Antibacterial Activity of Bacillus inaquosorum Strain T1 against pirAB^Vp-Bearing Vibrio parahaemolyticus: Genetic and Physiological Characterization

The shrimp aquaculture industry has been significantly impacted by acute hepatopancreatic necrosis disease (AHPND), resulting in significant financial losses annually. AHPND is caused by strains of the bacterial pathogen Vibrio parahaemolyticus, and treatment of AHPND involves the use of antibiotics, which leads to a rise in the number of antibiotic-resistant strains. Alternative treatments include the application of beneficial microorganisms having inhibitory activities against pathogens causing AHPND. In this study, we examined the ability of Bacillus inaquosorum strain T1 to inhibit the growth of an AHPND-causing Vibrio strain, and we show that this activity involves a gene cluster associated with antibacterial compound production. We found that gene expression is under stationary-phase control and that enhanced activity occurs upon inactivation of a global transition state regulator. Our approach for understanding the factors involved in producing B. inaquosorum strain T1 inhibitory activity will allow for the development of this strain as a tool for AHPND prevention and treatment.

ABSTRACT

Acute hepatopancreatic necrosis disease (AHPND) is caused by PirAB toxin-producing Vibrio parahaemolyticus and has devastated the global shrimp aquaculture industry. One approach for preventing the growth of AHPND-producing Vibrio spp. is through the application of beneficial bacteria capable of inhibiting these pathogens. In this study, we focused on the inhibitory activity of Bacillus inaquosorum strain T1, which hinders V. parahaemolyticus growth in coculture experiments in a density-dependent manner; inhibition was also observed using cell-free supernatants from T1 stationary-phase cultures. Using mariner-based transposon mutagenesis, 17 mutants having a complete or partial loss of inhibitory activity were identified. Of those displaying a total loss of activity, 13 had insertions within a 42.6-kb DNA region comprising 15 genes whose deduced products were homologous to nonribosomal polypeptide synthetases (NRPSs), polyketide synthases (PKSs), and related activities, which were mapped as one transcriptional unit. Mutants with partial activity contained insertions in spo0A and oppA, indicating stationary-phase control. The levels of expression of NRPS and PKS lacZ transcriptional fusions were negligible during growth and were the highest during early stationary phase. Inactivation of sigH resulted in a loss of inhibitor activity, indicating a role for σ^H in transcription. Disruption of abrB resulted in NRPS and PKS gene overexpression during growth as well as enhanced growth inhibition. Our characterization of the expression and control of an NRPS-PKS gene cluster in B. inaquosorum T1 provides an understanding of the factors involved in inhibitor production, enabling this strain’s development for use as a tool against AHPND-causing Vibrio pathogens in shrimp aquaculture.

IMPORTANCE The shrimp aquaculture industry has been significantly impacted by acute hepatopancreatic necrosis disease (AHPND), resulting in significant financial losses annually. AHPND is caused by strains of the bacterial pathogen Vibrio parahaemolyticus, and treatment of AHPND involves the use of antibiotics, which leads to a rise in the number of antibiotic-resistant strains. Alternative treatments include the application of beneficial microorganisms having inhibitory activities against pathogens causing AHPND. In this study, we examined the ability of Bacillus inaquosorum strain T1 to inhibit the growth of an AHPND-causing Vibrio strain, and we show that this activity involves a gene cluster associated with antibacterial compound production. We found that gene expression is under stationary-phase control and that enhanced activity occurs upon inactivation of a global transition state regulator. Our approach for understanding the factors involved in producing B. inaquosorum strain T1 inhibitory activity will allow for the development of this strain as a tool for AHPND prevention and treatment.

INTRODUCTION

Acute hepatopancreatic necrosis disease (AHPND) in juvenile penaeid shrimp, also known as early mortality syndrome (EMS), first emerged in China in 2009 (1) and has spread to Vietnam, Malaysia, Thailand, Mexico, and the Philippines, as well as throughout South America (1) and the United States (2). AHPND leads to 100% mortality and has resulted in annual losses for shrimp farming estimated to be over $1 billion (3). The disease is caused by strains of Vibrio parahaemolyticus that carry a plasmid harboring the gene for a Photorhabdus insect-related binary toxin, PirAB, also known as Pir-likeAB and PirVP (4–6). AHPND is lethal in both Penaeus monodon and Litopenaeus vannamei, the two most cultivated shrimp species in the aquaculture industry (3). Treatment has led to the inevitable increase in resistance to commonly used antibiotics, e.g., oxytetracycline, quinolones, and amoxicillin (7, 8). Thus, alternative approaches for preventing and treating this disease are warranted.

Colonization by AHPND-causing V. parahaemolyticus strains creates a major shift in the microbiota of the shrimp gut (9). In healthy shrimp, Rhodobacterales, Rhizobiales, and Planctomycetales are the major contributors to the gut microbiome. After infection, the Mycoplasmatales and Vibrionales are the dominant gut inhabitants (10). The practice of draining and disinfecting ponds between shrimp stocks may increase the risk of an AHPND outbreak because it removes beneficial bacteria. Manipulating microbial communities through the use of microbially mature water has been shown to increase the survival of fish larvae over that achieved by the use of filter-sterilized water (11). Thus, one approach to combating this disease is through the use of beneficial bacteria—probiotics—capable of positively modifying shrimp microbial communities and, at the same time, interfering with the growth of pathogens like the AHPND-causing strains. Probiotics are live microorganisms that, when administered in adequate amounts, confer health benefits to the host (12). These benefits occur through a variety of mechanisms, including the production of antimicrobial compounds (12).

A group of bacteria that has received attention for its probiotic potential is the Gram-positive spore-forming Bacillus spp., due to the antimicrobial activities of their structurally diverse secondary metabolites, including polyketides, aminoglycosides, and nonribosomal peptides, such as bacteriocins and lipopeptides (13, 14). Production of these antimicrobial compounds usually accompanies stationary-phase physiology, which is associated with adverse changes in their environment. The genes for the biosynthetic pathways for these activities are often organized as operons and include genes for nonribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs), which are modular in design and which are arranged in various combinations, resulting in the production of different compounds (15, 16). Many act through permeabilization and destruction of the cell membrane and other mechanisms (15).

The present study focuses on Bacillus inaquosorum strain T1, which we have found produces an inhibitory activity against AHPND-producing Vibrio parahaemolyticus strains (17). Originally characterized as subspecies of Bacillus subtilis (18) and recently promoted to species status (19), B. inaquosorum and Bacillus stercoris differ from closely related Bacillus subtilis and Bacillus spizizenii by the presence of an uncharacterized NRPS-PKS gene cluster (19). In this study, we demonstrate that T1 inhibitory activity is associated with a secreted product, and we demonstrate the involvement of the NRPS-PKS-encoding region in this activity. We found that NRPS-PKS gene expression was negligible during mid-exponential growth and highest during stationary phase and that stationary-phase regulators oppA (spo0K), spo0A, and sigH (spo0H) are involved in its control. Finally, disruption of the global transitional-phase regulator abrB resulted in derepressed NRPS-PKS expression during exponential phase, resulting in enhanced inhibitory activity. Our understanding of the factors involved in NRPS-PKS gene expression and control in strain T1 and our ability to manipulate the production of this activity may aid in the development of tools that could be used against AHPND in shrimp.

RESULTS

Assessing T1 inhibitory activity.

Using the soft-agar overlay assay (see Materials and Methods), T1 was shown to inhibit the growth of three strains of Vibrio parahaemolyticus—AHPND-causing strain D4 (Fig. 1A), isolated from Mexico, and AHPND-causing strain A3 (Fig. 1B), isolated from Vietnam, as well as a non-AHPND-causing strain, ATCC 17802 (Fig. 1C)—which was evident by a zone of clearance around the T1 colony. Because chloroform is toxic to T1, the growth inhibition observed for the Vibrio strains applied in the soft agar over the T1 colony is a consequence of the accumulation of a diffusible substance produced by T1. In contrast, B. subtilis strain SMY (20) did not inhibit the growth of any of the three strains (Fig. 1, right), highlighting a difference between B. inaquosorum and B. subtilis. Because strain D4 was more sensitive to T1 than A3, based on the size of the clearance zone, we used D4 as the test strain for subsequent studies.

FIG 1.

Soft-agar overlay assays for B. inaquosorum strain T1 (left) and B. subtilis strain SMY (right) against pirAB-harboring V. parahaemolyticus strains D4 (A) and A3 (B) and non-pirAB-harboring V. parahaemolyticus strain ATCC 17802 (C). Assays were done as described in Materials and Methods. A zone of clearance around a colony indicates the absence of growth of the Vibrio spp.

The ability of T1 to effect D4 growth was examined further by coculturing the two strains and monitoring D4 levels by measuring the toxR copy number (see Materials and Methods). For these growth experiments, overnight cultures of D4 were diluted to 2 × 10⁴ CFU/ml and combined with overnight cultures of strain T1g diluted to final densities of 2 × 10⁴, 2 × 10⁵, and 2 × 10⁶ CFU/ml (T1g and D4 at 1:1, 10:1, and 100:1, respectively) in 2216 broth, as described in Materials and Methods; the control culture did not contain T1g. After 3 h and 24 h, samples were collected and DNA was extracted for quantitative PCR (qPCR) analysis. At 3 h, toxR levels did not vary significantly for any of the treatments (Table 1). At 24 h, the levels increased approximately 10³- to 10⁴-fold for the control culture and cultures supplemented with T1g at the 1:1 and 10:1 ratios. On the other hand, at 24 h, the D4 toxR level in the culture containing T1g at a 100:1 ratio was between 750- and 1,250-fold lower than that for the control (no T1g) and the other T1-supplemented cultures and increased only approximately 70-fold compared to that at its 3-h time point (Table 1). At 24 h, T1g levels (based on gfp copy number) were not significantly different for all T1g-containing cultures, varying from 6 × 10⁷ ± 1.1 × 10⁷ copies/ml for the 1:1 culture to 1.6 × 10⁸ ± 0.6 × 10⁸ copies/ml for the 100:1 culture. Thus, while T1g reached similar cell densities for all three treatments, D4 growth inhibition due to T1g could be observed, but only when T1g was present at a 100-fold excess, suggesting both a relationship between the inhibitor and the culture density and an ability of D4 to overcome the inhibition.

TABLE 1.

Effect of T1 and A3-41 on D4 growth^a

Addition to D4 culture, final concn (no. of CFU/ml)	toxR copy no./ml at:
	3 h	24 h
None	1.6 × 105 ± 0.3 × 105	1.3 × 109 ± 0.5 × 109
T1, 2 × 104	8.8 × 105 ± 6.5 × 105	1.5 × 109 ± 0.9 × 109
T1, 2 × 105	8.8 × 105 ± 6.3 × 105	8.8 × 108 ± 1.5 × 108
T1, 2 × 106	1.7 × 105 ± 0.6 × 105	1.2 × 106 ± 0.2 × 106
A3-41, 2 × 104	7.5 × 105 ± 7.1 × 105	7.0 × 109 ± 2.0 × 109
A3-41, 2 × 105	1.4 × 106 ± 1.0 × 106	3.1 × 109 ± 0.7 × 109
A3-41, 2 × 106	1.9 × 105 ± 1.0 × 105	2.1 × 109 ± 1.6 × 109

^aThe average toxR copy numbers for D4 cultures grown in 2216 medium mixed with T1 or A3-41 after 3 and 24 h were determined as described in Materials and Methods. The initial cell density for strain D4 was 2 × 10⁴ CFU/ml. The results are the averages from four replicates.

A scenario that could explain the density-dependent activity observed in the coculture experiments, which would be consistent with the production of a diffusible substance inferred from the overlay assay, is that T1 produces and secretes a product that is inhibitory to D4 and accumulates during late exponential and/or stationary phase. To test this possibility, supernatant fractions from 18-h T1 cultures grown in 2216 broth were filtered and mixed with equal volumes of overnight D4 cultures freshly diluted into 2216, with D4 growth being monitored as described in Materials and Methods. As shown in Fig. 2, while all cultures grew to similar final densities, the growth of the D4 culture supplemented with the T1 cell-free supernatant was delayed for 5 to 6 h compared to the growth of the D4 control culture, while no delay was observed for the culture treated with the cell-free supernatant from an overnight D4 culture. D4 toxR levels in the T1 supernatant-treated culture were approximately 3 × 10⁵-fold lower than those in the untreated D4 culture (Table 2) at 5.5 h after inoculation. On the other hand, D4 levels in the D4 culture treated with its own supernatant declined only approximately 10-fold (Table 2), indicating that inhibition by the T1 supernatant was due to a factor specifically produced by T1 and was not likely due to the exhaustion of nutrients in the spent medium. Inhibition was also found to be concentration dependent, since the addition of the T1 cell-free supernatant at 25% resulted in an approximately 2-h lag (not shown). D4 growth was not affected when cell-free supernatants were prepared from mid-exponential-phase T1 cultures (data not shown).

FIG 2.

Effect of culture supernatant on D4 growth. The growth of an untreated D4 culture (●) and cultures mixed with an equal volume of cell-free supernatant fractions prepared from overnight cultures of wild-type T1 (■), D4 (▲), and A3-41 (♦) is shown. All cultures were grown in 2216 broth. The dotted line indicates the anticipated growth between 13 and 24 h. The graph shows the results of a representative experiment, and standard deviations were <10% and were omitted for clarity.

TABLE 2.

Effect of cell-free culture supernatant on D4 growth^a

Addition to D4 culture	toxR copy no./ml
50% 2216 medium	5.3 × 107 ± 0.7 × 107
50% D4 supernatant	5.0 × 106 ± 1.2 × 106
50% T1 supernatant	1.8 × 102 ± 0.7 × 102
50% A3-41 supernatant	1.5 × 107 ± 0.4 × 107

^aThe average toxR copy number for D4 cultures grown with supernatant fractions from 18-h cultures (50% of the total culture volume) was determined as described in Materials and Methods. Samples were taken 5.5 h after inoculation. The results are the averages from four technical replicates.

Identification of T1 genes involved in D4-inhibitory activity.

To determine the genetic basis for the T1 inhibitory activity, we generated a transposon insertion library using the mariner-derived himar1 transposase, as described in Materials and Methods. Over 3,000 transposon-containing mutants were screened for the loss of T1 inhibitory activity by the overlay assay. Seventeen mutants were identified as having a complete or partial loss of activity against D4, and insertion site locations could be identified for 16 mutants, and these locations are listed in Table 3; the results of overlay assays for mutants A2-18 (with a partial loss of activity), A3-41, A11-79, and A20-86 (with a full loss of activity) are shown in Fig. 3. The insertion sites for 13 mutants were found to be clustered in seven open reading frames within a 42.6-kb region of DNA positioned between the pcrA-ligA operon and phoA at the 5′ and 3′ ends (Table 3 and Fig. 4), respectively, found in B. inaquosorum and B. stercoris (19). BLASTp analysis revealed that the deduced orfA, orfB, and orfC products contained peptide synthesis, condensation, adenylation, and acyl carrier domains conserved among NRPSs. Domains conserved among type I PKSs, including ketoreductase, ketoacyl synthase, and acyl transferase, were evident for the orfD, orfE, and orfF products. Furthermore, homologies to Bacillus proteins involved in antimicrobial peptide and polyketide synthesis were found for the other open reading frame products, including a major facilitator superfamily (MFS) transporter (orf2), an endopeptidase processing enzyme (orf3), thioesterase (orf5), acyl coenzyme A (CoA) dehydrogenases (orf6 and orf9), and acyl carrier protein (orf8) (Fig. 4). The DNA sequence of this region was deposited in GenBank under accession number MT366812.

TABLE 3.

Summary of transposon insertion sites in T1 mutants^a

Mutant	Delivery vector	Transposon insertion site (position)	Gene, putative gene product	NRPS-PKS ORF coding sequence (positions)
A20-86	pDP384	8534	orfA, NRPS	7036 to 11532
S21-46	pEP4	9917	orfA, NRPS
A13-56	pDP384	18057	orf9, acyl-CoA dehydrogenase	16935 to 18062
A2-30	pEP4	19316	orfB, NRPS	18084 to 27179
A10-67	pDP384	19316	orfB, NRPS
S11-14	pEP4	21936	orfB, NRPS
S12-29	pEP4	21936	orfB, NRPS
A3-41	pDP384	24490	orfB, NRPS
A11-79	pDP384	27466	orfC, NRPS	27199 to 29868
S21-28	pEP4	28790	orfC, NRPS
A1-20	pDP384	32220	orfD, PKS	29861 to 34387
S35-30	pEP4	34499	orfE, PKS	34405 to 41949
A23-33	pDP384	44976	orfF, PKS	41951 to 48379
A2-18	pDP384	399b	oppA, ABC transporter substrate-binding protein	NA
S31-22	pEP4	427b	yoaZ, putative oxidative stress response factor	NA
A8-11	pDP384	68b	spo0A, sporulation master regulator	NA

^aShown are the delivery vector used to create the mutant, the location of the insertion site, the gene and its putative product, and the nucleotide location of the ORF coding region. The GenBank accession number for the DNA sequence is MT366812. For mutations within the NRPS-PKS gene cluster, the nucleotide position of the insertion site is indicated relative to the start of the pcrA gene, the first gene of the sequenced region. NRPS, nonribosomal peptide synthetase; PKS, polyketide synthase; NA, not applicable. The NRPS-PKS ORF coding region for a gene with multiple mutants is shown for the first entry corresponding to that gene.

^bThe insertion site location is relative to the initiation codon of the corresponding gene in B. subtilis (GenBank accession number NC_000964.3).

FIG 3.

Soft-agar overlay assays for the activity of select T1 mutants (left) and wild-type T1 (right) against strain D4. The assays were done as described in Materials and Methods.

FIG 4.

Genetic and transcription map of the NRPS-PKS region of strain T1 and mutant insertion sites. See the text for details.

In addition to evaluating mutant activities using the overlay assay, we examined the activity of mutant A3-41, grown in coculture with D4 at 1:1, 10:1, and 100:1 ratios. This mutant contains an insertion in the NRPS-like orfB gene, about halfway into the NRPS-PKS region (Table 3 and Fig. 4). Consistent with the overlay assay results, A3-41 did not significantly affect D4 levels at either the 3-h or the 24-h time point (Table 1) when included at any ratio. Furthermore, the cell-free supernatant prepared from A3-41 did not delay D4 growth like its wild-type T1 parent did (Fig. 2), and D4 levels at 5.5 h after inoculation in the presence of the A3-41 cell-free supernatant were comparable to those in the control D4 culture (Table 2). At 24 h, A3-41 levels (based on amyE copy number) were not significantly different for all A3-41-containing cultures and were similar to 24-h T1g levels, varying from 2.4 × 10⁸ ± 0.5 × 10⁸ copies/ml for the 1:1 culture to 4.7 × 10⁸ ± 1.2 × 10⁸ copies/ml for the 100:1 culture. Thus, the loss of inhibitory activity was directly due to elimination of the orfB product and was consistent with a role for the NRPS-PKS region in producing the anti-D4 activity.

Mutants with insertions outside of the NRPS-PKS cluster were also found, including those with insertions within spo0A (mutant A8-11; Table 3), the master transcriptional regulator required for activating early sporulation and early-stationary-phase genes in B. subtilis 168 (21), and oppA (mutant A2-18; Table 3). Also known as spo0K, oppA is the first gene of a five-gene operon in B. subtilis 168 responsible for the synthesis of the oligopeptide-binding permease, an ABC transporter that plays a role in a variety of stationary-phase activities, including initiation of sporulation and competence development (22). In both cases, gene disruption resulted in a partial loss of D4-inhibitory activity (Fig. 3 and data not shown). Mutant S31-22 (Table 3), which was also defective in D4-inhibitory activity, contained an insertion in a gene homologous to yoaZ in B. subtilis 168 (BSU_18790), a putative oxidative stress response factor whose function is unclear (23).

Transcription mapping of NRPS-PKS region.

RNA extracted from mid-exponential-phase and early-stationary-phase cultures of T1 grown in 2216 broth was examined by reverse transcriptase (RT) PCR, as described in Materials and Methods. Primers were designed to target the amino and carboxyl termini of consecutive predicted open reading frames (ORFs) (Table 4). Regions spanning all 17 predicted ORFs in the NRPS-PKS region, as well as the adjacent upstream pcrA operon and downstream phoA gene, were tested. The amplification of cDNA generated from both mid-exponential-phase (not shown) and stationary-phase (Fig. 5) RNA was observed for the entire NRPS-PKS region spanning from orf2 to orf10, indicating that the entire 42.6-kb region encompassed a single transcription unit. No amplification products were observed for the orf1-orf2 and orf10-phoA regions (Fig. 5), signifying the absence of a transcript spanning these genes and demarcating the 5′ and 3′ ends of the NRPS-PKS region transcriptional unit to be at orf2 and orf10, respectively. The absence of a product between orf10 and phoA was expected, since phoA is expressed in a direction opposite of that for the NRPS-PKS region. Both the orf10 and phoA stop codons are followed by sequences that could form, respectively, 25- and 27-base mRNA hairpin structures (ΔG = −17 kcal/mol for both, determined using the Vienna RNA Websuite [24]) with U-rich ends, which likely act as transcription terminators. Similarly, pcrA, ligA, and orf1 are part of a four-gene operon—pcrA is preceded by pcrB, and orf1 is referred to as yerH by Petit and Ehrlich (25)—with a transcription terminator found after orf1 (see below), so the cotranscription of orf1-orf2 was not anticipated.

TABLE 4.

Primers used in this study

Target	Sequence (5′–3′)a		Product size (bp)
	Forward primer	Reverse primer
Spr	TCTGATTACCAATTAGAATGAAT (2569R)	GAATATACGGAAATTATGACTTA (2570R)	722
gfp	CGACATTGTGTGGACAGGTAA	CCCGAAGGTTATGTACAGGAAAG	353
amyE	AGGCTGGGCAGTGATTGCTT	ACTTCCGCGGTCGCCTATTT	110
toxR	AATCCATGGATTCCACGCGTTAT	CACCAATCTGACGGAACTGAGATTC	103
Transposon insertion site	ATATTCATTCTAATTGGTAATCAGA (2569)	CTAAGTCATAATTTCCGTATATTC (2570)	Variable
pcrA→ligA	CGAACAAATATGAACCCGGAAT	GTCGAAATCACGGAGATCATCT	606
ligA→orf1	GGAACAACTGTCAAGAAACGAT	CAGATAATCGTCAGTAGAGAAAGAATC	505
orf1→orf2	CGAGGTCTTTAAGCGGTTCA	AGCAGGAGCACGATTTGATC	868
orf2→orfA	CTGTCGGCTCAGCATACAATAAGTAC	CAAGCGCGCCATCATATAAAGAG	604
orfA→orf3	GTCTGACAGTCAGTCTTGATTTGACT	TGCTCAGCCATATATTGCTCGTA	855
orf3→orf4	GTTCAGCAGCGAACAAGAATA	GTCTGGCTTCATATGACAAGT	702
orf4→orf6	GCAATACCGCCTCGCAT	GGCTCAGTCGTGAATTGAAT	1,166
orf6→orf7	GCATTTGGACGGTTCAAGAT	GGATCGACGATTAGATCCTTATA	627
orf7→orf9	GACCTGAGCCGGATTCAG	GTTCAGACAATGCAAACGCT	742
orf9→orfB	CGAGGTTTCTGATGACCAGATA	CGAACCGTTTGCTGCCAT	732
orfB→orfC	GCAGCGGATCATCGGTAT	GAAGCTCCGGATGCCTTTAAG	754
orfC→orfD	GAATCTGACGAGCAGGTCATA	GTCAAAGAAGATGAACAGGCTG	894
orfD→orfE	GAGCAACACCAAGGAATCCT	ACGAAGCAGCGTCTCATCTA	692
orfE→orfF	GGAGCAAGTCGGCATACAT	CAGTCTGCACCATCATGCT	819
orfF→orf10	GAAATTCGCTGCTCGCTGT	GTAGAGGAGAATGCTGATATCATTCAC	840
orf10→phoA	CAGCAGGCAGACGATTTAATG	CAGGCTGGACGACGTTA	680

^aThe primer name is given in parentheses.

FIG 5.

Transcription mapping of the NRPS-PKS region. The predicted open reading frames and gel electrophoresis of RT-PCR products generated using T1 RNA isolated from stationary-phase cultures (T₁) are shown. Primer pairs targeting open reading frame termini in the NRPS-PKS region as well as upstream and downstream genes, as indicated, are described in Materials and Methods and Table 4. Unmarked lanes, a 100-bp DNA size ladder; lanes G, PCR mixtures containing genomic DNA (without RT or DNase); lanes N, PCR mixtures with RNA and DNase but without RT; lanes C, RT-PCR of RNA after DNase treatment. The absence of the RT product for the orf1-orf2 and orf10-phoA regions is indicated by triangles; potential transcription terminators are indicated by a “T.”

Control of NRPS-PKS gene expression.

Several transposon mutants were obtained using plasmid pEP4, which, upon insertion, resulted in the creation of a transcriptional fusion with lacZ, thereby enabling the analysis of gene expression during growth and stationary phase. β-Galactosidase levels were determined for four mutants, A20-86, A3-41, A11-79, and A1-20, containing fusions with orfA, orfB, orfC, and orfD, respectively (Table 5). These mutants were grown in 2216 broth and harvested at mid-exponential phase (T₋₁), at the transition into stationary phase (T₀), and 1 to 2 h into stationary phase (T₁), and the results are shown in Table 6. During mid-exponential-phase growth, β-galactosidase levels were similar for all four strains and did not fluctuate significantly from the background levels obtained for wild-type T1g, which does not possess lacZ. Once the cultures entered stationary phase, lacZ expression increased between 17- and 42-fold compared to their basal levels at T₋₁ and were elevated an additional 50% 1 to 2 h into stationary phase. Thus, expression of the NRPS-PKS genes is linked to stationary-phase processes, consistent with the isolation of mutations in stationary-phase regulator genes.

TABLE 5.

Bacterial strains used in this study

Strain	Description	Source or reference
T1	Bacillus inaquosorum	Epicore Networks (U.S.A.) Inc.
SMY	Bacillus subtilis	Laboratory strain
T1g	amyE::rrnAp-gfp, T1 × AR13 DNA, Spr	This study
SSh1	sigH::erm, T1 × BKE00980 DNA, Emr Lnr	This study
SSb1	abrB::erm, T1 × BKE00370 DNA, Emr Lnr	This study
AR13	amyE::rrnAp-gfp Spr	BGSC
BKE00980	sigH::erm	BGSC
BKE00370	abrB::erm	BGSC
D4	Vibrio parahaemolyticus isolate 13-306 PirAB	4
A3	Vibrio parahaemolyticus isolate 13-028 PirAB	4
ATCC 17802	Vibrio parahaemolyticus	James Kaper
A1-20	orfDΩTnLacJump (orfD::lacZ) Spr	This study
A2-18	oppAΩTnLacJump Spr	This study
A2-30	orfBΩTnLacJump Spr	This study
A3-41	orfBΩTnLacJump (orfB::lacZ) Spr	This study
A8-11	spo0AΩTnLacJump Spr	This study
A10-67	orfBΩTnLacJump Spr	This study
A11-79	orfCΩTnLacJump (orfC::lacZ) Spr	This study
A13-56	orf9ΩTnKRM Spr	This study
A20-86	orfAΩTnLacJump (orfA::lacZ) Spr	This study
A23-33	orfFΩTnLacJump Spr	This study
S2-30	orfBΩTnKRM Spr	This study
S11-14	orfBΩTnKRM Spr	This study
S12-29	orfBΩTnKRM Spr	This study
S21-28	orfCΩTnKRM Spr	This study
S21-46	orfAΩTnKRM Spr	This study
S31-22	yoaZΩTnKRM Spr	This study
S35-30	orfEΩTnKRM Spr	This study
A20-86-A	abrB::erm orfAΩTnLacJump (orfA::lacZ), A20-86 × BKE00370 DNA, Emr Lnr Spr	This study

TABLE 6.

β-Galactosidase activities of T1 strains at different stages of growth^a

Strain	Relevant genotype	β-Galactosidase activity (nmol/min/OD600) at the following growth stageb :
		T−1	T0	T1
T1g		167 ± 29	108 ± 39	206 ± 79
A20-86	orfA::lacZ	140 ± 51	5,805 ± 158	8,660 ± 610
A3-41	orfB::lacZ	224 ± 53	3,898 ± 35	5,941 ± 180
A11-79	orfC::lacZ	180 ± 59	4,670 ± 1,060	7,012 ± 650
A1-20	orfD::lacZ	115 ± 24	3,905 ± 714	5,440 ± 124
A20-86-A	orfA::lacZ abrB::erm	7,288 ± 498	5,888 ± 296	7,240 ± 260

^aThe strains were grown in 2216 broth, and β-galactosidase levels were measured as described in Materials and Methods.

^bT₋₁, OD₆₀₀ of ∼1; T₀, OD₆₀₀ of ∼3; T₁, OD₆₀₀ of 4 to 5.

The orf1-orf2 intergenic region.

Examination of the 117-bp region between orf1 and orf2 revealed several potential promoter and regulatory sequences. We identified a sequence adjacent to the predicted orf1 stop codon having the capacity to form a 30-base mRNA hairpin structure (ΔG = −21 kcal/mol), followed by several consecutive U residues, which likely serve as a transcription terminator (Fig. 6). The sequences between 36 and 57 bp upstream from the orf2 initiation codon (Fig. 6) shared homology with σ^A (26) and σ^H (27) consensus sequences and may be used for orf2 transcription initiation. The loss of D4-inhibiting activity by strain SSh1 (Fig. 3), which contains an insertion in sigH, the σ^H structural gene, supports the involvement of σ^H in NRPS-PKS gene expression. Similarly, the partial loss of inhibiting activity observed for spo0A mutant A8-11 in the overlay assay also indicates a role for Spo0A in activating NRPS-PKS expression. A putative Spo0A binding site was identified 71 bp upstream of the σ^A and σ^H consensus sequences, overlapping the last two orf1 codons; this sequence includes internal G and C residues critical for Spo0A binding in B. subtilis 168 (21) (Fig. 6). The location of this site could allow for Spo0A-dependent activation of either σ^H or σ^A polymerases.

FIG 6.

The orf2 promoter region. Putative regulatory and promoter features for the DNA sequence between the last 13 codons and first 4 codons of the orf1 and orf2 open reading frames, respectively, are shown. Also shown are the consensus sequences and recognition sites for σ^A (double underlined), σ^H (single underlined), and Spo0A (boxed). Bases that agree with consensus sequences are capitalized.

Involvement of AbrB in NRPS-PKS expression.

AbrB is a key transitional-phase regulator in B. subtilis 168 that controls the expression of more than 100 stationary-phase genes, including those involved in antibiotic production (28). To examine a role for AbrB in NRPS-PKS expression, we constructed abrB mutant SSb1(abrB::erm). Like its parent T1 strain, SSb1 was found, using the overlay assay, to inhibit D4 growth (Fig. 3). A cell-free supernatant prepared from a stationary-phase SSb1 culture delayed D4 growth approximately 22 h, 4-fold longer than the delay observed for the cell-free T1 supernatant (Fig. 7, including the inset). Furthermore, the β-galactosidase activity of abrB mutant A20-86A (orfA::lacZ) was 52-fold higher than that of A20-86, its abrB⁺ parent, during mid-exponential phase (T₋₁) and comparable to the elevated levels in the transition (T₀) and stationary phase (T₁) (Table 6). Therefore, AbrB plays a role in controlling NRPS-PKS expression in T1.

FIG 7.

Deletion of abrB results in enhanced anti-D4 activity. The growth of an untreated D4 culture (●) and cultures mixed with an equal volume of cell-free supernatant fractions prepared from overnight cultures of wild-type T1 (▲) or SSb1 (abrB mutant) (■), as described in Materials and Methods, is shown. The dotted line indicates the anticipated growth between 12 and 26 h. The graph shows the results for a representative experiment, and standard deviations were <10% and were omitted for clarity. (Inset) Photograph of 26-h cultures. (A) D4 alone; (B) D4 treated with cell-free T1 supernatant; (C) D4 treated with cell-free SSb1 supernatant.

DISCUSSION

Our preliminary studies showed that B. inaquosorum strain T1 inhibits the growth of VirAP toxin-producing V. parahaemolyticus (S. E. Avery, A. M. Hise, and H. J. Schreier, unpublished data). Analysis of the T1 genome revealed that T1 is a strain of B. inaquosorum (H. J. Schreier, unpublished data) which is distinguishable from closely related B. subtilis and B. spizizenii by a 42.6-kb DNA region having the capacity to encode NRPSs and PKSs (19). In the present study, we used transposon mutagenesis to determine that this region is responsible for producing the inhibitory activity, and results from overlay assays and cell-free culture supernatant experiments indicate that this activity is from a secreted stationary-phase product whose synthesis is under the control of key stationary-phase regulators. Our study provides insight into the expression and control of the NRPS-PKS region found in B. inaquosorum and B. stercoris (19). Similar, although not identical, NRPS-PKS gene clusters have been identified in other Bacillus strains and species, including B. subtilis strain CW14 and Bacillus atrophaeus (Schreier, unpublished), which may be regulated in the same manner.

The synthesis and secretion of antimicrobial compounds by members of the genus Bacillus during the transition to stationary and early stationary phase are one strategy used by this group of bacteria to compete with other organisms for reduced resources as a prelude to sporulation (29). The elevated expression observed for NRPS-PKS genes by T1 during these periods (Table 6) is consistent with this strategy and, with their low or negligible expression during growth, can explain the results for the D4 and T1 coculture experiments. During exponential growth (i.e., 3 h after inoculation), the inability of any of the T1 treatments to inhibit D4 growth is likely the consequence of low-level inhibitor production during this period. Throughout the subsequent 21 h, however, the entry of T1 cultures into late exponential and early stationary phase resulted in as much as a 62-fold increase in NRPS-PKS gene expression (e.g., for strain A20-86; Table 6), with the accompanying synthesis and accumulation of the inhibitor. This resulted in a 700- to 1,200-fold decrease in D4 levels observed for growth in the presence of a 100-fold excess of T1 (Table 1). On the other hand, the absence of inhibition observed for D4 in cocultures of T1 at 1:1 and 10:1 ratios could be explained by differences in the generation times between D4 and T1 in 2216 broth, which were 40 and 60 min, respectively. For these cultures, D4 likely outgrew and attained stationary phase before T1 accumulated an amount of inhibitor sufficient to influence D4 growth. Furthermore, at the 1:1 and 10:1 ratios, D4 may be able to mitigate the effect of the inhibitor, as was observed by its ability to resume growth after treatment with cell-free T1 culture supernatants (see below).

Evidence that D4 was sensitive to a product of the NRPS-PKS region was obtained from overlay, coculture, and cell-free supernatant experiments, since inhibition did not occur using mutants containing insertions within NRPS-PKS region genes. Moreover, the inability of cell-free supernatants from orfB mutant A3-41 to inhibit D4 growth indicated that the effect was directly due to orfB and downstream genes and was not a consequence of toxic stationary-phase by-products, e.g., volatile organic or inorganic compounds (30), or nutrient depletion of the spent medium. While B. inaquosorum has the coding capacity to synthesize the peptide antibiotics subtilosin and bacilysin and the bacteriocin surfactin (19), the growth of V. parahaemolyticus strains A3, D4, and ATCC 17802 was not inhibited by either T1 NRPS-PKS mutants or B. subtilis strain SMY (GenBank accession number CP050532), which also possesses genes for these secondary products, arguing that none of these secondary metabolites are involved in T1 inhibitory activity. Furthermore, the B. inaquosorum gene clusters responsible for the antibacterial bacillomycin F and the antifungal fengycin are not associated with the NRPS-PKS region identified in our study (19), ruling out their activities as well.

Transcript mapping indicated that the NRPS-PKS region encodes one polycistronic message extending across all 15 genes, forming an operon, although our analyses cannot rule out the possibility that transcription of a subset of these genes might occur from internal promoters. The absence of a transcript spanning orf1 and orf2 indicated that initiation likely occurs from the first gene of the operon, orf2, and requires σ^H for activity (based on the results of the overlay assay of sigH mutant SSh1; Fig. 3), which may bind to sequences upstream of orf2 (Fig. 6). Recent RT-PCR studies have detected NRPS-PKS mRNA in stationary-phase cultures of sigH mutant SSh1 (S. P. Ruzbarsky, unpublished data), suggesting the involvement of another polymerase for transcription. One candidate is σ^A, since σ^A promoter sequences appear to overlap the putative σ^H promoter sequence (Fig. 6) and would provide a mechanism for NRPS-PKS expression during exponential growth. Promoters transcribed using both sigma factors under different physiological conditions in B. subtilis have been noted (27, 31).

Similar to many B. subtilis stationary-phase products and processes (29), the NRPS-PKS genes of T1 were found to be under the control of oppA, spo0A, and abrB, genes encoding transitional- and stationary-phase regulators. Disruption of oppA, the first gene of the opp operon, in mutant A2-18 resulted in a loss of inhibitory activity. In B. subtilis 168, this operon encodes an oligopeptide permease that functions as a receptor in signaling (22, 32), and impairment of oppA results in a loss of production of bacilysin (33), a nonribosomally synthesized antimicrobial. The opp operon is linked to the ComA competence response regulator and CSF (PhrC), the competence and sporulation factor, which also participates in B. subtilis quorum sensing (32). When CSF accumulates extracellularly due to a high cell density, the production of bacilysin is stimulated (34). Any involvement of the Com system or CSF in the control of NRPS-PKS expression through oppA is yet to be determined.

Isolation of a mutant with an insertion in spo0A (the A8-11 mutant) having decreased inhibitory activity indicated a role for this regulator in NRPS-PKS control. In B. subtilis 168, the spo0A gene product, Spo0A, is a master transcriptional regulator of early-stationary-phase processes, including NRPS and PKS gene activation, and the development of spores (35). In its phosphorylated state, Spo0A activates transcription initiation at both σ^A and σ^H promoters and is responsible for indirectly activating sigH transcription by repressing AbrB, the global transition state regulator (29, 35). While we did not address the nature of Spo0A participation in NRPS-PKS expression, we identified a putative Spo0A binding site upstream of orf2 that could be used to activate σ^H-dependent transcription and, possibly, σ^A-dependent transcription. Whether Spo0A is directly involved in activating NRPS-PKS transcription or is indirectly involved by elevating σ^H levels remains to be established.

Like Spo0A, AbrB is essential for controlling NRPS-PKS expression, since the level of orfA expression in abrB mutant A20-86-A was found to be elevated 52-fold during exponential growth compared to the level of orfA expression in its parent strain, A20-86 (abrB⁺) (Table 6). Enhanced activity was also found for the cell-free supernatant fraction prepared from the abrB mutant, SSb1 (Fig. 7). The involvement of AbrB in controlling the NRPS and PKS genes in Bacillus spp. is well documented, and AbrB can act by binding either directly to promoter sequences, interfering with transcription initiation, or indirectly through the repression of sigH (29). While the A/T-rich character of the orf2 promoter is typical of AbrB binding sites, TGGNA and TNCCA motifs associated with AbrB (36) are absent.

How is V. parahaemolyticus affected by the T1 inhibitor? The mode of action of many Bacillus NRPS- and PKS-derived antimicrobials is through membrane perturbation or depolarization (15), e.g., by fengycin (37) and iturin A (38); blockage of peptidoglycan biosynthesis, e.g., by bacilysin (30); or selective inhibition of protein synthesis, e.g., by difficidin (30). Regardless of the mechanism, both coculture and cell-free supernatant experiments demonstrated that D4 is capable of overcoming the T1 NRPS-PKS inhibitory activity, with growth levels reaching those obtained for untreated cultures. Along with D4, the bacteriostatic nature of the T1 inhibitor was also observed for A3 and ATCC 17802 strains, which yielded small colonies in overlay assay clearance zones after long-term incubation (Schreier, unpublished). Isolates obtained from these zones continued to remain sensitive to T1 (Schreier, unpublished), suggesting an ability to either degrade the inhibitor—Vibrio spp. secrete several classes of proteases, some of which may have activities against lipopeptides (39)—or modify their membrane components to be less sensitive to inhibitor activity as part of an induced stress response (40). The decreased sensitivity of A3 to T1 inhibition compared to the sensitivity of D4 and ATCC 17802 observed in the overlay assay might be explained by increased activities for either of these processes. While little is known about the stability and persistence of the NRPS and PKS products in the environment, it is also possible that factors other than enzymatic degradation, e.g., pH and medium components, may decrease their efficacy over time (41).

The use of Bacillus spp. as biological control agents has received a substantial amount of attention due to their antimicrobial characteristics and safety (13, 42), along with their spore-forming capability, which is advantageous for long-term storage (43). B. subtilis strains have been effective in controlling disease outbreaks due to Vibrio pathogens in a variety of aquaculture species, including shrimp (44–46), in addition to providing probiotic benefits (13). The ability of B. inaquosorum strain T1 to inhibit the growth of AHPND-causing Vibrio spp. suggested its potential for use as a tool in the prevention and the treatment of AHPND in shrimp aquaculture. Understanding the genetic basis of the inhibitory activity and its regulatory mechanisms allows for the development of T1 strains having desirable properties, such as enhanced NRPS-PKS expression during growth that is acquired by disabling abrB. Such strains could be used in aquaculture applications to complement probiotic strains having different antimicrobial activities. Our preliminary studies have shown that the daily addition of freshly prepared cell-free SSb1 supernatants to D4 cultures resulted in the continuous cessation of D4 growth over the course of a 72-h treatment (S. E. Avery, S. P. Ruzbarsky, and H. J. Schreier, unpublished data). Thus, supplementing feed with abrB mutant SSb1 might provide system water and animal microbiomes with a constant source of inhibitory activity, bypassing stability or degradation issues and restricting pathogen growth before it reaches virulent levels. Studies aimed at evaluating T1 and its derivatives to protect against AHPND-causing Vibrio in shrimp aquaculture systems are ongoing.

MATERIALS AND METHODS

Bacterial strains and culture media.

The bacterial strains used in this study and their sources are listed in Table 5. Strains were grown in Zobell 2216 marine broth (HiMedia Laboratories), tryptic soy broth (Sigma-Aldrich) or agar supplemented with 2% NaCl (TSB2 or TSA2, respectively), and lysogeny broth (LB) agar (47) with antibiotic, when appropriate. SOC and 2× YT have been described previously (47). Antibiotics were added to the media at concentrations of 100 μg spectinomycin (Sp)/ml, 1 μg erythromycin (Em)/ml, and 10 μg lincomycin (Ln)/ml. Strain T1g was constructed by transforming strain T1 by electroporation (described below) with DNA from B. subtilis strain AR13 (amyE::gfp) and selecting for spectinomycin resistance (Sp^r). The incorporation of gfp into amyE was confirmed by the loss of amylase activity on starch agar and by an increased size of amyE, as determined by PCR. The disruption of amyE did not affect D4-inhibitory activity, as determined by overlay or coculture assays (see below); T1 and T1g were interchangeable, and their choice for usage depended on the experiment. Strains SSh1 (sigH::erm) and SSa1 (abrB::erm) were constructed by transforming strain T1 with DNA from B. subtilis strains BKE00980 (sigH::erm) and BKE00370 (abrB::erm), respectively, selecting for Em resistance (Em^r) and Ln resistance (Ln^r); insertion of the erm cassette was confirmed by PCR. Vibrio parahaemolyticus strains D4, isolated from Mexico, and A3, isolated from Vietnam, were provided by Kathy Tang-Nelson, University of Arizona, and the presence of pirAB was confirmed by PCR.

Soft-agar overlay assay.

To evaluate inhibitory activity, 2.5 μl from an overnight B. inaquosorum culture was spotted onto 2216 agar, and the culture was incubated at 37°C for 18 h. In a chemical fume hood, uncovered plates were placed in a Pyrex dish along with a reservoir of chloroform (20 to 30 ml); the dish was then covered with plastic wrap to generate a chloroform atmosphere and facilitate cell death. After 30 min, the plates were removed, covered, and set at room temperature for 30 min to allow for chloroform evaporation. For each plate, 3 ml of semisolid 2216 agar (2216 broth with 0.75% Bacto agar) was heated until liquefied, cooled to 42°C, inoculated with 10 μl of an overnight V. parahaemolyticus culture, and immediately poured over the 2216 agar surface. The plates were incubated at 28°C overnight and examined for zones of clearance.

Coculture growth experiments.

Growth experiments examining the effect of strain T1g or mutant A3-41 on D4 growth were done by combining overnight cultures of D4 with T1g or A3-41 at various cell densities in fresh liquid 2216 broth and monitoring D4 growth by quantitative PCR (qPCR) (see below). The qPCR targets used to assess the densities of T1g, mutant A3-41, and strain D4 were gfp, amyE, and toxR, respectively, and the primers used are listed in Table 4. The inocula used for the cocultures were based on the number of CFU per milliliter for each strain in 2216 broth, which ranged from 2.0 × 10¹⁰ to 3.0 × 10¹⁰ CFU/ml for D4 and 2.0 × 10⁹ to 3.0 × 10⁹ CFU/ml for strains T1g and A3-41. Overnight cultures of T1g, A3-41, and D4 were prepared in 2216 broth and were used to inoculate 50 ml of 2216; overnight cultures of T1g and A3-41 were rinsed and suspended in fresh 2216 prior to their use as inoculants. The initial cell density for D4 was 2.0 × 10⁴ CFU/ml, and the initial cell densities for T1g and A3-41 ranged from 2.0 × 10⁴ to 2.0 × 10⁶ CFU/ml, to generate T1g- or A3-41-to-D4 ratios of 1:1, 10:1, and 100:1. Cultures were grown at 28°C and 240 rpm for 24 h. At 3 h after inoculation, 10 ml of each culture was centrifuged at 4°C and 4,000 × g for 10 min; at 24 h after inoculation, 0.5 ml of each culture was centrifuged at 4°C and 10,000 × g for 5 min. To assist in the recovery of low-density D4 cultures at the 3-h time point, autoclaved Aeromonas hydrophila was added to each sample to a final concentration of 10⁵ CFU/ml prior to centrifugation. Extraction of DNA from each sample was done using a Wizard genomic DNA purification kit (Promega) following the manufacturer’s specifications.

Cell-free culture supernatant experiments.

Overnight cultures grown in 2216 broth were centrifuged at 4°C and 5,000 × g for 10 min, and supernatant fractions were passed through a 0.2-μm-pore-size filter. The filtered supernatants were then added to fresh 2216 medium in 50 ml sidearm flasks to a final concentration of 50% in a total volume of 12 ml. The control flask received 12 ml of fresh 2216 broth. Each culture was then inoculated with D4 at a concentration of 2.0 × 10⁴ CFU/ml, and the cultures were grown at 28°C and 240 rpm, with growth being monitored using a Klett-Summerson colorimeter with a Wratten 54 filter. Two 1-ml samples were taken from each flask at 5.5 h after inoculation, followed by DNA extraction, as described above.

qPCR analysis.

qPCR was performed with 10-μl reaction mixtures containing 5 μl of PefeCTa SYBR green Fastmix (Quanta), 3.5 μl PCR-certified water, 0.25 μl of 1/10-diluted forward and reverse primers (0.5 μM each), and 1 μl of the sample to be quantified. All qPCRs were performed using an Applied Biosystems 7500 Fast real-time PCR machine. PCR-certified water was used as the no-template control to monitor for contamination. The primers used for qPCR are listed in Table 4. The DNA template for standard curves was prepared by PCR, performed with 50-μl reaction mixtures containing 25 μl Taq PCR Mastermix (Qiagen), 19 μl PCR-certified water, 2 μl of 1/10-diluted forward and reverse primers (0.5 μM each) for toxR, gfp, and amyE (Table 4), and 2 μl of DNA template. Chromosomal DNA from D4, T1g, and T1 was used as the DNA template for toxR, gfp, and amyE, respectively. The PCR products were then purified using a MinElute PCR purification kit (Qiagen), according to the manufacturer’s specifications, and the concentration (in nanograms per milliliter) was determined using a Qubit fluorometer (Thermo Fisher Scientific).

Plasmids.

mariner-derived himar1 delivery vectors pDP384 [TnKRMspec amp mls mariner-Himar1 ori(Ts)_Bs] and pEP4 [TnLacJump spec amp mls mariner-Himar1 ori(Ts)_Bs] (48) were used for the transposon mutagenesis of strain T1 and were provided by Daniel Kearns, Indiana University. The plasmids harbor a temperature-sensitive B. subtilis origin of replication [ori(Ts)_Bs]; an erythromycin resistance (Em^r) gene, located outside of the transposon sequences; and an Sp^r gene, contained within transposon sequences. Growth at the restrictive temperature and selection for Sp^r resulted in the identification of cells containing chromosomal insertions. Plasmid pEP4 differs from pDP384 by the presence of a promoter-less lacZ gene within transposon sequences, which is expressed when inserted downstream of an active transcription start site (48). Plasmids were purified using a Wizard DNA cleanup kit (Promega) according to the manufacturer’s standards.

Electroporation of T1 with delivery vectors and mutagenesis.

T1 was prepared for electroporation by growing T1 cells in 250 ml of 2× YT to an optical density at 600 nm (OD₆₀₀) of 0.8 at 37°C and 240 rpm, washing the cells three times in ice-cold 10% glycerol, suspending the cells in 1 ml ice-cold 10% glycerol, and storing the suspension at −80°C in 200-μl aliquots. For electroporation, one aliquot was mixed with purified plasmid (∼7 μg/ml), and then the mixture was incubated on ice for 5 min. The cells were then transferred to a 2-mm electroporation cuvette, and electroporation was done using the StA program of a Bio-Rad MicroPulser electroporation apparatus (∼1.8 kV for 2.5 ms). After electroporation, 0.5 ml SOC was added and the cells were incubated at 28°C with aeration for 2 h, followed by plating onto LB-Sp-Em agar. The presence of the transposase gene in several Sp^r Em^r transformants was confirmed by PCR, and one transformant was selected for mutagenesis. After growth in LB-Sp medium for 18 h at 42°C, the cells were plated onto LB-Sp agar at a dilution of 10⁻⁶. Approximately 3,000 transposon transformants were screened for decreased or lost D4-inhibitory activity by the presence of a reduced clearance zone (relative to that for a wild-type control) or the absence of a clearance zone, respectively, using an overlay assay in which TSA2 was substituted for 2216 agar, as described above.

Mutant characterization and identification of transposon insertion site.

Confirmation that Sp^r was due to transposon insertion was done by PCR using primer set 2569R/2570R (Table 4), followed by visualization via agarose gel electrophoresis. PCR was carried out in 25-μl reaction mixtures with Qiagen Taq polymerase and a Bio-Rad S1000 thermal cycler for 3 min at 94°C, followed by 30 cycles of 1 min at 94°C, 1 min at 55°C, and 1 min at 72°C and a final step at 72°C for 10 min. To ensure that a mutant had only one insertion, chromosomal DNA was prepared from that mutant and retransformed into strain T1 by electroporation, selecting for Sp^r. Insertion site identification was determined by amplification of the transposon and adjacent DNA using an inverse PCR strategy, as follows. One microgram of chromosomal DNA from the T1 mutants was digested with TaqI for 2 h at 37°C in 20 μl, and 1 μl from this reaction mixture was then ligated using a T4 DNA rapid ligation kit (Thermo Fisher Scientific) for 5 min at room temperature according to the manufacturer’s specifications. The ligation mixture was then used in a PCR mixture with primers 2569/2570 (Table 4) and Phusion DNA polymerase (New England BioLabs). PCR was carried out with 50-μl reaction mixtures for 3 min at 98°C, followed by 30 cycles of 1 min at 98°C, 1 min at 55°C, and 1 min at 72°C and a final step at 72°C for 10 min. The PCR products were purified using a Wizard PCR Preps DNA purification system (Promega), and DNA sequencing was performed using primer 2569, which anneals to transposon sequences adjacent to the insertion site. Comparison of DNA sequences interrupted by the transposon to database sequences was done by BLAST (49) and BLASTp (50) analyses.

Isolation of RNA.

Overnight cultures of T1 grown in 2216 broth were used to inoculate 10 ml of 2216 broth, and at ∼1 to 2 h before reaching stationary phase (T₋₁) and after reaching stationary phase (T₁), 1-ml samples were removed and centrifuged at 10,000 rpm and 4°C for 5 min, discarding the supernatant fraction. The cell pellets were suspended in 0.3 ml of 10 mM Tris-HCl (pH 8.0) and 10 mg lysozyme/ml (Sigma-Aldrich), and the suspension was incubated at 37°C for 1 h. The TRIzol reagent (1 ml; Thermo Fisher Scientific) was added, followed by the addition of 0.2 ml cold chloroform, and incubation was continued at room temperature for 2 to 3 min, and then the mixture was centrifuged at 14,000 × g and 4°C for 20 min. After centrifugation, 0.4 ml of the resulting aqueous phase was added to 0.5 ml isopropanol, and the mixture was incubated at room temperature for 10 min to precipitate the RNA. The precipitant was washed with 75% ethanol, dried, and suspended in 50 μl diethyl pyrocarbonate-treated water. RNA quality was assessed by ethidium bromide-agarose gel electrophoresis.

RT-PCR for transcription mapping experiments.

T1 RNA was treated with RNase-free DNase I (Thermo Fisher Scientific), prior to being used for reverse transcriptase (RT) PCR to eliminate residual genomic DNA; 50 to 100 ng RNA was used for each reaction mixture. RT-PCR was carried out using gene-specific primers (Table 4) at a final concentration of 0.5 μM and a SuperScript IV one-step RT-PCR system (Thermo Fisher Scientific). Reactions were performed using a reverse transcription step at 50°C for 10 min and a 2-min RT inactivation step at 98°C, followed by 25 to 30 cycles at 98°C for 10 s, 56°C for 10 s, and 30 s/kb at 72°C, with a final extension at 72°C for 5 min. Each primer set (Table 4) was tested with T1 chromosomal DNA as the template using Taq polymerase.

β-Galactosidase assays.

Overnight cultures were used to inoculate 2216 medium in sidearm flasks, and the cultures were grown at 37°C and 240 rpm, with the cell density being monitored by determination of the OD₆₀₀. Cells (1 ml) were harvested at mid-exponential phase (T₋₁), at the onset of stationary phase (T₀), and 1 to 2 h into stationary phase (T₁) and centrifuged at 10,000 × g and 4°C for 5 min, and the cell pellets were suspended in 1 ml modified Z buffer (51). Sodium dodecyl sulfate was added (final concentration, 0.05%), and the mixture was vortexed and then incubated at 37°C for 5 min, after which 0.2 ml of o-nitrophenyl-β-galactoside (4 mg/ml) was added and the components were mixed and incubated at 37°C for an additional 30 to 60 min. The reaction was stopped by the addition of 0.25 ml 2 M sodium carbonate, and the mixture was vortexed and placed on ice. After centrifugation at 10,000 × g at room temperature for 10 min, the OD₄₂₀ was measured. Specific activity is given as the number of nanomoles per minute per OD₆₀₀ unit.

Statistical analysis.

Analysis of the data was performed using one-way analysis of variance (ANOVA), and the significance level was a P value of <0.01.

Accession number(s).

The DNA sequence of the entire NRPS-PKS region was deposited in GenBank under accession number MT366812.