Physiological metabolic analysis and process optimization of hypoxia in promoting coenzyme Q₁₀ biosynthesis and accumulation in Rhodobacter sphaeroides HY01

Abstract

Coenzyme Q₁₀ fermentation by Rhodobacter sphaeroides (R. sphaeroides) is highly aerobic. However, its biosynthesis and accumulation are paradoxically induced under hypoxic conditions. While oxygen serves as both an induction signal and a key precursor, the mechanisms underlying Coenzyme Q₁₀ accumulation under hypoxia remain elusive, posing a significant bottleneck for yield improvement. This study systematically elucidated these mechanisms through transcriptomic analysis. Results revealed that metabolic pathways for precursor and Coenzyme Q₁₀ biosynthesis were, surprisingly, downregulated under hypoxia. Conversely, genes inhibiting morphological remodeling were downregulated, while those involved in cell membrane biosynthesis were upregulated. Furthermore, a significant positive correlation was observed between cellular morphology and Coenzyme Q₁₀ yield. Specifically, under varying hypoxic conditions, cell morphology and Coenzyme Q₁₀ yield also exhibited strong correlations, with higher yields were associated with larger cell volumes. These findings suggest that R. sphaeroides enhances Coenzyme Q₁₀ accumulation primarily by expanding the cell volume and membrane surface area rather than upregulating biosynthetic pathways. Leveraging this insight, the addition of unsaturated fatty acids further increasing the Coenzyme Q₁₀ yield by 18.9%. This study provides a novel strategy for enhancing Coenzyme Q₁₀ production through morphological engineering and process optimization.

1. Introduction

Coenzyme Q₁₀, also known as ubiquinone, is found in the cell membranes of prokaryotes and the inner mitochondrial membranes of eukaryotes. It primarily serves as an electron carrier in the respiratory chain, participating in ATP biosynthesis, and possesses antioxidant properties [1,2]. Coenzyme Q₁₀ is one of the world's largest nutritional supplements, widely used in medicine, food, and cosmetics [3]. It is primarily produced through microbial fermentation, with strains like Rhodobacter sphaeroides (R. sphaeroides), Agrobacterium tumefaciens, and Candida tropicalis being commonly used. Among these, R. sphaeroides is the natural host with the highest yield and is the strain currently employed in industrial production [4].

The biosynthesis of Coenzyme Q₁₀ in R. sphaeroides presents a complex relationship with oxygen. While oxygen is a mandatory substrate for quinone ring modification [1,5] and the fermentation process is inherently aerobic, Coenzyme Q₁₀ accumulation is paradoxically triggered under hypoxic conditions. This hypoxic state occurs during fermentation when the oxygen uptake rate (OUR) surpasses the oxygen transfer rate (OTR), causing the dissolved oxygen (DO) to drop to zero [6]. Consequently, oxygen acts as both a critical biosynthetic precursor and a regulatory signal, making its precise management essential for optimizing production [4]. To improve Coenzyme Q₁₀ production, researchers have conducted extensive studies on oxygen regulation. Zhu et al. achieved a shake flask fermentation level of 163.5 mg/L by simultaneously regulating the redox potential and OUR in R. sphaeroides [7]. Yen et al. found that oxygen limitation is beneficial for Coenzyme Q₁₀ biosynthesis [8]. Zhang et al. discovered that NADH is crucial for Coenzyme Q₁₀ biosynthesis, and increasing NADH levels can improve Coenzyme Q₁₀ production [9]. However, current studies mainly focus on two directions: (i) process optimization and control based on the OUR or the specific oxygen uptake rate (q_O2) during fermentation [6]; and (ii) modulation of cellular oxygen uptake intensity through strategies such as expressing oxygen-binding proteins. In contrast, systematic investigations into hypoxic-induced metabolic pathway reprogramming in industrial R. sphaeroides strains remain limited, and metabolomic evidence that directly reflects the physiological and metabolic state of cells is particularly scarce. In addition, oxygen plays a dual role in this system as both an “inducing signal” and a “key precursor,” meaning that hypoxia not only triggers regulatory responses but also imposes constraints on metabolic supply, thereby markedly increasing the regulatory complexity of CoQ₁₀ biosynthesis and accumulation in R. sphaeroides under hypoxic conditions. The underlying mechanism has not been systematically elucidated. Therefore, clarifying how oxygen limitation regulates intracellular metabolic pathways in R. sphaeroides is highly significant for targeted metabolic engineering and for optimizing industrial production processes precisely.

Moreover, no modifications have successfully increased Coenzyme Q₁₀ yield in industrial strains [10], indicating that the rate limiting steps affecting overproduction of Coenzyme Q₁₀ in R. sphaeroides have yet to be elucidated. Additionally, significant differences in Coenzyme Q₁₀ yield and cell morphology are observed under different levels of hypoxia, and different morphologies correspond to different Coenzyme Q₁₀ yields [6,11]. Yoshida et al. also found that Coenzyme Q₁₀ yield was higher under hypoxic condition, accompanied by an increase in cell membrane thickness [12]. The relationship between the cell morphology of R. sphaeroides and coenzyme Q₁₀ biosynthesis is also unknown.

This study employed transcriptomic analysis to investigate the gene expression changes in key pathways before and after hypoxia in the industrial production strain R. sphaeroides. The relationship among cell morphology, Coenzyme Q₁₀ production and oxygen supply was subsequently analyzed. Finally, the fermentation process was precisely optimized guided by transcriptomic findings.

2. Material and methods

2.1. Microorganism and culture conditions

The strain used in this study was R. sphaeroides, which was stored at −80 °C. The agar medium consisted of 15 g/L Yeast extract, 1 g/L KH₂PO₄, 2 g/L NaCl, 0.1 g/L FeSO₄, 0.5 g/L MgSO₄, 0.03 g/L MnSO₄, 0.001 g/L CoCl₂, 0.000404 g/L choline chloride, 20 g/L agar, 2 mL/L auxiliary solution (0.45 g/L thiamine, 0.045 g/L riboflavin, 0.083 g/L pyridoxine, 0.511 g/L niacin, 0.511 g/L nicotinamide, 0.025 g/L folic acid, 0.016 g/L biotin, and 0.008 g/L calcium pantothenate), and pH 7.0–7.2. The seed medium consisted of ammonium sulfate 3 g/L, aginomoto 0.7 g/L, corn steep powder 0.7 g/L, glucose 15 g/L, yeast extract 2 g/L, NaCl 1 g/L, K₂HPO₄ 0.7 g/L, KH₂PO₄ 0.7 g/L, NH₄Cl 1 g/L, MgSO₄ 3.6 g/L, ferrous sulfate 0.17 g/L, and pH 7.0. The fermentation medium consisted of ammonium sulfate 3.6 g/L, aginomoto 3.6 g/L, corn steep powder 3.6 g/L, KH₂PO₄ 2 g/L, MgSO₄ 5.6 g/L, ferrous sulfate 1.2 g/L, copper sulfate 2.5 g/L, yeast cream 0.6 g/L, Calcium carbonate 2.5 g/L, and pH 6.7. The feed medium consisted of glucose 600 g/L.

2.2. Fermentation in shake flask and bioreactor

To activate the strain, R. sphaeroides was cultured on solid agar medium at 32 °C for 48-72 h. Than, a total of 1-2 single colonies in the plate cultured were inoculated into 500 mL seed flask with a working volume of 100 mL and grown at 260 rpm and 32 °C for 28 h until the optical density at 700 nm (OD₇₀₀) reached 15. The resulting first seed culture was then transferred to second seed flasks or seed fermenters for further cultivation. The culture conditions for the second seed flasks were identical to those of the first seed. For the seed fermenters, the cultivation was carried out at 32 °C with an agitation speed of 400 rpm, an aeration rate of 1 vvm, and a pressure of 0.05 MPa, until the OD₇₀₀ exceeded 15. The well-cultured seeds were then inoculated into 500 mL shake flask (containing 50 ml of fermentation broth) and 50 L bioreactor (containing 30 L of fermentation broth), respectively, for shake flask and bioreactor fermentation. Shake flask fermentation conditions were set at 32 °C and 260 rpm for 48 h. Bioreactor fermentation conditions were set with a culture temperature of 32 °C, a rotation speed of 400 rpm, a pressure of 0.05 Mpa, and an aeration rate of 1 vvm. The pH was controlled at 6.7 using ammonia, and the sugar concentration was maintained at 3-5 g/L by adding 600 g/L glucose, and the fermentation was incubated for about 96 h.

2.3. Analytical methods

Biomass was quantified by measuring OD₇₀₀ and dry cell weight (DCW). OD₇₀₀ was measured using ultraviolet–visible spectrophotometry, with deionized water serving as the blank solution. The DCW detection method is as follows: 10 mL of culture broth was centrifuged at 8000 rpm for 5 min, the supernatant was discarded, and the cell was dried to a constant weight at 60 °C, then converted to standardized DCW (g/L). Coenzyme Q₁₀ was quantified using high-performance liquid chromatography (HPLC) [13]. The exhaust gas produced during fermentation was analyzed using a MAX300-LG mass spectrometer, and the OUR was calculated based on the collected data [11]. The cell size of R. sphaeroides was determined using Image-Pro Plus software (Media Cybernetics, Bethesda, MA) [6]. First, cell images acquired using a CCD camera were subjected to contrast enhancement. Next, binarization was performed to remove noise and achieve smoothing. Finally, the projected area of the cells was quantified. For each time point, measurements were performed in triplicate, and no fewer than 50 cells were counted per replicate.

2.4. Transcriptome analysis

During the fermentation process, three biological replicate samples were collected for transcriptomic analysis under varying oxygen levels: 20% DO (DO-20), hypoxia for 4 h (Hypoxic-4 h), and hypoxia for 24 h (Hypoxic-24 h). The samples were immediately subjected to low-temperature centrifugation (4 °C, 8000 rpm, 5 min) and washed with phosphate-buffered saline (PBS). The cells were then rapidly frozen in liquid nitrogen and stored at −80 °C for subsequent analysis. Total RNA was first extracted from the samples, and its concentration and purity were assessed using a Nanodrop 2000 spectrophotometer. Ribosomal RNA (rRNA) was then removed to isolate messenger RNA (mRNA). A fragmentation buffer was applied to randomly break the mRNA into fragments approximately 200 base pairs in length. Complementary DNA (cDNA) was biosynthesized from the fragmented mRNA via reverse transcription, and the resulting fragments were enriched through PCR to construct the final cDNA library. The cDNA library was sequenced using the Illumina HiSeq X platform by Shanghai Hongshi Biotechnology Co., Ltd., with gene annotation conducted based on the R. sphaeroides genome.

Differential gene expression analysis was performed using the DESeq software, with thresholds set at |log2Ratio| ≥1 and p-value ≤0.05. Gene expression data were visualized through heatmaps generated using TBtools, accompanied by hierarchical clustering analysis [14]. Additionally, KEGG pathway enrichment analysis was carried out using the hypergeometric test to identify significantly enriched pathways.

2.5. Statistical analysis

All data are expressed as the mean ± standard deviation. Statistical analyses were performed using one-way analysis of variance (ANOVA). Charting was carried out using Prism 7 (version 7.0, GraphPad Software Inc., San Diego, CA) and OriginPro 2021 (OriginLab, Northampton, MA).

3. Results

3.1. The physiological metabolic parameters in the fermentation process of R. sphaeroides

Oxygen is essential for microbial growth and product biosynthesis in aerobic fermentation. The control of oxygen concentration in the fermentation process is generally indexed by DO. For highly aerobic fermentations, there is a hypoxic condition where DO falls to zero, which does not favor product produce [15]. However, for some secondary metabolites, hypoxia serves as a signal for the initiation of product biosynthesis and accumulation [16].

Fig. 1 shows the physiological parameters during the biosynthesis of Coenzyme Q₁₀ by R. sphaeroides. Before 20 h, the fermentation has not entered a hypoxic state. R. sphaeroides relies on aerobic respiration to provide energy and precursors for cell growth, leading to rapid increases in cell density and OUR, accompanied by a rapid decrease in DO. The specific growth rate (μ) at 20 h is 0.1 (Fig. S7). After 20 h, DO drops to zero, and the fermentation enters a hypoxic state. Hypoxia suppresses the biosynthesis of energy and precursors, thereby inhibiting cell growth, with the μ decreasing to 0.03 at 28 h (Fig. S7). In the meantime, Coenzyme Q₁₀ titer begins to increase rapidly, with the biosynthesis rate rising from 6.9 mg/L/h at 18h to 52.6 mg/L/h at 36 h (Fig. S8). Fermentation shifts from growth to biosynthesis. Therefore, hypoxia serves as a signal for the initiation of Coenzyme Q₁₀ biosynthesis and accumulation. However, the underlying mechanism remains unclear.

Fig. 1.

Profiles of physiological and metabolic parameters during R. sphaeroides fermentation.

3.2. Global transcriptional changes stimulated by hypoxia

To explore the mechanism by which R. sphaeroides biosynthesizes and accumulates Coenzyme Q₁₀ under hypoxia, transcriptomic studies were conducted by sampling at different stages (DO-20, Hypoxic-4 h, and Hypoxic-24 h). Fig. 2 and Table 1 illustrate the differentially expressed genes (DEGs) at various stages of hypoxia. DEGs were defined as genes exhibiting expression level changes greater than a two-fold difference and a p-value less than 0.05. Compared to DO-20, there were 653 and 582 upregulated genes and 829 and 1441 downregulated genes after 4 h and 24 h of hypoxia, respectively. Additionally, compared to Hypoxic-4h, Hypoxic-24h showed 124 upregulated genes and 551 downregulated genes. R. sphaeroides responds to changes in oxygen levels by adjusting gene expression, a common mechanism used by microorganisms to adapt to hypoxic conditions [17].

Fig. 2.

The volcano plot of DEGs (A: Hypoxic-4 h VS DO-20, B: Hypoxic-24 h VS DO-20, C: Hypoxic-24 h VS Hypoxic-4 h).

Table 1.

Statistics of DEGs before and after hypoxia.

Group	Total DEGs	Up-regulated gene	Down-regulated gene
Hypoxic-4 h VS DO-20	1482	653	829
Hypoxic-24 h VS DO-20	2023	582	1441
Hypoxic-24 h VS Hypoxic-4 h	675	124	551

Further, DEGs were subjected to KEGG enrichment analysis (Figs. S1–6). Compared with DO-20, the up-regulated DEGs at Hypoxic-4 h and Hypoxic-24 h were mainly enriched in metabolic pathways including ABC transport, two-component systems, and oxidative phosphorylation. The down-regulated DEGs at Hypoxic-4 h and Hypoxic-24 h were mainly enriched in metabolic pathways including pentose phosphate pathway (PPP), tricarboxylic acid cycle (TCA), and biosynthesis of ubiquinone and other terpenoid quinones. Additionally, compared with Hypoxic-4 h, the up-regulated DEGs at Hypoxic-24 h were mainly enriched in metabolic pathways including cysteine and methionine metabolism, sulfur metabolism, and lipopolysaccharide biosynthesis. The down-regulated DEGs were mainly enriched in metabolic pathways including biosynthesis of ubiquinone and other terpenoid quinones, glycolysis (EMP), and ABC transport. In summary, the downregulation of genes encoding central carbon metabolic (CCM) pathways in R. sphaeroides after hypoxia affected the supply of energy, precursors, and cofactors, thereby inhibiting the growth of R. sphaeroides, which is consistent with the physiological parameter showing a decrease in R. sphaeroides μ after hypoxia. Notably, the downregulation of genes in the Coenzyme Q₁₀ biosynthesis pathway is inconsistent with the physiological parameter showing accumulation of Coenzyme Q₁₀ after hypoxia. At the same time, genes in the lipopolysaccharide biosynthesis pathway, which is related to cell morphology, were up regulated.

3.3. DEGs related to central carbon metabolism under hypoxia

Oxygen plays a critical role in CCM as the terminal electron acceptor, driving aerobic respiration, enhancing the TCA, and maximizing the production of ATP, reducing power, and precursors. However, hypoxia suppresses CCM, negatively impacting growth and biosynthesis.

Fig. 3 shows the changes in CCM gene expression in R. sphaeroides after hypoxia. Genes involved in EMP, TCA, and PPP were significantly down-regulated after hypoxia. In the EMP, compared with DO-20, the glk (glucokinase) gene was down-regulated 3-fold and 3.3-fold at Hypoxic-4 h and Hypoxic-24 h, respectively. glk catalyzes the conversion of glucose to glucose-6-phosphate, which requires ATP. The reduction in ATP synthesis after hypoxia inhibited glk. With the inhibition of glucose-6-phosphate biosynthesis, the PPP, which branches from fructose-6-phosphate, was also inhibited. Compared with DO-20, the tkt gene was down-regulated 1.5-fold at Hypoxic-24 h. At the same time, hypoxia also inhibited the activity of pyruvate dehydrogenase complex. The genes pdhB and pdhA, encoding its biosynthesis, were down-regulated 3-fold and 2.4-fold at Hypoxic-24 h, respectively. Since the pyruvate dehydrogenase complex catalyzes the conversion of pyruvate to acetyl-CoA, the inhibition of acetyl-CoA biosynthesis further inhibited the TCA, with almost all genes in the pathway being down-regulated. For instance, compared with DO-20, acnA and aarC were down-regulated 2.3-fold and 2-fold at Hypoxic-24 h, respectively.

Fig. 3.

DEGs related to CCM under different hypoxic time. Heatmap indicate the gene expression level. Glucose-6-P, Glucose-6-phosphate; Frucose-6-P, Fructose-6-phosphate; Glyceraldehyde-3P, glyceraldehye-3-phosphate; Glycerate-3P, 3-Phosphoglycerate; α-KG, α-Ketoglutaric acid; OAA, Oxaloacetate; Xylulose-5-P, Xylulose-5-phosphate; Ribulose-5P, Ribulose-5-phosphate; Sedoheptulose-7P, Sedoheptulose-7-phosphate; Erythrose-5-P, Erythrose-5-phosphate.

In summary, hypoxia inhibited R. sphaeroides CCM, thereby inhibiting cell growth, and R. sphaeroides shifted from growth to Coenzyme Q₁₀ biosynthesis. However, d-erythrose 4-phosphate, pyruvate, phosphoenolpyruvate, glyceraldehyde-3-phosphate and NADPH in the CCM are direct precursors and reducing power for Coenzyme Q₁₀ biosynthesis. Existing studies have shown that CCM regulation directly affects Coenzyme Q₁₀ biosynthesis. Xiao et al. found that Coenzyme Q₁₀ production was higher when using CaCO₃ to regulate pH compared to using NH₃·H₂O. Metabolic flux analysis revealed that more glucose-6-phosphate entered the PPP when using CaCO₃ to regulate pH, producing more NADPH and promoting Coenzyme Q₁₀ biosynthesis [18]. Zhang et al. enhanced Coenzyme Q₁₀ biosynthesis by overexpressing glyceraldehyde-3-phosphate dehydrogenase in the EMP pathway [19]. Xu et al. enhanced Coenzyme Q₁₀ biosynthesis by overexpressing zwf and gnd genes in the PPP to promote NADPH production [20].

3.4. DEGs related to coenzyme Q₁₀ biosynthesis under hypoxia

The main pathways for Coenzyme Q₁₀ biosynthesis in R. sphaeroides include the shikimate pathway, MEP pathway, and ubiquinone pathway. The shikimate pathway begins with the condensation of d-erythrose 4-phosphate and phosphoenolpyruvate, ultimately generating para-hydroxybenzoic acid (PHB) under the catalysis of chorismate lyase (UbiC). PHB is a direct precursor for ubiquinone biosynthesis, forming different types of ubiquinones by combining with isoprenoid side chains of varying lengths. Isoprenoid side chains are biosynthesized through the MEP pathway, where pyruvate and glyceraldehyde-3-phosphate are converted into isopentenyl pyrophosphate (IPP) by a series of enzymes, and the isoprenoid chain is elongated under the catalysis of farnesyl diphosphate synthase (ipsA), ultimately forming decaprenyl diphosphate for ubiquinone biosynthesis. Decaprenyl diphosphate and PHB are ultimately converted into Coenzyme Q₁₀ under the action of a series of quinone ring modification enzymes.

Following hypoxia, most genes in the shikimate, MEP, and ubiquinone pathways in R. sphaeroides were downregulated (Fig. 4). In the shikimate pathway, compared to DO-20, chorismate synthase (aroC), shikimate dehydrogenase (aroE), and shikimate kinase (aroK) were downregulated by 3.5-fold, 1.6-fold, and 0.9-fold, respectively, at Hypoxic-24 h. The steps catalyzed by aroE and aroK require NADPH and ATP, suppression of CCM may explain the downregulation of these genes. Interestingly, aroQ in the shikimate pathway was upregulated after hypoxia, it was undetectable at DO-20 but showed upregulation at 4 and 24 h of hypoxia, possibly explaining the initiation of Coenzyme Q₁₀ accumulation. Studies have reported that R. sphaeroides can provide essential precursors via the oxygen-independent shikimate pathway after the inhibition of CCM [21]. However, the downregulation of other genes may hinder PHB biosynthesis, thus affecting Coenzyme Q₁₀ yield. Dixson et al. found that adding PHB increased Coenzyme Q₁₀ yield by 8-fold [22]. Simultaneously, most genes in the MEP pathway were also downregulated under hypoxic conditions. This may also be due to the suppression of CCM, which reduced the production of pyruvate, glyceraldehyde-3-phosphate, ATP, and reducing power. Previous studies identified these genes as rate-limiting enzymes for metabolic flux, and their overexpression has been proven to enhance Coenzyme Q₁₀ yield [23]. For instance, Kim et al. overexpressed the dxs gene from Pseudomonas aeruginosa in Escherichia coli, doubling Coenzyme Q₁₀ yield [24]. Similarly, Cluis et al. found that overexpressing genes in the shikimate pathway combined with increased isoprenoid precursor supply boosted Coenzyme Q₁₀ yield by 2.6-fold [25]. Thus, the downregulation of MEP pathway genes after hypoxia may also hinder Coenzyme Q₁₀ biosynthesis.

Fig. 4.

DEGs related to Coenzyme Q₁₀ biosynthesis pathway. Heatmap indicate the gene expression level. SAM, S-adenosyl-l-methionine; SAH, S-adenosyl-l-homocysteine.

The final steps involve PHB and decaprenyl pyrophosphate forming 4-hydroxy-3-polyprenylbenzoate under the catalysis of 4-hydroxybenzoate polyprenyltransferase (UbiA), followed by entry into the ubiquinone pathway. Compared to DO-20, genes such as UbiG, UbiD, UbiA, UbiE, and UbiF were persistently downregulated after hypoxia, with respective reductions of 4.7-fold, 3.5-fold, 2.1-fold, 1.6-fold, and 1.4-fold at Hypoxic-24 h. Quinone modification involves three oxygen-dependent reactions, suggesting that hypoxia is a major factor in gene downregulation. UbiA, UbiG, and UbiE have also been identified as bottleneck enzymes affecting Coenzyme Q₁₀ production [26,27]. Their expression levels directly influence Coenzyme Q₁₀ yield. Zhang et al. found that overexpressing UbiA increased Coenzyme Q₁₀ production in E. coli [27]. Lu et al. achieved a 157% increase in Coenzyme Q₁₀ titer by co-expressing UbiG, UbiE, and UbiH in R. sphaeroides [26]. Additionally, studies have reported the existence of an oxygen-independent Coenzyme Q₁₀ biosynthesis pathway in R. sphaeroides [8], ensuring Coenzyme Q₁₀ biosynthesis under hypoxic conditions, catalyzed by UbiT, UbiU, and UbiV [28]. However, these genes were also downregulated following hypoxia.

In conclusion, the expression of aroQ under hypoxia may serve as a trigger for Coenzyme Q₁₀ accumulation in R. sphaeroides. However, the overall downregulation of biosynthesis pathway genes after hypoxia, which is unfavorable for Coenzyme Q₁₀ biosynthesis and inconsistent with the continuous increase in titer. In a separate study, E. adhaerens synthesizes and accumulates VB₁₂ under hypoxic conditions, with most genes involved in the biosynthetic pathway being upregulated [29]. For R. sphaeroides, gene expression levels in the Coenzyme Q₁₀ biosynthesis pathway may not be the rate-limiting step. Zhang et al. found that overexpressing idi, dxs, UbiC, UbiA, UbiF, UbiH, UbiE, and UbiG in R. sphaeroides did not increase Coenzyme Q₁₀ yield [10]. Additionally, overexpression of UbiB, UbiG, and UbiH in E. coli showed limited effects on Coenzyme Q₁₀ production. As Coenzyme Q₁₀ is a lipid-soluble molecule embedded in cell membranes and stored intracellularly [30]. Moreover, hypoxia induces morphological changes in R. sphaeroides [31,32]. Therefore, cell morphology and membrane area may influence its accumulation. Our previous studies also revealed a correlation between Coenzyme Q₁₀ biosynthesis, cell morphology, and oxygen supply [6].

3.5. DEGs related to cell morphology under hypoxia

Cells undergo morphological changes in response to external stress, and a close relationship between cell morphology and production yield always demostrated [33]^. Three critical factors related to cell morphology: the cell wall, the cytoskeleton, and cell division. The main component of the cell wall is peptidoglycan, whose rigidity determines cell shape and size. Reduced peptidoglycan biosynthesis weakens the cell wall, promoting the increase in cell volume and the accumulation of intracellular metabolites [34]. Peptidoglycan biosynthesis occurs in three stages: (1) biosynthesis of N-acetylmuramic acid pentapeptide through six “Mur” proteins (murA–murF); (2) catalysis of N-acetylmuramic acid pentapeptide and N-acetylglucosamine to form peptidoglycan monomer disaccharide-pentapeptide subunits, mediated by mraY and murG; and (3) assembly of peptidoglycan catalyzed by penicillin-binding proteins (PBPs). Compared to DO-20, the expression of murD was undetectable after 4 and 24 h of hypoxia. Additionally, pbpC was downregulated by 0.8-fold and 1.2-fold at 4 and 24 h of hypoxia, respectively (Table 2). PBPs are essential enzymes for peptidoglycan biosynthesis and modification, inhibiting PBPs causes cells to become spherical [35]. Therefore, R. sphaeroides downregulates genes in the peptidoglycan biosynthesis pathway, reducing cell wall rigidity and decreasing resistance to cell volume expansion, thereby providing more space for Coenzyme Q₁₀ accumulation. Zhang et al. showed that downregulating murD in E. coli increased cell volume and enhanced PHB accumulation by more than 90% [36].

Table 2.

DEGs related to cell morphology and membrane under hypoxia.

Gene	Description	Function metabolism	Log2FC
			DO-20/Hypoxic-4 h	DO-20/Hypoxic-24 h
murD	UDP-N-acetylmuramoyl-l-alanine-d-glutamate ligase	Peptidoglycan biosynthesis	ND	ND
murE	UDP-N-acetylmuramoyl-l-alanyl-d-glutamate-2,6-diaminopimelate ligase	Peptidoglycan biosynthesis	0.6	−0.5
murF	UDP-N-acetylmuramoyl-tripeptide-d-alanyl-d-alanine ligase	Peptidoglycan biosynthesis	0.2	−0.3
mraY	Phospho-N-acetylmuramoyl-pentapeptide-transferase	Peptidoglycan biosynthesis	0.3	−0.2
PBPs	Penicillin-binding protein	Peptidoglycan biosynthesis	−0.8	−1.2
MreB	Rod shape-determining protein MreB and related proteins		0.3	−1.2
FtsZ	Cell division protein	Cell growth and death	−1.4	−1.2
NtrX	Two-component system response regulator	Environmental Information Processing	ND	ND
NtrY	Two-component system sensor histidine kinase	Environmental Information Processing	ND	ND
Clsc	Phospholipase D family protein	Lipid metabolism	0.7	1.1
pgsA	CDP-diacylglycerol-glycerol-3-phosphate 3-phosphatidyltransferase	Lipid metabolism	1	1.9
pssA	CDP-diacylglycerol-serine O-phosphatidyltransferase	Lipid metabolism	0.7	0.5
fabI	Enoyl-ACP reductase FabI	Fatty acid biosynthesis	−1.8	−2.0
fabB	Beta-ketoacyl-ACP synthase I	Fatty acid biosynthesis	0.6	1

Note: FC (Fold change) = FPKM (DO-20)/FPKM (Hypoxic-4 h), FPKM (DO-20)/FPKM (Hypoxic-24 h).

FPKM: fragments per kilobase of transcript per million fragments mapped.

ND, not defined, because these genes transcript level were zero at hypoxic-4 h and hypoxic-24 h.

In addition to peptidoglycan, the cytoskeletal protein MreB and the cell division protein FtsZ are crucial for maintaining cell morphology [37]. MreB affects cell width and length. Knocking out MreB increases cell volume and makes cells spherical. FtsZ, the key enzyme regulating the cell division, forms a dynamic Z-ring at the cell midpoint to influence cell division and peptidoglycan biosynthesis [38]. Inhibition or deletion of FtsZ impedes cell division, increasing cell volume, while overexpression of FtsZ reduces cell size. Compared to DO-20, MreB and FtsZ in R. sphaeroides were downregulated 1.2-fold after 24 h of hypoxia (Table 2). Concurrent inhibition of MreB and FtsZ following hypoxia further altered R. sphaeroides morphology, increasing cell volume. MreB and FtsZ are common targets for morphological remodeling. In E. coli, deletion of MreB caused rod-shaped cells to become spherical, and partial suppression of MreB increased cell volume by 62%, enhancing PHB yield by 90% [39]. In Bacillus amyloliquefaciens, inhibiting MreB expression enlarged cell volume and increased poly-γ-glutamic acid yield by 55.7% [40]. In Halomonas campaniensis LS21, inhibiting FtsZ expression elongated cells and increased PHB accumulation [41]. Elhadi et al. demonstrated that reducing FtsZ and MreB expression in E. coli elongated and widened cells. The stronger the inhibition of FtsZ and MreB, the longer and larger the cells became, leading to higher PHB accumulation [42]. Although MreB and FtsZ are key genes influencing cell morphology and product accumulation, their downregulation reduces cell growth rate [43], which may explain the decreased μ of R. sphaeroides under hypoxia. Furthermore, Kimberly et al. identified NtrXY as a key regulatory gene for R. sphaeroides morphology. Deletion of NtrYX inhibited the expression of genes related to peptidoglycan biosynthesis and cell division [44]. Interestingly, NtrYX was normally expressed at DO-20 but undetectable after 4 and 24 h of hypoxia (Table 2). Changes in its expression may explain the downregulation of genes in the cell wall, cytoskeleton, and cell division pathways. In conclusion, the downregulation of morphology-related genes in R. sphaeroides under hypoxic conditions altered cell morphology, increasing cell volume and providing more space for Coenzyme Q₁₀ accumulation.

3.6. DEGs related to cell membrane under hypoxia

Changes in bacterial cell morphology can affect membrane expansion, and membrane composition can, in turn, influence cell morphology [45]. As the main site for biosynthesis, storage, and transport in prokaryotic organisms, the cell membrane plays a key role in productivity. Membrane engineering has also been proven feasible and widely applicable in industrial applications [37]. The cell membrane consists of the inner and outer membranes, with the membranes of Gram-negative bacteria primarily composed of a lipid bilayer and membrane proteins [34,46]. Phospholipids constitute the major component of the lipid bilayer, making them essential for determining cell morphology and function [47]. Genes regulating phospholipid biosynthesis include phosphatidylglycerol synthase A (pgsA), cardiolipin synthase (clsC), 1-acyl-glycerol-3-phosphate acyltransferase (plsC), and phosphatidylserine synthase A (pssA). Compared to DO-20, the expression of clsC, pgsA, and pssA increased by 1.9-fold, 1.1-fold, and 0.5-fold, respectively, after 24 h of hypoxia (Table 2). Research has shown that reduced cardiolipin levels contribute to the spherical morphology of R. sphaeroides, and deletion of the clsC gene decreases cell size [48]. In our previous studies, R. sphaeroides displayed regular spherical morphology before hypoxia but exhibited irregular expansion afterward [6], which may be associated with increased phospholipid biosynthesis. Other studies found that overexpressing clsC increased the specific yield of Coenzyme Q₁₀ by 15.01% (unpublished data).

Meanwhile, Kimberly et al. demonstrated that deletion of the NtrYX gene upregulates lipid biosynthesis-related genes under hypoxic conditions, leading to increased lipid accumulation and enhanced cell membrane thickness [49]. Niederman's study also revealed that hypoxia induces cell membrane remodeling in R. sphaeroides [50], while Yoshida et al. reported thicker membranes and higher Coenzyme Q₁₀ yields under hypoxia [12]. Therefore, changes in cell morphology following hypoxia create more space for membrane biosynthesis, while upregulation of phospholipid biosynthesis-related genes further promotes phospholipid production, increasing membrane area. Since Coenzyme Q₁₀ is primarily stored in the cell membrane, increased membrane area provides more space for product accumulation [40]. Furthermore, as Coenzyme Q₁₀ is lipid-soluble, increased phospholipid biosynthesis enhances its incorporation into the membrane, further promoting its accumulation. Upregulation of membrane phospholipid biosynthesis genes in E. coli increased β-carotene titer by 55% [51].

In addition to phospholipids, saturated and unsaturated fatty acid (UFA) are key components of the phospholipid bilayer. UFA influence membrane fluidity and curvature. Insufficient UFA content increases membrane rigidity, affecting cell morphology and growth [52]. Kim et al. found that fabI overexpression reduced UFA levels, while fabB overexpression increased them [53]. Compared to DO-20, fabI expression decreased by 1.8-fold and 2-fold at 4 and 24 h of hypoxia, respectively, while fabB expression increased by 0.6-fold and 1-fold (Table 2). Under hypoxia, R. sphaeroides enhanced UFA biosynthesis to increase membrane flexibility, accommodating morphological changes and meeting the demand for UFA due to increase membrane area. Kimberly's study also revealed that hypoxia increased total fatty acid content in R. sphaeroides, with the highest proportion of UFA [49]. Thus, by enhancing UFA biosynthesis, R. sphaeroides adapts to morphological changes and membrane expansion under hypoxia, further facilitating Coenzyme Q₁₀ accumulation. Adjusting fatty acid unsaturation and increasing UFA content has been widely used to improve fermentation yields, particularly for terpenoid-related products [[54], [55], [56]]. For instance, increasing UFA content in the cell membranes of Yarrowia lipolytica boosted lupeol yield 33-fold, achieving an organic phase titer of 381.67 mg/L [57]. However, studies focusing on enhancing Coenzyme Q₁₀ yield by adjusting UFA content in R. sphaeroides have not yet been reported.

3.7. R. sphaeroides morphology in relation to coenzyme Q₁₀ and oxygen supply

Transcriptomic studies have revealed that after hypoxia, genes related to peptidoglycan biosynthesis, cytoskeleton, and cell division in R. sphaeroides are downregulated, while genes related to phospholipid and UFA biosynthesis are upregulated. It is proposed that a relationship exists between the morphological changes and the biosynthesis and accumulation of Coenzyme Q₁₀ in R. sphaeroides following hypoxia, as R. sphaeroides increases cell volume and membrane area to provide additional space for Coenzyme Q₁₀ accumulation. To verify this, we further investigated the relationship between morphological changes and Coenzyme Q₁₀ biosynthesis in R. sphaeroides during fermentation. Fig. 5a shows that the cell area of R. sphaeroides did not change before 20 h (pre-hypoxia), but as the fermentation entered the hypoxic conditions (after 20 h), the cell area began to rapidly increase, eventually reaching 7368 pixels, approximately 9 times larger than the original area (797.2 pixels). A linear regression was further established to examine the relationship between Coenzyme Q₁₀ yield and cell area throughout the fermentation process (Fig. 5b), with a significant correlation (R² = 0.9695). Therefore, the morphological expansion of R. sphaeroides after hypoxia increases the storage space for Coenzyme Q₁₀, promoting continuous increases in its yield, consistent with the transcriptomic results.

Fig. 5.

Profiles of cell morphology in relation to Coenzyme Q₁₀ and oxygen supply. (a) Changing in cell area during fermentation of R. sphaeroides; (b) The relationship between Coenzyme Q₁₀ and cell area during the fermentation process. (c) OUR under different oxygen supplies during fermentation. (d) Cell area under different oxygen supplies during fermentation; (e) Coenzyme Q₁₀ titer under different oxygen supplies during fermentation.

Additionally, previous studies have found significant differences in Coenzyme Q₁₀ yield at different hypoxia levels [13]. We further studied the relationship between Coenzyme Q₁₀ biosynthesis and cell morphology under different oxygen supply levels (Fig. 5c). The high-oxygen group (650 rpm) had the lowest final yield, which was 44% and 53% lower than the medium-oxygen group (550 rpm) and low-oxygen group (500 rpm), respectively (Fig. 5e). Moreover, the final cell area of the high-oxygen group was the smallest (Fig. 5d). Therefore, oxygen supply level affects the morphological changes in R. sphaeroides, which in turn influences Coenzyme Q₁₀ biosynthesis. Yoshida's study also found that under lower oxygen supply, the yield of R. sphaeroides increased 3.6 times compared to the high-oxygen group, and the membrane area doubled [12]. Interestingly, the Coenzyme Q₁₀ yield and cell area of the low-oxygen group were consistently lower than those of the medium-oxygen group throughout the process. This suggests that other factors also influence cell morphology and yield. In summary, significant differences in cell size and Coenzyme Q₁₀ biosynthesis were observed under different hypoxia levels, with a strong correlation between the two, as larger cells were associated with higher yield. Therefore, employing cell morphology as a key indicator for optimizing and controlling Coenzyme Q₁₀ production is a feasible strategy. Wang et al. increased VB₁₂ yield by 14.7% through morphological control of E. adhaerens [51].

3.8. Effect of UFA on coenzyme Q₁₀ biosynthesis

After hypoxia, R. sphaeroides accumulates Coenzyme Q₁₀ by increasing its morphology and cell membrane area. UFA, as essential components of the cell membrane, are crucial for this process. The key gene involved in UFA biosynthesis, fabB, is continuously upregulated under hypoxia, satisfying the increased demand for UFA due to the enlarged cell membrane area. However, since the biosynthesis of UFA requires ATP, NADPH, and oxygen, excessive hypoxia may influence UFA biosynthesis by affecting ATP and NADPH production. This could be one of the reasons why Coenzyme Q₁₀ yield was lower under low oxygen supply compared to medium oxygen supply. However, gram-negative bacteria can assimilate exogenous UFA in addition to biosynthesizing them internally, altering their membrane phospholipid composition and thereby affecting cell morphology [58]. In R. sphaeroides, the addition of exogenous UFA can also increase the intracellular UFA content [53]. This study investigated the impact of exogenous UFA addition (3 mL/L) on Coenzyme Q₁₀ biosynthesis. The results (Fig. 6A) showed that the Coenzyme Q₁₀ yield after 48 h in the groups supplemented with soybean oil, peanut oil, corn oil, and palm oil were higher than the control group, with soybean oil having the best effect, increasing the yield by about 28% compared to the control group. This suggests that exogenous UFA addition can enhance Coenzyme Q₁₀ biosynthesis in R. sphaeroides, with soybean oil being the most effective. In the fermentation of Coenzyme Q₁₀ in Rhodotorula glutinis, the addition of PHB and soybean oil also increased the yield from 10 mg/L to 39.2 mg/L [59], and in the spinosad fermentation process, soybean oil increased the yield from 285.76 mg/L to 398.11 mg/L [60]. Further investigations in bioreactor examined the impact of soybean oil on the Coenzyme Q₁₀ biosynthesis of R. sphaeroides (Fig. 6B). The soybean oil addition group had a higher Coenzyme Q₁₀ biosynthesis rate, with the maximum biosynthesis rate (57.1 mg/L/h) being 9% higher than the control group (52.43 ± 1.5 mg/L/h). The final Coenzyme Q₁₀ yield reaching 2780 mg/L, an 18.9% increase compared to the control group. Therefore, the addition of soybean oil enhances Coenzyme Q₁₀ biosynthesis, with the mechanism possibly involving its influence on cell membrane biosynthesis.

Fig. 6.

Effect of UFA on Coenzyme Q₁₀ biosynthesis (A) Effect of UFA on growth and Coenzyme Q₁₀ biosynthesis in shake flasks; (B) The difference of physiological metabolic parameters in fermentation process between 3 mL/L soybean oil group and no soybean oil group.

4. Discussion

After oxygen limitation, R. sphaeroides initiates the biosynthesis and accumulation of Coenzyme Q₁₀. Transcriptome analysis revealed that CCM was inhibited following oxygen limitation, suppression of precursor and reducing power generation for Coenzyme Q₁₀ biosynthesis. Additionally, oxygen serves as a direct substrate for Coenzyme Q₁₀ biosynthesis. Consequently, the biosynthesis pathway of Coenzyme Q₁₀ was inhibited after hypoxia. However, physiological parameters indicated that Coenzyme Q₁₀ continued to be biosynthesized and accumulated rapidly under hypoxic conditions, suggesting that the biosynthesis pathway may not be rate-limiting. Zhang et al. demonstrated that overexpression of genes involved in precursor biosynthesis and the ubiquinone pathway in R. sphaeroides did not enhance Coenzyme Q₁₀ biosynthesis [10]. Similarly, overexpression of UbiB, UbiG, and UbiH in E. coli resulted in only a modest increase in Coenzyme Q₁₀ production [5].

Given that Coenzyme Q₁₀ is a lipophilic molecule embedded within the cell membrane and stored intracellularly [30], its biosynthesis and accumulate may be closely linked to cell morphology. Wang et al. established an OUR-stat strategy by adjusting oxygen supply and adding ammonium sulfate, which increased Coenzyme Q₁₀ yield by 15.4% compared to the control group alongside increased cell size [13]. Zhang et al. observed that the high-yield mutant R. sphaeroides VK-2-3 strain, obtained through mutagenesis, exhibited significant morphological changes characterized by irregular shape and rough surface, which may account for its enhanced Coenzyme Q₁₀ production [9]. Transcriptome analysis revealed that following hypoxia, genes related to cell wall biosynthesis, cytoskeleton organization, and cell division were downregulated (e.g., murD, mreB, ftsZ), reducing resistance to morphological changes in R. sphaeroides. Concurrently, genes involved in phospholipid and UFA biosynthesis were upregulated (e.g., clsC, fabB), promoting cell membrane biosynthesis and expansion. Previous studies have demonstrated that inhibition of murD, mreB, and ftsZ increases cell volume [33], while clsC serves as a key gene affecting R. sphaeroides morphology, knockout of clsC resulted in smaller cell morphology [48], whereas overexpression increased Coenzyme Q₁₀ specific yield by 15.01% (Unpublished data). Similarly, fabB overexpression was shown to elevate UFA content in R. sphaeroides [53]. Collectively, R. sphaeroides may increase cell volume and membrane area to provide additional space for Coenzyme Q₁₀ biosynthesis and accumulation following hypoxia. Additionally, Kimberly et al. demonstrated that morphology associated genes (murD, MreB, FtsZ, and ClsC) are regulated by the NtrYX. Deletion of NtrYX suppresses expression of peptidoglycan biosynthesis and cell division genes while enhancing lipid biosynthesis related gene expression, leading to increased lipid accumulation and elevated membrane thickness [44]. In R. sphaeroides, NtrYX exhibited normal expression at DO-20, but its expression dropped to zero after hypoxic-4h and hypoxic-24h, suggesting this regulatory shift may underlie the observed changes in morphology associated gene expression following hypoxia. Building on this, the morphological changes throughout the fermentation process were further investigated. As Coenzyme Q₁₀ biosynthesis commenced and accumulated, cell size increased concomitantly, with a statistically significant positive correlation between cellular morphology and Coenzyme Q₁₀ yield (R² = 0.9695). Moreover, under varying hypoxic conditions, cell morphology and Coenzyme Q₁₀ yield also exhibited strong correlations, higher yields were associated with larger cell volumes. Yoshida et al. similarly observed that R. sphaeroides developed thicker cell membranes under hypoxia while simultaneously producing higher levels of Coenzyme Q₁₀ [12]. These results verify the transcriptome findings, indicating that morphological changes in R. sphaeroides represent a potential mechanism for enhanced Coenzyme Q₁₀ biosynthesis and accumulation following hypoxia. Specifically, R. sphaeroides increases cell volume and membrane area to provide additional space for Coenzyme Q₁₀ biosynthesis and accumulate.

In prokaryotes, energy metabolism is localized to the cytoplasmic membrane, where the electron transport processes of the respiratory chain occur. An increased membrane surface area provides a greater capacity to accommodate components of the energy metabolizing machinery, thereby facilitating higher electron transport flux [61]. Under hypoxic conditions, many prokaryotes are induced to form ICM, effectively expanding the total membrane surface area [62]. This expansion allows for the incorporation of additional electron transport complexes (e.g., quinone pools), which enhances electron flux, promotes ATP biosynthesis, and mitigates energy deficits associated with oxygen limitation [50]. Consequently, R. sphaeroides may adapt to hypoxia by enlarging cell morphology and membrane surface area, thereby enabling the accumulation of more Coenzyme Q₁₀ to further augment electron transport flux and sustain ATP generation. Finally, as an essential component of the cell membrane, the biosynthesis of UFA is influenced by hypoxia, which may, in turn, affect membrane biosynthesis and consequently the accumulation of Coenzyme Q₁₀. Given that R. sphaeroides can utilize exogenous UFA to increase intracellular UFA content, this study demonstrated that supplementation with exogenous UFA enhanced Coenzyme Q₁₀ biosynthesis, resulting in an 18.9% increase in yield.

5. Conclusion

This study elucidates the potential mechanism of Coenzyme Q₁₀ biosynthesis and accumulation in R. sphaeroides under hypoxic conditions. Transcriptomic analysis revealed that, after hypoxia, genes involved Coenzyme Q₁₀ biosynthesis pathways were downregulated. In contrast, genes related to cell morphology inhibition were downregulated. Conversely, genes associated with cell membrane biosynthesis were upregulated. Moreover, a significant relationship was observed between cell morphology and Coenzyme Q₁₀ yield. Under hypoxia, R. sphaeroides may enhance Coenzyme Q₁₀ accumulation by increasing cell volume and cell membrane area, rather than enhancing its biosynthesis. Based on the transcriptomic results, the further enhancement of Coenzyme Q₁₀ biosynthesis through exogenous addition of UFA increased the yield to 2780 mg/L.

CRediT authorship contribution statement

Bo Li: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation. Yan Ge: Writing – review & editing, Writing – original draft, Data curation. Li Fu: Writing – review & editing. Huanan Guo: Writing – review & editing. Ali Mohsin: Writing – review & editing. Junming Li: Writing – review & editing. Jiequn Wu: Writing – review & editing. Biqin Chen: Writing – review & editing. Yingping Zhuang: Writing – review & editing, Supervision. Zejian Wang: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Funding acquisition.

Data availability statement

Data will be made available on request.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Biqin Chen is currently employed by Inner Mongolia Kingdomway Pharmaceutical Company.

Appendix. ASupplementary data

The following is the Supplementary data to this article: