semaphorin 3a inhibits erm protein phosphorylation in growth cone filopodia through inactivation of...

Semaphorin 3A Inhibits ERM Protein Phosphorylationin Growth Cone Filopodia Through Inactivation ofPI3K

Gianluca Gallo

Department of Neurobiology and Anatomy, Drexel University College of Medicine, 2900 Queen Lane,Philadelphia, Pennsylvania 19129

Received 29 November 2007; accepted 16 January 2008

ABSTRACT: Ezrin–radixin–moesin (ERM) pro-teins are involved in the linkage of membranes tothe actin filament (F-actin) cytoskeleton. Phosphory-lation of the C-terminus activates the F-actin bindingdomain of ERM proteins by preventing the action of anautoinhibitory domain. In this study, we investigatedwhether a growth cone collapsing signal, semaphorin 3A(Sema3A), alters the state of ERM C-terminus phospho-rylation. In the growth cones of dorsal root ganglionaxons, phosphorylated ERM proteins localize to filopo-dia. We report that Sema3A inhibits ERM protein phos-phorylation in growth cone filopodia. Significantly,Sema3A decreased ERM phosphorylation prior to theonset of growth cone collapse. Over-expression of the F-actin binding fragment of ERM proteins, which com-

petes with endogenous ERM proteins for binding toF-actin, inhibited filopodial initiation and dynamics.Sema3A has been previously shown to inhibit phospho-inositide 3-kinase (PI3K) activity. Inhibition of PI3Kresulted in the loss of phosphorylated ERM proteinsfrom growth cone filopodia, and treatment with a PI3Kactivating peptide blocked the effects of Sema3A onERM phosphorylation. Collectively, these observationsdemonstrate that inactivation of PI3K in response toSema3A results in decreased phosphorylation of ERMproteins in filopodia thereby contributing to growthcone collapse. ' 2008 Wiley Periodicals, Inc. Develop Neurobiol 68:

926–933, 2008

Keywords: actin; motility; repellent; guidance; collapse;axon

INTRODUCTION

Growth cone motility is determined by cytoskeletal

dynamics and regulated by extracellular signaling.

Filopodia and lamellipodia protrude from growth

cones and search the extracellular environment for

guidance signals en route to their target tissues. The

protrusion of lamellipodia and filopodia is a major

target of guidance signals. Signals that attract the

growth cone tend to promote protrusion, while signals

that repel the growth cone terminate protrusive activ-

ity (Gallo and Letourneau, 2004). Determining themolecular mechanisms by which guidance signalsregulate growth cone protrusive activity is of impor-tance to understanding neural development andregeneration.

Filopodia and lamellipodia are supported by an

underlying actin filament (F-actin) cytoskeleton. The

dynamics of F-actin thus determine the protrusive

behavior of growth cones. F-actin is under regulation

by a large variety of proteins, which are in turn regu-

lated by extracellular signals. Ezrin–radixin–moesin

(ERM) are a class of proteins that mediate linkage

between the plasma membrane and the F-actin cyto-

skeleton and regulate filopodia (reviewed in

Bretscher et al., 2002). The three members of the

ERM protein family exhibit a large degree of conser-

vation and may have redundant functions. The

binding of ERM proteins to F-actin is regulated by

Correspondence to: G. Gallo ([email protected]).Contract grant sponsor: NIH; contract grant number: NS048090.

' 2008 Wiley Periodicals, Inc.Published online 7 March 2008 in Wiley InterScience (www.interscience.wiley.com).DOI 10.1002/dneu.20631

926

phosphorylation in the C-terminus. Phosphorylation

of ERM proteins at this site removes an inhibitory

intra-molecular interaction exposing a site that allows

ERM proteins to bind F-actins. Thus, phosphorylation

allows ERM proteins to bind F-actin and form a link

between the membrane and the cytoskeleton. Anti-

sense depletion of ERM proteins results in a simplifi-

cation of growth cone morphology (Paglini et al.,

1998). However, to our knowledge, nothing is known

about whether ERM protein phosphorylation-medi-

ated interactions with F-actin are regulated by extrac-

ellular guidance signals in growth cones.

Sema3A is a guidance signal that causes growth

cone collapse, characterized by the termination of pro-

trusive activity. During development, a gradient of

Sema3A in the ventral spinal cord prevents nerve

growth factor responsive dorsal root ganglion axons

from entering the ventral spinal cord (Fu et al., 2000).

Furthermore, semaphorin signaling is of relevance to

many nonneuronal cellular systems (Neufeld et al.,

2007; Potiron et al., 2007; Toyofuku and Kikutani,

2007). In this study we determined the effects of

Sema3A signaling during growth cone collapse on

ERM protein phosphorylation. We report that Sema3A

causes a loss of phosphorylated ERM proteins from

growth cone filopodia prior to the initiation of growth

cone collapse, suggesting that ERM proteins are an

early target of Sema3A signaling. Over-expression of

the F-actin binding fragment of moesin, which com-

petes with endogenous ERM proteins for F-actin bind-

ing, caused partial growth cone collapse and prevented

protrusive activity. Finally, inhibition of phosphoinosi-

tide 3-kinase (PI3K) activity underlies growth cone col-

lapse in response to Sema3A (Atwal et al., 2003;

Chadborn et al., 2006; Cosker and Eickholt, 2007) and

the current data demonstrate it also mediates the loss of

phosphorylated ERM proteins from growth cone filo-

podia. Collectively these data indicate that inhibition of

ERM protein phosphorylation, and thus loss of binding

to F-actin, in response to Sema3A signaling is a com-

ponent of the mechanism of growth cone collapse.

RESULTS

Distribution of PhosphorylatedERM Proteins in Sensory Growth Cones

Phosphorylated ERM proteins in nonneuronal cells

are found in filopodia and stress fibers (Bretscher

et al., 2002). To determine the distribution of phospho-

rylated and total ERM proteins in neuronal growth

cones we stained cultures of DRG neurons with anti-

bodies to total ERM proteins and their C-terminus

phosphorylated form. ERM proteins were found

throughout the axon and the growth cone [Fig. 1(A)].

However, the phosphorylated form of ERM proteins

was predominantly localized to filopodia [Fig. 1(B),

left panel]. Thus, phosphorylated ERM proteins in

growth cones are found concentrated in protrusive

structures, suggesting they may have a role in regulat-

ing protrusive activity. Interestingly, on average only

60% of filopodia exhibited phosphorylated ERM pro-

teins, and within any given growth cone not all filo-

podia were positive for phosphorylated ERM pro-

teins. Since these data are obtained from fixed growth

cones, the presence or absence of phosphorylated

ERM proteins in individual filopodia may be reflec-

tive of whether the filopodium was extending, retract-

ing or quiescent at the time of fixation. This issue will

require further analysis.

Semaphorin 3A Induces the Loss ofPhosphorylated ERM Proteinsfrom Filopodia

Sema3A treatment causes growth cone collapse, char-

acterized by the retraction of lamellipodia and filopodia

and blockade of further protrusive activity. Sema3A

treatment caused collapse of *60% of growth cones

following a 10-min treatment and greater than 90% af-

ter 20 min (not shown; Gallo, 2006). To determine the

effects of Sema3A on ERM protein phosphorylation

we treated cultures with Sema3A and fixed them at

multiple time points after treatment, followed by stain-

ing with antibodies to detect phosphorylated ERM pro-

teins. In fully collapsed growth cones there was no de-

tectable staining for phosphorylated ERM proteins (not

shown). Importantly, in Sema3A treated growth cones

that had not yet undergone collapse, and exhibited filo-

podia and lamellipodia, there was little staining for

phosphorylated ERM proteins [Fig. 1(B), right panel].

To quantify the time course of decreases in the phos-

phorylation of ERM proteins in filopodia, we deter-

mined the percentage of filopodia that contained detect-

able phosphorylated ERM proteins on Sema3A treated

growth cones that had not collapsed. This analysis

revealed a loss of phosphorylated ERM proteins from

growth cones starting at 2.5 min following Sema3A

treatment [Fig. 1(C)]. In these experiments, we focused

the analysis on growth cones that exhibited a noncol-

lapsed morphology at all time points to determine

whether the loss of ERM phosphorylated proteins from

filopodia preceded growth cone collapse. The popula-

tions of growth cones that we sampled at different time

points following Sema3A treatment had filopodial

numbers similar to those of control growth cones

ERM Proteins and Sema3A 927

Developmental Neurobiology

(p > 0.1 at all time points), confirming that the sampled

growth cones had not undergone significant collapse at

the time of fixation. Furthermore, we determined the

localization of total ERM protein in neuronal filopodia

following Sema3A treatment. As determined by the

percentage of growth cone filopodia that contained

ERM proteins, Sema3A treatment did not affect the tar-

geting of ERM proteins to filopodia [Fig. 1(D)]. These

results demonstrate that Sema3A induces the loss of

phosphorylated ERM proteins from growth cone filo-

podial prior to the full collapse of growth cones and the

loss of filopodia.

Blockade of ERM Protein Binding toF-Actin Inhibits Growth Cone FilopodiaInitiation and Dynamics

Phosphorylation of the C-terminus allows ERM pro-

teins to bind F-actin (Bretscher et al., 2002) and

Sema3A decreases phosphorylation of ERM proteins

in filopodia (see Fig. 1). Thus, to determine the

effects of inhibiting ERM protein binding to F-actin

due to the loss of phosphorylation induced by

Sema3A signaling, we determined the effects of

expressing the moesin F-actin binding fragment in

neurons (C-moesin-GFP; Litman et al., 2000). Unlike

the full-length protein, which requires phosphoryla-

tion to expose the F-actin binding fragment, the C-

moesin-GFP binds available F-actins independent of

phosphorylation (Litman et al., 2000). All members

of the ERM family of proteins have conserved C-ter-

minus F-actin binding sites (Bretscher et al., 2002),

thus expression of C-moesin-GFP F-actin binding

fragment is predicted to inhibit binding of endoge-

nous ERM proteins to F-actin.

The morphology and dynamics of growth cones

were analyzed from timelapse videos of C-moesin-

GFP expressing neurons. Using the GFP signal as a

reporter for expression levels, neurons were divided

Figure 1 Sema3A treatment results in the loss of phosphorylated ERM proteins from growth

cone filopodia. (A) Example of a growth cone double labeled with phalloidin to reveal F-actin and

an antibody that recognizes total ERM proteins (ERM). ERM proteins are found throughout the

growth cone, including filopodia (arrowheads). (B) Phosphorylated ERM proteins are localized pre-

dominantly to growth cone filopodia. The left panel shows a growth cone double labeled for F-actin

and phosphorylated ERM proteins (p-ERM). The right panel shows a growth cone treated for 5 min

with Sema3A prior to fixation and staining for F-actin and p-ERM. Note the low or absent p-ERM

staining in filopodia. (C) Quantification of the percentage of growth cone filopodia scored positive

for p-ERM staining in a blind-analysis. Sema3A induces the loss of p-ERM staining. n ¼ 65–71

growth cones per group. The metric is derived on a single growth cone basis and the mean repre-

sents the mean % filopodia positive for p-ERM staining per growth cone. Welch t-test, one-tailed,between Sema3A treatment and control (pretreatment). (D) Quantification of the percentage of

growth cone filopodia scored positive for total ERM proteins, see (A). Sema3A (5-min treatment)

did not alter ERM protein localization to filopodia. [Color figure can be viewed in the online issue,

which is available at www.interscience.wiley.com.]

928 Gallo


into low and high expressing neurons (see Materials

and Methods for details). Relative to control GFP-

expressing neurons, the axons of neurons expressing

both low and high levels of C-moesin-GFP exhibited

small growth cones [Fig. 2(A)]. To analyze growth

cone filopodial dynamics, we measured rates of filo-

podial initiation, and the net dynamics of filopodial

extension and retractions. This analysis showed that

C-moesin-GFP expression inhibited filopodial initia-

tion [Fig. 2(B)]. Analysis of filopodia existing at

growth cones at the beginning of imaging revealed

that expression of C-moesin-GFP decreased the per-

centage of time that filopodia underwent either exten-

sion or retraction [dynamic index; Fig. 2(C)], inde-

Figure 2 Expression of the inhibitory C-moesin-GFP construct impairs filopodial initiation and

dynamics. (A) Examples of control (GFP) and C-moesin-GFP transfected growth cones. Expression

of C-moesin-GFP results in small growth cones with few filopodia and lamellipodia. The terms low

and high denote the relative expression levels determined, as described in the Materials and Meth-

ods. (B) Expression of C-moesin-GFP inhibits filopodial initiation relative to GFP-expressing con-

trol growth cones. n ¼ 10–11 growth cones per group. Welch t-test, one-tailed, comparing GFP to

C-moesin-GFP was used in this and subsequent panels. (C) The filopodia of C-moesin-GFP

expressing growth cones were less dynamic than GFP controls. The dynamic index is a measure of

the % time filopodia spent either extending or retracting. 24–43 filopodia sampled in each group

from the growth cones analyzed in panel (B). (D) Expression of C-moesin-GFP decreased both the

time filopodia spent extending and retracting, respectively. Low and high expressing C-moesin-

GFP groups were merged for this analysis. (E) Expression of C-moesin-GFP increased mean filopo-

dial lifespan (sec). Low and high expressing C-moesin-GFP groups were merged for this analysis.

(F) Expression of C-moesin-GFP resulted in shorter axon lengths than in GFP expressing controls.

Both low and high expressing C-moesin-GFP neurons exhibited shorter axons than GFP expressing

neurons. n ¼ 15–45 in each group.



pendently of expression level. Analysis of the time

filopodia spent retracting and extending, respectively,

demonstrated that both phases of filopodial dynamics

were decreased by the expression of C-moesin-GFP

[Fig. 2(D)]. In addition, analysis of filopodial life-

spans, seconds that filopodia persisted following

emergence and prior to retraction into the growth

cone, revealed that C-moesin-GFP expression

resulted in more persistent filopodia [Fig. 2(E)].

Finally, Analysis of axon lengths revealed that both

high and low levels of C-moesin-GFP expression

inhibited axon extension relative to expression of

GFP alone [Fig. 2(F)]. Collectively, these data dem-

onstrate that blocking the ability of endogenous ERM

proteins to bind F-actin inhibits filopodial initiation

and decreases the dynamics of filopodia.

Inactivation of PI3K by Semaphorin 3AMediates the Loss of ERMPhosphorylation

PI3K is an established component of the Sema3A sig-

naling pathway (Atwal et al., 2003; Chadborn et al.,

2006; Cosker and Eickholt, 2007; Orlova et al.,

2007). Sema3A inactivates PI3K, and PI3K inactiva-

tion contributes to the collapse of sensory growth

cones. Indeed, inhibition of PI3K alone results in

growth cone collapse, cytoskeletal reorganization and

axon retraction, similar to the effects of Sema3A

(Orlova et al., 2007). The mechanism of semaphorin-

induced PI3K inactivation involves the targeting of

the antagonistic PTEN phosphatase to the leading

edge of the growth cone (Chadborn et al., 2006). We

therefore determined whether inhibition of PI3K

decreases the phosphorylation of ERM proteins in

growth cone filopodia. Cultures were treated with

10 lM LY294002, a well characterized inhibitor of

PI3K (Djordjevic and Driscoll, 2002), for varied time

periods prior to fixation and staining for phosphoryl-

ated ERM proteins. As previously determined in our

laboratory (Orlova et al., 2007), live imaging of

growth cone responses to LY294002 revealed that

growth cones collapsed between 3 and 7 min follow-

ing treatment. Following a 2.5 to 5-min treatment

with LY294002 phosphorylated ERM protein stain-

ing in filopodia was decreased relative to DMSO

treated controls [Fig. 3(A)]. These data suggest that

inactivation of PI3K by Sema3A signaling may be

involved in the loss of phosphorylated ERM proteins

in growth cone filopodia.

We next sought to determine whether activation of

PI3K using a cell permeable (antennapedia peptide

conjugated) PI3K activating peptide, previously

Figure 3 PI3K regulates ERM phosphorylation in growth

cone filopodia. (A) Inhibition of PI3K using LY294002

(10 lM) decreased the percentage of p-ERM positive

growth cone filopodia. Data was collected and analyzed as

in Figure 1. Inset shows a growth cone treated for 5 min

with LY294002 and stained to reveal F-actin (red) and p-

ERM (green), note the lack of p-ERM staining in filopodia

(compare with Figure 1B). n ¼ 43–56 growth cones per

group. The control group was treated with DMSO for 5

min. Welch t-test, one-tailed. (B) Treatment with peptide

consisting of a PI3K activating peptide (PI3Kpep) and the

cell permeable antennapedia peptide sequence prevents

Sema3A induced loss of p-ERM in filopodia. All peptide

treatments were at 40 lg/mL for 40 min (Chadborn et al.,

2006). Control group is representative of antennapedia pep-

tide alone treatment (not different from no treatment, not

shown). The PI3Kpep increased the percentage of growth

cone filopodia positive for p-ERM proteins relative to con-

trols. Following treatment with the antennapedia peptide, a

5-min treatment Sema3A induced a decrease in the percent-

age of filopodia positive for p-ERM similar to Sema3A

treatment alone [see Fig. 1(C)]. However, in the presence

of the PI3Kpep Sema3A failed to induced the loss of p-

ERM from filopodia. n ¼ 65–103 growth cones per group.

Welch t-test, one-tailed. [Color figure can be viewed in the

online issue, which is available at www.interscience.wiley.

com.]

930 Gallo


shown to block Sema3A induced growth cone col-

lapse (Chadborn et al., 2006; data not shown), could

prevent the loss of phosphorylated ERM proteins

from growth cone filopodia. The activating peptide

consists of the phosphorylated EGFR sequence

that activates PI3K during growth factor signaling

(Williams and Doherty, 1999; Chadborn et al., 2006).

Cultures were pretreated with the activating peptide

for 40 min prior to a 5-min treatment with Sema3A.

Treatment with the PI3K activating peptide blocked

the decrease in the percentage of filopodia containing

phosphorylated ERM proteins induced by Sema3A

[Fig. 3(B)]. Treatment with the PI3K activating pep-

tide alone increased the percentage of filopodia con-

taining phosphorylated ERM proteins relative to con-

trol levels [Fig. 3(B)]. However, we did not observe

an effect of the PI3K activating peptide on the phos-

phorylation of ERM proteins in the axon or growth

cone body (not shown), indicating a filopodium spe-

cific effect of PI3K activation on ERM protein phos-

phorylation. Collectively these data demonstrate that

PI3K is required to maintain ERM phosphorylation in

growth cone filopodia and that inactivation of PI3K

by Sema3A mediates the loss in phosphorylated

ERM proteins from filopodia.

DISCUSSION

Growth cone collapse is a complex phenomenon

involving multiple signaling pathways affecting vari-

ous aspects of the cell biology of growth cones

(reviewed in Gallo and Letourneau, 2004). In this

study we addressed to the role of ERM proteins in

Sema3A signaling. To our knowledge, this is the first

demonstration that Sema3A signaling impinges upon

the phosphorylation state, and thus F-actin binding,

of ERM proteins in growth cones. The localization of

phosphorylated ERM proteins to filopodia in growth

cones indicates specific functions for ERM proteins

bound to F-actin in regulating protrusive dynamics.

The loss of phosphorylated ERM proteins in growth

cone filopodia, prior to collapse of the growth cone,

indicates that this is an early event in the process of

growth cone collapse.

Expression of the F-actin binding fragment of

ERM proteins demonstrated a role for ERM protein

binding to F-actin in filopodial initiation. In addition,

expression of this construct also inhibited the dynam-

ics of existing filopodia. Similarly, antisense downre-

gulation of ERM proteins also simplifies hippocam-

pal neuron growth cones (Paglini et al., 1998).

Sema3A induces the translocation of the PTEN phos-

phatase to the leading edge of growth cones, resulting

in the decreased phosphorylation of phosphoinositi-

des by PI3K (Chadborn et al., 2006). The current data

determine a link between PI3K signaling and ERM

protein phosphorylation. Inhibition of PI3K caused

growth cone collapse and the loss of phosphorylated

ERM proteins from filopodia, while activation

of PI3K blocked the effects of Sema3A on ERM

protein phosphorylation, and growth cone collapse

(Chadborn et al., 2006). Although the signaling

intermediates between PI3K and kinases responsible

for ERM phosphorylation are not clear, phosphoinso-

tide phosphorylation may regulate the targeting of

ERM proteins to membranes localizing them to sub-

cellular domains where they can be phosphorylated

(Bretscher et al., 2002). Our data extend the model of

Sema3A action involving decreases in phosphoinosi-

tide phosphorylation by the PI3K/PTEN pathway

(Atwal et al., 2003; Chadborn et al., 2006; Cosker

and Eickholt, 2007; Orlova et al., 2007) by demon-

strating a link to ERM phosphorylation and thus the

regulation of filopodial initiation and dynamics. The

current data indicate that the loss of phosphorylated

ERM proteins in growth cone filopodia in response to

Sema3A is a component of the mechanism that termi-

nates filopodial protrusion and initiation during

growth cone collapse in response to Sema3A.

The signaling elements up- and downstream of

PI3K and ERM proteins in mediating growth cone

collapse remain to be identified. The activity of the

RhoA GTPase and its downstream effector RhoA-ki-

nase plays a role in Sema3A-induced growth cone

collapse (Dontchev and Letourneau, 2002, 2003;

Gallo, 2006). The ERM protein moesin can inhibit

the activity of RhoA in epithelial cells (Speck et al.,

2003). Thus, inhibition of ERM protein phosphoryla-

tion and activation by Sema3A may contribute to

RhoA activation resulting in retraction of filopodia.

Indeed, introduction of constitutively active RhoA

alone in dorsal root ganglion growth cones causes

partial growth cone collapse (Gallo, 2006). Alterna-

tively, Haas et al. (2007) report that in central nerv-

ous system neurons RhoA-kinase can drive the phos-

phorylation of ERM proteins, and RhoA-kinase and

protein kinase C have also been reported to phospho-

rylate and activate ERM proteins in nonneuronal cells

(Bretscher et al., 2002). Activation of PKC has been

involved in Sema3A induced growth cone collapse

(Mikule et al., 2003). However, since PKC phospho-

rylates ERMs and it is activated by Sema3A, it is not

a likely candidate for driving the loss of ERM phos-

phorylation during growth cone collapse. Similarly,

activation of RhoA/RhoA-kinase by Sema3A would

not be expected to decrease ERM phosphorylation.

Future work will have to address the mechanisms by



which PI3K regulates ERM phosphorylation during

Sema3A.

L1 and its close homologs (e.g., CHL1) have been

identified as functional units in the Sema3A receptor

system (Castellani, 2002). Recently, Schlatter et al.

(2007) showed that CHL1 recruits ezrin to the mem-

brane, and that this recruitment contributes to

Sema3A induced growth cone collapse. Our data

demonstrate that the binding of ERM proteins to F-

actin is inhibited by Sema3A signaling. Jointly, these

observations suggest the speculation that recruitment

of ERM proteins to the membrane is required for

Sema3A signaling leading to growth cone collapse,

but that the function of membrane recruited ezrin in

growth cone collapse is independent of its ability to

bind F-actin. ERM proteins also act as scaffolds for

other signaling proteins (e.g., protein kinase A regula-

tory subunit; Bretscher et al., 2002). Thus, the

requirement for ezrin targeting to the membrane

through association with CHL1 during Sema3A sig-

naling may reflect a scaffold function of ERM pro-

teins in growth cone collapse, independent of their

phosphorylation-dependent F-actin binding. The anti-

phospho-ERM protein antibody used in this study

reports an all three ERM proteins. In future work, we

will address which ERM protein (e.g., ezrin) is tar-

geted to and phosphorylated in filopodia.

In conclusion, this study reports that Sema3A

inhibits ERM protein phosphorylation and activity.

Although the full mechanism of growth cone collapse

is not yet understood (Gallo and Letourneau, 2004),

these data provide insights into the mechanism by

which Sema3A causes the loss of filopodia during

growth cone collapse and suggest that semaphorins

may act through ERM proteins in nonneuronal cells.

MATERIALS AND METHODS

Culturing, Reagents, and Transfection

Embryonic day 10 chicken dorsal root ganglion explants

and dissociated cells were prepared and cultured as

described previously (Loudon et al., 2006). The C-moesin-

GFP construct was a kind gift of Dr. H. Furthmayr (Litman

et al., 2000), the GFP plasmid was from Clontech (Palo

Alto, CA). Plasmids were grown using standard protocols

and transfection performed as previously described in

Loudon et al. (2006). Neurons were transfected using

the Amaxa Nucleofector and reagents (Amaxa, Cologne,

Germany). Prior to use in experiments, cultures were incu-

bated for 24 h.

Sema3A was purchased from R&D Systems (Minneap-

olis, MN). LY294002 was purchased from Calbiochem

(Carlsbad, CA). The PI3Kpep (Arg-Gln-Ile-Lys-Ile-Trp-

Phe-Gln-Asn-Arg-Arg-Met-Lys-Trp-Lys-Lys-Ser-Asp-Gly-

Gly-Tyr(PO3H2)-Met-Asp-Met-Ser; Williams and Doherty,

1999; Chadborn et al., 2006) was custom synthesized at

97% purity by American Peptide Company (Sunnyvale,

CA), dissolved at 0.5 mg/mL in F12H medium and stored

at �708C. The antennapedia peptide was also obtained

from American Peptide Company and treated as the

PI3Kpep.

Immunocytochemistry

For immunolabeling cultures were fixed with either 0.25%

glutaraldehyde (total ERM proteins) or 8% paraformalde-

hyde (p-ERM) for 15 min, washed, blocked with 10% goat

serum in PBS, extracted with 0.1% triton X-100 in PBS and

stained using rhodamine phalloidin (Molecular Probes,

Portland OR) and 1:100 or 1:50 primary antitotal ERM

(Cell Signaling, Danvers, MA) or phospho-ERM antibodies

(Cell Signaling), respectively. Staining with primary anti-

bodies was performed at 48C overnight. FITC-conjugated

antirabbit secondary antibodies were used at 1:400 (1 h),

respectively. All antibodies were applied in 10% goat se-

rum with 0.1% triton X-100 in PBS. Omission of primary

antibodies resulted in undetectable staining.

Imaging and Classification ofExpression Levels

All imaging was performed 24-h post transfection and cell

plating using a Zeiss Axiovert 200M microscope (Zeiss,

Gottingen, Germany) equipped with a 100W mercury bulb

for fluorescent imaging, a heated three-plate stage insert for

live cell imaging, and an ORCA ER camera (Hamamatsu,

Bridgewater, NJ) in series with a computer workstation run-

ning Zeiss Axiovision software for image collection and

analysis. For timelapse imaging of C-moesin-GFP, we used

minimal light by closing the aperture diaphragm, a 0.3 neu-

tral density filter, and 100–200 ms exposures using 2 3 2

binning and a 6-s inter-frame interval. All images were

stored digitally.

To classify the neurons as low or high expressors of the

GFP-constructs, individual neuronal cell bodies were visu-

ally inspected using maximal light from the lamp with the

aperture diaphragm fully open. If GFP expression was evi-

dent in the cell body, the light coming from the lamp was

decreased to minimum by closing the aperture diaphragm.

If the neuronal cell body was detectable using both maxi-

mal and minimal light, then the neuron was classified as a

high expressing neuron. Conversely, if the neuron was only

detectable using maximal, but not minimal, light then the

neuron was classified as a low expressing neuron.

Analysis of Filopodial Dynamics andAxon Lengths

Analysis of growth cone dynamics from the timelapse mov-

ies was performed on digitally manipulated images with

932 Gallo


increased brightness and contrast to reveal the full GFP sig-

nal using Axiovision software. Comparisons of the GFP

signal to phase contrast images of the same growth

cones demonstrated that in both GFP and C-moesin-GFP

expressing neurons the GFP signal accurately reported on

the growth cone morphology. Images were acquired every

6 s. The growth cone was defined as in Loudon et al.

(2006). The dynamic index for filopodia was obtained by

determining the time that a filopodium spent either extending

or retracting divided by the total time that the filopodium

was observed. Filopodial formation rate was determined by

defining the emergence of a filopodium as a linear protrusion

from the growth cone greater than 1 lm in length. Axon

lengths were measured using Axiovision software from digi-

tal collages of images acquired using a 20x objective recon-

structing the whole morphology of the neuron.

The author thank Ms. L. Silver for technical assistance,

and Dr. P.C. Letourneau (University of Minnesota) for dis-

cussion of this project. Ms. Bonnie Marsick (University

of Minnesota) kindly provided the staining protocol for

phospho-ERM proteins.

REFERENCES

Atwal JK, Singh KK, Tessier-Lavigne M, Miller FD,

Kaplan DR. 2003. Semaphorin 3F anatagonizes neurotro-

phin-induced phosphatidylinositol 3-kinase and mitogen-

activated protein kinase signaling: A mechanism for

growth cone collapse. J Neurosci 23:7602–7609.

Bretscher A, Edwards K, Fehon RG. 2002. ERM proteins

and merlin: Integrators at the cell cortex. Nat Rev Mol

Cell Biol 3:586–599.

Castellani V. 2002. The function of neuropilin/L1 complex.

Adv Exp Med Biol 515:91–102.

Chadborn NH, Ahmed AI, Holt MR, Prinjha R, Dunn GA,

Jones GE, Eickholt BJ. 2006. PTEN couples Sema3A

signaling to growth cone collapse. J Cell Sci 119:951–

957.

Cosker KE, Eickholt BJ. 2007. Phosphoinositide 3-kinase

signaling events controlling axonal morphogenesis. Bio-

chem Soc Trans 35:207–210.

Dontchev VD, Letourneau PC. 2002. Nerve growth factor

and semaphorin 3A signaling pathways interact in regu-

lating sensory neuronal growth cone motility. J Neurosci

22:6659–6669.

Dontchev VD, Letourneau PC. 2003. Growth cones inte-

grate signaling from multiple guidance cues. J Histochem

Cytochem 51:435–444.

Djordjevic S, Driscoll PC. 2002. Structural insight into

substrate specificity and regulatory mechanisms of

phosphoinositide 3-kinases. Trends Biochem Sci 27:

426–432.

Fu SY, Sharma K, Luo Y, Raper JA, Frank E. 2000.

SEMA3A regulates developing sensory projections in the

chicken spinal cord. J Neurobiol 5:227–236.

Gallo G. 2006. RhoA-kinase coordinates F-actin organiza-

tion and myosin II activity during semaphorin-IIIA

induced axon retraction. J Cell Sci 119:3413–3423.

Gallo G, Letourneau PC. 2004. Regulation of growth cone

actin filaments by guidance cues. J Neurobiol 58:92–

102.

Haas MA, Vickers JC, Dickson TC. 2007. Rho kinase acti-

vates ezrin-radixin-moesin (ERM) proteins and mediates

their function in cortical neuron growth, morphology

and motility in vitro. J Neurosci Res 85:34–46.

Litman P, Amieva MR, Furthmayr H. 2000. Imaging of

dynamic changes of the actin cytoskeleton in microexten-

sions of live NIH3T3 cells with a GFP fusion of the F-

actin binding domain of moesin. BMC Cell Biol 1:1.

Loudon R, Silver L, Yee HF Jr., Gallo G. 2006. RhoA-

kinase and myosin II are required for the maintenance

of growth cone polarity and guidance by nerve growth

factor. J Neurobiol 66:847–867.

Mikule K, Sunpaweravong S, Gatlin JC, Pfenninger KH.

2003. Eicosanoid activation of protein kinase C epsilon:

Involvement in growth cone repellent signaling. J Biol

Chem 278:21168–21177.

Neufeld G, Lange T, Varshavsky A, Kessler O. 2007. Sem-

aphorin signaling in vascular and tumor biology. Adv

Exp Med Biol 600:118–131.

Orlova I, Silver L, Gallo G. 2007. Regulation of actomyosin

contractility by PI3K in sensory axons. Dev Neurobiol

67:1843–1851.

Paglini G, Kunda P, Quiroga S, Kosik K, Caceres A. 1998.

Suppression of radixin and moesin alters growth cone

morphology, motility, and process formation in primary

cultured neurons. J Cell Biol 143:443–455.

Potiron V, Nasarre P, Roche J, Healy C, Boumsell L. 2007.

Semaphorin signaling in the immune system. Adv Exp

Med Biol 600:132–144.

Schlatter MC, Buhusi M, Wright AG, Maness PF. 2007.

CHL1 promotes sema3A-induced growth cone collapse

and neurite elaboration through a motif required for

recruitment of ERM proteins to the plasma membrane.

J Neurochem (Online). doi:10.1111/j. 1471-4159.2007.

05013.x.

Speck O, Hughes SC, Noren NK, Kulikauskas RM, Fehon

RG. 2003. Moesin functions antagonistically to the Rho

pathway to maintain epithelial integrity. Nature 21:83–

87.

Toyofuku T, Kikutani H. 2007. Semaphorin signaling dur-

ing cardiac development. Adv Exp Med Biol 600:109–

117.

Williams EJ, Doherty P. 1999. Evidence for and against a

pivotal role of PI 3-kinase in a neuronal cell survival

pathway. Mol Cell Neurosci 13:272–280.



semaphorin 3a inhibits erm protein phosphorylation in growth cone filopodia through inactivation of...

Documents