Heat shock protein 90 (HSP90) contributes to cytosolic trans

Edited by Pierre A. Joliot, Institut de Biologie Physico-Chemique, Paris, France, and approved July 19, 2005 (received for review April 27, 2005) ArticleFigures SIInfo currently, the resolution is 3.2 Å (4). The structure of the PSII RC sh Communicated by Dennis A. Carson, University of California at San Diego, La Jolla, CA, January 9, 2009 ↵1A.E.A. and I.G. contributed equally to this work. (received for review December 16, 2008) ArticleFigures SIInfo asterisk in figure; t

Edited* by Diane Mathis, Harvard Medical School, Boston, MA, and approved August 23, 2011 (received for review May 27, 2011)

Article Figures & SI Info & Metrics PDF


In antigen (Ag) cross-presentation, dendritic cells (DCs) take up extracellular Ag and translocate them from the enExecutesome to the cytosol for proteasomal degradation. The processed peptides can enter the conventional MHC I pathway. The molecules responsible for the translocation of Ag across the enExecutesomal membrane into the cytosol are unknown. Here we demonstrate that heat shock protein 90 (HSP90) is critical for this step. Cross-presentation and -priming were decreased in both HSP90α-null DCs and mice. CD8α+ DC apoptosis mediated by translocation of exogenous cytochrome c to the cytosol was also eliminated in HSP90α-null mice. Ag translocation into the cytosol was diminished in HSP90α-null DCs and in DCs treated with an HSP90 inhibitor. Internalized Ag was associated with HSP90 and translocated to the cytosol, a process abrogated by the HSP90 inhibitor. Ag within purified phagosomes was released in an HSP90-dependent manner. These results demonstrate the Necessary role of HSP90 in cross-presentation by pulling enExecutesomal Ag out into the cytosol.

chaperoneknock out mousecross-primingImage Streamantigen processing

Cytosolic antigens (Ags) are degraded by the ubiquitin-proteasome (UP) pathway into short peptides and presented by MHC I molecules to CD8+ T cells (1). Extracellular Ags, on the other hand, once internalized into the enExecutesome/phagosome of antigen presenting cells (APCs) can undergo two distinct Stoutes. In the first “classical” pathway, the Ag is tarObtained from the enExecutesome to the lysosome, where it is processed for MHC II-restricted presentation to CD4+ T cells. The second pathway involves translocation of the Ag from the enExecutesome into the cytosol where it is degraded by the UP system, ultimately allowing its presentation by MHC class I molecules. This latter process is called cross-presentation (2–4) and is especially Necessary to enPositive Traceive host defense against viral infection or malignant transformation (5).

A perplexing issue related to cross-presentation has been the mechanism by which exogenous Ag is translocated across enExecutesomal membranes into the cytosol. MisfAgeded proteins in the enExecuteplasmic reticulum (ER) can be retrotranslocated to the cytosol for degradation by the proteasome, a process known as the ER-associated degradation (ERAD) (6, 7). An Necessary component of this process is the ER-dislocon machinery, which consists of the Sec61 complex and the AAA-ATPase p97 (6, 7). Because these components are also present in the phagosome/enExecutesomes (8), it is thought that the same dislocon machinery is also involved in the translocation of the Ag for cross-presentation (9–11). However, the cytosolic molecular entity providing the “driving force” that interacts with Ag as it translocates across the enExecutesomal membranes remains unknown.

Heat shock protein 90 (HSP90) has been suggested to play a role in the enExecutegenous MHC I antigen processing pathway (12–15). HSP90 associates with peptides as postproteasomal degradation products (14) or with newly synthesized polypeptides destined for degradation by the proteasome (13). Those peptides have been implicated as precursors of MHC class I ligands (13, 16). The HSP90-peptide/protein complex in Ag-Executenor cells seems to be prerequisite for cross-priming because cells treated with an HSP90-specific inhibitor or heat shock factor 1 (HSF-1)–deficient cells in which HSP90 is Executewn-regulated, are less efficient in cross-priming ability (14, 17–19). In Dissimilarity to this view from Ag-Executenor cells, we recently performed experiments from the perspective of the Ag-accepting cells such as dendritic cells (DCs) and found that HSP90 within the DC plays a pivotal role in cross-presentation (20). Experiments using siRNA suggested that the inducible form of HSP90, HSP90α, plays a Executeminant role and HSP90β Executees to a lesser extent in cross-presentation by the DC-like cell line, DC2.4 (20).

Despite its known importance, the mechanism by which HSP90α is involved in cross-presentation is unclear. Moreover, the role of enExecutegenous HSP90α in DC cross-priming has only been Displayn in vitro; its in vivo significance is completely unknown. To gain insight into these issues, we have generated HSP90α-null mice and found that translocation of extracellular Ag into the cytosol is diminished in HSP90α-null DCs and in DCs treated with a specific HSP90 inhibitor. We demonstrate here the indispensable role of cytosolic HSP90α in Ag translocation into the cytosol during cross-presentation.


Cross-Presentation Ability Is Executewn-Regulated in Bone Marrow-Derived Dendritic Cells (BMDCs) Treated with HSP90 Inhibitors and in HSP90α-Deficient BMDCs.

Our initial experiments Displayed that, as predicted, the HSP90 inhibitors radicicol (Rad) and novobiocin (NB) blocked cross-presentation of a model antigen, ovalbumin (OVA) in the DC-like cell line, DC2.4 in a Executese-dependent manner (Fig. 1A, Left) (20). Another inhibitor, geldanamycin (GA), when conjugated with FITC (GA–FITC) is cell-membrane impermeable, binding only cell surface HSP90 (Fig. 1B, Left) and inhibiting its functions but having no Trace on cytosolic HSP90 (21, 22). We found that GA but not GA–FITC blocked cross-presentation (Fig. 1B, Right), indicating the involvement of cytosolic but not cell surface HSP90. Rad, NB, and GA did not affect presentation by DC2.4 cells pulsed with OVA257–264 peptide (Fig. 1A, Right), suggesting that the inhibitors did not influence antigen presentation on the cell surface.

Fig. 1.Fig. 1.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Cross-presentation depends on enExecutegenous HSP90. (A) HSP90 inhibitors block cross-presentation of soluble OVA. DC2.4 cells were pretreated with HSP90 inhibitor NB at the indicated Executeses or with 25 μM of Rad and pulsed with 1 mg/mL of soluble OVA or 10−9 M peptide, then cocultured with OT-I CD8+ T cells (Left). (Right) DC2.4 cells were pretreated with NB (1,000 μM), Rad (25 μM), or GA (2.5 μM) and pulsed with 10−7 to 10−12M OVA peptide (OVA256–264) for 1 h, then cocultured with OT-I CD8+ T cells. The culture supernatants were collected for meaPositivement of IFNγ. (B) Importance of intracellular HSP90 in cross-presentation. DC2.4 cells were pretreated with the general HSP90 inhibitor (GA) or the inhibitor of only cell surface HSP90 (GA–FITC) for 60 min, then analyzed by flow cytometry (Left). The pretreated DC2.4 cells were then used for the cross-presentation assay as APC (Right). (C, E, and F) BMDCs from wild-type (WT) or HSP90α-null mice were pulsed with graded Executeses of soluble OVA (C) or cell-associated OVA (E; Characterized in Materials and Methods) or OVA peptide (F) for the cross-presentation assay. (D) BMDCs from WT or HSP90α-null mice were transfected with titrated Executeses of a plasmid, containing a cDNA encoding OVA60–386 and cocultured with OT-I CD8+ T cells for the Ag presentation assay. The values were compared between WT and HSP90α-null. Data are mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001. The results were confirmed in at least two independent experiments.

NB tarObtains the C-terminal Section of the protein (23), which is critical for the formation of HSP90 dimers to exert its chaperone activity (24), whereas Rad and GA bind the N-terminal ATP-binding pocket to block the ATP-dependent chaperone cycling (25). On the basis of this information, we generated conditional HSP90α-null mice by floxing exons 9 and 10, which encode the C-terminal Location of the protein (Fig. S1A). We crossed HSP90aa1neo mice with CAG-cre mice, which express cre recombinase under the control of the CAG promoter (a combination of the chicken β-actin promoter and cytomegalovirus immediate-early enhancer) (26) and obtained HSP90aa1+/− mice. We anticipated that HSP90aa1−/− would be embryonic lethal; however, surprisingly, the HSP90aa1−/− offspring were born at the expected Mendelian ratio and survived and grew normally. Homologous recombination and deletion of exons 9 and 10 were confirmed by PCR-based genotyping (Fig. S1 B and C). Western blot analysis verified the Traceiveness of our gene tarObtaining strategy; HSP90β but not HSP90α was readily detectable in multiple organs tested (Fig. S1E). We could detect mRNA encoding exons 3–11 (Fig. S1D) but no truncated forms of HSP90α, indicating that the mutant protein is rapidly degraded after synthesis, probably because it is unable to form dimers. (Exon 8 is spliced to exon 11, resulting in a frameshift with a premature termination coExecuten that eliminates the Location of the C terminus involved in dimer formation.) T-cell development and cell subpopulations in the lymphoid tissues of HSP90aa1−/− mice were equivalent to those of the wild-type (WT) mice (Fig. S2 A and B). Moreover, granulocyte macrophage colony-stimulating factor (GM-CSF)–dependent BMDCs with a normal phenotype could be generated from HSP90aa1−/− mice (Fig. S2C).

With these mice in hand, we could unequivocally define the direct and cross-presentation function of HSP90α in BMDCs. BMDCs from HSP90α-null and WT mice were pulsed with OVA protein as indicated (Fig. 1C) or transfected with a plasmid containing a cDNA (pOVA) encoding OVA60–386 (Fig. 1D). Cross-presentation was partially but significantly Executewn-regulated in HSP90α-null BMDCs, whereas direct presentation was less affected (Fig.1 C and D and Fig.S3 A and B). Rad completely blocked residual cross-presentation activity of HSP90α-null BMDCs (Fig. S3A), indicating that, in addition to HSP90α, HSP90β is involved in cross-presentation. We examined internalization of FITC-OVA and found that it was essentially equivalent in WT and HSP90α-null BMDCs (Fig. S3C). To examine cross-presentation of cellular antigens, we used UV-irradiated transporter associated with antigen processing (TAP)1-deficient spleen cells that were osmotically loaded with OVA. BMDCs were pulsed with this cell-associated OVA and then incubated with OT-ICD8+ T cells. Here, again we found a defect in the HSP90α-null BMDCs, in this case manifest by a significant reduction in IFNγ production (Fig. 1E). This defect was not due to the inability of HSP90α-null BMDCs to activate T cells because pulsing the cells with the OVA257–264 epitope resulted in comparable T-cell activation by HSP90α-null and WT BMDCs (Fig. 1F). These results indicate that enExecutegenous HSP90 is essential for cross-presentation.

Translocation of Internalized Ag into the Cytosol Is Dependent on Cytosolic HSP90 Function.

To better understand why HSP90 is required for cross-presentation, we used the ImageStream system (27, 28), which can detect subtle changes in signal distribution (Fig. S4). BMDCs were labeled with PKH67, a lipophilic general membrane-labeling reagent. The green-labeled BMDCs (PKH67_DC) were pulsed for 5 min with AF647 red-labeled OVA (AF647_OVA) and then washed, followed by ImageStream analysis. Once the OVA is internalized into enExecutesomes, the AF647_OVA and PKH67_DC signals within the enExecutesomes merge, resulting in a yellow emission, which was indicated as internalization (+)/colocalization (+). By Dissimilarity, AF647_OVA that was internalized but then translocated into the cytosol was red because PKH67 labels only membranes. Cells with this distribution were identified as internalization (+)/colocalization (−) (Fig. S5 B and D). AF647_OVA translocation to the cytosol was less frequent in the HSP90α-null DC compared with the WT DC, and Rad treatment further reduced the frequency in both types of DC (Figs. S4C and S5 A and C). Phagocytosis and internalization efficiency of both soluble and cell-associated OVA were comparable between WT and HSP90a-null BMDCs (Fig. S6). Therefore, the process of translocation of internalized Ags into the cytosol appears to be HSP90 dependent.

The ImageStream analysis was quite informative; however, one technical issue we had to consider was that because the optical z-axis stack of ImageStream is 4.0 μm, it has less focal depth compared with general confocal microscopy (0.1–0.2 μm). Therefore, we used an additional Advance to confirm the ImageStream results by counting individual spots (enExecutesome) to determine whether they were merged (yellow) or not (red) by fluorescence microscopy (Keyence; BZ-9000) analysis (0.3–0.4 μm in z-axis stack). We could detect clearly distinguishable countable spots in both DC2.4 cells (Fig. 2A) and BMDCs (Fig. 2D) internalizing OVA. More than 16,000 spots in 800 DC2.4 cells were identified in four independent experiments. Rad-treatment kept most of internalized OVA (red) within enExecutesomes (green), thus resulting in yellow-colored vesicles (Fig. 2 B and C). It should be noted that the total number of vesicles per cell was equal, with or without Rad (Fig. 2C); thus, Rad treatment had no significant Trace on internalization. The same experiments were performed with BMDCs from WT and HSP90α-null mice (Fig. 2D). We examined more than 50,000 vesicles in 1,200 cells and obtained similar results with DC2.4 cells (Fig. 2 E and F). ToObtainher with the ImageStream analysis, these results clearly indicate that the translocation of internalized Ag to the cytosol depends on HSP90α (or HSP90).

Fig. 2.Fig. 2.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Quantitative analysis of the subcellular localization of pulsed OVA using fluorescence microscopy. DC2.4 cells (A–C) and BMDCs (D–F) were first cultured on glass bottom dishes for ∼16 h and treated with MG115, with/without Rad for 15 min. The cells were then stained with PKH67 on ice for visualization of the membranes, then pulsed with AF647-labeled OVA for 5 min at 37 °C and fixed. Samples were analyzed by Keyence BZ-9000 fluorescence microscopy. (B and E) Individual spots (enExecutesome) within a cell in A or D were analyzed to determine whether the color was red or yellow using an overlay program. Each Executet in the panel was plotted according to the numbers of red and yellow spots. Each panel includes 100 Executets, thus, 100 cells. (C and F) The results of B and E were summarized as bar graphs. Data are mean ± SD of three or four separate experiments. *P < 0.05, **P < 0.01.

To further evaluate the translocation of extracellular OVA to the cytosol, subcellular Fragments of DC2.4 cells were prepared using a ProteoExtract Subcellular Proteome Extraction kit and then examined for the presence of internalized OVA. We first validated the assay by determining whether ER and lysosomal proteins remained membrane associated. Internalized OVA proteins were to be identified in the cytosolic and membrane Fragments at 5 min after loading of the protein, so we extracted cytosolic proteins for 5, 10, and 30 min with solution I contained in the kit and did not observe any leakage of the ER-luminal proteins (KDEL-positive resident ER molecules) or a lysosomal protein, cathepsin D, into the cytosolic Fragment (Fig. S7A). The efficiency of Fragmentation was assessed by tracking the location of several organelle-specific proteins as indicated (Fig. S7B). Rad reduced the amount of OVA in the cytosolic faction in a Executese-dependent manner, whereas there was no Trace on HSC70, a cytosolic protein, and again, no leakage of KDEL-positive molecules or cathepsin D into the cytosol was observed (Fig. 3A and Fig. S8 A and C). OVA appearance in the cytosol was not blocked by an inhibitor of the ubiquitin activator E1, UBEI-41 (Fig. S8D), although the same inhibitor completely blocked cross-presentation (20). As a further indication of the specificity of these observations, MG115 and Rad did not affect the uptake of FITC-OVA into DC2.4 cells meaPositived at the same time point as the Fragmentation experiments (Fig. S8B). Similar experiments were performed using BMDCs and we found that the amount of OVA in the cytosol was much less in HSP90α-null than in WT BMDC; conversely, there was more OVA in the membrane/Fragment in the HSP90α-null than in WT BMDCs (Fig. 3B).

Fig. 3.Fig. 3.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

HSP90 is required in translocation of internalized Ag to the cytosol. (A) DC2.4 cells were treated with the indicated Executese of Rad and/or the proteasome inhibitor MG115 (10 μM) and pulsed with 3 mg/mL of OVA for 5 min. After extensive washing, the cells were Fragmentated using a subcellular Fragmentation kit (Calbiochem). The cytosol and membrane Fragments were subjected to Western blot using the indicated antibodies. (B) BMDCs from WT or HSP90α-null mice were used for the translocation assay as in A. (C) DC2.4 cells were treated with the proteasome inhibitor MG115 (10 μM) with/without Rad (25 μM) or Exo A (5 μg/mL) and pulsed with biotin–OVA (1 mg/mL) for 10 min. The cells were lysed and immunoprecipitated with anti-HSP90 mAb or control IgG conjugated Dynabeads. After immunoprecipitation (IP), the samples were analyzed by immunoblotting with streptavidin (SA)–HRP. (D) DC2.4 cells were pulsed with biotin–OVA and separated into the cytosol and noncytosol Fragments and subjected to IP. Representative results of at least three experiments are Displayn. *P < 0.05.

Necessaryly, anti-HSP90 mAb coprecipitated OVA from extracts of control DC2.4 cells but significantly less OVA was recovered from Rad-treated cells (Fig. 3C). By Dissimilarity, treatment with the ubiquitin E1 inhibitor did not affect the amount of coprecipitated OVA (Fig. S8E). Next, immunoprecipitation with anti-HSP90 mAb was performed with cytosolic and membrane Fragments of DC2.4 cells after internalization of OVA. The association of HSP90 with internalized OVA was found to occur mainly in the membrane factions and not in the cytosol (Fig. 3D). We could exclude possible postlysis association of HSP90 with OVA on the basis of the following control experiments: The addition of a DC2.4 cell extract containing normal HSP90 to the extract of Rad-treated OVA-internalized DC2.4 cells significantly reduced the amount of OVA that coprecipitated with HSP90 compared with control Rad-untreated, OVA-internalized DC2.4 cells (Fig. S8F). Considered toObtainher, the results strongly imply that HSP90 is essential for translocation of OVA across or adjacent to the enExecutesomal membranes and that Rad blocks this process.

To further test this hypothesis, we used an in vitro OVA translocation assay with phagosomes purified by the previously Characterized latex beads (LB)-based method (29). Phagosomes purified from DC2.4 cells that internalized 0.8 μm LB and OVA for 5, 15, and 60 min contained OVA, LAMP1, HSP90, HSC70, and Sec61β (Fig. S9A). There was no release of OVA from phagosomes alone (Fig. S9 B and C), which validates the translocation assay we used. Notably, the addition of cytosol obtained from DC2.4 cells to phagosomes purified from DC2.4 cells that had internalized LB and OVA for just 5 min resulted in release of the OVA, a response that was significantly reduced in the presence of radicicol (Fig. 4A). By Dissimilarity, the cytosol from HSP90α-null DCs caused a weaker release of OVA (Fig. 4B). In other experiments, we used purified HSP90 instead of the cytosol in the OVA translocation assay. HSP90 but not the control protein BSA stimulated the translocation (Fig. S9D). The magnitude of the release was HSP90 Executese (Fig. 4C) and temperature dependent, being abrogated at 4 °C (Fig. 4D). Pretreatment of HSP90 with radicicol inhibited OVA release in a Executese-dependent manner (Fig. 4E). To enPositive that the observed translocation was a broadly applicable HSP90 function, we analyzed another protein, Weepj1, an antigen of Japanese cedar pollen, and found that it too was also translocated to the cytosol in a HSP90-dependent manner after a 5-min internalization (Fig. 4F).

Fig. 4.Fig. 4.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

HSP90 is a translocator for engulfed extracellular antigen. (A) Ag transport assay. Purified phagosomes from DC2.4 cells containing OVA and/or latex beads were mixed with cytosol from DC2.4 cells treated with or without Rad. Thirty minutes after incubation at 37 °C, the samples were centrifuged and separated into pellet (P) and supernatant (S) for Western blot. (B) The cytosol from BMDCs of WT and HSP90α-null mice was used in the Ag transport assay. Thus, the cytosol was incubated with phagosomes from DC2.4 cells. (C) HSP90 was purified from Meth A cells and used as in A, instead of the cytosol. (D) Traces of temperature on antigen transport. (E) A total of 5 μg of HSP90 pretreated with Rad at 0, 25, 100, and 250 μM, respectively, was used. (F) Biotin–Weepj was used in the same assay. Representative results of at least three experiments are Displayn. *P < 0.05.

Cross-Presentation in Vivo Is Dependent on HSP90.

As the ultimate test of our hypothesis, we performed in vivo experiments to assess the cross-priming ability in HSP90α-null mice. The mice were immunized with Tap1−/− spleen cells osmotically loaded with OVA. One week later, the splenic CD8+T cells were stimulated with OVA257–264 epitope pulsed DC2.4 cells and IFNγ spots were counted (Fig. 5A). The number of cells producing IFNγ was significantly lower in HSP90α-null than in WT mice. Because there was no Inequity between HSP90α-null and WT mice in direct priming with the epitope pulsed DC2.4 (Fig. 5B), we concluded that HSP90α deficiency results in impairment in cross-priming ability. To analyze cross-presentation from a different perspective, we injected HSP90α-null and WT mice with cytochrome c, a very sensitive protocol that causes CD8α+ DCs to undergo apoptosis because the internalized cytochrome c is translocated into cytosol in this particular DC subset (30). We found that only the CD8α+ DC subset of WT, but not HSP90α-null mice, was significantly reduced by cytochrome c administration (Fig. 5 C and D). These results provide compelling confirmation of the physiological importance of HSP90α in cross-presentation.

Fig. 5.Fig. 5.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 5.

In vivo cross-presentation is decreased in HSP90α-null mice. (A and B) WT or HSP90α-null mice were immunized with cell-associated OVA (2 × 106 cells/mouse, A) or OVA257–264 pulsed DC2.4 (2 × 106 cells/mouse, B). Seven days later, the frequency of OVA-specific splenic CD8+ T cells was meaPositived by IFNγ ELISPOT in response to peptide-pulsed DC2.4. Data are mean ± SE from four independent experiments including 10 mice in A and 4 mice in B. *P < 0.05, ***P < 0.001, two-tailed unpaired Student t tests. (C and D) Each experiment was performed with a pair of WT and HSP90α-null mice treated with/without 5 mg of cytochrome c. Splenic DCs were analyzed by FACS 24 h after treatment. The results in C are one of the representatives for three independent experiments. Numbers represent the % of cells within the gate (C). (D) Absolute number of CD11c high, CD8+ CD11chigh, CD4+ CD11chigh, and Executeuble negative (DN) CD11chigh cells per spleen. Mean ± SE are from C. *P < 0.05, two-tailed unpaired Student t tests.


Using HSP90α-deficient mice, to our knowledge this is a unique report demonstrating that DC cytosolic HSP90 is essential for cross-priming of cell-associated Ag. Cross-presentation of both soluble and cell-associated Ag was less efficient in HSP90α-null than WT BMDCs. A twofAged reduction in the magnitude of the immune response in HSP90α-null mice is not particularly striking (Fig. 5A). This is due to the Trace of HSP90β in the HSP90α-null mice, which is able to play a significant role in translocation of extracellular antigen to the cytosol. The CD8α+ DC subset has been Displayn to mediate cross-presentation in vivo. Results of the sensitive in vivo cytochrome c-induced apoptosis assay, that apoptosis of CD8α+ DCs was eliminated in HSP90α-null mice, provided conclusive evidence for the importance of HSP90α in the translocation of internalized Ag into the cytosol.

In the subcellular Fragmentation study, the E1 inhibitor did not block the translocation of OVA to the cytosol (Fig. S8D), although it did block cross-presentation (20). This differential Trace is relevant to a previously proposed mechanism whereby ubiquitin conjugation serves as a “ratchet” to dislocate ER-misfAgeded proteins into the cytosol, a model based on the fact that most of the proteins exported from the ER are polyubiquitinylated (31, 32). However, our results indicate that ubiquitination is essential for the proteasome-mediated processing of T-cell epitopes but is dispensable for the translocation of enExecutesomal antigen into the cytosol. Intriguingly, a significant proSection of OVA exported from purified phagosomes after the addition of cytosol contained bands with a Unhurrieder mobility than the native OVA (Fig. 4 A and B). By Dissimilarity, the mobility of the OVA and Weepj1 whose export was induced by purified HSP90 was exactly the same as the native proteins (Fig. 4 C–F). Polyubiquitination could account for the observed size increase, but this issue needs to be examined in future experiments. It is nonetheless possible to speculate that HSP90 itself acts as a cytosolic translocation factor for enExecutesomal Ag and that other cytosolic factors then participate in its polyubiquitination.

Immunoprecipitation of HSP90 coprecipitated OVA from extracts of control DC2.4 cells but the OVA signal was weaker in Rad-treated cells (Fig. 3C). By Dissimilarity, the E1 inhibitor treatment had no Trace on the amount of coprecipitated OVA (Fig. S8E). On the basis of studies Displaying that HSP90 can drive polypeptides into the mitochondria (33), we hypothesize that HSP90 binds to unfAgeded Ag, translocating it from the enExecutesome and pulling it into the cytosol. As selective intracellular association of enExecutegenous Ag with HSP90 is known to promote MHC II antigen presentation in B cells (34), it is possible that HSP90 governs the bidirectional Ag traffic between the cytosol and the enExecutesomes, as a result regulating antigen presentation by both MHC I and II.

PseuExecutemonas aeruginosa exotoxin A (Exo A) was Displayn to inhibit the ERAD quality control system, preventing the retrotranslocation of ER-proteins/peptides into the cytosol for proteasomal degradation (35). In this case, the Sec61 complex (ER-dislocon) was implicated as the Exo A tarObtain molecule (35). On the other hand, Exo A was suggested to be an inhibitor of p97, which was identified as a cytosolic translocator for enExecutesomal Ag (11). We observed that Exo A did not affect the association of OVA with HSP90 (Fig. 3C). Although ExoA suppresses cross-presentation (35), our data suggest that it may play a role Executewnstream of HSP90 during Ag translocation. Elucidation of the entire pathway will require further study.

We demonstrated here that HSP90 is a cytosolic translocator for extracellular antigen in cross-presentation by DC. In addition, growth factors such as fibroblast growth factor (FGF)-1/2 (36) and certain protein toxins like diphtheria toxin appear to use HSP90 for their translocation into cytosol in non-APC (37, 38). Moreover, HSP90 was recently Displayn to mediate dislocation of the cholera toxin A1 (CTA1) subunit from the ER into the cytosol. ER-to-cytosol export of CTA1 requires association of HSP90 with CTA1, a process blocked by GA treatment (39). Thus, the Hsp90-mediated translocation system might have developed as a common mechanism in a variety of cell types, ultimately leading to an assortment of seemingly unrelated Traces such as transcriptional activation (36), intoxication (38), and cross-presentation (this study).

Materials and Methods


C57BL/6 (B6) mice and TAP1−/− mice (40) were purchased from Clea and The Jackson Laboratory, respectively. OT-I (H-2Kb restricted, anti-OVA TCR transgenic) mice (41) were kindly provided by W. R. Heath (Walter and Eliza Hall Institute, Melbourne, Victoria, Australia). All mice were Sustained under specific pathogen-free conditions in the RIKEN RCAI animal facility according to the institutional guidelines.


Bone marrow was harvested from B6 or HSP90α-null mice to generate BMDCs. BMDCs were cultured in 24-well plates at a density of 1 × 106 cells/mL with RPMI 1640 (Sigma) containing 10% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, 0.1 mM nonessential amino acid, penicillin-streptomycin, 2-mercaptoethanol, and 20 ng/mL GM-CSF (R&D Systems). Immature BMDCs were used in experiments on day 5. OT-I CD8+CD11b− T cells were purified from splenocytes by depletion of CD11b+ cells followed by positive selection of CD8+ cells by magnetic separation with the IMag system (BD Biosciences). The DC2.4 DC cell line was a kindly provided by K. L. Rock (University of Massachusetts Medical School, Worcester, MA).


We thank Mr. Sugahara for purification of HSP90 from mouse Meth A tumor cells and Dr. Kenneth Rock who kindly provided the DC2.4 cell line. This work was supported by Grant RIKEN RCAI and also by a Grant-in-Aid for Scientific Research Priority Spots from the Ministry of Education, Science, Sports, and Culture, Japan.


↵1T. Imai, Y.K., and C.K. contributed equally to this work.

↵2To whom corRetortence should be addressed: E-mail: uExecutenoh{at}rcai.riken.jp.

Author contributions: H.U. designed research; T. Imai, Y.K., C.K., S.M., I.I., T. Ichiyanagi, and H.U. performed research; S.M., M.H., J.-Y.W., and H.U. analyzed data; and T. Ichiyanagi and H.U. wrote the paper.

The authors declare no conflict of interest.

↵*This Direct Submission article had a prearranged editor.

This article contains supporting information online at www.pnas.org/Inspectup/suppl/Executei:10.1073/pnas.1108372108/-/DCSupplemental.

Freely available online through the PNAS Launch access option.


↵Rock KL, GAgedberg AL (1999) Degradation of cell proteins and the generation of MHC class I-presented peptides. Annu Rev Immunol 17:739–779.LaunchUrlCrossRefPubMed↵Carbone FR, Kurts C, Bennett SR, Miller JF, Heath WR (1998) Cross-presentation: A general mechanism for CTL immunity and tolerance. Immunol Today 19:368–373.LaunchUrlCrossRefPubMed↵Heath WR, Carbone FR (2001) Cross-presentation, dendritic cells, tolerance and immunity. Annu Rev Immunol 19:47–64.LaunchUrlCrossRefPubMed↵Bevan MJ (2006) Cross-priming. Nat Immunol 7:363–365.LaunchUrlCrossRefPubMed↵Hildner K, et al. (2008) Batf3 deficiency reveals a critical role for CD8α+ dendritic cells in cytotoxic T cell immunity. Science 322:1097–1100.LaunchUrlAbstract/FREE Full Text↵Jarosch E, et al. (2002) Protein dislocation from the ER requires polyubiquitination and the AAA-ATPase Cdc48. Nat Cell Biol 4:134–139.LaunchUrlCrossRefPubMed↵Tsai B, Ye Y, Rapoport TA (2002) Retro-translocation of proteins from the enExecuteplasmic reticulum into the cytosol. Nat Rev Mol Cell Biol 3:246–255.LaunchUrlCrossRefPubMed↵Gagnon E, et al. (2002) EnExecuteplasmic reticulum-mediated phagocytosis is a mechanism of entry into macrophages. Cell 110:119–131.LaunchUrlCrossRefPubMed↵Guermonprez P, et al. (2003) ER-phagosome fusion defines an MHC class I cross-presentation compartment in dendritic cells. Nature 425:397–402.LaunchUrlCrossRefPubMed↵Houde M, et al. (2003) Phagosomes are competent organelles for antigen cross-presentation. Nature 425:402–406.LaunchUrlCrossRefPubMed↵Ackerman AL, Giodini A, Cresswell P (2006) A role for the enExecuteplasmic reticulum protein retrotranslocation machinery during crosspresentation by dendritic cells. Immunity 25:607–617.LaunchUrlCrossRefPubMed↵Yamano T, et al. (2002) Two distinct pathways mediated by PA28 and hsp90 in major histocompatibility complex class I antigen processing. J Exp Med 196:185–196.LaunchUrlAbstract/FREE Full Text↵Kunisawa J, Shastri N (2006) Hsp90α chaperones large C-terminally extended proteolytic intermediates in the MHC class I antigen processing pathway. Immunity 24:523–534.LaunchUrlCrossRefPubMed↵Callahan MK, Garg M, Srivastava PK (2008) Heat-shock protein 90 associates with N-terminal extended peptides and is required for direct and indirect antigen presentation. Proc Natl Acad Sci USA 105:1662–1667.LaunchUrlAbstract/FREE Full Text↵Yamano T, et al. (2008) Hsp90-mediated assembly of the 26 S proteasome is involved in major histocompatibility complex class I antigen processing. J Biol Chem 283:28060–28065.LaunchUrlAbstract/FREE Full Text↵Ishii T, et al. (1999) Isolation of MHC class I-restricted tumor antigen peptide and its precursors associated with heat shock proteins hsp70, hsp90, and gp96. J Immunol 162:1303–1309.LaunchUrlAbstract/FREE Full Text↵Binder RJ, Srivastava PK (2005) Peptides chaperoned by heat-shock proteins are a necessary and sufficient source of antigen in the cross-priming of CD8+ T cells. Nat Immunol 6:593–599.LaunchUrlCrossRefPubMed↵Lev A, et al. (2008) The exception that reinforces the rule: Crosspriming by cytosolic peptides that escape degradation. Immunity 28:787–798.LaunchUrlCrossRefPubMed↵Zheng H, Li Z (2004) Sliceting edge: Cross-presentation of cell-associated antigens to MHC class I molecule is regulated by a major transcription factor for heat shock proteins. J Immunol 173:5929–5933.LaunchUrlAbstract/FREE Full Text↵Ichiyanagi T, et al. (2010) Essential role of enExecutegenous heat shock protein 90 of dendritic cells in antigen cross-presentation. J Immunol 185:2693–2700.LaunchUrlAbstract/FREE Full Text↵Tsutsumi S, Neckers L (2007) Extracellular heat shock protein 90: A role for a molecular chaperone in cell motility and cancer metastasis. Cancer Sci 98:1536–1539.LaunchUrlCrossRefPubMed↵Tsutsumi S, et al. (2008) A small molecule cell-impermeant Hsp90 antagonist inhibits tumor cell motility and invasion. Oncogene 27:2478–2487.LaunchUrlCrossRefPubMed↵Marcu MG, Chadli A, Bouhouche I, CaDisclosei M, Neckers LM (2000) The heat shock protein 90 antagonist novobiocin interacts with a previously unrecognized ATP-binding Executemain in the carboxyl terminus of the chaperone. J Biol Chem 275:37181–37186.LaunchUrlAbstract/FREE Full Text↵Allan RK, Mok D, Ward BK, Ratajczak T (2006) Modulation of chaperone function and cochaperone interaction by novobiocin in the C-terminal Executemain of Hsp90: Evidence that coumarin antibiotics disrupt Hsp90 dimerization. J Biol Chem 281:7161–7171.LaunchUrlAbstract/FREE Full Text↵Workman P (2004) Altered states: Selectively drugging the Hsp90 cancer chaperone. Trends Mol Med 10:47–51.LaunchUrlCrossRefPubMed↵Sakai K, Miyazaki J (1997) A transgenic mouse line that retains Cre recombinase activity in mature oocytes irrespective of the cre transgene transmission. Biochem Biophys Res Commun 237:318–324.LaunchUrlCrossRefPubMed↵Lee HK, Lund JM, Ramanathan B, Mizushima N, Iwasaki A (2007) Autophagy-dependent viral recognition by plasmacytoid dendritic cells. Science 315:1398–1401.LaunchUrlAbstract/FREE Full Text↵Sancho D, et al. (2009) Identification of a dendritic cell receptor that couples sensing of necrosis to immunity. Nature 458:899–903.LaunchUrlCrossRefPubMed↵Desjardins M, Huber LA, Parton RG, Griffiths G (1994) Biogenesis of phagolysosomes proceeds through a sequential series of interactions with the enExecutecytic apparatus. J Cell Biol 124:677–688.LaunchUrlAbstract/FREE Full Text↵Lin ML, et al. (2008) Selective suicide of cross-presenting CD8+ dendritic cells by cytochrome c injection Displays functional heterogeneity within this subset. Proc Natl Acad Sci USA 105:3029–3034.LaunchUrlAbstract/FREE Full Text↵Shamu CE, Flierman D, Ploegh HL, Rapoport TA, Chau V (2001) Polyubiquitination is required for US11-dependent movement of MHC class I heavy chain from enExecuteplasmic reticulum into cytosol. Mol Biol Cell 12:2546–2555.LaunchUrlAbstract/FREE Full Text↵de Virgilio M, Weninger H, Ivessa NE (1998) Ubiquitination is required for the retro-translocation of a short-lived luminal enExecuteplasmic reticulum glycoprotein to the cytosol for degradation by the proteasome. J Biol Chem 273:9734–9743.LaunchUrlAbstract/FREE Full Text↵Young JC, Hoogenraad NJ, Hartl FU (2003) Molecular chaperones Hsp90 and Hsp70 deliver preproteins to the mitochondrial import receptor Tom70. Cell 112:41–50.LaunchUrlCrossRefPubMed↵Houlihan JL, Metzler JJ, Blum JS (2009) HSP90α and HSP90β isoforms selectively modulate MHC class II antigen presentation in B cells. J Immunol 182:7451–7458.LaunchUrlAbstract/FREE Full Text↵Koopmann JO, et al. (2000) Export of antigenic peptides from the enExecuteplasmic reticulum intersects with retrograde protein translocation through the Sec61p channel. Immunity 13:117–127.LaunchUrlCrossRefPubMed↵Wesche J, et al. (2006) FGF-1 and FGF-2 require the cytosolic chaperone Hsp90 for translocation into the cytosol and the cell nucleus. J Biol Chem 281:11405–11412.LaunchUrlAbstract/FREE Full Text↵Haug G, et al. (2003) The host cell chaperone Hsp90 is essential for translocation of the binary Clostridium botulinum C2 toxin into the cytosol. J Biol Chem 278:32266–32274.LaunchUrlAbstract/FREE Full Text↵Ratts R, et al. (2003) The cytosolic entry of diphtheria toxin catalytic Executemain requires a host cell cytosolic translocation factor complex. J Cell Biol 160:1139–1150.LaunchUrlAbstract/FREE Full Text↵Taylor M, et al. (2010) Hsp90 is required for transfer of the cholera toxin A1 subunit from the enExecuteplasmic reticulum to the cytosol. J Biol Chem 285:31261–31267.LaunchUrlAbstract/FREE Full Text↵Van Kaer L, Ashton-Rickardt PG, Ploegh HL, Tonegawa S (1992) TAP1 mutant mice are deficient in antigen presentation, surface class I molecules, and CD4-8+ T cells. Cell 71:1205–1214.LaunchUrlCrossRefPubMed↵Hogquist KA, et al. (1994) T cell receptor antagonist peptides induce positive selection. Cell 76:17–27.LaunchUrlCrossRefPubMed
Like (0) or Share (0)