Neuroprotective Peptide NAPVSIPQ Antagonizes Ethanol Inhibition of L1 Adhesion by Promoting the Dissociation of L1 and Ankyrin-G

Abstract

Background: Ethanol causes developmental neurotoxicity partly by blocking adhesion mediated by the L1 neural cell adhesion molecule. This action of ethanol is antagonized by femtomolar concentrations of the neuropeptide NAPVSIPQ (NAP), an active fragment of the activity-dependent neuroprotective protein (ADNP). How femtomolar concentrations of NAP antagonize millimolar concentrations of ethanol is unknown. L1 sensitivity to ethanol requires L1 association with ankyrin-G; therefore, we asked whether NAP promotes the dissociation of ankyrin-G and L1.

Methods: L1-ankyrin-G association was studied using immunoprecipitation, Western blotting, and immunofluorescence in NIH/3T3 cells transfected with wild-type and mutated human L1 genes. Phosphorylation of the ankyrin binding motif in the L1 cytoplasmic domain was studied after NAP treatment of intact cells, rat brain homogenates, and purified protein fragments.

Results: Femtomolar concentrations of NAP stimulated the phosphorylation of tyrosine-1229 (L1-Y1229) at the ankyrin binding motif of the L1 cytoplasmic domain, leading to the dissociation of L1 from ankyrin-G and the spectrin-actin cytoskeleton. NAP increased the association of L1 and EphB2 and directly activated EphB2 phosphorylation of L1-Y1229. These actions of NAP were reproduced by P7A-NAP, a NAP variant that also blocks the teratogenic actions of ethanol, but not by I6A-NAP, which does not block ethanol teratogenesis as potently. Finally, knockdown of EPHB2 prevented ethanol inhibition of L1 adhesion in NIH/3T3 cells.

Conclusions: NAP potently antagonizes ethanol inhibition of L1 adhesion by stimulating EphB2 phosphorylation of L1-Y1229. EphB2 plays a critical role in synaptic development; its potent activation by NAP suggests that ADNP may mediate synaptic development partly by activating EphB2.

Keywords: EphB2; Ethanol; L1; NAP; Phosphorylation; Y1229.

Figure 1.

NAP effect on the interaction of L1 with ankyrin-G (AnkG) and the spectrin-actin cytoskeleton. (A) L1 was immunoprecipitated using mAb 5G3 from whole cell lysates of NIH/3T3 cells expressing human L1 (2A2-L1_s) that had been treated with NAP at various concentrations; co-immunoprecipitated proteins were separated and blotted with antibodies to L1 or AnkG. Ankyrin-G was detected as a band of approximately 190 kD, but occasionally, a second band was observed at approximately 150 kD, likely representing a splice variant (60). L1 was detected as a doublet at 210 kD and 190 kD. Protein band densities were normalized to values for total L1, and values were expressed as mean ± SEM % of control from 6 - 14 independent experiments (F = 6.59, p < 0.0001). (B) L1 was immunoprecipitated using mAb 5G3 from whole cell lysates of 2A2-L1_s cells treated with 10⁻⁹ M NAP, and co-immunoprecipitated proteins were separated and blotted with antibodies to L1, AnkG, spectrin, or actin. Protein band densities were normalized to values for total L1, and values were expressed as mean ± SEM % control from 5-13 independent experiments (*t=4.4, *p=0.01; *t=2.6; *p=0.03; ***t=5.9, ***p=0.0004).

Figure 2.

L1 association with AnkG and ethanol inhibition of L1 adhesion in cells expressing wild-type L1 (L1-WT) or L1 in which Y1229 was mutated to phenylalanine (L1-Y1229F). (A) NIH/3T3 cells were transfected with L1-WT or L1-1229F in the absence and presence of 10⁻¹² M NAP or 100 μM octanol. L1 was immunoprecipitated from whole cell lysates using mAb 5G3, and coimmunoprecipitated proteins were separated and blotted with antibodies to L1 and AnkG. Densities of AnkG bands were normalized to values for L1 in corresponding experiments and then expressed as a percentage of values in untreated cells expressing L1-WT. Shown are mean ± SEM % association of L1 and AnkG from 4 - 13 independent experiments (F =2.94, p< 0.05); *** t=7.18, p < 0.001. (B) Co-localization of L1 (red) and AnkG (green) immunostaining in L1-WT and L1-Y1229F-expressing cells treated for 1 hour in the absence and presence of 10⁻¹² M NAP. Mander’s overlap coefficient indicated significant differences in the co-localization of L1 and AnkG under various experimental conditions (F = 3.2, p<0.05); ** t=5.26, p <0.01, n = 14-15. (C) and (D) The L1-Y1229F mutation abolished NAP (10⁻⁹ M) and okadaic acid (100 μM) antagonism of ethanol inhibition of L1 adhesion. NIH/3T3 cells were transiently transfected with L1-WT and L1-Y1229F. Cells were treated with 10⁻⁹ M NAP or 100 μM okadaic acid for one hour, and cells were harvested for cell aggregation assays performed in the absence and presence of 100 mM ethanol. Ethanol inhibition of L1 adhesion in L1-Y1229F expressing cells was expressed as a percentage of that obtained in L1-WT expressing cells (47.7 ± 4.4%). Shown are mean ± SEM relative levels of ethanol inhibition in (C) (F = 16.77; p < 0.0001); *** t=5.938, p*** = 0.0000, n = 12-32; (D) (F = 3.90; p < 0.05); ** t = 6.93 , p** =0.002, n = 5.

Figure 3.

NAP effect on tyrosine phosphorylation (pY) of the 16-amino acid FIQGY-peptide fragment derived from the L1-CD, containing a single tyrosine. (A) 2A2-L1_s cells were incubated in the absence and presence of 10⁻⁹ M NAP, and extracted cell lysates were incubated with FIGQY-peptide conjugated to agarose beads at room temperature for the indicated time (Methods). FIGQpY-peptide was measured using mouse anti-pY mAb (PY20). The magnitude of pY in FIGQY-peptide was quantified by densitometric analysis of anti-pY antibody-agarose bead bands separated by SDS-PAGE. Densities of FIGQpY-peptide bands were normalized to values obtained at time 0. Shown are mean ± SEM % FIGQpY-peptide compared to values at time 0 from 5 independent experiments (F =9.84, p< 0.0001). Peak phosphorylation occurred at 10 minutes following NAP treatment of intact cells. (B) Dose response curve for NAP phosphorylation of FIQGY-peptide by cell lysate from NAP-treated 2A2-L1_s cells. Reactions were carried out for 10 min at room temperature in the presence of the indicated concentrations of NAP. pY levels were normalized to values in cells that were not treated with NAP (lane 0). Shown is a representative gel from 12 experiments. Densitometry was obtained from 7 – 12 independent experiments (F = 2.48, p < 0.05). (C) Lysate of cerebral cortex from post-natal day 10 rat pups was incubated in the absence and presence of 10⁻¹² M NAP for 10 minutes. NAP treatment of rat brain lysates significantly increased phosphorylation of FIGQY-peptide (n=7; **t=2.888, ** p < 0.01). (D) 2A2-L1_s cells were incubated for 15 minutes at room temperature in the absence and presence of 10⁻⁹ M NAP and 20 uM PP2, and cell lysates were then incubated with FIGQY-peptide to determine the effect of drug treatments on pY; *t=3.45, *p=0.010, n=8; **t=3.55, **p=0.0085, n=8.

Figure 4.

NAP activation of EphB2 phosphorylation of L1. (A) Effect of NAP on association of L1 with EphB2. L1 was immunoprecipitated with mAb 5G3 from 2A2-L1_s cells in the absence and presence of 10⁻⁹ M NAP, and coimmunoprecipitated proteins were separated and blotted with antibodies to L1 and EphB2. Densities of EphB2 bands were normalized to those for L1, and values for NAP treatment were expressed as a percentage of control values. Shown is the mean ± SEM % increase in L1 association with EphB2 following NAP treatment derived from 8-9 independent experiments; *t=3.17,*p <0.05, n = 8. (B) Dose-dependent stimulation by NAP of tyrosine phosphorylation of FIGQY-peptide by recombinant EphB2 (F = 2.45, p<0.05). pY levels following NAP treatment were normalized to control values (0 NAP). Shown is the mean ± SEM % increase in pY levels following treatment with the indicated concentrations of NAP derived from 6-9 independent experiments (C) Stimulation of EphB2 phosphorylation of L1 by 10⁻⁹ M NAP and P7A-NAP (P7A), but not by 10⁻⁹ M I6A-NAP (I6A), SAL, or octanol (Oct). Shown is a representative gel and densitometric analysis from 7 independent experiments. pY levels for each drug treatment were normalized to values obtained in the absence of drugs (Control) (F = 2.84, p<0.05); **t = 4.01, **p = 0.0070; *t= 3.01, *p=0.0235, n=7.

Figure 5.

Effect of EphB2 knockdown on NAP antagonism of ethanol inhibition of L1 adhesion. (A) 2A2-L1_s cells were treated with an EphB2 siRNA or a scrambled siRNA. L1 and actin were used as loading controls, and densities of protein bands were normalized to those obtained in untreated cells (Control). Shown is a representative gel and mean ± SEM % changes in EphB2 expression derived from 9 independent experiments (F=9.33, p< 0.001); * t =1.07, p <0.0.0044, n=10. (B) EphB2 siRNA specifically reduced while scrambled siRNA had no effect on the phosphorylation of FIGQY-peptide from lysates of 2A2-L1_s cells treated with 10⁻¹² M NAP. Values for pY in NAP-treated siRNA-treated cells were normalized to values in control cells that were not treated with NAP (F = 2.77, p < 0.05); *t = 2.47, p =0.04, n=8; t = 2.54, p=0.039, n =8. (C) EphB2-siRNA specifically reduced while scrambled siRNA had no effect on NAP antagonism of ethanol inhibition of L1 adhesion. Values for ethanol inhibition of L1 adhesion in the presence of NAP were normalized to values obtained in the absence of NAP (38.6 ± 6.9%) (F = 7.17, p < 0.0001); t= 9.30, ***p =0.0000, n = 9; t = 7.41, ***p = 0.0001, n =9).

Figure 6.

NAP antagonizes ethanol inhibition of L1 adhesion by activating EphB2 phosphorylation of L1-Y1229, leading to the dissociation of L1 from ankyrin-G and the spectrin-actin cytoskeleton. L1 is sensitive to ethanol only when it is associated with ankyrin-G.