Hsp70-RAP46 interaction in downregulation of DNA binding by glucocorticoid receptor

Abstract

Receptor-associating protein 46 (RAP46) is a cochaperone that regulates the transactivation function of several steroid receptors. It is transported into the nucleus by a liganded glucocorticoid receptor where it downregulates DNA binding and transactivation by this receptor. The N- and C-termini of RAP46 are both implicated in its negative regulatory function. In metabolic labelling experiments, we have shown that the N-terminus of RAP46 is modified by phosphorylation, but this does not contribute to the downregulation of glucocorticoid receptor activity. However, deletion of a sequence that binds 70 kDa heat shock protein (Hsp70) and the constitutive isoform of Hsp70 (Hsc70) at the C-terminus of RAP46 abrogated its negative regulatory action. Surface plasmon resonance studies showed that RAP46 binds the glucocorticoid receptor only when it has interacted with Hsp70/Hsc70, and confocal immunofluorescence analyses revealed a nuclear transport of Hsp70/Hsc70 by the liganded receptor. Together these findings demonstrate an important contribution of Hsp70/Hsc70 in the binding of RAP46 to the glucocorticoid receptor and suggest a role for this molecular chaperone in the RAP46-mediated downregulation of glucocorticoid receptor activity.

Fig. 1. Mutation of the serine residues at the N-terminus of RAP46 does not abolish the inhibitory effect of RAP46 on DNA binding by the GR. (A) The sequences of the first 70 amino acids of the wild-type and the N-terminal serine mutant of RAP46 are illustrated. The amino acids exchanged are indicated by boxes. (B) Mutation of the serine residues at the N-terminus of RAP46 alters the secondary structure of the protein. Purified, bacterially produced fusion proteins GST–RAP46 and GST–RAP46mtSer were subjected to circular dichroism measurements. (C) Mutation of the N-terminal serine residues does not alter the negative effect of RAP46 on DNA binding by the GR. A hundred thousand COS-7 cells were transiently transfected with 0.9 µg of either an empty expression vector (vector) or an expression vector encoding glucocorticoid receptor (GR) and 0.9 µg of expression vectors encoding HA-tagged RAP46 or the mutant RAP46mtSer. One half of the transfected cells was used for EMSA and the other half was metabolically labelled with [³²P]orthophosphate for 4 h and immunoprecipitated with an anti-HA antibody. The immunoprecipitated proteins were resolved by SDS–PAGE and exposed to autoradiography (RAP46 phosphorylation) or immunoblotted using antibodies recognizing either the GR (GR immunoblot) or RAP46 as well as RAP46mtSer (RAP46 immunoblot). (D) Surfaces with an average of 2000 resonance units (RU) immobilized GST–RAP46 fusion proteins were obtained by covalent immobilization of an anti-GST antibody with the help of a standard amine coupling protocol, followed by capture of the different fusion proteins through antibody–antigen interaction. Injection of Hsp70 over the surface resulted in a strong interaction signal by 100 ng GST–RAP46 (1), whereas a decrease in the interaction signal was obtained when 100 ng GST–RAP46mtSer (2) were used for capture. No interaction was detected with 100 ng GST–RAP46dC47 (3).

Fig. 1. Mutation of the serine residues at the N-terminus of RAP46 does not abolish the inhibitory effect of RAP46 on DNA binding by the GR. (A) The sequences of the first 70 amino acids of the wild-type and the N-terminal serine mutant of RAP46 are illustrated. The amino acids exchanged are indicated by boxes. (B) Mutation of the serine residues at the N-terminus of RAP46 alters the secondary structure of the protein. Purified, bacterially produced fusion proteins GST–RAP46 and GST–RAP46mtSer were subjected to circular dichroism measurements. (C) Mutation of the N-terminal serine residues does not alter the negative effect of RAP46 on DNA binding by the GR. A hundred thousand COS-7 cells were transiently transfected with 0.9 µg of either an empty expression vector (vector) or an expression vector encoding glucocorticoid receptor (GR) and 0.9 µg of expression vectors encoding HA-tagged RAP46 or the mutant RAP46mtSer. One half of the transfected cells was used for EMSA and the other half was metabolically labelled with [³²P]orthophosphate for 4 h and immunoprecipitated with an anti-HA antibody. The immunoprecipitated proteins were resolved by SDS–PAGE and exposed to autoradiography (RAP46 phosphorylation) or immunoblotted using antibodies recognizing either the GR (GR immunoblot) or RAP46 as well as RAP46mtSer (RAP46 immunoblot). (D) Surfaces with an average of 2000 resonance units (RU) immobilized GST–RAP46 fusion proteins were obtained by covalent immobilization of an anti-GST antibody with the help of a standard amine coupling protocol, followed by capture of the different fusion proteins through antibody–antigen interaction. Injection of Hsp70 over the surface resulted in a strong interaction signal by 100 ng GST–RAP46 (1), whereas a decrease in the interaction signal was obtained when 100 ng GST–RAP46mtSer (2) were used for capture. No interaction was detected with 100 ng GST–RAP46dC47 (3).

Fig. 1. Mutation of the serine residues at the N-terminus of RAP46 does not abolish the inhibitory effect of RAP46 on DNA binding by the GR. (A) The sequences of the first 70 amino acids of the wild-type and the N-terminal serine mutant of RAP46 are illustrated. The amino acids exchanged are indicated by boxes. (B) Mutation of the serine residues at the N-terminus of RAP46 alters the secondary structure of the protein. Purified, bacterially produced fusion proteins GST–RAP46 and GST–RAP46mtSer were subjected to circular dichroism measurements. (C) Mutation of the N-terminal serine residues does not alter the negative effect of RAP46 on DNA binding by the GR. A hundred thousand COS-7 cells were transiently transfected with 0.9 µg of either an empty expression vector (vector) or an expression vector encoding glucocorticoid receptor (GR) and 0.9 µg of expression vectors encoding HA-tagged RAP46 or the mutant RAP46mtSer. One half of the transfected cells was used for EMSA and the other half was metabolically labelled with [³²P]orthophosphate for 4 h and immunoprecipitated with an anti-HA antibody. The immunoprecipitated proteins were resolved by SDS–PAGE and exposed to autoradiography (RAP46 phosphorylation) or immunoblotted using antibodies recognizing either the GR (GR immunoblot) or RAP46 as well as RAP46mtSer (RAP46 immunoblot). (D) Surfaces with an average of 2000 resonance units (RU) immobilized GST–RAP46 fusion proteins were obtained by covalent immobilization of an anti-GST antibody with the help of a standard amine coupling protocol, followed by capture of the different fusion proteins through antibody–antigen interaction. Injection of Hsp70 over the surface resulted in a strong interaction signal by 100 ng GST–RAP46 (1), whereas a decrease in the interaction signal was obtained when 100 ng GST–RAP46mtSer (2) were used for capture. No interaction was detected with 100 ng GST–RAP46dC47 (3).

Fig. 1. Mutation of the serine residues at the N-terminus of RAP46 does not abolish the inhibitory effect of RAP46 on DNA binding by the GR. (A) The sequences of the first 70 amino acids of the wild-type and the N-terminal serine mutant of RAP46 are illustrated. The amino acids exchanged are indicated by boxes. (B) Mutation of the serine residues at the N-terminus of RAP46 alters the secondary structure of the protein. Purified, bacterially produced fusion proteins GST–RAP46 and GST–RAP46mtSer were subjected to circular dichroism measurements. (C) Mutation of the N-terminal serine residues does not alter the negative effect of RAP46 on DNA binding by the GR. A hundred thousand COS-7 cells were transiently transfected with 0.9 µg of either an empty expression vector (vector) or an expression vector encoding glucocorticoid receptor (GR) and 0.9 µg of expression vectors encoding HA-tagged RAP46 or the mutant RAP46mtSer. One half of the transfected cells was used for EMSA and the other half was metabolically labelled with [³²P]orthophosphate for 4 h and immunoprecipitated with an anti-HA antibody. The immunoprecipitated proteins were resolved by SDS–PAGE and exposed to autoradiography (RAP46 phosphorylation) or immunoblotted using antibodies recognizing either the GR (GR immunoblot) or RAP46 as well as RAP46mtSer (RAP46 immunoblot). (D) Surfaces with an average of 2000 resonance units (RU) immobilized GST–RAP46 fusion proteins were obtained by covalent immobilization of an anti-GST antibody with the help of a standard amine coupling protocol, followed by capture of the different fusion proteins through antibody–antigen interaction. Injection of Hsp70 over the surface resulted in a strong interaction signal by 100 ng GST–RAP46 (1), whereas a decrease in the interaction signal was obtained when 100 ng GST–RAP46mtSer (2) were used for capture. No interaction was detected with 100 ng GST–RAP46dC47 (3).

Fig. 2. The Hsp70 binding domain is essential for RAP46-mediated downregulation of DNA binding by the GR. COS-7 cells were transiently transfected with 0.2 µg of GR-responsive MMTV-luciferase (pGL3MMTV) indicator gene and 10 ng of an internal control encoding the renilla luciferase under the control of the ubiquitin promoter. In addition, 0.9 µg of either an empty expression vector (vector) or an expression vector encoding glucocorticoid receptor (GR), or the mutant RAP46dC47 were cotransfected. Both RAP46 proteins were tagged with HA. The transfected cells were either left untreated or were metabolically labelled with [³²P]orthophosphate for 4 h. Whole-cell extracts were prepared for luciferase activity measurements (transactivation) or for EMSA and SDS–PAGE. In the latter case, the gels were immunoblotted using antibodies recognizing either the GR (GR immunoblot) or the HA tag on RAP46 or RAP46dC47 (RAP46 immunoblot). Immunoprecipitation with an anti-HA antibody was carried out on the cellular extracts from the metabolically labelled cells, followed by SDS–PAGE and autoradiography. The relative level of activation of the luciferase indicator plasmid in the transactivation assay was derived from the ratios of firefly over renilla luciferase activities obtained in three different experiments. For transactivation assays, dexamethasone (100 nM) was added to the cells immediately after transfection. For EMSA, the hormone was added to the reaction mixture before incubation of the cellular extracts with the radioactive DNA probe.

Fig. 3. Hsp70 binds to the GR in the absence or presence of hormone and its binding is required for the interaction of RAP46 with GR in SPR. (A) Sensograms showing injections of Hsp70 at 100 (1), 50 (2), 25 (3), 12.5 (4) and 5 ng/µl (5) with 1500 RU His-tagged GR produced in baculovirus coupled to the surface of a NTA chip. A dose-dependent increase in resonance units (Response, RU) can be seen. (B) SPR signals obtained when 100 ng of Hsp70 were injected in the absence (1) or presence (2) of 10^–7 M dexamethasone and 1500 RU His-tagged GR coupled to the NTA chip. A minor decrease in the interaction is observed in the presence of dexamethasone. (C) The consecutive addition of 100 ng RAP46 to a preformed GR–Hsp70 complex in the absence (1) or presence (2) of 10^–7 M dexamethasone resulted in an additional increase in RU, indicative of RAP46 binding. The values of the apparent rate constants for the Hsp70–GR interaction deduced from the experiments described in (A) were K_a = 5.4 × 10² M/s and K_d = 8.23 × 10^–4/s. From these data the apparent dissociation equilibrium constant was calculated to be K_d = 1.5 × 10^–6 M. All calculations were based on the assumption of monomolecular 1:1 binding characteristics, where A + B = AB. However, the apparent rate and dissociation equilibrium constants have to be seen as first approximations since a steady drift in the baseline (i.e. a time-dependent decline in GR binding; not shown) was observed. This is a well known phenomenon when working with non-covalent capturing of His-tagged proteins to the surface of a NTA sensor chip.

Fig. 3. Hsp70 binds to the GR in the absence or presence of hormone and its binding is required for the interaction of RAP46 with GR in SPR. (A) Sensograms showing injections of Hsp70 at 100 (1), 50 (2), 25 (3), 12.5 (4) and 5 ng/µl (5) with 1500 RU His-tagged GR produced in baculovirus coupled to the surface of a NTA chip. A dose-dependent increase in resonance units (Response, RU) can be seen. (B) SPR signals obtained when 100 ng of Hsp70 were injected in the absence (1) or presence (2) of 10^–7 M dexamethasone and 1500 RU His-tagged GR coupled to the NTA chip. A minor decrease in the interaction is observed in the presence of dexamethasone. (C) The consecutive addition of 100 ng RAP46 to a preformed GR–Hsp70 complex in the absence (1) or presence (2) of 10^–7 M dexamethasone resulted in an additional increase in RU, indicative of RAP46 binding. The values of the apparent rate constants for the Hsp70–GR interaction deduced from the experiments described in (A) were K_a = 5.4 × 10² M/s and K_d = 8.23 × 10^–4/s. From these data the apparent dissociation equilibrium constant was calculated to be K_d = 1.5 × 10^–6 M. All calculations were based on the assumption of monomolecular 1:1 binding characteristics, where A + B = AB. However, the apparent rate and dissociation equilibrium constants have to be seen as first approximations since a steady drift in the baseline (i.e. a time-dependent decline in GR binding; not shown) was observed. This is a well known phenomenon when working with non-covalent capturing of His-tagged proteins to the surface of a NTA sensor chip.

Fig. 3. Hsp70 binds to the GR in the absence or presence of hormone and its binding is required for the interaction of RAP46 with GR in SPR. (A) Sensograms showing injections of Hsp70 at 100 (1), 50 (2), 25 (3), 12.5 (4) and 5 ng/µl (5) with 1500 RU His-tagged GR produced in baculovirus coupled to the surface of a NTA chip. A dose-dependent increase in resonance units (Response, RU) can be seen. (B) SPR signals obtained when 100 ng of Hsp70 were injected in the absence (1) or presence (2) of 10^–7 M dexamethasone and 1500 RU His-tagged GR coupled to the NTA chip. A minor decrease in the interaction is observed in the presence of dexamethasone. (C) The consecutive addition of 100 ng RAP46 to a preformed GR–Hsp70 complex in the absence (1) or presence (2) of 10^–7 M dexamethasone resulted in an additional increase in RU, indicative of RAP46 binding. The values of the apparent rate constants for the Hsp70–GR interaction deduced from the experiments described in (A) were K_a = 5.4 × 10² M/s and K_d = 8.23 × 10^–4/s. From these data the apparent dissociation equilibrium constant was calculated to be K_d = 1.5 × 10^–6 M. All calculations were based on the assumption of monomolecular 1:1 binding characteristics, where A + B = AB. However, the apparent rate and dissociation equilibrium constants have to be seen as first approximations since a steady drift in the baseline (i.e. a time-dependent decline in GR binding; not shown) was observed. This is a well known phenomenon when working with non-covalent capturing of His-tagged proteins to the surface of a NTA sensor chip.

Fig. 4. In vitro cell-free system for studying the effect of RAP46 on DNA binding by the GR. One hundred thousand COS-7 cells in 3.4 cm culture dishes were transiently transfected by the Fugene transfection procedure with 0.9 µg empty expression vector or with an expression vector encoding the GR. Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR-response element as a probe. Where indicated, different amounts of bacterially purified GST or GST fusion proteins were added to the cell extracts. The fusion proteins consisted of GST alone or GST fused to either the wild-type RAP46 sequence (GST–RAP46) or RAP46 lacking the last 47 C-terminal amino acids (GST–RAP46dC47). (A) Inhibition of DNA binding by the GR is mediated by RAP46 and not RAP46C47. EMSA was carried out with cell extracts containing the GR and 1 µg GST, GST–RAP46 or GST–RAP46dC47. (B) Hsc70 relieves RAP46-mediated inhibition of DNA binding by the GR. COS-7 cells were transiently transfected with an empty expression vector (vector, lane 1), or an expression vector encoding either the GR (lanes 2–10) or an HA-tagged RAP46 (lanes 3–10). Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR response element as a probe. Where indicated, different amounts of purified Hsc70 were added to the whole-cell extracts. (C) Increased concentration of RAP46 downregulated DNA binding by the GR. Concentrations of GST–RAP46 ranging from 0.1 to 1000 ng was added to the cellular extract before the DNA binding reaction were carried out. (D) A RAP46–Hsc70 complex does not affect DNA binding by the GR. COS-7 cells were transiently transfected with either an empty expression vector (vector, lane 1) or an expression vector encoding the GR (lanes 2–5). Whole-cell extracts were prepared and subjected to EMSA. One microgram of bacterially produced GST (lanes 2 and 3) or GST–RAP46 fusion protein (lanes 4 and 5) were added to the cell extracts before providing the hormone and the probe. Where indicated, GST and GST–RAP46 were premixed with 2 µg of purified Hsc70 before addition to the cell extracts (lanes 3 and 5). In (A–D), dexamethasone (100 nM) was added to the EMSA reaction mixture prior to incubation with the radioactive probe.

Fig. 4. In vitro cell-free system for studying the effect of RAP46 on DNA binding by the GR. One hundred thousand COS-7 cells in 3.4 cm culture dishes were transiently transfected by the Fugene transfection procedure with 0.9 µg empty expression vector or with an expression vector encoding the GR. Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR-response element as a probe. Where indicated, different amounts of bacterially purified GST or GST fusion proteins were added to the cell extracts. The fusion proteins consisted of GST alone or GST fused to either the wild-type RAP46 sequence (GST–RAP46) or RAP46 lacking the last 47 C-terminal amino acids (GST–RAP46dC47). (A) Inhibition of DNA binding by the GR is mediated by RAP46 and not RAP46C47. EMSA was carried out with cell extracts containing the GR and 1 µg GST, GST–RAP46 or GST–RAP46dC47. (B) Hsc70 relieves RAP46-mediated inhibition of DNA binding by the GR. COS-7 cells were transiently transfected with an empty expression vector (vector, lane 1), or an expression vector encoding either the GR (lanes 2–10) or an HA-tagged RAP46 (lanes 3–10). Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR response element as a probe. Where indicated, different amounts of purified Hsc70 were added to the whole-cell extracts. (C) Increased concentration of RAP46 downregulated DNA binding by the GR. Concentrations of GST–RAP46 ranging from 0.1 to 1000 ng was added to the cellular extract before the DNA binding reaction were carried out. (D) A RAP46–Hsc70 complex does not affect DNA binding by the GR. COS-7 cells were transiently transfected with either an empty expression vector (vector, lane 1) or an expression vector encoding the GR (lanes 2–5). Whole-cell extracts were prepared and subjected to EMSA. One microgram of bacterially produced GST (lanes 2 and 3) or GST–RAP46 fusion protein (lanes 4 and 5) were added to the cell extracts before providing the hormone and the probe. Where indicated, GST and GST–RAP46 were premixed with 2 µg of purified Hsc70 before addition to the cell extracts (lanes 3 and 5). In (A–D), dexamethasone (100 nM) was added to the EMSA reaction mixture prior to incubation with the radioactive probe.

Fig. 4. In vitro cell-free system for studying the effect of RAP46 on DNA binding by the GR. One hundred thousand COS-7 cells in 3.4 cm culture dishes were transiently transfected by the Fugene transfection procedure with 0.9 µg empty expression vector or with an expression vector encoding the GR. Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR-response element as a probe. Where indicated, different amounts of bacterially purified GST or GST fusion proteins were added to the cell extracts. The fusion proteins consisted of GST alone or GST fused to either the wild-type RAP46 sequence (GST–RAP46) or RAP46 lacking the last 47 C-terminal amino acids (GST–RAP46dC47). (A) Inhibition of DNA binding by the GR is mediated by RAP46 and not RAP46C47. EMSA was carried out with cell extracts containing the GR and 1 µg GST, GST–RAP46 or GST–RAP46dC47. (B) Hsc70 relieves RAP46-mediated inhibition of DNA binding by the GR. COS-7 cells were transiently transfected with an empty expression vector (vector, lane 1), or an expression vector encoding either the GR (lanes 2–10) or an HA-tagged RAP46 (lanes 3–10). Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR response element as a probe. Where indicated, different amounts of purified Hsc70 were added to the whole-cell extracts. (C) Increased concentration of RAP46 downregulated DNA binding by the GR. Concentrations of GST–RAP46 ranging from 0.1 to 1000 ng was added to the cellular extract before the DNA binding reaction were carried out. (D) A RAP46–Hsc70 complex does not affect DNA binding by the GR. COS-7 cells were transiently transfected with either an empty expression vector (vector, lane 1) or an expression vector encoding the GR (lanes 2–5). Whole-cell extracts were prepared and subjected to EMSA. One microgram of bacterially produced GST (lanes 2 and 3) or GST–RAP46 fusion protein (lanes 4 and 5) were added to the cell extracts before providing the hormone and the probe. Where indicated, GST and GST–RAP46 were premixed with 2 µg of purified Hsc70 before addition to the cell extracts (lanes 3 and 5). In (A–D), dexamethasone (100 nM) was added to the EMSA reaction mixture prior to incubation with the radioactive probe.

Fig. 4. In vitro cell-free system for studying the effect of RAP46 on DNA binding by the GR. One hundred thousand COS-7 cells in 3.4 cm culture dishes were transiently transfected by the Fugene transfection procedure with 0.9 µg empty expression vector or with an expression vector encoding the GR. Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR-response element as a probe. Where indicated, different amounts of bacterially purified GST or GST fusion proteins were added to the cell extracts. The fusion proteins consisted of GST alone or GST fused to either the wild-type RAP46 sequence (GST–RAP46) or RAP46 lacking the last 47 C-terminal amino acids (GST–RAP46dC47). (A) Inhibition of DNA binding by the GR is mediated by RAP46 and not RAP46C47. EMSA was carried out with cell extracts containing the GR and 1 µg GST, GST–RAP46 or GST–RAP46dC47. (B) Hsc70 relieves RAP46-mediated inhibition of DNA binding by the GR. COS-7 cells were transiently transfected with an empty expression vector (vector, lane 1), or an expression vector encoding either the GR (lanes 2–10) or an HA-tagged RAP46 (lanes 3–10). Whole-cell extracts were prepared and subjected to an EMSA using radioactively labelled GR response element as a probe. Where indicated, different amounts of purified Hsc70 were added to the whole-cell extracts. (C) Increased concentration of RAP46 downregulated DNA binding by the GR. Concentrations of GST–RAP46 ranging from 0.1 to 1000 ng was added to the cellular extract before the DNA binding reaction were carried out. (D) A RAP46–Hsc70 complex does not affect DNA binding by the GR. COS-7 cells were transiently transfected with either an empty expression vector (vector, lane 1) or an expression vector encoding the GR (lanes 2–5). Whole-cell extracts were prepared and subjected to EMSA. One microgram of bacterially produced GST (lanes 2 and 3) or GST–RAP46 fusion protein (lanes 4 and 5) were added to the cell extracts before providing the hormone and the probe. Where indicated, GST and GST–RAP46 were premixed with 2 µg of purified Hsc70 before addition to the cell extracts (lanes 3 and 5). In (A–D), dexamethasone (100 nM) was added to the EMSA reaction mixture prior to incubation with the radioactive probe.

Fig. 5. Confocal immunofluorescence analysis of the intracellular localization of GR and Hsp70. Five hundred thousand COS-7 cells in 10 cm culture dishes were transiently transfected by the Fugene procedure with 4 µg of expression vectors encoding GR–GFP or GR(Δ491–515)–GFP and Hsp70. The Hsp70 was tagged with a decapeptide from human testis lactate dehydrogenase. Twenty-four hours after the transfection, the cells were treated for 16 h with vehicle alone (0.1% ethanol) (–Dex) or vehicle containing 0.1 µM dexamethasone (+ Dex) before harvesting, processing and visualization with laser confocal microscopy. The green fluorescence arises from the GFP-tagged receptors, whereas the red fluorescence comes from staining of Hsp70 with a rat antibody against the lactate dehydrogenase tag followed by an anti-rat antibody labelled with rhodamine. The yellow to orange colour indicates colocalization of the receptor and Hsp70.