Degradation of the low-calorie sugar substitute 5-ketofructose by different bacteria

Abstract There is an increasing public awareness about the danger of dietary sugars with respect to their caloric contribution to the diet and the rise of overweight throughout the world. Therefore, low-calorie sugar substitutes are of high interest to replace sugar in foods and beverages. A promising alternative to natural sugars and artificial sweeteners is the fructose derivative 5-keto-D-fructose (5-KF), which is produced by several Gluconobacter species. A prerequisite before 5-KF can be used as a sweetener is to test whether the compound is degradable by microorganisms and whether it is metabolized by the human microbiota. We identified different environmental bacteria (Tatumella morbirosei, Gluconobacter japonicus LMG 26773, Gluconobacter japonicus LMG 1281, and Clostridium pasteurianum) that were able to grow with 5-KF as a substrate. Furthermore, Gluconobacter oxydans 621H could use 5-KF as a carbon and energy source in the stationary growth phase. The enzymes involved in the utilization of 5-KF were heterologously overproduced in Escherichia coli, purified and characterized. The enzymes were referred to as 5-KF reductases and belong to three unrelated enzymatic classes with highly different amino acid sequences, activities, and structural properties. Furthermore, we could show that 15 members of the most common and abundant intestinal bacteria cannot degrade 5-KF, indicating that this sugar derivative is not a suitable growth substrate for prokaryotes in the human intestine. Key points • Some environmental bacteria are able to use 5-KF as an energy and carbon source. • Four 5-KF reductases were identified, belonging to three different protein families. • Many gut bacteria cannot degrade 5-KF. Supplementary Information The online version contains supplementary material available at 10.1007/s00253-021-11168-3.


Introduction
Several lines of evidence indicate that high sugar intake is a risk factor for the development of obesity, type 2 diabetes, and cardiovascular diseases (Malik et al. 2010;Schulze 2004). In addition, there is an increasing public awareness about the risk of dietary sugars (glucose, sucrose, and fructose) with respect to their caloric contribution to the diet and the rise of overweight of children and adults throughout the world (Schulze 2004;Ludwig et al. 2001). Therefore, low-calorie sugar substitutes are of high interest to replace sugar in foods and beverages. Based on the tendency of increased human sugar intake and the associated negative effects on human metabolism, the demand for alternative sweeteners has increased enormously. Especially naturally occurring low-calorie sweeteners have become the focus of the food industry. Among the already established synthetic sweeteners are saccharin, acesulfame-K, sucralose, aspartame, and neotame (Mattes and Popkin 2009). Disadvantages of these synthetically produced sweeteners can be, for example, expensive synthesis processes, artificial tastes, or undesirable side effects for consumers (Schiffman et al. 1979;Wiet and Beyts 1992). Two Jacqueline Schiessl and Konrad Kosciow contributed equally to this work.
other non-artificial sugar substitutes used in the food industry are the ketohexose tagatose and the sugar alcohol xylitol. Although both compounds have a high degree of sweetening, they can be partially metabolized by the human metabolism and therefore have a caloric value (Grembecka 2015;Vastenavond et al. 2011). Alternative sweeteners of natural origin are steviol glycosides, which have 30 to 150 times the sweetness of sugar. However, the compounds may have a bitter or licorice-like aftertaste (Soejarto et al. 1982). A promising alternative to natural sugars and artificial sweeteners is the bacterially produced fructose derivative 5-keto-D-fructose (D-threohexo-2,5-diulose, 5-KF). 5-KF is a natural sugar produced by acetic acid bacteria through oxidation of fructose by the membrane-bound enzyme fructose dehydrogenase (Ameyama et al. 1981;Kawai et al. 2013). The compound has the same natural taste as fructose without an artificial aftertaste and a sweetness comparable to sucrose (Herweg et al. 2018). In addition, 5-KF cannot or can only partially be metabolized by the human organism and thus has a very low-calorie content (Wyrobnik et al. 2009). The new potential sweetener has already been detected in various natural foods, including white wine, honey, and elderflower syrup (Burroughs and Sparks 1973;Blasi et al. 2008).
5-KF is produced by the oxidation of fructose by some representatives of acetic acid bacteria of the genus Gluconobacter. Some of the organisms, such as Gluconobacter (G.) japonicus, possess the enzyme fructose dehydrogenase, which can oxidize fructose to 5-KF (Ameyama et al. 1981;Yamada et al. 1966). The genes coding for fructose dehydrogenase were introduced into G. oxydans 621H ) and the resulting genetically modified strain was capable of producing 5-KF in a fed-batch fermentation process resulting in product titers of up to 490 g L −1 and product yields up to 98% (Herweg et al. 2018). Furthermore, a G. oxydans strain was developed for the efficient production of 5-KF from the cost-efficient and renewable feedstock sucrose (Hoffmann et al. 2020). But before 5-KF can be used as a sweetener, it should be tested whether the compound is degradable by microorganisms and whether it is metabolized by the human microbiota. Therefore, we tested several bacteria for their ability to degrade 5-KF and we analyzed the enzymes involved in this process.

HPLC analysis
For the analysis of substrate consumption and product formation analysis, 1-ml samples were taken at different time points and centrifuged at 13,000×g for 1 min. The supernatants were diluted 1:10 with H 2 O. HPLC analysis (Knauer Smartline HPLC system, Knauer GmbH, Berlin, Germany) was performed using an Aminex HPX-87H 300 mm × 7.8 mm column (Biorad, Munich, Germany) with 5 mM H 2 SO 4 at 65°C and a flow rate of 0.6 ml min -1 . Substrates and products were quantified by a refraction index detector (RI detector; Azura RID2.1 L, Knauer GmbH, Berlin, Germany) and a UV detector (Smartline 2600, Knauer, Berlin, Germany) at 210 nm by comparison to calibration curves. For product analysis of 5-KF reduction enzyme assays samples were analyzed by a SpectraSYSTEM HPLC system (Thermo Fisher Scientific Inc., Waltham, USA) using an amino phase column Eurospher II 100 NH2 (250 × 3.0 mm; 5 μm particle size) (Knauer GmbH, Berlin, Germany) with integrated precolumn, 90% acetonitrile as solvent, at 40°C and a flow rate of 0.6 ml min -1 . Products were quantified by a refraction index (RI) detector (Shodex RI-101) (Showa Denko Europe GmbH, Munich, Germany) and evaluated by the external standard method with ChromQuest 5.0 (Thermo Fisher Scientific Inc.).

Construction of expression systems for kfr and akr
The gene kfr from T. morbirosei (HA49_09215) and sdh from C. pasteurianum (CPAST_c38270), both encoding a putative s h i k i m a t e d e h y d r o g e n a s e , a n d t h e g e n e a k r (CPAST_c22030), encoding an aldo/keto-reductase (AKR) from C. pasteurianum, were amplified by PCR using Q5 High-fidelity DNA polymerase and the primers KFR_pASK5_for/ KFR_pASK5_rev, SDH_Cpast_for/ SDH_Cpast_rev, and AKR_Cpast_for/ AKR_Cpast_rev, respectively ( Table 1). As template genomic DNA of T. morbirosei DSM 23827 and genomic DNA of C. pasteurianum DSM 525 was used. The genes were cloned into a pASK-IBA.5 vector. The gene kfr from T. morbirosei Overexpression and purification of the proteins TM-KFR, CP-AKR, GOX0644, and GOX1432 For protein production, overnight cultures of E. coli (5 mL) harboring plasmids of interest were used to inoculate 1 L LB medium and were incubated at 37°C and 180 rpm in shaker flasks. After reaching an OD 600 of 0.4, protein production was induced by the addition of 0.2 μg mL -1 anhydrotetracycline and cells were further cultivated for 4 h. Cultures were harvested at OD 600 between 1.0 and 1.5 by centrifugation (9000×g, 4°C, 15 min). Lysis and purification were performed by sonication as previously described (Kosciow et al. 2016). For protein visualization, polyacrylamide gel electrophoresis was done according to (Laemmli 1970) and protein bands were detected via silver stain as described by (Blum et al. 1987). Analysis of the native conformation of the ketofructose reductases was performed by gel filtration chromatography using a HiLoad 16/60 Superdex 200 pg column (GE Healthcare, Chicago, USA) connected to an ÄKTApurifier system (GE Healthcare, Chicago, USA). The column was calibrated using the Gel Filtration Calibration Kit HMW (GE Healthcare, Chicago, USA). Equilibration was done with 50 mM Tris-HCl buffer pH 8, containing 150 mM NaCl.

Measurement of enzyme activities and kinetic parameters
The reduction of 5-KF and other substrates accompanied by the oxidation of NAD(P)H to NAD(P) + was recorded at 340 nm (ɛ=6.22 mM −1 cm −1 ). The reaction mixture with a final volume of 1 ml contained 250 μM NAD(P)H, 40 mM potassium phosphate buffer (pH 7), and varying substrate concentrations between 5 and 20 mM. One unit of enzyme activity corresponded to the oxidation of 1.0 μmol of substrate per min and mg protein. The pH optima were determined using the McIlvaine buffer system (McIlvaine 1921) between pH 3.0 and 9.0 containing disodium phosphate and citric acid. Temperature optima were analyzed in a temperature range between 20°C and 80°C using standard assay conditions. Nonlinear regressions of Michaelis-Menten data were used to calculate kinetic constants at optimal pH and temperate using 5-KF concentrations between 0.5 and 80 mM and NAD(P)H concentrations between 5 μM and 250 μM, respectively.
Enzyme assays analyzed via HPLC were performed with the same principle as described above. The enzyme assays had a final volume of 250 μL with 5 mM NADPH, 5 mM 5-KF, 40 mM potassium phosphate buffer pH 7 or pH 6 for CP-AKR and 2 μg enzyme. The reaction mixture was incubated at 30°C or 40°C for CP-AKR for three hours. The enzyme reaction assays were directly analyzed via HPLC, as described above.

Bioinformatic tools
The program Blastp at the NCBI database (https://blast.ncbi. nlm.nih.gov/Blast.cgi) was used to identify proteins and to compare amino acid sequences (Johnson et al. 2008). The term similarity is defined as percentage of the number of amino acids that were either identical between the query and the subject sequence or had similar chemical properties. Amino acid sequence alignments were performed with the program Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/) using default parameters. Kinetic parameters of purified enzymes were determined using nonlinear regression of the Michaelis-Menten data with the program GraphPad Prism Version 9.0.0

5-KF consumption in aerobic bacteria
It was already shown that T. citrea (former Erwinia citreus) is capable of reducing 5-KF. The organism possesses a 5-KF reductase, which catalyzes the reversible NADPH-dependent reduction of 5-KF to D-fructose (Schrimsher et al. 1988). We tested a close relative, T. morbirosei, for its ability to grow on 5-KF as a carbon and energy source. A complex medium was used for the growth experiments since no mineral medium is available for T. morbirosei and the growth requirements of this organism are unknown. In the control experiment without the addition of a sugar source, the cells grew to an optical density (OD 600 ) of 1.25 within 20 h (Fig. 1). The doubling time (t d ) was 2.3 h. Cultures with 5-KF revealed a similar growth pattern at the beginning of the experiment (t d = 1.8 h) with a short slowdown of growth after 16 h. In contrast to the negative control, a second growth was observed and the OD 600 increased to 2.1 (Fig. 1), indicating a typical pattern of diauxic growth. In line with this observation was the fact that the organism consumed significant amounts of 5KF only after 16 h. 5-KF was completely utilized after 25 h and cells turned into the stationary growth phase. It can therefore be assumed that components of the complex medium were first used for growth, as in the control without 5-KF. Only in the second phase, 5-KF was utilized, which was accompanied by an increased OD 600 . T. morbirosei was therefore able to use 5-KF as a substrate.
Some species of the genus Gluconobacter are known to produce 5-KF from fructose. The membrane-bound fructose dehydrogenase (Fdh), which is encoded by the fdhSCL genes, plays a crucial role in this process (Kawai et al. 2013). The enzyme oxidizes fructose at the hydroxyl group at position 5, thus forming 5-KF and transferring the released electrons into the respiratory chain of the organism. The resulting 5-KF diffuses via porins in the outer membrane into the extracellular space and accumulates in the culture supernatant. Hence, the question arose whether 5-KF can also be a substrate for growth. For this purpose, the Fdh-producing strain G. japonicus LMG 26773 was tested, which is able to form 5-KF. The organism was grown on a medium containing 0.6% yeast extract and 25 mM 5-KF (Fig. 2). It was found that G. japonicus LMG 26773 consumed large quantities of 5-KF after a lag phase of about 3 h and almost completely converted the substrate after 12 h. The OD 600 increased from 0.05 to 2.2 within 9 h and the doubling time was 1.3 h. The control culture without 5-KF addition showed only an increase in OD 600 from 0.05 to 0.2 (Fig. 2).
The model organism G. oxydans 621H does not possess the key enzyme Fdh for 5-KF production . Therefore, we tested a fdh knock-in mutant of G. oxydans (Hoffmann et al. 2020) for the expression of the fdh genes that allowed the production of the fructose dehydrogenase. Two other strains G. japonicus LMG 1281 and G. japonicus LMG 26773 naturally contained the fdh genes on their chromosomes (gene no. BAM93250 -BAM93252 and KXV40773 -KXV40775, respectively). As expected, all strains grew with fructose as substrate and accumulated 5-KF in the culture supernatant up to a concentration of 50 mM (Hoffmann et al. 2020). In the stationary phase, G. japonicus LMG 1281 and G. japonicus LMG 26773 completely consumed 5-KF within the period 40-98 h and 40-70 h, respectively (Fig. 3). In G. oxydans fdh the degradation of 5-KF was slow and only 10 mM 5-KF were utilized within 60 h.

5-KF consumption in anaerobic bacteria
All 5-KF degrading species described above were strict aerobes and needed O 2 as the final acceptor of their respiratory chain. Hence, the question arose whether only aerobic bacteria are able to ferment 5-KF or whether this ability is also observed in anaerobic bacteria. As a model organism we tested C. pasteurianum DSM 525 for its ability to use 5-KF as a substrate. The organism reached an OD 600 of 1.7 and a doubling time of 2.0 h with 25 mM glucose as its preferred carbon source. In contrast, C. pasteurianum grew to a final OD 600 of 1.4 within 17.5 h in the presence of 23 mM 5-KF, which corresponded to a growth rate of 0.33 h -1 and a doubling time of 2.1 h in the exponential phase (Fig. 4). HPLC analysis revealed a complete consumption of 5-KF after 17.5 h and the synthesis of the expected end products butyrate and acetate, indicating that this compound was metabolized by C. pasteurianum.
Without an additional carbon source, C. pasteurianum reached only a final OD 600 of 0.15 (Fig. 4).
Since we could show that the anaerobic bacterium C. pasteurianum is able to metabolize 5-KF, the next step was the investigation of whether the main players of the anaerobic human gut microbiota also have the ability to use 5-KF as a carbon and energy source. For consumers and the food industry, these experiments are important as they allow the determination of the effective caloric value of novel and potential sweeteners. Thus, 15 microorganisms were tested as representatives of important members of the human gut microbiota: . S1). However, it became evident that the tested organisms did not reach a higher OD 600 with 20 mM 5-KF as substrate when compared to the negative control without an additional carbon source (Fig. S5). Furthermore, HPLC analysis confirmed that 5-KF was not degraded by these bacteria. The results indicated that many important microbial species within the human gut are likely not able to metabolize 5-KF and to produce short-chain fatty acids, which could contribute to the total calorie intake of the human body.

Characterization of 5-KF reductases
As shown above, a couple of bacteria can grow with 5-KF as carbon and energy source. Others were able to use the compound in the stationary phase to ensure survival in their substrate-limited habitat. However, the knowledge about biochemical reactions that are involved in 5-KF degradation is limited. It is known that T. citrea contains a 5-KF reductase that can reduce 5-KF using NADPH as reductant. The corresponding enzyme was purified and characterized (Schrimsher et al. 1988). The N-terminal AS sequence was published and a BLASTx search (https://blast.ncbi.nlm.nih.gov/Blast.cgi) resulted in the identification of protein from T. morbirosei with 94% similarity. Hence, we could identify the corresponding gene (HA49_09215) and amino acid sequence (KGD75378.1). The corresponding DNA fragment was amplified by PCR and the gene was cloned into plasmid pASK-IBA.5, which allowed the precise fusion with the vector encoded Strep-tag II sequence. The protein was produced in E. coli and purified to apparent homogeneity with the help of a streptavidin affinity column (Fig. 5). The protein was referred to as TM-KFR.
As demonstrated by SDS-PAGE (Fig. 5), TM-KFR showed the expected protein band with a molecular mass of 38 kDa. The native protein revealed a single active peak when analyzed by gel filtration. The peak corresponded to 121.3 kDa, indicating that the enzyme is active as a trimer ( Table 2). The apparent K M value of the enzyme for 5-KF, estimated by using several concentrations of the substrate from 0.5 to 40 mM, was 6.4 ± 0.6 mM at pH 7.0 and 30°C (Table 2), while V max was 1522.8 ± 141.9 U mg -1 . The apparent K M value for NADPH was 20.4 ± 11.5 μM. For a detailed analysis of enzymatic activities, we first determined the substrate spectrum of the protein. It became evident that the enzyme only reduced 5-KF and could only utilize NADPH as cofactor. All other compounds tested (aldehydes, ketones, sugars, keto sugars, sugar acids, polyols, Table 3) were not attacked. Hence, the specificity towards 5-KF can be used to determine the presence and concentration of the alternative sweetener in various foods and beverages in a photometric assay. Such tests were performed and it was found that 5-KF is present in different white wines, honey, and vinegar in micromolar concentrations (not shown).
As shown above, G. oxydans was not able to use 5-KF as carbon and energy source in the exponential growth phase, but consumption of 5-KF was observed in the stationary phase (Fig. 3). To verify which enzymes are responsible for this effect, we tested several oxidoreductases that act on organic compounds containing keto groups for their ability to reduce 5-KF to fructose. All enzymes indicated in Fig. 6 were heterologously produced in E. coli and purified by Strep-tag affinity chromatography. It became evident that the 5- ketogluconate reducing enzyme GOX1432 (Zahid and Deppenmeier 2016) and the α-ketocarbonyl reductase GOX0644 ) were able to reduce 5-KF with NADPH as reductant (Fig. 6). The α-diketone reductase GOX0646 (Schweiger et al. 2013) reduced 5-KF at a very low rate. In contrast, the 2-ketogluconate reductase GOX0417 (Rauch et al. 2010) and the vinyl ketone reductases GOX0502 and GOX2684 (Schweiger et al. 2008) did not reduce 5-KF. The same was true for the α-ketocarbonyl reductases GOX1615 ) and GOX0313 (Schweiger et al. 2013) (Fig. 6).
SDS-PAGE analysis showed molecular masses of 34 kDa and 55 kDa for the proteins GOX0644 and GOX1432, respectively (Fig. 5), which were in good correlation with the estimated masses predicted from their amino acid sequences. The native conformation of both enzymes was analyzed via gel filtration, indicating a monomeric structure for protein GOX1432 and a dimeric conformation for protein GOX0644. Kinetic properties of the enzymes with 5-KF as substrate were examined with nonlinear regressions of Michaelis-Menten data. The K M and V max values of protein GOX0644 for 5-KF as substrate were 4.0 ± 0.8 mM and 32.7 ± 2.1 U mg -1 (Table 2). In case of protein GOX1432, a V max value of 230.3 ± 8.0 U mg -1 was calculated, which indicates a higher specific activity for 5-KF when compared to GOX0644. In contrast, protein GOX1432 showed a significantly lower affinity for this substrate (K M value: 42.0 ± 5.4 mM; Table 2). With NADPH as co-substrate of the reaction, K M values of 61.2 ± 11.4 μM and 19.5 ± 1.0 μM were determined, respectively. The substrate spectra of the enzymes were already published. Protein GOX0644 reduced αketoaldehydes, α-diketones, α-keto esters (Table 3), and the   1 Enzyme assays were performed at 30°C in 1 ml assays containing 40 mM potassium phosphate buffer pH 7 or pH 6 for CP-AKR, 250 μM NADPH and 20 mM substrate as indicated. None of the enzymes could reduce Darabinose, L-arabinose, glucose, mannose, ribose, tagatose, xylose, maltose, raffinose, sucrose, trehalose, and xylobiose sugar derivative 2,5-diketogluconate , which shows structural homologies to the newly discovered substrate 5-KF. Protein GOX1432 was previously described as mannitol dehydrogenase and reduced the sugars D-fructose and L-sorbose as well as the keto sugar 5-keto-D-gluconate (Table 3) (Zahid and Deppenmeier 2016). As demonstrated above, C. pasteurianum was able to grow on 5-KF. Since the three proteins KFR from T. morbirosei, GOX0644 and GOX1432 from G. oxydans, were characterized as 5-KF reducing enzymes, Blastp analyses (https://blast. ncbi.nlm.nih.gov/Blast.cgi) were performed to identify corresponding enzymes in C. pasteurianum. Two proteins encoding a putative shikimate dehydrogenase (AJA49865.1) and a putative aldo/keto reductase (AJA48273.1) were similar to the KFR from T. morbirosei and protein GOX0644, respectively (similarities of 48% and 67%). Both enzymes were produced in E. coli and purified via streptavidin affinity chromatography. While the putative shikimate dehydrogenase AJA49865.1 was inactive with 5-KF as substrate, high activity was detected for the aldo/keto reductase AJA48273.1 (referred to as CP-AKR). When analyzed by polyacrylamide gel electrophoresis and silver stain, a single band for CP-AKR could be detected that was in good accordance with the expected size of 35 kDa (Fig. 5). Gel filtration analysis revealed that the native enzyme formed homotetramers. In addition, the kinetic parameters were investigated, indicating a V max value for 5-KF of 48.9 ± 7.1 U mg -1 and a low K M value of 1.3 ± 0.2 mM. The K M for NADPH was 12.4 ± 3.8 μM. The CP-AKR was able to reduce aldehydes as well as different ketones and, additionally, the reduction of 5-ketogluconate and L-sorboson could be shown (Table 3).
Although all described enzymes exhibited 5-KF reductase activity, significant differences were determined concerning the substrate spectrum, the corresponding amino acid sequence, and the tertiary and quaternary structure. To test whether also the reaction mechanism is different, an HPLC analysis of the end products of all four enzymes after 5-KF reduction was performed (Fig. 7). D-fructose and L-sorbose were applied as analytical standards (Fig. 7a). It became evident that protein GOX0644 and its homologous protein CP-AKR produced L-sorbose as main product of 5-KF reduction (Fig. 7 c and d). In contrast, protein GOX1432 and TM-KFR from T. morbirosei, which did not show sequence similarities, reduced 5-KF to D-fructose ( Fig. 7b and e). A closer look at the structure of 5-KF indicates that both prochiral keto groups are homotopic and have therefore the same chemical reactivity. Each sp 2 -hydridized trigonal planar carbonyl C-atom can be reduced by the enzymes from the re or the si face. Hence, the stereoselectivity of the 5-KF reducing enzymes with respect to the hydrid transfer from dehydronicotinamide to the substrate is responsible for the formation of D-fructose or Lsorbose.
Involvement of the purified enzymes in the metabolism of 5-KF degrading bacteria C. pasteurianum was able to grow with 5-KF (Fig. 4). F u r t h e r m o r e , t h e a l d o / k e t o r e d u c t a s e C P -A K R (AJA48273.1) from C. pasteurianum showed high 5-KF reducing activity. qRT-PCR experiments revealed that the mRNA abundance of the corresponding gene was increased about 10-fold during growth with 5-KF (Fig. 8), indicating that CP-AKR had an important function in the degradation of this substrate and might be responsible for the formation of fructose, which is then channelled into the central metabolism of C. pasteurianum. G. oxydans did not use 5-KF as substrate for growth. Therefore, it was not possible to test whether the proteins GOX0644 and GOX1432 were Fig. 6 Analysis of oxidoreductases from G. oxydans. Activity with known substrates (white columns), activity with 5-KF (black columns). Assay consisted of 40 mM potassium phosphate buffer pH 7, 250 μM NADPH, 5 mM substrate and 2 μg enzyme. The assays were incubated at 30°C responsible for 5-KF reduction in vivo. However, in the 5-KF degrading bacterium G. japonicus LMG 26773, proteins with high similarity to the 5-KF reducing enzymes GOX0644 and GOX1432 were identified (Fig. S6). These enzymes were referred to as GJA0644 and GJA1432. The corresponding genes were analyzed by qRT-PCR and it was found that the mRNA concentrations were about 4-fold higher in 5-KF grown cells compared to cells, which used fructose as substrate (Fig. 8). The results indicate that the proteins GJA0644 and GJA1432 as counterparts of GOX0644 and GOX1432 are most probably involved in the first step of 5-KF degradation as performed by G. japonicus LMG 26773. In contrast to C. pasteurianum and G. japonicus LMG 26773, a change in the expression pattern of the gene encoding the 5-KF reductase TM-KFR was not observed in T. morbirosei (Fig. 8). However, for T. citrea, an almost identical 5-KF reductase has already been described and the enzyme was directly connected with the conversion of 5-KF into fructose as substrate for growth (Schrimsher et al. 1988).

Discussion
In recent years, it has been demonstrated that conventional sugars such as sucrose, glucose, and fructose can promote the development of food-related diseases (Malik et al. 2010;Schulze 2004;Ludwig et al. 2001;Malik et al. 2006;Johnson et al. 2007) such as type 2 diabetes, obesity, and cardiovascular diseases. It can also be observed that consumers are increasingly paying attention to a more conscious diet leading to an increased demand for calorie-free or lowcalorie substitutes as an alternative to traditional sweeteners (Marti et al. 2008;Ogden et al. 2006). A promising alternative to natural sugars and artificial sweeteners is the bacterially produced fructose derivative 5-KF. Tests with trained taste specialists have shown that the sweetening power and taste are equivalent to the properties of pure fructose. Thus, 5-KF has a stronger sweetening power than sucrose (Herweg et al. 2018;Stone and Oliver 1969). The compound revealed a similar intrinsic sweet threshold concentration in comparison to fructose. Moreover, there is no artificial, bitter or licorice-like aftertaste, which is noticeable when using some other alternative sweeteners such as stevia, saccharin, and acesulfame K (Soejarto et al. 1982;Herweg et al. 2018;Kuhn et al. 2004). In addition, 5-KF is a natural sugar derivative that is found, e.g., in honey, white wine, and vinegar , leading to the question of how 5-KF is degraded in nature.  depending on the stereoselectivity of the enzymes (Fig. 7). Our hypothesis is that the reduction of the si face of the planar C-carbonyl structure by the enzymes TM-KFR and GOX1432 leads to the formation of D-fructose. In the other case, the keto group is reduced at the re face by the 5-KF reducing enzymes GOX0644 and CP-AKR resulting in the production of Lsorbose.
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s00253-021-11168-3. of Microbiology and Biotechnology, University of Bonn, Germany) for technical assistance and Prof. Sigurd Höger (Institute of Organic Chemistry, University of Bonn, Germany) for support to analyze the stereospecificity of the 5-KF reducing enzymes.
Author contribution UD designed the study and wrote the paper. JS designed and performed most the experiments. JJH, LSG, and JH conducted same experiments. KK designed and performed experiments and contributed with TF to the study's conception, data analysis, and interpretation of the results. All authors analyzed the results and approved the final version of the manuscript.
Funding The study was supported by funding of the BMBF project IMPRES (FKZ 031B0370A). The funding agency was not involved in the research.
Data availability Data are available upon request from the authors.
Code availability Not applicable.

Declarations
Ethics approval This article does not contain any studies with human participants or animals performed by any of the authors.
Consent to participate Not applicable.
Consent for publication Not applicable.

Conflict of interest
The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.