Alteration of 28S rRNA 2'- O-methylation by etoposide correlates with decreased SMN phosphorylation and reduced Drosha levels

Abstract

The most common types of modification in human rRNA are pseudouridylation and 2'-O ribose methylation. These modifications are performed by small nucleolar ribonucleoproteins (snoRNPs) which contain a guide RNA (snoRNA) that base pairs at specific sites within the rRNA to direct the modification. rRNA modifications can vary, generating ribosome heterogeneity. One possible method that can be used to regulate rRNA modifications is by controlling snoRNP activity. RNA fragments derived from some small Cajal body-specific RNAs (scaRNA 2, 9 and 17) may influence snoRNP activity. Most scaRNAs accumulate in the Cajal body - a subnuclear domain - where they participate in the biogenesis of small nuclear RNPs, but scaRNA 2, 9 and 17 generate nucleolus-enriched fragments of unclear function, and we hypothesize that these fragments form regulatory RNPs that impact snoRNP activity and modulate rRNA modifications. Our previous work has shown that SMN, Drosha and various stresses, including etoposide treatment, may alter regulatory RNP formation. Here we demonstrate that etoposide treatment decreases the phosphorylation of SMN, reduces Drosha levels and increases the 2'-O-methylation of two sites within 28S rRNA. These findings further support a role for SMN and Drosha in regulating rRNA modification, possibly by affecting snoRNP or regulatory RNP activity.

Keywords: Cajal body; Drosha; SMN; rRNA modification; snoRNP.

Fig. 1.

Etoposide treatment increases the 2′-O-methylation of specific sites within 28S rRNA. (A) Primer extension using untreated RNA and a reducing amount of dNTPs, showing the induction of stop/pause signals corresponding to the 2′-O-methylation of 28S rRNA 2388 and 2352 at low (5 μM and 2.5 μM) dNTP concentrations. (B) Low dNTP primer extension using RNA from control or etoposide treated cells to evaluate 2388 methylation. HeLa cells were treated with 9 μM etoposide for 48 h in this experiment. (C) Quantification of 2388 methylation, normalized to the 2352 signal, relative to control, showing an increase in the relative amount of 2388 methylation in response to etoposide (n=3 biological repeats, *P<0.05). (D) Low dNTP primer extension using RNA from control or 48 h etoposide treated cells to evaluate 3923 methylation. (E) Quantification of 3923 methylation, normalized to the 3904 signal, relative to control, showing an increase in the relative amount of 3923 signal with 24, 48 and 72 h etoposide treatment (24 h and 72 h, n=3 biological repeats; 48 h, n=4 biological repeats. *P<0.05). (F) Low dNTP primer extension using RNA from control or 48 h etoposide treated cells to evaluate 18S rRNA 484 methylation. No significant difference was detected (n=3 biological repeats).

Fig. 2.

Etoposide mediated increase of scaRNA and snoRNA. (A) Reverse transcriptase quantitative real-time PCR analysis of scaRNA9 in RNA from untreated or 24, 48 or 72 h etoposide (9 μM) treated RNA from HeLa cells. 5.8S rRNA was used as the normalizer and data are shown relative to control, which is set as 1 (n=3 biological repeats, *P<0.05). (B) Quantification of selected scaRNAs, snoRNAs and U2 snRNA from untreated RNA or RNA isolated from cells treated with etoposide for 48 h. 5.8S rRNA was used as the normalizer (n=3 biological repeats, *P<0.05). The data are shown relative to those obtained from untreated control RNA, which is set as 1. (C) Quantification of selected protein coding mRNA, including that from host genes which contain intron-encoded scaRNA9 (9-Host, CEP295) and scaRNA9-like (9-Like Host, EIF1AX). RNA from untreated control cells or cells exposed to 9 μM etoposide for 48 h was analyzed. 5.8S rRNA was used as the normalizer (n=3 biological repeats, *P<0.05). The data are shown relative to those obtained from untreated control RNA, which is set as 1. For A–C, error bars represent standard error about the mean. (D) Western blot of lysate obtained from untreated or 48 h etoposide treated cells. Antibodies to Drosha (top panel), coilin, SMN and beta-tubulin (bottom panel) were used.

Fig. 3.

Hypophosphorylation of SMN by etoposide. (A) Western blot analysis of lysate from untreated or etoposide treated (9 μM for 48 h) HeLa cells. The blot was probed with antibodies to SMN (bottom) and beta tubulin (top). A slight downward mobility shift is seen in lanes 2 and 4. The estimated molecular weight of SMN (40 kDa) and beta tubulin (55 kDa) is shown. (B) Western blot to detect SMN using lysate treated with alkaline calf intestinal phosphatase (CIP) (lane 2). (C) Migration and detection of SMN using Phos-tag gels, which provide greater resolution of phosphorylated proteins compared to conventional SDS-PAGE. Low or hypophosphorylated SMN is indicated in the A region. More phosphorylated SMN is indicated in the B region. CIP treatment (lane 1) increases the amount of SMN in the A region, consistent with dephosphorylation. (D) Quantification of the signal in the A region divided by the signal in the B region for each condition tested, with the A/B ratio from untreated lysate set to 1. Etoposide treatment increases the amount of SMN in the A region relative to that in the B region by more than twofold compared to lysate from untreated cells (n=4 biological repeats, *P<0.05 compared to untreated, **P<0.05 compared to etoposide).

Fig. 4.

Etoposide treatment induces gem formation and disrupts SMN interaction with coilin. HeLa cells were either untreated or treated for 48 h with 9 μM etoposide. The cells were then processed and SMN (red) coilin (green) and nuclei (DAPI, blue) were detected. Arrows indicate co-localization of SMN and coilin in CBs. Arrowheads denote gems, which are SMN foci lacking coilin. Double arrowheads mark coilin foci lacking SMN. (B) Co-IP of coilin by SMN is decreased by etoposide. Untreated or etoposide-treated lysate was subjected to IP with control (IgG, lanes 3 and 4) or SMN (lanes 5 and 6) antibodies. Complexes were recovered by protein G beads, which were then extensively washed, boiled, then run on SDS-PAGE followed by western transfer and detection of coilin (top) or SMN (bottom) using the appropriate antibodies. The input signal represents 2% of that used in the IP reactions.

Fig. 5.

Etoposide-mediated induction of 28S rRNA 2388 2′-O-methylation is reduced by okadaic acid. (A) Low dNTP primer extension assay to analyze 2388 methylation using RNA from untreated cells or cells treated with 9 µM etoposide, 2 nM okadaic acid, and etoposide (9 µM)+okadaic acid (2 nM) for 48 h. (B) Quantification showing that etoposide+okadaic acid treatment decreases the relative amount of 2388 methylation compared to etoposide alone (n=4, P<0.05, * compared to untreated, ** compared to etoposide).

Fig. 6.

Okadaic acid alters the dynamics of full-length scaRNA9 and the mgU2-30 fragment. HeLa cells were transfected with scaRNA9 pcDNA 3.1+ for 24 h. 10 nM okadaic acid was added 7 h after transfection. RNA isolated from untreated and okadaic acid treated cells was then subjected to SDS-PAGE and northern blotting. ScaRNA9 and the mgU2-30 fragment were detected using a DIG labeled probe. Quantification was conducted using these and additional data by dividing the mgU2-30 fragment signal by the full-length scaRNA9 signal for each condition. The mgU2-30/full length scaRNA ratio for untreated cells was then set as 1. Okadaic acid increases the relative amount of the mgU2-30 fragment by approximately 1.7-fold (n=4 biological repeats, *P<0.05).

Fig. 7.

Drosha interacts with SMN and influences the modification of 28S rRNA A2388 and G3923. (A) SMN is associated with the Drosha complex. HeLa cells were transfected with FLAG-DGCR8, followed by lysis in KCl lysis buffer and IP with FLAG antibody (Flag) or control mouse antibody (IgG). After complex capture with protein G beads, beads were washed three times with KCl lysis buffer, followed by SDS-PAGE and western transfer. The membrane was probed with antibodies to SMN (top), Drosha (middle) and FLAG (to detect FLAG-DGCR8, bottom). Input represents 4.5% of the lysate used in the IP reactions. (B) Co-IP of endogenous Drosha with SMN. HeLa RIPA lysate was incubated with control antibody or SMN antibody, followed by complex capture with protein G beads. Beads were washed extensively then boiled and run on a SDS-PAGE followed by western transfer and detection of Drosha (top panel) or SMN (bottom panel) using the appropriate antibodies. A faint signal corresponding to endogenous Drosha is seen in lane 3, indicating that SMN and Drosha can form a complex. Reprobing of the same blot with SMN verifies the specificity of the reaction. Input represents 2% of that used in the IP reactions. (C) Low dNTP primer extension to detect 2388 methylation in RNA isolated from control siRNA or Drosha siRNA treated cells. An adjusted image is also shown to more easily visualize the increase in 2388 signal in the Drosha knockdown lane. Quantification was conducted by normalizing the 2388 signal to the 2352 signal and setting the control ratio value as 1. Drosha knockdown increases the relative amount of 2388 methylation by approximately 1.4-fold (n=3 biological repeats, *P<0.05). (D) Low dNTP primer extension to detect 3923 methylation in RNA isolated from control siRNA or Drosha siRNA treated cells. Quantification was conducted by normalizing the 3923 signal to the 3904 signal and setting the control ratio value as 1. Drosha knockdown increases the relative amount of 3923 methylation by a very small, but statistically significant amount (n=10 biological repeats, *P<0.05). (E) Drosha protein is reduced by Drosha siRNA. A western blot is shown. HeLa cells were transfected with negative control or Drosha siRNA for 48 h. The membrane was probed with Drosha antibody followed by probing with an antibody to β-tubulin.