Article

Asymmetric localization of Numb:EGFP in dividing neuroepithelial cells during neurulation in Danio rerio

Institut für Entwicklungsbiologie, Universität zu Köln, 50923 Köln, Germany.
Developmental Dynamics (Impact Factor: 2.38). 04/2006; 235(4):934-48. DOI: 10.1002/dvdy.20699
Source: PubMed
ABSTRACT
In the neural plate and tube of the zebrafish embryo, cells divide with their mitotic spindles oriented parallel to the plane of the neuroepithelium, whilst in the neural keel and rod, the spindle is oriented perpendicular to it. This change is achieved by a 90 degrees rotation of the mitotic spindle. We cloned zebrafish homologues of the gene for the Drosophila cell fate determinant Numb, and analyzed the localization of EGFP fusion proteins in vivo in dividing neuroepithelial cells during neurulation. Whereas Numb isoform 3 and the related protein Numblike are localized in the cytoplasm, Numb isoform 1 is localized to the cell membrane. Time-lapse analyses showed that Numb 1 is distributed uniformly around the cell cortex in dividing cells during plate and keel stages, but becomes localized at the basolateral membrane of some dividing cells during the transition from neural rod to tube. Using in vitro mutagenesis and Numb:EGFP deletion constructs, we showed that the first 196 amino acids of Numb are sufficient for this localization. Furthermore, we found that an 11-amino acid insertion in the PTB domain is essential for localization to the cortex, whereas amino acids 2-12 mediate the basolateral localization in the neural tube stage.

Full-text

Available from: Nico Scheer, Jan 27, 2015
RESEARCH ARTICLE
Asymmetric Localization of Numb:EGFP in
Dividing Neuroepithelial Cells During
Neurulation in Danio rerio
Alexander M. Reugels,
*
Barbara Boggetti, Nico Scheer,
and Jose´ A. Campos-Ortega
In the neural plate and tube of the zebrafish embryo, cells divide with their mitotic spindles oriented
parallel to the plane of the neuroepithelium, whilst in the neural keel and rod, the spindle is oriented
perpendicular to it. This change is achieved by a 90° rotation of the mitotic spindle. We cloned zebrafish
homologues of the gene for the Drosophila cell fate determinant Numb, and analyzed the localization of
EGFP fusion proteins in vivo in dividing neuroepithelial cells during neurulation. Whereas Numb isoform
3 and the related protein Numblike are localized in the cytoplasm, Numb isoform 1 is localized to the cell
membrane. Time-lapse analyses showed that Numb 1 is distributed uniformly around the cell cortex in
dividing cells during plate and keel stages, but becomes localized at the basolateral membrane of some
dividing cells during the transition from neural rod to tube. Using in vitro mutagenesis and Numb:EGFP
deletion constructs, we showed that the first 196 amino acids of Numb are sufficient for this localization.
Furthermore, we found that an 11–amino acid insertion in the PTB domain is essential for localization to the
cortex, whereas amino acids 2–12 mediate the basolateral localization in the neural tube stage.
Developmental Dynamics 235:934 –948, 2006.
© 2006 Wiley-Liss, Inc.
Key words: Danio rerio; zebrafish; neurulation; neuroepithelial polarity; spindle orientation; asymmetric localization;
Numb; Numblike; PTB domain; PRR
Accepted 27 December 2005
INTRODUCTION
Proper spatial and temporal specifica-
tion of cells during development is
crucial for the generation of cellular
diversity in the nervous system of
complex organisms (McConnell, 1991;
Edlund and Jessell, 1999). One way to
achieve cell diversification is by
means of asymmetric cell divisions
where the two daughter cells adopt
different fates. Such divisions can in-
volve extrinsic and/or intrinsic factors
(Horvitz and Herskowitz, 1992; Jan
and Jan, 1998). Extrinsic factors such
as the ligand Delta and its receptor
Notch act via cell– cell signalling to
specify distinct cell fates: Delta acti-
vates Notch in adjacent cells, which
finally directs them into alternative
developmental pathways (Campos-
Ortega, 1995; Artavanis-Tsakonas et
al., 1999; Wakamatsu et al., 2000). In-
trinsic factors are cell fate determi-
nants, such as the Drosophila tran-
scription factor Prospero or the
membrane-associated protein Numb
(Nb), which are localized asymmetri-
cally in dividing cells and hence are
segregated to only one daughter cell,
enabling this cell to adopt a different
fate from that of its sibling (Uemura
et al., 1989; Hirata et al., 1995).
Vertebrate homologues of the Dro-
sophila numb gene have been identi-
fied in mouse, rat, human, and
chicken (Zhong et al., 1996; Verdi et
al., 1996, 1999; Dho et al., 1999;
Wakamatsu et al., 1999) and database
analyses also reveal putative homo-
The Supplementary Material referred to in this article can be found at www.interscience.wiley.com/jpages/1058-8388/suppmat
Institut fu¨ r Entwicklungsbiologie, Universita¨ t zu Ko¨ln, Ko¨ln, Germany
Grant sponsor: Deutsche Forschungsgemeinschaft; Grant number: DFG, SFB 572; Grant sponsor: Fonds der Chemischen Industrie.
Nico Scheer’s present address is ARTEMIS Pharmaceuticals GmbH, Neurather Ring 1, 51063 Ko¨ln, Germany.
*Correspondence to: Dr. Alexander M. Reugels, Institut fu¨ r Entwicklungsbiologie, Universita¨ t zu Ko¨ ln, 50923 Ko¨ln
Germany. E-mail: alexander.reugels@uni-koeln.de
DOI 10.1002/dvdy.20699
Published online 21 February 2006 in Wiley InterScience (www.interscience.wiley.com).
DEVELOPMENTAL DYNAMICS 235:934–948, 2006
© 2006 Wiley-Liss, Inc.
Page 1
logues in other species, e.g., Xenopus
(see Fig. 2 and data not shown). Addi-
tionally, mice and humans (and prob-
ably other vertebrates) have a second
gene, called numblike (nbl), which
shows significant sequence similarity
to Drosophila numb (Zhong et al.,
1997; Salcini et al., 1997). Moreover,
in mouse and human, four different
isoforms of the Numb protein have
been identified (Dho et al., 1999; Verdi
et al., 1999). Since the mammalian ho-
mologues of Numb are functional in
Drosophila (Zhong et al., 1996; Verdi
et al., 1996), it is tempting to specu-
late that vertebrate homologues exert
similar functions during neurogenesis
as their Drosophila counterpart.
Numb is capable of interacting with
a wide variety of different proteins. Its
sequence can be subdivided into an
amino-terminal part, which contains a
phosphotyrosine-binding (PTB) do-
main and is highly conserved in all
identified Numb homologues, and a
more diverse carboxy-terminal seg-
ment that includes a proline-rich re-
gion (PRR). Close to the carboxy-ter-
minal end are two conserved sequence
motifs—aspartate-proline-phenylala-
nine (DPF) and asparagine-proline-
phenylalanine (NPF)—that serve as
binding motifs for the clathrin adaptor
-adaptin and Eps15 Homology (EH)
domain proteins, respectively, both of
which are involved in receptor-medi-
ated endocytosis (Berdnik et al., 2002;
Santolini et al., 2000). The PTB do-
main can interact with diverse pro-
teins such as the multipass trans-
membrane protein NIP (Numb
interacting protein, Qin et al., 2004);
the adapter protein PON (Partner of
Numb, Lu et al., 1998); the serine/
threonine kinase NAK (Numb-associ-
ated kinase, Chien et al., 1998), a neg-
ative regulator of Numb function; and
the RING type E3 ubiquitin ligase
LNX (Ligand of Numb X, Dho et al.,
1998), which is involved in ubiquitina-
tion and, hence, protein degradation.
Vertebrate Numb has been demon-
strated to be asymmetrically localized
as a cortical crescent in organisms
such as mouse, rat, and chicken
(Zhong et al., 1996; Wakamatsu et al.,
1999, 2000; Cayouette et al., 2001)
and there is good evidence that this
asymmetric localization is also associ-
ated with the adoption of different cell
fates by the daughter cells (Shen et
al., 2002; Cayouette and Raff, 2003).
In order to ensure the unequal segre-
gation of asymmetrically distributed
cell fate determinants, proper align-
ment of the mitotic spindle with re-
spect to the location of the determi-
nant is essential. Neuroectodermal
cells in Drosophila undergo planar mi-
toses, but in delaminated neuroblasts,
the spindle rotates so that it is perpen-
dicular to the neuroectoderm, result-
ing in a basal crescent of Prospero and
Numb that overlies one of the centro-
somes (Kaltschmidt et al., 2000). Sim-
ilarly, in the sensory bristle lineage,
the sensory organ precursor (SOP) di-
vides within the plane of the epithe-
lium, whereas the spindle in pIIb
changes its orientation and the cell
divides along the apicobasal axis (Gho
and Schweisguth, 1998; Roegiers et
al., 2001).
In a previous study, we followed the
mitotic behaviour of neuroepithelial
cells during neurulation in zebrafish
by analyzing strains carrying stable
transgenic insertions of the zebrafish
Histone2A.F/Z gene fused to a se-
quence encoding an enhanced variant
of the green fluorescent protein
(EGFP, Pauls et al., 2001). These in
vivo analyses revealed a stereotypic
orientation of the mitotic spindle in
each stage of neurulation. In the neu-
ral plate and tube of the zebrafish em-
bryo, the mitotic spindle is oriented
parallel to the plane of the neuroepi-
thelium, resulting in planar cell divi-
sions, whereas in the neural keel/rod
stages, the spindle rotates by 90° so
that it lies perpendicular to the apical
surface of the epithelium, resulting in
orthogonal cell divisions (Geldma-
cher-Voss et al., 2003; also see Kim-
mel et al., 1994; Papan and Campos-
Ortega, 1994; Concha and Adams,
1998).
Here we report the cloning of two
isoforms of zebrafish Numb that differ
by the presence of 11 additional amino
acids within the PTB domain, and the
identification of the Numb-related
protein Numblike. In order to analyze
their localization in dividing neuroep-
ithelial cells during neurulation in
vivo, we injected mRNA encoding ei-
ther of the Numb isoforms, or Numb-
like, fused to EGFP. These analyses
revealed that Numb isoform 1, which
has the 11–amino acid insertion in the
PTB domain (PTB
L
form), is localized
to the cell membrane in dividing cells,
whereas both the Numb isoform 3
without the insertion (PTB
S
form) and
Numblike (which also has a PTB
S
-
type PTB domain) are localized in the
cytoplasm. Time-lapse analyses showed
that the PTB
L
form is distributed
ubiquitously around the cell cortex of
dividing cells during the neural plate
and keel stages, but becomes localized
to the basolateral membrane of some
dividing cells during the transition
from the rod to the tube stage. Using
in vitro mutagenesis and Numb-
EGFP deletion constructs, we were
able to show that the first 196 amino
acids of Numb are sufficient to ensure
localization of the fusion protein to the
basolateral plasma membrane. Fur-
thermore, we found that the 11–amino
acid insertion in the PTB domain is
essential for cortical localization per
se, whereas amino acids 2–12 mediate
the basolateral localization, since de-
letion of these amino acids from the
PTB
L
form leads to a ubiquitous cor
-
tical localization of the corresponding
Numb:EGFP deletion construct in the
neural tube stage.
RESULTS
Two Zebrafish Homologues
of the Drosophila Cell Fate
Determinant Numb
In mouse and human, four different
isoforms of Numb have been described
(Dho et al., 1999; Verdi et al., 1999),
which are generated via alternative
splicing and differ by the presence of
an 11–amino acid insertion within the
PTB domain and/or the presence of an
additional 49 (in mouse) or 48 (in hu-
man) amino acids within the C-termi-
nal proline-rich region (PRR). The dif-
ferent isoforms were named Numb 1
(PTB
L
PRR
L
), Numb 2 (PTB
L
PRR
S
),
Numb 3 (PTB
S
PRR
L
), and Numb 4
(PTB
S
PRR
S
), respectively, with “L”
indicating the long form with inser-
tion and “S” standing for the short
form without insertion. Using a com-
bination of PCR, RT-PCR, and library
screening, we were able to clone
Numb isoforms 1 and 3, which have
predicted molecular masses of 73.0
and 71.7 kDa, respectively (Fig. 1),
from zebrafish. We also identified a
second Numb homologue, Numblike,
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 935
Page 2
with a predicted molecular mass of
66.9 kDa (Fig. 1).
Alignment with the different iso-
forms from mouse and human shows
that both zebrafish Numb isoforms
share a 51–amino acid insertion in the
PRR, but differ from each other by the
presence of the 11–amino acids inser-
tion in the PTB domain. Since we
could not clone the PRR
S
isoforms by
RT-PCR from total RNA isolated from
mixed stages up to 24 hr post-fertili-
zation (hpf), we analyzed the genomic
structure of zebrafish numb. We found
that two introns in the correct phase
for alternative splicing flank the exon
that encodes the PTB
L
insertion (in
-
trons 2 and 3 in Fig. 1). The same is
true for the exon that encodes the 51–
amino acid PRR
L
insertion (introns 8
and 9 in Fig. 1). Therefore, the exon/
intron structure of zebrafish Numb
should allow the generation of all four
isoforms known from mouse and hu-
man via alternative splicing. Whereas
overall homology with Drosophila
Numb is rather weak, the N-terminal
portion containing the so-called Numb
domain (Zhong et al., 1996) is highly
conserved (Fig. 2). Since both Numb
and Numblike show this high homol-
ogy to Drosophila Numb, we per-
formed a maximum-likelihood analy-
sis with homologues from various
species in order to determine whether
Numb or Numblike is more closely re-
lated to the Drosophila protein (for
details, see Experimental Procedures
section). This analysis indicated that
all vertebrate Numb proteins cluster
in a clade with Drosophila Numb,
whereas the different Numblike pro-
teins belong to a different clade. The
two clades are separated by a branch
with a bootstrap value of 81% (see
Supplemental Fig. 1). Of the two con-
served sequence motifs, aspartate-
proline-phenylalanine (DPF) and as-
paragine-proline-phenylalanine (NPF),
close to the carboxy-terminal end,
which serve, respectively, as binding
motifs for the clathrin adaptor
-adaptin and EH domain proteins
(Berdnik et al., 2002; Santolini et al.,
2000) and are present in all Numb
homologues identified so far, zebrafish
Numb and Numblike retain the NPF
motif, whereas they have aspartate-
alanine-phenylalanine (DAF in Numb)
or glutamic acid-histidine-phenylala-
nine (EHF in Numblike) in place of
DPF.
RT-PCR experiments revealed that
the transcripts of both numb isoforms
are present in all developmental
stages tested (64 cells up to 5 days;
Fig. 3). In situ hybridizations with a
numb-specific antisense probe further
showed that Numb is ubiquitously ex-
pressed during blastula and gastrula
stages. With the beginning of somito-
genesis, expression becomes concen-
trated at the midline. By the 18-
somite stage, a strong signal is found
at the midline from the head to the
tail region, and in the retina. How-
ever, at later stages (30 somites), ex-
pression in the neural tube ceases al-
most completely, while transcripts are
still present in the fore-, mid-, and
hindbrain and in the eyes (data not
shown).
Numb(PTB
L
PRR
L
):EGFP
Becomes Basolaterally
Localized During the
Transition From the Neural
Rod to the Neural Tube
Stage
The zebrafish, like other teleosts, un-
dergoes secondary neurulation. In the
following, we briefly describe neurula-
tion at the level of the 1st to 5th
somite (the prospective cervical spinal
cord), where all our observations were
made (refer to Kimmel et al., 1990,
1995; Schmitz et al., 1993; Papan and
Campos-Ortega, 1994). In the course
of convergence movements that follow
gastrulation, the neural plate folds in-
ward at the midline, between the 6-
and 10-somite stages (13.3 hpf), to
form the neural keel. During the 10- to
14-somite stage, the keel progres-
sively rounds up, forming the neural
rod, a massive cellular conglomerate
without a lumen, which is finally over-
laid by the adjacent epidermis at
about 16 hpf. The neurocoel, the lu-
men of the neural tube, forms second-
arily by cavitation of the neural rod, as
the neuroepithelial cells retract their
apical processes from the midline. Neu-
rocoel formation starts in the 17- to 18-
somite embryo, beginning ventrally in
the spinal cord and progressing towards
dorsal levels, and is completed by the
30-somite stage (approximately 24
hpf). Throughout neurulation, the ze-
brafish neuroepithelium remains es-
sentially pseudostratified, consisting
of columnar cells that extend apically
from the basement membrane. As in
many other epithelia, neuroepithelial
cells round up and divide apically
(Hinds and Ruffett, 1971).
In order to analyze the localization
of Numb and Numblike during neuru-
lation in vivo, we injected mRNA en-
coding either of the Numb isoforms, or
Numblike, fused to EGFP into the
yolk of wild-type embryos, starting
with mRNA transcribed from the
PTB
L
construct numb(PTB
L
PRR
L
):
egfp. As mentioned above, during the
neural plate stage, mitoses are planar,
i.e., divisions occur essentially paral-
lel to the plane of the epithelium. They
initially occur predominantly in an
anteroposterior direction and become
oriented preferentially in a mediolat-
eral direction at the onset of neural
Fig. 1. Schematic diagrams of Drosophila Numb (Nb) and zebrafish Numb and Numblike (Nbl).
Light grey boxes indicate the conserved Numb domains. Dark grey boxes mark the PTB domains.
The isoform-specific insertions within the PTB domain and the C-terminal proline rich region (PRR)
are highlighted in black. The small bold numbers mark intron positions within the corresponding
genomic sequence.
936 REUGELS ET AL.
Page 3
Fig. 2. Alignment of the Numb domains (positions 42–360) of various Numb homologues from Danio rerio (D.r.), Homo sapiens (H.s.), Mus musculus
(M.m.), Xenopus tropicalis (X.t.), Gallus gallus (G.g.), and Drosophila melanogaster (D.m.). Conserved amino acids are marked by asterisks. The PTB
domain as defined by the SMART analysis (Schultz et al., 1998; Letunic et al., 2004) is underlined. The 11–amino acid insertion within the PTB domain
of the Numb 1 isoforms is highlighted by grey shading. Positions that were used for the maximum likelihood analysis are marked by dots. Accession
numbers are: AY583653 (D.r. Nb 1), AY583654 (D.r. Nb 3), DQ022744 (D.r. Nbl), AF171938 (H.s. Nb 1), AF171940 (H.s. Nb 3), AF015041 (H.s. Nbl),
AF169192 (M.m. Nb 1), AF169191 (M.m. Nb 3), U96441 (M.m. Nbl), CR942503 (X.t. Nb), AF176086 (G.g. Nb), and M27815 (D.m. Nb).
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 937
Page 4
plate infolding (Concha and Adams,
1998; Geldmacher-Voss et al., 2003).
In non-dividing cells of the neural
plate, a punctate, highly dynamic
Numb(PTB
L
PRR
L
):EGFP labelling
pattern is apparent, consisting of sev-
eral bright dots positioned along the
membrane and within the cytoplasm
(Fig. 4A and Supplementary Movie 1).
When the cells round up for mitosis,
the vesicle-like cytoplasmic labelling
disappears, while the cortical label-
ling becomes more pronounced, result-
ing in a ubiquitous distribution of
Numb(PTB
L
PRR
L
):EGFP around the
cell membrane, with no obvious pref-
erence for an anteroposterior, medio-
lateral, or dorsoventral (apicobasal)
location. Upon mitosis, Numb is dis-
tributed equally to the daughter cells
(Fig. 4A at 50”). In the daughter cells,
the cortical labelling again becomes
punctate and the vesicle-like cytoplas-
mic labelling reappears (Supplemen-
tary Movie 1). This change in the dis-
tribution of Numb(PTB
L
PRR
L
):EGFP
from being punctate and highly dy-
namic in non-dividing cells to being
concentrated at, and restricted to, the
cortex in dividing cells can also be ob-
served in the neural keel/rod and tube
stages (see Supplementary Movies
2– 4). In the following, we will focus on
the localization of Numb(PTB
L
PRR
L
):
EGFP in dividing cells. In the neural
keel/rod stage, cell divisions switch
from planar to orthogonal, due to a 90°
rotation of the mitotic spindle. Cell
divisions occur at the region where the
two halves of the neural anlage meet,
i.e., at the apical surfaces of the neu-
roepithelial cells, in the middle of the
neural rod. After cytokinesis, the
daughter cells integrate into opposite
sides of the neuroepithelium, giving
rise to bilateral progeny (Papan and
Campos-Ortega, 1994, 1997; Concha
and Adams, 1998; Geldmacher-Voss
et al., 2003). As in the neural plate
stage, Numb(PTB
L
PRR
L
):EGFP is
ubiquitously distributed around the
cell cortex of mitotic cells in the neural
keel and rod, and hence is segregated
to both daughter cells (Fig. 4B and
Supplementary Movie 2). However, as
the embryo proceeds from the rod to
the tube stage, Numb(PTB
L
PRR
L
):
EGFP localization becomes polarized
for the first time (Fig. 4C and Supple-
mentary Movie 3). Formation of the
neurocoel begins ventrally in the spi-
nal cord and gradually progresses to-
wards dorsal levels, accompanied by
the change from the orthogonal orien-
tation of cell divisions in the neural
keel and rod to the planar orientation
in the neural tube (Schmitz et al.,
1993; Geldmacher-Voss et al., 2003).
This change in orientation occurs pro-
gressively, rather than suddenly.
Therefore, during the transition from
rod to tube, one can observe both cells
that divide with the mitotic spindle
oriented perpendicular to the plane of
the neuroepithelium and cells that
have their spindles oriented parallel
to the neuroepithelial plane (typical
for the neural tube stage). Interest-
ingly, cells that are going to divide in a
planar orientation (arrowhead in Fig.
4C) start to localize Numb(PTB
L
PRR
L
):EGFP to the basolateral cell
cortex when they round up at the mid-
line for mitosis (Fig. 4C, frames 0 to
8), producing a basolateral crescent of
the protein that is segregated to both
daughter cells upon cytokinesis (Fig.
4C, frame 10). Occasionally, cells can
be observed that are dividing in a pla-
nar orientation but still have
Numb(PTB
L
PRR
L
):EGFP localized
ubiquitously around the cell cortex.
However, we never saw orthogonally
dividing cells in which Numb was
clearly localized to the basolateral
membrane, a configuration that would
also result in asymmetric segregation.
The localization of Numb(PTB
L
PRR
L
):EGFP to the basolateral mem
-
brane becomes even more pronounced
at later stages in the neural tube; the
apical side of both dividing and non-
dividing neuroepithelial cells is com-
pletely devoid of labelling (Fig. 5 and
Supplementary Movie 4). In all, 432
mitoses were analyzed within a region
of the developing spinal cord of several
different embryos that comprised 2–3
neuromeres. Time-lapse movies were
compiled for 2–4 hr at up to seven
different z-levels throughout the
depth of the neural tube. Some 420
(97.2%) cell divisions showed a baso-
lateral localization of Numb:EGFP; in
11 cases (2.6%) the cells exhibited a
uniform distribution of Numb:EGFP.
Only in one case (0.2%) did we con-
sider the crescent to be mislocalized,
e.g., to be more pronounced at the an-
terolateral side (Table 1).
Both the N-Terminus and
the 11–Amino Acid Insertion
in the PTB Domain Are
Necessary for Basolateral
Membrane Localization
In contrast to Numb(PTB
L
PRR
L
):
EGFP, which is localized to the cell
membrane of dividing neuroepithelial
cells throughout neurulation, the vast
majority of the Numb isoform with the
short PTB domain, Numb(PTB
S
PRR
L
), localizes to the cytosol (Fig. 6A
and B). The distribution of the fusion
protein varies from homogenous to
more punctate, but an asymmetric lo-
calization of Numb(PTB
S
PRR
L
) dur
-
ing cell division was never observed.
Since the only difference between the
Fig. 3. Developmental expression of zebrafish Numb 1 and Numb 3. Total RNAs were treated with
DNase I, retrotranscribed using an oligo-dT primer, and amplified by PCR (30 cycles) with isoform-
specific primer pairs. Both isoforms are expressed in all stages. A fragment of -actin was amplified
as a control.
938 REUGELS ET AL.
Page 5
numb(PTB
L
PRR
L
):egfp construct and
the numb(PTB
S
PRR
L
):egfp clone is
the presence of exon 3, encoding the
additional 11 amino acids within the
PTB
L
domain, this insertion must be
essential for the cortical localization of
Numb(PTB
L
PRR
L
). As expected,
Numblike:EGFP, which also has a
PTB
S
-type PTB domain, is also ex
-
pressed in the cytoplasm (Fig. 6C).
This is in accordance with the results
of experiments in MDCK cells from
mouse, which indicated that Numb
PTB
L
isoforms are located at the cell
membrane, whereas PTB
S
forms are
predominantly localized in the cytosol
and the nucleus (Dho et al., 1999). In
contrast to these cell culture experi-
ments, we did not see any distinct nu-
clear localization of either Numb(PTB
S
PRR
L
):EGFP or Numblike:EGFP (ar
-
rows in Fig. 6B and C). Deletion anal-
yses in Drosophila have shown that
the first 227 amino acids of Numb are
sufficient to ensure its asymmetric lo-
calization (Frise et al., 1996; Knoblich
et al., 1997). However, since Drosoph-
ila Numb lacks a comparable amino
acid insertion in its PTB domain, it
appears that asymmetric localization
in zebrafish involves, at least in part,
different or additional mechanisms.
Therefore, several deletion constructs
were analyzed to identify those re-
gions in the Numb protein that are
necessary for the observed basolateral
localization in cells of the neural tube
(Fig. 6D–H). We found that the N-ter-
minal part (including the 11–amino
acid insertion in the PTB domain) of
the protein becomes basolaterally lo-
calized, whereas the C-terminal part
(amino acids 307 to 680) is localized in
the cytoplasm (Fig. 6D and E). We
further found that a fragment encom-
passing the first 196 amino acids (cor-
responding to positions 42–240 of the
alignment in Fig. 2) is efficiently local-
ized to the basolateral cortex in more
than 90% of the cell divisions analyzed
(n 116), in essentially the same way
as the corresponding full-length pro-
tein, indicating that these amino acids
are sufficient for localization (Fig. 6F
Fig. 4. A–C: Sequences of time-lapse frames of mitotic cells of embryos injected with numb(PTB
L
PRR
L
):egfp mRNA. The transparent grey line marks
the midline. All photographs are dorsal views. Anterior is towards the top. Scale bar 20 m. A: Two mitotic cells in the neural plate of a 5- to 6-somite
embryo (arrows in 0). Cell divisions are planar at this stage. Numb:EGFP is distributed ubiquitously around the cell cortex and segregated to both
daughter cells upon cell division (arrows in 500”). B: A mitotic cell in the neural rod of a 16-hpf-old embryo (arrow in 0). Cells in this stage divide
perpendicularly to the midline, due to a 90° rotation of the mitotic spindle. As in the neural plate stage, Numb:EGFP is localized ubiquitously around
the cell cortex and, therefore, segregated to both daughter cells (arrows in 300”). C: Two mitotic cells during the transition from neural rod to neural
tube in a 17-hpf-old embryo. Whereas one cell (arrow in 0) still divides perpendicular to the plane of the epithelium with Numb:EGFP being
ubiquitously localized (arrows in 8), the other cell (arrowhead in 0) already divides parallel to plane of the epithelium, as is typical for all cell divisions
in the neural tube (arrowheads in 10). Note that this cell starts to localize Numb:EGFP to the basolateral cell cortex when it rounds up at the midline
for mitosis (frames 0–8). See also Supplemental Material for the corresponding Supplementary Movies 1 to 3.
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 939
Page 6
and Table 1). As expected, the corre-
sponding construct without the inser-
tion, Numb(PTB
S
186 669):EGFP,
fails to localize to the cell cortex (data
not shown). The high level of conser-
vation in the N-terminal region in the
various Numb homologues (Fig. 2)
prompted us to analyze whether this
region is also involved in basolateral
protein localization. Interestingly,
upon deletion of amino acids 2–12 (po-
sitions 43–53 in Fig. 2) from the N-
terminus, cortical localization is im-
paired and basolateral localization is
lost. This results in a considerable
amount of protein being found in the
cytosol and in a ubiquitous distribution
around the membrane in dividing cells,
respectively (Fig. 6G). Deletion of amino
acids 2–22 (positions 43– 63 in Fig. 2)
enhances this phenotype, as shown in
Figure 6H. Taken together, these data
show that the 11–amino acid insertion
in the PTB
L
domain is essential for cor
-
tical localization, whereas the N-termi-
nal amino acids 2–12 are important for
basolateral localization in dividing cells
of the zebrafish neural tube.
Mutation of the 11–Amino
Acid Insertion Impairs
Basolateral Membrane
Localization and Spindle
Orientation
Alignment of the PTB domain of Dro-
sophila Numb with the zebrafish
PTB
L
domain shows that the insert in
zebrafish lies between amino acids
111 and 112 of the Drosophila homo-
logue, C-terminal to -helix 2 and N-
terminal to -strand 2 (Figs. 2 and
Fig. 5. A sequence of 15 confocal micrographs of a dorsal view of the neural tube of a 24-hpf-old embryo injected with numb(PTB
L
PRR
L
):egfp
mRNA. The transparent grey line in the first picture marks the neurocoel. Scale bar 20 m. Three cells (arrows, arrowheads, and asterisks in 0) one
after another round up at the neurocoel and subsequently undergo mitotic cell division. Note that Numb:EGFP becomes localized to the basolateral
cell cortex in all three cells, whereas the apical side is completely devoid of signal. Although Numb:EGFP is clearly asymmetrically localized in these
cells, it is distributed to both daughter cells upon division, since the cells divide parallel to the plane of the neuroepithelium. See also Supplemental
Material for the corresponding Supplementary Movie 4.
TABLE 1. Effects of the Expression of Various Numb:EGFP Constructs on Cortical Protein Localization and
Mitotic Orientation in the Neural Tube Stage (for Classification of the Orientation of the Cell Divisions, See
Experimental Procedures Section)
Expressed protein n
Protein localization in dividing cells (%)
Basolateral Ubiquitous Mislocalized
Numb(PTB
L
PRR
L
)
432 420 (97.2) 11 (2.6) 1 (0.2)
Numb(PTB
L
197–680)
116 106 (91.4) 8 (6.9) 2 (1.7)
Numb(AAAVFAGFFAA)
94 80 (85.1) 14 (14.9) 0 (0)
Numb(DRKAAKAAAKK) 156 98 (62.8) 42 (26.9) 16 (10.3)
Expressed protein n
Orientation of the cell division (%)
Planar Oblique Orthogonal
Numb(PTB
L
PRR
L
)
432 426 (98.6) 6 (1.4) 0 (0)
Numb(PTB
L
197–680)
116 115 (99.1) 1 (0.9) 0 (0)
Numb(AAAVFAGFFAA)
94 86 (91.5) 4 (4.3) 4 (4.3)
Numb(DRKAAKAAAKK) 156 135 (86.5) 14 (9.0) 7 (4.5)
940 REUGELS ET AL.
Page 7
7A). Interestingly, this is not in the
neighbourhood of the PTB-binding
groove, as is obvious from the solution
structure of the Drosophila Numb
PTB domain-Nak peptide complex
(Zwahlen et al., 2000; also see Li et al.,
1998). Because of the importance of
the PTB
L
insertion for cortical local
-
ization in zebrafish, we performed ala-
nine-scanning mutagenesis experi-
ments (Cunningham and Wells, 1989;
Bass et al., 1991; Wells, 1991) in order
to identify residues that are critical for
localization. The PTB
L
insertion con
-
sists of the amino acids DRKVFKG-
FFKK and is composed of five amino
acids with non-polar side chains (Val,
V; Phe, F; Gly, G), one residue with an
acidic side chain (Asp, D), and five
amino acids with basic side chains
(Arg, R; Lys, K; Fig. 7B). As revealed
by a charge distributional analysis us-
ing the SAPS program (Brendel et al.,
1992), insertion of these amino acids
generates a cluster of positive charges
over a stretch of 23 amino acids
(amino acids 63 to 85, corresponding
to positions 107–129 in Fig. 2) within
the PTB domain of Numb(PTB
L
PRR
L
), which is unique for the PTB
L
isoform. We found that neither dele-
tion of the three central amino acids
(DRKV_FFKK) nor substitution of the
three N-terminal charged residues
with alanines (AAAVFKGFFKK) nor
substitution of the central lysine to-
gether with the two C-terminal ly-
sines (DRKVFAGFFAA) considerably
reduced the ability of the correspond-
ing Numb:EGFP fusion proteins to lo-
calize to the basolateral membrane
during cell divisions in the neural
tube (Fig. 7D–F). A disturbance of the
localization was only observed when
all the charged amino acids were re-
placed by alanines (AAAVFAGFFAA,
Fig. 7G): In this case, 14.9% of the
mitoses showed a ubiquitous mem-
brane localization of Numb(AAAV-
FAGFFAA):EGFP, whereas only 2.8%
of embryos injected with mRNA cod-
ing for numb(PTB
L
PRR
L
):egfp exhib
-
ited either a ubiquitous or mislocal-
ized distribution of the fusion protein
(see above and Table 1). Interestingly,
expression of Numb(AAAVFAGFF-
AA):EGFP also slightly perturbed the
orientation of the mitotic spindle.
1
Of
94 mitoses analyzed, 8.6% were either
oblique or orthogonal to the plane of
the neuroepithelium. In contrast, of
the 432 cell divisions analyzed in em-
bryos injected with numb(PTB
L
PRR
L
):egfp mRNA, only 1.4% were
oblique to the neuroepithelial plane.
Orthogonal cell divisions were never
observed following injection of mRNA
coding for Numb(PTB
L
PRR
L
):EGFP
(Table 1). The effect on protein local-
ization and spindle orientation was en-
hanced when mRNA encoding Numb-
(DRKAAKAAAKK):EGFP was injected;
in this version all residues with non-
polar side chains were replaced by ala-
nines (Fig. 7H,J and Supplementary
Movie 5). Of 156 mitoses analyzed,
26.9% showed a ubiquitous localiza-
tion of the fusion protein and 10.3% a
mislocalization of the Numb(DRKAA-
KAAAKK):EGFP crescent. With re-
spect to the orientation of the mitotic
spindle, 13.5% of the cells divided ei-
ther obliquely or orthogonally to the
plane of the epithelium (Table 1). It is
tempting to speculate that mislocal-
ization of Numb and misorientation of
the spindle are correlated. However,
this seems not to be the case, since
cells can be found that localize Numb-
(AAAVFAGFFAA):EGFP or Numb-
(DRKAAKAAAKK):EGFP uniformly
1
Spindle orientation was deduced from the orien
-
tation of the cleavage plane in dividing neuro-
epithelial cells.
Fig. 6. A–H: Comparison of Numb(PTB
L
PRR
L
):EGFP, Numb(PTB
S
PRR
L
):EGFP, Nbl:EGFP, and five different Numb:EGFP deletion constructs with respect
to their localization in dividing cells. Confocal micrographs showing a dorsal view of mitotic cells in the neural tube of embryos injected with the corresponding
mRNA. Anterior is towards the top and the midline/apical surface is approximately in the middle of each photograph. See text for details. B–D: Note that the
nucleus in non-dividing cells is free of the signal (arrow), whereas in mitotic cells (arrowhead) the signal is found in the entire cell, after the nuclear membrane
has broken down. F: Amino acids 1–196 (including the PTB
L
insertion) are efficiently localized to the basolateral cortex. G: Deletion of amino acids 2–12
impairs cortical and abolishes basolateral localization in dividing cells. H: Impairment of cortical localization is enhanced after deletion of amino acids 2–22.
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 941
Page 8
around the cell cortex but still divide
planar to the neuroepithelial plane,
while others divide obliquely or or-
thogonally to the plane of the epithe-
lium. Additionally, some cells mislo-
calize the crescent of the fusion
protein and then divide parallel to this
crescent, while others divide perpen-
dicular to it. Furthermore, not all of
the cells that showed a disturbed cell
division pattern also displayed an ab-
normal localization of Numb(AAAV-
FAGFFAA):EGFP or Numb(DRKAA-
KAAAKK):EGFP. The latter was also
true for the few cases of misoriented
cell divisions seen upon injection of
numb(PTB
L
PRR
L
):egfp mRNA (Table
1). It is worth noting that expression
of Numb(AAAVFAGFFAA):EGFP re-
sulted in a relatively low overall level
of GFP fluorescence and also in the
occurrence of a large number of dot-
like aggregates of bright fluorescence
that most probably represent debris
from apoptotic cells (arrow in Fig. 7G),
something that is not found upon in-
jection of other Numb or Numblike
constructs. However, with the excep-
tion of the construct shown in Figure
7I, in which all 11 amino acids were
replaced by alanines, none of the
tested constructs markedly impaired
localization to the cell membrane.
DISCUSSION
Basolateral Localization of
Numb(PTB
L
PRR
L
):EGFP
During Neurulation
The polarity of the neuroepithelial
cells matures in the course of neuru-
lation. In a previous study, we ana-
lyzed the localization of several apical
markers, such as ASIP/PAR-3, aPKC,
-Catenin, and ZO-1. We found a
gradual increase in the apical concen-
tration of these markers during neu-
rulation; they were diffusely distrib-
uted at the cell membrane in the
neural plate stage, became concen-
trated at the apical pole of neuroepi-
thelial cells in ventral levels of the
neural keel, where the two halves of
the neural primordium are in apposi-
tion, and then continued to concen-
trate apically in the neural rod and
tube stages (Geldmacher-Voss et al.,
2003). In particular, localization of
ASIP:EGFP to the apical pole first ap-
pears during the neural keel/rod
stage. However, when the cells round
up at prophase, the protein is redis-
tributed uniformly around the cell cor-
tex, until apical localization becomes
evident again in telophase, in both
daughter cells. In contrast, in the neu-
ral tube stage apical localization of
ASIP:EGFP persists during mitosis.
Thus, apical localization of ASIP:EGFP
is first observed prior to the basolateral
localization of Numb(PTB
L
PRR
L
):
EGFP, which occurs as neurulation
progresses from the neural rod to the
tube stage (Fig. 4C), indicating that es-
Fig. 7. A: Solution structure of the Drosophila Numb PTB domain-Nak peptide complex as
described by Zwahlen et al. (2000). The PTB domain is coloured in grey, whereas the Nak peptide
is coloured in blue. Residues that have direct contacts to target peptides are coloured in green.
Alignment of the Drosophila and the zebrafish PTB domain indicates that the additional 11 amino
acids of the zebrafish PTB
L
domain (arrow) would be inserted between amino acids 111 and 112
(highlighted in red). Figure generated using RasMol Vers. 2.7.2.1. B: Sequence of the PTB
L
insertion. Amino acids with acidic or basic side chains (D, R, K) are coloured in black. Amino acids
with nonpolar side chains (V, F, G) are coloured in grey. C: Dorsal view of the neural tube of an
embryo injected with numb(PTB
L
PRR
L
):egfp mRNA (compare with Fig. 5). D–I: Alanine-scanning
mutagenesis of the 11–amino acid insertion. Confocal micrographs of mitotic cells in the neural
tube of embryos injected with mRNA made from different mutagenized numb(PTB
L
PRR
L
):egfp
constructs. See text for details. J: A sequence of five time-lapse frames of two mitotic cells, in the
neural tube of embryos injected with numb(DRKAAKAAAKK):egfp mRNA, showing mislocalization
of the fusion protein and misorientation of the cell division. The transparent red line marks the
midline. Anterior is towards the top. The upper cell (arrowhead in 0) exhibits a mislocalized
Numb(DRKAAKAAAKK):EGFP crescent and subsequently divides with an orientation that is clearly
oblique to the plane of the neuroepithelium (arrowheads in 300”). The lower cell (asterisks in 0)
divides parallel to the neuroepithelial plane, as is the normal case in neural tube stage (asterisks in
430”), but localizes Numb(DRKAAKAAAKK):EGFP ubiquitously around the cell cortex. Note the
apical localization of the fusion protein also in the non-dividing cell (arrow in 130”).
942 REUGELS ET AL.
Page 9
tablishment of apicobasal polarity is a
prerequisite for Numb localization.
Numb has been demonstrated to be
asymmetrically localized to the apical
plasma membrane in mouse cortical
ventricular zone cells (Zhong et al.,
1996) and in rat retinal neuroepithe-
lial cells (Cayouette et al., 2001). At
present, we cannot explain the appar-
ent discrepancy between the apical lo-
calization of Numb in mouse and rat
and the basolateral localization of
Numb(PTB
L
PRR
L
):EGFP in ze
-
brafish. It might reflect species-spe-
cific differences. However, since Numb:
EGFP localization has not yet been
confirmed via, e.g., immunostaining,
the localization of the fusion protein
might not necessarily reflect the local-
ization of the endogenous Numb. On
the other hand, EGFP fusions have
been used successfully for the analysis
of many different proteins, including
Numb and its adapter protein PON in
Drosophila (Bellaiche et al., 2001; Ro-
egiers et al., 2001; Justice et al., 2003;
Mayer et al., 2005). In addition, baso-
lateral localization of Numb(PTB
L
PRR
L
):EGFP in zebrafish resembles
the basal localization of Numb in Dro-
sophila neuroblasts and, furthermore,
asymmetric localization of Numb to
the basal cortex of mitotic cells has
also been reported for chick neuroepi-
thelial and neural crest cells (Waka-
matsu et al., 1999, 2000).
Although we find Numb(PTB
L
PRR
L
):EGFP to be clearly asymmetri
-
cally localized in the neural tube, it
appears not to be asymmetrically seg-
regated to the daughter cells, since all
the cells in this stage divide with the
spindle oriented parallel to the plane
of the epithelium. On the other hand,
slight deviations from a strictly paral-
lel orientation of the mitotic spindle
could still result in an unequal segre-
gation of the protein. Since nothing is
yet known about the localization of
the endogenous Numb, it is conceiv-
able that it is localized differently to
the EGFP-fusion protein, e.g., is more
restricted to the basal pole, which
would also favour differential segrega-
tion. Das et al. (2003) showed that in
zebrafish embryonic retinas, dividing
cells always retain contact with the
basal surface by means of a thin basal
process that can be inherited asym-
metrically by the daughter cells. At
present, it is not clear whether the
neuroepithelial cells in the neural
tube also leave thin basal processes
behind when they round up apically
for division. Occasionally, cell divi-
sions can be observed in which a basal
process appears to be maintained
throughout mitosis (cell division
marked by arrow in Supplementary
Movie 4), whereas in other cases no
basal process is visible (cell division
marked by the asterisk in Supplemen-
tary Movie 4). Nevertheless, this
would be one means by which to asym-
metrically segregate Numb. Future
work (e.g., single cell labelling) should
help to elucidate this possibility. Fi-
nally, Kosodo and co-workers (2004)
demonstrated that in the mouse
embryonic neuroepithelium, vertical
cleavage planes of mitotic neurogenic
cells (resulting in planar cell divi-
sions) nonetheless allow an unequal
distribution of the apical plasma
membrane to the daughter cells and,
hence, asymmetric cell division. They
showed that the so-called cadherin
hole, a region corresponding to the
prominin-1-positive apical membrane
proper, can be segregated unequally
or equally to the daughter cells upon
cytokinesis, regardless of the vertical
orientation of the cleavage plane.
These results indicate that the orien-
tation of the cleavage plane, and
hence the orientation of the mitotic
spindle, is in itself an insufficient cri-
terion for predicting whether a divi-
sion will be symmetric or asymmetric
(Kosodo et al., 2004).
The Role of the First 196
Amino Acids of Numb for
Localization
It is striking that the zebrafish Numb
homologue needs an additional 11
amino acids within its PTB domain for
asymmetric localization, whereas the
Drosophila Numb protein becomes lo-
calized although it lacks such an in-
sertion. Analysis of the 3D solution
structure of the PTB domain reveals
that this insertion is located far away
from the classical PTB binding groove
as described by Zwahlen et al. (2000)
and Li et al. (1998; Fig. 7A). In prin-
ciple, there are two possibilities. The
11–amino acid insertion could either
create (or mask) a new (or existing)
binding site at, or close to, the site of
its insertion, or it could influence the
phosphotyrosine-binding groove itself,
by altering the three-dimensional
structure of the whole PTB domain
and modifying the binding specificity
of the groove. The results from our
mutagenesis experiments do not
clearly favour either of these possibil-
ities. However, in Drosophila, the
PTB domain is not necessary for
asymmetric localization. Proteins
made from constructs that lack the
entire domain, or at least significant
parts of it, are still asymmetrically lo-
calized when expressed in mitotic neu-
roblasts (Frise et al., 1996). Con-
versely, the PTB domain is required
for the protein to fulfil its biological
function, since these constructs are
completely non-functional when over-
expressed in the Drosophila SOP lin-
eage (Frise et al., 1996), whereas the
cortical localization of Numb is dis-
pensable for function (Knoblich et al.,
1997). Furthermore, Numb isoform 4
(PTB
S
PRR
S
) from mouse, when ex
-
pressed in Drosophila embryos, is
asymmetrically localized and capable
of rescuing the Drosophila numb mu-
tant phenotype (Zhong et al., 1996).
Additionally, misexpression of the rat
isoform 2 (PTB
L
PRR
S
) in the SOP
lineage causes cell fate transforma-
tions identical to those produced by
ectopic expression of Drosophila
Numb (Verdi et al., 1996). Since both
PTB forms are functional in Drosoph-
ila, it is likely that their classical
binding grooves are identical. There-
fore, the 11–amino acid insertion most
probably creates or modifies a new or
existing binding site within the PTB
domain, rather than modifying the
phospotyrosine binding groove itself
via conformational changes.
How might the PTB
L
insertion func
-
tion? We showed that the first 196
amino acids of Numb are sufficient for
localization and, furthermore, that
the 11–amino acid insertion in the
PTB domain is essential for localiza-
tion to the cortex, whereas amino ac-
ids 2–12 mediate basolateral localiza-
tion in the neural tube stage (Fig. 6).
For Drosophila, Knoblich et al. (1997)
reported that amino acids 1 to 227,
which correspond to amino acids
1–194 in zebrafish (compare Fig. 2),
are also sufficient for asymmetric lo-
calization. A construct comprising the
first 76 amino acids of Drosophila
Numb fused to -Gal localizes to the
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 943
Page 10
cell membrane, but fails to localize
asymmetrically. Conversely, deletion
of amino acids 1– 41 or 41–118 elimi-
nates both localization to the mem-
brane and asymmetric localization
during mitosis, indicating that the
first 76 amino acids are sufficient to
direct cortical but not asymmetric lo-
calization (Knoblich et al., 1997). In-
terestingly, the first 41 amino acids of
the Drosophila sequence are absent in
all vertebrate homologues identified
so far (compare Figs. 1 and 2). A
charge distributional analysis of Dro-
sophila Numb using the SAPS pro-
gram (Brendel et al., 1992) identifies a
single cluster of positive charges en-
compassing amino acids 14 –57. As
previously mentioned, insertion of the
11 amino acids in the zebrafish PTB
domain also generates a cluster of pos-
itive charges (amino acids 63– 85; cor-
responding to positions 107–129 in
Fig. 2). It is, therefore, tempting to
speculate that the insertion in ze-
brafish serves the same function as
the N-terminus in Drosophila, creat-
ing a positively charged surface that
mediates cortical localization, e.g., by
binding to membrane lipids. In fact,
Dho et al. (1999) showed that the
PTB
L
domain from mouse Numb has a
slightly higher affinity for phosphati-
dylinositol 4-phosphate than the
PTB
S
form. However, in the deletion
constructs shown in Figure 7F and G,
the positive charge is considerably re-
duced, with no obvious effect on corti-
cal localization. This indicates that it
is not just the positive charge that fa-
cilitates localization to the membrane;
the structure and/or the presence of
specific residues (compare Fig. 7H)
probably also play a role. Since a ze-
brafish fusion protein consisting of
amino acids 55–100 (positions 99 –144
in Fig. 2) alone is not cortically local-
ized but instead remains in the cytosol
and the nucleus (data not shown),
membrane localization may also in-
volve interactions with other domains
within the first 196 amino acids.
Our deletion analysis showed that
amino acids 2–12 are necessary for ba-
solateral localization of Numb(PTB
L
PRR
L
):EGFP in the neural tube stage
(Fig. 6). Although this region is highly
conserved among all vertebrate Numb
homologues, database analyses have
not revealed any related sequence mo-
tifs in other proteins. Very recently,
Bhalerao et al. (2005) reported for
Drosophila that asymmetric localiza-
tion of Numb in dividing neuroblasts
is affected in a mutant in which the
serine at position 52 is changed to
phenylalanine (S52F). Interestingly,
this amino acid is not conserved in the
vertebrate Numb proteins; it lies
within a short stretch of three amino
acids that are present in Drosophila
Numb but are missing in all verte-
brate homologues (Fig. 2).
Nevertheless, since amino acids
43–56 of Drosophila Numb correspond
to amino acids 2–12 of the zebrafish
homologues (Fig. 2), identification of
this mutation supports our finding
that these amino acids are important
for asymmetric localization of Numb.
In addition, we recently obtained pre-
liminary results that indicate that
amino acids 176–196 of zebrafish
Numb might harbour a second asym-
metric localization domain (data not
shown). This would be compatible
with data from Knoblich et al. (1997),
which identify amino acids 217–227
as a possible site of a domain for
asymmetric localization in Drosophila
Numb. Thus, basolateral localization
of Numb:EGFP in the neural tube of
the zebrafish seems to depend on
three regions within the Numb do-
main: an 11–amino acid insertion
within the PTB domain, an asymmet-
ric localization domain at the N-termi-
nal end of the protein and, most prob-
ably, a second asymmetric localization
domain C-terminal to the PTB do-
main.
On the Function of Numb
Although great progress has been
made in uncovering the functions of
Numb proteins during vertebrate de-
velopment, their precise roles remain
controversial. Whereas some studies
propose that they promote neuronal
differentiation during neurogenesis
(Verdi et al., 1996; Wakamatsu et al.,
1999; Zilian et al., 2001), others have
reported that they are instead re-
quired for progenitor cell maintenance
(Petersen et al., 2002, 2004), or that
they promote the progenitor fate dur-
ing early neurogenesis but switch to
promoting neuronal fates at later
stages, probably by expressing differ-
ent isoforms (Verdi et al., 1996, 1999;
Shen et al., 2002). A function for
Numb has also been reported, inter
alia, in sensory axon arborization in
mice and in cell fate choice in rat ret-
inal neuroepithelial cells (Huang et
al., 2005; Cayouette and Raff, 2003). It
has been shown that expression of a
PTB
S
isoform in rat PC12 cells in
-
creases their vulnerability to cell
death induced by amyloid -peptide
by disrupting calcium homeostasis, an
effect that is potentially relevant to
the pathogenesis of Alzheimer’s dis-
ease (Chan et al., 2002). Furthermore,
novel roles for Numb isoforms have
also been described in the regulation
of neuronal differentiation in response
to nerve growth factor (NGF) and in
the dependency of neuronal cells on
NGF for survival (Pedersen et al.,
2002). Apart from their roles in the
nervous system, distinct expression
patterns of Numb isoforms have also
been reported in the endocrine lineage
of the developing pancreas (Yoshida et
al., 2003). However, due to the pres-
ence of different isoforms and its abil-
ity to interact with various proteins, it
is likely that Numb carries out differ-
ent roles in different developmental
contexts.
Numb is a negative regulator of
Notch signalling. Several groups have
shown that it can physically interact
with Notch (Spana and Doe, 1996;
Frise et al., 1996; Guo et al., 1996;
Wakamatsu et al., 1999; Zhong et al.,
1996). A recent model for Drosophila
ganglion mother cells proposes that
Numb inhibits Notch activity in the
“B” cell by blocking the ability of San-
podo, a four-pass transmembrane pro-
tein, to localize to the cell membrane.
In the “A” cell, which is devoid of
Numb, Sanpodo can localize to the
membrane, where it promotes Notch
signalling and, therefore, adoption of
the “A” cell fate (O’Connor-Giles and
Skeath, 2003). However, no verte-
brate homologue of Sanpodo has been
identified so far, and co-localization
experiments following injection of
mRNA coding for zebrafish Numb and
the intra-cellular domain of Notch, as
well as yeast two-hybrid interaction
studies, have not provided any evi-
dence for direct binding interactions
between Numb and Notch (data not
shown). Numb has been shown to be
involved in clathrin-mediated endocy-
tosis both in mammalian cells and in
Drosophila (Santolini et al., 2000;
944 REUGELS ET AL.
Page 11
Berdnik et al., 2002; Smith et al.,
2004). In Drosophila, for example, it
has been proposed that the regulation
of Notch involves Numb-dependent
endocytosis of Sanpodo (Hutterer and
Knoblich, 2005). Our finding that
Numb(PTB
L
PRR
L
):EGFP labelling is
found within vesicle-like structures in
non-dividing cells throughout neuru-
lation would be in accordance with
such a role of Numb in endocytosis.
Therefore, future studies should focus
on a more detailed analysis of this is-
sue.
Surprisingly, mutation of the 11–
amino acid insertion within the PTB
L
domain resulted in a considerable rise
in the incidence of misoriented cell di-
visions (Table 1). It is possible, and
perhaps likely, that this is a rather
unrelated effect, resulting from an al-
tered binding specificity of the protein,
and does not reflect a true biological
function of Numb in controlling the
orientation of the mitotic spindle.
Therefore, additional experiments
should be performed, in order to verify
the relevance of these results. So far,
little is known about the function of
Numb during neurogenesis in ze-
brafish. As expected from the expres-
sion pattern, we find Numb(PTB
L
PRR
L
):EGFP also to be basolaterally
localized in the developing eye (data
not shown), indicating a role for Numb
in eye development. In order to iden-
tify novel binding partners, we have
performed a yeast two-hybrid screen
and are now analyzing the potential
interaction partners (B. Boggetti, un-
published data). Among other experi-
ments, these studies will help to elu-
cidate the role of the different Numb
homologues during zebrafish develop-
ment.
EXPERIMENTAL
PROCEDURES
Animals
Zebrafish embryos were obtained from
spontaneous spawnings. Adult fish
were kept at 28.5°C in a 14-hr light/
10-hr dark cycle. The embryos were
staged according to Kimmel et al.
(1995). For time-lapse movies, em-
bryos were kept in a modified version
of Embryo Rearing Medium (ERM)
(Westerfield, 1994). Embryos were an-
aesthetized with MS222, manually
dechorionated and placed in holes cut
in a thin layer of 1.7% agarose gel (in
ERM) supported by a coverslip (see
Cooper et al., 1999, for a detailed de-
scription of the procedure).
Multilevel Confocal Time-
Lapse Imaging and
Statistical Analysis
An LSM 410 confocal microscope
(Zeiss) attached to an inverted Zeiss
Diavert microscope with 40 and 63
immersion objectives was used to col-
lect up to seven different vertical
stacks of images (z-series) at intervals
of 60 to 120 sec, each stack separated
by between 5 to 7 m from the next,
depending on the experiment. Z-series
were transferred to an Intel PC for
image processing using ImageJ and
Adobe Photoshop 6.0. The orientation
of cell divisions in neural tube stage
was determined by first labelling the
midline of the neural tube and defin-
ing it as and then drawing a line
between the separating daughter cells
at anaphase (perpendicular to the
plane of cleavage) and projecting it to
the midline. Cell divisions were clas-
sified as being parallel, oblique, and
perpendicular, based on measured an-
gles of 45°, and 90°, respectively,
with a deviation of 22.5° for each
class.
Cloning of Zebrafish Numb
and Numblike Homologues
Based on the published sequences for
Drosophila melanogaster numb and
Mus musculus m-numb, we designed
the degenerate PCR primers 5-
CAGTGGCAGACAGAYGARGARG-3
and 5-GCAGTGACNCCRCAYTC-3
to amplify a 468-basepair (bp) frag-
ment encoding amino acids 24 –179 of
a predicted zebrafish numb homo-
logue. Subsequently, a genomic li-
brary (Mobitec, Go¨ttingen, Germany)
was screened using a 327-bp subfrag-
ment (amplified with specific PCR
primers) as a probe, and a single
genomic clone was obtained, starting
in an intron 3 kb upstream of codon 79
and ending within codon 678. Based
on this sequence, the PCR primers 5-
CTGTGCGCCGGATCG-3 and 5-
CAAACACATTGGCCACCATC-3 were
designed to amplify a 1.3-kb cDNA
fragment encoding amino acids 117–
554. Several RACE and RT-PCR ex-
periments were performed to extend
the cDNA in 5 and 3 directions and,
finally, appropriate clones were com-
bined to give a continuous cDNA se-
quence of 2,373-bp encoding amino ac-
ids 1 to 680 of the zebrafish Numb
isoform 1 (PTB
L
PRR
L
, accession
number AY583653).
The PCR primers 5-AGAGGCTG-
TACGCGGAGGC-3 and 5-CGTCTC-
GGCAGATGTAGGAG-3 were used
for RT-PCR experiments to clone a
314-bp and a 281-bp fragment repre-
senting the PTB
L
and the PTB
S
iso
-
forms, respectively. A cDNA encoding
the 669 amino acids of the Numb iso-
form 3 (PTB
S
PRR
L
, accession number
AY583654) was then constructed by
replacing the 217-bp SphI fragment
from the Numb(PTB
L
PRR
L
) clone
with the corresponding PTB
S
frag
-
ment from the RT-PCR clone.
The PCR primers 5-CCTTCAGG-
GGCTTCCCGGCG-3 and 5-CCG-
GGGCAGGGAAGGGCTG-3 together
with the nested primers 5-CTCTCT-
GTGCACTCAGATCACGTC-3 5-GG-
GGTGCGCTTGTGAGCGGAGGC-3
were used for RT-PCR experiments to
clone a 473-bp and a 204-bp fragment,
respectively, of the putative PRR
S
iso
-
form. However, only fragments repre-
senting the PRR
L
isoform were ob
-
tained with cDNA from embryos up to
24 hr post fertilization.
Screening of the database of the
Washington University zebrafish EST
Project for putative Numb homo-
logues identified the IMAGE consor-
tium clone 2641374 (Lennon et al.,
1996). Sequencing of the clone re-
vealed that this clone represents the
full-length cDNA of the zebrafish ho-
mologue of the gene numblike from
Mus musculus (accession number
DQ022744).
RT-PCR
Total RNA was extracted from differ-
ent developmental stages with Trizol
(Invitrogen, Carlsbad, CA) according
to standard protocols. First-strand
synthesis was performed with the In-
vitrogen Superscript first-strand syn-
thesis system according to the manu-
facturer’s guidelines. To amplify a
525-bp fragment specific for numb 1
(PTB
L
PRR
L
), the following PCR
primers were used: 5-GACAGAAAG-
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 945
Page 12
GTCTTCAAGGGCTTCTTTAAA-3,
5-GGGGCCCGCCAGAGTCACCAC-
CACTGGG-3. For amplification of a
496-bp numb 3 (PTB
S
PRR
L
)-specific
fragment, the oligonucleotides 5-
GACGGCGGGTAAGAAGGCTGTGC-
GAGCG-3 and 5-GGGGCCCGCCA-
GAGTCACCACCACTGGG-3 were
used. As a control, a 558-bp fragment
of -actin was amplified using the
PCR primers 5-TGTTTTCCCCTC-
CATTGTTGG-3 and 5-TTCTCCTT-
GATGTCACGGAC-3.
A total of 30 cycles was used for all
PCR experiments.
GFP Fusion Proteins and
mRNA Injections
For mRNA injections, eight different
constructs were synthesized that en-
coded variants of Numb fused to the
enhanced variant of GFP (EGFP, Cor-
mack et al., 1996) and one that en-
coded Numblike fused to EGFP. An
expression vector (pCS2/EGFP) was
constructed by cloning the Eco47III-
XhoI fragment encoding the enhanced
GFP variant from the pEGFP-C1 vec-
tor (Clontech Laboratories, Inc., Palo
Alto, CA; Accession No. U55763) into
the StuIXhoI-digested pCS2 vec-
tor (Turner and Weintraub, 1994).
numb:egfp constructs were made by
amplifying the corresponding coding
region of zebrafish Numb by PCR from
the full-length cDNA clone using ap-
propriate primer pairs that carry arti-
ficial restriction sites, to allow the
cloning of the PCR fragment in-frame
into the pCS2/EGFP vector, either
N- or C-terminal to the EGFP. Con-
structs and primers were as follows:
numb(PTB
L
PRR
L
):egfp: Nb-5Bam
(BamHI): 5-GCTGGATCCGGAGGAG-
GTTCAGCGATG-3, Nb-3Eco2 (EcoRI):
5-GTTGAATTCGTAGCTGGATCT-
CGAATGTTTTG-3; egfp:numb(PRR
L
1-306): Nb-GFP5Bam (BglII): 5-CTG-
GGATCCACCATGCAGCGCAAGTCT-
GAC-3, Nb-Synth3 (XbaI); 5-TCT-
AGATTAAAGTTCGATCTCGAATGT-
TTTGTGGAGCTC-3
numb(PTB
L
308 680):egfp: Nb-
5Bam (BamHI), Nb-GFP3Eco (EcoRI):
5-CTGAATTCTGGAGGGCAGCTCG-
TTCATG-3;
numb(PTB
L
197– 680):egfp: Nb-
5Bam (BamHI), Nb-GFP3Eco2 (EcoRI):
5-TGGGAATTCGGAAAGATCCCT-
CTCGTGTG-3;
numb(PTB
L
2–12; 197– 680):egfp:
Nb-5Bam13 (BamHI), Nb-GFP3Eco2
(EcoRI);
numb(PTB
L
2–22; 197– 680):egfp:
Nb-5Bam23 (BamHI): 5-AGTGGA-
TCCGGATGCACCAGTGGCAGACCG-
ATG-3, Nb-GFP3Eco2 (EcoRI).
For numb(PTB
S
PRR
L
):egfp and
numb(PTB
S
186 669):egfp, respec
-
tively, PTB
S
variants were con
-
structed by replacing the 217-bp SphI
fragment from the PTB
L
clone with
the corresponding 184-bp SphI frag-
ment from the Numb(PTB
S
PRR
L
)
full-length clone (see above). The
numblike:egfp construct was made by
amplifying the corresponding coding
region from the numblike full-length
cDNA with the primer pair 5-
GAGAGATCTGAGATGAATAAGCT-
GCGTC-3 and 5-CTTGAATTCG-
TAGTTCAATCTCAAAAGTC -3 and
cloning it into the BamHI/EcoRI site
of the pCS2/EGFP vector. Capped
RNA was synthesized in vitro by tran-
scription with SP6 polymerase from
the constructs described above. mRNAs
were injected in 5-nl aliquots (1.5 to
2.5 ng) into the yolk of wild-type zy-
gotes.
In Vitro Mutagenesis
Using the QuikChange site-directed
mutagenesis kit from Stratagene (La
Jolla, CA), an alanine scanning mu-
tagenesis (Cunningham and Wells,
1989; Bass et al., 1991; Wells, 1991)
was performed to mutagenize the 11–
amino acid insertion within the PTB
L
domain. Reactions were performed ac-
cording to the manufacturer’s guide-
lines. Based on the numb(PTB
L
PRR
L
):egfp pCS2/EGFP vector, four
different constructs were synthesized
using the following oligonucleotides:
numb(AAAVFKGFFKK):egfp:5-GAA-
GCTGAAGACGGCCGCAGCGGTCTT-
CAAGGGCTTC-3 and 5-GAAGCC-
CTTGAAGACCGCTGCGGCCGTCTT-
CAGCTTC-3;
numb(DRKVFAGFFAA):egfp:5-CAG-
AAAGGTCTTCGCGGGCTTCTTTGC-
AGCAGCGGGTAAGAAGG-3 and 5-
CCTTCTTACCCGCTGCTGCAAAGA-
AGCCCGCGAAGACCTTTCTG-3;
numb(DRKV___FFKK):egfp:5-GAC-
GGACAGAAAGGTCTTCTTTAAAAA-
AGCGG-3 and 5-CCGCTTTTTTAAA-
GAAGACCTTTCTGTCCGTC-3;
numb(DRKAAKAAAKK):egfp:5-
CGGACAGAAAGGCCGCCAAGGC-
CGCCGCTAAAAAAGCGGG-3 and
5-CCCGCTTTTTTAGCGGCGGC-
CTTGGCGGCCTTTCTGTCCG-3.
In a second step, based on clone
numb(AAAVFKGFFKK):egfp, the clone
numb(AAAVFAGFFAA):egfp was con-
structed using the oligonucleotides
5-CCGCAGCGGTCTTCGCGGGCT-
TCTTTGCAGCAGCGGGTAAGA-
AGG-3 and 5-CCTTCTTACCCG-
CTGCTGCAAAGAAGCCCGCGAAGA-
CCGCTGCGG-3.
In a third step, the clone num-
b(AAAVFAGFFAA):egfp itself was
mutagenized again, to yield the clone
numb(AAAAAAAAAAA):egfp, using the
oligonucleotides 5-CGGCCGCAGCG-
GCCGCCGCGGCCGCCGCTGCAG-
CAGCGGG-3 and 5-CCCGCTGCTG-
CAGCGGCGGCCGCGGCGGCCGCT-
GCGGCCG-3.
Phylogenetic Analysis
The Numb domains of different Numb
homologues from various species were
used for a maximum likelihood analy-
sis. The sequences were pre-aligned
with ClustalW and then refined by vi-
sual inspection using the multiple
alignment sequence editor SeaView
(Galtier et al., 1996). The analysis was
restricted to those positions that were
alignable (dots in Fig. 2) and (in order
to reduce the number of sequences), in
the case of multiple isoforms, to the
PTB
S
form (isoform 3). A complex evo
-
lutionary model with gamma-distri-
bution was chosen to reduce long
branch attraction artefacts (LBA).
The evolutionary model best fitting
the data according to the Akaike In-
formation Criterion (AIC) was deter-
mined with the program ProtTest
1.2.6 (Abascal et al., 2005; Drummond
and Strimmer, 2001). ProtTest 1.2.6
proposed the JTTIG model as the
evolutionary model best fitting the
data (Jones et al., 1992). The molecu-
lar phylogenetic analysis under the
maximum likelihood criterion (200
bootstrap replicates) was performed
with the software Phyml 2.4.4 (Guin-
don and Gascuel, 2003).
Statistical Analysis of
Protein Sequences
Statistical analysis of the protein se-
quences was performed using the
946 REUGELS ET AL.
Page 13
SAPS program (Brendel et al., 1992)
available from the website of the
European Bioinformatics Institute
(http://www.ebi.ac.uk).
ACKNOWLEDGMENTS
This report is dedicated to the mem-
ory of Jose´ Campos-Ortega, whose sci-
entific passion and brilliance together
with his personal support were a con-
stant inspiration. We thank Elisabeth
Knust, Siegfried Roth, John Chandler,
and Paul Hardy for a critical reading
of the manuscript; Iris Riedl, Christel
Schenkel, and Thomas Wagner for ex-
pert technical assistance; Kerstin
Hoef-Emden for her help with the
maximum likelihood analysis; Bruce
Apple, Sumio Sugano, and Koichi
Kawakami for sharing their cDNA-li-
braries; and our colleagues for helpful
discussions.
REFERENCES
Abascal F, Zardoya R, Posada D. 2005.
ProtTest: selection of best-fit models of
protein evolution. Bioinformatics 21:
2104 –2105.
Artavanis-Tsakonas S, Rand MD, Lake RJ.
1999. Notch signaling: cell fate control
and signal integration in development.
Science 284:770 –776.
Bass SH, Mulkerrin MG, Wells JA. 1991. A
systematic mutational analysis of hor-
mone-binding determinants in the hu-
man growth hormone receptor. Proc Natl
Acad Sci USA 88:4498 –502.
Bellaiche Y, Gho M, Kaltschmidt JA,
Brand AH, Schweisguth F. 2001. Friz-
zled regulates localization of cell-fate de-
terminants and mitotic spindle rotation
during asymmetric cell division. Nat Cell
Biol 3:50 –57.
Berdnik D, Torok T, Gonzalez-Gaitan M,
Knoblich JA. 2002. The endocytic protein
alpha-Adaptin is required for numb-me-
diated asymmetric cell division in Dro-
sophila. Dev Cell 3:221–231.
Bhalerao S, Berdnik D, Torok T, Knoblich
JA. 2005. Localization-dependent and
-independent roles of numb contribute to
cell-fate specification in Drosophila.
Curr Biol 15:1583–1590.
Brendel V, Bucher P, Nourbakhsh I, Blais-
dell BE, Karlin S. 1992. Methods and
algorithms for statistical analysis of pro-
tein sequences. Proc Natl Acad Sci USA
89:2002–2006.
Campos-Ortega JA. 1995. Genetic mecha-
nisms of early neurogenesis in Drosoph-
ila melanogaster. Mol Neurobiol 10:75–
89.
Cayouette M, Raff M. 2003. The orienta-
tion of cell division influences cell-fate
choice in the developing mammalian ret-
ina. Development 130:2329 –2339.
Cayouette M, Whitmore AV, Jeffery G,
Raff M. 2001. Asymmetric segregation of
numb in retinal development and the in-
fluence of the pigmented epithelium.
J Neurosci 21:5643–5651.
Chan SL, Pedersen WA, Zhu H, Mattson
MP. 2002. Numb modifies neuronal vul-
nerability to amyloid beta-peptide in an
isoform-specific manner by a mechanism
involving altered calcium homeostasis:
implications for neuronal death in Alz-
heimer’s disease. Neuromol Med 1:55–
67.
Chien CT, Wang S, Rothenberg M, Jan LY,
Jan YN. 1998. Numb-associated kinase
interacts with the phosphotyrosine bind-
ing domain of Numb and antagonizes the
function of Numb in vivo. Mol. Cell Biol.
18:598 –607.
Concha ML, Adams, RJ. 1998. Oriented
cell divisions and cellular morphogenesis
in the zebrafish gastrula and neurula: a
time-lapse analysis. Development 125:
983–994.
Cooper MS, D’Amico LA, Henry CA. 1999.
Confocal microscopic analysis of morpho-
genetic movements. Methods Cell Biol
59:179 –204.
Cormack BP, Valdivia RH, Falkow S. 1996.
FACS-optimized mutants of the green
fluorescent protein (GFP). Gene 173:33–
38.
Cunningham BC, Wells JA. 1989. High-
resolution epitope mapping of hGH-re-
ceptor interactions by alanine-scanning
mutagenesis. Science 244:1081–1085.
Das T, Payer B, Cayouette M, Harris WA.
2003. In vivo time-lapse imaging of cell
divisions during neurogenesis in the de-
velopingzebrafishretina. Neuron37:597–
609.
Dho SE, Jacob S, Wolting CD, French MB,
Rohrschneider LR, McGlade CJ. 1998.
The mammalian numb phosphotyrosine-
binding domain. Characterization of
binding specificity and identification of a
novel PDZ domain-containing numb
binding protein, LNX. J Biol Chem 273:
9179 –9187.
Dho SE, French MB, Woods SA, McGlade
CJ. 1999. Characterization of four mam-
malian numb protein isoforms. Identifi-
cation of cytoplasmic and membrane-as-
sociated variants of the phosphotyrosine
binding domain. J Biol Chem 274:33097–
33104.
Drummond A, Strimmer K. 2001. PAL: an
object-oriented programming library for
molecular evolution and phylogenetics.
Bioinformatics 17:662–663.
Edlund T, Jessell TM. 1999. Progression
from extrinsic to intrinsic signaling in
cell fate specification: a view from the
nervous system. Cell 96:211–224.
Frise E, Knoblich JA, Younger-Shepherd
S, Jan LY, Jan YN. 1996. The Drosophila
Numb protein inhibits signaling of the
Notch receptor during cell-cell interac-
tion in sensory organ lineage. Proc Natl
Acad Sci USA 93:11925–11932.
Galtier N, Gouy M, Gautier C. 1996. SEA-
VIEW and PHYLO_WIN: Two graphic
tools for sequence alignment and molec-
ular phylogeny. Comp Appl Biosci 8:275–
282.
Geldmacher-Voss B, Reugels AM, Pauls S,
Campos-Ortega JA. 2003. A 90-degree
rotation of the mitotic spindle changes
the orientation of mitoses of zebrafish
neuroepithelial cells. Development 130:
3767–3780.
Gho M, Schweisguth F. 1998. Frizzled sig-
nalling controls orientation of asymmet-
ric sense organ precursor cell divisions in
Drosophila. Nature 393:178 –181.
Guindon S, Gascuel O. 2003. A simple, fast,
and accurate algorithm to estimate large
phylogenies by maximum likelihood.
Syst Biol 52:696 –704.
Guo M, Jan LY, Jan YN. 1996. Control of
daughter cell fates during asymmetric
division: interaction of Numb and Notch.
Neuron 17:27–41.
Hinds JW, Ruffett TL. 1971. Cell prolifer-
ation in the neural tube: an electron mi-
croscopic and golgi analysis in the mouse
cerebral vesicle. Z Zellforsch Mikrosk
Anat 115:226 –264.
Hirata J, Nakagoshi H, Nabeshima Y,
Matsuzaki F. 1995. Asymmetric segrega-
tion of the homeodomain protein Pros-
pero during Drosophila development.
Nature 377:627–630.
Horvitz HR, Herskowitz I. 1992. Mecha-
nisms of asymmetric cell division: two Bs
or not two Bs, that is the question. Cell
68:237–55.
Huang EJ, Li H, Tang AA, Wiggins AK,
Neve RL, Zhong W, Jan LY, Jan YN.
2005. Targeted deletion of numb and
numblike in sensory neurons reveals
their essential functions in axon ar-
borization. Genes Dev 19:138 –151.
Hutterer A, Knoblich JA. 2005. Numb and
alpha-Adaptin regulate Sanpodo endocy-
tosis to specify cell fate in Drosophila
external sensory organs. EMBO Rep 6:
836 –842.
Jan YN, Jan LY. 1998. Asymmetric cell
division. Nature 392:775–778.
Jones DT, Taylor WR, Thornton JM. 1992.
The rapid generation of mutation data
matrices from protein sequences. Com-
put Appl Biosci 8:275–282.
Justice N, Roegiers F, Jan LY, Jan YN.
2003. Lethal giant larvae acts together
with numb in notch inhibition and cell
fate specification in the Drosophila adult
sensory organ precursor lineage. Curr
Biol 13:778 –783.
Kaltschmidt JA, Davidson CM, Brown NH,
Brand AH. 2000. Rotation and asymme-
try of the mitotic spindle direct asym-
metric cell division in the developing
central nervous system. Nature Cell Biol
2:7–12.
Kimmel CB, Warga RM, Schilling TF.
1990. Origin and organization of the ze-
brafish fate map. Development 108:581–
594.
Kimmel CB, Warga RM, Kane, DA. 1994.
Cell cycles and clonal strings during for-
mation of the zebrafish central nervous
system. Development 120:265–276.
Kimmel CB, Ballard WW, Kimmel SR, Ul-
lmann B, Schilling TF. 1995. Stages of
LOCALIZATION OF NUMB:EGFP DURING NEURULATION 947
Page 14
embryonic development of the zebrafish.
Dev Dyn 203:253–310.
Knoblich JA, Jan LY, Jan YN. 1997. The N
terminus of the Drosophila Numb pro-
tein directs membrane association and
actin-dependent asymmetric localiza-
tion. Proc Natl Acad Sci USA 94:13005–
13010.
Kosodo Y, Roper K, Haubensak W, Marz-
esco AM, Corbeil D, Huttner WB. 2004.
Asymmetric distribution of the apical
plasma membrane during neurogenic di-
visions of mammalian neuroepithelial
cells. EMBO J 2311:2314 –2324.
Lennon G, Auffray C, Polymeropoulos M,
Soares MB. 1996. The I.M.A.G.E. Con-
sortium: an integrated molecular analy-
sis of genomes and their expression.
Genomics 33:151–152.
Letunic I, Copley RR, Schmidt S, Ciccarelli
FD, Doerks T, Schultz J, Ponting CP,
Bork P. 2004. SMART 4.0: towards
genomic data integration. Nucleic Acids
Res. 32:D142–144.
Li SC, Zwahlen C, Vincent SJ, McGlade
CJ, Kay LE, Pawson T, Forman-Kay JD.
1998. Structure of a Numb PTB domain-
peptide complex suggests a basis for di-
verse binding specificity. Nat Struct Biol
5:1075–1083.
Lu B, Rothenberg M, Jan LY, Jan YN.
1998. Partner of Numb colocalizes with
Numb during mitosis and directs Numb
asymmetric localization in Drosophila
neural and muscle progenitors. Cell 95:
225–235.
Mayer B, Emery G, Berdnik D, Wirtz-Peitz
F, Knoblich JA. 2005. Quantitative anal-
ysis of protein dynamics during asym-
metric cell division. Curr Biol 15:1847–
1854.
McConnell SK. 1991. The generation of
neuronal diversity in the central nervous
system. Annu Rev Neurosci 14:269 –300.
O’Connor-Giles KM, Skeath JB. 2003.
Numb inhibits membrane localization of
Sanpodo, a four-pass transmembrane
protein, to promote asymmetric divisions
in Drosophila. Dev Cell 5:231–243.
Papan C, Campos-Ortega JA. 1994. On the
formation of the neural keel and neural
tube in the zebrafish Danio (Brachy-
danio) rerio. Rouxs Arch Dev Biol 203:
178 –186.
Papan C, Campos-Ortega, JA. 1997. A
clonal analysis of spinal cord develop-
ment in the zebrafish. Dev Genes Evol
207:71–81.
Pauls S, Geldmacher-Voss B, Campos-Or-
tega JA. 2001. A zebrafish histone vari-
ant H2A.F/Z and a transgenic H2A.F/Z-
GFP fusion protein for in vivo studies of
embryonic development. Dev Genes Evol
211:603–610.
Pedersen WA, Chan SL, Zhu H, Abdur-
Rahman LA, Verdi JM, Mattson MP.
2002. Numb isoforms containing a short
PTB domain promote neurotrophic fac-
tor-induced differentiation and neurotro-
phic factor withdrawal-induced death of
PC12 Cells. J Neurochem 82:976 –986.
Petersen PH, Zou K, Hwang JK, Jan YN,
Zhong W. 2002. Progenitor cell mainte-
nance requires numb and numblike dur-
ing mouse neurogenesis. Nature 419:929
934.
Petersen PH, Zou K, Krauss S, Zhong W.
2004. Continuing role for mouse Numb
and Numbl in maintaining progenitor
cells during cortical neurogenesis. Nat
Neurosci 7:803–811.
Qin H, Percival-Smith A, Li C, Jia CY, Gloor
G, Li SS. 2004. A novel transmembrane
protein recruits numb to the plasma mem-
brane during asymmetric cell division.
J. Biol. Chem. 279:11304 –11312.
Roegiers F, Younger-Shepherd S, Jan LY,
Jan YN. 2001. Two types of asymmetric
divisions in the Drosophila sensory or-
gan precursor cell lineage. Nat Cell Biol
3:58 –67.
Salcini AE, Confalonieri S, Doria M, San-
tolini E, Tassi E, Minenkova O, Cesareni
G, Pelicci PG, Di Fiore PP. 1997. Binding
specificity and in vivo targets of the EH
domain, a novel protein-protein interac-
tion module. Genes Dev 11:2239 –2249.
Santolini E, Puri C, Salcini AE, Gagliani
MC, Pelicci PG, Tacchetti C, Di Fiore PP.
2000. Numb is an endocytic protein.
J Cell Biol 151:1345–1352.
Schmitz B, Papan C, Campos-Ortega JA.
1993. Neurulation in the anterior trunk
region of the zebrafish Brachydanio re-
rio. Rouxs Arch Dev Biol 202:250 –259.
Schultz J, Milpetz F, Bork P, Ponting CP.
1998. SMART, a simple modular archi-
tecture research tool: identification of
signaling domains. Proc Natl Acad Sci
USA 95:5857–5864.
Shen Q, Zhong W, Jan YN, Temple S. 2002.
Asymmetric Numb distribution is criti-
cal for asymmetric cell division of mouse
cerebral cortical stem cells and neuro-
blasts. Development 129:4843–4853.
Smith CA, Dho SE, Donaldson J, Tepass U,
McGlade CJ. 2004. The cell fate determi-
nant numb interacts with EHD/Rme-1
family proteins and has a role in endo-
cytic recycling. Mol Biol Cell 15:3698
3708.
Spana EP, Doe CQ. 1996. Numb antago-
nizes Notch signaling to specify sibling
neuron cell fates. Neuron 17:21–26.
Turner DL, Weintraub H. 1994. Expres-
sion of achaete-scute homolog 3 in Xeno-
pus embryos converts ectodermal cells to
a neural fate. Genes Dev. 8:1434 –1447.
Uemura T, Shepherd S, Ackerman L, Jan
LY, Jan YN. 1989. numb, a gene re-
quired in determination of cell fate dur-
ing sensory organ formation in Drosoph-
ila embryos. Cell 58:349 –360.
Verdi JM, Schmandt R, Bashirullah A, Jacob
S, Salvino R, Craig CG, Program AE, Lip-
shitz HD, McGlade CJ. 1996. Mammalian
NUMB is an evolutionarily conserved sig-
naling adapter protein that specifies cell
fate. Curr Biol 6:1134 –1145.
Verdi JM, Bashirullah A, Goldhawk DE,
Kubu CJ, Jamali M, Meakin SO, Lip-
shitz HD. 1999. Distinct human NUMB
isoforms regulate differentiation vs. pro-
liferation in the neuronal lineage. Proc
Natl Acad Sci USA 96:10472–10476.
Wakamatsu Y, Maynard TM, Jones SU,
Weston JA. 1999. NUMB localizes in the
basal cortex of mitotic avian neuroepi-
thelial cells and modulates neuronal dif-
ferentiation by binding to NOTCH-1.
Neuron 23:71–81.
Wakamatsu Y, Maynard TM, Weston JA.
2000. Fate determination of neural crest
cells by NOTCH-mediated lateral inhibi-
tion and asymmetrical cell division dur-
ing gangliogenesis. Development 127:
2811–2821.
Wells JA. 1991. Systematic mutational an-
alyzes of protein-protein interfaces.
Methods Enzymol 202:390 –411.
Westerfield M. 1994. The zebrafish book: a
guide for laboratory use of the zebrafish
(Brachydanio rerio). Edition 2.1. Eu-
gene: University of Oregon Press.
Yoshida T, Tokunaga A, Nakao K, Okano
H. 2003. Distinct expression patterns of
splicing isoforms of mNumb in the endo-
crine lineage of developing pancreas. Dif-
ferentiation 71:486 –495.
Zhong W, Feder JN, Jiang MM, Jan LY,
Jan YN. 1996. Asymmetric localization
of a mammalian numb homolog during
mouse cortical neurogenesis. Neuron 17:
43–53.
Zhong W, Jiang MM, Weinmaster G, Jan
LY, Jan YN. 1997. Differential expres-
sion of mammalian Numb, Numblike
and Notch1 suggests distinct roles dur-
ing mouse cortical neurogenesis. Devel-
opment 124:1887–1897.
Zilian O, Saner C, Hagedorn L, Lee HY,
Sauberli E, Suter U, Sommer L, Aguet
M. 2001. Multiple roles of mouse Numb
in tuning developmental cell fates. Curr
Biol 11:494-501.
Zwahlen C, Li SC, Kay LE, Pawson T, For-
man-Kay JD. 2000. Multiple modes of
peptide recognition by the PTB domain
of the cell fate determinant Numb.
EMBO J 19:1505–1515.
948 REUGELS ET AL.
Page 15
  • Source
    • "In neuroepithelial cells, members of the prototypical Par/aPKC complex become apically enriched prior to the onset of neurulation [52]. During subsequent events in the neurulation process, Numb localizes asymmetrically to the basolateral domain, consistent with a model of aPKC-dependent Numb polarization [53]. Numb localization depends on its N-terminal domain, and mutations that diminish the protein interaction capacity of this region result in defective convergence and extension morphogenic events as well as neural tube development [54]. "
    [Show abstract] [Hide abstract] ABSTRACT: The ability to dictate cell fate decisions is critical during animal development. Moreover, faithful execution of this process ensures proper tissue homeostasis throughout adulthood, whereas defects in the molecular machinery involved may contribute to disease. Evolutionarily conserved protein complexes control cell fate decisions across diverse tissues. Maintaining proper daughter cell inheritance patterns of these determinants during mitosis is therefore a fundamental step of the cell fate decision-making process. In this review, we will discuss two key aspects of this fate determinant segregation activity, cortical cell polarity and mitotic spindle orientation, and how they operate together to produce oriented cell divisions that ultimately influence daughter cell fate. Our focus will be directed at the principal underlying molecular mechanisms and the specific cell fate decisions they have been shown to control.
    Preview · Article · Dec 2015
  • Source
    • "s, somite, nc, notochord. nervous system of mice, chicken and zebrafish, Numb is broadly expressed and is also found in the mitotically active progenitor cells within the neural epithelium, while the expression of NumbL is restricted to the developing nervous system and found in post-mitotic differentiating cells [10,11,15,18]. The similarity observed in respect to the expression patterns of Numb and NumbL within different vertebrate species strongly argues for a conservation of function in the context of neurogenesis. "
    [Show abstract] [Hide abstract] ABSTRACT: Members of the vertebrate Numb family of cell fate determinants serve multiple functions throughout early embryogenesis, including an essential role in the development of the nervous system. The Numb proteins interact with various partner proteins and correspondingly participate in multiple cellular activities, including inhibition of the Notch pathway. Here, we describe the expression characteristics of Numb and Numblike (NumbL) during Xenopus development and characterize the function of NumbL during primary neurogenesis. NumbL, in contrast to Numb, is expressed in the territories of primary neurogenesis and is positively regulated by the Neurogenin family of proneural transcription factors. Knockdown of NumbL afforded a complete loss of primary neurons and did not lead to an increase in Notch signaling in the open neural plate. Furthermore, we provide evidence that interaction of NumbL with the AP-2 complex is required for NumbL function during primary neurogenesis. We demonstrate an essential role of NumbL during Xenopus primary neurogenesis and provide evidence for a Notch-independent function of NumbL in this context.
    Full-text · Article · Oct 2013 · BMC Developmental Biology
  • Source
    • "No difference in Numb expression pattern and intensity was observed when we compared in situ hybridized wild-type embryos with the ones injected with NBP MO (not shown). When we analyzed the localization of Numb1-GFP in NBP morphants, we could not detect the typical pattern with basal accumulation of Numb1 (Fig. 4D, Reugels et al., 2006). "
    [Show abstract] [Hide abstract] ABSTRACT: Numb is an adaptor protein implicated in diverse basic cellular processes. Using the yeast-two hybrid system we isolated a novel Numb interactor in zebrafish called NBP which is an ortholog of human renal tumor suppressor Kank. NBP interacts with the PTB domain of Numb through a region well conserved among vertebrate Kanks containing the NGGY sequence. Similar NBP and Numb morphant phenotype such as impaired convergence and extension movements during gastrulation, neurulation and epidermis defects and enhanced phenotypic aberrations in double morphants suggest that the genes interact genetically. We demonstrate that the expression of NBP undergoes quantitative and qualitative changes during embryogenesis and that the protein accumulates at the cell periphery to sites of cell-cell contact during gastrulation and later in development it concentrates at the basal poles of differentiated cells. These findings imply a possible role of NBP in establishing and maintaining cell adhesion and tissue integrity.
    Full-text · Article · Feb 2012 · Developmental Biology
Show more