Potential Role of APOBEC3 Family Proteins in SARS-CoV-2 Replication

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has acquired multiple mutations since its emergence. Analyses of the SARS-CoV-2 genomes from infected patients exhibit a bias toward C-to-U mutations, which are suggested to be caused by the apolipoprotein B mRNA editing enzyme polypeptide-like 3 (APOBEC3, A3) cytosine deaminase proteins. However, the role of A3 enzymes in SARS-CoV-2 replication remains unclear. To address this question, we investigated the effect of A3 family proteins on SARS-CoV-2 replication in the myeloid leukemia cell line THP-1 lacking A3A to A3G genes. The Wuhan, BA.1, and BA.5 variants had comparable viral replication in parent and A3A-to-A3G-null THP-1 cells stably expressing angiotensin-converting enzyme 2 (ACE2) protein. On the other hand, the replication and infectivity of these variants were abolished in A3A-to-A3G-null THP-1-ACE2 cells in a series of passage experiments over 20 days. In contrast to previous reports, we observed no evidence of A3-induced SARS-CoV-2 mutagenesis in the passage experiments. Furthermore, our analysis of a large number of publicly available SARS-CoV-2 genomes did not reveal conclusive evidence for A3-induced mutagenesis. Our studies suggest that A3 family proteins can positively contribute to SARS-CoV-2 replication; however, this effect is deaminase-independent.

SARS-CoV-2 has been evolving continually since its emergence in late 2019 [38][39][40][41][42][43][44].C-to-U mutations are among the most frequent mutations accumulated in the SARS-CoV-2 genome [39][40][41]43,44].This has been speculated to be due to the deaminase activity of APOBEC proteins.These mutations are thought to produce viruses that are more infectious and evade adaptive immunity.Recently, it has been reported that A3A protein is the potential source of C-to-U mutations in the SARS-CoV-2 genome [40].Another in vitro study focusing on APOBEC1 (A1), A3A, and A3G proteins suggested that A1 and A3A proteins (and, to a much lesser extent, A3G protein) have the capacity to mutate the SARS-CoV-2 genome; however, these mutations do not impact viral replication [45].Unexpectedly, the expression of these editing enzymes promoted SARS-CoV-2 replication and propagation [45].The authors suggested that APOBEC-induced mutations may provide a fitness advantage.Despite these illuminating studies, the role of A3 family proteins in SARS-CoV-2 replication is yet to be fully determined.
Figure 1.Validation of expression levels for ACE2 protein and A3 mRNAs in THP-1 parent and A3A-to-A3G-null THP-1 cells.(A) ACE2 protein expression levels on the surface of THP-1 parent and THP-1#11-4 (A3A-to-A3G-null THP-1) cells.The ACE2 gene was introduced by a retroviral vector, and the expression levels of the surface ACE2 protein were detected by an anti-ACE2 polyclonal antibody (red).The number in each graph shows the percentage of ACE2 + cells compared to those stained by isotype control (gray).(B) A3 mRNA expression levels in Calu-3 (gray), THP-1-ACE2 (blue), and THP-1-ACE2#11-4 (red) cells.A3 mRNA expression levels were quantified by RT-qPCR and normalized to TBP mRNA levels.Each bar represents the average of three independent experiments with Standard deviation (SD).Statistical significance was determined using the two-sided unpaired t-test.*, p < 0.05 compared to THP-1 parent cells.

Effect of A3 Family Proteins on SARS-CoV-2 Replication
We next investigated viral replication in THP-1 parent and THP-1#11-4 cells stably expressing ACE2 protein.We did not observe viral (Wuhan, BA.1, and BA.5) replication in cells lacking ACE2 protein expression, indicating that ACE2 is required for SARS-CoV-2 replication in THP-1 cells (Figure 2A).However, the replication of these variants was comparable between THP-1 parent and THP-1#11-4 cells during the 96-h time course (Figure 2A).The viral replication assay with the same time course was repeated until passage 5 (~20 days).In the THP-1 parent, the Wuhan, BA.1, and BA.5 variants showed continuous viral replication until passage 5 (Figure 2B).Surprisingly, the lack of A3A to A3G genes in THP-1 diminished the replication of all variants, and their viral RNAs became Validation of expression levels for ACE2 protein and A3 mRNAs in THP-1 parent and A3A-to-A3G-null THP-1 cells.(A) ACE2 protein expression levels on the surface of THP-1 parent and THP-1#11-4 (A3A-to-A3G-null THP-1) cells.The ACE2 gene was introduced by a retroviral vector, and the expression levels of the surface ACE2 protein were detected by an anti-ACE2 polyclonal antibody (red).The number in each graph shows the percentage of ACE2 + cells compared to those stained by isotype control (gray).(B) A3 mRNA expression levels in Calu-3 (gray), THP-1-ACE2 (blue), and THP-1-ACE2#11-4 (red) cells.A3 mRNA expression levels were quantified by RT-qPCR and normalized to TBP mRNA levels.Each bar represents the average of three independent experiments with Standard deviation (SD).Statistical significance was determined using the two-sided unpaired t-test.*, p < 0.05 compared to THP-1 parent cells.

Effect of A3 Family Proteins on SARS-CoV-2 Replication
We next investigated viral replication in THP-1 parent and THP-1#11-4 cells stably expressing ACE2 protein.We did not observe viral (Wuhan, BA.1, and BA.5) replication in cells lacking ACE2 protein expression, indicating that ACE2 is required for SARS-CoV-2 replication in THP-1 cells (Figure 2A).However, the replication of these variants was comparable between THP-1 parent and THP-1#11-4 cells during the 96-h time course (Figure 2A).The viral replication assay with the same time course was repeated until passage 5 (~20 days).In the THP-1 parent, the Wuhan, BA.1, and BA.5 variants showed continuous viral replication until passage 5 (Figure 2B).Surprisingly, the lack of A3A to A3G genes in THP-1 diminished the replication of all variants, and their viral RNAs became undetectable in passage 5 (Figure 2B).These data suggest that A3A to A3G proteins may be associated with long-term (~20 days) SARS-CoV-2 replication in THP-1 cells.

Effect of A3 Family Proteins on SARS-CoV-2 Infectivity
As mentioned above, the viral RNA of all variants tested became undetectable by 20 days postinfection, in contrast to that observed in THP-1 parent cells (Figure 2B).To know whether the Wuhan variant obtained from each passage was infectious, we performed a plaque assay in VeroE6/TMPRSS2 cells (Figure 3).Consistent with viral replication results (Figure 2B), the number of PFU obtained from the Wuhan variant infection in THP-1 parent cells was increased during the passage experiments (Figure 3).However, A3A-to-A3G gene disruption caused a decrease in viral infectivity, and it finally became undetectable (Figure 3).These data suggest that A3A to A3G proteins may contribute positively to the production of SARS-CoV-2 infectious particles from THP-1 cells.

Effect of A3 Family Proteins on SARS-CoV-2 Infectivity
As mentioned above, the viral RNA of all variants tested became undetectable by 20 days postinfection, in contrast to that observed in THP-1 parent cells (Figure 2B).To know whether the Wuhan variant obtained from each passage was infectious, we performed a plaque assay in VeroE6/TMPRSS2 cells (Figure 3).Consistent with viral replication results (Figure 2B), the number of PFU obtained from the Wuhan variant infection in THP-1 parent cells was increased during the passage experiments (Figure 3).However, A3A-to-A3G gene disruption caused a decrease in viral infectivity, and it finally became undetectable (Figure 3).These data suggest that A3A to A3G proteins may contribute positively to the production of SARS-CoV-2 infectious particles from THP-1 cells.

Effect of A3 Proteins on SARS-CoV-2 Mutagenesis
Finally, we asked whether A3 proteins contribute to C-to-U mutations in the SARS-CoV-2 genome.We performed whole-genome sequencing (WGS) analysis for viral RNA isolated from each passage to address this question.Our analysis revealed mutations in seven positions, only two of which were C-to-T mutations, but none were in the TCA or

Effect of A3 Proteins on SARS-CoV-2 Mutagenesis
Finally, we asked whether A3 proteins contribute to C-to-U mutations in the SARS-CoV-2 genome.We performed whole-genome sequencing (WGS) analysis for viral RNA isolated from each passage to address this question.Our analysis revealed mutations in seven positions, only two of which were C-to-T mutations, but none were in the TCA or TCT contexts, known as APOBEC targets (Figure 4).Additionally, there was no mutational burden difference between the cells with and without A3A to A3G genes (Figure 4).These results indicate that A3 proteins do not play a role in SARS-CoV-2 mutagenesis.To investigate whether the lack of A3-induced mutagenesis in SARS-CoV-2 is specific to THP-1 cells or is a general feature of SARS-CoV-2 infection, we conducted a bioinformatics analysis of 40,000 whole-genome SARS-CoV-2 sequences from NCBI.Our analysis involved the quantification of all possible 192 types of mutations followed by mutational signature deconvolution by the Non-negative Matrix Factorization (NMF).We investigated models with up to 20 mutational signatures and could not find any mutational signatures closely resembling single base substitution 2 (SBS2) [represented by TC(T/A)-to-TT(T/A)] or SBS13 [represented by TC(T/A)-to-TG(T/A)] as shown in Figure 5.However, we noted two signatures (S8 and S9), each represented by only one of the main SBS2 peaks (TCA-to-TTA in S8 and TCT-to-TTT in S9).Since such signatures containing only a single dominant peak have not been previously reported for any A3 enzymes, further studies are needed to provide evidence that these signatures are the footprint of A3 enzymes.

Discussion
The role of A3 enzymes in SARS-CoV-2 mutations has been implicated in several studies [39][40][41]43,44]; however, little is known about the functional relevance of A3 proteins in SARS-CoV-2 infection.In this study, we examined the effect of A3 proteins on SARS-CoV-2 replication by conducting infection in THP-1 cells lacking A3 enzymes.We showed that the Wuhan, BA.1, and BA.5 variants had comparable viral RNA production in THP-1-ACE2 parent and THP-1-ACE2 cells lacking the expression of A3A to A3G pro-

Discussion
The role of A3 enzymes in SARS-CoV-2 mutations has been implicated in several studies [39][40][41]43,44]; however, little is known about the functional relevance of A3 proteins in SARS-CoV-2 infection.In this study, we examined the effect of A3 proteins on SARS-CoV-2 replication by conducting infection in THP-1 cells lacking A3 enzymes.We showed that the Wuhan, BA.1, and BA.5 variants had comparable viral RNA production

Discussion
The role of A3 enzymes in SARS-CoV-2 mutations has been implicated in several studies [39][40][41]43,44]; however, little is known about the functional relevance of A3 proteins in SARS-CoV-2 infection.In this study, we examined the effect of A3 proteins on SARS-CoV-2 replication by conducting infection in THP-1 cells lacking A3 enzymes.We showed that the Wuhan, BA.1, and BA.5 variants had comparable viral RNA production in THP-1-ACE2 parent and THP-1-ACE2 cells lacking the expression of A3A to A3G proteins during 96 h of the time course (Figure 2A).However, A3 family proteins affected SAR-CoV-2 RNA production in the passage experiments for up to 20 days (Figure 2B).Further, the plaque assay results showed that A3 family proteins might contribute to the production of infectious virus particles in THP-1 cells (Figure 3).Notably, the effect of A3 family proteins on SARS-CoV-2 replication is independent of C-to-U mutations (Figure 4).Taken together, our findings suggest that A3 family proteins may influence SARS-CoV-2 replication in a deaminase-independent manner.
THP-1 cells were derived from the blood of a male patient with acute monocytic leukemia [72].In this study, we used THP-1 cells obtained from Dr. Cimarelli [73].However, it is important to note that multiple sublines of THP-1 exist, and the potential impact of this intra-cell line heterogeneity on the results presented here requires further investigation.THP-1 cells do not express ACE2 and TMPRSS2 and are not natural targets of SARS-CoV-2 (Figures 1A and 2A).It should be noted that adding TMPRSS2 expression to THP-1-ACE2 cells may give a different result that C-to-U mutations are induced on the SARS-CoV-2 genome.Therefore, the phenomenon observed in this study might be limited to THP-1 cells.
In summary, we found that A3 family proteins support SARS-CoV-2 replication in THP-1 independently of their enzymatic activity.Understanding the underlying mechanism may provide new insight into the interaction between A3 family proteins and coronaviruses.Further, A3 family proteins may be a potential therapeutic target for drug development to alleviate disease severity in respiratory diseases caused by coronaviruses.
Virus propagation was performed as previously described [75,76,[81][82][83].Briefly, VeroE6/TMPRSS2 cells (5 × 10 6 cells) were seeded in a T-75 flask the day before infection.The virus was diluted in virus dilution buffer [1M HEPES, DMEM (low glucose), Nonessential Amino acid (Gibco, Waltham, MA, USA Cat# 11140-050), 1% P/S], and the dilution buffer containing the virus was added to the flask after removing the initial medium.After 1 h of incubation at 37 • C, the supernatant was replaced with 15 mL of 2% FBS/DMEM (low glucose) and cell culture was continued to incubate at 37 • C until visible cytopathic effect (CPE) was clearly observed.Then, cell culture supernatant was collected, centrifuged at 300× g for 10 min and frozen at −80 • C as working virus stock.The titer of the prepared working virus was determined as the 50% tissue culture infectious dose (TCID 50 ) [78,82,84].The day before infection, VeroE6/TMPRSS2 cells (10,000 cells) were seeded in a 96-well plate and infected with serially diluted working virus stocks.The infected cells were incubated at 37 • C for 4 days, and CPEs were observed in the infected cells by a microscope.The value of TCID 50 / mL was calculated by the Reed-Muench method [85].

SARS-CoV-2 Infection
5 × 10 5 parent and A3A-to-A3G-null THP-1 cells were seeded into a 24-well plate, inoculated with SARS-CoV-2 (5000 TCID 50 ) and incubated at 37 • C for 1 h.After washing with phosphate-buffered saline (PBS), 1 mL of fresh cell culture medium was added.15 µL of cell culture supernatant was harvested at the indicated timepoints and used for RT-qPCR to quantify the viral RNA copy number (see "RT-qPCR" section) (Figure 2A,B).For passage experiments, 10 µL of the cell culture supernatant at 96 h postinfection was transferred to the next target cells.The passage of the cell culture supernatant was repeated until passage 5.

SARS-CoV-2 WGS
The WGS of the SARS-CoV-2 RNA genome was performed as previously described [95].Briefly, cDNA synthesis, viral sequence enrichment, library amplification, and indexing were performed using the QIAseq DIRECT SARC-CoV-2 kit (QIAgen, Hilden, Germany, Cat# 333891) according to the manufacturer's protocol.After multiplexing with QIASeq DIRECT UDI Set-A (QIAgen), a 25 µL library was prepared from each sample.The quality of the enriched libraries was evaluated by electrophoresis using the TapeStation 4150 system (Agilent Technologies, Santa Clara, CA, USA).The prepared libraries were subjected to sequencing using MiSeq reagent Micro and Nano Kits (Version 2, 300 cycles) in the MiSeq desktop sequencing system (Illumina, San Diego, CA, USA).The data analysis was done as previously performed [95].

SARS-CoV-2 Mutational Signature Analysis
We downloaded 40,000 whole-genome SARS-CoV-2 sequences from NCBI.These sequences were sampled between the years 2019 and 2022 and were highly diverse in terms of geographic distribution.A total of 38,830 of these sequences that had less than 5% insertion/deletion and/or non-A/C/G/T, were selected and aligned to the most common ancestor of SARS-CoV-2 as described by Kumar et al. [96].The frequencies of all the possible 192 mutation types (NnN-to-NmN where n mutates to m and N:A/C/G/T) were quantified for each sequence.These mutation counts were organized in a data matrix of (38,830 by 192) and used as input for analysis using the Non-negative Matrix Factorization (NMF) method.NMF is a matrix decomposition method used routinely to deconstruct mutational signatures in cancer and viral genomes [97-100].Here, NMF was used to investigate if the deconstruction of all SARS-CoV-2 sequences provides evidence for the existence of an APOBEC mutational signature similar to the known SBS2 and SBS13 signatures [97].We built NMF models with up to 20 components to investigate the presence of SBS2 and SB13.

Conclusions
Our results suggest that A3 family proteins may contribute to SARS-CoV-2 replication in THP-1 cells.This effect of A3 family proteins on SARS-CoV-2 replication is independent of deaminase activity.

Figure 1 .
Figure 1.Validation of expression levels for ACE2 protein and A3 mRNAs in THP-1 parent and A3A-to-A3G-null THP-1 cells.(A) ACE2 protein expression levels on the surface of THP-1 parent and THP-1#11-4 (A3A-to-A3G-null THP-1) cells.The ACE2 gene was introduced by a retroviral vector, and the expression levels of the surface ACE2 protein were detected by an anti-ACE2 polyclonal antibody (red).The number in each graph shows the percentage of ACE2 + cells compared to those stained by isotype control (gray).(B) A3 mRNA expression levels in Calu-3 (gray), THP-1-ACE2 (blue), and THP-1-ACE2#11-4 (red) cells.A3 mRNA expression levels were quantified by RT-qPCR and normalized to TBP mRNA levels.Each bar represents the average of three independent experiments with Standard deviation (SD).Statistical significance was determined using the two-sided unpaired t-test.*, p < 0.05 compared to THP-1 parent cells.

Figure 2 .
Figure 2. SARS-CoV-2 replication in THP-1 parent and A3A-to-A3G-null THP-1 cells.(A) Replication kinetics of the Wuhan, BA.1, and BA.5 variants produced from THP-1 parent and THP-1#11-4 (A3A-to-A3G-null THP-1) cells without (blue line) or with (red line) ACE2 protein expression.The SARS-CoV-2 N gene was quantified by RT-qPCR to monitor the viral RNA copy number across the indicated time points.Each timepoint represents the average of four independent experiments with SD. (B) Passage experiments.The SARS-CoV-2 N gene in the cell culture supernatants produced from THP-1-ACE2 or THP-1-ACE2#11-4 (A3A-to-A3G-null THP-1) cells at 96 h postinfection of each passage were quantified by RT-qPCR to monitor the viral RNA copy number of the Wuhan (gray), BA.1 (blue), and BA.5 (red) variants.Each bar represents the average of three independent experiments with SD.

Figure 2 . 15 Figure 3 .
Figure 2. SARS-CoV-2 replication in THP-1 parent and A3A-to-A3G-null THP-1 cells.(A) Replication kinetics of the Wuhan, BA.1, and BA.5 variants produced from THP-1 parent and THP-1#11-4 (A3A-to-A3G-null THP-1) cells without (blue line) or with (red line) ACE2 protein expression.The SARS-CoV-2 N gene was quantified by RT-qPCR to monitor the viral RNA copy number across the indicated time points.Each timepoint represents the average of four independent experiments with SD. (B) Passage experiments.The SARS-CoV-2 N gene in the cell culture supernatants produced from THP-1-ACE2 or THP-1-ACE2#11-4 (A3A-to-A3G-null THP-1) cells at 96 h postinfection of each passage were quantified by RT-qPCR to monitor the viral RNA copy number of the Wuhan (gray), BA.1 (blue), and BA.5 (red) variants.Each bar represents the average of three independent experiments with SD.Viruses 2024, 16, x FOR PEER REVIEW 5 of 15

Figure 5 .
Figure 5. Analysis of mutational signatures in publicly available SARS-CoV-2 genomes.All possible 192 mutation types (NnN-to-NmN where n mutates to m and N:A/C/G/T) were quantified in a total of 38,830 whole-genome SARS-CoV-2 sequences sampled between the years 2019 and 2022 and reported in NCBI.These mutation counts were organized in a data matrix of (38,830 by 192) and used as input for analysis using the NMF method.Models with up to 20 components were built.Only a model with 10 components is shown for simplicity.None of the models showed signatures closely related to SBS2 and SBS13.Positions of the four major C>T and C>G peaks in SBS2 and SBS13 are shown by arrows.

Figure 5 .
Figure 5. Analysis of mutational signatures in publicly available SARS-CoV-2 genomes.All possible 192 mutation types (NnN-to-NmN where n mutates to m and N:A/C/G/T) were quantified in a total of 38,830 whole-genome SARS-CoV-2 sequences sampled between the years 2019 and 2022 and reported in NCBI.These mutation counts were organized in a data matrix of (38,830 by 192) and used as input for analysis using the NMF method.Models with up to 20 components were built.Only a model with 10 components is shown for simplicity.None of the models showed signatures closely related to SBS2 and SBS13.Positions of the four major C>T and C>G peaks in SBS2 and SBS13 are shown by arrows.

Figure 5 .
Figure 5. Analysis of mutational signatures in publicly available SARS-CoV-2 genomes.All possible 192 mutation types (NnN-to-NmN where n mutates to m and N:A/C/G/T) were quantified in a total of 38,830 whole-genome SARS-CoV-2 sequences sampled between the years 2019 and 2022 and reported in NCBI.These mutation counts were organized in a data matrix of (38,830 by 192) and used as input for analysis using the NMF method.Models with up to 20 components were built.Only a model with 10 components is shown for simplicity.None of the models showed signatures closely related to SBS2 and SBS13.Positions of the four major C>T and C>G peaks in SBS2 and SBS13 are shown by arrows.