A Peptide Motif Covering Splice Site B in Neuroligin-1 Binds to Aβ and Acts as a Neprilysin Inhibitor

Abstract

The most common cause of dementia among elderly people is Alzheimer's disease (AD). The typical symptom of AD is the decline of cognitive abilities, which is caused by loss of synaptic function. Amyloid-β (Aβ) oligomers play a significant role in the development of this synaptic dysfunction. Neuroligin-(NL)1 is a postsynaptic cell-adhesion molecule located in excitatory synapses and involved in the maintenance and modulation of synaptic contacts. A recent study has found that Aβ interacts with the soluble N-terminal fragment of NL1. The present study aimed to elucidate the role of NL1 in Aβ-induced neuropathology. Employing surface plasmon resonance and competitive ELISA, we confirmed the high-affinity binding of NL1 to the Aβ peptide. We also identified a sequence motif representing the NL1-binding site for the Aβ peptide and showed that a synthetic peptide modeled after this motif, termed neurolide, binds to the Aβ peptide with high affinity, comparable to the NL1-Aβ interaction. To assess the effect of neurolide in vivo, wild-type and 5XFAD mice were subcutaneously treated with this peptide for 10 weeks. We observed an increase in Aβ plaque formation in the cortex of neurolide-treated 5XFAD mice. Furthermore, we showed that neurolide reduces the activity of neprilysin, the predominant Aβ-degrading enzyme in the brain. Accordingly, we suggest that neurolide is the NL1-binding site for Aβ peptide, and acts as an inhibitor of neprilysin activity. Based on these data, we confirm the involvement of NL1 in the development of AD and suggest a mechanism for NL1-induced Aβ plaque formation.

Keywords: Alzheimer’s disease; Amyloid-β; Neprilysin; Neuroligin-1.

Fig. 1

Both NL1 and NL1-derived peptide, neurolide, bind to Aβ. A, B For SPR binding experiments, recombinant NL1 was immobilized on a CM4 sensor chip and A positive control, NX1β, or B Aβ_1-42 were injected into the chip at the indicated concentrations and at a flow rate of 20 μL/min. Sensorgrams of blank-subtracted responses representative from four independent experiments are shown. RU, resonance units. C Ribbon diagram of NL1 (gray) with neurolide motif highlighted in blue and red, wherein the splice insert B is depicted in red (PDB: 5OJK) [13]. D, E Results from competitive ELISA showed the binding of immobilized NL1 (D) or neurolide (E) to serially diluted Aβ_1-40

Fig. 2

Both sNL-1 and neurolide modulate Aβ aggregation. The formation of Aβ amyloid aggregates was tracked by ThT fluorescence measurements either for 210 min (insert in (A)) or over 8 days, using 75 μM of Aβ_1-42 alone or co-incubated with either 1 μM sNL1 (A) or 10 μM of neurolide (B). Treatment with NL1 and neurolide, but not with 10 μM of NLp2 or NLp7 (C, D), showed a significant decrease in the formation of aggregates in comparison with Aβ alone (dashed lines for NLp7 and NLp2). Slopes differences between lineal regression lines for Aβ and NL1, F = 0.033, p = 0.033; Aβ and neurolide, F = 26.34, p = 0.0002; and Aβ and NLp7, F = 0.4, p = 0.4

Fig. 3

Effect of chronic administration of neurolide on memory in WT and 5XFAD mice. A Timeline of the in vivo experiments. Both WT and 5XFAD mice were s.c. injected (black arrow) with either vehicle or neurolide three times per week for 10 weeks (W) from the age of 2 months. B Y-maze test, performed at week 11, showed no neurolide effect on spontaneous alterations. C Number of arms entered either by WT or 5XFAD mice was not different within each genetic group; however, among neurolide-treated mice, 5XFAD transgene mice showed a significant decrease in the number of entrances. The neurolide effect on short-term (D) and long-term (E) memory was estimated using the novel object recognition test. The recognition index was calculated as 100 × Tn/(Tf + Tn), wherein Tf and Tn are the time spent on investigating a familiar and novel object, respectively. Number of animals is indicated in columns. All results are expressed as the mean ± SEM

Fig. 4

Region-dependent neurolide effect on plaque burden in the brain of 5XFAD mice. A The relative area of Aβ amyloid plaques (top row) and average plaque size (bottom row) measured in the cortex, hippocampus, striatum, thalamus, and hypothalamus of 5XFAD mice are depicted. Serial coronal sections of whole brain from 5XFAD mice treated with either vehicle (n = 13) or neurolide (n = 11) were stained with Thioflavine S, and the relative area (top row) and size of Aβ amyloid plaques (bottom row) were quantified using the ImageJ software. B Representative micrographs show Tioflavine S staining of brain sections from WT (top) and vehicle- (middle) and neurolide-treated (bottom) 5XFAD mice. Scale bar, 100 μm. C Quantification of GFAP-positive immunoreactivity in the cortex of 5XFAD mice after treatment with vehicle (n = 13) or neurolide (n = 11). Representative images of GFAP-positive immunoreactivity in cortex areas of vehicle- and neurolide-treated 5XFAD mice are depicted. Scale bar, 50 μm

Fig. 5

NL1 and neurolide effects on neprilysin activity. Both NL1 (A) and neurolide (B) decreased the cleavage activity of recombinant neprilysin in a dose-dependent manner. C Neurolide, but not other NL1-derived peptides, NLp4 and NLp7, down regulates the activity of recombinant neprilysin. D Neurolide (10 μM) downregulates the activity of recombinant neprilysin in the presence of Aβ_1-42 applied either in 0.5 μM or 5 μM. At A–D panels data from four to five independent experiments are shown and compared with vehicle alone. E Both NL1 and neurolide downregulate the activity of neprilysin, extracted from brains of WT mice (n = 5). Treatment of brain extracts with Thiorphan (100 nM) confirm the neprilysin activity in prepared brain fractions and used as a negative control. F Activity of neprilysin extracted from brains of 5XFAD mice (n = 6) is significantly lower compared to WT mice (n = 5) and is further significantly attenuated by treatment with 10 μM neurolide (n = 6). In E–F, neprilysin activity in brain extracts was normalized to the activity of synthetic neprilysin in each experiment, set as 100% (dashed line). Values are mean ± SEM