Fig 1.
Phylogenetic and sequence-based analysis of the SARS-related bat coronavirus papain-like protease (SR-PLP).
(a) Phylogeny of SARS-related beta-CoVs in the PLP gene (981 bp fragment) within the nonstructural protein 3. PLP genes characterized in the study are colored in red. The right-hand column shows the species classification of the included virus clades according to the International Committee on Taxonomy of Viruses (ICTV). Phylogenetic trees of SARS-related betacoronaviruses (CoVs) were calculated by the Neighbor Joining algorithm in Geneious under the assumption of a Tamura-Nei genetic distance model. Symbols correspond to the respective host species (human, civet and bat). The scale bar refers to the genetic distance. The SARS-outlier CoV (SO-CoV) was identified in a Ghanaian Hipposideros bat. SO-CoV belongs to a novel unclassified beta-CoV species. HCoV: human CoV, FRA: SARS Frankfurt strain, BtCoV: bat CoV. The accession numbers are as follows: HCoV_SARS/FRA: AY310120, Civet CoV_SARS: AY572034, BtCoV_Rp3: DQ071615, BtCoV_Rm1: DQ022305, BtCoV_Bulgarian: GU190215, BtCoV_Ganaian: MG916963, HCoV_MERS/EMC: JX869059. (b) Amino acid sequence alignment for the comparison of SR-PLP to SA-PLP. The alignment is based on the amino acid codes by the Blosum62 algorithm in the Geneious 6 software package. The SO-CoV derived PLP (SO-PLP) was included as an outlier PLP. Yellow boxes indicate conserved residues in all sequences. The boxes in light grey indicate conserved residues in only two sequences. Residues that form the catalytic center are indicated by grey arrows below the sequences. The catalytic cysteine, which was mutated to alanine in the course of this study, is highlighted in red. The ubiquitin-binding methionine at amino acid position 209, which was mutated to arginine (M209R) in this study, is marked in blue. Zinc-binding residues, important for the three dimensional PLP structure, are indicated by asterisks above the sequences. C1651 numeration refers to the position in the SARS-CoV pp1a already used before [46]. Residues framed in black indicate the binding sites of the inhibitor compound 3e, which was used in the course of this study. SA: SARS; SR: Bulgarian; SO-PLP: Ghanaian.
Table 1.
Amino acid identity and similarity matrix.
Fig 2.
SR-PLP had conserved protease cleavage-, DUB- and deISGylating activities.
(a) HEK-293T cells were transfected with empty vector plasmid (EV) or plasmids expressing either wild type (WT)- or catalytic mutant (CA) PLPs. SARS-CoV nsp2/3-GFP was cotransfected simultaneously. Lysates were harvested at 16 hours post transfection (hpt), and gene expression was analyzed by Western blotting. (b) A biosensor assay was applied for the detection of PLP cleavage activity. Cells were cotransfected with pGlo Firefly luciferase, and either WT-PLP, CA or EV plasmids in 96-wells. At 14 hpi, cells were incubated with GloSensor reagent and luminescence was detected. (c) To investigate the activity spectrum of a SA-PLP protease inhibitor, one hour after the GloSensor incubation, 12.5, 25 or 50 μM of compound 3e or DMSO were added. PLP activities were analyzed in relation to the different amounts of compound 3e at 4 h post treatment. Values were normalized to the respective DMSO-treated WT-PLP. Biosensor assays were performed in triplicate and repeated three times independently. Error bars indicate standard deviations of the means. Statistical significance between DMSO and inhibitor-treated cells or cells transfected with the CA-PLPs, respectively, was determined using one-way ANOVA and Sidak post hoc test. Statistically significant differences are indicated by asterisks (p> 0.05 not significant (ns), p≤ 0.05 significant (*), p≤ 0.01 very significant (**), p≤ 0.001 highly significant (***)). (d) HEK-293T cells were transfected with EV or plasmids expressing either WT- or CA-PLPs. For analysis of DUB activity WT-PLP plasmids were transfected in increasing amounts of plasmids (50 ng, 100 ng and 200 ng per 12-well). Control plasmids (200 ng) EV and CA-PLP were transfected for comparison. HA-ubiquitin (HA-ub) was coexpressed in all samples. Lysates were harvested at 18 hpt, and gene expression was analyzed by Western blotting. (e) For analysis of deISGylating activity pISG15-myc and the conjugation machinery (UbcH8, Ube1L, and Herc5) of ISG15 were coexpressed in all samples. Lysates were harvested at 18 hpt, and gene expression was analyzed by Western blotting. Western blot experiments were repeated three times independently and one representative blot is shown. β-Actin was applied as loading control.
Fig 3.
SR-PLP supported viral replication in the context of a recombinant, chimeric SARS-CoV (rSCV).
(a) A chimeric rSCV containing SR-PLP (SR-PLP-rSCV) was generated by reverse genetics to investigate SR-PLP functions in the context of a replicating virus. SR-PLP (purple) was inserted at the genomic position of SA-PLP (blue). The plaque morphology was analyzed in Vero E6 cells. Therefore, cells were infected with (b) rSCV and (c) SR-PLP-rSCV (MOI 0.01) and overlaid with a highly viscous medium. At 3 dpi, cells were fixed and stained with crystal violet. (d) To investigate structural integrity of the SR-PLP domain in the molecular context of the SARS-CoV nonstructural protein 3, cells were infected with either rSCV or SR-PLP-rSCV (MOI 0.0001) and treated with the SA-PLP protease inhibitor 3e (24-well format). After 1 h, cells were washed twice with PBS and either DMSO (0 μM) or in DMEM serially diluted compound 3e (3.125, 6.25, 12.5, 25 and 50 μM) was added. Supernatants were collected at 24 hpi. For virus quantification a real-time RT-PCR for genomic SARS-CoV RNA was performed. The experiment was done in triplicate and repeated twice. Error bars indicate the standard deviations of the means. Statistical significance between DMSO- and compound-treated cells was determined using one-way ANOVA and Sidak- or Games-Howell post hoc tests for rSCV and SR-PLP-rSCV, respectively. The 50% effective concentrations (EC50) were rSCV: IC50 = 2.36 μM and SR-PLP-rSCV: IC50 = 11.02 μM.
Fig 4.
Chimeric SR-PLP-rSCV grew less efficient in presence of type I interferon (IFN).
Virus growth was compared in (a) type I IFN-deficient (Vero) primate cells, and (b) IFN-competent (MA-104) primate cells. For virus growth kinetics, cells were infected with rSCV and SR-PLP-rSCV (MOI 0.01). Supernatants were taken at 0, 8, 14, 24 and 48 hpi, and viral replication was determined by a plaque titration assay. Growth experiments in Vero and MA-104 cells were done in triplicate and repeated twice. Error bars indicate the standard deviations of the means. Infectious particle production of rSCV and SR-PLP-rSCV was compared using SPSS Version 23.0.0.0 and a general linear regression model. Growth of both viruses did not significantly differ in Vero cells (p = 0.929 and R-square = 0.670). In MA-104 cells growth of the viruses significantly differed (p = 0.038 and R-square = 0.612). (c) Vero cells were treated with 100 IU/ml of recombinant pan-species IFN-α. At 16 h post IFN treatment, cells were infected with rSCV and SR-PLP-rSCV (MOI 0.01), respectively. Supernatants were taken at 24 hpi, and viral replication was determined by a plaque titration assay. The experiment was performed in triplicate and repeated twice. Error bars indicate the standard deviations of the means. One representative experiment is shown. (d) Virus growth was compared in IFN-competent human cells (Calu-3) as described above. The experiment was performed in triplicate. (e) To determine IFN-β expression, Calu-3 cells were infected with rSCV, SR-PLP-rSCV or IFN-inducing RVFV Cl 13 (control of IFN-β expression) at an MOI of 1. Total mRNA was extracted from cell lysates at 24 hpi. IFN-β expression was determined using quantitative real-time PCR analysis. The mean fold change in IFN-β expression was calculated using TATA-binding protein (TBP) expression as a reference gene and the 2−ΔΔCt analysis method [55]. The experiment was done in quadruplicates. Statistical significance between the indicated groups was determined using a two-sided t test. (f) Human airway epithelial cells (HAE) were infected with rSCV and SR-PLP-rSCV with an absolute infectious dose of 40,000 PFU. At 0, 48, 72 and 96 hpi samples were taken and viral replication was determined by a plaque titration assay. The experiment was done in duplicate and repeated three times independently. Statistical significance in (d-f) was determined using a two-sided t test.
Fig 5.
Protease-dependent and protease-independent IFN-antagonistic functions of PLPs.
(a) Primate cells (MA-104) were transfected with plasmids expressing 1 μg of WT- or CA-PLPs. EV was applied as a control. GFP-IRF-3 was coexpressed in each sample. At 24 hpi, cells were infected with IFN-inducing, recombinant Rift Valley fever virus clone 13 (RVFV Cl 13) at an MOI of 5. Cells were fixed at 8 hpi. Immunofluorescence pictures show representative results. White arrows indicate representative cells. (b) For quantification of IRF3 translocation events at least 4 representative pictures were taken. The total number of cells positive for GFP-IRF-3 and the respective PLP were determined and the number of GFP-IRF-3 nuclear translocation was calculated. The result represents one of three independently performed experiments. (c, d, e) An IFN-β promoter activation assay was conducted in HEK-293T cells to determine the anti-IFN activities of (c) SA-PLP and (d) SR-PLP and (e) SO-PLP. Cells were transfected with plasmids expressing MDA5 (IFN stimulator), Firefly- (detection of IFN-β promoter activity) and Renilla- (detection of the general transcription level) luciferases. 50 ng/24-well of EV were transfected and samples were applied as induction controls. WT- and CA-PLPs were coexpressed in a dose-dependent manner (10, 25 and 50 ng/24-well). At 17 hpt, cells were lysed and the luciferase activity was measured. Results are presented as induction relative to EV. The graph shows results of three independently performed experiments, which were all done in triplicate. Expression of the PLPs was confirmed by Western blotting. β-Actin was applied as a loading control. The error bars in each graph indicate the standard deviations of the means. Statistical significance between infected/induced EV samples and PLP-expressing cells was determined using one-way ANOVA and Dunnett-T3 post hoc test.
Fig 6.
The protease-independent IFN-antagonistic activity of SA-PLP is related to the interaction between SA-PLP and ubiquitin.
(a) A trans-cleavage assay was conducted in HEK-293T cells to determine the catalytic activity of SA-PLP and the SA-PLP mutant M209R, which had the ubiquitin-binding residue methionine (M) at amino acid position 209 mutated to arginine (R). (b) For analysis of DUB activity, HEK-293T cells were transfected with HA-ub and 200 ng of plasmids expressing either SA-PLP, the SA-PLP mutant M209R or the respective CA-mutants. Lysates were harvested at 18 hpt, and gene expression was analyzed by Western blotting. (c) IFN-β promoter activity was determined using an IFN-β promoter activation assay. 50 ng of SA-PLP encoding plasmids were coexpressed with the luciferase genes and IFN-inducing MDA5. Statistical significance between infected/induced EV samples and PLP-expressing cells was determined using one-way ANOVA and Dunnett-T3 post hoc test. Virus growth of rSCV and the modified rSCV-PLP/M209R, carrying the M209R mutation, was compared in type I IFN-deficient primate cells (Vero, d), and IFN-competent primate cells (MA-104, e). For virus growth kinetics, cells were infected with rSCV and rSCV-PLP/M209R at an MOI of 0.01. Each experiment was performed in triplicate. MA-104 growth experiment was repeated twice. Particle production of rSCV and rSCV-PLP/M209R was compared using SPSS Version 23.0.0.0 and a general linear regression model. Growth of both viruses did not significantly differ in Vero cells (p = 0.613 and R-square = 0.020). In MA-104 cells growth of the viruses significantly differed (p = 0.034 and R-square = 0.217). (f) In order to determine the sensitivity of rSCV and rSCV-PLP/M209R towards type I IFN, Vero cells were treated with 100 IU/ml of recombinant pan-species IFN-α. At 16 h post treatment, cells were infected with rSCV or rSCV-PLP/M209R (MOI 0.01) for 24 h. The experiment was performed in triplicate. Statistical significance between the indicated groups was determined using a two-sided t test. (g) Virus growth was compared in human-derived, IFN-competent cells (Calu-3). Cells were infected with rSCV or rSCV-PLP/M209R at an MOI of 0.01. The experiment was done in triplicate. Statistical significance was determined using a two-sided t test. (h) To determine IFN-β expression, Calu-3 cells were infected with rSCV, rSCV-PLP/M209R or IFN-inducing RVFV Cl 13 (control of IFN-β expression) at an MOI of 1. Total mRNA was extracted from cell lysates at 24 hpi. IFN-β expression was determined using quantitative real-time PCR analysis. The mean fold change in IFN-β expression was calculated using TATA-binding protein (TBP) expression as a reference gene and the 2−ΔΔCt analysis method [55]. The experiment was done in quadruplicates. Statistical significance between the indicated groups was determined using a two-sided t test.
Fig 7.
SR-PLP-rSCV grew less efficiently in IFN-competent, bat-derived lung cells.
(a) For virus growth kinetics Rhinolophus alcyone lung cells, which expressed the SARS-CoV receptor human angiotensin converting enzyme (ACE)-2, were infected with rSCV and SR-PLP-rSCV at an MOI of 0.1. Supernatants were taken at 0, 24, 48 and 72 hpi, and viral replication was determined by a plaque titration assay. The experiment was performed in triplicate. Error bars indicate the standard deviations of the means. (b) RhiLu-hACE2 cells were infected with rSCV, SR-PLP-rSCV or IFN-inducing RVFV Cl 13 (control of IFN-β mRNA expression) at an MOI of 1. Total mRNA was extracted from cell lysates 24 hpi. IFN-β expression was determined using quantitative real-time PCR analysis. The mean fold change in the IFN-β expression was calculated using β-actin expression as a housekeeping/reference gene and the 2−ΔΔCt analysis method [55]. The experiment was done in quadruplicates and repeated twice. One out of two experiments is shown. Statistical significance between the indicated groups was determined using a two-sided t test.
Fig 8.
Variability in the PLPs of SARS-related CoVs found in the reservoirs in China and Europe.
(a) Schematic representation of a PLP amino acid sequence alignment (315 amino acids) with European- and Chinese reservoir-associated SARS-related CoV strains which highlights the sequence diversity in the PLP domain. The alignment shows strains (represented as grey lines) with unique mutations (black vertical bars). The ubiquitin-binding surfaces are highlighted in orange (Ub1) and magenta (Ub2), respectively. The proposed localization of Ub1 and Ub2 is depicted in the cartoon below the alignment. SA- and SR-PLP, which were investigated in the presented study, are highlighted in red. Homology modelling for SR-PLP was done to compare the interaction surfaces between SA- and SR-PLP, respectively, and the two ubiquitin molecules of a K48-linked di-ubiquitin. Molecular graphics and analyzes were performed with the UCSF Chimera package [71] with the crystal structure of SARS-CoV PLP in complex with a K48-linked di-ubiquitin (PDB entry 5e6j) as template and the amino acid sequence of SR-PLP as a target. Putative amino acid contact points between SA- and SR-PLP and Ub1 (b) and Ub2 (c) highlight the conformational variability of both PLPs in ubiquitin-binding. The tertiary structure of SA-PLP is shown in dark blue, while the SR-PLP tertiary structure is shown in light blue.