Appl Microbiol Biotechnol (2001) 57:647–652
`DOI 10.1007/s00253-001-0845-z
`
`O R I G I N A L PA P E R
`
`F. Jørgensen · O. C. Hansen · P. Stougaard
`High-efficiency synthesis of oligosaccharides
`with a truncated β-galactosidase from Bifidobacterium bifidum
`
`Received: 17 July 2001 / Received revision: 13 September 2001 / Accepted: 14 September 2001 / Published online: 26 October 2001
`© Springer-Verlag 2001
`
`Abstract An exceptionally large β-galactosidase, BIF3,
`with a subunit molecular mass of 188 kDa (1,752 amino
`acid residues) was recently isolated from Bifidobacteri-
`um bifidum DSM20215 [Møller et al. (2001) Appl Envi-
`ron Microbiol 67:2276–2283]. The BIF3 polypeptide
`comprises a signal peptide followed by an N-terminal
`β-galactosidase region and a C-terminal galactose-bind-
`ing motif. We have investigated the functional impor-
`tance of the C-terminal part of the BIF3 sequence by de-
`letion mutagenesis and expression of truncated enzyme
`variants in Escherichia coli. Deletion of approximately
`580 amino acid residues from the C-terminal end con-
`verted the enzyme from a normal, hydrolytic β-galactosi-
`dase into a highly efficient, transgalactosylating enzyme.
`Quantitative analysis showed that the truncated β-galac-
`tosidase utilised approximately 90% of the reacted lac-
`tose for the production of galacto-oligosaccharides,
`while hydrolysis constituted a 10% side reaction. This
`9:1 ratio of transgalactosylation to hydrolysis was main-
`tained at lactose concentrations ranging from 10% to
`40%, implying that the truncated β-galactosidase be-
`haved as a “true” transgalactosylase even at low lactose
`concentrations.
`
`Introduction
`The enzyme β-galactosidase (EC 3.2.1.23), which will
`usually hydrolyse lactose to the monosaccharides D-glu-
`cose and D-galactose, may also catalyse the formation of
`galacto-oligosaccharides (Wallenfels 1951). The obser-
`vation that galacto-oligosaccharides stimulate the growth
`of bifidobacteria and other health-promoting intestinal
`bacteria (Minami et al. 1983; Rabiu et al. 2001; Rowland
`and Tanaka 1993) has renewed interest in the transferase
`reaction of β-galactosidases, and several reports have
`F. Jørgensen · O.C. Hansen · P. Stougaard (✉
`)
`Department of Enzyme Technology, Biotechnological Institute,
`Kogle Allé 2, 2970 Hørsholm, Denmark
`e-mail: pst@bioteknologisk.dk
`Tel.: +45-45-160444, Fax: +45-45-160455
`
`been published on enzymatic synthesis of galacto-oligo-
`saccharides (Huber and Hurlburt 1986; Mozaffar et al.
`1986; Nakao et al. 1994; Ohtsuka et al. 1990; Onishi et
`al. 1995; Petzelbauer et al. 2000; Rabiu et al. 2001;
`Smart 1991; Stevenson and Furneaux 1996; Yang and
`Tang 1988).
`During the normal hydrolytic reaction, β-galactosi-
`dase will hydrolyse lactose and transfer galactose to the
`hydroxyl group of water, resulting in the liberation of
`D-galactose and D-glucose. Some β-galactosidases, how-
`ever, are able to transfer galactose to the hydroxyl
`groups of the D-galactose or the D-glucose moiety in lac-
`tose, resulting in the formation of galacto-oligosaccha-
`rides. This transgalactosylation reaction has been shown
`in a number of studies to depend on the origin of the en-
`zyme, the reaction temperature and pH, and – in particu-
`lar – on a high concentration of lactose (Boon et al.
`2000; Burvall et al. 1979; Huber et al. 1976; Rustom et
`al. 1998; Yoon and Ajisaka 1996). Recently, successful
`attempts to improve oligosaccharide formation by site-
`directed mutagenesis of glycolytic enzymes have been
`reported (Hansson et al. 2001; Ly and Withers 1999;
`Mackenzie et al. 1998; Mayer et al. 2000).
`In an earlier report we described the screening of a
`number of Bifidobacterium strains and the identification
`and characterisation of an unusual, transgalactosylating
`β-galactosidase from Bifidobacterium bifidum, including
`gene cloning and heterologous expression of the enzyme
`in Escherichia coli (Møller et al. 2001). This β-galacto-
`sidase, BIF3, from Bifidobacterium bifidum DSM20215
`is exceptionally large (188 kDa), and the amino acid se-
`quence comprises an N-terminal signal sequence that is
`cleaved off during heterologous expression in E. coli.
`Amino acid sequence similarity searches have shown
`that the N-terminal part of the sequence is homologous
`to other β-galactosidase sequences and that the C-termi-
`nal part of the polypeptide is similar to galactose-binding
`protein domains.
`In this paper, we have analysed the functional signifi-
`cance of the C-terminal part of the enzyme. By deletion
`mutagenesis and expression of the truncated enzyme
`
`

`

`648
`
`variants in E. coli we have found that the enzyme can be
`engineered into a very efficient “transgalactosylase” that
`produces high levels of galacto-oligosaccharides even at
`low concentrations of lactose.
`
`Materials and methods
`
`Chemicals
`
`Chemicals were generally purchased from Sigma (St. Louis, Mo.,
`USA). Restriction enzymes and other enzymes used for DNA ma-
`nipulation were from New England Biolabs (Beverly, Mass.,
`USA), and were used according to the manufacturer’s instructions.
`
`Bacterial strains and growth media
`The plasmid pBIF3 containing a β-galactosidase from Bifidobac-
`terium bifidum DSM20215 (EMBL Nucleotide Sequence Data-
`base accession number AJ 224435) was isolated as described pre-
`viously (Møller et al. 2001). DNA cloning and expression of
`recombinant β-galactosidases was performed in E. coli JM105
`(Yanisch-Perron et al. 1985). E. coli cells were grown in LB medi-
`um (Miller 1972) supplemented with 100 µg/ml of ampicillin and
`solidified with 1.5% agar when appropriate.
`
`Construction of truncated BIF3 β-galactosidases
`
`The plasmid pBIF3 (Møller et al. 2001) was subjected to restric-
`tion enzyme digestion in order to remove parts of the β-galactosi-
`dase gene from the 3’-end of the coding region. Digestion with
`PstI, EcoRI, and partially with BglII removed approximately 450,
`1,280, 1,740, and 2,310 base pairs, respectively, of the β-galacto-
`sidase gene from the 3’-end. The resulting plasmids were deno-
`ted pBIF3-d1, pBIF3-d2, pBIF3-d3, and pBIF3-d4, respectively
`(Fig. 1).
`
`Preparation of cell extracts
`
`E. coli cells harbouring pBIF3 or deletion derivatives were grown
`in LB medium, harvested by centrifugation, resuspended in
`Z-buffer (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM
`MgSO4, and 50 mM β-mercaptoethanol, pH 7.0), and lysed in a
`FastPrep instrument (Qbiogene, Carlsbad, Calif., USA) according
`to the manufacturer’s procedure.
`
`Enzyme assays and analysis of oligosaccharides
`Hydrolysis of o-nitrophenyl-β-D-galactopyranoside (ONPG) fol-
`lowed by measurement of absorbance at 420 nm was used for de-
`termination of β-galactosidase activity (Miller 1972). Assays were
`performed in Z-buffer containing 0.5% Triton X-100, and the re-
`actions were stopped by addition of 1 M Na2CO3. Transgalactosy-
`lation assays were performed in a buffer containing 50 mM sodi-
`um citrate, and 100 mM Na2HPO4, pH 6.0. The 50-µl reaction
`volumes containing 0.4 M lactose, or as specified otherwise, were
`incubated for 20 h at 37°C. A 5-min incubation at 95°C was used
`to stop the enzyme reaction. The reaction mixtures were analysed
`by thin layer chromatography (TLC) on Silica-gel 60 plates
`(Merck) in a solvent containing butanol, 2-propanol, and water
`(3:12:4). Samples of 1 µl of diluted sample (1:1 dilution in water)
`were subjected to three runs. After drying, the sugars were visual-
`ised by spraying with an orcinol reagent, followed by incubation
`at 100°C for 5–10 min. Conversion of radioactively labelled
`[D-glucose-1-14C]-lactose, 2 GBq/mmol (Amersham Pharmacia
`Biotech) was performed in 20-µl reaction volumes containing 15%
`unlabelled lactose (w/v) and 370 Bq/µl of 14C-labelled lactose. Af-
`
`Fig. 1 Map of genes encoding BIF3 β-galactosidase and truncated
`enzyme variants derived from BIF3 by deletion mutagenesis.
`White bars indicate sequences coding for the signal sequence,
`black bars the putative galactose-binding motif, and dark-gray
`bars the sequence showing homology to known β-galactosidases.
`Relative β-galactosidase activities were measured in o-nitrophe-
`nyl-β-D-galactopyranoside (ONPG) assays using Escherichia coli
`lysates of overnight cultures. Restriction enzyme sites for BglII
`(B), EcoRI (E), and PstI (P) are indicated
`
`ter TLC as described above, the TLC plate was subjected to auto-
`radiography and quantitation in a Cyclone Storage Phosphor
`System (Packard Instrument Company, Meriden, Conn., USA) ac-
`cording to the procedure recommended by the manufacturer.
`Quantitative analysis of the carbohydrates formed during
`transgalactosylation was performed by high-performance liquid
`chromatography (HPLC) on a Shodex Sugar KS-801 column
`(300×8 mm; Showa Denko, Tokyo, Japan) using water as eluent
`and refractive index detection. The yield of galacto-oligosaccha-
`rides was calculated from the concentrations (in g/l) of lactose, ga-
`lactose and glucose measured by HPLC: [gos]=[lacinitial]-[lacfinal]-
`[glcfinal]-[galfinal]. The relative carbohydrate yields were calculated
`as final carbohydrate concentration as a percentage of initial lac-
`tose concentration.
`
`Results
`
`Construction and characterisation of truncated
`BIF3 β-galactosidases
`
`The large BIF3 polypeptide is composed of two regions,
`an N-terminal domain, which shows homology to other
`β-galactosidases, and a C-terminal part with homology
`to a sialidase-like galactose-binding motif (Møller et al
`2001). In order to investigate the role of the C-terminal
`part of the BIF3 β-galactosidase polypeptide, we con-
`structed four deletion mutants of the enzyme (BIF3-d1,
`BIF3-d2, BIF3-d3, and BIF3-d4) and analysed the
`truncated β-galactosidases for their hydrolytic activity
`(Fig. 1). These experiments showed that it was possible
`to delete almost one-third of the BIF3 enzyme and still
`retain hydrolytic activity. However, the relative activity
`of the deletion mutants was not reduced linearly (Fig. 1),
`since deletion mutant BIF3-d1 showed only 19% relative
`activity whereas BIF3-d2, containing a larger deletion,
`showed 30% relative activity. This discrepancy, possibly
`reflecting different folding of the enzyme, was not inves-
`tigated further.
`
`

`

`649
`
`Fig. 2 Transgalactosylating activity of BIF3 β-galactosidase and
`the deletion mutant BIF3-d3. Reaction mixtures containing 10%
`lactose and different amounts of recombinant enzyme produced in
`E. coli were incubated at 37°C for 20 h and analysed by thin layer
`chromatography. The position of lactose (lac), galactose (gal), and
`glucose (glc) in separate lanes is indicated by arrows. Similarly,
`arrows indicate the position of galacto-oligosaccharides (gos) in-
`cluding presumed disaccharides
`
`Determination of the transgalactosylating activity of
`the deletion mutants showed similar activities in extracts
`of E. coli cells containing the plasmids pBIF3, pBIF3-d1,
`and pBIF3-d2. Extracts of E. coli cells harbouring
`pBIF3-d3, however, showed an increased level of oligo-
`saccharide production and a relatively small release of
`galactose (Fig. 2). This result indicated that the truncated
`BIF3-d3 β-galactosidase does indeed hydrolyse lactose,
`but instead of transferring the galactose moiety onto the
`hydroxyl groups of water (which is done by normal hy-
`drolytic β-galactosidases) the enzyme was found to use
`virtually all galactose molecules in a transgalactosylating
`reaction.
`
`Quantification of transgalactosylating activity
`The transgalactosylating activity of the truncated β-
`galactosidase was analysed quantitatively by TLC of re-
`action mixtures containing radioactively labelled lactose
`as a substrate. Because the radioactively labelled lactose
`contained 14C at the C-1 position in the glucose part of
`the disaccharide, only glucose-containing products were
`detected (Fig. 3). The assay mixture containing 15% lac-
`tose was incubated at 37°C for 20 h. After separation of
`the reaction products by TLC the 14C-labelled carbohy-
`drates were quantitated with a Phospho-Imager and the
`distribution of 14C-labelled glucose between lactose, glu-
`cose, and oligosaccharides was determined (Fig. 3b).
`For low amounts of BIF3-d3 enzyme, the number of
`glucose molecules produced was quite similar to the
`
`Fig. 3A,B Transgalactosylating activity of the deletion mutant
`BIF3-d3 in the presence of 14C-labelled lactose. Reaction mixtures
`containing 15% 14C-labelled lactose and different amounts of re-
`combinant enzyme produced in E. coli were incubated at 37°C for
`20 h and analysed by thin layer chromatography. The 14C-labelled
`sugars were visualised by autoradiography of the chromatogram
`(A), and the relative amounts of 14C-labelled lactose (lac, open
`circles), glucose (glc, open squares) and galacto-oligosaccharides
`(gos, closed triangles) were quantitated with a Phospho-Imager
`(B)
`
`number of galacto-oligosaccharide molecules produced.
`This implies that almost every time a molecule of lactose
`is processed by the BIF3-d3 β-galactosidase enzyme, the
`galactose part is transferred to another molecule of lac-
`tose in a transgalactosylation reaction. Thus the BIF3-d3
`enzyme behaves like a true “transgalactosylase”. At
`higher concentrations of the BIF3-d3 enzyme, the ratio
`
`

`

`650
`
`Fig. 4 Transgalactosylating activity of the deletion mutant BIF3-d3
`in the presence of lactose. Reaction mixtures containing 10%,
`20%, or 40% lactose and identical amounts of recombinant en-
`zyme produced in E. coli were incubated at 37°C for 20 h and
`analysed by high-performance liquid chromatography (HPLC).
`The yield of galacto-oligosaccharides was calculated from the
`concentrations (in g/l) of lactose, galactose, and glucose measured
`by HPLC: [gos]=[lacinitial] – [lacfinal] – [glcfinal] – [galfinal]. The rel-
`ative carbohydrate yields shown in the figure were calculated as
`final carbohydrate concentration as a percentage of initial lactose
`concentration
`
`between radioactive oligosaccharides and glucose was
`reduced, probably due to formation of di-galactose com-
`pounds, which would be unlabelled and therefore invisi-
`ble on TLC. Approximately 32% of the 14C-labelled glu-
`cose were recovered as oligosaccharides using the opti-
`mal amount of enzyme.
`In another experiment 10%, 20%, and 40% of unla-
`belled lactose was used as a substrate in enzyme reac-
`tions with identical amounts of the BIF3-d3 enzyme. The
`concentrations of lactose, glucose, and galactose in the
`reaction mixtures were measured by HPLC and the con-
`centrations of oligosaccharides were calculated (Fig. 4).
`The yield of galacto-oligosaccharides was 39%, 44%,
`and 37% in reactions containing 10%, 20%, and 40%
`lactose, respectively. Only very low concentrations of
`galactose were detected, indicating that almost all galac-
`tose molecules derived from lactose were transferred on-
`to another sugar. The ratio of free galactose to glucose
`constituted 0.05–0.10 in all the reactions with 10%,
`20%, or 40% lactose. In other words, for every molecule
`of lactose hydrolyzed into galactose and glucose, at least
`nine lactose molecules were processed by the transgalac-
`tosylation pathway of the BIF3-d3 enzyme into glucose
`and galacto-oligosaccharides. The same ratio between
`liberated galactose and glucose was found when differ-
`ent amounts of enzyme were used in similar experiments
`with 10%, 20% or 40% lactose. These data supported the
`experiments with radioactively labelled lactose.
`
`Fig. 5A,B Hydrolytic activity of BIF3 β-galactosidase (white
`circles) and the deletion mutant BIF3-d3 (black circles) in the
`presence of glucose or galactose. ONPG assays were performed at
`37°C with addition of 0–10% glucose (A) or galactose (B)
`
`Characterisation of the hydrolytic activity
`of truncated β-galactosidases
`The hydrolytic activity of truncated β-galactosidases was
`measured in ONPG assays. All the deletion mutants had
`a temperature optimum at 35°C–40°C, similar to that of
`the native BIF3 enzyme. Likewise, all the truncated en-
`zyme variants showed a broad pH optimum between
`pH 5.0 and 7.0, similar to the pH optimum of the BIF3
`β-galactosidase (data not shown).
`The effect of exogenously added galactose and glu-
`cose on the hydrolytic activity of the truncated β-galac-
`tosidases and the native BIF3 enzyme was also investi-
`gated. Glucose had no effect on the hydrolytic activity of
`native BIF3 enzyme (or of the truncated BIF3-d1 or
`BIF3-d2 versions, data not shown), whereas BIF3-d3 ac-
`tivity was immensely stimulated (Fig. 5A). Addition of
`low concentrations of galactose inhibited the hydrolytic
`activity of the native BIF3 enzyme (and of BIF3-d1 and
`BIF3-d2, not shown), whereas the deletion mutant BIF3-d3
`was inhibited to a minor degree (Fig. 5B). The inhibitory
`and stimulatory effects of glucose and galactose were de-
`
`

`

`termined with ONPG as a substrate. Whether similar ef-
`fects would be observed with lactose as a substrate re-
`mains to be shown.
`Finally, the temperature stability of the truncated β-
`galactosidases was investigated and compared with that
`of the native BIF3 enzyme. E. coli lysates from over-
`night cultures were preincubated at 45°C and assayed for
`residual β-galactosidase activity in ONPG assays per-
`formed at 37°C. The truncated enzyme variants exhibited
`a temperature stability that was comparable to that of
`BIF3, although the deletion mutants BIF3-d1 and BIF3-d3
`displayed a slightly higher and slightly lower tempera-
`ture stability, respectively.
`
`Discussion
`
`Galacto-oligosaccharides have been investigated thor-
`oughly in recent years due to their ability to promote the
`growth of beneficial intestinal bacteria. However, be-
`cause of difficulties in producing large amounts of galac-
`to-oligosaccharides, only a few products have been test-
`ed rigorously and have reached the market. Two major
`enzymatic approaches have been proposed for produc-
`tion of galacto-oligosaccharides: galactosyl transferase-
`catalysed synthesis using nucleotide phospho-sugars as
`substrates, and β-galactosidase-catalysed synthesis using
`lactose as a substrate. However, due to availability and
`cost of substrates and enzymes, only the latter process is
`presently believed to be feasible.
`Transgalactosylation is a process in which the enzyme
`β-galactosidase hydrolyses lactose and instead of trans-
`ferring the galactose moiety to the hydroxyl group of
`water, the enzyme transfers galactose to an accepting al-
`cohol group of another carbohydrate, e.g., glucose, ga-
`lactose, lactose, or galactose-containing oligosaccha-
`rides. In normal β-galactosidase reactions, transgalac-
`tosylation is inferior to hydrolysis, since the number of
`hydroxyl groups of water is in great excess to the hy-
`droxyl groups of carbohydrates. In order to increase
`transgalactosylation, high concentrations of lactose are
`usually required (Huber et al. 1976), and due to the solu-
`bility of lactose, high temperature reactions have been
`desirable. Therefore, various thermophilic microorgan-
`isms have been investigated for thermostable β-galacto-
`sidases capable of producing oligosaccharides at high
`temperatures. For example, it has been reported that a β-
`galactosidase from Sterigmatomyces elviae produces
`39% oligosaccharides in a solution of 20% lactose at
`60°C (Onishi and Tanaka 1995), and that the enzyme
`from Saccharopolyspora rectivirgula can synthesise 41%
`oligosaccharides in 60% lactose at 70°C (Nakao et al.
`1994).
`Another approach to increase the yield of galacto-oli-
`gosaccharides has been to use protein engineering tech-
`nology. For example, knowing the three-dimensional
`structure of β-glucosidases from Agrobacterium and
`Pyrococcus furiosus, it was possible to change specific
`amino acids in the active site, thereby improving the
`
`651
`
`formation of oligosaccharides (Hansson et al. 2001;
`Mackenzie et al. 1998; Mayer et al. 2000).
`However, in most cases – including the BIF3 enzyme
`in this study – the three-dimensional structure of the
`β-galactosidase is not known, and therefore it would
`be very difficult to predict advantageous, site-specific
`changes in the enzyme. Nevertheless, without knowing
`the structure of the enzyme, we have improved the galac-
`to-oligosaccharide synthesis ability of the BIF3 β-galac-
`tosidase from B. bifidum using deletion mutagenesis.
`Like many other β-galactosidases, the native BIF3 en-
`zyme has been shown to transgalactosylate at high lac-
`tose concentrations (Møller et al. 2001). Here we show
`that removal of approximately 580 amino acids, includ-
`ing a putative galactose-binding motif, from the native
`BIF3 enzyme greatly improves the ability to synthesise
`galacto-oligosaccharides. The truncated enzyme, BIF3-d3,
`was shown not to require high lactose concentrations, in
`contrast to all previous reports on transgalactosylation
`using β-galactosidase enzymes (Fig. 4). Quantitative
`measurements showed that even at 10% lactose the en-
`zyme was able to produce at least 39% oligosaccharides
`at 37°C. The engineered, truncated enzyme acted as a
`true “transgalactosylase” in more than 90% of the en-
`zyme reactions, whereas hydrolysis was reduced to a
`10% side reaction (Figs. 3 and 4).
`Other remarkable features of the truncated BIF3-d3
`β-galactosidase include the relative galactose insensitivi-
`ty and the apparent stimulation of hydrolytic activity
`caused by the addition of glucose. Figure 5 shows that,
`in contrast to the native BIF3 enzyme, the hydrolytic ac-
`tivity of BIF3-d3 is inhibited by galactose to a much
`lesser degree. This effect is probably due to the fact that
`the exogenously added galactose acts as acceptor mole-
`cule in the concomitant transgalactosylation reaction, re-
`sulting in an increased ONPG-hydrolysing activity.
`However, since galactose also acts as an inhibitor for the
`BIF3-d3 enzyme, an overall reduction of activity is ob-
`served for high concentrations of exogenously added ga-
`lactose. Likewise, the addition of glucose stimulates
`ONPG hydrolysis, probably due to glucose acting as ac-
`ceptor in a transgalactosylation reaction with ONPG as
`donor. Similar stimulating and inhibiting effects of
`glucose and galactose have been observed in a study
`on transgalactosylation by thermostable β-glycosidases
`from Pyrococcus furiosus and Sulfolobus solfataricus
`(Petzelbauer et al. 2000).
`Comparison of the β-galactosidase BIF3 and the trun-
`cated enzyme variants with respect to temperature stabil-
`ity, temperature optimum, and pH optimum showed
`comparable temperature and pH profiles. The “transga-
`lactosylase” BIF3-d3 was the least-stable enzyme, but it
`showed pH and temperature optima similar to the native
`BIF3 β-galactosidase (data not shown).
`The BIF3 β-galactosidase contains a putative galac-
`tose-binding site in the C-terminal part of the molecule,
`but this domain is probably not involved in transgalac-
`tosylation, since the putative galactose binding site has
`been deleted in the truncated BIF3-d3 “transgalactosyl-
`
`

`

`652
`
`ase” (Fig. 1). However, since the thermostability of the
`BIF3-d3 enzyme is lower than that of the native β-ga-
`lactosidase, the truncated enzyme may be of a more-
`open structure, a feature that facilitates transgalactosy-
`lation.
`The catalytic event responsible for transgalactosylat-
`ion is the double displacement mechanism, which is
`characteristic for so-called retaining glycosyl hydrolases
`(Davies and Henrissat 1995; Sinnott 1990). Apparently,
`for the BIF3-d3 ß-galactosidase it is the reaction be-
`tween water and the enzyme-galactose intermediate that
`is affected, an observation difficult to understand in
`terms of steric hindrance. Instead, we speculate that
`binding between enzyme and acceptor may be a require-
`ment for the BIF3-d3-catalysed reaction. However, at the
`moment the molecular nature behind the increased trans-
`galactosylation of the BIF3-d3 β-galactosidase is un-
`known and remains to be elucidated.
`
`Acknowledgements The financial support from Arla Foods amba
`is gratefully acknowledged. We thank Bente Smith for excellent
`technical assistance.
`
`References
`
`Boon MA, Riet K van’t, Janssen AEM (2000) Enzymatic synthe-
`sis of oligosaccharides: product removal during a kinetically
`controlled reaction. Biotechnol Bioeng 70:411–420
`Burvall A, Asp N-G, Dahlqvist A (1979) Oligosaccharide forma-
`tion during hydrolysis of lactose with Saccharomyces lactis
`lactase (Maxilact). 1. Quantitative aspects. Food Chem 4:243–
`250
`Davies G, Henrissat B (1995) Structures and mechanisms of gly-
`cosyl hydrolases. Structure 3:853–859
`Hansson T, Kaper T, Oost J van der, Vos WM de, Adlercreutz P
`(2001) Improved oligosaccharide synthesis by protein engi-
`neering of beta-glucosidase CelB from hyperthermophilic
`Pyrococcus furiosus. Biotechnol Bioeng 73:203–210
`Huber RE, Hurlburt KL (1986) Reversion reactions of β-galactosi-
`dase (Escherichia coli). Arch Biochem Biophys 246:411–
`418
`Huber RE, Kurz G, Wallenfels K (1976) A quantitation of the
`factors which affect the hydrolase and transgalactosylase ac-
`tivities of β-galactosidase (E. coli) on lactose. Biochemistry
`15:1994–2001
`Ly HD, Withers SG (1999) Mutagenesis of glycosidases. Annu
`Rev Biochem 68:487–522
`Mackenzie LF, Wang Q, Warren RAJ, Withers SG (1998) Glyco-
`synthases: mutant glycosidases for oligosaccharide synthesis.
`J Am Chem Soc 120:5583–5584
`Mayer C, Zechel DL, Reid SP, Warren RAJ, Withers SG (2000)
`The E358S mutant of Agrobacterium sp. β-glucosidase is a
`greatly improved glycosynthase. FEBS Lett 466:40–44
`Miller HJ (1972) Experiments in molecular genetics. Cold Spring
`Harbor Laboratory, Cold Spring Harbor, NY
`
`Minami Y, Yazawa K, Tamura Z, Tanaka T, Yamamoto T (1983)
`Selectivity of utilization of galactosyl-oligosaccharides by bi-
`fidobacteria. Chem Pharm Bull (Tokyo) 31:1688–1691
`Møller PL, Jørgensen F, Hansen OC, Madsen SM, Stougaard P
`(2001) Intra- and extracellular ß-galactosidases from Bifido-
`bacterium bifidum and B. infantis: molecular cloning, heterol-
`ogous expression, and comparative characterization. Appl
`Environ Microbiol 67:2276–2283
`Mozaffar Z, Nakanishi K, Matsuno R (1986) Continuous produc-
`tion of galacto-oligosaccharides from lactose using immobi-
`lized β-galactosidase from Bacillus circulans. Appl Microbiol
`Biotechnol 25:224–228
`Nakao M, Harada M, Kodama Y, Nakayama T, Shibano Y,
`Amachi T (1994) Purification and characterization of a
`thermostable β-galactosidase with high transgalactosylation
`activity from Saccharopolyspora rectivirgula. Appl Microbiol
`Biotechnol 40:657–663
`Ohtsuka K, Tanoh A, Ozawa O, Kanematsu T, Uchida T, Shinke R
`(1990) Purification and properties of a β-galactosidase with
`high galactosyl transfer activity from Cryptococcus laurentii
`OKN-4. J Ferment Bioeng 70:301–307
`Onishi N, Tanaka T (1995) Purification and properties of a novel
`thermostable galacto-oligosaccharide-producing beta-galacto-
`sidase from Sterigmatomyces elviae CBS8119. Appl Environ
`Microbiol 61:4026–4030
`Onishi N, Yamashiro A, Yokozeki K (1995) Production of galacto-
`oligosaccharide from
`lactose by Sterigmatomyces elviae
`CBS8119. Appl Environ Microbiol 61:4022–4025
`Petzelbauer I, Reiter A, Splechtna B, Kosma, P, Nidetsky B (2000)
`Transgalactosylation by thermostable β-glycosidases from
`Pyrococcus furiosus and Sulfolobus solfataricus. Eur J Bio-
`chem 267:5055–5066
`Rabiu BA, Jay AJ, Gibson GR, Rastall RA (2001) Synthesis and
`fermentation properties of novel galacto-oligosaccharides by
`β-galactosidases from Bifidobacterium species. Appl Environ
`Microbiol 67:2526–2530
`Rowland IR, Tanaka R (1993) The effects of transgalactosylated
`oligosaccharides on gut flora metabolism in rats associated
`with a human faecal microflora. J Appl Bacteriol 74:667–674
`Rustom IYS, Foda MI, López-Leiva MH (1998) Formation of oli-
`gosaccharides from whey UF-permeate by enzymatic hydroly-
`sis – analysis of factors. Food Chem 62:141–147
`Sinnott ML (1990) Catalytic mechanisms of enzymic glycosyl
`transfer. Chem Rev 90:1171–1202
`Smart JB (1991) Transferase reactions of the β-galactosidase from
`Streptococcus thermophilus. Appl Microbiol Biotechnol 34:
`495–501
`Stevenson DE, Furneaux RH (1996) Synthesis of allyl β-D-
`galactopyranoside from lactose using Streptococcus thermo-
`philus β-D-galactosidase. Carbohydr Res 284:279–283
`Wallenfels K (1951) Enzymatische Synthese von Oligosacchari-
`den auf Disacchariden. Naturwissenschaften 38:306–307
`Yang ST, Tang IC (1988) Lactose hydrolysis and oligosaccharide
`formation catalyzed by beta-galactosidase. Kinetics and
`mathematical modeling. Ann N Y Acad Sci 542:417–422
`Yanisch-Perron C, Vieira J, Messing J (1985) Improved M13
`phage cloning vectors and host strains: nucleotide sequences
`of the M13mp18 and pUC19 vectors. Gene 33:103–119
`Yoon JH, Ajisaka K (1996) The synthesis of galactopyranosyl de-
`rivatives with β-galactosidases of different origins. Carbohydr
`Res 292:153–163
`
`