Repression of AGO1 by AGO2 via let-7 microRNAs facilitates embryonic stem cell differentiation

AGO2 represses AGO1 expression in mESCs

The recent observations of opposing functions of AGO1 and AGO2 in regulating pluripotency and differentiation in mESCs led us to examine whether there are mutual regulations between AGO1 and AGO2 (Liu et al. 2021a, 2024). We evaluated AGO1 and AGO2 protein levels in the wild-type (WT), AGO1Δ (knockout), and AGO2Δ mESCs by western blotting. Compared to those in the WT cells, AGO1 protein levels increased significantly in the AGO2Δ cells, which is consistent with previous observations (Ngondo et al. 2018), and AGO2 protein levels slightly elevated in the AGO1Δ cells (Fig. 1A). This result suggested that there are mutual regulations between AGO1 and AGO2 in mESCs. Here, we focused on the regulation of AGO2 on AGO1 due to the obvious (over twofold) increased AGO1 levels in the AGO2Δ cells.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

FIGURE 1.

AGO2 represses AGO1 in mESCs. (A) Western blotting in the WT, AGO1Δ, and AGO2Δ mESCs cultured in the 15% FBS + Lif medium. (B) Western blotting in the WT mESCs expressing a vector, WT AGO2, and an RNA-binding mutant AGO2 in the 15% FBS + Lif medium. Data information: In A and B, representative western blots are shown, and the quantifications represent the means (±SD) of three independent experiments. GAPDH levels were used for normalization in calculating the relative protein levels. One-way ANOVA was used to determine the significance of the difference, (*) P < 0.05.

To further test whether AGO2 regulates AGO1, we performed gain-of-function experiments. When an exogenous AGO2 was expressed in the WT mESCs, AGO1 levels decreased significantly (Fig. 1B), indicating that AGO2 can repress AGO1 expression in mESCs. To examine whether this repression by AGO2 is dependent on the small-RNA pathways, we performed the same ectopic expression experiment using a mutant AGO2, in which mutations in the conserved PAZ domain abolish AGO2's binding to small RNAs (Ma et al. 2004). Although expressed at a similar level as the WT AGO2, the mutant AGO2 could not decrease AGO1 levels in the WT mESCs (Fig. 1B). Importantly, in either the WT mESCs with ectopically expressed AGO2 or the AGO2Δ mESCs, neither AGO1 mRNA levels nor protein stabilities were significantly altered (Supplemental Fig. S1), arguing that the changes in AGO1 protein levels were not due to protein stability alterations from the potential availability of free miRNAs (Johnston et al. 2010; Winter and Diederichs 2011; Martinez and Gregory 2013). These results collectively suggest that AGO2 represses AGO1 expression in mESCs via its associated small RNA(s), possibly miRNAs (e.g., translational repression).

Let-7 miRNAs can repress AGO1 expression in mESCs

To examine which miRNAs are involved in repressing AGO1 in mESCs, we identified the conserved miRNA-binding sites in the 3′ UTR of AGO1 mRNA using TargetScan (Agarwal et al. 2015) and then evaluated the expression levels of the corresponding miRNAs in mESCs using published small-RNA-seq data sets (Liu et al. 2021a). This analysis revealed that AGO1 mRNA's 3′ UTR contains a highly conserved binding site for let-7 miRNAs, which are expressed in mESCs. Let-7 miRNAs are a group of highly conserved miRNAs that promote stem cell differentiation by repressing numerous stemness factors (Bussing et al. 2008). Previous studies in human somatic cell lines show that the let-7 miRNAs can regulate AGO1 through a unique mechanism: inducing translational readthrough by binding to the let-7 site in the 3′ UTR of AGO1 mRNA, generating a C-terminus extended AGO1 (Singh et al. 2019). However, this specialized mechanism is not present in mESCs, as mESCs ribosome profiling data sets indicate that ribosomes terminate at the predicted stop codon, but not the first in-frame stop codon in the 3′ UTR of AGO1 mRNA (Supplemental Fig. S2). Moreover, cross-linking followed by immunoprecipitation and qRT-PCR indicated that AGO2 binds to AGO1 mRNA in a let-7 miRNA-dependent fashion (Supplemental Fig. S3 and see below). Altogether, these observations argue that AGO1 is regulated by let-7 miRNAs through the canonical miRNA mechanism in mESCs.

Two lines of evidence indicated that let-7 miRNAs can repress AGO1 expression in mESCs. First, when let-7a was ectopically expressed in mESCs (Fig. 2A), AGO1 protein levels decreased, which was similar to the evolutionarily conserved let-7 target, LIN28a (Fig. 2B). Second, we generated let-7Δ mESCs using genome editing, in which most of let-7 miRNA family members were specifically knocked out or knocked down (Fig. 2C). Again, similar to LIN28a (Rybak et al. 2008), AGO1 protein levels also increased in the let-7Δ mESCs compared to those in the WT mESCs (Fig. 2D). To examine whether the let-7 miRNAs are involved in the AGO2-mediated repression of AGO1, we ectopically expressed AGO2 in the let-7Δ mESCs. Different from that in the WT mESCs, the ectopically expressed AGO2 failed to repress AGO1 in the let-7Δ mESCs (Fig. 2E). These results argued that AGO2 represses AGO1 via the let-7 miRNAs in mESCs.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

FIGURE 2.

Let-7 miRNAs can repress AGO1 expression in mESCs. (A) Ectopic expression of let-7a in mESCs. The let-7a levels were determined by qRT-PCR, and U6 RNA was used for normalization. (B) Western blotting in the WT mESCs expression a vector and let-7a miRNAs. (C) qRT-PCR quantification of miRNAs in the WT and the let-7Δ mESCs. U6 RNA was used for normalization. (D) Western blotting in the WT and the let-7Δ mESCs. (E) Western blotting in the WT and the let-7Δ mESCs expressing either a vector or AGO2. Data information: In B, D, and E, representative western blots are shown, and the quantifications represent the means (±SD) of three independent experiments. GAPDH levels were used for normalization in calculating the relative protein levels. One-way ANOVA was used to determine the significance of the difference, (*) P < 0.05, (n.s.) not significant (P > 0.05).

AGO1 is repressed by let-7 miRNAs in mESCs

An important caveat in interpreting the results from let-7 overexpression (Fig. 2A) and let-7 knockout (Fig. 2D) is the potential pleiotropic effects. Let-7 miRNAs regulate hundreds of mRNAs. Overexpression or knockout let-7 miRNAs alters numerous let-7:mRNA interactions, making it difficult to determine whether the altered AGO1 protein levels are a primary effect of let-7/AGO2-mediated regulation or caused by secondary effects from altered regulations of other let-7 targets.

To address this issue and to specifically determine whether AGO1 is repressed by let-7/AGO2 in mESCs, we mutated the conserved let-7 miRNA binding site in the 3′ UTR of AGO1 mRNA into a restriction enzyme site using genome editing (Fig. 3A). Thereby, we could identify mESCs with biallelic mutations using PCR followed by restriction enzyme digestion (Fig. 3B), and then we verified the mutations in the 3′-UTR mutant mESCs using Sanger sequencing (Fig. 3C). In the 3′-UTR mutant mESCs, the levels of let-7 miRNAs were not significantly altered compared to those in the WT mESCs (Fig. 3D), whereas the AGO1 protein levels were increased (Fig. 3E). Moreover, in contrast to the results in the WT mESCs, ectopic expression of AGO2 in the 3′-UTR mutant mESCs did not significantly decrease AGO1 levels (Fig. 3F). Collectively, these results argue that AGO1 is repressed by AGO2 via the let-7 miRNAs in mESCs.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

FIGURE 3.

AGO1 is regulated by let-7 miRNAs in mESCs. (A) Mutating the let-7 miRNA binding site in AGO1 mRNA 3′ UTR via genome editing. (B) Genotyping of the AGO1 3′-UTR mutant. The PCR was performed using the oligos (F and R) indicated in A. (C) Sanger sequencing to verify the biallelic mutation of the let-7 miRNA-binding site in AGO1 mRNA 3′ UTR. The red box indicates the let-7 miRNA-binding site. (D) qRT-PCR quantification of miRNAs in the WT and the AGO1 3′-UTR mutant mESCs. U6 RNA was used for normalization. The results represent the means (±SD) of three independent experiments. One-way ANOVA was used to determine the significance of the difference, (n.s.) not significant (P > 0.05). (E) Western blot in the WT and the AGO1 3′-UTR mutant mESCs cultured in the 15% FBS + Lif medium. (F) Western blotting in the WT and the AGO1 3′-UTR mutant mESCs expressing either a vector or AGO2. Data information: In E and F, representative western blots are shown. All the quantifications represent the means (±SD) of three independent experiments. GAPDH levels were used for normalization in calculating the relative protein levels. One-way ANOVA was used to determine the significance of the difference, (*) P < 0.05, (n.s.) not significant (P > 0.05).

Repression of AGO1 by let-7 miRNAs is important for mESC differentiation

To determine the biological significance of let-7/AGO2-mediated repression of AGO1 in mESCs, we examined the self-renewal and differentiation of mESCs using the colony formation assay (Mulas et al. 2019) and the exit pluripotency assay (Betschinger et al. 2013), respectively. We first evaluated the let-7Δ mESCs. Consistent with the pro-differentiation activity of let-7 miRNAs, the let-7Δ mESCs displayed increased self-renewal (Fig. 4A) and decreased differentiation (Fig. 4B). Moreover, the pluripotency markers, NANOG and OCT4, remained at higher levels in the differentiating let-7Δ mESCs than those in the differentiating WT mESCs (Fig. 4C), further indicating that the let-7Δ mESCs had defects in exit pluripotency.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

FIGURE 4.

let-7-mediated repression of AGO1 is important for mESC differentiation. (A) Colony formation assay for the WT and the let-7Δ mESCs. (B) Exit pluripotency assay for the WT and the let-7Δ mESCs. (C) Western blotting in the differentiating WT and the let-7Δ mESCs. (D) Colony formation assay for the WT and the AGO1 3′-UTR mutant mESCs. (E) Exit pluripotency assay for the WT and the AGO1 3′-UTR mutant mESCs. (F) Western blotting in the differentiating WT and AGO1 3′-UTR mutant mESCs. Data information: In A–B and D–E, the colony morphology and AP intensity were evaluated via microscopy. In total, 100–200 colonies were examined each time to determine the percentage of undifferentiated colonies. In C and F, representative western blots are shown. All the quantifications represent the means (±SD) of three independent experiments. GAPDH levels were used for normalization in calculating the relative protein levels. One-way ANOVA was used to determine the significance of the difference, (*) P < 0.05.

To determine whether let-7/AGO2-mediated repression of AGO1 contributes to the differentiation defects in the let-7Δ mESCs, we examined the AGO1 3′-UTR mutant mESCs. Although to a lesser extent, the 3′-UTR mutant mESCs also had elevated self-renewal (Fig. 4D), compromised differentiation (Fig. 4E), and higher pluripotency markers in differentiating cells (Fig. 4F). These results strongly argue that repressing AGO1 by AGO2 via let-7 miRNAs is important for the proper differentiation of mESCs.

In sum, we show that AGO1 is repressed by AGO2 and let-7 miRNAs in mESCs, and this repression is important for the proper differentiation of mESCs. Among the four AGOs in mammals, mESCs express AGO1 and AGO2, where AGO1 protein levels are ∼2%–5% of those of AGO2 (Liu et al. 2021b; Sala et al. 2023). Consistent with this relative protein abundance, AGO2Δ inhibits global miRNA activity, whereas AGO1Δ does not (Liu et al. 2021a, 2024). Thus, at the endogenous level, AGO2 is the major AGO protein in the miRNA pathway in mESCs. AGO2Δ mESCs and mESCs with mutations in the key components (e.g., DICER, DGCR8) of the miRNA biogenesis pathway all displayed defects in exit pluripotency and differentiation (Kanellopoulou et al. 2005; Murchison et al. 2005; Wang et al. 2007; Ngondo et al. 2018), implying that AGO2 and the global miRNA activity promote differentiation in mESCs. Recent studies suggest that this pro-differentiation activity of AGO2 is largely mediated by the pro-differentiation let-7 miRNAs (Liu et al. 2024). In contrast to AGO2, our recent study showed that the endogenous AGO1 functions in a small RNA-independent manner by modulating protein folding to maintain stemness in mESCs (Liu et al. 2021a). Altogether, these previous results suggest that the two AGOs expressed in mESCs have opposing roles in controlling stem cell fate: AGO2 promotes differentiation, whereas AGO1 maintains self-renewal. The results from this study show that AGO1 mRNA is a functional target of the conserved pro-differentiation let-7 miRNAs and AGO2. This finding reveals a regulatory mechanism between the two AGO paralogs with opposing functions in controlling stem cell fate decisions. In human, different expression levels of AGO1 and AGO2 can predict distinct disease outcomes (Pantazopoulou et al. 2020). For instance, in renal cancer, high AGO2 protein levels are associated with unfavorable patient outcomes, whereas high AGO1 protein levels are associated with favorable outcomes (Liu et al. 2024). Thus, we speculate that the repression of AGO1 by let-7 miRNAs and AGO2 may also occur during the pathogenesis in human cells.