Nuclear Transport of U1 snRNP in Somatic Cells: Differences in Signal Requh'ement Compared with Xenopus lae qs Ooeytes

The signal requirement for the nuclear import of U1 RNA in somatic cells from different species was investigated by microinjection of both digoxygenin-labeled wild type and mutant U1 RNA molecules and in vitro reconstituted U1 snRNPs. U1 RNA was shown to be targeted to the nucleus by a temperature-dependent process that requires the prior assembly of RNPs from the common proteins and the microinjected RNA. Competition in the cell between immunoaftinity-purified U1 snRNPs and digoxygeninlabeled U1 snRNPs reconstituted in vitro showed that the transport is saturable and should therefore be a mediated process. The transport of a karyophilic protein under the same conditions was not affected, indicating the existence of a U snRNP-specific transport pathway in somatic cells, as already seen in the Xenopus/aev/s ooeyte system. Surprisingly, the signal requirement for nuclear transport of U1 snRNP was found to differ between oocytes and somatic cells from mouse, monkey and Xenopus, in that the m3GGpppGcap is no longer an essential signaling component in somatic cells. However, as shown by investigation of the transport kinetics of m3GpppGand ApppG-capped U1 snRNPs, the m3GpppG-cap accelerates the rate of U1 snRNP import significantly indicating that it has retained a signaling role for nuclear targeting of U1 snRNP in somatic cells. Moreover, our data strongly suggest that cell specific rather than species specific differences account for the differential m3G-cap requirement in nuclear import of U1 snRNPs. the nuclear envelope, continual exchange of macromolecules between the nucleus and the cytoplasm takes place. For large components, such as proteins, RNAs, and RNA-protein complexes (RNPs), t this transport is signal mediated and saturable and hence a receptormediated process. (Feldherr et al., 1984; Goldfarb et al., 1986; Forbes, 1992). Up to now, the transport of proteins into the nucleus is the best understood of these processes (reviewed in Garcia-Bustos et al., 1991; Silver, 1991). It occurs in two separate steps: the initial binding of the protein to the nuclear pore complex, followed by an ATP-dependent translocation of the protein into the nucleus (Richardson et al., 1988; Newmeyer and Forbes, 1988). The information contained in nuclear location signals (NLS) allows the selective interaction of karyophilic proteins with import receptors (reviewed in Garcia Bustos et al., U. Fischer and J. Heinrich contributed equally to this work. K. van Zee's present address is Department of Horticulture, Oregon State University, Corvallis, Oregon 97331. Address all correspondence to R. Ltihrmann, Institfit ftir Molekularbiologie und Tumorforschung, Philipps-Universitat Marburg, Emil Mannkopff-Strasse 2, D-35037 Marburg, Germany. 1. Abbreviations used in this paper: m3G--cap, m3GpppG-cap; NLS, nuclear location signals; RNP, RNA-protein complexes. 1991; Silver, 1991; Nigg et al., 1991; Forbes, 1992). Some nuclear proteins bear signals composed of a single short basic sequence resembling the NLS of the SV-40 large T-antigen (Lanford and Butel, 1984; Lanford et al., 1986; Kalderon et al., 1984) while others contain a more complex bipartite NLS first defined in nucleoplasmin (Robbins et al., 1991). In addition to karyophilic proteins, RNA-protein complexes such as the spliceosomal U snRNPs U1, U2, U4, and U5 are a further major group of macromolecnles which are selectively targeted to the nucleus. The snRNAs U1-U5 are all transcribed by RNA polymerase II and share two structural motifs: the n~GpppG-cap structure (m~O-cap) containing the m2.2.7-trimethylguano sine (Reddy and Busch, 1988) and a single-stranded uridylic acid-rich sequence referred to as the Sin-binding site (Branlant et al., 1982; Liautard et al., 1982). Two classes of proteins bind to the individual snRNAs. One group of common (Sin) proteins designated B, B', D1, D2, D3, E, F, and G are present in all U snRNPs (Liihrmann et al., 1990). They bind to the Sm site of the snRNA forming the Sm core domain which is morphologically similar among all spliceosomal U snRNPs (Kasmer et al., 1990; Liihrmann et al., 1990). The second class is comprised of specific proteins which bind only to one particular U snRNA species (Liihrmann et al., 1990). © The Rockefeller University Press, 0021-9525/94/06/971/10 $2.00 The Journal of Cell Biology, Volume 125, Number 5, lune 1994 971-980 971 on July 8, 2017 jcb.rress.org D ow nladed fom The morphogenesis of the U snRNP particles requires the export of the nuclear encoded mTG-capped U snRNA to the cytoplasm followed by the association of these RNAs with the Sm proteins and hypermethylation of the mTG to the m3G-cap (DeRobertis, 1983; Mattaj and DeRobertis, 1985; Mattaj, 1988). The snRNP particle then returns to the nucleus. In Xenopus laevis oocytes the nuclear location signal of U1 snRNP is bipartite with the m3G-cap as one essential signaling component (Fischer and Liihrmann, 1990; Hamm and Mattaj, 1990). The second part of the signal is located within the Sm core domain but has not yet been precisely defined (Fischer et al., 1993). Competition studies revealed that U1 snRNP and karyophilic proteins do not compete for common transport factors and therefore follow different kinetic transport pathways (Michaud and Goldfarb, 1991, 1992; Fischer et al., 1993). Surprisingly, not all splieeosomal snRNAs have the same m3G-cap requirement for nuclear transport in oocytes. Whereas U1 and U2 cannot enter the nucleus without an intact m3G-cap, this structure has a much less pronounced influence on the transport of U4 and U5. The latter RNAs can enter the nucleus as ApppGcapped derivatives, although with reduced transport kinetics (Fischer et al., 1991). The differential requirement of the m3G-cap for UI and U5 transport is unlikely to be due to differences in the activity of the Sm core NLS of both snRNP types, since UI and U5 snRNPs compete for the same transport receptor. Rather the structure of the RNAs appears to be important for the requirement for an m3G-cap as an essential signaling component (Fischer et al., 1993; Jarmolowski and Mattaj, 1993). In the X./aev/s oocyte, the various snRNA molecules exhibit different signal requirements for nuclear transport. It may therefore be asked whether the requirements for transport for the individual snRNAs differ from one cell type to another. For karyophilic proteins, several cases have already been described where the NLS activity differs between cell types or stages of development (Slaviak et al., 1989; Standiford and Richter, 1992). For example, the adenovirus 5 EtA protein contains two NLSs, of which one is constitutively active while the other appears to be regulated during embryogenesis of X. /aev/s (Standiford and Richter, 1992). Among snRNAs, it has been reported that in somatic cells U2 RNA may be transported to the nucleus in an m3Gcap-independent manner after transfection into human 293 cells (Kleinschmidt and Pederson, 1990). However, this study left open both the possibility that the transfection procedure per se affected the transport process and the alternative interpretation that the accumulation in the nucleus of the RNA introduced was due to cell divisions that might have occurred in the course of their experiments. In this report we have studied the nuclear transport of U1 RNA molecules labeled with digoxygenin and U1 snRNP particles reconstituted in vitro. Both were microinjected into living somatic cells from different species. It is shown that the transport occurs in a temperature-dependent and saturable fashion, for which the assembly of the common proteins onto the U1 RNA is an essential prerequisite. The transport of U1 snRNP requires different limiting transport factors than those required for the transport of karyophilic proteins. The signal required to target the U1 snRNP-particle in somatic cells is shown to consist of the Sm core domain and the m3G-cap. However, in contrast to nuclear transport of U1 snRNP in the oocyte, the cap has no essential function, but accelerates the transport kinetics significantly. Materials and Methods In Vitro Transcription In vitro transcription of U snRNA genes. The clones XeUI, XeUIAD, and pSmII were linearized with BamH1 or in the case of pSmlI with Avai and used for in vitro generation of snRNAs. XeUI and XeU1AD RNAs were generated by T7-, and pSmII RNAs by SP6 transcription, respectively. In a typical 100-t~! transcription assay 10 ~g of linearized DNA template was incubated in buffer containing 40 mM Tris-HCi, pH 8.0, 5 mM MgCI2, 2 mM spermidine, 50 mM NaC1, 25 mM DTE, 0.1/~g//d BSA, 10 mM of the indicated cap dinucleotide, 2 mM each of GTP, ATP, and CTP, 1.5 mM UTP and 0.5 mM digoxygenin*UTP, and 1 U/~I polymerase. Transcription was allowed to proceed for 90 rain at 370C and stopped hereafter by adding 1 U//d R2~Ase free DNAse I and continuing the incubation for 15 rain at 37°C. RNA was phenul-extracted and precipitated with 3 vol of ethanol. Transcription efficiency was analyzed by separation of 2.5/~1 of the transcription assay in an 0.8% TAE agarose gel and visualizing the bands with 2 ~tg/mi ethidium bromide. The typical yield of one transcription reaction was between 5 and 30/tg RNA. In Vitro Reconstitution of Digoxygenin-labeled U snRNPs In vitro reconstitution of U snRNPs was carried out essentially as described in Sumpter et al. (1992) with the exception that digoxygenin-laheled instead of 32p-labeled U sn RNAs were used. In brief, snRNP proteins were isolated by incubation of 10 nag immunoaliinity purified U snRNPs with 30 ml DEAE cellulose (DE53; Whatman Laboratory Products Inc., Clinton, NJ) in buffer containing 150 mM KOAc, 140 mM NaCI, 5 rnM EDTA, and 0.5 mM DTE. After incubation for 15 min on ice and a further 15 rain at 37°C the DEAE-cellulose was pelleted by centrifugation. The supernatunt contained the snRNP proteins. The solution was dialyzed against buffer containing 20 mM Hepes-KOH, pH 7.9, 50 mM KCI, 5 mM MgC12, and 0.5 mM DTE. The native U snRNP proteins were concentrated using a Centriprep concentrator3 (Amicon Corp., Arlington Heights, IL) in the first concentration step and Centricon C3 (Amicon Corp.) in the second to obtain snRNP proteins in concentrations up to 5 mg/mi. The yield was typicaUy 1 mg native proteins from 10 nag U snRNP-particles. In vitro reconstitution was carried out by incubation of 0.5 ~g (,x,9 pmol) U snRNA with 5 ~g snRNP proteins in 3 ~tl reconstitution buffer for 30 rain at 30*(2 and 15 min at 37°C. The in vitro reconstituted U snRNP particles were microinjected without further purification. Cells and Microinjection Vero (African green monkey kidney) and 3T3 cells were grown in Dulbecco's modified Eagle medium (GIBCO BRL, Bethesda, MD) supplemented with 10% fetal calf serum, 5% CO2 at 37°C. Xenopus A6 cells were grown in L19 medium (GIBCO BRL) supplemented with I0% fetal calf serum, 5% CO2 at 26°C. For microinjection experiments Veto, A6, and 31"3 cells were plated on glass coverslips in 35 mm dishes with 3 mi medium and grown to 80% confluence. Microinjection was carried out as described by Graessmann et ai. (1980) using an Eppendorf micromauipulator 5170 and microinjector 5242 and Eppendoff Femrotips. The amount of RNA injected in a typical experiment varied between 1-5 × 105 molecules per cell. The injection volume was ,~200 fl per cell.