A strong tea infusion was prepared with CTC tea (1% w/v) for fifteen minutes so as to obtain alkaloids in the infusion as much as possible. Sucrose (6% w/v) was used as nutrient for the microbial culture inside the tea broth. Following protocols of Greenwalt et al. (2000) and Zhu et al. (2014), a total of five batches of kombucha batches were prepared and incubated under a sterile condition. Healthy batch of kombucha was selected after 30 days of incubation. The broth or kombucha (KB) and the thick floating SCOBY were taken out separately as different samples for further experiments. For obtaining extract, SCOBY was squeezed hard to separate SCOBY extract (SE) and the dry and hard remaining cellulose was discarded (Fig. 1 a, b, c).

Tests were performed with both KB (Kombucha broth) and SE (SCOBY extract) to detect alkaloids and caffeine. Alkaloids were tested by Mayer’s, Wagner’s and Marme’s tests (Santra et al. 2006) where formation of precipitate confirms the presence of alkaloids. Mayer’s reagent is used to detect alkaloids in natural products. Mayer’s reagent was freshly prepared by dissolving mercuric chloride (1.36 g) and of potassium iodide (5 g) in 100 ml of distilled water. Most alkaloids are precipitated by Mayer’s reagent and give a cream coloured precipitate. For Wagner’s test, 1 ml of diluted HCl and Wagner’s reagent (solution of iodine in potassium iodide) were added to 2 ml of sample and mixed. Formation of reddishbrown precipitate showed the presence of alkaloids. In 3 ml of each sample, a few drops of Marme’s reagent (20 g of KI in 20 ml of distilled water; 10 g of CdCl₂ in 50 ml distilled water) were added and diluted H₂SO₄ was then added to the mixture. Yellowish white precipitate shows presence of alkaloids. Caffeine detection was carried on following murexide test (Santra et al. 2006) where appearance of purple colour confirms the presence of caffeine.

Following the technique of column chromatography (Bhattcharya et al. 2009), we separated caffeine and related alkaloids from both KB and SE through a silica gel (200-400 mesh size) packed column. As analyte, 10 ml of each KB and SE were used and for mobile phase, two solvents (chloroform and dicholoromethane) were selected on the basis of affinity of targeted compounds towards solvents. Each fraction was collected individually, dried and tested for caffeine confirmatory test (the murexide test).

KB and SE (1 ml of each) were taken in individual test tubes and left for complete air drying. The dried samples were finally dissolved in 1 ml of ethanol prior to GC-MS analysis. The analysis was performed by GCMS-QP2010 Plus (Shimadzu Corporation, Kyoto Japan) attached with a Rxi-5 fused-silica capillary column (0.25 µm film thickness, 0.25 mm internal diameter and 30 m of length) following the protocol of Majumder et al. (2020).

The results of preliminary tests (alkaloid and caffeine detection tests) show the absence of any alkaloid in beverage or KB. Moreover, the murexide test for caffeine detection gave a positive result for SE only. These results clearly showed that SE contains the alkaloids of tea and one of the major components was caffeine unlike the komucha (KB). Furthermore, it confirmed the earlier reports of Greenwalt et al. (2000) and Malbaša et al. (2006) as they stated that caffeine and other alkaloids of tea get decreased during kombucha fermentation.

In column chromatographic separation of caffeine and related components, chloroform and dicholoromethane were used as mobile phase because our targeted compounds are already reported to be highly soluble in those two organic solvents (JaberVazdekis et al. 2006; Shalmashi and Golmohammad, 2010). Presence of caffeine was simply recognized in the fractions which showed white crystalline structures. Furthermore, the murexide test was again assessed to confirm presence of caffeine and related compounds and to check whether KB contains any traces of such compound or not. But, the murexide test also resulted negative for KB.

Meanwhile, results of the above experiments confirmed that sample SE contains caffeine and related alkaloids in it unlike the beverage. These result have not only proven that SCOBY needs caffeine and other stimulants for growing and stimulating the fermentation process but also suggested that alkaloids or stimulants (like caffeine), a major part of tea may get removed from the infusion as SCOBY picks all of them during kombucha fermentation. Further, to evaluate and judge the results of these tests, GC-MS analysis was performed as described earlier.

GC-MS chromatograms are given in Fig. 2 and peak reports of both KB and SE are represented in Table 1 and Table 2 respectively.

Expectedly, SCOBY extract (SE) showed richness in caffeine content with an amount of 8.70% while, in the beverage or KB, there was no traces of caffeine and related compounds. Moreover, besides caffeine, another alkaloid and major purine nucleotide of tea, guanosine (as guanosine hydrate) was detected as the major compound of SE with a peak area of 12.01%. Interestingly, the compound guanosine is also known to be present in tea leaves and that too as a precursor of caffeine (Negishi et al. 1992). Now, if we accept the fact that SCOBY has captured all the stimulants from broth, then presence of another SE component; e.g. thymine (4.08%), a pyrimidine plant stimulant (Šormová et al. 1960), should also be taken into consideration along with caffeine and guanosine. However, in KB, there were neither any traces of above mentioned compounds nor any other stimulants. So, this is definitely a remarkable confirmation to our interpretation that microflora of SCOBY biofilm has picked up all the tea stimulants from kombucha which have been detected by GC-MS (Fig. 3). Besides alkaloids, presence of an infamous flaw compound of wines, isovalericacid (5.45%); a toxic compound, 1,3-Dioxolane, 4-methylene- or DABCO (1.96%) and a susceptible oxidative agent, 1,3-dioxolane, 4-methylene (2.09%) in SE is also a good sign as these substandard and unacceptable components were completely absent in KB or might have been removed by the biofilm in the same way like the stimulants. However, isovaleric acid is also reported as a metabolite of Brettanomyces sp. (Licker et al. 1998), a common yeast found in SCOBY (Greenwalt et al. 2000).

The GC-MS analysis and other biochemical tests have helped to discover that both the beverage and the biofilm extract, may be a part of the same broth, but, are distinct to each other regarding their chemical characters. Moreover, the chemical profile has also been unfolded for SCOBY extract through this set of experiments. Acetobacter xylinum has been previously reported to produce the cellulose network stimulated by caffeine and other stimulants (Fontana et al. 1991; Greenwalt et al. 2000). But, through a biochemical analysis, it is very hard to identify responsible and specific organism(s) involved in picking up stimulants like caffeine and its precursor guanosine. Moreover, caffeine is also known to exhibit antimicrobial activity. So, chances may be less for other microorganisms to capture this big amount of caffeine from tea infusion. Furthermore, microbiological and metabolic pathway studies on SCOBY in future are needed to reveal the same.