Impact of age-associated increase in 2'-O-methylation of miRNAs on aging and neurodegeneration in Drosophila

Abstract

MicroRNAs (miRNAs) are 20- to ∼24-nucleotide (nt) small RNAs that impact a variety of biological processes, from development to age-associated events. To study the role of miRNAs in aging, studies have profiled the levels of miRNAs with time. However, evidence suggests that miRNAs show heterogeneity in length and sequence in different biological contexts. Here, by examining the expression pattern of miRNAs by Northern blot analysis, we found that Drosophila miRNAs show distinct isoform pattern changes with age. Surprisingly, an increase of some miRNAs reflects increased 2'-O-methylation of select isoforms. Small RNA deep sequencing revealed a global increase of miRNAs loaded into Ago2, but not into Ago1, with age. Our data suggest increased loading of miRNAs into Ago2, but not Ago1, with age, indicating a mechanism for differential loading of miRNAs with age between Ago1 and Ago2. Mutations in Hen1 and Ago2, which lack 2'-O-methylation of miRNAs, result in accelerated neurodegeneration and shorter life span, suggesting a potential impact of the age-associated increase of 2'-O-methylation of small RNAs on age-associated processes. Our study highlights that miRNA 2'-O-methylation at the 3' end is modulated by differential partitioning of miRNAs between Ago1 and Ago2 with age and that this process, along with other functions of Ago2, might impact age-associated events in Drosophila.

Keywords: 2′-O-methylation; aging; miRNA; miRNA sorting; neurodegeneration.

Figure 1.

Nbr-dependent miRNAs show distinct isoform patterns with age. Northern blots of different Nbr-dependent miRNAs with age. (A,B, left panel) miR-34-5p and miR-317-3p showed accumulation of short isoforms with age. (B, right panel) Both miR-34-5p and miR-317-3p also increased in total amount with age. (C,D, left panel) miR-305-5p and miR-263a-5p showed accumulation of long isoforms with age. (D, right panel) Both miR-305-5p and miR-263a-5p increased in total amount with age. (E,F) miR-11-3p, a single-form miRNA by Northern, accumulated with age. (B,D, left panels, F) Quantification of the different isoforms of each miRNA. (Right panels) Quantification of the total amount of each miRNA with age. Red arrowheads and arrows indicate the isoforms that increased with age. Mean ± SD. n = 3; (*) P < 0.05 (Student's t-test).

Figure 2.

Age-associated increase of long isoforms of miR-305, miR-263a/b, and miR-11 is associated with increased protection from oxidation/β-elimination. Northern blots of miR-305, miR-263a/b, miR-11, miR-34, and miR-317 without (left side of each miRNA panel) and with (right side of each miRNA panel) treatment by oxidation/β-elimination, with quantitation. (A) All isoforms of miR-34 were sensitive to oxidation/β-elimination. The longest isoform of miR-317 was protected (arrowheads), but there was a decrease in the ratio of the protected isoform with age. miR-305, miR-263a/b, and miR-11 show accumulation of a protected long isoform with age (arrows). (B) Ratio of oxidation/β-elimination-protected isoforms to unprotected isoforms with age. Mean ± SD. n = 3; (*) P < 0.05 (Student's t-test).

Figure 3.

The age-associated increase of the long isoforms of miR-305, miR-263a, and miR-11 is eliminated upon Hen1 and Ago2 mutation. Northern blots of each miRNA with or without oxidation/β-elimination at 3 d and 30 d in wild-type, Hen1^f00810, and Ago2^BL16608 animals. Red arrowheads indicate the isoforms that are protected after oxidation/β-elimination in wild type; these forms for miR-305, miR-263a/b, and miR-11 were no longer protected in Hen1^f00810 and Ago2^BL16608 mutant animals. The long forms of miR-317 remained protected in Hen1^f00810 and Ago2^BL16608 animals, indicating an alternative mechanism by which the long isoform accumulates for this miRNA.

Figure 4.

Increased Ago2 loading of long miRNA isoforms with age. Northern blots with quantitation of miRNAs associated with Ago1 versus Ago2 with age. (A) Northern blots for small RNAs on RNA isolated from Ago1-IP (left) and Ago2-IP (Flag-HA-Ago2) (right) . (B) Quantification of esi-2.1 level with age. Mean ± SD. n = 4; Student's t-test confirmed no significant differences with age. There is no change in esi-2.1 levels with age. (C–E) Quantification of miRNA isoforms loaded into Ago2 with age, normalized to esi-2.1 in Ago2-IP. Mean ± SD. n = 3; (*) P < 0.05 (Student's t-test). For all three miRNAs, the Ago2-loaded isoforms increase with age. (F) Western immunoblot for Ago2 and R2D2 with age. (G) Quantification of Ago2 protein level with age. Mean ± SD. n = 3; not significant by Student's t-test. (H) Quantification of R2D2 protein level with age. Mean ± SD. n = 4; not significant by Student's t-test.

Figure 5.

Ago1-IP versus Ago2-IP small RNA deep sequencing with age. (A) Normalized read number of total miRNAs in Ago1 or Ago2 with age. (B) Percentage of miRNAs and other small RNA classes in Ago1 and Ago2 with age. (C) miR-305-5p, miR-263a-5p, and miR-11-3p show a decrease in Ago1 but an increase of specific isoforms in Ago2 with age.

Figure 6.

Identification of Ago2-loaded miRNA isoforms whose loading increases with age. (A) Dot plot showing the ratio of preferentially Ago2-loaded isoforms at 3 d (X-axis) to 30 d (Y-axis). We selected miRNAs with values R₂₁ > 1.2 as candidates with greater loading into Ago2 at 3 d or 30 d (dashed lines parallel to the axes; miR-305-5p, miR-263a-5p, and miR-11-3p are seen) (full lists are in Supplemental Tables S2, S3). miR-34-18nt, miR-276a/b-5p-17nt, and miR-981-5p-21nt showed an extremely high R₂₁ ratio at 3 d and/or 30 d (>16) (Supplemental Tables S2, S3). This suggests much higher accumulation of these miRNA isoforms in Ago2-IP compared with Ago1-IP at 3 d and/or 30 d. This could happen because of biologically relevant up-regulation of these miRNA isoforms or cloning bias and degradation from the corresponding longer isoforms. These miRNA isoforms were removed from this figure because including these data obscured the distribution of the other data in the plot. miR-34-18nt was undetectable by Ago1-IP versus Ago2-IP Northerns. Therefore, most- likely, miR-34-18nt is an artifact of library preparation. (B) Fold change of the main isoform of preferentially Ago2-loaded miRNA isoforms with age (30 d/3 d). Red indicates miRNAs known from our Northern analyses that show an increase in Ago2 association with age. Black arrows here and green bars in Supplemental Figure 4 indicate abundant miRNAs that were tested for accumulation in Ago2 with age in C. (C) Northern analysis (top) and quantitation (bottom) of miRNAs highlighted by the analyses in B and Supplemental Figure 4.

Figure 7.

Mutations in Hen1 and Ago2 are associated with age-dependent brain degeneration and shorter life span. (A) Paraffin sections of wild-type, Hen1^f00810, Ago2^BL16608, Ago2⁴¹⁴/+, and Ago2⁴¹⁴ heads at 3 d and 30 d. Bar, 0.1 mm. The abundant vacuoles that are present in 30-d animals in the mutants are highlighted in yellow. (B) Quantification of brain vacuoles in lamina. (C) Quantification of brain vacuoles in retina. (D) Life span of wild-type, Hen1^f00810, and Ago2^BL16608 mutants. Mutant lines have been backcrossed into the genetic background of the control wild type (see the Materials and Methods). The life spans of the mutants are significantly different from wild type. P < 0.0001 for each compared with wild type; log rank analysis; χ² value with Hen1^f00810 is 327.7; χ² value with Ago2^BL16608 is 253.3.