Draft genome and physiological characterization of a newly isolated l-arabinose-utilizing Corynebacterium glutamicum CS176

Abstract

This study reports the physiological and genomic characterization of Corynebacterium glutamicum CS176, a newly isolated strain exhibiting a rare combination of traits: efficient l-arabinose utilization and l-glutamate production without chemical induction, even under excess biotin conditions. Genome analysis of the 3.10 Mb draft sequence (54.20% GC) revealed a 7.8 kb l-arabinose utilization gene cluster highly similar to those of arabinose-assimilating strains, suggesting acquisition via horizontal gene transfer. To link genotype with phenotype, the effects of temperature, oxygen availability, carbon sources, and biotin concentration on bacterial growth and l-glutamate production were systematically evaluated under controlled conditions. Optimal growth was observed at 30–37 °C under high oxygen conditions, whereas maximal l-glutamate production (up to 2.5 g/L in mixed substrates) occurred at 37–39.5 °C under medium–low oxygen conditions. Notably, CS176 maintained l-glutamate production across a wide range of biotin concentrations (0–200 µg/L), in contrast to the typical biotin-dependent regulation observed in C. glutamicum. Furthermore, glucose–arabinose co-utilization enhanced both growth and l-glutamate production, highlighting metabolic flexibility. Together, these findings identify CS176 as a promising strain that overcomes key limitations in conventional glutamate fermentation and provides insights for metabolic engineering and sustainable bioprocess development.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-54441-z.

Introduction

Corynebacterium glutamicum (C. glutamicum) was originally isolated as l-glutamate producer which can excrete large amounts of l-glutamate under proper conditions^1,2. This bacterium is Gram positive rod, non-spore forming, non-pathogenic and biotin auxotroph. Nowadays, C. glutamicum is widely used for industrial production of amino acids such as l-glutamate and l-lysine for seven million tons per year^3,4. The growth and productivity of C. glutamicum are significantly influenced by several environmental factors, including oxygen levels, temperature, and biotin concentration. Oxygen availability plays a crucial role in the metabolic activities of this microorganism, with aerobic conditions favoring robust growth and controlled oxygen limitation triggering glutamate production. Temperature, another critical parameter, plays a pivotal role in the enzyme activities and overall metabolic processes of C. glutamicum. Additionally, biotin, an essential cofactor for various carboxylase enzymes, impacts both growth and metabolite production. Controlled biotin limitation is a well-established strategy to enhance glutamate yield, as it induces stress responses that lead to glutamate accumulation and involves in fatty acid biosynthesis, and the composition of cell envelope that led to enhance l-glutamate excretion^2,5–8. Besides the biotin limitation, l-glutamate excretion can be induced by addition of antibiotics (penicillin, ciprofloxacin, ethambutol) or surfactants^9–13, temperature shift^14–19, or using lysozyme sensitive strain^20,21. Instead of l-glutamate and l-lysine, C. glutamicum has been used for industrial biotechnology²² to produce other amino acids like l-alanine, l-aspartate, l-isoleucine, l-leucine, l-threonine, and l-valine^23–30; organic acids and bioplastic monomer like lactic acid and succinic acid^31–34; alcohols like ethanol, isobutanol, 3-methyl-1-butanol, 2,3-butanediol^35–39; other value-added compounds like astaxanthin, carotenoids, diamine putrescine^28,40–43; and for environmental remediation^44–46.

Currently, lignocellulosic materials from agricultural waste become renewable carbon sources without a food competing material for biotechnological fermentation. They consist of cellulose, a homopolymer of glucose (40–50%), hemicellulose, a heteropolymer of hexose and pentose sugars (25–35%) and lignin (10–20%)^47,48. The lignocellulosic hydrolysates generate growth-inhibitory substances such as furfural, phenolic compounds, and weak acids (acetic, formic, or levulinic acid). l-arabinose is one type of pentose sugar derived from plant biomass and agricultural waste, particularly from hemicelluloses. The ability to utilize l-arabinose is significant in the context of converting renewable biomass into biofuels, chemicals, and other bioproducts. While many microorganisms can metabolize hexoses, the ability to catabolize pentoses like l-arabinose is rarer. The discovery of a strain that can metabolize l-arabinose opens up new avenues for bio-based production from agricultural waste. Interestingly, C. glutamicum can utilize a variety of hexose sugars, alcohols, and organic acids as sole or mixed carbon and energy sources for growth and bioproduction. Moreover, it can resist and even utilize some of inhibitory substances as carbon sources^49,50. However, for pentose sugars, no strain can utilize D-xylose as a carbon source⁵¹ and only one strain of C. glutamicum, the ATCC31831, has been reported as native l-arabinose utilizing strain⁵².

In this study, we investigated C. glutamicum CS176, a newly isolated l-arabinose-utilizing strain obtained from chicken feces-contaminated soil, which exhibits a rare and industrially relevant phenotype: the ability to utilize l-arabinose while maintaining l-glutamate production without chemical induction, even under excess biotin conditions. To elucidate the basis of this phenotype, we systematically characterized its growth and glutamate production under varying temperatures, oxygen levels, and biotin concentrations, and evaluated its performance on glucose, l-arabinose, and their mixtures. In parallel, draft genome sequencing was performed to identify genetic determinants associated with arabinose metabolism, including a comparative analysis of the arabinose utilization gene cluster with those of other C. glutamicum strains and related Corynebacterium species. By integrating physiological and genomic analyses, this study aims to provide insights into the metabolic flexibility and regulatory characteristics of CS176, which may help expand substrate utilization strategies and overcome current limitations in glutamate fermentation for industrial applications.

Materials and methods

Strain origin, isolation, and identification

Strain CS176 was originally isolated from environmental samples collected at Suwanvajokkasikit Farm, Kasetsart University, Bangkok, Thailand. Permission for sample collection was obtained from the farm management prior to sampling. Samples consisted of soil and naturally deposited manure collected without direct interaction with animals or disturbance to farm operations.

Bacterial isolation, soil sample was ground and added to D2 broth medium containing 10 µg/mL nalidixic acid and 2 µg/mL NaN₃, followed by overnight cultivation for enrichment. The overnight culture was diluted in sterile water using the standard serial dilution technique and plated on D2 agar medium containing 10 µg/mL nalidixic acid and 2 µg/mL NaN₃. Convex yellow colonies with entire margins, characteristic of the Corynebacterium genus, were selected and maintained on A1 agar medium (per liter: 10 g peptone, 10 g yeast extract, 5 g glucose, 1 g NaCl, and 15 g agar) for further study.

Genomic DNA of strain CS176 was extracted following the method of Green and Sambrook⁵³. The 16S rRNA gene was amplified by PCR using primers 27F and 1492R, and sequencing was performed using an ABI Prism 3730XL automated sequencer (BioBasic Inc., Canada). Sequence reads were assembled using CAP contig assembly in the BioEdit program⁵⁴.

Sequence similarity was calculated based on pairwise alignment using the EzBioCloud platform (https://www.ezbiocloud.net/)⁵⁵. Phylogenetic analysis was conducted using the Neighbor-Joining (NJ) method^56,57 in MEGA version 11⁵⁸ with 1,000 bootstrap replicates. Multiple sequence alignment was performed using Clustal W, and Mycobacterium smegmatis was used as an outgroup. Bootstrap values greater than 50% were indicated at branch nodes. The 16S rRNA gene sequence was deposited in the NCBI database under accession number MN826455.1

Physiological characterization

Strain CS176 was cultivated to evaluate growth conditions and L-glutamate production. For inoculum preparation, the strain was grown in AR medium (per litre: 10 g glucose, 2 g urea, 1 g yeast extract, 1 g casamino acid, 0.7 g (NH₄)₂SO₄, 0.5 g K₂HPO₄, 0.5 g KH₂PO₄, 0.5 g MgSO₄.7H₂O, 10 mg FeSO₄.7H₂O, 1 mg MnSO₄.7H₂O, 0.2 mg biotin, and 0.2 mg thiamine-HCl). A 1% of inoculum was transferred into minimal glucose medium (per litre: 10 g glucose, 3 g NH₄Cl, 2 g urea, 3 g K₂HPO₄, 1 g KH₂PO₄, 0.5 g casamino acid, 0.5 g MgSO₄.7H₂O, 10 mg FeSO₄.7H₂O, 1 mg MnSO₄.7H₂O, 20 mg cysteine, 1 mg thiamine-HCl, 0.03 mg biotin, and 1 ml metal mixture). Cultures were incubated in shake flasks at 200 rpm for 36–70 h under varying temperatures (30–40.5 °C) and oxygen conditions, adjusted by flask type, closure, and working volume.

Additionally, growth and l-glutamate production were evaluated using glucose, l-arabinose, and mixed carbon sources. The effect of biotin concentration (0–200 µg/L) on growth and l-glutamate production was also investigated.

Analytical methods

Bacterial growth was monitored using a Klett-Summerson photoelectric colorimeter or by measuring optical density at 600 nm (OD₆₀₀). Sugar consumption was analyzed by HPLC (Shimadzu, Japan) equipped with a refractive index detector. l-Glutamate production was quantified using a commercial assay kit (Yamasa Neo, Cosmo Bio, Japan) after cell removal.

Statistical analysis

All statistical analyses were performed using Python (v3.8.10). Two-way ANOVA was used to evaluate the effects of temperature, oxygen levels, carbon sources, and biotin concentration on growth and glutamate production. Tukey’s HSD test was applied for multiple comparisons.

Where significant interactions were observed, simple effects analysis was conducted. Homogeneity of variances was assessed using Levene’s test (p > 0.05). One-way ANOVA was used for independent factor analysis when appropriate. Statistical significance was defined at p < 0.05.

Genome analysis

Sequencing of genomic DNA of CS176 was performed at Hokkaido system science Co. Ltd. Assembly and annotation were performed at Department of Biological Chemistry, Faculty of Agriculture, Yamaguchi University.

The draft genome of C. glutamicum CS176 was sequenced using the Illumina Hiseq 2500 sequencing platform by the paired-end strategy (2 × 100 bp), which generated 10,013,404 reads pair. Adapter sequence was removed by using CutAdapt. The final coverage reached 257-fold for an estimated genome size of 3.10 Mb. The resulting sequence data was assembled with Velvet ver. 1.2.07⁵⁹. The final assembly yielded 113 contigs generating a genome size of 3.10 Mb. The contigs were ordered against the complete genome of C. glutamicum ATCC13032 [PMID: 12743753] using Mauve⁶⁰.

Protein-coding genes (ORFs) of draft genome assemblies were predicted using Glimmer version 3.02 with self-training dataset^61,62. tRNAs and rRNAs were predicted using ARAGORN and RNAmmer, respectively^63,64. Functional assignments of the predicted ORFs were based on a BLASTP homology search against the genome sequence of C. glutamicum ATCC13032, and NCBI non-redundant (NR) database⁶⁵. Functional assignment was also performed with BLASTP homology search against Clusters of Orthologous Groups (COG) databases⁶⁶.

This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession BAYH00000000. The version described in this paper is the first version, BAYH01000000. Illumina sequence reads have been deposited in the DDBJ Sequence Read Archive (DRA) under accession number DRA003758 (BioProject ID: PRJDB1732).

Results

Cell morphology and identification

CS176 was identified as a Gram-positive, rod-shaped, non-spore-forming bacterium. Scanning electron microscopy (SEM) revealed club-shaped cells occurring singly or in V-formations, with dimensions of approximately 0.7–1.2 × 1–2.5 µm (Fig. 1a). This strain was selected for further study based on its distinct morphological and biochemical traits, as well as its ability to utilize l-arabinose as the sole carbon source (see Supplementary Table S1).

Fig.1.

Morphological and phylogenetic characteristics of Corynebacterium glutamicum CS176. (a) Scanning electron microscopy (SEM) image of CS176 cells, showing the typical club-shaped morphology. (b) Neighbor-joining phylogenetic tree based on the 16S rRNA gene sequence, constructed with 1,000 bootstrap replications. The tree compares strain CS176 with type strains of the genus Corynebacterium and highlights clusters containing the arabinose gene cluster. Strains labeled in red harbor arabinose gene cluster type I, while those labeled in blue possess arabinose gene cluster type II. Bootstrap values greater than 50% are shown at branch nodes, and Mycobacterium smegmatis was used as the outgroup.

EzTaxon-e analysis assigned CS176 to the genus Corynebacterium, with the highest 16S rRNA sequence similarity to C. glutamicum ATCC13032(T) (99.86% identity; 1,453 bp; 99.9% coverage). Similarities to other type strains were below 99.0%. Phylogenetic analysis of the 16S rRNA gene sequence further confirmed this classification (Fig. 1b). Sequences from strains harboring ara gene clusters were also included, with type I and type II clusters indicated in red and blue, respectively.

Most Corynebacterium strains in the phylogenetic tree lacked ara gene clusters, indicating that arabinose utilization is not a conserved genus-level trait but rather strain specific. The scattered distribution of ara-positive strains among predominantly ara-negative strains suggests its acquisition by horizontal gene transfer (HGT) rather than vertical inheritance. This observation is consistent with previous reports indicating that the ability to metabolize arabinose is limited to specific Corynebacterium strains^52,67,68. These findings support the view that arabinose utilization genes contribute to metabolic diversification in Corynebacterium and may provide adaptive advantages under defined ecological conditions.

Effect of temperature and oxygen levels on the growth and glutamate production

In this work, the optimum temperature and levels of oxygen on CS176 growth and l-glutamate production were elucidated in shake-flask level using 500 ml flask containing minimal-glucose (1%) medium and 30 ug/L biotin without triggered substance for l-glutamate excretion (e.g. antibiotic, surfactant). Oxygen levels were varied by type of flask and cap, and culture volumes as followed: High oxygen was performed in baffle-flask with sponge cap containing 50 ml medium. Medium–high (MH) and medium–low (ML) oxygen were performed in normal-flask with sponge cap containing 100 and 200 ml medium, respectively and low oxygen was performed in normal-flask with stopped silicone cap containing 100 ml medium.

The optimum temperature for growing CS176 strain was 30–37 °C (Fig. 2 a-c) and can grew up to 40.5 °C when cultivated in the proper of oxygen level. At the optimum temperature, the maximum growth was observed in high oxygen level (Fig. 2, red line) and slightly decreased when the level of oxygen decreased. No growth was observed at high oxygen level when the temperature was higher than 38.5 °C (Fig. 2 d-f, red line) probably due to the co-effect of oxygen and high temperature stress. At MH (Fig. 2, blue line) and ML (Fig. 2, green line) oxygen levels, growth followed a similar trend but was slightly lower compared to high oxygen (Fig. 2, red line) while at low oxygen (Fig. 2, orange line), the growth was consistently lower across all temperatures. The maximum temperature was observed under ML oxygen conditions, with a normal growth curve at 39.5 °C and an extended lag period of more than 24 h at 40.5 °C (Fig. 2e and f green line, respectively).

Fig.2.

Growth (line charts; a–f) and l-glutamate production (bar chart; g) profiles of strain CS176 at various temperatures: (a) 30; (b) 35; (c) 37; (d) 38.5; (e) 39.5; and (f) 40.5 °C. Colors represent oxygen levels as follows: high (red), medium–high (blue), medium–low (green), and low (orange). Experiments were performed in minimal medium supplemented with 1% glucose and 30 µg/L biotin at 200 rpm for 36–64 h. Standard deviations are indicated by error bars.

For l-glutamate production (Fig. 2g), increasing of temperature up to 38.5 or 39.5 °C may enhanced glutamate excretion. The optimal temperature for glutamate produced by CS176 strain was 37–39.5 °C. At 40.5 °C, glutamate production declined due to reduced growth efficiency. Thus, raising the cultivation temperature is a viable strategy to improve glutamate yield. In addition to temperature, oxygen levels significantly influenced l-glutamate production. ML oxygen level consistently yielded the highest l-glutamate production across all studied temperatures (Fig. 2g, green bar chart). The highest glutamate yield, 1,326.26 ± 8.96 mg/L (1.32 g/L), was achieved at 38.5 °C after 36 h of cultivation.

Statistical analysis revealed a significant impact of both temperature (p < 0.05*) and oxygen levels (p < 0.01**) on specific growth rate (µ_max). Similarly, glutamate production was significantly affected by these variables. Post-hoc analysis demonstrated that medium–low oxygen level exhibited the highest glutamate yield, particularly at 38.5 °C. µ_max was maximized at high oxygen levels but did not correspond with increased glutamate production, suggesting a trade-off between growth rate and metabolite accumulation. The findings indicate that optimizing oxygen supply at moderate levels enhances glutamate biosynthesis more effectively than at extreme conditions.

Effect of carbon source on the growth and glutamate production

Here, we compared 2 level of biotin at 200 and 30 ug/L in minimal-glucose (1%) medium with 4 levels of oxygen conditions, 37 °C, 200 rpm. The result showed the same pattern of growth and l-glutamate production between 2 tested biotin levels (Fig. 3a and b, Table 1). No significant effect on either µ_max or glutamate production, indicating that within the tested concentration range, biotin does not play a major role in glutamate biosynthesis (p ˃ 0.05). The oxygen level did not significantly affect µ_max (p > 0.05) (but higher of oxygen led to higher of growth, indicating that growth rate remained stable across conditions. However, oxygen availability had a significant effect on glutamate production (p < 0.01**). The highest glutamate accumulation was observed under ML oxygen conditions. Tukey’s HSD test confirmed significant differences between ML and both high and low oxygen levels (p < 0.05*), suggesting that moderate oxygen limitation enhances glutamate accumulation.

Fig.3.

Growth (line charts) profiles of strain CS176 in various carbon sources and biotin levels: (a) glucose, 200 µg/L biotin [Gb200]; (b) glucose, 30 µg/L biotin [Gb30]; (c) arabinose, 30 µg/L biotin [Ab30]; and (d) mixed glucose–arabinose, 30 µg/L biotin [GAb30]. l-Glutamate production at the final time point is represented by a bar chart (e). Colors represent oxygen levels as follows: high (red), medium–high (blue), medium–low (green), and low (orange). Experiments were performed in minimal medium supplemented with 1% sugar at 37 °C and 200 rpm for 48 h. Standard deviations are indicated by error bars.

Table 1.

Growth performance and glutamate production of Corynebacterium glutamicum CS176 under various oxygen levels, carbon sources, and biotin concentrations.

Oxygen level	Carbon source	Biotin concentration (µg L-1)	Specific growth rate (umax, h-1)	Maximum growth (Klett unit) at culturing time (h)	Glutamate production (mg L-1)
High	Glucose	30	0.234 ± 0.009	463.00 ± 8.54 (20)	131.27 ± 11.32
		200	0.279 ± 0.015	465.00 ± 5.00 (20)	122.97 ± 6.51
	Arabinose	30	0.290 ± 0.010	485.00 ± 5.00 (36)	39.98 ± 3.88
	Glucose + Arabinose	30	0.350 ± 0.016	555.00 ± 10.00 (30)	463.27 ± 9.44
Medium high	Glucose	0	0.302 ± 0.020	358.00 ± 9.85 (36)	1,414.64 ± 37.71
		0.5	0.341 ± 0.027	364.20 ± 5.19 (28)	1,419.20 ± 19.05
		1	0.332 ± 0.019	364.67 ± 9.50 (48)	1,336.97 ± 37.60
		3	0.347 ± 0.030	388.33 ± 10.41 (48)	729.32 ± 12.90
		30	0.348 ± 0.028	413.33 ± 14.43 (28)	656.22 ± 7.36
		200	0.290 ± 0.011	409.83 ± 5.01 (48)	521.36 ± 16.76
	Arabinose	30	0.309 ± 0.014	428.33 ± 7.64 (36)	230.87 ± 6.422
	Glucose + Arabinose	30	0.303 ± 0.008	438.33 ± 7.64 (48)	2,173.02 ± 20.93
Medium low	Glucose	30	0.296 ± 0.009	332.64 ± 6.43 (36)	2,007.03 ± 18.51
		200	0.284 ± 0.011	318.33 ± 10.41 (48)	1,857.63 ± 13.27
	Arabinose	30	0.273 ± 0.006	373.33 ± 7.64 (36)	654.16 ± 14.78
	Glucose + Arabinose	30	0.313 ± 0.019	338.37 ± 7.68 (48)	2,521.60 ± 13.34
Low	Glucose	30	0.283 ± 0.004	301.00 ± 3.61 (48)	139.57 ± 7.08
		200	0.236 ± 0.005	299.67 ± 3.21 (48)	106.37 ± 8.65
	Arabinose	30	0.261 ± 0.008	288.33 ± 12.58 (48)	48.28 ± 6.86
merge this cell to “Low”	Glucose + Arabinose	30	0.303 ± 0.001	286.67 ± 5.77 (48)	122.97 ± 11.95

The data represent mean values ± standard deviations from replicate experiments.

Because of CS176 strain can utilize l-arabinose as sole carbon source, thus, their growth and l-glutamate production from 1% arabinose as sole and mixed with glucose (1% of each sugar) were elucidated compare with using glucose as sole carbon source at 4 levels of oxygen supplemented with 30 ug/L biotin. There were differences in growth and glutamate profiles among 3 carbon sources (Fig. 3 b-d, Table 1). l-Arabinose as sole carbon source resulted in minimal glutamate production regardless of oxygen levels (Fig. 3e, Ab30 group), implying that its metabolism via the pentose phosphate pathway (PPP) may not sufficiently drive glutamate biosynthesis without intermediates from glucose metabolism. While using glucose-arabinose mixture can promote the maximum growth at high oxygen level (Fig. 3d, red line) and increased µ_max from 0.234 ± 0.009 and 0.290 ± 0.010 in the presence of glucose and arabinose as sole carbon source, respectively, to 0.350 ± 0.016 h^-1. Two-way ANOVA showed that both of carbon source and oxygen level exhibited highly significant effects on µ_max and glutamate production, as did their interaction (p < 0.001***). For glutamate production, glucose-arabinose mixture exhibited a trend toward higher glutamate production, especially at high and MH oxygen levels, the glutamate was increased from 39.98 ± 3.88 (arabinose) and 123 ± 6.51 (glucose) to 463.27 ± 9.44 mg/L (glucose-arabinose mixture) under high oxygen and 230.87 ± 6.42 (arabinose) and 496.46 ± 11.80 (glucose) to 2,173.02 ± 20.93 mg/L (glucose-arabinose mixture) under MH oxygen level. The highest l-glutamate still observed at ML O₂ level with 2,521.62 ± 13.34 mg/L. Tukey’s HDS tests indicating that mixed glucose-arabinose combined with MH and ML oxygen conditions were significantly superior to all others. Moreover, simple effects analysis confirmed the effect of carbon sources on µ_max and glutamate production depended on the oxygen levels, with ML oxygen favoring higher glutamate production in specific carbon sources.

Effect of biotin on the growth and glutamate production

To further elucidate the effect of biotin on CS176 growth and l-glutamate production, biotin concentration was varied from 0–200 µg/L in minimal glucose medium under medium–high (MH) oxygen conditions at 37 °C and 200 rpm, without a cell-washing step. Biotin carried over from the inoculum was sufficient to support CS176 growth at 0 µg/L biotin in the main culture medium. All cultures exhibited a typical sigmoidal growth curve, reaching stationary phase after 17 h (Fig. 4a, line chart). Cultures supplemented with 30 and 200 µg/L biotin showed the highest biomass accumulation (Fig. 4a, red and green lines, respectively), while cultures with 0 and 0.5 µg/L biotin exhibited the lowest final growth (Fig. 4a, blue and orange lines, respectively). The µ_max was influenced by biotin concentration in the culture medium (Table 1). Under biotin-limited conditions (0 µg/L), µ_max was lowest at 0.302 ± 0.020 h⁻¹, indicating restricted cellular proliferation. Biotin supplementation led to a gradual increase in µ_max, reaching 0.341 ± 0.027 h⁻¹ at 0.5 µg/L and 0.347 ± 0.030 h⁻¹ at 3 µg/L. The highest µ_max was observed at 30 µg/L biotin (0.348 ± 0.028 h⁻¹), beyond which no substantial enhancement in growth was detected. One-way ANOVA revealed that biotin concentration had a statistically significant effect on µ_max (p < 0.05*). However, Tukey’s HSD post hoc test showed that most pairwise comparisons were not statistically significant (p > 0.05). In contrast to growth, the highest glutamate production occurred at biotin concentrations of 0 to 1.0 µg/L (Fig. 4b, blue, orange, and grey, respectively), with values of 1,414.64 ± 37.71, 1,419.20 ± 19.05, and 1,336.97 ± 37.60 mg/L, respectively. Biotin levels ≥ 3 µg/L (Fig. 4b, yellow) reduced glutamate accumulation by approximately half (p < 0.001***). This reduction in glutamate production at high biotin levels contrasts with the behavior observed in C. glutamicum ATCC130322.

Fig.4.

Growth (line chart; a) and glutamate production (bar chart; b) of strain CS176 at varied biotin concentrations. Colors represent biotin concentrations as follows: 0 µg/L (blue), 0.5 µg/L (orange), 1 µg/L (grey), 3 µg/L (yellow), 30 µg/L (red), and 200 µg/L (green). Experiments were performed under medium–high oxygen conditions at 37 °C and 200 rpm. Standard deviations are indicated by error bars.

Genome properties

Draft genome sequence of C. glutamicum CS176 is 3,100,840 bp long (113 contigs) with a 54.20% G + C content (Table 2). Of the 2,915 predicted genes, 2,844 were protein coding genes, and 71 were RNAs (5 rRNA genes, 65 tRNA genes, and 1 tmRNA gene). A total of 1,946 genes (68.42%) were assigned a putative function. The remaining genes were annotated as hypothetical genes. The properties and the statistics of the genome are summarized in Table 2. The distribution of genes into COGs functional categories is presented in Table 3. Of the 2,844 proteins, 2,170 (76.30%) were assigned to COG functional categories. The 256 proteins were assigned to multiple COG categories. The most abundant COG category was “General function prediction only” (237 proteins) followed by “Function unknown” (177 proteins), “Amino acid transport and metabolism” (158 proteins), “Transcription” (155 proteins), “Inorganic ion transport and metabolism” (150 proteins), and “Translation, ribosomal structure and biogenesis” (143 proteins). The genome map of C. glutamicum CS176 is illustrated in Fig. 5 demonstrates that the pattern of GC skew shifts from negative to positive along an ordered set of contigs with some exceptions. This suggests that the draft genome sequences were ordered almost exactly.

Table 2.

Genome statistics.

Attribute	Value	% of Total
Genome size (bp)	3,100,840	100.00
DNA coding (bp)	2,727,323	87.95
DNA G + C (bp)	1,680,774	54.20
DNA scaffolds	113	100.00
Total genes	2,915	100.00
Protein coding genes	2,844	97.56
RNA genes	71	2.44
Pseudo genes	0	0
Genes in internal clusters	N/A	N/A
Genes with function prediction	1,946	68.42
Genes assigned to COGs	2,170	76.30
Genes with Pfam domains	2,353	82.73
Genes with signal peptides	203	7.13
Genes with transmembrane helices	745	26.19
CRISPR repeats	0	0.00

Table 3.

Number of genes associated with general COG functional categories.

Code	Value	%age	Description
J	143	5.03	Translation, ribosomal structure, and biogenesis
A	1	0.04	RNA processing and modification
K	155	5.45	Transcription
L	121	4.25	Replication, recombination, and repair
B	0	0.00	Chromatin structure and dynamics
D	19	0.67	Cell cycle control, Cell division, chromosome partitioning
V	46	1.62	Defense mechanisms
T	37	1.30	Signal transduction mechanisms
M	87	3.06	Cell wall/membrane biogenesis
N	0	0.00	Cell motility
U	22	0.77	Intracellular trafficking and secretion
O	68	2.39	Posttranslational modification, protein turnover, chaperones
C	114	4.01	Energy production and conversion
G	139	4.89	Carbohydrate transport and metabolism
E	158	5.56	Amino acid transport and metabolism
F	61	2.14	Nucleotide transport and metabolism
H	95	3.34	Coenzyme transport and metabolism
I	57	2.00	Lipid transport and metabolism
P	150	5.27	Inorganic ion transport and metabolism
Q	27	0.95	Secondary metabolites biosynthesis, transport, and catabolism
R	237	8.33	General function prediction only
S	177	6.22	Function unknown
-	674	23.70	Not in COGs
-	256	9.00	Multiple COG assignment

The total is based on the total number of protein coding genes in the genome.

Fig.5.

Graphical circular map of a simulated draft Corynebacterium glutamicum CS176 genome. The simulated genome is a set of contigs ordered against the complete genome of C. glutamicum ATCC13032 using Mauve (PMID: 12743753; PMID: 19515959; PMID: 20593022). The circular map was generated using CGview (PMID: 15479716). Genome sequence of 21 C. glutamicum genomes were aligned by using nucmer (PMID: 14759262). The ara gene cluster of ATCC31831 strain (GenBank ID: AB447371) was also mapped.

Consistent with its observed ability to utilize l-arabinose, genome analysis revealed the presence of a six-gene cluster comprising araA, araD, araB, galM, araR, and araE (Fig. 5). This organization is identical to the l-arabinose utilization gene cluster previously reported in C. glutamicum ATCC31831⁵². The nucleotide sequences of the gene clusters in CS176 and ATCC31831 share 99–100% identity across the entire 7.8 kb region. No homologous cluster was detected in other publicly available C. glutamicum genome sequences (see Supplementary Table S2). Therefore, it may suggest that HGT event occurred before the divergence of these two strains.

Distribution of l-arabinose utilization gene clusters in the genus Corynebacterium

To investigate the distribution of l-arabinose utilization genes, homologous clusters were searched in available Corynebacterium genome sequences. A gene cluster with the same organization as that in CS176 and ATCC31831 was identified in Corynebacterium stationis strain GA-15 and was designated as type I (Fig. 6). In addition, a distinct gene cluster organization, designated as type II, was identified in four genomes: Corynebacterium sp. DNF00584, Corynebacterium sp. HMSC077B05, Corynebacterium sp. HMSC065A05, and Corynebacterium sp. CMW7794 (Fig. 6; Table S2). The type II cluster lacks the galM gene encoding aldose 1-epimerase and exhibits lower sequence identity relative to type I clusters. Phylogenetic analysis (Fig. 1b) revealed that C. glutamicum and C. stationis (encoding type I clusters) are not positioned within the same clade and are classified into distinct phylogenetic groups. In contrast, the four genomes encoding type II clusters form a closely related clade.

Fig.6.

Comparisons of l-arabinose utilizing gene clusters conserved in Corynebacterium strains. Green, pink, and grey colors correspond to the amino acid identity based on BLASTP. Figure was drawn by using genoPlotR (PMID: 20624783). Orange and cyan arrows indicate type I and type II gene cluster, respectively.

To explore the possible origin of the ara gene cluster, BLASTP searches were performed using the amino acid sequences of the six genes against 2,718 complete genomes excluding Corynebacterium species (PMID: 9254694). The proteins AraA, AraD, AraB, GalM, and AraR were consistently positioned within Actinobacterial clades, whereas AraE clustered within Firmicutes. Comparative analysis of gene organization revealed that the araBAD operon is present in several Actinobacterial genomes, including Arthrobacter arilaitensis Re117, Cellulomonas fimi ATCC484, and Isoptericola variabilis 225 (see Supplementary Table S3). However, a complete type I or type II gene cluster was not identified in any non-Corynebacterium genome examined.

Discussion

The present study reveals that glutamate production in strain CS176 arises from an integrated interaction among oxygen-transfer regime, temperature-dependent envelope dynamics, central carbon flux, and accessory genomic traits rather than from a single classical induction trigger. The observed hierarchy of intermediate aeration regimes (ML > MH > high (H) > low (L)) indicates the presence of a strain-specific oxygen window, consistent with earlier analyses showing that both excessive and severely restricted aeration impair glutamate productivity⁶⁹.

Our results demonstrate that both temperature and oxygen levels significantly influence the growth and glutamate production of C. glutamicum CS176. Optimal growth was observed at moderate temperatures (30–37 °C), particularly under higher oxygen availability. Total growth was the highest under high oxygen condition and maximum specific growth rate (µₘₐₓ) was achieved under MH oxygen condition (Table 1). However, this did not directly correlate with increased glutamate production. Instead, the highest glutamate concentration was consistently obtained under ML oxygen levels at 37 °C, but it was observed at 38.5 °C (followed by 39.5 °C and 37 °C, respectively), without the addition of chemical trigger substances.

Temperature also plays a crucial role in modulating both growth and glutamate production. Generally, C. glutamicum strains grow well within a temperature range of 25–37 °C, while temperatures above 40 °C significantly inhibit both growth and glutamate synthesis^70,71. Our findings confirm this, as CS176 displayed robust growth between 30–37 °C but showed reduced growth at higher temperatures, especially under high oxygen conditions. Although elevated temperatures may stimulate glutamate secretion, they also impose cellular stress, particularly when coupled with high oxygen availability. A temperature-triggered process developed by Delaunay et al.^14,72. demonstrated that a temperature-sensitive C. glutamicum strain 2262 produced high glutamate titers (up to 85 g/L from 208 g of consuming glucose) when the temperature was shifted from 33 °C to 39–40 °C during fed-batch fermentation. This process yielded a maximum specific production rate of 0.63 g/g/h and a glutamate yield of 0.46 g per gram of glucose. Although higher temperatures promote glutamate synthesis, they concurrently reduce growth rates and increase lactate accumulation. This temperature-triggered approach represents a promising alternative to traditional induction methods such as biotin limitation or surfactant addition for industrial glutamate production. Uy et al.¹⁶ further showed that elevating the temperature from 33 °C to 39 °C represses oxoglutarate dehydrogenase complex (ODHC) activity, reducing the growth rate while enhancing glutamate excretion to prevent toxic intracellular accumulation.

At the central metabolic level, glutamate overflow is probably governed by flux redistribution at the 2-oxoglutarate node. In C. glutamicum, the activity of ODHC is modulated by OdhI phosphorylation status, which is influenced by biotin limitation and temperature upshift conditions associated with glutamate production⁷³. Such regulation enables conditional redirection of carbon flux toward glutamate accumulation. Moderate oxygen limitation also promotes glutamate formation in C. glutamicum probably by attenuating flux through the TCA branch point at 2-oxoglutarate and reducing the activity of ODHC. Complete loss of ODHC activity, such as through odhA deletion, leads to constitutive glutamate overproduction even without induction⁷⁴. Flux analyses further demonstrate that reduced ODHC activity redirects carbon from complete oxidation via the TCA cycle toward glutamate accumulation⁷⁵. In contrast, conditions that sustain high ODHC activity, such as excess oxygen supply, favor biomass formation and suppress glutamate excretion. In CS176, intermediate oxygen-transfer regimes may create a metabolic state in which TCA throughput is sufficient to sustain energy generation while permitting partial restriction at the ODHC node, thereby promoting overflow. In contrast, excessive aeration may enhance respiratory flux and reduce metabolic pressure at this branch point, while severe oxygen limitation likely constrains overall carbon flux intensity.

To experimentally validate these mechanistic hypotheses in CS176, several targeted approaches are warranted. First, ODHC enzymatic activity should be measured directly in cell-free extracts prepared from cultures grown under varying dissolved oxygen (DO) levels, linking oxygen availability to flux restriction at the 2-oxoglutarate node. Second, ¹³C-metabolic flux analysis (¹³C-MFA) using [1-¹³C] or [U-¹³C] glucose as tracer substrates would provide quantitative resolution of carbon partitioning between the TCA cycle and glutamate biosynthesis under permissive and non-permissive aeration regimes. Finally, construction of an odhA deletion strain would determine whether or not constitutive ODHC inactivation drives glutamate overproduction independently of induction conditions in CS176.

Under mixed carbon (glucose-arabinose) utilization conditions, aeration intensity exerted a differential impact on growth and glutamate production. High aeration increased the maximum specific growth rate (µₘₐₓ) by approximately 1.5-fold and enhanced glutamate production by 3.5-fold by mixed carbon relative to the glucose condition, while MH aeration resulted in a modest reduction in µₘₐₓ (0.87-fold) but still markedly increased glutamate production (3.3-fold). In contrast, ML and low aeration conditions led to only marginal changes in both µₘₐₓ and glutamate production. These results indicate that lower aeration does not effectively promote glutamate production under mixed carbon conditions.

Mechanistically, higher oxygen availability likely enhances l-arabinose catabolism, thereby increasing ATP generation and precursor supply to support biomass formation and overall metabolic activity. In C. glutamicum, efficient pentose utilization requires proper respiratory activity to maintain redox balance and NAD⁺ regeneration, linking oxygen transfer to carbon flux distribution. Under high oxygen conditions, improved respiratory capacity may facilitate arabinose assimilation through the pentose phosphate pathway, providing reducing power (NADPH) and intermediates that support both growth and glutamate biosynthesis⁷⁶. Consequently, glucose-derived carbon can be more effectively redirected toward the TCA cycle and glutamate formation^22,28. In contrast, under low oxygen conditions, limited respiratory activity may constrain arabinose utilization and energy generation, thereby reducing metabolic stimulation of glucose flux toward glutamate production.

Draft genome sequencing identified a 7.8 kb gene cluster associated with l-arabinose utilization in strain CS176, showing high sequence identity to the corresponding region in Corynebacterium glutamicum ATCC31831, a naturally arabinose-utilizing strain⁵² (see Supplementary Table S4). The cluster comprises six genes, including the araBDA operon consisting of araB (l-ribulokinase), araD (l-ribulose-5-phosphate-4-epimerase), and araA (l-arabinose isomerase), which collectively mediate arabinose catabolism. Upstream of the operon, three additional genes—araE (l-arabinose transporter), araR (LacI-type transcriptional regulator), and galM (putative aldose 1-epimerase)—are divergently oriented relative to the araBDA operon. Comparative genomic analysis further demonstrated that this complete six-gene cluster was absent in the non-arabinose-utilizing reference strain C. glutamicum ATCC13032, whereas both CS176 and the naturally arabinose-utilizing strain ATCC31831 possessed the entire cluster with high sequence homology (see Table S4). These genes collectively encode proteins involved in arabinose transport, catabolism, and transcriptional regulation, suggesting that the cluster likely represents a candidate genetic determinant associated with efficient arabinose utilization in these strains. In particular, araE, predicted to encode an arabinose transporter, together with the catabolic genes araA, araB, and araD, may facilitate efficient uptake and assimilation of arabinose into central carbon metabolism. In addition, the regulatory gene araR may contribute to coordinated transcriptional regulation of arabinose-responsive genes, thereby supporting rapid adaptation to arabinose-containing environments, consistent with previous reports describing AraR-mediated regulation of araBDA and araE expression in C. glutamicum⁷⁷.

The functional significance of the araBDA operon was previously demonstrated through heterologous introduction of the operon from ATCC31831 into C. glutamicum strain R, which is naturally incapable of utilizing l-arabinose. The resulting recombinant strain, CRA16, acquired the ability to grow on l-arabinose as the sole carbon source⁵². Because this cluster is absent from most sequenced C. glutamicum genomes, its occurrence appears to be strain-specific rather than universally conserved within the species. Furthermore, the GC content of the arabinose utilization cluster in CS176 was 63.1%, which is approximately 8.9% higher than the genome average (54.2%), supporting the hypothesis that this region was acquired through HGT. The high sequence conservation observed between CS176 and ATCC31831 further suggests that this acquisition event likely occurred prior to the evolutionary divergence of these strains.

The absence of the arabinose utilization gene cluster in ATCC13032 further suggests that acquisition of this cluster may have contributed to the emergence of arabinose utilization capability. Accordingly, heterologous transfer of the complete gene cluster into industrial C. glutamicum strains, including ATCC13032, may confer arabinose utilization capability and improve pentose sugar metabolism. It remains unclear whether the extent to which the six-gene arabinose utilization cluster alone is sufficient to reproduce the rapid arabinose-utilizing phenotype of CS176. Additional factors, including transporter efficiency, transcriptional regulation, global carbon metabolism, and cofactor balance, may also be required to obtain the observed phenotype. In any case, this strategy could facilitate the development of engineered strains capable of efficiently utilizing lignocellulosic pentose sugars derived from renewable biomass for sustainable bioprocess applications^78–80.

Comparative genomic analysis additionally revealed the presence of two distinct l-arabinose utilization cluster architectures within the genus Corynebacterium. Type I clusters were identified in phylogenetically distant species, including C. glutamicum and C. stationis, whereas type II clusters were restricted to a closely related lineage. This contrasting distribution pattern suggests distinct evolutionary trajectories for the two cluster types. The limited distribution of type II clusters further implies that this configuration may represent a low-frequency accessory component of the Corynebacterium pan-genome. Phylogenetic analysis of individual cluster proteins indicated that most components are closely related to homologs from Actinobacteria, whereas AraE appears to have originated from Firmicutes, suggesting a mosaic evolutionary history involving multiple gene acquisition events. The widespread occurrence of araBDA operons among several Actinobacterial genera supports the possibility that an ancestral araBDA module was initially acquired from Actinobacteria and subsequently reorganized within Corynebacterium genomes, giving rise to the present cluster architectures. Notably, no complete type I or type II cluster organization was identified outside the genus Corynebacterium, indicating that these specific arrangements were likely assembled within the lineage following independent acquisition of constituent genes. Further genome sequencing of diverse Corynebacterium strains will provide deeper insight into the evolutionary origin and diversification of these l-arabinose utilization systems.

Biotin plays an essential role as a cofactor for enzymes involved in fatty acid biosynthesis pathways in biotin-auxotrophic C. glutamicum strains⁸¹. Biotin limitation generally restricts cell growth but promote glutamate excretion, whereas excess biotin suppresses glutamate production. In this work, we found the optimal biotin concentration for CS176 growth was 30 µg/L, while higher concentrations did not further enhance µmax. Although glutamate production was approximately twofold higher under biotin-limited conditions (0–0.5 µg/L), substantial glutamate production was still observed at higher biotin concentrations (30–200 µg/L). Thus, unlike C. glutamicum ATCC13032, glutamate production in CS176 was not strongly repressed under high-biotin conditions. Moreover, even in the naturally arabinose-utilizing strain C. glutamicum ATCC31831, glutamate production was not observed at either 20 or 200 µg/L biotin in the absence of inducers such as Tween 40 or penicillin under the conditions tested⁸². These observations suggest that CS176 possesses a distinct regulatory or physiological mechanism that permits glutamate production even under biotin-sufficient conditions.

In C. glutamicum ATCC13032, ODHC activity remains high during the late logarithmic phase under biotin-rich conditions, whereas under biotin-limited conditions ODHC activity decreases rapidly during stationary phase⁸³. In addition, deletion of dtsR1, which encodes a component of acetyl-CoA carboxylase, causes constitutive glutamate overproduction accompanied by reduced ODHA activity⁸⁴. Gain-of-function mutations in the mechanosensitive channel MscS, which mediates glutamate export, have also been shown to induce constitutive glutamate secretion⁸⁵. These findings indicate that glutamate production under biotin-sufficient conditions may be influenced by multiple factors, including biotin-dependent carboxylase activity, fatty acid biosynthesis, ODHC regulation, cell envelope properties, and glutamate export.

To explore possible genetic differences underlying the unusual biotin response of CS176, we compared representative genes involved in biotin uptake, biotin biosynthesis, biotin-dependent carboxylation, ODHC regulation, and glutamate export among CS176, ATCC13032, and ATCC31831 (Table S4). CS176 retained the canonical gene sets for biotin uptake (bioYMN), biotin biosynthesis (bioB and bioD), biotin-dependent carboxylases, including pyruvate carboxylase and acetyl-CoA carboxylase components, and glutamate production or export, including odhA, odhI, and mscS. Most of these genes showed high sequence conservation with their ATCC13032 counterparts. However, several genes, particularly bioD, birA, and accE, showed relatively lower sequence identity in CS176 and contained amino acid substitutions compared with the reference strains. Although these substitutions are potential candidates for the unusual maintenance of glutamate production under high-biotin conditions, their functional significance remains unknown.

At present, there is no direct evidence linking these sequence differences to the CS176 phenotype. Therefore, future studies should examine whether allelic differences in genes related to biotin metabolism, fatty acid biosynthesis, ODHC regulation, or glutamate export contribute to the inducer-independent glutamate production observed in CS176. Allele-replacement experiments or heterologous expression of candidate CS176 genes in ATCC13032 would be particularly useful for determining whether these mutations are sufficient to reproduce the high-biotin glutamate-producing phenotype.

Taken together, the data support a unified model in which CS176 represents a strain-specific adaptive configuration shaped by both physiological regulation and genomic innovation. At the metabolic level, an intermediate oxygen-transfer window optimizes flux redistribution at the 2-oxoglutarate node. At the evolutionary level, horizontally acquired accessory pathways expand ecological competence without fundamentally altering core overflow mechanisms. The integration of these layers defines the distinctive production phenotype of CS176 and highlights the interplay between central metabolism, cell envelope architecture, and genome plasticity in shaping strain-specific metabolic performance. The arabinose utilization capability of CS176 is of particular industrial relevance, as pentose sugars constitute a major fraction of lignocellulosic biomass hydrolysates, positioning this strain as a promising platform for sustainable amino acid production from renewable feedstocks. Future studies integrating transcriptomic and metabolic analyses, together with functional validation of the arabinose utilization cluster, will be important for further developing CS176 as an engineered production host.

Author contributions

ST supervised the research and served as the corresponding author. PF designed and performed the experiments, analyzed the data, prepared figures and tables. KM, MM and NK supported experimental design, genome analysis, and parts of the experiments of PF, respectively. PF prepared the drafted manuscript, which was collected by ST, KM, NK, and TY. All authors read and approved the manuscript.

Funding

This research was supported by the Faculty of Science, Kasetsart University, through the KU Science Graduate Research Fund. This work was also supported by the Advanced Low Carbon Technology Research and Development Program (ALCA) of the Japan Science and Technology Agency (JST; JPMJAL1106). Additional financial support was provided through the Core-to-Core Program funded by the Japan Society for the Promotion of Science (JSPS) and the National Research Council of Thailand (NRCT).

Data availability

The 16S rRNA gene sequence of strain CS176 has been deposited in the NCBI database under the accession number MN826455.1 (https://www.ncbi.nlm.nih.gov/nuccore/MN826455.1). The Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession number BAYH00000000, with the version described in this study corresponding to BAYH01000000 (https://www.ncbi.nlm.nih.gov/nuccore/BAYH00000000.1/). The raw Illumina sequencing reads have been deposited in the DDBJ Sequence Read Archive (DRA) under the accession number DRA003758, associated with BioProject ID PRJDB1732 (https://ddbj.nig.ac.jp/public/ddbj_database/dra/fastq/DRA003/DRA003758/)

Declarations

Competing interests

The authors declare no competing interests.