The Astrocyte Network in the Ventral Nerve Cord Neuropil of the Drosophila Third-Instar Larva

The astrocyte network in the ventral nerve cord neuropil of the Drosophila third-instar larva

Item Type Article

Authors Hernandez, Ernesto; MacNamee, Sarah E; Kaplan, Leah R; Lance, Kim; Garcia-Verdugo, Hector D; Farhadi, Dara S; Deer, Christine; Lee, Si W; Oland, Lynne A

Citation Hernandez, E, MacNamee, SE, Kaplan, LR, et al. The astrocyte network in the ventral nerve cord neuropil of the Drosophila thirdinstar larva. J Comp Neurol. 2020; 1– 21. https:// doi.org/10.1002/cne.24852

DOI 10.1002/cne.24852

Publisher WILEY

Journal JOURNAL OF COMPARATIVE NEUROLOGY

Download date 29/09/2021 07:02:47

Item License http://rightsstatements.org/vocab/InC/1.0/

Version Final accepted manuscript

Link to Item http://hdl.handle.net/10150/636762

The astrocyte network in the ventral nerve cord neuropil of the Drosophila third-instar larva Running title: Drosophila VNC astrocyte network Ernesto Hernandez1,2, Sarah E. MacNamee1,3, Leah R. Kaplan1,4, Kim Lance1, Hector D. Garcia- Verdugo1, Dara S Farhadi1,5, Christine Deer1,6, Si Woo Lee1, Lynne A. Oland1 Current addresses: 2Hernandez: University of Illinois at Chicago School of Medicine, 1601 Parkview Ave, Rockford, IL 61107; [email protected] 3MacNamee: Inscopix, 2452 Embarcadero Way, Palo Alto, CA 94303 4Kaplan: Consortium for Science, Policy & Outcomes, Arizona State University, 1800 I St NW, Suite 300, Washington, DC 20006 5Farhadi: University of Arizona, College of Medicine – Phoenix, 475 N 5th St, Phoenix, AZ 85004 6Deer: University of Arizona, University Information Technology Service, Research Technologies Group, Data Visualization Team

Corresponding author: Lynne A. Oland, Ph.D. Department of Neuroscience University of Arizona PO Box 210077 Tucson, AZ 85721-0077 520-621-7215 [email protected]

Data Sharing: The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgements: The authors are deeply grateful to Patty Jansma of the University of Arizona’s Core Imaging Facility who guided our students in use of the confocal and multi- photon microscopes; to Mark Borgstrom of the UITS Statistics Consultation Team who guided our statistical analyses; to Anu Sethuraman, a high school student who began the analysis of cross-segmental arborization; to Cathy Tran, who as an undergraduate carried out many of our immunocytochemistry experiments; and to Leslie Tolbert, who always provided insightful review and guidance of the work in this paper. The image rendering work was funded by University Information Technology Systems (UITS), Research Technologies Group, Visualization Team (funded by Technology Research Investment Fund/Space Exploration and Optical Sciences). The study was funded by NSF, grant #IOS-1353739 to LAO.

Abstract Understanding neuronal function at the local and circuit level requires understanding astrocyte

function. We have provided a detailed analysis of astrocyte morphology and territory in the

Drosophila 3rd-instar VNC where there already exists considerable understanding of the neuronal

network. Astrocyte shape varies more than previously reported; many have bilaterally

symmetrical partners, many have a high percentage of their arborization in adjacent segments,

and many have branches that follow structural features. Taken together, our data are consistent

with, but not fully explained by, a model of a developmental growth process dominated by

competitive or repulsive interactions between astrocytes. Our data suggest that the model also

should include cell-autonomous aspects, as well as use of structural features for growth.

Variation in location of arborization territory for identified astrocytes was great enough that a standardized scheme of neuropil division among the six astrocytes that populate each hemi- segment is not possible at the 3rd instar. The arborizations of the astrocytes can extend across

neuronal functional domains. The ventral astrocyte in particular, whose territory can extend well

into the proprioceptive region of the neuropil, has no obvious branching pattern that correlates

with domains of particular sensory modalities, suggesting that the astrocyte would respond to

neuronal activity in any of the sensory modalities, perhaps integrating across them. This study

sets the stage for future studies that will generate a robust, functionally-oriented connectome that

includes both partners in neuronal circuits – the neurons and the glial cells, providing the

foundation necessary for studies to elucidate neuron-glia interactions in this neuropil.

Key words:

glial cells, neuron-glia interaction

glial cells, neuron-glia interaction RRID:BDSC Cat# 30125, RRID:BDSC_30125 RRID:BDSC Cat# 4775, RRID:BDSC_4775 RRID:BDSC Cat# 6938, RRID:BDSC_6938 RRID:BDSC Cat# 64085, RRID:BDSC_64085 RRID:BDSC Cat# 5692, RRID:BDSC_5692 RRID:BDSC Cat# 38760, RRID:BDSC_38760 RRID:DSHB Cat# 1D4, RRID:AB_528235 RRID:Molecular Probes Cat# 6455, RRID:AB_2314543 RRID:Jackson ImmunoResearch Labs Cat# 115-167-003, RRID:AB_2338709 RRID:Molecular Probes Cat# A-21236, RRID:AB_141725 RRID:Bio-rad Cat # MCA1360, RRID:AB_322378 RRID:Cell Signaling Technology Cat # 3724, RRID:AB_1549585 RRID:Novus Cat # NBP1-06712, RRID:AB_1625981 RRID:Abcam Cat# ab6953, RRID:AB_955010 RRID:DSHB Cat# nc82, RRID:AB_2314866 RRID:Thermo Fisher Scientific Cat# A-11034, RRID:AB_2576217

Main Text Introduction Concerted efforts toward mapping neuronal connectomes have generated rapid advances in our

knowledge of brain circuitry and organization. As the number and details of various neuronal

connectomes in different species increase, it is critically important that glial cells, which help

determine and then regulate both anatomical and functional neuronal connectivity, be included in

these maps (Corty & Freeman, 2013; Clarke & Barres, 2013; Fields, et al., 2015). Given their association with synapses and their organization into domains that may modulate local neuronal

activity and synaptic interactions, astrocytes are particularly well-positioned to play a role in the

spatial segregation of information as well as the spread of information (Bushong, et al., 2002;

Halassa, et al., 2007). Here, we have interrogated the anatomical features of astrocytes that

comprise the glial network in the ventral nerve cord (VNC) neuropil of the third-instar larva. The

VNC neuropil features well-described neuronal sensori-motor maps (Schrader and Merritt, 2000;

Zlatic et al., 2009), and these territories are readily identifiable using FasII-positive axon tracts as

anatomical landmarks (Landgraf et al., 2003b). The 3rd instar is the stage at which the bulk of

studies that examine the physiology of neurons have been carried out (Choi, et al., 2004; Worrell

& Levine, 2008; Heckscher, et al., 2015; Ohyama, et al., 2015; Fushiki, et al., 2016; Clark, et al.,

2016; Tastekin, et al., 2018) as well as the only stage to date in which the electrophysiological characteristics of astrocytes and astrocyte-synapse interactions have been examined (MacNamee, et al., 2016.)

Of the glial subtypes commonly recognized in the Drosophila CNS (Hartenstein, 2011; Freeman,

2015; Kremer, et al., 2017), here we have focused on the astrocyte-like glia. While several of the glial subtypes in Drosophila may contribute to what are the functions of astrocytes in vertebrates, including Drosophila cortex, ensheathing and astrocyte-like glia, the morphological, developmental, molecular, and functional similarities between Drosophila astrocytes and vertebrate protoplasmic astrocytes are well-documented (Freeman, 2015; Ma, et al., 2016;

MacNamee, et al., 2016; Peco, et al., 2016; Yildirim, et al., 2019;). Morphologically, these species’ astrocytes share a highly branched structure, with processes that are strikingly complex and include veil-like distal processes (leaflets) found in the vicinity of neuronal synapses. During development, both vertebrate astrocytes and Drosophila astrocytes participate in the formation of synaptic networks by regulating synaptogenesis (Eroglu, et al., 2009; Molofsky, et al., 2012;

Allen, et al., 2012; Muthukumar, et al., 2014; Stork, et al., 2014) and by pruning synaptic spines

(MacDonald et al., 2006; Doherty, et al., 2009; Chung, et al., 2013). They contribute to the

determination of neuronal cell number and can promote, prune and/or dynamically regulate

synaptic connectivity through various mechanisms (vertebrates: Eroglu & Barres, 2010;

invertebrates: Corty & Freeman, 2013). In Drosophila, ablation of specific astrocytes during

pupal-stage synaptogenesis has both anatomical (reduced synapse number and density) and

behavioral (failed or partial eclosion) consequences (Muthukumar, et al., 2014).

Astrocytes functionally regulate synaptic strength in both vertebrate and fly nervous systems

(Theodosis, et al., 2008; Clark & Barres, 2013; Liu, et al., 2014; Verkhratsky & Nedergaard,

2018). Astrocytic processes reside near synapses, and Drosophila astrocytes express molecular

markers that are shared with vertebrate astrocytes, including the neurotransmitter transporters

Eaat1 (Rival, et al., 2004; 2006) and Gat (Muthukumar, et al., 2014). Astrocyte cell-body recordings in Drosophila show transporter-mediated responses to glutamatergic neuronal activity and cell-intrinsic physiology that is strikingly similar to that of vertebrate protoplasmic astrocytes (MacNamee, et al., 2016). Also, Drosophila astrocyte activity has been shown to regulate nervous system function at the level of synaptic signaling (Liu, et al., 2014; MacNamee, et al., 2016), like vertebrate astrocytes (Panatier, et al., 2011), as well as at the level of behavior

(Grosjean, et al., 2008; Stacey, et al., 2010).

Here, we sought to understand key features of the morphology of the astrocyte network in the

3rd-instar VNC. By analogy with the uniquely identified neurons in invertebrate nervous systems

that were so richly exploited to yield key insights into the function of individual neurons and the circuits in which they engage (e.g., Goodman, et al., 1985; Pfluger & Watson, 1988; Boyan &

Ball, 1993; Keshishian & Chiba, 1993), the small number of astrocytes in this Drosophila

nervous system may prove similarly useful. The flexibility and power of Drosophila as an experimental system create a unique opportunity to explore relationships between anatomical and functional astrocyte properties, to ask critical questions about how the individual astrocytes

in the network operate locally, how their spatial domains are established and regulated, how

astrocyte territories relate to functional neuropil compartments, and how astrocytes individually

affect and share information about neuronal activity.

As details of the neuronal connectome increase, it is essential that glial cells are included in these

connectome maps (Corty & Freeman, 2013; Fields, et al., 2015). The astrocyte data presented

here, integrated with the Drosophila neuronal connectome (Schneider-Mizell, et al., 2016;

Gerhard, et al., 2017), will serve as a foundation for studies that will advance our understanding of the role of neuron-astrocyte interactions in circuit function.

Methods Animals The following stocks were obtained from the Bloomington Drosophila Stock Center (Indiana

University, Bloomington, IN): repo-gal4 x UAS-IVS-mCD8:GFP, UAS-(FRT.stop)mCD8-GFP

(BDSC Cat# 30125, RRID:BDSC_30125), UAS-GFP.nls (BDSC Cat# 4775,

RRID:BDSC_4775), hs-70FLP (BDSC Cat# 6938, RRID:BDSC_6938) and hs-FLP; + ; UAS-

(FRT.stop)myr::smGdP-HA, UAS-(FRT.stop)myr::smGdP-V5, UAS-(FRT.stop)myr::smGdP-

FLAG (BDSC Cat# 64085, RRID:BDSC_64085) (Viswanathan et al., 2015) and w;UAS-CD4- tdGFP;looper-GAL4 (RRID: BDSC Cat# 38760, RRID:BDSC_38760). The alrm-GAL4 (w;

UAS-GFP.nls/+; Alrm-GAL4/+ (BDSC Cat# 5692, RRID:BDSC_5692) (Doherty et al., 2009) flies used to drive expression in Drosophila astrocytes were a gift from M. Freeman (Vollum

Institute, OHSU). Experimental animals used in this study were trans-heterozygous for

GAL4/UAS insertions. Stocks were reared on standard cornmeal-agar yeast fly food at 22°C on a

12-hour:12-hour light:dark cycle. For all experiments, we used 3rd-instar wandering larvae,

characterized by their crawling out of the food onto the walls of vials. For Multi-color FLP-out

(MCFO) and standard FLP-out experiments, animals were reared at 25oC.

FLP-out and Multi-color FLP-out labelling Table 1 provides details of the primary antibodies used.

Two FLP-out approaches were used, both using the astrocyte driver alrm-GAL4 to target

astrocytes (Doherty, et al., 2009). The FLP-out technique (Ito, et al., 1997) was used to drive

expression of membrane-targeted GFP in a small number of astrocytes. The second approach,

labeling multiple astrocytes with different fluorophores, was achieved using the multi-color FLP-

out variation (MCFO) (Nern, et al., 2015).

FLP-out labeling No heat shock was used for single color (GFP) FLP-out labeling, as hs-FLP induction at 25°C generated sufficiently sparse labeling. Larval CNSs were dissected in HL3.1 solution (Feng, et al., 2004) and placed in baskets with mesh bottoms. They were fixed in 4% paraformaldehyde in

0.1M phosphate buffer (PB, pH 7.2) for 45 minutes at room temperature (RT) and then washed 3 x 10 minutes with 0.1M PB. Tissue was incubated in a blocking-and-permeabilization solution containing 0.5% Triton X-100 and 2% BSA in 0.1M PB for 30 minutes, and then incubated in

the dark for 5 nights at 4oC on a rotator in PB with 0.5% Triton with primary antibodies: mouse

anti-FasII (DSHB Cat# 1D4, RRID:AB_528235) at 1:200 and rabbit anti-GFP Alexa 488

(Molecular Probes Cat# 6455, RRID:AB_2314543) at 1:1000. Tissue was washed 6 x 15 minutes

in 0.1M PB at RT, and then incubated in either goat anti-mouse Cy3 (Jackson ImmunoResearch

Labs Cat# 115-167-003, RRID:AB_2338709) at 1:500 or goat anti-mouse Alexa Fluor 647

(Molecular Probes Cat# A-21236, RRID:AB_141725) at 1:500 on a rotator overnight (ON) at

4oC. Tissue was washed 5 x 15 minutes with 0.1M PB at RT, transferred to microscope slides, and washed with 50% glycerol in HPLC water for 15 minutes, then placed in mounting solution

containing 80% glycerol in HPLC water.

Multi-color flip-out (MCFO) labeling Tissue was prepared in accord with Nern, et al. (2015). Larval CNSs were dissected in cold

phosphate-buffered saline (PBS), placed in mesh-bottomed baskets in Bouin’s fixative solution

(Ramachandran & Budnik, 2010), and microwaved (Pelco Laboratory Microwave Processor, model 3450) at low power (2 minutes on, 2 minutes off, 2 minutes on) at 18oC for a total time of

15 minutes. The tissue was washed 2 x 10 minutes in PBS with 1% Triton X-‐100 (1% PBST),

and washed 4 x 15 minutes in PBS, all on a shaker. Tissue was incubated on a shaker in a

blocking-and-permeabilization solution containing 5% normal donkey serum diluted in 1%

PBST for 2 hours at RT, then placed in primary antibody solution and microwaved at 18oC at

low power (3 minutes on, 2 minutes off, and 3 minutes on). Primary antibody solution included

mouse anti-V5 (Bio-rad Cat # MCA1360, RRID:AB_322378) at 1:200, rabbit anti-HA (Cell

Signaling Technology Cat # 3724, RRID:AB_1549585) at 1:200, and rat anti-FLAG (Novus Cat

# NBP1-06712, RRID:AB_1625981) at 1:200, all in 1% PBST. After microwaving, brains were

incubated for 4 hours at RT, and then moved to 4oC ON. Brains were washed 6 X 10 minutes

with 1% PBST and then incubated in secondary antibody solution containing goat anti-‐mouse

Alexa Fluor 647 at 1:500, goat anti-rabbit Alexa Fluor 488 (Thermo Fisher Scientific Cat# A-

11034, RRID:AB_2576217) at 1:500, and goat anti-rat Cy3 (Abcam Cat# ab6953,

RRID:AB_955010) at 1:500, all in 1% PBST, for 4 hours at RT before moving to 4oC ON.

Brains were then washed 9 x 10 minutes with 1% PBST, rinsed with PBS, and mounted as

described above.

In some cases, brain/VNC tissue was embedded in low-melting-point agarose and sectioned into

100-µm-thick cross-sections with a vibrating microtome.

For the FLP-‐out experiments, confocal images were collected on a Zeiss (Thornwood, NY) 510

LSCM. For MCFO experiments, confocal images were collected on a Zeiss 880 LSCM. Images were adjusted in Adobe Photoshop (San Jose, CA) using linear and/or gamma corrections to increase or decrease contrast.

Anatomical Analysis Analysis of the 51 isolated cells identified in the FLP-out gallery included cell-body position, arbor shape, volume of the cell, arbor territory (where it arborizes in the neuropil), and the

presence or absence of each category of astrocytic process extending outside the main arbor

territory. Overlap of arbor territories was examined using MCFO preparations (n=87

preparations in which 14 had sparse labeling but adjacent astrocytes).

To analyze cell-body position with respect to VNC segment, the segmentally repeating FasII-

positive transverse nerves were used to identify the boundaries between segments in the VNC

(Figure7A) (Landgraf, et al., 2003b). 3D projections rotated along the Z axis were used to

determine cell-body location (dorsal, lateral or ventral). In the case of the lateral astrocyte,

because the lateral astrocyte arises more dorsally and migrates (Peco, et al., 2014), we identified

a cell as a lateral astrocyte when its cell body location was more lateral than the dorsolateral

FasII tract.

Consistency of astrocyte cell-body location was determined from 3 preparations of the VNC cut

in cross-section in which GFP was targeted to the nuclei of astrocytes and the FasII-positive

tracks also were labeled immunocytochemically. Optical cross-sections through thoracic segment

3 and abdominal segment 1 were collected and the neuropil boundaries and the cell bodies of

each astrocyte were traced from each optical section using a WACOM drawing tablet. Segment

boundaries were determined by the FasII-positive transverse nerve position. An overlay of the

FasII tracts in cross-section provided a fiduciary matrix of the neuropil against which to chart the

location of each astrocyte. Reconstruct software (Harris, et al., 2015;

synapseweb.clm.utexas.edu/software-0) was used to generate surface-rendered images of the

neuropil boundaries and the position of each astrocyte cell body was added. Overlay of the cell

body-positions from all 3 preparations allowed determination of which positions were consistent and which were variable.

Cell volumes were computed using Amira v. 5.6 (http://www.amiravis.com) by using the

“LabelVoxel” command to create “material”, then the thresholding tool and brightness histogram

to tightly wrap this “material” to the fluorescent profile of the astrocyte, and finally the

“MaterialStatistics” command to find the volume of the resultant object. All parts of the cell, including the volume of protruding processes, were included.

The branching structure of astrocytes was analyzed using the Imaris ver 9.2.1 autopath

(marching cubes) algorithm, seeding from a central point in the cell body, but without the looping algorithm. We manually pruned the few obvious loops created in the renderings. 3D surface renderings, filament diagrams of branching structure, and animations also were created with Imaris.

To determine the degree to which astrocyte arbor territories were contained within a single

segment, the FasII-positive transverse nerve was used to identify the segment boundaries and the

FasII-‐tracts were used as fiducial points to create in each hemi-segment a box that enclosed most

of the neuropil (Figure 7). Viewed from the dorsal surface, the XY surface extended from

transverse nerve to transverse nerve and from the dorsolateral (DL) tract to the midline. In cross- section, the box extended from DL to midline and from the DL to the ventrolateral (VL) tract.

For each labeled cell, arbor volume within the box was determined and added to the within- segment volume of any part of the cell that extended into the region of same-segment neuropil from outside the box to the neuropil border. Neuropil borders were determined by faint background fluorescence. This value was subtracted from the total volume of the cell to give the volume of the cell outside of the segment that housed the cell body.

To analyze cell shape, confocal data were viewed in various LSM Viewer software display

modes, including 3D projections generated in both the XY and the XZ planes in order to fully

reveal the characteristics of the main arbor as well as presence or absence of common features

To measure the neuropil volume of each of the segments of the 3rd-instar VNC, the VNCs were

processed separately by gender (n=5 of each gender). The neuropil was labeled with anti-NC82,

which labels synapses. Because the antibodies for NC82 (DSHB Cat# nc82,

RRID:AB_2314866)and FasII were raised in the same species, so could not be used

simultaneously, we used “Looper” neuron tracts (period-positive median segmental interneurons

identified by Kohsaka, et al., 2014) as a proxy for the FasII-positive transverse nerve after having

determined that the Looper tracts, visible as GFP-positive tracts in w;UAS-CD4-tdGFP;looper-

GAL4 animals, essentially overlapped the FasII-positive transverse nerve (Figure 7A). Briefly, after fixation for 35 minutes in 4% PFA in 0.1M PB, the VNCs were rinsed, blocked and

permeabilized in 0.5% Triton/2% BSA in PB for 3 x 10 minutes, incubated in mouse anti-NC82

at 1:50 and rabbit anti-GFP 488 at 1:1000 in the dark for 5 nights at 4oC on a shaker, then rinsed,

incubated in goat anti-mouse Alexa 647 at 1:500 ON at 4oC on a shaker, rinsed, transferred to

50% glycerol, and finally mounted in 80% glycerol. Segment boundaries were positioned by

drawing a line across the looper tract and dropping a vertical plane to the ventral surface.

Segment neuropil volumes were computed using Amira v. 5.6 by using the “LabelVoxel” command to create “material”, then the thresholding tool and brightness histogram to tightly wrap this “material” to the fluorescent profile of the neuropil, and finally the “MaterialStatistics” command to find the volume of the resultant object.

Overlap of arbor territories was quantified using Imaris ver 8.4.1. We first identified adjacent

astrocytes that expressed different fluorophores (n=14 pairs) and isolated them from any nearby

(but not touching or overlapping) cells expressing the same fluorophores. The cells were reconstructed in 3 dimensions, using a threshold value that gave the best representation of the cell’s very fine distalmost branches without over-representing thicker processes. Voxels included, but not connected to any proximal branch, were deleted and the cell’s volume calculated. To measure the overlap between adjacent astrocytes, we did not use the Imaris co-

localization function as that function uses space overlap in two channels assuming strictly

overlapping voxels, but also takes all voxels in the Z stack, which was not useful. Instead we encapsulated the voxels in one channel by creating an iso-surface and masked all voxels in the second channel that did not fall within the volume enclosed by the first iso-surface and vice versa

(Figure 6). The volume of the resulting isolated overlap region was then calculated as a percent of the total volume of the reconstructed cells.

Statistics

Statistical analysis was carried out in SPSS (IBM SPSS Statistics for Windows, Version 26.0.

Armonk, NY: IBM Corp.) Comparison of neuropil volume across segments was run as a mixed

model analysis. Comparison of astrocyte volume across individual segments and by cell-body position, by thoracic vs abdominal segment and by shape were run as one-way ANOVAs followed by a post hoc Tukey HSD for multiple comparisons.

Results Orientation to the Drosophila ventral nerve cord

The Drosophila larval ventral nerve cord (VNC) is a segmented structure, with each VNC

segment corresponding to a segment of the body wall. The VNC has three thoracic and eight

abdominal segments, and its neuropil is divided at the midline into two continuous, tube-like neuropils that extend from T1 to A8/9 (Figure 1a). The organization of segments T1-A7 is rather

consistent, but differences arise in the terminal neuromeres (Campos-Ortéga & Hartenstein,

1997); for this reason, we have not included astrocytes found in the A8/9 segment in our analysis. The Drosophila VNC is organized in accordance with the ground plan of other invertebrate brains, wherein a distinct CNS compartment called “cortex” houses together the cell bodies of neurons and cortex glial cells as well as the processes of the cortex glial cells; this cortex is separated from the neuropil compartment that comprises only the processes of neurons and glial cells and contains synapses (Strausfeld, 1976; Kremer, et al., 2017). The 6 Drosophila astrocyte cell bodies that are associated with each hemi-segment of the VNC are located on the periphery of the synaptic neuropil at the interface with the cortex. From this position, astrocytic processes enter and interweave throughout the synaptic neuropil (Figure 1b). When the glia- specific repo-gal4 x UAS-IVS-mCD8:GFP reporter line is used to visualize the overall glial pattern in the 3rd-instar VNC, the thoracic segments, but not the abdominal segments, show clear

Figure 1. Ventral nerve cord (VNC) organizational plan. a View from the dorsal surface. Magenta, FasII-positive fiducial tracts run longitudinally through thoracic (T1-3) and abdominal (A1-8/9) segments; transverse nerves at the segment boundaries and segmental nerves extending laterally are also FasII-positive. b. Cross-section though T3. b. alrm-positive astrocytes (green) with cell bodies in dorsal, lateral and medial positions at the neuropil boundary and with processes infiltrating the neuropil. Cross- sectioned longitudinal FasII-positive fascicles (red, arrowheads; cell nuclei in blue (Syto); *, FasII-positive segmental nerve, the thickened region corresponding to the thoracic pocket area. Large axonal tracts in the neuropil are unlabeled, appearing as black regions. c. Schematic diagram, cross section of the VNC. Green stars: typical positions of astrocyte cell bodies. Magenta: FasII- positive tracts. d. Repo-positive, GFP-labeled glial cell bodies, likely exit glial cells, surround the base of the thoracic segmental nerves, forming pockets that are not present in the abdominal segments. e. 3D reconstruction of the neuropil of T3 and A1 segments showing the variability of position in astrocytes, particularly dorsal astrocytes. This figure shows not the number of astrocytes, but the positions at which astrocyte cell bodies were found in the antero-posterior and medio-lateral dimensions across 3 preparations; some positions were held in common by astrocytes in 2 or 3 preparations. The bilateral outpouchings in the ventral T3 segment are the pockets. Cross-section (e1) and tilted (e2) views from the rostral side; brackets indicate dorsal astrocytes. Scale bars: 50 µm in d1, and 25 µm in d2.

pockets comprising glial cells that surround the exit/entry point of the segmental nerve (Figure

1b,1d1-2). These cells are smaller than the astrocytes and have been identified by Enriquez, et al.

(2018) as the stem cells that give rise to the glia surrounding the adult thoracic neuropil. These pockets contain the segmental nerves and likely a few synapses as NC82 antibody revealed faint

labeling of this region. Other glial cells in the VNC are present, as would be expected with a

repo driver, but are only faintly labeled (Figure 1d2) as we optimized the labeling of the pocket- associated glial cells, which was very bright.

The boundaries between these segments and neuropil compartment boundaries within each hemi- segment can be charted by using Fasciclin II-positive axon tracts (FasII) as landmarks to demarcate known neuronal-processing compartments (Landgraf, et al., 2003b; Figure 1a,b,c). In

the dorsal-ventral axis, for example, the VNC neuropil is organized such that sensory inputs are processed within the ventral region of the neuropil and motor neuron dendrites and circuitry are

processed within the dorsal region (Schrader & Merritt, 2000; Rickert, et al., 2011). Peco, et al.

(2016) used this landmark system to map astrocyte domains in the 1st-instar larva.

The VNC as a whole, astrocytic and neuronal processes, and neuronal synapse number all

expand during larval growth (Hartenstein, et al., 2008; Kim, et al., 2009). Figure 2 shows the substantial increase in VNC size and in astrocyte size that occurs between the 1st and 3rd instars.

Figure 2. Growth of the VNC during larval development. a. 1st instar. b. 3rd instar. Dotted line: boundary of the VNC as viewed from the dorsal surface. Astrocytes, labelled by MCFO, show considerable arbor growth in size and complexity. Scale bars: 50 µm

Astrocyte characteristics We used two approaches to generate a collection of 3rd-instar larval VNC preparations with

sparse labeling of the astrocyte population (FLP-out) or with many astrocytes labeled but with

different colors (MCFO). Only those astrocytes that were labeled in isolation (i.e., astrocytes that

were not in contact any other surrounding, labeled cells in the same hemisphere) were selected

for detailed analysis of morphology (n=51).

Cell-body location Within a given hemi-segment of the VNC, as viewed from a dorsal-ventral perspective (Figure

1b), astrocyte cell bodies occupied characteristic, but not truly stereotyped positions surrounding

the neuropil: one cell body was located ventral to the neuropil, one lateral to the neuropil, and

four additional cells dorsal to the neuropil. These four dorsal cells were distributed along the

anterior-posterior axis. This general arrangement is in accordance with reports on 1st-instar VNC

(Stork, et al., 2014; Peco, et al., 2016). Figure 1e1-2 shows the neuropil volume of segments T3 and A1 as reconstructed from preparations labeled with the NC82 antibody to reveal synaptic neuropil. The positions of astrocyte cell bodies were plotted relative to the FasII tracts in 3

preparations in which astrocytes were labeled with alrm-targeted GFP. Each color of astrocyte represents a position occupied, some in more than one preparation. Although astrocyte cell bodies were consistently found in dorsal, lateral and ventral positions, there was considerable variation of position within those general regions.

Finally, 3 astrocytes had somas that were located along the midline between the two hemispheres of a segment, rather than fully ventrally (Table 2). The medially-located cell bodies observed in our dataset likely belong to the clone described by Peco, et al. (2016) that gives rise to a dorsal- medial astrocyte and the ventral astrocyte. Their less-ventral-than-usual position compared to the

Peco, et al. results in the 1st instar attests to considerable variance in the position of all the

astrocyte cell bodies at the 3rd instar, but also may reflect a failure of the more ventral cell of the

pair to fully migrate to the usual ventral position. We did not find any example of both a medial and a ventral cell in a single hemi-segment in either the FLP-out or MCFO preparations, but because these techniques generate a stochastic labeling pattern, we cannot definitively say that the presence of a medial astrocyte always means that there is no ventral cell, or vice versa.

Astrocyte morphology The FLP-out gallery of isolated GFP-labeled astrocytes was used for primary analysis of

astrocyte morphology. All astrocytes examined shared some common morphological features

(Figure 3) regardless of their location in thoracic or abdominal segments. The number of primary

processes was small (3-4). These processes rapidly became highly and densely branched upon

entering the neuropil. The gross shape of individual cells varied considerably, however.

Examples representing the range of this variation are shown in Figure 3 and quantified in Table

2; the morphologies shown are consistent with the range of morphology in our MCFO gallery.

Most astrocytes (49%) had globular arbors that were restricted to an approximately sphere-

shaped region (Figure 3a). This globular shape appeared in cells found in all three typical

regions. Some astrocytes took on more irregular shapes (43%) (Figure 3b). Cells with this

irregular shape likewise were found in all soma positions. In four cases (8%), the cells featured

two discrete arbor territories (Figure 3c), the second arbor entering a different functional

compartment as defined from the neuronal perspective (Rickert, et al., 2011, and see below).

This multi-arbor morphology was also observed in dye-filled dorsal astrocyte preparations in a previous study (MacNamee, et al., 2016, their Figure 1c), as well as in MCFO preparations (see

below), but across the three sample groups – MCFO, FLP-out and dye-filled cells - was never seen in the population of ventrally or medially located cells.

Figure 3. Astrocyte morphology. 3 basic morphological types, each 3D-reconstructed cell shown embedded in a frame with a1- c1 in X-Y orientation (from the dorsal surface) with FasII tracts (magenta). a2-c2 in X-Z orientation; axes in microns. a3-c3. Skeleton reconstruction for each cell in the rotation that best shows the morphological features that characterize the subtype. Cell bodies: teal circles. b. Fas II tracts: magenta. Arrowheads: transverse nerves at segment boundaries. Dotted line: midline. Videos showing rotation of 3D filament versions are included as supplemental material. We examined the branching pattern of 7 cells (Table 3), including cells located in both thoracic and abdominal segments, and including the 3 cells in Figure 3. Maximum branch order ranged

from 38 to 82, maximum dendritic length ranged from 13.1 to 21.22 µm, and maximum branch

angle ranged from 0.12o – 168.74o. These ranges are assumed not include details for the finest distal leaflets of astrocytes which, as noted earlier, are resolvable only at the EM level, so the values in Tables 2 and 3 should be considered only as reasonable estimates that likely underestimate both the absolute number of branches in the neuropil and the maximum total length and volume of the cells. The branching pattern seen across the cell types seems entirely consistent with expectation for cells with processes that insinuate into the spaces between

neuronal processes and interface with synapses.

Drosophila astrocytes often made glancing contact with FasII-positive tracts, simply by virtue of

their neuropil coverage, but most astrocytes also featured processes that traveled along and

branched around FasII-positive tracts, a feature also noted in the 1st instar (Peco, et al. 2016). The

extent of astrocyte process branching around a tract varied, and was not quantified in this study,

but at the electron-microscopy level, we never found tracts to be completely enclosed by

astrocyte processes (not shown). Figure 4a shows an example of the most extensive branching

around a tract observed. This type of loose enclosure was most often noted along the

dorsomedial tract, ventromedial, and centro-intermediate tracts, but did occur occasionally along

the dorso-lateral and ventro-lateral tracts. In a few cases, a single astrocyte sent processes along

more than one tract. Among the set of 51 cells in our gallery, only 10% of the astrocytes had no

processes associated with even a single axonal tract. In some cases, processes of an astrocyte

traveled outside of its home segment as defined by cell-body location, crossing into the adjacent

anterior or posterior segment (Figure 4b, 5d).

Variations in astrocyte morphology

While the gross morphology of astrocyte arbors ranged from a compact, globular form to a

patchy, irregular one, many astrocytes, even those with distinctly globular profiles, featured a

few additional processes that extended beyond the neuropil territory occupied by the main arbor.

These processes fell into three main categories, including: 1) processes that crossed the midline

to varying degrees; 2) protruding processes that were unbranched and traveled along the

boundary of the neuropil compartment; and 3) processes that exited the CNS along the path of

the segmental nerves.

Midline Crossing In total, 30 of 43 astrocytes (70%) exhibited some degree of midline crossing, with the midline

defined as a sagittal plane located between the left and right neuropil compartments. (Note: while

all of the single astrocytes in our dataset were isolated from neighboring cells within the

ipsilateral hemi-segment’s neuropil, there were 8 cells that were not fully separable at the midline from branches of labeled astrocyte(s) in the contralateral hemi-segment’s neuropil, thus making the degree of midline crossing difficult to determine; these cells were omitted from the midline-crossing analysis.) The degree of midline crossing ranged from extremely minimal, wherein an astrocytic process crossed the midline but did not arborize within the contralateral neuropil – essentially a stub (20%), to short (<10 µm) branched processes (57%), to considerable

contralateral arborization comprising 10-20% of the total cell volume and extending more than

10 µm into the contralateral neuropil (23%) (Figure 4c). The majority of the crossing processes

(77%) thus extended only short distances into the neuropil with relatively minor branching, if any. Astrocytes in dorsal/dorsomedial, ventral and medial cell-body positions had processes crossing the midline; no lateral cell did (Table 2). We expect that these midline crossing

processes do so at the level of the transverse commissures, although this is difficult to assess in our preparations because the transverse commissures are FasII-negative.

Figure 4. Additional morphological features of VNC astrocytes. a. Dorsal and orthogonal views of astrocyte processes associated with a Fas II-positive tract, here the dorsomedial tract. Dashed line: midline. b. Astrocyte arborizing primarily in lower segment may extend processes along a tract into an adjacent segment. Segment boundaries indicated by arrowheads in b-c. c. Some astrocytes extend processes across the midline. d and e. Ventral astrocytes extending a long, mostly unbranched process along the edge of the midline neuropil (arrowheads). f. A few astrocytes sent processes out into nerve roots for up to 100 µm. Scale bars: 10 µm.

Unbranched neuropil-boundary processes (Protruding processes)

Of the 51 cells examined, 20 (37.3%) featured a single, unbranched process that extended along

the neuropil-cortex boundary usually well beyond the main arborization territory (Figure 4d,e).

An unbranched neuropil boundary process was not observed with equal frequency across the

cell-body positions; it was observed most frequently in ventral cells (66.7%), sometimes in

lateral cells and medial cells (33.3% each), but never in dorsal cells. The process typically traveled along the medial edge of the neuropil. We saw no cases in which the protruding process extended into an adjacent segment.

Processes exiting the CNS In a few instances, we observed a single unbranched process that extended not only beyond the

main neuropil territory of that cell but left the CNS entirely (Figure 4e) and extended along the

segmental nerve for up to 100 µm, while the remaining processes of the cell entered the neuropil

and displayed characteristic astrocyte morphology. In two cases, these cells were located in

A8/9, and in one case the cell soma was located in A6. We also have noted in the MCFO gallery

2 additional instances of astrocytic cells with processes traveling outside of the CNS along the

segmental nerve.

In preparations with alrm-driven GFP-labeling of all astrocytes, however, faint labeled processes

were occasionally seen traveling out of the CNS along segmental nerves, mostly in abdominal

segments (not shown). The anatomical position of these astrocytic processes is very similar to

that described for a non-astrocytic glial cell that also derives from the longitudinal glioblast, the

“ensheathing/wrapping” glial cell (Peco, et al., 2016, their Figure 1G). But, because alrm-GAL4

is expressed transiently during development in non-astrocytic progeny of the longitudinal

glioblast and because we cannot assign the exiting process definitively to a particular astrocyte

soma in preparations with all astrocytes labeled, we cannot conclude that this is normal astrocyte

process behavior in all segments. Rather, we suggest that astrocytes may occasionally

outcompete and/or join the ensheathing/wrapping glial cell in associating with the segmental nerve. It is unclear why this might happen more often in the posterior abdominal segments.

Bilateral symmetry and between-segment symmetry of astrocyte branching In MCFO preparations (n = 87), we found 12 examples of astrocytes in adjacent hemi-segments that shared the same morphology, often even in small details such as a small unbranched process protruding in the same position or extending along the same neuropil boundary region for the same distance in their respective hemi-segment Figure 5a-c shows several examples. This pattern was not limited to cells appearing in the same segment, but also occasionally extended to astrocytes in adjacent segments in a single preparation. Figure 5d shows a preparation in which astrocytes in 3 segments extended an anterior-going process along the same FasII-positive dorsomedial tract. Given the stochastic nature of MCFO labeling, we cannot specify the true frequency at which cells on each side of a hemi-segment share morphology, nor do we know their lineage relationship.

The interface between astrocytes within and between segments: Tiling Peco, et al. (2016) in the 1st-instar and Stork, et al. (2014) in the 3rd-instar VNC reported that astrocytes tile the neuropil. Here we studied the details of the astrocyte tiling pattern, as the nature of the interface will determine how synaptic territory is divided, or shared, by adjacent astrocytes. Preparations such as those in Figures 2b, 5d and 6a,b show the tiling pattern at the 3rd instar.

Figure 5. Bilateral and between-segment morphological symmetry. a1 and a3, b1 and b3, and c show reconstructions of pairs of cells that share structural similarity. Panels a2 and a4 and b2 and b4 also show adjacent pairs of cells with major branches extending deep into each other’s territory. d. Astrocytes in adjacent segments extending similar processes rostrally to wrap the same longitudinal tract, dotted line. MCFO preparation. Scale bar in d: 50 µm.

To quantify the overlap, seven pairs of ventral astrocytes, 3 lateral-ventral pairs, and 1 lateral- lateral pair, all from the MCFO data set, were reconstructed in 3 dimensions in Imaris. Because the finest branches are vellate and often difficult to resolve, we set the threshold to optimize these distal-most processes. The range of overlap in territory between ventral cells was 0.5 to

5.1%, and between lateral and ventral cells from 1.2 to 3.0%. The single lateral-lateral pair (in adjacent segments) showed a 2.5% overlap. These values are within the range of overlap reported between adjacent vertebrate astrocytes (Bushong, et al., 2002).

Although the degree of overlap between adjacent astrocytes is clearly small, the characteristics of the interface between the branches of adjacent astrocytes will be a key determinant of whether a given synapse reciprocally interacts with one or two astrocytes. Figure 6a shows 3 cells, one dorsal (red), one lateral (blue), and one ventral (yellow) astrocyte in the same hemi-segment.

Branches of each cell appear to extend only short distances into the neuropil territory generally subtended by the adjacent cell. Similarly, Figure 6b shows 2 multi-arbor cells, a dorsal cell

(yellow) and a ventral cell (cyan) in which substantial branches form discrete domains within the territory of the adjacent cell.

Figure 6. Minimal overlap between adjacent astrocytes. a, a dorsal, a lateral, and a ventral astrocyte. b. a dorsal and a ventral astrocyte. Cell bodies: circles. Midline dotted. c1. Interface between the ventral (yellow, c3) and lateral (magenta, c2) astrocyte is a large dorsally projecting branch of the ventral cell that protrudes into the territory of the lateral cell (arrowheads) with minimal intermingling of fine branches. d. Method used to determine overlap. d1. Schematic of two adjacent astrocytes in different channels. d2-d3 Cell in each channel is isolated using a bounding box, reconstructed in 3D to create an iso-surface, and volume determined. d4 Processes in region of overlap isolated via masking process described in methods section. Volume of processes in overlap region determined. Percent overlap calculated (volume of overlapping processes divided by total volume of both cells). Scale bars in a: 10 µm; in c1-c3: 10 µm.

We examined the interface regions of our adjacent astrocytes at high-resolution and high magnification. Figure 6 c1-c3 shows 2 adjacent ventral astrocytes, reconstructed in 3D, in which the yellow cell at the bottom extends a branched process deep into the territory of the magenta cell above, clearly invading the general perimeter of the upper astrocyte. This gerrymander

pattern was also seen in the astrocytes in Figure 5a. Except at the edges of the invading process,

there is no fine-scale overlap of territory in which branches of the two astrocytes intermingle.

Segment specificity of 3rd-instar VNC astrocytes Using the FasII-positive transverse nerve as a segment boundary marker (Figure 7a, magenta),

we asked if astrocyte territories extend between segments, as do the axon terminal branches of

certain sensory neurons (e.g., lateral chordotonal neurons, and bipolar dendrite neurons

(Schrader & Merritt, 2000) and the dendrites of certain motor neurons (Choi, et al., 2004). For

each 3D-reconstructed astrocyte, the total volume of the arbor in the box and in the adjacent

within-segment neuropil (see Methods) was determined, and the percent of the arbor volume

outside of the segment calculated. To allow for the difficulty of defining volumetric segment

boundaries precisely, cells with less than 15% of their volume extending into an adjacent

segment were classified as arborizing within a single segment.

Figures 7b,c show the range we observed, from arborization totally within segment to

arborization shared almost equally between adjacent segments. Overall, only about 30% of the

Figure 7. Astrocytes can branch in more than one segment. a. Dorsal view of a VNC in which looper neurons expressed GFP (yellow) and transverse nerves were labeled with anti-FasII (magenta). Except in T1 and T2, which have a mid-segment FasII- positive nerve bundle, the transverse nerve and the laterally projecting looper axon tract essentially overlap. b. Astrocyte arborizes totally within a single segment. c. Astrocyte arborization shared nearly equally between segments. Boxes drawn to segment boundaries using FasII transverse nerve as a marker; dotted edge is midline. Scale bar in a: 50 µm.

astrocytes had arbors restricted to a single segment. Nearly all thoracic astrocytes arborized

extensively in adjacent segments, (15-48%), no matter their cell-body position. Whether located in thoracic or abdominal segments, lateral astrocytes in particular nearly always had more than

15% of their arbors in adjacent segments. Though an n of 1, one dorsal cell fully arborized in a segment adjacent to the segment that housed its cell body, indicating that there is not an inviolate rule that constrains astrocytes to arborize in the segment neuropil with which the cell body is associated. Table 4 shows the results for the 48 astrocytes examined, classified by their cell-body position. Medial cells were not included in this analysis.

Neuropil territories of individual astrocytes in 3rd-instar larval VNC relative to neuronal domains

During development, the six cells that become astrocytes in each hemi-segment of the VNC (Ito, et al., 1995; Pereanu, et al., 2005; Stork, et al., 2014; Peco, et al., 2016) derive from a final, symmetrical division of three progenitor cells located dorsal to the neuropil, occurring by embryonic stage 16. Then, two cells migrate laterally, with one reaching the lateral-most position relative to the neuropil and the other remaining dorsolateral. Finally, one cell migrates along the medial neuropil boundary before arriving in a position ventral to the neuropil (Peco, et al., 2016).

Astrocyte processes begin to invade the neuropil in the last hours of embryogenesis (~19 hrs

AEL).

By examining 2-cell clones labeled using a technique similar to FLP-out, Peco, et al. (2016) reported stereotypical parsing of the early 1st-instar VNC neuropil into three territories/domains,

astrocytic territories that roughly map onto the neuronal functional domains defined over the

years by dye fills of sensory and motor neurons and more recently by expression of fluorophores

in targeted subpopulations of sensory and motor populations (Merritt & Whittington, 1995;

Landgraf, et al., 2003a; Zlatic, et al., 2003; Grueber, et al., 2007). We asked whether this

predictable parsing of neuropil space by astrocytes is preserved at the 3rd instar. For each isolated astrocyte in our dataset, we charted domain territory by identifying the collection of major longitudinal FasII tracts shown in Figure 1a,b (ventromedial, VM; ventrolateral, VL; central intermediate, CI; dorsomedial, DM; dorsolateral, DL) contacted by a given astrocyte. This method also allowed us to consider the degree of overlap between a given astrocyte’s domain and the neuronal functional domains (Zlatic, et al., 2009; Burgos, et al. 2018). It is important to note we never found an arbor that could not be traced to a cell body, so the variation in arbor position cannot be explained as a simply a case of labeling of multiple cells. Comparison of the relationship of astrocyte territories to the FasII tracts (Figure 8 graphs) indicates that the territories typically occupied by the lateral and medial cells varied little; those of the ventral and dorsal cells had greater variability with only about 50% of the cells arborizing in the same tract- defined territory.

We also marked the territory of each individual astrocyte by drawing a perimeter at the tips of each cell’s distalmost branches. Figures 8a,b summarize the neuropil territories in the thoracic and abdominal segments typically occupied by astrocytes in each of the cell body positions and

show considerable overlap of the territories that could be occupied by the processes of astrocytes

whose cell bodies lie in the 3 cell-body positions – dorsal, lateral and ventral. Figures 8d-h show, for the thoracic and abdominal segments respectively, overlays of the tracing for the territory of each individual cell in each cell-body position (indicated by the stars). The transparent overlay in each panel represents the territory common to all of the cells in a given position.

Figure 8. Astrocyte territories and neuronal functional domains. a,b. Summary of astrocyte territories for each cell-body position. Colored overlay in each shows the territory held in common for each cell-body position in the thoracic and abdominal segments. c. Neuronal domains (modified from Zlatic et al., 2009 and Burgos et al., 2018). d-h. Overlaid tracings of the perimeter of each cell’s territory within each cell-body position (indicated by stars). Colored transparent overlay in each panel indicates the territory held in common by all of the astrocytes in a given cell-body position. d,f. Thoracic segments. e,g. Abdominal segments. h. Combined thoracic and abdominal tracing for lateral astrocytes. Graphs. For each cell-body position, relationship of astrocyte territories to FasII-positive longitudinal tracts. The territories of the 4 dorsally-located astrocytes usually arborized across the dorsal neuropil in accord with the medial to lateral position of their cell bodies; at least 2 most medially-located dorsal cells extended arbors out of the dorsal motor neuropil into the region of the proprioceptive

neuropil. 10% branched broadly across the dorsal region, and one astrocyte, with a dorsolateral

cell body position, branched in the territory typically occupied by the lateral astrocytes; its

territory suggests that it may correspond with the second lateral cell described in the 1st instar

(Peco, et al., 2016). The range of territories of ventrally-located astrocytes were unexpectedly

variable, given that a typical hemi- segment contains only a single ventrally-located astrocyte.

Most (5 of 6) astrocytes with laterally-located cell bodies had processes that were restricted to the lateral region of the neuropil. The astrocytes with medially-located cell bodies all branched in the same region, but the number of these cells examined was small (n = 3 cells).

Interestingly, compared to the 1st instar (Peco, et al., 2016), astrocyte territories in the 3rd instar

were quite variable, enough so that a standardized scheme of neuropil division among the six

astrocytes that populate each hemi-segment is not possible, except for the medial astrocytes.

Because the cells used for this analysis had to be isolated cells, it was not possible to determine

whether the territories together occupied by the dorso-medial cell and the ventral cell were

consistent even in the face of greater variability at the single-cell level. The range of territories

we observed to be occupied by individual astrocytes, however, argues strongly against true

stereotypy in astrocyte neuropil territory/domain.

Figure 8c shows the known neuronal functional domains, modified from Zlatic, et al., 2009 and

Burgos, et al. (2018). Given the broad and variable arborization pattern of astrocytes in each cell-

body position (Figure 8a,b), especially in multi-arbor astrocytes (Figures 3c, 6b (yellow cell)),

astrocyte arbors clearly can cross into more than one neuronal domain, potentially affecting activity across those functional domains, as was also seen in the adult Drosophila CNS (Kremer, et al., 2017).

Astrocyte volume relative to segment volume and cell-body position relative to the neuropil.

Finally, we asked whether astrocyte volume depends on the VNC segment in which that

astrocyte resides and/or on the position of the cell. Because the shape and size of the VNC

changes from rostral to caudal segments, we asked whether there is a proportional change in both

astrocyte and neuropil volume.

To determine the volume of each segment, GFP-labeled looper tracts in looper/NC-82 labeled preparations were used to define the anterior and posterior boundaries of a segment because antibody mis-match prevented simultaneous use of anti-FasII and anti-NC-82, a marker of synaptic neuropil. Looper tracts are bundles of period-positive axons belonging to a set of premotor, glutamatergic neurons with cell bodies clustered in the ventral cell body cortex at the midline of the VNC (Kohsaka, et al., 2014; Clark, et al., 2018). The looper neurons have a large axon that travels dorsally along the midline then branches and crosses near the dorsal surface of the neuropil toward the lateral edge. Because the path of the looper axon tract closely follows the

FasII-positive transverse nerve (Figure 7a) in all except segments T1 and T2, which have a thin

FasII tract crossing medial to lateral in roughly in the middle of the segment, we were able to use the GFP-positive looper tract to provide a visual marker of the rostral and caudal segment boundaries.

The remaining borders of the synaptic neuropil in each segment (dorsoventral and mediolateral) were defined by anti-NC82 labeling in each of the optical sections taken through the depth of the neuropil (Figure 9a; n=10), then reconstructed in 3 dimensions using Amira. Pairwise comparison of the volume of 3 thoracic and 3 abdominal segments revealed no significant volume differences except T1 was greater than A3 (p<0.006).

The cells in the FLP-out dataset were reconstructed in 3D using Amira software and their cellular volumes determined (note: the volume measurement is of the membrane-bound cell

Figure 9. Astrocyte and neuropil volume. a. Looper axon tracts used as boundary markers. Synaptic neuropil labeled with NC82 (not shown) and reconstructed in 3D. Upper three reconstructions are thoracic segments; lower three are for abdominal segments A3, 5 and 6. b. Neuropil volume/segment. c. Volume of astrocytes sorted by cell-body position, by segment, and by morphology. Each dot is an individual cell, with a total of 51 cells. d. Astrocyte volume by segment. Error bars: SEM.

volume, not the domain volume of an astrocyte). The volume of each astrocyte (n=51) is shown in Figure 9c in accord with cell-body position, with thoracic vs abdominal segment location, and with morphological classification. The mean volume of all Drosophila astrocytes examined was

6,834.2 µm3 + 3,468.9 (S.D., n = 51 cells), showing considerable variation across position and morphology (see also Table 2). Figure 9d shows astrocyte volume by segment. Overall, there were few significant differences: None related to cell-body position (dorsal, ventral, lateral, medial); irregular astrocytes had greater volume than multi-arbor astrocytes (p<0.036); and astrocytes in T3 had greater volumes than those in A4 (p<.001).

Discussion Here we have provided a detailed analysis of astrocyte morphology and territory in the

Drosophila 3rd-instar VNC, in which there are only six astrocytes/hemi-segment, two of them

uniquely identifiable by position. These astrocytes share morphological and molecular similarity

to vertebrate protoplasmic astrocytes (Ogata & Kosaka, 2002; Bushong, et al., 2002; Peco, et al.,

2016) and, to the extent it has been examined, functional similarity at synapses (MacNamee, et

al., 2016). Because it is now absolutely clear that understanding neuronal function at the local

and circuit level requires also understanding astrocyte function, these structural data provide a

necessary foundation for genetic and functional studies designed to elucidate neuron-glia interactions in this neuropil.

Astrocyte morphology Astrocytes in the VNC neuropil have most commonly been described as globular in shape

(Stork, et al., 2014; Peco, et al., 2016); indeed, we found about 50% of them to have the classic globular shape. The range of morphologies is much broader, however, including cells that have two or more discrete arbors (multi-arbor cells) and others that have rather irregular and wide- ranging arbors. These different morphologies are somewhat dependent on cell-body position within a segment, i.e. dorsal, lateral, ventral or medial. Nearly three-quarters of the ventral astrocytes are globular and none show a multi-arbor phenotype, lateral astrocytes tend to be globular, and dorsal astrocytes are mostly irregular in shape. These cell-body-position-specific shape patterns would be consistent with a competition model for defining territory during development. The processes of the 4 dorso/dorsolaterally-located astrocytes together insinuate among the neuronal processes of the motor neuropil, presumably competing strongly with each other for territory in the dorsal motor domain, while the lateral astrocytes would have

competition only with the most lateral dorsal astrocyte and with the ventral astrocyte, and the

ventral astrocyte mainly with the lateral astrocyte. In addition, the pattern of branching between two adjacent astrocytes, see below, also suggests competition in the form of inhibitory/repulsive cues. .

Yet, the wide variation in shape shown in this study suggests competition is not sufficient to rule out a contribution from other factors, including cell autonomous/genetic factors, use of

developmentally pre-existing structures, and perhaps timing of outgrowth. The first factor – cell

autonomy - is suggested by examples of astrocytes situated in adjacent hemispheres whose

arbors are bilaterally symmetrical, not just at the level of the large branches as has been reported

in detailed comparisons of the shapes of specific motor neurons (Vonhoff & Duch, 2010), but in

the case of these Drosophila astrocytes, down to higher branching orders. On the other hand, the

similarity of astrocyte shapes in adjacent segments, as shown in Figure 5d, could reflect use of

pre-existing structures, with ease of growth along, if not attraction to, structures such as axon

tracts. Processes growing in a particular vicinity of the neuropil during the same developmental

window are likely to encounter similar features, such as axonal tracts that were established

earlier in development, resulting in similar morphologies. Timing of outgrowth alone is likely to

have a minor effect, however, as the late-embryonic timing (~19 hrs AEL) of the initial invasion

of astrocyte processes seems to be similar across the neuropil (Stork, et al., 2014), requiring both

FGF signaling (Stork, et al., 2014) and expression of the transmembrane leucine-rich repeat

protein Lapsyn (Richier, et al., 2017) for extension of processes into the neuropil. Further, Stork,

et al. reported that disruption of FGF signaling in a subset of astrocytes or removal of some

astrocytes via stochastic apoptosis induction caused the remaining astrocytes to expand their

territories. Taken together, the variation in astrocyte cell shape shown here, though considerably

greater than previously reported, is consistent with a modification of the model originally

proposed by Stork, et al. (2014) in which the developmental growth process is dominated, but

not fully explained, by competitive or repulsive interactions between astrocytes. Our data suggest

that that model also should include cell-autonomous aspects affecting shape, as well as use of

structural features for growth.

In addition to overall shape, we report several common morphological features - midline crossing, unbranched neuropil processes, and loose association of astrocyte processes with axon tracts, some of which could have functional consequences.

Midline crossing was a common feature, occurring in 70% of dorsal and ventral astrocytes.

When midline crossing did occur, in nearly 80% of those cases the degree of crossing likely was too small to have a large effect on either neuronal or glial activity in the contralateral neuropil.

For the remaining ~20%, extensive branching deep into the contralateral neuropil could be

functionally significant in affecting neuronal activity or in coordinating astrocyte activity across

the midline, particularly in the later situation if there is gap-junctional communication between the midline-crossing process and astrocytes in the adjacent hemi-segment.. Approximately a third of the astrocytes examined electrophysiologically in a previous study were shown by dye- coupled to 1 to 4 other astrocytes (MacNamee, et al., 2016), suggesting that signals generated by neuronal activity in individual astrocytes could spread fairly widely in the neuropil. All the dye-

coupled cells in that report resided in the same hemi-segment and there were no instances of dye-

coupling to astrocytes across the midline, but the number of astrocytes recorded may not have

been sufficient to reveal infrequent instances of cross-midline dye coupling. Nor do we yet

understand whether the incidence of coupling is physiologically regulated, leaving open the

possibility that under some circumstances cross-midline coupling could be present.

Unbranched processes growing along the neuropil border could represent a developmental

failure or a remnant of migration, but the occasional symmetry of astrocyte across hemi- segments, each bearing a similar process of similar length, argues that this feature also could be a cell-autonomous one, function unknown.

As noted earlier, we found in our 3rd-instar Drosophila preparations that astrocytes often had

processes that extended well beyond their main arbor and even into an adjacent ipsilateral

segment, with processes traveling mainly along FasII-positive axon tracts. These tracts are established early in development (Hildago & Brand, 1997; Landgraf, et al. 2003b) and likely serve as a ready substrate for the growth of astrocyte processes. The astrocyte processes apparently do not invade the tracts to branch among the axons as do the processes of vertebrate fibrous astrocytes (Miller & Raff, 1984; Halassa, et al., 2007) or Schwann cells (Peters, et al.,

1976), nor did the processes completely ensheathe these tracts. While we did not have examples in our MCFO series that would allow us to determine whether 2 or more astrocytes could collaborate to fully ensheathe a segment of the tract, we never saw such ensheathment in the electron microscope, even in serially sectioned material examined in the context of our earlier study (MacNamee, et al., 2016). While the association between Drosophila astrocytes and the tracts with which they associate thus appears far looser than that between vertebrate astrocytes and axon tracts, the frequency of association suggests the possibility that these astrocytes may be functionally capable to signal reciprocally with axons and thus affect neuronal circuit function.

Astrocyte territory The territorial organization of astrocytes that comprise a glial network makes these cells well- suited to play a role in the spatial segregation or spread of information: in many of the species and brain regions examined to date, astrocytes have been found to occupy unique spatial

domains with minimal overlap (macaque: Distler, et al., 1993; mouse: Ogata & Kosaka, 2002;

rat: Bushong, et al., 2002, 2004; ferret: visual system (Lopez-Hildago & Schummers, 2014);

Drosophila: Stork, et al., 2014; Peco, et al., 2016). In mammalian sensory cortices and olfactory

bulb, there is experimental evidence that individual astrocyte activity maps align with neuronal

sensory field maps (Roux, et al., 2011; Lopez-Hildago & Schummers, 2014). Astrocytes are now

understood to regulate and refine the neuronal connectome in various ways (Fields, et al., 2015),

but there has been relatively little detailed study of the functional consequences of this

organizational pattern across species or brain regions.

We found first that territory overlap between adjacent 3rd-instar astrocytes in Drosophila VNC is minimal, at roughly 5%, consistent with findings in both vertebrate (Bushong, et al., 2002) and

Drosophila as reported previously (Stork, et al, 2014). Interestingly though, the overlap was not a simple interdigitation of distal processes within overlapping neuropil volumes of adjacent astrocytes. That arrangement would imply that synapses in that shared volume could be regulated by both astrocytes. Instead, if one envisions a sheath formed around an astrocyte at the tips of its distal branches, adjacent astrocytes would be found to create small peninsulas of branches protruding into the territory of the first. The second astrocyte thus would regulate any synapses that otherwise would be under control of the first. Gap-junctional communication between the

astrocytes, which could only occur in the small border where adjacent astrocyte territories meet,

could, of course, influence this dynamic. In our previous study in this system, about a third of the

astrocytes from which we recorded were found to be dye-coupled to one of more other astrocytes

(MacNamee, et al., 2016). Influence on neuronal function depends then on whether there is a

functional network between astrocytes via gap junctions vs an astrocyte-specific neuron-glia interaction (Giaume & Liu, 2012) and remains to be studied in the VNC.

Organization of the glial territories relative to neuronal domains By examining two-cell astrocyte clones in the early 1st-instar larvae, Peco, et al. (2016) revealed

stereotypical parsing of the VNC neuropil into three astrocyte territories. Here we used the FasII

landmarks that have been used to map neuronal morphologies to delineate the range of territories

occupied by astrocytes in each cell-body position (Figure 8). These maps of astrocyte territories

showed the typical neuropil territory occupied by 3rd-instar astrocytes in each cell-body position to be more variable than had been seen at the 1st instar, perhaps not surprising given the

extensive growth that occurs between these stages.

The range of territories we observed in these maps to be occupied by individual astrocytes argues

strongly against true stereotypy in astrocyte neuropil territory/domain. The variation was, in fact,

great enough that a standardized scheme of neuropil division among the six astrocytes that

populate each hemi-segment is not possible at the 3rd instar. Second, our finding that lateral

astrocytes can extend into the dorsal and/or ventral territory and that ventral astrocytes, though

typically arborizing in the sensory domain, can extend well into the intermediate region strongly

suggests that both populations – lateral and ventral astrocytes - could influence neurons across

neuronal functional domains.

Neuronal functional domains in the VNC neuropil have been well defined based on data gathered

from dye fills of sensory neurons, from cell-type-specific expression of fluorescent reporters to establish the characteristic domains of sensory axon terminals, motor neuron dendrites and interneuronal axons and dendrites, from high-resolution reconstruction of the VNC neuronal connectome from 3D electron-microscopic data sets (Merritt & Whitington, 1995; Landgraf, et al., 2003a; Zlatic, et al., 2003; Grueber, et al., 2007; Kim, et al., 2009; Rickert, et al., 2011;

Ohyama, et al., 2015; Schneider-Mizell, et al., 2016; Clark, et al., 2018), and from physiological

studies using activity monitors and activity drivers and silencers (Tessier & Broadie, 2009;

Kohsaka, et al., 2014, 2017). Those data show the dorsal neuropil to be dominated by

motoneuron processes, the ventral neuropil by sensory neuron processes, and the intermediate

region by proprioceptive and interneuron processes.

In the dorsal region, there is a myotopic map (Kim, et al., 2009) that appears during embryonic

development (Landgraf, et al., 2003), but there is no obvious subdivision of the region among the four dorsal astrocytes, the arbor of each extending widely across the dorsal neuropil and suggesting that the dorsal astrocytes do not respond only to activity in motor neurons driving a single muscle group. Within the ventral sensory region and in the mid-neuropil proprioceptive region, class-specific axon terminals are organized into subregions and often somatopically

(Pfluger, et al., 1988; Murphey, et al. 1989; Schrader & Merritt, 2000; Grueber, et al., 2007;

Ohyama, et al., 2015). The ventral astrocyte, whose territory can extend well into the

proprioceptive region, has no obvious branching pattern that correlates with these subregions,

suggesting that the astrocyte would respond to neuronal activity in any of the sensory modalities, possibly integrating across them.

Whether the physical presence of processes belonging to a single astrocyte and crossing neuronal domains has physiologically relevant consequences depends on whether the different regions of the astrocyte arbor are functionally discrete or well-connected from a signal-transmission perspective. Despite some elegant high-resolution calcium imaging and modeling studies of vertebrate (Shigetomi, et al., 2013; Kanemaru, et al., 2014; Asada, et al., 2015; Rungta, et al.,

2016; Bazargani & Attwell, 2016; Stobart, et al., 2018; Gordleeva, et al., 2019; Semyanov, 2019) and invertebrate (Zhang, et al., 2017) astrocytes responding to local neuronal activity, however, we still lack detailed understanding of calcium signal spread within astrocytic processes under

physiological conditions, particularly in the often vanishingly thin distal processes or leaflets that

are associated with synapses. Though their number was relatively small, the multi-arbor subtype may provide an interesting subject for determining whether calcium signals generated in response to activity in neurons of one domain remain local to or spread across the connecting process into the arbor that resides in a separate neuronal domain, potentially affecting neuronal activity in that second domain.

Astrocyte territories that cross into adjacent VNC segments In the VNC, the arbors of certain sensory and motor neurons extend beyond a single segment

(Kim et al. 2009). Interestingly, we found that a large percentage of astrocytes in the thoracic

segments have a significant portion of their arbor (15%-48% of their arbor) in an adjacent segment. Lateral astrocytes consistently (>80%) have such arbors. Because lateral astrocyte arbors can extend into all three neuronal domains and because they also can extend between segments, they could widely integrate and influence neuronal activity. Neuronal domains appear to form early in development (Merritt and Whitington, 1995; Schrader and Merritt, 2000,

Landgraf et al., 2003a), before the astrocytes begin to extend their processes into the neuropil, but the lack of adherence of arbors to specific neuronal domains suggests that astrocyte territories in the larval VNC are not driven solely by coordination with neuronal development, a

coordination that has been shown to occur between motor neurons and the neuropil glia

branching in the motor territory of the developing adult neuropil (Enriquez, et al., 2018). The

functionality of multi-segment astrocyte territories again will depend on the rate and extent of

signal spread through the full astrocytic arbor.

Summary

Combining this detailed analysis of astrocyte morphology and territory in the Drosophila 3rd- instar VNC with what we already understand structurally and functionally about the 3rd-instar neuronal networks sets the stage for future studies that will generate a robust, functionally- oriented connectome that includes both partners in neuronal circuits – the neurons and the glial cells. With only 6 astrocytes in a hemi-segment, two of them individually identifiable, and with clear evidence that astrocytes in this system modulate synaptic activity (MacNamee, et al.,

2016), the VNC astrocytes offer an unparalleled opportunity to explore in detail the functional relationships between astrocytic and neuronal networks in this experimentally powerful model system.

References Allen, N. J., Bennett, M. L., Foo, L. C., Wang, G. X., Chakraborty, C., Smith, S. J., & Barres, B.

A. (2012) Astrocyte glicans 4 and 6 promote formation of excitatory synapses via GluA1

AMPA receptors. Nature 486, 410-14. PMID 22722203, doi: 10.1038/nature11059.

Asada, A., Ujita, S., Nakayama, R., Oba, S., Ishii, S., Matsuki, N., & Ikegaya, Y. (2015) Subtle

modulation of ongoing calcium dynamics in astrocytic microdomains by sensory inputs.

Physiol Rep 3, e12454. PMID 26432942, doi: 10.14814/phy2.12454

Bazargani, N. & Attwell, D. (2016) Astrocyte calcium signaling: The third wave. Nat Neurosci

19, 182-9. PMID 26814587, doi: 10.1038/nn.4201

Boyan, G. S. & Ball, E. E. (1993) The grasshopper, Drosophila and neuronal homology

(advantages of the insect nervous system for the neuroscientist). Prog Neurobiol 41, 657-

682. PMID 8140256

Burgos, A., Honjo, K., Ohyama, T., Qian, C. S., Shin, G. J., Gohl, D. M., Silies, M., Tracey,

W.D., Zlatic, M., Cardona, A., & Grueber, W.B. (2018) Nociceptive interneurons control

modular motor pathways to promote escape behavior in Drosophila. eLife 7, e26016.

PMID 29528286, doi 10.7554/eLife.26016

Bushong, E. A., Martone, M. E., & Ellisman, M. H. (2004) Maturation of astrocyte morphology

and the establishment of astrocyte domains during postnatal hippocampal development.

Int J Devl Neurosci 22, 73-86. PMID 15036382, doi: 10.1016/j.ijdevneu.2003.12.008

Bushong, E. A., Martone, M. E., Jones, Y. Z., & Ellisman, M. H. (2002) Protoplasmic astrocytes

in CA1 stratum radiatum occupy separate anatomical domains. J Neurosci 22, 183-192.

PMID 11756501

Campos-Ortéga & J. A., Hartenstein, V. (1997) The embryonic development of Drosophila

melanogaster. Berlin: Springer.

Choi, J. C., Park, D., & Griffith, L. C. (2004) Electrophysiological and morphological

characterization of identified motor neurons in the Drosophila third instar larva central

nervous system. J Neurophysiol 91, 2353-2365. PMID 14695352, doi:

10.1152/jn.01115.2003

Chung, W.-S., Clarke, L. E., Wang, G. X., Stafford, B. K., Sher, A., Chakraborty, C., Houng, J.,

Foo, L. C., Thompson, A., Chen, C., Smith, S. J., & Barres, B. A. (2013) Astrocytes

mediate synapse elimination through MEGF10 and MERTK pathways. Nature 504, 394-

400. PMID 24270812, doi: 10.1038/nature12776

Clark, M. Q., McCumsey, S. J., Lopez-Darwin, S., Heckscher, E. S., & Doe, C. Q. (2016)

Functional genetic screen to identify interneurons governing behaviorally distinct aspects

of Drosophila larval motor programs. G3(Bethesda) 6, 2023-2031. doi:

10.1534/g3.116.028472, PMID 27172197

Clark, M. Q., Zarin, A. A., Carreira-Rosario, A., & Doe, C. Q. (2018) Neural circuits driving

larval locomotion in Drosophila. Neural Dev 13, 6. doi: 10.1186/s13064-018-0103-z,

PMID: 29673388

Clarke, L. E. & Barres, B. A. (2013) Emerging roles of astrocytes in neural circuit development.

Nat Rev Neurosci 14, 311-21. PMID 23595014, doi: doi: 10.1038/nrn3484

Corty, M. M. & Freeman MR (2013) Architects in neural circuit design: Glia control neuron

numbers and connectivity J Cell Biol 203, 395-405. PMID 24217617, doi:

10.1083/jcb.201306099

Distler, C., Weigel, H., & Hoffman, K. P. (1993) Glia cells of the monkey retina. I. Astrocytes. J

Comp Neurol 333, 134-147. PMID 8340493, doi: 10.1002/cne.903330111

Doherty, J., Logan, M. A., Taşdemir, O. E., & Freeman, M. R. (2009) Ensheathing glia function

as phagocytes in the adult Drosophila brain. J Neurosci 29, 4768-4781. PMID 19369546,

doi: 10.1523/JNEUROSCI.5951-08.2009

Enriquez, J., Rio, L.Q., Blazeski, R., Bellemin, S., Godemont, P., Mason, C., and Mann, R.S.

(2018) Differing strategies despite shared lineages of motor neurons and glia to achieve

robust development of an adult neuropil in Drosophila. Neuron 97, 538-554.e5. PMID

29395908, doi: 10.1016/j.neuron.2018.01.007

Eroglu, C. & Barres, B. A. (2010) Regulation of synaptic connectivity by glia. Nature 468, 223-

231. PMID 21068831, doi: 10.1038/nature09612

Eroglu, C., Allen, N. J., Susman, M. W., O’Rourke, N. A., Park, C. Y., Ozkan, E., Chakraborty,

C., Mulinyawe, S. B., Annis, D. S., Huberman, A. D., Green, E. M., Lawler, J.,

Dolmetsch, R., Garcia, K. C., Smith, S. J., Luo, Z. D., Rosenthal, A., Mosher, D. F. &

Barres, B. A. (2009) Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin

receptor responsible for excitatory CNS synaptogenesis. Cell 139, 380-92. PMID

19818485, doi: 10.1016/j.cell.2009.09.025.

Fields, R. D., Woo, D. H. & Basser, P. J. (2015) Glial regulation of the neuronal connectome

through local and long-distant communication. Neuron 86, 374-386. PMID 25905811,

doi: 10.1016/j.neuron.2015.01.014

Fishiki, A., Zwart, M. F., Kohsaka, H., Fetter, R. D., Cardona, A. & Nose, A. (2016) A circuit

mechanism of the propagation of waves of muscle contraction in Drosophila. eLife 5,

e13253. PMID 26880545, doi:10.7554/eLife.13253.001

Freeman, M. R. (2015) Drosophila central nervous system glia. CSH Perspect Biol. PMID:

25722465, doi:10.1101/cshperspect.a020552

Gerhard, S., Andrade, I., Fetter, R. D., Cardona, A., Schneider-Mizell, C. M. (2017) Conserved

neural circuit structure across Drosophila larval development revealed by comparative

connectomics. eLife 6, e29089. PMID 29058674, doi: 10.7554/elife.29089

Giaume, C. & Liu, X. (2012) From a glial syncytium to a more restricted and specific glial

networking. J Physiol (Paris) 106, 34-9. PMID 21979115, doi:

10.1016/j.physparis.2011.09.001

Goodman, C. S., Bastiani, M. J., Doe, C. Q. & Dulac, S. (1985) Growth cone guidance and cell

recognition in insect embryos. Dev Biol (NY) 3, 283-300. PMID 3077967

Gordleeva, S. Y., Ermolaeva, A. V., Kastalskiy, I. A. & Kazantsev, V. B. (2019) Astrocyte as

spatiotemporal integrating detector of neuronal activity. Front Physiol 10, 294. PMID

31057412, doi: 10.3389/fphys.2019.00294

Grenningloh, G., Rehm, E. J. & Goodman, C. S. (1991) Genetic analysis of growth cone

guidance in Drosophila: Fasciclin II functions as a neuronal recognition molecule. Cell

67, 45-57. PMID 1913818

Grosjean, Y., Grillet, M., Augustin, H. & Ferveur, J.-F. & Featherstone, D. E. (2008) A glial

amino-acid transporter controls synaptic strength and courtship in Drosophila. Nat

Neurosci 11, 54-61. PMID 18066061, doi: 10.1038/nn2019

Grueber, W. B., Ye, B., Yang, C.-H., Younger, S., Borden, K., Jan, L. Y. & Jan, Y. N. (2007)

Projections of Drosophila multidendritic neurons in the central nervous system: Links

with peripheral dendrite morphology. Development 134, 55-64. PMID 17164414, doi:

10.1242/dev.02666

Halassa, M. H., Fellin, T., Takano, H., Dong, J.-H. & Haydon, P. G. (2007) Synaptic islands

defined by the territory of a single astrocyte. J Neurosci 27, 6473-6477. PMID 17567808,

doi: 10.1523/JNEUROSCI.1419-07.2007

Harris, K.M., Spacek, J., Bell, M.E., Parker, P.H., Lindsey, L.F., Baden, A.D., Vogelstein, J.T.,

and Burns, R. (2015) A resource from 3D electron microscopy of hippocampal neuropil

for user training and tool development. Sci Data 2, 150046. PMID 4555877

Hartenstein, V. (2011) Morphological diversity and development of glia in Drosophila. Glia 59,

1237-1252. PMID 21438012, doi: 10.1002/glia.21162

Hartenstein, V., Spindler, S., Pereanu, W. & Fung, S. (2008) The development of the Drosophila

larval brain. Adv Exp Med Biol 628, 1-31. PMID 18683635, doi: 10.1007/978-0-387-

78261-4

Heckscher, E. S., Zarin, A. A., Faumont, S., Clark, M. Q., Manning, L., Fushiki, A., Schneider-

Mizell, C. M., Fetter, R. D., Truman, J. W., Zwart, M. F., Landgraf, M., Cardona, A.,

Lockery, S. R. & Doe, C. Q. (2015) Even-Skipped(+) interneurons are core components

of a sensorimotor circuit that maintains left-right symmetric muscle contraction

amplitude. Neuron 88, 314-29. PMID 26439528, doi: 10.1016/j.neuron.2015.09.009

Hildago, A. & Brand, A. H. (1997) Targeted neuronal ablation: the role of pioneer neurons in

guidance and fasciculation in the CNS of Drosophila. Development 124, 3253-3262.

PMID 9310320

Ito, K., Awano, W., Suzuki, K., Hiromi, Y. & Yamamoto, D. (1997) The Drosophila mushroom

body is a quadruple structure of clonal units each of which contains a virtually identical

set of neurons and glial cells. Development 124, 761-771. PMID 9043058

Kanemaru, K., Sekiya, H., Xu, M., Satoh, K., Kitajima, N., Yoshida, K., Okubo, Y., Sasaki, T.,

Moritoh, S., Hasuwa, H., et al. 2014. In vivo visualization of subtle, transient, and local

activity of astrocytes using an ultrasensitive Ca2+ indicator. Cell Rep 8, 311–318. PMID

24981861, doi: 10.1016/j.celrep.2014.05.056

Keshishian, H. & Chiba, A. (1993) Neuromuscular development in Drosophila: Insights from

single neurons and single genes. Trends Neuroscience 16, 278-283. PMID 7689772

Kim, M. D., Wen, Y. & Jan, Y.-N. (2009) Patterning and organization of motor neuron dendrites

in the Drosophila larva. Dev Biol 336, 213-221. PMID 19818341, doi:

10.1016/j.ydbio.2009.09.041

Kohsaka, H., Guertin, P. A., Nose, A. (2017) Neural circuits underlying fly larval locomotion.

Curr Pharm Des 23, 1722–1733. PMID 27928962, doi:

10.2174/1381612822666161208120835

Kohsaka, H., Takasu, E., Morimoto, T. & Nose, A. (2014) A group of segmental premotor

interneurons regulates the speed of axial locomotion in Drosophila larvae. Curr Biol 24,

2632-2642. PMID 25438948, doi: 10.1016/j.cub.2014.09.026

Kremer, M.C., Jung, C., Batelli, S., Rubin, G.M. & Gaul, U. (2017) The glia of the adult

Drosophila nervous system. Glia 65, 606-638. PMID: 28133822, doi:10.1002/glia.23115

Landgraf, M., Jeffrey, V., Fujioka, J. B., Jaynes, J. B. & Bate, M. (2003a) Embryonic origin of a

motor system: Motor dendrites form a myotopic map in Drosophila. PLoS Biol 1, e41.

PMID 14624243, doi: 10.1371/journal.pbio.0000041

Landgraf, M., Sánchez-Soriano, N., Technau, G. M., Urban, J. & Prokop, A. (2003b) Charting

the Drosophila neuropile: A strategy for the standardised characterization of genetically

amenable neurites. Dev Biol 260, 207-225. PMID 12885565, doi: 10.1016/S0012-

1606(03)00215-X

Liu, H., Zhou, B., Yan, W., Lei, Z., Zhao, X., Zhang, K. & Guo, A. (2014) Astrocyte-like glial

cells physiologically regulate olfactory processing through the modification of ORN-PN

synaptic strength in Drosophila. Eur J Neurosci 40, 2744-2754. PMID 24964821, doi:

10.1111/ejn.12646

Lopez-Hidalgo, N. & Schummers, J. (2014) Cortical maps: A role for astrocytes? Curr Opin

Neurobiol 24, 176-189. PMID 24419141, doi: 10.1016/j.conb.2013.11.001

Ma, Z., Stork, T., Bergles, D.E. & Freeman, M.R. (2016) Neuromodulators signal through

astrocytes to alter neural circuit activity and behavior. Nature 539, 428-432. PMID

27828941, doi: 10.1038/nature20145

MacDonald, J. M., Beach, M. G., Porpiglia, E., Sheehan, A. E., Watts, R. J. & Freeman, M. R.

(2006) The Drosophila cell corpse engulfment receptor Draper mediates glial clearance

of severed axons. Neuron 50, 869-881. PMID 16772169, doi:

10.1016/j.neuron.2006.04.028

MacNamee, S. E., Liu, K. E., Gerhard, S., Tran, C. T., Fetter, R. D., Cardona, A., Tolbert, L. P.,

Oland, L. A. (2016) Astrocyte glutamate transport regulates a Drosophila CNS synapse

that lacks astrocyte ensheathment. J Comp Neurol 524, 1979-1998. PMID 27073064,

doi:10.1002/cne.24016

Mathew, D., Popescu, A. & Budnik, V. (2003) Drosophila amphiphysin functions during

synaptic Fasciclin II membrane cycling. J Neurosci 23, 10710-10716. PMID 14627656

Merritt, D. J. & Whitington, P. M. (1995) Central projections of sensory neurons in the

Drosophila embryo correlate with sensory modality, some position, and proneural gene

function J Neurosci 15, 1755-1976. PMID 7891133

Miller, R. H. & Raff, M. C. (1984) Fibrous and protoplasmic astrocytes are biochemically and

developmentally distinct. J Neuroscience 4, 585-592. PMID 3753719

Molofsky, A. V., Krencik, R., Ullian, E. M., Tsai, H. H, Deneen, B., Richardson, W. D., Barres,

B. A. & Rowitch, D. H. (2012) Astrocytes and disease: A neurodevelopmental

perspective. Genes Dev 26, 891-907. PMID 22549954, doi: 10.1101/gad.188326.112

Murphey, R. K., Poissidente, D., Pollack, G. & Merritt, D. J. (1989) Modality-specific axonal

projections in the CNS of the flies Phormia and Drosophila. J Comp Neurol 290, 185-

200. PMID 2512333, doi: 10.1002/cne.902900203

Muthukumar, A. K., Stork, T. & Freeman, M. R. (2014) Activity-dependent regulation of

astrocyte GAT levels during synaptogenesis. Nat Neurosci 17, 1340-1350. PMID

25151265, doi: 10.1038/nn.3791

Nern, A., Pfeiffer, B. D. & Rubin, G. M. (2015) Optimized tools for multicolor stochastic

labeling reveal diverse stereotyped cell arrangements in the fly visual system. Proc Natl

Acad Sci USA 112, e2967-2976. PMID 25964354, doi: 10.1073/pnas.1506763112

Ogata, K. & Kosaka, T. (2002) Structural and quantitative analysis of astrocytes in the mouse

hippocampus. Neuroscience 113,221-233. PMID 12123700

Ohyama, T., Schneider-Mizell, C. M., Fetter, R. D., Aleman, J. V., Franconville, R., Rivera-

Alba, M., Mensh, B. D., Branson, K. M., Simpson, J. H., Truman, J. W., Cardona, A &

Zlatic M (2015) A multilevel multimodal circuit enhances action selection in Drosophila.

Nature 5120, 633-9. PMID 25896325, doi: 10.1038/nature14297

Panatier, A., Vallee, J., Haber, M., Murai, K. K., Lacaille, J. C. & Robataille, R. (2011)

Astrocytes are endogenous regulators of basal transmission at central synapses. Cell 146,

785-798. PMID 21855979, doi: 10.1016/j.cell.2011.07.022

Peco, E., Davla, S., Camp, D., Stacey, S. M., Landgraf, M., & van Meyel, D. J. (2016)

Drosophila astrocytes cover specific territories of the CNS neuropil and are instructed to

differentiate by Prospero, a key effector of Notch. Development 143, 1170-1181. PMID

26893340, doi: 10.1242/dev.133165

Pereanu, W., Shy, D. & Hartenstein, V (2005) Morphogenesis and proliferation of the larval

brain glia in Drosophila. Dev Biol 283, 191-203. PMID 15907832, doi:

10.1016/j.ydbio.2005.04.024

Peters, A., Palay, S. l. & Webster, H. deF. (1976) Fine Structure of the Nervous System: Neurons

and Their Supporting Cells. Oxford: Oxford University Press.

Pflüger, H. J. & Watson, A. H. (1988) Structure and distribution of dorsal unpaired median

(DUM) neurons in the abdominal nerve cord of male and female locusts. J Comp Neurol

268, 329-345. PMID 3360992, doi: 10.1002/cne.902680304

Pflüger, H. J., Braunig, P. & Hustert R (1988) The organization of mechanosensory neuropiles in

locust thoracic ganglia. Phil Trans R Soc (London) Ser B 321, 1-26.

Richier, B., de Miguel Vijandi, C., Mackensen, S., & Salecker, I. (2017) Lapsyn controls branch

extension and positioning of astrocyte-like glia in the Drosophila optic lobe. Nature

Communic 8, 317. PMID 28827667, doi:10.1038/s41467-017-00384-z

Rickert, C., Kunz, T., Harris, K.-L., Whitington, P., & Technau, G. M. (2011) Morphological

characterization of the entire interneuron population reveals principles of neuromere

organization in the ventral nerve cord of Drosophila. J Neurosci 31, 15870-15883. PMID

22049430, doi: 10.1523/jneurosci.4009-11.2011

Rival, T., Soustelle, L., Cattaert, D., Strambi, C., Iché, M. & Birman, S. (2006) Physiological

requirement for the glutamate transporter dEAAT1 at the adult Drosophila

neuromuscular junction. J Neurobiol 66, 1061-1074. PMID 16838372, doi:

10.1002/neu.20270

Rival, T., Soustelle, L., Strambi, C., Beeson, M.-T., Iché, M. & Birman, S. (2004) Decreasing

glutamate buffering capacity triggers oxidative stress and neuropil degeneration in the

Drosophila brain. Curr Biol 14, 599-605. PMID 15062101, doi:

10.1016/j.cub.2004.03.039

Roux, L., Benchenane, K., Rothstein, J. D., Bonvento, G. & Giaume, C. (2011) From the cover:

Plasticity of astroglial networks in olfactory glomeruli. Proc Natl Acad Sci (USA) 108,

18442-18446. PMID 21997206, doi: 10.1073/pnas.1107386108

Rungta, R. L., Bernier, L. P., Dissing-Olesen, L., Groten, C. J., LeDue, J. M., Ko, R., Drissler, S.

& MacVicar, B. A. (2016) Ca+2 transients in astrocyte fine processes occur via Ca+2

influx in the adult mouse hippocampus. Glia 64, 2093-2103. PMID 27479868, doi:

10.1002/glia.23042

Schneider-Mizell, C. M., Gerhard, S., Longair, M., Kazimiers, T., Li, F., Zwart, M. F.,

Champion, A., Midgley, F. M., Fetter, R. D., Saalfield, S. & Cardona, A. (2016)

Quantitative neuroanatomy for connectomics in Drosophila. eLife 5, e12059. PMID

2699, doi: 10.7554/eLife.12059

Schrader, S. & Merritt, D. J. (2000) Central projections of Drosophila sensory neurons in the

transition from embryo to larva. J Comp Neurol 425, 34-44. PMID 10940940

Semyanov, A. (2019) Spatiotemporal pattern of calcium activity in astrocytic network. Cell

Calcium 78, 15-25. PMID 30579813, doi: 10.1016/j.ceca.2018.12.007

Shigetomi, E., Bushong, E. A., Haustein, M. D., Tong, X., Jackson-Weaver, O., Kracun, S., Xu,

J., Sofroniew, M. V., Ellisman, M. H., & Khakh, B. S. (2013) Imaging calcium

microdomains within entire astrocyte territories and endfeet with GCaMPs expressed

using adeno-associated viruses. J Gen Physiol.141, 633–647. PMID 23589582, doi:

10.1085/jgp.201210949

Southern, J. A., Young, D. F., Heaney, F., Baumgartner, W. & Randall, R. E. (1991)

Identification of an epitope on the P and V proteins of simian virus 5 that distinguishes

between two isolates with different biological characteristics. J Gen Virol 72, 1551-1557.

PMID 1713260, doi: 10.1099/0022-1317-72-7-1551

Stacey, S. M., Muraro, N. I., Peco, E., Labbé, A., Thomas, G. B., Baines, R. A. & van Meyel, D.

J. (2010) Drosophila glial glutamate transporter Eaat1 is regulated by fringe-mediated

notch signaling and is essential for larval locomotion. J Neurosci 30, 14446-14457.

PMID 20980602, doi: 10.1523/JNEUROSCI.1021-10.2010

Stobart, J. L., Ferrari, K. D., Barrett, M. J. P., Stobart, M. J., Looser, Z. J., Saab, A.S. & Weber,

B. (2018) Long-term in vivo calcium imaging of astrocytes reveals distinct cellular

compartment responses to sensory stimulation. Cereb Cortex 28, 184-198. PMID

28968832, doi: 10.1093/cercor/bhw366

Stork, T., Sheehan, A., Tasdemir-Yilmaz, O. E. & Freeman, M. R. (2014) Neuron-glia

interactions through the Heartless FGF receptor signaling pathway mediate

morphogenesis of Drosophila astrocytes. Neuron 83, 388-403. PMID 35033182,

doi:10.1016/j.neuron.2014.06.026

Strausfeld, N. J. (1976) Atlas of an Insect Brain. New York: Springer-Verlag.

Tastekin, I., Khandelwal, A., Tadres, D., Fessner, N. D., Truman, J. W., Zlatic, M., Cardona, A,

& Louis, M. (2018) Sensorimotor pathway controlling stopping behavior during

chemotaxis in the Drosophila melanogaster larva. eLife 7, e38740. PMID 30465650.

Doi: 10.7554/eLife.38740

Tessier, C. R. & Broadie, K. (2009) Activity-dependent modulation of neural circuit synaptic

connectivity. Front Molec Neuroscience 2, 8. PMID 19668708, doi:

10.3389/neuro.02.008.2009

Theodosis, D. T., Poulain, D. A. & Oliet, S. H. R. (2008) Activity-dependent structural and

functional plasticity of astrocyte-neuron interactions. Physiol Rev 88, 983-1008. PMID

18626065, doi:10.1152/physrev.00036.2007

Verkhratsky, A. & Nedergaard, M. (2018) Physiology of astroglia. Physiol Rev 98, 239-389.

PMID 29351512, doi: 10.1152/physrev.00042.2016

Viswanathan, S., Williams, M. E., Bloss, E. B., Stasevich, T. J., Speer, C. M., Nern, A., Pfeiffer,

B. D., Hooks, B. M., Li, W.-P., English, B. P., Tian, T., Henry, G. L., Macklin, J. J.,

Patel, R., Gerfen, C. R., Zhuang, X., Wang, Y., Rubin, G. M. & Looger, L. I. (2015)

High-performance probes for light and electron microscopy. Nat Methods 12, 568-576.

PMID 25915120, doi: 10.1038/nmeth.3365

Vonhoff, F. & Duch, C. (2010) Tiling among stereotyped dendritic branches in an identified

Drosophila motoneuron. J Comp Neurol 518,2169-2185. PMID: 20437522, doi:

10.1002/cne.22380

Wagh, D. A., Rasse, T. M., Asan, E., Hofbauer, A., Schwenkert, I., Dürrbeck, H., Buchner, S.,

Dabauvalle, M.-C., Schmidt, M., Qin, G., Wichmann, C., Kittel, R., Sigrist, S. J. &

Buchner, E. (2006) Bruchpilot, a protein with homology to ELKS/DAST, is required for

structural integrity and function of synaptic active zones in Drosophila. Neuron 49, 833-

844. doi: 10.1016/j.neuron.2006.02.008

Worrell, J. W. & Levine, R. B. (2008) Characterization of voltage-dependent Ca2+ currents in

identified Drosophila motoneurons in situ. Neurophysiol 100, 868-878. PMID

18550721, doi: 10.1152/jn.90464.2008

Yildirim, K., Petri, J., Kottmeier, R. & Klämbt, C. (2019) Drosophila glia: Few cell types and

many conserved functions. Glia 67, 5-26. PMID 30443934, doi: 10.1002/glia.23459

Zhang, Y. V., Ormerod, K. G. & Littleton, J. T. (2017) Astrocyte Ca2+ influx negatively

regulates neuronal activity. eNeuro 4, e0340-16.2017. doi:

dx.doi.org/10.1523/eneuro.0340-16.2017

Zlatic, M., Landgraf, M. & Bate, M. (2003) Genetic specification of axonal arbors: atonal

regulates robo3 to position terminal branches in the Drosophila nervous system. Neuron

37, 41-51. PMID 12526771

Zlatic, M., Li, F., Strigini, M., Grueber, W. & Bate, M. (2009) Positional cues in the Drosophila

nerve cord: Semaphorins pattern the dorso-ventral axis. PLoS Biol 7, e1000135. PMID

19547742, doi: 10.1371/journal.pbio.1000135

* The data that support the findings of this study are available from the corresponding author

upon reasonable request.

Tables

Table 1: Primary antibodies

Antibody Source Concen. Characterization Mouse anti-FasII DSHB*, Cat# 1:200 Labeled a 97-kDa band in Western blot, which 1D4 concentrate was absent in FasII null mutants (Grenningloh, et al., 1991; Mathew et al., 2003). The labeling pattern observed in this study is identical to previous reports (Grenningloh, et al., 1991; Landgraf et al., 2003b; Mathew et al., 2003). Mouse anti-nc82, DSHB*, 1:50 Labels synaptic active zones (Wagh, et al., 2006). concentrate Cat#nc82 Panneural reduction of bruchpilot (BRP) expression by RNAi constructs results in loss of T- bars and absence of active zone labeling. Ectopic expression of nc82 resulted in nc82-positive spots matching the GFPBRP signal appeared. Western blots from fly heads overexpressing brp mRNA in all neurons showed strong enhancement of the 190 kDa signal present in WT CNS tissue. Mouse anti-‐V5 Bio-‐Rad 1:200 Raised against the V5 epitope, which is found in #MCA1360 the P and V proteins of the paramyxovirus SV5 (Southern, et al., 1991). Recognizes the sequence: Gly-Lys-Pro-Ile-Pro-Asn-Pro-Leu-Leu-Gly-Leu- Asp-Ser-Thr. Western blot with purified (25 ng) or recombinant (50 ng) proteins gave a detectable signal on immunoblot, with low background. Rabbit anti-‐HA Cell Signaling 1:200 Monoclonal antibody raised against influenza Technology hemagglutinin epitope (YPYDVPDYA). Detects #3724S exogenously expressed proteins fused to the HA epitope tag (Nern, et al., 2015). Rat anti-‐FLAG Novus 1:200 Recognizes epitope tag fused to targeted smGFP (DYKDDDDK Biologicals reporter vector (Nern, et al., 2015). Epitope Tag) #NBP1-‐06712

*Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA 52242

Table 2. Astrocyte volume by cell-body position and shape, and key morphological features

Cell- # of Glob. Irreg. Multi- Mean crossing unbranched with body cells shape shape arbor volume (µm3) midline boundary process position shape + SD process exiting CNS Dorsal 18 22% 61% 17% 5,419 + 2,557 83% 0% 0% Lateral 6 50% 33% 17% 8,650 + 4,864 0% 33% 17% Ventral 24 71% 29% 0% 7,574 + 3,668 79% 67% 0% Medial 3 33% 67% 0% 5,775 + 1,108 67% 33% 0%

Table 3. Branching patterns of astrocytes by morphological subtype

Characteristic Cell 41 Cell 15 Cell 8 Cell 71 Cell 31 Cell 53 Cell 47 Cell 8 (Fig. 3a) (Fig. 1b) (Fig. 1c) (Fig. 3a) Segment T3 A4 A6 T1 A2 A1 A4 A6 Morphology Globular Globular Globular Irregular Irregular Irregular Multi- Globular arbor # of branches 1417 1806 1807 3173 4222 2783 2653 1807 Branch order: 53 38 52 50 54 76 82 52 maximum Branch order: 21.8 + 18.2 + 19.7 + 23.4 + 26.3 + 22.7 + 31.1 + 19.7 + M + SD 11.2 7.6 10.2 9.9 10.4 14.7 19.7 10.2 Branch angle: 32.8o + 32.6o + 29.1o + 31.9o + 31.2o + 30.5o + 28.2o + 29.1o + M + SD 35.7o 31.8o 30.7o 31.9o 32.8o 32.8o 30.2o 30.7o Branch angle: 0.3o - 0.5o - 0.2o - 0.2o – 0.2o - 0.1o - 0.4o - 0.2o - range 164.5o 163.9o 154.4o 162.2o 160.7o 156.3o 168.7o 154.4o Branch angle: 16.0o 19.1o 15.7o 17.8o 16.1o 15.3o 15.7o 15.7o median Dendrite length 17.0 13.1 16.6 20.1 16.9 15.8 21.1 16.6 (µm): max Dendrite length 2.1 + 1.7 2.2 + 1.6 2.2 + 1.6 2.2 + 1.6 1.9 + 1.4 2.3 + 1.8 1.9 + 1.5 2.2 + 1.6 (µm): M + SD

Table 4. Segment-specificity of astrocyte arbors by cell-body location

Cell-body Segment location of % of cells with Range of % of location cell body arborization > 15% arbor outside of

in adjacent segment segment Dorsal (n=18) T1-T3 (n=6) 80% 16.9-32.9% A1-A6 (n=12) 42% 16.2-31.0% Lateral (n=6) T2-3 (n=2) 100% 29.5-39.7% A1-A6 (n=4) 80% 20.3-25.0%

Ventral T1-3 (n=10) 100% 17.2-48.2% (n=24) A2-A6 (n=14) 43% 21.6-47.4%

Figure Legends

Figure 1. Ventral nerve cord (VNC) organizational plan. a View from the dorsal surface.

Magenta, FasII-positive fiducial tracts run longitudinally through thoracic (T1-3) and abdominal

(A1-8/9) segments; transverse nerves at the segment boundaries and segmental nerves extending

laterally are also FasII-positive. b. Cross-section though T3. b1. alrm-positive astrocytes (green) with cell bodies in dorsal, lateral and medial positions at the neuropil boundary and with processes infiltrating the neuropil Cross-sectioned longitudinal FasII-positive fascicles (magenta,

arrowheads; cell nuclei in blue (Syto); *, FasII-positive segmental nerve, the thickened regions

corresponding to the thoracic pocket area. b2. Thick FasII-positive segmental nerves extending

laterally and ventrally (green). Large axonal tracts in the neuropil are unlabeled, appearing as

black regions. c. Schematic diagram, cross section of the VNC. Green stars: typical positions of

astrocyte cell bodies. Magenta: FasII-positive tracts. d. Repo-positive, GFP-labeled glial cell bodies, likely exit glial cells, surround the base of the thoracic segmental nerves, forming pockets that are not present in the abdominal segments. e. 3D reconstruction of the neuropil of

T3 and A1 segments showing the variability of position in astrocytes, particularly dorsal astrocytes. This figure shows not the number of astrocytes, but the positions at which astrocyte cell bodies were found in the antero-posterior and medio-lateral dimensions across 3 preparations; some positions were held in common by astrocytes in 2 or 3 preparations. The bilateral outpouchings in the ventral T3 segment are the pockets. Cross-section (e1) and tilted

(e2) views from the rostral side. Scale bars: 50 µm in d1, and 25 µm in d2.

Figure 2. Growth of the VNC during larval development. a. 1st instar; b. 3rd instar. Dotted line:

boundary of the VNC as viewed from the dorsal surface. Astrocytes, labelled by MCFO, show

considerable arbor growth in size and complexity. Scale bars: 50 µm.

Figure 3. Astrocyte morphology. 3 basic morphological types, each 3D-reconstructed cell shown

embedded in a frame with a1-c1 in X-Y orientation (from the dorsal surface) with FasII tracts

(magenta). a2-c2 in X-Z orientation; axes in microns. a3-c3. Skeleton reconstruction for each cell in the rotation that best shows the morphological features that characterize the subtype. Cell bodies: teal circles. b. Fas II tracts: magenta. Arrowheads: transverse nerves at segment boundaries. Dotted line: midline. Videos showing rotation of 3D filament versions are included as supplemental material.

Figure 5. Bilateral and between-segment morphological symmetry. a1 and a3, b1and b3, show reconstructions of pairs of cells that share structural similarity. Panels a2-a4 and b2-b4 also show 2 adjacent pairs of cells with major branches extending deep into each other’s territory. d.

Astrocytes in adjacent segments extending similar processes rostrally to wrap the same longitudinal tract, dotted line. MCFO preparation. Scale bar in d: 50 µm.

branch of the ventral cell that protrudes into the territory of the lateral cell (arrowheads) with

minimal intermingling of fine branches. d. Method used to determine overlap. d1. Schematic of

two adjacent astrocytes in different channels. d2-d3 Cell in each channel is isolated using a

bounding box, reconstructed in 3D to create an iso-surface, and volume determined. d4

Processes in region of overlap isolated via masking process described in methods section.

Volume of processes in overlap region determined. Percent overlap calculated (volume of

overlapping processes divided by total volume of both cells). Scale bars in a: 10 µm; in c1-c3: 10

µm.

Figure 7. Astrocytes can branch in more than one segment. a. Dorsal view of a VNC in which

looper neurons expressed GFP and transverse nerves were labeled with anti-FasII. Except in T1 and T2, which have a mid-segment FasII-positive nerve bundle, the transverse nerve and the laterally projecting looper axon tract essentially overlap. b. Astrocyte arborizes totally within a single segment. c. Astrocyte arborization shared nearly equally between segments. Boxes drawn to segment boundaries using FasII transverse nerve as a marker; dotted edge is midline. Scale bar in a: 50 µm.

domains (modified from Zlatic et al., 2009 and Burgos et al., 2018). d-h. Overlaid tracings of

the perimeter of each cell’s territory within each cell-body position (indicated by stars). Colored

transparent overlay in each panel indicates the territory held in common by all of the astrocytes

in a given cell-body position. d,f. Thoracic segments. e,g. Abdominal segments. h. Combined

thoracic and abdominal tracing for lateral astrocytes. Graphs. For each cell-body position,

relationship of astrocyte territories to FasII-positive longitudinal tracts.

Figure 9. Astrocyte and neuropil volume. a. Looper axon tracts used as boundary markers.

Synaptic neuropil labeled with NC82 (not shown) and reconstructed in 3D. Upper three

reconstructions are thoracic segments; lower three are for abdominal segments A3, 5 and 6. b.

Neuropil volume/segment. c. Volume of astrocytes sorted by cell-body position, by segment, and by morphology. Each dot is an individual cell, with a total of 51 cells. d. Astrocyte volume by

segment. Error bars: SEM.