COVID-19 : VirScan can determine which of more than 1000 different viruses have infected a person

0
173

A tool designed to detect viral history in a drop of blood has gotten an upgrade in the age of COVID-19.

VirScan, a technology that can determine which of more than 1,000 different viruses have infected a person, can now also detect evidence of infection from coronaviruses, including SARS-CoV-2.

In a paper published in Science, investigators from Brigham and Women’s Hospital and Harvard Medical School offer up a treasure trove of details about the antibody response to SARS-CoV-2 and how this response may differ in individuals who go on to have a more severe case of COVID-19.

“This may be the deepest serological analysis of any virus in terms of resolution,” said corresponding author Stephen Elledge, Ph.D., the Gregor Mendel Professor of Genetics at the Brigham and HMS.

“We now understand much, much more about the antibodies generated in response to SARS-CoV-2 and how frequently they are made.

The next question is, what do those antibodies do? We need to identify which antibodies have an inhibitory capacity or which, if any, may promote the virus and actually help it enter into immune cells.”

In their analysis, Elledge and colleagues looked in depth at antibody responses to SARS CoV-2 by using VirScan to analyze blood samples from 232 COVID-19 patients and 190 pre-COVID-19 era controls.

The team identified 800 sites of the virus that the immune system can recognize, known as epitopes.

Not all epitopes are created equal; some may be recognized by neutralizing antibodies, which can elicit a response that eliminates the infection.

However, if the body creates antibodies against other epitopes, it may launch a less effective response, giving the virus an advantage. In some cases, including the related coronavirus that causes SARS, viruses may even be able to benefit from the body’s antibody response, using antibodies to enter cells in a phenomenon known as antibody-dependent enhancement.

In the case of SARS-CoV-2, the team detected a range of antibody frequencies against various epitopes.

Many were public epitopes – regions recognized by the immune systems of large numbers of patients.

One public epitope was recognized by 79 percent of COVID-19 patients.

Others are considered private and recognized by only a few or even one individual.

Ten epitopes were in regions essential for viral entry and are likely recognized by neutralizing antibodies. The team used the most discriminatory epitopes to develop a rapid diagnostic test.

The team’s epitope findings may have important implications for vaccines. If the immune system’s response to public epitopes isn’t found to be protective – or even gives the virus an advantage – vaccines will need to target other regions of the virus to give the immune system a boost.

In addition, the team found that there are several epitopes conserved across coronaviruses, and that the immune system is likely to try to reuse antibodies against them when infected with SARS-CoV-2—a possible explanation for why so many serology tests for COVID-19 produce false positives.

The team further analyzed where and when different antibody responses occurred, finding that patients with severe COVID-19 were more likely to launch a stronger, broader response against SARS-CoV-2, possibly because their initial immune response failed to control the infection early.

Within hospitalized patients, males made more antibodies than females.

The researchers also compared the viral histories of hospitalized and non-hospitalized COVID-19 patients and found that hospitalized patients were much more likely to have had CMV and HSV-1, two common herpes viruses.

However, the researchers note that it is difficult to draw conclusions about causality given that the group of non-hospitalized patients was younger and consisted of a higher percentage of white people and women, a demographic group that generally have lower CMV infection rates.

Elledge envisions their studies as a stepping stone for identifying the most effective antibodies and eliciting them.

“Our paper illuminates the landscape of antibody responses in COVID-19 patients,” said Elledge.

“Next, we need to identify the antibodies that bind these recurrently recognized epitopes to determine whether they are neutralizing antibodies or antibodies that might exacerbate patient outcomes.

This could inform the production of improved diagnostics and vaccines for SARS-CoV-2.”


SARS-CoV-2 is composed of 4 major structural proteins: S (spike), M (membrane), N (nucleocapsid) and E (envelope) [47–49]. The spike protein is responsible for entry by binding the angiotensin-converting enzyme 2 (ACE 2) on the host cell [50, 51].

Accordingly, antibodies that bind the RBD and inhibit the interaction of the S protein with ACE 2 have been the center of attention.

Based on the critical role of the S protein in CoV infection, we focused our work on this protein, dissecting it into two sets of overlapping linear 12mer peptides (two-fold sequence coverage with 6AA overlap between the two sets; i.e 1–12, 7–18, 13–24,…).

The peptide array was prepared by hybridization of PNA-tagged peptide library onto a DNA microarray (Fig 1) [52]. This technology insures a high level of homogeneity across different arrays since individual arrays are prepared from the same library hybridized onto commercial DNA microarrays.

Furthermore, the arrays are designed to have each sequence present 23 times, thus insuring high accuracy by calculating the median of the observed fluorescence of the 23 spots.

Fig 1. Schematic representation of the 200 membered peptide-PNA epitope library design, DNA microarray generation and experimental approach to map the antibody response of COVID-19 patients to the S protein of SARS-CoV-2.

https://doi.org/10.1371/journal.pone.0238089.g001

The S protein of SARS-CoV-2 shares 76% homology with the SARS-CoV-1, [48, 53] and this homology has already been harnessed to predict epitopes based on experimental results from SARS-CoV-1. However, the different infection outcome of SARS-CoV-2 relative to SARS-CoV-1 originates in part from differences in the S protein. SARS-CoV-2 has better affinity to ACE 2 than SARS-CoV-1, yielding more efficient cellular entry [54, 55].

Furthermore, the presence of a furin cleavage site [56–58] in the S protein of SARS-CoV-2 (not present in SARS-CoV-1) coupled to an extended loop at the proteolytic site leads to higher cleavage efficacy thus facilitating its activation for membrane fusion [55, 59–61].

Analysis of 12 different plasma samples from SARS-CoV-2 infected patients and comparison to 6 samples from uninfected patients clearly highlighted a strong response to specific epitopes (Fig 2). The three linear epitopes most abundantly detected (SARS-CoV-2 S protein) were: 655–672, 787–822, and 1147–1158. None of these epitopes was singularly detected in all the positive samples tested, but each is detected in >40% of the positive patients.

The 655–672 epitope is the most abundantly detected in positive samples and corresponds to a peptide that is not part of a secondary structures (Fig 3A and 3B). The corresponding epitope had been also detected in SARS-CoV-1 [8] (89% homology for the 18mer peptide, Fig 4A–4C) and predicted bioinformatically for SARS-CoV-2 [27, 31, 35, 36]; however, it had yet to be observed experimentally. Interestingly, this epitope is just next to the reported S1/S2 cleavage site (Fig 4A–4C, furin/TMPRSS2) [50, 57].

The proteolytic cleavage of the loop 681–685 has been demonstrated to be necessary for the viral entry into the host cell [50]. Moreover, the proteolytic cleavage of the S protein could be a determinant factor for the capacity of the virus to cross species. For example, the S protein of Uganda bats MERS-like CoV is capable of binding human cells, but this is insufficient for entry [62]. However, if a protease (trypsin) is added the protein is cleaved and viral entry occurs.

Furthermore, the most closely related virus to SARS-CoV-2 is RaTG-13 from a bat found in Yunnan province in 2013 which does not contain the furin cleavage sequence [49]. Taken together, this evidence suggest that cleavage of the S protein is a barrier to zoonotic coronavirus transmission. Incorporation of the furin cleavage sites could have been acquired by recombination with another virus leading to human infection. In relation to the furin cleavage site, the pathogenic avian H5N1 contains such a furin cleavage site that leads to higher pathogenicity due to the distribution of furins in multiple tissues [63].

Most recently, high resolution structures analyzing the different conformation of the spike protein prior to and after furin-mediated proteolysis indicates that this proteolysis facilitates the conformational chage required for RBD exposure and binding to surface receptor [64]. We speculate that the binding of an antibody to the epitope 655–672 would sterically block the proteolysis of S1/S2 (vide infra) and should thus be broadly neutralizing, since this proteolysis is critical for infection.

Fig 2. Antibody response from 12 convalescent SARS-CoV-2 patients and 6 uninfected negative controls.

A: Domains of the spike protein (SP = Signal peptide, NTD = N-terminal domain, RBD = Receptor-binding domain, FP = Fusion Peptide, IFP = internal fusion protein, HR1 = Heptad repeat 1, HR2 = Heptad repeat 2) and heat map of antibody binding to the peptide fragments (black background intensity, red 5x background intensity and yellow 10x background intensity). Sample number are indicated on the left of the heat map. B: Fluorescence intensity of antibody binding from the 12 SARS-CoV-2 positive samples (left) and 6 SARS-CoV-2 negative samples (right). The fluorescence intensities are the median of 23 values followed by normalization to the background intensity. The immunodominant regions are highlighted with the corresponding residue numbers (the epitope numbers correspond to the column on top of the dash). See S1 Table in S1 File for quantification and summary of the data.

https://doi.org/10.1371/journal.pone.0238089.g002
Fig 3.

A) Localization of the three selected epitopes on the crystal structure of SARS-CoV-2 Spike protein (PDB ID: 6ZGE): red (epitope 655–672), green (epitope 782-798/811-822) and orange (epitope 1147–1158, the structure is undefined in the PDB). B) Expanded view of the 3 selected epitopes, N-linked glycan shown in purple.

https://doi.org/10.1371/journal.pone.0238089.g003
Fig 4. Selected epitopes localization in relation to the protease cleavage site of the spike protein.

A-B) The 655–672 epitope (red) and the two reported protease cleavage sites S1/S2: site 1 (685–686: blue) and site 2 (695–696: cyano). C) Sequence alignment of the S1/S2 cleavage sites for five different coronaviruses SARSCov2(2019), SARSCov (2003), HCovHKU1, HCovNL63, HCOVOC43 and HCov229E. D-E) The 787–822 epitope (green) and the S2’ cleavage site (815–816: magenta). F) Sequence alignment of the S2’ cleavage site. Figure generated from pdb ID: 6ZGE.

https://doi.org/10.1371/journal.pone.0238089.g004

Another epitope abundantly detected only in healing patients was the 787–822, a peptide segment extending at the periphery of the solvent exposed part of the protein (Fig 3A and 3B). It has also been experimentally observed in the SARS-CoV-1 [9, 13], SARS-CoV-2 [38, 39] and predicted bioinformatically [26, 27, 30, 31, 33, 36]. Interestingly, this epitope includes the S2’ cleavage site of the spike protein (Fig 4D–4F), which has been reported to activate the protein for membrane fusion via extensive irreversible conformational changes [53, 65].

This epitope also includes the fusion peptide (816–833, Fig 4D–4F) [66] which is highly conserved among coronaviruses [67, 68], suggesting a potential pan-coronavirus epitope at this location. It should be noted that a peptide-based fusion inhibitor was shown to exhibit broad inhibitory activity across multiple human CoVs [69] and that antibodies against that region have shown neutralizing activity in SARS-CoV-1 [70]. Taken together, the data support the fact that antibodies inhibiting this proteolytic cleavage should be neutralizing [61, 66].

Finally, the epitope 1147–1158 is found at the C terminus of the spike protein. The structural data reported thus far did not suggest a defined structure for this portion of the S protein. This epitope extends from the helix bundle 1140–1147 (Fig 3A and 3B) and had also been experimentally observed in SARS-CoV-1 [9] and predicted bioinformatically for SARS-CoV-2 [27, 31, 35].

One limitation of epitope mapping with a peptide array is that it is restricted to linear epitopes. Antibodies binding to the RBD have been shown to participate in interactions spanning multiple peptide fragments. Indeed, we did not observe a strong response to linear peptides in the RBD. A control experiment with AI334/CR3022 antibody [25, 71] showed only weak binding to 367–378 peptide sequence of the RBD.

To validate the results observed on the microarray, a peptide (655–672) was synthesized as a biotin conjugate for pull-down and ELISA experiments. The sequence corresponding to 655-672-biotin and a scrambled version of the biotinylated peptide were individually immobilized on agarose streptavidin beads. Beads were exposed to serum from patients that were either positive or negative for that epitope based on the microarray data and subsequently treated with anti-Human-IgG-FITC.

The fluorescence of the beads was quantified by confocal microscopy (Fig 5A). As can be seen in Fig 5B–5E, the beads with 655–672 peptide and positive serum sample showed higher fluorescence than the ones with either negative serum or using the scrambled peptide.

To further probe the binding of 655–672 peptide to antibodies of SARS-CoV-2 positive patients, the same 655–672 biotinylated peptide was used in an ELISA assay (Fig 6A). Three SARS-CoV-2 positive samples showing strong 655–672 signal (Samples 7, 8 and 9) and three SARS-CoV-2 negative samples (Samples 14, 15 and 17) were analyzed showing clear binding to the 655–672 peptide and not to the scrambled version (Fig 6B).

Fig 5.

A) Schematic representation of epitope validation (anti-Spike-655-672 IgG in the SARS-CoV-2 positive patients’ plasma). Microscope images of the beads with: B) Biotin 655–672 with Positive plasma; C) Biotin 655–672 with Negative plasma; D) Biotin-scrambled peptide with Positive plasma. E) FITC fluorescence quantification of B, C and D.

https://doi.org/10.1371/journal.pone.0238089.g005
Fig 6.

A) Schematic representation of the ELISA assay. B) ELISA assay with 3 different 655–672 positive samples with the 655–672 peptide and scrambled peptide and 3 negative samples. Error bars represent triplicate experiments.

https://doi.org/10.1371/journal.pone.0238089.g006

Next, an alanine scan was performed to assess the contribution of individual amino acids to the interaction with the antibodies of two of the COVID positive patients containing antibodies for this epitope (Sample number 1 and 6) at two different dilutions (1 to 100 and 1 to 400).

For this purpose, 17 different peptide-PNA conjugates were synthesized, replacing one amino acid at the time with Ala (Fig 7A) and measuring the intensity of the observed binding on the microarray. This analysis revealed the key role of 5 residues that, if converted to Ala, lead to dramatic loss of activity (amino acids in blue, Fig 7B).

Thus, the key amino acids crucial for binding with the antibodies common for these two plasma samples are H655, Y660, C662, G669 and C671. The binding of the antibody presents in sample 1 also seems to depend on the P665. The remarkable similarities between the two patients is notable considering a polyclonal response.

Fig 7. Alanine scan by hybridization of PNA-peptide conjugates Ala-1 to Ala-17 to a DNA microarray.

A) Peptide sequences of the 17 PNA-peptide conjugates used for the alanine scan, where one amino acid at the time is modified by an alanine. B) Heat map of the interaction of 655–672 positive patient plasma with the different peptides-PNA conjugates hybridized in the DNA array at two different dilutions (1 to 100 and 1 to 400). Heat map represents the normalized fluoresce average of 23 different spots in the array.

https://doi.org/10.1371/journal.pone.0238089.g007

It should be noted that the surface of SARS-CoV-2’s spike is heavily N-glycosylated by host-derived glycans (22 N-glycosylation sites) with a potential role in camouflaging immunogenic protein epitopes [72]. Position N657, which is part of the identified epitope (655–672) adjacent to the furin cleavage site is glycosylated. The alanine scan indicated that this position does not contribute significantly to epitope-antibody interaction.

A recently reported cryo EM structure with the 657-N-linked GlcNAc (6ZGE [64]) visible shows that the glycan point away from the epitope. Furthermore, a proteomic analysis showed that this position is unoccupied by a glycan in 16% of the monomers (i.e. nearly 50% of trimeric spike have an unoccupied N657) [72]. Glycosylation has also been reported at the edge of the other two prominent epitopes (N801, adjacent of epitope 787-798/811-822; N1158 at the edge of epitope 1147–1158).

The fact that antibodies against these epitopes were observed in convalescing patients suggest that the glycans do not effectively shield access to these segments of spike. In the case of N801, the glycan projects away from the identified epitope [64]; in the case of N1158, there is no structural information available to date.

Based on the importance of the furin-mediated proteolysis, we next asked if plasma from patient positive for epitope 655–672 could protect the spike protein against proteolysis. To this end, the spike protein was labeled with Dylight 549 for visualization following SDS-PAGE.

Treatment of labeled spike with furin for 20, 45 and 60 min afforded a progressive formation of two new lower molecular weight bands consistent with a single proteolytic event (one more intense band migrating at above the 70 KDa marker and a lower less intense band at ca. 70 KDa, S1 Fig in S1 File) and showed that 45 min was sufficient for nearly quantitative proteolysis under these conditions.

Addition of plasma prior to the analysis does slightly alter the migration of the spike protein on the gel due to the increased protein loading however, the fluorescence scan still enabled a selective and unambiguous identification of the spike protein on the gel (S2 Fig in S1 File). Performing the proteolytic experiment with furin in the presence of plasma from a patient positive for epitope 655–672 (sample 12, Fig 2) showed a complete protection against proteolysis while plasma from a patient negative for this epitope (sample 10) did not (Fig 8).

The comparison with a second sample from a patient positive for this epitope (sample 1) and the plasma from a healthy individual (sample 14) showed the same result (protection against proteolysis from sample 1 but not 14, S3 Fig in S1 File).

These experiments support the fact that antibodies binding to epitope 655–672 are protective against furin-proteolysis of spike. This protection could be highly relevant in mitigating ADE by preventing viral entry irrespectively of antibody-mediated cellular interactions.

Fig 8. Inhibition of furin-mediated proteolysis of spike.

Fluorescent scan of a SDS-PAGE with Dylight 549-labeled spike protein (lane 2) and treated with furin (lane 4). The same experiment was performed with the addition of plasma from patient 10 (negative for the epitope adjacent to the cleavage site, lane 5) and patient 12 (positive for the epitope adjacent to the cleavage site (lane 3). Lane 1 is a molecule weight marker.

https://doi.org/10.1371/journal.pone.0238089.g008

Conclusion
We have developed a peptide array for the epitope mapping of the spike protein of SARS-CoV-2. Using this array to profile healing plasma of twelve laboratory confirmed COVID-19 patients and six negative controls we have discovered three immunodominant linear regions, each present in >40% of COVID-19 patient (epitope 655–672 in 66%; epitope 782-798/811-822 in 40% and epitope 1147–1158 in 58%).

Two of these epitopes correspond to key proteolytic sites on the spike protein (655–672: S1/S2 and 782-798/811-822: S2’) which have been shown to be crucial for viral entry and play an important role in virus evolution and infection. We show biochemical evidence that serum positive for 655–672 epitope inhibits proteolysis of spike by furin.

The fact that antibodies binding adjacent to the protease cleavage sites were identified from COVID-19 patients raises the possibility that other mechanism than blocking the RBD-ACE2 interaction could be harnessed for neutralization and might mitigate antibody-dependent enhancement of viral entry. Full characterization of these antibodies is necessary, and efforts on this direction are on their way.

REFERENCE LINK : https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0238089


More information: “Viral epitope profiling of COVID-19 patients reveals cross-reactivity and correlates of severity,” Science (2020). science.sciencemag.org/cgi/doi … 1126/science.abd4250

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.