Journ

alof

Cell

Scie

nce

Axonal transport of neural membrane protein 35 mRNAincreases axon growth

Tanuja T. Merianda1, Deepika Vuppalanchi2, Soonmoon Yoo3, Armin Blesch4,5 and Jeffery L. Twiss1,*1Department of Biology, Drexel University, Philadelphia, PA 19104, USA2Department of Pharmacology and Toxicology and Stark Neurosciences Research Institute (SNRI), Indiana University School of Medicine,Indianapolis, IN 46202, USA3Nemours Biomedical Research, Alfred I. du Pont Hospital for Children, Wilmington, DE 19803, USA4Spinal Cord Injury Center, University Hospital Heidelberg, 69118 Heidelberg, Germany5Department of Neurosciences, University of California, San Diego, La Jolla, CA 92093, USA

*Author for correspondence (jeff.twiss@drexel.edu).

Accepted 10 October 2012Journal of Cell Science 126, 90–102� 2013. Published by The Company of Biologists Ltddoi: 10.1242/jcs.107268

SummaryMany neuronal mRNAs are transported from cell bodies into axons and dendrites. Localized translation of the mRNAs brings autonomyto these processes that can be vast distances from the cell body. For axons, these translational responses have been linked to growth and

injury signaling, but there has been little information about local function of individual axonally synthesized proteins. In the presentstudy, we show that axonal injury increases levels of the mRNA encoding neural membrane protein 35 (NMP35) in axons, with acommensurate decrease in the cell body levels of NMP35 mRNA. The 39 untranslated region (39UTR) of NMP35 is responsible for thislocalization into axons. Previous studies have shown that NMP35 protein supports cell survival by inhibiting Fas-ligand-mediated

apoptosis; however, these investigations did not distinguish functions of the locally generated NMP35 protein. Using axonally targetedversus cell-body-restricted NMP35 constructs, we show that NMP35 supports axonal growth, and overexpression of an axonally targetedNMP35 mRNA is sufficient to increase axonal outgrowth.

Key words: RNA transport, Axon growth, Axonal protein synthesis, Regeneration

IntroductionPolarized eukaryotic cells transport mRNAs from the nucleus to

subcellular domains, providing a precise spatial and temporal

control of protein levels (Jung et al., 2012). For rodent neurons,

the ends of their processes can be centimeters distance from the

cell body. mRNAs are actively transported into these distal

processes by microtubule-based transport mechanisms (Donnelly

et al., 2010). Although mRNAs and translational machinery were

initially detected in the post-synaptic processes of the neuron (i.e.

dendrites), it is now clear that the pre-synaptic or axonal

processes also have the capacity to synthesize proteins (Jung

et al., 2012). This has been particularly evident in growing axons,

both for developing and regenerating axons, but there are several

publications supporting the notion that mRNAs and ribosomes

are present in mature axons of neurons in the peripheral nervous

system (PNS) prior to injury (Ben-Yaakov et al., 2012; Hanz

et al., 2003; Koenig et al., 2000; Perlson et al., 2005; Yudin et al.,

2008; Zelena, 1970). Studies of cutaneous nerve terminals also

suggest that translation of mRNAs can be triggered by pain

invoking stimuli (Jimenez-Diaz et al., 2008; Melemedjian et al.,

2010). In developing axons, translation of new proteins is needed

for response to some guidance cues (Jung et al., 2011). However,

a study from the Letourneau lab raised questions on this need for

localized protein synthesis in developing axons (Roche et al.,

2009). Thus, despite that mRNA translation has been

demonstrated in axons, questions remain on what the locally

generated proteins do in the axon.

Neurons have proven a particularly appealing model to study

mRNA localization. Even in cultured neurons, axons can extend

millimeters from the cell body making it possible to physically

isolate these processes to purity (Willis and Twiss, 2011). RNA

profiling studies of axonal processes in cultured neurons have

shown increasing complexity, such that hundreds of different

mRNAs are now known to be transported into axons (Gumy et al.,

2011; Taylor et al., 2009; Zivraj et al., 2010). It is not known if

such a complex population of mRNAs is similarly transported

into axons in vivo. We have recently used a transgenic approach

to show that b-actin mRNA is transported into axons of neurons

in the peripheral and central nervous systems (CNS) (Willis et al.,

2011). Moreover, localized translation of b-actin and GAP-43

mRNAs support regeneration of PNS axons (Donnelly et al.,

2011). It seems likely that protein products of other axonal

mRNAs also contribute to regeneration of axons. Consistent with

this, mice lacking the b-actin gene apparently showed normal

development and regeneration of motor axons (Cheever et al.,

2011). This could indicate that b-actin is not needed for

regeneration, or that other locally synthesized proteins can

compensate for the loss of b-actin. Thus, in depth analyses for

regulation and function of other axonal mRNAs are needed.

The axonal transcriptome of cultured dorsal root ganglion

(DRG) neuron includes several transmembrane and secreted

proteins (Gumy et al., 2011; Willis et al., 2007). Although the

classic ultrastructure of rough endoplasmic reticulum (RER) and

Golgi apparatus has not been identified in axons, we have

90 Research Article

Journ

alof

Cell

Scie

nce

previously shown that growing axons have functional equivalents

of RER and Golgi apparatus for secretion of locally synthesizedproteins (Merianda et al., 2009). However, evidence formembrane insertion or secretion of individual axonally

synthesized proteins has been lacking. Here, we have focusedon neural membrane protein 35 (NMP35), which we previouslyidentified by cDNA array hybridization as an axonal mRNA incultured adult DRG neurons (Gumy et al., 2011; Willis et al.,

2007). NMP35 is a transmembrane protein that was initiallycloned in attempts to identify developmentally regulated genes inperipheral nerve, but as its name implies its expression is limited

to neurons (Schweitzer et al., 2002; Schweitzer et al., 1998).NMP35 has also been termed as the Fas apoptosis inhibitorprotein 2 (FAIM2) and Lifeguard (LFG) (Beier et al., 2005;

Fernandez et al., 2007). The designations FAIM2 and LFGstemmed from studies showing that this protein can preventapoptosis triggered by Fas ligand (FasL) (Beier et al., 2005;

Fernandez et al., 2007). However, these studies did not take intoaccount the possibility that NMP35 protein may be synthesizedlocally. In the studies here, we have used RNA targetingconstructs to test functions of cell body versus axonally

synthesized NMP35 protein.

NMP359s sequence was also shown to have homology to theglutamate binding subunit of the rat and Drosophila NMDA

receptor (GRINA1; previously termed glutamate binding proteinand dNMDARA1, respectively) (Schweitzer et al., 2002).Previous immuno-electron microscopy studies showed NMP35protein to be present in post-synaptic processes, but the protein

was also detected in efferent processes of spinal motor neurons(Schweitzer et al., 2002). Here, we show that NMP35 mRNAlocalizes to axons as well as dendrites of cultured neurons. This

mRNA is targeted to PNS axons in vivo, and the 39 untranslatedregion (39UTR) of NMP35 mRNA is both necessary andsufficient for its transport into axons. Axonal localization of

NMP35 mRNA increases after axonal injury. Moreover,expression studies indicate that axonal transport of NMP35mRNA with subsequent localized translation provides a function

in axonal growth that is distinct from cell body restricted NMP35(FAIM2, LFG) previously used for overexpression studies (Beieret al., 2005; Fernandez et al., 2007).

ResultsDifferential localization of membrane protein mRNAs inthe axons

Although initial cloning of NMP35 mRNA focused on

identifying Schwann cell gene products (Schweitzer et al.,1998), NMP35 protein expression was shown to be restricted toneurons (Schweitzer et al., 2002). We detected NMP35 as anaxonal mRNA using cDNA arrays to profile transcripts from

cultures of injury-conditioned adult DRGs (Willis et al., 2007).The transport of NMP35 mRNA into axons of DRG neurons wasnot significantly affected by neurotrophins, semaphorin, or

myelin-associated glycoprotein, stimuli that we previouslyshowed trigger robust changes in transport of other axonalmRNAs (Willis et al., 2007). Thus, we asked if the injury

conditioning could affect the axonal levels of NMP35 mRNA.For this, we cultured DRG neurons on porous membranes toisolate axonal processes (Zheng et al., 2001). Reverse

transcriptase coupled polymerase chain reaction (RT-PCR)confirmed purity of the axonal preparations showingamplification of b-actin mRNA from the axonal preparations

but not the cell body restricted c-actin and MAP2 mRNAs

(Fig. 1A). Both standard and quantitative RT-PCR (RTqPCR)showed increased NMP35 mRNA levels in the axons from theinjury-conditioned compared to naıve DRGs (Fig. 1B,C).

Although the overall levels of NMP35 mRNA did notappreciably change in these DRG cultures (supplementarymaterial Fig. S1A), the cell body levels of NMP35 mRNA inthe injury-conditioned DRG cultures was significantly decreased

compared to naıve cultures (Fig. 1B,C). By fluorescence in situ

hybridization (FISH), NMP35 mRNA showed granular signals inthe axon shaft of the cultured DRG neurons that extended distally

into the growth cones (Fig. 1D). Scrambled probes and no probecontrols showed minimal background labeling (supplementarymaterial Fig. S1B). Comparing processes of injury-conditioned to

naıve DRG neurons, axonal NMP35 mRNA levels showed asignificant increase and cell body NMP35 mRNA showed asignificant decrease (Fig. 1D,F) consistent with the RT-PCR

results above. Furthermore, immunolabeling showed a significantincrease in axonal signals and decrease in the cell body signalsfor NMP35 protein with injury conditioning (Fig. 1E,G).

Since NMP35 mRNA increased in axons of injury-conditioned

neurons in vitro, we asked if a similar increase in axonal NMP35mRNA and protein might be seen in vivo after axonal injury.Confocal imaging of DRG and sciatic nerve sections that were

processed for RNA-FISH showed clear cell body and axonalsignals for NMP35 mRNA (Fig. 2A,C). NMP35 mRNA andprotein appeared overall more abundant in the injured nerves andless abundant in the DRG compared with signals in uninjured

tissues at seven days after nerve crush (Fig. 2B,D). Since NMP35mRNA was only seen in the cell bodies and axons by these FISHanalyses, we used RT-PCR to compare levels of NMP35 mRNA

in nerves at seven days after a unilateral sciatic nerve crush.Axons in the crushed sciatic nerves showed fourfold moreNMP35 mRNA than those in the uninjured nerves (Fig. 2E,F).

Moreover, NMP35 mRNA levels were threefold lower in L4-5DRGs ipsilateral to the injury compared to contralateraluninjured ganglia (Fig. 2E,F). Taken together, these data

indicate that axotomy by nerve crush injury triggers a shift inNMP35 mRNA localization, with increased levels of NMP35mRNA and protein in injured or regenerating versus uninjuredaxons.

NMP35 protein has been shown to concentrate in post-synapticterminals of mature, neurons (Schweitzer et al., 2002). The DRGneurons used here only extend Tau-positive, MAP2-negative

axonal processes in culture, and both the centrally and distallyprojecting processes of these sensory neurons show axonalfeatures in vivo (Zheng et al., 2001). Thus, we asked if NMP35mRNA also localizes into axonal processes of fully polarized

CNS neurons. By FISH analyses, NMP35 mRNA is seen in boththe dendritic and axonal processes of cultured cortical neurons(supplementary material Fig. S2), indicating that NMP35 mRNA

can localize into both pre-synaptic and post-synapticcompartments consistent with previous analyses of its proteinlocalization (Schweitzer et al., 2002).

NMP35 mRNA’s 39UTR drives axonal localization in DRGneurons

Given the potential shift in transport of NMP35 mRNA with

axonal injury, we asked how this mRNA is localized into axons.UTRs, particularly the 39UTR, are frequently sites forlocalization elements (Andreassi and Riccio, 2009). Thus, we

Axonal translation of NMP35 increases axon growth 91

Journ

alof

Cell

Scie

nce

Fig. 1. NMP35 mRNA and protein are enriched in axons of injury-conditioned neurons in culture. Dissociated DRG cultures prepared from naıve and injured

rats at 7 days following sciatic nerve crush were used for fractionation of cell bodies and axons (A–C) or for FISH and immunofluorescence (D–G) after 24 hours in

vitro. (A–C) By RT-PCR, b-actin was detected only in the axonal processes, but c-actin and MAP2 mRNAs were restricted to the cell body samples showing purity of

the axonal preparations (A). By both standard (B) and quantitative RT-PCR (C) NMP35 mRNA levels shift from cell body to axon predominance in the neurons from

the injured animals. Amplification of GAPDH shows equivalent levels of total RNA in the standard RT-PCR; 12S rRNA was used for normalization of the RTqPCR.

(D,E) By FISH (D) and immunofluorescence (E), both NMP35 mRNA and protein appear more abundant in the axons (arrows) and reduced in cell bodies

of DRG neurons cultured from injury-conditioned compared with naıve animals. The cell body images in D and E show XYZ maximum projections constructed from

10 optical planes taken at 0.5 mm intervals; the inset panels show only the NMP35 mRNA or protein signals. (F,G) Quantification of FISH (F) and

immunofluorescence (G) signals for axons and cell bodies from matched exposure images as in D and E are shown. All image pairs of naıve and injury-conditioned

neurons are matched for exposure, gain, offset and post-processing (n$15 for cell body and n$20 for axons from at least three separate experiments; **P#0.01,

***P#0.001 by Student’s t-test). Values in C, F and G are means6s.e.m. Scale bars: 10 mm for main panels, 50 mm for insets.

Journal of Cell Science 126 (1)92

Journ

alof

Cell

Scie

nce

Fig. 2. Axotomy increases axonal localization of NMP35 mRNA in vivo. NMP35 mRNA and protein levels were examined in naıve versus crush-injured sciatic

nerves at 7 days after injury. The nerve was sampled just proximal to the crush injury with a corresponding level of naıve nerve sampled; L4-5 DRGs are

illustrated. (A) Representative exposure matched confocal image pairs of naıve versus injured nerve for FISH/immunofluorescence show increased signal for

NMP35 mRNA in axons of the injured sciatic nerve (arrows) and a decrease in NMP35 mRNA signal in the neuronal cell bodies of L4-5 DRGs compared with

uninjured nerve and DRG. (B) Representative exposure-matched confocal image pairs as in A show a similar increased signal for NMP35 protein in the sciatic

nerve and decreased signal for NMP35 protein in the L4-5 DRG neuronal cell bodies at 7 days after injury compared with uninjured nerve and DRG.

(C) Quantification of NMP35 mRNA intensity that overlapped with neurofilament heavy (NF H) protein signal in naıve and injured sciatic nerve and DRG

sections is shown. (D) Quantification of NMP35 protein intensity overlapping with NF H signal in uninjured and injured sciatic nerve versus DRG is shown.

(E,F) RNA was purified from sciatic nerve and L4-5 DRGs, both naıve and at 7 days post crush injury, and used for RT-PCR to gain a more quantitative estimate

of changes in NMP35 mRNA levels. Standard RT-PCR amplification of b-actin and GAPDH mRNAs shows relative equivalent loading between samples; absence

of MAP2 mRNA in the sciatic nerve is consistent with the axonal nature of the neuronal processes in the sciatic nerve (E). NMP35 mRNA appears to increase in

the nerve and decrease in the DRG following crush injury. RTqPCR similarly shows an increase in NMP35 mRNA in sciatic nerve and a decrease in DRG (F). 12S

rRNA was used to normalize RTqPCR data (n$20 from at least three separate experiments; ***P#0.001 by Student’s t-test for indicated data pairs). Values in C,

D and F are means6s.e.m. Scale bars: 5 mm for DRG images; 10 mm for sciatic nerve images.

Axonal translation of NMP35 increases axon growth 93

Journ

alof

Cell

Scie

nce

asked if NMP35 mRNA’s 39UTR is sufficient for localizing a

heterologous mRNA into axons of cultured DRG neurons. For

this, we generated a diffusion-limited myristoylated green

fluorescent protein (GFPmyr) reporter (Aakalu et al., 2001)

carrying the 39UTR of NMP35 mRNA (GFPmyr39NMP35). The

myristoylation element in GFP coding sequence decreases its

diffusion compared to non-modified GFP such that fluorescence

recovery after photobleaching (FRAP) can be used to test for

localized translation events by comparing recovery with and

without active protein synthesis (Vuppalanchi et al., 2010; Yudin

et al., 2008). Without a localizing UTR, the GFPmyr mRNA is

restricted to the neuronal cell body (Aakalu et al., 2001; Akten

et al., 2011; Ben-Yaakov et al., 2012; Perry et al., 2012;

Vuppalanchi et al., 2010; Willis et al., 2007; Yudin et al., 2008).

DRG neurons transfected with GFPmyr39NMP35 showed robust

GFP fluorescence in foci along the axon shaft and in the terminal

axon/growth cone (Fig. 3A; supplementary material Movies 1 and

2), suggesting that the localization of the endogenous NMP35

mRNA that we had seen by FISH analyses can be driven by the

39UTR of NMP35 mRNA. Upon photobleaching, the distal axon

showed recovery of GFP fluorescence that began to reach

statistical significance at 12 min post-bleach (Fig. 3B). This

kinetics of recovery is consistent with our previous studies

where increase in fluorescence after photobleaching in axons at

$750 mm from cell body, as used here, occurs before new GFP

protein could be transported from the neuronal cell body (Akten

et al., 2011; Vuppalanchi et al., 2010; Willis et al., 2007; Yudin

et al., 2008). To test if protein synthesis contributed to this

recovery, DRG cultures were pre-treated with translation inhibitor

anisomycin for 30 min prior to photobleaching. There was no

increase in the axonal GFP fluorescence above post-bleach levels

in the presence of anisomycin (Fig. 3A,B), indicating that the

recovery seen above is protein synthesis dependent. Furthermore,

FISH showed that axonal GFP mRNA was only detected if the

39UTR of NMP35 was included in the reporter constructs (see

Fig. 6A below). These data indicate that the 39UTR of NMP35

mRNA is sufficient for mRNA localization into the axons of

cultured DRG neurons.

Since the 39UTR of NMP35 mRNA could localize mRNA into

axons of cultured neurons, we asked if it might also be sufficient

for axonal localization in vivo. For this, we constructed NMP35-

AcGFP fusion protein/reporters with the 39UTR of NMP35 or

39UTR of Aequorea coerulescens GFP (AcGFP); we refer to the

monomeric AcGFP as simply ‘GFP’ throughout the remainder of

this text. These constructs were used to generate lentivirus (LV)

for in vivo expression (LV-NMP35-GFP-39NMP35 and LV-

NMP35-GFP-39GFP, respectively). To test for axonal mRNA

translation in vivo, LV-NMP35-GFP-39NMP35 and LV-NMP35-

GFP-39GFP were used to transduce the L4-5 DRGs in adult rats.

Ten days after injecting LV preparations, animals were subjected

to unilateral sciatic nerve crush injury at mid-thigh level,

,4.5 cm from the site of LV injection. GFP signals were then

assessed at the site of injection, in the DRG, and in the distal axon

proximal to the crush site (or comparable level in the uninjured

nerve). Tissue sections were evaluated by immunolabeling for

GFP and neuronal markers using confocal microscopy. At the

injection site, GFP signals were seen in Schwann cells and in

axons for both LV preparations (supplementary material Fig. S3).

GFP signals were clearly visible in the naıve and injured DRG

for LV-NMP35-GFP-39GFP and LV-NMP35-GFP-39NMP35

transduced animals, indicating in vivo transduction of these

neurons by LV (Fig. 4A–H). In examining the distal nerve just

proximal to the crush site, GFP signals were only seen in the LV-

NMP35-GFP-39NMP35 transduced animals (Fig. 4G,H). Despite

the enhanced detection afforded by anti-GFP immunolabeling,

we saw no GFP signals in the nerves of the LV-NMP35-GFP-

39GFP transduced animals (Fig. 4C,D). The absence of any GFP

signals in the nerves of the LV-NMP35-GFP-39GFP transduced

animals and the distance separating the injection and injury sites

(.4 cm) argues that the axonal GFP signals seen when 39UTR of

NMP35 mRNA was included in the LV construct are the result of

localized protein synthesis rather than transport or diffusion of

NMP35-GFP protein from the cell body into the axonal

compartment. Taken together, these data indicate that the 39UTR

of NMP35 transcript is both necessary and sufficient for axonal

localization in sensory neurons both in vitro and in vivo.

Fig. 3. 39UTR of NMP35 mRNA is sufficient for axonal mRNA localization in vitro. DRGs were transfected with diffusion-limited GFPmyr plus the 39UTR of

NMP35 mRNA to test for axonal-localizing ability of this non-translated region. (A) Representative images from FRAP sequences are shown for indicated pre-bleach

and post-bleach intervals for control (top row) and 80 mM anisomycin-treated cultures (bottom row). GFP fluorescence is displayed as a spectrum as indicated; the

boxed area represents the region of interest (ROI) for the terminal axon (arrow indicates terminal axon) that was subjected to bleaching. Recovery of GFP signals in

the ROI only occurs when protein synthesis is intact. Scale bar: 25 mm. (B) Quantification of fluorescence in the ROI over the indicated time period is shown for n$7

neurons analyzed over at least three separate transfection experiments. Average fluorescence is shown as a percentage of pre-bleach signals within each image

sequence; error bars represent s.e.m. for each experiment (*P#0.05, **P#0.01 and ***P#0.001 by one-way ANOVA compared with t50 min).

Journal of Cell Science 126 (1)94

Journ

alof

Cell

Scie

nce

The FRAP experiments above did not allow us to distinguish

changes in localized synthesis of NMP35 in axons of injury-

conditioned versus naıve neurons, since the transfected constructs

required at least 2 days for optimal expression. Over this

incubation period, the naıve neurons begin to transition from an

arborizing growth to an elongating axonal growth is similar to the

injury-conditioned phenotype (Smith and Skene, 1997). These in

vivo experiments brought an opportunity to use GFP protein

fluorescence as a surrogate for comparing NMP35 39UTR-

dependent effects on axonal GFP mRNA levels in injured and

naıve neurons. Quantification of the axonal GFP signals confirmed

a significant increase in GFP intensity in the LV-NMP35-GFP-

39NMP35 transduced animals after injury compared to the

uninjured animals (Fig. 4G–H). There was no axonal GFP signal

detected above background in nerves from the LV-NMP35-GFP-

39GFP transduced animals (Fig. 4C–D). In contrast to the

Fig. 4. 39UTR of NMP35 mRNA is needed for localization into sensory sciatic nerve. To determine whether the 39UTR of NMP35 mRNA is needed for

axonal localization in vivo, L4-5 DRGs were transduced with LV that expressed an NMP35-GFP fusion protein carrying the 39UTR of NMP35 or GFP mRNAs

(NMP-GFP-39NMP35 and NMP-GFP-39GFP, respectively). (A–H) Representative exposure-matched confocal images for GFP signals (green) and neurofilament

heavy (NF H) (red) are shown for sections of L4-5 DRGs and sciatic nerves for naıve and 7 day crush-injured nerves and DRGs are shown as indicated. Large

panels show merged images and inset panels show only the GFP channel. The right-hand strips for C, D, G and H show YZ images of merged GFP and NF signals

for the first strip, and GFP signals only for the second strip. GFP signals are seen along the periphery of DRG cell bodies (arrows) for both NMP35-GFP-39GFP-

(A,B) and NMP35-GFP-39NMP35- (E,F) transduced animals. This suggests membrane localization of NMP35-GFP protein. Sections of distal sciatic nerve show

GFP signals only in the NMP35-GFP-39NMP35 transduced animals (G,H), with a clear increase in GFP signals in the injured nerve (arrows). The nerves in the

NMP35-GFP-39NMP35-transduced animals are also suggestive of focal NMP35-GFP signals along the periphery of the axon. This is also seen in YZ projections

of image stacks that show intra-axonal NMP35-GFP and signals enriched along the periphery of the axoplasm (image strips in G and H; see also teased nerve

preparations in supplementary material Fig. S4). (I–L), GFP signal intensity that overlapped with NF H signals was quantified over multiple animals and is shown

for DRG and sciatic nerves for the NMP35-GFP-39GFP- (I,J) and NMP35-GFP-39NMP35- (K,L) transduced animals. These data confirm the absence of sciatic

nerve NMP35-GFP signals in the NMP35-GFP-39GFP-transduced animals (J) and an increase in sciatic nerve NMP35-GFP signals with nerve crush injury in the

NMP35-GFP-39NMP35-transduced animals (L) (n$25 from at least three separate transduced experiments; **P#0.01 and ***P#0.001 by Student’s t-test).

Values in I–L are means6s.e.m. Scale bars: 10 mm.

Axonal translation of NMP35 increases axon growth 95

Journ

alof

Cell

Scie

nce

endogenous NMP35 mRNA, we saw no decrease in GFP levels in

the DRG after injury in the LV-NMP35-GFP-39NMP35transduced animals but there was an increase in DRG GFPlevels in the LV-NMP35-GFP-39GFP transduced animals (Fig. 4I–

K). This may be attributed to overexpression of the NMP35-GFPfusion protein compared to the endogenous NMP35 mRNA.

The NMP35-GFP fusion protein used in these latter studiesalso allowed us to assess potential for localization of NMP35

protein in axons. In the DRGs of the LV transduced animals,NMP35-GFP signals were concentrated along the periphery ofthe neurons for both LV preparations, suggestive of membrane

localization of the fusion protein (Fig. 4). Projection of confocalimage stacks of distal nerve into YZ planes also showedenhanced signals along the periphery of several axons in theLV-NMP35-GFP-39NMP35 transduced preparations (Fig. 4G,H,

inset panels). Finally, teased nerve preparations showed GFPsignals that were enriched along the periphery of individualaxons in the LV-NMP35-GFP-39NMP35 transduced animals

(supplementary material Fig. S4C,D). The LV-NMP35-GFP-39GFP transduced animals again showed no detectable GFPsignals in these teased nerve preparations (supplementary

material Fig. S4A,B). Taken together, these data support theconclusion that the locally translated NMP35, which is targetedto axons through its 39UTR, is inserted into the axoplasmic

membrane.

Although the above data are consistent with an increase intransport of NMP35 mRNA into axons after injury, a shift in thestability of the transcript would also account for the increased

levels of NMP35 mRNA and protein in the axons of injured DRGneurons. Since severed DRG axons rapidly undergo Walleriandegeneration even in culture (Zheng et al., 2001), we could not

test survival of NMP35 mRNA directly in the isolated axons.Thus, we used cyclosporin A (CsA) to delay the degeneration ofsevered axons. Barrientos et al. (Barrientos et al., 2011) showedthat inhibition of the mitochondrial membrane permeability

transition pore with CsA extends the survival of severed axons(Barrientos et al., 2011). Using RNA isolated from CsA-treatedaxons of naıve and injury conditioned DRG cultures as a template

for RT-PCR, the endogenous NMP35 and GAPDH mRNAs wereeasily detected up to 6 hours after severing (supplementarymaterial Fig. S5). In contrast to the axonal GAPDH mRNA that

did not change with injury conditioning, the overall levels ofNMP35 mRNA were more abundant in the axons of the injuryconditioned versus naıve neurons at every time point. However,

there was no difference in stability of the NMP35 mRNA whenthe injury conditioned and naıve RNA levels at 4 and 6 hoursamples were normalized to the 2 hour time point(supplementary material Fig. S5B). These data suggest that a

change in stability of NMP35 mRNA does not explain theincrease in axonal NMP35 mRNA levels after injury.

Axonally synthesized NMP35 mRNA increases axonoutgrowth

Our experiments suggest that transport of NMP35 mRNA intoaxons is regulated by injury and/or axonal growth, with the

axonally generated NMP35 protein localizing to the axoplasmicmembrane. This led us to ask if the locally synthesized NMP35protein contributes to growth of axons. For this, we used small

interfering RNAs (siRNAs) to deplete adult DRG cultures ofNMP35 mRNA overall. Transfection with a siRNA pool to theNMP35 sequence depleted both the transcript and protein by

,75% compared to cultures transfected with non-targetingsiControl (Fig. 5A,B). Non-targeting siRNAs (siControl) had no

effect on NMP35 mRNA levels compared with non-transfectedcultures (data not shown). Depletion of NMP35 mRNAsignificantly decreased axonal outgrowth in the DRG cultures(Fig. 5C,D). This suggests that NMP35 contributes to axonal

outgrowth, but did not distinguish effects of NMP35 synthesisoverall from axonally synthesized NMP35. To address thisquestion, we performed rescue experiments using siRNA-

resistant NMP35 constructs that were cell body restrictedversus axonally targeted. We separately tested each siRNA ofthe siRNA pool used above (data not shown). From this, the

target sequence of a single siRNA oligonucleotide that gavemaximal NMP35 mRNA depletion was used to generate siRNA-resistant NMP35-GFP constructs. By RT-PCR, siNMP35significantly depleted NMP35 mRNA when the neurons were

co-transfected with GFP, but co-transfection with siRNA-resistant NMP35-GFP constructs resulted in NMP35 mRNAlevels comparable to siControl transfected neurons (Fig. 5E).

Immunoblotting confirmed expression of NMP35-GFP fusionproteins and GFP at near equivalent levels (Fig. 5E). Depletionof NMP35 did not appear to affect cell survival since cell

numbers for these co-transfections were statistically the same(data not shown). Consistent with the results above, siNMP35transfection resulted in a significant depletion of axonal

outgrowth compared to siControl-transfected neurons, and co-transfection with GFP alone did not rescue this growth deficit(Fig. 5F). Co-transfection of NMP35-GFP, either the cell bodyrestricted or axonally localizing form, rescued the axonal growth

deficit with siNMP35; however, axons were significantly longerwhen the axonally localizing NMP35-GFP-39NMP35 was used(Fig. 5F).

These siRNA studies suggested that simply increasing levels ofNMP35 mRNA in axons might be sufficient to support increasedaxonal outgrowth. To directly test this possibility, we used the LV

preparations to overexpress cell body restricted versus axonallylocalizing NMP35-GFP mRNA in DRG cultures without alteringendogenous NMP35 levels. FISH for GFP mRNA confirmedaxonal localization of the NMP35-GFP-39NMP35 mRNA, but not

NMP35-GFP-39GFP, in the LV-transduced cultures (Fig. 6A,B).Cell bodies of the transduced neurons showed no significantdifferences in GFP mRNA with the different LV preparations

(Fig. 6A,C). Axons of neurons transduced with the LV-NMP35-GFP-39NMP35 were significantly longer than control neuronstransduced with LV-NMP35-GFP-39GFP (Fig. 6D,E). The

average axon lengths with expression of the cell body restrictedNMP35-GFP-39GFP were not significantly different than culturestransduced with LV-GFP (Fig. 6D,E). Thus, consistent with the

siRNA rescue experiments above, increasing axonal levels ofNMP35, but not just simply the overall expression of NMP35, issufficient for increased axonal outgrowth.

DiscussionSeveral studies in different neuronal preparations have nowshown that axons of cultured neurons contain hundreds of

mRNAs (Gumy et al., 2011; Taylor et al., 2009; Willis et al.,2007; Zivraj et al., 2010). We previously showed that transport ofmRNAs into axons cultured from the adult rats used here can be

regulated by stimulation with axonal growth-promoting andgrowth-inhibiting cues; however, transport of several mRNAsinto DRG axons did not respond to neurotrophins, semaphorin, or

Journal of Cell Science 126 (1)96

Journ

alof

Cell

Scie

nce

myelin-associated glycoprotein (Willis et al., 2007). This

suggested that either the appropriate stimulus was not tested or

these mRNAs are constitutively transported into axons. Some of

these transcripts, such as Importin b1, RanBP1 and Stat3amRNAs, are resident in mature PNS axons with their local

translation in axons being triggered by injury (Ben-Yaakov et al.,

2012; Hanz et al., 2003; Yudin et al., 2008). Here, we show that

NMP35 mRNA is also regulated by injury. However, our data

suggest that transport of NMP35 mRNA into axons rather than its

translation within axons is injury-dependent. Axonal levels of

NMP35 mRNA are increased after axonal injury with no change

in its overall expression of the transcript. Moreover, the locally

generated NMP35 protein enhances axonal growth. Thus, this

shift in NMP35 mRNA levels from cell body predominant in

naıve neurons to axon predominant in axotomized neurons

suggests that injury triggers an inherent change in the neuron’s

ability to localize this mRNA.

Transport of mRNAs from the cell body into axons is an active

process, driven by specific RNA elements that serve as binding

sites for RNA binding proteins (RBP). These structural elements

have proven difficult to predict based on primary sequence alone,

but have most frequently been documented within the 39UTRs of

localized mRNAs (Andreassi and Riccio, 2009). Consistent with

this, NMP35 mRNA’s 39UTR shows no clear homology to

known localization elements. The localizing 39UTRs of

calreticulin and grp78/BiP mRNAs show striking interspecies

sequence identity among vertebrates (Ben-Yaakov et al., 2012;

Kislauskis et al., 1994; Vuppalanchi et al., 2010). In contrast, the

39UTR of rat NMP35 mRNA shows 83% identity with mouse,

less than 75% identity with primate, and essentially no identity

with other available vertebrate NMP35 39UTR sequences.

Nonetheless, the rat NMP35 mRNA’s 39UTR was sufficient for

localizing GFP reporter mRNA in cultured neurons and was

necessary for localizing an NMP35-GFP fusion protein mRNA

Fig. 5. Depletion of NMP35 mRNAs in the

DRG neurons attenuates axonal outgrowth.

(A) DRG cultures transfected with siRNA

targeting NMP35 (siNMP35) mRNA or a non-

targeting control (siControl) show reduced

NMP35 mRNA and protein at 72 hours post-

transfection by RT-PCR and immunoblotting,

respectively. (B) RTqPCR: siNMP35 reduced

endogenous NMP35 mRNA levels by ,80%.

(C,D) Depletion of NMP35 mRNA from DRG

cultures decreases axonal growth.

Representative neurofilament heavy (NF H)-

stained images are shown in C. Quantification

of average axon length for neurons transfected

with siNMP35 versus siControl after 72 hours

in vitro is shown in D. Scale bar: 100 mm.

(E,F) To test for potential off-target effects of

the siNMP35, neurons were co-transfected with

cell-body-restricted (NMP35-GFP-39GFP)

versus axonally targeted (NMP35-GFP-

39NMP35) constructs; transfection with GFP

was used as a control. (E) Representative RT-

PCR and immunoblotting for NMP35 mRNA

and NMP35-GFP mRNA and proteins. Upon

co-transfection with siNMP35 and NMP35-GFP

constructs, NMP35 mRNA levels were rescued

to near-endogenous levels. GFP mRNA levels

were relatively equivalent for the two NMP35-

GFP and the GFP transfections.

Immunoblotting confirmed expression of

NMP35-GFP protein. (F) Analysis of axonal

growth in these co-transfection experiments

showed the expected reduction in axon length

comparing siNMP35+GFP and siControl

transfections. Both of the NMP35-GFP

constructs rescued axonal outgrowth deficit

seen with siNMP35, with the axonally targeting

NMP35-GFP-39NMP35 increasing axonal

growth above the siControl-transfected cultures.

siNMP35+GFP showed significant reduction in

axonal outgrowth compared with siControl-,

NMP35-GFP-39NMP35- and NMP35-GFP-

39GFP-transfected neurons (n$25 from at least

three separate transfection experiments;

***P#0.001 by Student’s t-test for B, D and F).

Values in B, D and F are means6s.e.m.

Axonal translation of NMP35 increases axon growth 97

Journ

alof

Cell

Scie

nce

into adult peripheral axons in vivo. Similar to NMP35 mRNA,

CGRP mRNA shifts from cell body to axons in injured DRG

neurons (Toth et al., 2009) and sensorin mRNA shifts from cell

body to neurites in Aplysia sensory neurons upon synapse

formation (Andreassi et al., 2010; Lyles et al., 2006; Natera-

Naranjo et al., 2010). The overall levels of NMP35 mRNA do not

change with injury, only the ratio of axonal to cell body mRNA

increases comparing compared injured and naive DRG neurons.

With CGRP showing similar regulation (Toth et al., 2009), it is

appealing to hypothesize the existence of a cohort of axonal

mRNAs that share RBPs for enhanced targeting after axotomy.

Future studies will be needed to determine the RBP(s) that bind

to NMP35’s 39UTR for localizing this transcript, but this shift in

axonal levels of NMP35 mRNA without an apparent change in

overall expression of the transcript likely reflect either an

increase in availability or activity of this RBP after axotomy.

Such a shift in availability could be secondary to decreased levels

of other mRNAs as we have recently shown for GAP-43 and b-

actin mRNAs competing for limited quantities of ZBP1 (Donnelly

et al., 2011). However, in contrast to ZBP1, overexpression of the

axonally localizing NMP35-GFP-39NMP35 resulted in increased

axonal growth suggesting that the level of the RBP(s) needed for

axonal localization of NMP35 mRNA is not limiting.

Functions of locally synthesized proteins have largely been

extrapolated from what is known of the overall functions of these

proteins, but the locally generated proteins have been found to

have distinct functions in some circumstances. For example, the

microfilament protein b-actin allows for directional migration

rather than overall motility of migrating fibroblasts when it is

locally generated (Shestakova et al., 2001). Axonal translation of

Importin b1 and RanBP1 is used to generate an Importin a/bheterodimer for transporting axonal signaling proteins to the cell

body, with introduction of these proteins providing a precise

temporal indicator of axon injury (Ben-Yaakov et al., 2012; Hanz

et al., 2003; Perry et al., 2012; Yudin et al., 2008). Although

NMP35 mRNA was initially cloned from developing sciatic

Fig. 6. Axonal targeting of NMP35 mRNA is sufficient to increase axonal outgrowth. To determine whether targeting NMP35-GFP fusion protein for axonal

synthesis might contribute to growth, we transduced cultured DRG neurons with the LV preparations used in Fig. 4 to overexpress axonally targeted versus cell-

body-restricted NMP35 mRNA (NMP35-GFP-39NMP35 and NMP35-GFP-39GFP, respectively). (A–C) NMP35-GFP-39NMP35- and NMP35-GFP-39GFP-

transduced cultures were processed for FISH for GFP mRNA. Representative exposure-matched images of cell bodies and axon shafts are shown in A. Here, GFP

mRNA is only seen in axons of NMP35-GFP-39NMP35-transduced cultures. Quantification of axonal and neuronal cell body GFP mRNA signals for multiple

cultures is shown in B and C, respectively. There is no axonal GFP mRNA signal above background in the NMP35-GFP-39GFP-transduced cultures, whereas cell

body signals are not significantly different between the NMP35-GFP-39GFP- and NMP35-GFP-39NMP35-transduced cultures (**P#0.01 by Student’s t-test).

(D) Representative montage images of neurons expressing GFP, NMP35-GFP-39NMP35 and NMP35-GFP-39GFP for 72 hours and then stained for neurofilament

heavy (NF H) are shown. (E) Quantification of axon length of neurons transduced with GFP, NMP35-GFP-39NMP35 or NMP35-GFP-39GFP is shown (n$25

neurons in at least three separate transduction experiments; **P#0.01; ***P#0.001 by Student’s t-test). Scale bars: 10 mm (A); 200 mm (D).

Journal of Cell Science 126 (1)98

Journ

alof

Cell

Scie

nce

nerve (Schweitzer et al., 1998) and NMP35 protein concentratesat synapses in post-mitotic neurons (Schweitzer et al., 2002), the

possibility for localization of NMP35 mRNA was not consideredin these early expression studies. Our data linking increasedaxonal levels of NMP35 mRNA to increased axonal growthlikely accounts for the initial cloning of NMP35 mRNA from

sciatic nerve.

NMP35 was independently isolated by other groups as FAIM2or LFG and has been functionally implicated in cell survival

(Beier et al., 2005; Fernandez et al., 2007). Depletion of FAIM2/LFG from neurons increases sensitivity to Fas-mediatedapoptosis and developmental cell death, while overexpression

of FAIM2/LFG was protective from effects of FasL-mediateddeath of neurons (Beier et al., 2005; Fernandez et al., 2007;Hurtado de Mendoza et al., 2011). Although no role in neurite oraxonal growth was detected for the overexpressed FAIM2/LFG

proteins, these constructs did not contain the 39UTR that we haveshown localizes NMP35 mRNA into neuronal processes based onprimers and ESTs used for their cloning. Consequently, the

overexpression studies used by Beier et al. (Beier et al., 2005)and Fernandez et al. (Fernandez et al., 2007) did not examine theeffects of the locally synthesized protein that we show here plays

a role in axonal growth. Thus, we have uncovered a previouslyunrecognized function of NMP35 in promoting axonal growththat is distinct for the locally synthesized NMP35.

A similarly named but unrelated mRNA, FAIM (GenBank

accession number NM_080895), encodes a protein that alsoprotects neurons and other cell types from FasL-inducedapoptosis (Segura et al., 2007; Sole et al., 2004). Interestingly,

an alternatively spliced form of FAIM mRNA generates a shorterprotein that lacks 22 N-terminal amino acids (FAIMs) (Zhonget al., 2001). Similar to the axonally localizing NMP35 studied

here, overexpression of FAIMs increases neurite outgrowth (Soleet al., 2004). However, it should be emphasized that NMP35 andFAIMs show no sequence homology at the RNA or protein level

and the alternatively spliced FAIMs alters the 59 end of thetranscript. Furthermore, unlike NMP35/FAIM2, both long andshort forms of FAIM are cytoplasmic proteins without anypredicted transmembrane domains (Segura et al., 2007; Sole et al.,

2004).

In our hands, we do not see increased cell death with depletionof NMP35 from the DRG cultures under standard culture

conditions; thus, it is not clear if the role in axonal growth islinked to NMP35’s inhibition of Fas signaling. There have beenstudies showing that Fas activation can support neurite growth.

FasL was shown to increase neurite branching in hippocampalcultures, neurite outgrowth from DRG explants, and axonalregeneration after sciatic nerve crush (Desbarats et al., 2003;Zuliani et al., 2006). These studies could argue that NMP35 has

some function beyond Fas inhibition, since NMP35 is known toinhibit Fas signaling and our data indicate that NMP35 supportsrather than attenuates axon growth. However, Fas activation

through FasL was recently shown to increase secretion of NGF inSchwann cells (Mimouni-Rongy et al., 2011). Thus, the data fromDesbarats et al. (Desbarats et al., 2003) on FasL supporting nerve

regeneration may be a secondary effect of FasL on the Schwanncells. Further studies will be needed to determine molecularmechanisms that NMP35 utilizes to support growth of axons.

Nonetheless, our work clearly shows axonal targeting of anmRNA encoding a transmembrane protein can be used tomodulate axonal growth from adult neurons. Moreover, these

studies point to a previously unrecognized mechanism that

neurons can use to modify axonal transport of mRNAs beyondthe extracellular stimuli-induced transport and competition forRBPs that we have previously published (Donnelly et al., 2011;

Vuppalanchi et al., 2010; Willis and Twiss, 2010). For NMP35mRNA transport, the axonal injury must either increase the levelsor activity of the RBP(s) needed for targeting the transcript for

transport from the cell body into distal axons.

Materials and MethodsAnimal care and surgery

All animal experiments were conducted under Institutional Animal Care and UseCommittee (IACUC)-approved protocols at Alfred I duPont Hospital for Childrenand Drexel University. For nerve injury, animals were subjected to a conditioningsciatic nerve crush at mid-thigh level as previously described (Twiss et al., 2000).

Cell culture

For DRG culture, L4-6 ganglia were isolated from adult male Sprague Dawley ratsand dissociated using collagenase (500 U/ml; Sigma, St. Louis, MO) and trypsin-EDTA (0.05%; Cellgro, Manassas, VA) (Twiss et al., 2000). Cells were washedwith DMEM/F12 and then resuspended in DMEM/F12 medium containing N1supplement (Sigma), 10% horse serum (Hyclone, Salt Lake City), and 10 mMcytosine arabinoside (Sigma). Cells were cultured on poly-L-lysine(Sigma)+laminin (Millipore, Billerica, MA)-coated surfaces at 37 C with 5%CO2. For isolation of axons, dissociated cells were plated at moderate density onpolyethylene-tetrathalate (PET) membrane (8 mm pores; BD Falcon) inserts(Zheng et al., 2001); for FISH analyses, immunofluorescencea and neuriteoutgrowth assays, dissociated cells were plated at low density on coated glasscoverslips.

DRG cultures were transfected immediately after dissociation using theAMAXA Nucleofector apparatus (Lonza, Allendale, NJ). For this, dissociatedganglia were pelleted at 100 g for 5 min and resuspended in transfection solutionfrom the Rat Neuron Nucleofector kit (Lonza). A 3–5 mg portion of each plasmidDNA was transfected using the G-013 program. Cells were resuspended in culturemedia and plated; medium was replaced 4 and 24 hours later. For in vivotransduction with LV, an MOI of 100 was added to the dissociated DRGs 3–4 hours after plating cells; media was replaced the next day; LV was titered asdescribed previously (Vuppalanchi et al., 2010).

For culturing cortical neurons, cortices were dissected from P1-4 rat pups,incubated in papain for 30 min, and then dissociated by trituration in Hibernate Emedia (Brainbits, Springfield, IL). After gentle centrifugation, dissociated corticeswere plated on poly-L-lysine-coated glass coverslips in Neurobasal medium with10% fetal bovine serum, B27 supplement, 2 mM glutamine (Invitrogen, GrandIsland, NY), and penicillin/streptomycin (Cellgro) at 37 C and 5% CO2

(Vuppalanchi et al., 2010).

Isolation of axons

DRG neurons cultured on PET membranes for 16–20 hours were used for axonisolation. For this, the upper membrane surface containing cell bodies and non-neuronal cells was scraped using a cotton-tipped applicator as previously described(Willis and Twiss, 2010). For analysis of cell body, the axonal processes werescraped from the upper membrane surface. RNA was then isolated from thecellular elements remaining on the membrane as outlined below. These isolatedaxonal preparations were tested for purity by RT-PCR for c-actin, b-actin andMAP2 mRNAs (Merianda et al., 2009; Willis et al., 2005).

DNA and viral constructs

All PCR products used for cloning were sequence verified as individual clones.RNA from adult rat DRGs was used as template for reverse transcription usingiScript (BioRad, Hercules, CA); Pfu DNA polymerase (Stratagene, La Jolla, CA)was used for PCR amplifications.

Diffusion limited GFP reporter constructs (GFPmyr) (Aakalu et al., 2001) wereused to test for axonal localizing ability of NMP35 RNA sequences. 39UTR ofNMP35 (GenBank accession NM_144756) with terminal Not1 and Apa1restriction sites was PCR-amplified with the following primers: sense, 59-GCGG-CCGCAACCGGGAATGAGGAGCCCTCC-39 and anti-sense, 59-GGGCCCGAC-TGAGGGGACATGGCTGGAACTTG-39. PCR products were cloned into pTOPO2.1 (Invitrogen), and then sub-cloned into Not1 and Apa1 sites of the GFPmyr

expression vector replacing the 39UTR of aCAMKII (Aakalu et al., 2001).

For expression of NMP35 coding sequence, constructs were generated to encodean NMP35-GFP fusion protein using pAcGFP1-N3 (Clontech, Mountain View,CA). For this the complete 59UTR through coding sequence (i.e. next to lastcodon) was using the following primers, constructed with EcoR1 and BamH1restriction sites for cloning: sense, 59-GAATTCGAGACGCAGGCAGGCTGCG-GTGA-39 and anti-sense, 59-GGATCCAGCTTTTTGGCACCAACCGGGAA-39.

Axonal translation of NMP35 increases axon growth 99

Journ

alof

Cell

Scie

nce

This generated a plasmid encoding an mRNA with NMP35-GFP fusion protein and39UTR of GFP (pNMP35-GFP-39GFP). The GFP 39UTR was then replaced withNMP35’s 39UTR to generate pNMP35-GFP-39NMP35. NMP359s 39UTR wasgenerated as above, but with terminal Not1 sites for cloning downstream of GFP.

For generating LV, we first used a QuikChange XL Site-Directed Mutagenesiskit to mutate the polyadenylation signals in the NMP35 and GFP 39UTRs(Vuppalanchi et al., 2010). After sequence verification, the cDNA cassettescontaining the CMV promoter, 59UTR, NMP35-GFP fusion protein codingsequence were subcloned into the pENTR shuttle vector (Invitrogen). NMP35-GFPsequence was then recombined into a Gateway-compatible derivative of thepCDH-CMV–MCS1-Ef1a-copGFP (SBI System Biosci., Mountain View, CA) asdescribed previously (Vuppalanchi et al., 2010). LV was generated as describedpreviously (Blesch, 2004). Serial dilutions were used to determine MOI of theseLV preparations in DRG cultures.

To generate siRNA resistant NMP35-GFP constructs, four nucleotides in thesiRNA-targeted regions (CCGTATTCTTTGCAACATA) were mutated as aboveusing the following primers: sense, 59-GGGCATCCTATGCCGTG*TTCTTC*G-CAACG*TACCTGACTCTGGCTT-39 and anti-sense, 59-AAGCCAGAGTCAG-GTAC*GTTGCG*AAGAAC*ACGGCATAGGATGCCC-39 (* indicates mutatednucleotides).

RNA isolation and analyses

Total RNA was extracted from the fractionated DRG cultures using theRNAqueous Micro kit (Ambion, Austin, TX) as per the manufacturer’s protocol.For tissue preparations, RNA was isolated using the RNAqueous kit (Ambion) perthe manufacturer’s protocol. Isolated RNA was quantified using the VersaFluorfluorimeter (BioRad) with a RiboGreen assay (Invitrogen) (Willis et al., 2005).Tissue samples were normalized for RNA content. For the axonal preparations,flow-through from the affinity-based RNA isolation was used to measure theprotein content of the isolates by fluorimetry using the NanoOrange reagent(Invitrogen). Axonal RNA preparations were then normalized for protein contentto control for axon yield (Merianda et al., 2009; Willis et al., 2005; Willis et al.,2007). Normalized RNA samples (60 ng each) were used for reverse transcription(RT) using the iScript RT kit (BioRad).

Standard PCR was performed using HotStarTaq Mastermix (Qiagen, Valencia,CA). RNA isolated from adult rat brain was used as a positive control for RT-PCR.Negative control consisted of PCR using RT reactions processed without theaddition of enzyme (‘no RT’). Cycling parameters for standard PCR were: 15 min‘hotstart’ at 95 C and then 45 sec at 95 C, 45 sec at 58 C, and 3 min at 72 C for 30cycles.

Reverse transcribed pure axonal RNA samples were further processed forquantitative PCR (qPCR) using 26 SsoFast Evagreen Supermix on a CFX 384Touch qPCR instrument (Biorad). For the qPCR on axonal samples, we used signalsfor the 12S mitochondrial rRNA to normalize RNA content and RT efficiencybetween samples by differential cycle threshold (DCt) calculations (Willis and Twiss,2010). Primers for rat c-actin, b-actin, and MAP2 mRNAs and 12S mitochondrialrRNA have been published (Willis et al., 2005). Other primers used were as follows:glyceraldehyde-3-phosphate dehydrogenase (GAPDH; GenBank accession numberNM_017008) – sense, 59-GGAGAAACCTGCCAAGTATG-39 and antisense, 59-AGACAACCTGGTCCTCAGTG 39; NMP35 – sense, 59-ACCTGACTC-TGGCTTGCTGT-39 and antisense, 59-CGAGGAGGAGTCCACTGAAG-39.

Axonal mRNA decay analyses

Naıve and injury-conditioned DRGs were cultured on PET membranes for16 hours and then treated with 20 mM Cyclosporin A (CsA; LC Laboratories).After 20 mins, cell bodies and non-neuronal cells were removed as describedabove; severed axons were incubated at 37 C, 5% CO2 for up to 6 hr. Axonal RNAwas then isolated and processed for RT-PCR as described above. The qPCRanalyses were performed in quadruplicate over at least 3 separate experiments.

In situ hybridization and immunofluorescence analyses

Digoxigenin-labeled oligonucleotide or cRNA probes were used for FISH.Oligonucleotide probes were used to detect endogenous mRNAs as previouslydescribed (Vuppalanchi et al., 2010). Antisense oligonucleotides to nucleotides842–891 and 980–1029 of rat NMP35 (GenBank accession NM_144756) weredesigned using Oligo6 software (Molecular Biology Insights, Cascade, CO).BLAST analyses showed no homology to rat mRNAs deposited into GenBank.These were synthesized with 59-amino C6 modifier C6 at four thymidines peroligonucleotide, and then labeled with digoxigenin succinamide ester (Roche,Indianapolis, IN) per manufacturer’s protocol. Digoxigenin-labeled, scrambledprobes were used for specificity control. cRNA probes were used to detect GFPreporter mRNA. These were generated by in vitro transcription from linearizedpcDNA3-eGFP (Addgene, Plasmid 13031) with SP6 or T7 RNA polymerases anddigoxigenin-labeled nucleotide mixture (Roche). Oligonucleotide probes wereused at 1.66 ng/ml for cultured neurons and 50 ng/ml for tissue sections(Vuppalanchi et al., 2010). cRNA probes were used at 5 ng/ml probes. Afterhybridization and washing, samples were processed for immunofluorescence aspreviously described (Vuppalanchi et al., 2010). The following primary antibodies

were used: chicken anti-neurofilament heavy (NF H) (1:1000; Millipore), mouseanti-digoxigenin (1:200; Jackson ImmunoResearch, West Grove, PA), and Cy3-

conjugated mouse anti-digoxigenin (1:200; Jackson ImmunoResearch). Aftervigorous washing, secondary antibodies were applied for 2 hours. The following

secondary antibodies were used: FITC-conjugated donkey anti-chicken (1:200;Jackson ImmunoResearch) or AMCA-conjugated anti-chicken (1:200; Jackson

ImmunoResearch) and Cy3-conjugated anti-mouse (1:200; JacksonImmunoResearch). After series of washes, samples were mounted with Prolong

Gold Antifade (Invitrogen).

Immunofluorescence was performed as previously published (Merianda et al.,

2009), with the exception of either 4% paraformaldehyde or cold methanol fixedsamples were used for cultures (methanol fixation was used to visualize NMP35

protein in cultured neurons). Tissue sections were equilibrated in PBS and thenincubated in 20 mM Glycine followed by 0.25 M NaBH4 for 30 min to quench

autofluorescence (Vuppalanchi et al., 2010). The antibody to rat NMP35 wasgenerated in chicken using a synthetic NMP35 peptide (SYEEATSGEGLKAGAF;

Swissport NO. AAC32463) by GeneTel Labs (Madison, WI); this antibody wasused at 1:200 dilutions. Mouse anti-neurofilament antibody cocktail (1:400;Sigma) was used to detect axons; GFP antibody (1:200, Abcam) to enhance GFP

signals. Secondary antibodies were FITC or Texas red conjugated donkey anti-chicken, -rabbit or -mouse antibodies (1:300; Jackson ImmunoResearch). After

vigorous washes, samples were mounted as above. Cultured neurons were imagedwith Leica DMRXA2 or Zeiss Axioplan epifluorescent microscopes fitted with an

ORCA-ER charge-coupled device (CCD) camera (Hamamatsu, Bridgewater, NJ).All images were matched for acquisition parameters and post-processing. Tissue

sections were imaged using Leica TCS/SP2 or Zeiss LSM700 laser scanningconfocal microscope.

Protein isolation, electrophoresis and immunoblotting

DRG cultures were lysed for 20 min at 4 C in RIPA buffer supplemented with

protease inhibitor cocktail (Sigma). Lysates were cleared of debris bycentrifugation at 16,000 g for 15 min at 4 C and then normalized for protein

content using Bradford assay (BioRad). Normalized lysates were precipitatedovernight with acetone at 280 C, and then resuspended and denatured in Laemmli

sample buffer. Samples were resolved by standard SDS-PAGE and thenelectrophoretically transferred to PVDF membranes (Millipore). Membranes

were rinsed in Tris-buffered saline with 0.1% Tween 20 (TBST) and thenblocked in 5% nonfat dry milk. Blots were incubated in the following antibodies

overnight at 4 C: rabbit anti-LFG (1:500; Pro Sci, Poway, CA) or rabbit anti-GFP(1:2000; Abcam, Cambridge, MA). Blots were rinsed several times in TBST and

then incubated with HRP-conjugated anti-rabbit IgG (1:5000; Millipore) in theblocking buffer for 1 h at room temperature. Blots were washed for 30 min in

TBST and developed with ECLplus (GE Healthcare, Piscataway, NJ).

Fluorescence recovery after photobleaching

FRAP was used to test for axonal translation of GFPmyr reporter mRNAs withNMP35 39UTR as described (Vuppalanchi et al., 2010). For this, DRG cultures

that had been transfected with GFPmyr39NMP35 were initially evaluated for GFPexpression by epifluorescence microscopy. 406oil immersion objective (0.7 NA)

on an inverted Leica TCS/SP2 confocal microscope with an environmentalchamber maintained at 37 C was used with pinhole was set at 4 AU. For baseline

fluorescence intensity, neurons were imaged every 30 sec over 2 min with 488 nmlaser line set at 15% power before bleaching. For photobleaching, $150 mm2 ROI

comprising the terminal 70–120 mm of distal axon was exposed to 100% power ofthe 488 nm laser line for 40 frames at 1.6 sec intervals. Recovery was then

monitored every 30 sec over 30 min using 15% power of 488 nm laser line andGFP emission was collected over band filter of 498–530 nm with PMT energy,

offset, and gain matched for all collection sets. To assess role of protein synthesisin this recovery, cultures were pre-treated with 80 mM anisomycin for 30 min priorto photobleaching. ImageJ was used to calculate average pixels/mm2 in the ROIs of

the raw confocal images, which was then normalized to baseline intensity tocalculate percent recovery. Mean for this normalized intensity for each post-bleach

interval was calculated.

siRNA-based depletion

A pool of four synthetic siRNAs targeting rat NMP35 mRNA was designed using

Dharmacon’s siDESIGN Center (www.dharmacon.com/sigenome/default.aspx).For siRNA transfection, dissociated DRGs were cultured and then transfected with

200 nM siRNA after 24 hours using DharmaFECT3 transfection reagent in serumfree medium per manufacturer’s protocol (Dharmacon, Chicago, IL). Transfection

efficiency for siRNAs was monitored by co-transfection with a stable, fluorescent,non-targeting control siRNA with RISC-free modification (siGLO, Dharmacon).To test for non-specific effects of siRNA transfection, cultures were transfected

with non-targeting siCONTROL (Dharmacon). Efficiency of siRNA-baseddepletion was quantitated by RTqPCR for NMP35 mRNAs at 72 hours after

transfection as outlined above.

Journal of Cell Science 126 (1)100

Journ

alof

Cell

Scie

nce

In vivo expression of NMP35-GFPFor testing localization of NMP35-GFP in vivo, L4-5 DRGs were transduced withLV preparations as outlined above. 40 MOI (10 ml) of each LV preparation wasinjected into the L4-5 nerve roots adjacent to the DRG. 10 days later, animals weresubjected to sciatic nerve crush injury at mid-thigh as previously described (Twisset al., 2000). The L4-5 DRGs, nerve roots (injection site) and sciatic nerve (0.5 cmproximal to 0.5 cm distal to the crush site) were dissected fixed in 4%paraformaldehyde and processed for cryosectioning. Cryostat sections wereimmunolabeled as described above with anti-neurofilament (1:750; Millipore)and anti-GFP (1:200; Abcam) antibodies. Teased nerve preparations weregenerated from a segment of fixed sciatic nerve that was equilibrated in PBS. A30-ga needle was used to strip away the epineurium and teased apart the nerve intoindividual fibers. Teased nerves were air dried on Superfrostplus glass slides(Fisher, Pittsburgh, PA), hydrated in PBS, and then processed for immunostaining.Teased nerves were mounted under coverslips with Prolong Gold Antifade andanalyzed by confocal microscopy. Mean for GFP protein intensity (pixels/mm2) inDRG, distal and proximal nerves (both naıve and injured) was calculated usingImage J.

Axon outgrowth assaysAxonal length was assessed after 72 hours in culture by immunostaining withneuronal markers as above. Length of longest axon of randomly captured imageswas measured by a blinded observer using Image J as described (Donnelly et al.,2011). At least three separate culture preparations were analyzed for eachcondition.

Fluorescence intensity studiesIntensity was calculated from digital images that were matched for exposure time,gain, offset, and post-processing parameters. mRNA and protein signals in cellbody and axons was analyzed for intensity for n$15 for cell body and n$20 foraxons from at least 3 separate experiments. Neurofilament signals were used totrace neuronal cell body and axons. Image J was then used to calculate the averagepixels/mm2 in these areas as recently described (Vuppalanchi et al., 2012).

Statistical analysesGraphPad Prism 4 software package (La Jolla, CA) was used for statisticalanalyses. Student’s t-test was used to compare two means of independent groups.These included fluorescent intensities from FISH and immunofluorescence imagescalculated using Image J. One-way ANOVA with Bonferroni post-hoc test wasused to test for significance the FRAP studies where more independent groupswere compared as previously published (Akten et al., 2011; Ben-Yaakov et al.,2012; Perry et al., 2012; Vuppalanchi et al., 2010; Vuppalanchi et al., 2012; Yudinet al., 2008).

FundingThe initial studies in this work were funded by the ParalyzedVeterans of America Spinal Cord Research Foundation [grantnumber 2520 to T.T.M.]. Additional grant funding came from theNational Institutes of Health [grant number R01-NS041596 toJ.L.T.]; the Miriam and Sheldon G. Adelson Medical ResearchFoundation and the International Foundation for Research inParaplegia [grant number P116 to J.L.T.]. Deposited in PMC forrelease after 12 months.

Supplementary material available online at

http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.107268/-/DC1

ReferencesAakalu, G., Smith, W. B., Nguyen, N., Jiang, C. and Schuman, E. M. (2001).

Dynamic visualization of local protein synthesis in hippocampal neurons. Neuron 30,489-502.

Akten, B., Kye, M. J., Hao, T., Wertz, M. H., Singh, S., Nie, D., Huang, J.,

Merianda, T. T., Twiss, J. L., Beattie, C. E. et al. (2011). Interaction of survival ofmotor neuron (SMN) and HuD proteins with mRNA cpg15 rescues motor neuronaxonal deficits. Proc. Natl. Acad. Sci. USA 108, 10337-10342.

Andreassi, C. and Riccio, A. (2009). To localize or not to localize: mRNA fate is in39UTR ends. Trends Cell Biol. 19, 465-474.

Andreassi, C., Zimmermann, C., Mitter, R., Fusco, S., De Vita, S., Saiardi, A. and

Riccio, A. (2010). An NGF-responsive element targets myo-inositol monophosphatase-1mRNA to sympathetic neuron axons. Nat. Neurosci. 13, 291-301.

Barrientos, S. A., Martinez, N. W., Yoo, S., Jara, J. S., Zamorano, S., Hetz, C.,

Twiss, J. L., Alvarez, J. and Court, F. A. (2011). Axonal degeneration is mediatedby the mitochondrial permeability transition pore. J. Neurosci. 31, 966-978.

Beier, C. P., Wischhusen, J., Gleichmann, M., Gerhardt, E., Pekanovic, A.,Krueger, A., Taylor, V., Suter, U., Krammer, P. H., Endres, M. et al. (2005). FasL(CD95L/APO-1L) resistance of neurons mediated by phosphatidylinositol 3-kinase-

Akt/protein kinase B-dependent expression of lifeguard/neuronal membrane protein

35. J. Neurosci. 25, 6765-6774.

Ben-Yaakov, K., Dagan, S. Y., Segal-Ruder, Y., Shalem, O., Vuppalanchi, D., Willis,

D. E., Yudin, D., Rishal, I., Rother, F., Bader, M. et al. (2012). Axonaltranscription factors signal retrogradely in lesioned peripheral nerve. EMBO J. 31,1350-1363.

Blesch, A. (2004). Lentiviral and MLV based retroviral vectors for ex vivo and in vivo

gene transfer. Methods 33, 164-172.

Cheever, T. R., Olson, E. A. and Ervasti, J. M. (2011). Axonal regeneration andneuronal function are preserved in motor neurons lacking ß-actin in vivo. PLoS ONE

6, e17768.

Desbarats, J., Birge, R. B., Mimouni-Rongy, M., Weinstein, D. E., Palerme, J. S.

and Newell, M. K. (2003). Fas engagement induces neurite growth through ERKactivation and p35 upregulation. Nat. Cell Biol. 5, 118-125.

Donnelly, C. J., Fainzilber, M. and Twiss, J. L. (2010). Subcellular communicationthrough RNA transport and localized protein synthesis. Traffic 11, 1498-1505.

Donnelly, C. J., Willis, D. E., Xu, M., Tep, C., Jiang, C., Yoo, S., Schanen, N. C.,

Kirn-Safran, C. B., van Minnen, J., English, A. et al. (2011). Limited availability

of ZBP1 restricts axonal mRNA localization and nerve regeneration capacity. EMBO

J. 30, 4665-4677.

Fernandez, M., Segura, M. F., Sole, C., Colino, A., Comella, J. X. and Cena,

V. (2007). Lifeguard/neuronal membrane protein 35 regulates Fas ligand-mediated

apoptosis in neurons via microdomain recruitment. J. Neurochem. 103, 190-203.

Gumy, L. F., Yeo, G. S., Tung, Y. C., Zivraj, K. H., Willis, D., Coppola, G., Lam,

B. Y., Twiss, J. L., Holt, C. E. and Fawcett, J. W. (2011). Transcriptome analysis ofembryonic and adult sensory axons reveals changes in mRNA repertoire localization.RNA 17, 85-98.

Hanz, S., Perlson, E., Willis, D., Zheng, J. Q., Massarwa, R., Huerta, J. J.,

Koltzenburg, M., Kohler, M., van-Minnen, J., Twiss, J. L. et al. (2003).

Axoplasmic importins enable retrograde injury signaling in lesioned nerve. Neuron

40, 1095-1104.

Hurtado de Mendoza, T., Perez-Garcia, C. G., Kroll, T. T., Hoong, N. H., O’Leary,

D. D. and Verma, I. M. (2011). Antiapoptotic protein Lifeguard is required for

survival and maintenance of Purkinje and granular cells. Proc. Natl. Acad. Sci. USA

108, 17189-17194.

Jimenez-Diaz, L., Geranton, S. M., Passmore, G. M., Leith, J. L., Fisher, A. S.,

Berliocchi, L., Sivasubramaniam, A. K., Sheasby, A., Lumb, B. M. and Hunt,

S. P. (2008). Local translation in primary afferent fibers regulates nociception. PLoS

ONE 3, e1961.

Jung, H., O’Hare, C. M. and Holt, C. E. (2011). Translational regulation in growth

cones. Curr. Opin. Genet. Dev. 21, 458-464.

Jung, H., Yoon, B. C. and Holt, C. E. (2012). Axonal mRNA localization and local

protein synthesis in nervous system assembly, maintenance and repair. Nat. Rev.

Neurosci. 13, 308-324.

Kislauskis, E. H., Zhu, X. and Singer, R. H. (1994). Sequences responsible forintracellular localization of beta-actin messenger RNA also affect cell phenotype.

J. Cell Biol. 127, 441-451.

Koenig, E., Martin, R., Titmus, M. and Sotelo-Silveira, J. R. (2000). Cryptic

peripheral ribosomal domains distributed intermittently along mammalian myelinatedaxons. J. Neurosci. 20, 8390-8400.

Lyles, V., Zhao, Y. and Martin, K. C. (2006). Synapse formation and mRNAlocalization in cultured Aplysia neurons. Neuron 49, 349-356.

Melemedjian, O. K., Asiedu, M. N., Tillu, D. V., Peebles, K. A., Yan, J., Ertz, N.,

Dussor, G. O. and Price, T. J. (2010). IL-6- and NGF-induced rapid control ofprotein synthesis and nociceptive plasticity via convergent signaling to the eIF4F

complex. J. Neurosci. 30, 15113-15123.

Merianda, T. T., Lin, A. C., Lam, J. S., Vuppalanchi, D., Willis, D. E., Karin,

N., Holt, C. E. and Twiss, J. L. (2009). A functional equivalent of endoplasmicreticulum and Golgi in axons for secretion of locally synthesized proteins. Mol. Cell.

Neurosci. 40, 128-142.

Mimouni-Rongy, M., White, J. H., Weinstein, D. E., Desbarats, J. and Almazan,

G. (2011). Fas ligand acts as a counter-receptor in Schwann cells and induces thesecretion of bioactive nerve growth factor. J. Neuroimmunol. 230, 17-25.

Natera-Naranjo, O., Aschrafi, A., Gioio, A. E. and Kaplan, B. B. (2010).Identification and quantitative analyses of microRNAs located in the distal axons

of sympathetic neurons. RNA 16, 1516-1529.

Perlson, E., Hanz, S., Ben-Yaakov, K., Segal-Ruder, Y., Seger, R. and Fainzilber,

M. (2005). Vimentin-dependent spatial translocation of an activated MAP kinase ininjured nerve. Neuron 45, 715-726.

Perry, R. B.-T., Doron-Mandel, E., Iavnilovitch, E., Rishal, I., Dagan, S. Y., Tsoory,

M., Coppola, G., McDonald, M. K., Gomes, C., Geschwind, D. H. et al. (2012).Subcellular knockout of importin b1 perturbs axonal retrograde signaling. Neuron 75,

294-305.

Roche, F. K., Marsick, B. M. and Letourneau, P. C. (2009). Protein synthesis in distal

axons is not required for growth cone responses to guidance cues. J. Neurosci. 29,638-652.

Schweitzer, B., Taylor, V., Welcher, A. A., McClelland, M. and Suter, U. (1998).Neural membrane protein 35 (NMP35): a novel member of a gene family which is

highly expressed in the adult nervous system. Mol. Cell. Neurosci. 11, 260-273.

Schweitzer, B., Suter, U. and Taylor, V. (2002). Neural membrane protein 35/

Lifeguard is localized at postsynaptic sites and in dendrites. Brain Res. Mol. Brain

Res. 107, 47-56.

Axonal translation of NMP35 increases axon growth 101

Journ

alof

Cell

Scie

nce

Segura, M. F., Sole, C., Pascual, M., Moubarak, R. S., Perez-Garcia, M. J.,Gozzelino, R., Iglesias, V., Badiola, N., Bayascas, J. R., Llecha, N. et al. (2007).The long form of Fas apoptotic inhibitory molecule is expressed specifically inneurons and protects them against death receptor-triggered apoptosis. J. Neurosci. 27,11228-11241.

Shestakova, E. A., Singer, R. H. and Condeelis, J. (2001). The physiologicalsignificance of beta -actin mRNA localization in determining cell polarity anddirectional motility. Proc. Natl. Acad. Sci. USA 98, 7045-7050.

Smith, D. S. and Skene, J. H. (1997). A transcription-dependent switch controlscompetence of adult neurons for distinct modes of axon growth. J. Neurosci. 17, 646-658.

Sole, C., Dolcet, X., Segura, M. F., Gutierrez, H., Diaz-Meco, M. T., Gozzelino, R.,Sanchis, D., Bayascas, J. R., Gallego, C., Moscat, J. et al. (2004). The deathreceptor antagonist FAIM promotes neurite outgrowth by a mechanism that dependson ERK and NF-kapp B signaling. J. Cell Biol. 167, 479-492.

Taylor, A. M., Berchtold, N. C., Perreau, V. M., Tu, C. H., Li Jeon, N. and Cotman,C. W. (2009). Axonal mRNA in uninjured and regenerating cortical mammalianaxons. J. Neurosci. 29, 4697-4707.

Toth, C. C., Willis, D., Twiss, J. L., Walsh, S., Martinez, J. A., Liu, W. Q., Midha,R. and Zochodne, D. W. (2009). Locally synthesized calcitonin gene-related Peptidehas a critical role in peripheral nerve regeneration. J. Neuropathol. Exp. Neurol. 68,326-337.

Twiss, J. L., Smith, D. S., Chang, B. and Shooter, E. M. (2000). Translational controlof ribosomal protein L4 mRNA is required for rapid neurite regeneration. Neurobiol.

Dis. 7, 416-428.Vuppalanchi, D., Coleman, J., Yoo, S., Merianda, T. T., Yadhati, A. G., Hossain, J.,

Blesch, A., Willis, D. E. and Twiss, J. L. (2010). Conserved 39-untranslated regionsequences direct subcellular localization of chaperone protein mRNAs in neurons.J. Biol. Chem. 285, 18025-18038.

Vuppalanchi, D., Merianda, T. T., Donnelly, C., Pacheco, A., Williams, G., Yoo, S.,

Ratan, R. R., Willis, D. E. and Twiss, J. L. (2012). Lysophosphatidic aciddifferentially regulates axonal mRNA translation through 59UTR elements. Mol. Cell.

Neurosci. 50, 136-146.Willis, D. E. and Twiss, J. L. (2010). Regulation of protein levels in subcellular

domains through mRNA transport and localized translation. Mol. Cell. Proteomics 9,952-962.

Willis, D. E. and Twiss, J. L. (2011). Profiling axonal mRNA transport. Methods Mol.

Biol. 714, 335-352.

Willis, D., Li, K. W., Zheng, J. Q., Chang, J. H., Smit, A. B., Kelly, T., Merianda,

T. T., Sylvester, J., van Minnen, J. and Twiss, J. L. (2005). Differential transport

and local translation of cytoskeletal, injury-response, and neurodegeneration protein

mRNAs in axons. J. Neurosci. 25, 778-791.

Willis, D. E., van Niekerk, E. A., Sasaki, Y., Mesngon, M., Merianda, T. T.,

Williams, G. G., Kendall, M., Smith, D. S., Bassell, G. J. and Twiss, J. L. (2007).

Extracellular stimuli specifically regulate localized levels of individual neuronal

mRNAs. J. Cell Biol. 178, 965-980.

Willis, D. E., Xu, M., Donnelly, C. J., Tep, C., Kendall, M., Erenstheyn, M., English,

A. W., Schanen, N. C., Kirn-Safran, C. B., Yoon, S. O. et al. (2011). Axonal

Localization of transgene mRNA in mature PNS and CNS neurons. J. Neurosci. 31,

14481-14487.

Yudin, D., Hanz, S., Yoo, S., Iavnilovitch, E., Willis, D., Gradus, T., Vuppalanchi,

D., Segal-Ruder, Y., Ben-Yaakov, K., Hieda, M. et al. (2008). Localized regulation

of axonal RanGTPase controls retrograde injury signaling in peripheral nerve. Neuron

59, 241-252.

Zelena, J. (1970). Ribosome-like particles in myelinated axons of the rat. Brain Res. 24,

359-363.

Zheng, J.-Q., Kelly, T. K., Chang, B., Ryazantsev, S., Rajasekaran, A. K., Martin,

K. C. and Twiss, J. L. (2001). A functional role for intra-axonal protein synthesis

during axonal regeneration from adult sensory neurons. J. Neurosci. 21, 9291-9303.

Zhong, X., Schneider, T. J., Cabral, D. S., Donohoe, T. J. and Rothstein, T. L.

(2001). An alternatively spliced long form of Fas apoptosis inhibitory molecule

(FAIM) with tissue-specific expression in the brain. Mol. Immunol. 38, 65-72.

Zivraj, K. H., Tung, Y. C., Piper, M., Gumy, L., Fawcett, J. W., Yeo, G. S. and Holt,

C. E. (2010). Subcellular profiling reveals distinct and developmentally regulated

repertoire of growth cone mRNAs. J. Neurosci. 30, 15464-15478.

Zuliani, C., Kleber, S., Klussmann, S., Wenger, T., Kenzelmann, M., Schreglmann,

N., Martinez, A., del Rio, J. A., Soriano, E., Vodrazka, P. et al. (2006). Control of

neuronal branching by the death receptor CD95 (Fas/Apo-1). Cell Death Differ. 13,

31-40.

Journal of Cell Science 126 (1)102