Bergerac strains of Caenorhabditis elegans revisited: expansion of Tc1 elements imposes a significant genomic and fitness cost

Abstract

The DNA transposon Tc1 was the first transposable element to be characterized in Caenorhabditis elegans and to date, remains the best-studied transposable element in Caenorhabditis worms. While Tc1 copy-number is regulated at approximately 30 copies in the laboratory Bristol N2 and the vast majority of C. elegans strains, the Bergerac strain and its derivatives have experienced a marked Tc1 proliferation. Given the historical importance of the Bergerac strain in the development of the C. elegans model, we implemented a modern genomic analysis of three Bergerac strains (CB4851, RW6999, and RW7000) in conjunction with multiple phenotypic assays to better elucidate the (1) genomic distribution of Tc1 and (2) phenotypic consequences of transposable element deregulation for the host organism. The median estimates of Tc1 copy-number in the Bergerac strains ranged from 451 to 748, which is both (1) greater than previously estimated and (2) likely to be an underestimate of the actual copy-numbers since coverage-based estimates and digital droplet polymerase chain reaction results both suggest higher Tc1 numbers. All three Bergerac strains had significantly reduced trait means compared with the N2 control for each of four fitness-related traits, with specific traits displaying significant differences between Bergerac strains. Tc1 proliferation was genome-wide, specific to Tc1, and particularly high on chromosomes V and X. There were fewer Tc1 insertions in highly expressed chromatin environments than expected by chance. Furthermore, Tc1 integration motifs were also less frequent in exon than noncoding sequences. The source of the proliferation of Tc1 in the Bergerac strains is specific to Tc1 and independent of other transposable elements. The Bergerac strains contain none of the alleles that have previously been found to derepress transposable element activity in C. elegans. However, the Bergerac strains had several Tc1 insertions near or within highly germline-transcribed genes which could account for the recent germline proliferation.

Keywords: Caenorhabditis elegans; Bergerac; chromatin; copy-number variation; fitness; recombination; transposable element; transposon; whole-genome sequencing.

Fig. 1.

Proliferation of the Tc1 transposon in the Bergerac strains of C. elegans. a) Counts of the Tc1 transposon in our laboratory isolate of Bristol N2 and 3 Bergerac strains, CB4851, RW6999, and RW7000 using various computational and molecular methods. Computational methods included were ngs_te_mapper, RelocaTE, TEMP2, RetroSeq, TEFLoN, and coverage. Copy-number estimation via ddPCR was the sole molecular method. b) Counts of other/non-Tc1 transposable elements are displayed for N2, CB4851, RW6999, and RW7000. Tc2 was the only transposable element with a marked increase in the 3 Bergerac lines compared with N2, while Tc4 and Tc5 were found to be marginally higher in CB4851.

Fig. 2.

Comparison of Tc1 insertions predicted by McClintock component methods. The positional agreement of the McClintock component methods was assessed by assigning Tc1 calls to windows (1,000- and 100,000-bp windows for nonreference and reference calls, respectively) followed by a comparison of the calls made by each component method. The Venn diagrams compare the agreement of Tc1 calls for a) N2, b) CB4851, c) RW6999, and d) RW7000. Computational methods included were ngs_te_mapper2 (blue), RelocaTE (red), TEMP2 (green), RetroSeq (yellow), and TEFLoN (brown). The vast majority of Tc1 insertions were called by multiple methods.

Fig. 3.

Significant fitness reduction in Bergerac strains relative to N2. Means for 4 fitness-related traits, namely a) productivity, b) survivorship to adulthood, c) longevity, and d) developmental time observed in the N2 control, and the 3 Bergerac strains CB4851, RW6999, and RW7000. e) Decline in mean relative fitness of the 3 Bergerac strains relative to the N2 strain. For simplicity, the mean relative fitness value for each of the 4 traits in the Bristol N2 control was scaled to a value of 1 (not displayed). For all panels, error bars represent 1 SE. Significance was determined via ANOVA and is displayed as asterisks where P ≤ 0.05*/0.01**/0.001***. Exact P-values for the pairwise strain comparisons using Tukey–Kramer HSD can be found in Supplementary Tables 2–5.

Fig. 4.

Physical size reduction and aberrant behavior of the Bergerac strains. Mean values for 4 traits measured in the motility analyses, namely a) speed, b) body length, c) body area, and direction change d) in the N2 control, and the 3 Bergerac strains CB4851, RW6999, and RW7000. e) Mean relative changes between the 3 Bergerac strains and the N2 control. For simplicity, the mean relative value for each of the 4 traits in the Bristol N2 control was scaled to a value of 1 (not displayed). The stars on brackets summarize P-values for Tukey–Kramer HSD comparisons between Bergerac strains. Stars on top of bars reflect P-values for Tukey–Kramer HSD comparisons to N2. Exact P-values for the pairwise strain comparisons using Tukey–Kramer HSD can be found in Supplementary Tables 6–9.

Fig. 5.

The Tc1 insertion motif in the Bergerac strains. The vertical axis indicates the proportion of each base at 4 upstream and downstream positions from the insertion site. The motif for Tc1 insertion sites was based on 1,659 insertions in this study and matched that presented in previous literature. Tc1 invariably inserts between a T(−1) and an A(+1) with additional conserved A/T bases at positions −4 and +4.

Fig. 6.

The chromosomal distribution of Tc1 elements in the Bergerac strains relative to N2. a) Genomic map of Tc1 insertions within each strain. Nonreference Tc1 insertions are displayed in green, while reference Tc1 sites in Bristol N2 are shown in orange. b) The proportion of Tc1 elements by chromosome. Tc1 elements were significantly overrepresented on the X chromosome in CB4851 (χ² = 51.87, P = 5.92 × 10⁻¹³), and chromosomes V and X in RW6999 (χ² = 49.06, P = 2.16 × 10⁻⁹) and RW7000 (χ² = 50.88, P = 9.15 × 10⁻¹⁰).

Fig. 7.

The proportions of Tc1 elements located in arms, cores, and tips. Arms are associated with high recombination rates whereas cores and tips have low recombination rates. There were no significant differences in the distribution of Tc1 between these 3 domains (N2: χ² = 2.66, P = 0.21; CB4851: χ² = 0.47, P = 0.49; RW6999: χ² = 4.39, P = 0.11; RW7000: χ² = 4.81, P = 0.11).

Fig. 8.

The proportions of Tc1 insertions across different chromatin domains. a) Histone modification H3K27me3 is associated with repressed chromatin and lowly expressed genes. All of the Tc1 positions in N2 were found in this domain. Tc1 insertions were more abundant in regions associated with H3K27me3 than expected by chance in CB4851 (χ² = 7.44, P = 0.013) and RW6999 (χ² = 8.82, P = 0.0089), but not RW7000 (χ² = 2.82, P = 0.09). b) Histone modifications H3K9me1/2/3 are associated with repetitive DNA. All of the Tc1 positions in N2 were found in these domains. Tc1 insertions were overrepresented in H3K9me1/2/3 domains in strains RW6999 (χ² = 5.06, P = 0.049) and RW7000 (χ² = 7.06, P = 0.024), but not in CB4851 (χ² = 7.06, P = 0.44). c) Histone modifications H3K4me1/2/3, H3K27ac, H4K8ac, and H4K16ac are associated with promoters of highly expressed genes. In N2, Tc1 insertions were entirely absent from these domains. Tc1 insertions were underrepresented in domains associated with H3K4me1/2/3, H3K27ac, H4K8ac, H4K16ac, and promoters of in all 3 Bergerac strains: CB4851 (χ² = 10.91, P = 0.00287), RW6999 (χ² = 5.96, P = 0.0184), and RW7000 (χ² = 6.79, P = 0.0184). d) Histone modifications H3K36me3 and H3K79me1/2/3 are associated with highly expressed genes. In N2, Tc1 insertions were entirely absent from these domains. In all 3 Bergerac strains, Tc1 insertions were underrepresented in these domains (CB4851: χ² = 48.33, P = 3.60 × 10⁻¹²; RW6999: χ² = 75.28, P = 8.17 × 10⁻¹⁸; RW7000: χ² = 82.71, P = 2.85 × 10⁻¹⁹).

Fig. 9.

The proportion of Tc1 elements in exons, introns, and intergenic regions of Bergerac strains in comparison to Bristol N2. In N2, the Tc1 insertions were only found in introns and intergenic regions. The proportion of Tc1 insertions in exons, introns, and intergenic regions was significantly different between Bristol N2 and the Bergerac strains (G = 54.26, P = 6.54 × 10⁻¹⁰, df = 6), with the Bergerac strains showing a far larger and fewer proportion of Tc1 insertions in exonic regions and intronic regions, respectively. There was no significant difference among the 3 Bergerac strains in the proportion of Tc1 insertions in exons, introns, and intergenic regions (G = 4.28, P = 0.37, df = 4).