2.1 Bacterial strains and routine culture

Strains were isolated from yellowtail kingfish (Seriola lalandi) during 2015–2016 disease outbreaks in South Australia and Western Australia, identified and deposited at the Department of Primary Industries and Regional Development (DPIRD), Western Australia as Pdp AS-16-0540-1 and Pdp AS-16-0555-7, and subsequently deposited as QMA0505 and QMA0506 at The University of Queensland (UQ), where isolates were sequenced (Baseggio et al., 2021). Bacteria were cultured at 25°C on Tryptic soy agar (TSA) or in Tryptic soy broth (TSB), unless otherwise specified.

2.2 Preparation of electrocompetent cells

Initially, we tested best-performing conventional protocol in Pdp according to benchmarking performed by Cutrin et al. (1995) as described, except using TSB to grow cells instead of Brain heart infusion broth. Subsequently, cells were rendered competent using a salt-free protocol developed for preparation of electrocompetent streptococci (Framson et al., 1997) with minor modifications. The overnight TSB cultures were diluted to 10³ colony-forming unit (CFU)/mL (OD₆₀₀ = 0.025 in Eppendorf BioPhotometer 6131) in 150–300 mL TSB, and grown to early exponential phase (10⁴–10⁵ CFU/mL; OD₆₀₀ = 0.25–0.4). Harvested cultures were transferred to 50 mL tubes and chilled on ice; pelleted for 5 min at 2000g, 4°C; resuspended in 50–100 mL of ice-cold 0.625 M sucrose solution in water (pellets combined) and pelleted for 10 min at 18 500g, 4°C; resuspended in 10 mL of ice-cold 0.625 M sucrose (pellets combined), transferred into 15 mL tube, pelleted for 15 min at 18 500g, 4°C, and resuspended in 0.5–1 mL of the same buffer (or remaining supernatant if pellet was loose). This preparation was also attempted at room temperature using TSB supplemented with 1% salt to grow the culture. Competent cells were aliquoted by 50–90 μL into chilled 1.5 mL tubes and stored at – 80°C.

2.3 Transformation efficiency estimation

Competent cells were defrosted on ice, mixed with 100 ng of a plasmid carrying antibiotic resistance marker, pET-28a(+), pLZ12spec, or pUC19, in 10 μL H₂O, transferred to chilled 0.1 cm gap electro cuvettes (Sigma), and electroporated at the default P1 in Eppendorf Eporator (1250 V, 5 ms, 600 Ω). Cells were recovered in 0.5–1 mL of 0.25 M sucrose TSB for 2.5 h, 50 μL aliquots were spread on 75 μg/mL kanamycin (Km), 200 μg/mL spectinomycin, or 1 μg/mL ampicillin plates, and transformants were counted after 48 h. Viable cell counts were performed by Miles and Misra method (Miles et al., 1938) immediately after preparation, after defrosting, after electroporation, and after the recovery period.

2.4 Allelic exchange KO mutagenesis

2.4.2 Construction of the allelic exchange “KO minicircles”

Circular minimalistic AECs (“KO” minicircles) for inactivation of aip56 and ssrA/smpB genes were designed to contain homology arms spliced between selectable and counterselectable marker genes (Figures 1b and 2b). kanR2 gene (selectable marker conferring Km resistance; expressed under target gene transcription elements) and sacB gene with promoter and terminator (counterselection marker conferring sucrose sensitivity) were amplified using Km_F/R and sacB_F/R primer sets (Tables 1 and 2), from 0.5 ng of pET-28a(+) (Addgene; #2526) and pRE107 (Addgene; #43829) plasmids, respectively. The pET-28a reaction was treated postamplification with DpnI (NEB) to eliminate transformation background produced by pET-28a(+) plasmid used as a PCR template. The latter was extracted immediately before amplification to ensure that methylation required for DpnI restriction was retained. Homology regions (0.9–1.4 kb sequences upstream and downstream from aip56 and ssrA-smpB targets) were amplified from Pdp strain QMA0505 gDNA (CP061854-60) using chimeric primers containing GA overhangs homologous to the ends of kanR2 and sacB (aip56_up_F/R, aip56_down_F/R, ssrA_up_F/R, and ssrA_down_F/R reactions; Tables 1 and 2). Fragments (Figures 1a and 2a) were spliced into KO minicircles (Figures 1b and 2b) using NEBuilder HiFi DNA Assembly Cloning Kit (NEB) in 1:1:1:1 molar ratio in 30 μL reactions for 2 h (total of 0.75 pmol DNA, maximum recommended input). Schematic representations of minicircles in Figures 1b and 2b were created in Geneious Prime 2020.2.2.

2.5 Sequencing of mutagenesis clones and detection secondary of mutations

To detect secondary mutations, DNA was extracted from selected clones generated during aip56 mutagenesis (QMA0648, QMA0780-89) using Quick-DNADNA Miniprep Plus Kit (Zymo Research). Short-read (Illumina) library preparation and sequencing were performed at the University of Queensland Sequencing Facility (UQ, Brisbane, Australia). Sequencing libraries were prepared using the Nextera XT DNA Library Prep Kit (Illumina), purified with 30 µL of AxyPrep Mag PCR Clean-up beads (Axygen), quantified on the PerkinElmer LabChip GX with the DNA High Sensitivity Reagent Kit (PerkinElmer; CLS760672), and pooled in an equimolar ratio. Sequencing was performed using the Illumina NextSeq500 (NextSeq control software v4.0.0/Real-Time Analysis v2.11.3). The library pool was diluted and denatured according to the standard NextSeq protocol and sequenced to generate paired-end 151 bp reads using a 300-cycle NextSeq500/550 Mid Output Reagent Kit v2.5 (Illumina). After sequencing, fastq files were generated using bcl2fastq2 (v2.20.0.422). For better comparability/assessment of the initial frequency of mutation, WT strain QMA0505 was also resequenced along with mutagenesis clones as Illumina reads for it's published genome C06185-460 were obtained using different sequencing kit and instrument (Baseggio et al., 2021). Reads were trimmed using fastp 0.23.2 and paired during export into Geneious Prime 2023.0.4. Using the latter software, reads were mapped to the CP061854-60 genome (Baseggio et al., 2021) using Geneious assembler under default settings. Variants were identified using “Find variations/SNPs” command with minimum coverage set to 30 and minimum variant frequency set to 0.9. Where annotations differed between the polymorphism tracks, alignment of the reads was examined for strains/contigs lacking the identified variants, and the latter were assigned if lower than set coverage was the reason for the variant not being annotated. Where lower variant frequency was the reason, reads with and without the variant were counted and polymorphism was assigned for over 0.6% frequency cases. QMA0505 read alignment at each variant position was examined to identify whether mutation is de novo. Regions of no read coverage indicating putative large deletions were identified using “Find High/Low coverage” command with a number of sequences under the low coverage flag set to “0.” Where large regions of no coverage were found long-read (Nanopore) sequencing was carried out to confirm the putative large deletions. For the latter, libraries were prepared using Rapid Barcoding Kit 24 V14 (SQK-RBK114.24) and sequenced on a MinION Mk1C device equipped with FLO-MIN114 (R10) flow cell. Nanopore reads were assembled using Flye v2.9, annotated with Prokka v1.12, and exported to Geneious. Regions containing putative deletions were extracted from lon-gread assemblies and the CP061854-60 genome sequence, and aligned using Clustal Omega.

3.1 Salt-free preparation of electrocompetent Pdp

We aimed to inactivate two targets via gene KO mutagenesis: AIP56 toxin and SsrA-SmpB ribosomal rescue system, an established and recently proposed virulence factors in Pdp, respectively (Baseggio et al., 2021). For the purpose of obtaining isogenic KO mutants, we wanted to avoid mating with SM10/S17-1 λ pir donor strains of E. coli due to high off-target mutation rate produced by RP4::Mu conjugative transfer cluster (Ferrières et al., 2010; Strand et al., 2014). Transformation is a more straightforward and rapid approach compared to conjugation since allelic exchange DNA is delivered directly to the target organism. Furthermore, ssDNA entering the recipient's cell via conjugation activates bacterial stress response SOS which induces error-prone recombination (Baharoglu et al., 2010; Virolle et al., 2020). Consequently, transformation with double-stranded DNA (dsDNA) is likely to offer lower probability of secondary mutations. For transformation via electroporation, cells are typically washed in hypertonic solutions, as high osmotic pressure stabilizes the cell membrane and decreases cell lysis during electroporation. Most protocols rely on salt ions to create high osmotic pressure; however, salts cause arcing during electroporation and must be washed off in advance. Washed cells are typically resuspended in an electroporation solution containing glycerol that prevents freezing damage and allows storage (Wirth et al., 1989). However, many bacterial species and strains are not readily transformable by standard methods (Aune & Aachmann, 2010), and high transformation efficiency is required for the delivery of nonreplicating DNA in allelic exchange mutagenesis. Case in point, previous attempts to render Pdp cells electrocompetent using the outlined basic approach were not particularly successful. Multiple conditions were evaluated by Cutrin et al. in 1995, and according to the authors, excessive cell lysis occurred when cells were grown in a salty medium (standard for marine species) and prepared using low ionic strength buffers, and arcing occurred when cells were prepared using “isotonic for the bacteria” (i.e., high ionic strength) buffers (Cutrín et al., 2000). Of all the evaluated conventional methods, the highest efficiency of 9.8 × 10² CFU/μg DNA was achieved with a laborious multiple-buffer multiple-reagent protocol immediately after preparation without defrosting/prior storage (Cutrín, 1995). We attempted the same procedure in our laboratory and obtained around 10² CFU/μg transformation efficiencies with pET-28a(+) and pUC19 plasmids.

Since higher efficiencies are required for transformation with nonreplicating DNA, we attempted a rapid and straightforward single-buffer protocol originally developed by Dunny et al. (1991) for gram-positive bacteria. The method was subsequently applied by Framson et al. (1997) for group B streptococci and optimized to contain no salt in the buffer, which rendered this single-buffer protocol to be also a single reagent. The sole ingredient in the wash/electroporation buffer is 0.625 M sucrose, which works as a hypertonic stress agent (nonionic osmotic pressure), electroporation medium, and cryoprotectant. Following the above protocol, we have repeatedly achieved transformation efficiencies of 10⁶–10⁷ CFU/μg DNA in Pdp using pET-28a(+) and pUC19 vectors, and somewhat lower efficiencies of 10⁵–10⁶ CFU/μg DNA were obtained with streptococcal pLZ12spec vector. Notably, these efficiencies were estimated with defrosted cells previously stored at −80°C. We performed the protocol with only minor modifications: TSB was used instead of Todd–Hewitt broth (used to grow streptococci) with no supplementation with glycine, which is used to weaken cell walls in gram-positive bacteria but unnecessary for gram-negative preparations. Notably, supplementation of the growth medium with salt used to culture marine bacteria including Pdp was also routinely omitted for competent cell preparation to reduce the chances of arcing. Yet, there were no arcing events or reduction in transformation efficiency in several preparations from cultures grown in TSB supplemented with 1% salt (1.5% final salt concentration). This suggests that the method can potentially be used for strictly halophilic bacteria if salts from the growth medium are efficiently washed off by the sucrose solution. There was only minor reduction in the number of viable cells after defrosting and a 10-fold reduction after the electroporation. We have also tried room-temperature preparation and transformation, which may be used as a more convenient alternative to traditional ice-cold technique (Tu et al., 2016). This approach was functional but 10–100 times less efficient.

We have efficiently transformed sucrose-treated electrocompetent Pdp with minimalistic AEC generated by GA (KO minicircles) targeting aip56 and ssrA-smpB loci (see subsections 3.2 and 3.3), which demonstrates that achieved transformation efficiency is sufficient for allelic exchange mutagenesis with nonreplicating DNA. The most prominent difference between this protocol and the conditions evaluated by Cutrín et al. (1995) is the composition of the hypertonic (osmotic stress) buffer. The former has increased sucrose concentration and a complete absence of salt; thus, the osmotic pressure created is completely nonionic. This is highly advantageous because any unwashed salts can cause arcing, while washing/resuspension steps require extra time and handling, and may lower the stabilizing effect on the membranes created by a hypertonic buffer. Further, divalent cations from salts may be used as cofactors by nucleases targeting the transforming DNA (Wang & Griffiths, 2009). Indeed, salt-free preparations dramatically improve transformation efficiency in Vibrio parahaemolyticus, which was explained by deactivation of extracellular DNase abundantly secreted by Vibrio (Wang & Griffiths, 2009). In contrast, highly concentrated sucrose solution may be conveniently used as both hypertonic buffer and electroporation buffer as it creates nonionic osmotic pressure, which efficiently stabilizes cell membranes without adversely affecting the electroporation or augmenting DNase activity. Moreover, it has cryoprotective properties, which allow freezing of competent cells without the addition of glycerol.

3.2 “KO minicircles”: circular AECs generated by GA

Nonreplicating plasmids (suicide vectors) are most often used to deliver mutagenic DNA constructs in allelic exchange mutagenesis. Yet, suicide vectors are notorious for illegitimate integration into host chromosomes and plasmids and other kinds of off-target mutagenesis (Johnson et al., 2003). Illegitimate recombination is very common in bacteria, and it is facilitated in plasmid sequences by the abundance of repeats and common/universal motifs (Desomer et al., 1991; Li et al., 2018; Oliveira et al., 2010; de Vries & Wackernagel, 2002). Remarkably, some suicide plasmids illegitimately integrate at frequencies sufficient for random mutagenesis experiments (Desomer et al., 1991). In some circumstances, plasmid backbones may be necessary, such as when transformation is not available, so DNA is delivered by conjugation, or when transformation is low in efficiency, so very large quantities of DNA are required the plasmid is replicated in and extracted from E. coli for electroporation. However, whenever high-efficiency transformation and other factors permit, vector-free KO approach should be considered as an alternative offering lower off-target mutagenesis probability. Vector-free AECs in previous studies were Fusion PCR products circularized by ligation (Gomaa et al., 2017) and linear GA products (Wu et al., 2019). Here we combined these two approaches: we used GA as a more convenient alternative to Fusion PCR technique (Huang & Wilks, 2017; Rudenko & Barnes, 2018), but employed it to generate circular constructs/products, as they are more stable compared to linear DNA. The constructs, here dubbed “KO minicircles,” were generated to knock out  aip56 and ssrA-smpB ribosomal rescue loci (Figures 1 and 2). Minicircles were designed for marked nonpolar deletions and contained four fragments assembled in the following order: -> upstream homology arm -> selectable marker gene (kanR2, Km resistance) -> downstream homology arm -> counter-selectable marker transcript (sacB, sucrose sensitivity) ->. In-frame insertion of the selectable marker gene (without transcription elements) in the place of the KO target does not cause polar effects, facilitates mutant selection, and helps to maintain purity of the culture. Alternatively, if unmarked deletion is desired, a selectable marker gene (with transcription elements) can be placed outside the homology regions along with a counter-selectable marker. Overall, the minicircle approach is exceptionally versatile and may be used with any selectable and counter-selectable marker/s, and for other genetic modifications including site-directed mutagenesis and gene knock-in, e.g., insertion of the WT gene for the KO rescue.

3.3 aip56 and ssrA-smpB mutagenesis: Selection of single-crossover mutants

High frequency of single-cross mutants (meroploids with integrated minicircles) in four transformations (aip56 and ssrA-smpB minicircles, dsDNA, and ssDNA each) provided evidence that: (1) transformation efficiency is sufficient for delivery of nonreplicating allelic exchange vectors or vector-free constructs; (2) KO minicircles generated by GA readily integrate into chromosomes and plasmids; and (3) electroporation with KO minicircles is a straightforward and efficient way to deliver allelic exchange DNA. Only ~0.7 μg of assembly products (0.65 μg of aip56 and 0.78 μg of ssrA-smpB construct DNA) were used in mutagenesis, which were delivered either double-stranded or denatured. Many Km-resistant clones were obtained in all transformations, although around two times more in transformations with ssDNA (Table 3). Randomly chosen selected colonies were PCR-screened for the presence of crossover using primer sets where one primer binds to the Km sequence and the second one binds to the genomic sequence outside the homology region. Over 70% of PCR-screened colonies yielded products of expected size (Figure 3 and Table 3), which unambiguously showed successful integration of minicircles. In the case of aip56, both homology arms recombined equally well (Figure 3a,b), while, in the case of ssrA, the upstream region was apparently less recombinable (Figure 3d,c).

3.4 aip56 and ssrA-smpB mutagenesis: Selection of double-crossover mutants

To select double-crossover mutants, meroploids obtained from dsDNA-minicircle transformations were used, as more likely to be isogenic (Baharoglu et al., 2010; Virolle et al., 2020). Multiple aip56 KO clones were obtained with ease after two subcultures of single-crossover mutants in nonselective broth (TSB) and plating on agar supplemented with sucrose (lethal to single crossovers) and Km (lethal to WT). Randomly chosen clones were PCR-screened for the loss of aip56 gene (Table 3). Several aip56-negative clones were further confirmed to be KO mutants by amplification across the allelic exchange region yielding a shorter product compared to WT strain (Figure 4). PCR screening for aip56 deletion showed that many meroploids were escaping sucrose counterselection, which is often weak and inefficient (Cianfanelli et al., 2020; Lazarus et al., 2019), despite omission of salt supplementation in selective plates. Since salt inhibits sacB expression (Blomfield et al., 1991), the latter marker may be suboptimal for slight halophiles such as Pdp and completely inappropriate for obligate halophiles. Nonetheless, double-crossover mutants were selected from 75% of the broth cultures and accounted for 16.65%–83% of the screened colonies (Table 3). Such a success rate was unexpected since aip56 is encoded on a highly abundant, small (<10 kB) plasmid that was anticipated to be hard to eliminate (Freitas et al., 2022). Potentially, aip56 KO would have been harder, perhaps impossible, to achieve using a suicide vector, since integration of a long vector backbone sequence with plasmid replication elements is likely to compromise the aip56 plasmid stability and/or replication . AIP56 is a known toxin characterized at the protein level, which is a primary virulence factor of Pdp causing apoptosis in professional phagocytic cells (Pereira et al., 2014; do Vale et al., 2005). However, isogenic KO mutants lacking the gene have not been characterized yet, which hampers the research aiming at a complete understanding of Pdp pathogenicity (Freitas et al., 2022).

Although both ssrA-smpB homology arms were efficiently recombining (Figure 3 and Table 3), double-crossover mutants were not obtained when followed as per aip56 KO selection despite multiple subsequent attempts involving up to 24 independent single-crossover cultures. Also, based on the hypothesis that ssrA KO mutant (if viable) would be generally impeded in growth and more sensitive to oxidative stress, antibiotics, and high sugar concentrations, we tried the following modifications to the selective conditions: stationary broth cultures, broth cultures with sucrose and sucrose/Km, and lower Km and sucrose concentrations on selective plates, but still we were unable to select ssrA-negative clones. Notably, supplementation of broth cultures with sucrose and lower sucrose concentration on plates dramatically increased the number of single-crossover mutants escaping sacB counterselection, and thus, decreased chances to detect double-crossover mutants, if any were present at minor frequency. Thus, either some alternative counter-selectable markers are required to obtain ssrA KO mutant, or ssrA may be essential in Pdp. Indeed, ssrA-smpB is essential in obligatory intracellular pathogens (Karzai et al., 2000), and Pdp causes primarily intracellular infections and shows relatively slow growth on laboratory culture media.

3.5 Further aip56 mutagenesis and detection of secondary mutations via whole-genome sequencing

3.5.1 Mutagenesis clones used for sequencing

We aimed to confirm that electroporation of KO minicircles into sucrose-treated competent cells can indeed generate highly isogenic KO mutants. However, sequencing of a single KO clone is not sufficient for this purpose as secondary mutations can occur during culturing either completely at random or can be associated with a particular selective pressures or/and inactivation of a particular gene (a target mutation). Thus, we obtained more aip56 KO mutants from several independent broth cultures of a single meroploid clone (clone used to select the initial KO strain). Also, from the same individual broth cultures we have obtained clones that regained the WT genotype (“revertants”), which originate in single-crossover cultures in the same manner as KO clones except for the second crossover happening on the same rather than the opposite homology arm. Since our meroploid contained sacB counterselectable marker conferring sucrose sensitivity, it was possible to select both kinds of clones in parallel on the sucrose agar, which was followed by plating on Km agar to differentiate between revertant/WT and KO genotypes (kanamycin-sensitive and kanamycin-resistant phenotype respectively). Plenty of clones were no longer sensitive to sucrose in three out of four nonselective meroploid broth cultures, and 30%–53% of the clones selected on sucrose were resistant to Km (Table 4). In most cases, PCR screening of selected clones for the presence of aip56 and kanR2 genes has confirmed either revertant to WT genotype (positive for aip56/negative for kanR2) or a KO genotype (negative for aip56/positive for kanR2) (Table 4, Figure 5). To detect and track secondary mutations, we sequenced 11 clones generated during mutagenesis: three pairs of KO and revertant clones selected on sucrose from three independent cultures, the initial aip56 KO clone selected on sucrose/kanamycin agar, and four meroploid clones selected on Km: two from ds aip56-minicircle transformation and two from ss aip56-minicircle transformation (Table 5).

Table 4. Selection of aip56 KO mutants and revertants to WT from independent nonselective broth cultures of a single meroploid clone QMA0786.

Number of clones	Culture 1	Culture 2	Culture 3	Cultures 4–6
Selected on sucrose plates	~50	~100	~50	None
Picked and patched on sucrose plates and kanamycin plates	14	13	15	-
Resistant to kanamycin	5	4	8	-
PCR-screened for aip56 gene and kanR2 gene	5	4	10	-
PCR-positive for aip56/PCR-negative for kanR2 (revertants to WT)	2	1	4	-
PCR-negative for aip56/PCR-positive for kanR2 (KOs)	2	2	3	-

• Abbreviations: KO, knockout; WT, wild type.

Table 5. Sequenced clones generated during aip56 mutagenesis.

Strain	Genotype	Parent strain	Selection	Accession numbers
QMA0505	WT	-	N/A	SRR25336394
QMA0648	KO	QMA0786	Sucrose/kanamycin	SRR25322645
QMA0780	Revertant (WT)	QMA0786	Culture 1	SRR25337471
			Sucrose	SRR25316070
				JAUYUZ000000000
QMA0781	Revertant (WT)	QMA0786	Culture 2	SRR25337470
			Sucrose	SRR25316069
				JAUYVA000000000
QMA0782	Revertant (WT)	QMA0786	Culture 3	SRR25322644
			Sucrose
QMA0783	KO	QMA0786	Culture 1	SRR25337469
			Sucrose	SRR25316068
				JAUYVB000000000
QMA0784	KO	QMA0786	Culture 2	SRR25337468
			Sucrose	SRR25316067
				JAUYVC000000000
QMA0785	KO	QMA0786	Culture 3	SRR25337467
			Sucrose	SRR25316066
				JAUYVD000000000
QMA0786	Meroploid	QMA0505	ds-MC transformation kanamycin	SRR25322643
QMA0787	Meroploid	QMA0505	ds-MC transformation kanamycin	SRR25322642
QMA0788	Meroploid	QMA0505	ss-MC transformation kanamycin	SRR25322641
QMA0789	Meroploid	QMA0505	ss-MC transformation kanamycin	SRR25322640

• Abbreviations: ds, double-stranded; KO, knockout; N/A, not available; ss, single-stranded; WT, wild type.

3.5.2 Secondary mutations

Short reads obtained from Illumina sequencing (SRR25322640–45, SRR25336394, SRR25337467–70) were mapped onto WT reference genome CP061854–60 (Baseggio et al., 2021) to detect small variants and regions of no read coverage indicating putative large deletions. Except for two de novo substitutions in one of the ss-minicircle transformation meroploid clones, all small mutations were variations already existing within the original cell population that increased in frequency (Tables 5 and 6). Small mutations appeared to be due to random replication errors, unlikely to have functional significance, and were mostly copy-number variations within repetitive sequence regions. In contrast, several de novo large deletions were found in KO and revertant clones selected on sucrose plates (Tables 5 and 7). They were evident as areas of no read coverage such as in the case of intended aip56 deletion (Figure 6), and were subsequently confirmed via long-read Nanopore sequencing (SRR253160-66; JAUY(VA-D, UZ)000000000). Five out of six sucrose-selected clones contained one large deletion, which was unique and not linked to broth culture history, that is, different variants present in KO and revertant clones selected in parallel from the same culture (Tables 5 and 7). These five deletions were flanked by transposable elements (TE), which are recognized as primary mediators of evolution in Pdp. Indeed, TE movement is promoted by stress (Twiss et al., 2005) and 15% sucrose in the medium creates high-osmolarity environment stressful to bacteria (Cesar et al., 2020). On the other hand, the KO clone originally selected on Km/sucrose agar did not have any large deletion, which may be incidental or somehow linked to the presence of Km in the medium. Unlike the small variants, the identified large deletions ranging from 8.5 to 25.2 kb were likely to be phenotypically significant. In particular, two sucrose-selected aip56 KO clones, QMA0784 and QMA0785, contained deletions of the same genomic region except for a 3 kb difference. These were acquired independently as they were absent in both QMA0786 meroploid parent strain and two “sister” revertant clones QMA0781 and QMA0782 (Tables 5 and 7). Such independent deletions of the 22/25 kb sequence, which includes several known important genes happening twice, suggest that this change may be favored in aip56-deficient mutants cultured on agar with high sucrose concentrations.

Table 6. Small secondary mutations identified in aip56 mutagenesis clones.

Mutation	Reference nucleotides	CDS/location	Protein effect	Strains	De novo
(Tandem repeat) deletion	C1	16S rRNA	-	QMA0787	No
(GG)3 > (GG)2	25,133–25,134
(Tandem repeat) deletion	C1	Group II intron reverse transcriptase/maturase	Frameshift	QMA0780	No
(AA)3 > (AA)2	96,594–96,595
Substitution	C1	IS91 family transposase	Substitution	QMA0788	Yes
A > T	149,093		D > E
Substitution	C1	IS91 family transposase	Substitution	QMA0788	Yes
G > T	149,097		P > H
(Tandem repeat) deletion	C1	Outside CDS	-	QMA0648	No
(ACTCGCTT)12 > (ACTCGCTT)10	760,176–760,191			QMA0780
				QMA0785
				QMA0789
(Tandem repeat) insertion	C1	Outside CDS		QMA0781	No
(TCGCCAC)9 > (TCGCCAC)10	846,596			QMA0784
Deletion	C1	Hypothetical protein	Frameshift	QMA0781	No
−TC	858,221–858,222
(Tandem repeat) insertion	C1	Bifunctional methylenetetrahydrofolate dehydrogenase/methenyltetrahydrofolate cyclohydrolase FolD	Truncation	QMA0784	No
(TCGCTAC)12 > (TCGCTAC)14	864,195			QMA0787
				QMA0789
(Tandem repeat) insertion	C1	Outside CDS	-	QMA0782	No
(GAGGTTTC)9 > (GAGGTTTC)10	2,529,776			QMA0784
				QMA0788–89
Substitution/deletion	C2	Outside CDS	-	QMA0648	No
CT > G−	744,231–744,232			QMA0782–88

• Abbreviations: CDS, protein coding sequence; rRNA, ribosomal RNA.

Table 7. Large deletions identified in aip56 mutagenesis clones.

Mutation	Reference nucleotides	CDS within product/locus tag		Strains
		IC627_21610	Hypothetical protein
Deletion 1	C2	IC627_21615	IS1-like element ISPda1 family transposase	QMA0780
		IC627_21620	IS3 family transposase
11, 287 bp	1,109,902–1,121,188	IC627_21625	IS3 family transposase	(revertant, culture1)
		IC627_21630	Site-specific integrase
		IC627_21635	Phosphoribosyltransferase
		IC627_21640	DNA-protecting protein DprA
		IC627_21645	Recombinase family protein
		IC627_21650	DUF3404 domain-containing protein
		IC627_17620	IS1-like element ISPda1 family transposase
Deletion 2	C2	IC627_17625	Type VI secretion system tip protein VgrG	QMA0781
		IC627_17630	IS91-like element ISPda2 family transposase
26, 434 bp	368,062–394,495	IC627_17635	Hypothetical protein	(revertant, culture2)
		IC627_17640	Hcp family type VI secretion system effector
		IC627_17645	Penicillin-binding protein 1B
		IC627_17650	Exoribonuclease R
		IC627_17655	Molybdopterin adenylyltransferase
		IC627_17660	Isoprenylcysteine carboxylmethyltransferase family protein
		IC627_17665	M48 family metallopeptidase
		IC627_17670	Hypothetical protein
		IC627_17675	IS1-like element ISPda1 family transposase
		IC627_17680	Fic family protein
		IC627_17685	IS1-like element ISPda1 family transposase
		IC627_17690	Hypothetical protein
		IC627_17695	DUF3010 family protein
		IC627_17700	DEAD/DEAH box helicase
		IC627_17705	Hypothetical protein
		IC627_17710	Leucine-rich repeat domain-containing protein
		IC627_17715	DUF3820 family protein
		IC627_17720	NUDIX domain-containing protein
		IC627_17725	Galactosyl transferase
		IC627_17730	Cation:proton antiporter
		IC627_17735	NAD(P)H-dependent oxidoreductase
		IC627_17740	Iron-containing alcohol dehydrogenase
		IC627_17745	Gfo/Idh/MocA family oxidoreductase
		IC627_17750	Hypothetical protein
		IC627_17755	tRNA (pseudouridine(54)-N(1))-methyltransferase TrmY
		IC627_17760	IS1-like element ISPda1 family transposase
		IC627_17765	DUF2955 domain-containing protein
		IC627_17770	HlyD family secretion protein
		IC627_20615	NAD(P)H-hydrate epimerase
Deletion 3	C2	IC627_20620	Hypothetical protein	QMA0783
		IC627_20625	hypothetical protein
8, 547 bp	927,902–936,451	IC627_20630	Hypothetical protein	(KO, culture 1)
		IC627_20635	IS1 family transposase
		IC627_20640	IS1-like element ISPda1 family transposase
		IC627_20645	SCO family protein
		IC627_20650	DUF368 domain-containing protein
		IC627_20655	IS1 family transposase
		IC627_20660	IS1 family transposase
		IC627_20665	IS1-like element ISPda1 family transposase
		IC627_20670	Hypothetical protein
		IC627_20675	Glyoxalase
		IC627_20680	IS1 family transposase
		IC627_05605	IS1 family transposase
Deletion 4	C1	IC627_05610	Hypothetical protein	QMA0784
		IC627_05615	DUF2947 domain-containing protein
25, 214 bp	1,148,799–1,174,012	IC627_05620	NAD-dependent succinate-semialdehyde dehydrogenase	(KO, culture 2)
		IC627_05625	NUDIX domain-containing protein
		IC627_05630	IS1-like element ISPda1 family transposase
		IC627_05635	Phosphatase PAP2 family protein
		IC627_05640	FAD-binding oxidoreductase
		IC627_05645	Aquaporin Z
		IC627_05650	LysR family transcriptional regulator
		IC627_05655	Cation transporter
		IC627_05660	NYN domain-containing protein
		IC627_05665	CG2 omega domain protein
		IC627_05670	YggN family protein
		IC627_05675	Hypothetical protein
		IC627_05680	Phospholipase A
		IC627_05685	YgiQ family radical SAM protein
		IC627_05690	IS1-like element ISPda1 family transposase
		IC627_05695	YgiQ family radical SAM protein
		IC627_05700	AAA family ATPase
		IC627_05705	IS1-like element ISPda1 family transposase
		IC627_05710	IS1-like element ISPda1 family transposase
		IC627_05715	IS1-like element ISPda1 family transposase
		IC627_05720	IS1-like element ISPda1 family transposase
		IC627_05725	Hypothetical protein
		IC627_05730	IS1-like element ISPda1 family transposase
		IC627_05735	Hypothetical protein
		IC627_05740	Hypothetical protein
		IC627_05745	IS1-like element ISPda1 family transposase
		IC627_05615	DUF2947 domain-containing protein
Deletion 5	C1	IC627_05620	NAD-dependent succinate-semialdehyde dehydrogenase	QMA0785
22, 213 bp	1,149,105–1,171,327	IC627_05625	NUDIX domain-containing protein	(KO, culture 3)
		IC627_05630	IS1-like element ISPda1 family transposase
		IC627_05635	Phosphatase PAP2 family protein
		IC627_05640	FAD-binding oxidoreductase
		IC627_05645	Aquaporin Z
		IC627_05650	LysR family transcriptional regulator
		IC627_05655	Cation transporter
		IC627_05660	NYN domain-containing protein
		IC627_05665	CG2 omega domain protein
		IC627_05670	YggN family protein
		IC627_05675	Hypothetical protein
		IC627_05680	phospholipase A
		IC627_05685	YgiQ family radical SAM protein
		IC627_05690	IS1-like element ISPda1 family transposase
		IC627_05695	YgiQ family radical SAM protein
		IC627_05700	AAA family ATPase
		IC627_05705	IS1-like element ISPda1 family transposase
		IC627_05710	IS1-like element ISPda1 family transposase
		IC627_05715	IS1-like element ISPda1 family transposase
		IC627_05720	IS1-like element ISPda1 family transposase

• Abbreviations: CDS, protein coding sequence; KO, knockout.

Thus, it appears that prolonged exposure to high sucrose concentrations used for sacB counterselection might have created stressful conditions, which activated TE movement in Pdp. This indicates that alternative negative selection markers should be considered for TE-rich bacteria, and may potentially allow to obtain a viable ssrA-smpB mutant in Pdp. In contrast, preparation of the competent cells using high sucrose buffer and electroporation and integration of the allelic exchange DNA in the form of the minicircles had virtually no mutagenic effect (even when minicircles were delivered as ssDNA).