Cell, Vol. 110, 223–235, July 26, 2002, Copyright 2002 by Cell Press
Axonal Protein Synthesis Provides a Mechanism for
Localized Regulation at an Intermediate Target
Sabry, 1995; Alvarez et al., 2000; Lodish et al., 2000;
Schwartz and DeCamilli, 2000). However, other studies
over a period of four decades based on labeled amino
acid incorporation into isolated axon preparations, in-
hibitors of translation, and detection of various compo-
nents of protein translation machinery have led to in-
creasing interest in the idea of RNA translation by
vertebrate axons (Koenig, 1967; Bassell et al., 1998; Eng
reviewed by Alvarez et al., 2000). The further question
sized proteins such as guidance receptors to the cell
surface remains unknown.
By far the best-characterized intermediate target is
Goodman, 1996; Brose and Tessier-Lavigne, 2000; Ka-
prielian et al., 2001; Yu and Bargmann, 2001). In the
Drosophila and nematode ventral nerve cord, and the
vertebrate spinal cord, commissural axons initially grow
toward the midline, chemoattracted by netrins. After
crossing, they are prevented from recrossing, and they
are guided into longitudinal tracts on the contralateral
side, even though they had ignored the same tracts
before crossing. This scheme implies at least two types
of regulation in growth cones as they cross the midline
(Flanagan and Van Vactor, 1998; Stein and Tessier-
attractants, allowing them to escape on the other side.
A mechanism for this has been identified in a receptor-
receptor interaction between DCC, a netrin receptor,
and Robo, a receptor for midline Slit protein (Stein and
Tessier-Lavigne, 2001). (2) Axons gain responsiveness
to a constellation of new guidance cues, which prevent
Important progress has also been made in under-
standing this second type of regulation, beginning with
early observations of surface proteins localized to spe-
cific segments of axons (Bastiani et al., 1987; Dodd
et al., 1988). In Drosophila, Robo, Robo2, and Robo3
receptors are all upregulated on distal axon segments
after crossing and mediate responses to repellent mid-
line Slit protein, which prevents recrossing and also
serves as a cue for axons to select among contralateral
longitudinal tracts (Kidd et al., 1998, 1999; Rajagopalan
et al., 2000; Simpson et al., 2000). Similarly, in verte-
brates, receptors including L1 and EphB1 are upregu-
lated on distal axon segments after crossing, and regu-
latedresponsiveness tocuesinthe Slit,Sema,adhesion
molecule, and ephrin families appears to control cross-
ing and guide axons into correct longitudinal pathways
(Dodd et al., 1988; Stoeckli and Landmesser, 1995; Ber-
gemann et al., 1998; Burstyn-Cohen et al., 1999; Zou et
al., 2000; Imondi and Kaprielian, 2001; Kullander et al.,
2001; Yokoyama et al., 2001). While it is clear that multi-
ple proteins are upregulated on distal axon segments
after crossing and that this can play an important role
in guidance responses, the mechanism for this upregu-
lation is not known.
Here, axonal protein synthesis was studied by an ap-
Perry A. Brittis, Qiang Lu, and John G. Flanagan1
Department of Cell Biology
and Program in Neuroscience
Harvard Medical School
Boston, Massachusetts 02115
As axons grow past intermediate targets, they change
their responsiveness to guidance cues. Local upregu-
anisms for this are not clear. Here protein synthesis
is traced within individual axons by introducing RNAs
ons and growth cones can translate proteins and also
export them to the cell surface. As axons reach the
spinal cord midline, EphA2 is among the receptors
upregulated on at least some distal axon segments.
Midline reporter upregulation is recapitulated by part
of the EphA2 mRNA 3? untranslated region, which is
highly conserved and includes known translational
control sequences. These results show axons contain
all the machinery for protein translation and cell sur-
face expression, and they reveal a potentially general
and flexible RNA-based mechanism for regulation lo-
calized within a subregion of the axon.
To set up the pattern of projections in the nervous sys-
tem, axonal growth cones must find their targets by
migrating along pathways that may be long and com-
finding process is the presence of intermediate targets,
which can break the journey up into smaller segments
(Tessier-Lavigne and Goodman, 1996). Axons are ini-
tially guided toward such intermediate targets, then
grow past them and adopt a new trajectory on the other
to change their responsiveness to multiple guidance
cues as they pass from one side of an intermediate
target to the other. Comparable changes in respon-
siveness to multiple cues must also occur as projecting
axons reach their final target.
In principle, one mechanism to regulate axon respon-
siveness could be to synthesize proteins, such as cell
surface receptors, within the distal segment of the axon
after reachingan intermediatetarget. Whilecytoplasmic
polyadenylation and local translation has been studied
extensively as a mechanism for synapse regulation in
dendrites (Richter, 1999; Martin et al., 2000; Wells et al.,
2000; Job and Eberwine, 2001; Steward and Schuman,
2001), a long-prevailing view was that vertebrate axons
are not capable of protein synthesis, based on studies
that failed to detect ribosomes, and the ability of axons
to transport proteins from the cell body (see Tanaka and
proach using RNA constructs introduced by viral vector
or electroporation, combined with reporters that can be
visualized with subcellular resolution: fluorescent pro-
teins (FPs) to detect local translation, or membrane-
anchored alkaline phosphatase (AP) to detect cell sur-
face expression. This approach was used to show that
individual isolated axons and growth cones are capable
of not only protein translation, but also export to the
cell surface. We then went on to set up a model system
for electroporation of developing spinal cord, to investi-
lated region (3? UTR) of EphA2 receptor mRNA contains
a sequence that is highly conserved and contains a
cytoplasmicpolyadenylation element(CPE). UsingGFP,
as well as the Fluorescent Timer reporter, this 3? UTR
sequence was found to direct reporter expression to
distal axon segments, recapitulating the upregulation of
EphA2 at the midline. These results reveal an RNA-
based mechanism for regulation of protein expression
localized within a specific region of the axon.
of axon isolation, RNA was prepared from retinal strips
tion-polymerase chain reaction (RT-PCR) by similar
methodology to that used for single-cell RT-PCR. RNA
for histone-H1, a nuclear protein, was seen prominently
in the explants but could not be detected in the axon
carpets (Figure 1A, ii, insets). As in previous studies of
this type, purity of the axon preparation from all cell
body material cannot be guaranteed. Specificity here is
addressed by direct introduction of RNA constructs and
the use of visualizable reporters in individual axons.
synthesis by Sindbis-GFP virus infection. When exam-
ined after 6 hr, most axons in the culture contained
GFP fluorescence (Figure 1B, i–vi) (n ? 7 independent
cultures). In no case was a connection to a cell body
seen, and most fluorescent axons could be unambigu-
ously traced back to a cut end. Fluorescence was con-
centrated in granules, with additional diffuse labeling.
The granules were dispersed along the entire axon and
(Figure 1B, iii and v) and in the growth cone, even when
1B, iv and vi). All detectable GFP fluorescence was
blocked by inclusion of the protein synthesis inhibitor
cycloheximide along with the virus, in cut axon carpets,
or in cultures where the axons were left intact (data not
shown). These experiments show isolated axons and
growth cones can be infected by Sindbis RNA virus. By
detecting newly synthesized protein with resolution to
individual cut axons, they also show that isolated axons
and growth cones contain all the machinery needed for
To further test axonal translation in a context con-
taining no neuronal cell bodies, we tested a cut nerve
preparation. The optic nerve is a relatively simple struc-
ture, containing glial cells and long axons that project
from retinal ganglion cell bodies in the retina. Postnatal
day 3 (P3) mouse optic nerves were separated from the
retina and the rest of the brain and then cultured with
Sindbis-GFP virus. Nine hours later, GFP fluorescence
was seen in cells with a glial morphology and in GFP-
positive retinal axons (Figure 1C) (n ? 5 optic nerves).
Although we cannot rule out the possibility of protein
transfer from glial cells in this experiment, the results
even when separated from their cell body.
The presence of components potentially involved in
local translation was examined by immunolocalization.
In oocytes and dendrites, local translation is mediated
by CPE elements in the 3? UTR. These bind CPEB pro-
tein, a key component of a complex that regulates cyto-
plasmic polyadenylation and translation (Richter, 1999).
(Figure 1A, iv–vi). These results are consistent with the
idea that axonal translation could be regulated by a
mechanism involving CPEB and CPE elements. Axon
carpets were also tested for mRNA sequences by RT-
PCR, supporting the presence of RNA for the cytoskele-
tal protein ?-actin as reported previously (Bassell et al.,
Translation and Cell Surface Expression in
Individual Isolated Axons and Growth Cones
To assess the ability of developing vertebrate axons to
produce proteins locally, an approach based on an RNA
viral vector was tested. We speculated such a vector
might allow RNAs to be introduced directly into axons,
avoiding the problem of microinjecting thin vertebrate
axons, or using DNA vectors that rely on the nucleus.
Sindbis has a single-stranded RNA genome that func-
tions as an mRNA translated by host cell machinery. It
can infect a wide range of cells and is endocytosed by
binding to cell surface proteins including laminin recep-
tors (Strauss and Strauss, 1994). A comparable ap-
proach based on Sindbis virus was recently described
in an independent study to confirm protein synthesis in
dendrites (Aakalu et al., 2001), although a significant
difference is that rather than infecting cells and later
isolating neurites, here we infect only after neurite isola-
tion, showing that axons themselves can be infected
directly, and eliminating any contribution from the intact
Cut axons separated from their cell body in vitro are
known to survive and continue growing (Shaw and Bray,
uniform carpet of long axons can be obtained in culture.
Retinal explants were grown on laminin, then the axons
were cut, and the cell bodies were removed by aspira-
tion, leaving a carpet of isolated axons and growth
cones. To assess the effectiveness of this approach in
producing isolated axons, cultures were fixed and dou-
ble labeled for axonal protein GAP-43 and a nuclear
stain (Figure 1A, i–iii). All axons observed beyond the
cut were separated from their cell bodies, and in the
relatively low-density carpets used for the experiments
below, they could generally be traced back to a cut end.
Note that a few cells did migrate out of the explant in
some cultures. However, they remained close to the
explant (Figure 1A, iii) and had a glial-like morphology
(data not shown). Further supporting the effectiveness
RNA-Based Local Regulation in Axons
1998), as well as cell surface proteins NCAM and EphB2
(Figure 1A, ii, inset).
Although mRNA translation in vertebrate axons has
been proposed previously, this could not by itself ac-
count for local expression of key surface components
such as guidance receptors, unless axons can also lo-
cally export proteins. This was assessed here using an
AP reporter. A Sindbis construct was made encoding
native placental alkaline phosphatase (PLAP), a glyco-
protein with an amino-terminal secretion signal peptide
and a carboxy-terminal glycosyl phosphatidylinositol
tail that anchors it in the membrane (see Flanagan et
al., 2000). AP activity has been used extensively as a
reporter to determine cytoplasmic versus extracyto-
plasmic localization of fused proteins, since enzyme
activity requires export to an extracytoplasmic location
(reviewed by Manoil et al., 1990). After Sindbis-PLAP
treatment, isolated axon carpets showed prominent in
(Figure 1B, vii–xi). Staining was seen over axons and
growth cones, extending to the edges of the membrane,
including lamellipodia and filopodia. Some of the stain-
ing was concentrated in patches, though it is not known
aggregation. Although PLAP activity indicates translo-
cation to an extracytoplasmic location, it does not nec-
essarily demonstrate transport to the cell surface. This
was further tested using an antibody to detect surface
immunoreactivity of the PLAP reporter on unfixed, un-
permeabilized axons. Prominent surface binding was
to proteins on unfixed membranes (Figure 1B, xii–xiv).
These results indicate that axons and growth cones are
not only capable of local RNA translation as proposed
previously, but also contain all the machinery to translo-
cate proteins to an extracytoplasmic compartment and
insert them in the surface membrane.
low-density culture as used for reporter experiments. A few cells
migrated out (arrowhead) but were not observed beyond the cut
site. Most axons could be traced to a cut end (arrows). (iv–vi) E14
rat retinal axons grown on laminin and labeled with antibodies to
GAP-43 (red) and the translational regulation protein CPEB (green).
CPEB immunoreactivity is seen in granules within axon shafts and
growth cones (arrowheads).
(B) Reporter expression in cut axon carpets treated with Sindbis-
GFP or Sindbis-PLAP virus. (i) GFP fluorescence after Sindbis-GFP
infection. Axons were cut near the lower edge and extend toward
ing GFP fluorescence concentrated in granules in axon shafts
(arrow) and at the base of short collateral spikes (open arrowhead)
and growth cones (filled arrowhead). (vii) Staining for AP enzyme
activity after Sindbis-PLAP infection. (viii) Negative control virus.
(ix–xi) Higher magnification of cut distal axon segment. AP activity
is visible over the axon shaft (arrows) and growth cone, including
lamellipodia (arrowhead) and filopodia (open arrowhead). (xii–xiv)
Cut axons infected with Sindbis-PLAP, then treated unfixed and
unpermeabilized for surface staining with antibody to myc tag in the
right, overlap. Fluorescence is seen over the axon in a patched
(C) Reporter expression in cut optic nerve treated with Sindbis-GFP
virus. (i) Procedure. (ii and iii) Isolated optic nerve whole mount
showing GFP fluorescence in cut axons (arrows) coursing through
GFP-positive glial cells (arrowhead).
Figure 1. Local Protein Synthesis and Cell Surface Expression in
Isolated Axons and Growth Cones
(red) growing out from chick retinal explants were separated from
their cell bodies (blue) by cutting with a microknife. Inset: cut axon
carpet at low magnification with the explant removed. (ii) Cut axons,
double-labeled with GAP-43 antibody (red) for axons and cell bod-
ies, plus Bisbenzimide (blue) for nuclei. Inset: RT-PCR of RNA from
retinal strips or isolated axon carpets. (iii) Higher magnification of
3? UTR Conservation and Spinal Cord Expression
of EphA2 Receptor
We were next interested to test whether axonal protein
synthesis might provide a mechanism for local regula-
tion at an intermediate target. For these studies, we
used spinal commissural axons, the best-characterized
model for an intermediate target in vertebrates. Com-
missural interneurons in the spinal cord send out an
axon that grows ventrally toward the midline floor plate,
crosses the floor plate, and then turns to grow longitudi-
nally (Figure 2B, iv).
vant receptors. EphA2 has a relatively short 3? UTR that
berg and Hunter, 1990; Gilardi-Hebenstreit et al., 1992).
A species comparison revealed a strikingly conserved
region including 67 nucleotides with 100% identity (Fig-
ure 2A). Moreover, this sequence contains a motif pre-
cisely matching the CPE consensus UUUUUAU, located
a short distance upstream of the AAUAAA polyadenyla-
tion signal in the expected position to mediate transla-
EphA2 protein in developing rat was previously re-
lus (VF), which contains longitudinal tracts of spinal
upregulation was not specifically addressed. To investi-
gate this further, chick spinal cords were treated with
anti-EphA2 antibody. In transverse sections, prominent
fibrousEphA2 immunoreactivitywasseenin theVF(Fig-
ure 2B, i–ii). Labeling was also seen in some cell bodies,
particularly in ventricular and intermediate spinal cord,
consistent with expression in at least a subset of com-
missural neurons, which are widely distributed along
both dorsoventral and mediolateral axes including loca-
tions lining the ventricle (Dodd et al., 1988; Silos-Santi-
ago and Snider, 1992; Brittis et al., 1995b; Zou et al.,
2000; Kaprielian et al., 2001). However, EphA2 labeling
was not seen in ipsilateral commissural axon segments
extending ventrally toward the floor plate, and was low
or absent in the ventral commissure where axons cross
the floor plate. Comparable results were obtained in
developing rat (see Supplemental Figure S2 at http://
open book whole mounts viewed ventrally, EphA2 label-
ing was seen on the spinal cord surface in a fibrous
pattern with a predominantly longitudinal orientation,
consistent with upregulation on crossed fibers (Figure
2B, iii). When individual commissural axons were traced
longitudinal fiber tracts. (iii) Inverted open book spinal cord whole
mount, viewed ventrally, showing EphA2 immunofluorescence (red).
GFP-pA DNA was electroporated unilaterally (to the right here) to
trace axons. Inset shows higher magnification of VF. Axons cross
the midline, turn, and grow rostrally in longitudinal tracts that show
fibrous EphA2 immunoreactivity. (iv) Diagram. VF, ventral funiculus;
FP, floor plate; D, dorsal; V, ventral.
(C) Higher magnification of inverted open book. In laser confocal
sections, EphA2 immunofluorescence (red) overlaps GFP in growth
cones of crossed axon segments (i–iii), and cell bodies of commis-
sural neurons with axons that can be traced across the midline
(iv–vi). Arrow, crossed growth cone; arrowhead, cell body. Higher
resolution images of (i)–(iii) are in Supplemental Figure S1 at http://
Figure 2. EphA2Receptor: Conserved3? UTRSequence andImmu-
nolocalization in Spinal Cord
(A) EphA2 mRNA includes a region (red box) of 67 nucleotides in
the 3? UTR that is 100% conserved between mouse (upper se-
quence) and human (lower sequence), including a CPE motif (green,
underlined) andhexanucleotide polyadenylation signal(blue, under-
(B) EphA2 immunoreactivity in E5 spinal cord. In sections (i) and (ii),
EphA2 immunoreactivity (green) was strong in longitudinal axon
tracts including the VF (arrows), but was not seen in ipsilateral axon
segments growing ventrally toward the midline and was weak to
undetectable in crossing axons in the ventral commissure beneath
the floor plate (FP). Control GAP-43 antibody (red) shows labeling
in ipsilateral axon segments, axons crossing the floor plate, and
RNA-Based Local Regulation in Axons
by electroporation with DNA encoding a GFP reporter,
as described further below, laser confocal microscopy
revealed EphA2 immunoreactivity overlapping at least
some crossed axon segments and their associated cell
bodies, but not ipsilateral segments of crossed axons
ent on all commissural axons, nor whether there could
be additional expression on axons that project in ipsilat-
eral longitudinal tracts, the results indicate upregulation
of EphA2 on distal segments of at least a subset of
A Model System to Study Regulation
in Commissural Axons
To study the mechanism of regulation, we next wanted
to find a method to introduce RNA constructs into com-
missural axons. Electroporation has been used to intro-
duce DNA into spinal cord cells at early patterning
stages (Briscoe et al., 2000) and to introduce RNA into
cultured neurons (Teruel et al., 1999). In initial experi-
ments, we used GFP-pA DNA, containing a CMV tran-
tion region from SV40. Spinal cords were prepared in
an upside-down open book format. In every preparation
analyzed (n ? 50 spinal cords), a population of neurons
was labeled in dorsal and intermediate spinal cord and
hindbrain (Figure 3A). Although they are not necessarily
all commissural neurons, many had long axons that
longitudinal tracts (Figures 3A–3E). In addition to com-
missural axons that remained within the VF, many fibers
grew longitudinally within the VF for a short distance
and then curved away in a sigmoid path to enter more
dorsal longitudinal tracts (Figure 3D), as recently re-
ported for commissural axons iontophoretically labeled
with dye (Imondi and Kaprielian, 2001). Spinal cords
could be electroporated not only bilaterally, but also
unilaterally with different reporters (Figure 3C), allowing
cell bodies and axons from each side to be viewed
tralateral pathways. Altogether, these results indicate
that the specific conditions used here allowed efficient
electroporation of spinal commissural neurons with
high-resolution tracing of individual axons and growth
Figure 3. Electroporation of Reporter into Commissural Neurons
E5 chick spinal cords, whole mounted as an inverted open book,
ventricular face down, were electroporated with GFP-pA DNA.
(A) Overview of posterior hindbrain and spinal cord. Most labeled
cells are in dorsal and intermediate regions containing commissural
neurons, and many have long axons that can be traced across the
(B) Higher magnification of a commissural growth cone, showing
complex morphology characteristic of growth cones encountering
an intermediate target (Brittis et al., 1995a).
(C) Spinal cord electroporated with GFP-pA on the right and RFP-
pA on the left, showing unilateral labeling. This example was photo-
graphed a short time after electroporation, so axon labeling is not
(D) After entering the VF, some axons curve away in a sigmoid path
to enter more dorsal tracts.
across the midline and into contralateral longitudinal tracts (arrows;
growth cone seen on right). ML, midline; CB, cell body.
Upregulated Reporter Expression in Contralateral
This electroporation system was next used for expres-
conserved region of the EphA2 3? UTR might regulate
expression atthe midline,we testedGFP-3?EphA2 RNA,
containing a GFP reporter followed by 77 nucleotides
from the 3? UTR of EphA2 (Figure 2A). GFP was seen in
a population of cell bodies indistinguishable from those
labeled after electroporation of GFP-pA DNA. Axon la-
beling was also seen, but in a pattern different from the
DNA construct (Figures 4A–4C). To initially categorize
the distribution, we divided the neuroepithelium into
three zones as indicated in Figure 4P. Prominent GFP
fluorescence was seen in distal axon segments that had
reached or crossed the midline in zone 3 and also in
ulation of GFP reporter protein levels in distal axon seg-
ments that had reached or crossed the midline.
tion was used. GFP-3?EphA2 RNA was electroporated
simultaneously with RFP-3?pA DNA to provide a red
fluorescent internal control that would allow tracing of
the whole axon without a gap (Figure 5). Prominent dou-
ble labeling was seen in approximately 30% of labeled
neurons. In all double-labeled neurons, both RFP and
the axon was seen in individual axons where GFP label-
ing was localized to the distal segment: in two-color
overlaps, these axons appear yellow in the distal seg-
ment and red in the more proximal segment ipsilateral
to the midline. This is not due to a difference between
RFP and GFP, since GFP-pA DNA gave labeling along
the axon indistinguishable from RFP-pA DNA (data not
shown). Some GFP reporter expression appeared to
begin as axons enter the floor plate and reached high
levels as they exit the floor plate (Figure 5A). Expression
then appeared to persist in crossed axon segments as
they turn to grow longitudinally, especially in the most
distal portion including the growth cone (Figure 5B).
Mechanistically, our working model from these results
was that the conserved sequence in the EphA2 3?UTR,
including the CPE element, might upregulate polyade-
nylation and therefore translation in the distal axon seg-
ment. In contrast, the RFP-pA DNA construct would
specify immediate RNA polyadenylation, allowing un-
regulated translation throughout the axon. These dou-
ble-labeling results suggested that the elevated expres-
sion of GFP-3?EphA2 RNA in distal axon segments is
not solely a property of the protein itself, but rather is
dependent on the manner in which the protein is en-
coded. They also confirm that the gap in zone 2 is not
an optical or tissue preparation artifact.
This double-labeling procedure allowed us to directly
detectable in experiments using GFP-3?EphA2 RNA
only. RFP-labeled ipsilateral growth cones projecting
toward the midline never showed prominent green fluo-
rescence (Figure 6A). This confirms that GFP-3?EphA2
RNA is not simply expressed at high levels in all growth
cones, but rather is specifically upregulated in growth
cones that have reached or crossed the midline. When
ipsilateral and contralateral growth cone fluorescence
was compared quantitatively, ipsilateral growth cones
slightly above background (Figures 6B–6D). In compari-
son, a strong elevation of GFP fluorescence was seen
ingrowth conesthathad reachedorcrossed themidline
(p ? 0.0001; unpaired t test) (Figure 6D). The measured
increase was approximately 12-fold, which may be an
underestimate since some pixels were saturated in al-
most all contralateral growth cone images (see Figure
6C and legend).
Figure 4. Conserved 3? UTR RNA Sequences Confer Reporter Ex-
pression in Cell Bodies and Crossed Axon Segments
(A–O) RNAs were electroporated into E5 spinal cord. Right column
shows higher magnification of contralateral axon segments from
(P) Diagram showing longitudinal division of open book into three
conceptual zones. GFP was seen in cell bodies and immediately
proximal axon segments (filled arrowheads, located in zone 1) and
in contralateral axon segments entering longitudinal tracts (arrows,
located in zone 3), but little or no axon fluorescence was visible in
zone 2. ML, midline.
(Q) E13 rat dorsal spinal neuron with anti-CPEB (green) and anti-
GAP-43 (red). Granular CPEB staining is distributed throughout the
growth cone (arrows).
(R) Western blots show CPEB and Maskin immunoreactive bands
in chick dorsal spinal cord explant cultured on laminin.
Midline Upregulation Is Dependent
on a CPE Sequence
To investigate further the requirement for a specific se-
quence in the EphA2 3? UTR, the CPE sequence was
mutated, from UUUUUAU to GGCGGAG, giving con-
segments in zone 1. However, in zone 2 there was a
gap where little fluorescence could be seen in axons or
growth cones. Note that multisection confocal analysis
was used for all tracing experiments, so the gap is not
RNA-Based Local Regulation in Axons
Figure 5. Reporter Distribution in Commissural Axons Labeled by Double Electroporation with GFP-3?EphA2 RNA, Together with RFP-pA
DNA for Axon Tracing
Additional labeling with nuclear dye (blue) in (A, iii) and (B, iii).
(A) Commissural axon with growth cone emerging from the floor plate. Higher magnification shown in (iv)–(vi). Prominent RFP is seen along
the entire axon. GFP is high in the cell body (filled arrowhead) and growth cone (arrow), which appear yellow in the overlap, but is low between
(open arrowhead). Some elevated GFP expression appears to begin where the axon enters the floor plate and reaches high levels in the
emerging growth cone.
(B) Commissural axon with growth cone that has crossed the floor plate and turned to proceed longitudinally. Prominent GFP fluorescence
is seen in the cell body and in the contralateral axon segment, especially its distal portion.
struct GFP-3?EphA2mutCPE. GFP labeling was still seen in
cell bodies and immediately proximal axon segments.
However, little fluorescence was now seen in contralat-
was quantitated by the double-labeling procedure, the
more than 8-fold (p ? 0.0001) (Figure 6D). Expression
tive analysis this upregulation did not reach statistical
significance (p ? 0.21) (Figure 6D). Expression in the
cell body was not significantly affected by the presence
or absence of the CPE sequence (p ? 0.72; mutant
intensity slightly higher) (Figure 6D). The low expression
in ipsilateral growth cones was also not significantly
affected (p ? 1.0). These results argue against a model
and proportionately transported along the axon, since
the mutation strongly affected expression in contralat-
eral axons but not cell bodies. More importantly, these
mutation results demonstrate an RNA-based mecha-
nism where the conserved sequence of the EphA2 3?
required for effective upregulation of reporter expres-
sion in distal axon segments at the midline.
Since in oocytes the CPE operates via cytoplasmic
polyadenylation, the role of polyadenylation was ad-
dressed further. GFP-pAmutHexRNA, with the AAUAAA
polyadenylation signal mutated to ACCAAA, was not
expressed detectably in axons or cell bodies (data not
shown). When the requirement for this signal was cir-
cumvented by placing an artificial run of 25 A residues
immediately after the GFP reporter, to make construct
GFP-AAA, this restored translation, but there was no
obvious upregulation in distal axon segments. Labeling
was seen in cell bodies, with weak diffuse labeling
throughout the axon (Figures 4M–4O), presumably be-
cause, like the DNA constructs, GFP-AAA RNA allowed
unregulated translation throughout the axon. When a
longer tail of 200–300 A residues was added enzymati-
cally to GFP-3?EphA2 RNA, expression was seen at
greater intensity but in a similar pattern to GFP-AAA,
with expression throughout the axon (data not shown).
These results indicate RNA polyadenylation is needed
for protein expression in these experiments and appear
consistent with the idea that CPE-regulated polyadenyl-
ation may be involved in the midline upregulation in
distal axon segments.
If the mechanism for CPE recognition here is similar
to that in oocytes, axons might be expected to contain
control proteins such as CPEB and maskin. Both these
proteins were found in Western blot analysis of cultured
dorsal spinal cord (Figure 4R). Although this does not
show expression in commissural neurons at the single-
cell level, it does indicate these proteins are expressed
calization, CPEB antibodies showed a distribution com-
parable to that in retinal axons, in granules in the axon
shaft and growth cone (Figure 4Q).
Since theCPE motifin the EphA23? UTRwas required
for efficient midline upregulation, we wondered if a dif-
ferent mRNA containing CPE elements might also be
regulated at the midline. In oocytes, cyclin B1 is subject
to CPE-mediated translational control (Richter, 1999;
Mendez et al., 2000). When GFP-3?cycB1 RNA, con-
taining an 80 nucleotide region including the CPEs, was
electroporated into spinal cords, upregulation was seen
in distal axon segments, similar to that with GFP-
3?EphA2 RNA (Figures 4G–4I). This upregulation was
reduced when CPE mutations were introduced as de-
scribed previously (Richter, 1999; Mendez et al., 2000)
to make GFP-3?cycB1mutCPERNA (Figures 4J–4L), which
gave a pattern similar to GFP-3?EphA2mutCPE. These re-
sults suggest that the CPE element might be sufficient
for upregulation in axons at the midline. However, it is
likely that other parts of the highly conserved sequence
in the EphA2 3? UTR bind additional factors, which may
positively or negatively regulate RNA localization or
Fluorescent Timer Analysis of Protein Expression
Upregulated local protein synthesis within axons at the
midline would be expected to result in a preferential
localization of newly synthesized protein within these
axons. In contrast, if protein were synthesized only in
the cell body, and then transported to the growth cone,
it might be expected that soluble proteins in the cell
body would on average be newer than in the growth
cone. To investigate this, we used Fluorescent Timer, a
protein that changes its fluorescence from green to red,
at a rate that appears to be independent of protein
concentration and cellular environment (Figure 7A; Ter-
skikh et al., 2000). Since red and green may be detected
with different efficiencies, the Timer results give relative
rather than absolute time values, but red and green
become similar over a period on the order of 10 hr,
which appeared well suited for the experiments here. A
Fluorescent Timer coding sequence was linked to the
conserved region of the EphA2 3? UTR, to make con-
diately proximal axon segments (filled arrowhead) and in distal axon
segments that have reached or crossed the midline (arrow). How-
ever, little or no GFP fluorescence is visible in ipsilateral segments
of crossed axons, or in growth cones projecting toward the midline
(B) Diagram of locations quantitated.
(C) To quantitate GFP fluorescence, a region of interest (ROI) was
first selected using a contour selection tool on the red image, and
within this ROI (indicated by white line in [iii]) green fluorescence
was quantitated. Intensity of the confocal laser was set to avoid
artifactually positive conclusions: for axons, background was set
well above black to favor detection of any low green fluorescence
in ipsilateral axons, even though this meant contralateral axons
contained some saturated pixels (see [iv], showing intensity levels
in a spectrum from black, lowest, then violet through red). For cell
bodies, laser intensity was set lower, so that all cell bodies were
within the dynamic range with no pixels saturated (see iv).
(D) Quantitation results. Bars show mean and SEM.
Figure 6. Conserved 3? UTR Sequence Specifies Reporter Upregu-
lation on Contralateral versus Ipsilateral Growth Cones
GFP RNA constructs were electroporated, together with RFP-pA
DNA for axon tracing.
RNA-Based Local Regulation in Axons
struct FT-3?EphA2. After electroporation into spinal
cord, growth cones that had reached or crossed the
midline showed predominantly green fluorescence,
while cellbodies were shifted relativelymore toward red
fluorescence (p ? 0.02, unpaired t test) (Figure 7). RFP-
pA DNA could not be used for axon tracing here, but
since the density of labeled cells was kept low in this
experiment, some fluorescent distal axon segments
could be assigned to individual cell bodies, and in such
cases the axons had a higher proportion of green:red
of differentFluorescent Timer folding or degradation rates
in the axon versus the cell body, the Fluorescent Timer
results appear fully consistent with upregulation of new
protein synthesis within axons after they reach the
Two important properties of growing axons are their
ability to navigate far from the cell body and their ability
to change responsiveness to extracellular cues as they
grow toward and reach their ultimate targets. Here we
characterize mechanisms that may help explain both
these properties, showing that axons contain all the
machinery for local synthesis and cell surface expres-
sion of proteins, and identifying an RNA-based mecha-
nism that can locally regulate protein expression within
a specific segment of the axon.
Local Protein Synthesis in Axons
and Growth Cones
Protein translation in dendrites is well established as a
mechanism for synapse regulation (Martin et al., 2000;
Wells et al., 2000; Job and Eberwine, 2001; Steward
and Schuman, 2001). Translation has also been well
accepted to occur in large axons of invertebrates such
as Aplysia, although these have dendrite-like features
and are considered to differ in this regard from verte-
brate axons (Martin et al., 2000; Spencer et al., 2000).
In the case of vertebrate axons, it was widely believed
for several decades that they receive all their proteins
by anterograde transport from the cell body and are not
capable of translation (Tanaka and Sabry, 1995; Lodish
et al., 2000; Schwartz and DeCamilli, 2000). Although
many studies have long supported axonal translation,
the techniques have had inherent limitations: for exam-
ple, bulk precursor incorporation into preparations of
isolated cut axons can be due to low-level contamina-
tion with material from cell bodies; inhibitors of transla-
tion can directly or indirectly affect other cellular path-
ways; and the detection of ribosome-like particles or
various components of translation machinery does not
necessarily show the capacity for protein synthesis (re-
viewed by Alvarez et al., 2000). Nevertheless, recent
evidence has led to an increasing interest in axonal
mRNA translation and effects on axon growth and guid-
Figure 7. Fluorescent Timer Analysis
FT-3?EphA2 RNA was electroporated.
proportion of green:red fluorescence in crossed axon segment
(arrow) than cell body (arrowhead). In this example, the distal axon
segment showed almost exclusively green fluorescence.
(B) Quantitation. An ROI was created using a contour selection
toolbased on overall fluorescence, then within this ROI, green and
red fluorescence were quantitated. Laser intensity was set so that
unprocessed images were within the dynamic range with no pixels
saturated. Bars show mean and SEM.
ance (Bassell et al., 1998; Eng et al., 1999; Alvarez et
al., 2000; Koenig et al., 2000; Campbell and Holt, 2001).
Here, we have tested local translation by introducing
visualization of reporter proteins. This approach does
not rely exclusively on purification of the preparation
from cell body material, since key elements are direct
introduction of RNA and in situ visualization of newly
synthesized protein within individual cut axons and
growth cones. A similar approach was taken recently
in an independent study to confirm local translation in
dendrites (Aakalu et al., 2001), although here we further
guarded against a contribution from the cell body by
introducing RNAs only after the axons had been sev-
ered. Our results provide confirmation that protein can
be translated in individual isolated axons and growth
Within the axon, newly synthesized GFP reporter pro-
tein was localized in granules. These could correspond
to the granules containing mRNA and clusters of ribo-
some-like particles described by others (Bassell et al.,
1998; Koenig et al., 2000). The granules of new protein
were seen here in the axon shaft, at the base of lateral
microspikes, andin growth cones. Theywould therefore
seem suitably positioned to contribute newly synthe-
sized structural or regulatory proteins to regions of the
both at the axon tip and in collateral branching.
While many previous studies have proposed that in-
tracellular proteins such as cytoskeletal components
may be translated in axons, a key additional question
we have investigated this using AP as a cell surface
reporter, finding that axons and growth cones are capa-
ble of not only translation, but also export and cell sur-
to investigate the export machinery, electron micros-
copy shows intracellular membranous structures that
might represent a secretory apparatus (Koenig et al.,
2000; P.A.B. and J.G.F., unpublished data). Export of
locally synthesized proteins provides the potential for
axonal expression of both cell surface and secreted
In addition to roles in development, discussed below,
there could be a therapeutic relevance in our finding
that RNA vectors can be introduced into axons and
expressed locally, providing the potential for applica-
tions different from DNA vectors that require the nu-
lerian degeneration with RNA vectors encoding survival
proteins could promote temporary maintenance of cut
distal segments, providing a substrate for fasciculation
of regenerating axons (Davies et al., 1997) that might
help guide them back to appropriate targets.
Figure 8. Models for Regulated Local Protein Expression in Axons
(A) Regulation at the midline intermediate target. Upregulation of
as growth cones enter the midline floor plate (FP; orange) (2);
reaches high levels as they emerge from the floor plate (3); and
continues as the growth cone turns and enters longitudinal tracts
on the contralateral side (4). This pattern of regulation is seen for
EphA2inat leastsomecommissuralaxonsand canberecapitulated
by a conserved sequence from the 3? UTR of EphA2 mRNA.
axon segments. RNA (red) may be translated to produce cyto-
plasmic, secreted, or cell surface proteins (blue) locally in the axon.
Upregulated expression at the midline can be specified by a con-
served RNA sequence in the mRNA 3? UTR. Based on the require-
ment for a CPE sequence, likely mechanisms may include local
polyadenylation and consequent translation, regulated RNA trans-
port along the axon, or both.
tion. The ability to trace axons of individual electropor-
ated spinal neurons may have advantages for either
tracing of a large number of individual neurons, high
resolution imaging including both the cell body and
growth cone, and the potential for time lapse analysis
in live tissue.
An RNA-Based Mechanism for Localized
Regulation within Axons
Upregulation of receptor protein on the distal segment
appears to be an important mechanism to regulate axon
guidance. This mechanism has been shown to operate
for a number of receptors in commissural axons as they
cross the midline, including dRobo, dRobo2, dRobo3,
L1, EphB1, and EphA2 (Dodd et al., 1988; Kidd et al.,
1998; Simpson et al., 2000; Imondi and Kaprielian, 2001;
This receptor upregulation is believed to confer re-
sponsiveness to a new set of cues, allowing axons to
follow a new trajectory after crossing the midline. Al-
though the function of EphA2 was not directly investi-
crossing and contralateral longitudinal tract selection
by commissural axons (Bergemann et al., 1998; Kapriel-
Electroporation of DNA or RNA
into Commissural Neurons
The ability to introduce either DNA or RNA efficiently
into commissuralneurons provides apotentially general
tissue. In addition to introducing mRNAs, RNA electro-
poration might allow RNAi studies to block gene func-
RNA-Based Local Regulation in Axons
ian et al., 2001; Kullander et al., 2001; Yokoyama et al.,
2001). While a role for EphA2 in regulation of midline
crossing may be unlikely, since this receptor does not
bind the midline ligand ephrin-B3 efficiently, a role in
A ephrins are expressed in dorsal and intermediate spi-
nal cord (Gale et al., 1996).
We focused here on EphA2 in our studies of midline
regulation, because limited sequence is available for
other relevant 3? UTRs, which can be many kilobases
long and alternatively spliced, whereas the EphA2 3?
UTR is short and has been sequenced from both human
and mouse(Lindberg andHunter, 1990;Gilardi-Hebens-
are likely to be conserved, we found in a mouse-human
comparison that the EphA2 mRNA 3? UTR has striking
conservation of a stretch of 67 nucleotides with 100%
identity. Moreover, this sequence contains an element
of its precise fit to the UUUUUAU consensus and its
ter, 1999; Wickens et al., 2000).
When the conserved EphA2 3? UTR RNA sequence
was attached to a reporter and electroporated into em-
bryonic spinal cord, it directed upregulated protein ex-
pression to the distal segment of commissural axons
that had reached or crossed the midline floor plate.
Mutation of the CPE sequence specifically blocked this
nism to upregulate protein expression in distal axon
segments at an intermediate target. The mechanisms
and signaling pathways that mediate this upregulation
remain unknown, but based on the requirement for a
polyadenylation, regulated RNA translocation, or both
The molecular logic of axon guidance requires axons
to change their responsiveness to guidance cues as
or arrive at their final destination. One mechanism to
regulate responsiveness is transcriptional control, but
by itself this is unlikely to account for local changes in
the growth cone far from the cell body. Another class
of mechanisms is based on interactions of receptor pro-
teins or modulation of their downstream signaling path-
Lavigne, 2001). Regulation based on RNA motifs has
the advantage that RNA sequences provide an arbitrary
code independent of the structure of the encoded pro-
tein and can therefore be used flexibly to regulate multi-
ple proteins in different combinations. In the case of
regulation at intermediate targets, this could provide a
suitable mechanism for multiple receptors, as well as
cytoplasmic proteins, to be regulated coordinately, and
could also be flexible enough to regulate different pro-
teins at different targets.
In addition to intermediate targets, similar mecha-
nisms localized to the distal axon could apply to final
targets, including regulation of guidance, branching, or
synapse formation and function. Also, if regulation can be
localized not only to specific axon segments, as shown
here, but to one side of the axon shaft or growth cone,
these possibilities are mutually exclusive. The synthetic
and regulatory mechanisms characterized here could
be relevant to multiple aspects of axon development,
plasticity, and regeneration.
Reporter Expression in Cut Axons
E6 chick retinas placed on nitrocellulose filters vitreal side down
were cut in strips (100–300 ?m), then cultured 36 hr retina side
down on poly-L-lysine/laminin coverslips. Axons were cut with a
microknife in D-PBS, the explant was removed, areas proximal to
the cut were swept multiple times with a vacuum micropipette, and
removal of cell bodies was verified by phase microscopy. Sindbis
constructs were in pSinRep5 vector (Invitrogen). Axon carpets were
incubated with Sindbis-GFP virus 6 hr with periodic agitation, fixed
in 4% PFA, 0.01% glutaraldehyde, 3% sucrose, and mounted in
ProLong Antifade (Molecular Probes). For translation blocking, 35
?M cycloheximide was added with the virus. Sindbis-PLAP was
incubated with axons 9–12 hr, fixed in 60% acetone/10% formalin,
incubated 65?C overnight to inactivate endogenous AP and then
overnight in AP substrate (BCIP/NBT), postfixed in 4% PFA, and
viewed by light microscopy. Isolated optic nerves were incubated
with Sindbis-GFP virus 9 hr, fixed in 4% PFA, and whole mounted
in PBS. For surface staining, cut axons were treated unfixed with
monoclonal anti-myc tag antibody in medium 60 min with 0.1%
azide on ice, conditions that inhibit endocytosis or shedding of
surface proteins without killing cells, then were fixed and stained
with secondary antibody (negative controls were blank under the
same conditions). Images were acquired with Biorad Radiance 2000
trast were adjusted, and images were compiled for Figures with
Adobe Photoshop and Improvision Openlab software.
Electroporation of RNA or DNA into Spinal Cord
Chick E5 (stage 26) spinal cords were mounted as an upside-down
open book: opened along the dorsal midline and placed ventricular
side down on black nitrocellulose (Sartorius). 8 to 16 spinal cords
were testedfor eachRNA construct.The nitrocellulosewas trimmed
to exactly fit the length of the chamber (platinum block Petri dish
chamber, L8.3 ? W5.1 ? H3.0 mm; CUY 522; Protech) and placed
with pial side facing the cathode. For unilateral electroporations,
only one side was immersed. RNA was made by SP6 mMessage
mMachine kit (Ambion). The chamber was filled with DNA (50–75
?g/ml) or RNA (5 ?g/ml) or both, in 100 mM KAc, 30 mM HEPES-
KOH (pH 7.4), 2 mM MgAc, and given five 900 ms, 70V pulses, with
automatic resistance measurement and average automatic current
measurement of 0.51 A, using a square wave electroporator (CUY-
21; Protech International). Spinal cords were cultured 18–36 hr. To
preserve growth cone morphology, explants were fixed by carefully
adding to the medium an equal volume of 37? prewarmed 8% PFA,
3% sucrose in PBS. Specimens were counterstained with DAPI or
propidium and mounted in ProLong Antifade (Molecular Probes) for
?m z-series with a Biorad Radiance 2000.
For quantitation in Figures 6 and 7, unprocessed laser confocal
images were quantitated blind by a second investigator with Impro-
vision Openlab software. Laser intensity was set to avoid artifactu-
ally positive conclusions due to the limits of dynamic range (see
10 and 50 unless otherwise stated. For the GFP/RFP analysis,
growth cones were selected for quantitation if they could be unam-
biguously traced to a cell body using the red channel, and if both
red and green fluorescence were visible in the cell body. Cell bodies
were selected for quantitation based on the red channel only, to
avoid experimental bias in the green measurement. To control for
variations in shape and to allow measurement of growth cones
containing little or no green fluorescence, the red channel was used
to select a Region of Interest (ROI) containing each cell body, or
the distal 30 ?m of each axon, using the contour selection tool.
Within this ROI, average green intensity was calculated, and then
surrounding background intensity was subtracted. For the Timer
analysis, an ROI for each growth cone or cell body was selected
based on overall fluorescence, then average red and green intensity
were calculated within this ROI, and adjacent background was sub-
Plasmid construction, immunolocalization, RT-PCR, and Western
blot were by established techniques; see Supplemental Data at
actin, and other proteins in axons of sympathetic neurons in com-
partmented cultures. J. Neurosci. 19, 1–9.
axon guidance at the midline choice point. Cell 92, 429–432.
Flanagan, J.G., Cheng, H.J., Feldheim, D.A., Lu, Q., Hattori, M., and
Vanderhaeghen, P. (2000). Alkaline phosphatase fusions of ligands
or receptors as in situ probes for staining of cells, tissues and
embryos. Meth. Enzymol. 327, 19–35.
Gale, N.W., Holland, S.J., Valenzuela, D.M., Flenniken, A., Pan, L.,
Ryan, T.E., Henkemeyer, M., Strebhardt, K., Hirai, H., Wilkinson,
D.G., et al. (1996). Eph receptors and ligands comprise two major
specificity subclasses and are reciprocally compartmentalized dur-
ing embryogenesis. Neuron 17, 9–19.
Gilardi-Hebenstreit, P., Nieto, M.A., Frain, M., Mattei, M.-G., Ches-
tier, A., Wilkinson, D.G., and Charnay, P. (1992). An EPH-related
receptor protein tyrosine kinase gene segmentally expressed in the
developing mouse hindbrain. Oncogene 7, 2499–2506.
Imondi, R., and Kaprielian, Z. (2001). Commissural axon pathfinding
on the contralateral side of the floor plate: a role for B-class ephrins
in specifying the dorsoventral position of longitudinally projecting
commissural axons. Development 128, 4859–4871.
Job, C., and Eberwine, J. (2001). Localization and translation of
mRNA in dendrites and axons. Nat. Rev. Neurosci. 2, 889–898.
Kaprielian, Z., Runko, E., and Imondi, R. (2001). Axon guidance at
the midline choice point. Dev. Dyn. 221, 154–181.
Kidd, T., Brose, K., Mitchell, K.J., Fetter, R.D., Tessier-Lavigne, M.,
Goodman, C.S., and Tear, G. (1998). Roundabout controls axon
crossing of the CNS midline and defines a novel subfamily of evolu-
tionarily conserved guidance receptors. Cell 92, 205–215.
Kidd, T., Bland, K.S., and Goodman, C.S. (1999). Slit is the midline
repellent for the robo receptor in Drosophila. Cell 96, 785–794.
Koenig, E. (1967). Synthetic mechanisms in the axon-IV. In vitro
incorporation of [3H]precursors into axonal protein and RNA. J.
Neurochem. 14, 437–446.
Koenig, E., Martin, R., Titmus, M., and Sotelo-Silveira, J.R. (2000).
Cryptic peripheral ribosomal domains distributed intermittently
along mammalian myelinated axons. J. Neurosci. 20, 8390–8400.
Kullander, K., Mather, N.K., Diella, F., Dottori, M., Boyd, A.W., and
Klein, R. (2001). Kinase-dependent and kinase-independent func-
tions of EphA4 receptors in major axon tract formation in vivo. Neu-
ron 29, 73–84.
Lindberg, R.A., and Hunter, T. (1990). cDNA cloning and character-
ization of eck, an epithelial cell receptor protein-tyrosine kinase in
the eph/elk family of protein kinases. Mol. Cell. Biol. 10, 6316–6324.
Lodish, H., Berk, A., Zipursky, S.L., Matsudaira, P., Baltimore, D.,
and Darnell, J.E. (2000). Molecular Cell Biology (New York: W.H.
Lu, Q., Sun, E.E., Klein, R.S., and Flanagan, J.G. (2001). Ephrin-
B reverse signaling is mediated by a novel PDZ-RGS protein and
selectively inhibits G-protein coupled chemoattraction. Cell 105,
Magal, E., Holash, J.A., Toso, R.J., Chang, D., Lindberg, R.A., and
Pasquale, E.B. (1996). B61, a ligand for the Eck receptor protein-
cord neurons. J. Neurosci. Res. 43, 735–744.
Manoil, C., Mekalanos, J.J., and Beckwith, J. (1990). Alkaline phos-
phatase fusions: sensors of subcellular location. J. Bacteriol. 172,
Martin, K.C., Barad, M., and Kandel, E.R. (2000). Local protein syn-
thesis and its role in synapse-specific plasticity. Curr. Opin. Neuro-
biol. 10, 587–592.
Mendez, R., Hake, L.E., Andresson, T., Littlepage, L.E., Ruderman,
J.V., and Richter, J.D. (2000). Phosphorylation of CPE binding factor
by Eg2 regulates translation of c-mos mRNA. Nature 404, 302–307.
Rajagopalan, S., Vivancos, V., Nicolas, E., and Dickson, B.J. (2000).
Selecting a longitudinal pathway: Robo receptors specify the lateral
position of axons in the Drosophila CNS. Cell 103, 1033–1045.
Richter, J.D. (1999). Cytoplasmic polyadenylation in development
and beyond. Microbiol. Mol. Biol. Rev. 63, 446–456.
We thank Eric Lingueglia, Miriam Osterfield, Rikke Egelund, Elena
Pasquale, Seo-Hee Cho, Jamin DeProto, Jennifer Waters Shuler,
David Van Vactor, Marc Kirschner, Zaven Kaprielian, Richard Crane,
Joan Ruderman, Tom Rapoport, Dan Nowakowski, Xinmin Li, David
and Joel Richter for affinity-purified antibodies. This work was sup-
ported by grants from the NIH, an HHMI predoctoral fellowship
(P.A.B.), and a Klingenstein fellowship (J.G.F.).
Received: January 18, 2002
Revised: June 21, 2002
Published online: July 1, 2002
Aakalu, G.,Smith, W.B.,Nguyen, N.,Jiang, C.G.,and Schuman,E.M.
pal neurons. Neuron 30, 489–502.
Alvarez, J., Giuditta, A., and Koenig, E. (2000). Protein synthesis in
axons and terminals: significance for maintenance, plasticity and
regulation of phenotype with a critique of slow transport theory.
Prog. Neurobiol. 62, 1–62.
Bassell, G.J., Zhang, H., Byrd, A.L., Femino, A.M., Singer, R.H.,
Taneja, K.L., Lifshitz, L.M., Herman, I.M., and Kosik, K.S. (1998).
Sorting of ?-actin mRNA and protein to neurites and growth cones
in culture. J. Neurosci. 18, 251–265.
Bastiani, M.J., Harrelson, A.L., Snow, P.M., and Goodman, C.S.
(1987). Expression of fasciclin I and II glycoproteins on subsets of
axon pathways during neuronal development in the grasshopper.
Cell 48, 745–755.
Bergemann, A.D., Zhang, L., Chiang, M.-K., Brambilla, R., Klein, R.,
and Flanagan, J.G. (1998). Ephrin-B3, a ligand for the receptor
EphB3, expressed at the midline of the developing neural tube.
Oncogene 16, 471–480.
domain protein code specifies progenitor cell identity and neuronal
fate in the ventral neural tube. Cell 101, 435–445.
Brittis, P.A., Lemmon, V., Rutishauser, U., and Silver, J. (1995a).
Unique changes of ganglion cell growth cone behavior following
cell adhesion molecule perturbations—a time-lapse study of the
living retina. Mol. Cell. Neurosci. 6, 433–449.
Brittis, P.A., Meiri, K., Dent, E., and Silver, J. (1995b). The earliest
patterns of neuronal differentiation and migration in the mammalian
central nervous system. Exp. Neurol. 134, 1–12.
Brose, K., and Tessier-Lavigne, M. (2000). Slit proteins: key regula-
tors of axon guidance, axonal branching, and cell migration. Curr.
Opin. Neurobiol. 10, 95–102.
Burstyn-Cohen, T., Tzarfaty, V., Frumkin, A., Feinstein, Y., Stoeckli,
of commissural axons at the floor plate. Neuron 23, 233–246.
Campbell, D.S., and Holt, C.E. (2001). Chemotropic responses of
retinal growth cones mediated by rapid local protein synthesis and
degradation. Neuron 32, 1013–1026.
Davies, S.J.A., Fitch, M.T., Memberg, S.P., Hall, A.K., Raisman, G.,
and Silver, J. (1997). Regeneration of adult axons in white matter
tracts of the central nervous system. Nature 390, 680–683.
Dodd, J., Morton, S.B., Karagogeos, D., Yamamoto, M., and Jessell,
T.M. (1988). Spatial regulation of axonal glycoprotein expression on
subsets of embryonic spinal neurons. Neuron 1, 105–116.
RNA-Based Local Regulation in Axons
Schwartz, J.H., and DeCamilli, P. (2000). Synthesis and trafficking
of neuronal protein. In Principles of Neural Science, E.R. Kandel,
J.H. Schwartz, and T.M. Jessell, eds. (New York: McGraw-Hill), pp.
Shaw, G., and Bray, D. (1977). Movement and extension of isolated
growth cones. Exp. Cell Res. 104, 55–62.
Silos-Santiago, I.,and Snider,W.D. (1992).Development ofcommis-
sural neurons in the embryonic rat spinal cord. J. Comp. Neurol.
Simpson, J.H., Bland, K.S., Fetter, R.D., and Goodman, C.S. (2000).
a combinatorial code of Robo receptors controls lateral position.
Cell 103, 1019–1032.
Song, H.J., Ming, G.L., He, Z.G., Lehmann, M., McKerracher, L.,
Tessier-Lavigne, M., and Poo, M.M. (1998). Conversion of neuronal
tides. Science 281, 1515–1518.
Spencer, G.E., Syed, N.I., van Kesteren, E., Lukowiak, K., Geraerts,
W.P.M., and van Minnen, J. (2000). Synthesis and functional integra-
tion of a neurotransmitter receptor in isolated invertebrate axons.
J. Neurobiol. 44, 72–81.
Stein, E., and Tessier-Lavigne, M. (2001). Hierarchical organization
of guidance receptors: silencing of netrin attraction by slit through
a Robo/DCC receptor complex. Science 291, 1928–1938.
Steward, O., and Schuman, E.M. (2001). Protein synthesis at synap-
tic sites on dendrites. Annu. Rev. Neurosci. 24, 299–325.
Stoeckli, E.T., and Landmesser, L.T. (1995). Axonin-1, Nr-CAM, and
sural neurons. Neuron 14, 1165–1179.
Strauss, J.H., and Strauss, E.G. (1994). The alphaviruses: gene ex-
pression, replication, and evolution. Microbiol. Rev. 58, 491–562.
rearrangements during growth cone guidance. Cell 83, 171–176.
cent Timer”: protein that changes color with time. Science 290,
Teruel, M.N., Blanpied, T.A., Shen, K., Augustine, G.J., and Meyer,
T. (1999). A versatile microporation technique for the transfection
of cultured CNS neurons. J. Neurosci. Methods 93, 37–48.
Tessier-Lavigne, M., and Goodman, C.S. (1996). The molecular biol-
ogy of axon guidance. Science 274, 1123–1133.
Wells, D.G., Richter, J.D., and Fallon, J.R. (2000). Molecular mecha-
nisms for activity-regulated protein synthesis in the synapto-den-
dritic compartment. Curr. Opin. Neurobiol. 10, 132–137.
Wickens, M., Goodwin, E.B., Kimble, J., Strickland, S., and Hentze,
M.W. (2000). Translational control of developmental decisions. In
Translational Control of Gene Expression, N. Sonenberg, J.W.B.
Hershey, and M.B. Mathews, eds. (Cold Spring Harbor, NY: Cold
Spring Harbor Laboratory Press), pp. 295–370.
Yokoyama, N., Romero, M.I., Cowan, C.A., Galvan, P., Helmbacher,
F., Charnay, P., Parada, L.F., and Henkemeyer, M. (2001). Forward
signaling mediated by ephrin-B3 prevents contralateral corticospi-
nal axonsfrom recrossingthe spinalcord midline.Neuron 29,85–97.
Yu, T.W., and Bargmann, C.I. (2001). Dynamic regulation of axon
guidance. Nat. Neurosci. 4, 1169–1176.
Zou, Y., Stoeckli, E., Chen, H., and Tessier-Lavigne, M. (2000).
Squeezing axons out of the gray matter: a role for slit and semapho-
rin proteins from midline and ventral spinal cord. Cell 102, 363–375.
Note Added in Proof
in vitrogrowth cone adaptationto chemotactic cues,finds resensiti-
zation is blocked by protein translation inhibitors.