Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 May 12;533(7602):206-11.
doi: 10.1038/nature17977. Epub 2016 May 4.

Sex-specific Pruning of Neuronal Synapses in Caenorhabditis Elegans

Affiliations
Free PMC article

Sex-specific Pruning of Neuronal Synapses in Caenorhabditis Elegans

Meital Oren-Suissa et al. Nature. .
Free PMC article

Abstract

Whether and how neurons that are present in both sexes of the same species can differentiate in a sexually dimorphic manner is not well understood. A comparison of the connectomes of the Caenorhabditis elegans hermaphrodite and male nervous systems reveals the existence of sexually dimorphic synaptic connections between neurons present in both sexes. Here we demonstrate sex-specific functions of these sex-shared neurons and show that many neurons initially form synapses in a hybrid manner in both the male and hermaphrodite pattern before sexual maturation. Sex-specific synapse pruning then results in the sex-specific maintenance of subsets of these connections. Reversal of the sexual identity of either the pre- or postsynaptic neuron alone transforms the patterns of synaptic connectivity to that of the opposite sex. A dimorphically expressed and phylogenetically conserved transcription factor is both necessary and sufficient to determine sex-specific connectivity patterns. Our studies reveal new insights into sex-specific circuit development.

Figures

Extended Data Figure 1
Extended Data Figure 1. Adjacency of neuronal processes in hermaphrodites and males
Four TEM prints from wild-type adult hermaphrodite “JSE” and four from adult male “N2Y”, showing adjacency of neuronal processes. Tthese images were collected at MRC/LMB for Ref.6 and Ref.16, and the annotated images are now available online at www.wormimage.org, courtesy of David Hall. The set of processes directly adjacent to one another has been defined as the “neighborhood” of that process , and the placement of processes into specific neighborhoods is a major determinant of connectivity. Connections form only in one sex, although processes are adjacent in both sexes. a: Print #385, JSE series (JSE_122283), http://wormimage.org/image.php?id=122283&page=2 and print #620, N2Y series (PAG620), http://wormimage.org/image.php?id=103528&page=18, shows PHB>AVG adjacent processes, pseudo labeled in green (AVG) and red (PHB). b: Print #359, JSE series (JSE_122257), http://wormimage.org/image.php?id=122257&page=2, and print #500, N2Y series (PAG500), http://wormimage.org/image.php?id=103408&page=20, shows AVG>VD13 adjacent processes, pseudo labeled in green (AVG) and pink (VD13). c: Print #377, JSE series (JSE_122275), http://wormimage.org/image.php?id=122275&page=2, and print #800, N2Y series (PAG800), http://wormimage.org/image.php?id=103706&page=16, shows PHA>AVG adjacent processes, pseudo labeled in green (AVG) and orange (PHA). d: A table summarizing the number of EM sections in which direct adjacency of processes was observed. Over a 1000 PAG (preanal ganglion) serial sections were analyzed for each sex.
Extended Data Figure 2
Extended Data Figure 2
a: The connectivity diagram shown in Fig. 1a but now also including dimorphic connections of the LUA and PHC connections. The GRASP data that we show in this figure as well as the pruning data, sexual reversal data and mutant data shown in other Extended Figures, supports the original LUA connectivity data reported in Ref.3 and summarized in this schematic. However, a recent reassessment of the tracing of electron micrographs suggests that the connectivity assignments of the PHC and LUA neurons may have been swapped with each other (Scott Emmons, pers. comm.). b: Overview of LUA synaptic connections labeled in this paper. c: Visualizing LUA sexually dimorphic synapses. Quantification and fluorescent micrographs of GRASP transsynaptically labeled puncta between LUA>AVG and LUA>AVA, in L1 and adult hermaphrodites and males. The P, M and N letters next to fluorescent images denote labeling of Puncta, Neurite and Merge, respectively. Region of neurite overlap and observed synaptic puncta marked with white boxes. Gut; auto-fluorescence gut granules. Scale bars are 10µM for adult fluorescent panels and 5 µM for L1 panels. d: Fluorescent micrographs of the preanal ganglion region of transgenic animals expressing the presynaptic BirA::nrx-1 fusion in LUA (using the eat-4p9 promoter), and postsynaptic acceptor peptide::nlg-1 fusion in AVG. For more details see Extended data Figure 3. Scale bars are 10 µM We performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test) with Bonferroni correction for multiple comparisons. ****P < 0.0001, ***P < 0.001, **P < 0.01, NS; non significant. Magenta horizontal bars represent the median.
Extended Data Figure 3
Extended Data Figure 3. Transsynaptic labeling of dimorphic connections using iBLINC
a: Overview of synaptic connections labeled in this paper (Fig. 1 and this figure). Some connections were labeled with both GRASP and iBLINC, yielding similar results. We generally note that the number of synapses is roughly reproducible from animal to animal and the number of the fluorescent dots is roughly comparable to the number of synapses identified by the EM analysis. However, there is also some variance from animal to animal (quantified in Fig. 3), consistent with previous analysis . b: Labeling data not shown in Fig. 1. Fluorescent micrographs of the preanal ganglion region of transgenic animals expressing the presynaptic BirA::nrx-1 fusion in PHB (using the gpa-6 promoter), AVG (using the inx-18 promoter) and PHA (using the srg-13 promoter), and postsynaptic acceptor peptide::nlg-1 fusion in AVG and DA9 (using the acr-2 promoter). Transgenic worms also express the streptavidin detector fused to 2xsfGFP from the coelomocytes (unc-122 promoter) . Neuronal processes are labeled with cytoplasmic Cherry markers of the iBLINC pairs. Region of neurite overlap and observed synaptic puncta are marked with white boxes. Gut; auto-fluorescence gut granules. Scale bars are 10 µM. Anterior is left and dorsal is up.
Extended Data Figure 4
Extended Data Figure 4. Specificity of driver lines
a: A 2.6 kb gpa-6 promoter fragment fused to GFP is expressed consistently in PHB at all stages. There is also faint and variable expression in AWA. b: The 3.1kb flp-18 promoter fused to GFP is expressed consistently in AVA, and dimly and variably in AIY (85% of animals) and RIM (15% of animals), which were identified based on comparison to published flp-18 expression patterns . c: The 1.8 kb inx-18 2nd intron fused to codon optimized Cherry is expressed brightly and consistently in AVG, and dimly and variably in URXs. The AIY motif present in this fragment was deleted (see Methods). d: The 170 bp eat-4 promoter (eat-4p9) fused to GFP is expressed in LUAs and PVR.
Extended Data Figure 5
Extended Data Figure 5. Additional SDS avoidance assays
a: PHB silenced hermaphrodites move slower forward (quantified as absolute midbody speed) and as a result cover less of the plate (quantified as forward path range) compared with control hermaphrodites, while PHB silenced males do not show any difference from control males. In addition, the tail-bending wave is affected in PHB silenced hermaphrodites, but not in males (quantified as tail crawling frequency). Statistics were computed using Wilcoxon rank-sum, and correction for multiple testing (q-values) was computed across all measures (roughly 1404 tests) using the Benjamini–Hochberg procedure. b: Silencing of ASH neuronal activity using the histamine chloride channel 1(HisCl1) affects the animals’ chemosensory avoidance response. Males and hermaphrodites were assayed for affects of histamine on SDS avoidance behavior. We used the him-5 mutant background (which gives a high incidence of male progeny) as wild type. There is no difference between worms assayed in the presence and absence of histamine (Fig. 2a). The avoidance index of single animals was calculated as the fraction of reversal responses in 10 or more assays, depicted as black dots. Magenta vertical bars represent the median. L4 animals carrying the kyEx5104 [pNP424 (sra-6::HisCl1::SL2::mCherry)] transgene were grown on 10mM histamine-containing NGM plates for 24 hours. As a control, kyEx5104 animals were grown on NGM plates without histamine. ASH silencing reduces the head sensory response to SDS, thus in hermaphrodites the antagonizing activity of the PHBs inhibits the backward movement and the worms do not reverse. In males, no such antagonizing activity occurs and the worms reverse, albeit with reduced ability. c: Ablation of ASH neurons affects the animals’ chemosensory avoidance response in a similar manner to histamine-induced silencing. sra-6::gfp was used to identify the ASH neurons. d: Behavioral differences stem from dimorphic connectivity differences and not from amphid/ phasmid sensory function. tax-4; ceh-14 double mutants behave in a similar manner in both sexes. PHB silencing (gpa-6::HisCl1::SL2::GFP) doesn’t affect the worms behavior in both sexes. Silencing both the ASHs and PHBs in males showed no difference from ASH silenced males, but silencing ASHs and PHBs in hermaphrodites showed a significant difference from ASH silenced hermaphrodites, where we expect the PHBs to function in an antagonistic manner, and thus in its absence, ASH silenced hermaphrodites now show an increase in their ability to respond to SDS by reversing. e: PHB silencing in tax-4 mutant background. tax-4 is a subunit of a cyclic nucleotide gated channel expressed in chemosensory and thermosensory neurons , see panel g, tax-4 animals show a strongly reduced avoidance response to SDS . Silencing of PHBs in tax-4 hermaphrodies eliminated the antagonizing affect and animals are able to avoide SDS by backing. f: PHB silencing in tax-4 mutant background at the L1 stage. Lack of avoidance seen in tax-4 L1 males and hermaphrodites depends on PHB function. For all panels, we performed the nonparametric Mann-Whitney test with Bonferroni correction for multiple comparisons. ****P < 0.0001, **P < 0.01, *P < 0.05, NS; non significant. g:tax-4 expression pattern is identical in hermaphrodites and males. kyEx744 (tax-4p::TAX-4::GFP), was analyzed in adult male and hermaphrodites. Amphid neurons in the head (ADL, ASH, ASI, ASJ, ASK, AWB) were stained using DiD to facilitate cell identification. Neurons identified, shown in “Merge” panel are identical in both sexes and match published data. All neurons are bilaterally symmetric left-right pairs, and for simplicity only left cells are shown. Scale bars are 10 µM.
Extended Data Figure 6
Extended Data Figure 6. Additional behavioral analysis
a: Hermaphrodites in which LUAs have been ablated pause more frequently than mock-ablated hermaphrodites and LUA-ablated males. Error bars are S.E.M. b: LUA laser-ablated animals tested for the male’s vulva location efficiency. The behavioral data shown in panel a and b supports the reported connectivity data shown in Extended Figure 2. In panels a, b we performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test) with Bonferroni correction for multiple comparisons. Error bars in b are S.E.M. ***P < 0.001, *P < 0.05, NS; non significant. c: Summary of dimorphic behaviors induced by PHB sensory neurons and AVG/LUA interneurons.
Extended Data Figure 7
Extended Data Figure 7. Time course analysis of synapse pruning and development
Hermaphrodites and males were analyzed at the L1, L3, L4, young adult and gravid adult stages, and the number of synaptic puncta observed at each stage was plotted against developmental time points. Synaptic puncta in hermaphrodites are plotted in red, synaptic puncta in males are plotted in blue. a: PHB>AVG synapses are pruned in hermaphrodites at the L3 stage. b: PHB>AVA synapses are pruned in males at the L3 stage. c: LUA>AVG synapses are pruned earlier, starting at the L1 stage in hermaphrodites. Error bars are SEM, and for each time point depicted in graphs, at least 15 animals were analyzed.
Extended Data Figure 8
Extended Data Figure 8. Autonomy and non-autonomy of sex-specific synapse pruning
a: Cartoon summarizing sex changes effects on synapses. b: L1-stage connectivity is not affected by sex-reversal. c: Simultaneous sex reversal of both AVA and AVG. d: Sex reversal experiments. Masculinization of the postsynaptic cell AVA is sufficient to induce LUA>AVA synaptic puncta in hermaphrodites. The postsynaptic cell AVA was masculinized by expression of FEM-3 and the number of synaptic puncta was measured. Masculinization of AVG was not sufficient to induce synapses between LUA and AVA in hermaphrodites. Feminizing LUA, AVA and AVG by expression of TRA-2IC was sufficient to induce ectopic LUA>AVA puncta in males. We performed the nonparametric Mann-Whitney test with Bonferroni correction for multiple comparisons. ****P < 0.0001, ****P < 0.01, NS; Not significant. Magenta horizontal bars represent the median.
Extended Data Figure 9
Extended Data Figure 9. dmd-5 and dmd-11 expression, sequence and function
a: Quantification of dimorphic expression of dmd-5 and dmd-11 in AVG. Expression in hermaphrodites was off or extremely faint. Expression of inx-18p::FEM-3 derepressed dmd-5 and dmd-11 gene expression in hermaphrodite AVGs. Statistics calculated using Fischer’s Exact test. b: Quantification of the number of PHB-AVG synaptic puncta in L1 dmd-5(gk408945); dmd-11(gk552) double mutants, compared with wild-type L1 animals. At the L1 stage dmd-5 and dmd-11 do not affect PHB-AVG synapses, suggesting they are required for maintenance of mature synapses. c:dmd-5 single mutants and dmd-5; dmd-11 double mutants display similar alterations in AVG synaptic wiring. d:dmd-5 mutation suppresses the ectopic PHB>AVG synapses in AVG-masculinized animals. e: The PHB>AVA connection is non-autonomously partially stabilized in dmd-5; −11 mutants. f:dmd-11 genomic locus and gk552 deletion location. g:dmd-5 genomic locus and mutation description. ok1394 location was not curated, to determine location we used the following primers: Forward primer: cagaatgcctgtttctccgtc and Reverse: cactgcttttcccgttcaaac. ok1394 and tm1760 were both found to have an embryonic lethal phenotype that couldn’t be rescued with the genomic locus (data not shown), thus we searched for single point mutations of the “million mutation project”gk408945 is a missense substitution mutation of W54 to R, located in the second exon. Genomic analysis revealed that this mutation lies within the conserved DM domain (e.), with perfect conservation across evolution. The DM domain is an intertwined zinc-containing DNA binding module. The DM domain binds DNA as a dimer, allowing the recognition of pseudopalindromic sequences . h: DM-domain sequence conservation and location of gk408945 mutation. Conservation and multiple sequence alignment were done using UCSC Genome Browser (genome.ucsc.edu) and ClustalW. i:dmd-5 and dmd-11 are required for maintenance of AVG synapses. Fluorescent micrographs and quantification of synaptic puncta of LUA>AVG. Region of neurite overlap and observed synaptic puncta marked with white boxes. The P, M and N letters next to fluorescent images denote labeling of Puncta, Neurite and Merge, respectively. In b, d , e and i statistics were calculated using the nonparametric Mann-Whitney test. In panel c statistics were calculated using Kruskal-Wallis test with Dunn’s multiple comparison test. ****P < 0.0001, ***P < 0.001, *P < 0.05, NS; Not significant. Magenta horizontal bars represent the median. When using a parametric t-test, there is also a significant difference for the LUA>AVG synapse between dmd-5;dmd-11 mutant hermaphrodites and dmd-5;dmd-11 mutant hermaphrodites that over express DMD-5 (*P < 0.05). j: Summary of data. TRA-1 and DMD proteins are commonly thought to work as transcriptional repressors . Since dmd-5/11 are already dimorphically expressed in AVG in embryos and L1 stage animals (not shown in this schematic), there must be other timer mechanisms that control the onset of pruning. For example, DMD-5/11 may work together with a regulatory factor of the stage-specifically acting heterochronic pathway. Furthermore, we hypothesize that other neurons, such as the AVA neuron, may have its own complement of sex-specific dmd genes that control pruning.
Figure 1
Figure 1. Visualizing sexually dimorphic synapses
a: Connectivity of selected neurons at the adult stage, as inferred from serial section reconstructions of electron micrographs. Chemical synapses between sensory (triangles), inter- (hexagons) and motor (circles) neurons are depicted as arrows. Thickness of arrows correlates with degree of connectivity (number of sections over which en passant synapses are observed). The inset indicates where synaptic connections are formed. b: Visualizing sexually dimorphic synapses. Fluorescent micrographs of GRASP GFP signal in preanal ganglion region outlined in the inset in Fig. 1a. Neuronal processes are labeled with cytoplasmic codon optimized Cherry markers of the GRASP pairs. GRASP data is shown in this Figure, additional iBLINC data is shown in Extended Data Fig. 3. Expression pattern of the promoters used in this study to drive cell specific expression can be found in Extended Data Fig. 4. Quantification of data is shown in Fig. 3b; the number of fluorescent puncta (outlined with white boxes) is similar to those observed in the EM analysis. Gut; auto-fluorescence gut granules. Scale bars are 10 µM. In all images anterior is left, and dorsal is up. Blue and red color-coding is used in all figures to distinguish male from hermaphrodite, respectively.
Figure 2
Figure 2. Functional repurposing of dimorphic neurons
a: Chemosensory repulsion assays (see Methods for full description of behavioral assays). Scatter diagrams plotting the avoidance index of single animals. Each black dot represents the fraction of reversal responses (scored as reversing or not reversing) in 10 or more assays of a single animal. Magenta vertical bars represent the median. The left column indicates predictions of reversal behavior, based on previously published data that demonstrated a strong reversal drive from head neurons (thick arrows) that is counteracted by a forward drive mediated by the PHB neurons in the tail. Control experiments for silencing using histamine and additional SDS assays can be found in Extended Data Fig. 5 sra-6::HisCl1 and gpa-6::HisCl1 were used for ASH and PHB silencing, respectively. Summary of dimorphic behaviors induced by PHB sensory neurons can be found in Extended Data Fig. 6c. b: Changes in male movement and posture triggered by mate-contact. c: Mutant or laser-operated animals tested for the male’s vulva location efficiency. In lin-11 and ceh-14 mutants, the AVG and phasmid neurons, respectively, fail to differentiate11,23_ENREF_11. d: Initiation of backward movement in response to hermaphrodite contact is dependent on PHB activity, measured as contact response efficiency. Each dot represents one animal. In panels a and c we performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test) with Bonferroni correction for multiple comparisons. Error bars in c are S.E.M. In panel d we performed Fisher’s exact test. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, NS; non significant in all panels.
Figure 3
Figure 3. Synaptic pruning during development
a: Models for how sexually dimorphic connectivity patterns may arise during development. b–c: Quantification and fluorescent micrographs of synaptic pruning measured by the number of synaptic puncta observed using GRASP GFP (PHB>AVA, PHB>AVG, AVG>VD13, AVG>DA9) and iBLINC GFP (PHA>AVG) in L1 and adult hermaphrodites and males. Since VD neurons are born at the end of the L1 stage, Juvenile VD13 puncta were quantified at the L2 stage. We performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test) with Bonferroni correction for multiple comparisons. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, NS; non significant. Magenta horizontal bars represent the median. The P, M and N letters next to fluorescent images denote labeling of Puncta, Neurite and Merge, respectively. Region of neurite overlap and observed synaptic puncta marked with white boxes. Gut; auto-fluorescence gut granules. Scale bars are 5 µM. Note that the roughly two-fold increase in synapse number from L1 to L4 in the hermaphroditic PHB>AVA connection and the male PHB>AVG is in line with an overall increase in total synapse numbers seen between neurons between L1 and adult stage as deduced by recent reconstruction of an L1 stage animal (M. Zhen, pers. comm.). d: Summary of synaptic connection differences between juvenile, pre-L4 animals and adults.
Figure 4
Figure 4. Autonomy and non-autonomy of sex-specific synapse pruning
Either the presynaptic cell (PHB) or the postsynaptic cells (AVA, AVG) were masculinized (by expression of FEM-3) and feminized (by expression of TRA-2IC), and the number of synaptic puncta were quantified in hermaphrodites and males, respectively. We performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test) with Bonferroni correction for multiple comparisons. ****P < 0.0001, ***P < 0.001, **P < 0.01. Magenta horizontal bars represent the median. Results are summarized in Extended Data Fig. 8a. Note that a general trend in the sex-reversal experiments is that the change of the sex of the presynaptic neuron (PHB) appears to have a stronger effect than changing the sex of either of the postsynaptic cells. Changing the sex of both postsynaptic cells simultaneously did not enhance the effects (Extended Data Fig. 8c).
Figure 5
Figure 5. Sexually dimorphic expression and function of dmd-5 and dmd-11
a: Sex-specific expression of dmd-5 and dmd-11 reporter genes in AVG. There are no additional dimorphisms in the retrovasicular ganglion, apparent differences are due to differences in z-planes incorporated into the final Z-stack projection. Masculinization of AVG derepresses dmd-5 and dmd-11 expression in hermaphrodites AVG. Quantified in Extended Data Figure 9a. b: Vulva location efficiency is affected in dmd-5 (n=11); dmd-11 (n=20) and dmd-5;dmd-11 double mutant males (n=29) compared with wild-type males (n=15). Expression of either dmd-5 (n=10) or dmd-11 (n=10) in AVG (using the inx-18 promoter) of double mutant males rescues behavior defects. c, d:dmd-5 and dmd-11 are required for maintenance of AVG synapses. Fluorescent micrographs (c) and quantification of synaptic puncta of PHB>AVG (d). Region of neurite overlap and observed synaptic puncta marked with white boxes. Quantification of DMD effect on LUA>AVG synapses can be found in Extended Data Fig. 9i. In panels b, d we performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test) with Bonferroni correction for multiple comparisons. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. Error bars in b are S.E.M. Magenta horizontal bars represent the median. Scale bars are 10 µM.

Comment in

  • Neurobiology: Wired for sex.
    Portman DS. Portman DS. Nature. 2016 May 12;533(7602):188-9. doi: 10.1038/nature17898. Epub 2016 May 4. Nature. 2016. PMID: 27144358 No abstract available.

Similar articles

See all similar articles

Cited by 29 articles

See all "Cited by" articles

References

    1. Sulston JE, Schierenberg E, White JG, Thomson JN. The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev Biol. 1983;100:64–119. - PubMed
    1. Sammut M, et al. Glia-derived neurons are required for sex-specific learning in C. elegans. Nature. 2015;526:385–390. - PMC - PubMed
    1. Jarrell TA, et al. The connectome of a decision-making neural network. Science. 2012;337:437–444. - PubMed
    1. Emmons SW. The development of sexual dimorphism: studies of the Caenorhabditis elegans male. Wiley Interdiscip Rev Dev Biol. 2014;3:239–262. - PMC - PubMed
    1. Portman DS. Genetic control of sex differences in C. elegans neurobiology and behavior. Adv Genet. 2007;59:1–37. - PubMed

References for Methods

    1. Brenner S. The genetics of Caenorhabditis elegans. Genetics. 1974;77:71–94. - PMC - PubMed
    1. Wenick AS, Hobert O. Genomic cis-Regulatory Architecture trans-Acting Regulators of a Single Interneuron-Specific Gene Battery in C. elegans. Dev Cell. 2004;6:757–770. - PubMed
    1. Hobert O. PCR fusion-based approach to create reporter gene constructs for expression analysis in transgenic C. elegans. Biotechniques. 2002;32:728–730. - PubMed
    1. Bond SR, Naus CC. RF-Cloning.org: an online tool for the design of restriction-free cloning projects. Nucleic Acids Res. 2012;40:W209–W213. - PMC - PubMed
    1. Jansen G, et al. The complete family of genes encoding G proteins of Caenorhabditis elegans. Nat Genet. 1999;21:414–419. - PubMed

Publication types

MeSH terms

Feedback