Use of the informational spectrum methodology for rapid biological analysis of the novel coronavirus 2019-nCoV: prediction of potential receptor, natural reservoir, tropism and therapeutic/vaccine target

Introduction

Fears are mounting worldwide over the cross-border spread of the new strain of coronavirus (denoted as 2019-nCoV) originated in Wuhan, the largest city in central China, after its spread to Thailand and Japan. The newly emerging pathogen belongs to the same virus family as the deadly severe acute respiratory syndrome and Middle East respiratory syndrome coronaviruses (SARS-CoV and MERS-CoV, respectively). The World Health Organization (WHO) has recently published surveillance recommendations for a possible “large epidemic or even pandemic” of the novel coronavirus and it has issued guidelines for hospitals across the world. However, many questions about 2019-nCov remain unanswered: (i) what is the origin and/or natural reservoir of the virus? (ii) is it easily transmitted from human to human? and (iii) what are the potential diagnostic, therapeutic and vaccine targets? Currently, only nucleotide sequences of eight human 2019-nCoV isolates are available without any additional information about biological properties of the virus, beyond the morphology confirmation of the virion using electronic microscopy. This is likely not enough information to answer the important abovementioned questions.

The informational spectrum method (ISM), a virtual spectroscopy method for analysis of proteins, is based on the fundamental electronic properties of amino acids and requires only nucleotide sequence availability to investigate proteins¹. For this reason, ISM was previously used for analysis of novel viruses for which little or no information were available^2–5. Here, the 2019-nCoV was analyzed with ISM to identify its possible origin and natural host, as well as putative therapeutic and vaccine targets.

Methods

Results and discussion

In order to compare informational similarity between 2019-nCoV, SARS-CoV, MERS-CoV and Bat SARS-like CoV, the cross-spectra (CS) of S1 proteins from these viruses were calculated. Figure 1a shows the CS of 2019-nCoV, SARS-CoV and MERS-CoV. These CS contain only one dominant peak corresponding to the frequency F(0.257). Figure 1b displays the CS of S1 proteins from 2019-nCoV and Bat SARS-like CoV. Amplitudes in these latter CS are significantly lower than in those CS presented in Figure 1a. These results show that (i) S1 proteins from 2019-nCoV, SARS-CoV, MERS-CoV and Bat SARS-like CoV encode common information, which is represented with the frequency F(0.257), and (ii) S1 proteins from 2019-nCoV are remarkable more informationally similar with S1 from SARS-CoV and MERS-CoV than with S1 from Bat SARS-like CoV. This suggests that biological properties of 2019-nCoV are apparently more similar to SARS-CoV and MERS-CoV than to Bat SARS-like CoV.

Figure 1. Cross-spectrum (CS) of S1 proteins.

(a) CS of S1 from SARS-CoV, MERS-CoV and 2019-nCoV; (b) CS of Bat SARS-like CoV and 2019-NCov. The abscissa represents the frequencies from the Fourier transform of the sequence of electron-ion interaction potential corresponding to the amino-acid sequence of proteins. The lowest frequency is 0.0 and the highest is 0.5. The ordinate represents the signal-to-noise ratio (the ratio between signal intensity at one particular IS frequency and the main value of the whole spectrum, S/N).

To confirm this conclusion, the ISM-base phylogenetic tree for S1 proteins was calculated (Figure 2). In this calculation the amplitude on the frequency F(0.257) was used as the distance measure. As observed in Figure 2, all analyzed 2019-nCoV S1 amino acid sequences are grouped with SARS-CoV and MERS-CoV and separated from Bat SARS-like CoV. This indicates that 2019-nCoV are more phylogenetically similar to SARS-CoV and MERS-CoV than to Bat SARS-like CoV. This result differs from those obtained with the homology-based phylogenetic analysis, which showed that 2019-CoV are closely related to Bat SARS-like CoV (https://platform.gisaid.org/epi3/frontend#lightbox1296857287).

Figure 2. Informational spectrum method-based phylogenetic tree for S1 proteins from SARS-CoV, MERS-CoV, Bat SARS-like CoV and 2019-nCoV.

The frequency F(0.257) as the distance measure was used.

It has been previously shown that the dominant frequency in the informational spectrum of viral envelope proteins corresponds to interaction between the virus and its receptor^2,3,7,8. The ISM analysis showed that the frequency component F(0.257) is present in the CS of S1 SARS-CoV and its receptor angiotensin converting enzyme 2 (ACE2)⁹, but not in the CS of S1 MERS-CoV and its main receptor dipeptidyl peptidase 4 (DPP4)¹⁰. Of note is that both receptors ACE2 and DPP4 are expressed in airway epithelia. Presence of F(0.257) in the informational spectrum of MERS-CoV (Figure 1) suggests also possible interaction between this virus and the ACE2. The dominant peak on the frequency F(0.257) in the CS of S1 from SARS-CoV and MERS-CoV and ACE2 supports this possibility (Figure 3), although this has not been formally proved for MERS-CoV¹¹.

Figure 3. Cross-spectrum of ACE2 and S1 proteins from SARS-CoV and MERS-CoV.

The abscissa and the ordinate are as described in Figure 1.

As it is shown in Figure 1a, the frequency F(0.257) is also present in the informational spectrum of the 2019-nCoV, suggesting that ACE2 might be the receptor for this novel coronavirus too. Calculation of the CS for S1 protein from the 2019-nCoV and all ACE2 sequences available at the UniProt database revealed that the highest amplitudes on the frequency F(0.257) correspond to ACE2 from civet and chicken. This result indicates that these species can be included as potential candidates for the natural reservoir of the 2019-nCoV. However, it is possible that 2019-nCoV viruses use very different receptors in the natural host(s) and not only the ACE2 as it is the putative case in humans.

Finally, the S1 amino acid sequence from the 2019-nCoV was scanned to look for the domain that gives the highest contribution to the information represented by the frequency F(0.257) (Figure 4a). This analysis revealed domain 266–330 (numbering concerns the maturated protein) is essential for interaction of 2019-nCoV with ACE2. Of note is the striking homology between these domains of S1 proteins from 2019-nCoV and SARS-CoV, but not from MERS-CoV for which ACE2 is not the main receptor (Figure 4b).

Figure 4. Domain of S1 protein which is important for 2019-nCoV/ACE2 interaction.

(a) Mapping of the domain of S1 protein from 2019-nCoV (BetaCoV/Wuhan/IVDC-HB-01/2019) which gives the dominant contribution to the information represented with the frequency F(0.257). (b) Sequence homology between domains of S1 proteins from SARS-CoV and 2019-nCoV with essential contribution to the information corresponding to the frequency F(0.257).

Further, S1 spike proteins from SARS-CoV (Table 1) and 2019-nCoV (Table 2) were compared. The CS of S1 proteins from SARS-CoV (Figure 5a) and 2019-nCoV (Figure 5b) were assessed. Principal information encoded in S1 proteins from SARS-CoV and 2019-nCoV is represented with two different frequencies F(0.222) and F(0.478), respectively. This result indicates some potential difference(s) in the virus-host interaction of these two viruses although they apparently use the same receptor ACE2.

Figure 5. CS of S1 proteins from SARS-CoV and 2019-nCoV and actin proteins.

(a) CS of S1 proteins from human SARS-CoV; (b) CS of S1 proteins from 2019-nCoV; (c) CS of mammalian actin proteins. The abscissa and the ordinate are as described in Figure 1.

To identify the host proteins involved in the attachment and/or internalization of the 2019-nCoV, the UniProt database (https://www.uniprot.org) was screened by ISM for human proteins with the dominant peak on the frequency F(0.478). The list of human proteins that have a dominant peak in IS at the frequency F(0.478) are given in Table 3. According to the IS criterion, these proteins are potential candidate interactors with the 2019-nCoV S1 protein. Further, literature data mining was performed to identify which proteins presented in Table 3 might be involved in the processes of infection with human coronaviruses. This analysis revealed that the actin protein plays an important role in the early entry events during human coronavirus infections¹². Actin proteins were selected as the best candidate interactors for the 2019-nCoV among the host proteins that are characterized with frequency F(0.478). Figure 5c shows that CS of actins from different mammalian species (Table 4) contains the dominant peak on F(0.478), suggesting that these proteins probably encode the conserved information important for their biological function.

The data mining of the PubMed database (www.ncbi.nlm.nih.gov/pubmed/) also showed that actin protein plays an important role in the rapid virus cell-to-cell spread and dissemination of infection¹³. Additionally, the actin filament reorganization is a key step in lung inflammation induced by systemic inflammatory responses caused by infectious agents¹⁴. These findings indicate that interaction between actin proteins and the S1 could be involved in the infection and pathogenesis of 2019-nCoV. In consequence, the possibility to interfere on this interaction might represent a valid hypothesis for development of promising prevention and therapeutic strategies.

Interestingly, further data mining revealed that ibuprofen (FDA approved drug with excellent safety record) attenuates interleukin-1β-induced inflammation as well as actin reorganization¹⁵. Actin was also found to be the primary component by which ibuprofen can bind to the tissue in different organs¹⁶. This suggests that ibuprofen might impact the 2019-nCoV-induced disease by indirect interaction with actin proteins. Previously, ibuprofen was predicted as a candidate entry inhibitor for Ebola virus using the same in silico approach¹⁷, and this prediction was confirmed experimentally at a later time point^18,19. These results prompt the possibility to experimentally test the effects of ibuprofen on 2019-nCoV infection under in vitro and in vivo conditions.

In silico methods are considered very important tools to generate first hypotheses and identify first drug candidates against newly discovered agents, like in the case of 2019-nCoV, especially in the short-term. ISM, a technology based on electronic biology, allowed identifying potential importance of human actin proteins for viral infection/dissemination as well as one FDA approved drug that may have an indirect antiviral activity within weeks of the initial outbreak. However, additional experiments are required to confirm our initial findings.

In conclusion, results of the presented in silico analysis suggest the following: (i) the newly emerging 2019-nCoV is highly related to SARS-CoV and, to a lesser degree, MERS-CoV, and ACE2 is a likely receptor of it; (ii) civets and poultry are potential candidates for the natural reservoir of the 2019-nCoV, (iii) human actin proteins possibly participate in attachment/internalisation of 2019-nCoV, (iv) drugs which interact with actin proteins (e.g. ibuprofen) should be investigated as possible therapeutics for treatment of 2019-nCoV infection, and (v) domain 266-330 of S1 protein from the 2019-nCoV represents promising therapeutic and/or vaccine target. Further research on these issues are needed, including the development of reverse genetics and animal models to study the biology of 2019-nCoV.

Data availability