Growth and protein content of Cupriavidus necator on organic acids derived from fermented grey starch

Abstract

Abstract

Grey starch (GS) is a byproduct of potato processing and is conventionally valorised as low-grade animal feed or for biogas generation. In this manuscript, we systematically investigate the mixed culture fermentation of GS to organic acids and their subsequent aerobic upgrading to microbial protein via Cupriavidus necator. We focused on protein content, growth kinetics, biomass yield on individual, and mixed organic acids (C1–C6), using both synthetic media and real fermented effluents. High-throughput cultivation in microtiter plates was employed to evaluate performance across both straight-chain forms and isoforms of the acids. C. necator demonstrated growth on all tested individual substrates, although with distinct individual behaviour. Lactate, butyrate, and hexanoate supported the highest biomass yields, reaching up to 0.24 gCDW/gCOD (grammes of cell dry weight per grammes of COD fed). Lactate enabled the highest specific growth rate (0.6 h⁻¹) with 29 ± 4% of protein. The maximum protein content (70 ± 11%) was observed on acetate at an initial concentration of 2 g/L. Depending on the acid, higher initial concentrations (2 and 4 g/L) led to increased cell dry weight but reduced growth rates or inhibition in some cases. Real fermented GS, primarily composed of lactate and butyrate, proved to be a viable substrate for microbial protein production. Undiluted fermented GS yielded the highest protein content (70 ± 9%), while a 1/4 dilution (2.6 gCOD/L) enabled the fastest growth (0.84 h⁻¹) compared to all tested fermented GS and acid concentrations. These findings highlight the potential of GS-derived organic acids as feedstock for microbial protein production for feed and food.

Key points

• C. necator grows on all tested individual organic acids (C1–C6).

• Protein production varies by acid type and initial concentration.

• Grey starch was converted to organic acids for microbial protein production.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00253-026-13780-7.

Introduction

The rising world population, projected to reach 9.7 billion by 2050 (FAO 2018), will lead to an increase of about 67% in combined demand for pork, beef, poultry, and fish between 2019 and 2050 (Falcon et al. 2022), thus also increasing the amount of livestock and fish feed needed (Cottrell et al. 2020; Rauw et al. 2023). The rising prices of some protein sources such as fishmeal (Nunes et al. 2022) and the greater demand for its sustainability requirements have stimulated the deployment of alternative protein sources (Munialo et al. 2022). Microbial protein (MP) can be such a source, typically coming from a protein-rich microbial biomass grown on different substrates such as methane, glucose, C1 compounds or even wastewater. MP can be bacteria, algae, yeast, or fungi, preferably with more than 30% protein in their biomass to compete with current sources such as soy or meat (Ritala et al. 2017). High content of essential amino acids and their balanced profile in view of dietary requirements makes MP suitable for livestock or fish feed (Anupama and Ravindra 2000). It is suggested that MP can replace up to 19% plant-based protein in animal feed depending on MP production pathway and substrate (Pikaar et al. 2018) and up to 75% of fish meal in some cases (Kuhn et al. 2009; Woolley et al. 2023). Therefore, MP may represent a promising new route to supply the demand for protein (Matassa et al. 2016; Murali Sankar et al. 2023).

To make MP more affordable and sustainable, low-value substrates need to be considered. Microorganisms can be grown on many different substrates such as waste streams (Barbera et al. 2019; Sakarika et al. 2022), and they can achieve exceptionally high production rates, reaching more than 3000 tons MP per hectare per year when normalized to land area (Pikaar et al. 2017). Among these waste streams, an underexplored but attractive substrate is grey starch (GS) from the potato industry. GS is recovered from wastewater rich in starch after heating it above 40 °C, resulting in its grey colour (Duynie 2023). Typically, it contains around 15–20% of solids which are mostly composed of carbohydrates, reaching 60–75% (Bradshaw et al. 2002), as well as proteins, fats and other compounds (Duynie 2023). GS is typically used as input for biogas production or low-grade animal feed which offers less benefits than MP (Grommers and van der Krogt 2009; Graham and Ledesma-Amaro 2023).

The key to ensure that MP can be used for feed/food is the use of food-approved microbial cultures. A good candidate is Cupriavidus necator (formerly Ralstonia eutropha), a facultative chemolithotroph, capable of autotrophic and heterotrophic growth (Vandamme and Coenye 2004) on various substrates including organic acids (Vu et al. 2022; Cai et al. 2023; Morlino et al. 2023). C. necator has the potential to not only produce protein, but also polyhydroxyalkanoates (PHAs) (Cai et al. 2023), which can be beneficial in aquaculture feed (Laranja and Bossier 2020). C. necator is a well-studied microorganism and its many strains have been tested on different waste substrates, such as vegetable oils (Obruca et al. 2010), animal fats (Riedel et al. 2014), or organic acids (Joris et al. 2024). Streams such as GS cannot be used directly by wild-type strains of C. necator (Santolin et al. 2024; Sichwart et al. 2011) due to its inability to utilize the starch. Even when sugars are then released into solution, for example by hydrolysis, C. necator cannot utilize a broad range of sugars, only fructose, N-acetylglucosamine, and gluconate (Santolin et al. 2024). An approach to deal with this is to introduce a fermentation step necessary to convert the different constituents of GS into organic acids. These compounds serve as key intermediates and could be metabolized by Cupriavidus sp. (Riedel and Brigham 2020). Depending on the conditions or culture used, a mixture of different organic acids is formed, ranging from formate to hexanoate and lactate (Li et al. 2025). The performance of Cupriavidus necator on organic acids has been studied primarily in the context of polyhydroxyalkanoate (PHA) production. Only a few studies have addressed protein synthesis, and data on growth using individual organic acids are often fragmented across multiple sources, making direct comparison of substrates difficult (Yang et al. 2010; Sakarika et al. 2020; Vu et al. 2022).

This study systematically evaluates microbial protein production, growth kinetics, and biomass yield on individual and mixed organic acids (C1–C6), using both synthetic media and real effluent derived from fermented grey starch. While C. necator is well studied for polyhydroxyalkanoate (PHA) production on such substrates, its potential for protein synthesis from low-value waste streams remains underexplored. By examining microbial performance on complex waste-derived media and various organic acids, this work addresses a key gap in the valorisation of fermentation by-products and offers novel insights into sustainable bioprocessing for alternative protein generation.

Methods

Microorganism and culture media

Cupriavidus necator LMG 1199 was acquired from Belgian coordinated collection of microorganisms (BCCM/LMG, Ghent, Belgium). The organism was pre-grown on modified ammonium mineral salts (AMS) medium (Table S1) with acetate (3 g/L) as carbon source at pH 7 and 28 °C. Subsequently, the same modified AMS medium with different organic acids (Thermofisher Scientific, USA) was used for the well plate experiments.

Mixed culture fermentation of grey starch

GS was obtained from a potato processing plant (Agristo, Nazareth, Belgium). The GS feedstock was characterized (Table S2) and stored at −20 °C until use. Due to its high solids content, the raw GS was diluted 6.67-fold with tap water to reach targeted VSS concentration of 20 g/L before being fermented. Fermentations were performed in triplicate using a 1-L double-jacketed glass bioreactor without external inoculum. The reactor was maintained at 55 ± 1 °C, and pH was controlled between 6.0 and 6.3 using a 1 M NaOH solution. Each fermentation run lasted 24 h to maximize lactate production, based on preliminary batch fermentation tests (see SI). To ensure anaerobic conditions, the reactor was sparged with nitrogen gas at the start of the experiment. The resulting GS fermentate was centrifuged and filter sterilised before using in well plate experiments.

Well plate experiments

Two series of experiments were conducted to the establish performance of C. necator on selected organic acids and fermented GS. The first series aimed to estimate the maximum specific growth rate (µ_max) using sterile, flat-bottomed 96 well plates with lids (655,180, CELLSTAR®, Greiner). Modified AMS medium with vitamins and selected organic acids (acetate, formate, butyrate, valerate, propionate, hexanoate, or lactate) at concentrations of 4, 2, 1, 0.5, 0.25, 0.125, 0.0625, and 0.03125 g/L of their sodium salts were used. COD equivalent of tested acids, estimated using conversion factor for each acid, is presented in Table 1. The initial pH of the buffered medium was 7 as it is the optimal value for C. necator’s growth. After inoculation at optical density (OD) 0.03, the final effective volume in the wells was 200 µL. The 96 well plates were placed in a plate reader (Infinite M200 Pro, Tecan, Switzerland) at 30 °C and orbital rotation of 180 rpm, where the OD at 600 nm was measured every 15 min. For protein content estimation, experiments were repeated with the same conditions as detailed above, except the organic acids were used at concentrations of 4, 2, and 1 g/L and with an increased number of technical replicates to ensure reliability when using the well plate method for protein determination.

Table 1.

Calculated COD values expressed as gCOD/L for all organic acids tested across the full range of experimental concentrations. COD values were derived using theoretical COD conversion factor

	Lactate	Formate	Acetate	Propionate	Butyrate	Valerate	Hexanoate
COD conversion factor(gCOD/gACID)	1.07	0.35	1.07	1.51	1.82	2.04	2.20
Acid concentration (g/L)	Concentration (gCOD/L)
4.0	4.26	1.39	4.26	6.05	7.26	8.15	8.82
2.0	2.13	0.70	2.13	3.02	3.63	4.07	4.41
1.0	1.07	0.35	1.07	1.51	1.82	2.04	2.20
0.50	0.53	0.17	0.53	0.76	0.91	1.02	1.10
0.25	0.27	0.09	0.27	0.38	0.45	0.51	0.55
0.13	0.13	0.04	0.13	0.19	0.23	0.25	0.28
0.063	0.07	0.02	0.07	0.09	0.11	0.13	0.14
0.031	0.03	0.01	0.03	0.05	0.06	0.06	0.07

For experiments with GS fermentate, the same conditions and procedures were applied, apart from the substrate used for growth. The objectives remained also the same. The GS fermentate was used as a substrate in dilutions: 0, 1/2, 1/4, 1/8, 1/16, 1/32, 1/64, and 1/128.

Protein and cell dry weight analysis

OD was measured at 600 nm using a plate reader (Infinite M200 Pro, Tecan, Switzerland) and converted to cell dry weight (CDW) (g_CDW/L) using calibration curves. Total (crude) protein was analysed in quintuplicate using Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific Inc., USA). In certain cases, low OD values (< 0.15) prevented correct calculation of CDW, and these samples had to be excluded. This was the case for formate (all concentration except 4 g/L), acetate (concentrations < 1 g/L), butyrate (concentrations < 0.5 g/L), hexanoate (concentrations 4 g/L and < 0.5 g/L), lactate (concentrations < 0.5 g/L), valerate (concentrations 4 g/L and < 0.5 g/L), and propionate (concentration < 0.5 g/L). In case of fermented GS, dilutions below 1/4 (2.64 gCOD/L) were excluded because their protein values were at the detection limit of the assay, making further calculations unreliable.

Two methods were employed to quantify the protein content of the lactate-grown biomass: the direct protein quantification with the BCA Protein Assay Kit and indirectly via measuring the total nitrogen (TN) (NANOCOLOR Nitrogen TNb 220, Macherey-Nagel, Germany) and converting it to protein by multiplying with 5.35 (Jia et al. 2024). This approach was followed since the quantification of protein in one sample (sodium lactate, concentration 4 g/L) yielded consistently protein content above 100% (Table S3). To ensure that our approach would yield robust conclusions, we performed additional experiments with the same microbial strain and medium, and lactate concentration. The results of the latter experiment yielded comparable protein concentration and content regardless of the approach (Table S4), and therefore, due to the robustness and less interferences of the TN quantification, we decided to report the values of microbial protein as TN*conversion factor in this specific well plate experiment.

Calculation of kinetic parameters

The growth rate (µ) was calculated as outlined in Candry et al. (2018). To summarize, the OD of the inoculated wells was adjusted using the average OD of the wells that were not inoculated, and then it was pre-processed (noise, spikes and curve collapse removal) using PRECOG software (Fernandez-Ricaud et al. 2016). After log-transforming the OD data, the Gompertz equation was applied using the nls.lm optimization algorithm from the minpack.lm package in R, resulting in the estimation of μ. Concentrations of organic acids showing negligible growth as well as dilutions in case of GS fermentate were excluded from the calculation because the values were too low to correctly estimate the growth rate.

Analytical methods

Samples of grey starch fermentate (10 ml) were subjected to centrifugation at a speed of 10,000 rpm for a duration of 10 min. They were then filtered using a 0.22-µm filter. Prior to analysis, the samples were preserved in a freezer.

The pH of fermented GS was measured using Consort C3020 (Consort, Belgium) with Consort epoxy pH electrode SP10T with Pt1000 temperature sensor.

Ion chromatography (930 Compact IC Flex; Metrohm, CH) was used to quantify organic acids, and the procedure was performed as described in Sakarika et al. (2020). C2–C8 fatty acids (including isoforms C4–C6) were measured by gas chromatography (GC-2014, Shimadzu®, The Netherlands) according to Andersen et al. (2014).

In well plate experiments with fermented GS, samples from replicate cultures were pooled to obtain sufficient volume for chemical oxygen demand (COD) analysis. As a result, yield (gCDW/gCOD) derived from these measurements does not include standard deviations, as individual replicate variability could not be assessed. The quantification of chemical oxygen demand (COD) was carried out using COD testing kits (NANOCOLOR COD 1500, Macherey–Nagel) and UV–VIS spectroscope (NANOCOLOR Advance, Machery-Nagel).

Results

Kinetics of C. necator on organic acids as substrates

Growth on organic acids displayed several consistent aspects. C. necator reached the stationary phase at different times depending on the substrate (Fig. 1), ranging from the fastest on lactate (10.5 h) to the slowest on formate (21 h) and valerate (20.7 h). Increasing the concentration of organic acids generally led to higher biomass yields, with the exception of hexanoate and valerate, which showed signs of growth inhibition at elevated levels. Analysis of growth curves enabled the determination of maximum specific growth rates (Fig. 2, Table S5).

Fig. 1.

Growth curves of C. necator on individual organic acids used for kinetic analysis, including an example of Gompertz model fitting

Fig. 2.

Growth rates (h⁻¹) of C. necator grown on individual VFAs at varying concentrations

Formate-grown C. necator achieved a growth rate of 0.25 ± 0.02 h⁻¹ at a concentration of 1 g/L, with a similar growth rate of 0.22 ± 0.01 h⁻¹ observed at 2 g/L. The highest (4 g/L) and lowest concentrations (0.5 g/L) exhibited growth rates of 0.16 ± 0.01 h⁻¹ and 0.19 ± 0.01 h⁻¹, respectively. Growth rates on acetate were comparable across concentrations 2, 1, and 0.5 g/L reaching 0.33 ± 0.02, 0.34 ± 0.03, and 0.34 ± 0.01 h⁻¹, respectively. The lowest growth rate (0.25 ± 0.01 h⁻¹) was observed at 4 g/L. This may indicate substrate limitation at the lowest concentration and inhibition effect at the highest. Propionate demonstrated higher growth rates at lower concentrations, mirroring the trend observed with other acids in this work. Starting from a rate of 0.32 ± 0.01 h⁻¹ at concentration 0.5 g/L, the values decreased to 0.25 ± 0.01 h⁻¹ at 1 g/L and 0.19 ± 0.01 h⁻¹ at 2 g/L. The lowest observed growth rate of 0.13 ± 0.01 h⁻¹ was observed at a concentration of 4 g/L.

Growth on butyrate exhibited the highest rates of 0.41 ± 0.02 h⁻¹ and 0.40 ± 0.02 h⁻¹ at initial concentrations of 0.5 g/L and 1 g/L, respectively. The growth rate decreased to 0.3 ± 0.04 h⁻¹ at a concentration of 2 g/L. Growth curve at 4 g/L of butyrate displayed a diauxic-like growth pattern and therefore kinetics at this initial concentration were not calculated.

On varying concentrations of valerate, the growth rate was 0.27 ± 0.01 h⁻¹ at a concentration of 0.5 g/L. A lower growth rate was noted at 1 g/L (0.20 ± 0.01 h⁻¹) and at a concentration of 2 g/L, the growth rate further declined to 0.13 ± 0.1 h⁻¹. No growth was observed at 4 g/L. Progressively lower growth rates were also observed with increasing hexanoate concentrations. A growth rate of 0.39 ± 0.02 h⁻¹ was achieved at 0.5 g/L, while the value of 0.20 h⁻¹ was measured at 2 g/L. Negligible growth was observed at a concentration of 4 g/L. Lactate supported growth across all tested concentrations. The value of 0.60 ± 0.03 h⁻¹ was achieved at both concentrations of 0.5 g/L and 1 g/L. A growth rate of 0.57 ± 0.04 h⁻¹ was measured at a concentration of 2 g/L and 0.50 ± 0.02 h⁻¹ was observed at 4 g/L, which was the highest growth rate recorded at this concentration among all the tested acids.

Choice and concentration of organic acids influence the biomass concentration and yield

Biomass concentration varied between the organic acids with, as expected, higher initial acid concentrations supporting higher biomass concentrations in most of the cases (Fig. 3a). Similarly, biomass yield, calculated based on gCOD fed (gCOD_fed), was influenced by both the type and concentration of the organic acid (Fig. 3b).

Fig. 3.

a Cell dry weight of C. necator grown on individual organic acids, compared with values reported in previous studies (Yang et al. 2010; Jawed et al. 2022; Cai et al. 2023). No point in concentration bracket indicates OD too low for CDW calculation. b Biomass yields of C. necator grown on individual organic acids

Formate supported minimal growth, even at the highest tested concentration. Formate at 4 g/L supported CDW of 0.13 ± 0.01 g/L and biomass yield of 0.09 ± 0.01 gCDW/gCOD_fed. Although measurable values were observed only at 4 g/L, the available results indicate that formate yields lower performance relative to the other acids tested.

During growth on acetate, CDW was lower at lower initial acid concentrations, reaching 0.32 ± 0.07 g/L, 0.28 ± 0.04 g/L, and 0.15 ± 0.01 g/L at 4 g/L, 2 g/L, and 1 g/L acetate, respectively. The corresponding biomass yields were 0.08 ± 0.02 gCDW/gCOD_fed, 0.13 ± 0.01 gCDW/gCOD_fed, and 0.14 ± 0.01 gCDW/gCOD_fed. Hence, yield increased as concentration decreased, indicating more efficient substrate utilization at lower concentrations and inhibition at higher concentrations.

Growth of C. necator on propionate displayed a trade-off between biomass accumulation and substrate conversion efficiency. When propionate was used as a substrate, the CDW concentrations were 0.39 ± 0.06 g/L, 0.49 ± 0.05 g/L, 0.28 ± 0.04 g/L, and 0.14 ± 0.01 g/L at 4 g/L, 2 g/L, 1 g/L, and 0.5 g/L, respectively, with yields from 0.07 ± 0.01 to 0.22 ± 0.03 gCDW/gCOD_fed. The highest CDW was measured at 2 g/L, but lower values were observed at substrate concentrations below or above. For yield, it was the concentration of 1 g/L where the value was the highest (0.22 gCDW/gCOD_fed). This divergence between optimal concentrations for CDW and yield may suggest a trade-off between biomass production and substrate conversion efficiency as well as the effect of inhibition at higher concentrations and mirrors the trend observed for other acids such as biomass production on butyrate.

Inhibition patterns were observed when butyrate served as the substrate. It showed higher CDW at 4 g/L and 2 g/L (0.65 ± 0.01 g/L and 0.85 ± 0.01 g/L) compared to 1 g/L and 0.5 g/L (0.35 ± 0.01 g/L and 0.17 ± 0.01 g/L), with yields ranging from 0.09 ± 0.01 to 0.23 ± 0.01 gCDW/gCOD_fed. Biomass yield and CDW peak at 2 g/L, suggesting optimal conversion efficiency at this concentration. Subsequent decrease at the highest concentration of 4 g/L is consistent with the concentration-dependent inhibition trends observed for other organic acids in this study except lactate.

Shifting focus to the longer‑chain acids, hexanoate and valerate exhibited similar effects on growth and yields of C. necator, with clear evidence of inhibition at higher concentrations. In case of hexanoate, the highest CDW was at 2 g/L (0.95 ± 0.05 g/L) while yields were between 0.16 ± 0.01 and 0.22 ± 0.01 gCDW/gCOD_fed and peaking at 2 g/L. Valerate resulted in CDW concentrations of 0.64 ± 0.03 g/L, 0.29 ± 0.01 g/L, and 0.15 ± 0.01 g/L at 2 g/L, 1 g/L, and 0.5 g/L. The yields, which were relatively stable and very similar, ranged between 0.14 ± 0.01 and 0.16 ± 0.01 gCDW/gCOD_fed, with a slight peak at 2 g/L. The absence of growth at 4 g/L suggests that hexanoate and valerate are inhibitory at higher concentrations. The highest CDW at 2 g/L in both cases indicates that the concentration range near 2 g/L may represent an optimum for cell growth.

Interestingly, when lactate was used as a substrate there was a deviation from this inhibition pattern, as the highest concentration supported the greatest biomass accumulation. CDW of C. necator cultivated on lactate ranged from 0.75 ± 0.03 g/L at 4 g/L to 0.13 ± 0.01 g/L at 0.5 g/L, with the highest value observed at 4 g/L. Biomass yields were very similar at 0.5, 1 and 2 g/L lactate concentrations, reaching 0.24 ± 0.01,0.23 ± 0.01, and 0.22 ± 0.01 gCDW/gCOD_fed, respectively. The highest concentration (4g/L) displayed a lower yield of 0.18 ± 0.01 gCDW/gCOD_fed. This trend is consistent with other acids tested in our work, where higher concentrations result in increased CDW although for other acids this trend did not hold at the highest concentration (4 g/L). In this case, lactate was the only acid for which the highest concentration (4 g/L) also supported the highest CDW.

Different organic acids led to different protein content

The protein content of C. necator cultivated on various volatile fatty acids (VFAs) exhibited distinct trends across different substrate concentrations and types of acids (Fig. 4). The results of formate-grown C. necator provided limited data, showing measurable protein content (25 ± 6%) only at 4 g/L due to limited growth on lower concentrations. For acetate, the protein content increased from 30 ± 3% at 1 g/L, to 70 ± 11% at 2 g/L. The highest substrate concentration (4 g/L) showed a decrease again in protein content to 37 ± 14%. Propionate supported protein contents of 59 ± 2% and 50 ± 9% at 2 g/L and 1 g/L, respectively, with no data at 4 g/L because of limited growth. Cells cultivated on butyrate, valerate and hexanoate showed a decrease in protein content as the substrate concentration increased. The protein levels observed at initial concentration 1 g/L for all three substrates, were 36 ± 5% for butyrate, 49 ± 2% for valerate, and 41 ± 13% for hexanoate. Upon increasing the substrate concentration to 2 g/L, we measured protein content for butyrate- and for hexanoate-grown cultures, to be 32 ± 5% and 38 ± 15%, respectively. In contrast, protein content in cultures grown on valerate decreased, with protein content dropping by almost half to 26 ± 4%. Lactate stimulated protein synthesis to 29 ± 4% at 1 g/L, and to 39 ± 10% and 40% at 2 and 4 g/L of substrate, respectively.

Fig. 4.

Protein content of C. necator (%_CDW) grown on individual organic acids with comparison to other scientific works (Sakarika et al. 2020; Peng et al. 2022)

Fermented grey starch as new substrate for MP production

Grey starch was subjected to mixed culture fermentation at thermophilic conditions (55 ±  1 °C, pH = 6.0–6.3) and stopped after 24 h to maximize lactate production. The GS fermentate contained 10.55 gCOD/L in total with lactate (4.11 gCOD/L) being the main organic acid, followed by butyrate (3.58 gCOD/L) (Table S2). Other constituents included formate (0.17 gCOD/L), acetate (0.24 gCOD/L), and propionate (0.08 gCOD/L). It was used as substrate in dilutions: 0 (10.55 gCOD/L), 1/2 (5.28 gCOD/L), 1/4 (2.64 gCOD/L), 1/8 (1.32 gCOD/L), 1/16 (0.66 gCOD/L), 1/32 (0.33 gCOD/L), 1/64 (0.16 gCOD/L), and 1/128 (0.08 gCOD/L).

Total COD yields on substrate were 0.89 ± 0.11 gCOD gVSS⁻¹ consumed, from which lactate represented 0.49 ± 0.21 gCOD gVSS⁻¹. The obtained fermentate was then evaluated across three different substrate concentrations, corresponding to varying dilution levels (Fig. 5a). The different concentrations of fermentate affected growth rate of C. necator (Fig. 5c). The highest observed growth rate was 0.84 ± 0.23 h⁻¹ at a dilution to 2.64 gCOD/L. Growth rates at dilutions of 0.66 gCOD/L, 5.28 gCOD/L, and 1.32 gCOD/L were 0.69 ± 0.16 h⁻¹, 0.62 ± 0.06 h⁻¹, and 0.60 ± 0.03 h⁻¹, respectively. Undiluted GS fermentate (10.55 gCOD/L) still supported growth (0.36 ± 0,01 h⁻¹) although it was approximately half the rate observed with diluted fermentate.

Fig. 5.

a Growth curves of C. necator on fermented GS. b Example of Gompertz model fitting (dashed line) used for kinetic analysis. c Growth rates (h⁻¹) of C. necator grown on fermented GS at varying concentrations

The biomass production (Fig. 6a) and yield (Fig. 6b) from fermented GS were also influenced by different substrate concentrations. At the highest substrate concentration of 10.55 gCOD/L (undiluted), the resulting CDW was 0.17 ± 0.01 g/L, and biomass yield of 0.19 gCDW/gCOD. Upon dilution of fermented GS to a concentration of 5.28 gCOD/L, the CDW was 0.44 ± 0.02 g/L, with an improved yield of 0.31 gCDW/gCOD. Further dilution to substrate concentration of 2.64 gCOD/L, resulted in a CDW of 0.22 ± 0.01 g/L and a slightly lower yield of 0.25 gCDW/gCOD.

Fig. 6.

a Concentration of protein and CDW for C. necator grown at varying gCOD/L levels of fermented GS. b Biomass yields of C. necator grown on fermented GS

We observed changes in protein concentration following fermentate dilution. At the undiluted condition (10.55 gCOD/L), the protein content reached 70 ± 9% relative to CDW. Upon dilution to 5.28 gCOD/L, protein concentration remained similar, resulting in a protein content of 32 ± 6% due to high amount of CDW. This difference may be attributed to lower CDW at higher concentration of fermentate suggesting a potential inhibitory effect at higher concentrations. Additionally, high concentration of carbon-rich compounds in fermentate combined with low concentration of nitrogen led to a high C/N ratio of 24 and 12 in undiluted and 5.28 gCOD/L fermentate, respectively. This ratio favours the accumulation of PHA over protein synthesis, as nitrogen limitation is known to promote this process (Sánchez Valencia et al. 2021). Further dilution to 2.64 gCOD/L, lowered both carbon and nitrogen concentrations which supported protein content of 26 ± 3%.

Discussion

In the present study, growth on organic acids exhibited several consistent patterns. Increasing the concentration of organic acids generally resulted in a decline in growth, although the extent of inhibition varied among acids. This is likely the result of the inhibitory effect of organic acids at higher concentrations, as observed before (Warnecke and Gill 2005). Notably, while higher initial concentrations sometimes supported greater biomass accumulation, protein content did not always follow the same trend, displaying acid‑ and concentration‑specific differences. Although similar observations such as the effect of acid concentration on growth rate, biomass yield or protein content appear across the literature, direct comparison remains challenging because data on individual acids are scattered and prior investigations employed diverse cultivation methods. These discrepancies underline the need for systematic evaluations of organic acids, as addressed in the present work.

To place these general patterns in context, it is helpful to compare them with previous findings on individual acids and cultivation systems. Kim et al. (1992) noted in their study that C. necator grown in flasks was inhibited by propionic acid concentrations exceeding 1.5 g/L, while achieving maximum growth rate slightly above 0.20 h⁻¹, at concentration of 0.5 g/L. In contrast, in our work C. necator grew on propionate at concentrations above 1.5 g/L and displayed overall higher growth rate. Similar results were achieved in fed-batch system with propionic acid kept at 3 g/L, reaching a growth rate of 0.40 h⁻¹ as reported by Kobayashi et al. (2000). Compared to our results, the 2 times higher growth rate observed by Kobayashi et al. can be attributed to the better pH control and aeration exhibited by the fed-batch reactor in comparison to the well plate system in our work. Turning to formate and acetate, our findings align with previous work of Sakarika et al. (2020), where acetate-grown (1.72 g/L) C. necator showed μ_max = 0.41 h⁻¹ at pH 7 while growth on formate (1.80 g/L) yielded a lower μ_max of 0.29 h⁻¹ at the same initial pH. The different growth rates observed in the present study may reflect differences in cultivation conditions such as medium composition. Differences between acids were notable in lactate grown cultures. As with other acids, the negative trend between concentration and growth rate was observed in samples grown on lactate. However, the results indicate that inhibitory effect was lowest with lactate at the pH used here, which is well above its pKa. Supporting this observation, Sonnleitner et al. (1979) reported that a high concentration of lactic acid (20 g/L) also promoted strong growth of C. necator, with a maximum specific growth rate (μ_max) of 0.42 h⁻¹. The growth of C. necator on 4 g/L butyrate exhibited distinct time-based kinetics which is why we discuss this fact in greater detail. As mentioned before, growth kinetics at this initial concentration were not calculated because the curve displayed a diauxic-like growth pattern (Fig. 1). Following an initial increase in optical density between 10th and 22nd hour, the values reached a plateau. This period lasted approximately 6 h, after which the growth of C. necator resumed, as indicated by increasing OD₆₀₀. It can be speculated that the diauxic-like growth observed in C. necator may result from metabolic overflow triggered by elevated butyrate levels, potentially exceeding the assimilation capacity of the TCA cycle and PHB biosynthesis pathways. This could lead to the accumulation and export of intermediates such as pyruvate, thereby disrupting metabolic regulation and potentially triggering a secondary growth phase (Przybylski et al. 2013).

Beyond growth rate, biomass accumulation also varied across acids. In general when considering all acids tested, we observed lower CDW than those reported by Yang et al. (2010); Jawed et al. (2022) and Cai et al. (2023) at comparable substrate concentrations (Fig. 3). These differences may have been caused by different cultivation methods: the referenced studies used flasks, active aeration and greater culture volume compared to well plates used in our study. Although well plates are suitable for trend analysis, they are prone to oxygen limitations influenced by plate type and shaking mode (Zhang et al. 2008; Mansoury et al. 2021), which may result in lower overall biomass (Kaushik et al. 2020).

For most acids, we observed that increased concentrations of substrates, mostly up to 2 g/L, resulted in elevated CDW, whereas higher doses suppressed growth, consistent with toxicity and chain‑length effects reported previously. At elevated concentrations, specific substrates with long chain such as valerate and hexanoate strongly inhibited growth. This is in good agreement with findings of Vu et al. (2022) who reported increased toxicity with acid concentration and Jawed et al. (2022) who observed a negative relationship between chain length and growth performance of C. necator. Growth on formate was particularly limited, aligning with previous report (Sakarika et al. 2020; Jawed et al. 2022), which can be attributed to low energy density of substrate (1 g of formate has only 0.35 g COD compared to 1.07 g of COD in acetate) as well as low efficiency (only 20–35%) of Calvin cycle when using formate (Cotton et al. 2020). In the Calvin cycle cells must first oxidize formate to CO₂ and then re‑fix that CO₂ with a pathway that demands 9 ATP and 6 NAD(P)H per triose, leading to energy loss (Claassens et al. 2020; Hudson 2024). These metabolic constraints likely contribute to the reduced growth performance on formate observed in this study.

Extending from biomass to protein, substrate type and carbon balance influenced protein yields. The results observed in our work highlight the substrate-specific nature of protein accumulation and suggest that optimal substrate concentrations vary depending on acid type, possibly due to differences such as inhibition thresholds or carbon assimilation efficiency. This observation emphasizes the need for further research, since few studies have systematically evaluated protein production by C. necator in response to varying concentrations of individual organic acids. Sakarika et al. (2020) observed protein content of 46% on acetate and 52% on formate, both cultivated in flasks, which was higher than values found by in this study (Fig. 4) and can be attributed to different methods of cultivation. Sakarika et al. (2022), elsewhere reported protein contents ranging from 17 to 58% CDW using lactic acid from grass juice fermentation. Similar trends have been observed in other microbial systems. Peng et al. (2022) reported with the phototroph Rhodopseudomonas sphaeroides different effects of VFA on protein production. The initial concentrations used in the study were 0.93 g/L acetate, 0.74 g/L propionate, and 0.66 g/L butyrate. The protein content was found to be highest using acetate (60%), followed by propionate (55%) and butyrate (50%) which was similar to our work. The observed reduction in protein content at higher acid concentrations may be associated with increased PHA accumulation, which is typically enhanced under stress conditions, including nutrient limitation and reduced oxygen availability. In our study, elevated C/N ratios at higher acid concentrations (above 10 for all acids, except for formate at 6) likely favoured PHA over protein synthesis (Sánchez Valencia et al. 2021). Potential oxygen limitations and the inhibitory effects of specific acids may have further stressed the cells, promoting PHA accumulation, even at lower acid concentrations with more favourable C/N ratios (below 10) (Ahn et al. 2015; Zhang et al. 2022).

Research on the use of Cupriavidus necator for microbial protein production from complex, waste-derived media remains limited, representing a knowledge gap addressed in this study. Consistent with the broader findings in this work, the overall trends observed with fermented GS mirror acid-specific effects, particularly concentration-dependent inhibition of growth.

Compared to previous studies, the growth rates observed in our work are notably higher. A synthetic mixture of organic acids (acetic, butyric, and propionic acid) promoted a growth rate of 0.24 h⁻¹ (Wang et al. 2010). In another study, food scraps were fermented to a mixture containing acetic, propionic, butyric, and lactic acids, with concentrations of 0.42, 0.16, 1.40, and 1.51 g/L, respectively. The maximum growth rate achieved with this substrate was 0.018 h^–1 (Du et al. 2004). In case of biomass production, our results indicate that moderate dilution of the fermented GS enhances biomass yield, suggesting a potential inhibitory effect at higher substrate concentrations or improved metabolic efficiency at lower organic loads. In this context, larger-scale operation modes offer options to mitigate this negative effect such as different feeding regimes including pulse feeding and feeding controlled by pH or dissolved oxygen (Yamanè and Shimizu 1984).

Beyond growth performance, protein content in C. necator observed in this study aligns with previous reports of high protein accumulation on diverse substrates such as glucose or volatile fatty acids (VFAs). Joris et al. (2024) observed protein contents of 73.3 ± 0.5% CDW and 71.4 ± 2.4% CDW when C. necator was grown on glucose (9.6 g/L) and a mix of VFAs (acetate, propionate, valerate, butyrate, iso-valerate, and iso-butyrate; with acid concentrations 0.4 to 3 g/L), respectively, under lower dilution rates in a chemostat. Similarly, a glucose-utilising mutant, Cupriavidus necator CECT4623 (NCIMB 11599), grown on glucose, also reached 70% CDW of total protein (Ismail et al. 2024) and (Li et al. 2024) measured protein contents ranging from 44 to 71% in C. necator grown on synthetic biogas. These comparisons underscore the importance of optimizing carbon source, strain selection, and environmental parameters to maximize protein yield in microbial systems.

Conclusion

This study systematically assessed the growth and protein production of Cupriavidus necator on various organic acids and fermented grey starch. Growth was observed across all acids, with an inverse relationship between growth rate and acid concentration. Increasing acid levels led to increased CDW up to acid concentrations of 4 g/L. At this concentration lower growth was observed, indicating possible inhibition. The highest protein content of 70 ± 11%_CDW was reached at 2 g/L of acetate; only lactate and acetate supported significant production at 4 g/L (40% and 37%, respectively). Fermented GS, dominated by lactate and butyrate, proved to be a feasible substrate. Undiluted fermentate yielded the highest protein content (70 ± 9%), while dilution to 2.64 gCOD/L supported the fastest growth rate (0.84 ± 0.23 h⁻¹) compared to other concentrations of fermentate. These findings highlight the potential of grey starch as a feedstock to produce microbial proteins. Further studies are needed to evaluate additional parameters (e.g. pH, temperature) and operational modes at larger scale such as pulse or pH-controlled feeding to establish the attainable productivities and yields.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

M. Struk and K. Rabaey conceived and designed research. M. Struk conducted experiments, analysed data and wrote the manuscript. M. Sakarika and Á. Estévez provided analytical tools and guidance. K. Rabaey, R. Ganigué, M. Sakarika, Á. Estévez reviewed and edited the draft. All authors read and approved the manuscript.

Funding

This project has received funding from the European Union’s Horizon Europe research and innovation programme under the grant agreement No. 101081776, the UK Research and Innovation (UKRI) fund under the UK government’s Horizon Europe funding guarantee, the Swiss State Secretariat for Education, Research and Innovation (SERI), and from the National Key Research and Development Program supported by the Ministry of Science and Technology of the People’s Republic of China (No. 2023YFE0104900). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of neither of the aforementioned funding authorities. None of the cited funding authorities can be held responsible for them.

M. Sakarika was supported by Ghent University (BOF24/PDO/023). Á. Estévez is supported by a Marie Skłodowska-Curie Postdoctoral Fellowship (Grant Agreement ID: 101153341). R. Ganigué gratefully acknowledges support from the Special Research Fund of Ghent University (BOF.BAF.2024.0502.01).

Data availability

Data will be made available on request.

Declarations

Ethical approval

Not applicable. This article does not contain any studies with human participants or animals performed by any of the authors.

Competing interests

The authors declare no competing interests.