Evolutionary trajectory of SARS-CoV-2 and emerging variants

The emergence of a novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and more recently, the independent evolution of multiple SARS-CoV-2 variants has generated renewed interest in virus evolution and cross-species transmission. While all known human coronaviruses (HCoVs) are speculated to have originated in animals, very little is known about their evolutionary history and factors that enable some CoVs to co-exist with humans as low pathogenic and endemic infections (HCoV-229E, HCoV-NL63, HCoV-OC43, HCoV-HKU1), while others, such as SARS-CoV, MERS-CoV and SARS-CoV-2 have evolved to cause severe disease. In this review, we highlight the origins of all known HCoVs and map positively selected for mutations within HCoV proteins to discuss the evolutionary trajectory of SARS-CoV-2. Furthermore, we discuss emerging mutations within SARS-CoV-2 and variants of concern (VOC), along with highlighting the demonstrated or speculated impact of these mutations on virus transmission, pathogenicity, and neutralization by natural or vaccine-mediated immunity.


Background
Coronaviruses (CoVs) can infect humans and animals to cause mild to severe disease, including death [1]. CoVs are divided into four genera: alpha-and beta-CoVs predominantly originate in bats and infect other mammals, while gamma-and delta-CoVs originate in and largely infect avian species [2]. CoV infection in animals is generally associated with gastric symptoms [3], such as acute diarrhea in young pigs that are infected with porcine epidemic diarrhea virus (PEDV) and swine acute diarrhea syndrome coronavirus (SADS-CoV) [4,5]. While CoVs mainly circulate in animals, such as pigs, camels, cats, and bats [6], there have been at least seven documented instances where these viruses have spilled over into humans [7]. These events have led to the emergence of human coronaviruses (HCoVs) that are low and high pathogenic. The origin of the most recently emerged human coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is speculated to be associated with Rhinolophus bats, but the zoonotic transmission pathway remains unknown.
HCoV-229E, HCoV-OC43, HCoV-NL63 and HCoV-HKU1 represent endemic and low pathogenic HCoVs, and are responsible for one-third of common cold symptoms [8]. High pathogenic HCoVs such as severe acute respiratory syndrome coronavirus (SARS-CoV), Middle East respiratory syndrome coronavirus (MERS-CoV), and SARS-CoV-2 cause or have caused severe disease in humans with case-fatality rates of 10.9%, 34.3%, and 2.1%, respectively [9][10][11]. SARS-CoV, MERS-CoV and SARS-CoV-2 are beta-CoVs [12,13]. MERS-CoV belongs to the Merbecovirus subgenus, while SARS-CoV and SARS-CoV-2 belong to the SARS-related coronavirus (SARSr-CoV) species within the Sarbecovirus subgenus [14]. It remains unclear why most HCoVs evolved to largely cause minor illness while MERS-CoV continues to cause severe disease [15][16][17]. In this review, we have highlighted the origins of HCoVs and mapped positively selected for mutations within HCoV proteins to discuss the evolutionary trajectory of SARS-CoV-2. We have also discussed emerging mutations within SARS-CoV-2 and variants of concern (VOC), along with highlighting the demonstrated or speculated impact of these mutations on virus transmission, pathogenicity, and neutralization by natural or vaccine-mediated immunity.

Origin of human coronaviruses
All known HCoVs are speculated to have an evolutionary origin in bats or rodents [1,3,18] (Fig. 1), with five of seven HCoVs originating in bats [3,[19][20][21] (Table 1). Bats are speculated to be primordial hosts for all CoV lineages due to ubiquitous detection of diverse CoVs and constant CoV population growth, which contrasts epidemic-like growths observed in other animals [22]. Although bats and alpacas can serve as MERS-CoV reservoirs [23,24], dromedary camels are the major reservoir host and primary contributor to human infections [25][26][27][28] (Fig. 1). The full extent of wildlife or intermediate animal reservoirs of SARS-CoV-2 is currently unknown.
SARS-CoV-2 is believed to have originated in a seafood market in Wuhan, Hubei Province, China [29], although limited contact-tracing at the beginning of the pandemic does not allow for definitive characterization of the exact events that led to the first human-to-human transmission, including the index patient or initial animal contact. Nonetheless, it is speculated that the natural reservoirs of SARS-CoV-2 are Rhinolophus bats (Table 1) since diverse SARSr-CoVs have been detected in multiple Rhinolophus species [22,30,31], including RaTG13 in R. affinis [32]. RaTG13 is 96.2% identical to SARS-CoV-2 at the whole genome level [32]. Moreover, SARS-CoV-2 contains a polybasic furin-like cleavage site between S1 and S2 spike (S) protein subunits, similar to Rhinolophus CoV Fig. 1 Speculated animal origins of known human coronaviruses. HCoV species are organized chronologically (top to bottom) by their speculated dates of spill over into humans. Intermediate hosts (top to bottom) shown are alpacas, cattle, civet cats, dromedary camels, pangolins, and unknown (denoted as a question mark). Genome similarity to humans (A) indicates percentage similarity of CoV genomes detected in reservoir species with corresponding human CoV. Genome similarity to humans (B) indicates percentage similarity of CoV genomes detected in intermediate species with corresponding human CoV. Non-human CoVs that are highly pathogenic in animals, such as PEDV and SADS-CoV, are not shown here. Genomic percentage similarities were extracted from existing primary studies [20,21,32,56,60,[277][278][279][280][281][282][283]  . Indeed, if SARS-CoV-2 did transmit from animals to humans, further sampling in Hubei Province may identify more closely related SARSr-CoVs in archived animal specimens. Investigating the possibility of an infected person travelling to Wuhan and unwittingly spreading the virus will be more difficult in the absence of archived samples and records of travel history.
Despite the abundance of SARSr-CoVs and beta-CoVs in bat species [52,53], it is likely that additional reservoirs and intermediate hosts remain undetected [54]. Pigs, alpacas, and dromedary camels also maintain a variety of CoVs with the potential to transmit to humans [3,12,20,[55][56][57]. Independent insertions within RBDs of SARS-CoV, MERS-CoV, and SARS-CoV-2 suggest convergent evolution, which will likely lead to emergence of more pathogenic HCoVs [58]. Further sampling of bats, pangolins, and other species that share an ecological niche with bats may help piece together the puzzle surrounding the spill over of  [20,[308][309][310] SARS-CoV-2 into humans [59] and also help discover other CoVs with potential to infect humans. Aside from consistent spill over of MERS-CoV from camels [60], HCoVs have emerged through limited spill over events, followed by human-to-human transmission [3,61]. While challenging to predict, future spill over events are likely, due to the long history of CoV host shifting [62][63][64][65]. Anthropogenic factors such as urbanization and deforestation increase habitat overlap of humans and animals, providing increased zoonotic transmission opportunities [57,66]. Areas of high contact between humans, wildlife, and domesticated animals, such as live animal wet markets provide opportunity for viral recombination and adaptation to a broader range of animal species prior to transmission to humans [57]. Identifying existing CoV diversity in such areas will enhance our understanding of ecological opportunities for zoonosis and will help us better predict and prevent the emergence of future HCoVs.

Evolution of SARS-CoV-2 and its variants
Co-evolution of CoVs with their hosts is driven by genetic diversity that is selected through evolutionary pressures. CoV genetic diversity is made possible by a large genome (26.4-31.7 kb) [67], high mutation rate due to a low fidelity viral polymerase (~ 10 -4 substitutions per site per year) [68,69], and high recombination frequency (up to 25% for the entire genome in vivo) [70,71]. Mutations that confer greater fitness are selected for, leading to antigenic drift. Ratios of the rates of non-synonymous/synonymous mutations (dN/dS) greater than one, less than one and equal to one indicate positive selection, negative (purifying) selection and neutral evolution, respectively [72]. SARS-CoV-2 genomes are currently under purifying selection [73,74]. Despite observing little viral diversity at the beginning of the COVID-19 pandemic [75,76], positive selection with presumed advantages such as increased transmission rates has now been documented [77][78][79] (Fig. 2, Table 2). However, functional characterization of these mutations remains under-investigated.
Antigenic drift is most frequently observed in viral surface proteins that are highly exposed to selection pressures of the immune system, such as neutralizing antibodies [80]. Indeed, CoV spike genes, particularly the S1 and RBD coding regions, have the highest detected non-synonymous mutation rates [81,82], a trend observed across the majority of HCoVs (Fig. 2). For low pathogenic and endemic HCoVs, multiple positively selected for residues and polymorphic sites are found in the N-terminal domain (NTD) of S [83][84][85][86][87][88]. A notable exception is HCoV-HKU1, for which there is a shortage of sequencing data outside of the hemagglutinin esterase (HE) gene. Emerging data suggest that positively selected for and homoplastic sites have been observed within the SARS-CoV-2 NTD as well [78,[89][90][91]. Given the observations with other HCoVs (Fig. 2) and the detection of neutralizing epitopes within the SARS-CoV-2 NTD [91,92], we speculate that with continued circulation, vaccination and convalescent sera therapy, further positively selected for mutations in the NTD are likely to occur. Further retrospective research on the evolution of endemic HCoVs may help predict the likely evolutionary trajectory of SARS-CoV-2.
CoV genomic mutations give rise to virus variants, and closely related variants are grouped into clades. SARS-CoV-2 variants have been clustered into nine clades: L, V, S, G, GH, GR, GV, GRY and O [93,94] (Table 3), named after their most representative mutations [95]. Clade L dominated the beginning of the pandemic [38], prior to the appearances of clade S and the less defined clade O in early January, 2020 [73,93,96]. Clades V and G appeared in mid-January, followed by clades GH and GR at the end of February, clade GV at the end of June, and clade GRY in September, 2020 [94,97,98]. Clades L and V are likely extinct, while clades G, GH, GR, and GRY comprise the majority of global SARS-CoV-2 sequences currently [97,98]. Clade S has also been declining since the emergence of clade G [93]. Following rapid dissemination of clade G and its derivatives, such as B.1.1.7, B.1.351, P.1, and B.1.617.2 variants (Table 5), we may see the rise of other variants, selected by mounting population-level immunity and other yet unidentified factors [89,[99][100][101], highlighting the need for international genome surveillance efforts and global data sharing via the established GISAID resource [102].
D614G is usually accompanied by three other mutations which represent clade G [104,116,117] (Table 3). Of these mutations, P323L in the RNA-dependent RNA polymerase (RdRp), encoded by Nsp12 (Fig. 2, Table 2), is particularly interesting as CoV RdRp tends to be highly Mutations identified in human coronaviruses. Red dots within the genomes correspond to specific amino acid residues that have been strongly positively selected for such that a specific mutation has become dominant in the region where it emerged [74, 78, 83-91, 94-96, 99-101, 104, 111, 116, 117, 121, 123-125, 129, 131, 132, 135, 138-140, 146, 151-154, 158, 162, 278, 284-293]. Genomic regions highlighted by red bars correspond to deletions that have been selected for, while purple bars correspond to regions with significant polymorphisms within a CoV species. Beta-CoV Lineage B (Sarbecovirus) is represented within the blue shaded area, beta-CoV Lineage C (Merbecovirus) is represented within the yellow shaded area, beta-CoV Lineage A (Embecovirus) is represented within the red shaded area, and alpha-CoVs are represented within the green shaded area. Genome length in kilobases (kb) is noted on top. See Table 2 for more details conserved by purifying selection given its critical role in viral genome replication [118,119] (Table 4). P323 falls outside of the RdRp catalytic site and within a relatively uncharacterized interface domain that may interact with proteins that regulate viral polymerase function [120]. The correlation of this mutation with increased point mutations [121] elsewhere in the genome raises an intriguing hypothesis that P323L diminishes RdRp proofreading ability, leading to increased mutation rates. Moreover, P323L downregulates the association of Nsp12 with the Nsp8 primase subunit (Table 4), reducing polymerase activity and viral replication [122]. Decreased replication could decrease symptomology, leading to reduced COVID-19 detection and greater population-level spread. It is important to characterize the cumulative effect of all mutations, as any reduction in transmission due to P323L could be compensated for by the co-existing D614G mutation. Multiple factors may contribute to the success of clade G and its derivatives via rapid spread with low detection in human populations [104].
Another mutation of interest (L84S) lies within ORF8 [123,124,135], a protein implicated in evasion of host immune responses [136,137] (Table 4). ORF8 was under strong directional selection at the beginning of both SARS-CoV-2 [124] and SARS-CoV outbreaks [138], supporting the theory that it facilitates zoonotic transmission and adaptation in alternate hosts [139,140]. However, the over-representation of ORF8 deletions in SARS-CoV with no apparent effect on viral survival [138] suggests that ORF8 may be dispensable in humans [139], and L84S mutations may not be significant. While This table illustrates positively selected for residues across multiple human coronaviruses. Shaded boxes represent proteins not encoded by the specific CoV species. Text in bold highlight mutations and deletions that were positively selected for and showed population-level expansion, while non-bolded text represents highly polymorphic sites. Sites are indicated as nucleotide (nt) position or amino acid (aa) position. Empty cells in the table represent lack of evidence for positive selection or lack of publications on positive selection within these regions   S Spike glycoprotein Cleaved into S1 and S2 subunits. S1 binds host receptor (ACE2) while S2 mediates viral and host membrane fusion [333] ORF3a Orf3a viroporin Activates NF-kB and NLRP3 inflammasome to contribute to cytokine storm. Promotes viral release and may induce necrotic cell death [334][335][336] ORF3b Accessory protein ORF3b IFN-1 antagonist [337] E Envelope protein A viroporin involved in viral assembly, budding, and pathogenesis. Forms CoV envelope [338,339] M Membrane protein Forms viral membrane and induces N and S localization to the ER-Golgi-Intermediate compartment for virion assembly and budding [340] ORF6 Accessory protein ORF6 IFN-1 antagonist [144] ORF7a Accessory protein ORF7a SARS-CoV ortholog inhibits bone marrow stromal antigen 2 mediated tethering of virions to host plasma membrane [341] ORF7b Accessory protein ORF7b SARS-CoV ortholog attenuates viral replication [342] ORF8 Accessory protein ORF8 Inhibits IFN-1 activity and downregulates MHC-1 expression to evade host immunity [136,137,144] N Nucleocapsid Involved in immune evasion through IFN-1 antagonism, nucleocapsid formation, viral RNA replication, and virion assembly [144,145] ORF9b Accessory protein ORF9b Suppresses IFN-1 responses through inhibition of TOM70 [343] ORF9c Accessory protein ORF9c Interferes with IFN signalling, antigen presentation, and complement signalling. Induces IL-6 signalling [344] ORF10 Accessory protein ORF10 Interacts with a Cullin 2 RING E3 ligase complex to potentially modulate ubiquitination [345] clade GR [123]. RG203KR alters N protein morphology, resulting in increased intraviral protein binding affinity [132]. N-M interactions are necessary for CoV viral assembly [141,142], while N-envelope (E) interactions potentially increase production of virus-like particles [143]. Therefore, increased intraviral N protein binding affinities could contribute to increased viral replication. RG203KR may also confer immune evasion properties to SARS-CoV-2 considering the rapid expansion of clade GR and the role of N protein in antagonizing human antiviral immune responses [144,145] ( Table 4). The global prevalence of variant B.1.1.7 has generated clade GRY from clade GR [146]. Clade GV is associated with the European variant 20A. EU1 containing spike NTD mutation A222V [105,147]. A222 is located within a speculated B lymphocyte epitope [148] that may impact neutralization by human antibodies, consistent with observed SARS-CoV-2 re-infection with a clade GV variant [149]. The rise in prevalence of variant 20A.EU1 and clade GV is most likely associated with the relaxing of travel-associated restrictions across Europe near the end of the summer of 2020 considering the rapid decline in prevalence of global clade GV sequences in 2021 [97,150].
Early data suggest that RBD mutation N501Y emerged recurrently in multiple regions due to increased transmissibility, and is associated with multiple VOCs [89,99,100,158] (Table 5). SARS-CoV-2 N501 serves as one of six critical S residues required for binding to ACE2 [159] and N501Y increases viral infectivity through greater S-hACE2 binding affinity, likely due to stronger interactions with ACE2 residues Y41 and K353 [160]. Other critical residues within the SARS-CoV-2 RBD (L455, F486, Q493, S494, Y505) [73] should be closely monitored as mutations may increase SARS-CoV-2 transmission in humans and facilitate zooanthroponotic transfer to other species.
Early studies of the highly transmissible B.1.1.7 variant [77,161] originating in the United Kingdom described 17 co-occurring non-synonymous mutations or deletions [89], which are more than expected since the mutation rate of SARS-CoV-2 is estimated to be around 2.4 × 10 -3 per site per year [135]. In addition to N501Y, spike 69-70del, Y144del, and P681H mutations are speculated to be of functional significance [78,162] (Table 5). Spike NTD 69-70del variants have shown significant Table 5 SARS-CoV-2 variants of concern (as of July 22, 2021) Variant names are based on Rambaut et al.'s classification [347]. Other commonly used names are mentioned in brackets. Mutations mentioned here are nonsynonymous mutations that are speculated to confer some functional significance. These variants contain other mutations that may also contribute to viral advantages [89,[99][100][101]. Updated information about SARS-CoV-2 VOCs can be accessed through the GISAID resource (https:// www. gisaid. org). Dates of emergence are based on retrospective analyses. S, spike. del, deletion  [196,201,202,346] transmission expansion, with speculated increased resistance to antibody-mediated neutralization [92] likely associated with sequestration of a protruding spike loop [78]. Y144del confers antibody resistance due to loss of a negative surface charge [163,164]. Spike P681 is located in a known CoV mutational hotspot [83,101] directly adjacent to the SARS-CoV-2 S1/S2 furin cleavage site (aa 681-684) [89,165,166] which promotes virus entry into host cells [167]; mutation in this region may increase cleavability and membrane fusion to enhance infectivity. P681 is also within an antigenic epitope recognized by B and T lymphocytes, implicating host immune response alterations [168]. P681H may therefore represent adaptive evolution to evade host immunity, although confirmatory studies are required. Another speculated B.1.1.7 mutation at ORF8 (Q27stop) causes early protein termination [89]. Truncated ORF8 has been associated with milder symptoms [169], although increased mortality is also associated with the B.1.1.7 variant [79,170]. Emerging mutations in B.1.1.7 must be monitored and investigated, such as the sub-lineage VOC202102/02 that contains the RBD mutation E484K, which is associated with antibody resistance [171][172][173].
Mutations within the RBD (K417T, E484K, N501Y) have also been observed in the P.1 variant ( Table 5) that likely originated in Brazil and has since spread to other countries [101,[189][190][191]. In contrast, the P.2 variant only contains E484K, likely acquired through convergent evolution with P.1 [186,192]. Little is known about the P.1 variant, but based on emerging data [193], we speculate that the RBD mutations likely affect antibody-mediated neutralization and contribute to increased transmission as observed with B.1.351. Mutations shared between the B.1.1.7, B.1.351, and P.1 variants are speculated to have arisen independently, indicating convergent evolution [194] (Table 5). These variants also share Nsp6 3675-3677del, with unknown functional significance [194,195]. VOC B.1.617.2 was first identified in India in late 2020 and contains positively selected for mutations within the spike protein, namely, L452R, T478K, and P681R, along with the D614G mutation [196] ( Table 5). Mutation of the uncharged and hydrophobic leucine (L) residue into the positively charged and hydrophilic arginine (R) residue at spike position 452 allows for an increased electrostatic interaction with negatively charged ACE2 residues E35, E37, and D38, likely leading to the observed increase in S-hACE2 complex stability, viral infectivity, and virus replication [196,197]. Furthermore, abolition of the hydrophobic surface patch through the L452R mutation led to reduced antibody-mediated neutralization and cellular immune recognition [196][197][198]. Spike mutation T478K has also been shown to increase electrostatic interactions in the S-hACE2 complex and may increase binding affinity similar to the S477N mutation [199]. The mutation T478K is within a neutralizing epitope close to the immune evasion mutation E484K/Q that is present in multiple SARS-CoV-2 variants, including the ancestral B.1.617 lineage and current sub-lineages B.1.617.1 and B.1.617.2 [181,200,201]. T478K in combination with L452R may contribute to increased resistance to neutralization by monoclonal antibodies, convalescent sera, and vaccinated sera [201,202]. B.1.617.2 has increased replication efficiency in human airway systems relative to the B.1.1.7 lineage due to enhanced spike cleavability, which is likely augmented by the P681R mutation [201,203]. P681R is known to increase cell-to-cell fusion in the respiratory tract, potentially increasing transmissibility and pathogenicity in infected individuals [201,203]. B.1.617.2 may thus represent a VOC with similar resistance to antibody neutralization as B.1.351 and transmissibility beyond B.1.1.7 [200]. Recently discovered B.1.617.2 sequences containing the K417N mutation (AY.1/AY.2 lineages) must be monitored for altered antibody resistance and increased transmissibility [204].

Other variants of interest
Multiple emerging SARS-CoV-2 lineages are not considered VOCs but are still of interest and may become VOCs in the future. One variant, B.1.525, was first detected in December, 2020, in the United Kingdom and Nigeria and has since spread internationally. B.1.525 contains spike mutations 69-70del, E484K, Q677H, and F888L. Q677P/H has emerged in disparate variants and may affect spike cleavability similar to P681H [158,[216][217][218]. F888L lies between the fusion peptide and heptad repeat region of the S2 subunit [219] and may impact host cellular entry, similar to the impact of heptad repeat mutations in MERS-CoV [139,220].

Multiple factors will determine the evolutionary trajectory of SARS-CoV-2 and the COVID-19 pandemic
The future of SARS-CoV-2 and COVID-19 remains uncertain. Many virological, immunological, and social factors will influence the epidemiological trajectory of this virus. One particularly intriguing question that remains unanswered is whether SARS-CoV-2 will become endemic in the human population, like HCoVs NL63, OC43, HKU1, and 229E [232][233][234].
Currently, endemic HCoVs cause seasonal outbreaks [235], with increased circulation observed in the winter in temperate regions [232]. Cold temperatures are favourable for enveloped viruses [236], as lower temperatures enhance lipid ordering of the viral envelope, allowing the virus to remain protected outside the host for longer periods of time [237,238]. Low temperatures also enhance aerosol transmission of respiratory viruses by allowing virions to remain suspended in the air for a longer duration [239]. Furthermore, cold and dry environments can have immunosuppressive effects on a potential host, further increasing the chances of infection [240][241][242]. Evidence suggests decreased transmission of SARS-CoV-2 in warmer climates [243][244][245][246], likely due to degeneration of viral structural stability with increasing temperatures [247]. Decreased transmission of SARS-CoV-2 was not observed during the summer of 2020 [11,248] likely because of the sheer number of cases and an immunologically naïve population. For seasonality to have an observable impact on SARS-CoV-2 transmission, the basic reproduction number (R 0 ) must drop from its current estimate of around 2.5 to less than 1 [249]. In theory, SARS-CoV-2 R 0 should drop substantially when population herd immunity is reached through natural infection and vaccination, allowing for meteorological factors to influence viral transmission, leading to seasonal fluctuations. Other intervention mechanisms such as effective social distancing, quarantine, and contact-tracing will contribute towards reducing the R 0 for SARS-CoV-2 [250,251].
Multiple studies have demonstrated short-lasting immunity to endemic HCoVs, with waning of protective immunity and re-infections common within 80 days [85] to one year [252][253][254][255]. There is no observable association between endemic HCoV re-infection and infection severity [254]. Waning of humoral immunity within a year [256][257][258][259][260] and re-infection of immunocompetent patients [149] have been demonstrated for SARS-CoV-2, suggesting the possibility of annual outbreaks [233,261]. A weaker initial immune response and sharper decline of antibody levels have been reported in individuals with asymptomatic SARS-CoV-2 infections [257,258]. Thus, multiple exposures to SARS-CoV-2 may be required to develop sufficient immunity to prevent future re-infections, which may also be influenced by adaptive evolution of SARS-CoV-2 in the human population ( Table 5). The duration of protection through vaccination and natural exposures is being closely monitored, along with antigenic evolution of SARS-CoV-2 that may lead to immune escape. Indeed, the evolutionary trajectories of endemic HCoVs suggest that SARS-CoV-2 will evolve to co-exist with the human population. However, with roll-out of the first ever HCoV vaccines, predicting the evolutionary trajectory of SARS-CoV-2 remains challenging.
An important factor that may influence ongoing SARS-CoV-2 transmission is the potential for cross-protection by humoral and cellular immune responses induced by related endemic HCoVs. There is evidence of cross-protection within the same genera of HCoVs [233,262,263], but not between genera [264]. Thus, immunity against beta-CoVs HCoV-OC43 and HCoV-HKU1 may provide some protection against COVID-19 [265][266][267][268], while immunity against alpha-CoVs HCoV-229E and HCoV-NL63 will likely provide little to no protection. Antibody-dependent enhancement has not been observed for SARS-CoV-2 [269,270], ruling out the possibility of increased disease severity by cross-reactive antibodies generated against endemic HCoVs. The high frequency of CoV recombination during co-infections raises the additional concern that SARS-CoV-2 recombination with seasonal HCoVs could generate novel CoVs [131,271,272]. The role of HCoV co-infection has not been reported or extensively studied and will be especially important for immunocompromised and elderly individuals.

Conclusions
SARS-CoV-2 continues to evolve and adapt to the human population as highlighted by the emergence of novel variants. Mutations within the spike protein of SARS-CoV-2 variants confer increased transmissibility and some degree of resistance to antibody-mediated neutralization. However, recurrent attenuating mutations, such as P323L, L37F, G251V, and Q27stop have also been identified and are speculated to reduce disease severity. The appearance of attenuating mutations suggests that SARS-CoV-2 is evolving to become less pathogenic in humans. The current SARS-CoV-2 pandemic is driven by asymptomatic, pre-symptomatic, or otherwise unrecognized cases [273][274][275]. Reduced pathogenicity of SARS-CoV-2 combined with mounting population-level immunity will likely cause a reduction of severe cases of COVID-19, leading to an apparent abatement of the pandemic, followed by endemic circulation of low pathogenic SARS-CoV-2 variants. A similar evolutionary trajectory may have led to the establishment of current low-pathogenic endemic HCoVs [276].
Monitoring future emerging variants of SARS-CoV-2 is critical to determine control measures for the COVID-19 pandemic. Mutations speculated to reduce immune recognition, such as within the spike protein (S13I, 69-70del, W152L, A222V, K417N, N439K, S477N, T478K, E484K/Q, F490S, P681H/R) and nucleoprotein (RG203KR) should be studied for reduced sensitivity to natural or vaccine-induced immunity. Other factors, such as zoonotic and zooanthroponotic transmission of SARS-CoV-2, cross-protection through immunity against endemic HCoVs, and the possible creation of novel animal reservoirs through zooanthroponosis should continue to be investigated as they may have significant implications on the evolutionary trajectory of SARS-CoV-2 and the COVID-19 pandemic. travel-related cases, and vaccine readiness. J Infect Dev Ctries. 2020;14 (1)