Protocols for RecET‐based markerless gene knockout and integration to express heterologous biosynthetic gene clusters in Pseudomonas putida

Summary Pseudomonas putida has emerged as a promising host for the production of chemicals and materials thanks to its metabolic versatility and cellular robustness. In particular, P. putida KT2440 has been officially classified as a generally recognized as safe (GRAS) strain, which makes it suitable for the production of compounds that humans directly consume, including secondary metabolites of high importance. Although various tools and strategies have been developed to facilitate metabolic engineering of P. putida, modification of large genes/clusters essential for heterologous expression of natural products with large biosynthetic gene clusters (BGCs) has not been straightforward. Recently, we reported a RecET‐based markerless recombineering system for engineering P. putida and demonstrated deletion of multiple regions as large as 101.7 kb throughout the chromosome by single rounds of recombineering. In addition, development of a donor plasmid system allowed successful markerless integration of heterologous BGCs to P. putida chromosome using the recombineering system with examples of – but not limited to – integrating multiple heterologous BGCs as large as 7.4 kb to the chromosome of P. putida KT2440. In response to the increasing interest in our markerless recombineering system, here we provide detailed protocols for markerless gene knockout and integration for the genome engineering of P. putida and related species of high industrial importance.


Introduction
Cellular robustness and versatile metabolism of Pseudomonas putida have characterized this Gram-negative soil bacterium as an attractive workhorse of metabolic engineering for the bio-based production of chemicals and materials (Nikel et al., 2016;Nikel and de Lorenzo, 2018). Particularly, P. putida KT2440 has been officially classified as a generally recognized as safe (GRAS) strain (Federal Register, 1982) and exploited for the production of chemicals and products that human directly consumes, especially natural products including pharmaceuticals, nutraceuticals, and cosmetic ingredients (Loeschcke and Thies, 2015). Various genetic engineering elementsincluding counterselection markers (Galvao and de Lorenzo, 2005;Gross et al., 2006;Graf and Altenbuchner, 2011;Johnson et al., 2017), site-specific recombinases (Leprince et al., 2012;Ibrahim et al., 2015), homing endonuclease I-SceI (Martinez-Garcia and de Lorenzo, 2011;Martinez-Garcia et al., 2014;Chen et al., 2016), bacteriophage-derived recombinases homologous to Bet protein from k Red system (Aparicio et al., 2016;Luo et al., 2016), and clustered regularly interspaced palindromic repeat (CRISPR)/CRISPR-associated (Cas) systems (Aparicio et al., 2017;Cook et al., 2018;Sun et al., 2018) have been incorporated to facilitate gene knockout in P. putida. Although reduction of the genome size can be beneficial in some cases for better strain performance (Lieder et al., 2015), only few cases of deleting large genomic fragment have been reported (Martinez-Garcia et al., 2014;Aparicio et al., 2017). In addition, integration of heterologous biosynthetic gene clusters (BGCs) to P. putida chromosome for the production of heterologous secondary metabolites has relied on time-consuming homologous recombination with selection markers (Wenzel et al., 2005;Gross et al., 2006;Cao et al., 2012;Gong et al., 2016) and unpredictable transposonmediated random insertion (Glandorf et al., 2001;Chai et al., 2012;Loeschcke et al., 2013;Domrose et al., 2015Domrose et al., , 2017, requiring the development of rapid and reliable integration system for programmed introduction of heterologous BGCs to P. putida chromosome. Recently, we reported a RecET-based markerless recombineering system for deletion and integration of large-sized genes and clusters in P. putida (Choi et al., 2018). Although P. putida has chromosomes with high GC content (e.g. 61.6% in average for the chromosome of P. putida KT2440), conventional recombineering systems for P. putida chromosome modification use Bet recombinase derived from k Red system (Chen et al., 2016;Cook et al., 2018;Sun et al., 2018), which is efficient in invading double-stranded DNAs (dsDNAs) with low GC content (< 20%) (Rybalchenko et al., 2004), or a Bet recombinase homolog Ssr from P. putida DOT-T1E (Aparicio et al., 2016). Thus, we exploited Escherichia coli Rac prophage-derived RecET system (Choi et al., 2018) of which dsDNA invasion activity is independent of GC content of target loci (Noirot and Kolodner, 1998) considering the high GC content of P. putida. In addition, inducible expression of RecET proteinsof which leaky expression might reduce genome stabilityfrom a RecET vector pJB658-recET (Table S1) was tightly regulated by the P m /XylS system that requires m-toluic acid for the induction (Choi et al., 2018). Once the expression of RecET proteins was induced, introduction of linear donor dsDNAs encoding antibiotic resistance markers [i.e. the tetA(C) and aph(3ʹ)-IIa genes for tetracycline and kanamycin resistance, respectively] flanked by a pair of 100-bp homologies to target chromosomal loci ( Fig. 1A) was enough to integrate the antibiotic resistance markers to the target chromosomal loci. Moreover, a pair of mutant loxP sites (i.e. lox71 and lox66) placed at both ends of the antibiotic resistance marker [e.g. lox71-tetA(C)-lox66 cassette] allowed simple excision of the antibiotic resistance marker integrated to the chromosome upon the induction of Cre protein from the Cre vector pRK2Cre (Table S1) using isopropyl b-D-1-thiogalactopyranoside (IPTG). A mutant loxP site (i.e. lox72) left behind the excision of the antibiotic resistance marker is extremely inefficient in the Cre-induced recombination with lox71 or lox66 to be used for subsequent recombineering cycles (Lambert et al., 2007), allowing repeated use of the system for sequential modification of multiple target sites without significant interaction between engineered loci. Furthermore, instability of plasmid pJB658-recET (Table S1) enabled its rapid curing upon simple streaking of recombinant strains harbouring the RecET vector (mediating ampicillin resistance) on LB-agar medium without ampicillin (Choi et al., 2018). Similarly, temperature sensitivity of the plasmid pRK2Cre (Table S1) allowed its curing upon streaking recombinant strains harbouring the Cre vector (with kanamycin resistance) on LB-agar medium without kanamycin followed by incubation at 37°C. As a result, recombineered strains could be easily converted to markerless and plasmid-less strains (Choi et al., 2018), which is ideal for subsequent engineering and also prevents dissemination of antibiotic resistance genes to the environment (Jiang et al., 2017) upon accidental release of industrial strains. Using the RecET recombineering system, we demonstrated markerless deletion of chromosomal loci of P. putida KT2440 as large as 101.7 kb (i.e. 1.6% of P. putida KT2440 chromosome) in a single round of recombineering (Table S2) (Choi et al., 2018).
To establish a markerless recombineering system for the integration of heterologous BGCs to P. putida chromosome, we also devised a donor plasmid system (Choi et al., 2018). Most BGCs of heterologous natural products are longer than 6 kb (Loeschcke and Thies, 2015). However, the efficiency of gene integration via recombineering with linear donor dsDNA dramatically decreases as the insert size increases above 2-3.5 kb (Kuhlman and Cox, 2010), which includes the antibiotic resistance genes for selection [e.g. 1.7 kb lox71-tetA(C)-lox66 cassette]. In addition, increased length of linear donor dsDNA reduces the amplification efficiency while elevating the mutation rate during PCR. Moreover, transformation efficiency decreases together as the size of DNA increases (Hanahan, 1983;Ohse et al., 1995;Sheng et al., 1995). Thus, a donor plasmid system that does not replicate in P. putida was developed to overcome such issues (Choi et al., 2018). Donor DNAs in forms of plasmids can be maintained stably with extremely low mutation rates in cloning hosts and easily isolated in large quantities from them. Supercoiling of the isolated donor plasmid also enhances transformation efficiency during recombineering. The recombineering efficiency can be further enhanced by increasing the length of homology arms (HAs) in the donor plasmid. Furthermore, the sacB gene was placed next to the origin of replication (e.g. pUC ori not replicating in P. putida) in the donor plasmid to combine two-step double crossing-over procedure into a single step (Choi et al., 2018). Briefly, integration of entire donor plasmid to the target chromosomal locus via a single cross-over event on either HAs results in the expression of levansucrase from the sacB gene and kills the cell in the presence of sucrose and no NaCl. Thus, only recombinant cells that experienced double cross-over on both HAs, harbour the genes/clusters to be integrated and the lox71-tetA(C)-  (Table S3) are presented next to the primers with some fixed sequences, if available. bla, ampicillin resistance gene; tetA(C), tetracycline resistance gene; HA_L & HA_R, HAs placed left and right to the tetA(C) gene; MCS, multiple cloning site. pUC, pUC origin of replication; p15A, p15A origin of replication. The primer sequences are shown again in Table S3 for better visibility. lox66 cassette, and lost the sacB gene and ori survive, while unmodified cells are selected out due to the supplementation of appropriate antibiotic [e.g. tetracycline if the donor plasmid harbours the lox71-tetA(C)-lox66 cassette]. Knocking out 1-kb region on the pvdD gene using 1 lg of a donor plasmid pTetSac-DpvdD (Table S1) resulted in tens of times more recombinant strains with the pvdD gene knockout compared to when using 1 lg of linear donor dsDNA DpvdD1k100::tetA-lox, demonstrating the capacity of the donor plasmid system in improving the recombineering efficiency (Choi et al., 2018). Cloning of 7.4 kb heterologous violacein BGC to an adaptor plasmid pTetSac-DpvdD::Adaptor (Table S1) to construct a donor plasmid pTetSac-DpvdD::Violacein (Table S1) and subsequent markerless recombineering of P. putida KT2440 harbouring the RecET vector pJB658-recET (Table S1) using the donor plasmid resulted in markerless integration of the 7.4 kb heterologous BGC to the target locus (i.e. the pvdD gene) on P. putida KT2440 chromosome (Table S2) with successful biosynthesis of violacein (Choi et al., 2018).
In response to the request to provide more detailed protocols for our P. putida recombineering system, we report step-by-step procedures for the construction of linear donor dsDNAs and donor plasmids as well as recombineering P. putida with the RecET-based markerless recombineering system using the donor DNAs constructed. In addition, examples of engineering P. putida with the recombineering protocol from our previous report are briefly summarized.

Selection of recombineering targets for gene knockout and integration (1-3 h)
Chromosomal loci and episomal loci on single-copy megaplasmids (e.g. genes on a TOL plasmid pWW0 of P. putida mt-2 strain) to be deleted or used as platform sites for gene integration should be selected before proceeding to recombineering. In addition, DNA sequence of HAs and primers for HA construction should be prepared at this step. Two different procedures are used depending on the type of donor DNAs used for recombineering: linear donor dsDNAs for gene knockout (option A) and donor plasmids for gene knockout or integration (option B). For knocking out target genes, use of linear donor dsDNA accelerates recombineering procedure as its preparation steps are simpler while use of donor plasmid guarantees efficient and successful gene knockout.
Designing linear donor dsDNAs for gene knockout. i For each gene/region to be knocked out, define actual region to be deleted: either partial (e.g. in-frame or out-of-frame deletion for truncation) or complete deletion of the target gene (Fig. 1A). ii Retrieve DNA sequences for a pair of 100-bp regions neighbouring each deletion target. These sequences will be used to design HAs (Fig. 1A). It should be noted that HA of 50 bp is insufficient to engineer target sites on the P. putida chromosome. iii Design primer sets to amplify the lox71-tetA(C)-lox66 cassette that is flanked by the 100-bp HA pair (designed in step ii) from plasmid pTetAmulox (Tables S1, S3 and Fig. 1A). For example, two pairs of~70-nt primers can be designed as follows: Each primer in the first pair (primers LinearDonor+ and LinearDonorÀ) consists of a~20-nt segment complementarily binding to each end of the lox71-tetA (C)-lox66 cassette and a~50-nt segment from each inner half of the 100-bp HA (Table S3 and Fig. 1A). Each primer in the second pair (primers Lin-earDonor++ and LinearDonorÀÀ) consists of a~20nt segment from the 5ʹ ends of each primer in the first pair and a~50-nt segment spanning the rest half of each 100-bp HA (Table S3 and Fig. 1A).
Melting temperature (T m ) of each~20-nt segment for primer binding is recommended to be above 60°C to facilitate amplification by PCR. iv Design primers for confirming gene knockouts. For example, a pair of primers binding outside the HA with their 3ʹ ends oriented towards the deletion target (primers Check+ and CheckÀ) are good choices (Table S3). In addition, a pair of primers binding inside the tetA(C) gene with their 3ʹ ends oriented towards each near end of the gene [e.g. primers tetA(C)+ and tetA(C)À] are also useful (Table S3). v Order the synthesis of the designed primers.
Designing donor plasmids for gene knockout and adaptor plasmids for gene integration. i Define actual regions to be deleted for gene knockout or integration. ii Retrieve a pair of~1-kb DNA sequences flanking each deletion target, which will be used to design HA (Fig. 1B). iii Design two sets of primers (primer sets HA_L+/ HA_LÀ and HA_R+/HA_RÀ) that will be used to amplify the pair of~1-kb HA from the genomic DNA (gDNA) of P. putida (Table S3 and Fig. 1B). In addition, design two additional pairs of primers to amplify the (MCSÀ)lox71-tetA(C)-lox66 (primer pair lox71+/ lox66À or MCS+/lox66À) and the sacB-ori (primers sacB+ and oriÀ) cassettes respectively (Table S3 and Fig. 1B). Use plasmid pTetSac-DpvdD (Table S1) as a template to amplify the lox71-tetA (C)-lox66 cassette and plasmid pTetSac-DpvdD:: Adaptor (Table S1) as a template to amplify the MCS-lox71-tetA(C)-lox66 cassette (Fig. 1B). The primers may have additional sequences at the 5ʹ ends that provide homologies for Gibson assembly or restriction recognition sites for cloning by restriction enzyme digestion and subsequent ligation (Table S3 and Fig. 1B). T m of each primer binding sequence is recommended to be above~60°C to facilitate amplification by PCR as gDNA of P. putida has high GC content (e.g. 61.6% in average for P. putida KT2440). iv Design primers for confirming gene knockout and integration. For example, a pair of primers binding outside the HA with their 3ʹ ends oriented towards the deletion target are good choices (primers Check+ and CheckÀ). In addition, a pair of primers binding inside the tetA(C) gene with their 3ʹ ends oriented towards each near end of the gene [e.g. primers tetA(C)+ and tetA(C)À] are also useful. v Order the synthesis of the designed primers.

Preparation of linear donor DNA for gene knockout (5-6 h)
i Amplify the lox71-tetA(C)-lox66 cassette from plasmid pTetAmulox (Table S1) using primers Lin-earDonor+ and LinearDonorÀ (Table S3). The amplified product contains~50-nt HA at each end (Fig. 1A). ii Purify the amplified product after resolving by gel electrophoresis followed by gel extraction. iii Amplify the purified product using primers Lin-earDonor++ and LinearDonorÀÀ (Table S3), extending each end with the rest of the 100-nt HA. This results in the lox71-tetA(C)-lox66 cassette flanked by a pair of 100-nt HAs (Fig. 1A). Amplification in large quantity is recommended as more than 10 lg of the product is desired. iv Purify the amplified product (i.e. linear donor dsDNA) and elute using deionized water. Elution at high concentration (> 300 ng ll À1 ) is preferred as > 3 lg of linear donor DNA is used for each recombineering procedure.
Construction and preparation of donor plasmids for gene knockout and integration (2-3 and 4-6 days to prepare donor plasmids for gene knockout and integration, respectively) i Amplify each components of the donor/adaptor plasmid by PCR (Fig. 1B). For example, amplify the HA pair from P. putida gDNA using primer sets HA_L+/ HA_LÀ and HA_R+/HA_RÀ (Table S3 and Fig. 1B) and the MCS-lox71-tetA(C)-lox66 cassette from plasmid pTetSac-DpvdD::Adaptor (Table S1) using primers MCS+ and lox66À (Table S3 and Fig. 1B). For gene knockout, however, amplification (and subsequent use) of the lox71-tetA(C)-lox66 cassette, rather than the MCS-lox71-tetA(C)-lox66 cassette, from plasmid pTetAmulox (Table S1) using primers lox71+ and lox66À (Table S3) is recommended as repeated use of the MCS-lox71-tetA(C)-lox66 cassette unnecessarily introduces multiple MCS sites to the chromosome. In addition, the sacBori cassette harbouring an origin of replication with low copy number [e.g. sacB-ori (p15A) cassette from plasmid pTetSac15-DpvdD2::Adaptor], rather than the sacB-ori (pUC) cassette from plasmid pTetSac-DpvdD::Adaptor (Table S1), may be amplified using primers sacB+ and oriÀ (Table S3 and Fig. 1B). This is a useful strategy when constructing an adaptor plasmid for cloning genes/BGCs of which cloning in high copy number plasmids and subsequent (leaky) expression burdens the cloning host and interrupts cloning/maintenance of the resulting donor plasmids. ii Assemble the four amplified products (i.e. two HAs, (MCS-)lox71-tetA(C)-lox66 cassette, and sacB-ori cassette) to construct the donor plasmid for gene knockout or adaptor plasmid for the cloning genes of interest (Fig. 1B). For example, the two HAs and the (MCS-)lox71-tetA(C)-lox66 cassette can be assembled by overlapping PCR using primers HA_L+ and HA_RÀ (Table S3). The amplified product and the sacB-ori cassette can be subsequently joined by Gibson assembly (Gibson et al., 2009). Alternatively, restriction enzyme digestion followed by ligation may be used if the primers used for the amplification contain appropriate restriction recognition sites. Of course, certain variations can be made to facilitate the cloning procedure depending on their own lab protocols. Transform a cloning host (e.g. E. coli DH5a) and screen cloned colonies on LBagar medium supplemented with 10 lg ml À1 tetracycline at 37°C. iii Identify a construct (i.e. donor plasmid for gene knockout or adaptor plasmid for gene integration) with correct sequence, especially sequences for the MCS, lox71, and lox66. Store cell stock with the correct construct for further use. To prepare donor plasmids for gene knockout, directly go to step vi. iv Inoculate cells harbouring a correctly constructed adaptor plasmid into 5 ml of LB medium supplemented with 10 lg ml À1 tetracycline, incubate at 37°C with rotary shaking at 200 rpm and isolate the adaptor plasmid to clone genes of interest to be integrated to P. putida chromosome.
v Amplify the gene(s)/cluster(s) of interest using primers possessing appropriate restriction recognition sites at their 5ʹ ends. Digest the amplified product(s) and the adaptor plasmid isolated in step iv using appropriate restriction enzymes. Ligate the digested DNAs together to construct donor plasmid for gene integration. Alternatively, the primers for target gene amplification may possess homologies to the cloning site for direct assembly of the amplified product(s) and the digested adaptor plasmid by Gibson assembly. Any other variations can be used to facilitate the cloning procedure. Transform a cloning host with the assembly product (e.g. E. coli DH5a), screen colonies harbouring correct construct (i.e. donor plasmid for gene integration) on LB-agar medium supplemented with 10 lg ml À1 tetracycline at 37°C and identify colonies harbouring constructs with correct sequences. Store cell stocks for further use. vi To isolate donor plasmids for recombineering, inoculate cells harbouring the donor plasmids into 5-10 ml of LB medium supplemented with 10 lg ml À1 tetracycline, incubate at 37°C with rotary shaking at 200 rpm and isolate the donor plasmid using deionized water for the elution. Preparation of more than 3 lg of plasmids at concentrations higher than> 100 ng ll À1 is preferred.
Recombineering of P. putida for gene knockout and integration (shortest 4N + 1 day for N cycles) i Transform P. putida strain that will be applied for gene knockout or integration with the RecET vector pJB658-recET (Table S1 and Fig. 2A). Briefly, incubate P. putida strain in appropriate medium, such as 2-5 ml of LB medium with NaCl concentration reduced to 5 g l À1 [LB (5 g l À1 NaCl) medium], and conditions (e.g. 30°C, 200 rpm, with antibiotics, if required) to full density. This usually takes several hours. Harvest P. putida cells from 1 ml of the culture and wash the cells twice by resuspension to 1 ml of 300 mM sucrose solution and subsequent centrifugation at 4000 g for 2 min. Resuspend the pellet to 100 ll of 300 mM sucrose solution and add 0.1-1 lg of plasmid pJB658-recET (Table S1 and Fig. 2B). Electroporate the cells in 1-mm cuvette with electric pulse of 1.8 kV with capacitance and resistance of 25 lF and 200 Ω, respectievly. Recover the transformants in LB medium at 30°C and plate/streak the cells on LB-agar medium supplemented with 500 lg ml À1 of ampicillin. Alternatively, P. putida strains already harbouring plasmid pJB658-recET may be streaked on LB-agar medium supplemented with 500 lg ml À1 of ampicillin to isolate fresh colonies. Incubate the plates at 30°C.
ii Inoculate a single colony into 5 ml LB (5 g l À1 NaCl) medium supplemented with 500 lg ml À1 of ampicillin in 50 ml conical tube and incubate at 30°C for several hours with rotary shaking at 200 rpm. The screw cap of the conical tube is recommended to be slightly open and held with a small patch of adhesive tape for aeration. Meanwhile, prepare 50 ml of LB (5 g l À1 NaCl) medium supplemented with 500 lg ml À1 of ampicillin and 5 mM m-toluic acid in 250 ml Erlenmeyer flask pre-warmed at 30°C with rotary shaking at 200 rpm. m-Toluic acid stock solution dissolved in ethanol at 1 M concentration may be used for the supplementation. Dissolution of m-toluic acid to the medium can be facilitated by static incubation of the medium supplemented with m-toluic acid at 42°C with brief and intermittent manual shaking. Pre-warming the medium above 30°C can also facilitate the dissolution of m-toluic acid upon the supplementation. iii Transfer 1 ml of the culture with full growthoptical density measured at 600 nm (OD 600 ) should reach 5to the pre-warmed medium and incubate at 30°C with rotary shaking at 200 rpm until OD 600 reaches~2 to induce RecET protein expression ( Fig. 2A). This step usually takes~5 h, but the incubation time may depend on P. putida strains and physiological status of the overnight culture. iv Ice-chill the culture for 5 min (on ice with gaps filled with water) and harvest cells from 49 ml of the chilled culture by centrifugation at 2090 g for 10 min. v Wash the pellet from each 49 ml culture twice by resuspension to 2 ml of ice-cold 300 mM sucrose solution and subsequent centrifugation in two 1.5 ml microcentrifuge tubes at 4000 g for 2 min at 4°C. Centrifuge once more after two rounds of washing to completely remove the supernatant. vi Resuspend the washed pellet to 200 ll of ice-cold 300 mM sucrose solution for each OD 600 unit of the cultured cells (before the harvest). For example, resuspend the washed cell pellet from 49 ml of the culture with OD 600 = 2.00 to 400 ll (=2.00 9 200 ll) of ice-cold 300 mM sucrose solution. vii Aliquot 70 ll of the resuspended cells to 1.5-ml microcentrifuge tubes kept on ice. viii Add 3 lg or more linear donor dsDNA or 1 lg or more donor plasmid to each fresh aliquot ( Fig. 2A), mix by tapping and incubate on ice for 5 min. ix Electroporate the competent cell donor DNA mixture in 1-mm cuvette by applying electric pulse of 1.8 kV with capacitance and resistance of 25 lF and 200 Ω, respectively. Immediately resuspend the transformants to 900 ll of LB (5 g l À1 NaCl) and recover at 30°C for 2 h without shaking.
x For gene knockout using linear donor dsDNAs, plate the recovered cells on LB-agar plates supplemented with 50 lg ml À1 tetracycline ( Fig. 2A). For gene knockout/integration using donor plasmids, plate the recovered cells on LB-agar medium with 0 g l À1 NaCl, 100 g l À1 sucrose, and 50 lg ml À1 tetracycline ( Fig. 2A). To maintain the RecET vector for subsequent iterative recombineering, 500 lg ml À1 ampicillin may be supplemented together. Incubate statically at 30°C until colonies grow on the plate. Recombineered colonies usually grow within overnight but may grow slower depending on the engineering targets and engineered status of the cells. xi Examine the gene knockout/integration by colony PCR using primer sets appropriate for the confirmation [e.g. primers Check+ and CheckÀ or combinations with primers tetA(C)+ and tetA(C)À; Table S3] while making a master plate of the colonies for subsequent use (Fig. 2A). The sequence of the integrated genes may be optionally examined through amplification of the genes by colony PCR followed by sequencing of the amplified products. xii Inoculate a colony with desired modification into 2-5 ml of LB (5 g l À1 NaCl) medium supplemented with 50 lg ml À1 tetracycline and incubate at 30°C for several hours with rotary shaking at 200 rpm. Ampicillin (500 lg ml À1 ) may be added for stable maintenance of the RecET vector for additional rounds of recombineering. xiii Harvest the cells, prepare competent cells and transform with plasmid pRK2Cre (Table S1 and Fig. 2C) following the description in step i ( Fig. 2A). Plate the transformants on LB-agar plate supplemented with 50 lg ml À1 of kanamycin and 1 mM IPTG to induce the expression of Cre protein and excise out the tetA(C) gene via site-specific recombination between the lox71 and lox66 sites ( Fig. 2A). Thus, tetracycline should not be added to the LB-agar medium. Ampicillin (500 lg ml À1 ) may be supplemented together to maintain the RecET vector. Incubate the plates at 30°C until kanamycin-resistant colonies grow. This usually takes around one full day. xiv (Optional) To check the excision of the tetA(C) gene based on tetracycline resistance/sensitivity, streak the kanamycin-resistant colonies serially on LB-agar medium supplemented with 50 lg ml À1 tetracycline and LB-agar medium supplemented with 50 lg ml À1 kanamycin. Ampicillin (500 lg ml À1 ) may be added together to maintain plasmid pJB658-recET. Incubate the plates at 30°C until cell growth is observed from the kanamycin plate. This step usually takes several hours. This step may be skipped as the efficiency of the tetA(C) gene excision upon Cre protein induction reaches 100% (Choi et al., 2018). xv Cells growing on the kanamycin plate while not growing on the tetracycline plate are ones without the tetA(C) gene upon the introduction of the Cre vector pRK2Cre. Colony PCR may be optionally conducted to further confirm the tetA(C) gene excision. To repeat additional recombineering cycles, move on to the next step. To finish P. putida recombineering, go to step xix. xvi Streak colonies that lost the tetA(C) gene on LBagar medium supplemented with 500 lg ml À1 ampicillin and incubate at 37°C to isolate colonies cured of the Cre vector pRK2Cre ( Fig. 2A). Several hours of incubation are enough to obtain single colonies. xvii Check plasmid pRK2Cre curing by serially streaking the colonies on two LB-agar media each supplemented with 50 lg ml À1 kanamycin and 500 lg ml À1 ampicillin, respectively ( Fig. 2A). Incubate the plates at 30°C until cell growth is observed from the ampicillin plate. This usually takes several hours. A. Chromosomal regions of P. putida KT2440 examined for markerless deletion using the RecET recombineering system. Red, blue, and purple marks indicate regions knocked out using linear donor dsDNAs, donor plasmid, and both of them respectively. This figure includes the results of the previously reported (Choi et al., 2018) and updated genome engineering. B. Recombinant P. putida strains DpvdD::EGFP, DpvdD::Flaviolin, and DpvdD::Violacein harbouring heterologous biosynthetic genes/clusters of EGFP, flaviolin, and violacein on the chromosome (Table S2) show green, brown, and purple colours of each product respectively. These results have already been reported in our previous paper (Choi et al., 2018), but we took a new picture of engineered strains for better understanding of the protocols in this paper. xviii Cells growing on the ampicillin plate while not growing on the kanamycin plate are ones cured of plasmid pRK2Cre. Go to step ii for additional rounds of recombineering ( Fig. 2A). xix Streak colonies from step xv on LB-agar medium without antibiotic and incubate at 37°C to isolate colonies cured of both the RecET vector pJB658-recET and the Cre vector pRK2Cre ( Fig. 2A). Several hours of incubation is enough to obtain single colonies. xx Check the curing of plasmid pRK2Cre by serially streaking the colonies on three LB-agar media each supplemented with 500 lg ml À1 ampicillin, 50 lg ml À1 kanamycin, and no antibiotic ( Fig. 2A). Incubate the plates at 30°C until cell growth is observed from the plate without antibiotic. This usually takes several hours. xxi Cells not growing on the kanamycin plate are ones cured of the Cre vector. Store a cell stock of the recombineered strain and finish P. putida recombineering for gene knockout/integration ( Fig. 2A).
To examine the effect of the length of HAs in donor plasmids on recombineering efficiency, the pvdD gene knockout experiments were conducted using a donor plasmid pTetSac-DpvdD::Adaptor with a pair of 1-kb HAs (Table S1) and a series of pTetSac-DpvdD::Adaptor variants with pairs of 0.1-, 0.2-, 0.4-, 0.6-, and 0.8-kb HAs (Table S1). The number of tetracycline-resistant colonies appeared after each recombineering experiment drastically deceased as the length of HAs decreased from 0.6 kb to 0.4 kb (Fig. 4), while successful knockout of the pvdD gene was verified in all eight colonies randomly selected from each recombineering experimentor entire colonies if less than eight colonies grew. These results indicate that the use of donor plasmids with HA pairs longer than 0.6 kb is crucial to guarantee efficient modification of target genes on the P. putida chromosome. Furthermore, heterologous biosynthetic genes/ clusters of a fluorescent protein EGFP (1.2 kb), a polyketide flaviolin (1.2 kb), a terpenoid lycopene (3.5 kb), and an amino acid-derived secondary metabolite violacein (7.4 kb) were efficiently integrated to the pvdD gene locus of P. putida KT2440 as directed by the HAs of corresponding donor plasmids pTetSac-DpvdD:: EGFP, pTetSac-DpvdD::Flaviolin, pTetSac-DpvdD::Lycopene, and pTetSac-DpvdD::Violacein (Tables S1 and  S2) (Choi et al., 2018). Among four resulting recombinant strains, DpvdD::EGFP, DpvdD::Flaviolin and DpvdD::Violacein strains (Table S2) showed green, brown, and purple colours of EGFP, flaviolin, and violacein, respectively (Fig. 3B), supporting successful biosynthesis of each product.

Discussion
Pseudomonas putida has been considered as a promising host strain for the production of diverse secondary metabolites of importance (Loeschcke and Thies, 2015). The number of tetracycline (Tc)-resistant colonies appeared after recombineering using 1 lg of donor plasmids pTetSac-DpvdD100:: Adaptor, pTetSac-DpvdD200::Adaptor, pTetSac-DpvdD400::Adaptor, pTetSac-DpvdD600::Adaptor, pTetSac-DpvdD800::Adaptor, and pTetSac-DpvdD::Adaptor harbouring pairs of 0.1-, 0.2-, 0.4-, 0.6-, 0.8-, and 1.0-kb HAs respectively (Table S1). Successful knockout of the pvdD gene of all eight colonies randomly selected from each knockout experimentor all colonies for experiments with less than eight colonieswas verified. The values and error bars represent means and standard deviations of colony counts from triplicate experiments, respectively, while all the actual three data points are also shown. However, engineering tools for the expression of such heterologous natural product BGCs in P. putida have not been optimal. In particular, most gene knockout tools have not been optimal for deleting large regions on the chromosome, and integration of heterologous BGCs has relied on either time-consuming homologous recombination (7-10 days/cycle) (Cao et al., 2012;Martinez-Garcia and de Lorenzo, 2012) or fast yet unpredictable transposon-mediated random insertion (3-4 days/cycle) (Martinez-Garcia and de Lorenzo, 2012; Loeschcke et al., 2013;Domrose et al., 2017). Thus, development of rapid and reliable tools for knocking out and integrating largesized genes/clusters had been demanded. Our markerless recombineering system for P. putida, of which detailed protocols are provided in this paper, allows rapid (4N + 1 days for N cycles of recombineering) and efficient deletion of chromosomal regions and integration of heterologous BGCs. Examples of recombineering include, but the capacity of the system is not limited to, knocking out 17 different regions of 0.6-101.7 kb and integrating heterologous BGCs of 1.2-7.4 kb (Choi et al., 2018), successfully demonstrating the use of the recombineering system in constructing plasmid/marker-free overproducers of heterologous natural products. Engineering of P. putida using this recombineering system is expected to be expedited by adopting recent strategies of combining the recombinase and Cre vectors (i.e. plasmids pJB658-recET and pRK2Cre in this system, respectively) (Song and Lee, 2013) and engineering of multiple loci simultaneously (Jensen et al., 2015;Cho et al., 2017), facilitating the study of P. putida and related species and streamlining the construction of high performance P. putida overproducer strains.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article.