Isolation of a Hypomorphic skn-1 Allele That Does Not Require a Balancer for Maintenance

In Caenorhabditis elegans, the transcription factor SKN-1 has emerged as a central coordinator of stress responses and longevity, increasing the need for genetic tools to study its regulation and function. However, current loss-of-function alleles cause fully penetrant maternal effect embryonic lethality, and must be maintained with genetic balancers that require careful monitoring and labor intensive strategies to obtain large populations. In this study, we identified a strong, but viable skn-1 hypomorphic allele skn-1(zj15) from a genetic screen for suppressors of wdr-23, a direct regulator of the transcription factor. skn-1(zj15) is a point mutation in an intron that causes mis-splicing of a fraction of mRNA, and strongly reduces wildtype mRNA levels of the two long skn-1a/c variants. The skn-1(zj15) allele reduces detoxification gene expression and stress resistance to levels comparable to skn-1 RNAi, but, unlike RNAi, it is not restricted from some tissues. We also show that skn-1(zj15) is epistatic to canonical upstream regulators, demonstrating its utility for genetic analysis of skn-1 function and regulation in cases where large numbers of worms are needed, a balancer is problematic, diet is varied, or RNAi cannot be used.

skn-1 function by a degree similar to, or greater than RNAi, but is not restricted from some tissues. Importantly, a high enough fraction of homozygous skn-1(zj15) worms escape embryonic lethality so that they can be maintained without a balancer, and cultured as a large homogeneous population.
Mutagenesis, screening, whole genome sequencing, and mapping Roughly 6000 L4 to young adult QV160 worms were mutagenized in 50 mM ethyl methanesulfonate for 4 hr and allowed to recover and then lay F1 eggs on standard NA22 agar plates. Roughly 30,000 mutagenized F1 worms were divided among 15 10-cm NA22 plates, and allowed to lay a total of about 300,000 F2 worms. F2 worms were screened for Suppression of wdr-23 (Sow) by bulk sorting worms for loss of bright Pgst-4::GFP fluorescence in a COPAS BioSorter (Union Biometrica, Holliston, MA). Bulk-sorted F2 worms were then manually screened for viability, and low Pgst-4::GFP and Pgst-4::tdTomato fluorescence, and placed into 24-well plates. Sow mutants were also isolated from a selection screen for fast growth and reproduction by four to five rounds of seeding new plates, and allowing them to exhaust the bacteria food source. The largest worms from these plates were isolated and also screened for Pgst-4::GFP and Pgst-4::tdTomato fluorescence, and placed into 24 well plates. After ensuring true breeding in F3 progeny, strains were outcrossed to N2. These outcrossed lines were then crossed with Hawaiian mapping strain CB4856, and DNA was isolated from the progeny of recombinant F2s using a ReliaPrep gDNA Tissue Miniprep (Promega Corporation, Madison, WI) supplemented with extra RNAase. Genomic DNA was then labeled and sequenced by the University of Florida Interdisciplinary Center for Biotechnology Research in an Illumina MiSeq (San Diego, CA) according to manufacturer's recommendations. Raw sequence data were mapped to the N2 reference genome and Hawaiian and N2 SNP variant frequencies and EMS-induced mutations were identified with the CloudMap Workflow on the Galaxy server (Blankenberg et al. 2010;Giardine et al. 2005;Goecks et al. 2010;Doitsidou et al. 2010).
skn-1 transcript analysis skn-1 transcripts were amplified with Titanium Taq DNA polymerase (Clontech Laboratories, Mountain View, CA) from cDNA templates. Two exons flanking the sow(zj15) mutation were Sanger sequenced and analyzed with Geneious 6.0.6 software (Biomatters, New Zealand). Quantitative real-time RT-PCR was used to measure mRNA levels in L4 to young adult stage worms as described previously (Choe et al. 2009). Primers for quantitative PCR were: skn-1b forward, 59-CAAC AGGTGGATCAACACGG; reverse, 59-AGGCGTAGTTGGATGTT GGG and skn-1 a/c forward, 59-GGCAAATTTGACCGAGATGCA; reverse, 59-GAACAAAGTCTCTGGTTGAGCA. Note that the forward primer for skn-1a/c spans two exons of the wildtype cDNA, and does not match the sequence of cDNA from mis-spliced mRNA. Primers for quantitative PCR comparison of mis-spliced and normal skn-1a/c levels were: skn-1a/c normal forward 2, 59-GTTTATAATCAGGC AAATTTGACCG or skn-1(zj0015) mis-spliced forward, 59-AGGCA AATTGCATGTGACCG each paired with skn-1a/c reverse, 59-GAA CAAAGTCTCTGGTTGAGCA. Relative starting transcript levels of normal and mis-spliced skn-1a/c cDNA were calculated from Ct values and efficiency for each primer pair.

RNAi
RNAi was performed as described previously (Tang and Choe 2015). dsRNA producing bacteria were grown in Luria-Bertani broth containing selective antibiotic and then transferred to agar nematode growth medium (NGM) plates containing 3 mM isopropyl b-D-1thiogalactopyranoside (IPTG) or 0.2% b-lactose.

Growth and reproduction assays
Developmental stage was scored 48 hr after synchronized L1 worms were placed on food. For body length, worms at the same stage (young adult) were photographed, and body length was measured using Image J software (National Institutes of Health). Brood size assays were measured by counting the total number of eggs laid by each worm, and the total number of eggs hatched for 4 d, as described previously (Tang and Choe 2015).

Fluorescence assays
Pgst-4::GFP images were taken of adult worms with an Olympus BX60 (Center Valley, PA) microscope fitted with a Zeiss AxioCam MRm camera (Thornwood, NY). Fluorescence intensity was scored manually with a Zeiss Discovery V12 microscope.

Stress and longevity assays
Stress resistance assays were conducted on worms at L4 to young adult stage, when most of development is complete. Worms were exposed to 5 mM sodium arsenite, or 175 mM juglone. Death was recorded every 2-3 hr for up to 12 hr. For longevity assays, worms were transferred to plates containing 100 mg/ml 5-fluorodeoxyuridine (FUdR) at the L4 to young adult stage. Worms were considered dead if they did not display any movement in response to repeated prodding with a thin platinum wire.

Statistical analysis
Statistical significance was determined using a Student's t-test when two means were compared, and a one-way ANOVA with a Dunnet or Tukey's post-hoc test when three or more means were compared, a chi-square test for categorical data, and a log rank test when survival curves were compared. P values of , 0.05 were taken to indicate statistical significance, except for comparing more than two survival curves, in which Bonferroni adjustments were made to P values to account for repeated comparisons.

Reagent and data availability
All unique strains and reagents are available on request. Data are available at https://figshare.com/s/006267cd59845048237b.

RESULTS
A genetic screen identified a wdr-23 suppressor that mapped to an intron of skn-1 To screen for suppressors of wdr-23 (Sow), we conducted an EMS screen in wdr-23(tm1817) worms containing two transcriptional reporters for gst-4. Our primary screen was done for suppression of Pgst-4::GFP fluorescence, which is extremely bright in wdr-23 worms (Choe et al. 2009;Leung et al. 2013). This screen is difficult to perform manually because of the inherent difficulty in identifying rare worms with low fluorescence, and therefore a COPAS BioSorter was used. We screened roughly 300,000 F2s derived from 30,000 mutagenized F1s (60,000 genomes). The strain we screened also carried a single copy of Pgst-4::tdTomato inserted by a MOS transposon method (Frokjaer-Jensen et al. 2014); although not bright enough to use in the initial screen, we reasoned that this single copy reporter would allow us to eliminate mutations that only silenced repetitive arrays. After enriching for low GFP fluorescence, we manually screened worms for low tdTomato fluorescence. To isolate additional sow mutations, we also screened the same mutagenized lines for suppression of slow growth and reproduction by passing them through at least four generations on agar plates. wdr-23(tm1817) worms grow and reproduce extremely slowly (Tang and Choe 2015;Choe et al. 2009), and therefore suppressor mutations quickly dominated culture plates. Six true-breeding mutant lines were obtained from this selection screen and all but one also had low levels of the gst-4 reporter fluorescence.
All true-breeding mutant lines with low gst-4 reporter fluorescence levels were tested for detoxification, and skn-1 mRNA expression levels using real-time PCR. Two of these mutations [(zj15) and (zj21)] reduced skn-1 and target gene (gst-4, gst-10, and gst-30) mRNA levels by more than 50%, and complementation testing suggested that they were not caused by mutations to the same gene.
Both alleles were crossed with the Hawaiian mapping strain CB4856, and the genomic DNA of 59 (zj15) or 43 (zj21) recombinants was sequenced to an average coverage of 51X (zj15) or 41X (zj21) and mapped. sow-1(zj21) mapped to two regions, one corresponding to an intron mutation in skn-1, and the other to a broad (6-7 Mb) region of chromosome III, suggesting that at least two mutations are contributing to the phenotype; this allele was not characterized further. Alternatively, sow-1(zj15) mapped to a 2-Mb region of chromosome IV that includes an AT to GC mutation in the fourth intron of the longest skn-1 variant ( Figure 1A).
Direct Sanger sequencing of an RT-PCR product corresponding to this region indicates that two pools of skn-1a/c transcript are produced in sow-1(zj15) worms, some spliced normally, and some with an extra 5 bp from the mutated intron that causes a frame shift predicted to result in a truncated protein (251/533 amino acids for SKN-1c) that is missing all of the DNA binding domain ( Figure 1B). We next used quantitative PCR to determine the mRNA levels of skn-1 variants. The mRNA level of skn-1b was not changed by sow-1(zj15) (P = 0.5358), Figure 1 sow-1(zj15) maps to an intron of skn-1. (A) Schematic diagram of three skn-1 splice variant gene models; the mutated base is shown relative to exon 4 of skn-1a. (B) Direct RT-PCR product sequence results (top) and predicted proteins (bottom). sow-1(zj15) is an AT-GC mutation in an intron that is specific to skn-1a and c. Sequencing of cDNA reveals that a fraction of skn-1a/c transcripts have a splicing error that introduces an extra 5 bp (GCATG) from the beginning of this intron. The consequence of this spicing error is the addition of two amino acids and a stop codon that removes the entire c-terminus and DNA binding domain. (C) mRNA levels of skn-1b and skn-1a/c transcripts (mean 6 SE, n = 5 replicates of worms, ÃÃÃ P , 0.001 relative to wildtype). (D) Rescuing effects of skn-1 gDNA on Pgst-4::GFP expression during exposure to 2.8 mM acrylamide [n = 29-78 total worms, ÃÃÃ P , 0.001 relative to sow-1(zj15)]; given full rescue, we conclude that sow-1 is skn-1, and refer to the allele as skn-1(zj15) from this point forward.
but skn-1a/c mRNA was reduced 76% compared to wild type (P , 0.0001) ( Figure 1C). We also designed primers specific for each version of skn-1a/c cDNA (mis-spliced and wildtype). In three cDNA samples from sow-1(zj15), we found that the wildtype and mis-spliced variants are present at an average ratio of 1.00 6 0.12 to 0.65 6 0.04, respectively.
Some skn-1(zj15) worms escape embryonic lethality To our knowledge, all current loss-of-function skn-1 alleles are 100% penetrant maternal effect embryonic lethal. Alternatively, homozygote skn-1(zj15) worms appear superficially wildtype and reproduce and grow, indicating that at least some are viable. Larval developmental rate ( Figure 2A) and body length ( Figure 2B) were slightly decreased compared to wildtype N2 worms when skn-1(zj15) worms were measured. As shown in Figure 2C, skn-1(zj15) worms laid about half as many eggs as wildtype, and 42% of the eggs hatched. Therefore, each skn-1(zj15) worm produced an average of about 47 viable offspring. We next tested if transgenic expression of wildtype skn-1 genomic DNA could rescue the egg production and partial embryonic lethality phenotypes of skn-1(zj15); total eggs laid and number hatched were both rescued to nearly wildtype levels ( Figure 2C), suggesting that the skn-1 mutation is responsible for both phenotypes.

skn-1(zj15) is functionally comparable to skn-1 RNAi
To determine if skn-1(zj15) mutants are functionally comparable to skn-1 RNAi with respect to stress response, we measured the mRNA levels of core SKN-1 target genes (gst-4, gst-10, gst-30, gst-38, and gcs-1) under basal and stress-induced conditions. The basal mRNA levels of gst-4, gst-10, and gcs-1 were lower in skn-1(zj15) mutants to an extent similar to, or more than, with skn-1 RNAi ( Figure 3A). skn-1 RNAi in skn-1(zj15) worms did not further decrease the expression of any of the target genes. Exposure to 5 mM sodium arsenite for 1 hr induced expression of all the target genes in wild-type worms ( Figure 3B), and skn-1 RNAi or skn-1(zj15) largely suppressed their induction; skn-1 RNAi in skn-1(zj15) worms did not further decrease the expression of any target genes except for gcs-1 (P , 0.05).
tested if skn-1(zj15) could suppress these wdr-23 phenotypes. As expected, wdr-23 RNAi strongly increased survival of 5 mM sodium arsenite and 175 mM juglone ( Figure 5, C and D). The increased survival phenotypes were abolished in skn-1(zj15) mutants. Similarly, wdr-23 RNAi failed to increase lifespan in skn-1(zj15) mutants ( Figure  5E). Note that the survival trials in Figure 3, C, D, and E were run with those in Figure 5, C, B, and E, respectively, and that the control curves are the same.

DISCUSSION
In summary, we isolated a new skn-1 hypomorphic allele skn-1(zj15) that can be maintained as a homozygote without a genetic balancer. The effects of skn-1(zj15) on expression of SKN-1 target genes, stress resistance, and longevity were all similar to, or stronger than, skn-1 RNAi except that the effects of the allele were not limited by tissue type like RNAi (Figure 3 and Figure 4). Importantly, skn-1 RNAi did slightly enhance gst-4 reporter and survival phenotypes of skn-1(zj15) worms, consistent with residual functional SKN-1 being expressed. This is also reflected in the production of close to 50 viable offspring per homozygous hermaphrodite ( Figure 2C), compared to none from worms homozygous for currently available nonsense skn-1 alleles (WormBase 2015).
SKN-1 regulation and function are under intense investigation and epistasis is a useful tool for placing genes into ordered genetic pathways (Tullet et al. 2008;An et al. 2005;Li et al. 2011;Robida-Stubbs et al. 2012;Blackwell et al. 2015). The ability of skn-1(zj15) to suppress known upstream regulators ( Figure 5) demonstrates its utility as a tool for epistasis analysis without the technical problems inherent with nonsense mutation alleles that require balancers. We conclude that skn-1(zj 0015) is a strong hypomorphic allele with stress and longevity phenotypes as strong as, or stronger than, skn-1 RNAi. We propose that this allele could be used in cases where food sources are varied, constant dsRNA feeding is not appropriate, epistasis with an RNAi phenotype is needed, or when a large population is required. Potential users should, however, keep in mind that skn-1(zj15) is not a null allele when interpreting genetic interaction results as they would when using skn-1 RNAi.
The 76% reduction in wildtype skn-1a/c mRNA ( Figure 1) alone could cause the stress response and embryonic lethality phenotypes in skn-1(zj15) homozygotes. During outcrossing and mapping, skn-1(zj15) heterozygotes had strong Pgst-4::GFP induction on plates with acrylamide, consistent with the mutation being recessive for the stress response phenotype. However, we cannot rule out the possibility of truncated SKN-1 contributing in a dominant manner in skn-1(zj15) homozygotes, because the ratio of normal to misspliced SKN-1 protein is likely to be different than in heterozygotes.
Surprisingly, homozygous skn-1(zj15) worms have a total egg production defect that was not observed with a nonsense skn-1 allele or RNAi (Tang and Choe 2015). Rescue of this defect with wildtype skn-1 genomic DNA is consistent with this phenotype being caused by skn-1(zj15) and not a background mutation. It is unclear how skn-1(zj15) influences egg production, and care should be taken when interpreting reproductive phenotypes with this allele.

ACKNOWLEDGMENTS
C. elegans strains were provided by the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MN) supported by