Round 1-Random Mutation

1.Screening of mutant strains

Firstly, the atmospheric and room-temperature plasma mutation system (ARTP) was set out to screen mutants with higher CLPs yield. After ARTP treatment, we obtained 200 mutant strains. The anti-microbial activity of these strains was assessed by oxford cup method. We determined the size of the inhibition zone of these mutant strains against the plant pathogen Xanthomonas spp. QKHT-5 to characterize the strength of their inhibitory activity. After multiple rounds of screening, we obtained a series of mutant strains with stronger antibacterial activity than HMBY. Among them, the mutant strain HMBY-106 achieved the largest antibacterial diameter. As shown in Figure1, the inhibitory diameter of HMBY and HMBY-106 is (12±1) mm and (16±1) mm respectively, which suggest a significantly stronger antibacterial activity of HMBY-106 than HMBY.

We further validate the antibacterial activity of HMBY-106 by the method of minimum inhibition concentration. The minimum inhibitory dilution of HMBY and HMBY-106 is 25 and 26 respectively (Figure 2), which suggest a stronger antibacterial activity of HMBY-106.

2.Analysis of CLPs

To explore the anti-microbial metabolites, we searched for the secondary metabolite biosynthetic gene clusters across the genome. A complete set of NRPS modules for synthesizing iturin, surfactin, and fengycin (Figure 1b) was found in the genome of HMBY. In addition, the structure of CLPs was determined by LC-MS using the fermentation product of HMBY. Thirteen CLPs were revealed, comprising three iturin, five fengycin and five surfactin (Table 1).

To investigate if HMBY-106 synthesizes more lipopeptides than HMBY, we quantified the amount of the CLPs in HMBY and HMBY-106. HPLC detection of the fermentation product revealed the production of iturin, fengycin and surfactin by B. velezensis HBMY. In addition, HMBY-106 had around 2 times higher CLPs production compared to HMBY.

3.Determination of lipopeptide gene expression

To determine the regulation of CLPs biosynthesis genes expression in high-yield CLPs-producing mutant, qPCR was performed to assess the CLPs genes expression of HMBY and HMBY-106 (Figure 5). The results showed that the transcription levels of the NRPS genes for iturin A-D, surfactin AA-AD, and fengycin A-E in HMBY-106 are all higher than their levels in HMBY.

4.Cell growth viability analyses

Further biomass and cell growth viability analyses revealed the gene mutations on strain performance. We compared the growth curves of HMBY and mutants to assess the effect of gene mutations on bacterial growth (Figure 6). After their growth reached the plateau (stationary phase), the cell densities decreased quickly. HMBY and HMBY-106 exhibited similar growth rate, but the HMBY-106 declined more quickly after the plateau than HMBY. This indicate that the higher yield of CLPs result from HMBY-106 was not result of the bacterial density.

Round2-Targeted gene engineering – uncovering regulatory genes

1.Comparative genomics analysis of HMBY and HMBY-106

To investigate the specific genes that are responsible for the higher-yield CLPs of HMBY-106 than the wild-type strain, we sequenced and analyzed the whole genomes of HMBY and HMBY-106. Three genes (cdaA, relA and cheB) were found to acquire missense mutations in their protein-coding regions (Figure 1).

2.Targeted gene mutation of HMBY and CLPs yield

To deterimine whether all these gene mutations regulate CLPs synthesis, the CRISPR-Cas9 system gene editing vector pJOE8999 was employed to constructed the targeted mutant strains. The genes cdaA, rel and cheB in HBMY were respectively substituted by the mutant genes from HBMY-106. As shown in Figure 2, the amount of CLPs produced by the mutants of Rel and CdaA were higher than that of wide-type strain HMBY. CLPs yield of engineered Rel and CdaA mutant strain was increased compared to the wild-type strain. The mutation of CheB dramaticaly decrease CLPs production of HMBY. More importantly, we obtained the strain HMBY-rel, which is more productive than HMBY-106.

3.Cell growth viability analyses

In addition, the cell growth of the mutant strains and HMBY were assessed. The strains HMBY, HMBY-106 and HMBY-cheB exhibited similar growth rate, but the HMBY-106 and HMBY-cheB declined more quickly after the plateau than HMBY. The mutant strain HMBY-rel and HMBY-cdaA exihibited a lower growth rate than HMBY. This indicate that the higher yield of CLPs result from HMBY-rel and HMBY-cdaA was not due to the bacterial density.

4.Analysis of antibacterial activity

The anti-microbial activity of these strains was assessed by oxford cup method. We determined the size of the inhibition zone of these mutant strains against the plant pathogen Xanthomonas spp. QKHT-5 to characterize the strength of their inhibitory activity. After multiple rounds of screening, we obtained a series of mutant strains with stronger antibacterial activity than HMBY. Among them, the mutant strain HMBY-relA achieved the largest antibacterial diameter. As shown in Figure 4, the inhibitory diameter of HMBY-106 and HMBY-relA is (16.6±2.1) mm and (21.6±0.3) mm respectively, which suggest a significantly stronger antibacterial activity of HMBY-relA than HMBY-106.

5.Determination of lipopeptide gene expression

To determine the regulation of CLPs biosynthesis genes expression in high-yield CLPs-producing mutant, qPCR was performed to assess the CLPs genes expression of HMBY，HMBY-106，HMBY-relA, HMBY-CdaA and HMBY-CheB (Figure 5). The results showed that the transcription levels of the NRPS genes for iturin, surfactin , and fengycin in HMBY-relA are all higher than their levels in HMBY-106.

6.A regulatory model of CLPs biosynthesis

We proposed a new regulatory model that the mutations of relA and cdaA regulate the CLPs production. Rel synthesizes (p)ppGpp, a component that reduces cellular GTP levels, and upregulates amino acid biosynthesis. The mutation of rel may up-regulate the amount of amino acids. Alternatively, the NRPS genes may be up-regulated by the mutations of rel or cdaA. The increase supply of substrate amino acids and the up-regulated NRPS genes expression improve the CLPs biosynthesis.

Round 3-Metabolic Engineering

1. Construction of dual-gRNA and single-gRNA CRISPR-Cas systems for competing pathway knockout

To redirect the CLPs biosynthesis flux exclusively towards fengycin biosynthesis, we aimed to knockout the long gene clusters fragment responsible for the synthesis of competing lipopeptides surfactin (srfA-D) and iturin (ituA-D). Based on the pJOE8999 plasmid, we constructed a dual-gRNA CRISPR-Cas system, designated as pWLS. This vector was designed to express two guide RNAs simultaneously, targeting two specific sites, significantly improving the editing efficiency for long fragment. Additionally, a single-gRNA knockout vector, pJOE8999-ΔsrfB, was also constructed for comparative analysis.

The recombinant plasmids, pWLS-Δsrf, pWLS-Δitu, and pJOE8999-Δsrf, were successfully constructed via a one-step multi-fragment assembly strategy. Each plasmid contained the corresponding gRNA expression cassettes and the respective upstream and downstream homology arms (each approximately 1500bp) for the targeted gene cluster, serving as templates for homologous recombination after Cas9-induced double-strand breaks.

2. Transformation, Comparative Efficiency Analysis, and Mutant Verification

The constructed editing vectors (pWLS-Δsrf, pWLS-Δitu, and pJOE8999-ΔsrfB) were electroporated into the high-yielding strain HMBY-rel. Transformants were selected on LB agar supplemented with 5 µg/mL kanamycin and 0.2% D-mannose (to induce Cas9 expression) after incubation at 30°C for 2 days.

A comparative analysis revealed significant differences in performance between the single- and dual-gRNA systems. Following D-mannose induction, the viability of B. velezensis transformants carrying the single-gRNA vector was markedly lower than those with the dual-gRNA system, as clearly observed on solid medium plates.

Colony PCR verification further demonstrated the superior efficiency and accuracy of the dual-gRNA system. While the single-gRNA transformants showed messy, incorrect, or absent bands, the dual-gRNA system produced clear, correctly sized bands corresponding to the expected deletions.The editing vectors pWLS-Δsrf and pWLS-Δitu were sequentially introduced into HMBY-rel. Mutants from each transformation round were selected and screened via colony PCR. DNA sequencing confirmed the successful knockout of both the srf and itu gene clusters, yielding the final engineered strain HMBY-rel-ΔsrfΔitu. This result validates the high efficiency of our dual-gRNA system for long genomic fragment knockout.

3. Plasmid Curing and Ongoing Functional Analyses

For plasmid curing, transformants were replica-picked onto antibiotic-free LB plates, incubated overnight at 50°C to promote plasmid loss, and then streaked at 42°C to isolate single colonies. Successful plasmid curing was confirmed by the absence of growth on kanamycin-containing plates alongside normal growth on antibiotic-free LB plates.

Subsequent phenotypic and functional analyses are currently underway, including HPLC-MS/MS for lipopeptide quantification, gene expression profiling, growth curve monitoring, and antibacterial activity assays. These studies aim to comprehensively evaluate the effects of the srf and itu knockouts on fengycin production and overall strain performance.

Round 4-NRPS communication (COM) domain reprogramming

Our software platform offers a comprehensive computational solution for reprogramming NRPS communication (COM) domain. It integrates multiple bioinformatic tools—including MAFFT for sequence alignment, IQ-TREE for phylogenetic analysis, and InterProScan for domain annotation—into a unified and automated workflow. Key functionalities include fen sequence integration, functional domain mapping, and automated extraction of COM junction regions. Through multi-dimensional data analysis, the platform successfully identified four critical residue sites, providing precise targets for experimental validation. The predicted fen COM reprogramming will further be validated by our wet-lab results.

Designed to serve the iGEM community, this software provides a rational design pipeline from gene sequences to functional predictions. By incorporating structural modeling and machine learning strategies, it supports the reprogramming of NRPS machinery, helping teams efficiently develop high-performance microbial strains and strengthening the computational foundation of synthetic biology projects.