|Home | About | Journals | Submit | Contact Us | Français|
Insulin degradation enzyme (IDE) is a 110-kDa zinc metalloprotease found in the cytosol of all cells. IDE degrades insulin and a variety of small proteins including amyloid-β. Recently, IDE has been proposed as the receptor for varicella-zoster virus (VZV) attachment. During our reassessment, some of the original studies were repeated and expanded in scope. We first confirmed that IDE antibody reduced VZV spread. For additional controls, we repeated the same experiments with herpes simplex virus (HSV)-infected cells as well as uninfected cells. There was a visible reduction in HSV spread but less than seen in the VZV system. Of greater importance, IDE antibody also inhibited the growth of uninfected cells. Second, we repeated the coprecipitation assays. We confirmed that antibodies to VZV gE (open reading frame 68) coprecipitated IDE and that anti-IDE antibody coprecipitated gE. However, the detected gE protein was not the mature 98-kDa form; rather, it was a precursor 73-kDa gE form found in the endoplasmic reticulum. Additional control experiments included VZV-infected cell cultures treated with tunicamycin to block gE glycosylation in the endoplasmic reticulum; again, the anti-IDE antibody coprecipitated a 73-kDa gE product. Finally, Orbitrap mass spectrometry analysis of a chromatographically purified gE sample revealed four cellular proteins associated with the unfolded protein response: BiP (HSPA5), HSPA8, HSPD1, and PPIA (peptidyl-propyl cis-trans isomerase). We conclude that IDE protease binds to the 73-kDa gE precursor and that this event occurs in the cytosol but not as a receptor/ligand interaction.
Insulin degradation enzyme (IDE) has been proposed as a virus receptor. IDE is a zinc metalloprotease that is known to degrade a number of small proteins (<6 kDa) including insulin and amyloid-β (6, 18). Because of these observations, IDE has been intensively studied as a participant in the pathogenesis of diabetes and Alzheimer's syndrome (9, 30). In 2006, IDE was reported to be a cellular receptor of varicella-zoster virus (VZV), which mediated infection and cell-to-cell spread of the virus (19). This report was based on several experiments. Initially, a truncated VZV gE protein was immobilized onto protein A-Sepharose beads with anti-gE monoclonal antibody (MAb). When the beads were incubated with cell lysates, a protein identified as IDE was attached to the complex. The authors also performed coimmunoprecipitation assays. When gE was precipitated from VZV-infected melanoma cells, the investigators observed that IDE was also detectable in the precipitate. When a rabbit polyclonal anti-IDE antibody was added to a monolayer prior to VZV infection, VZV spread was subsequently inhibited by 30 to 45%. Knockdown of IDE by small interfering RNA (siRNA) inhibited VZV infection and cell-to-cell spread. Similarly, bacitracin, a compound that inhibits IDE, also inhibited VZV infection and cell-to-cell spread. Finally, these deficiencies were corrected by expression of exogenous IDE. In a subsequent publication, the IDE binding domain in gE was localized to amino acids 32 to 71 (20). Removal of that gE domain from a recombinant virus limited cell spread (1).
In this report, we present additional data about the VZV gE-IDE interaction. We first confirmed the interaction. But we found that the interaction occurred predominantly in the cytosol with a precursor nonglycosylated gE form (73 kDa) rather than the mature gE glycoprotein (98 kDa). In fact, the gE-IDE interaction occurred when VZV-infected cells were treated with tunicamycin to prevent the biosynthesis of mature glycosylated gE. The above results suggested that the VZV gE-IDE interaction may occur either in the endoplasmic reticulum (ER) or adjacent to the ER, a site where IDE is known to interact with other proteins such as amyloid-β.
The VZV-32 strain was first isolated in Texas from a child with chickenpox and is a low-passage laboratory strain. Its genome has been completely sequenced and falls within a European clade (29). Human melanoma cells designated MeWo were the cell substrate. Cells were grown in minimal essential medium with Earle's salts (E-MEM) supplemented with 10% fetal bovine serum (FBS). When cells were nearly confluent, they were inoculated with VZV-infected cells at a ratio of one infected cell to eight uninfected cells by previously described methods (14). In a few experiments, cell-free virus was the inoculum (16). For the experiments with herpes simplex virus type 1 (HSV-1), cells were inoculated with strain F stock at 500 or 1,000 PFU, kindly provided by Richard Roller (8).
Reagents for VZV gE included murine MAb 3B3 and a guinea pig monospecific polyclonal antibody; the guinea pig polyclonal antibody to VZV gE was obtained from the serum of a guinea pig immunized twice with gE protein purified by affinity (3B3) chromatography. For HSV-1 gE, a rat polyclonal antibody was obtained from David Johnson (Oregon Health and Science University) (23); for IDE, MAb IDE and a rabbit polyclonal antibody were purchased from Covance, Inc. MAbs 6B5, 258, and 158 from this laboratory detected VZV gI, gH, and gB, respectively (14). Fluorescently tagged secondary antibodies were conjugated to the Alexa Fluor 488 and 546 fluoroprobes (Invitrogen).
Primary antibody (5 μl) was added to 100 μl of infected cell lysate (25 cm2 of infected cells yields 1 ml of lysate) and diluted with 200 μl of radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 50 mM Tris, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride [PMSF], 0.1% Nonidet P-40, 1% deoxycholate [DOC], 0.1% SDS, protease inhibitors (Sigma P8340) adjusted to pH 7.4) and then placed on a rotator at 4°C overnight. Following antibody incubation, 50 μl of Sepharose-protein A beads in protein A storage buffer (150 mM NaCl, 10 mM Tris, pH 7.4) was added to each sample, which was then placed back on the rotator at 4°C for 4 h. The samples were then washed three times (beads were sedimented at 500 rpm for 2 min, supernatant was removed by pipette, and beads were resuspended with RIPA buffer and placed back on the rotator at 4°C for 5 min). After the last wash, the beads were resuspended in 100 μl of SDS reducing sample buffer (60 mM Tris, 10% glycerol, 3% SDS, 10% β mercaptoethanol, pH 6.8), and the samples were placed in boiling water for 10 min. After cooling, the beads were sedimented at 500 rpm for 2 min, and the supernatant was removed to another tube for later use.
Antigens within viral material lysates or immunoprecipitates were prepared for immunoblotting by previously described techniques using a SuperSignal chemiluminescent kit (Pierce) (32). The protein component of the lysate was separated by SDS-PAGE using 10 or 12% acrylamide gels, transferred to a polyvinylidene difluoride (PVDF) membrane and blotted with primary antibody diluted 1:2,000 to 1:10,000 in 1% nonfat milk in phosphate-buffered saline (PBS) with 0.1% Tween 20 (T-PBS) for 1 h after being blocked for 1 h in 5% nonfat milk. After three washes with T-PBS, the membrane was incubated in secondary antibody conjugated to horseradish peroxidase diluted 1:40,000 in 1% nonfat milk in T-PBS for 1 h at room temperature (RT). After five washes with T-PBS, the membrane was soaked in a chemiluminescent solution (peroxide-luminol) from Pierce for 5 min, blotted dry, and then exposed to X-ray film for various durations from 5 s to 12 h.
Infected and uninfected control cells were fixed and permeabilized with 2% paraformaldehyde and 0.05% Triton X-100 in 0.15 M phosphate-buffered saline, washed with PBS three times for 5 min, and then incubated with 5% nonfat milk in PBS for 1 h. Subsequently, the samples were incubated with one of the primary MAbs diluted 1:1,000 to 1:2,000 in 1% nonfat milk in PBS at RT for 1 h. After three rinses for 5 min each in PBS, the samples were treated with the appropriate secondary antibody diluted 1:1,250 in 1% nonfat milk in PBS and incubated at RT for 1 h. The specimens were subsequently rinsed three times for 5 min with PBS.
Our methods for confocal microscopy have been extensively described previously (7, 28). Following immunostaining with fluorescent secondary antibody (Molecular Probes) and rinsing with PBS, the coverslips were turned over onto a drop of H-1200 Vectashield (Vector Labs) on a glass slide and then sealed in place with nail polish. After the samples hardened, they were directly viewable on a Zeiss 510 confocal microscope.
Generation of a VZV gE-specific affinity chromatography column followed our previously described method (10). Briefly, 20 mg of MAb 3B3 was added to 10 ml of cyanogen bromide (CnBr)-activated Sepharose 4B beads (1.2 g of dry beads from Sigma swollen with 10 ml of 0.1 mM HCl and then activated with high-pH borate buffer [0.1 M sodium borate, 0.5 M NaCl, pH 8.3]) and rotated for 1 h at RT. The reaction was quenched with 1 M Tris (pH 8.0), beads were sedimented, and supernatant was removed. The beads were then incubated with 0.1 M Tris, pH 8.0, overnight at 4°C and then washed twice with the following (in order): 25 ml of buffer A (50 mM NaCl, 2 mM EDTA, 5 mM NaN3, 10% glycerol, pH 11), 25 ml of PBS, 25 ml of buffer B (0.5 M NaCl, 0.1 M sodium acetate, pH 4.0), and finally 25 ml of PBS. Monolayers of infected cells (1,350 cm2) were lysed with 35 ml of RIPA buffer and dislodged into PBS yielding 75 ml of lysate. The 3B3-linked Sepharose 4B beads were added to 75 ml of lysate and rotated overnight at 4°C. The beads were sedimented, supernatant was removed and then transferred back into a 10-ml column, and samples were washed with 75 ml of PBS augmented with 0.5% NP-40 to remove any remaining amount of unbound lysate. VZV gE protein was eluted into 2-ml fractions, using 7.7 mg of the peptide DQRQYGDVFKGD, corresponding to the gE 3B3 epitope, dissolved in 25 ml of PBS.
For our analyses, 10 μg per replicate of sample was reduced using 10 mM dithiothreitol (DTT) for 1 h at 56°C and alkylated with 55 mM iodoacetamide for 45 min at ambient temperature. After that, the sample was diluted 1:5, vol/vol, in 100 mM ammonium bicarbonate and digested with 2% trypsin (sequence grade modified; Promega) for 16 h at 37°C at pH 8.0. The reaction was quenched through acidification with 2% trifluoroacetic acid (Fluka). The resulting peptide mixture was desalted using reverse-phase C18 stop and go extraction (STAGE) tips and diluted in 0.1% trifluoroacetic acid prior to nano-high-performance liquid chromatography-mass spectrometry (HPLC-MS) analysis (31). Prior to injection in the spectrometer, the samples were separated by liquid chromatography using an Acclaim PepMap 100 column (C18, 3-μm particle size, and 100-Å pore size; Dionex Corp.) with a capillary bed length of 12 cm, an internal diameter of 100 μm, and self-packed with Reprosil-Pur C18-aq (Maisch GmbH). The gradient was 7% solvent B to 40% solvent B in 87 min, followed by 40 to 80% solvent B in 8 min. Solvent A was aqueous 2% acetonitrile in 0.1% formic acid, whereas solvent B was aqueous 90% acetonitrile in 0.1% formic acid.
All experiments were performed on a Dionex Ultimate 3000 nano-LC system connected to a linear quadrupole ion trap (LTQ)-Orbitrap mass spectrometer (ThermoElectron) equipped with a nanoelectrospray ion source (34). The mass spectrometer was operated in data-dependent mode to automatically switch between Orbitrap-MS and LTQ-tandem MS (MS/MS) acquisition. Full-scan MS spectra (from m/z 300 to 2,000) were acquired in the Orbitrap with resolution (R) of 60,000 at m/z 400 (after accumulation to a target of 1,000,000 charges in the LTQ). For accurate mass measurements the “lock mass” option was enabled in MS mode, and polydimethylcyclosiloxane (PCM) ions generated in the electrospray process from ambient air were used for internal recalibration during the analysis (27). Target ions already selected for MS/MS were dynamically excluded for 60 s. General mass spectrometry was conducted with an electrospray voltage of 1.5 kV, no sheath, and auxiliary gas flow. The ion selection threshold was 500 counts for MS/MS, and an activation Q-value of 0.25 and activation time of 30 ms were also applied for MS/MS.
MS/MS peak lists from individual RAW files were generated using the DTA SuperCharger package, version 1.29, available in the MSQuant validation tool. We used Mascot Daemon for multiple search submissions on a local Mascot server, version 2.1 (Matrix Science) (4, 5). Under our criteria, Mascot indicated a minimal score of 31 for a P value of ≤ 0.05. All data had a mass accuracy average of 3.1 ppm.
One of the key observations in the original report was that incubating VZV-infected melanoma cells with an IDE antibody at 1:100 inhibited VZV cell spread (19). We repeated this experiment with a cell-free virus inoculum and found that VZV cell spread was markedly reduced by the IDE antibody compared with control medium or medium containing a 1:100 concentration of preimmune rabbit serum (Fig. (Fig.1).1). VZV-infected cells incubated with IDE antibody showed virtually no cytopathic effect (Fig. 1E and F) while widespread VZV gE was evident in infected cells incubated in control medium or medium containing preimmune rabbit serum (Fig. 1A to D). We next repeated the same experiment looking at HSV-1 cell spread. Specifically, melanoma cells in 24-well plates were infected with HSV-1 at 500 PFU or 1,000 PFU and incubated in either 1:100 IDE antibody or 1:100 preimmune rabbit serum for 24 h; cells were then fixed, stained with crystal violet, and examined with an optical microscope (Fig. (Fig.2).2). Clearly, the well containing cells incubated with the IDE antibody showed fewer plaques of smaller size than wells containing cells incubated in normal medium or medium with 1:100 preimmune rabbit serum (Fig. (Fig.2A).2A). Sizes of each plaque were measured with NIH ImageJ software (http://rsb.info.nih.gov/nih-image/). The sizes of all plaques were then summed to assess the extent of HSV cell spread. The well incubated in IDE antibody showed 60% less plaque area than the well incubated with normal rabbit serum in infections at either 500 PFU or 1,000 PFU (Fig. (Fig.2B).2B). The latter result suggested that incubation with IDE antibody may diminish the ability of melanoma cells to host viral infection.
To pursue the hypothesis that incubation with the IDE antibody may be directly affecting the metabolism of the melanoma cells used to host viral infection in cell culture, we investigated the multiplication of MeWo cells incubated with the IDE antibody. Specifically, we split MeWo cells in half and then allowed them to grow to confluence over 72 h while being incubated with control medium or medium treated with a 1:100 dilution of a rabbit polyclonal antibody to IDE or a MAb to IDE. Triplicate samples were fixed at 48 and 72 h postplating and then stained with a nuclear stain. At least 10 images at a magnification of ×200 were taken at random locations within each well, resulting in almost 600 images, each containing 300 to 600 nuclei. Cellular density was then calculated by enumerating the number of nuclei in each image. The density of cells in wells incubated with IDE antibodies was about 20% less than the density in the control wells by 72 h (Fig. (Fig.3).3). This result supported the hypothesis that incubation of melanoma cells with IDE antibody diminished the ability of the cells to grow and likely diminished their ability to host viral infection.
Another key observation in the original report was that IDE coprecipitated with VZV gE but not with other VZV glycoproteins (19). We set out in an initial series of experiments to confirm this observation by screening VZV glycoproteins gE, gI, gH, and gB for binding to IDE by precipitating the glycoproteins with their respective monoclonal antibodies and then blotting the precipitates for IDE using the rabbit polyclonal antibody. Only gE and, to a minor extent, gI coprecipitated IDE (Fig. (Fig.4).4). It is likely that the minor extent of gI precipitation was due to the complex gI forms with gE (12). To further refine this observation, we precipitated IDE from VZV-infected cell lysates using two different gE antibodies, the same MAb 3B3 used in the experiment shown in Fig. Fig.11 or a guinea pig-monospecific polyclonal antibody, and again blotted for IDE. Both antibodies coprecipitated IDE (Fig. (Fig.5).5). Confirming the earlier report, IDE and gE bound together in infected cell lysates. Since the gE-IDE interaction may not occur at the cell surface, based on data shown in Fig. Fig.3,3, we postulated that IDE may attach to a partially processed gE form in the cytosol.
To consider this question and also investigate the specificity of the interaction, we carried out coprecipitation experiments between IDE and gE from both VZV and HSV-1. First, precipitating with antibodies against HSV gE and then blotting for IDE showed that HSV-1 gE did not coprecipitate IDE to the same extent as VZV gE (Fig. (Fig.6B,6B, lanes 1 and 3). This result was confirmed in the reverse coprecipitation experiment, i.e., precipitating with an IDE antibody and blotting for gE. The blot showed that HSV-1 gE was not coprecipitated with IDE. Surprisingly, the dominant form of coprecipitated VZV gE had a lower molecular weight (MW) than the mature form (Fig. (Fig.6A,6A, lanes 3 and 4). The antibody used in the blot for VZV gE was MAb 3B3. Its epitope has been mapped between residues 150 to 162 on the gE ectodomain (32). Given the surprising result, this experiment was repeated six times with the same result. The result was the same with the high-titered guinea pig polyclonal monospecific gE antiserum used in the experiment shown in Fig. Fig.5.5. In all experiments, the precursor 73-kDa gE form was the most prominent precipitated protein.
We have used MAb 3B3 in numerous VZV gE experiments published over the past 25 years, and this reagent has invariably immunoblotted all gE forms present in an infected culture (11). Because the result shown in Fig. Fig.66 was completely unexpected, we tested whether the VZV gE profile would be different if we varied the salt concentration in our precipitation reaction. To this end, we carried out the precipitation of IDE from infected cell lysates using buffers with twice or half the amount of normal (isotonic) salt, after which we again blotted for VZV gE. We found that high salt abolished the coprecipitation completely while low salt accentuated the interaction between the 73-kDa form of gE and IDE; in addition, at lower salt concentrations, the higher-molecular-weight forms of gE were precipitated in small amounts (Fig. (Fig.7).7). This finding is relevant because prior investigators used a low concentration of salt (10 times lower than normal) in their immunoprecipitation buffer (19). Our results suggested that higher-molecular-mass forms of gE, including the mature 98-kDa form, would coprecipitate with IDE at a very low salt concentration. Nevertheless, the 73-kDa form was still the most prominent.
In an earlier characterization of gE maturation, we analyzed gE in infected cultures treated with tunicamycin (24). Tunicamycin is a mixture of antibiotics which inhibits formation of polyisoprenyl N-acetylglucosaminyl pyrophosphates (38). In other words, tunicamycin blocks the formation and attachment of mature N-linked glycans onto a polypeptide backbone. We had previously documented the prominence of a 73-kDa gE protein in cultures treated with tunicamycin. Therefore, we postulated that IDE preferentially bound to the same precursor gE form as that seen in the coprecipitation experiments presented above. To test this hypothesis, we again treated VZV-infected cultures with tunicamycin under the same conditions as previously described and repeated the coprecipitation experiments exactly as described above. When the tunicamycin-treated extract was precipitated with IDE antibody, subjected to SDS-PAGE, and immunoblotted with anti-gE antibody, the precursor 73-kDa gE form was detected (Fig. (Fig.8A).8A). Furthermore, if the dominant 73-kDa form of gE in the tunicamycin-treated infected cell lysate was precipitated with a gE antibody and then blotted for IDE, IDE was coprecipitated with the immature form of gE (Fig. (Fig.8B8B).
In our prior gE biochemical studies with tunicamycin, we had never performed an imaging analysis of gE trafficking after tunicamycin treatment. To complete this aspect of our earlier research, we performed a confocal microscopy survey of gE trafficking with and without tunicamycin. Tunicamycin was added 5 h after infection, and the cultures were incubated for another 24 h. Then the monolayers were fixed, immunolabeled, and imaged by confocal microscopy (Fig. (Fig.9).9). The results clearly demonstrated that gE, in the presence of tunicamycin, was blocked in its transit out of the ER en route to the Golgi compartment (Fig. (Fig.9A).9A). In contrast, gE trafficking through the cytoplasm to the plasma membrane was easily detected in the nontreated cultures (Fig. (Fig.9B).9B). These images were in complete agreement with our prior biochemical analyses. We also observed during these imaging experiments that murine MAb 3B3 immunostained gE trapped within the ER more intensely than the guinea pig anti-gE antibody (compare Fig. 9C and D). This conclusion was reached after observing several Z-stacks of images (data not shown); Fig. Fig.99 represents one slice of each Z-stack. In summary, IDE coprecipitated with an immature form of VZV gE that was largely localized in the ER.
The fact that IDE interacted predominantly with the gE 73-kDa form found in the ER led to a further hypothesis. Namely, the gE-IDE interaction occurred in the cytoplasm at a site in or adjacent to the ER, possibly secondary to ER stress (39). We have previously documented that VZV glycoproteins are produced in such abundance in infected cultures and that biosynthesis of cellular glycoproteins is no longer detectable with radio sugar labeling (13). The overabundant production of glycoproteins within a cell is a well-known precipitant of ER stress (25).
To test for evidence of ER stress, we first purified the gE glycoprotein by immunoabsorbent chromatography and then analyzed this fraction by Orbitrap mass spectrometry (21, 34). This high-accuracy instrument uses a variation of ion cyclotron technology whereby the fragmented ions spiral down the length of the magnet (4, 5). The radius of the spiral is related to the mass/charge ratio. Ion cyclotron mass spectrometers exhibit very high mass discrimination due to the large number of orbits sampled (22). In addition to the detection of the gE protein and its complex partner gI, a number of other viral and cellular proteins were detected in the purified fraction (Table (Table1).1). Interestingly, IDE was not detected. Included in the set of detected cellular proteins were four members of the unfolded protein response (UPR) (2, 25). These included heat shock proteins HSPA5 (BiP), HSPA8, and HSPD1. The fourth protein was PPIA (peptidyl-propyl cis-trans isomer catalyst). The last results suggested that misfolded VZV gE was a candidate protein for ER retrotranslocation.
We have reexamined the previously described interaction of VZV gE and IDE and concur that IDE binds a gE product and that treatment of VZV-infected cells with an IDE antibody greatly inhibits cell-to-cell spread of the virus (19). However, we found that IDE primarily bound the 73-kDa unglycosylated gE form and that incubation with IDE antibodies also inhibited cell-to-cell spread of HSV-1 as well as the growth of uninfected melanoma cells. These observations have led us to hypothesize that IDE and the precursor gE form interact in the cytosol, most likely as a consequence of ER stress, and that the observation that IDE antibody inhibits VZV cell-to-cell spread is a separate and unrelated phenomenon. We suggest that treatment with IDE antibodies inhibits the function of IDE within the cell. The consequences of this inhibition are difficult to predict. Shii et al. found that microinjection of a monoclonal antibody to IDE into hepatoma cells inhibited insulin degradation by an average of 39% (35). More recently, Chou et al. found that IDE interacts with the intermediate filament proteins vimentin and nestin (3). In turn, vimentin and nestin can either activate or suppress the proteolytic IDE activity, depending upon substrate. Thus, any antibody-mediated inhibition of IDE activity would have unpredictable consequences on cytoskeletal remodeling and cellular division. In our case, it appears that IDE antibody exerted a modest (~20%) inhibition of melanoma cell proliferation.
In prior publications, we have characterized the biosynthesis of the predominant VZV glycoprotein gE in considerable detail (12, 24). Like many other glycoproteins, VZV gE has several distinguishable forms as the protein is synthesized by the polyribosomes and transported into the ER and subsequently through the Golgi stacks into the trans-Golgi network (TGN) (Fig. (Fig.10).10). These include a precursor form lacking N-linked oligosaccharides (73 kDa), a form with high-mannose N-linked and O-linked glycans (88 kDa), and a mature form containing both O-linked and N-linked complex-type glycans (98 kDa). In the course of the present investigation, we found that IDE bound the initial nonglycosylated precursor 73-kDa form of gE. Two questions arose from this observation: (i) Why did the other investigations of IDE and VZV interactions fail to detect the precursor form of gE, and (ii) where in the course of the infectious cycle would endogenous IDE encounter the precursor form of gE? Herein we consider these questions in turn.
Other investigations may not have detected the precursor form of gE interacting with IDE for several reasons: first, the interaction is very sensitive to salt concentration; second, different antibodies may not bind all forms of gE equally well; finally, small differences in conditions of SDS-PAGE may lead to an incomplete separation of the large gE/gI complex, with an apparent MW higher than expected (12). We showed that the coprecipitation of VZV gE forms was sensitive to salt concentration in the precipitation buffer. High salt eliminated the coprecipitation totally while low salt led to the coprecipitation of higher-molecular-mass forms of gE in addition to the 73-kDa form. Likely, most forms of gE would coprecipitate at the much lower salt concentration (15 mM) used in prior studies (19). Thus, the binding of the 73-kDa form of gE with IDE would be obscured.
With regard to antibody specificity, we have produced a large library of VZV antibodies that recognize the different forms of VZV gE. The best-characterized antibody against gE is MAb 3B3 (15). This is the only antibody reagent with a fully determined epitope, namely, gE amino acids 150 to 161 (32). The epitope is linear and extremely resistant to blockage or reducing conditions in that the antibody will attach even in the presence of 1% SDS (24). Since the epitope does not require secondary or tertiary structure, the antibody recognizes all gE forms from 73 kDa to 98 kDa (Fig. (Fig.10).10). Furthermore, the association and dissociation constants for the 3B3 antibody have been calculated and are similar to most antibody/epitope affinities (17). We also used a high-titered guinea pig monospecific polyclonal anti-gE antibody in this study. We have observed that the guinea pig anti-gE antibody attached less well to the 73-kDa gE form. Therefore, it is highly likely that other commercial antibodies to VZV gE also attach less well to the 73-kDa form. In the case of glycoproteins, many antibodies recognize conformational epitopes that require at least partial glycosylation. We suggest that differences between antibody reagents have been, in part, responsible for the differing results from prior reports.
The tunicamycin experiments were extremely informative in our validation of the important role of the 73-kDa gE form in the gE-IDE interaction. The easiest method by which to obtain the precursor gE is to grow the virus in the presence of tunicamycin, an inhibitor of complex-type glycosylation in the ER (11, 12). When coprecipitation was carried out with IDE antibody and tunicamycin-treated infected cell lysates, the gE precursor form (73 kDa) was identified (Fig. (Fig.10).10). This is the same form that was preferentially coprecipitated in untreated VZV-infected cells. The 73-kDa gE form is found in the ER in treated infected cells (11). Thus, the coprecipitation reaction occurred between IDE and a gE form found only in the ER. It is highly likely that the site of interaction on gE is the domain previously identified (1). This site is not located near a glycan attachment site and would be available on the precursor form or a glycosylated form (11). In summary, even though smaller amounts of the higher-MW gE forms are also coprecipitated by IDE, we postulate based on the data in this report that IDE primarily interacts with the ER form of gE.
Once we discovered this association with precursor gE, we postulated that the interaction may be related to the cellular pathway known as ER-associated degradation (ERAD). ERAD is a well-established mechanism by which misfolded proteins are retrotranslocated from the ER to the cytosol for degradation (39). Of interest, treatment of cells with tunicamycin can enhance ERAD of misfolded precursor glycoproteins, presumably because tunicamycin blockage results in their accumulation and subsequent ER stress, leading in turn to ERAD (39). Of further interest, ERAD substrates must remain in solution in order for them to be retrotranslocated (26). Molecular chaperones assist in the transportation; one of these is the heat shock protein HSPA5 or BiP (37). As shown in Results, we have detected BiP in a mass spectrometric analysis of purified gE samples. In the same experiment we have also identified two other heat shock proteins (HSPA8 and HSPD1) and a fourth protein (PPIA) commonly associated with ERAD. This result provides further evidence in support of the gE ERAD hypothesis. Further, we have documented that autophagy occurs in VZV-infected cells (36). ER stress is often a precipitating factor for subsequent autophagy (25).
Finally, there is an example of an ERAD substrate that is bound by IDE after translocation from the ER to cytosol (33). That protein is amyloid-β peptide. During Alzheimer's disease, amyloid-β protein is generated by proteolysis of amyloid precursor protein. Some amyloid-β peptide is retrotranslocated by both a proteasome-dependent and a proteasome-independent pathway. The cytosolic enzyme that degrades amyloid-β peptide in the latter pathway is IDE. However, since IDE usually cleaves only small peptides such as insulin and amyloid, any enzymatic effect of IDE on precursor gE as a substrate remains to be determined.
We thank Richard Roller (Iowa) and David Johnson (Oregon) for HSV-1 reagents, as well as valuable advice about carrying out the HSV experiments.
This research was supported in part by NIH grant AI53846.
Published ahead of print on 28 October 2009.