![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 19, Issue 2, 572-586, February 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1-Antitrypsin into Inclusion Bodies Is a Cell-protective Mechanism to Maintain Endoplasmic Reticulum Function
*Department of Biochemistry and Molecular Biology and Department of Physiology and Biophysics, University of Arkansas for Medical Sciences, Little Rock, AR 72205; and Dipartimento Universitario Clinico di Biomedicina, Universita' degli Studi di Trieste, Trieste I-34138, Italy
Submitted June 21, 2007;
Revised October 18, 2007;
Accepted November 16, 2007
Monitoring Editor: Thomas Sommer
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1-antitrypsin with E342K mutation has a high tendency to form intracellular polymers, and it is associated with liver disease. In the hepatocytes of individuals carrying the mutation, 1-antitrypsin localizes both to the endoplasmic reticulum (ER) and to membrane-surrounded inclusion bodies (IBs). It is unclear whether the IBs contribute to cell toxicity or whether they are protective to the cell. We found that in hepatoma cells, mutated 1-antitrypsin exited the ER and accumulated in IBs that were negative for autophagosomal and lysosomal markers, and contained several ER components, but not calnexin. Mutated 1-antitrypsin induced IBs also in neuroendocrine cells, showing that formation of these organelles is not cell type specific. In the presence of IBs, ER function was largely maintained. Increased levels of calnexin, but not of protein disulfide isomerase, inhibited formation of IBs and lead to retention of mutated 1-antitrypsin in the ER. In hepatoma cells, shift of mutated 1-antitrypsin localization to the ER by calnexin overexpression lead to cell shrinkage, ER stress, and impairment of the secretory pathway at the ER level. We conclude that segregation of mutated 1-antitrypsin from the ER to the IBs is a protective cell response to maintain a functional secretory pathway. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1-antitrypsin (AAT) and neuroserpin, provide well-studied examples of how small changes in protein conformation lead to cell toxicity (Kopito and Ron, 2000
; Perlmutter, 2002b; Carrell, 2005; Lomas, 2005). A variant 1-antitrypsin with E342K mutation (ATZ) has greatly increased tendency to form homodimers and higher order polymers compared with AAT (Huntington et al., 1999; Lomas et al., 2005). PiZZ individuals (homozygotes for ATZ; Perlmutter, 2002b) have reduced release of the serpin in the blood, which leads to an unbalanced activity of neutrophil proteases in the lung and damage to the alveolar tissue. Some PiZZ individuals have also liver disease due to intracellular retention of the mutated protein (Perlmutter, 2002a). Early ultrastructural studies of hepatic biopsies from children with the PiZZ mutation show accumulation of ATZ in dilated sections of the ER, indicating that the misfolded protein accumulates in this compartment (Feldmann et al., 1974; Yunis et al., 1976; Hultcrantz and Mengarelli, 1984). More recently, it has been found that ATZ also accumulates in membrane-surrounded inclusion bodies (IBs) that have the morphological appearance of autophagosomes (Teckman and Perlmutter, 2000). However, in livers from transgenic mice that express both ATZ and the autophagosome marker LC3-green fluorescent protein (GFP), the marker did not colocalize with the ATZ containing IBs (Kamimoto et al., 2006). In contrast, in cells where autophagic activity was inhibited, there was increased number of ATZ-containing IBs (Kamimoto et al., 2006). Therefore, although it seems that the autophagic response is important for ATZ disposal, it is not known to which extent ATZ-containing IBs correspond to autophagosomes. In addition to ATZ, aggregation of other proteins in the secretory pathway leads to formation of membrane-surrounded IBs that are separate from the ER, from pre-Golgi intermediates, or both. Examples are the Russell bodies formed in response to expression of mutated immunoglobulins in myeloma cells, IBs in response to accumulation of misfolded proinsulin in the β-cells of the Akita mouse, and myelin-like figures in Schwann cells in response to expression of mutated peripheral myelin protein-22 (Dickson et al., 2002; Zuber et al., 2004; Mattioli et al., 2006). It is not clear whether formation of the IBs is a cause of cell toxicity or a cell-protective mechanism. This issue is currently being investigated with respect to the IBs that originate in the cytoplasm from sequestration of aggregated proteins, such as aggresomes containing 
F508 mutant form of cystic fibrosis transmembrane conductance regulator and huntingtin filaments. The aggresomes are different from the ATZ-containing IBs or the Russell bodies because they are not membrane-surrounded (Kopito and Sitia, 2000). With respect to the aggresomes, it seems that cell toxicity actually precedes the formation of IBs (Bennett et al., 2005; Diaz-Hernandez et al., 2006) and that their formation reduces neuronal death (Arrasate et al., 2004). However, the question as to whether membrane-surrounded IBs, such as those containing ATZ, are themselves toxic or, rather, protective to the cell has not yet being addressed.
In the ER, ATZ interacts with a variety of ER chaperones, including calnexin, protein-disulfide isomerase (PDI), GRP-78/BiP, and with the N-glycan modifying enzymes UDP-glucose:glycoprotein glucosyltransferase and -mannosidase-like protein (Cabral et al., 2002; Oda et al., 2003; Schmidt and Perlmutter, 2005; Papp et al., 2006). Binding and release of ATZ from calnexin are necessary for ATZ degradation (Cabral et al., 2002; Oda et al., 2003; Sifers, 2004). Misfolded proteins are then retrotranslocated from the ER lumen to the cytosol and degraded by the ubiquitin-proteasome system known as ER-associated degradation (ERAD) (Meusser et al., 2005; Ruddock and Molinari, 2006). Here, we have defined the IBs as sites of ATZ localization segregated from the main ER. Whether calnexin or other chaperones function to control ATZ transit to the IBs and whether formation of ATZ-containing IBs is protective to the cell are major questions addressed in this article.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell Culture and Transfection
Mouse hepatoma Hepa 1-6, human hepatoma Huh-7, and mouse neuroblastoma Neuro2A (N2A) cell lines were cultured in DMEM with 10% bovine serum and 5% penicillin/streptomycin. Cells were transiently transfected with the indicated plasmids by using Lipofectamine 2000 according to the manufacturer's instructions.
cDNAs and Constructs
The enhanced green fluorescent protein (EGFP) expression vector pEGFP-N2 was purchased from Clontech (Mountain View, CA). The cDNA for human AAT, neuroserpin, and proinsulin were obtained from Origene Technologies (Rockville, MD). PDI-pCMV-SPORT6 was purchased from the American Type Culture Collection (Manassas, VA). Construction of the plasmids Myc-AAT-pcDNA3.1, HA-ATZ-pcDNA3.1, Myc-ATZ-pcDNA3.1, ATZ-GFP, neuroserpin Portland-pcDNA3.1, proinsulin-pcDNA3.1, calnexin-pcDNA3.1, and calnexin-GFP are described in Supplemental Material. Lysosomal-associated membrane protein 1 (LAMP1)-GFP was a kind gift from Dr. J. Lippincott-Schwartz and LC3-GFP was a kind gift from Dr. T. Yoshimori.
Immunoprecipitations
Transfected cells grown in 100-mm dishes were scraped in immunoprecipitation (IP) buffer (100 mM Tris-HCl, pH 7.4, 50 mM NaCl, 1% NP-40, and protease inhibitors [Complete Mini, 1 µM pepstatin, and 1 mM PMSF]) and incubated for 30 min at 4°C. Cell lysates were rotated for 1 h at 4°C with the indicated antibody. After addition of protein G beads, the samples were further incubated for 1 h at 4°C. After four washes with IP buffer and one wash with a high salt buffer (150 mM Tris, pH 7.4, 400 mM NaCl, and 1% NP-40 at pH 8.0), the beads were resuspended in sample buffer, boiled for 5 min, and centrifuged. Supernatants and cell extracts were loaded onto a 7% SDS-polyacrylamide gel electrophoresis (PAGE) gel.
Preparation of Cell Lysates
To measure the amount of cell ATZ, transiently transfected Hepa 1-6 cells were grown in 30-mm dishes, and, at the indicated times, cells were washed once in Kglu buffer (20 mM HEPES, pH 7.4, 120 mM potassium glutamate, 20 mM potassium acetate, 5 mM EGTA, and 1 mg/ml bovine serum albumin) and scraped from plates in Kglu buffer containing 1% Triton X-100 and proteases inhibitors (Complete Mini, 1 µM pepstatin, and 1 mM PMSF). Samples were passed five times through a 27-gauge1/2 needle and incubated 30 min at 4°C. Cell lysates were centrifuged at 7200 x g for 10 min, and the pellet was discarded. Cell lysates were electrophoresed on a 7% SDS-PAGE gel by using loading buffer with or without SDS and β-mercaptoethanol (denaturing and nondenaturing conditions, respectively).
Immunoblotting
Separated proteins were transferred to a nitrocellulose membrane probed with the indicated primary antibodies and secondary POD-conjugated antibodies. Enhanced chemiluminescence detention, densitometry, and protein determination were performed as described previously (Wang et al., 1997).
Sucrose Gradient Centrifugation
Transiently transfected Hepa 1-6 were grown in 100-mm dishes, washed once in Kglu buffer, and scraped from plates in Kglu buffer containing proteases inhibitors. Samples were passed 10 times through a 27-gauge1/2 needle. Postnuclear supernatant (PNS) was obtained by centrifugation at 100 x g for 5 min. The pellet obtained by further centrifugation of the PNS at 1000 x g for 10 min was resuspended in Kglu buffer containing protease inhibitors and loaded on the top of a sucrose density gradient (30–50%, wt/vol). The gradient was centrifuged for 2 h at 45,000 rpm in a Beckman Optima TLX ultracentrifuge (TLS-55 swinging rotor), and 19 fractions were collected. The fractions obtained were electrophoresed on a 7% SDS-PAGE gel by using loading buffer with SDS and β-mercaptoethanol.
Cell Sorting and Electron Microscopy
Hepa 1-6 cells plated in 100-mm dishes were transiently transfected with ATZ-GFP or pcDNA3.1 (mock-transfected cells). Forty-eight hours after transfection, cells were trypsinized and centrifuged at 300 x g for 3 min. The pellet was resuspended in a filter-sterilized sorting buffer (phosphate-buffered saline [PBS], 25 mM HEPES, pH 7.0, 2% fetal bovine serum [FBS], and 1 mM EDTA). Cells transiently expressing ATZ-GFP were sorted by flow cytometry using a FACSAria instrument (BC Biosciences, San Jose, CA). The brightest population of GFP-positive cells (top 30%) was recovered for the electron microscopy in 10 ml of complete growth medium and centrifuged for 10 min at 1500 x g. The pellets were rinsed with PBS and fixed with 2% glutaraldehyde in phosphate buffer, pH 7.4, for 1 h at room temperature. Cells were rinsed twice with PBS containing 0.01% sodium azide and processed for electron microscopy as described previously (Baldini et al., 1998).
Control cells were kept in complete medium (DMEM, 10% FBS, and 5% penicillin/streptomycin) during the sorting process (45 min).
Fluorescence Microscopy
Immunofluorescence staining was performed as described previously (Mohammad et al., 2007). Epifluorescence and confocal fluorescence images were captured with a CARV I spinning confocal imaging system (BioVision Technologies, Exton, PA) attached to an X-71 fluorescence microscope (Olympus, Tokyo, Japan). Images were collected using a CoolSNAP HQ camera (Photometrics, Tucson, AZ), and processed by using the IPLab software (Scanalytics, Fairfax, VA). For some experiments, ATZ-GFP distribution in Hepa 1-6 cells was visualized by total internal reflection microscopy (TIRFM) by using an X-71 inverted microscope equipped with an argon laser (488 nm). For these experiments, cells in DMEM were visualized 48 h after transfection. Images were collected using a CoolSNAP HQ camera and processed using MetaMorph software (Molecular Devices, Sunnyvale, CA).
Autophagy Assays
Before analysis, Hepa 1-6 cells were starved for 6 h in Hanks' balanced salt solution (HBSS) (HBSS, starvation medium) or maintained in normal growth media (Pattingre et al., 2005). Autophagy was inhibited by preincubation of the cells with 10 mM 3-methyladenine for 30 min before amino acid deprivation. To quantify the level of autophagy, the number of GFP-LC3 dots per cell was determined by manual counting.
Colocalization of ATZ with Calnexin
For the quantification of calnexin expression, fluorescence intensity was measured by manually tracing the outline of individual cells and by measuring the pixel intensity of the selected area by using the IPLab software. Confocal fluorescent images of the cells were analyzed for colocalization of ATZ with calnexin by using IPLab. Briefly, after autosegmentation of calnexin fluorescence (green channel) and Myc-ATZ fluorescence (red channel), the regions of interest were analyzed by measuring the Mander's overlap coefficient M1 (Manders et al., 1992). M1 is defined as the ratio of the summed intensities of pixels from the green image for which the intensity of the red is above zero. In our experimental conditions, M1 indicates pixel intensity overlap from channel 1 (green, calnexin) to channel 2 (red, Myc-ATZ).
Quantification of Hepa 1-6 Cell Area
To quantify the area of the cells, epifluorescence images were obtained with constant parameters of acquisition. The region of interest (ROI) was selected by manually drawing the cell margins after the F-actin staining for each individual cell. The intensity of the fluorescence in the different channels and the area of the selected ROIs were measured by using IPLab software. The amount of ATZ in each positive cell was calculated by subtracting the background (pixel intensity of the neighboring nontransfected cells) from the total pixel intensity of ATZ-GFP.
Quantification of the Increase of the Unfolded Protein Response (UPR) by Immunofluorescence
Epifluorescence images were obtained with constant parameters of acquisition. The ROI was selected by manually drawing the cell margins following KDEL staining for each individual cell. Fluorescence pixel intensity in the two different channels (blue, calnexin; red, KDEL) was measured by using IPLab software. As a positive control, UPR was induced by incubating the cells with 1 µg/ml tunicamycin for 16 h.
Quantification of Increase of Intracellular ATZ versus Overexpression of Calnexin and PDI
Epifluorescence images were obtained with constant parameters of acquisition. The ROI was selected by manually drawing the cell margins after calnexin or PDI staining. IPLab software was used to calculate the intensity of the fluorescence in the different channels (green, PDI or calnexin; red, Myc-ATZ) of the selected ROI. First, we measured the average pixel intensity of intracellular ATZ in cells expressing Myc-ATZ alone (n = 30). The -fold increase of intracellular ATZ was calculated by the following ratio: pixel intensity of Myc-ATZ (red channel) coexpressed with either exogenous calnexin or PDI/pixel intensity of Myc-ATZ in cells expressing ATZ alone. Calnexin and PDI overexpression was calculated by the following ratio: pixel intensity of calnexin or PDI in cotransfected cells/pixel intensity calnexin or PDI of the neighboring nontransfected cells (n = 3 per transfected cell).
Quantification of Area of ATZ-containing Compartment
After autosegmentation of Myc-ATZ fluorescence (red channel) in confocal images, the area of the regions of interest was measured using IPLab software.
Colocalization of ATZ with Insulin
Confocal fluorescent images were retrieved, and colocalization of ATZ with insulin was measured by using the Mander's overlap coefficient M1 (IPLab software) (Manders et al., 1992). For these experiments, the colocalization coefficient M1 indicates pixel intensity overlap from channel 1 (red, Myc-ATZ) to channel 2 (green, insulin).
Quantification of Neurite Outgrowth
To measure neurite outgrowth, cells were serum starved for 24 h. Cell F-actin staining visualized by epifluorescence microscopy was used to classify the cells into two categories: cells with neurites or cells without neurites. We considered neurites the processes that were at least half the length of the cell body. The length of the neurite was analyzed by using IPLab software.
Nocodazole Treatment
Hepa 1-6 cells were transiently transfected with ATZ-GFP, and 48 h after transfection cells, were incubated for 6 h in complete culture medium with 80 µM nocodazole or vehicle (dimethyl sulfoxide) at 37°C. After the incubation period, cells were washed, fixed, and processed for immunofluorescence.
Insulin and Proinsulin Measurements
To measure proinsulin release in the medium, transfected Hepa 1-6 cells grown in 30-mm dishes were washed with M2 medium (20 mM HEPES, pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, and 0.1 mg/ml bovine serum albumin) and incubated at 37°C for the indicated times in 1 ml of the same medium with 5 mM glucose. The medium was collected and centrifuged at 300 x g for 3 min to remove cell debris. To measure proinsulin and insulin in the cell, transiently transfected Hepa 1-6 and N2A cells were grown in 30-mm dishes, and, at the indicated times, cells were washed once in Kglu buffer and scraped from plates in Kglu buffer containing 1% Triton X-100 and proteases inhibitors. Cells were homogenized by passing them five times through a needle (27-gauge1/2), and then they were incubated for 30 min at 4°C. Cell extracts were obtained by centrifugation at 7200 x g for 10 min. All the measurements were done 48 h after transfection. Insulin and proinsulin levels in cell extracts and cell-free medium were measured using the human insulin and proinsulin enzyme-linked immunosorbent assay (ELISA) kit from Linco Research (St. Charles, MO) according to the manufacturer's instructions.
Statistical Analysis
All experiments were done at least twice, and data are expressed as mean ± SD from a single experiment unless noted otherwise. After checking that the measures realized were normally distributed, two-tailed Student's t tests were performed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The results shown here indicate that although this is the case for a population of transfected cells that had ATZ in a reticular pattern, ATZ segregates in a time-dependent manner to IBs that showed up as different organelles from the main calnexin-containing ER. To determine whether the antibody used in this study labels the ER, cells were transfected with calnexin-GFP. The distribution of GFP fluorescence and calnexin immunostaining was similar (Figure 1C). This shows that the calnexin immunostaining of Figure 1A reflects the actual distribution of the protein.
|
). When transiently transfected Hepa 1-6 cells were observed by TIRFM, ATZ showed up as a reticular pattern (Figure 1D, bottom) in addition to the brightly fluorescent IBs (Figure 1D, top, red asterisk). In addition, small, round accumulations of ATZ (Figure 1D, bottom, green arrowheads) were associated to the ATZ-positive tubules (Figure 1D, bottom, blue arrows). This suggests that the mutated protein concentrates in enlarged zones of the ER before its sequestration into the IB.
ATZ is prone to polymerization by a loop-sheet insertion mechanism (Lomas et al., 1992). By nondenaturing electrophoresis and Western blot analysis of the cell lysates, ATZ occurred as a 45-kDa band and a more abundant 98-kDa band (Figure 1, E and F). To determine whether the 98-kDa band corresponds to ATZ dimers, we coexpressed HA-ATZ and Myc-ATZ (Figure 1G). Anti-Myc antibodies coimunoprecipitated HA-ATZ, indicating that the mutated proteins form dimers and, perhaps, higher polymers. In the cell homogenates, both the amount of ATZ and the ratio ATZ dimer/ATZ monomer remained constant at the 24- and 48-h time points (Figure 1F). These data indicate that formation of IBs is a time-dependent cell response that is not due to changes of either the amount of ATZ at steady state or to increased formation of ATZ polymers.
Portland Neuroserpin Localizes to IBs in Neuroblastoma Cells
To determine whether IB formation is cell specific, we transiently expressed ATZ in N2A cells, and we monitored the distribution of ATZ in respect to that of the ER at 24 and 48 h after cell transfection (Figure 1H). Similarly to the Hepa 1-6 cells, most of the N2A cells formed IBs at the 48-h time point. This shows that the time-dependent IB formation is not specific to the hepatocyte. Diseases due to the accumulation of misfolded proteins have been also reported for other variants of the serpin family (Davis et al., 1999; Miranda et al., 2004). The disease familial encephalopathy with expression of polymerogenic neuroserpin variants leads to dementia, possibly by accumulation of mutant neuroserpins in neurons. The most severe form of dementia is caused by the Portland mutation (S52R) (Carrell, 2005). To determine whether Portland neuroserpin is accumulated in the same compartment as ATZ, the two proteins were coexpressed in neuroblastoma N2A cells. Some of the neuroserpin puncta (31.8% ± 24.7; n = 10 cells) colocalized with ATZ, indicating that at least a fraction of neuroserpin accumulates in IBs (Figure 1I).
ATZ in the IBs Is Sequestered from the Main ER
The ER chaperone calnexin binds ATZ in the ER (Cabral et al., 2002). At 48 h after transfection ATZ-containing IBs showed up as a distinct compartment from the main ER labeled by calnexin (Figure 1A). Immunoprecipitation experiments carried out with antibodies against calnexin and AAT showed that at 48 h, the interaction between ATZ and calnexin decreased by 
50% compared with that at 24 h (Figure 2A, i and ii, respectively). Thus, formation of the IBs correlates with decreased ATZ in a complex with calnexin. ATZ-containing IBs might exist as separate organelles from the main ER. In this case, it is possible that the two compartments, ER and IBs, fractionate differently. When PNS derived from Hepa 1-6 cells at 24 h after ATZ transfection was centrifuged onto a sucrose density gradient, the peak of ATZ colocalized with that calnexin (Figure 2Bi). At 48 h after transfection, the peak of ATZ was shifted from that of calnexin, indicating that at least a fraction of ATZ localizes to a compartment separate from the main ER. In the 48-h gradient, the fraction of ATZ migrating together with calnexin may reflect ATZ distribution to the ER in a population of transfected cells (30%; Figure 1B). To further explore whether IBs are separated from the ER, we used the microtubule-depolymerizing agent nocodazole. Consistent with reports that nocodazole induces the collapse of the ER (Lee et al., 1989), incubation of Hepa 1-6 cells with the drug lead to aggregation of calnexin fluorescence into large patches (Figure 3A). At 48 h after transfection, most of the cells had ATZ in the IBs and ATZ fluorescence was separate from the ER patches induced by the drug (Figure 3A, bottom). These data, together with those of Figures 1 and 2, indicate that IBs are organelles distinct from the main ER.
|
|
). In untransfected cells, calnexin- and KDEL-containing proteins had similar distribution (Supplemental Figure S4). In transfected Hepa 1-6 cells, KDEL immunoreactivity was both in the ER and concentrated in the IBs, showing that, although separated from the ER, these organelles retain some ER components (Figure 3B, arrowheads). In this respect, it has been reported that ATZ interacts with different cell chaperones that have the KDEL sequence, including PDI and GRP-78/BiP (Schmidt and Perlmutter, 2005; Papp et al., 2006). It is possible that these proteins accumulate in the IBs because they bind to ATZ. It has been proposed that a misfolded proinsulin accumulates mainly in ERGIC-53–positive pre-Golgi intermediates, and, to a lesser extent, in the ER (Zuber et al., 2004). Thus, ATZ-containing IBs may correspond to pre-Golgi intermediates. To determine whether this is the case, we used human HUH-7 cells that form IBs in response to ATZ expression like the Hepa 1-6 cells (Supplemental Figure S2) and where the intermediate compartment can be labeled by monoclonal antibodies against ERGIC-53 (Schweizer et al., 1993). However, ATZ did not colocalize with ERGIC-53, either in the perinuclear localization or in scattered cytoplasmic vesicles (Figure 3C, bottom). Autophagy has been involved in ATZ degradation (Teckman and Perlmutter, 2000; Kamimoto et al., 2006). The process occurs by sequestration of cytosol and organelles within double membrane vesicles, the autophagosomes, which then fuse with the lysosomes (Yorimitsu and Klionsky, 2005). We used the autophagosome marker LC3-GFP (Kabeya et al., 2000) to determine whether IBs are autophagosomes. ATZ did not colocalize with LC3-GFP (Figure 3D). Moreover, ATZ did not colocalize with the lysosomal marker LAMP1-GFP (Patterson and Lippincott-Schwartz, 2002) (Figure 3E) or with the dye LysoTracker Red that stains lysosomes (Supplemental Figure S4). This shows that IBs are neither classical autophagosomes nor lysosomes.
To further investigate the nature of the IBs and their relation to the ER, we carried out ultrastructural analysis of ATZ-expressing cells. To this end, Hepa 1-6 cells transfected with ATZ-GFP were sorted by fluorescence-activated cell sorting and analyzed by electron microscopy (Figure 4). Mock-transfected Hepa 1-6 cells were analyzed in parallel. At low magnification, the ATZ-expressing cells had fewer intact mitochondria and ER cisternae, and they had more clear vesicles and darker bodies (Figure 4B, white and black arrows, respectively) than the mock-transfected cells (Figure 4A). At higher magnification, the ER of control cells looked like tubules with enlarged tips (Figure 4C, arrows and arrowhead, respectively). In ATZ-expressing cells, ER tubules had a general appearance similar to control cells, but the tips of were more expanded than those in the control cells and they had a clear, electron-light core (Figure 4D, arrows and arrowheads, respectively). The ultrastructure shown here and the TIRFM of Figure 1D indicate that ATZ accumulates to subdomains of the ER. This is in agreement with immunoelectron microscopy studies showing colocalization of ATZ and calnexin in ER membranes (Teckman and Perlmutter, 2000) and immunoelectron microscopy showing that the liver disease-linked genetic variant null Hong Kong (NHK) of AAT localizes to the rough ER and to ER buds (Zuber et al., 2007). The numerous electron-light vesicles of ATZ-expressing cells showed up at higher magnification as round structures with no apparent connection with the ER tubules (Figure 4E), limited by only one membrane, and with and without attached ribosomes (Figure 4F, arrows and arrowheads, respectively). These structures are very similar to those described in the liver of PiZZ patients where ATZ was found in marked dilatations of the ER and smooth vesicles (Yunis et al., 1976; Hultcrantz and Mengarelli, 1984; Teckman and Perlmutter, 2000). Our electron microscopy (EM) data suggesting that ATZ induces formation of vesicles are in agreement with a recent work showing that NHK concentrates into vesicles other than the ER cisternae (Zuber et al., 2007). At higher magnification, most of the darker bodies (Figure 4B, black arrows) seemed to have multiple membranes (Figure 4G), some of which had the same morphological appearance as mitochondrial cristae. This observation is in agreement with previous studies showing mitochondrial injury in the liver of patients expressing ATZ (Teckman et al., 2004). In conclusion, data in Figures 1
–3 and the EM shown here indicate that ATZ is in a compartment that does not have calnexin and yet maintains some ER components such as ribosomes and luminal ER proteins. These observations suggest that in ATZ-expressing cells, the ER sheds to form ATZ-containing vesicles that are separate from the main compartment. The ultrastructural analysis of the ATZ-expressing Hepa 1-6 cells of Figure 4 also indicates that these cells recapitulate many structural features of hepatocytes from PiZZ patients.
|
). In addition, it has been proposed that dissociation of ATZ from calnexin is required for ATZ degradation (Cabral et al., 2002). We find that at 48 h, in most of the cells ATZ is out of the ER and sequestered into IBs that do not have calnexin. We reasoned that calnexin association to ATZ might control the formation of IBs by retaining the mutated protein in the main ER. To determine whether this is the case, we cotransfected calnexin and ATZ in the Hepa 1-6 cells. For the analysis, we selected a population of cotransfected cells where calnexin level was increased by approximately twofold at the 48-h time point (n = 40 cells), and we compared these cells to randomly chosen cells transfected with ATZ alone (n = 40 cells; Figure 5, A and B). Overexpression of calnexin shifted ATZ from the IBs to the ER, as indicated by the fourfold increase of ATZ colocalization with the chaperone (Figure 5, B and C). These results show that calnexin is involved in IB formation, and they suggest that release of ATZ from the chaperone is necessary for exit of the protein from the ER. To further investigate this possibility, we determined by immunoprecipitation experiments whether overexpression of calnexin increases binding of ATZ to the chaperone (Figure 5D). Transfection with exogenous calnexin did not induce a significant increase of the chaperone in the population of Hepa 1-6 cells (1.2 ± 0.35-fold; Figure 5Dii). Note that the measured calnexin level in cell transfected with the exogenous chaperone is lower than that reported in Figure 5A because the general cell population, rather than a selected group of cells as in Figure 5A, was used for the analysis. This is consistent with the observation that only a small percentage of cells is transfected (<20% of the general cell population). It has been reported by others that increased calnexin levels inhibit degradation of the NHK variant of AAT (Oda et al., 2003). However, changes of ATZ levels in the population of cells cotransfected with calnexin were again nondetectable compared with cells transfected with ATZ alone (1.2 ± 0.23-fold; Figure 5Dii). This is probably due to the low cotransfection rate of ATZ with calnexin (<38%). Immunoprecipitation with antibodies against calnexin lead to a sharp increase (5.3 ± 1.87-fold; Figure 5Di) of coimmunoprecipitated ATZ in samples from cells that were cotransfected with exogenous calnexin compared with those from cells transfected with ATZ alone. Note that the amount of immunoprecipitated calnexin in both samples is the same (Figure 5Di). This indicates that overexpression of the chaperone in the population of cotransfected cells leads to increased interaction of calnexin with ATZ. Together, the results of Figure 5, A and D, indicate that relocation of ATZ to the ER in the presence of overexpressed calnexin is dependent on lack of dissociation of ATZ from the chaperone. These data and those of Figure 2A indicate that release of ATZ from the chaperone is necessary for its exit from the main ER into the IBs.
|
). Moreover, overexpression of PDI in human embryonic kidney 293 cells inhibits NHK disposal (Hosokawa et al., 2006). In addition to the inhibition of ATZ degradation reported by others (Oda et al., 2003), we find that increased calnexin blocks IB formation. These observations raise the question as whether different cell chaperones have the same effect on ATZ levels and ATZ cell distribution. We measured the increase of intracellular ATZ in N2A cells in response to PDI and calnexin overexpression. In agreement with previous data, overexpressed PDI increased ATZ levels (Figure 5G). However, calnexin overexpression increased more effectively than that of PDI the cell level of ATZ (by 5.96-fold; p < 0.0001; Figure 5G). Next, we measured whether ATZ redistributed to the ER in the presence of increased PDI. IBs are defined as individual organelles of 2-µm2 area. ATZ, when localized in the endoplasmic reticulum, occupies a larger area than when the protein is in the IBs (Figure 5E). We reasoned that localization to organelles with an average area of <5 µm2 would discriminate ATZ in the IBs from that in the ER. Overexpression of calnexin shifted ATZ distribution from IBs to the larger main ER compartment (average area of 2.56 ± 1.42 and 19.89 ± 11.9 µm2, respectively; p < 0.0001; Figure 5, E and I). Differently, when PDI was overexpressed, ATZ remained in organelles with the same average area as the control (2.17 ± 1.64 and 2.65 ± 1.82 µm2; Figure 5, H and I). Moreover, PDI localized into the IB together with ATZ (Figure 5H), in agreement with our previous results that these organelles have some ER lumen components (Figure 3B). Thus, unlike calnexin, overexpression of PDI did not inhibit formation of IBs. These data show that calnexin has a specific role in IB biogenesis. Because expression of calnexin increased the amount of ATZ, it is possible that formation of the IBs is one route of ATZ disposal.
Formation of ATZ-containing IBs Occurs in Cells Capable of a Classical Autophagy Response
It has been reported that ATZ expression induces autophagy in the liver of patients and of transgenic mice but not in cells transiently expressing the protein (Teckman et al., 2000, 2004; Teckman and Perlmutter, 2000; Kamimoto et al., 2006). We reasoned that formation of the IBs or retention of ATZ in the ER induced by increased calnexin levels might result from a general inability of the hepatoma cells to activate the autophagy pathway. To monitor induction of autophagy, hepatoma cells were transfected with the autophagosomal marker LC3-GFP (Kabeya et al., 2000). Amino acid starvation is a classical stimulus to induce autophagy (Pattingre et al., 2005). On cell starvation, LC3-GFP changed its distribution with increased formation of "dot-like" autophagosomes in the cytoplasm as described by others (Mizushima et al., 2004; Pattingre et al., 2005) (Figure 6A). Conversely, the number of LC3-GFP–labeled autophagosomes did not significantly increase in response to starvation when cells were preincubated with the autophagy inhibitor 3-methyladenine (Teckman and Perlmutter, 2000) (Figure 6A). When Hepa 1-6 cells expressing either ATZ or ATZ and exogenous calnexin were starved, the number of LC3-GFP–labeled autophagosomes increased similarly to mock-transfected cells (Figure 6, B and C). In cells where autophagy was induced virtually none of the IBs had associated the autophagosome marker LC3-GFP in agreement with the data of Figure 3D. These experiments show that ATZ localization to IBs or ATZ retention in the ER induced by increased calnexin are not caused by lack of autophagy. Incubation of Hepa 1-6 cells expressing ATZ in the absence and in the presence 3-methyladenine did not change either the number of IBs (11.57 ± 11.76 and 12.69 ± 7.8, respectively) or their size (2.19 ± 2.36 and 2.78 ± 2.84 µm2, respectively). These data indicate that IBs and autophagosomes belong to two different pathways.
|
; Figure 5G). Thus, increased total amount of ATZ, rather than a change in ATZ distribution might induce cell shrinkage. To eliminate effects due to different levels of ATZ on the cell size, we measured the area of cells that had similar amounts of mutated serpin either with or without increased calnexin (ATZ = 1.10 ± 0.49 x 106 and 1.33 ± 0.53 x 106 pixels, respectively). Cells were stained to visualize at the same time calnexin, ATZ, and F-actin, and the cell margins were drawn manually on the merged image (Figure 7A). The mean area of the Hepa 1-6 cells transfected with ATZ was not significantly different from that of cells cotransfected with AAT and exogenous calnexin (406.25 ± 104.48 and 377.07 ± 122.27 µm2, respectively; Figure 7, A and B). This shows that when ATZ is in the IBs, cells have normal shape. In contrast, cells transfected with ATZ and exogenous calnexin had a size that was approximately half (203.49 ± 74.74 µm2) of that of cells transfected with ATZ alone. This shows that block of IB formation and retention of ATZ in the ER causes cell shrinkage.
|
To further explore the possibility that increased calnexin impairs constitutive secretion, we used endocrine cells that have both the constitutive and regulated secretory pathway. Unlike the constitutive pathway where proteins are constantly secreted, the regulated pathway stores proteins into organelles such as hormone-containing granules. This allows us to visualize directly proteins that have already traversed the ER, exited the trans-Golgi network, and are in granules poised for secretion. N2A cells have dense core granules at the tip of neurite-like processes where hormones are stored (Chevrier et al., 1991; Baldini et al., 2005). Expression of proinsulin in some endocrine cells other than the islet β-cells leads to targeting of mature insulin to dense core granules (Moore et al., 1983). To determine whether proinsulin is processed to the mature hormone in the N2A cells, we measured the amount of insulin and proinsulin by ELISA in transiently transfected Hepa 1-6 and N2A cells. In the Hepa 1-6 and N2A cell lysates the amount of proinsulin was comparable with 11.63 and 4.5 pmol/106 cells, respectively. In contrast, insulin was almost undetectable in the Hepa 1-6 cells compared with the N2A cells with 2.10 and 70.92 mU/106 cells of the mature hormone, respectively. This indicates that a fraction of the prohormone is converted to the mature form in the N2A cells. Insulin immunoreactivity in N2A cells occurred in a Golgi-like compartment near the nucleus and as punctuate staining at the tips of neurite-like processes and colocalized with the granule- associated protein Rab3A (data not shown) (Martelli et al., 2000). This indicates that N2A cells process proinsulin to its mature form and store it in dense core granules. Having established targeting of insulin to the granules in the N2A cells, we asked whether the redistribution of ATZ to the ER induced by increased calnexin levels blocked insulin traffic to the granules. In cells coexpressing ATZ and proinsulin, ATZ was in the IBs, and insulin immunoreactivity was evident in granules at the tips of the processes (Figure 7Di, bottom, arrows) and in the Golgi-like compartment (Figure 7Di, top, arrowhead), with little colocalization of the two proteins. This indicates that, when ATZ is localized to the IBs, proinsulin traverses the ER and the mature hormone is targeted to the granules. Differently, in N2A cells that coexpressed proinsulin together with ATZ and calnexin, the insulin immunoreactivity was retained in the ER together with ATZ as shown by the colocalization of ATZ and calnexin immunofluorescence (Figure 7Dii and E). Insulin-containing granules were not visible. Overexpression of calnexin by itself did not shift insulin distribution from the granule to the ER (Figure 7Diii). These experiments show that failure to sequester ATZ in IBs and consequent accumulation of the mutated ATZ in the ER induces a block in the secretory pathway at the step of protein exit from the ER.
Inhibition of IB Formation in N2A Cells Inhibits Neurite Formation
Neurite extension is a process that occurs by polarized addition of newly synthesized proteins and lipids at the tips of the neurites and requires an intact secretory pathway (Futerman and Banker, 1996). As an additional assay to determine the function of the IBs in the secretory pathway, we asked whether increased levels of calnexin blocked neurite formation in N2A cells that express ATZ. To measure the neurite outgrowth in ATZ expressing cells with and without exogenous calnexin, we selected cells with the same amount of ATZ (ATZ = 0.64 ± 0.12 x 106 and 0.70 ± 0.16 x 106 pixels, respectively). Calnexin, ATZ, and F-actin were visualized together (Figure 8A). The percentage of ATZ-expressing N2A cells that had neurites was 49.0 ± 9.7% (Figure 8Ai and B). This was not significantly different compared with cells expressing AAT and exogenous calnexin with 51.8 ± 10.2% of cells having neurites (Figure 8Aiii and B). In contrast, coexpression of ATZ and exogenous calnexin, inhibited formation of the neurites with 8.9 ± 6.3% of cells having the processes (Figure 8Aii, asterisk, and B). These data together with those in Figure 7 indicate that inhibition of IB formation impairs the secretory pathway at the ER level.
|
). It is possible that the lack of UPR activation in response to ATZ expression is due to segregation of ATZ out of the main ER to the IBs. In this case, retention of ATZ in the main ER in response to elevated calnexin levels might lead to UPR induction. To monitor UPR induction in Hepa 1-6 cells, we used tunicamycin, a drug that inhibits protein N-glycosylation and induces ER stress. UPR induction in response to ER stress can be monitored in HeLa cells by the increased levels of the ER chaperone GRP-78/BiP (Okada et al., 2002). Exposure of Hepa 1-6 cells to tunicamycin increased GRP-78/BiP immunoreactivity by approximately twofold (Figure 9A, lane 1 and 2), indicating that the drug induced a classical UPR in our experimental conditions. The level of GRP-78 did not change in cells expressing ATZ compared with cells transfected either with calnexin or with vector alone (Figure 9A, compare lanes 4 with lanes 3 and 1). This indicates that expression of ATZ or overexpression of calnexin does not induce the UPR. Conversely, coexpression of exogenous calnexin and ATZ lead to a significant increase of GRP-78 immunoreactivity compared with cells expressing ATZ alone (Figure 9A, lanes 5 and 4), indicating that the shift of ATZ from the IBs to the ER induces the UPR. To determine whether UPR induction occurred in the population of transfected cells, we monitored cells by immunofluorescence microscopy. Unfortunately, the antibody against GRP-78 could not visualize the mouse protein by immunofluorescence. Others have used an antibody against KDEL to monitor UPR induction by measuring increased GRP-78 levels by Western blot analysis (Lee, 2005). Tunicamycin treatment increased KDEL immunoreactivity in the Hepa 1-6 cells (Figure 9B). This indicates that induction of the UPR can be measured by the KDEL antibody by quantitative immunofluorescence microscopy. Expression of ATZ in the IBs did not increase KDEL immunoreactivity in the Hepa 1-6 cells (Figure 9Ci, iii, and v) consistent with the Western blot analysis of Figure 9A. Coexpression of ATZ and exogenous calnexin increased KDEL immunoreactivity by twofold compared with control cells expressing ATZ alone (Figure 9C, iii, iv, and v). We eliminated possible confounding effects due to different levels of ATZ on UPR induction by selecting for the analysis populations of cells with the same amount of ATZ (ATZ = 1.51 ± 0.66 x 106 and 1.52 ± 0.69 x 106 pixels in cells without or with increased calnexin, respectively). Conversely, a population of cells with no ATZ and similar levels of exogenous calnexin as the cells coexpressing the exogenous chaperone together with ATZ (calnexin = 2.97 ± 1.37 x 106 and 3.01 ± 1.175 x 106 pixels, respectively) did not have increased KDEL immunoreactivity (Figure 9Cii, iv, and v). In sum, both Western blot analysis and immunofluorescence experiments indicate that segregation of ATZ to the IBs prevents ER stress and the onset of the UPR.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Rudnick and Perlmutter, 2005). In addition, it is known that ATZ and other polymerogenic serpins form aggregates in "globules" or membrane-enclosed IBs in the liver and brain of affected individuals and in transfected cell models. The nature of the serpin-containing IBs in respect to the ER is unclear. Here, it is found that ATZ in the IBs does not have the same distribution as calnexin, an abundant ER chaperone, indicating that ATZ in the IBs is segregated from the main ER. In support of this possibility, IBs occurred in cells that otherwise maintained a largely intact ER, visualized both by the classical reticular pattern of calnexin distribution by fluorescence microscopy and by the overall normal appearance of ER cisternae at the ultrastructural level. It is possible that IBs are specialized ER sites that then develop to form separate organelles. In agreement with the second possibility, we find that microtubule disruption induced the collapse of the ER into structures that are clearly separated from the IBs and that ATZ migrated as a shifted peak from calnexin-containing ER by sucrose density gradient centrifugation. IBs might originate from shedding of the ER to form separate organelles. In this respect, ATZ-expressing cells have both ER tubules with enlarged, electron-light tips and numerous electron-light vesicles that do not maintain connections with the ER tubules. This suggests that ATZ would first concentrate at the tips of the ER cisternae and then form ATZ-containing vesicles that are separate from the ER. In respect to the possibility that IBs originate from the ER, ATZ concentrated in vesicles that could be visualized by TIRFM at the tips of ATZ-containing tubules, whereas there was no colocalization of ATZ with ERGIC-53 a marker of the ER-to-Golgi intermediate compartment (Schweizer et al., 1993). These data are in agreement with a recent observation that vesicles containing another polymerogenic serpin, the NHK variant, bud from the ER (Zuber et al., 2007). IBs, once separated from the ER, may yet retain some ER components such as luminal ER proteins (Figure 2) and ribosomes (Figure 4), as indicated by the rough ER-like ultrastructural appearance of ATZ-induced vesicles visualized both in transfected hepatoma cells (Figure 4) and in liver biopsies from patients with the disease (Yunis et al., 1976; Teckman and Perlmutter, 2000).
The work presented here shows that the ER chaperone calnexin functions in the cell localization of ATZ by controlling traffic of the protein from the ER to the IB. We have defined IBs as organelles of 2 µm2 area that do not have calnexin, maintain some other ER components such as the luminal protein PDI, and they are different from autophagosomes and lysosomes. We show that increased expression of calnexin leads to increased association of ATZ to the chaperone, retention of ATZ in the ER, and blocks formation of IBs both in hepatoma and neuroendocrine cells. This indicates that dissociation of ATZ from calnexin is a limiting factor for entry of ATZ into the IB. The role of calnexin in ATZ localization is specific because overexpression of another chaperone, PDI, did not block IB formation. The role of calnexin in ATZ disposal has been described by several studies where it has been reported that dissociation of ATZ from the chaperone is implicated in the degradation pathway of the mutated serpin (Liu et al., 1997, 1999; Cabral et al., 2000; Oda et al., 2003). We did not study the role of calreticulin in ATZ distribution because this soluble calnexin homologue has not been found to interact with the serpin (Cabral et al., 2002
