A reversible form of lysine acetylation in the ER and Golgi lumen controls the molecular stabilization of BACE1

Abstract

The lipid second messenger ceramide regulates the rate of beta cleavage of the Alzheimer's disease APP (amyloid precursor protein) by affecting the molecular stability of the beta secretase BACE1 (beta-site APP cleaving enzyme 1). Such an event is stimulated in the brain by the normal process of aging, and is under the control of the general aging programme mediated by the insulin-like growth factor 1 receptor. In the present study we report that BACE1 is acetylated on seven lysine residues of the N-terminal portion of the nascent protein. This process involves lysine acetylation in the lumen of the ER (endoplasmic reticulum) and is followed by deacetylation in the lumen of the Golgi apparatus, once the protein is fully mature. We also show that specific enzymatic activities acetylate (in the ER) and deacetylate (in the Golgi apparatus) the lysine residues. This process requires carrier-mediated translocation of acetyl-CoA into the ER lumen and is stimulated by ceramide. Site-directed mutagenesis indicates that lysine acetylation is necessary for nascent BACE1 to leave the ER and move ahead in the secretory pathway, and for the molecular stabilization of the protein.

Figure 1. BACE1 is transiently acetylated in the lumen of the ER

(A) CHO cells stably expressing BACE1 were treated with ceramide for the indicated time. BACE1 was immunoprecipitated from total cell lysates, separated on a 7% Tris/Acetate NuPAGE system, and then analysed by Western blotting using either anti-BACE1 or anti-acetylated lysine antibodies. Immature (im. BACE1) and mature (m. BACE1) BACE1 are indicated. (B) BACE1 was immunoprecipitated (IP) from ER and Golgi vesicles and then analysed with the indicated antibodies. ER (calreticulin) and Golgi (58 Golgi protein) markers are shown. For this experiment, samples were separated on a 4–12% Bis/Tris NuPAGE system, which allows a better resolution of total (mature and immature) BACE1. (C) BACE1 was immunoprecipitated from ER and Golgi vesicles as in (B) and then digested with Endo H prior to immunoblotting. (D) Intact ER vesicles were digested with trypsin for 30 min at 25 °C, in the presence or absence of 0.05% (v/v) Triton X-100. Digestion was halted using an anti-trypsin specific inhibitor. BACE1 was then immunoprecipitated using a monoclonal antibody against the N-terminal domain and analysed by Western blotting using the indicated antibodies. (E) BACE1–myc was purified using an anti-myc affinity column and incubated in the presence of HAT and [³H]acetyl-CoA for 1 h at 30 °C. BACE1 was then purified again and analysed on a scintillation liquid counter. *P<0.0005; n=4. (F) Primary neurons and human neuroblastoma (SH-SY5Y) cell lines were treated with ceramide for 6 days. The steady-state levels of BACE1 (top panel) were assessed by immunoblotting of total cell lysates, whereas the acetylation of BACE1 (bottom panel) was assessed following BACE1 immunoprecipitation. Electrophoresis was performed on a 4–12% Bis/Tris NuPAGE system. Neurons were prepared from wild-type mice as described in [8,9]. (G) Primary neurons were cultured in vitro up to 24 days as described previously [8,9] and then analysed for BACE1 acetylation with the indicated antibody. (H) BACE1 was immunoprecipitated from brain (cortex) extracts of 1-month-old wild-type and p44^+/+ mice [9] and then analysed for lysine acetylation with the appropriate antibody. Ac. Lys., acetylated lysine; IP, immunoprecipitate; Wb, Western blot.

Figure 2. Substitution of the lysine residues blocked the ability of BACE1 to be acetylated in vivo and in vitro

(A) CHO cells stably expressing the native (BACE1_WT) or mutated (BACE1_Ala) form of BACE1 were grown in the presence of ceramide and then analysed by Western blotting using an anti-BACE1 antibody (top panel). Cell lysates were separated on a 4–12% Bis/Tris SDS/PAGE system prior to Western blotting. For BACE1 acetylation, BACE1 was immunoprecipitated from total cell lysates and then analysed using an anti-acetylated lysine antibody (bottom panel). (B) Both BACE1_WT and BACE1_Ala were purified from stably transfected cells using a specific anti-BACE1 antibody cross-linked to BioMag Protein A, eluted by lowering the pH, and then incubated in the presence of HAT and [³H]acetyl-CoA for 1 h at 30 °C. BACE1 was then purified again and analysed on a scintillation liquid counter. *P<0.0005; n=3. IP, immunoprecipitate; Wb, Western blot.

Figure 3. Transient lysine acetylation controls both translocation along the secretory pathway and molecular stability of the nascent BACE1 protein

(A) BACE1_WT, BACE1_Ala, and BACE1_Gln cells were labelled with a mixture of radioactive methionine/cysteine for 10 min and then chased for 2 h. Pro-BACE1 was immunoprecipitated from total cell lysates with a specific antibody against the pro-domain region of BACE1. Unbound BACE1 (without the pro-domain) was then immunoprecipitated with an antibody against the C-terminal domain of BACE1. Immunoprecipitates were analysed by SDS/PAGE and autoradiography. (B) BACE1_WT, BACE1_Ala and BACE1_Gln cells were labelled with a mixture of radioactive methionine/cysteine for 10 min and then chased for 1.5 h. Cell surface proteins were biotinylated, separated with Immobilized Streptavidin and eluted by lowering the pH. BACE1 was then immunoprecipitated as above and analysed by reducing SDS/PAGE and autoradiography. The typical migration of cell-surface BACE1 on reducing SDS/PAGE is shown in Supplementary Figures 1A and 1B at http://www.BiochemJ.org/bj/407/bj4070383add.htm. Results are expressed as a percentage of total newly synthesized BACE1. *P<0.005; n=4. (C and D) BACE1_WT, BACE1_Ala and BACE1_Gln cells were labelled with a mixture of radioactive methionine/cysteine for 30 min. and then chased for 8, 16 and 24 h. BACE1 was immunoprecipitated from total cell lysates and analysed by SDS/PAGE and autoradiography. The half-life of BACE1 from five different readings is shown in (C), whereas a representative radiogram is shown in (D). (E) Cell lysates from BACE1_WT, BACE1_Ala and BACE1_Gln cells were separated either on a 7% Tris/Acetate (TA) or on a 4–12% Bis/Tris (BT) SDS/PAGE system, blotted on to a PVDF membrane, and probed with anti-BACE1 antibodies. Both immature (im. BACE1) and mature (m. BACE1) BACE1 are indicated. (F) Autoradiography of BACE1 immunoprecipitated from stably transfected cells immediately after a 30 min period of labelling with [³⁵S]methionine/cysteine (top panel). Samples were separated on a 4–12% Bis/Tris SDS/PAGE system. BACE1 and GAPDH (glyceraldehyde-3-phosphate dehydrogenase; control lane) mRNA levels were quantified after total RNA extraction and RT (reverse transcriptase)-PCR (bottom panel). IP, immunoprecipitate; WT, wild-type.

Figure 4. Both the lysine to alanine and lysine to arginine (loss-of-acetylation) substitutions generate an unstable BACE1 protein that is retained in the early secretory pathway and degraded by a proteasome-independent system.

(A) The steady-state levels of BACE1_Arg were analysed as described in Figure 3(E). (B) The half-life of BACE1_Arg was analysed as in Figure 3(C) and compared with BACE1_Ala (n=3). (C) The cellular distribution of the wild-type and mutant forms of BACE1 was analysed by SDS/PAGE and immunoblotting after separation of intracellular membranes on a 10–24% discontinuous Nycodenz gradient (described in [9,10]). Membranes were developed in parallel on the same film. The appropriate subcellular markers are indicated: calreticulin (ER), syntaxin (Golgi apparatus) and EEA1 (endosomes). A longer exposure of BACE1_WT and BACE1_Gln gradients showing the relative distribution of mature and immature BACE1 is found in Supplementary Figure 1(C) at http://www.BiochemJ.org/bj/407/bj4070383add.htm. (D) BACE1_Ala and BACE1_WT-expressing CHO cells were treated with two different proteasome inhibitors for 10 h. BACE1 steady-state levels were assessed by immunoblotting of total cell lysates on a 4–12% Bis/Tris NuPAGE system. The effects of proteasome inhibition on wild-type BACE1 have been described previously [26]. WT, wild-type.

Figure 5. The ER membrane has an acetyl-CoA transport activity

(A and B) Intact ER vesicles were incubated for 5 min at 30 °C with increasing concentrations of acetyl-CoA while maintaining [³H]acetyl-CoA constant. Translocation was measured as described in the Materials and methods section. The points of the double reciprocal plot (Lineweaver–Burk transformation (B) of the data shown in (A) were fitted by linear regression analysis to give a K_m of 14 μM and a V_max of 824 pmol/mg/5 min (control) (n=6). (C) ER and Golgi vesicles were assayed for transport of acetyl-CoA as described above. *P<0.0005; n=3 (D) Transport of CMP-sialic acid into ER and Golgi intact vesicles was measured as described in (A). The assay was performed at two different concentrations (5 and 10 μM) of solute. *P<0.0005; n=3 (E) Transport of ATP into ER intact vesicles obtained from control and ceramide-treated CHO cells was measured as described in (A). The assay was performed at two different concentrations (5 and 10 μM) of solute.

Figure 6. Acetyltransferase and deacetylase activities exist in the lumen of ER and Golgi apparatus respectively

(A) BACE1–myc was purified with an anti-myc affinity column, and then incubated with [³H]acetyl-CoA and ER or Golgi vesicles in the presence/absence of 0.2% (v/v) Triton X-100 for 1 h at 30 °C. Reaction was stopped by lowering the temperature; BACE1 was then purified again and analysed on a scintillation liquid counter. As control, BACE1 was also incubated with [³H]acetyl-CoA in the absence of ER/Golgi vesicles and with ER vesicles that had been boiled for 10 min prior to the reaction. *P<0.0005; n=4 (B) Purified BACE1–myc was first acetylated in vitro as described in Figure 1(E) and then purified again in order to eliminate unbound acetyl-CoA. For the in vitro deacetylation, acetylated BACE1 was incubated with ER or Golgi intact vesicles in the presence/absence of 0.2% (v/v) Triton X-100 for 1 h at 30 °C. Reaction was stopped by lowering the temperature; BACE1 was then immunoprecipitated and analysed on a liquid scintillation counter. As control, BACE1 was also incubated with Golgi vesicles that had been boiled for 10 min prior to the reaction in order to inactivate any enzymatic activity. *P<0.005; n=4. (C) Both the in vitro acetylation and deacetylation of BACE1 were repeated by using intact ER (for acetylation) or Golgi (for deacetylation) vesicles prepared from control and ceramide-treated CHO cells. The assays were performed at 30 °C and in the presence of 0.2% (v/v) Triton X-100, as described in (A) (for acetylation) or (B) (for deacetylation). *P<0.0005; n=4.

Figure 7. Schematic model of BACE1 transient lysine acetylation/deacetylation

During translation/translocation across the ER membrane (1), BACE1 undergoes acetylation in seven different lysine residues facing the lumen of the organelle. The reaction requires transfer of acetyl-CoA from the cytoplasm, where it is generated, to the lumen of the ER. The acetyl-CoA will then serve as donor of the acetyl group in the reaction of acetylation. This process is favoured by the second messenger ceramide, thereby increasing the concentration of acetyl-CoA in the ER. The enzyme that carries out the reaction, the acetyl-CoA:lysine acetyltransferase, is also an ER-resident protein with the catalytic site facing the lumen of the organelle. The acetylation of BACE1 provides conformational stability to the nascent protein (2). Once this has been achieved, BACE1 moves to the Golgi apparatus, where a Golgi-resident deacetylase will remove the acetyl groups (3). This event is most likely necessary in order to decrease the electron density of the globular domain of the protein (4) and allow conformational flexibility when the mature and active protein needs to shift from the ligand-free to the ligand-bound (and vice versa). In analogy with the many ER and Golgi resident glucosyltransferases already identified, both the acetyltransferase and the deacetylase are depicted here as membrane proteins. However, further studies are required to confirm this assumption.

Figure 8. The acetylated lysine residues are clustered in a disordered region of BACE1

(A) Disorder prediction of BACE1 protein sequence. The prediction was performed using DRIP-PRED analysis (http://www.sbc.su.se/∼maccallr/disorder/) of the Stockholm Bioinformatics Center, Stocholm University, Sweden. The colours used are from blue to red through light green and yellow, where dark blue indicates highly ordered regions. Underlined regions scored >0.5 in the prediction algorithm and are probably disordered (with the probability being higher for red regions). The transmembrane domain is indicated by a box and appears highly ordered; the acetylated lysine residues are circled. (B) Three-dimensional view of BACE1 with the acetylated lysine residues. The structure information, together with the graphic representation, was prepared using the Entrez Molecular Modeling Database (MMDB) available at http://www.ncbi.nlm.nih.gov/Structure/. The acetylated lysine residues are indicated. The white dots correspond to Lys²⁹⁹ and Lys³⁰⁷, which can also be found acetylated prior to ceramide treatment; the other lysine residues are shown in red. The arrowheads indicate the two catalytic aspartic acid groups of the enzyme. (C) Schematic view of the lysine residues that undergo acetylation in BACE1 and p53.