Cell, Vol. 114, 751–762, September 19, 2003, Copyright 2003 by Cell Press
Control of Myoblast Fusion by a Guanine
Nucleotide Exchange Factor, Loner,
and Its Effector ARF6
regulators of myoblast fusion identified genetically in
Drosophila are likely to provide insights into mammalian
myogenesis, as well as intercellular fusion in general.
The somatic musculature of the Drosophila embryo
is derived from the embryonic mesoderm. During mid-
embryogenesis, mesodermal cells expressing twist (twi)
acquire a myoblast cell fate. Subsequently, a subset
of myoblasts expressing lethal of scute is selected via
lateral inhibition to become muscle founder cells, while
the remaining twi-expressing cells become fusion-com-
petent (Baylies et al., 1998; Frasch, 1999). The founder
cellsare asourceof attractant(s)for surroundingfusion-
competent myoblasts, and fusion between these two
populations of cells leads to the formation of myotubes
that incorporate between 4 and 25 myoblasts. The sub-
sequent epidermal attachment of myotubes results in a
highly stereotyped, segmentally repeated pattern of 30
muscle fibers per hemisegment (Bate, 1993).
Myoblast fusion is a multistep process involving the
initial recognition and adhesion between muscle founder
cells and fusion-competent myoblasts, subsequent align-
ment of two adhering cells, and ultimate membrane
breakdown and fusion (Doberstein et al., 1997). Similar
ultrastructural changes associated with these events
occur in vertebrate and Drosophila muscle cells, as well
as in nonmuscle cells that undergo fusion (Wakelam,
1985; Knudsen, 1992; Hernandez et al., 1996; Blumen-
thal et al., 2003). Genetic studies in Drosophila have
begun to identify components of a possible signaling
cascade required for myoblast fusion (Figure 7, for
reviews see Paululat et al., 1999; Frasch and Leptin,
2000; Baylies and Michelson, 2001; Dworak and Sink,
2002; Taylor, 2002). dumbfounded (duf)/kin of Irregular-
chiasm-C (kirre) and roughest (rst)/irregular-chiasm-C
(irreC) encode paralogues of immunoglobulin (Ig) do-
main-containing transmembrane receptor-like proteins
that are specifically required in founder cells (Ruiz-
Gomez et al., 2000; Strunkelnberg et al., 2001). sticks
and stones(sns) and hibris(hbs) encodetwo paralogues
of transmembrane proteins with Ig domains that are
expressed, and in the case of sns required, in fusion-
competent myoblasts (Bour et al., 2000; Artero et al.,
2001; Dworak et al., 2001). It has been suggested that
DUF/KIRRE and RST/IRREC act as attractants for fu-
sion-competent myoblasts by interacting with SNS and
kelnberg et al., 2001). The antisocial (ants)/rolling peb-
bles (rols7) gene encodes a founder cell-specific intra-
by linking membrane fusion receptors and the cytoskel-
eton (Chen and Olson, 2001; Menon and Chia, 2001;
Rau et al., 2001). myoblast city (mbc) encodes a cyto-
skeleton-associated protein with homology to the hu-
man protein DOCK180 (Erickson et al., 1997). Recently,
a DOCK180-ELMO complex was shown to function as
factor (GEF) for the small G protein Rac in phagocytosis
(Brugnera et al., 2002). Interestingly, the Drosophila ho-
mologs of Rac, Drac1, and Drac2, play essential roles
in myoblast fusion (Hakeda-Suzuki et al., 2002). Despite
Elizabeth H. Chen, Brian A. Pryce, Jarvis A. Tzeng,
Guillermo A. Gonzalez, and Eric N. Olson*
Department of Molecular Biology
University of Texas Southwestern Medical Center
6000 Harry Hines Boulevard
Dallas, Texas 75390
Myoblast fusion is essential for the formation and re-
generation of skeletal muscle. In a genetic screen for
regulators of muscle development in Drosophila, we
discovered a gene encoding a guanine nucleotide ex-
change factor, called loner, which is required for myo-
and acts downstream of cell surface fusion receptors
by recruiting the small GTPase ARF6 and stimulating
guanine nucleotide exchange. Accordingly, a domi-
nant-negative ARF6 disrupts myoblast fusion in Dro-
sophila embryos and in mammalian myoblasts in cul-
ture, mimicking the fusion defects caused by loss of
Loner. Loner and ARF6, which also control the proper
membrane localization of another small GTPase, Rac,
are key components of a cellular apparatus required
for myoblast fusion and muscle development. In mus-
receptors; in other fusion-competent cell types it may
be triggered by different upstream signals.
Intercellular fusion is fundamental to the formation of
multicellular organisms and is required for processes as
and myogenesis (Hernandez et al., 1996; Blumenthal et
al., 2003). Recent studies have also revealed a role for
poietic stem cells (Blau, 2002). Whereas the mecha-
nisms involved in intracellular membrane fusion are un-
derstood in considerable detail, there is a dearth of
vide opportunities for its manipulation, which has obvi-
ous and important implications for tissue engineering
following injury, mononucleated myoblasts fuse to form
multinucleated muscle fibers. The process of myoblast
fusion is amenable to genetic dissection in the fruit fly
Drosophila melanogaster, in which muscle formation
involves a well-defined temporo-spatial sequence of
events thatareremarkablyconserved inmammalianmyo-
genesis (Wakelam, 1985; Knudsen, 1992; Doberstein et
al., 1997). Given the evolutionary conservation of the
cellular and molecular events of muscle development,
Figure 1. Myoblast Fusion Defect in loner
The somatic musculature in wild-type (A) and
lonerT1032embryos (B and C) are visualized by
a MHC-tauGFP reporter. Embryos are ori-
ented with dorsal up and anterior to the left
in this and all other figures.
(A) Ventrolateral view of a portion of a stage
14 wild-type embryo showing the segmen-
tally repeated pattern of its somatic muscu-
(B and C) Lateral view of a portion of a stage
13 (B) and stage 14 (C) lonerT1032embryo in
which myoblasts fail to fuse. Fusion-compe-
tent myoblasts extend filopodia (arrows) to-
ward elongated founder cells, suggesting
that adhesion between fusion-competent
myoblasts and founder cells is not affected.
Wild-type (D and F) and lonerT1032mutant em-
bryos (E and G) were stained with anti-KR (D
and E) and anti-DMEF2 (F and G) antibodies.
(D and E) Lateral view of stage 13 wild-type
and lonerT1032embryos stainedfor KR. In wild-
type embryos, KR is initially expressed in a
subset of founder cells, but is later turned on
in other nuclei of the multinucleated fibers as KR-positive founder cells fuse to neighboring myoblasts. Thus, KR staining appears as clusters
in the wild-type embryo (D). In the loner mutant embryo (E), KR is expressed in isolated, instead of clusters of, nuclei due to lack of fusion.
(F and G) Lateral view of stage 14 embryos showing similar number of DMEF2-expressing myoblasts in wild-type (F) and lonerT1032mutant
the discovery of the DUF/RST→ANTS→MBC→Rac sig-
naling pathway within founder cells, the complexity of
the fusion process predicts additional molecules that
function together with these components to accomplish
the events of fusion.
Here, we describe the discovery and mechanisms of
action of Loner, a GEF of the Sec7 family that acts
downstream of myoblast fusion receptors. Loner, which
is localized to subcellular sites of fusion, controls myo-
blast fusion by recruiting the small GTPase ARF6 and
promoting its guanine nucleotide exchange. The Loner/
ARF6 module acts in parallelto, and converges with, the
ANTS→MBC→Rac pathway. This fusigenic mechanism,
cells, has the potential to control fusion of other cell
types by coupling to different upstream effectors.
not shown). The phenotype resulting from the loner mu-
tation is therefore highly specific to the somatic muscu-
loner Is Specifically Required for Myoblast Fusion
In order to determine if the loner phenotype was due to
a specific defect in myoblast fusion or secondary de-
fects in myoblast fate determination or other develop-
mental processes, we examined several developmental
tion, including the specification of muscle founder cells
and myoblasts, the pattern of innervation by motor neu-
rons, and differentiation of the epidermis. Muscle
founder cell specification was assessed by expression
of Kru ¨ppel (Kr), which is initially expressed in a subset
of founder cell nuclei in wild-type embryos (Ruiz-Gomez
et al., 1997). Later, as neighboring myoblasts fuse with
KR-expressing founder cells, their nuclei also express
KR, resulting in clusters of KR-positive cells in the em-
bryo (Figure 1D). As shown in Figure 1E, KR was ex-
pressed in its characteristic positions in loner mutant
embryos, suggesting that founder cell fate was properly
specified. However, only one nucleus was present in
each “cluster” of KR-expressing cells, suggesting that
petent myoblasts. We also examined Dmef2, which is
expressed in the nuclei of all somatic, visceral, and car-
diac myoblasts (Lilly et al., 1994; Nguyen et al., 1994)
(Figure 1F). Wild-type and loner mutant embryos had
comparable numbers of DMEF2-expressing cells, sug-
gesting that the mutant myoblasts were properly speci-
fied despite a block in myoblast fusion (Figures 1F and
1G). In addition, antibody staining with fasciclin II (Gren-
ningloh et al., 1991) revealed a normal pattern of muscle
innervation by motor neurons and cuticle preparations
Identification of loner
We performed a genetic screen in Drosophila using an
MHC-tauGFP line to identify newgenes involved in skel-
etal muscle development (E.H.C. and E.N.O., unpub-
lished data). One complementation group on the third
chromosome containing two EMS mutant alleles, T1032
and T1057, showed a striking mutant phenotype in
Instead of mature, multinucleated muscle fibers, a large
number of unfused myosin-expressing myocytes were
present in mutant embryos (compare Figures 1A with
1B and 1C). Based on the failure of mutant myoblasts
to fuse with surrounding cells, we named this locus
loner. All skeletal muscles appeared to be affected in
loner mutant embryos. In contrast, there were no gross
defects in visceral muscles or the dorsal vessel (data
Loner, a Unique GEF Required for Myoblast Fusion
Figure 2. MolecularCharacterizationofloner
(A) Genomic organization of the loner gene.
Three alternatively spliced forms are shown.
Black boxes represent common exons, and
white boxes represent untranslated regions.
The three alternatively spliced exons are la-
beled in red and indicated as iso1, iso2,
(B) Schematic structure of Loner ISO1, ISO2,
and ISO3 proteins. All three isoforms share
an IQ motif, a Sec7 domain, a PH domain,
and a C-terminal coiled-coil domain. Loner
ISO1 is predicted to encode a 1325 amino
acid protein that also contains an N-terminal
coiled-coil domain. Loner ISO2 and Loner
ISO3 are predicted to encode 1315 and 1210
amino acid proteins, respectively.
(C) Comparison of the molecular structures
of Loner ISO1 and human Loner (hLoner), in-
between each of the conserved domains.
revealed apparently normal differentiation of epidermis
in loner mutant embryos (data not shown). Thus, the
unfused myoblast phenotype in loner mutant embryos
is likely due to a specific defect in myoblast fusion.
Since myoblast fusion is a multi-step process requir-
ing cell-cell recognition, adhesion, alignment, and co-
of these steps was blocked in loner mutant embryos.
Previous studies showed that in duf rst double-mutant
embryos, fusion-competent myoblasts extend filopodia
at random orientations and are not attracted by founder
cells (Ruiz-Gomez et al., 2000). In ants/rols7 mutant em-
ters around the founder cells and extend filopodia to-
ward their fusion targets, indicating that the fusion
process arrests after the initial attraction (Chen and
Olson, 2001; Menon and Chia, 2001; Rau et al., 2001).
This is consistent with ANTS functioning downstream
ysis of loner mutant embryos revealed that fusion-com-
petent myoblasts extendedfilopodia toward their fusion
targets, as in the ants mutant (Figures 1C), suggesting
that loner also functions after the initial recognition and
adhesion step required for fusion. Occasionally, we ob-
served miniature fibers that contained two nuclei, sug-
gesting that limited fusion can occur in the absence
predicted genes. We sequenced the coding regions of
several predicted genes in this chromosomal interval
for potential molecular lesions in the loner mutant and
detected pointmutations in a predictedgene, CG32434.
5? RACE experiments led to the identification of three
alternatively spliced forms of CG32434, represented by
three EST clones: RE02556, LP01489, and GH10594,
respectively. We refer to these forms of loner, which
vary only in their first exons, as isoforms 1, 2, and 3
(Figure 2A). Both loner alleles harbored nonsense muta-
tions in the third exon, a region common to all three
isoforms. lonerT1032contained a C to T mutation that
changed glutamine(Q)-414 of the predicted ISO1 (Q402
of ISO2 or Q299 of ISO3) to a stop codon. Remarkably,
lonerT1057contained a C to T mutation that changed the
adjacent Q415 of ISO1 (Q403 of ISO2 or Q300 of ISO3)
to a stop codon (Figures 2A and 2B).
To confirm that CG32434 corresponded to the loner
gene, we generated transgenic flies in which each of
the isoforms were expressed under the control of the
UAS promoter. Expression of any of the three isoforms
in the presence of the mesodermal specific twi-GAL4
driver completely rescued the fusion defects in loner
mutant embryos (Figures 3B, b–d). For embryos harbor-
ing the iso3 transgene, no mononucleated myoblasts
were observed, although there were some fiber-pat-
terning defects (Figure 3B, d). A transgene encoding
ISO2 under control of the ubiquitous tubulin promoter
not only rescued the fusion phenotype of the somatic
muscles, but also rescued loner mutants to adulthood
(data not shown). These results provide conclusive evi-
dence that CG32434 corresponds to loner.
Molecular Cloning of loner
loner was mapped to a small chromosomal region of
78A4-B1 by deficiency mapping (see Experimental Pro-
cedures). The proximal breakpoint of the complement-
ing deficiency Df(3L)Pc-cp2 (78B1-2; 78D) breaks within
the knockout gene, not only excluding it as a candidate
for loner, but also providing us with a precise distal
“molecular” boundary of the loner locus relative to other
Domain Structures of the Loner Protein
The three spliced forms of loner encode predicted pro-
teins of 1325, 1315, and 1210 amino acids, respectively.
of the former class have orthologs in all eukaryotes and
are probably involved in evolutionarily conserved as-
pects of membrane dynamics and protein transport.
Members of the latter class do not have orthologs in
S. cerevisiae, suggesting a function specific to higher
eukaryotes. Loner shares a common domain structure
with the low molecular weight Sec7 GEFs, containing a
PH domain besides the Sec7 domain. However, unlike
the other members of this class, Loner has a much
higher predicted molecular weight of over 100 kDa.
Database searches identified a highly homologous
putative mouse Loner protein of 916 amino acids
(mCP20090) and a potential human Loner protein of 963
amino acids (hCP438181) sharing 65% identity in the
Sec7 domain, 59% identity in the PH domain, and 65%
identity in the IQ domain with the fly protein (Figure 2C).
The mouse and human Loner proteins share 94% amino
acid identity. A human EST, KIAA0763, previously
named ARF-GEP100(Someya et al., 2001), is a shorter
form of the human Loner lacking 122 amino acids at the
N terminus. There is also a closely related predicted
C. elegans protein, 4E572, with an IQ motif and Sec7
domain, but without a PH domain. Loner and its or-
thologs represent a unique high molecular weight sub-
class of Sec7 GEFs that contain both Sec7 and PH do-
Figure 3. Phenotypic Rescue by Loner and the Requirement for its
GEF Activity and PH Domain In Vivo
(A) Wild-type and mutant Loner proteins were expressed under the
control of the twi promoter in transgenic rescue experiments. The
column on the right indicates whether or not these proteins rescued
the myoblast fusion defect in the loner mutant.
(B) Somatic musculature of loner mutant embryos expressing vari-
ous transgenes as indicated, stained with anti-MHC. Stage 14 em-
bryos are shown. Note that all three isoforms, iso1 (b), iso2 (c), iso3
(d), as well as iso1?IQ (h), rescued the myoblast fusion phenotype
of the loner mutant embryo. However, iso1?Sec7 (e), iso1Sec7E→K
(f), and iso1?PH (g) did not rescue the mutant phenotype.
Expression Pattern of loner
The embryonic expression pattern of loner was exam-
ined by in situ hybridization. loner exhibits a dynamic
at stage 4 (Figure 4A). Mesodermal expression of loner
is initiated at embryonic stage 11, at the onset of fusion
(Figure 4B). It is also expressed in the neuroectoderm.
As germ band shortening proceeds, the somatic meso-
dermal expression persists until stage 14 (Figures 4C–
4E). When fusion is completed at stage 15, loner is no
longer expressed in the mesoderm. Instead, strong ex-
4F). The temporal expression of loner in the somatic
mesoderm coincides precisely with the fusion process,
consistent with its requirement for myoblast fusion.
Their domain homologies suggest that Loner is a GEF
(Figure 2B). All three protein isoforms contain a Sec7
domain, an adjacent pleckstrin homology (PH) domain,
a C-terminal coiled-coil domain, and an IQ-motif. ISO1
also contains a coiled-coil domain at its N terminus.
Sec7 domains are regions of ?200 amino acids with
strong homology to the yeast protein Sec7p (Shevell et
al., 1994; Morinaga et al., 1997). Sec7 domains possess
GEF activity toward a family of ubiquitously expressed
small GTPases called ADP-ribosylation factors (ARFs),
which have been implicated in a variety of vesicular
transport and cytoskeleton rearrangement processes in
eukaryotic cells (Moss and Vaughan, 1998; Chavrier and
Goud, 1999; Donaldson and Jackson, 2000; Jackson
charged phospholipids of cell membranes and are able
to enhance GEF activity (Chardin et al., 1996; Paris et
al., 1997). IQ motifs are believed to mediate binding to
calmodulin (Rhoads and Friedberg, 1997).
Previously identified members of the Sec7 family are
subdivided into two major classes based on sequence
similarities and functional differences: high molecular
(45–50 kDa) (Jackson and Casanova, 2000). Members
Loner Is Localized to Discrete Cytoplasmic Foci
in Founder Cells
The subcellular distribution of the Loner protein in mus-
cle cells was determined by double-labeling experi-
ments with anti-Loner and anti-?-galactosidase (?-gal)
antibodies using the rP298 enhancer trap line, which
carries a P[LacZ] element insertion in the 5? promoter
of the founder cell-specific duf gene (Ruiz-Gomez et al.,
2000). Confocal microscopy demonstrated that Loner
was localized to the lacZ-expressing founder cells (Fig-
ure 4G). Furthermore, Loner protein expression was in-
creased in Notch (N) mutant embryos (Figure 4K), which
produce excess founder cells (Corbin et al., 1991; Fuer-
stenberg and Giniger, 1998; Rusconi and Corbin, 1998).
Interestingly, Loner is a cytoplasmic protein that aggre-
gates to discrete foci (Figures 4G and 4J). The punctate
appearance of Loner staining is reminiscent of that of
ANTS/ROLS7, a founder cell-specific adaptor molecule
Loner, a Unique GEF Required for Myoblast Fusion
that is localized to sites of muscle cell fusion (Chen and
Olson, 2001; Menon and Chia, 2001).
Loner Is Recruited to the Cell Membrane
by Founder Cell-Specific Receptors
The punctate distribution of Loner prompted us to ask
if its subcellular localization was, like ANTS, dependent
on fusion receptors (Chen and Olson, 2001; Menon and
Chia, 2001). To address this question, we established a
cell-culture-based assay in S2 cells to investigate the
interactions between fusion receptors and downstream
components of the fusion-signaling cascade. When ex-
pressed alone, DUF, a homophilic adhesion molecule
(Dworak et al., 2001), localized to membrane regions
between adhering cells, whereas ANTS localized in the
cytoplasm of S2 cells (Figures 5A and 5G). However,
when DUF and ANTS were coexpressed in S2 cells,
ANTS colocalized with DUF at the membrane region
between adhering cells (Figures 5H and 5I). The recruit-
ment of ANTS by DUF to cell-cell contacts in this assay
agrees with previous studies showing that the subcellu-
lar localization of ANTS is dependent on DUF function
(Chen and Olson, 2001; Menon and Chia, 2001). Using
interaction between DUF and Loner. While Loner local-
ized to distinct foci in the cytoplasm of S2 cells when
expressed alone (Figure 5J), it was recruited to mem-
brane regions of cell-cell contacts in the presence of
DUF (Figures 5B and 5C). This recruitment is highly spe-
cific, since coexpressing Loner with another cell adhe-
sion molecule, Delta, did not result in membrane local-
ization of Loner (Figure 5K). Thus, Loner, like ANTS, can
be recruited to membrane regions of cell-cell contact
by DUF. RST, a DUF-related fusion receptor that plays
redundant roles with DUF in myoblast fusion (Strunkeln-
berg et al., 2001), was also able to recruit Loner and
ANTS to cell-cell contacts in the S2 cell assay (Figures
5D–5F, and data not shown). We have not detected
direct interactions between Loner and DUF/RST (data
not shown), suggesting the involvement of additional
To substantiate a role for DUF/RST in the subcellular
localization of Loner, we examined Loner protein in
homozygous mutant embryos of a small deficiency,
Df(1)w67k30, which removed both duf and rst. As shown
in Figure 4M, instead of localizing to discrete foci in
founder cells, at least a portion of the Loner protein was
distributed throughout the cytoplasm and appeared as
rings that outlined the founder cells in duf rst double-
mutant embryos. Interestingly, some discrete foci of
Loner protein remained in founder cells, suggesting the
existence of different pools of Loner protein in the cyto-
Figure 4. loner Expression during Embryogenesis and Subcellular
Localization of the Loner Protein in Muscle Founder Cells
(A-F) Expression of loner transcripts in wild-type embryos detected
by RNA in situ hybridization.
(A) At stage 4, loner is expressed in several stripes along the antero-
posterior axis of the embryo.
toderm and begins to be expressed in the somatic mesoderm
(D and E) At stage 13 and 14, loner expression is clearly seen in the
somatic mesoderm (arrow). It is also expressed in the embryonic
CNS, which is not in focus.
(F)At stage15,loner expressiondisappearsfromthe somaticmeso-
derm. However, its expression in the CNS persists (arrowhead).
(H and I) Localization of ANTS in stage 13 wild-type (H) and loner
mutant (I) embryos. Muscle cells in two hemisegments are shown.
Note that the punctate subcellular localization of ANTS, labeled in
green, remained the same in the loner mutant embryo as in wild-
(G and J–N) Confocal images of stage 13 wild-type and mutant
embryos showing the subcellular localization of Loner protein.
(G) An embryo carrying rP298-lacZ and double-labeled with anti-
Loner (green) and anti-?-gal (red) antibodies. Muscle cells in five
hemisegments are shown. Loner staining appears as discrete foci
associated with founder cell nuclei.
(J–N) The nuclei of a subset of founder cells were labeled by anti-
Nautilus (NAU) (red). A cluster of muscle cells within a single hemi-
segment is shown in each image.
(J) Loner expression is associated with founder cells in a wild-
(K) In a Notch mutant embryo, where there are an increased number
of founder cells, elevated Loner expression is seen associated with
the founder cells.
(L and N) In sns and ants mutant embryos, the punctate pattern of
Loner expression remains unchanged.
(M)In aduf rstmutant embryo,Loner stainingis distributedthrough-
out the cytoplasm and appears to outline the founder cell nuclei.
Note that there is still some punctate staining of Loner.
of ANTS (red in O and Q) and Loner (green in P and Q). Muscle cells
in six hemisegments during the germ band extension stage are
shown. Note that Loner colocalized with ANTS at some foci (yellow),
but not the others.
Figure 5. Loner and ANTS Are Recruited to
the Cell Membrane by Founder Cell-Specific
Receptors, DUF and RST, in Transfected S2
Transgenes expressed are indicated above
(A, D, G, and J) Localization of DUF (A), RST
(D), ANTS (G), and Loner (J) in S2 cells when
they were expressed alone. Cells were
stained with anti-V5 to visualize V5-tagged
DUF (red) and RST (red), anti-FLAG for FLAG-
tagged ANTS (green). Note that DUF and RST
localized at the membrane region between
two adhering cells, whereas ANTS and Loner
localized in the cytoplasm of two randomly
(B and C) Localization of DUF (B) and Loner
(C) when they were coexpressed.
(E and F) Localization of RST (E) and Loner
(F) when they were coexpressed. Note that
brane region between two adhering cells.
(H and I) Localization of DUF (H) and ANTS
(I) when they were coexpressed. Note that
ANTS, like Loner, was recruited to the cell
membrane between two adhering cells.
(K) Localization of Delta (red) and Loner
(green) when they were coexpressed in
S2 cells. Note that Loner still remained cytoplasmic, even though Delta localized at the membrane of the two adhering cells.
(L) Loner (green) and ANTS (red) do not colocalize when coexpressed in S2 cells.
plasm. On the other hand, Loner localization was not
affected in embryos lacking the transmembrane protein
SNS, expressed specifically in fusion-competent myo-
blasts (Figure 4L). These results are consistent with the
recruitment of Loner by DUF and RST to discrete loci
at the founder cell membrane in vivo.
dependent on the fusion receptors DUF/RST, suggest
that DUF/RST recruit ANTS and Loner independently to
the subcellular sites of fusion and that Loner and ANTS
might function in parallel downstream of DUF and RST
during myoblast fusion (see Figure 7).
The GEF Activity of Loner Is Required
for Its Function In Vivo
To determine the functional significance of the con-
ties of a series of mutant Loner proteins. Wild-type
Loner, but not a Sec7-deletion mutant, or a “GEF-dead”
Loner mutant containing an E-to-K point mutation in the
conserved GEF domain (Shevell et al., 1994), rescued
the loner mutant phenotype (Figure 3B, b, e, and f).
Using the same rescue assays, we found that deletion
of the PH domain also abolished the ability of Loner to
rescue the myoblast fusion phenotype (Figure 3B, g),
whereas deletion of the IQ motif did not affect Loner
activity (Figure 3B, h). Taken together, these results
show that the Sec7 and PH domains, but not the IQ
domain, are essential for Loner function in the context
of myoblast fusion.
The Subcellular Localization of Loner
Is Independent of ANTS
It is intriguing that Loner and ANTS share several com-
mon features with respect to their intracellular localiza-
tion: both are expressed in discrete foci in founder cells;
both can be recruited to the cell membrane by DUF and
RST in S2 cells; and the subcellular localization of both
is altered in duf rst mutant embryos. However, they ap-
pear to be in different foci when coexpressed in S2 cells
(Figure 5L). This prompted us to investigate if Loner and
cells in vivo. Antibody double-labeling experiments re-
vealed that Loner and ANTS partially colocalized in
founder cells, with some foci containing both proteins
and others containing either one or the other of the two
proteins (Figures 4O–4Q). Given that the ANTS-positive
foci represent fusion sites (Menon and Chia, 2001), this
result suggests that at least some of the Loner-positive
foci also correspond to sites of muscle cell fusion.
To investigate if the localization of Loner and ANTS
of ANTS in loner mutant embryos. As shown in Figures
4I and 4N, Loner and ANTS maintain their pattern of
subcellular localization in the absence of the other pro-
tein. These results, combined with observations that
the subcellular localization of ANTS and Loner are both
Loner Has Specific Activity toward the Small
GTPase ARF6 In Vitro
The Sec7 family of GEFs regulates the ARF family of
small GTPases. Mammalian ARFs can be divided into
class I (ARF1-3), class II (ARF4, 5), and class III (ARF6).
membrane recycling, and cytoskeletal rearrangement
(Chavrier and Goud, 1999; Donaldson and Jackson,
2000). Several low molecular weight Sec7 GEFs related
to Loner, such as cytohesins, EFA6, as well as ARF-
Loner, a Unique GEF Required for Myoblast Fusion
Figure 6. GTPase Target of Loner and Its
Involvement in Drosophila and Mammalian
(A) Loner GEF stimulates guanine nucleotide
release on dARF6, but not dARF1. Activity
of 2 ug of GST alone, GST-Sec7, and GST-
sion proteins of Drosophila ARF family
GTPases, dARF6 and dARF1, respectively.
The activity is expressed as the percent of
initial3H-GDP remaining bound after 20 min
(B) Somatic musculature of a stage 14 wild-
the rP298-GAL4 driver and UAS-dARF6T27N
transgene, visualized by a MHC-tauGFP re-
porter. Note that there are many unfused
myoblasts in the latter embryo.
(C) Localization of Rac protein (green) in
The nuclei of a subset of founder cells were
of Rac protein aggregates are observed in
discrete foci, a few of which are indicated by
arrows, along the founder cell membrane in
wild-type, but not loner mutant, embryos.
(D) Regulation of myogenesis by ARF6.
10T1/2 fibroblasts were cotransfected with
expression vectors for MyoD and pcDNA as
blasts converted to differentiated myotubes
or myocytes were marked by anti-myosin an-
(E) The number of myosin-positive myotubes
(black bar) and mononucleated myocytes
(white bar) in cultures cotransfected of MyoD
with pcDNA, ARF6, and ARF6T27N, respec-
tively. The myotubes counted here contained
at least three fused muscle cells. Values rep-
resent the mean ? SD from at least three ex-
3H-GDP release from GST fu-
et al., 1998; Franco et al., 1999; Langille et al., 1999;
Someya et al., 2001), suggesting that ARF6 may be a
downstream small GTPase for Loner. Searches of Dro-
an apparent ARF6 homolog, Arf51F, which shares 97%
amino acid identity with mammalian ARF6. For simplic-
ity, we will refer to Arf51F as dARF6. To test if Loner
functions as a GEF for dARF6, we carried out in vitro
GDP release assays using GST fusion proteins con-
taining dARF6 and GST fusion proteins containing the
wild-type Sec7 domain of Loner or a Sec7 domain with
an E-to-K mutation that is known to abolish GEF activity
(Shevell et al., 1994). As shown in Figure 6A, the wild-
type Sec7 domain, but not the GEF-dead Sec7 domain,
catalyzed GDP/GTP exchange of dARF6. This activity
was highly specific, since no significant GDP release
was detected using the highly related GTPase dARF1
onstrate that Loner functions as a specific dARF6 GEF
whether dARF6 might also be involved in the same pro-
cess. In situ hybridization showed that dARF6 is ubiqui-
tously expressed in the embryo (data not shown). Since
no loss-of-function mutant of dARF6 exists, we engi-
neered transgenic flies carrying a dominant-negative
form of dARF6 (dARF6T27N) (D’Souza-Schorey et al.,
1998) under control of the UAS promoter. Expression of
the mutant ARF6 in founder cells severely perturbed
myoblast fusion throughout the internal layer of somatic
at their characteristic positions (Figure 6B). This pheno-
type is similar to, but less severe than that of loner
mutant embryos. The weaker phenotype resulting from
dARF6T27Nexpression could be due to incomplete inhibi-
the relatively late expression of the rP298-GAL4 line,
which does not start until stage 12, when myoblast fu-
sion has already initiated for many muscle fibers. Taken
together, our results suggest that ARF6 is an essential
downstream mediator of Loner activity during myoblast
fusion in the Drosophila embryo.
The Loner/ARF6 Module Is Required for the
Subcellular Localization of Rac
To further understand the function of Loner/ARF6 in
myoblast fusion, we investigated whether the Loner/
dARF6 Is Involved in Myoblast Fusion
The biochemical link between Loner and dARF6, com-
bined with the essential requirement for the GEF activity
of Loner in myoblast fusion, prompted us to investigate
ARF6 module impinged upon the ANTS→MBC→Rac
pathway. Previous studies in mammalian cultured cells
have suggested that ARF6 is involved in localizing Rac
to the plasma membrane, a prerequisite for Rac’s func-
tion in cytoskeletal rearrangement (Radhakrishna et al.,
1999). Since Rac is also required for myoblast fusion
in Drosophila, we tested if the Loner/ARF6 module is
required for the proper localization of Rac in founder
cells. In wild-type embryos, high levels of Rac protein
brane, which correspond to sites of fusion (Figure 6C
the specific aggregation of Rac at the fusion sites was
no longerobserved (Figure 6C).These datasuggest that
the Loner/ARF6 module converges with the ANTS→
MBC→Rac pathway at the small GTPase level, and
Loner/ARF6 are required, independent of ANTS, for the
proper subcellular localization of Rac in founder cells.
nova, 2000). Despite extensive molecular and biochemi-
cal characterization of ARF-GEFs in yeast and mamma-
are largely unknown. The only exception, to our knowl-
edge, is GNOM/Emb30, a high molecular weight GEF
specific for ARF1 that is required for the establishment
and maintenance of cell polarity in Arabidopsis (Shevell
et al., 1994; Geldner et al., 2003). At present, none of the
low molecular weight ARF-GEFs has been implicated in
a physiological process in vivo.
The Loner protein contains a Sec7 domain, a PH do-
mainand acoiled-coildomain. Sucha domainorganiza-
tion is reminiscent of low molecular weight ARF-GEFs.
However, Loner is distinguished from conventional low
molecular weight ARF-GEFs by its high molecular
the presence of an IQ motif at the N terminus. Therefore,
Loner and its mammalian homologs define a distinct
subclass of ARF-GEFs. Our studies have revealed a
physiological function for this subclass of ARF-GEFs
and have provided insights into their structure-func-
The Sec7 domain of Loner has specific GEF activity
toward Drosophila dARF6, a finding consistent with bio-
chemical studies of ARF-GEP100, a human homolog of
of Loner is required for myoblast fusion, since deletion
of the Sec7 domain or a “GEF-dead” version of the
Sec7 domain completely abolished the ability of Loner
to rescue the fusion defects in loner mutant embryos.
The PH domain is also required for Loner function, as
revealed by the loss of fusigenic activity of a Loner
deletion mutant lacking this region. The PH domain of
the low molecular weight ARF-GEFs such as cytohesins
has been implicated in phosphoinositide binding and
targeting to the plasma membrane (Chardin et al., 1996;
Paris et al., 1997).
Dominant-Negative ARF6 Disrupts Mammalian
To determine if a similar ARF6-mediated pathway might
control fusion of mammalian muscle cells, we investi-
gated whether expression of a dominant-negative form
of ARF6,ARF6T27N, couldinterfere withmuscle differenti-
ation in vitro, which is dependent on MyoD (Davis et al.,
1987). When 10T1/2 fibroblasts were transfected with
fected cells acquired a myoblast fate and differentiated
into myotubes (Figures 6D, a and 6E). Coexpression of
MyoD with wild-type ARF6 did not significantly affect
the differentiation of 10T1/2 cells into myotubes (Figure
6D, b and 6E). However, coexpression of ARF6T27Nwith
MyoD severely decreased the efficiency of myotube for-
mation, whereas myocyte differentiation was not af-
fected (Figures 6D, c and 6E). Only occasionally could
myotubes be observed in cells coexpressing MyoD and
ARF6T27N. This phenotype, in which myosin heavy chain
expression is induced but fusion is specifically blocked,
is analogous to the phenotype of loner mutant embryos
or embryos expressing dARF6T27N. Thus, ARF6 appears
play an important role in mammalian myoblast fusion,
as in Drosophila.
The Role of Loner and ARF6 in Myoblast Fusion
Myoblast fusion requires the initial recognition and ad-
hesion between founder cells and fusion-competent
myoblasts, followed by cytoskeleton rearrangements
that lead to the proper alignment of the two populations
of cells and eventual membrane coalescence. A path-
way involving ANTS and MBC has been proposed to
transduce fusion signals from the founder cell-specific
surface receptors DUF/RST to the small GTPase, Rac,
which controls actin cytoskeleton rearrangement (Hall,
1998; Chen and Olson, 2001). Our current studies have
revealed additional components of the cellular appara-
tus downstream of the fusion receptors and suggested
blast fusion sites. The small GTPase, ARF6, and its GEF,
Loner, act as essential effectors in this cellular appara-
tus (Figure 7).
Several lines of evidence suggest that Loner acts
downstream of the founder cell-specific receptors, DUF
and RST. First, Loner is recruited to the cell membrane
by DUF/RST in cell-culture assays, mimicking the be-
havior of ANTS, which acts downstream of DUF/RST
(Chen and Olson, 2001). Second, loner encodes a cyto-
plasmic protein that is expressed at sites of fusion in
opment in Drosophila, we discovered Loner, a GEF
required for myoblast fusion. Our results ascribe four
interdependent functions to Loner: (1) it acts as a down-
stream effector of myoblast fusion receptors; (2) it re-
cruits ARF6 to subcellular sites of fusion; (3) it promotes
guanine nucleotide exchange by ARF6; and (4) it im-
pinges on the ANTS→MBC→Rac pathway at the small
GTPase level. These findings establish Loner and ARF6
as key components of a cellular apparatus governing
myoblast fusion and suggest the involvement of ARF-
GEF signaling in intercellular fusion of other cell types.
Loner, a Member of the ARF-GEF Family
The ARF-GEFs constitute a large and diverse family of
proteins (Moss and Vaughan, 1998; Chavrier and Goud,
Loner, a Unique GEF Required for Myoblast Fusion
Figure 7. Model for Loner and ARF6 in My-
We propose that Loner is recruited to the cell
membrane by founder cell-specific recep-
tors, DUF/RST, likely through interaction with
protein “fusion complex,” including ANTS
and Loner, at the sites of fusion. The mem-
brane recruitment of Loner is independent of
the adaptor protein ANTS. Aggregation of
Loner at the sites of fusion in turn recruits
and activates ARF6 via its Sec7 domain. The
Loner/ARF6 module impinges on the ANTS→
MBC→Rac pathway at the small GTPase
level by controlling the proper membrane lo-
calization of Rac, a prerequisite for actin cy-
sion. Activated forms of ARF6 and Rac are
marked as ARF6* and Rac*, respectively.
set of Loner is altered in embryos lacking DUF/RST.
Lastly, in duf rst mutant embryos, myoblast fusion is
blocked at the initial recognition and adhesion between
fusion-competent myoblasts and their targets (Ruiz-
Gomez et al., 2000), while myoblast fusion is blocked
at a laterstep in loner mutant embryos(this study). Such
a temporal order is consistent with Loner functioning
downstream of DUF/RST.
We showed previously that the founder cell-specific
adaptor protein ANTS localized to fusion sites through
direct physical interaction with the fusion receptors,
DUF and RST (Chen and Olson, 2001; data not shown).
We demonstrate here that Loner is recruited by DUF/
RST to the cell membrane independent of ANTS. Thus,
the formation of the “fusion complex” is initially orga-
nized by the founder cell receptors. Our studies also
suggest that Loner functions in parallel to ANTS, since
Loner and ANTS are recruited independently to the cell
membrane by DUF/RST and the localization of ANTS
and Loner is independent of each other.
We propose that a major function of Loner is to recruit
its downstream GTPase, ARF6, to fusion sites defined
by DUF/RST. Given the role of ARF6 in cytoskeleton
organization (Chavrier and Goud, 1999), activation of
essential forcytoskeleton rearrangement,a prerequisite
for proper cell alignment in the fusion process.
zymes responsible for lipid modification, such as PIP5-
kinase and PLD (Brown et al., 2001), or in regulated
secretion events that it has been associated with in
mammalian systems (D’Souza-Schorey et al., 1998).
Implications Beyond Drosophila Myogenesis
Given the evolutionary conservation of muscle develop-
mental control mechanisms (Wakelam, 1985; Knudsen,
1992; Doberstein et al., 1997), it is likely that homologs
of genes involved in Drosophila myoblast fusion play
similar roles in mammalian skeletal muscle develop-
ment.Inthis regard,weshowedpreviously thatamouse
homolog of ants is expressed in the embryonic meso-
derm at the time of myoblast fusion, suggesting its po-
tential involvement in myogenesis (Chen and Olson,
2001). Here, we show that ARF6 is involved in mamma-
lian myoblast fusion, suggesting that the Loner/ARF6
module may play conserved roles in evolution.
Cell fusion is a universal and evolutionarily ancient
isms. Despite the diversity of cell types that undergo
cell-cell fusion (e.g., sperm-egg, osteoclasts, hemato-
poietic stem cells, muscle cells), the cellular events in-
volved in this process—cell recognition, adhesion, and
membrane merger—are common to all these cell types,
suggesting shared cellular mechanisms. In Drosophila
muscle cells, Loner and ARF6 are controlled by the cell
surface fusion receptors, DUF and RST. However,
Loner/ARF6 are expressed in a wide range of cell types.
Thus, they may represent a general fusigenic mecha-
nism coopted by different upstream effectors to control
intercellular fusion of diverse cell types. The recognition
that intercellular fusion is controlled by a G protein-
dependent mechanism involving Loner and ARF6 pro-
vides interesting opportunities for modulating this pro-
cess in a variety of therapeutic settings.
Small GTPases in Myoblast Fusion
It is intriguing that two small GTPases, ARF6 and Rac,
are both required for myoblast fusion. Our data suggest
that the Loner/ARF6 module impinges on the ANTS→
MBC→Rac pathway at the small GTPase level, and
Loner/ARF6 are required for the proper localization of
actin cytoskeleton rearrangements required for fusion.
cells demonstrating strong interactions between ARF6
and Rac in regulating the cortical actin cytoskeleton.
For example, ARF6 and Rac can bind to the same pro-
tein, POR-1, and ARF6 is required for the membrane
localization of Rac (D’Souza-Schorey et al., 1997; Rad-
hakrishna et al., 1999; Boshans et al., 2000). However,
for example, in its demonstrated activity toward en-
Mutations disrupting myoblast fusion were isolated in a genetic
screen for muscle development (E.H.C. and E.N.O., unpublished
data). loner mutants were initially defined by a lethal complementa-
tion group of two EMS-induced alleles, lonerT1032and lonerT1057. The
third chromosome deficiency kit, DK3, was used in complementa-
tion tests to map these mutants to the 77F3 to 78C8-9 region, de-
fined by Df(3L)ME107. Subsequently, the loner mutants were
mapped by overlapping small deficiencies in this region, kindly pro-
vided by Minx Fuller. Df(3L)Pc-MK (78A2;78C9) did not complement
either loner allele, whereas Df(3L)ri-XT1 (77E2; 78A4) and Df(3L)Pc-
cp2 (78B1-2; 78D) complemented both loner alleles, delimiting the
loner locus to 78A4-B1. Both alleles result in identical phenotypes
as homozygous embryos or transheterozygous embryos over defi-
ciency Df(3L)ME107. Therefore, we infer that they both behave as
The rP298-lacZ stock was generously provided by Akinao Nose,
the rP298-GAL4 driver by Devi Menon, the sns40-49/CyO by Renate
Renkawitz-Pohl. Df(1)w67k30, a small deficiency deleting both duf and
rst, N81k, and twi-GAL4 were obtained from the Bloomington stock
In overexpression studies using the rP298-GAL4 driver, males
carrying rP298-GAL4 were crossed with females carrying various
DMEM with 10% fetal bovine serum (FBS). Transfections were con-
ducted with Lipofectamine 2000 (Invitrogen) according to manufac-
turer’s instructions. 0.4 ug of plasmid was used for each well in a
12-well plate. Two days after transfection, cells were shifted to a
differentiation medium (DMEM with 2% horse serum). Cells were
subjected to immunocytochemistry five days after differentiation.
In situ hybridization of Drosophila embryos was performed using
standard protocols (Tautz and Pfeifle, 1989). DIG-labeled probe was
synthesized with the common coding region of the three isoforms
Rat polyclonal Loner antisera were generated against a carboxy-
terminal peptide RIPGRERKASRTDENGRS (Bio-Synthesis) and
used at 1:300 in combination with the TSA fluorescence system
(NEN Life Sciences). Embryo-staining procedures were performed
as described(Patel, 1994),using the followingadditional antibodies:
rabbit anti-MHC (1:1000) and mouse anti-MHC (1:10) (Kiehart and
Feghali, 1986); rabbit anti-DMEF2 (1:800) (Nguyen et al. 1994); rabbit
anti-KR (1:3000) (Gaul et al., 1987); rabbit anti-NAU (1:800) (Keller et
al., 1997); mouse anti-Rac1 (1:300) (BD Transduction Laboratories);
rabbit anti-?-gal (1:1500) (Cappel); and mouse anti-?-gal (1:1000)
(Promega). Secondary antibodies used were: Cy3 goat anti-rabbit
(1:300) (Jackson) and biotinylated antibodies made in goat (1:300)
S2 cells were fixed with 4% paraformaldehyde and stained with
the following primary antibodies: mouse anti-V5 (1:1000) (In-
vitrogen); rabbit anti-MYC (1:300) (Santa Cruz); rabbit anti-FLAG
(1:500) (Sigma); and mouse anti-Delta (1:20) (Developmental Studies
Hybridoma Bank). Secondary fluorochrome-conjugated antibodies
were used at 1:200 (Jackson).
10T1/2 cell myogenic conversion assays were performed as de-
scribed (Lu et al., 2000). Mouse antiskeletal myosin (MY32) (1:400)
(Sigma) was used to stain the differentiated skeletal muscle cells.
Fluorescent images were collected on a LSM410 Zeiss confocal
microscope and were processed with Adobe Photoshop 7.
Full-length EST clones RE02556, LP01489, and GH10594 (Berkeley
Drosophila Genome Project [BDGP]) were obtained from Re-
DNA sequences of loner alleles were determined by directly se-
quencing PCRproducts amplifiedfrom genomic DNAobtained from
homozygous mutant embryos selected by their lack of armadillo-
GFP expression, which was carried on the balancer chromosome.
When a mutation was uncovered, PCR and sequencing were re-
peated to confirm that the mutation was not due to PCR errors.
loner transgenes were prepared using standard subcloning pro-
cedures. For rescue constructs with full-length loner cDNAs, EcoRI
fragments including loner iso1, iso2, or iso3 from their respective
EST clones weresubcloned into transformation vectors,pUAST and
S102 (containing a tubulin promoter), respectively. loner deletion
and point-mutation constructs were prepared using standard PCR
procedures to introduce the necessary changes on their original
EST clones (Stratagene) and subcloned into pUAST and S102, re-
dARF6 (Arf51F) transgenes were prepared by amplifying the
dARF6 coding sequence from total embryonic cDNA. Standard PCR
procedures were used to introduce a T27→N point mutation (Stra-
tagene). Wild-type and mutant dARF6 were then subcloned into the
For bacterial expression, the coding sequences for Loner Sec7,
Loner Sec7E→K, dARF6, and dARF1 were amplified from the loner
iso1, loner iso1E→K, or total embryonic cDNA, and cloned in frame
into pGEX-2T (Pharmacia).
Constructs usedfor S2 celltransfection were preparedas follows.
FLAG-tagged Loner: loner iso1, iso2, and iso3 were amplified by
PCR with the 5? primer containing sequences encoding a FLAG tag.
They were then subcloned into the pAc-V5 His expression vector
(Invitrogen). V5-tagged RST: the full-length coding regions of rst
were amplified by PCR from embryonic cDNA and cloned in-frame
into the pAc-V5 His vector.
Constructs used for 10T1/2 cell transfection were prepared as
follows. The ARF6 and ARF6T27NcDNAs, generously provided by
Michael Roth, were amplified by PCR with the 3? primer containing
sequences encoding a MYC tag. They were then subcloned into the
pcDNA expression vector (Invitrogen).
All constructs were verified by sequence analysis.
We thank Drs. Susan Abmayr, Dan Kiehart, Hanh Nguyen, Michael
Roth, and the Developmental Studies Hybridoma Bank for cDNAs
and antibodies; Minx Fuller, Devi Menon, Akinao Nose, Renate Ren-
kawitz-Pohl and the Bloomington Stock Center for fly stocks. We
thank Alisha Tizenor for graphics and Jennifer Page for editorial
assistance. We appreciate Dennis McKearin, Duojia Pan, and Keith
Wharton for insightful discussions and critical reading of the manu-
script. E.H.C. was supported by a postdoctoral fellowship from the
Helen Hay Whitney Foundation. This work was supported by grants
from the National Institutes of Health and the D.W. Reynolds Center
for Clinical Cardiovascular Research to E.N.O.
Received: June 9, 2003
Revised: August 25, 2003
Accepted: August 25, 2003
Published: September 18, 2003
Adams, M.D., Celniker, S.E., Holt, R.A., Evans, C.A., Gocayne, J.D.,
Amanatides, P.G., Scherer, S.E., Li, P.W., Hoskins, R.A., Galle, R.F.,
et al. (2000). The genome sequence of Drosophila melanogaster.
Science 287, 2185–2195.
Artero, R.D., Castanon, I., and Baylies, M.K. (2001). The immuno-
globulin-like protein Hibris functions as a dose-dependent regulator
of myoblast fusion and is differentially controlled by Ras and Notch
signaling. Development 128, 4251–4264.
Bate, M. (1993). The mesoderm and its derivatives. In The develop-
ment of Drosophila melanogaster, M. Bate and A. Martinez Arias,
eds. (Cold Spring Harbor, NY: CSH Laboratory Press), pp. 1013–
Baylies, M.K., and Michelson, A.M. (2001). Invertebrate myogenesis:
Dev. 11, 431–439.
GST fusion proteins were prepared according to standard protocols
(Pharmacia). Guanine nucleotide exchange assays were performed
using bacterially expressed GST fusion proteins as described in
Debant et al. (1996). Each assay was performed in triplicate.
Cell Culture and Transfection
Drosophila S2 cells were transfected using the Effectene transfec-
tion reagent (Qiagen) according to manufacturer’s instructions. In
the case of Hm-Delta, expression of Delta was induced with 0.7 mM
CuSO4a day after the initial transfection. Cells were subjected to
immunocytochemistry two days after transfection.
Mouse 10T1/2 fibroblasts were maintained at low density in
Loner, a Unique GEF Required for Myoblast Fusion
Baylies, M.K., Bate, M., and Ruiz Gomez, M. (1998). Myogenesis: a
view from Drosophila. Cell 93, 921–927.
Blau, H.M. (2002). A twist of fate. Nature 419, 437.
Blumenthal, R., Clague, M.J., Durell, S.R., and Epand, R.M. (2003).
Membrane fusion. Chem. Rev. 103, 53–69.
Boshans, R.L., Szanto, S., van Aelst, L., and D’Souza-Schorey, C.
(2000). ADP-ribosylation factor 6 regulates actin cytoskeleton re-
modeling in coordination with Rac1 and RhoA. Mol. Cell. Biol. 20,
Bour, B.A., Chakravarti, M., West, J.M., and Abmayr, S.M. (2000).
Drosophila SNS, a member of the immunoglobulin superfamily that
is essential for myoblast fusion. Genes Dev. 14, 1498–1511.
Brown, F.D., Rozelle, A.L., Yin, H.L., Balla, T., and Donaldson, J.G.
(2001). Phosphatidylinositol 4,5-bisphosphate and Arf6-regulated
membrane traffic. J. Cell Biol. 154, 1007–1017.
Brugnera, E., Haney, L., Grimsley, C., Lu, M., Walk, S.F., Tosello-
dran, K.S. (2002). Unconventional Rac-GEF activity is mediated
through the Dock180-ELMO complex. Nat. Cell Biol. 4, 574–582.
Chardin, P., Paris, S., Antonny, B., Robineau, S., Beraud-Dufour, S.,
Jackson, C.L., and Chabre, M. (1996). A human exchange factor
for ARF contains Sec7- and pleckstrin-homology domains. Nature
Chavrier, P., and Goud, B. (1999). The role of ARF and Rab GTPases
in membrane transport. Curr. Opin. Cell Biol. 11, 466–475.
Chen, E.H., and Olson, E.N. (2001). Antisocial, an intracellular adap-
tor protein, is required for myoblast fusion in Drosophila. Dev. Cell
Corbin, V., Michelson, A.M., Abmayr, S.M., Neel, V., Alcamo, E.,
Maniatis, T., and Young, M.W. (1991). A role for the Drosophila
neurogenic genes in mesoderm differentiation. Cell 67, 311–323.
Debant, A., Serra-Pages, C., Seipel, K., O’Brien, S., Tang, M., Park,
S.H., and Streuli, M. (1996). The multidomain protein Trio binds
the LAR transmembrane tyrosine phosphatase, contains a protein
kinase domain, and has separate rac-specific and rho-specific gua-
nine nucleotide exchange factor domains. Proc. Natl. Acad. Sci.
USA 93, 5466–5471.
D’Souza-Schorey, C., Boshans, R.L., McDonough, M., Stahl, P.D.,
and Van Aelst, L. (1997). A role for POR1, a Rac1-interacting protein,
in ARF6-mediated cytoskeletal rearrangements. EMBO J. 16, 5445–
D’Souza-Schorey, C., van Donselaar, E., Hsu, V.W., Yang, C., Stahl,
P.D., and Peters, P.J. (1998). ARF6 targets recycling vesicles to the
plasma membrane: insights from an ultrastructural investigation. J.
Cell Biol. 140, 603–616.
Davis, R.L., Weintraub, H., and Lassar, A.B. (1987). Expression of
a single transfected cDNA converts fibroblasts to myoblasts. Cell
Doberstein, S.K., Fetter, R.D., Mehta, A.Y., and Goodman, C.S.
(1997). Genetic analysis of myoblast fusion: blown fuse is required
for progression beyond the prefusion complex. J. Cell Biol. 136,
Donaldson, J.G., and Jackson, C.L. (2000). Regulators and effectors
of the ARF GTPases. Curr. Opin. Cell Biol. 12, 475–482.
Dworak, H.A., and Sink, H. (2002). Myoblast fusion in Drosophila.
Bioessays 24, 591–601.
Dworak, H.A., Charles, M.A., Pellerano, L.B., and Sink, H. (2001).
Characterization of Drosophila hibris, a gene related to human
nephrin. Development 128, 4265–4276.
Erickson, M.R., Galletta, B.J., and Abmayr, S.M. (1997). Drosophila
myoblast city encodes a conserved protein that is essential for
myoblast fusion, dorsal closure, and cytoskeletal organization. J.
Cell Biol. 138, 589–603.
Franco, M., Peters, P.J., Boretto, J., van Donselaar, E., Neri, A.,
D’Souza-Schorey, C., and Chavrier, P.(1999). EFA6, a Sec7 domain-
containing exchange factor for ARF6, coordinates membrane recy-
cling and actin cytoskeleton organization. EMBO J. 18, 1480–1491.
Frank, S., Upender, S., Hansen, S.H., and Casanova, J.E. (1998).
ARNO is a guanine nucleotide exchange factor for ADP-ribosylation
factor 6. J. Biol. Chem. 273, 23–27.
Frasch, M. (1999). Controls in patterning and diversification of so-
Dev. 9, 522–603.
partnerships in Drosophila myoblast fusion. Cell 102, 127–129.
Fuerstenberg, S., and Giniger, E. (1998). Multiple roles for Notch in
Drosophila myogenesis. Dev. Biol. 201, 66–77.
Gaul, U., Seifert, E., Schuh, R., and Jackle, H. (1987). Analysis of
Kruppel protein distribution during early Drosophila development
reveals posttranscriptional regulation. Cell 50, 639–647.
Geldner, N., Anders, N., Wolters, H., Keicher, J., Kornberger, W.,
Muller, P., Delbarre,A., Ueda, T., Nakano, A., andJurgens, G. (2003).
The Arabidopsis GNOM ARF-GEF mediates endosomal recycling,
auxin transport, and auxin-dependent plant growth. Cell 112,
Grenningloh, G., Rehm, E.J., and Goodman, C.S. (1991). Genetic
as a neuronal recognition molecule. Cell 67, 45–57.
Hakeda-Suzuki, S., Ng, J., Tzu, J., Dietzl, G., Sun, Y., Harms, M.,
Nardine, T., Luo, L., and Dickson, B.J. (2002). Rac function and
regulation during Drosophila development. Nature 416, 438–442.
Hall, A. (1998). Rho GTPases and the actin cytoskeleton. Science
Hernandez, L.D., Hoffman, L.R., Wolfsberg, T.G., and White, J.M.
(1996). Virus-cell and cell-cell fusion. Annu. Rev. Cell Dev. Biol.
Jackson, C.L., and Casanova, J.E. (2000). Turning on ARF: the Sec7
family of guanine-nucleotide-exchange factors. Trends Cell Biol.
Keller,C.A.,Erickson, M.S.,andAbmayr,S.M. (1997).Misexpression
of somatic muscle fibers. Dev. Biol. 181, 197–212.
Kiehart, D.P., and Feghali, R. (1986). Cytoplasmic myosin from Dro-
sophila melanogaster. J. Cell Biol. 103, 1517–1525.
Knudsen, K.A. (1992). Fusion of myoblasts. In Membrane Fusion. J.
Wilschut and D. Hoekstra, eds. (New York: Marcel Decker, Inc.),
Langille, S.E., Patki, V., Klarlund, J.K., Buxton, J.M., Holik, J.J.,
Chawla, A., Corvera, S., and Czech, M.P. (1999). ADP-ribosylation
factor 6 as a target of guanine nucleotide exchange factor GRP1.
J. Biol. Chem. 274, 27099–27104.
Lilly, B., Galewsky, S., Firulli, A.B., Schulz, R.A., and Olson, E.N.
entiating mesoderm and musclecell lineages during Drosophila em-
bryogenesis. Proc. Natl. Acad. Sci. USA 91, 5662–5666.
Lu, J., McKinsey, T.A., Zhang, C.L., and Olson, E.N. (2000). Regula-
tionof skeletalmyogenesisby associationofthe MEF2transcription
factor with class II histone deacetylases. Mol. Cell 6, 233–244.
Menon, S.D., and Chia, W. (2001). Drosophila rolling pebbles: a
multidomain protein required for myoblast fusion that recruits
D-Titin in response to the myoblast attractant Dumbfounded. Dev.
Cell 1, 691–703.
sion of a cDNA encoding a bovine brain brefeldin A-sensitive gua-
nine nucleotide-exchange protein for ADP-ribosylation factor. Proc.
Natl. Acad. Sci. USA 94, 12926–12931.
Moss, J., and Vaughan, M. (1998). Molecules in the ARF orbit. J.
Biol. Chem. 273, 21431–21434.
Nguyen, H.T., Bodmer, R., Abmayr, S.M., McDermott, J.C., and
Spoerel, N.A. (1994). D-mef2: a Drosophila mesoderm-specific
embryogenesis. Proc. Natl. Acad. Sci. USA 91, 7520–7524.
Paris, S., Beraud-Dufour, S., Robineau, S., Bigay, J., Antonny, B.,
Chabre, M., and Chardin, P. (1997). Role of protein-phospholipid
interactions in the activation of ARF1 by the guanine nucleotide
exchange factor Arno. J. Biol. Chem. 272, 22221–22226.
Patel, N.H. (1994). Imaging neuronal subsets and other cell types in
whole-mount Drosophila embryos and larvae using antibody
probes. Methods Cell Biol. 44, 445–487.
for myoblast fusion in Drosophila embryogenesis. Mech. Dev. 83,
Radhakrishna, H., Al-Awar, O., Khachikian, Z., and Donaldson, J.G.
(1999). ARF6 requirement for Rac ruffling suggests a role for mem-
brane trafficking in cortical actin rearrangements. J. Cell Sci. 112,
Rau,A., Buttgereit,D., Holz,A.,Fetter, R.,Doberstein, S.K.,Paululat,
A., Staudt, N., Skeath, J., Michelson, A.M., and Renkawitz-Pohl,
R. (2001). rolling pebbles (rols) is required in Drosophila muscle
precursors for recruitment of myoblasts for fusion. Development
ulin recognition. FASEB J. 11, 331–340.
Ruiz-Gomez, M., Romani, S., Hartmann, C., Jackle, H., and Bate,
M.(1997). Specificmuscle identitiesare regulatedby Kruppelduring
Drosophila embryogenesis. Development 124, 3407–3414.
Ruiz-Gomez, M., Coutts, N., Price, A., Taylor, M.V., and Bate, M.
(2000). Drosophila dumbfounded: a myoblast attractant essential
for fusion. Cell 102, 189–198.
Rusconi, J.C., and Corbin, V. (1998). Evidence for a novel Notch
pathway required for muscle precursor selection in Drosophila.
Mech. Dev. 79, 39–50.
Shevell, D.E., Leu, W.M., Gillmor, C.S., Xia, G., Feldmann, K.A., and
Chua, N.H. (1994). EMB30 is essential for normal cell division, cell
expansion, and cell adhesion in Arabidopsis and encodes a protein
that has similarity to Sec7. Cell 77, 1051–1062.
Someya, A., Sata, M., Takeda, K., Pacheco-Rodriguez, G., Ferrans,
V.J., Moss, J., and Vaughan, M. (2001). ARF-GEP(100), a guanine
nucleotide-exchange protein for ADP-ribosylation factor 6. Proc.
Natl. Acad. Sci. USA 98, 2413–2418.
Strunkelnberg, M., Bonengel, B., Moda, L.M., Hertenstein, A., de
Couet, H.G., Ramos, R.G., and Fischbach, K.F. (2001). rst and its
paralogue kirre act redundantly during embryonic muscle develop-
ment in Drosophila. Development 128, 4229–4239.
Tautz, D., and Pfeifle, C. (1989). A non-radioactive in situ hybridiza-
tion method for the localization of specific RNAs in Drosophila em-
bryos reveals translational control of the segmentation gene hunch-
back. Chromosoma 98, 81–85.
Taylor, M.V. (2002). Muscle differentiation: how two cells become
one. Curr. Biol. 12, R224–R228.
Wakelam, M.J. (1985). The fusion of myoblasts. Biochem. J. 228,
The GenBank accession numbers for the three loner isoforms re-
ported in this paper are AY375487, AY375488, and AY375489.