Abstract
The family Coronaviridae is made up of a collection of viruses that cause prevalent diseases in humans and domestic animals (Holmes, 1990). The known human coronaviruses (HCoV-229E and HCoV-OC43) often cause cold-like symptoms that contrast with the recent infections caused by SARS CoV.
In general, the latent periods of coronaviruses can be relatively short in tissue culture (about 6 h) (Sturman & Takemoto, 1972). Infection can be cytocidal for the cells or, in some cases, persistent infection can result depending on the virus strain and cell type (Wege et al., 1982; Sturman & Holmes, 1983; Frana et al., 1985). Cell cultures infected with HCoV-229E were able to produce virus particles over weeks without any expression of cytopathic effects (Chaloner-Larsson & Johnson-Lussenburg, 1981), as only a portion of the cells are infected (Lucas et al., 1978; Holmes & Behnke, 1981; Lamontagne & Dupuy, 1984).
Coronaviruses have strong tissue tropism and will often grow in cells of the natural host species (Fleming et al., 1987, 1988; Sussman et al., 1987). The site of replication is in the cytoplasm of the infected cells (Wilhelmsen et al., 1981). The virus assembles by budding at the Golgi complex (David-Ferreira & Manaker, 1965; Tooze et al., 1984; Tooze & Tooze, 1985). The maturation site appears to be determined by the presence of one of the virus envelope glycoproteins, E1 (Sturman et al., 1980; Holmes et al., 1981a, b, 1984; Sturman & Holmes, 1983). The virus particles migrate through the Golgi complex where glycosylation and processing of the envelope glycoproteins take place. The mature virus particles are then transported in smooth-walled vacuoles to the cell periphery (Sturman & Holmes, 1983; Holmes et al., 1984).
Fusion occurs between the vacuolar walls and the plasma membrane to extrude the progeny virus particles. Extracellular virus particles are usually seen to accumulate in large quantities along the plasma membranes of the infected cells (Oshiro, 1973).
Unlike the known human coronaviruses, SARS CoV infections often result in severe disease. This electron microscopic study aims to determine if there are any unique features during the SARS CoV replication process that can be related to the severity of the disease seen during this outbreak.
Cells and virus.SARS CoV (strain 2003VA2774) was isolated from a SARS patient in Singapore by the Department of Pathology, Singapore General Hospital. The virus was subsequently grown in Vero E6 cells (ATCC #C1008) for 24 h in the Environmental Health Institute, Singapore. The titre of this virus stock was 1x107 p.f.u. ml-1 and was used to infect new cell monolayers for this study.
Electron microscopy.
The replication cycle of SARS CoV was followed at hourly intervals between 1 and 6 h post-infection (p.i.) and subsequently at intervals of 6 h until 30 h p.i. At the appropriate time, cells were fixed with 5 % glutaraldehyde and 2·5 % paraformaldehyde for 4 h. Following the fixation period, the infected monolayer was washed with cold PBS before fixation with 1 % osmium tetroxide. The cells were then dehydrated with a series of ethanol of ascending percentages and embedded in low-viscosity epoxy resin. The cells in resin were polymerized before ultramicrotomy was performed. Ultrathin sections were stained with 2 % uranyl acetate and post-fixed with 2 % lead citrate before viewing under the electron microscope (CM 120 BioTwin, Philips). Images were captured digitally with a Dual View digital camera (Gatan). Ultrastructural studies were performed in the Electron Microscopy Unit, Faculty of Medicine, National University of Singapore.
In a previous study by Ng et al. (2003), it was observed that SARS CoV was internalized and uncoated within 30 min of infection. In addition, there was smooth membrane induction within the smooth-membraned vacuoles containing these nucleocapsids.
In mock-infected cells, the morphology of the various cell organelles was normal (Fig. 1a). During the first hour after infection with SARS CoV (Fig. 1b), however, there were already several vacuoles containing membrane whorls (Fig. 1b, arrows). The other obvious ultrastructural change was the swelling of the Golgi sacs at the perinuclear region (Fig. 1b, arrowheads) and proliferation of the trans-Golgi vesicles (thick white arrow). However, the rough endoplasmic reticulum (Fig. 1b, thin white arrow) remained normal in morphology. Fused virus envelopes at the plasma membrane similar to that of the 2030 min p.i. cells were seen also (Fig. 1b, double arrowheads) (Ng et al., 2003).
|
By 3 h p.i., precursor viruses (50 nm) (Fig. 2a, arrows) were observed in the vicinity of the swollen trans-Golgi complex (arrowhead). The eclipse phase for SARS CoV was found to be only 4 h p.i. (Fig. 2b). A mixture of mature progeny virus particles (100 nm) (Fig. 2b, arrows) and nucleocapsids (50nm) (white arrows) was present in the swollen sacs by this time. Extracellular progeny virus particles (Fig. 2c, arrow) were first observed at 5 h p.i. in about 5 % of the cell population, indicating the end of the latent period. The presence of large vacuoles containing virus precursors and membrane whorls (Fig. 2c, arrowheads) was found consistently in the SARS CoV-infected cells. These membrane whorls, which could be the replication complexes where viral RNA synthesis occurs, were seen early in the replication process in close association with the newly internalized nucleocapsids. The morphology of the rough endoplasmic reticulum remained unaffected by the infection (Fig. 2c, white arrows).
|
SARS CoV grew rapidly in Vero E6 cells and a titre of 107 p.f.u. ml-1 was obtained easily. Nonetheless, it was still surprising to observe that SARS CoV infection was well advanced by 6 h p.i. Mature extracellular virus particles were present in about 2530 % of the cell population (Fig. 3a, arrowheads). This was a dramatic increase in the number of productive cells compared to that seen at 5 h p.i. (5 %), indicating a very rapid assembly process after the latent period. The common sight of swollen trans-Golgi sacs filled with membrane whorls (Fig. 3a, arrows) occupied a large area of the infected cytoplasm.
|
Some cells had spherical SARS CoV cores in the swollen Golgi sacs (Fig. 3b, arrowheads). Spherical cores were also found in the lumen of morphologically normal Golgi sacs (Fig. 3b, inset) and maturing virus particles appeared to be pinching off at the end of the Golgi sacs (Fig. 3b, inset, arrow). A few singly packed virus particles in small vesicles (Fig. 3b, white arrows) were very close to the cell periphery. These are most likely the transport vehicle for the progeny virus particles to the exterior.
The large vacuoles were most likely the very swollen trans-Golgi sacs (Fig. 4a, arrowheads). Doughnut-shaped and spherical nucleocapsids (50 nm) were found in these vacuoles. Membrane layers (Fig. 4a, arrows) associated closely with the virus cores was found in all infected cells. Mature extracellular virus particles with visible spikes on the envelopes were on the cell surface (Fig. 4a, white arrows).
|
Previous studies have reported that coronaviruses assemble by budding into the lumens of the Golgi complex (David-Ferreira & Manaker, 1965; Tooze et al., 1984; Tooze & Tooze, 1985). This mode of assembly was also seen in this study (Fig. 4b). The site of nucleocapsid invagination (Fig. 4b, thin arrows) into the swollen lumens of the Golgi complexes were seen frequently. Doughnut-shaped nucleocapsids were observed to arrange neatly on the cytoplasmic side of the lumen (Fig. 4b, thick arrows) while awaiting budding into the lumen.
In Fig. 4(c), a virus particle within a smaller transport vesicle (arrow) was seen in the process of fusing with the plasma membrane (arrowhead). Many mature virus particles were already exported to the cell surface (Fig. 4c, white arrows).
A low magnification of a typical infected cell at 12 h p.i. is shown in Fig. 5(a). Doughnut-shaped nucleocapsids (Fig. 5a, arrows), spherical cores (white arrows) and mature virus particles (arrowheads) were present in the enlarged vacuoles/swollen Golgi sacs. These different stages of maturation are illustrated more clearly in the inset. Both the 50 nm spherical cores (Fig. 5a, white arrows) and the mature virus particles (80120 nm) (arrowheads) were present. Besides virus-filled sacs at the periphery of the cells, swollen Golgi sacs at the perinuclear region (Fig. 5b, arrows) were also actively involved in producing progeny virus.
|
Another striking observation in SARS CoV-infected cells was the intense electron-dense mitochondria (15 h p.i.) (Fig. 6a, arrowheads). These mitochondria appeared to increase in electron density as infection progressed. The significance of this is not known. The usual swollen sacs were filled with membrane layers (Fig. 6a, arrows) and core/mature virus (white arrows). Crystalline arrays of extracellular virus were present all around the infected cells (Fig. 6a, double arrowheads).
|
By now, more than 50 % of the cell population was actively producing large amounts of virus particles. In Fig. 6(b), mature virus particles (arrows) were observed to pinch off the lumen of the swollen Golgi sacs close to the cell periphery before transportation to cell plasma membrane. The high magnification of the crystalline array of extracellular virus (Fig. 6c) clearly showed the knob-like spikes surrounding the virus particles (arrows).
On some occasions, the extracellular virus particles (arrowheads) were seen to re-enter the infected cells via coated pits (Fig. 6d, e, arrows). At this stage, it is not known if this mode of re-entry into infected cells would yield further productive cycles of replication. Between 15 and 21 h p.i., the percentage of infected cells increased to 70 %. With the longer infection periods, there was a parallel increase in the number of virus-filled vacuoles/swollen Golgi sacs in the infected cytoplasm.
By 24 h p.i., most mature virus particles were seen very close to the plasma membrane awaiting final exit (Fig. 7a). A virus particle (Fig. 7a, arrow) was seen being expelled through a fluke-like channel (arrowhead) created at the plasma membrane. Infected cell filapodia were also active sites of virus exit (Fig. 7b, arrows). Another virus in the small vesicle was seen just about to fuse with the cell plasma membrane (Fig. 7b, arrowhead).
|
Single virus particles pinching off from Golgi sacs into small vesicles for their final exit process (Fig. 8a, arrows) were common sights by 30 h p.i. By this stage, almost 100 % of the cells were infected and the cytoplasm of each cell was filled with very swollen sacs consisting mostly of mature virus particles (Fig. 8b, arrows). By now, very few of these sacs contained the membrane whorls (Fig. 8b, arrowheads).
|
Due probably to the advanced cytopathic effects in these infected cells, the mature virus particles no longer exit in singly packed vesicles. Instead, large bags packed with virus particles fused directly with the plasma membrane (Fig. 8b, double arrowheads). This mode provided fast release of the mature virus particles to the exterior of the dying cells. In this study, SARS CoV was found to replicate extremely well in Vero E6 cells reaching high titres. Despite the fact that it was a human host isolate, it grew well in the monkey kidney cells in vitro.
Previous studies have shown that most coronaviruses have a relatively short latent period of 6 h in tissue culture (Sturman & Takemoto, 1972). However, it may take a few days p.i. to achieve high virus yield (Luby et al., 1999). Positive immunofluorescence was obtained 714 days p.i. of a human enteric coronavirus in J774 cells (a mouse macrophage cell line). Although the infections could be cytocidal, it was not uncommon to establish persistent infections, as not all cells were infected at the same time (Lucas et al., 1978; Chaloner-Larsson & Johnson-Lussenburg, 1981; Holmes & Behnke, 1981; Lamontagne & Dupuy, 1984).
In this study, SARS CoV was found to have a latent period of only 5 h p.i. By this time, extracellular virus particles were seen (Fig. 2c). In addition, there was evidence of major ultrastructural changes in the infected cell cytoplasm. The Golgi sacs were extensively swollen due to the profuse accumulation of the maturing progeny virus particles within the lumens (Figs 14). There was also extensive induction of membrane whorls within the same vacuoles. The membranes evolved very early during infection (within 2030 min p.i.) (Ng et al., 2003) and could be the replication complexes and site of viral RNA synthesis.
Two types of nucleocapsids were observed during the earlier part of infection (up to 12 h p.i.) (Figs 2, 4a, b and 5). The doughnut-shaped structures probably represented the helical nucleocapsids, which then transformed into the 50 nm spherical core particles before final maturation. These two forms, including mature virus particles of 80120 nm in diameter, were often found within the same vacuole.
The reported mode of assembly of coronaviruses is by budding into the Golgi lumens (David-Ferreira & Manaker, 1965; Tooze et al., 1984; Tooze & Tooze, 1985). This was also demonstrated in Fig. 4(b, thin white arrows) where nucleocapsids were observed to bud into the swollen lumen of the Golgi complex. Subsequently, the mature virus particles in these large vacuoles were transported to the plasma membrane in pinched off, small, smooth-membraned vesicles (Figs 4c, 6b, 7 and 8a). This supported the previous observation for other coronaviruses (Oshiro, 1973). However, at late times of infection (30 h p.i.), the release of mature virus particles was speeded up by having large bags containing the virus particles fusing directly with the plasma membrane (Fig. 8b). At this stage, 100 % of the cells were infected with SARS CoV and advance cytopathic effects were evident.
From this study, it is concluded that SARS CoV grows faster than other known human coronaviruses, achieving 107 p.f.u. ml-1 within 24 h p.i. It induces dramatic ultrastructural changes, including the induction of membrane whorls within the same sacs as the virus precursor particles and causes intensive swelling of the Golgi sacs, which become enlarged vacuoles at later stages of infection. Although the ultrastructural changes observed here did not reveal any unique features in the SARS CoV replication cycle compared to other coronaviruses, SARS CoV, nevertheless, replicated quickly in vitro after isolation from the human host. Fifty per cent of the cells were infected within 12 h p.i. and this progressed to 100 % by 24 h p.i. There were large accumulations of both intracellular virus particles (in vacuoles) and extracellular virus particles by 1215 h p.i. The changes in its genomic make-up compared to other known coronaviruses could perhaps enable this new virus to grow equally well both in the human host and in tissue culture (Vero cells).
Electron microscopy work was performed using a grant from the University Research Grant, National University of Singapore (grant no: 182-000-055-112).References
David-Ferreira, J. F. & Manaker, R. A. (1965). An electron microscope study of the development of a mouse hepatitis virus in tissue culture cells. J Cell Biol 24, 5778.
Drosten, C., Gunther, S., Preiser, W. & 23 other authors (2003). Identification of a novel coronavirus in patients with severe acute respiratory syndrome. N Engl J Med 348, 19671976.
Fleming, J. O., Trousdale, M. D., Stohlman, S. A. & Weiner, L. P. (1987). Pathogenic characteristics of neutralization-resistant variants of JHM coronavirus (MHV-4). Adv Exp Med Biol 218, 333342.[Medline]
Fleming, J. O., el-Zaatari, F. A., Gilmore, W., Berne, J. D., Burks, J. S., Stohlman, S. A., Tourtellotte, W. W. & Weiner, L. P. (1988). Antigenic assessment of coronaviruses isolated from patients with multiple sclerosis. Arch Neurol 45, 629633.[Abstract]
Frana, M. F., Behnke, J. N., Sturman, L. S. & Holmes, K. V. (1985). Proteolytic cleavage of the E2 glycoprotein of murine coronavirus: host-dependent differences in proteolytic cleavage and cell fusion. J Virol 56, 912920.
Holmes, K. V. (1990). Coronaviridae and their replication. In Virology, 2nd edition, pp. 841856. Edited by B. N. Fields, D. M. Knipe, P. M. Howley, R. M. Chanock, T. P. Monath, J. L. Melnick, B. Roizman & S. E. Straus. New York: Raven Press.
Holmes, K. V. & Behnke, J. N. (1981). Evolution of a coronavirus during persistent infection in vitro. Adv Exp Med Biol 142, 287299.[Medline]
Holmes, K. V., Doller, E. W. & Behnke, J. N. (1981a). Analysis of the functions of coronavirus glycoproteins by differential inhibition of synthesis with tunicamycin. Adv Exp Med Biol 142, 133142.[Medline]
Holmes, K. V., Doller, E. W. & Sturman, L. S. (1981b). Tunicamycin resistant glycosylation of coronavirus glycoprotein: demonstration of a novel type of viral glycoprotein. Virology 115, 334344.[CrossRef][Medline]
Holmes, K. V., Frana, M. F., Robbins, S. G. & Sturman, L. S. (1984). Coronavirus maturation. Adv Exp Med Biol 173, 3752.[Medline]
Ksiazek, T. G., Erdman, D., Goldsmith, C. S. & 23 other authors (2003). A novel coronavirus associated with severe acute respiratory syndrome. N Engl J Med 348, 19531966.
Lamontagne, L. M. & Dupuy, J. M. (1984). Persistent infection with mouse hepatitis virus 3 in mouse lymphoid cell lines. Infect Immun 44, 716723.
Luby, J. P., Clinton, R. & Kurtz, S. (1999). Adaptation of human enteric coronavirus to growth in cell lines. J Clin Virol 12, 4351.[CrossRef][Medline]
Lucas, A., Coulter, M., Anderson, R., Dales, S. & Flintoff, W. (1978). In vivo and in vitro models of demyelinating diseases. II. Persistence and host-regulated thermosensitivity in cells of neural derivation infected with mouse hepatitis and measles viruses. Virology 88, 325337.[CrossRef][Medline]
Marra, M. A., Jones, S. J., Astell, C. R. & 56 other authors (2003). The genome sequence of the SARS-associated coronavirus. Science 300, 13991404.
Ng, M.-L., Tan, S.-H., See, E.-E., Ooi, E.-E. & Ling, A.-E. (2003). Entry and early events of severe acute respiratory syndrome coronavirus. J Med Virol 71, 323331.[CrossRef][Medline]
Oshiro, L. S. (1973). Coronaviruses. In Ultrastructure of Animal Viruses and Bacteriophages: an Atlas, pp. 331343. Edited by A. J. Dalton & F. Haguenau. Orlando: Academic Press.
Rota, P. A., Oberste, M. S., Monroe, S. S. & 32 other authors (2003). Characterization of a novel coronavirus associated with severe acute respiratory syndrome. Science 300, 13941399.
Ruan, Y. J., Wei, C. L., Ling, A.-E. & 17 other authors (2003). Comparative full length genome sequence analysis of 14 SARS coronavirus isolates and common mutations associated with putative origins of infection. Lancet 361, 17791785; erratum 361, 1832.[Medline]
Sturman, L. S. & Takemoto, K. K. (1972). Enhanced growth of a murine coronavirus in transformed mouse cells. Infect Immun 6, 501507.
Sturman, L. S. & Holmes, K. V. (1983). The molecular biology of coronaviruses. Adv Virus Res 28, 35112.[Medline]
Sturman, L. S., Holmes, K. V. & Behnke, J. (1980). Isolation of coronavirus envelope glycoproteins and interaction with the viral nucleocapsid. J Virol 33, 449462.
Sussman, M. A., Fleming, J. O., Allen, H. & Stohlman, S. A. (1987). Immune mediated clearance of JHM virus from the central nervous system. Adv Exp Med Biol 218, 399410.[Medline]
Tooze, J. & Tooze, S. A. (1985). Infection of AtT20 murine pituitary tumor cells by mouse hepatitis virus strain A59: virus budding is restricted to the Golgi region. Eur J Cell Biol 37, 203212.[Medline]
Tooze, J., Tooze, S. A. & Warren, G. (1984). Replication of coronavirus MHV-A59 in sac- cells: determination of the first site of budding of progeny virions. Eur J Cell Biol 33, 281293.[Medline]
Wege, H., Siddell, S. & ter Meulen, V. (1982). The biology and pathogenesis of coronaviruses. Curr Top Microbiol Immunol 99, 165200.[Medline]
Wilhelmsen, K. C., Leibowitz, J. L., Bond, C. W. & Robb, J. A. (1981). The replication of murine coronaviruses in enucleated cells. Virology 110, 225230.[CrossRef][Medline]
Received 14 July 2003; accepted 19 August 2003.