Dottie and Simons Neuronal and Epithelial Comp1

7/28/2019 Dottie and Simons Neuronal and Epithelial Comp1

1/10

Cell, Vol. 62, 63-72, July 13, 1990, Copyright 0 1990 by Cell Press

Polarized Sorting of Viral Glycoproteins to theAxon and Dendrites of Hippocampal Neuronsin CultureCarlos G. Dotti and Kai SimonsCell Biology ProgramEuropean M olecular B iology LaboratoryPostfach 10.2209Heidelberg, D-6900Federal Republ ic of Germany

SummaryCultured hippocampa l neurons were infected with atemperature-sensitive mutant of vesicular stomati t isvirus (VSV) and a wild-type strain of the avian influ-enza fowl plague virus (FPV). The intracellular distri-bution of viral glycoproteins was monitored by immu-nofluorescence microscopy. In mature, ful ly polarizedneurons the VSV glycoprotein (a basolateral protein inepithel ial MDC K cel ls) moved from the Golgi complexto the dendritic domain, whereas the hemagglutininprotein of FPV (an apically sorted protein in MD CKcells) was targeted preferentially, but not exclu sively,to the axon. The VSV glycoprotein appeared in cluste rson the dendritic surfac e, w hile the hemagglutinin wa sdistributed uniformly along the axonal mem brane.Based on the finding that the sam e viral glyc oproteinsare sorted in a polarized fashion in both neuronal andepithelial cells, we propose that the molecular mech a-nisms of surface protein sorting share commo n fea-tures in the two cel l types.IntroductionA neuronal cell has two different type s of neurites: a singleaxon and several dendri tes. The axon typical ly transmitssignals, whereas the dendrites generally receive them. Toperform these different functions, axons and dendriteshave to contain different membrane proteins (G. A. Bankerand J. E. Mazurkiewicz, Sot. Neurosci. abstract, 1982;Angelides et al., 1988; Jones et al., 1989). How the neurontargets proteins to their spec ific dendritic and axonal loca-tions is an interesting neurobiological problem that so farhas attracted little attention (see Kelly, 1988; Black andBaas, 1989). N eurons are not the only cells that polarizetheir cel l surfaces and have to target surface proteins todifferent plasm a mem brane doma ins. Epithelial cells arealso polarized into distinct apical and basolateral cellsurface domains (Simons and Fuller, 1985; Rodriguez-Boulan and Nelson, 1989). These two membrane do-mains are segregated by tight junctions that function asa fence to block lateral diffus ion of lipids (o f the exoplas-mic leaflet of the bilayer) and proteins between the apicaland basolateral sides of the plasma membrane (Gum-biner, 1987).

Although surface polarities in neurons and epitheliahave little in com mon functionally and morphologica lly,the molecular mechanism s of surface protein sorting mayshare com mon features . Stud ies on epithelial protein sort-

ing have demonstrated that two different scenarios areused. In Madin-Darby canine kidney (MDCK) cel ls, segre-gation of newly synthesized glycoproteins takes place inthe trans-Golgi network (TGN) (Griff iths and Simons,1986; Hughson et al., 1988). Apical and basolateral pro-teins are sorted into two different vesicular carriers, onefor the apical surface and another for the basolateral one(Bennett et al., 1988). Sorting in the apical direction is spe-cific, and little misso rting of basolateral proteins is ob-served . Apical proteins, on the other hand , ma y be deliv-ered in the wrong direction to the basolateral side, andthis has led to the postulation of the basolateral route asa default pathway, analogous to what has been proposedfor the consti tutive biosynthetic pathway from the TGN tothe plasma membrane in fibroblasts (Pfeffer and Roth-man, 1987; Wandinger-Ness and Simons, 1990). In hepa-tocytes , sorting is accomplished in a different way (Bartleset al., 1987). In these ep ithelial cells, seg regation of thesurface glycoproteins is delayed, taking place at the plas-ma membrane. Both newly synthesized apical and baso-lateral p roteins seem to travel together from the Golgicomplex to the basolateral domain. Sorting occurs afterendocytosis into endosomes. The apical proteins are de-l ivered by vesicular carr iers across the cel l by transcytosisto the apical side.

To explore w hether surfac e protein so rting in neuronsfollows any of the epithelial mod els, we have use d in thisstud y viral glycoproteins as tools to investigate proteintransport to the neuronal cell surfac e. S everal envelopedRNA viruses turn cel ls into virus factories after infection.Ho st cell protein s ynthe sis is inhibited; instead, v irus pro-teins are translated to assemble new virus particles. Thesurface glycoproteins of Semliki Forest virus, vesicularstomati t is virus (VSV), and influenza virus are assembledin the endoplasmic reticulum (ER), transported to theGolgi complex, and del ivered from the TGN to the cel l sur-face to direct budding of the nucleocapsid from theplasma m embrane (Simons and Warren, 1984; Dubois-Dalcq et al., 1984). These viral glycoproteins have beenused as probes to study biosynthetic protein traffic to thecell surface in both cultured fibroblasts and epithelialcel ls. In MD CK cel ls, the VSV glycoprotein is basolateral,whereas the influenza hemagglutinin (HA) is apical (Rod-riguez-Boulan and Pendergas t, 1980). In neurobiology,enveloped viruses have been mainly used as transneuro-nal tracers (Kuypers and Ugolini, 1990) and as tools forneuropathological analysis (Dubois-D alcq et al., 1982).

The question that we have asked in this study is whetherneuronal cells in culture infected with VSV or influenza vi-rus distribute the virus glycoproteins in a polarized fash-ion. For this purpose, we have used rat hippocampalneurons in culture. T hese cells undergo a well-defined se-quence of event s from a nonpolarized to a polarized stage(Dotti et al., 1988). After initially extending lamellipodia(developmen tal stage l), the cells send out short neurites(stage 2). After a period of 6-24 hr, the cells st art t o be-com e morphologically polarized. One of the cellular pro-


2/10

Cd l64

cesses begins to grow rapidly, and when i t has grown 10pm longer than any of the others (Goslin and Banke r,1989), this process is destined to become the final axon(stage 3). A few days later, the remaining neurites also be-gin to grow to form the dendrites (stage 4). By lo-14 daysin culture the axon and the dendrites have acquired all thefunctional and molecular properties that characte rize amature neuron (stage 5) (reviewed in Banker and Wax -man, 1988). Synaptogenesis starts when dendri tes de-velop, and during stage 5 extensive synapse formationhas occurred. Many of the axons run along the length ofthe dendri tes, forming numerous synapses en passant(Rothman and Cowan, 1981; Barlett and Banker, 1984b).Furthermore, morphological, functional, and immunocy-tochem ical criteria have been defined to differentiate theaxons from the dendri tes (G. A. Banker and J. E. Mazur-kiewicz, So t. Neurosci. abstract, 1982; Caceres et al .,1984,1986; Barlett and Banker, 1964a, 1984b; Shaw et al.,1985; Dot ti et al., 1987; Goslin et al., 1988; Banker andWaxm an, 1968; Baas et al ., 1988).

In this study, we infected the hippocampal neurons atstage s 3 and 5 with VSV and with influenza virus andstudied the distribution of the virus glycoproteins by im-munofluorescence microscopy. At stage 3, when the axonhad been extend ed, the VSV and the influenza glycopro-teins were delivered to the entire ce ll surfac e, includingthe short neuri tes and the axon. After synaptogenesis hadoccurred, the VSV glycoprotein was exclusively routed tothe dendrite and the cell body, but no axonal delivery wasobserved. In contrast, the influenza HA was sorted prefer-ential ly, al though not exclusively, to the axon. These find-ings pave the way for the use of viral glycoproteins asprobes fo r studying polarized sorting in neurons.ResultsNewly Synthesized VSV Glycoprotein Is Del iveredExclusively to the Dendri t ic DomainIn the fi rst series of experiments, we analyzed the intracel-lular distribution of the VSV glycoprotein in cultured hip-pocampal neurons (stage 5) infected w ith the tempera-ture-sensitive mutant VSV ts045. In cel ls infected with thisvirus the transport of the VSV glycoprotein to the cell sur-face is critically dependent on temperature (Bergmannet al., 1981; Bergmann and Singer, 1983; Griffiths et al.,1985). At the nonperm issive temperature of 39OC , viral in-fection occu rs, but the VSV glycoprotein is arrested in theER. Changing the temperature to 20C permits the trans-port of the glycoprotein to the Golgi co mplex where it ac-cumulates in the TGN . Synchronous transport from theGolgi to the cel l surface occurs after switching to the per-missive temperature of 32C.

When VSV t&45-infected hippocampal neurons werefixed in methanol after 3 .5 hr at 39OC and 90 min at 20Cand labeled with the anti-VSV glycoprotein antiserum ,positive imm unoreactivi ty was limited to the perinuclearregion (Figures lA-1C). The labeling had a reticular ap-pearance, resembling the staining of the Golgi com plexobtained with whea t germ agglutinin (data not show n).This pattern of staining was observed in 95% of all neu-

rons. In approximately 20% of the infected cel ls, labeledGolgi-like eleme nts extended bey ond the perinuclear re-gion into one of the cells dendrites (data no t show n). Thisfinding suppo rts prev ious data sugges ting the presenceof Golgi eleme nts in dendrites (De Cam illi e t al., 1986).Axons and dendrites were distinguished following mor-phological and immuno logical criteria. Morphologica lly,axons are very long, thin, and untapering. Dendrites areshort, typically no longer than 100-200 pm , thick , andtapering (Rothman and Cowan, 1981; Barlett and Banker,1984b). Immunological ly, dendrites but not axons reactpositively with anti-microtubule-as sociated protein 2 (anti-MAP2) antibodies (Cace res et al., 1984, 1986). After 3.5 hrat 39C and 90 min at 20C using the anti-MAP2 antibod-ies in the VSV ts045-infec ted cells, neither the thin anduntapering, MAPSnegative processes (axons) nor the moredistal segments of the MAP2 positively labeled processes(dendrites) appeared stained.

We next analyzed the intracellular distribution of the VSVglycoprotein at different time s after releasing the cellsfrom the 20C block (Figures lD-1F). After 1 hr at the per-missive temperature of 32OC, VSV glycoprotein was sti l lseen in the cell body, both intracellularly and at the cellsurface (see below). In some neuronal processes, VSVglycoprotein appeared in the form of t iny dots. This dottypattern appeared along the entire length of the neurites.Not al l of the numerous processes evident in the phase-contrast micrographs (Figure 1D) showed VSV glycopro-tein labeling (Figure 1F). By double labeling the infectedcells with an anti-MAP2 antiserum , the VSV glycoprotein-labeled n eurites could be identified as dendrites (com-pare labeled proces ses in Figures 1E and 1F). Quanti-tation of how many MAPP-positive processes had VSVglycoprotein imm unorea ctivity revealed that all the den-dri tes of infected cel ls were positive. Axons were usual lydevoid of VSV glycoprotein immunoreactivi ty (compareFigure 1D [phase-contrast] with Figure 1F). Of 615 VSVglycoprotein-positive processes, only 29 were MAP2 neg-ative. The absence of axonal labeling could n ot be attrib-uted to a difference in the thickness of the neuri tes be-cause numerous thick processes (bundles of axons) werenegative. We also ruled out that the lack of axonal labelingwas due to a delay in the transport of newly synthesizedVSV glycoprotein to this territory, by fixing ce lls at pro-longed times postinfection. Even after 7 hr at permissivetempe rature, axonal labeling could not be detected (datanot shown).In our next series of experiments we analyzed the inser-tion into the plasma membrane of the newly synthesizedVSV glycoprotein (Figure 2). For this purpose two differentprotocols were used. Infected cel ls were incubated in thepresence of anti-VSV glycoprotein antibodies at 4OC for 20min and then paraformaldehyde fixed. Alternatively, in-fected cells were fixed directly in 4% paraformaldehydeand 0.25% glutaraldehyde followed by incubation with pri-mary antibodies. With both methods similar results wereobtained. Surface labeling was in the form of small c lus-ters of immunoreactivi ty (Figures 2B-2C). The labelingwas restr icted to the cell body and dendrites, as judgedby MAP 2-positive labeling (Figure 2A ). Mo st (91%) of the


3/10

Sorting of Viral Glycoproteins to Axon and Dendrites65

Figure 1. Intracellu lar Distribution of the VS VNO45 Glycoprotein in Infected Mature (Stage 5)Hippocampal NeuronsCells w ere fixed in cold me thanol after a 1 hrinfection at 32% followed by 2.5 hr at 39% andthen incubated for 90 min at 20% (A-C) or afteran additional 60 min at the permissive temper-ature of 32% (D-F). Axons (arrows) and den-drites (arrowheads) are differe ntiated by theirmorph ology, phase-contrast micrographs (Aand D), and MAP 2 immunoreactiv ity (only den-drites are positive) (B and E). Incubation withthe anti-VSV glycoprotein antiserum revealedaccumulation of G protein in the ceil body atthe end of the 20% block (C) and in the cellbody and dendrites at the end of the 32% incu-bation (F). Axons are devoid of labeling. Bar:15 urn.

Figure 2. Distribution of the VS V ts045 Glyco-protein in Infected Mature (Stage 5) Hippocam-pal Neurons(A-C) After 1 hr at permissive temperature(conditions as in Figures lD-1F) cells werefixed in 4% paraformaldehyde-0.25% glutar-aldehyde and incubated with the anti-VS V gly-coprotein antiserum. Anti-MAP2 antiserum wasadded after permeabilization with Triton X-100.Dendrites and cell body are labeled with theanti-MAP 2 antibodies. VS V glycoprotein is ex-clusively present on the surface of a ll the MAP P-labeled processes in the form of small dots (ar-rowheads) (B). The punctate stain ing is seenalso on the cell body surface of the same cell63(D-F) Double labeli ng with anti-VSV glycopro-tein and anti-synaptophysin antibodies ofts045-infected cells. Comparison of the anti-VS V glycoprotein (E) and anti-synaptophysin(F) antibody labeling shows partia l colocaliza-tion of both proteins (arrows). Arrowhe ads in(E) show expression of VS V glycoprotein on thedendrit ic and cell body surface at s ites not la-beled with synaptophysin.

Bar: 10 urn.


4/10

Cd l66

Figure 3. Intracellu lar Distribution of the HAProtein of FPV-Infected Stage 5 NeuronsCells were fixed in cold methano l 4 hr (A-C) or5 hr (D-F) after infection, incubated with ananti-HA antiserum, and processed for immu-nofluoreescence. An anti-MAP 2 antiserum wasused to distinguish axons from dendrites (Band E). Four hours postinfection the HA pro-tein appears restricted to the cell body (C).Five hours postinfection all axons show intenseHA labeling in the form of small dots and vari-cosities (arrowheads). Dendrite s d 3, d4, d5,and d6 are clearly devoid of anti-HA labeling.Some l abeling appears restricted to the in it ia lsegment in dendrites d2 and dl (arrows). Bar:12 pm.

dendrites were positive for VSV glycoprotein. VSV glyco-protein insertion at the plasma membrane was detectedat 30-60 min after the temperature shift to 3PC. Axonswere devoid of anti-VSV glycoprotein immunoreactivi ty.Our finding of VSV glycoprotein clustering on the dendriticsurface is different from previous findings of uniform VSVglycoprotein distribution on the surface of VSV t&45+-fected fibroblasts and ATt20 cel ls (Bergmann et al ., 1983;Riva s and Moore, 1989). The nonneuronal cells (mos tlyastroc ytes ) prese nt in the hippocampal neuronal c ulturealso showed uniform surface appearance of the VSV gly-coprotein (data not shown). The observed clustering ofVSV glycoprotein on the dendritic surface and the lack ofaxonal surface staining did not change with longer infec-tion times.

Since dendri tes have special ized sites where synapticcontacts occur, we next investigated whether VSV glyco-protein clustering occurred at those sites. Infected cel lswere double labeled w ith anti-synap tophysin , a syna pticvesicle mark er, and anti-VSV glyc oprotein antibodies (Fig-ures 2D-2F ). Syn aptophys in is an integral membrane pro-tein of synaptic vesicles (Jahn et al ., 1985; Wiedenmannand Franke , 1985). In cultured hippocampal neurons itwas demonstrated to accumulate in presynaptic terminals(T. Fletcher, l? Cameron, l? De Camil li , and G. Banker,submitted), thus consti tuting an excel lent marker for syn-aptic contacts. When neurons, surface stained for the VSV

glycoprotein, were incubated with anti-synap tophysin anti-bodies after permeabilization with Triton X-100, cluste rs ofVSV glycoprotein colocalizing with synapto physin wereobserved (Figures 2E and 2F, arrows). However, synap-tophysin and the VSV glycoprotein did not comp letelyoverlap (Figures 2E and 2F, arrowheads ).Newly Synthesized Fowl Plague Vlrus HAIs Dal ivemd Predominantly, but NotExclusively, to the Axonal DomainIn fi lter-grown MD CK cel ls, VSV glycoprotein is trans-ported exclusive ly to the basolateral domain, whereas inMDC K cel ls infected with influenza virus, the major glyco-protein of the virus, HA, is delivered directly to the apicalsurface (Rodriguez-Boulan and Pendergas t, 1980; Matlinand Simon s, 1984). Because of the dendritic delivery ofVSV glycoprotein, which could be comparable to the baso-lateral sorting in epithelial cells, we analyzed whether HAwas sorted to the axon. In the next series of experiments,we infected mature hippocampal neurons (stage 5) with awild-type strain of the avian influenza fowl plague virus(FPV) (Figure 3).

When FW-infected neurons were fixed in methanolat 4 hr after infection, immunofluorescence microscopyshowed that 46% of the cel ls had positive immunoreac-tivity with anti-HA antibodies. In these c ells, HA labelingwas evident in the neuronal cell body (Figures 3A-3C).


5/10

Sorting of Viral Glycoproteins to Axon and Dendrites67

Figure 4. Surface Distribution of the HA Pro-tein of FPV-Infected Stage 5 Neurons 5 Hr afterInfection(A-C) Infected cells were incubated with anti-HA antiserum 30 min at 4% fixed with parafor-maldehyde. and processed for immunofluo-rescence. Dendrite s (arrowheads) and axons(arrows) are disting uished by phase-contrastmicroscopy (A) and MAP 2 immunoreactiv ity(6). Axons are the only processes labe led wit hthe anti-HA antibodies (C).(D-E) The infected cells were paraformalde-hyde-glutaraldehyde fixed before anti-HA la-beling. Axons, th in and untapering, are labeledfor HA (arrows in [El). N one of the th ick, taper-ing processes evident by phase-contrast mi-

The appearance of labeling was similar to that found inVSV ts045-infected cel ls f ixed after the 20C block (Figure1C). After 5 hr of infection, an increase in the cell bodylabel ing was accompanied by the appearance of punctateHA immunoreactivi ty in some of the neuri tes (Figures3D-3F ). The cell in this figure is representative of the vas tmajori ty of FPV-infected neurons. Axons (MAP2 negative,ax, arrowheads) showe d intense H A labeling (Figure 3F).The staining was present in dots and varicosities alongthe entire length of the axons. HA labeling of dendriteswas variable. In some dendri tes HA was clearly absent(d3, d4, d5, d6). Other dendrites showe d som e H A label-ing, restricted usually to the initial segmen t (dl, d2). Thiscould be derived from Golgi eleme nts. HA labeling canalso be seen furthe r into the dendrites in dl and d2. Oneinterpretation of this labeling is that the staining was notin dendrites but was derived from axons running in closeproximity, parallel to the dendrit ic surface. This type ofaxo-dendritic organization is com mon ly o bserved in thisculture (Rothman and Cowan, 1981; Gosl in et al ., 1988).Observation of the dendrite dl (Figure 3F) tends to sup-port this view . The intense labeling of the proximal den-dri te becomes puncti form and fol lows the trajectory of theaxon th at loops at the end of the dendrite (the axonal loop-ing is evident in the corresponding phase-c ontrast micro-graph). Fixation of cells at later time s postinfection did notshow any significant change of the described pattern: nei-ther a higher num ber of dendrites becam e labeled norwas there an increase in axonal labeling. Quantitation

showed that of 815 HA-positive processes, 77% were axonswhile 33% w ere classified as dendrites.

The insertion of FW HA into the axonal mem brane wasinvestigated nex t. Cells were either incubated with anti-HA antiserum at 4% and then fixed or f ixed directly inparaformaldehyde-glutaraldehyde. With the first metho d,HA labeling appeared in the form of sma ll dots on the ax-onal surface (Figures 4A-4C). However, in the latter case,when lateral mob il ity was blocked by fixation, the HA ap-peared distributed uniformly (Figures 4D and 4E). In ei-ther case the labeling was restricted to axon s, leavingmo st of the dendrites unlabeled. The cell bodies wereusually, but not alwa ys, labeled with the anti-HA anti-bodies. Only 10.7% of the processes positive for HA weredendrites. In those cells in which dendrites were positivefor surface HA, the labeling was only in one or two den-drites and usually restricted to the region close to the cellbody ( less than 10 mm aw ay). In the cel l in Figures 4A-4C,the absence of dendritic labeling is unequivocal. Only ax-ons (MAP2 negative, sma ll arrowheads) appeared s tained.The charac teristic uniform labeling of axons obtained af-ter aldehyde fixation is also clear on the right side of thesame figure (Figures 4D and 4E). Only the very thin anduntapering proces ses running along dendrites and cellbody are intensely labeled.Delivery of Viral Glycoprote ins during theEstablish ment of Neuronal PolarityHippocam pal cells establish morphological polarity early


6/10

Figure 5. lmmunofluorescence Localization ofIntracellu lar VS V Glycoprotein and FPV HA inHippocampa l Cells at Stages 2 and 3 of De-velopmentVS V ts045-infected cells were methanol f ixedafter 60 min at permissive temperature (as inFigures ID-IF). FPWn fected cells were fixedat 5 hr postinfection. VS V glycoprotein appearsin all the minor processes of stage 2 cells (B).VS V glycoprotein is c learly present in the axonof stage 3 cells (C). HA protein was also foundin all neurites of stage 2 cells (E) and in theaxon and dendrites (d) of stage 3 cells (F). Bar:10 urn.

after plating (Dotti et al., 1988). Already after 24 hr in cui-ture an axon can be differentiated from a dendrite on apurely morphological basis . In the final series of experi-ments, we analyzed whether polarized del ivery of thenewly synthe sized viral glycoproteins also occurred incells at the early stages of polarization (stage 3). Typica lstage 3 cel ls are shown in Figures 5C and 5F They haveonly one long neurite (axon) and several s hort p rocess es(dendrites). Infection of this type of cel l with VSV tsO45 orwith FPV and subsequent immunofluorescence local iza-tion of the intracellular viral proteins revealed no spec ificdelivery to axon and dendrites. Positive HA or VSV glyco-protein staining in the form of fine dots was observed inthe single axon and the multiple minor proces ses. In-fected sta ge 2 cells, which are not morphologically pola-rized, also showe d uniform delivery of VSV glycoproteinand HA into all the cellular proces ses (Figures 58 and 5E).DiscussionOne important consequence of this study is that an ap-proach that has proved useful in studies of cell surfaceglycoprotein sorting in cultured fibroblasts and epithelialcells can now also be used to investigate polarized glyco-protein transport in neurons. A prerequisite for this workwas the availability of a neuronal cell culture in which thecells polarize into axons and dendrites. Previous studiesusing viruses to infect neurons or cells with neuron-likeproperties have not been able to address the question ofsurface glycoprotein polarity because the cultures havenot been as stably polarized and as well differentiated asthe mature hippocampal neurons (D ubois-Dalcq et al.,1982; R ivas and Moore, 1989; Tooze et al ., 1989). We ob-served that when immature hippocampal neurons wereinfected at the stage when the axon had grown out, del iv-ery of the virus glycoproteins was not yet polarized. Whenmature neurons were infected, however, with VSV or withinfluenza virus, the distribution of the viral glycoproteinsbecam e polarized. The VSV glycoprotein was delivered todendri tes, whereas the influenza HA was mainly routed tothe axon.

Where is the sorting of the viral glycoproteins taking

place? W e do not have conclusive evidence on this point,but our results suggest that the sorting occurs in the Golgicomplex. For the VSV glycoprotein, the data are morecompelling than for the influenza HA. When th e neuronsare infected with VSV ts845 at the nonpermissive temper-ature and then shifted to 20%, the tempera ture-sensitiveVSV glycoproteins move from the ER to the Golgi complexin the cel l body, presumably accumulating in the TGN(Griffiths et al., 1985). After a shift to 32X, the glycopro-teins mo ve to the dendrites and the perikaryon cell sur-face. At no time are the VSV glycoproteins seen in theaxon. For the influenza HA, accumulation in the Golgicomplex was also observed, fol lowed by preferential del iv-ery to the axon, but there wa s also some transport intodendri tes. Previous results for newly synthesized axonalproteins also showed tha t they pass through the Golgicomplex in the cel l body before moving on by fast axonaltransport (Hammerschlag et al ., 1982; Vale et al ., 1988).For endogeneous dendri tic glycoproteins, less work hasbeen done. There has been some evidence that glycopro-teins might be synthe sized locally in dendrites as well asin the cell body, followed by delivery to the cell surface(Kreutzberg et al ., 1973; Steward et al ., 1988). Howeve r,more work is clearly required to substantiate whether somedendri tic glycoproteins could bypass the Golgi complex.

Our data on VSV and influenza glycoprotein sorting inneurons suggest a vesicular sorting mechanism similar towhat has been observed previously in MDC K ce l ls. Aftersorting in the TGN in MD CK cel ls, the influenza HA isreleased into apical carrier ves icles, whereas the VSV gly-coproteins are transported in vesicular carriers to thebasolateral side (Bennett et al., 1988). The hepato cytemodel (see Introduction) for glycoprotein transport seem sunlikely for hippocampal neurons. If the influenza HAwere first delivered to the soma to-dendritic domain andthen routed by a transcellular route to the axon, we shouldhave been able to follow the HA from the dendritic side tothe axon in the highly polarized neurons. Our data sug-gest that sorting occurred before surface del ivery. How-ever, this question will have to be analyzed further bydetailed studies of temperature-sensitive mutants of theinfluenza HA.


7/10

;;rt ing of Viral Glycoproteins to Axon and Dendrites

The small amount of influenza HA delivered to dendritescould be due to several causes. It may represent missort-ing either by inclusion of HA into dendritic carrier v esiclesor by delivery of axonal carrier vesic les to the wrong side(Pfeiffer et al., 1985). If the dendritic pathw ay were a de-fault pathway, then this dendri tic missorting would besimilar to wha t has been postulated for the basolateralpathway in MD CK cel ls. Alternatively, i t is also possiblethat viral infection leads to cytopathic effects in some cel lsthat influence sorting fidelity (Fuller et al., 1984).

After sorting into axonal and dendritic carrier ves iclesin the TGN has occurred, these have to be del ivered intothe axon and the dendrites. For the axonal vesicular inter-mediates, i t is l ikely that kinesin-mediated fast axonaltransport along microtub ules is the mode of delivery (Valeet al., 1985, 1986). We observed a punctate pattern of in-tracellular influenza HA staining throughout the lengthof axon, presum ably representing the vesicular carriersmoving in the anterograde direction. After 4 hr of infection,the HA wa s still in the Golgi region of the ce ll body,whereas after an additional 30-60 min, HA had beentransported al l the way to the end of the axon. Fast axonaltransport would easily fill the entire axon w ith the requiredspeed (Sheetz et al ., 1989).

Dendri t ic del ivery of VSV glycoproteins may also occurby a microtubule-m ediated interaction. The dendritic m i-crotubules are of opposite polarity, abou t 50% havingtheir plus ends in the cell body and the other half havingthem in the dendrit ic shafts (Baas et al ., 1988). The den-dritic carrier ves icles could use a MAP -1C dynein-like mo-tor (Paschal and Vallee, 1987) to mov e fro m the cell body(plus ends) toward the microtubule minus ends in the den-dri tes, as suggested by Black and Baas (1989) for or-ganelle translocation. This mechanism would preventdendri tic vesicles from entering the axonal traffic. The ax-onal vesicles could either la ck the capa city to bind to thedendritic microtubu les with their minus ends in the cellbody, or bind and move into the dendrites but then moveback to the cell body using the population of microtub ulesof opposite polarity. Th ese are possibilities that requirefurther study.

There was a str iking difference in the surface immuno-fluorescenc e after glutaraldehyde and paraformaldehydefixation of the influenza HA and of the VSV glycoprotein.The HA was diffusely spread over the entire axolemmalsurface, whereas the surface VSV glycoprotein stainingwas patched on the cel l body and on the dendri tes. Thesepatches often corresponded to si tes where synaptophysinstaining was observed. Synaptophysin is known to ac-cumulate in synaptic vesicles in presynaptic bulbs alongthe axon (Jahn et al ., 1985; Wiedenmann and Franke,1985; Tixier-Vidal et al., 1988). The VSV glycoproteinmight pa tch after s urface delivery in or close to dendriticspines (D ubois-Dalcq et al., 1982) by interactions withcytoskeletal elements (Siman et al ., 1985; Lazarides andNelson, 1985; Jones et al ., 1989). The influenza HA seemsto be freely diffusing along the axolem mal surface be-cause anti-HA antiserum patched the surface HA i f addedbefore fixation.

One important question is how the segregation of the vi-

ral glycoproteins in the plasma mem brane plane is main-tained. If the HA were freely m obile then i t should movefrom the axon (and the cell body) into dendrites by lateraldiffusion. Conversely, the VSV glycoprotein could movefrom the cell body into the axon. For the VSV glycoprotein,segregation could be maintained by anchoring to proteinsunderlying the plasma membrane. Howeve r, we have evi-dence that VSV progeny are released by budding into theextracellular medium (C. G. D. and K. S., unpublisheddata). This would imply that at least some of the VSV gly-coproteins must be mobile to form the patch for interactionwith the VSV nucleocaps id to drive the budding proce ss.Is it possible that there is a fence blocking lateral diffus ionin the plasma mem brane-for instanc e, in the axon hil-lock? The axon hi l lock is located at the boundary betweenthe axon and the soma to-dendritic domain, and it isknown to have a distinctive membrane organization (Mat-sum oto and Rose nbluth, 1985; Angelides et al., 1988). Inepithelial cells the tight junction mark s the boundary be-tween the apical and the basolateral plasma membranedomains and act s as a fence for both lipid and p roteindiffusion. In addition, interactions with ankyrin and fodrinare know n to organize basolateral proteins (e.g., the an-ion exchanger and the [Na+,K+] ATPase) in subdomainson the basolateral surface (Drenckha hn et al., 1985; Nel-son and Veshnock, 1987; Nelson and Hamm erton, 1989).Similar m echanisms are known to anchor sodium chan-nels and glutamate receptors in neurons (Siman et al.,1985; Srinivasan et al., 1988).

The final question that has to be answered by further ex-periments is when the sorting in the TGN becomes pola-rized during neuronal developm ent. Previous work onGAP43, the fatty acylated growth cone protein (Skene,1989) has suggested that already at stage 3, when m or-phological polarity has developed, preferential delivery ofGAP43 can be observed to the axon growth cone (Goslinet al ., 1990). Also, synaptophysin seems to become pola-r ized at this stage (T. Fletcher, l? Cameron, l? De Camil li ,and G. Banker, submitted). These results suggest that al lthe cellular compon ents neces sary for polarized s ortingare present in immature neurons and that assembly of thesorting and the del ivery mechanism s starts already dur-ing axon outgrowth. However, i t is not yet clear whetherthese findings only reflect a mas sive delivery of proteininto a rapidly growing axon (as is the cas e with the deliveryof newly synthesized glycoproteins to the leading edge ofa motile fibroblast; Bergmann et al., 1983) or whetherpolarized sorting in the TG N is indeed taking place. Ourown results show that al though there are differencesamong stage 3 neurons, no significant polarity in the deliv-ery of VSV glycoproteins and HA could be observed. Atthis stage, cell polarity can still be reversed. When theaxon of a stage 3 neuron is amputated, anyone of i ts re-maining neuri tes, which would otherwise become den-dri tes, have the capacity to differentiate into a new axon(Dotti and Banker, 1987; Goslin and Banker, 1989). Stabili-zation of neuronal polarity take s place later. If axon out-growth only requires massive membrane traffic directedinto one of the neurites but no polarized sorting, then thedifferentiation of the imma ture neuron could require gene


8/10

activation and synthesis of new proteins necessary fortransforming the TGN from a state making only one typeof carrier vesicle to the cell surface (cf. fibroblast) to an-other in which two types of vesicular intermediates aremade (cf. MDC K cel ls). Alternatively, the carr ier vesiclesand/or the microtubules could be transformed such thataxonal and dendritic segrega tion becom es possible bydifferential microtubule movem ent. This gene activationcould either be triggered by synapto genesis or precede it.Before it is possible to tes t unequivocally whether eitherof these two models is correct, more has to be learnedabout the molecular mach inery responsible for generat-ing and maintaining neuronal polarity. One very promisingprospec t is that the analysis of epithelial glycoprotein sort-ing might also provide clues to neuronal sorting mech a-n isms.Experimental Procedurescell CultumHippocampa l cells were prepared as described by Barlett and Banker(1984a). Briefly, the hippocampi of 17day-old rat embryos were mi-croscopically dissected, trypsinized (0.25% for 15 min). and furtherdissociated by repeate d passages through a constricted Pasteur pi-pette. Cells were plated onto polylysine-coated coverslips in dishescontaining MEM supplemented with 10% fetal calf serum and allowedto attach to the substratum for 30 min. The neuronal coverslips werethen transferred to culture dishes co ntaini ng a monola yer of astrocytesthat had been kept in serum-free medium (MEM supplemented withthe N2 mixture of Bottestein and Sato [lQ7Q]) for at least 24 hr. Thismedium will be referred to as neuronal medium. The coverslips werenext put with the neurons upside down, in c lose proximity toward theglia l monolayer. This coculture, neuron-glie, was previously shown toimprove the long-term survival of the neurons (Banker, 1980). Cellswere maintained in a humidifie d incubator at 3pC and 5% COP forlo-14 days before virus infections were performe d. For some experi-ments cells were kept in culture for only 24-48 hr before virus infection(stage 2 and stage 3 neurons).

InfectIon wlth V!W tW45 and FPVStocks of FPV and of VS V tsO45 were obtained as described in Matlinand Simons (1984) and Griff iths et a l. (1985) respectively. Coverslipswith neurons were removed from the culture dishes, r insed in Hanksbalanced salt solution (HBSS), and infected w ith 80-100 pfu per cellin neuronal medium . The temperature of infection differed dependingon which virus was used: 32% (permissive temperat ure) was the infec-tion temperature for the VS V mutant tsO45, and 3pC for FPV After 1hr of infection in a 5% CO? atmosphere the coverslips were rinsed inHBS S and returned to fresh neuronal medium for d ifferent t imes andconditions depending on the virus. VS V ts045-infected neurons werefurther incubated in a 5% COP incubator at the nonpermissive tem-perature of 39% for 2.5 hr. Cells w ere then transferred to dishes con-tain ing neuronal medium at 2ooC for 90 min. The temperature wasmaintained constant by using a temperature-controlled water bath.The pH of the medium was kept constant by having the neuronalmed ium mod i f ied wi th HEPES (10 mM) and low NaHCOs (4 mM). F i -nally, the coverslips were transferred to dishes containing mediumat the permissive temperature of 32oC and kept in a 5% COs at-mosphere for various times. Coverslips with FPVinfected neuronswere briefly r insed in HES S and transferred to fresh neuronal mediumat 3pC in a 5% COs incubator. Infection was carried out for d ifferenttimes before fixation and processing for immunofluorescence.

ImmunofluomecenceTo label intracellu lar spike VSV glycoproteins or FPV HA, coverslipswere briefly r insed in phosphate-buffered saline (PBS) and fixed in-20C methanol for 6 min. After f ixation, coverslips were allowed toair dry and then were rehydrated with PBS. Unspecific primary anti-body binding was prevented by a 30 min incubation in PB S containing

10% horse serum. The cells were then incubated in the presence ofprimary antibodies for e ither s ingle or double immunofluorescence.The fo llowing primary antibodies were used: rabbit anti-VSV glycopro-tein (Matlin et a l., 1982) rabbit anti-FPV HA (Matlin et a l., 1981) mousemonoclonal anti-MAP 2 (gift from Dr. Lester Binder), and mouse mono-clonal anti-synaptophysin (gift from Dr. Wiedenmann). Incubation withthe primary antibodies was for 1 hr at room temperature. The reactionwas stopped by three 5 min rinses in PB S followed by a 30 min incuba-tion with the corresponding secondary antibodies (FITC-conjugatedgoat anti-rabbit IgG and RlTCconjugated rat anti-mouse IgG, Di-anova Gmbh, FRG), previously immuno purif ied on nitmcetlu lose paper.

To label v ira l g lycoproteins on the neuronal surface, infected cellswere fixed in 4% paraformaldehyde-0.25% glutaraldehyde fo llowedby quenching of the aldehydes with sodium borohydride. After exten-sive washing the coverslips were incubated with anti-VSV glycoproteinor anti-HA antibodies in PB S without detergent. This incubation wasfollowed by incubation with an FITC-conjugated goat anti-rabbit IgGantiserum. Cells were then permeabilized with Triton X-100 (O.l%, 10min) and further incubated with anti-MAP 2 or anti-synaptophysin anti-bodies fo llowed by RITC-conjugated rat anti-mouse IgG antibodies,This fixation procedure only labeled the VS V glycoprotein on the cellsurface. When the anti-MAP 2 and anti-VSV glycoprotein antibodieswere added together, no cytoskeleton labeling was seen. A second wayto label surface vira l g lycoproteins was to incubate the infected cellswith the anti-v ira l spike glycoprotein antibodies at 4OC, which is knownto block endocytosis (von Bonsdorff et a l., 1985). Cells were then fixedin 4% paraformaldehyde and processed for immunofluorescence asbefore. lmmunofluorescence analysis was performed in a fluores-cence microscope (Axioph ot, Zeiss, FRG). Photog raphs were takenusing high sensitiv ity f i lm (TMax 3200, Kodak).AcknowledgmentsWe thank Hilkka V irta for providin g virus stocks, Dr. Lester Bin der forgiv ing us the anti-MAP 2 antiserum, and Dr. Bertram Wiedenmann forthe anti-synaptophys in antiserum . We are also gratefu l to Drs. GaryBanker, Jan De Mey, and Wieland Huttner for crit ical reading of themanuscript. Dr. Carlos G. Dotti is the recipient of an Alexander vonHumboldt Stiftung Fellowship.

The costs of public ation of this article were defraye d in part by thepayment of page charges. This artic le m ust therefore be herebymarked advertisement in accordance with 18 USC. Section 1 734solely to indicate th is fact.Received February 20, 1990; revised May 7, 1990.ReferencesAngelides, K. J., Elmer, L. W., Loftus, D.. and Elson, E. (1988). Distribu-tion and lateral mobility of voltage-dependent sodium channels in neu-rons J. Cell Bio l. 106, 19111925.Baas, I? W., Deitch, J. S., Black, M. M., and Banker, G. A. (198F1). Polar-ity orientation of microtubules in hippocampal neurons: uniformity inthe axon and nonuniformity in the dendrite. Proc. Natl. Acad. Sci. USA85, 8335-8339.Banker, G. (1980). Trophic interactions between glia l cells and hippocampal neurons in culture. Science 209, 809-810.Banker, G. A., and Waxman, A. 8. (1988). Hippocampal neuronsgenerate natural shapes in cell culture. In Intrinsic Determinants ofNeuronal Form and F unction, R. J. Lasek, ed. (New York: A. L iss), pp.61-82.Barlett. W. O., and Banker, G. A. (1984a). An electron microscopystudy of the development of axons and dendrites by hippocampal neu-rons in culture. I. Cells which develop without intercellu lar contacts, J.Neurosci. 4, 1944-1953.Barlett, W. O., and Banker, G. A. (1984b). An electron microscopystudy of the development of axons and dendrites by hippocampal neu-rons in culture. II. Synaptic re lationships. J. Neurosci. 4, 1954-1965.Bartles, J. R., Feracci, H. M.. Stieger, B., and Hubbard, A. L. (1987).Biogenesis of the rat hepatocyte plasma membrane in vivo: compari-son of the pathways taken by apical and basolateral proteins using sub-cellu lar fractionation. J. Cell Biol. 105, 1241-1251.


9/10

;prting of Viral Glycoproteins to Axon and Dendrites

Bennett, M. K., Wandinger-Ness, A., and Simon% K. (1988). Releaseof putative exocytic vesicles from perforated MDC K cells. EMBO J. 7,4075-4088.Bergmann, J. E., and Singer, S. J. (1983). lmmunoelectron micro-scopic studies of the intracellu lar transport of the membrane glycopro-tein (G) of vesicular stoma titis virus in infected Chines e hamster ovarycells. J. Cell Biol. 97; 1777-1787.Bergmann, J. E., Tokayasu, K. T., and Singer, S. J. (1981). Passage ofan integral m embrane protein, the vesicular stomatit is v irus glycopro-tein. through the Golgi apparatus en route to the plasma membrane.Proc. Natl. Acad. Sci. US A 78, 1746-1750.Bergmann, J. E., Kupfer, A., and Singer, S. J. (1983). Membrane inser-tion at the leading edge of motile f ibroblasts. Proc. Natl. Acad. Sci. USA80, 1367-1371.Black, M. M.. and Baas, P W. (1989). The basis of polarity in neurons.Trends Neuros ci. 72, 211-214 .Bottestein, J. E., and Sato, G. H. (1979). Growth of a rat neuroblastomacell l ine in serum-free supplemented medium . Proc. Natl. Acad. Sci.USA 76, 514-519.Caceres, A., Banker, G. A., Steward, O., Binder, L., and Payne, M .(1984). MAP 2 is localized to the dendrites of h ippocampal neuronswhich develop in culture. Dev. Brain Res. 73, 314-318.Caceres. A., Banker, G. A., and Binder, L. I. (1986). lmmunocytochemi-cal localization of tubulin and microtubule-associated protein 2 duringthe development of h ippocampal neurons in culture. J. Neurosci. 6.714-722.De Cam ill i, P, Moretti, M., Donini, S. D., Walter, U., and Lohman, S. M.(1986). Heterogeneous distr ibution of the CAM P receptor protein RII inthe nervous system: evidence for its intracellu lar accumulation onmicrotubules, microtubule-organizing center, and in the area of theGolgi complex. J. Cell Biol. 703, 189-203.Dotti, C. G., and Banker, G. A. (1987). Experimetally induced alterationin the polarity of developing neurons. Nature 330, 254-256.Dotti. C. G., Banker, G. A., and Binder, L. I. (1987). The expression anddistribution of the microtubule-associated proteins tau and micro-tubule-associated protein 2 in hippocampal neurons in the rat in s ituand in cell culture. Neuroscience 23, 121-130.Dotti, C. G., Sullivan, C. A., and Banker, G. A. (1986). The establish-ment of polarity by hippocampal neurons in culture. J. Neurosci. 8,1454-1468.Drenckhahn, D., Schluter, K., and Bennett, V. (1985). Colocalization ofband 3 with ankyrin and spectrin at the basal membrane of intercalatedcells in the rat k idney. Science 230, 1287-1289.Dubois-Dalcq. M., Rentier, B., Hooge-Peters, E.. Haspel, M. V., Kno-bier, R. L., and Holmes, K. (1982). Acute and persistent v ira l infectionsof d ifferentiated nerve cells. Rev. Infect. Dis. 4, 999-1014.Dubois-Dalcq, M.. Holmes, K., and Rentier, 8. (1984). Assembly of En-veloped RN A Viruses (V ienna: Springer-Verlag).Fuller, S. D., von Bonsdorff, C.-H., and Simons, K. (1984). Vesicular sto-matit is v irus infects and matures only through the basolateral surfaceof the polarized surface epithelia l cell l ine, MDCK. Cell 38, 65-77.Goslin, K., and Banker, G. A. (1989). Experimental observation on thedevelopment of polarity by hippocampal neurons in culture. J. CellBiol. 708, 1507-1516.Goslin, K., Schreyer, D. J.. Skene, J. H. l?, and Banker, G. A. (1988).Development of neuronal polarity: G AP-43 distinguishes axonal fromdendrit ic growth cones. Nature 336, 872-674.Goslin, K., Schereyer, D., Skene, J. H. P, and Banker, G. (1990).Changes in the distr ibution of GAP-43 during the development of neu-ronal polarity. J. Neurosci., in press.Griff iths, G., and Simon% K. (1986). The trans Golgi network: sortingat the exit s ite of the Golgi complex. Science 234, 438-443.Griff iths, G., Pfeiffer, S ., Simon% K., and Matlin, K. (1985). Exit of newlysynthesized membrane proteins from the trans cisterna of the Golgicomplex to the plasma membrane. J. Cell Biol. 707, 949-964.Gumbiner, B. (1987). Structure, b iochemistry, and assembly of epithe-lia l t ight junctions. Am. J. Physiol. 253, C749-C758.Hammerschlag, R., Stone, G. C., Balen, F. A., L indsey, J. D., and Ellis-

man, H. D. (1982). Evidence that newly synthesized proteins destinedfor fast axonal transport pass through the Golgi apparatus. J. Cell Biol.93, 568-575.Hughson, E., Wandinger-Ness, A., Gausepohl, H.. Griff iths, G., and Si-mans. K. (1988). The cell b io logy of enveloped virus infection of epithe-lia l t issues. In The Molecular Biology of Infectious Diseases, Centen-ary Symposium of the Pasteur Institute, M. Schwartz, ed. (Paris:Elsevier), pp. 75-89.Jahn, R.. Schiebler, W.. Ouimet, C., and Greengard, P (1985). A38,OOOdalton membrane protein (P 38) present insynapticvesicles. Proc. Natl.Acad. Sci. US A 88, 4137-4141.Jones, 0. W., Kunze, D. L., and Angelides, K. J. (1989). Localizationand mobility of w-conotoxin-sensitive Caz channels in hippocampalCA1 neurons. Science 244, 1189-1193.Kelly, R. B. (1988). The cell b io logy of the nerve terminal. Neuron 7,431-438.Kreutzberg, G. W., Schubert, P., T&h, L., and Rieske, E. (1973). In-tradendritic transport to postsynaptic sites. Brai n Res. 62, 399-404 .Kuypers, H. G. J. M., and Ugolin i, C. (1990). Viruses as transneuronaltracers. Trends Neurosci. 73, 7l-75.Lazarides, E., and Nelson, W. J. (1985). Expression and assembly ofthe erythroid membrane-skeletal proteins ankyrin (g lobin) and spectrinin the morphogenesis of chicken neurons. J. Cell. Biochem . 27, 423-441.Matlin, K. S.. and Simon& K. (1984). Sorting of an apical p lasma mem-brane glycop rotein occurs before it reaches the cell surface in culturedepithelia l cells. J. Cell Bi ol. 99, 2131-2139.Matlin, K. S., Reggio, H., Helenius, A., and Simon?., K. (1981). Infec-tious entry pathway of influenza virus in a canine kidney cell l ine. J.Cell Biol. 97, 601-613.Matlin, K. H., Reggio. H.. Helenius, A., and Simons. K. (1982). Pathwayof vesicular stomatit is v irus entry leading to infection. J. Mol. Biol. 756,609-631.Matsumoto, E., and Rosenbluth, J. (1985). Plasma m embrane struc-ture at the axonal h il lock, in it ia l segment and cell body of frog dorsalroot ganglion cells. J. Neurocytol. 74, 731-747.Nelson, W. J., and H ammerton, R. W. (1989). A membrane-cytoskel-etal complex containing Na+.K+-ATPase, ankyrin, and fodrin in Madin-Darby canine kidney (MDCK) cells: implications for the biogenesis ofepithelia l cell polarity. J. Cell Biol. 708, 893-902.Nelson, W. J.. and Veshnock, P J. (1987). Ankyrin binding to (Na+ +K+) ATPase and implications for the organization of membrane do-mains in polarized cells. Nature 328, 533-536.Paschal, B. M., and Vallee, R. B. (1987). Retrograde transport by themicrotubule-associated protein MAP 1C. Nature 330, 181-183.Pfeffer, S . R., and Rothman, J. E. (1987). Biosynthetic protein transportand sorting by the endoplasmic reticulum and Golgi. An nu. Rev. Bio-them. 56, 829-852.Pfeiffer, S., Fuller, S. D., and Simons, K. (1985). Intracellu lar sortingand basolateral appearance of the G protein of vesicular stomatit is v i-rus in MDC K cells. J. Cell B iol. 707, 470-476.Rivas, R. J., and Moore, H.-t? (1989). Spa tia l segregation of the secre-tory and constitutive secretory pathways. J. Cel l Bio l. 109, 51-60.Rodriguez-Boulan, E., and Nelson, W. J. (1989). Morphogenesis of thepolarized epithelia l cell phenotype. Science 245, 7 l8-725.Rodriguez-Boulan, E., and Pendergast, M. (1980). Polarized distr ibu-tion of v ira l envelope glycoproteins in the plasma membrane of in-fected epithelia l cells. Cell 20, 45-54.Rothman, S., and Cowan, W. M. (1981). A scanning electron micros-copy study of the in v itro development of d issociated hippocampalcells. J. Comp. Neural. 795. 141-155.Shaw, G., Banker, G. A., and Weber, K. (1985). An im munofluores-cence study of neurofilament expression by developing hippocampalneurons in tissue culture. Eur. J. Cell Bio l. 39, 205-216.Sheetz, M. P., Steuer, E. R., and Schroer, T. A. (1989). T he mecha-nisms and regulation of fast axonal transport. Trends Neurosci. 72,474-478.Sima n, R., Baudry, M.. and Lynch, G. (1985). Regu lation of g lutamate


10/10

Ce l l72

receptor b inding by the cytoskeletal protein fodrin. Nature 373, 22%228.Simons, K., and Fuller, S. D. (1985). C ell surface polarity in epithelia.Annu. Rev. Cell B iol. 7, 243-288.Sim on& K., and Warren, G. (1984). Semlik i Forest v irus: a probe formembrane traffic in the anima l cell. Adv. Prot. Chem. 36, 79-132.Skene, J. H. P (1989). Axonal growth associated proteins. Annu. Rev.Neurosci. 72, 127-156.Srin ivasan, Y., Elmer, L., Davis, J., Bennett, V., and Angelides, K.(1988). Ankyrin and spectrin associate with voltage-dependent sodiumchannels in brain. Nature 333, 177-180.Steward, O., Davis, L., Dotti, C., Philips, L.. Rao, A., and Banker, G.(1988). Protein synthesis and processing in cytoplasmic microdomainsbeneath postsynaptic s ites in CNS neurons. Mol. Neurobiol. 2, 227-261.Tixier-Vidal, A., Faivre-Bauman, A., Picart, R., and Wiedenmann, B.(1988). lmmunoelectron microscopy localization of synaptophysin in aGolgi subcompartment of developing hypothalamic neurons. Neuro-science 26, 847-881.Tooze, J., Hollinshead, M., Fuller, S. D., Tooze, S. A., and Huttner, W. B.(1989). Morphological and biochemical evidence showing neuronalproperties in AtT-20 cells and their growth cones. Eur. J. Cell Biol. 49,259-273.Vale, Ft. D., Reese, T. S., and Sheetz, M. P (1985). Identif ication of anovel force-generating protein, k inesin, involved in microtubule-basedmotil ity. Cell 42, 39-50.Vale, R. D., Scholey, J. M., and Sheetz, M. I? (1986). Kinesin: possiblebiological ro les for a new microtubule motor. Trends B iochem. Sci. 77,464-468.von Bonsdorff, C.-H., Fuller, S. D., and Simons, K. (1985). Apical andbasolateral endocytosis in Madin-Darby canine kidney (MDCK) cellsgrown on nitrocellu lose fi lters. EM BO J. 4, 2781-2792.Wandinger-Ness, A., and Simon % K . (1990). The polarized transportof surface proteins and lip ids in epithelia l cells, In Intracellu lar Traffick-ing of Proteins, J. Hanover and L. Steer, eds. (Cambridge, England:Cam bridge University Press), in press.Wiedenmann, B., and Franke, W. W. (1985). Identif ication and locali-zation of synaptophysin, an integral mem brane glycoprotein of M,38,000 characteristic of presynaptic vesicle. Cell 47, 1017-1028.

Dottie and Simons Neuronal and Epithelial Comp1

Documents