|Home | About | Journals | Submit | Contact Us | Français|
The mechanisms utilized by viruses to protect their transcripts from the cellular RNA decay machinery, as well as the biological relevance of this protection, are largely unknown. We demonstrate that Sindbis virus uses U-rich 3’ UTR sequences in its RNAs to recruit the cellular HuR protein during infections of both human and mosquito cells. HuR binds viral RNAs with high specificity and affinity. Furthermore, Sindbis virus induces the selective movement of HuR protein out of the nucleus of mammalian cells during infection thereby increasing the cytoplasmic pool of the protein available to the virus. Finally, knockdown of HuR results in a significant increase in the rate of decay of Sindbis virus RNAs and diminishes viral yields in both human and mosquito cells. Collectively these data indicate that Sindbis virus, and likely other alphaviruses, usurp the HuR protein to avoid the cellular mRNA decay machinery and maintain a highly productive infection.
Cellular RNA decay is a robust process by which the cell rapidly removes unwanted or aberrant RNAs from its transcriptome (Garneau et al., 2007). A significant proportion of cellular gene regulation rests at the level of post-transcriptional control via the selective degradation of mRNAs (Cheadle et al., 2005; Garcia-Martinez et al., 2004). Thus the cellular mRNA decay machinery serves as an effective control mechanism for the quantity and quality of mRNAs in the cytoplasm. The members of genus Alphavirus of the family Togaviridae are a group of geographically diverse single-stranded positive-sense RNA viruses (Strauss and Strauss, 1994). The genomic and subgenomic RNAs of the alphaviruses closely resemble the cellular mRNAs produced by RNA polymerase II, as they are 7meGpppG capped at their 5’ end and 3’ polyadenylated. Therefore these viral mRNAs likely have the capacity to interface with cellular RNA decay factors during infection. The goal of this study was to determine how these viral transcripts escape surveillance by the cellular mRNA decay machinery.
For the majority of cellular mRNAs the primary and rate-limiting step of degradation is the removal of the 3’ poly(A) tail by one or more cellular deadenylases (Xu et al., 2001; Wilson and Treisman, 1988). Deadenylation of mRNAs results in translational silencing, as well as serving to expose the 3’ end of the transcript to exonucleolytic degradation by the exosome (Schmid and Jensen, 2008) or prime the transcript for decapping and subsequent 5’-3’ exonucleolytic digestion by XRN1 (Franks and Lykke-Andersen, 2008). Therefore one effective method for viral transcripts to evade the cellular mRNA decay machinery would be to inhibit the deadenylation step. In fact Sindbis virus (SinV), the model Alphavirus, contains multiple elements in its 3’ UTR that we recently demonstrated are capable of independently repressing deadenylation (Garneau et al., 2008). Using both tissue culture-based assays and a cell-free RNA decay system (Sokoloski et al., 2008a; Sokoloski et al., 2008b), we have established that the Repeated Sequence Elements (RSEs) as well as a ~40 base U-Rich Element in conjunction with the 19nt 3’-terminal Conserved Sequence Element (URE/CSE) are capable of repressing deadenylation. The ability of the URE/CSE region to repress deadenylation in vitro was associated with the binding of a 38 kDa cellular trans-acting factor. Examination of the 3’ UTRs of other viruses within the Alphavirus genus reveals that while the overall 3’ UTR sequences may be divergent, the presence of the URE is well conserved in most Alphavirus species (Ou et al., 1982; Strauss and Strauss, 1994), with notable exceptions being O’nyong nyong, Chikungunya and Ross River viruses. Taken together, these data confirmed our hypothesis that RNA viruses, such as SinV, do in fact interface with the cellular mRNA decay machinery.
In this study we determined the mechanism by which SinV represses the degradation of its transcripts during infection of human and mosquito cells. The URE/CSE region of multiple alphaviral 3’ UTRs is bound specifically and with high affinity to the cellular HuR protein, a known regulator of cellular mRNA stability (Hinman and Lou, 2008; Abdelmohsen et al., 2009). This interaction occurs during SinV infection in both Aedes and human cells. While the mosquito HuR homolog (aeHuR) is largely cytoplasmic, in human cells SinV infection induces a dramatic translocation of the HuR protein from the nucleus to the cytoplasm where SinV viral RNAs accumulate. Knockdown of HuR protein in either Aedes or human cells significantly destabilizes SinV mRNAs and reduces viral yields. Taken together, these studies establish the HuR protein as a novel cellular factor required for efficient alphavirus infection. Furthermore these studies suggest that other viruses may also have evolved ways to interface with the cellular RNA decay machinery to stabilize their transcripts to promote a productive infection.
Given the major roles of RNA decay in gene regulation and disposal of unwanted transcripts, capped and polyadenylated positive-sense RNA viruses have likely developed some method of successfully interfacing with the cellular RNA decay machinery to stabilize their transcripts during the course of an infection. Stability elements of cellular mRNAs are often located within their 3’ UTRs and have been shown to regulate transcript-specific decay (Garneau et al., 2007). We have previously demonstrated that SinV RNAs, like cellular mRNAs, contain stability elements in their 3’ UTRs (Garneau et al., 2008). These stability elements serve to repress deadenylation in tissue culture models of SinV infection and cell-free systems. Interestingly, we determined that a major stability determinant within the SinV 3’ UTR is a ~60 base U-rich region at the 3’ end, termed the U-Rich Element/Conserved Sequence Element (URE/CSE). The URE portion of this region previously had no ascribed function despite being conserved, at least in nucleotide bias, amongst numerous members of the genus (Strauss and Strauss, 1994).
Given the conservation of the URE/CSE among the alphaviruses, we sought to determine if the repression of deadenylation observed with the SinV URE/CSE was indeed a common property of the URE/CSEs found in other members of the genus. We demonstrated previously that a cell-free mRNA decay system that we developed using mosquito cytoplasmic extracts faithfully reproduced SinV URE/CSE-mediated RNA stabilization (Opyrchal et al., 2005; Garneau et al., 2008; Sokoloski et al., 2008a). Therefore, we used this assay to evaluate the ability of the URE-bearing alphaviruses to repress deadenylation. Internally radiolabeled, capped and polyadenylated RNA substrates that contained either a nonspecific reporter sequence or an insertion of the ~60 base URE/CSE region of SinV, Venezuelan Equine Encephalitis (VEE), Eastern Equine Encephalitis (EEE), Western Equine Encephalitis (WEE) or Semliki Forest Virus (SFV) were incubated in a cell-free mRNA decay assay with mosquito cell cytoplasmic extract. As show in Figures 1A and B, while incubation of the control RNA substrate in the system resulted in the rapid removal of the poly(A) tail and accumulation of a deadenylated (A0) intermediate (more than 50% of the substrate was completely deadenylated within 9 minutes), the RNAs containing each of the alphavirus URE/CSE regions exhibited a 5-fold or greater stability relative to the deadenylation rate of the control RNA substrate. This strongly suggests that the URE/CSEs of many alphaviruses are indeed bona fide mRNA stability elements and that the ability to repress deadenylation is a conserved property of these viruses.
Using competition analyses in cell-free assays, we have previously shown that the repression of deadenylation associated with the 3’ UTR of SinV is mediated through the interaction of host protein(s) (Garneau et al., 2008). In addition, using UV cross-linking in conjunction with these competition assays we are able to correlate the binding of a 38 kDa cellular factor to the SinV 3’ UTR with the repression of deadenylation in our cell-free system. As shown in Figure 1C and D, the binding site for this 38kDa factor maps to the URE/CSE region of the SinV 3’ UTR. Interestingly, the URE/CSE region of the other four alphaviruses that repressed deadenylation in Fig. 1A all had a nearly identical pattern of UV cross-linked proteins as the SinV URE/CSE – including the 38 kDa band of interest (Figure 1E). While UV cross linking of the 38 kDa protein was resistant to the addition of 2.6 μg/μl heparin, the faster migrating 32 kDa protein band seen in Figs. 1D and E was effectively competed from the RNA by the addition of the polyanionic competitor suggesting that its interaction could be non-specific (data not shown).
Taken together, these data strongly suggest that many alphaviruses, despite considerable evolutionary divergence (Strauss and Strauss, 1994; Ou et al., 1982), may have maintained a similar strategy to evade the cellular mRNA decay machinery. Curiously, despite the apparent primary sequence heterogeneity of the Alphavirus 3’ UTRs as a whole, the conserved nucleotide bias of the URE (Figure S1A) was sufficient to repress deadenylation and cross-link to similar proteins in each case. The retention of function, despite the fluidity of primary sequence identity, underscores the potential impact of the virus / RNA decay machinery interface on positive-strand RNA virus biology.
In order to determine how the 38 kDa cellular factor functions to stabilize alphaviral RNAs, we needed to know its identity. To this end, we used the URE/CSE region of the 3’ UTR that we delineated as being necessary and sufficient for binding of the 38kDa protein (Fig. 1D) in an in vitro affinity purification strategy. Briefly, 5’-biotinylated RNA oligomers consisting of either the 3’ terminal 54 bases of SinV or a nonspecific control sequence were bound to streptavidin-agarose resin. C6/36 Aedes albopictus mosquito cell cytoplasmic extract was incubated with the resin and unbound proteins were removed by rigorous washing. Retained proteins were eluted, resolved using SDS-PAGE and detected by silver staining. As shown in Figure S2A, several host proteins specifically bound to the SinV USE/CSE RNA oligomer compared to the control. The 38 kDa band was excised and analyzed via tandem mass spectrometry following trypsin digestion.
Given the current lack of a complete Aedes albopictus genome, the Aedes aegypti genome (Nene et al., 2007) was utilized as a surrogate for the database search of the mass spectrometry data. The analysis revealed with high probability that the 38 kDa factor was a mosquito ELAV-superfamily member (AAEL008164) with notable homology to the mammalian HuR protein, a known mRNA stability factor (Hinman and Lou, 2008; Abdelmohsen et al., 2009). Given the high degree of homology between the Aedes aegypti hypothetical protein and human HuR ((~55% identical according to BLAST analysis; Figure S3), we have chosen to refer to the Aedes protein as aeHuR henceforth.
Since Drosophila anti-ELAV monoclonal antibodies failed to recognize Aedes ELAV proteins (data not shown), we first needed to develop reliable immunological reagents specific to aeHuR in order to confirm the identity of the 38 kDa factor implicated in viral RNA stability. Recombinant aeHuR protein was prepared in E. coli and used to generate polyclonal antisera in rabbits. As seen in Fig. S2B, these antibodies specifically detected a 38 kDa protein in a western blot of C6/36 mosquito cell cytoplasmic proteins. We next assessed whether this anti-aeHuR sera would recognize and specifically precipitate the 38 kDa protein cross-linked to the SinV URE/CSE that we previously correlated with repression of deadenylation. As seen in Fig. 2A, the 38 kDa cross-linked band was specifically immunoprecipitated using aeHuR antisera but not with control pre-immune sera. Similar data was obtained for immunoprecipitation of the 38 kDa protein cross-linked to the URE/CSE of the other 4 alphaviruses analyzed in Fig. 1 (Figure S4A).
Next we wished to examine whether the URE/CSE region of the SinV 3’ UTR was capable of interacting with the mammalian HuR protein. Using UV cross-link / immunoprecipitation assays with HeLa cytoplasmic extract we confirmed that this was indeed the case. The URE/CSE element of SinV (Fig. 2B), as well as the URE/CSE elements of VEE, EEE, WEE and SFV (Figure S4B), were capable of cross-linking to HuR protein. Non-specific, control RNAs fail to cross link to HuR in these assays (Garneau et al, 2008; data not shown). Therefore, the ability of HuR protein to interact with all five alphavirus 3’ UTRs is conserved across both mammalian and vector mosquito species.
Finally, while aeHuR and HuR are clearly capable of interacting with the alphaviral URE/CSE elements in cell-free assays, we sought to extend this observation to tissue culture cells during the course of an infection. Either 293T (human embryonic kidney) or Aag2 (Aedes aegypti) cells were infected with wild type SinV at an MOI of 5. At 12 hpi, formaldehyde was added to the cells to stabilize protein:RNA complexes. Cell lysates were prepared and immunoprecipitation analyses were performed using anti-aeHuR sera, anti-HuR (3A2) antibodies, or control pre-immune sera (to detect nonspecific interactions). Following reversal of the cross-linking, SinV genomic and subgenomic RNAs were detected in immunoprecipitated samples using RT-PCR. The retention of SinV RNA with specific anti-aeHuR and HuR antibodies (Figure 2C and 2D, respectively), but not the control pre-immune sera or with antibodies against unrelated proteins (e.g. DCP2, tubulin; data not shown), clearly indicated that SinV RNA indeed interacts with these ELAV-superfamily members during the course of an infection in tissue culture cells.
Taken together, these data identify a novel interaction between the SinV 3’ UTR URE/CSE element and HuR proteins in both cell-free assays and tissue culture models of infection. Furthermore, conservation of the interaction of the URE/CSEs of SinV, VEE, EEE, WEE and SFV with HuR proteins indicates that alphaviruses have evolved this interaction for an important reason - perhaps to successfully elude the host mRNA decay machinery.
The next question we wished to address was how effectively viral transcripts interact with the cellular HuR protein. We utilized Electrophoretic Mobility Shift Assays (EMSA) to determine the binding affinity of aeHuR and HuR for the URE region of the 3’ UTR of our set of alphaviruses. As shown in Figure 3A, recombinant aeHuR interacted with very high affinity to the URE/CSE of SinV (mean dissociation constant 0.16nM). Recombinant human HuR binds the URE/CSE of SinV with similar high affinity (Figure 3B). These high affinity interactions were also specific, as recombinant aeHuR or human HuR both failed to interact with a nonspecific control RNA. We next assayed RNA substrates containing the URE elements from VEE, EEE, WEE and SFV by EMSA and obtained dissociation constants for aeHuR binding in a similar range as obtained for the SinV URE (Figure 3C). Interestingly, the affinity observed for aeHuR and human HuR interactions with the tested alphavirus 3’ UTR elements was comparable to published affinities for HuR with cellular mRNAs (Nabors et al., 2001). Therefore, we conclude that five alphaviruses tested all contain a high affinity binding site for HuR from various species in their 3’ UTRs.
While the data presented above suggest that alphavirus transcripts bind HuR with a relative high affinity, the majority of HuR in a mammalian cell resides in the nucleus rather than the cytoplasm where it can be accessed by viral RNAs (Kim et al., 2008). The subcellular localization of the mosquito HuR homolog has never been examined. Therefore HuR could very well be limiting during an alphaviral infection and thus have only a minor role.
In order to begin to address this issue, we first assessed the subcellular localization of aeHuR in mosquito cells by immunofluorescence assays using the antibodies we developed (Fig. S2B). As shown in Figure 4A, aeHuR is disseminated throughout the Aag2 cell - with a significant amount present in the cytoplasm. The subcellular distribution of aeHuR in Aedes albopictus (C6/36) cells was similar to that observed in Aag2 cells (data not shown). Therefore, we conclude that unlike human cells, a substantial proportion of aeHuR is present in the cytoplasm of mosquito cells and should therefore be readily accessible to alphavirus transcripts during infection.
Given the difference in HuR subcellular localization between human and mosquito cells, we next addressed whether cytoplasmic HuR protein may indeed be a limiting factor during infection. Interestingly, HuR has been shown to relocalize from the nucleus to the cytoplasm in reaction to stimuli that cause a stress response in cells (Kim et al., 2008). Therefore we tested the hypothesis that SinV infection may cause HuR to relocalize to the cytoplasm. Human embryonic kidney cells (293T) were infected with SinV and the subcellular localization of HuR was assessed at 6 and 12 hours post infection (hpi). As seen in Fig. 4B, while HuR is largely nuclear at the start of the infection, there is a dramatic relocalization to the cytoplasm by 6 hpi. Furthermore at 12hpi, the majority of HuR has been shuttled out of the nucleus to the cytoplasm. There was a direct association between the cells that were infected with SinV (as determined by FISH analysis using a probe for the SinV E1 region) and relocalization of HuR to the cytoplasm (Fig. 4C). Similar data were obtained at MOIs of 5, 10, or 20 (data not shown). The relocalization of HuR from the nucleus to the cytoplasm during a SinV infection could also be demonstrated by subcellular fractionation and western blot analysis (Fig. 4D). The relocalization of HuR protein to the cytoplasm is a specific phenomenon as PABPN1 (Fig. 4D) as well as the abundant nuclear protein nucleophosmin (data not shown) both remain confined to the nucleus throughout the SinV infection. Finally, aeHuR maintained its relative nuclear/cytoplasmic distribution during SinV infection of mosquito cells (data not shown), suggesting that HuR relocalization is specific to mammalian cells that contain low levels of cytoplasmic HuR prior to infection.
In conclusion, these data demonstrate that both aeHuR and HuR are present within the cytoplasm of infected cells. In mosquito cells this is due to the natural cytoplasmic localization of aeHuR. Within human 293T cells, viral-induced relocalization of HuR serves to increase the available concentration of HuR in the cytoplasm.
Finally, while aeHuR and HuR bind to alphaviral 3’ UTRs with high affinity and the aeHuR-viral RNA interaction correlates with increased viral RNA stability in our cell-free RNA decay assays, it is still crucial to demonstrate whether HuR truly plays a role in viral RNA stability and the efficiency of viral replication in living cells. Therefore we used a shRNA knockdown approach to assess the effect of a reduction of the cellular levels of aeHuR and HuR on SinV infection.
In three independent experiments, 293T cells were transfected with either HuR-specific shRNA vectors or mock transfected using pLKO-1-puro vector DNA and 12 hrs later were infected with a variant of SinV that contains a temperature-sensitive mutation in its polymerase (ts6SinV). At 10 hpi (which was 22 hours post-transfection with the shRNA vectors), cells were switched to the non-permissive temperature to inhibit viral transcription and total RNA was recovered at various intervals to assess viral RNA half lives by qRT-PCR. During the infection, the level of HuR in cells was reduced an average of ~50-60% compared to mock treated 293T cells based on qRT-PCR (Figure 5A) or western blot analyses (Figure S5). The relative abundances at each time point for both viral RNA species over the three independent experiments were averaged and used to calculate the mean rate of decay for both the genomic and subgenomic RNA species in control 293T cells and HuR knockdown cell lines. These values were plotted with respect to time and an exponential regression curve was fitted to the data points. As shown in Figure 5B, the mean half life of both the genomic and subgenomic RNAs was significantly decreased in the HuR depleted 293T cells indicating an increased rate of viral RNA decay. A comparable increase in the rate of viral RNA decay was observed in a pool of stably transfected Aag2 mosquito cells that were knocked down for aeHuR (Fig. 5C and D). Similar data were obtained when samples were analyzed using an RNAse protection assay (data not shown).
To assay viral replication in a HuR deficient environment, 293T cells were treated with anti-HuR shRNA vectors, or mock treated as described above. Twenty-two hours after transfection, cultures were infected with ts6 SinV at an MOI of 5 and aliquots were removed over the next 15 hours for determination of virus yield by plaque titration. As shown in the one-step growth curve in Fig. 6A, the growth kinetics of SinV were significantly delayed in HuR deficient 293T cells and a >10 fold-repression in viral growth was observed. A statistically significant 5 fold reduction in the growth kinetics of SinV was also observed in a stable pool of Aag2 cells that were knocked down for aeHuR (Fig. 6D). Note that these HuR and aeHuR knockdown cells were viable and showed no apparent growth defects that could overtly account in an indirect way for any of the observations made in this study.
Taken together, these data demonstrate that both aeHuR and HuR are important cellular factors for efficient alphavirus infection in tissue culture cells. Even a modest ~50-60% reduction in aeHuR or HuR abundance resulted in a significant destabilization of SinV RNAs. In order to verify this using a complementary set of experiments, the URE region of the 3’ UTR of SinV that binds to the HuR protein (Fig. 3) was deleted. As seen in Figs 6B and 6E, this ΔURE ts6-SinV variant showed a significant repression in viral growth compared to ts6 SinV containing a wild type 3’ UTR in either 293T or mosquito Aag2 cells – similar to the repression observed in HuR knockdown cells in Figs. 6A and 6D. Furthermore, the ΔURE ts6 SinV variant virus did not demonstrate any additional growth defects in 293T or Aag2 cells that were knocked down for HuR (Figs 6C and 6F). Taken together, these results confirm that aeHuR and HuR are indeed viral RNA stability factors that act through a specific binding site in the viral 3’ UTR and help determine the outcome of an infection. Additionally, these data elucidate a previously unknown facet of Alphavirus biology that potentially could be exploited for the development of new anti-viral strategies.
The cellular HuR protein has been identified as an important stability factor for >50 cellular mRNAs (Wilusz and Wilusz, 2007). In response to cellular proliferation or stimulation by a variety of factors (stress, immune modulation, etc.), HuR protein will often relocalize from the nucleus to the cytoplasm and play a vital role in regulating the stability and translation of select populations of mRNAs (Zhang et al., 2009; Antic et al., 1999; Abdelmohsen et al., 2008; Fan and Steitz, 1998). This study demonstrates a novel function for the cellular HuR protein – it plays a significant role in supporting a SinV infection (and likely other alphavirus infections) in both mammal and vector host cells.
Our working model for how HuR promotes SinV infections is shown in Figure 7. HuR interacts with high affinity to a U-rich region near the 3’ end of SinV mRNAs and stabilizes the transcripts during infection. When SinV infects mammalian cells, HuR is largely nuclear. However by 6 hours post infection when levels of viral RNA synthesis are high, the protein has been induced by the virus to relocate to cytoplasm where it is readily available for binding to SinV transcripts. Knockdown of HuR protein results in reduced stability of SinV mRNAs and a significant reduction in the yields of progeny virions. Thus these studies suggest that HuR is an important host factor that is usurped by viruses to protect their transcripts from the major pathways of the cellular RNA decay machinery. Furthermore, these studies clearly document the importance of the interface between viral mRNAs and the cellular RNA decay machinery.
It is interesting to note that unlike mammalian HuR, a substantial amount of aeHuR protein is cytoplasmic in both Aag2 Aedes aegypti and C6/36 Aedes albopictus cells. This may reflect an innate difference in the way HuR functions in insects versus vertebrates in terms of finding its RNA substrates and helping to define RNA regulons (Keene, 2007). Thus it will be interesting to characterize the roles and regulation of aeHuR in the mosquito system, as further comparative analyses may give significant insight into its mechanisms of action. Curiously, Sindbis viral infection does not alter the subcellular localization of aeHuR in mosquito cell lines as it does in mammalian cells (data not shown). This may be important for preventing the dramatic changes in cellular gene expression influenced by relocalization of HuR (Zhang et al., 2009), thus reducing cytopathology in the mosquito allowing survival of the insect to ultimately serve as an effective vector.
The underlying mechanism for the relocalization of HuR during SinV infection is also unclear. We are currently pursuing two main hypotheses to gain insight into this area. First, a variety of cellular stresses such as heat shock (Gallouzi et al., 2000) and oxidative stress (Mazroui et al., 2008) cause HuR to rapidly shift from the nucleus to the cytoplasm. Perhaps the general stress induced by SinV infection is triggering signaling mechanisms along the same lines as these environmental stresses (McInerney et al., 2005). Alternatively, SinV could be specifically targeting HuR or its transport mechanisms to actively induce HuR protein relocalization. Interestingly, a significant fraction of SinV and SFV nsP2 protein is found in the nucleus of infected cells (Atasheva et al., 2007; Garmashova et al., 2006; Frolov et al., 2009) as well as bound to the ribosome (Ranki et al., 1979) and is associated with significant cytotoxicity. Given the nuclear localization of HuR as well as its role in regulating translation (Kawai et al., 2006), it is tempting to speculate that a viral factor such as the nsP2 protein may be specifically targeting HuR and promoting its relocalization.
The observation that four other alphaviruses (VEE, EEE, WEE and SFV) in addition to SinV interact with HuR suggests an evolutionary conservation of function that further supports the significance of HuR protein-RNA interactions to a productive alphavirus infection. HuR protein-RNA interactions have also been documented for several other viruses. HuR protein binds to the untranslated regions of Hepatitis C virus (HCV) (Spangberg et al., 2000; Harris et al., 2006) and has been recently shown to activate translation (Rivas-Aravena et al., 2009). Interestingly, one study has shown that siRNA knockdown of HuR expression in cells decreased RNA and protein expression from HCV viral replicons (Korf et al., 2005). These observations suggest that HuR could perhaps stabilize the non-polyadenylated transcripts of HCV and other Flaviviruses in a manner at least in part related to what we observed in this study with the alphaviruses. The potential role of HuR protein in infections with retroviruses or DNA viruses appears to be more complex than for the cytoplasmic RNA viruses. In Human Immunodeficiency Virus (HIV) infections, the HuR protein has been shown to interact with the viral reverse transcriptase (Lemay et al., 2008) and to have a negative impact on HIV Internal Ribosome Entry Site (IRES)-mediated translation (Rivas-Aravena et al., 2009). Herpesvirus Saimiri virus small HSUR RNAs can specifically interact with the HuR protein (Cook et al., 2004) although the functional impact of this interaction is not clear. Finally, HuR protein is up-regulated in neoplastic cells that contain Human Papilloma Virus (HPV sequences) (Cho et al., 2006; Sokolowski et al., 1999) and HuR protein has been associated with the post-transcriptional regulation of late HPV gene expression through interactions with the 3’ UTR of late HPV transcripts (Koffa et al., 2000). Determining the mechanistic role of HuR protein in these viral infections may not only give important insights into viral-host interactions but could also help further characterize the underlying mechanisms of HuR function in host cells.
BHK-21, Vero and 293T cell lines were cultured in HyQ MEM/EBSS media with 10% fetal bovine serum (FBS). Aedes aegypti Aag2 cells were maintained in Schneider’s Drosophila medium supplemented with 10% FBS. HeLa S3 spinner cells were maintained in JMEM with 10% Horse Serum. C6/36 Aedes albopictus suspension cells were cultured in SF-900II (Gibco) serum free media.
Full length SinV genomic RNAs were produced by in vitro transcription of either wild type SinV AR339 or the temperature sensitive ts6SinV AR339 clone (Barton et al., 1988; Bick et al., 2003), as previously described (Garneau et al., 2008). The ΔURE SinV variant which contained a 30 base deletion of the URE was constructed using the primers 5’-ATTTTGTTTTTAACATTTCA(37)GGGAATTC and 5’-TTATGCAGACGCTGCGTGGCATTATGC to jump from the CSE to the region upstream of the URE in the pToto1101/ts6 SinV AR339 vector using PCR (Garneau et al, 2008). Viral titers were determined by plaque titration on Vero cells.
The Hygromycin Phosphotransferase (hph) gene was isolated from pHyg (Gritz and Davies, 1983) via PCR using the primers 5’-CATACATGTTCATGAAAAAGCCTGAACTCACCGCG and 5’-CATCTCGAGCTATTCCTTTGCCCTCGGACGAGTG. PCR products were cut with PciI and XhoI and inserted into pBiEx-1 (Promega) to create pBiEx-hph. An Aedes aegypti U6 promoter-driven shRNA expression cassette was generated via PCR from pAedes1 (Konet et al., 2007) using the primers 5’-CATGGGCCCGAATGAATCGCCCATCGAGTTGATACGTC and 5’-CATGGCGCCAAAAAAAAAAGCTTCAGCTGGGTACCGGATCCATTTCACTACT CTTGCCTCTGCTCTTATATAG. The PCR product was cut with ApaI and SfoI and ligated into pBiEx-hph to create a selectable mosquito shRNA vector pAeSH that allows for the insertion of any shRNA into the multiple cloning site present downstream of the U6 promoter. Targeted shRNAs to Aedes aegypti aeHuR were designed and the following oligos were inserted into the BamHI and HindIII sites of pAeSH: 5’-GATCCCAAAGTGCTAGCAGCCGTATTCAAGAGATACGGCTGCTAGCACTTG TTA and 5’ AGCTTAACAAAGTGCTAGCAGCCGTATCTCTTGAATACGGCTGCTAGCACTT TGG to create pAeSH-aeHuR1.
Oligos containing sequences derived from the 3’ UTRs of VEE, EEE, WEE, and SFV (Fig. S1B) were inserted into the EcoRI and PstI sites of pGEM4-A60 (Garneau et al., 2008). Transcription templates were generated via digestion with HindIII or NsiI, for the non-adenylated and adenylated species respectively. Internally radiolabeled, capped RNA substrates were transcribed by SP6 polymerase and purified as described (Wilusz and Shenk, 1988).
293T cells were either transfected with a Mission shRNA vector (Sigma Aldrich) specific to HuR (TRCN0000017277) or mock transfected (lacking specific shRNA but still containing transfection reagent) using FuGene 6. At 12 hours post transfection cells were infected with ts6SinV at an MOI of 5. Following a 30 min adsorption period, fresh media was added and the cells were placed at 28°C for 10 hours. Pre-warmed media was then added and the cells were transferred to 40°C to shut off viral transcription. At the times indicated post shut off, total RNA was harvested using Trizol and quantified by qRT-PCR (Garneau et al., 2008) using the primers listed in Fig. S1B. Note that since the analysis was done at 10 hpi, shRNA knockdown of HuR was allowed to proceed for 22hrs post transfection.
Determination of SinV RNA decay in Aag2 cells was performed as described above using stable aeHuR-knockdown cell lines selected following transfection of the pAeSH-aeHuR1 or empty pAeSH vector using FuGene6. Stable cell pools were selected using 300U of Hygromycin B.
From the half life regression curves for each independent in vivo RNA decay experiment, the average genomic and subgenomic half lives observed in the 293T and HuR-deficient cell lines were determined. These average half lives were subjected to statistical analysis via a two-tailed Student’s T-test.
Abundance of aeHuR and HuR mRNAs was examined using qRT-PCR and the primers listed in Fig. S1B. The relative abundances between control transfected cells and shRNA treated cells were compared at time point zero.
Cytoplasmic extracts were derived from C6/36 cells and HeLa S3 cells as previously described (Sokoloski et al., 2008a,b). 100,000 CPM (4-50 fmoles) of internally radiolabeled, polyadenylated RNA was incubated in the presence of mosquito cytoplasmic extract as previously described (Opyrchal et al., 2005), aliquots were removed at the desired time points and recovered RNAs were analyzed on 5% acrylamide gels containing 7M urea. RNAs were analyzed by phosphorimaging.
100,000 CPM (4-50 fmoles) of RNA substrate were incubated in cell-free RNA decay assays for 1 min. using either HeLa or C6/36 extracts in the presence of 0.5 mM EDTA to inhibit RNA decay (Sokoloski et al., 2008b). Following cross-linking of RNA-protein complexes by 254nm UV light, the mixture was treated with RNAse A and RNAseONE, proteins cross-linked to short radiolabeled RNA oligomers were resolved using 10% SDS-PAGE and visualized by phosphorimaging.
For RNA cross-linking / immunoprecipitation assays, samples were incubated, irradiated and treated with RNAse as described above. Samples were diluted with NET2 buffer (40mM Tris-HCl, pH 7.5 / 200mM NaCl / 0.05% (v/v) NP-40) and clarified by centrifugation. Samples received either control sera or target-specific antibody and were incubated for one hr at 4°C. Protein G sepharose or Pansorbin were then added, antibody-antigen complexes were washed ~5 times with NET2 buffer and immunoprecipitated proteins were analyzed by 10% SDS-PAGE.
293T or Aag2 cells were infected with wild type SinV at an MOI of 5. Cells were released using trypsin (293T) or mechanical scraping (Aag2), washed with PBS, resuspended in a 1% formaldehyde solution in PBS and incubated for 10 min. at room temperature. The reaction was quenched using 0.25M glycine. Cell pellets were washed with PBS and resuspended in RIPA Buffer (50 mM Tris-Cl, pH 7.5 / 1% (v/v) NP-40 / 0.5% (w/v) Sodium Deoxycholate / 0.05% (w/v) SDS / 1 mM EDTA / 150 mM NaCl). Cells were disrupted by sonication on ice and insoluble materials removed via centrifugation. Aliquots received either anti-HuR, anti-aeHuR or control (pre-immune) antibodies and incubated at 4°C for one hr. Antibody bound complexes were recovered using protein G or protein A sepharose after 6X washes with RIPA buffer containing 1M urea. Formaldehyde cross-links were reversed by heating at 70°C for 45 min. Isolated RNAs were analyzed by RT-PCR to assess total SinV RNA.
5’ Biotinylated RNA oligos consisting of either the SinV URE/CSE fragment (5’-Bio- UAAUCAUUUAUUAUUUUCUUUUAUUUUAUUCACAUAAUUUUGUUUUUAA) or a nonspecific sequence (5’-Bio-GAAUCGAGCUCGGUACCCGGGGAUCCUCUAGAGUCGACCUGCAG) were incubated with 2.5 mgs of cytoplasmic extract. Protein-RNA complexes were bound to Streptavidin resin, pelleted by centrifugation and washed five times with 20mM HEPES, pH 7.9 / 225mM KCl / 2.5mM MgCl2 / 0.1% (v/v) NP-40 / 1mM DTT. Proteins were eluted by boiling in HSCB buffer, concentrated using methanol and resolved via 10% SDS-PAGE. Gels were stained and excised bands were analyzed by MALDI-TOF/TOF.
The Aedes aegypti open reading frame corresponding to AaeL_AAEL008164 was PCR-amplified using the primers 5’-CATGGATCCATGACCAACAAAGTGCTAGCAGCC and 5’-CATGAATTCTTAATGATCGGCCATTTCGGCG. The gel-purified fragment was cut with BamHI and EcoRI and inserted into pGEX-2TZQ (Qian and Wilusz, 1994) to make pGEX-aeHuR. Recombinant protein was prepared from BL21 DE3 E. coli and the GST tag was removed by thrombin cleavage. Recombinant human HuR was a gift from N. Curthoys (Colorado State University).
Separation of cells into nuclear and cytoplasmic fractions was performed as described in Hel et al (1998). Following separation from nuclei, the cytoplasmic fraction was centrifuged at 16,000 × g for 10 min at 4°C prior to analysis by western blotting. Antibodies to PABPN1 (K-18) and HuR (3A2) were purchased from Santa Cruz Biotechnology; GAPDH antibodies were obtained from Millipore. A polyclonal aeHuR antisera was raised by Bioo Scientific in rabbits using the above recombinant aeHuR. Serum specificity was determined via western blot analysis of C6/36 cytoplasmic extract separated on 10% SDS-PAGE.
Aag2 and 293T cells were grown on glass coverslips, fixed in 4% paraformaldehyde, permeabilized in methanol and rehydrated in 70% ethanol. Coverslips were blocked in 6% bovine serum albumin fraction V (BSA) in PBS for at least one hour and washed in PBS. Primary antibody (diluted in 0.6% (w/v) BSA in PBS) was added for one hr, washed in PBS and secondary antibody (diluted as above) was applied for one hr. After washing, the coverslips were mounted using ProLong Gold antifade reagent with DAPI. Antibodies used were HuR (3A2) (Santa Cruz Biotechnology (SCBT)), Nucleophosmin NA24 (Thermo Scientific), anti-aeHuR polyclonal serum (this study), Cy2 donkey anti-mouse Ig (Jackson ImmunoResearch), Cy5 goat anti-rabbit Ig (GE Healthcare) and Texas red anti-mouse Ig (SCBT).
For FISH analyses, cells were grown, fixed, permeabilized and rehydrated as above. Coverslips were washed in PBS followed by a treatment of 40% formamide/2X SSC. The oligonucleotide (5’Alexa488 labeled 5’-TGACATTTCAAGGAGCCGCAGCATTT) was diluted in 40% formamide/2X SSC/0.2% (w/v) BSA and added to the coverslip for two hrs and washed in 1X SSC.
Internally radiolabeled unadenylated alphavirus URE/CSE fragments (~1.25 fmols) were incubated in the presence of recombinant aeHuR or HuR at the indicated concentrations in Gel Shift Buffer (15mM HEPES, pH 7.9, 100mM KCl, 2.25mM MgCl2, 5% (v/v) Glycerol). The complexes were allowed to form for five minutes at room temperature prior to the addition of 2.6 ug/ul Heparin Sulfate. The addition of heparin prior to the incubation with recombinant protein gave similar results. Following a 5 min. incubation on ice, protein-RNA complexes were resolved on a 5% native acrylamide gel and analyzed by phosphorimaging. Values obtained for bound versus free RNA were plotted and dissociation constants were calculated from the slope of the linear regression line fitted to the data.
We wish to thank members of the Wilusz laboratories for suggestions, Dr. Alan Schenkel for microscopy assistance and Dr. Imed Gallouzi for helpful discussions and reagents. These studies were supported by NIH grant AI063434 and an NIAID award through the Rocky Mountain Regional Center of Excellence (U54 AI-065357) to J.W.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.