Engineering Success

Overview

This year, XMU-China has focused on advancing the development of chitooligosaccharides (COS) as a novel preservative and achieved the expected outcomes in this field (see our Results for more details).

To ensure our design meets the intended goals, we have made substantial efforts. On one hand, we conducted engineering on the LMT signal peptide to further improve its secretion performance. On the other hand, we verified the antimicrobial effect of degradation products catalyzed by chitosanase through multiple rounds of screening, and finally obtained the chitosanase that can efficiently degrade chitosan into COS with stronger antimicrobial activity. All relevant data have been documented, and we hope they can contribute to the iGEM community.

Part 1: Engineering LMT signal peptide for improving secretion performance

Introduction

The secretion of cellular proteins is a core process in protein biosynthesis and secretion mechanisms, holding profound significance for cellular function regulation and the biopharmaceutical field (1). Current research widely suggests that the secretion of most proteins relies on the mediation of specific signal peptides. Therefore, exploring and optimizing signal peptide sequences to enhance the secretion efficiency of target proteins, thereby increasing protein production yields, has become a key strategy for improving the economic benefits of protein biopharmaceuticals (1). Based on the efforts of our previous teams (XMU-China 2021 SALVAGE, XMU-China 2022 OMEGA, XMU-China 2023 NAIADS, XMU-China 2024 REPARO), our study is dedicated to the improvement of secretion performance of the existing signal peptide LMT (BBa_K5136066), achieving higher concentrations of secreted recombinant proteins.

Cycle 1: Identifying sequence features of LMT signal peptide

Design and Build

The secretion of most proteins is primarily achieved through the General Secretion (Sec) pathway (Figure 1A) (2). The typical sequence structure of signal peptides that guide secretion via this pathway is illustrated in Figure 1B. In this pathway, proteins first bind to SecB via the positively charged N-region of their signal peptide while SecB inhibits premature folding of the protein. Subsequently, SecA recognizes the hydrophobic H-region of the signal peptide and translocates the protein to the cell membrane, in which the SecYEG translocon acts as a channel, guiding the protein to the periplasmic space. After the signal peptidase (SPase) recognizes and cleaves at the cleavage site (AXA) in the C-region of signal peptide, the protein then is correctly folded in the periplasm (3).

Figure 1 Sec pathway-dependent secretion process and the corresponding sequence structures of signal peptide. (A) Secretion of recombinant proteins through Sec pathway. (B) The typical sequence features of signal peptide can be identified in the sequence of LMT.

The LMT signal peptide we discovered has the typical sequence features of signal peptide, which can be divided into 3 regions described above as well (Figure 1B). Thus, we turned to investigate the influence of some specific residues in this signal peptide on protein (here, superfolder green fluorescent protein (sfGFP)) secretion. Two specific variants of the N-terminal positively charged region (presumed to be the SecB binding region) and the C-terminal AXA cleavage site (SPase recognition site) were then constructed, namely LMT Δ2-4 and LMT A19V, respectively (Figure 2A). The sfGFP(BBa_K5136028) was chosen as the cargo protein to characterize the secretion performance of different variants. Gibson Assembly was performed to construct the expression module (BBa_25G6VM6I and BBa_25CWGZSX, respectively) of these variants at pSB1C3 vector in E. coli DH5α and positive transformants were selected and confirmed by colony PCR (Figure 2B) and sequencing.

Figure 2 Construction of the expression modules of LMT Δ2-4 and A19V. (A) Sequence alignments of LMT and two variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of expression modules of LMT Δ2-4 and A19V fused with sfgfp at pSB1C3 in DH5α. Target bands, 1383 bp and 1374 bp, can be observed at the position between 1500 bp and 1000 bp, respectively.

Test

Figure 3 Probe the key secA and secB genes in E. coli DH5α via colony PCR. DNA gel electrophoresis of colony PCR products to probe secA (A) and secB (B). Target bands, 2706 bp and 468 bp, can be observed at the position around 3000 bp and 500 bp, respectively. Specific primers were designed according to the genome sequence of DH5α (GCF_002899475.1).

Before characterization, we confirmed the presence of the key genes secA and secB of the Sec pathway in the chassis E. coli DH5α via colony PCR (Figure 3), which set the foundation of downstream works and tests.

Figure 4 Characterization of secretion performance of LMT and two variants. The OD₆₀₀ (A), fluorescence intensity of culture (B) and supernatant (C) was measured as time progressed. Relative fluorescence units (RFU) was defined as the fluorescence of bacterial culture subtracted the background fluorescence of growth media.

We then conducted the characterization experiments in E. coli DH5α. After inoculation, the fluorescence intensity (λ_ex = 475 nm, λ_em = 515 nm) of the culture and the supernatant after centrifugation as well as the OD₆₀₀ of each group were monitored as time progressed. Our results showed that the fluorescence intensity of the two variants, no matter for the culture or supernatant, was much lower than that of the wild-type (WT) LMT group (Figure 4B and 4C), although no significant growth defect due to the modifications was observed (Figure 4A).

Learn

Colony PCR yielded some non-specific bands, and the brightness of the secB bands were not that satisfying. Therefore, to obtain more unique secA and secB fragments in the future, the PCR conditions may need to be adjusted or the primers should be redesigned. Nevertheless, we successfully validated the presence of secA and secB genes.

Characterization experiments revealed that both the reduction of positive charges in N-region (LMT Δ2-4) and mutation on SPase cleavage site (A19V) are “lethal” to the function of LMT signal peptide, which highlights the importance of critical interactions between transportation machinery of the host and signal peptide during secretion process. Notably, the LMT Δ2-4 variant exhibited slightly higher culture and supernatant fluorescence intensity compared to A19V. Although the Δ2-4 variant deletes a significant portion of the positively charged amino acids in N-region, it only reduces, rather than completely blocks, the binding of SecB to the signal peptide. Therefore, the signal peptide may retain some degree of Sec pathway transport capability. However, the A19V variant disrupts the crucial AXA cleavage site, significantly impairing the specific recognition and cleavage by the SPase. This could even block the entire Sec secretion pathway's function, thereby greatly inhibiting the secretion of the cargo protein. Additionally, the folding of cargo sfGFP might be interfered due to these two modifications, which finally resulted in the lower fluorescence intensity of bacterial culture.

The importance of the N-terminal positively charged region and the AXA cleavage site of LMT signal peptide was validated. Based on the mechanistic understanding, we planned to design and construct a series of sequence mutants for the purpose of screening variant(s) with higher-performance on protein secretion.

Cycle 2: Enhancing the hydrophobicity of the H-region of LMT signal peptide

Design

Extensive literature reviews indicated that one of the crucial factors for signal peptides to successfully mediate protein secretion is their overall hydrophobicity, especially the hydrophobicity of the H-region core (2-4). This hydrophobicity is directly correlated with the efficiency of the signal peptide in traversing the hydrophobic biological membrane. Aiming to strengthen the interactions between the cell membrane and signal peptide to promote protein secretion, we focused on enhancing the hydrophobicity of the LMT signal peptide's H-region.

Build

We designed 4 LMT variants: G12L, M13L, C14L, and A15L, which replaced the residue in H-region with the hydrophobic leucine (L) residue (Figure 5A). Using the same method mentioned before, we constructed a series of expression modules of single-point mutants (BBa_25PJKLRW, BBa_25HA3X2Z, BBa_25IFJMYS and BBa_25A752GX) and confirmed by colony PCR (Figure 5B) and sequencing.

Figure 5 Targeted mutations in H-region of LMT signal peptide. (A) Sequence alignments of LMT and 4 variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of expression modules of LMT variants fused with sfgfp at pSB1C3 in DH5α. Target bands (1383 bp) can be observed at the position between 2000 bp and 1000 bp. (C) Each variant's fluorescence intensity of supernatant of 10 h was normalized to LMT WT, for the first characterization. (D) Fluorescence intensity of supernatant of M13L variant with LMT WT was measured as time progressed, for the second characterization.

Test

The same characterization method was utilized. For the first time of characterization, the LMT G12L and A15L variants showed lower fluorescence intensity of supernatant than wild-type LMT (Figure 5C). Given the poor reproducibility observed in C14L group, we evaluated M13L variant only for the second time of characterization with LMT as control, wondering whether the improvement was true or not. However, as time progressed, the M13L variant presented a little bit lower fluorescence intensity of supernatant than the wild-type (Figure 5D), inconsistent with the previous observation.

Learn

It is evident that not all mutations that enhance hydrophobicity will certainly improve secretion efficiency. In addition, there are other hydrophobic residues to be tested. An optimal range of hydrophobicity might exist, and exceeding this range may negatively affect cargo protein's folding, transmembrane transportation, or other crucial steps. Taking together, the strategy of enhancing the hydrophobicity of the H-region core might not be so feasible and robust for the case of LMT signal peptide. Therefore, we turned to consider the mutations in the C-region of LMT.

Cycle 3: Increasing the number of Q residues in C-region of LMT signal peptide

Design and Build

Demonstrated through high-throughput experiments, the number of glutamine residues (Q) in the C-region of most signal peptides is positively correlated with secretion efficiency (4). Therefore, we designed to increase the number of glutamine residues in the C-region of LMT and constructed four C-terminal point mutants (Figure 6A), A15Q, S16Q, A17Q, and F18Q as well as the corresponding expression modules (BBa_25C7FG1O, BBa_25CR4ZPR, BBa_25MZ2ZUA and BBa_25JOOXIU). The same cloning method was implemented and positive transformants were confirmed by colony PCR (Figure 6B) and sequencing.

Figure 6 Construction of the expression modules of the variants increasing the number of glutamine residues in the C-region. (A) Sequence alignments of LMT and four variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of four variants' expression modules fused with sfgfp at pSB1C3 in DH5α. Target bands (1383 bp) can be observed at the position between 1500 bp and 1000 bp.

Test

In characterization, none of the variants impaired the cell growth (Figure 7A). However, even though some variants (except A17Q) presented higher culture fluorescence intensity than the wild-type LMT (Figure 7B), all variants showed much lower fluorescence intensity of supernatant than LMT (Figure 7C), which means no improvements of these variants in secretion performance.

Figure 7 Characterization of secretion performance of LMT and the variants increasing the number of glutamine residues in the C-region. The OD₆₀₀ (A), fluorescence intensity of culture (B) and supernatant (C) was measured as time progressed.

Learn

The current findings indicated that increasing the number of Q residues in C-region may not effectively improve the secretion performance of LMT signal peptide. Especially, A17Q mutation disrupts the AXA cleavage site like A19V variant, thus resulting in the same “abortive” phenomenon of fluorescence intensity. With the failure experience of increasing the number of glutamine residues in the C-region, we then turned to focus on the key factor for SPase cleavage efficiency–the composition of residues in C-region (4).

Cycle 4: Increasing the number of A residues in LMT signal peptide

Design and Build

It was reported that the number of alanine (A) residues at the C-terminus of the signal peptide might be closely related to the cleavage efficiency of SPase, thus influencing secretion efficiency (4). Therefore, we hypothesized that increasing the number of C-terminal alanine residues could effectively enhance SPase cleavage efficiency of the signal peptide, resulting in improved secretion performance. We designed three C-terminal point mutants, M13A, C14A, and F18A (Figure 8A) and constructed corresponding expression modules (BBa_257PMILQ, BBa_258JKPVF and BBa_25VUQPBL) like before (Figure 8B).

Figure 8 Increase the number of alanine residues in LMT. (A) Sequence alignments of LMT and 3 variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of expression modules of LMT variants fused with sfgfp at pSB1C3 in DH5α. Target bands (1383 bp) can be observed at the position between 1500 bp and 1000 bp. (C) Fluorescence intensity of the supernatant was measured as time progressed. (D) Fluorescence intensity (at 10 h) of culture and supernatant of C14A and LMT WT were compared, for the second time of characterization. p-value: 0.0008 (***) and 0.0098 (**).

Test

As time progressed, the M13A and F18A variants showed a little bit higher fluorescence intensity of supernatant than wild-type LMT, while it can be easily observed that the C14A variant's performance was superior to other variants and even the wild-type LMT (Figure 8C). To confirm whether the improvement was true or not, we conducted the characterization experiments of LMT C14A for the second time with LMT as control. Stronger fluorescence intensity of both bacterial culture and supernatant was observed for the C14A variant when compared to the wild-type LMT signal peptide (Figure 8D), validating the effectiveness of this mutation in improving secretion performance.

Learn

The successful construction and characterization of the high-performance C14A variant confirmed the hypothesis that increasing the number of C-terminal A residues might effectively enhance the function of signal peptide to direct proteins out of the bacterial cell. The results (Figure 8C and 8D) suggested that the C14A mutation might promote protein secretion through mechanisms such as boosting SPase cleavage efficiency or optimizing the signal peptide conformation. Thus, C14A represents a promising optimization site for the LMT signal peptide with potential application values. Future studies should focus on an in-depth exploration of the SPase mechanism and the structure-activity relationship of C14A.

Conclusion

Through engineering iterations, we have identified the sequence features of LMT signal peptide and conducted several rounds of targeted mutations for the purpose of screening higher-performance variant(s) that can be used to secrete more recombinant proteins produced by the microbial cell factories. Ultimately, we successfully obtained an outstanding variant LMT C14A, indicating that increasing the number of C-terminal alanine residues can effectively promote protein secretion to some extent. In addition, this outstanding variant was further applied in secreting chitosanase for food preservation (see Proof of Concept for details), tightly linking the idea of iteration with our project.

In summary, engineering a biological system cannot be easy. We need to put our heart into it and move towards success through experiments and feedbacks step by step. Here, we have tried to demonstrate how we engineered LMT signal peptide for improving secretion performance through successive iterations of the DBTL (Design-Build-Test-Learn) cycle (Figure 9).

Figure 9 Summary of the different engineering steps completed before reaching the final goal of obtaining higher-performance variant(s) of LMT signal peptide.

Part 2: Screening for chitosanase to degrade chitosan to the products with the optimal antibacterial effect

Introduction

COS can act as a natural preservative (5), as they inhibit the growth of microorganisms on food surfaces, thereby delaying the spoilage process of aquatic products. In our experiments, we found that COS with different degrees of polymerization (DP) exhibit distinct antibacterial activities (6). Based on this feature, we aim to engineer chitosanase to ultimately obtain chitosanase that degrades chitosan to the products with the optimal antibacterial effect, which will be used in further preservation experiments.

Cycle 1: Evaluation of wild-type chitosanases

Design

To screen for high-efficiency chitosanases, we selected 5 reported endo-chitosanases (BamCsn, CsnMY002, McChoA, SaCsn46A, CsnCA) (7-10) and 1 exo-chitosanase (GlmA_Tk) (11) based on the characteristics of chitosan hydrolysis products (see our Design for details). We aimed to obtain the initial antimicrobial activity of the degradation products from each enzyme through preliminary toxicity tests, and then select one or several of these enzymes for random mutagenesis.

Build

Table 1 The part information of six wild-type chitosanases.

Name	Type	Description
BBa_25E8EHJF	coding	His tag-BamCsn
BBa_25JJYAF6	coding	His tag-CsnMY002-His tag
BBa_25JA5N69	coding	His tag-Tk Glm
BBa_25RYX5T9	coding	His tag-CsnCA WT-His tag
BBa_25WTOTOC	coding	ChoA WT
BBa_252ULD2I	coding	SaCsn46A WT

The targeted sequences BBa_25E8EHJF, BBa_25JJYAF6, BBa_25JA5N69, BBa_25RYX5T9, BBa_25WTOTOC, BBa_252ULD2I were individually inserted into pET-28a(+) to construct expression plasmids (Table 1), which were then transformed into E. coli BL21(DE3) to express each enzyme. The positive transformants were confirmed by kanamycin, colony PCR (Figure 10), and sequencing.

Figure 10 DNA gel electrophoresis of the colony PCR products of BamCsn (A), CsnMY002 (BA), GlmA_Tk (C), CsnCA (D), McChoA (E) and SaCsn46A (F) at pET-28a(+) in E. coli BL21(DE3). Target bands of BBa_25JA5N69 can be observed at the position between 2000 bp and 3000 bp, the other target bands can be observed at the position between 1000 bp and 2000 bp.

Test

The expression of chitosanases was confirmed by SDS-PAGE analysis (Figure 11). Chitosan powder was added to the lysate supernatant of the expression system, followed by incubation at 37 °C for 18 h. The resulting supernatant was collected as the degradation product. The degradation product was co-cultured with the test strain (E. coli BL21(DE3), OD₆₀₀ ~0.6) at a 1:1 ratio. The OD₆₀₀ of the co-culture system was measured at 2 h and 8 h. The antibacterial activity of the degradation products treated by the enzymes was evaluated using the relative ΔOD₆₀₀ value, calculated as (OD₆₀₀ at 8 h − OD₆₀₀ at 2 h) / OD₆₀₀ at 2 h. The empty vector pET-28a(+) without expressing any chitosanases was set as control.

Figure 11 SDS-PAGE analysis of BamCsn (A), CsnMY002 (B), GlmA_Tk (C), CsnCA (D), McChoA (E) and SaCsn46A (F). Target bands can be observed between 25 kDa and 35 kDa, while the target bands of GlmA_Tk can be observed between 80 kDa and 90 kDa.

Based on the results (Figure 12), we evaluated the antibacterial activity of the enzymatic degradation products from all six chitosanases. Among them, the enzymatic degradation products of CsnMY002, CsnCA, McChoA, SaCsn46A displayed relatively higher antibacterial activity.

Figure 12 Antibacterial activity of chitosan degradation products of wild-type BamCsn, CsnMY002, GlmA_TK, CsnCA, McChoA and SaCsn46A. Relative ΔOD₆₀₀ values were normalized to that of the empty vector pET-28a(+) group.

Learn

From the results of the first-round test, we found that the degradation products via two chitosanases (BamCsn and GlmA_Tk) showed no antimicrobial effect. In the GlmA_Tk group, the growth of the test bacteria was even stronger than that in the control group. We speculated that this was because the exo-chitosanase property of GlmA_Tk produces a large amount of glucosamine, which promotes the growth of the bacterial strain instead of inhibition. Therefore, we subjected SaCsn46A, CsnCA, CsnMY002 and McChoA, which showed good performance in the first-round test to error-prone PCR (EP-PCR) to obtain a large library of mutants, followed by a second-round screening.

Meanwhile, we optimized some conditions for our toxicity test: we adjusted the OD₆₀₀ value of the bacterial strain to 0.8 before incubating with the degradation products, and fixed the detection time points at 1.5 h and 8 h. These adjustments ensured the standardized operation of the experiment and avoided random errors.

Cycle 2: Random mutagenesis of chitosanase and large-scale preliminary screening

Design and Build

After the first-round test, we selected SaCsn46A, CsnCA, CsnMY002, and McChoA for EP-PCR followed by screening (see our Experiments for more details), aiming to accumulate extensive data on the toxicity of enzymatic degradation products. During this stage, we generated a large library of mutants, and the relevant agarose gel electrophoresis results can be found in our Results section.

Test

The resulting colonies from EP-PCR were all picked and the antibacterial assay was performed again with some adjusted conditions as described in Cycle 1.

Learn

The results showed that the degradation products of some mutants exhibited strong antibacterial activity (Figure 13), which was consistent with our expectations. We selected five mutants, namely SaCsn46A mutant 2 (SaCsn46A E151G D222G), SaCsn46A mutant 14 (SaCsn46A N29S K50R D150G D189G F247L), SaCsn46A mutant 16 (SaCsn46A V52A D134G K165I F175L S190C G210D K250E N252D), SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R), and SaCsn46A mutant 94 (SaCsn46A T249A L259P) (framed in Figure 13A). When their degradation products were mixed with test bacterial culture, the ratio of the system was significantly lower than that of the wild-type SaCsn46A enzyme, showing the optimal antibacterial effect (Figure 13A). Therefore, we subjected these five mutants to secondary screening.

Figure 13 Antibacterial activity of enzymatic degradation products from wild-type (WT) and mutant chitosanases. (A) SaCsn46A. (B) McChoA. (C) CsnCA. BL: sterile water, EV: empty vector pET-28a(+). Relative ΔOD₆₀₀ values were normalized to that of the empty vector pET-28a(+) group.

Cycle 3: Secondary screening of chitosanases

Design and Build

The mutation sites of every mutant we selected were confirmed via sequencing: SaCsn46A mutant 2 (E151G D222G), SaCsn46A mutant 14 (N29S K50R D150G D189G F247L), SaCsn46A mutant 16 (V52A D134G K165I F175L S190C G210D K250E N252D), SaCsn46A mutant 60 (D40G Q125R H234R), and SaCsn46A mutant 94 (T249A L259P). In order to check the antibacterial activity as well as functional change of these mutants, we decided to obtain the purified mutants.

Test

We then purified these enzyme mutants through an AKTA pure™ chromatography system. After normalizing the protein concentration, we performed chitosan degradation experiments and toxicity tests following the same procedure as described above.

Figure 14 Antibacterial activity of enzymatic degradation products of SaCsn46A WT and selected mutants. Relative ΔOD₆₀₀ values were normalized to that of the empty vector pET-28a(+) group.

The relative ΔOD₆₀₀ values of SaCsn46A mutant 2 (E151G D222G) and SaCsn46A mutant 60 (D40G Q125R H234R) were lower than other mutants as well as the wild-type SaCsn46A (Figure 14), indicating their enzymatic degradation products possessing superior antibacterial activity thus promising to further food preservation.

Learn

Based on the results of all rounds of screening, the chitosan degradation products through SaCsn46A mutant 2 (E151G D222G) and SaCsn46A mutant 60 (D40G Q125R H234R) exhibited the optimal activity. We therefore selected these two mutants for more detailed enzymatic characterizations.

Cycle 4: Enzymatic characterizations of outstanding mutants

Design and Build

After identifying SaCsn46A mutant 2 (E151G D222G) and SaCsn46A mutant 60 (D40G Q125R H234R) as the most effective chitosanases in our library, we aimed to obtain more detailed enzymatic parameters to guide subsequent degradation experiments and efficiently produce COS with strong antibacterial activity.

We chose to test the activity of these mutants under different temperature and pH conditions to determine their optimal degradation temperature and pH, as well as their stability profiles.

Test

Chitosanase activity was measured via the 3,5-dinitrosalicylic acid (DNS) method. Using D-glucose as the standard, reducing sugar content was determined by measuring the optical density (7) at 540 nm, with the reaction mixture incubated at 100 °C for 5 min. One unit (U) of activity was defined as the enzyme amount needed to release 1 mmol of D-glucose-equivalent reducing sugars per minute under these conditions.

Figure 15 Enzymatic properties of SaCsn46A mutant 2 and 60. (A) Effects of temperature on enzyme activity. (B) Effects of pH on enzyme activity. (C) Enzyme inactivation at different temperatures. (D) pH stability. The 100% relative activity of mutant 2 was 2.385 U/mg and the 100% relative activity of mutant 60 was 2.124 U/mg.

For pH profiling, assays were conducted at 45 °C across pH 3.0~9.0 using 0.05 M acetate (pH 3.0~6.0), phosphate (pH 6.0~8.0), or Tris-HCl (pH 9.0) buffers. pH stability was assessed by pre-incubating enzymes in ice bath (pH 3.0~9.0), then measuring residual activity under optimal conditions (mutant 2: pH 5, 30 min; mutant 60: pH 4, 30 min). Temperature profiling used different temperatures, while thermostability was tested by pre-incubating at 0~60 °C for 2 h, followed by measuring the residual activity at 45 °C (see our Experiments for more details).

Additionally, we analyzed the degradation products of both mutants using HPLC and TLC as well, and the relevant TLC and HPLC results can be found in our Results section.

Learn

SaCsn46A mutant 60 and mutant 2 presented different optimal reaction conditions and stabilities. Mutant 60 has an optimal temperature of approximately 40 °C (slight activity decreases but high retention at 45 °C) and an optimal pH condition of 4.0. While Mutant 2 showed unique temperature characteristics—activity drops sharply at 40 °C but recovers at 45 °C (similar to the activity at 35 °C)—with an optimal pH condition of 5.0. Both maintained good thermostability below 50 °C (minor activity increases around 40 °C), while mutant 2 is stable at pH 6~8 and mutant 60 at pH 5 (Figure 15).

Conclusion

To efficiently obtain COS with strong antibacterial activity for aquatic products preservation, XMU-China implemented a multi-cycle strategy for chitosanase screening and engineering. Initial evaluation of 6 wild-type chitosanases identified 4 (SaCsn46A, CsnCA, CsnMY002, McChoA) with the degradation products showed effective antibacterial activity, excluding BamCsn (weak activity) and GlmA_Tk (exo-chitosanase property promotes bacterial growth). After screening from the large library generated via EP-PCR, SaCsn46A mutant 2 (E151G D222G) and mutant 60 (D40G Q125R H234R) were selected as optimal candidates—both have good thermostability below 50 °C, distinct optimal pH/temperature, and chitosan degradation products that efficiently inhibit bacterial growth. These two outstanding mutants were further employed for food preservation experiments with the LMT C14A variant (see our Proof of Concept for more details).

Engineering a biological system requires iterative efforts. This study demonstrated the development of high-performance COS-producing chitosanases through successive DBTL (Design-Build-Test-Learn) cycles (Figure 16). Each step was guided by experimental feedback, ultimately forming a reliable technical route for low-cost, high-efficiency antimicrobial COS production—critical for advancing natural preservatives in the aquatic product industry.

Figure 16 Summary of the different engineering steps in screening for chitosanase to degrade chitosan to the products with the optimal antibacterial effect.

References

F.R. Schmidt, Recombinant expression systems in the pharmaceutical industry, Appl Microbiol Biotechnol 65, 363-372 (2004).
S. Kaushik, H. He, R.E. Dalbey, Bacterial Signal Peptides- Navigating the Journey of Proteins, Front Physiol 13, 933153 (2022).
H. Owji, N. Nezafat, M. Negahdaripour, A. Hajiebrahimi, Y. Ghasemi, A comprehensive review of signal peptides: Structure, roles, and applications, Eur J Cell Biol 97, 422-441 (2018).
S. Grasso, V. Dabene, M.M.W.B. Hendriks, P. Zwartjens, R. Pellaux, M. Held, S. Panke, J.M. van Dijl, A. Meyer, T. van Rij, Signal Peptide Efficiency: From High-Throughput Data to Prediction and Explanation, ACS Synth Biol 12, 390-404 (2023).
J. Wang, H. Chen, N. Wang, X. Pan, B. Xia, K. Xu, B. Zhong, Chitosan Oligosaccharides: A Natural Rich, High-Efficiency, and Safe Frontrunner for the Futural Fruits/Vegetables Preservation, Food Bioprocess Technol 18, 1104-1124 (2025).
L. Zhang, Y. Wang, S. Ling, M. Yuan, Q. Sun, X. Dong, Antibacterial mechanism of chitooligosaccharides against specific spoilage organisms in chilled processed fish paste products, Food Control 172, 111148 (2025).
B. Bhuvanachandra, D. Sivaramakrishna, S. Alim, G. Preethiba, S. Rambabu, M.J. Swamy, A.R. Podile, New Class of Chitosanase from Bacillus amyloliquefaciens for the Generation of Chitooligosaccharides, J Agric Food Chem 69, 78-87 (2021).
Y. Li, Y. Gou, Z. Liu, T. Xie, G. Wang, Structure-based rational design of chitosanase CsnMY002 for high yields of chitobiose, Colloids Surf B Biointerfaces 202, 111692 (2021).
J. Guo, Y. Wang, W. Gao, X. Wang, X. Gao, Z. Man, Z. Cai, Q. Qing, Gene Cloning, Functional Expression, and Characterization of a Novel GH46 Chitosanase from Streptomyces avermitilis (SaCsn46A), Appl Biochem Biotechnol 194, 813-826 (2022).
H. Su, J. Sun, C. Guo, Y. Wang, F. Secundo, H. Dong, X. Mao, Structure-based mining of a chitosanase with distinctive degradation mode and product specificity, Appl Microbiol Biotechnol 107, 6859-6871 (2023).
W.-S. Xia, D.-X. Lee, Purification and characterization of exo-β-d-glucosaminidase from commercial lipase, Carbohydr Polym 74, 544-551 (2008).