|Home | About | Journals | Submit | Contact Us | Français|
Several viruses express factors to silence host gene expression via widespread mRNA degradation. This phenotype is the result of the coordinated activity of the viral endonuclease SOX and the cellular RNA degradation enzyme Xrn1 during lytic Kaposi's sarcoma-associated herpesvirus (KSHV) infection. While most cellular transcripts are highly downregulated, a subset of host mRNA escapes turnover via unknown mechanisms. One of the most prominent escapees is the interleukin 6 (IL-6) mRNA, which accumulates robustly during KSHV lytic infection and is not subjected to SOX-induced degradation. Here we reveal that the IL-6 mRNA contains a dominant, cis-acting ~100-nucleotide element within its 3′ untranslated region (UTR) that renders it directly refractory to cleavage by SOX. This element specifically interacts with a cellular protein complex both in SOX-transfected cells and in KSHV-infected B cells. Using a directed RNA pulldown approach, we identified two components of this complex to be the AU-rich element (ARE) binding proteins AUF1 and HuR. Depletion of these proteins significantly reduced the protective capacity of the IL-6 RNA element in SOX-expressing cells. These findings suggest that SOX activity may be directly counteracted by select RNA regulatory complexes and reveal a novel mechanism contributing to the robust expression of IL-6 during KSHV replication.
Kaposi's sarcoma-associated herpesvirus (KSHV) is a human gammaherpesvirus and the etiologic agent of several neoplasms afflicting immunocompromised individuals. Among these are its namesake disease Kaposi's sarcoma (KS), as well as primary effusion lymphoma (PEL) and multicentric Castleman's disease (MCD). KS, in particular, has been a major focus of research, as it is one of the most common and aggressive cancers in untreated AIDS patients (1). Unlike more traditional virus-induced cancers, where viral latency (or genome integration) is a transforming event, cells latently infected with KSHV are heavily reliant on externally provided growth factors and do not display the classical hallmarks of transformed cells. Thus, although KSHV latency is an essential component of disease, lytic KSHV infection is likely to contribute importantly to pathogenesis as well. Although lytically infected cells are destined to die, they produce a number of secreted viral and host paracrine factors that play key roles in the maintenance of the tumor microenvironment (1, 2).
Despite the reliance on host factors induced during the lytic cycle for KS development, most host genes are prominently downregulated by the KSHV SOX protein. SOX, encoded by ORF37, is a lytic viral protein expressed with delayed early kinetics that promotes widespread destruction of host mRNAs, thereby effectively blocking cellular gene expression (3). SOX functions in cells as an mRNA endonuclease and works coordinately with the host Xrn1 5′ to 3′ exonuclease to execute RNA degradation (4). A murine gammaherpesvirus (MHV68) bearing a single point mutation that selectively compromises the mRNA degradation activity in MHV68 SOX exhibits defects in trafficking, as well as during peak latency establishment in infected mice, and fails to induce splenomegaly (5). These results suggest that SOX-induced host shutoff contributes to viral fitness and pathogenesis in vivo.
Unlike the majority of cellular mRNAs whose levels decline during lytic KHSV infection, the IL-6 mRNA is strongly induced and remains elevated throughout the viral life cycle (6, 7). IL-6 is a proinflammatory cytokine that plays a prominent role in the pathogenesis of KSHV-induced diseases. It is a critical component of the tumor microenvironment of KS and is present in high circulating levels in patients with PEL and MCD (8–10). It serves both as a key growth factor for infected cells (9, 11) and, together with IL-10, assists in protection of PEL cells from immune eradication through the inhibition of dendritic cell maturation (12). IL-6 is also a proangiogenic factor and thus likely contributes to the vascularity of KS lesions via its induction of the vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) (13, 14). Several factors contribute to its accumulation during infection, including transcriptional induction by AP-1 and the viral RTA protein and transcript stabilization by the viral Kaposin B protein (15–18). However, even in the absence of other viral factors, the IL-6 mRNA appears refractory to SOX-induced turnover (6).
Here, we confirm that IL-6 directly escapes cleavage by SOX and reveal that its capacity to escape maps to an RNA element within its 3′ untranslated region (UTR). Remarkably, this IL-6 segment can confer protection from SOX when fused to a heterologous RNA, indicating it actively represses SOX targeting. We applied a novel ribonucleoprotein (RNP) purification strategy to show that the cellular proteins AUF1 and HuR complex with this IL-6 element and demonstrated that upon depletion of each protein, the protective effect of the element is reduced. Our findings represent the first described SOX-resistance element and suggest a novel mechanism whereby select RNAs may accumulate during lytic KSHV infection in the background of widespread RNA degradation.
The KSHV-positive B cell line bearing a doxycycline-inducible version of the major lytic transactivator RTA (TREx BCBL-1-RTA) (19) was maintained in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS; Invitrogen), 200 μM l-glutamine (Invitrogen), 100 U/ml penicillin (Invitrogen), 100 μg/ml streptomycin (Invitrogen), and 50 μg/ml hygromycin B (Omega Scientific). Lytic reactivation was induced by treatment with 20 ng/ml 2-O-tetradecanoylphorbol-13-acetate (TPA; Sigma), 1 μg/ml doxycycline (BD Biosciences), and 500 ng/ml ionomycin (Fisher Scientific) for the indicated period of time.
HEK293T cells (ATCC) were grown in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) with 10% FBS. DNA transfections were carried out in 12-well plates at 80% cell confluence using 0.75 μg DNA and 2.2 μl Polyjet (SignaGen Laboratories) per well (1 μg DNA and 3 μl Polyjet per 6-well plates). For small interfering RNA (siRNA) transfections, HEK293T cells were transfected twice over a 48-h period with 70 μM siRNAs (scramble [89775908; IDT] sense, CUUCCUCUCUUUCUCUCCCUUGUdGdA, and antisense, UCACAAGGGAGAGAAAGAGAGGAAGGA; XrnI, AGAUGAACUUACCGUAGAAAAUGTA; AUF1 [s6724; Ambion] sense, GGAAGGUGAUUGAUCCUAATT, and antisense, UUAGGAUCAAUCACCUUCCCA; and HuR [s4610; Ambion] sense, GCGUUUAUCCGGUUUGACATT, and antisense, UGUCAAACCGGAUAAACGCAA) using 5.6 μl Lipofectamine 2000 (Invitrogen) per well of a 12-well plate. Twenty-four hours following the second siRNA transfection, the cells were split again and transfected 1 day later with the appropriate DNA plasmids.
The iSLK and KHSV-infected iSLK.219 cells bearing doxycycline-inducible RTA (20) were maintained in DMEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen). KSHV lytic reactivation of the iSLK.219 cells was induced by the addition of 0.2 μg/ml doxycycline (BD Biosciences) and 110.09 μg/ml sodium butyrate for 48 h.
The full-length IL-6 cDNA in pCMV-SPORT6.1 was obtained from Invitrogen. Sequence numbering for IL-6 refers to Homo sapiens interleukin 6 (interferon, beta 2), mRNA (GenBank accession number BC015511.1). The GFP-IL-6 fusion constructs GFP-IL-6 5′UTR, GFP-IL-6 CDS, GFP-IL-6 3′UTR, GFP-IL-6 3′UTR 689–790, GFP-IL-6 3′UTR 689–890, GFP-IL-6 3′UTR 689–990, GFP-IL-6 3′UTR 689–1090, GFP-IL-6 3′UTR 791–890, GFP-IL-6 1–450 and GFP-IL-6 586–686 were created by overlap PCR and cloned into EcoRI/XhoI sites of pCDNA3.1(+). The IL-6 966–1038 construct was generated by annealing two oligonucleotides and ligating into BamHI/XbaI sites of pCDNA3.1 (+). IL-6 nucleotides (nt) 791 to 890 were subcloned into the BamHI/XhoI sites of pBSSK(−) for in vitro transcription. The Csy4 recognition motif was ligated into the KphI/BamHI sites of pcDNA3.1(+) IL-6 966–1038 and the pBSSK(−) IL-6 791–890 using the following hybridized oligonucleotides: 5′-CGTTCACTGCCGTATAGGCAGCTAAGAAAG-3′ and 5′-GATCCTTTCTTAGCTGCCTATACGGCAGTGAAC-3′.
Total RNA was harvested using Zymo RNA extraction columns or TRIzol (Invitrogen) following the manufacture's manual. cDNAs were synthesized from 1 μg of total RNA using AMV reverse transcriptase (RT; Promega), diluted 1:5, and used directly for quantitative PCR (qPCR) analysis. TaqMan IL-6 gene expression assay (Applied Biosystems) with a 6-carboxyfluorescein (FAM)-labeled probe and forward/reverse primers and TaqMan rRNA control reagent (Applied Biosystems) with a VIC-labeled probe and forward/reverse primers for human 18S rRNA as a loading control were used for qPCR analysis. The qPCRs were performed using TaqMan gene expression mix (Applied Biosystems) in the presence of 900 nM IL-6 primers, 250 nM IL-6 probe, 50 nM 18S rRNA primers, 200 nM 18S rRNA probe, or 100 nM GFP primer, 20 nM GFP probe (54). The level of IL-6 mRNA was calculated using a mathematical model of relative expression in qPCR (21) to quantify the relative level of IL-6 mRNA in comparison to the 18S rRNA.
For the half-life studies, HEK293T cells were transfected with the indicated plasmids in 6-well plates. The cultures were split after 6 h into 12-well plates. Twelve hours later, the cells were treated with 5 μg/ml actinomycin D (ActD) for the indicated times. The extracted RNAs were subjected to qPCR analysis, and green fluorescent protein (GFP) or IL-6 mRNA levels were normalized to the level of 18S rRNA to correct for errors in sample loading. The log of normalized data was then plotted versus the time of treatment of ActD. The reported data are the means of a minimum of three independent experiments.
RNA was resolved on 1.2% agarose-formaldehyde gels and transferred onto a Nytran N membrane (Whatman) prior to Northern blotting with 32P-labeled DNA probes against IL-6, GFP, or 18S using either RediPrime II (GE Healthcare) or Decaprime II (Ambion). Northern blots were analyzed using either a Typhoon 8600 phosphorimager (Molecular Dynamics) or the Pharao phosphorimager (Bio-Rad). The quantification was performed using either Image J (http://rsbweb.nih.gov/ij/) or ImageLab 4.0.1 (Bio-Rad).
TREx BCBL-1-RTA cytoplasmic extracts were prepared using a slight variation of the NE-PER nuclear and cytoplasmic extraction reagents (Thermo Scientific). Briefly, 1 × 107 cells were harvested by washing twice with 1× PBS and centrifuged at 500 × g for 3 min, and the pellet was lysed by vortexing in 1,000 μl CER I. After a 10-min incubation on ice, 55 μl ice-cold CER II was added and the samples were vortexed for 5 s, incubated on ice for 1 min, vortexed again, and then centrifuged for 10 min at 10,000 × g in a microcentrifuge. Supernatant (cytoplasmic extract) was collected for electrophoretic mobility shift assays (EMSAs). Cytoplasmic extract from TREx BCBL-1-RTA cells (12 to 16 μg) was preincubated for 10 min at room temperature in binding buffer (0.5% Nonidet P-40, 1 mM dithiothreitol, 2 μg of yeast tRNA, 40 units of RNasin, 10% glycerol, and 10 to 40 fmol of 32P-labeled RNA). For competition assays, a 30-fold (0.327 pmol/μl), 100-fold (1.09 pmol/μl), or 200-fold (2.18 pmol/μl) excess of competitor RNA was added with the labeled RNA (10.9 fmol/μl). The reaction mixture was incubated at room temperature for 20 min, 1 μl of RNase T1 was added, and the samples were incubated for 10 min at room temperature to ensure complete digestion. The samples were separated on 5% native polyacrylamide gels, and the gels were dried and visualized using a Typhoon 8600 phosphorimager (Molecular Dynamics).
Csy4 H29A/S50C was expressed and purified using the same protocol as wild-type Csy4 (generously provided by R. Haurwitz, H.Y. Lee, and J. Doudna) (22, 55). Briefly, Csy4 H29A/S50C was expressed as a 6×His-maltose binding protein (6×His-MBP) fusion in BL21(DE3) cells. Cells were lysed, clarified, and bound in batch to nickel agarose resin. The tagged protein was eluted with 300 mM imidazole-containing buffer and dialyzed overnight with tobacco etch virus (TEV) protease. The His-MBP tag was removed from the sample by a second nickel resin step, and the Csy4 protein was further purified by size exclusion chromatography (Superdex 75). Csy4 H29A/S50C was incubated overnight with a maleimide-PEG2-biotin reagent (Pierce) that covalently attached to solvent-accessible cysteine residues. Excess biotin was removed by overnight dialysis. Samples were aliquoted and flash frozen. Plasmids expressing the Csy4 RNA binding motif fused to segments of IL-6 were in vitro transcribed using the T7 Maxiscript kit (Ambion). Transcribed RNA (20 μg) was mixed with purified recombinant Csy4 protein (200 pmol) and magnetic beads for 2 h in lysis buffer [10 mM HEPES (pH 8.0), 3 mM MgCl2, 5% glycerol, 1 mM dithiothreitol (DTT), 150 mM NaCl, 0.1% octyl β-d-glucopyranoside, 10 mM imidazole, 1× protease inhibitor]. Lysate from TREx BCBL-1-RTA cells (1 mg) was then added to the beads for 2 h, whereupon the beads were washed 7 times with lysis buffer containing 150 to 300 mM NaCl. RNA and its associated cellular proteins were released from the Csy4-bound beads by the addition of 500 mM imidazole for 2 h to activate the cleavage activity of Csy4 and then resolved by 10% SDS-PAGE.
Total protein lysates were prepared in radioimmunoprecipitation assay (RIPA) buffer [50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% (vol/vol) Nonidet P-40, 0.5% (wt/vol) sodium deoxycholate, 0.1% (wt/vol) SDS]. Where indicated, the NE-PER kit (Thermo Scientific) or the REAP protocol (23) were used interchangeably for cellular fractionation prior to RNA extraction. Western blots were performed with mouse monoclonal anti-GFP (1:2,000; BD Biosciences), monoclonal anti-HuR (1:200; Santa Cruz), rabbit polyclonal anti-AUF1 (1:3,000; Millipore), rabbit polyclonal anti-TTP (Abcam), mouse monoclonal anti-KSRP (kindly provided by Ching-Yi Chen), rabbit polyclonal anti-SOX J5803 (1:5,000), mouse monoclonal hnRNPC1/C2 (1:2,000; Abcam), or mouse monoclonal anti-Hsp90 (1:3,000; Stressgen Bioreagents).
Western blot quantifications were performed using the ChemiDoc system (Bio-Rad) and ImageLab 4.0.1 analysis software (Bio-Rad).
Although the majority of cellular mRNAs are depleted during lytic KSHV infection, previous microarray analyses indicated that a small population of messages escapes this fate (6, 7). The most prominent of these is IL-6, which is markedly induced upon lytic reactivation of KSHV-infected microvascular endothelial (TIME) cells, human foreskin fibroblasts (HFF), and HEK293T cells (6). We confirmed these observations by evaluating IL-6 mRNA levels over time by qPCR in a KSHV-positive B cell line harboring a doxycycline-inducible version of the viral major lytic transactivator protein RTA (TREx BCBL-1-RTA). Indeed, IL-6 levels steadily increased over the 36-h lytic reactivation time course, reaching a peak of ~7-fold induction over latently infected cells (Fig. 1A).
The robust accumulation of IL-6 during infection suggested it may be refractory to cleavage by the KSHV host shutoff factor SOX. To evaluate the ability of SOX to target IL-6 in the absence of other viral factors, we transfected HEK293T cells with plasmids expressing IL-6 and GFP (as an internal positive control for SOX activity) in the presence or absence of SOX. Northern blot analysis confirmed that although GFP mRNA was readily degraded in SOX-expressing cells, IL-6 mRNA was not depleted (Fig. 1B). To confirm that IL-6 is refractory to SOX-induced degradation, we measured its decay following transcription from actinomycin D-treated HEK293T cells. Again, we cotransfected a GFP plasmid as an internal control for SOX activity and measured mRNA levels by qPCR over an 8-hour time course. The half-life (t1/2) of IL-6 mRNA was not significantly altered in the presence of SOX, whereas the t1/2 of the GFP reporter mRNA was reduced from >8 h to 2.5 h in SOX-expressing cells (Fig. 1C). Thus, the IL-6 mRNA escapes SOX-mediated degradation.
The failure of SOX to target the IL-6 mRNA could be due to the absence of an appropriate SOX-targeting element or to the presence of specific cis-acting protective feature(s). Although the precise sequences that mediate SOX targeting remain unknown, it has been demonstrated that SOX endonucleolytically cleaves its targets in a directed manner (4). However, to date there have been no examples of RNA elements that can confer protection from SOX cleavage. We reasoned that such an element, when fused to an mRNA that is normally susceptible to SOX, should render the message refractory to SOX-induced degradation.
To determine whether the IL-6 transcript contains a protective element that could function in a dominant manner to prevent SOX-induced degradation, we fused its 5′ UTR, coding sequences (CDS), or 3′ UTR to GFP, a transcript whose targeting by SOX has been extensively characterized (Fig. 2A). The abundance of each chimeric mRNA in the presence or absence of SOX was then monitored in transfected HEK293T cells by Northern blotting with a probe against GFP. Indeed, while GFP fused to the IL-6 5′ UTR or CDS was readily degraded by SOX, targeting of GFP fused to the IL-6 3′ UTR was markedly reduced (Fig. 2B). These results suggest that the 3′ UTR of IL-6 contains a cis-acting element that dominantly protects the mRNA in SOX-expressing cells.
SOX has been shown to endonucleolytically cleave GFP at nt 147 (relative to the start of the coding sequence). After this initial cut, the 3′ cleavage fragment is rapidly degraded by the cellular exoribonuclease Xrn1 (4). The primary cleavage can be monitored by depletion of Xrn1, as this leads to stabilization of the SOX-cleaved RNA fragment (4). Given the above results, we reasoned that the IL-6 3′ UTR should inhibit the initial SOX cleavage event. In agreement with this prediction, no degradation intermediate of the GFP-IL-6 3′ UTR fusion was observed upon siRNA-mediated depletion of Xrn1 in cells expressing SOX (Fig. 2C). In contrast, the expected SOX-cleaved intermediate of the GFP-IL-6 5′ UTR fusion was detected in cells depleted of Xrn1 (Fig. 2C, right panel). These data confirm that the 3′ UTR of IL-6 contains a cis-acting element that confers protection from SOX-induced cleavage.
To identify the region in the 3′ UTR of IL-6 responsible for protecting GFP from cleavage by SOX, we deleted portions of the IL-6 3′ UTR in 100-nt increments, leaving nucleotides 689 to 790, 689 to 890, 689 to 990, or 689 to 1090 (numbering is based on human IL-6 mRNA, GenBank accession number BC015511.1) (Fig. 3A). As controls, we fused GFP to IL-6 sequences residing upstream of the 3′ UTR with lengths corresponding to the largest and smallest deletion constructs (IL-6 nt 1–450 and IL-6 nt 586–686, respectively). The abundance of each mRNA relative to 18S loading controls was then measured in the presence and absence of SOX (Fig. 3B).
GFP constructs containing IL-6 nt 689 to 890, 689 to 990, or 689 to 1090 were not depleted by SOX, while those containing IL-6 nt 689 to 790, 1 to 450, and 586 to 686 were readily degraded (Fig. 3B). Given that the IL-6 segment containing nt 689 to 890 conferred protection but the segment containing nt 689 to 790 did not, we hypothesized that the primary protective sequences may reside within IL-6 nt 791 to 890 (dashed segment in Fig. 3A, sequence is three-dimensional [3-D]). Indeed, fusion of this minimal 100-nt IL-6 segment to GFP is sufficient to increase the mRNA abundance 2- to 3-fold over the control GFP-IL-6 5′ UTR fusion in the presence of SOX (Fig. 3C), although full protection requires additional sequence context. Thus, while the 100-nt region is sufficient to confer partial protection from SOX-induced cleavage, we predict that the complete element is multipartite. With this in mind, we designated this 100-nt region as one component of the SOX-resistant element (SRE1). Interestingly, the IL-6 SRE1 contains 4 AU-rich elements and has been shown to regulate the normally rapid turnover of the IL-6 mRNA (24–26). Surprisingly, however, our data suggest that in SOX-expressing cells this element instead contributes toward protection from the host shutoff phenotype, even when fused to a heterologous transcript.
RNA regulatory elements often function through the recruitment of specific RNA binding proteins that influence the fate of the transcript. To determine if the protective region of the IL-6 mRNA is bound by proteins present in infected B cells, we performed RNA electrophoretic mobility shift assays (EMSAs) using extracts from lytically reactivated TREx BCBL-1-RTA cells (Fig. 4A). RNase T1 was added to digest any unbound single-stranded regions of probe to completion to ensure that the protein complex can exit the well and run into the gel. Complex formation was observed on SRE1-containing probes encompassing either the IL-6 full-length 3′ UTR (nt 689 to 1127) or the 200-nt segment sufficient to confer full protection (nt 689 to 890). In contrast, no robust mobility shifts occurred with a panel of control IL-6 probes spanning regions within the 5′ UTR, coding sequence, or 3′ UTR upstream of the SRE1. Binding specificity was further confirmed using competition assays, which showed that the protein complex was effectively competed off the IL-6 3′ UTR by unlabeled IL-6 mRNA (specific competitor) but not by unlabeled nonspecific RNA derived from pCDNA3.1 (Fig. 4B).
AU-rich elements (AREs) such as those found in the IL-6 protective segment recruit ARE binding proteins that control (often by accelerating) mRNA turnover. Interestingly, the region we identified as a protective element was previously shown to be directly targeted by several ARE binding proteins, including AUF1 (24), TTP (27), and KSRP (25). Furthermore, IL-6 mRNA abundance has recently been shown to be positively regulated by the HuR protein, albeit indirectly through suppression of the mRNA encoding the ARE binding protein TTP (28). To determine whether any of these ARE binding proteins are associated directly with the IL-6 SRE1, we performed RNA pulldown assays using a novel ribonucleoprotein (RNP) purification system recently developed by the Doudna lab (Fig. 5A) (55). This robust purification tool is based on the discovery that Csy4, a component of the bacterial CRISPR antiviral pathway, binds to a short (28-nt) hairpin within the unprocessed CRISPR RNA (crRNA hp) with 50 pM equilibrium dissociation constant (29). We used a mutant version of Csy4 (H29A/S50C) lacking the general base in the active site, which is normally unable to cleave but can be activated in the presence of high levels of imidazole to cleave at a precise location just after nucleotide 20, eluting the RNA with only the last 8 nucleotides of the tag. Thus, beads coupled to recombinant mutant Csy4 can be used to purify any RNA segment fused to the crRNA hp.
The crRNA hp was fused to the IL-6 SRE1 or a downstream segment of IL-6 that does not have canonical AREs but does possess a stretch of an AU-rich sequence (nt 966 to 1038). After in vitro transcription, the RNAs were incubated with Csy4-bound beads and then subsequently with lysate from reactivated TREx BCBL-1-RTA cells. After extensive washing, the RNA-protein complexes were released by the addition of imidazole and resolved by SDS-PAGE. Western blotting for several ARE binding proteins showed that the SRE1 bound HuR and AUF1 but not KSRP or TTP from the infected B cell lysate (Fig. 5B). A small amount of AUF1 binding to the downstream IL-6 segment was also detected, presumably because it contains at least one AU-enriched region. Both HuR and AUF1 are shuttling proteins but remain nuclear at steady state (30–33), while SOX-mediated mRNA degradation occurs in the cytoplasm (34). Thus, for either protein to have a role in protection of the IL-6 mRNA from SOX-mediated cleavage they would have to be present in the cytoplasm of KSHV-infected cells. Due to the very limited volume of cytoplasm in the TREx BCBL-1-RTA cells, we instead used SLK cells bearing an inducible version of RTA that were either uninfected (iSLK) or stably infected with KSHV (iSLK.219) (20). We isolated nuclear and cytoplasmic fractions from uninfected, latently infected, or lytically reactivated cells and monitored the abundance of HuR and AUF1 in each fraction by Western blotting. Western blots for the nuclear hnRNP C1/C2 protein and cytoplasmic HSP90 protein confirmed the integrity of each subcellular fraction. A representative experiment is shown in the left panel of Fig. 5C, and a graph showing quantitation across 5 independent replicates is shown in the right panel. In each experiment, we were able to detect both AUF1 and HuR in the cytoplasm, though, as expected, most of the population remained nuclear. There was no change in the total levels of each protein across the samples, suggesting that neither AUF1 nor HuR are induced upon infection. It should be noted that we sometimes observed a modest increase in the levels of cytoplasmic AUF1 and to a smaller extent HuR in response to KSHV latent infection and a further increase upon lytic reactivation (Fig. 5C, right panel). However, our data suggest that while a subpopulation of both proteins is present in the cytoplasm, they are not dramatically relocalized to that site during KSHV infection.
If the presence of HuR and AUF1 on the IL-6 SRE1 contribute to its protection against SOX-induced degradation, then the protective capacity of the IL-6 3′ UTR should be reduced in the absence of these proteins. We therefore used specific siRNAs to deplete ≥95% of each of these proteins individually from HEK293T cells (Fig. 6A) and then evaluated susceptibility of mRNAs containing or lacking the IL-6 3′ UTR to SOX (Fig. 6B). In cells treated with the control scramble siRNA, the GFP mRNA fused to the IL-6 3′ UTR was protected against SOX-induced decay. However, in cells depleted of either AUF1 or HuR, the IL-6 3′ UTR no longer conferred full protection from SOX. We hypothesize that HuR and AUF1 may function in the same pathway to confer protection rather than by independent mechanisms, as we did not consistently observe a further decrease in protection upon codepletion of both proteins (data not shown). GFP fused to the 5′ UTR of IL-6 was targeted by SOX with similar efficiency regardless of the siRNA transfection, indicating that depletion of HuR and AUF1 does not nonspecifically enhance SOX activity. Collectively, our data suggest that when bound to the SRE1, HuR and AUF1 help render RNAs inaccessible to SOX endonuclease.
Although cellular gene expression is significantly dampened by widespread mRNA degradation during lytic gammaherpesvirus infection, select transcripts escape this fate. We previously showed that there are multiple mechanisms mediating escape of mRNAs from SOX-induced depletion, including absence of a SOX targeting element and transcriptional upregulation (35). Here, we reveal that the IL-6 mRNA is protected by a unique mechanism: it contains a sequence within its 3′ UTR that inhibits cleavage by SOX. This sequence contains a SOX-resistant element (SRE) that helps confer protection even when fused to a heterologous transcript and thus prevents SOX targeting via a novel, dominant protective mechanism. One component of the IL-6 SRE, termed SRE1, maps to a regulatory region of the transcript that contains several AU-rich elements (AREs) previously shown to recruit host factors primarily involved in mRNA destabilization (24–26). Indeed, we confirmed that this element complexes with host proteins, including the ARE binding proteins AUF1 and HuR. However, our data suggest that in SOX-expressing cells, these interactions contribute to protection of the IL-6 mRNA, rather than mediating its degradation (Fig. 7). Although we do not yet know the mechanism(s) by which protection is conferred, it could, for example, be through recruitment of a SOX inhibitor, steric hindrance of SOX binding or cleavage, or shunting of the mRNA to a location or decay pathway inaccessible to SOX. Collectively, our findings demonstrate that RNA-protein interactions can directly influence the susceptibility of cellular mRNAs to cleavage by the gammaherpesviral SOX protein, even when they occur distal to the actual cleavage site.
AREs are among the most common determinants of mRNA stability in mammalian cells and play a prominent role in regulating turnover of the IL-6 message. AREs influence transcript stability through the recruitment of specific ARE binding proteins, often leading to enhanced mRNA deadenylation and decay. These elements are enriched in 3′ UTRs of many unstable mRNAs encoding tightly regulated proteins such as transcription factors, proto-oncogenes, and cytokines. They can be grouped into at least three classes depending on their sequence and rates of decay (36); IL-6 contains a series of nonclustered class I-like AREs. Presumably, the type and magnitude of the effect exerted by an ARE are determined by the cohort of ARE binding proteins recruited by that element. Several ARE binding proteins mediate mRNA destabilization by recruiting the degradation machinery, although some enhance mRNA stability (26).
The IL-6 AREs have been shown to bind AUF1, TTP, and KSRP, proteins that generally promote mRNA degradation (24, 25, 27), but IL-6 mRNA stability is also been shown to be indirectly positively regulated by HuR (28). Using a targeted RNA pulldown approach, we found that two of these proteins, AUF1 (also known as hnRNP D) and HuR, associate with the SRE1 core. Although AUF1 is predominantly nuclear, it is a shuttling protein and binds to target mRNAs in the cytoplasm where it enhances their decay (33, 36, 37). The mechanisms underlying its translocation remain incompletely understood and are complicated by the fact that AUF1 has 4 alternative splicing isoforms that may hetero-oligomerize (38–41). HuR is also a shuttling protein that primarily resides in the nucleus, though its cytoplasmic translocation is instead associated with stabilization and translational regulation of its mRNA targets (42). Interestingly, both AUF1 and HuR have been shown to interface with RNA viruses. AUF1 is relocalized to the cytoplasm upon poliovirus infection, whereupon it is cleaved by the viral 3CD proteinase (43). HuR is prominently relocalized during Sindbis virus infection and binds and stabilizes the viral mRNAs, and it has also been shown to bind RNAs of other alphaviruses and hepatitis C virus in a manner important for viral replication (44–47). In KSHV-infected cells, both proteins retain their prominent nuclear localization but are detectable in the cytoplasm as well. Thus, KSHV does not cause a robust translocation of either protein, although it may result in a modest increase in the cytoplasmic levels of AUF1. These observations argue against cytoplasmic relocalization of these proteins as a primary contributor to the stabilization phenotype during KSHV infection. Nonetheless, the fact that depletion of either protein significantly reduced the protective effect of the IL-6 3′ UTR in SOX-expressing cells argues that their binding to this region of the RNA contributes importantly to the escape mechanism. Presumably, the cytoplasmic concentrations of AUF1 and HuR are sufficient to mediate these effects, although it is also formally possible that they assemble onto the transcript while in the nucleus. We did not observe an additive effect on IL-6-mediated stabilization upon codepletion of AUF1 and HuR (S. Hutin and B. Glaunsinger, unpublished observations), suggesting that both proteins are operating in the same pathway.
Although the IL-6 escape element contains AREs, it cannot be generalized that ARE-bearing mRNAs are protected from SOX-induced degradation. In fact, a previous analysis of RNA features associated with escape from SOX found a negative correlation between the presence of an ARE and protection from SOX-induced decay (35). Furthermore, ARE-bearing mRNAs such as the granulocyte-macrophage colony-stimulating factor (GM-CSF) transcript are degraded by SOX, arguing against a simple failure of SOX to target unstable transcripts (6). We instead hypothesize that proteins associated with the IL-6 ARE, including AUF1 and HuR, coordinate with other factors or sequences within the escape element to confer protection from SOX. That said, in the context of lytic KSHV infection, there is an enrichment of ARE-bearing mRNAs in the escapee population (7). However, in this setting, the viral Kaposin B and vGPCR proteins are expressed (both of which have been shown to activate MK2 signaling), leading to stabilization of ARE mRNAs (16, 48). It is thus likely that multiple viral factors contribute to the overall manipulation of host gene expression during KSHV infection.
While the majority of protein complex binding to the IL-6 mRNA occurs within SRE1, additional flanking sequences are required to confer maximal protection from SOX. This is highlighted by the observation that fusion of SRE1 alone to the GFP reporter increases its abundance in SOX-expressing cells by approximately 2-fold, whereas inclusion of an additional 98 nt bordering the 5′ end of the SRE1 core leads to an ~4-fold increase. It is possible that neighboring sequences assist with RNP architecture or assembly in vivo or that they contribute to the protection via an independent mechanism. For example, the IL-6 mRNA is regulated by a series of miRNA binding sites, including miR-26 and let-7 (49–52). Most of the predicted miRNA target sides are surrounding SRE1, while miR-26 partially overlaps at the 3′ end. It has also been shown that the KSHV ORF57 protein contributes to IL-6 mRNA stability by interacting with a binding site for miR-608 (53), although this site is within the coding region of IL-6 and is thus distinct from the mechanism of protection against SOX. Given the partial overlap between miR-26 or potentially other miRNAs and the SRE1 core, an important future direction will be to determine whether they similarly influence targeting by SOX. In addition, we are currently working toward identifying additional proteins associated with the SRE1.
Collectively, our findings demonstrate that specific mRNA regulatory elements, together with their associated ribonucleoproteins, can render transcripts inaccessible to SOX. Presumably, such additional elements exist, and their discovery and characterization are anticipated to reveal new insight into the mechanisms by which cis-acting RNA elements function to control transcript stability.
We thank Rachel Elizabeth Haurwitz, Ho Young Lee, and Jennifer Doudna at the University of California, Berkeley, for their generous provision of biotinylated Csy4 and development and assistance with the Csy4-based purification system and the Glaunsinger lab for many fruitful discussions. We thank Jae Jung (UCLA) for the TREx-Rta BCBL-1 cell line and Ching-Yi Chen (University of Alabama) for providing anti-KSRP antibodies.
This research was funded by NIH grants F32 CA132308 to Y.L. and R01 CA160556, CA136367, and K01 CA117982 to B.A.G.
Published ahead of print 13 February 2013