Phenotypic tolerance for rDNA copy number variation within the natural range of C. elegans

Abstract

The genes for ribosomal RNA (rRNA) are encoded by ribosomal DNA (rDNA), whose structure is notable for being present in arrays of tens to thousands of tandemly repeated copies in eukaryotic genomes. The exact number of rDNA copies per genome is highly variable within a species, with differences between individuals measuring in potentially hundreds of copies and megabases of DNA. The extent to which natural variation in rDNA copy number impacts whole-organism phenotypes such as fitness and lifespan is poorly understood, in part due to difficulties in manipulating such large and repetitive tracts of DNA even in model organisms. Here, we used the natural resource of copy number variation in C. elegans wild isolates to generate new tools and investigated the phenotypic consequences of this variation. Specifically, we generated a panel of recombinant inbred lines (RILs) using a laboratory strain derivative with ~130 haploid rDNA copies and a wild isolate with ~417 haploid rDNA copies, one of the highest validated C. elegans rDNA copy number arrays. We find that rDNA copy number is stable in the RILs, rejecting prior hypotheses that predicted copy number instability and copy number reversion. To isolate effects of rDNA copy number on phenotype, we produced a series of near isogenic lines (NILs) with rDNA copy numbers representing the high and low end of the rDNA copy number spectrum in C. elegans wild isolates. We find no correlation between rDNA copy number and phenotypes of rRNA abundance, competitive fitness, early life fertility, lifespan, or global transcriptome under standard laboratory conditions. These findings demonstrate a remarkable ability of C. elegans to tolerate substantial variation in a locus critical to fundamental cell function. Our study provides strain resources for future investigations into the boundaries of this tolerance.

Fig 1. MY1 x SEA51 recombinant inbred line construction, genotyping, and rDNA copy number estimation.

A: Parental strains MY1 (hermaphrodite) and SEA51 (male) were crossed and F1s were self-fertilized to initiate production of 118 RILs. At the F2 generation, 60 GFP(+) (130-rDNA) and 60 GFP(-) (417-rDNA) worms were selected to self-reproduce for 10 generations. After 10 generations of single-worm selfing, worm strains were frozen and designated as RILs. The RILs were then propagated an additional 20 generations in bulk. B: Proportion of RILs with N2 genotype at each position of the C. elegans genome. Regions highlighted in red are significantly distorted from a 50:50 ratio with a Bonferroni corrected p-value of 0.05 or lower. The peak on chromosome I at approximately 2.3 Mb matches the location of the zeel-1/peel-1 toxin-antitoxin locus. C: rDNA copy number estimates from short read sequencing for RILs and propagated RILs linked to mIs13. The changes in copy number over the 20 generations of propagation were consistent with a normal distribution (Shapiro-Wilk, p = 0.17) and no lines rose above the 200-copy number level (see S1B Fig). Pink lines indicate copy number estimates of the included SEA51 parental strain controls. D: rDNA copy number estimates from short read sequencing for RILs and propagated RILs not linked to mIs13. The changes in copy number over the 20 generations of propagation were consistent with a normal distribution (Shapiro-Wilk, p = 0.08) and no lines fell below the 300-copy number level (see S1B Fig). Pink lines indicate copy number estimates of the included MY1 parental strain controls.

Fig 2. Near Isogenic Lines (NILs) with different rDNA copy numbers do not differ in rRNA expression.

A: The 45S rDNA array in C. elegans is located on the far right arm of chromosome I. The schematic depicts the chromosome I genotype for each NIL generated. Allele designations for the introgressed region of each NIL are indicated to the left. The relative amount of wild isolate DNA remaining linked to the rDNA array is indicated by the lighter colored regions, next to the strain name of the source wild isolate (MY16, JU775, MY1, or RC301). The copy number of rDNA units is noted in the darker color regions for each NIL. Wild type (WT) is indicated for reference, with no linked wild isolate DNA and approximately 100-130 rDNA copies in N2 and its derivatives. White chromosomal regions correspond to N2 genotype. B: pre-rRNA levels from the 45S transcript were quantified by RT-qPCR in NILs with minimal wild isolate DNA. Seven different probe pairs were used across the transcript template (see S3G Fig and S7 Table). rRNA levels were normalized to actin mRNA. No significant differences are present between any NILs and N2 (ANOVA with Tukey Honest Significant Difference test). See S3 Fig for further rRNA quantification.

Fig 3. rDNA copy number variation does not correlate with reduced competitive fitness or early life fertility.

A-B: Competitions were conducted between strains with high rDNA copy number (GFP(-); > 400 rDNA copies, alleles indicated at bottom) and marker strain SEA51 (GFP(+); 130 rDNA copies). A: Four independent replicates were conducted each of either competition between N2 and SEA51 or between the 417-rDNA NIL (allele catIR12) and SEA51, experiments propagated simultaneously. B: Four replicates each were conducted of competitions between SEA51 and either N2, the 417-rDNA NIL (allele catIR12), or the 420-rDNA NIL (allele catIR29), experiments propagated simultaneously. For all panels, bars represent the relative proportion of worms that are GFP(+) (green, SEA51) or GFP(-) (white) after ~10–11 generations. At least 1,000 worms were quantified for each bar (S14 File). C: Early life fertility was assessed for N2 and NILs with high rDNA copy number in three replicates, n = 20 individual worms per strain per replicate. Data were collected from 86 hours post-embryo of the worm’s life and thus represent progeny production through just the first day of reproductive adulthood. The data fail the Shapiro-Wilk normality test and are not normally distributed. Statistical tests represented in the figure are Pairwise Wilcoxon tests with Benjamini-Hochberg were performed to compare strains. * p < 0.1, **** p < 0.001.

Fig 4. Lifespan of C. elegans is robust to increases or decreases in rDNA copy number within the natural range.

Lifespan assays were conducted on N2 and NILs with various rDNA copy numbers. A: NILs with low rDNA copy number. B: NILs with high rDNA copy number. (n = 50) (S16 File) Lifespans for all five strains in A and B were performed simultaneously (N2 data are the same between the plots).

Fig 5. Few genes are expressed differentially between individual NILs and N2.

A: Principal component analysis (PCA) was conducted on RNA-seq datasets from synchronized Day 1 adults of four NILs and N2. The rlog transformed DEseq dataset was analyzed by PCA in the DEseq package. Datapoints are colored according to the rDNA copy number allele. B-E: Three biological replicates for each NIL and N2 were used to calculate differential expression of all genes using DESeq2. Horizontal pink line indicates an adjusted p-value cutoff of 0.05, vertical pink lines indicate log2 fold changes of 1.5. Genes whose expression differences are shared among strains or have particularly high significance of large log2 Fold Change are indicated with gene name.