RAM/Fam103a1 is required for mRNA cap methylation

Abstract

The 7-methylguanosine cap added to the 5' end of mRNA is required for efficient gene expression in eukaryotes. In mammals, methylation of the guanosine cap is catalyzed by RNMT (RNA guanine-7 methyltransferase), an enzyme previously thought to function as a monomer. We have identified an obligate component of the mammalian cap methyltransferase, RAM (RNMT-Activating Mini protein)/Fam103a1, a previously uncharacterized protein. RAM consists of an N-terminal RNMT-activating domain and a C-terminal RNA-binding domain. As monomers RNMT and RAM have a relatively weak affinity for RNA; however, together their RNA affinity is significantly increased. RAM is required for efficient cap methylation in vitro and in vivo, and is indirectly required to maintain mRNA expression levels, for mRNA translation and for cell viability. Our findings demonstrate that RAM is an essential component of the core gene expression machinery.

Graphical abstract

Figure 1

Fam103a1/RAM Isolated in Human RNMT Complexes (A) RNMT was detected by western blot in extracts from 293 cell lines expressing HA-RNMT and vector control. β-Tubulin (Tub) was detected as a loading control. (B) HA-RNMT complexes were purified using anti-HA antibodies, resolved by SDS-PAGE and stained with Coomassie Blue. Migration of HA-RNMT, antibody heavy chain (HC), light chain (LC) and Fam103a1/RAM are indicated. (C) Recombinant RNMT (Input) was mixed with GST and GST-RAM, and affinity purified on glutathione agarose. Proteins eluted were analyzed by western blot to detect RNMT, RAM, and GST. (D) Immunoprecipitations were performed on HeLa cell, primary T lymphocyte, and SAOS-2 cell extracts, using the antibodies indicated. Western blots were performed to detect RAM and RNMT in inputs, immunoprecipitates (IP), and immunodepleted extracts (ID extracts). (E) Gel filtration on a Superdex s200 10/30 column was used to resolve 1 mg HeLa cell extract, 1 μg recombinant RNMT, 1 μg RAM, and 1 μg of a 1:1 mixture of RNMT and RAM; 0.5ml fractions were collected following the void volume. Western blots were performed to detect RAM and RNMT. The migration of standards is indicated. (See also Figure S1.)

Figure 2

RAM Is Conserved in Vertebrates (A) The amino acid sequence of the H.sapiens RAM protein (NP_113640.1) and homologs in M. musculus (NP_080273.1), M. domestica (XP_001362351.1), O. anatinus (XP_001513424), T. guttata (XP_002199043.1), and X. tropicalis (NP_001037960.1) were aligned using EMBL-EBI ClustalW2 Multiple Sequences Alignment software, using the default parameters (Chenna et al., 2003). Amino acids identical in H.sapiens RAM protein and at least one other species are highlighted in gray, and those identical in all species investigated are indicated (^∗). (B) Immunofluorescence microscopy was used to detect RAM expression in HeLa cells. Cells were transfected with two independent siRNAs directed against RAM and a nontargeting control for 48 hr prior to fixation to confirm specificity of RAM staining. DAPI stain was used to detect nuclei.

Figure 3

RAM Is a RNA-Binding Protein and Promotes Cap Methylation (A) RAM and RNMT interaction with RNA was investigated by RNA band shift assay. 2 pmol RNMT and/or 0.5–2 pmol RAM were incubated with an excess of ³²P-capped transcript, and complexes were resolved by gel electrophoresis. (B) As in (A) except prior to gel electrophoresis, 2 pmol of RNMT and/or 2 pmol RAM were incubated with 700 ng anti-RNMT, GST, or RAM antibodies. The position of RAM and RNMT-RAM complexes is indicated. Assays were repeated three times, and a representative result is shown. (C) Cap methyltransferase assay was performed using a titration of recombinant RNMT, molarity indicated, or no protein was added (C). RNMT was incubated with ³²P-capped transcript and s-adenosyl methionine. Following the reaction, GpppG and m7GpppG were resolved by thin layer chromatography, as indicated. (D) The cap methyltransferase assay was performed with 20 nM RNMT and/or 20 nM RAM. (E) As in (D), except a titration of RAM was used. (F) Cap methyltransferase assay was performed on cellular HA-RNMT immunoprecipitated via the HA tag (lanes 3–5), or the same immunoprecipitation was performed on control cell extracts (lane 2). Prior to the assay, immunoprecipitates were incubated with anti-GST (lane 4) or anti-RAM antibodies (lane 5). (G) Cap methyltransferase assay was performed on 1 μg 293 cell extract (lanes 2–4). Prior to the assay, extracts were incubated with anti-GST (lane 3), or anti-RAM antibodies (lane 4). For Figures 3D–3G, mean relative cap methylation for four independent experiments and standard deviation is depicted (left panels). (See also Figure S2.)

Figure 4

RAM N Terminus Activates RNMT (A and B) 293 cells were transfected with combinations of pEGFP-RAM and pCDNA-HA-RNMT, deletion mutants, and vector controls. Immunoprecipitations were performed with anti-HA and anti-GFP antibodies. Western blots were performed to detect RAM and RNMT in inputs, anti-HA antibody immunoprecipitates, and anti-GFP antibody immunoprecipitates. (^∗) Indicates a cross-reacting band. (C) Cap methyltransferase assay was performed as in Figure 3D using 20 nM RNMT plus 20 nM GST or GST-RAM protein. (D) RNA band shift assay was performed as in Figure 3A using 2 pmol GST or GST-RAM protein. (E) Cap methyltransferase assay was performed as in (C) except for lane 6, in which RNMT was incubated with GST-RAM FL and 56–118 (all 20 nM). (F) Summary of RAM domain analysis. Deletion mutants used and their activity in RNMT binding, RNA binding, and RNMT activation are depicted. For Figures 4C and 4E, mean relative cap methylation for four independent experiments and standard deviation is depicted. (See also Figure S3).

Figure 5

RAM Is Required for Cellular RNMT Expression and Cap Methylation Expression of RAM was reduced in HeLa cells by transfection of two independent siRNAs (1 or 2) or control (c) siRNA for 48 hr. (A) Immunofluorescence microscopy was used to detect RNMT expression and DAPI stain was used to detect nuclei. (B) Western blots were performed to detect RAM, RNMT, and Tubulin. (C) Cells were transfected with pcDNA5-RAM (+) or pcDNA5 (c), 48 hr later they were transfected with RAM siRNA (1) or control (c), and 48 hr later they were lysed. Western blots were performed as in (B). (D) Relative cap methyltransferase activity was determined in cell extracts following siRNA transfection. (E) RAM expression was depleted by transfection of siRNA for 24 hr and RNMT expression was Doxycycline-induced (Dox) for the time course indicated. Western blots were performed as in (B). (F) Relative cap methyltransferase activity in cell extracts was determined following 24 hr RAM siRNA transfection and 2 hr RNMT induction (Dox). (G) Methyl cap levels on the four endogenous transcripts indicated was determined relative to total transcript level following 24 hr RAM siRNA transfection and 6 hr RNMT induction (Dox). (H) Cells were transfected with RAM siRNA, 24 hr later they were transfected with 0.1 ug (+) pcDNA5 RAM and RNMT, and 24 hr later they were lysed. Western blots were performed as in (B). (I) Relative cap methyltransferase activity was detected in cell extracts. (J) Methyl cap levels on the four endogenous transcripts indicated was determined relative to total transcript level. In charts, average result and standard deviation of at least three independent experiments are depicted. (See also Figure S4).

Figure 6

RAM Is Required for RNA Pol II Transcript Maintenance RAM siRNA (1) and controls (c) were transfected into HeLa cells for 24 hr and RNMT was induced with Doxycycline for 6 hr (+).RAM and RNMT transcripts (A) and c-Myc and RuvBL1 transcripts (B) were detected by real-time PCR. (C) 24 hr following RAM siRNA treatment 0.1 ug (+) pcDNA5 RAM and RNMT were transfected for 24 hr. RNA was harvested and the transcripts indicated were detected by real-time PCR. (D) Cells were prepared as in (C). Relative incorporation of ³⁵S-cysteine and ³⁵S-methionine into cellular proteins was determined. (E) Polysome profiles of cells transfected with RAM or control siRNA were determined. A representative result for five independent experiments is shown. (F) Cells were prepared as in (D). Cell counts were determined. (G) 48 hr following RAM siRNA, western blots were performed to detect full-length (FL) and cleaved (p89) PARP. In charts, average result and standard deviation of at least three independent experiments are depicted.

Figure 7

RAM Is Required for RNMT Translation and Stability (A) RAM and RNMT transcript level was determined by RT-PCR in HeLa cells transfected with two independent RAM-directed or control siRNAs, for 48 or 72 hr. (B) 293 cells were transfected with pcDNA3.1 Fg-RAM and HA-RNMT or relavent controls for 2 days. Western blots were performed to detect HA-RNMT (HA), RAM (^∗ indicates Fg-RAM), and Tubulin. RT-PCR was performed to detect Fg-RAM and HA-RNMT transcripts. (C) as (B), except pcDNA3.1 HA-RNMT-MTD (methyltransferase dead) replaced HA-RNMT. (D) RNMT or RNMT and RAM were in vitro translated in the same reaction. At the times indicated, RNMT protein levels were quantitated. At 30 min RNMT and RAM transcripts were quantitated by RT-PCR. Values were normalized to input DNA and the average and standard deviation of three independent experiments are depicted. (E) as (D), except HA-RNMT-MTD replaced HA-RNMT. (F) as (D), except RAM 56–118 replaced RAM.