DHHC8-dependent PICK1 palmitoylation is required for induction of cerebellar long-term synaptic depression

Abstract

The palmitoyl acyltransferase (PAT) DHHC8 is implicated in synaptic regulation but few DHHC8 substrates are known. Here we report that DHHC8 binds and palmitoylates the PDZ domain-containing protein PICK1 at a cysteine residue that is essential for long-term synaptic depression (LTD) in cultured mouse cerebellar Purkinje neurons. Cerebellar LTD is palmitoylation-dependent and induction of LTD requires DHHC8. Furthermore, PICK1 is a critical DHHC8 substrate whose palmitoylation is necessary for LTD. These results identify the first DHHC8 substrate required for a specific form of synaptic plasticity and provide new insights into synaptic roles of palmitoylation.

Figure 1.

The DHHC8 C-terminal PDZ ligand binds PICK1. A, DHHC8 schematic, showing predicted transmembrane domains (blue) and catalytic DHHC domain (red). A C-terminal 15 aa sequence (yellow, identical in DHHC5) terminates in a PDZ ligand (ISV-COO, orange). The yeast two-hybrid bait region is indicated. B, HEK293T cells were transfected with myc-tagged full-length PICK1, plus either GST alone (GST), a GST fusion of the DHHC8 C terminus (GST-8wt), or a GST fusion of the DHHC8 C terminus lacking the PDZ ligand (GST-8ΔC). Inputs (left) and GST pull-downs (right) were immunoblotted with the indicated antibodies. C, The DHHC5/8 common C terminus is sufficient to bind PICK1. HEK293T cells were cotransfected with myc-PICK1 plus either GST or a GST fusion of the DHHC5/8 C-terminal 15 aa (GST-15AA DHHC5/8). Inputs (left) and GST pull-downs (right) were immunoblotted with the indicated antibodies.

Figure 2.

DHHC8 palmitoylates PICK1 at Cys414, and PICK1 is endogenously palmitoylated in neurons. A, PICK1 Schematic, showing PDZ and BAR domains, plus C-terminal residues critical for actin binding (W413) and palmitoylation (C414, red underlined). B, HEK293T cells were transfected with His-myc-DHHC8 and either wild-type or point mutant (C414S) myc-PICK1. Palmitoyl-proteins were isolated by ABE and PICK1 levels in ABE samples were detected by immunoblotting (top). Lysates were blotted to detect total PICK1 (middle) and His-myc-DHHC8 (bottom) levels. Note that palmitoyl-PICK1 isolation requires hydroxylamine (NH₂OH), an essential ABE component. C, Immunoblots from multiple experiments as in Figure 2B (PICK1wt, 6 determinations per condition, PICK1–C414S, 3 determinations per condition) were quantified to detect palmitoylated and total PICK1 levels. Mean palmitoylated:total PICK1 signal is plotted as a percentage of maximum (PICK1wt + DHHC8). Error bars indicate SEM. *p < 0.05 compared with PICK1 + vector alone, t test. D, Neuronal PICK1 is endogenously palmitoylated. The indicated juvenile rat brain regions were homogenized and subjected to ABE to purify palmitoyl proteins. ABE samples and homogenates were blotted to detect PICK1. E, Cultured cortical neurons metabolically labeled with [³H]-palmitate were lysed and immunoprecipitated with the indicated antibodies. [³H] signal (top) and PICK1 protein levels (bottom) are shown.

Figure 3.

Cerebellar LTD in cultured PNs depends on palmitoylation and a specific PICK1 palmitoylation site. A, Wild-type mouse cultures were treated for >30 min with the indicated concentrations of 2-bromopalmitate or with vehicle (EtOH, <0.02%). After baseline recording of glutamate-evoked inward currents, LTD was induced by iontophoretic glutamate pulses in conjunction with depolarization. Six pairings were delivered at t = 0 min (horizontal bar) and 12 pairings at t = 22.5 min (horizontal bar, 2×). Representative single current traces were acquired at the indicated times. Scale bars, 1 s, 40 pA. PICK1 CKO mouse cultures were infected with Cre-expressing lentivirus at DIV1 and transfected with PICK1wt, PICK1–C414S, or PICK1–W413A cDNAs at DIV9–DIV10 before baseline recording and LTD induction at DIV13–DIV14. PICK1^−/−; PICK1wt plasmid (n = 6); PICK1^−/−; PICK1 CS mutant plasmid (n = 8); PICK1^−/−; PICK1 W413A actin-binding mutant plasmid (n = 7); WT: 2-Br-palmitate, 10 μm, preincubated (n = 5); WT: 2-Br-palmitate (20 μm, preincubated (n = 7); WT; vehicle (n = 6). Normalized current amplitude at t = 40 min was compared using the Mann–Whitney U test. This revealed a significant difference between the PICK1^−/−; PICK1 WT plasmid and PICK1^−/−; PICK1 CS mutant plasmid groups (p < 0.001) and between the WT; vehicle and WT, 2-Br-palmitate 20 μm groups (p < 0.001). B, For each condition in A, we attempted to induce chemical LTD by bath applying the PKC activator PDA (200 nm) at t = 0–10 min as indicated by the horizontal bar. mEPSCs were measured and are expressed as mean amplitudes. n = 10 cells/group. mEPSC amplitudes at t = 45 min were compared using the Mann–Whitney U test. This revealed a significant difference between the PICK1^−/−; PICK1 WT plasmid and PICK1^−/−; PICK1 CS mutant plasmid groups (p < 0.02) and between the WT; vehicle and WT, 2-Br-palmitate 20 μm groups (p < 0.02). C, DHHC8 palmitoylates PICK1–W413A. HEK293T cells were transfected with His-myc-DHHC8 plus either wild-type or actin-binding mutant (W413A) myc-PICK1. ABE samples were blotted to detect palmitoyl-PICK1 (top). Lysates were blotted to detect PICK1 (middle) and His-myc-DHHC8 (bottom) total expression.

Figure 4.

Cerebellar LTD in cultured PNs depends on DHHC8 catalytic activity, and palmitoyl-PICK1 is a key DHHC8 substrate that controls LTD. A, Wild-type mouse cultures were infected with lentiviruses expressing the indicated shRNAs. A subset of cultures infected with DHHC8 shRNA was subsequently transfected with the indicated DHHC8 rescue constructs. Baseline recording and LTD induction were performed as in Figure 3. Scale bars, 1 s, 50 pA. DHHC5 shRNA (n = 8); DHHC8 shRNA (n = 8); DHHC5 + DHHC8 shRNA (n = 7); empty vector (n = 7); DHHC8 shRNA; DHHC8 WT plasmid (n = 9); DHHC8 shRNA; DHHS8 transferase dead plasmid (n = 8); DHHC8 shRNA; DHHC8ΔC plasmid (n = 7). Normalized current amplitude at t = 40 min was compared using the Mann–Whitney U test. This revealed a significant difference between the empty vector and DHHC8 shRNA groups (p < 0.001) and between the DHHC8 shRNA; DHHC8 WT plasmid and DHHC8 shRNA; DHHC8 transferase dead plasmid groups (p < 0.001). B, For each condition in A, PDA (200 nm) was bath applied to induce chemical LTD and mEPSCs were measured. N = 10 cells/group. mEPSC amplitudes at t = 45 min were compared using the Mann–Whitney U test. This revealed a significant difference between the empty vector and DHHC8 shRNA groups (p < 0.05) and between the DHHC8 shRNA; DHHC8 WT plasmid and DHHC8 shRNA; DHHS8 transferase dead plasmid groups (p < 0.05). C, PICK1 CKO mouse cultures were coinfected with two lentiviruses, one expressing Cre and the second expressing DHHC8 shRNA and subsequently transfected with either PICK1wt or the palmitoylation-mimicking PICK1-CaaX. Baseline recording and LTD induction were performed as in Figure 3A. Scale bars, 1 s, 50 pA. PICK1^−/−; DHHC8 shRNA; PICK1 WT plasmid (n = 9); PICK1^−/−; DHHC8 shRNA; PICK1-CaaX plasmid (n = 10). Normalized current amplitude at t = 40 min was compared using the Mann–Whitney U test. This revealed a significant difference between the two groups (p < 0.001). D, For each condition in C, PDA was bath applied to induce chemical LTD and mEPSCs were measured. N = 10 cells/group. mEPSC amplitudes at t = 40 min were compared using the Mann–Whitney U test. This revealed a significant difference between the two groups (p < 0.02).