The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is thought to have a zoonotic origin, perhaps originating in a bat coronavirus with an unknown intermediate host. It is not only very infectious but has been found to infect a wide range of mammals, including dogs, cats, ferrets, hamsters, and minks, besides non-human primates. This range may be still higher, claims a recent preprint that appeared in January 2021 on the bioRxiv* server.
It is essential to identify all hosts to understand the sources of infection to human beings as well as the possibility of reverse transmission from humans to animals. The issue with such transmission to animals is that the virus may harbor in these non-human hosts and evolve new mutations. The mutant strains may prove to be capable of reinfecting humans, allowing the cycle to persist.
Sequence homology not predictive of host susceptibility
Earlier, researchers attempted to identify possible animal hosts for SARS-CoV-2 by assessing the degree to which the human angiotensin-converting enzyme 2 (ACE2) receptor for the virus shares its genetic, amino acid sequence, and structural similarity at the binding interface, with ACE2 in other species. This method has not proved uniformly successful in predicting the host range, suggesting that different ACE2 sequences may still allow viral binding. This, in turn, indicates that there may be many other unknown hosts among mammals.
Predictable and unpredictable trajectories
Again, unlike the earlier SARS-CoV, the receptor-binding domain (RBD) of the novel coronavirus binds to multiple residues within the binding domain of the ACE2 receptor, but interactions between these sites remain unclear. One way to explain how the viral RBD engages with different ACE2 sequences is intramolecular epistasis. This is a form of context-driven mutation, where an amino acid can have different functional effects based on the residues present at other sites. In this way, this type of epistasis defines the type of combinations possible to a given amino acid concerning a specific function so that evolution becomes predictable.
Intramolecular epistasis also allows new types of compensatory interactions to arise after a loss-of-function mutation. This results in the production of convergent amino acid sequences for the same molecular structure or function. Thus, while the number of interactions is limited for a particular function, making for predictable evolution, the potential for new interactions allows for unpredictability. This hypothesis could explain how the homology of ACE2 sequences is not the most reliable method to predict if the virus can infect another species.
Natural variation in SARS-CoV-2 binding is mirrored by diversity in ACE2 enzyme activities. (A-D) Flow cytometry was used to quantify SARS-CoV-1 and CoV-2 RBD-Fc association with human cells (HEK293T) expressing ACE2-eGFP orthologs from various mammalian species (mean luorescence intensity; MFI (19)). (E) SARS-CoV-2 S pseudovirus infection of HEK293T cells expressing ACE2 orthologs was quantified using a luciferase reporter system. (F) The carboxypeptidase activity of ACE2 was quantified using a fluorometric peptide incubated with solubilized HEK293T cells transfected with ACE2 orthologs. (G) Hydrolysis rate of ACE2 orthologs (Fluorescence units/minute).
Protective role of ACE2
ACE2 has a role in the functioning of the kidneys and cardiovascular system via the renin-angiotensin-aldosterone system (RAAS), along with ACE. The latter enzyme gives rise to a signaling peptide called angiotensin-II (Ang-II), which causes the blood vessels to constrict, triggers inflammation, and leads to fibrosis, all through the Ang-II/AT1R axis. ACE2 inhibits these effects and instead converts Ang-II to Ang-(1-7). This produces the opposite effects, namely, vasodilation, anti-inflammatory, and anti-fibrotic activity.
ACE2 is thus protective, and its loss in mice leads to worse cardiac function in obesity, impaired kidney function in diabetic mice, higher risk of death after a heart attack, and poor contractility of the heart muscle. Despite its protective role, ACE2 function is poorly conserved in mice. Though these animals have 50% higher ACE2 activity than humans, they do not show vasodilation in response to Ang-(1-7), indicating a disconnect between ACE2 and vasodilation in rodents. In fact, it seems that small changes in ACE2 function can be linked to large changes in physiological function.
In other words, the occurrence of natural variation across species has influenced the function of this gene, which could indicate that ACE2 function is dependent on the sequence recognized by the SARS-CoV-2.
Amino acid interactions mediating ACE2 activity govern SARS-CoV-2 binding to human and dog ACE2 (A) SARS-CoV-2 gains cellular entry through the viral spike protein receptor binding domain (RBD, red), which targets binding hotspots on the ACE2 receptor (blue) distal to the ACE2 active site (yellow) [6M17; (23)]. (B) Statistical phylogenetic analyses (dN/dS averages [dots] and ranges [grey]; PAML, HyPhy) of an alignment of mammalian ACE2 coding sequences reveals positive and negative selection (*) on RBD binding hotspots. Alignment of ACE2 residues across boreoeutherian mammals is shown. (C) SARS-CoV-2 is unable to bind mouse ACE2, which displays unique amino acid residues at positions within the RBD binding interface relative to other mammals. (D) Flow cytometry was used to quantify RBD association with human cells (HEK293T) expressing WT and mutant ACE2. (E) The effect of ACE2 mutations on ACE2 hydrolysis rates was measured using a fluorometric peptide (F-G) Interaction effects between amino acids on opposite ends of the viral binding interface determine 181 RBD binding (F-G) and ACE2 activity (H-I) in both human (F, H) and dog ACE2 (G, I).
Identification of deleterious mutations
The current study looked for variations that prevented binding to mouse ACE2.
They found six mutations inside the binding site that abolished binding in species such as humans, dogs, that are naturally susceptible to the virus and show ACE2 activity. These mutations thus disrupted binding between SARS-CoV-2 RBD and ACE2. These also reduced ACE2 hydrolysis by 50% to 60%. These mutations cause indirect structural effects on residues outside the active site of the enzyme.
This explains the discrepancy from earlier studies, which showed no direct correlation between inactivation of enzyme activity and viral binding. These previous studies looked at residues within the active site, though away from the binding site. In fact, when human ACE2 was modified to increase binding affinity, the changes reduced ACE2 activity simultaneously. The ACE2 viral binding domain is crucial for its catalytic function as well, which means that such mutations are unlikely to be selected unless there were compensatory mutations at other sites that kept ACE2 catalytic activity high.
This work also shows that missense alleles at these six sites that disrupt viral binding are rare in the human ACE2 due to the strategic importance of the ACE2 enzyme in cardiovascular health.
Epistasis determines mutational effect
An unusual phenomenon is the presence of compensatory interactions between different sites within the binding domain of ACE2, such that one mutation partially or completely reverses the detrimental effects of another at another site. This is called epistasis, or context-dependence, of the effects of a mutation.
For instance, a quadruple mutation denoted by L79T; T/M82S; Y83F; P84S is associated with a small effect because residues at one end of the binding cleft modulate the effects of residues on the other end. The combination of mutations was thus responsible for reversed effects on each other’s deleterious effects on enzyme activity at this site, which is called sign epistasis. This is probably a sign that there are many other hosts that can harbor this virus, to an unpredictable extent, and missed by purely sequence-based or bioinformatic analysis of the ACE2 binding site.
“This intramolecular epistasis makes the sequence-phenotype relationship unpredictable; [it] has generated sequence diversity driving ACE2 functional diversification but through similar conformations all bound by SARS-CoV-2.”
Only a few species (bat, mouse) have ACE2 mutations that have completely prevented viral binding while preserving or enhancing ACE2 activity. Such mutations may have been achieved by a conjoining of permissive mutations and constraints on the cardiovascular effects of such mutations
What are the implications?
Caution is therefore required to watch for reverse transmission of the virus from infected humans to animals, with subsequent evolution of novel strains that can evade the immune response and cause reinfection of humans. The efficiency of binding coupled with this intramolecular epistasis may have allowed it to infect a wide range of hosts.
Therapeutics may also be affected by these findings. The use of human recombinant soluble ACE2 (hrsACE2) could be used to neutralize the virus by binding to its spike protein, thus preventing it from engaging the host cell receptor. Since this also has catalytic activity, it can reduce angiotensin II and also COVID-19-associated inflammation. This latter effect is by increasing the production of Ang-(1-7) and reducing the hyperactivation of the RAAS induced via ACE2 downregulation by the virus. This could therefore mean an additional benefit from its use beyond the prevention of infection.
In actual life, however, hrsACE2 requires to be inactivated at its catalytic site in patients with cardiovascular conditions that could be worsened by increased sACE2 levels and activity. To reduce ACE2 activity according to need, non-human recombinant ACE2 could be used. Again, these six residues could be introduced to modify the ACE2 site in mink and other mammals with agricultural importance, to make them resistant to the infection.
bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.