p58−/− cells exhibit a lower ER substrate burden than wild-type cells. (A) Wild-type (top) and p58−/− (bottom) MEFs were treated with varying concentrations of DTT (0.1, 0.3, and 1 mM), TG (0.1, 1, and 10 μM), or arsenite (Ars; 10, 30, and 100 μM) for 30 min and labeled for 15 min with [³⁵S]methionine before analysis of total cell lysates by SDS-PAGE and autoradiography. Relative levels of [³⁵S]incorporation (quantitated using a phosphorimager) are shown below individual lanes of each panel. The level of synthesis in p58−/− cells relative to wild-type cells for each stressor (mean ± SD; n = 3) is shown below the respective bottom panels. Note that for both DTT and TG (but not Ars), p58−/− cells show about half the level of overall protein synthesis relative to p58+/+ cells. In this and subsequent experiments, Coomassie blue staining before autoradiography confirmed equal total protein recovery and loading (data not shown). (B) After treatment with DTT (1 and 10 mM) or TG (0.1, 1, and 10 μM) and [³⁵S]methionine labeling as described in A, cells were fractionated into cytosolic proteins (left) and glycoproteins (right) that were each analyzed by SDS-PAGE and quantified by phosphorimaging. The amount of ³⁵S incorporation in each lane relative to the untreated condition in that panel is indicated below the autoradiographs. Note that for both p58−/− and p58+/+ cells, glycoproteins are consistently attenuated to a slightly greater degree during stress than cytosolic proteins. Quantitation of the average ³⁵S incorporation into either cytosolic proteins or glycoproteins in p58−/− cells relative to wild-type cells was tabulated as described in A, and it is shown below the respective bottom panels. Note that although both cytosolic proteins and glycoproteins are made to lower relative amounts in p58−/− cells, the glycoproteins are affected to a greater extent. (C) The radiolabeled GCR was quantitated for all five ER stress conditions in B and normalized against the respective GCRs from nontreated cells for both genotypes. Plotted are the overall changes in GCR (mean ± SD) observed during ER stress in p58−/− and p58+/+ cells. (D) Wild-type and p58−/− MEFs were treated with the indicated concentrations of TG for 30 min, followed by RT-PCR to detect both spliced (spl) and unspliced (us) Xbp1 mRNA. Note the slight but consistently higher proportion of spliced Xbp1 seen in p58−/− cells at each stress condition. The image is presented in black-and-white inverted form for greater visual clarity.

p58^IPK contains a functional N-terminal signal sequence. (A) SignalP3.0 analysis of the murine p58^IPK N terminus predicts a likely ER targeting signal (solid line) and cleavage site (broken line). In this analysis, the probability of each residue being part of a targeting signal (based on an artificial neural network trained on known eukaryotic signal sequences) or suitable site for signal peptidase cleavage is plotted (Bendtsen et al., 2004). A horizontal line at the 0.5 probability value for the targeting signal plot is indicated, showing that nearly the entire putative signal sequence of p58^IPK (underlined) is above this cut-off. (B) Transcripts encoding the indicated constructs were translated in vitro in the presence or absence of ER-derived rough microsomes (RM). After synthesis, samples were divided into equal aliquots and treated as indicated with 0.5 mg/ml PK in the presence or absence of 1% Triton-X100 (det) to assess translocation into the lumen of microsomes. Note that calreticulin (Crt) contains a core domain that is resistant to digestion by PK. (C) In vitro translations of full-length murine p58 and Δs.s./p58 were analyzed adjacent to microsomes isolated from mouse liver, TM-treated murine liver, and dog pancreas. Detection of p58 was performed by immunoblotting. (D) NIH3T3 cells were transfected with constructs encoding p58^CHO, Δs.s./p58^CHO, or Prl/p58^CHO, each of which is appended at its C-terminus with an HA epitope tag and contains an inserted potential N-linked glycosylation site near the C terminus. Cell lysates were incubated in the absence or presence of PNGaseF before immunoblot with antibodies against the HA epitope or the ER resident glycoprotein TRAPα. (E) Lysates of NIH 3T3 cells transfected as indicated were treated with PNGaseF or EndoH followed by immunoblot against the HA epitope or the cell surface TfR. Resistance to EndoH digestion is indicative of glycan modifications that accompany trafficking of proteins past the cis-Golgi.

Endogenous p58^IPK is a component of the ER lumen. (A) Postnuclear hypotonic lysates from NIH 3T3 cells were treated with PK in the presence or absence of Triton X-100 (det) and probed by immunoblot against the indicated proteins. Antibodies against the transmembrane proteins TRAPα and TfR were directed against cytosolic epitopes. Note that BiP contains a protease-resistant core that is not digested fully even in the presence of detergent. (B) Postmitochondrial supernatants from homogenized livers of untreated or TM-injected mice (16 h) were sedimented over a cushion of 0.8 M sucrose. The microsomal pellets (P) and cytosolic supernatants (S) were probed by immunoblot against the indicated proteins. Efficacy of the TM was reflected in inhibition of TRAPα glycosylation and increased expression of p58^IPK and BiP, which are both up-regulated by ER stress. (C) Canine pancreatic microsomes were extracted using high salt, physiological salt, or varying concentrations of digitonin (dig) or Triton X-100 (TX). Microsomes were then separated by centrifugation into extracted (supernatants) and nonextracted (pellets) fractions, followed by immunoblot as indicated. (D) Murine liver microsomes prepared as described in B were treated with PK in the presence or absence of detergent (1% digitonin) and analyzed by immunoblot against the indicated proteins to assess their respective topologies. Note the complete protection from PK digestion of both BiP and p58^IPK in the absence of detergent. By contrast, TRAPα (detected using an antibody against its cytosolic tail) and HSP70 are equally accessible in the absence or presence of detergent. Note that both BiP and HSP70 contain protease-resistant cores that are not digested fully even in the presence of detergent. (E) Murine liver microsomes from either nontreated or TM-treated mice were solubilized in a physiological salt buffer containing 1% digitonin. The samples were separated by velocity sedimentation through a continuous sucrose gradient and individual fractions probed by immunoblot against the indicated proteins.

p58^IPK interacts with BiP in the ER lumen. (A) Liver microsomes from nontreated or TM-injected mice were solubilized in 1% Triton X-100 and immunoprecipitated with antiserum against BiP, an antiserum directed against the C terminus of p58^IPK, preimmune serum (PIS), or an irrelevant antibody against interferon-γ (irrel). Immune complexes were then probed by immunoblot (IB) for BiP, p58^IPK, or immunoglobulin (Ig)G light chain. (B) NIH 3T3 cells were transfected with empty vector, Prl/p58, p58, or Δs.s./p58. Cells were incubated in situ with the membrane-permeable cross-linker DSP at 2 mM, followed by lysis under denaturing conditions and immunoprecipitation with antibodies against BiP, the HA epitope, or an irrelevant antigen (interferon-γ). After cross-linker reversal, the immunoprecipitations were probed by immunoblot as indicated. Asterisk represents IgG heavy chain. Expression of Δs.s./p58 is much lower, likely due to degradation. Thus, a darker exposure is also shown to confirm the absence of an association between this protein and BiP. (C) NIH 3T3 cells were transfected with full-length p58^CHO or the same construct with the J-domain deleted (ΔJ). Overexpression of HA-tagged Crt served as a negative specificity control for BiP interaction. Cell lysates were prepared under nondenaturing conditions (no cross-linker was used) and immunoprecipitated with an HA antibody, followed by immunoblot as indicated. Multiple bands for p58^CHO constructs represent heterogeneity of glycosylation for the protein.

p58^IPK interacts with a secretory protein in the ER lumen. (A) Preprolactin was synthesized in vitro by using rabbit reticulocyte lysate and pancreatic microsomes. Untranslocated (pPrl) and translocated, signal sequence-cleaved material (Prl) are indicated. The microsomes were isolated by centrifugation, solubilized under nondenaturating conditions, and subjected to immunoprecipitation by using antibodies against PDI, p58^IPK, BiP, the β subunit of α glucosidase II (also known as 80K-H), and the HA epitope tag. Aliquots of the total translation products (lane 1), the isolated microsomes (lane 2), solubilized microsomes (lane 3), and immunoprecipitates were analyzed by SDS-PAGE and autoradiography. (B) Preprolactin was synthesized in vitro by using rabbit reticulocyte lysate and pancreatic microsomes as described above and adjusted to either 1% Triton X-100 (native conditions; lanes 1–5) or 1% SDS (denaturing conditions; lanes 6–10). After incubation of the denaturing sample at 37°C for 10 min, it was adjusted to 1% Triton X-100 and 0.1% SDS. Both samples were subjected to immunoprecipitation with the indicated antibodies. Aliquots of the input (lanes 1 and 6) and immunoprecipitates are shown.

p58^IPK overexpression alters PrP metabolism. (A) HeLa cells cotransfected with PrP and the indicated p58^IPK constructs were pulse labeled for 10 min and then chased for either 30 or 60 min before immunoprecipitation of PrP. The positions of immature PrP species (containing 0, 1, or 2 N-linked glycans) and post-ER PrP are indicated to the left. The post-ER species of PrP migrate as a heterogenous set of products, the most prominent of which migrates just above the doubly glycosylated immature PrP band. Right, products from pulse and 60-min chase time points (the equivalent of lanes 2 and 9) incubated overnight with either nothing or endoglycosidase H. Independent experiments showed that the expression levels of the different p58^IPK constructs were equal (data not shown; see also C). (B) HeLa cells cotransfected with either PrP (lanes 3–5) or Prl-PrP (lanes 6–8), and the indicated p58^IPK constructs were pulse labeled for 10 min before immunoprecipitation of PrP. In vitro translation products of PrP synthesized in the absence or presence of rough microsomes (RM) to generate markers for percursor (pre) and signal-cleaved (s.c.) species of PrP were analyzed on the same gel (but taken from a longer exposure of the autoradiograph). (C) HeLa cells cotransfected with PrP and either Crt or the indicated p58^IPK constructs without (left) or with BiP (right) were pulse labeled for 10 min before immunoprecipitation of PrP (top) or the cotransfected protein (bottom). Note that in the absence of BiP cotransfection, Prl/p58 overexpression results in decreased PrP translocation (lane 2). This effect is largely offset by BiP overexpression. (D) Pulse-chase analysis of Prl–PrP in HeLa cells cotransfected with either Δs.s./p58 or Prl/p58. Note the more rapid conversion of immature PrP species to mature species in the Prl/p58-expressing cells.

ER stress reduces the efficiency of p58^IPK translocation. (A) NIH 3T3 cells were transfected with p58^CHO or Prl/p58^CHO and treated with 100 nM TG for the indicated times. Lysates were digested with PNGaseF to resolve N-glycosylated from unmodified chains. Although ER stress does not affect either signal sequence cleavage or glycosylation of Prl/p58^CHO, TG inhibits signal sequence cleavage of p58^CHO, such that by 24 h all chains exist in the uncleaved (uncl) state. Of the uncleaved chains, approximately half are glycosylated (+CHO) and half unglycosylated (−CHO). (B) NIH 3T3 cells were transfected with p58^3CHO or Prl/p58^3CHO (identical to the constructs used in A except containing three C-terminal N-linked glycosylation sites to better separate glycosylated from nonglycosylated material) and treated with 100 nM TG for 24 h as indicated. Hypotonic lysates were prepared and treated with PK as described in Figure 3A, or PNGaseF as described above. p58^3CHO and Prl/p58^3CHO were identified by immunoblotting. Note that ER stress causes a substantial portion of p58^3CHO chains, but not Prl/p58^3CHO, to be unglycosylated, and these unglycosylated chains are sensitive to PK digestion, whereas glycosylated chains are resistant. Immunoblotting to detect a lumenal epitope of endogenous calnexin (CNX) confirmed the proper orientation and integrity of the membranous vesicles within the homogenate.