Cleavage of a Neuroinvasive Human Respiratory Virus Spike Glycoprotein by Proprotein Convertases Modulates Neurovirulence and Virus Spread within the Central Nervous System

Abstract

Human coronaviruses (HCoV) are respiratory pathogens that may be associated with the development of neurological diseases, in view of their neuroinvasive and neurotropic properties. The viral spike (S) glycoprotein is a major virulence factor for several coronavirus species, including the OC43 strain of HCoV (HCoV-OC43). In an attempt to study the role of this protein in virus spread within the central nervous system (CNS) and neurovirulence, as well as to identify amino acid residues important for such functions, we compared the sequence of the S gene found in the laboratory reference strain HCoV-OC43 ATCC VR-759 to S sequences of viruses detected in clinical isolates from the human respiratory tract. We identified one predominant mutation at amino acid 758 (from RRSR↓ G758 to RRSR↓R758), which introduces a putative furin-like cleavage (↓) site. Using a molecular cDNA infectious clone to generate a corresponding recombinant virus, we show for the first time that such point mutation in the HCoV-OC43 S glycoprotein creates a functional cleavage site between the S1 and S2 portions of the S protein. While the corresponding recombinant virus retained its neuroinvasive properties, this mutation led to decreased neurovirulence while potentially modifying the mode of virus spread, likely leading to a limited dissemination within the CNS. Taken together, these results are consistent with the adaptation of HCoV-OC43 to the CNS environment, resulting from the selection of quasi-species harboring mutations that lead to amino acid changes in viral genes, like the S gene in HCoV-OC43, which may contribute to a more efficient establishment of a less pathogenic but persistent CNS infection. This adaptative mechanism could potentially be associated with human encephalitis or other neurological degenerative pathologies.

Fig 1. Both variants rOC/ATCC and rOC/S_G758R are neuroinvasive and neurovirulent in 10 day-old BALB/c mice infected by the intranasal route.

10-day old BALB/c mice received 10^3.25 TCID₅₀/10μL of rOC/ATCC or rOC/S_G758R or PBS by the IN route. (A) Survival curves of mice in percentage (%) during 21 days post-infection. Difference between the two variants was significant (** P≤0.01) (B) Surviving BALB/c mice were weighed every 2 days after infection during 21 dpi to estimate weight variations. The weight variation was expressed in %, compared to day 0, which was set at 100%. Production of infectious viral particles was measured in brains (C) and spinal cords (D) every 2 days for 21 dpi. LOD represents the Limit of Detection of infectious viral particles. Results shown are the mean values (with standard deviations) of three independent experiments.

Fig 2. A delay in viral spread is observed in brain of rOC/S_G758R -infected mice compared to rOC/ATCC after intranasal inoculation in 10 day-old BALB/c mice.

Histological examination of virus spread within the brain of 10-day old BALB/c mice infected with 10^3.25 TCID₅₀/10μL of rOC/ATCC or rOC/S_G758R, or PBS by the IN route. (A) Detection of viral antigens in the olfactory bulb of infected mice at 5 and 7 dpi. (B) Detection of viral antigens in the hippocampus of infected mice at 7 and 9 dpi. Magnification 40x.

Fig 3. A decreased neurovirulence is observed for rOC/S_G758R variant in 21 day-old BALB/c female mice infected by the intracerebral route.

21 day-old BALB/c mice received 10^2.5 TCID₅₀/10μL of rOC/ATCC or rOC/S_G758R or PBS by the IC route. (A) Survival curves of mice in % during 21 day post-infection. Difference between the two virus variants was significant (* P≤0.05). (B) BALB/c mice were weighed every 2 days during 21 dpi to estimate weight variations, which were expressed in %, compared to day 0 (100%). Differences were significant (*** P≤0.001) when the three conditions (rOC/ATCC, rOC/S_G758R or PBS) were compared between 9 and 11 dpi. Evaluation of the clinical scores (percentage of mice at each level of the scale) of mice infected by rOC/ATCC (C) or rOC/S_G758R (D) based on neurological symptoms described in clinical score scale between level 0 and 3 (see Materials and Methods section). Production of infectious viral particles was measured in brains (E) and spinal cords (F) every 2 days for 21 dpi. LOD represents the Limit of Detection of infectious viral particles. Results shown are the mean values (with standard deviations) of three independent experiments.

Fig 4. A delay in viral spread is observed in brains of rOC/S_G758R -infected mice compared to rOC/ATCC after intracerebral infection in 21 day-old BALB/c female mice.

Histological examination of virus spread within the brain. 21 day-old BALB/c mice received 10^2.5TCID₅₀/10μL of rOC/ATCC or rOC/S_G758R, or PBS by the IC route. Detection of viral antigens in the olfactory bulb (A) or in the hippocampus (B) of infected mice at 5 and 7 dpi at magnitude X40. Black arrows indicate viral particles staining for the S protein.

Fig 5. Mutation in the spike glycoprotein of mutant virus delays viral spreading compared to the reference strain in mixed primary CNS cultures from BALB/c mice.

Mixed primary cultures from BALB/c mice brain were infected with rOC/ATCC or rOC/S_G758R at MOI 0.03. Viral spread was evaluated at 8, 24 and 48 hpi. Neurons were stained in green with a mAb against microtubule-associated protein 2 (MAP2) antibody and the S viral glycoprotein in red, was detected with a rabbit antiserum. Results are representative of three independent experiments. Magnification 200x.

Fig 6. The S glycoprotein harboring a predominant point mutation found in clinical isolates was cleaved more efficiently in supernatant of CNS cells.

Mixed primary cultures from BALB/c mice brain were infected with rOC/ATCC or rOC/S_G758R at MOI 0.03. Kinetics of infectious virus production in cell-associated (A) and in cell culture supernatant (free virus) (B) was performed. Release of free virus in the supernatant was significantly higher for rOC/S_G758R compared to rOC/ATCC (* P≤0.05). (C) Western blot analysis of whole cell lysates (C) or cell culture supernatant (D) (10 μg of proteins) revealed the presence of the uncleaved form of the S glycoprotein (180 kDa), and of a cleaved form at around 100 kDa (S1/S2). Results shown are the mean values (with standard deviations) of three independent experiments.

Fig 7. Cleavage of S glycoprotein is also observed in human LA-N-5 cells for mutant virus.

The differentiated human neuroblastoma cell line (LA-N-5) was infected with rOC/ATCC or rOC/S_G758R at MOI 0.1. Proteins in association with cell or in supernatant were extracted at 24 and 48 hpi, and kinetics of viral replication was evaluated over a period of 48 hpi for (A) cell-associated virus or (B) free virus (supernatant). Titers of cell-associated and free virus were significantly higher for rOC/S_G758R compared to rOC/ATCC (* P≤0.05 and ** P≤0.01). Western blot analysis of whole cell lysates (C) or cell culture supernatant (D) (10 μg of proteins) revealed the presence of the uncleaved form of the S glycoprotein (180 kDa), and of a cleaved form at around 100 kDa (S1/S2). Results shown are the mean values (with standard deviations) of three independent experiments.

Fig 8. Proprotein convertase as a potential player in S glycoprotein cleavage.

(A-B) Differentiated human neuroblastoma cell line (LA-N-5) was incubated before and after infection with different concentration of furin-like inhibitor (dec-RVLR-cmk; 0, 5, 10μM). Infection was performed with rOC/ATCC or rOC/S_G758R at MOI 0.1. Proteins in supernatant were extracted at 48 hpi and analyzed by Western blot (50 μg of proteins) of supernatant from LA-N-5 cells infected by reference virus (A) or mutant virus (B). Western blot was directed against viral S glycoprotein. (C-D) Incubation of synthetic peptide containing the sequence of the putative cleavage site of reference virus (C) and mutant virus (D) with different recombinant proprotein convertases in vitro. The % of cleaved peptide was quantified over time.

Fig 9. Morphology of viral crown peplomers is dependent on cleavage of the S glycoprotein but does not affect virus infectivity.

Observation of virus crown peplomers was made by Transmission Electron Microscopy (TEM). Mixed primary cultures from BALB/c mice brain were infected with rOC/ATCC or rOC/S_G758R at MOI 0.03 and supernatants were harvested at 48 hpi. After negative staining by PTA, viral particles were observed. (A) Left panel represents the “long S” morphology at magnification 50,000x and right panel indicates spike size in μm on the same viral particle. (B) Left panel shows the “short S” morphology at magnification 50,000x and right panel indicates spike size in μm on the same viral particle. Black arrows represent different sizes of spike glycoprotein, and white arrows show the viral hemagglutinin-esterase (HE) protein peplomers. (C) Mean size of spike glycoprotein (in nm) was evaluated for long S and short S types of viral particles. Measurements between 20 and 26 nm in length were considered as long S and those between 14 nm and 20 nm in length were considered as short S spikes on viral particles. The difference between long S and short S viral particles was significant (*** P≤0.001). (D) Quantification of 650 long S and shortS viral particles of rOC/ATCC or rOC/S_G758R shows a significant difference between both viruses (*** P≤0.001). (E) Infectivity assay between viruses: quantification of viral RNA (absolute quantity in RNA copy) over the number of infectious particles in viral stocks. (F) The amount of viral RNA copy found associated with LA-N-5 cells during the early steps of infection at 0.5, 2, 4, 8, 16 hpi. Results are shown in absolute number of viral RNA copy.