The elimination of two restriction enzyme genes allows for electroporation-based transformation and CRISPR-Cas9-based base editing in the non-competent Gram-negative bacterium Acinetobacter sp. Tol 5

ABSTRACT

Environmental isolates are promising candidates for new chassis of synthetic biology because of their inherent capabilities, which include efficiently converting a wide range of substrates into valuable products and resilience to environmental stresses; however, many remain genetically intractable and unamenable to established genetic tools tailored for model bacteria. Acinetobacter sp. Tol 5, an environmentally isolated Gram-negative bacterium, possesses intriguing properties for use in synthetic biology applications. Despite the previous development of genetic tools for the engineering of strain Tol 5, its genetic manipulation has been hindered by low transformation efficiency via electroporation, rendering the process laborious and time-consuming. This study demonstrated the genetic refinement of the Tol 5 strain, achieving efficient transformation via electroporation. We deleted two genes encoding type I and type III restriction enzymes. The resulting mutant strain not only exhibited marked efficiency of electrotransformation but also proved receptive to both in vitro and in vivo DNA assembly technologies, thereby facilitating the construction of recombinant DNA without reliance on intermediate Escherichia coli constructs. In addition, we successfully adapted a CRISPR-Cas9-based base-editing platform developed for other Acinetobacter species. Our findings provide genetic modification strategies that allow for the domestication of environmentally isolated bacteria, streamlining their utilization in synthetic biology applications.

IMPORTANCE

Recent synthetic biology has sought diverse bacterial chassis from environmental sources to circumvent the limitations of laboratory Escherichia coli strains for industrial and environmental applications. One of the critical barriers in cell engineering of bacterial chassis is their inherent resistance to recombinant DNA, propagated either in vitro or within E. coli cells. Environmental bacteria have evolved defense mechanisms against foreign DNA as a response to the constant threat of phage infection. The ubiquity of phages in natural settings accounts for the genetic intractability of environmental isolates. The significance of our research is in demonstrating genetic modification strategies for the cell engineering of such genetically intractable bacteria. This research marks a pivotal step in the domestication of environmentally isolated bacteria, promising candidates for emerging synthetic biology chassis. Our work thus significantly contributes to advancing their applications across industrial, environmental, and biomedical fields.

INTRODUCTION

Synthetic biology designs biological systems for applications beyond the realm of natural evolution. Although Escherichia coli has been used as the most tractable chassis over the past two decades, synthetic biology has recently sought a more robust, efficient, and diverse chassis capable of thriving in diverse environments and churning out complexes that E. coli can struggle to produce (1, 2). Genetic manipulation is essential not only for elucidating the biological functions of bacterial genes but also for effectively employing bacteria in synthetic biology applications encompassing industrial, environmental, and biomedical fields. CRISPR-Cas-based technologies have remarkably enhanced the efficiency of correlating genotype and phenotype (3–5) and constructed advantageous bacteria for synthetic biology applications (6–9). However, these techniques exhibit efficient performance in E. coli and some closely related enterobacteria, while functioning inefficiently in other bacterial species. Therefore, researchers need to develop a new tool or modify an existing tool for applying CRISPR-Cas-based technologies to bacterial species of interest (10–12). Species specificity of CRISPR-Cas-based systems is more pronounced in prokaryotes than in eukaryotes (13). Transformation and recombination efficiencies with exogenous DNA are postulated factors underlying the species specificity of CRISPR-Cas-based systems. Each bacterial species exhibits a distinct resistance to exogenous DNA. For instance, Acinetobacter baylyi ADP1 exhibits high competency, facilitating natural transformation and a strong inclination that is sufficient for homology-directed recombination without the assistance of a phage-derived recombinase (14). Conversely, other Acinetobacter species do not possess such traits. Non-naturally competent Acinetobacter baumannii strains are transformed via electroporation, and their CRISPR-Cas-based genome editing necessitates the use of recombinases derived from phages infecting Acinetobacter species (15). Even within the same Acinetobacter species, genetic manipulation and genome editing require different strategies.

Acinetobacter sp. Tol 5 is an environmentally isolated Gram-negative bacterium with intriguing characteristics for industrial and environmental applications. Tol 5 can metabolize various chemicals (16, 17) and exhibits high adhesiveness to various abiotic surfaces, ranging from hydrophobic plastics to hydrophilic glass and stainless steel through the bacterionanofiber protein AtaA (18). AtaA-mediated adhesiveness enables Tol 5 cells to function as immobilized biocatalysts for pollutant removal and chemical production (19, 20). We developed genetic manipulation tools for engineering Tol 5, including a replicable plasmid for expressing genes of interest and suicide plasmids for target gene deletion from Tol 5’s genome (18, 21). Nevertheless, these tools were designed based on transformation through bacterial conjugation because of the low transformation efficiency of Tol 5 via electroporation. Although bacterial conjugation is commonly employed for the efficient transformation of non-competent bacteria, it requires multiple steps to obtain the desired transformant and gene knockout mutant, rendering genetic manipulations laborious. Furthermore, conjugative plasmids must incorporate the sequence of the origin of transfer. Despite the availability of valuable plasmids from Addgene (https://www.addgene.org/), those devised for competent Acinetobacter strains are incompatible with the direct engineering of Tol 5 cells. In contrast, electroporation facilitates transformant acquisition and genome editing in fewer steps than those required for bacterial conjugation. Electroporation-based transformation and genome editing would expedite the construction of beneficial Tol 5 strains for synthetic biology applications.

In this study, we report on a genetically engineered Tol 5 strain capable of efficient transformation through electroporation. Restriction-modification (R-M) systems are recognized as prevalent barriers against foreign DNA invasion by bacteria (22). Some research reported the relationship between R-M systems and transformation efficiency via electroporation (23–25). Consequently, we hypothesized that the R-M systems in Tol 5 hinder electrotransformation. We identified four putative R-M systems within the Tol 5 genome by sequence similarity searches and excised their restriction enzymes (REases) using suicide plasmids through a conjugation-based approach. The resultant mutant strain not only underwent efficient transformation by electroporation but also proved amenable to in vitro and in vivo DNA assembly technologies. Two CRISPR-Cas9-based tools developed for other Acinetobacter strains were used to inactivate the ataA gene of Tol 5 (TOL5_00260). The CRISPR-Cas9-based genome-editing system failed to delete the target gene, while the CRISPR-Cas9-based base-editing system successfully introduced a mutation to inhibit the transcription of the target gene.

RESULTS

Influence of base modifications in a plasmid on electroporation-based transformation

R-M systems play a protective role in bacterial cells by distinguishing between their own DNA and foreign DNA, such as phage DNA, and frequently impede bacterial transformation using recombinant plasmids (23–25). We previously reported the complete genome sequence of Tol 5, which encompasses 4,402 protein-encoding genes (26). Sequence similarity search of Tol 5’s genome using REBASE (27) database (http://rebase.neb.com/rebase/rebase.html), the restriction enzyme database, revealed the presence of four putative R-M systems, including types I, IIG, III, and IV (Table 1). Types I, IIG, and III R-M systems shield host bacterial cells from foreign DNA lacking host-specific base modifications. Conversely, type IV R-M system defends against foreign DNA with specific base modifications that are not present in host bacteria. Therefore, we believe that these four R-M systems limit the electrotransformation of Tol 5 with plasmids propagated in E. coli. To validate this hypothesis, electroporation was conducted using an E. coli-Acinetobacter shuttle plasmid, pARP3, extracted from E. coli DH10B or Tol 5 transformant cells (Fig. 1). Electroporation with pARP3 sourced from E. coli DH10B yielded a transformation efficiency of a mere 3.7 colony-forming units (CFU)/µg-DNA. However, utilizing the pARP3 plasmid sourced from the Tol 5 transformant amplified this value to an impressive 4.1 × 10⁸ CFU/µg-DNA. E. coli DH10B has two DNA methyltransferases that target the N6 position of adenine (m6A): Dam methyltransferase, which methylates adenine in the GA*TC motif, and EcoKI methyltransferase, which methylates the second adenine in the AA*C(N6)GTGC motif and the third adenine in the GCA*C(N6)GTT motif. In addition, E. coli DH10B has a DNA methyltransferase that targets the C5 position of cytosine (m5C): Dcm methyltransferase, which methylates the second cytosine in the CC*WGG motif. Yet, Tol 5 cells barely received the pARP3 plasmid replicated in E. coli DH10B. The resultant transformation efficiency, which relied on the source of the pARP3 plasmid, implied that the R-M systems in Tol 5 employed adenine modifications divergent from E. coli DH10B. Single-molecule real-time (SMRT) sequencing, renowned for its precision in detecting m6A, revealed that the pARP3 plasmid from the Tol 5 transformant contained 26 methylated adenines, some of which deviated from the recognition sequences affiliated with Dam and EcoKI methyltransferase (Table S1). Although SMRT sequencing is less sensitive toward N4-methylcytosine (m4C) and performs poorly for C5-methylated cytosine (m5C) methylation (28), we also performed basecalling toward m4C and detected 305 candidates (Table S2). The results of electroporation and SMRT sequencing suggest that unique R-M systems are key limiting factors for the electrotransformation of Tol 5.

TABLE 1.

Restriction enzymes found in the genome of Acinetobacter sp. Tol 5

Type	Name	Gene locus tag	Best hit in REBASEa
I	AspTol5ORF4019P	TOL5_40190	DvuI (AAS96180)
IIG	AspTol5ORF3984P	TOL5_39840	Eco57I (AAA23389)
III	AspTol5ORF227P	TOL5_02270	Vsp69I (CCF17694)
IV	AspTol5ORF4146P	TOL5_41460	ZmoCP4Mrr (ABY40736)

^a Restriction enzyme database (http://rebase.neb.com/rebase/rebase.html).

Fig 1.

Transformation efficiency of Acinetobacter sp. Tol 5 when electroporated with the E. coli-Acinetobacter shuttle plasmid pARP3. The pARP3 plasmids, extracted from either E. coli DH10B transformants or Tol 5 transformants, were used to electroporate the wild-type Tol 5 strain. Data represent the means from five replicates ± SDs.

Identification of restriction nucleases hampering electroporation-based transformation of Acinetobacter sp. Tol 5

To investigate the influence of the R-M systems in Tol 5 on the transformation efficiency by electroporation, we constructed mutants deficient in a gene encoding a putative REase. Electroporation of these mutants with pARP3 sourced from E. coli DH10B yielded an increase in transformation efficiency to 1.0 ± 0.4 × 10³ CFU/µg-DNA and 1.4 ± 0.5 × 10² CFU/µg-DNA for mutants deficient in type I (REK1) and type III (REK3) REases, respectively (Fig. 2). Mutants lacking types II and IV REases (REK2 and REK4) exhibited unaltered transformation efficiency. This observation distinctly implies the role of type I and III R-M systems in the impeding electrotransformation using the E. coli-sourced pARP3 plasmid. In contrast, the deletion of types IIG and VI REase genes did not have a similar effect, suggesting the absence of sequence motifs recognized by these REases on the pARP3 plasmid. A combined deficiency in types I and III REases (double mutant, REK13) culminated in a transformation efficiency of 2.1 ± 1.5 × 10⁵ CFU/µg-DNA, surpassing individual mutant efficiencies. The transformation efficiency of the double mutant REK13 with the pARP3 plasmid replicated in E. coli DH10B was approximately 2,000-fold lower than that of the wild-type strain with the pARP3 plasmid replicated in Tol 5 (upper dashed line in Fig. 2).

Fig 2.

Effect of deleting restriction nuclease genes on the transformation efficiency of Acinetobacter sp. Tol 5. The pARP3 plasmid replicated in E. coli DH10B was electroporated into Tol 5 mutants lacking genes encoding certain restriction enzymes. The upper dashed line represents the transformation efficiency of Tol 5 wild-type strain with pARP3 plasmid replicated in Tol 5, which is shown as the rightmost bar in Fig. 1. The lower dashed line represents the transformation efficiency of Tol 5 wild-type strain with the pARP3 plasmid replicated in E. coli DH10B, the leftmost bar in Fig. 1. Data represent the means from five replicates ± SDs.

In vitro and in vivo recombinant DNA assembly for Acinetobacter sp. Tol 5

For the cell engineering of non-model bacteria, recombinant DNA is typically constructed within competent E. coli cells using a combination of linearized plasmids and insert DNAs. Subsequently, non-model bacteria are transformed with the resulting recombinant DNA extracted from E. coli cells. Constructing recombinant DNA directly within non-model bacteria can expedite the cell engineering process for synthetic biology applications. Commercially available E. coli competent cells show an extremely high transformation efficiency, approximately 10⁸–10¹⁰ CFU/µg-DNA. Although such competent cells are powerful tools for constructing DNA libraries and assembling multiple DNA fragments, routine DNA cloning and assembly do not require extremely high transformation efficiencies. Therefore, the REK13 mutant appears available for constructing recombinant DNA for cell engineering. Modern DNA assembly techniques, such as in-fusion cloning, Gibson assembly, and Golden Gate assembly, have largely superseded traditional DNA cloning, which relies on restriction enzymes and ligases. NEBuilder HiFi DNA assembly, an advancement from Gibson assembly, assembles linear DNA fragments in vitro using thermostable DNA polymerase, thermostable DNA ligase, and thermolabile T5 5´→3´ exonuclease. We attempted to transform the Tol 5 REK13 mutant with recombinant DNA assembled in vitro using the NEBuilder HiFi DNA assembly (Fig. 3A). The vector DNA fragment was amplified by inverse PCR from the pAPR3 plasmid, whereas the insert DNA fragment containing gfp gene was amplified by PCR using primers that included 20 bp overlaps of the ends of the vector DNA fragment. After incubation for in vitro DNA assembly, the reaction mixture was directly used for electroporation of Tol 5 (WT) or the Tol 5 REK13 mutant (REK13). While the electroporation of WT yielded no colony, the electroporation of the REK13 mutant generated 7.7 ± 0.58 × 10⁴ colonies per 1 pmol-vector DNA generated (Fig. 3B). An impressive 96.7% ± 0.74% of the colonies from the REK13 transformant displayed green fluorescence under blue Light Emitting Diode (LED) light (Fig. 3C), indicating successful DNA assembly. Hence, the Tol 5 REK13 mutant could be effectively transformed with synthetic recombinant DNA assembled in vitro, enabling the selection of desired clones from its transformants.

Fig 3.

Electrotransformation of Acinetobacter sp. Tol 5 cells with an in vitro-assembled plasmid DNA. (A) Scheme of the electrotransformation of Tol 5 with an in vitro-assembled DNA. PCR fragments including 20 bp overlaps at their ends undergo in vitro using NEBuilder HiFi DNA assembly. The vector pARP3 PCR fragment carries genes for gentamicin resistance (Gm^R) and ampicillin resistance (Amp^R), while the insert PCR fragment includes the gfp gene (GFP). The assembly reaction mixture is directly employed for Tol 5 electroporation. (B) Colony number appeared on Luria-Bertani (LB) agar plates with gentamicin and ampicillin after the electrotransformation. Data represent the means from three replicates ± SDs. (C) Percentage of GFP positive colony among those that appeared on LB agar plates with ampicillin and gentamicin. Data represent the means from three replicates ± SDs.

In addition to in vitro DNA assembly techniques, in vivo DNA assembly techniques have also been used to construct recombinant DNA for synthetic biology applications. The in vivo E. coli cloning method (also referred to as iVEC), for instance, provides a more streamlined approach, eschewing the need for costly enzyme solutions (29, 30). In typical iVEC methods, linear DNA fragments bearing 20–50 bp overlapping sequences at their ends are simply introduced into E. coli cells through calcium chloride treatment followed by heat shock or electroporation. We adapted this in vivo cloning method for the Tol 5 REK13 mutant using the iVEC protocol (Fig. 4A). Electroporation of DNA fragments with 20 bp overlapping sequences (838 bp insert and 6,905 bp vector), identical to those used in the NEBuilder HiFi DNA assembly (Fig. 3), yielded approximately 1.5 ± 0.58 × 10³ CFU/pmol-vector DNA on Luria-Bertani (LB) agar plates containing ampicillin and gentamicin (Fig. 4B). Although the colonies generated by in vivo DNA assembly were approximately 51-fold fewer than those generated by in vitro assembly, they all exhibited green fluorescence, indicating successful DNA assembly (Fig. 4C). Electroporation experiments using DNA fragments with 10 bp and 30 bp overlaps were also conducted. Prolonging overlap sequences to 30 bp augmented the colony count to 4.0 ± 0.98 × 10³ CFU/pmol-vector DNA, with a 98.3% ± 1.6% GFP positivity. Conversely, reducing the overlaps to 10 bp significantly diminished both the colony number (50 ± 86.7 CFU/pmol-vector DNA) and GFP fluorescence positivity (11.1% ± 19.2%). In summary, employing both in vitro and in vivo DNA assembly techniques with the Tol 5 REK13 mutant eliminated the need for intermediate E. coli constructs, thus facilitating the development of tailor-made strains for synthetic biology applications.

Fig 4.

Electroporation-based in vivo DNA assembly of Acinetobacter sp. Tol 5 strains. (A) Scheme of in vivo DNA assembly of Acinetobacter sp. Tol 5. PCR fragments including 10–30 bp overlaps at their ends are electroporated into a Tol 5 mutant strain deficient in types I and III restriction nucleases (REK13). The PCR fragment for an insert contains gfp gene (GFP), whereas the PCR fragment for a vector contains gentamicin resistance (Gm^R) and ampicillin resistance gene (Amp^R). (B) Colony number appeared on LB agar plates with gentamicin and ampicillin after the electrotransformation. Data represent the means from three replicates ± SDs. (C) Percentage of GFP positive colony among those that appeared on LB agar plates with ampicillin and gentamicin. Data represent the means from three replicates ± SDs.

CRISPR-Cas9-based technologies applied to Acinetobacter sp. Tol 5

Genome engineering, in addition to plasmid DNA transformation, is crucial for constructing bacteria that are useful for synthetic biology applications. Various CRISPR-Cas9-based technologies with species-specific modifications have been developed for deleting and inserting genes from diverse bacteria (10). Wang et al. developed highly efficient CRISPR-Cas9-based genome and base-editing methods optimized for A. baumannii (15). Although the CRISPR-Cas9-based systems for A. baumannii were likely to function in Tol 5, implementation was deterred because of the requirement for electrotransformation with plasmids carrying Cas9 and single guide RNA (sgRNA) genes. However, the deletion of the two restriction enzyme genes facilitated the transformation of the Tol 5 REK13 mutant with plasmid DNA through electroporation. Thus, we tested whether the CRISPR-Cas9 system for A. baumannii could cleave the genome of the Tol 5 REK13 mutant. The CRISPR-Cas9-based genome-editing system for A. baumannii uses two plasmids (pCasAb-apr and pSGAb-km) (15). The pCasAb-apr plasmid contains isopropyl β-D-thiogalactopyranoside (IPTG)-inducible Cas9 nuclease and IPTG-inducible RecAb recombination system. As A. baumannii also lacks the ability to perform non-homologous end-joining (NHEJ) like many other bacteria (31, 32), double-strand breaks (DSBs) caused by the Cas9 nuclease are repaired by homologous recombination. However, in the absence of the RecAb recombination system, the genome editing is not successful in A. baumannii, possibly due to its low capacity for homologous recombination (15). On the other hand, the pSGAb-km plasmid contains an expression cassette for the sgRNA under the control of a constitutive J23119 promoter, which includes two BsaI sites for cloning the 20 bp spacer sequence of a target gene. Tol 5 REK13 mutant cells harboring the pCasAb-apr plasmid were transformed with the pSGAb-km_ataA plasmid containing the 20 bp spacer sequence of ataA or empty pSGAb-km. The ataA gene encodes the adhesive nanofiber protein localized on the cell surface of Tol 5 (18). As we were able to delete the ataA gene using the conjugative plasmid-based method (18, 21), we targeted it to investigate the feasibility of the CRISPR-Cas9-based genome-editing system for A. baumannii in the Tol 5 REK13 mutant. Electroporation with the empty pSGAb-km yielded 4,426 ± 2,753 colonies, whereas that with pSGAb-km_ataA yielded a mere 9.5 ± 8.2 colonies (Fig. 5). This result implies that the Cas9-sgRNA complex targets and cleaves to the Tol 5 genome, and the subsequent DSB is lethal. Subsequently, we electroporated pSGAb-km_ataA_HR, containing the 20 bp spacer sequence of ataA and the repair template of ataA lacking a 274 bp sequence for homologous recombination, into Tol 5 REK13 mutant cells harboring the pCas9-apr plasmid. However, no difference was observed between electrotransformation with and without the repair template, suggesting that there was no homologous recombination after the Cas9-induced DSB. Colony PCR revealed no 274 bp deletion through homologous recombination with the repair template after Cas9-induced DSB. Although genome editing was not performed, the lethality of the Cas9-induced DSB confirmed the binding and cleaving activity of the Cas9-sgRNA complex on the genomic DNA of Tol 5.

Fig 5.

CRISPR-Cas9-based genome editing in Acinetobacter sp. Tol 5. (A) Scheme of CRISPR-cas9-based genome editing in Acinetobacter sp. Tol 5 REK13 mutant. The pSGAb-km_ataA_HR plasmid contains the sgRNA that directs Cas9 to the ataA gene and a repair template for homologous recombination following a Cas9-sgRNA-induced DSB. The pCasAb-apr plasmid supplies Cas9, which creates DSBs in the Tol 5 REK13 mutant’s genome in concert with sgRNA. The pSGAb-km_ataA_HR plasmid is also cleaved by the Cas9-sgRNA complex and subsequently serves as a template for homologous recombination to repair the genome injured by the DSB. (B) The CFUs of each transformation using different plasmids. The pSGAb-km plasmid, which harbors neither the sgRNA nor the repair template, was used as a negative control for the genome editing of ataA. Data represent the means from four replicates ± SDs. (C) Representative gel-like images of colony PCR amplicons displayed on the MCE-202 MultiNA. The attempted deletion of the ataA gene was examined using colony PCR using primers annealing to the outside of the flanking regions of the DSB used as homologous sites. No successful deletion was confirmed in the 12 tested colonies.

Wang et al. also developed a CRISPR-Cas9-based base-editing system for A. baumannii (15). The utilization of apolipoprotein B mRNA editing enzyme catalytic polypeptide 1 (APOBEC1) (33), which is merged with Cas9 (D10A) nickase, allows for the conversion of cytidine to thymidine without introducing a DSB. The APOBEC1 is a cytidine deaminase that catalyzes the conversion of cytosine to uracil on a displaced DNA strand. Concurrently, the Cas9(D10A) nickase cleaves the opposite, non-edited DNA strand. The resulting U:G hetero-duplex can be converted to a T:A base pair by subsequent DNA repair or replication processes. This bypasses the need for homologous recombination for targeting specific genes. We applied this CRISPR-Cas9-based base editing to introduce a stop codon within ataA in the Tol 5 REK13 mutant (Fig. 6A). Notably, the 20 nt-spacer sequence of ataA remained consistent with that used for genome editing (Fig. 6B). Electroporation of the Tol 5 REK13 mutant was conducted with pBECAb-apr_ataA [consisting of APOBEC1 fused to Cas9 (D10A) nickase and sgRNA], yielding several hundred colonies on an LB agar plate supplemented with apramycin. Subsequent colony PCR and sequencing identified 71.4% (10/16) colonies carrying the desired mutations that inhibited ataA transcription. The resultant DNA sequencing results revealed that the glutamine codon (CAA) at the 116th amino acid position in AtaA was switched to the stop codon (TAA) (Fig. 6C).

Fig 6.

CRISPR-Cas9-based base editing. (A) Scheme of CRISPR-Cas9-based base editing in Acinetobacter sp. Tol 5 REK13 mutant. The pBECAb-apr plasmid supplies the APOBEC1-fused Cas9 (D10A) nickase in Tol 5 REK13 mutant cells. The APOBEC1-fused Cas9 editor converts C to U, finally forming the T:A base pair. (B) Design of the spacer sequence targeting the ataA gene for knockout. The spacer sequence intended for a stop codon insertion is highlighted by red underline. The targeted cytosine within the glutamine (Q116) is indicated in orange. The protospacer adjacent motif sequence is highlighted in bold. (C) A representative sequencing result from a base-edited Tol 5 REK13 mutant colony.

DISCUSSION

In this study, by deleting two genes encoding REases, which form a part of the typical bacterial defense systems against foreign DNA, we successfully introduced recombinant DNA into Tol 5 cells with high efficiency. This advancement facilitated the use of in vivo/in vitro DNA assembly technologies and the CRISPR-Cas-based base-editing method previously developed for different Acinetobacter strains. Some research also reported that the deletion of REase genes increased the transformation efficiency via electroporation (23–25). REases play a protective role in bacterial cells by degrading invasive DNAs from phages, transferable plasmids, and transposons. As bacterial cells cannot differentiate recombinant DNAs from these invasive elements, REases eliminate them based on different base modifications. For synthetic biology applications, recombinant DNAs are commonly propagated in vitro or within E. coli cells. Cell-free DNA amplification technologies, such as PCR and rolling circle amplification, yield unmethylated products, whereas E. coli specifically modifies recombinant plasmid DNA using its own methyltransferases. This presents a challenge when the R-M systems of the target bacteria differ from E. coli, making the transformation process difficult using recombinant DNAs prepared in vitro or within E. coli cells. The REBASE database documents various methylation patterns across bacterial species, and metagenomic analysis unveils new methylation motifs, underscoring the diversity of R-M systems (34). This diversity explains why bacteria isolated from the environment are often resistant to transformation with recombinant DNAs prepared in vitro or within E. coli cells.

The base-editing method successfully inactivated the ataA gene; however, attempts at genome editing to remove it were unsuccessful. This may be due to the incompatibility of the recombination system within the pCasAb-apr plasmid with Tol 5 cells. When Wang et al. developed the pCasAb-apr plasmid, they tested three different exogenous recombination systems (Red from lambda phage, RecET from Rac prophage, and RecAb from A. baumannii) and identified that only RecAb could repair the DSBs generated by CRISPR-Cas9 in A. baumannii (15). For successful genome editing, we need to find an exogenous recombination system that facilitates homologous recombination in Tol 5 cells. As many bacteria lack the ability to perform NHEJ (31, 32), species specificity of the recombination system is a critical consideration when applying CRISPR-Cas9-based technologies to non-model bacteria.

The deletion of types I and III REases genes increased the transformation efficiency with the pARP3 plasmid. In contrast, the deletion of type IIG and VI REase genes did not show a similar effect, suggesting the absence of sequence motifs recognized by these REases on the pARP3 plasmid. However, the transformation efficiency in the double mutant REK13 with the pARP3 plasmid replicated in E. coli DH10B did not correspond to that of the wild-type strain with the pARP3 plasmid replicated in Tol 5 (Fig. 2). This suggests the involvement of methylation-directed defense systems against foreign DNA in Tol 5, distinct from conventional R-M systems. The bacteriophage exclusion (BREX) system, a relatively recently identified defense mechanism (35), could be an influencing factor. This system, which modifies phage DNA to differentiate it from bacterial DNA, might also target plasmid DNA, as suggested by transformation experiments with the BREX-lacking Bacillus subtilis BST7003 (35). Two sets of gene clusters homologous to components of the BREX system (BCX74278–BCX74285 and BCX76169–BCX76174) were found in the complete genome sequence of Tol 5. The BREX systems in Tol 5 may hamper transformation via electroporation.

The ease of genetic manipulation in laboratory E. coli strains has significantly propelled synthetic biology research over the past two decades. However, these strains do not always serve as ideal hosts for industrial and environmental applications, prompting the search for new bacterial chassis (1, 2). Although new chassis for synthetic biology need to be genetically tractable, many environmental isolates remain intractable. Phages are the most abundant biological entities on Earth, with the particle count estimated to exceed 10 times that of bacteria (36). Given the prevalence of phages, environmental bacteria have evolved robust defense systems against them. These defense systems often target and eliminate recombinant DNAs constructed in vitro or within E. coli cells. Methods for in vivo/in vitro DNA assembly directly using target bacteria eliminate the need for intermediate E. coli constructs. The traditional cell engineering of non-competent bacteria, which relies on conjugation-based transformation, involves laborious and time-consuming procedures. Our strategy, which involves the deletion of REase genes using conjugation methods, paves the way for the adoption of efficient electroporation-based transformation, in vivo/in vitro DNA assembly, and CRISPR-Cas9 technologies in the non-competent bacteria. This approach could significantly expedite the cell engineering process in beneficial non-model bacteria, advancing their application in synthetic biology.

MATERIALS AND METHODS

Bacterial culture

Acinetobacter sp. Tol 5, which is closely related to Acinetobacter bereziniae, and its derivative strains were grown in basal salt (BS) medium supplemented with 0.05% (vol/vol) toluene or LB medium as previously described (18). E. coli strains were grown in an LB medium containing the appropriate antibiotics at 37°C. Antibiotics were used at the following concentrations when required: ampicillin (100 µg/mL), apramycin (100 µg/ mL), gentamicin (10 µg/mL), and kanamycin (50 µg/mL).

Construction of a deletion mutant in Acinetobacter sp. Tol 5

Deletion mutants of Tol 5 presented in Table S3 were constructed using the plasmid-based counterselection method as described previously (37). The primers and plasmids used for the construction of mutant strains are presented in Tables S4 and S5, respectively. Briefly, approximately 1-kb regions upstream and downstream regions of a target gene were amplified by PCR and assembled with BamHI-cut pJQ200sk using the NEBuilder HiFi DNA assembly system (New England Biolabs, Ipswich, MA) to generate a suicide plasmid. Tol 5 or its derivative mutant was mated with E. coli S17-1 harboring the suicide plasmid on an LB agar plate for 24 h at 28°C. The cells were collected in 1 mL of 0.85% NaCl solution, plated on BS agar plates containing gentamicin (100 µg/mL), and incubated with toluene vapor for 2 days at 28°C. Chromosomal integration of the suicide plasmid was confirmed by colony PCR using primers annealing to the outside and inside of the flanking region of the target gene, which were used as homologous sites for recombination. For the deletion of a target gene, the resulting single-crossover mutant was spread on a BS agar plate containing 5% sucrose and incubated with toluene vapor for 2 days at 28°C. Deletion of the target gene was confirmed by PCR using primers annealing to the outside of the flanking region of the target gene used as homologous sites for recombination.

Electroporation

Tol 5 and its derivative mutants were inoculated into 5 mL of 2 × YT medium in a 50 mL conical tube and incubated at 28°C with shaking at 125 rpm. After overnight incubation, the cells were harvested by centrifugation at 8,000 × g for 3 min at 4°C, rinsed with 10 mL 10% (wt/vol) glycerol solution three times, and resuspended in 500 µL 10% (wt/vol) glycerol solution. One hundred microliter of the cell suspension was mixed with DNA and transferred to an ice-cold 2-mm-gap cuvette. Electroporation was performed using a Gene Pulser II system (Bio-Rad, Hercules, CA) set at 25 µF, 3,000 V, and 200 Ω. Immediately after electroporation, 0.9 mL of 2 × YT medium at room temperature was added to the cells. The resuspended cells were transferred into a 15 mL conical tube and incubated at 28°C for 2 h with shaking at 125 rpm. After incubation, the cells were spread on a selective plate.

SMRT sequencing

The pARP3 plasmid for methylation analysis was extracted from 200 mL of an overnight culture of Tol 5 transformant cells using the HiSpeed Plasmid Midi kit (QIAGEN) according to the manufacturer’s instructions. The extracted plasmid was linearized by SalI digestion and purified using the Wizard SV PCR cleanup system (Promega, Madison, WI). SMRT sequencing was performed using Sequel IIe (Pacific Biosciences, Menlo Park, CA). The pARP3 sequence was determined using reads covering the full-length plasmid. Reads were mapped to the pARP3 sequence with pbmm2 1.4.0 in SMRT Tools (Pacific Biosciences). Detection of DNA methylation signatures was performed using ipdSummary in the SMRT Tools program (Pacific Biosciences) with the --identify m6A, m4C, --methylFraction, and --maxAligments 15,000 options.

In vitro and in vivo DNA assembly

For in vitro and in vivo DNA assembly, DNA fragments were amplified using KOD-Plus-Ver.2 (Toyobo, Osaka, Japan), and the primers are presented in Table S5. All the resulting amplicons were purified using the Wizard SV PCR cleanup system (Promega). The DNA fragment for the vector was amplified by PCR using the primer set Inv-pARP3-fwd/Inv-pARP3-rev and the pARP3 plasmid as the DNA template. A DNA fragment containing a constitutive promoter, a ribosome-binding site, and the gfp gene was artificially synthesized as a template for amplifying DNA fragments for inserts.

For in vitro assembly, the DNA fragment containing 20-bp-overhangs at the ends was amplified using the primer set 20bp-eGFP-fwd/20bp-eGFP-rev. After purification, 0.18 pmol of the insert DNA fragment and 0.02 pmol of the vector DNA fragment were mixed with NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs) to make a total volume of 10 µL. After 1-h incubation at 50°C, 1 µL of the reaction mixture was electroporated into Tol 5 or REK13 mutant cells. Transformants were selected on LB agar plates containing ampicillin and gentamicin.

For in vivo assembly, the DNA fragments containing 10 bp-, 20 bp-, and 30 bp-overhangs at the ends were amplified using primer sets 10bp-eGFP-fwd/10bp-eGFP-rev, 20bp-eGFP_fwd/20bp-eGFP-rev, and 30bp-eGFP-fwd/30bp-eGFP-rev, respectively. After the purification using Wizard SV PCR cleanup system (Promega), 0.18 pmol of the insert DNA fragment and 0.02 pmol of the vector DNA fragment were directly electroporated into REK13 mutant cells. Transformants were selected on LB agar plates containing ampicillin and gentamicin.

Genome editing and base editing

CRISPR-Cas9-based genome and base editing were performed according to the methods described by Wang et al. (15). Briefly, two oligonucleotides (5′-TAGTAACGCACAAGGTGGTACTTC-3′ and 5′-TTGCGTGTTCCACCATGAAGCAAA-3′) were synthesized for 20 bp spacer sequence before a protospacer adjacent motif site (5′-NGG-3′) in the ataA gene. The two oligonucleotides were phosphorylated, annealed, and cloned into the BsaI sites of pSGAb-km and pBECAb-apr using a Golden Gate Assembly Kit (New England Biolabs) to generate pSGAb_ataA and pBECAb_ataA, respectively. To prepare the plasmid-based repair template of ataA for homologous recombination after Cas9-mediated DSB, 214 bp upstream and 284 bp downstream of the DSB site in ataA were amplified by PCR using the primer sets RTemp-upst_fwd/RTemp-upst_rev and RTemp-dwst_fwd/RTemp-dwst_rev, respectively. The resulting amplicons were assembled with SalI- and KpnI-cut pSGAb-ataA using the NEBuilder HiFi DNA assembly kit (New England Biolabs) to generate pSGAb-ataA_HR.

For genome editing, pCasAb-apr was electroporated into Tol 5 REK13 mutant. The resulting transformant was inoculated into 5 mL of 2 × YT medium containing 100 µg/mL apramycin in a 50 mL conical tube and incubated at 28°C with shaking at 125 rpm. After overnight cultivation, the cells were prepared for electroporation as described above. Then, 0.1 pmol of pSGAb-km, pSGAb-ataA, or pSGAb-ataA_HR was electroporated into Tol 5 REK mutant cells harboring pCas9Ab-apr. Electroporated cells were plated on LB agar plates containing apramycin, kanamycin, and IPTG (0.1 mM) and incubated at 28°C overnight. Deletion of the ataA gene was examined by colony PCR using the primers Edit-Check-F and Edit-Check-R2. Colony PCR amplicons were loaded onto an MCE-202 MultiNA (Shimadzu, Kyoto, Japan).

For base editing, pBECAb-ataA was electroporated into the Tol 5 REK13 mutant. Electroporated cells were plated onto LB agar plates containing apramycin. Base editing of the ataA gene was confirmed by colony PCR using the primer set Edit-Check-F/Edit-Check-R2 and subsequent amplicon sequencing. Successfully base-edited mutants were cultivated in LB medium and streaked onto an LB agar plate containing 5% sucrose to cure the pBECAb-ataA-spacer plasmid using SacB-based counterselection.

DATA AVAILABILITY

The genome sequence of Acinetobacter sp. Tol 5 was deposited in DDBJ/GenBank under the accession numbers AP024708 and AP024709.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.00400-24.

Tables S1 and S2. aem.00400-24-s0001.xlsx.

DNA modifications detected by the SMRT sequencing.

Tables S3 to S5. aem.00400-24-s0002.pdf.

Bacterial strains, plasmids, and primers used in this study.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.