Active ERK1 is dimerized in vivo: bisphosphodimers generate peak kinase activity and monophosphodimers maintain basal ERK1 activity

Abstract

ERK1 and ERK2 are widely involved in cell signalling. Using a recombinant approach, it has been shown that exogenous ERK2 is capable of dimerization and that preventing dimerization reduces its nuclear accumulation on stimulation. Dimerization occurs on phosphorylation; the dimer partner of phosphorylated ERK2 may be either phosphorylated or unphosphorylated. It has been assumed that monophosphodimers are hemiactive. Here we show that ERK1 is capable of dimerization both in vivo and in vitro. Dimerization of human recombinant ERK1 in vitro requires both ERK1 phosphorylation and cellular cofactor(s); it leads to the formation of a high molecular weight complex that can be dissociated by treatment with beta-mercaptoethanol. We demonstrate for the first time in both sea urchin embryos and human cells that native ERK forms dimers and that high ERK kinase activity is largely associated with bisphosphodimers, not with monophosphodimers or phosphorylated monomers. The activity of the bisphosphodimer is about 20-fold higher than that of the phosphorylated monomer in vitro and the bisphosphodimer shows 5- to 7-fold higher in vivo activity than the basal activity attributable to the monophosphodimer. Thus phosphorylation of both partners in the dimer is a hallmark of ERK activation. Judgments made about ERK kinase activity associated with phosphorylated monomers are at best a proxy for ERK activity.

Fig. 1. Human ERK1 activated in vitro dimerises in the presence of cellular co-factors. The highest activity is associated with bisphosphodimers

A Western blotting of recombinant hERK1 with anti-dualphosphorylated ERK antibody: Lane 1. unactivated protein. Lane 2. unactivated protein incubated in cell extract for 15 min. Lane 3. hERK1 after 1 hour activation by recombinant MEK, not incubated in cell extract. Lane 4. hERK1 treated as for Lane 3, but with an additional incubation in cell extract for 15 min. Lane 5. hERK1 treated as for Lane 4, followed by dephosphorylation using hCL100. Lane 6. hERK1 treated as for Lane 4, followed by additional activation by MEK overnight at 4°C . Lane 7. hERK1 sample treated identically to the sample in Lane 6, but dissolved in sample buffer containing 5% β-mercaptoethanol. Lane 8. Control: low ERK1 activity extract (30 min post-fertilisation, 15 μg total cellular protein loaded); active ERK1 is undetectable. -AB Control: hERK1 sample treated as for Lane 4, but primary antibody was omitted. The positions of molecular mass markers are shown. B MAP kinase assays with myelin basic protein (MBP) using samples identical to those in A. Relative protein kinase activity is also shown from densitometry. C Western blotting using anti-ERK antibody of samples identical those of lanes 2,4,5,6 and -AB in A. The positions of molecular mass markers are shown. D Western blotting using anti-GST antibody of samples identical to 4,5,6 and -AB in A. The positions of molecular mass markers are shown. E GST-ΔPEHD-hERK1, a dimerisation deficient mutant, does not show enhanced activity when incubated in whole-cell extract. 6 pM of each ERK1 recombinant protein was used per sample. Western blotting with anti-ERK antibody: Lane 1. unactivated ΔPEHD-hERK1. Lane 2. ΔPEHD-hERK1 after 1 hour activation by active MEK not incubated in cell extract. Lane 3. ΔPEHD-hERK1 treated as for Lane 2, but with an additional incubation in cell extract for 15 min. Lane 4. ΔPEHD-hERK1 treated as for Lane 3, followed by additional activation by MEK overnight at 4°C. Lane 5. Equal amount of recombinant control hERK1 treated as for Lane 3. F MAP kinase assays with MBP using samples identical to those in E. Relative protein kinase activity is also shown from densitometry. Note that the monomeric mutant protein is effectively activated by MEK (eight-fold over baseline), however incubation in cell extract does not result in additional increase in its activity. The same amount of activated control wild type protein forms a complex in cell extract with much higher kinase activity.

Fig. 1. Human ERK1 activated in vitro dimerises in the presence of cellular co-factors. The highest activity is associated with bisphosphodimers

A Western blotting of recombinant hERK1 with anti-dualphosphorylated ERK antibody: Lane 1. unactivated protein. Lane 2. unactivated protein incubated in cell extract for 15 min. Lane 3. hERK1 after 1 hour activation by recombinant MEK, not incubated in cell extract. Lane 4. hERK1 treated as for Lane 3, but with an additional incubation in cell extract for 15 min. Lane 5. hERK1 treated as for Lane 4, followed by dephosphorylation using hCL100. Lane 6. hERK1 treated as for Lane 4, followed by additional activation by MEK overnight at 4°C . Lane 7. hERK1 sample treated identically to the sample in Lane 6, but dissolved in sample buffer containing 5% β-mercaptoethanol. Lane 8. Control: low ERK1 activity extract (30 min post-fertilisation, 15 μg total cellular protein loaded); active ERK1 is undetectable. -AB Control: hERK1 sample treated as for Lane 4, but primary antibody was omitted. The positions of molecular mass markers are shown. B MAP kinase assays with myelin basic protein (MBP) using samples identical to those in A. Relative protein kinase activity is also shown from densitometry. C Western blotting using anti-ERK antibody of samples identical those of lanes 2,4,5,6 and -AB in A. The positions of molecular mass markers are shown. D Western blotting using anti-GST antibody of samples identical to 4,5,6 and -AB in A. The positions of molecular mass markers are shown. E GST-ΔPEHD-hERK1, a dimerisation deficient mutant, does not show enhanced activity when incubated in whole-cell extract. 6 pM of each ERK1 recombinant protein was used per sample. Western blotting with anti-ERK antibody: Lane 1. unactivated ΔPEHD-hERK1. Lane 2. ΔPEHD-hERK1 after 1 hour activation by active MEK not incubated in cell extract. Lane 3. ΔPEHD-hERK1 treated as for Lane 2, but with an additional incubation in cell extract for 15 min. Lane 4. ΔPEHD-hERK1 treated as for Lane 3, followed by additional activation by MEK overnight at 4°C. Lane 5. Equal amount of recombinant control hERK1 treated as for Lane 3. F MAP kinase assays with MBP using samples identical to those in E. Relative protein kinase activity is also shown from densitometry. Note that the monomeric mutant protein is effectively activated by MEK (eight-fold over baseline), however incubation in cell extract does not result in additional increase in its activity. The same amount of activated control wild type protein forms a complex in cell extract with much higher kinase activity.

Fig. 1. Human ERK1 activated in vitro dimerises in the presence of cellular co-factors. The highest activity is associated with bisphosphodimers

A Western blotting of recombinant hERK1 with anti-dualphosphorylated ERK antibody: Lane 1. unactivated protein. Lane 2. unactivated protein incubated in cell extract for 15 min. Lane 3. hERK1 after 1 hour activation by recombinant MEK, not incubated in cell extract. Lane 4. hERK1 treated as for Lane 3, but with an additional incubation in cell extract for 15 min. Lane 5. hERK1 treated as for Lane 4, followed by dephosphorylation using hCL100. Lane 6. hERK1 treated as for Lane 4, followed by additional activation by MEK overnight at 4°C . Lane 7. hERK1 sample treated identically to the sample in Lane 6, but dissolved in sample buffer containing 5% β-mercaptoethanol. Lane 8. Control: low ERK1 activity extract (30 min post-fertilisation, 15 μg total cellular protein loaded); active ERK1 is undetectable. -AB Control: hERK1 sample treated as for Lane 4, but primary antibody was omitted. The positions of molecular mass markers are shown. B MAP kinase assays with myelin basic protein (MBP) using samples identical to those in A. Relative protein kinase activity is also shown from densitometry. C Western blotting using anti-ERK antibody of samples identical those of lanes 2,4,5,6 and -AB in A. The positions of molecular mass markers are shown. D Western blotting using anti-GST antibody of samples identical to 4,5,6 and -AB in A. The positions of molecular mass markers are shown. E GST-ΔPEHD-hERK1, a dimerisation deficient mutant, does not show enhanced activity when incubated in whole-cell extract. 6 pM of each ERK1 recombinant protein was used per sample. Western blotting with anti-ERK antibody: Lane 1. unactivated ΔPEHD-hERK1. Lane 2. ΔPEHD-hERK1 after 1 hour activation by active MEK not incubated in cell extract. Lane 3. ΔPEHD-hERK1 treated as for Lane 2, but with an additional incubation in cell extract for 15 min. Lane 4. ΔPEHD-hERK1 treated as for Lane 3, followed by additional activation by MEK overnight at 4°C. Lane 5. Equal amount of recombinant control hERK1 treated as for Lane 3. F MAP kinase assays with MBP using samples identical to those in E. Relative protein kinase activity is also shown from densitometry. Note that the monomeric mutant protein is effectively activated by MEK (eight-fold over baseline), however incubation in cell extract does not result in additional increase in its activity. The same amount of activated control wild type protein forms a complex in cell extract with much higher kinase activity.

Fig. 2. ERK activity in vivo is associated with its homodimers. Monomers at 44 kDa are largely inactive

A Schematic diagram of ERK1 activity during the first mitotic cell cycle of sea urchin embryos. An inset with the original data (Philipova and Whitaker, 1998) is also shown. B Sea urchin embryos Active ERK1 was immunoprecipitated during the first mitotic cell cycle at time points that corresponded to maximum (6min, NEB) and minimum activity (Unfertilised, 30 min) using an anti-dualphosphorylated ERK antibody and detected by Western blotting with a second, different anti-dualphosphorylated ERK antibody or with an anti-MEK antibody (NEB sample). Arrows: active ERK1 dimers. Cell extract: Whole-cell extracts are rich in ERK1 monomers as a Western blot of a 30 min post-fertilisation cell extract detected with anti-ERK antibody demonstrates. It was necessary to load 150 μg total cellular protein in order to obtain a detectable band at 91 kDa for comparison with the anti-dualphosphorylated ERK antibody immunoprecipitate. The positions of molecular mass markers are shown. C HeLa cells: Active ERK was immunoprecipitated from cell extracts using an anti-dualphosphorylated ERK antibody, and proteins detected on Western blots using a second anti-dualphosphorylated ERK antibody, an anti-ERK antibody or an anti-MEK antibody. The immunoprecipitate was also run in the absence of β-mercaptoethanol (β-ME) before blotting (Protein complex). The positions of molecular mass markers are shown.

Fig. 2. ERK activity in vivo is associated with its homodimers. Monomers at 44 kDa are largely inactive

A Schematic diagram of ERK1 activity during the first mitotic cell cycle of sea urchin embryos. An inset with the original data (Philipova and Whitaker, 1998) is also shown. B Sea urchin embryos Active ERK1 was immunoprecipitated during the first mitotic cell cycle at time points that corresponded to maximum (6min, NEB) and minimum activity (Unfertilised, 30 min) using an anti-dualphosphorylated ERK antibody and detected by Western blotting with a second, different anti-dualphosphorylated ERK antibody or with an anti-MEK antibody (NEB sample). Arrows: active ERK1 dimers. Cell extract: Whole-cell extracts are rich in ERK1 monomers as a Western blot of a 30 min post-fertilisation cell extract detected with anti-ERK antibody demonstrates. It was necessary to load 150 μg total cellular protein in order to obtain a detectable band at 91 kDa for comparison with the anti-dualphosphorylated ERK antibody immunoprecipitate. The positions of molecular mass markers are shown. C HeLa cells: Active ERK was immunoprecipitated from cell extracts using an anti-dualphosphorylated ERK antibody, and proteins detected on Western blots using a second anti-dualphosphorylated ERK antibody, an anti-ERK antibody or an anti-MEK antibody. The immunoprecipitate was also run in the absence of β-mercaptoethanol (β-ME) before blotting (Protein complex). The positions of molecular mass markers are shown.

Fig. 2. ERK activity in vivo is associated with its homodimers. Monomers at 44 kDa are largely inactive

A Schematic diagram of ERK1 activity during the first mitotic cell cycle of sea urchin embryos. An inset with the original data (Philipova and Whitaker, 1998) is also shown. B Sea urchin embryos Active ERK1 was immunoprecipitated during the first mitotic cell cycle at time points that corresponded to maximum (6min, NEB) and minimum activity (Unfertilised, 30 min) using an anti-dualphosphorylated ERK antibody and detected by Western blotting with a second, different anti-dualphosphorylated ERK antibody or with an anti-MEK antibody (NEB sample). Arrows: active ERK1 dimers. Cell extract: Whole-cell extracts are rich in ERK1 monomers as a Western blot of a 30 min post-fertilisation cell extract detected with anti-ERK antibody demonstrates. It was necessary to load 150 μg total cellular protein in order to obtain a detectable band at 91 kDa for comparison with the anti-dualphosphorylated ERK antibody immunoprecipitate. The positions of molecular mass markers are shown. C HeLa cells: Active ERK was immunoprecipitated from cell extracts using an anti-dualphosphorylated ERK antibody, and proteins detected on Western blots using a second anti-dualphosphorylated ERK antibody, an anti-ERK antibody or an anti-MEK antibody. The immunoprecipitate was also run in the absence of β-mercaptoethanol (β-ME) before blotting (Protein complex). The positions of molecular mass markers are shown.

Fig. 3. In vitro phosphorylation by MEK of immunoprecipitated low-activity cellular ERK1 dimer (_*-o) results in the appearance of the shifted fully activated dimerised ERK1 (_*-_*)

Total ERK1 was immunoprecipitated from 30 min extracts from sea urchin embryos using an anti-ERK antibody and then treated in vitro with either active MEK (IP + MEK) or an inactive (kinase dead) mutant of MEK (IP + k.d.MEK); a control IP in the absence of the anti-ERK antibody is also shown (Control IP –AB). A and B show a comparison between Western blots of the same samples with (A) anti-ERK antibody or (B) anti-dualphosphorylated ERK antibody. In both cases, the immunoprecipitates are compared with an untreated aliquot of the 30 min extract used for immunoprecipitation. Note that immunoprecipitates using anti-dualphosphorylated ERK antibody are highly enriched in the 91 kDa dimeric form. (C) Relative activities of samples treated identically as indicated in B, washed in MAP kinase buffer and used in protein kinase assays. The sample activated by MEK shows 6 fold higher activity than those either untreated or treated with kinase dead MEK.

Fig. 4. A. The two ERK1 homodimers (_*-o and _*-_*) co-purify with the ERK1 monomer

Western blots of a fraction with high MAP kinase activity (Fraction 23) eluted from a Phenyl-Sepharose column during ERK1 purification from mitotic sea urchin embryo extracts probed with (a) an anti-ERK antibody and (b) an anti-dualphosphorylated ERK antibody. 4μl of the fraction loaded for each blot. Blot b was detected using the high-sensitivity ECL Advance Kit: a faint band of active monomer is visible (arrow). B, C. Phosphorylation by MEK and dephosphorylation using CL100 dual specificity phosphatase cause the predicted gel shifts of the ERK1 dimers in the phenylsepharose column fraction. 0.5 μl of fraction 23 was used for each treatment; samples were blotted and probed with (B) anti-ERK antibody and (C) anti-dualphosphorylated ERK antibody. Samples treated with CL100 in the presence of vanadate to block phosphatase activity (CL100+Na₃VO₄ or with mutant inactive (kinase dead) MEK (k.d.MEK) serve as controls. Treatment with CL100 led to the loss of the more slowly migrating band, while treatment with MEK enhanced the band. Note that after MEK treatment, dimeric forms accumulate in preference to the monomer.

Fig. 4. A. The two ERK1 homodimers (_*-o and _*-_*) co-purify with the ERK1 monomer

Western blots of a fraction with high MAP kinase activity (Fraction 23) eluted from a Phenyl-Sepharose column during ERK1 purification from mitotic sea urchin embryo extracts probed with (a) an anti-ERK antibody and (b) an anti-dualphosphorylated ERK antibody. 4μl of the fraction loaded for each blot. Blot b was detected using the high-sensitivity ECL Advance Kit: a faint band of active monomer is visible (arrow). B, C. Phosphorylation by MEK and dephosphorylation using CL100 dual specificity phosphatase cause the predicted gel shifts of the ERK1 dimers in the phenylsepharose column fraction. 0.5 μl of fraction 23 was used for each treatment; samples were blotted and probed with (B) anti-ERK antibody and (C) anti-dualphosphorylated ERK antibody. Samples treated with CL100 in the presence of vanadate to block phosphatase activity (CL100+Na₃VO₄ or with mutant inactive (kinase dead) MEK (k.d.MEK) serve as controls. Treatment with CL100 led to the loss of the more slowly migrating band, while treatment with MEK enhanced the band. Note that after MEK treatment, dimeric forms accumulate in preference to the monomer.

Fig. 5. Newly phosphorylated ERK1 accumulates as fully active (_*-_*) homodimers

Incorporation of γ-[³²P]ATP into purified ERK1 (fraction 23 from the phenylsepharose column) during in vitro phosphorylation by MEK. 0.5μl of the fraction was used per reaction as a substrate for in vitro MEK assays. Kinase reactions were carried out for 2, 4 and 6 min at 30°C, and then terminated by addition of sample buffer. Molecular mass markers are also shown. Note that radioactivity is incorporated predominantly into the band above 116kDa. The radioactivity at the top of the gel represents insoluble material that has not entered the gel.