Muscular dystrophy associated mutations in caveolin-1 induce neurotransmission and locomotion defects in Caenorhabditis elegans.

Abstract

Mutations in human caveolin-3 are known to underlie a range of myopathies. The cav-1 gene of Caenorhabditis elegans is a homologue of human caveolin-3 and is expressed in both neurons and body wall muscles. Within the body wall muscle CAV-1 localises adjacent to neurons, most likely at the neuromuscular junction (NMJ). Using fluorescently tagged CAV-1 and pre- and post-synaptic markers we demonstrate that CAV-1 co-localises with UNC-63, a post-synaptic marker, but not with several pre-synaptic markers. To establish a model for human muscular dystrophies caused by dominant-negative mutations in caveolin-3 we created transgenic animals carrying versions of cav-1 with homologous mutations. These animals had increased sensitivity to levamisole, suggesting a role for cav-1 at the NMJ. Animals carrying a deletion in cav-1 show a similar sensitivity. Sensitivity to levamisole and locomotion were also perturbed in animals carrying a dominant-negative cav-1 and a mutation in dynamin, which is a protein known to interact with caveolins. Thus, indicating an interaction between CAV-1 and dynamin at the NMJ and/or in neurons.

Full text

Page 1

ORIGINAL PAPER
Muscular dystrophy associated mutations in caveolin-1 induce
neurotransmission and locomotion defects in Caenorhabditis
elegans
Scott Parker Æ Æ Helen S. Peterkin Æ Æ Howard A. Baylis
Received: 17 April 2007/Accepted: 11 June 2007/Published online: 13 July 2007
? Springer-Verlag 2007
Abstract
underlie a range of myopathies. The cav-1 gene of
Caenorhabditis elegans is a homologue of human caveolin-
3 and is expressed in both neurons and body wall muscles.
Within the body wall muscle CAV-1 localises adjacent to
neurons, most likely at the neuromuscular junction (NMJ).
Using fluorescently tagged CAV-1 and pre- and post-syn-
aptic markers we demonstrate that CAV-1 co-localises
with UNC-63, a post-synaptic marker, but not with several
pre-synaptic markers. To establish a model for human
muscular dystrophies caused by dominant-negative muta-
tions in caveolin-3 we created transgenic animals carrying
versions of cav-1 with homologous mutations. These ani-
mals had increased sensitivity to levamisole, suggesting a
role for cav-1 at the NMJ. Animals carrying a deletion in
cav-1 show a similar sensitivity. Sensitivity to levamisole
and locomotion were also perturbed in animals carrying a
dominant-negative cav-1 and a mutation in dynamin, which
is a protein known to interact with caveolins. Thus, indi-
cating an interaction between CAV-1 and dynamin at the
NMJ and/or in neurons.
Mutations in human caveolin-3 are known to
Keywords
Levamisole ? Muscular-dystrophy
Aldicarb ? Caveolin ? Dynamin ?
Introduction
Caveolae are 50–100 nm invaginations located at the
plasma membrane of many cell types. They are enriched
in lipid raft associated molecules: cholesterol, sphingo-
lipids, glycosphingolipids and many signalling and recep-
tor proteins (Anderson 1998). Although identified over
50 years ago, the function of these structures has remained
elusive. However, it is known that they play a role in cell
signalling, and clathrin-independent endocytosis (Parton
and Richards 2003). Caveolae require the caveolin protein
for formation. Removal of caveolin from cell lines or from
knock-out mice causes a complete loss of caveolae (Drab
et al. 2001). Furthermore, introduction of caveolin into
cells that do not normally produce caveolae stimulates
caveolar bio-genesis (Lipardi et al. 1998). Mammals have
three caveolin genes: caveolin-1 and caveolin-3 are very
similar and are essential for caveolae formation. Caveolin-
2 forms heterodimers with caveolin-1. Caveolin-3 is mus-
cle specific, and several mutations in this gene are associ-
ated with muscle pathologies including rippling muscle
disease (RMD), limb-girdle muscular dystrophy (LGMD)
type 1C, distal myopathy and hyperCkemia (Woodman
et al. 2004). The mechanisms by which caveolin-3 muta-
tions alter muscle function are unknown, but it is known
that caveolin-3 is associated with the trafficking of dys-
ferlin, a protein known to be involved in muscular dys-
trophy diseases (Hernandez-Deviez et al. 2006) and
dystrophin, a protein which leads to muscular dystrophy
when defective (McNally et al. 1998). A range of mutations
associated with myopathies in the caveolin-3 gene are
known (McNally et al. 1998; Minetti et al. 1998; Woodman
et al. 2004). Of these, most are single amino acid substi-
tutions and a number are known, or believed to be, domi-
nant-negative mutations.
S. Parker ? H. S. Peterkin ? H. A. Baylis (&)
Department of Zoology, University of Cambridge,
Downing Street, Cambridge CB2 3EJ, UK
e-mail: hab@mole.bio.cam.ac.uk
Present Address:
S. Parker
Department of Molecular Microbiology and Immunology,
Saint Louis University Medical School,
1402 South Grand Blvd, St Louis, MO 63104, USA
123
Invert Neurosci (2007) 7:157–164
DOI 10.1007/s10158-007-0051-5

Page 2

Caenorhabditis elegans has two caveolin genes; cav-1
and cav-2, that are homologous to their mammalian coun-
terparts (Tang et al. 1997). The CAV-1 protein is equally
similar to both caveolin-1 and caveolin-3 from mammals.
Previous results (Scheel et al. 1999) showed that cav-1 is
expressed predominantly in the nervous system in adult
worms; however, no neuronal function has been elucidated.
To investigate the role of cav-1, we used fluorescently
tagged proteins to elucidate the localisation of CAV-1. We
show that it is expressed in neurons and body-wall muscles.
To deplete CAV-1, we generated two mutations that are
analogous to previously described dominant-negative
mammalian caveolin-3 mutations identified in patients with
either LGMD or RMD. In conjunction with RNAi and
analysis of a deletion mutant, this approach revealed that
CAV-1 functions at the neuromuscular junction in C. ele-
gans. We also investigated possible interactions with dyn-
amin.Theroleof dynamin
endocytosis is well established [reviewed by (Takei et al.
2005)]. Other evidence also suggests that dynamin plays an
integral part in caveolae function. We show that in C. ele-
gans, cav-1 and dynamin (dyn-1) interact and induce a
locomotion phenotype. Taken together, these data suggest
that C. elegans would make an appropriate model for the
study of caveolinopathies and dystrophies.
inclathrin-meditated
Materials and methods
Manipulation of C. elegans, RNAi and phenotypic
analysis
Caenorhabditis elegans was cultured using standard media
and methods (Fleming et al. 1997; Sulston and Hodgkin
1988).
Levamisole and aldicarb assays were conducted on
NGM plates containing 100 lM or 400 lM levamisole, or
1 mM aldicarb. E. oli OP50 was seeded overnight on the
plates. Briefly, larval stage 4 (L4) worms were placed on
plates and monitored for complete paralysis (i.e. animals
were incapable of locomotion when prodded) every 10 min
over a 100 min period. Radial dispersion assays were
conducted on seeded NGM plates. Worms were placed at a
marked spot and left for 3 min, the distance from the
starting location was scored. All assays were conducted at
20?C unless indicated otherwise. Twenty worms were used
in each assay. Assays were repeated five times.
For fluorescence studies a Leica SP laser confocal
microscope was used with excitation at 488 nm or, for co-
localisation studies, a Leica SP confocal microscope
equipped with 458 and 514 nm laser lines was used.
DNA and dsRNA constructs were injected into the
animals’ gonads. To make transgenic animals, DNA con-
structs were microinjected at 7 ng/ll. rol-6(su1006) was
used as a marker to isolate transformants, except for
locomotion assays where a mec-7 GFP plasmid was used
(pHAB200).
GFP/YFP/CFP fusion and site directed mutagenesis
CAV-1::GFP fusion proteins were generated with GFP
inserted close to the 5¢ or 3¢ end of the gene with 4 kb of
wild-type DNA upstream of the ATG start codon and 2 kb
of DNA downstream of the stop codon. A GFP-only
transcriptional fusion was made with the same 4 kb of
upstream DNA driving GFP. C-terminal CFP or YFP fu-
sions were generated using the PCR fusion technique
(Hobert 2002). In these constructs YFP/CFP is coupled to a
let-858 3¢ UTR.
All point mutations were generated using the Quick-
Change site directed mutagenesis kit (Stratagene).
Statistics
Unpaired t tests were used to compare the mean values
from two groups. P values are indicated in the results.
Mean values are plotted graphically, error bars indicate the
standard error of the mean.
Results
Caveolin-1 is expressed in the nervous system
and body-wall muscles
In order to explore the function of cav-1 we made two cav-
1::GFP fusions in which GFP was inserted close to the N-
or C-terminal coding regions. These constructs included
4 kb of DNA upstream of the cav-1 coding sequence, thus
including all of the putative, native cav-1 promoter. Fol-
lowing transformation, both constructs yielded identical
expression patterns. CAV-1::GFP was expressed in most, if
not all, cells in embryos. It was predominantly localised to,
or close to, the membrane in these cells although some
small intracellular bodies were also visible (Fig. 1 a). This
widespread expression continued into the early larval
stages, but with maturation, CAV-1::GFP became re-
stricted to most, but not all, neurons (Fig. 1b). Staining was
seen in the ventral nerve cord (VNC), and was also fre-
quently observed in the commissures. Of particular interest
was the punctate nature of labelling which is most obvious
in the VNC, this staining suggests possible localisation to
synapses. A similar expression pattern with an anti-CAV-1
antibody was observed by Scheel (1999).
The identification of cells or cell types expressing
fluorescent proteins is most readily performed when the
158Invert Neurosci (2007) 7:157–164
123

Page 3

protein is present in the cell body. We considered that the
full expression pattern of cav-1 might be being masked due
to the highly localised nature of CAV-1::GFP. Therefore,
we analysed a construct that contained a transcriptional
fusion of the cav-1 promoter alone driving GFP. This
construct revealed that body wall muscles also express cav-
1 (Fig. 1c). C. elegans differs from most other animals
because motor neurons do not leave the nerve fascicle to
innervate their target. Instead, body wall muscles extend
muscle arms that intercept neuronal process running on the
outside of the bundle in the nerve cord and form en-passant
neuromuscular junctions (NMJs) (Hedgecock et al. 1990;
Hedgecock and Hall 1991). Thus, if CAV-1::GFP is tar-
geted to muscle arms and/or synaptic regions in body wall
muscles it is likely that it will appear to be present close to
the nerve cord.
To test the localisation of CAV-1 in the VNC we looked
for co-localisation with known pre- and post-synaptic
markers. To this end, we generated CFP translational
fusions to the markers and looked for co-localisation in
conjunction with a CAV-1::YFP fusion. The following pre-
synaptic markers were used: (1) synaptotagmin (snt-1),
which is localised to regions known to be rich in synapses
and appears to be associated with synaptic vesicles (Nonet
et al. 1993; Richmond and Broadie 2002). (2) Endophilin
(unc-57), a pre-synaptic protein required for endocytosis of
synaptic vesicles and synaptic vesicle recycling in C. ele-
gans (Schmidt et al. 1999; Schuske et al. 2003). (3) Syn-
aptobrevin (snb-1), expressed in GABAergic neurons under
the control of the unc-47 promoter (McIntire et al. 1997;
Nonet et al. 1998). No substantial co-localisation was de-
tected with any of these pre-synaptic markers (Fig. 2a, b,
and unpublished data). Thus we conclude that cav-1 is
unlikely to be presynaptic within the VNC.
Since CAV-1 is also expressed in muscle cells, we
reasoned that it could be localised to the muscle arms
containing the post-synaptic regions of NMJs, we therefore
tested for co-localisation with UNC-63. UNC-63 is a nic-
otinic acetylcholine receptor (nAChR) a-subunit that is
expressed predominantly in body wall and vulval muscles
with some neuronal expression (Culetto et al. 2004; Ho-
sono and Kamiya 1991). Confocal images of worms car-
rying both cav-1::YFP and unc-63::CFP revealed that
CAV-1::YFP and UNC-63::CFP co-localise in some parts
of the nervous system and in structures that are either part
of, or close to, the VNC (Fig. 2c, d). The latter localisation
suggests that CAV-1 and UNC-63 co-localise in muscle
arms. This pattern may include localisation to the post-
synaptic regions of NMJs and extrasynaptic positioning.
Within the nervous system as a whole, obvious area
without co-localisation also exist, these are particularly
evident in neurons that express only unc-63 and not cav-1
or vice versa, confirming that cav-1 is not expressed in all
neurons.
The hypothesis that CAV-1 is located post-synaptically
at NMJs, reminiscent of mammalian caveolin-3 expression
(Carlson et al. 2003), is supported because CAV-1 failed to
co-localise with any of the pre-synaptic markers, but did
Fig. 1 Expression of cav-1 in C. elegans. a cav-1::gfp translational
fusion constructs under the control of the native cav-1 promoter
reveal that CAV-1::GFP is expressed in embryonic cells and localises
to the cell periphery. Intracellular CAV-1 positive vesicles can also be
detected (arrowed). b With maturation, CAV-1::GFP becomes
restricted to neuronal structures, it is particularly evident in the
ventral nerve cord (arrowed). c A transcriptional GFP fusion driven
by the native cav-1 promoter allows for clearer identification of cells
expressing GFP (and therefore cav-1). In this case, body wall muscle
cells (arrows) and neuronal processes are visible
Fig. 2 CAV-1 co-localises with UNC-63. CAV-1::YFP (green and
indicated by block-headed arrows) failed to co-localise with the pre-
synaptic markers, SNT-1::CFP (synaptotagmin a) and UNC-57::CFP
(endophilin b) (both in red and indicated by open-headed arrows).
c CAV-1::YFP (green) co-localises in some places (yellow), indicated
by the block-headed arrow, with the post-synaptic nAChR subunit,
UNC-63::CFP (red). Areas of no co-localisation are detected and are
showed by the open-headed arrows. d The ventral nerve cord (VNC)
presented with good co-localisation (yellow). Embryos in (b) are
indicated by ‘‘emb’’. Auto-fluorescent gut granules are also shown in
red
Invert Neurosci (2007) 7:157–164159
123

Page 4

localise with UNC-63. Cav-1 also appears to be expressed
elsewhere in the nervous system—but it is unlikely that it is
at the pre-synaptic membranes of NMJs as it does not co-
localise with presynaptic markers in the motor neurones of
the ventral nerve cord.
CAV-1 functions in acetylcholine signalling
To test the function of cav-1 in C. elegans, and specifically
in the neuromuscular system, we aimed to perturb CAV-1.
We used a dominant-negative transgenic approach to mimic
the effects of muscular dystrophy associated mutations.
Two reports have shown that in at least four unrelated cases
an ala46thr substitution in human caveolin-3 prevented the
localisation of the protein to the plasma membrane in a
dominant-negative fashion (Betz et al. 2001; Herrmann
et al. 2000; Vorgerd et al. 1999). A second dominant-neg-
ative mutation, asp27glu, was reported in human caveolin-3
in patients with RMD, protein reduction was detected by
western blots, supporting a dominant-negative effect
(Fischer et al. 2003). The regions of human caveolin-3
containing these mutations are well conserved in C. elegans
CAV-1 (Tang et al. 1997), although the equivalent to ala46
is gly126. Site directed mutagenesis was used to create the
equivalent changes in CAV-1, gly126thr and asp108glu, to
produce cav-1G126Tand cav-1D108E, respectively. Both
mutations were made in the cav-1 and cav-1::GFP con-
structs and expressed under the control of the native cav-1
promoter. Confocal analysis revealed that cav-1::GFPD108E
had a substantial reduction in detectable GFP whilst cav-
1::GFPG126Thad only a small reduction in fluorescence
(data not shown). Worms expressing the dominant-negative
constructs grew well and did not have any gross pheno-
types. We also tested a cav-1 deletion mutant (ok2089), this
strain grows poorly for reasons that are currently unclear.
Aldicarb and levamisole assays were used to test the
function of CAV-1 at the synapse. Worms are gradually
paralysed by these two compounds. Previous analysis has
shown that resistance to levamisole, a nAChR agonist,
results from mutations affecting levamisole sensitive post-
synaptic nAChRs, or putative downstream components at
the NMJ. Whereas strains carrying mutations that reduce
pre-synaptic ACh release are resistant to aldicarb, which is
an inhibitor of acetylcholine esterase. It should be noted
that post-synaptic changes can also alter aldicarb sensitiv-
ity. The use of these two compounds to dissect synaptic
function is well established (Nguyen et al. 1995; Nonet
et al. 1998; Nurrish et al. 1999; Rand and Russell 1985).
Wild type (N2) and cav-1 dominant-negative worms
were subjected to 100 lM levamisole or 1 mM aldicarb
and assayed for paralysis. Both the cav-1G126T(HB544)
and cav-1D108E(HB542) worms were more sensitive to
levamisole compared to N2s (Table 1, Fig. 3a). The in-
creased sensitivity is most pronounced in animals carrying
the D108E mutation (HB542), which also had the most
substantial decrease in cav-1::GFP expression. Cav-
1(ok2089) worms were also more sensitive (Fig.3c). Thus,
depletion of cav-1 increases levamisole sensitivity at the
NMJ, this again indicates a post-synaptic role for CAV-1.
Aldicarb treatment had no affect on HAB542 or 544
(Fig. 3b) or on cav-1(ok2089) (data not shown). Depletion
of cav-1 using dsRNA had no effect on the animals’ sen-
sitivity to either levamisole or aldicarb (Table 1).
Cav-1 and dynamin interact at the genetic level
Dynamin associates with mammalian caveolin-1 (Kim and
Bertics 2002) and mediates caveolar fission (Oh et al.
1998). We used dyn-1(ky51) a temperature-sensitive allele
of dyn-1 which contains a mutation within the GTPase
Table 1 Effect of aldicarb and levamisole treatment on cav-1 RNAi and cav-1 dominant-negative animals
StrainGenotypeTreatmentResponse to aldicarb Response to levamisole
+/–T50%A
+/– T50%L
N2 Wild-type– 5571
N2Wild-type
cav-1 RNAin/e 55n/e72
N2Wild-type
CAV-1D108E
CAV-1G126T
cat RNAin/e 55n/e71
HB542
–n/e55+++45
HB544
–n/e 55++59
CX51
dyn-1(ky51)
–42 30
CX51
dyn-1(ky51)cat RNAi n/e 41 n/e31
CX51
dyn-1(ky51)
dyn-1(ky51), CAV-1D108E
cav-1 RNAi– – 55n/e 29
HB552
–– –55++18
T50%Aand T50%Lrepresent the average time for 50% of the animals to become paralysed on aldicarb and levamisole, respectively. +/– is a
measure of the sensitivity (+) or resistance (–) derived from comparison of the T50% values to the relevant control strain: + (or –) 2–10 min
change, ++ (or –) 11–20 min change, +++ (or – –) >21 min change, n/e (no effect) <2 min change
160Invert Neurosci (2007) 7:157–164
123

Page 5

domain. CX51, dyn-1(k51) worms exhibit a reversible,
uncoordinated phenotype when shifted to higher tempera-
tures (‡25?C) (Clark et al. 1997). We detected some
co-localisation between CAV-1::YFP and DYN-1::CFP,
although the over-expression of cav-1::yfp disrupted the
DYN-1::CFP localisation in neurons compared to DYN-
1::YFP alone (data not shown).
To assess whether dyn-1 and cav-1 act in the same
processes we tested the effect of the cav-1D108Emutation
on dyn-1 mutant worms by introducing the construct into
dyn-1(ky51) animals (making HB552). dyn-1(ky51) worms
presented with a large increase in sensitivity to levamisole
and a small increase in sensitivity to aldicarb, compared to
wild-type animals (Table 1). Introduction of the cav-1D108E
construct into dyn-1(ky51) worms resulted in a further in-
crease in sensitivity to levamisole but a decrease in sensi-
tivity to aldicarb, compared to the dyn-1(ky51) control.
When cav-1 was targeted by RNAi in the CX51 strain, the
phenotype mimicked the increase in aldicarb sensitivity
observed in the HB552 strain. However, this was not the
case with sensitivity to levamisole. The observation that
dominant-negative and RNAi levamisole assays gave dif-
ferent results might be explained by the reduced efficiency
of RNAi in the C. elegans nervous system (Fraser et al.
2000; Simmer et al. 2002).
Caveolin interacts with dynamin to regulate locomotion
Locomotion in C. elegans is achieved through sinusoidal
body-wave formations resulting from neuronal regulation
of 95 body-wall muscles (Chalfie et al. 1985; Sulston 1983;
Wakabayashi et al. 2004). Radial dispersion assays were
used to investigate locomotion phenotypes in control and
dyn-1 mutant animals carrying the cav-1D108Econstruct.
Transformants were selected by expression of a GFP
marker (pHAB200). Control animals carry the GFP marker
only; cav-1D108Eanimals carry the GFP marker and the
dominant-negative construct. Strains were tested for loco-
motion deficiencies at 20?C, i.e. below the fully restrictive
temperature for dyn-1(ky51) function however, previous
results also suggest that the function of DYN-1 is com-
promised at 20?C (Clark et al. 1997).
cav-1D108Ehad little to no effect on locomotion com-
pared to control animals (Fig. 4). Compared to controls, the
dyn-1(ky51) mutant had approximately 20% reduction in
dispersion; however, expression of cav-1D108Ein the dyn-
1(ky51) background (Table 1; dyn-1 D108E, HAB200)
furtherreducesdispersion
(P = 0.0033). RNAi of cav-1 in wild type worms (N2s) has
no effect on locomotion or drug sensitivity compared to
controls (Table 1 and data not shown), similarly RNAi
seems to have little effect on the levels of CAV-1::GFP in
transgenic animals (not shown). But, cav-1 RNAi in a dyn-
1 background resulted in a reduction in animal dispersion
(Table 1). (P < 0.0001 compared to the chloramphenicol
acetyltranferase control cat). RNAi and dominant-negative
worms have similar locomotion phenotypes, suggesting
by approximately45%
a
b
100
0
20
40
60
80
100
0 20 40
Time (min)
6080100
0
0
1 n
o
d
e
s
y l a r a
p
%
µ
M
e l o
s i
m
a
v
e l
CAV-1D108E
CAV-1G126T
Wild-type
0
20
40
60
80
0 204060 80100
CAV-1D108E
CAV-1G126T
Wild-type
Time (min)
b r a
c i d l a
M
m
1 n
o
d
e
s
y l a r a
p
%
0
20
40
60
80
100
0 2040
Time (min)
60 80100
cav-1(ok2089)
Wild-type
c
0
0
4 n
o
d
e
s
y l a r a
p
%
µ
M
e l o
s i
m
a
v
e l
Fig. 3 Animals carrying cav-1 dominant-negative constructs or cav-1
mutant alleles are sensitive to levamisole. a Wild-type (N2) worms
transformed with the dominant-negative constructs cav-1D108E(strain
HB542) and cav-1G126T(strain HB544) are sensitive to 100 lM
levamisole exposure. The time taken for 50% paralysis (T50%L) is
approximately 45, 55 and 70 min in cav-1D108E, cav-1G126Tand N2s,
respectively. b Neither strain exhibited any sensitivity to aldicarb
compared to the N2 controls. c cav-1(ok2089) worms are also more
sensitive to levamisole and have a T50%Lof 43 min. Error bars
indicate SEM
Invert Neurosci (2007) 7:157–164161
123