Development of a sensor for disulfide bond formation in diverse bacteria

ABSTRACT

In bacteria, disulfide bonds contribute to the folding and stability of proteins important for processes in the cellular envelope. In Escherichia coli, disulfide bond formation is catalyzed by DsbA and DsbB enzymes. DsbA is a periplasmic protein that catalyzes disulfide bond formation in substrate proteins, while DsbB is an inner membrane protein that transfers electrons from DsbA to quinones, thereby regenerating the DsbA active state. Actinobacteria including mycobacteria use an alternative enzyme named VKOR, which performs the same function as DsbB. Disulfide bond formation enzymes, DsbA and DsbB/VKOR, represent novel drug targets because their inhibition could simultaneously affect the folding of several cell envelope proteins including virulence factors, proteins involved in outer membrane biogenesis, cell division, and antibiotic resistance. We have previously developed a cell-based and target-based assay to identify molecules that inhibit the DsbB and VKOR in pathogenic bacteria, using E. coli cells expressing a periplasmic β-Galactosidase sensor (β-Gal^dbs), which is only active when disulfide bond formation is inhibited. Here, we report the construction of plasmids that allows fine-tuning of the expression of the β-Gal^dbs sensor and can be mobilized into other gram-negative organisms. As an example, when expressed in Pseudomonas aeruginosa UCBPP-PA14, which harbors two DsbB homologs, β-Gal^dbs behaves similarly as in E. coli, and the biosensor responds to the inhibition of the two DsbB proteins. Thus, these β-Gal^dbs reporter plasmids provide a basis to identify novel inhibitors of DsbA and DsbB/VKOR in multidrug-resistant gram-negative pathogens and to further study oxidative protein folding in diverse gram-negative bacteria.

IMPORTANCE

Disulfide bonds contribute to the folding and stability of proteins in the bacterial cell envelope. Disulfide bond-forming enzymes represent new drug targets against multidrug-resistant bacteria because inactivation of this process would simultaneously affect several proteins in the cell envelope, including virulence factors, toxins, proteins involved in outer membrane biogenesis, cell division, and antibiotic resistance. Identifying the enzymes involved in disulfide bond formation in gram-negative pathogens as well as their inhibitors can contribute to the much-needed antibacterial innovation. In this work, we developed sensors of disulfide bond formation for gram-negative bacteria. These tools will enable the study of disulfide bond formation and the identification of inhibitors for this crucial process in diverse gram-negative pathogens.

INTRODUCTION

More than three decades ago, the enzymes that introduce disulfide bonds were discovered, thanks to a protein fusion that was developed as a complementary approach to studying membrane protein topology (1, 2). Froshauer et al. generated hybrid proteins of the membrane protein MalF with β-Galactosidase (β-Gal, LacZ). Cytoplasmic domain fusions exhibited high levels of β-Gal activity, whereas the periplasmic domain fusions expressed low activity. This approach complemented the use of PhoA fusions, which behaved oppositely to β-Gal and strengthened the evidence of membrane topology (1, 3, 4). One fusion, MalF-LacZ102 (aka β-Gal^dbs for disulfide bond-sensitive β-Gal), displayed low levels of β-Gal and was found to be an in-frame hybrid of 198 amino acids of MalF attached by a three-amino acid linker (Gly-Asp-Pro) to the residue 8 of β-Gal (Fig. 1) (1). The fusion occurred at the third transmembrane segment in the middle of the long hydrophilic domain, thus placing β-Gal in the periplasm (1, 5). When β-Gal^dbs is expressed in the periplasm of cells with an active disulfide bond formation system, β-Gal^dbs is inactivated with disulfide bonds, while an inactive disulfide bond machinery would produce active β-Gal^dbs (Fig. 1). This fusion has been a crucial tool for the seminal discovery of the disulfide bond formation pathway in Escherichia coli, through two decades later to identify inhibitors of this process.

Fig 1.

β-Gal^dbs (MalF-LacZ102) monitors disulfide bond formation in the bacterial periplasm. (A) In E. coli, DsbA and DsbB participate in introducing aberrant disulfide bonds (represented as yellow lines) into periplasmic β-Galactosidase; hence, misfolding and inactivating it. (B) The removal or inhibition of either dsbA or dsbB disrupts the formation of disulfide bonds, allowing β-Galactosidase to be folded.

Disulfide bond formation plays a role in protein folding for both eukaryotes and prokaryotes. In bacteria, disulfide bonds contribute to the folding and stability of proteins involved in crucial cellular processes in the cell envelope (6–8). In E. coli, disulfide bond formation is catalyzed by DsbA and DsbB enzymes, which work together to introduce disulfide bonds into many proteins (2, 9). DsbA is a periplasmic protein, a member of the thioredoxin family, which catalyzes the formation of disulfide bonds into substrate proteins through its Cys-X-X-Cys active site (2). DsbB is a membrane protein that regenerates DsbA’s activity by transferring electrons to quinones (9–12). Some pathogenic bacteria, like Pseudomonas aeruginosa, harbor more than one copy of the DsbA-DsbB proteins (7, 13); whereas, actinobacteria, cyanobacteria, and δ-proteobacteria use an alternative enzyme named VKOR (for vitamin K epoxide reductase) that performs the same function as DsbB (14). DsbB and VKOR share no protein sequence identity, but they exhibit similar structural features and contain a quinone cofactor to generate a disulfide bond (15, 16).

The viability of mutants lacking disulfide bond-forming machinery varies from organism to organism. E. coli dsb mutants are viable aerobically but not anaerobically (17). Overall, disulfide bond formation is required for virulence but not for in vitro growth of gram-negative bacteria (6, 7); whereas, it is essential in actinobacteria (18–20). Disulfide bond-forming enzymes represent a compelling new drug target because their inhibition could simultaneously affect several proteins localized in the cell envelope, including virulence factors, proteins involved in outer membrane biogenesis, cell division, and antibiotic resistance (6, 7, 21–23). Thus, we have previously developed a cell-based and target-based assay to find molecules that inhibit the membrane proteins, DsbB and VKOR, of pathogenic bacteria (24, 25). This assay uses E. coli ΔdsbB mutant expressing the β-Gal^dbs in which dsbB function is complemented with a plasmid carrying either dsbB or vkor genes from pathogens (24). We then perform parallel screens of compounds to look for inhibitors of each of the enzymes (24–26). The screens are performed in parallel to provide reciprocal controls that eliminate inhibitors that influence β-Gal activity by acting directly on E. coli DsbA or affecting membrane protein assembly as those molecules would likely appear as inhibitors of both DsbB and VKOR-expressing strains. We have identified several classes of DsbB and VKOR inhibitors using this method (24, 25, 27) and we have sought to expand the small molecule search by using cell-based and target-based approaches with gram-negative organisms for which antimicrobial resistance is posing a threat.

Here, we report the construction of a series of biosensor plasmids that allow fine-tuning of the expression of β-Gal^dbs and are mobilizable into other gram-negative bacteria. A plasmid carrying β-Gal^dbs introduced into P. aeruginosa UCBPP-PA14 behaves similarly as in E. coli. The strains expressing β-Gal^dbs respond to molecular inhibition of the two DsbB proteins and provide a basis to look for novel inhibitors of both enzymes using P. aeruginosa cells. These vectors represent tools to further study oxidative protein folding in other gram-negative pathogens and to identify inhibitors of DsbAB proteins against these organisms.

RESULTS

Use of tightly regulated promoters to control β-Gal^dbs expression

Previous attempts to move the β-Gal^dbs into other organisms including P. aeruginosa were unsuccessful because moving the plasmid resulted in toxicity, which prevented growth of transconjugants. Dwyer et al. have shown that periplasmic LacZ is toxic to E. coli, likely due to the presence of 16 cysteines, which cause aberrant disulfide bonds introduced by DsbA, rendering misfolded β-Gal (5). The original construct (Fig. 2A, pNG102) includes the maltose-binding protein, MalE, and β-Gal^dbs under the maltose promoter (P_mal) (1), which was then inserted into the chromosome at the lambda att site (28). We reasoned that the difficulty in introducing the β-Gal^dbs construct was perhaps a combination of the basal expression levels and/or the co-expression of MalE upstream of β-Gal^dbs. We thus cloned only β-Gal^dbs into four low-copy number plasmids (p15A ori) with promoters and regulators that have been engineered to have low background, low cross-reactivity, and high dynamic range (29). The promoters are inducible with cuminic acid (Cuma, PL23), vanillic acid (Van, PL25), 2,4-diacetylphophloroglucinol (DAPG, PL26), and anhydrotetracycline (aTc, PL31).

Fig 2.

Fine-tuning levels of β-Gal^dbs sensor in E. coli. (A) β-Gal activity of the λ::P_mal-MalE-β-Gal^dbs expressed in E. coli ΔdsbB strain grown in M63 0.2% glucose supplemented with 50 µg/mL of essential amino acids (except Cys) grown at 30°C for 10 h and induced with varying concentrations of maltose. An aliquot of cells was then used to quantify β-Gal by following the hydrolysis of ONPG at 28°C using whole cells. (B) β-Gal activity of the P_CymR-β-Gal^dbs expressed in E. coli ΔdsbB strain induced with Cuma. (C) β-Gal activity of the P_VanR-β-Gal^dbs expressed in E. coli ΔdsbB strain induced with Van. (D) β-Gal activity of the P_PhlF-β-Gal^dbs expressed in E. coli ΔdsbB strain induced with DAPG. (E) β-Gal activity of the P_Tet-β-Gal^dbs expressed in E. coli ΔdsbB strain induced with aTc. Data represent the average ± SD of four independent experiments. Plasmid maps were created with BioRender.com.

We transformed the plasmids (PL23, PL25, PL26, PL31) into E. coli ΔdsbB mutant in which β-Gal^dbs would be active, and quantified the β-Gal activity in M63 minimal medium supplemented with 0.2% glucose, 50 µg/mL of all essential amino acids, except cysteine to prevent thiol rearrangement, and serial dilutions of inducers. We used whole cells to measure the velocity of o-nitrophenyl-β-galactoside (ONPG) hydrolysis without lysing cells (no chloroform-SDS step) because ONPG, while unable to enter the cytoplasm, is permeable to the outer membrane, and β-Gal^dbs is in the periplasm. All four plasmids (Fig. 2B and E) impart higher β-Gal activity compared to P_mal-β-Gal^dbs induced with maltose (Fig. 2A). To avoid overexpression and toxicity from P_mal-β-Gal^dbs, we mildly induced adding maltose in media containing glucose to weaken the expression of β-Gal (Fig. 2A). The four new vectors allow titration of β-Gal^dbs expression at desired levels to sensitize or strengthen the assay (Fig. 2). In addition, all the concentrations used to induce the β-Gal^dbs resulted in similar growth compared to the growth obtained in strains harboring the P_mal-β-Gal^dbs (Fig. S1). In the expression of the aTc-inducible β-Gal^dbs vector (PL31), the biosensor plasmid with the highest expression, we see a decrease in growth of the ΔdsbB mutant only at a very high induction with 10 µM aTc (Fig. S2).

Expression of β-Gal^dbs in P. aeruginosa UCBPP-PA14

P. aeruginosa is one of the six pathogens for which new antibacterial agents are most desperately needed (30, 31) and is often associated with extremely difficult-to-treat infections in immunosuppressed patients for whom current antibiotic treatments fail to work (32–35). Recalcitrant infections include acute pneumonia, ulcerative keratitis, bacteremia, urinary tract, intra-abdominal, chronic airway, and wound infections (32, 36).

Some P. aeruginosa virulence factors known to contain disulfide bonds include the pilin protein PilA (37, 38), the elastolytic metalloprotease elastase (LasB), exotoxin A (39–41), the chitin-binding protein (CbpD), the immunomodulating metalloprotease IMPa and proteases PaAP and PrpL, as well as type-III secretion components required for the delivery of the effectors ExoT and ExoU (37). Disulfide bond-forming enzymes in P. aeruginosa include two dsbA and two dsbB homologs. Disulfide bond formation is mainly driven by DsbA1, which is oxidized by both DsbB1 and DsbB2 proteins (13, 25, 37). The deletion of dsbA1 and the double deletion of dsbB1dsbB2 cause a decrease in virulence in pneumonia and keratitis mice models (25, 42). Thus, inhibitors of the disulfide bond-forming enzymes would affect the folding of several virulence factors.

To use the β-Gal^dbs biosensor system in other gram-negative bacteria, we subcloned the four regulators and promoters together with β-Gal^dbs into pJN105, a broad host range (BHR) plasmid. pJN105 plasmid harbors a gentamicin resistance cassette (43), a mob site to allow conjugal delivery from E. coli strains harboring RK2 transfer machinery (44), and a BHR origin of replication (pBBR) (43) that despite the origin similarity with plasmids of gram-positive bacteria, it resides in gram-negative bacteria and does not replicate by a rolling-circle mechanism (45).

We then moved the inducible plasmids (PL60, PL61, PL62, and PL63) into P. aeruginosa UCBPP-PA14 wild type as well as the single and double dsbB1 and dsbB2 mutant strains. Both P. aeruginosa DsbB proteins maintain oxidized DsbA1; hence, the single dsbB deletions are rescued by the remaining paralog and reactivate DsbA1 (13, 25). Indeed, high levels of β-Gal can be observed when we induce the expression of β-Gal^dbs in both the ΔdsbA1 and ΔdsbB1ΔdsbB2 mutants grown in M63 0.2% glucose medium supplemented with X-Gal (Fig. 3A). However, little to no activity is seen in the wild-type or single ΔdsbB1 or ΔdsbB2 mutants grown in X-Gal minimal media (Fig. 3A). When we quantified the β-Gal activity of the wild type and mutants carrying the biosensor inducible with Cuma (PL60), the Miller Units reflected the results seen in the X-Gal plates (Fig. 3B). Furthermore, when the ΔdsbB1ΔdsbB2 mutant was grown in different concentration of inducers, we observed a dose-dependent induction of β-Gal^dbs (Fig. 3C through F). The biosensors with the best inducible range were the Cuma- and aTc-inducible plasmids (Fig. 3C). The Van-inducible plasmid was the least induced (Fig. 3D). The Cuma- and aTc-inducible biosensors also displayed lower background activity in the absence of the inducer (depicted as 10⁻⁴, Fig. 3C through F) compared to what is observed in E. coli (Fig. 2C through F). Thus, the concentration of the inducer can be adjusted to the desired sensitivity of the assay when screening for inhibitors of the Dsb proteins (see below).

Fig 3.

Mobilizable β-Gal^dbs biosensors for gram-negative bacteria. (A) Phenotype of β-Gal^dbs biosensors in dsb mutants. Aliquots (10 µL) of overnight cultures were spotted onto M63 0.2% glucose supplemented with 120 µg/mL X-Gal and inducer (either 50 µM Cuma, 20 µM Van, 50 µM DAPG, or 50 nM aTc). Plates were incubated at 30°C for 2 d. (B) Low levels of β-Gal^dbs [limit of detection: 5 Miller Units (MU)] are detected in wild type, ΔdsbB1 and ΔdsbB2, where DsbA1 is still functional, while the ΔdsbA1 and double ΔdsbB1ΔdsbB2 mutants activate β-Gal due to lack of DsbA or its oxidation. P. aeruginosa UCBPP-PA14 wild type and dsb mutant strains carrying the biosensor plasmid PL60 were used to quantify β-Gal^dbs by ONPG hydrolysis in live cells that were grown for 18 h at 37°C in M63 0.2% glucose supplemented with 50µg/mL of essential amino acids (except Cys), antibiotics, and either 5 or 50 µM of Cuma to induce β-Gal^dbs. Data represent the average ± SD of at least two independent experiments. (C-F) The β-Gal activity is dose-dependent on the concentration of inducer in the double ΔdsbB1ΔdsbB2 mutant. β-Gal^dbs was measured as in (B) but with a serial dilution of inducers to induce β-Gal^dbs. Data represent the average ± SEM of at least two independent experiments, and non-linear regression was used to model the response. Plasmid maps were created with BioRender.com.

Inhibition of DsbB1 and DsbB2 using P. aeruginosa β-Gal^dbs sensor

As a proof of concept for a small molecule screen, we tested the ΔdsbB1 and ΔdsbB2 mutants against an inhibitor of the two proteins that we previously discovered, compound 12, a dichloro-pyridazinone (24). We used both agar (Fig. 4A) and liquid (Fig. 4B) assays to determine the inhibition of DsbB1 and DsbB2 proteins independently using the Cuma-inducible biosensor. The adaptation of a 384-well plate agar assay using X-Gal for the ΔdsbB1 mutant gave a dark blue color at high concentrations of compound 12 when β-Gal^dbs was induced at 5 µM Cuma (Fig. 4A, left). A higher concentration of β-Gal^dbs inducer, 25 µM Cuma, was needed for the ΔdsbB2 mutant to show observable activity, with even high concentrations of compound 12 only causing light blue pigmentation (Fig. 4A, right). We were also able to adapt a 96-well plate liquid assay from the E. coli assay by growing P. aeruginosa in M63 defined medium supplemented with 50 µg/mL of all essential amino acids except cysteine. As with the E. coli assay, the ONPG hydrolysis was measured without lysing P. aeruginosa cells. The quantification of β-Gal activity (Fig. 4B) reflected what was observed in the agar plates; the ΔdsbB1 mutant displays higher β-Gal activity at lower concentrations of compound 12 (EC50 = 0.19 µM) than the ΔdsbB2 mutant (EC50 = 0.74 µM). This is consistent with our previous findings using indirect assays for inhibition of DsbB proteins by measuring elastase activity in the supernatant of P. aeruginosa grown with the drug (25). Although compound 12 behaves as a strong inhibitor of the E. coli strains expressing P. aeruginosa DsbB1 (0.56 µM) or DsbB2 (0.28 µM), when using P. aeruginosa cells, the drug behaves as a stronger inhibitor of DsbB2 (ΔdsbB1 strain, IC50 = 0.01 µM) than DsbB1 (ΔdsbB2 strain, IC50 = 8.6 µM), and the drug does not reach full inhibition of elastase when the DsbB1 protein is present (Fig. S3) (25). This divergence between hosts was one of the motivations for the development of a target-based whole cell-based method using the native host, in this case, P. aeruginosa itself.

Fig 4.

β-Gal^dbs expression allows detection of DsbB inhibition in P. aeruginosa. (A) Inhibition of DsbB2 (ΔdsbB1, left) and DsbB1 (ΔdsbB2, right) with compound 12 can be observed by the generation of blue color when expressing β-Gal^dbs in the mutants. 384-well plates were prepared with 50 µL of M63 0.2% glucose supplemented with X-Gal and either 5 (ΔdsbB1) or 25 µM (ΔdsbB2) Cuma to induce β-Gal^dbs. A twofold serial dilution of 6-mM compound 12 was made in DMSO, and 1 µL of the diluted drug was pipetted into the agar surface. Bacteria (10 µL) were then added on top and incubated for 2 d at 30°C. Images represent the result of at least two independent experiments. (B) Quantification of β-Gal in ΔdsbB1 and ΔdsbB2 mutants grown in the presence of compound 12. β-Gal^dbs was measured by ONPG hydrolysis in live P. aeruginosa dsbB cells that were previously grown for 18 h at 37°C in M63 0.2% glucose supplemented with 50 µg/mL of essential amino acids (except Cys), 5 µM (ΔdsbB1) or 25 µM (ΔdsbB2) Cuma, and serially diluted compound 12. Data represent the average ± SD of three independent experiments, and non-linear regression was used to model the response.

DISCUSSION

The MalF-LacZ102 hybrid protein (β-Gal^dbs) has been an instrumental tool in studying disulfide bond formation in E. coli for over three decades. The use of LacZ allows for simple genetic selections and screens (3). Basal expression of the hybrid protein in M63 glucose minimal medium without maltose allowed the development of a heterologous platform to search for inhibitors of DsbB and VKOR proteins via high-throughput screening using E. coli (24, 25). However, the misfolding of β-Gal^dbs in the periplasm results in toxicity and cell death due to aberrant disulfide bonds formed by DsbA when the fusion was highly induced from P_mal-β-Gal^dbs in the presence of maltose and glycerol in the medium (5). Indeed, our previous attempts to use the β-Gal^dbs in other gram-negative organisms failed likely due to this toxicity problem. In this work, we showed that it is possible to express the β-Gal^dbs in a heterologous host, such as P. aeruginosa, when using tightly regulated promoters. The expression of β-Gal^dbs under the Cuma-inducible promoter was tolerated in P. aeruginosa cells, and no growth defect was observed at the maximum inducer concentration tested. Deletion of dsbA or both dsbB proteins in P. aeruginosa cells expressing the β-Gal^dbs allows the folding of periplasmic β-Gal and hence produces blue colonies in the presence of X-Gal, similar to what is observed for the single deletions of dsbA or dsbB in E. coli. We were also able to adapt our E. coli high- (384-well plate) and medium-throughput (96-well plate) methods expressing the β-Gal^dbs sensor in P. aeruginosa. The Cuma- and aTc-inducible promoters gave a good dynamic range to look for inhibitors of Dsb proteins in P. aeruginosa. The amount of inducer can be optimized to a desired level to increase the assay’s sensitivity. For instance, with more β-Gal^dbs expressed to a level where there is already a pale blue color (when DsbAB cannot misfold all β-Gal^dbs produced), one would detect weaker inhibitors, while less expression of the fusion would find stronger inhibitors (24).

Considering that disulfide bonds are required for the stability of virulence factors, toxins, antibiotic resistance, and cell envelope biogenesis proteins, lack of disulfide bond formation in many gram-negative pathogens results in virulence attenuation as well as antibiotic and phage susceptibility (6, 7, 21, 23, 46). However, the pathway is not essential for aerobic laboratory growth (17), thus facilitating the development of methods to search for molecules that identify molecules on-target in pathogenic bacteria, also known as target-based and whole cell-based approaches (47). The biosensor plasmids generated in this work could not only help to identify such Dsb inhibitors but could also help to study the variations of disulfide bond formation found in other bacteria (7, 14). Some gram-negative organisms encode two or more sets of DsbAB proteins that seem to play different roles but can substitute for each other in oxidizing particular substrates. Examples of organisms with multiple Dsb proteins include P. aeruginosa (13), Salmonella enterica serovar Typhimurium (48), Neisseria meningitidis (49, 50), and Campylobacter jejuni (51). In addition, some proteobacterial DsbA proteins such as Legionella pneumophila DsbA can perform both reduction and oxidation of disulfide bonds (52). The redundancy of Dsb proteins could perhaps provide a specialized activity for folding certain substrates that require priority under some environmental conditions. Using the single dsb mutants expressing the β-Gal^dbs, one could look for an array of growth conditions that inactivate the Dsb protein and render Lac+ cells. Similarly, one could also look for intragenic L. pneumophila dsbA mutations to either loose or gain a more oxidizing nature with the use of the β-Gal^dbs by looking for pale blue or white colonies, respectively. In addition, the β-Gal^dbs can be used to study bacterial interactions between organisms that share the same niche, such as the gut, to determine whether other bacteria could affect the oxidative protein folding of their neighbors through the secretion of quinones (53) or other redox molecules.

The use of translational LacZ fusions has advanced the study of membrane topology, protein targeting and secretion, as well as oxidative protein folding in E. coli since the 1970s (3). The use of the BHR plasmid-borne β-Gal^dbs disulfide bond biosensor we describe here should enable this analysis in a greater diversity of bacteria.

MATERIALS AND METHODS

Strains and growth conditions

The strains and plasmids used in this study are listed in Table 1. The primers used in this study are listed in Table 2. To construct PL23, 25, 26, and PL31, plasmid backbones without eYFP were amplified by PCR with PR41 and PR42 using Q5 Polymerase (New England Biolabs) and pAJM657, pAJM773, pAJM847, and pAJM011 as templates. The β-Gal^dbs insert was amplified with PR43 and PR44 using HK325 genomic DNA as a template. Insert and vectors were independently assembled using BsaI-HFv2 Golden Gate Assembly (NEBridge, New England Biolabs). Ligated products were transformed into HK320 electrocompetent cells and were plated on NZ Kanamycin, inducer (Cuma, Van, DAPG, or aTc), and X-Gal. One blue colony was used to purify the plasmid and sequence the insert using primers PR51 and PR52. To construct PL60-PL63, high-fidelity assembly (NEBuilder, New England Biolabs) was used using three PCR products. First, PR96 and PR97 were used to amplify the backbone with β-Gal^dbs without araC and arabinose promoter using pLEM2 vector. Second, the four regulators were amplified with PR98 and PR99, while the four promoters were amplified with PR100 and PR101 using pAJM657, pAJM773, pAJM847, and pAJM011 as templates. The three inserts were assembled independently using NEBuilder and were transformed into DH10β competent cells (New England Biolabs). Plasmids were sequenced with PR102 and PR46 to confirm the regulators and promoters. PL60 was then transformed into E. coli S17λpir electrocompetent cells and was used to conjugate the vector into P. aeruginosa strains. Nalidixic acid was used to counter-select the E. coli donor and gentamicin to select for the plasmids.

TABLE 1.

List of strains

ID	Genotype	Reference
E. coli strains
HK320	HK295 ΔdsbB	(54)
HK325	HK295 ΔdsbB λPmal-malE ’Φ(malF-lacZ)102	(54)
S17-1 λpir	F-, thi, pro, hsdR, [RP4-2 Tc::Mu Km::Tn7 (Tp Smr)],(Lambda-pir)	Kolter lab
P. aeruginosa strains
CL418	P. aeruginosa UCBPP-PA14	Lory lab
CL384	P. aeruginosaΔdsbA1	(25)
CL339	P. aeruginosa ΔdsbB1	(25)
CL355	P. aeruginosa ΔdsbB2	(25)
CL356	P. aeruginosa ΔdsbB1ΔdsbB2	(25)
LL81	P. aeruginosa UCBPP-PA14, pJN105-Cuma-β-Galdbs	This study
LL264	P. aeruginosa ΔdsbB1, pJN105-Cuma-β-Galdbs	This study
LL265	P. aeruginosa ΔdsbB2, pJN105-Cuma-β-Galdbs	This study
LL266	P. aeruginosa ΔdsbB1ΔdsbB2, pJN105-Cuma-β-Galdbs	This study
LL498	P. aeruginosa ΔdsbA1, pJN105Cuma-β-Galdbs	This study
LL499	P. aeruginosa UCBPP-PA14, pJN105Van-β-Galdbs	This study
LL500	P. aeruginosa ΔdsbA1, pJN105Van-β-Galdbs	This study
LL501	P. aeruginosa ΔdsbB1, pJN105Van-β-Galdbs	This study
LL502	P. aeruginosa ΔdsbB2, pJN105Van-β-Galdbs	This study
LL503	P. aeruginosa ΔdsbB1 ΔdsbB2, pJN105Van-β-Galdbs	This study
LL504	P. aeruginosa UCBPP-PA14, pJN105Tet-β-Galdbs	This study
LL505	P. aeruginosa ΔdsbA1, pJN105Tet-β-Galdbs	This study
LL506	P. aeruginosa ΔdsbB1, pJN105Tet-β-Galdbs	This study
LL507	P. aeruginosa ΔdsbB2, pJN105Tet-β-Galdbs	This study
LL508	P. aeruginosa ΔdsbB1 ΔdsbB2, pJN105Tet-β-Galdbs	This study
LL509	P. aeruginosa UCBPP-PA14, pJN105DAPG-β-Galdbs	This study
LL510	P. aeruginosa ΔdsbA1, pJN105DAPG-β-Galdbs	This study
LL511	P. aeruginosa ΔdsbB1, pJN105DAPG-β-Galdbs	This study
LL512	P. aeruginosa ΔdsbB2, pJN105DAPG-β-Galdbs	This study
LL513	P. aeruginosa ΔdsbB1 ΔdsbB2, pJN105DAPG-β-Galdbs	This study
Plasmids
pNG102	Pmal-malE ’Φ(malF-lacZ)Hyb102, Ampr	(1)
pAJM657	eYFP, CymR, Cuma promoter, Kmr, p15A ori	(29)
pAJM773	eYFP, VanR, Kmr, p15A ori	(29)
pAJM847	eYFP, PhlF, DAPG promoter, Kmr, p15A ori	(29)
pAJM011	eYFP, TetR, Tet promoter, Kmr, p15A ori	(29)
pJN105	AraC, ara promoter, Gmr, mob, pBBR ori	(44)
pLEM2	pJN105-Pmal- malE ’Φ(malF-lacZ)102, Gmr	This study
PL23	pAJM657-malF-lacZ102, Kmr	This study
PL25	pAJM773-malF-lacZ102, Kmr	This study
PL26	pAJM847-malF-lacZ102, Kmr	This study
PL31	pAJM011-malF-lacZ102, Kmr	This study
PL60	pJN105-Cuma-malF-lacZ102, Gmr	This study
PL61	pJN105-Van-malF-lacZ102, Gmr	This study
PL62	pJN105-Tet-malF-lacZ102, Gmr	This study
PL63	pJN105-DAPG-malF-lacZ102, Gmr	This study

TABLE 2.

Primer list

ID	Sequence 5’ to 3’
PR41.pAJMR_fwd	ggctacggtctcccaaattccagaaaagaggc
PR42.pAJMR_rev	ggctacggtctccttcccctctttctctagtattaaac
PR43.MalF_fwd	ggctacggtctccggaaatactagatggatgtcattaaaaagaaac
PR44.LacZ_rev2	ggctacggtctcctttggtaccgagtcatttttgacaccagac
PR46. pAJMRFZ_rev	gggatcccccatcatcactttgttg
PR51. pAJMseq_fwd	tgagacacaaccaattattgaaggcc
PR52. pAJMseq_rev	ccgccattggaccaaaacgaaaa
PR96.pLEM2_fwd	atggatgtcattaaaaagaaacattggtg
PR97.pLEM2_rev	attgcgttgcgctcactg
PR98.Reg_fwd	ggcagtgagcgcaacgcaatgccggtgatgccggccac
PR99.Reg_rev	ttgtgtctcatttcgttttggtccaatggcggcg
PR100.Prom_fwd	caaaacgaaatgagacacaaccaattattg
PR101.Prom_rev	ttctttttaatgacatccatctagtatttcccctctttc
PR102.Gmprom_fwd	cgcgttggccgattcatt

All E. coli strains were grown in NZ or M63 0.2% glucose and supplemented with 50 µg/mL of 19 essential amino acids (Ser, Val, Ile, Leu, Trp, Tyr, Met, Asp, Glu, Ala, Arg, Lys, Asn, Gln, Phe, Gly, Thr, His, and Pro) at 30–37°C when indicated. All P. aeruginosa strains were grown in lysogeny broth (LB) or M63 0.2% glucose supplemented with 50 µg/mL of 19 essential amino acids either broth or agar media at 30–37°C when indicated. The antibiotic concentrations used were nalidixic acid 10 µg/mL, kanamycin 40 µg/mL, gentamicin 3 µg/mL for E. coli and 10 µg/mL for P. aeruginosa.

Kanamycin was purchased from Sigma (USA). Nalidixic acid and gentamicin were purchased from GoldBio (USA). Inducers were purchased from Sigma (cuminic acid, vanillic acid, and anhydrotetracycline hydrochloride) and ChemCruz (2,4-diacetylphlorogucinol). Molecules were dissolved in dimethyl sulfoxide (DMSO) to make 100 mM (Cuma, Van, DAPG) and 1 mM (aTc) stocks, respectively. Compound 12 was purchased from Enamine (EN300-173996 purity 95%, Ukraine) and was dissolved in DMSO to 10 mg/mL (34.72 mM).

β-Galactosidase assays

β-Galactosidase assays were carried out by determining the velocity of hydrolysis of o-nitrophenyl-β-galactoside (ONPG, Sigma) in microtiter plates using a protocol previously described with slight modifications (24, 27, 55). Briefly, cultures were grown in M63 0.2% glucose medium with proper antibiotics at 37°C overnight. Cells were then inoculated to an OD600 of 0.01 into fresh M63 medium containing 0.2% glucose, 50 µg/mL of 19 essential amino acids (except cysteine), and proper antibiotics. Diluted bacteria (200 µL) were transferred to a 96-well plate (Corning), and a twofold serial dilution of the inducers was performed from columns 12 to 2. The plate was then sealed with a breathable film. The growth plate was incubated for 10 h at 30°C and fast-orbital shaking inside a Synergy H1 plate reader (BioTek). After growth, absorbance at 600 nm was read in a Synergy H1 plate reader (BioTek). Then, 100 µL of bacteria from the growth plate was transferred to the assay plate. Note that no cell lysis step is required because β-Gal^dbs is in the periplasm. The reaction was started by adding 100 µL of the ONPG buffer to the cells (a mixture of 8 mL of Z-buffer with 4 mL of 4 mg/mL ONPG dissolved in Z-buffer). The addition of β-mercaptoethanol was also omitted to avoid interference with the disulfide bonds formed in β-Gal^dbs. The absorbance at λ420 nm was measured every minute for 1–2 h to follow the kinetics of ONPG hydrolysis at 28°C in a Synergy H1 plate reader (BioTek). The velocity of the reaction was calculated by performing linear regression using GraphPad Prism Software. The slopes were then used together with OD600 and the following constants, 1.81 (CF1), 2.45 (CF2), and 3.05 (CF3), to calculate the Miller Units as reported before (55). P. aeruginosa β-Galactosidase assays were conducted in microtiter plates like the E. coli protocol with few modifications in the growth conditions. Cultures were grown in M63 0.2% glucose medium supplemented with 100 µL of NZ broth and proper antibiotics at 37°C overnight. Cells were then inoculated to an OD600 of 0.01 into fresh M63 medium containing 0.2% glucose, 50 µg/mL of 19 essential amino acids (except cysteine), proper antibiotics, and either a fixed concentration or a serial dilution of Cuma when indicated. Diluted bacteria (200 µL) were transferred to a 96-well plate (Corning). Compound 12 (100 µM) was added to column 12 and was twofold serially diluted to column 2. The plate was sealed with a breathable film and incubated for 18 h at 37°C and 700 rpm in an Incu-mixer MP (Benchmark). After growth, the same steps were performed like the E. coli assay indicated above. Non-linear regression using GraphPad Prism Software (Boston, USA) was used to model the data by using the equation [Agonist] vs response—variable slope (four parameters).

Agar assays

Drug testing was performed as previously described with slight modifications (24, 25). A liquid dispenser (BioTek) fitted with a small-bore tubing cartridge was used to dispense 50-µL aliquots of hot agar medium (M63 medium containing 0.2% glucose and 0.9% agar, supplemented with gentamicin [10 µg/mL], Cuma [5 or 25 µM], and X-Gal [120 µg/mL]) to 384-well tissue culture-treated plates (BD Falcon). In order to prevent agar solidification in the Wellmate tubing (at too-low temperatures) or inactivation of the antibiotics and X-Gal (at too-high temperatures), the medium was maintained in a 60°C oven. In addition, the Wellmate tubing was pre-warmed by washing with sterile hot water immediately prior to loading the agar medium. After the agar solidified, the plates were stored in a humidified sealed container at 4°C for no more than 2 d. Compound 12 was added by pipetting 1 µL of serial dilutions into the agar’s surface (final concentration of DMSO, 1.7%), and 10 µL of diluted P. aeruginosa cells (OD600 of 0.01) was dispensed with a liquid dispenser (BioTek). 384-well plates were sealed with a breathable film and then incubated in humidity boxes at 30°C for 2 days and then stored at 4°C for 2 days to determine the minimal concentration to produce a blue color. Photographs were taken using a white light-emitting diode (LED) transilluminator.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00433-23.

Supplemental material. jb.00433-23-s0001.pdf.

Figure S1 to S3.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.