Genomic sequences of the strains belonging to Fna were analyzed to identify a representative strain on which further analysis can be performed. Nucleotide sequences of 65 genomes of Fna deposited under BioProject PRJNA549513 (Zepeda-Rivera et al., 2024) were downloaded from the National Center for Biotechnology Information (NCBI) Nucleotide database (https://www.ncbi.nlm.nih.gov/nucleotide/) in nucleotide FASTA format. In order to select a representative Fna strain (implicated in CRC) for further analysis, Fna Clade 2 (Fna C2) was considered owing to its enrichment in CRC tumors (Zepeda-Rivera et al., 2024). The nucleotide sequences of the strains belonging to Fna C2 were analyzed using the FastANI tool (https://github.com/ParBLiSS/FastANI) in order to evaluate pairwise average nucleotide identity (ANI) and subsequent selection of the representative strain.

Amino acid sequences corresponding to proteome of the selected strain were downloaded from the NCBI Genome database in FASTA format. Publicly available tools with distinct algorithms were used for the identification of potential cNLS on proteins. These include NLStradamus (Nguyen Ba et al., 2009) (http://www.moseslab.csb.utoronto.ca/NLStradamus/), which operates on a simple HMM (Hidden Markov Model) algorithm; NucImport (Mehdi et al., 2011) (UQ eSpace), which is a combination of three Bayesian network models; and NucPred (Brameier et al., 2007) (https://nucpred.bioinfo.se/nucpred/), which utilizes regular expressions from NLSs and assigns a likelihood score using an ensemble model consisting of over 100 sequence-based predictors. All these tools provided scores in the range of 0 to 1, and the values were noted for each protein in the proteome of Fna. In order to arrive at a cutoff score that might be a good indicator of an actual cNLS, we utilized these tools for assessment of NLS sequences identified in experimentally characterized bacterial NMs ( Supplementary Table 1 ). The results indicated that at least one of the three tools yielded a score of ≥0.6 for most of the experimentally characterized bacterial NMs with a cNLS ( Supplementary Table 1 ). Thus, we have considered a cutoff score of 0.6, in at least one tool, to predict a probable set of NMs. A subset of the predicted NMs with a more stringent prediction score of ≥0.7 was further considered for detailed analysis. This subset of proteins was analyzed using NCBI BLASTp (https://blast.ncbi.nlm.nih.gov/Blast.cgi) for assessing their presence within the Fn genomes and across other bacteria, and analyzing their NLS sequence conservation.

The structure PDB ID 5HUY was obtained from RCSB Protein Data Bank (https://www.rcsb.org/) (RCSB PDB: homepage), which corresponds to an importin α bound to the NLS of a small terminase from HCMV. The modeled structures for the identified candidate NM proteins in Fusobacteria were obtained from the AlphaFold Protein Structure Database (https://alphafold.com/) (Varadi et al., 2024; Varadi et al., 2022; Jumper et al., 2021). Structures that were unavailable in the database were modeled using the AlphaFold monomer model with the DeepMind Colab Jupyter notebook with default settings (Jumper et al., 2021). The binding interaction between candidate proteins (in Fna possessing NLS) and importin α was analyzed using the HADDOCK 2.4 (High Ambiguity Driven protein–protein DOCKing) (Honorato et al., 2024) (https://rascar.science.uu.nl/haddock2.4/) web server with default parameters. HADDOCK deploys ab initio docking methods by incorporating data from known or predicted protein interfaces into ambiguous interaction restraints (AIRs) to guide the docking process. Both proteins were submitted in PDB format in the HADDOCK interface specifying the constraints for docking including active residues of importin α based on experimental sources and the NLS residues of candidate proteins identified in previous steps. The pocket residues from the major groove of importin α provided as constraints were Ser105, Trp142, Asn146, Ser149, Gly150, Thr155, Q181, Trp184, Asn188, Gly191, Asp192, Trp231, Asn228, Asn235, Arg238, Asn239, and Trp273. Furthermore, the modeled complexes were shortlisted based on the HADDOCK score, which accounts for the weighted sum of various energy components that differ with respect to each docking stage: rigid body (it0), semi-flexible refinement (it1), and explicit solvent refinement (water). The default values considered are as follows:

where Evdw = van der Waals intermolecular energy, Eelec = electrostatic intermolecular energy, Edesol = desolvation energy, Eair = distance restraints energy [only unambiguous and AIR (ambig) restraints], and BSA = buried surface area (https://rascar.science.uu.nl/haddock2.4/). Following the visualization of the top four scoring complexes in https://www.pymol.org/, the most optimal complex was chosen for detailed analysis. These structures were further evaluated by calculating contact distances (default cutoff, 5 Å) using MAPIYA (Badaczewska-Dawid et al., 2022) (https://mapiya.lcbio.pl/). Furthermore, to investigate the probable impact of mutation on the interaction between importin α and RBPL34, the conserved Asp192 was substituted by Ala. The mutated structure was then modeled using Swiss-Model (https://swissmodel.expasy.org/) (Waterhouse, 2018) with the structure corresponding to Mus musculus importin alpha2 (PDB ID: 3TPO) as a template.

In order to identify potential nucleomodulins in Fna, a representative genome was selected based on the ANI metric. ANI values below 90% indicate that the genomes are from different species, while values greater than 95% suggest that the genomes belong to the same species (Rodriguez-R et al., 2023). As mentioned earlier, the reclassification of Fn subspecies has been proposed, as the ANI values when compared between subspecies are below 95%, indicating significant genomic heterogeneity (Kook et al., 2017). This is evident in the case of Fna genomes including Clade 1 (C1) and Clade 2 (C2), which generated ANI values ranging from 91% to 99% when compared pairwise using FastANI. Based on these values, Fna strain SB031 (2,432,266 bp, GenBank Assembly ID: GCA_037897425.1), isolated from CRC tissue, which exhibited ANI > 96% with other members of Fna C2, was selected for subsequent analysis ( Supplementary Figure 1 ).

As discussed in the Methods section, multiple tools are available to identify cNLS within a protein (NLStradamus; Brameier et al., 2007; Mehdi et al., 2011). NLStradamus, Nucimport, and NucPred were utilized for the prediction of cNLS to identify potential NMs within the proteome (2,204 proteins) of Fna SB031. The results showed the presence of potential cNLS in 330 proteins with a score ≥0.6 in one or more tools (scores from 3 tools and predicted NLS sequence details are provided in Supplementary Table 2 ).

We selected the five proteins that showed a score greater than 0.7 for all the three tools for a detailed analysis of their probable interactions with the importin protein ( Table 1 ). Additionally, we also analyzed the secondary structure of the experimentally characterized NMs and their corresponding cNLSs using JPred4 (https://www.compbio.dundee.ac.uk/jpred/) (Drozdetskiy et al., 2015) in order to understand the potential structural features that bacterial cNLSs can adopt, and the results are provided in Supplementary Figure 2 . A detailed sequence and structural analysis on NLS of the predicted NMs is provided below.

The results of NLS prediction tools indicated the set of residues 26–44 as a potential cNLS NGRKVLKRRRVRGRAKLSA in ribosomal protein L34 (RBPL34) from Fusobacterium. The sequence is conserved across the genomes of all strains of F. nucleatum subsp. animalis, F. nucleatum subsp. polymorphum (Fnp), and F. nucleatum subsp. vincentii (Fnv) often implicated in CRC and oral diseases. Furthermore, a sequence-based search for homologs within complete bacterial genomes indicated that the same NLS is not observed in F. varium, F. necrophorum, and F. ulcerans. Interestingly, on further analysis, it was found that the C-terminal region of RBPL34 from these species also shows a high score for the presence of NLS within their sequences. Therefore, these proteins may also show nuclear localization probability but utilizing a different cNLS sequence. Analysis of RBPL34 homologs in bacteria other than Fusobacterium indicated that despite a high sequence similarity in the N-terminal region, the C-terminal part (comprising the identified NLS sequence in Fusobacterium) of the sequences shows divergence and, in some cases, the presence of a different NLS sequence from the one observed in Fna. Earlier studies show a higher abundance of Fusobacterium subspecies like Fnn, Fnv, and Fnp in these oral disease conditions and indicate their role in influencing the host disease microenvironment and association with changes in gene expression and signaling pathways (Chen et al., 2022; Galaski et al., 2024; Wolf et al., 2025). The presence of cNLS in RBPL34 from other subspecies of Fusobacterium indicates that they might play a role in influencing gene expression in oral conditions like plaque formation, periodontitis, and oral cancer. Additionally, experimental studies have also implicated certain strains of subspecies Fnn in the development and manifestation of CRC (Abed et al., 2016; Yu et al., 2017). Thus, the presence of candidate NLS sequences in RBPL34 of these species might indicate their potential role in explaining the association of Fn with different disease mechanisms.

As mentioned above, the cNLS-containing proteins need to bind to the importin α major groove to enter the nucleus (Sankhala et al., 2016). In order to further corroborate if the identified cNLS in the protein shows a propensity to interact with importin α protein, we docked the model built using AlphaFold for RBPL34 from Fna SB031 with importin α protein (PDB ID: 5HUY) using HADDOCK 2.4 (as described in the Methods section). Most of the experimentally characterized structures available for importin only show the binding of the NLS from the candidate nuclear protein without accounting for the remaining part of the protein. Recent studies have indicated that the importin–NLS binding depends on the context within the protein in addition to the NLS itself (Friedrich et al., 2006). Thus, we chose to utilize the whole structure instead of only docking the cNLS to the importin structure. We analyzed the energy landscape and postures for RBPL34 obtained in the docking clusters and selected the best cluster for further analysis ( Supplementary Figure 3 , Supplementary Table 3 ) with a HADDOCK score of −125.53 (a score with a more negative value is considered better).

We further analyzed the obtained docked complex to check if the contacting residues between importin and RBPL34 are similar to the interactions with Trp and Asn residues as well as with Asp192 in importin as have been reported in experimentally characterized importin NLS structures (Sankhala et al., 2016). The helix of the RBPL34 protein containing the predicted NLS residues 26–44 is seen to interact with ARM repeats 2, 3, and 4 in the major groove of importin α ( Figure 2a ). Figure 2b shows the electrostatic interactions between the NLS and importin in the docked complex. As observed in the experimentally resolved structures, the conserved Asn146 and Asn188 of ARM 2 and 3 show an electrostatic/hydrogen bonding interaction with polar residues Lys29 and Arg33 in the cNLS of RBPL34. Furthermore, the conserved Asp192 in importin shows salt bridge formation with Lys25 and Lys29 on the NLS of RBPL34. Figure 2c indicates that the hydrophobic Trp array from ARM 2, 3, and 4 (corresponding to Trp142, Trp184, and Trp231) shows cation–pi interactions with the long carbon side chains of basic residues in the cNLS (including Lys29, Lys32, Arg33, Arg34, and Arg39) of RBPL34.

In order to assess the viability of the identified interactions as well as possible conformational dynamics of the importin complex with RBPL34, we performed explicit solvent MD simulations (see the Methods section). The results corresponding to radius of gyration (for both simulation replicates) showed the reduction in the first 2 ns followed by the stabilization of the Rg value ( Supplementary Figure 4A ). To further understand the changes in the secondary structure of the NLS sequence, we analyzed the trajectory using DSSP, and the results indicated that the secondary structure fluctuated between helix and loops for most residues although we observe higher occurrence of loop structures towards the end of the simulation ( Supplementary Table 4 ). These results indicated that the secondary structure of different NLS residues switches between loops and helices, indicating conformational changes in the NLS region as well ( Supplementary Figure 4B ).

On visualization, it was observed that the ARM helices rearranged, which might be conformational changes to accommodate the whole RBPL34 protein ( Figure 3 ). Earlier studies have shown that the Armadillo helices of importin show a spring like movement to accommodate cNLS and the protein cargo (Sankhala et al., 2017). Studies have further shown that the conformational changes are more prominent in case of complete cargo binding since the ARM helices 2–4 show a movement to accommodate the larger protein without steric clashes and with favorable interactions. Thus, the changes in radius of gyration as well as the overall orientation of helices observed on visualization might be an indication of these interactions.

We observed that despite the changes in importin helices, the interactions between the conserved acidic patch containing Asp192 and the basic residues remained intact. Thus, we further analyzed the change in distance between these residues to assess if the interaction between the importin and NLS remained stable across both replicates of the MD simulation. Previous studies indicate that the residue Asp192 is conserved in importin α and located in the acidic patch of ARM helix responsible for stabilizing the basic residues within the NLS. During the simulation studies, it was observed that the Asp192 formed persistent electrostatic contacts with both Lys25 and Lys29 residues (observed to interact with Asp192 as shown in Figure 2 ) of the NLS with average distances of 0.3 nm/3 Å ( Supplementary Figure 4C ). Additionally we observed the distances between the conserved Trp184 and the pi interactions observed with Arg33, Arg34, Arg35, and Arg37. Thus, the distance between these residues was calculated by utilizing the center of mass of residues 33–37 in NLS and the CE2 atoms of Trp84, and it was observed to stabilize at 0.5 nm/5 Å ( Supplementary Figure 4D ). This simulation outcome indicates that the NLS interactions with the ARM helices of importin might be stable, thereby further corroborating the prediction that RBPL34 might function as a NM. This indicates that despite the changes observed in importin when analyzed with the complete protein RBPL34, the interacting residues between NLS and importin ARM 2, 3, and 4 are retained.

In order to further corroborate the role of the conserved residue Asp192 for binding of NLS in RBPL34, we mutated the Asp192 residue to an Ala192 residue and docked and minimized the generated complex. Furthermore, this mutated complex was also subjected to an explicit solvent simulation in order to understand the stability of the docked mutated complex. The docking results showed a reduction in overall HADDOCK score from –125.53 in the unmutated complex to –67.22 in the mutated complex. The clusters obtained by this docking as well as the energy landscape has been provided in Supplementary Table 5 and Supplementary Figure 5 . Based on the analysis, we chose Cluster 2 for further simulations. The snapshots and distance calculations between the conserved importin residues as well as the RBPL34 NLS were calculated as discussed before in this complex as well. The results showed that the distances between the conserved residues and NLS increased across the simulation, and the final posture of RBPL34 in the mutated complex did not form favorable interactions with the NLS ( Supplementary Figure 6 ). The simulation of the mutated complex shows that the important electrostatic interactions as well as pi interactions between ARM repeats and NLS are not observed in the mutated complex, indicating that the predicted NLS and interactions with acidic patch on importin might be necessary for a stable complex. These results can be further corroborated by experimentally mutating the protein and elucidating the change in nuclear import, which is not part of the scope of the present manuscript. The posture of NLS sequence observed across the simulation places Arg41 close to the Ala CB atom, indicating a hydrophobic interaction. Thus, no electrostatic contacts or interactions with conserved Trp residues in ARM helices are observed in the mutated complex ( Supplementary Figure 6 , Supplementary Table 6 ).

Thus, our in silico studies indicate that the Fn protein RBPL34 possesses all features of an NM and might be capable of entering the host nucleus. In addition to our findings, other groups have also indicated the role of ribosomal proteins as NMs using experimental studies. Lu et al. (2024) observed ribosomal proteins acting as NMs in Mycoplasma (Lu et al., 2024) and indicated that these proteins may be secreted into extracellular vesicles, which can further carry and release them into host cells across different organs. The proteins in host cells can then enter the nucleus to exert effects on host gene expression (Lu et al., 2024). Furthermore, RBPL34 has also been observed in the EV proteome of Fusobacterium, indicating its transfer to host cells through extracellular vesicles. It has been suggested in recent studies that Fusobacteria also form extracellular vesicles that aid in colonization and have also been reported to be associated with CRC progression (Zheng et al., 2024). In addition, studies have also implicated eukaryotic RBPL34 in influencing cancer progression (Luo et al., 2016; Liang et al., 2023). Although these inferences from the in silico sequence and structural analysis indicate that RBPL34 in Fusobacterium might influence disease progression by entering the host cell nucleus, they can be further corroborated by conducting laboratory experiments to assess the localization of these proteins in the host.

The NLS prediction tools identified the presence of cNLS sequence as EKTGKFKKKKRFGKSL consisting of residues 344–359 of the protein with GenBank accession ID WYD17513 in the chosen strain of Fna. The protein sequence was conserved (100% identity) across the homologs in Fna Clade C2 and was observed to co-occur with the gene encoding TnpA. The analysis of the protein using the Conserved Domain Database (https://www.ncbi.nlm.nih.gov/cdd/) as well as HHpred (https://toolkit.tuebingen.mpg.de/tools/hhpred) indicates the protein (WYD17513) to be a homolog of TnpB, which is a family of proteins observed in IS200/605 and IS607 transposons in bacteria and in many cases, as observed in Fna, flanks the TnpA transposase protein (Karvelis et al., 2021). Although the TnpB was earlier considered to be an accessory protein to transposase TnpA, it has now been characterized as an RNA-guided DNA nuclease and is also considered to be an ancestor of CRISPR-associated Cas nucleases (Karvelis et al., 2021). The TnpB protein is considered important in retaining the transposon in a bacteria post-ligation (Karvelis et al., 2021). Studies show that there is a non-coding RNA observed within the same locus as TnpB proteins called omegaRNA or ωRNA, which aids in RNA-guided cleavage of DNA substrates. The 3′ end of the transposon is also called the right end (RE) and demarcates the 3′ end of the ωRNA. The RE is followed by a guide sequence encoded outside of the transposon that enables recognizing the target sequence for cleavage (Karvelis et al., 2021; Nety et al., 2023). The TnpB protein observed in Fusobacterium is followed by a non-coding 200-nucleotide inter-genic region indicating the possibility of ωRNA and guide RNA presence in this case as well.

Interestingly, the homologs for this Fusobacterium TnpB were observed in plasmids from gut pathogens like Clostridium botulinum and Clostridium difficile with two to three mismatches in the NLS where FKK residues are substituted with residues INR in the plasmid. Similar changes in NLS are also observed in viral homologs from Caudovirales as well as Clostridium phages. Despite changes in the sequence, cNLS was also predicted within these proteins by the NLS prediction tools indicating that the homologs also retain the capacity for nuclear localization. On further analysis using HHpred and BLAST, the results showed that the NLS region was observed as an insertion within the Fusobacterium and Clostridium RNA-guided endonucleases TnpB. The region corresponding to NLS was not observed in experimentally characterized TnpB structures as well as in HMMs used for guided TnpB proteins catalogued in Conserved Domain, TIGRFAM, and PFAM databases ( Supplementary Figure 7 ). All these observations may suggest that TnpB proteins with insertion of NLS sequences observed in gut pathogens like Fusobacterium and C. difficile might lead to their entry into the nucleus and influence gene expression. Our results show insertions of NLS in currently available genome sequences of bacteria C. botulinum, C. difficile, and Fusobacterium and the viral homologs in Fusobacterium and Clostridium phages. Despite this, it is possible that the TnpB sequence with NLS might also be observed in other bacteria due to horizontal transfer as the protein evolves.

Experimental studies have shown that fusing an NLS sequence with a TnpB with no NLS of its own showed significant changes in the host nucleus (Karvelis et al., 2021). The authors designed plasmids encoding the TnpB protein and fused it with the nuclear localization sequence and reRNA sequences, and these constructs were transfected into human HEK293T cells (Karvelis et al., 2021). The targeted cleavage sites in the experiment showed mutations at frequencies of 10%–20%, thereby indicating repair events and genome editing (Karvelis et al., 2021). Thus, this study showed that the presence of NLS in the protein TnpB provides it the capability of human cell genome editing, indicating its potential use in gene editing as a simpler alternative to CRISPR-Cas. In addition to having a nuclear localization property, an NM needs to be released out of the Fusobacterium and reach the host cells to exert its effect. Studies in Acinetobacter baumannii have indicated the presence of transposases with NLSs, which provides them a nuclear localization property (Moon et al., 2012a). Studies with fused GFP showed its presence in the nucleus of COS-7 cells (Moon et al., 2012a; Moon et al., 2012b). Furthermore, GFP fused with A. baumannii transposase showed presence in the nucleus of A549 cells and induced DNA methylation of E-cadherin gene, thereby reducing its expression using epigenetic modifications (Moon et al., 2012a). The study also showed presence of transposases in extracellular vesicles produced by the bacterium through which they could be delivered to host cells and finally targeted to the nucleus (Moon et al., 2012a). Although these observations in other bacteria might be indicative of the packaging of TnpB in Fusobacterium EVs and the capability of entering the host nucleus, further experimental studies would be required to prove the nuclear localization as well as gene editing capabilities of TnpB.

In order to further assess the binding of NLS in transposase with importin α, the interactions between the two proteins were assessed using protein–protein docking studies (described in the Methods section). The docked complex showed a HADDOCK score of −100.337 where the TnpB cNLS showed electrostatic, salt bridge, and cation–pi interactions with the importin major groove ARM repeats 1–4 ( Figure 4 ). The details of the obtained clusters after the docking of importin-TnpB and their energy landscape have been provided in Supplementary Figure 8 and Supplementary Table 7 .

As observed in experimentally characterized importin NLS structures, the Trp142 from ARM 2 and Trp231 from ARM 4 showed cation–pi and hydrophobic interactions with side-chain carbon atoms of Lys352 and Glu344 within the NLS of TnpB protein. Furthermore, the conserved residues Asn146 from ARM 2 and Asn235 from ARM 4 also showed polar interactions with the NLS. In addition to these interactions, the salt bridge is observed between Glu344 in NLS and Arg238 from the importin ARM 4. These interactions are in agreement with those required for cNLS binding to the major groove of importin α. Thus, the insertion of sequence corresponding to predicted NLS in the TnpB from Fusobacteria (as compared to Tnp sequences in other bacteria) as well as the interactions observed in the docked complex indicate that this protein might be able to localize in the nucleus and impact gene editing through its endonuclease function specific to certain sequences.

A Cas9 CRISPR endonuclease from Fna with a cNLS sequence NSRRRLKRRKWRLNLLEEIF was observed in fewer strains of the Fusobacterium genus. This indicates that only the homologs of Cas9 with NLS, observed in limited number of strains (of the Fusobacterium genus), can be considered as candidate NMs. Homologs possessing the NLS are also observed in bacterial pathogens including Listeria monocytogenes, Treponema, and Streptococcus pyogenes with four to five mismatches in the cNLS sequences. In addition, analysis of gene context of these Cas9 genes in genomes of all these bacteria showed the presence of a complete CRISPR Cas system including the repeat regions.

Another protein observed in a few Fusobacterium strains corresponds to WYD16313.1 in Fna with predicted cNLS ENLKLIKKKKRYKNSSVYYL corresponding to residues 82–101 on the protein. The homologs of the protein are observed in only a few sequenced strains of Fna and F. polymorphum with three mutations in the NLS region. Interestingly, the protein has matches with a >50% identity in Fusobacterium phages and other human gut phages belonging to the Caudoviricetes family. Furthermore, it was observed that the proteins in Fusobacterium strains belonging to species other than animalis as well as in Fusobacterium phages had insertions that were missing in the Fna protein. Despite these insertions, the cNLS was retained with just 2 mismatches in all these strains as well. The cNLS prediction tools also indicated the presence of NLSs in these proteins.

Although the protein WYD16313.1 has been annotated as Replication initiator protein A (RepA), the domain-based annotations using HHPred and Interpro (www.ebi.ac.uk/interpro/) indicate the presence of the N-terminal domain (NTD) of RepA in the first 106 residues while no domain predictions are observed for the remaining 170 C-terminal residues. It is known that while the NTD in RepA proteins is in a conserved DNA binding winged helix domain, the C-terminal domain shows significant sequence changes across bacterial genera (Schumacher et al., 2014).

Thus, in order to further corroborate that the protein is indeed a RepA, we analyzed the flanking genes of this protein within the Fna strain. Interestingly, the genomic neighborhood of the identified RepA protein in Fusobacterium genomes included a helicase and a helicase loader (DnaB and DnaC) protein as well as a large number of recombinase and nuclease proteins. A similar set of genes was also observed in the vicinity of each other in the Fusobacterium phages ( Figure 5 ). Furthermore, the NLS of Rep proteins in adeno-associated virus (AAV) has been shown to aid in the translocation of these proteins to the nucleus of infected human cells (Cassell and Weitzman, 2004). Thus, the homologs of RepA (in Fna) identified in the current study may have the potential to translocate to host cell nucleus with the help of their predicted cNLS, which can further be experimentally validated.

The protein RepA has been characterized in plasmids and is known to initiate replication using its winged helix domain for DNA binding. The protein finally binds the primase helicase DnaB/DnaC for the initiation of plasmid replication (Wegrzyn et al., [[NoYear]]). Thus, the observations indicate that the winged helix domain within the Fusobacterium RepA might interact with human DNA after entering the human nucleus. The role of the winged helix domain in human transcription and its effect on DNA-bound complexes have already been observed in earlier studies (Harami et al., 2013). Furthermore, the studies show that the characterized RepA proteins interact with DnaB and DnaC proteins to initiate replication, and the recombinase and nuclease proteins are utilized to bring about variations in the target DNA. Thus, all these proteins are shown to create a complex that can help in the initiation of plasmid replication. Therefore, it is possible that these proteins can form a complex and be carried together as a cargo using the RepA protein’s NLS to enter the human nucleus. The role of these proteins in human cells needs to be explored using further experimental studies.