ABSTRACT

   Microbial transformation is a favored approach for environmental
   remediation. However, the effectiveness of microbial remediation has
   been limited by the lack of chassis cells with satisfactory contaminant
   degradation performance. Pseudomonas putida B6-2, with a wide substrate
   spectrum and high solvent tolerance, is a chassis strain with great
   potential for application in environmental remediation. Here, guided by
   bioinformatic analyses and genome-scale metabolic model (GEM)
   predictions, we successfully optimized P. putida B6-2 by rationally
   reducing its nonessential genetic components and generating a more
   robust genome-streamlined strain, P. putida BGR4. Several improvements
   were observed compared with the original P. putida B6-2 strain,
   including a 1.4 × 10^5-fold increase in electroporation efficiency, an
   8.3-fold increase in conjugation efficiency, improved glycerol
   utilization capability, and increased phenol utilization after
   heterologous expression of the phenol monooxygenase encoded by
   dmpKLMNOP. Additionally, P. putida BGR4 exhibited enhanced tolerance to
   several stressors, including starvation, oxidative stress, and DNA
   damage. Transcriptomic analysis revealed that genome streamlining led
   to the upregulation of genes involved in the “carbon metabolism” and
   “tricarboxylic acid cycle” pathways in P. putida BGR4, which likely
   contributed to the superior phenotype of P. putida BGR4 in terms of
   carbon source utilization and contaminant degradation capabilities.
   Furthermore, the absence of four prophages was identified as a
   potential cause of the enhanced stress resistance observed in P. putida
   BGR4. Overall, we developed a combined genome-streamlining strategy
   involving bioinformatic analyses and GEM predictions and generated a
   more robust chassis strain, P. putida BGR4, which expands the
   repertoire of chassis cells for environmental remediation.

IMPORTANCE

   Despite the development of many chassis cells, there is still a lack of
   robust chassis cells with satisfactory contaminant degradation
   performance. Targeted genome streamlining is an effective way to
   provide powerful chassis cells. However, genome streamlining does not
   always lead to the improved phenotypes of genome-streamlined chassis
   cells. In this research, a novel procedure that combined bioinformatic
   analyses and GEM predictions was proposed to guide genome streamlining
   and predict the effects of genome streamlining. This genome
   streamlining procedure was successfully applied to Pseudomonas putida
   B6-2, which was a chassis cell with great potential for application in
   environmental remediation and resulted in the generation of a more
   robust chassis cell, P. putida BGR4, thereby providing a superior
   chassis cell for efficient and sustainable environmental remediation
   and a valuable framework for guiding the genome streamlining of strains
   for other applications.

   KEYWORDS: Pseudomonas putida, genome streamlining, genome-scale
   metabolic models (GEMs), stress tolerance, environmental remediation

INTRODUCTION

   With the rapid development of industry and agriculture, an increasing
   number and variety of refractory contaminants are being produced every
   year, posing a major threat to public health ([34]1). Strategies to
   degrade these refractory contaminants have received widespread
   attention. Advantageous properties, such as sustainability and
   environmental friendliness, have made bacteria favorable candidates for
   environmental remediation. However, most microbes that contain only the
   catabolic genes for a single compound cannot be effectively used in
   most pollution scenarios involving multiple refractory contaminants
   ([35]2). Artificial combinations of microorganisms with different
   abilities to degrade refractory contaminants may solve this problem,
   but negative interactions between the components of the microbial
   consortium would disturb the microbial community and limit the
   contaminant degradation efficiency of the consortium ([36]3).

   The emergence and rapid development of synthetic biology has made it
   possible to construct robust chassis cells with combined pollution
   degradation capabilities. Several existing natural contaminant
   degradation pathways can be recruited together and assembled in the
   chassis cells by molecular biological techniques to produce functional
   synthetic pathways, thus expanding the substrate range of the chassis
   cells and providing a favorable solution for combined pollution
   remediation ([37]4). To date, there are only limited reports on the
   construction of chassis cells for the remediation of combined pollution
   using synthetic biology ([38]2). Pseudomonas strains are model
   organisms that are endowed with metabolic, physiological, and stress
   endurance traits that could help them fulfill the demands of combined
   pollution remediation ([39]5), and Pseudomonas putida KT2440 is the
   most commonly used chassis strain in such studies ([40]6[41]–[42]8).
   However, as a chassis strain for combined pollution degradation, P.
   putida KT2440 has certain limitations, as it can degrade only a limited
   number of simple aromatic compounds ([43]9) and cannot degrade
   refractory contaminants such as polycyclic aromatic hydrocarbons (PAHs)
   and dioxin analogs, which are toxic and pose serious health risks
   ([44]10, [45]11). Therefore, it is necessary to develop new chassis
   cells with enhanced contaminant degradation capabilities.

   P. putida B6-2 was a “superstar” for the mineralization of PAHs and
   dioxin-like compounds ([46]12). Compared with P. putida KT2440 (Table
   S1), P. putida B6-2 possesses several advantageous traits that make it
   more suitable for bioremediation, such as a wide substrate spectrum,
   diverse contaminant degradation pathways, and strong solvent tolerance
   ([47]12[48]–[49]14). Therefore, we expect P. putida B6-2 to be a
   promising chassis strain with significant potential for combined
   pollution remediation. However, during the process of using P. putida
   B6-2 as a chassis strain for heterologous expression of functional
   contaminant degradation pathways to enable the remediation of more
   refractory contaminants, it was observed that the transformation
   efficiency of P. putida B6-2 was remarkably low. This limitation
   prevents P. putida B6-2 from expressing functional contaminant
   degradation pathways and degrading a broader range of contaminants,
   thereby hindering its broader application for combined pollution
   remediation.

   Targeted genome streamlining has been proven to be an effective
   approach for enhancing the gene editing efficiency of various organisms
   such as Escherichia coli ([50]15), P. putida ([51]16), and Schlegelella
   brevitalea ([52]17). Moreover, targeted genome streamlining has also
   led to several enhanced characteristics in genome-streamlined chassis
   cells, such as increased growth rates and biomass yield ([53]15,
   [54]18, [55]19), enhanced fitness under stressful conditions ([56]20,
   [57]21), and boosted heterologous gene expression ([58]16, [59]18,
   [60]21, [61]22). Therefore, we aimed to optimize P. putida B6-2 by
   targeted genome streamlining to address its limitations and further
   enhance its capabilities, thereby obtaining a more resilient and
   efficient chassis cell for environmental remediation.

   During the process of targeted genome streamlining, various
   bioinformatic methods are employed to analyze genomes for redundant
   components, such as prophages ([62]18, [63]20), genomic islands (GIs)
   ([64]16, [65]17), and other mobile genetic elements ([66]15, [67]23).
   These redundant components are then sequentially eliminated. However,
   the outcomes of genome streamlining have been unpredictable. Even if
   the elimination of a single fragment could lead to an improved
   phenotype, the successive elimination of multiple fragments might also
   yield an unpredictable phenotype, such as decreased growth ([68]19,
   [69]24) and reduced production ([70]16, [71]25). Moreover, the specific
   mechanisms underlying phenotypic changes after genome streamlining have
   not been fully characterized. Genome streamlining inherently involves
   uncertainty, underscoring the need for a tool capable of accurately
   predicting the effects of genome streamlining on strains.

   Genome-scale metabolic models (GEMs) are chemical stoichiometric
   matrices generated by linking mass-balanced metabolic reactions with
   gene-protein-reaction associations and provide a comprehensive
   numerical summary of whole-cell biochemistry at the system level
   ([72]26, [73]27). GEMs have been successfully applied to increase
   biosynthesis yields ([74]28, [75]29) and bioremediation efficiencies
   ([76]26, [77]30). However, to our knowledge, there are no reports to
   date on the use of GEMs for the rational targeted genome-streamlining
   strategies ([78]31). Flux balance analysis (FBA) is a widely used
   biochemical network analysis method ([79]32). By selecting a specific
   objective function (such as biomass) in GEMs for FBA, it is feasible to
   rapidly predict the influence of genome streamlining on the target
   traits of genome-streamlined strains. This approach allows the
   simulation and preliminary forecasting of the impacts of different
   genome-streamlining strategies on target traits within GEMs to prevent
   “blind” genome streamlining.

   Therefore, in this study, we systematically eliminated redundant
   elements within the P. putida B6-2 genome based on bioinformatic
   analyses and GEM predictions, aiming to construct an optimal chassis
   cell for environmental remediation ([80]Fig. 1). Furthermore, the
   electroporation and conjugation efficiencies, growth characteristics,
   contaminant degradation capabilities, and stress resistance of P.
   putida B6-2 and its genome-streamlined strains were evaluated to
   systematically characterize the effects of genome streamlining.
   Finally, a transcriptomic analysis of P. putida B6-2 and P. putida BGR4
   was conducted to provide valuable insights into the mechanisms
   underlying the improved phenotypes of the genome-streamlined strains.

Fig 1.

   [81]Visual workflow illustrates steps for genome streamlining in P.
   putida B6-2, from initial analysis and prediction of genome features to
   automated model construction, manual curation, model evaluation, and
   flux simulation.
   [82]Open in a new tab

   Schematic of the genome-streamlining workflow based on bioinformatic
   analyses and GEM predictions. (a) The genome of P. putida B6-2. (b)
   Various bioinformatic analysis and prediction methods were used to
   identify the fragments to be deleted. (c) Several genome-streamlined
   strains were constructed based on different genome-streamlining
   strategies. (d) Draft GEMs were then generated by using an automated
   modeling tool called gapseq ([83]33). (e) Several manual curation steps
   (including gap filling, model revision, and parameter calibration) were
   conducted to improve the quality of the GEMs, and more details about
   manual curation steps could be found in Materials and Methods. (f)
   Memote ([84]34) was used to score the GEMs, and the qualitative (such
   as predicting the ability of strains to utilize carbon sources) and
   quantitative (such as predicting the maximum growth rates [MGRs] of
   strains) predictive abilities of the GEMs were tested. (g) The GEMs
   were subjected to FBA in MATLAB using the COBRA Toolbox v.3.0 ([85]35).
   (h) The genome-streamlined strains were constructed in vivo with the
   optimal genome-streamlining strategy. (i) The target traits were
   evaluated, and changes in transcription levels between P. putida B6-2
   and P. putida BGR4 were tracked via transcriptomic analysis.

RESULTS

Determination of deletion targets

   The presence of prophages and GIs in the genome of a chassis cell is
   redundant, as their replication and expression consume the energy of
   the chassis cell, while the presence of redundant nucleases also
   hampers the gene editing efficiency of the chassis cell ([86]16,
   [87]18, [88]21). Therefore, prophages, GIs, and redundant
   nuclease-encoding genes were targeted for deletion. Specifically,
   PHASTER ([89]36) and Prophage Hunter ([90]37) were used to predict
   prophages, while IslandViewer 4 ([91]38) was utilized to analyze GIs.
   Additionally, the genome sequences were subjected to annotation with
   the RAST server ([92]39) and the Pseudomonas Genome Database ([93]40)
   to identify redundant nuclease-encoding genes. The identified deletion
   regions are detailed in Table S2. Overall, four prophages, two GIs and
   three nuclease-encoding genes were deleted, for a total size of 259.7
   kb, accounting for 4.1% of the P. putida B6-2 genome.

Construction and evaluation of the GEMs

   Once the deletion targets were designed, they were divided into four
   portions, with each portion comprising approximately 1% of the total
   genome content. Subsequently, four genome-streamlined strains,
   designated P. putida BGR1, P. putida BGR2, P. putida BGR3, and P.
   putida BGR4, were designed based on the schematic diagram shown in
   [94]Fig. 2a. The GEMs for these strains were then automatically
   constructed using the gapseq tool; detailed information about the GEMs
   is shown in Table S3.

Fig 2.

   [95]Genome-streamlining steps in P. putida B6-2, carbon source
   utilization analysis, specific growth rate comparison, FBA results for
   metabolic pathways, and increased growth rates across different
   substrates.
   [96]Open in a new tab

   Construction of genome-streamlined strains and GEM-predicted results.
   (a) A pipeline for the construction of genome-streamlined strains. (b)
   Results of the wild_type_B6-2 model’s prediction of the carbon source
   utilization capability of P. putida B6-2 and its ability to
   quantitatively predict the MGR of P. putida B6-2 on glucose. TP, ture
   positive; FP, false positive; FN, false negative; TN, ture nagetive.
   (c) KEGG enrichment analysis results of differential flux reactions
   between model wild_type_B6-2 and model BGR4 after FBA on LB. (d)
   Differential flux reactions enriched in the “carbon metabolism” pathway
   in FBA results and strictly constrained reactions enriched in the
   “carbon metabolism” pathway in flux variability analysis (FVA) results
   of model BGR4 and model wild_type_B6-2 on LB. The green arrows
   represent reactions that were strictly constrained in FVA results. The
   gray arrows indicate that there was no significant difference in FBA
   results. G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; FBP,
   fructose-1,6-bisphosphate; DHAP, dihydroxyacetone phosphate; 6PG,
   6-phosphogluconate; KDPG, 2-keto-3-deoxy-6-phosphogluconate; Ru5P,
   ribulose-5-phosphate; R5P, ribose-5-phosphate; X5P,
   xylulose-5-phosphate; S7P, sedoheptulose-7-phosphate; E4P,
   erythrose-4-phosphate; G3P, glyceraldehyde-3-phosphate; 3PG,
   3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate;
   Pyr, pyruvate; AcCoA, acetyl-coenzyme A; OAA, oxaloacetate; CIT,
   citrate; ICT, isocitrate; KG: 2-oxo-glutarate; SUC, succinate; FUM,
   fumarate; MAL, malate; Prpp, 5-phospho-alpha-D-ribose 1-diphosphate.
   (e) The enhanced MGRs of P. putida BGR4 in the presence of LB, 28
   positive carbon sources, and phenol (shown in pink) compared with those
   of P. putida B6-2 predicted by GEMs.

   The quality of the GEMs constructed by gapseq was assessed using the
   model wild_type_B6-2 as an example; this model received a score of 85
   in the Memote report (File S4) and achieved an accuracy ratio of 72% in
   predicting the utilization of 78 carbon sources by P. putida B6-2 in
   the carbon source utilization capacity experiment ([97]Fig. 2b; Tables
   S4 and S5). Subsequently, the model wild_type_B6-2 was used to predict
   the maximum growth rate (MGR) of P. putida B6-2 on glucose to assess
   its quantitative predictive capability, and the difference between the
   in silico prediction and the in vitro measurements was lower than 20%,
   indicating that the model wild_type_B6-2 has reliable quantitative
   predictive capability ([98]Fig. 2b). The above results indicated that
   all GEM quality validation experiments consistently demonstrated that
   the model wild_type_B6-2 constructed by gapseq was reliable.

The GEM-predicted results showed that, compared with P. putida B6-2, P.
putida BGR4 exhibited higher MGRs on all the tested carbon sources

   Once GEMs of reliable quality were obtained, FBA was performed in
   MATLAB to predict the effect of genome streamlining on target traits.
   For the chassis cells, higher growth rates indicated that the chassis
   cells could utilize carbon sources more efficiently for growth,
   allowing faster acquisition and utilization of resources and gaining
   competitive advantages. Therefore, growth rate was chosen as the target
   trait in this study. Then, biomass was set as the objective function to
   solve for the MGRs on LB (default gap-filling medium) and 28 positive
   carbon sources (from the carbon source utilization capacity experiment)
   in the GEMs. The GEM-predicted results showed that genome streamlining
   did not reduce but rather slightly increased the MGRs of the
   genome-streamlined strains (Table S6). Moreover, compared with P.
   putida B6-2 and the other genome-streamlined strains, P. putida BGR4
   had the highest MGRs in the presence of LB and 28 positive carbon
   sources. The model wild_type_B6-2 and model BGR4 were then subjected to
   FBA on LB to figure out the reason why the MGR of P. putida BGR4 was
   higher. KEGG enrichment results showed that differential flux reactions
   were mainly enriched in pathways such as “biosynthesis of secondary
   metabolites,” “carbon metabolism,” and “fatty acid metabolism”
   ([99]Fig. 2c). Further analysis of the “carbon metabolism” pathway
   suggested that the higher MGR of model BGR4 on LB might be attributed
   to enhanced pathways such as the “pentose phosphate (PP) pathway,”
   “Embden-Meyerhof-Parnas (EMP) pathway,” and “butanoate metabolism”
   pathway ([100]Fig. 2d). Furthermore, flux variability analysis (FVA)
   was used to help identify which reactions were strictly constrained
   (with nearly fixed fluxes) under conditions of maximal growth rate. The
   results indicated that nine reactions in the “carbon metabolism”
   pathway were strictly constrained, which were also enriched in the “PP
   pathway,” the “EMP pathway,” and “butanoate metabolism” pathway
   ([101]Fig. 2d).

   The ability to heterologously express contaminant degradation gene
   clusters is crucial for chassis cells for combined pollution
   remediation purposes. Therefore, by incorporating the phenol transport
   reaction and the phenol hydroxylation reaction into the GEMs, the MGRs
   of P. putida B6-2 and its genome-streamlined strains on phenol were
   also predicted. The prediction results showed that P. putida BGR4 still
   had the highest MGR on phenol (Table S6). In summary, compared with P.
   putida B6-2, P. putida BGR4 had enhanced MGRs in the presence of LB, 28
   positive carbon sources, and phenol, with improvements ranging from
   1.15% to 4.46% ([102]Fig. 2e). Consequently, the genome-streamlining
   strategy for P. putida BGR4 was adopted for the construction of the
   genome-streamlined strains in the laboratory.

Construction and characterization of the genome-streamlined strains

   The deletion targets were deleted using the traceless deletion method
   with the pK18mobsacB plasmid (Fig. S1). To verify the accurate deletion
   of the targeted regions, PCR and DNA sequencing of the amplified
   fragments were performed using the corresponding primers (Fig. S2). The
   growth characteristics of P. putida B6-2 and its genome-streamlined
   strains were subsequently evaluated. Growth was tested in both LB
   medium and mineral salt medium (MSM) supplemented with a carbon source
   that induced different metabolic pathways, i.e., glycerol for
   gluconeogenesis or glucose for glycolysis. Prior to this, the
   morphologies of P. putida B6-2 and P. putida BGR4 were examined by
   transmission electron microscopy (TEM), and no noticeable differences
   were observed (Fig. S3). The growth curves of P. putida B6-2 and its
   genome-streamlined strains in LB medium and MSM supplemented with 2 g/L
   glucose showed that the growth rates and growth tendencies of the five
   strains were similar, with no significant differences ([103]Fig. 3a and
   b). Drop assays were also performed to evaluate the ability of the
   strains to grow on solid LB plates and solid mineral salt plates
   supplemented with 2 g/L glucose (Fig. S4a), and the results were
   consistent with the growth curve results mentioned above. However, the
   growth curves of P. putida B6-2 and P. putida BGR4 in glycerol were
   different, with P. putida BGR4 showing a 12-h reduction in the lag
   period, a higher growth rate, and a 0.7-fold increase in the maximum
   optical density at 600 nm (OD[600]) ([104]Fig. 3c). Additionally, the
   ability of P. putida B6-2 and P. putida BGR4 to degrade phenol after
   heterologous expression of the phenol monooxygenase encoded by
   dmpKLMNOP from Cupriavidus pinatubonensis JMP134 ([105]41) was
   evaluated. P. putida BGR4 pBBR-dmpKLMNOP exhibited a 24-h reduction in
   lag period, a higher growth rate, and a 1.1-fold increase in maximum
   phenol degradation rate compared with P. putida B6-2 pBBR-dmpKLMNOP
   ([106]Fig. 3d). In conclusion, the findings above suggested that genome
   streamlining did not hinder the metabolic capacity of the
   genome-streamlined strains but actually enhanced the glycerol
   utilization ability of P. putida BGR4 and the phenol degradation
   capability of P. putida BGR4 pBBR-dmpKLMNOP.

Fig 3.

   [107]Comparative analysis depicts growth curves, phenol degradation and
   transformation efficiencies for P. putida B6-2 and its
   genome-streamlined strains. Each plot highlights performance metrics
   over time.
   [108]Open in a new tab

   Gross physiological properties of P. putida B6-2 and its
   genome-streamlined strain(s). The growth curves of P. putida B6-2 and
   its genome-streamlined strain(s) in LB medium (a) and MSM supplemented
   with 2 g/L glucose (b) or 10 g/L glycerol (c). (d) The phenol
   degradation (shown as a dashed line) and growth (shown as a solid line)
   curves of P. putida B6-2, P. putida B6-2 pBBR-dmpKLMNOP, and P. putida
   BGR4 pBBR-dmpKLMNOP. (e) The electroporation efficiencies for the
   pBBR1MCS2 plasmid in P. putida B6-2 and its genome-streamlined strains.
   The vertical coordinate is the log value of the electroporation
   efficiency. CFU, colony-forming unit. (f) The conjugation efficiencies
   for the plasmid pK18mobsacB-Δvdh into the vdh gene locus of P. putida
   B6-2 and its genome-streamlined strains. The data are presented as the
   means ± SDs from three independent experiments. One-way analysis of
   variance (ANOVA) was performed for statistical analysis. ns no
   significant difference, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P
   < 0.0001.

   The ability to accept and uptake exogenous DNAs is important for an
   ideal chassis cell. Therefore, the ability of P. putida B6-2 and its
   genome-streamlined strains to uptake foreign DNA by electroporation was
   tested using the broad host range vector pBBRMCS2. Compared with P.
   putida B6-2, the genome-streamlined strains exhibited varying degrees
   of improvement in electroporation efficiency, with P. putida BGR4
   demonstrating the greatest increase, surpassing P. putida B6-2 by five
   orders of magnitude ([109]Fig. 3e). The enhanced electroporation
   abilities of the genome-streamlined strains were mainly attributed to
   the knockout of two deoxyribonuclease genes, endA-1 and endA-2, while
   the knockout of the hsdR gene did not significantly affect
   electroporation efficiency (Fig. S5). In addition, bacteria can also
   obtain foreign DNA through conjugation. Gong et al. reported that the
   vdh gene locus was a favorable chromosomal integration site ([110]6).
   Hence, the suicide plasmid pK18mobsacB was used to construct a
   chromosomal integration vector targeting the vdh gene locus. Compared
   with P. putida B6-2, both P. putida BGR3 and P. putida BGR4 exhibited
   significantly greater conjugation efficiencies, with P. putida BGR4
   exhibiting the greatest increase (8.3-fold) ([111]Fig. 3f). The
   deletion of the endA-1 and endA-2 genes, as well as the
   endonuclease-encoding gene of prophage 3, might explain the enhanced
   conjugation ability of P. putida BGR3 and P. putida BGR4. Taken
   together, the above results demonstrated that genome streamlining
   greatly enhanced the electroporation and conjugation efficiencies of P.
   putida BGR4.

Transcriptomic analysis indicated more active metabolism within P. putida
BGR4

   As mentioned previously, the GEMs predicted altered carbon metabolic
   capabilities between P. putida B6-2 and P. putida BGR4. Therefore, a
   transcriptomic analysis was subsequently performed to investigate how
   genome streamlining affects the basal physiological metabolism of P.
   putida BGR4. As the volcano plots showed, 595 differentially expressed
   genes (DEGs) were upregulated and 1,003 DEGs were downregulated in P.
   putida BGR4 compared with P. putida B6-2 ([112]Fig. 4a). After KEGG
   pathway enrichment analysis, DEGs were found to be significantly
   enriched in the “biosynthesis of secondary metabolites,” “carbon
   metabolism,” “ribosome,” “oxidative phosphorylation,” and
   “tricarboxylic acid (TCA) cycle” pathways ([113]Fig. 4b). Among these,
   the enrichment of DEGs in the “carbon metabolism” pathway indicated
   altered carbon metabolic capabilities between P. putida B6-2 and P.
   putida BGR4, which was consistent with the GEM-predicted results.
   Moreover, the expression of the majority of the DEGs involved in
   central carbon metabolism pathways (Fig. S6) was also significantly
   upregulated, indicating a greater metabolic capacity of P. putida BGR4
   ([114]Fig. 4c).

Fig 4.

   [115]Comparison of gene expression between P. putida BGR4 and P. putida
   B6-2 in LB medium: Volcano plot, KEGG enrichment, and carbon metabolism
   pathway map showing up/downregulated DEGs with directional arrows.
   [116]Open in a new tab

   Transcriptomic analysis of DEGs between P. putida B6-2 and P. putida
   BGR4 in LB medium. Volcano map (a)and KEGG pathway enrichment analysis
   (b)of DEGs in P. putida BGR4 compared with P. putida B6-2. (c)DEGs
   enriched in the “carbon metabolism” pathway in P. putida BGR4 compared
   with P. putida B6-2. The pink arrows represent the exclusive
   upregulation of DEGs in the pathway. The dark-blue arrows represent the
   exclusive downregulation of DEGs in the pathway. The light-blue arrows
   indicate the presence of both upregulated and downregulated DEGs in the
   pathway. The gray arrows indicate that there was no significant
   difference. W, P. putida B6-2 cultured in LB medium; R4, P. putida BGR4
   cultured in LB medium.

   The “oxidative phosphorylation” pathway is the primary pathway for
   generating the energy currency ATP to sustain basic life activities.
   Most of the DEGs involved in the “oxidative phosphorylation” pathway,
   such as the genes encoding some subunits of the complex of NAD(P)H
   dehydrogenase, cytochrome c oxidase, and succinic dehydrogenase, as
   well as seven of the eight subunits of ATP synthetase, were
   significantly upregulated ([117]Fig. 5a). Moreover, P. putida BGR4 had
   higher ATP content, but there was no significant difference compared
   with that of P. putida B6-2 (Fig. S7). The TCA cycle is the main
   pathway for energy production, while ribosomes are responsible for
   protein synthesis. The overall DEGs in the “TCA cycle” and “ribosome”
   pathways were significantly upregulated, which also indicated more
   active metabolism within P. putida BGR4 ([118]Fig. 5b and c).

Fig 5.

   [119]Differential gene expression in oxidative phosphorylation
   pathways, with complexes I–V of the electron transport chain showing
   up- and downregulation. Ribosome-associated genes are visualized with
   changes in expression.
   [120]Open in a new tab

   Analysis of the DEGs of P. putida BGR4 in LB medium compared with P.
   putida B6-2. (a) Fold changes and heatmap of transcripts in the
   “oxidative phosphorylation” pathway. (b) Fold changes in transcript
   expression in the “ribosome” pathway. (c) Heatmap of transcripts in the
   “TCA cycle” pathway. The components and structures of the “oxidative
   phosphorylation,” “TCA cycle,” and “ribosome” pathways are depicted
   based on the KEGG database. The gray blocks represent areas with no
   significant difference. W, P. putida B6-2 cultured in LB medium; R4, P.
   putida BGR4 cultured in LB medium.

The genome-streamlined strains had a stronger ability to cope with stress

   The ability of chassis cells to withstand a stressful environment is
   the fundamental prerequisite for successful expression and functioning
   of contaminant degradation genes. Therefore, the tolerance of P. putida
   B6-2 and its genome-streamlined strains to acidic pH, alkaline pH,
   heat, osmotic pressure, oxidative stress, and DNA damage was measured
   using drop assays. The results showed that there was no significant
   difference in the tolerance to acidic pH, alkaline pH, heat, or osmotic
   pressure among P. putida B6-2 and its genome-streamlined strains (Fig.
   S4b through d). However, when exposed to oxidative stress and DNA
   damage induced by chemical agents (H[2]O[2], nalidixic acid [Nalid]
   [[121]42] or 4-nitroquinoline-1-oxide [4-NQO] [[122]43, [123]44]), the
   genome-streamlined strains exhibited a greater survival rate than that
   of P. putida B6-2 ([124]Fig. 6a), indicating that genome streamlining
   enhanced the tolerance of the genome-streamlined strains to oxidative
   stress and DNA damage.

Fig 6.

   [125]Dilution plating depicts colony growth under control and stress
   conditions for B6-2 and BGR strains in the presence of LB, H₂O₂,
   Nalidixic acid, and 4-NQO. UV exposure survival is depicted for B6-2
   and BGR strains at different energy levels.
   [126]Open in a new tab

   Survival of P. putida B6-2 and its genome-streamlined strain(s) under
   different stresses. (a) Drop assays to compare the fitness of P. putida
   B6-2 and its genome-streamlined strains under treatment with different
   chemical agents. (b) Plate tests to compare the tolerance of P. putida
   B6-2 and its genome-streamlined strains to UV irradiation. (c)
   Long-term survival tests comparing the ability of P. putida B6-2 and P.
   putida BGR4 to cope with starvation. The data are presented as the
   means ± SDs from six independent experiments. Two-tailed Student’s t
   tests were performed for statistical analysis. ns, no significant
   difference, *P < 0.01 and **P < 0.001.

   In addition, UV irradiation and starvation are common adverse
   conditions encountered by chassis cells in the environment. Therefore,
   P. putida B6-2 and its genome-streamlined strains were streaked on the
   surface of LB plates and exposed to different intensities of UV
   radiation (254 nm). The results revealed that the genome-streamlined
   strains P. putida BGR2, P. putida BGR3, and P. putida BGR4 showed
   greater tolerance to UV irradiation ([127]Fig. 6b). The starvation
   experiments were carried out by long-term cultivation without changing
   the medium. Compared with that of P. putida B6-2, the survival rate of
   P. putida BGR4 decreased significantly slower ([128]Fig. 6c), showing
   that genome streamlining improved the starvation tolerance of P. putida
   BGR4.

The presence of prophages reduced the resistance of P. putida B6-2 to 4-NQO

   As mentioned above, in the drop assays, the genome-streamlined strain
   P. putida BGR4 showed greater resistance to the typical reactive oxygen
   species (ROS) inducer and the potent mutagen 4-NQO, which was added to
   solid plates at a concentration of 0.6 mM. To explore how genome
   streamlining affected the tolerance of P. putida BGR4 to the ROS and
   DNA damage caused by 4-NQO, a transcriptomic analysis was performed.
   Prior to this, P. putida B6-2 and P. putida BGR4 were cultured in LB
   medium supplemented with different concentrations of 4-NQO to determine
   the appropriate 4-NQO treatment concentration and duration. By
   combining the results of the drop assays ([129]Fig. 7a) and the growth
   curves ([130]Fig. 7b), the 4-NQO treatment concentration and duration
   were selected as 0.6 mM and 4 h, respectively. Then, a transcriptomic
   analysis was performed.

Fig 7.

   [131]Dilution plating compares colony growth between B6-2 and BGR4
   after 6 h. Growth curve depicts effect of 4-NQO on B6-2 and BGR4
   strains, with statistical significance indicated. Heatmap depicts
   clustering and treatment effects on genomic regions.
   [132]Open in a new tab

   Cell survival and transcriptional analysis in the presence or absence
   of 4-NQO. (a) Drop assays of P. putida B6-2 and P. putida BGR4 grown in
   LB medium supplemented with 0.3 mM, 0.6 mM, or 1.2 mM 4-NQO for 6 h.
   (b) Growth curves of P. putida B6-2 and P. putida BGR4 in LB medium
   without or supplemented with 0.6 mM 4-NQO. (c) KEGG pathway enrichment
   analysis of DEGs in the treatment group. (d) Transcriptional analysis
   of the deleted regions between the control group and treatment group.
   W, P. putida B6-2 cultured in LB medium; R4, P. putida BGR4 cultured in
   LB medium; WQ, P. putida B6-2 cultured in LB medium supplemented with
   0.6 mM 4-NQO; R4Q, P. putida BGR4 cultured in LB medium supplemented
   with 0.6 mM 4-NQO. The data are presented as the means ± SDs from three
   independent experiments. Two-tailed Student’s t tests were was
   performed for statistical analysis. ns, no significant difference, *P <
   0.05, **P < 0.01, and ***P < 0.001.

   First, the overall difference in gene expression between the control
   group (P. putida B6-2 and P. putida BGR4 cultured in LB medium) and the
   treatment group (P. putida B6-2 and P. putida BGR4 cultured in LB
   medium supplemented with 0.6 mM 4-NQO) was investigated. The results of
   cluster analysis showed that the transcripts in the same group were
   clearly correlated (Fig. S8), indicating that the experimental
   treatment and group setting were reasonable.

   As shown in the volcano plots, 1,059 DEGs were upregulated and 1,377
   DEGs were downregulated under 4-NQO treatment in P. putida BGR4
   compared with P. putida B6-2 (Fig. S9). However, KEGG pathway
   enrichment analysis revealed that DEGs were not significantly enriched
   in the oxidative damage- or the DNA damage repair-related pathways
   ([133]Fig. 7c). Moreover, the gene expression of prophage 2, prophage
   3, and prophage 4 in P. putida B6-2 greatly increased upon treatment
   with 4-NQO ([134]Fig. 7d).

   Prophages can interrupt the lysogenic state and enter the lytic cycle
   in response to chemical inducers, forming free virions that can lyse
   bacteria and cause bacterial death ([135]45). Therefore, we
   hypothesized that 4-NQO might induce the prophages of P. putida B6-2
   into the lytic cycle and cause bacterial cell death ([136]Fig. 8a). A
   series of experiments were subsequently performed. First, virions were
   observed in the 4-NQO-treated bacterial culture of P. putida B6-2,
   demonstrating that 4-NQO could indeed induce a prophage-to-virion
   transition in P. putida B6-2 ([137]Fig. 8b). Then, a phage infection
   test was carried out with mitomycin C ([138]46) as the positive
   control. The results showed that the supernatant from P. putida B6-2
   treated with 4-NQO did not form phage plaques on P. putida B6-2 lawns
   (data not shown), whereas the supernatant from mitomycin C treatment
   did (Fig. S10), indicating that 4-NQO-induced virions could not infect
   P. putida B6-2, unlike mitomycin C-induced virions. Flow cytometry was
   used to quantify the production of virions, and the supernatant
   produced by centrifugation after mitomycin C treatment of P. putida
   B6-2 was used as the positive control. The flow cytometry results
   showed that P. putida B6-2 produced more virions after 4-NQO treatment
   ([139]Fig. 8c), which was consistent with the transcriptomic analysis
   in which the gene expression of the prophages of P. putida B6-2 greatly
   increased after 4-NQO treatment ([140]Fig. 7d). In summary, the above
   experimental results indicated that the prophages within P. putida B6-2
   could be induced by 4-NQO and transformed into the lytic cycle, leading
   to the formation of additional virions and causing bacterial death,
   thereby reducing the tolerance of P. putida B6-2 to 4-NQO.

Fig 8.

   [141]Illustration depicts effect of 4-NQO on prophage activation,
   leading to virion assembly and release. Microscopy image highlights
   virions with arrows. Bar graph compares virus content percentages under
   different conditions.
   [142]Open in a new tab

   The induction of prophages and detection of virions. (a) The proposed
   schematic diagram of a prophage within P. putida B6-2 being induced by
   4-NQO into the lytic cycle. (b) Morphology of the virion observed by
   TEM in the 4-NQO-treated bacterial culture of P. putida B6-2. (c) Flow
   cytometry analysis of the virion contents in the control group and
   treatment group. The data are presented as the means ± SDs from three
   independent experiments. W, P. putida B6-2 cultured in LB medium; R4,
   P. putida BGR4 cultured in LB medium; WQ, P. putida B6-2 cultured in LB
   medium supplemented with 0.6 mM 4-NQO; R4Q, P. putida BGR4 cultured in
   LB medium supplemented with 0.6 mM 4-NQO. One-way ANOVA was performed
   for statistical analysis. ns, no significant difference, *P < 0.05, **P
   < 0.01, and ***P < 0.001.

DISCUSSION

   In recent years, genome streamlining has become an effective strategy
   for constructing ideal chassis cells. Various genome-streamlined
   chassis cells, such as those of E. coli ([143]23), Lactococcus lactis
   ([144]20), and Streptomyces avermitilis ([145]47), have been
   successfully developed and utilized as platforms for product synthesis
   ([146]48). However, little attention has been given to the construction
   of chassis cells that degrade contaminants, and Liang’s report
   ([147]16) is the only study describing the construction of a chassis
   cell suitable for bioremediation by genome streamlining of P. putida
   KT2440. In this study, P. putida B6-2 was optimized by genome
   streamlining to further enhance its properties. Interestingly, both
   Liang’s report and our research indicated that genome streamlining
   improved the transformation efficiency, carbon source utilization
   capacity, and pollutant degradation capacity of the genome-streamlined
   strains, demonstrating the potential application of genome streamlining
   in constructing chassis strains suitable for bioremediation.

   Although there have been numerous studies about target genome
   streamlining, these studies have been conducted in a somewhat blinded
   manner. Researchers could evaluate target traits only after laborious
   construction of genome-streamlined strains in the laboratory. However,
   genome streamlining does not always lead to improved phenotypes.
   Therefore, there remains a demand for a tool that can evaluate the
   effects of genome streamlining. In this study, GEMs were used to
   predict the effects of genome streamlining on the MGR, and the results
   of the GEM predictions showed that the genome-streamlined strain P.
   putida BGR4 had enhanced and the highest MGRs across all tested carbon
   resources ([148]Fig. 2e). Therefore, strain BGR4 was chosen for in vivo
   construction. Subsequent growth experiments confirmed that P. putida
   BGR4 exhibited improved utilization of glycerol, as well as enhanced
   utilization of phenol after heterologous expression of the dmpKLMNOP
   gene cluster ([149]Fig. 3c and d). Moreover, analysis of the FBA
   results of model wild_type_B6-2 and model BGR4 provided insights into
   why the genome-streamlined strain BGR4 has enhanced metabolic
   capabilities ([150]Fig. 2c and d). In fact, GEMs have the potential to
   guide the construction of chassis cells not only for contaminant
   degradation purposes but also for other applications. For example, GEMs
   can be employed to guide the genome streamlining of chassis cells for
   biosynthetic purposes, with biomass serving as the objective function
   to solve for the maximum production rates. However, importantly, GEMs
   themselves have limitations. For example, the influence of global
   carbon regulation on carbon source utilization is not considered by
   GEMs ([151]49). Within P. putida B6-2, prophage 1, prophage 3, and GI-1
   contained several transcriptional regulators, prophage 3 specifically
   possessed a carbon storage regulator (Table S2). Both types of
   regulators can influence the carbon metabolic capacity of strains
   ([152]50). The deletion of prophages and GIs in P. putida BGR4 might
   change its global carbon regulation system compared with that in P.
   putida B6-2. However, the GEMs did not account for the effect of those
   regulatory factors on carbon metabolism capability, which might lead to
   inconsistencies between the GEM-predicted results and the experimental
   results. For example, the GEMs predicted that P. putida BGR4 would also
   have enhanced glucose utilization ability, but the experimental results
   showed no significant difference in glucose utilization capacity
   between P. putida B6-2 and P. putida BGR4. In future studies,
   transcriptomic, proteomic, and metabolomic data can be incorporated
   into the GEMs to update and calibrate it, allowing the GEMs to more
   accurately reflect the actual metabolic state of the cell after genome
   streamlining ([153]49). This, in turn, helps identify redundant genes
   and pathways and predicts the impact of further genome streamlining.

   Many studies have reported that genome streamlining leads to many
   improved phenotypes, but few studies have reported the mechanisms
   underlying these improved phenotypes. In this study, we attempted to
   explain the reasons behind the superior phenotypes of the
   genome-streamlined strains by employing transcriptomic analysis. After
   genome streamlining, P. putida BGR4 exhibited enhanced utilization of
   not only glycerol but also phenol after heterologous expression of
   phenol monooxygenase. P. putida strains can convert glycerol to
   dihydroxyacetone phosphate ([154]51), which then enters the glycolytic
   pathway and the TCA cycle (Fig. S11). C. pinatubonensis JMP134 can
   utilize phenol as the sole carbon source and convert phenol to catechol
   through the monooxygenase enzyme complex DmpKLMNOP ([155]41). Catechol
   is then further catalyzed by a ring-cleaving enzyme to produce
   succinyl-CoA and acetyl-CoA, which ultimately enter the TCA cycle (Fig.
   S11). Transcriptomic analysis of P. putida B6-2 and P. putida BGR4
   revealed that most of the DEGs involved in the downstream pathways of
   the EMP pathway and the TCA cycle were upregulated, while only one DEG
   involved in a reversible reaction was downregulated ([156]Fig. 4c),
   which suggested that the upregulation of DEGs involved in the
   downstream pathway of EMP and the TCA cycle led to the enhanced
   utilization of glycerol by P. putida BGR4 and phenol by P. putida BGR4
   pBBR-dmpKLMNOP. In fact, not only phenol but most aromatic compounds
   (such as benzene, naphthalene, and their analogs) are metabolized by
   bacteria via the catechol pathway and eventually enter the TCA cycle
   ([157]4). Therefore, we considered P. putida BGR4 to be suitable for
   remediation of aromatic compound contamination.

   Additionally, the genome-streamlined strains were more resistant to
   stress than P. putida B6-2. After the elimination of prophage 1, the
   genome-streamlined strains exhibited increased tolerance to H[2]O[2]
   and Nalid ([158]Fig. 6a). Under the stress of 4-NQO and UV irradiation,
   as the number of eliminated prophages increased, the resistance of the
   genome-streamlined strains to UV irradiation and 4-NQO gradually
   strengthened. After the elimination of the four prophages, P. putida
   BGR3 and P. putida BGR4 exhibited similar highest levels of 4-NQO and
   UV irradiation tolerance. Moreover, P. putida BGR3 and P. putida BGR4
   consistently displayed similar highest stress tolerance abilities in
   all the stress tolerance tests. The above results suggested that the
   elimination of prophages enhanced the stress tolerance of the
   genome-streamlined strains, while the elimination of GIs might not.
   Transcriptomic analysis of the control group and treatment group
   revealed that the gene expression of prophage 2, prophage 3, and
   prophage 4 of P. putida B6-2 was strongly upregulated after 4-NQO
   treatment ([159]Fig. 7d). Subsequent experiments confirmed that the
   prophages within P. putida B6-2 could indeed be induced by NQO,
   resulting in decreased tolerance of P. putida B6-2 to 4-NQO. In
   summary, the above results suggest it is the elimination of prophages
   that results mainly in improved stress resistance in P. putida BGR4,
   which is consistent with the findings of Esteban ([160]52) and Qiao
   ([161]20), who showed that the elimination of prophages increased the
   stress tolerance of genome-streamlined strains. Therefore, it is
   suggested that the stress tolerance of chassis cells may be improved by
   genome streamlining of prophages within chassis cells.

Conclusions

   Compared with the original strain P. putida B6-2, the
   genome-streamlined chassis strain P. putida BGR4 exhibited improvements
   in various physiological characteristics, such as electroporation and
   conjugation efficiencies, carbon source utilization, contaminant
   degradation capabilities, and stress tolerance. These desirable traits
   make P. putida BGR4 an optimum chassis strain for environmental
   remediation. Moreover, our study demonstrated the effectiveness of
   targeted genome streamlining based on bioinformatic analyses and GEM
   predictions in obtaining an optimized chassis cell for contaminant
   degradation. Additionally, the insights gained from the transcriptomic
   analysis, along with the established targeted genome streamlining
   procedure, can provide a valuable framework for guiding the genome
   streamlining of other strains for diverse applications.

MATERIALS AND METHODS

Bacterial strains and growth conditions

   The bacterial strains and plasmids used in this work are listed in
   [162]Table 1. The sequences of all the oligonucleotides used in this
   study are listed in Table S7. Bacteria were routinely grown in LB
   medium (10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl). LBS
   medium (10 g of tryptone, 5 g of yeast extract, 10 g of NaCl, and 150 g
   of sucrose per liter of distilled water, sterilized at 115°C for 15
   min) was used for double-crossover integration screening. The MSM
   consisted of (g/L) 2.00 KH[2]PO[4], 3.28 Na[2]HPO[4]·12H[2]O, 0.10
   MgSO[4], 1.00 (NH[4])[2]SO[4], and trace metals (0.050 CaCl[2]·2H[2]O,
   0.050 CuCl[2]·2H[2]O, 0.008 MnSO[4]·H[2]O, 0.004 FeSO[4]·7H[2]O, 0.100
   ZnSO[4], 0.100 MoNa[2]O[4]·2H[2]O, and 0.050 K[2]WO[4]·2H[2]O). The
   solid media were prepared from liquid media supplemented with 1.5%
   (wt/vol) agar. If necessary, 2 g/L glucose, 10 g/L glycerol, or 2 mM
   phenol was added to the MSM as the sole carbon source. Phenol (≥99.5%,
   purity) was purchased from Sinopharm Chemical Reagent Co. Ltd.
   (Shanghai, China). Kanamycin (50 µg/mL) was added as needed. P. putida
   B6-2 and its genome-streamlined strains were cultured at 30°C, while E.
   coli cells were grown at 37°C. 2,6-Diaminopimelic acid (2,6-DAP) (0.3
   mM) was required when cultivating E. coli WM3064. Other supplements
   were added to the media when needed as follows: 30 mM H[2]O[2], 0.1 mM
   Nalid, and 0.6 mM 4-NQO. Cell growth was monitored by measuring the
   OD[600].

TABLE 1.

   Detailed information on the bacterial plasmids and strains used in this
   study
   Strain or plasmid Description Source
   Strains
    P. putida
     B6-2 Wild-type strain Professor Hongzhi Tang’s laboratory ([163]12)
     BGR1 B6-2 mutant (ΔhsdR, Δprophage 1) This study
     BGR2 B6-2 mutant (ΔhsdR, Δprophage 1, Δprophage 2) This study
     BGR3 B6-2 mutant (ΔhsdR, Δprophage 1, Δprophage 2, ΔendA-1, ΔendA-2,
   Δprophage 3,
   Δprophage 4) This study
     BGR4 B6-2 mutant (ΔhsdR, Δprophage 1, Δprophage 2, ΔendA-1, ΔendA-2,
   Δprophage 3,
   Δprophage 4, ΔGI-1, ΔGI-2) This study
     B6-2ΔhsdR B6-2 mutant (ΔhsdR) This study
     BGR2ΔendA-1ΔendA-2 B6-2 mutant (ΔhsdR, Δprophage 1, Δprophage 2,
   ΔendA-1, ΔendA-2) This study
     B6-2 pBBR-dmpKLMNOP B6-2 containing the plasmid pBBR-dmpKLMNOP This
   study
     BGR4 pBBR-dmpKLMNOP BGR4 containing the plasmid pBBR-dmpKLMNOP This
   study
    E. coli
     DH5α supE44 lacU169 (80dlacZΔM15) hsdR17 TransGen Biotech
     WM3064 2,6-Diaminopimelic acid auxotroph Laboratory stock
   Plasmids
    pBBRMCS2 Empty vector for expression; Kan^r Laboratory stock
    pK18mobsacB Empty vector for allelic exchange; Kan^r Laboratory stock
    pBBRMCS2-dmpKLMNOP pBBRMCS2 derivative constructed for the expression
   of the phenol degradation gene cluster from C. pinatubonensis JMP134
   Ren et al. ([164]41)
    pK18mobsacB-ΔhsdR pK18mobsacB derivative containing PCR product
   covering up- and downstream
   regions of hsdR This study
    pK18mobsacB-Δprophage 1 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions of prophage 1 This study
    pK18mobsacB-Δprophage 2 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions of prophage 2 This study
    pK18mobsacB-ΔendA-1 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions of endA-1 This study
    pK18mobsacB-ΔendA-2 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions of endA-2 This study
    pK18mobsacB-Δprophage 3 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions of prophage 3 This study
    pK18mobsacB-Δprophage 4 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions of prophage 4 This study
    pK18mobsacB-ΔGI-1 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions
   of GI-1 This study
    pK18mobsacB-ΔGI-2 pK18mobsacB derivative containing PCR product
   covering up- and downstream regions
   of GI-2 This study
    pK18mobsacB-Δvdh pK18mobsacB derivative containing PCR product
   covering up- and downstream regions
   of vdh This study
   [165]Open in a new tab

GEM construction, evaluation, and simulation

   Gapseq ([166]https://github.com/jotech/gapseq) was used for
   automatically reconstructing the GEMs. Gapseq uses a curated reaction
   database and a novel gap-filling algorithm. The resulting GEMs could be
   directly employed for FBA-based metabolic flux simulations of microbial
   growth. The GEMs in this study were constructed based on a recent
   genomic sequence of P. putida B6-2 (Refseq ID: [167]CP015202.1)
   ([168]53). First, the FASTA sequence of P. putida B6-2 was submitted to
   gapseq for semiautomated construction of the draft model; next, gap
   filling was conducted in the default LB medium. Subsequently, a
   literature comparison and model revision were conducted to ensure that
   the reactions and metabolic pathways in the model were consistent with
   known biological knowledge. Additionally, the parameters in the model
   were further calibrated: the upper and lower bounds of the reactions
   were adjusted to better reflect the state of the actual biological
   system; system constraints were defined with the upper bound for each
   metabolic reaction set to 1,000 and the lower bounds set to −1,000 and
   0 for reversible and irreversible reactions, respectively, except for
   exchange reactions; the lower bounds of exchange reactions related to
   oxygen were set to −10, and the lower bounds of exchange reactions
   related to the medium compounds were set to −100, while the remaining
   bounds were set to 0. Models BGR1–4 were constructed by deleting or
   modifying genes, reactions, and metabolites of the model
   wild_type_B6-2. Moreover, the biomass objective function (BOF) for
   models BGR1-4 was derived by modifying the BOF of model wild_type_B6-2
   based on the percentage of genome streamlining content. Specifically,
   the components in the BOF that included nucleotides or
   deoxyribonucleotides were proportionally adjusted. For example, in the
   BOF for the model wild_type_B6-2, the coefficient for ATP was
   40.1654758653685, whereas in the BOF for BGR1, the coefficient for the
   ATP component was changed to 39.7638211067148 (40.1654758653685 *
   0.99). All adjusted components included ATP, GTP, CTP, UTP, dATP, dGTP,
   dCTP, and dTTP. Detailed information and comparisons of metabolites and
   reactions among the five models are shown in Files 2 and S3. The SBML
   files of 5 GEMs could be accessed at
   [169]https://github.com/fansiqing/models.git.

   To ensure the reliability of the GEM-predicted results, the quality of
   the GEMs was evaluated using the model wild_type_B6-2 as an example.
   Memote was used to score the model wild_type_B6-2. The carbon source
   utilization capacity experiment was conducted as follows: 71 carbon
   sources from the GEN III Micro-Plate test and 7 reported positive
   carbon sources ([170]12) were simulated in the model wild_type_B6-2.
   The predicted results were subsequently compared with the experimental
   results to evaluate the accuracy of the model wild_type_B6-2 in
   predicting the ability of P. putida B6-2 to utilize carbon sources. To
   ensure accurate simulation of the ability of P. putida B6-2 to utilize
   78 carbon sources, additional transport reactions were added as free
   diffusion reactions within the model wild_type_B6-2 in MATLAB. Taking
   the addition of the phenol transport reaction as an example, the
   commands used were as follows: model = addReaction(model,
   'EX_cpd00127_e0', 'cpd00127[c0] <=>'). The Memote reports for models
   BGR1-4 are presented in Files S5–S8. Models with scores above 70 in
   both assessments were considered relatively reliable. If the model
   scored below 70, further refinement of the model was recommended
   ([171]27). In the quantitative assessment of the MGR on glucose,
   regression analysis was applied to calculate the specific growth rate
   of P. putida B6-2 on glucose, and the glucose maximum uptake rate of
   the model wild_type_B6-2 was referenced from P. putida KT2440 and
   defined as 6.1 mmol/g/h ([172]54).

   FBA was subsequently performed in MATLAB R2022b v.9.13 using the COBRA
   toolbox v.3.0 and the Gurobi linear solver to solve for the MGR. The
   commands used were as follows: model =
   changeObjective(wild_type_B6-2.xml, 'bio1'); solution =
   optimizeCbModel(model, 'max'). FVA was performed in the same
   environment as FBA. During FVA, the objective function value was fixed
   at the optimal value obtained through FBA under the condition of
   maximum growth rate, in order to analyze the upper and lower bounds of
   fluxes for each reaction. A reaction was defined as strictly
   constrained reaction if the relative change (relative change = |maxFlux
   − minFlux| / max(|maxFlux|, |minFlux|)) was less than a predefined
   threshold of 0.01. Here, minFlux represents the minimum flux, and
   maxFlux represents the maximum flux.

Traceless deletion of redundant elements in P. putida B6-2

   The plasmid pK18mobsacB ([173]55) was used to construct the
   genome-streamlined strains (Fig. S1). First, homology arms of
   approximately 1,000 bp flanking the deleted regions were amplified and
   ligated with linearized pK18mobsacB vectors to construct knockout
   vectors. The vectors were subsequently transformed into E. coli WM3064.
   Conjugation of E. coli WM3064 with the P. putida strains resulted in
   single-crossover mutants. A single colony was picked into LB medium
   supplemented with 50 µg/mL kanamycin, followed by PCR to identify
   positive transformants. Next, the single-crossover mutant was cultured
   in LB medium without antibiotics at 30°C for 24 h, after which a small
   amount of the bacterial solution was coated on LBS solid medium. Then,
   a single colony was picked and cultured in LB medium with or without
   kanamycin. A single colony that could not grow in medium supplemented
   with kanamycin was chosen for analysis. PCR was used to confirm whether
   the single colony was a knockout mutant.

Assessment of physiological characteristics

   The electroporation efficiency was assessed by electroporation of the
   broad host range plasmid pBBRMCS2 ([174]56) into P. putida B6-2 and the
   other four genome-streamlined strains at 1.4 kV for 2.5 ms, followed by
   CFU counting after 2 d of incubation at 30°C. The number of
   transformants was calculated as CFU per µg DNA. To measure conjugation
   efficiency, P. putida B6-2, the genome-streamlined strains and E. coli
   WM3064 pK18mobsacB-Δvdh were grown in LB medium to OD[600] = 0.6. Then,
   the bacterial cells were harvested (4,500 × g for 10 min) and washed
   twice with PBS (137.0 mM NaCl, 2.7 mM KCl, 10.0 mM Na[2]HPO[4], and 2.0
   mM KH[2]PO[4]; pH = 7.4). The donor and acceptor strains were mixed (6
   mL:2 mL), and the mixture was resuspended in 50 µL of PBS. Then, 50 µL
   of the mixture was added to the center of an LB plate supplemented with
   2,6-DAP and incubated for 12 h at 30°C. The bacteria were washed with 1
   mL of PBS, and 10 µL of the bacterial mixture was collected and mixed
   with 90 µL of PBS. Then, 100 µL of the bacterial mixture was plated on
   an LB plate containing 50 µg/mL kanamycin and incubated for 1 d. Single
   colonies on the plate were counted to determine the conjugation
   efficiency. The Biolog GEN III Micro-Plate test was used to evaluate
   the metabolic phenotypes of P. putida B6-2. The culture steps were
   performed according to the manufacturer’s instructions. After the
   plates were incubated at 30°C for 48 h, the absorbance at 600 nm of the
   samples in the 96-well plate was measured to determine the metabolic
   phenotypes of P. putida B6-2. Well 1, where no carbon source was
   introduced, was designated the control. After subtraction of the
   control value, a positive OD[600] value was taken as an indication of
   the ability to utilize the substrate. Conversely, a negative or zero
   OD[600] value was considered to indicate a lack of ability to utilize
   the substrate.

RNA extraction and RNA-seq analysis

   Overnight cultures of P. putida B6-2 and P. putida BGR4 were inoculated
   into fresh LB medium with an initial OD[600] = 0.05. When the OD[600]
   of the cultures reached 0.6, the cultures in the treatment group were
   treated with 0.6 mM 4-NQO for 4 h, while the cultures in the control
   group were not treated. After completion of the treatment, samples for
   RNA extraction and RNA-seq were then immediately collected from the
   bacterial cultures. Three biological replicates of each sample were
   used to ensure data reliability. For RNA extraction, 1 mL of bacterial
   culture was collected at 13,000 × g for 2 min at 4°C, after which the
   medium was discarded, and the tubes were frozen in liquid nitrogen. The
   pellets were resuspended in 100 µL of lysozyme solution (20 mg/mL
   lysozyme in Tris-EDTA buffer, pH = 8.0). Total RNA was isolated using
   an Omega E.Z.N.A. Bacterial RNA Kit (Omega Bio-Tek, USA).

   RNA sequencing was performed by Beijing Novogene Bioinformatics
   Technology Co. Ltd. (China). A total of 12 samples (four treatments
   with three replicates) were sequenced. The DEG analysis was performed
   using the DESeq R package (1.18.0). The resulting P values were
   subsequently adjusted using Benjamin and Hochberg’s approach to control
   for the false discovery rate. Genes with | log[2](fold change) | > 0
   and padj (adjusted P) < 0.05 were defined as DEGs. KOBAS software was
   used to test the statistical enrichment of DEGs in KEGG pathways. The
   figures were mainly generated with GraphPad Prism 8.01 (GraphPad
   Software Inc., San Diego, CA) or R using the heatmap package, and the
   parameters were modified when necessary.

Determination of ATP content

   The Beyotime ATP Test Kit S0026 was used to extract and determine the
   ATP content. The sample preparation and sampling time were in
   accordance with the RNA extraction method. The dosage of the bacterial
   solution was 0.1 OD[600]. The detection steps were carried out
   following the manufacturer’s instructions.

Stress resistance

   Acid stress and alkaline stress were achieved by adjusting the pH of
   the LB medium with HCl or NaOH, while osmotic stress was achieved by
   adding different concentrations of NaCl to the LB medium. Oxidative
   stress and DNA damage were induced by the chemical agents H[2]O[2],
   Nalid, and 4-NQO. Nalid is a DNA-damaging antibiotic that can cause
   irreversible chromosome fragmentation, and 4-NQO is a potent mutagen
   that can induce transitions and transversions; both of these agents are
   ROS-producing agents and can cause DNA damage ([175]42[176]–[177]44).
   When needed, 0.1 mM Nalid or 0.6 mM 4-NQO was added to the LB medium to
   induce oxidative stress and DNA damage. After the preparation of the
   above liquid media, 1.5% (wt/vol) agar was added, and the mixture was
   sterilized to prepare agar plates. Thermal stress was induced by
   incubating the plate at 38°C. For the above stress, overnight cultures
   of P. putida B6-2 and its genome-streamlined strains were diluted in
   PBS to an OD[600] = 0.5, and 3 µL of each dilution was spotted onto
   agar plates. For oxidative stress and DNA damage induced by H[2]O[2],
   activated cultures (OD[600] = 0.05) of P. putida B6-2 and its
   genome-streamlined strains were treated with 30 mM H[2]O[2] for 1 h and
   then diluted in PBS, after which 3 µL of each dilution was spotted onto
   LB agar plates. The UV irradiation tolerance test was performed as
   follows: 50 µL of activated cultures of P. putida B6-2 and its
   genome-streamlined strains (OD[600] = 0.1) were spotted onto one side
   of the plate, after which the side containing 50 µL of culture was
   rotated 90° along the horizontal plane to form even lines. The plate
   was dried and exposed to UV light from a 254-nm lamp (CL-1000
   Ultraviolet Crosslinker, 100 µJ/cm) at various intensities ranging from
   20 J to 40 J, at a distance of 30 cm. Specifically, plates containing
   4% (wt/vol) or 5% (wt/vol) NaCl were incubated at 30°C for 42 h and
   then photographed. Unless otherwise stated, the plates were
   photographed after 18 h of incubation at 30°C. In the starvation test,
   activated cultures of P. putida B6-2 and P. putida BGR4 were inoculated
   into fresh LB medium with an initial OD[600] = 0.05 and subsequently
   cultured at 30°C for a total period of 15 d without the addition of
   fresh nutrients. CFU counts were determined by serial dilution of the
   cultures in PBS and plating 10 µL of each dilution on LB agar plates in
   six independent experiments.

Flow cytometry

   For the quantitative analysis of virions, sample preparation was
   performed with the RNA extraction method described above. After
   completion of the treatment, 1 mL of bacterial culture was centrifuged
   at 13,000 × g for 5 min, and the supernatant was collected for the
   quantitative analysis of virions. Moreover, overnight-grown cultures of
   P. putida B6-2 were treated with 2 µg/mL mitomycin C for 6 h, and the
   resulting supernatant was used as a positive control. Samples for flow
   cytometry analysis were prepared and detected according to the
   manufacturer’s instructions for the CytoFLEX flow cytometry instrument.
   Briefly, the supernatant was treated with pure and chilled ethanol to
   achieve a final ethanol concentration of 70% (vol/vol), and the mixture
   was then fixed overnight. SYBR dye (Invitrogen, 10,000×, DMSO dilution)
   was added to the mixture at a final concentration of 5×, after which
   the mixture was incubated for 10 min at 80°C in the dark. The
   excitation wavelength used to measure the fluorescence intensity of
   SYBR was 488 nm, and the emission wavelength was 509 nm. At least
   100,000 events were counted for analysis. The data were analyzed with
   FlowJo 10.8.1.

Phage infection test

   The detailed procedures for the phage infection test were previously
   described by Esteban et al. ([178]52). Briefly, overnight cultures of
   P. putida B6-2 were treated with 2 µg/mL mitomycin C for 6 h or with
   0.6 mM 4-NQO for 4 h starting from an initial OD[600] of 0.6. Then, 1
   mL of culture was treated with chloroform, vortexed, and centrifuged at
   13,000 × g for 5 min. Twenty-five microliters of the supernatant was
   spotted onto bacterial lawns prepared by mixing 100 µL of the overnight
   culture of P. putida B6-2 with 15 mL of LB medium containing 0.7%
   (wt/vol) agar and pouring the mixture onto an LB agar plate
   supplemented with 5 mM CaCl[2] and 5 mM MgCl[2]. The control sample was
   LB medium supplemented with the same concentrations of mitomycin C and
   chloroform. The plate was then photographed after 18 h of incubation at
   30°C.

Imaging of bacteria and virions

   A JEM-1400 Flash TEM (JEOL, Japan) was used to observe the morphology
   of P. putida B6-2, P. putida BGR4, and virions. Colonies of P. putida
   B6-2 and P. putida BGR4 on 0.5% (wt/vol) LB agar plates were dipped in
   20 µL of ddH[2]O and resuspended to produce samples. To observe the
   morphology of virions induced by 4-NQO, the sample preparation
   procedure was the same as that for RNA extraction. Finally, 4 µL of
   each sample was transferred to a TEM carrier and stained with 1%
   (wt/vol) uranium acetate for detection.

Phenol degradation assays

   Ten milliliters of activated cultures of P. putida B6-2, P. putida B6-2
   pBBR-dmpKLMNOP, and P. putida BGR4 pBBR-dmpKLMNOP in LB medium was
   harvested, washed twice with PBS, and then resuspended in MSM
   supplemented with 2 mM phenol. The strains were then passaged for an
   additional two generations in MSM supplemented with 2 mM phenol.
   Subsequently, the cells were harvested by centrifugation, washed twice
   with PBS, and resuspended in MSM supplemented with 2 mM phenol at an
   initial OD[600] = 0.05. Cultures were incubated at 30°C and 200 rpm on
   a shaker. One milliliter of each culture was harvested at regular time
   intervals for the detection of the OD[600] and phenol concentration.
   Analysis of the phenol concentration was performed by a GC-2014C gas
   chromatograph equipped with a flame ionization detector and an AOC-20i
   autoinjector (SHIMADZU, Shanghai, China). All the samples were
   centrifuged at 10,000 × g for 2 min. The supernatant was collected
   after pretreatment with a 0.45-µm filter. The detection program was as
   follows: the initial temperature was set at 60°C, the temperature was
   gradually increased to 160°C at a rate of 20°C/min, and the temperature
   was subsequently held at 160°C for 5 min.

ACKNOWLEDGMENTS