Engineering Pseudomonas putida for isoprenoid production by manipulating endogenous and shunt pathways supplying precursors
The soil bacterium Pseudomonas putida is a promising platform for the production of industrially valuable natural compounds. In the case of isoprenoids, the availability of biosynthetic precursors is a major limiting factor. In P. putida and most other bacteria, these precursors are produced from pyruvate and glyceraldehyde 3-phosphate by the methylerythritol 4-phosphate (MEP) pathway, whereas other bacteria synthesize the same precursors from acetyl-CoA using the unrelated mevalonate (MVA) pathway.
Here we explored different strategies to increase the supply of isoprenoid precursors in P. putida cells using lycopene as a read-out. Because we were not aiming at producing high isoprenoid titers but were primarily interested in finding ways to enhance the metabolic flux to isoprenoids, we engineered the well-characterized P. putida strain KT2440 to produce low but detectable levels of lycopene under conditions in which MEP pathway steps were not saturated. Then, we compared lycopene production in cells expressing the Myxococcus xanthus MVA pathway genes or endogenous MEP pathway genes (dxs, dxr, idi) under the control of IPTG-induced and stress-regulated promoters. We also tested a shunt pathway producing isoprenoid precursors from ribulose 5-phosphate using a mutant version of the Escherichia coli ribB gene.
The most successful combination led to a 50-fold increase in lycopene levels, indicating that P. putida can be successfully engineered to substantially increase the supply of metabolic substrates for the production of industrially valuable isoprenoids.
KeywordsDXS Isoprenoids Lycopene MEP pathway Pseudomonas putida
Isoprenoids (also known as terpenoids) constitute the largest class of natural products, with tens of thousands of different structures found in all kingdoms of life . Many of them are high value compounds due to their diverse range of applications in areas such as human nutrition and pharmaceutical industry . Most economically-relevant isoprenoids are typically found at low levels in their natural sources. This, together with the increasing demand for novel isoprenoid structures with new or improved bioactivities, has spurred many metabolic engineering and synthetic biology projects focused on boosting isoprenoid production. Much effort has been devoted towards their synthesis in microbial hosts, mostly employing Escherichia coli and Saccharomyces cerevisiae. However, interference with the endogenous cell metabolism and toxic effects associated with the production of most isoprenoids, among others, are important issues that often prevent reaching commercially relevant yields .
In this context, the stress-resistant bacterium Pseudomonas putida is a promising alternative platform for isoprenoid production and storage. Some P. putida strains can tolerate much higher level of isoprenoids than other microorganisms (e.g. its growth is not affected by concentrations of the monoterpene geranic acid that are toxic to E. coli or S. cerevisiae ). The high tolerance to otherwise toxic compounds of P. putida cells is in part due to fine-tuning of lipid fluidity to adjust membrane functions, activation of a general stress-response system, increased energy generation and induction of efflux pumps . Additionally, P. putida is amenable to targeted genetic modifications and it has been targeted in biofactories design and biodegradation studies due to its robust metabolism, among others characteristics. Several industrial processes based on this bacterium are being exploited for the production of fine chemicals (e.g. 2-quinoxalinecarboxylic acid, 5-methylpirazine-2-carboxylic acid, chiral amines and 4-(hydroxypyridin-3-yl)-4-oxobutyrate), and the production of many different compounds, including isoprenoids, is under study . P. putida has actually been used for the production of some isoprenoids using biotransformation approaches to obtain oxidation products of the plant monoterpene limonene  or de novo biosynthesis of the monoterpene geranic acid  or the carotenoids zeaxanthin and β-carotene [8, 9, 10]. Despite the low number of available studies, the achievement of substantial yields of isoprenoids such as geranic acid and zeaxanthin and the high tolerance of P. putida to some of these compounds, usually toxic for other microbial models, illustrate the high potential of this bacterium to become a platform of choice for the industrial production of valuable isoprenoids. An important conclusion of these studies is that enhancing the supply of isoprenoid precursors is crucial for success as the bacterium only produces minor amounts of such precursors .
Here we have compared the performance of previously known and novel strategies to improve the supply of IPP and DMAPP for isoprenoid biosynthesis in KT2440, the best characterized P. putida strain. Using as read-out the production of lycopene, a red carotenoid pigment with great commercial and nutritional importance, we increased isoprenoid precursor levels by introducing an exogenous MVA pathway for IPP and DMAPP synthesis, by producing DXP via nDXS, and by expressing DXS and other MEP pathway enzymes under the control of different promoters.
Pseudomonas putida KT2440 can be engineered for lycopene production
Cell density (turbidity at 600 nm) and lycopene accumulation (measured by HPLC) in cells cultured during 24 h in LB
Cell density (OD600nm)
Lycopene accumulation (ng/ml culture)
5.417 ± 0.791
0.000 ± 0.000
5.293 ± 0.623
0.000 ± 0.000
4.930 ± 1.191
1.223 ± 0.185
6.157 ± 0.551
0.106 ± 0.026
Once we confirmed the ability for lycopene synthesis of P. putida cells, the crtE–crtB–crtI operon with its own promoter was inserted into the Tn7 att site in the chromosome of KT2440, using a Tn7-based delivery system , to yield the strain KTLYC. Strains KT2440 and KTLYC were grown for 24 h in LB and lycopene accumulation was then measured by HPLC. Growth of both strains was basically identical (Fig. 2a) whereas, as expected, lycopene was only detected in KTLYC (Table 1). However, the amounts produced in this strain (0.106 ± 0.026 ng/ml culture) were an order of magnitude lower than those found in KT2440 cells containing the plasmid p421-LYC (Table 1), perhaps due to differences in copy number or DNA context (genomic vs. plasmidial). In any case, these results allowed to conclude that, under our experimental conditions, the intermediate steps of the endogenous MEP pathway are not saturated in strain KTLYC. We therefore selected this strain for exploring ways of increasing the supply of MEP-derived isoprenoid precursors (IPP and DMAPP) in P. putida cells.
Lycopene production in P. putida KTLYC is enhanced by expressing an exogenous MVA pathway
Up-regulation of endogenous DXS greatly improves isoprenoid precursor supply in KTLYC cells
GGPP required for lycopene production is formed by adding three IPP molecules to one DMAPP acceptor, whereas bacterial isoprenoids are most often derived from farnesyl diphosphate (FPP), that is formed from two IPP and one DMAPP . Because an unbalanced production of isoprenoid precursors has been shown to be detrimental for bacterial growth , we next investigated whether enhancing the interconversion of IPP and DMAPP might positively impact cell growth and lycopene accumulation in strains overexpressing DXS. The IPP/DMAPP ratio is regulated by the enzyme isopentenyl diphosphate isomerase (IDI) that interconverts IPP and DMAPP molecules (Fig. 1). The idi gene from E. coli was cloned into the p424-DXS vector and the resulting p424-IDI-DXS plasmid was introduced into the strain KTLYC. Positive transformants were grown with or without IPTG as described above (Fig. 4c) and both cell growth and lycopene accumulation were recorded at 24 h (Fig. 4d). In the absence of IPTG, lycopene production tended to be slightly higher in cells harboring p424-IDI-DXS compared to p424-DXS, but the difference was not statistically significant. Induction with increasing concentrations of IPTG did not improve cell growth or lycopene production on cells harboring p424-IDI-DXS (Fig. 4d) compared to p424-DXS (Fig. 4b). On the contrary, growth of KTLYC(p424-IDI-DXS) cells was reduced compared to that of IDI-lacking strains at IPTG concentrations higher than 0.1 mM. Addition of 1 mM IPTG to the medium also reduced the capacity of the cells for lycopene accumulation (Fig. 4). In summary, IPP/DMAPP imbalance might influence lycopene biosynthesis but it appears not to be a major factor limiting lycopene yields in P. putida cells. While DXS activity is much more relevant to regulate flux to isoprenoid biosynthesis, a fine control of its levels is critical to prevent deleterious effect on cell growth.
Use of the stress-regulated promoter PkatA to express IDI and DXS has a modest effect on lycopene production
With the aim of adjusting enhanced isoprenoid precursor supply to actual cell metabolism and growth, we next used the PkatA promoter  to construct a synthetic PkatA-idi-dxs operon. KTLYC strains harboring the plasmids pPkatA-IDI-DXS or p424-IDI-DXS, as well as the empty pSEVA424 vector, were cultured in LB without IPTG and lycopene accumulation levels were analyzed every 24 h during 5 days. After the first 24 h (Fig. 6b) KTLYC(pPkatA-IDI-DXS) cells showed fourfold higher levels of lycopene than KTLYC(p424) but half as much compared to KTLYC(p424-IDI-DXS). Neither KTLYC(pPkatA-IDI-DXS) nor KTLYC(p424-IDI-DXS) cells produced more lycopene at longer incubation times (Fig. 6c). It is therefore possible that the Ptrc promoter present in the p424-IDI-DXS construct leads to higher enzyme levels than the PkatA promoter even without the IPTG inducer, hence being more efficient to increase the MEP pathway flux.
Depletion of pyruvate and GAP might contribute to the cell growth defects associated with DXS overexpression
Ru5P, the substrate of nDXS, is an intermediate of the pentose phosphate pathway, which is not very active in P. putida [6, 23]. In complete (LB) medium, the addition of 50 mM glucose had no effect on lycopene production when KTLYC(p424-nDXS) cells were induced with 1 mM IPTG (Fig. 7). In minimal (M9) medium, similar levels of lycopene were produced with glucose, citrate, or succinate. It is interesting, however, that whereas lycopene accumulation in KTLYC(p424-DXS) cells was much lower in M9 than in LB medium, KTLYC(p424-nDXS) cells showed similar lycopene levels in both media (Fig. 7). We therefore conclude that the main factor limiting isoprenoid production in the KTLYC(p424-nDXS) strain is not Ru5P availability but the capacity of nDXS to synthesize DXP.
Combination of DXS and DXR further increases lycopene accumulation
Improving the supply of IPP and DMAPP for the biotechnological production of value-added isoprenoids in microorganisms has been the target of many studies. However, many other factors need to be taken into account to achieve an efficient production of isoprenoids without affecting the fitness of the cell. They include (A) depletion of metabolic substrates, (B) toxicity of accumulated intermediates or end products, and (C) saturation of the cell storage capacity. Strategies to deal with the last two problems are abundant, but less attention has been paid to the first, despite the negative growth impact of diverting the pools of central metabolites such as pyruvate and GAP to the synthesis of IPP and DMAPP. While most studies aimed to improve IPP and DMAPP supply have been done with well-characterized organisms such as E. coli (a model for bacteria that use the MEP pathway) and S. cerevisiae (a model for yeasts, which only use the MVA pathway), other microorganisms that show additional advantages for the production and accumulation of high isoprenoid titers remain scarcely explored. This is the case with P. putida, a bacterium that shows a very high tolerance to toxic compounds such as monoterpenes [4, 5]. An additional advantage of P. putida is that glucose is metabolized to produce balanced amounts of the MEP pathway substrates pyruvate and GAP [6, 10, 23]. These central metabolic substrates are produced in non-equivalent amounts in other bacteria such as E. coli, where this imbalanced availability of pyruvate and GAP can limit the activity of the MEP pathway . Increasing the levels of both pyruvate and GAP in P. putida actually led to an increase (1.3-fold) in carotenoid (β-carotene) production , suggesting that the supply of metabolic substrates limits the MEP pathway flux. Our data reported here further support the conclusion that ensuring a proper supply of pyruvate and GAP is also important to upregulate isoprenoid production while preventing major interferences with overall microbial metabolism leading to defective growth.
To investigate different ways of stimulating the metabolic flux through the endogenous MEP pathway to isoprenoid precursors, we generated a P. putida strain (KTLYC) with an exogenous 3-gene operon to produce lycopene as a read-out. By transforming this strain with plasmids to either overexpress endogenous MEP pathway genes or incorporate alternative shunt pathways, we could multiply lycopene production (i.e. IPP and DMAPP supply) by 3 (with nDXS; Fig. 8), 7 (with the MVA pathway; Fig. 3), 25 (with DXS; Fig. 4) and 50 (with DXS and DXR; Fig. 10). It is most likely, however, that further optimization of at least some of these strategies or a combination of several could result in even higher increases. Additive and synergistic effects could also be achieved by improving the supply of MEP pathway substrates. A recent example is the rewiring of the P. putida core metabolism for glucose assimilation from the native Entner–Doudoroff to the linear Embden–Meyerhof–Parnas glycolysis, which improved the production of both pyruvate and GAP . The metabolic plasticity shown by P. putida, which is able to metabolize a broad variety of carbon sources including fatty acids, glycerol and aromatic compounds , provides an additional opportunity to redesign the pools of substrates required for the synthesis of IPP and DMAPP. Feeding the bacteria with different carbon sources might help accumulate higher pools of pyruvate, GAP (i.e. glucose) or acetyl-CoA (i.e. fatty acids) and the combination of different carbon sources could help release the competition of the productive (exogenous) pathways with the housekeeping (endogenous) metabolism. The hierarchical utilization of different carbon sources  might limit our ability to combine fully active metabolic pathways to simultaneously feed productive and housekeeping pathways in the cell. However, our knowledge of the mechanisms controlling this hierarchy are constantly improving [27, 28], providing new engineering opportunities.
Among the novel strategies tested here, the use of nDXS to produce DXP from Ru5P was most successful as it improved lycopene accumulation up to threefold without negatively affecting cell growth (Fig. 8). While this increase is still far from that achieved with the canonical DXS enzyme (Fig. 4), it is important to keep in mind that nDXS was identified as a mutant enzyme which acquired the ability to produce DXP as a result of a point mutation and it is not optimized for DXP synthesis. Our estimation is that only about 10% of the Ru5P substrate is transformed into DXP by nDXS. Increasing the DXS activity of the mutant enzyme should be feasible and even fast if using the new automatized platforms available for adaptive laboratory evolution approaches [29, 30]. An optimized nDXS would allow a complementary strategy using the canonical DXS to push the MEP pathway using pyruvate and GAP up to the limits that show no effect on growth with an additional DXP input obtained from the transformation of Ru5P by nDXS. The availability of the crystal structure of the wild-type ribB protein  should allow to infer how the newly selected mutations could affect the DXS activity of nDXS, an information that could be very valuable to eventually create synthetic enzymes able to produce DXP from Ru5P or from substrates that do not interfere with the endogenous MEP pathway or other essential metabolic processes in the bacteria. As a complementary strategy, the supply of Ru5P might be enhanced by rewiring the pentose phosphate pathway. Ru5P is a C5 substrate and hence it is converted into DXP by nDXS without losing carbon (unless the reaction catalyzed by DXS, which starts with two C3 precursors and involves a decarboxylation). Together, the use of optimized versions of nDXS either alone or in combination with other strategies might increase IPP and DMAPP production to new high levels.
Another novel strategy that we tested was the fine regulation of DXS expression by using an endogenous promoter that responded to cell growth stress (Fig. 6). Using the PkatA promoter we only achieved a fourfold increase in lycopene accumulation compared to empty plasmid controls, suggesting that the transcriptional activity of this promoter in the conditions tested is too low to effectively boost the MEP pathway and hence to reach the stress conditions that could eventually repress transcription. It is likely that the use of stronger but still stress-responsive promoters might work better. The PkatA promoter might be engineered to increase its transcriptional activity while preserving its regulatory traits. Alternatively, analysis of the transcriptome of KTLYC(p424-DXS) cells grown in the presence of different concentrations of IPTG should allow to identify genes that are highly expressed when cells grow normally but are repressed when cell growth is compromised in an IPTG-dependent fashion (i.e. when DXS activity is too high). The promoters of such genes might be then used in synthetic operons to increase the supply of DXS (and IDI, DXR, nDXS, etc.) and hence the availability of isoprenoid precursors under optimal growth conditions but down-regulate them when growth-repressing effects appear. This will give time to the cells to readjust their metabolism and retake isoprenoid overproduction afterwards.
The more “classical” approaches that we tested, i.e. the use of an exogenous MVA pathway to produce IPP and DMAPP from acetyl-coA and the overexpression of MEP pathway genes, led us to conclude that we cannot directly extrapolate the results obtained with one bacterium (e.g. E. coli) to engineer another bacterium (e.g. P. putida). For example, using the MVA pathway typically works better than DXS up-regulation in E. coli [32, 33, 34] but we observed the opposite behavior in P. putida. A similar observation was previously reported in cyanobacteria growing in photoautotrophic conditions to produce isoprene . The presence of the bacterial MVA pathway operon improved lycopene accumulation by sevenfold (Fig. 3), similar to the tenfold increase reported to cause the same operon on monoterpene production in P. putida DSM 12264 cells . By contrast, the presence of a plasmid containing DXS was able to promote lycopene accumulation in KTLYC cells by 15-fold in the absence of any induction (Fig. 4b). The almost complete lack of effect of IDI overproduction on lycopene synthesis when the limiting role of DXS is released in P. putida (Fig. 4d) is also different to that observed in E. coli .
Production of isoprenoids has been extensively engineered in classical microbial chassis such as E. coli and S. cerevisiae but much less effort has been made in P. putida despite the unique characteristics that make this organism a very attractive host for the industrial production of these metabolites. The present work analyzed different ways of enhancing the supply of isoprenoid precursors (IPP and DMAPP) in P putida strain KT2440 using lycopene as read-out. The most successful combination led to a 50-fold increase of lycopene accumulation. Under our experimental conditions, the production of limonene in P. putida cells engineered to synthesize this monoterpene was also boosted by the same MEP pathway constructs that improved lycopene levels, supporting the conclusion that the increases we observed in downstream isoprenoid products are indicative of a major up-regulation of the metabolic flux to IPP and DMAPP. Our work therefore represents an initial but important step towards engineering P. putida cells to produce high isoprenoid titers. The unique metabolic plasticity of P. putida together with its astonishing capacity to tolerate high levels of toxic isoprenoids such as monoterpenes and its amenability to genetic engineering make this bacterium a promising candidate to become a favorite microbial cell factory to produce high-value isoprenoids.
Bacterial strains and culture media
Escherichia coli and P. putida cells were cultured at 37 °C and 30 °C, respectively, with strong aeration. LB broth (10 g/l tryptone; 5 g/l yeast extract, 10 g/l NaCl) was used as complete growth medium. M9 minimal salts medium  was supplemented with trace elements , and either 30 mM citrate, 30 mM succinate or 50 mM glucose as the carbon source. When needed, antibiotic were added at the following concentrations: kanamycin 25 μg/ml, gentamicin 25 μg/ml, streptomycin 50 μg/ml (or 500 μg/ml when the host strain harboured a gentamicin-resistance determinant), ampicillin 100 μg/m. Cell density was monitored by measuring turbidity (optical density) at 600 nm (OD600).
To generate strain KTLYC, a 3788 bp DNA fragment containing the P. ananatis genes crtE, crtB and crtI and their own promoter was PCR-amplified from the vector pACCRT-EIB  and cloned into the NotI site of the Tn7-based delivery vector pTn7-M . The generated plasmid, named pTn7-M-LYC, was transferred to P. putida strain KT2440 by tetra-partite mating using E. coli cc118λpir (pTnS1) as the supplier of Tn7 transposase and E. coli HB101 (pRK600) as the transfer function donor, as previously described . The correct insertion of the transposon at the att Tn7 site of the P. putida genome was checked by sequencing.
Plasmid p421-LYC contains the crtE, crtB and crtI genes and their own promoter from P. ananatis. To generate this vector, the same 3788 bp DNA fragment inserted in the genome of the strain KTLYC was cloned in the NotI site of the vector pSEVA421 . The pMiS1-ges-mva vector  contains the MVA pathway from M. xanthus under the control of the promoter rhaPBAD; transcription from this promoter can be induced by addition of L-rhamnose. To overproduce P. putida or E. coli enzymes involved in IPP and DMAPP synthesis, genes coding for these enzymes were cloned downstream of the IPTG-inducible promoter Ptrc in the vector pSEVA424 . To construct the p424-DXS vector, a DNA fragment including the rbs region of the vector pET23 and the coding region of P. putida dxs gene was cloned between the sites EcoRI and HindIII of plasmid pSEVA424. Plasmid p424-nDXS was made in a similar way using a version of the E. coli ribB gene harboring the G108S mutation . The p424-DXS plasmid was used as a backbone to clone the idi gene from E. coli (upstream of dxs, in the EcoRI site, rendering plasmid p424-IDI-DXS) or the P. putida dxr gene (downstream of dxs, between HindIII and SpeI, rendering p424-DXS-DXR). The Ptrc promoter in the vector p424-IDI-DXS was replaced with the P. putida PkatA promoter after PCR amplification using primers katA-PacI-F (5′-T G T T A A T T A A G A C G T T C T G T C T G C T A C) and katA-AvrII-R (5′-C T C C T A G G G C C G G C T A A T C G G) and cloning between the PacI and AvrII sites to create plasmid pPkat-IDI-DXS. All plasmid constructs were sequenced and the plasmids transferred to P. putida by triparental mating using pRK600 as a donor of transfer functions.
Quantification of lycopene accumulation
An overnight culture of the strain of interest was diluted to a final turbidity (OD600) of 0.05 in fresh LB medium. The indicated amount of IPTG or L-rhamnose was added to induce transcription from promoters Ptrc or Pram respectively. Cells were allowed to grow at 30 °C with vigorous aeration. After 24 h, cell growth was measured (OD600) and 5 ml aliquots were taken to lycopene extraction with acetone and quantification by measuring absorbance at 472 nm, as described at Rodríguez-Villalón et al. . Three independent assays were performed. When indicated, lycopene levels were also measured by HPLC in an Agilent 1200 series HPLC system (Agilent Technologies), as previously described .
The KTLYC(p424-IDI-DXS) strain was grown in LB medium at 30 °C in aerated flask and after 1 h the indicated IPTG concentration was added. At the start of the stationary phase (OD600 = 2.2), 1 ml of samples were collected, harvested by centrifugation and immediately frozen at − 70 °C. RNA was purified from cell pellet using RNeasy RNA Purification Kit (Qiagen) following the manufacturer’s instructions. Purified RNA was treated with RNase-free DNase I (Ambion), as specified by the manufacturer. RNA was quantified using a NanoDrop (Thermo Scientific) and its integrity was analyzed by agarose gel electrophoresis. The absence of DNA was confirmed by real-time PCR using primers for rpoN, as previously described .
Oligonucleotides used for real-time RT-PCR
We thank Victor de Lorenzo for pSEVA vectors, Jens Schrader for plasmids pMiS1-ges and pMiS1-ges-MVA and Fernando Rojo for the P. putida strains. Technical support from Mª Rosa Rodríguez Goberna is greatly acknowledged.
SHA, JPG and MRC conceived the whole study and designed the research; MRC supervised the study; SHA, JPG and DMS performed experiments; SHA, JPG, DMS and MRC analyzed and discussed data; SHA and MRC wrote the paper and JPG revised it. All authors read and approved the final manuscript.
This work was funded by Grants PCIN-2015-103 and BIO2017-90877-REDT from the Agencia Estatal de Investigación and 2017SGR-710 from Generalitat de Catalunya to MRC. We also thank the financial support of the MINECO Severo Ochoa Programme for Centres of Excellence in R&D 2016-2019 (SEV-2015-0533) and the Generalitat de Catalunya CERCA Programme to CRAG. JPG was supported by a Marie Curie International Outgoing Fellowship within the 7th European Community Framework Programme (Project Number 627639). The funding bodies had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
- 1.Dictionary of Natural Products 28.1. http://dnp.chemnetbase.com/faces/chemical/ChemicalSearch.xhtml. Accessed 17 July 2019.
- 11.Boronat A, Rodríguez-Concepción M. Terpenoid biosynthesis in prokaryotes. In: Schrader J, Bohlmann J, editors. Biotechnology of isoprenoids. Cham: Springer; 2014.Google Scholar
- 36.Sambrook J, Russell D. Molecular cloning: a laboratory manual. Cold Spring Harbor: Cold Spring Harbor Laboratory Press; 2001.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.