Mosaic analysis with a repressible cell marker (MARCM) for Drosophila neural development

Mosaic analysis with a repressible cell marker (MARCM) for Drosophila neural development
Research Update
TRENDS in Neurosciences Vol.24 No.5 May 2001
Techniques & Applications
Mosaic analysis with a repressible cell marker (MARCM)
for Drosophila neural development
Tzumin Lee and Liqun Luo
We have modified an FLP/FRT-based
genetic mosaic system to label either
neurons derived from a common progenitor
or isolated single neurons, in the Drosophila
CNS. These uniquely labeled neurons can
also be made homozygous for a mutation of
interest within an otherwise phenotypically
wild-type brain. Using this new mosaic
system, not only can normal brain
development be described with
unprecedented single cell resolution, but
also the underlying molecular mechanisms
can be investigated by identifying genes
that are required for these developmental
The ‘MARCM’ system for performing
mosaic analysis in the CNS
Mosaic analysis involves the generation
of homozygous mutant cells from
heterozygous precursors via mitotic
recombination (reviewed in Ref. 4). For
most genes, one wild-type allele in a
diploid cell is sufficient for normal
function. Therefore, by creating
homozygous mutant cells in
heterozygous tissues, one can knock out
gene function in a small subset of cells
and then examine their phenotypes in
an otherwise phenotypically wild-type
organism. Such analysis is particularly
Since Ramon y Cajal and his
contemporaries first investigated the
diverse and intricate neuronal
morphologies in the brain1, significant
progress has been made in describing the
formation of neural circuits based on the
projection patterns of single neurons
(reviewed in Ref. 2). However, such
descriptive works only intensify our desire
for understanding how complex neural
circuits are established. In order to
visualize single neurons in the developing
brain, and also to investigate the
molecular mechanisms underlying the
observed phenomena, we have developed
a novel genetic system for performing
mosaic analysis in the Drosophila CNS
(Ref. 3).
Homozygous mutant,
uniquely labeled
Heterozygous cell
mitotic recombination
Homozygous wild type
TRENDS in Neurosciences
Fig. 1. Genetic basis of the mosaic analysis with a repressible cell marker (MARCM) system. After site-specific
mitotic recombination, a heterozygous mother cell can give rise to two daughter cells in which the chromosome arms
distal to the recombination site become homozygous. Driven by the tubulin 1 α promoter, GAL80 is ubiquitously
expressed and efficiently suppresses GAL4-dependent expression of a UAS-marker gene (i.e. the GAL80 pink
rectangle binds to the GAL4 orange box at the UAS site, preventing transcription). If tubP-GAL80, but not GAL4 or
UAS-marker, is inserted on the chromosome arm carrying the wild-type (+) gene of interest, the daughter cell
homozygous for the mutant gene (x) no longer contains tubP-GAL80 (pink rectangle). Therefore, the marker gene can
be specifically turned on by GAL4 (orange box) in homozygous mutant cells. Adapted from Ref. 3.
useful to determine the stage-specific,
cell-autonomous roles of a gene that has
pleiotropic functions in diverse tissues
and at different developmental stages.
In conventional mosaic analysis,
homozygous mutant cells are identified
as either unstained cells or cells that are
stained with twice the intensity of
heterozygous cells5. Because neurons
are densely packed and neuronal
processes are fasciculated and often
extend over long distances, it is
impossible to analyze such mutant cells
in the CNS using conventional mosaic
The system that we developed, called
MARCM (mosaic analysis with a
repressible cell marker)3, allows one to
label homozygous mutant cells uniquely
in mosaic tissues, which is essential for
performing mosaic analysis in the
complicated nervous system. To achieve
this, the yeast GAL80 protein was
introduced into the GAL4-UAS binary
expression system in Drosophila6. The
MARCM system initially contains cells
that are heterozygous for a transgene
encoding the GAL80 protein, which
inhibits the activity of the transcription
factor GAL4 (Ref. 7). Following
FLP/FRT-mediated mitotic
recombination8, the GAL80 transgene is
removed from one of the daughter cells,
thus allowing expression of a GAL4driven reporter gene specifically in this
daughter cell and its progeny (Fig. 1). If
there is a mutation located on the
chromosome arm in trans to the
chromosome arm containing the GAL80
transgene, the uniquely labeled GAL80negative (GAL80−) cells should be
homozygous for this mutation.
Therefore, one can specifically label the
homozygous mutant cells in a mosaic
tissue using the MARCM system. To
examine the entire projection patterns of
such labeled neurons, we have designed
a membrane-targeted green fluorescent
protein (GFP) (mCD8-GFP) as a marker
in the MARCM system. This article
describes three different examples that 0166-2236/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S0166-2236(00)01791-4
Research Update
TRENDS in Neurosciences Vol.24 No.5 May 2001
A multi-cellular Nb clone
Single cell/two-cell clones
TRENDS in Neurosciences
Fig. 2. Three distinct clone sizes of mushroom body (MB) neurons. (a) and (b) In the Drosophila CNS, a neuroblast
(Nb) generates a series of ganglion mother cells (GMC or G in Fig.) via asymmetric divisions, and each GMC gives
rise to two post-mitotic neurons (N). (a) In the MARCM (mosaic analysis with a repressible cell marker) system, if
a Nb becomes GAL80-negative (GAL80–) after Flipase (FLP)-mediated mitotic recombination, all neurons
subsequently derived from this GAL80 – Nb are specifically labeled and appear as a multicellular Nb clone. (b) If a
GMC loses GAL80 after mitotic recombination, two neurons derived from the GAL80– GMC are labeled and
become a two-cell clone. By contrast, if mitotic recombination occurs in a dividing GMC, only one of the two postmitotic neurons will be labeled. (c) and (d) Composite confocal images of MARCM clones of MB neurons. Using a
membrane-targeted green fluorescent protein (GFP) (mCD8-GFP), the entire morphologies of uniquely labeled
MARCM clones in intact brains could be observed. (c) A MB Nb clone generated by inducing mitotic
recombination in a newly hatched larva (NHL) consists of hundreds of neurons at the adult stage. There are five
axon bundles (lobes) in the adult MB: γ, β’ and β projecting towards the midline and α’ and α projecting dorsally.
(d) Unique labeling of single-cell and two-cell clones of MB neurons generated in NHL reveals that each cell body
extends a single process from which dendrites (arrowhead) branch out. The scale bar unit is in microns. Adapted
from Ref. 10.
illustrate the utility of the MARCM
Example 1: cell lineage analysis
Formation of the Drosophila CNS
involves a stereotypic set of neuroblasts
(Nbs), each of which undergoes a series of
asymmetric divisions to produce a
characteristic clone of neurons. Each
asymmetric division of a Nb generates
another Nb and ganglion mother cell
(GMC), which subsequently gives rise to
two neurons after an additional cell
division9. Consequently, three sizes of
neuronal clones can be labeled using the
MARCM system (Fig. 2)3. A multicellular
clone is derived from a GAL80− Nb,
whereas a two-cell clone is the product of
an isolated GAL80− GMC. If mitotic
recombination occurs in a dividing GMC,
one of the two post-mitotic neurons might
lose GAL80 and become a single-cell clone
(Fig. 2).
The MARCM system can be used for
cell lineage analysis by taking advantage
of the following properties. First, one can
use heat shock to induce FLP expression at
a specific stage, and only precursors that
divide actively at that stage are subject to
mitotic recombination. Therefore, patterns
of clones induced at various stages should
reflect the spatial patterns of neurogenesis
at the corresponding stages. Second, all
neurons in a multicellular Nb clone are
derived from a common progenitor, so
examination of the compositions of Nb
clones and their projection patterns
reveals the roles of cell lineage in the
construction of neural circuits. Third, in
single-cell and two-cell clones, one can
visualize morphological differentiation of
single neurons. Induction of single-cell and
two-cell clones at various time points
allows the projection patterns of neurons
that are generated at different stages to be
MARCM analysis of the development
of mushroom bodies (MB), the insect
center for olfactory learning and memory,
provides one example for such cell lineage
analysis. Examining the Nb clones of MB
neurons at different stages illustrates
how adult MBs acquire their mature
morphologies after production of postmitotic neurons through larval and pupal
stages10. The axon projection patterns of
individual MB neurons are further
revealed by examining single-cell and
two-cell clones systematically induced at
different stages10. There are three distinct
types of neurons and their axons
fasciculate into three different sets of MB
lobes10,11. Interestingly, they are
sequentially derived from common
progenitors, and the MB neurons that are
generated before the mid-third instar
stage remodel their projections after
pruning larval-specific processes during
early metamorphosis10. Similar strategies
can be applied to study the development
of other neural structures in the
Drosophila CNS.
Example 2: functional analysis of candidate
pleiotropic genes
Because one can selectively create clones
of MB neurons that are homozygous for a
mutation of interest within an otherwise
phenotypically wild-type organism, the
MARCM-labeled MB neurons have
recently become model neurons for
determining the cell-autonomous
functions of pleiotropic genes in neuronal
development. For example, the actinbinding protein encoded by short stop
(kakapo) is important for extension and
correct guidance of axons3 (Fig. 3b). The
Drosophila small GTPase, RhoA, was
found to be required for proliferation of
neuroblasts and restriction of dendritic
growth12 (Fig. 3c,d). Trio, a guanine
nucleotide exchange factor that activates
the small GTPase, Rac, was shown to be
autonomously required in MB neurons
for extension and correct guidance of
their axons13. Mutations in the
Drosophila homolog of the mammalian
Lissencephaly 1 gene, Lis1, or a
cytoplasmic dynein heavy chain,
caused defects in neuroblast
proliferation, dendritic elaboration, and
axonal transport14. In addition, the
Dachshund nuclear protein is
autonomously required for proper
differentiation of pupal-born MB
neurons, as shown by the selective
Research Update
TRENDS in Neurosciences Vol.24 No.5 May 2001
Calyx (dendrites)
Cell bodies
Fig. 3. Requirement of short stop and RhoA for different aspects of MB morphogenesis. Composite confocal
images of neuroblast (Nb) clones of mushroom body (MB) neurons that were generated in newly hatched larva
(NHL) and examined at the wandering third-instar stage. Compared with a wild-type clone (a), the short stop mutant
clone (b) has ectopic axon bundles (arrowheads) and most axons in the peduncle stop at the bifurcation point
(arrow). By contrast, the RhoA mutant clone (c) contains much fewer cell bodies with over-extended dendrites
(arrowheads) but no detectable axon defect. A dendritic marker, the Nod-β-galactosidase chimeric protein [red in
(d)], was used to confirm the dendritic nature of the overshooting processes. Scale bar is in microns. Adapted from
Refs 3 and 12.
loss of a specific axonal lobe contributed
by the pupal-born MB neurons in spite
of normal numbers of neurons in
mutant clones15.
If required, the non-cell autonomous
effects of a mutant clone can be
addressed by specifically labeling its
immediate wild-type sibling cells using
‘reverse MARCM’12. If the mutation of
interest is located on the same
chromosome arm as the GAL80
Third instar
transgene, the MARCM-labeled cells will
be equivalent to the homozygous wildtype part of a ‘twin spot’ in conventional
mosaic analysis5. With respect to cell
lineage, they are closest to the mutant
cells and should be the ideal subjects for
examining any non-cell autonomous
effects of a mutation. However, a clearcut interpretation of the phenotypes
might require visualizing both sides of
the ‘twin spot’.
18 hrs APF
Fig. 4. Cell-autonomous requirement of ultraspiracle (usp) for remodeling of γ neurons. Composite confocal images
of single-cell and two-cell clones of wild-type (a)–(c) or usp– (d)–(f) γ neurons that were fixed at the late larval (a) and
(d), early pupal (b) and (e), or adult stages (c) and (f). Wild-type γ neurons acquired different stage-specific axon
projections [arrows, compare (c) with (a)] after pruning the larval-specific axonal branches [arrows in (a)] as well as
most dendrites [arrowhead, compare (b) with (a)] during early metamorphosis. By contrast, no remodeling was
observed in usp mutant γ neurons (d)–(f) such that the larval-specific bifurcation of axons (arrows) was retained
through metamorphosis (e) and into the adult stage (f). Note that remodeling of dendrites (arrowheads) was also
blocked. Adapted from Ref. 16.
One alternative for manipulating gene
functions specifically in the MARCMlabeled cells involves ectopic expression of
transgenes that encode gain-of-function
proteins. In the absence of GAL80, GAL4
can activate expression of all the
transgenes that are driven by the UAS
promoter. Therefore, multiple marker
genes can be visualized, and distinct
aspects of neuronal morphologies can be
visualized simultaneously12,14, or mutant
phenotypes could be rescued by selectively
expressing various transgenes in the
MARCM mutant clones12,14. In
combination with MARCM analysis of
loss-of-function mutations, these other
features greatly facilitate molecular
genetic studies of complex developmental
Example 3: genetic screen to identify new
Following analysis of wild-type neuronal
development using the MARCM system,
one can in theory systematically
investigate the underlying molecular
mechanisms by screening for random
mutations that disrupt the normal
developmental processes. Indeed, a genetic
mosaic screen has been used to identify
genes required for remodeling of MB
neurons16. Using single-cell mosaics, it was
shown that all single-cell and two-cell
clones of MB neurons generated in newly
hatched larvae altered their axon
projection patterns after pruning of the
larval-specific branches followed by
outgrowth of the adult-specific process
during early metamorphosis (Fig. 4a–c). In
order to elucidate the molecular
mechanisms that control remodeling of
neuronal processes, a genetic mosaic screen
was conducted and mutations that block
MB remodeling were identified (Fig. 4d–f).
The first mutation was found to be in the
ultraspiracle gene, encoding a co-receptor
for the nuclear hormone ecdysone16.
Identifying additional mutations should
Research Update
shed more light on the molecular
mechanisms of neuronal remodeling.
Future improvement and extension of the
MARCM system
The MARCM system has noteworthy
limitations. If the gene of interest is
expressed in precursor cells, knocking out
a gene in MARCM clones does not
guarantee an immediate and complete
loss of its encoded protein, the perdurance
of which might complicate phenotypic
analysis of mutant cells. Similarly, the
homozygous mutant cells do not express
the marker immediately because of the
perdurance of the GAL80 protein, so the
current MARCM system is not useful to
examine events that happen shortly after
the induction of mitotic recombination.
Engineering a more labile GAL80 version
might overcome this difficulty.
Development of various means of
controlling the patterns of mitotic
recombination, which are required for
better targeting of MARCM analysis to
distinct types of neurons, will also extend
the use of the MARCM system.
Although this review is focused on
Drosophila brain development, MARCMbased mosaic analysis could in theory also
be applied to study the function of neural
genes in behavior. It could also be applied
to study many developmental processes,
in which positive marking of mutant cells
is advantageous and sometimes essential.
For example, MARCM has recently been
used to studying Drosophila
spermatogenesis17, asymmetric cell
division in the adult sensory organ
precursor18, and planar cell polarity
in adult wing19. Finally, given that both
the GAL4-UAS expression system20 and
the FLP/FRT (Ref. 21) or CRE/LOX
(Ref. 22) mitotic recombination systems
TRENDS in Neurosciences Vol.24 No.5 May 2001
function in mice, it might be possible to
adapt the MARCM system to study
neuronal development in the mouse brain.
We acknowledge grant support from the
National Institutes of Health, and thank
J. Reuter and anonymous reviewers for
their suggestions that improved the
1 Cajal, S.R. (1911) Histology of the Nervous System
of Man and Vertebrates, Oxford University Press,
2 Cowan, W.M. (1998). The emergence of modern
neuroanatomy and developmental neurobiology.
Neuron 20, 413–426
3 Lee, T. and Luo, L. (1999) Mosaic analysis with a
repressible cell marker for studies of gene
function in neuronal morphogenesis. Neuron 22,
4 Ashburner, M. (1989) Drosophila: a Laboratory
Handbook, Cold Spring Harbor Laboratory Press
5 Xu, T. and Rubin, G.M. (1993) Analysis of genetic
mosaics in developing and adult Drosophila
tissues. Development 117, 1223–1237
6 Brand, A.H. and Perrimon, N. (1993) Targeted
gene expression as a means of altering cell fates
and generating dominant phenotypes.
Development 118, 401–415
7 Ma, J. and Ptashne, M. (1987) The carboxyterminal 30 amino acids of GAL4 are recognized
by GAL80. Cell 50, 137–142
8 Golic, K.G. and Lindquist, S. (1989) The FLP
recombinase of yeast catalyzes site-specific
recombination in the Drosophila genome. Cell 59,
9 Goodman, C.S. and Doe, C.Q. (1993) in The
Development of Drosophila melanogaster (Bate,
M. and Martinez-Arias, A., eds), pp. 1131–1206,
Cold Spring Harbor Laboratory Press
10 Lee, T. et al. (1999) Development of the Drosophila
mushroom bodies: sequential generation of three
distinct types of neurons from a neuroblast.
Development 126, 4065–4076
11 Crittenden, J.R. et al. (1998) Tripartite mushroom
body architecture revealed by antigenic markers.
Learn. Mem. 5, 38–51
12 Lee, T. et al. (2000) Essential roles of Drosophila
RhoA in the regulation of neuroblast proliferation
and dendritic but not axonal morphogenesis.
Neuron 25, 307–316
Awasaki, T. et al. (2000) The Drosophila trio plays
an essential role in patterning of axons by
regulating their directional extension. Neuron 26,
Liu, Z. et al. (2000) Drosophila Lis1 is required for
neuroblast proliferation, dendritic elaboration
and axonal transport. Nat. Cell Biol. 2, 776–783
Martini, S.R. et al. (2000) The retinal
determination gene, dachshund, is required for
mushroom body cell differentiation. Development
127, 2663–2672
Lee, T. et al. (2000) Cell autonomous requirement
of the USP/EcR-B ecdysone receptor for
mushroom body neuronal remodeling. Neuron 28,
Kiger, A.A. et al. (2000) Somatic support cells
restrict germline stem cell self-renewal and
promote differentiation. Nature 407, 750–754
Roegiers, F. et al. (2001) Two types of asymmetric
divisions in the Drosophila sensory organ
precursor cell lineage. Nat. Cell Biol. 3, 58–67
Winter, C. et al. Drosophila Rho-associated kinase
(Drok) links Frizzled–mediated planar cell
polarity signaling to the actin cytoskeleton. Cell
(in press)
Ornitz, D.M. et al. (1991) Binary system for
regulating transgene expression in mice:
targeting int-2 gene expression with yeast
GAL4/UAS control elements. Proc. Natl. Acad.
Sci. U. S. A. 88, 698–702
Dymecki, S.M. (1996) Flp recombinase promotes
site-specific DNA recombination in embryonic
stem cells and transgenic mice. Proc. Natl. Acad.
Sci. U. S. A. 93, 6191–6196
Gu, H. et al. (1994) Deletion of a DNA polymerase
beta gene segment in T cells using cell typespecific gene targeting. Science 265, 103–106
Tzumin Lee
Dept of Cell and Structural Biology,
University of Illinois, Urbana-Champaign,
Urbana, IL 61801, USA.
Liqun Luo*
Dept of Biological Sciences, Neurosciences
Program, Stanford University, Stanford,
CA 94305, USA.
*e-mail: [email protected]
Letters to the Editor
Trends in Neurosciences : a forum for comment
The Editor welcomes correspondence from readers of Trends in Neurosciences on any articles published. Please state
clearly whether or not you wish the letter to be considered for publication. If a letter is considered suitable for publication, it
will be sent to the author of the original article for their response; both the letter and reply will be published together. Please
note: submission does not guarantee publication.
Please address Letters to:
Dr Siân Lewis, Editor
Trends in Neurosciences, 84 Theobald’s Road, London, UK WC1X 8RR.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF