Non-muscle myosin IIA transports a Golgi glycosyltransferase to the endoplasmic reticulum by binding to its cytoplasmic tail

Abstract

The mechanism of the Golgi-to-ER transport of Golgi glycosyltransferases is not clear. We utilize a cell line expressing the core 2 N-acetylglucosaminyltransferase-M (C2GnT-M) tagged with c-Myc to explore this mechanism. By immunoprecipitation using anti-c-Myc antibodies coupled with proteomics analysis, we have identified several proteins including non-muscle myosin IIA (NMIIA), heat shock protein (HSP)-70 and ubiquitin activating enzyme E1 in the immunoprecipitate. Employing yeast-two-hybrid analysis and pulldown experiments, we show that the C-terminal region of the NMIIA heavy chain binds to the 1-6 amino acids in the cytoplasmic tail of C2GnT-M. We have found that NMIIA co-localizes with C2GnT-M at the periphery of the Golgi. In addition, inhibition or knockdown of NMIIA prevents the brefeldin A-induced collapse of the Golgi as shown by the inhibition of the migration of both Giantin, a Golgi matrix protein, and C2GnT-M, a Golgi non-matrix protein, to the ER. In contrast, knockdown of HSP70 retains Giantin in the Golgi but moves C2GnT-M to the ER, a process also blocked by inhibition or knockdown of NMIIA. Also, the intracellular distribution of C2GnT-M is not affected by knockdown of β-coatomer protein with or without inhibition of HSPs, suggesting that the Golgi-to-ER trafficking of C2GnT-M does not depend on coat protein complex-I. Further, inhibition of proteasome results in accumulation of ubiquitinated C2GnT-M, suggesting its degradation by proteasome. Therefore, NMIIA and not coat protein complex-I is responsible for transporting the Golgi glycosyltransferase to the ER for proteasomal degradation. The data suggest that NMIIA is involved in the Golgi remodeling.

Fig. 1. The proteins pulled down from the lysates of Panc1-bC2GnT-M (c-Myc) cells with anti-c-Myc Abs

(A): Coomassie blue stained SDS-PAGE gel analyzed on the proteins in the complexes pulled down from the lysate of Panc1-bC2GnT-M (c-Myc) cells with mouse anti-c-Myc Abs. (B): Tryptic peptides derived from the 250 kDa protein band in (A) matched to the amino acid sequence of NMIIA heavy chain. SDS-PAGE was performed on 10% gel under non-reducing condition. Proteomics analysis of these four bands identified: Non-muscle myosin IIA and IIB in 1^st and 2^nd bands, cytoskeletal keratin, type II in 2^nd and 3^rd band, and gelsolin in 4^th band. The tryptic peptides are underlined. (C): Western blots of the cell lysates, the supernatants, and the complexes pulled down from the lysates of Panc1-bC2GnT-M (c-Myc) cells with control IgG, anti-c-Myc, or anti-NMIIA Abs. Top panel: NMIIA heavy chain western blot. Bottom panel: C-Myc western blot. The signal intensities of the bands were quantitated by Image J software. The efficiency of the pulldown was assessed by the ratio: amount of protein detected in each IP condition/ Input (in %). For this calculation, we considered the Input (40 μg of protein) and the amount of protein employed for the immunoprecipitation for each cell line (0.8 mg). About 90% of NMIIA and 57% of c-Myc were pulled down by anti-NMIIA Abs, 76% of NMIIA and 97% of c-Myc were pulled down by anti-c-Myc Abs. The results shown are representative of three independent experiments.

Fig. 2. Interactions of NMIIA heavy chain with hC2GnT-M cytoplasmic tail

(A): Western blot of NMIIA pulled down from Panc1-bC2GnT-M (c-Myc) cell lysates with biotinylated hC2GnT-M N-terminal peptide: 1-21 aa (lane 1), 1-7 aa (lane 2), and 7-21 aa (lane 3). The complexes pulled down with these peptides were run on 8% SDS-PAGE gel and blotted with rabbit anti-NMIIA Abs. Lane 4: Panc1-bC2GnT-M (c-Myc) cell lysate treated only with Streptavidin dynabeads. (B): NMIIA western blot of the recombinant C-terminal region (1682-1960 aa) of NMIIA heavy chain pulled down with biotinylated hC2GnT-M (1-21 aa) peptide (lane 2) and control peptide (7-21 aa) (lane 1). (C): Yeast two-hybrid analysis of the interaction of hC2GnT-M peptide (1-10 aa) and the C-terminal 454 aa peptide (1507-1960 aa) of NMIIA heavy chain. Yeast cells at 1.5×10⁷cells/10 μl suspension were plated on growth medium lacking leucine, tryptophan and histidine in the presence of 100 ng/ml Aureobasidin A. Compared with the negative control (SV40-T antigen + Lamin), the 454 aa peptide (1507-1960 aa of the NMIIA heavy chain) showed a moderate interaction with hC2GnT-M (1-10 aa) peptide, which was reduced after WKR⁶ was replaced with AAA⁶. SV40-T + p53 is a positive control. (D) hC2GnT-M peptide (1-14 aa) deletion mutants - Δ2,3; Δ4-6; Δ7-9; Δ10-12; Δ13,14. (E) Yeast two-hybrid analysis of the interaction of the C-terminal 454 aa peptide (1507-1960 aa) of NMIIA heavy chain and hC2GnT-M peptides (C2GnT-M^(1-14), C2GnT-M^(Δ2,3), C2GnT-M^(Δ4-6), C2GnT-M^(Δ7-9), C2GnT-M^(Δ10-12) and C2GnT-M^(Δ13,14). All mutants except C2GnT-M^(Δ4-6) show the growth intensity comparable to that of the wild-type peptide in the presence of histidine and absence of Aureobasidin A.

Fig. 3. Confocal immunofluorescence microscopic images of C2GnT-M, Giantin, and NMIIA in Panc1-bC2GnT-M (c-Myc) cells

(A–C): bC2GnT-M (green) co-localizes with Giantin (red), a Golgi matrix protein; (D–G): bC2GnT-M co-localizes with NMIIA at the Golgi periphery (arrows indicate peri-Golgi staining of NMIIA). (H–K): Treatment with a NMIIA inhibitor Blebbistatin abolishes co-localization of NMIIA with bC2GnT-M in the Golgi; bC2GnT-M is retained in the Golgi while NMIIA is localized in the cytoplasm. Bars, 10 μm. G & K are enlarged images of F & J, respectively (arrows indicate peri-Golgi co-staining of NMIIA and C2GnT-M). (L): NMIIA western blot of the complexes pulled down with anti-c-Myc Abs from lysates of Panc1-bC2GnTMc(c-Myc) cells treated with DMSO (-) or Blebbistatin (+). The amount of protein employed for the immunoprecipitation for both lane is 0.7 mg.

Fig. 4. NMIIA inhibition or knockdown followed by BFA treatment of Panc1-bC2GnTM (c-Myc) cells retains C2GnT-M and Giantin in the Golgi

(A–D): Confocal immunofluorescence images of C2GnT-M (green) and NMIIA (red) in cells treated with DMSO (A), Blebbistatin (B), BFA (C), and Blebbistatin followed by BFA (D). (E–H): Confocal images of C2GnT-M (green) and Giantin (red) in cells treated with scrambled siRNA (E), NMIIA siRNA (F), scrambled siRNA followed by BFA (G), and NMIIA siRNA followed by BFA (H). The images were overlaid with DIC. Bars, 10 μm. (I–L): Confocal images of C2GnT-M and Giantin in cells treated with DMSO (I) or BFA (J), and confocal images of C2GnT-M and an ER marker, protein disulfide isomerase (PDI) in cells treated with DMSO (K) or BFA (L). (M & N): Quantification of the intensity of bC2GnT-M fluorescence in the cytoplasm vs. that in the Golgi (=100%) as shown in A–D and E–H, respectively. Data were obtained from 50 cells in three independent experiments and expressed as mean percentage ± SEM; *, P<0.001 for B, C, or D vs. A and F, G, or H vs. E. (O) NMIIA western blot of the lysates of Panc1-bC2GnT-M cells (50 μg) treated with scrambled or NMIIA-specific siRNAs.

Fig. 5. NMIIA inhibition or knockdown prevents HSP inhibition or HSP-70 knockdown-induced shifting of C2GnT-M cellular localization from Golgi to ER

(A–F): Confocal immunofluorescence images of bC2GnT-M (green) and Giantin (red) in Panc1-bC2GnT-M cells treated with DMSO (A & D), KNK437 (B & E), and Blebbistatin followed by KNK437 (C & F). D–F are enlarged images of A–C, respectively. Bars, 10 μm. (H–K): Confocal images of bC2GnT-M and Giantin in cells treated with scrambled siRNAs (H), scrambled siRNAs followed by KNK437 (I), HSP-70 siRNAs (J), and NMIIA siRNAs followed by KNK437 (K). Bars, 10 μm. (G & L): Fluorescence intensity of C2GnT-M in the cytoplasm relative to that in the Golgi (=100%) in A–C and H–K was quantified and analyzed as described above. *, P<0.001 for B or C vs. A, and I or J vs. H; and **, P<0.01 for K vs. H. (M): HSP-70 western blot of the lysates of cells (50 μg) treated with scrambled or HSP-70 specific siRNAs. KNK437 treatment or HSP-70 knockdown retained Giantin in the Golgi and distributed bC2GnT-M to the ER. Golgi retention of both was observed when NMIIA was inhibited or knocked down coupled with inhibition of HSPs. (N & O): Confocal images of C2GnT-M and an ER marker, protein disulfide isomerase (PDI) in cells treated with DMSO (N) or KNK437 (O). (P) HSP-70 western blot of the complexes pulled down with non-specific IgG and anti-c-Myc Abs from cell lysates. The amount of protein employed for the immunoprecipitation was 0.8 mg.

Fig. 6. β-COP knockdown does not affect intracellular distribution of C2GnT-M in Panc1-bC2GnT-M (c-Myc) cells with or without treatment with KNK437

(A–D): Confocal images of bC2GnT-M and Giantin in cells treated with scrambled siRNAs (A), scrambled siRNAs followed by KNK437 (B), β-COP siRNAs (C), and β-COP siRNAs followed by KNK437 (D). Bars, 10 μm. (E): Quantification of the C2GnT-M fluorescence signal in the cytosolic compartment relative to that of the Golgi (=100%). The analysis was performed on the complete confocal stacks of 50 cells from three independent experiments. Data are expressed as mean percentage ± SEM *, P<0.001 for B vs. A, and D vs. C. The bC2GnT-M distribution was not affected after β-COP knockdown. KNK437 treatment has distributed bC2GnT-M to the ER even in the β-COP deficiency. (F): β-COP western blot of the lysates of cells (50 μg) treated with scrambled or β-COP-specific siRNAs. (G–J): Confocal images of β-COP and Giantin or C2GnT-M and Giantin in cells treated with DMSO or KNK437. Note that in control (G), β-COP distributes throughout the cell, but locates primarily in the Golgi area after KNK treatment (H). Unlike β-COP, C2GnT-M moves away from the Golgi following KNK437 treatment (I & J).

Fig. 7. Proteasomal degradation of C2GnT-M in Panc1-bC2GnT-M (c-Myc) cells

(A): Western blot of bC2GnT-M in the lysates of cells (40 μg) treated with DMSO or MG-132. Cell lysates were run on 8 % SDS-PAGE with β-mercaptoethanol and then probed with antic-Myc Abs. (B): Densitometric analysis of the protein bands in (A). Data represent bC2GnTM expressed relative to the density of γ-tubulin band and each bar represents mean ± SEM from three independent experiments. *, P<0.001 for MG-132 vs. DMSO. (C): Ubiquitin western blot of the complexes pulled down with anti-c-Myc Abs from the lysates of cells treated with DMSO or MG-132. Equal amounts of proteins in the lysates (0.8 mg) were used for anti-c-Myc Co-IP. (D & E): Confocal images of bC2GnT-M and NMIIA in cells treated with DMSO (D) or MG-132 (E). (F): NMIIA western blot of the lysates of cells (40 μg) treated with DMSO, Blebbistatin or MG-132. The 250 kDa band corresponded to NMIIA and γ-tubulin was used as a control for equal loading. (G): Densitometric analysis of the protein bands in (F). Data show the amounts of NMIIA relative to that of γ-tubulin and each bar represents mean ± SEM from three independent experiments. *, P<0.001 for Blebbistatin or MG-132 vs. DMSO. (H): Ubiquitin western blot of anti-c-Myc (bC2GnT-M) immunoprecipitate from the lysates of Panc1-bC2GnT-M (c-Myc) cells treated with DMSO or Blebbistatin. Co-IP samples were run on 12% SDS-PAGE with β-mercaptoethanol and then probed with anti-ubiquitin Abs. The amount of protein employed for the immunoprecipitation for both lane was 0.7 mg.

Fig. 8. Proposed role of NMIIA in the Golgi-to-ER transport of C2GnT-M

Treatment of Panc1-bC2GnT-M cells with BFA, which inhibits the ER-to-Golgi transport, results in the collapse of the Golgi. Inhibition (with KNK437) or knockdown (with siRNA) of HSP-70, which is required for the ER-to-Golgi transport of C2GnT-M, retains C2GnT-M in the ER but Giantin in the Golgi. Inhibition (with Blebbistatin) or knockdown (with siRNA) of NMIIA retains C2GnT-M and Giantin in the Golgi. Subsequent treatment with BFA or KNK437 still retains both C2GnT-M and Giantin in the Golgi. Therefore, NMIIA is required for the Golgi-to-ER transport of C2GnT-M. Inhibition or knockdown of NMIIA prevents the Golgi-to-ER movement of C2GnT-M even in the presence of reagents that are known to stimulate the exodus of this enzyme from the Golgi. Note that unlike BFA, KNK437 does not alter the Golgi morphology.

Fig. 9. Intracellular trafficking and distribution of C2GnT-M

C2GnT-M is synthesized in the ER, where HSP-70 helps fold C2GnT-M properly, and then transported to the Golgi, where the enzyme is retained by binding to a protein yet to be identified. C2GnT-M returns to the ER in a NMIIA-dependent but COPI-independent process. This process is mediated by the interaction of the cytoplasmic tail (1-6 aa) of C2GnT-M with the C-terminal portion (1682-1960 aa) of NMIIA heavy chain. C2GnT-M is degraded in the proteasome in a process mediated by ubiquitination.