Introduction

Bacillus amyloliquefaciens is a species of Gram-positive aerobic endospore-forming bacteria used to produce various commercially important enzymes and metabolites (Zhang et al. 2011b). It is also a group of plant growth-promoting rhizobacteria (PGPR) that are capable of promoting plant growth and suppressing soil pathogens through the secretion of various secondary metabolites (Xu et al. 2013). B. amyloliquefaciens such as strains FZB42 (Chen et al. 2007) and SQR9 (Li et al. 2014) have been used commercially as biofertilisers and biocontrol agents in agriculture (Chen et al. 2007; Chowdhury et al. 2015; Huang et al. 2015; Li et al. 2014; Xu et al. 2013). As more genome sequences of B. amyloliquefaciens are released, including strains FZB42 (Chen et al. 2009), SQR9 (Zhang et al. 2015), LL3 (Geng et al. 2011), DSM7 (Rückert et al. 2011), NJN-6 (Yuan et al. 2015), CAUB946 (Blom et al. 2012) and YAUB9601-Y2 (He et al. 2012), simple and rapid genetic manipulation tools are needed to gain insight into gene function and improve these strains via genetic engineering.

Marker-free gene deletion is considered an ideal genetic manipulation with the advantage of no polar effect, guaranteeing safety in case of gene flow by evicting the resistance marker gene. Site-specific recombination systems and counter-selectable markers (CSMs) are generally employed to realize this marker-free manipulation. Site-specific recombination systems, such as Cre/lox (Yan et al. 2008) derived from phage P1 and Flp/FRT (Hoang et al. 1998) first discovered in yeast, allow the excision of a selection marker by catalysing reciprocal site-specific recombination between two lox or FRT sites. However, the lox or FRT site will be left at the replaced locus. Thus, subsequent rounds of manipulations in the same strain may result in the recombination between the lox or FRT sites. In contrast with site-specific recombination systems, CSM-based marker-free manipulation depends upon the host’s endogenous recombination systems. Under defined growth conditions, a CSM leads to the death of the cells harbouring it. Mutants losing both CSM and the resistance marker will survive and be easily selected (Ueki et al. 1996; Wu and Kaiser 1996; Zhang et al. 2006). Wu et al. (2015) used the Cre/lox system to replace the native promoter of the bacilysin operon with the constitutive promoters P repB and P spac in strain FZB42, leaving no resistance marker. Zhang et al. (2014) combined the CSM upp, encoding uracil phosphoribosyl-transferase (UPRTase) and a temperature-sensitive plasmid to successfully remove the 47-kb bae cluster from the genome of B. amyloliquefaciens LL3. However, it was necessary to delete the upp gene of the host strain prior to use of upp as a CSM for marker-free manipulation. Thus, a host-genotype-independent CSM is more desirable to perform marker-free manipulation in B. amyloliquefaciens.

A host-genotype-independent CSM pheS first developed in Escherichia coli (Kast 1994) has been adapted for use as a CSM in several bacteria (Barrett et al. 2008; Carr et al. 2015; Kast and Hennecke 1991; Kristich et al. 2007; Xie et al. 2011; Zhou et al. 2015). The pheS gene encodes the highly conserved phenylalanyl-transfer (RNA) tRNA synthetase α-subunit, and a PheS protein containing an A294G substitution can aminoacylate phenylalanine analogues such as p-chloro-phenylalanine (p-Cl-Phe) (Kast and Hennecke 1991). Presumably, the incorporation of p-Cl-Phe into cellular proteins causes cell death. Therefore, by expressing this point mutant of pheS, it was feasible to perform negative selection in the presence of p-Cl-Phe (Kast 1994). The advantage of this CSM was its versatile utility in a wild-type host, omitting any pre-mutation. To our knowledge, the strategy has not been developed in Bacillus species, and no effective and host-genotype-independent CSM has been employed in B. amyloliquefaciens. In this work, we demonstrate that a mutant of pheS (pheS *) from B. subtilis could be used as a highly effective CSM in three B. amyloliquefaciens strains. A pheS * and overlap-PCR-based method was employed to perform an unmarked deletion of the target gene or large-scale DNA fragments in these strains. Additionally, the potential application of pheS * to other Bacillus species was preliminarily evaluated.

Materials and methods

Bacterial strains, media, growth conditions and phenotypic characterisation

The bacterial strains and plasmids used in this study are listed in Table 1. Luria-Bertani (LB) medium consisted of (g/L) peptide (10), yeast extract (5) and NaCl (5). Minimal medium-glucose-yeast (MGY) medium contained (g/L) glucose (5), yeast extract (4), NH4NO3 (1), NaCl (0.5), K2HPO4 (1.5), KH2PO4 (0.5) and MgSO4 (0.2). MGY-Cl medium was an MGY-based medium supplemented with 5 mM p-Cl-Phe (Sigma, lot no. SHBC0245V, Nanjing, China), which was added to the media prior to autoclaving at 115 °C. Solid medium was obtained by adding 15 g/L agar to the liquid medium. All strains were grown at 37 °C. Unless otherwise indicated, the final concentrations of antibiotics were as follows (mg/L): ampicillin (Amp), 100; chloramphenicol (Cm), 5; and erythromycin (Em), 200 for E. coli and 5 for Bacillus. Amylase activity was detected by growing B. amyloliquefaciens overnight on LB plates with 1 % starch and then staining the plate with iodine.

Table 1 Strains and plasmids used in this study
Full size table

DNA manipulation techniques

Oligonucleotide synthesis (Supplementary Table S1, Online Resource) and DNA sequencing were performed by Sangon Biotech Co., Ltd. (Shanghai, China). The isolation and manipulation of recombinant DNA were performed using standard techniques. All enzymes were commercial preparations. Phusion DNA high-fidelity polymerase was purchased from NEB (Shanghai, China). B. amyloliquefaciens transformation was performed through the artificial induction of genetic competence (Chen et al. 2016). The transformation of other Bacillus was carried out according to published protocols (Anagnostopoulos and Spizizen 1961; Brown and Carlton 1980; Peng et al. 2009; Turgeon et al. 2006; Waschkau et al. 2008; Wemhoff and Meinhardt 2013).

Fusion of multiple DNA fragments by overlap PCR

Fusion of multiple DNA fragments by overlap PCR was carried out as described by Shevchuk et al. (2004). In brief, overlaps of approximately 40 nucleotides were introduced between each of 2 fragments through primers. The fragments were amplified and gel purified. The reaction mixture of step A contained 9.5 μL of water, 4 μL of Phusion buffer (×5), 2 μL of deoxynucleoside triphosphate (dNTP) mix (2.5 mM each), 4 μL of gel-purified fragments (approximately 100 ng each) and 0.5 μL of Phusion DNA polymerase. The cycling parameters were an initial denaturation at 98 °C for 3 min and subsequent steps of 98 °C for 15 s, annealing at 55 °C for 10 s and extension at 72 °C for 3 min for 15 cycles total. The reaction mixture of step B contained 32.5 μL of water, 10 μL of Phusion buffer, 4 μL of dNTP mix, 1 μL of forward and reverse primers (20 mM) specific for the expected fragment, 1 μL of the unpurified PCR product from step A and 0.5 μL of Phusion DNA polymerase. The cycling parameters were an initial denaturation at 98 °C for 3 min and subsequent steps of 98 °C for 10 s, annealing at 58 °C for 10 s and extension at 72 °C for 3 min for 30 cycles total. The resulting product was purified using an AxyPrep DNA Gel Purification and Extraction Kit (Axygen, Hangchow, China), and the purified product was directly transformed into the strains.

Construction of a positive and negative selection cassette P bc -pheS * -cat

The P bc -pheS *-cat (PC) cassette (Fig. 1a) was constructed in three steps using overlap PCR. First, the site-directed mutant of pheS * (GCA309GGC) was generated by overlap extension with primer pairs pheSF/pheSmR and pheSm-F/pheS-R using the pheS gene of B. subtilis 168 as a template. The GCA309GGC mutation was introduced by the primer pheSm-F. Next, as intracellular PheS* should be maintained at a level high enough to induce p-Cl-Phe sensitivity, the strong promoter P bc (http://parts.igem.org/Part:BBa_K090504), which drives the transcription of the uracil phosphatase gene in B. cereus, was amplified from B. cereus strain American Type Culture Collection (ATCC) 14579 using the primer pair P bc -F/P bc -R and was assembled to the 5′ end of pheS * using primers P bc -F/pheS-R to generate P bc -pheS *. The Shine-Dalgarno (SD) sequence (aaggagg) was introduced 8-bp upstream of the start codon of pheS * using primer P bc -R. Finally, the cat gene, together with its putative SD sequence, was amplified from plasmid pNW33N and fused to the 3′ end of the P bc -pheS * with primers cm-F, cm-R, pheS *-F and pheS *-R. Thus, pheS * and cat were arranged into an operon under the control of promoter P bc . This 1.9-kb synthetic positive and negative selection cassette was named PC and cloned into the vector pMD19-T (TaKaRa) to yield the plasmid pTPC.

Fig. 1
figure 1

Graphic representation of the construction of the PC cassette (a) and the strategy of PC cassette and overlap PCR-based marker-free DNA fragment deletion in B. amyloliquefaciens (b). a Primer binding sites are denoted by small arrows, whereas primers containing overlapping sequences are denoted by bent arrows. Overlap PCR was used to introduce a GCA-to-GGC point mutation into codon 309. b To delete a target fragment (dotted line), an approximately 500-bp fragment just downstream of the target region is used as a direct repeat sequence (DR) and is placed ahead of the PC cassette. Then, transformants with an insertion of the fragment containing the DR and PC cassette just upstream of the target region are selected on Cm. Finally, recombination between two DR sequences excises both the PC cassette and the target fragment, and the resulting mutant is selected on 5 mM p-Cl-Phe. P bc , the promoter of the uracil phosphatase gene from B. cereus. SD Shine-Dalgarno sequence, cat chloramphenicol resistance gene, LF left flanking region, RF right flanking region, Chr chromosome

Full size image

Assessment of the p-Cl-phe resistance of different Bacillus strains

Overnight-cultured B. amyloliquefaciens strains SQR9, FZB42 and NJN-6 and other Bacillus species (B. subtilis 168, B. licheniformis ATCC 14580, B. pumilus NJLZ-8, B. thuringiensis DSM 2046, B. cereus ATCC 14579 and B. megaterium MS941) were diluted 100-fold into freshly prepared LB broth and shaken at 37 °C. When the optical density at 600 nm (OD600) reached 1.0, 100 μL of 10−4-fold dilutions was plated on MGY medium containing different concentrations (0, 5, 10, 20 mM) of p-Cl-Phe and incubated overnight at 37 °C.

Assessment of the p-Cl-phe sensitivity of different Bacillus strains containing pheS *

The strains containing pheS * were grown in LB broth supplemented with corresponding antibiotics to OD600 1.0, and 100-μL aliquots of 100- and 10−1-fold dilutions were spread on MGY plate containing the corresponding antibiotics and different concentrations of p-Cl-Phe and the plates were incubated overnight at 37 °C.

Deletion of amyE gene in B. amyloliquefaciens strain SQR9

The left flanking (LF) region (∼800 bp), direct repeat (DR) sequence (∼500 bp) and right flanking (RF) region (∼800 bp) were amplified from strain SQR9 using the primer pairs 9aLF-F/9aLF-R, 9aDR-F/9aDR-R and 9aRF-F/9aRF-R, respectively. The PC cassette was amplified with the primer pair 9aPC-F/9aPC-R using pTPC as a template. These four fragments were fused using overlap PCR in the order of LF, DR, PC cassette and RF. The resulting 4.0-kb amyE deletion amplicon (PCR product) was directly transformed into strain SQR9, and the transformants were selected on LB plates containing Cm. CmR colonies were cultivated to an OD600 of 1.0 without Cm, and a 100-μL aliquot of a 10-fold dilution of the cultures (approximately 105 cells) was plated on MGY-Cl medium. Mutants growing on MGY-Cl were further confirmed by PCR, DNA sequencing and amylase activity analysis.

Deletion of the 37-kb bmy cluster in B. amyloliquefaciens strain SQR9

Deletion of the bmy cluster was carried out using the same strategy as described for amyE deletion. The LF, DR, PC cassette and RF fragments were amplified using the primer pairs 9bLF-F/9bLF-R, 9bDR-F/9bDR-R, 9bPC-F/9bPC-R and 9bRF-F/9bRF-R, respectively. The four fragments were fused to generate the 4.0-kb bmy deletion amplicon, which was subsequently transformed into B. amyloliquefaciens SQR9.

Deletion of amyE gene in B. subtilis strain 168

Deletion of amyE in the B. subtilis strain 168 was also performed according to the method used for the deletion of amyE in the B. amyloliquefaciens strain SQR9. LF, DR, PC cassette and RF fragments were amplified using the primer pairs 8aLF-F/8aLF-R, 8aDR-F/8aDR-R, 8aPC-F/8aPC-R and 8aRF-F/8aRF-R, respectively. The four fragments were fused to generate the 4.0-kb PCR product, which was directly transformed to B. subtilis strain 168.

Construction of the vector pNZT1-pheS *

The fragment containing the P bc -pheS * cassette was amplified with the primer pair Salpbes-F/Pstpbes-R using the PC cassette as a template, digested by SalI and PstI sites and inserted into the corresponding sites of the temperature-sensitive vector pNZT1 (Zakataeva et al. 2010). This produced a vector pNZT1-pheS *, which is able to replicate in both E. coli and Bacillus spp. at or below 30 °C.

Results

p-Cl-phe resistance of wild-type B. amyloliquefaciens strains

To determine whether p-Cl-Phe can be used in counter-selection with its marker pheS * in B. amyloliquefaciens, the resistance of three wild-type strains to different concentrations of p-Cl-Phe was tested as described in the “Materials and methods” section. Strains with survival rates of >80, 50–80 and <50 % were defined as strongly, moderately and weakly resistant to p-Cl-Phe, respectively. As shown in Fig. 2, all three tested wild-type strains of B. amyloliquefaciens (SQR9, FZB42 and NJN-6) showed strong resistance to 5 mM p-Cl-Phe, with survival rates of 97, 95 and 85 %, respectively. In the presence of 10 mM p-Cl-Phe, the B. amyloliquefaciens strain FZB42 showed strong resistance, with a survival rate of 87 %, whereas strains B. amyloliquefaciens SQR9 and NJN-6 showed moderate resistance, with survival rates of 58 and 79 %, respectively. In the presence of 20 mM p-Cl-Phe, strain B. amyloliquefaciens FZB42 still showed strong resistance and a survival rate of 80 %, whereas the strains B. amyloliquefaciens SQR9 and NJN-6 showed weak resistance, with survival rates of 12 and 1 %, respectively. These results lay the foundation for pheS * application in these strains.

Fig. 2
figure 2

p-Cl-Phe resistance of B. amyloliquefaciens. B. amyloliquefaciens strains SQR9, FZB42 and NJN-6 were grown to an OD600 of 1, diluted 104-fold and plated on MGY medium containing different concentrations of p-Cl-Phe. The different letters above bars indicate significant differences (P < 0.05). Data represent means ± standard deviation (SD) (n = 3). The survival rate of each strain on MGY plate was defined as 100 %

Full size image

Generation of a positive and negative selection cassette PC

In previous work, efficient negative selection systems have been developed based on a point-mutant pheS * gene encoding an A294G substitution in E. coli PheS (Kast 1994), A314G substitution in Streptococcus mutans PheS (Xie et al. 2011) or A312G substitution in Enterococcus faecalis PheS (Kristich et al. 2007). To identify the amino acid residue for mutagenesis, a multiple sequence alignment of the PheS proteins from a variety of species was performed by Clustal W2 (Larkin et al. 2007). As shown in Fig. 3, PheS from Bacillus species have a valine at position 294, and the relevant amino acid position of the alanine is 309, which therefore could be substituted (A309G) to create p-Cl-Phe sensitivity. Moreover, the pheS of B. subtilis shares 82 % nucleic acid sequence identity and 92 % amino acid sequence identity with the pheS of B. amyloliquefaciens, suggesting that they probably can replace each other. To reduce undesirable homologous recombination at the wild-type copy of the pheS locus in B. amyloliquefaciens during allelic replacement, the pheS of B. subtilis was engineered and used as a CSM in B. amyloliquefaciens. Codon 309 of B. subtilis pheS was changed from GCA to GGC (A309G) through overlap PCR to generate pheS *. Then, pheS * and cat (chloramphenicol resistant gene) were assembled into an operon under the control of promoter P bc , as described in the “Materials and methods” section, to give the PC cassette.

Fig. 3
figure 3

Multiple sequence alignment of the C-terminal region of PheS from different strains. C-terminal region of strains E. coli K-12, Veillonella atypica OK5, Streptococcus mutans UA159, Burkholderia pseudomallei K42, Enterococcus faecalis OG1, B. subtilis 168, B. amyloliquefaciens strains FZB42, B. licheniformis ATCC 14580, B. pumilus NJLZ-8, B. thuringiensis DSM 2046, B. cereus ATCC 14579 and B. megaterium MS941 were aligned using Clustal W2. The arrow indicates the conserved alanine residue that can be subjected to mutagenesis to create p-Cl-Phe sensitivity. The stars refer to the conserved amino acids

Full size image

PC cassette and PCR-based chromosome marker-free deletion strategy in B. amyloliquefaciens

As shown in Fig. 1b, this strategy depends on three key strategies. First, an approximately 500-bp fragment just downstream of the target region to be deleted is used as a DR sequence and added ahead of the PC cassette. Next, two flanking regions (LF, RF) (∼800 bp on each side), DR and a PC cassette were fused by overlap PCR and transformed into B. amyloliquefaciens strains. Thus, the DR and PC cassette are inserted just upstream of the target region via a double-crossover recombination event, which is selected by Cm. Notably, no target region (or only a very short region) was deleted in this step, ensuring the successful insertion of the DR and PC cassette. Last, the target region together with the PC cassette is excised via a single recombination between the two DR sequences, as selected by p-Cl-Phe. The elongated DR sequences will lead to the efficient eviction of the fragment between them, even if the fragment is large.

Knockout of the amyE gene in B. amyloliquefaciens SQR9

As a proof of concept, we used this strategy to delete the amyE gene in strain SQR9. The amyE deletion amplicon containing the LF, DR, PC cassette and RF was assembled by PCR and transformed into strain SQR9. The DR and PC cassette were inserted upstream of the amyE gene, generating the mutant CmR strain SQR9A. The transformation efficiency was approximately 2.5 × 102 CFU/μg DNA. To investigate the effectiveness of pheS *, approximately 106 cells of mutant strain SQR9A were spread on an MGY plate supplemented with Cm and varying concentrations of p-Cl-Phe. As shown in Fig. 4a, 100 % of the cells of mutant strain SQR9A were killed in the presence of 5 mM p-Cl-Phe, demonstrating that pheS * can act as an effective CSM in strain SQR9. The mutant strain SQR9A was cultured in LB broth without Cm, and approximately 105 cells were plated on an MGY-Cl plate to select the cells that had lost the PC cassette together with the amyE gene. Approximately 400 colonies were obtained per plate. PCR and amylase activity analyses (Fig. 4b, d) showed that all 50 randomly chosen clones had undergone the expected deletion. The final mutant was further verified by DNA sequencing (data not shown) and named as SQR9AS.

Fig. 4
figure 4

Confirmation of the PC cassette-based DNA fragment deletion in B. amyloliquefaciens SQR9. a Conditional growth inhibition mediated by PC cassette in B. amyloliquefaciens SQR9A. The strain SQR9A carrying the PC cassette was cultivated to an OD600 of 1 in LB broth containing Cm, and 100-μL aliquots without dilution were plated on MGY plate supplemented with Cm and the indicated concentration of p-Cl-Phe. The plates were incubated at 37 °C overnight. b Lanes 1 and 4 DNA marker. Lanes 2 and 3 show the amplicons of the amyE region in strain SQR9 (lane 2) and strain SQR9AS (lane 3) using outside oligonucleotides 9a–VF/9a–VR as primers. c Lane 1 DNA marker. Lane 2 shows the length of bmy region in strain SQR9BS using outside oligonucleotides 9b–VF/9b–VR as primers. Lanes 3 and 4 show a 1.8-kb fragment from bmy cluster in strain SQR9 and strain SQR9BS using the primer pair 9b–MF/9b–MR. d Strains SQR9 and SQR9AS were grown on LB plates with 1 % starch overnight and stained with iodine to detect the α-amylase activity, which was indicated by the transparent plaque

Full size image

Deletion of the 37-kb bmy cluster in B. amyloliquefaciens SQR9

To demonstrate the capacity of this strategy to remove large fragments, the 37-kb bmy cluster that is responsible for the synthesis of bacillomycin D (Zhang et al. 2015) was targeted for deletion in strain SQR9. As described in the “Materials and Methods” section, the bmy deletion amplicon was generated by PCR and introduced into strain SQR9. The transformation efficiency was approximately 2 × 102 CFU/μg DNA. The CmR clones carrying the PC cassette were designated SQR9B. Subsequently, approximately 105 cells of mutant strain SQR9B were spread on an MGY-Cl plate to select the cells that had lost the PC cassette together with the bmy cluster. Approximately 40 colonies were obtained from each plate, which is less than in the amyE deletion. PCR analysis (Fig. 4c) showed that all 50 randomly selected clones were correct. The final mutant was further checked by DNA sequencing (data not shown) and designated SQR9BS. Thus, this strategy performed very well in deleting a large chromosomal fragment in strain SQR9.

Application of the pheS * in other B. amyloliquefaciens strains

To determine whether the application of pheS * can be extended to other B. amyloliquefaciens strains, we further tested the effectiveness of pheS * in B. amyloliquefaciens strains FZB42 and NJN-6. The amyE genes of both strains were successfully deleted using the strategy mentioned above (Fig. 1b). The deletion amplicons were assembled by PCR and transferred into strains FZB42 and NJN-6, generating mutant strains FZB42A and NJN-6A, which harboured the PC cassette. The transformation efficiencies of this step for strain FZB42 and NJN-6 were approximately 3 × 102 and 0.3 × 102 CFU/μg DNA, respectively. When 105 cells were spread on an MGY plate containing p-Cl-Phe and Cm, mutant FZB42A showed pronounced sensitivity to 10 mM p-Cl-Phe, and mutant NJN-6A exhibited pronounced sensitivity to 5 mM p-Cl-Phe (Supplementary Figs. S1 and S2, Online Resource).

Approximately 105 cells of mutant FZB42A or NJN-6A were plated on an MGY-Cl plate containing 5 mM p-Cl-Phe to select the cells that had lost the PC cassette together with the amyE gene. Approximately 400 colonies were obtained per plate, and 50 clones were randomly chosen for each strain. Of these, 100 % had undergone the expected deletion (Supplementary Figs. S3 and S4, Online Resource). Notably, although the strains FZB42 and NJN-6 harbouring pheS * are less sensitive to p-Cl-Phe than strain SQR9 (Fig. 4a), pheS *-based counter-selection was equally efficient (100 %) in these three strains. The final amyE-deletion mutant strains verified by DNA sequencing were designated FZB42AS and NJN-6AS, respectively. The effectiveness of pheS * was thus validated in three B. amyloliquefaciens strains, so it is reasonable to expect that pheS * could be used as a CSM in other B. amyloliquefaciens strains.

Evaluation of the potential use of pheS * in other Bacillus species

  1. (i)

    p -Cl-Phe resistance of other Bacillus species

To evaluate the potential use of pheS * in other Bacillus species, the p-Cl-Phe resistance of Bacillus species, including B. subtilis, B. licheniformis, B. pumilus, B. thuringiensis, B. cereus and B. megaterium, was determined on MGY medium containing different concentrations of p-Cl-Phe. As shown in Supplementary Fig. S5, Online Resource, all tested strains of B. subtilis 168, B. licheniformis ATCC 14580, B. pumilus NJLZ-8, B. thuringiensis DSM 2046, B. cereus ATCC 14579 and B. megaterium MS941 showed strong resistance to 5 mM p-Cl-Phe, with survival rates of 82, 96, 93, 95, 93 and 89 %, respectively. In the presence of 10 mM p-Cl-Phe, the strains B. licheniformis ATCC 14580, B. pumilus NJLZ-8, B. thuringiensis DSM 2046 and B. cereus ATCC 14579 showed strong resistance, with survival rates of 87, 91, 86 and 90 %, respectively, whereas strains B. subtilis 168 and B. megaterium MS941 showed weak resistance, with survival rates of 43 and 24 %, respectively. In the presence of 20 mM p-Cl-Phe, the strains B. licheniformis ATCC 14580, B. pumilus NJLZ-8 and B. cereus ATCC 14579 still showed strong resistance, with survival rates of 83, 88 and 85 %, respectively, whereas B. thuringiensis strain DSM 2046 showed moderate resistance and a survival rate of 69 %. The strains B. subtilis 168 and B. megaterium MS941 showed weak resistance (both with survival rates of 1 %).

  1. (ii)

    PC cassette and PCR-based marker-free deletion of amyE gene in B. subtilis 168

The PC cassette was first inserted at the amyE locus following the strategy shown in Fig. 1b, generating mutant 168A. When approximately 106 cells were spread on an MGY plate supplemented with p-Cl-Phe and Cm, mutant 168A exhibited prominent sensitivity to 5 mM p-Cl-Phe (Fig. 5). Next, mutant 168A was grown to OD600∼1.0 in LB broth without Cm, and approximately 105 cells of mutant 168A were plated on an MGY plate containing 5 mM p-Cl-Phe. Approximately 150 clones were obtained per plate. All of the 50 randomly tested clones lost the PC cassette through recombination between the two DR sequences (Supplementary Fig. S6, Online Resource), and the final mutant was designated 168AS.

  1. (iii)

    Expression of PheS * in B. cereus

Fig. 5
figure 5

p-Cl-Phe sensitivity of B. subtilis strains 168 and 168A. The strains 168 and 168A carrying the PC cassette were cultivated to an OD600 of 1 in LB broth, and 100-μL aliquots of each dilution were plated on MGY plate supplemented with indicated concentration of p-Cl-Phe. To prevent the loss of PC cassette, Cm was added in the cultivation of strain 168A. The plates were incubated at 37 °C overnight

Full size image

To investigate whether pheS * could be used as a CSM in B. cereus ATCC 14579, P bc -pheS * was ligated to vector pNZT1, generating the plasmid pNZT1-pheS *. Plasmids pNZT1 and pNZT1-pheS * were transformed to B. cereus ATCC 14579. The strain carrying pNZT1-pheS * was sensitive to 5 mM p-Cl-Phe on MGY plates (Fig. 6), indicating that pheS * could be an effective CSM in B. cereus.

Fig. 6
figure 6

p-Cl-Phe sensitivity of B. cereus strains ATCC 14579 (pNZT1) and ATCC 14579 (pNZT1-pheS *). The strains ATCC 14579 (pNZT1) and ATCC 14579 (pNZT1-pheS *) were cultivated to an OD600 of 1 in LB broth containing Em, and 100-μL aliquots of each dilution were plated on MGY plate supplemented with Em and indicated concentration of p-Cl-Phe. The plates were incubated at 30 °C overnight

Full size image

Discussion

In this study, a host-genotype-independent CSM pheS * was generated for B. amyloliquefaciens. The pheS * was demonstrated to be extremely effective through unmarked gene knockout or large-fragment deletion in three important B. amyloliquefaciens strains. The pheS gene encodes the α-subunits of the Phe-tRNA synthetase. Kast and Hennecke (1991) found that replacement of the Ala294 of PheS with a smaller Gly residue renders E. coli sensitive to p-Cl-Phe and demonstrated that this site-directed mutant of pheS could act as an effective CSM in E. coli. However, to extend this strategy to other bacteria, at least three points should be noted. First, the wild-type host cell is resistant to p-Cl-Phe. For example, the majority of wild-type sphingomonads failed to grow under addition of 0.1 mM p-Cl-Phe, so pheS cannot be used as a CSM for these strains (Kaczmarczyk et al. 2012). On the other hand, host cells harbouring mutant pheS must be sensitive to p-Cl-Phe, enabling efficient negative selection. Finally, undesirable homologous recombination at the wild-type copy of the pheS locus should be reduced during gene knockout, which can be accomplished by introducing a series of silent mutations into the mutant pheS.

In this work, the p-Cl-Phe resistance of B. amyloliquefaciens was assessed, and all three strains tested showed strong resistance to 5 mM p-Cl-Phe. To reduce the undesirable homologous recombination at the wild-type copy of the pheS locus in B. amyloliquefaciens, the pheS of B. subtilis with 82 % nucleic acid sequence identity to B. amyloliquefaciens was mutated and used as a CSM in B. amyloliquefaciens. A strong promoter P bc was used to drive pheS *, generating high intracellular levels of PheS* to create p-Cl-Phe sensitivity. Fortunately, P bc -pheS * rendered the B. amyloliquefaciens strains that harboured it pronounced sensitivity to p-Cl-Phe.

Many Bacillus species are of considerable interest to industry and agriculture. Both gene function research and strain improvement require effective CSM to enable efficient unmarked genetic manipulation in these species. Although several CSMs have been developed for Bacillus species (Brans et al. 2004; Liu et al. 2008; Wemhoff and Meinhardt 2013; Zhang et al. 2011a; Zhang et al. 2006), only mazF was demonstrated as a host-genotype-independent CSM in B. subtilis. Other CSMs require pre-mutation of the host; for instance, to use the upp gene as a CSM in B. subtilis (Fabret et al. 2002) and B. amyloliquefaciens (Zhang et al. 2014), the native upp must first be deleted. Before employing blaI (Brans et al. 2004) as a CSM, the endogenous P lysA promoter of B. subtilis should be replaced with the P blaP promoter. The B. subtilis chromosomal araR locus must be replaced with a promoterless neomycin-resistance gene (neo) fused to the ara promoter prior to the use of araR as a CSM (Liu et al. 2008). The pheS * generated in this work was demonstrated as an effective host-genotype-independent CSM in B. amyloliquefaciens as well as in B. subtilis 168 and B. cereus ATCC 14579, which will facilitate functional genome research and industrial application of these Bacillus spp. In addition, plasmids pTPC and pNZT1-pheS * will be available from BGSC (http://www.bgsc.org) under the accession numbers ECE355 and ECE356, respectively.