Identification of the Mechanisms Causing Reversion to Virulence in an Attenuated SARS-CoV for the Design of a Genetically Stable Vaccine

Abstract

A SARS-CoV lacking the full-length E gene (SARS-CoV-∆E) was attenuated and an effective vaccine. Here, we show that this mutant virus regained fitness after serial passages in cell culture or in vivo, resulting in the partial duplication of the membrane gene or in the insertion of a new sequence in gene 8a, respectively. The chimeric proteins generated in cell culture increased virus fitness in vitro but remained attenuated in mice. In contrast, during SARS-CoV-∆E passage in mice, the virus incorporated a mutated variant of 8a protein, resulting in reversion to a virulent phenotype. When the full-length E protein was deleted or its PDZ-binding motif (PBM) was mutated, the revertant viruses either incorporated a novel chimeric protein with a PBM or restored the sequence of the PBM on the E protein, respectively. Similarly, after passage in mice, SARS-CoV-∆E protein 8a mutated, to now encode a PBM, and also regained virulence. These data indicated that the virus requires a PBM on a transmembrane protein to compensate for removal of this motif from the E protein. To increase the genetic stability of the vaccine candidate, we introduced small attenuating deletions in E gene that did not affect the endogenous PBM, preventing the incorporation of novel chimeric proteins in the virus genome. In addition, to increase vaccine biosafety, we introduced additional attenuating mutations into the nsp1 protein. Deletions in the carboxy-terminal region of nsp1 protein led to higher host interferon responses and virus attenuation. Recombinant viruses including attenuating mutations in E and nsp1 genes maintained their attenuation after passage in vitro and in vivo. Further, these viruses fully protected mice against challenge with the lethal parental virus, and are therefore safe and stable vaccine candidates for protection against SARS-CoV.

Fig 1. Generation of chimeric membrane genes after 16 serial passages of SARS-CoV lacking E protein in cell culture.

(A) Representation of membrane chimeric genes generated after passage of SARS-CoV-∆E (∆E) and SARS-CoV∆[E,6-9b] (∆[E,6-9b] 16 times in Vero E6 cells and ∆E virus in DBT-mACE2 cells. Top, ∆E p1 represents the genomic sequence of viruses lacking E protein at passage 1. Grey boxes indicate E gene with a partial deletion highlighted in light grey (∆). Transcription-regulating sequences (TRSs) of the different genes are shown in blue boxes and the membrane gene (M) is shown in red. Chimeric membrane genes generated after 16 serial passages (p16) are formed by a partial duplication of membrane gene fused to part of SARS-CoV leader sequence (green boxes). TMD1, TMD2 and TMD3 represent the three different transmembrane domains contained within the first part of membrane gene. ATG and STOP represent the start and stop codon of potential proteins, respectively. EITL, RSVL and SLVL represent the last four amino acids of chimeric proteins that form PDZ-binding motifs. (B) Amino acid sequences corresponding to chimeric proteins generated after SARS-CoV-∆E passage in cell culture. New amino acids are shown in red. Presence of native membrane and chimeric membrane proteins was analyzed by Western blot at 24 hpi using two polyclonal antibodies generated to recognize either all membrane proteins, chimeric or not (C) or a unique sequence in the chimeric protein generated after SARS-CoV-∆E passage in DBT-mACE2 cells (D).

Fig 2. Growth kinetics of SARS-CoV-∆E viruses containing chimeric proteins.

Subconfluent monolayers of Vero E6 and DBT-mACE2 cells were infected with wt, ΔE, MCH-Vero and MCH-DBT viruses at a moi of 0.001. Culture supernatants collected at 4, 24, 48 and 72 hpi were titrated by plaque assay.

Fig 3. Virulence and viral growth of SARS-CoV-∆E after serial passage in cell culture.

16-week-old BALB/c mice were intranasally inoculated with 100,000 pfu of wt, ΔE, MCH-Vero and MCH-DBT viruses. (A) Weight loss and survival were monitored for 10 days. Data represent two independent experiments with 5 mice per group. (B) Viral titer in lungs was determined at 2 and 4 days post infection (n = 3, each day). Error bars represent standard deviations. (C) Lung tissue sections from mice infected with the different recombinant viruses were prepared and stained with hematoxylin and eosin at 2 and 4 dpi. Three independent mice per group were analyzed. Original magnification was 20x and representative images are shown.

Fig 4. Identification of E protein domain involved in chimeric protein generation in cell culture.

(A) The SARS-CoV genome is shown at the top, and the expanded region shows the E protein sequence and its different regions. Grey boxes indicate a set of recombinant SARS-CoVs including mutations or deletions in different regions of E protein. Mutations and deletions are shown in red. Recombinant viruses were passaged 16 times in Vero E6 cells and the presence of chimeric genes and E protein PBMs was analyzed by sequencing, using specific primers. (B) Schematic indicating the presence and sequence (green) or absence (red) of PDZ-binding motifs at the end of the different E proteins (E PBM) at passages 0 (p0) and 16 (p16). Presence or absence of chimeric proteins after 16 serial passages in cell culture are indicated by (–) and (+), respectively. (C) Representation of recombinant viruses generated of SARS-CoV-∆E-MCH-EPBM (MCH-EPBM) and SARS-CoV∆E-3aCH-3aPBM (3aCH-3aPBM). Top, ∆E represents the genomic sequence of viruses lacking E protein. Grey boxes indicate E gene with a partial deletion highlighted in light grey (∆). Transcription-regulating sequences (TRSs) of the different genes are shown in blue boxes and the membrane gene (M) is shown in red. TMD1 represents the transmembrane domain of the indicated proteins. DLLV and SVPL represent the last four amino acids forming different PDZ-binding motifs.

Fig 5. Stability of SARS-CoV-∆E after serial passage in mice.

A recombinant virus lacking the E protein (SARS-CoV-∆E) was passaged in BALB/c mice using an initial moi of 100,000 pfu. After 10 serial passages the last third of the genome was sequenced using specific primers. (A) Sequence corresponding to native 8a protein (8a) and to mutated protein generated after passage in mice (8a-dup). New incorporated amino acids are highlighted in light blue. (B) Predicted structures for 8a protein at passage 0 (8a) and 10 (8a-dup) in mice. Presence of new incorporated amino acids in 8a-dup and PBM are highlighted in light blue and red, respectively.

Fig 6. Virulence and viral growth of SARS-CoV-∆E after serial passage in mice.

16-week-old BALB/c mice were intranasally inoculated with 100,000 pfu of wt, ΔE and ∆E-8a-dup viruses. (A) Weight loss and survival were monitored for 10 days. Data represent two independent experiments with 5 mice per group. (B) Viral titer in lungs was determined at 2 and 4 days post infection (n = 3, each day). Error bars represent standard deviations. Statistically significant data are indicated with one asterisk (P < 0.05). (C) Lung tissue sections from mice infected with the different recombinant viruses were prepared and stained with hematoxylin and eosin at 2 and 4 dpi. Three independent mice per group were analyzed. Original magnification was 20x and representative images are shown. (D) 16-week-old BALB/c mice were intranasally inoculated with 100,000 pfu of wt or ΔE-8a-dup viruses and were monitored daily for weight loss and survival.

Fig 7. Analysis of p38 MAPK activation and inflammatory cytokines expression during infection with recombinant SARS-CoV-∆E virus after passage in mice.

(A) Vero E6 cells were mock-infected or infected with the wt, ∆E and ∆E-8a-dup viruses at a moi of 0.3 and the presence of active phosphorylated (p-p38) and total (p38) p38 MAPK was detected by Western blot analysis at 24 hpi. Actin was used as control. Lines 1, 2 and 3 indicated three independent experiments analyzed in each case. (B) Phospho and total p38 MAPK amounts were quantified by densitometric analysis. The graph shows the phosphorylated p38/total p38 MAPK ratio in Vero E6 cells infected with wt, ΔE and ∆E-8a-dup viruses at 24 hpi. Statistically significant data compared to mock-infected cells are indicated with one (P < 0.05) or two (P < 0.01) asterisks. (C) Expression of inflammatory cytokines in lungs of infected mice evaluated by RT-qPCR at 2 dpi. Three independent experiments were analyzed with similar results in all cases. Error bars represent the means of three experiments analyzed for each condition. (D) Vero E6 cells were mock-infected or infected with the wt or ∆E-8a-dup viruses at a moi of 0.3 and the presence of active phosphorylated (p-p38) and total (p38) p38 MAPK was detected by Western blot analysis at 24 hpi. Actin was used as control. Lines 1, 2 and 3 indicated three independent experiments analyzed in each case. (E) Phospho and total p38 MAPK amounts were quantified by densitometric analysis. The graph shows the phosphorylated p38/total p38 MAPK ratio in Vero E6 cells infected with wt and ∆E-8a-dup viruses at 24 hpi. Statistically significant data compared to mock-infected cells are indicated with one (P < 0.05) asterisk. (F) Expression of inflammatory cytokines in lungs of infected mice evaluated by RT-qPCR at 2 dpi. Three independent experiments were analyzed with similar results in all cases. Error bars represent the means of three experiments analyzed for each condition.

Fig 8. Generation and growth kinetics of SARS-CoVs lacking small regions within the nsp1 protein.

(A) Sequence alignment of nsp1 from SARS-CoV and MHV. The conserved amino acids in these proteins are indicated in red. Gray boxes represent the amino acids deleted within the nsp1 protein in ΔA, ΔB, ΔC and ΔD viruses. (B) Mutant viruses growth kinetics. Subconfluent monolayers of Vero E6 and DBT-mACE2 cells were infected with wt or SARS-CoV-nsp1* viruses at a moi of 0.001. At different times post infection, virus titers were determined by plaque assay on Vero E6 cells. Error bars represent standard deviations of the mean of three independent experiments.

Fig 9. Virulence and viral growth of SARS-CoV-nsp1* mutants in vivo.

BALB/c mice were intranasally infected with 100,000 pfu of wt, and SARS-CoV-nsp1-ΔA, -ΔB, -ΔC and -ΔD viruses (5 mice per group). (A) Animals were monitored daily for weight loss and survival. (B) Viral titers in lungs were determined at 2 and 4 days post infection (3 mice per group and time point). Error bars represent the standard deviations from three independent mice in each case. (C) Lung tissue sections from mice infected with the different recombinant viruses were prepared and stained with hematoxylin and eosin at 2 and 4 dpi. Three independent mice per group were analyzed. Original magnification was 20x and representative images are shown.

Fig 10. Expression of IFN-β and ISGs in SARS-CoV-nsp1* attenuated mutants infected cells.

DBT-mACE2 cells were mock-infected or infected at a moi of 0.125 with SARS-CoV-nsp1* attenuated mutants or SARS-CoV (wt). Cellular RNAs were extracted at 48 hpi and the expression of the indicated genes was determined by RT-qPCR. In each case, the corresponding mRNA expression levels were plotted as fold change relative to expression levels in uninfected cells. (A) IFN-β. (B) ISGs such as IRF1, DDX58 and STAT1, and 18S rRNA as a control. Error bars represent standard deviations of the means from three experiments. Statistically significant data compared to uninfected cells are indicated with one (P < 0.05) or two (P < 0.01) asterisks.

Fig 11. Protection conferred by immunization with SARS-CoV-nsp1* mutants.

16-week old BALB/c mice were mock-immunized or immunized with 6000 pfu of SARS-CoV-ΔC and -ΔD mutants, and challenged at day 21 post-immunization with 100,000 pfu of SARS-CoV-wt virus (5 mice per group). Weight loss (A) and survival (B) were recorded daily.

Fig 12. Viral growth of SARS-CoV-∆3 and virulence after serial passage in mice.

(A) Subconfluent monolayers of Vero E6 and DBT-mACE2 cells were infected with wt, ΔE and Δ3 viruses at a moi of 0.001. Culture supernatants collected at 4, 24, 48 and 72 hpi were titrated by plaque assay. (B) 16-week-old BALB/c mice were intranasally inoculated with 100,000 pfu of wt, ΔE and Δ3 viruses. Viral titer in lungs was determined at 2 and 4 days post infection (n = 3, each day). Error bars represent standard deviations. (C) Weight loss and survival were monitored for 10 days (5 mice per group).

Fig 13. Generation and growth kinetics of SARS-CoV mutants with deletions in both nsp1 and E genes.

(A) SARS-CoV genome is shown in the top, and the expanded region shows the nsp1 and E genes. White boxes represent the amino acids stretches deleted in both proteins in each virus. (B) Mutant viruses growth kinetics. Subconfluent monolayers of Vero E6 and DBT-mACE2 cells were infected with wt, SARS-CoV-nsp1ΔD-ΔE and SARS-CoV-nsp1ΔD-EΔ3 at passage 1 and 10 (-p1 and -p10C) viruses at a moi of 0.001. At different times post infection, virus titers were determined by plaque assay on Vero E6 cells. Error bars represent standard deviations of the mean using data from three independent experiments.

Fig 14. Virulence and virus growth of SARS-CoV including two safety guards.

BALB/c mice were intranasally infected with 100,000 pfu of wt, SARS-CoV-nsp1ΔD-ΔE and SARS-CoV-nsp1ΔD-EΔ3 at passage 1 and 10 (-p1 and -p10C) viruses (5 mice per group). (A) Animals were monitored daily for weight loss and survival. (B) Viral titers in lungs were determined at 2 and 4 days post infection (3 mice per group and time). Error bars represent the standard deviations from three independent mice in each case. (C) Lung tissue sections from mice infected with the different recombinant viruses were prepared and stained with hematoxylin and eosin at 2 and 4 dpi. Three independent mice per group were analyzed. Original magnification was 20x and representative images are shown.

Fig 15. Virulence and virus growth of SARS-CoV with two safety guards after serial passage in mice.

(A) Subconfluent monolayers of Vero E6 and DBT-mACE2 cells were infected with wt and SARS-CoV-nsp1ΔD-EΔ3-p10M at a moi of 0.001. At different times post infection, virus titers were determined by plaque assay on Vero E6 cells. Error bars represent standard deviations of the mean using data from three independent experiments. (B) BALB/c mice were intranasally infected with 100,000 pfu of wt virus, or SARS-CoV-nsp1ΔD-EΔ3-p1 and -p10M viruses (5 mice per group). Animals were monitored daily for weight loss and survival. (C) Viral titers in lungs were determined at 2 and 4 dpi (3 mice per group and time). Error bars represent the standard deviations from three independent mice in each case. (D) BALB/c mice were intranasally infected with 100,000 pfu of wt or SARS-CoV-nsp1ΔD-EΔ3-p20M viruses (5 mice per group). Animals were monitored daily for weight loss and survival.

Fig 16. Protection conferred by immunization with SARS-CoV-double mutants.

BALB/c mice (16-week-old) were mock-immunized or immunized with 6000 pfu of SARS-CoV-nsp1ΔD-EΔ3 viruses at passage 1 and 10 (-p1, -p10C and -p10M), and challenged at day 21 post-immunization with 100,000 pfu of wt virus (5 mice per group). Weight loss (A) and survival (B) were recorded daily.