De Novo Mutations in Protein Kinase Genes CAMK2A and CAMK2B Cause Intellectual Disability

Abstract

Calcium/calmodulin-dependent protein kinase II (CAMK2) is one of the first proteins shown to be essential for normal learning and synaptic plasticity in mice, but its requirement for human brain development has not yet been established. Through a multi-center collaborative study based on a whole-exome sequencing approach, we identified 19 exceedingly rare de novo CAMK2A or CAMK2B variants in 24 unrelated individuals with intellectual disability. Variants were assessed for their effect on CAMK2 function and on neuronal migration. For both CAMK2A and CAMK2B, we identified mutations that decreased or increased CAMK2 auto-phosphorylation at Thr286/Thr287. We further found that all mutations affecting auto-phosphorylation also affected neuronal migration, highlighting the importance of tightly regulated CAMK2 auto-phosphorylation in neuronal function and neurodevelopment. Our data establish the importance of CAMK2A and CAMK2B and their auto-phosphorylation in human brain function and expand the phenotypic spectrum of the disorders caused by variants in key players of the glutamatergic signaling pathway.

Keywords: AMPAR; CAMK2; CAMK2A; CAMK2B; NMDAR; de novo mutations; intellectual disability; synaptic plasticity.

Figure 1

Molecular Genetic Findings in Individuals with CAMK2A and CAMK2B Variants (A) Schematic of CAMK2A and CAMK2B protein domain organizations and corresponding mRNA structure (derived from PDB features for Q9UQM7 and Q13554) indicating the positions of 12 missense variants, 1 stop-gain variant, 1 frameshift deletion, and 5 splice site variants found in affected individuals, together with the variant CAMK2A p.Pro138Ala reported in the literature as de novo., (B and C) Representation of the structure of a single human CAMK2A subunit obtained from the corresponding full-length holoenzyme structure present in the protein data bank (PDB ID: 3SOA) (B) and homology model of a single human CAMK2B subunit without the F-actin binding domain (C). For both structures, the catalytic domain is represented in green, the autoregulatory domain in yellow, and the association domain in cyan. The location of each single point variant (in magenta) in CAMK2A and CAMK2B, respectively, is indicated in the 3D structure (two different orientations), showing that for CAMK2A seven of the missense variants are located in the catalytic domain, and two variants in the autoregulatory domain. No variants were found in the association domain (B). For CAMK2B, three of the missense variants are located in the catalytic domain and one in the autoregulatory domain. No variant was found in the association domain (C). Structure representations were made with PyMol. Correspondence between the nomenclatures of amino acid changes: F98S, p.Phe98Ser; G109D, p.Glu109Asp; A112V, p.Ala112Val; P138A, p.Pro138Ala; E183V, p.Glu183Val; P212L, p.Pro212Leu; P235L, p.Pro235Leu; H282R, p.His282Arg; T286P, p.Thr286Pro; E110K, p.Glu110Lys; P139L, p.Pro139Leu; E237K, p.Glu237Lys; K301E, p.Lys301Glu.

Figure 2

Transfection of HEK293T Cells with the Different CAMK2 Mutants Shows Changes in Stability as well as Phosphorylation at Thr286/287 (A) Schematic overview of the in vitro and in vivo assays. (B and C) Top, Representative western blots of HEK293T cells transfected with either CAMK2A or CAMK2B constructs, probed with an antibody against CAMK2A, CAMK2B, and RFP. Below, quantification of the normalized protein levels of CAMK2A or CAMK2B, showing instability for CAMK2A^{p.(Glu183Val)}, CAMK2A^{p.(His282Arg)}, CAMK2B^{p.(Glu110Lys)}, and CAMK2B^{p.(Pro139Leu)} proteins. (D and E) Top, Representative western blots of HEK293T cells transfected with either CAMK2A or CAMK2B constructs, probed with a specific antibody against the phosphorylation site Thr286/287 and an antibody against CAMK2A and CAMK2B, respectively. Below, quantification of the normalized levels of CAMK2A-Thr286 phosphorylation and normalized levels of CAMK2B-Thr287 phosphorylation. Number in the box and whisker plot graphs indicates the n per construct. Error bars indicate the minimum and maximum of all data. Individual data points are shown in the box and whisker plots. Correspondence between the nomenclatures of amino acid changes: F98S, p.Phe98Ser; G109D, p.Glu109Asp; A112V, p.Ala112Val; P138A, p.Pro138Ala; E183V, p.Glu183Val; P212L, p.Pro212Leu; P235L, p.Pro235Leu; H282R, p.His282Arg; T286P, p.Thr286Pro; E110K, p.Glu110Lys; P139L, p.Pro139Leu; E237K, p.Glu237Lys; K301E, p.Lys301Glu.

Figure 3

Increased or Decreased Levels of the CAMK2B but Not the CAMK2A Protein In Vivo Cause Deficits in Neuronal Migration in Mice (A) Representative images of E14.5 in utero electroporated P0 brains (overexpression of CAMK2), with the cortical plate (CP, dashed lines) and the subventricular zone (SVZ, arrow) indicated. tdTomato-positive cells represent the successfully transfected neurons. (B) Left: Quantification of the neuronal migration pattern observed in the different conditions. The cortex was divided into 10 bins of equal size and the percentage of tdTomato⁺/RFP⁺ cells per bin was counted. Right: Analysis of the percentage of targeted cells that reach the outer layers of the cortex measured as the sum of bin 1 to 4. (C) Representative images of E14.5 in utero electroporated P0 brains (knockdown of endogenous CAMK2), with the cortical plate (CP, dashed lines) and the subventricular zone (SVZ, arrow) indicated. RFP-positive cells represent the successfully transfected neurons. (D) Left: Quantification of the neuronal migration pattern observed in the different conditions. The cortex was divided into 10 bins of equal size and the percentage of tdTomato⁺/RFP⁺ cells per bin was counted. Right: Analysis of the percentage of targeted cells that reach the outer layers of the cortex measured as the sum of bin 1 to 4. Number in the box and whisker plot graphs indicates the number of pictures analyzed per construct. Error bars indicate the minimum and maximum of all data. Individual data points are shown in the box and whisker plots.

Figure 4

Transfection of CAMK2A Mutations with Changes in Thr286 Phosphorylation In Vivo Induces Migration Deficits (A) Schematic overview of CAMK2A with the location of the variants with increased (green), decreased (red), or unchanged (gray) Thr286 phosphorylation. (B) Representative images of E14.5 in utero electroporated P0 brains. tdTomato-positive cells represent the successfully transfected neurons. (C–E) Left: Quantification of the neuronal migration pattern observed in the variants with unchanged (C), decreased (D), or increased (E) Thr286 auto-phosphorylation. Right: Analysis of the percentage of targeted cells of the different constructs that reach the outer layers of the cortex measured as the sum of bin 1 to 4. Dotted line indicates the WT level. Number in the box and whisker plot graphs indicates the number of pictures analyzed per construct. Error bars indicate the minimum and maximum of all data. Individual data points are shown in the box and whisker plots.

Figure 5

Transfection of the CAMK2A^{p.(Thr286Pro)} Mutation In Vivo Reveals a Constitutive Active Phenotype (A) Representative images of E14.5 in utero electroporated P0 brains. tdTomato-positive cells represent the successfully transfected neurons. (B) Left: Quantification of the neuronal migration pattern observed when overexpressing CAMK2A^{p.(Thr286Pro)}. Right: Analysis of the percentage of targeted cells of the different constructs that reach the outer layers of the cortex measured as the sum of bin 1 to 4. (C) Left: Quantification of the neuronal migration pattern observed when overexpressing CAMK2A^{p.(Thr286Ala)} or CAMK2A^{p.(Thr286Pro)}. Right: Analysis of the percentage of targeted cells of the different constructs that reach the outer layers of the cortex measured as the sum of bin 1 to 4. (D) Left: Quantification of the neuronal migration pattern observed when overexpressing CAMK2A^{p.(Thr286Pro)/p.(Lys42Arg)}. Right: Analysis of the percentage of targeted cells of the different constructs that reach the outer layers of the cortex measured as the sum of bin 1 to 4. Dotted line indicates the WT level. Number in the box and whisker plot graphs indicates the number of pictures analyzed per construct. Error bars indicate the minimum and maximum of all data. Individual data points are shown in the box and whisker plots.

Figure 6

Transfection of CAMK2B Mutations In Vivo Causes Changes in the Migration Pattern of the Targeted Neurons (A) Schematic overview of CAMK2B with the location of the mutations with increased (green), decreased (red), or unchanged (gray) Thr286 phosphorylation. (B) Representative images of E14.5 in utero electroporated P0 brains. tdTomato-positive cells represent the successfully transfected neurons. (C) Left: Quantification of the neuronal migration pattern observed when overexpressing the mutations. Right: Analysis of the percentage of targeted cells of the different constructs that reach the outer layers of the cortex measured as the sum of bin 1 to 4. Dotted line indicates the WT level. Number in the box and whisker plot graph indicates the number of pictures analyzed per construct. Error bars indicate the minimum and maximum of all data. Individual data points are shown in the box and whisker plots.