Improvement of L-Tryptophan Production in Escherichia coli Using Biosensor-Based, High-Throughput Screening and Metabolic Engineering

Abstract

The demand for L-tryptophan (L-Trp) has been rapidly increasing across various industries, including pharmaceuticals, food, and animal feed. However, traditional production methods have been unable to efficiently meet this growing demand. Hence, this study aimed to develop strategies for enhancing L-Trp production in Escherichia coli. Firstly, an L-Trp-producing strain was selected and subjected to atmospheric and room temperature plasma (ARTP) mutagenesis to generate a mutant library. This was followed by high-throughput screening using an L-Trp-specific riboswitch and a yellow fluorescent protein (YFP)-based biosensor in a flow cytometric cell sorting (FACS) system. Among the screened mutants, GT3938 exhibited a 1.94-fold increase in L-Trp production. Subsequently, rational metabolic engineering was applied to GT3938 by knocking out the L-Trp intracellular transporter gene (tnaB), enhancing the expression of the aromatic amino acid exporter (YddG) and optimizing precursor supply pathways. The resulting strain, zh08, achieved an L-Trp titer of 3.05 g/L in shake-flask fermentation, representing a 7.71-fold improvement over the original strain. This study demonstrated an effective strategy for industrial strain development by integrating biosensor-assisted, high-throughput screening with rational metabolic engineering.

1. Introduction

L-tryptophan (L-Trp), one of the eight essential amino acids for humans and animals, is widely used in food additives, animal feed, and pharmaceuticals [1,2]. According to QYResearch, the global L-Trp market was valued at $682 million in 2023 and is expected to reach $1915 million by 2030, with a compound annual growth rate (CAGR) of 13.7% from 2024 to 2030. However, traditional methods of L-Trp production, such as casein hydrolysis, enzymatic conversion, and chemical synthesis, can no longer meet market demands [3,4,5]. As a result, microbial fermentation has become the dominant approach for industrial-scale L-Trp production, especially due to its high productivity and cost-effectiveness [6,7,8].

In particular, fermentation processes employing Escherichia coli have gained popularity as the main approach for L-Trp production. However, since wild-type E. coli typically produce low L-Trp titers, research efforts have focused on developing more efficient L-Trp production strains, with metabolic engineering emerging as a key strategy for this purpose [9]. Indeed, in recent years, various techniques have been employed to optimize metabolic flux, such as by knocking out degradation genes (e.g., tnaA), weakening competing pathways (e.g., tyrosine and phenylalanine branches), and introducing key enzymes that are resistant to feedback inhibition (e.g., trpE^fbr) [10,11,12]. For instance, Hou et al. introduced aroG^S211F, trpE^{Q71K/S94N/C465Y}-trpABCD, and serA^H344A/N364A expression cassettes into the E. coli genome and achieved an L-Trp titer of 43.0 g/L by balancing growth and biosynthesis [13]. Similarly, through the metabolic engineering of E. coli, Cheng et al. enhanced L-Trp production to 38.8 g/L [14]. Nevertheless, in the face of the complex metabolic regulatory network, it is very time-consuming and labor-intensive to rely solely on metabolic engineering to improve L-Trp production [9]. At this point, biosensor-based, high-throughput screening technology is gaining more and more attention due to its rapid and efficient advantages [15,16,17].

Indeed, L-Trp biosensors enable real-time monitoring of intracellular metabolite concentrations by coupling a biorecognition element to a signal transduction system, thereby facilitating dynamic metabolic regulation and efficient strain screening [18]. In this case, some commonly used recognition elements include transcription factors [19], enzymes, and riboswitches [20,21]. For example, Gong et al. performed rational mutagenesis on the amino acid residues I57 and V58 in the ligand-binding domain of E. coli TrpR, a transcriptional regulator of L-Trp metabolism. The resulting variants, V58E and V58K, exhibited a 10-fold increase in biosensor expression, leading to the development of the TrpR1-PtrpO1 biosensor, with an improved dynamic range and ligand specificity for accurate L-Trp detection [19,22]. Furthermore, Liu et al. engineered a biosensor incorporating a highly selective riboswitch—a cis-acting regulatory RNA element located in the non-coding region of mRNA—and an L-Trp-specific aptamer conjugated to yellow fluorescent protein (YFP). When L-Trp was bound to the riboswitch, it induced a conformational change, thereby modulating mRNA translation and activating YFP expression. This engineered p15-ribo727 biosensor also demonstrated a low detection threshold, high sensitivity, as well as a broad dynamic range, thus underscoring its effectiveness for high-throughput screening [21,23].

In this study, a high-throughput screening platform based on an L-Trp biosensor was developed to efficiently construct L-Trp-producing E. coli strains. Using this platform, a mutant strain with a 1.94-fold increase in L-Trp titer was successfully identified. Whole-genome sequencing subsequently revealed ptsN as a key gene in L-Trp production. Further optimization of the synthetic pathway by enhancing extracellular secretion of L-Trp and improving precursor supply based on the ptsN gene led to the development of the final strain, zh08, which exhibited a 7.71-fold increase in L-Trp titer. This study demonstrated the successful application of biosensor technology for the efficient construction of high-yield L-Trp-producing E. coli strains.

2. Materials and Methods

2.1. Strains and Plasmids

Gene knockout and integration in E. coli were performed using the CRISPR/Cas9 system according to the procedure described by Jiang et al. [24]. The strains and plasmids used in this study are listed in Table S1.

2.2. Plasmid and Homologous Fragment Construction

The 727 gene fragment was synthesized using Gibson assembly with primers 727-1-F, 727-2-R, 727-3-F, and 727-4-R, along with the KOD One™ PCR Master Mix (Bailinke Biotechnology, Beijing, China). In addition, the YFP gene (Constructed in the lab, previously) was amplified by PCR from a previously constructed plasmid harboring the gene using primers YFP-F and YFP-R. The plasmid pACYC184-727 was then generated using linearized pACYC184 as the vector, with 500 bp homology arms also incorporated. For the construction of the P₄-yddG homologous fragment, the upstream gene was amplified with the primers 500-yddG-F and P₄-F-yddG-R, while the P₄ promoter [25] was amplified using the primers w-promoter-F and yddG-3-P₄-R. Similarly, the primers 800-yddG-F and 800-yddG-R were used to amplify the downstream gene. These different fragments were subsequently assembled using Gibson assembly to create the desired repair fragments. Finally, the pTarget-tnaA plasmid was also constructed by generating pTargetF fragments carrying the tnaA-targeted N20 sequence through reverse PCR. These fragments were then self-ligated with the BM seamless cloning kit (Biomed, Beijing, China). Other plasmids were constructed using a similar approach. All primers used in this study are listed in Table S2.

2.3. Media and Growth Conditions

E. coli cultures were grown at 30 °C or 37 °C in LB liquid medium (10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl) or on LB agar plates. When required, antibiotics were added to the culture medium at the following concentrations: 34 μg/mL chloramphenicol, 50 μg/mL kanamycin, and 50 μg/mL spectinomycin.

For 96-deep-well microtiter plate fermentation, individual colonies were inoculated into 2 mL square V-bottom deep-well plate (Corning Costar 3960, Tewksbury, MA, USA) containing 500 μL of LB medium and incubated at 37 °C and 800 rpm for 12 h. The strains were subsequently transferred to 96-deep-well plates containing 1 mL of Trp medium (27.5 g/L glucose, 7.5 g/L K₂HPO₄·3H₂O, 2.0 g/L MgSO₄·7H₂O, 1.6 g/L (NH₄)₂SO₄, 0.077 g/L FeSO₄·7H₂O, 2.0 g/L citric acid, 1.0 g/L yeast extract, 1 mL/L ionic liquid) and fermented at 37 °C and 800 rpm for 48 h. The ionic liquid contained 4.5 g/L MnSO₄·H₂O, 20.0 g/L Na₂SO₄, 6.4 g/L ZnSO₄·7H₂O, 0.6 g/L CuSO₄·5H₂O, 4 g/L CoCl₂·6H₂O.

For shake-flask fermentation, single colonies selected from LB solid medium were inoculated into LB medium, and following overnight incubation, the resulting cultures were transferred to 250 mL flasks containing 15 mL of Trp medium for a 48 h incubation at 37 °C and 250 rpm.

2.4. ARTP Mutagenesis

Log-phase bacterial cultures were harvested and resuspended in 1 mL of sterile water to achieve an OD₆₀₀ between 0.1 and 0.2. A 10 μL aliquot of the suspension was then pipetted onto the center of a sterile metal carrier plate, which was then placed into the appropriate groove of the ARTP mutagenizer (ARTP-IIIS, Yuanqing Tianmu Biotechnology, Wuxi, China). The power supply was set to 120 W, with a helium flow rate of 10 standard liters per minute (SLM) and at varying exposure times. Following irradiation, the metal plates were immediately transferred into LB liquid medium and incubated at 37 °C and 250 rpm for 30 min to allow recovery. The reactivated bacterial cultures were subsequently spread onto LB agar plates for overnight incubation at 37 °C. Each set of experiments was performed in triplicate.

Lethality was eventually calculated using the following formula:

Lethality (%) = (number of viable bacteria in the control group − number of viable bacteria in the treatment group)/number of viable bacteria in the control group) × 100

2.5. Flow Cytometric Cell Sorting (FACS)

Following ARTP mutagenesis and recovery, the bacterial suspension was diluted to an OD of approximately 0.1 before being subjected to FACS using a BD FACS flow cytometer (BD Biosciences, San Jose, CA, USA). In this case, the optimal FACS parameters were as follows: a nozzle size of 85 μm, a sorting mode set to Purity 3-way, a pressure of 30 psi, a passage setting of B515, and a sorting rate of 15 cells per second. In addition, the maximum emission and excitation wavelengths were 514 nm and 527 nm, respectively. The top 0.1% of YFP-positive single cells were selected based on YFP fluorescence intensity and transferred onto LB agar plates. These plates were then incubated at 37 °C until single colonies were obtained, with selected clones eventually cultured in 96-deep-well plates using a fully automated companion clone selector (QP Expression, Genetix, CA, USA) for subsequent fermentation experiments.

2.6. High-Throughput Pre-Treatment and Analysis Using the UTPA Platform

A pre-established ultra-throughput pre-treatment and analysis (UTPA) platform, consisting of two integrated modules, namely an automated sample pre-treatment workstation and an acoustic droplet ejection (ADE)-open port interface (OPI)-mass spectrometry (MS) system, was employed. In Module A, a customized sample pre-treatment device was developed based on the Beckman i7 liquid handling workstation (Beckman Biomek, Brea, CA, USA), and it was equipped with a 96-channel pipettor, a 384-channel pipettor, two grippers, and two shakers. Additional modifications included an automated centrifuge (Hettich Rotanta 460 Robotic, Kirchlengern, Germany), a positive pressure filter (Amplius-Positive Pressure ALP V3, Rostock, Germany), a consumable hotel (Thermo Fishier Cytomat 10, Dreieich, Germany), and a robotic arm (Beckman Scara, Brea, CA, USA). In Module B, the Echo MS system was used for ADE-OPI-MS analysis (SCIEX, Concord, ON, Canada). The ADE transducer was connected to a SCIEX triple quadrupole 6500+ mass spectrometer equipped with an OptiFlow ion ESI source via an OPI.

Bacterial cultures in 96-deep-well plates were pre-treated using Module A of the UTPA platform. Specifically, they were centrifuged at 4000 rpm for 10 min at 4 °C, with the resulting supernatants subsequently aspirated and transferred to new 96-well plates. This was followed by the addition of 80 μL of ultrapure water and, after vortexing, the plates were transferred to Module B for ADE-OPI-MS analysis. In this case, 2 mM NH₄F in a 50:50 (v/v) mixture of MeOH and ACN was used as the carrier liquid at a constant flow rate of 500 μL/min. For each sample, 2.5 nL (1 droplet) of solvent was ejected via ADE into the OPI and captured by the carrier liquid stream, which then transported it to the ESI ion source through a transfer capillary. An additional delay time of 1 s was applied between sample injections and added to the default transmission time of 1 s to ensure sufficient baseline separation. L-Trp was eventually detected using the 6500+ mass spectrometer in multiple reaction monitoring (MRM) mode under positive ionization polarity. The optimal MS parameters for flow injection analysis were also set as follows: MRM transition m/z 205→146.1, temperature at 300 °C, spray voltage at 5500 V, ion source gas 1 at 90 psi, ion source gas 2 at 70 psi, curtain gas at 25 psi, collision-activated dissociation gas at 9 psi, declustering potential at 40 V, collision energy at 20 V, and dwell time at 95 ms.

2.7. Genome-Wide Mutation Analysis

Sequencing data were processed with Breseq (v0.35.5). Initially, the sequencing reads were aligned to the E. coli ATCC 8739 reference genome using Bowtie 2. Multiple ancestral mutations detected during the preliminary analysis were then incorporated into a revised reference genome. An iterative realignment process was subsequently performed by recursively mapping ancestor reads to the updated reference until Breseq ceased reporting novel mutations. All predicted mutations underwent manual visual verification to eliminate false positives, with ambiguous genomic regions also further validated using Integrative Genomics Viewer (IGV) to ensure accuracy.

2.8. Analysis of OD₆₀₀ and L-Trp Accumulation

OD₆₀₀ was measured with a spectrophotometer (ND-2000, Thermo Fisher Scientific, Waltham, MA, USA). For L-Trp quantification, fermentation broth samples were diluted to an appropriate concentration and filtered through a 0.22 μm membrane. This was followed by centrifugation at 12,000 rpm and 4 °C for 10 min, with 1 mL of the resulting sample subsequently analyzed by HPLC (1200 series, Agilent Technologies, Santa Clara, CA, USA) equipped with an Innoval C18 column (4.6 mm × 250 mm, 5 μm, Bonna-Agela, Tianjin, China). The following conditions were applied for HPLC detection: a wavelength of 280 nm, a temperature of 35 °C, and a flow rate of 0.8 mL/min. In addition, the mobile phase consisted of 80% (v/v) of 0.1% phosphoric acid as phase A and 20% methanol as phase B.

3. Results and Discussion

3.1. Construction and Validation of the L-Trp Biosensor

As described earlier, Liu et al. previously developed an L-Trp-specific biosensor based on an aptamer and validated its performance through detailed characterization. Overall, that biosensor demonstrated a dynamic detection range of 0.2–10 mM for L-Trp along with high sensitivity, with minimal cross-reactivity with the other 19 amino acids. To achieve high-throughput screening of L-Trp-producing strains, an L-Trp biosensor, designated as pACYC184-727, was constructed in this study based on the method reported by Liu et al. This biosensor consisted of two key components: (i) An L-Trp-specific riboswitch, which underwent conformational changes upon L-Trp binding, thereby activating the expression of downstream genes. (ii) Yellow fluorescent protein (YFP), which served both as a reporter gene for biosensor characterization and as a selection marker for identifying highly productive strains during HTS [21].

To verify the functionality of the biosensor, the DH5α strain carrying pACYC184-727 was employed to test the function of the L-Trp biosensor treated with L-Trp. YFP fluorescence was observed under a fluorescence microscope. The results showed a significant increase in fluorescence in E. coli DH5α cells harboring pACYC184-727 upon L-Trp supplementation (Figure 1A). Furthermore, to evaluate the biosensor’s response, the pACYC184-727 plasmid was introduced into various L-Trp-producing strains. In this case, the fluorescence intensity of YFP in four different strains correlated positively with their L-Trp titers (Figure 1B). These results indicated that the biosensor could effectively translate intracellular L-Trp levels into optical signals in a concentration-dependent manner, thus making it a valuable tool for high-throughput screening of L-Trp-producing strains.

3.2. Determination and Optimization of ARTP Mutagenesis Conditions

The parental strain G01 was mutagenized as described in Section 2.4 to determine the optimal mutagenesis duration. Overall, strain lethality remained largely unchanged at mutagenesis times of 0, 10, 20, 30, and 40 s. However, at 45, 50, 55, 60, 65, 70, 75, and 80 s, the recorded strain lethality was 14.67%, 52.73%, 74.59%, 90.33%, 93.67%, 97.33%, 97.67%, and 97.70%, respectively (Figure 2). To achieve the target lethality rate (≥90%) [26] while ensuring sufficient mutant strain recovery, both the lethality curve trend and subsequent strain selection requirements were considered, and based on these factors, the optimal mutagenesis duration was determined to be 60 s.

3.3. High-Throughput Screening of Mutant Strains with Enhanced L-Trp Production

Following the optimization of the ARTP mutagenesis conditions, a mutant library was generated from strain G01 and subjected to high-throughput screening. The workflow of the screening platform is illustrated in Figure 3A. Initially, approximately 5000 cells were screened from the G01 mutant library containing approximately 5–10 million mutants using FACS, and the top 50 mutants were selected from 5000 mutants using 96-deep-well plate fermentation. Subsequently, the top 12 mutants were screened through repeated shake-flask fermentation until strains GT39, GT04, GT03, and GT23, which exhibited increases in L-Trp production per OD₆₀₀ of 76.96%, 42.28%, 28.22%, and 20.03%, respectively, were obtained (Figure 3B). Among these, GT39 demonstrated an L-Trp titer of 0.85 g/L, which represented a 1.43-fold increase compared with the parental strain G01 (Figure 3C). As a result, GT39 was selected as the starting strain for a second round of mutagenesis and screening. In the second round, approximately 5000 cells were screened from the GT39 mutant library containing approximately 5–10 million mutants using FACS. Again, the top 50 mutants were selected from 5000 mutants using 96-deep-well plate fermentation. Next, the top 17 mutants were selected for repeated shake-flask fermentation (Figure 3D). Finally, the mutant strain GT3938 was identified, and it exhibited an increase in L-Trp titer per OD₆₀₀ of 43.45%, with an absolute L-Trp titer reaching 1.03 g/L (Figure 3E).

To assess the genetic stability of GT3938, the strain was passaged three consecutive times before evaluating its growth and L-Trp production capacity. The OD₆₀₀ values after the three passages were found to be 16.00, 15.80, 15.97, and 15.80, while the L-Trp titers were 1.03, 1.06, 1.05, and 1.12 g/L, respectively (Figure 3F). Overall, the average L-Trp titer was 1.07 g/L, thus suggesting that strain GT3938 maintained stable growth and L-Trp production levels across passages, both of which were indicative of good genetic stability.

In this study, the high-throughput screening platform, based on the L-Trp biosensor, effectively facilitated the rapid identification of mutant strains with enhanced L-Trp production. Specifically, the final mutant strain, GT3938, exhibited a 1.94-fold increase in L-Trp titer as well as a 1.69-fold increase in OD₆₀₀ compared with the parental strain G01, while also displaying excellent genetic stability. Although high-throughput screening of high L-Trp-producing strains has been selected, the coupling technology of L-Trp biosensor combined with FACS still has many shortcomings, which limits its practical application. For example, the stability and environmental adaptability of L-Trp biosensors need to be improved [27], the output signal is inaccurate [28], and the sensitivity and detection threshold need to be improved [29]. Meanwhile, FACS can only analyze intracellular products and a non-specific environment is highly likely to lead to false positive results in screening, as reported by Han et al. [30]. Overall, the high-throughput screening technique based on L-Trp biosensors can be combined with molecular docking, deep learning, and droplet microfluidics in the future to further realize the accurate screening of high L-Trp-producing strains [31,32,33].

3.4. Sequencing Analysis and Validation of the L-Trp Mutant GT3938

To further elucidate the mechanisms underlying the high L-Trp titer of strain GT3938, genomic DNA was extracted from both G01 and GT3938 strains for whole-genome resequencing. Using the parental strain G01 as a reference, more than 99% of the sequencing reads from GT3938 aligned with the reference genome. Genome analysis subsequently revealed 13 single nucleotide variant (SNV) mutations, 8 insertion–deletion (InDel) mutations, and 6 copy number variation (CNV) mutations in the GT3938 genome (Tables S3–S5). The sequencing data were further analyzed using Breseq (version 0.35.5). In this case, the reads were mapped to the E. coli ATCC 8739 reference genome using Bowtie2, and any identified mutations were manually curated using Breseq and IGV to eliminate false positives as well as ambiguous regions. Through this analysis, we focused on some genes that may be related to L-Trp synthesis, such as tyrR, ptsI, ptsH, and ptsN. However, after a search of the literature, we learned that the effects of tyrR [34], ptsI [35], and ptsH [36] genes on the synthesis of L-Trp have already been reported, but not the ptsN gene. Therefore, we focused more to the ptsN gene and functional validation experiments were subsequently conducted.

Firstly, a ptsN-knockout strain (GT3938 ΔptsN), a strain carrying the plasmid pACYC184-P₄-ptsN (GT3938-1, GT3938 carrying pACYC184-P₄-ptsN), and an overexpression strain (zh01, GT3938 ΔackA::P₄-ptsN) were constructed, with the unmodified GT3938 serving as a control. These strains were then evaluated through 48 h shake flask fermentation. The results showed that knocking out ptsN decreased the L-Trp production to 0.17 g/L with an OD₆₀₀ of only 4.57, thus suggesting that ptsN is crucial for both growth and L-Trp synthesis in E. coli. Conversely, ptsN overexpression significantly increased L-Trp production by 21.36%, reaching 1.25 g/L with an OD₆₀₀ of 15.74. At the same time, the L-Trp titer of the strain GT3938-1 was 1.18 g/L with an OD₆₀₀ of 15.53 (Figure 4). These results indicated that overexpression of the ptsN gene significantly increased the titer of L-Trp. The ptsN gene, an atypical component of the phosphotransferase system (PTS), encodes the EIIANtr protein, which differs from classical sugar-transporting PTS components, such as EIIAGlc, involved in glucose uptake [37,38]. Instead, as a non-sugar-transporting component of PTS, EIIANtr mainly acts as a key regulator in nitrogen metabolism, carbon–nitrogen homeostasis, and stress responses, with its primary functions being: (i) to regulate nitrogen assimilation by modulating the phosphorylation status of related genes such as gltB and glnA; (ii) to balance carbon and nitrogen metabolism, including inhibition of the pyruvate dehydrogenase complex (PDH) and redirection of carbon flux; (iii) to participate in osmotic stress response and antibiotic resistance in bacteria. Furthermore, ptsN deletion has been linked to impaired stress tolerance, including decreased resistance to high salinity environments and potential alterations in the expression of efflux pump genes (e.g., acrAB) [39,40,41,42]. In summary, ptsN plays a critical role in L-Trp biosynthesis in E. coli. Therefore, its regulation and associated metabolic pathways should be key targets when employing metabolic engineering strategies for E. coli optimization.

3.5. Enhancing L-Trp Efflux to Improve Production

Although strain GT3938 achieved a 1.94-fold increase in L-Trp titer after high-throughput screening, further enhancement remained possible. Consequently, strain zh01 was used as a chassis to focus on improving L-Trp efflux efficiency (Figure 5A).

One of the primary limitations to higher L-Trp titer was the inefficient secretion of intracellularly synthesized L-Trp [5]. To address this, the L-Trp transporter gene tnaB was knocked out to create strain zh02 (zh01 ΔtnaB). A comprehensive review of the available literature subsequently identified YddG as the only known L-Trp efflux protein in E. coli. Therefore, the yddG gene was overexpressed in zh02 to generate strain zh03 (zh02 P₄-yddG). These strains were then evaluated via 48 h shake-flask fermentation, with the results showing that strain zh02 could achieve a 6.4% increase in L-Trp titer, reaching 1.33 g/L with an OD₆₀₀ of 15.06. Furthermore, strain zh03 demonstrated a 19.55% increase to reach 1.59 g/L with an OD₆₀₀ of 14.86 (Figure 5B). Altogether, the results confirmed that, in addition to ptsN overexpression, further enhancement of L-Trp production could be achieved by disrupting intracellular L-Trp transport and promoting L-Trp efflux. At the same time, the absence of significant differences in OD₆₀₀ between strains suggested that these genetic modifications did not negatively impact cell growth.

3.6. Improvement of L-Trp Production by Increasing Introcellular PRPP and L-Ser Supplies

In the L-Trp biosynthesis pathway, L-serine (L-Ser) and phosphoribosyl pyrophosphate (PRPP) are two essential precursors. Among these, PRPP is primarily synthesized via the pentose phosphate pathway, and its availability directly influences L-Trp production. A comprehensive literature review along with an in-depth analysis of metabolic pathways involved in L-Trp synthesis identified several key genes involved in PRPP synthesis, including zwf (encoding glucose-6-phosphate dehydrogenase), gnd (encoding 6-phosphogluconate dehydrogenase), rpiA (encoding ribulose phosphate isomerase A), and prs (encoding PRPP synthetase). Similarly, L-Ser biosynthesis requires the simultaneous action of three enzymes, namely phosphoglycerate dehydrogenase (PGDH) encoded by the serA gene, phosphoserine phosphatase (PSP) encoded by the serB gene, and phosphoserine aminotransferase (PAST) encoded by the serC gene. Altogether, these enzymes catalyze the conversion of 3-phosphoglyceric acid into L-Ser (Figure 5A).

In order to improve the supply of precursors, PRPP and L-Ser, genes prs, zwf, gnd, rpiA, serB, and serC were categorized according to PRPP supply and serine supply. The genes were optimized for expression using promoter engineering or gene copy number amplification. Moreover, only one promoter modification was performed for each gene. For gene copy number, some modifications were overexpressed at the original position of the gene without changing the copy number, and other modifications were overexpressed at other positions and only one copy was added. Eventually, the following strains were generated: zh04 (zh03 P₄-prs), zh05 (zh04 P₄-zwf), zh06 (zh05 P₄-gnd), zh07 (zh06 Δedd::P₄-rpiA), and zh08 (zh07 ΔadhE::P₄-serB-serC). These engineered strains were then evaluated via 48 h shake-flask fermentation, with the results showing that the L-Trp titer increased to 1.77 (OD₆₀₀ = 14.68), 2.01 (OD₆₀₀ = 13.32), 2.30 (OD₆₀₀ = 12.29), 2.81 (OD₆₀₀ = 11.18), and 3.05 g/L (OD₆₀₀ = 11.37) for strains zh04, zh05, zh06, zh07, and zh08, respectively (Figure 5C). However, compared with strain GT3938, strain zh08 exhibited a 28.94% reduction in OD₆₀₀. This could be attributed to incompatibility between promoter strength and certain genes or excessive gene copy number, which might have led to overexpression of certain genes, resulting in intracellular resource imbalances that disrupted normal cellular functions. For instance, previous studies by Hou et al. demonstrated that, while increasing the copy number of the CRISPR-associated transposase system improved L-Trp production (7.7% more for a 4-copy ZH-4 strain compared with a 3-copy ZH-3 strain), a significant decrease in biomass and glucose consumption was also noted [13]. However, despite the metabolic trade-off, strain zh08 achieved an L-Trp titer of 3.05 g/L in this study, with this yield representing a 2.96-fold increase compared with strain GT3938. Overall, these results confirmed that the applied metabolic modifications effectively enhanced L-Trp production capacity in E. coli.

4. Conclusions

In this study, an L-Trp biosensor was employed as an auxiliary tool to construct a mutant library using ARTP technology before establishing a high-throughput screening system via FACS. These strategies enabled the successful identification of GT3938, a high-titer mutant strain with an L-Trp production that was 2.94 times higher than that of the parental strain G01. To further elucidate the mechanism of L-Trp production in GT3938, whole-genome sequencing and validation were performed for strains GT3938 and G01. This analysis identified ptsN as a key gene influencing L-Trp biosynthesis and based on this result, targeted metabolic engineering strategies were employed, with L-Trp production eventually reaching 3.05 g/L, which represented a 7.71-fold improvement over strain G01. In summary, the approach developed in this study—constructing a high-throughput screening platform by integrating L-Trp biosensor-assisted screening with metabolic engineering—provides a robust strategy for developing high-yield L-Trp-producing strains.