A streptomycin resistance marker in H. parasuis based on site-directed mutations in rpsL gene to perform unmarked in-frame mutations and to verify natural transformation

Bacteria strains, plasmids and culture conditions

The bacteria strains, plasmids and primers used in this study are listed in Tables S1 and S2. Plasmids were propagated in E. coli DH5α or S17-1 (pir) and bacteria were grown in liquid Lysogeny Broth (LB, Difco, USA) medium or on LB agar (Invitrogen, Shanghai, China) plate. When required, the medium was supplemented with kanamycin (Kan; 50 µg ml⁻¹) or ampicillin (Amp; 100 µg ml⁻¹) from Sigma-Aldrich, USA. H. parasuis was routinely cultured in Tryptic Soy Broth (TSB; Difco, Franklin Lakes, NJ, USA) or on Tryptic Soy agar (TSA; Difco, Franklin Lakes, NJ, USA) supplemented with 5% inactivated bovine serum (Solarbio, Beijing, China) and 0.1% (w/v) nicotinamide adenine dinucleotide (NAD; Sigma-Aldrich, St. Louis, MO, USA) (TSB++ and TSA++). Where necessary, the media were supplemented with streptomycin (SM; 25 µg ml⁻¹) or kanamycin (Kan; 50 µg ml⁻¹). Unless otherwise stated, all strains were grown at 37 °C.

DNA manipulations

Bacterial genomic DNA was extracted using a TIANamp Bacteria DNA Kit (TIANGEN, Beijing, China). Small-scale plasmid DNA preparations were generated using a E.Z.N.A™ Plasmid Miniprep Kit (Omega Bio-Tek, Norcross, GA, USA). The DNA fragments were amplified in a C1000™ Thermal Cycler using the 2×Taq Plus Master Mix (Vazyme, Jiangsu Sheng, China) or I-5^TM 2 ×High Fidelity Master Mix (MCLAB, South San Francisco, CA, USA). Purification of DNA fragments from the PCR reaction and the restriction digests were performed using the DNA Fragment Purification Kit Ver.4.0 (TaKaRa, Shiga, Japan). In-fusion segments were recombined to linearized pK18mobsacB (EcoRI/BamHI) using a ClonExpress II One Step Cloning Kit (Vazyme, Jiangsu Sheng, China). Restriction enzymes were purchased from TaKaRa. Point mutations were performed according to manual of Fast mutagenesis system Kit (TransGen Biotech, Beijing, China).

Antibiotic and Ethyl Methane Sulfonate (EMS) susceptibility assays

Assays on the minimal inhibitory concentration were carried out as described by Li (Li et al., 2016b). Briefly, overnight-grown bacteria were grown in TSB++ with twofold serially diluted SM (from 2 µg/mL to 128 µg/mL; MIC-S) in a 96-well microtiter plate. The plates were incubated at 37 °C for 24 h, after which, the results were scored for growth or no growth. The experiments were independently performed at least three times in triplicate. The MIC of EMS (from 0.5 mM to 40 mM; MIC-E) was analyzed in sterilized tubes. The results were also scored for growth or no growth. Moreover, the MIC-S of EMS induced SC1401 derivatives were also determined using the above method.

The minimal bactericidal concentrations of EMS for H. parasuis (MBC-E) were determined using the method modified from Li and Zhang (Zhang & Mah, 2008; Li et al., 2016b). Overnight-grown H. parasuis in sterilized tubes were exposed to different concentration of EMS (from 0.5 mM to 40 mM) for at least 12 h. Then, fresh TSB++ was used to replace the EMS medium and incubated for additional 16 h. The viability of bacteria was assessed by transferring a spot of the culture onto a TSA++ plate and incubation for 16 h. The experiments were independently performed at least three times in triplicate. Bacteria were serially passaged in TSB++ with MIC-E of EMS for five more generations for further determining the optimum inducing concentration of EMS.

In addition, the Oxford cup assay was used to evaluate the streptomycin resistance of wildtype H. parasuis SC1401, Hps32, Hps33 and derivatives 1401D88, 1401D43. Inocula were prepared and spread uniformly on TSA++ plates. One hundred microlitres of different concentrations (8,192 µg ml⁻¹ and 4,096 µg ml⁻¹ per well) of SM were added to the Oxford cups which were placed at equal distances above agar surfaces. The zone of inhibition for each concentration was measured after an extra 16 h-incubation at 37 °C. The experiments were independently performed at least three times in triplicate (Shang et al., 2014).

Ethyl Methane Sulfonate discontinuously inducing method

Wild type SC1401 (whose genomic sequence has been deposited in GenBank under the accession number NZ_CP015099.1), which was SM sensitive and used as the parental strain, was cultured in TSB++ with an optimum concentration (10 mM) of EMS (AR, Biotopped, China) at 37 °C with agitation for 12 h. The culture (F1) was transferred into 5 mL of fresh TSB++ without EMS, in a “Healing Resurgence” step. The F2 was subcultured and induced by 10 mM of EMS again with the same method for F1, F2, F3, F4 generations. Each generation was plated onto TSA++ with appropriate SM antibiotic and incubated for 24 h. The visible single colonies were propagated and the DNA extracted for further identification.

Sequence analysis

The 375-bp sequence of rpsL (locus tag: A4U84_RS04600) were amplified from the visible single colonies’ genomic DNA with primers rpsL-P1/P2 (Table S2). The PCR products were then purified. The products were cloned to pMD-19T(simple) for sequencing by BGI, Beijing, China. The RpsL protein sequences of H. parasuis were obtained by translating gene sequences using DNAMAN V6. Gene and protein sequences were then analyzed using BLAST in a comparison with that of wild type SC1401.

Site-directed mutagenesis

To further delineate mutations in rpsL, rather than other random mutations in whole genome, principally contributing to SM resistance induced. The entire rpsL coding sequence and corresponding flanking regions (975 bp) were amplified from genomic DNA of wild type SC1401 using primers rpsL-PM-P1/P2 and subsequently cloned into suicide vector pK18mobSacB, generating wild type pkrpsL. Afterwards, rpsL-43-P1/P2 and rpsL-88-P1/P2 were used to amplify the whole pkrpsL backbone with respective point mutations at codon 43rd (A128G) and 88th (A263G) in ORF of rpsL with 2 ×Transtart Fastpfu PCR SuperMix. Then the products were digested by DMT enzyme to move methylated template plasmids. Plasmids were subsequently mobilized into DMT competent cells by the CaCl₂ method. The mutants were confirmed by sequencing.

The wild type pkrpsL (control), mutant plasmids pkrpsL43 (A128G) and pkrpsL88 (A263G) were transformed into wild type strain SC1401 by electroporation. The electroporation parameters were: 1800 v/mm, 25 µF, 200 Ω. The cells were screened by TSA++ containing 25 µg ml⁻¹ of SM (although additional 10% sucrose together with antibiotic for screening transformants may be advantageous in order to prevent illegitimate recombinations, we found that once we obtained SM-resistant transformants, nearly all of them were expected derivatives after direct sequencing, even without the inclusion of 10% sucrose in the medium.). The visible single colonies were propagated and the DNA extracted and sequenced for further identification.

Construction of plasmids pkTLR and pkALR

Plasmids pkTLR and pkALR were constructed for tfox and arcA gene deletions and counter-selection in H. parasuis 1401D88, respectively. Here, we use pkTLR as an example to describe the process (Fig. 1). The 750-bp upstream homologous arm region and the 767-bp down homologous arm region of tfox were amplified using primers tfoxL-P1/P2 and tfoxR-P1/P2 from the genome of SC1401 (or 1401D88), respectively. These two flanking fragments were then purified using Qiaquick spin column (Qiagen, Hilden, Germany) and integrated by overlap extension PCR with tfoxL-P1/tfoxR-P2, after which the in-fusion segment was recombined to linearized pK18mobsacB (EcoRI/BamHI) using ClonExpress II. One Step Cloning Kit (Vazyme, Jiangsu Sheng, China). After confirmation by PCR and sequencing, the resulting plasmid, pkTLR, was mobilized into S17-1 (λpir) by the CaCl₂ method.

Construction of unmarked in-frame targeted mutant 1401D88ΔtfoxΔarcA of H. parasuis 1401D88

An unmarked tfox mutant was constructed using a two-step selective method (Fig. 2; Fig. 3). DNA was mobilized from E. coli S17-1 (λpir) to H. parasuis using a filter mating technique adapted by Oswald with some modification (Oswald et al., 1999). Briefly, S17-1 (λpir) harboring pkTLR and 1401D88 strains were grown to exponential phase in SOC and TSB++, respectively. The bacteria cultures were centrifuged at 4,350×g for 5 min; the pellet was resuspended in 1mL of TNM-buffer (1 mM Tris–HCl pH 7.2, 10 mM MgSO₄, 100 mM NaCl). Aliquots corresponding to 0.5 ml of donor and 1 ml of recipient (each adjusted to an OD₆₀₀ = 1) were mixed and plated onto a nitrocellulose filter membrane (0.45 µm pore size, 2.5 cm diameter; Millipore, Darmstadt, Germany). The NC membrane with bacterial mixture was placed onto TSA++ plates and incubated at 30 °C for 12 h. Eventually, the cultures were scraped up and spread on selective TSA++ supplemented with 25 µg ml⁻¹ of SM and 50 µg ml⁻¹ of Kan and incubated at 37 °C for 24 h. The transformants were then examined for sucrose sensitivity by plating on TSA++ supplemented with 10% sucrose (UP; Amresco, Solon, OH, USA) and 25 µg ml⁻¹ of SM. PCR was carried out to confirm the appropriate deletion in colonies resistant to sucrose and sensitive to kanamycin. Afterwards, the mutants were continuously passed to 20 times on TSA++ and primers tfox-P1/P2, tfoxL-P1/tfoxR-P2 and Hps-P1/P2 were adopted for further confirming stability of the desired mutant.

The effectiveness of our novel protocol for generating a double-gene mutant was confirmed when we further deleted the arcA gene from 1401D88△tfox. Likewise, the process was repeated as above, using a different primer set. The upstream and downstream homologous arms regions of arcA were amplified using primers arcAL-P1/P2 and arcAR-P1/P2, respectively. Primers arcA-P1/P2, arcAL-P1/arcAR-P2 and Hps-P1/P2 were adopted for further confirming the stability of the desired mutant.

The whole extract of 1401D88△tfox△arcA was further analyzed by western blotting. Briefly, the whole-cell lysates of the 1401D88, 1401D88△tfox△arcA mutant strain and SC1401 strain were analyzed by 12% SDS-PAGE electrotransferred to a polyvinylidene fluoride membrane (PVDF membrane). After being activated in methanol for 30 s and blocked with 5% BSA in TBST at room temperature (RT) for 30 min to 1 h, the membrane was incubated at RT for 1 h or overnight with mouse antiserum against recombinant Tfox or ArcA as the primary antibody, respectively. The membrane was rinsed 3–5 times with no less than 25 mL of TBST between incubations. Horseradish peroxidase (HRP)-conjugated rabbit-anti-mouse IgG (Earth, Hong Kong) were used as the secondary antibody (1:10,000 dil.). The membrane was finally developed with Immun-Star Western C Kit (Biorad, USA) according to the manufacturer’s instructions (Ding et al., 2016).

Using genomic DNA of 1401D88 and 1401D43 to verify natural transformation

The genomic DNAs of H. parasuis SC1401 derivatives 1401D88 and 1401D43 via site-directed point mutations in rpsL gene (as indicated in ‘DNA manipulations’) that confer SM resistance to H. parasuis were used to verify natural transformation capacity of H. parasuis strain SC1401; genomic DNA of SC1401△htrA, HPS32, HPS33 and the plasmid DNA pKBHK were used as controls. The protocol for natural transformation was performed as previously described with some modifications (Bigas et al., 2005). Briefly, recipient bacteria were grown in TSB++ to log phase or to an OD₆₀₀ = 1.8 (about 2.8 ×10⁹ cfu/mL). The bacteria were then spotted on TSA++, cultured overnight at 37 °C and resuspended in TSB++ at 20 ×10¹⁰ cfu/mL. A 20 µL aliquot of suspension was supplemented with 1 µg of donor DNA. The mixture was incubated for 10 min at 37 °C and then spread in a small area on TSA++. Cultures were incubated for an additional five hours at 37 °C to induce expression of antibiotic resistance. Afterwards, bacteria were harvested and resuspended and plated on TSA++ with SM in concentrations ranging from 25 µg ml-1 to no antibiotic. Bacteria were incubated for 36 h.

After incubation, the colonies on TSA++ and on selective plates were counted. Transformation frequencies were determined from the number of antibiotic-resistant CFU mL⁻¹ divided by the total CFU mL⁻¹ scored on nonselective agar. Transformation efficiency was evaluated by calculating number of transformed CFU per µg of donor DNA.

Statistical analysis

The statistical analysis was performed using R studio (loaded with R V3.3.2) for Windows (RStudio Team, 2015). Comparison of several test series was evaluated by analysis of variance (ANOVA). The significance of differences between groups was calculated using Student’s t-test. A P value < 0.05 was considered to be statistically significant (*), and <0.01 highly significant (**).

Determination of MIC & MBC of streptomycin and ethyl methane sulfonate in H. parasuis

To investigate an appropriate inhibiting concentration of SM and EMS, we determined minimal inhibitory concentration of SM (MIC-S) and EMS (MIC-E) for H. parasuis SC1401. The results showed that MIC-S for SC1401 to be 16 µg ml⁻¹; however, 24 µg ml⁻¹ of SM was bactericidal (Table 1). Hence 25 µg ml⁻¹ of SM was employed in later experiments to screen out SM-resistant derivatives from wild type SC1401. By comparison, EMS was shown to have fairly strong antibacterial activity. As shown in Fig. 4, we could see flocculent bacteria when the culture was exposed to 10 mM of EMS, but not in the TSB++ with 20 mM of EMS. Therefore, twenty millimoles per liter of EMS was determined to be the MIC-E of H. parasuis SC1401 (Table 1 and Fig. 4).

To determine optimal inducing, subinhibitory concentrations of EMS, we tested the minimal bactericidal concentrations (MBC-E). Overnight-grown seed broth was subcultured in fresh TSB++ with different concentration of EMS (from 0.5 mM to 40 mM) and subcultured again with TSB++ with no EMS. The results showed that 30 mM of EMS was the definitive MBC-E (Table 1).

The EMS-induced derivatives of SC1401 and MIC-S assay

As shown in Table 1, thirty mili-moles per liter of EMS has strong bactericidal activity (this concentration was determined to be the MBC-E), whereas 10 mM of EMS was shown to be the MIC-E (Fig. 4 and Fig. S1). Therefore, in consideration of maximizing the inductive effect of EMS, we used a discontinuous induction method in our experiments. The culture in a 10 mM of EMS-containing TSB++ was transferred into 5 mL of fresh TSB++ without any EMS, in a “Healing Resurgence” step of our protocol. Each generation was plated onto TSA++ with appropriate SM antibiotic and incubated for 24 h. In this way, we obtained SM-resistant generations in F7, F9 and F11. Among them, mutations in rpsL at codon 43rd (A128G) and 88th (A263G) were confirmed by sequencing.

By comparison, these two kinds of mutations conferred complete resistance to SM (Table 1). The two derivatives can survive well at a concentration of at least 4,096 µg ml⁻¹ of SM. As a control, H. parasuis wild type strains SC1401, HPS32, HPS33 can only survive in less than 16 µg ml⁻¹ of SM/TSB++. We also performed the Oxford cup assay to visually demonstrate different levels of SM-resistance. As shown in Fig. 5, 1401D43 and 1401D88 can form complete bacteria lawns on TSA++ with both 100 µl of 8,192 µg ml⁻¹ and 4,096 µg ml⁻¹ of SM per well, while inhibition zones can be easily recognized for wild type H. parasuis, indicating the successfully artificially EMS-induced products were far more resistant to SM than wild type SC1401.

rpsL hot mutations

It has been well established in procaryotic organisms that streptomycin hinders the functioning of ribosome by binding to 16S rRNA helices and ribosomal protein S12, encoded by rrs and rpsL, respectively. SM also interacts with decoding site and shifts 16S rRNA helix in the direction of ribosomal protein S12 which in turn induces miscoding by stabilising the closed confirmation of 30S ribosomal subunit (Suriyanarayanan et al., 2016). We directly sequenced the rpsL and rrs genes for wild-type strains Hps32, Hps32 and artificially induced SM-resistant strains. We used BLAST and these reads to query the rpsL and rrs sequences of H. parasuis SC1401. The results of DNA sequencing showed that all these SM-resistant strains harbored rpsL mutations. Comparatively, point mutations at codon 88th (AAA → AGA; K88R) and at codon 43rd (AAA → AGA; K43R) confer much higher SM-resistance on the EMS-induced derivatives than H. parasuis wild type strains Hps32, Hps32, in which random mutations of synonymous codon have been introduced (Fig. 6), indicating that the mutations at codon 43rd and 88th are the core regions blocking SM binding in H. parasuis, whilst other point mutations occurred less frequently. The mostly highly-resistant SM mutations occur in these two codons. However, we haven’t found a derivative with simultaneous mutations in these two sites. By comparison, rrs mutation could be found in derivative 1401D43 (in which only G920A have been identified), HPS32 and HPS33 (in which almost fifteen point-mutations have been identified, data not shown), but obviously not a greater factor of inducing the high resistance to SM (Table 1 and Fig. 5). We didn’t find any mutations in rrs in 1401D88.

Identification of successful construction of pkTLR and pkALR

Further investigation of Kan-resistant colonies showed that the in-fusion segments of up/downstream regions of tfox was successfully integrated into linearized pK18mobsacB. PCR runs with primers tfoxL-P1/tfoxL-P2, tfoxR-P1/tfoxR-P2 and tfoxL-P1/tfoxR-P2 demonstrated that the products were amplified as expected, which was followed by sequencing for further confirmation (Fig. S2). The plasmid pkTLR was then used to create unmarked in-frame tfox mutations in H. parasuis. Likewise, the plasmid pkALR was generated and identified as above. PCR runs were carried out using primers arcAL-P1/arcAL-P2, arcAR-P1/arcAR-P2 and arcAL-P1/arcAR-P2.

Construction of an unmarked mutation of tfox and arcA genes in H. parasuis 1401D88

Based on a two-step natural transformation mutagenesis method, we used a ‘dual’ process with one round of SM selection and a subsequent sucrose counterselection. The in-fusion segment was successfully transferred to the chromosomal DNA of recipient strain and subsequently integrated into the host tfox loci through homologous recombination, yielding the in situ deletion. Here, the method of EMS induced high level-resistance to SM plays an important role in weeding out donor strain S17-1 (λpir), thus allowing us to generate unmarked mutants in H. parasuis that are multi-drug susceptible.

The plasmid pkTLR was integrated into the two flanking regions of tfox gene at the genomic site, so that the original tfox gene was replaced seamlessly. SM and Kan positive selection was highly successful, and the PCR tests validated both the insertion of the kan cassette and SacB gene in most of the Kan-resistant clones (Fig. S3A). In the second round of homologous recombination, the backbone of pk18mobSacB was removed, followed by sucrose counterselection and kanamycin susceptibility testing. PCRs were carried out to verify the appropriate deletion in colonies resistant to 10% sucrose and 25 µg ml⁻¹ of SM but sensitive to Kan (Fig. S3B).

The second round is to remove arcA gene from 1401D88△tfox. Likewise, pkALR was used to create in-frame arcA knock-out mutant. PCR runs were carried out to confirm the appropriate deletion in transformants resistant to sucrose and 25 µg ml⁻¹ of SM but sensitive to 20 µg/mL Kan.

Confirmation of 1401D88△tfox△arcA by western blotting

Western blotting confirmed that tfoX and arcA can be detected in the whole-cell extract of wild type H. parasuis SC1401 and 1401D88, but not in 1401D88△tfox△arcA (Fig. 7). This result further confirms the successful deletion of tfox and arcA genes in strain 1401D88.

Verification of natural competence

Point mutations in rpsL which don’t confer negative effect on growth were assumed to be more effective in verifying natural competence (see Fig. S4 for comparison of growth rates of H. parasuis mutants and wild type SC1401). To confirm this hypothesis, we tested the natural frequencies of different donor DNA, including genomic DNA of H. parasuis 1401D43, 1401D88, SC1401△htrA (in which the Kan cassette had been stably inserted.) and the plasmid DNA pKBHK which was previously used to creat htrA knock-out mutant in H. parasuis and to screen natural competent cells, as well as the naturally SM-resistant strains HPS32 and HPS33. The results showed that transformation frequencies of 1401D43, 1401D88, SC1401△htrA genomic DNA after an extended growth period (13 h) on TSA++ were considerably higher than those for pKBHK construct. Genomic DNA of HPS32 and HPS32 couldn’t confer SM-resistance to SC1401, which also in support of the elucidation highlighted in the “Introduction” that spontaneous mutations in this species occur at a fairly low level. The transformation frequencies of derivative genomic DNA reached a fairly high level (about 2.458 × 10⁻³) compared to SC1401△htrA genomic DNA. However, the highest transformation frequency of genomic DNA of SC1401△htrA reached to 4.748 × 10⁻⁴ (Fig. 8). Furthermore, genomic DNA of 1401D43, 1401D88 can be actively taken up by recipient cells. The transformation frequencies of 1401D43, 1401D88 DNA derived from point mutations in SC1401 are significantly higher than SC1401△htrA genomic DNA which carries a Kan^R cassette in the htrA loci. In independent repeated experiments, more than thirty transformants were directly sequenced for further identification of desired mutations in rpsL.