Transmission dynamics of heritable silencing induced by double-stranded RNA in Caenorhabditis elegans

Abstract

Heritable silencing effects are gene suppression phenomena that can persist for generations after induction. In the majority of RNAi experiments conducted in Caenorhabditis elegans, the silencing response results in a hypomorphic phenotype where the effects recede after the F1 generation. F2 and subsequent generations revert to the original phenotype. Specific examples of transgenerational RNAi in which effects persist to the F2 generation and beyond have been described. In this study, we describe a systematic pedigree-based analysis of heritable silencing processes resulting from initiation of interference targeted at the C. elegans oocyte maturation factor oma-1. Heritable silencing of oma-1 is a dose-dependent process where the inheritance of the silencing factor is unequally distributed among the population. Heritability is not constant over generational time, with silenced populations appearing to undergo a bottleneck three to four generations following microinjection of RNA. Transmission of silencing through these generations can be through either maternal or paternal gamete lines and is surprisingly more effective through the male gametic line. Genetic linkage tests reveal that silencing in the early generations is transmitted independently of the original targeted locus, in a manner indicative of a diffusible epigenetic element.

Figure 1.—

Genomic region showing the oma-1 locus with linked morphological markers and the dsRNA triggers. We used two morphological markers, unc-24 and dpy-20, to follow the oma-1 locus during crosses. Locations of the triggers used for injections are shown. Long dsRNA triggers are shown aligned against the physical map of the oma-1 locus. Together, fragments A1 and A2 span the coding region of the oma-1 gene. A1 includes exons 1–4 and is 721 bases long. A2 includes exons 5 and 6 and is 685 bases long.

Figure 2.—

Pedigree selection scheme for determining of persistence of heritable silencing. Young adult hermaphrodites of the strain oma-1(zu405), dpy-20(e1282ts)IV; him-5(e1467)V were injected with dsRNA trigger and allowed to recover at room temperature. Individual I₀ animals were plated onto petri dishes containing fresh lawns of OP-50 bacteria and grown at 23° for 3 days, when they were scored for viable progeny. Criteria for selecting animals to pedigree were (1) groups where all observed siblings had viable progeny and (2) from the sibling group, the individuals with largest brood sizes. Both these criteria were our indicators of a strong silencing response. We designated I₀'s (labeled “A1,” “B1,” and “C1”) and individually plated 14 F₁ animals from each. Three days later, we scored F₁ animals for viable progeny. We chose B1 as the best sibling group and picked plates B1.9, B1.11, and B1.14 as our source for L4 larvae (14 F₂ animals each). Three days later we scored F₂ for viable progeny. We chose B1.11 as the best sibling group, and plates B1.11.2, B1.11.5, and B1.11.11 as our source for L4 F₃ animals each. For F₄'s, we chose F₃ sibling group B1.11.2 and picked 14 animals from plates B1.11.2. 2, B1.11.2.7, and B1.11.2.13. We repeated this pedigree selection protocol using each of the two long dsRNA triggers, A1 and A2 (see Figure 1). For each generation and both triggers we scored 42 animals. We depict viability of progeny in sibling groups with a white background and inviability with a gray background. Solid circles represent plates selected at each generation to further pedigrees.

Figure 3.—

Multigenerational assays for oma-1(zu405) silencing. Following the protocol described in Figure 2 leads to robust silencing in the F₁, F₂, and F₃ generations followed by a severe drop to zero silencing in the F₄ generation. At each generation, we plotted three subpopulations of 14 animals each. All animals analyzed had viable progeny (100% observed in F₁, F₂, and F₃), and all F₄ animals had no viable progeny. The error bars represent one standard deviation for each sibling group.

Figure 4.—

Dependence of heritable silencing on injected trigger concentration. We tested a fivefold serial dilution of the dsRNA induction trigger using the same pedigree selection scheme described in Figure 2. At each concentration, we picked three injected animals to select 14 F₁ animals, placed them on individual plates, and incubated for 3 days at 23°. We used the same criteria to select the descendant populations. We found animals injected with concentrations of 50 to 0.4 ng/μl had an equivalent silencing response: all animals scored in generations F₁, F₂, and F₃ had viable progeny, while all animals of the F₄ generation had no viable progeny. Silencing efficacy of concentrations of 0.08 and 0.016 ng/μl were less effective at both the overall silencing frequency (<100%) and in the persistence of the silencing response. Animals injected with 0.0032 ng/μl showed no silencing response.

Figure 5.—

Comparison of silencing efficiency between early and late-born progeny. We injected animals and selected F₁ progeny by birth order and determined their silencing capacity. We segregated animals from the same brood as (1) early born animals (born the first 24 hr after the injection) and (2) late-born animals (born the second 24 hr after injection). We found injected concentrations of 50, 10, 2, and 0.4 ng/μl show no significant difference in silencing between early and late-born siblings. In contrast, at concentrations of 0.08 and 0.016 ng/μl, there is a significant difference between the early born animals (solid line) and late-born animals (dashed line). Early born progeny of injection concentrations of 0.08 and 0.016 ng/μl had silencing frequencies of 67.4 and 46.7%, respectively, while the late-born progeny for both concentrations has a silencing frequency of 0%. Bars represent 1 SD.

Figure 6.—

Tests for silencing transmission of oocytes and linkage to the chromosomal locus exposed to dsRNA. (A) Schematic of crosses designed to follow chromosome origin from oocyte transmission. We injected animals that were morphologically dumpy by carrying a homozygous recessive allele of dpy-20(e1282ts). The dpy-20 allele is linked to the oma-1 locus and marks the origin of the chromosome. Injected animals self-fertilized and we used the F₁ hermaphrodites to cross with males not exposed to dsRNA and with a wild-type copy of the dpy-20 gene. The cross-progeny were non-dumpy heterozygous (only cross-progeny are non-dumpy). We then allowed heterozygous animals to produce self-fertilized progeny. Dumpy animals are dpy-20(e1282ts) homozygous and non-dumpy animals are either homozygous wild type or heterozygous for dpy-20(e1282ts). F₃ animals descended from F₂ cross-progeny inherited chromosomes from ancestors exposed or not exposed to dsRNA. We scored the F₃ animal's capacity for producing viable progeny. (B) Results of linked heritable silencing assay. We followed the genetic scheme described in A and individually plated F₃ animals carrying exposed or nonexposed chromosomes to dsRNA. Homozygous animals carrying the wild-type dpy-20 allele inherited their oma-1 locus from nonexposed animals. To determine if silencing was segregating with the origin of the chromosomes, we used animals having broods of >10 progeny. We used two ranges in brood size (10–30, open bars; >31, shaded bars) as indicators of the efficacy of the silencing achieved. Error bars are 1 SD. The silencing efficacy of F₃ animals demonstrates that (1) the transmission through the oocyte is sufficient to transmit silencing capacity and (2) the silencing capacity is unlinked to the origin of the oma-1 locus.

Figure 6.—

Tests for silencing transmission of oocytes and linkage to the chromosomal locus exposed to dsRNA. (A) Schematic of crosses designed to follow chromosome origin from oocyte transmission. We injected animals that were morphologically dumpy by carrying a homozygous recessive allele of dpy-20(e1282ts). The dpy-20 allele is linked to the oma-1 locus and marks the origin of the chromosome. Injected animals self-fertilized and we used the F₁ hermaphrodites to cross with males not exposed to dsRNA and with a wild-type copy of the dpy-20 gene. The cross-progeny were non-dumpy heterozygous (only cross-progeny are non-dumpy). We then allowed heterozygous animals to produce self-fertilized progeny. Dumpy animals are dpy-20(e1282ts) homozygous and non-dumpy animals are either homozygous wild type or heterozygous for dpy-20(e1282ts). F₃ animals descended from F₂ cross-progeny inherited chromosomes from ancestors exposed or not exposed to dsRNA. We scored the F₃ animal's capacity for producing viable progeny. (B) Results of linked heritable silencing assay. We followed the genetic scheme described in A and individually plated F₃ animals carrying exposed or nonexposed chromosomes to dsRNA. Homozygous animals carrying the wild-type dpy-20 allele inherited their oma-1 locus from nonexposed animals. To determine if silencing was segregating with the origin of the chromosomes, we used animals having broods of >10 progeny. We used two ranges in brood size (10–30, open bars; >31, shaded bars) as indicators of the efficacy of the silencing achieved. Error bars are 1 SD. The silencing efficacy of F₃ animals demonstrates that (1) the transmission through the oocyte is sufficient to transmit silencing capacity and (2) the silencing capacity is unlinked to the origin of the oma-1 locus.

Figure 7.—

Comparison of silencing transmission capacity for oocytes and sperm. (A) Genetic scheme. We first selected one injected animal (I₀) to produce F₁ and F₂ descendants by self-fertilization. We then selected one F₂ animal with a large brood size to separately assess sperm and oocyte transmission of silencing. (B) Results of sperm/oocyte comparison. F₃ cross-progeny animals were scored individually (7 animals for oocyte transmission and 6 for sperm transmission). F₄ descendants were scored in groups of siblings rather than individually since our previous experiments (Figure 2) indicated a consistent F₄ bottleneck. In the oocyte experiments, only 2 of 7 F₃ animals had progeny: 1 animal had 48 progeny and 1 animal had 1 progeny. None of the F₄ animals had F₅ progeny. In contrast, in the sperm-transmission experiments, all 6 F₃ animals had viable progeny, with brood sizes ranging from 28 to 80. We used groups of F₄ animals from sperm transmission and found 26/26 plates with 10 animals to a plate to have viable progeny. To analyze the F₅ silencing, we checked the F₄ plates for fertile F₅ animals. We found 2/26 plates with viable F₅ progeny.

Figure 7.—

Comparison of silencing transmission capacity for oocytes and sperm. (A) Genetic scheme. We first selected one injected animal (I₀) to produce F₁ and F₂ descendants by self-fertilization. We then selected one F₂ animal with a large brood size to separately assess sperm and oocyte transmission of silencing. (B) Results of sperm/oocyte comparison. F₃ cross-progeny animals were scored individually (7 animals for oocyte transmission and 6 for sperm transmission). F₄ descendants were scored in groups of siblings rather than individually since our previous experiments (Figure 2) indicated a consistent F₄ bottleneck. In the oocyte experiments, only 2 of 7 F₃ animals had progeny: 1 animal had 48 progeny and 1 animal had 1 progeny. None of the F₄ animals had F₅ progeny. In contrast, in the sperm-transmission experiments, all 6 F₃ animals had viable progeny, with brood sizes ranging from 28 to 80. We used groups of F₄ animals from sperm transmission and found 26/26 plates with 10 animals to a plate to have viable progeny. To analyze the F₅ silencing, we checked the F₄ plates for fertile F₅ animals. We found 2/26 plates with viable F₅ progeny.

Figure 8.—

Transmission as a function of the oma-1 genetic background. (A) We examined the transmission of the silencing character in two loss-of-function backgrounds of oma-1. The genetic scheme shows the missense mutation zu405te36 as the starting point for the experiment. The same scheme was followed to examine the silencing efficacy, starting with the loss-of-function nonsense mutation zu405te33. Loss-of-function strains (zu405te36 and zu405te33) of oma-1 were injected with dsRNA trigger A2 and crossed to gain-of-function zu405 strains. Sperm transmission experiments were done at 16° and oocyte transmission at 23°. (B) Calculated frequencies of silencing were measured by the frequency of homozygous zu405 F₃ animals with viable F₄ progeny.

Figure 8.—

Transmission as a function of the oma-1 genetic background. (A) We examined the transmission of the silencing character in two loss-of-function backgrounds of oma-1. The genetic scheme shows the missense mutation zu405te36 as the starting point for the experiment. The same scheme was followed to examine the silencing efficacy, starting with the loss-of-function nonsense mutation zu405te33. Loss-of-function strains (zu405te36 and zu405te33) of oma-1 were injected with dsRNA trigger A2 and crossed to gain-of-function zu405 strains. Sperm transmission experiments were done at 16° and oocyte transmission at 23°. (B) Calculated frequencies of silencing were measured by the frequency of homozygous zu405 F₃ animals with viable F₄ progeny.

Figure 9.—

Explicit test of male transmission of silencing. (A) Males during copulation transmit both sperm and seminal fluid. The male-derived silencing efficacy can be explained by at least two models: (1) The silencing factor is inside the sperm and (2) the silencing factor is transmitted through the male in the seminal fluid. (B) Schematic of crosses designed to test the male transmission of silencing through the sperm or the seminal fluid. It was critical for this experiment that we identify self-progeny animals that had been fertilized after their parent hermaphrodite had received male sperm and seminal fluid. To ensure this, we transferred the parent hermaphrodites each day and scored only self-progeny that derive from mothers that had previously produced cross-progeny. Operationally, this was carried out by mating individual F₁ male silencing carriers with five naive hermaphrodites for 6–12 hr, transferring the hermaphrodites to individual fresh plates to allow egg laying for 1 day (first brood), and transferring hermaphrodite mothers to a second plate for an additional day (second brood). Of 50 mated hermaphrodites, 6 met the criterion that they had some cross-progeny on the first day of transfer and some self-progeny on the second day of transfer. The self-progeny broods on these six plates from the second transfer consist of self- and cross-progeny that were fertilized subsequent to the transfer of sperm and seminal fluid from males to the mother hermaphrodite. We then compare silencing transmission to self-progeny and cross-progeny from these broods. The boxes summarize the viability of F₃ and F₄ cross- and self-progeny from first and second transfers at 25°. The data show that carrier males transfer the silencing trait to cross-progeny and not to self-progeny. This is consistent with a signal intrinsic to sperm and not one carried in the seminal fluid. The asterisk indicates that a single viable F₃ larva was produced from 1 of the 36 F₂ animals in this experiment; this animal yielded no F₄ progeny and may have represented a rare “spontaneous rescue” affecting ∼1 in 10⁴ progeny of oma-1(zu405) mothers.

Figure 9.—

Explicit test of male transmission of silencing. (A) Males during copulation transmit both sperm and seminal fluid. The male-derived silencing efficacy can be explained by at least two models: (1) The silencing factor is inside the sperm and (2) the silencing factor is transmitted through the male in the seminal fluid. (B) Schematic of crosses designed to test the male transmission of silencing through the sperm or the seminal fluid. It was critical for this experiment that we identify self-progeny animals that had been fertilized after their parent hermaphrodite had received male sperm and seminal fluid. To ensure this, we transferred the parent hermaphrodites each day and scored only self-progeny that derive from mothers that had previously produced cross-progeny. Operationally, this was carried out by mating individual F₁ male silencing carriers with five naive hermaphrodites for 6–12 hr, transferring the hermaphrodites to individual fresh plates to allow egg laying for 1 day (first brood), and transferring hermaphrodite mothers to a second plate for an additional day (second brood). Of 50 mated hermaphrodites, 6 met the criterion that they had some cross-progeny on the first day of transfer and some self-progeny on the second day of transfer. The self-progeny broods on these six plates from the second transfer consist of self- and cross-progeny that were fertilized subsequent to the transfer of sperm and seminal fluid from males to the mother hermaphrodite. We then compare silencing transmission to self-progeny and cross-progeny from these broods. The boxes summarize the viability of F₃ and F₄ cross- and self-progeny from first and second transfers at 25°. The data show that carrier males transfer the silencing trait to cross-progeny and not to self-progeny. This is consistent with a signal intrinsic to sperm and not one carried in the seminal fluid. The asterisk indicates that a single viable F₃ larva was produced from 1 of the 36 F₂ animals in this experiment; this animal yielded no F₄ progeny and may have represented a rare “spontaneous rescue” affecting ∼1 in 10⁴ progeny of oma-1(zu405) mothers.

Figure 10.—

Relaxed stringency of early selection allows some persistence of silencing in the F₄ generation. (A) Schematic of assay. We designed a selection process to evaluate the relationship between the strength of the silencing response measure by the silencing frequency in a particular pedigree, to the persistence of the silencing across generations. Degrees of silencing efficacy were determined by the silencing frequency and brood size of the selected animals. We used the frequency of silencing to classify pedigrees as transmitting at highest, intermediate, or low silencing efficacy. When then used brood size as a second criterion to guide the selection of individuals to analyze the silencing frequency of the next generation. In the “highest silencing efficacy” group, we selected from plates with the largest brood sizes (>90). In the intermediate silencing efficacy group, we selected individuals from plates with broods between 30 and 80. Animals where most siblings have no viable progeny represent low-silencing-efficacy groups and were not used. (B) Intermediate silencing efficacy populations overcome the F₄ bottleneck. We followed the less stringent selection scheme of intermediate silencing efficacy and found that 7/10 F₄ sibling groups had at least some viable F₅ progeny. This is in contrast to the F₄ bottleneck that we observed when we used the “highest silencing efficacy” selection (data in Figure 3 and data not shown). Error bars represent 1 SD. As the manner in which animals are chosen to carry forward the silencing trait is critical in determining the behavior of descendant populations, we describe the selection process for the intermediate silencing efficacy group in some detail as follows: The F₄ animals, classified as descendants of continuous intermediate silencing efficacy selection, were derived from one of five injected animals. Of the original five injected animals, we picked all viable progeny and arbitrarily assigned each a color (purple, red, green, orange, and blue). Three days later, all injected animals had viable progeny. We individually plated the F₁ animals and scored the frequency of viable F₂ progeny. Only the orange F₁ family had no viable progeny (n = 15). All F₁ plates from blue (n = 40), red (n = 54), purple (n = 91), and green (n = 20) had viable progeny. We selected F₂ animals from eight F₁ families: two blue, one green, two purple, and three red. Each F₁ family gave rise to an F₂ sibling group (designated by two letters). From the blue family, the BD group had 100% plates with viable progeny while the BE group had only 7.8%. From the green family, GF had 80%; from the purple families, both PH and PJ had 100% transmission; and from the red families, RA had 94.7%, RB had 80.7%, and RC had 100%. The RA and RB lineages fulfilled the criteria for selection of intermediate silencing efficacy. To extract the populations with smaller brood sizes, we removed the F₂ animals at day 2. On day 3, we scored the F₂ plates. This allowed us to better assign a generation to animals by increasing the age difference between F₃ adults and young F₄ larvae. Two days after removing the F₂ adults, we surveyed all plates of F₂ animals with F₃ broods. Plates with large brood sizes had depleted the bacterial lawns. Plates with “smaller” F₃ broods were not depleted of bacteria (fewer worms on plate, more food per worm) and their growth was uninterrupted. We used F₄ animals from small broods to represent the intermediate silencing efficacy groups.

Figure 10.—

Relaxed stringency of early selection allows some persistence of silencing in the F₄ generation. (A) Schematic of assay. We designed a selection process to evaluate the relationship between the strength of the silencing response measure by the silencing frequency in a particular pedigree, to the persistence of the silencing across generations. Degrees of silencing efficacy were determined by the silencing frequency and brood size of the selected animals. We used the frequency of silencing to classify pedigrees as transmitting at highest, intermediate, or low silencing efficacy. When then used brood size as a second criterion to guide the selection of individuals to analyze the silencing frequency of the next generation. In the “highest silencing efficacy” group, we selected from plates with the largest brood sizes (>90). In the intermediate silencing efficacy group, we selected individuals from plates with broods between 30 and 80. Animals where most siblings have no viable progeny represent low-silencing-efficacy groups and were not used. (B) Intermediate silencing efficacy populations overcome the F₄ bottleneck. We followed the less stringent selection scheme of intermediate silencing efficacy and found that 7/10 F₄ sibling groups had at least some viable F₅ progeny. This is in contrast to the F₄ bottleneck that we observed when we used the “highest silencing efficacy” selection (data in Figure 3 and data not shown). Error bars represent 1 SD. As the manner in which animals are chosen to carry forward the silencing trait is critical in determining the behavior of descendant populations, we describe the selection process for the intermediate silencing efficacy group in some detail as follows: The F₄ animals, classified as descendants of continuous intermediate silencing efficacy selection, were derived from one of five injected animals. Of the original five injected animals, we picked all viable progeny and arbitrarily assigned each a color (purple, red, green, orange, and blue). Three days later, all injected animals had viable progeny. We individually plated the F₁ animals and scored the frequency of viable F₂ progeny. Only the orange F₁ family had no viable progeny (n = 15). All F₁ plates from blue (n = 40), red (n = 54), purple (n = 91), and green (n = 20) had viable progeny. We selected F₂ animals from eight F₁ families: two blue, one green, two purple, and three red. Each F₁ family gave rise to an F₂ sibling group (designated by two letters). From the blue family, the BD group had 100% plates with viable progeny while the BE group had only 7.8%. From the green family, GF had 80%; from the purple families, both PH and PJ had 100% transmission; and from the red families, RA had 94.7%, RB had 80.7%, and RC had 100%. The RA and RB lineages fulfilled the criteria for selection of intermediate silencing efficacy. To extract the populations with smaller brood sizes, we removed the F₂ animals at day 2. On day 3, we scored the F₂ plates. This allowed us to better assign a generation to animals by increasing the age difference between F₃ adults and young F₄ larvae. Two days after removing the F₂ adults, we surveyed all plates of F₂ animals with F₃ broods. Plates with large brood sizes had depleted the bacterial lawns. Plates with “smaller” F₃ broods were not depleted of bacteria (fewer worms on plate, more food per worm) and their growth was uninterrupted. We used F₄ animals from small broods to represent the intermediate silencing efficacy groups.