It is well known that ORF7b and ORF8 are located in a genome region characterized by a high mutation rate and thus considered a mutational hotspot. Genetic mutations in this region might correlate with zoonotic events and pandemic waves [
50,
51]. SARS-CoV-2 ORF7b encodes for a 43 aa protein showing more than 80% similarity with its SARS-CoV-1 homolog [
29], and it plays an important role in interferon downregulation [
52]. Its structure has not been fully elucidated, but it has been suggested that this protein could be considered a single-pass transmembrane protein acting as viroporin with a multimeric structure [
31,
53]. The SARS-CoV-2 ORF8 gene is common to several SARS-related coronaviruses [
54], but it is highly divergent from its homolog in SARS-CoV-1 [
55]. The ORF8 gene encodes for a 121 aa accessory protein, with an N-terminal transmembrane peptide followed by an Ig-like domain [
23]. The ORF8 protein is expressed at the endoplasmic reticulum (ER) level, where it causes ER stress [
24], and its dimeric form elicits a pro-inflammatory role, stimulating the IL-17 receptor [
25,
26]. A recent study [
56] demonstrated that ORF8 can induce ER reshaping by its involvement in the formation of mixed disulfide complexes with ER proteins. This remodeling is advantageous for ORF8 to escape from degradation, and it activates ER stress, alters ER homeostasis, and accelerates protein trafficking through the ER. Moreover, ORF8 is also involved in the disruption of epigenetic regulation via histone mimicry and the degradation of MHC-I [
27]. Overall, ORF8 appears to be widely involved in immune evasion by suppressing the type 1 interferon antiviral response [
24] and deregulating the TGF-β pathway [
54]. It was speculated that the acquisition of a new ORF8 via recombination could play a consistent role in transmission from animals to humans [
57]. A discrete number of genomic deletions in ORF8, ranging from 1 to 382 nt [
55,
57,
58], as well as truncated forms [
21,
58], have been reported for SARS-CoV-2 (see graphical representation in
Supplementary Figure S1). Partial or complete loss of ORF8 was also reported for SARS-CoV-1, with some sequences exhibiting a 415 nt deletion similar to our 426 deletion [
59], and such losses could be involved in the zoonotic transition, favoring an adaptive change [
59]. Half of the mutations within the ORF8 gene, identified in the first 6 months of the pandemic, were detrimental for the encoded protein [
60]. More frequent mutations include nonsense mutations (for instance, Q27*) [
21] and deletion events, some of which completely abolished the expression of the gene product [
20,
61]; in addition, extensive deletion might involve upstream ORF7b and, in some cases, even ORF7a [
62]. In SARS-CoV-2, ORF8 is involved in many processes and seems to be far from dispensable in vivo [
22]. Massive deletion events do not represent a novelty for the Coronaviridae because this phenomenon was already described for the homolog in the variant of SARS-CoV-1 at the end of the epidemic era [
59]. Interestingly, a study of a deletion of 382 nt (Δ382) showed that this mutation produces a truncated form of ORF7b, abolishing its activity [
59]. According to the literature, hairpins in the ORF8 transcript region could play a role in genomic rearrangements during viral genome replication [
57]. The so-called Δ382 variant, first described in Singapore in January/February 2020, was reported to induce milder clinical symptoms compared to the wild-type SARS-COV-2 strains [
20,
21]. In this paper, we describe a consistent deletion found in 6 samples from infected patients that truncates ORF7b, abolishes the ORF8 transcripts, and removes the TRS located between the ORF7b and ORF8 sequences. The analysis with Swiss-Model [
63] and Protter [
64] of a hybrid protein generated by 426del suggests a single alpha-helix secondary structure and transmembrane localization. The truncated peptide seems to retain the transmembrane domain, although in a curiously inverted configuration (
Supplementary Figure S2), with the N-terminal inside the membrane. Due to a lack of knowledge about the ORF7b protein and the small dimension of our hybrid, its role in SARS-CoV-2 infected patients carrying the 426 deletion remains unknown. Liu and coworkers [
56] demonstrated that use of reducing agents supported the release of ORF8 from mixed disulfide complexes, favoring its degradation and resolution of ER stress. Our 6 cases showed particular clinical and viral genome characteristics: all cases shared the BA.5.1 variant, confirming some similar observations from Germany and Switzerland; and 3 of 6 patients displayed the presence of immunodeficiency or autoimmune disease, whereas the remaining 3 cases did not show evident immunological disorders. Immune system derangement generally allows for a larger number of in vivo viral replication cycles, thus determining a longer time of infection with the onset of more frequent genome mutations or deletions. Strikingly, our 6 cases exhibited a persistence of infection even in immunocompetent patients. The absence of ORF7b and ORF8 might enable the deleted mutation to escape from the interferon response and innate immunity; however, we have no information about the replication and transmission rate effectiveness of our mutations. Although variants with ORF7b and ORF8 region deletions have been described to different extents during the infection waves, the number of sequenced ORF7b/8 deleted strains is very small, and in some cases, these deletions were present in the same patient, with the full-length variant suggesting evolution of a viral strain [
25]. In addition, we should take into account that these large deletions are considered to be rare, not because they do not appear frequently, but several sequencing pipelines are not able to automatically detect these deletions, thus determining an underestimation of phenomenon. Starting from this observation and according to the studies by Zinzula and DeRonde and coworkers [
58,
65], we speculate that this deletion might be considered a possible pathway for viral evolution, which appeared during the succession of the different variants as an alternative route for viral spreading, although its development and evolutionary success might be overcome by the continuous and rapid appearance of new variants with better transmissibility and replication. Moreover, it should be noted that the low frequency of this deletion could be related to other cases of functional abolishment of ORF8 (as the already mentioned stop codon Q27* and the more recently appeared Q18* and G8*). Hisner and colleagues [
66] have also noticed that the BA.5 variants are characterized by a mutation in the TRS (C27889T) that probably interferes with a proper generation of the ORF8 subgenomic RNA and thus with its expression. Interestingly, Hisner [
66] hypothesized an evolutionary trend which moves towards a functional disappearance of ORF8 expression starting with BA.5 variants and carrying on with the XBB variants bearing the G8* mutation. Hence, it is highly probable that the 426 nt deletion was not able to overcome the competition of other variants that were de facto deleted for these gene. Notwithstanding the small impact of these mutations, the onset of ORF7b/8 deleted, or functionally deleted, mutant variants is playing a role in the evolutionary balance between host and virus.