Development of a High-Titer Glucose-Based Fermentation Process for Di-Rhamnolipid-Rich Rhamnolipid Production by Pseudomonas aeruginosa YM4

Abstract

Rhamnolipids are promising biosurfactants with diverse applications, but their efficient production from glucose remains challenging. In this study, a glucose-based fermentation process was developed using Pseudomonas aeruginosa YM4 through sequential evaluation of nitrogen source, pH, C/N ratio, glucose concentration, and Fe²⁺ supplementation in a 5 L bioreactor. A mixed NH₄⁺/NO₃⁻ nitrogen source resulted in higher rhamnolipid production than either NH₄⁺ or NO₃⁻ alone. Maintaining the culture pH at 7.2 prevented excessive acidification and promoted glucose consumption and rhamnolipid production. The highest rhamnolipid titer (29.7 g/L) was achieved using 100 g/L glucose, a C/N ratio of 35, and 10.8 μM Fe²⁺. Fermentation kinetics indicated that biomass accumulated rapidly during the early fermentation stage, whereas rhamnolipid production continued after cell growth had slowed. Under the selected conditions, YM4 produced a rhamnolipid mixture containing 73.8% di-rhamnolipids and showed substantially higher rhamnolipid production than three additional P. aeruginosa strains cultivated under identical conditions. This study developed a glucose-based fermentation process for producing di-rhamnolipid-rich products and demonstrated the efficient rhamnolipid-producing capability of P. aeruginosa YM4 under the evaluated conditions.

1. Introduction

Rhamnolipids (RLs), a group of glycolipid biosurfactants mainly produced by Pseudomonas strains, have attracted growing attention as renewable alternatives to synthetic surfactants due to their excellent surface activity, biodegradability, and low toxicity [1,2,3]. Owing to these properties, RLs have been explored for applications in petroleum recovery, environmental remediation, agriculture, cosmetics, and pharmaceuticals [2,3].

Structurally, RLs consist of one or two rhamnose molecules linked to one or two β-hydroxy fatty acids, forming mono-rhamnolipid and di-rhamnolipid congeners with different physicochemical properties and application characteristics [4,5]. Among these congeners, di-rhamnolipids generally exhibit stronger emulsifying properties and biological activities, increasing interest in the production of di-rhamnolipid-rich RL products [6,7,8].

Despite their promising applications, large-scale RL production remains constrained by high production costs and limited fermentation efficiency [9,10]. Among the factors influencing fermentation performance, the carbon source is particularly important because it affects cell growth, precursor availability, and RL composition [11,12]. Vegetable oils are commonly used substrates for RL production because they provide fatty acid precursors that support high product titers [8,12]. However, their poor water solubility can complicate substrate feeding, oxygen transfer, and downstream product recovery, increasing process complexity and production costs. In contrast, glucose is inexpensive and readily available, making it an attractive alternative carbon source for RL fermentation [12]. Consequently, glucose has attracted considerable interest as an alternative carbon source for RL production.

However, establishing an efficient glucose-based fermentation process remains challenging. Compared with lipid substrates, glucose often results in lower RL production, and fermentation performance varies among strains and cultivation conditions [12,13]. Previous studies have shown that nitrogen source, pH, nutrient balance, and trace metal supplementation all influence RL production [14,15,16,17,18]. Nevertheless, these factors have mostly been investigated independently, whereas their combined effects on glucose utilization, biomass formation, and RL production during controlled bioreactor fermentation remain poorly understood. Consequently, systematic process development for glucose-based RL production is still limited.

Pseudomonas aeruginosa YM4 is an RL-producing strain previously isolated in our laboratory [8]. In our previous study, YM4 produced more than 20 g/L RLs using soybean oil or glycerol as carbon sources and naturally produced di-rhamnolipid-rich RLs [8]. These characteristics suggest that YM4 may be suitable for further evaluation in glucose-based RL fermentation. Its performance during glucose-based fermentation, however, has not yet been systematically evaluated. Moreover, it is unknown whether refining the fermentation process can simultaneously achieve efficient glucose utilization, high RL productivity, and a di-rhamnolipid-rich product profile.

Therefore, this study aimed to establish a glucose-based fermentation process for P. aeruginosa YM4 by evaluating key fermentation factors affecting RL production. Nitrogen source and fermentation pH were first investigated to identify the major factors limiting glucose-based RL production. The C/N ratio, initial glucose concentration, and Fe²⁺ supplementation were subsequently evaluated in a 5 L bioreactor to improve fermentation performance. Fermentation kinetics were further analyzed to characterize the relationship between biomass formation and RL accumulation, and the RL composition produced by YM4 was compared with those of three additional P. aeruginosa strains. This study developed a glucose-based fermentation strategy for producing di-rhamnolipid-rich RLs and provides a basis for further development of glucose-based RL production processes.

2. Materials and Methods

2.1. Bacterial Strains and Seed Preparation

Four Pseudomonas aeruginosa strains, including YM4, PAO1, HN, and HEB, were maintained in our laboratory. P. aeruginosa YM4 (CCTCC No. M2017494) was generated by atmospheric and room-temperature plasma (ARTP) mutagenesis of a parental strain originally isolated from crude-oil-contaminated soil, as described previously [8]. P. aeruginosa PAO1, a well-characterized reference laboratory strain, was used as a comparative strain. Strains HN and HEB were previously isolated and screened in our laboratory for RL production and maintained in our collection. YM4 was used as the primary production strain, while PAO1, HN, and HEB were used for comparative evaluation of RL production and composition under glucose-based fermentation conditions.

LB medium was used for strain activation and seed preparation and contained 10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl. Bacterial stocks stored at −80 °C were streaked onto LB agar plates and incubated at 37 °C for 24 h. A single colony was inoculated into liquid LB medium and cultivated at 37 °C and 200 rpm for 12 h. For bioreactor cultivation, the seed culture was grown in LB medium until the OD₆₀₀ reached approximately 3–4 and was subsequently used as the inoculum.

2.2. Fermentation Media

The basal medium for shake-flask fermentation contained 80 g/L glucose, 7.888 g/L Na₂HPO₄·12H₂O, 1.972 g/L KH₂PO₄, 0.2 g/L MgSO₄·7H₂O, 0.1 g/L CaCl₂, 1 g/L NaCl, 1 g/L KCl, 0.5 g/L yeast extract, and 2 mL/L trace-element solution. The basal medium used for bioreactor cultivation contained 80 g/L glucose, 0.5 g/L KH₂PO₄, 0.2 g/L MgSO₄·7H₂O, 0.1 g/L CaCl₂, 1 g/L NaCl, 1 g/L KCl, 0.5 g/L yeast extract, and 2 mL/L trace-element solution.

The trace-element stock solution comprised 0.500 g/L FeSO₄·7H₂O, 0.250 g/L CuSO₄·5H₂O, 2.875 g/L ZnSO₄·7H₂O, and 1.691 g/L MnSO₄·H₂O and was added to the fermentation medium at 2 mL/L unless otherwise specified.

For nitrogen-source evaluation, the basal medium was supplemented with 37.5 mM (NH₄)₂HPO₄, 75 mM NaNO₃, 12 g/L soy peptone (FP410, Angel Yeast Co., Ltd., Yichang, China), or a mixed nitrogen source consisting of 18.75 mM (NH₄)₂HPO₄ and 37.5 mM NaNO₃.

The mixed nitrogen source supplied equal molar amounts of NH₄⁺-N and NO₃⁻-N. The total nitrogen concentration of the inorganic nitrogen source treatments was maintained at 75 mM, whereas the soy peptone concentration was selected based on the manufacturer-specified nitrogen content.

2.3. Shake-Flask Evaluation of Nitrogen Sources

Fermentation medium (50 mL) was inoculated with 2% (v/v) seed culture and incubated at 37 °C and 200 rpm for 96 h. Glucose-based RL production was then evaluated using (NH₄)₂HPO₄, NaNO₃, soy peptone, or the mixed ammonium/nitrate formulation described in Section 2.2 as nitrogen sources.

The initial glucose concentration was 80 g/L. At the end of cultivation, the RL titer, dry cell weight, residual glucose concentration, and culture pH were determined.

2.4. Bioreactor Cultivation

Fermentations were performed in a 5 L stirred-tank bioreactor (BioCore QF-05, Shanghai Chuyi Biotechnology Co., Ltd., Shanghai, China) with a working volume of 3 L. The cylindrical vessel had an inner diameter of 160 mm, a straight-section height of 240 mm, and a total height of 285 mm. Agitation was provided by a single six-blade straight turbine impeller with a diameter of 70 mm. The bioreactor was inoculated with 10% (v/v) seed culture and operated at 37 °C. Industrial-grade oxygen was supplied at 200 mL/min, and the agitation speed was maintained at 300 rpm. Dissolved oxygen was not actively controlled during fermentation. The pH was automatically controlled by the addition of 2 M NaOH or 1 M H₂SO₄. Silicone antifoam (1 g/L) was added as required to suppress excessive foaming [8]. Samples were collected at predetermined intervals to determine RL concentration, residual glucose concentration, and biomass. Unless otherwise specified, fermentation was terminated when the residual glucose concentration decreased below 5 g/L. For glucose concentration experiments in which glucose was not completely consumed, fermentation was terminated after 96 h.

2.5. Nitrogen-Source Evaluation Under Controlled pH Conditions

To compare the effects of nitrogen-source composition under controlled pH conditions, the four nitrogen-source formulations described in Section 2.2 were evaluated in the 5 L bioreactor. The initial glucose concentration was maintained at 80 g/L, and the fermentation pH was controlled at 7.0. The other cultivation conditions were maintained as described in Section 2.4.

RL concentration, dry cell weight, residual glucose concentration, fermentation time, and volumetric productivity were determined. Based on the overall fermentation performance, the mixed ammonium/nitrate formulation was used in subsequent experiments.

2.6. Effect of Fermentation pH

The effects of controlled pH on glucose utilization, biomass formation, and RL production were investigated in a 5 L bioreactor. Fermentation was carried out in medium containing 80 g/L glucose and the mixed ammonium/nitrate nitrogen source, with the pH maintained at 6.5, 7.0, 7.2, 7.4, and 7.6 in separate runs. Other cultivation conditions were maintained as described in Section 2.4.

2.7. Effect of the C/N Ratio

Different C/N ratios were evaluated for their effects on biomass formation and RL production using the mixed ammonium/nitrate nitrogen source. All fermentations were performed with an initial glucose concentration of 80 g/L at pH 7.2. The concentrations of (NH₄)₂HPO₄ and NaNO₃ were adjusted proportionally to obtain C/N ratios of 18, 29, 35, 41, and 53 while maintaining equal molar contributions of NH₄⁺-N and NO₃⁻-N. The C/N ratio was defined as the molar ratio of carbon atoms supplied by glucose to nitrogen atoms supplied by inorganic nitrogen sources.

2.8. Effect of the Initial Glucose Concentration

The effect of initial glucose concentration on RL production and glucose utilization was evaluated using initial glucose concentrations of 60, 80, 100, 120, and 150 g/L. For each glucose concentration, the concentrations of (NH₄)₂HPO₄ and NaNO₃ were adjusted proportionally to maintain a C/N ratio of 35, while maintaining equal molar contributions of NH₄⁺-N and NO₃⁻-N. The fermentation pH was controlled at 7.2, and the other cultivation conditions were maintained as described in Section 2.4. Cells collected at the end of fermentations containing 60 and 80 g/L initial glucose were additionally examined by transmission electron microscopy to compare intracellular inclusion formation.

2.9. Effect of Fe²⁺ Concentration

The effect of Fe²⁺ concentration on RL production was evaluated using a medium containing 100 g/L glucose. All fermentations were performed at pH 7.2. Fe²⁺ concentrations of 3.6, 7.2, 10.8, and 14.4 μM were obtained by adjusting the amount of FeSO₄·7H₂O, while the concentrations of the remaining trace elements were kept constant.

2.10. Fermentation Under the Selected Conditions

The selected nitrogen-source composition, pH, C/N ratio, glucose concentration, and Fe²⁺ concentration identified above were combined for fermentation under the selected conditions in the 5 L bioreactor. The final fermentation medium contained 100 g/L glucose, 23.5 mM (NH₄)₂HPO₄, 47 mM NaNO₃, 0.5 g/L KH₂PO₄, 0.2 g/L MgSO₄·7H₂O, 0.1 g/L CaCl₂, 1 g/L NaCl, 1 g/L KCl, 0.5 g/L yeast extract, and 2 mL/L trace-element solution. FeSO₄·7H₂O was adjusted to provide a final Fe²⁺ concentration of 10.8 μM, while the remaining trace elements remained unchanged. Fermentation was conducted at 37 °C with the pH controlled at 7.2. Samples were collected periodically to determine residual glucose, NH₄⁺, dry cell weight, and RL concentrations.

2.11. Comparison of RL Production Among Different Strains

To compare strain-dependent differences in glucose-based RL production, P. aeruginosa YM4, PAO1, HN, and HEB were cultivated using the selected fermentation medium described in Section 2.10. The fermentation pH was maintained at 7.2, and the other cultivation conditions were maintained as described in Section 2.4. The RL products produced by the four strains were further analyzed by high-performance liquid chromatography coupled with an evaporative light-scattering detector (HPLC-ELSD) to compare the proportions of mono-rhamnolipids and di-rhamnolipids.

2.12. Analytical Methods

RLs in the fermentation broth were quantified by HPLC-ELSD, as previously described by Li et al. [8]. Briefly, fermentation broth samples were centrifuged to remove cells. The resulting supernatants were appropriately diluted and filtered through a 0.22 μm membrane. HPLC-ELSD analysis was performed using an Agilent 1260 Infinity HPLC system (Agilent Technologies, Palo Alto, CA, USA) equipped with an evaporative light-scattering detector (ELSD 2000ES, Alltech Associates, Inc., Deerfield, IL, USA) and a C18 column (4.6 × 150 mm, 5 μm; Sepax Technologies, Suzhou, China). Acetonitrile and water containing 0.05% formic acid were used as the mobile phases, with a linear gradient of acetonitrile from 30% to 100% over 40 min at a flow rate of 1 mL/min. The ELSD drift temperature was set at 103 °C, and the nebulizer gas flow rate was 2.8 L/min. RLs were quantified using a commercial RL standard (AGAE-R90, catalog No. A791255126058; AGAE Technologies, Corvallis, OR, USA; purity, 90%), containing approximately 55.4% di-rhamnolipids and 44% mono-rhamnolipids according to the supplier. Mono-rhamnolipid and di-rhamnolipid fractions were assigned based on the reported retention times of the major rhamnolipid congeners, and their relative proportions were calculated from the integrated peak areas. The calibration curve was established over an RL concentration range of 0.2–1.4 mg/mL using logarithmically transformed concentration and peak-area data, yielding the regression equation y = 1.8978x + 2.222 with an R² of 0.9964, where x and y represent the logarithms of RL concentration and HPLC-ELSD peak area, respectively.

Glucose concentration was measured using an SBA-40C bioanalyzer equipped with a glucose oxidase electrode (Institute of Biology, Shandong Academy of Sciences, Jinan, China). The NH₄⁺ concentration in the fermentation broth was determined using the Nessler reagent spectrophotometric method. Biomass was determined as either dry cell weight (DCW) or OD₆₀₀, depending on the experiment. For DCW determination, 10 mL of fermentation broth was centrifuged at 12,000 rpm and 4 °C for 10 min, and the resulting cell pellet was dried at 70 °C overnight and weighed gravimetrically. For OD₆₀₀ determination, appropriately diluted fermentation broth samples were measured spectrophotometrically at 600 nm.

Volumetric RL productivity (mg/L/h) was calculated by dividing the final RL concentration by the corresponding fermentation time. Each fermentation condition was evaluated in an individual 5 L bioreactor run, and the reported error bars represent the standard deviations of repeated analytical measurements.

2.13. Transmission Electron Microscopy

Cells collected at the end of fermentation were fixed, dehydrated, embedded, and sectioned according to standard procedures for transmission electron microscopy. Ultrathin sections of approximately 70–90 nm were prepared using a Leica EM UC7 ultramicrotome (Leica Microsystems GmbH, Vienna, Austria), stained with uranyl acetate and lead citrate, and examined using a Hitachi HT7800 transmission electron microscope (Hitachi High-Tech Corporation, Tokyo, Japan) operated at 80 kV. Because the chemical composition of the observed intracellular structures was not independently determined, they are referred to as intracellular inclusions in this study.

2.14. Fermentation Kinetic Analysis

The major fermentation variables were defined as glucose concentration (S, g/L), dry cell weight (X, g/L), and RL concentration (P, g/L). Biomass growth was described using the logistic growth model, whereas RL formation was evaluated using the Luedeking–Piret model [19]:

dX/dt = μ_mX(1 − X/X_m)

(1)

dP/dt = α(dX/dt) + βX

(2)

where μ_m is the maximum specific growth rate, X_m is the maximum biomass concentration, α and β are the growth-associated and non-growth-associated product-formation coefficients, respectively. The corresponding model parameters and goodness-of-fit values were obtained by nonlinear regression and are summarized in Table S1.

The apparent biomass yield and RL yield based on glucose consumption were calculated as

Y_X/S = (X_f − X₀)/(S₀ − Sf)

(3)

Y_P/S = (P_f − P₀)/(S₀ − Sf)

(4)

where X₀ and X_f are the initial and final biomass concentrations, P₀ and P_f are the initial and final RL concentrations, and S₀ and S_f are the initial and final glucose concentrations, respectively.

To further characterize the temporal changes in biomass formation and RL production, the instantaneous biomass formation rate (r_X, g DCW/(L·h)) and instantaneous RL production rate (r_P, g RL/(L·h)) were calculated from the first derivatives of the fitted biomass and RL concentration profiles:

r_X = dX/dt

(5)

r_P = dP/dt

(6)

The maximum values of r_X and r_P, together with the corresponding fermentation times, were determined from the derivative curves. Kinetic model fitting and data processing were performed using Origin 2021 software (OriginLab Corporation, Northampton, MA, USA).

3. Results and Discussion

3.1. Effect of Nitrogen Source and pH

In our previous studies, strain YM4 produced more than 20 g/L RL under shake-flask conditions when soybean oil or glycerol was used as the carbon source and NaNO₃ was supplied as the nitrogen source [8]. To determine whether the preference for NaNO₃ observed in previous oil-or glycerol-based fermentations remained applicable to glucose fermentation, four nitrogen sources, including NaNO₃, soy peptone, (NH₄)₂HPO₄, and a mixed inorganic nitrogen source containing equimolar NH₄⁺ and NO₃⁻ were compared. Nitrate is generally considered a favorable nitrogen source for RL production when oils are used as carbon substrates, whereas ammonium nitrogen has often been associated with lower RL production [15,16]. When NaNO₃ was used as the sole nitrogen source, strain YM4 produced only 0.3 g/L RLs and reached a final biomass concentration of 0.9 g/L (Table 1). Cultures grown with peptone reached a higher biomass concentration (4.7 g/L), but the RL titer remained low at 1.5 g/L. Using (NH₄)₂HPO₄ as the nitrogen source resulted in an RL titer of 3.2 g/L and a biomass concentration of 3.9 g/L. In contrast, the mixed nitrogen source resulted in an RL titer of 14.1 g/L and a DCW of 3.8 g/L (Table 1). These results indicate that RL production was not solely associated with biomass formation.

Table 1. Effect of nitrogen sources on RL production and biomass formation by P. aeruginosa YM4 under the tested fermentation conditions.

Nitrogen Source	Shake Flask Scale				Bioreactor Scale
	RL Titer	Productivity	pH	DCW	RL Titer	Productivity	Time	DCW
	g/L	mg/L/h		g/L	g/L	mg/L/h	h	g/L
NaNO3	0.3 ± 0.1	3.1 ± 0.9	4.0 ± 0.2	0.9 ± 0.2	10.1 ± 1.0	206.1 ± 20.2	49	4.6 ± 0.3
Soy peptone	1.5 ± 0.5	15.6 ± 5.2	4.8 ± 0.2	4.7 ± 0.5	4.2 ± 0.7	68.9 ± 11.4	61	10.8 ± 0.8
(NH4)2HPO4	3.2 ± 0.3	33.3 ± 3.1	5.5 ± 0.1	3.9 ± 0.8	14.4 ± 0.5	177.8 ± 6.1	81	6.1 ± 0.4
Mixed nitrogen sources	14.1 ± 0.8	146.9 ± 8.3	6.5 ± 0.3	3.8 ± 0.4	18.2 ± 1.5	252.8 ± 20.8	72	5.2 ± 0.3

The fermentation medium contained 80 g/L glucose, and the total nitrogen concentration was maintained at 75 mM. Values represent mean ± SD from repeated analytical measurements of a single fermentation run.

The final culture pH varied among the nitrogen treatments. Despite phosphate buffering, NaNO₃, soy peptone, and (NH₄)₂HPO₄ cultures acidified to pH 4.0, 4.8, and 5.5, respectively, whereas the mixed nitrogen culture maintained a higher final pH of 6.5 and showed the highest RL production. These findings suggest that acidification may represent an important limitation during glucose-based fermentation.

To further evaluate the effect of acidification, nitrogen-source experiments were repeated in a 5 L bioreactor with pH controlled at 7.0. Under controlled-pH conditions, RL titers increased to 10.1, 14.4, and 18.2 g/L for NaNO₃, (NH₄)₂HPO₄, and the mixed nitrogen source, respectively. The increase in RL production from 0.3 to 10.1 g/L under controlled-pH conditions suggests that acidification substantially affected RL production in shake-flask fermentation, rather than nitrogen source composition alone determining RL production. The mixed nitrogen source still achieved the highest RL titer and a productivity of 252.8 mg/L/h, whereas soy peptone promoted biomass formation (10.8 g/L) but resulted in an RL titer of only 4.2 g/L. Therefore, the mixed nitrogen source was selected for subsequent pH evaluation.

3.2. Effect of Fermentation pH on Glucose Utilization and RL Production

The nitrogen-source experiments indicated that culture acidification was likely an important limitation in glucose-based fermentation. As shown in Figure 1A, both biomass accumulation and RL production were strongly affected by fermentation pH. At pH 6.5, only 6.3 g/L RL was produced. Increasing the pH to 7.0–7.4 increased RL production, with the highest titer (20.3 g/L) obtained at pH 7.2. Further increasing the pH to 7.6 reduced both biomass and RL production. Glucose consumption showed a similar trend (Figure 1B). At pH 7.2, glucose was almost completely consumed within approximately 62 h, whereas more than 20 g/L remained after 74 h at pH 6.5.

Figure 1. Effect of fermentation pH on (A) RL titer and DCW and (B) glucose consumption by P. aeruginosa YM4 cultivated in a 5 L bioreactor with 80 g/L glucose and a mixed NH₄⁺/NO₃⁻ nitrogen source. Data are presented as the mean ± SD of repeated analytical measurements.

Previous studies have identified fermentation pH as an important factor affecting cell growth and RL production. For example, Soares dos Santos et al. [17] reported that pH 7.0 supported the highest RL production by P. aeruginosa PA1 under batch fermentation, whereas deviations from this value reduced productivity. Earlier glucose-based studies also reported reduced growth and RL production when the fermentation pH deviated from the favorable range [18]. Consistent with these findings, YM4 showed the highest RL production at pH 7.2. In addition, the present results suggest that the favorable pH facilitated glucose utilization and that culture acidification was an important constraint in glucose-based RL fermentation.

3.3. Effect of C/N Ratio on Biomass Formation and RL Production

After selecting pH 7.2 as the fermentation condition, the C/N ratio was varied from 18 to 53 with 80 g/L glucose to determine a suitable carbon–nitrogen balance for RL production. Glucose catabolism provides metabolic intermediates and reducing power for RL precursor synthesis in P. aeruginosa [20], and appropriate medium composition can improve RL production by balancing carbon utilization, precursor supply, and cell growth [21].

As shown in Figure 2, RL production and biomass both increased as the C/N ratio rose from 18 to 35, reaching 20.8 g/L and an OD₆₀₀ of 15.1, respectively. Further increases to 41 and 53 reduced both RL production and biomass formation, indicating that both lower and higher C/N ratios were unfavorable for RL fermentation under the tested conditions. Thus, the higher RL production at C/N 35 was associated with balanced biomass formation and product accumulation.

Figure 2. Effect of C/N ratio on RL production and biomass formation by P. aeruginosa YM4 cultivated in a 5 L bioreactor with 80 g/L glucose at pH 7.2. The C/N ratio ranged from 18 to 53. Data are presented as the mean ± SD of repeated analytical measurements.

Notably, the favorable C/N ratio was identical to that previously determined for YM4 with soybean oil, despite the known effects of carbon source on RL titer and congener distribution [6]. This suggests that a similar C/N balance may be favorable for YM4 with both glucose and soybean oil under the respective tested conditions [8]. A C/N ratio of 35 was therefore used in subsequent experiments to evaluate the effect of initial glucose concentration.

3.4. Effect of Glucose Concentration on RL Production

Increasing the initial substrate concentration is a common strategy to increase product titer, provided that the additional carbon can be efficiently consumed. However, previous studies have shown that increasing glucose concentration does not always improve RL production. Reduced production at higher glucose concentrations has been reported for several Pseudomonas strains [22,23]. To identify the initial glucose concentration that resulted in the highest RL titer, fermentations were performed with glucose concentrations ranging from 60 to 150 g/L while maintaining pH at 7.2 and a C/N ratio of 35.

Increasing the glucose concentration from 60 to 100 g/L increased the RL titer from 15.7 to 24.1 g/L (Figure 3). At 100 g/L glucose, most of the glucose was consumed, with only 3.4 g/L remaining at the end of fermentation. Increasing the glucose concentration beyond 100 g/L did not further improve RL production.

Figure 3. Effect of the initial glucose concentration on RL production and biomass formation by P. aeruginosa YM4 cultivated in a 5 L bioreactor at pH 7.2 with a C/N ratio of 35. Initial glucose concentrations ranged from 60 to 150 g/L. Fermentation was terminated when the residual glucose concentration decreased below 5 g/L or after 96 h when glucose was not completely consumed. Data are presented as the mean ± SD of repeated analytical measurements.

At 120 and 150 g/L glucose, the final RL titers decreased, while substantial amounts of substrate remained unused. Residual glucose reached 64.9 g/L at an initial concentration of 150 g/L. These results suggest that YM4 was unable to efficiently utilize the additional glucose supplied at higher initial concentrations under the tested conditions. Increasing the initial glucose concentration beyond 100 g/L did not further increase RL production under the tested conditions.

Cell morphology also changed with increasing glucose concentration. When the glucose concentration increased from 60 to 80 g/L, OD₆₀₀ rose from 10.3 to 15.3. TEM observations further showed intracellular inclusions in cells cultivated under increased glucose supply (Figure S1). Although the chemical nature of these inclusions was not determined, their occurrence may be associated with increased intracellular carbon storage under carbon-rich conditions [24]. Together with the increased residual glucose concentration and lower RL production, these results indicate that further increasing glucose supply was not beneficial to RL production.

3.5. Effect of Fe²⁺ Concentration on RL Production

After determining the carbon and nitrogen conditions, the effect of Fe²⁺ concentration was investigated. Iron is required for multiple cellular processes and has been reported to affect RL production and the expression of RL biosynthetic genes in P. aeruginosa. Shatila et al. [14] showed that both iron form and concentration influenced RL production and the expression of rhlA, rhlB, and rhlR. Jiang et al. [25] also identified a low FeSO₄ concentration as favorable for high-productivity RL fermentation.

As shown in Figure 4, increasing the Fe²⁺ concentration from 3.6 to 10.8 μM increased the RL titer from 24.1 to 29.7 g/L. During the same range, OD₆₀₀ increased from 15.3 to 17.2. However, when Fe²⁺ was further increased to 14.4 μM, RL production decreased, whereas biomass accumulation was not markedly affected. The different responses of biomass and RL production at the highest Fe²⁺ concentration indicate that the effect of Fe²⁺ was not solely related to changes in cell growth. Instead, RL formation appeared to be sensitive to the tested range of iron availability. This observation is consistent with previous reports showing concentration-dependent effects of iron on RL production [14]. The highest RL titer was obtained at 10.8 μM Fe²⁺. Based on the preceding experiments, the conditions selected for further characterization were therefore 100 g/L glucose, a mixed NH₄⁺/NO₃⁻ nitrogen source at C/N 35, pH 7.2, and 10.8 μM Fe²⁺.

Figure 4. Effect of Fe²⁺ concentration on RL production and biomass formation by P. aeruginosa YM4 cultivated with 100 g/L glucose at pH 7.2 and a C/N ratio of 35. The Fe²⁺ concentration was varied from 3.6 to 14.4 μM by adjusting the concentration of FeSO₄·7H₂O in the trace-element solution. Data are presented as the mean ± SD of repeated analytical measurements.

3.6. Fermentation Profile and Kinetics Under the Selected Conditions

The fermentation process was subsequently characterized under the selected fermentation conditions. Glucose consumption, biomass formation, RL production, NH₄⁺ concentration, instantaneous biomass formation rate, and instantaneous RL production rate were monitored to characterize the temporal relationship between biomass formation and RL accumulation.

As shown in Figure 5A, biomass accumulated rapidly during the early stage of fermentation, while RL accumulation remained relatively limited. The instantaneous biomass formation rate (r_X) reached a maximum of 0.20 g DCW/(L·h) at approximately 7.4 h (Figure 5B). As fermentation proceeded, the NH₄⁺ concentration decreased and subsequently remained at approximately 3 mM, while biomass accumulation slowed and RL production continued. The instantaneous RL production rate (r_P) reached a maximum of 0.51 g RL/(L·h) at approximately 42.7 h, considerably later than the maximum biomass formation rate. These profiles indicate a temporal separation between rapid biomass formation and the period of highest RL production.

Figure 5. Fermentation profiles and related kinetic parameters under the selected fermentation conditions. (A) Time-course profiles of glucose consumption, NH₄⁺ consumption, RL production, and biomass formation together with the corresponding fitted concentration profiles. (B) Instantaneous biomass formation rate (rx, g DCW/(L·h)) and instantaneous RL production rate (r_P, g RL/(L·h)), obtained from the first derivatives of the fitted biomass and RL concentration profiles.

Independent kinetic model fitting showed that the Logistic model adequately described biomass growth (R² = 0.979), while the Luedeking–Piret model showed good agreement with the RL production data (R² = 0.994). The positive growth-associated and non-growth-associated coefficients suggested contributions from both growth-associated and non-growth-associated RL formation. Detailed kinetic parameters and apparent fermentation yields are summarized in Table S1. Changes in NH₄⁺ concentration may help explain the performance of the mixed nitrogen source. The rapid decline in NH₄⁺ during early fermentation suggests that ammonium mainly supported initial biomass formation, whereas nitrate may have contributed to nitrogen availability during later fermentation. However, nitrate consumption was not directly measured in this study.

At 91 h, the RL titer reached 29.7 g/L, corresponding to a volumetric productivity of 326.4 mg/L/h and a final dry cell weight of 6.0 g/L. Compared with the initial 5 L fermentation using the mixed nitrogen source at pH 7.0, sequential evaluation of pH, C/N ratio, glucose concentration, and Fe²⁺ increased the RL titer from 18.2 to 29.7 g/L and improved volumetric productivity from 252.8 to 326.4 mg/L/h. Overall, the fermentation kinetics indicate rapid biomass formation during the early stage, followed by continued RL accumulation after growth slowed. These observations also help explain why the highest biomass obtained during nitrogen-source screening did not correspond to the highest RL titer.

3.7. RL Production and Composition Among Different P. aeruginosa Strains

To determine whether the higher RL production of YM4 was mainly associated with strain characteristics or the selected fermentation conditions, three additional P. aeruginosa strains, PAO1, HN, and HEB, were cultivated under the same conditions.

As shown in Figure 6A, considerable differences were observed among the four strains. PAO1 and HN produced approximately 5–6 g/L RLs, whereas HEB produced 16.2 g/L RLs. In comparison, YM4 produced 29.7 g/L under the same fermentation conditions. These results indicate that the higher RL production of YM4 was not solely attributable to the selected medium composition and process conditions, but was also strongly influenced by strain background. The substantially higher production of YM4 under the same culture conditions supports its suitability for glucose-based RL fermentation.

Figure 6. RL production profiles of different P. aeruginosa strains under the selected fermentation conditions. (A) RL titers and OD₆₀₀ values of the four strains (HN, PAO1, HEB, YM4). (B) Relative proportions of mono- and di-rhamnolipids calculated from their integrated HPLC-ELSD peak areas relative to the total RL peak area.

The RL composition also differed among the strains (Figure 6B). Under glucose-based cultivation, di-rhamnolipids accounted for 73.8–95.2% of the total integrated RL peak area in the HPLC-ELSD chromatograms, with proportions of 95.2%, 85.6%, 78.9%, and 73.8% for HN, PAO1, HEB, and YM4, respectively. Thus, although YM4 achieved the highest total RL titer among the four strains, it did not exhibit the highest relative proportion of di-rhamnolipids. Zhao et al. [6], for example, reported differences in mono- and di-rhamnolipid composition when glucose and soybean oil were used by the same P. aeruginosa strain. These results indicate that total RL production and congener distribution are strain-dependent characteristics that should be considered separately. In the present study, the main advantage of YM4 was its substantially higher total RL production while still producing a di-rhamnolipid-rich product mixture.

Our previous studies also showed that the proportion of di-rhamnolipids produced by YM4 depended on the carbon substrate, reaching approximately 65% with soybean oil and 86% with glycerol [8]. PAO1 and HN similarly produced lower di-rhamnolipid fractions with soybean oil than those observed here with glucose. The present results suggest that glucose fermentation resulted in a di-rhamnolipid-rich profile among the strains examined, although the extent of this effect remained strain-dependent. To place the performance of YM4 in the context of previously reported glucose-based processes, published studies were compared in Table 2. Most reported glucose-based fermentations with P. aeruginosa produced less than 10 g/L RLs. YM4 produced 29.7 g/L within 91 h, corresponding to a volumetric productivity of 326.4 mg/L/h, indicating its potential for high-titer glucose-based RL production. Despite the favorable fermentation performance at the 5 L scale, further process development is required before larger-scale application. In particular, dissolved-oxygen dynamics and oxygen-transfer requirements should be characterized during scale-up, as dissolved oxygen was not actively controlled in the present study and may influence glucose metabolism, biomass formation, and RL production. Process economics, downstream recovery, and robustness at larger scales also remain to be evaluated. In addition, because P. aeruginosa is an opportunistic pathogen, appropriate containment, sterilization, and waste treatment strategies would be essential for any future large-scale implementation.

Table 2. RL production by different strains using glucose-based fermentation.

Strain	Glucose (g/L)	Nitrogen Source	Titer (g/L)	Productivity (mg/L/h)	Time (h)	Scale	Ref.
YM4	100	(NH4)2HPO4 NaNO3	29.7	326.4	91	5 L bioreactor	This study
PAO1			6.2	64.6	96
HN			5.1	53.1	96
HEB			16.2	172.3	94
S2	40	NH4NO3	5.3	36.8	144	5 L bioreactor	[18]
BC1	40	NaNO3	1.3	10.8	120	10 L bioreactor	[26]
SG	30	NaNO3	7.1	49.3	144	6 L bioreactor	[6]
MK002729.1	20	NH4Cl	2.6	27.1	96	shake flask	[27]
P. aeruginosa	20	(NH4)2SO4	4	33.3	120	shake flask	[28]
MR01	12	(NH4)2SO4	0.3	3.6	84	shake flask	[29]
NCIM 5514	10	NH4NO3	2.9	30.2	96	shake flask	[30]
	10	KNO3	2.7	28.1	96
PAO1	50	NH4Cl	0.4	4.2	96	shake flask	[23]
TMN	40	NaNO3	0.3	1.8	168	shake flask	[31]
EM1	40	NaNO3	8.6	51.4	168	shake flask	[32]

4. Conclusions

A glucose-based fermentation process for P. aeruginosa YM4 was established through sequential evaluation of nitrogen source, pH, C/N ratio, initial glucose concentration, and Fe²⁺ supplementation. The selected conditions resulted in an RL titer of 29.7 g/L with a volumetric productivity of 326.4 mg/L/h, and the produced RL mixture contained 73.8% di-rhamnolipids. Fermentation profiling and kinetic analysis showed rapid biomass formation during the early cultivation stage, followed by continued RL accumulation after growth slowed. The kinetic analysis further indicated contributions from both growth-associated and non-growth-associated RL formation. YM4 also achieved substantially higher RL titers than the three additional P. aeruginosa strains tested under the same conditions. These results support further investigation of glucose-based RL fermentation, although oxygen-transfer control, process economics, downstream recovery, scale-up performance, and biosafety require further evaluation.