Identification of mutant phenotypes associated with loss of individual microRNAs in sensitized genetic backgrounds in Caenorhabditis elegans

Summary

MicroRNAs (miRNAs) are small, non-coding RNAs that regulate the translation and/or the stability of their mRNA targets. Previous work showed that for most miRNA genes of C. elegans, single gene knockouts did not result in detectable mutant phenotypes [1]. This may be due, in part, to functional redundancy between miRNAs. However, in most cases, worms carrying deletions of all members of a miRNA family do not display strong mutant phenotypes [2]. They may function together with unrelated miRNAs or with non-miRNA genes in regulatory networks, possibly to ensure the robustness of developmental mechanisms. To test this, we examined worms lacking individual miRNAs in genetically sensitized backgrounds. These include genetic backgrounds with reduced processing and activity of all miRNAs or with reduced activity of a wide array of regulatory pathways [3]. Using these two approaches, mutant phenotypes were identified for 25 out of 31 miRNAs included in this analysis. Our findings describe biological roles for individual miRNAs and suggest that use of sensitized genetic backgrounds provides an efficient approach for miRNA functional analysis.

Results

Loss of individual miRNAs can enhance or suppress alg-1 developmental defects

Genetic analysis demonstrates that development of worms, flies, fish and mice requires miRNAs [4-9]. While it is clear that normal development of the worm requires miRNA biogenesis, functions have been described for only a few individual miRNAs. In C. elegans, the majority of individual miRNAs are not required for viability or for development, as most loss of function miRNA mutants display no obvious developmental abnormalities [1]. To test the hypothesis that the absence of phenotypes in miRNA mutant worms is due to functional redundancy with other miRNAs, we determined if loss of specific miRNAs in a sensitized genetic background resulted in observable mutant phenotypes. For a sensitized background, we used a loss of function allele of alg-1 (argonaute like 1), which is one of two Argonaute-encoding genes that function in the miRNA pathway in worms. Compared to wild-type worms, alg-1 mutants have lower levels of mature miRNAs [6, 10]. Whereas worms that lack alg-1 and alg-2 activities die during embryogenesis [6], alg-1 single mutant worms are viable, displaying developmental timing defects, molting defects and early adult-stage lethality [11-14].

A subset of 25 miRNA deletion alleles covering 31 miRNA-encoding genes were chosen for analysis (Table 1). As a subset of miRNAs are clustered in the genome, some alleles affect multiple miRNAs. For example, nDf58 is a deletion of three miRNAs: mir-54, mir-55, and mir-56 (mir-54-56). We selected miRNAs that show evolutionary conservation of miRNA family members based on their seed sequence or a developmentally-regulated expression pattern [15-19]. First, all of the strains carrying deletion alleles were outcrossed with wild-type N2 worms (see Table S1 available online). A set of multiply mutant strains was then constructed that carried a single miRNA deletion allele and the gk214 allele of alg-1. We performed phenotypic analysis of our collection of 25 alg-1;mir multiply mutant strains. This analysis focused on readily-observable, quantifiable, developmental phenotypes: embryonic, larval, and adult lethality, gross morphology and motility, dauer formation, gonad migration, and alae formation. A summary of the phenotypic analysis is shown in Table 1, which includes the identification of novel enhanced or synthetic phenotypes for 19 of the 25 strains analyzed.

Table 1.

Phenotypic Characterization of miRNA mutants in alg-1 sensitized genetic background

		Developmental Timing	Gonad Migration	Embryonic Lethality	Adult Lethality
Strain	Genotype	% Incomplete alae formationa	% Abnormalb	% Unhatchedc	% Dead at 72hrd
N2	wild type	0%	0%	0%	0%
RF54	alg-1(gk214)	61%	8%	3%	63%
RF70	mir-1(n4102); alg-1(gk214)	57%	25% **	3%	53%
RF129	mir-34(n4276) alg-1(gk214)	59%	13%	5%	72%
RF420	mir-51(n4473); alg-1(gk214)	31% **	7%	8% *	51%
RF411	mir-52(n4114); alg-1(gk214)	3% **	0%	6%	16% **
RF398	mir-53(n4113); alg-1(gk214)	60%	17%	1%	57%
RF410	mir-54-55(nDf45) alg-1(gk214)	4% **	11%	2%	12% **
RF89	mir-54-56(nDf58) alg-1(gk214)	23% **	4%	3%	5% **
RF133	mir-57(gk175); alg-1(gk214)	51%	5%	8% **	73%
RF137	mir-59(n4604); alg-1(gk214)	69%	23% **	10% **	83% **
RF153	mir-72(n4130); alg-1(gk214)	56%	2%	5%	48%
RF81	mir-73-74(nDf47) alg-1(gk214)	75%	7%	3%	40% **
RF178	mir-77(n4285); alg-1(gk214)	54%	8%	12% **	59%
RF65	mir-83(n4638); alg-1(gk214)	51%	25% **	3%	77%
RF141	mir-85(n4117); alg-1(gk214)	48%	4%	3%	54%
RF77	mir-124(n4255); alg-1(gk214)	69%	18% *	2%	69%
RF145	mir-228(n4382); alg-1(gk214)	39% **	5%	13% **	53%
RF93	mir-234(n4520); alg-1(gk214)	53%	5%	3%	57%
RF182	mir-235(n4504); alg-1(gk214)	56%	7%	0%	80% **
RF85	mir-237(n4296) alg-1(gk214)	47%	12%	3%	75%
RF163	mir-238(n4112); mir-239a-b(nDf62) alg-1(gk214)	22% **	3%	1%	11% **
RF60	mir-240 mir-786(n4541) alg-1(gk214)	50%	13%	7% **	45% *
RF186	mir-244(n4367); alg-1(gk214)	13% **	9%	5%	53%
RF149	mir-246(n4636); alg-1(gk214)	44%	7%	8% **	71%
RF368	mir-247 mir-797(n4505) alg-1(gk214)	41%	25% **	4%	57%
RF343	mir-259(n4106); alg-1(gk214)	34% **	28% **	4%	59%

^*p < 0.05,

^**p < 0.01 by the chi-square test, as compared to alg-1 single mutants.

^aAlae were scored at the L4m using DIC microscopy. n > 39 (range: 39-204) worms scored for each strain

^bGonad morphology was scored in young adult worms using DIC microscopy. n > 41 (range: 41-262) worms scored for each strain

^cEmbryos were transferred to a new plate and scored after 16-24 hours for the presence of unhatched embryos. n > 78 (range: 78-548) embryos scored for each strain.

^dSynchronized L1-stage worms were transferred to plates to initiate development. Lethality was scored 72 hours after plating at 20°C, n > 76 (range: 76-172) worms scored for each strain.

First, we observed that loss of mir-51, mir-57, mir-59, mir-77, mir-228, mir-240 mir 786, or mir-246 in the alg-1 sensitized background resulted in a significant increase in the percentage of embryonic lethality, from 3% in alg-1 single mutants to 7-13% in the alg-1;mir multiple mutants (Table 1). Consistent with a role in embryogenesis, embryonic expression has been reported for mir-51, mir-57, mir-59, and mir-228 [17, 20]. However, no embryonic expression has been reported for mir-77, mir-240 mir-786, or mir-246. The point at which these embryos arrested was not determined and because the increase in embryonic lethality was relatively modest, transgenic rescue was not performed. Recently, it has been shown that worms lacking all six members of the mir-51 miRNA family display a penetrant embryonic lethal phenotype [2], demonstrating an essential role for this miRNA family in embryogenesis.

Second, we observed that loss of mir-1, mir-59, mir-83, mir-124, mir-247 mir-797, or mir-259 in the alg-1 sensitized background resulted in a significant increase in the percentage of worms that had defective distal tip cell migration as determined by the gonad morphology in young adult worms (Table 1). In contrast to wild-type animals that execute a single, developmentally-regulated reflex of each gonad arm, 8% of alg-1 single mutants displayed abnormal distal tip cell migration, with the distal tip cell executing an extra turn during development. The distal tip cell of the gonad arms migrated normally to the anterior or posterior, made a dorsal turn, initiated migration back to the midline normally but then executed an extra turn back away from the midline. The extra turn was observed to occur predominantly, but not exclusively, in the posterior gonad arm. In all six alg-1; mir strains, this enhanced gonad migration phenotype was rescued by the extra-chromosomal expression of a wild-type genomic fragment containing the miRNA stem-loop sequence along with flanking sequence (Table S1). The expression patterns of most of these miRNAs are consistent with a possible role in the regulation of secreted or membrane-bound guidance molecules: mir-1 is expressed in the body wall muscles [21] mir-59 and mir-83 are expressed in the intestine [20] and mir-247 mir-797 is expressed in the distal tip cell [20]. Proteins from the body wall muscle and intestinal cells are required for guidance cues in the body wall basement membrane over which the distal tip cells migrate [22, 23].

Third, we observed that loss of mir-59 and mir-235 resulted in an increased level of adult lethality compared to alg-1 single mutants (Table 1). alg-1 single mutants appeared healthy at the L4-to-adult transition but 63% die as young adults (Table 1). About 62% of alg-1 worms enter a supernumerary molt in the adult stage (Table S2). Nearly all of the adult worms that enter the molt subsequently die with a “bag of worms” phenotype (Table S2), likely due to a failure to properly complete the molting cycle that compromises the worm’s ability to lay embryos. Execution of a supernumerary molt is a heterochronic phenotype also observed in worms missing let-7 family members [24, 25]. This adult lethality was enhanced in mir-59;alg-1 and mir-235;alg-1 worms, with 83% and 80% adult lethality, respectively (Table 1). The increase in adult lethality observed in these strains is primarily due to an increase in the percentage of worms that enter a supernumary lethargus (Table S2).

In contrast to this enhancement of the alg-1 adult lethality phenotype, we found that loss of other miRNAs suppressed alg-1 phenotypes. In particular, loss of mir-73-74, mir-238/mir-239a-b, mir-240 mir-786, or certain mir-51 family members (mir-52, mir-54-55, and mir-54-56) in the alg-1 background resulted in a reduced level of adult lethality when compared to alg-1 single mutants (Table 1). This decrease in adult lethality is primarily due to a decrease in the percentage of worms that enter a supernumary lethargus (Table S2). Similarly, we observed that loss of mir-228, mir-238/mir-239a-b, mir-244, mir-259 or mir-51 family members (mir-51, mir-52, mir-54-55, and mir-54-56), resulted in a reduced penetrance of developmental timing defects, as assayed by the formation of an adult-specific cuticle structure called alae (Table 1). About 60% of alg-1 single mutants showed incomplete alae formation owing to a partial reiteration of larval stage programs, which is likely due to reduced levels of lin-4 and let-7 family miRNAs [6]. Incomplete alae formation indicates an inappropriate larval cell fate for some of the cells in the hypodermis at the L4-to-adult transition. Introduction of a transgene with the mir-54-56, mir-238, mir-244, or mir-259 genomic loci into the respective alg-1;mir strain, restored the level of incomplete alae formation to that observed in alg-1 single mutants (Table S1). Four strains, mir-52; alg-1, mir-54-55 alg-1, mir-54-56 alg-1 and mir-238; mir-239a-b alg-1 displayed both reduced adult lethality as well as reduced alae formation defects (Table 1).

Loss of individual miRNAs confers enhanced or synthetic sterility with knockdown of a set of chromatin regulator “hub” genes

We next tested whether miRNA genes functionally interact with broadly-acting pathways of gene regulation. A small set of genes, termed “hub” genes, have been identified that show a high level of connectivity to many developmental processes [3]. These hub genes encode chromatin regulatory proteins and interact with a wide range of signaling pathways that are critical for normal development, including the Wnt, EGF, and Notch pathways [3]. Therefore, we hypothesized that the reduced activity of these hub genes could provide a sensitized genetic background in which to reveal additional relationships between individual miRNA mutants and other developmental processes. To address this question directly, we determined whether individual miRNA deletion alleles displayed enhanced or synthetic interactions upon knockdown of one of five hub genes by RNAi: egl-27, din-1, hmg-1.2, dpy-22, and trr-1 (Table 2). We performed phenotypic assays on eleven strains carrying miRNA deletion alleles to identify embryonic lethal or sterile phenotypes upon knockdown of hub genes.

Table 2.

Summary of hub genes and synthetic interactions with developmental pathways

gene	descriptiona	Enhancement with developmental regulatory componentsb
din-1	ortholog of human transcriptional co-repressor SHARP/SPEN, encodes large RNA-binding protein of RRM superfamily	EGF, Wnt, Notch, ephrin receptor
dpy-22	ortholog of the human transcriptional Mediator protein TRAP230	EGF, Wnt, Notch, E2F/synMuvB, ephrin receptor, RAC1/GTPase
egl-27	ortholog of human MTA1, part of nucleosome remodelling and histone deacetylation (NURD) complex	EGF, Notch, E2F/synMuvB, ephrin receptor
hmg-1.2	DNA binding protein with HMG box	EGF, Wnt, Notch, ephrin receptor
trr-1	Ortholog of component of the NuA4/Tip60 histone acetyltransferase complexes,atypical protein kinase of TRAAP subfamily	EGF, Wnt, Notch, E2F/synMuvB, GTPase

^ainformation from www.wormbase.org

^bsee [3] for details

We found that RNAi knockdown of egl-27, din-1, hmg-1.2, dpy-22, and trr-1 in wild-type worms did not result in a significant number of worms that were sterile as compared to the negative control RNAi (Table 3). However, four miRNA mutant strains displayed a synthetic sterile phenotype following knockdown of individual hub genes. When egl-27 or hmg-1.2 activity were knocked down in mir-1 mutant worms, 24% and 29% of worms were sterile, respectively (Table 3). No sterility was observed when mir-1 worms were fed bacteria containing an empty vector construct. Like mir-1 worms, mir-59 worms showed a synthetic sterile phenotype with knockdown of egl-27 or hmg-1.2 (Table 3). When either egl-27 or trr-1 activity was knocked down in mir-247 mir-797 mutant worms 29% and 40% of worms were sterile, respectively, compared to 3% in the negative control group (Table 3). The strongest synthetic phenotype was observed in mir-240 mir-786 worms following knockdown of hmg-1.2 or trr-1, with 76% and 83% sterility (Table 3). These worms exhibited germline defects (Figure S1). For all strains that displayed a sterile phenotype upon hub gene knockdown, there was no significant embryonic lethality (Table S3). Few, if any, embryos were found on plates with mir-240 mir-786 worms following knockdown of hmg-1.2 or trr-1 activity (Table S3). These data suggest that mir-1, mir-59, mir-240 mir-786, and mir-247 mir-797 function in regulatory pathways that are essential for germline development. Of these, only mir-240 mir-786 has been shown to be expressed in the gonad using miRNA promoter reporter transgenes, with expression observed in the uterus, spermatheca, and gonadal sheath [20]. It is possible that mir-1, mir-59, and mir-247 mir-797 are expressed in the germline but this expression is undetectable using reporter transgenes or that these miRNAs function to control developmental signals required for fertility from outside of the germline.

Table 3.

Phenotypic Characterization of miRNA mutants following knockdown of hub gene activity by RNAi.

		% Sterile Worms
Strain	Genotype	empty vector RNAi	din-1 RNAi	dpy-22 RNAi	egl-27 RNAi	hmg1.2 RNAi	trr-1 RNAi
N2	wild-type	0%	11%	4%	3%	7%	14%
RF71	mir-1(n4102)	0%	16%	4%	24% *	29% *	16%
RF124	mir-34(n4276)	3%	2%	0%	7%	3%	0%
RF90	mir-54-56(nDf58)	0%	4%	0%	0%	5%	21%
RF15	mir-59(n4604)	7%	24%	8%	21% *	26% *	27%
RF66	mir-83(n4638)	0%	5%	5%	0%	11%	7%
RF78	mir-124(n4255)	0%	4%	0%	0%	14%	24%
RF94	mir-234(n4520)	0%	7%	0%	9%	11%	9%
RF86	mir-237(n4296)	2%	0%	0%	0%	0%	12%
RF61	mir-240 mir-786(n4541)	0%	17%	13%	4%	76% **	83% **
RF126	mir-246(n4636)	0%	5%	3%	15%	16%	33%
RF24	mir-247 mir-797 (n4505)	3%	11%	9%	29% **	13%	40% *

^*p < 0.05,

^**p < 0.01 by the chi-square test as compared to wild-type worms with the corresponding hub gene RNAi.

Discussion

The goal of this work was to test the hypothesis that miRNAs have overlapping or redundant functions and to identify phenotypes associated with the loss of individual miRNAs. We found that loss of individual miRNAs resulted in developmental abnormalities in an alg-1 genetic background, which has lower total miRNA activity, and in backgrounds with reduced hub gene activity, in which multiple regulatory pathways are compromised. Our analysis described mutant phenotypes associated with loss of 25 out of the 31 miRNAs included in our study.

These results identified only a limited number of phenotypes in the alg-1;mir strains. This indicates that only a subset of miRNA-regulated pathways are sufficiently sensitized in alg-1 worms, while others may require a further reduction in the activity of miRNAs to become sensitized. Additional knockdown of alg-2 activity, which is the other Argonaute-encoding gene that acts in the miRNA pathway, could be used to enhance the sensitivity of this assay. However, complete loss of alg-1 and alg-2 results in embryonic lethality [6]. The synthetic sterility observed with individual miRNAs and hub genes indicates that pathways that are essential for germline development are sensitized in these backgrounds. Additional work is needed to identify the specific regulatory pathways and targets that are controlled by individual miRNAs.

Our assays identified multiple miRNAs that interact with either alg-1 or individual hub genes to give similar phenotypes. One parsimonious model for this observation is that these sets of miRNAs may regulate shared mRNA targets. For example, mir-1, mir-59, and mir-240 mir-786 all interact with hmg-1.2 (RNAi) to give a synthetic sterile phenotype. To examine this hypothesis, we identified predicted mRNA targets using Targetscan [19] or mirWIP [26]. Both prediction algorithms identified possible shared targets for pairs of these miRNAs (Table S4), with three targets that are possible shared targets for three miRNAs (miR-1, miR-59, and miR-786): daf-12, hbl-1, and cfim-2. Interestingly, cfim-2 encodes an mRNA cleavage and polyadenylation factor that is required for fertility [27]. This analysis of target predictions provides a platform to identify the biologically-relevant targets and to test the hypothesis that multiple miRNAs regulate shared targets. Alternatively, the phenotypes we observe may reflect miRNA regulation of distinct targets in convergent developmental pathways.

Interestingly, loss of a subset of miRNAs, mir-228, mir-238;mir-239a/b, mir-244, mir-259 or certain mir-51 family members (mir-51, mir-52, and mir-54-56) resulted in a suppression of alg-1 alae formation defects. This could indicate regulation of specific targets in the developmental timing pathway. In this model, some miRNAs could function antagonistically to the lin-4 or let-7 family miRNAs, which are largely responsible for the alg-1 developmental timing defects [6]. An alternative explanation for the observed suppression of alg-1 phenotypes by certain miRNA gene mutations could be that these miRNAs regulate specific targets that control miRISC activity, such that in their absence, miRISC activity is enhanced. For example, the miRISC cofactor, nhl-2 [13], is predicted to be a target of miR-244 and miR-228 using TargetScan [19] and mirWIP [26] miRNA target prediction algorithms, suggesting a possible mechanism for suppression. It is also possible that the observed suppression may be due to a more general effect on miRISC activity. If individual miRNAs compete for loading into miRISC, then loss of a highly abundant miRNA, such as miR-52 [17, 19], may allow for greater activity of remaining miRNAs, including the let-7 family. However, this can not account for the observed suppression in all alg-1;mir strains as loss of abundant miRNAs does not suppress alg-1 defects in all cases. Additionally, strong suppression of alg-1 defects is observed with loss of weakly expressed miRNAs, such as mir-238 or mir-259 [19]. Interestingly, Targetscan identifies alg-2 as a candidate target for miR-259 [19].

Few phenotypes have been described for individual miRNA mutants [see 28]. In addition, few phenotypes are observed in strains that lack most or all miRNA family members, suggesting that the lack of detectable phenotypes for individual miRNA mutants is not due primarily to overlapping function with related miRNA family members [2]. Our results indicate that miRNAs may function together with unrelated miRNAs or non-miRNA genes, perhaps functioning together to ensure the robustness of developmental mechanisms. The use of sensitized genetic backgrounds is a fruitful approach to identify phenotypes associated with the loss of specific miRNAs. Genetic backgrounds can be selected that have compromised activity of specific regulatory pathways, which may render them sensitive to the loss of individual miRNAs. It is possible that functions of some miRNAs may be revealed only when sufficient genetic or environmental variation is introduced.

Experimental Procedures

General Methods and strains

C. elegans strains were maintained under standard conditions as previously described [29]. Worms were kept on NGM plates seeded with E. coli strain AMA1004 [30]. The wild type strain used was var. Bristol N2 [31]. RF54 was used as the alg-1(gk214)X control in phenotypic analysis of mir; alg-1 strains. All strains were kept at 20°C unless otherwise indicated. miRNA mutant strains were first outcrossed to wild-type N2 (Table S1). For building multiply mutant strains, presence of the miRNA deletion alleles and the alg-1 allele in F2s were identified by performing PCR with primers that amplified the genomic region flanking the deletion mutation. Sequences for primers used for genotyping can be found in Table S1. Fluorescence and DIC microscopy was performed using a Nikon Eclipse 80i equipped with a Photometrics CoolSNAP HQ2 monochrome digital camera and RS Image software (Roper Scientific).

RNAi experiments

Four L4-stage worms were placed on RNAi plates (NGM with 0.2% (w/v) lactose and 100 μg/ml ampicillin ) seeded with bacteria to knockdown egl-27, din-1, hmg-1.2, dpy-22, or trr-1. Empty vector (L4440) was used as the negative control. Bacteria for RNAi experiments were isolated from the Ahringer RNAi library [32]. Worms were placed at 20°C overnight. 48 F1 embryos were transferred to a new RNAi plate, seeded with the same bacteria. The next day live worms and unhatched embryos were scored. The surviving larval worms were cloned into 96 well culture plates, with each well containing RNAi liquid media (M9 containing 0.2% Lactose and 100 μg/ml ampicillin) plus bacteria. The bacteria were from frozen bacterial pellets taken from 5 mL stationary phase overnight cultures of appropriate bacteria strain in LB containing 0.2% lactose and 100 μg/ml ampicillin. 96-well plates were placed at 20°C for 96 hours and then wells were scored for the presence of F2 progeny. Wells in which the F1 worm died were not scored.

Transgene Rescue experiments

To create transgenic animals, germline transformation was performed as described [33]. Injection mixes contained 5-25 ng/μl of the rescue plasmid, 25-100 ng/μl of a co-injection marker (myo-2::gfp or myo-2::dsRed) containing plasmid, and pRS413 plasmid for a final DNA concentration of 150 ng/μl (Table S1). Transgenic animals expressing GFP or dsRed in the pharynx were assayed for alae formation or for gonad migration defects.