A GFP-lacZ Bicistronic Reporter System for Promoter Analysis in Environmental Gram-Negative Bacteria

Here, we describe a bicistronic reporter system for the analysis of promoter activity in a variety of Gram-negative bacteria at both the population and single-cell levels. This synthetic genetic tool utilizes an artificial operon comprising the gfp and lacZ genes that are assembled in a suicide vector, which is integrated at specific sites within the chromosome of the target bacterium, thereby creating a monocopy reporter system. This tool was instrumental for the complete in vivo characterization of two promoters, Pb and Pc, that drive the expression of the benzoate and catechol degradation pathways, respectively, of the soil bacterium Pseudomonas putida KT2440. The parameterization of these promoters in a population (using β-galactosidase assays) and in single cells (using flow cytometry) was necessary to examine the basic numerical features of these systems, such as the basal and maximal levels and the induction kinetics in response to an inducer (benzoate). Remarkably, GFP afforded a view of the process at a much higher resolution compared with standard lacZ tests; changes in fluorescence faithfully reflected variations in the transcriptional regimes of individual bacteria. The broad host range of the vector/reporter platform is an asset for the characterization of promoters in different bacteria, thereby expanding the diversity of genomic chasses amenable to Synthetic Biology methods.


Introduction
The process of gene regulation in a living organism is vitally important for its adaptation to changing conditions in the environment. In fact, the analysis of the complete genomes currently available reveals that a large amount of the genome encodes sequences related to gene regulation and transcription [1,2,3,4,5]. In the case of prokaryotes, the comparison between organisms with different lifestyles shows that the generalists (such as free living environmental bacteria) usually have a higher proportion of their genome content dedicated to gene regulation than the specialists (e.g., endosymbionts, [6]). Moreover, the coordination of gene expression involves several steps that are controlled by transcriptional factors (TFs, [7,8,9]). In the last few years, there has been tremendous progress in the study of regulatory networks, and evidence has shown that, among other things, gene expression is also largely susceptible to stochastic variations from cell to cell [10,11,12], primarily because there are typically few units of the reactant molecules (TF, promoters, etc.) found in the cell cytoplasm [12,13,14]. The biochemical processes underlying gene regulation are driven by the collision of the reactants; therefore, the low amount of the reactants makes the system prone to a higher level of noise in the final output. In addition, reactions endowed with low kinetic constants contribute to the increased noise of the cell [11].
While living organisms can indeed use noise to control crucial differentiation programs, such as competence in Bacillus subtilis [15,16], it is usually deleterious for the proper functioning of intracellular circuits. In fact, the improper function of some synthetic circuits in bacteria is a consequence of high levels of noise during the different processes of gene regulation [17,18,19,20]. For the integration of stochasticity in decision-making switches, the resulting network architectures are usually associated with the generation of multi-stability, where it is possible for the cell to obtain multiple stable states [16,21,22]. In a multi-stable system, the presence of noise is crucial as small fluctuations during gene regulation determine the fate of the network and thus its final steady state [16,21,23].
An analysis of the stochastic process in cellular systems is fundamental not only for the understanding of the underlying process in a given regulatory network, but also for the proper characterization of molecular components that regulate synthetic circuits [24,25,26]. Thus, single-cell methodologies based on fluorescent proteins, such as GFP, are crucial for the analysis of noise. While classical approaches based on the enzymatic quantification of a reporter gene (e.g., lacZ) only provide information concerning the gene expression process within a population, fluorescent reporter assays offer high resolution information at the single cell level [11]. Ideally, a single reporter system should be instrumental to examine both of these aspects, but indicator products (e.g., fluorescent proteins) that are best suited for use in single cells perform poorly in cultures because of quenching and other optical interferences. Similarly, the best population-level reporters (e.g., lacZ and luxAB) give diffuse signals that prevent the examination of particular bacteria. One possible solution is the combination of two reporter products in a single transcriptional unit [27,28] or hybrid functional polypeptide [29]. Unfortunately, these constructs are only amenable for model bacteria, such as E. coli, and the toolbox for less standardized microorganisms, such as environmental Gram-negative bacteria, is much more scarce [30,31]. For instance, the soil bacterium Pseudomonas putida KT2440 is endowed with the remarkable capability to degrade aromatic compounds [31,32]. As this organism is naturally adapted to deal with toxic pollutants, it is a more versatile bacterium for in situ applications, such as biosensing xenobiotics [33,34] and the mineralization of chemical contaminants. Yet, the number of tools available for studying gene expression in this bacterium -as well as in other Gram-negative microorganisms, is scarce as compared to E. coli.
In this paper, a dual promoter probe system based on the expression of GFP and lacZ, the two most widely used reporters in bacteria [35], is used for the examination of environmental Gramnegative bacteria. The system is designed to allow the insertion of an artificial operon (GFP-lacZ) into a specific position in the bacterial chromosome, creating a stable monocopy reporter system. The resulting strains can be then assayed for both GFP and lacZ expression, allowing the characterization of the target promoters in monocopy gene dose and stoichiometry. This system has been instrumental for studying the regulation of the benzoate and catechol degradation pathways in P. putida, which show monostable behavior during the entire catabolic process. We anticipate that the applications of this system will be expanded from the fundamental research of regulatory networks to more applied endeavors in Synthetic Biology.

Construction of a Dual GFP-lacZ Reporter System
Initially, we created a bicistronic system for promoter probing based on the expression of GFP and lacZ genes using as a reference a system for the single copy integration of a lacZ reporter that was previously described by Kessler and co-workers [36]. Briefly, a stable version of GFP [37] and a modified version of the lacZ gene, which had a premature amber stop codon TAG at the 3' region, was inserted into the suicide vector pRV1 harboring a promoterless synthetic operon ( Fig. 1A; [36]). A streptomycin/spectinomycin (Sm/Sp) resistance marker, which was located in a divergent orientation, also contained a premature stop codon ( Fig. 1A; [36]). The presence of the amber codons made the two resulting proteins non-functional unless the host strain harbored a supF tRNA. In addition, because of the narrow host range of pBR322 ori, the pRV1 vector is unable to replicate in organisms unrelated to E. coli [36]. Moreover, the RP4 oriT transfer origin allowed for plasmid mobilization through conjugation to target bacterium, such as P. putida. The target promoter is cloned into the pRV1 vector using EcoRI and BamHI restriction enzyme sites located upstream of the GFP gene (Fig. 1A). To integrate the GFP-lacZ reporter cassette, the target strain should contain a homologous fragment in the chromosome to mediate double homologous recombination (see below, [36]). The mini-Tn10-based transposon pLOF-hom.fg. was used to modify the transferred strain. In this transposon, a homologous fragment containing the lacZ gene, which lacked the ATG start codon, and a Sm/Sp marker disrupted by a kanamycin (Km) gene were placed in the transposable element (Fig. 1B). As explained in the next section, recombination between the versions of the Sm marker and the lacZ gene allowed the integration of the reporter system through the reconstitution of fully functional genes in the chromosome.

Generation of a Non-fluorescent P. Putida Strain
The soil bacterium P. putida produces a fluorescent siderophore pyoverdine for the capitation of iron ions from the environment [38,39]. To assay for GFP expression in P. putida, we generated a bacterial strain that was unable to produce pyoverdine, as this siderophore may mask the fluorescence from GFP. The steps for mutagenization of P. putida are summarized in Figure 2. First, a random mutant library was created using a mini-Tn5-based vector containing a removable Km marker (Fig. 2, step I). The pyoverdine production was assayed in minimal media plates under ultraviolet (UV) light, and the two colonies that were unable to fluoresce were selected and named P. putida UV1 and UV2 (Fig. 2, step II). Subsequently, the Km resistance markers of the two strains were removed by expressing the resolvase coding gene parA, generating strains P. putida MEG1 and MEG2 (derived from P. putida UV1 and UV2, respectively). As shown in Figure 2, upon removal of the Km marker, the strains did not recover the fluorescent phenotype (step III). Once we generated bacterial strains suitable to analyze GFP, we introduced the homologous fragment into the chromosome of P. putida by targeting P. putida MEG1 with the pLOF-hom.fg. transposon as indicated in the Materials and Methods section. The resulting strain, harboring the homologous fragment stably integrated in the chromosome, was named P. putida MEG3 and used in further analyses.

Assembly of Monocopy Promoter Fusions to the GFP-lacZ System
As previously discussed, P. putida is endowed with the remarkable capability to degrade a number of aromatic compounds [32]. In this organism, benzoate is degraded through the b-ketoadipate pathway, which is highly distributed among the Pseudomonas genera and related organisms [40]. In this pathway, benzoate is oxidized to catechol, which is then subjected to an intradiol cleavage to generate cis,cis-muconate (cis,cis-muc, Fig. 3). Subsequently, cis,cis-muconate is degraded to form b-ketoadipate, which is further degraded to tricarboxylic acid cycle (TCA) intermediates. The first steps of benzoate metabolization are performed for the enzymes encoded in the ben operon [40]. This operon is expressed from the Pb promoter, which is activated by the AraC-type regulator BenR bound to benzoate ( Fig. 3; [41]). The metabolization of cis,cis-muconate depends upon the action of the cat genes. These genes are activated by the LysR-type regulator CatR, a TF that responds to the presence of cis,cis-muconate [42,43]. Finally, the final steps of metabolization are performed by the pca genes, which are controlled by the b-ketoadipate-induced IclR-type regulator PcaR [40,44].
To validate the performance of the bicistronic reporter system described here, we analyzed the dynamic properties for the initial steps that control the expression of the b-ketoadipate pathway in P. putida. First, the promoter regions of the ben and cat operons were cloned into the pRV1 reporter vector. The resulting constructs where introduced into P. putida MEG3 by tri-parental mating as previously described ( Fig. 4; [45]). After plasmid transference to P. putida, the cells were selected in minimal media using Sm/Sp as the resistance marker. Following two homologous recombination events, a fully functional Sm/Sp marker and lacZ gene were generated in the chromosome of the recipient bacterium, and the Km resistance phenotype of the strain was lost (Fig. 4). The resulting strains containing the correct insertion of the Pband Pcbased systems were named P. putida MEG3-Pb and MEG-Pc, respectively, and used in further assays.
Population and Single-cell Analysis of Promoter Activity in P. Putida After constructing P. putida reporter strains containing monocopy bicistronic cassettes, we investigated the expression profiles of Pb and Pc in response to benzoate. As shown in Figure 5, P. putida MEG3-Pb and MEG-Pc simultaneously expressed GFP and LacZ proteins in the presence of benzoate on the agar plates. To quantify the promoter activity in response to benzoate present in the liquid media, an overnight culture of cells was diluted in minimal media containing succinate as the sole carbon source and incubated for a few hours. At the mid-exponential phase, 1 mM benzoate was added to the growth media, and the cells where incubated for an additional 4 hours. Subsequently, samples were taken and analyzed using flow cytometry to quantify GFP expression, and a b-galactosidase assay was used to quantify the expression of lacZ. As shown in Figure 6A, Pb and Pc showed a higher level of induction in response to benzoate as assayed using GFP. In general, Pb presented a higher basal level and a lower maximal activity compared with Pc (Fig. 6A). Interestingly, the analysis of the promoter activities using the b-galactosidase assay showed the same overall results, indicating that the synthetic construction worked faithfully as a bicistronic unit (Fig. 6B). The primary difference between the two reporters was observed in the fold-change detected in the activity of the two promoters depending on the reporter examined. In the case of GFP, a 22fold induction of Pb was observed, whereas Pc presented a fold change of 100. However, when lacZ was used as the reporter, we observed a fold induction of 10 and 67 for Pb and Pc, respectively. These results indicated that, under the conditions specified, GFP provides a higher resolution of the changes in transcription rates in response to the inducer. This circumstance seems to improve the signal-to-noise ratio and thus provides a systematically higher induction as compared with that of lacZ in the same cells, where both reporters are expressed simultaneously.
We next analyzed the behavior of the two promoters in a single cell using GFP. The promoter activities were monitored along the induction curve using flow cytometry as described in the Materials and Methods section. As shown in Figure 6C, at the population level, the Pc promoter showed a more noticeable induction curve than Pb, as the former started at a lower basal value and reached a higher maximal activity. When we analyzed the distribution of fluorescence from cell to cell within the population, we found that both promoters presented similar dynamic behaviors for GFP expression. Only single populations were observed at each time point, and a continuous transition in the fluorescence levels was observed (Fig. 6D-E). These results suggested that a mono-stable regulatory device controls the b-ketoadipate pathway for benzoate metabolism in P. putida which prevents the appearance of inactive subpopulations during biodegradation.

Conclusion
Here, we describe the implementation and validation of a dual bicistronic reporter system based on the GFP and lacZ genes. This novel tool permits the characterization of the regulatory networks in many environmental bacteria for which only few genetic tools are available. By integrating the system in monocopy in the chromosome of the target bacterium, we avoided the deleterious effects associated with the stochastic variation of copy number observed in plasmid-based systems [17,18,19]. In fact, the noise generated in plasmid-based systems might interfere with the accurate analysis of the target network [46]. While the Tn7 transposon-based system also provide an alternative to the insertion of synthetic constructs in single copy, this method only utilizes native sites existing in the bacterium [47]. However, the novel system introduced here exploits different genomic locations, as the integration site can be randomly generated using a mini-Tn10 delivery transposon [48]. In addition, the combinatory use of this system with existing Tn7-based tools would permit the construction of stable strains with isogenic modifications, as different inserts would be placed at specified positions. We used this system to investigate the expression of the pathway for benzoate metabolism in the soil bacterium P. putida. Usually, the genetic characterization of the regulatory networks in this and other atypical organisms is performed in the model organism E. coli [49,50], but this approach might miss important host-specific features. For example, using the new tool in the native host, we were able to monitor cat pathway expression in response to benzoate. Under the experimental conditions used in this study, benzoate was converted to cis,cis-muconate to trigger Pc activity (Fig. 3). Moreover, the novel system presented here can be easily applied to other organisms and would provide valuable informa-

Bacterial Strains
E. coli CC118 [45] or its variant CC118supF [36] were used as the hosts for the plasmid constructs. E. coli HB101 (pRK600) was used as a helper strain for tri-parental mating as previously described [45]. The induction experiments were performed in M9 minimal medium [51] with 2 mM MgSO 4 and 25 mM of succinate as the sole carbon source. This minimal medium was additionally supplemented with Sm (50 mg ml -1 ) and Km (50 mg ml -1 ) to ensure plasmid retention. The benzoate was purchased from Sigma-Aldrich.

Plasmid Construction
For cloning purposes, DNA fragments were amplified using the polymerase chain reaction (PCR) containing 50-100 ng of the template, 50 pmol of each primer and 2.5 U of Pfu DNA polymerase (Stratagene) in a 100 ml reaction volume. The mixtures were subjected to 30 cycles of 1 min at 95uC, 1 min at 55uC and 2 min at 72uC. The primers used were purchased from Sigma. The plasmid DNA purification, restriction enzyme digestion and general cloning procedures were conducted using standard protocols [51]. For the analysis of the stochastic effects of the TOL promoters under inducing conditions, we developed a system for the insertion of transcriptional GFP fusions in the chromosome of P. putida. This system is a modified version of a lacZ-based system described by Klesser et al., which comprises a Figure 2. Isolation of a non-fluorescent P. putida strain. (i) P. putida KT2442, a rifampicin resistance variant of P. putida KT2440, was mutagenized with mini-Tn5 transposon pUT-ResKm. (ii) Single colonies were selected in minimal media and screened for the lack of fluorescence. Two colonies were selected and named P. putida UV1 and UV2. (iii) The Km resistance marker was removed from the strains by expressing ParA resolvase, which recognizes the two res sites flanking the Km marker. After marker elimination, the strains retained the non-fluorescence phenotype. (iv) The homologous fragment placed in the mini-Tn10 transposon was mobilized to the marker-less P. putida MEG1 strain (a deviant of P. putida UV1), generating the P. putida MEG3 strain. This strain was used as a host for the bicistronic reporter system with the target promoters. doi:10.1371/journal.pone.0034675.g002 suicide vector pBK16 (Fig. 1A) and a homology fragment placed in a mini-Tn10 transposon (Fig. 1B, [36]). To generate a variant of pBK16 having a fluorescent reporter, the gfp tir gene from the pGREENtir [37] plasmid was PCR amplified using the primers 5GFP (5'-GAA TTC ATC GGA TCC TGA TTA ACT TTA TAA GGA GG-3') and 3GFP (5'-ATA GAT CTT TAT TAT TTG TAT AGT TCA TCC ATG CC-3'). The resulting PCR fragment was digested with the restriction enzymes EcoRI and BglII and cloned into a pBK16 vector that was previously digested with EcoRI and BamHI. The resulting vector pRV1 performed as a bicistronic gfp/lacZ reporter system (Fig. 1A). The promoters of the ben (Pb) and cat (Pc) pathways where PCR amplified using the primer pairs PBf (5'-TGG ATG AAT TCG ACA GTA CCC TCC-3')/PBr (5'-GCG CGG ATC CGG CCA GGG TCT CCC TTG-3') and PCf (5'-GAG AGA ATT CAG GCC CAG TTC CAG CTC G-3')/PCr (5'-GCG CGG ATC CTG TTG CCA GGT CCC GTC AG-3'), respectively, and cloned as EcoRI/ BamHI fragments into pre-digested pRV1 vectors. The resulting reporter vectors were introduced into P. putida MEG3 as described below.

Construction of the Non-fluorescent P. Putida Strains
To perform experiments using GFP in P. putida, we generated variants of this organism that were unable to auto-fluorescence. Briefly, overnight cultures of P. putida KT2442 were mutagenized with the pUT-ResKm transposon (a modified version of pUT-Km [52] with res sites flanking the Km marker) by tri-parental mating [45]. The P. putida transposition library was plated on M9 agar plates supplemented with 0.2% citrate and Km. After overnight incubation, the colonies were screened under ultraviolet (UV) light at 254 nm for the identification of strains with no autofluorescence (Fig. 2). Two strains that did not fluoresce under UV light were selected and named P. putida UV1 and UV2. Subsequently, the plasmid pJMSB8 [53] was transferred to P. putida UV1 and P. putida UV2 by tri-parental mating. This plasmid expressed the gene for the ParA enzyme, which catalyzes the recombination between adjacent Res sequences [53]. The pUT-ResKm transposon used to create strains UV1 and UV2 contained a Km gene flanked by two res sequences, thus the expression of ParA in this system expression eliminated the expression of the resistance marker. Upon insertion of the ParA coding plasmid, single colonies of strains UV1 and UV2 were analyzed for the loss of resistance to Km. Two strains demonstrating successful antibiotic removal were confirmed and named P. putida MEG1 and MEG2 (derived from UV1 and UV2, respectively) and used for further analysis (Fig. 2).

Construction of the P. Putida MEG3 Reporter Strains
For the analysis of the TOL promoter activities at the single cell level, we used a modified version of the pBK16 homologous recombination system (see above, [36]). To generate a P. putida variant that was able to accommodate the GFP reporter system in the chromosome, we first mutagenized P. putida MEG1 using the pLOF-hom.fg. mini-Tn10 transposon [36]. This transposon contains a lacZ gene lacking the ATG start codon cloned in the opposite direction of an Sm/Sp resistance marker truncated by the insertion of a Km gene (Fig. 1B). The resultant transposon library was plated on M9 minimal media supplemented with 0.2% citrate and Km. A single colony that was able to grow in Km and confirmed as the resultant of two transposition events was named MEG3 and used in further experiments. For the generation of P. putida strains containing transcriptional fusions to GFP in the chromosome, pRV1-variants with different promoters were transferred to P. putida MEG3 (Fig. 4). After tri-parental mating, the strains were plated on M9 minimal media supplemented with 0.2% citrate and Sm to select for homologous recombination events between the Sm/Sp gene from the pRV1 vector and the homologous fragments placed in the chromosome of MEG3 strain (Fig. 4). As a control, a P. putida strain lacking the homology fragment was used as the recipient for conjugation. No Smresistant colonies were observed, ruling out the possibility for plasmid integration on alternative sites. Finally, single colonies were assessed for the second recombination event (i.e., between the two variants of the lacZ gene) by analyzing their sensitivity to Km. Strains containing the correct insertion of the reporter system in the chromosome were named P. putida MEG3-P (where P stands for the identity of the cloned promoter) and used for the single-cell analysis. . Catabolic pathway for benzoate degradation in P. putida. Benzoate activates the BenR regulator, which controls the expression of the ben operon. The enzymes encoded by the ben operon metabolize the conversion of benzoate to cis,cis-muconate (cis,cis-muc), which is the signal for CatR activation. CatR controls the expression of the cat operon for the metabolization of cis,cis-muconate, which is further converted to bketoadipate. Finally, b-ketoadipate is converted to TCA intermediates through the action of the pca pathway (gray arrows). doi:10.1371/journal.pone.0034675.g003 Figure 4. Design of bicistronic GFP-lacZ reporter strains. (i) pRV1 containing the cloned promoters (labeled as P) was introduced into the P. putida MEG3 strain by conjugation. Colonies were selected in minimal media using the Sm antibiotic. (ii) The homology regions of the suicide vector pRV1 were recombined with the two segments placed in the chromosome, generating a functioning Sm/Sp resistance marker and lacZ gene. (iii) Finally, the correct insertion of the segments generated a strain with a stable reporter system that was sensitive to the Km antibiotic. doi:10.1371/journal.pone.0034675.g004

GFP Analysis at the Single-cell Level
For the quantification of GFP expression at the single-cell level, P. putida MEG3 strains containing different promoter fusions to GFP were inoculated into M9 media supplemented with 25 mM succinate. After overnight growth, the cultures were washed twice, diluted 1:20 in fresh M9 media containing 25 mM succinate and incubated for an additional 4 hours. After this pre-incubation, the cultures were distributed into new flasks containing different concentrations of benzoate and incubated with air shaking. Each hour after induction, 500 mL samples were spun down, the cells resuspended in 500 mL of PBS and stored on ice until analysis. The GFP distribution in the cell population was analyzed by flow cytometry using a GALLIOS cytometer (Perkin Elmer). For each sample, 15,000 events were analyzed. The data processing was performed using Cyflogic software (http://www.cyflogic.com/). Calculations of the mean fluorescence from the different replicas and standard deviations were calculated with the statistical package of Microsoft Excel (2010).

b-galactosidase Assay
To perform the LacZ activity assay, single colonies of reporter strains were grown overnight in M9 media supplemented with 25 mM succinate at 30uC. The overnight cultures were subsequently diluted 1:20 in fresh M9 media containing 25 mM Figure 6. Assay of promoter activities in response to benzoate. Overnight cultures were diluted in fresh minimal media supplemented with succinate as the sole carbon source. At the mid-exponential phase, the cells were exposed to 1 mM benzoate, and the GFP and the lacZ expression was assayed. A The GFP expression was measured in P. putida MEG3-Pb and MEG3-Pc after 4 hours of induction using flow cytometry. succinate and cultured for an additional 4 hours. Subsequently, benzoate was added to the media, and the cells were incubated for several hours. b-galactosidase activities were assayed in permeabilized whole cells according to Miller's method with minor modifications [54].