The 70-kDa Heat Shock Proteins Associate with Glandular Intermediate Filaments in an ATP-dependent Manner
1995; Elsevier BV; Volume: 270; Issue: 2 Linguagem: Inglês
10.1074/jbc.270.2.915
ISSN1083-351X
AutoresJian Liao, Lori A. Lowthert, Nafisa Ghori, M. Bishr Omary,
Tópico(s)Bee Products Chemical Analysis
ResumoKeratin polypeptides 8 and 18 (K8/18) are intermediate filament proteins expressed preferentially in glandular epithelia. We describe the identification, by co-immunoprecipitation from normal human colonic tissues and cultured cell lines, of the 70-kDa heat shock protein (hsp) and its related heat shock cognate protein as K8/18-associated proteins (hsp/c). The association is significant but sub-stoichiometric and occurs preferentially with the soluble rather than the cytoskeletal K8/18 fractions. Heat stress increases the level of soluble K8/18 in association with an increase in hsp70 levels and an increase in the stoichiometry of K8/18-hsp70 association. Identity of the associated proteins was confirmed by microsequencing of a tryptic digest of the purified associated protein and by using anti-hsp/c70-specific antibodies. The K8/18-hsp/c70 complex can be dissociated in a Mg-ATP-dependent manner that requires ATP hydrolysis. Binding of hsp to K8/18 can be reconstituted using purified bovine hsp70 and human K8/18 immunoprecipitates that have been depleted of bound hsp/c70 and increases slightly in the presence of ATP. The reconstituted K8/18-hsp70 complex can be again released in the presence of Mg-ATP. In addition, hsp70 binds to K8/18 without having a significant effect on in vitro filament assembly when added during or after assembly. Using an overlay assay, hsp70 binds exclusively to K8 in the presence of ATP. Our results show direct association of the hsp/c70 proteins with K8/18. This interaction may serve, at least in part, to regulate the function of these two abundant protein groups. Keratin polypeptides 8 and 18 (K8/18) are intermediate filament proteins expressed preferentially in glandular epithelia. We describe the identification, by co-immunoprecipitation from normal human colonic tissues and cultured cell lines, of the 70-kDa heat shock protein (hsp) and its related heat shock cognate protein as K8/18-associated proteins (hsp/c). The association is significant but sub-stoichiometric and occurs preferentially with the soluble rather than the cytoskeletal K8/18 fractions. Heat stress increases the level of soluble K8/18 in association with an increase in hsp70 levels and an increase in the stoichiometry of K8/18-hsp70 association. Identity of the associated proteins was confirmed by microsequencing of a tryptic digest of the purified associated protein and by using anti-hsp/c70-specific antibodies. The K8/18-hsp/c70 complex can be dissociated in a Mg-ATP-dependent manner that requires ATP hydrolysis. Binding of hsp to K8/18 can be reconstituted using purified bovine hsp70 and human K8/18 immunoprecipitates that have been depleted of bound hsp/c70 and increases slightly in the presence of ATP. The reconstituted K8/18-hsp70 complex can be again released in the presence of Mg-ATP. In addition, hsp70 binds to K8/18 without having a significant effect on in vitro filament assembly when added during or after assembly. Using an overlay assay, hsp70 binds exclusively to K8 in the presence of ATP. Our results show direct association of the hsp/c70 proteins with K8/18. This interaction may serve, at least in part, to regulate the function of these two abundant protein groups. INTRODUCTIONIntermediate filaments (IF) 1The abbreviations used are: IFintermediate filamentshscheat shock cognate proteinhspheat shock proteinHSPheat shock protein familyMAbmonoclonal antibodyPBSphosphate-buffered salinePAGEpolyacrylamide gel electrophoresisS and P fractionssoluble and pellet fractionsIFAPintermediate filament-associated proteinsBSAbovine serum albuminATP γ Sadenosine 5′-O-(thiotriphosphate). comprise one of the three major cytoskeletal protein groups that include microfilaments and microtubules (for reviews, see Steinert and Roop, 1988; Klymkowsky et al., 1989; Skalli and Goldman, 1991; Fuchs and Weber, 1994). IF expression is generally tissue specific with the keratin IF subfamily being preferentially expressed in epithelial cells. Tissue-specific expression within epithelial subtypes also extends to keratins, which are comprised of at least 20 members (catalogued as K1-K20) that are found in cells as heteropolymeric pairs (Moll et al., 1982, 1990). Although each epithelial cell type may express several keratins, they usually express one pair as their dominant IF network. For example, glandular "simple" type epithelia, as found in the gastrointestinal tract, liver and pancreas express K8 and K18 (K8/18), keratinocytes express K5/14, and esophageal epithelia express K4/13 as their major keratin pair. IF proteins are highly abundant in cells that express them, as noted for K8/18, which make up ∼ 5% of the total protein in cultured colonic cell lines (Chou et al., 1993), and for keratinocyte keratins, which account for ∼ 30% of the cellular protein (Sun et al., 1979).A number of intermediate filament-associated proteins (IFAP) have been described that may be involved in regulating the function of IF or that may utilize binding to IF to regulate their function (Foisner and Wiche, 1991). These IFAP may interact directly or indirectly with IF and include kinases, membrane-associated proteins, and nuclear proteins. Examples of IF-kinase interactions include the association of protein kinase C-related kinases with K8/18 (Omary et al., 1992a) and vimentin (Spudich et al., 1992; Murti et al., 1992) and the association of cGMP-dependent protein kinase with vimentin (Wyatt et al., 1991; MacMillan-Crow and Lincoln, 1994). Other IFAP include ankyrin and lamin B (Georgatos et al., 1987; Djabali et al., 1991), desmoplakin (Stappenbeck et al., 1993), a 140-kDa desmosomal antigen (Cartaud et al., 1990), IFAP 300 (Skalli et al., 1994), plectin (Foisner et al., 1991), gyronemin (also called filamin) (Brown and Binder, 1992), and KAP85 (Chou et al., 1994).Another highly abundant protein family is the heat shock protein family (HSP) with major classes including HSP25, −60, −70, −80, −90, and −110 (numbers correspond to approximate Mr of class members) (Gething and Sambrook, 1992; Georgopoulos and Welch, 1993; Hendrick and Hartl, 1993; Becker and Craig, 1994). An important function of HSP proteins is to act as molecular chaperones, and in doing so they form complexes with a variety of cellular proteins (e.g. see Schlesinger, 1990). In mammalian cells, two cytosolic members of the HSP70 family are the heat-inducible hsp70 (Mr 72,000) and the constitutively expressed hsc70 (Mr 73,000) (Becker and Craig, 1994). Several indirect reports and reviews have suggested an interaction of HSP70-like proteins with IF, although most describe an association of HSP70 with microtubules (e.g. Schlesinger, 1990; Gupta, 1990). For example, 1) a ubiquitous 70-kDa protein termed β-internexin was initially identified as a microtubule-associated protein and shown indirectly to associate with IF (Napolitano et al., 1985) and then was subsequently shown to be identical to hsp70 (Green and Liem, 1989); 2) hsp70 was shown to co-purify with microtubules in murine mastocytoma cells (Ohtsuka et al., 1986) and with retinal microtubule and IF fractions (Clark and Brown, 1986); 3) in vitro interaction of purified yeast HSP70 with vimentin and nuclear lamins in an ATP-dependent manner has been reported (Georgatos et al., 1989); 4) treatment of cultured rat brain cells with the natural product withangulatin A resulted in increased vimentin and hsp70 levels in the detergent-nonextractable fraction, although direct association was not shown (Lee et al., 1993). More recently, α-crystallins, which are related to the HSP25 family, were shown to co-immunoprecipitate with vimentin in bovine lens and to inhibit in vitro filament assembly of vimentin and glial fibrillary acidic protein (Nicholl and Quinlan, 1994).During our studies of the post-translational modification of K8/18, we noted a 70-kDa band, visible by Coomassie staining, that co-immunoprecipitated with K8/18 (Chou et al., 1993). In this report, we show that this 70-kDa band corresponds to hsp/c70 and that its association with K8/18 increases with heat stress. K8/18-hsp/c70 association occurs preferentially with the soluble K8/18 fraction, with release of hsp in the presence of Mg-ATP. Furthermore, K8/18 and hsp70 binding and Mg-ATP-mediated release can be reconstituted using purified hsp70 and K8/18 immunoprecipitates that have been depleted of bound hsp/c70. In vitro binding of hsp70 to keratins using an overlay assay occurs exclusively with K8. Addition of purified hsp70 to K8/18 during or after filament assembly does not affect assembly. The potential functional significance of the K8/18-hsp70 association is discussed.MATERIALS AND METHODSReagentsPurified bovine brain hsp, α-actinin, and ATP γ S were purchased from Sigma. Monoclonal antibodies (MAb) to K8/18 that recognize different K8/18 epitopes (not shown) were CK5 and 8.13 ascites (Sigma) and L2A1 (Chou and Omary, 1991), which was used as ascites or covalently coupled to protein A-agarose. Other MAb used were M2B3, which recognizes a proliferation-associated cell surface glycoprotein (Omary et al., 1992c), and I4D4 (Omary et al., 1992b), which recognizes an epithelial cell adhesion marker termed KS-1 or more recently referred to as Ep-CAM (Litvinov et al., 1994). The latter two MAb were coupled to cyanogen-activated Sepharose (Pharmacia Biotech Inc.) as recommended by the manufacturer. Antibodies to hsp70(1197) and hsp/c70 (BRM-22) were purchased from Amersham Corp. and Sigma, respectively. 36 S protein-labeling mix (Met/Cys, 1200 Ci/mmol) and ENHANCE ® were from DuPont NEN.Cell Culturing, Metabolic Labeling, and Colonic TissuesHT29, T84, Caco2, and SK-CO-1 cell lines (human colon) were obtained from the American Type Culture Collection and were grown at 37°C (10% CO2) in RPMI 1640 medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, 100 μg/ml streptomycin, and 1 mM glutamine. For heat stress, cells at ∼ 50% confluency were grown at 42°C for 16-24 h. Colonic biopsies (2-3 mm in diameter) were obtained from patients undergoing routine colonoscopy for polyp screening under a protocol approved by the Medical Committee for the Protection of Human Subjects on Research at Stanford University.Metabolic labeling of HT29 cells with [ 36 S]Met/Cys was done by pre-incubating cells (1 h) in Met-free RPMI 1640 medium supplemented with 10% fetal calf serum dialyzed against 0.15 M NaCl, followed by addition of the 36 S label (100 μ Ci/ml) for 1, 5, or 10 min. Cells were then immediately placed over ice, followed by removal of the labeling medium and rinsing with cold phosphate-buffered saline (PBS) containing 15 μg/ml unlabeled Met. In some cases, cells were chased for 1-20 h by incubating in normal growth medium after rinsing off the label.ImmunoprecipitationCells were solubilized (30-60 min, 4°C) in 1% Nonidet P-40 in PBS containing 0.1 mM phenylmethylsulfonyl fluoride, 25 μg/ml aprotinin, 10 μM leupeptin, 10 μM pepstatin, and 5 mM EDTA (buffer A). Colonic biopsies were homogenized using a Dounce in buffer A (100 strokes) and then were allowed to solubilize with rocking for 1 h. Nonsolubilized material was pelleted (16,000 × g, 30 min, 4°C), and the solubilized material was used for immunoprecipitation. Alternatively, HT29 cells were disrupted by nitrogen cavitation (150 p.s.i., 5 min, 4°C) in PBS, 5 mM EDTA followed by pelleting (16,000 × g, 30 min, 4°C) to generate the soluble (S) fraction. The insoluble fraction was solubilized in buffer A (1 h) followed by repelleting to generate the detergent-solubilized P fraction. Immunoprecipitates were then obtained using the S fraction (directly or after adding detergent to a final concentration of 1% Nonidet P-40), buffer A-solubilized cells or biopsies, or the detergent-solubilized P fraction. For this, 200-500 μl of lysate (from 2-5 × 106 cells or 2 biopsies) were mixed with 20 μl of agarose- or Sepharose-coupled antibody or 3 μl of ascites. The ascites-antigen complexes were then collected using protein A-Sepharose conjugated to rabbit anti-mouse antibody.Protein AnalysisImmunoprecipitates were analyzed using SDS-polyacrylamide gel electrophoresis (PAGE) (Laemmli, 1970). Isoelectric focusing (Chou et al., 1992) and Western blotting (Towbin et al., 1979) were as described except that an enhanced chemiluminescence system (ECL) was used as recommended by the manufacturer (Amersham). Relative levels of K8/18 and hsp/c70 were determined using densitometric scanning of Coomassie stained-destained gels (LKB Ultrascan XL enhanced laser densitometer). Serial dilutions of bovine serum albumin (BSA) as a protein standard were analyzed on the same SDS-PAGE gel containing the proteins of interest.Filament Assembly and Fast Protein Liquid Chromatography Purification of K8/18Purified K8/18 were obtained prior to filament assembly using a two step procedure. The first step involved high salt extraction exactly as described (Achtstaetter et al., 1986; Chou et al., 1993). This generated a highly enriched K8/18, which is then solubilized in 6 M urea, 50 mM Tris-HCl, 2 mM EGTA (pH 7.5) (buffer B). Non-solubilized material was pelleted (16,000 × g, 2 min) followed by filtering of the solubilized material (0.2-μm filters) and then passage over a Mono-Q Resource ™ column assembled in a fast protein liquid chromatography apparatus (Pharmacia). K8/18 were then eluted using a 0-0.5 M guanidine HCl gradient (in buffer B) with the K8/18 peak being eluted between 130 and 260 mM guanidine HCl. Average yield of purified keratin was ∼ 50 μg per confluent 100-mm dish of HT29 cells. The eluted keratins were dialyzed against buffer B (4°C, 12-24 h) and then stored at 4°C till further use.Filament assembly was initiated by dialyzing K8/18 (150 μg/ml) against 50 mM Tris-HCl, 2 mM EGTA (pH 7.5) for 36 h with two buffer changes. Assembly was also done in the presence of bovine hsp70 or BSA. Alternatively, filaments were assembled for dialysis for 36 h and then incubated with BSA or hsp70 for 24 h at 4°C. Filament formation was assessed by negative staining followed by electron microscopy as described (Chou et al., 1993). Analysis of the soluble K8/18 fraction after assembly in the presence or absence of hsp70 or BSA was assessed by ultracentrifugation (300,000 × g, 1 h, 4°C) and then SDS-PAGE of equivalent fractions of the soluble and pellet fractions.Release of hsp/c70 from K8/18 and Reconstitution of K8/18-hsp70 BindingK8/18 immunoprecipitates were obtained from 1% Nonidet P-40 detergent lysates (of cells grown at 37 or 42°C) in the absence or presence of 5 mM Na-ATP, Na-ATP plus MgCl2 (Mg-ATP), Mg-AMP, Mg-ADP, Mg-GTP, ATP γ S, ATP γ S plus MgCl2, or MgCl2 alone. Alternatively, immunoprecipitates were isolated first and then incubated with 5 mM Mg-ATP for 15-90 min, followed by washing 2 times with PBS to remove released hsp/c70. K8/18 immunoprecipitates depleted of hsp/c70 (K8/18-hsp − ) were then incubated with 50 μl of PBS containing 2 μg of bovine hsp70 (60 min) in the presence or absence of 5 mM Na-ATP, Mg-ATP, or MgCl2 followed by washing to remove unbound hsp70 and then analysis by SDS-PAGE.Keratin-hsp70 Binding Using an Overlay AssayA highly enriched keratin fraction was prepared by high salt extraction (Chou et al., 1993) followed by separation using preparative SDS-PAGE. Individual K8, K18, and K19 bands were visualized using copper stain (Bio-Rad), cut and electroeluted, and then concentrated using Centricon ™ microconcentrators (Amicon). SDS-PAGE of individual K8, K18, K19, and α-actinin was then done, followed by transfer to two duplicate polyvinylidene fluoride membranes. One membrane was stained with Coomassie Blue, and the second was blocked with 3% BSA in PBS for 48-96 h, followed by incubation with 1 μg/ml of bovine hsp70 for 2 h in the presence of 5 mM Na-ATP in PBS containing 0.05% Tween 20 and 0.1% BSA (buffer C). After four washes with 0.05% Tween 20 in PBS (10 min/wash), the membrane was incubated with anti-hsp70 antibody in buffer C for 2 h, washed 3 times, and then incubated with peroxidase-conjugated goat anti-mouse IgG. Bound hsp was visualized using ECL.RESULTSAssociation of K8/18 with HSP70In carrying out immunoprecipitations of K8/18 from the colonic epithelial cell line HT29, we consistently noted co-immunoprecipitation of a 70-kDa band with K8/18. This is exemplified in Fig. 1A (lanea), where a 70-kDa species is noted to co-immunoprecipitate with K8/18 using our high capacity anti-K8/18 MAb L2A1. As a control, immunoprecipitation of other antigens using well described antibodies to Ep-CAM and M2B3 antigens (Fig. 1A, lanesb and c, respectively) did not co-immunoprecipitate either K8/18 or the 70-kDa band. The Ep-CAM glycoprotein (Mr ∼ 38,000) is abundant in HT29 cells and can be visualized by Coomassie staining (Fig. 1A, laneb), whereas the M2B3 antigen (Fig. 1A, lanec) can only be visualized after radiolabeling (Omary et al., 1992c). Purification of the 70-kDa band shown in Fig. 1A followed by trypsinization, high pressure liquid chromatography separation of the generated peptides, and microsequencing of one of the peaks provided the sequence TTPSYVAFTDTE. This sequence showed 100% homology with amino acids 37-48 of human hsp70 and hsc70 proteins (Dworniczak and Mirault, 1987). The association of K8/18 with hsp/c70 by co-immunoprecipitation was noted also in normal human colonic biopsies (Fig. 1A, lanef) and in all colonic tissue culture cell lines tested including T84, Caco2, and SK-CO-1 (not shown). The identity of a 160-kDa band that co-immunoprecipitated with K8/18 in normal colonic tissue (Fig. 1A, lanef, openarrowhead) was not investigated further. Another Coomassie-stained band that migrates between K8 and K18 (e.g.Fig. 1A, lanesa, e, and f) is usually seen with variable intensity and likely corresponds to degraded K8 or a related keratin (based on tryptic peptide mapping, not shown).To confirm the specificity of association of hsp/c70 with K8/18, we used three MAb that recognize different K8/18 epitopes. All three MAb resulted in co-immunoprecipitation of hsp/c70 with K8/18 (Fig. 1B, lanesb-d), indicating that hsp/c70 co-immunoprecipitation is not due to cross-reaction of MAb L2A1. In addition, the identity of the 70-kDa band as hsp70 and/or hsc70 was further confirmed by blotting of K8/18 immunoprecipitates using two different MAb to HSP70 proteins (Fig. 1B, lanese and f). We were not able to adequately test if anti-hsp/c70 antibodies co-immunoprecipitate K8/18 since the two antibodies we used did not efficiently immunoprecipitate hsp/c70 (not shown).Effect of Heat Stress on K8/18 Solubility and Association with HSP70We studied the effect of heat stress on K8/18 association with hsp/c70 and asked how the association partitions between the soluble and cytoskeletal K8/18 pools. As shown in Fig. 2A, the amount of HSP70 associated with K8/18 increased ∼ 2.5-fold after heat stress (lanes3 and 4), and this correlated with the increased expression of hsp/c70 (lanes1 and 2). This was confirmed by blotting of K8/18 immunoprecipitates and detergent lysates obtained from normal and heat-stressed cells using anti-HSP70 antibodies (not shown). Analysis of K8/18 immunoprecipitates obtained from the S and P fractions of HT29 cells (normalized to nearly equal amounts of immunoprecipitated keratins) showed increased association of hsp/c70 with K8/18 upon heat stress in both fractions (Fig. 2A, lanes5-8). Of note, a band indicated by the dottedarrow in lanes3, 6, and 8 is co-immunoprecipitated with K8/18 from heat-stressed cells. This band, which may represent a hyperphosphorylated form of K8 that was previously observed in mitotically arrested cells (Chou and Omary, 1994), becomes more noticeable with increasing duration of heat stress (not shown).Figure 2:Hsp/c70 associates preferentially with the soluble K8/18 fraction, with increased association, and with increased K8/18 solubility induced by heat stress. Panel A, HT29 cells (37°C) or heat-stressed cells (42°C, 16 h) were solubilized in Nonidet P-40 followed by analysis using SDS-PAGE of the soluble lysates after pelleting (lanes1 and 2) or K8/18 immunoprecipitates obtained from the lysates (lanes3 and 4). Alternatively, cells (37 and 42°C) were disrupted by nitrogen cavitation followed by ultracentrifugation. The pellet (P fraction) and soluble supernatant (S fraction) were used for immunoprecipitation with MAb L2A1 (lanes5-8) as described under "Materials and Methods." Panel B, equal numbers of HT29 cells (37 and 42°C) were solubilized directly in Laemmli sample buffer (Total) or disrupted followed by ultracentrifugation and isolation of the S fraction and then SDS-PAGE analysis (lanesa-d). Equal fractions (based on cell number) of the S preparations were also used for immunoprecipitation (i.p.) of K8/18 (lanese and f). Duplicate samples were analyzed by Coomassie staining or by immunoblotting using MAb L2A1. +, indicates heat-stressed cells; -, indicates cells grown at 37°C.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Given that HSP70 association with K8/18 occurs preferentially with the soluble K8/18 pool with increased association noted after heat stress, we asked if heat stress results in an increase in the soluble K8/18 pool. As shown in Fig. 2B, the soluble K8/18 fraction increases after heat stress when analyzed by immunoprecipitation (lanese and f) or by Western blotting of the total soluble fraction before immunoprecipitation (lanesc and d) or after immunoprecipitation (lanese and f). For the analysis shown in Fig. 2B, equal numbers of cells grown at 37 and 42°C were used, which had nearly equal total keratin (Fig. 2B, lanesa and b).The molar stoichiometry of hsp/c70 associated with soluble K8/18 was ∼ 1:10 and 1:25 for cells grown at 42 and 37°C, respectively, as estimated after Coomassie staining using BSA standards (not shown). This assumes binding of hsp/c70 to either K8 or K18 and an average Mr of 70,000 for hsp/c and 50,000 for keratin. This stoichiometry decreases nearly 3-fold when analyzing the detergent-solubilized fraction (e.g.Fig. 1B or 2A).Binding of hsc70 and hsp70 to the Soluble and Cytoskeletal K8/18 FractionsThe 70-kDa K8/18-associated species were occasionally resolved in our gel system into two very closely migrating bands, thereby suggesting that both hsp70 and hsc70 associated with K8/18 (e.g.Fig. 2B, lanef). To confirm this, we used isoelectric focusing/SDS-PAGE and antibodies that recognized hsp70 or hsp/c70 to study K8/18 association with hsp/c70 after isolation of the complex from the soluble and cytoskeletal pellet fractions. As shown in Fig. 3, the anti-hsp70 MAb recognized a single hsp spot (panelsc and d), whereas the anti-hsp/c70 MAb recognized two spots corresponding to hsp70 and hsc70 (panelsa and b). The pI values for hsp70 and hsc70 were 5.8 and 5.5, respectively, which are similar to values previously reported for bovine hsp/c70 (Ungewickell, 1985). The hsc70 (Mr 73,000) migrated slightly slower than the hsp70 (Mr 72,000) species as noted by Coomassie staining (Fig. 3, panelse and f). Hence, both hsp70 and hsc70 associate with soluble and cytoskeletal K8/18.Figure 3:Two-dimensional gel and Western blot analysis of the associated 70-kDa heat shock proteins isolated from soluble and insoluble keratin pools. HT29 cells were disrupted by nitrogen cavitation followed by centrifugation to isolate the soluble and pellet fractions. K8/18-hsp/c70 co-immunoprecipitates were obtained from the S and P fractions followed by isoelectric focusing (IEF) in the first dimension and then SDS-PAGE in the second dimension. Two-dimensional gels were then analyzed by Western blotting using MAb BRM-22 (anti-hsp/c70) and MAb 1197 (anti-hsp70). Blots used in panelsa and b were also stained to confirm the position of hsc and hsp, which are indicated by arrows, and are shown in panelse and f.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Since immunoprecipitation examines steady state levels of partitioning between the S and P fractions, we examined the K8/18-hsp/c70 association in terms of newly synthesized K8/18 and hsp/c70. As shown in Fig. 4, the majority of newly synthesized K8/18 (after a 1-min pulse label) is seen in the soluble fraction and is also associated with some newly synthesized hsp/c70. Labeling for 5 or 10 min resulted in rapid redistribution of K8/18 from the soluble to the cytoskeletal pools, with similar association with hsp/c70 (Fig. 4). As noted in Fig. 2, the steady state level for the various labeling time points, as determined by Coomassie staining of the immunoprecipitates from the S and P fractions, showed preferential association with the soluble K8/18 pool (Fig. 4B). Labeling for 10 min followed by 1-20 h chase gave results that were similar to the 10-min time point in Fig. 4A (not shown). This indicated that the K8/18-hsp/c70 association occurs not only at the early stages of biosynthesis but also subsequent to biosynthesis.Figure 4:Hsp associates with K8/18 early during biosynthesis. HT29 cells were incubated in Met-free medium for 1 h and then labeled with [ 36 S]methionine for 1, 5, or 10 min (100 μCi/ml). After labeling, cells were immediately placed over ice and then rinsed with cold PBS containing 15 μg/ml unlabeled Met. Labeled cells were disrupted using nitrogen cavitation followed by isolation of the S and P fractions and then immunoprecipitation of K8/18-hsp/c70. Immunoprecipitates were analyzed by SDS-PAGE and Coomassie staining (panelB) followed by fluorography (panelA).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Effect of hsp70-K8/18 Binding on Filament AssemblyGiven recent findings that the HSP25-related α-crystallins inhibited glial fibrillary acidic protein and vimentin in vitro filament assembly in an ATP-independent manner (Nicholl and Quinlan, 1994), we tested the effect of purified hsp70 on K8/18 in vitro filament assembly. This was done by adding bovine hsp70 to K8/18 proteins during assembly (i.e. before dialysis of the 6 M urea, which allows for filament assembly) or after filaments have assembled. As shown in Fig. 5, addition of bovine hsp70 to K8/18-assembled filaments for 24 h followed by ultracentrifugation resulted in binding of a small fraction of hsp70 to the cytoskeletal pellet pool in a dose-dependent manner. To exclude the possibility of protein trapping during assembly, no binding of BSA to the formed filamentous fraction was noted (Fig. 5, Pfraction, lane5). Analysis of the formed filaments by electron microscopy showed no effect of hsp70 (or BSA) on filament assembly (not shown). Similarly, inclusion of hsp70 during filament assembly resulted in binding of some of the hsp70 to the insoluble K8/18 pool in a dose-dependent fashion (Fig. 6, panelA) and did not significantly interfere with filament assembly (Fig. 6, panelB). No binding of BSA to the filamentous fraction was noted (Fig. 6, panelA, lane1). As a control, incubation of hsp70 in 6 M urea, followed by dialysis (i.e. under filament assembly conditions) and then ultracentrifugation, maintains all the hsp70 in the soluble fraction (not shown). Of note, hsp70 did not significantly alter the partitioning of K8/18 into the soluble fraction when added after (Fig. 5) or during (Fig. 6) filament assembly.Figure 5:Assembled K8/18 filaments bind to hsp70. K8/18 was purified from HT29 cells using high salt extraction of cells to enrich for keratin proteins followed by purification using anion exchange chromatography (Mono-Q column) as described under "Materials and Methods." Purified keratins (15 μg in 100 μl of 50 mM Tris-HCl (pH7.5), 2 mM EGTA, containing 6 M urea) were dialyzed to remove the urea and allow filament assembly. Assembled filaments were incubated with the indicated amount of bovine brain hsp70 or BSA as a control protein (24 h, 22°C), followed by centrifugation to isolate the soluble (S) and insoluble (P) fractions (300,000 × g, 30 min). Equivalent fractions were then analyzed by SDS-PAGE.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 6:Hsp70 does not significantly alter solubility or in vitro assembly of keratin intermediate filaments. K8/18 were co-purified by ion exchange chromatography and then used for in vitro assembly (30 μg in 200 μl) alone or in the presence of the indicated amounts of BSA or bovine brain hsp70 in 6 M urea, 50 mM Tris-HCl (pH 7.5), 2 mM EGTA. Filament assembly was initiated by dialysis against 50 mM Tris-HCl, 2 mM EGTA (pH 7.5) for 36 h with buffer change every 12 h. Dialyzed samples were analyzed by electron microscopy (panelB) or were pelleted (300,000 × g, 30 min) followed by SDS-PAGE analysis of equivalent portions of the S and P fractions (panelA). Filament assembly was similar for all the treatments and is represented by the two micrographs shown in panelB. Smallarrow shows hsp70 particles, and bar = 60 nm. Similar hsp70 particles were noted by electron microscopy if hsp70 in the 6 M urea containing buffer, in the absence of K8/18, was dialyzed under assembly conditions (not shown).View Large Image Figure ViewerDownload Hi-res
Referência(s)