Báo cáo khoa học: Coiled–coil interactions modulate multimerization, mitochondrial binding and kinase activity of myotonic dystrophy protein kinase splice isoforms pptx

Coiled–coil interactions modulate multimerization,
mitochondrial binding and kinase activity of myotonic
dystrophy protein kinase splice isoforms
Rene
´
E. M. A. van Herpen, Jorrit V. Tjeertes, Susan A. M. Mulders, Ralph J. A. Oude Ophuis,
Be
´
Wieringa and Derick G. Wansink
Department of Cell Biology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, the Netherlands
Myotonic dystrophy protein kinase (DMPK) was first
identified and cloned as a protein kinase in the quest
to establish the molecular basis of disease in myotonic
dystrophy, now more than a decade ago. Study of the
structure–function relationship of domains in DMPK
and homologous kinases, such as the myotonic dystro-
phy kinase-related Cdc42-binding kinase (MRCKa ⁄ -b-
c) [1,2], ROCK-I ⁄ -II [3] and Citron kinase [4], placed
DMPK in the large AGC group of protein kinases
[5–7]. MRCKs, ROCKs and Citron kinase regulate
and reorganize the actin-based cytoskeleton as effectors
of the small GTPases Cdc42 or Rho [8]. Their kinase
activity controls the status of myosin regulatory light
chain phosphorylation, either directly or indirectly via
regulation of myosin phosphatase activity, thereby
affecting stress fiber formation, smooth muscle contrac-
tion or cytokinesis [9–11]. Although myosin phospha-
tase targeting subunit 1 (MYPT1) has been identified as
a substrate for DMPK [5,12], the effects of DMPK-
mediated phosphorylation on actomyosin dynamics
Keywords

coiled-coil domain; multimerization;
myotonic dystrophy protein kinase; protein–
protein interaction; Rho kinase family
Correspondence
D. G. Wansink, Department of Cell Biology
(code 283), NCMLS, Radboud University
Nijmegen Medical Centre, PO Box 9101,
6500 HB Nijmegen, the Netherlands
Fax: +31 24 3615317
Tel: +31 24 3613664 ⁄ 14329
E-mail:
Website:
(Received 24 November 2005, revised 11
January 2006, accepted 16 January 2006)
doi:10.1111/j.1742-4658.2006.05138.x
The myotonic dystrophy protein kinase polypeptide repertoire in mice and
humans consists of six different splice isoforms that vary in the nature of
their C-terminal tails and in the presence or absence of an internal Val–
Ser–Gly–Gly–Gly motif. Here, we demonstrate that myotonic dystrophy
protein kinase isoforms exist in high-molecular-weight complexes controlled
by homo- and heteromultimerization. This multimerization is mediated by
coiled–coil interactions in the tail-proximal domain and occurs independ-
ently of alternatively spliced protein segments or myotonic dystrophy
protein kinase activity. Complex formation was impaired in myotonic dys-
trophy protein kinase mutants in which three leucines at positions a and d
in the coiled-coil heptad repeats were mutated to glycines. These coiled-coil
mutants were still capable of autophosphorylation and transphosphoryla-
tion of peptides, but the rates of their kinase activities were significantly
lowered. Moreover, phosphorylation of the natural myotonic dystrophy
protein kinase substrate, myosin phosphatase targeting subunit, was pre-

served, even though binding of the myotonic dystrophy protein kinase to
the myosin phosphatase targeting subunit was strongly reduced. Further-
more, the association of myotonic dystrophy protein kinase isoform C to
the mitochondrial outer membrane was weakened when the coiled–coil
interaction was perturbed. Our findings indicate that the coiled-coil domain
modulates myotonic dystrophy protein kinase multimerization, substrate
binding, kinase activity and subcellular localization characteristics.
Abbreviations
CM
, coil mutant; DMPK, myotonic dystrophy protein kinase; DSP, dithiobis (succinimidyl propionate); ER, endoplasmic reticulum;
MOM, mitochondrial outer membrane; MRCK, myotonic dystrophy kinase-related Cdc42-binding kinase; MYPT, myosin phosphatase
targeting subunit.
1124 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS
have not yet been studied in detail. In addition, the role
of other domains in DMPK, and their possible involve-
ment in the regulation of catalytic kinase activity, has,
to date, remained elusive. Homology comparison can
help, as members of the DMPK family have, to some
degree, a similar domain arrangement [3,5,13].
In all members of the DMPK family of protein kin-
ases, a conserved leucine-rich N terminus of  70
amino acids precedes the catalytic kinase domain,
which is followed by a characteristic coiled-coil region
at the C-terminal end [1,4,14]. In DMPK, next to these
shared protein domains, two alternatively spliced
domains were identified (a) a five-amino acid Val–Ser–
Gly–Gly–Gly (VSGGG) sequence and (b) the DMPK
C terminus [15]. In vitro study of one of the six major
DMPK splice isoforms has revealed that a relationship
must exist between kinase activity and the state of

multimerization promoted by the N terminus and the
coiled-coil domain [16,17]. However, the consequences
of multimerization and association with other proteins
for in vivo activity regulation of DMPK are not clear.
Activation of Citron kinase and ROCK-I is mediated
by RhoA binding to a Rho-binding domain located in
the C-terminal part of the coiled-coil region 1[4,8,19]. In
this respect, the coiled-coil region fulfils a regulatory
role, as RhoA binding relieves inhibition imposed by
the C terminus on the kinase domain [20]. Furthermore,
the coiled-coil segment seems to carry out a special role
in regulating the multimeric state of ROCK-I and
MRCKa, thereby regulating their kinase activity
[14,20,21]. Dimer formation mediated by the N terminus
of MRCKa is followed by transautophosphorylation
and also contributes to regulation of the MRCKa cata-
lytic activity [14]. For DMPK, it has been reported that
myotonic dystrophy protein kinase-binding protein
enhances DMPK catalytic activity [22]. In addition,
binding of the Rho-family member, Rac-1, and phos-
phorylation by Raf-1, serve as activating events [23].
We have recently reported that the different alternat-
ive C termini anchor DMPK isoforms in distinct intra-
cellular membranes, targeting DMPK isoforms A and
B to the endoplasmic reticulum (ER) and DMPK C
and D to the mitochondrial outer membrane (MOM)
[5,24]. Specific elements in the coiled-coil domain
exclusively affect the mitochondrial but not the ER
targeting behavior [24]. Short DMPK isoforms E and
F, containing a two-amino acid C terminus following

the coiled-coil domain, adopt a cytosolic localization.
Furthermore, the VSGGG motif, unique among AGC
kinases, regulates DMPK autophosphorylation, in-gel
migration behavior and, probably, folding [5].
To fully understand how the individual DMPK iso-
form structure relates to function, we focus here on
the significance of the coiled-coil domain in the regula-
tion of multimerization, kinase activity and localiza-
tion behavior. Using biochemical and cell biological
approaches, we demonstrate that the tendency of
DMPK to multimerize to higher-molecular weight
complexes relies on typical structural sequence proper-
ties of the coiled-coil segment, independent of kinase
activity or the presence of other alternatively spliced
domains. Reciprocal effects were also seen because
coiled–coil interactions modulated, but did not abolish,
autophosphorylation ability, transphosphorylation
activity towards peptide substrates in vitro, complex
formation with the DMPK substrate MYPT2 and, in
the case of DMPK C, localization to mitochondria
in vivo.
Results
Individual DMPK isoforms exist in high-molecular-
weight complexes
DMPK isoforms (Fig. 1) differ in localization, enzy-
matic activity and autophosphorylation, owing to
alternatively spliced domains [5,24]. As a first step
towards a better understanding of the role of the
coiled-coil segment in the differential structure–func-
tion properties of the different isoforms, we performed

gel filtration experiments to obtain a size estimate of
the complexes in which DMPK isoforms can reside.
Cleared lysates of cells expressing DMPK A, C, E or
F were applied to a Superose gel-filtration column
calibrated with standard molecular weight markers.
Elution profiles were traced with western blotting.
Full-length DMPK isoforms A and C (predicted
molecular mass  70 kDa; apparent molecular mass
 75 kDa on an SDS ⁄ PAGE gel) eluted as large com-
plexes, with the main signal exceeding 440 kDa in
molecular mass (Fig. 2). Breakdown products of these
large DMPKs, inevitably formed by in vitro product
handling in the experimental procedures used,
appeared predominantly in the same fractions as the
BSA marker protein ( 67 kDa). Splice isoforms E
and F (predicted molecular mass  60 kDa; apparent
molecular mass  68 kDa on an SDS ⁄ PAGE gel) were
also found in high-molecular-weight complexes, but
did not yield any breakdown products because they
lack the long C-terminal tail domains that confer pro-
teolytic vulnerability (see Fig. 1) [5,15]. The elution
profiles of isoforms E and F, which only differ in the
presence of a VSGGG motif, were comparable, sug-
gesting that the VSGGG motif has no role in complex
formation (note the characteristic doublet signal for
isoforms containing a VSGGG motif, which is related
R. E. M. A. van Herpen et al. DMPK coiled-coils mediate multimerization
FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS 1125
to autophosphorylation) [5]. Taken together, these
experiments reveal that all individual DMPK isoforms

occur in large, multimeric complexes.
The coiled-coil domain mediates DMPK
multimerization
To investigate whether multimerization is involved in
the formation of large complexes, and to identify the
protein domains involved, we introduced various
N-terminal and C-terminal truncation mutations into
DMPK isoform E (Fig. 1). COS-1 cells were doubly
transfected with full-length HA-DMPK E and one
VSV-tagged truncation mutant, and extracts were tes-
ted in co-immunoprecipitation experiments (Fig. 3A).
HA-DMPK E did not precipitate the N-terminal
region fused to the kinase domain, irrespective of the
presence of the VSGGG motif. Similarly, the kinase
domain alone, VSV-DMPK E(60–375), did not inter-
act with full-length HA-DMPK E. In contrast, the
C-terminal region of DMPK E, containing the coiled-
coil domain, with or without the VSGGG motif [i.e.
constructs E(340–537) and E(402–537)], did associate
with HA-DMPK E. The specificity of this interaction
was also observed in a reverse immunoprecipitation
using the anti-VSV immunoglobulin (Fig. 3B). Inde-
pendently, gel filtration confirmed that VSV-DMPK
E(402–537) participated in high-molecular-weight com-
plexes (Fig. 2). Together, our findings suggest that
only the coiled-coil domain is relevant for interaction.
DMPK homo- and heteromultimerization occurs
independently of kinase activity and alternatively
spliced domains
To confirm that DMPK self-association indeed occurred

independently of kinase activity, a kinase dead mutant
was tested in a co-immunoprecipitation experiment. As
evident from the results shown in Fig. 4A, both trunca-
tion mutants containing the coiled-coil domain copre-
cipitated with HA-DMPK E(K100A), a kinase-inactive
variant impaired in ATP binding owing to a lysine to
alanine mutation in the kinase domain [5].
DMPK splice isoforms are expressed in a cell-type-
specific manner. The main isoforms in skeletal muscle,
Fig. 2. Myotonic dystrophy protein kinase (DMPK) isoforms reside
in high-molecular-weight complexes. Size exclusion chromatogra-
phy was performed on cell-free extracts of transfected COS-1 cells
containing DMPK A, C, E, F or VSV-DMPK E(402–537). Fractions
were analyzed on western blots using an anti-DMPK immuno-
globulin. Molecular mass markers ribonuclease A (14 kDa), BSA
(67 kDa), aldolase (150 kDa) and ferritin (440 kDa) were used to cal-
ibrate the fraction volume positions of differently sized proteins in
the column eluate (indicated on top). Molecular mass markers for
SDS ⁄ PAGE are indicated at the left (30, 65 and 83 kDa). All full-
length isoforms and also truncation mutant VSV-DMPK E(402–537)
were found in large multimeric complexes. During the procedure,
loss of the C terminus occurred for isoforms A and C, which has
been described previously [5,15].
Fig. 1. Myotonic dystrophy protein kinase (DMPK) isoforms and truncation mutants. Structural domain organization of DMPK splice isoforms
A, C, E and F, and N-terminal and C-terminal truncation mutants of DMPK E used in this study, are shown. The N terminus, Ser ⁄ Thr protein
kinase domain, alternatively spliced VSGGG motif, coiled-coil domain and alternatively spliced tail regions are indicated. Truncation mutants
contain an N-terminal VSV-tag. Numbers refer to amino acid numbering in full-length DMPK E and indicate the first and last amino acid of
the DMPK segment contained within the mutants. The ability to multimerize is indicated (see the text).
DMPK coiled-coils mediate multimerization R. E. M. A. van Herpen et al.
1126 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS

heart and brain are DMPK A–D, whereas DMPK E
and F predominate in smooth muscle tissue [15]. It
was previously shown that the VSGGG motif in
DMPK A, C and E enhances autophosphorylation [5].
To study a potential modulatory effect of the VSGGG
motif on DMPK self-association, the interaction
between DMPK isoforms E (includes a VSGGG
sequence) and F (no VSGGG sequence) was examined.
Again, when His tags and HA tags were used to dis-
criminate between the different splice isoform partners,
we observed homomultimerization by DMPK isoforms
E and F, as well as heteromultimerization between
isoforms E and F (Fig. 4B). To investigate whether
different tail regions affected DMPK self-association
(Fig. 1), we tested isoform combinations A with C, A
with E, and C with E, and again found all of these iso-
AB
*
Fig. 3. The coiled-coil domain mediates self-association of myotonic dystrophy protein kinase (DMPK) E. (A) Full-length HA-DMPK E was
co-expressed with VSV-tagged DMPK truncation mutants in COS-1 cells, as indicated on top (+, present; ), not present). Immunoprecipita-
tions (IP) were carried out using anti-HA-coated beads. DMPK E interacted only with mutants containing a coiled-coil domain [i.e. E(340–537)
and E(402–537)]. (B) The coiled–coil mediated interaction was confirmed by reverse IPs with anti-VSV on lysates containing untagged DMPK
E and VSV-DMPK E(340–537) or VSV-E(402–537). Precipitated proteins were detected on western blots with anti-HA, anti-VSV or anti-DMPK,
as indicated. The asterisk indicates a nonspecific signal detected by the anti-HA immunoglobulin in whole cell lysates.
A
BC
*
Fig. 4. Multimerization of myotonic dystrophy protein kinase (DMPK) is independent of kinase activity or alternatively spliced domains. (A)
Involvement of kinase activity in DMPK complex formation was investigated by immunoprecipitation (IP) using lysates expressing the DMPK
inactive mutant E(K100A) and truncation mutants VSV-DMPK E(340–537) or E(402–537) as indicated on top (+, present; ), not present).

(B) To investigate the effects of the VSGGG motif on multimerization, HA- and His-tagged versions of DMPK E and F were used in IPs, as
indicated. The asterisk indicates a nonspecific signal detected by the anti-HA immunoglobulin in whole cell lysates. (C) Involvement of the C
terminus in multimerization was examined by the expression of combinations of YFP–DMPK A or C and His-DMPK C or E, as indicated. IPs
were performed using anti-HA or anti-YFP, and western blots were probed with anti-DMPK, anti-His, anti-HA or anti-YFP, as indicated.
R. E. M. A. van Herpen et al. DMPK coiled-coils mediate multimerization
FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS 1127
forms to interact independently of their C-terminal
sequence (Fig. 4C; data not shown). Taken together,
these results provide strong evidence that DMPK mul-
timerization is an intrinsic property of all DMPK
isoforms, and that homomultimers as well as hetero-
multimers can be formed when the 402–537 region is
present.
Mutations in the coiled-coil region impair DMPK
multimerization
We used coils, paircoil and multicoil algorithms to
compute the probability of coiled-coil formation across
the stretch of amino acids between positions 340
through 537 in DMPK E [25–27]. Nine heptad-repeat
sequences were identified by these algorithms, predict-
ing a coiled-coil probability of almost 1 for the entire
segment between amino acids 469 and 531 (Fig. 5A).
Combined with the knowledge that 3.5 amino acids
are needed to complete one turn in a coil, this suggests
that 18 turns make up the entire DMPK a-helical coil
[28]. At the a and d positions of every heptad, known
to make up the typical hydrophobic interface between
interacting coiled-coil domains [28], over 50% of the
residues are of hydrophobic nonaromatic nature (i.e.
leucine, isoleucine or valine) (Fig. 5B). Lysine residues,

which are commonly found at electrostatic residue
positions e and g in the heptad, are absent in the
DMPK coiled coil. Instead, arginines were found at
these positions in heptads IV, VI and VIII. Database
comparison of the DMPK coiled-coil region showed
homology to distinct parts of the large coiled-coil
region of myosin heavy chain, MRCKa ⁄ -b ⁄ -c and
ROCK-I ⁄ -II ( 25% identity;  60% similarity) [5].
In order to approach the anticipated structural role
of the coil experimentally, we mutated amino acid resi-
dues at the a and d positions within heptad repeats II,
III and VII, because these are known to be of crucial
importance for coiled-coil formation. Leucine to gly-
cine changes at positions 477, 487 and 515 were intro-
duced, because, according to the paircoil algorithm,
these would lower the coiled-coil probability to < 0.5
(Fig. 5A). Transfection in COS-1 cells showed that
expression levels of the HA-DMPK E mutant with
point mutations L477G, L487G and L515G [hereafter
designated HA-DMPK E coil mutant (HA-DMPK
E
CM
)] and HA-DMPK E were similar (Fig. 6A).
HA-DMPK E
CM
migrated more slowly in the gel,
indicating that its protein conformation had indeed
changed as a result of the Leu to Gly mutations. This
altered migration was specifically caused by the L477G
mutation, as the two other mutations did not contrib-

ute to the effect (data not shown).
Immunoprecipitation of HA-DMPK E
CM
was less
efficient than of HA-DMPK E. When using a stepwise
increasing series of antiserum concentrations, the
amount of precipitated HA-DMPK E
CM
was consis-
tently lower than that of HA-DMPK E, indicating im-
paired self-association of HA-DMPK E
CM
(Fig. 6B).
Co-immunoprecipitation with differentially tagged
DMPK E confirmed that DMPK E
CM
could not
engage in homodimerization. This was confirmed when
HA-DMPK E
CM
was unable to associate with and pull
down full-length His-DMPK E under conditions of
excess anti-HA beads (Fig. 6C). Furthermore, the
association of DMPK E
CM
with truncation mutants
DMPK E(340–537) and E(402–537) was not observed
(data not shown), indicating that leucines 477, 487 and
Fig. 5. Prediction of coiled-coil probability and heptad repeats in myotonic dystrophy protein kinase (DMPK). (A) Computational analysis of
the coiled-coil forming probability of the DMPK segment spanning amino acids 340–537 using the programs

COIL, PAIRCOIL and MULTICOIL.
When leucine at positions 477, 487 and 515 were replaced by glycine, as in the DMPK E coil mutant (DMPK E
CM
), the predicted coil
between amino acids 469–531 dropped below 50% probability (
PAIRCOIL). (B) Assignment of a heptad repeat register (a–g) for amino acids
469–531 of DMPK based on predictions made by
COIL, PAIRCOIL and MULTICOIL. The DMPK coiled-coil domain contains nine heptad repeats,
indicated by Roman numerals. Leucine to glycine mutations in DMPK E
CM
are boxed.
DMPK coiled-coils mediate multimerization R. E. M. A. van Herpen et al.
1128 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS
515 within the coiled-coil domain are essential for the
self-association behavior of DMPK E.
Coiled-coil mutations reduce, but do not abolish,
DMPK kinase activity
Earlier work of our group and others has demonstra-
ted that DMPK E phosphorylates MYPT1 [5] (data
not shown) and thereby inhibits myosin phosphatase
activity [12]. We tested here whether impaired coiled–
coil interactions would affect DMPK substrate binding
and kinase activity. MYPT2, a paralogue of MYPT1
[29] was co-expressed with DMPK E or E
CM
in COS-1
cells and their interaction analyzed by immunoprecipi-
tation. A small, but significant, fraction of MYPT2
was complexed to DMPK E, but no interaction could
be detected with the coil mutant (Fig. 7A). Owing to

incomplete reduction of the reversible chemical cross-
linker dithiobis (succinimidyl propionate) (DSP), used
to stabilize the DMPK–MYPT2 interaction and only
included for DMPK–MYPT2 co-expression studies, a
considerable fraction of DMPK E complexes migrated
as high-molecular-weight structures in the gel (asterisks
in Fig. 7A). This was also observed for lysates that
contained DMPK E
CM
, albeit at a lower signal inten-
sity. The most simple explanation for the latter obser-
vation is that, despite the perturbed coiled-coil
domain, other parts of the protein could still be
involved in association behavior.
We next examined, in an in vitro kinase assay, whe-
ther the association (i.e. as assessed by immunoprecipi-
tation) between DMPK and MYPT2 was a
prerequisite for MYPT2 phosphorylation. Much to
our surprise, MYPT2 was phosphorylated by DMPK
E and DMPK E
CM
at almost similar efficiency
(Fig. 7B). In addition, autophosphorylation was still
present in DMPK E
CM
, but this was two- to threefold
lower than in wild-type DMPK E (Fig. 7B). More
quantitatively, we examined how coil mutations affec-
ted kinase activity in an assay based on the preferred
DMPK peptide substrate, KKRNRRLTVA [5]. Under

these conditions, both peptide phosphorylation and au-
tophosphorylation by DMPK E
CM
was approximately
threefold lower than peptide phosphorylation and au-
tophosphorylation by DMPK E (Fig. 7C). By exam-
ination of steady-state levels and structural intactness
of DMPK protein products, we ruled out that altered
proteolytic processing was underlying this effect (data
not shown). Combined, these results thus suggest that
the coil region must have a facilitating, rather than an
essential, role in the determination of DMPK activity
and specificity.
Fig. 6. Coiled-coil mutations in myotonic dystrophy protein kinase (DMPK) E impair self-association. (A) Expression of the HA-DMPK E coil
mutant (HA-DMPK E
CM
) in COS-1 cells was examined by western blotting using an anti-HA immunoglobulin. Using HA-DMPK E as a refer-
ence, HA-DMPK E
CM
displayed an altered gel mobility (+, present; ), not present). The upper band of the HA-DMPK E
CM
doublet co-migra-
ted with a nonspecific signal (marked with an asterisk). (B) Immunoprecipitations on extracts containing HA-DMPK E or HA-DMPK E
CM
were
carried out with a series of increasing concentrations of anti-HA immunoglobulin. DMPK products in cell lysates (CL) and immunoprecipitates
(IP) were probed with anti-HA on western blots. Loading was 1% of total for the cell lysate and 2% of total for the immunoprecipitates.
Brackets indicate positions of heavy chains of the anti-HA immunoglobulin. The graph shows the immunoprecipitation efficiency for each
protein, determined by densitometrical scanning of the DMPK signal and plotted against the concentration of anti-HA immunoglobulin used
in the immunoprecipitation. (C) Immunoprecipitations were performed on lysates containing His-DMPK E together with either HA-DMPK E

or HA-DMPK E
CM
(+, present; ), not present). DMPK proteins were probed with an anti-HA or anti-His immunoglobulin on a western blot.
R. E. M. A. van Herpen et al. DMPK coiled-coils mediate multimerization
FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS 1129
The coiled-coil domain stabilizes DMPK C
interaction with mitochondria
Immediately C-terminal to the coiled-coil segment,
spaced by a stretch of only five amino acids, is the
alternatively spliced DMPK C terminus, which deter-
mines subcellular targeting to the ER (mDMPK A),
the MOM (mDMPK C), or the cytosol (DMPK E) [5].
We have reported that the presence of the coiled-coil
region in DMPK C is essential for MOM anchoring,
but it remained unclear whether this effect should be
attributed to structural integrity of the entire coiled-
coil domain, or to properties of any particular amino
acid segment therein [24]. We compared the localiza-
tion of YFP-tagged isoforms A, C and E, containing
the L477G, L487G and L515G mutations, with that of
the wild type YFP–DMPK isoforms (Fig. 8). As
shown in Fig. 8A–F, YFP–DMPK C
CM
was parti-
tioned over MOM and the cytosol, whereas nonmutat-
ed YFP–DMPK C was located uniquely at
mitochondria. To us this suggests that coiled coil-
mediated associations contribute to DMPK–MOM
Fig. 7. The myotonic dystrophy protein kinase E coil mutant (DMPK E
CM

) displays reduced kinase activity. (A) Myosin phosphatase targeting
subunit 2 (MYPT2) interaction with DMPK E depends on an intact coiled-coil domain. HA-DMPK E or E
CM
were co-expressed with VSV-MYPT2
in COS-1 cells (+, present; ), not present). Lysates were prepared in the presence of the cross-linker, dithiobis (succinimidyl propionate), and
used in immunoprecipitations with anti-HA beads. Proteins in the cell lysate (input) and after immunoprecipitation (IP) were probed on a western
blot with anti-DMPK or anti-MYPT immunoglobulin. The asterisk marks slow-migrating complexes resistant to decrosslinking, which contain
DMPK. (B) DMPK E
CM
showed reduced autophosphorylation, but was capable of phosphorylating MYPT2. Immunopurified MYPT2 was used in
a kinase assay with purified HA-DMPK E, E
CM
or E(K100A) (+, present; ), not present). Western blotting was used to validate the input of
DMPKs and MYPT. Phosphorylation of MYPT2 and autophosphorylation of DMPK were visualized by autoradiography and quantified by phos-
phoimager analysis. The MYPT2 background signal, caused by copurifying kinase activity and nearly equal to the signal observed using HA-
DMPK E(K100A), was subtracted [5]. (C) DMPK E
CM
displayed reduced kinase activity towards a peptide substrate. Immunopurified HA-DMPK
E, E
CM
and E(K100A) were used in a kinase assay with preferred peptide substrate, KKRNRRLTVA [5] (+, present; ), not present). Input of
DMPK in the assay was validated by western blotting.
DMPK coiled-coils mediate multimerization R. E. M. A. van Herpen et al.
1130 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS
binding strength, but are not essential. Although con-
sidered less likely, an alternative explanation could be
that abnormal properties of the coiled-coil structure
render the tail in DMPK C less avid to engage in
MOM binding. We also examined whether the coil
mutations had any effect on the distribution of DMPK

A (present in the ER membrane) and DMPK E (cyto-
solic variant). Figure 8G–I show that the locations of
the DMPK A
CM
and E
CM
remained unchanged, corro-
borating previous findings that unique properties of
the A and C tails drive localization [24].
Discussion
The results presented here provide evidence for the
contention that the intact coiled-coil region, presuma-
bly by means of its unique helical ⁄ structural proper-
ties, is a key factor in aggregation behavior and a
modifier of biological properties of the adjacent
domains (i.e. the kinase and tail domains) in DMPK.
The coiled-coil region thus codetermines the unique
structure–function characteristics of each of the six
major DMPK isoforms.
The coiled-coil domain mediates DMPK
homo- and heteromultimerization
Sizing experiments with gel filtration chromatography
revealed that full-length DMPK isoforms reside in
high-molecular-weight multimeric complexes. Pure
DMPK dimers may exist, but form only a minor frac-
tion of total protein. Given the elution profile, and
ABC
DE F
GH I
Fig. 8. Myotonic dystrophy protein kinase (DMPK) C association with mitochondria is stabilized by the coiled-coil domain. (A–C) YFP–DMPK

C colocalized with mitochondria stained with a cytochrome c oxidase antibody. (D–F) The YFP–DMPK C coil mutant (YFP–DMPK C
CM
)was
not only located at mitochondria but was also found dispersed throughout the cytosol. (G) The cytosolic distribution of YFP–DMPK E
CM
is
not altered by introduction of the mutations L477G, L487G and L515G (see YFP–DMPK E in the insert). (H–I) Expression of YFP–DMPK A
and A
CM
in N2A cells showed identical endoplasmic reticulum (ER) localizations for both proteins. Bars, 10 lm.
R. E. M. A. van Herpen et al. DMPK coiled-coils mediate multimerization
FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS 1131
based on an average complex size of 0.5 MDa and a
molecular mass for individual isoforms of 60–70 kDa,
we estimate that the DMPK complex must contain
around six monomers. Fewer monomers may be pos-
sible if a considerable fraction of the complex is made
up of different DMPK interacting proteins (i.e. other
than DMPK itself).
Conservation of multimerization capacity in trunca-
tion mutant DMPK E(402–537), and impaired
self-association of DMPK E
CM
, provide strong experi-
mental evidence that the coiled-coil sequence is
uniquely responsible for complex formation. More-
over, our data rule out dominant involvement of the
N-terminal region, kinase domain and amino acids
402–468 flanking the coiled-coil domain, including the
alternatively spliced VSGGG motif or membrane

anchors in the C-tail region.
The tendency to multimerize via coiled–coil associ-
ation, but the apparent lack of effect of other protein
motifs present in the protein, distinguishes DMPK
from some of its closest relatives. The N-terminal
region of MRCKa mediates dimerization of the kinase
domain independently of the coiled-coil domain [14].
In the case of ROCK-II, removal of the large coiled-
coil domain still results in the presence of a dimeric
protein. Here, dimerization may be driven by a small
coiled-coil region in the N-terminal region [30].
Although homology exists among the N termini of
DMPK, ROCK-I ⁄ -II and MRCKa ⁄ -b ⁄ -c (i.e. a leucine
zipper-like motif is found in all), we consider it unli-
kely that the N terminus has a strong role in DMPK
self-association, as mutant DMPK E(1–375) did not
multimerize under the experimental conditions used. A
supporting role for the N terminus cannot be com-
pletely ruled out, however.
To provide evidence that it is the typical 3D coiled-
coil organization and not the linear peptide sequence
of the segment that is important for DMPK complex
formation, we mutated three hydrophobic residues at
the putative heptad positions a and d, known to stabil-
ize the hydrophobic interface between helices forming
the coiled coil [31,32]. These mutations strongly influ-
enced the in-gel-migration behavior of DMPK, and
the single mutation L477G had already resulted in a
remarkable migration shift, indicative of structural
alterations introduced within the coiled-coil domain.

Most likely, the introduced glycine residues break up
the helical coiled-coil conformation [33], whereas the
hydrophobicity changes alter the folding behavior
within the coil [28]. From our experiments, it became
clear that mutated positions a and d strongly reduced
the self-association of DMPK E, providing evidence
that it is the coiled-coil structure proper that deter-
mines the self-association tendency. Although residual
self-association of DMPK E
CM
could be demonstrated
through covalent cross-linking, we conclude that this
mutant provides a proficient tool to study multimeriza-
tion-related functions of DMPK.
Multimerization modulates DMPK kinase activity,
substrate binding and localization
We observed that DMPK E
CM
autophosphorylation
and the transphosphorylation activity towards a pep-
tide substrate was two to threefold reduced when com-
pared with the corresponding activities of wild-type
DMPK E. To us this suggests that DMPK auto-
phosphorylation is largely an intermolecular reaction
in a homo- or heteromultimeric complex of DMPK
isoforms. We cannot rule out, however, the alternative
possibility that distortion of the coiled-coil structure
affects conformational flexibility in the kinase domain
itself and that this feature is needed for efficient intra-
molecular autophosphorylation. More detailed under-

standing of the DMPK structure is needed to be able
to distinguish between these possibilities. In DMPK-
like kinases, multimerization capacity is apparently a
prerequisite for proper kinase activation: the kinase
activity of Rho-kinase and MRCKa is also partly
dependent on the presence of a coiled-coil domain
[14,18,20] and others have shown that multimerization
of the human DMPK A isoform is correlated with
increased activity [16,17].
In contrast to the findings discussed above, we
found that DMPK E
CM
was able to phosphorylate its
natural substrate, MYPT2. Our inability to detect
effects of the coil mutation on MYPT2 phosphoryla-
tion may be a result of the experimental conditions
used. In our assay system, only a limited amount of
 0.2 lm MYPT2 was present, in contrast to the
excess of 30 lm peptide used in the peptide kinase
assay [5]. However, because we were not able to stabil-
ize the DMPK E
CM
MYPT2 interaction with a chem-
ical cross-linker (Fig. 7A), and the amount of MYPT2
bound to DMPK E
CM
was clearly lower than with
wild-type DMPK E, we assume that the DMPK E
CM
MYPT2 binding is short-lived. The DMPK coiled-coil

domain may be important in strengthening the binding
between DMPK and MYPT2 or in bringing protein
sequences, crucial for cross-link formation, in close
proximity. Whether it is the coiled-coil region in
MYPT2 that plays a role in the DMPK MYPT inter-
action will be investigated in future studies [5] (D. G.
Wansink et al. unpublished results).
ER or MOM targeting of full-length mouse DMPK
A and C, respectively, critically depends on the final 45
DMPK coiled-coils mediate multimerization R. E. M. A. van Herpen et al.
1132 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS
C-terminal amino acids [24]. Here, we show that muta-
tional disruption of the structural integrity of the
coiled-coil segment relocalizes DMPK C – but not
DMPK A – to the cytosol. One explanation for this
observation would be that loss of multimerization cau-
ses increased exposure of YFP–DMPK C to the proteo-
lytic machinery. In turn, clipping could be associated
with increased release of the YFP reporter moiety into
the cytosol. Western blotting and analysis of YFP–
DMPK products in our transfected cell lines, however,
ruled out this possibility (data not shown). Evidently,
structural intactness of the coiled-coil domain stabilizes
DMPK C association to the MOM, but is less critical
for ER association of DMPK A. It has been reported
that formation of a coiled-coil structure utilizes a two-
step model where protein folding and multimerization
are coupled events [28]. If this model holds, and given
the fact that the membrane anchors are situated at the
very C-terminal end of newly produced DMPK poly-

peptide chains, this would make it likely that DMPK
anchoring into ER or mitochondrial membranes occurs
when isoforms are already in a multimeric state. Thus,
multimerization per se may have co-operative effects
and promote MOM binding, but, based on the evidence
provided, we cannot exclude the possibility that the
coiled-coil mutations perturb proper DMPK C-tail con-
formation, thereby reducing the membrane-anchoring
affinity of individual polypeptide chains. Currently, not
enough is known about the molecular events involved
in targeting tail-anchored proteins to mitochondria [34]
to discriminate between these possibilities. It is evident
that the DMPK coiled-coil domain itself has no target-
ing properties [24]; however, it remains possible that,
once targeted to mitochondria, the coiled-coil domain
has some affinity for the mitochondrial membrane, as
recently demonstrated for the coiled-coil domain in
mitochondrial targeting of DLP1 ⁄ Drp1 [35].
Taken together, our results provide evidence that
the coiled-coil domain is crucial for homo- and hetero-
multimerization of DMPK isoforms and that multime-
rization has a function in substrate binding,
phosphorylation and subcellular targeting properties of
individual DMPK isoforms. Whether DMPK complex
formation is actively regulated in vivo (e.g. to modulate
DMPK activity and downstream effects), remains to
be investigated.
Experimental procedures
Cell culture and transfection
Neuro-2A (N2A) and COS-1 cells were cultured and trans-

fected as described previously [24].
Expression plasmids and site-directed
mutagenesis
Expression vectors for HA-, His- and EYFP-tagged DMPK
A–F have been described previously [5,24]. Expression plas-
mids, encoding VSV-tagged DMPK truncation constructs,
were obtained by cloning PCR fragments amplified from
template pSGmDMPK E [15] with the use of Pfu poly-
merase and specific primers. EcoRI and XhoI sites were
incorporated in the forward and reverse primers, respect-
ively (underlined, see below). DNA fragments were cut with
EcoRI and XhoI, gel purified and ligated into EcoRI and
XhoI polylinker sites of plasmid pSG8VSV. The sequence
of all PCR fragments was verified by DNA sequencing.
The following primers were used: pSGVSVDMPK
E(1–375): 5¢-ATA
GAATTCATGTCAGCCGAAGTGCG-
3¢ and 5¢-ATT
CTCGAGTCAAGTGAGCCGGTCCTCCA-
3¢; pSGVSVDMPK E(1–400): 5¢-ATA
GAATTCATGTCA
GCCGAAGTGCG-3¢ and 5¢-AAT
CTCGAGTCAGAAGG
GCAGGCGCAC-3¢; pSGVSVDMPK E(60–375): 5¢-ATA
GAATTCAGGCTTAAGGAGGTCCGA-3¢ and 5¢-ATT
CTCGAGTCAAGTGAGCCGGTCCTCCA-3¢; pSGVSVD
MPK E(340–537): 5¢-ATT
GAATTCTTTGGCCTTGATTG
GGA-3¢ and 5¢-ATA
CTCGAGCTAGGGATCTGCGGCT-

3¢; pSGVSVDMPK E(402–537): 5¢-ATA
GAATTCGGCTA
CTCCTACTGCTGCAT-3¢ and 5¢-ATA
CTCGAGCTAGG
GATCTGCGGCT-3¢.
To generate the HA-DMPK E coil mutant expression
vector, pSGHADMPK E
CM
, three amino acid mutations
were introduced into full-length pSGHADMPK E with use
of the QuikChange Site-Directed Mutagenesis Kit (Strata-
gene, La Jolla, CA, USA), according to the manufacturer’s
protocol, in successive mutagenesis steps. The following
primers were used: for L477G, primers 5¢-CAGCTCCAGG
AAGCCGGGGAAGAAGAGGTTC-3¢ and 5¢-GAACCT
CTTCTTCCCCGGCTTCCTGGAGCTG-3¢; for L487G,
primers 5¢-TCACCCGGCAGAGCGGGAGCCGCGAGC
TGGAG-3¢ and 5¢-CTCCAGCTCGCGGCTCCCGCTCTG
CCGGGTGA-3¢; for L515G, primers 5¢-GTCCGAAACC
GAGACGGGGAGGCGCATGTTC-3¢ and 5¢-GAACATG
CGCCTCCCCGTCTCGGTTTCGGAC-3¢.
To generate expression plasmids pEYFP-DMPK A
CM
,
C
CM
and E
CM
, the following cloning steps were carried out.
First, an L515G mutation was introduced into

pSGHADMPK A and C, as described above, resulting in
pSGHADMPK A(L515G) and C(L515G). Then, two frag-
ments [an AflII–BspEI fragment – common to all DMPK
isoforms and including mutations L477G and L487G – iso-
lated from pSGHADMPK E
CM
, and a BspEI–BsrGI frag-
ment – specific for each individual DMPK isoform, and
including mutation L515G isolated from pSGHADMPK
A(L515G), C(L515G) and E
CM
] were ligated into a
pSGmDMPK E vector digested with AflII and BsrGI,
resulting in pSGDMPK A
CM
,C
CM
and E
CM
. Finally,
BglII-flanked cDNAs from pSGHADMPK A
CM
,C
CM
and
R. E. M. A. van Herpen et al. DMPK coiled-coils mediate multimerization
FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS 1133
E
CM
were ligated into the BglII site of plasmid pEYFP-C1

(Clontech, Mountain View, CA, USA).
MYPT2 was cloned by RT-PCR using mouse skeletal
muscle RNA and primers 5¢-ATAGAATTC
ATGGCGGA
GCTGGAGCA-3¢ and 5¢-ATACTCGAG
CTACTTGGAC
AGTTTGCTGATGACT-3¢ (start and stop codons under-
lined) in a pSG8-based vector in-frame with a His and a
VSV tag sequence.
Gel filtration chromatography
Expression plasmids pSGDMPK A, C, E, F or
pSGVSVDMPK E(402–537) were transfected into COS-1
cells grown on 10 cm dishes and cultured for an additional
24 h prior to analysis. Then, cells were collected and the
cleared cell extracts were prepared by centrifugation after
lysis of cells in RIPA buffer [50 mm Hepes, pH 7.5,
150 mm NaCl, 1 mm EDTA, 25 mm NaF, 1% (w ⁄ v) Triton
X-100, 1% (w ⁄ v) sodium desoxycholate, 0.1% (w ⁄ v) SDS,
1mm phenylmethanesulfonyl fluoride) on ice. Gel filtration
was performed on a SMART system Superose 6 HR10 ⁄ 30
column (GE Healthcare, Roosendaal, the Netherlands)
equilibrated with RIPA buffer and calibrated with molecu-
lar mass markers ferritin (440 kDa), alcohol dehydrogenase
(150 kDa), BSA (67 kDa) and ribonuclease A (14 kDa).
Cell extracts (10 lL) were applied to the column at a flow
rate of 40 lLÆmin
)1
. Fractions of 40 lL were collected and
the protein content in each fraction was subjected to
SDS ⁄ PAGE analysis.

Confocal microscopy
N2A cells were grown on glass coverslips and transfected
with expression plasmids pEYFP-DMPK A, C, E, A
CM
,
C
CM
or E
CM
. After 24 h, cells were fixed in NaCl ⁄ P
i
(PBS),
containing 2% (w ⁄ v) formaldehyde, and either mounted
directly or processed for immunofluorescence using stand-
ard procedures [5]. A rabbit anti-(cytochrome c) oxidase
immunoglobulin was used to visualize mitochondria. Con-
focal images were generated on a Bio-Rad MRC1024 con-
focal laser-scanning microscope (Bio-Rad, Hercules, CA,
USA) equipped with an argon ⁄ krypton laser, using a
60 · 1.4 NA oil objective and lasersharp2000 acquisition
software. Images were further processed with adobe photo-
shop 7.0 (Adobe, San Jose, CA, USA).
Immunoprecipitation and in vitro kinase assay
COS-1 cells were transfected with DMPK and ⁄ or MYPT2
expression plasmids, or with empty vector DNA
(pSG8DEco), cultured for  24 h, washed with ice-cold
PBS and lysed on ice in RIPA buffer or kinase assay
lysis buffer [50 mm Tris ⁄ HCl, pH 7.5, 150 mm NaCl, 1%
Nonidet P-40, 25 mm NaF, 1 mm sodium pyrophosphate,
0.1 mm Na

3
VO
4
,2lm Microcystin LR (ALEXIS, Lausen,
Switzerland), 1 mm phenylmethanesulfonyl fluoride, supple-
mented with protease inhibitor cocktail (Roche, Mannheim,
Germany)]. The reversible chemical cross-linker, DSP
(0.5 mgÆmL
)1
), was added to the lysis buffer in DMPK–
MYPT2 co-expression studies only. Co-immunoprecipita-
tions were performed with an anti-HA (12CA5) or an anti-
VSV monoclonal antibody coupled to protein A–Sepharose
beads. Immunoprecipitates were washed four times in
RIPA buffer, the final wash was removed and beads were
mixed with Laemmli sample buffer and used for
SDS ⁄ PAGE and western blot analysis on poly(vinylidene
difluoride) membrane, as detailed below.
Kinase assays were performed as described previously
[5]. Briefly, beads derived from the immunoprecipitation
were washed and aliquoted into the desired number of
kinase reactions. The final wash was removed and 30 lm
peptide substrate KKRNRRLTVA or 1 lg of immuno-
purified His-VSV-MYPT2 in kinase assay buffer was
added [5]. The reaction was started by the addition of
[
32
P]ATP[cP] ( 2 lCi) and the mixture was incubated
at 30 °C (60 min for the peptide, 20 min for MYPT2).
Phosphate incorporation into the peptide substrate under

the conditions used was linear, and was analyzed as
described previously [5].
SDS ⁄ PAGE and western blotting
Proteins were separated by SDS ⁄ PAGE, transfered to
poly(vinylidene difluoride) membrane (GE Healthcare),
then analyzed by immunodetection. For visualization of
DMPK, DMPK-specific antibody (B79) or anti-HA
immunoglobulin was used [5,15]; truncation mutants were
detected using an anti-VSV immunoglobulin. Horseradish
peroxidase-conjugated secondary antibodies (Jackson
ImmunoResearch Laboratories, Soham, UK) were used
followed by enhanced chemiluminescence (ECL) and
exposure to film (Kodak X-OMAT AR; Kodak, Paris,
France).
Bioinformatics
Computational algorithms used to predict heptad sequences
in DMPK included the program coil (v2.1, mtidk matrix
and window 21) [26], paircoil (probability cut-off 0.5) [27]
and multicoil (cut-off 0.5 and window 28) [25].
Acknowledgements
We would like to express our gratitude to Dr J.
Schalkwijk (Department of Dermatology, RUMC Nij-
megen) for advice on and help with gel-filtration
chromatography. This study was supported by the
DMPK coiled-coils mediate multimerization R. E. M. A. van Herpen et al.
1134 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS
Prinses Beatrix Fonds, the Stichting Spieren voor Spi-
eren, the American Muscular Dystrophy Association
and the Association Franc¸ aise contre les Myopathies.
References

1 Leung T, Chen X-Q, Tan I, Manser E & Lim L
(1998) Myotonic dystrophy kinase-related Cdc42-bind-
ing kinase acts as a Cdc42 effector in promoting
cytoskeletal reorganization. Mol Cell Biol 18, 130–140.
2 Ng Y, Tan I, Lim L & Leung T (2004) Expression of
the human myotonic dystrophy kinase-related Cdc42-
binding kinase {gamma} is regulated by promoter DNA
methylation and Sp1 binding. J Biol Chem 279, 34156–
34164.
3 Riento K & Ridley AJ (2003) Rocks: multifunctional
kinases in cell behaviour. Nat Rev Mol Cell Biol 4, 446–
456.
4 Madaule P, Eda M, Watanabe N, Fujisawa K, Mats-
uoka T, Bito H, Ishizaki T & Narumiya S (1998) Role
of citron kinase as a target of the small GTPase Rho in
cytokinesis. Nature 394, 491–494.
5 Wansink DG, van Herpen REMA, Coerwinkel-Dries-
sen MM, Groenen PJTA, Hemmings BA & Wieringa
B (2003) Alternative splicing controls myotonic dystro-
phy protein kinase structure, enzymatic activity and
subcellular localization. Mol Cell Biol 23, 5489–5501.
6 Manning G, Plowman GD, Hunter T & Sudarsanam S
(2002) Evolution of protein kinase signaling from yeast
to man. Trends Biochem Sci 27, 514–520.
7 Caenepeel S, Charydczak G, Sudarsanam S, Hunter T
& Manning G (2004) The mouse kinome: discovery and
comparative genomics of all mouse protein kinases.
Proc Natl Acad Sci USA 101, 11707–11712.
8 Bishop AL & Hall A (2000) Rho GTPases and their
effector proteins. Biochem J 348 Part 2, 241–255.

9 Amano M, Fukata Y & Kaibuchi K (2000) Regulation
and functions of Rho-associated kinase. Exp Cell Res
261, 44–51.
10 Leung T, Chen X-Q, Manser E & Lim L (1996) The
p160 RhoA-binding kinase ROKa is a member of a kin-
ase family and is involved in the reorganization of the
cytoskeleton. Mol Cell Biol 16, 5313–5327.
11 Yamashiro S, Totsukawa G, Yamakita Y, Sasaki Y,
Madaule P, Ishizaki T, Narumiya S & Matsumura F
(2003) Citron Kinase, a Rho-dependent kinase, induces
di-phosphorylation of regulatory light chain of myosin
II. Mol Biol Cell 14, 1745–1756.
12 Muranyi A, Zhang R, Liu F, Hirano K, Ito M, Epstein
HF & Hartshorne DJ (2001) Myotonic dystrophy pro-
tein kinase phosphorylates the myosin phosphatase tar-
geting subunit and inhibits myosin phosphatase activity.
FEBS Lett 493, 80–84.
13 Groenen P & Wieringa B (1998) Expanding complexity
in myotonic dystrophy. Bioessays 20, 901–912.
14 Tan I, Seow KT, Lim L & Leung T (2001) Intermole-
cular and intramolecular interactions regulate catalytic
activity of myotonic dystrophy kinase-related Cdc42-
binding kinase a. Mol Cell Biol 21, 2767–2778.
15 Groenen PJTA, Wansink DG, Coerwinkel M, van den
Broek W, Jansen G & Wieringa B (2000) Constitutive
and regulated modes of splicing produce six major myo-
tonic dystrophy protein kinase (DMPK) isoforms with
distinct properties. Hum Mol Genet 9, 605–616.
16 Bush EW, Helmke SM, Birnbaum RA & Perryman MB
(2000) Myotonic dystrophy protein kinase domains med-

iate localization, oligomerization, novel catalytic activity,
and autoinhibition. Biochemistry 39, 8480–8490.
17 Zhang R & Epstein HF (2003) Homodimerization
through coiled-coil regions enhances activity of the
myotonic dystrophy protein kinase. FEBS Lett 546,
281–287.
18 Shimizu T, Ihara K, Maesaki R, Amano M, Kaibuchi
K & Hakoshima T (2003) Parallel coiled–coil associ-
ation of the RhoA-binding domain in Rho-kinase.
J Biol Chem 278, 46046–46051.
19 Dvorsky R, Blumenstein L, Vetter IR & Ahmadian MR
(2004) Structural insights into the interaction of ROCKI
with the switch regions of RhoA. J Biol Chem 279,
7098–7104.
20 Amano M, Chihara K, Nakamura N, Kaneko T,
Matsuura Y & Kaibuchi K (1999) The COOH terminus
of Rho-kinase negatively regulates Rho-kinase activity.
J Biol Chem 274, 32418–32424.
21 Chen X-Q, Tan I, Ng CH, Hall C, Lim L & Leung T
(2002) Characterization of RhoA-binding kinase
ROKalpha implication of the pleckstrin homology
domain in ROKalpha function using region-specific
antibodies. J Biol Chem 277, 12680–12688.
22 Suzuki A, Sugiyama Y, Hayashi Y, Nyu-i N, Yoshida
M, Nonaka I, Ishiura S, Arahata K & Ohno S (1998)
MKBP, a novel member of the small heat shock protein
family, binds and activates the myotonic dystrophy pro-
tein kinase. J Cell Biol 140, 1113–1124.
23 Shimizu M, Wang W, Walch ET, Dunne PW & Epstein
HF (2000) Rac-1 and Raf-1 kinases, components of dis-

tinct signaling pathways, activate myotonic dystrophy
protein kinase. FEBS Lett 475, 273–277.
24 van Herpen RE, Oude Ophuis RJ, Wijers M, Bennink
MB, van de Loo FA, Fransen J, Wieringa B & Wansink
DG (2005) Divergent mitochondrial and endoplasmic
reticulum association of DMPK splice isoforms depends
on unique sequence arrangements in tail anchors. Mol
Cell Biol 25, 1402–1414.
25 Wolf E, Kim PS & Berger B (1997) MultiCoil: a pro-
gram for predicting two- and three-stranded coiled coils.
Protein Sci 6, 1179–1189.
26 Lupas A, Van Dyke M & Stock J (1991) Predicting
coiled coils from protein sequences. Science 252, 1162–
1164.
R. E. M. A. van Herpen et al. DMPK coiled-coils mediate multimerization
FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS 1135
27 Berger B, Wilson DB, Wolf E, Tonchev T, Milla M &
Kim PS (1995) Predicting coiled coils by use of pairwise
residue correlations. Proc Natl Acad Sci USA 92, 8259–
8263.
28 Mason JM & Arndt KM (2004) Coiled coil domains:
stability, specificity, and biological implications. Chem-
biochem 5, 170–176.
29 Ito M, Nakano T, Erdodi F & Hartshorne DJ (2004)
Myosin phosphatase: structure, regulation and function.
Mol Cell Biochem 259, 197–209.
30 Doran JD, Liu X, Taslimi P, Saadat A & Fox T (2004)
New insights into the structure-function relationships
of Rho-associated kinase: a thermodynamic and hydro-
dynamic study of the dimer-to-monomer transition and

its kinetic implications. Biochem J 384, 255–262.
31 Wagschal K, Tripet B, Lavigne P, Mant C & Hodges
RS (1999) The role of position a in determining the
stability and oligomerization state of alpha-helical coiled
coils: 20 amino acid stability coefficients in the hydro-
phobic core of proteins. Protein Sci 8, 2312–2329.
32 Tripet B, Wagschal K, Lavigne P, Mant CT & Hodges
RS (2000) Effects of side-chain characteristics on stabi-
lity and oligomerization state of a de novo-designed
model coiled-coil: 20 amino acid substitutions in posi-
tion ‘d’. J Mol Biol 300, 377–402.
33 O’Neil KT & DeGrado WF (1990) A thermodynamic
scale for the helix-forming tendencies of the commonly
occurring amino acids. Science 250, 646–651.
34 Borgese N, Colombo S & Pedrazzini E (2003) The tale
of tail-anchored proteins: coming from the cytosol and
looking for a membrane. J Cell Biol 161 , 1013–1019.
35 Pitts KR, McNiven MA & Yoon Y (2004) Mitochon-
dria-specific function of the dynamin family of protein
DLP1 is mediated by its C-terminal domains. J Biol
Chem 279, 50286–50294.
DMPK coiled-coils mediate multimerization R. E. M. A. van Herpen et al.
1136 FEBS Journal 273 (2006) 1124–1136 ª 2006 The Authors Journal compilation ª 2006 FEBS