SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We have examined the requirements for the initiation factors (eIFs) eIF4A and eIF2 to translate Sindbis virus (SV) subgenomic mRNA (sgmRNA) in the natural hosts of SV: vertebrate and arthropod cells. Notably, this viral mRNA does not utilize eIF4A in SV-infected mammalian cells. However, eIF4A is required to translate this mRNA in transfected cells. Therefore, SV sgmRNA exhibits a dual mechanism for translation with respect to the use of eIF4A. Interestingly, SV genomic mRNA requires eIF4A for translation during the early phase of infection. In sharp contrast to what is observed in mammalian cells, active eIF2 is necessary to translate SV sgmRNA in mosquito cells. However, eIF4A is not necessary for SV sgmRNA translation in this cell line. In the SV sgmRNA coding region, proximal to the initiation codon is a hairpin structure that confers eIF2 independence only in mammalian cells infected by SV. Strikingly, this structure does not provide independence for eIF4A neither in mammalian nor in mosquito cells. These findings provide the first evidence of different eIF requirements for translation of SV sgmRNA in vertebrate and invertebrate cells. These observations can help to understand the interaction of SV with its host cells.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Protein synthesis in mammalian cells is mainly regulated at the initiation phase. The canonical mechanism of initiation of mRNA translation commences with recognition of the m7GpppN cap structure present at the 5′ end by the eIF4F complex formed by the cap recognition factor eIF4E, the DEAD-box helicase eIF4A and the scaffolding protein eIF4G (Sonenberg and Hinnebusch, 2009; Jackson et al., 2010). In addition, eIF2 interacts with the initiator Met-tRNAMeti and GTP leading to the formation of the ternary complex Met-tRNAMeti-eIF2-GTP. This complex together with eIF3, eIF1, eIF1A and perhaps eIF5 recognize the small 40S ribosomal subunit to form the 43S pre-initiation complex (Lorsch and Dever, 2010). The interaction between eIF3 and eIF4G promotes the joining of the pre-initiation complex to the 5′ end of mRNAs. Unwinding of the secondary structure at the proximal 5′ end of mRNA is accomplished by eIF4F in conjunction with eIF4B or eIF4H, preparing the mRNA to receive the incoming pre-initiation complexes. The pre-initiation complex then scans the mRNA leader sequence in the 5′–3′ direction until the initiation codon is found. This scanning involves unwinding of secondary structure within the leader sequence by the eIF4A subunit of eIF4F, a process stimulated by eIF4B or eIF4H (Parsyan et al., 2011). Other RNA helicases, including DHX29, may also participate in this process (Pisareva et al., 2008; Abaeva et al., 2011; Parsyan et al., 2011). On reaching the initiation codon, scanning of the pre-initiation complex is complete and the 48S initiation complex is generated. Once codon-anticodon base pairing is established between the Met-tRNAMeti and the AUG initiation codon, eIF5 triggers GTP hydrolysis. Several factors, including eIF1, eIF1A, eIF2-GDP and eIF3, exit the 40S subunit when the 60S ribosomal subunit is joined to form the 80S ribosome. This is promoted by eIF5B-GTP, which in turn leaves the ribosome upon GTP hydrolysis. The initiation phase ends when the initiator Met-tRNAMeti is bound at the P site of the 80S ribosome. A number of cellular and viral mRNAs contain internal ribosome entry site (IRES) elements within their 5′ untranslated regions (5′UTRs) directing translation by a non-canonical mechanism (Balvay et al., 2009; Belsham, 2009; Fitzgerald and Semler, 2009). In these cases the formation of the initiation complex may occur without the participation of some eIFs and, in some cases, the 80S ribosome is formed directly at the initiation codon in the absence of scanning from the 5′ end.

A variety of translation initiation mechanisms has been reported for viral mRNAs (Niepmann, 2009; Sanz et al., 2009; Walsh, 2010; Hertz and Thompson, 2011). In some cases viral mRNAs are translated following the canonical mode of initiation, while in other instances, efficient translation of viral mRNAs takes place in the absence of one or more eIFs (Pestova et al., 2001; Terenin et al., 2008; Skabkin et al., 2010; Kim et al., 2011). Alphaviruses have two different mRNAs that are translated at different times during the infectious cycle: the genomic 49S mRNA (gmRNA) is translated early during infection following a canonical mechanism and gives rise to the non-structural proteins (nsP1–4), whereas the subgenomic 26S mRNA (sgmRNA) directs the synthesis of structural proteins during the late phase of infection (Strauss and Strauss, 1994). Both mRNAs are capped at their 5′ end and contain a poly(A) tail. Protein synthesis directed by the sgmRNA of alphaviruses can occur when eIF2 is highly phosphorylated (Gorchakov et al., 2004; McInerney et al., 2005; Ventoso et al., 2006) and does not require intact eIF4G in infected cells (Castello et al., 2006; Sanz et al., 2009). Notably, this mRNA exhibits a dual mechanism of translation; it follows the canonical mode when transfected in culture cells or in cell-free systems whereas eIF2 and intact eIF4G are dispensable in virus-infected cells (Sanz et al., 2009). Sindbis virus (SV) is a representative member of the Alphavirus genus that can infect a variety of vertebrate and invertebrate cells (Strauss and Strauss, 1994). The inhibition of host translation and the final outcome of SV infection are different in mammalian and mosquito cells (Karpf and Brown, 1998). How many eIFs participate in sgmRNA translation in SV-infected cells and the exact mechanism of initiation under physiological conditions is still unknown. In reconstituted in vitro systems, the SV sgmRNA can be translated only in the presence of Ligatin (or MCT-1/DENR), DHX29 and eIF3 and under these conditions 80S ribosomes interact directly with the AUG initiation codon (Skabkin et al., 2010). Recent studies have demonstrated that some viral proteins can modify the requirements for eIFs. Thus, hantavirus N protein (Mir and Panganiban, 2008), influenza virus PB2 (Burgui et al., 2007) or poliovirus 2Apro (Redondo et al., 2011) can enable efficient viral mRNA translation in the absence of some eIFs. Therefore, analysis of viral mRNA translation in transfected cells, cell-free systems or reconstituted systems, must take into consideration that they do not accurately reflect the physiological conditions under which viral mRNAs are translated in virus-infected cells. In this regard, we now demonstrate that translation of SV sgmRNA does not require eIF4A, but this factor is necessary for efficient translation in transfected mammalian cells or in cell-free systems. This has been analysed by utilization of the selective inhibitor of eIF4A, hippuristanol (hipp) (Bordeleau et al., 2006; Lindqvist et al., 2008). We have also found that eIF4A is not required for SV sgmRNA translation in infected mosquito cells. Notably, eIF2 is involved in the translation of sgmRNA in SV-infected insect cells. Finally, we have examined in both cell lines the role of a hairpin structure (DLP) located downstream of the initiation codon. This structure confers eIF2 but not eIF4A independence for SV sgmRNA translation initiation in SV-infected mammalian cells, whereas it has no function in uninfected cells or in SV-infected mosquito cells. These findings provide insights into the host factors required during the SV life cycle.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Requirement for eIF4A in several virus-infected cells

Translation of SV sgmRNA does not require active eIF2 or intact eIF4G in infected cells (Gorchakov et al., 2004; Castello et al., 2006; Ventoso et al., 2006; Sanz et al., 2009). On the other hand, translation of several mRNAs bearing picornavirus IRES elements can take place when eIF4G has been cleaved by picornavirus proteases 2Apro or Lpro (Jackson et al., 2010; Castello et al., 2011). Notably, expression of poliovirus (PV) 2Apro in SV-infected cells efficiently cleaves eIF4G, but does not inhibit translation of SV sgmRNA, despite the fact that this mRNA is capped at its 5′ end (Castello et al., 2006). Therefore, it was of interest to determine whether eIF4A participates in SV sgmRNA translation, as occurs with the majority of cellular and viral mRNAs. To this end, hipp, a selective inhibitor of eIF4A (Lindqvist et al., 2008), was employed. Concentrations of hipp as low as 0.3 μM drastically blocked (∼ 90%) cellular mRNA translation in uninfected BHK cells (Fig. 1A and E). This result indicates that cellular mRNAs employ eIF4A during initiation and when eIF4A is blocked, it cannot be replaced by other putative cellular helicases to restore the initiation of cellular protein synthesis (Parsyan et al., 2011). On the other hand, the presence of hipp even at concentrations of 0.5 μM had no inhibitory effect on the translation of SV sgmRNA (Fig. 1B and E), since the capsid protein (C), as well as the viral glycoproteins were synthesized at control levels at 7 h post infection (hpi). Besides, hipp had a similar behaviour on the translation of vesicular stomatitis virus (VSV) mRNAs (Fig. 1C and E), in line with the findings that translation of these mRNAs does not require eIF4E or eIF4G (Connor and Lyles, 2002; Welnowska et al., 2009). Our present results demonstrate that eIF4A is not necessary to translate SV sgmRNA or VSV mRNAs, making it unlikely that any of the three components of the eIF4F complex participate in SV or VSV protein synthesis at late 10times of infection. As a control, encephalomyocarditis virus (EMCV)-infected BHK cells were analysed. In this case, inhibition of viral protein synthesis by ∼ 90% was observed when 0.3 μM hipp was present at 6 hpi (Fig. 1D and E). This observation is in agreement with the ability of hipp to block picornavirus IRES-driven translation (Bordeleau et al., 2006; Lindqvist et al., 2008). In conclusion, there is a sharp contrast between the drastic blockade of translation by hipp in control or in EMCV-infected cells, as compared with the lack of inhibition in SV- or VSV-infected cells.

figure

Figure 1. Effect of hipp on translation in virus-infected BHK cells.

A–D. BHK cells were either mock-infected (A) or infected with SV (B), VSV (C) or EMCV (D) with a multiplicity of infection (MOI) of 5 pfu per cell. At 5.5 hpi (or 4.5 hpi in the case of EMCV infection), cells were treated with hipp at the indicated concentrations for 30 min. From 6 to 7 hpi (from 5 to 6 hpi in the case of EMCV infection) cultures were labelled with [35S]Met-Cys in the absence or presence of the inhibitor at the same concentrations as before. Radiolabelled proteins were separated by SDS-PAGE, followed by autoradiography.

E. The percentage of cellular and viral protein synthesis in cells treated with hipp compared with untreated cells was calculated from values obtained by densitometric scanning of the corresponding bands. The protein bands analysed were actin (mock-infected cells), C (SV-infected cells), P (VSV-infected cells) and P12A (EMCV-infected cells).

Download figure to PowerPoint

To further assess the potential lack of participation of eIF4A in SV-infected cells, we analysed the subcellular localization of this factor by immunofluorescence assays during the late phase of infection. For this purpose, colocalization of eIF4A with the viral C protein or with TIA (T-cell intracellular antigen) was tested at 5 and 8 hpi in the presence or absence of hipp. In control uninfected BHK cells, TIA was mainly located within the nucleus, while eIF4A was present in the cytoplasm. Previous results have demonstrated that hipp treatment induces stress granules in HeLa and MEFs cells (Mazroui et al., 2006). Addition of 0.5 μM hipp to BHK cells induced the formation of stress granules since a fraction of TIA was then present in punctate bodies, where a fraction of eIF4A was also present (Fig. 2). This may indicate that the inhibition of translation by hipp could be due at least in part to the redistribution of a fraction of eIF4A in stress granules, although a large fraction of eIF4A remains distributed in the cytoplasm. In Semliki forest virus-infected cells there is an evanescent induction of stress granules, such that at early times during the late phase, the presence of stress granules is observed in some cells, while these granules are destroyed as infection progresses (McInerney et al., 2005). Under our experimental conditions, we did not observe the formation of stress granules in the infected cells and only a small percentage (about 8%) of infected cells contained stress granules when hipp was present at 5 hpi (Fig. 2). Consequently, two different patterns of eIF4A and TIA distribution were observed at this time. In one of these patterns, TIA had been released from the nucleus and appeared mainly in stress granules and also in part in the cytoplasm. eIF4A was present both in the cytoplasm and in stress granules together with TIA, while the viral capsid protein was diffused in the cytoplasm. In contrast, most of the infected cells treated with hipp did not contain stress granules (Fig. 2). In this latter case, viral C protein typically appeared in a region close to the nucleus where viral protein synthesis takes place and ribosomes appear concentrated (Sanz et al., 2009). This is in good agreement with previous studies that identified interactions between alphavirus C protein and ribosomes (Ulmanen et al., 1976; Soderlund and Ulmanen, 1977). Notably, in infected cells that did not contain stress granules, eIF4A was distributed in the cytoplasm, but its accumulation was lower in the regions where translation was occurring (Fig. 2). A similar pattern of eIF4A distribution was observed in the infected cells at 8 hpi, irrespective of whether or not they had been treated with hipp since at this time stress granules were no longer assembled. For a more accurate assessment of the degree of colocalization of eIF4A and SV C protein, we have employed the ImageJ program with the Just Another Colocalization Plugin (JaCoP), which provides the Pearson's correlation coefficient (values higher than 0.5 denote colocalization) (Bolte and Cordelieres, 2006). In SV-infected cells, the Pearson's coefficient for eIF4A and C was 0.34, indicating that there was no colocalization between those two proteins. This finding suggests that eIF4A may not be involved in viral mRNA translation.

figure

Figure 2. Subcellular localization of eIF4A, SV C protein and TIA-1 in mock and SV-infected BHK cells in the presence or absence of hipp. BHK cells were seeded on glass coverslips and mock-infected or infected with SV (5 pfu per cell). At 3.5 or 6.5 hpi, cells were not treated or treated for 90 min with 0.5 μM hipp. At 5 or 8 hpi, respectively, cells were fixed, permeabilized and processed for immunofluorescence using anti-eIF4A (green), anti-SV C protein (cyan) and anti-TIA-1 (red). Images were acquired on a confocal microscope and subsequently processed with Huygens 4.1 software. Merged images represent the simultaneous visualization of eIF4A, C and TIA-1.

Download figure to PowerPoint

A good marker of sites of protein synthesis in cells is elongation factor eEF2. In uninfected cells, this factor was widespread in the cytoplasm, even when BHK cells had been treated with 0.5 μM hipp (Fig. S1). This result indicates that eEF2 is not redistributed to stress granules after hipp treatment. As described above, the subcellular localization of SV C protein indicates sites where protein synthesis takes place. Thus, eEF2 colocalized with SV C protein in SV-infected cells (Fig. S1), sustained by a Pearson's correlation coefficient of 0.72. This observation supports the conclusion that the location of C protein provides a good indication of the foci where viral translation occurs.

Participation of eIF4A in the translation of SV sgmRNA from different SV replicons

Cells transfected with SV replicons mimic in many instances the events that occur in SV-infected cells. Thus, cellular translation is also abrogated in these cells and the requirements for eIFs to translate sgmRNA are similar to those observed in SV-infected cells (Sanz et al., 2009). The use of SV replicons that synthesize luciferase (luc) is useful to precisely quantify protein synthesis in this system. One possibility to account for the lack of eIF4A requirement in SV-infected cells is that a viral protein replaces this function. Alternatively, it is possible that a cellular protein replaces eIF4A in the context of viral infection. To test these possibilities, we have analysed two different SV replicons: one of them (rep C+luc) encodes for C+luc and contains the genuine leader sequence of sgmRNA followed by the C and luc sequences (Fig. 3A), whereas the other type of SV replicons synthesize luc directed by the EMCV or PV IRESs (rep LEMCV-luc and rep LPol-luc, Fig. 4A).

figure

Figure 3. Effect of hipp on the translation of sgmRNA from a SV replicon.

A. Schematic representation of the replicons employed. nsPs, non-structural proteins; sPs, structural proteins; SG.P., subgenomic promoter (represented by an arrow).

B and C. BHK cells were transfected with Lipofectamine 2000 and in vitro transcribed mRNAs: rep C+luc (B) or Mengo-luc replicon (C). At 3.5, 5.5 and 7.5 hpt, cells were not treated or treated with 0.5 μM hipp or 100 μg ml−1 cycloheximide for 90 min. At the indicated times, untreated and treated cells were collected in luciferase lysis buffer and luc activity was measured. Values obtained from cycloheximide-treated cells were used to subtract the amount of luc synthesized prior to hipp addition (3.99 × 106 RLU, 14.74 × 106 RLU, 11.48 × 106 RLU for rep C+luc at 5, 7 and 9 hpt respectively; 0.46 × 106 RLU, 1.62 × 106 RLU, 4.42 × 106 RLU for Mengo-luc replicon at 5, 7 and 9 hpt respectively). Luciferase activity results are means ± SD of three representative experiments performed in triplicate. RLU, relative light units. The percentage values of hipp-treated cells relative to their respective untreated cells are indicated in the figure.

Download figure to PowerPoint

figure

Figure 4. Action of 2Apro on the translation of SV sgmRNAs bearing picornavirus IRESs in hipp-treated BHK cells.

A. Schematic representation of the replicons employed. The arrow represents the subgenomic promoter.

B and C. BHK cells were co-transfected with Lipofectamine 2000 and a mixture of in vitro transcribed mRNAs composed of rep LEMCV-luc (B) or rep LPol-luc (C) and rep C+2A or rep C+2A(G60R). At 5.5 hpt, cells were treated with the indicated concentrations of hipp (0, 0.2 and 0.5 μM) or 100 μg ml−1 cycloheximide for 90 min. Cells were then harvested in luciferase lysis buffer and luc activity determined. Values obtained from cycloheximide-treated cells were used to subtract the amount of luc synthesized prior to hipp addition [15.71 × 106 RLU for rep LEMCV-luc + rep C+2A; 0.23 × 106 RLU for rep LEMCV-luc + rep C+2A(G60R); 19.48 × 106 RLU for rep LPol-luc + rep C+2A; 0.29 × 106 RLU for rep LPol-luc + rep C+2A(G60R)]. Luc activity results are displayed as means ± SD of three representative experiments performed in triplicate and plotted using a logarithmic scale on the ordinate axis. The percentage values of hipp-treated cells relative to their respective untreated cells are indicated in the figure.

Download figure to PowerPoint

Transfection of BHK cells with rep C+luc gave rise to increased synthesis of luc as replication progressed (Fig. 3B) reaching the maximum values at 7 h post transfection (hpt). The addition of 0.5 μM hipp at different times during replication did not inhibit luc synthesis, indicating that eIF4A does not participate in translation in this system. As a control, a replicon from Mengovirus containing the luc gene in place of the viral structural proteins (see scheme in Fig. 3A) (Fata-Hartley and Palmenberg, 2005) was transfected in BHK cells and luc activity measured at different times in presence or absence of 0.5 μM hipp. In this case, hipp strongly inhibited the synthesis of luc (Fig. 3C) in agreement with the results obtained above in cells infected with EMCV (Fig. 1D), which is closely related to mengovirus.

Several laboratories and the results obtained in Fig. 1D have described that translation directed by picornavirus IRES is dependent on eIF4A. Therefore, it was of interest to determine whether protein synthesis directed by EMCV or PV IRESs is eIF4A-dependent in the context of SV replication. We have reported that sgmRNAs bearing picornavirus IRESs, such as EMCV or PV IRESs, are inefficiently translated in SV replicating cells (Sanz et al., 2010). However, the presence of PV 2Apro potent stimulates this translation. In line with these observations, BHK cells co-transfected with rep LEMCV-luc or rep LPol-luc and a proteolytically inactive 2A protease [rep C+2A(G60R)] gave rise to low levels of luc synthesis, whereas co-transfection of these replicons with rep C+2A induced a strong stimulation (about 36-fold) of luc activity (Fig. 4B and C). Strikingly, addition of 0.5 μM hipp potently diminished luc synthesis (Fig. 4B and C), indicating that eIF4A is not replaced by a SV or cellular protein in this system. Moreover, these observations demonstrate that PV 2Apro does not lead to independence from eIF4A, whereas 2Apro expression does promote translation in the absence of eIF2 (Redondo et al., 2011). In addition, the eIF4A present in SV replicating cells does not become irreversibly modified or inactivated, since it can be used to translate mRNAs derived from SV replicons and participates in translation of sgmRNAs bearing EMCV or PV IRESs. Therefore, translation of only genuine SV sgmRNA is independent of eIF4A in the context of an active infection, whereas protein synthesis promoted by picornavirus IRESs requires eIF4A.

Dual mechanism for the requirement of eIF4A in SV sgmRNA translation

In principle, it is possible that SV sgmRNA does not require the helicase activity of eIF4A since it contains a short unstructured 5′UTR. Alternatively, this sgmRNA might exhibit a dual mechanism for its translation in SV-infected cells compared with when sgmRNA is translated out of an infection context, as previously reported with respect to utilization of eIF2 or eIF4G (Sanz et al., 2009). To test this possibility, BHK cells were transfected with in vitro synthesized non-replicating mRNAs such as sgmRNA C+luc or IRES EMCV-luc (Fig. 5A) in presence or absence of hipp. In addition, to analyse the translation of SV gmRNA, we made use of a construct bearing the luc gene inside the region encoding the early protein nsP3 (gmRNA SV-Luc, see scheme in Fig. 5A). Transfection of this mRNA leads to luc synthesis prior to SV replication. As observed in Fig. 5B, protein synthesis directed by IRES EMCV-luc mRNA was sensitive to inhibition by hipp. Translation of SV gmRNA was also inhibited by hipp although to a lesser extent as compared with IRES EMCV-luc mRNA. Curiously, in vitro synthesized sgmRNA C+luc showed a similar behaviour. This observation points to the idea that in contrast to the findings observed in infected cells (Fig. 1), SV sgmRNA requires eIF4A for efficient translation when transfected in BHK cells. We employed CrPV IGR-luc as a mRNA control that does not use eIF4A. Notably, translation of this mRNA in BHK cells was strongly stimulated by hipp, most probably due to the fact that this inhibitor blocks translation of cellular mRNAs. In this manner, competition of CrPV IGR-luc mRNA with cellular mRNAs for components of the protein synthesizing machinery is abrogated in the presence of hipp. Moreover, this finding clearly indicates that hipp has no deleterious effects on the different steps of translation apart from initiation.

figure

Figure 5. Transfection of BHK cells with in vitro synthesized mRNAs. Effect of hipp treatment.

A. Schematic representation of the mRNAs employed.

B. sgmRNA C+luc, gmRNA SV-Luc, IRES EMCV-luc and CrPV IGR-luc mRNAs synthesized in vitro by T7 RNA polymerase were transfected in BHK cells with Lipofectamine 2000. Different amounts of hipp (0, 0.4 and 0.8 μM) or 100 μg ml−1 cycloheximide were added 30 min later and cells were incubated for 90 min before harvesting to analyse luc activity. Values obtained from cycloheximide-treated cells were used to subtract the amount of luc synthesized prior to hipp addition (0.085 × 106 RLU for sgmRNA C+luc; 0.003 × 106 RLU for gmRNA SV-Luc; 0.04 × 106 RLU for IRES EMCV-luc; 0.003 × 106 RLU for CrPV IGR-luc). Luciferase activity results are means ± SD of three independent experiments performed in triplicate. The percentage values of hipp-treated cells relative to their respective untreated samples are indicated in the figure.

Download figure to PowerPoint

The requirement of eIF4A for the translation of SV sgmRNA or SV gmRNA was also assayed in cell-free systems (rabbit reticulocyte lysates, RRL). In agreement with the findings observed in transfected cells, these two mRNAs were inhibited by hipp (Fig. S2), indicating that both mRNAs (genomic and subgenomic) require eIF4A for efficient translation. These findings reveal that the eIF requirements obtained in cell-free systems or in cells transfected with sgmRNA do not accurately reflect those observed under physiological conditions in SV-infected cells. In conclusion, SV sgmRNA exhibits a dual mechanism for its translation with regard to the eIF4A requirement.

Functioning of eIFs in mosquito cells infected by SV

Alphaviruses are arthropod-borne viruses that can infect two different hosts in nature, vertebrates and arthropods, with a different outcome of infection in each case (Strauss and Strauss, 1994). The most common invertebrate host for alphaviruses is mosquitoes, which are persistently infected without apparent disease. In contrast, when vertebrates are infected by a mosquito bite, a fatal encephalitic disease may occur in some instances. The most striking difference during alphavirus infection between vertebrate and mosquito cells lies in the inhibition of protein synthesis. Thus, cytolytic infection of vertebrate cells in culture leads to the inhibition of protein synthesis and ultimately to cell death, whereas no blockade of cellular translation occurs in mosquito cells persistently infected by alphaviruses. However, some differences may exist when different mosquito cell subclones are infected by SV (Miller and Brown, 1992). Very little is known about the participation of eIFs in the translation of SV sgmRNA in mosquito cells. Recently, activation of the PI3K-Akt-TOR pathway has been observed in arthropod cells but not in BHK cells upon infection with SV (Patel and Hardy, 2012). The stimulation of this pathway phosphorylates 4E-BP1, enhancing the formation of the eIF4F complex (Gingras et al., 1999).

Then, we wanted to examine two key eIFs, eIF2 and eIF4A, in the translation of SV sgmRNA in Aedes albopictus C6/36 cells. The kinetics of protein synthesis in mock or SV-infected mosquito cells is shown in Fig. 6A. In good agreement with previous results, there was little inhibition of cellular protein synthesis in the infected cells, such that both cellular and viral mRNA translation coexisted for several hours and even days. To analyse the participation of eIF2, cells were treated with 200 μM sodium arsenite (Ars), a treatment that leads to robust phosphorylation of eIF2α (Sanz et al., 2009; Welnowska et al., 2011). To our surprise, the translation of SV sgmRNA was strongly inhibited, to an extent similar to that observed with cellular protein synthesis (Fig. 6B and D). This result strongly contrasts with the findings observed in several laboratories, including our group, indicating that eIF2α phosphorylation does not block alphavirus sgmRNA translation in mammalian cells (Gorchakov et al., 2004; McInerney et al., 2005; Ventoso et al., 2006; Sanz et al., 2009). Another important difference found in mosquito cells was that after SV infection, the phosphorylation of eIF2α did not increase (Fig. 6B, middle panel). This result agrees with the fact that insect cells do not encode an orthologue of the PKR gene present in vertebrates (Ventoso, 2012). The effect of thapsigargin was also analysed in these cells, but we found that this compound did not stimulate the phosphorylation of eIF2α (results not shown). The participation of eIF4A was tested by using hipp as described in the above experiments for mammalian cells. This compound potently blocked cellular protein synthesis in mosquito cells, indicating that eIF4A from these cells is sensitive to hipp. As occurred with BHK cells, translation of SV sgmRNA was resistant to hipp treatment (Fig. 6C and D). This finding suggests that eIF4A and most probably the eIF4F complex do not participate in the translation of SV sgmRNA. In conclusion, the mechanism of sgmRNA translation may exhibit differences or similarities between mosquito and vertebrate cells with regard to the requirement for eIFs. One important difference is that this translation requires eIF2 in mosquito cells, while the requirement for eIF4A is similar in both cell types.

figure

Figure 6. Infection of insect cells with SV: effect of Ars and hipp treatments.

A. A. albopictus (C6/36) cells were either mock-infected or infected with SV (MOI of 5 pfu per cell). Protein synthesis was determined by labelling for 1 h with [35S]Met-Cys at the indicated hpi. Samples were processed by SDS-PAGE and autoradiography.

B. A. albopictus (C6/36) cells were infected as in (A). At 15 hpi, cells were treated with 200 μM Ars for 15 min or left untreated and then labelled for 45 min with [35S]Met-Cys in the absence or presence of the compound. Samples were submitted to SDS-PAGE and autoradiography (upper panel). Phosphorylated eIF2α (middle panel) and total eIF2α (lower panel) were analysed by Western blot using the same samples and specific anti-phospho-eIF2α and anti-eIF2α antibodies.

C. C6/36 cells were infected as in (A). At 14.5 hpi, cells were treated with 0.5 μM hipp for 30 min or left untreated. From 15 to 16 hpi, cultures were labelled with [35S]Met-Cys in the absence or presence of the inhibitor. Samples were separated by SDS-PAGE, followed by autoradiography.

D. The percentage of cellular (actin) and viral C protein synthesis in cells treated with Ars or hipp compared with untreated cells was calculated from values obtained by densitometric scanning of the corresponding bands from experiments (B) and (C). The values displayed are the mean ± SD of three independent experiments performed in triplicate.

Download figure to PowerPoint

Analysis of the DLP ability to confer factor-less translation

Both the structure and the context in which SV sgmRNA translation takes place are important determinants of translation initiation (Sanz et al., 2009). An important structural motif in alphavirus sgmRNA, and one that strongly stimulates its translation, is a hairpin located between 25 and 64 nucleotides downstream of the AUG initiation codon (Frolov and Schlesinger, 1994), subsequently named DLP (Ventoso et al., 2006). In fact, DLP enhances mRNA translation in BHK and MEFs cells because its integrity is crucial to translate this mRNA following eIF2α phosphorylation (McInerney et al., 2005; Ventoso et al., 2006). Thus, in PKR−/− MEFs, where the increase in eIF2α phosphorylation is much lower than in PKR+/+ MEFs following alphavirus infection, translation of sgmRNA lacking DLP takes place at almost control levels, whereas little viral protein synthesis is found in PKR+/+ MEFs after infection with this SV variant (ΔDLP) (Ventoso et al., 2006). Consequently, it was of interest to determine whether DLP may also confer independence from other eIFs such as eIF4A. This possibility was evaluated using PKR−/− MEFs, where the translation of this SV variant is not inhibited. First, we wanted to assure that the presence of DLP confers eIF2 independence. Thus, PKR−/− MEFs were infected with wt SV or with the SV variant (ΔDLP) that contains a restructured DLP hairpin (Fig. 7A), and eIF2α phosphorylation was induced using thapsigargin (Tg). Infection of these cells with wt SV or with the ΔDLP variant produced significant amounts of late viral proteins, which was inhibited by Tg only in cells infected with ΔDLP SV but not in those infected with wt SV (Fig. 7B and C). This indicates that ΔDLP SV sgmRNA requires eIF2 for its translation. Therefore, the integrity of the DLP provides eIF2 independent translation in PKR−/− MEFs. Next, the role of DLP to provide independence to eIF4A was analysed by treating infected PKR−/− MEFs with hipp. Interestingly, addition of 0.5 μM hipp had little effect on protein synthesis both in wt SV (about 15%) and in ΔDLP SV (about 25%), suggesting that the DLP does not confer independence to eIF4A.

figure

Figure 7. Analysis of the DLP element present in SV sgmRNA in infected PKR−/− MEFs cells treated with Tg or hipp.

A. Schematic representation of the 5′ end of SV sgmRNA (wt; upper panel, ΔDLP mutant; lower panel) that includes the leader sequence and the DLP structure downstream of AUGi. Secondary structure predicted by the RNAfold program is shown. The initiation AUG codon at position 50 is highlighted in grey and the in-frame AUGs at positions 71 and 107 are highlighted in blue. The nucleotides replaced in the ΔDLP mutant are indicated in red.

B. PKR−/− MEFs cells were either mock-infected or infected with wt or ΔDLP SV (MOI of 5 pfu per cell). At 4.5 hpi, cells were treated with 1 μM Tg or 0.5 μM hipp for 30 min or left untreated. From 5 to 6 hpi, cultures were labelled with [35S]Met-Cys in the absence or presence of the corresponding inhibitors at the same concentrations as before. Radioactive samples were processed by SDS-PAGE and autoradiography (upper panel). In parallel, phosphorylated eIF2α (middle panel) and total eIF2α (lower panel) were detected by Western blot analysis using specific antibodies.

C. The percentage of cellular (actin) and viral C protein synthesis in cells treated with Tg or hipp compared with untreated cells was calculated from values obtained by densitometric scanning of the corresponding bands and are the mean ± SD of three independent experiments performed in triplicate.

Download figure to PowerPoint

To further investigate the activity of DLP in translation, a new construct was engineered placing the luc gene after the capsid sequence and containing the disrupted DLP (rep C+luc ΔDLP). Transfection of PKR−/− MEFs with rep C+luc ΔDLP gave rise to slightly lower levels of luc compared with control rep C+luc (Fig. 8A), indicating that DLP does not enhance translation in cells devoid of PKR. Treatment with 0.5 μM hipp had little inhibition of luc synthesis either with rep C+luc (14% inhibition) or with rep C+luc ΔDLP (16% inhibition). Therefore, hipp has almost no effect on the initiation of translation of sgmRNA in SV replicating cells irrespective of the presence of the DLP. The activity of hipp was also tested in PKR−/− MEFs transfected with sgmRNA C+luc and sgmRNA C+luc ΔDLP (Fig. 8B). In agreement with the results obtained above (see Fig. 5B) hipp blocked (86% inhibition) translation of control sgmRNA C+luc in PKR−/− MEFs. The inhibition of luc synthesis (87%) directed by sgmRNA C+luc ΔDLP by hipp was similar to that of controls. In conclusion, the behaviour of both mRNAs is similar with regard to the effects of hipp. Moreover, our present results indicate that DLP does not stimulate translation in PKR−/− MEFs irrespective of viral replication, or in mosquito cells (see below). Thus, DLP is not a translational enhancer sensu stricto; instead, this element should be regarded as a motif that confers eIF2-independent translation of sgmRNA in SV-infected mammalian cells.

figure

Figure 8. Protein synthesis from sgmRNAs that have the DLP disrupted in PKR−/− MEFs treated with hipp.

A. PKR−/− MEFs cells were transfected with Lipofectamine 2000 and in vitro transcribed rep C+luc or rep C+luc ΔDLP (or Mengo-luc replicon as sensitive control) mRNAs. At 3.5 hpt, cells were treated with 0.5 μM hipp or 100 μg ml−1 cycloheximide for 90 min, or left untreated. At 5 hpt untreated and treated cells were collected in luciferase lysis buffer and luc activity was measured.

B. sgmRNA C+luc or sgmRNA C+luc ΔDLP, IRES EMCV-luc and CrPV IGR-luc (as sensitive and resistant controls respectively) mRNAs synthesized in vitro by T7 RNA polymerase were transfected in PKR−/− MEFs cells with Lipofectamine 2000. 0.5 μM hipp or 100 μg ml−1 cycloheximide were added 30 min later and cells were incubated for 90 min before harvesting to analyse luc activity.

Both in (A) and in (B), values obtained from cycloheximide-treated cells were used to subtract the amount of luc synthesized prior to hipp addition (0.60 × 106 RLU for rep C+luc; 0.44 × 106 RLU for rep C+luc ΔDLP; 0.37 × 105 RLU for Mengo-luc replicon; 0.044 × 104 RLU for sgmRNA C+luc; 0.018 × 104 RLU for sgmRNA C+luc ΔDLP; 0.042 × 105 RLU for IRES EMCV-luc; 0.229 × 104 RLU for CrPV IGR-luc). Luciferase activity results are means ± SD of three representative experiments performed in triplicate. The percentage values of hipp-treated cells relative to their respective untreated cells are indicated in the figure.

Download figure to PowerPoint

To further compare the behaviour of SV sgmRNA in mosquito cells, we examined the translation of mRNAs lacking DLP. As described above, this structure leads to eIF2 independent translation in mammalian cells. Since there is no eIF2α phosphorylation in mosquito cells, both wt sgmRNA and the variant ΔDLP would be expected to be translated to similar extents in mosquito cells. Indeed, this seemed to be the case when mosquito cells were infected with wt SV or the variant ΔDLP, since translation of sgmRNA was similar in both instances at 16 hpi. In addition, ΔDLP SV sgmRNA was sensitive to Ars but not to hipp treatment (Fig. 9A and B) in the same way as wt SV sgmRNA (Fig. 6). Thus, initiation necessitates active eIF2 and therefore, DLP is not functional in mosquito cells. In conclusion, the action of DLP on translation is dependent on the host cell examined.

figure

Figure 9. Analysis of the DLP structure in infected A. albopictus (C6/36) cells treated with Ars or hipp.

A. C6/36 cells were infected with ΔDLP SV (5 pfu per cell). At 15 hpi, cultures were treated with 200 μM Ars for 15 min or left untreated and then labelled for 45 min with [35S]Met-Cys in the absence or presence of the compound (left panel). In parallel, at 14.5 hpi cells were treated with 0.5 μM hipp or left untreated and labelled with [35S]Met-Cys from 15 to 16 hpi in the absence or presence of the inhibitor (right panel). Radiolabelled proteins were separated by SDS-PAGE followed by autoradiography.

B. The percentage of viral C protein synthesis in cells treated with Ars or hipp compared with untreated cells was calculated from values obtained by densitometric scanning of the corresponding bands.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Viral mRNAs must be translated by the cellular protein synthesizing machinery, in such a way that they have to compete with a vast repertoire of cellular mRNAs (Bushell and Sarnow, 2002; Komar et al., 2012). In the case of early viral mRNAs, the synthesis of viral proteins at low levels is usually sufficient for their participation in viral replication. However, at late times of infection when structural proteins are made, viral mRNAs usually need to usurp cellular translation, in order to direct the protein synthesizing capacity of the cell to translate viral mRNAs that will produce virion components. In this regard, late viral RNAs have evolved structures to maximize their translation and to efficiently compete with cellular mRNAs. One strategy could be to evolve viral mRNAs with structures that make them independent of some eIFs. Thus, following viral infection, one or more eIFs that are necessary for cellular translation are inactivated, leading to the inhibition of cellular protein synthesis while viral mRNA translation dominates. For example, in SV-infected mammalian cells, eIF2α becomes phosphorylated, but the initiation of SV sgmRNAs continues by a hitherto unknown mechanism. In the present work we have demonstrated that following entry in the cell, SV gmRNA is immediately translated using eIF4A, as well as other canonical eIFs (Castello et al., 2006). In contrast, eIF4A does not participate in SV sgmRNA translation and most probably the entire eIF4F complex is not required for this initiation at late times of infection (Castello et al., 2006; Sanz et al., 2009). These observations are very surprising because SV sgmRNA contains a cap structure, a poly(A) tail and a 5′UTR almost devoid of secondary structure. In fact, only SV sgmRNA with genuine 5′ region exhibits this independence for some eIFs in a replication context. Interestingly, the same mRNA is translated by a canonical mechanism that requires eIF4A, as well as eIF2 and eIF4G (Sanz et al., 2009), following transfection in mammalian cells. This dual mechanism for translation initiation with respect to some eIFs results in an ability of SV sgmRNA to be translated in different cellular contexts in which the availability of initiation factors varies, as occurs in nature during SV life cycle. In conclusion, SV sgmRNA capacity to initiate translation by distinct mechanisms facilitates SV replication in different hosts.

It is still difficult to envisage the exact mechanism by which SV sgmRNA is translated without some eIFs in a context of infection. In the case of eIF2, perhaps this factor is substituted by eIF2A (Ventoso et al., 2006) or by Ligatin (Skabkin et al., 2010). However, the participation of DLP remains obscure in this regard. Thus, SV sgmRNA containing the DLP element does not require active eIF2 in infected mammalian cells whereas this factor is necessary to translate this mRNA in transfected cells, despite the fact that in both instances the sgmRNA structure is the same. It has been suggested that this structure serves to slow down the passage of ribosomes in this region and stalls ribosomes at the initiation codon, giving them the opportunity to initiate at low concentrations of eIFs (Frolov and Schlesinger, 1994; 1996). However, this possibility does not explain the dual mechanism exhibited by SV sgmRNA (Sanz et al., 2009). We believe that the DLP may have an active role in initiation of translation, and one that is different from the model previously suggested (Frolov and Schlesinger, 1994; 1996). Strikingly, our results show differences in SV sgmRNA translation between PKR−/− MEFs and C6/36 cells. Both cell lines lack PKR and consequently SV replication cannot induce eIF2α phosphorylation. However, after induction of eIF2α phosphorylation by Tg or Ars, SV sgmRNA translation is always abolished in mosquito C6/36 cells even if the DLP is present, and in PKR−/− MEFs only when the DLP structure is disrupted. In contrast, hipp treatment does not inhibit viral translation in any situation. These results suggest a role of DLP in combination with some specific mammalian elements to replace exclusively eIF2. These not well-defined components may be absent or inefficient in mosquito cells. Remarkably, Tg treatment does not reduce viral protein synthesis to any extent in wt SV-infected PKR−/− MEFs (see Fig. 7B), suggesting that the translation initiation mechanism might operate in an eIF2-independent mode even when this factor is in an unphosphorylated and active state, opposite to ΔDLP SV sgmRNA which requires eIF2 for translation. As mentioned above, initiation factors eIF2A and Ligatin have been proposed to replace eIF2 in SV sgmRNA translation. However, their relationship with the DLP structure in SV-infected mammalian cells requires further investigation. Finally, this structure is not responsible for eIF4A-independent translation of SV sgmRNA neither in infected mammalian nor in insect cells, and most probably it does not provide independence for any other eIF different from eIF2.

Cellular protein synthesis is drastically abrogated by alphavirus infection of mammalian cells. Several hypotheses were put forward in the past to explain this phenomenon. One suggestion indicated that the modification of membrane permeability by one or more viral proteins led to an increased ionic concentration in the cytoplasm (Garry et al., 1979; Garry, 1994), as initially suggested for picornaviruses, which, like alphaviruses, encode proteins that enhance membrane permeability (Carrasco and Smith, 1976; Nieva et al., 2012). Another hypothesis was that the capsid protein interacted with ribosomes, promoting the selective inhibition of cellular, but not alphavirus sgmRNA translation (van Steeg et al., 1981). However, SV replicons devoid of the coding region for structural proteins inhibited cellular translation with a similar kinetics and efficiency as non-defective replicons (Frolov and Schlesinger, 1994). Moreover, levels of sgmRNA were low from this replicon devoid of structural proteins, suggesting that mRNA competition was not the cause of the inhibition of cellular translation. This observation suggested that non-structural SV proteins and/or viral RNA replication were responsible for the shut-off of host protein synthesis. The possibility that one eIF, such as eIF2, becomes inactivated after alphavirus infection has been thought to account for the shut-off phenomenon. This eIF2 phosphorylation would specifically block cellular translation, since protein synthesis directed by sgmRNA does not require active eIF2 (Gorchakov et al., 2004; Ventoso et al., 2006; Sanz et al., 2009). Thus, several molecular events can influence cellular translation. One of these events is phosphorylation of eIF2 mediated by PKR, while other unidentified mechanisms not mediated by PKR are also present (Gorchakov et al., 2004). Indeed, alphavirus infection of PKR−/− MEFs blocks cellular translation, even though eIF2 remains unphosphorylated (McInerney et al., 2005; Ventoso et al., 2006). It must be remarked that after several decades of investigation of the shut-off phenomenon in alphavirus-infected cells, we still do not precisely understand its exact molecular mechanism. For this reason, the analysis of cellular protein synthesis in mosquito cells may provide clues to understand this ablation of cellular translation. There are several differences between mammalian and mosquito cells, such as the temperature of incubation and the culture medium employed. We have tested protein synthesis at 37°C in a CO2 incubator and have found that also under these conditions no inhibition of host translation occurs in mosquito cells by SV infection (results not shown). Interestingly, in mosquito cells, SV infection does not seem to induce eIF2α phosphorylation (our current results and Ventoso, 2012). Moreover, translation of SV sgmRNA is blocked by arsenite, a compound that induces eIF2α phosphorylation. This result was totally unexpected and suggests that the DLP element does not provide eIF2-independence by itself. However, in these invertebrate cells eIF4A is not necessary for initiation of translation of SV sgmRNA. Most probably, the eIF4F complex is not involved in the initiation of sgmRNA translation in these cells, as occurs in mammalian cells (Castello et al., 2006). The initiation events on this viral mRNA that is capped without the participation of the eIF4F complex remain a puzzle. Perhaps, this complex is substituted by a viral protein, as has been described for hantavirus N protein (Mir and Panganiban, 2008). On the other hand, the role of eIF2 can be replaced in mammalian cells but not in mosquito cells, suggesting that one or more components specific to vertebrate cells, together with DLP, carry out this function. Future efforts to reveal more details of the exact mode by which SV sgmRNA initiates translation will provide further insights into how a capped mRNA with a short 5′UTR almost devoid of secondary structure is translated without the participation of several eIFs.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Cell lines and viruses

The cell lines used in this work were BHK-21 (baby hamster kidney, obtained from ATCC), PKR−/− MEFs (mouse embryo fibroblasts) (Yang et al., 1995) and A. albopictus C6/36 (obtained from ATCC). BHK-21 and PKR−/− MEFs cells were grown at 37°C, 5% CO2 in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal calf serum (FCS) or 10% FCS respectively. C6/36 cells were cultured at 28°C without CO2 in M3 medium supplemented with 10% FCS. Viral infections of BHK-21 and PKR−/− MEFs cells were carried out in DMEM without serum for 1 h at 37°C, whereas infections of C6/36 cells were performed at 28°C. This medium was then removed and infection continued in DMEM with 5% FCS, DMEM with 10% FCS or M3 medium with 10% FCS, respectively, at the same temperatures as before. Infections with wt SV, ΔDLP SV, VSV and EMCV were carried out at a multiplicity of 5 pfu per cell.

Inhibitors

The following chemical inhibitors were used in this work at the indicated concentrations: hippuristanol (Bordeleau et al., 2006), thapsigargin (Sigma), sodium arsenite (Riedel-de Haën) and cycloheximide (Sigma).

Plasmids and recombinant DNA procedures

The plasmids employed in this work are listed and described in Table S1. The plasmid pT7 rep C+luc ΔDLP was specifically designed for this work. It was made by inserting the HpaI/AatII-digested product from pT7 SV ΔDLP (Ventoso et al., 2006) in the same sites of rep C+luc. To obtain the corresponding sgmRNA by in vitro transcription, the plasmid pT7 C+luc ΔDLP was made using pT7 rep C+luc ΔDLP as template in a PCR reaction with oligonucleotides 5′SacI-T7prom (GCGCGCGAGCTCTAATACGACTCACTATAGATAGTCAGCATAGT) and 3′Aat (CGTTCTTGACGTCGAAC). The PCR product was then digested with SacI and AatII restriction endonucleases and inserted in the same sites of rep C+luc.

In vitro RNA transcription

Linearized plasmids were used as templates for in vitro RNA transcription using T7 RNA polymerase (New England Biolabs). Reactions containing 10 ng μl−1 template DNA, 1000 U ml−1 of T7 RNA polymerase, 1× RNAPol Reaction Buffer, 1 mM m7G(5′)ppp(5′)G RNA Cap Structure Analog (New England Biolabs), 0.5 mM ATP, 0.5 mM CTP, 0.5 mM UTP, 0.25 mM GTP and 0.32 U μl−1 RNaseOUT Recombinant Ribonuclease Inhibitor (Invitrogen), were incubated for 2 h at 37°C. The template DNA was then digested with 0.1 U μl−1 RNase-free Recombinant DNase I (Takara) for 10 min at 37°C. The transcription mixture always contained the m7G(5′)ppp(5′)G cap analogue except when T7 Rluc ΔEMCV IGR-Fluc and pTM1-luc were used as templates.

RNA transfection

RNAs transcribed in vitro were transfected using Lipofectamine 2000 (Invitrogen) according to the supplier's recommendations.

In vitro translation

In vitro translation was carried out in nuclease-treated RRL (Promega) according to the manufacturer's instructions. One hundred nanograms of in vitro transcribed mRNAs were added to translation mix. Protein synthesis was estimated by measuring luc activity.

Measurement of luciferase activity

Cells were lysed in a buffer containing 0.5% Triton X-100, 25 mM glycylglycine (pH 7.8) and 1 mM dithiothreitol. Luc activity was determined using a Monolight 2010 luminometer (Analytical Luminiscense Laboratory) and the Luciferase Assay System (Promega).

Analysis of protein synthesis

Protein synthesis was analysed at the indicated times by replacing the growth media with 0.2 ml of DMEM without methionine-cysteine supplemented with 1 μl of EasyTagTM EXPRESS 35S Protein Labelling mix, [35S]Met-Cys (11 mCi ml−1, Perkin Elmer) per well of an L-24 plate. The cells were then collected in sample buffer, boiled for 5 min and analysed by autoradiography of SDS-polyacrylamide gels (15%). Protein synthesis was quantified by densitometry using a GS-800 Calibrated Densitometer (Bio-Rad).

Western blotting

Cells were collected in sample buffer, boiled for 5 min and processed by SDS-PAGE. After electrophoresis, proteins were transferred to a nitrocellulose membrane. Specific rabbit polyclonal antibodies raised against phospho-eIF2α (Ser 51) (Cell Signaling Technology) or total eIF2α (Santa Cruz Biotechnology) were used at 1:1000 dilution. Anti-rabbit immunoglobulin G antibody coupled to peroxidase (Amersham) was used at a 1:5000 dilution. Protein bands were visualized with the ECL detection system (Amersham).

Immunofluorescence assays

Fixation, permeabilization and confocal microscopy were performed as described previously (Madan et al., 2008), using a confocal LSM510 lens coupled to an Axio Imager.Z1 microscope (Zeiss). Primary antibodies used were: rabbit polyclonal anti-C SV (Sanz et al., 2009), mouse monoclonal anti-eIF4A (a generous gift from Dr H. Trachsel, Institute for Biochemistry and Molecular Biology, University of Berne, Switzerland) and mouse monoclonal anti-eIF4AII (sc-137147, Santa Cruz Biotechnology), goat polyclonal anti-eEF2 (sc-13003, Santa Cruz Biotechnology) and goat polyclonal anti-TIA-1 (sc-1751, Santa Cruz Biotechnology). Specific antibodies conjugated to Alexa 488, Alexa 555 or Alexa 647 (A-21202, A-21432 and A-31573 respectively; Invitrogen) were employed as secondary antibodies. Image processing was performed with Huygens 4.1 software (Scientific Volume Imaging B.V.). The Pearson's correlation coefficients were obtained using the ImageJ software JACoP plug-in (Bolte and Cordelieres, 2006).

Secondary structure prediction

RNA optimal secondary structures were predicted using the RNAfold webServer (Hofacker, 2003): http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

This study was supported by a DGICYT (Dirección General de Investigación Científica y Técnica. Ministerio de Educación y Ciencia. Spain) grant (BFU2009-07352). M.G.-M. is holder of FPI (Formación de Personal Investigador) fellowship. J.P. is funded by the Canadian Cancer Society Research Institute. The Institutional Grant awarded to the Centro de Biología Molecular ‘Severo Ochoa’ (CSIC-UAM) by the Fundación Ramón Areces is acknowledged.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
cmi12079-sup-0001-fS1.pdf1411K

Fig. S1. Subcellular localization of eEF2 and SV C protein in mock and SV-infected BHK cells. BHK cells were seeded on glass coverslips and mock-infected or infected with SV (5 pfu per cell). At 3.5 or 6.5 hpi, cells were treated for 90 min with 0.5 μM hipp or left untreated. At 5 or 8 hpi, cells were fixed, permeabilized and processed for immunofluorescence using anti-eEF2 (red) and anti-SV C protein (cyan) respectively. Images were acquired on a confocal microscope and subsequently processed with Huygens 4.1 software. Merged images represent the dual visualization of eEF2 and C.

cmi12079-sup-0002-fS2.pdf89K

Fig. S2.In vitro translation of SV mRNAs in RRL. Effect of hipp treatment. One hundred nanograms of sgmRNA C+luc or gmRNA SV-Luc mRNAs synthesized in vitro by T7 RNA polymerase were added to RRL in the presence of different concentrations of hipp (0, 0.8 and 1.6 μM) and incubated for 90 min at 30°C. Luc synthesis was estimated by measuring luc activity. The percentage values of hipp-treated samples relative to their respective untreated samples are represented in the figure and are the mean ± SD of three representative experiments performed in triplicate.

cmi12079-sup-0003-tS1.doc207K

Table S1. Nomenclature of plasmids, replicons, sgmRNAs and mRNAs.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.