Seebohm G et al. (2014), Structural basis of PI(4,5)P2-dependent regulat...

XB-ART-50773

Pflugers Arch 2014 Oct 01;46610:1885-97. doi: 10.1007/s00424-013-1424-8.

Show Gene links Show Anatomy links

Structural basis of PI(4,5)P2-dependent regulation of GluA1 by phosphatidylinositol-5-phosphate 4-kinase, type II, alpha (PIP5K2A).

Seebohm G , Wrobel E , Pusch M , Dicks M , Terhag J , Matschke V , Rothenberg I , Ursu ON , Hertel F , Pott L , Lang F , Schulze-Bahr E , Hollmann M , Stoll R , Strutz-Seebohm N .

???displayArticle.abstract???
Ionotropic glutamate receptors are the most important excitatory receptors in the central nervous system, and their impairment can lead to multiple neuronal diseases. Here, we show that glutamate-induced currents in oocytes expressing GluA1 are increased by coexpression of the schizophrenia-associated phosphoinositide kinase PIP5K2A. This effect was due to enhanced membrane abundance and was blunted by a point mutation (N251S) in PIP5K2A. An increase in GluA1 currents was also observed upon acute injection of PI(4,5)P2, the main product of PIP5K2A. By expression of wild-type and mutant PIP5K2A in human embryonic kidney cells, we were able to provide evidence of impaired kinase activity of the mutant PIP5K2A. We defined the region K813-K823 of GluA1 as critical for the PI(4,5)P2 effect by performing an alanine scan that suggested PI(4,5)P2 binding to this area. A PIP strip assay revealed PI(4,5)P2 binding to the C-terminal GluA1 peptide. The present observations disclose a novel mechanism in the regulation of GluA1.

???displayArticle.pubmedLink??? 24389605
???displayArticle.pmcLink??? PMC4159565
???displayArticle.link??? Pflugers Arch

Species referenced: Xenopus
Genes referenced: canx pip4k2a pip4k2b

???attribute.lit??? ???displayArticles.show???

	Fig. 1. Effects of PIP5K2A or PI(4,5)P2 but not of PIP5K2A(N251S) on properties of GluA1 currents in oocytes. a Representative current traces measured in Xenopus oocytes in response to superfusion with 300 μM glutamate and 100 μM cyclothiazide. All currents were measured at −70 mV. Steady-state currents (indicated by arrows) were used for analysis. b GluA1 current amplitudes in oocytes expressing GluA1, GluA1 + PIP5K2A, or GluA1 + PIP5K2A(N251S) normalized to the GluA1 currents. Numbers of oocytes vary between 23 and 35, and significant changes are indicated by *p < 0.001, p < 0.01, or p < 0.05. c GluA1 current amplitudes in oocytes expressing GluA1 or GluA1 + PIP5K2A are shown. For bars 2 and 4, PI(4,5)P2 was acutely injected 5 min prior to measurements. Numbers of oocytes varied between 12 and 17, and significant changes are indicated by p < 0.01. Striped bars indicate results after injections of oocytes with water-soluble PI(4,5)P2. d Concentration–response curves for glutamate of GluA1, GluA1 + PIP5K2A, and GluA1 + PIP5K2A(N251S) expressed in oocytes. Mean values ± SEM were calculated from independent measurements (n = 5–6). e Oocytes were injected with BAPTA prior to the experiments. GluA1 current amplitudes were determined in oocytes expressing GluA1 or GluA1 + PIP5K2A, without or after acute injection with a water-soluble PI(4,5)P2 analog. Numbers of oocytes varied between 17 and 26, and significant changes are indicated by p < 0.01 or p < 0.05. Striped bars indicate results from oocytes injected with water-soluble PI(4,5)P2 prior to recording
	Fig. 2. Effect of PI(4,5)P2 on properties of GluA1 currents and of PIP5K2A on GluA1 plasma membrane expression. a Overlay of GluA1 currents in the absence (black trace) and presence (red trace) of PI(4,5)P2 diC8; traces represent averages of 10–20 repeated applications. b Mean peak amplitudes (±SEM) of peak currents evoked by 10-ms application of 1 mM glutamate are shown. Under control conditions, the mean peak current amplitude was 22 ± 4 pA. When PI(4,5)P2 diC8 was present in the pipette solution, the mean peak current was significantly larger (70 ± 11 pA, *p = 0.0009, n = 10, Mann–Whitney test). To compensate for patches of different sizes, current densities were calculated (control: 32 ± 6 pA/pF, PI(4,5)P2 diC8: 66 ± 13 pA/pF; p = 0.0019, n = 10, Mann–Whitney test). c Representative pictures taken from HEK293 cells transfected with GluA1-EGFP alone or GluA1-EGFP together with the kinases PIP5K2A or PIP5K2A(N251S). Dashed lines indicate the position of intensity analysis. d Analysis of GFP fluorescence intensity at the plasma membrane. Line plots were analyzed and the mean data ± SEM is shown. The estimated position of the membrane (arrow) and the side of the cytosol are indicated. Number of cells analyzed: GluA1: n = 53, GluA1 + PIP5K2Awt: n = 55, GluA1 + PIP5K2A(N251S): n = 66. e Western blot of membrane GluA1 expression level in HEK293 cells after coexpression with either PIP5K2A or PIP5K2A(N251S) or GluA1 expressed alone. As control, only transfection reagent was transfected (lane 1). β-Na+/K+-ATPase was used as control protein
	Fig. 3. Reduced PI(4,5)P2 production by PIP5K2A(N251S) could be a result of structural changes at the catalytic site of the mutant. a Representative Western blot of either mock-transfected HEK293 cells or HEK293 cells transfected with PIP5K2A or PIP5K2A(N251S) cDNA. α-Calnexin was used as a marker for proper loading. Bar graph showing the total protein abundance of PIP5K2A protein of mock-transfected HEK293 cells and HEK293 cells transfected with PIP5K2A or PIP5K2A(N251S) cDNA. The band intensity was quantified by densitometry. Number of experiments used for analysis, n = 7. b ELISA assay for detection of PI(4,5)P2 in HEK293 cells expressing either PIP5K2A or PIP5K2A(N251S). Standard curve shows absorbance values obtained versus amount of standard PI(4,5)P2 concentrations. Below, graph shows concentrations of PI(4,5)P2 in mock-transfected cells in comparison to cells expressing either PIP5K2A or PIP5K2A(N251S). Significant change is indicated by **p < 0.001 or p < 0.05 (n = 6–7). c–f Molecular dynamics simulations of PIP5K2A wt and PIP5K2A(N251S). c, d A PIP5K2A wt homology model is depicted. ATP/ADP, PI(4,5)P2, and PI(5)P were docked into the ATP binding site of the PIP5K2A wt (turquoise) and mutant homology model (red) in a similar position as proposed for the related kinase PIP5K2B, resulting in the complexes PIP5K2A-ADP-PI(4,5)P2, PIP5K2A-ATP-PI(5)P, PIP5K2A(N251S)-ADP-PI(4,5)P2, and PIP5K2A(N251S)-ATP-PI(5)P. MD simulations covering 2,000 ps were performed on all four models. Overlays of the final structures are shown in c and d. e, f The affinity of a ligand to its receptor increases with increasing numbers of atom contacts as a result of increased van der Waals interactions. The number of contacts of the respective ligands ATP/ADP and PI(5)P/PI(4,5)P2 with their receptors in PIP5K2A wt (red) and PIP5K2A(N251S) (turquoise) were determined. The combined interactions of both ligands with the respective kinase determine the stability of the enzyme-ligand complexes. The combined number of contacts of both ligands with the kinase receptors is shown in f. The marked differences in contacts between ATP/PI(5)P and ADP/PI(4,5)P2 indicate an increased interaction and thus an increased affinity for ATP/PI(5)P
	Fig. 4. Alanine scan and PIP strip assay reveal the binding of PI(4,5)P2 to the C-terminal domain of GluA1. a Glutamate + cyclothiazide-evoked current amplitudes in oocytes expressing GluA1 or GluA1 point mutants (K813A through G823A) normalized to the GluA1 currents. Statistically significant current changes of the mutants compared to GluA1 wt are indicated by *p < 0.005 or p < 0.01. b GluA1 or GluA1 point mutant current amplitudes in oocytes with (right bars) or without (left bars) coexpression of PIP5K2A. Blue bars indicate that the PI(4,5)P2 effect is present; magenta bars indicate an impaired PI(4,5)P2 effect. Numbers of oocytes measured, n = 10–35. Statistically significant current changes of GluA1 by PIP5K2A are indicated by **p < 0.005 or p < 0.05. c Confidence level of a predicted α-helical secondary structure of a GluA1 peptide and 3D model of the region of GluA1 chosen for the alanine scan are shown. The same color code used as for the bar diagram in b, where blue indicates amino acids probably not involved in PI(4,5)P2 binding while magenta indicates amino acids that may play a role in PI(4,5)P2 binding. d N-terminal FITC-labeled GluA1 peptide shows binding to PIP, PIP2, and PIP3. PIP strip membranes were used for the detection of peptide binding to various lipids. 1 = lysophosphatidic acid, 2 = lysophosphatidylcholine, 3 = phosphadidylinositol, 4 = PtdIns(3)P, 5 = PtdIns(4)P, 6 = PtdIns(5)P, 7 = phosphatidylthanolamine, 8 = phosphatidylcholine, 9 = spingosine 1-phosphate, 10 = PtdIns(3,4)P2, 11 = PtdIns(3,5)P2, 12 = PtdIns(4,5)P2, 13 = PtdIns(3,4,5)P3, 14 = phosphatidic acid, 15 = phosphatidylserine, 16 = blank
	Fig. 5. Molecular dynamics simulations of GluA1 alone and in complex with PI(4,5)P2. a A GluA1 wt homology model was built for residues 410–841. In the case of GluA1 in the presence of 4 PI(4,5)P2, one phospholipid per subunit was manually docked in close proximity to the residues (Lys833, Lys837, Arg838, and Lys840) identified in the alanine scan as PI(4,5)P2-relevant residues (Fig. 4e). GluA1-PI(4,5)P2 was inserted into a PEA/PCH/PSE (each 33 %) membrane. A 2,000-ps MD simulation was performed (details are given in “Material and methods”). The final averaged structure is shown in a. A close-up of the partial inner pore with backbone structures (yellow) within the surface (blue) and cut at the center of the pore is shown in the upper right figure. The position of the membrane core is indicated. The inner pore following the selectivity filter (sf) forms a bowl-like structure (green dotted line). The lower figure shows PI(4,5)P2 (in CPK colors), and the effective residues and non-effective residues identified in the alanine scan are colored magenta and blue, respectively. b The mean RMSD ± SEM of the complete models compared to the initial structures are shown. c The RMSFs of residues compared to the averaged structure are shown

References [+] :

Altschul, Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. 1997, Pubmed

	Fig. 1. Effects of PIP5K2A or PI(4,5)P2 but not of PIP5K2A(N251S) on properties of GluA1 currents in oocytes. a Representative current traces measured in Xenopus oocytes in response to superfusion with 300 μM glutamate and 100 μM cyclothiazide. All currents were measured at −70 mV. Steady-state currents (indicated by arrows) were used for analysis. b GluA1 current amplitudes in oocytes expressing GluA1, GluA1 + PIP5K2A, or GluA1 + PIP5K2A(N251S) normalized to the GluA1 currents. Numbers of oocytes vary between 23 and 35, and significant changes are indicated by *p < 0.001, p < 0.01, or p < 0.05. c GluA1 current amplitudes in oocytes expressing GluA1 or GluA1 + PIP5K2A are shown. For bars 2 and 4, PI(4,5)P2 was acutely injected 5 min prior to measurements. Numbers of oocytes varied between 12 and 17, and significant changes are indicated by p < 0.01. Striped bars indicate results after injections of oocytes with water-soluble PI(4,5)P2. d Concentration–response curves for glutamate of GluA1, GluA1 + PIP5K2A, and GluA1 + PIP5K2A(N251S) expressed in oocytes. Mean values ± SEM were calculated from independent measurements (n = 5–6). e Oocytes were injected with BAPTA prior to the experiments. GluA1 current amplitudes were determined in oocytes expressing GluA1 or GluA1 + PIP5K2A, without or after acute injection with a water-soluble PI(4,5)P2 analog. Numbers of oocytes varied between 17 and 26, and significant changes are indicated by p < 0.01 or p < 0.05. Striped bars indicate results from oocytes injected with water-soluble PI(4,5)P2 prior to recording
	Fig. 2. Effect of PI(4,5)P2 on properties of GluA1 currents and of PIP5K2A on GluA1 plasma membrane expression. a Overlay of GluA1 currents in the absence (black trace) and presence (red trace) of PI(4,5)P2 diC8; traces represent averages of 10–20 repeated applications. b Mean peak amplitudes (±SEM) of peak currents evoked by 10-ms application of 1 mM glutamate are shown. Under control conditions, the mean peak current amplitude was 22 ± 4 pA. When PI(4,5)P2 diC8 was present in the pipette solution, the mean peak current was significantly larger (70 ± 11 pA, *p = 0.0009, n = 10, Mann–Whitney test). To compensate for patches of different sizes, current densities were calculated (control: 32 ± 6 pA/pF, PI(4,5)P2 diC8: 66 ± 13 pA/pF; p = 0.0019, n = 10, Mann–Whitney test). c Representative pictures taken from HEK293 cells transfected with GluA1-EGFP alone or GluA1-EGFP together with the kinases PIP5K2A or PIP5K2A(N251S). Dashed lines indicate the position of intensity analysis. d Analysis of GFP fluorescence intensity at the plasma membrane. Line plots were analyzed and the mean data ± SEM is shown. The estimated position of the membrane (arrow) and the side of the cytosol are indicated. Number of cells analyzed: GluA1: n = 53, GluA1 + PIP5K2Awt: n = 55, GluA1 + PIP5K2A(N251S): n = 66. e Western blot of membrane GluA1 expression level in HEK293 cells after coexpression with either PIP5K2A or PIP5K2A(N251S) or GluA1 expressed alone. As control, only transfection reagent was transfected (lane 1). β-Na+/K+-ATPase was used as control protein
	Fig. 3. Reduced PI(4,5)P2 production by PIP5K2A(N251S) could be a result of structural changes at the catalytic site of the mutant. a Representative Western blot of either mock-transfected HEK293 cells or HEK293 cells transfected with PIP5K2A or PIP5K2A(N251S) cDNA. α-Calnexin was used as a marker for proper loading. Bar graph showing the total protein abundance of PIP5K2A protein of mock-transfected HEK293 cells and HEK293 cells transfected with PIP5K2A or PIP5K2A(N251S) cDNA. The band intensity was quantified by densitometry. Number of experiments used for analysis, n = 7. b ELISA assay for detection of PI(4,5)P2 in HEK293 cells expressing either PIP5K2A or PIP5K2A(N251S). Standard curve shows absorbance values obtained versus amount of standard PI(4,5)P2 concentrations. Below, graph shows concentrations of PI(4,5)P2 in mock-transfected cells in comparison to cells expressing either PIP5K2A or PIP5K2A(N251S). Significant change is indicated by **p < 0.001 or p < 0.05 (n = 6–7). c–f Molecular dynamics simulations of PIP5K2A wt and PIP5K2A(N251S). c, d A PIP5K2A wt homology model is depicted. ATP/ADP, PI(4,5)P2, and PI(5)P were docked into the ATP binding site of the PIP5K2A wt (turquoise) and mutant homology model (red) in a similar position as proposed for the related kinase PIP5K2B, resulting in the complexes PIP5K2A-ADP-PI(4,5)P2, PIP5K2A-ATP-PI(5)P, PIP5K2A(N251S)-ADP-PI(4,5)P2, and PIP5K2A(N251S)-ATP-PI(5)P. MD simulations covering 2,000 ps were performed on all four models. Overlays of the final structures are shown in c and d. e, f The affinity of a ligand to its receptor increases with increasing numbers of atom contacts as a result of increased van der Waals interactions. The number of contacts of the respective ligands ATP/ADP and PI(5)P/PI(4,5)P2 with their receptors in PIP5K2A wt (red) and PIP5K2A(N251S) (turquoise) were determined. The combined interactions of both ligands with the respective kinase determine the stability of the enzyme-ligand complexes. The combined number of contacts of both ligands with the kinase receptors is shown in f. The marked differences in contacts between ATP/PI(5)P and ADP/PI(4,5)P2 indicate an increased interaction and thus an increased affinity for ATP/PI(5)P
	Fig. 4. Alanine scan and PIP strip assay reveal the binding of PI(4,5)P2 to the C-terminal domain of GluA1. a Glutamate + cyclothiazide-evoked current amplitudes in oocytes expressing GluA1 or GluA1 point mutants (K813A through G823A) normalized to the GluA1 currents. Statistically significant current changes of the mutants compared to GluA1 wt are indicated by *p < 0.005 or p < 0.01. b GluA1 or GluA1 point mutant current amplitudes in oocytes with (right bars) or without (left bars) coexpression of PIP5K2A. Blue bars indicate that the PI(4,5)P2 effect is present; magenta bars indicate an impaired PI(4,5)P2 effect. Numbers of oocytes measured, n = 10–35. Statistically significant current changes of GluA1 by PIP5K2A are indicated by **p < 0.005 or p < 0.05. c Confidence level of a predicted α-helical secondary structure of a GluA1 peptide and 3D model of the region of GluA1 chosen for the alanine scan are shown. The same color code used as for the bar diagram in b, where blue indicates amino acids probably not involved in PI(4,5)P2 binding while magenta indicates amino acids that may play a role in PI(4,5)P2 binding. d N-terminal FITC-labeled GluA1 peptide shows binding to PIP, PIP2, and PIP3. PIP strip membranes were used for the detection of peptide binding to various lipids. 1 = lysophosphatidic acid, 2 = lysophosphatidylcholine, 3 = phosphadidylinositol, 4 = PtdIns(3)P, 5 = PtdIns(4)P, 6 = PtdIns(5)P, 7 = phosphatidylthanolamine, 8 = phosphatidylcholine, 9 = spingosine 1-phosphate, 10 = PtdIns(3,4)P2, 11 = PtdIns(3,5)P2, 12 = PtdIns(4,5)P2, 13 = PtdIns(3,4,5)P3, 14 = phosphatidic acid, 15 = phosphatidylserine, 16 = blank
	Fig. 5. Molecular dynamics simulations of GluA1 alone and in complex with PI(4,5)P2. a A GluA1 wt homology model was built for residues 410–841. In the case of GluA1 in the presence of 4 PI(4,5)P2, one phospholipid per subunit was manually docked in close proximity to the residues (Lys833, Lys837, Arg838, and Lys840) identified in the alanine scan as PI(4,5)P2-relevant residues (Fig. 4e). GluA1-PI(4,5)P2 was inserted into a PEA/PCH/PSE (each 33 %) membrane. A 2,000-ps MD simulation was performed (details are given in “Material and methods”). The final averaged structure is shown in a. A close-up of the partial inner pore with backbone structures (yellow) within the surface (blue) and cut at the center of the pore is shown in the upper right figure. The position of the membrane core is indicated. The inner pore following the selectivity filter (sf) forms a bowl-like structure (green dotted line). The lower figure shows PI(4,5)P2 (in CPK colors), and the effective residues and non-effective residues identified in the alanine scan are colored magenta and blue, respectively. b The mean RMSD ± SEM of the complete models compared to the initial structures are shown. c The RMSFs of residues compared to the averaged structure are shown