Dystrobrevin requires a dystrophin-binding domain to function
in
Caenorhabditis elegans
Karine Grisoni, Kathrin Gieseler and Laurent Se
´
galat
CGMC, CNRS-UMR, Universite
´
Lyon, Villeurbanne, France
Dystrobrevin is one of the intracellular components of the
transmembrane dystrophin–glycoprotein complex (DGC).
The functional role of th is complex in normal an d patho-
logical situations has not yet been clearly established.
Dystrobrevin disappears from t he muscle m embrane in
Duchenne muscular dystrophy (DMD), which results from
dystrophin mutations, as well as in limb girdle muscular
dystrophies (LGMD), which results f rom mutations affect-
ing other members of the DGC complex. These findings
therefore s uggest that dystrobrevin may play a pivotal role in
the progression of these clinically related diseases. In this
study, w e u sed t he Ca enorhabditis elegans mod el t o a ddress
the question of the relationship between dystrobrevin
binding to dystrophin and dystrobrevin function. Deletions
of the dystrobrevin protein w ere performed and the ability
of the mutated forms to bind t o dystrophin was tested both
in vitro and in a two-hybrid assay, as well as their ability t o
rescue dystrobrevin (dyb-1) mutations in C. elegans. The
deletions affecting the second helix of the D yb-1 coiled-coil
domain abolished the binding of dystrobrevin to dystrophin
both in vitro and in the two-hybrid assay. These deletions
also abolished t he rescuing activity of a functional transgene

a-Dystrobrevin binds to dystrophin via a coiled-coiled
motif present in both proteins, and to the PDZ domain
containing syntrophins [11,12]. Indirect evidence suggests
that dystrobrevin may also bind to other members of the
DGC [13]. Although no enzymatic activity has yet been
assigned to dystrobrevins, there a re several i ndications that
they may play a role in signalling mechanisms. First,
dystrobrevins are tyrosine-phosphorylated proteins [5,14].
Secondly, in the absence of a-dystrobrevin, the signalling
molecule, neuronal nitric o xide synthase (nNO S) disappears
from the muscular membrane [10].
In addition, two lines of evidence suggest that dystro-
brevin may p lay a key role i n the muscle degeneration
observed in DMD and sarcoglycanopathies; first, dystro-
brevin immunostaining decreases greatly in DMD and in
several sarcoglycanopathies [15]. Secondly, although t he
DGC components (with the exception of NOS) are not
affected by the absence of dystrobrevin in adbn mice, musc le
degeneration occurs.
The nematode Caenorhabditis elegans has homologues of
most of the DGC proteins (L. Se
´
galat, unpublished results).
There is one dystrophin- and one dystrobrevin-like gene in
thegenomeofC. elegans (dys-1 and dyb-1, respectively)
[16,17]. C. elegans dystrophin and dystrobrevin are able t o
bind to each other in vitro [18] in the same way as their
mammalian counterparts [12], and they also bind to
syntrophin [18]. dys-1 and dyb-1 mutants d isplay a similar
behavioural phenotype c onsisting of hyperactivity, exagger-

EXPERIMENTAL PROCEDURES
Construction of deleted forms of Dyb-1
for
in vitro
binding experiments
Deletions were carried out on the dyb-1 coding sequence,
using clone AN450 [encoding Dyb-1 amino acids 3 90–543
fused in frame to the GST coding sequen ce; plasmid pGEX
3X (Pharmacia)] [18]. AN450 DNA (500 ng) was cut with
the restriction enzyme MfeI. The cut DNA was then
distributed among several tubes incubated with 0.05 lLof
BAL31 exonuclease for various times (typically 0–10 min).
The r eactions were stopped by adding EGTA to 4 m
M
and
heating a t 6 5 °C f or 10 min. DNA was purified on a Wizard
column (Promega) and the action of BAL31 was checke d by
loading an aliquot o f each tube onto a n agarose gel column.
The DNA corresponding to the deletions required was
treated b y applying T4 DNA polymerase in the presence of
nucleotides to create blunt ends, w hich were ligated and the
plasmids were transformed in Escherichia coli DH5. Clones
were picked randomly and analysed using sequencing
procedures. Any clones carrying a frame shift were rejected.
Construct 6¢4 was built using similar procedures, but using
the enzyme HindIII instead of MfeI. The amino acids
removed in the deletions were 489–499 (clone 2¢5), 487–513
(clone 5 ¢1), 489–528 ( clone 5 ¢2B), 471–517 (clone 5 ¢5B), 478–
543 ( clone 2 ¢1), and 391–450 (clone 6¢4). Clones 2¢1and6¢4
have been described previously [18], but clo ne 2¢1was

)1
, phenylmethanesulfo-
nyl fluoride (1 mM)]. The
35
S-labelled Dys-1 C-terminal
end was synthesized using a coupled in vitro transcription
and translation kit (Promega) with cDNA yk12c11 [16].
The p reparations were incub ated for 2 h at 4 °C. GST
controls were performed using 1–2 times the amount of
fusion protein. After five washes with binding buffer, the
labelled proteins were e luted by boiling t he preparation f or
3 min in gel loading buffer. Gels were dried, exposed
overnight and revealed using a radiographic analyser
(Fuji BAS-1500). Band intensity was quantitated using
the analyser software on at least three independent
experiments.
Constructs for the yeast two-hybrid assay
The C-terminal end of Dys -1 (amino a cids 2857–3674) was
fused t o t he DNA binding domain (DNA-BD) of t he Gal4
protein. For t his purpose, a 2,4 kb dys-1 cDNA fragment
(yk12c11) was cloned into the polylinker of pAS2-1
(Clontech) with respect to the reading frame.
Dyb-1 fragments and deletions were PCR a mplified using
clones AN 450, 2¢5, 5¢1, 5¢2B, 5 ¢5B and 6¢4inpGEX3Xas
templates (see b elow) and cloned into pACT2 (Clontech) in
frame w ith the activation domain ( AD) of the Gal4 protein
and the HA epitope. The resulting constructs were called
AD-AN450, AD-2¢5, AD-5¢1, AD-5¢2B, AD-5¢5B, and
AD-6¢4, respectively. All DNA constructs were checked by
performing DNA sequencing.

units) were
spun down and frozen at )70 °C for at least one hour.
The yeast pellet was resuspended in 60 lLofsample
buffer [21]. After boiling the mixture for 5 min, and
centrifuging f or 30 s a t 1 3 000 g,10lL of supernatant
was loaded onto each lane of a 0.1% SDS/10%
polyacrylamide gel. Proteins were transferred onto a
BA83 nitrocellulose membrane (Schleicher & Schuell) in
transfer buffer (Tris 25 m
M
, glycine 190 m
M
, SDS 0.01%,
ethanol 20%) for 1 h at 100 V. AD-Dyb-1 fusion proteins
were detected using a rabbit polyclonal anti-(Dyb-1) Ig
[19] at a dilution 1 : 500. Peroxidase-coupled anti-(rabbit
IgG) Ig (Biorad) was used at a dilution of 1 : 3000. Blots
1608 K. Grisoni et al. (Eur. J. Biochem. 269) Ó FEBS 2002
were revealed using the ECL+ kit (Amersham) as
recommended by the supplier.
Functional assay in
C. elegans
First, a dyb-1 functional construct was obtained by m odi-
fying a previously built dyb-1:gfp construct [19]. The
dyb-1:gfp construct, which has been previously described,
was shortened on the 5¢ end to leave 2.9 kb of upstream
sequence, and various restriction e nzyme s ites were removed
and a dded by performing synonymous point mutations to
yield the construct dyb-1:gfp VII, which h as single Age Iand
MluI sites at codons 390 and 543. This co nstruct encod es a

(clone 5¢5B). Clone 2¢1, lacking amino acids 478–543, was
used as a negative control [18]. Clone 6¢4, lacking amino
acids 391–450, was used as a second positive control [18].
The constructs were used to produce Dyb-1–GST chimeric
proteins in E. coli, which were affinity purified on gluthati-
one–Sepharose beads and subjected to in vitro binding with
35
S-labelled Dys-1. Clones 2¢5and5¢2B bound to Dys-1 at
levels that were not significantly different from those of the
positive controls AN450 (Fig. 2) and 6¢4 (gel not shown). In
contrast, the binding activity of clones 5¢1and5¢5B was
weaker (Fig. 2). The difference between clones 2¢5, 5¢1and
Fig. 1. Deletions used in this s tudy. The ‘WT’ line represents the amino-
acid sequence of the wild-type Dyb-1 protein in the predicted coiled-
coil domain region. The predicted helices forming the domain are
shown by hatched boxes. Numbers above the wild-type sequence
indicate the amino-acid coordinates of t he helices. Deletions are s hown
below the wild-type se quenc e. Numbers i ndicate the coordinates o f the
breakpoints. D eletions we re ge ne rated b y e xonuclease digestion . No te
that the 6¢4 de letion extends on the left side further tha n sh own on the
drawing. For in vitro binding experiments, the corresponding DNAs
were cloned into the pGEX vector to produce Dyb-1–GST fusion
proteins [18]. T he righ t column g ives t he binding affinity of the v arious
constructs to
35
S-labelled Dys-1 in arbitrary units (mean ± SD). One
unit is defined as t he autoradiogram intensity obtained with the neg-
ative control GST. Asterisks indicate values significantly different from
wild-type. Constructs 5¢1, 5¢5B and 2¢1 have significantly reduced
affinity to Dys-1.

and t he 6 ¢4 fragment resulted i n t he gr owth of yeasts on His -
plates, which can occur only if Dyb-1 binds to Dys-1
(Fig. 4). Similar r esults were obtained when a shorter Dys-1
fragment (amino acids 3402–3674) encompassing the syn-
trophin-binding domain and the coiled-coil domain (cor-
responding to BB810 in [18]) was used (not shown). These
results indicate that all the deletions affecting the second
helix of Dyb-1, including the shortest deletion (clone 2¢5),
greatly reduce t he interactions between Dys-1 and D yb-1 in
the yeast system.
Functional complementation of Dyb-1 deletions
in
C. elegans
The only functional assay available for dystrobrevin resides
in functional complementation. To test whether the dele-
tions of various parts of the coiled-coil domain had an effect
on the in vivo function of Dyb-1, we created transgenes
carrying the s ame deletions as those tested in v itro andinthe
yeast system. We transferred the deletions into the vector
dyb-1:gfp V II, w hich is a functional transgene consisting of
genomic Dyb-1 sequences (Fig. 5). Because the deletions
are derivatives of clone AN450, a cDNA fragment that
encompasses several exons, it was necessary first to check
whether removing introns 7 and 8 had any effect on the
rescuing activity of dyb-1:gfp VII. When the 1.2-kb AgeI–
MluI genomic fragment of dyb-1:gfp VII was replaced by
the 450 -bp A N450 cDNA fragment encoding the same
amino acids (Fig. 5), rescue of dyb-1(cx36) animals still
occurred in t wo out of thre e lines transgenic lines (Table 1),
which i ndicates that r emoving i ntrons 7 and 8 did not impair

Fig. 4. Yeast two-hybrid assay. A plate containing SD media minus
leucin, t ryp tophan and histidine was seeded with yeast c arrying DNA-
BD-Dys-1 and various AD-Dyb-1 p lasmids or empty pACT2 (as a
negative control), an d incubated at 30 °C for 3 days. Gro wth on m edia
devoid of histidine can occur only if D ys-1 and Dyb-1 interact and the
HIS3 reporter gene is transactivated. Only t he wild-type construct A D-
AN450 and the AD-6¢4 construct promoted growth in this assay. The
other constructs, all carrying deletions in the second h e lix of the Dyb-1
coiled-coil domain (H2), were unable t o p romote gr owth in this assay,
which indicates that the H2 helix is a critical prerequisite for Dys-1/
Dyb-1 interactions to b e possible.
1610 K. Grisoni et al. (Eur. J. Biochem. 269) Ó FEBS 2002
between th e binding of dystrobrevin to dystrophin and
functional activity of dystrobrevin. The C. elegans mode l
organism was particularly well suited for the latter part
because transgenic a nimals can be quickly obtained in this
species. As long as the catalytic, enzymatic, or other
functional activity of the dystrobrevin protein will remain
unknown, the only way of making functional investi-
gations will continue to be through complementation of
mutations.
A preliminary in vitro study on Dys-1–Dyb-1 interactions
pointed to the second helix (H2) of the predicted coiled-coil
domain of Dyb-1 [18]. Here, we refined this analysis by
studying additional d eletions that s ubdivide the H2 domain.
Our results confirm the previously published data and show
that H2 is involved in the interaction with Dys-1 in vitro.
Within this domain, the first half seems to be particularly
critical as deletions infringing on th is side lead to a d ecrease
in binding. Constructs 2¢5, 5¢2B and 5¢1 are of particular

in dark. T he deletions were transfered into
dyb-1:gfp VII by PCR using the unique Age I
and MluI restriction sites o f dyb-1:gfp VII. As
a result, introns 7 an d 8 of dyb-1:gfp VII we re
removed in t he se constructs. These construct
were injected in dyb-1(cx36) mutants to assay
their ability to re sc ue the mutant phenotype.
Table 1. Rescuing activity o f Dyb-1 deletions. Constructs were injec ted in dyb-1(cx36) animals a long with the t ransformation marker KP13 [22].
dyb-1(cx36) animals display a behavioral phenotype consisting of hyperactivity, exaggerated be nding of the head w h en moving forward , and a
tendency to hyp ercontract when m oving backwards. + + +, t ransgenic animals n ot distinguishable from wild-type animals; + +, transgenic
animals resemble w ild-type but remain slightly hyperactive and bend their h ead more than wild-type; +, transgenic animals remain hyperactive and
bend their h ead, but can be distingu ished from nontransgenic siblings in blind tests; ±, some t ransgenic animals show a slightly improved behavior
but transgenics cannot be reco gnized with certainty in b lind tests; –, no m odification of the phenotype c ou ld be observed.
Construct
Number of
transgenic lines
Number of
rescuing lines
Rescue
in best line(s)
dyb-1:gfp VII 3 2 + + +
AN450 in dyb-1:gfp VII 3 2 + +
2¢5 in dyb-1:gfp VII 5 1 ±
5¢1 in dyb-1:gfp VII 6 0 –
5¢2B in dyb-1:gfp VII 4 0 –
5¢5B in dyb-1:gfp VII 8 0 –
6¢4 in dyb-1:gfp VII 4 3 +
Ó FEBS 2002 Dystrophin–dystrobrevin interactions in C. elegans (Eur. J. Biochem. 269) 1611
5¢2B) had no or very limited functional a ctivity in t he worm.
Howev er, the 6¢4 construct, which removes the syntroph in-

4. Carr, C., Fischbach, G.D. & Cohen, J.B. (1989) A novel 87,000-
M
r
protein associated with acetylcholine receptors in Torpedo
electric organ and vertebrate skeletal muscle. J. Cell Biol. 109,
1753–1764.
5. Wagner, K.R., Cohen, J.B. & Huganir, R.L. (1993) The
87K postsynaptic membrane pro tein from torpedo is a protein-
tyrosine kinase substrate homologous to dystrophin. Neuron 10,
511–522.
6. Sadoulet-Puccio, H.M., Khurana, T.S., Cohen, J.B. & Kunkel,
L.M. (1996) Cloning and characterization of the human
homologue o f a dystrophin r elat ed phosphoprotein found at the
Torpedo electric o rgan post-synaptic membrane. Hum. Mol.
Genet. 5, 489–496.
7. Blake, D., Nawrotzki, R., Peters, M., Froehner, S. & Davies, K.
(1996) Isoform diversity of dystrobrevin, the murine 87-kDa
postsynaptic protein. J. Biol . Chem. 271, 7 802–7810.
8. Peters, M., Adams, M. & Froehner, S. (1997) Differential
association of syntrophin pairs with the dystrophin complex.
J. Cell Bio l. 138, 81–93.
9. Blake, D., Nawrotzki, R., Loh, N., Gorecki, D. & Davies, K.
(1998) b-Dystrobrevin, a member of the dystrophin-related
protein family. Proc. Natl A cad. Sci. USA 95, 241–246.
10. Grady, R.M., G range, R.W., Lau, K.S., Maimone, M.M., Nichol,
M.C., Stull, J .T. & Sanes, J.R. ( 1999) R ole f or a lpha-dystrobrevin
in the pathogenesis of dystrophin-dependen t muscular dystro-
phies. Nat. Cell Biol. 1, 215–220.
11. Peters, M., Sadoulet -Puccio, H., Grady, R., Kram arcy, N .,
Kunkel,L.,Sanes,J.,Sealock,R.&Froehner,S.(1998)

galat, L. (1999) In vitro
interacti ons of Caenorhabditis elegans dystrophin with
dystrobrevin and syntrophin. FEBS Lett. 461 , 59–62.
19. Gieseler,K.,Mariol,M.,Bessou,C.,Migaud,M.,Franks,C.,
Holden-Dye,L.&Se
´
galat, L. (2001) Molecular, genetic, and
physiological characterisation of dystrobrevin-like (dyb-1)
mutants of Caenorhabditis elegans. J. Mol. Biol. 307, 1 07–117.
20. Gieseler, K., Grisoni, K . & Se
´
galat, L. (2000 ) Genetic suppression
of phenotypes arisin g from m utations in d ystrophin -related genes
in Caenorhabditis elegans. Curr. Biol. 10 , 1092–1097.
21. Harlow,E.&Lane,D.(1988)Antibodies, a Laboratory Manual.
Cold Spring Harbor Laboratory Press, Cold Spri ng Har bor, N ew
York.
22. Se
´
galat, L., Elkes, D.A. & Kaplan, J.M. (1995) Modulation of
serotonin-controlled behaviors by Go in Caenorhabditis elegans.
Science 267, 1648–1651.
23. Mello, C. & Fire, A. (1995) DNA transformation. Methods C ell
Biol. 48, 451–482.
1612 K. Grisoni et al. (Eur. J. Biochem. 269) Ó FEBS 2002