Engineering Escherichia coli for Aromatic Compound Biosynthesis: Integrating Metabolic Engineering and Synthetic Biology

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Major Comments:

1. The manuscript covers well-trodden ground. While comprehensive, its novelty is unclear. Please clarify in the introduction how this review differs from existing excellent reviews (e.g., Shen et al. 2020, Tang et al. 2024, Hirasawa et al. 2025, all cited). What unique perspective or recent (2024-2026) emphasis does this review offer?
2. The review focuses heavily on strain engineering but largely ignores downstream processing, fermentation scale-up challenges, and economic feasibility. A section on techno-economic barriers (product toxicity, extraction yields, purification costs) would significantly strengthen practical relevance.
3. The review exclusively focuses on E. coli. While justified, a brief comparison with other industrially relevant hosts (Corynebacterium glutamicum, Saccharomyces cerevisiae, Pseudomonas putida) would contextualize why E. coli remains preferred for specific product classes and where alternatives may be superior.
4. Figure 1 mentions glucose, glycerol, and xylose, but the review lacks systematic discussion of how different carbon sources affect flux distribution, cofactor balance, and product profiles. Lignocellulosic hydrolysates and other renewable feedstocks deserve attention given sustainability goals.
5. The manuscript mentions cofactor balance as a bottleneck but does not systematically discuss cofactor engineering strategies across the described pathways. This is a major gap.
6. Product toxicity is mentioned as a bottleneck but not substantively addressed. Expand this section with specific cases where toxicity-limited titers and how engineering overcame them.
7. Figures 1-6 are visually dense and difficult to read, especially in grayscale printing. Many abbreviations are undefined in legend. Provide high-resolution, color figures with complete legends. Consider splitting complex figures into multiple panels.
8. The introduction mentions "biosensor-driven dynamic control" and CRISPR-based regulation, but the review lacks detailed case studies. Discuss specific biosensors and their application outcomes.
9. Metabolic burden is mentioned but not analyzed. Discuss strategies to decouple growth from production with quantitative examples from aromatic compound production.
10. Specify which isoenzymes (AroF, AroG, AroH) are feedback-inhibited by which aromatic amino acids. This is a critical context for engineering feedback resistance but is omitted. 
11. The distinction between tyrosol and hydroxytyrosol (antioxidant activities, production routes) is confusing. Clarify that hydroxytyrosol is the hydroxylated derivative and specify which pathways produce which compound.
12. It is advised to extend the future directives to lines 772-779: "Furthermore, we examine emerging approaches, including CRISPR-based regulation, biosensor-driven dynamic control, membrane engineering, and synthetic microbial consortia," and cite the below references as they provide environmental context, suggest CRISPR-based regulation, support One Health/sustainability framing, and nanoparticle-based strategies. (https://doi.org/10.47278/journal.abr/2023.044) (http://dx.doi.org/10.71081/cvj/2025.054) (https://doi.org/10.47278/journal.abr/2025.005).
13. It is advised to extend the future directives: "Despite significant progress, challenges related to pathway regulation, cofactor balance, metabolic burden, and product toxicity remain critical bottlenecks. Additionally, biofilm formation by multidrug-resistant E. coli poses challenges in both clinical and bioprocessing settings, and understanding factors that influence biofilm formation may inform the design of more robust production strains. " Cite the reference below, as it addresses biofilm formation relevant to industrial fermentation of metabolic burden and strain robustness (https://doi.org/10.47278/journal.ijab/2024.185).

Comments on the Quality of English Language

Several grammatical errors and awkward phrasings throughout. Professional English language editing.

Author Response

We acknowledge Reviewer 1's thorough critical evaluation, which has substantially strengthened the manuscript. We address each major comment below. All changes are incorporated in the revised manuscript, and the relevant sections are identified for each response. Modified text is reproduced in the response where appropriate and highlighted in red in the new version of the manuscript.

The manuscript covers well-trodden ground. While comprehensive, its novelty is unclear. Please clarify in the introduction how this review differs from existing excellent reviews (e.g., Shen et al. 2020, Tang et al. 2024, Hirasawa et al. 2025, all cited). What unique perspective or recent (2024–2026) emphasis does this review offer?

Authors' Response: We agree that the novelty statement was insufficiently articulated. We have revised the Introduction (Section 1) to explicitly position our work relative to these three reviews. Shen et al. [1] provide an extensive product-by-product survey of metabolically engineered microorganisms; Tang et al. [2] focus primarily on synthetic biology tools for aromatic amino acid biosynthesis; and Hirasawa et al. [3] offer a comparative analysis of E. coli and C. glutamicum strategies. The present review is distinguished by three features: (i) organization around metabolic nodes (CHA, L-Tyr, L-Phe, L-Trp) rather than by product class or toolbox, providing a systems-perspective framework for understanding how engineering decisions at one node propagate to downstream branches; (ii) incorporation of literature from 2023–2026, including several key papers published after the above reviews (e.g., β-arbutin hyperproduction at 30.5 g/L [32], de novo 5-HTP at 5.1 g/L [16], deoxyviolacein at 12.18 g/L [37], and mandelic acid at 9.58 g/L using CRISPRi [55]); and (iii) expanded coverage of emerging themes, specifically biosensor-driven dynamic control, CRISPRi flux regulation, OMV-based secretion, and synthetic microbial consortia, that are reshaping microbial cell factory design. This is now stated explicitly in the Introduction.

The review focuses heavily on strain engineering but largely ignores downstream processing, fermentation scale-up challenges, and economic feasibility. A section on techno-economic barriers (product toxicity, extraction yields, purification costs) would significantly strengthen practical relevance.

Authors' Response: We acknowledge this important gap. We have added a dedicated paragraph in Section 4 (Future Perspectives) (new lines 487-497) that explicitly discusses the disconnect between laboratory achievements and industrial realization, addressing: (i) the impact of product toxicity and hydrophobicity on downstream extraction and purification costs (e.g., curcumin, indigo); (ii) the challenges of using complex or renewable feedstocks (e.g., lignocellulosic hydrolysates) in scale-up; and (iii) a call for incorporating techno-economic analysis (TEA) and life-cycle assessment (LCA) from early strain design stages, citing relevant precedents in the metabolic engineering literature [63]. We also note, in the interest of scope management, that a comprehensive quantitative TEA section falls beyond the primary focus of this strain-engineering review; we have instead highlighted this as a critical future direction and identified specific bottlenecks where engineering decisions directly affect downstream processing. Additionally, product toxicity, one of the principal techno-economic barriers, is now systematically addressed in a new paragraph in Section 2.1, with specific case studies (see response to Comment 1.7 below).

The review exclusively focuses on E. coli. While justified, a brief comparison with other industrially relevant hosts (Corynebacterium glutamicum, Saccharomyces cerevisiae, Pseudomonas putida) would contextualize why E. coli remains preferred for specific product classes and where alternatives may be superior.

Authors' Response: We have added a new paragraph in Section 1 (Introduction) comparing E. coli with three alternative hosts (new lines 56-67). Specifically: C. glutamicum is noted for its industrial track record in aromatic amino acid overproduction and its absence of acetate overflow, making it particularly suited for L-Tyr and L-Phe production at scale [3]; S. cerevisiae is advantaged for plant-derived polyphenols requiring cytochrome P450 enzymes, which are more efficiently folded in the eukaryotic endomembrane system; and P. putida excels in tolerance to toxic aromatic intermediates owing to its robust membrane composition and expanded efflux pump repertoire [5]. The paragraph concludes by contextualizing why E. coli remains the most versatile chassis for the compound classes reviewed here: rapid growth, genetic tractability, and compatibility with synthetic biology tools, while acknowledging the complementary strengths of alternative hosts.

Figure 1 mentions glucose, glycerol, and xylose, but the review lacks systematic discussion of how different carbon sources affect flux distribution, cofactor balance, and product profiles. Lignocellulosic hydrolysates and other renewable feedstocks deserve attention.

Authors' Response: We have added a dedicated paragraph in Section 2.1 (Architecture and Control of the SHK Pathway) that systematically discusses the impact of carbon source on flux distribution and product profiles (new lines 107-121). The paragraph covers: (i) glucose metabolism via PTS, which creates a PEP competition constraint; (ii) glycerol as a non-PTS substrate that circumvents this limitation; (iii) xylose utilization via the PPP for E4P enrichment, as exploited in co-substrate strategies [12]; and (iv) emerging use of lignocellulosic hydrolysates and agricultural byproducts (e.g., whey powder [13]), with a note on the engineering challenges of inhibitor tolerance and feedstock variability. The paragraph explicitly connects carbon source selection to transporter engineering (Glf^Zm) and cofactor management.

The manuscript mentions cofactor balance as a bottleneck but does not systematically discuss cofactor engineering strategies across the described pathways. This is a major gap.

Authors' Response: We fully agree; this is a critical gap that we have now addressed. A new paragraph has been added in Section 2.1 dedicated to cofactor engineering strategies across SHK pathway branches (new lines 122-136). It covers: (i) transhydrogenase overexpression (pntAB or udhA) for NADPH/NADH interconversion; (ii) enhancement of the oxidative PPP for NADPH regeneration; (iii) expression of heterologous NADPH-generating enzymes; and (iv) engineering of cofactor-independent enzyme variants [4,5,63]. Compound-specific cases are cited: BH4 regeneration for 5-HTP [16], malonyl-CoA supply for polyketides [61], and NADPH demand for L-DOPA [18] and curcumin [19] hydroxylase reactions. This paragraph is explicitly cross-referenced in the relevant compound sections.

Product toxicity is mentioned as a bottleneck but not substantively addressed. Expand this section with specific cases where toxicity limited titers and how engineering overcame them.

Authors' Response: We have added a comprehensive paragraph on product toxicity in Section 2.1 (new lines 137-152), presenting specific cases with quantitative data: (i) 2-PE, with an inhibitory concentration of approximately 2–3 g/L despite biosynthetic capacity of up to 9 g/L, addressed through separated-strain co-culture and in situ extraction [11]; (ii) curcumin hydrophobicity managed by OMV-based export, achieving 978 mg/L [19]; (iii) indigo toxicity mitigated by membrane engineering and two-stage fermentation, achieving 12.9 g/L [58]; and (iv) CAD toxicity addressed by deletion of ten endogenous reductases/dehydrogenases combined with in situ product recovery, achieving 3.8 g/L [21]. This paragraph provides the compound-specific, strategy-linked discussion of toxicity that was missing from the original manuscript.

Figures 1–6 are visually dense and difficult to read, especially in grayscale printing. Many abbreviations are undefined in legends. Provide high-resolution, color figures with complete legends. Consider splitting complex figures into multiple panels.

Authors' Response: We acknowledge this legitimate concern regarding figure quality. All figure legends in the revised manuscript have been revised to include complete definitions of all abbreviations used therein. Specifically: Figure 1 now defines all transporter abbreviations (PTS, GalP, Glf^Zm, GlfP, XylFHG), all glycolytic and PPP intermediates (G6P, F6P, G3P, Gly3-P, PYR, AcCoA, X5P, E4P), all SHK pathway intermediates (DAHP, DHQ, DHS, SHK, S3P, EPSP, CHA), and all aromatic derivatives shown (QA, PCA, PAL, GA, PDC, HPP). The legend language has also been standardized across all figures: “Solid line arrows designate one enzymatic reaction; dashed arrows illustrate two or more enzymatic reactions.” Regarding resolution and color: we confirm that high-resolution color figures will be submitted as separate files per MDPI guidelines. We have also included a legend for each figure indicating that the figures were created with BioRender, which generates print-quality vector graphics. The legend for Figure 6 has been reorganized to more clearly delineate the content of each panel (A–H), improving readability. Splitting of complex figures into additional panels is under consideration for the final version, depending on editorial guidelines.

The introduction mentions 'biosensor-driven dynamic control' and CRISPR-based regulation, but the review lacks detailed case studies. Discuss specific biosensors and their application outcomes.

Authors' Response: We have expanded the discussion of specific biosensor and CRISPRi implementations in Section 4 (Future Perspectives) (new lines 454-461). Specifically: (i) the salicylate-responsive sensor-reporter system (based on the NahR transcription factor) used by Qian et al. [61] to rapidly identify optimal expression levels for SHK pathway enzymes is now explicitly described; (ii) tyrosine-responsive biosensors used for high-throughput screening of L-Tyr overproducers are mentioned in the context of dynamic regulation; and (iii) CRISPRi applications for repression of pheA and trpE branches to redirect flux toward 4-hydroxymandelate (9.58 g/L) [55] and 5-HTP [13] are now presented as concrete case studies with quantitative outcomes, referenced to Sections 3.2.2 and 3.3.1, respectively.

Metabolic burden is mentioned but not analyzed. Discuss strategies to decouple growth from production with quantitative examples from aromatic compound production.

Authors' Response: A new paragraph dedicated to metabolic burden analysis has been added in Section 4 (Future Perspectives) (new lines 462-475). It defines metabolic burden in the context of heterologous pathway expression and presents four categories of decoupling strategies with specific examples from aromatic compound production: (i) two-stage fermentation (growth phase followed by production phase, as used for indigo [20] and 2-PE [11]); (ii) dynamic regulation systems sensing growth rate or metabolite levels; (iii) chromosomal integration and auto-inducible systems that eliminate plasmid maintenance cost, exemplified by the CAD production strain achieving 3.8 g/L with a fully integrated, inducer-free system [21]; and (iv) synthetic microbial consortia that distribute biosynthetic burden across specialized strains [59]. The paragraph also identifies quantitative assessment via proteomic profiling and flux balance analysis as an important future research direction.

Specify which isoenzymes (AroF, AroG, AroH) are feedback-inhibited by which aromatic amino acids. This is a critical context for engineering feedback resistance but is omitted.

Authors' Response: This information has been explicitly incorporated in Section 2.1 (Architecture and Control of the SHK Pathway), where we now state: “These isoenzymes are subject to distinct feedback inhibition: AroF is feedback-inhibited by L-Tyr, AroG by L-Phe, and AroH by L-Trp.” (new lines 86-88) Additionally, in Section 4 (Future Perspectives), the opening paragraph reinforces this specificity when discussing the use of feedback-resistant variants (new lines 442-448). The same information is reflected in the revised legend of Figure 6D, which now explicitly attributes inhibition specificities to each isoenzyme.

The distinction between tyrosol and hydroxytyrosol (antioxidant activities, production routes) is confusing. Clarify that hydroxytyrosol is the hydroxylated derivative and specify which pathways produce which compound.

Authors' Response: We agree that the original text created ambiguity by discussing both compounds in overlapping sections. We have addressed this in two ways: (i) a new clarifying paragraph has been added in Section 2.3 (Chorismate as a metabolic node) that explicitly defines the structural relationship: Tyrosol is 2-(4-hydroxyphenyl)ethanol (new lines 215-223), while hydroxytyrosol is the catechol derivative 2-(3,4-dihydroxyphenyl)ethanol, and distinguishes their biosynthetic routes: Tyrosol can be synthesized from CHA-derived intermediates without requiring L-Tyr as an obligate intermediate, whereas hydroxytyrosol requires an additional hydroxylation step by HpaBC acting on tyrosol [29]; and (ii) in Section 3.1.1 (new lines 253-254), the text now explicitly states that “Hydroxytyrosol, the hydroxylated derivative of tyrosol bearing a catechol ring, is typically produced by the additional action of HpaBC on tyrosol [29].” The two compounds are no longer conflated.

It is advised to extend the future directives and cite references provided via URLs (https://doi.org/10.47278/journal.abr/2023.044; http://dx.doi.org/10.71081/cvj/2025.054; https://doi.org/10.47278/journal.abr/2025.005; https://doi.org/10.47278/journal.ijab/2024.185) in the context of CRISPR-based regulation, One Health/sustainability framing, nanoparticle-based strategies, and biofilm formation.

Authors' Response: We thank Reviewer 1 for these suggestions and have carefully evaluated each reference. Regarding the three references from the journal Advances in Biological Research (ABR) and the Clinical Veterinary Journal (CVJ): Upon review, the content of these publications, while addressing related microbiological themes, does not overlap directly with the specific engineering strategies for aromatic compound production in E. coli that are the focus of this review. Including them would not add substantive scientific content to the current manuscript and could be perceived as a tangential citation. We have therefore not included them. Regarding the reference on biofilm formation in E. coli from the International Journal of Applied Biology (https://doi.org/10.47278/journal.ijab/2024.185): We agree with the reviewer's general point that biofilm formation and strain robustness are relevant topics for industrial fermentation. However, the specific connection to the engineering strategies discussed in this review is indirect, and citing it solely in the context of “understanding factors that influence biofilm formation” would require a more thorough discussion of biofilm biology than is appropriate within our review's scope. We have instead added a general note in Section 5 (Conclusions) acknowledging strain robustness, including biofilm-related properties, as an area of future investigation in the context of industrial production strain design, without citing sources outside the direct scope of the review (new lines 538-552). We trust the Editor will appreciate our effort to maintain the scientific coherence and citation integrity of the manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript provides a current review of metabolic engineering strategies for aromatic compound production in Escherichia coli, focusing on the shikimate pathway and its key components – chorismate, L-tyrosine, L-phenylalanine, and L-tryptophan. The authors provide a significant collection of modern examples, including specific product titers and approaches such as dynamic regulation, CRISPR interventions, and synthetic consortia. The organization of the material by metabolic nodes makes the review easy to understand and useful for specialists.

However, the manuscript contains several substantive shortcomings.

1. The review remains primarily descriptive and lacks a critical analysis of the strategies presented. The authors report high titers for numerous compounds (arbutin 30.5 g/L, rosmarinic acid 5.8 g/L, indigo 12.9 g/L, etc.), but do not discuss which of these results are truly reproducible on an industrial scale and which remain laboratory achievements. There is no analysis of typical limitations—product toxicity, cofactor imbalances, metabolic burden – as applied to specific classes of compounds.
2. In the section on L-phenylalanine, the production of (S)-mandelic acid is described almost verbatim twice, using the same references (53 and 56). This duplication should be eliminated.
3. The Conclusions are formulated too generally and largely repeat the introduction and abstract. Instead of listing known challenges (regulation, cofactors, toxicity), the authors should specify which strategies described in the review were most effective in overcoming each of these barriers, and which ones fell short.
4. The text contains factual inaccuracies. For example, on page 5, the ubiC enzyme is listed as "ubichistomase lyase," whereas the correct name is chorismate pyruvate lyase (or ubiquinone biosynthesis lyase).
5. The discussion of future prospects (Section 4 and Conclusion) contains generalities ("machine learning," "DBTL loops") without addressing the specific challenges of the shikimate pathway. It would be useful to specify which specific parameters for aromatic compounds can be improved using these methods.

Author Response

We thank Reviewer 2 for the careful and precise reading of the manuscript. The comments are highly constructive and have led to concrete improvements.  All changes are incorporated in the revised manuscript, and the relevant sections are identified for each response. Modified text is reproduced in the response where appropriate and highlighted in red in the new version of the manuscript.

The review remains primarily descriptive and lacks a critical analysis of the strategies presented. The authors report high titers for numerous compounds (arbutin 30.5 g/L, rosmarinic acid 5.8 g/L, indigo 12.9 g/L, etc.), but do not discuss which of these results are truly reproducible on an industrial scale and which remain laboratory achievements.

Authors' Response: This is a well-founded criticism that we have addressed throughout the revised manuscript. We have added contextual qualifications at key points: (i) for dopamine (27 mg/L, shake flask), we explicitly note that “these are primarily laboratory achievements; translation to industrial scale would require cofactor regeneration systems, efflux pump engineering, and careful metabolic burden management” (Section 3.1.3) (new lines 279-287); (ii) for resveratrol (2340 mg/L from p-coumaric acid supplementation), we note that “these results represent high-performance laboratory demonstrations; bioreactor-scale production from glucose remains challenging and has been reported only at lower titers” (Section 3.1.4) (new lines 293-295); (iii) in the Conclusions (Section 5), we now explicitly contrast compound classes where scale-up has been demonstrated (e.g., salidroside at 16.8 g/L in a 5 L bioreactor [40] (new lines 556-558); L-DOPA at 25.53 g/L in bioreactor fermentation [18]) against those where results remain confined to shake flask or laboratory scale new lines 280-283). Additionally, the new downstream processing paragraph in Section 4 frames the general issue of the laboratory-to-industry gap and the role of TEA in guiding engineering priorities (new lines 487-497).

In the section on L-phenylalanine, the production of (S)-mandelic acid is described almost verbatim twice, using the same references (53 and 56). This duplication should be eliminated.

Authors' Response: We thank the reviewer for identifying this error. Upon revision, we confirmed that the previous references [53] and [56] in the original manuscript both cited Lukito et al. 2019 (Adv. Synth. Catal. 361:3560–3568), with the first entry erroneously attributed to whole-cell biocatalysis for (S)-mandelic acid and the second correctly attributed to (R)-mandelic acid from the same study. We have corrected this: reference [10] now properly cites Lukito et al. 2021 (Bioresour. Bioprocess. 8:22, DOI: 10.1186/s40643-021-00374-6) for (R)-mandelic acid production from multiple substrates, and reference [54] cites Lukito et al. 2019 for (S)-mandelic acid via whole-cell biocatalytic cascade. The duplicated descriptive passage for (S)-mandelic acid has been removed from Section 3.2.2, and the narrative now clearly distinguishes the two stereoisomers, their respective enzymes, and their reference studies (new lines 338-353).

The Conclusions are formulated too generally and largely repeat the introduction and abstract. Instead of listing known challenges (regulation, cofactors, toxicity), the authors should specify which strategies described in the review were most effective in overcoming each of these barriers, and which fell short.

Authors' Response: The Conclusions section (Section 5) has been substantially rewritten. Rather than restating known challenges, the revised Conclusions now provide a critical synthesis that directly addresses the reviewer's request: (i) feedback-resistant enzyme variants (AroG^fbr, AroF^fbr, PheA^fbr) and PTS replacement by Glf^Zm are identified as the most universally effective strategies for deregulating pathway entry and improving precursor supply; (ii) compound-specific toxicity solutions are evaluated comparatively (OMV-export for curcumin; membrane engineering for indigo; in situ recovery for 2-PE and CAD; two-stage fermentation for violacein); (iii) CRISPRi and biosensor-guided optimization are highlighted as particularly effective for fine-tuning competing branch point fluxes; (iv) fully de novo plant pathway reconstruction is characterized as less robust and scalable in its current state, limited by cofactor imbalances and enzyme activity; and (v) the gap between laboratory titers and industrially validated processes is identified as the most pressing challenge, with a clear forward-looking statement about integrating strain engineering with downstream process design.

The text contains factual inaccuracies. For example, on page 5, the ubiC enzyme is listed as 'ubichistomase lyase,' whereas the correct name is chorismate pyruvate lyase (or ubiquinone biosynthesis lyase).

Authors' Response: We sincerely thank the reviewer for catching this nomenclatural error. The enzyme encoded by ubiC is correctly named chorismate pyruvate-lyase (EC 4.2.99.21; also known as para-hydroxybenzoate synthase), which catalyzes the conversion of chorismate to 4-hydroxybenzoate (4-HBA) and pyruvate. The erroneous term “ubichistomase lyase” has been corrected throughout the manuscript. In the revised text (Section 2.3), the enzyme is referred to as “chorismate pyruvate-lyase (UbiC, encoded by ubiC),” with a parenthetical note clarifying the synonymous term “ubichorismate lyase” used in some older literature (new line 206).

The discussion of future prospects (Section 4 and Conclusion) contains generalities ('machine learning,' 'DBTL loops') without addressing the specific challenges of the shikimate pathway. It would be useful to specify which specific parameters for aromatic compounds can be improved using these methods.

Authors' Response: We agree that the original treatment of these topics was too generic. Section 4 now specifies concrete applications of ML and DBTL cycles to the SHK pathway: (i) ML-accelerated optimization of enzyme variant selection for improved cofactor specificity or reduced feedback sensitivity; (ii) ML-guided promoter and RBS strength combination screening for balanced multi-gene pathway expression, which is particularly relevant given the complex flux distribution across SHK branches; (iii) ML-assisted optimization of fermentation process parameters (carbon feeding strategy, dissolved oxygen, temperature profiling) for scale-up of aromatic compound production; and (iv) a reference to the successful application of automated DBTL in lycopene and amino acid production in E. coli as the nearest precedent [64] (new lines 498-509). These specific applications now ground the forward-looking discussion in the actual engineering bottlenecks identified throughout the review.

Reviewer 3 Report

Comments and Suggestions for Authors

This is a clearly written and well-organized review of the current status of using E. coli as a bioengineered organism for practical biosynthesis applications.  There are only a few minor corrections in the text that are suggested  revisions

Abstract

This journal requires that the Abstract must be 200 words maximum.  Your abstract is 279 words in length.

Line                 Recommended revisions

209. Muconic acid (MA) is an important -------

134–135          ------- and A(H1N1), which is currently used as the main anti-influenza antiviral against these viruses.   [correct wording recommended]

173      extension [4,5,7,20] (Figure 2).  Space needed

176      HBA, 3,4-dehydroxybenzoicacid (DHB), protocatechuic acid, and its ester with methanol (methylparaben, MP) ---

209  Muconic acid (MA) is an important precursor ------

232      4.09 g/ L of MA acid [16].

336, 337   sam5  (not sam5). 

389   In E. coli, the first---  no italics for the word  ‘In’

410      synthase (CURS1, from C. longa).  The word  'from' should not be italicized.

458        2-Phenylethanol (2-PE) is a fragrance -----

 490     --- Amycolatopopsis orientalis was achieved by overexpressing key enzymes of the SHK pathway

498      --- balancing, improved yields; while whole-cell biocatalysis---

501–502                      This long sentence needs clarification. Please consider the following.

The alternative use of HmaS homolog from Actinosynnema mirum for MA synthesis, and the subsequently enhanced SHK pathway, along with the supply of PEP and E4P and  the application CRISPR interference (CRISPRi) to repress competing pathways to redirect flux toward MA production, resulted in the production of 9.58 g/L of MA (using 0.5% glycerol and 0.05% glucose as carbon sources), the highest reported for microbial production [51].

514–515             The production of L-PHG by E. coli cascade biocatalysis from racemic mandelic acid--- (removed ‘in’)

533–534             Finally, integrating all expression units into the chromosomal DNA  created an auto-inducible system for antibiotic- and inducer-free production.         

536      This engineered strain was batch cultured, and in situ product----

557–558.    A similar strategy, using the flavonoid pinocembrin 7-O-methyltransferase from Eucalyptus nitida, gave a yield of 558 mg/L pinostrobin.

569      Solid line arrows designate one enzymatic reaction; dashed arrows illustrate two or more enzymatic reactions.

593–596             Recommended edit of the following complex sentence

Production of 5-HTP in the strain of E. coli C1T7-593 S337A/F318Y (with optimized promoter distribution of the 5-HPT hydroxylation by L-Trp  hydroxylase and BH4 synthesis and regeneration pathways) resulted in an increment of 60.3 % of 5-HTP compared to the initial strain.

598      Further scale-up to a 5 L fermenter improved yield at 1.6 g/L [60].

598–600                5-HPT production of 5.1 g/L was achieved if fed-batch cultures used glycerol as the carbon source in the  L-Trp hydroxylation pathway, including co-expressing human TPH and BH4 biosynthesis  and regeneration system in E. coli [61] (Figure 5).

620      ---- from Pseudomonas balearica DSM 6083)

621      --- from Pseudomonas putida

[It would be helpful if the genus name is spelled out the first time it is used in the text, where possible]

663–664          Solid line arrows illustrate one enzymatic reaction; dashed arrows illustrate two or more enzymatic reactions.

695      ----- from Z. mobilis have ---

701. --- remains a cornerstone strategy (Figure 6D); it is increasingly ---

718–719   ----- such as E. coli–E. coli co-culture systems for violacein production (Figure 6B),---

735      A. The development of parallel catabolic pathways ----- ( The letter A. was missing).

Please check carefully to be certain E. coli is italicized each time it occurs in this figure caption.

Author Response

We are grateful to Reviewer 3 for the detailed line-by-line editorial corrections. All suggested revisions have been incorporated. We address them systematically below. All changes are incorporated in the revised manuscript, and the relevant sections are identified for each response. Modified text is reproduced in the response where appropriate and highlighted in red in the new version of the manuscript.

This journal requires that the Abstract must be 200 words maximum. Your abstract is 279 words in length.

Authors' Response: The Abstract has been substantially condensed and now contains 190 words, within the required limit. The revised Abstract retains all essential elements: the scope of the review (SHK pathway; E. coli; high-value aromatic compounds), the organizational framework (metabolic nodes), the key engineering approaches covered, the emerging strategies addressed, the persistent challenges identified, and the overarching conclusion. Redundant phrasing and overly detailed enumerations have been eliminated.

Muconic acid (MA) is an important ——— 134–135 ——— and A(H1N1), which is currently used as the main anti-influenza antiviral against these viruses. [correct wording recommended]

Authors' Response: The sentence at the relevant position has been revised for clarity. The revised text now reads: “SHK is the chiral precursor for the antiviral oseltamivir phosphate (Tamiflu®), used against influenza A (H5N1) and A(H1N1).” (new lines 186-187). The connection between shikimate and muconic acid, which appeared confusingly in the same paragraph in the original, has been reorganized so that each compound is discussed in its dedicated section.

Line 173: extension [4,5,7,20] (Figure 2). Space needed.

Authors' Response: Corrected. The mention to Figure 2 was moved to the previous new line 195.

Line 176: HBA, 3,4-dehydroxybenzoicacid (DHB)... — space missing.

Authors' Response: Corrected. The compound name has been corrected to “3,4-dihydroxybenzoic acid (DHB, protocatechuic acid)” (new lines 200-201) throughout the manuscript, with proper spacing and the inclusion of the common name protocatechuic acid.

Line 209: Muconic acid (MA) is an important precursor ——— Line 232: 4.09 g/L of MA acid [16].

Authors' Response: Both corrected. The introductory sentence for muconic acid (MA) has been revised to avoid the redundant phrase “MA acid”, which now reads simply “4.09 g/L of MA [12].” (new line 212- 213). The first introduction of MA in Section 2.3 reads: “cis,cis-Muconic acid (MA) is an important precursor for adipic acid and terephthalic acid production.” (new lines 208-209).

Lines 336, 337: sam5 (not sam5). Correct formatting required.

Authors' Response: Corrected. Throughout the manuscript, the gene name sam5 now appears consistently in italics as sam5, in accordance with standard bacterial gene nomenclature (new line 274).

Line 389: In E. coli, the first — no italics for the word 'In'.

Authors' Response: Corrected. The word “In” at the beginning of all the sentences including “In E. coli” has been revised. Only the binomial name E. coli is italicized.

Line 410: synthase (CURS1, from C. longa). The word 'from' should not be italicized.

Authors' Response: Corrected. The word “from” preceding species names has been de-italicized throughout the manuscript. Only genus and species names are italicized, not the prepositions introducing them (new line 299).

Line 458: 2-Phenylethanol (2-PE) is a fragrance ——

Authors' Response: The paragraph introducing 2-PE has been revised for clarity and now begins: “2-Phenylethanol (2-PE) is a fragrance compound widely used in food and cosmetics.” The sentence structure and flow have been improved as indicated (new line 329).

Line 490: — Amycolatopopsis orientalis was achieved by overexpressing key enzymes of the SHK pathway.

Authors' Response: Corrected. The genus name has been corrected from “Amycolatopopsis” (erroneous) to “Amycolatopsis orientalis” throughout the manuscript (new line 339).

Line 498: — balancing, improved yields; while whole-cell biocatalysis — Lines 501–502: This long sentence needs clarification. [Suggested rewording provided by reviewer.]

Authors' Response: The original complex sentence at previous lines 501–502 has been replaced with the clearer formulation provided by the reviewer, with minor grammatical adjustment: “An alternative HmaS homolog from Actinosynnema mirum was employed for de novo mandelic acid (MA) synthesis, combined with enhancement of the SHK pathway, supply of PEP and E4P, and application of CRISPR interference (CRISPRi) to repress competing pathways. This strategy produced 9.58 g/L of MA using 0.5% glycerol and 0.05% glucose as carbon sources, the highest reported titer for microbial production [55].” (Section 3.2.2) (new lines 344-348). The sentence at the previous line 498 has been restructured with a period before “while whole-cell biocatalysis” to split it into two independent clauses (new line 342).

Lines 514–515: The production of L-PHG by E. coli cascade biocatalysis from racemic mandelic acid — (removed 'in').

Authors' Response: Corrected. The superfluous preposition “in” has been removed, and the sentence now reads: “L-Phenylglycine (L-PHG) was produced by E. coli cascade biocatalysis from racemic mandelic acid through co-expression of mandelate racemase, D-mandelate dehydrogenase,” (new lines 354-356).

Lines 533–534: Finally, integrating all expression units into the chromosomal DNA created an auto-inducible system for antibiotic- and inducer-free production.

Authors' Response: The sentence has been revised to improve clarity: “chromosomal integration of all expression units for antibiotic- and inducer-free production” (Section 3.2.3) (new lines 363-364).

Line 536: This engineered strain was batch cultured, and in situ product——

Authors' Response: Corrected. “in situ” now appears in italics throughout the manuscript as per scientific convention.

Lines 557–558: A similar strategy, using the flavonoid pinocembrin 7-O-methyltransferase from Eucalyptus nitida, gave a yield of 558 mg/L pinostrobin.

Authors' Response: We thank the reviewer for this correction. The titer for pinostrobin has been corrected from 153 mg/L (as stated in the original) to 558 mg/L in the revised manuscript (Section 3.2.3), consistent with the value reported by Hanko et al. [17] (new lines 369-370).

Line 569: Solid line arrows designate one enzymatic reaction; dashed arrows illustrate two or more enzymatic reactions.

Authors' Response: The arrow definition language has been standardized across all figure legends (Figures 1–6) to read: “Solid line arrows designate one enzymatic reaction; dashed arrows illustrate two or more enzymatic reactions.” This is now consistent with the reviewer's suggested wording.

Lines 593–596: Production of 5-HTP in the strain of E. coli C1T7-S337A/F318Y (with optimized promoter distribution of the 5-HPT hydroxylation by L-Trp hydroxylase and BH4 synthesis and regeneration pathways) resulted in an increment of 60.3% of 5-HTP compared to the initial strain. [Recommended edit provided.]

Authors' Response: The sentence has been revised in Section 3.3.1 following the reviewer's recommended wording, with minor grammatical editing: “Production in E. coli C1T7-S337A/F318Y with optimized promoter distribution for L-Trp hydroxylase and BH4 synthesis/regeneration pathways resulted in a 60.3% increase in 5-HTP compared to the initial strain, reaching 1.3 g/L 5-HTP from 2.0 g/L L-Trp using whey powder as substrate and inducer.” (new lines 397-401).

Line 598: Further scale-up to a 5 L fermenter improved yield at 1.6 g/L [60].

Authors' Response: Corrected. The sentence now reads: “Further scale-up to a 5 L fermenter improved the 5-HTP yield to 1.6 g/L [13].” (Section 3.3.1) (new line 401).

Lines 598–600: 5-HPT production of 5.1 g/L was achieved if fed-batch cultures used glycerol as the carbon source in the L-Trp hydroxylation pathway, including co-expressing human TPH and BH4 biosynthesis and regeneration system in E. coli [61] (Figure 5).

Authors' Response: The sentence has been revised for clarity: “A titer of 5.1 g/L 5-HTP was achieved by fed-batch culture using glycerol as carbon source, via co-expression of human TPH and the BH4 biosynthesis and regeneration system in E. coli [16] (Figure 5).” (Section 3.3.1) (new lines 401-403).

Lines 620, 621: — from Pseudomonas balearica DSM 6083 / — from Pseudomonas putida [It would be helpful if the genus name is spelled out the first time it is used in the text.]

Authors' Response: Corrected. The first mention of Pseudomonas species in this section uses the full genus name: “Pseudomonas balearica DSM 6083”. Subsequent mentions of the genus use the abbreviated “P. putida.” (new lines 412-413).

Lines 663–664: Solid line arrows illustrate one enzymatic reaction; dashed arrows illustrate two or more enzymatic reactions.

Authors' Response: Corrected and standardized across all figure legends (see response to Comment 3.16 above).

Line 695: — from Z. mobilis have — — remains a cornerstone strategy (Figure 6D); it is increasingly —

Authors' Response: This section was restructured for clarity: “and replacing the PTS with more efficient transporters (Glf^Zm from Z. mobilis) (Figure 6D) [4].” (new lines 478-479).

Lines 718–719: — such as E. coli–E. coli co-culture systems for violacein production (Figure 6B) —

Authors' Response: Corrected. The sentence now reads: “...such as E. coli–E. coli co-culture systems for violacein production (Figure 6B)...” with the binomial name properly italicized in both instances (new line 484).

Line 735: A. The development of parallel catabolic pathways — (the letter A. was missing).

Authors' Response: Corrected. The panel label “A.” has been reinstated at the beginning of the Figure 6 legend description. All panel labels (A through H) have been verified and are present and correctly formatted in the revised manuscript.

Please check carefully to be certain E. coli is italicized each time it occurs in this figure caption.

Authors' Response: Verified and corrected (new line 515). All instances of E. coli in the captions of Figures 1–6 have been carefully checked and are now consistently italicized. A global review of E. coli italicization throughout the manuscript was conducted, and additional instances of non-italicized E. coli were corrected.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Professional English language editing is still required.

Comments on the Quality of English Language

Several grammatical errors and awkward phrasings throughout. Professional English language editing.

Author Response

Comment — Quality of the English language: Several grammatical errors and awkward phrasings throughout. Professional English language editing is still required.

Authors' Response: We thank Reviewer 1 for reiterating this concern and have carefully considered it. We respectfully note, however, that this comment, both in the first and second rounds of review, has not been accompanied by any specific examples of grammatical errors or instances of awkward phrasing. In the absence of concrete examples, it is not possible to identify and correct specific passages, nor to objectively assess the nature or extent of the purported issues. We also note that Reviewer 2 and Reviewer 3, who independently evaluated the same manuscript in both rounds of review, explicitly selected the option "The English is fine and does not require any improvement" on the review form and did not raise any linguistic concerns in their written comments. This consistent assessment across two independent evaluations, both in the first and second rounds, indicates that the manuscript meets the linguistic standards expected for an international scientific publication. Furthermore, the manuscript was carefully reviewed for language quality prior to its original submission. During the revision process, particular attention was paid to the clarity and precision of all newly incorporated text. We also draw attention to point (IV) of the editorial checklist provided by the Section Managing Editor, which states: "If the reviewer(s) recommended references, critically analyze them to ensure that their inclusion would enhance your manuscript. If you believe these references are unnecessary, you should not include them." We apply the same principle of critical evaluation to all reviewer recommendations, including linguistic ones, particularly when they are not supported by specific, verifiable examples. We remain fully committed to the clarity and accessibility of our manuscript and are confident that it meets the publication standards of Microbiology Research.

We trust that the responses provided above, together with the revised manuscript, will meet the requirements for publication in Microbiology Research. We remain available to address any further questions that the Editor may have.

Sincerely yours,

Adelfo Escalante
Corresponding Author
On behalf of Silvana M. Tapia Cabrera and Francisco Bolívar