAβ₄₂ aggregates in the presence of Zn^2 + is characterized by two dissociation constants which differ substantially: 12 nM (K_D1) and 3 μM (K_D2).

Role of histidine residues in zinc-induced interaction of nucleic acids with Aβ₁₆ peptides

Optical density at 405 nm (A₄₀₅) of the mixtures of Aβ₁₆ with NAs exhibited a rise upon increase of NA concentration in the presence of 1 mM Zn^2 + (Fig. 3A) but not Ca^2 + or Mg^2 + (A₄₀₅ was less 0.003 throughout the NA concentration range tested). Zinc per se did not bring about any turbidity to the NA or Aβ₁₆ solutions under experimental conditions used. The turbidity of Zn^2 +-Aβ₁₆-NA mixtures can solely be accounted for by a formation of large Aβ₁₆/NA aggregates that unequivocally points at the occurrence of zinc-induced Aβ₁₆ interactions both with RNA and DNA molecules.

Three histidine residues (H6, H13, and H14) of Aβ₁₆ are known to participate in the coordination of a zinc ion in the Zn^2 +-Aβ₁₆ complex [32, 36]. To test whether peptide’s ability to interact with NA depends on coordination of Zn^2 + by peptide’s histidine residues, we carried out a turbidimetric analysis of rDNA interactions with Aβ₁₆ mutants H6R-Aβ₁₆ and H6A-H13A-Aβ₁₆ (Table 1). Fig. 3B shows A₄₀₅ values for the mixtures of rDNA with Aβ₁₆, H6R-Aβ₁₆, and H6A-H13A-Aβ₁₆ peptides at various Zn^2 + concentrations. As seen, the substitutions of either H6 or both H6 and H13 with other amino acid residues completely abolished the formation of aggregates. Moreover, peptide rAβ₁₆ representing MBD of rat Aβ and lacking the histidine residue at position 13 of the polypeptide chain (Table 1) demonstrated the substantially reduced ability to form zinc-induced aggregates with rDNA (Fig. 3B). One may therefore conclude that histidine residues H6 and H13 play a key role in providing a particular Zn^2 +-dependent interface enabling Aβ₁₆-NA interactions.

This conclusion is supported by the results of biosensor analysis of the binding of various Aβ₁₆ peptides to rDNA immobilized on the surface of biosensor chip. The biosensor signal (denominated in RU – response units) is equal to the differences in values of refractive indices from flow cells with and without (a reference cell) immobilized DNA and directly proportional to the amount of peptides bound to DNA [31]. As seen from Fig. 4, Aβ₁₆ binds to rDNA only in the presence of Zn^2 +. The peptide’s ability to bind rDNA in the presence of Zn^2 + was substantially decreased and completely abolished by H6R and H6A-H13A mutations, respectively. No binding was also observed for rAβ₁₆ lacking H13 (Fig. 4).

DISCUSSION

To date, several pathogenic mechanisms involving iAβ have been proposed depending on the particular intracellular localization of Aβ (e.g., [37]). One of these hypothetical mechanisms assumes unspecific unfavorable interactions of iAβ with genomic DNA [7] and is based on the reported iAβ localization in cell nuclei [8–10 , 12]. Our results on fAβ₄₂-DNA interactions in the absence of Zn^2 + are consistent with this hypothesis. The β-sheet rich Aβ₄₂ aggregates were found to bind various types of DNA molecules. Additionally, the binding of RNA to fAβ₄₂ aggregates points at a possibility that iAβ can also interfere with the gene expression by affecting the mRNA transport. The observed variation of binding efficiencies among NAs tested (Table 2) can be attributed to substantial differences in length and conformation of NA molecules used in thestudy.

Though a possible role of iAβ in AD pathogenesis has actively being discussed [2 , 37], it has never being considered in conjunction with an intracellular zinc homeostasis and its regulation by MTs. It should be noted that mechanisms of Zn^2 + transfer from MTs to Aβ peptides as well as a character of Aβ aggregates induced by this transfer have drawn some attention but in connection to the extracellular MT pool [38, 39]. Interestingly, oxidation conditions were shown to induce Zn^2 + release from MT via cysteine oxidation and consequently promote Aβ₄₀ aggregation in vitro [38]. Oxidative and nitrosative stresses are known to liberate iZn^2 + from MTs intracellular [26, 27]. Additionally, the translocation of MTs to a cell nucleus and intranuclear release of Zn^2 + have been shown under conditions of nitrosative stress [28]. These alterations in intracellular Zn^2 + homeostasis under conditions of oxidative and/or nitrosative stresses have motivated us to examine how Zn^2 + can modulate the Aβ₄₂ interaction with NAs, especially in view of the fact that oxidative stress can drive Aβ₄₂ to a cell nucleus [10]. Our results demonstrate that the addition of Zn^2 + to fibrillar Aβ₄₂ aggregates leads to a significant enhancement of fAβ₄₂ interactions with NAs (Fig. 1A and Table 2). Moreover, zinc-induced amorphous Aβ₄₂ aggregates were also shown to bind RNA and DNA molecules (Fig. 1C). These findings suggest that stress-related alterations of intracellular Zn^2 + homeostasis can potentially provoke unfavorable interactions of iAβ₄₂ with NAs either by producing zinc-induced Aβ₄₂ aggregates able to interact with NAs or by enhancing the DNA/RNA binding to pre-formed Aβ₄₂aggregates.

It is clear that zinc-dependent Aβ₄₂-NA interactions are reliant on a formation of Zn^2 +/Aβ₄₂ complexes since divalent metals unable to form complexes with Aβ (such as Ca^2 + and Mg^2 +) have induced no Aβ₄₂-NA interactions. Histidine residues located in the Aβ MBD are commonly accepted to play a determinative role in Zn^2 +/Aβ complexing via both intra- and intermolecular coordination of zinc [24 , 40–42]. In the latter case, Zn^2 + is suggested to trigger Aβ oligomerization by bridging peptides via zinc coordination involving E11/H14 and H6/H13 pairs of interacting peptide molecules [42]. However, other coordination modes may also be realized depending on particular environmental conditions inasmuch as seven residues of the Aβ MBD (D1, E3, H6, D7, E11, H13, H14) can potentially chelate Zn^2 + [43].

The substitution of histidine residues H6 and H13 has been found to have a remarkable impact on the zinc-mediated binding of Aβ₁₆ peptides to rDNA (Figs. 3B and 4). The effect of these substitutions is consistent with the earlier suggestion that Aβ₁₆ oligomerization can be a decisive factor in the zinc-induced Aβ₁₆-DNA interaction [31]. Indeed, both H6 and H13 residues are crucial for Aβ₁₆ oligomerization induced by Zn^2 + [42]. Moreover, peptide rAβ₁₆ lacking H13 residue can form zinc-induced dimers but its oligomerization is hindered by the particular orientation of peptides in the dimer [44]. One may suggest that Zn^2 + binding to Aβ₁₆ peptides (representing the Aβ MBD) constitutes an anion-binding site, while Aβ₁₆ oligomerization can results in the occurrence of entity bearing several such anion-binding sites and thereby able to produce a stable complex with NAs via multiple interactions with phosphate groups of a DNA or RNA polymer.

In contrast to Aβ₁₆ peptides, the substitution of H6 with R had no effect on the binding of rDNA to fibrillar aggregates of the full-length peptide (H6R-fAβ₄₂). DNA molecules bind to fAβ₄₂ and H6R-fAβ₄₂ aggregates with similar efficiency (92% and 94%, respectively) in the presence of Zn^2 +. It seems most likely that a Zn^2 + coordination mode not involving the H6 residue prevails when Aβ₄₂ peptides are tightly and orderly parked in a fibril. This suggestion agrees with the recent report showing by means of solid-state NMR spectroscopy that Zn^2 + binding to pre-formed Aβ fibrils brings more order to the side chains of particular residues but not to H6 [45]. However, the H6R substitution obviously affects the ability of the full-length peptide to undergo zinc-induced aggregation: higher Zn^2 + concentrations are required to engage the same number of H6R-Aβ₄₂ peptides into aggregation, compared to that for Aβ₄₂ peptides (Fig. 1B). This observation is consistent with the Aβ oligomerization mechanism involving the residue H6 and segment ¹¹EVHH¹⁴ which have been suggested to form two zinc-mediated interaction interfaces in Aβ [42]. When H6 residue is absent, other Zn^2 + coordination modes can apparently be realized [43] prompting peptide oligomerization and formation of amorphous Aβ₄₂ aggregates but with a lesser efficiency. Nonetheless, once aggregates form, they can bind DNA polymers (Fig. 1C). One may assume that binding occurs via multiple interactions of phosphate groups with anion-binding sites constituted on the aggregate’s surface by zinc ions complexed with Aβ MBDs, apparently irrespectively of a particular Zn^2 + coordinationmode.

It should be noted that ¹³HHQK¹⁶ region of the Aβ MBD was shown to constitute a site responsible for the interaction of Aβ₄₂ aggregates with another biological polyanion, heparin [46, 47]. This site can also be responsible for the observed NA interactions with fAβ₄₂ aggregates in the absence of Zn^2 + (with K_D of 1.5 μM in the case of rDNA, Table 3). In the presence of Zn^2 +, two DNA-binding sites are situated on the surface of fAβ₄₂ aggregates (Table 3). While one of them can be attributed to Zn^2 +/MBD complexes, another can also arise from Zn^2 +-fAβ₄₂ interactions. Zn^2 + was shown to break up the D23-K28 salt bridge in the Aβ loop region by driving these residues to different structural conformations but not disrupting the β-sheet structure of Aβ₄₂ fibrils [45]. As a result, lysine residues can be exposed to the surface of fAβ₄₂ aggregates, thus forming a stretch of positive charges constituting the second DNA-bindingsite.

In summary, we report that zinc ions significantly enhance binding of various NAs including RNA and single- and double stranded DNA molecules to fibrillar (β-sheet rich) Aβ₄₂ aggregates. Amorphous (zinc-induced) Aβ₄₂ aggregates were shown to also bind DNA and RNA molecules. The histidine residues H6 and H13 play a vital role in the zinc-induced interaction of Aβ₁₆ peptides (representing the Aβ metal-binding domain) with DNA. However, the H6R mutation did not affect DNA binding to Aβ₄₂ aggregates that may be attributed to the polymorphism of Zn^2 + coordination modes in the full-length Aβ peptides. Our findings suggest that the stress-related disturbance of intracellular Zn^2 + homeostasis can potentially provoke unfavorable interactions of iAβ with NAs either by enhancing the DNA/RNA binding to pre-formed β-sheet structured iAβ aggregates or by producing zinc-induced iAβ aggregates able to interact with NAs. Since conditions of oxidative and/or nitrosative stresses are implicated in AD pathogenesis, the interplay between intraneuronal zinc, iAβ, and NAs can affect gene expression thus contributing to molecular mechanisms by which iAβ exerts its neurotoxicity.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the Russian Science Foundation grant No 14-24-00100. Authors are grateful to the Orekhovich Institute of Biomedical Chemistry (Russia) for the opportunity to use Zetasizer Nano ZS and Biacore-3000 instruments.

Authors’ disclosures available online ().

Appendix

References

Hardy

, Bogdanovic

, Winblad

, Portelius

, Andreasen

, Cedazo-Minguez

, Zetterberg

(2014) Pathways to Alzheimer’s disease. J Intern Med 275, 296–303.

Gouras

, Tampellini

, Takahashi

, Capetillo-Zarate

(2010) Intraneuronal beta-amyloid accumulation and synapse pathology in Alzheimer’s disease. Acta Neuropathol 119, 523–541.

Grundke-Iqbal

, Iqbal

, George

, Tung

, Kim

, Wisniewski

(1989) Amyloid protein and neurofibrillary tangles coexist in the same neuron in Alzheimer disease. Proc Natl Acad Sci U S A 86, 2853–2857.

D’Andrea

, Nagele

, Wang

, Peterson

, Lee

(2001) Evidence that neurones accumulating amyloid can undergo lysis to form amyloid plaques in Alzheimer’s disease. Histopathology 38, 120–134.

Cuello

, Canneva

(2008) Impact of intracellular beta-amyloid in transgenic animals and cell models. Neurodegener Dis 5, 146–148.

Wirths

, Bayer

(2012) Intraneuronal Abeta accumulation and neurodegeneration: Lessons from transgenic models. Life Sci 91, 1148–1152.

Jimenez

(2010) Protein-DNA interaction at the origin of neurological diseases: A hypothesis. J Alzheimers Dis 22, 375–391.

Johnstone

, Babbey

, Stephenson

, Paul

, Santerre

, Clemens

, Williams

, Little

(1996) Nuclear and cytoplasmic localization of the beta-amyloid peptide (1-43) in transfected 293 cells. Biochem Biophys Res Commun 220, 710–718.

Buckig

, Tikkanen

, Herzog

, Schmitz

(2002) Cytosolic and nuclear aggregation of the amyloid beta-peptide following its expression in the endoplasmic reticulum. Histochem Cell Biol 118, 353–360.

10.

Ohyagi

, Asahara

, Chui

, Tsuruta

, Sakae

, Miyoshi

, Yamada

, Kikuchi

, Taniwaki

, Murai

, Ikezoe

, Furuya

, Kawarabayashi

, Shoji

, Checler

, Iwaki

, Makifuchi

, Takeda

, Kira

, Tabira

(2005) Intracellular Abeta42 activates p53 promoter: A pathway to neurodegeneration in Alzheimer’s disease. FASEB J 19, 255–257.

11.

Ahn

, Song

, Jung

, Chay

, Chung

, Yang

, Shin

(2000) Detection of beta-amyloid peptide aggregation using DNA electrophoresis. Anal Biochem 284, 401–405.

12.

Hegde

, Anitha

, Latha

, Mustak

, Stein

, Ravid

, Rao

(2004) First evidence for helical transitions in supercoiled DNA by amyloid beta peptide (1-42) and aluminum: A new insight in understanding Alzheimer’s disease. J Mol Neurosci 22, 19–31.

13.

, Ren

, Qu

(2007) Time-dependent DNA condensation induced by amyloid beta-peptide. Biophys J 92, 185–191.

14.

Barrantes

, Rejas

, Benitez

, Jimenez

(2007) Interaction between Alzheimer’s Abeta1-42 peptide and DNA detected by surface plasmon resonance. J Alzheimers Dis 12, 345–355.

15.

Geng

, Zhao

, Ren

, Qu

(2010) Alzheimer’s disease amyloid beta converting left-handed Z-DNA back to right-handed B-form. Chem Commun (Camb) 46, 7187–7189.

16.

Barrantes

, Camero

, Garcia-Lucas

, Navarro

, Benitez

, Jimenez

(2012) Alzheimer’s disease amyloid peptides interact with DNA, as proved by surface plasmon resonance. Curr Alzheimer Res 9, 924–934.

17.

Maloney

, Lahiri

(2011) The Alzheimer’s amyloid beta-peptide (Abeta) binds a specific DNA Abeta-interacting domain (AbetaID) in the APP, BACE1, and APOE promoters in a sequence-specific manner: Characterizing a new regulatory motif. Gene 488, 1–12.

18.

Camero

, Ayuso

, Barrantes

, Benitez

, Jimenez

(2013) Specific binding of DNA to aggregated forms of Alzheimer’s disease amyloid peptides. Int J Biol Macromol 55, 201–206.

19.

Gupta

, Monica

, Berrocal

, Rao

(2013) Studies on the mechanism of the DNA nicking property of amyloid-beta40: Implications in Alzheimer’s disease. J Alzheimers Dis 33, 1059–1071.

20.

Lahiri

, Maloney

(2010) Beyond the signaling effect role of amyloid-ss42 on the processing of APP, and its clinical implications. Exp Neurol 225, 51–54.

21.

Woerner

, Frottin

, Hornburg

, Feng

, Meissner

, Patra

, Tatzelt

, Mann

, Winklhofer

, Hartl

, Hipp

(2016) Cytoplasmic protein aggregates interfere with nucleocytoplasmic transport of protein and RNA. Science 351, 173–176.

22.

Nagele

, D’Andrea

, Anderson

, Wang

(2002) Intracellular accumulation of beta-amyloid(1-42) in neurons is facilitated by the alpha 7 nicotinic acetylcholine receptor in Alzheimer’s disease. Neuroscience 110, 199–211.

23.

Ayton

, Lei

, Bush

(2013) Metallostasis in Alzheimer’s disease. Free Radic Biol Med 62, 76–89.

24.

Faller

, Hureau

, Berthoumieu

(2013) Role of metal ions in the self-asbly of the Alzheimer’s amyloid-beta peptide. Inorg Chem 52, 12193–12206.

25.

Colvin

, Bush

, Volitakis

, Fontaine

, Thomas

, Kikuchi

, Holmes

(2008) Insights into Zn2+ homeostasis in neurons from experimental and modeling studies. Am J Physiol Cell Physiol 294, C726–C742.

26.

Sensi

, Rapposelli

, Frazzini

, Mascetra

(2008) Altered oxidant-mediated intraneuronal zinc mobilization in a triple transgenic mouse model of Alzheimer’s disease. Exp Gerontol 43, 488–492.

27.

Cuajungco

, Lees

(1998) Nitric oxide generators produce accumulation of chelatable zinc in hippocampal neuronal perikarya. Brain Res 799, 118–129.

28.

Spahl

, Berendji-Grun

, Suschek

, Kolb-Bachofen

, Kroncke

(2003) Regulation of zinc homeostasis by inducible NO synthase-derived NO: Nuclear metallothionein translocation and intranuclear Zn2+ release. Proc Natl Acad Sci U S A 100, 13952–13957.

29.

Kim

, Kim

, Rhie

, Yoon

(2015) The role of oxidative stress in neurodegenerative diseases. Exp Neurobiol 24, 325–340.

30.

Togo

, Katsuse

, Iseki

(2004) Nitric oxide pathways in Alzheimer’s disease and other neurodegenerative dementias. Neurol Res 26, 563–566.

31.

Khmeleva

, Mezentsev

, Kozin

, Tsvetkov

, Ivanov

, Bodoev

, Makarov

, Radko

(2013) Zinc-induced interaction of the metal-binding domain of amyloid-beta peptide with DNA. J Alzheimers Dis 36, 633–636.

32.

Tsvetkov

, Kulikova

, Golovin

, Tkachev

, Archakov

, Kozin

, Makarov

(2010) Minimal Zn(2+) binding site of amyloid-beta. Biophys J 99, L84–L86.

33.

Suprun

, Zaryanov

, Radko

, Kulikova

, Kozin

, Makarov

, Archakov

, Shumyantseva

(2015) Tyrosine based electmical analysis of amyloid-beta fragment (1-16) binding to metal(ii) ions. Electrochim Acta 179, 93–99.

34.

Stine

, Jungbauer

, Yu

, LaDu

(2011) Preparing synthetic Abeta in different aggregation states. Methods Mol Biol 670, 13–32.

35.

Wang

, Jiang

(1996) A novel two-site binding equation presented in terms of the total ligand concentration. FEBS Lett 392, 245–249.

36.

Zirah

, Kozin

, Mazur

, Blond

, Cheminant

, Segalas-Milazzo

, Debey

, Rebuffat

(2006) Structural changes of region 1-16 of the Alzheimer disease amyloid beta-peptide upon zinc binding and in vitro aging. J Biol Chem 281, 2151–2161.

37.

Ohyagi

(2008) Intracellular amyloid beta-protein as a therapeutic target for treating Alzheimer’s disease. Curr Alzheimer Res 5, 555–561.

38.

Durand

, Meloni

, Talmard

, Vasak

, Faller

(2010) Zinc release of Zn(7)-metallothionein-3 induces fibrillar type amyloid-beta aggregates. Metallomics 2, 741–744.

39.

Pedersen

, Hureau

, Hemmingsen

, Heegaard

, Ostergaard

, Vasak

, Faller

(2012) Rapid exchange of metal between Zn(7)-metallothionein-3 and amyloid-beta peptide promotes amyloid-related structural changes. Biochemistry 51, 1697–1706.

40.

Kozin

, Mezentsev

, Kulikova

, Indeykina

, Golovin

, Ivanov

, Tsvetkov

, Makarov

(2011) Zinc-induced dimerization of the amyloid-beta metal-binding domain 1-16 is mediated by residues 11-14. Mol Biosyst 7, 1053–1055.

41.

Miller

, Ma

, Nussinov

(2010) Zinc ions promote Alzheimer Abeta aggregation via population shift of polymorphic states. Proc Natl Acad Sci U S A 107, 9490–9495.

42.

Istrate

, Kozin

, Zhokhov

, Mantsyzov

, Kechko

, Pastore

, Makarov

, Polshakov

(2016) Interplay of histidine residues of the Alzheimer’s disease Abeta peptide governs its Zn-induced oligomerization. Sci Rep 6, 21734.

43.

Miller

, Ma

, Nussinov

(2012) Metal binding sites in amyloid oligomers: Complexes and mechanisms. Coord Chem Rev 256, 2245–2252.

44.

Istrate

, Tsvetkov

, Mantsyzov

, Kulikova

, Kozin

, Makarov

, Polshakov

(2012) NMR solution structure of rat abeta(1-16): Toward understanding the mechanism of rats’ resistance to Alzheimer’s disease. Biophys J 102, 136–143.

45.

Mithu

, Sarkar

, Bhowmik

, Chandrakesan

, Maiti

, Madhu

(2011) Zn(++) binding disrupts the Asp(23)-Lys(28) salt bridge without altering the hairpin-shaped cross-beta structure of Abeta(42) amyloid aggregates. Biophys J 101, 2825–2832.

46.

Watson

, Lander

, Selkoe

(1997) Heparin-binding properties of the amyloidogenic peptides Abeta and amylin. Dependence on aggregation state and inhibition by Congo red. J Biol Chem 272, 31617–31624.

47.

McLaurin

, Fraser

(2000) Effect of amino-acid substitutions on Alzheimer’s amyloid-beta peptide-glycosaminoglycan interactions. Eur J Biochem 267, 6353–6361.

Zinc-Mediated Binding of Nucleic Acids to Amyloid-β Aggregates: Role of Histidine Residues

Abstract

Keywords

INTRODUCTION

MATERIALS AND METHODS

2.1 Materials

Aβ peptide and nucleic acid solutions

Preparation of Aβ42 aggregates

2.3.1 Nucleic acids binding to Aβ42 aggregates

Turbidimetry and biosensor analysis

Dynamic light scattering

Thioflavin T fluorescence assay

Statistical analysis

RESULTS

The binding of nucleic acids to fibrillar Aβ42 aggregates

The binding of nucleic acids to zinc-induced Aβ42 aggregates

Affinity of rDNA to fAβ42 aggregates in the presence and absence of Zn2 +

Role of histidine residues in zinc-induced interaction of nucleic acids with Aβ16 peptides

DISCUSSION

Footnotes

ACKNOWLEDGMENTS

Appendix

References

Preparation of Aβ₄₂ aggregates

2.3.1 Nucleic acids binding to Aβ₄₂ aggregates

The binding of nucleic acids to fibrillar Aβ₄₂ aggregates

The binding of nucleic acids to zinc-induced Aβ₄₂ aggregates

Affinity of rDNA to fAβ₄₂ aggregates in the presence and absence of Zn^2 +

Role of histidine residues in zinc-induced interaction of nucleic acids with Aβ₁₆ peptides