Isolation of specific neurons from C. elegans larvae for gene expression profiling

Abstract

Background: The simple and well-described structure of the C. elegans nervous system offers an unprecedented opportunity to identify the genetic programs that define the connectivity and function of individual neurons and their circuits. A correspondingly precise gene expression map of C. elegans neurons would facilitate the application of genetic methods toward this goal. Here we describe a powerful new approach, SeqCeL (RNA-Seq of C. elegans cells) for producing gene expression profiles of specific larval C. elegans neurons.

Methods and results: We have exploited available GFP reporter lines for FACS isolation of specific larval C. elegans neurons for RNA-Seq analysis. Our analysis showed that diverse classes of neurons are accessible to this approach. To demonstrate the applicability of this strategy to rare neuron types, we generated RNA-Seq profiles of the NSM serotonergic neurons that occur as a single bilateral pair of cells in the C. elegans pharynx. These data detected >1,000 NSM enriched transcripts, including the majority of previously known NSM-expressed genes.

Significance: This work offers a simple and robust protocol for expression profiling studies of post-embryonic C. elegans neurons and thus provides an important new method for identifying candidate genes for key roles in neuron-specific development and function.

Figure 1. Isolation of C. elegans neurons from larval animals by Fluorescence Activated Cell Sorting (FACS).

(A) Expression of the pan-neural GFP marker, evIs111, in an L2 stage larval animal. Anterior is to left. (B) Procedure for generating dissociated suspensions of larval cells for primary culture and for isolation of GFP-marked cells (green) by FACS. (C) Primary culture of L1-stage cells 36 hr after plating. Arrows point to elongated processes extending from GFP-labeled neurons. Scatterplot of FACS profile for cells dissociated from the wild-type (N2) reference strain (D) and from the pan-neural evIs111 line (E). Propidium Iodide (PI) marks dead cells (blue). GFP-labeled cells (green) were isolated by FACS (outlined with box). The majority (∼90%) of FACS-derived cells (F) are marked with GFP (G) in primary cultures examined within 2 hr of FACS isolation. Note that neurons are loosely attached which likely accounts for the displacement of DIC vs GFP images of individual cells in these micrographs. Scale bars = 10 micron.

Figure 2. Specific sensory and motor neurons are accessible to isolation by FACS from multiple larval stages.

Primary cultures of L1 stage cells 24 hr after dissociating from transgenic lines expressing (A) mec-4::mCherry to mark ALM and PLM neurons (red) and (B) srh-142p::dsRed to label ADF sensory neurons (red). (C) GABA motor neuron marked with unc-47::mCherry (red) and cultured for 24 hour after dissociating from L4 larval animals. (D) Primary culture of A-class motor neurons marked with unc-4::GFP (green) and isolated by FACS from L2 stage larvae. (E) del-1::GFP labels VB motor neurons (green) in the ventral nerve cord of an L2 stage larva. Anterior to left. (F) FACS profile of cells dissociated from del-1::GFP L2 larvae. Propidium Iodide (PI) marks dead cells (blue). Viable GFP-labeled cells (green) are outlined with the box. Scale bars are 10 microns.

Figure 3. Isolation of NSM serotonergic neurons by FACS.

(A) Head region of an adult hermaphrodite depicting the NSML neuron (orange) in the pharynx (green) (WormAtlas). Anterior is to left. (B) Confocal image of NSML in L1 larval stage. Anterior is to left. (C) FACS profile of cells dissociated from L1 stage larvae expressing tph-1::GFP in NSM neurons. Propidium Iodide (PI) marks dead cells (blue). Viable GFP-labeled cells (green) are outlined with the box. Images of FACS-isolated NSM neurons (D) 24 hr, (E), 48 hr and (F) 72 hr after plating. Scale bars are 10 microns.

Figure 4. 5′-to-3′ coverage.

(A) The average relative coverage for each library is shown for normalized transcript length. The Y-axis denotes the average coverage of each position in the transcript normalized to the total number of mapped reads. Coverage for coding sequence genes (B) tph-1 and (C) bas-1 and for (D) a small nuclear RNA (snoRNA) gene, ZK643.9. Lighter vs darker shades of color depict results for independent NSM (blue) and reference (green) samples.

Figure 5. Differential RNA-Seq analysis detects transcripts that are highly expressed in NSM neurons.

Pairwise comparisons of (A) NSM and (B) Reference data sets with expression values represented as log₂(FPKM) and Spearman rank-order correlation coefficients, r_s. (C) Volcano scatter plot of log₁₀(p-value) vs log₂(fold change) of transcript expression in NSM neurons relative to the reference sample derived from all L1 larval cells. Significantly enriched or depleted transcripts (≥2.4 fold, p<0.012) are indicated in red. (D) Schematic of NSM serotonergic presynaptic terminus depicting genes that are highly enriched (E) in the NSM RNA-Seq profile. See text for additional information about specific genes.

Figure 6. GFP reporters validate NSM-enriched RNA-Seq data set.

Transgenic animals expressing promoter::GFP reporter genes for NSM-enriched transcripts. Confocal images of adult hermaphrodites, anterior to left, ventral down. A–B, Bright tph-1::GFP (vsIs45) expression is limited to NSML and NSMR; note dim tph-1::GFP signal in ADF neurons (arrows). C, Cartoon depicting NSM neurons marked with tph-1::GFP in the anterior pharyngeal bulb. Y39B6A.19::GFP (D, E) and F41E7.3::GFP (F, G) are specifically expressed in neurons, including NSM, head neurons and ventral cord motor neurons (asterisks). Y39B6A.19::GFP is also expressed in sensory neuron, PDE (E). Scale bar is 25 microns.