CaMKII Association with the Actin Cytoskeleton Is Regulated by Alternative Splicing

Supplemental Material

This article contains the following supporting material:

• Figure 1 - CaMKIIβ but not α bundles actin filaments, as assessed by electron microscopy. Kinase (170 nM) was mixed with polymerized actin (4 μM). Shown are images after staining with 2% uranyl actetate. Scale bars: 50 nm.

• Figure 2 - Essentially all CaMKIIβ transcripts in mature rat brain contain the exon v3 sequence.
  A CaMKIIβ-specific PCR product (lane 1) was cut with SacII (lane 2; cuts in exon v1) or BamHI (lane 3; cuts in exon v3). The products were subjected to nested PCR and cut with the same enzymes, SacII (lane 4) and BamHI (lane 5). Bands were quantified using Scion Image, as illustrated. ~7% of the transcripts correspond to βe or βe’ (lacking v1), but no transcripts lacking v3 were detected in the 1^st round of amplification. Even in the 2^nd round, only ~30% of the PCR products were BamHI resistant. Analysis of clones derived from the BamHI resistant band indicated that they were due to PCR mutagenesis and not to alternative splice products lacking exon v3. Of 47 clones analyzed, 6 still contained the BamHI site, 3 were not related to CaMKIIβ, 25 contained point mutations in the BamHI site, and 13 had deletions that did not correspond to exon/intron borders, in 10 cases also disrupting the reading frame. Thus, no CaMKIIβ splice variants that lack variable exon v3 were detectable in mature brain. By contrast, CaMKIIβe, a splice variant lacking exon v1 and constituting the dominant variant before birth (Brocke, et al., 1995), was significantly expressed even in mature brain.
  CaMKIIβ-specific primers b1f (TGAATTCGCCACCACGGTGACCTGCA) and b4r (ACTCGAGCTCCAAACACCAACTCTGT) were used in the first round, nested primers b2f (TTCCACTACCTGGTCTTCGA) and b3r (CGAGCAGTGGAAATGGACAT) in the second round of amplification. First round products were purified using “QIAquick” (Qiagen), digested with restriction enzymes as indicated, and purified from gels using “MinElute” (Qiagen).

• Figure 3 - Actin phosphorylation by CaMKIIβ
  Phosphorylation of actin and CaMKII autophosphorylation were detected by ³²P-incorporation and autoradiography. (A) The actin-sized band phosphorylated by basal activity of purified CaMKIIβ but not α was not observed when actin was omitted (-) or substituted with BSA. Exposures without (upper panel) or with (lower panel) enhancer screen are shown. (B) Ca²⁺/CaM-stimulated or basal phosphorylation of actin was seen in presence of purified CaMKIIβ, but not when kinase was omitted (upper panels). Presence of similar amount of actin in the reactions was confirmed by Western-blot analysis with an actin-specific antibody (lower panels). The radioactive ATP had 2-fold higher specific activity in the basal phosphorylation reaction compared to the stimulated reaction. Nevertheless, ratio of basal to stimulated actin phosphorylation by CaMKIIβ was higher compared to the experiment shown in Figure 6. The difference between the experiments was that the kinase/actin mix was added to the stimulated reaction first in Figure 6, and to the basal reaction first here. The resulting 1 min delay in start of the reaction may cause partial occlusion of the ³²P signals by basal phosphorylation utilizing the unlabelled ATP present in the actin polymerization buffer. (C) Basal activity of GFP-CaMKIIβ wild type (β) in extracts of transfected Cos-7 cells phosphorylated actin more than extracts with the ATP-binding impaired mutant K43R (K43R) or mock transfected extracts (cos) (left panel). Reactions were adjusted for total Cos-7 protein and kinase amount. As loading control, relative kinase and actin amounts in the reactions were tested by Western-blot analysis with specific antibodies (left panel).