New Biotechnology Volume 00, Number 00 January 2016
RESEARCH PAPER
Research Paper
Thermostable b-galactosidases for the synthesis of human milk oligosaccharides Birgitte Zeuner1, Christian Nyffenegger1, Jørn Dalgaard Mikkelsen and Anne S. Meyer Q1 Center for BioProcess Engineering, Department of Chemical and Biochemical Engineering, Technical University of Denmark, Building 229, DK-2800 Kgs., Lyngby, Denmark
Human milk oligosaccharides (HMOs) designate a unique family of bioactive lactose-based molecules present in human breast milk. Using lactose as a cheap donor, some b-galactosidases (EC 3.2.1.23) can catalyze transgalactosylation to form the human milk oligosaccharide lacto-N-neotetraose (LNnT; Galb(1,4)-GlcNAc-b(1,3)-Gal-b(1,4)-Glc). In order to reduce reaction times and be able to work at temperatures, which are less welcoming to microbial growth, the current study investigates the possibility of using thermostable b-galactosidases for synthesis of LNnT and N-acetyllactosamine (LacNAc; Gal-b(1,4)-GlcNAc), the latter being a core structure in HMOs. Two hyperthermostable GH 1 b-galactosidases, Ttb-gly from Thermus thermophilus HB27 and CelB from Pyrococcus furiosus, were codon-optimized for expression in Escherichia coli along with BgaD-D, a truncated version of the GH 42 b-galactosidase from Bacillus circulans showing high transgalactosylation activity at low substrate concentrations. The three b-galactosidases were compared in the current study in terms of their transgalactosylation activity in the formation of LacNAc and LNnT. In all cases, BgaD-D was the most potent transgalactosidase, but both thermostable GH 1 b-galactosidases could catalyze formation of LNnT and LacNAc, with Ttb-gly giving higher yields than CelB. The thermal stability of the three b-galactosidases was elucidated and the results were used to optimize the reaction efficiency in the formation of LacNAc, resulting in 5–6 times higher reaction yields and significantly shorter reaction times.
Introduction Q2 Human milk oligosaccharides (HMOs), which are abundant in human breast milk (present at levels up to 8 g/L) are known to be of major importance for infant health of development [1]. The HMOs share similar structural features: They are all based on a reducing end lactose molecule, which can be substituted with sialic acid or fucose and elongated with N-acetylglucosamine (GlcNAc) and galactose (Gal) units; these elongated molecules may in turn also be substituted with sialyl or fucosyl residues [1,2]. Important elongated HMO core structures include
Corresponding author: Meyer, A.S. ([email protected]) 1
These two authors contributed equally.
http://dx.doi.org/10.1016/j.nbt.2016.01.003 1871-6784/ß 2016 Published by Elsevier B.V.
lacto-N-tetraose (LNT; Gal-b(1,3)-GlcNAc-b(1,3)-Gal-b(1,4)-Glc) and lacto-N-neotetraose (LNnT; Gal-b(1,4)-GlcNAc-b(1,3)-Galb(1,4)-Glc) [2]. Approaches to synthesize LNT and LNnT include chemical synthesis [3–6], in vivo production in Escherichia coli [7], the use of glycosyl transferases requiring nucleotide sugars [8], as well as the use of glycosidases with transglycosylation activity [9]. More than a decade ago, it was shown that lacto-N-triose II (LNT2; GlcNAc-b(1,3)-Gal-b(1,4)-Glc) could be elongated to LNnT by transgalactosylation catalyzed by the GH 42 b-galactosidase from Bacillus circulans using lactose as donor substrate [9]. Using the GH 35 b-galactosidase from Bacillus circulans and p-nitrophenyl b-galactopyranoside (pNP-b-Gal) as donor, LNT (and LNT isomers, most likely LNnT) was also reported to be formed from www.elsevier.com/locate/nbt
Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
1
NBT 850 1–6 RESEARCH PAPER
New Biotechnology Volume 00, Number 00 January 2016
Research Paper
LNT2, but the reaction was not performed with lactose as donor substrate [9]. Similarly, the use of lacto-N-biosidase (EC 3.2.1.140) from Aureobacterium sp. L-101 which could catalyze the transfer of the lacto-N-biose disaccharide to lactose to form LNT also required pNP activation of the donor substrate [9]. In the quest for finding routes to HMO production which involve cheap substrates only, we have recently shown that LNT2 can be produced from chitin degradation products and lactose by two novel b-N-acetylhexosaminidases at molar yields of 2–8% [10]. Previously, LNT2 synthesis has relied on nucleotide, methyl, or pNP-activated donors or chemical synthesis [9,11], all keeping the price of LNT2 too high for a feasible large-scale production. With the possibility of using chitin degradation products as donor substrates and achieving appreciable yields, the b-galactosidase-catalyzed route to production of LNnT with lactose as a cheap donor substrate is becoming increasingly interesting from an economic point of view. In the only published report on b-galactosidase-catalyzed formation of LNnT from lactose and LNT2, Murata and co-workers [9] reached a molar yield based on LNT2 of 19% by performing the reaction for 36 hours at 408C. Such reaction conditions are however problematic with regard to microbial safety. In order to speed up the reaction and be able to work at a temperature, which is less welcoming to microbial growth, the current study investigates the possibility of using thermostable b-galactosidases for LNnT synthesis. A b-glycosidase from Thermus thermophilus HB27, known as Ttb-gly or TTP0042, has high thermal stability [12,13] and has been shown to catalyze the transgalactosylation of GlcNAc to form N-acetyllactosamine (Gal-b(1,4)-GlcNAc; LacNAc) using pNP-bGal as donor [13]. b-Glycosidase from Pyrococcus furiosus, known as CelB, has even higher thermal stability than Ttb-gly [14]. This b-galactosidase has not previously been used for transgalactosylation of GlcNAc moieties, but has been shown to efficiently catalyze the formation of galactooligosaccharides at 958C [15]. Thus, the current work was also undertaken with the objective of comparing these two thermostable b-galactosidases to the benchmark enzyme used for synthesis of LNnT as well as for LacNAc, the GH42 b-galactosidase from Bacillus circulans which is found in the popular commercial Biolacta1 preparation [9,16]. The Biolacta1 preparation contains several different isozymes, but the one used in the current study (BgaD-D) is the one having the highest transgalactosylation activity at low substrate concentrations [17,18]. Ttb-gly, CelB, and BgaD-D (Table 1) were codon-optimized for expression in E. coli, His6-tagged for purification purposes, and compared in the current study in terms of their transgalactosylation activity in the formation of LacNAc and LNnT at elevated temperatures (Fig. 1). In
addition, a sequential reaction with b-N-acetylhexosaminidase and b-galactosidase to form LNnT from lactose and the chitin degradation product N,N0 -diacetylchitobiose ((GlcNAc)2; GlcNAc-b(1,4)GlcNAc) was tested (Fig. 2).
Materials and methods Chemicals Lacto-N-triose II (GlcNAc-b(1,3)-Gal-b(1,4)-Glc; LNT2), lacto-Nbiose (Gal-b(1,3)-GlcNAc), lacto-N-tetraose (Gal-b(1,3)-GlcNAcb(1,3)-Gal-b(1,4)-Glc; LNT), lacto-N-neotetraose (Gal-b(1,4)GlcNAc-b(1,3)-Gal-b(1,4)-Glc; LNnT), and N,N0 -diacetylchitobiose (GlcNAc-b(1,4)-GlcNAc; (GlcNAc)2) were purchased from Carbosynth Ltd. (Compton, United Kingdom). For analytical confirmation, LNT and LNnT were also purchased from Elicityl SA (Crolles, France). N-Acetylglucosamine (GlcNAc), N-acetyllactosamine (Gal-b(1,4)-GlcNAc; LacNAc), allo-N-acetyllactosamine (Galb(1,6)-GlcNAc; allo-LacNAc), p-nitrophenyl b-galactopyranoside (pNP-b-Gal), b-lactose, and all other chemicals were purchased from Sigma–Aldrich (Steinheim, Germany).
Cloning, expression, and purification of b-galactosidases Genes encoding the three b-galactosidases BgaD-D, Ttb-gly, and CelB, each possessing an N-terminal His6-tag linked via a thrombin recognition site, were codon-optimized for expression in Escherichia coli (E. coli), synthesized and inserted into the vector pJ411 by DNA2.0 (Menlo Park, CA, USA). E. coli BL21 (DE3) was transformed with the resulting plasmids and selected for kanamycin resistance. BgaD-D was expressed as follows: transformants were grown in lysogenic broth (LB) supplemented with 50 mg/ml kanamycin at 378C. When reaching an OD600 of 0.6, the temperature was reduced to 258C and expression induced by the addition of IPTG to a final concentration of 1 mM. Cells were grown for 15 hours post induction before harvest. Ttb-gly and CelB were expressed in a similar manner, except that the cells were allowed to grow for 5 hours after induction before being harvested. Cells were centrifuged and the pellets resuspended in binding buffer (BgaD-D: 20 mM Na-phosphate buffer, 500 mM NaCl, 20 mM imidazole, pH 7; Ttb-gly and CelB: 40 mM EPPS buffer, 500 mM NaCl, 24 mM imidazole), followed by sonication to open the cells and centrifugation to remove cell debris. The supernatant containing CelB or Ttb-gly was pre-purified by incubation for 20 min at 808C resulting in precipitation of non-thermostable proteins, followed by another centrifugation step. The supernatant containing the b-galactosidases was passed through a 0.45 mm filter before being loaded onto a 1 ml Ni2+sepharose HisTrap HP column (GE Healthcare, Uppsala, Sweden),
TABLE 1
An overview of the b-galactosidases used in the current work. All three b-galactosidases have an N-terminal His6-tag and were codonoptimized for expression in E. coli. Trivial name
Organism
Strain
GH family
UniProt entrya
Reference b
BgaD-D
Bacillus circulans
ATCC 31382
GH 42
E5RWQ2 truncated
[18]
Ttb-gly (TTP0042)
Thermus thermophilus
HB27
GH 1
Q746L1/Q9RA61
[12,13]
CelB
Pyrococcus furiosus
DSM 3638
GH 1
Q51723
[23]
a b
In addition to the amino acid sequence given in this entry, the enzymes used in the current work have an N-terminal His6-tag followed by a thrombin cleavage site. BgaD-D corresponds to amino acids 36-847 in this sequence [18].
2
www.elsevier.com/locate/nbt Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
NBT 850 1–6 RESEARCH PAPER
Research Paper
New Biotechnology Volume 00, Number 00 January 2016
FIGURE 1
Transgalactosylation reactions carried out with b-galactosidases BgaD-D, Ttb-gly, and CelB in the current work: (a) formation of LacNAc with GlcNAc as acceptor and lactose as donor, (b) formation of LNnT with LNT2 as acceptor and lactose as donor.
¨ KTA purifier (GE Healthequilibrated with binding buffer, using an A care, Uppsala, Sweden). Unbound material was washed off the column with 5 column volumes (CV) of binding buffer. The bgalactosidases were eluted with elution buffer (binding buffer with
500 mM imidazole) in a gradient from 0 to 100% elution buffer in 15 CV. Protein purity was confirmed by SDS–PAGE and the concentration determined by the Lambert–Beer law using baseline corrected absorption at 280 nm in combination with an extinction coefficient
FIGURE 2
Sequential reaction with b-N-acetylhexosaminidase and b-galactosidase to form LNnT from lactose and the chitin degradation product (GlcNAc)2. www.elsevier.com/locate/nbt 3 Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
NBT 850 1–6 RESEARCH PAPER
New Biotechnology Volume 00, Number 00 January 2016
of 175,000 M1 cm1 (BgaD-D), 104,000 M1 cm1 (Ttb-gly) and 131,000 M1 cm1 (CelB), respectively [19].
Hydrolytic activity
Research Paper
Hydrolytic activity of the three b-galactosidases was determined at 408C using 0.43 mM enzyme and 3.3 mM p-nitrophenyl b-galactopyranoside (pNP-b-Gal) in a 10 mM phosphate-citrate buffer (pH 6.0). The reaction was stopped at relevant time points over a course of 4 min by addition of Na2CO3. The release of pNP was measured ¨ dig, at 410 nm in an Infinite M200 microplate reader (Tecan, Gro Austria) and quantified against a pNP standard curve. 1 unit (U) equals the amount of enzyme required to release 1 mmol pNP per minute. All reactions were performed in triplicates.
Thermal stability Thermal stability of the three b-galactosidases was determined by incubating the enzymes at relevant temperatures (50–998C depending on the enzyme; see Table 2) for up to 60 min in a thermomixer (Eppendorf, Hamburg, Germany) at 500 rpm. At relevant time points up to 60 min, the thermal incubation was terminated by rapid cooling in an ice bath and the remaining hydrolytic activity was determined at the same conditions as described above. All reactions were performed in triplicates.
Transgalactosylation of N-acetylglucosamine and lacto-N-triose II For initial studies, GlcNAc was used as acceptor in order to form LacNAc (Fig. 1a). The three b-galactosidases were added in equimolar amounts (0.5 mM in the reaction) to a reaction mixture of 50 mM lactose and 500 mM GlcNAc in 27 mM phosphate-citrate buffer (pH 6.0). The reaction volume was 750 mL. The reaction took place at 408C as well as at the highest temperature at which the enzymes were completely stable for at least 1 hour, namely 508C for BgaD-D, 658C for Ttb-gly, and 908C for CelB. The reaction was stopped by immediate cooling on ice quickly followed by 3 min of centrifugation in a 5 kDa Vivaspin filter (Sartorius AG, ¨ ttingen, Germany) at 15,000 g to remove the enzyme from Go
the reaction mixture, since heat inactivation did not apply to the thermostable enzymes. The reaction was monitored for up to 2 hours by HPAEC-PAD analysis. For synthesis of LNnT (Fig. 1b), the reaction mixture contained 20 mM lactose and 100 mM LNT2 in a 27 mM phosphate-citrate buffer (pH 6.0). The reaction volume was 12.5 mL. Again, the three b-galactosidases were added in equimolar amounts (0.5 mM in the reaction). The reaction conditions were the optimal ones found in the initial studies: 30 min at 508C for BgaD-D, 30 min at 658C for Ttb-gly, and 10 min at 908C for CelB. The reactions were stopped as described above and the samples were analyzed by HPAEC-PAD.
Sequential formation of lacto-N-neotetraose from lactose and N,N0 -diacetylchitobiose The metagenomic b-N-acetylhexosaminidase HEX2 was produced as previously described [10]. It was then added to a reaction mixture of 100 mM (GlcNAc)2 and 500 mM lactose in 160 mM phosphate-citrate buffer (pH 6.0) and allowed to react for 2 hours at 258C. This reaction was stopped by heat inactivation at 958C for 3 min. After cooling to the reaction temperature of the following reaction (508C for BgaD-D, 658C for Ttb-gly, and 908C for CelB), one of the b-galactosidases was added (0.5 mM in the reaction) and the reaction took place at the set reaction temperature for 10 min (CelB) or 30 min (BgaD-D and Ttb-gly). The b-galactosidase reactions were stopped as described above. Samples were analyzed by HPAEC-PAD.
HPAEC-PAD analysis Separation and quantification of substrates and reaction products in the transgalactosylations were carried out by HPAEC-PAD analysis using the method described previously [10] for the transgalactosylation of LNT2 and a modified version with only 25 mM NaOH and an extended re-equilibration time for the transgalactosylation of GlcNAc. External standards of LacNAc, lacto-N-biose, allo-LacNAc, GlcNAc, and lactose were used for the analysis of the samples in transgalacto-sylation of GlcNAc, while external standards of LNnT, LNT, LNT2, and lactose were used for the analysis of the transgalactosylation of LNT2 to verify the identity of the compound produced. The amounts of LacNAc and LNnT produced were quantified by an external standard curve.
TABLE 2
Thermal stability of the three His6-tagged b-galactosidases expressed in E. coli. An asterisk (*) indicates that no thermal inactivation was detected (kD = 0); hence, no t½ could be calculated. Temp. [8C]
t½ [min] BgaD-D
50
*
55
273 291
60
28 3
65
2.4 0.4
70
0.6 0.2
Ttb-gly
CelB
*
80
280 117
90
205 42
*
95
11 1
*
99
1.8 0.8
155 30
4
Results & discussion Production of b-galactosidases All tagged b-galactosidases showed good expression levels when expressed recombinantly in the host E. coli BL21 (DE3) in combination with IPTG induction. The thermostable enzymes Ttb-gly and CelB were incubated at 808C for 20 min in order to precipitate non-thermostable enzymes prior to affinity chromatography without rendering the b-galactosidases inactive. This simple step was capable of removing a large fraction of impurities. The enzymes were then subjected to immobilized Ni2+ affinity chromatography to yield high purity. When assessing purity by SDS–PAGE, dimers and higher oligomers were observed for the thermostable enzymes when incubating samples at 998C in the presence of SDS.
Characterization of b-galactosidases The sequence identities between the three b-galactosidases were low. In particular, BgaD-D (GH 42) had only 7% sequence identity
www.elsevier.com/locate/nbt Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
NBT 850 1–6 New Biotechnology Volume 00, Number 00 January 2016
Transgalactosylation of N-acetylglucosamine N-Acetyllactosamine (LacNAc) is an important structural element in certain HMOs as well as a core structure in glycolipids and glycoproteins which are important in cellular recognition, including the Lewis blood group antigens sialyl-LewisX. It has been hypothesized that LacNAc analogues could be used in cancer treatment [16,22]. The LacNAc unit is also found at the nonreducing end of LNnT and as a consequence, the formation of LacNAc was studied in order to characterize the transgalactosylation potential of the three b-galactosidases BgaD-D, Ttb-gly, and CelB when using a GlcNAc unit as acceptor (Fig. 1a). The reactions were run at 408C, the temperature used by Murata and co-workers [9], and the yields were then compared to those obtained at the highest temperature where the enzymes were stable for at least 1 hour, that is, 508C for BgaD-D, 658C for Ttbgly, and 908C for CelB (Fig. 3). For all three b-galactosidases, increasing the temperature had a major effect on the LacNAc yield as well as on the time required to reach the maximum yield (Fig. 3). For BgaD-D, a LacNAc yield of 6.7% was reached after 1 hour of reaction at 408C, whereas it was 32% after 30 min at 508C. For Ttbgly, the yield increased from 2.7% after 1 hour at 408C to 16% after 30 min at 658C. For CelB, which has not previously been used for formation of LacNAc, the yield increased from 0.9% after 1 hour at 408C to 5.4% after 10 min at 908C (Fig. 3). Thus, BgaD-D was the most efficient b-galactosidase in the formation of LacNAc, reaching a LacNAc yield which was twice as high as that obtained with Ttb-gly even if the reaction temperature was 158C lower (Fig. 3). Importantly, the LacNAc yield obtained with BgaD-D was almost
40%
BgaD-D 40°C BgaD-D 50°C Ttß-gly 40°C Ttß-gly 65°C CelB 40°C CelB 90°C
Molar yield of LacNAc
35% 30% 25% 20% 15%
5% 0%
0
20
40
60 80 Reacon me [min.]
100
120
140
FIGURE 3
Formation of LacNAc by b-galactosidases BgaD-D, Ttb-gly, and CelB at 408C and at the highest temperature at which the enzymes show complete stability for extended reaction times (Table 2). Molar yield is based on the limiting donor substrate lactose; the acceptor substrate GlcNAc is used in 10fold excess. Error bars indicate standard deviation of duplicate reactions.
5-fold improved, when the reaction temperature was increased from 408C to 508C, and at the same time the optimal reaction time was halved. In all cases, maximum yields were reached within the first hour of reaction (Fig. 3).
Transgalactosylation of lacto-N-triose II Using the optimal reaction times established for each of the three b-galactosidases, that is, 10–30 min, at the highest temperature used in the formation of LacNAc (Fig. 3), LNT2 was used as acceptor in order to form LNnT when using lactose as donor (Fig. 1b). Since excess of acceptor substrate generally facilitates
8% 7% 6% 5% 4% 3% 2% 1% 0% 50°C 30 min
65°C 30 min
90°C 10 min
BgaD-D
Ttβ-gly
CelB
FIGURE 4
Molar yield of LNnT obtained with b-galactosidases BgaD-D, Ttb-gly, and CelB at the highest temperature at which the enzymes show complete stability for extended reaction times (Table 2) and the reaction times chosen as optimal from Fig. 3, that is, 10 min for CelB and 30 min for BgaD-D and Ttb-gly. Substrate concentrations were 100 mM LNT2 and 20 mM lactose. Thus, the more expensive substrate, LNT2, is used in 5-fold excess, that is, when calculating the molar yields on LNT2, the numbers are 5 times lower. Error bars indicate standard deviation of duplicate reactions.
www.elsevier.com/locate/nbt 5 Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
Research Paper
10%
Molar yield of LNnT
with Ttb-gly and 9% with CelB. The two thermostable GH 1 bgalactosidases had a sequence identity of 26% to each other (data not shown). When dosed on equimolar basis and assayed at 408C and pH 6.0, BgaD-D showed slightly higher hydrolytic activity (331 4 U/ mmol enzyme) than Ttb-gly (304 2 U/mmol enzyme) and CelB (207 3 U/mmol enzyme). However, Ttb-gly and especially CelB showed markedly higher thermal stability than BgaD-D (Table 2). Wishing to decrease the reaction time significantly from the extended ones used previously and realizing that hydrolase-catalyzed transglycosylation often reach the maximum yield within the first few hours of reaction [20], the thermal stability of the enzymes was studied for up to 1 hour. BgaD-D was completely stable for at least 1 hour at 508C. At 558C, the t½ of BgaD-D was 4.5 hours, but then the stability decreased rapidly with increasing temperature, and was less than 1 min at 708C (Table 2). This result is in line with observations made on BgaD-D isolated from the Biolacta1 preparation [21]. In contrast, Ttb-gly was completely stable for more than 1 hour at 658C. The t½ was approx. 4.5 hours at 808C, and then decreased with increasing temperature to approx. 2 min at 998C (Table 2). The recombinant Ttb-gly used in the current work was thus more thermostable than previously observed: Dion and co-workers reported a half-life of 10 min at 908C for Ttb-gly expressed in E. coli [12]. CelB was the more thermostable of the three b-galactosidases tested: It was stable for at least 1 hour at 908C and 958C, and had a t½ of more than 2.5 hours at 998C (Table 2). The native CelB expressed in Pyrococcus furiosus was reported previously to be even more stable, having a half-life of approx. 82 hours at 1008C [14].
RESEARCH PAPER
NBT 850 1–6 RESEARCH PAPER
Research Paper
glycosidase-catalyzed transglycosylation [20], a 5-fold excess of LNT2 was used in the reaction to favor transgalactosylation. Again, BgaD-D gave the highest yield of LNnT (Fig. 4). BgaD-D gave a molar yield of LNnT of 7.1% followed by Ttb-gly with a yield of 5.2% and CelB with a yield of 1.0% (Fig. 4). Calculating the molar yield on the more expensive LNT2 acceptor, the yields of LNnT were 1.4%, 1.0%, and 0.2%, respectively. Thus, the relative efficiency of the three b-galactosidases in the transgalactosylation of a GlcNAc unit was similar in the formation of LacNAc and LNnT, although LacNAc yields were 3–5 times higher than LNnT yields. Murata and co-workers [9] used equimolar amounts of LNT2 acceptor and lactose donor in their Biolacta1-catalyzed formation of LNnT reaching a molar yield of 19% after 36 hours of reaction at 408C. In conclusion, the two thermostable b-galactosidases from the GH family 1 did not improve the yield LacNAc or LNnT even if they could operate at higher temperatures. The GH42 b-galactosidase from B. circulans is already used for glycan synthesis due to its superior transgalactosylation properties. The data obtained establish that elevated yields are obtained with a reaction temperature of 508C rather than 408C. The sequential reaction where the metagenomic b-N-acetylhexosaminidase was used to catalyze the production of LNT2 from lactose and the chitin degradation product (GlcNAc)2 (Fig. 2), was run in two steps rather than one-pot in order to maximize the yield of LNT2 by avoiding lactose hydrolysis. The concentration of LNT2 in the reaction mixture reached 8 mM after the b-N-acetylhexosaminidase reaction. This corresponds to only 8% of the LNT2 present in the reaction where chemically synthesized LNT2 was used (Fig. 4), and was not sufficient to drive
New Biotechnology Volume 00, Number 00 January 2016
the transgalactosylation of LNT2 into LNnT forward: no LNnT was detected in the samples (data not shown). With 5-fold excess lactose being used for the b-N-acetylhexosaminidase reaction, the acceptor:donor ratio was approx. 1:60, whereas it was 5 in the reaction with chemically synthesized LNT2 (Fig. 4). In conclusion, higher concentrations of LNT2 need to be reached in the first part of the reaction in order to drive the second part of the reaction forward; that is, the b-N-acetylhexosaminidase step should be optimized further before such a sequential reaction to form LNnT can be realized.
Conclusions The three b-galactosidases all catalyzed the formation of LacNAc from GlcNAc and lactose as well as the formation of LNnT from LNT2 and lactose. The studies of thermostability and optimal temperature emphasized the importance of keeping the reaction temperature as high as possible without compromising enzyme stability. In all cases, BgaD-D was the most potent transgalactosidase. However, if a higher working temperature is desired, especially Ttb-gly is a good choice. The study also showed that b-galactosidases of very low similarity (low sequence identity and different GH families) could catalyze the same transgalactosylation reactions.
Acknowledgements We thank ‘The Strategic Research Council’ (DSF) for the financial support to the strategic project ‘Enzymatic production of human milk oligosaccharides’.
References [1] Bode L. Human milk oligosaccharides: every baby needs a sugar mama. Glycobiology 2012;22:1147–62. ¨ ller-Werner B, Jelinek J, et al. [2] Thurl S, Munzert M, Henker J, Boehm G, Mu Variation of human milk oligosaccharides in relation to milk groups and lactational periods. Br J Nutr 2010;104:1261–71. [3] Takamura T, Chiba T, Ishihara H, Tejima S. Chemical modification of lactose. XIII. Synthesis of lacto-N-tetraose. Chem Pharm Bill 1980;28:1804–9. [4] Aly MRE, Ibrahim ESI, El Ashry ESH, Schmidt RR. Synthesis of lacto-N-neotetraose and lacto-N-tetraose using the dimethylmaleoyl group as amino protective group. Carbohydr Res 1999;316:121–32. ´ goston K, Pe´rez IF, Boutet J, Hederos M, et al. A method for [5] Bajza I, Dekany G, A preparation of the tetrasaccharide lacto-N-neotetraose (LNnT) containing Nacetyllactosamine. Patent WO2011/100980, August 2011. [6] Hederos M, Dekany G, Demko´ S, Kova´cs I, Bajza I. Manufacture of lacto-Ntetraose. Patent WO2012/155916, November 2012. [7] Baumga¨rtner F, Conrad J, Sprenger GA, Albermann C. Synthesis of the human milk oligosaccharide lacto-N-tetraose in metabolically engineered, plasmid-free E. coli. ChemBioChem 2014;15:1896–900. [8] Johnson KF. Synthesis of oligosaccharides by bacterial enzymes. Glycoconjugate J 1999;16:141–6. [9] Murata T, Inukai T, Suzuki M, Yamagishi M, Usui T. Facile enzymatic conversion of lactose into lacto-N-tetraose and lacto-N-neotetraose. Glycoconjugate J 1999;16:189–95. [10] Nyffenegger C, Nordvang RT, Zeuner B, Lezyk M, Difilippo E, Logtenberg MJ, et al. Backbone structures in human milk oligosaccharides: trans-glycosylation by metagenomic b-N-acetylhexosaminidases. Appl Microbiol Biotechnol 2015; 99:7997–8009. [11] Matahira Y, Tashiro A, Sato T, Kawagishi H, Usui T. Enzymic synthesis of lacto-N-triose II and its potential analogues. Glycoconjugate J 1995;12: 664–71. [12] Dion M, Fourage L, Hallet J-N, Colas B. Cloning and expression of a b-glycosidase gene from Thermus thermophilus. Sequence and biochemical characterization of the encoded enzyme. Glycoconjugate J 1999;16:27–37.
6
[13] Sandoval M, Ferreras E, Pe´rez-Sa´nchez M, Berenguer J, Sinisterra JV, Hernaiz MJ. Screening of strains and recombinant enzymes from Thermus thermophilus for their use in disaccharide synthesis. J Mol Catal B: Enzym 2012;74:162–9. [14] Kengen SWM, Luesink EJ, Stams AJM, Zehnder AJB. Purification and characterization of an extremely thermostable b-glucosidase from the hyperthermophilic archaeon Pyrococcus furiosus. Eur J Biochem 1993;213:305–12. [15] Hansson T, Adlercreutz P. Optimization of galactooligosaccharide production from lactose using b-glycosidases from hyperthermophiles. Food Biotechnol 2001;15:79–97. [16] Bridiau N, Maugard T. A comparative study of the regioselectivity of the b-galactosidases from Kluyveromyces lactis and Bacillus circulans in the enzymatic synthesis of N-acetyl-lactosamine in aqueous media. Biotechnol Prog 2011;27:386–94. [17] Warmerdam A, Paudel E, Jia W, Boom RM, Janssen AEM. Characterization of bgalactosidase isoforms from Bacillus circulans and their contribution to GOS production. Appl Biochem Biotechnol 2013;170:340–58. [18] Song J, Imanaka H, Imamura K, Minoda M, Katase T, Hoshi Y, et al. Cloning and expression of a b-galactosidase gene of Bacillus circulans. Biosci Biotechnol Biochem 2011;75:1194–7. [19] Wilkins MR, Gasteiger E, Bairoch A, Sanchez JC, Williams KL, Appel RD, et al. Protein identification and analysis tools in the ExPASy server. Methods Mol Biol 1999;112:531–52. [20] Zeuner B, Jers C, Mikkelsen JD, Meyer AS. Methods for improving enzymatic trans-glycosylation for synthesis of human milk oligosaccharide biomimetics. J Agric Food Chem 2014;62:9615–31. [21] Song J, Abe K, Imanaka H, Imamura K, Minoda M, Yamaguchi S, et al. Causes of the production of multiple forms of b-galactosidase by Bacillus circulans. Biosci Biotechnol Biochem 2011;75:268–78. [22] Fukuda M. Cell surface glycoconjugates as onco-differentiation markers in hematopoietic cells. Biochim Biophys Acta 1985;780:119–50. [23] Voorhorst WGB, Eggen RIL, Luesink EJ, de Vos WM. Characterization of the celB gene coding for b-glucosidase from the hyperthermophilic archaeon Pyrococcus furiosus and its expression and site-directed mutation in Escherichia coli. J Bacteriol 1995;177:7105–11.
www.elsevier.com/locate/nbt Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
RESEARCH PAPER
Research Paper
Thermostable b-galactosidases for the synthesis of human milk oligosaccharides Birgitte Zeuner1, Christian Nyffenegger1, Jørn Dalgaard Mikkelsen and Anne S. Meyer Q1 Center for BioProcess Engineering, Department of Chemical and Biochemical Engineering, Technical University of Denmark, Building 229, DK-2800 Kgs., Lyngby, Denmark
Human milk oligosaccharides (HMOs) designate a unique family of bioactive lactose-based molecules present in human breast milk. Using lactose as a cheap donor, some b-galactosidases (EC 3.2.1.23) can catalyze transgalactosylation to form the human milk oligosaccharide lacto-N-neotetraose (LNnT; Galb(1,4)-GlcNAc-b(1,3)-Gal-b(1,4)-Glc). In order to reduce reaction times and be able to work at temperatures, which are less welcoming to microbial growth, the current study investigates the possibility of using thermostable b-galactosidases for synthesis of LNnT and N-acetyllactosamine (LacNAc; Gal-b(1,4)-GlcNAc), the latter being a core structure in HMOs. Two hyperthermostable GH 1 b-galactosidases, Ttb-gly from Thermus thermophilus HB27 and CelB from Pyrococcus furiosus, were codon-optimized for expression in Escherichia coli along with BgaD-D, a truncated version of the GH 42 b-galactosidase from Bacillus circulans showing high transgalactosylation activity at low substrate concentrations. The three b-galactosidases were compared in the current study in terms of their transgalactosylation activity in the formation of LacNAc and LNnT. In all cases, BgaD-D was the most potent transgalactosidase, but both thermostable GH 1 b-galactosidases could catalyze formation of LNnT and LacNAc, with Ttb-gly giving higher yields than CelB. The thermal stability of the three b-galactosidases was elucidated and the results were used to optimize the reaction efficiency in the formation of LacNAc, resulting in 5–6 times higher reaction yields and significantly shorter reaction times.
Introduction Q2 Human milk oligosaccharides (HMOs), which are abundant in human breast milk (present at levels up to 8 g/L) are known to be of major importance for infant health of development [1]. The HMOs share similar structural features: They are all based on a reducing end lactose molecule, which can be substituted with sialic acid or fucose and elongated with N-acetylglucosamine (GlcNAc) and galactose (Gal) units; these elongated molecules may in turn also be substituted with sialyl or fucosyl residues [1,2]. Important elongated HMO core structures include
Corresponding author: Meyer, A.S. ([email protected]) 1
These two authors contributed equally.
http://dx.doi.org/10.1016/j.nbt.2016.01.003 1871-6784/ß 2016 Published by Elsevier B.V.
lacto-N-tetraose (LNT; Gal-b(1,3)-GlcNAc-b(1,3)-Gal-b(1,4)-Glc) and lacto-N-neotetraose (LNnT; Gal-b(1,4)-GlcNAc-b(1,3)-Galb(1,4)-Glc) [2]. Approaches to synthesize LNT and LNnT include chemical synthesis [3–6], in vivo production in Escherichia coli [7], the use of glycosyl transferases requiring nucleotide sugars [8], as well as the use of glycosidases with transglycosylation activity [9]. More than a decade ago, it was shown that lacto-N-triose II (LNT2; GlcNAc-b(1,3)-Gal-b(1,4)-Glc) could be elongated to LNnT by transgalactosylation catalyzed by the GH 42 b-galactosidase from Bacillus circulans using lactose as donor substrate [9]. Using the GH 35 b-galactosidase from Bacillus circulans and p-nitrophenyl b-galactopyranoside (pNP-b-Gal) as donor, LNT (and LNT isomers, most likely LNnT) was also reported to be formed from www.elsevier.com/locate/nbt
Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
1
NBT 850 1–6 RESEARCH PAPER
New Biotechnology Volume 00, Number 00 January 2016
Research Paper
LNT2, but the reaction was not performed with lactose as donor substrate [9]. Similarly, the use of lacto-N-biosidase (EC 3.2.1.140) from Aureobacterium sp. L-101 which could catalyze the transfer of the lacto-N-biose disaccharide to lactose to form LNT also required pNP activation of the donor substrate [9]. In the quest for finding routes to HMO production which involve cheap substrates only, we have recently shown that LNT2 can be produced from chitin degradation products and lactose by two novel b-N-acetylhexosaminidases at molar yields of 2–8% [10]. Previously, LNT2 synthesis has relied on nucleotide, methyl, or pNP-activated donors or chemical synthesis [9,11], all keeping the price of LNT2 too high for a feasible large-scale production. With the possibility of using chitin degradation products as donor substrates and achieving appreciable yields, the b-galactosidase-catalyzed route to production of LNnT with lactose as a cheap donor substrate is becoming increasingly interesting from an economic point of view. In the only published report on b-galactosidase-catalyzed formation of LNnT from lactose and LNT2, Murata and co-workers [9] reached a molar yield based on LNT2 of 19% by performing the reaction for 36 hours at 408C. Such reaction conditions are however problematic with regard to microbial safety. In order to speed up the reaction and be able to work at a temperature, which is less welcoming to microbial growth, the current study investigates the possibility of using thermostable b-galactosidases for LNnT synthesis. A b-glycosidase from Thermus thermophilus HB27, known as Ttb-gly or TTP0042, has high thermal stability [12,13] and has been shown to catalyze the transgalactosylation of GlcNAc to form N-acetyllactosamine (Gal-b(1,4)-GlcNAc; LacNAc) using pNP-bGal as donor [13]. b-Glycosidase from Pyrococcus furiosus, known as CelB, has even higher thermal stability than Ttb-gly [14]. This b-galactosidase has not previously been used for transgalactosylation of GlcNAc moieties, but has been shown to efficiently catalyze the formation of galactooligosaccharides at 958C [15]. Thus, the current work was also undertaken with the objective of comparing these two thermostable b-galactosidases to the benchmark enzyme used for synthesis of LNnT as well as for LacNAc, the GH42 b-galactosidase from Bacillus circulans which is found in the popular commercial Biolacta1 preparation [9,16]. The Biolacta1 preparation contains several different isozymes, but the one used in the current study (BgaD-D) is the one having the highest transgalactosylation activity at low substrate concentrations [17,18]. Ttb-gly, CelB, and BgaD-D (Table 1) were codon-optimized for expression in E. coli, His6-tagged for purification purposes, and compared in the current study in terms of their transgalactosylation activity in the formation of LacNAc and LNnT at elevated temperatures (Fig. 1). In
addition, a sequential reaction with b-N-acetylhexosaminidase and b-galactosidase to form LNnT from lactose and the chitin degradation product N,N0 -diacetylchitobiose ((GlcNAc)2; GlcNAc-b(1,4)GlcNAc) was tested (Fig. 2).
Materials and methods Chemicals Lacto-N-triose II (GlcNAc-b(1,3)-Gal-b(1,4)-Glc; LNT2), lacto-Nbiose (Gal-b(1,3)-GlcNAc), lacto-N-tetraose (Gal-b(1,3)-GlcNAcb(1,3)-Gal-b(1,4)-Glc; LNT), lacto-N-neotetraose (Gal-b(1,4)GlcNAc-b(1,3)-Gal-b(1,4)-Glc; LNnT), and N,N0 -diacetylchitobiose (GlcNAc-b(1,4)-GlcNAc; (GlcNAc)2) were purchased from Carbosynth Ltd. (Compton, United Kingdom). For analytical confirmation, LNT and LNnT were also purchased from Elicityl SA (Crolles, France). N-Acetylglucosamine (GlcNAc), N-acetyllactosamine (Gal-b(1,4)-GlcNAc; LacNAc), allo-N-acetyllactosamine (Galb(1,6)-GlcNAc; allo-LacNAc), p-nitrophenyl b-galactopyranoside (pNP-b-Gal), b-lactose, and all other chemicals were purchased from Sigma–Aldrich (Steinheim, Germany).
Cloning, expression, and purification of b-galactosidases Genes encoding the three b-galactosidases BgaD-D, Ttb-gly, and CelB, each possessing an N-terminal His6-tag linked via a thrombin recognition site, were codon-optimized for expression in Escherichia coli (E. coli), synthesized and inserted into the vector pJ411 by DNA2.0 (Menlo Park, CA, USA). E. coli BL21 (DE3) was transformed with the resulting plasmids and selected for kanamycin resistance. BgaD-D was expressed as follows: transformants were grown in lysogenic broth (LB) supplemented with 50 mg/ml kanamycin at 378C. When reaching an OD600 of 0.6, the temperature was reduced to 258C and expression induced by the addition of IPTG to a final concentration of 1 mM. Cells were grown for 15 hours post induction before harvest. Ttb-gly and CelB were expressed in a similar manner, except that the cells were allowed to grow for 5 hours after induction before being harvested. Cells were centrifuged and the pellets resuspended in binding buffer (BgaD-D: 20 mM Na-phosphate buffer, 500 mM NaCl, 20 mM imidazole, pH 7; Ttb-gly and CelB: 40 mM EPPS buffer, 500 mM NaCl, 24 mM imidazole), followed by sonication to open the cells and centrifugation to remove cell debris. The supernatant containing CelB or Ttb-gly was pre-purified by incubation for 20 min at 808C resulting in precipitation of non-thermostable proteins, followed by another centrifugation step. The supernatant containing the b-galactosidases was passed through a 0.45 mm filter before being loaded onto a 1 ml Ni2+sepharose HisTrap HP column (GE Healthcare, Uppsala, Sweden),
TABLE 1
An overview of the b-galactosidases used in the current work. All three b-galactosidases have an N-terminal His6-tag and were codonoptimized for expression in E. coli. Trivial name
Organism
Strain
GH family
UniProt entrya
Reference b
BgaD-D
Bacillus circulans
ATCC 31382
GH 42
E5RWQ2 truncated
[18]
Ttb-gly (TTP0042)
Thermus thermophilus
HB27
GH 1
Q746L1/Q9RA61
[12,13]
CelB
Pyrococcus furiosus
DSM 3638
GH 1
Q51723
[23]
a b
In addition to the amino acid sequence given in this entry, the enzymes used in the current work have an N-terminal His6-tag followed by a thrombin cleavage site. BgaD-D corresponds to amino acids 36-847 in this sequence [18].
2
www.elsevier.com/locate/nbt Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
NBT 850 1–6 RESEARCH PAPER
Research Paper
New Biotechnology Volume 00, Number 00 January 2016
FIGURE 1
Transgalactosylation reactions carried out with b-galactosidases BgaD-D, Ttb-gly, and CelB in the current work: (a) formation of LacNAc with GlcNAc as acceptor and lactose as donor, (b) formation of LNnT with LNT2 as acceptor and lactose as donor.
¨ KTA purifier (GE Healthequilibrated with binding buffer, using an A care, Uppsala, Sweden). Unbound material was washed off the column with 5 column volumes (CV) of binding buffer. The bgalactosidases were eluted with elution buffer (binding buffer with
500 mM imidazole) in a gradient from 0 to 100% elution buffer in 15 CV. Protein purity was confirmed by SDS–PAGE and the concentration determined by the Lambert–Beer law using baseline corrected absorption at 280 nm in combination with an extinction coefficient
FIGURE 2
Sequential reaction with b-N-acetylhexosaminidase and b-galactosidase to form LNnT from lactose and the chitin degradation product (GlcNAc)2. www.elsevier.com/locate/nbt 3 Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
NBT 850 1–6 RESEARCH PAPER
New Biotechnology Volume 00, Number 00 January 2016
of 175,000 M1 cm1 (BgaD-D), 104,000 M1 cm1 (Ttb-gly) and 131,000 M1 cm1 (CelB), respectively [19].
Hydrolytic activity
Research Paper
Hydrolytic activity of the three b-galactosidases was determined at 408C using 0.43 mM enzyme and 3.3 mM p-nitrophenyl b-galactopyranoside (pNP-b-Gal) in a 10 mM phosphate-citrate buffer (pH 6.0). The reaction was stopped at relevant time points over a course of 4 min by addition of Na2CO3. The release of pNP was measured ¨ dig, at 410 nm in an Infinite M200 microplate reader (Tecan, Gro Austria) and quantified against a pNP standard curve. 1 unit (U) equals the amount of enzyme required to release 1 mmol pNP per minute. All reactions were performed in triplicates.
Thermal stability Thermal stability of the three b-galactosidases was determined by incubating the enzymes at relevant temperatures (50–998C depending on the enzyme; see Table 2) for up to 60 min in a thermomixer (Eppendorf, Hamburg, Germany) at 500 rpm. At relevant time points up to 60 min, the thermal incubation was terminated by rapid cooling in an ice bath and the remaining hydrolytic activity was determined at the same conditions as described above. All reactions were performed in triplicates.
Transgalactosylation of N-acetylglucosamine and lacto-N-triose II For initial studies, GlcNAc was used as acceptor in order to form LacNAc (Fig. 1a). The three b-galactosidases were added in equimolar amounts (0.5 mM in the reaction) to a reaction mixture of 50 mM lactose and 500 mM GlcNAc in 27 mM phosphate-citrate buffer (pH 6.0). The reaction volume was 750 mL. The reaction took place at 408C as well as at the highest temperature at which the enzymes were completely stable for at least 1 hour, namely 508C for BgaD-D, 658C for Ttb-gly, and 908C for CelB. The reaction was stopped by immediate cooling on ice quickly followed by 3 min of centrifugation in a 5 kDa Vivaspin filter (Sartorius AG, ¨ ttingen, Germany) at 15,000 g to remove the enzyme from Go
the reaction mixture, since heat inactivation did not apply to the thermostable enzymes. The reaction was monitored for up to 2 hours by HPAEC-PAD analysis. For synthesis of LNnT (Fig. 1b), the reaction mixture contained 20 mM lactose and 100 mM LNT2 in a 27 mM phosphate-citrate buffer (pH 6.0). The reaction volume was 12.5 mL. Again, the three b-galactosidases were added in equimolar amounts (0.5 mM in the reaction). The reaction conditions were the optimal ones found in the initial studies: 30 min at 508C for BgaD-D, 30 min at 658C for Ttb-gly, and 10 min at 908C for CelB. The reactions were stopped as described above and the samples were analyzed by HPAEC-PAD.
Sequential formation of lacto-N-neotetraose from lactose and N,N0 -diacetylchitobiose The metagenomic b-N-acetylhexosaminidase HEX2 was produced as previously described [10]. It was then added to a reaction mixture of 100 mM (GlcNAc)2 and 500 mM lactose in 160 mM phosphate-citrate buffer (pH 6.0) and allowed to react for 2 hours at 258C. This reaction was stopped by heat inactivation at 958C for 3 min. After cooling to the reaction temperature of the following reaction (508C for BgaD-D, 658C for Ttb-gly, and 908C for CelB), one of the b-galactosidases was added (0.5 mM in the reaction) and the reaction took place at the set reaction temperature for 10 min (CelB) or 30 min (BgaD-D and Ttb-gly). The b-galactosidase reactions were stopped as described above. Samples were analyzed by HPAEC-PAD.
HPAEC-PAD analysis Separation and quantification of substrates and reaction products in the transgalactosylations were carried out by HPAEC-PAD analysis using the method described previously [10] for the transgalactosylation of LNT2 and a modified version with only 25 mM NaOH and an extended re-equilibration time for the transgalactosylation of GlcNAc. External standards of LacNAc, lacto-N-biose, allo-LacNAc, GlcNAc, and lactose were used for the analysis of the samples in transgalacto-sylation of GlcNAc, while external standards of LNnT, LNT, LNT2, and lactose were used for the analysis of the transgalactosylation of LNT2 to verify the identity of the compound produced. The amounts of LacNAc and LNnT produced were quantified by an external standard curve.
TABLE 2
Thermal stability of the three His6-tagged b-galactosidases expressed in E. coli. An asterisk (*) indicates that no thermal inactivation was detected (kD = 0); hence, no t½ could be calculated. Temp. [8C]
t½ [min] BgaD-D
50
*
55
273 291
60
28 3
65
2.4 0.4
70
0.6 0.2
Ttb-gly
CelB
*
80
280 117
90
205 42
*
95
11 1
*
99
1.8 0.8
155 30
4
Results & discussion Production of b-galactosidases All tagged b-galactosidases showed good expression levels when expressed recombinantly in the host E. coli BL21 (DE3) in combination with IPTG induction. The thermostable enzymes Ttb-gly and CelB were incubated at 808C for 20 min in order to precipitate non-thermostable enzymes prior to affinity chromatography without rendering the b-galactosidases inactive. This simple step was capable of removing a large fraction of impurities. The enzymes were then subjected to immobilized Ni2+ affinity chromatography to yield high purity. When assessing purity by SDS–PAGE, dimers and higher oligomers were observed for the thermostable enzymes when incubating samples at 998C in the presence of SDS.
Characterization of b-galactosidases The sequence identities between the three b-galactosidases were low. In particular, BgaD-D (GH 42) had only 7% sequence identity
www.elsevier.com/locate/nbt Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
NBT 850 1–6 New Biotechnology Volume 00, Number 00 January 2016
Transgalactosylation of N-acetylglucosamine N-Acetyllactosamine (LacNAc) is an important structural element in certain HMOs as well as a core structure in glycolipids and glycoproteins which are important in cellular recognition, including the Lewis blood group antigens sialyl-LewisX. It has been hypothesized that LacNAc analogues could be used in cancer treatment [16,22]. The LacNAc unit is also found at the nonreducing end of LNnT and as a consequence, the formation of LacNAc was studied in order to characterize the transgalactosylation potential of the three b-galactosidases BgaD-D, Ttb-gly, and CelB when using a GlcNAc unit as acceptor (Fig. 1a). The reactions were run at 408C, the temperature used by Murata and co-workers [9], and the yields were then compared to those obtained at the highest temperature where the enzymes were stable for at least 1 hour, that is, 508C for BgaD-D, 658C for Ttbgly, and 908C for CelB (Fig. 3). For all three b-galactosidases, increasing the temperature had a major effect on the LacNAc yield as well as on the time required to reach the maximum yield (Fig. 3). For BgaD-D, a LacNAc yield of 6.7% was reached after 1 hour of reaction at 408C, whereas it was 32% after 30 min at 508C. For Ttbgly, the yield increased from 2.7% after 1 hour at 408C to 16% after 30 min at 658C. For CelB, which has not previously been used for formation of LacNAc, the yield increased from 0.9% after 1 hour at 408C to 5.4% after 10 min at 908C (Fig. 3). Thus, BgaD-D was the most efficient b-galactosidase in the formation of LacNAc, reaching a LacNAc yield which was twice as high as that obtained with Ttb-gly even if the reaction temperature was 158C lower (Fig. 3). Importantly, the LacNAc yield obtained with BgaD-D was almost
40%
BgaD-D 40°C BgaD-D 50°C Ttß-gly 40°C Ttß-gly 65°C CelB 40°C CelB 90°C
Molar yield of LacNAc
35% 30% 25% 20% 15%
5% 0%
0
20
40
60 80 Reacon me [min.]
100
120
140
FIGURE 3
Formation of LacNAc by b-galactosidases BgaD-D, Ttb-gly, and CelB at 408C and at the highest temperature at which the enzymes show complete stability for extended reaction times (Table 2). Molar yield is based on the limiting donor substrate lactose; the acceptor substrate GlcNAc is used in 10fold excess. Error bars indicate standard deviation of duplicate reactions.
5-fold improved, when the reaction temperature was increased from 408C to 508C, and at the same time the optimal reaction time was halved. In all cases, maximum yields were reached within the first hour of reaction (Fig. 3).
Transgalactosylation of lacto-N-triose II Using the optimal reaction times established for each of the three b-galactosidases, that is, 10–30 min, at the highest temperature used in the formation of LacNAc (Fig. 3), LNT2 was used as acceptor in order to form LNnT when using lactose as donor (Fig. 1b). Since excess of acceptor substrate generally facilitates
8% 7% 6% 5% 4% 3% 2% 1% 0% 50°C 30 min
65°C 30 min
90°C 10 min
BgaD-D
Ttβ-gly
CelB
FIGURE 4
Molar yield of LNnT obtained with b-galactosidases BgaD-D, Ttb-gly, and CelB at the highest temperature at which the enzymes show complete stability for extended reaction times (Table 2) and the reaction times chosen as optimal from Fig. 3, that is, 10 min for CelB and 30 min for BgaD-D and Ttb-gly. Substrate concentrations were 100 mM LNT2 and 20 mM lactose. Thus, the more expensive substrate, LNT2, is used in 5-fold excess, that is, when calculating the molar yields on LNT2, the numbers are 5 times lower. Error bars indicate standard deviation of duplicate reactions.
www.elsevier.com/locate/nbt 5 Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
Research Paper
10%
Molar yield of LNnT
with Ttb-gly and 9% with CelB. The two thermostable GH 1 bgalactosidases had a sequence identity of 26% to each other (data not shown). When dosed on equimolar basis and assayed at 408C and pH 6.0, BgaD-D showed slightly higher hydrolytic activity (331 4 U/ mmol enzyme) than Ttb-gly (304 2 U/mmol enzyme) and CelB (207 3 U/mmol enzyme). However, Ttb-gly and especially CelB showed markedly higher thermal stability than BgaD-D (Table 2). Wishing to decrease the reaction time significantly from the extended ones used previously and realizing that hydrolase-catalyzed transglycosylation often reach the maximum yield within the first few hours of reaction [20], the thermal stability of the enzymes was studied for up to 1 hour. BgaD-D was completely stable for at least 1 hour at 508C. At 558C, the t½ of BgaD-D was 4.5 hours, but then the stability decreased rapidly with increasing temperature, and was less than 1 min at 708C (Table 2). This result is in line with observations made on BgaD-D isolated from the Biolacta1 preparation [21]. In contrast, Ttb-gly was completely stable for more than 1 hour at 658C. The t½ was approx. 4.5 hours at 808C, and then decreased with increasing temperature to approx. 2 min at 998C (Table 2). The recombinant Ttb-gly used in the current work was thus more thermostable than previously observed: Dion and co-workers reported a half-life of 10 min at 908C for Ttb-gly expressed in E. coli [12]. CelB was the more thermostable of the three b-galactosidases tested: It was stable for at least 1 hour at 908C and 958C, and had a t½ of more than 2.5 hours at 998C (Table 2). The native CelB expressed in Pyrococcus furiosus was reported previously to be even more stable, having a half-life of approx. 82 hours at 1008C [14].
RESEARCH PAPER
NBT 850 1–6 RESEARCH PAPER
Research Paper
glycosidase-catalyzed transglycosylation [20], a 5-fold excess of LNT2 was used in the reaction to favor transgalactosylation. Again, BgaD-D gave the highest yield of LNnT (Fig. 4). BgaD-D gave a molar yield of LNnT of 7.1% followed by Ttb-gly with a yield of 5.2% and CelB with a yield of 1.0% (Fig. 4). Calculating the molar yield on the more expensive LNT2 acceptor, the yields of LNnT were 1.4%, 1.0%, and 0.2%, respectively. Thus, the relative efficiency of the three b-galactosidases in the transgalactosylation of a GlcNAc unit was similar in the formation of LacNAc and LNnT, although LacNAc yields were 3–5 times higher than LNnT yields. Murata and co-workers [9] used equimolar amounts of LNT2 acceptor and lactose donor in their Biolacta1-catalyzed formation of LNnT reaching a molar yield of 19% after 36 hours of reaction at 408C. In conclusion, the two thermostable b-galactosidases from the GH family 1 did not improve the yield LacNAc or LNnT even if they could operate at higher temperatures. The GH42 b-galactosidase from B. circulans is already used for glycan synthesis due to its superior transgalactosylation properties. The data obtained establish that elevated yields are obtained with a reaction temperature of 508C rather than 408C. The sequential reaction where the metagenomic b-N-acetylhexosaminidase was used to catalyze the production of LNT2 from lactose and the chitin degradation product (GlcNAc)2 (Fig. 2), was run in two steps rather than one-pot in order to maximize the yield of LNT2 by avoiding lactose hydrolysis. The concentration of LNT2 in the reaction mixture reached 8 mM after the b-N-acetylhexosaminidase reaction. This corresponds to only 8% of the LNT2 present in the reaction where chemically synthesized LNT2 was used (Fig. 4), and was not sufficient to drive
New Biotechnology Volume 00, Number 00 January 2016
the transgalactosylation of LNT2 into LNnT forward: no LNnT was detected in the samples (data not shown). With 5-fold excess lactose being used for the b-N-acetylhexosaminidase reaction, the acceptor:donor ratio was approx. 1:60, whereas it was 5 in the reaction with chemically synthesized LNT2 (Fig. 4). In conclusion, higher concentrations of LNT2 need to be reached in the first part of the reaction in order to drive the second part of the reaction forward; that is, the b-N-acetylhexosaminidase step should be optimized further before such a sequential reaction to form LNnT can be realized.
Conclusions The three b-galactosidases all catalyzed the formation of LacNAc from GlcNAc and lactose as well as the formation of LNnT from LNT2 and lactose. The studies of thermostability and optimal temperature emphasized the importance of keeping the reaction temperature as high as possible without compromising enzyme stability. In all cases, BgaD-D was the most potent transgalactosidase. However, if a higher working temperature is desired, especially Ttb-gly is a good choice. The study also showed that b-galactosidases of very low similarity (low sequence identity and different GH families) could catalyze the same transgalactosylation reactions.
Acknowledgements We thank ‘The Strategic Research Council’ (DSF) for the financial support to the strategic project ‘Enzymatic production of human milk oligosaccharides’.
References [1] Bode L. Human milk oligosaccharides: every baby needs a sugar mama. Glycobiology 2012;22:1147–62. ¨ ller-Werner B, Jelinek J, et al. [2] Thurl S, Munzert M, Henker J, Boehm G, Mu Variation of human milk oligosaccharides in relation to milk groups and lactational periods. Br J Nutr 2010;104:1261–71. [3] Takamura T, Chiba T, Ishihara H, Tejima S. Chemical modification of lactose. XIII. Synthesis of lacto-N-tetraose. Chem Pharm Bill 1980;28:1804–9. [4] Aly MRE, Ibrahim ESI, El Ashry ESH, Schmidt RR. Synthesis of lacto-N-neotetraose and lacto-N-tetraose using the dimethylmaleoyl group as amino protective group. Carbohydr Res 1999;316:121–32. ´ goston K, Pe´rez IF, Boutet J, Hederos M, et al. A method for [5] Bajza I, Dekany G, A preparation of the tetrasaccharide lacto-N-neotetraose (LNnT) containing Nacetyllactosamine. Patent WO2011/100980, August 2011. [6] Hederos M, Dekany G, Demko´ S, Kova´cs I, Bajza I. Manufacture of lacto-Ntetraose. Patent WO2012/155916, November 2012. [7] Baumga¨rtner F, Conrad J, Sprenger GA, Albermann C. Synthesis of the human milk oligosaccharide lacto-N-tetraose in metabolically engineered, plasmid-free E. coli. ChemBioChem 2014;15:1896–900. [8] Johnson KF. Synthesis of oligosaccharides by bacterial enzymes. Glycoconjugate J 1999;16:141–6. [9] Murata T, Inukai T, Suzuki M, Yamagishi M, Usui T. Facile enzymatic conversion of lactose into lacto-N-tetraose and lacto-N-neotetraose. Glycoconjugate J 1999;16:189–95. [10] Nyffenegger C, Nordvang RT, Zeuner B, Lezyk M, Difilippo E, Logtenberg MJ, et al. Backbone structures in human milk oligosaccharides: trans-glycosylation by metagenomic b-N-acetylhexosaminidases. Appl Microbiol Biotechnol 2015; 99:7997–8009. [11] Matahira Y, Tashiro A, Sato T, Kawagishi H, Usui T. Enzymic synthesis of lacto-N-triose II and its potential analogues. Glycoconjugate J 1995;12: 664–71. [12] Dion M, Fourage L, Hallet J-N, Colas B. Cloning and expression of a b-glycosidase gene from Thermus thermophilus. Sequence and biochemical characterization of the encoded enzyme. Glycoconjugate J 1999;16:27–37.
6
[13] Sandoval M, Ferreras E, Pe´rez-Sa´nchez M, Berenguer J, Sinisterra JV, Hernaiz MJ. Screening of strains and recombinant enzymes from Thermus thermophilus for their use in disaccharide synthesis. J Mol Catal B: Enzym 2012;74:162–9. [14] Kengen SWM, Luesink EJ, Stams AJM, Zehnder AJB. Purification and characterization of an extremely thermostable b-glucosidase from the hyperthermophilic archaeon Pyrococcus furiosus. Eur J Biochem 1993;213:305–12. [15] Hansson T, Adlercreutz P. Optimization of galactooligosaccharide production from lactose using b-glycosidases from hyperthermophiles. Food Biotechnol 2001;15:79–97. [16] Bridiau N, Maugard T. A comparative study of the regioselectivity of the b-galactosidases from Kluyveromyces lactis and Bacillus circulans in the enzymatic synthesis of N-acetyl-lactosamine in aqueous media. Biotechnol Prog 2011;27:386–94. [17] Warmerdam A, Paudel E, Jia W, Boom RM, Janssen AEM. Characterization of bgalactosidase isoforms from Bacillus circulans and their contribution to GOS production. Appl Biochem Biotechnol 2013;170:340–58. [18] Song J, Imanaka H, Imamura K, Minoda M, Katase T, Hoshi Y, et al. Cloning and expression of a b-galactosidase gene of Bacillus circulans. Biosci Biotechnol Biochem 2011;75:1194–7. [19] Wilkins MR, Gasteiger E, Bairoch A, Sanchez JC, Williams KL, Appel RD, et al. Protein identification and analysis tools in the ExPASy server. Methods Mol Biol 1999;112:531–52. [20] Zeuner B, Jers C, Mikkelsen JD, Meyer AS. Methods for improving enzymatic trans-glycosylation for synthesis of human milk oligosaccharide biomimetics. J Agric Food Chem 2014;62:9615–31. [21] Song J, Abe K, Imanaka H, Imamura K, Minoda M, Yamaguchi S, et al. Causes of the production of multiple forms of b-galactosidase by Bacillus circulans. Biosci Biotechnol Biochem 2011;75:268–78. [22] Fukuda M. Cell surface glycoconjugates as onco-differentiation markers in hematopoietic cells. Biochim Biophys Acta 1985;780:119–50. [23] Voorhorst WGB, Eggen RIL, Luesink EJ, de Vos WM. Characterization of the celB gene coding for b-glucosidase from the hyperthermophilic archaeon Pyrococcus furiosus and its expression and site-directed mutation in Escherichia coli. J Bacteriol 1995;177:7105–11.
www.elsevier.com/locate/nbt Please cite this article in press as: Zeuner, B. et al., Thermostable b-galactosidases for the synthesis of human milk oligosaccharides, New Biotechnol. (2016), http://dx.doi.org/10.1016/ j.nbt.2016.01.003
