Gene expression modulation by the linker of nucleoskeleton and cytoskeleton complex contributes to proteostasis

Abstract Cellular mechanisms that act in concert to maintain protein homeostasis (proteostasis) are vital for organismal functionality and survival. Nevertheless, subsets of aggregation‐prone proteins form toxic aggregates (proteotoxicity) that in some cases, underlie the development of neurodegenerative diseases. Proteotoxic aggregates are often deposited in the vicinity of the nucleus, a process that is cytoskeleton‐dependent. Accordingly, cytoskeletal dysfunction contributes to pathological hallmarks of various neurodegenerative diseases. Here, we asked whether the linker of nucleoskeleton and cytoskeleton (LINC) complex, which bridges these filaments across the nuclear envelope, is needed for the maintenance of proteostasis. Employing model nematodes, we discovered that knocking down LINC components impairs the ability of the worm to cope with proteotoxicity. Knocking down anc‐1, which encodes a key component of the LINC complex, modulates the expression of transcription factors and E3 ubiquitin ligases, thereby affecting the rates of protein ubiquitination and impairing proteasome‐mediated protein degradation. Our results establish a link between the LINC complex, protein degradation, and neurodegeneration‐associated proteotoxicity.

Animals were counted daily. The displayed results represent three independent repeats, except for anc-1, in which animals were counted in two independent repeats. Statistical significance was determined by the logrank test with a Bonferroni correction. ** p-value < 0.01, **** p-value < 0.0001. Error bars indicate SEM.
(B) Paralysis of animals that express Aβ3-42 in their body-wall muscles (strain CL2006). The worms were treated from hatching with RNAi towards one of the genes that encode LINC components or fed with control bacteria (EV). On day 1 of adulthood, the worms were transferred onto dcr-1 RNAi, to restore the expression of LINC components during adulthood. n=238-240 per condition, from two independent repeats.
Statistical significance was determined by the logrank test with a Bonferroni correction.
(C, D) Morphologies of anterior intestinal cell nuclei (C), and their representative images (D), in lmn-1p::emr-1::GFP expressing worms (strain YG002). The worms were treated with RNAi towards the indicated LINC components. lmn-1 RNAi serves as a positive control for defective nuclear morphology. The morphologies were categorized into three classes, as detailed in the main text. Briefly, the nuclear morphology classes are evenly distributed GFP (class I, white arrows), GFP puncta and/or convoluted nuclear shape (class II, yellow arrows), and abnormal nuclear shapes (class III, blue arrows). For each condition, n=60 intestinal nuclei were counted from the anterior intestines of ten worms. Scale bar, 10 µm.

Treatment (RNAi)
Expression log 2 (fold-change) relative to EV a n c -1 a n c -1  (B) Quantitative real-time PCR (qPCR) measuring anc-1 transcript levels in day 1 adult CF512 animals that were treated with one of two RNAi constructs that target anc-1; One from the C. elegans ORF-RNAi feeding library (Vidal) and the second is a construct we generated against the 3'UTR region. Statistical significance was determined by one-way ANOVA with Dunnett's multiple comparisons test. ** p-value < 0.01, **** p-value < 0.0001. Error bars indicate SEM.
(C) Mean paralysis of wild type animals (WT, strain N2), and Aβ worms (strain CL2006  (E-G) Western blots of Aβ species in homogenates of day 3 adult Aβ worms that were either left untreated or exposed to Aβ or unc-84 RNAi (E, representative). Mean total Aβ signals (F) and mean high-MW Aβ aggregates (G       Enrichment (upregulation, NES > 0) or Depletion (downregulation, NES < 0) of GO biological processes based on the differential gene expression measured in wild type (strain N2) and polyQ35-YFP (strain AM140) animals treated with anc-1 RNAi.
Adjusted p-value threshold set at 0.05. "Pathway" is the GO annotation gene-set; "Gene ranks" are the positions of the genes that make up the corresponding Pathway relative to all the genes measured by RNA-seq, which are sorted by their log2(fold-change) from highest to lowest. "NES" are normalized enrichment scores; "pval" is the p-value prior to multiple comparisons correction; "padj" is the FDR-adjusted p-values.    (B) Expression fold-change of genes that code Skp1 homologs in day 6 adult Aβ worms, as measured by qPCR, shows that anc-1 knockdown regulates the expression of SCF complex ubiquitin ligase components. Three independent repeats.

s k r-1 4 s k r-1 5 s k r-1 3 s k r-2 1 s k r-1 6 s k r-1 2 c u l-1 s k r-1 9 s k p t-1 s e l-1 0 s k r-1 s k r-2 0 s k r-2 s k r-1 8 li n -2 3 s k r-1 7 s k r-3 c u l-6 s k r-4 s k r-5 s li -1 s ia h -1 g e i-1 7 s e l-1 1 p d r-1 p rp -1 9 c y n -4 c h n -
(C) Mean paralysis of Aβ worms that were treated from hatching with mixes of RNAi towards anc-1, cul-1, and EV bacteria. n=119-124 per condition from one experiment shows that anc-1 RNAi leads to a small non-significant increase in Aβ proteotoxicity when cul-1 is knocked down. Statistical significance was determined by the logrank test with a Bonferroni correction. Error bars indicate SEM.
(D) Expression fold-change of Cullin coding genes in day 6 adult polyQ35-YFP worms, as measured by qPCR. Three independent repeats show that the expression of cul-1 and cul-6 is not regulated by T-box transcription factors.
(E, F) In vitro caspase-like proteasome activity of day 6 adult wild type (E) and polyQ35-YFP (F) worm homogenates, and the mean slopes of three independent repeats, show that the knockdown of anc-1 impairs this activity.
(G, H) In vitro trypsin -like proteasome activity of day 6 adult wild type (G) and polyQ35-YFP (H) worm homogenates, and the mean slopes of three independent repeats, show no significant difference between untreated worms and their counterparts that were fed with anc-1 RNAi bacteria.
Supplemental Tables   Table S1. Related to Figure 3. Enrichment of transcription factor binding motifs that are regulated by anc-1 RNAi in both wild type and polyQ35-YFP animals.

Supplemental Movies
Movie S1.

Caenorhabditis elegans
Standard methods were used for the maintenance and manipulation of C. elegans (Stiernagle, 2006).

RNA interference (RNAi)
NG plates supplemented with 100 µg/ml ampicillin were seeded with bacterial cultures. To induce RNAi, the seeded plates were treated with 120 µl 100 mM Isopropyl-

Paralysis assay
Synchronized CL2006 eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. On the first day of adulthood, 120 randomly picked animals per treatment were transferred onto 60 mm NG-ampicillin plates seeded with bacteria, 24 animals per plate. The worms were assessed daily for paralysis by gently tapping their noses with a platinum wire. If worms did not move, they were further incentivized by placing a hot platinum wire in their proximity. Living worms that failed to move their central body regions were scored as paralyzed and removed from the plates. To avoid scoring old animals as paralyzed, paralysis assays were terminated at day 12 of adulthood.

Nuclear morphology and chromatin condensation
For each treatment, synchronized YG002 eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. Starting at day 1 of adulthood, 50 worms per condition were picked daily to avoid progeny. On the indicated days, the worms were washed twice with M9 to remove bacteria followed by two washes with water, and incubated on ice for ten minutes. Samples were fixed with 4% paraformaldehyde in MRWB buffer (80 mM KCl, 20 mM NaCl, 10 mM Na2EGTA, 5 mM spermidine, 25% methanol). The worms were frozen in -80C, thawed, and incubated for 30 minutes on ice with occasional agitation. The worms were washed three times in PBS, and 10-20 animals were mounted per condition in 5% (w/v) n-propyl gallate (Sigma-Aldrich, P3130), 100 mM Tris-HCl pH 9.0, 70% (v/v) glycerol, and 20 µg/ml 4',6-Diamidino-2-phenylindole (DAPI), for fluorescence microscopy imaging. DIC and fluorescence images were acquired using a Zeiss Axio Observer.Z1 inverted microscope equipped with an AxioCam HRm Zeiss camera, with optical sectioning using using Zeiss ApoTome.2, Plan-Apochromat 63x/1.40 oil immersion objective, and the Zeiss ZEN 2 (blue edition) software. For nuclear morphology, we imaged the anterior intestines across the worm diameter, in 2 µm slices. The nuclei were categorized into classes, based on nuclear features that were set by Haithcock and colleagues, as detailed in the main text (Haithcock et al., 2005). For nuclear morphology, we imaged anterior intestinal nuclei in 200 nm slices. Chromatin condensation index was calculated using a MATLAB tool designed to measure chromatin in vivo, developed by Sosnik and colleagues (Sosnik, Vieira, Webster, Siegfried, & McCusker, 2017). In brief, Chromatin condensation index is defined as the number of edges detected using the Sobel edge detection method within a nuclear cross-section relative to the nuclear area. For more details, see work by Irianto and colleagues (Irianto, Lee, & Knight, 2014).

Lifespan assay
Synchronized CF512 eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. CF512 animals are heat-sensitive sterile; to avoid progeny, the eggs were placed at 25 °C. On the first day of adulthood the worms were moved to 20 °C, and 120 randomly picked animals per treatment were transferred onto 60 mm NG-ampicillin plates seeded with bacteria, 12 animals per plate. Worms that failed to move when tapped twice gently with a platinum wire and placing a hot platinum wire in their proximity were scored as dead. Survival rates were recorded every 2-3 days until the last surviving worm perished.

Thrashing assay
Synchronized eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. On each time-point, 20 randomly picked animals were sequentially placed in 5 µl M9 buffer drop on top of a microscope slide and allowed thirty seconds of recovery. Afterwards, the number of body bends of each worm was counted for the duration of thirty seconds. One body bend is the formation of an angle between the head and tail of the worm in a pre-selected lateral movement direction.

Native Agarose Gel Electrophoresis (NAGE)
We performed the assay as described by its developers (Holmberg & Nollen, 2013 Afterwards, the buffer volume was minimized, and the worms were flash frozen the worms in liquid nitrogen. After thawing the samples, a small amount of 0.5 mm zirconium oxide beads (Next Advance, ZrOB05) was added to the collected worms. The worms were homogenized at 4 °C using a Bullet Blender® (Next Advance) set to speed nine -three times for ten seconds each. Worm debris was removed by centrifuging at 780 x g for three minutes at 4 °C and transferring away the soluble upper post-debris fraction.
The protein content of the post-debris fraction was quantified using the BCA Protein

Counting polyglutamine foci
Synchronized eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. On each time-point, worms were picked onto NGampicillin free from bacteria, immobilized in 20 mM sodium azide diluted in M9 buffer, and their entire bodies were individually imaged in a Nikon AZ100 microscope equipped with a Nikon DS-Fi2 camera, using the software NIS-Elements BR (version 4.13). The fluorescent polyQ35-YFP foci were quantified using the software worMachine (Hakim et al., 2018).

SDS-PAGE and western blot
For each condition, 10,000 to 15,000 synchronized eggs were placed on NGampicillin plates seeded with HT115 bacteria and supplemented with IPTG. The worms were washed daily with M9 to discard of progeny. On the indicated ages, the worms were washed with M9 to remove bacteria, and resuspended twice in PBS supplemented with 1:1000 Protease Inhibitor Cocktail Set III (Merck / Calbiochem, 539134).
When testing the induction of the HSR, CL2070 worms were exposed heat at 33 °C for 3 h, and were then collected as described. When testing the induction of the UPR ER , SJ4005 worms were exposed to 10 µg/ml Tunicamycin at 20 °C for 3 h, and were then collected as described. When testing the induction of the UPR mt , SJ4100 worms were treated with cco-1 RNAi from hatching, and were collected as described.
Afterwards, the buffer volume was minimized, and the worms were flash frozen in liquid nitrogen. After thawing the samples, a small amount of 0.5 mm zirconium oxide beads (Next Advance, ZrOB05) was added to the collected worms. The worms were homogenized at 4 °C using a Bullet Blender® (Next Advance) set to speed eight, for three minutes, and then set to speed nine and for an additional two minutes. The homogenates were centrifuged at 780 x g for three minutes at 4 °C and the worm debris

Immunofluorescence
For each treatment, synchronized eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. Starting at day 1 of adulthood, 50 worms per condition were picked daily to avoid progeny. On the indicated days, the worms were washed twice with M9 to remove bacteria followed by two washes with water, and incubated on ice for ten minutes. Samples were fixed with 4% paraformaldehyde in MRWB buffer (80 mM KCl, 20 mM NaCl, 10 mM Na2EGTA, 5 mM spermidine, 25% methanol). The worms were frozen in -80C, thawed, and incubated for 30 minutes on ice with occasional agitation. The worms were washed twice in Tris-Triton buffer (100 mM Tris-HCl pH 7.4, 1% Triton X-100, 1 mM EDTA), incubated for 2 h at 37 °C in Tris-Triton buffer supplemented with 1% β-Mercaptoethanol, with mild agitation, washed in BO3 buffer (25 mM H3BO3, 12.5 mM NaOH), and incubated for 15 minutes in BO3 buffer supplemented with 10 mM DTT, with mild agitation. The worms were then washed once in BO3 buffer, and incubated in BO3 buffer supplemented with 0.3% H2O2 for 15 minutes at room temperature, with mild agitation. After washing once with BO3 buffer, and for 15 minutes at room temperature with washing buffer (1X PBS, 0.1% BSA, 0.5% Triton X-100, 1 mM EDTA), the worms were incubated for one hour at room temperature in blocking buffer (1X PBS, 1% BSA, 0.5% Triton X-100, 1 mM EDTA) and stained overnight at 4 °C with the primary antibody in blocking buffer. The worms were washed three times in washing buffer, and stained for two hours at room temperature with fluorophore conjugated secondary antibodies followed by ten washes in washing buffer. 10-20 animals were mounted per condition in 5% (w/v) n-propyl gallate

RNA isolation
Total RNA was isolated using NucleoSpin® RNA kit (MACHEREY-NAGEL, 740955). For each time-point, 6,000 synchronized eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. The worms were washed daily with M9 to discard of progeny. On each time-point, we washed the worms with M9 to remove bacteria, left a minimal buffer volume, and froze them immediately in liquid nitrogen. The samples were thawed on ice, resuspended in 350 µl buffer RA1 (kit reagent) supplemental with 5.5 µl 1M DTT, and added a small amount of 0.5 mm zirconium oxide beads (Next Advance, ZrOB05). The worms were homogenized at 4 °C using a Bullet Blender® (Next Advance) set to speed eight, for three minutes, and then set to speed nine and for an additional two minutes. Homogenates were transferred to microcentrifuge tubes and centrifuged at 14,000 x g for 5 minutes. The supernatant was transferred to NucleoSpin® Filter (NucleoSpin® RNA kit reagent), and the next steps followed the manufacturer's standard protocol. Following RNA purification, the samples were stored at -80 °C.

Quantitative real-time PCR (qPCR)
Following RNA isolation, the RNA integrity was assessed by loading and running

RNA sequencing (RNA-Seq)
Following RNA isolation, the RNA integrity and quantity was assessed using the Agilent 2200 TapeStation System. The RNA purity was assessed by loading 1.5 µl of each sample onto a NanoDrop™ 2000c Spectrophotometer and reading the A260 nm emission, as well as the A260/A280 and A260/A230 ratios. Isolated RNA samples were prepared and sequenced using the Illumina HiSeq 2500 by the Technion Genome Center (Technion, Haifa, Israel) according to the CEL-Seq2 protocol, as described by its developers (Hashimshony et al., 2016). During CEL-Seq sample preparation, each sample is marked with a primer containing a unique barcode. Each initial RNA sample was barcoded using a pool of three different CEL-Seq primers. Using several primers to create technical replicates for each sample reduces the possibility of technical variability caused by differences in barcode efficiency. Processing was done with a modified CEL-Seq2 pipeline (yanailab.github.io/celseq2), based on the one described by its developers.
After adapter and quality trimming over 90% of the reads remained. Mapping to the C. elegans genome, Ensembl: WBcel235, release 36 (June, 2017), joint with the ERCC sequences (CEL-Seq2 control Spike-In), was done using Tophat2 version 2.1.1, ((Kim et al., 2013), ccb.jhu.edu/software/tophat) with parameters: -N 2, --read-edit-dist 2, -G protein_coding_ERCC.gtf. The set parameters mean up to 2 mismatches, and up to an edit distance of 2, using annotations of protein coding genes (from Ensembl WBcel235 release 36), and the ERCC Spike-In. A mapping percentage of over 94% was reached.
Gene counting was done using HTseq-count ( (Anders, Pyl, & Huber, 2015), htseq.readthedocs.io) version 0.6.1, with parameters: gff_file=protein_coding_ERCC.gtf, umi=false. The set parameters mean that the counting was done using the same annotations as in the mapping step (protein coding and ERCC Spike-In) and ignoring UMIs. This was done since UMIs are too short to capture the full repertoire of transcripts in the samples and using them would cause a flattening of the signal for genes with a higher expression. Normalization and differential expression analysis done using the R/Bioconductor R Core Team, 2018) (Love, Huber, & Anders, 2014). Genes with low expression (mean counts < 2) were discarded from the analysis. The similarity between replicates, and the relations between the samples, were evaluated by hierarchically clustering the Euclidean or the Pearson correlation distance measures of the samples and visualizing in heatmaps. Principal Component Analysis (PCA) plots were also used to ascertain the similarity between samples. The RNA sequencing data files that were generated in this study are available in the NCBI Gene Expression Omnibus (GEO) under the accession number GEO: GSE126585.

RNA-Seq results presentation and pathway analysis
Euler diagrams were generated using the R package eulerr version 5.1.0. GO Biological Processes Gene Set Enrichment Analysis (GSEA) was performed using the R/Bioconductor package fgsea version 1.8.0 (Sergushichev, 2016). KEGG GSEA was performed using the R/Bioconductor packages clusterProfiler version 3.8.1 (Yu, Wang, Han, & He, 2012) and org.Ce.eg.db version 3.6.0. RNA-Seq measured genes were ranked by their log2 fold-change between control and treatment and were used to perform a GSEA in medium sized gene-sets (GO: 15-500 members, KEGG: greater than 15 members), using 10,000 permutations. Protein Class overrepresentation test was done using the online PANTHER classification system interface, set to perform a Binomial test followed by a Bonferroni correction (Released 2017-12-05), using the PANTHER database version 13.1 (2018-02-03), and the PANTHER Protein Class annotation data set (Thomas et al., 2003). The gene list that was selected to be tested for overrepresented protein classes, included all the genes whose expression fold-change passed the statistical significance criteria (p-value < 0.05). The reference/background list was composed of all the genes (20,057) in the PANTHER database for C. elegans.

Binding motif enrichment analysis
The presence of T-box transcription factor binding motifs in the promoter regions of genome-wide protein coding genes was determined using the R/Bioconductor package Biostrings version 2.50.0. The enrichment of T-box transcription factor binding motifs was determined using the R/Bioconductor package PWMEnrich version 4.16.0. The promoter regions of C. elegans genes were defined as the sequences between 500 bp upstream-, and 100 bp downstream, to the TSS, based on previous findings (Narasimhan et al., 2015;Niu et al., 2011). The promoter sequences were obtained using the R/Bioconductor package biomaRt version 2.38.0 (Durinck, Spellman, Birney, & Huber, 2009). The T-box transcription factors binding motifs were obtained from the CIS-BP database version 1.02 (Weirauch et al., 2014). The Binding motifs heatmap was composed using the R package pheatmap version 1.0.10. The genes that are presented in the heatmap possess at least one T-box transcription factor binding motifs, whose presence was scored with a p-value < 0.01. The "Function" of selected genes is based on our interpretation of their Gene Ontology annotations. The genes were hierarchically clustered using the 'hclust' function with a binary distance measure set in the function 'dist'.

Proteasome activity assay
For each treatment, 5,000 synchronized eggs were placed on NG-ampicillin plates seeded with HT115 bacteria and supplemented with IPTG. The worms were washed daily with M9 to discard of progeny. On day six of adulthood, the worms were washed with M9 to remove bacteria, and resuspended them twice in ice-cold PBS. Afterwards, a small amount of 0.5 mm zirconium oxide beads (Next Advance, ZrOB05) was added to the collected worms. The worms were homogenized at 4 °C using a Bullet Blender® (Next Advance) set to speed nine, three times for ten seconds each. Worm debris was removed by centrifuging at 780 x g for three minutes at 4 °C and transferring away the soluble upper post-debris fraction. The protein content of the post-debris fraction was quantified using the BCA Protein Assay Kit (Pierce™, 23227) according to its suggested protocol.
Triplicate wells of an opaque 96-well plate (Thermo Scientific, Nunc F96 MicroWell, 137101) were each loaded with 10 µg of protein diluted in a 26S proteasome activity assay buffer (50 mM Tris-HCl pH 7.5, 40 mM potassium chloride, 5 mM magnesium chloride, 1 mM DTT, 0.5 mM ATP, 0.1% BSA). Each well had a total volume of 90 µl. (Sigma-Aldrich, M7449) and incubated for 5 minutes at room temperature prior to measurement as a control. Results at time 0 of each experiment were defined as baseline.

Statistical Analysis
Statistical details of experiments that include the statistical tests used, significance, error bars, repeats, exact values of n, and what they represent, can be found in the corresponding figure legends. Data are presented as mean ± SEM. Statistical analyses were performed using GraphPad Prism version 7.01 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com, and R version 3.5.1 (R Core Team, 2018). p-values are indicated by * for p < 0.05, * * for p < 0.01, * * * for p < 0.001, **** for p < 0.0001, and "ns" for p > 0.05. "Statistically significant" was defined as pvalue < 0.05.