microRNA regulation of the embryonic hypoxic response in Caenorhabditis elegans

Layered strategies to combat hypoxia provide flexibility in dynamic oxygen environments. Here we show that multiple miRNAs are required for hypoxic survival responses during C. elegans embryogenesis. Certain miRNAs promote while others antagonize the hypoxic survival response. We found that expression of the mir-35 family is regulated by hypoxia in a HIF-1-independent manner and loss of mir-35–41 weakens hypoxic survival mechanisms in embryos. In addition, correct regulation of the RNA binding protein, SUP-26, a mir-35 family target, is needed for survival in chronic hypoxia. The identification of the full mRNA target repertoire of these miRNAs will reveal the miRNA-regulated network of hypoxic survival mechanisms in C. elegans.

Oxygen (O 2 ) is crucial for cellular survival. Its availability has had an enormous impact on shaping life in the variety of ecosystems, and many organisms have evolved to live and reproduce under the ambient O 2 concentration (21%). The correct level of O 2 in tissues during embryogenesis is particularly important for the successful execution of a variety of developmental events 1 . However, abnormally low O 2 levels (hypoxia) during development can lead to permanent defects and even premature termination of embryogenesis 2,3 . Therefore, embryos have developed molecular strategies to cope with inherent as well as exogenous hypoxic conditions 4,5 . The core hypoxic pathway relies on stabilization of the hypoxia-inducible factor HIF-1, which regulates the transcription of many protein-coding genes and microRNAs (miRNAs) responsible for the hypoxic response 6,7 . In addition, hif-1-independent hypoxic pathways have been identified in different models 8,9 . This suggests that other hif-1 independent pathways may exist to regulate aspects of the hypoxic response. miRNAs are small (18-25nt) non-coding RNA molecules that were first discovered in Caenorhabditis elegans 10 . They function mainly as negative post-transcriptional regulators by binding to the 3' untranslated region (UTR) of target mRNAs 11 . miRNAs are grouped in distinct families that share the same seed sequence, which is an important determinant for mRNA target specificity. Therefore, members of a miRNA family can conceptually act in a redundant manner. To date, miRNAs have been implicated in a range of developmental paradigms in C. elegans 10,[12][13][14] , in addition to stress-related phenomena such as the regulation of lifespan 5,[15][16][17] . Moreover, miRNAs 18 and other non-coding RNA species 19 function in the hypoxic response 20 . Their active turnover and ability to simultaneously regulate more than one target are properties that make miRNAs potentially useful stress response molecules 21,22 . Although they appear to regulate the majority of mRNAs in a cell 23 , information is often lacking on their specific role in different stress responses. We therefore sought to identify miRNAs that are important for the regulation of the hypoxic response in C. elegans.
We screened 40 C. elegans miRNA mutants covering 22 miRNA families for the ability to control embryonic survival in hypoxia. We found that loss of 14 miRNA families caused embryos to be either sensitive or resistant to hypoxic exposure. Focusing on the mir-35 family, we found that expression of all members of this family is induced in embryos exposed to hypoxia in a hif-1-independent manner. The mir-35 family is required for hypoxic survival as deletion mutants that remove mir-35-41 exhibit > 90% embryonic lethality under hypoxia. Finally, we identified sup-26, which encodes an RNA binding protein, as a target of the mir-35 family that is, at least in part, required for embryonic survival in chronic hypoxia. Our study is the first to offer insights into miRNA regulation of the embryonic hypoxic response in C. elegans and further work will identify targets of these miRNAs that together provide appropriate survival mechanisms in hypoxic conditions.

Results and Discussion
Multiple miRNAs are important for the embryonic hypoxic response. In an effort to identify miRNAs important for the control of hypoxic responses in C. elegans embryos, we screened a collection of mutants that lack individual or multiple miRNAs for embryonic survival under hypoxia ( Figure S1, Table S1 and Table S2). We exposed semi-synchronized populations of embryos to 0.5% O 2 for 24 hours at 20 o C and after recovery for 24 hrs at 21% O 2 embryos were scored for survival ( Figure S1). To exclude phenotypes caused by background mutations, we outcrossed the mutants and used multiple alleles or strains where possible. Using this approach, we were able to identify multiple gene families that regulate the ability of embryos to survive a hypoxia insult (Table S1). Loss of certain miRNAs or miRNA families led to hypoxia sensitivity (mir-2, mir-35, mir-44, mir-49, mir-51, mir-60, mir-63 and mir-67) and others to hypoxia resistance (let-7, mir-58, mir-67, mir-79, mir-237, mir-246, mir-359). Finally, we found that 20 miRNA mutant strains exhibited similar embryonic survival rates to wild type animals (Table S2). These data suggest that multiple miRNAs are required for C. elegans embryos to respond appropriately to hypoxia.
The mir-35-41 family regulates the embryonic response to hypoxia. We focused our studies on the mir-35-41(nDf50) mutant strain due to its high sensitivity to hypoxia ( Fig. 1 and Table S1). The mutant strain carrying the nDf50 deficiency lacks 7 (mir-35-41) of the 8 members of mir-35 family (Fig. 1A). The embryonic lethality observed in this mutant under ambient O 2 concentration (21% O 2 ) at 20 °C is approximately 50%, in agreement with previous reports 24 , however, lethality is increased to over 90% when freshly-laid embryos are exposed to 0.5% O 2 for 24 hrs at 20 °C (Fig. 1B). Hypoxia sensitivity was also phenocopied in the independently isolated mir-35-41(gk262) mutant strain ( Fig. 1A-B). A higher percentage of nDf50 mutant embryos die younger in hypoxia than in normoxia, implying that hypoxia affects development at an earlier stage in this mutant (Fig. 1C). Indeed, when we subjected older embryos (6-9 h old) to hypoxia we observed significantly lower embryonic lethality compared to when we subjected early embryos (0-3 h old) ( Figure S2A). The similarity between the elevated nDf50 embryonic lethality we found in hypoxia and the known high lethality of this mutant in high temperature 24 may reflect common components of hypoxia and heat stress such as challenges in protein folding and stability and induction of heat shock proteins 25 . However, when we subjected both mir-35-41 deletion mutant strains to 2 g/l sodium sulfite, which mimics hypoxic stress 26 , we also observed a significant increase in embryonic lethality ( Figure S3A) indicating that mir-35 family has a specific role in survival of embryos in hypoxia.
The mir-35-41 cluster is located within an intron of a worm specific gene (Y62F5A.9), and both nDf50 and gk262 lesions affect exonic sequences of this gene (Fig. 1A). We therefore performed rescue experiments by expressing either mir-35 alone or the entire mir-35-41 cluster under the control of mir-35-locus upstream sequence located within the intron of its host gene ( Fig. 1 and Figure S2B-C). We rescued embryonic lethality in both normoxic and hypoxic conditions using these strategies ( Fig. 1D and S2C). This indicates that the mir-35 family is required for the hypoxic response of embryos and implies that a single mir-35 family member can rescue phenotypes of mir-35-41 mutant animals, in accordance to previous reports 24,27,28 . However, we wished to exclude the possibility that the hypoxia-induced embryonic lethality of mir-35-41 mutant embryos is due to general sensitivity of strains that exhibit a high embryonic lethality in ambient O 2 conditions. Therefore, we analyzed the hypoxic response of mutants of three unrelated genes (plk-1, spd-2 and mex-1) that exhibit high normoxic embryonic lethality. We exposed embryos of these mutant strains to 0.5% O 2 and did not observe any significant hypoxia-induced enhancement of embryonic lethality ( Figure S3B, Table S3). These data point to a specific role for the mir-35 family in the embryonic hypoxic response.
Hypoxic induction of the mir-35 family is independent of HIF-1. The mir-35 family is predominantly expressed during embryogenesis 24 . We confirmed this expression pattern by constructing a transgene using a 602 bp intronic promoter driving expression of yfp fused to a nuclear localization signal (NLS) (Fig. 2). This promoter region was sufficient to rescue the nDf50 hypoxic phenotype when driving the mir-35-41 cluster ( Figure S2). We detected YFP expression in virtually all cells starting from ∼ 20 cell embryos and persisting throughout mid-and late embryogenesis (Fig. 2). We also detected expression in L1 animals but the signal drops in later larval stages. Thus, the rescuing mir-35 intronic promoter drives expression throughout embryogenesis in most, if not all cells.
We have shown that the mir-35 family is required for hypoxic survival of embryos and is expressed throughout embryogenesis. Next, we asked whether the level of each mir-35 family member is regulated by hypoxia. We therefore subjected early embryos to 0.5% O 2 for 20 mins or 4 hrs, to test for acute and chronic responsiveness, and measured the levels of the mature sequence of each family member by quantitative real-time PCR (qRT-PCR). We found that the expression of the mature sequence of all eight members of the mir-35 family was induced by approximately 2-fold after 4 hrs but not after 20 mins (Fig. 3A-B). Whereas, expression of the Y62F5A.9 host gene is not induced ( Figure S4C) and expression of all three reference miRNAs (mir-34, mir-86 and mir-1829c) is stable in these conditions (geNorm M value < 0.5, Coefficiency of variation < 0.2). Interestingly, the expression of predicted primary and precursor forms of the mir-35-41 locus measured by a previously described qPCR-based method 29 , using primers flanking the middle member of the cluster (mir-38), is decreased after 4h in hypoxia ( Figure S4D). This suggests post-transcriptional regulation of this locus upon hypoxia. As previous studies in C. elegans have shown that HIF-1 stabilization and general downregulation of translation are important requirements for animals to survive hypoxic insults 25,30,31 , such regulation of the mir-35 family could be dependent on HIF-1 transcription. Therefore we subjected hif-1(ia4) mutant embryos to 4 hrs of 0.5% O 2 and quantified induction of the mir-35 family. We still detected induction of mir- [35][36][37][38][39][40][41][42] indicating that this is independent of HIF-1 ( Figure S4A). Previous work has shown that C. elegans embryos are more sensitive to hypoxia when they are exposed directly to low O 2 rather than in utero 32 . This prompted us to ask whether hypoxia-induced mir-35-42 induction may differ in utero. We therefore, subjected late L4 animals to 0.5% O 2 for 15 hrs and developing embryos were subsequently extracted from the gonad. Using qRT-PCR analysis we observed an approximate 4-fold induction in expression of all mir-35 family members except for mir-38 and mir-41 ( Figure S4B). These data suggests that post-transcriptional mechanisms within the mother regulate the differential expression of the mir-35 family under hypoxic stress. Taken together, these results show that embryonic expression of the mir-35 family is induced by hypoxia in a hif-1-independent manner and that mir-38 and mir-41 are differentially regulated by hypoxia in utero.
During our qRT-PCR analysis, we observed additional evidence that the mir-35 cluster is potentially post-transcriptionally regulated. We observed that the expression levels of the mature miRNAs differ between members of the mir-35 family. In particular, based on the measured Cq values, mature mir-38 and mir-41 are much less abundant than the other members, 117-and 215-fold less abundant respectively   [35][36][37][38][39][40][41] promoter in wild type embryos exposed to 21% O 2 (black bars) or 0.5% O 2 (orange bars) for 20 mins (E) or 4 hrs (F). Data are presented as means of at least 3 independent repetitions and error bars represent ± SD. Students t-test was used to assess for statistical significance. *p ≤ 0.05, **≤ 0.01, ***p ≤ 0.001, ****≤ 0.0001, n.s. = not significant.
(in accordance with previous reports 33 ) when compared to the average abundance of the other 6 members (Fig. 3C). This observation further supports the possibility that mir-38 and mir-41 are differentially processed after transcription of the locus.
We next asked whether hypoxic induction of the mir-35 family is due to enhanced transcription of the mir- [35][36][37][38][39][40][41] locus. To answer this question, we used a strain carrying an integrated gfp transgene driven by genomic region upstream of the mir-35 locus 34 . We subjected embryos of this strain to 0.5% O 2 for 4 hrs and found that the intensity of mir-35-41-promoter driven GFP was similar to that of embryos cultivated in normoxia (Fig. 3D). To confirm these data at the RNA level we subjected the same strain to 0.5% O 2 for 20 mins or 4 hrs and found that the level of gfp mRNA is not elevated in hypoxia in either condition (Fig. 3E-F). Taken together, these results suggest that the induction of the mir-35 family in hypoxia is controlled by a post-transcriptional mechanism such as increased protection by RNA binding proteins 35,36 . sup-26 is a potential mir-35-41 direct target. Although the involvement of mir-35 family in embryonic development has been observed in the past 24,37 , there is a lack of information as to the downstream regulatory targets of these miRNAs during embryogenesis. Using miRNA target prediction logarithms (mirSOM, TargetScan) we extracted a list of potential direct targets of the mir-35 family. We screened this list for genes that are potentially related to hypoxia based on published literature. We noticed that the human homolog (RBMS1) of one of the candidates, namely SUP-26, is involved in the brain ischemic response 38 . The 3'UTR of sup-26 contains a sequence complementary to the seed sequence of all the mir-35 family members. In addition, SUP-26 co-purifies with mir- [35][36][37][38][39][40][41][42] in embryos as part of miRISC 39 . SUP-26 is an RNA binding protein that acts in the sex determination pathway in C. elegans 40 .
As a potential target of the mir-35 family one would expect sup-26 to be expressed during embryogenesis, the period at which the mir-35 family predominates. We therefore generated a transgene using the sup-26 promoter to drive nuclear-localized YFP followed by the native sup-26 untranslated region (UTR). We detected ubiquitous expression of YFP throughout embryogenesis indicating that sup-26 is expressed in a common temporal and spatial window to that of the mir-35 family (Fig. 4).
To test the possibility that sup-26 is a direct mir-35 target we conducted sensor experiments in a heterologous tissue, the pharynx, after failing to do so in embryos due to specific transgene toxicity (data not shown). We drove expression of mir-35 in the pharynx in combination with a RFP reporter under the control of the unrelated unc-54 3'UTR and a GFP reporter under the control of the sup-26 3'UTR (wild type or mir-35 binding site mutated) (Fig. 5B-D). We found that mir-35 robustly downregulates GFP expression (sup-26 3'UTR) and not the RFP sensor (control 3'UTR) (Fig. 5B-D). Further, regulation via the sup-26 3'UTR is dependent on the predicted mir-35 binding site (Fig. 5B-D) strongly suggesting a direct interaction between mir-35 and the sup-26 3'UTR. Interestingly, a genetic interaction between these two molecules was recently observed independently by the Ambros group 41 .
Finally, we asked whether sup-26 plays a role in the hypoxic response of embryos as one may expect as a potential mir-35 family target. We therefore subjected two independent sup-26 loss-of-function alleles, gk403 and gk426, to embryonic hypoxia ( Figure S5). Both mutants exhibited the wild type level of embryonic lethality when exposed to 0.5% O 2 for 24 hrs, however, when we exposed embryos to chronic hypoxia (45 hrs), we observed a significant increase in embryonic lethality ( Figure S5 and data not shown). This increase was reduced to the wild type levels when the sup-26 product was reintroduced to the sup-26(gk403) mutant ( Figure S5B). These data suggest that sup-26 plays a role in the embryonic response to hypoxia. However, since reduced SUP-26 levels resulted in embryonic lethality under low oxygen conditions, we hypothesize that the mir-35 family negatively regulates sup-26 in order to keep its expression to a physiological level, but not to eliminate its expression, for optimal survival of the embryos in hypoxia. Thus, defective regulation of sup-26 levels may be one of the causes of hypoxic lethality exhibited by loss of the mir-35 family. Other mir-35 target genes are also very likely to be involved in this process since the embryonic lethality of the sup-26 mutants is not significantly increased by the milder oxygen conditions (0.5% for 24 hrs). miRNAs, such as miR-210 were previously shown to play a pivotal role in hypoxic response in mammals and other organisms by regulating key factors and also being regulated by hif-1 42 . Our study offers insights into the role of additional miRNA families in hypoxia in a highly attractive experimental system, the C. elegans embryo, and we hope that further work based on these findings will extend our current knowledge on the role of miRNAs in protecting tissues when challenged with decreased oxygen tension.

Methods
Strains used. All strains used in this study are listed in Table S4.

Molecular biology and transgenic lines.
The oligonucleotides sequences used in this study are listed in Table S5.
Embryonic lethality in hypoxia. See Supplementary information for details. Briefly, plates containing embryos were placed for 24 hrs at 0.5% O 2 (nitrogen balanced). After 24 hrs recovery in ambient O 2 conditions, embryonic lethality was scored ( Figure S1). For the in utero hypoxia experiment, mid-L4 animals were placed in 0.5% O 2 for 15 hrs and eggs were removed from mothers by bleaching followed by immediate RNA extraction. mRNA isolation and qRT-PCR analysis. miRNA qPCRs were performed as described in 42 . All qRT-PCR primers were designed using the on line tool described in 43 . See Supplementary information for further details.

Statistical analysis. See Supplementary information.
Microscopy. Animals were anaesthetized on 5% agarose pads using 20 mM NaN 3 . Images were taken using a fluorescence microscope and the Zen software (Zeiss, AXIO Imager M2).