A method of producing a human milk oligosaccharide (HMO)

TECHNICAL FIELD

The disclosure relates to the production of sialylated human milk oligosaccharides , more particularly to the production of nutritionally enriched formula in milk. In an embodiment, the disclosure also relates to fucosylated HMO . In a further embodiment, the disclosure relates to the production of sialylated as well as fucosylated HMO .

BACKGROUND

Breast milk (or mother' s milk) produced by mammary glands located in the breast of a human female to feed a young child is the primary source of nutrition for new-borns . Breast milk comprises of various complex proteins , lipids, carbohydrates and other, biologically active components . While breastfeeding is the ideal food for infants, as it provides all the energy and nutrients a child needs, not all parents can (e . g . due to low milk supply) or want to breastfeed.

According to the W. H . O. only 41% of infants aged 0-6 months are exclusively breastfed. A remaining part of the infants are nurtured by formula milk specially developed for infants or combine breastmilk and infant formula .

Infant formulas are manufactured food products for feeding babies and infants until about 12 months of age . Follow-up and toddler formulas are also available up to usually 3 years of age. These include follow-on toddler formulas for children aged 6 months to 3 years and are usually divided into several subcategories, with each subcategory having a different composition ratio according to the children' s developmental needs. Transition formulas in the US are marketed for children from age 9 to 24 months and toddler milk is sold for children aged 12 to 26 months. Infant formula as described in present application shall apply for all nutritional formulas for children aged 0 to 3 years of age. The infant formulas are usually prepared from powder mixed with water for bottle- feeding or cup-feeding. Infant formulas are essential food supplements and are recommended by health professionals when the parent cannot or will not breastfeed due to the mother' s health condition (e.g. a HIV-positive mother) , or the baby' s condition (e.g. the infant has a birth defect or inborn error of metabolism, such as galactosemia, making breastfeeding difficult or impossible) . It is also possible that allergies, lactation insufficiency or risk of malnutrition play a role in choosing formula over breastmilk, but a societal structure and pressures, personal beliefs, experiences, or preferences may also prompt the use of formula. Therefore, it is essential that nutritious formula milk is available that resembles breast milk as closely as possible.

Breast milk is characterised by the presence of human milk oligosaccharides (HMOs) in the range of 5-15 g/L. HMOs are sugar molecules and are the third most abundant solid components of human milk after lactose and fat. Approximately 200 structurally different HMOs are known, and the composition is individual to each mother and varies over the period of lactation. HMOs are considered to play an important role in

03278-PCT the health of the breastfed infant as they function as antimicrobial agents and prebiotic agents in the gut of the breastfed infant, blocking pathogenic adhesion by changing the surface of the receptor cells .

HMOs are lactose-decorated oligosaccharides and are composed of up to five monosaccharides : β-d-glucose (glucose / Glc) , β-d-galactose (galactose / Gal) , β-d-N-acetyglucosamine (N- acetylglucosamine / GlcNAc) , α-l-fucose ( fucose / Fuc) and sialic acid (SA) , with α-d-N-acetylneuraminic acid (N- acetylneuraminic acid / Neu5Ac) being the predominant form of

SA.

Three maj or HMO categories are present in breast milk: 1 ) fucosylated neutral HMOs (35-50% ) ; 2 ) sialylated acidic HMOs

( 12-14% ) , and 3 ) non-fucosylated neutral HMOs ( 42-55% ) . 2 fucosyllactose (2 ' -FL) is part of the fucosylated, while

Lacto-N-neotetraose (LNnT) is part of the non-fucosylated neutral HMOs .

Sialylated acidic HMOs are thought to be involved in the brain and cognitive development of an infant . The two simplest ones are 3' -sialyllactose (3 ' -SL) and 6' -sialyllactose ( 6' -SL) , in which a sialic acid molecule is linked to the galactosyl subunit of lactose at the position 3 or 6, respectively. The concentration range of 3' -SL is between 90-350 μg/mL in the first 35 lactation days , which then increases to 90-840 μg/mL from 0. 5 to 8 months .

Fucosylated HMOs , e . g. 2 ' -fucosyllactose ( 2 ' -FL) and fucosylated-LNnT, as well as Lacto-N-neotetraose (LNnT) are thought to play role in the development of healthy gut microbiota and are essential for the health and development of infants .

There has been increased interest in HMOs over the past few years, in particular the synthesis of these complex carbohydrates . Many attempts have been made to produce individual HMOs via organo-chemical synthesis and, due to its stereoselectivity, via enzymatic means as well as by fermentation .

WO2018122225A1 describes the in vivo production of sialylated compounds using an engineered E. coli strain.

Mcj arrow et al . (Paul Mcj arrow, Jean Garman, S . H. ; Amelsfort,

A. Van. Dairy Process and Product, December 13, 2002. ) developed an enzymatic process using the sialidase from

Arthrobacter ureafaciens to produce 3' -SL through in vi tro process . The reaction used K (kappa) -casein glycomacropeptide

(CGMP) as sialic acid donor .

CN106190938A describes the biosynthesis of 3' -SL using a recombinant E. coli by providing a successful gene knockout scheme .

US10533164B2 describes a mutant enzyme (EC 3. 2. 1. 18 ) having trans-sialidase activity for producing a range of trans- sialylated mono- and oligosaccharides .

EP1003366 describes a method for producing sialyloligosaccharides in a dairy source or in a cheese processing waste stream using Trypanosoma cruzi trans- sialidase (TcTS ) .

EP2707380B1 describes a method to produce one or more different HMOs using T. cruzi to produce sialylated HMOs and Thermotoga maritima to produce fucosylated HMOs or their mixtures in a stepwise process .

Nyffeneger et . Al (Nyffenegger, C . ; Nordvang, R. T . ; Jers ,

C . ; Meyer, A. S . ; Mikkelsen, J. D. Design of Trypanosoma

Rangeli Sialidase Mutants with Improved Trans-Sialidase

Activity. PLoS One 2017, 12 (2 ) , e0171585 . https : //doi . org/10 . 1371 /j ournal . pone . 0171585 . ) investigated mutated Trypanosoma rangeli enzymes in aqueous solutions containing D-lactose, CGMP in a phosphate-citrate buffer at 30 °C for producing enzymes with reduced hydrolytic and increased trans-sialidase activity.

One of the challenges associated with the development of HMOs for infant formula is that HMOs are large and complex molecules . Furthermore, there is a limited availability of suitable carbohydrate-processing enzymes (Leloir-type glycosyltransferases (GTs, EC 2. 4 . 1. - ) or glycosidases ) due to problems with expression, solubility and high cost . Another challenge is that while individual HMOs have been successfully produced, there is a need for producing complex mixtures of naturally occurring oligosaccharides or derivatives .

A number of HMOs has in the recent years added to the approved novel foods authorized to be placed on the marked in the

European Union pursuant to Regulation (EU) 2015/2283 , including 3' -sialyllactose, 2 ' -fucosyllactose, and Lacto-N- neotetraose (LNnT ) . Therefore, a need exists to produce HMOs in a complex medium like milk or milk-derived products for infant formulas . Furthermore, there is a need for providing milk or milk-derived products containing two or more different types of HMOs, such as sialylated as well as fucosylated HMOs . SUMMARY

It is an obj ect to provide a method for producing a human milk oligosaccharide ((HMO)) in milk, comprising the step of incubating an aqueous solution comprising

- at least one sialic acid donor, at least one milk or milk-derived product comprising lactose, and at least one trans-sialidase enzyme .

The present inventor has surprisingly experienced that the activity of the trans-sialidase enzyme was not substantially inhibited by complex media constituted by the milk or milk- derived product . Furthermore, it was experimentally experienced that the decomposition by hydrolysation of the produced HMOs was reduced when a milk or milk-derived product was used, suggesting that milk or the milk-derived product inhibited the hydrolyzation of the trans-sialidase .

In a possible implementation form of the first aspect, the milk has been pretreated by de-creaming, homogenization, pasteurization, and/or sterilization.

It is advantageous to pretreat the milk product prior to use to remove e . g. bacteria or to prepare the milk ready for human consumption by traditional milk-processing methods .

In a possible implementation form of the first aspect, the milk-derived product is a whey, whey-derived product, or aqueous lactose solution and the obtained incubated aqueous solution is adding milk. The milk or the milk-derived product may be subj ected to one or more pretreatment steps, including de-creaming to produce a milk having a lower cream content, homogenization to obtain a uniform dispersion of ccrreeaamm particles in the milk, pasteurization to reduce the amount of possible pathogens, or sterilization ttoo reduce or eliminate the presence of microorganisms in the milk. In a certain embodiment of the invention, the milk is selected among the group comprising whole milk, skimmed milk, and mixtures thereof . Generally, the fat content is in the range of 0 to 7% , suitably between

0. 1 and 5% . The protein content of the pretreated milk may be in the range of 1 to 7% by weight, such as 2 to 5% by weight .

In a certain embodiment of the invention the milk-derived product is selected among the group comprising whey and whey- derived products, such as whey permeate or an aqueous solution of lactose . When the whey or whey-derived product is used in the incubation of the present invention, it is subsequently mixed with a milk having a protein content of at least 1% , and suitably at least 2% by weight to produce the infant formula or a precursor thereof .

The source of the milk or milk-derived product may be selected among large group of mammal milk, such as bovine milk, caprine milk, ovine milk, buffalo milk, equine milk, camel milk, yak milk, reindeer milk, or mixtures thereof . In a suitable embodiment of the invention, the source of the milk or milk- derived product is bovine milk. Alternatively, the source of milk or milk-derived product may be 'alternative milk' or more commonly known as 'plant-based beverage' or 'plant-based drink' or 'plant-based milk' , which are popular vegan, plant- based alternatives to dairy milk products. The plant-based milk or plant-based milk-derived product may be selected among large group of plant-based milk, such as almond milk/drink, coconut milk/drink, rice milk/drink, soy milk/drink, hemp milk/drink, ooaatt milk/drink, pea milk/drink, peanut milk/drink, plant-based milk/drink produced from grains (e.g. barley, fonio, maize etc.) , pseudocereals (e.g. amaranth or buckwheat etc.), legumes (e.g. lupin, soy etc.) , nuts (e.g. brazil nut, cashew nut etc.), seeds (e.g. chia seed, flax seed, hemp seed etc.) or other, plant-based alternatives such as coconut, potato or tiger nut. The plant-based milk or plant-based milk-derived product may also be created by mixing two or more types of plant-based milk together.

In a possible implementation of the present invention the sialic acid donor is selected among glycomacropeptide s

(GMPs) . GMPs are generally rich in sialic acids, which can be donated to the enzymatic process . Suitably the glycomacropeptide is K-casein glycomacropeptide (CGMP) .

Besides being a suitable donor of sialic acid, CGMP is a suitable protein source for infants and thus improves the beneficial effect of the infant formula. Furthermore, CGMP contains a relatively high level of large neutral amino acids, such as threonine and isoleucine.

In a possible implementation form of the first aspect, the sialidase and/or trans-sialidase is derived from Trypanosoma rangeli .

In a possible implementation form of the first aspect the trans-sialidase derived from Trypanosoma rangeli is Tr15. The trans-sialidase used in the present method may be derived from a number of organisms, including Trypanozoma cruzi (TcTS ) and Trypanosoma rangeli (TrSA) . The trans-sialidase enzyme from Trypanozoma cruzi (TcTS ) sshhoowwss a high trans-sialidase activity with low hydrolytic activity and has been used for enzymatic glycan sialylation. The drawback of using T. cruzi for industrial production of food-grade HMOs is that the organism constitutes a virulence factor possibly causing Chagas disease . Therefore, it may be difficult to obtain an approval by the suitable authorities for using the trans- sialidase of TcTS for use in commercial production of food ingredients .

Trypanosoma rangeli is a non-pathogenic relative of T. cruzi, having about 70% sequence identity with T. cruzi . Thus , sialidases and/or trans-sialidases derived from Trypanosoma rangeli (TrSA) are generally preferred. Modifications made on the TrSA sialidase enzyme derived from Trypanosoma rangeli resulted in increased trans-sialidase activity in aqueous solutions as previously described by Nyffenegger et al . , with the modified TrSA enzyme resembling the trans-sialidase properties and activity of TcTS . By engineering the sialidase (TrSA) produced by Trypanosoma rangeli, engineered sialidase enzymes can be obtained with trans-sialidase properties and activities, resembling those of TcTS . Therefore, in a certain implementation, the modifications of sialidase enzyme derived from Trypanosoma rangeli (TrSA) termed Tr15 and Tr16 are preferred. Tr15 is generally preferred over Tr16 because of a lower activation energy. In a possible implementation form of the present invention the activation energy of said trans-sialidase enzyme is in the range of 12.5 to 25.5 kJ/mol .

In a second aspect of the present invention, t thhee aqueous solution in addition to at least one sialic acid donor, at least one milk or milk-derived product comprising lactose, and at least one trans-sialidase enzyme further comprises at least one fucose donor, at least one fucose acceptor, and at least one trans-fucosidase enzyme . The simultaneous activity of the trans-sialidase enzyme and the trans-fucosidase enzyme result in tthhee formation of 3' -sialyllactose as well as fucosylated-Lacto-N-neotetraose ( fucosylated LNnT) .

In a possible implementation of the second aspect, the fucose donor is 2 ' -fucosyllactose (2 ' FL) and the fucose acceptor is Lacto-N-neotetraose (LNnT) . The resulting product after the reaction with the trans-fucosidase enzyme is fucosylated LNnT and 3' -S, 2' -FL.

It was surprisingly found that both the trans-fucosylation and trans-sialylation reactions occur without inhibiting each other' s enzymatic activity, producing sialylated-HMOs as well as fucosylated-HMOs in milk .

In a possible implementation of the second aspect, the trans- fucosidase enzyme is derived from Tannerella forsythia . In particular, the trans-fucosidase enzyme is a wild-type or modified α ( 1, 2 ) -trans-fucosidase derived from Tannerella forsythia . In a possible implementation of the invention the trans- sialidase and/or trans-fucosidase are immobilized. When immobilized enzymes are used in the present method it is secured that enzymes are not appearing in the final product or only appears in a limited amount . Furthermore, immobilized enzymes may be used again for a subsequent batch of milk of milk-derived product . In a certain embodiment of the present invention, the immobilized enzymes are retained after the reaction by a mesh and reused in a subsequent batch. The liquid part, i . e . the treated milk or milk-derived product may be collected and optionally stored before spray drying.

The immobilization may be effected e . g. by cross-linkage, affinity-tag binding, adsorption onto a surface, entrapment, or covalent bonding. In a certain embodiment of the invention, the crosslink enzyme aggregation (CLEA) immobilization technology is used because it is currently regarded food grade safe .

It is advantageous to immobilize the enzymes as they may be reused for further reactions and it also ensures that the final solution doesn' t have enzymes present .

In an alternative embodiment, the incubation is performed in a dialysis module comprising a first and a second chamber separated by an ultrafiltration membrane, wherein a first chamber is used for retaining the trans-sialidase and/or the trans-fucosidase enzymes and a second chamber is used for the aqueous solution containing the reactants .

The ultrafiltration membrane allows the reactants, such as

CGMP, 3' -SL and lactose, to pass and the enzyme ( s ) to be retained. By maintaining the enzyme ( s ) in one chamber of the dialysis cell and the aqueous solution containing the reactants in the other chamber, it is avoided that the enzyme ( s ) appear in any substantial amount in the final product . Furthermore, the enzyme ( s ) may be reused for another batch.

In a certain embodiment of the invention, the temperature is maintained below normal room temperature . In particular, the temperature of the trans- sialidase and/or the trans- fucosidase reaction is in a range of 3 °C to 10°C, such as about 5 °C .

This is advantageous as milk is often handled and stored at low temperatures during production, hence the method allows for e . g . in situ production of HMOs in e . g. milk production factories .

In a certain embodiment, the method includes the step of post- treating the incubated aqueous solution by pasteurizing. The pasteurization generally involves the heating of the aqueous solution to about 72 °C for 15-25 seconds . Once the milk has been pasteurized, it is then cooled to less than about 5 °C .

A further post-treatment step may include the step of freeze- drying or spray-drying the incubated and/or pasteurized aqueous solution to obtain HMO-enriched dry milk infant formula .

In a possible implementation form of the first aspect, the method further comprising the step of wet-blending or dry- blending of the aqueous solution or the freeze-dried or spray- dried dry milk with ingredients conventionally used in infant formula for obtaining an HMO-enriched dry milk infant formula .

In a further embodiment, an HMO-enriched dry milk infant additive is obtainable by the method disclosed.

In a further embodiment, an HMO-enriched dry milk infant formula is obtainable by the method disclosed.

The advantage of being able to produce an HMO-enriched dry milk infant formula or additive is that the production time is shortened, and the method allows for in situ production of

HMOs in e . g. milk production factories .

BRIEF DESCRIPTION OF THE DRAWINGS

In the following detailed portion of the present disclosure, the aspects, embodiments and implementations will be explained in more detail with reference to the example embodiments shown in the drawings, in which:

Fig. la shows the trans-sialidase reaction scheme in accordance with an embodiment;

Fig. lb shows the reaction mechanism for TfFucl , Tr15 in a one pot reaction (a) , the reaction mechanism of TfFucl transfucosylation of LNnT and 3' -SL using 2 ' -FL as fucose donor (b) , and the one pot reaction mechanism of Tr15 transsialylation of LNnT, 2 ' -SL and lactose using cGMP as sialic acid donor (c) ; Fig. 2 shows the concentration change of 3 ' -SL and SA over time when using Tr15 in milk and in aqueous lactose solution in accordance with an embodiment ;

Fig 3. shows the results of the dose response of trans- sialylation of organic whole milk and aqueous lactose solution by various concentrations of Tr15 over time in accordance with an embodiment;

Fig. 4 shows the concentration change of 3 ' -SL and SA over time when using Tr16 in milk and aqueous lactose solution in accordance with an embodiment;

Fig 5. shows the results of the dose response of trans- sialylation of organic whole milk and aqueous lactose solution by various concentrations of Tr16 over time in accordance with an embodiment;

Fig. 6. shows the transfucosylation of LNnT by free (non- immobilized) TfFucl at two different enzyme concentrations ;

Fig. 7. shows the transfucosylation of LNnT by immobilized

TfFucl at two different enzyme concentrations .

DETAILED DESCRIPTION

The general reaction scheme for a sialylation reaction using a trans-sialidase enzyme is shown below.

Sialidase enzymes Tr15 and Tr16 derived from Trypanosoma rangeli (TrSA) were obtained by cloning and expressing in

Pichia pastoris by a previously described method by

Nyffenegger et al . which is hereby incorporated by reference .

The modified sialidase enzymes Tr15 and Tr16 comprise of 15 and 16 amino acid mutations, respectively, and were previously found to display a 3' -SL production on the same level as the native enzyme in aqueous solutions, however, with reduced hydrolytic activity. Hydrolytic activity is generally not desired because 3 ' -SL will be degraded to sialic acid. Thus, by decreasing the hydrolytic activity, the ability of the enzyme producing 3' -SL is increased. Tr15 comprises of energetic (T39A, F59N, D285G) and loop mutations (Loop 197-203, VTNKKKQ) in addition to TrSA _5mut mutations (M96V, A98P, S120Y, G249Y, Q284P) previously described by Nyffenegger et al . Tr16 comprises of the mutations as described above for Tr15 , but with an additional structural mutation ( I37L) .

The TcTS from the human pathogen Trypanozoma cruzi shows a high trans-sialidase activity with low hydrolytic activity and has been used for enzymatic glycan sialylation . While

TcTS is capable of catalyzing the reaction, the practical use may be hampered by the drawback that T. cruzi is a human pathogen and therefore may be less desirable for industrial production of food-grade HMOs . Trypanosoma rangeli is a non- pathogenic relative of T. cruzi, showing about 70% sequence identity with T. cruzi .

Surprisingly, it was found that the Tr15 and Tr16 enzymes also show high degrees of trans-sialylation activities in a complex media like milk and therefore it becomes possible to produce sialylated HMOs directly in milk or milk-derived products . In addition, the unwanted hydrolysis of the produced HMOs exemplified by 3' -SL to SA is lower for the Tr15 and Tr16 in milk compared ttoo aann aqueous llaaccttoossee solution, suggesting that the hydrolysis reaction of the Tr15 and Tr16 is inhibited by components in milk.

Milk is a nutrient-rich liquid food, with a pH of around 6. 4 to 6. 8 , comprising of lipids, proteins, salts, minerals and vitamins, sugars and carbohydrates and various enzymes , The typical nutritional composition of cow milk is 87. 8% water, 3.2% protein, 3. 9% fat (including saturated fatty acids, monounsaturated fatty acids and polyunsaturated fatty acids ) , 4 . 8% carbohydrates, and 14 mg cholesterol and 120 mg Calcium. Due to the complex physical and chemical structure of milk it was surprising to discover that the milk structure did not inhibit the enzymes . Moreover, it was found that the desired reactions not only take place by producing a comparable level of sialylated-HMOs found in breast milk, but after a short reaction time of 5 minutes, the concentration of sialylated- HMOs exceeded the concentration of 3' -SL found in human breast milk. The reaction mechanism is not fully understood, but at present it is thought that CGMP is connected to the Thr amino acid of the enzymes via an O-linked glycosyl-bridge and the lactose is connected to the enzyme via either a Thr or Ser group . The addition of trans-sialidase enzyme to the reaction mixture results in the production of 3' -sialyllactose from the lactose available in milk. Due to hydrolysis, a small amount of sialic acid is produced, both via the hydrolysis of 3' -sialyllactose and via the hydrolysis of the sialic acid donor CGMP . Thus , production of sialic acid is generally not desired as it reduces either the concentration of one of the reactants or the product .

Fucosylated HMOs and N-Glycans have been shown to regulate gut microbe in breast-fed infants , hence playing a role in the health and development of an infant . Similar to sialylated HMOs, the concentration of fucosylated HMOs fluctuate during lactation and the fucosylation levels of N-glycans in breast milk were shown to be increased during late lactation. The changes in the levels of fucosylated HMOs and milk N-glycans may be correlated with the growth of Bifidobacterium spp. and Lactobacillus spp. in the gut of infants during early and later lactation, respectively.

According to a certain hypothesis, it is possible to produce 3' -sialyl, 2 ' -fucosyllactose in aqueous solutions and milk using only the enzymes Tr15 or Tr16.

CGMP is added to an aqueous solution comprising lactose and milk as a sialic acid donor. 2' fucosyllactose (2' -FL) is added to the solution as a sialic acid acceptor.

Using Tr15, Tr16 and TcTS it should be possible to produce not only 3'-SL but the pharmaceutically relevant compound 3'- sialyl, 2' -fucosyllactose (3'-S,2'-FL) .

It is hypothesized that due to the enzymes' selectivity for sialylation, the 3'-SL rraattee is high and is close to the results demonstrated in other embodiments . However, the addition of 2' -FL presents a more complex mixture and it is thought that 2' -FL fucosylated the 3' -sialyllactose at the 2'- position at a higher rate than at the 3' -position due to the hinderance of the active site at the 3' -position.

According to another hypothesis a two-step: and one-pot reaction: produces biologically relevant HMOs including 3' -sialyl, 2' - fucosyllactose (3' -S, 2' -FL) .

In a one-pot embodiment it is hypothesized that it is possible to produce 3'-SL, 3'-S,2'-FL as well as fucosylated-lacto-N- neotetraose (2'-FLNnT) in vitro using Tr15, Tr16 or TcTS and a (1, 2) -transfucosidase.

The schematic setup of the one-pot embodiment is shown below:

Fig. lb shows the most probable reaction mechanisms of TfFucl transfucosylation, Tr15 trans-sialylation and a one pot reaction of using both TfFucl and Tr15 in accordance with an embodiment .

The transfucosylation of LNnT wwaass studied using TfFucl transfucosylase (Fig. lb, (a) ) , testing 2' -FL as fucose donor and LNnT as acceptor molecule in order to investigate TfFucl and its ability to use different types of acceptor molecules. In addition to the above, tests were performed on different acceptor molecules and TfFuc 1 ability to perform transfucosylation at low temperatures (5 °C) . LC-MS analysis wwaass performed to study and quantify TfFucl transfucosylation products, identifying 2 main peaks characterized by the the m/z of 852. 15 , the mass of the fucosylated product . MS2 spectra was performed in order to characterize the products formed, identifying two main peaks :

Peak 1, characterized by the fragment m/z 364 and hence the product formed is LNFP III ; Peak 2 showed the fragment m/z

325 , which is the fragment characteristic of LNFP I .

The yield of the total fucosylated products (LNFP I and LNFP III ) increased a long with increasing reaction time peaking after 2 h reaction with a yield of 13.7 ± 0.5 % . At longer reaction times ( 6 h) the yield of fucosylated LNnT decreased to 10. 1 ± 1. 0 % .

The decrease in yield over time is due to the natural hydrolytic reaction, as fucosidases hydrolyze the fucose donor .

Transsialylation of LNnT was performed using engingeered transsialidase from T. rangeli , Tr15, which is known to be able to use several HMO core structures as sialic acid acceptors (Fig. lb, (b) ) . Previous studies investigated different acceptor-substrates, such as as LNT, LNFP V, LNnFP

V as well as GOS, and they were all sialylated by Tr15.

Previously, LNnT was also studied, but no sialylation was measured. The inventors of present application however were able to measure transsialylase activity at 5 °C using three different enzyme dosages . The reaction evolution of Tr15 transsialylation of LNnT was observed using cGMP as sialic acid donor . Three different Tr15 dosage were used: 132 μg/mL, 13.2 μg/mL, and 1.32 μg/mL .

The LC-MS spectra showed a peak with the mass m/z 997. 41 , which corresponds to the mass of the sialylated LNnT molecule . The results of the LC-MS spectra were also confirmed by NMR analysis, confirming that the product after Tr15 reaction is indeed sialylated LNnT with C3 sialylation of the external Gal . The transialylation reaction was tested for three dosages of Tr15 and for all of them it was possible to observe formation of sialylated LNnT . Product yield increased over reaction time reaching up to 20 % of yield after 6 h using the highest enzyme dosage ( 132 μg/mL) .

A one pot reaction wwaass also investiaged (Fig. lb, (c) ) , performing bbootthh sialylation and fucosylation reactions simultaneously, using Tr15 and TfFucl . One pot reactions were run both in an aqueous system and in organic whole milk. LC-MS was used also in the one pot reactions to characterize the products and surprisingly the two enzymes worked together, as the product profile, MS and MS2 spectra looked the same as the one in the single enzymes reactions, tthhuuss ,, the main reaction products obtained in milk are 3' SL sialylated LNnT,

LNFP I and LNFP III .

A control reaction was run to understand if TfFucl was able to transfucosilate 3' -SL to obtain 3F3S-L . Four different concentrations of 3' -SL were studied varying from the value of concentration which was found as maximum in the one pot reaction ( 0.2 mM) to the concentration of 50 mM which was shown to be a good acceptor concentration for the reaction with LNnT . The amount of 3F3SL obtained increased at increased

3 ' SL concentration.

Examples

Example 1

The general experimental setup is as follows :

Organic whole milk was purchased from Lidl and the lactose content in whole milk of 112 mM was assessed by LC-MS with a dilution of 1000 times in 70% v/v acetonitrile prior to analysis .

Casein glycomacropeptide (CGMP) was obtained from the commercial product Lacprodan CGMP-20 and was purified from molecular weight impurities prior to use by ultrafiltration on a 10 kDa regenerated cellulose membrane (RC70PP; AlfaLaval , Naskov, Denmark) and was concentrated to 37. 1 g/L . The sialic acid content was determined after 48 h desialylation using the unspecific sialidase from Arthrobacter urea fa ci ens

(Sigma-Aldrich, Steinheim, Germany) to a concentration of 10 mM, which was found to be the sum of α2 , 3- and α2, 6-bounded sialic acid with distribution of 50 : 50.

Sialidase enzymes Tr15 and Tr16 from Trypanosoma rangeli TrSA were obtained by cloning and expressing in Pichia pastoris by a previously described method by Nyffenegger et al . which is hereby incorporated by reference . A 5L scale of recombinant sialidases (Tr15 and Tr16) in P. pastoris were performed according to Zeuner et al . with a total fermentation time of

112 hour . The Tr15 and Tr16 enriched fermentation broths were recovered by centrifugation at 5300 x g at 5 °C for 1 hour and the supernatants were subj ected to sterile filtration and concentration by ultrafiltration, using a cross-flow bioreactor system with a 30 kDa cutoff polyethersulfone membrane (Millipore, Sartorius, Denmark) , as described by

Zeuner et al . The enzyme aliquots were stored at -80 °C until further use .

Protein concentrations were found to be 16.2 mg/mL and 20. 5 mg/mL for Tr15 and Tr16, respectively, using the Bradford method (Thermo Fischer Scientific) and Bovine Serum Albumin as standard. It is however understood that other suitable protein concentrations may be used in accordance with an embodiment .

Trans-sialidase from Trypanosoma cruzi, TcTS, was cloned and expressed in Pichia pastoris using the same fermentation process as described above for Tr15 and Tr16. Protein concentration using the Bradford method was found to be 5. 8 mg/mL . It is however understood that other suitable protein concentrations may be used in accordance with an embodiment .

Trans-sialylation reactions were run in both aqueous lactose solutions and milk using the same reaction conditions for both.

The enzymes may or may not be immobilized; the immobilization helps the retention and re-usability of the enzyme and hence is advantageous for industrial use . Trans-sialylation reaction of organic whole milk and aqueous solution using Tr15 and Tr16 were performed in an unbuffered reaction mixture with a lactose : CGMP ratio of 3. 8. It is however understood that other suitable acceptor : donor ratios are also acceptable and that a reaction buffer may or may not be used.

Arrhenius equation Eq 1 was linearized to Eq 2 in order to determine the enzyme activation energy: where k is the initial 3' -sialyllactose formation rate (mM/s ) , Ea is the activation energy (J/mol ) , R is the gas constant ( 8.3145 J/ (mol*K) ) , TT is the temperature ( (KK)) and A is a constant .

Small and large scale studies have been carried out in accordance with an embodiment of the invention.

Fig. 2 shows the results of the trans-sialylation reaction using 15 μg/mL Tr15 in milk and aqueous lactose solution. It is shown that despite the complex nature of milk, a comparable level of 3' -SL was produced and the milk as a solution did not appear to inhibit the enzymatic activity of Tr15. During this reaction, the hydrolytic activity of the enzyme was only measurable after one hour of reaction time for the milk trans- sialylation. Further, surpris ingly, the hydrolytic activity observed in the aqueous ssoolluuttiioonn and milk sshhoowwss that the hydrolytic activity of the enzyme is more active in aqueous solutions than in milk as the concentration of SA was found to be higher after one hour of reaction time in the aqueous solution than in milk .

Fig. 3 shows the dose response ccuurrvvee for using 3 separate concentrations of Tr15 in milk aqueous lactose solution.

Surprisingly, already after 10 mins of reaction time of 15 μg/mL Tr15 , 1. 8 mM of 3' -SL was obtained, which corresponds to 1160 μg/mL, which is in the higher end of the 3' -SL concentration range of breast milk (i . e . 261-1245 μg/mL) .

The dose response experiment confirmed that the concentration of 3' -SL was a function of the concentration of enzyme available in the reaction mixture . The decrease of 3 ' -SL was observed for the higher dosage of Tr15 ( 150 μg/mL) after 20 minutes of reaction time due to the remaining hydrolytic activity of the enzyme . This is hypothesized to be due to two simultaneous reactions that take place : 1 ) the trans- sialylation of lactose and 2 ) the hydrolysis of the 3' -SL product or the sialic acid donor CGMP . Thus, the higher the enzyme concentration, the higher the 3' -SL concentration and so hence the more prominent the hydrolytic activity is .

Fig. 4 shows the results of the trans-sialylation reaction using 15 μg/mL Tr16 in milk and aqueous lactose solution . Trans-sialylation of whole organic milk by Tr16 led to comparable 3' -SL concentrations to the standard trans- sialylation reaction of an aqueous solution, however 1 hour after the reaction began, the 3' -SL concentration was higher in the aqueous solution. It is also apparent that the hydrolytic activity of the enzyme is mostly present during the trans-sialylation of the aqueous ssoolluuttiioonn,, suggesting that the newly formed 3' -SL or the CGMP is protected by components in milk.

Fig. 5 shows the dose-response curve for using 3 separate concentrations of Tr15 in milk and lactose solution. The trans-sialylation using Tr16 appear to be slower than the trans-sialylation using Tr15 and only 0. 1 mM of 3' -SL were formed after 10 mins of reaction with 15 μg/mL of Tr16

(corresponding to 63 μg/mL) .

The dose response experiment of the trans-sialylation of organic whole milk by Tr16 verified that the concentration of

3' -SL was a function of the concentration of the enzyme in the reaction mixture .

Example 2

In another embodiment it is the obj ect of the invention to provide a method for producing sialylated galactooligosaccharides . Galactooligosaccharides (GOS ) are oligosaccharides composed of different galactosyl residues ( from 2 to 9 units ) and a terminal glucose linked by β-glycosidic bonds, such as β- ( 1—

2 ) , β- ( 1— 3 ) , β- ( 1— 4 ) , and β- ( 1— 6) and belong to the group of prebiotics . GOS effects include the stimulation of immune functions, absorption of essential nutrients and e . g . the production of the powerful anti-oxidant H2 gas or synthesis of certain vitamins .

It is thought that the addition of galactooligosaccharides to infant formula helps relieve symptoms often associated with formula feeding, e . g. constipation and hard stools . GOS are thought to improve stool frequency and relive the symptoms related to constipation. It is known from previous studies that trans-sialylation of GOS by Tr15 and Tr16 is possible in aqueous ssoolluuttiioonnss b buutt it has never been tried in complex solutions such as milk to produce a readily-available product for use as an infant formula .

Cup Oligo P GOS (by Nissin Sugar using biotechnology) , mainly comprised by low-degree polymerization (DP) in the DP range of 2 to 6 was dissolved in both aqueous solutions and in milk at the same final concentration. 7 .5 μL Tr15 enzyme was used to provide a Tr15 concentration of 150 pg/ml . 50 μL milk and

50 μL aqueous solutions were prepared and the GOS was added. The resulting concentrations were measured to be 16 mM lactose and 11. 4 mg/mL GOS . Further, 292.5 μL of CGMP was added to a final concentration of 31. 1 g/L . TThhee trans-sialylation reaction was run for 30 mins at 5 °C . The reaction results are summarized in table 1.

*> Sia DP2 in the GOS reaction

Table 1 : Concentration of sialylated compounds

A lower yield of 3' -SL concentration was observed for the reaction in the presence of GOS both in aqueous solution and in milk . The amount of 3 ' -SL without the addition of GOS was

44% and 45% higher than in the case of adding the GOS for aqueous and milk solutions, respectively. In the aqueous solution with GOS, 85% of the total sialylated compounds were

3' -SL and Sia DP2 ; the rest of the 15% is due to sialylated-

GOS . In the milk solution with GOS, 88% of the total sialylated compounds were due to 3' -SL; the rest of the 12% was due to sialylated GOS .

Even though it may appear that the addition of GOS in both solutions lower the overall sialylation yield when compared to the sialylation yield in absence of GOS, the ratio between the concentration of the sialylated compounds and the concentration of the sialic acid remained constant in both cases and equal to 6 for aqueous solution with or without GOS and 7 for milk with or without GOS .

Example 3

Trans-sialylation of organic whole milk by Tr15 and Tr16 was performed in a 350 μL reaction mixture containing 15 μg/mL of enzyme, 16 mM lactose and 31. 1 g/L CGMP (corresponding to 4 . 2 mM of 2 , 3-bounded sialic acid) to achieve the acceptor donor ratio of 3. 8 . Single reactions were run in a thermomixer for each reaction time, varying from 0 to 120 min in an unbuffered system at pH 6.7 , 25 °C and 1400 rpm. The reactions were thermally inactivated in a thermomixer at 90 °C for 10 mins .

Already after 10 minutes of reaction 1.7 mM of 3' -SL, corresponding to 1160 mg/L were obtained using Tr15 , which is comparable to the level of 3' -SL produced during a standard trans-sialylation of aqueous solutions .

After 10 minutes of reaction using Tr16, 0. 1 mM of 3' -SL were formed, corresponding to 63 μg/mL, which is comparable to 3' -

SL concentration of a standard trans-sialylation of aqueous solutions .

Table 2 . Change of concentration of 3 ' -SL and sialic acid

(SA) over time in milk using Tr15 at the concentration of 15 μg/mL .

Table 3. Change of concentration of 3'-SL and sialic acid

(SA) over time in milk using Tr16 at the concentration of 15 μg/mL .

Example 4

Trans-sialylation of organic whole milk was tested at 5 °C,

10 °C, 25 °C, 30 °C and 40°C for 150 μg/mL Tr15 and 150 μg/mL Tr16:

Both Tr15 and Tr16 showed trans-sialidase activity at 5 °C and

150 μg/mL Tr15 formed 1.3 mM of 3'-SL after 5 minutes of reaction, which correspond to 802 μg/mL and yielded to a total of 0.28 mg of 3'-SL in the final sample. Tr16 reaction was slower at 5°C as only 0.02 mM of 3'-SL were produced after 5 mins reaction at 5 °C.

TcTS was used as a benchmark enzyme to compare trans-sialidase activity of Tr15 and Tr16. After 3 hours reaction at 5 °C only 0.04 mM of 3'-SL were obtained by TcTS being 33 times lower than the 3'-SL concentration obtained in 5 minutes with Tr15:

Table 4 . Concentration of 3' -SL after 3 h of transsialylation of organic whole milk by TcTS .

Example 5

Activation energy was determined to assess the trans- sialylation reaction energy for Tr15, Tr16 and TcTs :

Table 5 . Activation energy of the trans-sialylation reaction of aqueous solution and organic whole milk by Tr15 , Tr16 and

TcTS . The activation energy values are the average of triplicates .

No statistical difference was observed between the trans- sialylation reaction energy obtained for the reaction in milk or aqueous solutions with Tr15 showing the lowest activation energy, i . e . 22. 5 and 16.5 kJ/mol for trans-sialylation in aqueous solution and milk, respectively, which consequently lead to a faster reaction.

Example 6

An upscale of the trans-sialidase reaction of organic whole milk using Tr15 was performed to study the industrial applicability of the production of 3' -SL-enriched milk.

Table 6. Concentration o 3' -SL and SA in a 0.5 L laboratory upscale of trans-sialylation reaction of organic whole milk by Tr15.

The amount of 3' -SL produced during the reaction was increased up to 30 minutes without the presence of secondary hydrolysis .

After 30 minutes, about 1. 1 g of 3 ' -SL was produced and the concentration of 3' -SL produced after 5 minutes of the reaction by Tr15 was already comparable to the concentration of 3' -SL in breast milk.

Example 7

The stability of 3' -SL in three different pasteurization temperature was assessed: known concentration of 3' -SL was dissolved in water and milk ( 0.5 mM initial concentration) and was pasteurized using the Holder method at 62.5 °C for 30 mins, at 80 °C for 1 minute and at 90 °C for 1 minute . After pasteurization, the milk underwent a drying process using freeze-drying equipment .

Table 7. Stability of 3' -SL after pasteurization and freeze- drying processes . Concentrations of 3' -SL before and after pasteurization and after freeze-drying are reported for two systems : 1 mM 3'-SL dissolved in water and 0.5 mM 3'-SL dissolved in organic whole milk.

Freeze-drying of a 5 minutes trans-sialylation reaction of o organic whole milk by Tr15 at 5 C was performed to study the stability of 3'-SL in the reaction mixture. The concentration of 3'-SL was measured before (1.39 ± 0.06 mM) and after (1.40

± 0.06 mM) the freeze-drying process . No statistical difference between the two concentrations was found, therefore the freeze-drying process is a safe process which does not affect the stability of 3'-SL.

Example 8

A one-pot reaction was performed. The enzymes trans-sialidase and trans- fucosidase were provided in the same aqueous reaction mixture together with all the reactants, i.e. CGMP, lactose, 2' -FL and LNnT.

Reaction conditions:

7.5 μL enzyme Tr15 -> final concentration 1.32 μg/mL

20 μL enzyme Tffucl -> final concentration 0.2 mg/mL = 5.4 μM

50 μL 2' -FL -> 5 mM final concentration

50 μL LNnT -> 50 mM final concentration

50 μL lactose -> 12.3 mM

292.5 μL CGMP -> 3.2 mM SA

Reaction media: water.

Temperature 5 °C

Reaction time: 2h and overnight The products formed are 3'-SL, fucosylated LNnT and 3' -S,2' -

FL. The latter is obtained from the transfucosidase reaction by Tffucl so 2' -FL is the donor and 3'-SL (produced by Tr15) is the acceptor.

Example 9

TfFucl immobilization was tested by confinement in dialysis membrane and by immobilization on silica support.

The immobilization yield of TfFucl were tested at a concentration of 0.1 mg/mL and 0.5 mg/mL.

Table 8. Comparison between two immobilization concentration and the retained concentration.

Both concentrations allowed a high enzymatic retention on the support .

Example 10 The transfucosidase activity was measured using 0. 1 μM or 0.5 μM enzyme either in free form or immobilized, as shown on

Figs . 6 and 7 . 750 μL 5 mM 2 ' FL and 750 μL 50 mM LNnT was added and the transfucosidation was performed at 25 °C .

Table 9. Transfucosylation activity of free and immobilizec

TfFucl using different concentrations of enzyme in the reaction.

TfFucl immobilization using a TfFucl concentration of 1. 91 μM retains only the 8. 5% of the transfucosidase activity compared to the free enzyme . If the immobilization takes place with a higher TfFucl concentration ( 9.77 μM) only 0. 42% of the activity is retained.

Interestingly, by increasing the concentration of TfFucl immobilized the concentration of fucosylated LNnT decrease, thus some enzyme inactivation takes place when more enzyme is immobilized, likely due to incorrect immobilization of the enzyme itself . It is hypothesized that the active site is being immobilized facing the support or multilayer immobilization which the decrease the total amount of enzyme that is able to react . Example 11

LNnT transglycosylation and recycling conditions were tested using 12.5 mg of 1. 91 μM immobilized TfFucl, 750 μL 10 mM

2 ' FL ( final concentration = 5 mM) and 750 μL 100 mM LNnT

( final concentration = 50 mM) . The enzymes were left to react for 30 minutes , and the supernatant was collected using a centrifuge after the reaction time passed. The pellets were washed with water three times and new substrates were added to repeat the reaction .

Table 10. Biocatalytic productivity of TfFucl after recycling

(table shows standard deviations ) .

The biocatalytic productivity provides information on the amount of product produced per amount of TfFucl . The presence of the washing step between different cycles did not influence the biocatalytic productivity. It is therefore possible to reuse the enzyme .

Exampel 12

Materials

Engvang organic whole milk was purchased in Lidl (Lidl Denmark K/S, Kolding) . Content of lactose in whole milk of 112 mM was assessed by liquid chromatography coupled to mass spectroscopy (LC-MS ) analysis, dilution of 1000 times in 70% v/v acetonitrile prior to analysis was needed in order to be inside the standard curve . Casein glycomacropeptide (CGMP) , in the form of the commercial product Lacprodan CGMP-20, was a gift from Aria Foods Ingredients Group (Viby, Denmark) .

CGMP was purified from molecular weight impurities prior to use by ultrafiltration on a 10 kDa regenerated cellulose membrane (RC70PP; AlfaLaval, Nakskov, Denmark) . The cGMP solution was concentrated to 37. 1 g/L with a sialic acid

(SA) content of 10 mM determined after 48 h desialylation using the unspecific sialidase from Arthrobacter ureafaciens

(Sigma-Aldrich, Steinheim, Germany) . This 10 mM concentration is the sum of the α2, 3- and α2, 6-bounded sialic acid which distribution was found in literature to be 50 : 50. The 3' - sialyllactose standard was purchased from Carbosynth

(Compton, United Kingdom) and all the other chemicals were purchased ffrroomm Sigma-Aldrich ((SStteeiinnhheeiimm,, Germany) . 2 ' - fucosyllactose and Lacto-N-neotreose were a gift from Dupont and DSM, respectively.

Enzymes preparation

The sialidase from Trypanosoma rangeli (Tr15 ) was mutated with 15 amino acid substitution and it was cloned and expressed in Pichia pastoris as described previously (B .

Zeuner, J. Muschiol, J. Hoick, M. Lezyk, M. R. Gedde, C . Jers ,

J. D. Mikkelsen, A. S . Meyer, Substrate specificity and transfucosylation activity of GH29 α-l-fucosidases for enzymatic production of human milk oligosaccharides , N.

Biotechnol . 41 (2018 ) 34-45 ) . Tr15 production was performed in a 5 L scale fermentation in P. pastoris according to Zeuner et al . (B . Zeuner, J. Hoick, V. Perna, J. D. Mikkelsen, A. S .

Meyer, Quantitative enzymatic production of sialylated galactooligosaccharides with an engineered sialidase from

Trypanosoma rangeli, Enzyme Microb . Technol . 82 (2016) 42-

50 ) . The total time for the fermentation process was 112 h. Tr15 enriched fermentation broths was recovered by centrifugation at 5300 x g 5 °C for 1 h and the supernatant was subj ected to sterile filtration and concentrated by ultrafiltration, using a cross-flow bioreactor system with a

30 kDa cutoff polyether sulf one membrane (Millipore,

Sartorius, Denmark) , as described by Zeuner et al . The enzyme aliquots were stored at -80 °C until further use . Tr15 protein concentration was 16. 2 mg/mL, and it was determined with

Bradford method (Thermo Fisher Scientific, Hvidovre ) using

Bovine Serum Albumin (BSA) as standard.

The transfucosidase from Tannerella forsythia, TfFucl, was expressed in Escherichia coli according to Zeuner et al . (B .

Zeuner, J. Muschiol, J. Hoick, M. Lezyk, M. R. Gedde, C . Jers ,

J. D. Mikkelsen, A. S . Meyer, Substrate specificity and transfucosylation activity of GH29 α-l-fucosidases for enzymatic production of human milk oligosaccharides , N.

Biotechnol . 41 (2018 ) 34-45 ) . TfFucl was grown in lysogenic broth (LB) supplemented with 100 μg/ mL ampicillin at 37 °C . When reaching an OD600 of 0. 6-0. 9, the temperature was reduced to 18 °C and expression was induced with 0. 1 mM IPTG. Cells were grown for 20 h before being harvested. E . coli cells were harvested by centrifugation (5300g, 15 min, 4 °C) and the pellet was resuspended in binding buffer (20 mM sodium phosphate buffer, 20 mM imidazole, 500 mM NaCl, pH 7. 4 ) . E coli cells were lysed by sonication ( 10 rounds of 10 s at 50% amplitude with 50 s breaks in between using a Q500 sonicator

(QSonica, Newton, USA) ) and centrifugated to remove cell debris (20000g, 20 min, 4 °C) . The resulting supernatant was passed through a 0. 45 pm filter before being loaded onto a 2 mL Ni ²⁺ Sepharose 6 FF column (GE Healthcare, Uppsala, Sweden) for immobilized mmeettaall affinity chromatography ( IMAC) purification . Unbound material was washed off with 6 column volumes (CV) binding buffer . TfFucl was eluted with 3 CV elution buffer (20 mM sodium phosphate buffer, 250 mM imidazole, 500 mM NaCl, pH 7 . 4 ) in 0.5 CV fractions . The IMAC- purified TfFucl was desalted on PD-10 columns (GE Healthcare, Uppsala, Sweden) using a glycerol-free desalting buffer (20 mM sodium acetate buffer, 100 mM NaCl, PH 6 ) to avoid interference of glycerol with HPLC-MS analysis . Protein purity was confirmed by SDS-PAGE and protein concentration was determined with Bradford reagent using BSA standards .

Transfucosidase reaction

Transfucosylation of LNnT by TfFucl was performed in a 105 μL reaction mixture containing 1.2 μM of TfFucl, 5 mM 2 ' -FL and

50 mM LNnT in order to have an acceptor donor ratio of 10.

Single reactions were run in a thermomixer for each reaction time, varying from 0 to 6 hours at 5 ”C and 1400 rpm. The reactions were then halted by thermal inactivation at 90 °C for 10 minutes . Fucosylated LNnT concentration was determined as 2 ' -FL equivalents by LC-MS analysis .

In order to understand if TfFucl was able to use other types of acceptor the transfucosylation of 3' -SL by TfFucl was tested. A 105 μL reaction mixture was performed containing

1.2 μM of TfFucl, 5 mM 2 ' -FL and 50 mM of 3' -SL in order to have an acceptor donor ratio of 10. Single reactions were run in a thermomixer for 2 h at 5 °C and 1400 rpm. The reactions were then halted by thermal inactivation at 90 °C for 10 minutes . In order to study the lower value of acceptor donor ratio able to produce fucosylated 3' -SL reactions with different concentration of 3' -SL were run; specifically it was used 0.2 , 1 and 10 mM 3 ' -S corresponding to acceptor donor ration of 0. 04 , 0.2 and 2 , respectively.

Fucosylated 3' -SL concentration was determined as 2 ' -FL equivalents by LC-MS analysis as described below.

One pot reaction and dose response

Transsialylation and transfucosylation of organic whole milk and aqueous solution by Tr15 and TfFucl were performed in a

450 μL reaction mixture containing 1. 32 μg/mL of Tr15 , 1.2 μM of TfFucl , 12.3 mM lactose ( from the actual milk or from the lactose dissolved in aqueous solution) , 22. 1 g/L of CGMP corresponding to 3.2 mM of 2 , 3-bounded sialic acid in order to have an acceptor donor ratio of 3. 8 , 5 mM 2 ' -FL and 50 mM

LNnT . Single reactions were run in a thermomixer for each reaction time, varying from 0 to 6 hours at 5 °C and 1400 rpm. The reactions were then halted by thermal inactivation at 90 °C for 10 minutes . Transsialylation dose response experiments were run in the same way as described above using Tr15 concentrations of 13.2 μg/mL and 132 μg/mL . Product concentrations were determined by LC-MS .

Transsialidase reaction evolution and dose response Transsialylation of organic whole milk and aqueous solution by Tr15 and Tr16 was performed in a 350 μL reaction mixture containing 1. 32 μg/mL of enzyme, 12. 3 mM lactose ( from the actual milk or from the lactose dissolved in aqueous solution) and 22. 1 g/L of CGMP corresponding to 3.2 mM of 2 , 3-bounded sialic acid in order to have an acceptor donor ratio of 3. 8 . Single reactions were run in a thermomixer for each reaction time, varying from 0 to 6 h at 5 °C and 1400 rpm. The reactions were then halted by thermal inactivation at 9900 °C for 10 minutes .

The dose response experiments were rruunn i inn t thhee same way as described above using enzyme concentrations of 13.2 μg/mL and 132 μg/mL .

LC-MS analysis Tr15 transsialylation and TfFucl transfucosylation products were separated and quantified using LC-MS .

For all the methods aliquots of 5 μL were onto a Hypercarb column ( 150 mm x 2 . 1 mm; 3 μm, Thermo Fischer Scientific, Sunnyvale, CA, USA) . Chromatography was performed on a Dionex UltiMate 3000 UPLC (Thermo Fischer Scientific, Sunnyvale, CA, USA) operated at 0.3 mL min ^-1 and 70 °C with a two-eluent system consisting of eluent A, 0. 1% formic acid in water, and eluent B, acetonitrile . The elution was performed as follow: 0-5 min, isocratic 100% A; 5-38 min, linear gradient to 62.5% A,

27.5% B; 38-41 min, linear gradient to 40% A, 60%B; hereafter going directly to 41-60 min isocratic gradient with 40% A,

60% B followed directly by 60-80 min isocratic equilibration with 100% A.

The HPLC was connected to an ESI-iontrap (model Amazon SL from Bruker Daltonics, Bremen, Germany) and the electrospray was operated in ultrascan mode with target mass settings of

600 m/z and a scan range from 100 to 2000 m/z. The spray settings were: capillary voltage of 4.5 kV, end plate offset

0.5 kV, nebulizer pressure at 3.0 bar, dry gas flow at 12.0

L min ^-1 , and dry gas temperature at 280 °C.

Quantification

Quantification of the precursor ion was performed using Bruker

Compass QuantAnalysis software (Bruker Daltonik GmbH) , defining one method able to determine the concentration of all substrates and products in one. All ions were observed as

[M - H]-. 2' -FL quantification was performed by defining an

Extracted Ion Chromatogram (EIC) on All MS of masses m/z 533.00, m/z 487.04, m/z 408.95, m/z 324.94, m/z 204.72 and m/z 160.43 with a width of ± 0.5, retention time 19.8 min with a window of 1 min. LNnT quantification was performed by defining an EIC on All MS of masses m/z 742.08, m/z 706.14, m/z 627.97 and m/z 381.90 with a width of ± 0.5, retention time 27.6 min with a window of 1 min. Fucosylated LNnT quantification was performed by defining an EIC on All MS of masses m/z 852.15, m/z 774.08, m/z 689.97, m/z 528.01, m/z

409.04, m/z 852.15, m/z 810.00, m/z 671.96 and m/z 486.99 with a width of ± 0.5, retention time 32 min with a window of 1 min. 3'-SL quantification was performed by defining an EIC on All MS of masses m/z 632.09 and m/z 289.86 with a width of

± 0.5, retention time 43.8 min and a window of 1 min.

Sialylated LNnT quantification was performed by defining an EIC on All MS of masses m/z 997.15, m/z 536.01, and m/z 289.87 with a width of ± 0.5, retention time 54 min with a window of

1 min. Fucosylated 3'-SL quantification was performed by defining an EIC on All MS of masses m/z 778.17 and m/z 289.95 with a width of ± 0.5, retention time 56.1 min with a window of 1 min.

In all the cases peak detection was done using algorithm version 2.1, S/N threshold 1, area threshold 0.1, intensity threshold 0.1, skim ratio 0.1 and smoothing width 1.

Calibration curves of 2' -FL, LNnT and 3'-SL were performed using 8 levels of concentrations varying from 1 to 100 μM and fitting the data with a quadratic equation.

Fucosylated LNnT and fucosylated 3'-SL were quantified as 2' -

FL equivalents while sialylated LNnT was quantified as 3'-SL equivalents . Throughout the work all the product concentrations , obtained by enzymatic reactions were converted into yields, which in general were calculated as molar yield of the product based on the total molar donor content. In the specific fucosylated LNnT and fucosylated 3'- SL yields were calculated as molar yield of the fucosylated

LNnT or fucosylated 3'-SL based on the total molar fucosyl content (5 mM) in the donor, 2' -FL, while sialylated LNnT and

3'-SL yields were calculated as molar yield of the sialylated

LNnT or 3'-SL based on the total molar sialyl content (3.2 mM) in the donor, 3'-SL. Results and Discussions - Transfucosylation of LNnT

In the current study TfFucl transfucosylation was tested using 2 ' -FL as fucose donor and LNnT as acceptor molecule in order to see if TfFucl was able to use a different type of acceptor molecule .

In addition to test TfFucl on a different acceptor molecule,

TfFucl ability to perform transfucosylation at low temperature (5 °C) was assessed as well . Being able to work at 5 ° C is a requirement to the final scope of coupling different reactions in the same system media : milk. TfFucl was indeed able to perform the transfucosidase reaction at 5 °C (Figure 8 ) , even if its optimal temperature has been shown to be 40 °C .

LC-MS analysis was used to study and quantify TfFucl transfucosylation products (Figure 8 and 9 ) . After TfFucl reaction it was possible to find 2 peaks characterized by the m/z of 852. 15 , the mass of the fucosylated product (Fig. 9 ) . In order to understand which products were formed MS2 spectra were also obtain from the LC-MS (Figure 2c) . Peak 1 ( Figure 9 ) is characterized by the fragment m/z 364 and hence the product formed is LNFP III (Figure 1 ) [ 11 ] ; Peak 2 ( Figure 9 ) present the fragment m/z 325 which is instead the fragment characteristic of LNFP I .

The yield of the total fucosylated products (LNFP I and LNFP III ) increased a long with increasing reaction time peaking after 2 h reaction with a yield of 13.7 ± 0.5 % (Figure 8 ) . At longer reaction time ( 6 h) the yield of fucosylated LNnT decreased to 10. 1 ± 1. 0 % (Figure 8 ) . The decrease in yield is not a surprise ; TfFucl is first of all a fucosidase, thus its natural reaction is a hydrolytic reaction, and hence fucose will hydrolyzed from 2 ' -FL . Fucosidases will tend to hydrolyze the fucose donor and hence at any time point during reaction both reactions will take place at the same time . The longer the reaction time the higher is the concentration of fucose donor or fucose product and thus hydrolysis will take over the transfucosylation.

Transsialylation of LNnT

In this current study Tr15 ability of transsialylate LNnT was tested again at 5 °C using three different enzyme dosages (Figure 10 and 11 ) .

The LC-MS spectra ( figure 11 ) showed a pick with the mass m/z 997. 41, which correspond to the mass of the sialylated

LNnT molecule . The LC-MS analysis was also confirmed by NMR analysis and it was confirmed that the product after Tr15 reaction is indeed sialylated LNnT with C3 sialylation of the external Gal (Figure lb, bl ) .

The transialylation reaction was tested for three dosages of Tr15 and for all of them it was possible to observe formation of sialylated LNnT (Figure 9 ) . Product yield increased over reaction time reaching up to 20 % of yield after 6 h using the highest enzyme dosage ( 132 μg/mL) .

One pot reaction

The idea driving the study was to be able to perform different type of reaction to produce HMOs molecules directly in milk in order to be able to have a HMO enriched milk ready as final ingredient or to be spray dried for infant formula production . Both enzyme have just been shown to be able to work at 5 C and produce the more complex HMO. In a previous study we have also shown that Tr15 was able to produce 3' -SL both in milk and in an aqueous solution containing lactose [ 14 ] , The one pot was therefore ran both in an aqueous media and in organic whole . In both cases Tr15 and TfFucl were able to perform the same reactions as in the singular case ( Figure 12 ) . One pot reactions were run both in an aqueous system and in organic whole milk and the reaction evolution is comparable for both systems (Figure 12a and 12b) . In both systems the transfucosylation maximum yield was obtained after one hour reaction and equal to 14% after which the fucosylated product yield decreased mostly due, as already discussed, to the hydrolysis that took over the transglycosylation reaction,

The yields of the transsialylation products : 3' SL and sialylated LNnT steadily increased over reaction time and reached a maximum after 6 hours of reaction.

LC-MS was used also in the one pot reaction to characterize the products and even when the two enzymes worked together the product profile, MS and MS2 spectra looked the same as the one in the single enzymes reactions (Figure 9 and 10 ) .

The products obtained in the one pot were only the ones that the two enzymes could produce alone . It was expected to be able to produce products with both functionalities : sialic acid and fucose, but unfortunately these products were not obtained.

A control reaction was run to understand if TfFucl was able to transfucosilate 3' -SL to obtain 3F3S-L (Figure lb, a2 ) .

Four different concentrations of 3 ' -SL were studied varying from the value of concentration which was found as maximum in the one pot reaction ( 0.2 mM) to the concentration of 50 mM which was shown to be a good acceptor concentration for the reaction with LNnT . The amount of 3F3SL obtained increased at increased 3 ' SL concentration. At the lowest concentration 0.2 mM (Figure 11 ) not detectable levels of 3F3SL were obtained and it is hard to obtain the two functionality product in the one pot set up as designed in the current study.

MS and MS2 spectra of the products peaks : m/z 852. 15 (F-LNnT,

1 and 2 ) , red line; m/z 632. 07 (3' -SL, 3 ) , blue line and m/z

997.22 (S-LNnT) , green line . The peaks reported occurs at 60 min reaction for the one pot ran in organic whole milk.

In conclusion,

1. TfFucl is able to use LNnT and 3 ' SL as fucose acceptor molecule in the transfucosylation reaction and the reaction can take place at 5 C

2. Tr15 is able to use LNnT as sialic acid acceptor in the sialylation reaction

3. The one pot reaction in milk can be performed and the main obtained products are 3' SL sialylated LNnT, LNFP I and

LNFP III

4 . In the one pot reaction the cocnetration of 3' SL in not high enough to have the formation of 3S3FL