ABSTRACT
Nonhomologous end joining (NHEJ) is critical for genome stability because of its roles in double-strand break repair. Ku and ligase D (LigD) are the crucial proteins in this process, and strains expressing Ku and LigD can cyclize linear DNA in vivo. Here, we established a proof-of-concept single-homology-arm linear DNA recombination for gene inactivation or genome editing by which cyclization of linear DNA in vivo by NHEJ could be used to generate nonreplicable circular DNA and could allow allelic exchanges between the circular DNA and the chromosome. We achieved this approach in Dietzia sp. strain DQ12-45-1b, which expresses Ku and LigD homologs and presents NHEJ activity. By transforming the strain with a linear DNA single homolog to the sequence in the chromosome, we mutated the genome. This method did not require the screening of suitable plasmids and was easy and time-effective. Bioinformatic analysis showed that more than 20% of prokaryotic organisms contain Ku and LigD, suggesting the wide distribution of NHEJ activities. Moreover, an Escherichia coli strain also showed NHEJ activity when the Ku and LigD of Dietzia sp. DQ12-45-1b were introduced and expressed in it. Therefore, this method may be a widely applicable genome editing tool for diverse prokaryotic organisms, especially for nonmodel microorganisms.
IMPORTANCE Many nonmodel Gram-positive bacteria lack efficient genetic manipulation systems, but they express genes encoding Ku and LigD. The NHEJ pathway in Dietzia sp. DQ12-45-1b was evaluated and was used to successfully knock out 11 genes in the genome. Since bioinformatic studies revealed that the putative genes encoding Ku and LigD ubiquitously exist in phylogenetically diverse bacteria and archaea, the single-homology-arm linear DNA recombination by the NHEJ pathway could be a potentially applicable genetic manipulation method for diverse nonmodel prokaryotic organisms.
INTRODUCTION
Bacterial genome editing such as deletion, mutation, and insertion of genes is an efficient method to understand gene functions and modify metabolic activities. The most often used gene recombination methods in bacteria include the bacteriophage recombination system using the linear double-stranded DNA (dsDNA) as the recombination substrate (1–3), the single- or double-crossover homologous recombination (HR) using the circular plasmid DNA as the substrate (4, 5), and the transposase-mediated transposition recombination (6, 7). Recently, several genome editing technologies have emerged, which are mediated by the targeted nucleases, including zinc finger nucleases (ZFNs) (8), transcription activator-like effector nucleases (TALENs) (9), and clustered regularly interspaced short palindromic repeat(s) (CRISPR)-associated Cas9 endonuclease (10). All these methods generally require suitable plasmids and a laborious process such as cloning of genes into plasmids. Moreover, it is nearly impossible to find compatible plasmids for diverse prokaryotic organisms in nature, especially for the nonmodel microorganisms. Therefore, developing efficient and suitable genetic manipulation methods is still of great necessity.
Generally, a genetic manipulation method is developed according to a natural biological process. For example, the HR method was developed following a natural HR process to repair DNA double-strand breaks (DSBs) (11), which evolves in all cellular organisms (12). Another DSB repairing process is the nonhomologous end joining (NHEJ) process, which was first discovered in mammalian cells (13). During the eukaryotic NHEJ process, the ends of broken DNA are approximated by the DNA-end-binding protein Ku (Yku70/Yku80 in yeast) and then joined by an ATP-dependent DNA ligase IV/XRCC4/XLF (LXX) complex (LigD, Dnl4/Lif1/Nej1 in yeast) (13). Consequently, a homologous DNA template is not needed (13, 14). Recently, NHEJ was also discovered in prokaryotes with the evidence of the circulation of a linear plasmid DNA in Ku- and LigD-containing mycobacteria (15, 16) and the circulation of a linear DNA in Escherichia coli expressing mycobacterial Ku and LigD (17). In addition, the Ku homodimer and LigD are crucial proteins in the bacterial NHEJ process (18–20), and their presence could indicate NHEJ activities in bacteria (21–23). NHEJ also offers protection to bacteria when only a single copy of the genome is available, such as after sporulation or during stationary phase under environmental stresses (24, 25).
Similar to the HR DSB repairing process, the NHEJ process may also be used for genetic manipulation. In this case, a target homologous sequence can be amplified, PCR fused with a selective marker to form a linear DNA fragment, and then transformed into the recipient bacterium, which expresses Ku and LigD. If the linear DNA can be cyclized by the NHEJ pathway in vivo, the resultant circular DNA may act as a circular-DNA-incompatible plasmid in bacterial genetic manipulation (26). Because the linear/circular DNA without the replication region cannot be replicated, the transformed cells surviving against selective stresses such as antibiotic stress should be mutants with insertions in the target sequences. Consequently, the target gene can be knocked out by using the NHEJ pathway.
Here, we propose a novel and much simpler gene knockout method by using the NHEJ pathway in Dietzia sp. strain DQ12-45-1b. Dietzia strains, found in diverse environments, are powerful n-alkane degraders and potential pathogens. They can utilize a wide range of compounds as the sole carbon source, such as hydrocarbons and aniline (27–29). Moreover, a number of Dietzia strains can be used as potential probiotics to inhibit fecal Mycobacterium avium subsp. paratuberculosis in vitro (30) and might be useful for the treatment of patients with Crohn's disease (31, 32). They are nonmodel Gram-positive bacteria and lack efficient genetic manipulation systems, but they express genes encoding Ku and LigD. The NHEJ pathway in Dietzia sp. DQ12-45-1b was evaluated and was used to successfully knock out 11 genes in the genome. Since bioinformatic studies revealed that the putative genes encoding Ku and LigD ubiquitously exist in phylogenetically diverse bacteria and archaea, the single-homology-arm linear DNA recombination by the NHEJ pathway could be a potentially applicable genetic manipulation method for diverse nonmodel prokaryotic organisms. Compared to previous gene manipulation methods, this method did not need any plasmid or phage systems and could be applied in the Ku/LigD-expressing bacteria mediated by linear DNA fragment. This process will be useful for studying bacteria without an established genetic manipulation method.
RESULTS
NHEJ activity in Dietzia sp. DQ12-45-1b and E. coli expressing Ku and LigD.Genes encoding the Ku and LigD homologs Dt-Ku and Dt-LigD were identified by screening the genome of Dietzia sp. DQ12-45-1b. These homologs showed 57% and 44% amino acid identities with Mt-Ku and Mt-Lig in Mycobacterium (18–20), respectively. To evaluate their NHEJ activities in strain DQ12-45-1b in vivo, the linear DNA fragments LA18 with short overhang ends, HD18 with long overhang complementary ends, and HB18 with long overhang noncomplementary ends (Fig. 1) were transformed into the strain DQ12-45-1b. In addition, negative controls were used in which the LA18, HD18, and HB18 fragments were transformed into E. coli DH5α, which did not express components of the NHEJ pathway. No streptomycin-resistant colonies were detected in the negative controls, while transformation of LA18, HD18, and HB18 fragments into strain DQ12-45-1b generated 2,130, 359, and 229 streptomycin-resistant colonies per microgram of linear DNA, respectively, indicating that the streptomycin-resistant gene originally embedded in the linear DNA fragments was functional in strain DQ12-45-1b (see Fig. S1 in the supplemental material). Since the streptomycin gene-containing DNA could be replicated only if it was in circular form, these results suggested that the linear DNA fragments were successfully cyclized in the strain, indicating that the NHEJ pathway was active in vivo in strain DQ12-45-1b. Among the three linear DNA fragments, it was more difficult for linear DNA with longer 5′-overhang ends to cyclize.
Sequences of DSB junction in LA18 (A), HD18 (B), and HB18 (C). The 5′-A tail is in red. HindIII and BamHI digestion sites are shown in green and purple, respectively. The “CCC” label in LA18 added by PCR is in blue. The nontemplated inserted nucleotides are in pink, and the number of deleted nucleotides is shown at the junction site.
In addition, the three linear DNA fragments were cyclized differently by NHEJ activity. In the case of LA18 with short overhang ends, 61 positive colonies were sequenced and 7 different NHEJ joints were detected (Fig. 1A). The overhung “A” at the ends was deleted before ligation in all colonies, suggesting the presence of a 5′ to 3′ exonuclease activity in strain DQ12-45-1b. Nontemplated insertions and sequence deletions were detected in 38 and 23, respectively, of the 61 colonies. Among the 38 colonies (62% of the 61 colonies) with insertion, three types of nontemplated insertion sequences were identified. Two of them were detected in 25 colonies composed of G+C nucleotides (Fig. 1A). Among the 23 colonies with sequence deletions, 12 colonies contained short-scale (9-bp) deletions and the remaining 11 colonies contained 175- to 308-bp-long deletions. It is notable that the deletions, being short or long, happened at the ends of the side without the “CCC” tag, suggesting that the repeated G+C sequence might protect the linear DNA from exonuclease activity. In contrast, for HD18 and HB18 with long overhang ends, long unidirectional deletions of up to 148 bp were detected at both ends (Fig. 1B and C). The results further confirmed the DNA polymerase and exonuclease activities in the NHEJ process in strain DQ12-45-1b (15).
To evaluate the NHEJ activities of Ku and LigD in strain DQ12-45-1b, a ku mutant strain was constructed. It was notable that the ligD mutant failed to be obtained even though two different methods had been tried to disrupt the ligD gene from the strain (data not shown). The result suggested that ligD might be an essential gene for strain DQ12-45-1b. The linear DNA L53 with a single adenine overhang at 3′ end was transformed into a wild-type L53, a ku gene mutant, and a mutant with ku complementation to observe the NHEJ activities. No ampicillin-resistant colonies were obtained from the ku mutant cells, while there were 463 and 351 ampicillin-resistant colonies per microgram of linear DNA in the wild-type cells and the mutant with ku complementation (see Fig. S2 in the supplemental material), indicating that the linear DNA L53 was cyclized in the cells. In addition, the L53 fragment was also transformed into E. coli DH5α, and no ampicillin-resistant colonies were detected, suggesting that there was no circular plasmid contaminant.
To further confirm the cyclization capability of Dt-Ku and Dt-LigD, we transformed pUC19, pUC19-Ku, pUC19-LigD, and pUC19-Ku-LigD first into E. coli DH5α following electrotransformation of P28 and successfully recovered colonies in the presence of kanamycin, while no kanamycin-resistant colony was detected in the bacteria, including pUC19, pUC19-Ku, and pUC19-LigD. From the kanamycin-resistant colonies, plasmids were extracted and digested by NdeI to further verify the circulation results. Three bands with the expected sizes of about 1.2 kb, 5.2 kb, and 3.8 kb were observed, corresponding to the digestion products of pUC19-Ku-LigD and cyclized P28 (see Fig. S3 in the supplemental material), respectively. These results supported that Dt-Ku and Dt-LigD could function in E. coli to cyclize the linear DNA and both activities were needed. The electrotransformation of E. coli was 2.4 × 104 CFU/μg DNA. The cyclization efficiency was calculated as 228 CFU/μg DNA in E. coli DH5α expressing Dt-Ku and Dt-LigD. The ratio of Amp/Kan-resistant colonies was about 100:1. All the assays were performed in triplicates.
Recombination of linear DNA in Dietzia sp. DQ12-45-1b via the NHEJ pathway.Since linear DNA could be cyclized in strain DQ12-45-1b by NHEJ activity in vivo, we assumed that it could act as a nonreplicable plasmid that could further be used to inactivate a target gene by homologous recombination. To verify this hypothesis, we constructed a linear alk-Km DNA fragment containing a single-homology-arm DNA fragment of the alkW1 gene and the kanamycin resistance cassette. After transforming 200 ng of this linear alk-Km DNA fragment into Dietzia sp. DQ12-45-1b, colonies were recovered in the presence of kanamycin. Among them, eight colonies were randomly selected and subjected to PCR amplification with primers alkW1-F and alkW1-R to amplify the alkW1 gene. DNA fragments with the expected size (about 3.6 kb) were amplified from four of the eight colonies, indicating that one-half of the colonies presented the correct insertion into the alkW1 gene (Fig. 2A). Sequencing of these DNA fragments further confirmed the successful insertion of linear DNA into the alkW1 gene in the genome (Fig. 2B). The recombinant efficiency was approximately 3 × 10−6.
Identification of the alkW1 mutant in Dietzia sp. DQ12-45-1b by single-homology-arm linear DNA. (A) Ethidium bromide-stained agarose gel showing typical results of PCR identification of eight colonies. Lanes: M, DNA marker; 0, blank control; 1 to 8, different colonies. The bands designated with the black arrow are the alkW1 mutant colonies with correct sizes. The bands designated with the red arrow are the unchanged alkW1 colonies. (B) Schematic diagram showing the linear DNA recombination pattern. P45, promoter p45 (119 bp); Hm, single-homology arm (410 bp); Km, kanamycin (795 bp).
We also examined whether the linear fragment was cyclized before recombination. We collected cells 4 h after transformation and extracted the total DNA of the cells. Then, we amplified it using primers P45-BR and 143-F. An ∼540-bp fragment, which could be amplified from the cyclized fragment only, was found (see Fig. S4 in the supplemental material). The result suggested the cyclization of linear DNA fragment before recombination.
To further verify the gene insertion, we carried out Western blotting with the proteomes of Dietzia sp. DQ12-45-1b and the alkW1 mutant, which were grown in glucose-tryptone-yeast extract (GPY) medium (see Materials and Methods) or a mineral medium with hexadecane as the sole carbon source (MF+C16). AlkW1 was induced in Dietzia sp. DQ12-45-1b by hexadecane but was not detected in the alkW1 mutant strain cultured either in GPY or with hexadecane (Fig. 3A). As expected, the growth of the alkW1 mutant strain in hexadecane as the sole carbon source was inhibited within 8 days of incubation (Fig. 3B). These results proved that the alkW1 gene was successfully inactivated by the single-homology-arm DNA fragment integration. Using the single-homology-arm linear DNA recombination, we also successfully disrupted nine genes of Dietzia sp. DQ12-45-1b (Table 1), as confirmed by the PCR analysis (see Fig. S5 in the supplemental material), the presence of green fluorescence (Fig. S6), and sequencing (see File S1 in the supplemental material). The recombinant efficiencies were approximately about 1.2 × 10−6 to 6.5 × 10−5 (Table 1). Besides, the successful inactivation of another gene, alkX, the regulator of alkW1, using this method (33), suggested the reliability of this approach. In addition, the mutant was continuously cultured for 30 reinoculations in GPY medium without antibiotics with no change in its genotype (data not shown), which demonstrated that the insertion was stable and heritable.
Functional characterization of the alkW1 mutant strain. (A) Western blot analysis of the AlkW1 protein expression in Dietzia sp. DQ12-45-1b and the alkW1 mutant strain. Cells grew in GPY as a negative control; WT cells grew in MF+C16 as a positive control; WT, wild-type Dietzia sp. DQ12-45-1b; MT, alkW1 mutant strain; GPY, cells grown in GPY medium; MF+C16, cells grown in minimal medium containing hexadecane. (B) Growth curve of strain DQ12-45-1b and alkW1 mutant strain in minimal medium with hexadecane as the sole carbon source.
List of mutant genes in Dietzia sp. DQ12-45-1b
Distribution of Ku and LigD in prokaryotic organisms.By searching the ku and ligD homologous genes in the available 25,270 bacterial and 528 archaeal genomes using KEGG Orthology (KO) annotation, we identified 6,118 ku homologous genes in 5,098 bacterial and 13 archaeal genomes belonging to 14 bacterial phyla and Euryarchaeota. We also identified 18,952 ligD homologous genes in 7,631 bacterial and 64 archaeal genomes belonging to 26 bacterial phyla and 6 archaeal phyla (Fig. 4). Among them, 4,783 bacterial and 13 archaeal genomes contained both genes. Strikingly, about 76% and 77% of the total Actinobacteria genomes analyzed contained ku and ligD homologous genes, respectively, indicating that NHEJ might be common in Actinobacteria. Phylogenetic analysis showed that ku genes could be grouped into six clusters (Fig. 5), largely matching their taxonomic classification, suggesting the conservative property of the ku genes. In contrast, the ligD genes could be grouped into five clusters and their phylogenetic tree topology was very different from their taxonomic classification, suggesting a possible vigorous gene transfer of ligD genes among microorganisms (Fig. 5).
Distribution of ku and ligD homologous genes in microbial genomes at the phylum level. (A) Distribution of ligD in microbial genomes at the phylum level; (B) distribution of ku in microbial genomes at the phylum level.
Phylogenetic relationships of Ku and LigD based on amino acid sequences in prokaryotic organisms. The phylogenetic trees were constructed using the neighbor-joining method in MEGA6 (59). The trees were bootstrapped with 500 replicates, indicated at the respective nodes.
The ku genes were generally juxtaposed with ligD genes (Fig. 6A), for which a functional association between Ku and LigD was suggested (21). In contrast, the Dt-Ku- and Dt-LigD-coding genes were not adjacent in Dietzia sp. DQ12-45-1b but were within 9.8 kb of each other (Fig. 6A). Similar gene arrangements were detected in other Dietzia species such as Dietzia alimentaria 72, Dietzia cinnamea P4, and Dietzia sp. strain UCD-THP (see Fig. S7A in the supplemental material), whose Ku shared 72 to 81% amino acid identity with Dt-Ku. All the Ku protein sequences retrieved had a Ku core functional domain, which could bind DNA ends and transiently bring them together (Fig. 6B), suggesting Ku functions in these strains (21). All the LigD sequences presented an ATP-dependent DNA ligase domain (Fig. 6C), which was essential in NHEJ (13, 14). The Ku core and ATP-dependent DNA ligase domain found in all Ku and LigD proteins suggested intact NHEJ functions in all these strains containing both ku and ligD. Besides, Dt-LigD and its homologs from other Dietzia species contained a polymerase domain and a phosphoesterase domain from the N terminus to the C terminus, which was similar to that of the homolog Mt-LigD in Mycobacterium (Fig. 6C and S7B).
Organization of genes encoding Dt-Ku and Dt-LigD and their protein domain architecture. (A) Protein domain architecture of Ku. Ku core, the Ku core domain; HC2, histone H1-like nucleoprotein domain; Rho, helix-extension-helix domain from the bacterial transcription termination factor Rho. (B) Protein domain architecture of LigD. (C) Gene organization of ku and ligD operons. The direction of the arrow indicates the direction of the transcription. The numbers above the arrows are the GenBank accession IDs of their corresponding genes. Sc, Streptomyces coelicolor; Bs, Bacillus subtilis; Bh, Bacillus halodurans; Mt, Mycobacterium tuberculosis; Pa, Pseudomonas aeruginosa; Ml, Mesorhizobium loti; Af, Archaeoglobus fulgidus.
DISCUSSION
As the major DSB repair pathway, HR has been extensively studied in E. coli and widely used for gene manipulation. This classical genetic engineering method uses incompatible or suicide plasmids and requires a time-consuming process. A phage-mediated HR, usually referred to as “recombineering,” constructs linear DNA with short homologies (34) and was conveniently applied in genomic manipulation of E. coli, Mycobacterium, Salmonella, and Shigella (35–38). Unfortunately, the known phage systems are not applicable for recombineering many other bacteria of interest. For example, 666 novel species were published in 2013 (LPSN [list of prokaryotic names with standing in nomenclature]; http://www.bacterio.net/), a lot of which are of great importance in bioremediation, chemical production, and human health. However, it is barely possible to find compatible phage systems as well as shuttle vectors for these strains. Although ZFNs, TALENs, and CRISPR have been developed for precise genome editing in eukaryotes (8–10), they also need shuttle vectors to express a nuclease in the host strain. The lack of shuttle vectors expressing these systems is additionally hindering their use for diverse prokaryotes. In this study, we proposed a novel and much simpler gene inactivation method by using the NHEJ pathway mediated by Ku and LigD proteins in Dietzia sp. DQ12-45-1b without additional vectors or genome editing systems.
After the linear DNA with single homology arm was transformed into Dietzia sp. DQ12-45-1b cells, it performed self-cyclization in vivo, recombined with the target homologous sequence in the genome by HR, and produced the mutant strain (Fig. 7) (33). Although the efficiency was not very high, around 10−5 to 10−6, when targeting the genes in Dietzia sp. DQ12-45-1b, it may be acceptable for those newly discovered and important bacteria that lack a compatible genetic manipulation system. Although NHEJ was first found in mammalian cells and thought to be specific to eukaryotic organisms, a number of studies showed that NHEJ also exists in prokaryotic organisms (15, 16, 18–23, 39). Ku was proved to be essential in NHEJ in strain DQ12-45-1b. It was interesting that we tried to disrupt the ku and ligD genes using two methods, the single-homology-arm recombination method as described in this paper and the double-homologous recombination method described previously (35), but neither ku nor ligD mutants were obtained using the single-homology-arm linear DNA recombination method. In contrast, the ku gene could be replaced using the double-homologous recombination method, which expressed additional recombination enzymes. The result suggested unknown roles of Ku in the recombination process using single-homology-arm linear DNA. It was notable that no ligD mutant was obtained using either of the two methods. LigD, as the main ligase of NHEJ, could repair DSBs that had a lethal effect on cell mitosis unless repaired in time (40–43). According to previous reports, ligD could be knocked out from Mycobacterium smegmatis, because there was another ATP-dependent DNA ligase (LigC) that could provide a backup function of LigD-independent error-prone repair of blunt-end DSBs in the strain (15). However, in Dietzia sp. DQ12-45-1b, only one ligD was identified, and its disruption might be lethal. Furthermore, there was no cyclization of linear plasmid DNA in the ku mutant of strain DQ12-45-1b, implying that ku might be critical in the process of linear DNA cyclization.
Mechanism of single-homology-arm linear DNA for target gene mutant. The linear DNA transformed into bacterial cells is self-cyclized and then recombined with the chromosomal gene.
In addition, 5,111 and 7,695 of the 25,798 bacterial and archaeal genomes available in the Integrated Microbial Genomes (IMG) database contain ku and ligD homologous genes, respectively. The distribution of ku and ligD genes may suggest the wide distribution of the NHEJ pathway in microorganisms. Consequently, the introduction of mutagenesis during the cyclization might have played an important role in genome diversification, which could increase the natural microbial diversity and adaptability. In addition, the wide distribution of ku and ligD genes may also suggest that the single-homology-arm linear DNA recombination by using the NHEJ pathway could be commonly used as a gene inactivation method for diverse prokaryotic organisms, especially for newly discovered but important microorganisms. Especially, more than 3/4 of the genomes of Actinobacteria, with high GC content and refractory to genetic manipulation, express the ku and ligD genes, suggesting that the single-homology-arm linear DNA recombination is a potentially good method, at least for Actinobacteria. However, more studies are needed to test the versatility of this method.
For organisms lacking Ku and LigD, the single-homology-arm linear DNA recombination method requires the construction of a compatible and inducible Ku and LigD system, i.e., one plasmid needs to be constructed to express Ku and LigD. In this case, the expression of Ku and LigD could be controlled by appropriate inducible promoters as described for the expression of Mt-Ku and Mt-LigD in E. coli (17). Once the plasmid expressing Ku and LigD is constructed, it can be used for disrupting any gene without constructing new plasmids. Thus, this method is still simpler than the classic approaches requiring incompatible or suicide plasmids.
This method presents two major concerns. One is how to improve the cyclization efficiency, and the other is how to reduce the risk of linear DNA degradation by nucleases in vivo. In prokaryotic organisms, NHEJ activity was particularly efficient during the stationary phase to counteract DSBs induced by heat, desiccation, and other factors (44). However, the electroporation efficiency reached its highest value when the recipient cells were in the early exponential phase (45). A balance is therefore necessary to increase the cyclization and recombination activities and to increase the transforming efficiency. Several methods have been developed, including heating the cells, to increase both the transforming efficiencies and the NHEJ process (45), which may promote the efficiency of this method.
Unlike in yeast, linear dsDNA in bacteria such as E. coli can be rapidly degraded by nucleases (26). This phenomenon was also observed in strain DQ12-45-1b during sequence deletions when cyclizing the linear DNA LA18, HD18, and HB18 in vivo (Fig. 1). In E. coli, recB and recC encode the major nucleases degrading the dsDNA and are essential for the normal activities of the bacterium. Deletion of recB and recC resulted in poor growth of E. coli, producing up to 80% of nonviable cells (12). To inhibit the nuclease activities and promote the recombination efficiency, several phage systems such as thr RecET system and λ Red system were developed, which again require plasmids (36, 46). For example, in the λ Red system, the exo, bet, and gam genes are under the lac promoter control on a multicopy plasmid. Gam function could inhibit RecBCD nuclease to protect the transformed linear DNA from being degraded. Then, with the help of Exo and Bet, gene recombination was conducted (12). Another trial was to recombine linear DNA in wild-type E. coli containing the RecBCD nuclease (47). In this case, special sites were engineered in the linear DNA that decreased the RecBCD activities (47). Similarly, when we added a triplet cytidine (CCC) tag at one terminus of the linear DNA (LA18) by PCR, the long deletions were unidirectional and occurred at the opposite terminus with lower G+C contents (Fig. 1), suggesting that addition of the CCC tag could protect the target sequence from being digested by nucleases and increased the recombination efficiency. Additionally, electroporation itself has been suggested to reduce DNA degradation by RecBCD nuclease and to allow recombination with linear dsDNA (48).
If this method is reliable, it could be used not only to inactivate one gene by inserting the selective cassette into the chromosome (see Fig. S8A in the supplemental material) but also for gene overexpression and exogenous gene insertion. For example, for overexpression of a gene, the cassette consists of a selectable integrated gene that would be used as the single-homology-arm DNA (Fig. S8B). For exogenous DNA insertion, the cassette could be linked to an exogenous DNA before the homologous arm (Fig. S8C).
In conclusion, we propose an easy and fast method for gene inactivation in Dietzia sp. DQ12-45-1b, using a linear DNA with single homology arm. This method uses the NHEJ pathway and might be applicable to prokaryotic organisms that harbor genes encoding Ku and LigD, especially for the newly identified and nonmodel species. Because the linear DNA with cassettes and homologous DNA could be amplified by fusion PCR conventionally, these procedures should be rapid, convenient, and reliable.
MATERIALS AND METHODS
Bacterial strains, plasmids, media, and primers.The bacterial strains and plasmids used in this study are listed in Table 2. All primers are listed in Table 3. Dietzia sp. DQ12-45-1b, isolated from oil production water, has been intensively studied in our laboratory (28, 33, 45, 49–52). Dietzia sp. DQ12-45-1b was cultured in GPY medium (1% [wt/vol] glucose, 1% [wt/vol] tryptone, 0.5% [wt/vol] yeast extract) at 30°C on a rotary shaker at 150 rpm. The alkane hydroxylase-rubredoxin fusion gene (alkW1) (49) mutant of Dietzia sp. DQ12-45-1b (alkW1 mutant) was grown in GPY medium with 40 mg/liter kanamycin. To determine their functions in alkane degradation, the alkW1 mutant strain and Dietzia sp. DQ12-45-1b were grown in minimal medium (4) supplemented with 0.1% (vol/vol) hexadecane as the sole carbon source at 30°C, as described previously (49). E. coli DH5α was grown in lysogeny broth (LB) medium at 37°C. E. coli strains and Dietzia strains harboring plasmids were grown with appropriate antibiotics (ampicillin, 100 μg/ml; kanamycin, 30 μg/ml; streptomycin, 30 μg/ml).
List of strains and plasmids used in the studya
Primers used in this worka
The vector pNV18-PEK (Table 2) was constructed as follows. First, the promoter p45 DNA fragment was amplified from plasmid pNV18-Dsred (53) with primers p45-EF and p45-BR, the egfp gene from pK18-egfp (our unpublished result) with primers egfp-BF and egfp-HR, and the kanamycin (Km) resistance gene from plasmid pK18 (54) with primers Km-HF and Km-PR. Then, the promoter p45 was cloned into pXL1801 (33) with EcoRI and BamHI restriction sites, egfp with BamHI and HindIII restriction sites, and the Km resistance gene with HindIII and PstI restriction sites sequentially, yielding the plasmid pNV18-PEK. The plasmid pJV53-PA was also constructed. Briefly, a large fragment of vector pJV53 was first amplified with primers 53-HF and 53-ER (Table 3). The resulting 3,013-bp linear fragment lacking the acetamidase promoter and Chec9c 60-61 was digested with EcoRI and HindIII. The promoter p45 was amplified from pNV18-PEK with primers p45-EF and p45-(amp)-R (Table 3), and the ampicillin resistance gene was amplified from pUC19 with primers amp-(p45)-F and amp-HR (Table 3). The p45 promoter and ampicillin resistance gene were fused by PCR with primers p45-EF and amp-HR, resulting in PA DNA fragment. Then, PA was ligated into linear pJV53 by the EcoRI and HindIII restriction sites to obtain the plasmid pJV53-PA.
Sequence analysis.Genes encoding the Ku and LigD homologs in the genome of Dietzia sp. DQ12-45-1b, designated Dt-Ku and Dt-LigD, were identified by comparing the proteome of strain DQ12-45-1b (our unpublished result) against the nonredundant (NR) database of protein sequences at the National Center for Biotechnology Information (NCBI) using BLASTP (55). The conserved domains of proteins were determined by comparing the protein sequences against the conserved domain database (CDD) at NCBI (56). To identify the Ku and LigD homologous genes in prokaryotic organisms, we searched the KEGG Orthology (KO) (57) for IDs K10979 and K01971, which indicated Ku and LigD, respectively, in the Integrated Microbial Genomes (IMG) system (58) against a total of 25,270 bacterial genomes and 528 archaeal genomes (as of April 2015). Sequence alignment and phylogenetic analysis were performed in MEGA6 (59).
NHEJ assay in Dietzia sp. DQ12-45-1b.To identify the circular efficiency of different linear DNA ends, three types of linear DNA were generated. The linear DNA fragment with 5′-A tails (LA18) was amplified from 4 ng plasmid pNV18-Sm (53) using LA Taq (TaKaRa, Tokyo, Japan) under the conditions suggested by the manufacturer with the primers pNV18-F and pNV18-R, followed by treatment with DpnI digestion (TaKaRa, Tokyo, Japan) for 4 h at 37°C to eliminate the circular plasmid pNV18-Sm template. Second, the plasmid pNV18-Sm was digested by HindIII to produce 5′-overhang DSB fragments HD18 and by HindIII and BamHI simultaneously to produce 5′-overhang DSB fragments HB18 overnight, respectively (Fig. 1). All three linear DNA fragments contained the intact streptomycin resistance gene (see Fig. S9 in the supplemental material). Gel purification was performed to remove the undigested plasmids and collect the linear plasmid only with a DNA purification kit (Tiangen Biotech, Beijing, China). Five microliters purified linear DNA fragments that could be replicated in the bacteria after cyclization, LA18 (108 ng), HD18 (437 ng), and HB18 (188 ng), were then separately transformed into 100 μl competent cells of strain DQ12-45-1b (∼1010 CFU/ml; the transformation efficiency was 106 CFU/μg DNA). After transformation, cells were recovered at 30°C for 4 h, followed by plating on LB agar containing streptomycin. Positive clones were numbered and picked after 3 days incubation. The method used for competent cell preparation and electrotransformation was previously described (45).
The resultant streptomycin-resistant colonies on LB agar were then suspended in 10 μl double-distilled water and incubated at 98°C for 30 min to release DNA from the cells, followed by centrifugation at 20,000 × g for 5 min. The DNA-containing supernatant was then subjected to PCR using rTaq (TaKaRa, Tokyo, Japan) for 26 cycles with the primers IdenF and IdenR to amplify the joint region of the circular DNA. The PCR product was purified and cloned into pGEM-T (Promega, Madison, WI, USA). Recombinant plasmids were transformed into E. coli DH5α for white-blue screening, and the white colonies were sequenced to assess NHEJ activities.
To determine the role of the ku gene in Dietzia sp. DQ12-45-1b, we constructed a ku-inserted mutant with the streptomycin resistance gene using a double-homologous-recombination method as described previously (see Fig. S10 in the supplemental material) (35, 60, 61); the primers used are listed in Table 3. The Ku expression vector for complementation studies were constructed as follows: the ku gene was amplified from Dietzia sp. DQ12-45-1b chromosomal DNA with the primers Ku-BF and Ku-HR (Table 3), and then the 1,046-bp DNA fragment was ligated into pNV18-PEK using BamHI and HindIII restriction sites to construct a new plasmid, pNV18-PKK (Table 2). Vector pNV18-PKK was introduced into ku mutant cells via electroporation as described previously to obtain the complementary mutant cell (45). The effects of the ku gene on the circular function of linear DNA in Dietzia sp. DQ12-45-1b were identified on the ku-complementary mutant cell using the linear plasmid DNA pJV53-PA (L53) with a p45 promoter and an ampicillin resistance gene that could be replicated in the bacteria after cyclization (see Fig. S11 in the supplemental material), which was amplified from plasmid pJV53-PA with the primers p45-F and 53-ER (Table 3) followed by being processed with DpnI digestion (TaKaRa, Tokyo, Japan) for 4 h at 37°C to eliminate the circular plasmid pNV18-Sm template before being transformed into cells as described above, by taking the treatments used on the WT cells and ku mutant cells transformed with pNV18-PEK as the negative controls (Table 2). After the 3-day incubation, ampicillin resistance colonies were counted as described above.
Expression of Dt-Ku and Dt-LigD and NHEJ assay in E. coli.The genes coding for Dt-Ku protein and Dt-LigD were amplified from the genome of strain DQ12-45-1b using PrimeSTAR HS DNA Polymerase (TaKaRa, Tokyo, Japan) according to the manufacturer's instructions with the primers KuF and KuR for Dt-Ku and LigDF and LigDR for Dt-LigD. The PCR products were purified and digested with the KpnI and EcoRI for Dt-Ku and with HindIII and KpnI for Dt-LigD. The plasmid pUC19-Ku and pUC19-LigD were then constructed by cloning Ku and LigD DNA sequences into the corresponding sites of the plasmid pUC19. Similarly, the plasmid pUC19-Ku-LigD was constructed by cloning these two fragments into the HindIII and EcoRI sites of the plasmid pUC19, using standard methods (62), and then transformed into E. coli DH5α to construct the strains expressing Dt-Ku or Dt-LigD singly and both of them, respectively.
To assess Dt-LigD and Dt-Ku activities, a heterogeneous expression test was performed. A linear DNA fragment, P28 (3,818 bp, containing the pBR322 origin and kanamycin resistance gene), which can be replicated in E. coli after cyclization, was amplified from pET-28a(+) (Qiagen, Hilden, Germany) with the primers pET-28F and pET-28R. Specifically, 4 ng pET-28a(+) was used as the template for PCR with PrimeSTAR HS DNA Polymerase (TaKaRa, Tokyo, Japan) according to the manufacturer's instructions. Then, the PCR product was treated with DpnI digestion (TaKaRa, Tokyo, Japan) for 4 h at 37°C to degrade the circular plasmid DNA, and only the 3.8-kb band was collected for further purification with a DNA purification kit (Tiangen biotech, Beijing, China). pUC19, pUC19-Ku, PUC19-LigD, and pUC19-Ku-LigD (100 ng) were then transformed by 42°C heat shock for 90 s into 100 μl E. coli DH5α. After transformation, 900 μl of room temperature Super Optimal broth with catabolite repression (SOC) medium was added to the tubes, and the cells were allowed to recover at 37°C with shaking at 300 rpm for 1 h. Cells (50 μl) were cultured on LB agar containing 100 μg/ml ampicillin. The bacteria expressing no target protein or expressing Ku, LigD, and Ku-LigD were prepared into competent cells and then electrotransformed with fragment P28 (500 ng) according to standard instructions (49). All the cells were then collected by centrifugation (1,500 × g, 5 min) and cultured on LB agar containing 100 μg/ml ampicillin and 40 μg/ml kanamycin. The transformation of pUC19 and P28 into E. coli DH5α was used as a negative control. The kanamycin-resistant colonies indicated that the P28 was cyclized, allowing the expression of the kanamycin-resistant gene. Plasmids were then extracted from the positive colonies and digested with NdeI, followed by electrophoresis to determine the lengths of digested fragments.
Gene disruption in Dietzia sp. DQ12-45-1b using single-homology-arm linear DNA fragment.We used the alkW1 gene as an example to determine whether NHEJ could be used for target gene manipulation. First, the homologous DNA fragment (from nucleotide [nt] 143 to nt 552 of the alkW1 gene) was amplified from the genome of strain DQ12-45-1b using primers 143-F and 552-R. For easy identification, a triplet cytidine (CCC) tag was additionally designed at the 5′ end of primer 552-R. Second, the kanamycin resistance cassette P45-EGFP-Km with a P45 promoter and an egfp gene as the reporter was amplified from the plasmid pNV18-PEK using the primers P45-F and Km-R. Finally, the homologous DNA and P45-EGFP-Km cassette were fused and amplified by fusion PCR (63) using the primers P45-F and 552-R. The generated fragment (alk-Km) with homologous DNA of alkW1 and kanamycin resistance cassette was purified, and 5 μl of the purified product (500 ng) was transformed into Dietzia sp. DQ12-45-1b as previously described (45). Cells were then grown on LB plate with kanamycin for 4 days. The kanamycin-resistant colonies were subjected to PCR amplification of gene alkW1 with the primers alkW1-F and alkW1-R, which was further verified by sequencing. The recombinant efficiency was expressed as the number of recombinants per microgram DNA divided by the cell competency, which was indicated by the transformation efficiency with plasmid pNV18-Sm. The expression of AlkW1 in the cells cultured in liquid GPY medium or minimal medium with hexadecane as the sole carbon source was detected by Western blot analysis as described previously (33).
To verify whether this method could be used for other genes, we selected seven genes encoding histidine kinase of the two-component system and two other genes (Table 1) for disruption. The homologous linear DNA fragments were constructed and transformed as described above. The kanamycin-resistant colonies were selected and verified by PCR, fluorescence microscopic analysis, and sequencing.
ACKNOWLEDGMENTS
We thank Guang Hu for generously providing the antibody of polyclonal mouse anti-AlkW1, Hui Fang for her kindly help in bioinformatic analysis, and Xiao-Yu Qin for her help in Ku knockout, and we thank Yue-Qin Tang for reading the manuscript and for discussions.
This work was supported by the National Natural Science Foundation of China (31200099, 31225001, and 31300108) and the National High Technology Research and Development Program (“863” Program: 2012AA02A703 and 2014AA021505).
FOOTNOTES
- Received 12 April 2018.
- Accepted 4 July 2018.
- Accepted manuscript posted online 20 July 2018.
- Address correspondence to Xiao-Lei Wu, xiaolei_wu{at}pku.edu.cn.
S.L. and Y.N. contributed equally to this work and are co-first authors.
Citation Lu S, Nie Y, Wang M, Xu H-X, Ma D-L, Liang J-L, Wu X-L. 2018. Single-homology-arm linear DNA recombination by the nonhomologous end joining pathway as a novel and simple gene inactivation method: a proof-of-concept study in Dietzia sp. strain DQ12-45-1b. Appl Environ Microbiol 84:e00795-18. https://doi.org/10.1128/AEM.00795-18.
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.00795-18.
REFERENCES
- Copyright © 2018 American Society for Microbiology.
