Expanding genetic engineering capabilities in Vibrio natriegens with the Vnat Collection

Abstract

Vibrio natriegens, with its exceptionally fast growth rate, has great promise as a revolutionary chassis for synthetic biology, yet the realization of its full potential has been limited by the lack of robust, standardized genetic tools. Here, we present the Vnat Collection, a comprehensive, modular toolkit specifically engineered to overcome these limitations. Leveraging optimized Golden Gate cloning strategies, we introduce improved junction sequences and a highly efficient dropout part system, achieving up to a 300-fold increase in assembly efficiency. Our toolkit significantly expands the synthetic biology toolbox by providing a wide array of characterized inducible promoters, enabling precise, orthogonal gene regulation, and novel operon connectors to streamline the construction of multi-gene pathways critical for metabolic engineering. Furthermore, we enhance genome editing workflows through refined NT-CRISPR methods, incorporating homology-flanked targeting constructs and demonstrating a simplified protocol that eliminates intermediate purification steps. With over 220 rigorously validated modular components, the Vnat Collection establishes an advanced standard for genetic engineering of V. natriegens, empowering researchers to efficiently harness this organism’s unparalleled potential for diverse biotechnology applications.

Graphical Abstract

Graphical Abstract.

Introduction

In the last decade, Vibrio natriegens has emerged as a fast-growing bacterial chassis for synthetic biology and biotechnological applications [1–3]. Traditionally, ideal chassis organisms in these fields are characterized by their short generation times, high substrate uptake rates, broad substrate spectra, carbon-efficient biosynthesis, and easy genetic manipulability [4, 5]. With synthetic biology evolving towards field-deployable applications, the utilization of more robust chassis thriving outside of standard laboratory conditions, such as marine environments, becomes increasingly important [4]. The marine bacterium V. natriegens naturally excels in most of these criteria, featuring an impressive doubling time of <10 min in optimal conditions, abundant ribosomes, and high cell densities [1]. However, limited genetic accessibility has restricted the full exploitation of V. natriegens as a biotechnological chassis. To address this, an expanding suite of genetic tools has been developed (Fig. 1), including protocols for DNA introduction via electroporation, conjugation, and chemical transformation [6–8, 9]. Additionally, genome engineering has been achieved through recombination, multiplex genome editing with natural transformation (MuGENT), transposon mutagenesis [6, 7, 10–12, 13] and CRISPR-based approaches, such as NT-CRISPR [14–16, 17]. Efforts to modify gene regulation in V. natriegens began in 2016 with the establishment of CRISPRi for targeted transcriptional repression [6, 12, 18, 19], while a genome-scale model was released in 2023 [20]. Systematic optimization of genetic parts and plasmids has progressed through characterization studies [8, 21–23] and resources like the Marburg Collection [24], which represents the most comprehensive collection of characterized genetic parts in V. natriegens to date.

Figure 1.

Timeline of genetic tool development for V. natriegens over the past six decades.

Designed as a Golden Gate Assembly (GGA) toolbox, the Marburg Collection enables modular, multi-gene plasmid construction from a standardized part library. This collection has facilitated diverse research applications, including recombinant protein production [25], fundamental genetic studies [26], and the development of additional genetic tools in V. natriegens [14, 16, 18]. Unlike pre-assembled backbone systems, the Marburg Collection allows de novo plasmid assembly, providing flexibility in choosing the origins of replication and selection markers. This adaptability has enabled its use beyond V. natriegens, such as Pseudomonas aeruginosa, Agrobacterium tumefaciens, and Lactiplantibacillus plantarum [27–29], and has facilitated applications across various research settings [30, 31].

NT-CRISPR complements plasmid-based engineering by enabling precise and scarless genome modifications, including deletions, insertions, and single-base changes [14]. This two-step approach involves genome editing via natural transformation and homologous recombination, followed by CRISPR–Cas9-mediated counterselection of unedited cells. This allows for markerless genome editing in V. natriegens, ranging from single-base edits to the insertion of multi-kilobase sequences, such as those needed for chromosome fusion [26]. Further, NT-CRISPR has aided the exploration of efficient natural plasmid transformation in V. natriegens [32] as well as gene regulation based on CRISPRi [18].

Both the Marburg Collection and NT-CRISPR surpass earlier genetic engineering methods [6, 7] in efficiency, flexibility, and part diversity. However, extensive use of both tools has unveiled opportunities for further optimization. In this study, we present a series of improvements to both the Marburg Collection and NT-CRISPR, addressing limitations encountered in their broader application. To enhance the flexibility of GGA in V. natriegens, we expand the available toolkit with a set of orthogonal inducible promoters and introduce a streamlined strategy for routine operon construction. We also resolve a key technical bottleneck by redesigning dropout parts to overcome compatibility issues with restriction enzymes used in part replacement. In parallel, we optimize the NT-CRISPR workflow to increase its time efficiency, simplify the selection of genetic parts, and support future expansion of the toolkit. Together, these enhancements, released as the Vnat Collection, strengthen the utility of V. natriegens as a genetic chassis and provide practical resources for the synthetic biology community.

Materials and methods

Bacterial strains and media

All V. natriegens bacterial strains used in this study are listed in Supplementary Table S12. The base V. natriegens strain used in this study, derived from DSMZ 759 (ATCC 14048), carries ΔVNP1 + 2 and Δdns mutations, improving plasmid transformation efficiency, osmotolerance, and pyruvate production [33]. For simplicity, this strain is referred to as ‘base strain (BS)’ in this study. Unless otherwise specified, V. natriegens was cultured in LBv2 medium, either in liquid or on solid plates. For testing of the inducible promoters, medium was inoculated directly from glycerol stocks of three independent biological replicates of V. natriegens. Glycerol stocks of V. natriegens were prepared by mixing 750 μl of stationary-phase cultures (grown in LBv2 with the appropriate antibiotic for ∼5 h at 37°C and 220 rpm) with equal volumes of 50% glycerol and stored at −80°C. If required, kanamycin and chloramphenicol were added to LBv2 or minimal media for V. natriegens at a final concentration of 200 and 4 μg/ml (2 μg/ml for solid medium), respectively. Vibrio natriegens plates were stored at room temperature to maintain colony viability [7]. For genetic manipulation, E. coli DH5α and DH10β were used for transformation with GGA product, handling plasmids up to 5 and 5–12 kb, respectively. Escherichia coli cells were cultured in LBI medium supplemented with the required antibiotic: 50 μg/ml kanamycin for level 0* and level 1 cloning, and 25 μg/ml chloramphenicol for level 0 and level 2 cloning. Glycerol stocks of E. coli were prepared by mixing 850 μl of stationary-phase cultures (grown in LBI with the appropriate antibiotic for ∼16 h at 37°C and 220 rpm) with 750 μl of 50% glycerol and stored at −80°C. LBI medium contained tryptone (10 g/l), yeast extract (5 g/l), and NaCl (10 g/l), while LBv2 medium included an additional V2 salt mixture (NaCl 11.9 g/l, KCl 0.3 g/l, MgCl₂ 2.2 g/l) following Weinstock et al. [7]. All media were autoclaved at 121°C for 15 min. Solid media were prepared by adding 1.5% (w/v) agar before autoclaving.

For minimal medium experiments, V. natriegens was inoculated into MOPS2 or M9 minimal media. MOPS2, derived from Neidhardt medium [34], was supplemented with 2% (w/v) (342.2 mM) NaCl to mimic ocean salinity. The final 1× MOPS2 media (pH 7.2) contained MOPS (42 mM), Tricine (4 mM), K₂SO₄ (0.276 mM), CaCl₂ (0.5 μM), MgCl₂ (0.525 mM), NaCl (total 392.2 mM), FeSO₄ (10 μM), K₂HPO₄ (1.32 mM), NH₄Cl (9.5 mM), and the micronutrients, (NH₄)₆Mo₇O₂₄ (3 nM), H₃BO₃ (0.2 μM), CoCl₂ (1.5 nM), CuSO₄ (0.48 nM), MnCl₂ (4 nM), and ZnSO₄ (0.48 nM) (for preparations of micronutrients stock see [34]). Carbon sources included either 5 g/l (61 mM) sodium acetate (acetate-MOPS2) or 5 g/l (33 mM) arabinose (arabinose-MOPS2). Components were sterilized by 0.22 μm filtration. Agar plates were prepared by adding 3 g of agar to 100 ml ddH₂O, microwaving in a sterile container, and cooling to ∼70°C before adding MOPS2 components. Sodium acetate stock solution (20% w/v) was neutralized to pH 7 with HCl.

M9 minimal medium, based on SubtiWiki formulations, was supplemented with 2% (w/v) NaCl to mimic ocean salinity. The final 1× M9 media (pH 7.0) contained Na₂HPO₄ (6.78 g/l), KH₂PO₄ (3.0 g/l), NH₄Cl (1.0 g/l), sodium chloride NaCl (20.5 g/l), CaCl₂·2H₂O (0.0147 g/l), MnCl₂·4H₂O (1.0 mg/l), ZnCl₂ (1.7 mg/l), CuCl₂ (0.39 mg/l), CoCl₂·6H₂O (0.6 mg/l), Na₂MoO₄·2H₂O (0.6 mg/l), MgSO₄·7H2O (0.246 g/l), FeCl₃ (0.0081 g/l), trisodium citric acid dihydrate C₆H₅Na₃O₇·2H₂O (0.0294 g/l), and glucose as carbon source (w = 0.1%, 1 g/l). All components were filter-sterilized (0.22 μm).

Plasmid assembly

All plasmids used in this study are listed in Supplementary Table S14. The oligonucleotides employed for constructing new Vnat Collection parts are provided in Supplementary Table S13. Overhang sequences for assembling level 0 and level 0* parts following the Vnat Collection standard are detailed in Supplementary Table S11. Plasmid assembly followed the GGA method as described by Pryor et al. [35], with modifications. GGA reactions were carried out in 15 μl volumes, containing 15 fmol of each plasmid, 1.5 μl T4 ligase buffer [New England Biolabs (NEB)], 0.75 μl Hi-T4™ DNA ligase (400 U/μl, NEB), and 0.5 μl BsaI-HF^®v2 (20 U/μl, NEB) for level 0* and 1 constructs or 0.5 μl BsmBI-v2 (10 U/μl, NEB) for level 0 and 2 constructs. Alternatively, BsmBI was replaced with equal volumes of its isoschizomer Esp3I (10 U/μl, NEB) when cycling GGA reactions at 37°C. Reactions with BsaI-HF^®v2 were cycled between 37°C and 16°C for 1.5 and 3 min, respectively, for 25 cycles (unless stated otherwise). Reactions with BsmBI-v2 were cycled between 42°C and 16°C for 5 min each for 60 cycles (unless stated otherwise). A final incubation at 37°C or 42°C for 20 min completed restriction of input parts. For the assembly of level 2 dropout plasmids (without curated overhangs), an additional 10-min ligation step at 16°C was added at the end of the last cycle. The resulting GGA products were incubated at 80°C to deactivate the enzymes and kept at 4°C indefinitely until the transformation step. For dropout-part replacement, 0.5 μl BpiI (20 U/μl, NEB) was added to the one-pot GGA reaction mix in addition to BsaI-HF^®v2 or Esp3I.

To assemble the homology flank plasmids, the homology flanks were amplified from the genome of the base strain with Q5 polymerase (NEB) using primers listed in Supplementary Table S5. The reaction conditions are listed in Supplementary Tables S6 and S7. To ensure compatibility with the Vnat Collection cloning standard, primers (listed in Supplementary Table S9) were designed to remove BsmBI sites internal to the homology flanks, replacing them with synonymous codons appearing at similar frequencies in V. natriegens (https://www.kazusa.or.jp/codon/cgi-bin/showcodon.cgi?species=691). If required, 1 μl of DpnI (20 U/μl, NEB) was added to 50 μl polymerase chain reaction (PCR) product for template digest, before incubating at 37°C for 1 h. The PCR product was subsequently purified with E.Z.N.A.^® Cycle Pure Kit (Omega Biotek) and was cloned into the level 0* entry vector pVC_V*_01 from the Vnat Collection. The resulting plasmid was transformed into E. coli, verified through colony PCR, and extracted with E.Z.N.A. Plasmid DNA Mini Kit (Omega Biotek).

Vibrio natriegens competent cells and transformation

Chemically competent V. natriegens Δdns (base strain) cells were prepared as described by Stukenberg et al.[24]. Briefly, overnight cultures were inoculated from glycerol stocks into LBv2 medium and incubated at 37°C with shaking at 220 rpm. The next day, cultures were diluted to an initial OD₆₀₀ of 0.01 and grown to mid-exponential phase (OD₆₀₀ = 0.5–0.7). Cells were harvested by centrifugation, washed with cold TB buffer (10 mM Pipes, 15 mM CaCl₂, 250 mM KCl, pH 6.7, and 55 mM MnCl₂), and resuspended in TB buffer with 350 μl dimethyl sulfoxide (DMSO). After incubation on ice, aliquots were snap-frozen in liquid nitrogen and stored at −80°C. Plasmid transformation, described by Stukenberg et al. [24], was performed by incubating 50–150 ng of DNA with competent cells on ice for 30 min, followed by a 45-s heat shock at 42°C. After recovery in LBv2 medium at 37°C for 1 h with shaking at 600 rpm, cells were pelleted, resuspended in a small volume (∼100 μl) of medium, and plated on antibiotic-containing LBv2 agar.

Escherichia coli competent cells and transformation

Chemically competent E. coli cells were prepared following Xiyan with minor modifications [36]. Overnight cultures in LBI medium (37°C, 220 rpm shaking) were diluted 1:1000 into 300 ml of fresh LBI and grown at 37°C until OD₆₀₀ reached 0.5–0.7. Cells were harvested, washed in cold TFB1 buffer, pelleted, resuspended in TFB2 buffer, and stored at −80°C in 50 μl aliquots. The culture was transferred to pre-chilled Falcon tubes and incubated on ice for 10 min before being centrifuged at 3000 × g and 4°C. After removing the supernatant, the pellet was resuspended in 10 ml of TFB1 buffer containing potassium acetate (3 g/l), MnCl₂·4H₂O (10 g/l), RbCl (100 mM), CaCl₂ (10 mM), and glycerol (177 g/l). Cells were pelleted by centrifugation at 3000 × g and 4°C, and the supernatant was removed. The pellet was resuspended in 6 ml of TFB2 buffer, containing Na-MOPS (10 mM, pH 7), RbCl (10 mM), CaCl₂ (35 mM), and glycerol (151 g/l). The competent cells were aliquoted in 50 μl and stored at −80°C for future use. For transformation, 5 μl of GGA product was added to 50 μl of competent cells, incubated on ice for 30 min, heat shocked at 42°C for 45 s, and then cooled on ice for another 10 min. Cells were recovered in LBI medium at 37°C with shaking at 600 rpm for 60 min, then plated onto LBI agar with the appropriate antibiotic.

Vibrio natriegens strain development and genetic engineering

Vibrio natriegens mutants were generated following the NT-CRISPR protocol according to Stukenberg et al.with minor modifications [14]. Guide RNAs (gRNAs) for the NT-CRISPR plasmids (listed in Supplementary Table S10) were designed with the ‘Find CRISPR Sites’ function in Geneious Prime 2021 with algorithms developed by Doench et al. [37]. gRNA oligonucleotides were annealed by mixing 1.5 μl of the complementary oligonucleotides (100 μM) with 5 μl T4-DNA Ligase buffer (New England Biolabs) in a reaction volume of 50 μl. The mixture was heated to 95°C for 15 min in a PCR machine and was left in the machine to cool down slowly to room temperature for ∼1 h. The annealed oligonucleotides were then cloned into the NT-CRISPR plasmid (pST_116) via GGA.

To perform the NT-CRISPR workflow, V. natriegens cells containing the respective NT-CRISPR plasmid were inoculated from glycerol stock in 5 ml LBv2 medium with 100 μM IPTG and 4 μg/ml chloramphenicol overnight in a glass culture tube at 30°C at 220 rpm. Natural transformation was initiated by diluting 3.5 μl of the overnight culture in 350 μl sea salt solution (28 g/l sea salts from Sigma Aldrich, catalog number S9883-500G) with 100 μM IPTG. Unless indicated otherwise, 100 ng of transforming DNA (tDNA) was supplied as a circular plasmid (listed in Supplementary Table S14). Samples were incubated without shaking at 30°C for 5 h prior to addition of 1 ml LBv2 with 200 ng/ml aTc to induce the expression of the CRISPR–Cas9 system. The cultures were further incubated at 30°C at 300 rpm for 1 h. Finally, the cells were diluted 1:100 and 20 μl were plated on LBv2 agar with 400 ng/ml aTc and 2 μg/ml chloramphenicol (one plating per replicate).

For the V. natriegens landing pad strains, the P_Tet-mScarlet-I cassette is induced only in the final NT-CRISPR steps upon aTc addition, ensuring minimal metabolic burden while enabling direct visual selection of edited colonies when needed. Landing pad strains can be readily transformed with NT-CRISPR plasmids targeting mScarlet-I (e.g. pST_116_mScarlet) to facilitate gene replacements. When landing pad strains were stored as glycerol stocks at −80°C, colony counts remained consistent whether the workflow was initiated from fresh colonies or directly from glycerol stocks (data not shown). This eliminates the need for prior cloning and transformation steps, streamlining the workflow by reducing hands-on time in the early stages of the protocol.

The genetic modification of the resulting colonies was verified using colony PCR. Positive colonies underwent plasmid curing by growing colonies in 100 μl LBv2 medium for 5 h. The cells were streaked out to obtain single colonies and individual colonies were patched on LBv2 solid medium with and without 2 μg/ml chloramphenicol. Colonies that only grew on the medium without antibiotics were considered plasmid-cured and stored as glycerol stocks. Genomic DNA from cured strains was extracted, purified, and verified via Sanger sequencing.

Plate reader assays

Growth kinetics of V. natriegens were measured in black flat-bottom 96-well plates (Greiner, catalogue number 655096) using a BioTek Synergy HTX plate reader. Overnight cultures were pre-grown in the same medium used for the assay. Since most inducible promoters tested are growth-phase independent, strains carrying mScarlet-I reporter constructs were grown to stationary phase (5 h for LBv2 and M9, 24 h for acetate-MOPS2) before dilution into fresh medium (1:2000 for LBv2 and M9, 1:100 for acetate-MOPS2). Inducers were added either at the start of the experiment (LBv2 and acetate-MOPS2) or after 50 min (M9) (details in Supplementary Tables S1 and S2). To minimize evaporation, outer wells and gaps between wells were filled with ddH₂O. For P_BDA promoter experiments [24], V. natriegens was grown to OD₆₀₀ = 0.5 before dilution to 0.01 in arabinose-supplemented LBv2 medium.

Plates (100 μl culture per well) were sealed with a lid and parafilm for NT-CRISPR-related assays, while experiments on inducible promoters and operons were measured without a lid. Incubation was at 37°C with continuous orbital shaking at 425 or 807 cpm (details in Supplementary Table S3). Fluorescence was recorded for mScarlet-I (λ_Ex 530 ± 25 nm, λ_Em 620 ± 15 nm) and sfGFP (λ_Ex 485 ± 20 nm, λ_Em 530 ± 25 nm). OD₆₀₀ values were blank-subtracted using the average of at least three blank wells containing fresh medium. Data were computationally aligned to a uniform starting OD₆₀₀: 0.025 for acetate-MOPS2, 0.015 for LBv2, and arabinose-MOPS2, with no adjustment for M9 experiments. Growth rates were determined by linear regression between OD₆₀₀ = 0.01–0.1 (M9), 0.015–0.1 (LBv2), and 0.025–0.1 (acetate-MOPS2) in R. Dose–response curves for inducible promoters in M9 were based on mScarlet-I fluorescence 5 h post-induction. The minimum values for the normalized mScarlet-I signal in the dose–response curves were determined by dividing the minimal mScarlet-I increment [1] by the mean blank-subtracted OD₆₀₀ measured 5 h after induction (∼0.2).

Flow cytometry and computational analysis

For each inducible promoter, four biological replicates (two biological and two technical) were grown at four inducer concentrations in a 96-well plate. Precultures were prepared as described for plate reader experiments, grown in M9 medium with 2 μg/ml chloramphenicol. After 5 h, cultures were diluted 1:2000 in fresh M9 medium (2 μg/ml chloramphenicol, 100 μl per well) and incubated in a BioTek Synergy HTX plate reader with double orbital shaking (807 cpm). Induction with four different inducer concentrations occurred after 50 min.

Five hours post-induction, 100 μl of culture was fixed with 12.1 μl ice-cold formaldehyde (4.0% final concentration) and incubated at 4°C for 1 h. Fixed samples were centrifuged (10 000 × g, 10 min), resuspended in fresh M9 medium, and stored at 4°C for up to 7 days. Before flow cytometry, cells were stained with HOECHST dye (Thermo Fisher Scientific, R37165) following manufacturer instructions and diluted in 0.5× phosphate buffered saline + 5 mM ethylenediaminetetraacetic acid (EDTA) (NaCl 4 g/l, KCl 0.1 g/l, Na₂HPO₄ 0.71 g/l, KH₂PO₄ 0.12 g/l, 5 mM EDTA) to achieve appropriate cell densities.

Measurements were performed on a BD LSRFortessa flow cytometer, with HOECHST (λ_Ex 355 nm, λ_Em 450 ± 50 nm) and mScarlet-I (λ_Ex 561 nm, λ_Em 582 ± 15 nm) fluorescence measured with appropriate lasers/filters. Flow rate was set to 12 μl/min, and a minimum of 10 000 events were recorded. Data analysis was conducted using a custom R script based on the Bioconductor package flowCore (version 2.16.0).

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT to improve the readability and language clarity of the manuscript. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the content of the published article.

Results

Enhancing promoter control in V. natriegens

To enhance orthogonal and multifactorial control of gene expression in V. natriegens, we sought to expand the Vnat Collection with well-characterized inducible promoters. In total, we tested nine promoters by incorporating them upstream of an mScarlet-I reporter cassette on level 1 plasmids (Fig. 2A; plasmid assembly plans in Supplementary Table S14). This allowed direct measurement of promoter activity through fluorescence. Out of the nine promoters, P_3B5B, P_Cin, P_Tet, P_Tac, P_Ttg, P_PhlF, and P_SalTTC were previously characterized in E. coli [38], while P_{Tet_MC} stemmed from the Marburg Collection and P_Cin+1 was developed as a variant of P_Cin as described below. Detailed experimental conditions and information on the respective inducers are provided in Supplementary Tables S1 and S2.

Figure 2.

Characterization of inducible promoters in V. natriegens. (A) Schematic of the level 1 plasmid design used for promoter characterization, featuring inducible promoters and their respective regulator genes upstream of an mScarlet-I reporter cassette. (B) Dose–response curves and growth rates for characterized inducible promoters. Each promoter (P_Ttg, P_Cin, P_Cin+1, P_PhlF, P_Tet, P_{Tet_MC (Marburg Collection)}, P_Tac, P_3B5B, and P_Sal) is associated with a repressor (TtgR, CinR, PhlF, TetR, LacI, PcaU, or NahR) encoded in the same construct in divergent orientation from the target promoter. Fluorescence was measured 5 h post-induction in M9 minimal medium. Growth rates were calculated from OD₆₀₀ = 0.01 until OD₆₀₀ = 0.1. Data represent the mean of four replicates, with error bars indicating standard deviation (SD). The fold change was calculated based on the normalized mScarlet-I signal without induction and the highest observed signal. The dashed line indicates instrument sensitivity as defined in the ‘Materials and methods’ section. P_{Tet_MC} originates from the Marburg Collection [24], while all other promoters were previously characterized in E. coli [38]. Detailed experimental conditions and inducer concentrations are provided in Supplementary Tables S1, S2, and S3. (C) Orthogonality assay for inducible promoters, measuring mScarlet-I fluorescence 5 h post-induction in M9 minimal medium. Induction was performed 50 min after growth initiation using the second-highest effective inducer concentration [as determined in panel (B)]. Data represent the mean of four replicates.

Initial plate reader assays produced dose–response curves for each inducer concentration (Fig. 2B). Most promoters exhibited tight regulation, with minimal expression in the absence of an inducer. Notably, P_3B5B, P_Cin, P_Cin+1, P_Tet, P_{Tet_MC}, P_Tac, and P_Ttg demonstrated strong induction and low basal activity, making them ideal candidates for controlled expression. However, P_PhlF and P_SalTTC displayed higher basal expression in V. natriegens, a phenomenon not observed in E. coli [38]. Interestingly, DAPG activated P_PhlF at 1000-fold lower concentrations than in E. coli [38]. Unexpectedly, we observed that P_Cin exhibited poor induction efficiency in V. natriegens, while showing tunable expression in E. coli [38]. We speculated that the binding of the CinR repressor to the P_Cin operator site is excessively strong in V. natriegens, effectively blocking transcription even in the presence of an inducer. To address this, we modified the last four base pairs of the P_Cin promoter, creating the P_Cin+1 variant, which significantly improved its dynamic range, achieving up to 1000-fold induction. Overall, P_Ttg, P_Cin+1, P_Tet, P_Tac, and P_3B5B exhibited the highest fold-change induction and minimal leakiness, making them the preferred choice for future applications in V. natriegens.

We next assessed whether inducers affected bacterial growth. Growth rate measurements in M9 minimal medium revealed no significant impact at most tested concentrations, except for naringenin and salicylate, which inhibited growth at concentrations ≥0.05 mM and ≥200 μM, respectively (Fig. 2B). To assess time-dependent promoter activity, we monitored mScarlet-I expression over 8 h. Only P_PhlF exhibited time-dependent expression, likely due to growth-phase effects or leakiness (Supplementary Fig. S1). Flow cytometry analysis confirmed that all promoters exhibited homogeneous population behavior, except for naringenin-induced P_Ttg, which caused significant morphological variability at 1 mM (Supplementary Figs S2 and S3). This also complicates growth rate measurements shown in Fig. 2B for this promoter. In conclusion, we recommend the use of naringenin at concentrations below 100 μM, which is sufficient for full induction of this promoter.

We evaluated promoter orthogonality by measuring mScarlet-I expression at maximum induction levels, testing both the addition of cognate and non-cognate inducers for each promoter (Fig. 2C; relative crosstalk shown in Supplementary Fig. S4). P_Tet, P_Tac, and P_3BB5 exhibited strong orthogonality, with minimal crosstalk. However, naringenin and OHC14 slightly induced unintended reporter gene expression for all non-cognate promoters, an observation that was not made in E. coli [38]. This may be explained by the roles of naringenin, a flavonoid, and OHC14, an N-acyl homoserine lactone, as signalling molecules associated with quorum sensing and biofilm formation in many Gram-negative bacteria, including various Vibrio species [39–41, 42]. Thus, it is possible that these compounds broadly influence cellular metabolism and modulate the quorum sensing pathway in V. natriegens, potentially explaining the unintended promoter activation. Furthermore, as previously demonstrated, the promoters P_PhlF and P_Sal exhibit poor regulation, with notable basal activity even in the absence of their respective inducers, DAPG and sodium salicylate. This lack of tight regulation renders P_PhlF and P_Sal impractical for use in V. natriegens. In summary, the orthogonal promoters P_Tet, P_Tac, and P_3B5B are best suited for independent titration of multiple target genes in V. natriegens, enabling precise and concurrent regulation of synthetic genetic circuits.

New operon connectors for expanded gene assembly

The original collection design [24] allows up to five independent transcription units (TUs) to be assembled into a level 2 plasmid, each comprising a promoter, ribosome binding site (RBS), coding sequence (CDS), and terminator. Additionally, 3′- and 5′-connectors link individual TUs to the level 1 plasmid backbone, providing fusion sites for level 2 plasmid assembly. By selecting specific connector parts, individual TUs can be arranged in a desired sequence and orientation within the level 2 plasmid. Here, we introduce a new set of operon connectors, enabling the assembly of operons with up to five genes (RBS and CDS) under a single promoter (Fig. 3A and B). These new operon connectors replace the original 5′-connector and promoter parts with 5′-operon connectors, while the 3′-operon connectors substitute for 3′-connector and terminator parts, seamlessly linking consecutive genes within the same operon. The Vnat Collection now includes five 5′- and five 3′-operon connectors, providing a flexible, modular system for arranging genes within an operon.

Figure 3.

Design and experimental validation of operon connectors. (A) The Vnat Collection introduces a novel framework for assembling level 2 plasmids containing operons with up to five genes. These genes (synonym operon units, OUs) replace classical TUs within a level 2 plasmid, flanked by 5C-connectors, 3C-connectors, and the plasmid backbone. (B) Newly developed operon connectors enable seamless linking of operon units (OU) within an operon by replacing the 3′-connector and terminator regions in genes 1–4, as well as the 5′-connector and promoter regions in genes 2–5. (C) As a proof of concept, an operon was constructed to produce deoxyviolacein and sfGFP. The operon included five CDSs (vioA, vioB, vioC, vioE, and sfGFP), each paired with an RBS. Functionality was demonstrated by the visible production of deoxyviolacein and sfGFP expression, quantified using a plate reader assay in LBv2 medium (details in Supplementary Table S3). Data from three biological replicates were analysed; error bars represent SD of the mean.

To validate the operon design, we assembled a level 2 plasmid comprising five ‘operon units’ (RBS and CDS flanked by one or two operon connectors) encoding the deoxyviolacein biosynthesis genes and sfGFP, all driven by the P_Tet promoter (Fig. 3C; plasmid assembly plans in Supplementary Table S14). Correct assembly was confirmed via sequencing. Upon induction with aTc, the operon supported production of VioA, VioB, VioE, and VioC, resulting in the synthesis of the violet pigment deoxyviolacein [16], which was moderately but consistently visible in the engineered strain (Fig. 3C). Prior experiments from our group (data not shown) indicate that pigmentation intensity is highly sensitive to the expression levels of vioA, vioB, and vioE, suggesting that the observed phenotype reflects suboptimal but functional pathway flux. Inducible expression was further confirmed by monitoring sfGFP fluorescence in a plate reader, which showed a ∼30-fold increase above background (Fig. 3C). Note that the fold change in sfGFP expression was lower than that of mScarlet-I under control of the same P_Tet promoter (Fig. 2B), likely due to V. natriegens’ elevated background fluorescence in the green channel, as observed previously [24]. Overall, these results demonstrate that the operon connectors enable reliable assembly and coordinated expression of multigene pathways.

Improved dropout parts with curated restriction sites

Dropout parts streamline plasmid assembly by acting as placeholders for single components or entire TUs. Dropout parts comprise a complete expression cassette to produce a fluorescent protein (sfGFP or mScarlet-I), flanked by outward-facing Type IIS restriction sites. These sites remain intact in assembled dropout plasmids, allowing subsequent replacement via GGA and convenient identification of correct colonies based on color. Plasmids constructed with a dropout part effectively serve as an extended backbone template, reducing the number of input components required in a GGA and thereby enhancing cloning efficiency (Fig. 4A). For example, assembling a level 1 plasmid using a preassembled dropout plasmid and a promoter part in a two-part assembly was 100 times more efficient than assembling the same plasmid from eight individual parts (Fig. 4A; plasmid assembly plans in Supplementary Table S14).

Figure 4.

Enhanced dropout parts for streamlined plasmid assembly. Dropout parts serve as placeholders to facilitate the construction of preassembled level 1 and level 2 plasmids, incorporating a fluorescent protein expression cassette (sfGFP or mScarlet-I) for visual colony identification. The updated design replaces outward-facing BsaI sites with BpiI sites to facilitate efficient dropout replacement. (A) Unlike the standard assembly of level 1 plasmids, which requires eight input parts, the use of a level 1 dropout plasmid enables GGA with just two parts. This strategy enhances assembly efficiency up to 100-fold, as demonstrated with the assembly of a level 1 plasmid containing an mScarlet-I cassette. (B) Incorporating BpiI recognition sites significantly improves the efficiency for the assembly of dropout plasmids, achieving a more than 300-fold increase over the original design. All efficiency data were obtained from two (A) or three (B) technical replicates for colony forming units (CFU). Plasmid assembly plans are detailed in Supplementary Table S14.

The original collection design used the same restriction enzymes for dropout plasmid assembly and dropout replacement, reducing efficiency of cloning dropout plasmids due to the risk of final digestion of the assembled dropout plasmids. To overcome this, we redesigned the dropout system to use distinct restriction enzymes: BsaI or Esp3I for plasmid assembly and BpiI for dropout replacement (Fig. 4B). BpiI, an isoschizomer of BbsI, recognizes GAAGAC (5′→3′) and works efficiently alongside the collection’s standard enzymes (BsaI for level 1 and Esp3I for level 2 assembly) in a one-pot reaction. We recommend Esp3I over BsmBI for one-pot reactions with BpiI, as both enzymes function optimally at 37°C.

To ensure compatibility with the existing collection, we verified that commonly used level 0 and 0* parts lack BpiI sites. Additionally, curated dropout parts are now available in two colours and in various positions of level 0 parts (pos. 2, 3, 4, 5, and 2–5). The updated design significantly improves assembly efficiency, increasing cloning success by over 300-fold (Fig. 4B). These findings highlight both the efficiency boost from reducing input parts with dropout plasmids and the successful integration of distinct restriction enzymes in single-pot reactions.

Streamlined NT-CRISPR workflow for rapid genome editing

The NT-CRISPR method, introduced by Stukenberg et al. in 2022, enables rapid and scarless genome editing in V. natriegens, producing mutants in under a week [14]. It builds upon MuGENT system [11] by integrating natural transformation and homologous recombination with CRISPR-based counterselection. This way, high efficiency with as little as 1 ng of tDNA and editing success rates of up to 99.99 % can be achieved [14]. However, generating the tDNA and assembling NT-CRISPR plasmids remained labour-intensive, involving multiple cloning and purification steps (Fig. 5B).

Figure 5.

Schematics of the original and optimized landing pad NT-CRISPR workflow. (A) In the NT-CRISPR workflow, the NT-CRISPR plasmid is introduced into the target strain via transformation. V. natriegens’ natural transformation is induced with IPTG, enabling the uptake of tDNA, which integrates into the genome through homologous recombination. However, only a minority of cells are successfully edited. Counterselection is achieved by inducing CRISPR activity with aTc, which eliminates non-edited cells. The optimized landing pad strategy introduces specialized strains containing a chromosomally integrated TU with an aTc-inducible mScarlet-I cassette at the target loci. During NT-CRISPR, the mScarlet-I TU is replaced by the tDNA, supplied as an unpurified GGA product. Homologous recombination between the homology flanks on the tDNA plasmid and the genome of V. natriegens enables the targeted insertion of up to five TUs at a specific genomic locus. Upon plating, aTc induction activates the CRISPR system to eliminate most unedited cells and simultaneously induces mScarlet-I expression, marking surviving unedited cells with red fluorescence. This approach enables efficient dual-colour screening for successfully edited colonies. (B) Comparison of the original NT-CRISPR workflow from Stukenberg et al. [14] and the new landing pad strategy for genome engineering in V. natriegens developed in this study.

Design of the landing pad strategy

To streamline the NT-CRISPR process, we implemented a landing pad strategy incorporating two key innovations (Fig. 5). First, we developed V. natriegens strains with defined genomic landing pads, where an aTc-inducible mScarlet-I cassette replaces the original chromosomal genes. Thus, successful NT-CRISPR edits eliminate this marker, while non-edited cells retaining mScarlet-I can be visually identified by red fluorescence. This fluorescence-based screening significantly improves selection efficiency and reduces reliance on colony PCR validation. Second, we strategically targeted seven genomic loci for substitution with the landing pads, including genes related to arabinose metabolism (araA, araFGH), flagella assembly (fli), polyhydroxyalkanoate production (pha), overflow metabolism (pta), prophage induction (VNP1), and biofilm formation (vps) (Supplementary Table S4). Deletion of these genes has been reported to enhance resource allocation and improve microbial fitness in biotechnological applications. For example, deleting the vps operon reduced fermentation viscosity [43], while removing VNP1 increases tolerance to hypo-osmotic stress [33], making V. natriegens more robust for industrial-scale use.

To further simplify genomic insertions, the landing pad strategy integrates with the Vnat Collection to allow users the creation of level 2 tDNA plasmids with arbitrary cargo (multiple TUs) flanked by homology regions matching the targeted genomic loci (Fig. 5). These homology regions, stored in level 0* plasmids, replace the 5C- or 3C-connectors in level 2 plasmids, enabling seamless insertion of desired TUs in a 5′-homology-cargo-3′-homology architecture. Homology flank orientation is determined relative to the target coding sequence, while TU orientation is configurable via connector part selection in level 1 plasmid assembly. Overall, the landing pad strategy reduces the time required for genomic insertions in V. natriegens from 5 to 4 days and simplifies the workflow by requiring significantly fewer experimental steps (Fig. 5B). Notably, unlike the original NT-CRISPR protocol, which relies on purified linear PCR amplicons as tDNA, our method allows for the direct use of GGA products as circular tDNA (see below), bypassing the need for multiple PCR steps and plasmid amplification in E. coli.

Testing the landing pad strategy

To validate the landing pad strategy, we integrated an insulated, aTc-inducible mScarlet-I expression cassette (Fig. 6) into the seven genomic loci mentioned earlier (cf. Supplementary Table S4), using a Δdns VNP1 + 2 deletion as the V. natriegens base strain (see the ‘Materials and methods’ section). Fluorescence measurements across a range of aTc concentrations confirmed consistent mScarlet-I expression in all engineered mutants, demonstrating that the genetic insulator effectively prevented interference from native promoter regions (Fig. 6A). Notably, maximal induction was achieved at just 10 ng/ml aTc, 10-fold lower than in plasmid-based reporters (cf. Fig. 2B), probably caused by differences in the regulatory environment and copy numbers of the respective transcription factors.

Figure 6.

Expression and growth analysis of V. natriegens mutants carrying an insulated P_Tet-mScarlet-I cassette at different genomic loci. (A) Relative fluorescence units normalized to OD₆₀₀ at 5 h. (B) Growth rates, calculated from the average logarithmic slope between OD₆₀₀ = 0.015 and 0.10. Data were obtained from four replicates measured across three independent days. Raw data were blank-corrected (average of three blanks) and computationally aligned to an initial OD₆₀₀ of 0.015.

Although we hypothesized that these knockouts might improve growth due to carbon sink reductions, no significant differences were observed in rich medium (LBv2) (Fig. 6B, kinetics in Supplementary Fig. S8). However, when tested in minimal medium with acetate as the sole carbon source, the vps::P_Tet-mScarlet-I strain showed a significantly higher growth rate and reduced variability between replicates than the control (VNP1::P_Tet-mScarlet-I), particularly at 10 ng/ml aTc induction (Supplementary Fig. S6, kinetics in Supplementary Fig. S9). This suggests that exopolysaccharide (EPS) biosynthesis imposes a significant burden, limiting optimal growth, as observed in Xanthomonas campestris, where EPS production is similarly costly [44]. At 100 ng/ml aTc, additional mutants (araAC::P_Tet-mScarlet-I, araFGH::P_Tet-mScarlet-I, phaBAPC::P_Tet-mScarlet-I, pta::P_Tet-mScarlet-I) displayed improved growth rates, likely because the burden caused by higher mScarlet-I expression amplifies advantages of eliminating superfluous carbon sinks. In contrast, the flagellar knockout (fli::P_Tet-mScarlet-I) did not show significantly improved growth, likely because flagella synthesis is already downregulated in acetate medium due to acetyl-phosphate inhibition [45, 46]. Future studies integrating proteomic analyses could further elucidate how these genetic modifications enhance carbon utilization efficiency, guiding further strain optimization for bioproduction applications.

Pre-assembled tDNA dropout plasmids for genomic integration and knockouts

To simplify tDNA assembly, we developed pre-assembled level 2 dropout plasmids containing 5′- and 3′-homology flanks for each of the seven target loci (cf. Supplementary Table S4), along with an mScarlet-I dropout cassette (Figs 7A, 1). This dropout cassette serves as both a fluorescent selection marker during cloning and as a placeholder, which can be replaced with various cargo (Figs 7A, 2–5). As cargo, we first introduce a short, seven-amino-acid linker sequence (GGSGGSA) (https://parts.igem.org/Part:BBa_K243004), which bridges the 5′- to the 3′-homology flanks. This creates a link between the remaining CDS in the 5′- and 3′-homology flanks to form a non-functional minigene, enabling in-frame genomic knockouts that minimize polar effects on the expression of adjacent genes (Figs 7A, 2). Alternatively, the dropout cassette can be replaced with up to five TUs into each targeted locus (Figs 7A, 3), while for applications requiring fewer than five TUs, the unused positions can be bridged with an end linker part (pVC_0*_EL) to link the last TU to the 3′-homology flank (Figs 7A, 4).

Figure 7.

Offering versatile tDNA plasmid designs with the homology flank plasmid strategy and gene expression utilizing the native P_BDA promoter. (A) This schematic illustrates different tDNA plasmid configurations, enabling flexible genomic modifications in V. natriegens. Some 5′ homology flanks include native promoter sequences, allowing targeted expression or transcriptional regulation studies. (1) Dropout tDNA plasmid serves as a placeholder for later part replacement. (2) Knockout tDNA plasmid utilizes a seven-amino-acid linker (GGSGGSA) to create an in-frame minigene by fusing the remaining amino acids in the 5′ and 3′ homology flanks (pVC0_EL_5_EL5C1-3C5O). (3) Multi-TU insertion tDNA plasmid allows genomic integration of up to five TUs. (4) Native promoter-driven tDNA plasmid—leverages 5′ homology flank promoter sequences for controlled gene expression or transcriptional regulation analysis. (5) Insulated TU insertion tDNA plasmid designed for targeted integration of one or more TUs using end-linkers (EL); landing pad strains are generated using an insulated P_Tet-mScarlet-I as the TU1. (B) Schematic overview of the different strains used to characterize the utilization of the native P_BDA promoter to drive mScarlet-I expression in the araAlocus. (C) Normalized fluorescence units at 10 h. Cells were grown to stationary phase (∼5 h) and diluted 1:1000 into fresh LBv2 medium. Cultures were then grown to mid-exponential phase (OD₆₀₀ ∼0.5) before a 1:100 dilution into induction media with varying arabinose concentrations. (D) Heatmap of mScarlet-I production kinetics over 4 h after cultures reached OD₆₀₀ = 0.1, providing complementary insight into the timing and stability of P_BDA-driven induction under different conditions. Fluorescence measurements compared the mutant to the parental base strain carrying either an empty or P_BDA-mScarlet-I plasmid, as well as a ΔaraA mutant with a P_BDA-mScarlet-I plasmid. Data were collected from four biological replicates over two independent experiments. Raw data were blank-subtracted using three blanks and computationally aligned to start at OD₆₀₀ = 0.015.

Interestingly, for two of the dropout plasmids [phaBAPC::dropout (mScarlet-I) and pta::dropout (mScarlet-I)], we observed lower fluorescence values than from the other dropout plasmids (Supplementary Fig. S7), likely due to native promoter interference from the 5′-homology flanks. To mitigate undesired interference with the expression of cargo genes, we recommend introducing a genetic insulator, consisting of two transcriptional terminators, in the 5′-connector of TU1s (Figs 7A, 5), as in our previous design to characterize integration loci (cf. Fig. 6). This insulator effectively blocks transcriptional readthrough from native promoters, ensuring stable and independent TU expression.

Exploitation of native promoters

A key advantage of the landing pad strategy is its ability to harness native chromosomal promoters for heterologous gene expression, leveraging endogenous regulatory networks. This is achieved by positioning the RBS and CDS of the first TU near the 5′-homology flank, which contains or is positioned downstream of a native promoter. This setup is facilitated by using short 5′-connectors and a non-functional ‘placeholder’ promoter from the Vnat Collection when constructing the level 1 TU plasmid, allowing native promoters to drive expression while permitting customization of the RBS, CDS, and terminator. As a proof of concept, we targeted the araA locus to utilize the chromosomal P_BDA promoter. Deleting araA has been shown to enhance P_BDA sensitivity to arabinose induction in E. coli [47]. To minimize background fluorescence while maintaining the CDS close to the native promoter, we integrated a TU containing a non-functional placeholder promoter (P_Dummy) upstream of an RBS, mScarlet-I CDS, and terminator (pKM1_004). We then compared the fluorescence level in the mutant to those in a V. natriegens base strain carrying an empty plasmid (GFN1-127), a P_BDA-mScarlet-I plasmid, and a ΔaraA mutant with a P_BDA-mScarlet-I plasmid (Fig. 7A–C). To account for growth-phase dependency of the P_BDA promoter [24], cultures were synchronized at OD₆₀₀ = 0.005 from exponentially growing cells.

The ΔaraA mutants exhibited significantly increased P_BDA promoter activity, demonstrating enhanced sensitivity to arabinose (Fig. 7C). While wild-type cells required 5 mg/ml arabinose for maximal induction, ΔaraA mutants reached peak expression at just 1 μg/ml, representing a 5000-fold reduction in the required arabinose concentration. This likely results from increased arabinose stability in the ΔaraA strain, as araA encodes the enzyme catalysing the first step of l-arabinose catabolism [48, 49]. The observed increase in sensitivity aligns with previous findings in E. coli, where araBAD knockouts led to increased arabinose sensitivity [47, 50]. Additionally, fluorescence intensity was ∼10-fold lower from chromosomally integrated mScarlet-I than from plasmid-based expression, consistent with the medium-copy nature of pMB1-M plasmids [24]. However, despite enhanced arabinose sensitivity, high induction levels were only observed at later growth stages and higher OD₆₀₀, regardless of araA deletion (Fig. 7D). This suggests that the P_BDA promoter remains subject to carbon catabolite repression, with its activation increasing as alternative carbon sources are depleted in late exponential growth. These findings highlight the importance of considering regulatory influences such as catabolite repression when exploiting native promoters. To further refine their application in V. natriegens, strategies such as targeted mutations in catabolite repression pathways or controlled co-feeding of carbon sources may be necessary to achieve consistent promoter activation.

Editing efficiency for different tDNA forms

To further streamline the NT-CRISPR protocol, we evaluated the feasibility of using circular plasmids as tDNA instead of linearized plasmids or PCR amplicons. Given that V. natriegens can take up circular plasmids via overexpression of the competence master regulator tfoX [32] and that circular plasmids can be used for homologous recombination [51], we hypothesized that circular plasmid DNA could serve effectively as tDNA in genome editing. To test this, we performed NT-CRISPR to replace the flagella operon (fli, 30.8 kb) with an aTc-inducible mScarlet-I expression cassette, using three tDNA forms in equimolar quantities (50 fmol): purified PCR product (as per the original NT-CRISPR protocol), purified circular plasmid, and ScaI-linearized plasmid (Fig. 8A). The quantification of CFU showed similar numbers of edited cells for PCR-derived and linearized plasmids, whereas circular plasmid tDNA resulted in an ∼50% reduction in CFU (Fig. 8B). This suggests that while circular DNA is taken up and used for recombination, its efficiency is lower than that of linear DNA.

Figure 8.

Editing efficiency assessment using different tDNA forms in the NT-CRISPR system. (A) Schematic representation of the different tDNA forms used for integrating the P_Tet-mScarlet-I TU into the fli locus. (B) Colony forming units (CFUs) counted across four replicates (two biological replicates from two independent experiments). (C) Colony PCR results using primers outside the homology flank (n = 80, representing 20 colonies per biological duplicate from two independent experiments). (D) NT-CRISPR plate using unpurified GGA product as tDNA, plated with different dilution factors (20 μl per plate). (E) Verification of araA deletion by streaking the same colony on LBv2 (growth) and 0.5% (w/v) arabinose MOPS minimal medium (no growth). Statistical analysis in panel (B) was performed with an unpaired two-way t-test.

Since both the NT-CRISPR and tDNA plasmids confer chloramphenicol resistance, we verified successful genomic modification through colony PCR on 80 randomly selected colonies from each tDNA type. Strikingly, 97.5% (78/80) of colonies showed correct integrations for both linear and circular tDNA forms (Fig. 8C). Despite a lower CFU count, circular tDNA achieved high integration efficiency, simplifying NT-CRISPR by eliminating the need for PCR amplification and purification. These findings align with recent work by Specht et al. [32], showing that circular plasmid uptake in V. natriegens relies on tfoXand type IV pilus-mediated transformation [52]. The data suggest that circular plasmid DNA may undergo degradation or shearing before being taken up as single-stranded DNA for recombination, consistent with natural transformation mechanisms in Gram-negative bacteria [53].

Building on this observation, we hypothesized that unpurified GGA products could serve as viable tDNA. To evaluate this, we replaced araA with an insulated P_Tet-mScarlet-I cassette using unpurified GGA product as tDNA. Successful mutants were identified by their inability to grow on arabinose as the sole carbon source (Fig. 8E and Supplementary Fig. S5). The NT-CRISPR workflow was performed using 5 μl (10 fmol) of unpurified GGA product (Fig. 8A), yielding ∼5 × 10⁷ CFU/ml after CRISPR counterselection, ∼2.5-fold lower than using 50 fmol of purified plasmid (cf. Fig. 8B). This aligns with findings by Stukenberg et al. [24], which demonstrated a positive correlation between tDNA concentration and CFU count. To improve efficiency, the NT-CRISPR workflow could be optimized by increasing tDNA concentration, enhancing the efficiency of individual assembly steps, or reducing dilution during final plating (Fig. 8D). However, in most cases, obtaining just a few positive colonies is sufficient for subsequent experiments. Thus, using GGA products as direct tDNA eliminates the need for intermediate E. coli transformations, reducing metabolic burden and avoiding toxic effects associated with certain constructs or homology flanks (Fig. 5). This approach removes the necessity for PCR amplification and plasmid purification, significantly improving both time and cost efficiency in V. natriegens genome editing.

Curating the junction design of the original collection

The original 4-bp junctions in the Marburg Collection followed the Phytobricks/MoClo/Loop assembly specification, with additional junctions introduced between parts 6–7 (3′-connector to ori) and 7–8 (ori to antibiotic resistance marker ABR), allowing modular backbone assembly and multi-TU level 2 constructs (Fig. 9A) [54]. However, these junctions were selected before comprehensive ligase fidelity profiling, leading to suboptimal assembly efficiency [55]. Using the NEB Ligase Fidelity Viewer, we found that the original overhangs resulted in a predicted ligation fidelity of just 19% for an 8-part level 0 assembly and as low as 16% for a full 10-part assembly under standard GGA conditions. Even with static incubation at 37°C, the predicted fidelity improved only marginally to 36%. The primary issue stemmed from the palindromic 6–7 junction (AGCT), which promoted misassembly through self-ligation and inversion, depleting functional part concentrations. Assembly errors involving ori and ABR would likely prevent colony formation, potentially masking the inefficiencies but still reducing overall cloning success.

Figure 9.

Comparison of original and curated 4-bp fusion sites and their impact on assembly efficiency. (A) Overview of the original and revised fusion sites for level 1 and level 2 plasmids. The part 6–7 and 7–8 junctions were optimized for higher fidelity by replacing AGCT with CGAA (6–7) and TGCT with ATAA (7–8). Fusion sites are shown as 5′- to 3′-sequences of 5′-assembly overhangs. (B) Proportion [%] of phenotypically correct CFUs relative to the total CFU count for different level 1 and 2 test constructs. These include level 1 constructs expressing mScarlet-I, mCherry, and lux cassettes and a level 2 construct containing an operon for deoxyviolacein and sfGFP production. GGA and transformation were performed on at least three independent days (minimum of three technical replicates) and partially in different labs (for details, see Supplementary Table S8). Plasmid assembly plans are listed in Supplementary Table S14. Error bars indicate SD. Statistical significance was determined using an unpaired t-test: *P < .001, P < .01, and °P < .05; NS indicates no significant difference.

To resolve these issues, we optimized junction sequences using the NEB GetSet tool and Ligase Fidelity Viewer, replacing AGCT with CGAA (6–7 junction) and TGCT with ATAA (7–8 junction) (Fig. 9A). These revised overhangs significantly improved predicted ligation fidelity, reaching 100% for standard eight-part assemblies and 84% when including the 4a–4b GATG junction. While existing 4a and 4b parts remain unchanged, we recommend adopting AGGT as the 4a–4b junction, aligning with the original MoClo design [56] and providing 100% predicted fidelity (where GGT encodes glycine for in-frame fusions).

For the Vnat Collection, we cloned new level 0 and level 0* parts incorporating these revised junctions, ensuring compatibility with level 1 and 2 assemblies. To compare the original and curated junction sets, we assembled and transformed test constructs expressing mScarlet-I, mCherry, luminescence (lux) cassette, or a deoxyviolacein and sfGFP operon (plasmid assembly plans in Supplementary Table S14). Surprisingly, despite the significant predicted fidelity difference, the proportion of correctly assembled colonies was comparable between the original and curated junction sets (Fig. 9B). This suggests that while improved junctions reduce misassembly risks, in vivo selection mechanisms may still favour correct constructs during transformation.

Although the revised junctions did not lead to significant improvements in correct CFU counts for five out of six test constructs, they provide potential benefits in specific cases. The updated fusion sites, located between parts 6–7 and 7–8, connect the ori and ABR cassette, both essential for colony survival on antibiotic-containing media. While the theoretical fidelity improvements were not as pronounced in practice, this highlights the importance of experimental validation. To ensure robust and flexible plasmid construction, the Vnat Collection includes all parts with the curated junctions, offering a standardized and optimized assembly framework for V. natriegens genetic engineering.

Part management and selection software

Managing a modular part library for GGA presents challenges in construct design, part organization, and assembly planning. To streamline this process, we developed a free and accessible suite of software tools that support the Vnat Collection. These tools, stand-alone components of the CloneCoordinate DNA construction suite, are built in Google Sheets. The pre-populated registry provides detailed construct and part information for all elements in the collection, allowing users to register custom parts in their own copies. The Golden Gate Queuer facilitates assembly planning by ensuring part compatibility, dynamically filtering available options via drop-down menus linked to the Registry. Newly built level 1 assemblies can be stored and used as inputs for level 2 cloning, maintaining design consistency. Valid constructs generated by the Queuer tool are compatible with any GGA protocol. Additionally, users can integrate their designs into the full CloneCoordinate suite (manuscript in preparation), which provides support from construct planning to sequence validation. This system includes automated calculations, checklists, and data tracking for wet lab steps, enhancing reproducibility and efficiency. Custom entries from the standalone tools can be imported into CloneCoordinate at any time, ensuring seamless part management and assembly execution. Construct management tools can be viewed, and copied for individuals’ use, from a publicly available Google Drive folder [https://drive.google.com/drive/folders/1m1hSjnyTVoYwYlI7ZALlG57rUb08i8mA].

Conclusion

The Vnat Collection addresses key limitations in existing genetic toolkits for V. natriegens by expanding the range of available genetic parts and enhancing compatibility between systems such as the Marburg Collection and NT-CRISPR. These improvements streamline genetic engineering in V. natriegens, reinforcing its potential as a next-generation chassis in synthetic biology.

A major contribution of this work is the expanded repertoire of inducible promoters. By testing seven promoters previously characterized in E. coli, we identified several, P_Ttg, P_Cin+1, P_Tet, P_Tac, and P_3BB5, that exhibit strong regulation and high fold-change induction in V. natriegens. Additionally, orthogonality testing confirmed that P_Tet, P_Tac, and P_3BB5 allow independent control of multiple genes, facilitating complex genetic circuits. However, our results highlight that compounds such as naringenin and OHC14 may have broader physiological effects in V. natriegens, and certain promoters (P_PhlF and P_Sal) exhibit high basal expression, limiting their applicability.

We also introduced operon connectors, enabling the seamless assembly of operons containing up to five genes under a single promoter. This advance simplifies the construction of multi-gene pathways in V. natriegens—key for metabolic engineering applications in this chassis. Further, our improved dropout part system significantly enhances GGA efficiency, with a 300-fold increase in assembly efficiency for level 1 dropout plasmids. The use of distinct restriction enzymes for plasmid assembly and dropout replacement mitigates cloning inefficiencies and improves part exchangeability.

The development of the landing pad strategy further refines the NT-CRISPR workflow by incorporating a dual-colour fluorescent screening system for edited cells and enabling the direct use of unpurified GGA products as tDNA. These modifications accelerate the genetic engineering process, reducing experimental complexity while maintaining high efficiency. Additionally, our combination of NT-CRISPR and promoter characterization demonstrated that an araA knockout reduces the arabinose requirement for full P_BDA promoter activation by 5000-fold, a finding with potential applications in bioprocess optimization.

The Vnat Collection employs an optimized GGA approach, addressing prior inefficiencies caused by suboptimal junction sequences. By introducing curated level 0 and level 0* parts, we improved predicted assembly fidelity to 100% for standard eight-part assemblies. However, experimental validation showed only modest improvements in actual assembly efficiency, underscoring the need to complement theoretical predictions with empirical testing.

Taken together, the Vnat Collection aligns with the broader efforts to domesticate V. natriegens for biotechnological applications [6–8, 10, 11, 14, 20, 32]. Its development parallels advances made for other next-generation chassis such as Pseudomonas putida and Bacillus subtilis, which have seen similar expansions in genetic tools [57–59], regulatory systems [60, 61], and genome engineering capabilities [57, 59, 61]. With its rapid growth and metabolic versatility, V. natriegens stands as a promising candidate for marine-based synthetic biology applications. We present the Vnat Collection as a community resource, facilitating the exchange of genetic parts and protocols to enable rapid, high-throughput genome engineering. By addressing key limitations and standardizing genetic tools, this collection contributes to the broader effort of fully realizing V. natriegens as a next-generation synthetic biology chassis.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This research was supported by the Australian Government through the Australian Research Council’s Future Fellowship funding scheme (project number FT230100283), and by the Forrest Research Foundation (Perth, Australia) through a Forrest Research Foundation Scholarship. Additional support was provided by the Deutsche Forschungsgemeinschaft (DFG), Germany, through project GRK 2937 and Williams College, USA. The views expressed herein are those of the authors and are not necessarily those of the Australian Government or the Australian Research Council. Funding to pay the Open Access publication charges for this article was provided by Australian Research Council.

Data availability

Genetic part sequences, oligonucleotides, plasmid maps, and plasmid assembly details are provided in the Supplementary data. Custom R scripts used for data analysis and figure preparation were tested on R version 2021b. All data supporting the figures and computational scripts are available from the corresponding author upon reasonable request. The complete Vnat Collection will be deposited at Addgene and will be available for distribution. Construct management tools can be viewed, and copied for individuals’ use, from a publicly available Google Drive folder: https://drive.google.com/drive/folders/1m1hSjnyTVoYwYlI7ZALlG57rUb08i8mA.