ALKBH5 is a mammalian RNA demethylase that impacts RNA metabolism and mouse fertility

Abstract

N(6)-methyladenosine (m(6)A) is the most prevalent internal modification of messenger RNA (mRNA) in higher eukaryotes. Here we report ALKBH5 as another mammalian demethylase that oxidatively reverses m(6)A in mRNA in vitro and in vivo. This demethylation activity of ALKBH5 significantly affects mRNA export and RNA metabolism as well as the assembly of mRNA processing factors in nuclear speckles. Alkbh5-deficient male mice have increased m(6)A in mRNA and are characterized by impaired fertility resulting from apoptosis that affects meiotic metaphase-stage spermatocytes. In accordance with this defect, we have identified in mouse testes 1,551 differentially expressed genes that cover broad functional categories and include spermatogenesis-related mRNAs involved in the p53 functional interaction network. The discovery of this RNA demethylase strongly suggests that the reversible m(6)A modification has fundamental and broad functions in mammalian cells.

Figure 1. ALKBH5 Catalyzes Demethylation of m⁶A-Containing ssRNA

(A) Proposed oxidative demethylation of m⁶A to adenosine in RNA by ALKBH5. (B) ALKBH5 (0.3 nmol) completely demethylated m⁶A in ssRNA (1 nmol) at pH 7.5 and 16°C overnight as revealed by HPLC analysis of the digested substrates. (C) Demethylation activity of m⁶A-containing ssRNA for the wild-type ALKBH5 (WTALKBH5) and ALKBH5 mutants, H204A and H266A, at pH 7.5 and 16°C for 12 hr. Error bars indicate ± SEM (n = 3). (D) Michaelis-Menten plot of the steady-state kinetics of ALKBH5-catalyzed demethylation of m⁶A in ssRNA at pH 7.5 at room temperature. Error bars indicate ± SEM (n = 3). See also Figure S1 and Table S1.

Figure 2. Demethylation of m⁶A in mRNA by ALKBH5 and Its Effects on Selected mRNA Splicing Factors in HeLa Cells

(A) Top panel, western blotting of ALKBH5 upon knockdown and overexpression of ALKBH5 in HeLa cells. GAPDH served as a loading control. Bottom panel, quantification of the m⁶A/A ratio in mRNA using LC-MS/MS. Both sets of data were assessed using Student's t test. Error bars indicate ± SEM (n = 8). (B) RNase A-sensitive subcellular localization of ALKBH5 (green) in nuclear speckles. RNA processing factors SM and SC35-pi (red) were used as a nuclear speckle marker. DAPI (blue): DNA staining. Scale bar: 10 μm. (C) ALKBH5-dependent nuclear speckle staining of SC35-pi. ALKBH5 deficiency led to diminished signal of SC35-pi (red), which could be restored by over-expression of GFP-tagged WT ALKBH5, but not by the empty vector or the inactive mutant H204A. DAPI (blue): DNA staining. Scale bar: 10 μm. Representative images from one of the three independent experiments were shown for (B) and (C). See also Figure S2.

Figure 3. ALKBH5-Mediated Demethylation Affects mRNA Export

(A) Poly(A)-RNA-FISH using oligo(dT)₅₀-digoxigenin probe was performed in control and ALKBH5-deficient cells. mRNA (red) was mainly detected in the nucleus of control cells while in the cytoplasm of ALKBH5-deficient cells. Overexpression of GFP-tagged WT ALKBH5, but neither the empty vector nor the inactive ALKBH5 H204A, could rescue the nuclear mRNA export induced by endogenous ALKBH5 depletion. Arrows indicate ALKBH5-deficient cells. (B) ALKBH5-dependent nuclear speckle staining of ASF/SF2. ALKBH5 deficiency led to diminished signal of the ASF/SF2 (red), which could be efficiently restored by overexpression of GFP-tagged WT ALKBH5, but not by the empty vector or H204A mutant. (C) Aberrant localization of SRPK1 in ALKBH5-depleted HeLa cells. SRPK1 (red) was mainly detected in the nucleus of control cells. It was relocalized to form dot-like signals in the cytoplasm after ALKBH5 knockdown. The native localization could be rescued by overexpression of GFP-tagged WT ALKBH5, but not by the empty vector or inactive H204A mutant. DAPI (blue): DNA staining. Scale bar: 10 mm. Representative images from one of three independent experiments were shown. See also Figure S3.

Figure 4. Gene Ontology and Enrichment Analysis of DEGs from HeLa RNA-Seq

A total of 2,740 DEGs at either gene or isoform levels were subjected to DAVID GO analysis. An enrichment map was constructed by using Cytoscape installed with the Enrichment Map plugin. Red node represents each enriched GO pathway (p < 0.01, FDR q < 0.05, overlap cutoff > 0.5). Node size is proportional to the total number of genes in each pathway. Edge thickness represents the number of overlapping genes between nodes. GO pathways of similar functions are sorted into one cluster, marked with circles and labels. Gene numbers in each cluster are labeled. See also Figure S4 and Tables S2, S3, and S4.

Figure 5. Spermatogenic Defects in Alkbh5-Deficient Mice

(A) Relative expression of Alkbh5 in different organs. The testis is the organ with the highest expression of Alkbh5 mRNA. Data are presented as an expression relative to the expression in heart, set to one, and normalized to Gapdh expression in each organ. Error bars indicate ± SEM (n = 3 individual mice, each with duplicate cDNA synthesis and triplicate RT-qPCR). (B) Overview of the Alkbh5-targeting strategy. (C) Left panel, representative testes from 12-week-old WT and Alkbh5^−/− males. Right panel, average testis weight (mg) from 12-week-old WT (117.4 ± 16.0 mg, n = 12) and Alkbh5^−/− (61.4 ± 15.5 mg, n = 20) males. Error bars indicate ± SEM. (D) Hematoxylin and eosin stained sections from WT and Alkbh5^−/− formalin-fixed and paraffin-embedded testes and cauda epididymis. Note the striking presence of round cells in the Alkbh5^−/− cauda epididymis. (E) Representative images of WT and Alkbh5^−/− spermatozoa after staining with hematoxylin and eosin. (F) Characterization of spermatozoa number and motility for WT and Alkbh5^−/−. Error bars indicate ± SEM (n = 4, wild-type, and n = 4, Alkbh5^−/−, individual mice). (G) Immunofluorescence staining reveals an increased primary:secondary spermatocyte ratio in Alkbh5^−/− as compared to WT testicular tubuli. Sertoli cell nuclei were stained bright green for H3K27me3. Pachytene-stage primary spermatocytes were stained with phosphorylated H2AX in red, where the foci represent sex bodies specific to the pachytene stage. Round spermatides (secondary spermatocytes) are seen as smaller round nuclei positioned further toward the lumen of seminiferous tubuli. The DNA was stained blue with DAPI. (H) Increased cell death through apoptosis in Alkbh5-deficient testes compared to WT was apparent from TUNEL staining of formalin-fixed and paraffin-embedded testicular sections. See also Figure S5.

Figure 6. Transcriptome Analysis of Mouse Testicular Cells

(A) Quantification of the m⁶A/A ratio in mouse testicular cells by LC-MS/MS. Data were assessed using Welch t test to account for difference between variances due to the increased variation between Alkbh5^−/− mice in the representation of spermatogenic cell. Error bars indicate ± SEM (n = 2 WT and n = 5 Alkbh5^−/− mice). (B) A simplified scheme showing the functional interaction network between p53 and Alkbh5 transcriptome involved in apoptosis. A total of 127 spermatogenesis-related, differentially expressed genes were applied to Cytoscape analysis (version 2.8.2) installed with Reactome FI plugin. A total of 78 genes fall into p53 functional interaction network via the connection of 60 linker genes. Circle, differentially expressed genes from RNA-seq of Alkbh5; diamond, linker. (C) RT-qPCR validation of spermatogenesis- and apoptosis-related genes and isoforms found to be differentially expressed in Alkbh5-deficient testicular cells compared to those in the WT testicular cells. Expression of all genes was normalized to Actb, and WT expression was set to 1. Error bars indicate ± SEM (n = 3–5 individual mice, each with triplicate RT-qPCR). See also Figure S6 and Tables S5 and S6.