Regulation of synaptic nlg-1/neuroligin abundance by the skn-1/Nrf stress response pathway protects against oxidative stress

Abstract

The Nrf family of transcription factors mediates adaptive responses to stress and longevity, but the identities of the crucial Nrf targets, and the tissues in which they function in multicellular organisms to promote survival, are not known. Here, we use whole transcriptome RNA sequencing to identify 810 genes whose expression is controlled by the SKN-1/Nrf2 negative regulator WDR-23 in the nervous system of Caenorhabditis elegans. Among the genes identified is the synaptic cell adhesion molecule nlg-1/neuroligin. We find that the synaptic abundance of NLG-1 protein increases following pharmacological treatments that generate oxidative stress or by the genetic activation of skn-1. Increasing nlg-1 dosage correlates with increased survival in response to oxidative stress, whereas genetic inactivation of nlg-1 reduces survival and impairs skn-1-mediated stress resistance. We identify a canonical SKN-1 binding site in the nlg-1 promoter that binds to SKN-1 in vitro and is necessary for SKN-1 and toxin-mediated increases in nlg-1 expression in vivo. Together, our results suggest that SKN-1 activation in the nervous system can confer protection to organisms in response to stress by directly regulating nlg-1/neuroligin expression.

Figure 1. SKN-1 expression and WDR-23 sub-cellular localization in motor neurons.

A) Schematic of bec-1 and skn-1 genetic loci. Locations of gfp tags used in this study indicated. B) Co-localization of nls-gfp driven by the 7.3 kb Pbec-1 promoter (vjEx254) with Punc17::rfp (nuIs321) in cholinergic motor neurons. Arrowheads indicate cholinergic motor neurons; arrow indicates GABAergic neuron. C–D) Left, Localization of WDR-23a::GFP (vjEx423; C) and WDR-23b::GFP (vjEx426; D) in ventral cord motor neurons expressed under the snb-1 promoter fragment. Neuronal cell bodies are identified by the unc-17 promoter driving rfp (nuIs321) for reference. Right, WDR-23a::GFP (vjIs26; C) and WDR-23b::GFP (vjEx7; D) co-localization with the synaptic vesicle associated protein mCherry::SNB-1 (vjEx339) in motor neuron axons using the unc-129 promoter fragment. E) WDR-23a::GFP (vjIs26) co-localization with the mitochondrial marker (INVOM::RFP; vjEx663) in motor neuron axons in wild type, unc-104/kinesin mutants (E′) and drp-1/Drp1 mutants (E″). Scale bars represent 10 µm.

Figure 2. WDR-23 regulates SKN-1 abundance in motor neurons.

A) Representative images of SKN-1a::GFP driven by the unc-17 promoter (vjIs45) in cholinergic motor neurons of wild type, left, and wdr-23 mutants, right. Cell bodies identified by presence of RFP expressed under the unc-17 promoter (nuIs321). B) Quantification of SKN-1a::GFP (vjIs45) in wild type (N = 28) and wdr-23 (N = 33) mutants. Cells were classified as low, medium or high expressing neurons based on average SKN-1a::GFP fluorescence per cell (see methods). Scale bar represents 10 µm.

Figure 3. Analysis of whole transcriptome RNA sequencing of wdr-23 mutants expressing neuronal WDR-23.

A) Correlation plot of RNAseq reads per kilobase per million reads (RPKM) for WDR-23 rescue compared to wdr-23 mutants, for 2,285 genes previously identified to be significantly different between wild type and wdr-23 mutants . Correlation plot includes 2,118 genes, as successful reads were not obtained for all genes in the rescuing line. WDR-23 rescue denotes wdr-23 mutants with expression of WDR-23a (using snb-1/synaptobrevin promoter, nuIs225). Genes that do not rescue (indicated in blue) have similar RPKM for wdr-23 mutants and the WDR-23 rescue, whereas genes that rescue (indicated in red) have statistically higher RPKM for wdr-23 mutants. B-C) Characterization of 810 genes significantly different in wdr-23 versus WDR-23 rescue (nuIs225). B) Right, Breakdown by GO term of genes with predicted functions identified by DAVID (492 genes). Left, Breakdown by GO term of rescued genes expressed specifically in the nervous system (43 genes). C) Distribution by tissue expression of genes with WormMart reported expression patterns (173 genes).

Figure 4. nlg-1 transcription is regulated by neuronal SKN-1.

A) Schematic of Pnlg-1::gfp reporters; vjIs47 and vjIs48 contain a 3.6 kb nlg-1 promoter driving gfp. Location of SKN-1 binding site and mutated site indicated. B) Top, Representative images of Pnlg-1::gfp (vjIs47) expression in the ventral cord of indicated strains. WDR-23a and WDR-23b rescues indicate isoform specific cDNA driven by the nlg-1 promoter (vjEx447 and vjEx436, respectively) and expressed in wdr-23 mutants. Bottom, Representative images of the extra-chromosomal transgene Pnlg-1(Δbs)::gfp (vjEx391) in the indicated strains. C) Quantification of Pnlg-1::gfp (vjIs47) and Pnlg-1(Δbs)::gfp (vjEx391) in indicated strains. Sample sizes indicated. Values normalized to wild type (vjIs47) or deleted binding site (vjEx391), respectively. D) Representative images of Pnlg-1::gfp (vjIs48) heterozygotes in indicated strains. skn-1 indicates zu67 loss of function allele, unless otherwise indicated. E–F) Normalized quantification of Pnlg-1::gfp (vjIs48) heterozygotes (E), Pnlg-1::gfp (vjIs48) homozygotes and Pnlg-1(Δbs)::gfp (vjEx756, F) in indicated strains. All nlg-1 reporters were imaged using the same microscope settings, and expression of reporters was similar between strains (total fluorescence vjIs48 in wild type animals: 3280.6±155.6; total fluorescence of vjEx756 in wild type animals: 2919.7±241.7). G) Electrophoretic mobility shift assay for binding of full-length SKN-1a to the consensus sequence at −396 bp. Lysate was either unprogrammed control or programmed to express SKN-1a. Competitor is 20, 100 and 200 fold molar excess of unlabeled probe. Scale bar represents 10 µm; error bars represent ±sem; ***p<0.001.

Figure 5. Synaptic NLG-1 levels are regulated by SKN-1.

A) Representative images of Pnlg-1::nlg-1-gfp (vjIs105) in the dorsal (top) and ventral (bottom) cords of the indicated strains. skn-1(gf) indicates lax188 gain-of-function allele. B) Quantification of punctal fluorescence (peak), cord fluorescence (cord) and interpunctal interval (IPI) of Pnlg-1::nlg-1-gfp (vjIs105) in dorsal (wild type n = 31, skn-1(gf) n = 29) and ventral cords (wild type n = 32, skn-1(gf) n = 26) of the indicated strains. Cord fluorescence could not be measured for vjIs105 in the dorsal cord because it was not above background levels. Scale bar represents 10 µm; error bars represent ±sem; ***p<0.001.

Figure 6. nlg-1 transcription is regulated by the SKN-1 pathway.

A–B) Representative images (A) and quantification (B) of animals expressing the Pnlg1::gfp (vjIs47) or Pnlg-1(Δbs)::gfp (vjEx756) reporters exposed to control, 120 µM juglone, 5.0 mM sodium arsenite or 35 µM thimerosal. L4 stage animals were exposed to drug overnight for 14 hours, followed by 2–4 hour recovery prior to imaging. Values of stressed animals were compared to non-stressed animals of the same genotype to determine significant changes in response to stress. Sample sizes indicated. C–D) Representative images (C) and quantification (D) of Pnlg-1::nlg-1-gfp (vjIs105) in the dorsal cord of animals exposed to control (n = 30) or 120 µM juglone (n = 29) treatments. E–F) Representative images (E) and quantification (F) of Pnlg-1::gfp (vjIs47) in indicated strains exposed to control or 120 uM juglone. Sample sizes indicated. Scale bar represents 10 µm; error bars represent ±sem; **p<0.01, ***p<0.001.

Figure 7. NLG-1 is protective to juglone.

A–B) Survival curves of indicated strains on 200 uM juglone. C) Survival curve of indicated strains on 400 uM juglone. D) Model for SKN-1-dependent transcription of nlg-1. Stress activated the SKN-1 pathway, which increases nlg-1 expression; increased NLG-1 activity promotes organism survival. Error bars represent ±sem. Log rank tests were used to identify significant differences between genotypes and reported in Table S3. Curves depict average values for four replicates of n = 40 per genotype.