Alternative splicing controls myotonic dystrophy protein kinase structure, enzymatic activity, and subcellular localization

Abstract

Transcripts of the myotonic dystrophy protein kinase (DMPK) gene, a member of the Rho kinase family, are subject to cell-type specific alternative splicing. An imbalance in the splice isoform profile of DMPK may play a role in the pathogenesis of DM1, a severe multisystemic disorder. Here, we report how structural subdomains determine biochemical properties and subcellular distribution of DMPK isoforms. A newly developed kinase assay revealed that DMPK is a Lys/Arg-directed kinase. Individual DMPK isoforms displayed comparable transphosphorylation activity and sequence preference for peptide substrates. However, DMPK autophosphorylation and phosphorylation of MYPT1 (as putative in vivo target of DMPK), were dependent on presence of an alternatively spliced VSGGG motif and the nature of the C terminus. In-gel effects of the VSGGG motif on the migration behavior of full-length kinase provide evidence for a model in which this motif mediates 3-D-conformational changes in DMPK isoforms. Finally, different C termini endow DMPK with the ability to bind to either endoplasmic reticulum or mitochondria or to adopt a cytosolic location. Our results suggest that DMPK isoforms have cell-type and location dependent substrate specificities with a role in organellar and cytoarchitectural dynamics.

FIG. 1.

DMPK: domain organization and homology to serine/threonine kinase family members. (A) Major DMPK isoforms A to F have an N-terminal leucine-rich domain (L), a serine/threonine kinase domain, a protein kinase C-terminal domain containing the hydrophobic phosphorylation motif, and a coiled coil region. Differences between isoforms originate from alternative splicing, conserved between humans and mice: (i) a VSGGG-sequence can be present (isoforms A, C, and E) or absent (isoforms B, D, and F) and (ii) three different C-terminal tails occur. Minor splice form DMPK G, only present in humans, carries a fourth type of C terminus, of which the N-terminal half is identical to tail 1. (B) Sequence comparison between mDMPK, rMRCKα, mROCK-I, mNDR1, and mPKBα. Only the first 412 aa of DMPK are shown, since no relevant homology exists with the other kinases beyond this point. Identical amino acids (in at least three of the five kinases) are shown in white on a black background, and similar amino acids are shown in black on a grey background. The kinase domain is indicated with a dotted line below the sequence, the VSGGG sequence is underlined, and the hydrophobic phosphorylation motif is doubly underlined. The total number of amino acids for each full-length protein is indicated in parentheses; note that rMRCKα and mROCK-I are very large proteins compared to mDMPK (accession numbers: P54265, T14039, S74244, AAH09658, and P31750). (C) Sequence identity between mDMPK, rMRCKα, ROCK-I, mNDR1, and mPKBα. The N terminus (aa 1 to 70), the kinase domain (aa 71 to 339), and the protein kinase C-terminal domain (aa 340 to 405) of DMPK were compared with the corresponding parts in rMRCKα, mROCK-I, mNDR1, and mPKBα using ClustalW. The relative sequence identity for each domain is expressed as a percentage relative to mDMPK. Values for the protein kinase C-terminal domain without the VSGGG motif (as in DMPK B, D, and F) are listed in parentheses. Similar results were obtained with rMRCKβ and mROCK-II (data not shown).

FIG. 1.

DMPK: domain organization and homology to serine/threonine kinase family members. (A) Major DMPK isoforms A to F have an N-terminal leucine-rich domain (L), a serine/threonine kinase domain, a protein kinase C-terminal domain containing the hydrophobic phosphorylation motif, and a coiled coil region. Differences between isoforms originate from alternative splicing, conserved between humans and mice: (i) a VSGGG-sequence can be present (isoforms A, C, and E) or absent (isoforms B, D, and F) and (ii) three different C-terminal tails occur. Minor splice form DMPK G, only present in humans, carries a fourth type of C terminus, of which the N-terminal half is identical to tail 1. (B) Sequence comparison between mDMPK, rMRCKα, mROCK-I, mNDR1, and mPKBα. Only the first 412 aa of DMPK are shown, since no relevant homology exists with the other kinases beyond this point. Identical amino acids (in at least three of the five kinases) are shown in white on a black background, and similar amino acids are shown in black on a grey background. The kinase domain is indicated with a dotted line below the sequence, the VSGGG sequence is underlined, and the hydrophobic phosphorylation motif is doubly underlined. The total number of amino acids for each full-length protein is indicated in parentheses; note that rMRCKα and mROCK-I are very large proteins compared to mDMPK (accession numbers: P54265, T14039, S74244, AAH09658, and P31750). (C) Sequence identity between mDMPK, rMRCKα, ROCK-I, mNDR1, and mPKBα. The N terminus (aa 1 to 70), the kinase domain (aa 71 to 339), and the protein kinase C-terminal domain (aa 340 to 405) of DMPK were compared with the corresponding parts in rMRCKα, mROCK-I, mNDR1, and mPKBα using ClustalW. The relative sequence identity for each domain is expressed as a percentage relative to mDMPK. Values for the protein kinase C-terminal domain without the VSGGG motif (as in DMPK B, D, and F) are listed in parentheses. Similar results were obtained with rMRCKβ and mROCK-II (data not shown).

FIG. 2.

Analysis of DMPK isoforms and mutants by Western blotting. Individual mouse DMPK cDNA products were analyzed by Western blotting using an anti-DMPK antibody. (A) Expression of isoforms DMPK A to F. The presence or absence of the VSGGG motif, the type of C-terminal tail, and the theoretical molecular weights are summarized below the lanes. The apparent molecular weights of the main bands are indicated (arrows). All isoforms with a VSGGG motif showed two distinct bands (arrowheads), whereas isoforms without a VSGGG motif showed only one (sometimes broad) band. (B) Expression of His-DMPK A through F kd mutants. All kd isoforms showed only one band (i.e., the lower band in A). (C) The effect of in vivo treatment with the phosphatase inhibitor OA (1 μM, 1 h) was tested for DMPK E, His-DMPK E kd, and DMPK E^VA. OA treatment changed the doublet ratios of DMPK E and E^VA but did not change the migration of His-DMPK E kd. Note that due to the His tag (2.4 kDa) DMPK E kd migrated slightly more slowly than the regular DMPK E isoform. Shown are results obtained by immunoprecipitation. (D) Ser379 mutations alter DMPK E mobility. DMPK E and mutants DMPK E^VA and E^VD were expressed and treated with OA prior to lysis as in panel C. Lysis was done at −80°C in 10% TCA-10 mM DTT-acetone to preserve intracellular phosphorylation status.

FIG. 3.

Development of an in vitro protein kinase assay for DMPK. Peptides were tested for phosphorylation by HA-DMPK E as described in Materials and Methods. The extent of phosphorylation of each peptide is expressed as the percentage relative to KKRNRRLSVA (peptide 2). The serine or threonine phosphoacceptor (position 0) is shown in boldface type. Bars indicate standard errors. For comparison, peptides are assembled in two groups (A and B).

FIG. 4.

Differential kinase activity and (auto)phosphorylation of DMPK isoforms and mutants. (A) HA-DMPK isoforms A to G, E^VA and E^VD were expressed in COS-1 cells. Cell extracts were prepared and tested for HA-DMPK by Western blotting (upper panel). Extracts were used in a kinase assay with substrates KKRNRRLSVA (peptide 2) or KKLRRTLSVA (peptide 17) (middle panel). All DMPK isoforms and mutants showed comparable activities towards these peptides. Immunoprecipitated DMPK (after performing assay with peptide 2) was blotted to PVDF membrane to control input (Coomassie brilliant blue [CBB]staining) and examine DMPK autophosphorylation. All isoforms were autophosphorylated (i.e., upper bands in VSGGG-containing isoforms), but isoforms lacking VSGGG showed lower autophosphorylation activity. (B) HA-DMPK isoforms A to G were used in a kinase assay with full-length MYPT1 protein as substrate. Input of MYPT1 and DMPK was verified by Western blotting (upper panel; note aspecific anti-HA band, unrelated to DMPK, in vector-only lane [lane v]). Phosphorylation of MYPT1 and autophosphorylation of DMPK became evident by autoradiography, which was quantified by phosphorimager analysis (lower two panels; values in arbitrary units, after background [lane v] subtraction). (C) COS-1 cells expressing HA-DMPK isoforms and mutants were cultured in the presence of [³²P]orthophosphate to examine in vivo phosphorylation of DMPK. DMPK was immunoprecipitated, separated by SDS-PAGE, and blotted to PVDF membrane. The blot was used for autoradiography (middle panel) and phosphorimager analysis (lower panel). Values are presented in arbitrary units, after background (vector-only lane [lane v]) subtraction. Note that kd mutants E and G were less well expressed (upper panel), which partly explains their low signals on autoradiography.

FIG. 4.

Differential kinase activity and (auto)phosphorylation of DMPK isoforms and mutants. (A) HA-DMPK isoforms A to G, E^VA and E^VD were expressed in COS-1 cells. Cell extracts were prepared and tested for HA-DMPK by Western blotting (upper panel). Extracts were used in a kinase assay with substrates KKRNRRLSVA (peptide 2) or KKLRRTLSVA (peptide 17) (middle panel). All DMPK isoforms and mutants showed comparable activities towards these peptides. Immunoprecipitated DMPK (after performing assay with peptide 2) was blotted to PVDF membrane to control input (Coomassie brilliant blue [CBB]staining) and examine DMPK autophosphorylation. All isoforms were autophosphorylated (i.e., upper bands in VSGGG-containing isoforms), but isoforms lacking VSGGG showed lower autophosphorylation activity. (B) HA-DMPK isoforms A to G were used in a kinase assay with full-length MYPT1 protein as substrate. Input of MYPT1 and DMPK was verified by Western blotting (upper panel; note aspecific anti-HA band, unrelated to DMPK, in vector-only lane [lane v]). Phosphorylation of MYPT1 and autophosphorylation of DMPK became evident by autoradiography, which was quantified by phosphorimager analysis (lower two panels; values in arbitrary units, after background [lane v] subtraction). (C) COS-1 cells expressing HA-DMPK isoforms and mutants were cultured in the presence of [³²P]orthophosphate to examine in vivo phosphorylation of DMPK. DMPK was immunoprecipitated, separated by SDS-PAGE, and blotted to PVDF membrane. The blot was used for autoradiography (middle panel) and phosphorimager analysis (lower panel). Values are presented in arbitrary units, after background (vector-only lane [lane v]) subtraction. Note that kd mutants E and G were less well expressed (upper panel), which partly explains their low signals on autoradiography.

FIG. 5.

Subcellular localization of DMPK isoforms and mutants in COS-1 cells. COS-1 cells were transfected with expression vectors encoding DMPK isoforms and mutants and grown for 24 h. Cells were fixed with 2% formaldehyde and stained with a DMPK-specific antibody. The isoform or mutant shown is indicated in the lower left of each panel. Bar,10 μm. The insets in A, C, and E show a higher magnification of parts indicated with an arrow. The bar in the insets is 2 μm.

FIG. 6.

Tail 1 and tail 2 target DMPK isoforms to the ER or mitochondria, respectively. HA-DMPK isoforms were transfected to N2A cells to examine their subcellular localization in more detail. Colocalization with an ER-resident GFP mutant demonstrated localization of DMPK A to the ER (A to C). Similarly, colocalization with cytochrome c oxidase, a mitochondrial protein, showed that DMPK C is confined to mitochondria (D to F). Bars, 10 μm.

FIG. 7.

Alternatively spliced protein domains. The VSGGG motif and the C terminus determine differential functioning of DMPK isoforms.