Inducible and tunable gene expression systems for Pseudomonas putida KT2440

Inducible and tunable expression systems are essential for the microbial production of biochemicals. Five different carbon source- and substrate-inducible promoter systems were developed and further evaluated in Pseudomonas putida KT2440 by analyzing the expression of green fluorescent protein (GFP) as a reporter protein. These systems can be induced by low-cost compounds such as glucose, 3-hydroxypropionic acid (3HP), levulinic acid (LA), and xylose. 3HP-inducible HpdR/PhpdH was also efficiently induced by LA. LvaR/PlvaA and XutR/PxutA systems were induced even at low concentrations of LA (0.1 mM) and xylose (0.5 mM), respectively. Glucose-inducible HexR/Pzwf1 showed weak GFP expression. These inducer agents can be used as potent starting materials for both cell growth and the production of a wide range of biochemicals. The efficiency of the reported systems was comparable to that of conventional chemical-inducible systems. Hence, the newly investigated promoter systems are highly useful for the expression of target genes in the widely used synthetic biology chassis P. putida KT2440 for industrial and medical applications.

www.nature.com/scientificreports/ systems and the intactness of the promoter systems were analyzed. These low-cost substrates are already used as potent starting materials for both cell growth and the production of a wide range of biochemicals. Although the microbial production of industrially relevant biochemicals using Glu-and LA-inducible expression systems has been previously reported 20,21 , the promoters have not been characterized in detail. To our knowledge, this is the first report describing the characterization of substrate-inducible and tunable promoter systems for P. putida KT2440 at the single-cell level to broaden their applicability.

Results and discussion
In the present study, the strength of each promoter was evaluated using a GFP-based reporter assay. The level of GFP fluorescence intensity was linearly correlated with the concentration of the respective inducers added. Therefore, the useful range of inducer concentration and homogeneity (cell populations with similar expression levels) of all expression systems were analyzed using flow cytometry. Furthermore, amicability between promoters/inducers is necessary when different expression systems are used to employ complex metabolic pathways containing multiple genes. Thus, the cross-talk among promoters/inducers was characterized. Here, five different carbon source-and substrate-inducible promoter systems, HexR/P zwf1 induced by glucose, LvaR/P lvaA by LA or 4HV, HpdR/P hpdH by 3HP or LA, MmsR/P mmsA by 3HP or LA, and XutR/P xutA by xylose, were developed (Fig. 1).
Glu-inducible HexR/P zwf1 system. The GFP expression of the strain HRZ01 harboring pHRZ-eGFP + was evaluated in M9Y medium supplemented with different concentrations of Glu (0.5 to 20 mM). As expected, the addition of Glu resulted in a higher fluorescence intensity compared to that in the control condition without Glu (Fig. 2a). The fluorescence intensity increased with increasing concentrations of Glu in the medium. However, no difference was observed when the concentration of Glu exceeded 5 mM. The system was not induced by the addition of Xyl, LA, 4-hydroxyvalerate (4HV), or 3-hydroxypropionic acid (3HP). Furthermore, the addition of either Xyl, LA, 4HV, or 3HP in combination with Glu (10 mM each) did not affect the induction of the HexR/ P zwf1 system (Supplementary Figure S1).
An intermediate of the Entner-Doudoroff pathway (EDP) involves in the activation of the HexR/P zwf1 system. P. putida KT2440 almost exclusively uses the EDP for glucose metabolism, rather than the Embden-Meyerhof-Parnas pathway (EMPP) 5 . The transcriptional repressor HexR in P. putida KT2440 negatively regulates the expression of the glucose metabolism genes, such as gapA gene and zwf-pgl-eda and edd-glk-gltR2 operons, by binding to consensus sequences found in upstream of target genes 22,23 . This HexR-mediated repression is blocked by the binding of 2-keto-3-deoxy-6-phosphogluconate (KDPG), an exclusive metabolic intermediate of the EDP, to HexR via its sugar isomerase domain, thus inducing the expression of target genes after releasing the repressor (HexR) from the operator sites [22][23][24] . Therefore, the added Glu was efficiently catabolized via the EDP, thus producing KDPG to bind the HexR protein, resulting in the expression of GFP ( Supplementary Fig. 1). A previous study showed that an EDP-activated Escherichia coli strain harboring this system efficiently produced 2,3-butanediol (BDO) from Glu with a 71% increased titer compared to the control strain 20 . Moreover, and that the system was comparable to the chemical (anhydrotetracycline) inducible expression system (P LtetO-1 ). No considerable difference was observed on the efficiency of both the systems to produce BDO (13.95 g/L vs 14.17 g/L) 20 . Thus, Glu, which is a low-cost carbon source for growth and a starting material in the production of diverse chemicals, could be used as an efficient inducer for the production of target chemicals.
LA-inducible LvaR/P lvaA system. The induction of the LvaR/P lvaA system was analyzed by estimating the fluorescence intensity of the strain LRL01 harboring pLRL-eGFP + with different concentrations of LA (0.1 to 20 mM) or 4HV (1-20 mM). The addition of LA significantly increased (P < 0.05) the fluorescence intensity compared to that in the control (Fig. 2b). The system was inducible even at low concentrations of LA (0.1 mM). Notably, without the inducer, the system was tightly controlled (no leaky expression). The fluorescence intensity increased with increasing concentrations of LA up to 10 mM. Furthermore, the system was also induced by the addition of 4HV, an intermediate of LA catabolism 25 . However, in contrast to LA, 4HV showed non-dosedependent (all-or-none) expression of the reporter protein (Fig. 2c). In comparison, LA induced a stronger expression than 4HV (Supplementary Figure S2). The LvaR/P lvaA system was not induced by Glu, Xyl, or 3HP. However, the expression was slightly affected when the system was induced with LA and Glu (Supplementary Figure S2). This might be due to glucose-mediated carbon catabolite repression 25 . Interestingly, the expression was higher when the system was induced with a mixture of LA and 4HV, compared to that observed with either LA or 4HV alone ( Supplementary Fig. 2). When compared with the chemical inducible system (P m by 3 MB; strain XSM01), the LvaR/P lvaA system (LRL01) showed at least 3.6-fold higher expression levels (Fig. 3). In contrast, the system showed 1.5-fold lower expression than the IPTG-inducible P tac system (LIT01). However, the basal expression (without inducer) was sixfold higher in LIT01 than in LRL01.
LA, a five-carbon versatile platform chemical, is identified as one of the "Top-12' of the most valuable sugarbased building blocks by the United States Department of Energy 26 . It can be obtained through acid-catalyzed dehydration and the hydrolysis of renewable cellulosic biomass especially from cellulose and hemicellulose 27 . LA catabolism, mediated by a seven-gene lva operon encoded by lvaABCDEFG, was previously elucidated in P. putida KT2440 25 . The lva operon is regulated by LvaR (transcriptional activator). LvaR activates the expression of downstream genes cloned under the control of P lvaA in the presence of LA and/or 4HV. The LvaR/P lvaA system exemplifies an interesting phenomenon in which the system can be induced by both the substrate and product. Because it is inexpensive, LA can be used as a starting material to produce a wide range of chemicals, including polymers, plasticizers, fuels, resins, pharmaceuticals, anti-freeze agents, and solvents [27][28][29][30] . The lvaABdeleted P. putida KT2440 strain produced a maximum of 50 g/L of 4HV with 97% conversion from LA using the LA-inducible LvaR/P lvaA system without the addition of any chemical inducers 21  www.nature.com/scientificreports/ source for strain KT2440 21,25 . As the lva operon-disturbed strain did not grow on LA, the added LA could not be intrinsically catabolized by the other endogenous system 21,25 . It is noteworthy that in strain KT2440, LA is catabolized to the central metabolites, acetyl-CoA and propionyl-CoA. Thus, the LA-inducible LvaR/P lvaA system could be useful for producing a wide range of chemicals from LA.  www.nature.com/scientificreports/  www.nature.com/scientificreports/ LA-inducible HpdR/P hpdH system. The fluorescence intensity of strain HRH01 harboring pHRH-eGFP + increased with increasing concentrations of 3HP (Fig. 2d). The addition of Glu or Xyl along with 3HP did not strongly affect induction (P > 0.05; Supplementary Figure S3). In addition, the system was not induced by the addition of either Glu or 4HV. However, this system was induced by the addition of LA ( Supplementary Figure S3). Therefore, the efficiency of the system was further evaluated using LA as an inducer (Fig. 2e). Compared with the HpdR/3HP system (HRH01 induced by 3HP), the HpdR/LA system (HRH01 induced by LA) was not very sensitive to the lowest concentration of LA (< 1 mM). However, the fluorescence increased with increasing concentrations of LA in a dose-dependent manner ( Supplementary Fig. 4). It may be advantageous to have an HpdR/P hpdH system that can be induced with a low-cost substrate, such as LA. To our knowledge, this is the first description of the induction of an HpdR/P hpdH system by LA. 3HP is a C3 biotechnologically important platform chemical. It is used as a potent building block for deriving both commodity and specialty chemicals 26 . It can be used as a carbon and energy source by Pseudomonas species, possessing 3HP catabolic pathways 31,32 . Zhou and coworkers reported that the 3HP-catabolism in P. denitrificans is activated by LysR-type transcriptional regulators in the presence of 3HP 32 . In P. putida, the 3HP catabolic genes and their respective transcriptional regulators (HpdR or MmsR) are arranged in two operons: (i) hpdH and its regulator hpdR in one operon, and (ii) mmsA and mmsB, and their regulator mmsR in another 31 . Upon binding of a 3HP molecule to HpdR, the resulting conformational changes in DNA enhance RNA polymerase binding in the promoter sequence of the downstream genes, thus allowing the transcription of the genes. The orthogonality of the P hpdH and P mmsA of P. putida KT2440 was demonstrated in E. coli and Cupriavidus necator 31 . Moreover, the study of cross-reactivity of some chosen compounds that are structurally similar to 3HP with the P hpdH system showed at least 5 to 30% relative induction level 31 . In the present study, LA (10 mM) exhibited around 82% relative induction level of the natural inducer 3HP (10 mM) with the HpdR/P hpdH system (Supplementary Figure S4). The relative induction level reached 99.8% with 20 mM of LA. This is advantageous because the system can be activated by two different potential substrates, C3 (3HP) and C5 (LA), which can be used as the starting materials to produce a wide range of chemicals. Furthermore, as both substrates can be completely catabolized by P. putida KT2440, a complete deletion or disruption of the catabolic pathways for LA and/or 3HP could be considered to achieve higher performance of the synthetic pathways.
Moreover, the induction of the MmsR/P mmsA closely related promoter system was also analyzed by estimating the fluorescence intensity of the constructed strain MRM01 with different concentrations of LA or 3HP (1-20 mM). Strain MRM01 did not show any considerable increase in fluorescence intensity when induced with LA or 3HP compared to the control. The system was not tightly controlled, and a highly leaky regulation was observed (Supplementary Figure S5).
Xyl-inducible XutR/P xutA system. The induction of P xutA with XutR was analyzed by estimating the fluorescence intensity of the strain XRX01 cultivated in M9Y medium supplemented with different concentrations of Xyl (0.1-20 mM). XutR/P xutA was strongly induced by the addition of Xyl (Fig. 2f); even at concentrations as low as 0.5 mM, a fluorescence intensity four-fold stronger than that in the control was observed (Supplementary Figure S6). The system was tightly controlled (no leaky expression, P < 0.01). Moreover, a dose-dependent fluorescence intensity was observed for up to 10 mM Xyl. Interestingly, the system was not induced by other C5 (arabinose) or C6 (Glu and mannose) sugars or other commonly tested inducers (LA, 4HV, and 3HP) (Supplementary Fig. 6). This is advantageous when a synthetic pathway necessitates the use of multiple genes under the control of different inducers to avoid cross-reactivity of the system, thus achieving smooth regulation of the metabolic pathways. In the present study, XutR/P xutA showed a slower induction compared to that of the other systems studied here ( Supplementary Fig. 6). However, the system showed the least basal expression (not leaky) and strong expression on induction, suggesting the non-specific transport of Xyl into the cell. The slower induction might be due to the absence of specific Xyl transporters, such as XylE found in E. coli, in P. putida KT2440 33 . A similar shortcoming was observed for other sugars, such as arabinose and rhamnose, for P araE or P rhaB in the strain KT2440 9 .
As P. putida KT2440 is not capable of catabolizing xylose, three different xylose utilization pathways, such as xylose isomerase, Dahms, and Weimberg pathways were implemented in KT2440 34,35 . These pathways catabolize xylose into various intermediates and introduce them into the pentose phosphate pathway, glycolysis, or tricarboxylic acid (TCA) cycle, respectively [33][34][35] . In P. fluorescens SBW25, the Xyl catabolic gene xutA (encoding xylose isomerase) is regulated by the xylose utilization regulator XutR 36 . XutR functions as a dimer. XutR/PA-II binding affinity is enhanced in the presence of Xyl and activates the promoter in a Xyl dose-dependent manner 36 . Thus, by engineering the strain for the catabolism of Xyl, a wide range of chemicals including amino acids, rhamnolipids, and xylitol could be produced 35 . After Glu, Xyl is the second most abundant sugar in nature and is a potential feedstock for the microbial production of biochemicals. Cellulosic hydrolysate can be efficiently used as it generally contains glucose and xylose in the major portion 37 . Therefore, the Xyl-inducible system could serve as an effective inducible expression system to produce various chemicals from this low-cost carbon source/substrate.

Intercomparison of different substrate-inducible expression systems. Intercomparison results
indicated that the maximum levels of expression were achieved by HpdR/P hpdH with the addition of 3HP and LA, XutR/P xutA with the addition of Xyl, and by LvaR/P lvaA with the addition of LA (Fig. 3). The saturating concentration of inducers for the highest expression level varied among the promoter systems: 5 mM Glu for HexR/ P zwf1 , 10 mM LA for LvaR/P lvaA , 10 mM Xyl for XutR/P xutA , and 10 mM 3HP or 20 mM LA for HpdR/P hpdH . The saturating concentration of inducers were found to be within the normal range of widely used concentrations as reported in a comprehensive study on the characterization of inducible promoters using 3 MB (P m ), rhamnose (P rhaB ), arabinose (P araB ), salicylate (P sal ), and IPTG (P T7-lac and P lacUV5 ) in P. putida KT2440 9 . The LvaR/P lvaA and www.nature.com/scientificreports/ XutR/P xutA showed the lowest basal level of expression in the absence of corresponding inducers. In contrast, MmsR/P mmsA was found to be the leakiest of all the tested expression systems. When compared with chemicalinducible systems, such as IPTG inducible-P LlacO1 and P tac or 3-methylbenzoate inducible-P m , our systems were shown to be equally effective (Fig. 3).

Conclusion
Five different carbon source-or substrate-inducible promoter systems were developed for the tunable control of gene expression in P. putida KT2440. The use of inexpensive and amicable substances to induce promoter systems could be a feasible strategy for the large-scale production of biochemicals. Here, we report a group of promoter systems that can be induced by low-cost substrates, such as Glu, Xyl, and LA. These inducer agents can be used as potent starting materials for both cell growth and the production of a wide range of biochemicals. Expression systems such as HpdR/P hpdH or LvaR/P lvaA could be used to produce target biochemicals from LA as a substrate.
As the xylose-inducible XutR/P xutA system was not affected by other substrates such as Glu or LA, this system could be efficiently used to produce biochemicals from xylose. This could even occur along with Glu or LA supplemented as a carbon source(s) or co-substrate(s). To produce a wide range of chemicals from glucose, the HexR/P zwf1 system could be used. Another 3HP-inducible system (MmsR/P mmsA ) did not seem to be regulated, resulting in similar GFP expression levels with or without 3HP or LA. The selection of the expression system depends on the target product, synthetic pathway, and substrate. As the efficiency of the reported systems was comparable with that of conventional chemical inducible systems, their limitations such as high cost, cell toxicity, and difficulties in downstream processes could be avoided. Hence, the newly investigated promoter systems have advantages for industrial and medical applications in P. putida KT2440.

Methods
The wild-type P. putida KT2440 strain was used to analyze the efficiency of inducible promoter systems. Five different promoter systems -(i) HexR regulating P zwf1 ; (ii) LvaR regulating P lvaA ; (iii) HpdR regulating P hpdH ; (iv) MmsR regulating P mmsA ; and (v) XutR regulating P xutA -were analyzed by expressing an enhanced GFP (eGFP + ) under the control of each promoter system and induced by the corresponding inducer. In addition, to compare the efficiency of the constructed systems, commonly used P LlacO1 , P tac , and P m promoter systems were constructed. All promoter systems were constructed uniformly using pPROBE_P yqjFmut _eGFP + (pBBR1-ori, Km R : a broadhost-range expression vector). All plasmids had the same 5' untranslated region (UTR) of the pPROBE plasmid. Each constructed plasmid was transformed into the electrocompetent cells of P. putida KT2440 for the GFP assay. The details of the constructed strains and plasmids, as well as selected promoter systems and their sources for PCR amplification, are provided in Supplementary Tables S1 and S2, respectively. The oligonucleotides used for PCR amplification of each promoter system are listed in Supplementary Table S3. The recombinant strains were cultured in Luria-Bertani medium (5 g yeast extract, 10 g peptone, and 10 g NaCl per liter) at 30 °C and 200 rpm overnight. Then, they were subcultured (initial optical density at 600 nm [OD 600 ] set to 0.1) in 20 mL of M9Y medium. When the OD 600 reached approximately 0.4, each culture (180 µL) was inoculated into a clear bottom Corning 96-well plate containing different concentrations of the corresponding inducers; they were then incubated at 30 °C with shaking in a microplate fluorescence reader (Infinite F200 PRO, Tecan, Grődig, Austria) to measure the GFP fluorescence intensity (gain of 30 at a wavelength of 485/535 nm) of the constructed systems. Fluorescence intensity was normalized based on the OD 600 value of the culture. Subsequently, the culture was diluted appropriately with phosphate-buffered saline, and flow cytometry analysis of GFP fluorescence was performed by fluorescence-activated cell sorting (FACSCalibur Flow Cytometer, BD Bioscience, CA, USA). Approximately 2 × 10 5 cells were analyzed per sample. The inducers and their concentrations tested for each promoter system are listed in Supplementary Table S4. All data represent the mean of two different experiments. Data were subjected to statistical analysis using SPSS (Version 11) software (SPSS Inc., Chicago, IL) to determine the level of significance. The detailed methods are provided in the supplementary information.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.