Journal of Visualized Experiments www.jove.com Video Article Isolation and Purification of Drosophila Peripheral Neurons by Magnetic Bead Sorting

The Drosophila peripheral nervous system (PNS) is a powerful model for investigating the complex processes of neuronal development and dendrite morphogenesis at the functional and molecular levels. To aid in these analyses, we have developed a strategy for the isolation of a subclass of PNS neurons called dendritic arborization (da) neurons that have been widely used for studying dendrite morphogenesis1,2. These neurons are very difficult to isolate as a pure population, due in part to their extremely low occurrence and their difficult-to-reach location below the tough chitinous larval cuticle. Our newly developed method overcomes these challenges, and is based on a fast and specific cell enrichment using antibody-coated magnetic beads. For our magnetic bead sorting studies, we have used age-matched third instar larvae expressing a mouse CD8 tagged GFP fusion protein (UAS-mCD8-GFP)3 under the control of either the class IV dendritic arborization (da) neuron-specific pickpocket (ppk)-GAL4 driver4 or the control of the pan-da neuron-specific GAL421-7 driver5. Although this protocol has been optimized for isolating PNS cells which are attached to the inner wall of the larval cuticle, by varying a few parameters, the same protocol could be used to isolate many different cell types attached to the cuticle at larval or pupal stages of development (e.g. epithelia, muscle, oenocytes etc.), or other cell types from larval organs depending upon the GAL4-specific driver expression pattern. The RNA isolated by this method is of high quality and can be readily used for downstream genomic analyses such as microarray gene expression profiling studies. This approach offers a powerful new tool to perform studies on isolated Drosophila dendritic arborization (da) neurons thereby providing novel insights into the molecular mechanisms underlying dendrite morphogenesis.


Removal of loosely adherent non-specific cells: (2-3 minutes)
[This step aids the clearing of loosely adherent non-specific tissues such as fat bodies and CNS.]

Dissociating the tissue into a single cell suspension: (18-20 minutes)
[Critical Step: Over-dissociation may cause the loss of the cell-surface marker leading to poor cell yield and low cell viability. The larval tissue can be dissociated by either mechanical dissociation (sonication, douncing), enzymatic dissociation (trypsin, collagenase etc.) or a combination of both. As these larval tissues are difficult to dissociate, we found that a combination of both mechanical and enzymatic dissociation yielded the best results.] 1. Place 10-12 larvae on the center of a Sylgard coated 35mm Petri plate. Position them slightly away from each other. [Critical Step: Discard any larvae that do not seem to be at the appropriate developmental stage] 2. Cut open the anterior tip of the larvae using a pair of fine dissection scissors for all the larvae. 3. Using a pair of dull Dumont No. 5 forceps invert the larvae inside-out. Insert one forcep inside the larval cuticle all the way to the posterior end. Pinch the tips of the forceps together to grab the posterior end of the cuticle (Figure 1a) (try pressing the cuticle down on the Sylgard surface to make it easier). Using the second pair of forceps, push the larval cuticle inside out. [Critical Step: Try practicing this method a few times before attempting the cell isolation experiment. Try to get the larvae completely inverted, to ensure that all the soft tissues are exposed to the solution for easy dissociation.] 4. After dissecting 3-4 larvae, transfer them immediately to fresh, ice-cold PBS (place the tube on ice) in a 1.6 ml microfuge tube. 5. Repeat steps 3.1 to 3.4 until all the required larvae are collected (30-40 larvae in this protocol). [Critical Step: Anticipate 10-20% loss during dissection and dissociation, and plan accordingly.] 1. Take the 1.6 ml microfuge tube containing the inverted larval cuticles and replace the supernatant with approximately 700-800 μl of fresh ice-cold PBS. Typically the sample will be highly enriched for fluorescent cells.

Representative Results:
Magnetic bead sorting was used to isolate Drosophila da neurons (Figure 1). The RNA purified from these isolated da neurons (Figure 2a) was found to be of excellent quality as indicated by the presence of sharp 5.8S, 18S and 28S ribosomal RNA peaks when analyzed on an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.) (Figure 2b). Beginning with 30-40 third instar larvae we were capable of isolating on average 300-500 class-IV da neurons using the ppk-GAL4 driver, and 1500-2000 da neurons (Class I,II,III & IV) using the pan-da neuron-specific GAL4 21-7 driver. To assess the neuronal-specific enrichment of our isolated cells we performed quantitative reverse transcription PCR (qRT-PCR) using two neuronal gene-specific markers (elav and futsch). These analyses revealed significant fold enrichment of both marker genes indicating a highly specific enrichment for da neurons as compared to flow through using our protocol ( Figure 3). Finally, the isolated RNA from both pan-da neurons and class-IV da neurons was used to perform transcriptional expression profiling on Agilent Drosophila melanogaster whole-genome oligo microarrays (4 x 44K) ( Figure 4). These analyses identified numerous previously implicated regulators of da neuron dendrite morphogenesis in addition to a broad spectrum of previously uncharacterized molecules and putative signal transduction pathways that potentially play important functional roles in da neuron development. Studies designed to assess the potential role(s) of these previously uncharacterized molecules in mediating da neuron development, and specifically dendrite morphogenesis, are presently underway.

Discussion
The protocol presented here is optimized for the isolation and purification of peripheral neurons which adhere tightly to the inner surface of the Drosophila third instar larval cuticle using a magnetic bead cell sorting strategy. While we have used this protocol to specifically isolate Drosophila da neurons, applications of this protocol to the isolation of other cell types that adhere to the cuticle in larval or pupal stages of development (e.g. epithelia, muscle, other peripheral neurons) can be adapted by varying a few parameters and using distinct GAL4,UAS-mCD8-GFP reporter transgenes which label the cell type or types of interest. Moreover, this protocol can be used in both loss-of-function and gain-of-function approaches where a gene of interest may be cloned into a UAS-mCD8-GFP transgene that can be coupled with a GAL4 transgene to direct either gene-specific loss-of-function (e.g. UAS-RNAi) or gain-of-function to a cell type of interest. For example, in the case of a transcription factor one may wish to identify potentially up-or down-regulated genes upon loss-of-function or gain-of-function expression in a cell type of interest. By isolating total RNA from the purified cell type of interest via this protocol and using this RNA to perform microarray expression profiling it is possible to identify differentially regulated genes that may represent downstream targets of transcriptional regulation that play a role in mediating phenotypic changes within the cell.
For successful cell sorting it is essential to give careful attention to the critical steps highlighted in the above protocol. Examples of common problem areas that may require some further troubleshooting and optimization, depending upon cell type, include (1) low cell yield and (2) cell clumping during magnetic bead isolation. In the first case, one may try reducing the concentration of Liberase Blendzyme 3, and compensate by increasing mechanical dissociation via douncing. In the second case, one may try reducing the magnetic field strength by applying a single or multiple layers of adhesive lab tape over the magnet.