Hsp27 Regulates Akt Activation and Polymorphonuclear Leukocyte Apoptosis by Scaffolding MK2 to Akt Signal Complex*

We have shown previously that Akt exists in a signal complex with p38 MAPK, MAPK-activated protein kinase-2 (MK2), and heat shock protein 27 (Hsp27) and MK2 phosphorylates Akt on Ser-473. Additionally, dissociation of Hsp27 from Akt, prior to Akt activation, induced polymorphonuclear leukocyte (PMN) apoptosis. However, the role of Hsp27 in regulating Akt activation was not examined. This study tested the hypothesis that Hsp27 regulates Akt activation and promotes cell survival by scaffolding MK2 to the Akt signal complex. Here we show that loss of Akt/Hsp27 interaction by anti-Hsp27 antibody treatment resulted in loss of Akt/MK2 interaction, loss of Akt-Ser-473 phosphorylation, and induced PMN apoptosis. Transfection of myristoylated Akt (AktCA) in HK-11 cells induced Akt-Ser-473 phosphorylation, activation, and Hsp27-Ser-82 phosphorylation. Cotransfection of AktCA with Hsp27 short interfering RNA, but not scrambled short interfering RNA, silenced Hsp27 expression, without altering Akt expression in HK-11 cells. Silencing Hsp27 expression inhibited Akt/MK2 interaction, inhibited Akt phosphorylation and Akt activation, and induced HK-11 cell death. Deletion mutagenesis studies identified acidic linker region (amino acids 117–128) on Akt as an Hsp27 binding region. Deletion of amino acids 117–128 on Akt resulted in loss of its interaction with Hsp27 and MK2 but not with Hsp90 as demonstrated by immunoprecipitation and glutathione S-transferase pulldown studies. Co-transfection studies demonstrated that constitutively active MK2 (MK2EE) phosphorylated Aktwt (wild type) on Ser-473 but failed to phosphorylate AktΔ117–128 mutant in transfixed cells. These studies collectively define a novel role of Hsp27 in regulating Akt activation and cellular apoptosis by mediating interaction between Akt and its upstream activator MK2.

cytes (PMN) 3 take part in host defense mechanisms against infection and inflammatory diseases. Inappropriate termination of PMN activation or failure to remove apoptotic PMNs results in inflammation. This apoptotic process has been suggested to represent an in vivo mechanism limiting oxidant-induced tissue injury caused by PMNs at the sites of inflammation. Although PMNs are constitutively committed to apoptosis from the time they enter circulation, the rate of apoptosis is not fixed. We reported that interleukin-8, granulocytemacrophage colony-stimulating factor, LTB 4 , and bacterial lipopolysaccharide (LPS) delay constitutive PMN apoptosis through the activation of the serine/threonine kinase Akt (1,2). We demonstrated that p38 mitogen-activated protein kinase (MAPK) activity is required for Akt phosphorylation and activation (3). Additionally, we showed that Akt exists in a signaling module with p38 MAPK, MAPK-activated protein kinase-2 (MK2), and heat shock protein 27 (Hsp27) (3).
Heat shock proteins represent a group of chaperone proteins that protect the cells against a variety of stresses. Besides being involved in functioning as a chaperone, Hsp27 has also been shown to regulate stability of the cytoskeleton, cell motility (4 -7), and apoptosis (8 -13). When overexpressed in tumor cells, Hsp27 increases their tumorigenicity by overexpressing MMP-9 expression and down-regulating Src tyrosine kinase Yes expression (14 -16) and protects against apoptotic cell death triggered by various stimuli, including cytotoxic drugs and ligation of the Fas/Apo-1/CD95 death receptor (17)(18)(19). Mice overexpressing Hsp27 were protected from lethal ischemia/reperfusion injury compared with their negative littermates (20). Possible mechanisms of Hsp27 anti-apoptotic activity are proposed to result from its activity as a molecular chaperone. Hsp27 binds to and inactivates the pro-apoptotic molecules Smac, caspase 3, caspase 9, and cytochrome c (21)(22)(23)(24)(25). Hsp27-mediated suppression of Bid translocation to the mitochondria correlates with an inhibition of cytochrome c release (25). Hsp27 has also been shown to promote survival pathways by modulating IKK complex stability and activity. Parcellier et al. (26) demonstrated that Hsp27 mediates NF-KB activation and cell survival by promoting the proteasomal degradation of polyubiquitinated IKB. Phosphorylated Hsp27 has been shown to bind an adaptor protein Daxx and to inhibit Fas-mediated apoptosis (27). Additionally, phosphorylation of Hsp27 has been shown to be required for proper maintenance of cell adhesion and inhibition of renal epithelial cell apoptosis (28). Furthermore, Sheth et al. (13) showed that introduction of recombinant Hsp27 caused delay of PMN apoptosis; however, mechanisms regulating this delay of PMN apoptosis were not determined.
We recently demonstrated direct protein/protein interaction between Akt/Hsp27 (3,8). The physical association of Hsp27 with Akt is a critical determinant of PMN survival, as removal of Hsp27 from the Akt signal module prevented Akt phosphorylation and activation and resulted in accelerated PMN apoptosis suggesting an important role for Hsp27 in regulating Akt activity (8). MK2 has been shown to bind and phosphorylate Hsp27 (29), and MK2 is PDK2 for Akt in human PMNs (3). Recently, Zheng et al. (30) demonstrated that MK2 is required for p38 MAPK and Hsp27 interaction; however, association of Hsp27 and Akt was not dependent on MK2. Hence we hypothesized that Hsp27 regulates Akt activation and apoptosis by scaffolding MK2 to the Akt signal complex.
Akt contains an N-terminal pleckstrin homology (PH) domain and a catalytic kinase domain (residues 1-116 and 148 -411 respectively) linked by a highly acidic linker region (residues 117-147). The C-terminal tail region lies between residues 412 and 480. Phosphoinositides are known to bind the PH domain of Akt and recruit it to the plasma membrane for full activation by PDK1 and PDK2 (28,(31)(32)(33). In the present study we show that amino acids 117-128 within the acidic linker region of Akt are required for interaction with Hsp27. An in-frame deletion mutant of Akt (Akt ⌬117-128 ), lacking the Hsp27 binding region, interacts with Hsp90 but not with Hsp27 and MK2. Disruption of Akt/Hsp27 interaction prevents MK2 association with Akt, resulting in loss of MK2-mediated Akt Ser-473 phosphorylation, activation, and induction of apoptosis. These studies demonstrate for the first time that Hsp27 regulates Akt activation and cellular apoptosis by scaffolding MK2 to the Akt signal complex.
Isolation of PMNs and Culture Conditions-PMNs were isolated from venous blood obtained from healthy volunteers as described previously (3,8). PMN preparations routinely contained Ͼ95% PMNs, as determined by morphology, and were Ͼ99% viable by trypan blue dye exclusion. PMNs were suspended in RPMI 1640 medium supplemented with 10% fetal calf serum, L-glutamine, penicillin, and streptomycin and incubated for the indicated times at 37°C in 5% CO 2 .
Generation of pUseAkt ⌬117-128 Mutant by Site-directed Mutagenesis-In-frame deletion of amino acids 117-128 from pUseAktwt cDNA construct was carried out using the Transformer site-directed mutagenesis kit from BD Biosciences according to the manufacturer's instructions. The in-frame deletion primer was 5Ј-AGGCAGGAAGAAGAGTCAGGGG-CTGAAGAG-3Ј, and the selection primer for pUseAktwt (mutating the KpnI site) was 5Ј-GTTAAGCTTGAATCCGA-GCTCG-3Ј. Cloning and mutation were confirmed by DNA sequencing.
Gel Filtration Chromatography-The system used for the size exclusion chromatography experiments consisted of a HiPrep 26/60 Sephacryl S-300 HR prepacked chromatography column connected to an AKTA purifier 10 liquid chromatography system (Amersham Biosciences), equipped with a Frac-900 automated fraction collector, and controlled by the UNICORN version 4.00 software (Amersham Biosciences). Prior to chromatography, the column was equilibrated in 50 mM Tris-Cl, pH 7.4, 1 mM EDTA, 150 mM NaCl, 1.5 mM MgCl 2 , 5% glycerol, 0.5% Triton X-100. For the chromatographic separation of the samples, control and fMLP (0.3 M)-stimulated PMNs were harvested and resuspended in Akt lysis buffer containing 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% (v/v) Triton X-100, 0.5% (v/v) Nonidet P-40, 1 mM EDTA, 1 mM EGTA, 20 mM sodium orthovanadate, 10 M p-nitrophenol phosphate, 20 mM NaF, 5 mM PMSF, 21 g/ml aprotinin, and 5 g/ml leupeptin. The cell lysate was cleared by centrifugation at 14,000 rpm for 15 min. Next, 3-5 ml of total cleared cell lysate was injected onto the column with a manual injection through a 50-ml capacity superloop system (Amersham Biosciences). Isocratic elution with 2.5 column volumes of the same buffer was performed. Both a constant flow rate of 1.0 ml/min and the absorbance profiles at A 230nm and A 280nm were monitored during the entire chromatographic procedure. Fractions of 1.0 ml were collected, and 100 l of every fifth fraction was used for Western blot analysis. All procedures were performed at 4°C.
Isoelectric Focusing (IEF) Electrophoresis-PMN lysates (25 g) were subjected to IEF electrophoresis. Proteins were separated based on their isoelectric point. Precast IEF (NOVEX) gels (Invitrogen) were run according to manufacturer's instructions. Gels were fixed in 100 ml of 20% methanol solution for 20 min. Gel was rinsed in 1ϫ SDS-running buffer for 20 min and transferred onto nitrocellulose membrane using semidry transfer apparatus (Invitrogen) for 20 min at 20 V. Nitrocellulose membranes were washed with Krebsϩ buffer and then immunoblotted with anti-Hsp27 and anti-Ser(P)-82Hsp27 antisera.
Two-dimensional PAGE-Protein lysates (50 g) were diluted into urea/thiourea rehydration buffer (Genomic Solutions), and proteins were separated by two-dimensional PAGE. IPG strips, pH 3-10 (Invitrogen), were rehydrated overnight with protein samples. Proteins were separated on the basis of their isoelectric point by IEF using the ZOOM IPG Runner (Invitrogen) with a maximal voltage of 2000 V and 50 A per gel. Following IEF, IPG strips were incubated twice in equilibration buffer I (6 M urea, 130 mM dithiothreitol, 30% glycerol, 45 mM Tris base, 1.6% SDS, 0.002% bromphenol blue; Genomic Solutions) and once in equilibration buffer II (6 M urea, 135 mM iodoacetamide, 30% glycerol, 45 mM Tris base, 1.6% SDS, 0.002% bromphenol blue; Genomic Solutions) for 10 min. Equilibrated IPG strips were applied to 4 -12% BisTris gradient gels (Invitrogen), and proteins were separated in the second dimension based on their molecular size using NuPAGE MES/ SDS buffer (Invitrogen) at 200 V for 40 min. Following electrophoresis, gels were transferred onto nitrocellulose and immunoblotted with anti-Hsp27 antibody.
GST Pulldown Assay-PMNs (2 ϫ 10 7 ) were untreated or treated with isotype control antibody or Hsp27 antibody for 2 h at 37°C. The cells were lysed with 200 l of lysis buffer containing 1% (v/v) Nonidet P-40, 10% (v/v) glycerol, 137 mM NaCl, 20 mM Tris-HCl, pH 7.4, 1 g/ml aprotinin, 1 g/ml leupeptin, 5 mM phenylmethylsulfonyl fluoride, 20 mM NaF, 1 mM sodium pyrophosphate, 1 mM sodium orthovanadate, and 1% (v/v) Triton X-100. Appropriate GST beads (10 l) were added to the lysates and incubated at 4°C for 1 h with shaking. The beads were washed three times with Krebs buffer, and 15 l of 2ϫ Laemmli buffer was added to each tube. The samples were boiled for 3 min and then subjected to 10% SDS-PAGE. Proteins were transferred onto nitrocellulose and immunoblotted with appropriate antibodies.
GST Pulldown Assay with Recombinant Proteins-Appropriate GST beads (10 l) were added to 50 l of kinase buffer (20 mM HEPES, 10 mM MgCl 2 , 10 mM MnCl 2 ) containing 50 ng of appropriate recombinant protein. The samples were incubated at 4°C for 1 h with shaking. The beads were washed three times with Krebs buffer, and 15 l of 2ϫ Laemmli buffer was added to each tube. The samples were boiled for 3 min and then subjected to 10% SDS-PAGE and immunoblotting. Recombinant GST-Akt-PH domain (1-149 amino acids) was first conjugated with glutathione-Sepharose beads by incubating the two at 4°C for 1 h with shaking. These beads were then incubated with 50 ng of recombinant Hsp27 as described above.
cDNA Transfections-HK-11or HEK-293 cells were plated on 6-well trays a day prior to performing transfections, to achieve 60% confluence. On the day of transfection appropriate cells were washed with serum-free RPMI 1640 medium. One g of appropriate cDNA was transfected into these cells using GenePORTER reagent according to the manufacturer's protocol (Gene Therapy Systems). Twenty four hours after transfection cells were lysed in 100 l of Akt lysis buffer containing 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% (v/v) Triton X-100, 0.5% (v/v) Nonidet P-40, 1 mM EDTA, 1 mM EGTA, 20 mM sodium orthovanadate, 10 M p-nitrophenol phosphate, 20 mM NaF, 5 mM PMSF, 21 g/ml aprotinin, and 5 g/ml leupeptin, and proteins concentrations were determined. Protein lysates (50 g) were subjected to SDS-PAGE and immunoblot analysis or to immunoprecipitation studies.
Assessment of Cell Viability-Cell viability was measured by assessment of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2Htetrazolium bromide (MTT) reduction (Sigma) as described previously (36,37). The soluble form of MTT was reduced by mitochondria of live cells, resulting in a water-insoluble salt. Product formation was monitored by reading absorbance at 540 nm using a microplate reader.
Akt Immunoprecipitation Assays-Appropriate cells were lysed in Akt lysis buffer containing 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% (v/v) Triton X-100, 0.5% (v/v) Nonidet P-40, 1 mM EDTA, 1 mM EGTA, 20 mM sodium orthovanadate, 10 M p-nitrophenol phosphate, 20 mM NaF, 5 mM PMSF, 21 g/ml aprotinin, and 5 g/ml leupeptin. Following centrifugation at 15,000 ϫ g for 15 min at 4°C, cleared lysates were incubated with 20 l of anti-Akt PH domain agarose beads or with mouse isotype control antibody beads as described previously (8). Immunoprecipitated proteins were eluted with 40 l of 2ϫ Laemmli dye. Samples were boiled for 2 min; beads were precipitated by a quick spin in a picofuge, and 40 l supernatant containing eluted proteins was subjected to SDS-PAGE and immunoblot analysis.
Akt Kinase Assay-HK-11 cells were transfected with pUSE vector or pUSEAktCA (c-Myc-tagged myristoylated constitutively active Akt; Upstate Biotechnology, Inc.) or pUSEAktCA along with Hsp27siRNA. Transfected lysates were subjected to anti-Akt immunoprecipitation (as described above). Immunoprecipitated Akt was subjected to an in vitro Akt kinase assay using histone H2B as substrate as described previously (8).
Annexin V binding was performed as described previously (8). Briefly, 1 ϫ 10 6 PMNs were washed and resuspended in 100 l of RPMI 1640 medium without fetal calf serum. The cells were prewarmed for 5 min at 37°C; anti-Hsp27 antibody (10 g) or isotype control antibody (10 g) was then added, and the cells were further incubated for 2 h at 37°C. PMNs were washed once with RPMI 1640 medium without fetal calf serum and centrifuged at 400 ϫ g for 2 min. Cells were resuspended in ice-cold 1ϫ Binding Buffer (10 mM Hepes/NaOH, pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2.5 mM CaCl 2 ) at a concentration of 10 6 cells/0.3 ml. Next 3 l of annexin conjugate ApopNexin FITC was added to the cell suspension and incubated at room temperature for 15 min in the dark. The cells were immediately examined using Zeiss Axiovert 100 microscope and LSM 510 software.

Enhanced Akt/Hsp27 Interaction during LPS Delay of PMN
Apoptosis-Cultured PMNs undergo constitutive apoptosis such that only about 30% are viable after 24 h. However, treatment of PMNs with lipopolysaccharide (LPS) delays constitutive PMN apoptosis, such that about 70% are viable after 24 h (2). We determined whether increased PMN viability at 24 h in the presence of LPS could be attributed to increased Akt/Hsp27 interaction. Freshly isolated PMNs (zero time control) and PMNs in culture for 24 h with or without LPS were lysed. Lysates were subjected to anti-Akt immunoprecipitation and immunoblotted with anti-Hsp27 antibody. Immunoprecipitates were also immunoblotted with anti-Akt to determine equal immunoprecipitation in every condition. Additionally, as control, zero time PMN lysates were subjected to isotype control antibody immunoprecipitation and immunoblotted with anti-Hsp27 and anti-Akt antisera (Fig. 1, 4th lane). As expected, no Akt or Hsp27 binding was detected in PMN lysates immunoprecipitated with isotype control antibody (Fig. 1, 4th lane). Akt/Hsp27 interaction was detected in zero time PMN lysates subjected to anti-Akt immunoprecipitation (Fig. 1, 3rd lane). Akt-Hsp27 association was markedly reduced in PMNs cultured for 24 h in the absence of LPS (Fig. 1, 2nd lane); however, this association was maintained in PMNs cultured for 24 h in the presence of LPS (Fig. 1, 1st lane). These results suggest there is a need to determine the cause and effect relationship between Akt/Hsp27 interaction and PMN viability, suggesting the importance of Akt/Hsp27 interaction in the regulation of PMN apoptosis during inflammation.
Separation of Akt Signal Components in Control and fMLPstimulated PMNs-We have previously shown in PMNs that Akt exists in a complex with Hsp27, MK2, and p38 MAPK and that p38 MAPK-dependent kinase MK2 regulates Akt activation by phosphorylating Ser-473 on Akt (8). In addition, we demonstrated that upon fMLP stimulation, Hsp27 dissociates from the Akt signal complex, whereas MK2 and p38 MAPK continue to associate with Akt (3,8). To estimate percentage of Hsp27 or MK2 or p38 MAPK from PMNs that associate with Akt, we subjected control and fMLP-stimulated PMN lysates to gel filtration chromatography. Fractions collected were immunoblotted with anti-Hsp27, anti-MK2, anti-p38 MAPK, and anti-Akt ( Fig. 2A). The results demonstrated that Hsp27 eluted in the same fractions as Akt in control PMN lysates. In addition, MK2 and p38 MAPK eluted in the same fractions as Hsp27 and Akt; however, both MK2 and p38 were detected in additional fractions in the absence of Akt and Hsp27. Furthermore, p38 MAPK was detected in some fractions in the absence of Akt, Hsp27, and MK2. These results suggest that under control conditions, almost 100% of Hsp27 and Akt associate with MK2, whereas a small percentage of MK2 exists in a complex with p38 MAPK in the absence of Akt and Hsp27 (Fig. 2A). These results Lysates from human PMNs freshly isolated or in culture for 24 h in the presence and absence of LPS (100 ng/ml) were subjected to anti-Akt immunoprecipitation (IP) and immunoblotted (IB) with anti-Hsp27 antibody. Lysates from freshly isolated PMNs were also subjected to isotype control immunoprecipitation as control. As a control, immunoprecipitates were immunoblotted with anti-Akt antibody to demonstrate equal amounts of immunoprecipitations in each condition.
were confirmed by performing Akt immunoprecipitations and immunoblotting with anti-Hsp27 and anti-MK2 antisera (data not shown). In fMLP-stimulated PMNs, we demonstrated that Hsp27 eluted out prior to Akt (Fig. 2B). Thus, only a small fraction of Hsp27 associates with Akt, whereas a large amount of Hsp27 was not in a complex with Akt in fMLP-stimulated PMNs. MK2 and p38 MAPK continue to associate with Akt after fMLP stimulation.
Phosphorylation Induces an Acidic Shift in Hsp27 pI without Altering Its Molecular Size in Human PMNs-The oligomerization status of Hsp27 in human PMNs has not been demonstrated to date. To determine whether Hsp27 existed at different isoelectric points by the virtue of its phosphorylation, control and fMLP-stimulated PMN lysates were subjected to IEF gel electrophoresis and immunoblotting with anti-Hsp27 antibody. The results demonstrated two species of Hsp27 with different isoelectric points in both control and fMLP-stimulated PMN lysates. However, the intensity of the more acidic form of Hsp27 was enhanced in fMLP-stimulated lysate compared with control lysate (Fig. 3A, left panel). Immunoblot analysis with anti-Ser(P)-82Hsp27 antibody confirmed that the increase in the acidic form of Hsp27 correlated with enhanced Hsp27 Ser-82 phosphorylation (Fig. 3A, right panel). Having demonstrated that two forms of Hsp27 exist at different isoelectric points, we next sought to determine the size of these two forms of Hsp27. To accomplish this goal, PMN lysates were subjected to two-dimensional PAGE. PMN proteins were separated in the first dimension on pI and separated on the second dimension by the virtue of their molecular weight or size. These two-dimensional polyacrylamide gels were subsequently subjected to immunoblotting with anti-Hsp27 antibody. Results demonstrated two species of Hsp27; however, both these species were identical in size (Fig. 3B). Collectively, IEF and twodimensional PAGE Hsp27 immunoblot analysis of control and fMLP-stimulated PMNs lysates demonstrated that phosphorylation of Hsp27 in PMNs induced a shift in pI without altering size of Hsp27. Thus, under these current experimental conditions, Hsp27 does not exist as an oligomeric protein in PMNs.
Hsp27 Depletion Induces PMN Apoptosis by Inhibiting Akt Association with Its Upstream Kinase MK2-We have previously shown that loss of Akt/Hsp27 association in human Eluted protein fractions were collected as described under "Materials and Methods." Protein fractions were subjected to SDS-PAGE and immunoblotting with anti-Akt, anti-Hsp27, anti-MK2, and anti-p38 MAPK antisera. Results show that Akt and Hsp27 elute out in the same fractions in control PMN lysates, although a majority of Hsp27 elutes out prior to Akt in fMLP-treated PMN lysates, suggesting its dissociation from Akt. MK2 and p38 MAPK associate with Akt in the presence and absence of fMLP. FIGURE 3. Phosphorylation induces an acidic shift in Hsp27 pI without altering its molecular size in human PMNs. A, human PMN lysates generated from control (C) and 5-min fMLP (F5) (0.3 M)treated PMNs were subjected to IEF electrophoresis and immunoblotted (IB) with anti-Hsp27 and anti-Ser(P)-82Hsp27 antisera as described under "Materials and Methods." IEF immunoblotting demonstrated two species of Hsp27 with different isoelectric points in both control and fMLP-stimulated PMN lysates. Increased abundance of the more acidic species of Hsp27 was seen in fMLP-treated lysates. Immunoblot analysis with anti-Ser(P)-82Hsp27 antibody confirmed that the increase in the acidic form of Hsp27 correlated with enhanced Hsp27-Ser82 phosphorylation. B, control and 5-min fMLP (0.3 M)-treated PMN lysates were subjected to two-dimensional PAGE (2D-PAGE) immunoblot analysis with anti-Hsp27 antibody. Results demonstrated two species of Hsp27 as seen in IEF gels; however, both these species were of identical size. Collectively, these results demonstrate that phosphorylation of Hsp27 induced an acidic pI shift of Hsp27 without altering its size in human PMNs.
PMNs results in loss of Akt activation and PMN apoptosis (8).
To determine whether loss of Akt phosphorylation and activation in the absence of Hsp27 results from loss of interaction between Akt and its upstream activator kinase MK2, PMNs were incubated in the presence or absence of anti-Hsp27 antibody or isotype control antibodies as described previously (8). Successful introduction of FITC-tagged antibodies into PMNs was confirmed by confocal microscopy and trypan blue quenching as described previously by our group (8). After 4 h of appropriate antibody incubation cell lysates generated were subjected to isotype control or anti-Akt immunoprecipitation to validate disruption of Akt/Hsp27 interaction in the presence of anti-Hsp27 antibody. Fig. 4A demonstrates association of Akt/Hsp27 in isotype control-treated cells but not in cells treated with anti-Hsp27 antibody (2nd and 3rd lanes, top  panel). In addition, as expected nonspecific binding was not detected in isotype control immunoprecipitation (Fig. 4A, 1st  lane, top panel). Equal Akt immunoprecipitation was documented by immunoblotting with anti-Akt antibody (Fig. 4A,  2nd and 3rd lanes, bottom panel). Having demonstrated loss of Akt/Hsp27 interaction in anti-Hsp27-treated cells, we next sought to determine association of Akt/MK2 in the presence and absence of Akt/Hsp27 interaction. PMN lysates generated above were subjected to GST or GST-Aktwt pulldown assay. Proteins were resolved by SDS-PAGE followed by immunoblot analysis with anti-MK2 antibody (Fig. 4B). Fig. 4B demonstrates association of Akt/MK2 in isotype control antibody-treated but not in anti-Hsp27 antibody-treated PMN lysates. No binding was detected in control GST beads (Fig. 4B, 2nd and 4th lanes). Isotype control antibody-treated PMN lysate was run as a positive control (Fig. 4B, 5th lane). These results suggest that disruption of Akt-Hsp27 association results in loss of Akt interac-tion with the Hsp27-interacting protein MK2. As demonstrated previously, incubation of PMNs with anti-Hsp27 but not isotype control antibody induced apoptosis as shown by enhanced annexin V binding in anti-Hsp27-treated PMNs (Fig. 4C). These results suggest that disruption of Akt/Hsp27 association results in loss of Akt/ MK2 and induction of PMN apoptosis.
Silencing Hsp27 Expression Disrupts Akt/MK2 Interaction-We further validated the role of Hsp27 in scaffolding MK2 to Akt and regulating cellular apoptosis by an Hsp27 knockdown approach. Effects of silencing Hsp27 on Akt/MK2 interaction were determined. Cell lysates from HK-11 cells transfected with scrambled or Hsp27 siRNA were immunoblotted with anti-Hsp27 and anti-Akt antisera to demonstrate specific silencing of Hsp27 expression. Fig. 5A shows that Hsp27 siRNA but not scrambled siRNA significantly inhibited Hsp27 expression without altering Akt expression. GST pulldown studies performed with above lysates demonstrated Akt/ MK2 interaction in scrambled siRNA-treated cells but not in Hsp27 siRNA-treated cells (Fig. 5B). As expected no binding was detected in GST-Sepharose control (Fig. 5B, lane 3). These results were confirmed by Akt immunoprecipitation studies in scrambled siRNA and Hsp27 siRNA-transfected HK-11 cells (Fig. 5C). Collectively these results demonstrate an important role of Hsp27 expression in regulating Akt/MK2 interaction ( Fig. 5B and Fig. 3C).
Disruption of Akt/MK2 Interaction by Silencing Hsp27 Expression Inhibits Akt Phosphorylation and Activation-Because MK2 acts as PDK2 for Akt, we next determined whether loss of Akt/MK2 interaction would regulate Akt activation. HK-11 cells were transfected with vector or c-Myc-tagged constitutively active myristoylated Akt (AktCA) in the presence of scrambled siRNA or Hsp27 siRNA. AktCA cDNA construct is myristoylated and hence targeted to the plasma membrane allowing constitutively activated PDK2 to phosphorylate Akt on Ser-473. Fig. 6A shows that Hsp27 siRNA selectively silenced Hsp27 expression without altering c-Myc-tagged Akt expression (Fig. 6A, 2nd lane, panels 1 and 2), whereas scrambled siRNA had no effect on Hsp27 or Akt expression (Fig. 6A,  1st lane, panels 1 and 2). Hsp27 siRNA but not scrambled siRNA inhibited AktSer-473 phosphorylation of transfected myristoylated Akt and also inhibited AktCA-induced Hsp27 Ser-82 phosphorylation (Fig. 6A, 2nd lane, panels 3 and 4). Additionally, Hsp27 siRNA but not scrambled siRNA inhibited the ability of immunoprecipitated Akt to phosphorylate histone H2B from AktCA transfected HK-11 cells (Fig. 6B). Loss of Akt/MK2 interaction and subsequent loss of Akt phosphoryla- . Hsp27 depletion induces PMN apoptosis by inhibiting Akt association with its upstream kinase MK2. A, human PMNs were treated with isotype control (IC) antibody or Hsp27 antibody for 4 h at 37°C as described previously (8). PMN lysates were subjected to anti-Akt or isotype control immunoprecipitation (IP) followed by immunoblot (IB) analysis with anti-Hsp27 antibody. Results demonstrate Akt/Hsp27 interaction in isotype control-treated PMNs but not in anti-Hsp27 treated PMNs (n ϭ 3). PMN lysates obtained above were incubated with GST or GST-Akt-Sepharose. B, proteins were separated by SDS-PAGE and immunoblotted with anti-MK2 antibody (n ϭ 3). Isotype control-treated PMN lysate was included as a positive control for MK2 (5th lane). Results show that depletion of Hsp27 inhibits Akt/MK2 interaction. C, PMNs were treated with mouse isotype control antibody or monoclonal anti-Hsp27 antibody for 4 h at 37°C. Isotype control and anti-Hsp27 antibody-treated cells were stained with FITC-conjugated annexin V and viewed by confocal microscopy. Annexin V-positive staining was observed in anti-Hsp27 antibody-treated cells. JULY 27, 2007 • VOLUME 282 • NUMBER 30 tion and activation in the presence of Hsp27 siRNA suggest a critical role for Hsp27 in the regulation of Akt activation.

Regulation of Akt Activation by Hsp27
Disruption of Akt/MK2 Interaction by Silencing Hsp27 Expression Induces HK-11 Cell Death-Silencing Hsp27 expression inhibits Akt/MK2 interaction and Akt activation. Thus, we determined effects of silencing Hsp27 expression on HK-11 cell viability.
HK-11 cells alone or HK-11 cells in the presence of Gene-PORTER reagent, scrambled siRNA, or Hsp27 siRNA were incubated at 37°C in 5% CO 2 for 24 h, and cells were subjected to MTT cell viability assay. Hsp27 siRNA but not scrambled siRNA significantly inhibited Hsp27 without altering Akt expression (Fig. 7, bottom panel). Additionally, Hsp27 siRNA but not scrambled siRNA significantly decreased HK-11 cell viability. GenePORTER reagent by itself had no effect on cell viability (Fig. 7, top panel). These results suggest that Hsp27 is a survival protein, and its regulation of HK-11 survival may be mediated by scaffolding MK2 to Akt, allowing Akt activation to occur, thereby promoting cell survival.
Akt PH Domain Is Not Required for Interaction with Hsp27-We have demonstrated previously that Hsp27 directly interacts with Akt (8). To identify Hsp27-binding site(s) on Akt, we determined the ability of GST-Akt-(1-149)-Sepharose, which contains the Akt PH domain (1-116 amino acids) and the acidic linker region (117-149) to interact with recombinant Hsp27. Recombinant GST-Akt-(1-149) was conjugated to glutathione-Sepharose as described under "Materials and Methods." Fig. 8A demonstrates that GST-Akt-(1-149) (2nd lane), but not glutathione-Sepharose (1st lane), specifically interacts with recombinant Hsp27 suggesting a role for the PH domain and/or the acidic linker region to interact with Hsp27. Recombinant GST-Akt-(1-149) and recombinant Hsp27 were run as positive controls (Fig. 8A, 3rd and 4th lanes). To determine the role of Akt PH domain in interacting with Hsp27, recombinant Akt ⌬PH (mutant lacking N-terminal PH domain or 1-116 amino acids) was precipitated with GST or GST-Hsp27. Fig. 8B demonstrates interaction of recombinant Akt ⌬PH with GST-Hsp27 but not with GST-Sepharose. Recombinant Akt ⌬PH was run as positive control (Fig. 8B, lane 3). These results suggest that Akt-PH domain is not required for its interaction with Hsp27.  3). B, above lysates were subjected to GST or GST-Aktwt pulldown assay to determine Akt/MK2 interaction in the presence of scrambled or Hsp27 siRNA (n ϭ 3). C, above lysates were also subjected to mouse anti-Akt immunoprecipitation (IP) and immunoblotted with rabbit polyclonal anti-MK2 and rabbit polyclonal anti-Akt antisera (n ϭ 3). The results demonstrate loss of Hsp27 expression abrogates Akt/MK2 interaction. Lysates were immunoprecipitated with anti-Akt antibody and subjected to an in vitro Akt kinase assay using histone H2B as substrate (n ϭ 3). Autoradiograph demonstrates that transfection of AktCA induces increased Akt-mediated histone H2B phosphorylation in the presence of scrambled siRNA but not in the presence of Hsp27siRNA. These results demonstrate that silencing Hsp27 expression inhibits Akt phosphorylation and activation.
Hsp27 Directly Interacts with the Acidic Linker Region of Akt-In-frame deletion mutant of Akt was generated such that amino acids 117-128 from the acidic linker region of Akt were deleted. In a GST pulldown assay recombinant Hsp27 was precipitated by GST-Aktwt but not by GST or GST-Akt ⌬117-128 -Sepharose (Fig. 9A, bottom panel). Transferred gel was stained with Coomassie Blue to serve as a loading control for GST beads utilized in the assay. Additionally, HEK-293 cells were co-transfected with c-Myc-tagged pUseAktwt or pUseAkt ⌬117-128 along with Hsp27 cDNA. After transfection cells were lysed, and lysates were subjected to anti-c-Myc immunoprecipitation and immunoblotted with anti-Hsp27 antibody. Fig. 9B (bottom panel) demonstrates interaction between Aktwt and Hsp27, whereas no interaction is detected between Akt ⌬117-128 and Hsp27. To demonstrate overexpression of c-Myc-tagged Akt constructs and Hsp27, total lysates were immunoblotted with anti-c-Myc and anti-Hsp27 antisera (Fig. 9B, top and middle panel). Furthermore, these results were confirmed in human PMNs by performing a GST pulldown assay. GST-Akt but not GST or GST-Akt ⌬117-128 interacted with Hsp27 from PMN lysates (Fig. 9C). Transferred gels were stained with Coomassie Blue to demonstrate equal loading of GST-Akt and GST-Akt ⌬117-128 -Sepharose. These results collectively indicate that amino acids 117-128 in the acidic linker region of Akt are required for interaction with Hsp27.
We next determined the ability of GST, GST-Aktwt, and GST-Akt ⌬117-128 to interact with Hsp90 from PMN lysates. Hsp90 has been shown previously to directly interact with Akt via amino acids 229 -309 (38). To confirm that deletion of 117-128 amino acids from Akt did not alter protein conformation and its ability to interact with Hsp90, we subjected GST, GST-Akt, and GST-Akt ⌬117-128 to a GST pulldown assay with PMN lysate. Fig. 9D demonstrates that as expected both GST-Aktwt and GST-Akt ⌬117-128 but not GST interacted with Hsp90 from PMN lysates. PMN lysate run as positive control immunoreacted with anti-Hsp90 antibody (Fig. 9D, lane 1). Transferred gels were stained with Coomassie Blue to demonstrate equal loading of GST-Akt and GST-Akt ⌬117-128 -Sepharose. These results collectively indicate that Hsp27 interacts with Akt via amino acids 117-128, and the deletion of this region does not severely alter its conformation and binding to its known binding protein Hsp90.
MK2 Fails to Interact with Akt ⌬117-128 and Phosphorylate It on Ser-473 ex Vivo-Having disrupted Akt/Hsp27 interaction by deleting Hsp27 binding region on Akt, we next determined the ability of GST, GST-Aktwt, and GST-Akt ⌬117-128 to precipitate MK2 from PMN lysates. Fig. 10A shows that GST-Aktwt, but not GST or GST-Akt ⌬117-128 , interacted with MK2 from PMN lysate. Transferred gels were stained with Coomassie Blue to demonstrate equal loading of GST-Akt and GST-Akt ⌬117-128 -Sepharose. Having shown loss of MK2 binding to Akt ⌬117-128 , we next determined the ability of constitutively activated MK2 (MK2EE), a known upstream activator of Akt, to phosphorylate Aktwt and Akt ⌬117-128 in transfected HEK-293 cells. Fig. 10B shows that MK2EE phosphorylates Aktwt on Ser-473 but fails to phosphorylate Akt ⌬117-128 . These results establish a novel role for Hsp27 in regulating Akt activation by scaffolding MK2 to Akt signal complex.

DISCUSSION
The first indication that Hsp27 might regulate Akt activity was reported by Konishi et al. (39), who demonstrated an association of Hsp27 and Akt in COS-7 cells. Subsequently, we demonstrated for the first time that Akt exists in a signaling complex with p38 MAPK, MK2, and Hsp27 in human PMNs (3). More recently p38 MAPK-dependent activation of Akt has been demonstrated in a variety of cells, including cardiomyocytes, keratinocytes, and endothelial cells (40 -47). Thus, p38 MAPK regulation of Akt activation is not unique to PMNs. In this study we show that Hsp27 directly interacts with Akt through its acidic linker region. The physical association of Hsp27 with Akt is a critical determinant of PMN survival, as removal of Hsp27 from the Akt signal module prevents Akt activation and results in accelerated PMN apoptosis (8). Moreover, MK2 in addition to phosphorylating Hsp27 (29) acts as PDK2 for Akt in human PMNs, phosphorylating Ser-473 (3).
Based on these studies, we postulated that Hsp27 regulates PMN apoptosis by acting as a scaffolding protein in the Akt signaling complex. Hsp27 binds to p38 MAPK, MK2, and other proteins required for activation of Akt. In this study we show for the first time relevance of Akt/Hsp27 interaction to constitutive apoptosis and LPS delay of apoptosis. Akt/Hsp27 interaction was documented in freshly isolated PMNs and was markedly reduced in PMNs cultured for 24 h, a time point at which only about 30% of PMNs are viable in culture (2). In contrast, Akt-Hsp27 association was markedly enhanced in PMNs cultured for 24 h in the presence of LPS. We have shown previously that LPS delays constitutive PMN apoptosis with 70% viable PMNs at 24 h (2). In addition, we have shown that Hsp27 is a member of the Akt signal complex, and removal of Hsp27 from the complex results in loss of Akt activation and induction of PMN apoptosis. These findings indicate a role for Hsp27 in the regulation of inflammation by modulation of Akt activation and PMN apoptosis.
Isolation of Akt signal complex in the presence and absence of fMLP by gel filtration chromatography demonstrated that Hsp27 dissociated from Akt upon fMLP stimulation as demonstrated previously by our group (3). In addition, IEF gel electrophoresis/immunoblot analysis demonstrated an acidic pI shift in Hsp27 in fMLP-stimulated PMNs compared with control, coinciding with Hsp27Ser-82 phosphorylation. Moreover, twodimensional PAGE immunoblotting demonstrated that this acidic pI shift in Hsp27 occurred without altering the molecular size of Hsp27. Thus, phosphorylation of Hsp27 does not alter the molecular size of Hsp27 in PMNs. Therefore, Hsp27 dissociation from Akt upon PMN activation is dependent on Hsp27 phosphorylation and not on the molecular size of Hsp27.
Heat shock proteins are induced under various stress conditions to combat stress-induced apoptosis caused by aggregation of denatured proteins. Heat shock proteins bind to and refold denatured proteins and chaperone them to appropriate cellular compartments or target them to the proteasome for degradation. Hsp27 binds actin and stabilizes the cellular architecture (48). The protective effect of Hsp27 on the cytoskeleton is mediated by the ability of Hsp27 to bind denatured actin and prevent its aggregation (49). Ariggo et al. (50) demonstrated FIGURE 9. Hsp27 directly interacts with the acidic linker region of Akt. A, GST, GST-Aktwt, or GST-Akt ⌬117-128 (117-128 amino acids deleted) was subjected to a pulldown assay with recombinant Hsp27 (50 ng). Transferred gel was stained with Coomassie Blue to serve as a loading control of GST beads utilized in the assay. Recombinant Hsp27 associated with GST-Aktwt but failed to bind GST or GST-Akt ⌬117-128 -Sepharose. B, HEK-293 cells were transfected with Hsp27 along with either c-Myc-tagged Aktwt or in-frame Akt deletion mutant Akt ⌬117-128 (117-128 amino acids deleted). Transfected lysates were immunoprecipitated (IP) with anti-c-Myc antibody and immunoblotted (IB) with anti-Hsp27 antibody (middle panel). Transfected lysates were also immunoblotted with anti-Hsp27 and anti-c-Myc antibodies to demonstrate expression of Hsp27 and c-Myctagged Akt constructs (top and bottom panels). C, PMN lysate was incubated with GST, GST-Akt, and GST-Akt ⌬117-128 and subjected to a GST pulldown assay and immunoblotted with anti-Hsp27 antibody (n ϭ 3). Transferred gels were stained with Coomassie Blue to demonstrate equal loading of GST-Akt and GST-Akt ⌬117-128 -Sepharose. D, PMN lysate was incubated with GST, GST-Akt, and GST-Akt ⌬117-128 and subjected to a GST pulldown assay and immunoblotted with anti-Hsp90 antibody (n ϭ 3). Transferred gels were stained with Coomassie Blue to demonstrate equal loading of GST-Akt and GST-Akt ⌬117-128 -Sepharose. Results show that Akt ⌬117-128 mutant binds Hsp90 but fails to interact with Hsp27. that Hsp27 inhibited apoptosis by maintaining the redox equilibrium of the cell. In addition, overexpression of Hsp27 inhibited oxidant stress-induced cellular damage by increasing levels of cellular glutathione (51). Overexpression of various heat shock proteins has been shown to induce a survival response. Contrary to most heat shock proteins, Hsp60 (52,53) has been shown to induce apoptosis by interaction with caspase-3, whereas Hsp27 (21-25), Hsp70 (54 -56), and Hsp90 (57) have been shown to inhibit apoptosis by binding to and inactivating specific components of the apoptosis machinery. In general heat shock protein-induced protection from stress was shown to act at the level of the apoptosome by preventing downstream activation of caspases. This study defines a novel role of Hsp27 in the regulation of cellular apoptosis by acting as a scaffolding protein.
We have recently demonstrated that Hsp27 exists in a complex with Akt, p38 MAPK, MK2, and Hsp27, and MK2 acts as PDK2 for Akt (3). In addition, we demonstrated that MK2binding protein Hsp27 dissociated from Akt signal complex after Akt activation (3). However, disruption of Akt/Hsp27 interaction prior to Akt activation resulted in loss of fMLPstimulated Akt activation and induced PMN apoptosis (8). These studies indicated the importance of Akt/Hsp27 interaction to Akt activation and PMN survival. Recently Zheng et al. (30) demonstrated that MK2 was required for p38 MAPK/ Hsp27 interaction. However, MK2 was not required for association of Hsp27 with Akt. Given that Akt/Hsp27 interaction was not regulated by MK2 and that MK2 serves as PDK2 for Akt, we hypothesized that Hsp27 modulated interactions within the Akt-Hsp27-MK2 signal complex by acting as a scaffolding protein. This study defined a novel scaffolding role of Hsp27 in the regulation of Akt activation and cellular apoptosis.
Disruption of Akt/Hsp27 interaction by depleting Hsp27 from human PMNs resulted in loss of Akt interaction with both Hsp27 and MK2 and induction of PMN apoptosis. Because MK2 is an Hsp27-binding protein, we next determined whether depletion of Hsp27 nonspecifically removed MK2 from the Akt signal complex. To address this question, siRNA technology was utilized to specifically silence Hsp27 expression from HK-11 cells to determine effects of silencing Hsp27 on Akt/ MK2 interaction and HK-11 cell viability. Hsp27 siRNA but not scrambled siRNA-transfected HK-11 cells demonstrated a marked decrease in Hsp27 expression, loss of Akt/MK2 interaction, and induction of HK-11 cell death. These results suggested a role for Hsp27 in regulating Akt/MK2 interaction and Akt activation. Cell death induced by silencing Hsp27 was not entirely surprising as Schepers et al. (58) have shown that silencing Hsp27 expression resulted in a 2-fold increase in VP-17-induced apoptosis in acute myeloid leukemia cells. However, induction of cell death by modulating Akt/MK2 interaction was a novel observation. To determine more directly the role of Hsp27 in regulating Akt activation, HK-11 cells were transfected with myristoylated-c-Myc-tagged Akt (AktCA) in the presence and absence of Hsp27 siRNA. Immunoblotting studies demonstrated that loss of Hsp27 expression inhibited AktSer-473 phosphorylation and activation. These studies clearly identified a role for Hsp27 in regulating Akt activation. Hsp27 negatively regulates protein kinase C␦ activity and inhibits cell apoptosis (59); however, we show for the first time that Hsp27 positively regulates Akt activation and promotes cell survival.
We have demonstrated direct protein/protein interaction between Hsp27 and Akt (3,8). This study identified amino acids 117-128 in the acidic linker region of Akt to mediate interaction with Hsp27. An in-frame Akt deletion mutant (Akt ⌬117-128 ) interacted with Hsp90␣ but failed to interact with Hsp27 and MK2. This finding is in accordance with the literature, which demonstrated that Akt interacted with Hsp90␣ via amino acids 229 -309 (38).
These studies suggested that loss of Akt ⌬117-128 interaction with MK2 and Hsp27 was not because of altered conformation of the mutant. The importance of Hsp27 in regulating Akt/ MK2 interaction and AktSer 473 phosphorylation was documented by co-transfecting HEK-293 cells with constitutively active MK2 (MK2EE) and Hsp27 along with Aktwt or Akt ⌬117-128 . MK2EE phosphorylated Aktwt on Ser 473 as expected; however, it failed to phosphorylate Akt ⌬117-128 on Ser 473 , a critical site known to regulate Akt activation (60). Furthermore, the physiologic consequence of loss of Akt/MK2 interaction and loss of Ser 473 phosphorylation resulted in cellular apoptosis. These studies collectively define a novel role of Hsp27 in regulating Akt activation and apoptosis by mediating interaction between Akt and its upstream activator MK2.