Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), first reported in Wuhan (China) in late 2019, soon spread around the globe. To date, more than 700 million people have been infected and 6.97 million have died worldwide. Despite the availability of vaccines, SARS-CoV-2 remains a cause for concern. Like other RNA viruses, SARS-CoV-2 is endowed with a high mutation rate and high viral load and accumulates mutations with each replication cycle. As a result, viral variants that differ in one or more nucleotides are continuously being generated in infected hosts (1–3). The nomenclature of SARS-CoV-2 variants plays a critical role in facilitating their clear identification, tracking, and global collaboration in understanding the evolutionary dynamics of the virus. For Nextstrain clade naming, a new major clade earns its designation once it attains a 20% frequency on a global scale, regardless of the time frame. When computing these frequencies, it is crucial to ensure a relatively uniform sampling of sequences across different times and geographical locations due to the considerable disparity in sequencing efforts among countries. Clade names are formulated using a standardized protocol, typically derived from the year of emergence and the subsequent available letter in the alphabet. Additionally, a newly identified clade must exhibit a minimum of two mutations differentiating it from its parent major clade. This systematic approach ensures consistency and accuracy in clade designation, facilitating the clear identification and tracking of viral lineages in genomic surveillance studies. SARS-CoV-2 has undergone several genetic mutations leading to the emergence of various clades since its first identification in 2019. A recent study using Belgian data revealed an enhanced immune escape capability exhibited by the Omicron variant compared to the Alpha and Delta variants, resulting in a substantial reduction in the protective efficacy conferred by both acquired immunity and vaccination. Furthermore, a decline in vaccine effectiveness over time was observed, underscoring the significance of booster doses to sustain long-term immunity (4).

In the Netherlands population, the effectiveness of primary and booster vaccination against SARS-CoV-2 infection was estimated overall and in four risk groups defined by age and medical conditions during the Delta and Omicron BA.1/BA.2 periods. The findings underscored the advantages of booster vaccinations in reducing infection rates, particularly within at-risk groups (5).

Investigations of the complete of SARS-CoV-2 proteome have often been limited to analyzing only a few sequences or individual sequences, rather than taking advantage of a large sequencing dataset. However, a direct RNA sequencing approach was employed to assess the SARS-CoV-2 transcriptome in Vero E6 cells, and mass spectrometry was used to explore the proteome and phosphoproteome of these virus-infected cells (6). Furthermore, a proteomic analysis of proteins extracted from nasopharyngeal swabs of 12 patients diagnosed with COVID-19 identified 13 different SARS-CoV-2 proteins. Additionally, host proteome analysis revealed that several key host proteins were uniquely expressed in patients with COVID-19 (7). In a separate study, distinct epitopes of seven different proteins were identified using the complete SARS-CoV-2 virus genome obtained from the NCBI database. However, the 12 protein sequences of the genome were formatted as FASTA files using RefSeq accessions (8). Furthermore, a proteome-wide study of SARS-CoV-2 assessed its potential to induce autoimmune diseases by segmenting the proteome into peptides and identifying shared peptides with experimentally confirmed human T-cell and B-cell epitopes (9).

The SARS-CoV-2 genome consists of a single-stranded, unsegmented, positive-polarity RNA molecule [(+) ssRNA] 29,903 nucleotides in length encoding 13 ORFs (10). Two-thirds of the viral genome corresponds to ORF1a and ORF1b, which express the two polyproteins pp1a and pp1ab, the latter though a − 1 ribosomal frameshift, and that are processed by two viral proteases into 16 non-structural proteins (nsp). ORF1a encodes nsp1 to nsp11 and ORF1b comprises nsp12 to nsp16. Non-structural proteins make up the replication and transcription machinery and are responsible for the maintenance of the viral genome (11). Some nsp proteins are targets for antiviral drugs such as nsp12, the RNA-dependent RNA polymerase (RdRp), and nsp5, the 3C-like protease (Mpro, 3CLpro) (12). In addition, nsp3, the papain-like protease (PLpro) is also a therapeutical target for antivirals (13).

The structural proteins are, namely, the surface glycoprotein or spike (S), the envelope protein (E), the membrane glycoprotein (M), and the nucleocapsid phosphoprotein (N). The S protein has been shown to play a major role in virus attachment and entry into cells, being a key antigen for development of vaccines and neutralizing antibodies, and as a pharmacological target (14–16). The E protein plays a key role in the pathogenesis of the virus affecting the binding of SARS-CoV-2 to the tight junction proteins (17). The M protein is responsible for maintaining the shape of the virion by spanning the membrane bilayer and facilitates budding of the viral particles from the host cells (18). Interestingly, the M protein was found to elicit IgM response during the acute phase of SARS-CoV-2 infection (19). The nucleocapsid function is to maintain the genome structure inside the envelope (20). The N protein has been identified as an important target for T-cell response, making it a suitable candidate for next-generation COVID-19 vaccines against emerging variants (21–23).

There are nine accessory proteins, ORF3a, 3d, 6, 7a, 7b, 8, 9b, 14, and 10, which are produced from the encoding accessory genes ORF3a, ORF3d, ORF6, ORF7a, ORF7b, ORF8, ORF9b, ORF14 and ORF10, respectively (24). Although the exact functions of SARS-CoV-2 accessory proteins remain to be determined, studies of other coronaviruses suggest that they are not essential for viral replication but can modulate replication and pathogenesis through interaction with host pathways including antiviral activation (25, 26) and viral translation (27).

The existing literature predominantly concentrates on isolated mutations within the S protein rather than examining mutations as a network involving multiple proteins. Individually, mutations in SARS-CoV-2 might not pose significant risks, however their collective effect in tandem with other mutations could amplify the virus’s transmissibility and virulence. Consequently, relying solely on information about distinct segments of the virus might provide an incomplete understanding. Therefore, a comprehensive study of mutations across the entire SARS-CoV-2 genome and proteome becomes essential. Such a holistic approach is critical in understanding the virus’s mechanisms for evading vaccines and is instrumental in the development of effective vaccines and therapies.

Mutations (including deletions) that alter protein sequence, may affect physicochemical properties and folding conformation of proteins resulting in changes in biological functions. Amino acid positions in proteins could be assigned a single conservation number based on the alignment of homologous proteins for quantification of amino acid substitutions (28, 29).

Given that structural variations due to mutations could affect vaccine effectiveness and drug function, and thus the severity of COVID-19, this work involved an extensive genome-wide mutation combination analysis of SARS-CoV-2 in vaccinated patients from Italian and Spanish populations as two independent datasets. Analyzing more than 7,500 proteomes from each population increases the statistical power and reliability of the results. Our objective was to identify mutations across the entire viral genome, rather than focusing on individual mutations, that may be associated with vaccine failure, in addition to analyze the physicochemical properties of the resulting amino acid changes. This study involved aligning SARS-CoV-2 genomic sequences from both vaccinated and unvaccinated COVID-19 patients using the Wuhan genome as reference. We examined the frequency of specific mutation sets and analyzed their physicochemical properties to understand how these mutations may affect the structure and function of viral proteins. This approach provides a comprehensive view of the genetic diversity of SARS-CoV-2 variants circulating in different geographic regions and contributes to a deeper understanding of the underlying mechanisms of the vaccine effectiveness, which is crucial for informing public health strategies and vaccine development efforts.

Mutations and deletions in SARS-CoV-2 proteins can significantly alter their structure and function. This study investigated the impact of mutations of interest (MOI) across the entire SARS-CoV-2 proteome on vaccine escape using two datasets. The analysis revealed several mutations and combinations of residues that may influence vaccine coverage, particularly concerning the Delta (clade 21J) and Omicron BA.2 (clade 21L) variants.

For the Delta variant, mutations at P2046L (ORF1a), P2287S (ORF1a), L829I (ORF1b), R158- (S) and Q9L (N) were identified as potentially crucial for vaccine effectiveness. Similarly, mutations such as L3201F (ORF1a), A27- (S) and G30- (N) in the Omicron variant were found to impact vaccine effectiveness. However, deletions A27- (S), I68- (S) and G30- (N), along with the R346T mutation in the spike protein of the Omicron variant, may further compromise vaccine effectiveness.

Observations of mutations in the spike protein, particularly A27-, I68-, and R158- located within the N-terminal domain (NTD), and the R346T mutation within the Receptor Binding Domain (RBD), align with previous studies. Molecular dynamics simulations have shown the critical involvement of NTD residues in interactions with monoclonal antibodies (42), suggesting potential immune evasion risks for viruses carrying mutations in these regions. Furthermore, mutations within the RBD have been shown to affect the binding affinity to the ACE2 receptor, indicating potential shifts in the binding free energy of the RBD-ACE2 complex and modified chemical interactions, leading to increased stability (43).

Limited published data on vaccine efficacy or immunity related to the A27-(S) deletion were found. However, the spike 68–76 deletion within the NTD was identified in a human hepatoma cell clone termed Huh7.5-adapted-SARS2, indicating genetic adaptations. This modified version of SARS-CoV-2 effectively infiltrated A549 lung cancer cells, inducing cellular damage, a capability absent in the original strain, which exhibited no infectivity toward A549 cells. Additionally, the Spike 68–76 deletion variant displayed increased susceptibility to IFN-α2b treatment in comparison to the wild-type SARS-CoV-2 strain. However, the Spike 68–76 deletion was not found in SARS-CoV-2 isolates obtained from VERO E6 cells (44). In the context of vaccine stability and effectiveness, it suggests that despite the presence of the 68–76 deletion in vaccine batches (CoronaVac), it might not drastically alter the vaccine’s effectiveness. The R158- (S) mutation in combination with E156G/157 deletion and L452R mutation has been suggested to exhibit higher infectivity in spike-pseudotyped viruses (45). However, experimental evidence points to the 156–158 deletion notably diminishing the neutralization capacity against antibodies present in the sera of convalescent COVID-19 patients and vaccinated individuals (42, 46, 47).

The R346T change in the RBD is a key mutation for neutralization escape, enhanced fusogenicity, and enhanced S protein processing. Structural modeling suggests that R346T appears to disrupt salt bridge formation between the S protein and class III monoclonal antibodies (e.g., Cilgavimab), lowering effectiveness (48, 49). However, in our study the set of sequences containing R indicates poorer vaccine coverage than T for the Omicron clades 22B/22E.

The outcomes related to spike mutations impacting the infectivity of SARS-CoV-2 appear to exhibit a wide-ranging scope. Conversely, the available data regarding the influence of mutations occurring in other viral proteins seems comparatively constrained.

An analysis conducted on 244 SARS-CoV-2 positive samples, gathered during the second wave of the pandemic, indicated that mutations P2046L and P2287S in the nsp3 (ORF1a) gene might contribute to persistent symptomatic COVID-19 infections post-vaccination (50). Additionally, an investigation involving severe, moderate, and mild COVID-19 cases, encompassing individuals who were either partially or fully vaccinated (with Covishield/Covaxin) or unvaccinated, revealed a marginal association of the P2287S mutation with disease severity (51). The nsp3 protein in the SARS-CoV-2 virus constitutes a crucial component of the viral replicase complex, contributing significantly to multiple functions associated with viral replication (52), transcription (53), and modulation of the host immune response (54), but no specific data on the impact of the L3201F mutation on these functions have been found in the literature.

A study aimed at modeling the fitness of several SARS-CoV-2 lineages by combining the effect of individual mutations introduced a scalable hierarchical Bayesian regression model to analyze all available SARS-CoV-2 genomes. The study identified the L829I (ORF1b) mutation in nsp12, which promotes an amino acid change in the RdRp (RNA-dependent RNA polymerase) thumb subdomain that could affect the function of the enzyme. RdRp plays a critical role in replicating and transcribing the viral genome in RNA viruses like SARS-CoV-2 (55).

The N protein consists of different structural components, namely, an N-arm, an N-terminal RNA-binding domain, a linker region containing serine/arginine-rich loops (SR-rich region), a C-terminal RNA-binding domain, and a C-tail (56–61). Some regions of the N-arm (amino acids 1–46) have been identified as immunodominant epitopes. Specifically, five monoclonal antibodies were developed through a study of epitopes targeting the N protein of SARS-CoV-2. The research revealed that one particular antibody had a specific affinity for the N-arm region of the N protein (62), suggesting the possible involvement of mutations identified in this study such as Q9L and G30- in immune evasion.

As mentioned, some of the substitutions and deletions associated with vaccine failure in our study seem to be in line with previous studies. However, others have not been previously linked to vaccine coverage or immunity. Therefore, a genome-wide analysis of SARS-CoV-2 mutations, and their effects on the proteome, could help to understand the molecular basis of viral vaccine escape, in connection with vaccine and therapeutic drug development.

The effect of COVID-19 booster-dose vaccination against the Omicron variant has been reviewed by a study that identified a total of 27 published studies supporting the effectiveness of booster dose vaccine (63). Our results are consistent with the improved effectiveness of the booster dose against the Omicron variant in Dataset_2, where the vaccines show high coverage further improved by the administration of the third dose for clades 21L and 22B. However, in Dataset_1 where the vaccine is not effective for Omicron clades, administration of the third dose showed a slight protection for clades 21L, 22B/22E, and 22F/23A. We hypothesize that the discrepancies observed between these datasets might be attributed to the molecular composition of the administered vaccines.

Full implementation of SARS-CoV-2 vaccines is a major goal facing the COVID-19 pandemic. A comparative analysis of COVID-19 vaccine characteristics, adverse events, efficacy and effectiveness reported that all vaccines up to 22nd September 2021 appeared to be safe and effective tools against all variants of concern to prevent severe COVID-19, hospitalization, and death. However, the evidence varies greatly depending on the vaccines considered (64). In addition, the effectiveness of BNT162b2/Comirnaty vaccine (Pfizer) against the Omicron variant has been reported as 60% (65). Conversely, the effectiveness of the Spikevax/mRNA-1273 vaccine (Moderna) was published for symptomatic and asymptomatic cases, without specifying effectiveness for different clades of SARS-CoV-2 (66–68). In this line, in our study, the set of sequences associated with vaccine failure were mainly related to nucleic acid-based vaccines developed by BioNTech-Pfizer and Moderna-Lonza. However, Omicron cases where the vaccine exhibited a protective effect showed a high percentage of the Pfizer brand.

Investigation of the physicochemical attributes of amino acids enables an understanding of the intricate dynamics between viral proteins and the host immune system. This exploration provides valuable insights into viral pathogenicity, contributes to vaccine design and shapes strategies for drug development. In our study a significant number of substitutions, evaluated by conservation scores, showed a robust conservation index, indicating a strong correlation with amino acid physicochemical properties in approximately one-third of the changes. Furthermore, a comparable proportion of these substitutions exhibited lower conserved similarities in both datasets, implying an equal prevalence of such disparities among substitutions analyzed using conservation scores. Analysis of the physicochemical properties of amino acid changes revealed a predominant occurrence of hydrophobic and polar amino acid substitutions in both datasets. Substitutions in the Omicron variant (clades 21L, 22B/22E, and 22F/23A) were predominantly characterized by hydrophobic and polar properties. However, the non-defining mutations of each Omicron clade were mainly polar and small. The 21J (Delta) clade sequence set featured mainly amino acid substitutions with hydrophobic properties. The non-defining mutations of 21J (Delta) were distinguished by prevalent hydrophobic and aliphatic amino acid substitutions. This is agreed with the research that showed the significant role of hydrophobic residues in the spike protein, enhancing interactions in the Delta variant (69). Recently, it was highlighted that the defining mutations in the Delta and Omicron variants markedly impact hydrophobicity, polarity, and charge distribution in all regions of the N-protein (70).

The role of missense mutations and deletions in SARS-CoV-2 has been recognized as pivotal for vaccine effectiveness and residue interactions, highlighting the need to elucidate the molecular basis of these substitutions and deletions for advancing vaccine and drug development. Our investigation identified six mutations significantly associated with reduced vaccine coverage, such as P2046L (ORF1a), P2287S (ORF1a), L3201F (ORF1a), L829I (ORF1b), R346T (S), and Q9L (N), along with the four deletions A27- (S), I68- (S), R158- (S), and G30- (N). Analysis of whole proteome sequences of SARS-CoV-2 derived from COVID-19 patients revealed a correlation between non-clade-defining mutations and vaccine effectiveness. Currently approved vaccines primarily target the spike protein. Thus, changes in this protein could challenge vaccine effectiveness. Our findings support a proteome perspective in SARS-CoV-2 vaccine design, which could improve vaccine effectiveness. In addition, we found that amino acid substitutions exhibited predominantly hydrophobic and polar properties. Understanding the physicochemical properties of amino acid substitutions is crucial, as it reveals how these modifications affect protein structure, function, and interactions. This understanding provides valuable insights into disease mechanisms and the identification of potential therapeutic targets.