Skip to main content
NCBI home page
Food Science and Biotechnology logoLink to Food Science and Biotechnology
. 2025 Feb 15;34(10):2225–2235. doi: 10.1007/s10068-025-01834-x

Enhancement of iminosugar production, 1-deoxynojirimycin and 1-deoxymannojirimycin, in recombinant Corynebacterium glutamicum

Inonge Noni Siziya 1,2,3, Hyo Jung Lim 4, Dong-Hyun Jung 3, Suhyeon Baek 5, Sanggil Lee 5,6, Myung-Ji Seo 1,2,4,✉
PMCID: PMC12064549  PMID: 40351729

Abstract

The iminosugars 1-deoxynojirimycin (1-DNJ) and 1-deoxymannojirimycin (1-DMJ) were produced by recombinant Corynebacterium glutamicum (CgTYB) bearing the 1-DNJ-producing GabT1-Yktc1-GutB1 (TYB) gene cluster from Bacillus velezensis MBLB0692. The enhanced iminosugar biosynthesis in CgTYB cultures increased both α-mannosidase inhibition (AMI) and α-glucosidase inhibition (AGI) activity. Individual cultures harboring GabT1, Yktc1, and GutB1 genes were found to be able to produce both 1-DNJ and 1-DMJ in mixed cultures of one pot analysis. This suggests that the genes necessary for epimerization and reduction are intrinsic to the cells rather than close to the 1-DNJ-producing cluster. Glucose was the preferred carbon source and a provision of 10 g/L glucose increased AGI and AMI to 69.8% and 70.1%, respectively. The 1 mM IPTG produced 55.0% AGI and 72.1% AMI, and 20% ethanolic permeabilization produced 62.8% AGI and 74.2% AMI. Batch fermentation increased iminosugar yields from 18.7 to 229.9 mg/L 1-DNJ, and 41.8 to 63.7 mg/mL 1-DMJ.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-025-01834-x.

Keywords: Iminosugar, 1-Deoxynojirimycin, 1-Deoxymannojirimycin, Corynebacterium glutamicum

Introduction

Iminosugars are low molecular weight carbohydrate mimics with an endocyclic nitrogen in place of the oxygen atom typical in furanose and pyranose saccharide structures. These sugar mimics can be obtained from chemical and biological synthesis, with a greater part of the latter being extracted from plants, fungi and microorganisms (Gao et al., 2016). As glycomimetics, they have significant biological activity, particularly as potent inhibitors of carbohydrate active enzymes (CAZymes), which supports their potential as functional foods. These properties have led to the expansion of research of sugar analogues as CAZyme modulators with varied pharmacological potential, such as 1-deoxynojirimycin (1-DNJ) inhibiting α-glucosidase in the treatment of diabetes (Gao et al., 2016), obesity and inflammation (Lee et al., 2021b), and biofilm formation (Yoo et al., 2019); 1-deoxymannojirimycin (1-DMJ) inhibiting α-mannosidases in the treatment of cancer (Lu et al., 2006) and viral infections (Balzarini, 2007); and deoxygalactonojirimycin inhibiting α-galactosidase in the treatment of Fabry disease (Asano et al., 2000). Their benefit as inhibitors of carbohydrate processing enzymes is expressed in the inhibition of enzymes like α-glucosidases and α-mannosidases which are present in the Golgi apparatus and the rough endoplasmic reticulum. However, it has been noted that an inhibitor of a plant mannosidase may not necessarily inhibit the same enzyme from a mammal, and there are differences in enzymes across species, as well (Mohan and Masanori, 1995).

Iminosugars have been isolated from species of Streptomyces and Bacillus and distinct genes within the biosynthetic pathway for 1-DNJ have been identified in both, notably the transaminase gabT1, phosphatase yktc1 and dehydrogenase gutB1, which make up the TYB gene cluster (Clark et al., 2011; Lee et al., 2021a; Seo et al., 2013). The heterologous expression of the cluster in Escherichia coli produced mannojirimycin (MJ) but it was suggested that the genes responsible for the biosynthetic progression from MJ to 1-DMJ and 1-DNJ were located elsewhere in the genome, as the TYB cluster was not flanked by possible candidates for their synthesis (Clark et al., 2011).

Microorganisms have long been modified, via metabolic and genetic engineering, for the biosynthesis of useful compounds. In addition, optimization and characterization aid in the further improvement of bacterial efficiency in biosynthesis (Quin and Schmidt-Dannert, 2014). There have been several studies on iminosugar isolation from microorganisms but few on optimization and enhancement methods to increase microbial biosynthesis, particularly from recombinant strains. Li et al. (2023) used metabolic engineering with homologous recombination to efficiently produce 1-DNJ from B. amyloliquefaciens HZ-12, in which glucose transportation, fructose-6-phosphate supply, and 1-DNJ biosynthetic cluster modification were used to increase 1-DNJ. Zhu et al. (2013) enhanced α-glucosidase inhibition (AGI) activity in B. subtilis using response surface methodology (RSM) with the factors of fermentation time, temperature, substrate concentration and pH. Culture optimization for 1-DNJ production by B. methylotrophicus K26 was performed by Lee et al. (2018) using both one-factor-at-a-time (OFAT) and RSM, with controlled variables of sucrose concentration, yeast extract concentration and culture temperature. Nguyen et al. (2021) applied random mutagenesis to B. subtilis subsp. inaquosorum to produce a mutant with increased DNJ production.

Chassis selection allows for the use of a microbial host that is more suited to the production of a desired compound. E. coli is notably the most well characterized bacteria, with a short doubling time and the ability to grow under various conditions. A wide variety of techniques and tools have been developed for genetic manipulations using E. coli, and consequently, it has a US and European market share of 34% for recombinant protein products (Pontrelli et al., 2018). However, the disadvantages of E. coli include high endotoxin levels and control of secretions. On the other hand, gram positive bacteria like Corynebacterium and Bacillus hosts do not produce endotoxins and naturally secrete proteins. Corynebacterium glutamicum has generally regarded as safe (GRAS) status and has been used in the wide-scale production of organic compounds, alcohols and amino acids (Cankar et al., 2023).

There is currently limited research on the improved production of sugar mimics from recombinant microbial cell factories. In this study, we used the OFAT method to enhance biosynthesis of iminosugars in recombinant C. glutamicum containing individual and clustered B. velezensis genes from the DNJ biosynthetic pathway.

Materials and methods

Strains, media and cultivation conditions

Corynebacterium glutamicum ATCC13032 containing the C. glutamicum/E. coli shuttle vector pXMJ19 was previously constructed with the TYB gene cluster and individual genes (Siziya et al., 2024). The In-Fusion cloning products were: CgP, with the empty pXMJ19 vector; CgTYB, with the DNJ-producing TYB cluster from B. velezensis MBLB0692; CgGabT1, with the gabT1 gene; CgYktc1, with the yktc1 gene; and CgGutB1 with the gutB1 gene. Bacterial cultures were grown under agitation (180 rpm) in Luria Bertani (LB) media (Becton, Dickinson and Co, Franklin Lakes, NJ, USA) for E. coli and brain heart infusion (BHI) media (MBcell, Seoul, South Korea) for C. glutamicum, supplemented with 34 µg/mL and 25 µg/mL chloramphenicol at 37 °C and 30 °C, respectively. Chemical induction with 0.4 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was performed at an optical density of 0.7–1.0 (OD600), measured using a UV–Vis spectrophotometer (Shimadzu, Kyoto, Japan). The TYB cells were cultured for seven days and harvested for further testing. Transformed cells with individual genes were harvested after 24 h in BHI media (180 rpm, 30 °C).

Whole cell preparation

Recombinant C. glutamicum was cultured in BHI (containing 2 g glucose, 10 g tryptone, 5 g NaCl, 17.5 g brain heart infusion and 2.5 g Na2HPO4 per liter) and CGXII minimal media (containing 42 g 3-morpholinopropanesulfonic acid (MOPS), 80 g glucose, 20 g (NH4)2SO4, 5 g urea, 1 g KH2PO4, 1 g K2HPO4, 10 mg CaCl2, 0.25 g MgSO4·7H2O, 10 mg FeSO4·7H2O, 10 mg MnSO4·H2O, 1 mg ZnSO4·7H2O, 0.2 mg CuSO4, 0.02 mg NiCl2·6H2O, 0.2 mg biotin, and 30 mg protocatechuic acid (PCA) per liter) with 25 µg/mL chloramphenicol (30°C, 150 rpm). IPTG induction was performed at an OD600 between 0.7 and 0.8 and incubated further at 30°C at 200 rpm for 7 days. For whole cell suspensions, cells were harvested at 8000 rpm for 20 min at 4 °C, and gently washed three times with 1 × PBS. Cells were then suspended in phosphate buffer (pH 7.0) for subsequent experiments. Individually, whole cell cultures of CgGabT1, CgYktc1 and CgGutB1 were grown at 30°C (200 rpm) after IPTG induction. After 24 h, the three cultures were mixed into one flask and incubated further for 7 days.

Enzyme inhibition assays

Alpha-glucosidase inhibition (AGI) was measured as described previously by Lee et al. (2018). The culture supernatant (137 μL) was added to tubes with 30 μL of 0.1 M potassium phosphate buffer (pH 7.2) and 30 μL of 10 mM p-nitrophenyl α-glucopyranoside (pNPG, Sigma-Aldrich, Burlington, MA, USA), and the reaction was started by the addition of 3 μL of 100 U/mL α-glucosidase from B. stearothermophilus (Sigma, USA). The mixture was reacted at 37°C for 10 min, and absorbance was measured at 405 nm, noting the release of p-nitrophenol (pNP).

Alpha-mannosidase inhibition (AMI) was measured as described by Evans et al. (1983) with some modifications. Reaction mixtures contained 40 µL 0.05 M trisodium citrate pH 4.5, 40 µL 2 mM p-nitrophenyl α-d-mannopyranoside (Sigma, USA), 40 µL of 5.5 µg/mL mannosidase from Canavalia ensiformis (Jack bean) (Sigma, USA) and 40 µL sample. The reaction mixture was incubated at 25 °C for 15 min. 80 µL of 0.1 M NaOH was added to quench the reaction and the mixture was read at 405 nm.

Evaluation of cell growth and enzyme inhibition

Culture supernatants of CgTYB were grown in CGXII media with and without glucose, at 30 °C for 24 h in 250 mL flasks with agitation (180 rpm). Cell growth was measured using a UV–Vis spectrophotometer (UV-1280, Shimadzu, Japan) at 600 nm. Measurements were taken during the 24 h period and supernatants from centrifugation (1580R, LaboGene, Denmark) at 8000 rpm for 20 min were analyzed for their AGI and AMI activity. A time-course profile for cell density and enzyme inhibition was constructed.

Effects of carbon source, pyridoxal phosphate, IPTG concentration, and permeabilization on AGI and AMI activities

The effect of carbon sources in the culture media was measured against AGI and AMI for 7-day cultures. Sucrose, glucose, maltose, galactose, sorbitol, lactose, soluble starch and corn starch were incorporated into the culture media at a concentration of 2 g/L to select the ideal carbon source. The ideal carbon source was then provided in the media at the initial inoculation as well as supplemented four hours after IPTG induction to 10 g/L, and AGI and AMI were determined. Additionally, supplementation of pyridoxal phosphate (PLP, final concentration 10 µM), was tested to determine its effect on iminosugar production. The effect of IPTG concentration on iminosugar production and AGI and AMI activity was determined. Cultures were grown to an OD600 between 0.7 and 0.8 and induced with IPTG to final concentrations of 0.4, 1.0, 2.5 and 6.25 mM. The inhibitory activity was measured using the 7-day culture supernatants. The effect of permeabilization on AMI and AGI was determined using 0–100% ethanol, and 0.1–0.9% polysorbate 20 (Tween20), polysorbate 80 (Tween80) and Triton X-100. Whole cells were permeabilized according to the method described by Yuan and Heinzle (2009), and resuspended in culture media for 7-day growth under previously mentioned conditions.

Batch fermentation of CgTYB

CgTYB was grown in a lab-scale 5 L fermentor (Jar-Fermentor, Kobiotech, South Korea) in CGXII media with the ideal carbon source and IPTG concentration. The seed culture was prepared overnight at 30 °C in 300 mL media at 180 rpm (25 µg/mL chloramphenicol). This was inoculated into 3 L of optimal modified CGXII media containing 10 g/L glucose and 10 g/L ammonium sulfate at an inoculation ratio of 10% (v/v) with 25 µg/mL chloramphenicol and 1 mM IPTG. Batch fermentation was conducted at 30°C (200 rpm), and aeration rates were set up at 3 volumes of air per unit volume of culture media per minute (vvm), with pH maintained at 7.0 ± 0.3. Fermentation lasted 24 h, and periodic sampling was performed.

Ion exchange (IEX) chromatography for iminosugar preparation

Cultures were prepared for iminosugar isolation as described by Siziya et al. (2024). Concentrated samples were loaded onto Amberlyst 15 (H+) cation exchange column (5 × 32 cm), washed with DI and eluted with 0.5 M NH4OH. Alkaline fractions were then loaded onto Dowex 1 × 2 (Cl−) anion exchange column and eluted with DI. The pH of the samples was adjusted to 7.0 and the eluents were freeze-dried and stored at 4 °C for subsequent analyses.

High-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD)

The biosynthesized iminosugars were analyzed using HPAEC according to Yoshihashi et al. (2010) with some modifications. The IEX products were filtered (0.22 μm polyvinylidene difluoride, PVDF) at room temperature before analysis with a DX-300 system (Dionex, Sunnyvale, CA, USA) fitted with a pulsed amperometric detector (Model PAD-2), CarboPac™ MA-1 column (0.4 × 25 cm) and electrochemical detector (ED50). Standard solutions of 1-DNJ (1 mg/mL) and 1-DMJ (0.05 mg/mL) were prepared at varying dilutions for calibration curve construction at concentrations ranging from 50 to 500 μg/mL. The mobile phase consisted of eluent A (0.2 mM NaOH) and eluent B (1 M NaOH) with a gradient program as follows: 100% Eluent A at 0–5 min, 10% at 20 min, 0% at 28 min and 100% at 35–45 min. The flow rate was set at 0.3 mL/min with a compartment temperature of 30 °C.

Liquid Chromatography Tandem Mass Spectrometry (LC/MS/MS)

The LC/MS/MS analysis was performed on lyophilized purified IEX products using a Waters UPLC system with an Acquity PDA detector and an Acquity UPLC BEH HILIC column (1.7 µM × 2.1 × 100 mm, Waters Corporation, Milford, MA, USA). The mobile phase comprised eluent A (DI with 0.1% formic acid) and eluent B (acetonitrile with 0.1% formic acid) run at 30 °C with a flow rate of 0.4 mL/min. Sample injection volume was 2 µL. The MS/MS parameters were: desolvation gas flow, 500 L/hr; cone, 20 L/hr; collision gas flow, 0.15 mL/min; desolvation temperature, 500 °C; and source temperature, 150 °C.

Statistical analysis

Data was presented as mean ± SD (standard deviation) and analyzed using Prism 5.0 program (GraphPad software, USA) and SigmaPlot 13 (Systat Software Inc., CA, USA). Statistical analyses (analysis of variance, ANOVA) were performed using both programs, and where significant mean separations were obtained, Tukey’s test and Student’s t test at p < 0.05 were used.

Results and discussion

Effect of carbon source on AGI and AMI activities

Whole cell biosynthesis uses naturally-occurring and engineered cells to produce complex molecules directly from simple substrates, offering a scalable and efficient approach for the production of biologically relevant compounds (Cankar et al., 2023). Unlike E. coli, C. glutamicum does not produce endotoxins, making it safer for pharmaceutical and food-related applications. It is more robust, tolerating higher salt and stress conditions, and thrives in aerobic environments without producing unwanted byproducts. Additionally, it efficiently utilizes a wide range of sugars and has GRAS status, simplifying regulatory approval for food and feed production (Cankar et al., 2023). The choice of carbon source in microbial cultures influences cell growth, metabolic activity and product formation. Glucose is a common carbon source due to its efficiency in supporting rapid cell growth and its role as a key metabolite in many biosynthetic pathways. Other carbon sources, such as lactose, sucrose, and maltose, can be used to modulate metabolic flux toward desired production levels (Jouhten et al., 2009) and also reduce the risk of catabolite repression from high glucose concentrations. In Fig. 1(A), the effects of carbon sources on AGI and AMI activities of CgTYB are shown. In BHI media containing 2 g/L glucose, both AGI and AMI were observed in the supernatants but compared to the control, both sucrose and glucose increased inhibition activity. Cultures in the more complex carbohydrates, corn and soluble starch, did not have any notable inhibitory activity. Inhibition was greatest in sucrose, glucose, and maltose-fed cultures, with the greatest activity exhibited in 10 g/L glucose-supplemented media with 42.0% AGI and 69.9% AMI.

Fig. 1.

Fig. 1

Open in a new tab

(A) Effect of carbon sources on AGI and AMI activities by CgTYB, induced with 0.4 mM IPTG at 30 °C in CGXII minimal media. Controls contained initial concentration of 2 g/L glucose; carbon sources were 10 g/L. (B) Time-course profile of cell density, AGI and AMI activities of CgTYB in 1 L flasks. Cultures were grown at 30 °C in CGXII minimal media with 10 g/L glucose and aliquots were obtained periodically for 120 h. Black circles, white squares and black squares indicate cell density, AMI and AGI, respectively

Figure 1(B) displays the cell density over a period of 120 h alongside AMI and AGI activities. There was an increase in inhibition with time, though AMI was greater overall. At approximately 72 h, there was a notable decrease in the cell density that briefly appeared to affect the AMI, which then continued to rise. The increasing production of inhibitors with time was similar to that of previous studies comparing DNJ concentration and cell biomass in B. amyloliquefaciens HZ-12 (Cai et al., 2017), and DNJ concentration and dry cell weight in S. lavendulae (Wu et al., 2019).

Effect of IPTG concentration on AMI and AGI activities

Whole cell biosynthesis is dependent on genetic, metabolic and environmental factors. Temperature, pH, dissolved oxygen, and carbon source availability help modulate biosynthetic efficiency and host health (Pontrelli et al., 2018). Downstream processing is impacted by the physicochemical properties of iminosugars, requiring tailored purification strategies to ensure high recovery and purity of the final product.

The expression of recombinant genes is often controlled by inducible promoters, with IPTG being a common inducer. Target gene expression is triggered by inactivating the lac repressor, and allowing transcription to proceed. IPTG concentration plays an important role in determining the level of gene expression and protein production. While higher IPTG concentrations can increase protein production, excessive levels lead to metabolic stress and reduce overall cell viability, limiting productivity (Donovan et al., 1996). Optimizing IPTG concentration balances enhanced production with bacterial health, ensuring efficient and scalable protein synthesis. The IPTG concentration was optimized for enzyme inhibition (Fig. 2(A)), and at all concentrations, AMI was greater than AGI. Inhibitory activity was highest in 1 mM IPTG-treated samples for both AMI and AGI, with significant differences (p < 0.05) present between the 1 mM IPTG-treated bacteria and concentrations of 0.4 mM, 2.5 mM and 6.25 mM. Thus, 1 mM IPTG was selected as the ideal induction concentration with respective AGI and AMI activities of 55.0% and 72.1%.

Fig. 2.

Fig. 2

Open in a new tab

Effect of (A) IPTG concentration and (B) permeabilization on AGI and AMI by CgTYB in minimal media, and effect of PLP-supplementation on (C) AMI and (D) AGI activity of culture supernatants of CgP (black circle), CgTYB (white circles), and CgTYB-PLP (black triangles). In (B) negative and positive controls indicate CgP and CgTYB with no permeabilization, respectively

Effect of permeabilization treatment on AGI and AMI activities

The sensitivity of enzymes to environmental factors and their limited reusability pose challenges in industrial applications and microbial cells often hinder substrate access due to their rigid cell walls and membranes (Chen, 2007). Permeabilization is a convenient technique to enhance substrate and product diffusion. In a previous study, catalytic activity of recombinant C. glutamicum was increased following ethanolic permeabilization (Siziya et al., 2021). Permeabilization is typically performed using chemical agents such as detergents, organic solvents, and surfactants, but while chemical permeabilization is effective, it can impact cell growth and enzyme stability, particularly during fermentation. Other techniques, such as osmotic shifts, electropermeabilization, and freezing/thawing, have also been explored, but may cause significant damage to cell structures (Jamur and Oliver, 2010). Permeabilized cells are efficient and reusable biocatalysts, reducing the need for toxic chemicals and solvents in industrial processes (Chen, 2007).

Permeabilization was able to increase the inhibitory performance of the whole cell biocatalysts, suggesting increased movement of substrates and release of products. Organic solvent treatment of different concentrations of ethanol (EtOH) however, produced different results (Fig. 2(B)). There were no significant differences between the inhibitory activities of CgTYB cells treated with the three concentrations (0.1, 0.5, 0.9%) of the three detergents, but ethanol treatments produced lower activity from 40 to 100% EtOH treatments. 20% EtOH permeabilized cells were ideal as they had the greatest inhibitory activity, but while the treatment was significantly different from the untreated control, the margin was not steep.

As nonionic surfactants, Tween 20, Tween 80, and Triton X-100 differ significantly in their chemical structure, properties, and applications (Kang et al., 2019). Tween 20 contains a lauric acid chain and is highly hydrophilic and gentle on proteins and cells, making it ideal for use in immunoassays, preventing nonspecific protein binding without significantly disrupting cell membranes (Amidzadeh et al., 2014). In contrast, Tween 80 is more lipophilic and has a longer and more unsaturated fatty acid chain (oleic acid). It is frequently used to emulsify lipids in pharmaceutical and food applications and is more effective at solubilizing fats and oils (Nielsen et al., 2016). Triton X-100 is an octylphenol ethoxylate without a fatty acid chain, and is a stronger, more lipophilic detergent, compared to the Tweens. It is used in more demanding applications such as solubilizing membrane proteins and lysing cells, where stronger disruption is needed (Jamur and Oliver, 2010). Even though detergents are effective, their permeabilized cells are not reusable due to cell lysis from excessive damage to the cell membrane (Chen, 2007).

When C. glutamicum is used as a whole-cell biocatalyst, it functions as a living factory, converting substrates into desired products, with retention and release largely depend on several factors such as production conditions, cell viability, stress, membrane permeability, and transport systems (Chen, 2007). Before purification by IEX-chromatography, the inhibitory activity of CgTYB cultures was 13.7% AGI and 33.9% AMI; and the AGI and AMI activities of CgTYB from 20% EtOH-permeabilized cultures were 62.8 and 74.2%, respectively (data not shown). These values changed after chromatographic purification, as shown in Table 1, which presents the distribution of AGI, AMI, and iminosugar content from permeabilized and non-permeabilized samples after IEX-chromatography, and MS/MS analysis of purified freeze-dried samples. From the purified samples, CgTYB-Perm had higher AGI activity while in AMI, permeabilization led to lower activity. Interestingly, iminosugars were successfully synthesized in one-pot analysis with containing the individual genes of the TYB cluster. This supports the hypothesis that the genes responsible for epimerization and reduction are intrinsic to cells rather than located near the cluster (Clark et al., 2011). Permeabilization did not improve one-pot iminosugar biosynthesis and reduced productivity instead. Optimizing the permeabilization times for the cultures, or culturing them sequentially, could help to improve their group performance. However, while sequential culturing may help to improve the process for permeabilized one-pot cells, it could increase the time and labor required for production.

Table 1.

Effect of permeabilization on distribution of iminosugar inhibitor activity and concentration

Measure CgP CgP_Perm CgTYB CgTYB_Perm OnePot OnePot_Perm
AGI (%) 0.0 ± 0.0D 0.0 ± 0.0D 33.0 ± 3.3C 54.0 ± 2.1B 62.3 ± 3.1A 39.9 ± 3.9C
AMI (%) 0.0 ± 0.0D 0.0 ± 0.0D 14.9 ± 2.7A 8.2 ± 2.5BC 10.3 ± 2.5AB 3.7 ± 0.2CD
Iminosugar (mg/L): MS/MS 0.0 ± 0.0B 0.0 ± 0.0B 1.9 ± 0.2A 0.4 ± 0.2B 0.4 ± 0.2B 0.3 ± 0.1B
Open in a new tab

CgP C. glutamicum harboring the empty pXMJ19 plasmid, Perm permeabilized, CgTYB C. glutamicum harboring TYB gene cluster, OnePot single pot cultures of recombinant C. glutamicum whole cells containing the individual genes from the TYB gene cluster, Iminosugar represents both 1-DNJ and 1-DMJ from the LC/MS/MS analysis as both are present at 164 m/z

Different uppercase superscripts indicate significant difference (p < 0.05) between groups under the measurements of AGI, AMI and iminosugar content. Data represent the means (± standard deviation) of three independent assays with triplicate samples

Effect of PLP supplementation on AGI and AMI activities

The supplementation of PLP increased the inhibitory activity of CgTYB. The pNP released was measured for both AMI and AGI activities (Fig. 2(C), (D)) activity and lower concentrations of pNP were detected in CgTYB culture supplemented with 10 µM PLP (CgTYB-PLP) compared to CgTYB and control CgP. There was a greater effect on AGI, implying greater inhibition with a nearly constant concentration of released pNP over time compared to the control. For AMI, the graph suggests a weaker inhibition pattern by CgTYB and CgTYB-PLP as pNP was released over time with a lower mean difference when compared to CgP. The AGI of CgTYB cultures before purification was 13.7% and this increased to 33.0% after IEX-chromatography. However, AMI was 33.9% before purification and reduced to 14.9% after IEX-chromatography. These changes following IEX-treatment were likely due to the concentration of 1-DNJ during the chromatographic isolation, and the loss of 1-DMJ (Table 2). As DNJ is a weakly basic alkaloid, chromatography via cation exchange resins have been used to isolate it but with low purification efficiency (Ma et al., 2024). The IEX method has been used in previous studies for 1-DNJ isolation (Clark et al., 2011), but not 1-DMJ, and it is necessary to develop a method for improving the isolation and purification efficiency of 1-DMJ as well. Freeze-dried purified samples showed a reduction in iminosugar content in CgTYB-PLP compared to CgTYB. Pyridoxal 5-phosphate contains a slightly basic pyridine ring and a slightly acidic phenolic hydroxyl group (Liang et al., 2019). The presence of PLP in varied pH conditions, as are required for ion-exchange chromatography, suggest that its presence could interfere with the purification process if it binds with the resins or influences the ionic environment (Smith et al., 1983). The transaminase GabT1 is important for the first notable step in the DNJ biosynthetic pathway, but GabT1 is a low activity enzyme that typically aggregates at relatively low concentrations and exhibits a yield that is dependent on PLP-addition following induction. Arciola et al. (2018) found that GabT1 yield increased fivefold after a 10 µM PLP addition to the media. Their wild type GabT1 from Paenibacillus polymyxa SC2 performed multiple turnover reactions of amino donor L-glutamate in the presence of exogenous PLP without the preferred amino acceptor, fructose-6-phosphate. The enzyme showed no dependence on PLP for the full reaction, but four times higher dependence in the half-reaction, supporting their hypothesized branched ping-pong mechanism for the enzyme. Previously, our research also showed a similar mechanism for the recombinant GabT1 from C. glutamicum (Siziya et al., 2024).

Table 2.

Effect of PLP supplementation on distribution of iminosugar inhibitor activity and concentration

Measure CgP CgTYB CgTYB-PLP
AGI (%) 0.0 ± 0.0C 33.0 ± 3.3B 44.1 ± 6.1A
AMI (%) 0.0 ± 0.0B 14.9 ± 2.7A 3.3 ± 1.2B
Iminosugar (mg/L): MS/MS 0.0 ± 0.0C 1.9 ± 0.2A 1.3 ± 0.0B
Open in a new tab

CgP C. glutamicum harboring the empty pXMJ19 plasmid, CgTYB C. glutamicum harboring TYB gene cluster, PLP pyridoxal phosphate supplementation, Iminosugar represents both 1-DNJ and 1-DMJ from the LC/MS/MS analysis as both are present at 164 m/z

Different uppercase superscripts indicate significant difference (p < 0.05) between groups under the measurements of AGI, AMI and iminosugar content. Data represent the means (± standard deviation) of three independent assays with triplicate samples

Time profiling of CgTYB growth and its AGI and AMI activities during batch fermentation

CgTYB cells were cultured for 24 h in a 5 L fermenter with controlled pH in optimized media (Supplementary Fig. S1). Both AGI and AMI increased with increasing cell density (Supplementary Table S1 and Fig. S1). The AGI activity of flask and fermenter cultures were 33.0% and 42.9%, respectively, and the AMI activity of the respective cultures were 14.9% and 44.0%. The iminosugar content of the CgTYB flask culture was 1.9 mg/L and batch fermentation increased this to 8.9 mg/L. AGI activity increased but was maintained steadily in the first 12 h before increasing more with time. Overall, the inhibition properties of the fermenter culture were greater, achieving inhibition levels in 24 h that required more time for flask cultures. AMI was 31.3% in the fermenter supernatant at 24 h and 32.1% from the flask supernatant at 60 h. AGI reached 19.2% after 24 h in the fermenter culture and after 72 h in the flask culture. This supports the use of CgTYB for scaled-up production with shortened incubation time for iminosugar production. Permeabilized cells were not used with batch fermentation because of the possibility of cell death or reduced viability due to the weakened cytoplasmic membrane and shear stress from the impeller.

From MS/MS analysis, batch fermentation increased the iminosugar yields. CgTYB increased both AGI and AMI activity in the large-scale fermenter, in a shorter time (24 h) than was required for flask cultures (7 days). A comparison was made between CgTYB cell cultures and mixed cultures of CgGabT1, CgYktc1, and CgGutB1 in one-pot fermentation. While previously, one pot analysis with purified enzymes of GabT1, Yktc1 and GabT1 did not produce the expected 1-DNJ iminosugar, maintaining the enzymes within their cells to encourage catalysis in a native environment was successfully shown. Presently, microbial iminosugar biosynthesis has a varied production titer range of 17–6,181 mg/L in wild type strains, 359–10,500 mg/L in mutants, and 0.2–1,633 mg/L in recombinant strains (Lim et al., 2024). Random mutagenesis has shown the greatest enhancement but as it affects the entire genome rather than the specific protein of interest, there is a lack of control and genomic instability as host cells undergo genetic alterations (Badran and Liu, 2015). Wu et al. (2019) increased the 1-DNJ titer of 90 mg/L to 296.6 mg/L in S. lavendulae via precursor feeding and metabolic inhibitors, and showed that precursor glucose could increase 1-DNJ by up to 96.4%. Cai et al. (2017) increased their DNJ production in B. amyloliquefaciens HZ-12 3.8-fold in solid-state fermentation, to 870 mg/kg. Previously, we obtained 187 mg of crude product from 1 L of CgTYB which was found to contain 1.9 mg of DMJ in the HPAEC analysis (Siziya et al., 2024). In this study, iminosugars ranged from 0.3 to 8.9 mg/L which was an improvement but still on the lower end of production compared to other strains. The titer yields can be improved through process optimization like RSM, and metabolic and genetic engineering. Zhu et al. (2013) optimized culture conditions using RSM for B. subtilis B2 to increase AGI activity by 21% and with a DNJ production content of 146 mg/L. An optimized fermentation process for 1-DNJ from Streptomyces sp. SID9135 produced a yield of 640 mg/L (Paek et al., 1997) and metabolic engineering in recombinant E. coli had a 1-DNJ titer of 273 mg/L (Rayamajhi et al., 2018). Li et al. (2023) used metabolic engineering of B. amyloliquefaciens with homologous recombination to obtain an enhanced yield of 267.4 mg/L, and optimized batch fermentation further improved the production, with a final yield of 1632.5 mg/L.

Analysis of iminosugar products

While AGI and AMI assays may provide a basis for analyzing the inhibitory activity of iminosugars, whole-cell cultures may have additional metabolites that affect the two assay methods. Analyses like HPAEC and MS/MS can provide a clearer picture as to the production of these compounds and aid in effectively and accurately taking steps to increase their biosynthesis. CgTYB exhibited both AGI and AMI activity, and MS/MS analysis showed the presence of a compound at 164 m/z (Supplementary Fig. S2), the same nominal mass of the proton adduct (+ 1) for both 1-DNJ (C6H13NO4, 163.17 g/mol) and 1-DMJ (C6H13NO4, 163.17 g/mol), even though the comparison was made with a 1-DMJ standard. The addition of a proton or loss of an electron leads to the adduct formation, a common feature in mass spectrometry (Blumer et al., 2021). AGI increased after purification and concentration, suggesting better yields due to binding of the molecules to the resin. As mentioned above, prior to IEX-chromatography, sample supernatant AGI and AMI activities in CgP were both 0%, and in CgTYB, were 13.7% and 33.9%, respectively (data not shown). As the MS/MS could not be used to isolate or identify whether the 164 m/z compound was DNJ or DMJ, we reported them collectively as DNJ/DMJ iminosugar products along with their AMI and AGI activities. It is notable that MJ dehydrate, at 162 m/z, could also be responsible in part for the AMI activity of the sample supernatants and IEX-purified concentrates.

As the purification process has been used for 1-DNJ isolation, the various eluents were tested via HPAEC for the presence of other iminosugars (Fig. 3(A)), such as 1-DMJ and MJ-dehydrates, which could be released during the washing steps, compared to 1-DNJ. A number of unidentified peaks were found and, in several eluents, 1-DMJ peaks were present to varying degrees. HPAEC analysis showed iminosugars present in IEX-purified samples (Fig. 3(B)) and the 1-DMJ peaks were close to an unknown compound at 8.8 min. Batch-fermented samples had the highest 1-DNJ content, which, when compared to one-pot samples, suggests that the one-pot cultures having high AGI was likely due to the presence of other inhibiting metabolites or interfering intermediates. Table 3 shows the range on iminosugar content for both 1-DNJ and 1-DMJ from HPAEC analysis. Production yield of 1-DNJ ranged from 12.6 to 229.9 mg/L, and 1-DMJ from 30.5 to 63.7 mg/L. Both DNJ and DMJ content were highest in the batch-fermented cultures. Interestingly, permeabilization did not lead to increased iminosugar yields, even though AGI was enhanced. This suggests other compounds, or intermediates, that contributed to the inhibitory activity. With the exception of mixed one-pot cultures, the distribution of iminosugars appeared to favor 1-DMJ production. However, it was notable that the range of improvement was incomparable to 1-DNJ, in which batch fermentation produced 229.9 mg/L compared to the 63.7 mg/L of 1-DMJ.

Fig. 3.

Fig. 3

Open in a new tab

HPAEC chromatograms of (A) iminosugar standards and (B) iminosugars produced by tested recombinant C. glutamicum samples

Table 3.

Iminosugar content from HPAEC analysis

Cultures Iminosugars
1-DNJ (mg/L) 1-DMJ (mg/L)
CgTYB 18.7 ± 2.5CD 41.8 ± 0.0B
CgTYB_Perm 12.6 ± 0.9D 36.0 ± 1.6C
OnePot 48.3 ± 2.9B 30.5 ± 4.0C
OnePot_Perm 22.7 ± 1.9C 34.5 ± 2.2C
CgTYB-PLP 19.4 ± 1.4CD 35.9 ± 1.2C
CgTYB_Batch 229.9 ± 4.4A 63.7 ± 0.8A
Open in a new tab

CgTYB C. glutamicum harboring TYB gene cluster, Perm permeabilized, OnePot single pot cultures of recombinant C. glutamicum whole cells containing the individual genes from the TYB gene cluster, PLP pyridoxal phosphate supplementation, Batch batch fermentation

Different uppercase superscripts indicate significant difference (p < 0.05) between groups under the measurements of AGI, AMI and iminosugar content. Data represent the means (± standard deviation) of three independent assays with triplicate samples

In this study, 1-DNJ and 1-DMJ production from recombinant C. glutamicum harboring clustered and individual TYB genes was determined. Initial inhibitory values of 13.7% AGI and 33.9% AMI from CgTYB cultures altered by IEX-chromatography purification to 33.0% AGI and 14.9% AMI, suggesting that the process was more suited for 1-DNJ isolation than 1-DMJ. Physical and chemical alterations influenced both the bacterial growth and production yields of the cultures, with increasing AGI and AMI with time. Enhanced yields of 229.9 mg/L 1-DNJ and 63.7 mg/L 1-DMJ from batch fermentation showed that scaling up biosynthetic production of sugar mimics for industrial applications is feasible. For future research, metabolic engineering can be used to manipulate the biosynthesis of one compound over the other.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 173 KB) (172.7KB, docx)

Author contributions

I.N.S.: Methodology, data analysis and writing original draft. H.J.L. and S.B.: Data analysis. D.-H.J. and S.L.: Reviewing and editing manuscript. M.-J.S.: Conceptualization, supervision, funding acquisition, reviewing and editing manuscript.

Funding

This work was supported by Incheon National University Research Grant in 2023.

Declarations

Conflict of interest

The authors declare that there is no conflict of interest.

Ethical approval

Not applicable.

Consent for publication

All authors read and approved the final manuscript and potential publication.

Footnotes

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Amidzadeh Z, Behbahani AB, Erfani N, Sharifzadeh S, Ranjbaran R, Moezi L, Aboualizadeh F, Okhovat MA, Alavi P, Azarpira N. Assessment of different permeabilization methods of minimizing damage to the adherent cells for detection of intracellular RNA by flow cytometry. Avicenna Journal of Medical Biotechnology. 6: 38-46 (2014) [PMC free article] [PubMed] [Google Scholar]
  2. Arciola JM, Horenstein NA. Characterization of the PLP-dependent transaminase initiating azasugar biosynthesis. Biochemical Journal. 475: 2241-2256 (2018) [DOI] [PubMed] [Google Scholar]
  3. Asano N, Ishii S, Kizu H, Ikeda K, Yasuda K, Kato A, Martin OR, Fan JQ. In vitro inhibition and intracellular enhancement of lysosomal α‐galactosidase A activity in Fabry lymphoblasts by 1‐deoxygalactonojirimycin and its derivatives. European Journal of Biochemistry. 267: 4179-4186 (2000) [DOI] [PubMed] [Google Scholar]
  4. Badran AH, Liu DR. Development of potent in vivo mutagenesis plasmids with broad mutational spectra. Nature Communications. 6: 8425 (2015) [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Balzarini J. The α (1,2)-mannosidase I inhibitor 1-deoxymannojirimycin potentiates the antiviral activity of carbohydrate-binding agents against wild-type and mutant HIV-1 strains containing glycan deletions in gp120. FEBS Letters. 581: 2060-2064 (2007) [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Blumer MR, Chang CH, Brayfindley E, Nunez JR, Colby SM, Renslow RS, Metz TO. Mass spectrometry adduct calculator. Journal of Chemical Information and Modeling. 61: 5721-5725 (2021) [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cai D, Liu M, Wei X, Li X, Wang Q, Nomura CT, Chen S. Use of Bacillus amyloliquefaciens HZ-12 for high-level production of the blood glucose lowering compound, 1-deoxynojirimycin (DNJ), and nutraceutical enriched soybeans via fermentation. Applied Biochemistry and Biotechnology. 181: 1108-1122 (2017) [DOI] [PubMed] [Google Scholar]
  8. Cankar K, Henke NA, Wendisch VF. Functional food additives/ingredients production by engineered Corynebacterium glutamicum. Systems Microbiology and Biomanufacturing. 3: 110-121 (2023) [Google Scholar]
  9. Chen RR. Permeability issues in whole-cell bioprocesses and cellular membrane engineering. Applied Microbiology and Biotechnology. 74: 730-738 (2007) [DOI] [PubMed] [Google Scholar]
  10. Clark LF, Johnson JV, Horenstein NA. Identification of a gene cluster that initiates azasugar biosynthesis in Bacillus amyloliquefaciens. ChemBioChem. 12: 2147-2150 (2011) [DOI] [PubMed] [Google Scholar]
  11. Donovan RS, Robinson CW, Glick B. Review: optimizing inducer and culture conditions for expression of foreign proteins under the control of the lac promoter. Journal of Industrial Microbiology. 16: 145-154 (1996) [DOI] [PubMed] [Google Scholar]
  12. Evans SV, Fellows LE, Bell EA. Glucosidase and trehalase inhibition by 1,5-dideoxy-1,5-imino-D-mannitol, a cyclic amino alditol from Lonchocarpus sericeus. Phytochemistry. 22: 768-770 (1983) [Google Scholar]
  13. Gao K, Zheng C, Wang T, Zhao H, Wang J, Wang Z, Zhai X, Jia Z, Chen J, Zhou Y. 1-Deoxynojirimycin: occurrence, extraction, chemistry, oral pharmacokinetics, biological activities and in silico target fishing. Molecules. 21: 1600 (2016) [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Jamur MC, Oliver C. Permeabilization of Cell Membranes. In: Oliver, C., Jamur, M. (eds) Immunocytochemical Methods and Protocols. Methods in Molecular Biology, vol 588. pp 63–66. Humana Press (2010) [DOI] [PubMed]
  15. Jouhten P, Pitkänen E, Pakula T, Saloheimo M, Penttilä M, Maaheimo H. 13C-metabolic flux ratio and novel carbon path analyses confirmed that Trichoderma reesei uses primarily the respirative pathway also on the preferred carbon source glucose. BMC Systems Biology. 3: 104 (2009) [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kang S, Lim HS, Gao Y, Kang J, Jeong HY. Evaluation of ethoxylated nonionic surfactants for solubilization of chlorinated organic phases: Effects of partitioning loss and macroemulsion formation. Journal of Contaminant Hydrology. 223: 103475 (2019) [DOI] [PubMed] [Google Scholar]
  17. Lee H, Jung D-H, Seo D-H, Chung W-H, Seo M-J. Genome analysis of 1-deoxynojirimycin (1-DNJ)-producing Bacillus velezensis K26 and distribution of Bacillus sp. harboring a 1-DNJ biosynthetic gene cluster. Genomics. 113: 647-653 (2021a) [DOI] [PubMed] [Google Scholar]
  18. Lee H, Kim W, Jang S, Seo M-J. Anti-obesity and anti-inflammatory effects of the desalted Salicornia europaea L. fermented by Bacillus velezensis K26. Food Engineering Progress. 25: 169-177 (2021b) [Google Scholar]
  19. Lee H, Shin H-H, Kim HR, Nam Y-D, Seo D-H, Seo M-J. Culture optimization strategy for 1-deoxynojirimycin-producing Bacillus methylotrophicus K26 isolated from Korean fermented soybean paste, Doenjang. Biotechnology and Bioprocess Engineering. 23: 424-431 (2018) [Google Scholar]
  20. Li X, Zhang M, Lu Y, Wu N, Ji Z, Zhan Y, Ma X, Chen J, Cai D, Chen S. Metabolic engineering of Bacillus amyloliquefaciens for efficient production of α-glucosidase inhibitor 1-deoxynojirimycin. Synthetic and Systems Biotechnology. 8: 378-385 (2023) [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Liang J, Han Q, Tan Y, Ding H, Li J. Current advances on structure-function relationship of pyridoxal 5'-phosphate-dependent enzymes. Frontiers in Molecular Biosciences. 6: 4 (2019) [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lim HJ, Siziya IN, Seo M-J. Verification and applications of 1-deoxynojirimycin biosynthesis: Recent advances and future prospects. Process Biochemistry. 141: 118-131 (2024) [Google Scholar]
  23. Lu Y, Xu Y-Y, Fan K-Y, Shen Z-H. 1-Deoxymannojirimycin, the α1,2-mannosidase inhibitor, induced cellular endoplasmic reticulum stress in human hepatocarcinoma cell 7721. Biochemical and Biophysical Research Communications. 344: 221-225 (2006) [DOI] [PubMed] [Google Scholar]
  24. Ma J, Ye Y, He R, Xiong Y, Xiao R, Wang K, Zhang Y, Wu X. Enrichment of deoxynojirimycin in mulberry using cation exchange resin: Adsorption/desorption characteristics and process optimization. Food Chemistry. 463: 141281 (2024) [DOI] [PubMed] [Google Scholar]
  25. Mohan P, Masanori B. Anti-AIDS Drug Development (1st ed.). CRC Press. (1995) [Google Scholar]
  26. Nguyen KN, Kim Y, Maibunkaew S, Park J, Nguyen MT, Oh D-B, Kwon O. Enhanced production of 1-deoxynojirimycin in Bacillus subtilis subsp. inaquosorum by random mutagenesis and culture optimization. Biotechnology and Bioprocess Engineering. 26: 265-276 (2021) [Google Scholar]
  27. Nielsen CK, Kjems J, Mygind T, Snabe T, Meyer RL. Effects of Tween 80 on growth and biofilm formation in laboratory media. Frontiers in Microbiology. 7: 1878 (2016) [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Paek N-S, Kang D-J, Choi Y-J, Lee J-J, Kim T-H, Kim K-W. Production of 1-deoxynojirimycin by Streptomyces sp. SID9135. Journal of Microbiology and Biotechnology. 7: 262-266 (1997) [Google Scholar]
  29. Pontrelli S, Chiu T-Y, Lan EI, Chen FY-H, Chang P, Liao JC. Escherichia coli as a host for metabolic engineering. Metabolic Engineering. 50: 16-46 (2018) [DOI] [PubMed] [Google Scholar]
  30. Quin MB, Schmidt-Dannert C. Designer microbes for biosynthesis. Current opinion in Biotechnology. 29: 55-61 (2014) [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Rayamajhi V, Dhakal D, Chaudhary AK, Sohng JK. Improved production of 1-deoxynojirymicin in Escherichia coli through metabolic engineering. World Journal of Microbiology and Biotechnology. 34: 77 (2018) [DOI] [PubMed] [Google Scholar]
  32. Seo M-J, Nam Y-D, Lee S-Y, Park S-L, Yi S-H, Lim S-I. Isolation of the putative biosynthetic gene cluster of 1-deoxynojirimycin by Bacillus amyloliquefaciens 140N, its production and application to the fermentation of soybean paste. Bioscience Biotechnology Biochemistry. 77(2): 398-401 (2013) [DOI] [PubMed] [Google Scholar]
  33. Siziya IN, Kim Y-S, Seo D-H. Whole cell biosynthesis of luteolin glycosides by engineered Corynebacterium glutamicum harboring the amylosucrase gene. Food and Bioproducts Processing. 127: 349-359 (2021) [Google Scholar]
  34. Siziya IN, Lim HJ, Baek S, Lee S, Seo M-J. Mannosidase-inhibiting iminosugar production by recombinant Corynebacterium glutamicum harboring the 1-deoxynojirimycin biosynthetic gene cluster. International Journal of Biological Macromolecules. 278: 134858 (2024) [DOI] [PubMed] [Google Scholar]
  35. Smith GP, Samson D, Peters TJ. A fluorimetric method for the measurement of pyridoxal and pyridoxal phosphate in human plasma and leucocytes, and its application to patients with sideroblastic marrows. Journal of Clinical Pathology. 36: 701-706 (1983) [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Wu H, Guo Y, Chen L, Chen G, Liang Z. A novel strategy to regulate 1-deoxynojirimycin production based on its biosynthetic pathway in Streptomyces lavendulae. Frontiers in Microbiology. 10: 1968 (2019) [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yoo Y, Seo D-H, Lee H, Cho E-S, Song N-E, Nam T-G, Nam Y-D, Seo M-J. Inhibitory effect of Bacillus velezensis on biofilm formation by Streptococcus mutans. Journal of Biotechnology. 298: 57-63 (2019) [DOI] [PubMed] [Google Scholar]
  38. Yoshihashi T, Do HTT, Tungtrakul P, Boonbumrung S, Yamaki K. Simple, selective, and rapid quantification of 1‐deoxynojirimycin in mulberry leaf products by high‐performance anion‐exchange chromatography with pulsed amperometric detection. Journal of Food Science. 75: C246-C250 (2010) [DOI] [PubMed] [Google Scholar]
  39. Yuan Y, Heinzle E. Permeabilization of Corynebacterium glutamicum for NAD(P)H-dependent intracellular enzyme activity measurement. Comptes Rendus Chimie. 12: 1154-1162 (2009) [Google Scholar]
  40. Zhu YP, Li XT, Teng C, Sun BG. Enhanced production of α-glucosidase inhibitor by a newly isolated strain of Bacillus subtilis B2 using response surface methodology. Food and Bioproducts Processing. 91: 264-270 (2013) [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary file1 (DOCX 173 KB) (172.7KB, docx)

Articles from Food Science and Biotechnology are provided here courtesy of Springer

RESOURCES