Nuclear Export of Influenza Virus Ribonucleoproteins: Identification of an Export Intermediate at the Nuclear Periphery

Virology 282, 215–220 (2001) doi:10.1006/viro.2001.0833, available online at http://www.idealibrary.com on

RAPID COMMUNICATION Nuclear Export of Influenza Virus Ribonucleoproteins: Identification of an Export Intermediate at the Nuclear Periphery Kai Ma, Ann-Marie M. Roy, 1 and Gary R. Whittaker 2 Department of Microbiology and Immunology, Cornell University, Ithaca, New York 14853 Received December 6, 2000; returned to author for revision January 4, 2001; accepted January 10, 2001 A critical phase of the influenza virus life cycle is the regulated translocation of genomic ribonucleoproteins (vRNPs) from the nuclear interior, across the nuclear envelope, and into the cytoplasm. Two viral proteins, M1 and NS2, have previously been implicated as mediators of vRNP export. We show here that vRNP nuclear export is prevented by leptomycin B (LMB), an inhibitor of the cellular factor CRM1. In LMB-treated cells, vRNPs were found in a peripheral nuclear location that localized with the nuclear lamina. vRNPs were not colocalized with either M1 or NS2. In situ extraction of cells late in infection also revealed a peripheral localization of nuclear vRNPs, whereas early in infection vRNPs were dispersed throughout the nuclear interior. We believe that vRNPs at the nuclear periphery represent a novel intermediate in the influenza virus nuclear export pathway. © 2001 Academic Press

nuclear export proposed that M1 promoted vRNP nuclear export by forming a vRNP-M1 complex in the nucleus (9). We have subsequently shown that M1 is indeed a crucial part of the overall vRNP nuclear export event, as expression of recombinant M1 can induce export of nucleartrapped vRNPs (1). However, we and others have also shown that vRNP export can occur even though the vast majority of M1 is trapped in the nucleus due to its hyperphosphorylation (11, 16). The small NS2 protein of influenza has been proposed to be the mediator of vRNP nuclear export (10). NS2 contains a functional nuclear export signal (NES) in in vitro assays, and current models suggest that nuclear export of the influenza genome is mediated by a vRNP–M1–NS2 complex (10). However, data from our laboratory show that a phosphorylation/ M1-mediated triggering of vRNP nuclear export can occur in cells containing undetectable levels of NS2 (1). It is possible therefore that the vRNPs utilize additional signals that are part of the viral nuclear export pathway. In this paper, we examine the nuclear export of influenza vRNPs and show that the virus is sensitive to LMB, an inhibitor of the cellular CRM1 nuclear export receptor. Our data indicate the presence of a novel nuclear export intermediate that is part of an intranuclear trafficking pathway for the viral ribonucleoproteins.

INTRODUCTION Influenza A virus is unusual in that it is an RNA virus that replicates in the nucleus and buds from the plasma membrane (7). It must therefore have defined mechanisms for nuclear export and translocation and assembly at the plasma membrane. The virus has a segmented genome comprising eight distinct RNA molecules, individually packaged into viral ribonucleoprotein complexes (vRNPs). The viral nucleoprotein (NP) forms a core around which the RNA is wrapped in a helical fashion (13). Whereas nuclear import of influenza virus has been extensively studied (see 15 for a review), relatively little is known about its nuclear export. During virus replication, the vRNPs are bound to the insoluble matrix of the nucleus (8). As such, they are relatively resistant to biochemical extraction, but can be released along with cellular chromatin (1). Following the completion of virus replication, the vRNPs must be released from the “nuclear matrix” before they can be translocated across the nuclear pore. How this is achieved is unknown, but it has been suggested that one of the roles of the viral matrix (M1) protein is to bind histone molecules in the nucleus and displace the vRNPs from the chromatinassociated fraction (17). The original model for vRNP

RESULTS AND DISCUSSION 1

Present address: Program in Biomedical Sciences, University of California, San Francisco, CA. 2 To whom reprint requests should be addressed at C5141 Veterinary Medical Center, Dept. Microbiology & Immunology, Cornell University, Ithaca, NY 14853. Fax: (607) 253-3384. E-mail: [email protected].

In influenza virus-infected cells, the onset of vRNP nuclear export coincides with the expression of late proteins (including M1) (9), but the cellular receptors responsible for export have not been identified. To ad215

0042-6822/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

216

RAPID COMMUNICATION

FIG. 1. Dose-response curve of the effect of LMB on vRNP localization. Influenza virus-infected cells were treated with varying concentrations of LMB 3 h p.i. and examined by immunfluorescence microscopy at a late time of infection (8 h). vRNPs were localized and scored for nucleo-cytoplasmic localization: entirely nuclear (nuc), entirely cytoplasmic (cyto), or distributed in both nuclear and cytoplasmic compartments (nuc/cyto). vRNPs present in each localization are given as a percentage of total.

dress this issue, we tested the effects of the drug leptomycin B (LMB), a specific inhibitor of the cellular nuclear export receptor CRM1 (6). We first performed a doseresponse experiment using LMB in an immunofluorescence assay of vRNP trafficking. In the absence of drug, ⬎95% of the cells showed NP entirely in the cytoplasm or distributed between the cytoplasm and the nucleus; less than 5% of the cells contained nuclear vRNPs. By increasing levels of LMB up to 2 ng/ml, 95% of cells now had nuclear vRNPs (Fig. 1). The effects of LMB on nuclear export of influenza vRNPs were dose-dependent between 0.1 and 2 ng/ml. At LMB levels ⬎2 ng/ml, the overall level of vRNP nuclear retention was unaffected, suggesting the presence of a small degree of LMBinsensitive infection in these cells. Influenza virus-infected cells from the immunofluorescence assay are shown in Fig. 2A. vRNPs were almost exclusively cytoplasmic at late times of infection in the absence of LMB (Fig. 2Aa), indicating that extensive nuclear export of the virus genome had occurred in these

cells. We also examined the localization of the influenza M1 and NS2 proteins, which have previously been shown to play a role in nuclear export of the vRNPs. In the absence of LMB, M1 and NS2 localized both to the nucleus and the cytoplasm, and the cytoplasmic localization overlapped with the vRNPs (Figs. 2Ab, 2Ac, 2Ae, and 2Af). In the presence of LMB, vRNP localization was drastically altered. vRNPs were now localized exclusively to the nucleus (Fig. 2Ag), showing that vRNP nuclear export is sensitive to LMB and indicating that the functional receptor for vRNPs is CRM1. However, in contrast to other nuclear substrates that are LMB sensitive, the influenza vRNPs were apparently not evenly distributed through the nucleus and appeared to be concentrated around its outside edge. Under the same conditions, M1 and NS2 were synthesized, indicating that viral mRNA export is LMB insensitive. M1 was largely retained in the nucleus (Fig. 2Ah), with a fraction of M1 present in the cytoplasm. In contrast to vRNPs, the nuclear pool of M1 was distributed evenly throughout the nucleus except for the nucleoli. Overlay of the vRNP and M1 labeling showed that they did not show significant colocalization (Fig. 2Ai). NS2 also displayed nuclear retention compared to untreated cells, but again no obvious colocalization with vRNPs was observed (Figs. 2Ak and 2Al). These data demonstrate that M1 and NS2 behave quite differently to vRNPs in the presence of LMB. To confirm that LMB treatment has a direct effect on vRNP transport and did not result in artifacts due to dissociation of NP from the viral RNA, we analyzed vRNP integrity by glycerol gradient centrifugation and Western blot analysis (5). LMB had no apparent effect on the distribution of NP, which migrated as authentic vRNP complexes and not as soluble NP (data not shown). Although M1 and NS2 were somewhat relocalized to the nucleus with LMB treatment, their localization is similar to that obtained when the proteins are expressed alone (3, 1). It is well known that vRNPs and M1 undergo stable interaction in the cytoplasm to initiate the process of virus budding (7), accounting for the cytoplasmic M1

FIG. 2. Effect of LMB on nucleo-cytoplasmic localization of influenza vRNPs, M1 and NS2. (A) Influenza virus-infected cells were treated with or without LMB (2 ng/ml) at 3 h p.i. and examined by immunfluorescence microscopy at a late time of infection (8 h). vRNPs were localized with anti-NP antibodies (red) and influenza M1 or NS2 localized with anti-M1 or anti-NS2 antibodies (green). A merge of the red and green channels is also shown, with colocalization of vRNP/M1 and vRNP/NS2 signal represented in yellow. (B) Influenza virus-infected cells were treated with LMB at 3 h p.i. and examined by confocal microscopy at a late time of infection (8 h) using anti-sera specific for vRNPs (a). (b) shows double-label confocal microscopy, using anti-nuclear pore complex antibody (NPCs; green) together with anti-vRNP sera (red). (c) shows double-label confocal microscopy with anti-lamin A/C antibodies (Lamina; green) together with anti-vRNP sera (rd). Merged signal is represented in yellow. (b) and (c) represent a portion of the nucleus and are shown magnified in comparison to (a). Bar ⫽ 1 ␮m. FIG. 3. In vivo extraction of vRNPs, M1, and NS2 from influenza-infected cells. (A) Influenza virus-infected cells were treated with or without LMB (2 ng/ml) at the time of infection (0 h) and examined by immunfluorescence microscopy at a late (8 h) or early (3 h) time of infection. Cells were extracted with 0.1% Triton X-100 (TX-100) or with PBS prior to fixation and vRNPs localized with anti-NP antibodies. (B) Influenza virus-infected cells were treated with LMB at the time of infection (0 h) and examined by immunfluorescence microscopy at a late (8 h) time of infection. TX-100-extracted, cells were subjected to double-label microscopy with anti-NP antibodies, combined with anti-M1 or anti-NS2 antibodies. Following capture, images were subject to no-neighbors deconvolution using Slide Book software.

RAPID COMMUNICATION

217

218

RAPID COMMUNICATION

localization seen in Fig. 2Ab. In the presence of LMB, cytoplasmic M1–vRNP interaction is not possible. In this situation it is likely that M1 reverts to its “default” localization (mostly nuclear with some cytoplasmic localization). As M1 is a small protein (27 kDa), it is also possible that it could still diffuse out of the nucleus, as it is below that threshold for diffusion across the nuclear pore. A similar scenario can be envisioned for NS2 (14 kDa), with the NS2 protein diffusing out of the nucleus if interaction with a receptor is prevented. Overall, these data argue against the formation of a stable vRNP–M1–NS2 complex in the nucleus that uses CRM1 as an export receptor. To further investigate the localization of vRNPs in the presence of LMB, we performed confocal microscopy. No vRNP signal could be detected in the nuclear interior in the presence of LMB (Fig. 2Ba), and the vRNPs were apparently trapped at the nuclear periphery. To determine the precise localization of the LMB-trapped vRNPs, we performed double-label confocal microscopy, with vRNPs in combination with antibodies against nuclear pore complex (NPC) proteins or against nuclear lamina proteins. In LMB-treated cells, vRNPs were visible directly below the nuclear pores and showed no colocalization with NPCs by confocal microscopy (Fig. 2Bb). The localization of vRNPs below the nuclear envelope suggested an association with the nuclear lamina. We therefore localized the nuclear lamina using anti-sera specific for lamins A and C. In LMB-treated cells, vRNPs and lamins A/C showed substantial overlap by confocal microscopy (Fig. 2Bc). We routinely observed that a small degree of the outermost nuclear lamina was vRNP-free (green signal, Fig. 2Bc) and that the inner fraction of the lamina overlapped with the vRNPs (yellow signal, Fig. 2Bc). A fraction of the vRNPs also resided in a localization interior to the lamina, which was not coincident with the nuclear lamina (red signal, Fig. 2Bc). To further examine the subcellular localization of LMBtrapped vRNPs, we performed thin-section electron microscopy. Cells treated with or without LMB showed no obvious differences in the morphology of the nucleus or nuclear envelope, and an intermediate could not be observed in LMB-treated cells (data not shown). However, the inability to visualize an export intermediate most likely represents the lack of staining of vRNP structures by conventional electron microscopy techniques. A major difference was apparent at the plasma membrane and in the absence of LMB, large numbers of budding virions were seen that were not present after LMB treatment (data not shown), confirming that LMB had a significant effect on influenza virus infection due a defect in nuclear export. To determine the in vivo localization of vRNPs in the presence and absence of LMB, we performed immunofluorescence microscopy of influenza infected cells that were permeabilized with 0.1% Triton X-100 (TX-100) be-

fore fixation or were treated with PBS. As expected at a late time of infection, vRNPs were cytoplasmic in the absence of LMB (Fig. 3Aa) and localized to the nuclear periphery in the presence of LMB when treated with PBS (Fig. 3Ae). TX-100 treatment resulted in an extraction of vRNPs from the cytoplasm. Interestingly, a peripheral nuclear localization of vRNPs was now also visible in absence of LMB (Fig. 3Ab). This distribution was essentially unchanged when cells were treated with LMB and extracted with TX-100 (Fig. 3Af). These data indicate that in situ extraction of infected cells with TX-100 exposes a peripheral nuclear pool of vRNPs that are present both in the presence or absence of LMB. Thus, the LMB-trapped nuclear export intermediate identified previously may not require LMB treatment but appears to always be present in infected cells undergoing vRNP export. However, this intermediate is normally invisible without pre-extraction and accumulates drastically upon LMB treatment. To address the issue of movement of vRNPs from the nuclear interior prior to nuclear export, we repeated the extraction at an early time of infection when vRNA is actively being transcribed and replicated in the nuclear interior (7). In all cases examined, in the presence or absence of LMB and with or without extraction, vRNPs were present throughout the nuclear interior and showed no sign of a peripheral nuclear localization (Figs. 3Ac– 3Ah). These data show that the peripherally localized vRNPs are not part of the nuclear targeting during active virus replication or transcription but are specific to the late phase of virus replication when the vRNPs are undergoing nuclear export. Our extraction studies (Fig. 3Ad) show that early in infection, vRNPs tend to concentrate in a small number of bright spots in the nucleus (typically 2–3 spots). The identification of these spots is currently unknown, and we are currently investigating whether they colocalize with any known nuclear subcompartment. Under TX-100 extraction conditions, neither M1 nor NS2 displayed a prominent peripheral localization (Fig. 3B). However, following TX-100 extraction, a limited degree of overlap of the M1 and NS2 signal occurred with the peripherally localized vRNPs (Figs. 3Bc and 3Bf). As with nonextracted cells (Fig. 2A), the bulk of the M1 and NS2 was in an intranuclear location and did not colocalize with the vRNPs. These data suggest the possibility that whereas the majority of the M1 and NS2 is not part of a vRNP nuclear export complex, a small subset of M1 and NS2 are present in the TX-100-resistant fraction as part of the viral export complex. Overall, our data show that nuclear export of the genomic influenza vRNPs is sensitive to LMB and that vRNPs are trapped in a novel localization in the nuclear periphery. We believe that vRNPs are trapped in an intermediate state that represents a transition state for RNP traffic through the nuclear interior. Our data show that at early times of infection, when vRNPs are being actively transcribed/replicated, they are present in the

RAPID COMMUNICATION

219

nuclear interior. Upon induction of vRNP export late in infection, vRNPs move out of their interior localization prior to translocation through nuclear pores. The vRNPs normally pass through a transient localization in the nuclear periphery, but in the presence of LMB, the vRNPs all become trapped in this intermediate state directly beneath the nuclear envelope, which appears to overlap with the nuclear lamina. Our data suggest that subsequent to their appearance in the nuclear periphery, vRNPs are translocated through the nuclear pore by the exportin CRM1. Unexpectedly, our microscopy data suggest that vRNP nuclear export does not involve the bulk of the viral M1 and NS2 proteins. However, data from extracted cells indicate the possibility that a few molecules of M1 or NS2 are associated with each vRNP. In addition, it remains possible that M1 and or NS2 are required for a vRNP nuclear export event before or after vRNP–CRM1 interaction. Our localization data suggest a possible interaction of vRNPs with components of the nuclear lamina prior to translocation across the nuclear pore. It is well known that the nuclear lamina is required for nuclear integrity and DNA replication (2). In interphase cells, the nuclear lamina is situated in a peripheral nuclear location and acts as a scaffold for attachment of the baskets of the nuclear pores and the association of chromatin. However, there is little direct evidence to functionally link the nuclear lamina with nuclear transport pathways. Limited data suggest that the nuclear lamina is not part of nuclear import pathways (4); however, there is no information on its role in nuclear export. It would appear likely that the intranuclear architecture is crucial for nuclear export, and proteins have been identified that connect the nuclear pore with the nuclear interior (14), but these are not part of the nuclear lamina. Given that influenza vRNPs accumulate as a functional nuclear export intermediate and that the vRNPs are both physically large and abundantly expressed (facilitating morphological and biochemical examination of cells), they are an attractive model to study the process of intranuclear transport and the isolation of associated receptor molecules.

were detected using the NP-specific mouse monoclonal antibody H10, L16–4R5; IgG2a (ATCC). Biochemical analysis of influenza virus-infected cells shows that this antibody has negligible association with free nucleoprotein late in infection, and so labeling represents the localization of vRNP complexes (1). Influenza M1 protein was detected using rabbit polyclonal sera raised against purified M1 (8-071), and NS2 protein was detected using a mono-specific polyclonal antibody (generous gift of Dr. P. Palese. Mt. Sinai School of Med.). NPCs were detected using the mouse monoclonal antibody Mab 414; IgG1 (BD Transduction Labs) and nuclear lamina detected using a mouse monoclonal antibody against LaminA/C; IgG2b (Santa Cruz Biochemicals). For double-labeling with monoclonal antibodies, we used the following isotype-specific secondary antibodies; FITC-anti-mouse IgG1 or IgG2b (BD Pharmingen), followed by biotin-antimouse IgG2a (BD Pharmingen) and Cy3-streptavidin (Sigma). Lack of cross reactivity of the secondary antibodies was confirmed by using the primary and secondary antibodies in their incorrect combination and with non-infected cells. Cells were analyzed using a Zeiss Axioplan 2 fluorescence microscope and Slide Book software (Intelligent Imaging Innovations Inc). For quantitation cells in the immunofluorescence assay, all cells within random fields of view were scored and 100 cells counted for each sample. Confocal microscopy was performed using an Olympus FluoView confocal station (Olympus). FITC was excited with the 488 nm line of an Argon laser and Cy3 was excited with the 568 nm line of a Krypton laser.

MATERIALS AND METHODS

1. Bui, M., Wills, E., Helenius, A., and Whittaker, G. R. (2000). The role of influenza virus M1 protein in nuclear export of viral ribonucleoproteins. J. Virol. 74, 1781–1786. 2. Goldberg, M., Harel, A., and Gruenbaum, Y. (1999). The nuclear lamina: Molecular organization and interaction with chromatin. Crit. Rev. Eukaryot. Gene Expression 9, 285–293. 3. Greenspan, D., Krystal, M., Nakada, S., Arnheiter, H., Lyles, D. S., and Palese, P. (1985). Expression of influenza virus NS2 nonstructural protein in bacteria and localization of NS2 in infected eucaryotic cells. J. Virol. 54, 833–843. 4. Jenkins, H., Ho¨lman, T., Lyon, C., Lane, B., Stick, R., and Hutchinson, C. (1993). Nuclei that lack a lamina accumulate karyophilic proteins and assemble a nuclear matrix. J. Cell Sci. 106, 275– 285. 5. Kemler, I., Whittaker, G., and Helenius, A. (1994). Nuclear import of microinjected influenza virus ribonucleoproteins. Virology 202, 1028–1033.

Cells, viruses, and reagents Influenza A virus (strain WSN) was prepared as described and Mv1 Lu cells (NBL-7; ATCC) used for experimental infections (12). Cells were infected with 5–10 pfu/cell in all experiments. Leptomycin B (LMB) was kindly provided by Dr. M. Yoshida, Univ. of Tokyo, and was used at a concentration of 2 ng/ml unless otherwise stated. Indirect immunofluorescence microscopy Immunofluorescence microscopy was carried out essentially as described previously (12). Influenza vRNPs

ACKNOWLEDGMENTS We thank Ruth Collins for a critical reading of this manuscript and Liz Wills for performing electron microscopy. We also thank Minoura Yoshida and Peter Palese for generous contributions of reagents. These studies were funded in part by a research grant from the American Lung Association and by the US Department of Agriculture Animal Health and Disease Program.

REFERENCES

220

RAPID COMMUNICATION

6. Kudo, N., Matsumori, N., Taoka, H., Fujiwara, D., Schreiner, E. P., Wolff, B., Yoshida, M., and Horinouchi, S. (1999). Leptomycin B inactivates CRM1/exportin 1 by covalent modification at a cysteine residue in the central conserved region. Proc. Natl. Acad. Sci. USA 96, 9112–9117. 7. Lamb, R. A., and Krug, R. M. (1996). Orthomyxoviridae: The viruses and their replication. In “Fields Virology” (B. N. Fields, D. M. Knipe, and P. M. Howley, Eds.), pp. 1353–1395. Lippincott-Raven, Philadelphia. 8. Lo´pez-Turiso, J. A., Martı´nez, C., Tanaka, T., and Ortı´n, J. (1990). The synthesis of influenza virus negative-strand RNA takes place in insoluble complexes present in the nuclear matrix fraction. Virus Res. 16, 325–336. 9. Martin, K., and Helenius, A. (1991). Nuclear transport of influenza virus ribonucleoproteins: The viral matrix protein (M1) promotes export and inhibits import. Cell 67, 117–130. 10. O’Neill, R. E., Talon, J., and Palese, P. (1998). The influenza virus NEP (NS2 protein) mediates the nuclear export of viral ribonucleoproteins. EMBO J. 17, 288–296. 11. Rey, O., and Nayak, D. (1992). Nuclear retention of M1 protein in a

12.

13.

14.

15. 16.

17.

temperature-sensitive mutant of influenza A/WSN/33 virus does not affect nuclear export of viral ribonucleoproteins. J. Virol. 66, 5815–5824. Roy, A.-M. M., Parker, J. S., Parrish, C. R., and Whittaker, G. R. (2000). Early stages of influenza virus entry into Mv-1 lung cells: Involvement of dynamin. Virology 267, 17–28. Ruigrok, R. W. H. (1998). Structure of influenza A, B and C viruses. In “Textbook of Influenza” (K. G. Nicholson, R. G. Webster, and A. J. Hay, Eds.), pp. 29–42. Blackwell Science, Oxford. Strambio-de-Castilla, C., Blobel, G., and Rout, M. P. (1999). Proteins connecting the nuclear pore complex with the nuclear interior. J. Cell Biol. 144, 839–855. Whittaker, G. R., Kann, M., and Helenius, A. (2000). Viral entry into the nucleus. Annu. Rev. Cell Dev. Biol. 16, 627–651. Whittaker, G., Kemler, I., and Helenius, A. (1995). Hyperphosphorylation of mutant influenza virus matrix (M1) protein causes its retention in the nucleus. J. Virol. 69, 439–445. Zhirnov, O. P., and Klenk, H.-D. (1997). Histones as a target for influenza virus matrix protein M1. Virology 235, 302–310.