Synthetic Biofilm Reactor with Independent Supply of Gas and Liquid Phase for Studying Chain Elongation with Immobilized Clostridium kluyveri at Defined Reaction Conditions

Abstract

In this study, we explore the use of C. kluyveri in synthetic biofilms for the production of 1-butyrate and 1-hexanoate, investigating the impact of inoculation temperature during biofilm formation and the presence of yeast extract. Therefore, a novel synthetic biofilm reactor has been designed and constructed. Prior to investigating synthetic biofilms in this reactor, we carried out preliminary batch experiments in anaerobic flasks containing an inoculated agar hydrogel fixed at the bottom and overlaid medium. For the operation of the novel synthetic biofilm reactor, specific volumes of inoculated agar hydrogel were dispensed into a cylindrical mold with a diameter of 102 mm, forming the synthetic biofilm with a height of 4 mm, which was then transferred into the biofilm reaction chamber onto the support grid. The biofilm support grid separates the gas phase (CO₂, N₂) above the synthetic biofilm from the aqueous phase (medium) below. Our results show that C. kluyveri remains metabolically active at biofilm preparation temperatures of up to 45 °C, with extended lag phases observed at 70 °C. The synthetic biofilm demonstrated efficient chain elongation in batch processes, converting ethanol and acetate into 1-butyrate and 1-hexanoate, with final concentrations of 2.7 g L⁻¹ and 10.1 g L⁻¹, respectively, with yeast extract in the circulating liquid medium of the synthetic biofilm reactor setup. The maximum estimated space-time yields for 1-butyrate and 1-hexanoate, referenced to the biofilm volume, were 1.331 g L⁻¹ h⁻¹ and 4.947 g L⁻¹ h⁻¹, respectively. Experiments without yeast extract lead to final concentrations of 2.0 g L⁻¹ 1-butyrate, and 7.3 g L⁻¹ 1-hexanoate and maximum estimated space-time yields, referenced to the biofilm volume, were 0.332 g L⁻¹ h⁻¹ and 1.123 g L⁻¹ h⁻¹, respectively. The use of synthetic biofilms, even without yeast extract, eliminates the need for significant cell growth during chain elongation. However, product concentrations were lower without yeast extract.

1. Introduction

The chain elongation of short-chain carbon substrates, such as acetate and ethanol, into longer-chain hydrocarbons like 1-butyrate and 1-hexanoate holds significant relevance due to its potential to address critical challenges in sustainable chemistry, bioeconomy, and environmental management [1,2]. This process, mediated by anaerobic microbial fermentation, has garnered increasing interest as a promising route for the production of medium-chain fatty acids (MCFAs), which serve as valuable platform chemicals and precursors for biofuels, bioplastics, and various industrial applications [3,4,5]. From a sustainability perspective, chain elongation offers an attractive solution for valorizing low-value or waste carbon sources, including agricultural residues, food waste, and industrial effluents, into high-value products [6,7]. By converting these substrates into MCFAs, this biotechnological approach contributes to waste reduction and the circular bioeconomy, aligning with global efforts to reduce dependence on fossil resources [8,9]. Furthermore, the chain-elongating process is inherently energy-efficient and environmentally friendly, as it relies on microbial consortia operating under mild conditions, such as low operating pressure and temperatures near to room temperature. The conversion of ethanol, a readily available bio-based product, and acetate, often a by-product of acetogenic processes, into longer-chain hydrocarbons exemplifies the integration of different biochemical processes to achieve sustainable resource utilization [6,10]. Additionally, the ability to produce tailored chain lengths through this process is of great importance. Medium-chain fatty acids such as 1-butyrate and 1-hexanoate possess unique physicochemical properties, making them suitable for diverse applications, including antimicrobial agents, lubricants, and additives for food and feed industries [11,12].

C. kluyveri is widely recognized as the model organism for the chain elongation of short-chain carbon substrates into MCFAs like 1-butyrate and 1-hexanoate [1,13,14]. MCFAs are produced via the reverse β-oxidation pathway, enabling the sequential elongation of two-carbon units to form longer-chain fatty acids, utilizing acetate as the carbon donor and ethanol as the electron donor. More than 25% of the cellular carbon of C. kluyveri is derived from CO₂ during chain elongation [15]. The robustness of C. kluyveri further enhances its suitability as a model organism. C. kluyveri tolerates fluctuations in pH, substrate concentrations, and other environmental conditions, making it adaptable to a wide range of feedstocks, including those derived from industrial or agricultural waste streams [14,16,17]. This adaptability is crucial for scaling up chain elongation processes using inexpensive and heterogeneous substrates.

The use of yeast extract as a supplement in the cultivation of C. kluyveri is common practice due to its ability to support growth and chain elongation [18,19,20,21]. Yeast extract provides a rich and readily available source of essential nutrients, including vitamins, amino acids, and trace elements [22]. It acts as a complex growth factor, supplying precursors for the biosynthesis of coenzymes and other cellular components necessary for key metabolic pathways, including reverse β-oxidation [23,24]. For example, it provides B vitamins, such as riboflavin and biotin, which are vital cofactors in many enzymatic reactions within C. kluyveri [22]. While yeast extract is not strictly necessary for C. kluyveri cultivation in all cases, its inclusion simplifies medium preparation and ensures consistent performance, particularly in non-defined or complex feedstocks [16,25]. This is especially relevant in research and industrial applications, where maximizing productivity and minimizing variability are critical. But yeast extract is also a cost factor that has to be considered when thinking about economical and industrial relevant production processes [26,27]. That is what makes it important to consider implementing chain elongation without yeast extract in the medium.

So far, only a few results have been published on chain elongation with C. kluyveri without yeast extract in batch processes. Bornstein and Barker (1948) found that yeast extract is important for the optimal growth of C. kluyveri, but the same biomass formation as with yeast extract supplementation can be achieved through the addition of specific vitamins, namely biotin and para-aminobenzoic acid (PABA). They suggest that these two vitamins are key components of yeast extract responsible for supporting the growth of C. kluyveri [28]. Gildemyn et al. (2017) found that C. kluyveri was able to maintain similar production rates without yeast extract, provided that defined growth factors were added. They supplemented the medium with precise amounts of trace elements, vitamins, and a selenite–tungstate solution, and observed comparable growth and 1-hexanoate production to that achieved with yeast extract [18]. Yan and Dong (2018) found that no 1-hexanoate production occurred when yeast extract and the vitamins were omitted from the medium. This highlights the essential role of these components in supporting the metabolic activity of C. kluyveri [25]. San-Valero et al. (2020) observed that omitting yeast extract from the medium did not extend the lag phase of growth of C. kluyveri, but it did reduce 1-hexanoate production [29]. It has to be noted that, in their experiments, not only acetate and ethanol were provided as substrates, but 1-butyrate was also included as an additional substrate in the medium.

It is therefore beneficial to overcome the limited cell density of C. kluyveri cultivated in suspension, e.g., in stirred tank bioreactors, caused by the omission of yeast extract. One promising approach is the initial immobilization of cells on carriers inside the bioreactor using the ability of many microorganisms to form biofilms [30,31].

As an example, Shen et al. (2014) [32] performed a continuous syngas fermentation process in a monolithic packed-bed reactor with C. carboxidivorans, forming the biofilm naturally. After two days of biofilm formation, the fermentation process was started. They found that the syngas fermentation performance was not only dependent on the gas–liquid mass transfer efficiency, but also on the biofilm formation. Their biofilm reactor showed higher syngas utilization efficiency and productivity compared to a bubble column reactor with suspended cells. In addition, Shen et al. (2014) [33] studied the continuous syngas fermentation with C. carboxidivorans in a hollow fiber membrane biofilm reactor with the biofilm growing on the shell side of the hollow fiber membrane for 48 h before starting the process. The gaseous phase was supplied through the lumen side of the membranes while the liquid media passed the shell side. They achieved a maximum ethanol concentration of nearly 24 g L⁻¹ with an ethanol-to-acetate ratio of 4.79. Furthermore, Shen et al. (2017) [34] carried out continuous experiments in a horizontally oriented packed bed biofilm reactor (rotating biological contactor) for syngas utilization with C. carboxidivorans. A gear motor was used to rotate a semi-submerged cage containing the carrier material, ensuring uniform liquid contact with all surfaces. The gas phase was supplied to the headspace, and biofilm growth was completed within 48 h. They achieved an ethanol concentration of 7.0 g L⁻¹ and a productivity of 6.7 g L⁻¹ d⁻¹, representing a 3.3-fold increase compared to a continuous stirred tank reactor under identical operation conditions.

Riegler et al. (2019) [35] operated a multi-purpose bioreactor as a packed-bed and a trickle-bed biofilm reactor for the conversion of CO₂/H₂ to acetate with immobilized C. aceticum. Biofilm formation and gas conversion occurred simultaneously in the continuous processes. They reported the immobilization of up to 8.2 g cells per L packed-bed volume of a macroporous ceramic carrier material, using the natural ability of C. aceticum to form biofilms. The continuously operated trickle-bed biofilm reactor showed an up to 3-fold higher acetate space-time yield compared to the packed-bed biofilm reactor at hydrogen conversions of up to 96%.

Reports on biofilm formation of C. kluyveri are rare. Zhang et al. (2019) [36] used dried wheat straw as a carrier material in anaerobic shake flasks. To inoculate the wheat straw with C. kluyveri, the researchers soaked wheat straw with medium containing varying bacterial concentrations. Biofilm formation occurred simultaneously with product generation in the batch process, and it was not possible to pinpoint a specific time at which biofilm development was complete. They found that C. kluyveri could withstand 2.4 times higher ammonia concentrations when immobilized compared to suspended cells, showing that immobilized C. kluyveri exhibit increased robustness. In a further study, Zhang et al. (2021) [37] investigated the 1-hexanoate production with C. kluyveri and wheat straw in anaerobic flasks [36]. No statement could be made about the duration of completed biofilm formation. They tested different acetate/ethanol ratios as substrates and achieved 1-hexanoate concentrations of up to 17.0 g L⁻¹.

Compared to biofilms formed naturally, synthetic biofilms are not reliant on biofilm formation and cell growth inside the bioreactor. For example, Kheyrandish et al. (2015) [38] produced acetone, butanol, and ethanol from potato waste starch with immobilized and suspended C. acetobutylicum with the same cell dry weight concentrations at the start of the process. They used calcium alginate-polyvinyl alcohol boric acid beads with immobilized cells in a batch operated 5 L stirred tank bioreactor. Although cell immobilization resulted in a lower butanol concentration compared to batch fermentation with suspended cells, it enabled the separation and reuse of cells for subsequent cycles. During repeated batch fermentation, butanol concentrations totaling 1.5 times the amount produced with suspended cells were achieved across three sequential cycles.

Morandeira et al. (2019) [39] entrapped Halomonas sp. in hydrogel spheres based on agar, alginate, and alginate–polyvinyl in order to degrade choline in a fixed-bed bioreactor. They demonstrated that the complete biodegradation of choline chloride from aqueous solutions was achievable using this setup.

The formation of natural biofilms in biofilm reactors is a time-consuming process that poses challenges for industrial and research applications requiring rapid implementation [32,33,34]. It therefore makes sense to focus research more intensively on synthetic biofilms [40]. Currently, there is limited evidence in the literature demonstrating the successful establishment of synthetic biofilms with specific organisms. Research on chain elongation with immobilized C. kluyveri, either naturally or synthetically, is rare. Only Zhang et al. [36,37] provided an example of natural biofilm formation with C. kluyveri in anaerobic shaking flasks, but no additional studies with this organism have been reported to date.

To address the prolonged timeframes associated with natural biofilm development, the use of defined synthetic biofilms has emerged as a promising alternative [38,39]. Synthetic biofilms offer the potential to bypass the lengthy maturation periods in natural biofilm formation. Despite their advantages, synthetic biofilms with C. kluyveri have not yet been fabricated or applied in bioreactors. Therefore, we have recently demonstrated that C. kluyveri can be immobilized in agar hydrogels [41]. Agar, composed of agarose and agaropectin, was selected to enable the rapid entrapment of cells under anaerobic conditions, while being biologically inert. Unlike, for example, alginate, which requires a cross-linking agent and is therefore subject to diffusion limitations, agar solidifies upon cooling without the need for cross-linking, ensuring efficient immobilization. The viability and metabolic activity of C. kluyveri in this synthetic biofilm were already shown in anaerobic shake flasks [41].

To achieve well-defined conditions in contrast to anaerobic shake flasks for investigating anaerobic C. kluyveri cells immobilized in planar synthetic biofilms, we decided to design a new laboratory-scale biofilm reactor setup. In this setup, the synthetic biofilm with defined biofilm height and surface area separates the gas phase for CO₂ supply and the aqueous phase (medium), both streaming across the top (gas phase) and bottom surfaces (liquid phase) of the synthetic biofilm. The recirculating medium serves a dual purpose: providing substrates (ethanol and acetate) and removing products from the biofilm. First, it is essential to identify the optimal temperature conditions for entrapping C. kluyveri within the agar hydrogel and to assess whether there are deviations in productivity compared to suspended cells. A suitable synthetic biofilm preparation procedure provides the basis for subsequent investigations on chain elongation with immobilized C. kluyveri for the production of 1-hexanoate in the synthetic biofilm reactor under defined reaction conditions, both with and without yeast extract in the medium.

2. Materials and Methods

2.1. Microorganisms and Growth Conditions

The bacterial strain that has been worked with in this study was Clostridium kluyveri DSM 555, obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). A modified version of Hurst medium, originally described by Hurst and Lewis (2010) [42], prepared by supplementing the original formulation with of 0.4 g L⁻¹ L-cysteine-HCl as a reducing agent and 2.5 g L⁻¹ sodium bicarbonate as a buffer, following the protocol described by Schneider et al. (2021) [43], was used for all experiments. The medium was further supplemented with 10 g L⁻¹ potassium acetate and 20 mL L⁻¹ ethanol as heterotrophic substrates for C. kluyveri, and the pH was adjusted to pH 6.8. The medium was prepared anaerobically, as described in the literature [44,45]. This medium was utilized for preculture preparation and batch experiments conducted in anaerobic flasks and the synthetic biofilm reactor.

2.2. Preculture

The strain acquired from the DSMZ (Braunschweig, Germany) was initially grown in a modified Hurst medium. Glycerol was subsequently added, and the cultures were preserved at −80 °C as frozen stocks until needed for preculture preparation. To prepare the precultures, 2.5 mL of the frozen stock culture was thawed and transferred into anaerobic flasks equipped with a nitrogen-filled headspace at a total pressure of 1.0 bar. The transfer was carried out using a syringe (BD Discardit II, Becton Dickinson, Franklin Lakes, NJ, USA) and sterile needles (Sterican 0.9 × 70 mm, B. Braun, Melsungen, Germany) through a butyl rubber stopper (Glasgerätebau Ochs, Bovenden, Germany). C. kluyveri was grown heterotrophically in 500 mL anaerobic flasks, each containing 100 mL of modified Hurst medium. The medium was supplemented with 10 g L⁻¹ potassium acetate and 20 mL L⁻¹ ethanol as carbon sources. Additionally, 2.5 g L⁻¹ Na₂HCO₃ was included to provide bicarbonate/carbon dioxide as an inorganic carbon source [15,46], and 0.4 g L⁻¹ L-cysteine-HCl was added as a reducing agent. The cultures were maintained at 37 °C in a shaking incubator (Wisecube WIS-20R, Witeg Labortechnik GmbH, Wertheim, Germany) set to 100 rpm with a 2.5 cm orbital eccentricity for 5 days. Cells were harvested during the exponential growth phase by centrifugation at 3620 rcf for 10 min (Hettich Zentrifuge, Rotina 50 RS, Andreas Hettich GmbH, Tuttlingen, Germany). The collected cells were subsequently resuspended in anaerobic phosphate-buffered saline (PBS, pH 7.4) for use in inoculating liquid cultures or synthetic biofilm preparations. The inoculum cell quantity per anaerobic flask was kept identical for all flasks within an experimental series by evenly distributing the resuspended pellet across all parallelized flasks.

2.3. Preparation of Synthetic Biofilms

An amount of 100 mL of the hydrogel for the synthetic biofilm was prepared by dissolving 18 g L⁻¹ agar in an anaerobic bottle in a modified Hurst medium containing all supplements except L-cysteine-HCl, as previously described, followed by autoclaving at 121 °C for 20 min. To form the synthetic biofilm, the anaerobic bottle containing the hydrogel was opened and cooled to room temperature under a nitrogen atmosphere in an anaerobic work bench (MB-Labstar ECO, M. Braun Inertgas-Systeme, Garching, Germany). When the temperature dropped to 45 °C, 0.4 g L⁻¹ L-cysteine-HCl was added by pipetting and the hydrogel was inoculated with the cell suspension prepared in PBS by pouring it into the liquid hydrogel. Mixing was performed by shaking the bottle containing the inoculated hydrogel. Specific volumes of the inoculated hydrogel were then dispensed into self-constructed sterile stainless steel molds with 102 mm diameter (Supporting Information Figure S4) or 500 mL anaerobic flasks using a positive displacement pipette (Transferpettor Digital 2000–10,000 µL; BRAND GmbH + CO KG, Wertheim, Germany), depending on the experimental setup, resulting in mean heights of 4 mm for the synthetic biofilm. The molds or flasks were then cooled to room temperature under a nitrogen atmosphere in the anaerobic work bench for solidifying the synthetic biofilms.

2.4. Experiments in Anaerobic Flasks

Experiments in 500 mL anaerobic flasks were conducted either with a synthetic biofilm at the bottom or in suspension culture. For experiments with immobilized C. kluyveri cells, anaerobic flasks containing synthetic biofilm at the bottom were overlaid with 100 mL of modified Hurst medium. The anaerobic flasks were then sealed airtight under a nitrogen atmosphere and pressurized with a gas mixture of 1:9 N₂:CO₂ at a total pressure of 1.5 bar. For experiments with suspension cultures, airtight anaerobic flasks containing 100 mL of modified Hurst medium with all described components, except for L-cysteine-HCl, were pressurized with the same gas mixture (1:9 N₂:CO₂) at a total pressure of 1.5 bar. The flasks, along with their contents, were brought to the desired inoculation temperature using water baths. Via syringes and needles, they were then supplemented with 0.4 g L⁻¹ L-cysteine-HCl and inoculated with the prepared cell suspension in PBS. Immediately after that, they were cooled down to 37 °C at room temperature for 10 to 30 min and maintained at 37 °C in a shaking incubator (Wisecube WIS-20R, Witeg Labortechnik GmbH, Wertheim, Germany) set to 100 rpm with a 2.5 cm orbital eccentricity throughout the experiments.

2.5. Experiments with the Synthetic Biofilm Reactor

The previously steam-sterilized (121 °C, 20 min) biofilm reactor chamber (Figure 1) was transferred into the anaerobic work bench and opened under a nitrogen atmosphere. The solidified synthetic biofilm was then removed from the mold using a custom-built tool and transferred into the opened biofilm reactor chamber. Subsequently, the chamber was resealed to ensure gas-tightness and removed from the anaerobic work bench.

Subsequently, 600 mL of modified Hurst medium, containing all additional components except L-cysteine-HCl, was pumped into the biofilm reactor via a septum. Then, 0.4 g L⁻¹ of L-cysteine-HCl was added via the same septum. Throughout the batch processes, the biofilm reactor chamber’s headspace was continuously sparged with 5 L h⁻¹ of a gas mixture of 4:1 N₂:CO₂ at a total pressure of 1.0 bar. The pH of the liquid was maintained at pH 6.8 using 2 M NaOH or 3 M H₂SO₄. The temperature was controlled to 37 °C. Experiments using the modified Hurst medium with 1.0 g L⁻¹ yeast extract (the standard formulation) were compared to those without yeast extract (0.0 g L⁻¹).

2.6. Biofilm Reactor Setup

The biofilm reactor consists of three main modules: the biofilm module, the level control module, and the process control unit. These modules are interconnected via hoses, gas lines, and apparatuses according to the schematic shown in Figure 1(left) and Figure 2.

The biofilm reactor chamber (A02) is depicted in Figure 1(right) as an exploded view for a detailed explanation of its construction. Visible components include the bottom section with a tangential liquid inlet (1), the flat gasket (2) for a liquid-tight connection between the biofilm support grid (3) and the bottom section, the optional microfiltration membrane (4) for optional cell retention, the lower casting frame (5) for securing the support grid and for casting the biofilm, the upper casting frame (6), the synthetic biofilm (7), the lower part of the lid (8), the flat gaskets (9 and 11) for gas-tight installation of the sight glass (10), and the sight glass flange (12) for securing the sight glass. Technical drawings can be found in the Supporting Information (Figures S1–S5).

The premixed gas is supplied with a volumetric flow rate of 5 L h⁻¹ from gas cylinders containing nitrogen (N₂), and carbon dioxide (CO₂), with the dosing of each gas controlled by calibrated independent mass flow controllers (F-201CV, Bronkhorst High-Tech B.V., Veenendaal, Netherlands) to the biofilm reactor chamber’s headspace with an approximate volume of 200 mL. Before the gas mixture enters the biofilm reactor, it passes through a sterile filter (Inline-filter PTFE, 0.2 µm, Carl Roth, Karlsruhe, Germany) (F01), and a laboratory bottle filled with deionized water (Schott Duran 250 mL, VWR International, Radnor, PA, USA) is used as a gas humidification unit (A01). The gas then enters the biofilm reactor chamber (A02), which is housed in a self-designed water bath from a stainless steel container (GN container, GastroHero GmbH, Hamm, Germany) (B01), with controlled temperature via a heat exchanger and temperature regulation system (VMS-C10 Advanced, VWR International, Radnor, PA, USA) (W01). After passing through the biofilm reactor chamber, the gas exits on the other side. Subsequently, the gas flows through a laboratory bottle (Schott Duran 250 mL, VWR International, Radnor, PA, USA) as a self-designed overflow return container (A03), a self-made condensate trap (laboratory bottle) (A10), a self-constructed Liebig exhaust cooler (W03), a laboratory bottle containing silica gel as exhaust drying unit (Silica gel, Carl Roth, Karlsruhe, Germany) (A11), another sterile filter (F02), a mass flow measurement unit (F-111CM, Bronkhorst High-Tech B.V., Veenendaal, The Netherlands) (FI), and a micro-gas chromatograph (490 Micro GC) equipped with a 1 m Cox HI column at 80 °C and a thermal conductivity detector (both Agilent Technologies, Waldbronn, Germany) (OGA) for exhaust gas analysis, before being discharged into a special exhaust system for disposal. The gas atmosphere of the biofilm reactor chamber is connected to those of the self-designed level control module made from a laboratory bottle (Schott Duran 500 mL, VWR International, Radnor, PA, USA) (A04) and the process control unit (technical drawings in the Supporting Information Figures S6–S8) (A05) consisting of a closed container with a pH electrode (405-DPAS-SC-K8S/120, Mettler Toledo, Columbus, OH, USA) connection for automated acid and base addition and sampling septum (diameter 12 mm, Infors AG, Bottmingen, Switzerland) to ensure uniform pressure throughout the system.

The liquid collected in the overflow return container is directly pumped into the process control unit using a peristaltic pump (Ismatec Reglo Digital, Cole-Parmer GmbH, Vernon Hills, IL, USA; abbr. P02). The circulating liquid (0.5 L h⁻¹; liquid volume in the biofilm reactor chamber approx. 150 mL) exits the biofilm reactor chamber at the bottom and flows directly into the level control module, which can be adjusted vertically using a lift platform (H01) to regulate the liquid level in the biofilm module. Excess liquid is siphoned off via an immersion tube and pumped into the process control unit using another peristaltic pump (Ismatec Reglo Digital, Cole-Parmer GmbH, Vernon Hills, IL, USA; abbr. P01). The process control unit is placed in a water bath (B02), with the temperature control achieved using a heat exchanger (W02), both manufactured by Aerne Analytic (Weißenhorn, Germany). The pH of the circulating liquid is also regulated in the process control unit through the operation of two peristaltic pumps (100 series with 114DV flip-top single-channel pumphead, Watson-Marlow Fluid Technology Solutions, Wilmington, DE, USA) (P04 and P05) connected to a pH probe (405-DPAS-SC-K8S/120, Mettler Toledo, Columbus, OH, USA) (QIC).

Acid or base is added as needed from dedicated supply containers (A07 and A08), which are connected to a pressure-balancing module (Plastigas, Linde plc, Dublin, Ireland) (A09) to ensure oxygen-free pressure equilibration within these vessels. From the process control module, the circulating liquid phase is pumped back into the biofilm module using a peristaltic pump (Masterflex L/S Digital Miniflex, Cole-Parmer GmbH, Vernon Hills, IL, USA) (P03). Before re-entering the biofilm module, the liquid passes through a three-way valve (self-assembled with materials from Landefeld, Kassel, Germany) (V05) that allows for the reactor system to be emptied at the end of the process. The self-constructed drainage apparatus (A06) enables both the liquid and any residual gas in the containers and hoses to be discharged from the process loop.

2.7. Analytical Methods

For cell dry weight (CDW) determination, the optical density (OD₆₀₀) of the samples obtained via the septum in the process control unit with single-use syringes (BD Discardit II; Becton Dickinson, Franklin Lakes, NJ, USA) and sterile needles (Sterican 0.9 × 70 mm, B. Braun, Melsungen, Germany) was determined at a wavelength of 600 nm using a UV-Vis spectrophotometer (Genesys 10S UV-Vis, Thermo Scientific, Neuss, Germany). This measurement was used to estimate the CDW concentrations based on a linear correlation factor determined before (0.47 ± 0.04 g L⁻¹) [21]. Organic acids and alcohols were analyzed using HPLC (1100 Series, Agilent Technologies, Santa Clara, CA, USA) equipped with a refractive index (RI) detector and an Aminex HPX-87H ion exchange column (Bio-Rad, Munich, Germany). The separation was performed with 5 mM H₂SO₄ as the mobile phase, operating at a constant flow rate of 0.6 mL min⁻¹, and with the column temperature maintained at 60 °C. Prior to injection into the HPLC, the samples were passed through a 0.2 µm cellulose filter (Chromafil RC20/15 MS, Macherey-Nagel GmbH & Co. KG, Düren, Germany). Time-dependent product formation rates were estimated through sigmoidal model fitting using non-linear regression with the Gompertz curve [47].

3. Results and Discussion

3.1. Variation in the Inoculation Temperature

The effect of inoculation temperature on C. kluyveri was investigated in anaerobic flasks containing suspended cells. The tested inoculation temperatures were 25 °C, 45 °C, and 70 °C. Immediately after inoculation, the anaerobic flasks were transferred to an incubator operated at 37 °C. The incubation temperature of 37 °C was achieved between 10 and 30 min. Ethanol and acetate were provided as heterotrophic substrates in the anaerobic fermentation medium. The gas phase in the flasks consisted of a N₂:CO₂ mixture at a 1:9 ratio, with a total pressure of 1.5 bar. Over a batch process duration of 7 days, the CDW concentration and pH were monitored, and the organic acids and alcohols produced were analyzed (Figure 3).

The suspended C. kluyveri cells become metabolically active after variable lag phases from 12 to 72 h, while the lag phase increases with increasing inoculation temperature, carrying out chain elongation by converting ethanol and acetate into 1-butyrate and 1-hexanoate. During substrate conversion, the pH decreases from pH 6.4 to pH 5.7, where it stabilizes across all experiments. The final concentrations of CDW, acetate, ethanol, 1-butyrate, and 1-hexanoate are consistent across experiments, with CDW at ~1.9 g L⁻¹, acetate at ~1.8 g L⁻¹, ethanol at ~8.4 g L⁻¹, 1-butyrate at ~3.5 g L⁻¹, and 1-hexanoate at ~4.8 g L⁻¹. The lag phase for experiments conducted at inoculation temperatures of 25 °C and 45 °C is approximately 12 h, while higher inoculation temperatures lead to longer lag phases. For an inoculation temperature of 70 °C, the lag phase extends to approx. 70 h.

The incomplete consumption of ethanol and acetate is likely due to the drop in pH, as C. kluyveri remains metabolically active only up to approximately pH 5.7. At higher inoculation temperatures, the initial active dry mass concentration of the cells is likely reduced, which may account for the observed delays at temperatures above 45 °C. Inoculation temperatures up to 45 °C do not appear to impact the process or affect cell viability.

It was clearly demonstrated that C. kluyveri remains metabolically active at inoculation temperatures up to 70 °C, exhibiting comparable activity for temperatures up to 45 °C. This suggests that cell immobilization agents designed to create synthetic biofilms, which solidify upon cooling, are suitable for use with C. kluyveri. To avoid compromising cell activity due to heat stress and reduced cell viability, it is recommended to use agents that solidify at temperatures below 45 °C, ensuring that C. kluyveri is not exposed to inoculation temperatures exceeding this threshold.

3.2. Batch Processes with Synthetic Biofilm Compared to Suspended C. kluyveri Cells

Based on the results from Section 3.1, the batch conversion of ethanol and acetate was investigated in 500 mL anaerobic flasks using C. kluyveri immobilized in 1.8% (w/v) agar hydrogel, inoculated at 45 °C with a CDW concentration of 0.07 g L⁻¹ related to the 100 mL of liquid volume, and compared to 100 mL suspended cells inoculated with the same CDW concentration in anaerobic flasks. The heterotrophic substrates, ethanol and acetate, were added to the anaerobic fermentation medium. An amount of 100 mL medium was added without inoculation to the anaerobic flasks with the synthetic biofilm. The gas phase in all anaerobic flasks consisted of a N₂:CO₂ mixture at a 1:9 ratio, with a total pressure of 1.5 bar. The results are shown in Figure 4.

During ethanol and acetate conversion, the pH decreases from pH 6.4 to approximately pH 5.8, where it stabilizes for both the immobilized and the suspended cells. An increase in CDW concentration was observed only in the experiments with suspended cells, as the immobilized cells remained within the immobilization matrix and did not detach. Ethanol and acetate conversion is slightly faster between 24 and 48 h with the suspended cells. The final substrate conversions are slightly increased in the batch processes with suspended cells, resulting in lower substrate concentrations (acetate: 3.1 ± 0.4 g L⁻¹ compared to 2.3 ± 0.2 g L⁻¹; ethanol: 6.0 ± 0.1 g L⁻¹ compared to 5.3 ± 0.3 g L⁻¹). The final concentrations of 1-butyrate (2.3 ± 0.1g L⁻¹) are identical within the estimation error, but more 1-hexanoate was produced with the suspended cells (7.0 ± 0.2 g L⁻¹ compared to 6.4 ± 0.1g L⁻¹). As observed before, the incomplete consumption of ethanol and acetate is likely attributed to the decrease in pH.

The slightly lower final product concentrations and delayed substrate-to-product conversion with the synthetic biofilm may be caused by a reduced initial viable cell concentration of C. kluyveri due to a potential viability loss during the immobilization procedure. This could prolong biomass accumulation in the synthetic biofilm and, thus, the metabolic activity. Unlike freely suspended cells, immobilized cells may experience constrained growth due to spatial limitations, nutrient diffusion barriers, and altered physiological states, which may result in slower conversion kinetics. A reduced final cell density in the biofilm, structural limitations, and local early pH drops may further contribute to lower conversion efficiency. Diffusion limitations within the biofilm could exacerbate localized pH variations, negatively impacting metabolic activity. However, these limitations are likely minimal, as the biofilm is very thin (4 mm) and process times are very long.

We were able to demonstrate that C. kluyveri immobilized in a planar synthetic biofilm made of agar hydrogel with a thickness of 4 mm shows metabolic activities nearly identical to that of suspended cells.

3.3. Product Formation Without Yeast Extract

The influence of yeast extract on C. kluyveri within a synthetic biofilm was investigated using a biofilm reactor. Experiments were conducted with and without yeast extract in the cultivation medium and for synthetic biofilm preparation. The synthetic biofilm was inoculated with 0.82 g L⁻¹ C. kluyveri at the start. Ethanol and acetate served as heterotrophic substrates in the anaerobic fermentation medium, while the reactor was continuously gassed at 5 L h⁻¹ with a 1:4 CO₂:N₂ mixture at a total pressure of 1.0 bar. Temperature and pH were maintained at 37 °C and pH 6.8, respectively. The pH and temperature were monitored online, and CDW concentration, the conversion of the substrates, and the production of organic acids were analyzed offline after sampling from the liquid phase (Figure 5).

C. kluyveri utilizes acetate and ethanol via chain elongation to produce 1-butyrate and 1-hexanoate in both experimental approaches. However, substrate consumption and product formation are more pronounced in the presence of yeast extract.

In both setups, ethanol is fully consumed by the end of the batch fermentation. Acetate, however, remains at 1.0 g L⁻¹ with yeast extract and at 1.5 g L⁻¹ without. Final product concentrations are lower in the absence of yeast extract: 1-butyrate reaches 2.0 g L⁻¹ (vs. 2.7 g L⁻¹ with yeast extract) and 1-hexanoate reaches 7.3 g L⁻¹ (vs. 10.1 g L⁻¹). Additionally, final product concentrations are achieved approximately 80 h earlier in the experiments with yeast extract, being at around 160 h.

Yeast extract plays a crucial role in the production of 1-hexanoate by C. kluyveri. In its absence, C. kluyveri is forced to utilize C2 compounds for biomass formation in addition to CO₂, which consequently reduces the carbon available for product synthesis and leads to lower product yields [22,28,48].

The CDW concentration in suspension also increases earlier and has a maximum at ~90 h compared to ~110 h without yeast extract with values of ~0.10 g L⁻¹ vs. ~0.04 g L⁻¹. Biomass formation and product synthesis do not compete in the presence of yeast extract because product synthesis provides adenosine triphosphate and reduction equivalents needed for growth; therefore, the biomass formation is higher. Without yeast extract, cells must produce previously supplied molecules on their own, which slows growth if essential components are missing. Metabolic activity continues at a reduced level, since adenosine triphosphate and reduction equivalents are still required for maintenance metabolism. Due to the observed cells in suspension, it can be assumed that part of the chain elongation was carried out by those cells. Since the maximum space-time yield for 1-butyrate occurs both for the experiment with (~54 h) and without yeast extract (~80 h) prior to the strongest increase in CDW concentration, it can be concluded that 1-butyrate was primarily produced by the cells in the biofilm. The further chain elongation to 1-hexanoate reaches its maximum space-time yield at approximately 80 h and 110 h. During this time, the number of cells in suspension is also maximal, suggesting that these cells likely contributed to the production of 1-hexanoate. The liquid volume is 19.1 times greater than the volume of the synthetic biofilm. Therefore, the maximum measured CDW concentrations in suspension correspond to theoretical CDW concentrations in the synthetic biofilm volume of 1.91 g L⁻¹ and 0.76 g L⁻¹, respectively.

To make a quantitative estimate of the contribution of cells in suspension to product formation, both an optimistic and a pessimistic approach are presented below.

In the optimistic scenario, cell growth in the synthetic biofilm is assumed to be equivalent to how it would be in suspension. Furthermore, it is assumed that cell growth in the biofilm occurs simultaneously with that in the liquid medium throughout the process. These assumptions are supported by the results shown in Figure 4, where very similar activities were observed for experiments with immobilized cells and cells in suspension, while immobilized cells remained in the biofilm. Based on the growth recorded in these experiments, where the CDW concentration increased from 0.07 g L⁻¹ to 0.4 g L⁻¹—corresponding to a multiplication factor of 5.7—the maximum CDW concentration for the yeast extract approach can be estimated using this factor. Given the initial biomass concentration in the gel of 0.82 g L⁻¹, the maximum CDW concentration would be approximately 4.67 g L⁻¹ and would be reached after ~90 h. This would suggest that the contribution of cells in suspension after 90 h would be at ~29% for the yeast extract approach. Accordingly, the contribution to product formation would also be at similar levels. For the approach without yeast extract, no growth data are available, but the growth should be smaller since growth in suspension is also observed to a lesser extent. It can be assumed that the ratio here is similar and, thus, the contribution to product formation might also be in the same order of magnitude.

For the pessimistic scenario, it is assumed that no cell growth occurs in the biofilm and that the cell concentration remains unchanged. This would mean that the contribution of cells in suspension to product formation would correspond to 70% (with yeast extract) or to 48% (without yeast extract).

By incorporating results from the literature [41], where it has been shown that the cell number of C. kluyveri in agar hydrogel increases significantly even at a starting concentration of 3.0 g L⁻¹, exhibiting similar cell proliferation to standard growth, the optimistic scenario is considered significantly more likely.

Furthermore, the cell leaking from the synthetic biofilm contrasts with the experiments conducted in anaerobic flasks with C. kluyveri, where no increase in CDW concentration in suspension was observed (cf. Figure 4). A possible explanation is that shear forces exerted by the liquid on the synthetic biofilm are higher in the biofilm reactor, leading to biofilm abrasion. Consequently, cells could be released into the suspension and proliferate there.

The space-time yield (STY), particularly its maximum value (STY_max), serves as an indicator of process efficiency for the approaches with and without yeast extract. The data for 1-butyrate and 1-hexanoate for the approaches with and without yeast extract are shown in

Table 1. Maximum space-time yield (STY_max) for 1-butyrate and 1-hexanoate per total reaction volume (631.4 mL) and per volume of the synthetic biofilm (31.4 mL) for the experiments with and without yeast extract.

Experiment	Reference Volume	1-Butyrate	1-Hexanoate
With yeast extract	631.4 mL	0.066 g L−1 h−1	0.246 g L−1 h−1
	31.4 mL	1.331 g L−1 h−1	4.947 g L−1 h−1
Without yeast extract	631.4 mL	0.017 g L−1 h−1	0.056 g L−1 h−1
	31.4 mL	0.332 g L−1 h−1	1.123 g L−1 h−1

. Relative to the biofilm volume, STY_max is logically significantly higher than when referenced to the total reaction volume. For the approach with yeast extract, this results in an increase in STY_max for 1-butyrate and 1-hexanoate by factors of 20.2 and 20.1, respectively. The approach without yeast extract yields increases by factors of 19.5 and 20.1 for 1-butyrate and 1-hexanoate. However, as previously discussed, not all cells involved in product formation were located within the biofilm. Consequently, the actual product formation rates relative to the biofilm volume are likely only about two-thirds to one-half of these values. Nevertheless, these STY_max values for C. kluyveri remain remarkable. This highlights the potential of fermentation processes using synthetic biofilms with C. kluyveri with and without yeast extract as a promising strategy for future investigations.

When comparing the cultivation of immobilized cells with yeast extract in the biofilm reactor to that in anaerobic flasks, the key differences—aside from the significantly higher initial cell concentration—are pH control in the biofilm reactor and the continuous supply of a gas mixture containing CO₂. A direct comparison between anaerobic flasks and the biofilm reactor reveals that complete ethanol consumption and a significantly more extensive conversion of acetate occur only in the biofilm reactor. The cessation of acetate conversion is directly linked to the depletion of ethanol, as both substrates are required for the metabolism of C. kluyveri. The complete consumption of ethanol in the biofilm reactor can be attributed to pH control, which prevents the pH from dropping into a range intolerable for C. kluyveri, as observed in the anaerobic flask experiments. This also explains the approximately 2 g L⁻¹ lower final 1-hexanoate concentration in the anaerobic flasks. For complete substrate conversion, pH control is therefore essential. Additionally, the liquid volume in the anaerobic flasks is 5.0 times larger than the biofilm volume, whereas in the biofilm reactor, it is 19.1 times larger. Consequently, the productivity of the cells within the biofilm is significantly higher in the biofilm reactor compared to the anaerobic flasks.

The results clearly demonstrate that cells within the synthetic biofilm are capable of producing 1-butyrate and 1-hexanoate, even in the absence of yeast extract. Synthetic biofilms of C. kluyveri show great potential for use in industrial processes without yeast extract, offering a cost-saving advantage. By increasing the initial biomass, the need for active cell growth during the process can be eliminated, further enhancing process efficiency.

4. Conclusions

C. kluyveri remains unaffected by inoculation temperatures up to 45 °C during the preparation of a synthetic biofilm made of agar hydrogel. This highlights C. kluyveri as an ideal candidate for immobilization within agar hydrogel to form a synthetic biofilm. The newly designed lab-scale biofilm reactor setup allows for fully controlled investigations of synthetic biofilms in contact with a defined gas phase on the top and a defined liquid phase at the bottom, offering significant advantages over conventional anaerobic bottle experiments.

Immobilized in a synthetic biofilm, C. kluyveri can produce 1-hexanoate without the addition of yeast extract in the medium, enabling more cost-effective processes. While this results in lower product concentrations, this limitation can be addressed by inoculating the synthetic biofilm with larger amounts of C. kluyveri cells at the start of the process. This approach minimizes substrate conversion into biomass and enhances the conversion of substrates into longer-chain fatty acids, such as 1-hexanoate.

The transfer and application of synthetic biofilms with agar hydrogels in scalable bioreactor configurations like, e.g., membrane bioreactors, fixed-bed biofilm reactors, or trickle-bed biofilm reactors, is presently under study. The synthetic biofilm approach can be adapted for large-scale bioreactors by tailoring biofilm preparation to reactor designs. In trickle-bed reactors, suitable carriers and coating methods must ensure defined hydrogel film thickness. In membrane reactors, applying the biofilm to the membrane surface supports substrate diffusion and stability. Scaling up requires optimizing mass transfer, shear forces, and biofilm stability for industrial performance.

In addition to the technical advancements, the potential for process economy improvements is substantial. The cost-effective approach of using synthetic biofilms, particularly in the absence of yeast extract, offers exciting prospects for industrial applications. Optimizing biofilm inoculation strategies and reactor design could further reduce production costs, making the process more competitive for the large-scale production of valuable chemicals such as 1-hexanoate. These efforts will contribute to the development of more sustainable and economically viable biotechnological processes.