Alternative splicing modulates the frequency-dependent response of CaMKII to Ca(2+) oscillations

Abstract

Ca(2+) oscillations are required in various signal trans duction pathways, and contain information both in their amplitude and frequency. Remarkably, the Ca(2+)/calmodulin(CaM)-dependent protein kinase II (CaMKII) can decode such frequencies. A Ca(2+)/CaM-stimulated autophosphorylation leads to Ca(2+)/CaM-independent (autonomous) activity of the kinase that outlasts the initial stimulation. This autonomous activity increases exponentially with the frequency of Ca(2+) oscillations. Here we show that three beta-CaMKII splice variants (beta(M), beta and beta(e)') have very similar specific activity and maximal autonomy. However, their autonomy generated by Ca(2+) oscillations differs significantly. A mechanistic basis was found in alterations of the CaM activation constant and of the initial rate of autophosphorylation. Structurally, the splice variants differ only in a variable 'linker' region between the kinase and association domains. Therefore, we propose that differences in relative positioning of kinase domains within multimeric holoenzymes are responsible for the observed effects. Notably, the beta-CaMKII splice variants are differentially expressed, even among individual hippocampal neurons. Taken together, our results suggest that alternative splicing provides cells with a mechanism to modulate their sensitivity to Ca(2+) oscillations.

Fig. 1. Structure of β-CaMKII splice variants. (A) Primary structure of the β-CaMKII splice variants in a schematic representation (left), and CaMKII holoenzyme structure (Kolodziej et al., 2000) (right). (B) The variable region of β-CaMKII. Amino acid sequences missing in β_e′ are boxed. The lines above the sequence indicate the three β_M-specific repeats. (C) Intron/exon borders of the three β_M-specific exons (black boxes) as determined by further sequence analysis of the murine β-CaMKII locus (for complete β_M exon sequences see Bayer et al., 1999; more intron sequences at DDBJ/EMBL/GenBank accession Nos AF416336 and AF416337).

Fig. 2. (A) Autophosphorylation at the autonomy site T287 can be induced by Ca²⁺/CaM for all three β variants, as detected by immunoblotting using a phospho-T287-specific antibody (top). Similar phos phorylation was observed for α (at the homologous T286). No T287 phosphorylation was detected before stimulation with Ca²⁺/CaM (bottom). (B) Immunodetection of total immobilized β-CaMKII splice variants with the same specific kinase activity (3000 fmol of ATP/min), using the CB-β1 antibody. Below: purified β-CaMKII as standard. (C) Specific activities of β splice variants (means ± SEM) in a peptide substrate phosphorylation assay. Stimulated activity was induced by 300 nM CaM; autonomous activity was measured in the absence of Ca²⁺/CaM, but after autophosphorylation. Autonomy is indicated relative to the stimulated activity. Basal activity (not stimulated and without autophosphorylation) for all variants was <1.5%, as indicated. The experiments were performed with HA-tagged kinase isoforms immobilized on anti-HA antibody-coated microtiter plates.

Fig. 3. Differential response of β-CaMKII splice variants to Ca²⁺ oscillations. Different frequencies of Ca²⁺ oscillations evoke different levels of autonomous activity for β, β_M and β_e′. Curves were fitted to the single-exponential function y = ae^bx + c. The factor of exponential increase, b, for β_e′ is 10% greater than that for β, and 50% greater than that for β_M. All data points are means ± SEM.

Fig. 4. Calmodulin activation constant and initial rate of autophosphorylation as parameters that can modulate the response of CaMKII to Ca²⁺ oscillations. (A) Calmodulin concentration–response curves reveal a higher K_a for β_e′ than for β and β_M (47.5 ± 6.9, 22.4 ± 2.5 and 21.4 ± 1.7 nM, respectively). (B) The initial rate of autophosphorylation is higher for β_M than for β, as measured by the resulting autonomous activity. All data points are means ± SEM.

Fig. 5. β_M-CaMKII is expressed in neurons. (A) RT–PCR with β_M- specific primers at increasing cycle numbers from oligo-dT-primed cDNA prepared from mRNA from different rat tissues. The positions of 300 and 400 bp marker bands are indicated. (B) In situ hybridization of cultured hippocampal neurons with a digoxigenin-labeled β_M- specific RNA probe (left) (nucleotides 133–354 in Bayer et al., 1999) and the corresponding sense probe control (right). Color coding represents different expression levels, as indicated. (C) Immuno cytochemistry for β-CaMKII and MAP-2 in the cultured neurons. Nearly all neurons (MAP-2-positive cells) are positively stained with an antibody specific for all splice variants of β-CaMKII. Scale bars = 40 µm.