C-terminal tagging impairs AGO2 function

Abstract

MicroRNA-mediated gene silencing is a conserved mechanism of post-transcriptional gene regulation across metazoans. It depends on base pairing between small RNAs and mRNAs, and on protein complexes including the RNA-induced silencing complex (RISC), where Argonaute 2 (AGO2) plays a central role. A full understanding of RNA silencing requires reliable molecular tools to study AGO2 and RISC. Affinity tagging and antibody-based methods can introduce artefacts, and both the N- and C-terminal domains of AGO2 are critical for its function. While N-terminal tags are frequently used, and a recent study in mice showed altered activity in N-terminal HaloTag-AGO2 fusions, the consequences of C-terminal tagging remain underexplored. CRISPaint, a CRISPR-Cas9-based technique, enables endogenous C-terminal tag fusions without requiring homology arms. Using this system, we generated the first C-terminal HaloTag fusion of AGO2 (AGO2HALO) in human A549 cells. We found that the AGO2HALO fusion protein exhibits reduced binding with TNRC6A, with no effect on cell viability. However, it significantly impairs RNA cleavage, silencing activity, and nuclear localization. We further compared AGO2-EGFP and EGFP-AGO2 using transient transfection. N-terminally tagged AGO2 retained wild-type-like function and localization, while C-terminally tagged AGO2 was impaired in siRNA and miRNA silencing, nuclear import, and P-body localization. These results demonstrate that a C-terminal HaloTag compromises AGO2 functionality and is unsuitable for studying RISC biology. Our findings highlight the importance of validating tagging strategies to avoid misleading conclusions due to tag-induced functional defects. Pre-print, bioRxiv.

Keywords: AGO2; Argonaute; CRISPR-Cas9; HaloTag; RNAi; microRNAs; protein tagging.

Figure 1.

Creation and genomic characterisation of AGO2-HaloTag cell Lines(A) schematic of WT AGO2 and the C terminal AGO2-HaloTag fusion (including T2A site, puromycin resistant (PuroR) and Poly(A) sections) genotype to be generated by CRISPaint editing of A549 cells. Arrows indicate locations of forward and reverse primers designed to confirm editing. Blue = AGO2 WT last intron forward and reverse; Light Teal = HaloTag1 forward and reverse; Dark Teal = HaloTag2 forward and reverse.(B) agarose gel loaded with PCR products of A549 WT and two AGO2-HaloTag (AGO2-HaloTag C5 and AGO2-HaloTag C10) lines amplified with indicated combinations of AGO2 WT and HaloTag primers, as indicated in (A). Red arrows indicate gDNA containing HaloTag sequence which was purified and submitted for sequencing. Circled numbers 1–3 indicate gDNA containing C terminal non-HaloTagged AGO2 product which was purified and submitted for sequencing. (C) sequence (generated from TOPO-seq) alignments of WT and two AGO2-HaloTag clones at the AGO2-HaloTag junction. From several submitted TOPO clones, two variants of AGO2-HaloTag (one long and one short) were identified in AGO2-HaloTag cells. Asterisk (*) indicates STOP codon. (D) schematic to show known functionally important domains of AGO2, with a focus on C-terminal PIWI domain. CRISPaint mediated AGO2-HaloTag fusion generated a long and a short variant. (E) chromatograph of C terminal AGO2 sequence identified in WT, AGO2-HaloTag C5 and AGO2-HaloTag C10 cells (circled numbers 1–3 in (B)) showing the additional and premature STOP codon in both AGO2-HaloTag lines. (F) abundance of untagged AGO2 mRNA transcript in A549 WT, two UnTagged (UT C1 and UT C2) and two AGO2-HaloTag (AGO2-HaloTag C5 and AGO2-HaloTag C10) cells. AGO2 (non-HaloTagged) mRNA abundance normalized to B Actin mRNA abundance and made relative to levels in WT cells. Data represent mean ± SEM of experiments; n = 3 (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001). (G & H) Western blot of whole-cell lysates from A549 WT, two UnTagged (UT C1 and UT C2) and two AGO2-HaloTag (AGO2-HaloTag C5 and AGO2-HaloTag C10) cell lines probed with antibodies against AGO2, HaloTag, and beta Actin.

Figure 2.

Initial characterisation of AGO2-HaloTag cells Lines(A) doubling time of A549 WT, two UnTagged (UT C1 and UT C2) and two AGO2-HaloTag (AGO2-HaloTag C5 and AGO2-HaloTag C10) measured using IncucyteZoom over 108 hours. Data represent mean ± SEM; n = 3 (ns p > 0.05). (B) Representative Western blot of whole-cell lysates from indicated cell lines (harvested during log-phase) probed with antibodies against AGO1, AGO2, AGO3, AGO4, Vinculin. (C) Representative Western blot of whole-cell lysates from indicated cell lines (harvested during log-phase) probed with antibodies against DDX6, TNRC6A, LIMD1, and beta Actin. (D) densitometry of AGO1/2/3/4 (normalized to loading control (Vinculin)) in indicated cell lines. Data represents mean ± SEM; n = 3 (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001 according to one-way ANOVA test with Dunnett’s multiple comparisons). (E) densitometry analysis of drosha, exportin 5, Dicer, DGCR8, DDX6 andTNRC6A, (normalized to loading control (beta Actin or vinculin)) in indicated cell lines. Data represent mean ± SEM; n = 3 (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001).

Figure 3.

Characterising AGO2 interactions, AGO2 localization and miR-451a levels in AGO2-HaloTag clones. (A) endogenous co-immunoprecipitation of AGO2 with Dicer and TNRC6A in WT, UnTagged C1, UnTagged C2, AGO2-HaloTag C5 and AGO2-HaloTag C10 cells. Note the increased AGO2-Dicer and decreased AGO2-TNRC6A co-immunoprecipitation in AGO2^HALO cells compared to in WT and UnTagged cells. Histogram shows mean densitometry for TNRC6A IP normalized to AGO2 IP densitometry relative to WT A549 cells. Data represents mean of four independent repeats with * indicating p ≤ 0.05 according to one-way ANOVA test with Dunnett’s multiple comparisons. (B) immunoblots of nuclear and cytoplasmic fractions from WT. UnTagged C1, UnTagged C2, AGO2-HaloTag C5 and AGO2-HaloTag C10 cells. Whole cell lysate (WCL) from WT A549 was also blotted alongside. Beta tubulin and histone H3 were probed as loading controls for cytoplasmic and nuclear fractions, respectively. (C) MiR-451a abundance (normalized to U6 RNA and made relative to relevant WT) in indicated cell lines measured by RT-qPCR using the 2^–∆∆Ct method. Data represent mean ± SEM; n = 2 (*p ≤ 0.05 according to one-way ANOVA test with Dunnett’s multiple comparisons). (D) pri-miR-451 abundance (normalised to beta actin mRNA levels) in indicated cell lines measured by RT-qPCR using the standard curve method. Data represent mean ± SEM; n = 2 (ns = not significant according to one-way ANOVA test with Dunnett’s multiple comparisons).

Figure 4.

Creation of AGO2 knockout A549, firefly luciferase assays and comparison of N- and C-terminally tagged AGO2. (A) Generation of AGO2 knockout A549 cells. Immunoblot of two AGO2^−/− clones alongside two control clones for AGO2 and beta actin as a loading control. (B) Relative firefly/Renilla luciferase activity in indicated cell lines transfected with reporter plasmids expressing firefly and Renilla luciferase and a siRNA against firefly luciferase (esiFFLuc) or ‘non-targeting’ control (esiGFP). The ratio between firefly and Renilla luciferase activity was measured 24 h after transfection. Data represent mean ± SEM; n = 3 (ns p > 0.05, **** p ≤ 0.0001 according to two-way ANOVA test with Sidak’s multiple comparisons). (C) Plasmid constructs for expression of untagged AGO2, EGFP-AGO2 and AGO2-EGFP. EGFP-VO is a control. (D) Relative firefly/Renilla luciferase activity in AGO2^−/− A549 transfected with reporter plasmids, AGO2 plasmids and an siRNA against firefly luciferase (esiFFLuc) or ‘non-targeting’ control (scr). The ratio between firefly and Renilla luciferase activity was measured 24 h after transfection. Data represent mean ± SEM; n = 3 (ns p > 0.05; ***p ≤ 0.001, ****p ≤ 0.0001 according to one-way ANOVA test with Dunnett’s multiple comparisons). (E) Relative Renilla/firefly luciferase activity of a miR-100 targeted (T) reporter relative to a non-targeted (NT) reporter in AGO2^−/− A549 transfected with reporter plasmid, 15 nM miR-100 mimic and the indicated AGO2 constructs or EGFP-VO. Data represent mean ± SEM; n = 3 (ns p > 0.05; **p ≤ 0.01 according to one-way ANOVA test with Dunnett’s multiple comparisons). (F) Immunoblots of nuclear and cytoplasmic fractions from AGO2^−/− A549 transfected with EGFP-VO, AGO2 UT, EGFP-AGO2 or AGO2-EGFP. Beta tubulin and histone H3 were probed as loading controls for cytoplasmic and nuclear fractions, respectively. (G) Fluorescence microscopy images for AGO2^−/− A549 transfected with EGFP-VO, EGFP-AGO2 or AGO2-EGFP. Cell nuclei are stained with DAPI and merged images show DAPI and EGFP signal combined. Scale bars indicate 10 µm. (H) UnTagged C1 and AGO2-HaloTag C10 cells treated with 100 nM HaloTag-TMR ligand visualized using confocal microscopy.

Figure 5.

Structural insights of the C-terminal of AGO2 for an explanation of impaired function upon HaloTag insertion. (A) Schematic composition of AGO2 showing 7 main domains and motifs. (B) C-terminal residue A859 contributes to miRNA binding (PDB code: 4OLB). Residues shown in stick format and residue type and sequence number annotated. Dashed lines show inter-atom distances 5.0 Å. (C) Surface representation of AGO2 (4OLB) with domains coloured as in (A). Bound miRNA shown in spheres with 5ʹ-3ʹ direction indicated. The approximate location of the buried C-terminal residue A859 is indicated. (D) Surface representation of AGO2 (4OLB) showing sites of tryptophan binding and the N-terminal most residue (A22) seen in the electron density. Residues 1–21 were not observed in the data. (E) pLDDT values for predictions of native AGO2 (solid black line) and C-terminal halo-tagged AGO2 and relative solvent accessible surface area (QASA, Ų; red line) calculated using PDB entry 4OLB [63].