RNAi pathway is functional in peripheral nerve axons

The FASEB Journal • Research Communication

RNAi pathway is functional in peripheral nerve axonsAlexander K. Murashov,*,1 Vishnu Chintalgattu,*,2 Rustem R. Islamov,#,2

Teresa E. Lever,†,2 Elena S. Pak,* Paulina L. Sierpinski,* Laxmansa C. Katwa,*and Michael R. Van Scott**Department of Physiology, and †Department of Communication Sciences and Disorders,East Carolina University, Greenville, North Carolina, USA; and #Department of Histology, KazanState Medical University, Kazan, Russia

ABSTRACT Recent observations demonstrated thattranslation of mRNAs may occur in axonal processes atsites that are long distances away from the neuronalperikaria. While axonal protein synthesis has beendocumented in several studies, the mechanism of itsregulation remains unclear. The aim of this study was toinvestigate whether RNA interference (RNAi) may beone of the pathways that control local protein synthesisin axons. Here we show that sciatic nerve containsArgonaute2 nuclease, fragile X mental retardation pro-tein, p100 nuclease, and Gemin3 helicase—compo-nents of the RNA-induced silencing complex (RISC).Application of short-interfering RNAs against neuronal�-tubulin to the sciatic nerve initiated RISC formation,causing a decrease in levels of neuronal �-tubulin IIImRNA and corresponding protein, as well as a signifi-cant reduction in retrograde labeling of lumbar motorneurons. Our observations indicate that RNAi is func-tional in peripheral mammalian axons and is indepen-dent from the neuronal cell body or Schwann cells. Weintroduce a concept of local regulation of axonaltranslation via RNAi.—Murashov, A. K., Chintalgattu,V., Islamov, R. R., Lever, T. E., Pak, E. S., Sierpinski,P. L., Katwa, L. C., Van Scott, M. R. RNAi pathway isfunctional in peripheral nerve axons. FASEB J. 21,656–670 (2007)

Key Words: RNA interference � axonal protein synthesis� siRNA � sciatic nerve � DRG

Rna interference (rnai) is a natural mechanismthat acts to selectively suppress gene expression (1–4).The RNAi machinery appears to have appeared early inevolution to protect the eukaryotic genome from en-dogenous transposable elements and from viral infec-tions (5). Recent observations demonstrated that inaddition to its protective action, RNAi plays an impor-tant role during cell growth and differentiation (6) andin early development (7, 8). Regulation of proteinsynthesis via RNAi is based on the expression of a groupof small endogenous RNA molecules, called microRNAs(miRNAs). miRNAs are produced by special enzymescalled Drosha and Dicer (9, 10), with subsequentinitiation of the RNA-induced silencing complex(RISC). RISC is a multiprotein complex that is consid-ered to be the effector mechanism of the RNAi pathway

(11–13). A common architecture for RISC includes astrand of either miRNA or short interfering RNA(siRNA) and several proteins, including Argonaute2(AGO2), fragile X mental retardation protein (FMRP),and Tudor-SN/p100 mammalian homologue (14–17).The strand of miRNA or siRNA retained within RISCserves as a template to identify the complementarynucleotide sequence of mRNAs within the cell. Finally,through a catalytic process, RISC cleaves target mRNAs,rendering them incapable of protein synthesis (18).

Recent studies have revealed that the miRNA path-way influences a variety of processes, including earlydevelopment (19), cell proliferation (20), and celldeath (21). Some miRNAs were identified in the verte-brate nervous system (22, 23) and in neuronal cells(24–27). It was also shown that miRNAs play an impor-tant role in brain development (28) and neuronalmaturation (29, 30). A novel study has indicated thatbrain-specific miRNAs may regulate dendritic spinedevelopment in mammalian brain (31). The involve-ment of miRNAs in dendritic development is particu-larly interesting, given the indication that selectedmRNAs in neurons are delivered to sites of synapticcontact that are remote from the neuronal cell body(32, 33). Although activity of the RNAi pathway wasdemonstrated in dendrites, the possibility of regulationof axonal protein synthesis via RNAi awaits thoroughinvestigation.

Recent observations demonstrated that translation ofmRNAs may occur locally in axonal processes at siteslong distances away from the neuronal cell body (34–37). Even though axonal local protein synthesis hasbeen documented in a number of studies, the mecha-nism of its regulation remains unclear. A recent obser-vation revealed that some cytoskeletal proteins (�-actin,�-tubulin, peripherin, vimentin, gamma-tropomyosin3, and cofilin-1), heat-shock proteins (HSP27, HSP60,HSP70, grp75, alphaB crystallin), and proteins associ-ated with neurodegenerative diseases [ubiquitin C-terminal hydrolase L1, rat ortholog of human DJ-1/

1 Correspondence: Department of Physiology, Brody Schoolof Medicine, East Carolina University, 600 Moye Blvd., Green-ville, NC 27834, USA. E-mail: [email protected]

2 These authors contributed equally to this work.doi: 10.1096/fj.06-6155com

656 0892-6638/07/0021-0656 © FASEB

Park7, gamma-synuclein, superoxide dismutase (SOD)1] are synthesized in axonal fibers (38). The presenceof mRNAs and protein synthesis in axons suggests theexistence of a process that regulates protein expressionlocally. A potential mechanism of regulation of axonalprotein synthesis may be RNAi.

In the current study, we asked whether RNAi may beone of the pathways that control local protein synthesisin axons. Here we show that murine sciatic nervecontains a multiprotein complex consisting of AGO2nuclease, FMRP, p100 nuclease, and Gemin3 (a DEAD-box RNA helicase), which was induced by applicationof siRNAs. siRNA duplexes against neuronal �-tubulinapplied onto the proximal stump of the sciatic nerveinitiated RISC formation, with a concomitant dramaticdecrease in levels of �-tubulin mRNA and a correspond-ing protein product in the sciatic nerve fibers, as well asa significant reduction in retrograde labeling of lumbarmotor neurons with Fluorogold. Our observations in-dicate that RNAi is functional in peripheral mamma-lian axons and is independent from the neuronal cellbody or Schwann cells. Here we propose a new conceptof regulation of axonal protein synthesis via the RNAipathway.

MATERIALS AND METHODS

Animals

Experiments were performed on 8-wk-old male ICR miceobtained from Charles River Laboratories (Wilmington, MA,USA). Animals were housed one per cage under standardlaboratory conditions, with a 12 h light/dark schedule andunlimited access to food and water. Animal protocols wereapproved by the Animal Care and Use Committee of EastCarolina University, an AAALAC-accredited facility.

Surgery, ligatures, and colchicine treatment

Animals were anesthetized with ketamine (18 mg/ml) -xyla-zine (2 mg/ml) anesthesia (0.5 ml/10 g of body wt, i.p.). Asshown in Fig. 1, the sciatic nerve was cut in the midthigh anda 7 mm silastic tube with a 1.47 mm inner diameter wasapplied to the proximal nerve stump (39). Saline, antineuro-nal �-tubulin III siRNA (accession #AF312873), nonspecificcontrol siRNA, or a mixture of siRNA with Fluorogold waspipetted into the tube (7 �l total volume). The lower end ofthe tube was sealed with petroleum jelly (Vaseline) and gluedto the surrounding skeletal muscles with tissue adhesive (3M228 Vetbond 228). The incision was closed with wound clips.In sham operations, sciatic nerves were exposed but nottouched. In some experiments, to investigate whether theRNAi process in sciatic nerve may exist and function inde-pendently from the neuronal cell body, a ligation or localapplication of colchicine was used at the time of the surgery.Ligatures or colchicine (5 mM) were applied at the proximalpart of sciatic nerve, �2 cm above the end of the sciatic nervestump submersed in the pouch, for the duration of theexperiment (24 h).

Retrograde labeling and quantification oflabeled motoneurons

For retrograde labeling of neurons, FITC-conjugated siRNA(Dharmacon, Inc.; Chicago, IL, USA; 2 �l of 20 �M of siRNA

in 7 �l of PBS) and/or 7 �l of 5% Fluorogold (Fluorochrom;Denver, CO, USA) in saline was applied to the proximalstump of sciatic nerves. In some experiments, to determinewhether siRNA is taken up and/or transported by the sciaticnerve, siRNA targeted at a nonmammalian sequence of fireflyLuciferase and labeled with FITC or Cy3 (Dharmacon; 2 �l of20 �M) was applied onto the sciatic nerve. As a negativecontrol, some sciatic nerves were treated with unconjugatedFITC or Cy3. The doses for siRNA and Fluorogold werechosen based on the manufacturer’s instruction (Dharma-con) and our published observations (39). To inhibit proteinsynthesis in axons, 10 �g/ml cycloheximide (39), an inhibitorof protein synthesis (Sigma; St. Louis, MO, USA), was appliedonto the sciatic nerve proximal stump. In some experiments,siRNAs were applied to the nerve stump 6 h before applica-tion of a mixture containing Fluorogold to allow for theinitiation of RNAi machinery. Cryostat serial transversal 30�m sections of lumbar spinal cord were analyzed with fluo-rescence microscopy. Fluorogold-labeled neurons were rec-ognized by their size, shape, and location and counted inevery section, without allowing for split nucleoli, according toa protocol described elsewhere (40, 41). The images wereanalyzed for the mean cross-sectional area of neurons usingImageJ (NIH Image). In previous studies, no difference in themean cross-sectional area of labeled cells between controland treated animals was observed, and Abercrombie’s correc-tion for neuronal counts was not required (39). All values arerepresented as the mean � se. Statistical analysis was per-formed with one-way ANOVA and Newman-Keuls MultipleComparison Test using Prism (GraphPad Prism version 3.00for Windows, GraphPad Software; San Diego, CA, USA).

Spinal cord and sciatic nerve collection

After sacrifice, we carefully verified that the silastic tube wasstill in place and that the solution did not leak out. Animalswith misplaced tubes or leaked out solutions were not usedfor further analysis. Mice were overdosed with sodium pento-barbital, and spinal cords and sciatic nerves were quicklydissected out, snap frozen in liquid nitrogen, and stored at–80°C. For histology, animals were perfused with cold PBS,followed by cold 4% paraformaldehyde in PBS (pH 7.4).Lumbar spinal cords and sciatic nerves were removed andpostfixed in 4% paraformaldehyde for 2 h, cryoprotected in

Figure 1. Schematic of experimental procedure for retro-grade labeling. See Materials and Methods for details.

657RNAI IN AXONAL FIBERS

30% sucrose overnight, and embedded in TBS tissue freezingmedium (Triangle Biomedical Science; Durham, NC, USA).Spinal cords and sciatic nerves were cut into 10 �m serialsections (42).

In experiments using sciatic nerve in vivo, it is impossible toseparate axonal fibers from Schwann cells. That is why in ourbiochemical and histological experiments we used a portionof sciatic nerve well above the site of siRNA application(unless otherwise noted), so that the Schwann cells of sciaticnerves used in analyses were never in physical contact withsiRNAs. We collected an �2 cm portion of sciatic nerveproximal stump �0.5 cm above the treatment site. Therefore,all changes in mRNA and RISC protein levels observed in thestudy occurred in the portion of sciatic nerve located wellabove the level of siRNA treatment (Fig. 1). We assumed thatthe probability for Schwann cells to transport siRNA and/orRISC proteins across two membranes or to express RISC inresponse to axonal uptake of siRNA was extremely low.

Real-time RT-polymerase chain reaction (RT-PCR)

To evaluate the expression of neuronal �-tubulin III mRNA(accession #AF312873) in the sciatic nerve, real-time quanti-tative RT-PCR was performed. Total RNA was isolated fromintact, sham-operated, or siRNA-treated samples of sciaticnerve (n�6) (43, 44). Samples from the experimental andcontrol tissue were run three times in duplicate using athermoscript one-step quantitative RT-PCR platinum Taq kit(Invitrogen, Life Technologies; Carlsbad, CA, USA) in theCepheid RT-PCR thermocycler (Sunnyvale, CA, USA), asdescribed previously (44, 45). The real-time PCR reactionswere carried out in the presence of mouse gene-specificprimers for neuronal �-tubulin III (accession #AF312873;forward 5�-GAGGACAGAGCCAAGTGGAC-3� and reverse 5�-CAGGGCCAAGACAAGCAG-3�) (46), and GAPDH (forward5�-AGATCCACAACGGATACATT-3� and reverse 5�-5�-TC-CCTCAAGATTGTCAGCAA-3�) (45). The GAPDH served asan internal control to monitor cDNA synthesis efficiency.Relative quantitation of gene expression was based on themodel by Pfaffl (47), which is based on the relative expressionof a target gene vs. a reference gene. Quantification of geneexpression was performed by comparing amplified productsto the generated �-actin standard curve (45). �-TubulinmRNA copy numbers from intact sciatic nerves were taken as100%. This value was calculated by taking an average of genecopy numbers amplified from the three sets of PCR reactionsperformed in duplicate.

In situ hybridization

The protocol for in situ hybridization was adapted from apreviously described protocol (48). Briefly, sections of sciaticnerves were postfixed in 4% paraformaldehyde solution inPBS (pH 7.0) at room temperature for 5 min, washed in PBStwice for 5 min, and incubated in 100% methanol � 0.3%hydrogen peroxide solution for 10 min at 4°C. After twowashes in PBS for 10 min, slides were prehybridized at 42°Cfor 1 h in hybridization buffer containing 600 mM sodiumchloride, 50 mM sodium phosphate buffer (pH 7.0), 5.0 mMEDTA (Sigma), 0.02% Ficoll (Sigma), 0.02% BSA (Sigma),0.02% polyvinylpyrrolidone (Sigma), 200 ng/ml sheared anddenatured salmon sperm DNA (Sigma), and 40% formamide(Super Pure, Fisher Scientific; Pittsburgh, PA, USA). Hybrid-ization was performed at 42°C in the same buffer with theaddition of dextran sulfate to 7%, tRNA (baker’s yeast) to 0.1mg/ml, poly-A to 10 �g/ml, and biotinylated antisense (5�-biotin-TCTGACCAAAGATAAAGTTGTCGGGCCTGAAT-AGGTGTCCAAAGG-biotin-3�) or sense oligonucleotide

probe against neuronal �-tubulin III (accession #AF312873) toa final concentration of 30 ng/�l of hybridization mixture.After overnight incubation, sections were washed in 2� salinesodium citrate (SSC) at 40°C for 10 min, washed at roomtemperature for 10 min in 1� SSC, then for 10 min in 0.25�SSC. Sections were subsequently washed in PBS and incu-bated for 1 h with an avidin-biotin-peroxidase (ABC) reagent(Vector Laboratories; Burlingame, CA, USA). To visualize thesignal, anti-horseradish peroxidase, Cy3-conjugated antibod-ies (Chemicon International, Inc.; Temecula, CA, USA) wereapplied for 2 h in 1:100 dilution. As a negative control, somesections were pretreated with RNase (42) before viewing onan Olympus IMT-2 fluorescent microscope or a Zeiss LSM 510confocal laser scanning microscope.

Protein lysates

Proteins were extracted from the sciatic nerves or spinal cordsand subfractionated into s100 and p100 fractions accordingto a protocol described elsewhere (49). Samples of spinalcords and sciatic nerves were homogenized in ice-cold lysisbuffer containing 10 mM potassium acetate, 2 mM magne-sium acetate, 2 mM DTT, 5 mM HEPES (pH 7.3), 20 �Mcytochalasin B, 0.5%, aprotinin, 2 �g/ml leupeptin, 2 �g/mlpepstatin and RNAsin (1 U/�l) and incubated for 10 min onice. Samples were centrifuged at 1500 g for 15 min at 4°C.After centrifugation, the supernatant (cytoplasmic lysate) wascentrifuged at 100,000 g for 45 min at 4°C, yielding superna-tant (s100) and pellet (p100) fractions. The p100 fraction wasgently resuspended in 1 volume of buffer B and stored at–80°C until ready to use.

Immunoblot analysis

The solubilized proteins were loaded at 20 �g per lane andseparated by relative size by SDS-PAGE analysis in precast 5%Tris-HCl gels (Bio-Rad; Hercules, CA, USA), then the sepa-rated proteins were transferred to Immobilon P membranes(Millipore Corporation; Bedford, MA, USA). Membraneswere probed with primary antibodies according to our stan-dard immunoblotting laboratory procedures. Secondaryhorseradish peroxidase-conjugated antibodies (Roche Molec-ular Biochemicals; Indianapolis, IN, USA) were used at1:5000 dilution. Protein bands were visualized using a chemi-luminescence detection system (ECL kit; Amersham; Arling-ton Heights, IL, USA).

Immunoprecipitation

Following procedures described before (42), immunoprecipi-tation was done in a buffer containing 100 mM KCl, 5 mMEDTA, 10 mM HEPES (pH 7.3), 1 mM DTT, 2 �g/ml ofleupeptin, 2 �g/ml of pepstatin, 0.5% aprotinin, and0.5%Triton-X100. Five micrograms of protein from eachsample of p100 fraction was immunoprecipitated with anti-bodies specific for AGO2 or p100 overnight and precipitatedusing BSA preblocked protein A-Sepharose beads (BD,PharMingen; San Diego, CA, USA). After overnight incuba-tion, the beads were washed four times with 1% BSA of theprevious buffer and centrifuged at 10,000 rpm for 5 min at4°C. The supernatant was removed and 40 �l of X2 WesternStop solution was added. Samples were boiled for 5 min,centrifuged again, then the supernatant was collected andsubjected to electrophoretic analysis on precast 5% Tris-HClgels (Bio-Rad). Controls included incubation with nonim-mune serum, incubation without primary antibodies, andcompetition of antibodies with recombinant target proteins.

658 Vol. 21 March 2007 MURASHOV ET AL.The FASEB Journal

Immunohistochemistry

Sections of sciatic nerves and lumbar spinal cords wereprocessed as described previously (42, 50). Serial sectionswere stained using the Elite ABC Kit (Vector Laboratories).Tissue antigens were visualized with DAB Substrate Kit forPeroxidase (Vector Laboratories). After staining, sectionswere air dried and permanently mounted with DPX (Sigma).Incubations without primary or secondary antibodies, or ABCreagent, were used as negative controls. Slides were examinedwith an Olympus IMT-2 fluorescent microscope. Images wererecorded using the Spot digital camera system (DiagnosticInstruments; Sterling Heights, MI, USA).

Fluorescence and confocal microscopy

Sections of sciatic nerves and spinal cords were incubatedwith primary antibodies (1:100 dilution) overnight at 4°C ona rocker. Secondary FITC-, TX Red-, CY3-, or CY5-conjugatedIgG (Jackson ImmunoResearch Laboratories, Inc.; WestGrove, PA, USA) were applied in a 1:100 dilution for 1 h atroom temperature. Sections were then washed with PBST andstained with 4�,6�-diam idino-2-phenylidole (Sigma) for 5min. Sections were rinsed with dH2O and mounted usingantifading Gel/Mount (Biomeda; Foster City, CA, USA).Images were captured using a Spot digital camera. A ZeissLSM 510 confocal laser scanning microscope was used fordouble- and triple-stained specimens and for subcellularresolution.

siRNA reagents

�-Tubulin duplex was purchased from Dharmacon and ready touse in transfection experiments. The siRNA target sequence5�-GACAGAGCCAAGTGGACTCAC-3� (accession #AF312873)was validated in previous RNAi gene silencing experiments (51).The �-tubulin siRNA duplex consisted of the sequence 5�-CAGAGCCAAGUGGACUCACdTdT/dTdTGUCUCGGUU-CACCUGAGUG-5�, and the nonspecific control siRNA duplexwas 5�-ACUCUAUCGCCAGCGUGACUU/UUUGAGAUAGCG-GUCGCACUGP-5� (Dharmacon). Duplexes were dissolved to afinal concentration of 20 �M in universal buffer (20 mM KCl,6 mM HEPES pH 7.5, and 0.2 mM MgCl) according to themanufacturer’s instruction. For some experiments, the du-plexes were labeled with FITC using a Silencer™ siRNALabeling Kit (Ambion; Austin, TX, USA), according to themanufacturer’s manual.

General list of antibodies

The following antibodies were used for immunodetectionprocedures: rabbit polyclonal anti-AGO2-specific antibody(kindly donated by Tom Hobman, University of Alberta,Canada; this antibody was successfully used in an earlierstudy, see ref. 14). Guinea pig polyclonal anti-p100 antibody(kindly donated by Tom Keenan, VA Polytechnic Instituteand State University, Blacksburg, VA, USA) was tested forspecificity (52) and successfully used elsewhere (14). Mousemonoclonal anti-FMRP antibody (7G-1; DevelopmentalStudies Hybridoma Bank) was tested for specificity and suc-cessfully used previously (14, 53). Mouse monoclonal anti-Gemin3-specific antibody was from ImmuQuest Ltd., (Cleve-land, UK). Anti-Gemin3 antibody (gift from Dr. GideonDreyfuss, University of Pennsylvania, Philadelphia, PA, USA)was tested for specificity and successfully used in previousstudies (54, 55). Mouse monoclonal neuron-specific beta IIItubulin antibody [TUJ-1] was provided by Covance ResearchProducts, Inc. (Denver, PA, USA).

DRG culture

Primary dissociated cultures were prepared from DRGs ofadult ICR mice according to a protocol described elsewhere(38). Some of the rat DRG cultures were also obtainedcommercially from Cambrex BioScience (Walkersville, MD,USA). For immunolocalization studies, rat DRGs were cul-tured at low density on poly-l-lysine/laminin-coated cover-slips according to the manufacturer’s protocol. In someexperiments, DRG cultures were transfected with siRNAduplexes using DharmaFECT3 (Dharmacon) according tothe manufacturer’s instructions.

To isolate axonal proteins, we used a culture method forisolation of DRG axons already described (38). Briefly, disso-ciated DRGs were plated into tissue culture inserts containingporous membrane (8 �m pores; BD Falcon, Bedford, MA,USA) coated with poly-l-lysine/laminin. Axons were isolatedafter 16–20 h in culture by carefully scraping cellular contentsfrom the upper membrane surface. For isolation of the cellbody compartment, the under surface of the membrane wasscraped in an identical manner. Samples obtained were usedfor protein isolation and subsequent Western blot analysis.

Statistical analysis

Statistical analysis was performed with one-way ANOVA andNewman-Keuls Multiple Comparison Test using Prism(GraphPad Prism version 3.00 for Windows, GraphPad Soft-ware; San Diego, CA, USA). All values are represented as themean � se.

RESULTS

Immunodetection of components of RISC in sciaticnerve and spinal cord lysates

The aim of this experiment was to determine whetherthe known components of mammalian RISC (AGO2,FMRP, and p100) were present in the sciatic nervefibers. For these experiments we chose to target neuro-nal �-tubulin because it is one of the abundant cytoskel-etal proteins known to be synthesized locally in theaxonal fibers (38). Proximal stumps of sciatic nervestransected at midthigh level were incubated with anti-tubulin siRNA, nonspecific control siRNA, or vehiclefor 24 h. For protein extraction we used a 2 cm portionof sciatic nerve 0.5 cm above the site of treatment (Fig.1). For spinal cord protein extracts we used lumbarspinal cord from the corresponding animals. Proteinsamples isolated from sciatic nerves and spinal cordswere resolved by SDS/PAGE electrophoresis and trans-ferred to membranes, which were subsequently probedwith antibodies against AGO2, FMRP, and p100. Theseantibodies have been successfully used elsewhere (14).Immunoblot analysis revealed the presence of AGO2,FMRP, and p100 proteins in samples from the spinalcord (Fig. 2, upper panel) and sciatic nerves (Fig. 2,lower panel). In sciatic nerves, pretreatment with anti-tubulin siRNA visibly increased the level of p100 expres-sion, which may be a sign of induction by siRNA. Thedata of this experiment demonstrated that the three


known components of RISC (AGO2, FMRP, and p100)were present in sciatic nerve fibers.

Components of RNAi machinery, AGO2, FMRP,p100, and Gemin3 are present in sciatic nerve fibers

To confirm our finding at the histological level, weperformed immunofluorescence on sciatic nerve fi-bers to investigate the distribution of the known(AGO2, FMRP, and p100) and suspected (Gemin3)components of RISC. Longitudinal sections of an �2cm portion of sciatic nerve proximal stump (�0.5 cmabove the site of siRNA treatment) were incubatedwith primary antibodies against target proteins andsubsequently incubated with secondary antibodiesconjugated with Texas Red, FITC, or CY3. Thesections were analyzed using a Zeiss LSM 510 confo-cal laser scanning microscope. Triple immunostain-ing of these sections revealed immunoreactivity inthe sciatic nerve for p100 (blue color-coded), AGO2(red), FMRP (green), and Gemin3 (green) proteins(Fig. 3, upper and lower panels). A merge of theimages showed coexpression of p100 (blue), AGO2(red), Gemin3 (green) (Fig. 3, upper panel) as well

as coexpression of p100 (blue), AGO2 (red), andFMRP (green) (Fig. 3, lower panel). These data dem-onstrated that AGO2, FMRP, p100, and Gemin3 coex-

Figure 2. Immunoblot analysis provides evidence of AGO2,FMRP, and p100 proteins in spinal cords and sciatic nerves.Protein lysates isolated from sciatic nerves or spinal cordswere resolved by electrophoresis, transferred to membranes,then probed with antibodies against AGO2 (1:1000), FMRP(1:1000), and p100 (1:1000). 20 �g of protein was loaded pereach lane. Immunoblot analysis revealed AGO2, FMRP, andp100 proteins in samples from the spinal cords (upper panel)and sciatic nerves (lower panel). Note that in sciatic nerves,pretreatment with siRNA increased the level of p100 expres-sion.

Figure 3. Confocal immunofluorescence for p100, AGO2,Gemin3, and FMRP proteins in sciatic nerve fibers. Longitu-dinal sections of sciatic nerves were incubated with primaryantibodies against p100, AGO2, and Gemin3 or FMRP andsubsequently incubated with secondary antibodies conjugatedwith Texas Red, FITC, or CY5. All primary antibodies wereapplied in a concentration of 1:100. Upper panel shows triplestaining against p100 (blue), AGO2 (red), and Gemin3(green). Lower panel shows triple staining against p100(blue), AGO2 (red), and FMRP (green). The fluorescentsignals indicate that the major components of RISC (p100,AGO2, Gemin3, and FMRP) are expressed in sciatic nervefibers (indicated by white arrows).


press in peripheral nerve fibers, providing the neces-sary substrates for the formation of RISC.

Immunostaining of dissociated DRG neurons revealsthe presence of RISC proteins in axons

To investigate whether RISC proteins are present inaxons of dissociated neuronal cultures, we performedimmunofluorescence on rat DRG neurons plated at lowdensity according to a protocol described elsewhere(37, 38). The preparations were stained with antibodiesagainst AGO2, p100, FMRP, and Gemin3, as well as withTUJ1 antibodies against neuronal �-tubulin and anti-bodies against growth-associated protein 43 (GAP43).In some experiments, DRG cultures were treated withantitubulin siRNA or nonspecific control for 24 hbefore immunostaining. Immunofluorescence showedcolocalization of p100 with �-tubulin (Fig. 4A), p100with FMRP (Fig. 4B), p100 with AGO2 (Fig. 4C), andp100 with Gemin3 (Fig. 4D). A merge of the imagesshowed a complete overlap between immunofluores-cence of the proteins of interest in axons as well as incell bodies. Incubation of DRG cultures with siRNAsdemonstrated successful uptake of siRNA duplexes.FITC-conjugated siRNAs were detected in axons andbodies of DRG neurons. Triple immunofluorescence ofDRG cultures treated with antitubulin siRNA con-firmed colocalization of p100 with �-tubulin (Fig. 4E),p100 with GAP43 (Fig. 4F), p100 with AGO2 (Fig. 4G),and p100 with FMRP (Fig. 4H). These results clearlydemonstrated the existence of RISC proteins in axons.

To confirm this finding at the protein level, weperformed immunoblot analysis on axonal proteinsisolated from DRG cultures according to a methoddescribed previously (38). This method allows theseparation of DRG processes from cell body and non-neuronal cells by culturing neurons on a porous mem-brane that allows passage of axons but restricts the cellbody and non-neuronal cells to the upper membranesurface. Proteins extracted from the cell body andaxonal preparations were separated by 7.5% SDS-PAGE(10 �g/lane) and transferred to PVDF membrane. Todemonstrate the purity of the DRG axonal preparation,membranes were probed with antibodies against MAP-2and GAP43. MAP2 protein resides in the cell soma anddendrites but does not extend into the axon (37),whereas GAP43 is localized to the growing axons andsoma (56). Our immunoblot showed that MAP2 waspresent in the cell body fraction but absent from axonalpreparation, whereas GAP43 was more abundant in theaxonal preparation than in the cell body fraction, thusconfirming the purity of the axonal fraction. Theimmunoblot analysis revealed the presence of RISCproteins (p100, AGO2, FMRP, and Gemin3) in boththe cell body and axonal fractions (Fig. 5). Togetherwith the immunofluorescence experiments on sciaticnerve fibers and dissociated DRG cultures, these dataargue that axons possess proteins specific for RISC.

Figure 4. RISC proteins in dissociated DRG cultures imagedby confocal microscopy. Cultures of rat DRG neurons werecolabeled with antibodies to p100 (blue) and to A) neuronal�-tubulin (red); B) FMRP (red); C) AGO2 (red); D) Gemin3(red). Merges of the images indicate colocalization of thestudied proteins in soma and axons. E–H) Fluorescent imagesof neurons in DRG cultures treated with antitubulin siRNAconjugated with FITC (green). E) Colocalization of p100(blue) and �-tubulin (red). F) colocalization of p100 (blue)and GAP43 (red). G) Colocalization of p100 (blue) andAGO2 (red). H) Colocalization of p100 (blue) and FMRP(red). Merged images reveal colocalization of investigatedproteins and siRNA to soma and axons. White arrows indicatecell bodies, arrowheads show axons.


AGO2, FMRP, p100, and Gemin3 form complexeswithin sciatic nerve in response to antitubulin siRNA

The purpose of this experiment was to determinewhether AGO2, FMRP, p100, and Gemin3 can formprotein complexes in sciatic nerve fibers in response totreatment with siRNA against neuronal �-tubulin. Theproteins were isolated from �1 cm distal and �1 cmproximal parts of the proximal stump dissected out 0.5cm above the site of siRNA application. The controltreatments included saline and nonspecific siRNA. Pro-tein lysates were incubated with AGO2 or p100 antibod-ies overnight and precipitated using BSA preblockedprotein A-Sepharose beads. The precipitated proteinswere separated according to relative size by SDS-PAGEanalysis in 5% gels, transferred to membranes, andprobed with antibodies against AGO2, FMRP, p100,and Gemin3. In all treatment conditions, p100 wasshown to coprecipitate with FMRP and AGO2 (Fig. 6,upper panel), whereas AGO2 was shown to coprecipi-tate with p100 and Gemin3 (Fig. 6, middle panel). Asexpected, expression of the multiprotein complex wasmore pronounced in response to treatment by siRNA,indicating that the presence of siRNA induces RISCformation (14).

To investigate whether RNAi may act independentlyfrom the neuronal cell body, experiments with ligatureand colchicine treatment were performed. The ligaturewas used to mechanically separate the sciatic nerveproximal stump from the neuronal cell bodies. Colchi-cine, which causes depolymerization of microtubules

(57), was used to block axonal transport pharmacolog-ically without affecting the nerve capacity to conductaction potentials. Ligation or colchicine were appliedfor 24 h at the proximal stump of the sciatic nerve �1.5cm above the site of the treatment with siRNAs. Exper-iments showed that neither the ligature nor colchicineaffected RISC protein levels in the proximal stumpabove or below the ligature/colchicine application site.This observation indicated that RISC forms locally andindependently from the neuronal cell body.

Figure 5. Immunoblot analysis for p100, AGO2, Gemin3, andFMRP proteins in axonal protein fraction. Protein wereextracted from the neuronal cell body (N) and axonalpreparations (A) from mouse DRG cultures. Samples wereloaded 10 �g per lane and separated by 7.5% SDS-PAGE.MAP2, which resides in the cell soma and dendrites but isrestricted from axons, was present in the cell body fractionbut absent from axonal preparation. By immunoblotting, aband corresponding to GAP43 was more abundant in theaxonal preparation than in the cell body fraction. RISCproteins p100, AGO2, FMRP, and Gemin3 were present inboth cell body and axonal fractions.

Figure 6. AGO2, FMRP, p100, and Gemin3 form a complexwithin sciatic nerves in response to application of siRNA ontoproximal sciatic nerve stump. Upper panel: immunoprecipi-tation with p100 antibodies. 5 �g of protein of each sample ofp100 fraction were incubated with specific antibodies againstp100, resolved in 5% Tris-HCl gels, and transferred to mem-branes. Membranes were probed with primary antibodiesagainst FMRP (1:1000) and AGO2 (1:1000) according to ourstandard immunoblotting laboratory procedures. p100 wasshown to coprecipitate with FMRP and AGO2 in response tosiRNA treatment while AGO2 was shown to coprecipitate withp100 and Gemin3. Middle panel: immunoprecipitation withAgo2 antibodies. Samples were precipitated with antibodiesagainst AGO2, resolved in 5% gels, transferred to mem-branes, and probed with primary antibodies against p100(1:1000) and Gemin3 (1:500). As shown, AGO2 was copre-cipitated with p100 and Gemin3 in response to siRNA treat-ment. Lower panel: double immunostaining with antibodiesagainst S-100 (1:100) to label Schwann cells and with antibod-ies against FMRP (1:100) demonstrated nonoverlapping pat-tern of expression in axonal fibers where the proximal stumpwas treated with siRNA. White arrows indicate Schwann cells.


To see whether Schwann cells may contribute to theexpression of RISC proteins, we performed doubleimmunostaining with antibodies against S-100 protein(to label Schwann cells) and with antibodies againstRISC proteins. Immunofluorescence demonstrated anonoverlapping pattern of expression of S-100 andRISC proteins. For example, FMRP, an abundant com-ponent of RISC, was detected in axonal fibers but not inSchwann cells (Fig. 6, lower panel). Taken together,these findings indicate that peripheral axons containAGO2, FMRP, p100, and Gemin3 proteins, which formmultiprotein RISC in response to siRNA treatment.Moreover, our data show that induction of RISC takesplace locally and independently from the neuronal cellbody or Schwann cells.

Uptake and transport of siRNA by sciatic nervein vivo

These experiments were performed to determinewhether siRNA can be taken up and transported withinthe peripheral mammalian nerve. We selected siRNAtargeted at luciferase, a nonmammalian protein, to use.siRNA firefly luciferase is a common control usedelsewhere (58, 59) that does not interfere with expres-sion of mammalian proteins. The sciatic nerves ofanesthetized animals were severed and anti-luciferasesiRNA labeled with FITC or CY3 was applied onto theproximal nerve stump. The mice were allowed tosurvive for 24 h. After the specified survival period, thesciatic nerves and lumbar spinal cords were removed,frozen, and sectioned on a cryostat. Localization of thesiRNA in sciatic neurons was analyzed using a confocallaser scanning microscope. Analysis of these sectionsrevealed localization of the majority of siRNAs at the tipof the severed sciatic nerve, but some fluorescence wasalso observed along fibers and distributed in the directionof the spinal cord (Fig. 7). This conclusion was strength-ened by our further observation of the uptake and trans-port of siRNA against �-tubulin and a nonspecific siRNAcontrol. siRNA against �-tubulin and control siRNA wereconjugated with FITC and applied onto severed sciaticnerves for 24 h. Both specific and control siRNAs weretaken up by sciatic nerve fibers (Fig. 7). However, nolabeling was observed in lumbar motor neurons, indicat-ing that siRNAs were not transported all the way to thespinal cord. Taken together, these results reveal thatsiRNAs can be taken up by the sciatic nerve.

Application of antitubulin siRNA decreasesexpression of �-tubulin in sciatic nerve fibers

The aim of this experiment was to determine whetherlocal treatment with siRNA could diminish the localexpression of protein in sciatic nerve fibers. The sciaticnerves of anesthetized animals were severed and asilastic tube containing siRNA against �-tubulin mRNAor nonspecific control was applied onto the proximalnerve stump. Twenty-four hours after surgery, the micewere euthanized and sciatic nerves were removed for

analysis. Expression of neuronal �-tubulin protein inthe distal part of the proximal stump of the sciaticnerve was examined by immunostaining using TUJ1, amonoclonal antibody (mAb) against neuron-specific�-tubulin protein. Tissue antigens were visualized witha DAB Substrate Kit for Peroxidase. A dramatic differ-ence in staining for �-tubulin protein was observed inexperimental vs. control sciatic nerves. Nerve stumpsincubated with antitubulin siRNA had almost no �-tu-bulin protein in the distal part of the sciatic nerveproximal stump in comparison with controls (Fig. 8).Furthermore, antitubulin siRNA-treated nerves also ex-hibited less �-tubulin protein in the distal part of thesciatic nerve proximal stump compared with the prox-imal part of the nerve. These results indicated thatsiRNA taken up into peripheral neurons inhibited localprotein synthesis.

We used immunoblot analysis to confirm this obser-vation. Proteins were extracted from proximal anddistal parts of sciatic nerve proximal stumps treatedwith antitubulin siRNA or control siRNA for 24 h,according to our experimental procedure describedabove. The proteins were separated by relative size bySDS-PAGE gels, transferred to PVDF membranes, andprobed with TUJ1 antibodies against neuron-specific�-tubulin, as well as with antibodies against �-actin. Theexperiments revealed markedly decreased levels of�-tubulin in the distal part of the sciatic nerve proximalstump (Fig. 8B), whereas expression of another cy-

Figure 7. Uptake and retrograde transport of siRNA in sciaticnerves. Upper left image: the proximal stump of a sciaticnerve was incubated for 24 h in FITC-labeled anti-luciferasesiRNA. Arrows show uptake of siRNA (green fluorescence)into sciatic nerve fibers. Upper right image: sciatic stumpincubated with FITC alone as a negative control. Lower leftimage: arrows show uptake and retrograde transport of siRNA(green fluorescence) into sciatic nerve fibers. Red indicatesimmunofluorescence for tubulin protein. Lower right image:anti-�-tubulin siRNA is indicated by green/yellow fluores-cence (arrows) in sciatic nerve fibers. Red indicates immuno-fluorescence for tubulin protein.


toskeletal protein �-actin was unchanged. ANOVA ofnormalized to �-actin mean intensity values showed asignificant reduction in �-tubulin expression aftersiRNA treatment (Fig. 8C). These results demonstratedthat application of antitubulin siRNA markedly dimin-ishes �-tubulin levels in the treated sciatic nerve. Thesedata suggest that siRNA treatment specifically inhibitsprotein synthesis in axons.

Real-time RT-PCR shows reduction of �-tubulinmRNA in sciatic nerve after treatment withantitubulin siRNA

To verify our previous finding at the mRNA level, weused RT-PCR. The goal of this experiment was toexamine levels of �-tubulin mRNA in sciatic nervestreated with antitubulin siRNA or with nonspecificsiRNA using real-time RT-PCR. Total RNA was ex-tracted from proximal and distal parts of sciatic nervestreated with antitubulin siRNA or control siRNA for24 h according to our experimental procedure de-scribed earlier. The real-time PCR reaction was carriedout in the presence of gene-specific �-tubulin forwardand reverse primers. Relative levels of �-tubulin mRNAexpression were normalized to �-actin control. Relativegene expression of �-tubulin mRNA in the siRNA-treated nerves was significantly lower (59930�13750n�9) than in sciatic nerves treated with control siRNA(113100�15190 n�9; P�0.0195) (Fig. 8D). These re-sults demonstrate that application of antitubulin siRNAmarkedly diminishes levels of �-tubulin mRNA in thetreated sciatic nerves. These data provide further evi-dence that RNAi is active in mammalian axons.

In situ hybridization analysis indicates decrease ofexpression of �-tubulin mRNA in sciatic nervestreated with antitubulin siRNA

To confirm our previous observation at the histologicallevel, in situ hybridization and immunohistochemistrywere used. The aim of this experiment was to investi-gate whether locally applied siRNA could specificallydiminish the expression of both �-tubulin mRNA andprotein in sciatic nerve fibers. The sciatic nerves ofanesthetized animals were severed and incubated withsiRNA against �-tubulin or nonspecific control. Twenty-four hours after surgery, the mice were euthanized and

Figure 8. Decrease in immunohistochemical staining against�-III tubulin protein and reduction of neuronal �-tubulinmRNA in sciatic nerve fibers treated with antitubulin siRNA.A) Immunohistochemical staining against neuronal tubulinusing TUJ1 antibodies (1:1000). Left image shows a sciaticnerve treated with nonspecific control siRNA. Right imageshows a sciatic nerve treated with siRNA targeted at �-tubulinmRNA. Black arrow indicates a decrease in tubulin immuno-reactivity in the distal end of the proximal nerve stump afterincubation with antitubulin siRNA. B) Western blot analysis ofneuronal �-III tubulin levels in sciatic nerve using TUJ1antibodies. PN, proximal part of the proximal sciatic nervestump treated with nonspecific control siRNA; PT, proximalpart of the proximal sciatic nerve stump treated with antitu-bulin siRNA; DN, distal part of the proximal sciatic nervestump treated with nonspecific control siRNA; DT, distal partof the proximal sciatic nerve stump treated with antitubulinsiRNA. Note the lower level of �-III tubulin in distal part ofthe proximal sciatic nerve stump treated with antitubulinsiRNA. Immunostaining for �-actin shows no change in

�-actin expression in response to siRNA application. C) Bargraph represents mean optical density (OD) values of neuro-nal �-III tubulin registered from autoradiographs. Data areexpressed as mean of five separate measurements normalizedto �-actin. Level of �-III tubulin was significantly lower in thedistal part of sciatic nerve stumps treated with siRNA(P0.01). D) Real-time quantitative RT-PCR of �-tubulinmRNA. Graph indicates relative change in �-tubulin geneexpression. Abbreviations are the same as in the previousgraph. Levels of �-tubulin mRNA were significantly lower inthe distal part of sciatic nerve stumps treated with siRNAwhen compared with nonspecific control siRNA (P0.05).


sciatic nerves were removed for analysis. The expres-sion of neuronal �-tubulin mRNA in sciatic nervesincubated with antitubulin siRNA and with nonspecificcontrol siRNA was analyzed using fluorescent in situhybridization. After hybridization, sections were immu-nostained with TUJ1 antibody against neuron-specific�-tubulin protein. The sections were examined using aZeiss LSM 510 confocal laser scanning microscope.Analysis of these sections revealed a dramatic decreasein the expression of �-tubulin mRNA and correspond-ing protein in experimental vs. control animals (Fig. 9).These results confirmed that application of antitubulinsiRNA onto sciatic nerve results in degradation of�-tubulin mRNA and inhibition of local protein synthe-sis of �-tubulin in sciatic nerve fibers.

Local treatment with antitubulin siRNA inhibitsretrograde labeling of lumbar motor neurons

The experiments described above demonstrated forma-tion of RNAi effector complexes and specific inhibitionof �-tubulin synthesis in sciatic nerve fibers treated withcorresponding siRNA. In the following experiment, weasked whether siRNA may affect the functionality of thesciatic nerve. A specific function of peripheral nerve isanterograde and retrograde axonal transport. Thistransport depends on the integrity of axon microtu-bules built of dimers of -tubulin and �-tubulin. Thehypothesis tested here was that the reduction in �-tu-bulin synthesis in sciatic nerve fibers treated with siRNAshould impair retrograde transport. For this experi-ment, retrograde labeling with Fluorogold was used.Mice were anesthetized, the sciatic nerve was severed atmidthigh level, and the proximal nerve stump wasinserted into a 7 mm silastic tube. �-Tubulin-specificsiRNA was placed in the tube (2 �l of 20 �M in 7 �l ofPBS). After a 6 h incubation period to allow forinitiation of RNAi machinery, a second surgery wasperformed in which the tube around the nerve wasdrained and refilled with fresh antitubulin siRNA (2 �lof 20 �M solution) and a fluorescent tracer, Fluorogold(5 �l of 7% FG). Control animals received equivalenttreatment and doses of nonspecific siRNA and Fluoro-gold. After the second surgery, the lower end of thetube was again sealed with petroleum jelly and theincision was closed. Eighteen hours after the secondsurgery, mice were euthanized and perfused. The sci-atic nerve and lumbar spinal cord were removed andsectioned on a cryostat. Serial coronal 30 �m sectionswere analyzed with fluorescent microscopy using awide-band UV filter. Fluorogold-labeled motor neuronsidentified by their size, shape, and location (Fig. 10upper panel) were counted in every section withoutallowing for split nucleoli according to a protocoldescribed elsewhere (39).

The experiments revealed that the total number ofFluorogold-labeled motoneurons in animals that re-ceived antitubulin siRNA was �4.4-fold lower than incontrol subjects (Fig. 10, lower panel). Performance ofa standard t test revealed a strong significance in these

Figure 9. Decrease of mRNA and protein levels of neuronal�-tubulin in sciatic nerve fibers after application of antitubu-lin siRNA onto the proximal stump of sciatic nerve. Fluores-cent in situ hybridization for �-tubulin mRNA and immuno-fluorescence for �-tubulin protein in sciatic nerves. Upperpanel shows sciatic nerve fibers of the proximal stump treatedwith nonspecific control siRNA. Lower panel shows sciaticnerve fibers of the proximal stump treated with antitubulinsiRNA. Green fluorescence shows the distribution of retro-gradely transported FITC-conjugated specific or nonspecificsiRNAs in sciatic nerve fibers. Red fluorescence indicatesimmunodetection of tubulin protein with TUJ1 antibodies(1:100). Blue and purple signals indicate �-tubulin mRNA.White arrows indicate the location of fluorescent signalswithin the nerve fibers. Note that the levels of �-tubulin mRNAand tubulin protein are abundant in the upper panel anddiminished in the lower panel.


results (204�26 n�4, and 912�107 n�8, respectively;P0.0001). We also showed that application of theprotein synthesis inhibitor cycloheximide onto theproximal stump of the nerve resulted in a significantdecrease in the backlabeling (490�93, n�4; P0.01),similar to siRNA treatment (Fig. 10, lower panel).

Collectively, the data showed that local inhibition of�-tubulin synthesis using either siRNA or cyclohexi-mide resulted in a decrease in retrograde labeling. Thedata indicated that application of antitubulin siRNAaffected sciatic nerve functionality locally, dramaticallydecreasing the retrolabeling of lumbar motor neuronsvia inhibition of local tubulin synthesis.

DISCUSSION

Studies of axonal growth and remodeling of peripheralnerves in response to developmental clues and injuryindicate that translation can occur outside the soma, inaxons and nerve terminals (34–38). Recent observationrevealed that several proteins and the encoding mRNAsfor the cytoskeletal proteins �-actin, �-tubulin, periph-erin, vimentin, gamma-tropomyosin 3, and cofilin-1were present in the axonal fibers (38). In addition tothe cytoskeletal elements, several heat-shock proteins(HSP27, HSP60, HSP70, grp75, alphaB crystallin), res-ident endoplasmic reticulum proteins (calreticulin,grp78/BiP, ERp29), proteins associated with neurode-generative diseases (ubiquitin C-terminal hydrolase L1,rat ortholog of human DJ-1/Park7, gamma-synuclein,SOD 1), antioxidant proteins (peroxiredoxins 1 and 6),and metabolic proteins (e.g., phosphoglycerate kinase 1(PGK 1), alpha enolase, and aldolase C/Zebrin II) werefound among the axonally synthesized proteins (38).The presence of RNAs and protein synthesis in axonsimplies the existence of a process that regulates proteinsynthesis locally. One of the potential mechanisms inregulating axonal protein synthesis may be RNAi.

In recent years, RNAi has developed as an effectivetool to specifically knock down gene expression in awide variety of cell types. Several studies have investi-gated in vivo delivery of siRNA to target diseasedneuronal genes of mammals. For example, intrathecaldelivery of siRNA targeting the pain-related cationchannel P2X3 in rats resulted in degradation of P2X3mRNA, inhibition of P2X3 protein synthesis, and di-minished neuropathic pain responses (60). In anotherstudy, mice implanted with U87 glioma cells into thecaudate-putamen nucleus received a weekly intrave-nous injection of siRNA targeting the human epider-mal growth factor receptor (EGFR), which resulted inreduced tumor expression of immunoreactive EGFRand an 88% increase in survival of the animals (61).RNAi was also shown to inhibit polyglutamine-inducedneurodegeneration caused by mutant ataxin-1 in micewith spinocerebellar ataxia type 1 after intracerebellarinjection of recombinant adeno-associated virus vectorsexpressing short hairpin RNAs (62). However, deliver-ing siRNA to target cells in vivo has been only partiallysuccessful in inhibiting target gene expression, andcurrent research efforts in this area are focused ondeveloping effective delivery systems, which will facili-tate the study of RNAi in neurons in vivo.

Although recent studies have shown that RNAi ma-chinery exists and is functional in neurons (24–27), noresearch had been performed to investigate whetherRNAi effector complexes and their components arepresent and can be activated in peripheral nerve axonalfibers. To answer this intriguing question, in the cur-rent study we investigated the hypothesis that applica-tion of siRNA to peripheral nerves will initiate RNAi. Totest our hypothesis, we used the sciatic nerve of mice toinvestigate the presence of the known (AGO2, FMRP,and p100) and suspected (Gemin3) protein compo-

Figure 10. Reduction in retrograde labeling of lumbar motorneurons with Fluorogold after application of antitubulinsiRNA onto the proximal stump of sciatic nerve. Upper panel:coronal section of the lumbar spinal cord showing Fluoro-gold-labeled motoneurons (arrows) 18 h after application ofa mixture of Fluorogold with siRNA to the proximal stump ofthe transected sciatic nerve. Top left image: sciatic nerveproximal stump treated with nonspecific control siRNA. Topright image: sciatic nerve proximal stump treated with anti-tubulin siRNA. Lower graph: total number of Fluorogold-labeled lumbar motor neurons after application of nonspe-cific siRNA, cycloheximide, or antitubulin siRNA onto sciaticnerves. C, sciatic nerves treated with control siRNA; CH,sciatic nerves treated with cycloheximide; siRNA, sciaticnerves treated with antitubulin siRNA.


nents of RISC and the ability of these proteins to formeffector complexes, the uptake and retrograde trans-port of siRNA, and the action of siRNA on proteinsynthesis and retrograde labeling.

In experiments using sciatic nerve in vivo, it isimpossible to separate axonal fibers from Schwanncells. That is why for all our biochemical and histolog-ical analyses we used �2 cm portions of sciatic nervescollected well above the site of siRNA application (�0.5cm above the upper edge of the tube with siRNA).Thus, the Schwann cells of sciatic nerve used in ouranalyses were never in physical contact with siRNAs.Therefore, all changes in the levels of mRNA and RISCproteins observed in the study were due to the directeffect of siRNA on axons. The likelihood of Schwanncells transporting RISC proteins across two membranesand/or expressing RISC in response to axonal siRNAremains far from being feasible.

Components of RISC are present and formcomplexes in peripheral nerve fibers

There is mounting evidence in support of a commonarchitecture for RISC that contains either siRNA ormiRNA and the following protein components: AGO2,FMRP, and p100 (14–17). However, miRNAs have alsobeen identified in multiprotein complexes containingeIF2C2 (eukaryotic initiation factor 2C2; the humanhomologue of AGO2), Gemin3 (a DEAD-box RNAhelicase), and Gemin4 (a protein of unknown functionthat binds with Gemin3) (25, 26, 63). Considering thatthe RNA helicase responsible for unwinding siRNAsand miRNAs has yet to be identified, we were curious toknow whether Gemin3 may be the elusive RNA helicasecomponent of RISC. Therefore, we investigated theexistence of Gemin3 in peripheral axons along with thebona fide protein components of RISC (AGO2, FMRP,and p100). We also examined whether these fourproteins assemble into RISC upon application of siRNAto peripheral nerves.

Immunoblot analysis revealed AGO2, FMRP, andp100 proteins in samples from the sciatic nerves, thusproviding evidence that the three known componentsof RISC are present in peripheral nerves. This findingwas further supported by detection of immunoreactivityin the sciatic nerve for AGO2, FMRP, and p100 pro-teins. The presence of Gemin3 in the sciatic nerve wasalso detected using immunofluorescence. Coexpres-sion of the proteins under investigation also was evi-dent. We then used immunoprecipitation to determinewhether the four proteins under investigation form acomplex in peripheral nerve fibers. For all treatmentconditions, p100 was shown to coprecipitate with FMRPand AGO2, whereas AGO2 was shown to coprecipitatewith p100 and Gemin3. As expected, expression of themultiprotein complex was more pronounced in re-sponse to treatment by siRNA, indicating that thepresence of siRNA induces RISC formation (14).

To investigate whether RNAi may act independentlyfrom the neuronal cell body, we conducted experi-

ments with ligature and colchicine treatment. Theligature was used to mechanically separate the sciaticnerve from neuronal cell bodies. Colchicine, whichcauses depolymerization of microtubules, was used toblock axonal transport without affecting the nerve’scapacity to conduct action potentials. Experimentsshowed that neither the ligature or colchicine treat-ment affected protein complexes in peripheral nerve,indicating that RISC forms locally and independentlyof the neuronal cell body. To see whether Schwanncells may contribute to RISC expression, we performeddouble immunostaining with antibodies againstSchwann cells and with antibodies against RISC pro-teins. The results showed that RISC was localized withinaxonal fibers and was not induced in surroundingSchwann cells. Taken together, these findings indicatethat peripheral axons contain AGO2, FMRP, p100, andGemin3 proteins, which form multiprotein RISC inresponse to siRNA treatment. Moreover, our data showthat formation of RISC takes place locally and indepen-dently from the neuronal cell body or Schwann cells.

RISC proteins are present in axonal preparationsfrom DRG-dissociated cultures

To further confirm our findings, we used a uniquemethod allowing preparation of axonal protein frac-tion from DRG-dissociated cultures (37). This methodwas successfully used previously to characterize expres-sion of several proteins and their encoding mRNAs inaxonal fibers (38). Dissociated DRGs are plated intospecial tissue culture inserts containing a porous mem-brane that allows passage of axons but restricts the cellbody and non-neuronal cells to the upper membranesurface. Thus, axons and bodies are isolated from theopposite surfaces of the membrane and subsequentlyused for protein isolation and Western blot analysis.Our analysis showed that RISC proteins p100, AGO2,FMRP, and Gemin3 are abundant in both the cell bodyand axonal fractions.

Moreover, immunofluorescence performed on disso-ciated DRG neurons unequivocally confirmed our find-ing that key proteins involved in the RNAi pathway arepresent in axons. Recently, a new study also reportedpresence of AGO3, AGO4, and Dicer in developingDRG axons and growth cones in culture (64). Thisfurther implicates the RNAi pathway as a regulatorymechanism for local protein expression in axons.

Uptake and retrograde transport of siRNAs by thesciatic nerve

Spinal motor neurons and sensory neurons in DRGextend their axons over long distances; therefore, theefficiency of axonal transport in both anterograde andretrograde directions is a crucial parameter in neuronalfunction. Retrograde axonal transport of target-derivedneurotrophic factors is critical for survival, axongrowth, and path finding during development andregeneration. Neurotrophins, such as nerve growth


factor (NGF), brain-derived neurotrophic factor(BDNF), and neurotrophin-3 (NT3), are delivered tothe perikarya by retrograde axonal transport by recep-tor-mediated internalization (65–69). In our experi-ments, we observed the uptake of specific and nonspe-cific fluorescently labeled siRNAs by the sciatic nerve.While the mechanism of siRNA uptake and deliveryremains to be elucidated, further investigation ofsiRNA retrograde delivery along the sciatic nerve mayprovide insight into targeted retrograde delivery ofsiRNA for the inhibition of protein expression inspecific populations of neurons.

Action of siRNA on protein synthesis andretrograde labeling

While recent studies have shown that RNAi machineryexists in neurons (24–27) and dendrites (31), noresearch had investigated whether siRNAs applied toperipheral nerve axons can affect local protein synthe-sis. Our experiments showed that application of antitu-bulin siRNA onto the proximal sciatic nerve stumpsignificantly reduced the level of �-tubulin mRNA,resulting in markedly diminished synthesis of �-tubulinprotein, whereas the ligatures or colchicine did notinterrupt formation of RISC. These data were con-firmed using immunohistochemistry, in situ hybridiza-tion, and real-time RT-PCR. Furthermore, antitubulinsiRNA was shown to sufficiently impair microtubulestructure and function, as evidenced by a significantlyreduced number of retrogradedly labeled motor neu-rons (Fig. 11).

Our data provide the first direct evidence that theRNAi machinery exists and is functional in peripheralmammalian axons and is independent from the neuro-nal cell body or Schwann cells. We hypothesize thatRNAi in peripheral nerves may be a mechanism forregulating local axonal protein synthesis. Further stud-ies should provide important insight into the RNAipathway that regulates local protein synthesis in axons.The ability to specifically modulate expression of ax-

onal proteins using RNAi may have significant clinicalpotential in the treatment of neurodegenerative dis-eases, specifically those involving peripheral nerves.

We are grateful to Dr. Jeffery Twiss for helpful comments andsuggestions and to Dr. Dianna Willis for sharing DRG cultureprotocol. We would like to acknowledge the following scientistsfor their kind donation of antibodies for use in our research: Dr.Tom Hobman (University of Alberta, Canada; rabbit polyclonalanti-AGO2 antibody), Dr. Tom Keenan (Virginia PolytechnicInstitute and State University, Blacksburg, VA, USA; guinea pigpolyclonal anti-p100 antibody), and Dr. Gideon Dreyfuss (Uni-versity of Pennsylvania, Philadelphia, PA, USA; anti-Gemin3antibody). We also thank Dr. Douglas Weidner (East CarolinaUniversity, Greenville, NC, USA) for assistance with confocalmicroscopy. This work was partially supported by Rosnaukacontract #02.442.11.7319 (R.R.I).

REFERENCES

1. Ambros, V. (2003) MicroRNA pathways in flies and worms:growth, death, fat, stress, and timing. Cell 113, 673–676

2. Cerutti, H. (2003) RNA interference: traveling in the cell andgaining functions? Trends Genet. 19, 39–46

3. Hannon, G. J. (2002) RNA interference. Nature 418, 244–2514. Scherr, M., and Eder, M. (2004) RNAi in functional genomics.

Curr. Opin. Mol. Ther. 6, 129–1355. Shi, Y. (2003) Mammalian RNAi for the masses. Trends Genet. 19,

9–126. McManus, M. T., and Sharp, P. A. (2002) Gene silencing in

mammals by small interfering RNAs. Nat. Rev. Genet. 3, 737–7477. Carrington, J. C., and Ambros, V. (2003) Role of microRNAs in

plant and animal development. Science 301, 336–3388. Wienholds, E., Koudijs, M. J., van Eeden, F. J., Cuppen, E., and

Plasterk, R. H. (2003) The microRNA-producing enzyme Dicer1is essential for zebrafish development. Nat. Genet. 35, 217–218

9. Bernstein, E., Caudy, A. A., Hammond, S. M., and Hannon, G. J.(2001) Role for a bidentate ribonuclease in the initiation step ofRNA interference. Nature 409, 363–366

10. Lee, Y., Ahn, C., Han, J., Choi, H., Kim, J., Yim, J., Lee, J.,Provost, P., Radmark, O., Kim, S., and Kim, V. N. (2003) Thenuclear RNase III Drosha initiates microRNA processing. Nature425, 415–419

11. Hammond, S. M., Bernstein, E., Beach, D., and Hannon, G. J.(2000) An RNA-directed nuclease mediates post-transcriptionalgene silencing in Drosophila cells. Nature 404, 293–296

Figure 11. Schematic representation of the pro-posed local RNAi mechanism in axonal fibers.The schematic shows that application of antitu-bulin siRNA onto the sciatic nerve induces thecascade of events including RISC formation,degradation of �-tubulin mRNA, inhibition of�-tubulin protein synthesis, and impairment ofretrograde axonal transport.


12. Hammond, S. M., Boettcher, S., Caudy, A. A., Kobayashi, R., andHannon, G. J. (2001) Argonaute2, a link between genetic andbiochemical analyses of RNAi. Science 293, 1146–1150

13. Sharp, P. A., and Zamore, P. D. (2000) Molecular biology. RNAinterference. Science 287, 2431–2433

14. Caudy, A. A., Ketting, R. F., Hammond, S. M., Denli, A. M.,Bathoorn, A. M., Tops, B. B., Silva, J. M., Myers, M. M., Hannon,G. J., and Plasterk, R. H. (2003) A micrococcal nucleasehomologue in RNAi effector complexes. Nature 425, 411–414

15. Doench, J. G., Petersen, C. P., and Sharp, P. A. (2003) siRNAscan function as miRNAs. Genes Dev. 17, 438–442

16. Hutvagner, G., and Zamore, P. D. (2002) A microRNA in amultiple-turnover RNAi enzyme complex. Science 297, 2056–2060

17. Zeng, Y., Yi, R., and Cullen, B. R. (2003) MicroRNAs and smallinterfering RNAs can inhibit mRNA expression by similarmechanisms. Proc. Natl. Acad. Sci. U. S. A. 100, 9779–9784

18. Tijsterman, M., and Plasterk, R. H. (2004) Dicers at RISC; themechanism of RNAi. Cell 117, 1–3

19. Reinhart, B. J., Slack, F. J., Basson, M., Pasquinelli, A. E.,Bettinger, J. C., Rougvie, A. E., Horvitz, H. R., and Ruvkun, G.(2000) The 21-nucleotide let-7 RNA regulates developmentaltiming in Caenorhabditis elegans. Nature 403, 901–906

20. Chen, J. F., Mandel, E. M., Thomson, J. M., Wu, Q., Callis, T. E.,Hammond, S. M., Conlon, F. L., and Wang, D. Z. (2005) Therole of microRNA-1 and microRNA-133 in skeletal muscleproliferation and differentiation. Nat. Genet. 38, 228–233

21. Brennecke, J., Hipfner, D. R., Stark, A., Russell, R. B., andCohen, S. M. (2003) bantam encodes a developmentallyregulated microRNA that controls cell proliferation andregulates the proapoptotic gene hid in Drosophila. Cell 113,25–36

22. Kim, J., Krichevsky, A., Grad, Y., Hayes, G. D., Kosik, K. S.,Church, G. M., and Ruvkun, G. (2004) Identification of manymicroRNAs that copurify with polyribosomes in mammalianneurons. Proc. Natl. Acad. Sci. U. S. A. 101, 360–365

23. Krichevsky, A. M., King, K. S., Donahue, C. P., Khrapko, K., andKosik, K. S. (2003) A microRNA array reveals extensive regula-tion of microRNAs during brain development. RNA 9, 1274–1281

24. Caudy, A. A., Myers, M., Hannon, G. J., and Hammond, S. M.(2002) Fragile X-related protein and VIG associate with theRNA interference machinery. Genes Dev. 16, 2491–2496

25. Nelson, P. T., Hatzigeorgiou, A. G., and Mourelatos, Z. (2004)miRNP:mRNA association in polyribosomes in a human neuro-nal cell line. RNA 10, 387–394

26. Dostie, J., Mourelatos, Z., Yang, M., Sharma, A., and Dreyfuss, G.(2003) Numerous microRNPs in neuronal cells containingnovel microRNAs. RNA 9, 180–186

27. Krichevsky, A. M., and Kosik, K. S. (2002) RNAi functions incultured mammalian neurons. Proc. Natl. Acad. Sci. U. S. A. 99,11926–11929

28. Giraldez, A. J., Cinalli, R. M., Glasner, M. E., Enright, A. J.,Thomson, J. M., Baskerville, S., Hammond, S. M., Bartel, D. P.,and Schier, A. F. (2005) MicroRNAs regulate brain morphogen-esis in zebrafish. Science 308, 833–838

29. Miska, E. A., Alvarez-Saavedra, E., Townsend, M., Yoshii, A.,Sestan, N., Rakic, P., Constantine-Paton, M., and Horvitz, H. R.(2004) Microarray analysis of microRNA expression in thedeveloping mammalian brain. Genome Biol. 5, R68

30. Sempere, L. F., Freemantle, S., Pitha-Rowe, I., Moss, E.,Dmitrovsky, E., and Ambros, V. (2004) Expression profiling ofmammalian microRNAs uncovers a subset of brain-expressedmicroRNAs with possible roles in murine and human neuronaldifferentiation. Genome Biol. 5, R13

31. Schratt, G. M., Tuebing, F., Nigh, E. A., Kane, C. G., Sabatini,M. E., Kiebler, M., and Greenberg, M. E. (2006) A brain-specificmicroRNA regulates dendritic spine development. Nature 439,283–289

32. Eberwine, J., Belt, B., Kacharmina, J. E., and Miyashiro, K.(2002) Analysis of subcellularly localized mRNAs using in situhybridization, mRNA amplification, and expression profiling.Neurochem. Res. 27, 1065–1077

33. Kiebler, M. A., and DesGroseillers, L. (2000) Molecular insightsinto mRNA transport and local translation in the mammaliannervous system. Neuron 25, 19–28

34. Brittis, P. A., Lu, Q., and Flanagan, J. G. (2002) Axonal proteinsynthesis provides a mechanism for localized regulation at anintermediate target. Cell 110, 223–235

35. Giuditta, A., Kaplan, B. B., van Minnen, J., Alvarez, J., andKoenig, E. (2002) Axonal and presynaptic protein synthesis:new insights into the biology of the neuron. Trends Neurosci. 25,400–404

36. Lee, S. K., and Hollenbeck, P. J. (2003) Organization andtranslation of mRNA in sympathetic axons. J. Cell Sci. 116,4467–4478

37. Zheng, J. Q., Kelly, T. K., Chang, B., Ryazantsev, S., Rajasekaran,A. K., Martin, K. C., and Twiss, J. L. (2001) A functional role forintra-axonal protein synthesis during axonal regeneration fromadult sensory neurons. J. Neurosci. 21, 9291–9303

38. Willis, D., Li, K. W., Zheng, J. Q., Chang, J. H., Smit, A., Kelly, T.,Merianda, T. T., Sylvester, J., van Minnen, J., and Twiss, J. L.(2005) Differential transport and local translation of cytoskel-etal, injury-response, and neurodegeneration protein mRNAs inaxons. J. Neurosci. 25, 778–791

39. Murashov, A. K., Islamov, R. R., McMurray, R. J., Pak, E. S., andWeidner, D. A. (2004) Estrogen increases retrograde labeling ofmotoneurons: evidence of a nongenomic mechanism. Am. J.Physiol. 287, C320–C326

40. Wagey, R., Lurot, S., Perrelet, D., Pelech, S. L., Sagot, Y., andKrieger, C. (2001) Phosphatidylinositol 3-kinase activity in mu-rine motoneuron disease: the progressive motor neuropathymouse. Neuroscience 103, 257–266

41. McMurray, R., Islamov, R., Murashov, A. K. (2003) Raloxifeneanalog LY117018 enhances the regeneration of sciatic nerve inovariectomized female mice. Brain Res. 980, 140–145

42. Murashov, A. K., Ul Haq, I., Hill, C., Park, E., Smith, M., Wang,X., Goldberg, D. J., and Wolgemuth, D. J. (2001) Crosstalkbetween p38, Hsp25 and Akt in spinal motor neurons aftersciatic nerve injury. Brain Res. Mol. Brain Res. 93, 199–208

43. Islamov, R. R., Chintalgattu, V., McMurray, R. J., Pak, E. S.,Murashov, A. K., and Katwa, L. C. (2003) Differential expressionof endothelin receptors in regenerating spinal motor neuronsin mice. Brain Res. Mol. Brain Res. 116, 163–167

44. Islamov, R. R., Chintalgattu, V., Pak, E. S., Katwa, L. C., andMurashov, A. K. (2004) Induction of VEGF and its Flt-1 receptorafter sciatic nerve crush injury. NeuroReport 15, 2117–2121

45. Chintalgattu, V., Nair, D. M., and Katwa, L. C. (2003) Cardiacmyofibroblasts: a novel source of vascular endothelial growthfactor (VEGF) and its receptors Flt-1 and KDR. J. Mol. CellCardiol. 35, 277–286

46. Sharifi, N., Reuss, A. E., and Wray, S. (2002) Prenatal LHRHneurons in nasal explant cultures express estrogen receptorbeta transcript. Endocrinology 143, 2503–2507

47. Pfaffl, M. W. (2001) A new mathematical model for relativequantification in real-time RT-PCR. Nucleic Acids Res. 29, e45

48. Murashov, A. K., and Wolgemuth, D. J. (1996) Sense andantisense transcripts of the developmentally regulated murinehsp70.2 gene are expressed in distinct and only partially over-lapping areas in the adult brain. Brain Res. Mol. Brain Res. 37,85–95

49. Siomi, M. C., Zhang, Y., Siomi, H., and Dreyfuss, G. (1996)Specific sequences in the fragile X syndrome protein FMR1 andthe FXR proteins mediate their binding to 60S ribosomalsubunits and the interactions among them. Mol. Cell Biol. 16,3825–3832

50. Murashov, A. K., Talebian, S., and Wolgemuth, D. J. (1998) Roleof heat shock protein Hsp25 in the response of the orofacialnuclei motor system to physiological stress. Brain Res. Mol. BrainRes. 63, 14–24

51. Yu, J. Y., DeRuiter, S. L., and Turner, D. L. (2002) RNAinterference by expression of short-interfering RNAs and hair-pin RNAs in mammalian cells. Proc. Natl. Acad. Sci. U. S. A. 99,6047–6052

52. Keenan, T. W., Winter, S., Rackwitz, H. R., and Heid, H. W.(2000) Nuclear coactivator protein p100 is present in endoplas-mic reticulum and lipid droplets of milk secreting cells. Biochim.Biophys. Acta 1523, 84–90

53. Brown, V., Jin, P., Ceman, S., Darnell, J. C., O’Donnell, W. T.,Tenenbaum, S. A., Jin, X., Feng, Y., Wilkinson, K. D., Keene,J. D., Darnell, R. B., and Warren, S. T. (2001) Microarrayidentification of FMRP-associated brain mRNAs and altered


mRNA translational profiles in fragile X syndrome. Cell 107,477–487

54. Gubitz, A. K., Mourelatos, Z., Abel, L., Rappsilber, J., Mann, M.,and Dreyfuss, G. (2002) Gemin5, a novel WD repeat proteincomponent of the SMN complex that binds Sm proteins. J. Biol.Chem. 277, 5631–5636

55. Pushkin, S., Cubits, A. K., Massenet, S., and Dreyfuss, G. (2002)The SMN complex, an assemblyosome of ribonucleoproteins.Curr. Opin. Cell Biol. 14, 305–312

56. Shen, Y., Mani, S., Donovan, S. L., Schwob, J. E., and Meiri, K. F.(2002) Growth-associated protein-43 is required for commis-sural axon guidance in the developing vertebrate nervoussystem. J. Neurosci. 22, 239–247

57. Mader, K., Andermahr, J., Angelov, D. N., and Neiss, W. F.(2004) Dual mode of signalling of the axotomy reaction:retrograde electric stimulation or block of retrograde transportdifferently mimic the reaction of motoneurons to nerve tran-section in the rat brainstem. J. Neurotrauma 21, 956–968

58. Guo, S., Tschammer, N., Mohammed, S., and Guo, P. (2005)Specific delivery of therapeutic RNAs to cancer cells via thedimerization mechanism of phi29 motor pRNA. Hum. Gene.Ther. 16, 1097–1109

59. Soutschek, J., Akinc, A., Bramlage, B., Charisse, K., Constien, R.,Donoghue, M., Elbashir, S., Geick, A., Hadwiger, P., Harborth,J., et al. (2004) Therapeutic silencing of an endogenous gene bysystemic administration of modified siRNAs. Nature 432, 173–178

60. Dorn, G., Patel, S., Wotherspoon, G., Hemmings-Mieszczak, M.,Barclay, J., Natt, F. J., Martin, P., Bevan, S., Fox, A., Ganju, P., etal. (2004) siRNA relieves chronic neuropathic pain. Nucleic AcidsRes. 32, e49

61. Zhang, Y., Zhang, Y. F., Bryant, J., Charles, A., Boado, R. J., andPardridge, W. M. (2004) Intravenous RNA interference gene

therapy targeting the human epidermal growth factor receptorprolongs survival in intracranial brain cancer. Clin. Cancer Res.10, 3667–3677

62. Xia, H., Mao, Q., Eliason, S. L., Harper, S. Q., Martins, I. H.,Orr, H. T., Paulson, H. L., Yang, L., Kotin, R. M., and Davidson,B. L. (2004) RNAi suppresses polyglutamine-induced neurode-generation in a model of spinocerebellar ataxia. Nat. Med. 10,816–820

63. Mourelatos, Z., Dostie, J., Paushkin, S., Sharma, A., Charroux,B., Abel, L., Rappsilber, J., Mann, M., and Dreyfuss, G. (2002)miRNPs: a novel class of ribonucleoproteins containing numer-ous microRNAs. Genes Dev. 16, 720–728

64. Hengst, U., Cox, L. J., Macosko, E. Z., and Jaffrey, S. R. (2006)Functional and selective RNA interference in developing axonsand growth cones. J. Neurosci. 26, 5727–5232

65. Markus, A., Patel, T. D., and Snider, W. D. (2002) Neurotrophicfactors and axonal growth. Curr. Opin. Neurobiol. 12, 523–531

66. Ginty, D. D., and Segal, R. A. (2002) Retrograde neurotrophinsignaling: Trk-ing along the axon. Curr. Opin. Neurobiol. 12,268–274

67. Kuruvilla, R., Ye, H., and Ginty, D. D. (2000) Spatially andfunctionally distinct roles of the PI3-K effector pathway duringNGF signaling in sympathetic neurons. Neuron 27, 499–512

68. Delcroix, J. D., Valletta, J. S., Wu, C., Hunt, S. J., Kowal, A. S.,and Mobley, W. C. (2003) NGF signaling in sensory neurons:evidence that early endosomes carry NGF retrograde signals.Neuron 39, 69–84

69. Ye, H., Kuruvilla, R., Zweifel, L. S., and Ginty, D. D. (2003)Evidence in support of signaling endosome-based retrogradesurvival of sympathetic neurons. Neuron 39, 57–68

Received for publication April 12, 2006.Accepted for publication September 25, 2006.


RNAi pathway is functional in peripheral nerve axons

Documents

Transcript of RNAi pathway is functional in peripheral nerve axons