SARS-CoV-2: High-resolution microscopy study in human nasopharyngeal samples

Background The novel coronavirus SARS-CoV-2 is the etiological agent of COVID-19. This virus has become one of the most dangerous in recent times with a very high rate of transmission. At present, several publications show the typical crown-shape of the novel coronavirus grown in cell cultures. However, an integral ultramicroscopy study done directly from clinical specimens has not been published. Methods Nasopharyngeal swabs were collected from two Cuban individuals, one asymptomatic and RT-PCR negative (negative control) and the other from a COVID-19 symptomatic and RT-PCR positive for SARS CoV-2. Samples were treated with an aldehyde solution and processed by Scanning Electron Microscopy (SEM), Confocal Microscopy (CM) and, Atomic Force Microscopy (AFM). Improvement and segmentation of coronavirus images were performed by a novel mathematical image enhancement algorithm. Results The images of the negative control sample showed the characteristic healthy microvilli morphology at the apical region of the nasal epithelial cells. As expected, they do not display virus-like structures. The images of the positive sample showed characteristic coronavirus-like particles and evident destruction of microvilli. In some regions, virions budding through the cell membrane were observed. Microvilli destruction could explain the anosmia reported by some patients. Virus-particles emerging from the cell-surface with a variable size ranging from 80 to 400 nm were observed by scanning electron microscopy (SEM). Viral antigen was identied in the apical cells zone by CM. Conclusions The integral microscopy study showed that SARS-CoV-2 has a similar image to SARS-CoV. The application of several high-resolution microscopy techniques to nasopharyngeal samples awaits future use.


Background
The family Coronaviruses comprises a wide-ranging of viruses that infect many animal species including human beings (1). In December 2019, a new coronavirus disease emerged in China. Rapidly, the disease caused by the SARS CoV-2 extended to the whole world with a record in the number of cases and fatalities (2)(3)(4)(5). SARS-CoV-2 is 96.2% identical to a bat coronavirus at the whole-genome level and it belongs to the species of SARS-CoV (6). Due to the genetic similarity to SARS-CoV, several characteristics have been assumed, but not clari ed. Therefore, more deep investigations with a precise speci c morphological description of this novel virus are necessary (5,7). SARS-CoVs are enveloped viruses with a diameter ranging from 60 to 400nm. The envelope is constituted by the envelope and the spikes (S-protein) proteins. The viral RNA is covered by the nucleocapsid. The overall structure looks similar to other viruses of the Coronaviridae family. The S-protein forms a clovershaped trimer, with three S1 heads and a trimeric S2 stalk. The envelope-anchored spike protein guides coronavirus entry into host cells. This state of the spike on the mature virions is called "prefusion" (1,(8)(9)(10). After viral replication and before the viral release, the spike proteins move from the Golgi apparatus to cell membrane, and the ribonucleoprotein core interact with this S-rich membrane. Not microscopic evidence of this nal stage has previously been shown for SARS-CoV-2, but following its similarity with SARS-CoV, this characteristic has been assumed (8).
The structural details of the SARS-CoV-2 virus are essential towards understanding its historical resembles that of other coronaviruses (mode of infection, mechanism of entry at tissue site of infection, and the replication process in the infected cells) (11,12). High-resolution microscopic studies are essential in identifying the etiological agent of several outbreaks. This technology is a useful tool in guiding subsequent laboratory and epidemiologic investigations mostly in new epidemic emergencies (13). Particularly, in the case of SARS-CoV-2, the SEM images provide fundamental data of the structural aspects of the virus and must be a guiding point in therapeutic developments, for instance, in advanced antiviral drugs and monoclonal antibodies therapies (11).
The application of AFM to viruses-associated to human pathologies might have a signi cant impact on the diagnosis and treatment. Nowadays, scanning probe microscopy is a well-established technique for the rapid visualization of pathogens, including viruses at high resolution (14).

Methods
This study aimed to describe the morphologic characteristics of the SARS-CoV-2 present in human nasopharyngeal specimens using high-resolution microscopy.
Clinical specimens. Two nasopharyngeal swabs tested by Real Time-PCR to SARS CoV-2 were studied.
The rst one, with a negative Real-Time -PCR result, was collected from a contact of a con rmed COVID-19 patient (with a second negative PCR result after 21 days). While the other sample belongs to a con rmed COVID-19 patient and resulted in a positive Real -Time-PCR. These samples were received and tested at the National Reference Laboratory of Viral Respiratory Infections of the Institute of Tropical Medicine for virologic diagnostic. Real Time-PCR was performed as previously described (15).
Inactivation of clinical specimens. 200 µL of the clinical specimens were inactivated for 12 hours in a solution of 25% formaldehyde and 5% glutaraldehyde before microscopy study. Inactivated samples were processed at the Center for Advanced Studies of Cuba by SEM, CM, and AFM.
Scanning Electron Microscopy. Ten microliters of the inactivated clinical specimen were placed in a glasscoverslip and dry-in air oven overnight. Then, the coverslips were xed with 5% glutaraldehyde and dehydrated through a series of increasing concentrations (25%-100%) of ethanol. Coverslips were further Background correction of the images was performed in both, control and samples images using the Olympus Flowview FV-ASW. Software version 3.1 (Olympus, Japan). Improvement and segmentation of coronavirus images by mathematics algorithm (16) . In this section, we will slightly expose some details of the mathematical algorithms used in the enhancement of microscopic images of the novel coronavirus. A second paper will deeply explain all related to these algorithms.
The Gaussian lter was used to diminishing the noise in the original images. In this case, the best performance was obtained using σ = 1.5. The used window size was 3x3. A larger dimensional window caused a loss of information in the microphotographs (I). The h-dome transformation extracts light structures without involving any size or shape criterion. The only parameter (h) related to the height of these structures. In the case of coronavirus S-spikes enhancement, this parameter was of very importance, which will be analyzed deeply in a next publication.

Results
The SEM images showed a general view of SARS-CoV-2 infected human cells. The virus-induced morphological changes demonstrated by the destruction of epithelial cells microvilli in the positive sample (Fig. 1). Uninfected cells show a typical rough surface morphology under the scanning electron microscope and no pseudohyphae are visible either on the cell edge or surfaces (Fig. 1A).
The viral budding through the cell membrane was evident particularly at the edge of these cells. Different stages of the budding process were observed ( Figure 1A-B). some areas show a high density of S-spike budding and pseudohyphae projections appear to be related to an active zone of viral budding. The active budding and release of virions are more evident in the apical membrane of the epithelial cells. The whole viruses are well visualized in the lateral and basal membranes. Figure 2 shows the different stages of the viral budding process and urchin-shape of virions after the improvement and segmentation of coronavirus images by mathematic algorithms.
The viral particle size varied from 86 to 400 nm, but typical virions, observed in the cell-free space, had a size of approximately 80 nm (data no shown). At a higher magni cation, crown-shape is evident in some viral particles, resembling the typical morphology of the coronaviruses (Fig. 1C-1D). S-trimers showed a typical tips clover-shape in budding and mature virions.
Non-lymphocyte-like structures were observed in any of the 50 viewed-elds, suggesting a low immune local in ammatory cell response (data not shown).
The clover-like structures on the virus surface were further con rmed by AFM. The extrusion process of SARS-CoV 2 particles from infect cells is shown in Fig. 3. It can be seen as a group of viruses in budding and others about to come out (arrows), which indicates the presence of different populations. This con rms the obtained results through SEM. Figure 4 shows the results of the immuno uorescence staining by confocal-microscopy. Image A represents the negative control of SARS-CoV-2. Image B, originated from the COVID-19 con rmed patient. The cells of the patient express the SARS-CoV-2 virus antigens, the antibodies present in the hyperimmune serum recognized the antigens. The positive signal of green-clouds was observed due to the presence of the viral antigens (white arrows). Clusters of antigens over the cell surface and in the freecells space detected by CM, like to be as the virus aggrupation found by scanning electron and atomic force microscopy.

Discussion
High-resolution SEM and AFM allow a three-dimensional holistic view of the virus and the infected cell surfaces. High-resolution microscopy is gaining popularity in other areas of the life sciences (14,(17)(18)(19).
Most of these studies were on puri ed macromolecules. However, the AFM has also become a virologic standard in recent years (14).
Images reconstruction provides a very e cient method to extract regional maxima and minima from grayscale SEM-images. Furthermore, the reconstruction technique extends to the determination of maximal structures, which will be call h-domes and h-basins. The h-dome transformation extracts light structure without involving any size or shape criteria. The only parameter (h) is related to the height of these structures. The improved images showed a wild quantity of details of virions and hypothetical cycle life stages of SARS-CoV-2 that were not seen in raw images.
The microvilli destruction observed in the infected epithelial nasal cells could be associated with the anosmia described by some patients. The active budding of the virus particles is observed mostly in the apical zone of the cells. It could be associated with the polarized translocation of immature viral particles from the Golgi apparatus as was described for SARS-CoV (6,11,20). Profuse complete-virus release areas were observed in some cells ( gure 1D). In the same ones and rear of these active zones, a pseudohyphaformations were recorded with the hugest quantity of S-spike inserted in the membrane. This pseudohypha formation was reported during the multiplication of SARS-CoV in Vero cells (14).
One of the advantages of this nding is the possibility to use SEM for the follow up of long-term PCRpositive asymptomatic individuals (Mondeja, Valdes et al., unpublished results). If the virus is actively replicating in these individuals, the presence of cluster spike in the cell membrane, pseudohypha formations, and or viral particle budding, will be noted. The lack of those ndings in the presence of serials PCR -positive samples could be explained by the ampli cation of viral RNA fragments released from infected cells. However, this could be proved by a large-cohort study comparing the SEM and cellassisted culture in this important group of patients.
The non-homogeneous diameter of virus particles was observed by SEM and/or AFM, possibly due to the same unsynchronized replication of the virus. However, in the majority of the cells, virion clusters rather than isolated particles were observed. This active replication could explain the high contagious rate of the virus that associated with an e cient mechanism of viral infection and replication conducting to total cell destruction(4). Billions of virus particles are released for a relatively long time, without much effect in the cell structure(2).
The AFM technique has shown to be a good tool for virus study. Details of the viral structure and the con rmation of the SEM nding was archived. The results of this technique in the SARS-CoV-2 study were similar to the morphological analysis of SARS-CoV previously reported (11,14). One explanation of the difference between SEM and AFM results of the virus-spike observation could be the previous sputtering of the sample with gold before AFM performance. For obvious reasons and the high risk of the work with non-inactivated samples carried SARS-CoV-2, the aldehyde-inactivation and gold coat were used. Fortunately, the inactivation methods do not affect the viral and cell morphology for SEM consequently this procedure can be an option to the microscopy study of clinical samples in unknow or high virulent infectious pathogens.
Confocal studies con rmed the presence of SARS-CoV-2 in clinical specimens and the distribution of the virus into the cell membrane. This is a very useful technique for virus identi cation using a polyclonal hyperimmune serum of convalescence patients. It could be implemented as a diagnostic assay.
It is important to remark that this high-resolution microscopy study was done directly in positive SARS-CoV-2 clinical samples rather than viruses grown in cell cultures as the previous investigation of the topology of virions made in the past for SARS-CoV (11,14). However, our observation is very close to describing the real cellular pathology and damage of SARS-CoV-2 to the respiratory tissue in the patients.
On the other hand, the use of hyperimmune serum of a convalescent patient could be able to introduce some background in the virus detection by confocal microscopy. This could be possible to crossreactivity with another respiratory virus that could be present in the sample. If they use a different cellular receptor to ACE2 (e.g. in uenza A virus), the coinfection is possible, and a false positive signal will be recorded and misunderstand as SARS-CoV-2 (8,(21)(22)(23).
Previous microscopy studies of SARS-CoV-2 were focused on viral infected cell cultures or applied only one of these advanced electron microscopy techniques to the study of clinical specimens (14). Our investigation on the morphology of the novel coronavirus SARS-CoV-2 is based on viruses on clinical samples using high-resolution microscopy (SEM, AFM, Confocal microscopy).
Further investigations should be aimed at the nanostructure of SARS-CoV-2 by High-Resolution SEM and imaging processing to bring us a better understanding of the viral structure in clinical samples. This will enhance the development of new inhibitors and innovative materials to control or destroy in-vivo viruses that could be used in protective clothes and medical devices.

Conclusions
The integral microscopy study showed that SARS-CoV-2 has a similar image to SARS-CoV. The application of several high-resolution microscopy techniques to clinical samples can help to answer important questions its replicative cycle and immunopathogenic mechanism of this novel coronavirus, relevant for the development of new treatments against this disease.