Derlin-2 and Derlin-3 are regulated by the mammalian unfolded protein response and are required for ER-associated degradation

Abstract

Proteins that are unfolded or misfolded in the endoplasmic reticulum (ER) must be refolded or degraded to maintain the homeostasis of the ER. Components of both productive folding and ER-associated degradation (ERAD) mechanisms are known to be up-regulated by the unfolded protein response (UPR). We describe two novel components of mammalian ERAD, Derlin-2 and -3, which show weak homology to Der1p, a transmembrane protein involved in yeast ERAD. Both Derlin-2 and -3 are up-regulated by the UPR, and at least Derlin-2 is a target of the IRE1 branch of the response, which is known to up-regulate ER degradation enhancing alpha-mannosidase-like protein (EDEM) and EDEM2, receptor-like molecules for misfolded glycoprotein. Overexpression of Derlin-2 or -3 accelerated degradation of misfolded glycoprotein, whereas their knockdown blocked degradation. Derlin-2 and -3 are associated with EDEM and p97, a cytosolic ATPase responsible for extraction of ERAD substrates. These findings indicate that Derlin-2 and -3 provide the missing link between EDEM and p97 in the process of degrading misfolded glycoproteins.

Figures

Figure 1.

Identification of Derlin-2 and -3. (A) The levels of CGI-101, BiP, and β-actin mRNA are determined by microarray analysis in HeLa cells treated with or without 2 μg/ml tunicamycin for 8 h. Fold induction was determined, with the means from six independent experiments presented with SDs (error bars). CGI-101 is identical to Derlin-2. (B) Hydropathy plots of human Derlin-2 (CGI-101), -3 (FLJ43842), and -1 are shown. Hydrophobicity and hydrophilicity (expressed by positive and negative numbers, respectively) of the amino acid sequences of three human Derlins were obtained according to the method of Kyte and Doolittle (1982). Black bars mark hydrophobic regions that span the membrane. (C) Amino acid sequence alignment of human Derlin-2, Derlin-3 tv1, Derlin-3 tv2, Derlin-1, and yeast Der1p is shown. Identical amino acids are indicated by white letters in black boxes. Two transcriptional variants for Derlin-3 are deposited in the data bank. Derlin-3 tv1 lacks the 30 COOH-terminal amino acids present in Derlin-3 tv2.

Figure 2.

Structure and tissue distribution of Derlin-2 and -3 mRNA. (A) Schematic structures of transcripts deposited in the data bank for human (available from GenBank/EMBL/DDBJ under accession no. NM_024295) and mouse (NM_024207) Derlin-1, human (NM_016041) and mouse (NM_033562) Derlin-2, and human (NM_001002862) and mouse (NM_024440) Derlin-3 are shown. Black underlines indicate the respective region of cDNA probe used for Northern blot analysis. (B) A nylon membrane onto which ∼2 μg each of poly A+ RNA prepared from eight human tissues was blotted after separation through gel electrophoresis (Human MTN blot) was hybridized with a DIG-labeled cDNA probe specific to human Derlin-1, Derlin-2, Derlin-3, or β-actin. Migration positions of molecular weight markers are indicated on the left of each panel. (C) A second nylon membrane onto which ∼1 μg each of poly A+ RNA prepared from 12 human tissues was blotted after separation through gel electrophoresis (Human 12-lane MTN blot) was hybridized as in B.

Figure 3.

Involvement of the IRE1–XBP1 pathway in the induction of Derlin-1 and -2 in response to ER stress. (A) IRE1α+/+, IRE1α−/−, XBP1+/+, and XBP1−/− MEFs were treated with 10 μg/ml tunicamycin (Tm) for the indicated periods. Total RNAs were isolated and analyzed by Northern blot hybridization using a DIG-labeled cDNA probe specific to mouse Derlin-2, Derlin-1, EDEM, BiP, or GAPDH. Closed and open arrowheads indicate the migration positions of 28S ribosomal RNA (4.7 kb) and 18S ribosomal RNA (1.9 kb), respectively. (B) 293T or HeLa cells were treated with 2 μg/ml tunicamycin (Tm) for the indicated periods. Total RNAs were analyzed as in A using a DIG-labeled cDNA probe specific to human Derlin-2, Derlin-3, EDEM, BiP, or GAPDH.

Figure 4.

Characterization of Derlin-2 and -3. (A) HeLa cells were trans-fected with plasmid to express each Derlin tagged with the c-myc epitope at the respective NH2 terminus. Transfected cells were fixed and stained with anti-myc and anti-Sec61β antibodies. (B) HEK293 cells were trans-fected with plasmid to express each Derlin tagged with the c-myc epitope at the respective COOH terminus. Postnuclear supernatant of transfected cells was incubated with increasing amounts of trypsin (0, 4, 8, and 16 μg for lane 1, 2, 3, and 4, respectively) for 15 min at 4°C. Immunoblotting analysis of the samples was performed using anti-myc, anti–NH2 terminus of calnexin (CNX[N]), anti–COOH terminus of calnexin (CNX[C]), and anti-calreticulin (CRT) antibodies. (C) HEK293 cells were transfected with plasmid to express Derlin-2 (a) or Derlin-3 tv1 or tv2 (b) tagged with the c-myc epitope at either the NH2 (N) or COOH (C) terminus. 36 h later, transfected cells were pulse labeled with 35S-methionine and cysteine for 15 min and then chased for the indicated periods. Cells were lysed with buffer containing 1% NP-40 and subjected to immunoprecipitation analysis using anti-myc antibody.

Figure 5.

Effects of overexpression of Derlin-2 and -3 on degradation of NHK and NHK(QQQ). (A and B) HEK293 cells were mock-transfected or transfected with plasmid to express Derlin-2 or Derlin-3 tv2 tagged with the c-myc epitope at the respective NH2 terminus together with plasmid to express NHK (A) or NHK(QQQ) (B). Transfected cells were pulse-chased and subjected to immunoprecipitation analysis using anti–α1-PI antibody as in Fig. 4 C. Migration positions of NHK, NHK(QQQ), Derlin-2, and Derlin-3 are indicated. The radioactivity of each NHK band was determined and expressed as relative to the summation of radioactivity of nine NHK bands obtained in each experiment. The relative radioactivity of each band was normalized with the value at chase period 0 h. The means from three independent experiments with SDs (error bars) are plotted against the chase period (bottom).

Figure 6.

Effects of knockdown of Derlin-2 and -3 on degradation of NHK. (A) HEK293 cells were untransfected or transfected with the shRNA vector pSUPER alone or pSUPER carrying a sequence corresponding to a part of Derlin-2 or -3. Total RNAs were isolated 64 h after transfection and analyzed by Northern blot hybridization using a DIG-labeled cDNA probe specific to human Derlin-1, Derlin-2, Derlin-3, or GAPDH. (B) HEK293 cells were transfected with pSUPER alone or pSUPER carrying a sequence corresponding to a part of Derlin-2, Derlin-3, or both together with plasmid to express NHK. Pulse-chase and subsequent immunoprecipitation were performed as in Fig. 5 A 64 h after transfection. The results of three independent experiments are shown (left). Normalized radioactivity of each NHK band was determined and is presented as in Fig. 5 A (right). Error bars depict means ± SD. (C) HEK293 cells were transfected with (+) or without (−) plasmid to express Derlin-1 or -2 tagged with the c-myc epitope at the respective NH2 terminus with (+) or without (−) plasmid to express Derlin-3 tv2 tagged with the HA epitope at the NH2 terminus. 36 h later, transfected cells were labeled with 35S-methionine and cysteine for 2 h, lysed with buffer containing 1% NP-40, and subjected to immunoprecipitation analysis using anti-myc or anti-HA antibody as indicated. Migration positions of Derlin-1, Derlin-2, and Derlin-3 tv2 are marked. (D) HEK293 cells were transfected with pSUPER alone or pSUPER derivatives as in B together with plasmid to express the wild-type (wt) α1-PI. Pulse-chase and subsequent immunoprecipitation from cell lysate as well as from media were performed as in B. Migration positions of high-mannose and complex wild-type α1-PI are indicated.

Figure 7.

Association of Derlin-2 and -3 with p97, EDEM, and NHK. (A) HEK293 cells were mock-transfected or transfected with plasmid to express each Derlin tagged with the c-myc epitope at the respective COOH terminus alone (a) or with plasmid to express FLAG-tagged p97 (b). HEK293 cells were also transfected with plasmid to express each Derlin tagged with the c-myc epitope at the respective NH2 terminus (c). 24 h later, transfected cells were labeled with 35S-methionine and cysteine for 1 h, lysed with buffer containing 1% NP-40, and subjected to immunoprecipitation analysis using anti-myc or anti-p97 antibody as indicated. Migration positions of endogenous and FLAG-tagged p97 are marked. The asterisk shows a nonspecific band. The short open arrowhead indicates p97 coimmunoprecipitated with Derlins, whereas long open arrowheads indicate Derlins coimmunoprecipitated with p97. (B, a) HEK293 cells were mock-transfected or transfected with plasmid to express each Derlin tagged with the c-myc epitope at the respective NH2 terminus together with plasmid to express HA-tagged EDEM. 24 h later, transfected cells were labeled with 35S-methionine and cysteine for 2 h, lysed, and subjected to immunoprecipitation analysis using anti-myc or anti-HA antibody as indicated. The migration position of EDEM is marked. The asterisk shows a nonspecific band. The short open arrowhead indicates EDEM coimmunoprecipitated with Derlins, whereas long open arrowheads indicate Derlins coimmunoprecipitated with EDEM. (b and c) HEK293 cells were transfected with (+) or without (−) plasmid to express HA-tagged EDEM with or without plasmid to express Derlin-1 or -2 tagged with the c-myc epitope at the respective NH2 terminus. Transfected cells pulse labeled and then lysed as in a were subjected to immunoprecipitation analysis using anti-HA or anti-p97 antibody as indicated. (c) The amount of Derlin-1 expression plasmid to transfect HEK293 cells was twice that of Derlin-2 expression plasmid. Migration positions of p97, EDEM, and Derlins coimmunoprecipitated with EDEM or p97 are indicated. (C) HEK293 cells were transfected with (+) plasmid to express NHK and with or without plasmid to express Derlin-1 or -2 tagged with the c-myc epitope at the respective NH2 terminus. Transfected cells were pulse labeled and then lysed as in panel a. Equal amounts of cell lysate were subjected to immunoprecipitation analysis using anti–α1-PI, anti-p97, or anti-myc antibody as indicated. Migration positions of p97, NHK, and Derlins are indicated.

Figure 8.

Effects of knockdown of Derlin-1 on degradation of NHK. (A) HEK293 cells were untransfected or transfected with pSUPER alone or pSUPER carrying a sequence corresponding to a part of Derlin-1, -2, or -3. 64 h later, transfected cells were subjected to immunoblotting analysis using antibody against Derlin-1, Derlin-2, or GAPDH. (B) HEK293 cells were transfected with pSUPER alone or pSUPER carrying a sequence corresponding to a part of Derlin-1 or -3 together with plasmid to express NHK. 64 h later, transfected cells were pulse-chased, lysed, and subjected to immunoprecipitation as in Fig. 6 B (left). The results of three independent experiments are shown. Normalized radioactivity of each NHK band was also determined and is presented as in Fig. 5 A (right). Error bars depict mean ± SD.