Rotavirus NSP1 Associates with Components of the Cullin RING Ligase Family of E3 Ubiquitin Ligases

ABSTRACT

The rotavirus nonstructural protein NSP1 acts as an antagonist of the host antiviral response by inducing degradation of key proteins required to activate interferon (IFN) production. Protein degradation induced by NSP1 is dependent on the proteasome, and the presence of a RING domain near the N terminus has led to the hypothesis that NSP1 is an E3 ubiquitin ligase. To examine this hypothesis, pulldown assays were performed, followed by mass spectrometry to identify components of the host ubiquitination machinery that associate with NSP1. Multiple components of cullin RING ligases (CRLs), which are essential multisubunit ubiquitination complexes, were identified in association with NSP1. The mass spectrometry was validated in both transfected and infected cells to show that the NSP1 proteins from different strains of rotavirus associated with key components of CRL complexes, most notably the cullin scaffolding proteins Cul3 and Cul1. In vitro binding assays using purified proteins confirmed that NSP1 specifically interacted with Cul3 and that the N-terminal region of Cul3 was responsible for binding to NSP1. To test if NSP1 used CRL3 to induce degradation of the target protein IRF3 or β-TrCP, Cul3 levels were knocked down using a small interfering RNA (siRNA) approach. Unexpectedly, loss of Cul3 did not rescue IRF3 or β-TrCP from degradation in infected cells. The results indicate that, rather than actively using CRL complexes to induce degradation of target proteins required for IFN production, NSP1 may use cullin-containing complexes to prevent another cellular activity.

IMPORTANCE The ubiquitin-proteasome pathway plays an important regulatory role in numerous cellular functions, and many viruses have evolved mechanisms to exploit or manipulate this pathway to enhance replication and spread. Rotavirus, a major cause of severe gastroenteritis in young children that causes approximately 420,000 deaths worldwide each year, utilizes the ubiquitin-proteasome system to subvert the host innate immune response by inducing the degradation of key components required for the production of interferon (IFN). Here, we show that NSP1 proteins from different rotavirus strains associate with the scaffolding proteins Cul1 and Cul3 of CRL ubiquitin ligase complexes. Nonetheless, knockdown of Cul1 and Cul3 suggests that NSP1 induces the degradation of some target proteins independently of its association with CRL complexes, stressing a need to further investigate the mechanistic details of how NSP1 subverts the host IFN response.

INTRODUCTION

Protein ubiquitination is among the most widely used posttranslational modifications and regulates many aspects of cell biology, including cell signaling, DNA damage responses, and protein degradation (1). Conjugation of ubiquitin to a target protein occurs by the sequential activities of three types of enzymes: a ubiquitin-activating enzyme (E1), a ubiquitin-conjugating enzyme (E2), and a ubiquitin-ligating enzyme (E3). E3 ubiquitin ligases are classified on the basis of the presence of either a HECT domain or a RING-like motif. E3s with a HECT domain accept the ubiquitin moiety from an E2 and then mediate the transfer of ubiquitin directly to a substrate. E3 ubiquitin ligases containing a RING domain are more common, and these E3s serve as a bridge between an E2 enzyme carrying ubiquitin and the substrate protein being targeted for degradation (1–3). The RING domain consists of eight conserved residues, generally Cys and His, which coordinate two Zn²⁺ ions in a cross-brace arrangement that creates a platform for E2 binding. The classical RING finger domain consists of a signature Cys₃-His-Cys₄ motif, although several different variants of the RING domain have been identified, including Cys₃-His₂-Cys₃ and Cys₄-His-Cys₃ motifs (1).

E3 ubiquitin ligases are the main determinant of substrate specificity and may exist as single proteins or as multisubunit complexes. The largest group of E3 ubiquitin ligases is that of the cullin RING ligases (CRLs). CRLs are multisubunit complexes nucleated by a cullin (Cul) protein that serves as a scaffold onto which an E3 ubiquitin ligase (Rbx1), an adaptor subunit, and a substrate receptor are assembled (4, 5). A variable number of linker proteins may increase the complexity of these complexes. Together, these protein complexes promote the transfer of ubiquitin from a bound E2 protein to the target substrate. CRLs can be subdivided according to the interaction between their Cul scaffold subunits and their adaptors. For example, CRL1 complexes (also known as Skp–Cul–F-box, or SCF, complexes) have an adaptor protein called Skp1 that forms a link between the Cul1 scaffolding protein and the substrate receptor (an F-box protein). On the other hand, CRL3 complexes, which have a Cul3 scaffolding protein, contain a single BTB (broad complex, tram track, bric-a-brac fold) protein that functions as both the adaptor protein and the substrate receptor (4).

Because ubiquitination is important in so many cellular functions, it is not surprising that many viruses have evolved ways to hijack the host ubiquitination machinery to replicate and spread (6). In fact, there are many cases in which viruses express proteins that directly interact with components of one or more host CRL complexes in order to usurp CRLs to benefit the virus (7–9). Group A rotaviruses are nonenveloped, double-stranded RNA (dsRNA) viruses that belong to the family Reoviridae and are a leading cause of severe, life-threatening gastroenteritis in young children (10). Relatively little is known about how rotavirus interacts with the ubiquitin-proteasome system in infected cells, but a functional proteasome appears to be required for the virus to establish a productive replication cycle (11).

Rotavirus replicates in the cytoplasm of an infected cell and, like most viruses, has several means by which it can avoid host defenses. The nonstructural protein NSP1 is known to serve as an antagonist of the innate immune response, primarily by inhibiting the production of type I interferon (IFN). When host pattern recognition receptors (PRRs) interact with viral nucleic acids or proteins, a signaling cascade is initiated that causes the activation of transcription factors responsible for the production of IFN, including IFN regulatory factor 3 (IRF3), IRF7, and nuclear factor κB (NF-κB) (12–14). IRF3 and IRF7 are activated by phosphorylation and form homodimers or heterodimers that translocate to the nucleus. NF-κB is activated when the inhibitor of κB (IκB) is phosphorylated, which targets IκB for ubiquitination by an SCF complex containing the β-transducin repeat-containing protein (β-TrCP). Ubiquitination of IκB leads to its degradation by the proteasome, revealing a nuclear localization signal present on NF-κB that allows translocation to the nucleus (15). NF-κB and IRF transcription factors bind to specific gene-regulatory sequences to induce the production of IFN and other antiviral proteins.

Rotavirus NSP1 has been shown to inhibit the production of IFN by inducing the degradation of host factors required for IFN induction, but the host protein targeted for degradation varies depending on the strain of the virus. IRF3 was first identified as a target of NSP1-mediated degradation and was also shown to directly interact at the C-terminal end of NSP1 (16–19). IRF5, IRF7, and IRF9 were subsequently identified as additional targets of degradation by NSP1 due to the structural similarity of their IRF association domain (or dimerization domain) to that of IRF3 (20, 21). β-TrCP has also been shown to be a target of some NSP1 proteins, and the loss of β-TrCP prevents the nuclear translocation of NF-κB, thereby preventing IFN induction (22, 23). As noted earlier, the NSP1 proteins from different rotavirus strains do not always target each of these host proteins for degradation; instead, target recognition appears to be variable. Studies comparing the gene and protein sequences of NSP1 proteins from different rotavirus isolates have shown that NSP1 is highly variable, in particular at the C terminus of the protein (23, 24). It has been proposed that the C-terminal variation is responsible for the differences in substrate recognition (22, 23, 25).

In spite of the variability observed among NSP1 proteins, a conserved region at the N terminus of NSP1 is predicted to form a RING domain. The Cys₄-His-Cys₃ motif, conserved in all group A rotavirus NSP1 proteins, is a variant of the classical RING and shares homology with other RING-containing viral proteins that function as ubiquitin ligases (26–28). The degradation of IRF and β-TrCP targets normally induced by NSP1 is blocked in the presence of proteasome inhibitors, suggesting the ubiquitin-proteasome system is involved in NSP1-mediated host protein degradation (17, 20). For these reasons, NSP1 has been hypothesized to induce degradation of host proteins by acting as an E3 ubiquitin ligase. Unfortunately, there is a lack of direct evidence to support this hypothesis.

In the present study, NSP1 proteins from a variety of different rotavirus isolates were shown to associate with components of CRL complexes. The association of NSP1 with individual components of CRL complexes, including the Cul3, Cul1, Rbx1, and BTB proteins, was examined, and Cul3 was identified as a direct interacting partner of NSP1. The ability of NSP1 to induce degradation of IRF3 and β-TrCP target proteins when Cul3 or Cul1 was knocked down by small interfering RNA (siRNA) was also examined, but unexpectedly, both NSP1 targets were still degraded in the absence of Cul3 or Cul1. These results suggest that NSP1 binding to CRL complexes in the host cell may be used to induce degradation of other target proteins or that NSP1 uses CRLs for an as yet undefined function. This study highlights the need for further investigation to define the mechanism of NSP1-induced host protein degradation so that there can be a better understanding of the role of NSP1 in inhibiting the IFN response.

MATERIALS AND METHODS

Cell culture.

Human 293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Corning) supplemented with 5% fetal bovine serum (FBS) (Seradigm) and 1% nonessential amino acids (NEAA) (Thermo Fisher Scientific). African green monkey kidney MA104 cells were maintained in medium 199 (M199) (Corning) supplemented with 5% FBS. All the cells were cultured at 37°C with 5% CO₂.

Viruses.

Rotavirus strains SA11-4F, SA11-5S, OSU, WI61, UK, and DS-1 were propagated in MA104 cells as previously described (27). Virus titers were determined by plaque assay (27). All viruses were activated by incubation with 5 μg/ml trypsin for 60 min prior to infection.

Antibodies.

Polyclonal antisera to NSP1 from the simian SA11-4F and SA11-5S strains of rotavirus were produced by Pacific Immunology Corporation (Ramona, CA). Peptides corresponding to amino acids 480 to 496 of SA11-4F NSP1 (C-TEEFELLISNSEDDNE) and amino acids 257 to 273 of SA11-5S NSP1 (C-RDELELYSDLKNDKNKL) were conjugated to the carrier protein keyhole limpet hemocyanin. Each peptide was used to immunize individual New Zealand White rabbits. NSP1-specific antisera were collected and affinity purified using the immunizing peptide, and each tested negative for cross-reactivity with other rotavirus proteins. Additional affinity-purified NSP1 antibodies were specific for the following strains of rotavirus: UK (25), OSU (23), WI61 (provided by John Patton), and DS-1 (provided by John Patton). All NSP1 antibodies were used at a 1:1,000 dilution.

Polyclonal VP6 antiserum was produced by Thermo Fisher Scientific (Rockford, IL) in guinea pigs by immunization with VP6, which was purified as previously described (29). The VP6 antibody was used at a 1:1,500 dilution. Rabbit polyclonal antibodies to Halo (Promega; G9281), IRF3 (Cell Signaling; cs-11904), β-TrCP (Cell Signaling; cs-4394), Cul1 (Novus; NB100-91724), Cul3 (Novus; NB100-58788), MBP (Santa Cruz Biotechnology; sc-808), and PCNA (Santa Cruz Biotechnology; sc-7907) were used at a 1:1,000 dilution. Monoclonal antibodies to FLAG M2 (Sigma-Aldrich; F1804) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Santa Cruz Biotechnology; sc-32233) were used at 1:1,000 and 1:750 dilutions, respectively.

Cloning.

To generate plasmids expressing NSP1 from the rotavirus strains SA11-4F and SA11-5S tagged at the N terminus with HaloTag (Promega), the open reading frame (ORF) of each NSP1 was amplified using primers that added terminal SacII and SbfI restriction sites. The PCR products and the mammalian expression plasmid pHTN HaloTag CMV-neo (Promega) were digested with SacII and SbfI restriction enzymes and ligated together. To generate plasmids expressing NSP1 from the rotavirus strains OSU, WI61, UK, and DS-1 tagged at the N terminus with HaloTag, the ORF of each NSP1 was amplified using primers that added a 15-bp overlap with the plasmid for Gibson assembly cloning. The pHTN plasmid was digested with the restriction enzyme EcoRV, and the NSP1 ORF PCR products were inserted into the digested pHTN plasmid using Gibson assembly master mix (New England BioLabs).

To generate plasmids expressing human Cul1, Cul3, Rbx1, KEAP1, KLHL12, and SPOP tagged at the N terminus with hemagglutinin (HA), the ORF of each was amplified from the following templates, respectively: pDONOR201-Cul1 (Dana Farber Harvard Cancer Center PlasmID Repository), pCMV6-XL5-Cul3 (Origene), pENTR223-Rbx1 (Dana Farber Harvard Cancer Center PlasmID Repository), pCMV6-XL5-KEAP1 (Origene), pCMV6-XL5-KLHL12 (Origene), and pCMV6-XL5-SPOP (Origene). Primers were designed to amplify the ORF and add a 15-bp overlap with the plasmid for Gibson assembly cloning. The pCMV-HA plasmid (Clontech) was digested with the restriction enzyme EcoRI, and the ORF PCR products were cloned into the digested pCMV-HA plasmid using Gibson assembly master mix (New England BioLabs). The plasmids pcDNA3-DN-hCul1-Flag (Addgene plasmid number 15818) and pcDNA3-DN-hCul3-Flag (Addgene plasmid number 15820), which express C-terminally Flag-tagged dominant-negative Cul1 and Cul3, respectively, were gifts from Wade Harper (30).

To generate a plasmid for bacterial expression of SA11-4F NSP1 containing a dual His-MBP tag at the N terminus, the NSP1 ORF was amplified from pCI-NSP1 (SA11-4F) (25). Primers were designed to amplify the NSP1 ORF and add a 15-bp overlap with the plasmid for Gibson assembly cloning. A modified pET-32 bacterial expression plasmid (generously provided by Donald Ronning, University of Toledo) was digested with PshAI, and the NSP1 ORF was cloned into the digested plasmid using Gibson assembly master mix (New England BioLabs). All plasmid sequences were verified by sequencing. Primer sequences are available upon request.

Transfections.

For experiments examining the association between HaloTag-NSP1 proteins and endogenous CRL1 and CRL3 complex proteins in mammalian cells, 9.0 × 10⁶ 293T cells were plated in 10-cm dishes. The following day, the cells were transfected with 10 μg of plasmid DNA using 30 μl of PolyJet transfection reagent according to the manufacturer's instructions (SignaGen, Rockville, MD). At 8 h posttransfection (p.t.), the medium containing the transfection reagent was removed, and fresh complete medium was added to the cells, which were then incubated for an additional 16 h. For experiments that treated cells with proteasome inhibitors, at 8 h p.t., fresh complete medium containing 20 μM MG132 (R&D Systems) was added to the cells, which were then incubated for an additional 16 h. The cells were harvested at 24 h p.t. by pelleting, washed once with phosphate-buffered saline (PBS), and stored at −80°C until analysis.

For experiments examining the association between Halo-tagged NSP1 proteins and exogenously expressed, HA-tagged CRL1 and CRL3 complex proteins in mammalian cells, 3.3 × 10⁶ 293T cells were plated in 60-mm dishes. The following day, the cells were transfected with 5 μg of plasmid DNA (2.5 μg of Halo-tagged NSP1 plasmid and 2.5 μg of HA-tagged Cul1, Cul3, KEAP1, KLHL12, SPOP, or Rbx1 plasmid) using 15 μl of PolyJet transfection reagent, according to the manufacturer's instructions (SignaGen, Rockville, MD). At 16 h p.t., the medium containing the transfection reagent was removed and fresh complete medium was added to the cells, which were then incubated for an additional 32 h. Cells were harvested at 48 h p.t. by pelleting, washed once with PBS, and stored at −80°C until analysis. The same transfection protocol was used for experiments examining the association between Halo-tagged NSP1 proteins and Flag-tagged dominant-negative Cul1 and Cul3.

To assess the levels of endogenous IRF3 and β-TrCP in cells transfected with plasmids expressing Halo-tagged NSP1 proteins and wild-type or dominant-negative Cul3, 1.5 × 10⁶ 293T cells were plated in 6-well dishes. The following day, the cells were transfected with 2 μg of plasmid DNA (1 μg of Halo-tagged NSP1 plasmid and 1 μg of Cul3 plasmid) using 6 μl of PolyJet transfection reagent according to the manufacturer's instructions (SignaGen, Rockville, MD). At 16 h p.t., the medium containing the transfection reagent was removed and fresh complete medium was added to the cells, which were then incubated for an additional 8 h. Cells were harvested at 24 h p.t. by pelleting, washed once with PBS, and stored at −80°C until analysis.

Infections.

For immunoprecipitation assays, 9.0 × 10⁶ 293T cells were plated in 10-cm dishes. The following day, the cells were washed three times with serum-free DMEM and then inoculated with rotavirus (strain SA11-4F, SA11-5S, OSU, WI61, UK, or DS-1) at a multiplicity of infection (MOI) of 5 with plaque-determined titers of virus. After a 1-h inoculation, DMEM containing 5% FBS and 1% NEAA was added to the cells. At 10 h postinfection (p.i.), the 293T cells were washed once with PBS, pelleted, and stored at −80°C until immunoprecipitation.

For immunoblot assays to assess protein levels, 5.0 × 10⁵ MA104 cells were plated in 6-well dishes. The following day, the cells were washed three times with serum-free M199 and then inoculated with rotavirus at an MOI of 5. After an inoculation of 1 h, M199 containing 5% FBS was added to the cells. At 10 h p.i., the MA104 cells were washed once with PBS, pelleted, and stored at −80°C until analysis by immunoblotting.

HaloTag pulldown assays.

HaloTag pulldown assays were performed following the manufacturer's instructions in the HaloTag Mammalian Pull-Down Systems Technical Manual (Promega). Briefly, cell pellets were thawed and lysed in 300 μl mammalian lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% Triton X-100, 0.1% sodium deoxycholate, 1× protease inhibitor cocktail [Promega]). After a brief incubation on ice, the lysates were passed through a 25-gauge needle 8 times and centrifuged at 21,100 × g for 10 min at 4°C. Thirty microliters of this whole-cell lysate (WCL) was transferred to a new tube, and an equal amount of 2× SDS sample buffer (200 mM Tris-HCl [pH 6.8], 20% glycerol, 2% SDS, 0.04% Coomassie blue R-250, 2% 2-mercaptoethanol) was added before boiling at 95°C for 10 min. The remaining supernatant was added to 200 μl of prewashed HaloLink resin (Promega) and incubated at room temperature (RT) for 30 min. The resin was washed 4 times with resin equilibration/wash buffer (1× Tris-buffered saline [TBS], pH 7.5, 0.05% Igepal CA-630) and centrifuged between the wash steps at 6,200 × g for 1 min. HaloTag-NSP1 protein complexes were eluted from the HaloLink resin by the addition of 80 μl 2× SDS sample buffer and boiled at 95°C for 10 min.

Immunoprecipitation (IP).

Cul3 was immunoprecipitated following the manufacturer's instructions in the Pierce Cross-link IP kit (Thermo Scientific). Briefly, 40 μl of protein A agarose per sample was added to a spin column and washed with IP lysis/wash buffer (25 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% Igepal CA-630, 5% glycerol). Ten micrograms of Cul3 antibody (Novus) per sample was added to the prewashed protein A agarose and incubated at RT for 1 h. The Cul3 antibody was cross-linked to the protein A agarose by adding the cross-linking reagent disuccinimidyl suberate (DSS) to a final concentration of 25 mM and incubating at RT for 1 h. Anti-Cul3-cross-linked protein A agarose was washed twice with elution buffer (50 mM glycine [pH 2.8]) to remove non-cross-linked antibody and to quench the reaction and then equilibrated with IP lysis/wash buffer. Pellets of rotavirus-infected 293T cells were resuspended in 500 μl of IP lysis/wash buffer containing 1× protease inhibitor cocktail (Thermo Fisher Scientific). The cell lysates were thawed and centrifuged at 21,100 × g for 10 min; 25 μl of this WCL was transferred to a new tube, and an equal amount of 2× SDS sample buffer was added before boiling at 95°C for 10 min. The remaining cell lysate was incubated with the anti-Cul3-cross-linked agarose at 4°C for 2 h. The bound proteins were eluted by the addition of 50 μl 2× SDS sample buffer and boiling at 95°C for 10 min.

Protein knockdown with siRNA.

Dharmacon ON-TARGETplus siRNAs targeted to human Cul1 (Smartpool L-004086-00) and Cul3 (Smartpool L-010224-00) were purchased from GE Healthcare Life Sciences. The nontargeting siRNA used as a negative control was siGENOME nontargeting siRNA pool 2 (D-001206-14). For siRNA knockdown experiments, 5.0 × 10⁵ MA104 cells were plated in 6-well dishes. The following day, the cells were transfected with 25 mM Cul1 siRNA, Cul3 siRNA, or nontargeting siRNA using 4 μl of GenMute transfection reagent (SignaGen, Rockville, MD) according to the manufacturer's instructions. At 16 h p.t., the medium containing the transfection reagent was removed, and fresh complete medium was added to the cells, which were then incubated for an additional 6 h. At 24 h p.t., the transfected MA104 cells were retransfected using the same protocol. At 62 h after the first transfection, the cells were washed three times with serum-free M199 and then inoculated with rotavirus at an MOI of 5 for 1 h, after which M199 containing 5% FBS was added. At 10 h p.i., the cells were washed once with PBS, pelleted, and stored at −80°C until analysis by immunoblotting.

SDS-PAGE and immunoblotting.

For immunoblot assays to assess protein levels in whole-cell lysates, infected or transfected cells were lysed in 200 μl radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 150 mM Tris-HCl [pH 8.0], 1.0% Igepal CA-630, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 1× protease inhibitor cocktail [Thermo Scientific]), briefly sonicated, and then diluted with equal amounts of 2× SDS sample buffer. Proteins were resolved by electrophoresis in 10% Tris-Tricine gels and transferred to a nitrocellulose membrane, which was then blocked by incubating with Odyssey PBS blocking buffer (LiCor) at RT for 1 h. Tween 20 (0.1%) and primary antibodies were added for incubation at 4°C overnight. The membranes were washed three times in Tris-buffered saline (50 mM Tris-HCl [pH 7.5], 150 mM NaCl) containing 0.1% Tween 20 (TBS-T) and then incubated with secondary antibodies conjugated to IRDye680 or IRDye800 (LiCor) at a 1:15,000 dilution in TBS-T containing 1% milk at RT for 2 h. The membranes were then washed three times in TBS-T and imaged using the Odyssey infrared imaging system (LiCor). The experimental bands of interest were quantified using Image Studio software (LiCor) and normalized to the loading control protein used in the specific experiment (either GAPDH or actin). In all quantitative analyses, either mock-infected cells or empty-vector controls were set to 100%, and samples within that experiment were compared to the control sample.

Mass spectrometry of proteins associating with HaloTag-NSP1.

HaloTag pulldown assays were performed with HaloTag-NSP1 from the SA11-4F, SA11-5S, and OSU strains of rotavirus, along with a HaloTag-only control. Proteins that associated with pulldown complexes were eluted in 50 mM Tris-HCl (pH 7.5) containing 1% SDS and sent to the Proteomics Shared Resource facility at Oregon Health Sciences University for tandem mass spectrometry (MS-MS) analysis to identify proteins.

Bacterial protein expression and purification.

The pET-Cul3-Rbx1 plasmid (generously provided by Andrew Mesecar, Purdue University) (31) and p32MBP-NSP1 (SA11-4F) were used to transform BL21(DE3) bacterial cells (New England BioLabs). The bacteria were cultured in LB medium (10 g/liter tryptone, 5 g/liter NaCl, 5 g/liter yeast extract, and 1.5 g/liter Tris base) containing 100 μg/ml ampicillin and incubated at 37°C until the optical density reached a reading between 0.4 and 0.6 nm. Isopropyl β-d-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.2 mM, and the cells were then incubated for 16 h at 16°C before harvesting by centrifugation. The cells were resuspended in His buffer A [50 mM sodium phosphate buffer, 300 mM NaCl, 10 mM imidazole, 1 mM tris(2-carboxyethyl)phosphine (TCEP)] containing 1× DNase I, 1× RNase A, 1× protease inhibitor cocktail (Thermo Fisher Scientific), and 1 mM phenylmethanesulfonyl fluoride (PMSF). The cells were lysed using a French press and then centrifuged at 25,000 × g for 45 min. The supernatant was loaded onto a 5-ml HisTrap HP column (GE Healthcare Life Sciences) that had been equilibrated with His buffer A. To remove nonspecifically bound proteins, the column was washed with 10 column volumes of His buffer A. The bound His-tagged proteins were eluted with a linear gradient using His buffer B (50 mM sodium phosphate buffer, 300 mM NaCl, 250 mM imidazole, 1 mM TCEP). A Bradford assay (Thermo Scientific) was performed to determine which fractions contained the purified protein, and those fractions were then pooled.

Pooled pET-Cul3-Rbx1 fractions from the HisTrap column were loaded for a second time onto a 5-ml HisTrap HP column to remove nonspecifically bound proteins. Fractions containing Cul3-Rbx1 were pooled and dialyzed overnight in exchange buffer (20 mM Tris-HCl [pH 7.5], 100 mM KCl, 1 mM TCEP, 10% glycerol), followed by storage at −80°C. To determine the purity and protein concentrations, the purified protein and bovine serum albumin (BSA) standards were resolved by SDS-PAGE and stained with Coomassie blue R-250 in 50% methanol and 10% glacial acetic acid.

Pooled p32MBP-NSP1 (SA11-4F) fractions from the HisTrap column were dialyzed overnight in MBP buffer A (20 mM Tris-HCl [pH 7.5], 200 mM NaCl, 1 mM EDTA, and 1 mM TCEP). The pooled fractions were loaded on to a 5-ml MBPTrap HP column (GE Healthcare Life Sciences) that was previously equilibrated with MBP buffer A. The column was washed with 10 column volumes of MBP buffer A. The MBP-tagged NSP1 protein was eluted from the column using MBP buffer B (20 mM Tris-HCl [pH 7.5], 200 mM NaCl, 1 mM EDTA, 1 mM TCEP, 10 mM maltose). A Bradford assay (Thermo Scientific) was performed to determine which fractions contained the purified protein, and those fractions were then pooled. The pooled fractions containing NSP1 were dialyzed overnight in exchange buffer. To determine the purity and protein concentrations, the purified protein and BSA standards were resolved by SDS-PAGE and stained with Coomassie blue R-250 in 50% methanol and 10% glacial acetic acid.

In vitro binding assay.

To assess the binding between NSP1(SA11-4F) and Cul3, amylose resin (NEB) was incubated at RT overnight with 1% BSA to block nonspecific binding. The amylose resin was washed 8 times with MBP buffer A to remove all traces of BSA. The recombinant proteins MBP-NSP1(SA11-4F) (MBP-tagged NSP1 from the SA11-4F strain of rotavirus) and His-Cul3-Rbx1 were added to 50 μl of pretreated amylose resin. Binding buffer (10 mM Tris-HCl [pH 8,0], 150 mM NaCl, 2 mM MgCl₂, 0.5% Igepal CA-630, 5% glycerol, 100 μg BSA) was added for a total volume of 50 μl. The reaction mixtures were incubated at 4°C for 1.5 h and then washed 4 times with MBP buffer A. The protein complexes were eluted from the resin, resolved on SDS-PAGE, and immunoblotted with antibodies to detect NSP1 and Cul3 proteins.

RESULTS

Rotavirus NSP1 associates with proteins that function in the host ubiquitin-proteasome pathway.

To identify host proteins that associate with NSP1, HaloTag-NSP1 was transiently expressed in 293T cells for 24 h. The transfected cell lysates were used in a pulldown assay specific for the HaloTag, and samples were analyzed by mass spectrometry for associating proteins. Three different NSP1 proteins were investigated and compared to a pulldown with HaloTag alone in order to ascertain if there were similarities and differences among NSP1s. NSP1 from SA11-4F rotavirus has been shown to induce the degradation of IRF3, IRF5, IRF7, and IRF9 (21, 25). NSP1 from SA11-5S rotavirus is identical to SA11-4F except for a C-terminal truncation of 17 amino acids, which interferes with its ability to induce IRF degradation (25). NSP1 from OSU does not induce IRF degradation but rather inhibits the induction of IFN by causing the degradation of β-TrCP (22, 23). The mass spectrometry results are presented in Table S1 in the supplemental material, after removing proteins containing more than two spectral counts in the HaloTag control and known contaminants. More than 100 proteins were found to associate with HaloTag-NSP1 in the pulldown assays, and many of the host proteins clustered within groups of common functions, including nuclear export, RNA binding, translation, mitochondrial functions, and ubiquitination.

The host proteins involved in the ubiquitin-proteasome system that were found to associate with the three NSP1 proteins tested included Cul3 (cullin 3), Ube3C (ubiquitin-protein ligase E3C), Uba52 (ubiquitin A-52 residue ribosomal protein fusion product 1), and Pgam5 (phosphoglycerate mutase family member 5) (see Table S1 in the supplemental material). β-TrCP protein 2 (FBW1B) was identified only in the HaloTag-NSP1 pulldown from the OSU strain of rotavirus, which was expected based on previous studies that identified an association between OSU NSP1 and β-TrCP (17, 23). In addition to Cul3, which is the scaffolding subunit of CRL3 complexes, several proteins that serve as components or regulatory factors of CRLs were identified, including CAND1 (cullin-associated NEDD8-dissociated protein 1), CAND2, and TCEB2 (transcription elongation factor B, polypeptide 2). Although the mass spectrometry results did not identify each of the ubiquitin-associated proteins in association with all the NSP1s tested, it served as a useful guide to deciding which protein associations to validate by other methods.

NSP1 proteins associate with Cul3 and Cul1.

We chose to investigate the association of Cul3 with NSP1 because a high number of Cul3 peptides were observed to associate with all three NSP1 proteins tested and because additional components of CRL complexes were also identified. Our interest was supported by a study that investigated host protein associations with 70 viral immunomodulatory proteins from a panel of 30 different DNA and RNA viruses, one of which was NSP1 from the OSU rotavirus strain, which was shown to associate with Cul3 by mass spectrometry but was not further validated (32). To confirm the mass spectrometry results, HaloTag-NSP1 proteins from six different strains of rotavirus (simian SA11-4F, simian SA11-5S, porcine OSU, human WI61, bovine UK, and human DS-1) were transiently expressed in 293T cells. These NSP1 proteins were chosen to cover diverse targets of degradation, including IRF3, -5, -7, and -9 (SA11-4F and WI61) (21, 25); IRF3 and -7 (UK); IRF5 and -7 (DS-1) (25); β-TrCP (OSU) (22); and no target (SA11-5S) (25). The cell lysates were used in a HaloTag pulldown and immunoblotted for endogenous Cul3, which associated with all of the NSP1 proteins tested except the C-terminally truncated SA11-5S NSP1 (Fig. 1A, top). Because Cul3 was associated with SA11-5S NSP1 by mass spectrometry, but with fewer peptides identified than were found for SA11-4F or OSU NSP1, it is possible that the C-terminal deletion in NSP1 results in a weaker or less stable association with Cul3 that was undetectable by immunoblotting. A lower level of Cul3 associated with HaloTag-NSP1 from the SA11-4F strain of rotavirus was repeatedly observed, which was also reflected in a lower number of Cul3 peptides identified by mass spectrometry with SA11-4F NSP1. Of note, the NSP1 proteins tested did not induce the degradation of Cul3 in whole-cell lysates, indicating that Cul3 is not a target of NSP1-mediated degradation.

FIG 1.

Full-length HaloTag-NSP1 proteins associate with endogenous (A and B) and exogenously expressed (C and D) Cul1 and Cul3. (A and B) 293T cells were transfected with a plasmid encoding one of the HaloTag-NSP1 proteins or with the HaloTag vector alone as indicated and harvested at 24 h p.t. (A) or medium was removed from the cells and replaced with fresh medium containing 20 μM the proteasome inhibitor MG132 at 8 h p.t., and the cells were incubated for an additional 16 h and then harvested (B). (C and D) 293T cells were cotransfected with a plasmid encoding one of the HaloTag-NSP1 proteins and HA-tagged Cul3 (C) or HA-tagged Cul1 (D). For all the panels, the HaloTag-NSP1 proteins, along with their interacting partners, were captured using HaloLink resin. Proteins in the WCL and HaloTag pulldown (PD) were resolved by SDS-PAGE and analyzed by immunoblotting for Halo, Cul3, Cul1, IRF3, and β-TrCP. Actin or GAPDH was used as a loading control. The upper band of the Cul1 triplet and the Cul3 doublet represents the Nedd8-conjugated form of Cul1 and Cul3, respectively. The blots are representative of the results of three independent experiments.

As a control, the whole-cell lysates from cells transfected with each HaloTag-NSP1 were examined for the degradation of endogenous IRF3 and β-TrCP (Fig. 1A, bottom). Only HaloTag-NSP1 from SA11-4F and UK rotaviruses induced IRF3 degradation. WI61 was expected to induce IRF3 degradation based on previous studies (25) but did not, suggesting the possibility that the HaloTag interfered with WI61 NSP1 binding to or degradation of IRF3. None of the HaloTag-NSP1 proteins tested induced degradation of endogenous β-TrCP, although OSU and WI61 NSP1 have previously been shown to reduce levels of exogenously expressed Flag-tagged β-TrCP (22, 23, 25).

OSU NSP1 has been shown to associate with and induce the degradation of β-TrCP, which inhibits the production of IFN by preventing nuclear translocation of NF-κB (22, 23). Because β-TrCP is a component of CRL1 complexes, endogenous Cul1 and β-TrCP were examined for their association with the same panel of six NSP1 proteins (Fig. 1A, top). β-TrCP associated with OSU, WI61, UK, and DS-1 NSP1 proteins, which is a finding similar to previous reports (22, 23). Cul1 associated with NSP1 from the OSU, WI61, and DS-1 strains of rotavirus. UK NSP1 was expected to pull down Cul1 because of its association with β-TrCP, but it did not pull down Cul1. Also of note, the HaloTag-NSP1 proteins tested did not induce the degradation of Cul1 or β-TrCP in whole-cell lysates (Fig. 1A, bottom).

To determine if any of the endogenous host proteins examined would associate with NSP1 in the presence of a proteasome inhibitor, the HaloTag-NSP1 proteins were transiently expressed in 293T cells and treated with MG132 (Fig. 1B, top). The results of the pulldown in the presence of MG132 were the same as the results in the absence of MG132, suggesting that inhibition of the proteasome did not enhance the association of Cul3 or Cul1 with NSP1. The degradation of IRF3 induced by SA11-4F and UK NSP1 proteins was inhibited in the presence of MG132, whereas the protein levels of Cul3, Cul1, and β-TrCP were unchanged in whole-cell lysates (Fig. 1B, bottom). From these data, Cul3 was identified as a host protein that associated with all five full-length NSP1 proteins tested, leading us to further investigate its role in NSP1 activity.

NSP1 proteins from multiple rotavirus strains associate with components of the CRL3 and CRL1 complexes.

Having identified several NSP1 proteins that associated with endogenous Cul3 and Cul1, we sought to gain a better understanding of where within CRL complexes NSP1 might function. Cul3 and Cul1 serve as the scaffolding subunits of CRL3 and CRL1, respectively. 293T cells were cotransfected with the same panel of six HaloTag-NSP1 proteins and HA-tagged Cul3 or Cul1. The cell lysates were used in a HaloTag pulldown and immunoblotted for associated proteins using an antibody to HA. HA-tagged Cul3 associated with all six NSP1 proteins tested, including the C-terminally truncated SA11-5S NSP1 (Fig. 1C). Although endogenous Cul3 was not detected in association with SA11-5S NSP1 by immunoblotting, it is possible that overexpression of Cul3 allowed a detectable level to associate with the SA11-5S NSP1. HA-tagged Cul1 also associated with all of the NSP1 proteins tested (Fig. 1D). Again, it is possible that overexpression allowed a detectable level of Cul1 to associate with NSP1 proteins, whereas endogenous levels were lower and thus not detectable by immunoblotting.

Rbx1 is the E3 ubiquitin ligase component of both CRL1 and CRL3 complexes. The levels of Rbx1 were unchanged in cells expressing HaloTag-NSP1, indicating that NSP1 does not target Rbx1 for degradation (Fig. 2A, bottom). Because Rbx1 contains a RING domain, NSP1 could potentially substitute for Rbx1 in CRL complexes. To determine if Rbx1 remained a part of CRL complexes in the presence of NSP1, the association of HaloTag-NSP1 with HA-tagged Rbx1 was examined using the HaloTag pulldown assay, as before. HA-tagged Rbx1 associated with all six of the NSP1 proteins tested (Fig. 2A, top). Although this observation does not indicate that Rbx1 directly interacts with NSP1, it suggests that Rbx1 remains a part of the CRL complex in the presence of NSP1 and that NSP1 does not substitute for Rbx1 in CRL complexes.

FIG 2.

HaloTag-NSP1 proteins associate with overexpressed CRL3 complex proteins. 293T cells were cotransfected with a plasmid encoding one of the HaloTag-NSP1 proteins or the HaloTag vector alone and HA-tagged Rbx1 (A), KEAP1 (B), KLHL12 (C), or SPOP (D). Cells were harvested at 48 h p.t., and the HaloTag fusion proteins, along with their interacting partners, were captured using HaloLink resin. Proteins in the WCL and HaloTag PD were resolved by SDS-PAGE and analyzed by immunoblotting for Halo and HA. (A to C) Actin was used as a loading control. (D) PCNA was used as a loading control. The blots are representative of the results of three independent experiments.

OSU NSP1 has previously been shown to associate with and induce degradation of β-TrCP, the substrate recognition component of the CRL1 complex (22, 23). Therefore, it seemed possible that NSP1 could also associate with the BTB substrate recognition proteins in the CRL3 complex. BTB proteins may contain a BTB domain only but frequently also contain additional domains, including the MATH (meprin and tumor necrosis factor receptor-associated factor homology), kelch, NPH3 (nonphototropic hypocotyl 3), ion transport, and zinc finger domains (33). The shared feature of all BTB proteins is involvement in recruiting a substrate to the CRL3 complex for ubiquitination and subsequent degradation. We chose to examine the association of HaloTag-NSP1 with three HA-tagged BTB proteins, each of which has a slightly different domain structure. KEAP1 and KLHL12 contain a BTB-kelch domain, whereas SPOP contains a MATH-BTB domain. Using the HaloTag pulldown assay as before, the association of HA-tagged BTB proteins KEAP1, KLHL12, and SPOP with the panel of six HaloTag-NSP1 proteins was examined. HA-tagged KEAP1 associated with all of the NSP1 proteins tested (Fig. 2B, top). The levels of KEAP1 were similar in whole-cell lysates, suggesting that NSP1 did not induce the degradation of KEAP1 (Fig. 2B, bottom). HA-tagged KLHL12 (Fig. 2C, top) and HA-tagged SPOP (Fig. 2D, top) associated with the NSP1 proteins from the SA11-4F, OSU, WI61, and DS-1 strains of rotavirus, but not with NSP1 from the SA11-5S or UK strain. The differences in association between NPS1 and the BTB proteins examined may be due, in part, to the transient interactions of some BTB proteins in the CRL3 complex. The variability in the association of NSP1 with BTB proteins suggests that while NSP1 is found with components of the CRL3 complex, NSP1 does not directly interact with CRL3 through BTB proteins.

The N terminus of Cul3 directly interacts with NSP1.

In order to identify the region of the Cul3 scaffold protein involved in the association with NSP1, a dominant-negative Cul3 mutant (Cul3 N418) that expresses only the first 418 amino acid residues of the protein, was utilized. The deletion of the C terminus from Cul3 prevents the interaction with Rbx1 and, by extension, with the E2 ubiquitin-conjugating enzyme but retains the ability to associate with BTB proteins (34). To determine if NSP1 associated with the N-terminal region of Cul3, 293T cells were cotransfected with HaloTag-NSP1 proteins and Flag-tagged Cul3 N418. The cell lysates were used in a HaloTag pulldown and immunoblotted for associated Cul3 N418 using an antibody to Flag (Fig. 3A). All six NSP1 proteins examined associated with the dominant-negative Cul3 mutant. Because Rbx1 is unable to associate with this mutant of Cul3 (34), these results suggest that NSP1 does not associate with the Rbx1 component of CRL complexes and further supports the idea that NSP1 does not compete with Rbx1 for binding to Cul3. The same results were observed when a Flag-tagged dominant-negative Cul1 mutant (Cul1 N452) was used to demonstrate the association of HaloTag-NSP1 proteins with the N terminus of Cul1 (Fig. 3B). Again, the Rbx1 binding site is absent from this Cul1 mutant, but the binding site remains for the Skp1 adaptor protein and, by extension, the interacting F-box protein β-TrCP (30, 35), suggesting that the N terminus of Cul1 associates with NSP1.

FIG 3.

HaloTag-NSP1 proteins associate with dominant-negative (DN) Cul3 and Cul1 mutants. 293T cells were cotransfected with a plasmid encoding one of the HaloTag-NSP1 proteins or the HaloTag vector alone and Flag-tagged DNCul3 (A) or DNCul1 (B). Cells were harvested at 48 h p.t., and the HaloTag fusion proteins, along with their interacting partners, were captured using HaloLink resin. Proteins in the WCL and HaloTag PD were resolved by SDS-PAGE and analyzed by immunoblotting for Halo, Flag, and GAPDH (loading control). The blots are representative of the results of three independent experiments.

To examine if NSP1 directly binds to Cul3, MBP-tagged NSP1 from the SA11-4F strain of rotavirus was expressed in bacteria and purified using affinity chromatography. Similarly, a His-tagged Cul3-Rbx1 complex was expressed in bacteria and affinity purified. A binding assay was performed by immobilizing MBP-NSP1 on amylose resin and incubating with or without His-Cul3-Rbx1 for 1 h. The protein complexes were eluted from the beads, and the NSP1–Cul3-Rbx1 interaction was examined by immunoblotting with antibodies specific for SA11-4F NSP1 and Cul3 (Fig. 4). His-tagged Cul3-Rbx1 was pulled down in the presence of MBP-tagged NSP1, but not in the presence of MBP alone, suggesting the interaction was specific and direct. In a reciprocal experiment, the His-tagged Cul3-Rbx1 protein was immobilized on cobalt resin and MBP-tagged NSP1 was pulled down (Fig. 4). Together, these data indicate that NSP1 directly interacts with Cul3.

FIG 4.

Recombinant NSP1 directly interacts with Cul3. MBP-NSP1(SA11-4F) (750 ng) was immobilized on amylose resin by incubating for 1 h at 4°C. Recombinant His-Cul3-Rbx1 protein was incubated with (+) or without (−) MBP-NSP1(SA11-4F) for 1 h at 4°C. In the reciprocal experiment, 20 μg of recombinant His-Cul3-Rbx1 was immobilized on cobalt resin by incubating for 1 h at 4°C. Recombinant MBP-NSP1 (SA11-4F) protein was incubated with (+) or without (−) His-Cul3-Rbx1for 1 h at 4°C. After incubation, bound proteins were eluted from the resin using SDS loading buffer and analyzed by immunoblotting for Cul3 and the viral protein NSP1 (SA11-4F). The same protein concentrations used in binding assays were resolved by SDS-PAGE and immunoblotted for NSP1 and Cul3 as shown (Input). The blots are representative of the results of three independent experiments.

NSP1 proteins associate with endogenous CRL3 complex during rotavirus infection.

To determine if rotavirus infection affected the levels of Cul3, the levels of Cul3, Cul1, IRF3 and β-TrCP were examined in rotavirus-infected cells. MA104 cells were infected with six different strains of rotavirus (SA11-4F, SA11-5S, OSU, WI61, UK, and DS-1), and at 10 h p.i., cells were harvested. Whole-cell lysates were separated by SDS-PAGE, and protein levels were examined by quantitative immunoblot analysis (Fig. 5). An antibody to detect the viral protein VP6 was used to verify that cells were infected. The levels of Cul3 and Cul1 were unchanged in the infected cells compared to mock-infected cells. The IRF3 protein was undetectable in SA11-4F- and UK-infected cells, which was consistent with the loss of IRF3 in cells transfected with HaloTag-NSP1 proteins from the two rotavirus strains (Fig. 1A). β-TrCP levels were substantially reduced in OSU-, WI61-, UK-, and DS-1-infected MA104 cells compared to the level of β-TrCP in mock-infected cells (Fig. 5). This result was in contrast to the HaloTag-NSP1-transfected 293T cells, where β-TrCP levels were unchanged (Fig. 1B). This could be due to differences in the context of viral infection, leading to loss of β-TrCP that is not fully recapitulated in transfected cells. Overall, these results indicated that the levels of Cul1 and Cul3 were unchanged in rotavirus-infected cells and that the target substrate proteins IRF3 and β-TrCP were degraded as expected.

FIG 5.

Rotavirus infection does not alter the cellular level of Cul3 or Cul1. MA104 cells were mock infected or infected with rotavirus strain SA11-4F, SA11-5S, OSU, WI61, UK, or DS-1 (MOI = 5). Cells were harvested at 10 h p.i., and total cellular protein was resolved by SDS-PAGE, followed by immunoblotting for endogenous Cul3, Cul1, IRF3, β-TrCP, and GAPDH proteins. An antibody to detect the viral protein VP6 was used to verify that cells were infected. The blots are representative of the results of three independent experiments.

To investigate the ability of NSP1 to associate with the CRL3 complex during rotavirus infection, 293T cells were infected with the six different strains of rotavirus. At 9 h p.i., cells were harvested and lysates were incubated with a Cul3 antibody cross-linked to protein A agarose beads. The immunoprecipitated Cul3 and associated proteins were resolved by SDS-PAGE, and immunoblots were used to detect NSP1. NSP1 was expressed in all of the rotavirus-infected whole-cell lysates, and Cul3 was present at approximately equivalent levels in the mock- and rotavirus-infected cells (Fig. 6, left lanes). Cul3 was immunoprecipitated from all infected cell lysates, and NSP1 from the five wild-type viruses tested was shown to coimmunoprecipitate with Cul3 (Fig. 6, right lanes). The C-terminally truncated NSP1 from the SA11-5S strain of rotavirus did not coimmunoprecipitate with Cul3, similar to what was observed in the HaloTag-NSP1 pulldown of endogenous Cul3 in transfected cells (Fig. 1A). These results indicate that the association of NSP1 with Cul3 occurs across a number of wild-type rotavirus strains, including animal (SA11-4F, OSU, and UK) and human (WI61 and DS-1) virus isolates. The failure of NSP1 from SA11-5S to associate with Cul3 in infected cells may indicate that the C terminus of NSP1 is a part of the Cul3-interacting domain. Alternatively, the C terminus may play an important role in folding of the NSP1 protein, and deletion of that domain could potentially cause a conformational change that limits the interaction of NSP1 with Cul3.

FIG 6.

Rotavirus NSP1 associates with Cul3 during viral infection. 293T cells were either mock infected or infected with the indicated strains of rotavirus (MOI = 5) and harvested at 10 h p.i. The cell lysates were incubated with Cul3 antibodies cross-linked to protein A agarose for 1 h at 4°C. The protein complexes that immunoprecipitated with Cul3 were eluted from the protein A agarose. Proteins in the WCL and immunoprecipitation elution were resolved by SDS-PAGE, followed by immunoblot analysis using antibodies specific for Cul3 and individual NSP1 proteins. GAPDH was used as a loading control for WCL fractions. The blots are representative of the results of two independent experiments.

Induction of IRF3 and β-TrCP degradation by NSP1 are independent of its association with Cul1 and Cul3.

Rather than acting as an E3 ubiquitin ligase itself, NSP1 may usurp CRL complexes to induce degradation of IRF3 and β-TrCP. To determine if loss of Cul3 or Cul1 had an impact on NSP1-mediated IRF3 and β-TrCP degradation in rotavirus-infected cells, an siRNA approach was used. MA104 cells were transfected twice with no siRNA, a nontargeting control siRNA, or a pool of 4 siRNAs targeting Cul1 or Cul3. At 62 h after the first transfection, cells were infected with different rotavirus strains, and at 10 h p.i. (72 h p.t.), cells were harvested. Whole-cell lysates were separated by SDS-PAGE, and protein levels were examined by quantitative immunoblot analysis. Knockdown of Cul3 in MA104 cells resulted in reduced Cul3 protein levels, which on average were about 25% of the Cul3 levels observed in cells treated with nontargeting siRNA (set at 100%) (Fig. 7A). Cul1 protein levels were unaffected in cells treated with Cul3 siRNA, indicating that the knockdown was specific and that there was no compensatory increase in Cul1 when Cul3 levels were reduced. If Cul3 were required for NSP1-mediated degradation of the target protein IRF3 or β-TrCP, an increase in these target proteins would be expected when Cul3 was knocked down by siRNA. Unexpectedly, when Cul3 protein levels were reduced by siRNA, the levels of IRF3 in SA11-4F- and UK-infected cells did not return to the levels observed in control siRNA-treated cells, as IRF3 was nearly undetectable by immunoblotting (Fig. 7A). There was no significant loss of IRF3 in cells infected with SA11-5S, OSU, WI61, or DS-1 virus, as was expected based on previous results, nor did loss of Cul3 dramatically alter IRF3 levels in cells infected with these viruses. Similarly, the levels of β-TrCP in control nontarget siRNA-treated cells and mock-infected cells were set at 100%. As expected, β-TrCP levels were reduced to below 20% in OSU-, WI61-, UK-, and DS-1-infected cells. The levels of β-TrCP did not return to that observed in control siRNA-treated cells, although there was a small observable increase in UK- and DS-1-infected cells. These results were reproducible in several independent experiments and were observed in both MA104 (Fig. 7) and 293T (data not shown) cells.

FIG 7.

Knockdown of Cul1 or Cul3 by siRNA does not prevent NSP1-induced degradation of IRF3 and β-TrCP. (A) MA104 cells were transfected twice (24 and 48 h postseeding) with no siRNA (None), control nontarget siRNA (siNT), or siRNA targeting Cul3 (siCul3). At 62 h p.t., the cells were mock infected or infected with rotavirus strain SA11-4F, SA11-5S, OSU, WI61, UK, or DS-1 at an MOI of 5. (B) MA104 cells were transfected twice (24 and 48 h postseeding) with no siRNA (None), control nontarget siRNA (siNT), or siRNA targeting Cul1 (siCul1). At 62 h p.t., the cells were mock infected or infected with rotavirus strain SA11-4F, SA11-5S, OSU, or UK at an MOI of 5. For both panels, cells were harvested at 10 h p.i., and total cellular proteins were resolved by SDS-PAGE, followed by quantitative immunoblot analysis for endogenous Cul1, Cul3, IRF3, β-TrCP, and GAPDH (loading control). The band intensities were normalized to the loading control (GAPDH) and expressed as a percentage of those of the proteins in the siNT-transfected cells and mock-infected cells, which were set at 100%. An antibody to detect the viral protein VP6 was used to verify that cells were infected. The blots are representative of the results of three independent experiments.

Knockdown of Cul1 by siRNA in MA104 cells resulted in reduced levels of Cul1, which on average were about 35% of the Cul1 levels observed in cells treated with nontargeting siRNA (set at 100%) (Fig. 7B). Cul3 protein levels were unaffected in cells treated with Cul1 siRNA, indicating that the knockdown was specific, and there was no compensatory increase in Cul3 when Cul1 was reduced. As with Cul3, if Cul1 were required for NSP1-mediated degradation of the target protein IRF3 or β-TrCP, an increase in these target proteins would be expected when Cul1 was knocked down by siRNA. IRF3 levels in Cul1 siRNA-treated cells remained low in SA11-4F- and UK-infected cells, and β-TrCP levels in Cul1 siRNA-treated cells remained low in OSU-, WI61-, UK-, and DS-1-infected cells. The Cul1 siRNA treatments were also performed in 293T cells with the same outcome (data not shown), suggesting that Cul3 and Cul1 are not critical in the NSP1-mediated degradation of IRF3 and β-TrCP. Double-knockdown experiments to reduce Cul1 and Cul3 levels at the same time did not rescue the IRF3 or β-TrCP target, suggesting that the Cul1 and Cul3 interactions with NSP1 are not functionally redundant (data not shown). Because low levels of Cul3 and Cul1 remained after each siRNA treatment, there could have been sufficient amounts remaining in the cell to function with NSP1. Alternatively, there could be other, unknown target substrates of NSP1 that are degraded by an NSP1-CRL complex in infected cells.

To further support the siRNA knockdown data, we turned again to the dominant-negative Cul3 mutant, which does not contain the Rbx1 binding site and, by extension, also does not bind to the E2 ubiquitin-conjugating enzyme. If NSP1 utilized the CRL3 complex to induce degradation of IRF3 and β-TrCP, then in the presence of dominant-negative Cul3, the complex would no longer function and the levels of IRF3 and β-TrCP would be expected to remain similar to mock levels. 293T cells were cotransfected with the panel of six HaloTag-NSP1 proteins and Flag-tagged wild-type Cul3 or dominant-negative Cul3 (Cul3 N418). Cells were harvested at 24 h p.t., and proteins were separated by SDS-PAGE, followed by quantitative immunoblotting for the HaloTag, Flag, IRF3, and β-TrCP (Fig. 8, top). In cells expressing Flag-tagged wild-type Cul3, the level of endogenous IRF3 was set at 100% when coexpressed with the HaloTag empty vector, and this was used for comparison to cells coexpressing HaloTag-NSP1 (Fig. 8, middle). IRF3 levels were reduced to approximately 40% of the empty-vector control in the presence of HaloTag-NSP1 from SA11-4F and UK rotaviruses, but not the other NSP1 proteins tested, similar to what was observed in earlier experiments (Fig. 1A). In the same way, in cells expressing Flag-tagged dominant-negative Cul3, IRF3 levels were set at 100% when the HaloTag vector was coexpressed, and this was used for comparison to cells expressing HaloTag-NSP1 (Fig. 8, middle). The levels of IRF3 in the presence of the dominant-negative Cul3 were also approximately 40% of those of the vector control in the presence of HaloTag-NSP1 from SA11-4F and UK rotaviruses, but not the other NSP1 proteins tested.

FIG 8.

A DN Cul3 mutant does not affect the ability of NSP1 to induce degradation of target proteins. Plasmids encoding one of the HaloTag-NSP1 proteins or the HaloTag vector alone were cotransfected into 293T cells with Flag-tagged Cul3 (wild type) or DNCul3. Cells were harvested at 24 h p.t., and total cellular protein was resolved by SDS-PAGE and analyzed by immunoblotting for HaloTag and Flag-tagged proteins, as well as endogenous IRF3, β-TrCP, and GAPDH (loading control). IRF3 and β-TrCP protein levels were quantified using the Odyssey infrared imaging system. The band intensities were normalized to the loading control (GAPDH) and expressed as a percentage of the protein in HaloTag empty-vector (Vector)-transfected cells, which was set at 100% (denoted by the dashed lines). The mean values from three experiments are shown, and the error bars represent the standard errors of the mean (SEM).

The levels of β-TrCP were also examined in cells expressing either wild-type or dominant-negative Cul3 (Fig. 8, top). In cells expressing Flag-tagged wild-type Cul3, the level of endogenous β-TrCP was set at 100% when coexpressed with the HaloTag empty vector, and this was used for comparison to cells coexpressing HaloTag-NSP1 (Fig. 8, bottom). β-TrCP levels were reduced to below 50% of those for the empty-vector control in the presence of HaloTag-NSP1 from OSU, WI61, and UK rotaviruses, and β-TrCP levels were reduced to 60% of those for the control in the presence of HaloTag-NSP1 from DS-1. In cells expressing the dominant-negative Cul3, there was a small but reproducible increase in β-TrCP levels compared to cells expressing the wild-type Cul3. This increase was more notable in cells transfected with HaloTag-NSP1 from WI61 and UK rotaviruses. However, because of transfection inefficiency, the expression of dominant-negative Cul3 may not have been adequate to inactivate all of the endogenous Cul3 protein, leaving sufficient Cul3 in the cell to function with NSP1 to mediate the degradation of the IRF3 and β-TrCP targets. The results from the dominant-negative Cul3 experiments, along with the siRNA experiments, suggest that NSP1 mediates the degradation of IRF3 and β-TrCP independently of its interaction with Cul3.

DISCUSSION

The NSP1 protein from rotavirus has been proposed to act as an E3 ubiquitin ligase that is responsible for inducing the degradation of IRF proteins and β-TrCP, but direct evidence for E3 ligase activity has been lacking. The need for experimental data to support the hypothesis that NSP1 functions as an E3 ligase is reiterated by this study, which showed that NSP1 associates with CRL complexes, which are essential E3 ubiquitin ligase complexes used in the regulation of numerous host cell processes. Our initial prediction was that NSP1 could usurp CRL3 or CRL1 complexes to induce degradation of IRFs and β-TrCP, but the data unexpectedly showed that this was not the case. While this study does not yet provide a mechanistic understanding of NSP1-induced host protein degradation, it does represent an exciting new interaction of NSP1 with the host cell ubiquitination machinery.

The depletion of Cul3 was expected to prevent the NSP1-mediated degradation of IRF3 and β-TrCP targets. Although there was a slight recovery of IRF3 or β-TrCP in Cul3 siRNA-treated cells (Fig. 7A), the results did not conclusively show that Cul3 is required for NSP1-mediated protein degradation. Similar results were observed when Cul1 was knocked down by siRNA. The inability to completely knock down Cul3 or Cul1 may have left a small amount of Cul3 that was sufficient for NSP1 to usurp the CRL3 complex to induce target degradation. Alternatively, there could be redundancy of CRL3 and CRL1, but it is uncommon for a single substrate protein to be targeted for degradation by both complexes. A previous study that performed a genome-wide RNA interference (RNAi) screen in MA104 cells to examine host proteins that are essential for rotavirus entry and replication identified Cul3 as important for viral infection (36). However, Cul3 was identified only in the screen and was not validated in follow-up assays in that study. Our observations suggest that Cul3 knockdown by siRNA did not have an impact on viral replication, as VP6 levels were unchanged in siRNA-treated cells (Fig. 7), stressing the importance of follow-up validation studies with large-scale screens. We also examined virus titers in infected cells treated with Cul3 siRNA but did not observe an appreciable change in virus replication (data not shown). This result was expected, since the highly permissive cell culture model systems used in the laboratory do not require NSP1 for viral replication to occur (37). Similarly, natural rotavirus variants that express truncated or no NSP1 protein replicate to levels similar to those of their parental strains in cell culture. There is not currently a system in which to study the effects of altering NSP1 activity on viral replication. Because knocking down or eliminating the expression of NSP1 does not impact viral titers in permissive cell culture models, it is not surprising that knocking down or eliminating a cofactor or host protein that is used by NSP1, such as Cul3, would also not result in a change to virus titers.

The primary role of the NSP1 protein during rotavirus infection is the inhibition of the host type I IFN response. Because the NSP1 amino acid sequence varies greatly among different isolates of group A rotaviruses, it appears that the step at which NSP1 inhibits the IFN response also varies among different NSP1 proteins. As such, numerous targets of NSP1-mediated degradation have been identified, including IRF proteins (16, 19–21), β-TrCP (17, 23), MAVS (mitochondrial antiviral signaling protein) (38, 39), TRAF2 (tumor necrosis factor [TNF] receptor-associated factor 2) (40), and several other host proteins (41–43). Cul3 and Cul1 proteins are not targets of NSP1-mediated degradation (Fig. 1), but by interacting with NSP1, they might play a role in its activity. The ability of NSP1 to induce the degradation of host proteins and the presence of a conserved RING domain at the N terminus of NSP1 suggest that the mechanism by which NSP1 induces degradation of so many different proteins may be shared. Alternatively, NSP1 may regulate an unidentified protein that controls the levels of many different proteins that have been observed to be lost in rotavirus-infected cells, but this idea needs further exploration. Ultimately, in vitro ubiquitination assays will be the best indicator of how NSP1 functions to degrade host proteins, as those assays will define the minimal components required for ubiquitination by NSP1.

CRL complexes are responsible for nearly 20% of all protein ubiquitination in the cell; thus, it is no surprise that viruses hijack CRLs for survival. Human immunodeficiency virus, Kaposi's sarcoma-associated herpesvirus, hepatitis B virus, and Rift Valley fever virus have all been shown to utilize the CRL1 complex (44). For example, the ORF2 protein of hepatitis E virus and the Vpu protein of human immunodeficiency virus type 1 sequester the CRL1 complex by preventing β-TrCP-mediated degradation of IκBα, thus inhibiting NF-κB activation (45, 46). The only viruses currently known to use the CRL3 complex are viruses within the family Poxviridae. Ectromelia virus encodes two BTB-kelch proteins, EVM150 and EVM167, which interact with Cul3 and are required for viral replication and NF-κB signaling (47, 48). Our study is the first to report that Cul3 interacts with rotavirus NSP1. Although we have not defined the role of Cul3 in the ability of NSP1 to induce degradation of IRF3 and β-TrCP, further study is warranted to understand how rotavirus uses CRLs for its own survival.

Cullins are posttranslationally modified by the ubiquitin-like protein Nedd8. Neddylation of Cul3 occurs near the Rbx1 binding site and is required to form the binding interface that recruits the E2-conjugating enzyme (4). CAND1 is an inhibitory protein that wraps around a deneddylated cullin to regulate the CRL complex. The presence of Nedd8 on a cullin blocks CAND1 binding, suggesting that CAND1 binds to a cullin only after Nedd8 has been removed (4). Identification of the neddylation status of Cul3 in rotavirus-infected cells may lead to clues to how NSP1 usurps this complex for its own function. CAND1 was identified in our mass spectrometry results, suggesting that CRLs may be in the deneddylated form when interacting with NSP1, but this idea requires further investigation. Additionally, CRLs are capable of forming homomultimers and possibly heteromultimers, thereby expanding their conformational space and target substrates. Whether NSP1 functions with monomeric or multimeric forms of CRLs remains to be explored. Further studies of the biochemical details of NSP1 interactions with CRL complexes may lead to a determination of the molecular structure of the NSP1 protein, which would certainly aid in understanding the function of NSP1 in IFN-β inhibition. From this study, we were unable to determine the function of the interaction of NSP1 with Cul3. We have not completely discarded the potential of NSP1 using the CRL3 complex to induce protein degradation, but possibly involving targets of NSP1 different from those that were investigated here. This study highlights the need to better define the mechanism of viral IFN antagonist proteins, such as NSP1, as these proteins play important roles in the pathogenesis of viral infection.