Methylation of Xenopus CIRP2 regulates its arginine- and glycine-rich region-mediated nucleocytoplasmic distribution

Abstract

Cold-inducible RNA-binding protein (CIRP) was originally found in mammalian cells as a protein that is overexpressed upon a temperature downshift. Recently, we identified a Xenopus homolog of CIRP, termed xCIRP2, as a major cytoplasmic RNA-binding protein in oocytes. In this study we found by yeast two-hybrid screening that the Xenopus homolog of protein arginine N-methyltransferase 1 (xPRMT1) interacted with xCIRP2. We found that an arginine- and glycine-rich region of xCIRP2, termed the RG4 domain, was a target of xPRMT1 for methylation in vitro. xCIRP2 expressed in cultured cells accumulated in the nucleus as does mammalian CIRP. Interestingly, the RG4 domain was necessary for nuclear localization of xCIRP2. RG4-mediated nuclear accumulation of xCIRP2 was diminished in the presence of transcription inhibitors, suggesting that nuclear localization of xCIRP2 was dependent on ongoing transcription with RNA polymerase II. Analysis of interspecies heterokaryons revealed that xCIRP2 was capable of nucleocytoplasmic shuttling and the RG4 domain functioned as a nucleocytoplasmic shuttling signal. Methylation by overexpressed xPRMT1 caused cytoplasmic accumulation of xCIRP2. Possible implications of the relationship between regulation of intracellular localization and multiple functions of xCIRP2 will be discussed.

Figure 1

xPRMT1 interacts with xCIRP2. (A) Interaction of xCIRP2 and clone 4.32 in yeast two-hybrid analysis. AH109 yeast cells were transformed with each plasmid and spread on the minimal SD medium lacking adenine, histidine, leucine and tryptophan. (B) cDNA cloning of xPRMT1. The deduced amino acid sequence of cloned xPRMT1 is aligned with previously reported amino acid sequences, namely, human PRMT1 (accession no. CAA71764) (40), C.elegans NM_075508 (NP_507909.1), D.melanogaster CG6554 (XP_082348), A.thaliana pam1 (AAL36326) and S.cerevisiae Hmt1p/Rmt1p (NP_009590) (41,65,66). Amino acids identical in all six sequences are indicated by asterisks.

Figure 2

In vitro arginine methylation of xCIRP2 protein by recombinant xPRMT1. (A) Physical interaction of xCIRP2 and xPRMT1 was analyzed by GST pull-down assay. One microgram of GST (lanes 1 and 2) or GST-xCIRP2 (lanes 3 and 4) was incubated with 2 µg of xPRMT1 in 100 µl of 20 mM Tris–HCl (pH 7.5), 100 mM NaCl without (lanes 1 and 3) or with (lanes 2 and 4) 80 µM AdoMet. Half of each reaction was analyzed by SDS–PAGE and the gel was stained with silver. One hundred nanograms of GST, GST-xCIRP2 and xPRMT1 proteins were electrophoresed in lanes 5, 6 and 7, respectively. (B) One microgram of recombinant HA-xCIRP2 protein (lanes 2, 4, 6 and 8) and 0.5 µg of xPRMT1 protein (lanes 3, 4, 7 and 8) were incubated in 20 µl of 20 mM Tris–HCl (pH 7.5) containing [³H]AdoMet. Proteins were separated by SDS–PAGE and stained with Coomassie Brilliant Blue (CBB staining). The gel was then dried and subjected to fluorography (Methylation). (C) One microgram of recombinant HA-xCIRP2 (lanes 1 and 5), HA-xCIRP2ΔRG4 (lanes 2 and 6), GST (lanes 3 and 7) and GST-RG4 (lanes 4 and 8) were methylated and visualized as in (B). (D) Schematic diagrams of recombinant xCIRP2 protein and its mutants used for in vitro methylation analysis. RNP1 and RNP2 are the conserved sequences in RRM (67).

Figure 3

Identification of the domain, which functions as a nuclear localization signal. (A) Fluorescent micrographs showing subcellular localization of GFP and the GFP-xCIRP2 derivatives. HeLa S3 cells were transfected with plasmid DNA to express GFP fusion xCIRP2 derivatives. The localization of the GFP fusion proteins was examined under a fluorescence microscope. (B) Schematic diagrams of xCIRP2 protein and its deletion mutants expressed in HeLa S3 cells. GFP is fused at the N-terminus of xCIRP2.

Figure 4

RG-rich domains function as nuclear localization signals. (A) Sequence alignment of RG-rich domains function as NLSs. Sequences shown correspond to amino acids 88–130 in xCIRP2 (25), 89–137 in human CIRP (21), 197–252 in yeast Nab2p (NAB35) (43) and 267–291 in human hnRNP A1 (M9) (8). Amino acids that are identical in at least three NLSs are indicated by asterisks. (B) Fluorescent micrographs show subcellular localization of the GFP-xCIRP2ΔRG4, GFP-Xenopus RG4 and GFP-human RG4. Schematic diagrams of the GFP-xCIRP2ΔRG4 and GFP-Xenopus RG4 are shown in Figure 3B.

Figure 5

Nucleocytoplasmic shuttling of xCIRP2. (A) xCIRP2 accumulates in the cytoplasm of cultured cells by treatment with transcriptional inhibitors. The GFP-xCIRP2 expression vector was transfected into HeLa S3 cells. The cells were treated with 10 µg/ml actinomycin D or 0.1 mM DRB at 37°C for 4 h, and then DRB was removed by changing the medium with fresh medium. The culture was further incubated at 37°C for 2 h. The subcellular localization of the GFP fusion proteins was analyzed by fluorescence microscopy. (B) xCIRP2 and the RG4 domain shuttle between nucleus and cytoplasm. Expression vectors of GFP-xCIRP2 (upper panels) and GFP-xRG4 (lower panels) were transfected into HeLa cells. Two days later, the cells were fused with NIH 3T3 cells to form heterokaryons and incubated in the medium containing 100 µg/ml of cycloheximide for 1 h. Heterokaryons formed between HeLa cells (human) and NIH 3T3 cells (mouse) were examined by florescence microscopy of GFP fusion proteins and nuclei stained with Hoechst 33342. The arrows indicate murine NIH 3T3 cells. ‘Phase’ panels show the phase contrast images of the heterokaryons.

Figure 6

xCIRP2 accumulates in the cytoplasm of HeLa S3 cells upon arginine methylation by xPRMT1. (A) The expression vector of GFP-xCIRP2-HA alone (left panel) or together with that of myc-tagged xPRMT1 (right panel) was transfected into HeLa cells. Forty-eight hours after transfection, cells were fixed and the myc-xPRMT1 was stained with Alexa Fluor 488-conjugated anti-mouse IgG. The subcellular localization of the GFP-xCIRP2-HA and myc-tagged xPRMT1 was analyzed by fluorescence microscopy. (B) HeLa cells were transfected with the expression vector of GFP-xCIRP2-HA without (lanes 1–3) or with (lanes 4–6) that of xPRMT1. Cytoplasmic fractions (C, lanes 1 and 4), nuclear extracts (NE, lanes 2 and 5) and nuclear pellets (NP, lanes 3 and 6) were prepared and GFP-xCIRP2-HA was detected by immunoblotting with anti-HA antibodies (top panel). The lower graph compares the amount of GFP-xCIRP2-HA in the cytoplasmic fractions and nuclear extracts. The amount of GFP-xCIRP2-HA in nuclear extracts either in the absence or presence of xPRMT1 was set to be 100%. (C) The cytoplasmic fractions were subjected to further fractionation by ultracentrifugation at 100 000 g. GFP-xCIRP2-HA in the supernatant (S100) and the pellet (P100) was detected by immunoblotting as in (B). (D) Cells were incubated in the medium containing 20 µM AdOx for 24 h prior to transfection and AdOx was included in the culture medium until cell fixation. Transfection was performed as in (A).