Optogenetic action potentials and intrinsic pacemaker interplay in retrogradely identified midbrain dopamine neurons

Dissecting the diversity of midbrain dopamine (DA) neurons by optotagging is a promising addition to better identify their functional properties and contribution to motivated behavior. Retrograde molecular targeting of DA neurons with specific axonal projection allows further refinement of this approach. Here, we focus on adult mouse DA neurons in the substantia nigra pars compacta (SNc) projecting to dorsal striatum (DS) by demonstrating the selectivity of a floxed AAV9‐based retrograde channelrhodopsin‐eYFP (ChR‐eYFP) labeling approach in DAT‐cre mice. Furthermore, we show the utility of a sparse labeling version for anatomical single‐cell reconstruction and demonstrate that ChR‐eYFR expressing DA neurons retain intrinsic functional properties indistinguishable from conventionally retrogradely red‐beads‐labeled neurons. We systematically explore the properties of optogenetically evoked action potentials (oAPs) and their interaction with intrinsic pacemaking in this defined subpopulation of DA neurons. We found that the shape of the oAP and its first derivative, as a proxy for extracellularly recorded APs, is highly distinct from spontaneous APs (sAPs) of the same neurons and systematically varies across the pacemaker duty cycle. The timing of the oAP also affects the backbone oscillator of the intrinsic pacemaker by introducing transient “compensatory pauses”. Characterizing this systematic interplay between oAPs and sAPs in defined DA neurons will also facilitate a refinement of DA neuron optotagging in vivo.


| INTRODUCTION
Dissecting the diversity of midbrain dopamine (DA) neurons by optotagging is a promising addition to better identify their functional properties and contribution to motivated behavior (Pallikaras et al., 2022;Steinberg et al., 2013;van der Merwe et al., 2023).Based on the selective expression of channelrhodopsin (ChR) in DA neurons, previous approaches (Cohen et al., 2012;Matsumoto et al., 2016;Mohebi et al., 2019;Stauffer et al., 2016) were based on the ability to reliably evoke short latency action potentials (APs) by light illumination in combination with the key assumption that the shape of the optogenetically evoked AP (oAP) remains unaltered and thus can be used as an identifier for a particular DA neuron in the context of multi-electrode in vivo recordings.This assumption has not yet been explicitly tested in vitro, where the cellular origin of sAPs and oAPs is experimentally determined.In addition, potential effects of oAPs on intrinsic pacemaking have also not yet been studied in DA neurons.A systematic characterization of excitability changes across the pacemaker duty cycle of DA SN neurons in vitro has been recently probed by phase setting curves (Higgs et al., 2023).In this elegant study, the role of intrinsic conductances mediated by, for example, Kv4 and SK channels in limiting spike timing pertubations have been identified.Therefore, we expect that also oAPs interact with the intrinsic duty cycle.
In contrast, the techniques for selective molecular optotagging of DA midbrain neurons are well established: floxed viral constructs (often delivered by AAVs) for expression of ChR (or other optogenetic tools) are used in combination with DAT-cre reporter mice, where cre recombinase expression is selective to DA midbrain neurons (Lammel et al., 2015).As an alternative, double transgenic approaches can be utilized (DAT-cre + floxed ChR2 transgenic mice (Kim et al., 2015)).While these approaches reliably and selectively label the majority of DA midbrain neurons, they do not discriminate between distinct subpopulations by single cell sequencing approaches (Poulin et al., 2014(Poulin et al., , 2018) ) or axonal projections (de Jong et al., 2019;Lammel et al., 2008Lammel et al., , 2011)).These emerging strategies have the capacity to single-out DA subpopulation with particular affiliations to corticostriatal networks and in turn help to define their functional roles (Azcorra et al., 2023;Heymann et al., 2020).In this context, retrograde molecular labelling approaches are one attractive option, for example, by infusing a floxed AAV construct into a particular target region of the striatum of a DAT-cre mouse.Not all AAV serotypes are suitable for this retrograde approach (Tervo et al., 2016), but AAV9 has already been successfully applied to retrogradely label dorsal striatum (DS)-projecting DA neurons located in the substantia nigra (SN) for anatomical analysis (Crittenden et al., 2016).Based on this work, we have studied the selectivity and functional impact of this retrograde AAV9 DAT-cre method and provide two protocols for either sparse labelling suitable for single cell anatomical reconstructions or bulk labelling for functional analysis.We use the latter for a detailed interrogation of the interaction between oAPs and sAPs in defined DS-projecting DA SN neurons.

| Animals
Adult male and female DAT-Cre mice (Slc6a3 tm1(cre)Xz ; Jackson Laboratory stock number: 020080 (Zhuang et al., 2005) backcrossed to C57BL/6N mice for 6 generations), at the average age of 18 weeks, were used in this study.All procedures involving animals were approved by the Regierungspräsidium Darmstadt (FU/2026).Animals were kept at a 12 h light-dark cycle, group housed whenever possible, and had unlimited access to food and water.
For dual tracing animals underwent a second surgery were 100 nl of red retrobeads (Lumafluor, USA) diluted 1:30 in artificial cerebrospinal fluid (aCSF, Harvard Apparatus, USA) were injected bilaterally in the DLS.

| Confocal imaging
Pictures of brain slices were taken with a laser-scanning confocal microscope (Nikon Eclipse90i, Nikon, Japan) operated using AR-NIS-Elements (Nikon, Japan).
For neuronal reconstruction, high-magnification imaging was performed using a 60Â oil immersion objective.Images were acquired as Z-stacks, with dimensions of 1024 Â 1024 pixels and a resolution of 0.41 μm/ pixel in z-stacks spaced 0.4 μm apart.To facilitate later connection of neuronal segments, Z-stacks from the same brain slice had an overlap of approximately 20%.Manual tracing and imaging of tree structures were carried out in different brain slices, considering other labeled structures and measuring the distance between trees.

| Single cell reconstruction and analysis
Using Neurolucida360 (MBF Bioscience, USA) eYFPlabeled midbrain DA neurons were reconstructed.To account for any shrinkage that occurred during the immunohistology process, a correction factor for the z-axis was calculated and applied.
Inbuilt Neurolucida360 and Neurolucida Explorer (MFB Bioscience, USA) analysis tool were used to asses different parameters of the reconstructed neurons.

| Cell counting
For counting TH-eYFP overlapping cells four different z-stack scans with 1 μm spacing in three different slices of each midbrain were taken using a 60Â immersion oil objective.All positive neurons in the 488 nm channel (eYFP) stack were counted before the neurons in the 561 nm channel (TH) were counted using Fiji (Schindelin et al., 2012).For the counting of sparse labelled brains, the cells in the whole midbrains of bilaterally injected animals were counted.

| Statistical analysis
All statistical analysis were performed using GraphPad Prism 10.1 (GraphPad Software, USA) and MATLAB (MathWorks, USA).Normally distributed data sets are shown as mean ± SEM, while non-normally distributed data sets are shown as median ± the 95% confidence interval (95% CI).Detailed description of statistical tests can be found in Table 1.

| Retrograde viral infection of midbrain DA neurons using AAV9 enables molecular targeting of DA subpopulations
To selectively infect midbrain DA neurons based on their axonal projection site viral particles were injected in the DS.Using DAT-Cre mice combined with AAV9-double-floxed-hChR2(H134R)-eYFP, viral expression was limited to tyrosine hydroxylase (TH) expressing midbrain neurons (Figure 1A).Immunohistochemical staining of striatal injection sites showed that eYFP expression (eYFP + ) is limited to TH-positive (TH + ) terminals (Figure 1B).Using higher magnification, eYFP + TH + terminals were observed located next to eYFPnegative (eYFP À ) TH + terminals in the striatum (Figure 1C).On the level of the midbrain, robust retrograde infection of DA neurons was seen, located in the medial part of the SN pars compacta (Figure 1D).At higher magnification, eYFP + TH + DA neurons were observed next to non-infected eYFP À TH + DA neurons (Figure 1E).In total, 1477 eYFP + TH + neurons were counted across four midbrains (ca.370 eYFP + TH + per mouse), while no eYFP + TH À neurons were identified  (Figure 1F).This data demonstrated the high DA selectivity and robust expression of the AAV9-based retrograde labelling approach.
3.2 | Pacemaking and action potential properties of dorsolateral striatumprojecting DA SN neurons were unaffected by chronic AAV9-mediated channelrhodopsin expression We compared the functional properties of AAV9 and fluorescent retrobeads double retrogradely labeled dorsolateral striatum (DLS)-projecting DA SN neurons with those only labeled with red retrobeads (RB).About four weeks after unilateral viral injection of AAV9-double-floxed-hChR2(H134R)-eYFP in the dorsolateral striatum (DLS), animals underwent a second surgery to label ipsi-and contralateral DLS-projecting DA SN neurons with RB (Figure 2A).After 3 days of recovery from the 2nd tracing, electrophysiological in vitro experiments were performed (Figure 2B). Figure 2C shows a representative example of injection sites in the DLS.While after unilateral injection, viral expression of eYFP was observed only on the ipsilateral side, RB injection sites as expected were observed in both hemispheres.Using whole cell in vitro patch clamp recordings, we observed spontaneous pacemaking in both eYFP + (Figure 2C, top left) and eYFP À (Figure 2C, bottom left) DLS-projecting DA SN neurons in current clamp identified via RBs (Figure 2C).No differences between eYFP + and eYFP À red beads containing (RB + ) DA neurons were observed regarding mean firing frequencies, coefficient of variation, the voltage at threshold, minimal afterhyperpolarization (minAHP), and the action potential with at threshold (Table 1, Figure 2D-H).

| Striatal injections of 10 nanoliter of diluted AAV9 constructs enabled reliable sparse labeling of individual DA SN neurons suitable for somato-dendritic single cell reconstruction
We also aimed to adapt the AAV9-based retrograde tracing of DA neurons for sparse labelling suitable for single cell anatomical reconstruction.We tested a range of volumes (10-40 nl) and dilutions (3.7 * 10 13 vector genomes per milliliter undiluted, 1:3, 1:10, 1:30, 1:100 diluted in aCSF, data not shown) and identified the combination most reliable for sparse labelling of DS-projecting DA SN neurons: 10 nl of 1.2 * 10 12 AAV9 vector genomes per milliliter was the most promising combination and thus used to compare the morphology of individual DA neurons from three different striatal projection sites.Figure 3 shows a representative example of this sparse labelling approach.While no eYFP + terminals were visible under low magnification in the striatum (Figure 3B), we identified eYFP expression in the distal axon under high magnification (Figure 3C).While this demonstrates that eYFP immunoreactivity is to some degree detectable also in the striatum, we did not attempt a full reconstruction of the distal axon.Regarding somato-dendritic labeling, we found on average about 5 labeled DA SN neurons per animal (Figure 3E) and only in one of 12 animals no single cell was detected.Thus, we succeeded in establishing a reliable AAV9-based sparse retrograde labelling protocol.
3.4 | Somato-dendritic single cell reconstructions of DA neurons revealed projection-specific differences in dendritic architecture  compare the different complexities of dendritic branching, we plotted the corresponding dendrograms (Figure 4B, right column).Although we only reconstructed a small number of individual cells (i.e., n = 3 for each projection), DA SN neurons projecting to lNAcc displayed a significantly lower dendritic complexity according to Sholl analysis compared to DA SN neurons projecting to DLS or DMS (Figure S4A, Figure S4B, for further data see Table 2).These results demonstrate that our AAV9-based sparse labelling protocol can be successfully applied to reveal projection-specific morphological differences among DA neurons.Future studies are needed to test whether this is also possible for the distal axon in the striatum.

| Optogeneticallly evoked action potentials (oAPs) are different from spontaneous action potential (sAPs) in DS-projecting DA SN neurons
We carried out in vitro whole cell current clamp recordings of spontaneously firing DS-projecting DA SN neurons, retrogradely infected with AAV9-double-floxed-hChR2(H134R)-eYFP in DAT-cre mice.When activating ChR for 5 ms by wide-field LED stimulation (3 mW at 470 nm) single oAPs were detected (Figure 5A-C).As expected, the probability of eliciting these oAPS increased with increasing duration (1, 5, 10 ms) or increasing stimulus intensity (0.3, 1.0, 3.0 mW; Figure 5D).Also, oAP latency from stimulus onset decreased (from 7.5-3 ms) with increasing stimulus intensity.However, stimulus duration per se did not affect oAP latency (Figure 5E).Regarding the shape of the APs, we observed striking differences between oAPs and the immediately preceding sAPs of the same cell (sAP pre , Figure 5F).To quantify these differences, basic sAP pre and oAP parameters were compared (also with the first sAP following the oAP (sAP post )).The peak of the afterhyperpolarization (minAHP) was significantly decreased for the oAP compared with sAP pre and sAP post (Figure 5G).The minAHP of the oAP did not change across the normalized ISI (ISI norm ) of the spontaneous pacemaker (Figure 5H; linear regression: r = À.15, slope = À1.76,p = 0.11, n = 110).Also, the duration of the AP half width was significantly shorter for the oAP compared with sAP pre or sAP post (Figure 5I).A weak, but significant correlation (r = .23,slope = .09,p = 0.016, n = 110) was observed between the oAP half width and its position across the ISI norm (Figure 5J).Finally, the threshold of the oAP (defined as depolarization velocity >2.5 mV/ms) was significantly lower (difference between medians (sAP pre vs. oAP): À11.5 mV; difference between medians (sAP post vs. oAP): À12.3 mV) compared with sAP pre and sAP post (Figure 5K) Also, the oAP threshold was strongly dependent on its position within the ISI norm (Figure 5L, r = .76,slope = 1.29, p = <0.0001,n = 110).

| Recovery of pacemaker rhythm after oAPs
We also assessed the impact of the oAP on the pacemaker rhythm by comparing the spontaneous action potentials after the oAP (i.e., AP post) ; Figure 6A, black trace) with the expected APs based on the ISI norm (Figure 6A, red trace).We calculated the temporal differences between the AP post and the three expected APs.This indicated the   6B), which was maximal for the first AP post and absent for the third AP post .The temporal distortion can also be expressed as relative ISI changes (Figure 6C), where the first two ISIs after the oAP (ISI 1/2 ) were longer and the third ISI unaffected (Figure 6D).Thus, oAPs induce a transient and systematic perturbation of the intrinsic pacemaker rhythm of DA neurons.

| Differences between the first derivatives of oAPs and sAPs
To predict potential effects of oAPs for extracellular recordings, we calculated and analyzed the corresponding first derivative of our whole-cell recordings (Figure 7A, Figure S7A).By aligning the peaks of the first derivatives of sAP pre and oAPs, we detected significant waveform differences (Figure 7B-J).In particular, the peak, corresponding to the maximal depolarization velocity (max dV/dt, Figure S7B), the valley, corresponding to the maximal repolarization velocity (min dV/dt, Figure S7B) and the height were significantly larger for oAPs, while the peak-to-valley width was significantly shorter for oAPs.Given the results, we expect strong differences between extracellularly recorded oAPs and sAPs.Furthermore, we compared sAP-oAP pair from four different DA neurons (Figure 7K) and plotted their respective properties in a two-dimensional space (Figure 7L-M).Here, sAPs and oAPs exhibited only minimal overlap.While sAPs formed homogeneous clusters for each DA neuron, oAPs were scattered distant their respective sAP clusters (Figure 7K-M).These data illustrate that it is likely to be non-trivial to identify matching oAP-sAPs pairs originating from the same individual DA neuron.
We extended this approach to our entire sAP-oAP dataset (Figure 8A, n = 18, oAP = 110, sAP = 2123).For quantitative analysis of the waveform features, we calculated the centers of their respective sAP clusters and determined their differences to individual corresponding oAPs (Figure 8B).These differences for all four waveform parameters were significantly different compared with that of individual sAPs to their cluster center (Figure 8C).This confirms the divergence of sAP and oAPs in feature space.Finally, we asked whether this differences measure varied with the position of the oAP in the ISI norm .
Figure 8D-E illustrates that the difference between 1st derivative waveforms of oAPs and sAPs are strongly position dependent and are maximal during the early ISI norm .As shown in Figure 8F these position-dependent differences are at least one order of magnitude larger compared with variability within sAPs.
Overall, these results demonstrate that in vitro waveform properties of oAPs display significant and ISI-position-dependent differences compared with their corresponding sAPs and also affect the timing of the intrinsic pacemaker.

| DISCUSSION
In the initial section of our study we demonstrate that AAV9-mediated retrograde molecular labelling of projection-defined DA SN neurons in DAT-cre mice is highly selective and does not modifyeven after prolonged in vivo expressionthe pacemaking properties of the DLS-projecting DA subpopulation when compared with conventional short-term retrograde labeling with retrograde beads (Lammel et al., 2008).We are aware that differences might still be present at the level of synaptic inputs and/or neuromodulation.These factors need to be addressed in follow-up in vivo studies, for example, by recording, optotagging and juxtacellular filling of AAV9-retrogradely labeled DA neurons.Furthermore, we show that our sparse labelling protocol gives reliable results and enabled the full reconstruction of the somatodendritic domain of three different projection-defined DA subtypes (Figures 3, 4).We give also anecdotal evidence that this approach might also be suitable to reconstruct the huge axonal tree of these neurons in the striatum, a very labor-intensive work not attempted here.However, in comparison to previous approaches where highly diluted sindbis viral infusion were made in the midbrain (Matsuda et al., 2009), our approach can easily be targeted to a particular projections system of interest and showed no signs of toxicity (Figure 3E).
We have then focused on the in vitro properties of optogenetically-evoked action potentials in DS-projecting DA SN neurons and their reciprocal interaction with intrinsic pacemaker activity.We believe this to be relevantalso for in vivo recordingsas there are two issues that need to be resolved for reliable DA optotagging: on one hand sufficient and selective channelrhodopsin expression to evoke short latency oAPs in the target neurons and on the other hand, the assignment of the evoked oAP to one of the multiple spontaneously active DA units observed in the same recording.While the first feature has been achieved in several studies (Cohen et al., 2012;Matsumoto et al., 2016;Mohebi et al., 2019;Stauffer et al., 2016) by restricting channelrhodopsin to DA neurons by the use of DAT-cre (or THcre (Lammel et al., 2015) driver mouse lines, the second condition is less certain.The matching of oAPs to sAPs is usually based on a degree of similarity between their waveforms (Cohen et al., 2012;Matsumoto et al., 2016;Mohebi et al., 2019;Stauffer et al., 2016).However, this assumption has never been explicitly tested i.e. individual cells need to identified independent of optotagging.By filling the neuron during whole-cell patch-clamp recording this is easily achieved in brain slices (Evans et al., 2017), but more challenging in vivo, where juxtacellular or in vivo whole-cell options are available (Farassat et al., 2019;Otomo et al., 2020).Here, we focus on in vitro brain slices and demonstrate that short latency oAPs are evoked robustly ( p > 0.9) by short widefield LED illumination in a reasonable intensity range (see Figure 5D, E).We document that the intracellular AP waveform as well as its 1st derivative -a proxy for extracellular APs -displays significant differences in comparison to the sAPs originating from the same neurons.Moreover, these waveform differences are such that a reliable assignment between oAPs and sAPs from the same cell cannot be easily made.However, some systematic oAP-sAP relations did emerge (see Figure 8).First, the differences between oAP waveform and the preceding sAP linearly scaled with the temporal position of the oAP in the duty cycle of the background pacemaker.This implies that oAPs are more similar to sAPs the later they are evoked during the interspike interval.This systematic relation between oAP and sAP waveforms might also be present in vivo and could facilitate cell identification.In addition, we identified oAP effects on pacemaker timing in the sense of a "compensatory pause" (similar to the one observed in the beating heart after a ventricular extrasystole) in the DA SN pacemaker.This pause was most obvious when the oAP was elicited early during the pacemaker duty cycle.Future in vivo optotagging experiments are needed to clarify whether this oAP-induced pause can also be identified in the context of irregular or bursty firing patterns.
In summary, our in vitro study suggests a refined use of oAP timing and timing-sensitive oAP waveform data to improve DA optotagging.
AUTHOR CONTRIBUTIONS Niklas Hammer: Conceptualization; data curation; formal analysis; investigation; methodology; project administration; software; visualization; writing-original draft; writing-review and editing.Pascal Vogel: Data F I G U R E 8 oAPs are distant from their respective sAPs-cluster center and the distance scales with the ISI norm .(a) Cluster distribution of sAPs (left) and oAPs (right) of all recorded neurons when the spike waveform features valley (min dV/dt) and peak (max dV/dt) (top), as well as spike height (peak-to-valley amplitude) and peak-to-valley width (bottom) are plotted against each other.Cells are shown in different colors.Gray shadings (right) indicate the corresponding clusters of sAPs.(b) Illustration of calculating the distance of a spike to its cluster center.Left, waveform of a oAP (blue) and the average waveform of all sAPs (gray) from the same cell.The difference between the oAPs and the averaged sAP waveform is indicated as x (peak distance) and y (valley distance).Right, valley (min dV/dt) and peak (max dV/dt) of all oAPs (blue squares) and sAPs (gray circles) from the same cell shown on the left.The blue square which peak and valley distances to the cluster center is indicated as x and y, corresponds to the blue oAP shown on the left.The center of the cluster (orange cross) reflects the average peak plotted against the average valley across all sAPs from this cell.(c) Absolute distance from all sAPs (gray) and oAPs (blue) to their cluster centers regarding the spike waveform features peak (max dV/dt), valley (min dV/dt), peak-to-valley width and height (peak-tovalley amplitude).Boxplots represent median (line), 25th and 75th percentiles (box), and 5th and 95th percentiles (whiskers).*** p < 0.00001, Wilcoxon rank sum test.(d) Waveform of three oAPS (blue) with their corresponding averaged sAPs (gray).The shown oAPs differ in their temporal distance to the preceding sAPs and are sorted from short to long intervals.Note that the shape of the oAP waveform scales with the ISI norm .Scalebar 20 mV/ms, 2 ms.(e) Absolute distance from all oAPs to their cluster centers plotted as a function of the normalized preceding ISI.Note that the change in valley (left) and peak-to-valley width (right) of the oAPs scales with the ISI norm .Red diamonds, triangles and circles indicate the oAPs shown in D. (a) Absolute distance from all oAPs (blue) and their corresponding sAP pre (gray) to their cluster centers for the waveform features valley (left), peak-to-valley width (middle), and valley width (right).Absolute distance values are sorted according to their ISI norm and are assigned to four categories: 0-25th, 25th-50th, 50th-75th, 75th-100th percentiles.Lines show the averaged absolute distance ± SEM for all four groups.curation; formal analysis; software; writing-review and editing.Sanghun Lee: Data curation; investigation.Jochen Roeper: Conceptualization; funding acquisition; project administration; resources; supervision; validation; writing-original draft; writing-review and editing.

F
I G U R E 1 Retrograde viral infection of midbrain DA neurons using AAV9 enables molecular targeting of DA subpopulations.(a) 300 nl of AAV9-double-floxed-hChR2(H134R)-eYFP (titer of 1.8 Â 10 13 vg/ml) were infused in the DS.(b) Injection site at 0.74 mm anterior in reference to bregma showing eYFP + axonal terminals of midbrain DA neurons in the DS.(c) 60Â magnification of the area indicated in B of striatal injection site showing eYFP + and TH + DA terminals next to eYFP À TH + DA terminals.(d) Midbrain section containing DA neurons labelled trough viral injection in the DS, showing eYFP + TH + DA neurons in the medial SNc.(e) High magnification of midbrain sections show eYFP + , TH + DA neurons (yellow arrows), next to eYFP À TH + neurons (white/red arrows).(f) On average, 370 cells were counted per animals.All 1477/1477 eYFP + cells in the midbrain were TH + (N = 4).Red = TH, green = eYFP.F I G U R E 2 Pacemaking and action potential properties of DLS-projecting DA SN neurons were unaffected by chronic AAV9-mediated channelrhodopsin expression.(a) DAT-Cre animals were injected unilaterally in the dorsolateral striatum (DLS) using AAV9-double-floxed-hChR2(H134R)-eYFP and bilaterally with RB in the DLS to label midbrain DA neurons based on their axonal projection site.(b) Timetable of the experimental approach.AAV9-double-floxed-hChR2(H134R)-eYFP was injected unilaterally at least 28 days prior bilateral injection of RB in the DLS.Three days after beads injection electrophysiological in vitro recordings were performed.(c) Representative injection site showing bilateral injection of RB and unilateral eYFP terminals of midbrain DA neurons in the DLS.(d) Upper left panel: spontaneous pacemaking activity of a DLS-projecting DA neuron; upper right panel: representative RB + , TH + , and eYFP + midbrain DA neuron, lower left panel: spontaneous pacemaking activity of a DLS-projecting midbrain DA neuron; lower right panel: representative RB + , TH + and eYFP À DA neuron.Scalebar 10 mV, 1 s.blue = TH, green = eYFP, white = NB, red = RB.
Using the above described projection-specific sparse labeling approach in three different subareas of the striatum (i.e., DLS, DMS and lateral shell of nucleus accumbens [lNAcc]), we compared the somato-dendritic morphology of the respective eYFP+ TH + DA neurons using z-stacks of confocal imaging and Neurolucida 360 for reconstruction (Figure4A).As shown for individual examples in Figure4B, DA neurons projecting to DLS (top, blue), DMS (middle, red) and lNAcc (bottom, green) were fully reconstructed including the proximal axon (Figure4B, left column;(Henny et al., 2012)).To

F
I G U R E 3 Striatal injections of 10 nanoliter of diluted AAV9 constructs enabled reliable sparse labeling of individual DA SN neurons suitable for somato-dendritic single cell reconstruction.(a) Schematic representation of experimental approach.Small volumes of 10 to 25 nl of highly diluted AAV9-DIO-eYFP were injected in different subareas of the striatum enabling sparse labeling of different subpopulations of midbrain DA neurons based on their axonal projection sites.(b) Striatal injection site in low magnification.(c) High magnification of the indicated area in B shows sparse eYFP + TH + terminal next to the majority of eYFP À TH + terminals.(d) Midbrain section at low magnification showing, that only a small amount of midbrain DA neurons were labeled using the sparse labelling protocol.(e) On average 5.5 with a range of 0 to 11 DA neurons were labeled per animal when injected bilaterally in the striatum.All eYFP + neurons located in the midbrain were TH positive.Red = TH, green = eYFP.

F
I G U R E 4 Somato-dendritic single cell reconstructions of DA neurons revealed projection-specific differences in dendritic architecture.(a) Workflow of reconstruction.After injection and sufficient time for expression of eYFP in vivo, midbrain blocks were processed immunohistochemically.Single cells were reconstructed using Neurolucida360.(b) the left column shows single reconstructed neurons in coronal plane projecting to DLS, DMS and lNAcc.Corresponding dendrograms of the reconstructed neurons are displayed in the right column.
T A B L E 2 Somatic und dendritic size and complexity measures of reconstructed midbrain DA neurons.