Regulation of presynaptic Ca2+, synaptic plasticity and contextual fear conditioning by a N-terminal β-amyloid fragment

Abstract

Soluble β-amyloid has been shown to regulate presynaptic Ca(2+) and synaptic plasticity. In particular, picomolar β-amyloid was found to have an agonist-like action on presynaptic nicotinic receptors and to augment long-term potentiation (LTP) in a manner dependent upon nicotinic receptors. Here, we report that a functional N-terminal domain exists within β-amyloid for its agonist-like activity. This sequence corresponds to a N-terminal fragment generated by the combined action of α- and β-secretases, and resident carboxypeptidase. The N-terminal β-amyloid fragment is present in the brains and CSF of healthy adults as well as in Alzheimer's patients. Unlike full-length β-amyloid, the N-terminal β-amyloid fragment is monomeric and nontoxic. In Ca(2+) imaging studies using a model reconstituted rodent neuroblastoma cell line and isolated mouse nerve terminals, the N-terminal β-amyloid fragment proved to be highly potent and more effective than full-length β-amyloid in its agonist-like action on nicotinic receptors. In addition, the N-terminal β-amyloid fragment augmented theta burst-induced post-tetanic potentiation and LTP in mouse hippocampal slices. The N-terminal fragment also rescued LTP inhibited by elevated levels of full-length β-amyloid. Contextual fear conditioning was also strongly augmented following bilateral injection of N-terminal β-amyloid fragment into the dorsal hippocampi of intact mice. The fragment-induced augmentation of fear conditioning was attenuated by coadministration of nicotinic antagonist. The activity of the N-terminal β-amyloid fragment appears to reside largely in a sequence surrounding a putative metal binding site, YEVHHQ. These findings suggest that the N-terminal β-amyloid fragment may serve as a potent and effective endogenous neuromodulator.

Keywords: Ca regulation; fear memory; neuromodulation; synaptic plasticity; β-amyloid.

Figure 1.

A N-terminal Aβ fragment encompassing residues 1–15 is a potent and highly effective agonist at α7-nAChRs on presynaptic-like axonal varicosities. A, B, Averaged [Ca²⁺]i responses (F/F₀) in varicosities of NG108–15 cells expressing α7-nAChRs (A) or mouse hippocampal synaptosomes (B) to 100 nm Aβ_1–42 (A, n = 49; B, n = 16) were compared with responses to 100 nm Aβ_1–15 (A: n = 178; p < 0.05; B: n = 26; NS), followed sequentially by K⁺-induced depolarization. Time-series traces are means ± SEM at individual time points. The Ca²⁺ responses to 100 nm Aβ_1–42 are similar in magnitude to that observed with nicotine (data not shown; see Khan et al., 2010). C, Representative CD spectra for Aβ_1–42 and Aβ_1–15. D, Averaged peak Ca²⁺ responses in varicosities expressing α7-nAChRs to 100 fm (n = 22 or n = 33), 1 pm (n = 26 or n = 21), 100 pm (n = 19 or n = 32), 1 nm (n = 30 or n = 15), and 100 nm (n = 44 or n = 29) Aβ_1–42 or Aβ_1–15, respectively. All pairs (Aβ_1–42 vs Aβ_1–15) are significantly different (p < 0.05, t test). E, 4–20% gradient Tris-Tricine PAGE of Aβ_1–15 (2 nmol). Positions of molecular weight standards (data not shown) are as marked in kilodaltons. Insets show a comparison of Aβ_1–15 (molecular weight, 1827 kDa) before and after (from whole lane) filtration through an Amicon 3 kDa cutoff filter. Dimers and larger oligomers will be excluded by the 3 kDa filter. F, Fibril–aggregate formation of Aβ_1–42, Aβ_1–15, or Aβ_42–1 control peptide, each at 200 nm, assessed in triplicate by a fluorimetric thioflavin (ThT) assay.

Figure 2.

Structure–function analysis of the N-terminal Aβ domain and fragment. A, Averaged peak Ca²⁺ responses in varicosities of NG108–15 cells expressing α7-nAChRs to 100 nm Aβ_1–42 (n = 44), Aβ_1–28 (n = 10), and Aβ_1–15 (n = 29), and the control peptides Aβ_42–1 (n = 24) and Aβ_15–1 (n = 38). Peak Ca²⁺ responses to Aβ_1–15 in the presence of 50 nm α-bungarotoxin (α-BgTx, n = 12), Aβ_1–15 (n = 20), and Aβ_1–28 (n = 11) on Y188S α7-nAChRs. Peak Ca²⁺ responses to E22Q Aβ_1–42 (n = 29) and E22G Aβ_1–42 (n = 19). B, Peak Ca²⁺ responses via α7-nAChRs to rodent Aβ_1–42 (n = 49), rodent Aβ_1–15 (n = 39), F4A Aβ_1–15 (n = 27), R5A Aβ_1–15 (n = 46), H6A Aβ_1–15 (n = 34), D7A Aβ_1–15 (n = 40); H13A/H14A Aβ_1–15 (n = 25), Aβ_1–12 (n = 48), Aβ_1–16 (n = 51), and Aβ_10–15 (n = 70). C, Averaged peak Ca²⁺ responses to mixtures of Aβ_1–15 and Aβ_1–42 at various concentrations, as noted. *p < 0.05 (Bonferroni post hoc tests) NB. Dashed lines indicate the baseline (background) and respective average maximal responses for either Aβ_1–42 and Aβ_1–15, as indicated.

Figure 3.

N-terminal Aβ fragment augments LTP in hippocampal slices in the absence or presence of elevated full-length Aβ. Hippocampal slices were superfused with aCSF containing vehicle (control) or various concentrations of Aβ_1–15 without (B–D) or with (E, F) Aβ_1–42, followed by the induction of LTP in the CA1 region via theta burst stimulation (TBP: four trains of 100 Hz pulses delivered at 5 Hz repeated three times every 15 s for a total of 3 bursts) or HFS (two 1 s trains of 100 Hz separated by 20 s) through the Schaffer collaterals and expressed as normalized fEPSP slope values. A, Control input/output curves, before treatment. B, Recording during and after the theta burst following 57 fm, 57 pm, or 100 nm Aβ_1–15 for 20 min, with the start of each burst marked with an arrow. Note the change in time scale (dashed lines) for the bursts: PTP marked with a solid bar. C, TBP-induced LTP with color-coded insets showing example fEPSPs for control aCSF (black), femtomolar Aβ_1–15 (red) or picomolar Aβ_1–15 (green) for baseline and LTP. The period of Aβ_1–15 pretreatment is marked by the open bar. D, Average fEPSP slope values for the end of the plateau (50–60 min post-tetanus), as noted by the solid bar in C (*). E, HFS-induced LTP, with color-coded insets showing example fEPSPs for control aCSF (black), 500 nm Aβ_1–42 (blue), or 500 nm Aβ_1–15 followed by 500 nm Aβ_1–42 (green) for baseline and LTP; periods of peptide pretreatment are marked by the bars. F, Average fEPSP slope values for the end of the plateau (50–60 min post-tetanus), as noted by the solid black bar in E (*). Data are the means ± SD, n = 6 slices/group derived from three experiments. Calibration: horizontal, 10 ms; vertical, 0.4 mV. *p < 0.05 (Bonferroni post hoc tests).

Figure 4.

N-terminal Aβ fragment rescues LTP deficits in APPswe mouse hippocampal slices. Hippocampal slices from APPswe or wild-type (WT) littermates were superfused with aCSF containing vehicle (Control) or 500 nm Aβ_1–15. A, HFS-induced LTP, with color-coded insets showing example fEPSPs for slices from WT mice without or with pretreatment with Aβ_1–15 (red) or slices from APPswe mice without or with pretreatment with 500 nm Aβ_1–42 (blue) for baseline and LTP; period of Aβ_1–15 pretreatment marked by the open bar. B, Average fEPSP slope values for the end of the plateau (50–60 min post-tetanus), as noted by the solid bar in A (*). Data are the means ± SD, n = 4 slices/group. Calibration: horizontal, 10 ms; vertical, 0.4 mV. *p < 0.05 (Bonferroni post hoc tests).

Figure 5.

Bilateral delivery of picomolar N-terminal Aβ fragment into the dorsal hippocampus enhances contextual fear conditioning. Mice were trained for contextual fear conditioning under a single-trial paradigm using mild shock. A–C, Twenty-four hours before testing, 1.83 ng/L (100 pm) Aβ_1–15 or 1.83 μg/L (100 nm) Aβ_1–15 (A); 4.5 ng/L (100 pm) Aβ_1–42 or 1.83 ng/L (100 pm) Aβ_1–15 (B); and 1.83 ng/L (100 pm) Aβ_1–15, 1.83 ng/L (100 pm) Aβ_1–15 + 100 nm MLA or 100 nm MLA (C); or sterile saline were bilaterally injected into the dorsal hippocampi. Freezing was measured via TSE videotracking software. Conditioned freezing was assessed by two trained observers. Data are the means ± SEM, n = 6–9 mice/group. *p < 0.005 compared with saline control; ^ap < 0.05 comparing Aβ_1–15 to Aβ_1–42 (Bonferroni post hoc tests).