mCloverBlaster: a tool to make markerless deletions and fusions using lambda red and I-SceI in Gram-negative bacterial genomes

Abstract

This study introduces mCloverBlaster as a genetic tool to create deletions and transcriptional and translational fusions in bacterial genomes using recombineering. The major advantage of this system is that it can be used to make deletions and fusions without leaving a selectable marker on the chromosome. mCloverBlaster has a kanamycin resistance cassette with an I-SceI restriction site flanked by fragments of the gene for the mClover3 fluorescent protein including direct repeats of mClover3 sequence on both sides of the kanamycin resistance gene. The mCloverBlaster sequence is introduced into the chromosome using lambda red recombineering, expression of I-SceI creates a double stranded break in the kanamycin resistance cassette that initiates a recombination event that can occur in the mClover3 repeats. This recombination results in the simultaneous removal of the kanamycin resistance gene and the restoration of a functional mClover3 gene that can be used as a reporter. Here, this system was used to replace the rcsB stress response gene in Serratia marcescens. The strain was tested for mClover3 fluorescence as a reporter for rcsB gene expression in the wild type and isogenic ΔgumB mutant and evaluated for pigment biosynthesis. In summary, mCloverBlaster is a molecular genetic tool to make markerless mClover3 fusions and gene deletions.

1. Introduction

Recombineering technologies such as the lambda red system have revolutionized molecular biology across many genetic systems from prokaryotic to eukaryotic and viral research (1, 2). A common use for this approach is to generate mutations by replacing a gene with an antibiotic resistance cassette. A disadvantage of this approach is that subsequent use of plasmids or other manipulations in the mutant strain can be restricted because of the innate antimicrobial tolerance and acquired resistance elements of many clinical isolates, and even some environmental bacterial isolates. Therefore, researchers have developed several methods to recycle selectable markers, that is, to use a selectable marker to make a genetic manipulation on a chromosome, plasmid, or artificial chromosome, and then remove the resistance marker using a recombination event. In this way, the same selectable marker can be used serially to make double and triple mutants or introduce a plasmid with the same resistance marker previously used to make a chromosomal mutation. A major solution to recycling markers has been to flank the selectable marker with short sequences, such as frt or loxP, recognized by recombinase enzymes followed by expression of the recombinase (3, 4). In this fashion, the recombinase targets recombination at the flanking repeats and the selectable marker is excised from the DNA. One drawback of this is the frt or loxP site is retained on the chromosome, and this is of little use except as a priming site.

An alternative approach to recycling selectable markers is the use of a dual purpose selectable and counter-selectable marker such as the Saccharomyces cerevisiae URA3 gene (orthologous to pyrF in Escherichia coli), flanked by repeating DNA. After targeted integration into the chromosome using URA3 as the selectable marker, the microorganism is subject to counter-selection by an antimetabolite (5-fluoroorotic acid), and the subset of microbes that had experienced a recombination between the flanking DNA will have lost the selectable marker and therefore survive counter selection. This method has been used to mutate countless genes in various yeasts, and was named the “ura-blaster” method (5). The original ura-blaster cassette has the URA3 gene of S. cerevisiae flanked by direct repeats of hisG DNA from Salmonella (6). Resolution of the cassette, following selection on medium with 5-fluoroorotic acid, leaves the microorganism with a single hisG sequence that can be useful as for hybridization or PCR, but is otherwise not functional.

In this study, we co-opted the ura-blaster concept of repeated sequences flanking a selectable marker for positive selection and I-SceI meganuclease from S. cerevisiae for counter selection. However, the recycling recombination yields an intact gene that can be used as a transcriptional or a translational reporter. Specifically, the improved GFP fluorescent reporter protein mClover3 (7) was used because of its versatile functionality. Herein, we describe the generation of the mCloverBlaster and its use in the deletion and replacement of the rcsB gene with mClover3 in a clinical isolate of the opportunistic Gram-negative pathogen Serratia marcescens.

2. Materials and methods

2.1. Bacterial cultures.

Bacteria were maintained at −80°C and streaked out on Lysogeny Broth (8) agar plates and grown and incubated at 30°C as noted in the text or in LB broth on a TC-7 tissue culture rotor at 30°C. Kanamycin was used at 50 μg/ml for Escherichia coli or 100 μg/ml for S. marcescens; tetracycline and gentamycin were used 10 μg/ml. Strains are listed in Table 1.

Table 1.

S. marcescens strains used in this study

Name	Description	Source or Reference
K904	S. marcescens keratitis isolate	(16)
ΔgumB	K904 with gumB deletion allele	(28)
ΔrcsB	K904 with rcsB replaced by mClover	This study
ΔgumB ΔrcsB	ΔgumB with rcsB replaced by mClover	This study

2.2. Plasmid construction and genetic manipulations.

Plasmids are listed in Table 2. A plasmid to direct the replacement of rcsB allele with mCloverBlaster was generated using two synthetic double stranded DNA fragments (G-block fragments from Integrated DNA Technologies), listed in Table 3 as nucleotide number 4767 and 4768. The two fragments were designed with overlapping regions to direct recombination with each other and a shuttle vector in S. cerevisiae as previously described (9). The two fragments and shuttle vector pMQ132 (which had been linearized by the SmaI restriction enzyme (New England Biolabs)), were used to transform S. cerevisiae by chemical transformation (10), and colonies were selected on synthetic complete medium without uracil (10). The resulting plasmid, pMQ739, has the following rationally designed sequences in order: 504 bp of DNA upstream of the rcsB open reading frame (ORF), a ribosome binding site, an AvrII restriction site, the first 388 bp of mClover3 (7), an 1105 bp kanamycin resistance cassette with the nptII gene, an I-SceI meganuclease site, the last 583 bp of mClover3, a BamHI restriction site, and 501 bp downstream of the rcsB ORF. The plasmid was validated by PCR analysis and sequencing.

Table 2.

Plasmids used in this study.

Name	Description	Source or Reference
pRS416	shuttle vector, CEN6/ARSH4, URA3	(Sikorski and Hieter, 1989)
pMQ117	oripSC101-ts, oriT, aacC1, CEN6/ARSH4, URA3	(Shanks et al., 2009)
pMQ125	orip15a, oripRO1600, aacC-1, PBAD-lacZ	(Shanks et al., 2009)
pMQ132	oripBBR1, oriT, aacC1, CEN6/ARSH4, URA3	(Shanks et al., 2009)
pMQ337	orip15a, oripRO1600, aacC-1, PBAD-I-SceI	(Shanks et al., 2009)
pMQ538	pMQ125 with lambda red recombination genes	(Brothers et al., 2019)
pMQ614	pMQ132 + rcsB from S. marcescens	(Brothers et al., 2019)
pMQ739	pMQ132 + rcsB::mClover-kan	This study
pmCloverBlaster-kan	pRS416 + mClover-kan	This study

Table 3.

Nucleotides used in this study

Name	Sequence
4767
cctcttcgctattacgccagctggcgaaagggggatgtgctgcaaggcgattaagttgggtaacgccagggttttcccagtcacgacg ttgtaaaacgacggccagtgccaagcttgcatgcctgcaggtcgactctagagctacattcgggtgaactataatctgaacggcgcattaatcgacgcc gtgctgatgctgatcgaacagcaaatggccgcgctggaacaggaagaaagcccgctttcattgagctcagaagatatccaactctatgaaaaacaattgaa atcaagtgattactatgggctgtttgtcgatacagtacccgacgatgtcaaaaaactgtatactgaggcgggcagcagtgatttcaatgcgctgtcacaaacc gcacaccgcctgaaaggcgtgtttgccatgttaaatctgcttcccggcaagcagctgtgcgaatcgttagaacagcgcatcgcagaaggtgatgcgcccga gatcgagaataacatcagtcagattgattttttcgtcagcagactgctgaagcaaggtagccaacaacatgaataacctgaacgtaattattgctgatgaccat cctatcgtactgtttggcatccggaagtcacttgagcaaattgaatgggaggacctaggaatggtatcaaaaggggaagaactttttaccggggtagtaccaa ttcttgttgaattggatggtgacgtgaacggccacaagttcagcgtacggggagagggcgaaggtgacgccacgaatgggaaacttacactgaaatttatct gcacaaccggaaaactccccgttccctggccaacattggttacgacattcgggtacggggttgcctgcttcagtcggtacccggaccatatgaagcaacac gatttctttaagtctgccatgccggagggttatgtccaggaacggactatttcattcaaagacgacggaacatataaaacccgtgcggaagttaagtttgagg gcgataccttggtaaatcgcattgaactcaagggcattggaaagccagtccgcagaaacggtgctgaccccggatgaatgtcagctactgggctatctgga caagggaaaacgcaagcgcaaagagaaagcaggtagcttgcagtgggcttacatggcgatagctagactgggcggttttatggacagcaagcgaaccg gaattgccagctggggcgccctctggtaaggttgggaagccctgcaaagtaaactggatggctttcttgccgccaaggatctgatggcgcaggggatcaa gatctgatcaagagacaggatgaggatcgtttcgcatgattgaacaagatggattgcacgcaggttctccggccgcttgggtggagaggctattcggctatg actgggcacaacagacaatcggctgctctgatgccgccgtgttccggctgtcagcgcaggggcgcccggttctttttgtcaagaccgacctgtccggtgcc ctgaatgaactccaagacgaggcagcgcggctatcgtggctggccacgacgggcgttccttgcgcagctgtgctcgacgttgtcactgaagcgggaagg gactggctgctattgggcgaagtgccggggcaggatctcctgtcatctcaccttgctcctgccgagaaagtatccatcatggctgatgcaatgcggcggctg catacgcttgatccggctacctgcccattcgaccaccaagcgaaacatcgcatcgagcgagcacgtactcggatggaagccggtcttgtcgatcaggatga tctggacgaagagcatcaggggctcgcgccagccgaactgttcgccaggctcaaggcgcggatgcccgacggcgaggatctcgtcgtgacccatggcg atgcctgcttgccgaatatcatggtggaaaatggccgcttttctggattcatcgactgtggccggctgggtgtggcggaccgctatcagga
4768
gactgggcacaacagacaatcggctgctctgatgccgccgtgttccggctgtcagcgcaggggcgcccggttctttttgtcaagaccgacctg tccggtgccctgaatgaactccaagacgaggcagcgcggctatcgtggctggccacgacgggcgttccttgcgcagctgtgctcgacgttgtcactgaag cgggaagggactggctgctattgggcgaagtgccggggcaggatctcctgtcatctcaccttgctcctgccgagaaagtatccatcatggctgatgcaatg cggcggctgcatacgcttgatccggctacctgcccattcgaccaccaagcgaaacatcgcatcgagcgagcacgtactcggatggaagccggtcttgtcg atcaggatgatctggacgaagagcatcaggggctcgcgccagccgaactgttcgccaggctcaaggcgcggatgcccgacggcgaggatctcgtcgtg acccatggcgatgcctgcttgccgaatatcatggtggaaaatggccgcttttctggattcatcgactgtggccggctgggtgtggcggaccgctatcaggac atagcgttggctacccgtgatattgctgaagagcttggcggcgaatgggctgaccgcttcctcgtgctttacggtatcgccgctcccgattcgcagcgcatcg ccttctatcgccttcttgacgagttcttctgagcgggactctggggtagttacgctagggataacagggtaatataggccgcgaagttcctattccgaagttcc gaaatttatctgcacaaccggaaaactccccgttccctggccaacattggttacgacattcgggtacggggttgcctgcttcagtcggtacccggaccatatg aagcaacacgatttctttaagtctgccatgccggagggttatgtccaggaacggactatttcattcaaagacgacggaacatataaaacccgtgcggaagtta agtttgagggcgataccttggtaaatcgcattgaactcaagggcattgatttcaaagaggatgggaatatcctgggtcacaagttggaatataattttaattcac actacgtgtacattaccgctgataaacagaaaaactgcataaaggctaacttcaaaatacgccataatgtggaggacgggtcagtacaactcgcagaccatt accaacagaacacaccaataggtgacggaccagtgcttctgcccgataaccattatctgtcccatcaaagcaagctctcgaaagatccgaacgagaaacg ggatcacatggtattgttggagtttgtcactgccgctggaattacacacggaatggatgagctgtataagtaaggatccagcccgcctcacggcaggttaac gccaaacaggcacggcatccgccgtgcctgtttgcttttcagacctcgctgcggttgtagcgtacctgctggctgtaataagccagcgtctgctccagcgtct ccagcgtaaccggcttcgacaggcaattatccatgcccgcctcgaggcagcgctgtttctcttccgccaacgcgttcgccgtcacgccgatcaccggcgcg ctgaactgcatctgccgcagttcctgcgtcagacgatagccgtccatgttcggcatattgacgtcggtcagcacaatatcgacgtggtgctgtttcagcacgc cgagcgcgtcgacgccgtcgttggcggtcaccacctgataccccagcgaccccagctgatcggacaacaggcgacggttgatcggatgatcgtcaacca ccagcagatgaatatcgccgttgtccgccgcactggccttggcgggcaccggcagctgtaccagcgcctctggggtaccgagctcgaattcgtaatcat ggtcatagctgtttcctgtgtgaaattgttatccgctcacaattccacacaacatacgagccggaagcataaagtg
4876	tcgacggtatcgataagcttgatatcgaattcctgcagatggtatcaaaaggggaagaac
4877	caccgcggtggcggccgctctagaactagtggatcccttacttatacagctcatccattc

All nucleotides are listed as 5’ to 3’. Bolded sequence directs recombination with vector pMQ132. Underlined sequences direct recombination between dsDNA sequences.

To generate a pmCloverBlaster-kan, the mClover3 and kanamycin cassette from pMQ739 were amplified with oligonucleotide primers 4876 and 4877 and introduced into pRS416, which had been cut with SmaI. The amplicon and linearized plasmid were used for yeast in vivo recombination cloning as noted above. The entire plasmid was sequenced using the PacBio platform (SNPsaurus, Inc) and the total assembled sequence was deposited at Genbank (accession number MT23788).

To replace rcsB with mCloverBlaster, the rcsB::mCloverBlaster region was amplified from pMQ739 that had been linearized by digestion with SacII and StuI by PCR using Phusion high-fidelity polymerase. The resulting amplicon was concentrated by ethanol precipitation and used for electroporation with S. marcescens strains K904 and ΔgumB that had been prepared for lambda red recombination. Bacteria with pMQ538 to express lambda red recombination genes (1) were grown to OD₆₀₀=0.5 and arabinose was added to 1 mM followed by incubation at 30°C for 30 minutes on ice for 45 minutes, and subsequent washes with cold sterile water and 10% glycerol. One to three μg of DNA was used in the electroporation in 1 mm gap electroporation cuvettes using a MicroPulser Electroporator (Bio-Rad) on the Ec1 setting with a single pulse and 1.8 kV (11). Bacteria were grown in LB broth at 30°C for 2–20 hours then plated on LB plates with 100μg/ml kanamycin. Candidates were evaluated by PCR as detailed below.

2.3. Phenotypic analysis.

To evaluate mClover3 fluorescence, independent single colonies were used to inoculate test tubes with 5ml of LB broth and grown with aeration for 20 hours at 30°C. Samples (150μl) were loaded into black sided and clear-bottomed 96 well plates (Nunc 165305) and optical density at 600 nm (OD₆₀₀) and fluorescence was measured using a BioTek Synergy 2 plate reader. Relative fluorescence units (RFU) were defined as the ratio of arbitrary fluorescence units to OD₆₀₀. A total of 6 independent colonies were tested in two separate experiments.

Prodigiosin was measured by growing colonies as noted above. After 20 hours of growth at 30°C, OD₆₀₀ was determined using a 1 cm path length cuvette and prodigiosin was extracted as previously described by Slatter et al (12), and measured by absorbance at 534 nm using a Spectramax M3 spectrophotometer. The ratio of A₅₃₄/OD₆₀₀ was used as metric to evaluate prodigiosin biosynthesis. Six independent colonies were tested in two separate experiments.

Plasmid maintenance was determined by growing the WT with pMQ337 (13) and pMQ538 (14) bacteria grown on plates with gentamicin for plasmid selection. Single colonies were transferred to test tubes with 5 ml of LB medium without gentamicin and grown with aeration (200 rpm) at 37°C for 18–20 h. Cultures were serial diluted and plated on LB agar plates with and without gentamicin.

3. Results and Discussion

3.1. Use of mCloverBlaster to delete the rcsB gene of S. marcescens.

The mCloverBlaster cassette was designed in silico (Fig 1). It has the first 388 base pairs of mClover3 followed by the kanamycin/neomycin resistance gene, nptII, followed by an I-SceI meganuclease site, and the latter 586 base pairs of the mClover3 gene. This importantly includes 254 bp of repeated mClover3 sequence (Fig 1, dark green) on each side of the kanamycin marker that can be used for recombination to remove the kanamycin cassette. The use of a repeated region of DNA was inspired by the hisG DNA flanking the URA3 gene in the widely used “ura-blaster” approach (6), in which the repeated hisG region is used to recycle the URA3 selectable marker, leaving one copy of hisG as a scar on the chromosome after recombination between the hisG repeats. Here, however, recombination between the repeated regions in mClover3 will result in a functional mClover3 gene that can be used as a transcriptional or translational reporter. Another critical component of the mCloverBlaster is an I-SceI meganuclease site proximal to the kanamycin resistance gene that can initiate a double stranded break to facilitate conjugation between the repeated mClover3 sequences.

Figure 1.

Schematic diagram of strategy to replace chromosomal rcsB gene with mcloverBlaster.

To test the utility of this approach, we targeted the rcsB transcription factor gene in the Gram-negative bacterium S. marcescens. A schematic representation of this approach is shown in Fig 1. The transcription factor is a known regulator of the shlA cytolysin, the flagellar regulon, capsular polysaccharides, and pigment biosynthesis (14, 15). In this study we chose to delete the rcsB gene in the genome of the wild-type ocular clinical isolate K904 (16) and an isogenic ΔgumB mutant (Table 1). The GumB protein is predicted to inhibit the RcsB protein indirectly through the RcsC and RcsD phosphotransfer proteins (17); however, it is unknown whether transcription of rcsB is altered in a gumB mutant.

To target rcsB, we cloned DNA upstream and downstream of the rcsB ORF, 504 bp and 501 bp respectively, along with the mCloverBlaster onto plasmid pMQ132 (see Materials and Methods). The resulting rcsB::mCloverBlaster amplicon was generated by PCR and used for electroporation into bacteria expressing the lambda red recombination machinery (using plasmid pMQ538) (7). The resulting kanamycin resistant colonies were screened for replacement of rcsB with the mCloverBlaster cassette (Fig 2). After loss of the pMQ538 plasmid by growth without gentamicin, plasmid pMQ337 (13), which expresses the I-SceI gene under control of the L-arabinose-inducible P_BAD promoter (18), was introduced into the bacteria and mCloverBlaster replaced rcsB (Fig 1). Following expression of I-SceI, single colonies were screened for loss of kanamycin resistance, and susceptible colonies were verified for loss of the kanamycin gene by PCR. A shift in the migration pattern of the rcsB region, consistent with removal of the nptII resistance gene, was observed (Fig 2). The approach was used in an isogenic ΔgumB mutant with the same outcome (Fig 2).

Figure 2.

Evaluation of chromosomal manipulations and fluorescence. Agarose gel of PCR products demonstrating expected migration patterns of the bacterial chromosome as shown in Fig 1.

Because plasmid loss is required for this protocol, we evaluated plasmid maintenance of pMQ337 and pMQ538 in the WT. A plasmid maintenance frequency of 0.70±0.51 and 0.44±0.23 (n≥10) was measured for pMQ337 and pMQ538 respectively following one round of growth without selection in LB broth at 37°C. This indicates that after a single over-night growth experiment without selection, around 30–65% of the resulting colonies will have lost the plasmid, suggesting that the desired plasmid loss can easily be achieved by patching out 10–20 single colonies.

3.2. Evaluation of rcsB mutants

If the recombination proceeded as planned, then the resulting strains should express the intact mClover3 gene under control of the rcsB promoter. Therefore, fluorescence was measured from cultures with a plate reader. Relative green fluorescence (RFU) measured, following overnight growth in LB medium at 30°C, was close to zero for the wild-type strain without mClover (31±127 RFU) and detectable in the rcsB::mClover derivative (954±145 RFU). Together these data indicated the successful generation of the intact mClover gene in place of rcsB and elimination of the kanamycin resistance marker.

S. marcescens is known for its production of the red pigment prodigiosin, which is a highly regulated secondary metabolite noted for its anti-tumor and autophagy-inducing properties (19–22). Here we evaluated the impact of altered Rcs system activity on prodigiosin biosynthesis, as we have previously demonstrated that constitutive activation of the Rcs system through mutation of gumB highly reduces pigment production. It would therefore be expected that mutation of rcsB should inactivate the Rcs system and increased pigmentation will result. This prediction was supported by our findings (Fig 3), where the rcsB::mClover3 and ΔgumB rcsB::mClover3 double strains produce more pigment than the corresponding strain with a functional rcsB gene. These results suggest a negative regulatory role for the Rcs system in control of prodigiosin biosynthesis that will be explored in greater detail in subsequent studies. Moreover, they serve as a proof-of-concept for the efficacy of the mCloverBlaster approach.

Figure 3.

Mutation of rcsB correlates with increased prodigiosin biosynthesis. Bacteria of indicated genotypes were grown in LB medium at 30°C for 18 hours and prodigiosin pigment was extracted and measured by absorbance. Representative images of extracted pigment (A) and quantified prodigiosin normalized by culture turbidity (B) are shown. Averages and standard deviations are shown, n=6. Asterisks indicate p<0.05 as measured by ANOVA with Tukey’s post-test.

3.3. Generation of a mCloverBlaster plasmid.

To enhance convenience of the mCloverBlaster, it was cloned into a multicloning site on plasmid pRS416 (23), so that the mCloverBlaster cassette is flanked by numerous restriction sites to facilitate ligation-mediated cloning approaches. The resulting plasmid was named pmCloverBlaster-kan and was sequenced using the PacBio platform to verify the construct.

This approach could easily be modified for other reporter genes, such as mcherry or lacZ, that could be used as fusions (24) and other selectable markers. Furthermore, rather than I-SceI, the system could be modified by using a selectable/counterselectable marker such as tetA (25, 26), or a combination of both a selectable and counterselectable marker, such as the “janus cassette” with rpsL for counter selection and aphIII gene for selection with kanamycin (27).

4. Conclusions

This study has generated a new approach to gene replacement and modification in Gram-negative bacteria that could be used directly for a wide number of approaches, including deleting genes and generating fusions for transcriptional or subcellular localization analysis. Furthermore, the tools generated for this study could easily be modified for specific projects, such as incorporation of different resistance cassettes and genes for different fluorescent proteins.

Highlights.

• Recombineering based method for markerless deletions and insertions

• Combination of I-SceI, Lambda Red, and mClover

• Characterize role for RcsB in prodigiosin biosynthesis regulation

• Evaluate role of IgaA-protein regulation of rcsB gene expression