A case of recto-sigmoid endometriosis mimicking carcinoma.

Introduction

Fermented beverages undergo diverse biochemical changes driven by microbial activity. Among them, kombucha is traditionally prepared from sweetened tea using a symbiotic culture of bacteria and yeast (SCOBY) and has attracted increasing research interest because of its dynamic fermentation patterns and complex chemical composition. During fermentation, yeasts hydrolyze sucrose and generate fermentable sugars and small amounts of ethanol, while acetic acid bacteria (AAB) and other associated microorganisms contribute to the formation of organic acids and additional metabolites (Jayabalan et al., 2014 ; Villarreal-Soto et al., 2018 ). As fermentation proceeds, acidity increases and the chemical composition changes markedly. Because these metabolic changes depend strongly on the substrate, kombucha-type fermentation provides a useful framework for examining how plant matrices influence fermentation behavior and compositional change.

Black tea has traditionally served as the primary substrate for kombucha fermentation. Recently, various plant materials have been used to diversify kombucha products (Anantachoke et al., 2023 ; Emiljanowicz and Malinowska-Pańczyk, 2020 ). Differences in phenolic composition, carbohydrate content, and other phytochemicals can shape microbial growth and metabolite formation, leading to distinct fermentation patterns and compositional profiles. These observations highlight the importance of characterizing fermentation outcomes when non-traditional substrates are introduced. In addition, for plant matrices rich in phenolic-related constituents, kombucha-type fermentation may provide a suitable system for examining whether mixed microbial metabolism is associated with changes in substrate-derived compounds (de Miranda et al., 2022 ; Villarreal-Soto et al., 2018 ).

Chenopodium formosanum Koidz., commonly known as Taiwanese quinoa, is a native annual plant widely cultivated in Taiwan and traditionally consumed as food or used in alcoholic fermentation (Tsai et al., 2010 ). While research has largely focused on its seeds, which contain phytochemicals such as betalains, rutin, and β-glucans (Chyau et al., 2018 ; Tsai et al., 2010 ), the leaves have received limited attention and are often discarded during harvest, making them an underutilized agricultural by-product (Chiu et al., 2025 ; Lin et al., 2024 ).

Although the leaves are less utilized, they contain phytochemicals associated with antioxidant-related properties (Chiu et al., 2025 ; Lin et al., 2023 , 2024 ; Wu et al., 2021 ). Previous fermentation-related studies on C. formosanum leaves have mainly focused on single-strain fermentation, such as Aspergillus oryzae fermentation, and have shown changes in phenolic-related parameters and antioxidant activity (Lin et al., 2023 ). Because C. formosanum leaves also contain rutin and other phenolic-related constituents, a kombucha-type fermentation system may be useful for examining whether mixed bacterial–yeast fermentation is associated with changes in the detectable phenolic-related profile of the leaf infusion during fermentation (Chyau et al., 2018 ; Lin et al., 2023 ; Tsai et al., 2010 ). However, little is known about how C. formosanum leaf infusion (CFLI) behaves in a defined mixed bacterial–yeast fermentation system involving yeasts, AAB, and lactic acid bacteria (LAB). Compared with single-strain fermentation, such a mixed bacterial–yeast system may involve coordinated microbial growth, sugar conversion, acid production, and changes in phenolic-related constituents. Therefore, evaluating CFLI in a defined kombucha-type system may help clarify how this underutilized plant matrix supports microbial dynamics and fermentation-associated compositional changes. In Taiwan, C. formosanum leaves have been used as food-related plant materials, supporting the relevance of evaluating this underutilized leaf material as a substrate for fermented beverage development.

In this study, a controlled kombucha-type fermentation system was established using individually propagated and taxonomically identified microbial strains. This approach was adopted not only to reduce variability associated with undefined microbial communities but also to provide a more controlled basis for examining substrate-dependent microbial dynamics and compositional change during fermentation. CFLI was prepared under selected brewing conditions and subsequently fermented. Changes in pH, titratable acidity, microbial populations, organic acids, sugar composition, and phenolic-related parameters were monitored throughout the fermentation period. In addition, Drosophila melanogaster was used to provide exploratory biological observations associated with the final CFLI and fermented sample matrices. The objective of this work was to characterize stage-dependent microbial and compositional changes in an underutilized plant matrix during defined kombucha-type fermentation and to clarify how a leaf-derived substrate behaves in a controlled mixed bacterial–yeast system relevant to food processing.

Materials and methods

Preparation of CFLI

Fresh C. formosanum leaves were obtained from Heart Sheltered Farm (Yunlin County, Taiwan), washed, and dried in a hot-air oven at 40 °C to constant weight. The dried leaves were stored at room temperature until use. For infusion preparation, dried leaves were placed in tea bags and extracted with deionized water at a solid-to-liquid ratio of 1:50 (w/v) at 90 °C for 5, 10, 20, or 30 min. After extraction, tea bags were removed, and the infusions were cooled and stored at − 20 °C before analysis.

Microbial strains, inoculum preparation, and establishment of OD 600 –CFU calibration curves

Yeast ( Kluyveromyces marxianus and Saccharomyces bayanus ), AAB ( Komagataeibacter xylinus and Gluconobacter oxydans ), and LAB ( Leuconostoc mesenteroides ) were used in this study. All strains were provided and taxonomically identified by the Agricultural Technology Research Institute (Miaoli County, Taiwan) based on 16S rRNA gene sequencing for bacteria and ITS sequencing for yeasts. According to the provider, this five-strain combination was selected from preliminary screening in sweetened black tea fermentation based on its acceptable fermentation performance and sensory characteristics under the tested conditions. The five-strain combination was used to represent the major functional microbial groups involved in the present defined kombucha-type fermentation system, including yeasts as sugar-utilizing fermentative microorganisms, acetic acid bacteria as contributors to oxidative metabolism and organic acid formation, and lactic acid bacteria as an additional acid-tolerant bacterial group associated with fermented plant-based systems.

Single colonies were inoculated into the corresponding liquid media (Table S1 ) and incubated at 30 °C to obtain cultures for fermentation. Cultures were first adjusted to an OD 600 of 1.0 using the corresponding broth. Serial dilutions were then prepared in phosphate-buffered saline and plated using pour or spread plate methods (Table S1 ). Plates were incubated at 30 °C for 3–5 days. Only calibration curves with coefficients of determination (R 2 ) greater than 0.99 were used for microbial enumeration.

Fermentation conditions, sampling, and physicochemical analysis

CFLI prepared under the selected hot-brewing condition (90 °C for 20 min) was supplemented with sucrose (10%, w/v) and boiled for 5 min. After cooling, the infusion was transferred into 250-mL glass bottles (working volume 200 mL) and inoculated with each propagated strain at 1% (v/v), yielding an initial cell density of approximately 1 × 10 6  CFU/mL for each strain. Fermentation was conducted under static conditions at 30 °C for 15 days, and samples (CFL kombucha, CFLK) were collected on days 0, 3, 6, 9, 12, and 15.

For analysis, liquid samples and homogenized samples (combined liquid and cellulose pellicle) were prepared. Microbial populations were enumerated by plate counting as described above. Subsequently, samples were heat-treated at 80 °C for 10 min, centrifuged, and stored at − 20 °C until further analysis. pH was measured using a calibrated pH meter (JENCO, USA). Titratable acidity was determined by titration with 0.1 N NaOH using phenolphthalein as an indicator and expressed as acetic acid equivalents (%).

HPLC analysis

HPLC analysis was performed using a Hitachi L-2000 system equipped with a diode array detector (L-2455) or a refractive index detector (L-2490) (Hitachi, Tokyo, Japan). Samples were diluted as necessary and filtered through 0.22 μm membrane filters before analysis.

Rutin, acetic acid, and glucuronic acid were analyzed using a Mightysil RP-18 column under the chromatographic conditions summarized in Table S2. Sugar composition (sucrose, glucose, and fructose) was determined using an Ultra Amino column with isocratic elution of acetonitrile–water (75:25, v/v) and refractive index detection. Quantification was performed using calibration curves constructed with authentic standards. Regression equations derived from the standard curves were used to calculate analyte concentrations from peak areas, and final concentrations were determined after correction for dilution factors. Only calibration curves with coefficients of determination (R 2 ) greater than 0.99 were accepted. Concentrations below the lowest calibration level used for quantification were reported as not detected (ND). For acetic acid, the lowest calibration level was 250 µg/mL (Table S2).

DPPH assay

DPPH radical scavenging activity was determined according to a previously reported method (Dibacto et al., 2021 ). Absorbance was measured at 517 nm. For CFLI samples prepared under different brewing conditions, IC 50 values were calculated from concentration–response curves. For CFLK samples, the inhibition rate (%) was determined at a fixed dilution ratio (1:10, v/v) to ensure comparability among fermentation time points.

Determination of total phenolic and flavonoid contents

Total phenolic content (TPC) and total flavonoid content (TFC) were determined using the Folin–Ciocalteu and aluminum chloride colorimetric methods, respectively, as previously reported (Fattahi et al., 2014 ). Absorbance was measured at 720 nm for TPC and 415 nm for TFC. Gallic acid and quercetin were used as reference standards, respectively, and results were expressed as μg gallic acid equivalents (GAE)/mL and μg quercetin equivalents (QE)/mL, respectively.

D. melanogaster experiments

Fly strains, husbandry, and diet preparation

Day 6 CFLK samples were selected for the supplementary biological assay because this time point represented an active stage of fermentation within the defined consortium. At day 6, yeast and LAB reached peak or near-peak populations, while AAB remained at a high level. This time point also corresponded to marked sucrose depletion, transient accumulation of glucose and fructose, and evident acidification of the system. Therefore, day 6 was selected based on the overall fermentation stage and microbial activity rather than on the maximum value of any single phenolic-related parameter.

Male w 1118
D. melanogaster aged 2–3 days after eclosion were maintained in transparent plastic vials (40 flies per vial) at 25 °C under a 12 h light/12 h dark cycle. Flies were fed a standard laboratory basal diet containing glucose, yeast powder, cornmeal, agar, and sucrose. For preparation of the samples used in the Drosophila assay, CFLI was concentrated under reduced pressure and freeze-dried. Day 6 CFLK samples were heat-treated at 80 °C for 10 min and then centrifuged at 3000 rpm for 10 min at 4 °C. Only the CFLK supernatants were collected, concentrated under reduced pressure, and subsequently freeze-dried. Therefore, the assay used freeze-dried CFLI powder and freeze-dried powder prepared from heat-treated and centrifuged day 6 CFLK supernatants. For CFLK, this preparation procedure was intended to reduce the contribution of live microorganisms and suspended microbial biomass in the fly diet. Experimental diets were prepared by supplementing the basal medium with freeze-dried CFLI powder (1.2 mg/mL) or freeze-dried day 6 CFLK powder (1.2 mg/mL), expressed as the concentration of added sample powder in the diet. The concentration of 1.2 mg/mL was selected as an empirically determined supplementation level for the fly model, with reference to a previous cell-based study (Lin et al., 2023 ), and was not intended as a direct cell-to-fly dose conversion. Diets were refreshed every 3–4 days.

Longevity and climbing ability assays

Flies were assigned to three groups: control (basal diet), CFLI (1.2 mg/mL), and CFLK (1.2 mg/mL). Only the higher supplementation level was used in the longevity and climbing assays as a representative exposure condition to observe whether measurable effects could be detected during long-term dietary treatment. Survival was recorded twice weekly until all flies had died. Climbing ability was evaluated using a negative geotaxis assay, and the percentage of flies climbing 5 cm within 18 s was recorded. Each experimental group consisted of three independent vials (40 flies per vial), and all assays were performed using these biological replicates.

Determination of antioxidant enzyme activities

Flies were divided into five groups: control (basal diet), CFLI (0.3 and 1.2 mg/mL), and CFLK (0.3 and 1.2 mg/mL). Two supplementation levels were used for antioxidant enzyme assays because these biochemical parameters were considered suitable for comparing responses at different concentrations and for observing whether a concentration-dependent tendency might be present. At days 10 and 30, flies were homogenized in ice-cold phosphate buffer and centrifuged. Supernatants were used to determine superoxide dismutase (SOD) and catalase (CAT) activities using commercial assay kits (Abbkine, China). Enzyme activities were normalized to total protein content determined by the Bradford method.

Statistical analysis

Experiments were conducted with three independent biological replicates (n = 3). Data are presented as mean ± standard deviation (SD). Statistical analyses were performed using IBM SPSS Statistics (version 20.0). Differences among multiple groups were evaluated by one-way analysis of variance (ANOVA) followed by Duncan’s multiple range test. For microbial population data, differences between liquid and homogenized samples at the same fermentation time point were evaluated using Student’s t-test. Survival curves were analyzed using the Kaplan–Meier method and compared using the log-rank test. Climbing ability was analyzed using one-way ANOVA followed by Dunnett’s post hoc test, with the control group as the reference. SOD and CAT activities were analyzed by one-way ANOVA followed by Duncan’s multiple range test. In all analyses, p < 0.05 was considered statistically significant.

Results and discussion

Selection of a suitable brewing time for CFLI as a fermentation substrate

The effects of different brewing times at 90 °C on phenolic-related parameters and antioxidant activity of CFLI are shown in Table  1 . TFC increased significantly from 5 to 10 min and remained unchanged thereafter. TPC increased progressively up to 20 min, with no significant difference between 20 and 30 min. Rutin content increased continuously throughout the tested brewing period, reaching its highest level at 30 min. The IC 50 value for DPPH radical-scavenging activity decreased significantly from 5 to 20 min, with no further significant change at 30 min. Based on these results, 20 min was selected as the brewing time for preparation of CFLI for subsequent fermentation experiments. Table 1 Effect of brewing time on phenolic-related parameters and antioxidant activity of Chenopodium formosanum leaf infusion Analyte Brewing time (min) 5 10 20 30 TFC (µg QE/mL) 462 ± 18 a 625 ± 19 b 650 ± 43 b 650 ± 12 b TPC (µg GAE/mL) 193 ± 8 a 224 ± 20 b 246 ± 8 c 247 ± 6 c Rutin (µg/mL) 101 ± 10 a 107 ± 3 a 187 ± 16 b 258 ± 13 c DPPH (IC 50 , mg/mL) 0.86 ± 0.05 a 0.74 ± 0.14 a 0.55 ± 0.09 b 0.56 ± 0.07 b Values are expressed as mean ± SD (n = 3). Different lowercase letters within the same row indicate significant differences among brewing times (p < 0.05). IC 50 values were calculated from concentration–response curves. TFC, total flavonoid content; TPC, total phenolic content; QE, quercetin equivalent; GAE, gallic acid equivalent

Table 1 Effect of brewing time on phenolic-related parameters and antioxidant activity of Chenopodium formosanum leaf infusion Analyte Brewing time (min) 5 10 20 30 TFC (µg QE/mL) 462 ± 18 a 625 ± 19 b 650 ± 43 b 650 ± 12 b TPC (µg GAE/mL) 193 ± 8 a 224 ± 20 b 246 ± 8 c 247 ± 6 c Rutin (µg/mL) 101 ± 10 a 107 ± 3 a 187 ± 16 b 258 ± 13 c DPPH (IC 50 , mg/mL) 0.86 ± 0.05 a 0.74 ± 0.14 a 0.55 ± 0.09 b 0.56 ± 0.07 b Values are expressed as mean ± SD (n = 3). Different lowercase letters within the same row indicate significant differences among brewing times (p < 0.05). IC 50 values were calculated from concentration–response curves. TFC, total flavonoid content; TPC, total phenolic content; QE, quercetin equivalent; GAE, gallic acid equivalent

Effect of brewing time on phenolic-related parameters and antioxidant activity of Chenopodium formosanum leaf infusion

Values are expressed as mean ± SD (n = 3). Different lowercase letters within the same row indicate significant differences among brewing times (p < 0.05). IC 50 values were calculated from concentration–response curves. TFC, total flavonoid content; TPC, total phenolic content; QE, quercetin equivalent; GAE, gallic acid equivalent

Although rutin continued to increase at 30 min, the absence of further significant changes in TFC, TPC, and DPPH radical-scavenging activity after 20 min indicated that the major phenolic-related and antioxidant responses had largely stabilized. Similar effects of brewing or extraction time on phenolic-related parameters have been reported for tea and herbal infusions (Fernando and Soysa, 2015 ; Salamatullah et al., 2021 ; Sentkowska et al., 2016 ). Therefore, the 20-min brewing condition was considered a practical preparation condition for evaluating subsequent fermentation-associated compositional changes.

Microbial population dynamics during fermentation

OD 600 –CFU calibration curves for all inoculated strains showed strong linearity (R 2  > 0.99; Table S3). Initial microbial counts were close to the intended inoculation levels. The combined yeast population and the combined AAB population were each approximately 10 6  CFU/mL, and the LAB population was also approximately 10 6  CFU/mL.

As shown in Fig.  1 A, yeast populations in liquid samples increased significantly from day 0 to day 3 and then declined progressively thereafter. In homogenized samples, yeast counts increased further and reached the highest level at day 6, followed by significant decreases at days 9, 12, and 15. Significant differences between liquid and homogenized samples were observed at days 6 and 12, whereas no significant differences were detected at the other time points. Fig. 1 Changes in A yeast, B acetic acid bacteria (AAB), and C lactic acid bacteria (LAB) populations in liquid and homogenized samples (liquid + cellulose pellicle) during fermentation of Chenopodium formosanum leaf infusion. Values are expressed as mean ± standard deviation (n = 3). Different lowercase and uppercase letters indicate significant differences among fermentation times within liquid and homogenized samples, respectively, as determined by one-way ANOVA followed by Duncan’s multiple range test (p < 0.05). Asterisks indicate significant differences between liquid and homogenized samples at the same fermentation time point, as determined by Student’s t-test (p < 0.05)

Changes in A yeast, B acetic acid bacteria (AAB), and C lactic acid bacteria (LAB) populations in liquid and homogenized samples (liquid + cellulose pellicle) during fermentation of Chenopodium formosanum leaf infusion. Values are expressed as mean ± standard deviation (n = 3). Different lowercase and uppercase letters indicate significant differences among fermentation times within liquid and homogenized samples, respectively, as determined by one-way ANOVA followed by Duncan’s multiple range test (p < 0.05). Asterisks indicate significant differences between liquid and homogenized samples at the same fermentation time point, as determined by Student’s t-test (p < 0.05)

AAB populations showed more pronounced differences between liquid and homogenized samples (Fig.  1 B). In liquid samples, AAB counts at days 0 and 3 did not differ significantly, but decreased significantly from day 6 onward. In homogenized samples, AAB populations were significantly higher at days 3, 6, and 9 than at day 0, with no significant differences among these three time points, and then declined significantly at days 12 and 15. Significant differences between liquid and homogenized samples were observed from day 3 to day 15.

LAB populations also changed significantly during fermentation in both sample types (Fig.  1 C). In liquid samples, LAB counts increased significantly to day 3, remained lower at day 6, and then declined progressively from day 9 to day 15. In homogenized samples, LAB populations increased significantly to day 3 and reached the highest level at day 6, followed by a gradual decline thereafter. Significant differences between liquid and homogenized samples were observed at days 3, 6, 12, and 15, whereas no significant differences were detected at days 0 and 9.

The differences between the liquid and homogenized samples indicate that the two sample types did not reflect identical microbial distribution during fermentation. The homogenized samples included both the fermentation broth and the cellulose pellicle, whereas the liquid samples represented only the broth fraction. Therefore, microorganisms retained in or associated with the pellicle structure were more likely to be represented in the homogenized samples. This pattern is consistent with the structural heterogeneity of kombucha fermentation systems, in which AAB are often enriched in the pellicle-associated region because of their aerobic growth characteristics (Jayabalan et al., 2014 ; Wang et al., 2022 ). In the present study, AAB showed the most pronounced differences between liquid and homogenized samples, whereas yeast and LAB exhibited more limited or stage-dependent differences between the two sample types. These results suggest that microbial enumeration based only on the liquid fraction may not fully reflect microbial distribution in this defined kombucha-type fermentation system.

Because homogenized samples contained both the fermentation broth and cellulose pellicle, they were used as the main basis for interpreting the overall microbial dynamics of the defined kombucha-type system. Taken together, the microbial profiles suggested three partially overlapping stages: an early active proliferation phase, a middle metabolic transition phase, and a late decline phase. The middle transition phase was defined cautiously because yeast and LAB reached their highest levels at day 6, while AAB remained relatively high through day 9. In this interpretation, yeast and LAB mainly contributed to early growth and sugar conversion, whereas AAB showed a more sustained contribution during the middle stage. This stage-based interpretation provides a basis for linking microbial dynamics with subsequent sugar utilization, acidification, organic acid accumulation, and phenolic-related changes.

Acidification, sugar utilization, and organic acid accumulation during fermentation

As shown in Fig.  2 , the initial pH of CFLI (approximately 5.4) decreased sharply during the first 3 days and then remained relatively stable from day 6 to day 15. In contrast, titratable acidity increased continuously throughout the fermentation period. Fig. 2 Changes in pH and titratable acidity during kombucha fermentation of Chenopodium formosanum leaf infusion. Values are expressed as mean ± standard deviation (n = 3). Different uppercase letters indicate significant differences in pH values, and different lowercase letters indicate significant differences in titratable acidity among fermentation times (p < 0.05)

Changes in pH and titratable acidity during kombucha fermentation of Chenopodium formosanum leaf infusion. Values are expressed as mean ± standard deviation (n = 3). Different uppercase letters indicate significant differences in pH values, and different lowercase letters indicate significant differences in titratable acidity among fermentation times (p < 0.05)

Sucrose was the predominant sugar before fermentation and decreased markedly throughout the fermentation period, particularly during the early stage. In contrast, glucose and fructose increased during the early stage, reaching their highest levels at day 6. Thereafter, both monosaccharides declined progressively during the later stages of fermentation (Table  2 ). Table 2 Time-course changes in selected phytochemical contents, antioxidant activity, organic acids, and sugar composition during kombucha fermentation of Chenopodium formosanum leaf infusion Items Fermentation time (day) 0 3 6 9 12 15 Acetic acid (µg/mL) ND 2922 ± 404 a 5318 ± 533 b 9302 ± 1986 c 18,369 ± 3750 d 24,480 ± 1247 e Glucuronic acid (µg/mL) 9141 ± 940 a 9443 ± 192 a 9967 ± 885 ab 11,506 ± 1415 bc 12,285 ± 1516 c 13,257 ± 1069 c Sucrose (µg/mL) 95,014 ± 6869 a 80,863 ± 6261 b 50,593 ± 6312 c 15,938 ± 2084 d 12,364 ± 1654 e 10,457 ± 2084 e Glucose (µg/mL) 761 ± 73 a 1004 ± 134 b 7876 ± 681 c 3080 ± 382 d 2917 ± 509 d 1629 ± 421 e Fructose (µg/mL) 298 ± 28 a 667 ± 50 b 2799 ± 345 c 2543 ± 573 c 2030 ± 264 d 802 ± 192 b TFC (µg QE/mL) 596 ± 17 a 604 ± 22 ab 643 ± 16 abc 667 ± 16 bc 692 ± 38 c 708 ± 57 c TPC (µg GAE/mL) 211 ± 8 a 210 ± 9 a 215 ± 10 a 268 ± 11 b 285 ± 11 bc 303 ± 12 c Rutin (µg/mL) 172 ± 5 a 237 ± 13 b 251 ± 13 bc 267 ± 11 c 296 ± 7 d 327 ± 7 e DPPH (Inhibition rate, %) 50 ± 1 a 51 ± 2 ab 53 ± 2 ab 55 ± 3 bc 58 ± 4 c 63 ± 2 d Values are expressed as mean ± standard deviation (n = 3). Different lowercase letters within the same row indicate significant differences among fermentation times (p < 0.05). DPPH inhibition rate was determined at a fixed dilution ratio of 1:10 (v/v). QE, quercetin equivalent; GAE, gallic acid equivalent; ND, not detected. For acetic acid, ND indicates that no quantifiable peak was detected above the lowest calibration level used for quantification (250 µg/mL)

Table 2 Time-course changes in selected phytochemical contents, antioxidant activity, organic acids, and sugar composition during kombucha fermentation of Chenopodium formosanum leaf infusion Items Fermentation time (day) 0 3 6 9 12 15 Acetic acid (µg/mL) ND 2922 ± 404 a 5318 ± 533 b 9302 ± 1986 c 18,369 ± 3750 d 24,480 ± 1247 e Glucuronic acid (µg/mL) 9141 ± 940 a 9443 ± 192 a 9967 ± 885 ab 11,506 ± 1415 bc 12,285 ± 1516 c 13,257 ± 1069 c Sucrose (µg/mL) 95,014 ± 6869 a 80,863 ± 6261 b 50,593 ± 6312 c 15,938 ± 2084 d 12,364 ± 1654 e 10,457 ± 2084 e Glucose (µg/mL) 761 ± 73 a 1004 ± 134 b 7876 ± 681 c 3080 ± 382 d 2917 ± 509 d 1629 ± 421 e Fructose (µg/mL) 298 ± 28 a 667 ± 50 b 2799 ± 345 c 2543 ± 573 c 2030 ± 264 d 802 ± 192 b TFC (µg QE/mL) 596 ± 17 a 604 ± 22 ab 643 ± 16 abc 667 ± 16 bc 692 ± 38 c 708 ± 57 c TPC (µg GAE/mL) 211 ± 8 a 210 ± 9 a 215 ± 10 a 268 ± 11 b 285 ± 11 bc 303 ± 12 c Rutin (µg/mL) 172 ± 5 a 237 ± 13 b 251 ± 13 bc 267 ± 11 c 296 ± 7 d 327 ± 7 e DPPH (Inhibition rate, %) 50 ± 1 a 51 ± 2 ab 53 ± 2 ab 55 ± 3 bc 58 ± 4 c 63 ± 2 d Values are expressed as mean ± standard deviation (n = 3). Different lowercase letters within the same row indicate significant differences among fermentation times (p < 0.05). DPPH inhibition rate was determined at a fixed dilution ratio of 1:10 (v/v). QE, quercetin equivalent; GAE, gallic acid equivalent; ND, not detected. For acetic acid, ND indicates that no quantifiable peak was detected above the lowest calibration level used for quantification (250 µg/mL)

Time-course changes in selected phytochemical contents, antioxidant activity, organic acids, and sugar composition during kombucha fermentation of Chenopodium formosanum leaf infusion

Values are expressed as mean ± standard deviation (n = 3). Different lowercase letters within the same row indicate significant differences among fermentation times (p < 0.05). DPPH inhibition rate was determined at a fixed dilution ratio of 1:10 (v/v). QE, quercetin equivalent; GAE, gallic acid equivalent; ND, not detected. For acetic acid, ND indicates that no quantifiable peak was detected above the lowest calibration level used for quantification (250 µg/mL)

Acetic acid was not detected before fermentation, but appeared at day 3 and increased progressively throughout the fermentation period. In contrast, glucuronic acid was already detectable before fermentation. Its concentration showed only minor changes during the early stage and increased gradually during the later stage. Compared with acetic acid, glucuronic acid showed a smaller overall increase during fermentation (Table  2 ).

Consistent with the stage-based microbial pattern described above, sugar utilization and acid accumulation also changed progressively during fermentation. In the homogenized samples, yeast and LAB increased during the early stage. This increase was accompanied by a marked decline in sucrose and transient increases in glucose and fructose, indicating that sucrose was rapidly hydrolyzed and converted into fermentable monosaccharides during the active stage of fermentation. Such a pattern has been reported in kombucha fermentation, in which sucrose is hydrolyzed into glucose and fructose, which subsequently support microbial growth and metabolism (Jayabalan et al., 2014 ; Villarreal-Soto et al., 2018 ). As fermentation proceeded, the marked decrease in pH and continuous increase in titratable acidity indicated progressive acidification of the system. The rapid acidification during the early stage is consistent with organic acid formation associated with mixed microbial metabolism in kombucha fermentation (Malbaša et al., 2008 ). Because pH and titratable acidity reflect different aspects of acidification, the continued increase in titratable acidity despite relatively stable pH values indicates that total acid accumulation persisted during the later stage of fermentation (Geraris Kartelias et al., 2023 ). This continued acidification may also have contributed to the later decline in yeast and LAB populations. Previous studies have shown that acid accumulation can inhibit yeast growth and that LAB may become less competitive under increasingly acidic conditions (Chen et al., 2016 ; Li et al., 2023 ; Lund et al., 2020 ).

In contrast, AAB showed a broader high-population phase in the homogenized samples, extending from days 3 to 9, suggesting a more sustained contribution during the middle and later stages of fermentation. This pattern was accompanied by progressive accumulation of acetic acid, consistent with a greater contribution of AAB-associated oxidative metabolism as fermentation progressed. In kombucha systems, AAB are known to contribute substantially to acetic acid formation and are closely associated with the later oxidative phase of fermentation (Antolak et al., 2021 ; Martínez Leal et al., 2018 ). The different pattern observed for glucuronic acid should be interpreted cautiously. The detection of glucuronic acid before fermentation indicates that the CFLI matrix already contained a measurable background level of this compound. Chenopodium species have been reported to contain glucuronic acid-bearing triterpenoid saponins (Grabowska et al., 2024 ), suggesting that plant-derived constituents may partly explain the initial glucuronic acid level. In addition, kombucha-related systems may generate glucuronic acid through microbial oxidative metabolism (Trček et al., 2021 ). Therefore, the gradual increase in glucuronic acid during fermentation should be interpreted conservatively as reflecting the combined effects of the initial plant matrix background and possible fermentation-associated formation.

Changes in phenolic-related parameters and antioxidant activity during fermentation

TFC, TPC, and rutin content increased progressively during fermentation (Table  2 ). DPPH radical-scavenging activity also increased over time, but the magnitude of change was smaller than that observed for TFC, TPC, and rutin (Table  2 ). The continuous increase in rutin further indicates that fermentation was associated with changes in the detectable phenolic-related profile of CFLI.

Similar increases in phenolic-related parameters have been reported in kombucha and other plant-based fermented systems (Bueno-Rojas et al., 2025 ). These changes may be associated with the release of bound phenolic compounds or the transformation of existing constituents during microbial fermentation (Xia et al., 2019 ; Zhou et al., 2022 ). However, the smaller increase in DPPH radical-scavenging activity suggests that the change in antioxidant activity was not strictly proportional to the measured phenolic indices. Together, these findings indicate progressive fermentation-associated changes in the phenolic profile of CFLI.

Supplementary exploratory observations in D. melanogaster

D. melanogaster is commonly used as a model organism to evaluate oxidative stress–related parameters and age-associated physiological changes (Ogienko et al., 2022 ). Day 6 CFLK samples were selected for this exploratory assay because this time point represented an active stage of defined consortium fermentation. At this stage, yeast and LAB reached peak or near-peak populations, AAB remained at a high level, sucrose had markedly declined, glucose and fructose showed transient accumulation, and acidification was evident. Therefore, day 6 was selected based on the overall fermentation stage rather than on the maximum value of any single phenolic-related parameter.

The overall survival curves did not differ significantly among the control, CFLI, and CFLK groups (Fig.  3 A). At later time points, the CFLI and CFLK groups showed slightly higher survival rates than the control group, but this difference was not statistically significant. All groups declined to near zero by the end of the assay. Climbing ability decreased with age in all groups (Fig.  3 B). Compared with the control group, higher climbing percentages were observed in the CFLI group at several intermediate time points and in the CFLK group over a broader assay period. Fig. 3 Exploratory biological observations in Drosophila melanogaster associated with dietary exposure to Chenopodium formosanum leaf infusion (CFLI) and day 6 fermented CFLI kombucha-type sample (CFLK). CFLI was prepared as a freeze-dried powder. Day 6 CFLK samples were heat-treated at 80 °C for 10 min and centrifuged at 3000 rpm for 10 min at 4 °C; only the supernatants were collected, concentrated under reduced pressure, and freeze-dried before incorporation into the fly diet. A Survival rate and B climbing ability of flies fed a control diet, CFLI (1.2 mg/mL), or CFLK (1.2 mg/mL). C Superoxide dismutase (SOD) activity and D catalase (CAT) activity in flies after 10 and 30 days of dietary exposure. Survival curves were analyzed using the Kaplan–Meier method and compared using the log-rank test. Values in panels B–D are expressed as mean ± standard deviation. For climbing ability, * and # indicate significant differences compared with the control group for CFLI and CFLK, respectively, at the same time point (p < 0.05). For SOD and CAT activities, different letters indicate significant differences among treatment groups (p < 0.05). Because a sucrose-supplemented, matrix-matched unfermented CFLI control was not included, these data should be interpreted as exploratory responses to the final sample matrices rather than evidence of fermentation-specific physiological effects

Exploratory biological observations in Drosophila melanogaster associated with dietary exposure to Chenopodium formosanum leaf infusion (CFLI) and day 6 fermented CFLI kombucha-type sample (CFLK). CFLI was prepared as a freeze-dried powder. Day 6 CFLK samples were heat-treated at 80 °C for 10 min and centrifuged at 3000 rpm for 10 min at 4 °C; only the supernatants were collected, concentrated under reduced pressure, and freeze-dried before incorporation into the fly diet. A Survival rate and B climbing ability of flies fed a control diet, CFLI (1.2 mg/mL), or CFLK (1.2 mg/mL). C Superoxide dismutase (SOD) activity and D catalase (CAT) activity in flies after 10 and 30 days of dietary exposure. Survival curves were analyzed using the Kaplan–Meier method and compared using the log-rank test. Values in panels B–D are expressed as mean ± standard deviation. For climbing ability, * and # indicate significant differences compared with the control group for CFLI and CFLK, respectively, at the same time point (p < 0.05). For SOD and CAT activities, different letters indicate significant differences among treatment groups (p < 0.05). Because a sucrose-supplemented, matrix-matched unfermented CFLI control was not included, these data should be interpreted as exploratory responses to the final sample matrices rather than evidence of fermentation-specific physiological effects

SOD and CAT activities are shown in Fig.  3 C and D. At day 10, both enzyme activities were higher in the CFLI- and CFLK-treated groups than in the control group, and the 1.2 mg/mL groups showed higher enzyme activities than the corresponding 0.3 mg/mL groups. At day 30, SOD and CAT activities were lower than those at day 10 in all groups, although treated groups still showed higher enzyme activities than the corresponding control group.

These observations indicate that dietary exposure to CFLI- and CFLK-derived sample matrices was associated with measurable biological responses under the present assay conditions. However, these results should be interpreted cautiously. CFLK was prepared from CFLI supplemented with sucrose and fermented using the defined microbial consortium, whereas the CFLI group did not include an equivalent sucrose-supplemented, matrix-matched unfermented control processed in parallel. Therefore, the observed responses cannot be attributed specifically to fermentation-associated transformations of C. formosanum leaf constituents. Residual sucrose, accumulated glucose and fructose, acidity, organic acids, caloric differences, and other matrix-level changes may have contributed to the observed responses. Accordingly, the D. melanogaster assay is presented as an exploratory observation of the final sample matrices rather than as evidence of fermentation-specific physiological effects.

Overall interpretation and relevance of the defined fermentation system

Overall, this study showed that CFLI can support a defined kombucha-type fermentation process characterized by coordinated changes in microbial populations, sugar utilization, organic acid accumulation, and phenolic-related parameters. Compared with previous studies based on single-strain fermentation, the present work employed a defined mixed bacterial–yeast fermentation system to characterize stage-dependent fermentation-associated changes in CFLI. These findings indicate that CFLI can serve as a suitable plant-derived matrix for controlled mixed-culture fermentation and highlight its potential relevance to fermented beverage development.

The D. melanogaster assay provided supplementary exploratory observations under the present assay conditions. As discussed above, these observations should be interpreted as responses to the final sample matrices rather than as fermentation-specific physiological effects. Therefore, the main contribution of this study lies in the characterization of fermentation-associated microbial and compositional changes under defined microbial conditions. Future studies should include matrix-matched controls, targeted metabolite tracking, strain-specific contribution analysis, and product-oriented evaluation to further support the development of this defined fermented beverage system.

Supplementary Information

Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 22 KB)

Supplementary file1 (DOCX 22 KB)