Abstract
ABSTRACT. Heat stress alters epithelial barrier function, and heat stress preconditioning protects epithelial function from injury. Hsp27 is a small stress protein that has previously been shown to modulate actin assembly. Thus, by regulating actin filaments associated with cell junctions, hsp27 could alter epithelial function. To begin to address this hypothesis, the regulation and distribution of a human hsp27-green fluorescence fusion protein (EGFPhHsp27) that is expressed in cultured renal epithelial cells was assessed. EGFPhHsp27, like the endogenous hsp27, associated with the cytoskeleton in heat-stressed and chemically ATP-depleted cells, and both proteins were regulated similarly. Confocal microscopy of intact and detergent-lysed cells revealed novel distribution patterns in which EGFPhHsp27 associated with basolateral, but not apical, cell borders in injured cells. Double labeling studies revealed EGFPhHsp27 and actin filament colocalization in ATP-depleted cells. However, during heat shock, granules of EGFPhHsp27 were found at sites of cell-cell contact and in the cell body, but colocalization with actin was not apparent. Thus, heat stress and ATP depletion induce distinct patterns of hsp27 redistribution in epithelial cells, and sites of cell-cell and cell-substrate attachment are unique in their ability to recruit hsp27 during injury. The association of EGFPhHsp27 with basolateral cell boundaries supports a potential role for hsp27 in protection or regulation of epithelial cell-cell and cell-substrate attachments.
Heat shock preconditioning enhances renal epithelial cell survival (1) and protects barrier function (2,3) from subsequent more severe injury. Several mechanisms have been investigated to explain this phenomenon, including inhibition of apoptotic cascades (4) and expression of high molecular weight stress proteins hsp70 (1,5–8) and hsp90 (9,10). However, alteration to the cytoskeleton underlies epithelial injury by many mechanisms (11,12). Thus, actin-associated proteins responsive to thermal stress could also regulate epithelial injury or repair. One such protein is the mammalian small heat shock protein, hsp27. Nonphosphorylated, but not phosphorylated, hsp27 behaves like an actin-capping protein that is capable of inhibiting actin polymerization in vitro (13,14), and altered hsp27 levels or expression of hsp27 phosphorylation site mutants induce changes in cellular actin polymer content and actin-dependent behavior (15,16). On the basis of these findings, previous investigators have suggested that hsp27 phosphorylation can enhance actin filament polymerization and cytoskeletal stability by reducing its ability to inhibit actin polymerization (17–19).
Although the above model of hsp27 function is consistent with in vitro and in vivo data, it is not clearly supported by the reported intracellular distribution of hsp27 in stressed and injured cells. For example, the hypothesis predicts that hsp27 associates with actin filaments in unstressed cells but may be released during injury. Instead, hsp27 is predominately found in the cytosolic fraction of unstressed cells and associates with the detergent-insoluble cell fraction, or cytoskeleton, in response to various injuries (20–23). Constitutive association of hsp27 with the apical microvillar border of proximal tubule epithelial cells and redistribution of hsp27 to intracellular compartments during injury have been observed in renal tissues (20,24). However, hsp27 in unstressed cells was detergent soluble, and association of hsp27 with specific cellular structures was not detected during injury. Thus, there is agreement that the factors that regulate hsp27 distribution and its association with the cytoskeleton are critical aspects of hsp27 function, but neither the mechanisms that mediate hsp27 redistribution nor their significance are fully understood.
In this study, we examined the distribution of a human wild-type hsp27-green fluorescence protein (EGFPhHsp27) in control and injured epithelial cells using two- and three-dimensional confocal imaging techniques. Novel and distinct distribution patterns of total and detergent-insoluble EGFPhHsp27 were detected, including colocalization of EGFPhHsp27 with actin-filament arrays at basolateral cell boundaries after ATP depletion but not heat shock. Biochemical studies confirm that the EGFPhHsp27 fusion protein and endogenous canine hsp27 are regulated in a similar manner under these conditions. The association of EGFPhHsp27 with basolateral sites of cell-cell and cell-substrate attachment suggests that hsp27 could directly affect the function or regulation of cell junctions and cell-substrate attachments.
Materials and Methods
Cell Culture
MDCK cells were purchased from American Type Culture Collection (Manassas, VA) and cultured as described elsewhere (25). An expression vector for EGFPhHsp27 was produced by cloning a human wild-type hsp27 cDNA into a commercial vector (EGFP-C1; Clontech, Palo Alto, CA). MDCKs were transfected using lipofectamine (Invitrogen Life Technologies, Carlsbad, CA) according to the manufacturer’s instructions. Transfected cells were cultured in Dulbecco’s modified Eagle medium (DMEM) that contained 600 μg/ml G418 sulfate (Invitrogen Life Technologies) for 4 wk and were then inspected by fluorescence microscopy. Cells were trypsinized briefly, and a fluorescence colony (MDCKe27) was subcloned by scraping with a sterile pipette tip while removing approximately 100 μL of medium. Expression and integrity of the EGFPhHsp27 fusion protein was confirmed by Western blotting using antibodies that recognize human hsp27 (SPA-800; StressGen Biotechnologies, Victoria, BC, Canada). An MDCK cell line that was stably transfected with the empty cloning vector was used for control studies.
For biochemical studies, equal volumes of a trypsinized cell suspension were plated into wells of a 24-well plate. For microscopy, MDCKe27 cells were mixed with parental cells at a ratio of 1:5 and plated onto 22-mm coverslips in Petri dishes. Cells were cultured 2 d after plating to allow monolayer formation. For ATP depletion experiments, medium was removed and cells were incubated for 1 h at 37°C in saline containing 10 mM 2-deoxy glucose and 1 μM antimycin A (reagents were purchased from Sigma Chemical Co., St Louis, MO, except as indicated). In other studies using similar methods, actin filament alterations in cultured renal epithelial cells similar to that described in the present study (see Results) correlated with a reduction in ATP level to <10% of their initial values (26,27). For heat shock studies, culture medium was replaced with DMEM medium containing 20 mM HEPES (pH 7.3). Cultures were sealed with parafilm and placed in a water bath (Fisher Scientific, IsoTemp 202; Fisher Scientific, Pittsburgh, PA) for 30 min at 37°C (controls), 42°C (mild heat shock), or 45°C (severe heat shock). Trypan blue exclusion studies after overnight incubation at 37°C indicated that about 80% of cells survived the severe heat shock and that negligible cell death occurred after ATP depletion or mild heat shock (not shown).
Cell Fractionation and Western Blotting
Treated and control cells were rinsed with ice-cold phosphate-buffered saline (PBS) and then harvested on ice in 500 μl of detergent lysis buffer (DLB: 20 mM Tris, 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid [EDTA], 1 mM ethyleneglycotetraacetic acid [EGTA], 0.5% Triton X100, pH 7.4) by vigorous scraping. Lysed cells were centrifuged at 16,000 × g for 5 min at 4°C. Supernatants were incubated with 2 volumes of ethanol overnight at −20°C and then recentrifuged. Both pellets were dissolved in 50 μl of sample buffer (2 mM EDTA, 70 mM sodium dodecyl sulfate [SDS], 20% glycerol, 20 mM dithiothreitol, 0.05% bromophenol blue, 20 mM Tris, pH 8.0) followed by brief sonication and boiling for 2 min. Equal sample volumes were loaded and resolved on 13% acrylamide gels using a Miniprotean II system (BioRad Laboratories, Hercules, CA), and proteins were transferred to nitrocellulose membranes using a BioRad Mini Trans-Blot. Membranes were probed with human hsp27-specific antibodies (SPA-800; Stressgen) or a monoclonal antibody (8A7, gift of Dr. M. J. Welsh, University of Michigan) detecting the conserved crystallin domain of hsp27 and αB-crystallin (28) and peroxidase-labeled secondary antibodies (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Blots were developed with a chemiluminescence detection system (Amersham Pharmacia Biotech, Inc., Piscataway, NJ) using Kodak x-ray film, and films were digitized by using a flatbed scanner (ScanJet 4C; Hewlett Packarad, Avondale, PA).
Analysis of hsp27 Phosphorylation
Previously described methods for detection of changes in hsp27 phosphorylation by isoelectric focusing (IEF) gel electrophoresis (29) were used here. To analyze total cell hsp27, cells were rinsed with ice-cold PBS and suspended in 50 μl of extraction buffer (EB: 9 M urea, 2% NP-40, 67 mM (β-d)-glycerophosphate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 2% 2-mercaptoethanol, 1 mM EDTA, pH 7.6). For analysis of triton-insoluble hsp27, cells were scraped in 500 μl of ice-cold DLB that contained protease and phosphatase inhibitors and centrifuged at 16,000 × g for 4 min, and pellets were dissolved in 50 μl of EB. IEF gels containing ampholytes (pH range, 5 to 7; Sigma Chemical Co) were prepared by using the Bio-Rad 111 Mini-IEF Cell. Equal volumes of sample were focused at 100 V for 15 min, 200 V for 15 min, and 450 V for 90 min. Proteins were transferred to nitrocellulose membrane by using a semi-dry transfer apparatus (Model FB-SDB-2020; Fisher Scientific), and hsp27 was detected as described above for Western blotting.
Fluorescence Microscopy
Cells on glass coverslips were rinsed with PBS and then fixed either with or without a 2-min pretreatment with DLB in PBS that contained 4% paraformaldehyde for 30 min (all at 21°C). Some preparations were also incubated with 0.2 μg/ml rhodamine phalloidin for 15 min. Anti ZO-1 immunostaining was conducted by using a monoclonal antibody (MAB1520; Chemicon International Inc., Temecula, CA) and Texas-red conjugated secondary antibodies (Southern Biotechnology Associates, Inc., Birmingham, AL) as described elsewhere (25). Preparations were mounted by using Prolong (Molecular Probes Inc., Eugene, OR) and sealed with nail polish. Alternatively, cells were cultured on membranes with 3-μm pores in MilliCell chambers (Sigma Chemical Co.) and then lysed, fixed, and cryopreserved overnight in PBS that contained 20% sucrose. The membrane was excised with a razor blade, incubated in PBS that contained 20% sucrose and 30% Tissue Freezing Medium (Electron Microscopy Sciences, Washington, PA), and then frozen in an ethanol/dry ice bath. Using a cryotome (Frigocut N 2800, Reichert-Jung, Germany), 7-μm-thick sections were cut perpendicular to the membrane and mounted on Superfrost slides (Fisher Scientific). Imaging was conducted with a Zeiss LSM510 confocal microscope (Carl Zeiss Inc., Thornwood, NY) with a 60 × 1.2 NA water immersion lens. Confocal pinholes were adjusted to an Airy disk diameter of 1, generating images 0.14 μm/pixel (X and Y dimensions) and 0.2 μm/plane (Z dimension). Dual-labeled specimens were imaged by using a multiscan mode in which the 512 individual scan lines of an image were scanned twice, illuminating and imaging each of the two fluorescence probes sequentially instead of simultaneously. This method eliminated the tendency of the signal from one fluorescence probe to corrupt images of the other probe.
Results
EGFPhHsp27 and hsp27 Redistribution in MDCKs by Heat Shock and ATP Depletion
In control parental MDCKs, hsp27 is detected in the detergent-soluble (Cnt-S), but not -insoluble (Cnt-P), cell fraction (Figure 1A). ATP depletion (1 h) and severe heat shock (30 min at 45°C) cause a dramatic increase in detergent-insoluble hsp27 and a decrease in detergent-soluble hsp27, but little change in hsp27 solubility is detected in cells after mild heat shock (30 min at 42°C). In MDCKs that express EGFPhHsp27 (cell line MDCKe27), the fusion protein (>43 kD) is also predominately soluble in control cells and cells after mild heat shock. Detergent-insoluble EGFPhHsp27 is detected after 1 h of chemical ATP depletion or 30 min of severe, but not mild, heat shock (Figure 1B). Relatively more EGFPhHsp27 than canine hsp27 is retained in the soluble fraction in these experiments, perhaps because excess EGFPhHsp27 is expressed by transfected cells. Figure 1C shows this blot after reprobing with a monoclonal antibody (8A7) that recognizes EGFPhHsp27, the endogenous canine hsp27, and a protein of approximately 22 kD (αB-crystallin). Normal association of endogenous hsp27 and αB-crystallin with the detergent-insoluble cell fraction occurs in cells expressing EGFPhHsp27. Thus, canine hsp27 and EGFPhHsp27 are specifically detected by immunoblotting with appropriate antibodies, and the solubility of EGFPhHsp27 and canine hsp27 is regulated similarly by MDCKs. Overexpression of EGFPhHsp27 may saturate the ability of the triton-insoluble cell fraction to recruit EGFPhHsp27, but it does not otherwise appear to alter hsp27 regulation.
Figure 1. Redistribution of endogenous canine hsp27 and a human hsp27-green fluorescent protein fusion protein in MDCKs. (A) Western blot of detergent-insoluble (P) and -soluble (S) cell fractions isolated from parental MDCKs after no treatment (Cnt), 1 h of ATP depletion (-ATP), or heat stress for 30 min at 42°C (42C) or 45°C (45C) probed with a primary antibody specific for canine hsp27. Hsp27 coprecipitates with detergent-insoluble proteins in ATP-depleted cells and cells subject to heat stress at 45°C but not in controls or cells heat stressed at 42°C. (B) Western blot of cell fractions isolated from MDCKs expressing EGFPhHsp27 probed with a primary antibody specific for human hsp27, showing similar results. (C) Western blot of cell fractions isolated from MDCKs expressing EGFPhHsp27 probed with a primary antibody that recognizes hsp27 and αB-crystallin.
Posttranslational Modification of Total and Triton-Insoluble hsp27 in MDCKs
Phosphorylation of hsp27 can be detected as altered migration of hsp27 in IEF gels (28,30–32). Three isoelectric isoforms of canine hsp27 occur; the most basic of which is nonphosphorylated. Singly and doubly phosphorylated isoforms are progressively more acidic (32). In our studies, untreated and ATP-depleted parental MDCKs expressed a single predominate isoform (presumably nonphosphorylated). Cells show increasing amounts of two more acidic isoforms after mild and severe heat shock (Figure 2A). Triton-insoluble pellets isolated from cells after ATP depletion or severe heat shock contain hsp27 isoforms representative of those observed in the total cell pool (Figure 2B, lanes 2 and 4). As expected, little hsp27 is detected in pellets that have been isolated from untreated cells or cells subject to mild heat shock (Figure 2B, lanes 1 and 3). Overall IEF behavior of EGFPhHsp27 in MDCKe27 cells is similar to that observed for canine hsp27 in parental MDCKs (Figure 2, C and D), although the specific IEF points of EGFPhHsp27 and canine hsp27 isoforms are not identical (not shown). Most EGFPhHsp27 in untreated and ATP-depleted MDCKe27 cells focuses at a single location, and a single more acidic isoform is also detected in heat-shocked cells (Figure 2C). Again, triton-insoluble EGFPhHsp27 (Figure 2D) displays an IEF migration pattern that is indistinguishable from that of total cell samples (Figure 2C). Only one phosphorylated isoform of EGFPhHsp27 is detected, which is consistent with the results of earlier studies that examined phosphorylation of an exogenous human hsp27 in nonhuman cell lines (15). Together, these data indicate that EGFPhHsp27 and the endogenous canine hsp27 are regulated similarly in MDCKs under our experimental conditions.
Figure 2. Isoelectric focusing patterns of endogenous canine hsp27 and a human hsp27-green fluorescent protein fusion protein in MDCKs. (A) Isoelectric focusing (IEF) gel of total cell protein in MDCKs. The most basic hsp27 isoform (presumably nonphosphorylated) predominates in control (Cnt) and ATP-depleted (-ATP) cells. More acidic (phosphorylated) isoforms are detected in heat-stressed cells (42C, 45C). (B) IEF gel showing detergent-insoluble canine hsp27 in MDCKs. (C) IEF gel showing the isoelectric focusing pattern of total cell EGFPhHsp27 in MDCKe27 cells. A single isoform is seen in control and ATP-depleted cells, and a more acidic isoform is also detected in heat-stressed cells. (D) IEF gel showing the isoelectric focusing pattern of detergent-insoluble EGFPhHsp27. Gels are oriented so that the most basic proteins are at the bottom of each image. Blots were probed antibodies that recognize canine hsp27 (A and B) or are specific for human hsp27 (C and D).
EGFPhHsp27 Localization in MDCKs during Heat Stress or Chemical ATP Depletion
To specify the localization of hsp27 in epithelial cells, total cell EGFPhHsp27 distribution was first examined in MDCKe27 cells that were fixed without detergent lysis. In untreated cells (Figure 3A) and cells after mild heat shock (Figure 3B), homogeneous distribution of EGFPhHsp27 within the cell body is observed. Decreased fluorescence intensity at cell boundaries after mild heat shock probably reflects changes in cell shape (flattening) and reduced cell-cell attachment. After severe heat shock, the distribution of EGFPhHsp27 is much less homogeneous (Figure 3C). Fluorescence granules of EGFPhHsp27 are seen throughout the cell and at sites of cell-cell contact (Figure 3C, arrow). Finally, after ATP depletion (Figure 3D), association of EGFPhHsp27 with sites of cell-cell contact is much more apparent than in cells after heat shock (Figure 3, compare C with D). EGFPhHsp27 is also excluded from the nucleus under all experimental conditions. To provide internal negative controls for fluorescence, these experiments were conducted by using cocultures of MDCKe27 and parental MDCK cells (see Methods). The shape of fluorescence cell boundaries in Figure 3 (A, C, and D) indicates the presence of nonfluorescence parental MDCKs in these images. No fluorescence was detected in cultures of parental MDCKs alone (not shown).
Figure 3. Localization of EGFPhHsp27 in intact cells. (A) Homogeneous distribution of EGFPhHsp27 is seen in control cells. (B) Largely homogeneous EGFPhHsp27 is seen in cells incubated at 42°C for 30 min before fixation. (C) Cytoplasmic aggregates and some recruitment of EGFPhHsp27 to sites of cell-cell contact (arrow) is seen in cells incubated at 45°C for 30 min before fixation. (D) Recruitment of EGFPhHsp27 to sites of cell-cell contact is highly evident in cells after 1 h of chemical ATP depletion. Scale bar, 20 μm.
To determine the location of detergent-insoluble EGFPhHsp27, cells were detergent-lysed before fixation and then examined. Little EGFPhHsp27 is seen in untreated cells after detergent lysis (Figure 4A), which confirms that EGFPhHsp27 is soluble in nonstressed cells. After mild heat shock, detergent-insoluble EGFPhHsp27 is detected at low levels in a novel distribution pattern that consists of small granules at sites of cell-cell contact (Figure 4B, middle and green). Actin filament arrays are associated with cell boundaries in this optical section (Figure 4B, left and red) but are not clearly associated with EGFPhHsp27. In ATP-depleted cells, detergent-insoluble EGFPhHsp27 is strikingly associated with cell boundaries; granules are also seen throughout the cell body (Figure 4C). In these cells, colocalization with phalloidin-stained actin is evident, but only at cell boundaries (Figure 4C, yellow). EGFPhHsp27 within the cell body (Figure 4C, green) shows little association with phalloidin-stained structures. Finally, cells also retain more EGFPhHsp27 after severe heat shock (Figure 4D) than after mild heat shock, and EGFPhHsp27 is found in larger and more distinct granules throughout the cell and at sites of cell-cell contact. Intriguingly, comparison of actin and EGFPhHsp27 distributions at sites of cell-cell contact reveals that granules of EGFPhHsp27 (Figure 4D, arrows) alternate rather than colocalize with areas of high phalloidin staining. In an overlay image, the resulting codistribution pattern resembles a string of alternating red and green beads along sites of cell-cell contact (Figure 4D). For all studies, cells without phalloidin staining were also examined, yieldeding identical EGFPhHsp27 distributions (not shown).
Figure 4. Localization of detergent-insoluble EGFPhHsp27 and actin. Dual-channel confocal optical sections show actin (left panel and red in the overlay image) and detergent-insoluble EGFPhHsp27 (middle panel and green in the overlay image). (A) Control cells exhibit little detergent-insoluble EGFPhHsp27. (B) Small localized regions containing detergent-insoluble EGFPhHsp27 are seen at sites of cell-cell contact in cells heat stressed at 42°C. (C) Detergent-insoluble EGFPhHsp27 in ATP-depleted cells showing aggregates throughout the cell body and strong colocalization with actin filament arrays at sites of cell-cell contact (yellow). (D) Detergent-insoluble EGFPhHsp27 in cells heat stressed at 45°C throughout the cell body and at discrete sites associated with cell-cell contacts. Scale bar, 10 μm.
To test whether the EGFP protein alone could be detergent-insoluble in some conditions, MDCKs stably expressing EGFP were examined. Intact cells fixed without detergent lysis show fluorescence throughout the cell body and nucleus (Figure 5A). Control cells (Figure 5, B and E) and cells subject to either 1 h of chemical ATP depletion (Figure 5, C and F) or severe heat shock (Figure 5, D and G) were detergent lysed, fixed, and double labeled with fluorescence phalloidin. No detergent-insoluble EGFP was detected after any experimental treatment (Figure 5 B through D). Alterations in actin filament organization, such as formation of actin aggregates in ATP-depleted cells (Figure 5F), and a reduction in basal actin filament bundles after heat stress (Figure 5G) were observed, which is consistent with observations that have been made by previous investigators using similar protocols (26,33).
Figure 5. The enhanced green fluorescent protein alone does not remain in epithelial cells after detergent lysis. (A) Untreated MDCKs stably expressing EGFP were fixed without detergent lysis, showing homogeneous EGFP distribution. (B through D) Detergent-insoluble EGFP is not detected in untreated cells (B) or cells after ATP depletion (C) or severe heat shock (D). Also detected in these double-labeled cells is actin filament organization in untreated cells (E) and cells after ATP depletion (F) or severe heat shock (G) showing typical alterations. Scale bar, 10 μm.
The relationship between actin and detergent-insoluble EGFPhHsp27 was determined in three dimensions by using optical sectioning methods. Figure 6 is a series of confocal images taken with 0.2-μm changes in focal position of the junction between two cells heat shocked at 42°C. Granular structures that contain detergent-insoluble EGFPhHsp27 are indicated by arrows in the most basal optical section (but individual granules may be most apparent in a neighboring image plane). These granules are only visible in panels 1 to 3. In panels 6 through 13, a second, more apical site that shows EGFPhHsp27 is seen. These granules are smaller and more uniform than those observed at the basal surface. The inset shows a model of the distribution of basal coarse granules (1) and finer granules at lateral and apical sites of cell-cell contact (2).
Figure 6. Distribution of detergent-insoluble EGFPhHsp27 in epithelial cells after mild heat shock. Fourteen confocal images of a single cell-cell contact were obtained with a 0.2-μm focal change between images. Granules of EGFPhHsp27 associated with the boundary between the two cells are indicated with arrows in the first (most basal) image. A more apical site containing EGFPhHsp27 is also seen in higher images. Inset shows a diagram of EGFPhHsp27 distribution at basal (1) and apical (2) sites of cell-cell contact. Scale bar, 5 μm.
The greater amounts of EGFPhHsp27 that were detected in ATP-depleted and severely heat-shocked cells compared with untreated cells and cells subject to mild heat shock allowed double labeling and computer-assisted reconstruction of XZ images. Figures 7A and 7B show two examples of XZ images that are oriented so the image plane passes vertically through a site of cell-cell contact (the orientation of the image plane is illustrated in Figure 7B). Actin (red and lower gray scale image) typically appeared as large globular structures containing little or no detectable EGFPhHsp27 (upper gray scale image). Occasionally, vertical columns of actin are seen (Figure 7B, arrows) that also lack EGFPhHsp27. In ATP-depleted cells, detergent-insoluble EGFPhHsp27 is found along the lateral sites of cell-cell contact and at the basal surface of cells (Figure 7, C and D), but little EGFPhHsp27 is observed at the apical cell surface (the plane of section is illustrated in Figure 7D). In contrast, filamentous actin is present at both the apical and basolateral cell surfaces (Figure 7C, red). Arrows in Figure 7C mark the most apical extent of the EGFPhHsp27 fluorescence that is associated with sites of lateral cell-cell contact. In a separate experiment, ATP-depleted MDCKe27 cells were also stained with antibodies to reveal the distribution of the tight junction protein ZO1. Detergent-insoluble EGFPhHsp27 at the basolateral cell border does not extend above the level of apical tight junctions (Figure 7D, arrows). Finally, previous studies have shown that fluorescence objects can display different apparent intensities of fluorescence at different orientations with respect to the optical axis (34). To address this potential problem, MDCKe27 cells exposed to chemical ATP depletion and severe heat shock were detergent lysed, fixed, and cryosectioned perpendicular to the substrate plane. Basolateral EGFPhHsp27 is observed in cryosectioned cells after ATP depletion (Figure 7E) and is particularly bright at the most apical site of cell-cell contact (arrows). In contrast, no similar distribution of EGFPhHsp27 is observed at sites of cell-cell contact after heat shock (Figure 7F).
Figure 7. Three-dimensional distribution of actin, Z01, and detergent-insoluble EGFPhHsp27 in epithelial cells after severe heat shock or ATP depletion. (A and B) Two examples of XZ images through sites of cell-cell attachment after sever heat shock showing actin (red and lower gray-scale images) and EGFPhHsp27 (green), showing lack of colocalization. (B) Illustration of the plane of section for panels A and B. (C) XZ image of ATP-depleted MDCKe27 cells showing colocalization of actin (red and lower gray-scale images) and EGFPhHsp27 (geen) at basolateral but not apical cell borders. (D) Double labeling of ATP-depleted MDCKe27 cells showing EGFPhHsp27 (green) and Z01 (red). Localization of EGFPhHsp27 is restricted to sites below tight junctions. (D) Plane of section for images shown in panels C through F. (E) Cryosection showing XZ distribution of EGFPhHsp27 in ATP-depleted MDCKe27 cells without optical sectioning. (F) Cryosection showing EGFPhHsp27 in severely heat-shocked MDCKe27 cells. Scale bar, 10 μm.
Discussion
The most significant new finding of this study is evidence of the direct and specific association of hsp27 with sites of epithelial cell-cell and cell-substrate contact in injured epithelial cells. Because association of EGFPhHsp27 with lateral cell boundaries is seen in cells fixed before lysis (Figure 3) as well as after lysis, this association represents recruitment of hsp27 from the cellular pool to these sites and is not a result of washing away detergent-soluble protein from homogeneously distributed hsp27. A relationship between hsp27 and the regulation of cell-cell and cell-substrate attachment has been suggested by studies in which changes in hsp27 expression or phosphorylation correlate with altered cell adhesion, junctional integrity, and cell migration (reviewed in (18). However, our results provide the first direct evidence that hsp27 is a component of cell-cell and cell-substrate attachments in injured cells. Our results also confirm and extend those of a previous study indicating that up to one third of cellular hsp27 can associate with a basolateral membrane cell fraction in vascular endothelial cells (17). However, there are notable differences between the previous and current results. For example, studies conducted by Piotrowicz and Levin (17) showed constitutive association of hsp27 with the basolateral membrane in endothelial cells, and the abundance of membrane-associated hsp27 did not change when cells were stimulated with phorbol ester. In contrast, EGFPhHsp27 is found in our studies at sites of epithelial cell-cell and cell-substrate contact during hyperthermia or ATP depletion, but not in untreated cells. The previous study also did not specifically address whether membrane-associated hsp27 was a component of the triton-insoluble cell fraction, and because the previous study used only biochemical assays, the subcellular localization pattern of membrane-associated hsp27 was not determined. Our work extends the previous study by showing that basolateral EGFPhHsp27 is a component of the detergent-insoluble cell fraction and that discrete granular structures found at apical and basal sites of epithelial cell-cell contact are specific sites of hsp27 association in cells during mild thermal injury. Sites of both cell-cell and cell-substrate attachment show colocalization of actin and hsp27 in ATP-depleted cells in our studies.
The intracellular distribution of hsp27 was also previously examined in renal epithelial cells subject to ischemia/ATP-depletion using immunofluorescence staining methods. Studies examining hsp27 distribution in sectioned rat kidney tissues provided evidence for the colocalization of hsp27 with actin filament arrays, most notably at or near the apical microvillar border in unstressed renal proximal convoluted tubule epithelial cells (17,21, but see (24 for an exception). Redistribution of hsp27 accompanied ischemic injury, resulting in loss of hsp27 from the apical compartment and its association with the triton-insoluble cell fraction. Thus, the previous and current studies support the view that hsp27 redistribution and differential regulation of its association with the cytoskeleton accompany renal cell injury. However, there are also findings that are unique to the present study. For example, in contrast to our own results, association of hsp27 with specific actin filament arrays was not detected during injury in previous studies, and hsp27 detected at the apical microvillar border of unstressed cells did not associate with the detergent insoluble cell fraction. It is of interest to consider the basis of these differences. Previous investigators examined hsp27 in intact renal proximal tubules. We examined cultured MDCK epithelial cells, a cell line derived from canine distal convoluted tubule epithelial cells. MDCKs are more resistant to some types of cell injury than are epithelial cells of proximal tubule origin (35). Thus, the association of hsp27 with sites of cell-cell and cell-substrate attachment observed in our studies may partially account for the observed greater resistance to injury of distal versus proximal convoluted tubule epithelia. Alternatively, our results may reflect a relatively poor state of differentiation of MDKCs in culture, and our cell cultures may model hsp27 function during development or repair of epithelial tissues, rather than the condition of fully differentiated cells. Differences between the results of our own and previous studies using immunostaining methods may indicate that highly antigenic hsp27 domains are important to the association of hsp27 with the triton-insoluble cell fraction or cell junctions and are unavailable for antibody binding in intact cells. Finally, cells expressing the EGFPhHsp27 fusion protein in our studies likely express higher than normal overall levels of hsp27. Thus, the cellular response of MDCKe27 cells to heat shock and ATP depletion in our studies may model that of preconditioned cells expressing high levels of hsp27 before injury. In contrast, preconditioning of renal tissues was not used in the in vivo studies conducted by Aufricht et al. (20) or Smoyer et al. (29). Further studies examining hsp27 redistribution in cells subject to injury after heat stress preconditioning or with no preconditioning will likely resolve this issue. However, the differences in results between the present and previous studies may indicate that association of hsp27 with basolateral cell boundaries is an element of cellular cytoprotection rather than a response to initial cellular injury.
Although our results clearly demonstrate alteration in the ability of epithelial cell-cell and cell-substrate attachment to recruit regulatory factors during thermal and ATP depletion-induced injury, these results only address the function of hsp27 if the EGFPhHsp27 fusion protein accurately mimics the behavior of endogenous hsp27. We believe there is ample evidence that this is the case. Borrelli has shown that expression of EGFPhHsp27 provides thermal protection to A549 human lung carcinoma and L929 murine fibroblast cells (Michael J. Borrelli, personal communication, March 2001). Here, we demonstrate that EGFPhHsp27 and canine hsp27 associate with cytoskeleton similarly in MDCKs after chemical ATP depletion and severe heat shock. However, neither protein is detected significantly in pellets obtained from untreated cells or cells after mild heat shock (Figure 1). Our studies also reveal similar overall changes in the regulation of canine hsp27 and the EGFPhHsp27 fusion protein in response to heat stress and ATP depletion (Figure 2). Although EGFPhHsp27 is only singly phosphorylated in MDCKs in response to thermal injury, this is consistent with previous work showing that nontagged human hsp27 is predominately singly phosphorylated in response to heat shock in a Chinese hamster cell line (15). Finally, we show that expression of the EGFPhHsp27 fusion protein in MDCKs does not disrupt normal regulation of canine hsp27 (Figure 1C). Thus, there is no evidence that addition of the EGFP tag alters the function of the human hsp27 protein in our studies.
In summary, our results indicate that EGFPhHsp27 is an accurate model of hsp27 function in renal epithelial cells and that studies of the intracellular distribution of EGFPhHsp27 are an important complement to immunostaining studies. The novel distribution patterns of total and detergent-insoluble EGFPhHsp27 protein observed in our studies provide new evidence for a specific association of hsp27 with structures mediating cell-cell and cell-substrate attachment, and they suggest that hsp27 could play a role in the regulation and function of these structures during injury.
Acknowledgments
These studies were supported by grants from the National Kidney Foundation of Michigan, Inc., the National Institute of Environmental Health and Safety (NIH R01 ES11196–01), and the National Institute on Aging (NIH R03-AG19847–01) to E. A. S. and a grant from the National Cancer Institute (NIH R01 CA71650) to M. J. B. We thank Drs. R. Benndorf and J. M. Weinberg for critical reading of this manuscript and helpful comments, Dr. S. A. Ernst for access to his cryotome, and Ms. J. Edwards for instruction in its use.
- © 2002 American Society of Nephrology
References
Jump to section
More in this TOC Section
Cited By...
- Hsp72 Interacts with Paxillin and Facilitates the Reassembly of Focal Adhesions during Recovery from ATP Depletion
- Hsp27 Associates with Actin and Limits Injury in Energy Depleted Renal Epithelia
- Transitioning to Therapy in Ischemic Acute Renal Failure
- Site-Specific Alteration of Actin Assembly Visualized in Living Renal Epithelial Cells during ATP Depletion