Introduction

Corynebacterium glutamicum (C. glutamicum) is widely used for the production of L-amino acids, vitamins, organic acids, fuel ethanol and other products1,2,3, and manipulation of its genes is essential for metabolic engineering and synthetic biology4, 5. Traditional strategies of counter-selectable systems6,7,8 and Cre/loxP site-specific recombination9,10,11 have been successfully used in the model strain of C. glutamicum ATCC13032, for gene deletion; however, the genetic manipulation tools are still limited for the non-model strain, such as C. glutamicum ATCC14067, which is also widely used to study of amino acid production12,13,14,15.

The widely used gene deletion approach of the conventional sacB counter-selectable system in C. glutamicum ATCC13032 is based on two rounds of homologous recombination, during which only 2% of the events correspond to double-crossover events and which usually requires more than 10 days16. Moreover, it does not work in C. glutamicum ATCC140676, 17, which may be due to specific or unclear genetic information among different Corynebacteria 12. The Cre/mutant loxP system is useful for sequential gene deletion and large-scale genome engineering in C. glutamicum 9,10,11. Cre recombinase can catalyze reciprocal site-specific recombination between the mutated loxP sites, lox71 and lox66, and can generate lox72, which cannot be recognized by Cre. However, the deletion based on the Cre/mutant loxP system involves two plasmids and two rounds of transformation in C. glutamicum 17, which is also tedious and laborious. Gene deletion via a linear double-strand DNA (dsDNA) cassette with a native recombineering system, which exists in several species18, 19, can greatly reduce the workload; however, double-crossover events rarely occur without the integrative vector in C. glutamicum 20.

The homologous recombination via the phage recombinase has revolutionized bacterial genetics since 199821. It is easy and efficient to construct gene deletion mutants by recombineering via dsDNA in gram-positive and gram-negative bacteria, including Lactobacillus plantarum 22 and Pseudomonas syringae 23. The λ red system and RecET system from E. coli are typical for recombineering24, 25. The λ red recombination system is composed of three phage-encoded proteins, Exo, Beta and Gam25 and the RecET system is composed of only two phage-encoded proteins, RecE and RecT26. The proteins λ Exo and RecE can act as the 5′-3′dsDNA-dependent exonuclease that can resect a linear dsDNA to expose a 3′-ended single-stranded DNA (ssDNA) tail; ssDNA-annealing proteins, λ Beta and RecT, which are members of the single-stranded DNA annealing protein (SSAP) family, also called recombinase, bind to the ssDNA tail and promote the annealing of complementary DNA strands, strand exchange and strand invasion.

The RecFACS system (RecT-mediated single-strand recombination via fluorescence-activated cell sorting) is a fast method to introduce genomic mutations and achieves ultrahigh-throughout detection and isolation of productive recombinants in C. glutamicum ATCC1303227. However, the Lrp-based biosensors can only detect intracellular several amino acids and there are no currently available optical sensors for the most biotechnologically relevant compounds28. Therefore, RecFACS is limited for other small moleculars in general and the screening of recombinants is still problematic in C. glutamicum. The recombineering with dsDNA needs to be employed in C. glutamicum to accomplish effective recombineering and simple selection.

Thus, in this study, we explored the recombination activity of the orthologous exonuclease-recombinase pairs, RecE + RecT (RecET) from Rac phage and Exo + Beta from Lamda phage, and RecET-like pairs, Orf47 + Orf48 from the A118 phage of Listeria monocytogenes (L. monocytogenes) and OrfB + OrfC from Legionella pneumophila (L. pneumophila)29, in C. glutamicum ATCC14067. We also designed a self-excisable linear dsDNA cassette combining the Cre/loxP system to perform markerless deletion via RecET recombineering system. This strategy provides a new simple and efficient markerless deletion strategy for C. glutamicum ATCC14067.

Results

The linear dsDNA recombination efficiency of the different orthologous exonuclease-recombinase pairs in C. glutamicum ATCC14067

The orthologous exonuclease-recombinase pairs Exo/Bet, RecE/RecT (RecET), Orf47/Orf48 and OrfB/OrfC, which have been identified to perform the function of linear dsDNA recombineering in E.coli 29, were selected to test the recombineering activity in C. glutamicum ATCC14067. A 0.5 μg linear dsDNA cassette of CrtB/400-Kan was used for the verification (Supplementary Fig. S1). The exonuclease-recombinase pairs of RecET and OrfB/OrfC could catalyze linear dsDNA recombination (Table 1). Orf47/Orf48 showed weak activity, with only ~13 kanamycin resistant colonies, and Exo/Bet did not produce any recombinants in C. glutamicum ATCC14067 (Table 1). No recombinant was observed in the control, which contained the linear dsDNA cassette in C. glutamicum ATCC14067 without an exonuclease-recombinase pair (Table 1), indicating that linear dsDNA could not lead to recombination in C. glutamicum ATCC14067 via the cell’s own recombination system and that the recombination was mediated by the exonuclease-recombinase pairs. The replicative plasmid pZ9, which contained kanamycin resistance cassette, was used to explore the DNA uptake capability, and all transformations produced 3.1 × 103 ~ 3.8 × 103 kanamycin resistant colonies (Table 1), which suggested that expression of the exonuclease-recombinase does not affect DNA uptake. This result is similar to what has been observed for C. glutamicum ATCC1303227.

Table 1 Comparison of the recombination efficiencies of different orthologous exonuclease-recombinase pairs for linear dsDNA.
Full size table

Optimization of the RecET recombineering

The RecET pair showed the highest recombination efficiency in C. glutamicum ATCC14067, we next varied the conditions to enhance the recombination efficiency. The recombineering frequency of the RecET system is dependent on the length of DNA homology30 and the length of homology arms were optimized to improve the linear dsDNA recombination efficiency in C. glutamicum ATCC14067. 0.5 µg CrtB-Kan cassettes with the homology lengths from 100 to 2,000 bp (Supplementary Table S1) were used to investigate the change in recombination frequency. The number of kanamycin resistant colonies increased as homology arms length increased from 100 bp to 800 bp (Fig. 1a). Beyond a homology region length of 800 bp, there was no significant increase or even decrease in the number of colony (Fig. 1a), which is similar with the RecTEPsy recombineering in Pseudomonas syringae 31. The homology arms length may have reached a threshold length for efficient recombineering via RecET.

Figure 1
figure 1

Optimization of linear dsDNA recombineering parameters. (a) Effect of the homology length of linear dsDNA on recombination efficiency. 100, 200, 300, 400, 500, 800, 1000, 1200, 1500, 2000 bp left and right homology arms of the crtB gene combining with the Kan cassette were used for the investigation. 0.5 μg dsDNA cassettes were used for electroporation. The recovery time was 5 h. (b) Effect of the quantity of linear dsDNA on recombination efficiency. 0.1–4 µg CrtB/800-Kan cassettes were used for electroporation. The recovery time was 4 h. (c) Effect of the recovery time on recombination efficiency. The recovery time was 0 h to 5 h. 0.5 μg CrtB/800-Kan cassettes were used for electroporation. The total number of colony of per OD600 also is shown. (d) Effect of the induction time on recombination efficiency. The RecET recombinases were induced for 2–8 h. 0.5 μg CrtB/800-Kan cassettes were used for electroporation and the recovery time was 4 h. OD600 of the harvested culture is also shown. (e) Effect of the phosphothiolated linear dsDNA on recombination efficiency. 0.5 μg CrtB/800-Kan cassettes were used for electroporation. The recovery time was 4 h. All of the Kmr cfu is the number of kanamycin resistance colony per mL. Datas are the means of at least three experiments with standard deviations by error bars.

Full size image

Then, we assayed the effect of linear dsDNA concentration on recombineering efficiency. 0.1–4 µg CrtB/800-Kan cassettes were used for the electroporation. There was a significant increase in the number of colonies with increasing of dsDNA substrates concentration, with as many as ~653 recombinants when adding the 4 µg CrtB/800-Kan cassettes (Fig. 1b). This result was different from the recombineeing in Pseudomonas syringae 31 and Lactobacillus plantarum 22 that the frequency decreased with higher dsDNA substrate. Adding 0.5 µg and 1.0 µg CrtB/800-Kan cassettes achieved ~100 and 200 recombinants and it was suitable for next analysis. So, 0.5 µg or 1.0 µg dsDNA cassettes were used further.

The recovery cultivation time also affected the recombineering efficiency22, therefore we varied recovery cultivation time from 0 h to 5 h. The results showed that the total number of recombinants was increased with the increasing recovery time, while the number of colonies per OD600 did not obviously increase after 1 h (Fig. 1c). In order to save time and obtain suitable colonies, 4 h recovery time was applicable. To improve the recombineering efficiency, the induction time was optimized (from 2 h to 8 h). The optimal induction time was 5 h, at which point the OD600 was ~1.0 (Fig. 1d).

To protect the linear dsDNA from being degraded by a host nuclease, the 5′ end phosphorylated dsDNA could further improve the recombineering efficiency22, 32. The dsDNA substrate of CrtB/800-Kan cassette for recombineering was phosphorylated which further improved the recombineering efficiency at least as twice as much in C.glutamicum ATCC14067 (Fig. 1e). The results showed that the colony can reach to 1.41 ± 0.18 × 103 colonies per mL (Supplementary Fig. S2) under the condition of 800 bp homology arms, 1 μg phosphorylated dsDNA, 5 h induction time and 4 h recovery time. These work established the RecET recombineering condition in C. glutamicum ATCC14067.

Scheme for markerless deletion of a gene in C. glutamicum ATCC14067

The RecET system with linear dsDNA can effectively perform targeted gene replacement by selection marker in C. glutamicum ATCC14067. However, a simple selection marker rescue strategy is necessary for metabolic engineering and synthetic biology in C. glutamicum. Therefore, we redesigned a self-excisable linear dsDNA cassette containing the site-specific Cre/loxP recombination system for recombineering via RecET (RecET-Cre/loxP system).

First, a PBS-Cre-Kan plasmid was constructed containing the Cre expression cassette and kanamycin resistance expression cassette (Fig. 2). The Cre expression cassette is under control of theophylline riboswitch E* (thoE* RBS), which can be induced with 1 mM theophylline (Fig. 2). The kanamycin resistance expression cassette was used as a screening marker. Then, the PBS-Cre-Kan plasmid was used to amplify the generic fragment of Cre-Kan cassette, which had the 34 bp sequence of lox71 added on the left and lox66 added on the right, with primers C-K-lox66 and C-K-lox71. The C. glutamicum ATCC14067 genome was used for the amplification of ~800 bp left (GeneX-L containing the lox71 sequence) and right (GeneX-R containing the lox66 sequence) homologous fragments flanking the gene to be deleted (Fig. 2). Finally, all the three fragments, the generic fragment of Cre-Kan cassette and the left and right homologous fragments, were used for subsequent fusion PCR to generate ~4350 bp linear self-excisable dsDNA cassettes. We named the dsDNA cassettes were ΔArgR-cassette, ΔCrtB-cassette, ΔNcgl1221-cassette and ΔProB-cassette, respectively (Fig. 3a and Supplementary Fig. S3). 1 µg of each self-excisable linear dsDNA cassette was transformed into C. glutamicum ATCC14067 expressing RecET proteins. The linear self-excisable dsDNA cassettes accomplished the recombination via RecET system in C. glutamicum ATCC14067 (Fig. 2). To excise the selection marker and Cre expression cassette, 1 mM theophylline was added to the recombinants to induce the expression of Cre to perform excision between lox71 and lox66, which produced the markerless deletion strain for the subsequent gene deletion (Fig. 2).

Figure 2
figure 2

Scheme for gene markerless deletion via a self-excisable cassette in C. glutamicum ATCC14067. The theoE*-RBS is theophylline riboswitch E* which can induce the expression of Cre with 1 mM theophylline. The homology-L/R represents the left/right homology arms of the targeted gene, which are approximately 800 bp. GeneX-L represents the targeted gene’s left homology arm and 34 bp lox71 sequence, and GeneX-R represents the 34 bp lox66 sequence and the targeted gene’s right homology arm. Cre-Kan cassette contains the 34 bp lox71 sequence, the Cre expression cassette, the kanamycin resistance cassette and the 34 bp lox66 sequence. ΔGeneX represents the targeted gene to be deleted. GeneX represents the targeted gene. All primers are listed in Supplementary Table S3.

Full size image
Figure 3
figure 3

Markerless deletion in C. glutamicum ATCC14067. (a) ΔArgR-cassette construction. The ΔArgR-cassette contains ArgR-L (argR homology left and the 34 bp lox71 sequence), Cre-Kan cassette and ArgR-R (34 bp lox66 sequence and argR right homology). (b) ΔArgR-cassette recombineering with RecET. The primer pairs of ΔargR-JD-S/A were used for the verification of colony PCR. WT is C. glutamicum ATCC14067. (c) Cre excision by the addition of 1 mM theophylline: at least four recombinants were used to evaluate the excision efficiency. (d) Identification markerless deletion of the argR gene. MargR represents the strain in which the argR gene is replaced by the 34 bp sequence of lox72. (e) Identification markerless deletion of the crtB gene in MargR. MargR-crtB represents the strain in which the argR and crtB genes are replaced by the sequence of 34 bp lox72. PR is the ΔargR-JD-S/A primer pair. PC is the ΔcrtB-JD-S/A primer pair. All primers are listed in Supplementary Table S3. Kmr, kanamycin resistance. Cmr, chloramphenicol resistance.

Full size image

Markerless deletion with the RecET-Cre/loxP system

1 µg each of the ΔArgR-cassette, ΔCrtB-cassette, ΔNcgl1221-cassette and ΔProB-cassette were transformed into ATCC14067-recE/T competent cells, yielding numbers of Kan-resistant colonies with different gene recombineering (Fig. 3b). At least 30 of the kanamycin resistant colonies were analyzed by PCR using the primer pairs of ΔargR-JD-S/A, ΔcrtB-JD-S/A, ΔNcgl1221-JD-S/A and ΔproB-JD-S/A, respectively (Fig. 3b). At least 94% of the transformants successfully replaced the targeted gene with the Cre-Kan cassette, and none of the correct transformants were mutated by the sequencing (Table 2). The correct recombinants were called MCargR, MCcrtB, MCNcgl1221 and MCproB. These results are in agreement with the linear dsDNA substrate recombineering in M. smegmatis and M. tuberculosis, where >90% of the drug-resistant colonies correctly replaced the targeted gene33. Then, 1 mM theophylline was added to perform marker excision (Fig. 3c). The streak results show that after excision by Cre, the single colonies of the four genes can only grow on plates without kanamycin (Fig. 3d, Supplementary Fig. S4). The PCR results showed that only a ~500 bp fragment was amplified from the single colonies (Fig. 3d, Supplementary Fig. S4). Both of these results indicate that the Cre executed the excision with an efficiency approaching 98% (Table 2), yielding the markerless deletion recombinants MargR, McrtB, MNcgl1 221 and MproB.

Table 2 RecET recombineering and Cre excision of different genes in C. glutamicum ATCC14067.
Full size table

This method achieved markerless deletion of a single gene, and the MargR (ΔargR::lox72) strain was used for the second gene deletion and marker rescue. 1 µg each of the ΔCrtB-cassette, ΔNcgl1221-cassette and ΔProB-cassette were recombined in MargR to generate the new recombinant strains MargR-MCcrtB, MargR-MCNcgl1221 and MargR-MCproB. The PCR results showed that ≥97% of the transformants were correct (Table 2), and the sequencing results showed that none of the transformants were mutated, except for one transformation site of MargR-MCproB (Table 2). The screening and PCR results showed that the second genes of crtB, Ncgl1221, and proB underwent markerless deletion after Cre expression was induced. The double-deletion markerless recombinants MargR-crtB, MargR-MNcgl1221 and MargR-MproB were obtained (Fig. 3e and Supplementary Fig. S5) and Cre excision efficiency exceeded 95% (Table 2). These results are similar to the single gene markerless deletion.

Discussion

Recombineering with exonuclease-recombinase pairs via linear dsDNA is a powerful method for gene modification. In this study, we employed the RecET recombineering system for dsDNA recombination in C. glutamicum ATCC14067. Besides, we developed a system containing a linear self-excisable dsDNA cassette and RecET system (RecET-Cre/loxP system). It is the first effective and simple strategy for gene markerless deletion in C. glutamicum ATCC14067.

Although recombination activity of RecET was lower than that in E. coli 34, the recombineering activity can reach up to 1.41 ± 0.18 × 103 colonies per mL in C. glutamicum ATCC14067 under optimized recombination conditions (Supplementary Fig. S2). There was no recombinant without the help of RecET, which indicated that the linear dsDNA could’t complete recombination with its own recombination system in C. glutamicum ATCC14067, and the RecET from the E. coli phage effectively improved the recombination efficiency. It also suggests that the RecET from the E. coli phage can across a species barrier to be functional29 in C. glutamicum ATCC14067.

We observed that the recombineering efficiency of the longer 4.4 kb self-excisible dsDNA of ΔCrtB-cassette was significantly lower than the 2.8 kb CrtB/800-Kan dsDNA cassette (Supplementary Fig. S6). Both of them contained the same 800 bp homology arms, but the insertion lengths of the region between the homologous flanking of the two cassettes were 2755 and 1235 bp, respectively (crtB gene is 918 bp). This result is consistent with the observation in the RecTEPsy recombineering in Pseudomonas syringae 31 and the insertion length between the homologous flanking sequences also strongly affects the RecET recombineering efficiency in C. glutamicum ATCC14067.

With the RecET-Cre/loxP system, the linear dsDNA cassette for recombineering can be generated via two-steps PCR and only requires one round of recombination to obtain the mutated recombinants, which it only takes 2–3 days and the gene replacement frequency can reach up to ≥94%. The subsequent markerless deletion only requires an extra 2–3 days. It is more effective and simpler for recombination in C. glutamicum ATCC14067 than the traditional counter-selectable system16. This recombineering system also provides a new strategy to perform gene manipulation for Corynebacteria and builds a solid basis for producer construction.

Materials and Methods

Strains and growth condition

C. glutamicum ATCC14067 was grown in BHI liquid medium (37 g L−1 brain heart infusion (Becton, Dickinson and company)) at 30 °C, 250 rpm. E. coli DH5α was used as the cloning host for plasmid manipulation, and it was cultured in Luria-Bertani medium (10 g L−1 peptone, 5 g L−1 yeast extract, 10 g L−1 NaCl) at 37 °C. If necessary, antibiotics were added in the following concentrations: kanamycin 50 mg L−1 (50 Km) or chloramphenicol 15 mg L−1 (15 Cm) for E. coli and Kanamycin 25 mg L−1 (25 Km) and 7.5 chloramphenicol mg L−1(7.5 Cm) for C. glutamicum ATCC14067. All bacterial strains used in this study are listed in Supplementary Table S2.

Plasmids and linear dsDNA cassette

The genes of the exonuclease-recombinase pairs are assembled into the E. coli-C. glutamicum shuttle-inducible vector pEC-XC99E, which can be induced by 1 mM isopropyl-β-d-thiogalactoside (IPTG), to construct the expression plasmids pEC-exo/bet, pEC-orf47/orf48, pEC-orfB/orfC and pEC-recE/recT. PBS-crtB-L/R-Kan plasmid was constructed for the amplification of different homology lengths of linear CrtB-Kan cassettes (Supplementary Fig. S1). The 5′ end phosphorylation of dsDNA was performed by T4 polynucleotide kinase (Thermo Scientific, Waltham, MA).

The theophylline riboswitch E*(theoE*-RBS)35, 36 sequence was amplified using the primer pair, theoE-S and theoE-A, from the synthetic theoE*-RBS sequence. Then the cre gene and the theoE*-RBS sequence are assembled into pEC-XC99E to construct the plasmid pEC-theoE-cre. For construction of PBS-Cre-Kan, firstly, the DNA fragment of Cre cassette was amplified from plasmid pEC-theoE-cre with primers theoE-Cre-S and theoE-Cre-A. Secondly, the plasmid pK18mobSacB was used for the amplification of the Kan cassette with primers kan-S and kan-A. Finally, the Cre cassette and the Kan cassette are assembled into pBluescript II SK(+) to construct the generic plasmid PBS-Cre-Kan (Supplementary Fig. S7). NEB Builder HiFi DNA Assembly Master Mix (New England BioLabs, Boston, MA) is used to assembly all the plasmids, and the PCR products are generated by KOD DNA polymerase (TOYOBO, Japan). Successful PCR products were confirmed by agarose gel electrophoresis. Plasmid construction and DNA fragments were confirmed by DNA sequencing (Sangon Biotech, Shanghai, China). All plasmids used in this study are listed in Supplementary Table S2. All the primers used in this study are listed in Supplementary Table S3.

Preparation of the competent cells

A fresh single colony of wild-type C. glutamicum ATCC14067, or C. glutamicum ATCC14067 carrying pEC-XC99E derivatives encoding recombinases, was inoculated from a BHI plate into fresh BHI liquid medium containing the relevant antibiotic and grown overnight at 30 °C. Each culture was then inoculated into Epo medium (5 g L−1 yeast, 10 g L−1 tryptone, 10 g L−1 NaCl, 4 g L−1 isoniazide, 25 g L−1 glycine, 1 g L−1 Tween80) containing the relevant antibiotic with or without 1 mM IPTG at an initial OD600 of 0.3. When the cultures reached an OD600 of ~0.9, the cells were collected to make competent cells. The cells were cooled on ice for 15 min and then harvested at 4,000 rpm at 4 °C for 10 min. Following washed twice in 30 ml of ice-cold 10% (v/v) glycerol, cells were suspended in 1/70 (v/v) 10% glycerol.

Recombineering assay

80 μl of fresh unfrozen electrocompetent cells carrying the exonuclease-recombinase pairs or pEC-XC99E mixed with linear dsDNA cassette and incubated on ice for 5 min. Then the cell-plasmid DNA/dsDNA mixture was transferred to an ice-cold electroporation cuvette (0.1-cm electrode gap). Electroporation was performed with a Bio-Rad Micropulser set by three times 1.8 KV/cm (Ec1) pulse. The cell-plasmid DNA/dsDNA mixture was inoculated into 6 ml of the resuscitation medium LBHIS (2.5 g L−1 yeast, 5 g L−1 tryptone, 5 g L−1 NaCl, 18.5 g L−1 Brain Heart Infusion, 91 g L−1 sorbitol) containing the relevant antibiotics. Then cells were placed in a hot bath for 6 min at 46 °C and incubated for hours at 30 °C, 250 rpm. LBHIS-Km25-Cm7.5 solid medium was used for recombinant selection and the cells were cultured at 30 °C for 36~48 h for colony-forming unit (cfu) determination. For exonuclease-recombinase pairs’ recombination activity analysis, the empty plasmid pEC-XC99E with linear dsDNA was used as a positive control. The exonuclease-recombinase pairs’ expression plasmids without the addition of linear dsDNA were used as a negative control. The replicative plasmid pZ9, which contains kanamycin resistant, was used to explore the competence for uptake efficiency of DNA.

Cre excision assay

The Cre expression cassette is under the control of thoE*-RBS, which can be induced by the addition of 1mM theophylline. To allow Cre-mediated intermolecular excision and accomplish the deletion of kanamycin resistance selection marker, the correct transformants were inoculated into BHI medium containing 1 mM theophylline and cultured for 24 h, at 30 °C, 250 rpm. The cells were then streaked onto the BHI-Cm7.5 plate with 1 mM theophylline and incubated at 30 °C for 16 h. Then, the single colonies were streaked into the BHI-Cm7.5-Km25 and BHI-Cm7.5 plate to explore the excision efficiency. To further identification the makerless excision, PCR identification and sequencing were used for the single colonies which can only grow on plates without kanamycin (BHI-Cm7.5).