In Vivo Transduction of Central Neurons Using Recombinant Sindbis Virus

Abstract

A new recombinant virus which labeled the infected neurons in a Golgi stainlike fashion was developed. The virus was based on a replication-defective Sindbis virus and was designed to express green fluorescent protein with a palmitoylation signal (pal-GFP). When the virus was injected into the ventrobasal thalamic nuclei, many neurons were visualized with the fluorescence of palGFP in the injection site. The labeling was enhanced by immunocytochemical staining with an antibody to green fluorescent protein to show the entire configuration of the dendrites. Thalamocortical axons of the infected neurons were also intensely immunostained in the somatosensory cortex. In contrast to palGFP, when DsRed with the same palmitoylation signal (palDsRed) was introduced into neurons with the Sindbis virus, palDsRed neither visualized the infected neurons in a Golgi stain-like manner nor stained projecting axons in the cerebral cortex. The palDsRed appeared to be aggregated or accumulated in some organelles in the infected neurons. Anterograde labeling with palGFP Sindbis virus was very intense, not only in thalamocortical neurons but also in callosal, striatonigral, and nigrostriatal neurons. Occasionally there were retrogradely labeled neurons that showed Golgi stain-like images. These results indicate that palGFP Sindbis virus can be used as an excellent anterograde tracer in the central nervous system.

Keywords

Sindbis virus green fluorescent protein DsRed palmitoylation Golgi stain-like labeling neuronal tracer rat brain

Sindbis virus, which is an enveloped virus with a positive-strand RNA genome and which belongs to the alphavirus genus of the Togavirus family, causes acute encephalomyelitis in mice (for review cf. Griffin 1998). The virus has the advantages of a high infection efficiency to a broad range of hosts, including neural cells, and the ability to express large amounts of gene products (for review cf. Xiong et al. 1989; Schlesinger 1993; Piper et al. 1994; Strauss and Strauss 1994; Lundstrom 1999). A replication-defective Sindbis virus expression system has been developed for transduction of mammalian cells and is known to generate high levels of protein expression (Bredenbeek et al. 1993). Recently, the Sindbis virus expression system was used for in vivo delivery of a reporter gene to neurons in the neostriatum, hippocampus, and cerebral cortex (Altman-Hamamdzic et al. 1997; Gwag et al. 1998; Lendvai et al. 2000).

The introduction of palmitoylation signals into a reporter protein was demonstrated to be effective in targeting the protein to the cell membrane of cultured cells using adenovirus vectors (Moriyoshi et al. 1996). Okada et al. (1999) and Tamamaki et al. (2000), using retrorvirus and adenovirus vectors, respectively, revealed that green fluorescent protein (GFP) tagged with a palmitoylation site (palGFP) was useful in visualizing the infected neurons in a Golgi stain-like manner. The palGFP expression is expected to be more useful in visualizing neurons when it is combined with the high-expression activity of the Sindbis virus system. In this study we developed a recombinant Sindbis virus that was designed for the infected cells to express membrane-targeted fluorescent proteins, palGFP or DsRed (red fluorescent protein) tagged with a palmitoylation site (palDsRed), and applied the virus to in vivo gene delivery of the membrane-targeted fluorescent protein in the rat brain.

Materials and Methods

These experiments were approved by the Committee for Animal Care and Use of the Graduate School of Medicine at Kyoto University and that for Recombinant DNA Study in Kyoto University.

Construction of Recombinant Sindbis Viruses

The reporter proteins were designed as the fusion protein of N-terminal icosapeptide (palmitoylation signal) of growth-associated protein-43 (GAP43) and enhanced GFP (Clontech; Palo Alto, CA) or DsRed (Clontech). The cDNA fragment encoding GFP with the palmitoylation signal (palGFP) was obtained from plasmid pGG16 (expression vector of palGFP; a generous gift from Dr. T. Kagawa, National Institute for Physiology, Okazaki, Japan; Tamamaki et al. 2000). The cDNA fragment encoding DsRed with the palmitoylation signal (palDsRed) was constructed as follows (Figure 1). After digestion of pDsRed1-N1 (Clontech) with restriction enzyme PstI, the resulting PstI-PstI fragment containing the 5′-end sequence of DsRed was inserted into the PstI site of pBlueScript II which lost the XhoI site (Stratagene; La Jolla, CA) and was cloned as pBS-DsRedN. The XhoI-NcoI fragment containing the sequence for the palmitoylation signal was obtained from pGG16 and replaced the XhoI-NcoI fragment of the pBS-DsRedN, resulting in pBS-palDsRedN. The XhoI-PstI fragment of the original pDsRed1-N1 was replaced with the XhoI-PstI fragment of pBS-palDsRedN, resulting in the expression vector p-palDsRed1-N1.

pGG16 or p-palDsRed1-N1 was double digested with XhoI and NotI, and a DNA construct containing the sequence for palGFP or palDsRed was blunt-ended. The construct was inserted into the PmaCI site of pSinRep5 (Invitrogen; Carlsbad, CA). The production of Sindbis virus was performed according to the instructions with the Sindbis Expression System (Invitrogen), as follows. The capped transcript of recombinant RNA was synthesized from the pSinRep5 containing the construct (pSinRep5-palGFP or pSinRep5-palDsRed). Sindbis viral particles were obtained by co-transfecting baby hamster kidney (BHK) cells electrophoretically with the recombinant RNA transcript and DH (26S) 5′SIN helper RNA encoding the structural protein. The viral particles in the culture supernatant were concentrated to a 20–55% sucrose interface by ultracentrifugation (160,000 × g, 90 min). The titer was adjusted to 2 × 10¹⁰ infective U/ml and the virus was stored in aliquots at −80C until used for delivery to brain tissue. The resulting Sindbis virus was replication-deficient and had the least chance of production of parent viral particles in the infected cells (Bredenbeek et al. 1993).

Cell Culture and Transfection of Expression Vectors

Chinese hamster ovary (CHO) cells (American Type Culture Collection CCL-61) were grown in Dulbecco's modified Eagle's medium (Gibco, Gaithersburg, MD, and Invitrogen; Rockville, MD) supplemented with 10% (v/v) fetal bovine serum. One microgram of expression vector with cytomegalovirus (CMV) promoter, pEGFP-N3, pGG16, pDsRed1-N1, or p-palDsRed1-N1, was mixed with 3.6 μl of FuGENE 6 reagent (Roche; Basel, Switzerland) in 100 μl of serum-free medium and then incubated with CHO cells at 80% confluency. The cells were fixed with 4% formaldehyde in 0.1 M sodium phosphate (pH 7.0) 36 hr after the transfection. The nuclei of some fixed cells were stained with 1 μg/ml of 4′,6-diamidino-2-phenylindole-2HCl (DAPI) in 5 mM phosphate-buffered 0.9% (w/v) saline, pH 7.4 (PBS). The cells were coverslipped and observed under an Axiophot fluorescence microscope (Zeiss; Oberkochen, Germany) with an appropriate filter set for GFP (excitation 450–490 nm, emission 515–565 nm), DsRed (excitation 540–552, emission ≥575 nm), or DAPI (excitation 359–371, emission 397–490 nm).

Injection of Viruses, Fixation, and Immunohistochemistry

Thirty-six Wistar rats (200–300 g) were deeply anesthetized with chloral hydrate (350 mg/kg body weight). The virus (1–2 μl of 2 × 10¹⁰ infectious U/ml) was injected through glass micropipettes equipped with Picospritzer II (General Valve Corporation; East Hanover, NJ) into the ventrobasal nuclei of the thalamus, somatosensory cortex, neostriatum, and substantia nigra. The rats were allowed to survive for 4.5 hr to 28 days after the injection.

The rats were anesthetized again with chloral hydrate (350 mg/kg body weight) and perfused transcardially with 200 ml of PBS, followed by 300 ml of 4% (w/v) formaldehyde in 0.1 M sodium phosphate, pH 7.0. The brains were removed, cut into several blocks, and placed for 2 hr at 4C in the same fixative. After cryoprotection with 30% (w/w) sucrose in PBS, the blocks were cut into 40-μm-thick frontal sections on a freezing microtome. Some sections were mounted on a glass slide, coverslipped with 50% (v/v) glycerol and 2.5% (w/v) triethylene diamine (anti-fading reagent) in PBS, and observed with the epifluorescence microscope as described above.

The other sections were immunostained with an affinity-purified rabbit antibody to GFP (Tamamaki et al. 2000) or a rabbit antibody to DsRed (Clontech). The sections were incubated overnight at room temperature with 0.1 μg/ml of anti-GFP antibody or 0.2 μg/ml of anti-DsRed antibody in PBS containing 0.3% (v/v) Triton X-100, 0.5% (w/v) Λ-carrageenan, and 0.5% (v/v) donkey serum (PBS-XCD). After a rinse with PBS containing 0.3% (v/v) Triton X-100 (PBS-X), they were incubated for 1 hr with 10 μg/ml of biotinylated anti-rabbit IgG donkey antibody (Chemicon; Temecula, CA) in PBS-XCD and then for 1 hr with avidin-biotinylated peroxidase complex (ABC Elite; Vector, Burlingame, CA) in PBS-X. Finally, the sections were reacted for 10–30 min with 0.2% (w/v) diaminobenzidine-4HCl and 0.003% (v/v) H₂O₂ in 50 mM Tris-HCl, pH 7.6. The sections were mounted on a glass slide, dehydrated with an ethanol series, cleared in xylene, and coverslipped with organic mounting medium MX (Matsunami; Kishiwada, Japan). Adjacent sections were stained with Cresyl violet for cytoarchitecture.

Results

The differences in intracellular localization between GFP, palGFP, DsRed and palDsRed were examined with CHO cultured cells and expression vectors pEGFP-N3, pGG16, pDsRed1-N1, and p-palDsRed1-N1, all of which used a CMV promoter for protein synthesis in mammalian cells (Figure 2). palGFP appeared to be more associated with cell membranes (arrowheads in Figure 2a) than GFP (Figure 2b), while palDsRed showed aggregated fluorescence in the cytoplasm (Figure 2c). Under a Nomarski differential interference contrast microscope, the form of palDsRed-expressing cells was spherical rather than spindle-shaped (arrowheads in Figure 2c). Double fluorescence observation showed that both palGFP and palDsRed were not located in DAPI-positive cell nuclei (not shown). In contrast, both GFP- and DsRed-expressing CHO cells were spindle-shaped and displayed diffuse fluorescence throughout the cytoplasm and cell nuclei (Figures 2b and 2d). This indicates that the addition of exactly the same palmitoylation signal had clearly different effects on the respective soluble molecules, GFP and DsRed.

It was surprising to us that CHO cultured cells could not be infected with the Sindbis virus vectors (not shown). This might be caused by the very low susceptibility of CHO cells to Sindbis virus, although Sindbis virus was reported to have a broad host range in animal cells (Xiong et al. 1989).

Because the gene expression by Sindbis virus system was reported to reach peak between 24 and 48 hr after infection in the brain (Altman-Hamamdzic et al. 1997; Gwag et al. 1998), we first observed the injection sites 36 hr after injection of the virus into the rat thalamus (Figure 3). Many large cells, probably neurons, showed intense fluorescence of GFP (Figure 3a) or DsRed (Figure 3c) in the ventrobasal thalamic nuclei. palGFP was distributed not only in the neuronal cell body but also throughout the fine dendritic processes (Figure 3b), whereas palDsRed was mostly localized in the neuronal cell bodies. Both proteins were not found in smaller cell bodies, such as those of glial cells. These results suggest that neurons were preferentially infected with Sindbis virus at 36 hr after infection.

To determine the best survival time of rats for more precise neuronal labeling, we analyzed the change of GFP immunoreactivity at the injection sites of the thalamus from 4.5 hr to 28 days after inoculation of palGFP Sindbis virus (Figure 4). Many small immunoreactive cells (arrows in Figure 4a) and amorphous or bushy deposits showing intense immunoreactivity (arrowheads) were found in the injection site 4.5 hr after the injection. Nine hours after the injection, larger immunoreactive cells with long processes, probably neurons, appeared among the amorphous immunopositive staining (Figure 4b). From 18 to 72 hr after the injection, immunoreactivity was very intense and was mostly confined to neuronal structures such as cell bodies, dendrites, and axons (Figures 4c-4e). However, at 7 days of survival, neurons showed degenerative changes, such as beaded dendrites (arrowheads in Figure 4f) and shrinkage of the cell bodies (arrow). Immunoreactive neurons had almost completely disappeared by 28 days after the injection, except for remnant punctate immunoreactivity (arrowheads in Figure 4h). Therefore, survival times of 18 hr to 3 days (72 hr) were suitable for neuronal labeling in the injection sites.

The labeling of palDsRed was compared with that of palGFP by co-injection of the same titers of palGFP and palDsRed Sindbis viruses (Figures 5a-5d). Thirty-six hours later at the injection site in the ventrobasal thalamic nuclei, DsRed immunoreactivity was localized to the cell bodies and appeared to be aggregated or accumulated in some cytoplasmic structures (Figure 5c). Although no DsRed immunoreactivity was observed in the cerebral cortex, many GFP-immunoreactive axon fibers were observed in the somatosensory areas (Figures 5b-5d). When the sections were observed without immunostaining under the fluorescence microscope, a few neurons were found to express both palGFP and palDsRed at the injection site, although their fluorescence was weaker than that of neurons expressing either palGFP or palDsRed alone (not shown). The latter finding might suggest a competition between palGFP and palDsRed protein synthesis. In those co-expressing neurons, palDsRed was localized to cell bodies, whereas palGFP was distributed in fine dendrites as well as in cell bodies, as shown in Figure 3b.

When the palDsRed virus was injected into the cerebral cortex (Figure 5f), intense DsRed immunoreactivity was mostly found in the cell bodies and much weaker immunoreactivity was found along the apical dendrites (arrows). In contrast, when the palGFP virus was injected into the cerebral cortex (Figure 5e), many intensely immunoreactive pyramidal cells showed Golgi stain-like images with heavily spiny dendrites (arrow) and many axon collaterals (arrowheads) around the injection site. These results suggest that palDsRed is not as efficiently targeted to cell membranes as palGFP but is mostly aggregated or accumulated in some cytoplasmic structures.

The characteristics of palGFP Sindbis virus as an anterograde tracer were examined in several brain regions (Figure 6). When the virus was injected into the motor cortex, many projecting axon fibers were labeled in the corpus callosum (arrows in Figure 6a) and the terminals were in the contralateral motor cortex (Figure 6b). After injection of the virus into the neostriatum (Figure 6c) or substantia nigra (Figure 6f), intensely GFP-immunoreactive axon fibers and terminals were found in the reticular part of the substantia nigra (Figures 6d and 6e) or in the neostriatum (Figure 6h), respectively. In the case of nigrostriatal projection, patchy distribution of labeled axon fibers in the neostriatum was observed as reported previously (Gerfen et al. 1987). These results indicate that pal-GFP Sindbis virus can be used as an anterograde tracer in many types of neurons, such as glutamatergic, GABAergic, and dopaminergic ones.

In addition to intense anterograde labeling with palGFP Sindbis virus, retrogradely labeled neurons were occasionally seen (Figure 7). Five to seven pyramidal neurons in Layer VI (Figure 7a) or four to six medium-sized spiny neurons in the caudate putamen (Figure 7b) were labeled retrogradely after injection of 2 μl of 2 × 10¹⁰ infectious U/ml palGFP virus solution into the ventrobasal thalamic nuclei or substantia nigra, respectively. In both cases, the labeled neurons showed Golgi stain-like images with richly spiny dendrites (arrows in Figure 7c and 7d) and axon collaterals surrounding the cell bodies.

Discussion

The present study showed the advantages of palGFP Sindbis virus as in vivo anterograde and retrograde neuronal tracers with Golgi stain-like labeling of neurons. In contrast, palDsRed Sindbis virus was not suitable as an anterograde or retrograde tracer. DsRed immunoreactivity was mostly localized or aggregated inside of the cell bodies of neurons that were infected with palDsRed Sindbis virus. Very recently, DsRed was shown by analytical centrifugation to form a tetramer in living cells (Baird et al. 2000), and its car-boxy terminus was shown by X-ray crystallography to participate in tetramer formation (Yarbrough et al. 2001). Although we added the palmitoylation signal to the relatively free amino-terminus of DsRed, the added peptide might interfere with the tetramer formation and result in the aggregation of the protein or in the unexpected intracellular localization.

Altman-Hamamdzic et al. (1997), using a replication-deficient Sindbis virus containing the lacZ gene, reported that β-galactosidase activity in the injected site was very strong 24–48 hr after injection of the virus into mouse neostriatum, and decreased gradually until 14 days after infection. Gwag et al. (1998) also reported that β-galactosidase activity was very strong 24–48 hr after injection of lacZ Sindbis virus into rat hippocampus, and decreased gradually until 21 days after infection. These time courses of reporter gene expression were quite consistent with the present results. GFP immunoreactivity at the injection site was very intense 18–72 hr after injection of palGFP Sindbis virus, and decreased gradually from 7 to 28 days after infection. The present results further indicate that the decrease in GFP immunoreactivity was caused by neuronal death rather than by simple cessation of palGFP production, because degenerative changes were observed in palGFP-expressing neurons at 7–28 days after infection. This suggests that overproduction of reporter proteins, which was driven by the strong subgenomic promoter of Sindbis virus, would consume important resources of the infected neurons to bring them to ruin, or that the virus would start a mechanism of programmed cell death as a native Sindbis virus induces apoptosis in neurons (Griffin et al. 1994).

At a very early stage of palGFP Sindbis virus infection, GFP immunoreactivity was located in small cells and amorphous bushy structures at the injection site (Figure 3a). The amorphous or bushy immunoreactivity was similar to that reported in the previous study using a recombinant adenovirus encoding palGFP (Figures 1C and 2A in Tamamaki et al. 2000). The immunoreactivity was considered to be labeling of glial cells and processes, because adenovirus was more infectious to glial cells than to neurons. Furthermore, in the present study, many small cells, probably glial cells, were labeled with palGFP at 4.5–9 hr after infection. However, in contrast to adenovirus infection, the amorphous and cellular immunoreactivity in presumed glial cells and processes almost disappeared within 36 hr after the viral injection. Taking the degenerative changes of the infected neurons at later stages into consideration, this result might indicate that the infected glial cells died more quickly than the infected neurons. This could be accounted for by the argument that the resources of glial cells were much less than those of neurons because of the smaller size of glial cells, or that apoptosis would be induced much faster in glial cells than in neurons. Therefore, the stages at 18–72 hr after infection were the period when the infected glial cells disappeared and the infected neurons expressed intense GFP immunoreactivity without discernible degenerative changes, and thus the time suitable for neuronal tracing.

The palGFP Sindbis virus was proved to be an excellent tracer for anterograde axonal labeling (Figure 5). It could be applied to thalamocortical and cortico-cortical glutamatergic neurons, striatonigral GABAergic neurons, and nigrostriatal dopaminergic neurons in the present experiments, indicating that the Sindbis virus does not choose a target neuron. One of the possible advantages of viral vectors is that the virus can label axons sparsely but intensely; in the best case, intensely stained axons from a single neuron can be traced. On the other hand, in the case of the previous axonal tracers, the smaller the number of labeled neurons, the weaker the intensity of axonal labeling. This is not the case for viruses because of their proliferative property. The recombinant Sindbis virus produces reporter proteins massively, although it can not produce mature viral particles. In addition, using palGFP Sindbis virus, we can determine, by their Golgi stain-like images, the morphological type of the neurons that send axon fibers to the target. This has been impossible with ordinary anterograde tracers, such as Phaseolus vulgaris leukoagglutinin. We hope that the palGFP Sindbis virus will be helpful in analyzing neuronal connections of the central nervous system by its outstanding properties.

Footnotes

Acknowledgments

Supported by Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan (12308039, 12680731, 13035020, 13035026, 13041024, 13041025, and 13210071).

We are grateful for the photographic help from Mr Akira Uesugi and Ms Keiko Okamoto.

References

Altman-Hamamdzic

Groseclose

J-X

Hamamdzic

Vrindavanam

Middaugh

Parratto

Sallee

(1997) Expression of β-galactosidase in mouse brain: utilization of a novel nonreplicative Sindbis virus vector as a neuronal gene delivery system. Gene Ther 4:815–822

Baird

Zacharias

Tsien

(2000) Biochemistry, mutagenesis, and oligomerization of DsRed, a red fluorescent protein from coral. Proc Natl Acad Sci USA 97:11984–11989

Bredenbeek

Frolov

Rice

Schlesinger

(1993) Sindbis virus expression vectors: packaging of RNA replicons by using defective helper RNAs. J Virol 67:6439–6446

Gerfen

Herkenham

Thibault

(1987) The neostriatal mosaic: II. Patch- and matrix-directed mesostriatal dopaminergic and non-dopaminergic systems. J Neurosci 7:3915–3934

Griffin

(1998) A review of alphavirus replication in neurons. Neurosci Behav Rev 22:721–723

Griffin

Levine

Ubol

Hardwick

(1994) The effect of alphavirus infection on neurons. Ann Neurol 35:S23–27

Gwag

Kim

Ryu

Won

Cho

Y-G

Sung

(1998) A neuron-specific gene transfer by a recombinant defective Sindbis virus. Mol Brain Res 63:53–61

Lendvai

Stern

Chen

Svoboda

(2000) Experience-dependent plasticity of dendritic spines in the developing rat barrel cortex in vivo. Nature 404:878–881

Lundstrom

(1999) Alphaviruses as tools in neurobiology and gene therapy. J Recept Signal Transduct Res 19:673–686

10.

Moriyoshi

Richards

Akazawa

O'Leary

DDM

Nakanishi

(1996) Labeling neural cells using adenoviral gene transfer of membrane-targeted GFP. Neuron 16:255–260

11.

Okada

Lasford

Weimann

Fraser

McConnel

(1999) Imaging cells in the developing nervous system with retro-virus expressing modified green fluorescent protein. Exp Neurol 156:394–406

12.

Piper

Slot

Stahl

James

(1994) Recombinant Sindbis virus as an expression system for cell biology. Methods Cell Biol 43:55–78

13.

Schlesinger

(1993) Alphaviruses—vectors for the expression of heterologous genes. Trends Biotechnol 11:18–22

14.

Strauss

(1994) The alphaviruses: gene expression, replication, and evolution. Microbiol Rev 582:491–556

15.

Tamamaki

Nakamura

Furuta

Asamoto

Kaneko

(2000) Neurons in Golgi-stain-like images revealed by GFP-adenovirus infection in vivo. Neurosci Res 38:231–236

16.

Xiong

Levis

Shen

Schlesinger

Rice

Huang

(1989) Sindbis virus: an efficient, broad host range vector for gene expression in animal cells. Science 243:1188–1191

17.

Yarbrough

Wachter

Kallio

Matz

Remington

(2001) Refined crystal structure of DsRed, a red fluorescent protein from coral, at 2.0-Å resolution. Proc Natl Acad Sci USA 98:462–467