Rumen fluid used for the in vitro incubation was collected from adult goats immediately after routine slaughter at a licensed local slaughterhouse. As the study did not involve any live-animal handling, experimental intervention, anesthesia, or euthanasia performed for research purposes, formal Animal Ethics Committee approval was not required under institutional policy for the use of abattoir by-products. Nevertheless, permission to collect rumen contents was obtained from the slaughterhouse management, and sampling was conducted only after slaughter as part of normal commercial operations, without altering animal handling or welfare procedures.

Rumen contents were collected within 15 min post-slaughter using an aseptic technique, transported in sealed sterile containers, and maintained under CO₂ flushing to preserve anaerobiosis and minimize contamination. All personnel used appropriate personal protective equipment (laboratory coats, gloves, protective eyewear, and masks as required), and all procedures involving rumen fluid and microbial cultures were performed in accordance with institutional biosafety and hygiene protocols and under Biosafety Level 2 practices (e.g., controlled access, disinfected work surfaces, safe handling of biological materials, and appropriate waste segregation).

All biological waste (rumen residues, culture materials, and disposables) was decontaminated (e.g., autoclaving and/or approved chemical disinfection) before disposal according to institutional and national biohazard waste management regulations. No human participants were involved; therefore, informed consent was not applicable.

This study was conducted from April to December 2024 at the Faculty of Animal Science, Feed Industry Technology Laboratory, Ruminant Nutrition Laboratory, and Non-Ruminant Nutrition Laboratory, Universitas Andalas, Indonesia. SCFA analysis was performed at the Indonesian Research Institute for Animal Production (Balitnak), Bogor, Indonesia.

The probiotic strains used in this study were S. harbinensis LH991 and P. kudriavzevii B-5P, obtained from the culture collection of the Feed Industry Technology Laboratory, Faculty of Animal Science, Universitas Andalas. Both strains were previously isolated from fermented fish (Budu) and molecularly identified by 16S rRNA gene sequencing for S. harbinensis and internal transcribed spacer region sequencing for P. kudriavzevii, showing ≥99% sequence similarity to reference strains.

Stock cultures were preserved at −80°C in media supplemented with 20% (v/v) glycerol. Before experimentation, S. harbinensis was revived in de Man, Rogosa, and Sharpe broth and incubated anaerobically at 37°C for 24–48 h until the mid-log phase (OD₆₀₀ ≈ 0.8). P. kudriavzevii was cultured in yeast peptone dextrose medium at 35°C–37°C for 24–48 h to reach the exponential phase.

Cell concentrations were adjusted to approximately 1 × 10¹² colony-forming units (CFU)/mL for each strain. The inoculum was prepared by mixing bacterial and yeast cultures at a 1:1 ratio by CFU count. The final inoculum dose used for incubation was 4% (v/v) of the total fermentation volume.

Rumen fluid was collected immediately after slaughter from adult goats (approximately 45–55 kg body weight) fed a conventional diet consisting of forage and concentrate at a ratio of approximately 60:40. Rumen contents were collected within 15 min post-slaughter, filtered through four layers of sterile cheesecloth, and maintained under continuous CO₂ flushing to ensure anaerobic conditions.

The rumen fluid pH at collection ranged from 6.5 to 6.8. The filtered rumen fluid was maintained at 39°C before incubation and used within 1 h of collection.

The incubation substrate consisted of 60% forage and 40% dry matter (DM) concentrate. The forage component primarily consisted of dried grass, whereas the concentrate comprised corn meal, soybean meal, rice bran, and mineral–vitamin premix.

The chemical composition of the substrate (% DM) was as follows: dry matter, 90.2%; crude protein, 14.8%; neutral detergent fiber, 42.5%; acid detergent fiber, 27.1%; ether extract, 4.2%; and ash, 8.1%. The substrates were dried at 60°C, ground to pass through a 1-mm sieve, and stored in airtight containers before use.

A central composite design (CCD) within the response surface methodology (RSM) framework was used to evaluate the effects of three independent variables: C concentration (glucose; 0.1–0.3 g), N concentration (yeast extract; 5–15 g), and incubation time (24–72 h). A face-centered CCD consisting of 20 experimental runs, including six axial points and six replicates at the center point, was applied.

The response variables were individual SCFA concentrations (acetate, propionate, butyrate, iso-butyrate, valerate, and iso-valerate). A second-order polynomial model was fitted using Design-Expert software version 13 (Stat-Ease Inc., Minneapolis, MN, USA).

Each anaerobic fermentation unit contained 2.5 g of substrate, 200 mL of buffer solution (Tilley and Terry method), and 50 mL of rumen fluid. The prepared probiotic inoculum was added according to treatment specifications, followed by glucose and yeast extract as C and N sources, respectively.

Incubations were performed at 39°C under anaerobic conditions without agitation. At the end of the incubation period, microbial activity was halted by immersing the tubes in an ice-water bath.

Fermentation contents were centrifuged at 1,509 × g for 5 min at 4°C. Supernatants were collected and stored at −18°C until analysis. SCFA concentrations were determined using gas chromatography (GC; Shimadzu GC-2014, Japan) equipped with a flame ionization detector and a fused silica capillary column (30 m × 0.32 mm × 0.25 μm). Injector and detector temperatures were set at 250°C. The oven temperature was programmed from 100°C to 180°C at 10°C/min and held for 5 min. N was used as the carrier gas. SCFA standards were used for calibration, and results were expressed in millimolar units.

Data were expressed as mean ± standard error of the mean. RSM modeling was performed using Design-Expert software version 13 (https://www.statease.com, Stat-Ease, Inc., USA). A one-way analysis of variance followed by Tukey’s post hoc test was conducted using SPSS version 29.0 (IBM Corp., NY, USA) to compare treatment means. Model adequacy was evaluated using R², adjusted R², predicted R², and lack-of-fit tests. Statistical significance was set at p < 0.05.

The quadratic polynomial models generated for each SCFA were significant (p < 0.05). The coefficient of determination (R²) ranged from 0.82 to 0.94, indicating good agreement between the predicted and observed values. Adjusted R² and predicted R² values were in close agreement (<0.2 difference), confirming the model’s reliability. Lack-of-fit tests were non-significant (p > 0.05), demonstrating adequate model fitness. Residual plots showed a random distribution, supporting the validity of the model.

The C concentration, N concentration, and incubation time had a significant influence on acetic acid productivity (Table 1). Among the tested C concentrations, 0.2 g yielded the highest productivity, with an average of 151.087 mM. In contrast, a lower C concentration (0.1 g) consistently produced lower productivity, while a 0.3 g C concentration, although yielding moderately high results, did not exceed the performance observed at 0.2 g. The optimal N concentration was 10 g, yielding the highest incubation output when combined with 0.2 g of C and 48 h of incubation. N levels of 5 g generally led to the lowest productivity, whereas 15 g yielded moderate but less consistent results. In comparison, incubation for 24 h led to moderate results, whereas treatments for 72 h showed relatively high productivity. The most favorable conditions for maximizing incubation productivity were identified as 0.2 g of C, 10 g of N, and 48 h of incubation.

The response surface plots (Figure 1) in three-dimensional form showed color variations. The highest acetic acid production was indicated by the red area, corresponding to C concentration of 0.2 g, N concentration of 15 g, and incubation time of 24 h. The graph shows several regions in different colors, with the blue area indicating lower acetic acid productivity and the red area indicating higher productivity.

The highest propionic acid productivity of 75.177 mM (Table 2) was observed at a low C concentration (0.1 g), a high N concentration (15 g), and a longer incubation time (72 h). N plays a crucial role, with higher concentrations generally correlated with increased productivity. Incubation time also affects productivity, with longer incubation times primarily enhancing output at favorable C and N concentrations.

The results showed that butyric acid productivity was affected by C and N concentrations and incubation time (Table 3). Maximum productivity (17.443 mM) was recorded at 0.1 g of C, 15 g of N, and 24 h of incubation. Higher N concentration and shorter incubation time resulted in increased productivity, particularly when C was low. Conversely, lower productivity was associated with longer incubation times and low N or high C concentrations. These observations emphasize the importance of optimizing N levels and incubation duration to enhance butyric acid production. The response surface plots (Figure 3) support this result, where the green region rather than the red region represents the highest productivity, corresponding to the conditions of 0.1 g C, 15 g N, and 24 h of incubation.

The productivity of isobutyric acid was affected by C and N concentrations and incubation time (Table 4). The highest productivity occurred at a high N concentration of 15 g and a longer incubation time of 72 h, with a low C concentration of 0.1 g. Moderate C (0.2 g) and 48-h incubation produced consistent results, whereas short incubation (24 h) generally led to lower productivity. The outcome depends on the interaction of all three factors. The response surface plots (Figure 4) show that the highest isobutyric acid production was indicated by the green area, corresponding to a C concentration of 0.1 g, a N concentration of 15 g, and an incubation time of 72 h.

The average productivity varied with C and N concentrations and incubation time (Table 5). The highest productivity was observed at 0.1 g of C, 15 g of N, and 24 h (3.340 mM), followed by 0.2 g of C, 10 g of N, and 48 h (4.527 mM), suggesting that specific combinations are more effective. In general, a high N concentration (15 g) and shorter incubation time (24 h) produced better results. The interaction among all three factors played a crucial role in productivity optimization. The response surface plots (Figure 5) showed that the highest valeric acid production was indicated by the green area, corresponding to C concentration of 0.1 g, N concentration of 15 g, and incubation time of 24 h.

Achieving maximum productivity depends on a specific combination of C and N concentrations and incubation duration (Table 6). The highest iso-valeric acid productivity was produced at a high N concentration (15 g), particularly when paired with either extended (72 h) or shortened (24 h) incubation times, yielding productivity of 8.690 mM and 6.903 mM, respectively. Consistently good results were also obtained with moderate C concentration (0.2 g) and 48 h of incubation. As illustrated in the response surface plots (Figure 6), the optimal production zone shown in green corresponds to a C concentration of 0.1 g, a N concentration of 15 g, and an incubation time of 72 h.

For the first time, this study demonstrates that a co-culture of S. harbinensis and P. kudriavzevii exhibits nutrient-dependent metabolic switching under rumen-simulated conditions. The yeast likely contributes to redox balance and precursor supply, while the LAB directs C flux toward SCFA synthesis. This synergistic interaction explains the enhanced production of propionate and branched-chain SCFAs in low C and N-rich environments.

Moderate C availability (0.2 g) favored acetate production, reflecting overflow metabolism under sufficient glycolytic flux. In contrast, C limitation combined with N enrichment redirected metabolism toward propionate, butyrate, and iso-SCFAs, likely through enhanced amino acid catabolism and Wood–Werkman and Ehrlich pathway activation. Extended incubation further intensified iso-SCFA formation, supporting their classification as secondary metabolites.

Most postbiotic optimization studies rely on monocultures or synthetic media. The present rumen-based co-culture system more accurately mimics host-associated fermentation and reveals metabolic dynamics that are not observed in conventional de Man, Rogosa and Sharpe and Yeast Peptone Dextrose systems systems. The findings extend previous reports by demonstrating that SCFA spectrum modulation can be controlled through nutrient management rather than strain replacement.

Glucose is a fundamental energy source for microbial incubation, driving glycolysis and leading to pyruvate production, a key precursor for the biosynthesis of SCFAs. Acetic acid production was highest at a glucose concentration of 0.2 g. A decline in acetate production occurred when the glucose level exceeded or fell below this concentration. As a primary substrate for S. harbinensis and P. kudriavzevii, C is essential to support acetogenic metabolism without triggering substrate inhibition or metabolic overflow. Excessive C input can lead to acidification of the medium, disrupting microbial balance and subsequently slowing down acetate biosynthesis [12].

In contrast to acetic acid, propionic acid reached the highest production at a glucose concentration of 0.1 g by S. harbinensis and P. kudriavzevii. Excess glucose can trigger acid stress and inhibit the growth of propionate-producing microbes such as Propionibacterium spp. [13]. Furthermore, glucose overload induces acid stress, resulting in the inhibition of key enzymes that play a role in the succinate–propionate pathway [14]. Low glucose availability may help establish a favorable redox environment for propionate synthesis under anaerobic conditions.

Butyric and isobutyric acid production by S. harbinensis and P. kudriavzevii was optimal at a low glucose concentration of 0.1 g. The limited C supply in the environment promotes more efficient energy conservation through butyrate incubation, as observed in Clostridium species [15]. Moreover, branched-chain fatty acids such as iso-butyrate, which are often derived from amino acid catabolism, are not severely dependent on high carbohydrate availability, further supporting the efficiency of low C conditions for production [16, 17].

In contrast to the production of other SCFAs, the production of valeric and iso-valeric acids by S. harbinensis and P. kudriavzevii demonstrated dual production peaks at C concentrations of 0.1 and 0.2 g. N levels and incubation time also influenced the peaks. This suggests that C availability is not the primary determining factor for longer-chain SCFAs, but the interaction between C, N, and incubation time plays a more critical role [18].

The results showed that N source concentration, provided in the form of yeast extract, plays a significant role in the biosynthesis of SCFAs. Across all measured SCFAs, acetic, propionic, butyric, isobutyric, valeric, and iso-valeric acids, N availability had a significant effect on both acid production and formation time.

The highest productivity for most SCFAs by S. harbinensis and P. kudriavzevii was observed at a yeast extract concentration of 15 g/L, particularly when combined with low-to-moderate C concentration and optimized incubation times. Peak concentrations of propionic and iso-butyric acids were achieved under high N conditions (15 g/L) when paired with extended incubation (72 h) and a low C concentration (0.1 g). Similarly, butyric and valeric acids demonstrated higher productivity under the same high N conditions but with shorter incubation time (24 h), specifically at low C concentration.

The optimal N concentration for acetic acid production was 10 g, particularly when combined with 0.2 g of C and a 48-h incubation time. A lower N concentration (5 g) consistently resulted in the lowest productivity, whereas a higher level (15 g) produced moderate productivity. These results suggest that N functions as a regulatory factor in balancing cell growth and metabolite production. Excessive N may promote biomass accumulation at the expense of product formation, whereas insufficient N can limit enzymatic activity and reduce cell viability [19].

For propionic acid, a high N concentration (15 g) combined with low C (0.1 g) and an extended incubation time (24–72 h) resulted in the highest productivity (75.177 mM). This is in line with previous studies indicating that elevated N availability can enhance propionate biosynthesis by upregulating the expression of propionyl-CoA carboxylase and other key enzymes associated with the Wood–Werkman cycle [20, 21]. Based on the results, N-rich conditions may support prolonged metabolic activity and gradual product accumulation over time.

Branched-chain fatty acids, such as iso-butyrate and iso-valerate, are known to originate from the catabolism of branched-chain amino acids, including valine and leucine, and may benefit from the amino acid-rich composition of yeast extract [22]. Both acids exhibited peak productivity at an N concentration of 15 g, specifically under low C conditions and short to moderate incubation times. This suggests that high N levels support the metabolic vitality required for activation of butyrate synthesis pathways. Iso-butyrate, in particular, demonstrated a preference for longer incubation (72 h) under high N conditions, implying that N may also extend microbial activity and metabolite synthesis under controlled environments.

The production of valeric and iso-valeric acids increased under high N conditions (15 g). The highest valeric acid productivity (4.527 mM) was achieved at an N concentration of 10 g and an incubation time of 48 h, suggesting a slightly different optimal balance for this compound. However, under most other conditions, 15 g of N produced either higher or comparable results. Iso-valeric acid reached the peak productivity (8.690 mM) at 15 g N, particularly when combined with either 24 or 72 h incubation times. These results underscore the critical role of N in sustaining microbial metabolism across varying incubation times.

High N concentrations (particularly 15 g) generally promoted greater acid productivity across all acid types, although the degree of impact varied by acid type and incubation time. N is a key building block for amino acids, nucleic acids, and coenzymes, which are essential for cell growth and enzymatic functions. However, excess N can sometimes favor biomass over metabolite production, requiring a fine-tuned balance depending on the target product and microbial strain [23].

Incubation time is a factor that influences microbial growth, substrate utilization, and metabolite accumulation. The results showed that incubation time (24, 48, and 72 h) significantly affected the production of various organic acids, with each acid demonstrating a different optimal incubation time depending on the C and N concentrations.

The highest acetic acid productivity by S. harbinensis and P. kudriavzevii was obtained after 48 h incubation with 0.2 g C and 10 g N. A shorter incubation time (24 h) resulted in moderate productivity, while extending the time to 72 h did not lead to further improvement and, in some cases, even reduced productivity. This pattern is consistent with a previous study by Song et al. [24], who reported that maximum acetic acid accumulation occurred near the late exponential growth phase, when acetaldehyde dehydrogenase activity peaked before a drop in pH inhibited metabolism. The decline observed at 72 h can be attributed to substrate depletion or product inhibition effects common in prolonged incubation [25].

The production of propionic acid increased with extended incubation up to 72 h, particularly under low C (0.1 g) and high N (15 g) conditions. This suggests that S. harbinensis and P. kudriavzevii require prolonged metabolic activity to convert intermediates through the Wood–Werkman cycle. Zhang et al. [26] similarly reported that propionate productivity peaks during the early to late stationary phase (48–96 h), driven by Nicotinamide Adenine Dinucleotide + hydrogen accumulation necessary for acrylate reduction to propionate.

The maximum butyric acid production (17.443 mM) occurred at 24 h with 0.1 g C and 15 g N, whereas productivity decreased when incubation was extended to 72 h. This pattern indicates that S. harbinensis and P. kudriavzevii achieved optimal butyrate synthesis during the late exponential phase, after which butyrate kinase activity was inhibited by acid accumulation and pH decline. Similar results were reported by Esquivel-Elizondo et al. [27], who observed a significant decrease in butyrate levels due to a metabolic transition toward solventogenesis or gas production during extended incubation.

Isobutyric acid showed a different pattern, indicating maximum productivity at 72 h under low C (0.1 g) and high N (15 g) conditions. The delayed production suggests that isobutyric acid formation is a secondary metabolic pathway relying on valine-derived precursor accumulation. The biosynthesis of isobutyric acid and valeric acid from the catabolism of branched-chain amino acids in Saccharomyces cerevisiae through the Ehrlich pathway comprises three reaction steps: transamination, decarboxylation, and oxidation [28].

Valeric acid showed dual optima at 24 h (0.1 g C, 15 g N) and 48 h (0.2 g C, 10 g N). This indicates a strong interaction between incubation time and substrate composition, with valerate synthesis depending on the acetate and butyrate precursor balance. Factors such as feedstock characteristics, biocatalyst type, reactor configuration, initial pH, organic loading rate, temperature, and retention time significantly influence the productivity of volatile fatty acids [29].

Iso-valeric acid followed a similar trend to isobutyric acid, achieving the highest productivity at 72 h with high N (15 g), while moderate productivity was obtained at 24 and 48 h. The extended optimum corroborates the results of Thierry et al. [30], who stated that iso-valeric acid and α-ketoisocaproic acid were the main products of leucine catabolism, which intensified during prolonged incubation.

To the best of our knowledge, this is the first study to optimize postbiotic production using a rationally selected LAB–yeast consortium, revealing the synergistic metabolic conversions that enhance SCFA yields under controlled rumen in vitro conditions. Unlike previous studies that focused on single metabolites or single-source inocula, this study uniquely models and optimizes a full SCFA spectrum using a multifactor RSM approach, enabling the prediction of metabolite shifts under varying nutrient conditions.

Acetate is maximized at moderate C (0.2 g) and N (10 g). Propionate, butyrate, isobutyric, and iso-valeric acids require very low C (0.1 g) and high N (15 g). This demonstrates that nutrient-driven metabolic routing can alter pathways depending on environmental conditions. Ad originality: The study reveals, for the first time, distinct nutrient-dependent metabolic switching in a mixed-culture system, highlighting how substrate limitation and N enrichment drive different SCFA pathways selectively. This study reveals, for the first time, distinct nutrient-dependent metabolic switching in a mixed-culture system, highlighting how substrate limitation and N enrichment drive different SCFA pathways selectively. Combining LAB and yeast strains with complementary metabolic roles can significantly enhance SCFA biosynthesis, offering a new strategy for developing enriched postbiotic formulations. High N with prolonged incubation disproportionately enhances iso-SCFA synthesis. Branched-chain SCFA formation in co-culture systems suggests that N-rich environments promote amino acid catabolism through the Ehrlich pathway.

Higher propionate production, which is associated with improved energy efficiency in ruminants, and increased butyrate levels that support gut barrier integrity and exhibit anti-inflammatory effects, as well as balanced acetate production, are associated with improved energy efficiency in ruminants. This creates a scalable blueprint for cost-effective, nutrient-based optimization of postbiotics. This study proposes a resource-efficient approach for industrial postbiotic production with implications in animal feed, functional foods, and microbial fermentation technologies by identifying low C, high N conditions that maximize energy-dense SCFAs. The integration of rumen-derived microbial ecosystems with controlled nutrient optimization provides an innovative bioprocessing platform that mimics the dynamics of natural host fermentation, an approach that has not been previously applied in postbiotic research.