The strains and plasmids used in this study are listed in Table 1. The pBASE_Bha vector (pFW004) is a derivative of the pBASE6 vector designed for use in S. aureus (Geiger et al., 2012) but adapted to H. halodurans by replacing the Staphylococcus-specific secY antisense sequence with one specific for targeting the secY mRNA of H. halodurans (Supplementary file S1). See section 2.6 and Table 2 for cloning strategy. pBASE_Bha is the “empty vector” not containing an insert in its multiple cloning site. All further pBASE_Bha vectors used in this study [pBASE_Bha_ΔyqeY (pFW006) and pBASE_Bha_rpsU** (pFW032)] are derivatives of pBASE_Bha. Escherichia coli NEB5α (#C2987, New England Biolabs, Ipswich, MA, United States) transformed with pBASE_Bha-based vectors were grown in Luria Bertani (LB) Broth (J75852-A1, Thermo Fisher Scientific, Fair Lawn, NJ, United States) medium supplemented with carbenicillin (Carb, #C1389, Sigma-Aldrich, St. Louis, MO, United States) at a final concentration of 100 μg mL⁻¹ for plasmid isolation. E. coli TOP10 + pPAMC125 cells transformed with pBASE_Bha-based vectors for in vivo methylation were grown in LB medium supplemented with chloramphenicol (Cm, #C0378, Sigma-Aldrich) at a final concentration of 5 μg mL⁻¹ in addition to Carb for plasmid isolation. A 3 mg mL⁻¹ working stock of Cm was established in ≥99.8% ethanol from a 100 mg mL⁻¹ stock (also in ≥99.8% ethanol). If not stated otherwise, H. halodurans was grown in LB Broth medium supplemented with Na₂CO₃ to a final concentration of 94.4 mM to obtain a pH of 10. The LB broth medium was set up with ddH₂O to 90% of the final volume, autoclaved and then supplemented with sterile-filtered Na₂CO₃ solution to the final volume. If not stated otherwise, E. coli and H. halodurans were cultivated at 37°C on solid media and with shaking at 220 rpm in liquid media.

For the allelic replacement process, H. halodurans was grown on LB agar (#113002232, MP Biomedicals, Santa Ana, CA, United States) or in LB media at different pH values noted in section 3.2. Either Na₂CO₃ was added to a final concentration of 94.4 mM (for pH 10), 100 mM NaHCO₃ (for pH 8.5 LB liquid media) or 50 mM NaHCO₃ and 50 mM NaCl (for pH 8.5 LB agar, which avoids NaHCO₃ crystallization). Sterile-filtered solutions of Na₂CO₃ (944 mM) or NaHCO₃ (1 M) were added to the autoclaved, volume-adjusted media and agar. Media and agar were supplemented with Cm (final concentration of 3 μg mL⁻¹) or anhydrotetracycline (ATc, #10009542, Cayman Chemical Company, Ann Arbor, MI, United States) (final concentration of 100 ng mL⁻¹) as detailed elsewhere. A 100 μg mL⁻¹ ATc working stock in 70% ethanol was prepared from a 2 mg mL⁻¹ stock (also in 70% ethanol) and stored at −20°C with protection from light.

Plasmid DNA for in vitro methylation or ensuing transformation into E. coli + pPAMC125 was isolated from E. coli NEB5α according to the manufacturer’s instructions using the QIAprep® Spin Miniprep Kit (#27106, Qiagen, Hilden, Germany). For in vivo methylation of the plasmid DNA, the construct was transformed into chemically competent bacteria of the engineered E. coli + pPAMC125 strain (Wallace and Breaker, 2011) using standard heat shock procedures. Plasmid DNA was subsequently isolated as described above. To determine the approximate concentration of in vivo methylated plasmid in the preparation, the DNA concentration determined by A₂₆₀ spectrophotometric measurement was divided by two, assuming that the ratio of the pPAMC125 and the plasmid of interest is 1:1 in the plasmid preparation.

HaeIII methyltransferase of Haemophilus aegyptius (#M0224S, New England Biolabs) was used for in vitro methylation of plasmid DNA as described in Figure 2A and in the text. Glycogen, RNA grade (#R0551, Thermo Fisher Scientific) was used to increase plasmid DNA precipitation.

Protoplast transformations of H. halodurans to introduce plasmids were conducted as described in Wallace and Breaker (2011). Briefly, exponentially growing H. halodurans cells cultured in LB (pH 10) were harvested from 6 mL at 3,000 rcf for 10 min, the cell pellet was resuspended in 0.1 volumes of ASMMP buffer [1 volume of 4× BD Difco^™ Dehydrated Culture Media: Antibiotic Medium 3, pH 7.5 (#DF0243-17-8, Thermo Fisher Scientific) and 1 volume of 2× SMM solution [1 M sucrose, 0.04 M maleate, 0.04 M MgCl₂ (pH 6.5)]] as described elsewhere (Chang and Cohen, 1979; Kudo et al., 1990). Then cells were treated with 5 mg mL⁻¹ lysozyme from chicken egg white (#4403, MilliporeSigma, Burlington, MA, United States) for 1 h at 37°C with agitation at 100 rpm, pelleted at 2,600 rcf for 10 min, once washed with ASMMP buffer and resuspended in ASMMP buffer to an optical density (OD) at 600 nm of 10. 500–1,000 ng methylated plasmid DNA was mixed with an equal volume of 2× SMM solution (see above), then 150 μL of the protoplast preparation was added to the plasmid DNA mix, incubated for 1 min at room temperature. Following, the protoplast plasmid DNA mix was transferred into 450 μL of a 30% (w/v) polyethylene glycol 8000 (PEG) (#PR-V3011, Thermo Fisher Scientific) solution (PEG dissolved in 1× SMM buffer), gently mixed by inversion and incubated for 2 min at room temperature. Then, 1.3 mL ASMMP buffer was added and gently mixed by inversion. Protoplasts were pelleted at 2,600 rcf for 10 min, very gently resuspended in 300 μL ASMMP buffer and incubated for 90 min at 30°C with agitation at 100 rpm. After recovery, protoplasts were pelleted in a table-top centrifuge for 1 min, 150 μL of the supernatant was discarded and cells were gently resuspended in the remaining volume. Subsequently, the suspension was carefully plated on M-NB regeneration medium agar (1 volume 2× Nutrient Agar, 1 volume 1 M sodium succinate, pH 7.8). 50 mL of 2× Nutrient Agar contained 0.5 g peptone (#211910, Polypeptone^™ Peptone), 0.3 g beef extract (#211520, Bacto™ Beef Extract, desiccated) and 1.5 g agar (#DF0140-01-0, BD Bacto^™ Dehydrated Agar) (all purchased from Thermo Fisher Scientific). After autoclaving of the 2× Nutrient Agar an equal volume of the 1 M sodium succinate solution was added and agar mix was supplemented with Cm (final concentration of 3 μg mL⁻¹). Agar plates were incubated at 30°C for ≥2 days until colonies appeared. Colony PCRs to check for successful transformation using primers pBASE_MCS_F and pBASE_MCS_R primers (see Table 2 for all oligonucleotides) were conducted using the DreamTaq Green PCR Master Mix (2×) (#K1082, Thermo Fisher Scientific) as recommended by the manufacturer. Colony PCRs were of poor quality or failed with other commercially available PCR kits.

For generation of plasmids pFW004-006 and pFW032 (Table 1) “in vivo E. coli cloning” (iVEC)’ was conducted as described previously (Wencker et al., 2021) using a previously reported strategy (Nozaki and Niki, 2019). In brief, DNA fragments overlapping by ~25–35 bp were generated by PCR, DpnI-treated (#FD1703, Thermo Fisher Scientific), purified using a QIAquick^® PCR Purification Kit (#28106, Qiagen) and transformed into chemically competent E. coli NEB5α cells in the presence of 1× Quick Ligation Reaction Buffer (part of M2200S, B2200SVIAL, New England Biolabs) following a standard heat-shock protocol. For example, PCRs to generate pBASE_Bha (pFW004) were conducted by (i) amplifying the pBASE6 vector as two fragments overlapping in the AmpR cassette and with overlaps to the Bha antisense secY sequence and (ii) amplifying the Bha antisense secY sequence from genomic DNA of H. halodurans.

Colony PCRs to check for successful chromosomal modification (either deletion of yqeY or introduction of point mutations into rpsU) using appropriate primers (Table 2) were performed according to standard procedures. Primers should be targeted to bind approximately 20 to 40 bp up- and downstream of the ~1 kb flanking regions to ensure amplification arises only from genomic sequences and not from the plasmid.

To reduce both the time and effort needed to genetically manipulate H. halodurans C-125, we revised the protoplast transformation method previously established in our laboratory (Wallace and Breaker, 2011). A key step in this protocol is the in vivo methylation of the plasmid of interest in an engineered E. coli strain prior to transformation. This step overcomes the ability of H. halodurans to degrade a foreign plasmid via DNA restriction enzymes naturally encoded in its genome. The engineered E. coli strain harbors the pPAMC125 vector with genes for two DNA adenine-specific methyltransferases (BH4003 and BH4004) and a gene for HaeIII DNA 5-cytosine methyltransferase (BH3508). These enzymes methylate plasmids passaged through the strain, which enables protection from H. halodurans C-125 restriction endonucleases upon transformation. This in vivo methylation step improves transformation efficiency into H. halodurans C-125 by up to 1,000-fold relative to untreated plasmids isolated from standard E. coli cloning strains (Wallace and Breaker, 2011). However, in vivo methylation requires additional time and work to transform the plasmid into the E. coli + pPAMC125 strain and subsequent plasmid isolation, which takes ~3 days.

An important DNA methyltransferase for in vivo methylation is a type II DNA 5-cytosine methyltransferase that is encoded by the prophage gene BH3508 (M.BhaII, AYT26_RS17595) (Roberts et al., 2003). BH3508 and its corresponding restriction endonuclease BH3509 are embedded in a prophage within the H. halodurans C-125 genome and are absent in other H. halodurans strains such as the NCBI reference isolate LB-1 or the strain DSM 497 (protein BLAST analysis, data not shown). These methyltransferases modify the internal cytosine residue at the C5 position of the sequence GGCC. The homolog HaeIII methyltransferase of Haemophilus aegyptius is commercially available (Slatko et al., 1988). Therefore, we reasoned that purified HaeIII methyltransferase could be exploited to prepare methylated plasmid DNA in vitro for subsequent transformation into H. halodurans with high efficiency. Previous studies in, e.g., Helicobacter pylori and Streptomyces griseus successfully employed in vitro methylation of plasmids to increase transformation efficiency (Donahue et al., 2000; Kwak et al., 2002). To demonstrate this method for H. halodurans with the HaeIII methyltransferase, in vitro methylation reactions were established with 1–2 μg plasmid DNA following the manufacturer’s instructions for methylation of genomic DNA (Figure 2A). For comparison of transformation efficiencies, we transformed the in vivo and in vitro methylated pBASE_Bha vector in parallel into H. halodurans using the protoplast transformation method (Wallace and Breaker, 2011). Although, transformation with the in vitro methylated plasmid is about four times less efficient than with the in vivo methylated vector (Figure 2B), it commonly yielded nearly 100 transformants per μg of plasmid DNA, and no clones were observed to be false positives (data not shown). Thus, in vitro methylation of plasmids allows for substantial transformation success with H. halodurans in less than 1 day compared to 3 days for the in vivo methylation process.

Next, we modified the allelic replacement procedure previously used with S. aureus to generate genomic alterations in H. halodurans. The DNA construct used is an E. coli/Gram-positive shuttle vector with a temperature-sensitive origin of replication (pE194ts), a chloramphenicol resistance cassette (cat gene), and a “payload” sequence designated to replace the chromosomal sequence. This payload region carries ~1 kb of sequences on each flank of the deletion, insertion, or mutation site. This extensive homology between the plasmid and the target site for genomic modification promotes HR at the desired location. The initial recombination event is expected to integrate the plasmid into the genome at the desired location. A second recombination event then either leads to the (allelic) replacement of the original targeted DNA region with the plasmid-derived sequence or restores the original genomic wild-type sequence and removes the plasmid sequence.

The vector also contains an antisense sequence of the essential gene secY. Transcription of this “secY as” RNA is regulated by an anhydrotetracycline (ATc)-inducible promotor, and its production enables counter-selection against bacteria still containing the plasmid. Counter-selection is necessary to isolate transformants that have eliminated the plasmid vector. Isolation of plasmid-free cells by exposure to ATc occurs because the antisense RNA forms a duplex with the secY mRNA and prevents its translation. SecY is a membrane protein and a component of the SecYEG translocase necessary for protein secretion (Oswald et al., 2021). Protein secretion, and therefore secY expression, is essential for bacterial growth and survival (Ji et al., 2001).

If the desired genetic changes occur, the plasmid will have integrated into the genome at the locus of interest via homologous recombination at the non-permissive temperature for plasmid replication, i.e., 43°C. Subsequently, cultivation of the bacteria in the presence of ATc allows for the selection of bacteria that have undergone a second homologous recombination event to eliminate the non-payload portion of the vector from the chromosome (Figure 1). Unless the portion of the plasmid carrying the secY antisense RNA gene is eliminated or mutated, ATc treatment will be lethal to the cell. Importantly, to adapt the S. aureus pBASE6 vector for use in H. halodurans, we replaced the Staphylococcus-specific secY antisense sequence with one specific for targeting the secY mRNA of H. halodurans to create pBASE_Bha (Figure 3A; Supplementary file S1). The sequence spans the first half of secY (BH0154, AYT26_RS00900) and includes its Shine-Dalgarno (SD) region (Figure 3B).

In addition to preparing the plasmid for transformation into H. halodurans, the allelic replacement procedure was optimized for the high pH media preference of this species. With pH 10 growth media, 50 ng mL⁻¹ ATc did not sufficiently induce secY antisense transcription to permit counter-selection. Our initial attempt to delete a gene (yqeY [BH1355, AYT26_RS07050]) from the genome of H. halodurans using the allelic replacement method with inducible counter-selection was inefficient. Of 95 colonies that emerged after counter-selection (step 5, Figure 3C) on LB pH 10 agar supplemented with 50 ng mL⁻¹ ATc, only two were chloramphenicol-sensitive (~2%) indicating that the remaining 93 still contained the pBASE_Bha-based vector and had not undergone the second homologous recombination event. Only one of the two chloramphenicol-sensitive candidates contained the desired deletion (Table 3). From previous experience with the counter-selection method in S. aureus, this high number of failed counter-selections was unexpected, and a higher number of initial candidates was expected. We speculated that uptake of active ATc was insufficient at pH 10, perhaps because it is insufficiently stable (recommended pH for storage is 7.2). To compensate, ATc concentration was increased to 100 ng mL⁻¹ and pH 8.5 growth medium was used. These conditions permitted robust bacterial growth while yielding sufficient secY antisense transcript production to enable counter-selection. We attempted to obtain single chromosomal deletions for 10 additional genes with the adjusted pH and ATc concentration. The efficiency of counter-selection now ranged from ~6 to 57%, with the majority of attempts (8 out of 10) yielding in at least 20% of chloramphenicol-sensitive clones (Table 3).

The complete workflow for the allelic replacement procedure is briefly outlined below (Figure 3C), with procedural recommendations and possible modifications highlighted. All steps involving ATc should be performed with minimal light exposure to prevent degradation of photosensitive ATc (Baumschlager et al., 2020). LB (pH 8.5) + ATc (100 ng mL⁻¹) plates should be prepared fresh by spreading the required volume of ATc (delivered in sterile phosphate-buffered saline [PBS]) on the agar. When step 4 is performed with a short incubation (~10 h), plates for step 5 should be incubated at 37°C instead of 30°C to ensure sufficient colony size the next day. For screening on agar supplemented with chloramphenicol in step 6, it is recommended that the incubation proceeds at least overnight at 37°C to exclude false negative results (no growth on Cm despite presence of vector) due to potentially slightly delayed growth in the presence of the antibiotic.

Notably, the number of candidates per screening (clones that only grow on plain agar) as well as the ratio of successfully chromosomally modified clones to clones with the wild-type sequence are highly variable and seem to depend on the target sequence. Of candidates which underwent the second homologous recombination event and have eliminated the plasmid from the chromosome, 4–100% had the desired knockout, depending on the gene to be deleted. The remaining clones were wild type for the targeted gene (Table 3). We note that the theoretical likelihood either to obtain the desired mutation or to restore the wild-type allele via the second homologous recombination event is 50% each, due to the nature of the procedure. Therefore, ensuing screening via PCR is required (step 7, Figure 3C). We present two examples using the above-described methods to illustrate the successful deletion of H. halodurans yqeY (BH1355, AYT26_RS07050) and insertion of point mutations into rpsU (BH1354, AYT26_RS07045) to introduce two early stop codons to functionally disrupt RpsU (ribosomal protein S21) (Figure 4). Successful deletion of the yqeY gene was confirmed by colony PCR analysis (Figure 4B) whereas successful introduction of mutations into the rpsU gene was confirmed by ensuing sequencing of the PCR products (Figure 4D).