Volume 14, Issue 6 p. 2679-2685
Brief Report
Open Access

An automated DIY framework for experimental evolution of Pseudomonas putida

David R. Espeso

David R. Espeso

Systems Biology Program, Centro Nacional de Biotecnología-CSIC, Campus de Cantoblanco, Madrid, 28049 Spain

Search for more papers by this author
Pavel Dvořák

Pavel Dvořák

Department of Experimental Biology, Faculty of Science, Masaryk University, Brno, 62500 Czech Republic

Search for more papers by this author
Tomás Aparicio

Tomás Aparicio

Systems Biology Program, Centro Nacional de Biotecnología-CSIC, Campus de Cantoblanco, Madrid, 28049 Spain

Search for more papers by this author
Víctor de Lorenzo

Corresponding Author

Víctor de Lorenzo

Systems Biology Program, Centro Nacional de Biotecnología-CSIC, Campus de Cantoblanco, Madrid, 28049 Spain

*For correspondence. E-mail [email protected]; Tel. +34 91 585 4536; Fax +34 91 585 4506.

Search for more papers by this author
First published: 13 October 2020
Citations: 4

Funding information

This work was funded by the SETH Project of the Spanish Ministry of Science RTI 2018-095584-B-C42, the MADONNA (H2020-FET-OPEN-RIA-2017-1-766975), BioRoboost (H2020-NMBP-BIO-CSA-2018), SYNBIO4FLAV (H2020-NMBP/0500) and MIX-UP (H2020-Grant 870294) Contracts of the European Union and the S2017/BMD-3691 InGEMICS-CM Project of the Comunidad de Madrid (European Structural and Investment Funds) as well as by the Czech Science Foundation (19-06511Y).

Summary

Adaptive laboratory evolution (ALE) is a general and effective strategy for optimizing the design of engineered genetic circuits and upgrading metabolic phenotypes. However, the specific characteristics of each microorganism typically ask for exclusive conditions that need to be adjusted to the biological chassis at stake. In this work, we have adopted a do-it-yourself (DIY) approach to implement a flexible and automated framework for performing ALE experiments with the environmental bacterium and metabolic engineering platform Pseudomonas putida. The setup includes a dual-chamber semi-continuous log-phase bioreactor design combined with an anti-biofilm layout to manage specific traits of this bacterium in long-term cultivation experiments. As a way of validation, the prototype was instrumental for selecting fast-growing variants of a P. putida strain engineered to metabolize D-xylose as sole carbon and energy source after running an automated 42 days protocol of iterative regrowth. Several genomic changes were identified in the evolved population that pinpointed the role of RNA polymerase in controlling overall physiological conditions during metabolism of the new carbon source.

Introduction

The development of do-it-yourself (DIY) technical solutions (Moe-Behrens et al., 2013; de Lorenzo and Schmidt, 2017) for performing adaptive laboratory evolution (ALE) experiments (Portnoy et al., 2011; Dragosits and Mattanovich, 2013; LaCroix et al., 2017) is expanding the capabilities of researchers to integrate this attractive technique in their regular laboratory workflows. Some examples include the development of automatic microbial cultivation platforms operating mini-chemostats (Amanullah et al., 2010; Bergenholm et al., 2019), turbidostats (Marlière et al., 2011; Wong et al., 2018; McGeachy et al., 2019) or segregostats (Sassi et al., 2019). Yet, in any circumstance ALE experiments have to be designed taking into account the biological constrains of the evolving microbe and the target to achieve. One of such microorganisms of interest is the soil bacterium Pseudomonas putida (in particular strain KT2440) which, because of its distinct management of oxidative stress, has emerged as a prime host of engineered redox reactions (Nikel et al., 2014, 2016; Nikel and de Lorenzo, 2018). On this background, we set out to design and implement a DIY framework specifically developed for applying flexible ALE protocols to this bacterium for the sake of increasing its performance as whole-cell catalyst.

The construction details and every step of the implementation of the evolutionary device are fully disclosed in the Supplementary Information (Materials, Equipment and other procedural features: see Figs S1–S25 and Tables S1–S3). The reader is encouraged to access such accompanying particulars for a more complete comprehension of the technical solution hereby presented. The experimental setup was inspired in the turbidostat scheme proposed by Marlière et al. (2011) but was redesigned considering a number of constrains linked to the intrinsic biological features of the KT2440 strain of P. putida. One first consideration is that the specimen of interest belongs to a bacterial species that naturally sticks to surfaces and builds considerable amounts of biofilms (Auerbach et al., 2000; Espinosa-Urgel et al., 2000; Tolker-Nielsen et al., 2000; Espeso et al., 2018). Biofilm formation is operationally problematic, because it clogs culture conduits and selects for surface super-sticker variants. A second constraint is that P. putida KT2440 is strictly aerobic (Nikel and de Lorenzo, 2013; Kampers et al., 2019), and proper aeration is required to ensure culture viability and vitality during the long-term experiment. Furthermore, the evolutionary platform must ensure isolation of the manipulated culture to avoid contamination by microorganisms that may displace the template strain. Finally, growth media quality should be secured at all times for maximizing cell division and foster DNA replication – thereby increasing chances of mutations.

With these criteria in mind, an experimental setup was designed and assembled to execute a basic protocol for sustaining bacterial growth for long periods of time. Figure 1A shows the thereby implemented workflow process as a block diagram. The sketch illustrates a recurrent cycle in which a semi-continuous incubation of a culture is executed by a period of time defined by the user. The protocol included a control loop in the reactor incubation step (blue box) where the workflow was stalled in a periodic subroutine of incubation steps followed by optical density measurements at regular intervals (tsampling). Such a recurrent sequence ended when an upper threshold OD600 value was reached, allowing the workflow in this manner to keep advancing. Figure 1B shows the fluidic layout implemented for succeeding with this protocol. The basic setup includes (i) a bioreactor coupled to a photodetector to obtain OD600 readings, (ii) an auxiliary chamber to allow biofilm cleaning with NaOH and H2O, (iii) a rack of pumps to deliver the different chemicals and (iv) a group of valves to set the logic of the liquid transport through the circuit. The tubing is connected in a circular fashion with two independent waste outputs and venting connections to ensure an uninterrupted cell culture with sufficient aeration. The design was complemented with electronic and control layers, consistently designed to make possible the synchronized actuation of all these devices (Figs S1–S15). Additionally, this basic arrangement was complemented with the manufacturing of 3D-printed supports to spatially arrange pumps and valves (Figs S16–S18) and the assembly of an online optical density chamber to gather OD600 lectures (Figs. S19–S25).

Details are in the caption following the image

Schematic representation of the DIY device for experimental evolution of Pseudomonas putida.

A. Block diagram showing the process workflow to implement conceptualized from the ALE protocol taken to analyse as example. The workflow shows the different high-level actions to perform, their relative order of execution, timing, recurrence loops (i.e. blue square) and decision taking points (on yellow).

B. Conceptual scheme showing the actuators, sensors and vessels used to design the fluidic layer of the ALE experimental device. A set of peristaltic pumps, compressors and valves is in charge of transporting different chemicals to clean/wash the vessels and feed a bacterial culture constantly monitored by an optical sensor reader.

To test the efficacy of the thereby constructed DIY platform, we used a derivative of P. putida KT2440 that had been engineered to grow on D-xylose, a pentose abundant in hydrolysates of lignocellulosic materials (Chen et al., 2017). The construct at stake (named P. putida mk-1, Table S3) bears a large number of genomic modifications for increasing stability, raising the intracellular levels of ATP and NAD(P)H (Martínez-García et al., 2014a,b) and avoiding misrouting of intermediates during d-xylose metabolism. Specifically, P. putida mk-1 lacks flagella and other energy-draining cellular devices and has a deletion of gcd (thereby lacking glucose dehydrogenase). In addition, the strain bears a chromosomal implant of a synthetic xylABE operon encoding XylA (xylose isomerase), XylB (xylulokinase) and XylE (xylose-proton symporter) from Escherichia coli (Dvořák and de Lorenzo, 2018). To this end, the DNA segment bearing xylABE was assembled in a mini-Tn5 transposon vector (Martínez-García et al., 2014a,b) as described in the Supplementary information. During the construction of the test strain, the mobile element mini-Tn5 Sm:: [PEM7 → xylABE] was randomly inserted throughout the genome of strain P. putida EM42 ∆gcd (Table S3). The organization of the mini-Tn5 transposon was such that the xylABE operon could be expressed from the synthetic PEM7 promoter engineered in the mobile element as well as from readthrough transcription of nearby promoters close to the site of insertion.

Selection of the best grower clone on d-xylose as sole carbon source yielded the aforementioned strain P. putida mk-1 with the business DNA segment inserted in the midst of the locus PP_2260 (a putative glycerol-phosphate ABC transporter ATP-binding protein; Fig. 2C). Whether there was a benefit in the interruption of that ORF is unknown, but insertion of the DNA segment with [PEM7 → xylABE] in the chromosome secured the stable inheritance of the knocked-in trait during the course of the bioreactor experiment (Fig. 2A).

Details are in the caption following the image

Evolution of an engineered Xyl+ strain of P. putida along a 42 days protocol of iterative regrowth.

A. Optical density evolution during the 45 days period of the ALE experiment. The sawtooth pattern of the graph corresponds to the culture dilution dynamic (semi-continuous culture) used by the device, programmed to hold the optical density within an exponential growth regime with optical densities within the range [0.1–0.5].

B. Independent ALE validation experiments. Growth curve assays using shake flasks were performed to estimate the growth rates of template (mk-1, blue) and evolved (mk-2, red). For the tested conditions, mk-2 sample exhibited a 60% increment respect to template strain. The plot shows a fitting of three independent biological replicates. Asterisks indicate that both regressions passed t-test at 5% confidence (P < 0.05).

C. Mutations detected by whole genome sequencing of the P. putida mk-2 sample. A scheme of the P. putida mk-1 chromosome is depicted showing relevant genes and genomic changes detected after the evolution procedure. Genomic coordinates of PP_2260 (locus of mini-Tn5 insertion) and rpoC refer to P. putida EM42 ancestral strain. Inverted repeats ME-I and ME-O, defining the edges of mini-Tn5, are also shown by black arrowheads with xylABE cluster in between. Locations of detected mutations are denoted by red asterisks. PEM7 sequence features −35 and −10 boxes in high case and underlined text, while deletion found in mk-2 genome appears underlined in red colour. Single nucleotide changes found in rpoC and xylE appear in brackets: wild type and mutated codon are depicted with mutated site in high case. The amino acid change and position in the polypeptide are also shown below.

Next, strain P. putida mk-1 was inoculated in an intermediate reactor chamber with an operative volume of 20 ml and containing around 109 cultivated cells with an OD600 bounded within the range [0.1–0.5]. During a 45 days period, cells were recurrently incubated and diluted (Fig. 2A) using M9 minimal medium supplemented with 0.2% (w/v) d-xylose and 60 μg ml−1 streptomycin. Under these simple conditions, the setup selects for faster growers which – in case of appearance – should bear mutations that increase overall physiological fitness and/or improve nesting of the implanted metabolic segment in the background biochemical network of P. putida. The progress of the experiment is shown in Figure 2A. At the end of the corresponding period of time, an increase in growth rate of the population present in the culture became clearly noticeable, same as the fact that beneficial changes occurred probably mainly during the initial phase of the experiment (Fig. 2A, Fig. S26). To examine the basis of such a change, samples were collected from the reactor and further inspected. First, the evolved sample (hereafter called mk-2) was verified as an authentic descendant of the original strain. For this, the mk-2 sample was plated on LB and M9 + 0.2% (w/v) citrate agar dishes and cells were streaked out to discard any contamination. Strain clonal identity was confirmed through PCR of the genome with primers 5ʹCTTCAGCTCTTCGCTGTACA3ʹ and 5ʹGCGTGCGCTACAACCTTAC3ʹ that amplify the region surrounding the deletion of the glucose dehydrogenase gene (PP_1444) present in the template strain and which acted as a diagnostic signature. Second, the growth rate of the evolved culture on d-xylose as the only C source was re-assessed in respect to the precursor strain performing independent growth curve assays in Erlenmeyer flasks. Regression slopes comparing the two (Fig. 2B) indicated that the evolved specimen grew a 60% faster than template strain. Finally, the genomes of the original P. putida mk-1 strain and the evolved counterpart mk-2 were sequenced to find mutations that could account for the observed shift in the growth phenotype.

While no modifications became apparent in the bacteria of reference, the faster-growing derivative bore 3 conspicuous changes in its chromosome. The first modification of the evolved genome was found in the rpoC gene of P. putida, which encodes the βʹ subunit of RNA polymerase. The rpoC of gene of mk-2 had a point mutation C → T in codon 51 (cCt → cTt) causing a quite drastic change Pro51Leu. The emergence of this modification acted in fact as a descriptor of the efficacy of the evolutionary experiment. This is because as a large number of rpoC mutations have been reported in the course of laboratory evolution studies aiming to increase E. coli growth rate (Cheng et al., 2014; Wytock et al., 2018; Kavvas et al., 2020). Therefore, the Pro51Leu change plausibly reflects a similar adaptation in P. putida. The other two mutations were identified in the implanted xylose cluster. One of them involved a single base change G → T in the codon 33 of the xylE gene (gGt → gTt) which translates into a Gly33Val amino acid change. The xylE product is a xylose-proton symporter of E. coli composed by several transmembrane domains connected by periplasmic/cytoplasmic amino acid stretches (Davis and Henderson, 1987), and Gly33 is located in a periplasmic side close to the H+ coupling site Asp27 (Madej et al., 2014). While a number of loss-of-function mutations have been reported for xylE (Sun et al., 2012), to the best of our knowledge no changes are known to enhance xylose transport. The Gly33Val observed in mk-2 could be one of them, an issue that deserves further studies. Finally, a 13-bp deletion removing part of the PEM7 −10 box was observed also in the faster-growing culture. As mentioned above, PEM7 is a strong synthetic promoter engineered for driving the expression of xylABE cluster in P. putida mk-1 strain. The loss of part of the −10 box (Collado-Vides et al., 1991; Meysman et al., 2014) is expected to reduce the promoter activity in the mk-2 cells. Selective pressure to curb PEM7 strength might be related to the fact that overproduction of the XylE transporter is toxic to P. putida cells (unpublished data). The coexistence in mk-2 of mutations anticipated to both decrease xylE transcription and improve XylE efficiency could reflect a solution to the conflict between the negative effects of overproducing a membrane protein and the need to secure a sufficient inflow of the carbon source for a faster growth. Note that – as discussed above – even complete elimination of the PEM7 promoter of the genomic implant [PEM7 → xylABE] could still deliver expression of the operon owing to readthrough transcription from promoter(s) outside the mini-Tn5 insertion (Fig. 2C).

Whether the effects of these three mutations found in the evolved, faster-growing sample are additive, synergistic or altogether independent is beyond the scope of this technical note and will be the subject of subsequent studies. Correspondingly, further rational engineering cuts and ALE with the constructed system are being considered to remove additional metabolic bottleneck(s) that could have prevented achieving even faster growth of P. putida recombinant on the non-native substrate during the evolution experiment (Elmore et al., 2020). Yet, the data presented above accredits the power of the simple and affordable DIY setup described here to generate phenotypes of considerable biotechnological interest in the synthetic biology chassis and metabolic engineering platform Pseudomonas putida. Besides the enhancement of catabolic traits, the authors foresee the use of the bespoken device also for the evolution of biosynthetic pathways in P. putida and other bacterial cell factories. As additional modules, e.g., for absorbance or fluorescence quantification, can be easily integrated into the presented setup, we entertain the use of this framework also for the accelerated evolution of industrially relevant strains equipped with genetically encoded product-responsive biosensors (Mahr et al., 2015).

Acknowledgements

Authors are indebted to Alfonso Jaramillo, Rui Rodrigues and Philippe Marliere for fruitful discussions and advice on automation of protocols, hardware support and electronic device design, assembly and maintenance. Lee Cronin and Soichiro Tsuda are gratefully acknowledged for their valuable help with 3D printing technology and Arduino programming. This work was funded by the SETH Project of the Spanish Ministry of Science RTI 2018-095584-B-C42, the MADONNA (H2020-FET-OPEN-RIA-2017-1-766975), BioRoboost (H2020-NMBP-BIO-CSA-2018), SYNBIO4FLAV (H2020-NMBP/0500) and MIX-UP (H2020-Grant 870294) Contracts of the European Union and the S2017/BMD-3691 InGEMICS-CM Project of the Comunidad de Madrid (European Structural and Investment Funds) as well as by the Czech Science Foundation (19-06511Y).

    Conflict of interest

    The authors declare no competing financial interest.