T H E J O U R N A L O F C E L L B I O L O G Y
© The Rockefeller University Press $8.00
The Journal of Cell Biology, Vol. 175, No. 1, October 9, 2006 147–157
The basic function of neurons is to receive, integrate, and trans-
mit signals. To do so, most neurons develop polarity by forming
a single axon and multiple dendrites (Craig and Banker, 1994;
Winckler and Mellman, 1999; Da Silva and Dotti, 2002; Horton
and Ehlers, 2003). Neurons have the remarkable ability to polar-
ize even in symmetrical in vitro environments (Dotti et al., 1988;
Craig and Banker, 1994). The processes of their polarization have
been extensively studied using hippocampal neurons. These cells
fi rst form several immature neurites that are capable of becoming
either axons or dendrites. One of the neurites then acquires axo-
nal characteristics, whereas the others later become dendrites.
Hippocampal neurons must use a robust internal mechanism that
guarantees polarization, as they generate a single axon and multi-
ple dendrites even when polarity is altered by axonal amputation
(Dotti and Banker, 1987; Goslin and Banker, 1989).
Recent studies have begun to defi ne the signaling path-
ways involved in neuronal polarization. Esch et al. (1999) re-
ported that the extracellular signals laminin and neuron-glia
cell adhesion molecule can specify which neurite will become
an axon. As effectors of spatial signals, rearrangements of the
cytoskeleton are important, as actin fi lament instability (Bradke
and Dotti, 1999) and tubulin assembly by collapsin response
mediator protein-2 (Inagaki et al., 2001; Arimura and Kaibuchi,
2005) are reported to initiate axon formation. Recent work has
shown that spatially localized intracellular signaling pathways,
including phosphoinositide-3-kinase (PI 3-kinase), phosphati-
dylinositol (3,4,5) triphosphate, the mPar3–mPar6–aPKC complex
(with the exception of some neurons in Drosophila melanogaster;
Rolls and Doe, 2004), Cdc42, Rap1B, STEF/Tiam1, Rac, Akt,
adenomatous polyposis coli, and glycogen synthase kinase-3β,
are involved in axon specifi cation for neuronal polarity formation
(Shi et al., 2003, 2004; Menager et al., 2004; Schwamborn
and Puschel, 2004; Jiang et al., 2005; Nishimura et al., 2005;
Yoshimura et al., 2005), and PI 3-kinase is implicated as an
upstream molecule in these events (Shi et al., 2003; Arimura
and Kaibuchi, 2005; Wiggin et al., 2005).
In spite of this progress, the mechanism and logic of
how the polarized distribution of intracellular signals originates
in the absence of external asymmetric cues remain elusive.
Shootin1: a protein involved in the organization
of an asymmetric signal for neuronal polarization
Michinori Toriyama,1 Tadayuki Shimada,1 Ki Bum Kim,1 Mari Mitsuba,1 Eiko Nomura,1 Kazuhiro Katsuta,1
Yuichi Sakumura,2 Peter Roepstorff,3 and Naoyuki Inagaki1
1Division of Signal Transduction and 2Bioinformatics Unit, Nara Institute of Science and Technology, Ikoma 630-0192, Japan
3Department of Biochemistry and Molecular Biology, University of Southern Denmark, DK-5230 Odense M, Denmark
cellular signals can induce neuronal polarization, it re-
mains unclear how these polarized signals are organized
without asymmetric cues. We describe a novel protein,
named shootin1, that became up-regulated during polar-
ization of hippocampal neurons and began fl uctuating
accumulation among multiple neurites. Eventually, shootin1
accumulated asymmetrically in a single neurite, which
led to axon induction for polarization. Disturbing the
eurons have the remarkable ability to polarize
even in symmetrical in vitro environments. Although
recent studies have shown that asymmetric intra-
asymmetric organization of shootin1 by excess shootin1
disrupted polarization, whereas repressing shootin1
expression inhibited polarization. Overexpression and
RNA interference data suggest that shootin1 is required
for spatially localized phosphoinositide-3-kinase activity.
Shootin1 was transported anterogradely to the growth
cones and diffused back to the soma; inhibiting this
transport prevented its asymmetric accumulation in
neurons. We propose that shootin1 is involved in the
generation of internal asymmetric signals required for
M. Toriyama and T. Shimada contributed equally to this paper.
Correspondence to Naoyuki Inagaki: firstname.lastname@example.org
K.B. Kim’s present address is Laboratory of Cell Signal Transduction, School of
Life Science and Biotechnology, Korea University, Seoul, 136-701 Korea.
Abbreviations used in this paper: 2DE, 2D electrophoresis; CMFDA,
5- chloromethylfl uorescein diacetate; DIV, day in vitro; E, embryonic day; miRNA,
microRNA; mRFP, monomeric red fl uorescent protein; P, postnatal day; PI 3-
The online version of this article contains supplemental material.
JCB • VOLUME 175 • NUMBER 1 • 2006 148
During the polarization of cultured hippocampal neurons,
undifferentiated neurites undergo competitive elongation with
each other. When one of them exceeds the others by a critical
length, it rapidly elongates to become an axon (Goslin and
Banker, 1989). This observation led to the proposal that a posi-
tive feedback loop and negative regulation among neurites are
necessary for neuronal polarization (Goslin and Banker, 1989;
Andersen and Bi, 2000; Banker, 2003). A locally acting positive
feedback loop may amplify a small stochastic increase in sig-
nals until it exceeds a threshold to induce an axon, and negative
regulation may also be important to prevent the formation
of surplus axons. However, little is known about the molecular
basis of such regulation.
To approach this problem, we performed proteome analyses
of cultured hippocampal neurons using highly sensitive large-
gel 2D electrophoresis (2DE), which can detect ?11,000 protein
spots over a dynamic range of 1–105 (Inagaki and Katsuta, 2004).
We describe a novel brain-specifi c protein, named shootin1. Our
data suggest that shootin1 organizes its own polarized distribution
to break neuronal symmetry through the PI 3-kinase pathway.
Identifi cation of shootin1 by double
Cultured hippocampal neurons are a well-established system to
study spontaneous neuronal polarization (Dotti et al., 1988;
Craig and Banker, 1994). They extend several minor processes
during the fi rst 12–24 h after plating (stages 1–2). One of these
processes then begins to elongate continuously to become an
axon (stage 3). The transition from stage 2 to 3 is the initial step
of polarization (Fig. 1 A).
To identify proteins involved in neuronal polarization, we
performed two separate proteome analyses of cultured rat hip-
pocampal neurons using a 93- × 103-cm large-gel 2DE (Inagaki
and Katsuta, 2004). One was to detect proteins up-regulated
during neuronal polarization (Fig. 1 A): we screened ?6,200
protein spots on 2DE gels and detected 277 that were consis-
tently up-regulated during the transition from stage 2 to 3 (n ≥ 3).
The second analysis screened proteins enriched in axons (Fig. 1 B).
Hippocampi dissected from embryonic day (E) 18 rat embryos
were cut into ?1-mm blocks and cultured on plastic dishes,
where they formed complicated networks of radial axons in
2 wk. The explants’ somatodendritic parts were then separated
from the axon networks, and both were compared by 2DE. By
screening ?5,200 protein spots, we detected 200 spots enriched
in the axon samples (n ≥ 3).
A total of 23 spots were detected by both screenings.
Tryptic digestion and mass spectrometry of one of them, lo-
cated at a molecular mass of 60 kD and pI = 5.3 in gels (Fig. 1,
A and B), identifi ed 10 peptides whose sequences corresponded
to the human cDNA sequence KIAA1598 encoding a 5′-truncated
ORF of 446 amino acids. A BLAST search identifi ed four
human EST clones (BI598285, BG720033, BE568283, and
BI457767) and suggested that 10 additional amino acids are
present in the complete ORF. We then cloned the cDNAs for the
rat and human ORFs and termed them shootin1.
Rat and human shootin1 encode proteins of 456 amino
acids and predicted molecular masses of 52.4 and 52.6 kD, re-
spectively (Fig. 1 C). Domain searching revealed that shootin1
contains three coiled-coil domains and a single proline-rich region
(Fig. 1 D). It does not show signifi cant homology to previously
known polypeptides, however, suggesting that it belongs to a
novel class of proteins. Database searches also identifi ed a mouse
orthologue of shootin1 (Fig. 1 C) and partial ORFs in Macaca
fascicularis, chick, zebrafi sh, and Fugu rubripes. Invertebrate
homologues of shootin1 were not found in the databases.
Thus, shootin1 is probably a late addition to the genome during the
evolution of animals.
Shootin1 is brain specifi c and highly
up-regulated during polarization
We raised an antibody against recombinant shootin1. It recog-
nized a 60-kD band, corresponding to the apparent Mr of na-
tive and recombinant shootin1, in immunoblots of rat cultured
hippocampal neurons (Fig. 1 E, arrowhead). Consistent with
the 2DE data for the metabolically labeled protein (Fig. 1 A),
the level of shootin1 expression increased remarkably during
stage 2/3 transition (14.4-fold increase; n = 4; P < 0.005)
and remained high until day in vitro (DIV) 14, thereafter re-
turning to a low level by DIV28 when expression of the pre-
synaptic protein synaptophysin increased (Fig. 1 E). Immunoblot
analysis of various rat tissues detected shootin1 in postnatal
day (P) 4 and adult brains but not in other tissues, suggesting
that shootin1 is a brain-specifi c protein (Fig. 1 F). Expression
of shootin1 was relatively low on E15, peaked around P4,
and decreased to a low level in the adult brain (Fig. 1 G). Thus,
the expression of shootin1 is up-regulated, both in hippocam-
pal neurons and in brain, during the period of axon formation
Shootin1 accumulates in axonal growth
cones during the stage 2/3 transition
Next, we examined the localization of shootin1 in cultured hip-
pocampal neurons. Immunocytochemical analysis showed a
faint and diffuse staining of endogenous shootin1 in early stage
2 neurons (18–24 h in culture; unpublished data). In late stage 2,
moderate amounts of shootin1 appeared in some growth
cones of minor processes (Fig. 1 H). We used a volume marker,
5-chloromethylfl uorescein diacetate (CMFDA), to measure the
relative concentration of shootin1: it was calculated by using
CMFDA as an internal standard (shootin1 immunoreactivity/
CMFDA staining). The relative concentration of shootin1 accu-
mulated in the growth cones of late stage 2 neurons was 2–4
times higher than that in the cell body (Fig. 1 H, arrowheads).
In stage 3, shootin1 accumulated strongly in axonal growth cones
(Fig. 1 I, arrows): 100% of axonal growth cones showed accu-
mulation (n = 19). The relative concentration of shootin1 in the
axonal growth cones of stage 3 neurons was ?10 times higher
than that in the other regions. Notably, the accumulation seen at
late stage 2 in minor processes mostly disappeared in stage 3
(Fig. 1 I, arrowheads), with only 12% of the processes showing
accumulation (n = 68). Shootin1 concentration in the cell body
remained low throughout stages 2 and 3 (Fig. 1, H and I, asterisks).
SHOOTIN1 AS A NEURONAL POLARITY SIGNAL • TORIYAMA ET AL. 149
The accumulation of shootin1 in axonal growth cones was
observed until stage 5 (unpublished data).
In stage 2, shootin1 shows fl uctuating
accumulation in multiple growth cones,
concurrent with neurite elongation
To analyze the localization of shootin1 in living neurons, we
monitored fl uorescent images of EGFP-shootin1 expressed in
hippocampal neurons under the cytomegalovirus promoter
every 5 min. Although relatively high levels of EGFP-shootin1
appeared in the soma, indicating that the expression exceeds
the endogenous levels, its distribution in neurites was virtually
identical to that of endogenous shootin1 (see the following
paragraph). Consistent with the immunocytochemical data, we
observed accumulation of EGFP-shootin1 in the growth cones
of minor processes in late stage 2 neurons (Fig. 2 A). As re-
ported previously (Goslin and Banker, 1989), minor processes
showed competitive extension and retraction before polarization.
Surprisingly, “hotspots” of EGFP-shootin1 accumulation re-
peatedly appeared and disappeared in the growth cones of
individual neurites (n = 11 cells; Fig. 2 A and Video 1, available
Most of the neurites elongated in conjunction with EGFP-shootin1
accumulation and, conversely, retracted as EGFP-shootin1
Figure 1. Identifi cation, structure, expression, and intracellular localization of shootin1. (A) Differential 2DE analysis of proteins in stage 2 (cultured for
14 h) and stage 3 (cultured for 62 h) hippocampal neurons. The arrows indicate the protein spot of shootin1 enriched in the stage 3 sample (stage 3/2 = 3.2;
n = 12; P < 0.001). (B) Differential 2DE analysis of proteins in cell body/dendrite and axon samples. The arrows indicate the same protein spot
shown in A, which is also enriched in the axon samples (axon/somatodendrite = 1.6; n = 7; P < 0.005). (C) Amino acid sequence of rat, human, and
mouse shootin1. Sequences of the peptides identifi ed by mass spectrometry analysis are underlined. (D) Schematic representation of rat shootin1, showing
three coiled-coil domains (CC1–3) and a single proline-rich region. (E) Immunoblot analysis of purifi ed recombinant shootin1 and shootin1 in cultured rat
hippocampal neurons at different stages. Immunoblot data of synaptophysin are also shown. (F) Immunoblot analysis of shootin1 in adult rat tissues and
P4 brain. (G) Immunoblot analysis of shootin1 in rat brains at various developmental stages. (H and I) Immunofl uorescent localization of shootin1 in late
stage 2 (H) and stage 3 (I) hippocampal neurons. Neurons were double stained with anti-shootin1 antibody (red) and a volume marker CMFDA (green).
Quantitative profi les show the relative fl uorescence intensities of shootin1 and CMFDA and relative concentration of shootin1 (shootin1 immunoreactivity/
CMFDA staining). Arrows, arrowheads, and asterisks denote axonal growth cones, minor process growth cones, and cell bodies, respectively. Bars, 20 μm.
JCB • VOLUME 175 • NUMBER 1 • 2006 150
disappeared (Fig. 2 B). To measure relative concentration of
EGFP-shootin1 in growth cones, we used the volume marker
monomeric red fl uorescent protein (mRFP): it was calculated
by using mRFP as an internal standard (EGFP-shootin1/mRFP).
By quantifying EGFP-shootin1 and mRFP in growth cones and
neurite elongation speed, we found a clear dose dependency of
neurite elongation rate on shootin1 concentration in the growth
cones of stage 2 neurons (Fig. 2 C).
Shootin1 accumulates asymmetrically
in a single neurite before polarization
We continued observations until the neurons entered stage 3.
Because long exposure to UV light damaged the cells, images were
recorded every 30 min (n = 3; Fig. 2, D–G; and Video 2, avail-
able at http://www.jcb.org/cgi/content/full/jcb.200604160/DC1).
After stage 2, when EGFP-shootin1 accumulation fl uctuated in
individual neurites (Fig. 2 D), the neurons entered a phase in
which one of the neurites was 10–15 μm longer than the others
(Fig. 2, E and G; and Fig. S1). In most cases, this neurite would
later become an axon (Goslin and Banker, 1989, 1990). In the
longest neurites, accumulation of EGFP-shootin1 stabilized
in the growth cone (Fig. 2 E and Fig. S1, neurite 1, arrows).
Simultaneously, the level of EGFP-shootin1 in its sibling neurites
decreased dramatically (neurites 2–5). In this period, the mean
number of neurites that showed EGFP-shootin1 accumulation
decreased to 1.13 (n = 30). The longest neurites then underwent
rapid elongation and the cells entered stage 3 (Fig. 2, F and G;
and Fig. S1). Consistent with the immunocytochemical data
(Fig. 1 I), EGFP-shootin1 remained highly concentrated in axonal
growth cones during stage 3, whereas it disappeared from the
growth cones of minor processes (Fig. 2, F and G; and Fig. S1).
The dynamic shift of shootin1 accumulation into the nascent
axon raises the possibility that it provides an intracellular asym-
metric signal for neuronal polarization.
Excess levels of shootin1 disturb
its asymmetric distribution and induce
formation of surplus axons
To examine whether the asymmetric accumulation of shootin1
in a single neurite is important for neuronal polarization, we
overexpressed EGFP-shootin1 or myc-tagged shootin1 (myc-
shootin1) in hippocampal neurons under the stronger β-actin
promoter. A high level of EGFP-shootin1 was detected in the
soma, with its frequent transport from the soma to growth cones
(Fig. 3 A, arrowheads; and Video 3, available at http://www.jcb.
org/cgi/content/full/jcb.200604160/DC1). This in turn resulted
in more continuous accumulation of EGFP-shootin1 in multiple
growth cones (Fig. 3 A, arrows; compared with the dynamic
fl uctuation of a lower level of EGFP-shootin1 in Fig. 2, A and B,
and Video 1) and ectopic accumulation of myc-shootin1 in
minor process growth cones in stage 3 neurons (Fig. 3 B,
arrowheads). These results suggest that the limited amount
of shootin1 is essential for its asymmetric accumulation in a
We further cultured the neurons with overexpressed myc-
shootin1 until DIV7. Remarkably, 47 ± 2.1% (n = 3; 71 neu-
rons examined; P < 0.0001, compared with myc-GST) of the
neurons bore more than one (two to four) axons that were
immunostained by the axon-specifi c markers anti–tau-1 (Fig. 3 C)
and anti-synaptophysin (Fig. S2 A, available at http://www.
jcb.org/cgi/content/full/jcb.200604160/DC1) antibodies but were
immunonegative for the dendrite-specifi c marker anti-MAP2
Figure 2. Dynamic accumulation of EGFP-
shootin1 in growth cones of hippocampal
neurons. (A) A stage 2 hippocampal neuron
expressing EGFP-shootin1 was observed un-
der a time-lapse fl uorescence microscope
every 5 min. The full video is presented in
Video 1 (available at http://www.jcb.org/
cgi/content/full/jcb.200604160/DC1). (B) The
pictures represent a series of enlarged im-
ages of neurite 1 in A taken every 5 min.
Asterisks indicate the front edge of the neurite.
(C) Correlation between neurite elongation
speed and EGFP-shootin1 levels in growth
cones. Stage 2 hippocampal neurons express-
ing EGFP-shootin1 (green) and mRFP (red) was
observed under a time-lapse fl uorescence mi-
croscope every 5 min. Relative levels of EGFP-
shootin1 and mRFP in growth cones were
quantifi ed using Multi Gauge (n = 315). Rela-
tive concentration of EGFP-shootin1 in growth
cones was calculated by using mRFP as an in-
ternal standard (EGFP-shootin1/mRFP), and
the neurite elongation speeds during the next
5 min were measured. (D–F) A hippocampal
neuron expressing EGFP-shootin1 was ob-
served by time-lapse fl uorescence microscopy
every 30 min. The full video is presented in
Video 2. Arrows indicate EGFP-shootin1 accu-
mulation in the growth cone of the nascent
axon. (G) Elongation of neurites 1–5 shown in D. The circles denote growth cones with apparent shootin1 accumulation, and the diamonds indicate growth
cones without the accumulation. The green shade denotes the period when the nascent axon (neurite 1) showed exclusive and continuous accumulation
before rapid elongation. Bars, 20 μm.
SHOOTIN1 AS A NEURONAL POLARITY SIGNAL • TORIYAMA ET AL. 151
antibody (Fig. S2 B). In contrast, only 2.5 ± 1.4% (n = 3; 81
neurons examined) of control neurons with overexpressed myc-
GST formed supernumerary axons. On DIV4, 32 ± 1.8% of
neurons overexpressing shootin1 bore multiple axons (n = 3;
209 neurons examined; P < 0.002, compared with GST) that
were immunoreactive for tau-1 and anti-synaptophysin antibod-
ies but were immunonegative for anti-MAP2 antibody. On the
other hand, 10 ± 2.5% of control neurons overexpressing GST
bore multiple axons (n = 3; 191 neurons examined). At 50 h in
culture, 21 ± 0.9% of neurons overexpressing shootin1 bore
multiple axon-like neurites (n = 3; 226 neurons examined;
P < 0.001, compared with GST) that were immunoreactive for
tau-1, whereas only 6 ± 1.2% of neurons overexpressing GST
bore multiple axon-like neurites (n = 3; 173 neurons examined).
We also quantifi ed the length of the neurites. Neurites labeled
by axonal markers were markedly longer than dendrites (Fig. 3 D).
Interestingly, the sum of the length of neurites in neurons over-
expressing shootin1 was similar to that in control neurons on
DIV7 (Fig. 3 E) and DIV4 (not depicted). Hippocampal neu-
rons elongate axons rapidly (43 μm/d) from stages 3 to 5 (DIV7;
Dotti et al., 1988). We consider that the limited amount of struc-
tural components produced in cell bodies similarly limits the
total neurite elongation in shootin1-overexpressing and control
neurons. A similar limitation of neurite growth in neurons with
multiple axons was reported previously (Jiang et al., 2005).
Multiple axons were also induced by nontagged shootin1 co-
transfected with EGFP (43 ± 2.6%; n = 3; 67 neurons examined;
P < 0.001, compared with EGFP), whereas a small population
(1.6 ± 1.6%; n = 3; 61 neurons examined) of control neurons
expressing EGFP formed supernumerary axons, thereby ruling
out the possibility that tagging myc to shootin1 infl uences the
effects. These results suggest that the asymmetric accumulation
of shootin1 is involved in neuronal polarization.
Repressing shootin1 expression inhibits
We next suppressed shootin1 expression using a vector-based
RNAi system that expresses microRNA (miRNA). To ensure a
high level of expression of miRNA before polarization, hippo-
campal neurons prepared from E18 rat embryo and transfected
with the expression vector of a miRNA designated against
shootin1 or a control miRNA were plated on polystyrene plates
without any coating. After 20 h for the induction of the miRNA
expression, the cells were collected and cultured on coverslips
coated with polylysine and laminin. The shootin1 miRNA re-
duced the level of neuronal shootin1 (Fig. 3 F, arrows), in com-
parison to control neurons (arrowheads) and neurons transfected
with the control miRNA. Repression of shootin1 expression by
the miRNA led to signifi cant suppression of neuronal polari-
zation at 50 and 70 h in culture, whereas the control miRNA
had no such effect (Fig. 3 G). On the other hand, 100% of neu-
rons transfected with the shootin1 miRNA (n = 25) became
polarized on DIV7. As the 20-h delay in neuronal plating
might affect time course of neuronal polarization after plating,
we also performed similar experiments using E17 rat embryo.
Essentially equivalent data were obtained with E17 rat embryo
Figure 3. Effects of shootin1 overexpression
and RNAi on polarization of hippocampal
neurons. (A) A neuron overexpressing EGFP-
shootin1 was observed every 5 min under a
time-lapse fl uorescence microscope. The full
video is presented in Video 3 (available
200604160/DC1). Arrows indicate EGFP-
shootin1 accumulated at a high level in growth
cones. Arrowheads indicate frequent transport
of shootin1 from the cell body to the growth
cones. (B) Stage 3 neurons overexpressing
myc-shootin1 were immunostained by anti-myc
antibody (red). Arrowheads indicate myc-
shootin1 aberrantly accumulated in the growth
cones of minor processes. (C) Hippocampal
neurons overexpressing myc-shootin1 were
cultured for 7 d and then double immuno-
stained by anti-myc (green) and anti–tau-1
(red) antibodies. Arrowheads indicate axons
labeled by anti–tau-1 antibody. (D) Quantifi ca-
tion of the neurite lengths of the neurons (DIV7)
with surplus axons. The lengths of the longest
and secondary axons and dendrites were
measured in neurons overexpressing myc-
shootin1 (n = 26) or myc-GST (n = 27). Data
present neurite length as means ± SE.
(E) Quantifi cation of the total neurite lengths
of the neurons (DIV7) overexpressing myc-
shootin1 or myc-GST. (F) Neurons transfected
with the miRNA or a control miRNA vector were cultured for 50 h. They were then fi xed and immunostained with anti-shootin1 antibody. The vectors for
miRNA and control miRNA expressions are designed to coexpress EGFP. (G) Neurons prepared from E18 or E17 embryo and transfected with the miRNA or
a control miRNA vector were cultured. Data present percentages of neurons bearing axons (stage 3 neurons) as means ± SE (**, P < 0.02; ***, P < 0.002;
n = 3; a total of 752 neurons were examined). Black dots present percentages of nontransfected neurons bearing axons. (H and I) Quantifi cation of the
total neurite lengths of the neurons overexpressing myc-shootin1 or myc-GST (H) or expressing shootin1 miRNA or control miRNA (I). Data present neurite
length as means ± SE (***, P < 0.0005; n = 3; a total of 761 neurons for H and 568 neurons for I were examined). Bars, 50 μm.
JCB • VOLUME 175 • NUMBER 1 • 2006 152
(Fig. 3 G). The signifi cant suppression of neuronal polarization
by shootin1 RNAi provides evidence that shootin1 is involved
in neuronal polarization.
Shootin1 accumulation in growth cones
stimulates neurite elongation during
the stage 2/3 transition
As described, shootin1 showed fl uctuating accumulation in
growth cones concurrent with neurite elongation in stage 2
neurons, raising the possibility that shootin1 accumulation in
growth cones stimulates neurite elongation. During the stage
2/3 transition, neurites of hippocampal neurons show dynamic
elongation and retraction without a remarkable increase in total
neurite length (Goslin and Banker, 1989). In addition, the stage
2/3 transition is a critical period of neuronal polarization.
Therefore, we examined the effect of shootin1 overexpression
and RNAi during this period (24 and 48 h in culture). In contrast
to the data of DIV7 (Fig. 3 E) and DIV4, shootin1 overexpres-
sion induced a signifi cant increase in total neurite length during
this period (Fig. 3 H). Furthermore, repression of its level by
RNAi resulted in a signifi cant decrease in it (Fig. 3 I). Along
with the time-lapse data, these results suggest that shootin1
accumulation in growth cones stimulates neurite elongation
during the transition from stage 2 to 3.
Shootin1 is anterogradely transported to
the growth cones with wave-like structures
and diffuses back to the soma
We next asked how shootin1 accumulates asymmetrically in
hippocampal neurons. As already noted (Fig. 3 A, arrowheads),
the series of time-lapse imaging revealed active transport of
shootin1 from the cell body to the growth cones in stages 2 and
3 neurons (Fig. 4 A). The shootin1 transport was observed along
minor processes and axons. Ruthel and Banker (1998, 1999)
reported wave-like anterograde movement of growth cone–like
structures along minor processes and axons of cultured hippo-
campal neurons. The transport rate of these “waves” was ?3
μm/min, similar to that of slow axonal transport component b,
which transports actin (Lasek, et al., 1984; Brown, 2003). In
addition, waves were enriched in F-actin and their movement
was reversibly blocked by the actin-disrupting agent cytochalasin.
Therefore, Ruthel and Banker (1998, 1999) suggested that
actin and other cytoskeletal components are transported as
waves from the cell body to neurite tips via an actin-dependent
mechanism. Shootin1 traveled as discrete boluses with growth
cone–like structures at a mean rate of 1.0 ± 0.1 μm/min (n = 12),
which is similar to the speed of wave transport. We occasion-
ally observed transient retrograde transport of GFP-shootin1.
However, as in the case of the wave, retrograde transport
was rare and short-lived, quickly reverting to anterograde move-
ment. In addition, the boluses of shootin1 were enriched for
F- actin (Fig. 4 B) and the transport was arrested by the actin-
disrupting agent cytochalasin D within 5 min (Fig. S3 A, available
at http://www.jcb.org/cgi/content/full/jcb.200604160/DC1), as
reported for the waves. Blebbistatin, an inhibitor of myosin II
(Straight et al., 2003), also stopped shootin1 transport (Fig.
4 C). These re sults suggest that shootin1 is anterogradely trans-
ported with the wave-like structure by an actin- and myosin-
Within 2 h of the cessation of the transport by blebbistatin
or cytochalasin D, shootin1 accumulation in the axonal growth
cones of stage 3 neurons disappeared (Fig. 4 D and Fig. S3 B,
arrows) and a relatively high level of shootin1 was observed in
the soma, axonal shaft, and minor processes (arrowheads). To
examine whether shootin1 returned back from the axonal growth
cones to the cell bodies by diffusion or was locally degraded in
the growth cones and newly synthesized in the cell body, we
used the photoconvertible reporter Kaede (Ando et al., 2002)
to distinguish old shootin1 from newly synthesized shootin1.
Figure 4. Shootin1 is anterogradely transported to the growth cones with
wave-like structures and diffuses back to the soma. (A) Distal movements
of EGFP-shootin1 within neurite shafts from the cell body to a growth
cone. The arrows indicate boluses of EGFP-shootin1. (B) A stage 3 hippo-
campal neuron double stained with anti-shootin1 antibody and Rhodamine
phalloidin. Shootin1- and F-actin–enriched wave is indicated by arrowheads.
(C) Serial time-lapse images showing the effect of 50 μM blebbistatin on
anterograde transport of EGFP-shootin1. Blebbistatin was applied to the
medium for between 35 and 40 min. EGFP-shootin1 in the neurite shaft is
indicated by arrowheads. (D) Disturbance of shootin1 transport by bleb-
bistatin inhibits its accumulation in axonal growth cones. Stage 3 hippo-
campal neurons (cultured for 48 h) were incubated with 50 μM blebbistatin
for 1 h and stained with anti-shootin1 antibody. (E) Shootin1 diffuses from
the axonal growth cone to the cell body. Kaede-shootin1 expressed in
stage 3 hippocampal neurons was converted from green to red using UV
light. 1 h after the cessation of shootin1 transport by 50 μM blebbistatin,
distributions of red Kaede-shootin1 and newly synthesized green Kaede-
shootin1 were examined. Bars: (A and B) 20 μm; (C–E) 50 μm.
SHOOTIN1 AS A NEURONAL POLARITY SIGNAL • TORIYAMA ET AL. 153
Kaede-shootin1 expressed in stage 3 hippocampal neurons was
converted from green to red using UV light, and shootin1 trans-
port was blocked by blebbistatin (Fig. 4 E). 1 h after the cessa-
tion of shootin1 transport, the accumulation of the red
Kaede-shootin1 in the axonal growth cones decreased (Fig. 4 E,
yellow arrows), whereas the red fl uorescence of Kaede-shootin1
increased in the soma and shaft (yellow arrowheads). On the
other hand, we could not detect new synthesis of green Kaede-
shootin1 in the soma (Fig. 4 E, blue arrowhead). These data
suggest that shootin1 passively diffuses back from the growth
cones to the cell bodies.
Inhibition of shootin1 transport prevents
its asymmetric accumulation in neurons
We next asked whether the anterograde transport of shootin1 is
involved in its asymmetric accumulation in hippocampal neu-
rons. As shown in Fig. 5 A and Fig. S3 C, cessation of shootin1
transport in stage 2 neurons by blebbistatin or cytochalasin D
prevented accumulation of shootin1 in multiple growth cones.
Stage 2 neurons were cultured for 36 h in the presence of bleb-
bistatin or cytochalasin D. As described, in control neurons,
shootin1 accumulates asymmetrically in growth cones of na-
scent axons during this period. On the other hand, shootin1 did
not accumulate in single neurites in the presence of these drugs
(Fig. 5 B and Fig. S3 D). Cessation of shootin1 transport in al-
ready polarized stage 3 neurons also prevented accumulation of
shootin1 in axonal growth cones, as described (Fig. 4 D and Fig.
S3 B). These data indicate that the actin- and myosin-dependent
anterograde transport of shootin1 is necessary for its asymmet-
ric accumulation in single growth cones.
Shootin1 regulates the localization of
PI 3-kinase activity in hippocampal neurons
Recent studies indicate that PI 3-kinase is located at a critical
upstream position in signaling pathways for neuronal polariza-
tion (Arimura and Kaibuchi, 2005; Wiggin et al., 2005). We fi -
nally examined whether shootin1 interacts with the PI 3-kinase
pathway. The physiological association of shootin1 and PI 3-
kinase was examined by coimmunoprecipitation assay. When
shootin1 was immunoprecipitated from P5 rat brain lysates,
coimmunoprecipitation of the p85 subunit of PI 3-kinase was
detected (Fig. 6 A). Shootin1 was also reciprocally coimmuno-
precipitated with p85, indicating that it associates with p85 in
vivo. PI 3-kinase activity, indirectly visualized by the phosphor-
ylation of Akt at Ser473 (P-Akt), was enriched in the axonal
growth cones of stage 3 neurons (Fig. 6 B, arrows) as reported
(Shi et al., 2003) and preferentially colocalized there with
shootin1 (Fig. 6 B, insets). We exogenously coexpressed
shootin1 and p85 in HEK293T cells but could not detect co-
immunoprecipitation between shootin1 and p85 (not depicted).
Thus, shootin1 may interact with PI 3-kinase through unidenti-
fi ed neuronal proteins. As shown recently (Yoshimura et al.,
2006), overexpression of constitutively active PI 3-kinase
(Myr-PI 3-K p110) induced formation of multiple axons (Fig. 6 H
and Fig. S4 B, available at http://www.jcb.org/cgi/content/full/
jcb.200604160/DC1), as in the case of shootin1 overexpression.
Jiang et al. (2005) also reported that overexpression of constitu-
tively active Akt (Myr-Akt), a downstream kinase of PI 3-K, in-
duced formation of multiple axons. On the other hand, inhibition
of PI 3-kinase activity by 20 μM LY294002, a specifi c inhibitor
of PI 3-kinase, delayed neuronal polarization (Fig. 6 C), as re-
ported previously (Menager et al., 2004) and as in the case of
shootin1 RNAi. These results suggest that shootin1 interacts
with PI 3-kinase and is involved in a similar pathway mediating
Next, we examined whether shootin1 functions upstream
of PI 3-kinase or vice versa. Shootin1 RNAi decreased its level
in axonal growth cones, which in turn inhibited accumulation of
PI 3-kinase activity there (Fig. 6 D, arrows), suggesting that
shootin1 in axonal growth cones is required for accumulation of
PI 3-kinase activity there. Conversely, myc-shootin1 over-
expression induced its ectopic accumulation in the growth cones
of minor processes, which in turn resulted in ectopic accumula-
tion there of P-Akt (Fig. 6 E, arrowheads), thereby suggesting
that accumulation of shootin1 can recruit PI 3-kinase activity.
On the other hand, inhibition of PI 3-kinase activity by
LY294002 did not affect the accumulation of shootin1 in axonal
growth cones (Fig. 6 F, arrows). Shootin1 overexpression or
RNAi did not change the activity of PI 3-kinase in hippocampal
Figure 5. Inhibition of shootin1 transport prevents its asymmetric accu-
mulation in hippocampal neurons. (A) Stage 2 hippocampal neurons
treated with 50 μM blebbistatin for 1 h were double stained with anti-
shootin1 antibody (red) and a volume marker CMFDA (green). Quantitative
profi les show the relative fl uorescence intensities of shootin1 and CMFDA
and the relative concentration of shootin1 (shootin1 immunoreactivity/
CMFDA staining) in neurites 1–4. The arrowheads and asterisks denote the
minor processes and cell body, respectively. (B) Hippocampal neurons were
treated with 50 μM blebbistatin at 14 h in vitro, further cultured for 36 h,
and double stained with anti-shootin1 antibody (red) and CMFDA (green).
Quantitative profi les show the relative fl uorescence intensities of shootin1
and CMFDA and the relative concentration of shootin1 in neurites 1–6.
The asterisks denote the cell body. Bars, 50 μm.
JCB • VOLUME 175 • NUMBER 1 • 2006 154
neurons (Fig. 6 G), ruling out the possibility that the expression
level of shootin1 changes the total activity of PI 3-kinase in
neurons. These results suggest that shootin1 regulates subcellular
localization of PI 3-kinase activity in hippocampal neurons.
We further examined the functions of shootin1 and PI 3-
kinase within the cell polarity pathways. Inhibition of PI 3-kinase
activity by LY294002 led to a reduction in the percentage of
neurons with multiple axons induced by shootin1 overexpression
(Fig. 6 H and Fig. S4 A). On the other hand, multiple axon
formation by overexpression of constitutively active PI 3-kinase
was not inhibited by shootin1 RNAi (Figs. 6 H and Fig. S4 B).
Collectively, these results provide evidence that shootin1 func-
tions upstream of PI 3-kinase and is required for spatially
localized PI 3-kinase activity, which is essential for neuronal
polarization (Shi et al., 2003).
We have identifi ed a novel brain-specifi c protein, shootin1,
using highly sensitive 2DE-based proteomics. The spatiotemporal
localization of shootin1 in hippocampal neurons changed dy-
namically during polarization: it became up-regulated, began
fl uctuating accumulation among multiple neurites, and eventu-
ally accumulated asymmetrically in a single neurite, which led
to axon induction for polarization. Disturbing the asymmetric
organization of shootin1 by excess shootin1 induced formation
of multiple axons, whereas repressing shootin1 expression
inhibited polarization. These results suggest that shootin1 plays
a critical role in neuronal polarization.
Shootin1 in generation of an asymmetric
signal for neuronal polarization
How does the symmetric localization of shootin1 shift to
asymmetric localization before polarization? Does shootin1
spontaneously organize its own polarized distribution without
asymmetric cues? Two types of regulation have been postulated
for a mechanism that accounts for spontaneous neuronal polar-
ization (Craig and Banker, 1994; Andersen and Bi, 2000). One
is a positive feedback loop acting locally in neurites, where a
stochastic increase in signals is enhanced until their level ex-
ceeds a threshold to induce an axon. If accumulation of a partic-
ular molecule in growth cones stimulates neurite elongation and
if its accumulation increases in proportion to neurite length,
such a molecule can constitute the positive feedback loop to en-
hance its own signal for axon formation (Goslin and Banker,
1990; Craig and Banker, 1994). Our data suggest that shootin1
accumulation in growth cones stimulates neurite elongation
during the stage 2/3 transition. In addition, shootin1 may accu-
mulate in growth cones in a neurite length–dependent manner,
as it is actively transported from the cell body to growth cones
and its retrograde diffusion to the cell body should vary inversely
with neurite length (Goslin and Banker, 1989). Consistent with
this notion, inhibiting the anterograde transport of shootin1 dis-
turbed its asymmetric accumulation in neurons. Thus, shootin1
Figure 6. Shootin1 regulates the localization
of PI 3-kinase activity in hippocampal neurons.
(A) Brain lysates from P5 rat brain were incu-
bated with anti-shootin1 antibody, anti-p85
antibody, or control IgG. The immunoprecipi-
tates were analyzed by immunoblotting with
anti-shootin1 and anti-p85 antibodies as indi-
cated. (B) Stage 3 hippocampal neurons were
incubated with DMSO as control for 10 h and
double stained with anti-shootin1 antibody
and anti–P-Akt (Ser473) antibody. Arrows indi-
cate an axonal growth cone. Note that PI 3-
kinase activity was preferentially colocalized
with shootin1 in the axonal growth cone (insets).
(C) Neurons were cultured in the normal
medium or in the presence of 20 μM LY294002
or DMSO. Data present percentages of neu-
rons bearing axons (stage 3 neurons) as means
± SE (**, P < 0.005; ***, P < 0.002; n = 3;
a total of 699 neurons were examined).
(D) Stage 3 neurons transfected with the miRNA
against shootin1 were double stained with
anti-shootin1 antibody (blue) and anti–P-Akt
antibody (red). The vector for the miRNA
expressions is designed to coexpress EGFP
(green). Arrows indicate a shootin1-immuno-
negative axonal growth cone without remark-
able accumulation of P-Akt. (E) Shootin1
overexpressed in stage 3 neurons accumulated
ectopically in minor processes together with P-Akt (arrowheads). (F) Stage 3 hippocampal neurons were incubated with 20 μM LY294002 for 10 h and
double stained with anti-shootin1 antibody and anti–P-Akt antibody. Arrows indicate a P-Akt–immunonegative axonal growth cone with remarkable
shootin1 accumulation. (G) Primary cultured hippocampal neurons were transfected by pCAGGS-myc-shootin1 or pCAGGS-myc-GST (control; left) or were
transfected with the miRNA against shootin1 or a control miRNA (right) using Nucleofector. The effi ciency of transfection was >80%. After 36 h in culture,
cell lysates were collected and immunoblotted by anti–P-Akt antibody to monitor PI 3-kinase activity. (H) Hippocampal neurons overexpressing myc-shootin 1
were cultured for 7 d in the presence of 20 μM LY294002 or DMSO (n = 3; 198 neurons examined). Neurons transfected with pCAGGS–Myr–PI 3- kinase
p110 plus the miRNA against shootin1 or pCAGGS–Myr–PI 3-kinase p110 plus a control-miRNA were also cultured for 7 d (n = 3; 225 neurons
examined). Data present percentages of neurons with multiple axons as means ± SE. **, P < 0.001. Bars, 20 μm.
SHOOTIN1 AS A NEURONAL POLARITY SIGNAL • TORIYAMA ET AL. 155
is a good candidate molecule for the requisite positive feedback
loop for axon induction (Fig. 7 A).
The second component is negative regulation to prevent the
formation of surplus axons (Craig and Banker, 1994; Andersen
and Bi, 2000). Because the concentration of shootin1 in the cell
body is very low and limited throughout stages 2 and 3, the afore-
mentioned positive feedback loop may elicit competition among
neurites for a limited amount of shootin1: an increase in shootin1
in one growth cone may lead to a decrease in sibling growth
cones (Fig. 7 B). Indeed, sustained accumulation of shootin1 in
one growth cone reduced levels in its sibling neurites. Further-
more, excess levels of shootin1 prevented its redistribution and
induced the formation of surplus axons. Thus, the negative regu-
lation to prevent formation of surplus axons may be achieved by
competition among neurites for a limited amount of shootin1.
Fig. 7 C shows our current model in which shootin1 gen-
erates an asymmetric signal for neuronal polarization. In this
model, shootin1 up-regulation triggers the aforementioned pos-
itive and negative regulation, which shifts the symmetry of neu-
rons from a stable to an unstable state. Eventually, shootin1
accumulates asymmetrically in a single neurite, through ampli-
fi cation of its stochastic signals and competitive accumulation
among neurites, thereby leading to neuronal polarization.
Shootin1 and previously described
mechanisms for neuronal polarization
Recent papers indicate that PI 3-kinase is located at an upstream
position in signaling pathways for neuronal polarization involv-
ing many molecules, such as phosphatidylinositol (3,4,5) tri-
phosphate, the mPar3–mPar6–aPKC complex, Cdc42, Rap1B,
STEF/Tiam1, Rac, Akt, adenomatous polyposis coli, glycogen
synthase kinase-3β, and collapsin response mediator protein-2
(Shi et al., 2003; Arimura and Kaibuchi, 2005; Wiggin et al.,
2005). We found that shootin1 interacts with PI 3-kinase and is
required for spatially localized PI 3-kinase activity in hippo-
campal neurons. Furthermore, a series of overexpression and
loss of function studies suggested that shootin1 functions up-
stream of PI 3-kinase in regulating neuronal polarity. Thus,
shootin1 may be involved in the organization of polarized PI 3-
kinase activity (Fig. 7 C), which is essential for neuronal polar-
ization (Shi et al., 2003).
Recently, Jacobson et al. (2006) showed time-lapse imag-
ing of the motor domain of kinesin-1 in cultured hippocampal
neurons. As in the case of shootin1, the kinesin-1 motor domain
transiently accumulated in different minor processes of stage 2
neurons and selectively and continuously accumulated in the
nascent axon during the stage 2/3 transition, thereby indicating
that it serves as a very early marker for the symmetry breaking
event. The present data showed that shootin1 accumulation in
growth cones was induced by the actin- and myosin-dependent
wave-like transport and stimulated neurite elongation. On the
other hand, the accumulation of the kinesin-1 motor domain
may be dependent on microtubule and was not related to neurite
elongation. It is intriguing to analyze how these molecules inter-
act during polarization.
In addition to internal signals for polarization, additional
external cues are likely to adjust the orientation of an axon and
dendrites in situ. Although the identities of such cues in the
brain are not yet clear, Esch et al. (1999) reported that the spa-
tially asymmetric extracellular signals of laminin and neuron-glia
cell adhesion molecule can specify which neurite will become
an axon under experimental conditions. The present study does
not rule out the possibility that shootin1 is modifi ed by other
molecules. By regulating the activity of shootin1, additional
molecules might further adjust the orientation of an axon and
dendrites in situ.
In conclusion, we have identifi ed shootin1, a novel protein
involved in neuronal polarization. Based on the present fi nd-
ings, we proposed a model in which shootin1 expression trig-
gers the positive and negative regulation required for neuronal
symmetry breaking. Although we cannot rule out the involve-
ment of other potential mechanisms, our data provide an insight
into how internal asymmetric signals are generated during neu-
Materials and methods
Cultures and metabolic labeling
Hippocampal neurons prepared from E18 rat embryos were cultured as
described previously (Inagaki et al., 2001). For hippocampal explant cul-
ture, the hippocampi dissected from E18 rat embryos were cut into blocks
(?1 mm), carefully washed to remove dissociated cells, and cultured on
polylysine- and laminin-coated plastic dishes. The explants started to ex-
tend radial axons on the dishes within 12 h. On DIV14, the radial axons
formed complicated networks around the explants. The axonal networks
were usually devoid of cell bodies, dendrites, and nonneuronal cells (Fig.
1 B), although we occasionally observed migration of neurons from blocks
onto axonal networks. Such cells were rigorously removed under a micro-
scope using pipette tips. Explants containing somatodendritic parts were
separated from radial axons by applying streams of medium to the explants
with a pipette, and the explants were then collected in microcentrifuge tubes.
Figure 7. A model for the involvement of shootin1 in generation
of an asymmetric signal for neuronal polarization. (A) A positive
feedback loop between shootin1 accumulation in growth cone
and shootin1-induced neurite elongation. (B) Competition among
neurites for a limited amount of shootin1. (C) Shootin1 up-regulation
triggers local positive feedback loops (red arrows) and negative
regulations (blue arrows). Eventually, shootin1 will be asymmetri-
cally accumulated in a single neurite and recruit PI 3-kinase activity
there, thereby leading to neuronal polarization.
JCB • VOLUME 175 • NUMBER 1 • 2006 156
Removal of the explants and dissociated cells from axon networks was veri-
fi ed by microscopy.
For quantitative 2DE, stages 2 and 3 neurons were metabolically la-
beled with the culture medium containing 13% of the normal levels of me-
thionine and cysteine plus Pro-mix L-[35S] in vitro cell labeling mix (containing
?70% L-[35S]methionine and ?30% L-[35S]cysteine; GE Healthcare) for 4 h.
Hippocampal explants were labeled with the same medium for 24 h.
Highly sensitive gel 2DE and protein identifi cation by mass spectrometry
2DE was performed as reported previously (Oguri et al., 2002), using a
93- × 103-cm large-gel system (Inagaki and Katsuta, 2004). For differen-
tial 2DE, neurons or explants were metabolically labeled and protein spots
separated by 2DE gels were visualized by autoradiography. For protein
identifi cation, unlabeled protein samples from 2-wk-old rat brains were
separated by the 2DE gel and visualized by silver staining. The protein
spots corresponding to the radio-labeled ones were then excised from gels
and in-gel digested as described previously (Nomura et al., 2004). Matrix-
assisted laser desorption/ionization mass spectrometry was performed using
a Voyager Elite equipped with delayed extraction (Applied Biosystems).
Database searches were conducted using the Mascot program (Matrix
Science) and National Center for Biotechnology Information databases.
Cloning of shootin1
cDNA encoding KIAA1598 was provided by T. Nagase and O. Ohara
(Kazusa DNA Research Institute, Chiba, Japan). Full-length cDNA of hu-
man shootin1 was obtained by PCR of KIAA1598 with the primers 5′-G C-
G G A T C C A T G A A C A G C T C G G A C G A A G A G A A G C A G C T G C A G C T C A T T A C C A G-
T C T G A A G and 5′-G C G G A T C C C T A C T G G G A G G C C A G T A T T C . cDNA en-
coding rat shootin1 was amplifi ed by PCR from a rat brain cDNA library
(CLONTECH Laboratories, Inc.) with the primers 5′-C C G C T C G A G A T G A A-
C A G C T C G G A C G A G G A G A A G and 5′-C C G C T C G A G T T A C T G G G A G G-
C C A G G A T T C C C T T C A G . The cDNAs were then subcloned into pCMV
(Stratagene), pCAGGS with a β-actin promoter (provided by J. Miyazaki,
Osaka University, Osaka, Japan; Niwa et al., 1991), pEGFP (CLONTECH
Laboratories, Inc.), pGEX (GE Healthcare), and pKaede-MC1 (MBL Inter-
national Corporation) vectors.
Protein and antibody preparation
Recombinant shootin1 was expressed in Escherichia coli as a GST fusion
protein and purifi ed on a glutathione–Sepharose column (GE Healthcare),
after which GST was removed from shootin1 by PreScission protease (GE
Healthcare). Rabbit polyclonal anti-shootin1 antibody was raised against
the recombinant shootin1 and affi nity purifi ed before use.
Immunocytochemistry, immunoblot, and immunoprecipitation
Immunocytochemistry, CMFDA staining, Rhodamine phalloidin staining,
and immunoblot were performed as described previously (Inagaki et al.,
2001). For immunoprecipitation, P4 or P5 rat brains were extracted by
addition of lysis buffer (50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl,
1% Triton X-100, 0.1% SDS, 0.1% sodium deoxycholate, 2 mM phenyl-
methylsulfonyl fl uoride, 5 μg/ml leupeptin, 10 mM NaF, 1 mM Na3VO4,
and 10 mM β-glycerophosphate) and centrifuged at 100,000 g for 30 min
at 4°C. The supernatants were incubated with antibodies overnight at 4°C,
and immunocomplexes were then precipitated with protein G–Sepharose
4B (GE Healthcare). After washing out beads with RIPA buffer, immuno-
complexes were analyzed by immunoblot.
Fluorescent and phase-contrast images of neurons were acquired at room
temperature using a fl uorescent microscope (Axioplan 2; Carl Zeiss Micro-
Imaging, Inc.) equipped with a Plan-NEOFLUAR 40×, 0.75 NA, or 20×,
0.50 NA, objective, a charge-coupled device camera (AxioCam MRm; Carl
Zeiss MicroImaging, Inc.), and imaging software (AxioVision 3; Carl Zeiss
MicroImaging, Inc.). Time-lapse microscopy was performed at 37°C using
a fl uorescent microscope (Axiovert S100; Carl Zeiss MicroImaging, Inc.)
equipped with a Plan-NEOFLUAR 40×, 1.3 NA oil iris objective, CSNAP,
and Deltavision 2 (Applied Precision) software or Axiovert 200M (Carl Zeiss
MicroImaging, Inc.) equipped with a Plan-NEOFLUAR 40×, 0.75 NA ob-
jective, LSM 510 scan module (Carl Zeiss MicroImaging, Inc.), and LSM
510 META software (Carl Zeiss MicroImaging, Inc.). The acquired images
were analyzed with Multi Gauge (Fujifi lm) or LSM510 META software.
Transfection and RNAi
Neurons or HEK293T cells were transfected with cDNA or RNA by the cal-
cium phosphate method (Inagaki et al., 2001), Nucleofector (Amaxa), or
Lipofectamine 2000 (Invitrogen) before or after plating. For vector-based
RNAi analysis, we used BLOCK-iT Pol II miR RNAi expression vector kit
(Invitrogen). The targeting mRNA sequence T G A A G C T G T T A A G A A A C T G G A
corresponds to nucleotides 138–158 in the coding region of rat shootin1,
whereas the control vector pcDNA 6.2-GW/EmGFP-miR-neg encodes an
mRNA not to target any known vertebrate gene.
Antibodies against myc, tau-1, synaptophysin, MAP-2, α-tubulin, the p85
subunit of PI 3-kinase, and monoclonal (587F11) phospho-Akt (Ser473)
were obtained from MBL International Corporation, Boehringer, Progen,
Sigma-Aldrich, Sigma-Aldrich, Upstate Biotechnology, and Cell Signaling
Technology, respectively. CMFDA, Rhodamine phalloidin, blebbistatin,
cytochalasin D, and LY294002 were obtained from Invitrogen, Invitrogen,
BIOMOL Research Laboratories, Inc., Calbiochem, and Calbiochem, re-
spectively. cDNA encoding Myr-PI 3-K p110 was obtained from Upstate
Biotechnology. mRFP was provided by R. Tsien (University of California,
San Diego, La Jolla, CA).
Online supplemental material
Fig. S1 shows serial time-lapse images of EGFP-shootin1 accumulation in
neurites 1 and 2 of Fig. 2 (D–G). Fig. S2 shows DIV7 hippocampal neu-
rons overexpressing myc-shootin1, which are immunostained by anti-
synaptophysin or anti–MAP-2 antibody. Fig. S3 shows the effects of
cytochalasin D on shootin1 distribution in hippocampal neurons. Fig. S4
shows that inhibition of PI 3-kinase activity suppresses formation of
shootin1-induced multiple axons, but repression of shootin1 expression by
RNAi does not inhibit formation of PI 3-kinase–induced multiple axons.
Video 1 is a time-lapse video of a stage 2 hippocampal neuron expressing
EGFP-shootin1 as described in Fig. 2 A. Video 2 is a time-lapse video of a
hippocampal neuron expressing EGFP-shootin1 taken from stages 2 to 3
as described in Fig. 2 (D–G). Video 3 is a time-lapse video of a hippocam-
pal neuron overexpressing EGFP-shootin1 as described in Fig. 3 A. Online
supplemental material is available at http://www.jcb.org/cgi/content/
We thank Dr. Goro Eguchi and the advisors of Recognition and Formation,
Precursory Research for Embryonic Science and Technology, for suggestions;
Drs. Hiroshi Itoh, Sadao Shiosaka, Kazunori Imaizumi, and Shin Ishii for dis-
cussion; Drs. Mu-ming Poo and Ian Smith for reviewing the manuscript; Drs.
Takahiro Nagase and Osamu Ohara for providing the cDNA encoding
KIAA1598; and Ms. Kumiko Motegi for secretarial assistance. We also thank
Drs. Jun-ichi Miyazaki and Roger Y. Tsien for providing pCAGGS and mRFP
This research was supported in part by Precursory Research for Embry-
onic Science and Technology at Japanese Science and Technology Corpora-
tion; the Ministry of Education, Sports, Culture, Science and Technology
(18300107); and the Japan Society for the Promotion of Science KAKENHI
(18016020 and 18022028), the Suntory Institute for Bioorganic Research,
and the Osaka Medical Research Foundation for Incurable Diseases.
Submitted: 26 April 2006
Accepted: 18 September 2006
Andersen, S.S., and G.Q. Bi. 2000. Axon formation: a molecular model for the
generation of neuronal polarity. Bioessays. 22:172–179.
Ando, R., H. Hama, M. Yamamoto-Hino, H. Mizuno, and A. Miyawaki. 2002. An
optical marker based on the UV-induced green-to-red photoconversion of
a fl uorescent protein. Proc. Natl. Acad. Sci. USA. 99:12651–12656.
Arimura, N., and K. Kaibuchi. 2005. Key regulators in neuronal polarity. Neuron.
Banker, G. 2003. Pars, PI 3-kinase, and the establishment of neuronal polarity.
Bradke, F., and C.G. Dotti. 1999. The role of local actin instability in axon
formation. Science. 283:1931–1934.
Brown, A. 2003. Axonal transport of membranous and nonmembranous cargoes:
a unifi ed perspective. J. Cell Biol. 160:817–821.
Craig, A.M., and G. Banker. 1994. Neuronal polarity. Annu. Rev. Neurosci.
Da Silva, J.S., and C.G. Dotti. 2002. Breaking the neuronal sphere: regulation
of the actin cytoskeleton in neuritogenesis. Nat. Rev. Neurosci.
SHOOTIN1 AS A NEURONAL POLARITY SIGNAL • TORIYAMA ET AL. 157 Download full-text
Dotti, C.G., and G.A. Banker. 1987. Experimentally induced alteration in the
polarity of developing neurons. Nature. 330:254–256.
Dotti, C.G., C.A. Sullivan, and G.A. Banker. 1988. The establishment of polarity
by hippocampal neurons in culture. J. Neurosci. 8:1454–1468.
Esch, T., V. Lemmon, and G. Banker. 1999. Local presentation of substrate
molecules directs axon specifi cation by cultured hippocampal neurons.
J. Neurosci. 19:6417–6426.
Goslin, K., and G. Banker. 1989. Experimental observations on the devel-
opment of polarity by hippocampal neurons in culture. J. Cell Biol.
Goslin, K., and G. Banker. 1990. Rapid changes in the distribution of GAP-43
correlate with the expression of neuronal polarity during normal devel-
opment and under experimental conditions. J. Cell Biol. 110:1319–1331.
Horton, A.C., and M.D. Ehlers. 2003. Neuronal polarity and traffi cking. Neuron.
Inagaki, N., and K. Katsuta. 2004. Large gel two-dimensional electrophoresis:
improving recovery of cellular proteome. Current Proteomics. 1:35–39.
Inagaki, N., K. Chihara, N. Arimura, C. Menager, Y. Kawano, N. Matsuo, T.
Nishimura, M. Amano, and K. Kaibuchi. 2001. CRMP-2 induces axons
in cultured hippocampal neurons. Nat. Neurosci. 4:781–782.
Jacobson, C., B. Schnapp, and G.A. Banker. 2006. A change in the selective
translocation of the Kinesin-1 motor domain marks the initial specifi ca-
tion of the axon. Neuron. 49:797–804.
Jiang, H., W. Guo, X. Liang, and Y. Rao. 2005. Both the establishment and the
maintenance of neuronal polarity require active mechanisms: critical
roles of GSK-3β and its upstream regulators. Cell. 120:123–135.
Lasek, R.J., J.A. Garner, and S.T. Brady. 1984. Axonal transport of the cyto-
plasmic matrix. J. Cell Biol. 99:212s–221s.
Menager, C., N. Arimura, Y. Fukata, and K. Kaibuchi. 2004. PIP3 is involved
in neuronal polarization and axon formation. J. Neurochem. 89:109–118.
Nishimura, T., T. Yamaguchi, K. Kato, M. Yoshizawa, Y. Nabeshima, S. Ohno,
M. Hoshino, and K. Kaibuchi. 2005. PAR-6-PAR-3 mediates Cdc42-
induced Rac activation through the Rac GEFs STEF/Tiam1. Nat. Cell
Niwa, H., K. Yamamura, and J. Miyazaki. 1991. Effi cient selection for high-
expression transfectants with a novel eukaryotic vector. Gene. 108:193–199.
Nomura, E., K. Katsuta, T. Ueda, M. Toriyama, T. Mori, and N. Inagaki. 2004.
Acid-labile surfactant improves in-sodium dodecyl sulfate polyacryl-
amide gel protein digestion for matrix-assisted laser desorption/ionization
mass spectrometric peptide mapping. J. Mass Spectrom. 39:202–207.
Oguri, T., I. Takahata, K. Katsuta, E. Nomura, M. Hidaka, and N. Inagaki. 2002.
Proteome analysis of rat hippocampal neurons by multiple large gel two-
dimensional electrophoresis. Proteomics. 2:666–672.
Rolls, M.M., and C.Q. Doe. 2004. Baz, Par-6 and aPKC are not required for axon
or dendrite specifi cation in Drosophila. Nat. Neurosci. 7:1293–1295.
Ruthel, G., and G. Banker. 1998. Actin-dependent anterograde movement of
growth-cone-like structures along growing hippocampal axons: a novel
form of axonal transport? Cell Motil. Cytoskeleton. 40:160–173.
Ruthel, G., and G. Banker. 1999. Role of moving growth cone-like “wave”
structures in the outgrowth of cultured hippocampal axons and dendrites.
J. Neurobiol. 39:97–106.
Schwamborn, J.C., and A.W. Puschel. 2004. The sequential activity of the
GTPases Rap1B and Cdc42 determines neuronal polarity. Nat. Neurosci.
Shi, S.H., L.Y. Jan, and Y.N. Jan. 2003. Hippocampal neuronal polarity speci-
fi ed by spatially localized mPar3/mPar6 and PI 3-kinase activity. Cell.
Shi, S.H., T. Cheng, L.Y. Jan, and Y.N. Jan. 2004. APC and GSK-3beta are in-
volved in mPar3 targeting to the nascent axon and establishment of neu-
ronal polarity. Curr. Biol. 14:2025–2032.
Straight, A.F., A. Cheung, J. Limouze, I. Chen, N.J. Westwood, J.R. Sellers, and
T.J. Mitchison. 2003. Dissecting temporal and spatial control of cytokine-
sis with a myosin II inhibitor. Science. 299:1743–1747.
Wiggin, G.R., J.P. Fawcett, and T. Pawson. 2005. Polarity proteins in axon speci-
fi cation and synaptogenesis. Dev. Cell. 8:803–816.
Winckler, B., and I. Mellman. 1999. Neuronal polarity: controlling the sorting
and diffusion of membrane components. Neuron. 23:637–640.
Yoshimura, T., Y. Kawano, N. Arimura, S. Kawabata, A. Kikuchi, and K.
Kaibuchi. 2005. GSK-3β regulates phosphorylation of CRMP-2 and neu-
ronal polarity. Cell. 120:137–149.
Yoshimura, T., N. Arimura, Y. Kawano, S. Kawabata, S. Wang, and K. Kaibuchi.
2006. Ras regulates neuronal polarity via the PI3-kinase/Akt/GSK-3β/
CRMP-2 pathway. Biochem. Biophys. Res. Commun. 340:62–68.