Report

Inhibition of Ran Guanosine Triphosphatase-Dependent Nuclear Transport by the Matrix Protein of Vesicular Stomatitis Virus

See allHide authors and affiliations

Science  20 Jun 1997:
Vol. 276, Issue 5320, pp. 1845-1848
DOI: 10.1126/science.276.5320.1845

Abstract

Transport of macromolecules into and out of nuclei, essential steps in gene expression, are potential points of control. The matrix protein (M protein) of vesicular stomatitis virus (VSV) was shown to block transport of RNAs and proteins between the nucleus and cytoplasm of Xenopus laevis oocytes. The pattern of inhibition indicated that M protein interfered with transport that is dependent on the ras-like nuclear guanosine triphosphatase (GTPase) Ran-TC4 and its associated factors. This inhibition of nuclear transport by M protein explains several observations about the effects of VSV infection on host cell gene expression and suggests that RNA export is closely coupled to protein import.

Macromolecules are in constant flux between the nucleus and cytoplasm; RNAs are transported between their sites of synthesis, maturation, and function; and proteins are both imported into, and exported from, the nucleus (1). This intracellular traffic, essential to the expression of nuclear genes, is a potential target for control by cells and viruses (2, 3). We now show that a single protein of VSV can specifically shut down both RNA export and protein import, demonstrating a close coupling of nuclear transport in the two opposite directions.

RNAs and proteins are transported between the nucleus and cytoplasm by signal-mediated, energy-requiring processes that involve many components of the nuclear pore complex (NPC) (4) and a large number of soluble factors (1, 5). Export of RNA is dependent on bifunctional RNA binding proteins (such as the Rev protein of HIV-1) that recognize specific sequences in the RNA and interact with other export factors or components of the NPC (1, 3). RNA-binding proteins like Rev (3), CBP20, the 20-kD component of the 5′ cap binding complex (6) and heterogeneous nuclear ribonucleoprotein particle (hnRNP)-A1 (7) exit the nucleus with the RNA transport substrates; subsequently, these proteins are again imported (8) by means of soluble transport factors like importins-α and -β (5, 6) or homologs of importin-β (7). Both RNA export and protein import are blocked by the lectin wheat germ agglutinin (WGA) and by antibodies specific for proteins of the NPC (nucleoporins) (4, 8, 9).

Key components of protein import systems (1, 5, 7) are the GTPase Ran-TC4 (Ran), several Ran-binding proteins, and the Ran-specific effectors RCC1 (the nuclear GTP:GDP exchange factor for Ran) and RanGAP1 (the cytoplasmic Ran GTPase activating protein) (10, 11). Inactivation of this system blocks import of proteins (12) and small nuclear ribonucleoprotein particles (snRNPs) (13), and export of most RNAs with the notable exception of tRNAs (14, 15); movement of RNAs within the nucleus also is impaired (15). However, the precise function or functions of the Ran GTPase in RNA transport remains to be established (11).

Infection by VSV, an RNA virus that replicates in the cytoplasm, results in rapid inhibition of host cell gene expression (16). The earliest observed effect is the cessation of maturation of small nuclear RNAs (snRNAs or U RNAs) (17). We have proposed (15) that this block in snRNA metabolism results from inhibition of export of snRNA precursors out of the nucleus (18). The putative viral inhibitor of RNA export would likely be contained in the infecting virions, since the effects on snRNA metabolism are extremely rapid and do not require formation of functional viral mRNA (17). A prime candidate for such an inhibitor is the matrix (M) protein of VSV, present in 1600 to 1800 copies per virion (16); synthesis of this protein in transiently transfected cells interferes with nuclear gene expression (19, 20). We now present data showing that M protein can inhibit export of nuclear pre-snRNA and other cellular RNAs and import of proteins.

We studied the ability of M protein to block nuclear transport by synthesizing the protein in Xenopus laevis oocytes (21-24). After the oocytes were cytoplasmically injected with m7G-capped, polyadenylated transcripts coding for M protein (M mRNA), the protein accumulated both in the nucleus and the cytoplasm for up to 48 hours, as assayed either by immunoblotting (Fig.1A) or by direct labeling of injected oocytes with [35S]methionine (25). Differences in the amounts of M protein in the two compartments reflect, at least in part, the approximately 10-fold difference between volumes of the nucleus and cytoplasm. After about 20 hours of synthesis of M protein, we assayed for RNA export by injection of various DNA templates or RNAs made in vitro (21).

Figure 1

Inhibition of export of small RNAs in X. laevis oocytes by VSV M protein. (A) Synthesis of VSV M protein in X. laevisoocytes injected with M mRNA. ∼15 ng of m7G capped, polyadenylated VSV M protein mRNA (lanes 3 to 8) or the complementary antisense sequence (lanes 1 and 2) were injected into the cytoplasms of X. laevis oocytes. After incubation at 18°C for the times indicated, total nuclear and cytoplasmic proteins (one oocyte equivalent, each) were assayed by immunoblots probed with the VSV M protein-specific monoclonal antibody 23H12. (B) Inhibition of nuclear export of pre-snRNAs in oocytes containing M protein. A mixture of genes encoding X. laevis U1, U2, and U3 RNAs was injected into the nuclei of oocytes that had been pre-injected with M mRNA, or control antisense RNA (Fig. 1A); for labeling of RNAs, α-32P GTP was injected 2 hours later. After incubation for 24 hours, oocytes were separated into nuclear (N) and cytoplasmic (C) fractions and one oocyte-equivalent of RNA of each fraction was analyzed by electrophoresis in an 8% polyacrylamide gel (21). (C) Analysis of the 5′ cap structures of nuclear snRNAs synthesized in the presence or absence of M protein. Total nuclear RNAs (Fig. 1B, lanes 1 and 3) were incubated with precursor-specific antibodies to m7G (22, 23) and RNAs in the precipitates (P) and supernatants (S) were analyzed as in Fig. 1B. (D) Inhibition by M protein of 5SrRNA but not tRNA export. Oocytes were injected with X. laevis 5S rRNA and pre-tRNATyr genes and were labeled and analyzed as in Fig. 1B.

Expression of M protein reduced the amounts (15 to 20 times less) of newly transcribed U1 and U2 RNAs in the cytoplasm (Fig. 1B). This reduction contrasts with only a two to three times lower accumulation of newly made RNA, an indication that M protein inhibited export of the newly made transcripts, but had little on their synthesis (26). Blockage of the export of m7G-capped precursor U1 and U2 RNAs by M protein was confirmed by immunoprecipitation with antibodies to m7G (22,23) that do not recognize the modified m2,2,7G-cap that is generated on export of these RNAs to the cytoplasm (Fig. 1C) (18). The nuclear cap hypermethylation (22) and accumulation of U3 RNA were unaffected by the M protein, indicating that the inhibition of snRNA metabolism by M protein is specific for those snRNAs that must be transported across the nuclear envelope (15, 18, 22).

M protein also inhibited export of 5S ribosomal RNA (rRNA) made either from endogenous (25) or exogenously introduced genes, an indication that the inhibition of export was not limited just to m7G-capped snRNAs (Fig. 1D, top). In contrast, M protein had no effect on the export of tRNAs, such as tRNATyr (Fig.1D, bottom) or tRNAAsn (25), which utilize a transport system that functions even in the absence of functional RCC1 (15). Intranuclear maturation of tRNAs (splicing of tRNATyr and 5′ and 3′ end trimming of both tRNAs) also was not affected by M protein.

The effect of M protein on export of mRNA was assayed with a derivative of major late pre-mRNA from adenovirus (21). Although splicing of the injected pre-mRNA was normal, as indicated by reduction in precursor and a concomitant accumulation of the excised intron lariat (Fig. 2, lanes 2 and 6), export of the spliced mRNA product was severely compromised in oocytes containing M protein. At 1 hour, when about half of the mRNA was exported in control oocytes (lanes 6 and 8), no cytoplasmic mRNA was detected in oocytes treated with M protein (lane 4). Export of a small amount of the polyadenylated mRNAs in the treated oocytes was observed after a 4-hour incubation period (lane 5). However, even at 4 hours, export of the deadenylated spliced mRNA, which is formed when mRNAs are sequestered in the nucleus (9), was not detected; as expected (Fig. 1B), co-injected pre-U1Sm and pre-U5 RNAs were not exported.

Figure 2

Inhibition of export but not splicing of pre-mRNA by VSV M protein. A mixture of 32P-labeled, m7G-capped RNAs containing about 12 fmol of a derivative of adenovirus major late pre-mRNA plus about 5 fmol each of pre-U1Sm RNA and pre-U5 RNAs were injected along with about 5 fmol of γ-methyl-pppG-capped U6 RNA into nuclei of oocytes that had been pre-injected with M mRNA (Fig. 1). One and 4 hours later, the oocytes were fractionated and RNAs were analyzed as in Fig. 1B. The presence of unspliced pre-mRNA and spliced intron lariat in the cytoplasm was due to saturation of nuclear retention of these RNAs and was not seen in experiments in which nonsaturating amounts of pre-mRNAs were injected (9). Deadenylation of pre-mRNA and spliced mRNA in the nucleus is accentuated when export of these molecules is blocked (9). U6 RNA is retained in the nucleus (21) and pre-U1Sm RNA, which lacks an Sm protein binding site, remains in the cytoplasm upon export (18).

The cytoplasmic accumulation of 18S and 28S rRNAs synthesized from endogenous oocyte rRNA genes was also sensitive to M protein (Fig. 3, lane 4) as was the maturation of these rRNAs. Both 45S precursor rRNA and abnormal processing intermediates (asterisks in Fig. 3) accumulated in the nucleus (lane 2), at the expense of mature forms. Both 18S and 28S rRNAs were occasionally detectable in the nuclei of some (Fig. 3) batches of oocytes (25); however, these mature rRNAs were never found in the cytoplasm, showing that M protein affects multiple steps in ribosome biogenesis. Because both maturation and export of rRNAs require transport of newly made ribosomal proteins into the nucleus (27), these results indicated that protein import was inhibited.

Figure 3

Inhibition of maturation and export of rRNA by VSV M protein. Oocytes that had been injected with M mRNA or control mRNA (Fig. 1B) were then injected in the cytoplasm with about 0.5 μCi α[32P]GTP and incubated for 24 hours. After fractionation, the newly synthesized endogenous rRNAs were analyzed by electrophoresis in a 1.2% agarose gel containing formaldehyde.

When tested directly, M protein was at least as effective as WGA in inhibiting the nuclear uptake of cytoplasmically injected X. laevis karyophilic proteins (21) (Fig.4A, lanes 5 to 7). The M protein is also an extremely effective inhibitor of RNA import. U6 RNA and NL15 RNA are imported via the pathway used for NLS-containing proteins but U5 RNA uses a different pathway, specific for snRNPs (24, 28). Both of these Ran-dependent pathways (1, 12, 13) are blocked in the presence of M protein (Fig. 4B, lane 2).

Figure 4

Inhibition of import into oocyte nuclei by VSV M protein. (A) Inhibition of protein import. [35S]methionine-labeled X. laevis nuclear proteins were injected into the cytoplasms of oocytes that had been pre-injected with M mRNA or control antisense mRNA (Fig. 1B) or with ∼200 ng of WGA (2 to 3 hours previously). Twenty hours later, the labeled proteins in the cytoplasmic or nuclear fractions were analyzed by electrophoresis in a 10% polyacrylamide gel containing 0.1% SDS. (B) Inhibition of ribonucleoprotein (RNP) import. A mixture of approximately 5 fmol each of m7G-capped U5 RNA. γ-Methyl-capped U6 RNA and uncapped NL15 RNA (an RNA selected for nuclear localization) (24) were injected into the cytoplasms of oocytes that had been injected with M mRNA or control antisense RNA (Fig. 1B). The oocytes were fractionated 24 hours later for RNA analysis.

Immunoblot analyses of several key transport factors indicated that M protein did not appreciably alter the amounts or the distributions between the nucleus and cytoplasm of RCC1, Ran and importin-β (Fig.5A, lanes 1 to 4). However, it did increase the amount of importin-α that was associated with the nuclear fraction. Most of this importin-α co-migrated with a minor cytoplasmic form present in control oocytes (Fig. 5B, lanes 1 and 4) and both of these slowly migrating forms were phosphorylated (Fig. 5C) (29). Although phosphorylation of importin-α has been implicated in transport (30), it is unclear whether the changes that we observed play an obligate role in the inhibition of transport or are a consequence of the mechanism that is responsible (31). Clearly, the inhibition of transport is unlikely to be caused by depletion of the major form of importin-α from the cytoplasm (Fig.5B, compare lanes 1 and 3). Because M protein inhibits export of mRNA, it is unlikely that the slowly migrating form of importin-α is due to induction of an isoform of this protein.

Figure 5

Effects of VSV M protein on cellular transport factors. (A) Intracellular distributions of Ran, RCC1 and importin-β. Twenty-four hours after injection of M mRNA or control antisense RNA into oocyte cytoplasms, the nucleo-cytoplasmic distributions of several transport factors were analyzed by immunoblotting with antibodies directed against X. laevis RCC1, human Ran, or rat importin-β (21). For importin-β, lanes 1 and 3 each have one oocyte-equivalent of protein, whereas lanes 2 and 4 each have three oocyte-equivalents of protein; for RCC1 and Ran all lanes have one oocyte equivalent of protein. (B) Fractionation and mobilities of importin-α in the presence of M protein or WGA. Proteins in the nuclear and cytoplasmic fractions were analyzed as in (A), with antibody directed against human importin-α. In lanes 5 to 8, oocytes had been pre-injected in the cytoplasm with buffer (lanes 5 and 6) or ∼200 ng of WGA (lanes 7 and 8). Lanes 1, 3, 5, and 7 (cytoplasms) each have one oocyte-equivalent of protein, whereas lanes 2, 4, 6, and 8 (nuclei) each have three oocyte-equivalents of protein. (C) Dephosphorylation of slowly migrating forms of importin-α. Oocyte proteins were prepared and analyzed as in (B), lanes 1, 3, and 4 except that a portion of each sample was incubated for 25 min at 37°C with three units calf intestinal alkaline phosphatase prior to electrophoresis. Lanes 1 to 4 each have one oocyte-equivalent of protein whereas lanes 5 and 6 each have three oocyte-equivalents of protein.

In contrast to M protein, WGA increased the amount of the major, rapidly migrating form of importin-α in the nuclear fraction (Fig.5B, lanes 5 to 8), perhaps reflecting accumulation of docked import complexes at the cytoplasmic side of the NPC (32). The association of importin-α with the nuclear fraction in the presence of M protein might also result from non-productive binding of transport complexes at the NPC rather than from accumulation within the nucleus; however, the two inhibitors appear to work by different mechanisms in that different forms of importin-α accumulate, and WGA inhibits export of tRNAs (8) but M protein does not (9) (Fig. 1D).

How might M protein affect both the import of proteins (and snRNPs) and the export of several classes of RNAs? These features, including the continued export of tRNAs, are identical to those observed when the Ran system is inactivated (12, 15, 31). Thus, we propose that the pleiotropic effects of M protein on active transport between the nucleus and cytoplasm (33) result from interference with transport that is dependent on the Ran GTPase system. It is not known if M protein has enzymatic activity or affects cellular enzymes that impact on the Ran system. Injection of 20 times less M mRNA into oocytes still leads to inhibition of transport; the resulting number of M protein molecules, as detected by immunoblotting (34), is comparable to the number of NPCs per oocyte nucleus and well below the number of molecules of Ran or RCC1. Similarly, VSV infection at a multiplicity of 10 would introduce less than 2 × 104molecules of M protein per cell (16), as compared to 107 molecules of Ran.

Inhibition of Ran-dependent protein import could block RNA export secondarily by depleting the nucleus of shuttling factors that are required for RNA export (8). Alternatively, a change in the activity of Ran or a Ran-specific factor may inactivate other components of the transport machinery that are required for mobilization of RNP export complexes prior to translocation through the NPC (15). The differential effects of M protein on the cytoplasmic appearance of snRNAs, mRNAs, and tRNAs (Figs. 1D and 2) also raises the possibility that additional Ran-independent RNA export pathways exist (8, 33).

It is possible that initiation of VSV transcription releases M protein from the nucleocapsids of infecting virions (16), thereby causing the inhibition of export of host snRNAs and mRNAs. In this model, synthesis of the 47-nucleotide VSV leader RNA would release the virion-bound M protein (35), which could explain the apparent paradox that blockage of snRNA maturation depends on VSV RNA synthesis but not on formation of functional VSV mRNA (17).

Inhibition of nucleocytoplasmic transport could account for changes in protein and RNA synthesis observed upon VSV infection (16) or introduction of M protein into cells (19, 36). Reduced accumulation of mRNA, due to turnover of transcripts that are sequestered in the nucleus, would cause an apparent inhibition of both mRNA and protein synthesis (37). The immediate blockage of export of host cell mRNAs would contribute to the rapid and efficient establishment of infection. Also, the increase (10 to 20 times greater) in translation of a transfected reporter mRNA observed in the presence of M protein mRNA (36) might be due to a lack of competing endogenous mRNAs.

Uninfected cells may use molecules similar to M protein to regulate transport between the nucleus and cytoplasm. Thus, elucidation of the mechanism of action of M protein may reveal novel ways by which gene expression could be controlled through the import and export of nuclear factors and RNAs.

  • * To whom correspondence should be addressed. E-mail: dahlberg{at}facstaff.wisc.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article