Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 May 5.
Published in final edited form as: Curr Biol. 2014 Apr 17;24(9):1024–1030. doi: 10.1016/j.cub.2014.03.041

Trim9 regulates activity-dependent fine-scale topography in Drosophila

Limin Yang 1,2, Ruonan Li 2, Takuya Kaneko 1, Kendra Takle 3, Rei K Morikawa 4, Laura Essex 1, Xin Wang 1, Jie Zhou 5, Kazuo Emoto 4, Yang Xiang 2, Bing Ye 1,*
PMCID: PMC4030378  NIHMSID: NIHMS579948  PMID: 24746793

SUMMARY

Topographic projection of afferent terminals into two-dimensional maps in the central nervous system (CNS) is a general strategy used by the nervous system to encode the locations of sensory stimuli. In vertebrates, it is known that while guidance cues are critical for establishing a coarse topographic map, neural activity directs fine-scale topography between adjacent afferent terminals [1–4]. However, the molecular mechanism underlying activity-dependent regulation of fine-scale topography is poorly understood. Molecular analysis of the spatial relationship between adjacent afferent terminals requires reliable localization of the presynaptic terminals of single neurons as well as genetic manipulations with single-cell resolution in vivo. Although both requirements can potentially be met in Drosophila melanogaster [5, 6], no activity-dependent topographic system has been identified in flies [7]. Here we report a topographic system that is shaped by neuronal activity in Drosophila. With this system, we found that topographic separation of the presynaptic terminals of adjacent nociceptive neurons requires different levels of Trim9, an evolutionarily conserved signaling molecule [8–11]. Neural activity regulates Trim9 protein levels to direct fine-scale topography of sensory afferents. This study offers both a novel mechanism by which neural activity directs fine-scale topography of axon terminals and a new system to study this process at single-neuron resolution.

RESULTS AND DISCUSSION

The presynaptic terminals of the three nociceptive neurons in each larval hemi-segment are arranged in dorsal-to-ventral topography

We exploited the nociceptive neurons in Drosophila larva, termed the class IV dendritic arborization (C4da) neurons [12, 13], as a potential system for molecular and genetic analysis of fine-scale topography because each C4da neuron can be unambiguously identified. The dendrites of these neurons form an array of detectors for noxious stimuli on the larval body wall and respond to noxious heat, harsh mechanical poke, and intense short-wavelength light [14–17]. In each hemi-segment of a larva, the dendrites of three C4da neurons covers the body wall in a complete but non-overlapping fashion [12] (Figure 1A and S1A) as a result of homotypic repulsion among dendrites [18]. While this dendritic array conceivably allows the nociceptive circuit to spatially resolve noxious stimuli, it was unknown as to whether afferent terminals of C4da neurons are topographically arranged in the CNS.

Figure 1. The presynaptic terminals of the three nociceptive neurons in each larval hemi-segment are arranged in dorsal-to-ventral topography.

Figure 1

Open in a new tab

(A) Cartoons showing the tile-like array of the dendritic territories of the three C4da neurons in the body wall of each hemi-segment. Top: side view. Bottom: drawing of a fillet preparation made by opening up the larva along the dorsal midline. The dendritic territories of the dorsal (D) neuron ddaC, middle (M) neuron v’ada, and ventral (V) neuron vdaB are represented as blue, purple, and green tiles, respectively. (B) Schematic representation of the transgenes for multi-color labeling of single C4da neurons by the Potts’ assay. The asterisks indicate stop codons. Random recombination between the two FRT sites in each cassette, induced by heat-shock, leads to stochastic expression of tdTomato and GFP in C4da neurons. If the recombination occurs in the tdTomato containing cassette only, the neuron expresses only tdTomato (red). The same is true for the GFP cassette. If both cassettes undergo recombination, the neuron appears as yellow. The neurons that do not express either tdTomato or GFP express CD2, which can be labeled by immunostaining with antibodies conjugated with another fluorophore (shown in blue in panel C). (C) Presynaptic arbors of pairs of C4da neurons labeled by the Potts’ assay. The dendritic territory of each clone is schematically represented in the cartoons of larva. The presynaptic terminals of the D, M, and V C4da neurons in each hemi-segment are arranged in dorsal-to-ventral topography, which is visible in the side view, but not the top view. The V neurons usually project a commissural branch (yellow triangles), which extends dorsally after passing the C4da neuropil. Scale bar: 10 μm for top view and 2 μm for side view. Grayscale images with the green and red channels separated are shown in Figure S1E. (D) Locations of single C4da terminals (green) along the dorsal-ventral axis, as visualized by the MARCM technique incorporated with the ppk-tdTomato as the reference for the C4da neuropil (magenta). Scale bar: 10 μm for top view and 2 μm for side view. (E) Statistical analysis of the C4da topography. Each dot represents the topographic index of a C4da MARCM clone. The error bars indicate mean ± SEM. (F) A cartoon showing the organization of D, M, and V axon terminals in the VNC. The axons of the three C4da neurons in each hemi-segment project to the VNC in one nerve. On entering the VNC, the axon of the dorsal neuron (blue) immediately separates from those of the middle (magenta) and ventral (green) neurons. It enters the C4da neuropil through the dorsal boundary and terminates in the dorsal portion of the C4da neuropil. In contrast, the middle and ventral axonal terminals are topographically indistinguishable until they reach the C4da neuropil.

We examined the spatial relationship among the presynaptic terminals of the three C4da neurons in each hemi-segment. In the Drosophila ventral nerve cord (VNC), synaptic connections reside in the neuropil, which is roughly at the center of the VNC [19] (Figure S1B). The neuropil in each hemi-segment is divided into areas that are responsible for different sensory modalities with the C4da terminals residing in the most ventral and medial part of the neuropil [20]. The axon terminals of C4da neurons collectively form a ladder-like structure along the anterior-posterior axis (Figure S1C) [21, 22]. The presynaptic terminals of the three C4da neurons in a hemi-segment, which are the dorsal neuron ddaC (D), middle neuron v’ada (M), and ventral neuron vdaB (V), are confined in a compact, synapses-enriched neuropil (Figure S1C–S1D) [11], termed “C4da neuropil”, with a dorsal-ventral distance of only 4.45 ± 0.85 μm (n = 33) (Figure S1B). Upon entering the VNC, the D axon was immediately separated from the M and V axons by projecting dorsally (Figure S1B), and then entered the C4da neuropil through the dorsal boundary.

To determine the relative locations of presynaptic terminals of the adjacent C4da neurons in such a small space, we designed a multi-color clonal labeling technique, termed Potts’ assay, based on the genetic mosaic approach FLP-out [23] (Figure 1B). By controlling the expression level and timing of the flippase through heat-shock, the FRT sites in each FLP-out cassette can recombine in random cells, leading to the expression of a fluorescent protein. The FLP-out cassettes are driven by the UAS-promoter. Consequently, by using a C4da-specific GAL4 driver, ppk-GAL4 [13], we were able to restrict the randomly labeled cells to only C4da neurons. We integrated two different FLP-out cassettes for tdTomato and GFP, receptively. Thus, FRT/Flippase-based random recombination leads to a stochastic expression of tdTomato and GFP in C4da neurons, allowing sparse labeling of random C4da neurons with green, red, or yellow fluorescence. Using this technique, we found that, consistent with the dendritic field coverage, the presynaptic terminals of the D, M and V neurons were located in the dorsal, middle and ventral portion of the C4da neuropil, respectively (Figure 1C and S1E).

The Potts’ assay does not allow loss-of-function analysis of genes in single neurons, which is required for studying the molecular mechanism of fine-scale topography. We thus used the mosaic analysis with a repressible cell marker (MARCM) [24], in conjunction with ppk-GAL4, to locate the synaptic terminals of single C4da neurons in the C4da neuropil (Figure 1D). We integrated a transgene that expresses the red fluorescent protein tdTomato, directly driven by the C4da-specific ppk promoter (ppk-tdTomato) [13], into the MARCM system. In this modified MARCM system, the position of a C4da terminal in a C4da neuropil can be determined by comparing to the reference ppk-tdTomato (Figure 1D). Because each presynaptic terminal forms a convoluted 3-dimensional (3-D) structure inside the C4da neuropil, it is insufficient to describe the topography of C4da terminals with 2-D analysis. We thus designed a 3-D image quantification algorithm to automatically determine the relative position of each presynaptic terminal, quantified as a topographic index (TI), inside a C4da neuropil (see details in Experimental Procedures, and Figure S1G). Statistical analysis of TI showed that the presynaptic terminals of the D, M, and V neurons ended, respectively, in the dorsal, middle, and ventral portions of the C4da neuropil in early 3rd instar larvae (Figure 1E). Thus, the presynaptic terminals of the nociceptive C4da neurons form a continuous topographic arrangement in the VNC (Figure 1F).

Analysis of the C4da topography at distinct developmental stages showed that the M and V terminals were indistinguishable from each other in the first instar stage, but are significantly separated in the early 2nd instar stage (Figure S1F), suggesting that the C4da topography is established gradually during development.

Topographic separation of middle and ventral terminals requires neuronal activity in C4da neurons

Neural activities, including spontaneous and sensory input-evoked activities, directs fine-scale topography in vertebrate sensory systems [1–4, 25]. The retinotopic map in adult flies has been an important system for molecular analysis of fine-scale topography [26], but the establishment of fly retinotopic map is independent of neural activity [7]. We tested whether the Drosophila nociceptive map is regulated by neural activities, by inhibiting or activating individual C4da neurons in each hemi-segment. Expression of the inward rectifier potassium channel Kir2.1 [27, 28] in C4da neurons robustly inhibited both spontaneous and light-evoked activity (Figure 2 A and 2B), providing us a tool to inhibit C4da neurons. The ppk promoter is active in as early as the stage 16 embryos (data not shown). Inhibiting single M neurons with Kir2.1 in the Potts’ assay, starting at the 1st instar stage (by inducing FLP/FRT-mediated recombination with heat-shock), shifted the presynaptic terminals of M neurons to the ventral portion of the C4da neuropil (Figure S2A). In contrast, Kir2.1 expression in either D or V neurons did not change the topography of their presynaptic terminals. Results from MARCM experiments quantitatively confirmed the results of Potts’ analysis (Figure 2C). The topographic separation of M and V terminals, but not that of D from M or V, was eliminated by Kir2.1 expression as a result of a ventral shift of the M terminals (Figure 2C). In these experiments, Kir2.1 expression was detectable by immunostaining at the 1st instar stage (data not shown). Replacing Kir2.1 with a constitutively open mutant of Drosophila rectifier potassium channel 1 (dORKΔ-C) [28] had similar effects (Figure 2C). Conversely, enhancing the neural activity of single V neurons by thermal activation of dTrpA1 [29] resulted in a dorsal shift, while enhancing activity in M neurons did not change their topographic location (Figure 2D). These results suggest that neural activity regulates fine-scale topography of the M and V terminals. While starting Kir2.1 expression in early 1st and 2nd instar larval stages led to a significant ventral shift of the M terminals, starting Kir2.1 expression in early 3rd instar larvae did not change the topographic locations of the M terminals (Figure S2B).

Figure 2. Topographic separation of middle and ventral terminals requires neuronal activity in C4da neurons.

Figure 2

Open in a new tab

(A) Kir2.1 inhibits spontaneous activities of C4da neurons. Shown are representative traces (top) and quantification (bottom) of the spontaneous action potentials of wild-type (n=23) and Kir2.1-expressing (n=5) C4da neurons. (B) Kir2.1 inhibits light-evoked activities of C4da neurons. Shown is the statistical analysis of firing frequency changes of C4da neurons stimulated with 340, 387, 466, 531, and 628 nm light. The recordings were from M and V neurons, and were combined for statistical analysis. (C) Inhibiting neural activity in single M neurons leads to a ventral shift of the M terminals. The MARCM technique was used to express mCD8::GFP as the wild-type control (wt), Kir2.1::GFP, dORKΔ-NC, and dORKΔ-C. dORKΔ-NC is a non-conducting pore mutant of dORK, which is used as a negative control for dORKΔ-C. Left: representative images of the dorsal-ventral view of single C4da terminals (green) and the C4da neuropil marked by ppk-tdTomato (magenta). Right: statistical analysis of topographic index. Scale bar: 2 μm. (D) Enhancing neural activity in single V neurons by thermal activation of dTrpA1 results in a dorsal shift of the V terminals. Error bars indicate mean ± SEM.

Decreasing or increasing neural activity did not affect the branching or extension of C4da neurons (Figure S2C–E and data not shown). Moreover, while Kir2.1 expression led to an increase in the volume of both M and V synaptic terminals – measured as a ratio of the total number of voxels in the 3D image of the clone to that of the neuropil (Figure S2F), it only affected the topographic locations of the M terminals (Figure 2C), suggesting that there is no correlation between the changes in the volume and topographic locations of presynaptic terminals. This notion is confirmed by the observation that dORKΔ-C expression, which also preferentially affects the locations of the M terminals (Figure 2C), did not affect the volume of either M or V terminals (Figure S2F). Thermal activation of the neurons with dTrpA1 led to a small but significant reduction in the volume of presynaptic terminals (Figure S2G). We also analyzed the dorsal and ventral boundaries of each presynaptic terminal. The distance between the dorsal and ventral boundaries of each presynaptic terminal changed in a manner that is consistent with the changes in the volume of the terminals (Figure S2H and S2I). Taken together, these results demonstrate at single-neuron level that neural activity does not regulate topographic maps by simply restraining the size of axonal arbors [30, 31]. It is noteworthy that the topographic index is a better measurement of the topography of a synaptic terminal, compared to the boundary analysis. If the synaptic terminals are of regular shapes, analyzing the locations of dorsal and ventral boundaries would provide additional information regarding the sizes of the terminals. However, due to the convoluted morphology of the presynaptic terminals, there are often “holes” in the 3-D images. As a result, a change in boundary location does not necessarily reflect a change in voxel density in the 3-D space. Different from the boundary analysis, the topographic index reflects the voxel density and is not affected by the shape of the presynaptic terminal.

Drosophila Trim9 regulates the topographic projection of C4da sensory afferents

To investigate the molecular mechanisms underlying the topographic projection of C4da afferents, we tested a number of genes known to regulate presynaptic arbor development and found that the Drosophila ortholog of the tripartite motif protein Trim9, dTrim9 (also termed anomalies in sensory axon patterning, or Asap) [11], regulates the topographic projections of C4da terminals. Trim9 is a member of the TRIM protein family, which share a RING domain in the N-terminal region followed by two B-boxes and a coiled-coil domain [32]. In mammals, Trim9 is specifically expressed in the nervous system [8, 9] and is required for axon branching in response to DCC signaling [10]. The C. elegans ortholog of dTrim9, MADD-2, regulates axon branching and has a mild effect on ventral attractive guidance [10]. In Drosophila, the protein levels of dTrim9 regulate the formation of the contra-lateral projections in the D and V neurons but not the M neurons [11]. dTrim is expressed at higher levels in D and V neurons than in the M neuron [11], providing a molecular basis to differentiate the C4da neurons in each hemi-segment. We thus tested the possibility that dTrim9 regulates the topographic projections of C4da neurons. Consistently, we found 48.8% of dTrim9−/− V neurons lack the contra-lateral projections, compared to 2.5% in wild-type V neurons. Conversely, overexpressing one copy of a dTrim9 transgene in the M neuron led to 11.8% of these neurons forming contra-lateral projections, compared to 0% in wild-type M neurons. Presynaptic terminals of single V neurons homozygous for dTrim9 null mutations shifted dorsally to the middle portion of the C4da neuropil, while those of dTrim9 null M neurons remained indistinguishable from the wild-type M neurons (Figure 3A). The terminals of dTrim9 null D neurons were not affected. We analyzed both neurons with normal patterns of presynaptic terminals and those with contra-lateral projection defects, and found both displayed a dorsal shift of the V terminals, suggesting that the topographic defect is separable from defects in contra-lateral projections. Conversely, overexpressing dTrim9 led to a ventral shift of the M terminals (Figure 3A and Figure S3A). In both dTrim9 loss-of-function and overexpression neurons, the topographic separation between the M and V terminals was abolished (Figure 3A). These results suggest that dTrim9 regulates the topography of the M and V terminals.

Figure 3. dTrim9 regulates activity-dependent topography of C4da sensory afferents.

Figure 3

Open in a new tab

(A) dTrim9 regulates topographic projections of C4da terminals. Left: representative images of the dorsal-ventral view of single C4da terminals (green) in C4da neuropils (magenta). Right: statistical analysis of topographic index. (B) Neuronal inhibition requires dTrim9 to alter the topography of presynaptic terminals. Left: representative images of the presynaptic terminals of dTrim9−/− M and V neurons that overexpress Kir2.1. Right: statistical analysis of topographic index. The data on Kir2.1 is the same as that in Figure 2C; that on wild type and dTrim9−/− is the same as that in Figure 3A. Error bars indicate mean ± SEM. Scale bars: 2 μm.

Neural activity regulates Trim9 levels to control fine-scale topography

Because inhibiting neuronal activity resulted in a topographic defect that was opposite to dTrim9 loss-of-function but similar to dTrim9 overexpression, we did genetic epistasis tests to determine whether the topographic effects of neuronal inhibition requires dTrim9 function. Using the MARCM technique, we overexpressed Kir2.1 in dTrim9−/− C4da neurons. Presynaptic terminals of dTrim9−/− V neurons with inhibited activity shifted dorsally to the middle position (Figure 3B), which phenocopied dTrim9−/− mutant neurons that had normal activity. Overexpressing dORKΔ-C in dTrim9−/− C4da neurons led to similar results (data not shown). Overexpressing both dTrim9 and dTrpA1 (with thermal activation at 30 °C) led to a ventral shift of the M terminals, which is similar to overexpressing dTrim9 only (Figure S3B). These results suggest that the topographic defects induced by activity inhibition require dTrim9. Again, no correlation was observed between the sizes of presynaptic terminals and their topography (Figure S3C–G).

Because dTrim9 levels determine the locations of the M and V terminals, we tested the hypothesis that neuronal activity regulates dTrim9 levels in C4da neurons. Immunostaining with an anti-dTrim9 antibody [11] showed that expressing Kir2.1 or dORKΔ-C in M neurons led to an increase in dTrim9 levels, eliminating the difference in dTrim9 levels between the inhibited M neurons and the wild-type V neurons in the same hemi-segment (Figure 4A and 4B). Conversely, increasing neural activity in single V neurons by thermal activation of dTrpA1 led to a reduction of dTrim9 levels, which greatly reduced the difference between the M and V neurons (Figure 4 B and 4D). Taken together, these results suggest that dTrim9 levels are regulated by neural activity and direct the fine-scale topography of the C4da presynaptic terminals (Figure 4E).

Figure 4. Neural activity regulates dTrim9 protein levels.

Figure 4

Open in a new tab

(A) dTrim9 protein levels are increased in M neurons inhibited by Kir2.1. A Kir2.1::eGFP transgene was specifically expressed in single M neurons by the MARCM technique. Kir2.1-expressing M neurons were identified by GFP fluorescence (not shown). The V neuron in the same hemi-segment was identified by the C4da-specific marker ppk-tdTomato. Scale bars: 6 μm. (B) dTrpA1-mediated thermal activation (30 °C) of V neurons leads to reduced dTrim9 levels. A dTrpA1 transgene specifically was expressed in single V neurons by MARCM. The MARCM clones were identified by GFP expression (not shown). (C–D) Statistical analysis of the ratio of dTrim9 immunofluorescence intensity between the M and V neurons in the same hemi-segment. Data presented as mean ± SEM. (E) A model that summarizes the findings in this study. The presynaptic terminals of the M and V C4da neurons project to the VNC in a way that matches the dorsal-to-ventral locations of dendritic fields on the larval body wall. Neuronal activity differentially regulated the protein levels of dTrim9 in the M and V neurons, which in turn direct the dorsal-to-ventral topography of the presynaptic terminals.

The assembly of topographic maps in vertebrates is known to be controlled by genetic programs as well as neural activity [1]. In this study we demonstrated at the level of single identified neurons that neuronal activity directs the fine-scale topography of the Drosophila somatosensory system. Using this system, we have identified an intracellular signaling molecule that mediates the regulation of neural activity in establishing the topography of adjacent afferent neurons. This study not only offers a novel model that explains how neural activity directs topographic projections of sensory afferents, but also provides an instrumental system for investigating the interactions between neural activity and genetic programs in neural circuit assembly.

EXPERIMENTAL PROCEDURES

Potts’ and FLP-out assays

For Potts’ assays (named after William Potts, the Detroit police officer who invented the traffic light), we generated transgenic lines that carry FLP-out cassettes UAS-FRT-rCD2-stop-FRT-mCD8::GFP and UAS-FRT-rCD2-stop-FRT-CD4::tdTomato. These transgenes were integrated into flies together with a Flippase transgene driven by a heat-shock promoter and the C4da-specific driver ppk-GAL4. For clonal overexpression of Kir2.1 in the Potts’ assay, a UAS-FRT-rCD2-stop-FRT-Kir2.1::eGFP FLP-out cassette was used in place of UAS-FRT-rCD2-stop-FRT-mCD8::GFP. Eggs were collected, allowed to develop at 25 °C for 24 hrs to become early 1st instar larvae, and heat-shocked at 37.5 °C for 20 minutes. The larvae were then allowed to develop for 2 days to become early 3rd instar larvae, and dissected for immunostaining. A sample useful for analyzing topographic mapping contains one GFP-positive and one tdTomato-positive C4da neuron in a hemi-segment. It is important that no clone in the adjacent segments is labeled, as these clones extend terminals into neighboring C4da neuropils. To obtain successful labeling, the clone rate was kept low and a large number of larvae were screened.

For FLP-out assays, flies carrying either UAS-FRT-rCD2-stop-FRT-mCD8::GFP (for labeling wild-type neurons) or UAS-FRT-rCD2-stop-FRT-Kir2.1::eGFP (for labeling neurons that express Kir2.1 to inhibit neuronal activity) were crossed with flies carrying hs-FLP and ppk-GAL4. Procedures for inducing single C4da clones are the same as described above. The FLP-out assays can be used to analyze topography because the C4da neuropils can be labeled by immunostaining the rCD2 in the FLP-out cassette.

For overexpressing dTrim9 at different developmental stages (Figure S3A), a modified FLP-out technique with an excisable GAL80 [33] was used to express a UAS-dTrim9 transgene under the control of ppk promoter.

MARCM for analyzing topographic mapping

MARCM experiments were done as previously described [34]. We dissected only size-matched 3-day-old 3rd-instar larvae to ensure consistency of the developmental stages of the analyzed animals. Three hours after egg laying the eggs were heat-shock at 37.5 °C for 15~20 minutes. For using MARCM to overexpress dTrpA1 in single C4da neurons, the eggs were kept at 30 °C after heat-shock and dissected two-and-a-half days later. The same procedure was followed for the control experiments.

Immunostaining, imaging, image preprocessing, and image analysis

Third-instar larvae were immunostained as described [35] with minor modifications. The anti-dTrim9 antibodies were preabsorbed with the fillet preparation from dTrim9 null allele asap91 3rd-instar larvae. All images were collected as 3-D stacks using a Leica SP5 confocal system (Leica Microsystems) equipped with a 63X oil lens (Leica, Plan-Apochromat, NA = 1.4). The axial sampling step (z-step) is 0.3 μm. Images were collected with minimum signal saturation. Three steps were necessary to pre-process images for analyzing the topographic index, volume, and boundary location of presynaptic terminals. First, confocal image stacks were deconvolved with Huygens software (Scientific Volume Imaging). Next, the VNC in each stack was aligned to uniform orientation with the 3-D image analysis software Amira (Visualization Science Group). Third, the 3-D image stacks were cropped to contain only the C4da neuropil with the single MARCM or FLP-out clones in them. After preprocessing, the image stack was automatically analyzed by custom-designed software for topographic index, volume ratio, and boundary location (see below).

To quantify dTrim9 levels in C4da neurons, the mean fluorescence intensity of dTrim9 immunosignals in the cell bodies were measured with the LAS AF Lite software. Neurolucida software (Visage Imaging) was used to quantify the length and branch numbers of presynaptic arbors.

Algorithm for analyzing the topographic index and volume of presynaptic terminals

The neuropil channel for topographic index (TI) and volume ratio calculation was obtained by combining the raw neuropil signal with the clone channel signal. Since it was typically dim, the neuropil channel was enhanced by iterative histogram normalization: The maximum-intensity parameter during normalization was iteratively adjusted so that the mean foreground intensity was increased to 80. The extraction of Signals of C4da neuropil and clone was extracted from background by using Robust Adaptive Threshold Selection (RATS) [36] (http://rsb.info.nih.gov/ij/plugins/rats/index.html).

RATS is a segmentation method for extracting the foreground object out of a gray level image based on robust and adaptive thresholding. It selects the thresholds by recursively dividing the image using quadtree structure and then automatically calculating the thresholds using intensity values as well as their gradients in the local region. The thresholds for all local regions are then bilinearly interpolated across the entire image. The advantage of the RATS approach is its local adaptability which suits well to microscopic images with contrast variation among different local regions. RATS was applied to each slice of the 3D image stack. The minimum region size (also called leaf size) of the quadrant was set to 5 pixels per side. Figure S1G exemplifies the flow of image processing using a dorsal neuron with MARCM analysis. Images with the FLP-out clone were processed in the same way.

The TI of each clone voxel was calculated by measuring its relative position between dorsal and ventral neuropil boundaries: TIi = di/(di + vi), where di was the distance of the voxel i to the dorsal boundary and vi was its distance to the ventral boundary. As such, the voxels at the dorsal side of the neuropil had TIs closer to 0; the voxels at the ventral side of the neuropil had TIs closer to 1. The overall topographic index of a clone was the averaged sum calculated by TI = ΣTIi/n, where n was the total number of clone voxels in the 3D image stack. Note that an overall TI isin; (0, 1) can never be 0 or 1. The volume ratio was the ratio between the volumes of clone and neuropil, with the volume being measured by the number of foreground voxels in the 3D volume. The volume of a clone or neuropil was represented by the total number of voxels in the 3D images of the clone or neuropil. Thus, the result of this analysis is not affected by the “holes” in the 3D image of a presynaptic terminal caused by the convoluted morphology of the terminal.

The average dorsal boundary position of each clone was calculated by taking the mean of the normalized positions of the dorsal-most voxels of the clone. To obtain the normalized position, we measured the distance (di) between the dorsal-most voxel of the clone and that of the neuropil at the same x position, and the distance between the dorsal-most and ventral-most voxels of the neuropil at that x position (Di). The normalized boundary position (Bi) at a particular x position is calculated as the ratio between the two distances: Bi = di/Di. The average boundary position of a clone was the averaged sum of the Bi for all x position.

The software for quantifying TI, volume ratio, and boundaries was developed as an NIH ImageJ plugin.

Electrophysiology

Extracellular recording of C4da neuronal activities was done as described previously with minor modifications [16]. Briefly, age-synchronized early 3rd- instar larvae carrying both ppk-GAL4 and UAS-mCD8::GFP transgenes were dissected to make fillet preparations in the external saline solution composed of (in mM): NaCl 120, KCl 3, MgCl2 4, CaCl2 1.5, NaHCO3 10, trehalose 10, glucose 10, TES 5, sucrose 10, HEPES 10. The osmolality was 305 mOsm kg−1; the pH was 7.25. After gentle proteinase (type XXIII, Sigma) digestion of muscles, the GFP-positive (i.e., C4da) neurons were identified using a Zeiss D1 microscope with a 40X/1.0 NA water-immersion objective lens with the assistance of a CoolSNAP K4 CCD camera (Photometrics). Gentle negative pressure was delivered to suck the soma into a recording pipette (5-μm tip opening) containing external saline solution. Extracellular recordings of action potentials were obtained in voltage-clamp mode at a holding potential of 0 mV, a 2 kHz low-pass filter, and a sampling frequency of 20 kHz with a 700B amplifier (Molecular Devices). Spontaneous activities of C4da neurons were obtained from a 10-min gap-free recording. For AITC stimulation, AITC was applied to the chamber to a final concentration 100 μM. For light-evoked responses, a 300-W xenon light source was connected to the microscope with a liquid light guide to provide light stimulation through the lens, yielding an evenly illuminated light spot of 600-μm diameter that covered an entire C4da neuron. Light intensity used (in mW/mm2) was 2.84 for 340 nm, 14.6 for 387 nm, 103.8 for 466 nm, 57.7 for 531 nm, 9.3 for 628 nm. The duration (5 sec) of light illumination was controlled by a shutter in the xenon lamp house triggered by Digidata 1440A (Molecular Devices). Band-pass excitation filters (Semrock) were used to select light wavelength. For each recording trace, average frequency during the 5 sec immediately before light exposure was used as control. Five-sec light stimulation was controlled by a TTL-triggered shutter (Sutter Instruments) in the xenon lamp house.

Statistical analysis

Pairwise comparisons were performed with t-tests between two groups, and one way ANOVA analyses with Sidak correction were used for comparing three or more groups. N.S.: not significant, P > 0.05; * P < 0.05; ** P < 0.01; *** P < 0.001.

Supplementary Material

01

HIGHLIGHTS.

  • Nociceptive afferents in Drosophila larva form fine-scale topography.

  • Neural activity regulates larval nociceptive topography.

  • Trim9 directs the fine-scale topography of larval nociceptive afferents.

  • Neural activity regulates Trim9 expression to direct fine-scale topography.

Acknowledgments

We thank Dr. Barry Dickson for generously sharing fly stocks. We also thank Drs. Charles Zucker, Shawn Xu, Hisashi Umemori, and Jung Hwan Kim for critical comments on earlier versions of the manuscript, and Yonghua Wang for making the constructs for Potts’ analysis. This work was supported by a grant from the Natural Science Foundation of China (31200839) to L.Y., grants from Worcester Foundation and NIH (R01MH103133) to Y.X., a grant from NIH to J.Z. (R15MH099569), and grants from NIH (R01MH091186), the Whitehall Foundation, and the Pew Charitable Trusts to B.Y.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.Goodman CS, Shatz CJ. Developmental mechanisms that generate precise patterns of neuronal connectivity. Cell. 1993;72(Suppl):77–98. doi: 10.1016/s0092-8674(05)80030-3. [DOI] [PubMed] [Google Scholar]
  • 2.Katz LC, Shatz CJ. Synaptic activity and the construction of cortical circuits. Science (New York, N Y. 1996;274:1133–1138. doi: 10.1126/science.274.5290.1133. [DOI] [PubMed] [Google Scholar]
  • 3.Ruthazer ES, Cline HT. Insights into activity-dependent map formation from the retinotectal system: a middle-of-the-brain perspective. J Neurobiol. 2004;59:134–146. doi: 10.1002/neu.10344. [DOI] [PubMed] [Google Scholar]
  • 4.Pfeiffenberger C, Yamada J, Feldheim DA. Ephrin-As and patterned retinal activity act together in the development of topographic maps in the primary visual system. J Neurosci. 2006;26:12873–12884. doi: 10.1523/JNEUROSCI.3595-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Clandinin TR, Zipursky SL. Afferent growth cone interactions control synaptic specificity in the Drosophila visual system. Neuron. 2000;28:427–436. doi: 10.1016/s0896-6273(00)00122-7. [DOI] [PubMed] [Google Scholar]
  • 6.Jefferis GS, Marin EC, Stocker RF, Luo L. Target neuron prespecification in the olfactory map of Drosophila. Nature. 2001;414:204–208. doi: 10.1038/35102574. [DOI] [PubMed] [Google Scholar]
  • 7.Hiesinger PR, Zhai RG, Zhou Y, Koh TW, Mehta SQ, Schulze KL, Cao Y, Verstreken P, Clandinin TR, Fischbach KF, et al. Activity-independent prespecification of synaptic partners in the visual map of Drosophila. Current biology: CB. 2006;16:1835–1843. doi: 10.1016/j.cub.2006.07.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Berti C, Messali S, Ballabio A, Reymond A, Meroni G. TRIM9 is specifically expressed in the embryonic and adult nervous system. Mechanisms of development. 2002;113:159–162. doi: 10.1016/s0925-4773(02)00013-8. [DOI] [PubMed] [Google Scholar]
  • 9.Tanji K, Kamitani T, Mori F, Kakita A, Takahashi H, Wakabayashi K. TRIM9, a novel brain-specific E3 ubiquitin ligase, is repressed in the brain of Parkinson’s disease and dementia with Lewy bodies. Neurobiology of disease. 2010;38:210–218. doi: 10.1016/j.nbd.2010.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hao JC, Adler CE, Mebane L, Gertler FB, Bargmann CI, Tessier-Lavigne M. The tripartite motif protein MADD-2 functions with the receptor UNC-40 (DCC) in Netrin-mediated axon attraction and branching. Developmental cell. 2010;18:950–960. doi: 10.1016/j.devcel.2010.02.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Morikawa RK, Kanamori T, Yasunaga KI, Emoto K. Different levels of the Tripartite motif protein, Anomalies in sensory axon patterning (Asap), regulate distinct axonal projections of Drosophila sensory neurons. Proceedings of the National Academy of Sciences of the United States of America. 2011;108:19389–19394. doi: 10.1073/pnas.1109843108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Grueber WB, Jan LY, Jan YN. Tiling of the Drosophila epidermis by multidendritic sensory neurons. Development. 2002;129:2867–2878. doi: 10.1242/dev.129.12.2867. [DOI] [PubMed] [Google Scholar]
  • 13.Grueber WB, Ye B, Moore AW, Jan LY, Jan YN. Dendrites of distinct classes of Drosophila sensory neurons show different capacities for homotypic repulsion. Curr Biol. 2003;13:618–626. doi: 10.1016/s0960-9822(03)00207-0. [DOI] [PubMed] [Google Scholar]
  • 14.Hwang RY, Zhong L, Xu Y, Johnson T, Zhang F, Deisseroth K, Tracey WD. Nociceptive Neurons Protect Drosophila Larvae from Parasitoid Wasps. Current Biology. 2007;17:2105–2116. doi: 10.1016/j.cub.2007.11.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Zhong L, Hwang RY, Tracey WD. Pickpocket Is a DEG/ENaC Protein Required for Mechanical Nociception in Drosophila Larvae. Current Biology. 2010;20:429–434. doi: 10.1016/j.cub.2009.12.057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Xiang Y, Yuan Q, Vogt N, Looger LL, Jan LY, Jan YN. Light-avoidance-mediating photoreceptors tile the Drosophila larval body wall. Nature. 2010;468:921–926. doi: 10.1038/nature09576. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Kim SE, Coste B, Chadha A, Cook B, Patapoutian A. The role of Drosophila Piezo in mechanical nociception. Nature. 2012;483:209–212. doi: 10.1038/nature10801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Grueber WB, Jan LY, Jan YN. Different levels of the homeodomain protein cut regulate distinct dendrite branching patterns of Drosophila multidendritic neurons. Cell. 2003;112:805–818. doi: 10.1016/s0092-8674(03)00160-0. [DOI] [PubMed] [Google Scholar]
  • 19.Landgraf M, Sanchez-Soriano N, Technau GM, Urban J, Prokop A. Charting the Drosophila neuropile: a strategy for the standardised characterisation of genetically amenable neurites. Developmental biology. 2003;260:207–225. doi: 10.1016/s0012-1606(03)00215-x. [DOI] [PubMed] [Google Scholar]
  • 20.Zlatic M, Li F, Strigini M, Grueber W, Bate M. Positional cues in the Drosophila nerve cord: semaphorins pattern the dorso-ventral axis. PLoS Biol. 2009;7:e1000135. doi: 10.1371/journal.pbio.1000135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Grueber WB, Ye B, Yang CH, Younger S, Borden K, Jan LY, Jan YN. Projections of Drosophila multidendritic neurons in the central nervous system: links with peripheral dendrite morphology. Development. 2007;134:55–64. doi: 10.1242/dev.02666. [DOI] [PubMed] [Google Scholar]
  • 22.Ye B, Zhang Y, Song W, Younger SH, Jan LY, Jan YN. Growing dendrites and axons differ in their reliance on the secretory pathway. Cell. 2007;130:717–729. doi: 10.1016/j.cell.2007.06.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Struhl G, Basler K. Organizing activity of wingless protein in Drosophila. Cell. 1993;72:527–540. doi: 10.1016/0092-8674(93)90072-x. [DOI] [PubMed] [Google Scholar]
  • 24.Lee T, Luo L. Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis. Neuron. 1999;22:451–461. doi: 10.1016/s0896-6273(00)80701-1. [DOI] [PubMed] [Google Scholar]
  • 25.Chandrasekaran AR, Plas DT, Gonzalez E, Crair MC. Evidence for an instructive role of retinal activity in retinotopic map refinement in the superior colliculus of the mouse. J Neurosci. 2005;25:6929–6938. doi: 10.1523/JNEUROSCI.1470-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Sanes JR, Zipursky SL. Design principles of insect and vertebrate visual systems. Neuron. 2010;66:15–36. doi: 10.1016/j.neuron.2010.01.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Baines RA, Uhler JP, Thompson A, Sweeney ST, Bate M. Altered electrical properties in Drosophila neurons developing without synaptic transmission. J Neurosci. 2001;21:1523–1531. doi: 10.1523/JNEUROSCI.21-05-01523.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Nitabach MN, Blau J, Holmes TC. Electrical silencing of Drosophila pacemaker neurons stops the free-running circadian clock. Cell. 2002;109:485–495. doi: 10.1016/s0092-8674(02)00737-7. [DOI] [PubMed] [Google Scholar]
  • 29.Hamada FN, Rosenzweig M, Kang K, Pulver SR, Ghezzi A, Jegla TJ, Garrity PA. An internal thermal sensor controlling temperature preference in Drosophila. Nature. 2008;454:217–220. doi: 10.1038/nature07001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Cline HT, Constantine-Paton M. NMDA receptor antagonists disrupt the retinotectal topographic map. Neuron. 1989;3:413–426. doi: 10.1016/0896-6273(89)90201-8. [DOI] [PubMed] [Google Scholar]
  • 31.Simon DK, Prusky GT, O’Leary DD, Constantine-Paton M. N-methyl-D-aspartate receptor antagonists disrupt the formation of a mammalian neural map. Proceedings of the National Academy of Sciences of the United States of America. 1992;89:10593–10597. doi: 10.1073/pnas.89.22.10593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Reymond A, Meroni G, Fantozzi A, Merla G, Cairo S, Luzi L, Riganelli D, Zanaria E, Messali S, Cainarca S, et al. The tripartite motif family identifies cell compartments. The EMBO journal. 2001;20:2140–2151. doi: 10.1093/emboj/20.9.2140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Gordon MD, Scott K. Motor control in a Drosophila taste circuit. Neuron. 2009;61:373–384. doi: 10.1016/j.neuron.2008.12.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ye B, Kim JH, Yang L, McLachlan I, Younger S, Jan LY, Jan YN. Differential regulation of dendritic and axonal development by the novel kruppel-like factor dar1. J Neurosci. 2011;31:3309–3319. doi: 10.1523/JNEUROSCI.6307-10.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ye B, Petritsch C, Clark IE, Gavis ER, Jan LY, Jan YN. Nanos and Pumilio are essential for dendrite morphogenesis in Drosophila peripheral neurons. Curr Biol. 2004;14:314–321. doi: 10.1016/j.cub.2004.01.052. [DOI] [PubMed] [Google Scholar]
  • 36.Wilkinson MHF. Optimizing Edge Detectors for Robust Automatic Threshold Selection: Coping with Edge Curvature and Noise. Graphical Models and Image Processing. 1998;60:385–401. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS579948-supplement-01.pdf (731.6KB, pdf)

RESOURCES