New rabies virus variants for monitoring and manipulating activity and gene expression in defined neural circuits

Abstract

Glycoprotein-deleted (ΔG) rabies virus is a powerful tool for studies of neural circuit structure. Here, we describe the development and demonstrate the utility of new resources that allow experiments directly investigating relationships between the structure and function of neural circuits. New methods and reagents allowed efficient production of 12 novel ΔG rabies variants from plasmid DNA. These new rabies viruses express useful neuroscience tools, including the Ca(2+) indicator GCaMP3 for monitoring activity; Channelrhodopsin-2 for photoactivation; allatostatin receptor for inactivation by ligand application; and rtTA, ER(T2)CreER(T2), or FLPo, for control of gene expression. These new tools allow neurons targeted on the basis of their connectivity to have their function assayed or their activity or gene expression manipulated. Combining these tools with in vivo imaging and optogenetic methods and/or inducible gene expression in transgenic mice will facilitate experiments investigating neural circuit development, plasticity, and function that have not been possible with existing reagents.

Figure 1. Production of rabies viral vectors encoding multiple genes

(A, B) Generation of SADΔG-mCherry (A) and SADΔG-BFP (B) rabies viruses. (A) Neurons in layer 6 of the rat primary visual cortex were visualized following injection of SADΔG-mCherry into the dLGN. (B) Neurons in layers 2/3 and 4 of rat visual cortex were labeled with SADΔG-BFP by injection into nearby visual cortex. Scale bars, 100 μm. (C) Each open reading frame requires transcription start and stop sequences to be inserted before and after the open reading frame in the rabies genome. To insert an additional exogenous gene, mCherry with additional transcription start and stop sequences (black with red outlines) followed by mCherry was inserted between the last codon of GFP and its transcription end sequence. (D, E, F) Neurons in cortical slice cultures infected with SADΔG-GFP-mCherry. All of the neurons infected with SADΔG-GFP-mCherry express both mCherry (D, F) and GFP (E, F), indicating reliable transcriptional regulation and expression of both gene products. Scale bar, 100 μm. (G) Plasmid created for efficient introduction of one or two trasngenes into the rabies genome. The pSADΔG-F3 rabies genome vector has two multiple cloning sites (MCS-1 and MCS-2) for insertion of genes of interest. For expression of a single gene from rabies vectors, use of one site in MCS-1 and one site in MCS-2 enables insertion of a single ORF and deletion of stop and start transcription cassettes. For cloning of two transgenes, insertion of each gene in MCS-1 and MCS-2 allows reliable co-expression from the rabies vectors.

Figure 2. Monitoring of neural activity with GCaMP3-expressing ΔG rabies virus

(A) Retinotopic organization of the striate and extrastriate cortical regions. The retinotopic map from intrinsic imaging was overlaid on the image of surface blood vessels. Location of the border between V1 and AL was identified based on the representation of the vertical meridien (nasalmost visual fields) to allow targetting of viral injections to AL (red X). SADΔG-GCaMP3-DsRedX was injected into the lateral extrastriate cortical area AL of the mice and the corresponding retinotopic location in V1 was noted as the expected location of retrogradely infected neurons (red square). (B) Z-stack of SADΔG-GCaMP3-DsRedX-infected neurons visualized in vivo with two-photon imaging of V1, 9 days after rabies injection. AL-projecting V1 neuronal cell bodies and processes could seen in imaging planes extending from the cortical surface to a depth of 1.5 mm (C) Top view of two-photon laser-scanning images of SADΔG-GCaMP3-DsRedX-infected neurons at depth of 370 μm from the cortical surface. V1 neurons were retrogradely labeled with SADΔG-GCaMP3-DsRedX and co-expressed GCaMP3 (green) and DsRedX (red). Note that GCaMP3 could be seen in dendrites and axons, as well as cell bodies. Scale bar, 25 μm. (D) Orientation selectivity of SADΔG-GCaMP3-DsRedX-infected V1 neurons. Panels in D1-D4 correspond to neuronal cell bodies (D1-D2) or dendrites (D3-D4) labeled 1-4 in panel C. Orientation tuning curves are plotted as the mean change in fluorescence of the cell body (D1 and D2) or dendritic segments (D3 and D4) during the entire stimulus period, in response to square-wave gratings presented at various orientations in a random order. (E) Changes in fluorescence over time, in response to drifting gratings at the preferred orientation. Time 0 indicates the onset of the visual stimulus, which lasted for 4 seconds, as indicated by the black bar. Note that fluorescence was modulated at temporal frequencies that correspond to the temporal frequencies of the drifting gratings. These temporal modulations in phase with the visual stimuli can also be seen in supplemental movies. Values in D and E represent means ± S.E.M. of ΔF/F values across 5 repetitions of the visual stimulus.

Figure 3. Photoactivation of neurons infected with ChR2-mCherry-expressing ΔG rabies virus

(A) Generation of action potentials in ChR2-mCherry-expressing neurons by blue light pulses. Intracellular recordings were made from a layer 5 neuron in S1 barrel cortex 6, 8 and 10 days after injection of SADΔG-ChR2-mCherry in postnatal nine-day-old mice. During current-clamp, action potentials were reliably generated by blue light pulses of 5 Hz and 2 ms duration on day 8 and 10, but not day 6. (B) The recorded neuron, filled with biocytin and stained with streptavidin-Cy2 (green), was positive for mCherry (red). Scale bar, 30 μm.

Figure 4. Allatostatin-induced silencing of neural activity with AlstR-expressing ΔG rabies virus

(A-C) Whole-cell current clamp recordings from a pyramidal neuron in a brain slice from the barrel cortex of an 18-day-old mouse, 7 days after injection of SADΔG-GFP-AlstR. (A) Before application of allatostatin (AL), representative traces of responses to depolarizing current pulses (+50, +100, +150, and +200 pA, 750 ms duration) show that multiple action potentials are generated with as little as +50pA. (B) Application of the ligand AL at 1 μM inactivated the SADΔG-GFP-AlstR-infected neuron. +200 pA of current could no longer generate action potentials in the presence of AL. (C) The effects of AL were partially reversed by washout of AL with normal ACSF. +150 pA induced action potentials after washout of AL. (D) Photographs of the recorded neuron, as well as neighboring infected neurons. The recorded neuron was filled with biocytin and stained with streptavidin-Cy3 (red), and was positive for GFP derived from SADΔG-GFP-AlstR (green). Scale bar, 30 μm.

Figure 5. Doxycycline-dependent gene expression and conditional transsynaptic spread with SADΔG-GFP-rtTA rabies virus

(A, B) Following Biolistic transfection of isolated neurons in rat cortical slice culture with both pTetO-CMVmin-Histone2B-mCherry-F2A-B19G and pCMMP-TVA, application of EnvA-pseudotyped SADΔG-GFP-rtTA resulted in selective infection of transfected neurons. (A) The presence of both dox and rtTA expressed from the rabies virus allowed mCherry expression in EnvA-SADΔG-GFP-rtTA-infected neurons as shown by the yellow (GFP and mCherry) appearance of the soma of the “postsynaptic neuron”. The expression of rabies glycoprotein in this same postsynaptic neuron allowed complementation of the ΔG rabies so that the virus could spread to and express GFP in numerous presynaptic neurons. (B) Under the same conditions in the absence of dox (1.0 μg/ml), the EnvA-SADΔG-GFP-rtTA rabies virus also selectively infected scattered, isolated TVA-expressing neurons. But without dox there was no mCherry or rabies glycoprotein expression and therefore no complementation or transsynaptic spread of the rabies virus. Only three transfected and primarily infected neurons (green cells indicated by arrows) could be found in the brain slice in the absence of dox. Scale bars, 100 μm (right figures), 50 μm (left figures).

Figure 6. Recombination-dependent gene expression by ER^T2CreER^T2 or FLPo ΔG rabies viruses

(A) Temporal regulation of gene expression using inducible Cre recombinase. SADΔG-GFP-ER^T2CreER^T2 activates tamoxifen-dependent expression of DsRed in HEK293t cells transfected with pCALNL-DsRed (recombination indicator). Top and bottom panels show results in the presence and absence of tamoxifen (+4HOT, 1 nM), respectively. Left panels show GFP expression from the rabies genome, middle panels show Cre-induced DsRed expression in the presence of 4HOT (top) but not in the absence of 4HOT (bottom); and overlays in right panels. Scale bar, 100 μm. (B) SADΔG-FLPo-DsRedX-infected cells expressed a red fluorescent DsRedX in HEK293t cells 3 days after virus infection. Scale bar, 100 μm. (C, D) Recombination induced in SADΔG-FLPo-DsRedX-infected cells. X-gal staining showed that HeLa cells stably expressing frt-STOP-frt LacZ cassette (recombination indicator) expressed LacZ in the presence of SADΔG-FLPo-DsRedX (C), but not in the absence of SADΔG-FLPo-DsRedX (D). Scale bar, 300 μm.