Rabies virus entry into cultured rat hippo campalneurons

Rabies virus entryL E W I S and L E N T Z

Rabies virus entry into cultured rat hippo campalneuronsP E T E R L E W I S a n d T H O M A S L . L E N T Z *

Department of Cell Biology, Yale University School of Medicine, 333 Cedar St. P.O. Box 208002, New Haven, CT 06520-8002

Received 9 April 1998; revised 29 June 1998; accepted 22 July 1998

Summary

Rabies virus entry into cultured hippocampal neurons was investigated by immunofluorescence and electron microscopy.Hippocampal neurons were susceptible to rabies virus infection and became filled with viral antigen 1 day after infection.Infection was inhibited by the lysosomotropic agents chloroquine and ammonium chloride. To study entry, neurons wereadsorbed with rabies virus at 48C and warmed to 378C for short periods of time prior to fixation and localization of viral antigenby immunofluorescence microscopy. By 5 min at 378C, viral antigen was localized to puncta in the cell body and dendrites andin synapses along dendrites. Little viral antigen was present in axons. Cells adsorbed with rabies virus were incubated withtracers for early endosomes. The endocytic tracers or markers Lucifer Yellow, transferrin receptor, dextran, and wheat germagglutinin co-localized with rabies virus, indicating that rabies virus enters an endosome compartment shortly after uptake.Rabies virus also co-localized with LysoTracker Red, an acidotropic probe, indicating that some of the virus-containingendosomes are acidified. Rabies virus also co-localized with synapsin I, a synaptic vesicle marker, in nerve terminals but didnot co-localize with lysosomal glycoprotein. By electron microscopy, after adsorption of virus and warming for 10 min, virusparticles were present in coated pits, coated vesicles, and vacuolar membrane compartments in processes and axon terminals.It is concluded that rabies virus enters the somatodendritic domain and axon terminals of cultured hippocampal neurons byadsorptive endocytosis and is located in endosomes shortly after uptake.

Introduction

The pathogenesis of rabies has been investigated inanimals by immunoflourescence and electron micro-scopy (Murphy, 1977; Baer & Lentz, 1991; Iwasaki,1991). These studies have demonstrated the retrogradetransport of rabies virus from the periphery to the spi-nal cord, the ascent in the spinal cord, and the spreadin the brain. Although the pathways and sites of rabiesvirus infection are characterized, less is known aboutthe cellular events of rabies virus entry and transportin neurons. Enveloped viruses such as rabies enter cellsby direct fusion with the plasma membrane or by ad-sorptive or receptor-mediated endocytosis (Paulson,1985; Marsh & Helenius, 1989; Lentz, 1994). Rabies vi-rus appears to enter chicken embryo-related (CER)cells by adsorptive endocytosis, since virus particleshave been observed in coated pits and vesicles (Supertiet al., 1984). Endocytosis of virus particles by nerveterminals has been observed by electron microscopy(Iwasaki & Clark, 1975; Charlton & Casey, 1979; Baer &Lentz, 1991). In adsorptive endocytosis, internalizedvirus particles are delivered to endosomes. The acidicinterior of the endosome, generated by an ATP-driven

proton pump, triggers fusion of the viral membranewith the endosome membrane, allowing the viral nu-cleocapsid to escape into the cytoplasm. In rabies, aconformational change in the G protein mediates fu-sion with membranes (Gaudin et al., 1991).

It is assumed that rabies virus enters neurons by en-docytosis, but except for the electron microscopic stud-ies, entry into neurons has received little study. In addi-tion, the properties of endosomes in neurons are not aswell characterized as those in other cell types. The pre-sent study was undertaken to determine whether rabiesvirus enters hippocampal neurons by adsorptive endo-cytosis and resides in an endosome compartment. Sev-eral studies have shown that neurons of the hippocam-pus are infected by rabies virus in vivo (Johnson, 1965;Murphy et al., 1973; Iwasaki & Clark, 1975; Charlton &Casey, 1979; Iwasaki et al., 1985). Techniques for theculture of hippocampal neurons are well established(Banker & Cowan, 1977; Bartlett & Banker, 1984). In thisstudy, it was found that cultured hippocampal neuronsare highly susceptible to infection by rabies virus. Hip-pocampal neurons were adsorbed with rabies virus and

0300–4864 © 1998 Kluwer Academic Publishers

*To whom correspondence should be addressed.

Journal of Neurocytology 27, 559–573 (1998)

incubated with tracers for early endosomes, and thevirus and tracers or markers for endosomes were lo-calized at early time periods by immunofluorescencemicroscopy. Rabies virus and tracers or markers forendosomes were co-localized in the cell body and den-drites and in axon terminals. By electron microscopy,virus particles were present in coated pits and vesiclesand in vacuoles similar to endosomes in other cells.These findings show that rabies virus enters hippocam-pal neurons by adsorptive endocytosis and resides inan endosome compartment shortly after uptake.

Materials and methods

MATERIALS

Fluorescein isothiocyanate (FITC)-conjugated anti-rabiesglobulin was purchased from Becton Dickinson; rhodamine-conjugated wheat germ agglutinin from Vector Laboratories;Texas Red-conjugated dextran, LysoTracker Red DND-99,and rabbit antibody against Lucifer Yellow from MolecularProbes; Texas Red-conjugated donkey anti-rabbit and TexasRed-conjugated goat anti-mouse antibodies from Jackson Im-munoResearch; Lucifer Yellow CH from Sigma; chloroquinephosphate from Research Biochemicals International; andammonium chloride (NH4Cl) from J. T. Baker. Mouse mono-clonal antibody against rat lysosomal glycoprotein 120 andmouse monoclonal antibody against rat neural cell adhesionmolecule (NCAM) L1 (ASCS4) were gifts from Dr Ira Mell-man; mouse monoclonal antibody against transferrin recep-tor was a gift from Dr David Sheff; and rabbit antibodyagainst rat synapsin I and mouse monoclonal antibodyagainst rat microtubule-associated protein-2 (MAP-2) weregifts from Dr Pietro De Camilli.

CELL CULTURE

Rat hippocampal cell cultures were prepared from hippo-campi of 17–18-day-old fetal rats and cultured on coverslipsat low density (3000–12,000 cells/cm2) and maintained asdescribed (Banker & Cowan, 1977; Bartlett & Banker, 1984;Mundigl et al., 1993). Some cultures were grown over a mono-layer of astroglial cells. All hippocampal neurons were ma-ture when used (stage 5 according to Dotti et al., 1988), but thetime in culture ranged from 6 to 82 days.

VIRUS

The challenge virus standard (CVS) strain of rabies virus wasused in these studies. Virus was propagated in mouse neu-roblastoma (N18) cells in tissue culture flasks as described(Lentz et al., 1985). Clarified tissue culture supernatants hadan infection potency of 5 3 105 fluorescent focus units/ml andwere used for infection of hippocampal neurons. More con-centrated virus preparations were obtained by incubating cul-ture supernatants overnight at 48C with gentle stirring in thepresence of polyethylene glycol 6000 (8%), 2.2% NaCl, and0.01% phenyl-methyl-sulfonyl fluoride. The solution was cen-trifuged for 1 h at 15,000 3 g in a GSArotor in a Sorvall RC-2Bcentrifuge. The pellets were resuspended to 5 ml in tissueculture media. This preparation had an infection potency of 5

3 107 fluorescent focus units/ml and was used for adsorbingvirus to the surface of neurons.

INFECTION OF HIPPOCAMPAL NEURONS

For infection of hippocampal neurons, virus culture super-natant was added to medium in wells of a 12-well tissueculture plate (final concentration 5 3 103 fluorescent focusunits). Coverslips with cells were added to the wells andincubated at 378C for 6 h to 4 days prior to fixation. Somecultures were infected in the presence of NH4Cl (20 mM) orchloroquine (200 lM), both lysosomotropic agents that raisethe pH of endosomes and lysosomes. At the end of the incu-bation period, cultures were fixed and incubated with FITC-anti-rabies virus antibody as described below.

ADSORPTION OF VIRUS AND LABELLING WITHTRACERS

For adsorption of virus, cells on coverslips were incubatedfor 1 h at 48C in concentrated virus preparation diluted 1:100–1:250 in medium (2–5 3 105 fluorescent focus units). Cyclo-heximide (100 lg/ml) was included to prevent viral replica-tion. Tracers (Lucifer Yellow CH, 0.5 mg/ml; rhodamine-wheat germ agglutinin, 20 lg/ml; Texas Red-dextran, 0.5mg/ml) were added and the cultures warmed to 378C.LysoTracker Red DND-99 (2 lM), a fluorescent acidotropicprobe for labelling acidic compartments, was added withvirus. They were incubated for various times up to 30 min,then rinsed twice with cold buffer (sodium phosphate, 120mM, pH 7.4) and fixed for 30 min with freshly prepared 4%paraformaldehyde in 120 mM sodium phosphate, pH 7.4, in4% sucrose. At this stage, LysoTracker Red-labelled cells wererinsed twice and mounted on microscope slides for photog-raphy of the LysoTracker Red fluorescence. Each experimentwas performed 3–5 times on cultures of different ages.

IMMUNOFLUORESCENCE

All preparations were then washed three times for 5 min ineach of the following solutions: 120 mM sodium phosphate,pH 7.4; 150 mM sodium chloride, 10 mM sodium phosphate,pH 7.4; 500 mM sodium chloride, 20 mM sodium phosphate,pH 7.4. Next they were permeabilized and non-specific bind-ing blocked by incubation for 30 min in the same solutionwith the addition of 0.3% Triton X-100 (0.01% saponin forlysosomal glycoprotein) and 10–16% goat serum (Triton solu-tion). Cells exposed to Texas Red-dextran and rhodamine-wheat germ agglutinin were incubated with FITC-labelledanti-rabies virus antibody (1:50–1:250) in Triton solution for2–4 h. Lucifer Yellow, transferrin receptor, lysosomal glyco-protein, and synapsin I were localized using primary andsecondary antibodies. Primary antibodies (rabbit anti-LuciferYellow [1:200], mouse monoclonal anti-transferrin receptor[1:50], mouse monoclonal anti-lysosomal glycoprotein 120[1:25], rabbit anti-synapsin I [1:50], mouse monoclonal anti-MAP-2 [1:200], mouse monoclonal culture supernatant anti-NCAM L1 [1:1] antibodies) and FITC-anti-rabies antibodywere diluted in the Triton solution and incubated for 4 h toovernight. Cells were washed for 30 min in three changes of500 mM sodium chloride, 20 mM sodium phosphate, pH 7.4and were then incubated for 2 h in Triton solution with sec-ondary antibodies (Texas Red-conjugated goat anti-mouseIgG [1:50–1:80], Texas Red-conjugated donkey anti-rabbit

560 L E W I S and L E N T Z

IgG [1:50–1:100]). Cells were then washed three times for 5min each in 500 mM sodium chloride, 20 mM sodium phos-phate, pH 7.4; twice for 5 min in 150 mM sodium chloride, 10mM sodium phosphate, pH 7.4; once for 5 min in 120 mMsodium phosphate, pH 7.4; and briefly in 5 mM sodiumphosphate, pH 7.4 before mounting on slides with Fluorsavereagent (Calbiochem).

Preparations were viewed on a Zeiss Axiophot immuno-fluorescence microscope with 10x, 40x, or 63x objectives andwere photographed with Kodak T-MAX 400 film. Negativeswere scanned using a Polaroid SprintScan. The images wereadjusted for contrast and intensity using Adobe Photoshopsoftware and exported to Deneba Canvas software for as-sembly of figures and addition of labels. When double im-munofluorescence was performed, the two fluorescence im-ages were treated in the same manner. Some images weremerged by superimposing in the same position using AdobePhotoshop to determine the extent of co-localization of virusand tracers or markers. Figures were printed on a CodonicsNP-1600 printer.

ELECTRON MICROSCOPY

For electron microscopy, virus was adsorbed for 1 h at 48C tothe surface of hippocampal neurons grown on coverslips andcells warmed to 378C for 10 min. Cells were rinsed and fixedwith 1.5% glutaraldehyde, 4% sucrose in 100 mM cacodylatebuffer, pH 7.2 for 60 min. Coverslips were rinsed in cacody-late buffer and post-fixed for 60 min with cold 2% osmiumtetroxide. Subsequent processing, sectioning, and stainingwere performed as described by Mundigl et al. (1993). Sec-tions were examined with a Philips 410 electron microscope.

Results

INFECTION OF HIPPOCAMPAL NEURONS AND EFFECTOF LYSOSOMOTROPIC AGENTS

In order to determine the course of infection and replica-tion of virus in hippocampal neurons, cells were in-fected with rabies virus tissue culture supernatant andfixed and stained with FITC-anti-rabies antibody at 6and 12 h and at daily intervals up to 4 days. At 6 h, rabiesantigen was localized primarily in the perikaryon. Anti-gen occurred in small round accumulations in the cyto-plasm (Fig. 1A). A low-intensity diffuse fluorescencewas present in the cell body and proximal regions ofprocesses. At 12 h, the intensely fluorescent foci werepresent in both the cell body and the proximal regions ofprocesses (Fig. 1B). The foci of antigen within the soma(Figs 1Aand B) represent the counterpart of the classicalNegri bodies. Some varicosities along processes con-taining antigen appear to represent nerve terminals. By1 day, antigen completely filled the cell body and theentire extent of processes (Fig. 1C). Cells maintained asimilar appearance for 2–4 days post-infection (Fig. 2),after which they lysed and died. Infectivity and distri-bution of viral antigens was the same in cultures of dif-ferent ages.

Cells were infected in the presence of the lysosomo-

tropic weak bases chloroquine and NH4Cl. These agentslargely blocked infection of hippocampal neurons. Evenafter 2 days of infection, when untreated cells were filledwith viral antigen and were intensely fluorescent (Fig.2), treated cells showed only faint diffuse fluorescencelocalized largely to the cell body (Figs 1D and E).

ADSORPTION OF VIRUS TO CELL SURFACE

To coat the surface of cells with virus, cells were incu-bated in the presence of concentrated virus for 1 h at48C, a temperature at which endocytosis is inhibited. Tostudy uptake, cells were then warmed to 378C. To deter-mine whether incubation in the cold adversely affectedcells, cells were cooled for 1 h at 48C and warmed andcell morphology observed for 3 days. Cells appearednormal morphologically (Fig. 3A) and showed normaldistribution of the dendritic marker MAP-2 (Fig. 3B).Second, cells were adsorbed with virus in the cold andthen warmed. After 2 days, cells were fixed and stainedwith FITC-anti-rabies virus antibody. Cells were filledwith viral antigen (Fig. 3C), showing that cooling didnot affect the ability of cells to take up virus and sup-port viral replication.

To observe distribution of adsorbed virus, cells wereincubated with virus for 1 h at 48C and immediatelyfixed and stained. At the end of the adsorption period,viral antigen was localized uniformly over the surfaceof the cell body and processes (Fig. 3E). Fluorescencewas most intense over processes with larger proximalregions. Fine, thin processes extending throughout thefield showed low-intensity fluorescence. Some proc-esses, particularly the larger ones, also showed intensefluorescence in localized, round, or oval regions alongtheir length (Fig. 3E). In phase micrographs, these re-gions appeared as moderately dense swellings (Fig.3F). They were often present where processes contactedanother cell or crossed one another and are consideredto represent axon terminals synapsing on dendrites.Synapses on the cell soma and dendrites are numerousafter 7 days of culture (Fletcher et al., 1994). In oldercultures, processes were more numerous and highlybranched, and varicosities representing synapses weremore numerous (Figs 4B and 8C).

VIRUS LOCALIZATION AFTER WARMING TO 378C

Rabies virus antigen showed a similar pattern of local-ization from 5 to 30 min after adsorption of virus andwarming of cells to 378C in the presence of cyclohexi-mide to block replication. Antigen was localized mostintensely in small puncta in the cell body and along thebroader processes (Fig. 4A). There was less diffuse sur-face fluorescence than at the end of the adsorptionperiod. Swellings or varicosities along processes con-tained viral antigen (Figs 4A and B). Thin processescrossing the field showed low-intensity diffuse fluores-cence along their length. Glial cells, when present in the

Rabies virus entry 561

Fig. 1. Hippocampal neurons (6-day culture with glia) after infection with rabies virus tissue culture supernatant. (A) Fluores-cence microscopy after 6 h of infection and fixation and incubation with FITC-labelled anti-rabies antibody. Viral antigen ispresent primarily in small, round accumulations in the perikaryon. 3 470. (B) At 12 h of infection. Antigen is present inaccumulations in the cell body and proximal regions of processes. Some foci appear to be located on the surface of the processes.3 470. (C) At 24 h of infection. Viral antigen fills the cell body and is present in more distal regions of processes. 3 470.(D) Infection of 21-day culture for 2 days in the presence of NH4Cl. Low levels of antigen are present in the cell body. 3 305.(E) Infection of 15-day culture for 2 days in the presence of chloroquine. Antigen is present in low-intensity, diffuse accumula-tions in the cell body and in puncta along processes. 3 305.

cultures, contained less viral antigen than hippocampalneurons (Figs 4D and 6A).

Dendrites and axons can be distinguished morpho-logically (Bartlett & Banker, 1984; Cáceres et al., 1986;Cameron et al., 1991). Dendrites are thick at the base,taper distally, and are shorter than axons. The axonmay arise from the cell body or the proximal region ofa dendrite. Axons are thin at their origin, taper less, andare longer than dendrites. Based on the morphology ofprocesses, rabies virus appears to be most intenselylocalized to dendrites and much less so to axons at boththe end of the adsorption period (Fig. 3E) and afterwarming to 378C (Fig. 4A).

To confirm the identity of axons and dendrites anddifferential binding of rabies virus at these sites, cellswere adsorbed with virus, warmed for 5 min, fixed,and incubated with antibodies against rabies virus andMAP-2, a marker for dendrites, and NCAM L1, amarker for axons. In cells stained for rabies virus andMAP-2, rabies virus and MAP-2 were co-localized overthe cell body and larger processes with broad proximalregions (Fig. 4C). Small fine processes showed low-in-tensity staining for rabies virus and no staining forMAP-2 (Fig. 4C). In cells stained for rabies virus andNCAM L1, thick processes showing intense localiza-

tion of virus did not stain for NCAM L1 (Fig. 4D). Mostthin processes showing little or no staining for rabiesvirus were intensely stained for NCAM L1 (Fig. 4D).Thus, it is concluded that broad processes to whichvirus is most intensely localized are dendrites based onstaining for MAP-2 and absence of staining for NCAML1. Fine, thin processes that show little rabies viruslocalization are axons based on staining for NCAM L1and absence of staining for MAP-2.

LOCALIZATION OF RABIES VIRUS AND TRACERS ORMARKERS FOR ENDOSOMES

After adsorption of virus and incubation at 378C in thepresence of tracers, virus and the endosome tracers ormarkers Lucifer Yellow (Fig. 5A–C), transferrin receptor(Fig. 5D–F), Texas Red-dextran (Fig. 6A and B), and rho-damine-wheat germ agglutinin (Fig. 6C and D) werelocalized by fluorescence microscopy. The tracersshowed a similar pattern of localization from 5 to 30 minafter warming (Figs 5B, E and 6B, D). In the cell body,tracers were localized to small puncta that were often soclosely packed as to give an intense, uniform fluores-cence. Tracers were also present in puncta, and varicosi-ties that are considered to be synapses along dendrites.The greatest degree of co-localization of rabies virus and

FPO

Fig. 2. Hippocampal neurons (cultured 15 days) infected with rabies virus tissue culture supernatant, fixed after 2 days, andstained with FITC-labelled rabies antibody. Rabies virus completely fills the cell body and processes of the neurons. 3 200.


Fig. 3. (A, B) Hippocampal neuron (14-day culture) cooled at 48C for 1 h, warmed to 378C, and cultured for 3 days. (A) Phasemicrograph. Neuron exhibits normal morphology. Varicosities along processes represent sites of synaptic contact and are morenumerous in longer-term cultures. (B) Localization of MAP-2, a marker for dendrites, in the same cell. Dendrites are broadproximally where they arise from the cell body. 3 580. (C, D) Hippocampal neuron (13-day culture) adsorbed with rabies virusat 48C for 1 h, warmed to 378C, and cultured for 2 days. (C) Localization of rabies virus antigen. Cell bodies and processes arefilled with antigen, demonstrating that cooled and warmed cells support viral replication. (D) Corresponding phase micrograph.3 380. (E, F) Hippocampal neuron (32-day culture) fixed after 1 h incubation in rabies virus suspension at 48C. (E) Fluorescencemicroscopy after staining with FITC-rabies virus antibody. Antigen is present in an intense and diffuse localization on the cellbody and proximal regions of large processes. Antigen is intensely localized in small swellings or varicosities along processes(circles). Thin processes (arrows) crossing the field have less antigen and few swellings along them. (F) Corresponding phasemicrograph of cells. The sites of intense antigen localization on processes appear as moderately dense varicosities (circles). 3 465.

FPO

Fig. 4. (A, B) Neuron (82-day culture) adsorbed with virus in the cold and warmed to 378C for 30 min in the presence ofcycloheximide. (A) Viral antigen is present in puncta most numerous in the cell body and large processes arising from the soma(single arrows). Fluorescent varicosities occur along processes (circles). Fine, thin processes (double arrows) throughout the fieldshow faint diffuse fluorescence and few fluorescent puncta or swellings. (B) Corresponding phase micrograph. 3 440. (C) Co-lo-calization of rabies virus and MAP-2, a marker for dendrites. Neuron (11-day culture) adsorbed with virus in the cold andwarmed to 378C for 5 min. Rabies virus was localized with FITC-anti-rabies antibody and MAP-2 with anti-MAP-2 antibody andTexas Red-anti-mouse IgG. Photographs were taken at wavelengths for fluorescein and Texas Red, and the images were merged.In the merged image, rabies virus localization alone will appear green, MAP-2 alone will appear red, and sites of co-localizationwill appear yellow. Rabies virus and MAP-2 were co-localized in the cell body and some processes, particularly those with broadproximal regions. Rabies virus is weakly localized to a long thin process (arrows) lacking MAP-2. 3 665. (D) Co-localization ofrabies virus and NCAM L1, a marker for axons. Neuron (13-day culture) adsorbed with virus in the cold and warmed to 378C for5 min. Rabies virus was localized with FITC-anti-rabies antibody and NCAM L1 with anti-NCAM L1 antibody and TexasRed-anti-mouse IgG. Photographs were taken at wavelengths for fluorescein and Texas Red, and the images were merged. In themerged image, rabies virus localization alone will appear green, NCAM L1 alone will appear red, and sites of co-localization willappear yellow. Some co-localization of rabies virus and NCAM L1 occurs in the cell soma (right). Rabies virus but not NCAM L1is localized to larger processes. Fine, thin processes show localization of NCAM L1 but not rabies virus. Underlying glial cells(upper left) show low-intensity fluorescence for rabies virus antigen. 3 795.


FPO

Fig. 5. Co-localization of rabies virus and Lucifer Yellow (A, B, C) and rabies virus and transferrin receptor (D, E, F). (A–C).Hippocampal neuron (6-day culture over glia) adsorbed with rabies virus, warmed to 378C for 10 min in the presence of LuciferYellow, and stained with rabbit anti-Lucifer Yellow antibody and Texas Red-conjugated donkey anti-rabbit antibody andFITC-labelled anti-rabies antibody. (A) Localization of rabies virus antigen. (B) Localization of Lucifer Yellow. (C) Merged imageof (A) and (B). There is extensive co-localization (yellow) of rabies virus and Lucifer Yellow. Both are present in the cell bodyand in varicosities (arrowheads) along dendrites. An axon (Ax) shows faint fluorescence for rabies virus and Lucifer Yellow. 3590. (D–F) Hippocampal neuron (82-day culture) adsorbed with rabies virus, warmed to 378C for 30 min, and stained withmouse anti-tranferrin receptor antibody and Texas Red conjugated goat anti-mouse antibody and FITC-labelled anti-rabiesvirus antibody. (D) Localization of rabies virus antigen. (E). Localization of transferrin receptor. (F) Merged image of (D) and(E). Extensive co-localization (yellow) occurs in puncta in the cell body and dendrites. Co-localization occurs in varicositiesalong dendrites (circles). Some puncta along processes and in the cell body contain either rabies virus or transferrin receptor,but not both. Fine processes (axons) throughout the field show low-intensity fluorescence for rabies virus and none fortransferrin receptor. 3 485.


endosome tracers occurred in the puncta in the cell bodyand dendrites and in the varicosities along dendrites(Figs 5 and 6A–D). Some puncta contained either rabiesvirus or marker, but not both. Axons along their lengthshowed low-intensity fluorescence for tracers (Fig. 5).The pattern of localization of virus and tracers was thesame in cultures of all ages, although older culturescontained more processes and synapses. Glial cellscontained less tracer or marker than neurons. Some con-tained a few fluorescent puncta and a diffuse cyto-plasmic fluorescence (Fig. 6B). Rabies virus did not co-localize with tracers or markers in glial cells (Fig. 6A).

Lysosomes were localized with antibody against ly-sosomal glycoprotein. They appeared as small roundstructures most abundant in the perinuclear region ofthe cell body (Fig. 6F). There was little or no co-local-ization with rabies virus, which was present in differ-ent foci (Fig. 6E).

In order to determine whether virus-containing com-partments were acidified, neurons were adsorbed withrabies virus and incubated with LysoTracker Red, anacidotropic tracer. Although LysoTracker Red is re-ported not to undergo diffusion after aldehyde fixation,some diffusion of dye was observed after processing ofcells for rabies fluorescence. Because of this, afterwarming of cells and fixation, cells were viewed andphotographed for LysoTracker Red before significantdiffusion occurred. The coverslips were then processedwith FITC-rabies antibody and the same fields relo-cated and rephotographed for rabies fluorescence. Al-though cells underwent slight changes in cell shapeduring processing with rabies antibody, the same re-gions were easily identified. Co-localization of rabiesvirus and LysoTracker Red occurred in puncta in thecell body, especially clusters of fine puncta (Fig. 7).Similarly, co-localization was observed in the punctaand swellings along dendrites. Some puncta were reac-tive for viral antigen alone.

LOCALIZATION OF VIRUS AND SYNAPSIN I

Synapsin I was localized to puncta in the cell body andalong dendrites (Fig. 8B). The varicosities along den-drites contained synapsin I. Rabies virus co-localizedwith synapsin I in nearly all of these varicosities (Fig.8A–C). Elsewhere, there was less co-localization. Mostpuncta in the cell body contained either rabies virus orsynapsin I, but not both.

ELECTRON MICROSCOPY

Cells adsorbed with virus in the cold and warmed for10 min were observed by electron microscopy. Mor-phologically, cells, organelles, and vesicle-filled nerveterminals were intact (Fig. 9A). Virus particles wereround in cross section and bullet-shaped in longitudi-nal section. Virus particles were observed in coated pitslocated at synapses (Fig. 9B) and along processes (Fig.9C). They also occurred within coated vesicles in axonterminals and processes (Fig. 9D). Virus particles also

were present in larger membranous vacuoles in nerveterminals (Fig. 9A and E). Smaller intracellular vesiclesand vacuoles usually contained a single virus particle,but larger vacuoles often contained several virus parti-cles (Fig. 9D).

Discussion

REPLICATION IN HIPPOCAMPAL NEURONS

The present observations confirm that hippocampalneurons are highly susceptible to rabies virus infection.Cells support replication of virus, as evidenced by rapidincrease in the amount of cytoplasmic antigen through-out the cell body and processes. At 1–3 days of infection,antigen completely fills the cell body, after which cellslyse and die. When neurons were adsorbed with highconcentrations of virus in the cold, antigen was presentin a uniform distribution over the surface of the cellbody, dendrites, and axon terminals. This distributionmost likely reflects a greater number of viral bindingsites at these regions. Although the cellular receptor forrabies virus in the CNS is unknown, it has been postu-lated that the nicotinic acetylcholine receptor acts as arabies receptor on skeletal muscle (Lentz et al., 1982).Hippocampal neurons contain neuronal nicotinic ace-tylcholine receptors and a-bungarotoxin-binding recep-tors (Alkondon et al., 1994; Barrantes et al., 1994; Adams& Freedman, 1997), and it is possible that these couldserve as rabies virus receptors in the central nervoussystem.

RABIES VIRUS ENTRY INTO ACIDIC ENDOSOMES

When hippocampal neurons with adsorbed virus werewarmed to 378C for times up to 30 min, virus waspresent in small puncta located in the cell body anddendrites and within swellings or varicosities along thedendrites representing axon terminals synapsing ondendrites. Axons along their length contained less viralantigen and few fluorescent swellings. The identity ofdendrites and axons was confirmed by the localizationof MAP-2, a marker for dendrites, and NCAM L1, amarker primarily located on axons. Rabies virus co-lo-calized with MAP-2 on processes with broad proximalregions but did not co-localize with NCAM L1, whichwas located on fine thin processes.

To determine whether virus entered endosomes,cells were adsorbed with virus and incubated withtracers for endosomes. It was found that within 5 minof warming to 378C, there was significant co-localiza-tion of virus with the markers Lucifer Yellow, transfer-rin receptor, Texas Red-dextran, and rhodamine-wheatgerm agglutinin. Virus and markers were co-localizedprimarily in puncta in the cell body and dendrites andin the nerve terminals on dendrites. Based on the ob-served co-localization of virus with tracers for earlyendosomes, it is concluded that shortly after uptake,virus resides in an endosome compartment. Thus, ra-


bies virus enters hippocampal neurons by a process ofadsorptive endocytosis. This conclusion is supportedby electron microscopic observations showing pres-ence of virus particles in coated pits at the cell surface.Within the cells, particles were present in coated ves-

icles and larger membrane compartments morphologi-cally similar to endosomes in other cell types. Virusalso co-localized with LysoTracker Red, a probe foracidic compartments, indicating that at least some ofthe virus-containing endosomes are acidified.

Fig. 7. Localization of rabies virus and LysoTracker Red. Cells (28-day culture) were adsorbed with rabies virus, incubated withLysoTracker Red, and fixed 5 min after warming to 378C. Cells were then photographed with fluorescence for LysoTraker Red(C) and by phase microscopy (D). They were then processed with FITC-rabies antibody and the same field reexamined withfluorescence for fluorescein (A) and phase microscopy (B). Clusters of puncta in the cell body and proximal dendrites showco-localization of virus and LysoTracker Red. Along the processes, puncta and swellings show co-localization (circles). 3 410.

Fig. 6. Co-localization of rabies virus and Texas Red-dextran (A, B), rabies virus and rhodamine-wheat germ agglutinin (C, D),and rabies virus and lysosomal glycoprotein (E, F). (A, B) Hippocampal neuron (13-day culture over glia) adsorbed V withrabies virus, warmed to 378C for 30 min in the presence of Texas Red-dextran, and stained with FITC-rabies antibody. (A)Localization of rabies virus. (B) V Localization of Texas Red dextran. Co-localization is most prominent in varicosities alongdendrites (circles). The neuron overlies a flattened glial cell (G) that shows diffuse, low-intensity fluorescence for dextran butnot for rabies virus. 3 435. (C, D). Neuron (24-day culture) adsorbed with rabies virus, warmed to 378C in the presence ofrhodamine-wheat germ agglutinin for 30 min, and stained with FITC-rabies antibody. (C) Localization of rabies virus antigen.(D) Localization of wheat germ V agglutinin. Puncta in the cell body and varicosities along processes (circles) contain bothrabies virus and wheat germ agglutinin. 3 565. (E, F). Neuron (40-day culture) adsorbed with rabies virus, warmed to 378C for30 min, and stained with mouse anti-lysosomal glycoprotein antibody and Texas Red-conjugated goat anti-mouse antibodyand FITC-rabies antibody. (E) Localization of rabies virus. (F) Localization V of lysosomal glycoprotein. Lysosomal glycoproteinis localized to small round structures most numerous in the cell body. Rabies virus is localized to puncta in the processes andcell body and is not co-localized with lysosomal glycoprotein. 3 665.


The present observations are in agreement with pre-vious studies showing high endocytic activity in thesomatodendritic domain of the neuron (Cameron et al.,1991; Parton et al., 1992). They are also consistent withthe existence of a dendroaxonal transcytotic routingpathway in neurons (de Hoop et al., 1995). In the axon,internalization of markers takes place in the axon termi-nals. There is little virus binding or endocytic activityalong the axon shaft. At sites of entry, endocytic markersenter tubulovesicular endosomes. Sites of rabies virusentry coincide with those for endocytic tracers, withentry occurring in the somatodendritic domain andaxon terminals. In the nerve terminals, endosomes arebelieved to be involved in the recycling of synaptic ves-icles (De Camilli & Takei, 1996). Because these en-dosomes are also labelled by endocytic tracers, theymay serve as a trafficking station for both recycling sy-naptic vesicles and internalized substances. Further evi-dence for the entry of virus by adsorptive endocytosis isthe inhibition of infection by lysosomotropic agents.Chloroquine and NH4Cl inhibited infection of hippo-campal neurons by rabies virus. These agents have alsobeen shown to inhibit infection of CER cells (Superti etal., 1984), NS 20 murine neuroblastoma cells (Tsiang &Superti, 1984), and IMR-32 human neuroblastoma cells(Lentz et al., 1997) by rabies virus. These agents raise thepH of endosomes and lysosomes, preventing fusion ofthe viral envelope with the endosomal membrane.

SYNAPSIN II

Rabies virus and synapsin I (Fletcher et al., 1991), asynaptic vesicle protein, were localized. In most hippo-campal neurons, synapsin I was localized in puncta inthe cell body and along processes. The swellings alongdendrites contained synapsin I, further indicating thatthese regions represent axon terminals on dendrites.Rabies virus showed the greatest co-localization withsynapsin in these varicosities, showing that one site ofentry of rabies virus into hippocampal neurons is theaxon terminal. In addition, because virus in the termi-nals co-localized with endocytic tracers, virus residesin endosomes in the nerve terminal. It is concluded thatvirus entry into cultured hippocampal neurons takesplace primarily at the cell soma and dendrites and ataxon terminals.

The subsequent fate of the virus after entry of virusinto endosomes is unknown. Either the virus fuseswith the endosome at the site of uptake or the en-dosome containing the virion is transported to otherregions of the neuron prior to fusion. Virus andLysoTracker Red, a probe for acidic compartments,were co-localized, indicating that some virus-contain-ing endosomes are acidified. Because fusion of rabiesvirus with membranes occurs at low pH (Mifune et al.,1982; Perrin et al., 1982) and acidic endosomes are pre-sent in both cell body and processes, fusion of viruswith endosomes potentially can occur at these regions

Fig. 8. Co-localization of rabies virus (A) and synapsin I (B).Neurons (82-day culture) were adsorbed with rabies virus andwarmed to 378C for 30 min. Synapsin was localized with rab-bit anti-synapsin antibody and Texas Red-conjugated donkeyanti-rabbit antibody and rabies virus with FITC-rabies anti-body. Co-localization occurs primarily in varicosities alongprocesses (circles). Some puncta in the cell body and dendritesshow co-localization, but others do not. (C) Correspondingphase micrograph. The varicosities along processes (syn-apses) showing co-localization for rabies virus and synapsin I(circles) appear as moderately dense swellings along proc-esses. 3 435.


Fig. 9. Electron micrographs of 45-day hippocampal neurons warmed for 10 min at 378C after adsorption of virus. (A) Axonterminals and processes. Terminals contain numerous synaptic vesicles, and processes contain both synaptic vesicles andmicrotubules. Extracellular virus particles are located adjacent to an invagination on a terminal (1). Within the terminals,membranous vacuoles are present (2). Some of the intra-axonal membrane compartments contain virus particles (3). 3 71,500.(B) Synapse containing pre-and post-synaptic densities and synaptic vesicles. A coated pit (arrow) containing two virusparticles is located on the outer surface of the axon terminal. 3 102,000. (C) Coated pit (arrow) containing a virus particlelocated on the surface of a nerve cell process. 3 78,500. (D) Virus particle present within a coated vesicle (arrow) in a process.3 115,500. (E) Virus particles within a vacuole located in a vesicle-filled terminal. 3 72,000.


near the site of uptake. If fusion occurs in axon termi-nals, the viral nucleocapsid must be transported to theperikaryon in order for replication to occur. Studieshave shown that pseudorabies virus (swine a-herpes-virus) and herpes simplex virus are transported retro-gradely as free nucleocapsids (Marchand & Schwab,1986; Lycke et al., 1988; Card et al., 1993). Anterogradetransport of herpes simplex virus nucleocapsids alsooccurs (Penfold et al., 1994). Transport of the nucleocap-sids is mediated by microtubules (Sodeik et al., 1997).While these studies indicate that free nucleocapsids canbe transported, they may not be applicable to rabiesbecause herpes viruses enter cells by fusion with theplasma membrane. Transport of virus within an en-dosome would also provide a mechanism for relocali-zation of the viral genome to a specific region of cyto-plasm (Marsh & Helenius, 1989).

In vivo, rabies virus is carried from the periphery tothe spinal cord by retrograde transport within periph-eral nerves. Infection then ascends in the spinal cord tothe brain, where early infection is selective for certain

regions such as rhinencephalic structures and cerebel-lar Purkinje cells. Subsequently, infection spreads tomost regions of the brain (Johnson, 1965). The presentfindings indicate that one mechanism of virus spreadfrom neuron to neuron is by endocytosis of extracellu-lar virus particles by axon terminals. In addition, virusenters dendrites and the cell soma of cultured hippo-campal neurons. Thus, it is possible that virus shedfrom infected cells or budding from neurons could betaken up by adjacent uninfected cells at any of thesesites.

Acknowledgments

We would like to thank Olaf Mundigl, Laurie Daniell,and Bettina Winckler for preparing the hippocampalcultures used in these studies and Pietro De Camilliand Ira Mellman for advice and suggestions. This workwas supported by grants NS34275 (to T. Lentz) andNS36251 (to P. De Camilli) from the National Institutesof Health.

References

A D A M S , C . E . & F R E E D M A N , R . (1997) Nicotinic an-tagonist a-bungarotoxin binding to rat hippocampal neu-rons containing nitric oxide synthase. Brain Research 776,111–6.

A L K O N D O N , M . , R E I N H A R D T, S . , L O B R O N , C . ,H E R M S E N , B . , M A E L I C K E , A . & A L B U Q U E R -Q U E , E . X . (1994) Diversity of nicotinic aceylcholinereceptors in rat hippocampal neurons. II. The rundownand inward rectification of agonist-elicited whole-cellcurrents and identification of receptor subunits by in situhybridization. Journal of Pharmacology and ExperimentalTherapeutics 271, 494–506.

B A E R , G . M . & L E N T Z , T. L . (1991) Rabies pathogenesisto the central nervous system. In: The Natural History ofRabies (edited by B A E R , G . M . ), pp. 105–20. Boca Ra-ton: CRC Press.

B A N K E R , G . A . & C O WA N , W. M . (1977) Rat hippocam-pal neurons in dispersed cell culture. Brain Research 126,397–425.

B A R R A N T E S , G . E . , W E S T W I C K , J . & W O N N A -C O T T, S . (1994) Nicotinic acetylcholine receptors in pri-mary cultures of hippocampal neurons: pharmacologyand Ca1 permeability. Biochemical Society Transactions 22,294S.

B A RT L E T T, W. P. & B A N K E R , G . A . (1984) An electronmicroscopic study of the development of axons and den-drites by hippocampal neurons in culture. I. Cells whichdevelop without intercellular contacts. Journal of Neuro-science 4, 1944–53.

C Á C E R E S , A . , B A N K E R , G . A . & B I N D E R , L . (1986)Immunocytochemical localization of tubulin and micro-tubule-associated protein 2 during the development ofhippocampal neurons in culture. Journal of Neuroscience6, 714–22.

C A M E R O N , P. L . , S Ü D H O F, T. C . , J A H N , R . & D EC A M I L L I , P. (1991) Colocalization of synaptophysin

with transferrin receptors: implications for synaptic ves-icle biogenesis. Journal of Cell Biology 115, 151–64.

C A R D , J . P. , R I N A M A N , L . , LY N N , R . B . , L E E , B . -H . , M E A D E , R . P. , M I S E L I S , R . R . & E N Q U I S T,L . W. (1993) Pseudorabies virus infection of the rat cen-tral nervous system: ultrastructural characterization ofviral replication, transport, and pathogenesis. Journal ofNeuroscience 13, 2525–39.

C H A R LT O N , K . M . & C A S E Y, G . A . (1979) Experimentalrabies in skunks. Immunofluorescence light and electronmicroscopic studies. Laboratory Investigation 41, 36–44.

D E C A M I L L I , P. & TA K E I , K . (1996) Molecular mecha-nisms in synaptic vesicle endocytosis and recycling. Neu-ron 16, 481–6.

D E H O O P, M . , V O N P O S E R , C . , L A N G E , C . , I K -O N E N , E . , H U N Z I K E R , W. , & D O T T I , C . G . (1995)Intracellular routing of wild-type and mutated poly-meric immunoglobulin receptor in hippocampal neuronsin culture. Journal of Cell Biology 130, 1447–59.

D O T T I , C . G . , S U L L I VA N , C . A . & B A N K E R , G . A .(1988) The establishment of polarity by hippocampalneurons in culture. Journal of Neuroscience 8, 1454–68.

F L E T C H E R , T. L . , C A M E R O N , P. , D E C A M I L L I , P. &B A N K E R , G . (1991) The distribution of synapsin I andsynaptophysin in hippocampal neurons developing inculture. Journal of Neuroscience 11, 1617–26.

F L E T C H E R , T. L . , D E C A M I L L I , P. & B A N K E R , G .(1994) Synaptogenesis in hippocampal cultures: evi-dence indicating that axons and dendrites become com-petent to form synapses at different stages of neuronaldevelopment. Journal of Neuroscience 14, 6695–706.

G A U D I N , Y. , T U F F E R E A U , C . , S E G R E TA I N , D . ,K N O S S O W, M . & F L A M A N D , A . (1991) Reversibleconformational changes and fusion activity of rabies vi-rus glycoprotein. Journal of Virology 65, 4853–9.

I WA S A K I , Y. (1991) Spread of virus within the central nerv-


ous system. In: The Natural History of Rabies (edited byBAER, G. M.), pp. 121–32. Boca Raton: CRC Press.

I WA S A K I , Y. & C L A R K , H . F. (1975) Cell to cell transmis-sion of virus in the central nervous system II. Experimen-tal rabies in mouse. Laboratory Investigation 33, 391–9.

I WA S A K I , Y. , L I U , D . - S . , YA M A M O T O , T. & K O N N O ,H . (1985) On the replication and spread of rabies virus inthe human central nervous system. Journal of Neuropathol-ogy and Experimental Neurology 44, 185–95.

J O H N S O N , R . T. (1965) Experimental rabies. Studies ofcellular vulnerability and pathogenesis using fluorescentantibody staining. Journal of Neuropathology and Experi-mental Neurology 24, 662–74.

L E N T Z , T. L . (1994) Molecular interaction of viruses withhost cell receptors. In: Adhesion Molecules (edited byW E G N E R , C . D . ), pp 223–51. London: Academic Press.

L E N T Z , T. L . , B U R R A G E , T. G . , S M I T H , A . L . ,C R I C K , J . & T I G N O R , G . H . (1982) Is the acetyl-choline receptor a rabies virus receptor? Science 215,182–4.

L E N T Z , T. L . , C H E S T E R , J . , B E N S O N , R . J . J . ,H AW R O T, E . , T I G N O R , G . H . & S M I T H , A . L .(1985) Rabies virus binding to cellular membranes meas-ured by enzyme immunoassay. Muscle & Nerve 8, 336–45.

L E N T Z , T. L . , F U , Y. & L E W I S , P. (1997) Rabies virusinfection of IMR-32 human neuroblastoma cells and ef-fect of neurochemical and other agents. Antiviral Re-search 35, 29–39.

LY C K E , E . , H A M A R K , B . , J O H A N S S O N , M . , K R O -T O C H W I L , A . , LY C K E , J . & S V E N N E R H O L M , B .(1988) Herpes simplex virus infection of the human sen-sory neuron. An electron microscopy study. Archives ofVirology 101, 87–104.

M A R C H A N D , C . F. & S C H WA B , M . E . (1986) Binding,uptake and retrograde axonal transport of herpes virussuis in sympathetic neurons. Brain Research 383, 262–70.

M A R S H , M . & H E L E N I U S , A . (1989) Virus entry intoanimal cells. Advances in Virus Research 36, 107–51.

M I F U N E , K . , O H U C H I , M . & M A N N E N , K . (1982)Hemolysis and cell fusion by rhabdoviruses. FEBS Let-ters 137, 293–7.

M U N D I G L , O . , M AT T E O L I , M . , D A N I E L L , L . ,T H O M A S - R E E T Z , A . , M E T C A L F, A . , J A H N , R . &D E C A M I L L I , P. (1993) Synaptic vesicle proteins andearly endosomes in cultured hippocampal neurons: dif-ferential effects of Brefeldin A in axons and dendrites.Journal of Cell Biology 122, 1207–21.

M U R P H Y, F. A . (1977) Rabies pathogenesis. Brief review.Archives of Virology 54, 279–97.

M U R P H Y, F. A . , H A R R I S O N , A . K . , W I N N , W. C . &B A U E R , S . P. (1973) Comparative pathogenesis of ra-bies and rabies-like viruses. Infection of the central nerv-ous system and centrifugal spread of virus to peripheraltissues. Laboratory Investigation 29, 1–16.

PA RT O N , R . G . , S I M O N S , K . & D O T T I , C . G . (1992)Axonal and dendritic endocytic pathways in culturedneurons. Journal of Cell Biology 119, 123–37.

PA U L S O N , J . C . (1985) Interactions of animal viruses withcell surface receptors. In: The Receptors (edited byCONN, P. M.), pp. 131–219. Orlando: Academic Press.

P E N F O L D , M . E . T. , A R M AT I , P. & C U N N I N G H A M ,A . L . (1994) Axonal transport of herpes simplex virionsto epidermal cells: evidence for a specialized mode ofvirus transport and assembly. Proceedings of the NationalAcademy of Sciences USA 91, 6529–33.

P E R R I N , P. , P O RT N O Ï , D . & S U R E A U , P. (1982) Étudede l’adsorption et de la pénétration du virus rabique:interactions avec les cellules BHK21 et des membranesartificielles. Annales de Virologie (Institut Pasteur) 133E,403–22.

S O D I E K , B . , E B E R S O L D , M . W. & H E L E N I U S , A .(1997) Microtubule-mediated transport of incoming her-pes simplex virus 1 capsids to the nucleus. Journal of CellBiology 136, 1007–21.

S U P E RT I F. , D E R E R M . , & T S I A N G , H . (1984) Mecha-nism of rabies virus entry in CER cells. Journal of GeneralVirology 65, 781–9.

T S I A N G , H . & S U P E RT I , F. (1984) Ammonium chlorideand chloroquine inhibit rabies virus infection in neu-roblastoma cells. Archives of Virology 81, 377–82.


Rabies virus entry into cultured rat hippo campalneurons

Documents

Transcript of Rabies virus entry into cultured rat hippo campalneurons