Enzymatic removal of a cephalosporin methyl ester blocking group

Over 7000 microorganisms were screened to find an enzyme source for the hydrolysis of a C₄ methyl ester blocking group on 7-aminodesacetoxycephalosporanic acid (7-ADCA). Only one culture, Streptomyces capillispira Mertz and Higgens nov. sp., produced an enzyme that catalysed the reaction. Enzyme synthesis in a defined mineral salts medium was repressed by NH₃ and amino acids. Under optimum fermentation conditions, the maximum rate of substrate hydrolysis was 6 × 10^?10 mol min^?1 mg^?1 cell. The enzyme was recovered from the mycelia and partially purified by gel filtration. Kinetic studies by pH-stat titration indicated that the pH optimum was 7.5–8.5, the temperature optimum was 25–30°C, and the substrate K_m value was 2.3 mg ml^?1. The reaction products, 7-ADCA and methanol, were weak competitive inhibitors of the enzyme with K₁ values of 6.63 and 0.188 mg ml^?1, respectively. The enzyme also hydrolysed cefaclor and cephalexin methyl esters but did not hydrolyse cephalosporin ethyl esters. With further improvements in enzyme yields and stability, enzymatic deblocking of cephalosporins could provide an alternative to chemical deblocking processes.     

The synthesis of 4-methylumbelliferyl alpha-ketoside of N-acetylneuraminic acid and its use in a fluorometric assay for neuraminidase

4-Methylumbelliferyl α-ketoside of N-acetylneuraminic acid was synthesized by reacting the sodium salt of 4-methylumbelliferone with the 2-chloro-2-deoxy derivative of peracetylated methyl N-acetylneuraminate, followed by preparative silica gel chromatography, deblocking, and purification by gel filtration on Sephadex G-25. The final product was isolated as either the sodium or ammonium salt, and its suitability as a substrate for neuraminidase was evaluated. The optimal pH values for various neuraminidases were 5.6 in acetate buffer (Arthrobacter ureafaciens), 5.0–5.1 in acetate buffer (Clostridium perfringens), and 4.4 in phosphate-citrate buffer (human fibroblasts). K_m values for these enzymes at the optimal pH were 6 × 10^?4m (Arthrobacter), 1 × 10^?4m (Clostridium), and 3 × 10^?4m (human fibroblasts).     

Activity of imidazole on the hydrolysis of cyclic AMP and cyclic GMP by bovine heart and rat liver cyclic nucleotide phosphodiesterases.

The effects of imidazole on the hydrolysis of cyclic AMP and cyclic GMP by crude and partially purified phosphodiesterases obtained from bovine heart and rat liver were studied in order to determine if imidazole has an activity on cyclic nucleotide hydrolysis under conditions which might explain its ability to antagonize the effects of several hormones. Imidazole-Cl (40 mm, pH 7.4) had no effect on the hydrolysis of cyclic AMP or cyclic GMP at substrate levels below 10 μm by the crude enzymes but increasing stimulation was observed with increasing substrate concentrations reaching a twofold stimulation at 1 mm cyclic nucleotide. Three phosphodiesterases with varying substrate specificities were partially purified from bovine heart by ammonium sulfate precipitation and diethyl aminoethyl cellulose chromatography. With these enzymes imidazole had less stimulatory activity and some inhibitory effect on the hydrolysis of 10^?4m cyclic AMP and cyclic GMP but was without significant effect on the hydrolysis of 10^?6m cyclic AMP or cyclic GMP. The stimulatory activity of imidazole on the hydrolysis of high levels of cyclic nucleotide was dependent on the presence of phosphodiesterase activator. The stimulatory effect of the activator and imidazole plus activator on the hydrolysis of 10^?4m cyclic GMP by the rather cyclic GMP-specific enzyme could be eliminated by the addition of ethylene glycol-bis-(β-aminoethyl ether)N,N′-tetraacetate (EGTA) and restored by Ca²⁺. Imidazole was without effect on the binding of cyclic AMP to a cyclic AMP-dependent protein kinase from bovine heart. The lack of effect of imidazole on the hydrolysis of physiological levels of cyclic AMP or cyclic GMP suggests that the activity of imidazole to antagonize the effects of various hormones is probably not due to a direct action of imidazole on the hydrolysis of cyclic AMP or cyclic GMP.     

The hydrolysis of 3-(2-furylacryloyl)-glycyl-l-leucine amide by thermolysin

The kinetics of the hydrolysis of 3-(2-furylacryloyl)-glycycl-l-leucine amide by thermolysin has been reinvestigated. It was found that the K_m for the enzyme substrate interaction is 2.5 × 10^?3m at pH 7.2. This K_m is an order of magnitude less than what has been previously assumed to be the K_m for the enzyme-substrate interaction. The normally recommended assay has 1–3 × 10^?3m substrate and is based on the assumption that the substrate concentration is much less than the K_m. Our data indicate that this assumption appears to be invalid. The hydrolysis of 3-(2-furylacryloyl)-glycyl-l-leucine amide results in a maximum decrease in absorbance at 322 nm. The change in absorbance is nearly 10-fold greater at 322 nm than the change in absorbance at 345 nm where the hydrolysis has been customarily followed. By following the hydrolysis of the substrate at 10^?4m at 322 nm it is possible to work under conditions where the substrate concentration is much less than the K_m.     

Esteroproteolytic enzymes from the submaxillary gland. Kinetics and other physicochemical properties.

Two esteroproteolytic enzymes (A and D) have been isolated from the mouse submaxillary gland and shown to be pure by ultracentrifugation, immunoelectrophoresis, acrylamide-gel electrophoresis, and amino acid analyses. The enzymes have molecular weights of approximately 30,000 and are structurally and antigenically related. Narrow pH optima between 7.5 and 8.0 are exhibited by both enzymes. The “pK₁'s” are between 6.0 and 6.5 and the “pK₂'s” are near 9.0. A marked preference for arginine-containing esters is shown by both enzymes. The maximum specific activity of enzyme A on p-tosylarginine methyl ester (TAME) at pH 8 was 2500–3000 μm min^?1 mg^?1 and for enzyme D, 400–600 μm min^?1 mg^?1. With TAME as substrate, the K_m for enzyme A was 8 × 10^?4m at 25 °C and 6 × 10^?4m at 37 °C. For D, K_m was 3 × 10^?4 at 25 °C and 2 × 10^?4m at 37 °C.An apparent activation of enzyme D by tosylarginine (TA), a product of TAME hydrolysis, and all α-amino acids examined was due to removal of an inhibitor by chelation. This effect could be duplicated by 8-hydroxyquinoline and diethyldithiocarbamate but not by EDTA. Enzyme A was not affected by these substances to any remarkable extent. Several divalent ions proved to be potent inhibitors of enzyme D. Both enzymes are inactivated by the active site reagents diisopropyl phosphofluoridate and tosyllysine chloromethylketone but much less rapidly than is trypsin. Nitrophenyl-4-guanidionobenzoate reacts with a burst of nitrophenol liberation but with a rapid continuing hydrolysis. One active site per molecule is indicated. Enzyme D is inactivated by urea, reversibly at 10 m and with maximal permanent losses at 6 m. Autolysis of the unfolded form by the native enzyme when they coexist at intermediate urea concentrations appears to occur.Identity of enzyme D and the epithelial growth factor binding protein is demonstrated.     

S-adenosylmethionine: Protein-carboxyl methyltransferase from erythrocyte

Protein methylase II (S-adenosylmethionine:protein—carboxyl methyltrans-ferase), which modifies free carboxyl residues of protein, was purified from both rat and human blood, and properties of the enzymes were studied. The pH optima for the reaction were dependent on the substrate proteins used; pH 7.0 was found with endogenous substrate, 6.1 with plasma, 6.5 with γ-globulin, and 6.0 with fibrinogen. The molecular weight of the enzymes from both rat and human erythrocytes were identical (25,000 daltons) determined by Sephadex G-75 chromatography. Partially purified enzyme from rat erythrocytes showed three peaks on electrofocusing column at pH 4.9, 5.5 and 6.0. The K_m values of the enzymes from rat and human erythrocytes showed 3.1 × 10^?6m and 1.92 × 10^?6m at pH 6.0, 1.96 × 10^?6m and 1.78 × 10^?6m at pH 7.2, respectively, for S-adenosyl-l-methionine. It is also found that S-adenosyl-l-homocysteine is a competitive inhibitor for protein methylase II with K_i value of 1.6 × 10^?6m.     

Kinetics of the enzymatic hydrolysis of cellulose: 1. A mathematical model for a batch reactor process

A mathematical model for enzymatic cellulose hydrolysis, based on experimental kinetics of the process catalysed by a cellulase [see 1,4-(1,3;1,4)-β-d-glucan 4-glucanohydrolase, EC 3.2.1.4] preparation from Trichoderma longibrachiatum has been developed. The model takes into account the composition of the cellulase complex, the structural complexity of cellulose, the inhibition by reaction products, the inactivation of enzymes in the course of the enzymatic hydrolysis and describes the kinetics of d-glucose and cellobiose formation from cellulose. The rate of d-glucose formation decelerated through the hydrolysis due to a change in cellulose reactivity and inhibition by the reaction product, d-glucose. The rate of cellobiose formation decelerated due to inhibition by the product, cellobiose, and inactivation of enzymes adsorbed on the cellulose surface. Inactivation of the cellobiose-producing enzymes as a result of their adsorption was found to be reversible. The model satisfactorily predicts the kinetics of d-glucose and cellobiose accumulation in a batch reactor up to 70–80% substrate conversion on changing substrate concentration from 5 to 100 g l^?1and the concentration of the enzymic preparation from 5 to 60 g l^?1.     

Assay for human erythrocyte pyridoxine kinase

An accurate and rapid method for the assay of pyridoxine kinase in human erythrocytes has been developed. The procedure involves the separation of the radioactive product from the substrate with Dowex 50 resin in a test tube. Using the assay designed, we found that human red blood cells have a pyridoxine kinase activity of 1.381 nmole/min/g of hemoglobin (n = 25, SE = 0.051), and the enzyme has a K_m of approximately 1.72 × 10^?6m for pyridoxine. Pyridoxine phosphate was identified as the main product of the assay reaction catalyzed by human erythrocyte pyridoxine kinase in crude hemolysates.     

Factor X activator from Vipera lebetina snake venom,molecular characterization and substrate specificity

Our studies of the venom from the Levantine viper Vipera lebetina have demonstrated the existence of both coagulants and anticoagulants of the hemostatic system in the same venom. We showed that V. lebetina venom contains factor X activator (VLFXA) and factor V activator, fibrinolytic enzymes. VLFXA was separated by gel filtration on Sephadex G-100 superfine and ion exchange chromatography on CM-cellulose and on TSK-DEAE (for HPLC) columns. VLFXA is a glycoprotein composed of a heavy chain (57.5 kDa) and two light chains (17.4 kDa and 14.5 kDa) linked by disulfide bonds. VLFXA has multiple molecular forms distinguished by their isoelectric points. The differences in their pI values may be caused by dissimilarities in the respective charged carbohydrate content or in the primary sequence of amino acids. We synthesized 6–9 amino acid residues containing peptides according to physiological cleavage regions of human factor X and human factor IX. The peptides (Asn-Asn-Leu-Thr-Arg-Ile-Val-Gly-Gly – factor X fragment, and Asn-Asp-Phe-Thr-Arg-Val-Val-Gly-Gly – factor IX fragment) were used as substrates for direct assay of VLFXA. Cleavage products of peptide hydrolysis and the molecular masses of cleavage products of human factor X were determined by MALDI-TOF MS. The MALDI-TOF MS was highly efficient for the recovery and identification of peptides released by VLFXA hydrolysis. We can conclude that VLFXA cleaves the Arg⁵²-Ile⁵³ bond in the heavy chain of human factor X and the Arg²²⁶-Val²²⁷ bond in human factor IX precursor. VLFXA could not activate prothrombin nor had any effect on fibrinogen, and it had no arginine esterase activity toward benzoylarginine ethyl ester.     

Detection of propeptides of AlpA and AlpB lytic endopeptidases from Lysobacter sp. XL1 by the sandwich enzyme immunoassay based on monoclonal antibodies

Extracellular lytic endopeptidases AlpA and AlpB of the Gram-negative bacterium Lysobacter sp. XL1 have a high degree of homology and are synthesized as preproproteins consisting of a signal peptide, a propeptide, and the mature protein. In the present work, two monoclonal antibodies against the AlpA propeptide (ProA) and eleven antibodies against the AlpB propeptide (ProB) have been obtained. The affinity constants for antibodies to ProA were 2.9 × 10⁹ and 3.5 × 10⁹ M^?1, and those for antibodies to ProB were from 1.5 × 10⁸ to 2.2 × 10⁹ M^?1. The antibodies showed no immune cross-reactivity with each other and with mature forms of the enzymes. On the basis of monoclonal antibodies, a sandwich enzyme immunoassay has been developed, which makes it possible to detect these propeptides in the dissolved native form. The linear range of the detection of ProA was 1.5–100.0 ng/mL with an error of measurement of 6%, and that of the determination of ProA was 0.2–6.25 ng/mL with an error of measurement of 6%. By using the assay, propeptides ProA and ProB were detected in cell lysates of Lysobacter sp. XL1 in an amount of 1.18 ± 0.03 and 0.096 ± 0.002 ng per 1 OD540 of bacterial culture, respectively. The immunochemical assay for the detection of different forms of AlpA and AlpB can be useful in solving the problems associated with their secretion into environment.     

A smiple specific assay for uracil tRNA methylases of enteric bacteria: use of ribothymidine-deficient tRNA

A simple quantitative assay that is about 95% specific for uracil tRNA methylases of E. coli and A. aerogenes has been developed. tRNA was isolated from a strain of E. coli carrying the trm^? mutation. These organisms have a low level of uracil methylase and consequently produce tRNA with a selective deficiency of ribothymidine. This RNA acted as a specific substrate for uracil tRNA methylases, when exposed to cell extracts from E. coli or A. aerogenes containing tRNA-methylating enzymes of multiple specificities. This assay can be used to screen organisms for trm^? mutations and for studies with inhibitors.     

Synthesis,characterization, and reactions of 2′-, 3′-, and 5′-(2-bromoethyl)-AMP: Potential affinity labels for nucleotide binding sites in enzymes

Two new adenosine analogs, 2′-(2-bromoethyl) adenosine monophosphate and 3′-(2-bromoethyl) adenosine monophosphate, were synthesized, purified by semipreparative high-pressure liquid chromatography, and completely characterized. A new synthesis of 5′-(2-bromoethyl) adenosine monophosphate is presented which facilitates the preparation of radioactive reagent with label either in the ethyl group or the purine ring of the nucleotide derivative. The reactive moiety of these derivatives, a bromoalkyl group, has the ability to react with the nucleophilic side chains of several amino acids. The second-order, pH-independent rate constants for reaction with the side chains of the amino acids cysteine, lysine, histidine, and tyrosine were determined as 3×10^?4, 6×10^?6, 3×10^?7, and <1×10^?7 M^?1 sec^?1, respectively. These data could be use in estimating the rate enhancement observed in modification of a protein by these affinity-labeling reagents. 5′-(S-(2-hydroxyethyl)cysteine) adenosine monophosphate, the derivative expected from exhaustive digestion of protein in which a cysteinyl residue is modified by 5′-(2-bromoethyl) adenosine monophosphate, and S-2-hydroxyethyl)cysteine, the derivative anticipated upon acid hydrolysis of such a modified protein, were synthesized, characterized, and their elution positions from an amino acid analyzer determined. These bromoethyl AMP derivatives are potential affinity labels for enzymes that bind 2′-, 3′-, or 5′-nucleotides such as TPN, coenzyme A, or ADP, respectively.     

Binding kinetics of sulfatide with influenza A virus hemagglutinin

Association of a sulfated galactosyl ceramide, sulfatide, with the viral envelope glycoprotein hemagglutinin (HA) delivered to the cell surface is required for influenza A virus (IAV) replication through efficient translocation of the newly synthesized viral nucleoprotein from the nucleus to the cytoplasm. To determine whether the ectodomain of HA can bind to sulfatide, a secreted-type HA (sHA), in which the transmembrane region and cytoplasmic tail were deleted, was generated by using a baculovirus expression system. The receptor binding ability and antigenic structure of sHA were evaluated by a hemagglutination assay, solid-phase binding assay and hemagglutination inhibition assay. sHA showed subtype-specific antigenicity and binding ability to both sulfatide and gangliosides. Kinetics of sHA binding to sulfatide and GD1a was demonstrated by quartz crystal microbalance (QCM) analysis. QCM analysis showed that the sHA bound with the association rate constant (k _on) of 1.41?×?10⁴ M^?1 sec^?1, dissociation rate constant (k _off) of 2.03?×?10^?4 sec^?1 and K _d of 1.44?×?10^?8 M to sulfatide immobilized on a sensor chip. The k _off values of sHA were similar for sulfatide and GD1a, whereas the k _on value of sHA binding to sulfatide was 2.56-times lower than that of sHA binding to GD1a. The results indicate that sulfatide directly binds to the ectodomain of HA with high affinity.     

Chromophoric peptide substrates for activity determination of animal aspartic proteinases in the presence of their zymogens: A novel assay

Solid-phase synthesis was used for the preparation of pyroglutamyl-histidyl-p-nitrophenylalanyl-phenylalanyl-alanyl-leucine amide (I) and glycyl-glycyl-histidyl-p-nitrophenylalanyl-phenylalanyl-alanyl-leucine amide (II), two water-soluble and sensitive chromophoric substrates of chicken pepsin, hog pepsin A, and bovine spleen cathepsin D. The kinetic constants of hydrolysis of the p-nitrophenylalanyl-phenylalanyl bond of the substrates were measured by difference spectrophotometry at 308 nm (Δ? = 860 m^?1 cm^?1) and by ninhydrin colorimetry (substrate I, ?₅₇₀ = 2.31 × 10⁴m^?1 cm^?1). The pH optimum of cleavage is 5 for the pepsins and 3.7 for cathepsin D. Since all three proteinases still have a significant activity at pH 5.5–6 a new, simple assay was designed for submicrogram quantities of pepsins in the presence of pepsinogens without interference of the latter. The method is particularly suitable for the analyses of the zymogen activation mixtures.     

Effect of hydrolytic enzymes and protein-modifying reagents on gonadotropin receptors in bovine corpus luteum cell membranes

Preincubation of membranes with various concentrations of pronase, trypsin, lipase, phospholipase A from Vipera russelli and from Crotalus durissus terrificus, phospholipase C from Bacillus cereus and from Clostridium welchii, acetic anhydride, 2,4-dinitrofluorobenzene and tetranitromethane resulted in a dose-dependent inhibition of ¹²⁵I-labeled human choriogonadotropin binding. At the submaximal concentrations of enzymes and at both submaximal and maximal concentrations of protein-modifying reagents, the losses were always greater with ¹²⁵I-labeled human choriogonadotropin than with ¹²⁵I-labeled human lutropin. The inhibition of binding was a consequence of changes in the membranes rather than changes in the hormone caused by the agents being carried over to the final incubation. Inhibition of binding was non-competitive and irreversible.In untreated membranes, the ¹²⁵I-labeled human choriogonadotropin binding was homogenous (K_d = 1.7 · 10^?10 M; N = 60 fmol/mg protein). Treatment of membranes with various enzymes and protein-modifying reagents except tetranitromethane resulted in heterogeneous binding. The number of available high affinity receptors was greatly reduced in every case. However, the affinity of these sites were either unchangedd (trypsin, lipase, phospholipase A from V. russelli, dinitrofluorobenzene and the tetranitromethane) or decreased (pronase and acetic anhydride). the newly appeared second receptor site had a K_d which varied from 3.2 · 10^?10 to 7.1 · 10^?9d M depending on the agent used, and the receptor numbers were low in all cases except acetic anhydride.Receptor occupancy conferred the receptors with marked protection against various hydrolytic enzymes, dinitrofluorobenzene and tetranitromethane. These data suggest the inhibition of binding by the above agents was primarily a consequence of changes in the receptor molecules themselves.     

A critical reexamination of the continous spectrophotometric assay for adenosine deaminase

The continuous spectrophotometric assay for adenosine deaminase has been reinvestigated, using both adenosine and 9-β-d-arabinofuranosyladenine as substrates. This assay is based on the reported decrease in absorbance at or near 265 nm between the adenine nucleoside substrate and the hypoxanthine nucleoside product. In the substrate concentration range 1,5 – 8.0 × 10^?4m, the progress of the reaction is associated with an anomalous sigmoidal dependence of absorbance on time, and the overall change in absorbance decreases with increasing substrate concentration. Near 8 × 10^?4m substrate, the deamination proceeds with no change in absorbance, while at higher concentrations, small increases in absorbance are observed. These effects, if ignored, generate initial “rate” data exhibiting an apparent substrate inhibition whieh, however, is completely an artifact induced by the spectral anomalies. Over the entire concentration range 5 × 10^?6–1 × 10^?3m, alternative assay methods (e.g., discontinuous detection of the product, ammonia) yeld normal Michaelis-Menten kineties. The anomalous behavior manifested in the continuous spectrophotometric assay is due to large negative deviations from Beer's law. These deviations are observed for all four of the nucleosides tested, viz., adenosine, 9-β-d-arabinofuranosyladenine, inosine, and 9-β-d-arabinofuranosylhypoxanthine. The departure from Beer's law is detectable anywhere in the concentration range 5 × 10^?6–1 × 10^?3m, but is most marked at concentrations above 1 × 10^?4m. Thus, the continuous spectrophotometric assay for adenosine deaminase should be utilized withextreme caution, and should not be employed at concentrations exceeding 1 × 10^?4m, irrespective of the K_m value for the substrate. Specific recommendations are given for future assays.     

On the impact of low concentrations of chlorophos in the period of early ontogenesis on digestive carbohydrases of underyearlings of roach Rutilus rutilus

The remote effects of brief exposure to low doses of chlorophos (1 × 10^?6 ?1 × 10^?2 mg/l) in the period of early ontogenesis of roach Rutilus rutilus have been studied. Diverse changes of the activity of carbohydrases and of the kinetic characteristics of the hydrolysis of di-and polysaccharides in experimental under-yearlings are revealed. Especially interesting are the data on the greater effect of superlow concentrations on changes of the kinetic characteristics; in some cases adaptive changes in the affinity of enzymes to the substrate are recorded.     

Proteases as catalysts of enantioselective α-amino ester hydrolysis: 2. Pronase-catalysed hydrolysis of tryptophan methyl ester

The influence of pH, temperature, EDTA and metal ions on the rate of hydrolysis of methyl esters of d- and l-tryptophan by pronase from Streptomyces griseus has been studied. pH, Ca²⁺ and Co²⁺ ions have different effects on the hydrolysis of d- and l-substrates, which permits the enantioselectivity of the reaction to be regulated. Maximum enantioselectivity was observed at pH 7 in the presence of 10^?3–10^?2m Co²⁺. The data obtained are compared with the literature data for the hydrolysis of other low molecular weight substrates by pronase. It is suggested that leucine aminopeptidase makes a considerable contribution to the observed esterase activity of pronase.     

Four GH11 xylanases from the xylanolytic fungus Talaromyces versatilis act differently on (arabino)xylans

The filamentous fungus Talaromyces versatilis produces a wide range of cellulolytic and hemicellulolytic enzymes such as xylanases. The recent accessibility to the T. versatilis genome allows identifying two new genes, xynE and xynF, encoding glycoside-hydrolases from family GH11. Both genes were cloned and expressed in the methylotrophic yeast Pichia pastoris in order to compare these new xylanases with two other GH11 xylanases from T. versatilis (XynB and XynC) that were previously reported. High-level expression of recombinant enzymes was obtained for the four enzymes that were purified to homogeneity. The XynB, XynC, XynE and XynF enzymes have molecular masses of 34, 22, 45 and 23 kDa, an optimal pH between 3.5 and 4.5 and an optimal temperature between 50 °C and 60 °C. Interestingly, XynF has shown the best thermal stability at 50 °C for at least 180 min with a weak loss of activity. The four xylanases catalysed hydrolysis of low viscosity arabinoxylan (LVAX) with K _m(app) between 11.5 and 23.0 mg.mL^?1 and k _cat/K _m(app) 170 and 3,963 s^?1 mg^?1.mL. Further investigations on the rate and pattern of hydrolysis of the four enzymes on LVAX showed the predominant production of xylose, xylobiose and some (arabino)xylo-oligosaccharides as end products. The initial rate data from the hydrolysis of short xylo-oligosaccharides indicated that the catalytic efficiency increased with increasing degree of polymerisation of oligomer up to 6, suggesting that the specificity region of XynE and XynF spans at least six xylose residues. Because of their attractive properties, T. versatilis xylanases might be considered for biotechnological applications.     

Production of 6-kestose by the filamentous fungus Gliocladium virens as affected by sucrose concentration

The filamentous fungus Gliocladium virens is able to produce fructooligosaccharides (FOS), fructose-containing sugars, used as functional ingredients to improve nutritional and technological properties of foods. In this work we evaluated FOS production by G. virens when grown in a wide range of sucrose concentrations (10–400 g l^?1). High sucrose concentrations increased both biomass and FOS production, including 6-kestose, a trisaccharide comprising β (2 → 6) linked fructosyl units, with enhanced stability and prebiotic activity when compared to the typical FOS β (2 → 1) linked. The highest 6-kestose yield (3 g l^?1) was achieved in media containing 150 g l^?1 sucrose after 4–5 days of culture, production being 90% greater than in media containing 10, 30, or 50 g l^?1 sucrose. After 5 days, FOS production declined markedly, following complete sucrose depletion in the medium. Although most of the β-fructofuranosidases preferentially catalyze sucrose hydrolysis, FOS production in G. virens grown in high sucrose concentration, might be attributed to a reverse hydrolysis by these enzymes. In conclusion, high sucrose concentrations increase growth of G. virens whilst 6-kestose accumulation in the medium seems to be controlled both by specific properties of β-fructofuranosidases and on the sucrose concentration.