Ancestral SARS-CoV-2 is not a natural virus because it cannot support sustained transmission in any animal species: Ancestral SARS-CoV-2, with its metastable D614/Furin Cleavage Site phenotype, cannot be maintained outside of a laboratory environment

The origin of SARS-CoV-2 (CoV-2) remains currently unknown and because investigation inside China is currently impossible, only indirect methods of analysis are available. Here I use the metastable nature of the CoV-2 Spike Protein to infer the timing of the first human infection as well as the first zoonotic event. First, the progenitor of SARS-CoV-2 is determined to be D614/Null Furin Cleavage Site (FCS), based on the time to Most Recent Common Ancestor (tMRCA) for that phenotype within the Sarbecoviruses of 3500 BCE to 900 ADE. Next, I document that the D614/FCS phenotype is metastable under all experimental conditions. This is due to the shedding of the furin-cleaved S1 protein, generating non-infective particles, and is therefore an intrinsic viral property and is thus host species independent. The D614G substitution solves the S1 instability problem. Under all experimental conditions, the D614/FCS phenotype evolves into either the G614/FCS phenotype or reverts to the D614/Null FCS phenotype. In cell culture and animal-to-animal experiments these changes all begin in the first inoculation and become fixed within four passages. In human-to-human transmission, the G614/FCS phenotype appears within the first five mutations. Given that all initial patient CoV-2 infections were the D614/FCS phenotype and assuming the tMRCA for the COVID-19 pandemic is approximately November 13, 2019, the first introduction into a human can be no earlier than September 9, 2019. Any earlier human infection would have already contained the G614/FCS phenotype in the first sequences at the beginning of the epidemic if the virus had been present in a non-human intermediate or reservoir host in the wild for any length of time before the beginning of the outbreak. Similarly, at this time there has been no animal host identified in which the D614/FCS phenotype can be stably transmitted for more than a week; all infections rapidly evolve to the G614/FCS phenotype. Based on these findings, a zoonotic origin for COVID-19 requires a hypothesis that explains the highly adapted receptor binding domain present at the outset that supported human-to-human transmission, the insertion of the first FCS in any Sarbecovirus virus in more than 900 years, and the first animal-to-human transmission within literally days of that FCS insertion. The alternative and more parsimonious conclusion is that these steps were achieved by serial passage of a D614/Null FCS virus in non-human primates or humanized mice for ACE2 adaption, the laboratory insertion of the FCS, and an almost immediate laboratory-acquired infection, sometime after September 9, 2019.

Version 2 deletes nothing. It adds a section on D614G reversions late in the pandemic and their implications. It adds a section on an experiment of nature that supports the D614/FCS metastability thesis. The feline coronavirus exists in two states, one with an FCS and one without. The FCS variant has mutations in the N-terminal adjacent section of S1 that stabilize the spike protein.