WO2015192247A1

WO2015192247A1 - Enzymatic production of prebiotic oligosaccharides from potato pulp

Info

Publication number: WO2015192247A1
Application number: PCT/CA2015/050566
Authority: WO
Inventors: Salwa KARBOUNE; Nastaran KHODAEI
Original assignee: The Royal Institution For The Advancement Of Learning/Mcgill University
Priority date: 2014-06-21
Filing date: 2015-06-19
Publication date: 2015-12-23

Abstract

The present disclosure relates to a process of isolating non-digestible oligosaccharides from potato pulp by extracting rhamnogalacturonan content from the potato pulp, digesting the extracted rhamnogalacturonan content with a multi-enzymatic mixture to yield the non- digestible oligosaccharides from the extracted rhamnogalacturonan content; and isolating the non-digestible oligosaccharides. Also provided herewith are pharmaceutical, nutraceutical and food applications for these protein-enriched preparations.

Description

ENZYMATIC PRODUCTION OF PREBIOTIC OLIGOSACCHARIDES FROM POTATO

PULP

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority from U.S. provisional patent application 62/015,420 filed on June 21 , 2014 and herewith incorporated in its entirety.

This application also contains a sequence listing in a computer-readable format. The content of the sequence listing is herewith incorporated in its entirety.

TECHNICAL FIELD

The present disclosure relates to a process of isolating non-digestible oligosaccharides from potato pulp.

BACKGROUND

Potato pulp is as a major low-value product generated primarily upon starch production or potato processing. Despite its interesting phytochemical composition, potato pulp was mainly used as cattle feed (0bro et a/., 2004). Major components of potato pulp are cell wall polysaccharides, which include cellulose, hemicellulose and pectic polysaccharides. Pectic polysaccharides (56%) being the major components of potato cell wall have low economic value because of its low gelling properties associated with its high content of neutral sugars (0bro et a/., 2004). Indeed, unlike pectic polysaccharides from other sources, such as citrus and apple, potato pectin consists of high proportion of rhamnogalacturonan I (RG I) (72%) and smaller amount of homogalacturonan (HG) (Oomen et a/., 2002). Second distinguished structural property of potato pectin arise from the high proportion of neutral β-linked galactan side chains (representing 46% of the total pectin) linked to RG I backbone; while RG I originating from currently commercial sources of pectin are mostly substituted with arabinan side chains (Oomen et a/., 2002). In spite of the interesting structural properties of pectic polysaccharides containing galactan-rich RG I and their low cost, so far they have not been used as sources for the production of bioactive molecules.

Several bioactive effects of galactan-rich RG I have been put forward. β-(1 ,4)-galactan rich potato fibers with high molecular weights, obtained upon enzymatic treatment of pulp, have shown higher bifidogenic activity than fructooligosaccharides (Thomassen et a/., 201 1). In addition, metabolic studies of galactan-rich molecules originating from pectic polysaccharides have highlighted their health beneficial effects in stimulating immune system and preventing metastasis (Kelly, 1999; Nangia-Makker et a/. , 2002).

Moreover, progress in the understanding of the relationship between the structures of non- digestible oligosaccharides (NDOs) and their health attributes has given rise to a need for new sources of well-defined NDOs. NDOs are non-digestible food ingredients that beneficially affect the host by selectively stimulating the growth and/or activity of one or a limited number of bacteria in the colon and thus improve the intestinal health (Van Laere et al., 2000). Recently, plant cell wall polysaccharides have received a lot of attention as source of novel NDOs. Olesen et al. (1998) have shown that enzymatically solubilized potato cell wall fibers were fermentable and their consumption increased end-expiratory of H₂. In particular, the fermentability of arabinogalactan of RG I has been linked to an increase in the beneficial microflora (Lactobacillus; Bifidobacterium) and to the production of short chain fatty acids (Kelly, 1999).

There is thus a need to be provided an approach for the biogeneration of non-digestible oligosaccharides.

SUMMARY

The present disclosure provides a process for making a non-digestible mixture of oligosaccharides and oligomers of potato-derived rhamnogalacturonan I (RG I), the resulting non-digestible mixture as well as contemplated uses of the non-digestible mixture. The non- digestible mixture is obtained by hydrolyzing a source of potato-derived RG I with a multi- enzyme mixture preferentially comprising an arabinanase enzymatic activity and a galactanase enzymatic activity.

In a first aspect, the present disclosure provides a process for making a non-digestible mixture of oligosaccharides and oligomers from a potato pulp. Broadly, the process comprises (a) providing a source of potato-derived rhamnogalacturonan I (RG I), wherein the source of potato-derived RG I is a potato pulp or a RG l-enriched fraction obtained from the potato pulp; (b) hydrolyzing the source of potato-derived RG I with a multi-enzymatic mixture to yield an hydrolyzed mixture; and (c) isolating the non-digestible mixture from the hydrolyzed mixture. The non-digestible mixture comprises oligosaccharides having a degree of polymerization between 2 and 12 and oligomers having a degree of polymerization between 13 and 70. In an embodiment, the potato pulp is substantially free of starch or proteins and/or consists essentially of potato cell wall materials. In another embodiment, the process further comprises, prior to step (a), treating the potato pulp with an endo- polygalacturonase (for example, from Aspergillus n/ger) to obtain the RG l-enriched fraction. In still another embodiment, the process further comprises, prior to step (a), submitting the potato pulp to a microwave-assisted alkaline extraction to obtain the RG l-enriched fraction. In some embodiments, the microwave-assisted alkaline extraction comprises suspending the potato pulp in an alkaline solution (for example a solution of KOH) to obtain an alkaline potato pulp mixture and submitting the alkaline potato pulp mixture to a microwave treatment (for example wherein the power of the microwave treatment is between 20 and 220 W/g of the alkaline potato pulp mixture). In still another embodiment, step (c) of the further comprises precipitating or filtering (for example, using ultrafiltration) the hydrolyzed mixture to isolate the non-digestible mixture. In some embodiments, the multi-enzymatic mixture has a (low) arabinanase enzymatic activity and a (high) galactanase enzymatic activity. In an embodiment, the multi-enzymatic mixture comprises at least one of Gamanase and Depol. In another embodiment, the multi-enzymatic mixture comprises or consists of Gamanase and Depol. When both Gamanase and Depol are used, the ratio (U/U) of Gamanase and Depol enzymatic activity can be of about 1 :1 .

In a second aspect, the present disclosure provides a non-digestible mixture of oligosaccharides and oligomers obtained by the process described herein.

In a third aspect, the present disclosure provides a prebiotic composition or a food composition comprising the non-digestible mixture of described herein and a carrier.

In a fourth aspect, the present disclosure provides a process of extracting non-digestible oligosaccharides from a potato. Broadly, the process comprises (a) providing a potato pulp fraction free of starch and protein; (b) extracting rhamnogalacturonan content from said pulp fraction; (c) digesting the extracted rhamnogalacturonan content with a multi-enzymatic mixture to yield the non-digestible oligosaccharides from said extracted rhamnogalacturonan content; and ( d) isolating the non-digestible oligosaccharides. In an embodiment, the potato pulp fraction consist of cell wall materials. In still another embodiment, the rhamnogalacturonan content is extracted by treatment with endo-polygalacturonase (from A Niger for example) or by alkaline-microwave extraction. In an embodiment, the alkaline- microwave extraction comprises suspending cell wall materials in an alkaline solution (for example a solution of KOH 1 .5 M) and heating the suspended cell wall materials with microwave heating (using a microwave power of 90 W). In still another embodiment, the non- digestible oligosaccharides are isolated by filtration (for example by suing ultrafiltration trough a membrane, such as a glass microfiber filter grade GF/D). In another embodiment, the multi- enzymatic mixture comprises at least one of Galactanase and Depol. In still a further embodiment, the process further comprises the initial steps of (a1) washing potato pulp with a phenol-acetic acid-water solution, (a2) gelatinizing the starch of said potato pulp at 95°C and (a3) hydrolyzing by incubation with thermostable a-amylase (from Bacillus licheniformis for example) and amyloglucosidase (from A. niger for example) said starch to obtain a potato pulp fraction free of starch and protein. In yet another embodiment, the process further comprises the step of (a4) recovering cell wall materials from the potato pulp fraction free of starch and protein by filtration and freeze drying. In a fifth aspect, the present disclosure provides a non-digestible oligosaccharides preparation obtained by the process described herein.

In a sixth aspect, the present disclosure provides a prebiotic composition comprising the non- digestible oligosaccharides preparation described herein and a carrier.

In a seventh aspect, the present disclosure provides a food additive comprising the non- digestible oligosaccharides preparation described herein and a carrier.

BRIEF DESCRIPTION OF THE DRAWINGS

Reference will now be made to the accompanying drawings.

Figures 1A-B₃ illustrate the solubility of galactan, RG-enriched polysaccharides and HG (5% (w/v)) (A) and changes in sheer stress as function of sheer rate and viscosity as function of temperature for solutions containing 1 % of galactan (Bi), RG-enriched polysaccharides (B₂) and HG (B₃) at 25°C♦, 49°C■ and 73°C A and for water (25 °C) *. Chart equations at 25, 49 and 73°C were y=0.0004x ⁰⁹⁹⁷, y=0.0014x^{0 9603} and y=0.0014x^{0 9469} for galactan, y=0.001x ⁰²⁶⁵, y=0.01 x^{0 6462} and y=0.0022x^{0 9149} for RG-enriched polysaccharides, and y=0.0005x ⁰²⁴, y=0.0004x ⁰⁸⁷¹ , y=0.0009x^{0 9274} and y=0.0004x^{1 0822} for HG, respectively.

Figures 2A-F illustrate the particle size distribution for emulsions prepared at pH 4 (A), pH 6

(B) and pH 8 (C) containing galactan , RG-enriched polysaccharides and HG as emulsifier. Increase in particle diameter upon incubation of emulsions containing galactan (D), RG-enriched polysaccharides (E) and HG (F) for 2 h at 25, 49, 73 and 97°C. Figures 3A-G illustrate the HPAEC-PAD chromatograms of various hydrolysates from potato rhamnonogalacturonan I using selected biocatalysts: (A) Depol 670L, (B) Gamanase 1 .5L,

(C) logen HS 70, (D) Newlase II, (E) Pectinex Ultra SPL, (F) Viscozyme L and (G) endo-1 , 4- β-galactanase. Low (DP of 3-4) to medium (DP of 5-6) molecular weight oligosaccharides correspond to Peaks # 2-5 and 6-7, respectively.

Figures 4A-H illustrate the polymerization of various oligosaccharides from potato rhamnonogalacturonan I obtained under selected reaction conditions. Results are shown as the relative proportion of each class of oligosaccharides, in function of enzyme concentration (0.5, 4 or 7 U) and type of enzymes : (A) Depol 670L, (B) Gamanase 1 .5L, (C) logen HS 70,

(D) Newlase II, (E) Pectinex Ultra SPL, (F) Viscozyme L, (G) endo-1 ,4-p-galactanase and (H) a mixture of endo-1 ,5-a-arabinanase and endo-1 ,4-p-galactanase.

Figure 5 provides schematic representation of TIM- 1 system. 010: Stomach compartment, 020: Peristaltic valve, 030: Duodenum compartment, 040: Jejunum compartment, 050: Ileum compartment, 060: Pressure sensor, 070: Stomach secretion, 080: Duodenum secretion, 090: Jejunum secretion, 100: Ileum secretion, 1 10: Pre-filter, 120: Semi-permeable membrane, 130: Filtrate pump, 140: pH electrode, 150: Level sensor, 160: Temperature sensor, 170: Dosing port.

Figure 6 illustrates the schedule of the continuous fermentation steps, the numbers indicate days of the fermentation (BC: Beads colonization, RG I: Extracted polysaccharides from potato pulp using microwave-assisted alkaline treatment, Depol-based oligo-RG I: Generated oligosaccharide from potato RG I upon treatment with Depol 670L, Gal-based oligo-RG I: Generated oligosaccharide from potato RG I upon treatment with endo-p-1 ,4-galactanase, FOS: Fructo-oligosaccharide, 25% CH: Fermentation with 25% of carbohydrate source in the medium as a negative control).

DETAILED DESCRIPTION

The present disclosure provides an enzymatic approach for the biogeneration of non- digestible oligosaccharides/oligomers (galacto and galacto- (arabino)oligosaccharides/oligomers) from the pectic rhamnogalacturonan I (RG I) obtained from the potato (e.g., the potato pulp). The present disclosure also provides composition comprising such non-digestible oligosaccharides/oligomers as well as contemplated use of such non-digestible oligosaccharides/oligomers, especially as prebiotics and food additives.

Potato pectic polysaccharides show significance differences, in terms of composition and structural properties, depending on the source and extraction method. Such variations affect their application as a source of food ingredients, nutraceuticals and pharmaceutical molecules. Given its high galactan-rich rhamnogalacturonan I (RG I) content (75% of pectic polysaccharides) and low homogalacturonan (HG) content compared to other food processing by-products, potato (Solarium tuberosum L.) pulp represents a distinct pectin-rich by-product. Comparatively, in other sources such as citrus and apple, RG I accounts for only ~20% of pectic polysaccharides. A second distinguishing structural property of potato pectin is its high proportion of (1→4)-p-galactan side chains (67% of RG I) linked to the RG I backbone. High proportion of β-linked galactan side chains in potato makes it a good source for production of non-digestible oligosaccharides/oligomers. Galactan-rich RG I can be extracted from potato pulp which is a major and low-cost waste product of the potato- processing. However, because of the presence of linkages between cell wall components, the extraction of pectic polysaccharides being trapped in the cell wall structure, such that the pectin network, is challenging. Moreover, because of the complexity, the diversity and the multiple side chains of potato pectic polysaccharides, their controlled hydrolysis into well- defined oligosaccharides is challenging, requiring the synergistic actions of glycosyl- hydrolase combinations identified based on the molecular structure of parent polysaccharides. Process for making mixtures of rhamnogalacturonan-derived oligosaccharides and oligomers

The present disclosure provides processes and methods of making a non-digestible mixture of oligosaccharides and oligomers from a potato pulp. Potato pulp is a major waste product of the potato-processing industry. Approximately 0.75 kg of potato pulp is generated to produce 1.0 kg of starch. On a dry weight basis, the non-starch components of potato pulp include fibers (88.3%), proteins (6.3%), ashes (5.0%) and fat (0.4%). The moisture (8.7- 87.0%) and starch content (13.1-37.0% by dry weight) of potato pulp varies depending on the starch extraction method and any further drying process. Concerns about rising production costs and environmental pollution have led to considerable efforts have been made towards recovering and adding value to this waste. Accordingly, potato pulp has been used in a variety of ways: as cattle feed, in syrup production, as a substrate for alcohol and vitamin B₁₂ production, and for the generation of biogas. Although potato pulp contains considerable quantities of pectic polysaccharides, it remains largely unexploited because pectin extracted from potato has low gelling properties due to its high content of neutral sugars.

Potato pulp also comprises a large amount of rhamnogalacturonan I which, as shown herein, can be advantageously converted into oligosaccharides and oligomers. In the context of the present disclosure, the term "rhamnogalacturonan I" or "RG I" refers to a plant pectic polysaccharide having a backbone composed of alternating disaccharide of [→2)- a-L- rhamonose (Rha)-(1→ 4)-a-D-GalA]. The GalA residues in the backbone of RG I are partially methyl-esterified and O-acetylated at C-2 or C-3. Up to 80% of the Rha residues are substituted at 0-4 with different amounts of arabinan, galactan and arabinogalactan side chains or individual a-L-arabinose (Ara) and β-D-galactose (Gal) residues. Galactan side chains are polymers of (1→4) linked β-D-Gal, while arabinan side chains have a backbone of a-L-(1→5) Ara substituted with a-L-(1→3) Ara, a-L-(1→2) Ara or chains of a-L-(1→3) Ara. Arabinogalactan type I consist of a backbone of p-D-(1→4) Gal with a-L- Ara or arabinan branches attached to 0-3 of Gal residues, whereas arabinogalactan type II has a backbone of p-D-(1→3) Gal with β-6-linked galactan or arabinogalactan branches.

The methods and processes described herein are for making a mixture of non-digestible oligosaccharides and oligomers of the potato-derived RG I. As used in the present disclosure, the term "a mixture of non-digestible oligosaccharides and oligomers" refer to a mixture of at least one oligosaccharide and at least one oligomer obtained from the enzymatic degradation of RG I obtained from a potato pulp. In the context of the present disclosure, the term "oligosaccharide(s)" refers to an enzymatically-obtained fragment of the potato's RG I having a degree of polymerization between 2 and 12. In the context of the present disclosure, the term "oligomer(s)" refers to an enzymatically-obtained fragment of the potato's RG I having a degree of polymerization between 13 and 70. The degree of polymerization of a polymeric saccharide (such as an oligosaccharide and an oligomer) refers to the number of monosaccharide that the polymer contains. For example, an oligosaccharide will contain between 2 and 12 monosaccharide units while an oligomer will contain between 13 and 70 monosaccharide units. The term "oligosaccharides" refers to a plurality of single oligosaccharide each having a degree of polymerization between 2 and 12, whereas the term "oligomers" refers to a plurality of single oligomer each having a degree of polymerization between 3 and 70. The degree of polymerization can be determined by methods known in the art such as, for example, high performance size exclusion chromatography and/or mass spectrometry. The monomeric units of the oligosaccharides and of the oligomers can be the same or different. Monomeric units that can be present in the oligosaccharides and oligomers include, but are not limited to galactose, galacturonic acid, rhamnose, arabinose, glucose, xylose and/or mannose. In some embodiments, the weight percent of the galactose (Gal) units (with respect to the total weight of the polymeric saccharide) in the oligosaccharides and oligomers is at least 50%, 60%, 70%, 80% or 85%. In other embodiments, the weight percent of the galacturonic acid (GalA) units (with respect to the total weight of the polymeric saccharide) in the oligosaccharides and oligomers is at least 0.5%, 1 %, 2%, 5%, 10%, 15%, 20%, 25%, 30%, 35% or 40%. In still other embodiments, the weight percentage of each of the rhamnose (Rha), arabinose (Ara), glucose (Glc), xylose (Xyl) and mannose (Man) (with respect to the total weight of the oligosaccharides and oligomers) in the oligosaccharides and oligomers is less than 5%, 4%, 3%, 2% or 1 %.

The oligosaccharides and oligomers are considered non-digestible because they cannot be hydrolyzed by human gastrointestinal enzymes or under acidic conditions present in the stomach. Therefore, much like RG I, the oligosaccharides and oligomers described herein have a limited ability of being digested (e.g., capable of being absorbed via the gastrointestinal tract of a mammal). In some embodiments, less than 10% (in weight percentage) of the oligosaccharides and of the oligomers are digested.

The first step of the process is to provide a source of potato-derived RG I. In the processes described herein, it is preferred that RG I is highly branched with β-linked galactan side chains. In some embodiments, RG I is considered to be highly branched when it contains between 50 to 60% (in weight percentage) of galactan side chains. This can be achieved by either providing a potato pulp (preferably pre-treated so as to substantially remove its starch and proteins) or by providing a RG l-enriched fraction obtained from a potato pulp. As used herein, the expressions "rhamnogalacturonan l-enriched fraction" or "RG l-enriched fraction" refers to an isolated fraction of the potato pulp enriched for RG I (e.g., having a higher percentage in RG I than the potato pulp). The RG l-enriched fraction can be obtained, for example, by enzymatic treatment (e.g., endo-polygalacturonase treatment and, in some embodiments, by using an endo-polygalacturonase from Aspergillus n/ger) or by alkaline treatment or using a microwave-assisted alkaline treatment.

Once the source of RG I has been obtained or provided, it is submitted to an enzymatic treatment (e.g., enzymatic digestion or hydrolysis) to generate a mixture of oligosaccharides and oligomers. The enzymatic digestion is preferably performed by a multi-enzyme mixture having galactanase activity and arabinanase activity. In the context of the present disclosure, a "multi-enzyme mixture" refers to a combination of more than one enzymes capable of hydrolyzing bonds between two saccharide subunits and which are isolated from an organism. The term "multi-enzyme mixture" specifically excludes combination of enzymes independently purified and eventually mixed together. In some embodiments, the multi- enzyme mixture has a low arabinanase activity and a high galactanase activity. The presence of arabinanase is the enzyme mixture is important because it improves some debranching of the RG I to, ultimately, provide access of the galactanase to its substrate (to generate oligosaccharides and oligomers). However, the presence of arabinanase in the multi-enzyme mixture must be limited so as not to limit the excessive hydrolysis of RG I into low molecular weight oligosaccharides and monosaccharides. In other embodiments, the multi-enzyme mixture has a high ratio of galactanase-to-arabinanase activity (U/U). In an embodiment, the enzymatic ratio of the enzyme mixture is (galactanase activity)/(arabinanase activity) and is at least above 10, 1 1 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, 40, 50, 60, 70, 80 90, 100, 1 10, 120, 130, 140, 150 or 160 U/U, preferably not more than 160 U/U. The multi-enzyme mixture can optionally have pectinase and/or rhamnogalaturonase activity. In some embodiments, the enzymatic ratio (U/U) of the multi-enzyme mixture is :

(galactanase activity + arabinanase activity) / (rhamnogalacturonase activity) and is at least higher than 10, 1 1 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 31 , 32, 33, 34, 35, 40, 50, 60, 70, 80, 90, 100, 110, 120 U/U. The multi-enzyme mixture can be, for example, Depol (Depol 670L), Gamanase (Gamanase 1.5L), logen (logen HS 70), Newlase (Newlase II), Pectinex (Pectinex Ultra SPL), Viscozyme (Viscozyme L) as well as mixtures thereof. In some embodiments, the multi-enzyme mixture consists of a combination of Depol and Gamanase. In other embodiments, the multi-enzyme mixture consists of Depol.

In some embodiments, it is preferable or warranted to conduct the enzymatic hydrolysis/digestion of the RG l-enriched fraction so as to limit or prevent the formation of monomers (also referred to a monomeric units or monosaccharides). In such circumstances, it is possible to use the following multi-enzyme mixtures: Depol, Gamanase, Newlase, and/or Pectinex. Once the RG l-enriched fraction has been digested, the mixture of oligosacccharides and oligomers that have been generated are isolated from the fraction. In the context of the present disclosure, the expression "isolated" does not necessarily indicate that the mixture of oligosaccharides and oligomers are in a purified or substantially purified form. The expression "isolated" indicates that the mixture of oligosaccharides and oligomers are separated from some of the components of the RG l-enriched fraction, for example, intact or undigested RG I. This isolation can be achieved by precipitating the undigested/unhydrolyzed RG I so as to remove the polysaccharide from the mixture. This isolation can also be achieved by filtrating the RG I- enriched fraction so as to select oligosaccharides and oligomers having a specific range of molecular size (for example, having a molecular size of less than 10 kDa, preferably between 6 and 8 kDa). The filtration step can be conducted, for example, on a glass filter or on an ultrafiltration cartridge. The isolation step can optionally be followed by a freeze-drying step to further concentrate the mixture of oligosacccharides and oligomers. The isolation can also be performed by other methods known in the art, for example, size exclusion chromatography. In some embodiments, the isolation is performed using filtration (optionally ultrafiltration) and/or precipitation (preferably with ethanol).

Optionally, the process can comprise making the potato pulp. In the context of the present disclosure, the potato pulp is preferably provided as substantially devoid of starch and/or proteins. As used in the context of the present disclosure, a potato pulp "substantially devoid" or "substantially free" of starch and/or proteins refer to a potato pulp which has been treated so as to reduce its content in starch and/or proteins. In some embodiments, the starch content of the potato pulp is less than 1 % (weight percentage) and the protein content of the potato pulp is less than 0.1 % (weight percentage). In an embodiment, the potato pulp essentially consists of potato cell wall materials. In the context of the present disclosure, the expression "essentially consists of potato cell wall materials" indicates that more than 50% and/or less than 70% (in weight percentage) of the potato pulp consists of potato cell wall material. In some embodiments, the potato pulp comprise between 50% and 70% of potato cell wall materials (in weight percentage). In some embodiments, the potato pulp can be obtained by washing it, gelatinizing it (for example at a temperature of 95°C) and hydrolyzing it (for example with an enzymatic treatment using a thermostable a-amylase and an amyloglucosidase). The resulting mixture can be submitted to filtration and freeze-drying to generate a potato pulp substantially free of starch and/or proteins. Alternatively, a de- starched potato pulp can also be obtained commercially.

In addition, the process can comprise making the RG l-enriched fraction from the potato pulp. In such embodiment, the conditions for making the rhamnogalacturonan-enriched fraction from the potato pulp are adjusted so as to limit the debranching of RG I. In one embodiment, the RG l-enriched fraction is obtained by enzymatically digesting the potato pulp with an endo-polygalacturonase. The endo-polygalacturonase is preferentially provided in a substantially purified form (e.g., it is not provided in a multi-enzyme mixture). In some embodiments, the endo-polygalacturonase used to make the RG l-enriched fraction is obtained from Aspergillus niger. In an alternative or complementary embodiment, the RG I- enriched fraction is obtained by a microwave-assisted alkaline extraction. In such embodiment, the potato pulp is first submitted to an alkaline treatment (preferably by mixing the potato pulp with an alkaline solution, such as a KOH solution) and then submitted to a microwave treatment (preferably by submitting the alkaline potato pulp mixture to a microwave power between about 20 to 220 W/g of mixture). In some embodiments, the microwave treatment is at least about 20, 30, 40, 50, 60, 70, 80, 90, 100, 1 10, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210 or 220 W/g of potato pulp mixture and/or no more than 220, 210, 200, 190, 180, 170, 160, 150, 140, 130, 120, 1 10, 100, 90, 80, 70, 50, 40, 30 or 20 W/g of potato pulp mixture. Some embodiments of the microwave-assisted alkaline extraction method are provided in Example I. RG I can also be obtained using the methods presented in Khodaei, N., & Karboune, S. (2013) or Khodaei, N., & Karboune, S. (2014) . Such methods include, but are not limited to an alkaline (NaOH and KOH) or an enzymatic approach. RG-I is preferably obtained using an enzymatic approach (endo-polygalacturonase from A. n/ger).

Compositions comprising mixtures of rhamnogalacturonan-derived oligosaccharides and oligomers

The non-digestible mixture of oligosaccharides and oligomers described herein can be formulated as prebiotic compositions and/or food additive. As used herein, the term "prebiotic composition" means a non-digestible food ingredient that beneficially affects the host by selectively stimulating the growth and/or activity of one or a limited number of bacteria in the colon that can improve the health of the host. On the other hand, the "food additive" refers to a non-digestible food ingredient that is combined with another food ingredient. The food additive may or may not be considered a prebiotic composition.

The non-digestible mixture can be combined with a carrier which can be, in some embodiments, a "food carrier" or a "pharmaceutical carrier". In the context of the present disclosure, the term "carrier" is an acceptable solvent, suspending agent or any other inert vehicle for delivering the non-digestible mixture to a mammal. A "food carrier" is a vehicle added to food, while a "pharmaceutical carrier" is a vehicle used in a pharmaceutical composition. The carrier is typically liquid or solid. A pharmaceutical carrier is generally selected to provide for the desired bulk, consistency, flowability, etc., when combined with components of a composition, in view of the intended administration mode. Typical pharmaceutical carriers include, but are not limited to binding agents (e.g., pregelatinized maize starch, polyvinylpyrrolidone or hydroxypropyl methylcellulose, etc.); fillers (e.g., lactose and other sugars, microcrystalline cellulose, pectin, gelatin, calcium sulfate, ethyl cellulose, polyacrylates or calcium hydrogen phosphate, etc.); lubricants (e.g., magnesium stearate, talc, silica, colloidal silicon dioxide, stearic acid, metallic stearates, hydrogenated vegetable oils, corn starch, polyethylene glycols, sodium benzoate, sodium acetate, etc.); disintegrants (e.g., starch, sodium starch glycotate, etc.); and wetting agents (e.g., sodium lauryl sulphate, etc.). Additional suitable pharmaceutical excipients include starch, cellulose, talc, glucose, lactose, gelatin, malt, rice, flour, chalk, silica, magnesium stearate, sodium stearate, glycerol monostearate, sodium chloride, dried skim milk, glycerol, propylene glycol, water, ethanol, and the like.

In an embodiment, the non-digestible mixture is adapted for oral or anal delivery. The indigestible mixture can thus be administered orally (e.g., buccal cavity), sublingually and/or rectally (e.g., by suppositories or washings).

Useful carriers for the formulation of the non-digestible mixture are preferably solids or liquids; thus, the compositions can take the form of tablets, pills, capsules, syrups, creams, suppositories, liquid formulas (such as, for example, infant formula), powders, enterically coated or other protected formulations, sustained release formulations, solutions, suspensions, elixirs, aerosols, and the like.

The indigestible mixture (or its associated compositions/formulations) can be administered via any of the usual and acceptable methods known in the art, either singly or in combination with other agents (pharmaceutical or food). For example, the indigestible mixture can be formulated and/or administered in combination with other prebiotics, with probiotics, with antibiotics and the like. The indigestible mixture can be administered before or after the other agent(s). Alternatively, the indigestible mixture can be administered simultaneously with the other agent(s). The administration can be conducted in a single unit dosage form with continuous therapy or in a single dose therapy ad libitum.

In some embodiments, the indigestible mixture is administered when relief of symptoms is specifically required or perhaps imminent. Alternatively, the indigestible mixture is administered as continuous or prophylactic treatment. For example, the indigestible mixture can be administered prior to or shortly after the administration of antibiotics to favor microbial balance in the gastro-intestinal tract of the mammal.

Use of mixtures of rhamnogalacturonan-derived oligosaccharides and oligomers as prebiotics

Several studies have indicated health promoting effects of pectin derived compounds, such as the oligosaccharides and oligomers described herein. For example oligosaccharides obtained from apple pomace, orange and bergamot peel have been reported to enhance the growth of Bifidobacterium and Lactobacillus, while not supporting the growth of Bacteroides or Clostridium. Samples tested mainly originated from the cell wall's homogalacturonan (HG) region; however, neutral-sugar-rich carbohydrates have shown prebiotic properties as well. For example, fermentation of the highly branched arabinan-containing hairy region of pectin was shown to result in a greater quantity of SCFA, specifically propionate, as well as a lesser pH decrease, than that seen with commercial pectin.

The fermentability of the different oligosaccharides obtained from apple pomace in a medium inoculated with fecal samples was tested and it was reported that gluco-oligosaccharides, galacto-oligosaccharides and xylo-oligosaccharides were those most fermentable by intestinal bacteria followed by arabino-oligosaccharides and oligogalacturonides.

The structural properties of potato pectic polysaccharides containing galactan-rich RG I, make it an excellent source of non-digestible oligosaccharide (NDOs), in particular (arabino) galacto-oligosaccharides and galacto-oligosaccharides. Currently available commercial galacto-oligosaccharides are produced from lactose and therefore contain considerable amounts of Glc and lactose, making them non-consumable to consumers suffering from diabetes mellitus or lactose intolerance.

Potato pectic polysaccharides have been mainly recognized for their bioactive properties, in particular their fermentability and their bifidogenic activity. It was shown that enzymatically solubilized potato fibres are fermentable and their consumption increased end-expiratory H₂ levels. It has also been shown that consumption of soluble potato fibers obtained by enzymatic treatment of potato pulp decreased weight gain in rats. Having extracted high-MW polysaccharides (up to 400 kDa) from potato pulp using pectin lyase and polygalacturonase, it was found that the extracted polysaccharides, assumed to be mainly RG I with galactan side chains, promoted Bifidobacterium and Lactobacillus growth. The increase in number of the Bifidobacterium when RG l-rich and HG-rich fractions were used was higher than results obtained for FOS (DP 2-8 Orafti®P95) and RG l-rich fraction had better bifidogenic properties than HG-rich fraction. Moreover, potato pulp polysaccharides extracted through treatment with pectin lyase and polygalacturonase and treated with endo-1 ,4-p-galactanase produced two fractions of <10 and > 10 kDa. It was reported that both fractions promoted the growth of Bifidobacterium longum and Lactobacillus acidophilus and generally did not support the growth of Clostridium perfringens. Stimulation of B. longum with hydrolyzed samples was significantly better than with FOS, Gal or unhydrolyzed galactan.

As such, the non-digestible mixture and associated compositions and preparations can be used as prebiotics in mammals (for example in humans). The non-digestible mixture and associated compositions and preparations can be used to favor Bifidobacterium and/or Lactobacillus growth in the gastro-intestinal track of the mammal. Alternatively or complementarily, the non-digestible mixture and associated compositions and preparations can be used to limit or impede Bacteroides and/or Clostridium growth in the gastro-intestinal track of the mammal. The non-digestible mixture can be used, for example, to decrease weight, to limit weight gain, to limit or prevent the adhesion of pathogenic microorganisms to the gastro-intestinal tract wall, to limit or prevent tumor growth and/or to enhance or stimulate the immune system.

EXAMPLE I - MICROWAVE-ASSISTED ALKALINE EXTRACTION OF GALACTAN-RICH RHAMNOGALACTURONAN I FROM POTATO CELL WALL BY-PRODUCT

Materials. Potato galactan was obtained from Megazyme (Wicklow, Ireland). Soybean oil, KOH, sodium acetate, acetic acid, dextran standards, Gal, Ara, Rha, Glc, Xyl, Man, GalA, trifluoroacetic acid, thermostable a-amylase from B. licheniformis, amyloglucosidase from A. niger and HG from oranges were purchased Sigma Chemical Co. (St-Louis, MO). All salts were obtained from Fisher Scientific (Mississauga, ON).

Preparation of potato cell wall material. In order to prepare cell wall material, protein was removed from the potato pulp (Lyckeby Starch AB) as previously described (Khodaei & Karboune, 2013). Starch in the potato pulp was gelatinized at 95°C for 30 s and hydrolyzed upon treatment with thermostable a-amylase from B. licheniformis and amyloglucosidase from A. niger (>300 units/ml) at a ratio of 1 : 10 and 1 :2.3 (v:w) respectively. Cell wall material was recovered by filtration and freeze dried.

Microwave-assisted alkaline extraction of RG I from cell wall. Microwave-assisted alkaline extraction of RG I was optimized using RSM and a CCRD. Concentration of KOH (0, 0.5, 1 .0, 1 .5, and 2 M), extraction time (0, 2, 4, 6, and 8 min), microwave power (14, 32, 50, 68, and 86 W) and solid/liquid ratio (0.2, 1 .1 , 2.0, 2.9, and 3.8%, w/v) were chosen as independent variables. The volume of liquid was maintained constant at 30 ml. As a result, power to substrate ratio varied between 23.3 and 833.3 W/g cell wall. The complete design (30 experiments) includes sixteen factorial points (levels ±1), eight axial points (levels ± a), and six replicates in center point, and the experiments were run in random order. The yield (%, w/w), the MW distribution (>600 kDa, %) and the monosaccharide profile (Gal, Ara, Rha, Glc, Xyl, Man, GalA) of extracted polysaccharides were the quantified as responses.

Isolated cell wall material was suspended in KOH solution to yield the selected solid/liquid ratio. The suspension was prepared in a cylindrical Pyrex tube and microwave heated for selected extraction time under a specific power using a Prolabo Synthewave 402 Microwave Reactor. After incubation, samples were filtered under vacuum using glass microfiber filter grade GF/D with pore size of 2.7 μηι (Whatman^®). pH of filtrate was adjusted to 7 using hydrochloric acid, dialyzed against water at 4°C using membranes with a 6-8 kDa cut-off and freeze dried. Yield (%, w/w) was calculated as the amount of recovered extracts (before dialysis) over the initial amount of cell wall material, multiplied by 100. Monosaccharide composition and MW distribution of extracted polysaccharides were also analyzed.

Statistical analysis. Analysis of the experimental results and calculation of predicted responses were carried out using Design Expert software (Version 8.0, Stat-Ease, Inc. Minneapolis, MN, USA). The ANOVA of the model, including lack of fit, Fisher's F-test, its associated probability and correlation coefficient was investigated. The variations of the responses [yield (%, w/w), MW distribution (>600 kDa, %) and monosaccharide profile (μηιοΙ/g extract)] as function of the independent variables, including concentration of KOH (x-i), extraction time (x₂), power (x₃) and solid/liquid ratio (x₄), were explained by the following empirical quadratic polynomial equation (Eq. 1.1):

^ = £_e H-∑A¾ -∑¾ ¾£f -∑&.¾² ₍u₎

The coefficients of the polynomial equations were represented by βα (constant term), β, (linear coefficient), β„ (quadratic coefficient), and β_/, (interaction coefficient). Y represents the response and x, and x represent independent variables. Response surfaces were developed using fitted polynomial equations for most significant interaction terms.

Sugar content and monosaccharide composition analyses. Uronic acid and total sugar contents were measured using sulfamate/m-hydroxydiphenyl and phenol-sulfuric acid colorimetric assays, respectively.

In order to determine the monosaccharide composition, extracted polysaccharides were first hydrolyzed using a two-step procedure as previously described by Khodaei & Karboune, 2013. Extracted polysaccharides were incubated at 60°C for 24 h in HCI/methanol mixture (1 :4, v/v) at a ratio of 0.6% (w/v) and thereafter boiled for 1 h in trifluoroacetic acid solution at a ratio of 1 :8 (v/v). Hydrolyzed samples were analyzed with HPAEC-PAD (Dionex), and a CarboPac PA20 column (3 ^χ 150 mm). Isocratic elution was performed with 5 mM NaOH (0.5 ml/min). L-Rha, L-Ara, D-Gal, D-Glc, D-Xyl and D-Man were used as standards.

MW distribution. MW distribution of the extracted polysaccharides was investigated by HPSEC on three columns in series (TSK G3000 PWXL, TSK G4000 PWXL and TSK G5000 PWXL, Tosoh Bioscience, Montgomeryville, PA) connected to a Waters HPLC system equipped with refractive index detector and Breeze software for the data analysis. NaCI (0.1 M) was used as mobile phase at a flow rate of 0.5 ml/min and 30°C. Dextrans with MW of 50, 150, 270, 410 and 670 kDa were used as standards for calibration. Isolation and characterisation of RG l-enriched fraction. The extracted polysaccharides were subjected to size exclusion chromatography on Superdex 200 16/600 (GE Healthcare, NJ) column connected to AKTA purifier system (GE Healthcare). The isocratic elution with 0.05 M sodium acetate at pH 5 containing 0.2 M NaCI at a flow rate of 1 ml/min was used. The fractions were analyzed using phenol-sulfuric acid colorimetric assay. The recovered RG I- enriched fraction was analyzed for their monosaccharide content.

Assessment of functional properties of RG l-enriched polysaccharides. Extracted RG I- enriched polysaccharides were assessed for their functional properties as compared to potato galactan and orange HG as standards.

Solubility. To determine solubility, 5% (w/v) suspensions of RG l-enriched polysaccharides, potato galactan and oranges HG, in 0.1 M sodium acetate (pH 6), were prepared. Samples were incubated at 25, 42, 59, 76 and 93°C for 2 h, and mixed at room temperature for 1 h. After centrifugation at 14,000 ^χ g for 15 min, the amount of total sugars in the supernatants was measured using phenol-sulfuric acid colorimetric assay. Solubility was expressed as the amount of total sugars in the recovered supernatant over its initial concentration, multiplied by 100.

Viscosity. Viscosity of solutions containing 1 % (w/v) RG l-enriched polysaccharides, potato galactan or orange HG was measured at 25, 49 and 73 °C using an AR2000 controlled- stress rheometer (TA, Crawley, U.K.) equipped with 60 mm acrylic parallel plate. Samples were heated for 2 h at selected temperatures before analysis and the shear rate varied between 1 and 100 s Ostwald-de Waele equation (Eq. 1.2) was used in order to define the flow behavior of the liquids.

In which 3u/5y is shear rate (s ), τ is shear stress (Pa), K is the flow consistency index (Pa.sⁿ) and n is flow behavior index.

Emulsifying properties. In order to assess emulsifying properties, 0.2% (w/v) solutions of RG l-enriched polysaccharides, potato galactan and orange HG were prepared and adjusted to selected pH values of 4, 6 and 8. Polysaccharide solutions were mixed with soybean oil (5%, v/v) and passed three times through a homogenizer (EmulsiFlex-C5, Avestin Inc.) at operational pressure of 5000 - 7000 kPa. Droplet size was measured using Delsa™Nano Submicron Particle Analyzer (Beckman Coulter). Measurements were performed at 25°C immediately after homogenization and after 2 h of incubation at 25, 49, 73 and 97°C. Droplet size distribution was expressed by relative volume (v/v, %) of particles of each size in sample. In addition, the extent of droplet size increase was estimated as the droplet diameter (nm) after incubation over the value obtained before incubation. The droplet size distribution and the extent of droplet size increase in emulsions, prepared with potato galactan and as standards, were also measured.

Yield, MW and monosaccharide profile of extracted polysaccharides. RSM was used in order to investigate the effects of selected extraction parameters, including X -concentration of KOH (M), x₂-extraction time (min), x₃-power (W) and x₄-solid/liquid ratio (%, w/v), on the yield of recovered polysaccharides (%), the proportion of high-MW polysaccharides (>600 kDa, %) and the monosaccharide profile of isolated polysaccharides. The extraction of galactan-rich RG I was monitored by assessing Gal, Ara and Rha contents in the polysaccharidic extracts. While Glc, Xyl and Man monosaccharide contents were used as indicators for the extraction of hemicellulosic polysaccharides as side-products. In addition, the extent of the hydrolysis of HG region under alkaline condition was evaluated by quantifying the GalA content in the polysaccharidic extracts (Table 1). Alkaline extraction of cell wall is expected to isolate intact RG I through extensive hydrolysis of HG region by β-elimination and oxidative peeling. The experimental design was performed based on the CCRD. The range and center point values of these parameters were set based on preliminary trials. The investigated solid/liquid ratios (0.2-3.8%, w/v) were within the same range (2-4%) as most of previous studies on microwave-assisted extraction of polysaccharides; however, some have used higher solid/liquid ratios of 6.3 and 8.0%, respectively, for the microwave-assisted extraction of pectin. On the other hand, the investigated extraction times (0-8 min) were shorter than those reported in other studies (1-30 min); and the power values (14-86 W, corresponding to 23.3- 833.3 W/g cell wall) were within the same range reported for the extraction of pectic polysaccharides (15 to 1200 W/g).

Table 1 summarizes the results of the yield (%, w/w), the proportion of high-MW polysaccharides (>600 kDa, %) and the monosaccharide profile upon selected microwave- assisted alkaline extractions. The highest yield of 52.0% was achieved upon treatment n° 8, in which the concentration of KOH, time and power were at + 1 level of design (1.5 M, 68 W, 6 min); while solid/liquid ratio was at -1 level of design (1.1 %) (Table 1). Therefore, harsher extraction conditions of treatment n° 8, in terms of power and extraction time, combined with low solid/liquid ratio (corresponding to high power/substrate ratio of 206.1 W/g cell wall) may have led to an efficient opening of the cell wall network and hence to a higher yield.

However, the low total monosaccharide content (3,889.8 μηιοΙ/g) of the polysaccharidic extract and the low relative molar proportion of Gal (44%) associated with the high molar proportion of GalA (28%) obtained with treatment n° 8, reveal the extensive defragmentation . Table 1 CCRD design and responses for yield (%), high molecular weight polyss kDa, %) and monosaccharide composition (μηιοΙ/g). x₂-

Std x₄-Solid/liquid^a 00 kDa Gal Ara Rha Glc Man GalA

KOH Time Power Power/solid Yield >6

Xyl (μηιοΙ/g)

order (%) (%) (%) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g)

(M) (min) (W) (W)/g

1 0.5 2 32 1.1 97.0 28.8 57.0 2590.0 (55) 262.2 (6) 86.2 (2) 257.3 (5) 51 .9 (1) 73.2 (2) 1392.7 (30)

2 1.5 2 32 1.1 97.0 30.3 1.0 3997.0 (60) 432.8 (6) 103.5 (2) 739.1 (1 1) 315.2 (5) 237.4 (4) 851 .1 (13)

3 0.5 6 32 1.1 97.0 32.4 40.9 2270.0 (53) 246.0 (6) 66.0 (2) 32.1 (1) 84.0 (2) 22.4 (1) 1537.2 (36)

4 1.5 6 32 1.1 97.0 39.9 7.1 2289.0 (47) 296.0 (6) 148.0 (3) 670.0 (14) 180.0 (4) 182.6 (4) 1148.8 (23)

5 0.5 2 68 1.1 206.1 37.6 20.8 3194.4 (61) 230.4 (4) 113.3 (2) 141 .7 (3) 51 .0 (1) 86.9 (2) 1454.0 (28)

6 1.5 2 68 1.1 206.1 39.6 6.3 3396.4 (64) 300.5 (6) 67.0 (1) 396.6 (8) 168.3 (3) 115.0 (2) 835.0 (16)

7 0.5 6 68 1.1 206.1 41 .9 13.1 3245.9 (59) 266.0 (5) 174.5 (3) 145.0 (3) 23.5 (0) 171 .7 (3) 1468.3 (27)

8 1.5 6 68 1.1 206.1 52.0 11 .3 171 1.0 (44) 240.0 (6) 144.0 (4) 443.4 (1 1) 110.0 (3) 161 .7 (4) 1079.7 (28)

9 0.5 2 32 2.9 36.8 33.4 27.9 2963.0 (57) 130.0 (3) 164.0 (3) 196.1 (4) 31 .9 (1) 77.2 (1) 1615.4 (31)

10 1.5 2 32 2.9 36.8 27.4 8.0 5381.8 (65) 402.7 (5) 188.6 (2) 951 .4 (1 1) 355.0 (4) 251 .8 (3) 776.7 (9)

11 0.5 6 32 2.9 36.8 31 .6 10.5 3173.6 (58) 226.6 (4) 121 .7 (2) 182.7 (3) 21 .8 (0) 107.7 (2) 1602.0 (29)

12 1.5 6 32 2.9 36.8 37.4 7.8 3807.9 (58) 365.5 (6) 180.3 (3) 983.9 (15) 226.0 (3) 241 .9 (4) 802.6 (12)

13 0.5 2 68 2.9 78.2 31 .2 8.4 3326.3 (57) 223.6 (4) 179.0 (3) 283.0 (5) 24.2 (0) 191 .9 (3) 1614.2 (28)

1 1.5 2 68 2.9 78.2 20.2 6.7 4558.0 (58) 513.0 (6) 107.0 (1) 1241.7 (16) 437.0 (6) 176.2 (2) 865.3 (1 1)

15 0.5 6 68 2.9 78.2 27.6 1.5 4482.0 (60) 392.4 (5) 215.8 (3) 638.0 (9) 37.4 (1) 436.7 (6) 121 1.5 (16)

16 1.5 6 68 2.9 78.2 30.1 20.9 3424.9 (46) 629.6 (8) 147.8 (2) 1818.4 (24) 370.0 (5) 408.5 (5) 692.3 (9)

17 0.0 4 50 2.0 83.3 37.8 23.0 2887.6 (46) 214.0 (3) 172.3 (3) 288.8 (5) 9.1 (0) 128.1 (2) 2556.2 (41)

18 2.0 4 50 2.0 83.3 42.1 2.5 3605.0 (48) 423.0 (6) 154.0 (2) 1327.6 (18) 432.16 (6) 332.7 (4) 1200.0 (16)

19 1.0 0 50 2.0 83.3 29.5 16.7 3106.0 (77) 207.0 (5) 59.0 (1) 59.5 (1) 103.0 (3) 22.2 (1) 473.0 (12)

20 1.0 8 50 2.0 83.3 41 .0 8.3 2266.0 (53) 394.4 (9) 140.6 (3) 464.1 (1 1) 75.9 (2) 153.3 (4) 762.6 (18)

21 1.0 4 1 2.0 23.3 38.9 28.5 2831.0 (57) 142.6 (3) 67.7 (1) 152.6 (3) 43.2 (1) 57.6 (1) 1633.6 (33)

22 1.0 4 86 2.0 1 3.3 39.9 13.7 3175.5 (63) 268.2 (5) 135.3 (3) 238.8 (5) 47.7 (1) 160.5 (3) 982.2 (20)

23 1.0 4 50 0.2 833.3 23.7 24.3 2763.0 (57) 342.9 (7) 131 .0 (3) 356.2 (7) 94.9 (2) 144.4 (3) 983.0 (20)

2 1.0 4 50 3.8 3.9 15.5 12.9 6217.4 (60) 597.9 (6) 281 .9 (3) 1706.6 (16) 189.2 (2) 349.3 (3) 1079.4 (10)

25 1.0 4 50 2.0 83.3 32.6 10.5 3616.9 (52) 473.0 (7) 217.0 (3) 1098.4 (16) 164.0 (2) 305.3 (4) 1015.0 (15)

26 1.0 4 50 2.0 83.3 29.7 16.0 4052.0 (54) 476.0 (6) 218.0 (3) 1221.1 (16) 201.6 (3) 299.5 (4) 1014.0 (14)

27 1.0 4 50 2.0 83.3 27.9 13.0 4061.0 (56) 519.1 (7) 193.1 (3) 1100.0 (15) 164.0 (2) 310.6 (4) 878.7 (12)

28 1.0 4 50 2.0 83.3 30.0 10.8 4255.5 (58) 421 .8 (6) 183.8 (3) 1008.3 (14) 161.6 (2) 274.5 (4) 1012.0 (14)

x₂-

Std x₄-Solid/liquid^a >600 kDa

ower/solid Yield Gal Ara Rha Glc Man GalA

KOH Time Power P Xyl (μηιοΙ/g)

order (%) ιοΙ/g) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g) (μηιοΙ/g)

(M) (min) (W) (W)/g (%) (%) (μη

29 1.0 4 50 2.0 83.3 28.6 11 .5 4049.0 (54) 478.0 (6) 220.0 (3) 1090.0 (15) 129.1 (2) 304.0 (4) 1176.2 (16) 30 1.0 4 50 2.0 83.3 31 .2 12.4 4060.0 (54) 474.9 (6) 217.0 (3) 1267.3 (17) 146.2 (2) 322.7 (4) 1050.5 (14) ^a Constant extraction volume of 30 ml was used; ^b Relative molar proportion of the monosaccharides (%).

of extracted polysaccharides and the debranching of their neutral side chains (estimated at 37.8% of total extracted polysaccharides) under harsher extraction conditions; indeed, GalA/Rha and Gal/Rha ratios at treatment n° 8 were only 7.5 and 1 1.9, respectively (Table 1).

The results also show that with treatment n° 7 (68 W, 1.1 %, 6 min) in which lower KOH concentration (0.5 M) was used as compared to treatment n° 8, the yield and the defragmentation/debranching of extracted polysaccharides were less significant, estimated at 41.9% and 12.1 %, respectively, with GalA/Rha and Gal/Rha ratios of 8.4 and 18.6, respectively. On the other hand, treatment n° 20 (1.0 M KOH, 50 W, 2.0%, 8 min), with longer extraction time and lower KOH concentration than that of n° 8, resulted in a lower yield (41.0%), but similar defragmentation/debranching of extracted polysaccharides (31.9%) with GalA/Rha and Gal/Rha ratios of 5.4 and 16.1 , respectively. These results show the significant effects of extraction time and concentration of KOH on the yield of extracted polysaccharides and their defragmentation/debranching. However, the use of milder extraction conditions, corresponding to treatment n° 1 (0.5 M KOH, 32 W, 1.1 %, 2 min), resulted in the highest proportion of high-MW polysaccharides (>600 kDa, 57.0%) and in GalA/Rha and Gal/Rha ratios of 16.0 and 30.0, respectively; these results reveal the lower break down of polysaccharides and the limited debranching of its neutral side chains under milder conditions. Comparing treatments n° 2/4 and 10/12 reveals the significant effect of the extraction time (2 to 6 min), under constant microwave power, on the debranching of galactan side chains ; indeed, the Gal contents at treatments n° 2 and 10 (3,997.0-5,381.8 μηιοΙ/g extract) were higher than at treatments n° 4 and 12 (2,289.0 - 3,807.9 μηιοΙ/g extract). The highest amount of Gal and Rha (6,217.4 and 281.9 μηιοΙ/g extract, respectively; 60 and 3%) and relatively high amount of Ara (597.9 μηιοΙ/g extract, 6 %) were obtained upon treatment n° 24, in which the highest solid/liquid ratio of 3.8% (w/v), power of 50 W (corresponding to 44 W/g cell wall) as well as moderate incubation time of 4 min and KOH concentration of 1.0 M were used. The high isolation of galactan-rich RG I at treatment n° 24 was accompanied by the high recovery of the hemicellulose polysaccharides (Glu+Xyl+Man, 2245.1 μηιοΙ/g), representing 22% of total sugars, and the moderate debranching of galactan side chain (Gal/Rha of 22.1). The highest Gal/Rha ratios of 50.7-52.6 were obtained upon treatments n° 6 and 19 characterizing by shorter extraction time of 0-2 min and higher power (83.8-206.1 W/g cell wall). As shown by the hemicellulosic monosaccharide contents (Glc, Xyl and Man), the combined use of high KOH concentration (≥ 1 M) and high solid/liquid ratio (≥ 2%) (treatments n° 10, 14, 16, 18, 24-30) favored the isolation of hemicellulosic polysaccharides (1 ,567.7-2,596.9 μηιοΙ/g; 19-35%). On the other hand, treatment n° 17, in which only water was used, led to the highest amount of GalA (2,556.2 μηιοΙ/g extract; GalA/Rha of 14.8), indicating the importance of alkaline conditions for the removal of HG region through β-elimination reaction. When concentration of KOH was increased to 0.5 M (treatments n° 1 , 3, 5, 7, 9, 11 , 13, 15) and 1.0 M (treatments n° 19-30), the content of GalA decreased to 1 ,211.5-1 ,615.4 μηιοΙ/g extract (16-36%) and 473.0 - 1633.6 μηιοΙ/g extract (10-33%), respectively. However, no significant changes in the GalA contents (692.3 - 1 , 148.8 μηιοΙ/g extract; 9-28%) was observed as a result of further increase of KOH concentration to 1.5 M (treatments n° 2, 4, 6, 8, 10, 12, 14, 16).

Mathematical models were fitted to the experimental data using Design-Expert software version 8.0.2. and significance and adequacy of the models were tested with ANOVA (Table 2). Neglecting the insignificant terms by backward elimination regression (a =0.05), ANOVA results indicate that the quadratic model was statistically significant for the description of the variations of both yield (F-value of 27.2 and p-value of < 0.0001) and the proportion of high-MW polysaccharides (F-value of 26.7 and p-value of < 0.0001). For these two responses, lack of fit was not significant relative to pure error with F value of 1.5 and 3.3, respectively. The significance of each coefficient was determined using the F value and P value as shown in Table 2. As expected, solid/liquid ratio (x₄, F value of 67.4; P value of <0.0001) and extraction time (x₂, F value of 48.3; P value of <0.0001) were the most significant linear terms for the extraction yield. High extraction yield is expected to be achieved with an increase in the incubation time. The results also show that the concentration of KOH (x^ F value of 92.6; P value of <0.0001) and power (x₃, F value of 40.7; P value of <0.0001) were the most significant linear terms for the proportion of high-MW polysaccharides. In terms of quadratic terms, solid/liquid ratio (x , F value of 45.3; P value of <0.0001) and concentration of KOH (x , F value of 44.7; P value of <0.0001) were the most important variables for yield, whereas power (x₃ ², F value of 1 1.4; P value of 0.0042) was the most significant for the proportion of high-MW polysaccharides.

Table 2 The analysis of variance for response surface model for yield, molecular weight distribution and monosaccharide content (Gal, Ara, Rha and GalA).

Yield (%) ^a >600 kDa (%) ^a Gal ^mol/g)^a Ara ^mol/g)^a Rha ^mol/g)^a GalA ^mol/g)^a

F Value P Value F Value P Value F Value P Value F Value P Value F Value P Value F Value P Value

27.2 < 0.0001 26.7 < 0.0001 25.1 < 0.0001 20.5 < 0.0001 22.6 < 0.0001 22.6 < 0.0001

Model

4.6 0.0483 92.6 < 0.0001 13.3 0.0024 65.3 < 0.0001 0.8 0.3873 164 < 0.0001 xrKOH (M)

48.3 < 0.0001 6.3 0.0238 26.3 0.0001 7.3 0.0164 19.6 0.0005 1.5 0.2431

X2-Time (min)

4.6 0.048 40.7 < 0.0001 1 .4 0.251 11 .7 0.0038 8 0.0127 9.4 0.0078 x₃-Power (W)

X4-Solid/liquid 67.4 < 0.0001 31.6 < 0.0001 138.5 < 0.0001 31 .2 < 0.0001 78.1 < 0.0001 0.4 0.5147

(%)

Xf X2 24.5 0.0002 32.3 < 0.0001 45.8 < 0.0001 6 0.0266 3.3 0.0878 1.8 0.1956

0.4 0.5168 77.5 < 0.0001 28.1 < 0.0001 0.1 0.713 37.7 < 0.0001 0.4 0.553

Xf X3

14.2 0.0019 61.7 < 0.0001 8.7 0.01 16.9 0.0009 1.5 0.2417 4 0.0627

Xi X4

0 0.8442 6.2 0.0248 2.8 0.1 15 4.7 0.0471 13.9 0.002 2.6 0.1301

X2 X3

3.7 0.0725 0 0.8488 4.8 0.0454 10.2 0.0061 4.3 0.0547 6.9 0.0191

X2 X4

57.7 < 0.0001 8.5 0.0106 0 0.9535 25.9 0.0001 2.4 0.1445 0.4 0.5157

X3 X4

44.7 < 0.0001 0 0.9937 14.2 0.0019 25 0.0002 15.1 0.0015 82.8 < 0.0001

X1²

12.8 0.0028 0 0.9034 42.6 < 0.0001 31 < 0.0001 80.8 < 0.0001 21 .1 0.0003

X2²

39.6 < 0.0001 1 1.4 0.0042 24.6 0.0002 74.2 < 0.0001 78.3 < 0.0001 8.4 0.0109

X3²

45.3 < 0.0001 5.5 0.0331 5.6 0.0324 0 0.8908 0.2 0.6993 0 0.9183

X4²

1 .5 0.3438 3.3 0.0986 1 .9 0.2521 2.1 0.209 1.1 0.4746 1.9 0.2476

Lack of Fit

a R² = 0.93, 0.93, 0.92, 0.90, 0.91 and 0.91 for yield, molecular weight population, Gal, Ara, Rha and GalA respectively. P < 0.05 indicates statistical significance.

Among all interactive effects, interaction between power and solid/liquid ratio (x₃ x F value of 57.7; P value of < 0.0001) seems to be the most significant in the yield model, and the negative sign of its term (Eq. 1.3) reveals its antagonistic effect. Interaction between concentration of KOH and power (x_? x₃, F value of 77.5; P value of < 0.0001) was the most significant in the high-MW polysaccharides model and was of synergistic type as demonstrated by the positive sign of this interactive term (Eq. 1.4). The fitted quadratic model for yield and high-MW polysaccharides are given by Equations 1.3 and 1.4 using coded variable names. Insignificant terms were dropped except for those that were necessary for maintaining the model's hierarchy.

Yield (%) = 30.0 -t- Q$Xi + Z.8x₃ -t- Q.9 ₃ - 3.3x_t +- 2.£ i%₂ ~ l.¾_i¾ - 3.8¾χ₄ λ- Ζ, ² λ- 1Αχ₂ ²λ- ΖΛχ ² - Z&x ² | ₃\

> 6QQ kDa (%) = 13.1— 7.Zx_x— 1.83_₂— 4.3χ₃ - 3.2¾ -t- 4.1κ₁3ε₃ -t- 6.9x₁x_s +- 6. Is,! 3,4 -t- .2¾₃¾₄ -t- . 18 s,₃ ² +^■ 1.& ¾²

(1.4)

The results of ANOVA for the monosaccharide contents of extracted polysaccharides are shown in Table 2. Quadratic model was the most significant for Gal (F-value of 25.1 and p- value of < 0.0001), Ara (F-value of 20.5 and p-value of < 0.0001), Rha (F-value of 22.6 and p-value of < 0.0001) and GalA (F-value of 22.6 and p-value of < 0.0001) contents. Lack of fit was not significant with F value of 1.9, 2.1 , 1 .1 and 1 .9 for Gal, Ara, Rha and GalA contents respectively, suggesting a good fit. Solid/liquid ratio was the most significant linear term in the Gal (x₄, F value of 138.5; P value of <0.0001) and Rha (x₄, F value of 78.1 ; P value of <0.0001) content models. In the Ara (xi , F value of 65.3; P value of <0.0001) and GalA (xi , F value of 164.0; P value of <0.0001) content models, concentration of KOH was the most significant linear term. The variable with the largest quadratic effect on Gal (x_? ², F value of 42.6; P value of <0.0001) and Rha (x,², F value of 80.8; P value of <0.0001) contents was the extraction time; while for Ara (x , F value of 74.2; P value of <0.0001) and GalA (x ², F value of 82.8; P value of <0.0001) contents, power and concentration of KOH had the most significant quadratic effect respectively. The most important interactions for the Gal, Ara, Rha and GalA contents were the ones between KOH concentration/ time (x_? x_2, F value of 45.8; P value of < 0.0001), power/ solid to liquid ratio (x₃ x_4, F value of 25.9; P value of < 0.0001), KOH concentration/Power (x_? x F value of 37.3; P value of < 0.0001) and time/ solid to liquid ratio (x₂ x F value of 6.9; P value of 0.0191), respectively. These interactive effects were of antagonistic types in the Gal, Rha and GalA content models (Eq. 1.5, 1.7 and 1.8), whereas that of Ara content model (Eq. 1.6) exhibited a synergistic effect. Quadratic models were simplified by removing terms which were not statistically significant and can be described by the following equations using coded terms.

Gal (timol/g) = 4Q1&.7 - 198.23-_! - 278.4¾ - 64.8¾ - 63&9fc_t - 449.9a_-13.₂ - 352.3:£ι¾ +^■ +^■ 14&.lx₂x₄ - 191.23._!* - 331.3x₂ ²- Z&2.0¾² 4- 119.7s.

(1.5)

Am (μ-riQl/g) = 472.6 4- &7.53_i +^■ 22. &¾ + 28.5x,₃ + 4&.&s,₄ - 25.23.₁3_₃ 4- .1*1 ¾ +- 22.1¾x₃ 4- 32.7x.₂3.₄ 4- 52.1¾ι₄ - 38.9 - 43.4 ¾² - 67.2 ¾²

(1.6)

Rha (fimol/g) = 2G&.7 - 3-.Q¾i -I- 14.7x₂ -I- 9.4x₃ -I- 29.3x₄ - 2 . x.iX3 +" 15.1x₂¾ - 11. 3-!²- 27.7¾² - 27.3x₃ ²

(1.7)

GalA (μίΏϋΙ/g) = 1021. - 31 .8X_! 4- 29.9¾ - 7S.4x₃ - 16.4¾j - 9.QX2X* 4- 2Q9.7 Xi ³ - 105.4 x₂ ² -t" 67.1 x₃ ³

(1.8)

These results indicate that the Gal content or the debranching of galactan side chains can be modulated by compromising between the concentration of KOH (x-i) and the extraction time (x₂), which exhibited an adverse interaction in the Gal content model. The concomitant increase of solid/liquid ratio (x₄) and power (x₃) is expected to accelerate the physical rupture of cell wall and increase concentration of arabinan side chains released into the liquid phase. Rha content, which is found in the backbone of RG I, is mainly affected by concentration of KOH (M) and power (W); indeed, increasing power and KOH concentration at the same time accelerates hydrolysis of RG backbone resulting in lower amount of Rha. Although the KOH concentration exhibited the most important linear effect on the GalA content, it didn't show any statistically significant interactive effects. However, a good compromise between solid/liquid ratio (x₄) and extraction time (x₂) needs to be considered to achieve an efficient removal of HG region.

Effects of microwave assisted-alkaline extraction parameters. 2D contour plots (data not shown) generated from the predicted model of yield (%, w/w) at KOH concentration of 1.5 M (a factorial point) show the interactive effect of solid/liquid ratio and power upon selected extraction times. The extraction yield increases as the power increases and the solid/liquid ratio decreases; and increasing extraction time resulted in higher yields. The effect of power on the yield can be attributed to the efficient rupture/opening of cell wall under high microwave energy. Higher extraction time may have also led to a better interaction between the extracting agent and the cell wall material and to a high thermal accumulation. In addition, the suspension of the cell wall in a high volume of alkaline solution (corresponding to smaller solid/liquid ratio) may have improved the diffusion rate of KOH solution into the cell wall structure and hence the absorption of microwaves by the swollen cell wall. Indeed, the efficient swelling of the cell wall particles suspended in the ionic solution can improve the direct absorption of the microwaves by the cell wall polysaccharides. An increase in the yield by increasing power and time and decreasing solid/liquid ratio has been reported elsewhere. However, other few studies have shown a different pattern.

The 2D contour plots (not shown) of the predicted Ara content illustrating the interactions between solid/liquid ratio and power at a KOH concentration of 1 .5 M. As expected, increasing the solid/liquid ratio increased significantly the absolute amount of arabinan side chains in the extracted polysaccharides. Increasing power up to 45-50 W and time up to 4 min increased the Ara content in the polysaccharide extracts; however, further increase of power to 68 W and extraction time to 8 min led to a decrease in the content of Ara. These results reveal the debranching of arabinan side chains, when time and power were increased.

The interactive effect of KOH concentration and extraction time on the predicted Gal content at a microwave power of 50 W exhibited a hyperbolic trend (data not shown). The horizontal lines of this hyperbolic trend reveal the pronounced effect of KOH concentration on the Gal content of the extracted polysaccharides at shorter extraction times (<4 min); indeed in this investigated space, increasing concentration of KOH resulted in a commitment increase of the Gal content of extracted polysaccharides. These results can be attributed to the ability of alkaline solution to open the cell wall structure and to solubilize the galactan-rich RG I. At longer extraction times, no significant effect of the KOH concentration on the Gal content was observed. The vertical lines of the hyperbolic curves showed the significant negative effect of extraction time on the Gal content of extracted polysaccharides at higher KOH concentrations (>1 .2 M). The results also indicate that the extraction time exhibited no or little effect on the Gal content of the extracted polysaccharides within the shorter time range (>3.0 min). The decrease in the Gal content with an increase in the extraction time at higher KOH concentrations can be attributed to the debranching of galactan side chains and/or to the defragmentation of RG I into oligomers. This negative effect of extraction time seems to be less pronounced at high solid/liquid ratios. These results can be attributed to the lower steric mobility of galactan side chain and/or to the higher diffusional mass limitations at high solid/liquid ratio (2.9%) as compared to low ratio.

The 2D contour plots (not shown) of GalA content illustrating the interaction between extraction time and solid/liquid ratio at a microwave power of 50 W indicate that increasing extraction time up to 4-5 min increased the GalA content of extracted polysaccharides up to a maximum value, which decreased thereafter. These results can be explained by the fact that at shorter extraction times, the rate of the release of cell wall HG polysaccharides was higher as compared to the rate of their hydrolysis by β-elimination. However, further exposure of extracted HG to KOH caused their defragmentation. As expected, the higher is the KOH concentration, the lower is the GalA content of the extracted polysaccharides. Effect of solid/liquid ratio on GalA content seems to be dependent on the extraction time used. At shorter times (>4 min), GalA content of the extracted polysaccharides did not increase significantly with the solid/liquid ratio. However, at higher extraction time, GalA content as indicator of HG presence decreased with the increase of solid/liquid ratio. This effect may be explained by the presence of polymer/polymer interactions at high solid/liquid ratio, which may have favored the elimination of HG.

Optimum conditions for extraction of Gal. Using the predictive models, the optimum extraction conditions, which result in the highest amount of Gal and the lowest amount of GalA in the polysaccharide extracts (RG-enriched fraction), were identified. The identified optimum conditions corresponded to KOH concentration at + 1 level (1 .5 M), solid/liquid ratio at + 1 level (2.9%), extraction time at - 1 level (2.0 min) and power close to - 1 level (36 W, corresponding to 41 .4 W/g cell wall) of design. Comparing experimental and predicted values at the identified optimal conditions can be used to validate the predictive models (Table 3). These results indicate that there is no significant (P <0.05) difference between the experimental data and the predicted values, and all experimental results are within the range of the estimated values with 95% prediction intervals. Using the identified optimal conditions, yield and productivity of RG-enriched polysaccharides were estimated at 21 .9% (w/w) and 192 g/l.h. The extracted polysaccharides contained 71 .0% of monosaccharides originating from RG I (Gal, Ara, Rha). In our previous study, lower yield of 9.5% and productivity of 1 .2 mg/l.h were obtained upon the enzymatic extraction of RG I from potato cell wall with a purity of 85.0% (Khodaei & Karboune, 2014). However, similar yields of 18- 33% were reported for the microwave-assisted alkaline extraction of soluble polysaccharides from sugar beet pulp. In contrast, others have reported a low yield of 1 .5% for the extraction of pectin from orange peel through microwave assisted-alkaline extraction. Furthermore, similar yields (15.7-32.4%) were obtained upon the microwave-assisted acidic extractions of pectin from orange peel, sugar beet pulp and apple pomace. In contrast, others have achieved higher yields of 59 - 84% upon the microwave-assisted acidic extraction of pectin from berry fruits. Table 3. Optimum condition and corresponding responses, and sugar composition of the fractions.

x₂-Time (min) 2.0

x₃-Power (W) 36.0

x₄-Solid/liquid (%) 2.9

Power (W)/g cell wall 41 .4

Responses

Experimental Predicted Fraction 1 Fraction 2

Yield (%) 21 .9 26.8

>600 kDa (%) 78.0 56.7

<600 kDa (%) 19.1 14.1

Polysaccharide (%) 100

Gal ^a 4802.5(63.1)^b 5383.8 (63.3) 4538.2 (75.3) 264.3 (17.5)

Ara 434.8 (5.7) 431 .6 (5.1) 380.7 (6.3) 54.1 (3.6)

Rha 170.0 (2.2) 186.8 (2.2) 156.1 (2.6) 13.9 (0.9)

Glc 1000.9 (13.2) 1097.0 (12.9) 414.6 (6.9) 586.3 (38.8)

Xyl 373.5(4.9) 359.2 (4.2) 289.8 (4.8) 83.7 (5.5)

Man 226.7 (3.0) 245.9 (2.9) 160.5 (2.7) 66.2 (4.4)

GalA 597.1 (7.9) 806.3 (9.5) 84.7 (1 .4) 443.3 (29.3)

Monosaccharide composition is expressed in μηιοΙ per g of extract

b Relative molar proportion of monosaccharides.

Under the identified optimal conditions, no depolymerisation of extracted polysaccharides into oligomers was detected. The MW distribution of extracted polysaccharides showed two major populations (>600 kDa and <600 kDa) with higher proportion of 78.0% for the high-MW polysaccharides (>600 kDa). Lower proportions of high molecular-weight RG l-type polysaccharide population (>500 kDa) of 42.4-62.2% were obtained upon the enzymatic treatment of potato cell wall (Khodaei & Karboune, 2014). Others have reported that the MW of pectin extracted from lime flavedo/albedo/pulp decreased from 559 to 13 kDa, when the microwave-assisted acid treatment time was increased from 2.5 to 10 min. Pectin isolated from sugar beet by microwave-assisted acid extraction exhibited also high MW of 532- 1200 kDa. However, low MW of alkaline-soluble pectin from sugar beet extracted with microwave method (83 - 299 kDa) was obtained.

The results also indicate that the identified optimal conditions resulted in high content of Gal (4802.5 μηιοΙ/g extract, 63.1 % mol/mol) in the extracted RG l-enriched polysaccharides. Similar Gal content of 66.7-71.8% has been reported for RG I polysaccharides obtained though enzymatic treatment. RG-enriched fraction contained hemicellulosic polysaccharides, representing 21.1 % of total molar monosaccharide content. The high molar proportions of Glc (13.2%), Xyl (4.9%) and Man (3.0%) suggest the extraction o, xyloglucan, heteromannans and heteroxylans as hemicellulosic polysaccharides. Higher molar proportions of hemicellulosic-based monosaccharides (Glc, 22%; Xyl, 18%; Man, 12%) were obtained upon the conventional alkaline extraction of potato RG I from cell wall. In addition, the low molar proportion of GalA (7.9%) in the RG I extract confirms the efficient removal of HG region by microwave-assisted alkaline extraction as compared to the enzymatic extraction (GalA of 1 1.3-15.2.

Extracted polysaccharides, under identified optimum conditions, were fractionated on preparative scale size exclusion chromatography (Table 3). Distribution of polysaccharides was similar to results obtained with HPSEC (two populations of polysaccharides i.e. >600 kDa and <600 kDa). High (fraction#1) and low (fraction#2) MW polysaccharides were collected and analyzed for their monosaccharide profile. High-MW fraction #1 exhibited higher molar proportions of monosaccharides originating from RG I, including Gal (75.3%), Ara (6.3%) and Rha (2.6%) as compared to low-MW fraction#2 (17.5, 3.6 and 0.9%, respectively). The high Gal content in the high-MW fraction#1 confirmed that major part of galactan side chains were not debranched and were linked to the backbone of RG I. Low- MW fraction#2 consisted of higher molar proportions of Glc (38.8%), Xyl (5.5%) and Man (4.4%) as compared to high-MW fraction#1 (6.9, 4.8 and 2.7%, respectively). These results reveal that most of hemicellulosic polysaccharides were present in low-MW fraction#2. The high molar proportion of GalA in low-MW fraction#2 (29.3%) reveals the extensive break down of HG.

Functional properties of extracted polysaccharides - Solubility and viscosity. Figure 1A displays the solubility of extracted RG-enriched polysaccharides and of standards (potato galactan and oranges HG) at selected temperatures. RG-enriched polysaccharides exhibited a solubility of 68.8% at 25 °C and 100% at 42-93 °C. The solubility of RG-enriched polysaccharides was higher than that of galactan (53.2-93.7%) and HG (12.8-19.0%). The solubility of polysaccharides is, generally, dependent on their structures. Uniform polysaccharides exhibit low solubility, while branched polysaccharides are more soluble. Lower solubility of HG may be attributed to the fact that it is a linear polysaccharide with one type of the linkage and monosaccharide unit. RG-enriched polysaccharides with heterogeneous structure, consisting of different monosaccharide units and linkages, showed the best solubility.

Figures 1 B B₃ shows changes in shear stress (Pa) by changing shear rate (s^~ ) from 1 to 100. For all samples (1 % w/v) at all tested temperatures (25-73 °C), the flow behavior index (n) was close to one revealing their Newtonian behavior, except for RG-enriched polysaccharides at 49°C which exhibited a pseudoplastic behaviour with flow behavior index of n<1 (0.6). Others have reported that pectin gels, prepared with 40% saccharose and 5% (w/w) berry pectin obtained by conventional extraction, exhibited a Newtonian behavior, while those extracted with microwave method showed a flow behavior index of 0.7. On the other hand, others have shown that pectin had a Newtonian behavior at low concentration, and its pseudoplastic nature increased when its concentration was increased from 7.5 to 12.5.

In terms of the effect of temperature on the viscosity, it can be seen (not shown) that galactan and RG-enriched polysaccharides showed similar pattern as their viscosity increased by increasing temperature to 49°C and decreased at higher temperature of 73°C . In contrast, HG showed the opposite trend as the viscosity of solution decreased slightly by increasing temperature to 49°C and increased by further increase of temperature to 73°C. In addition, RG-enriched polysaccharides resulted in the highest viscosity followed by galactan and HG. Others have reported that the relative viscosity of apple and citrus pectin at concentration of 0.05 - 0.5% didn't vary with temperature; while at concentration of 0.5- 0.75%, viscosity decreased by increasing temperature from 0 to 50 °C. Moreover, others have reported that pectin obtained by microwave extraction produced higher viscosity at lower temperatures as compared to the pectin extracted with conventional extraction method.

Emulsifying properties. Droplet size distribution and stability of emulsions prepared with RG- enriched polysaccharides were assessed and compared with commercial potato galactan and oranges HG as standards. Galactan and HG were selected in order to highlight the effect of neutral sugar and GalA contents on the emulsifying properties. Because of deproteination and deacetylation/demethylation during extraction, the measured emulsifying properties can only be correlated to polysaccharides. Others have reported that the pectins linked to protein and the acetylated ones exhibit higher emulsifying properties. Figures 2A-C show droplet size distribution in the polysaccharide-enriched emulsions at selected pHs. At pH 6 and 8, RG- enriched emulsion showed sharper peaks with more homogeneous droplet size distribution (polydispersity index of 0.03- 0.04); as compared to pH 6 (943.3 nm) and 8 (900.2 nm), the diameter of the oil droplets of the RG-enriched emulsion at pH 4 (670.2 nm) was lower with a polydispersity index of 0.29. However, the HG-enriched emulsion showed lower droplet diameter distributions at pH 6 and 8 (447.8-594.7 nm, polydispersity index of 0.27-0.29) as compared to pH 4 (984.8 nm, polydispersity index of 0.33). These results may be explained by the higher charge of HG at neutral pH reducing the self-aggregation of HG chains. As expected, the pH did not affect significantly the droplet size distribution of galactan-enriched emulsions; smaller droplets of 460 to 508.3 nm were obtained with polydispersity index of 0.04-0.28.

Similarly, others have reported that droplet diameters of soybean soluble polysaccharide and gum arabic-based emulsions were almost independent of pH varying from 3 to 6; however, an increase was observed at pH 6 for sugar beet pectin. It has been shown that sugar beet pectin (550 nm) had higher emulsifying properties followed by soybean soluble polysaccharide (660 nm) and gum arabic (820 nm) using emulsifier/oil ratio of 0.1 , 0.3 and 0.7 g/ml and two passes through a high-pressure homogenizer at 50000 kPa. Although lower emulsifier/oil of 0.04 g/ml and three passes at pressure of 5000 - 7000 kPa were used in the current example, the droplet diameter of RG-enriched emulsion at pH 4 (670.2 nm) was comparable to those reported for sugar beet pectin (550 nm) and soybean soluble polysaccharide (660 nm). Others have reported smaller droplet diameter of 310 nm for gum arabic-based emulsion at emulsifier/oil ratio of 2.5 and higher droplet diameter of 800 nm for citrus pectin-based emulsion at emulsifier/oil ratio of 0.4 at pH of 3.5.

Figures 2D-F show the extent of droplet size increase at different pHs and temperatures. As compared to galactan, heat treatment of RG-enriched emulsions resulted in lower increases of droplet diameters, indicating their high stability. HG-enriched emulsions were highly affected by the heat treatment at pH 8 and 6. These overall results show that the polysaccharide composed of either neutral sugar or GalA can only maintain the stability of emulsion at lower extent than polysaccharides containing both fractions. Highly branched structure of RG-enriched polysaccharides, promoting the steric hindrance, can also explain the stability of their emulsions. Effect of the polysaccharide stuctures on the emulsion stability was studied and a decrease in the stability of emulsions upon removal of neutral side chains (arabinan and galactan) of pectic polysaccharides was observed.

Effects of four independent variables, including KOH (M), extraction time (min), power (W) and solid/liquid ratio (%, w/v), were studied using RSM with 5-level and 4-factor CCRD. The effects of extraction parameters and their interactions were elucidated through the developed quadratic mathematical models. A compromise between the combined effects of high KOH concentration/solid to liquid ratio and low power/extraction time was found to be a key for the efficient extraction of galactan-rich RG I and the limitation of their debranching. Using the identified optimal conditions, high yield and productivity of RG-enriched isolates were obtained. RG-enriched isolates was fractionated into two populations of polysaccharides, high (>600 kDa) and low (<600 kDa) MW ones. Galactan-rich RG I was recovered in the high-MW fraction (78%). Galactan-rich RG I polysaccharides exhibited higher solubility and emulsifying stability as compared to potato galactan and oranges HG. In addition, galactan- rich RG I polysaccharides showed Newtonian behavior at 25 and 73°C and pseudoplastic behavior at 49°C.

EXAMPLE II - ENZYMATIC PRODUCTION OF OLIGOSACCHARIDES/OLIGOMERS FROM POTATO GALACTAN-RICH RHAMNOGALACTURONAN I PECTIC

POLYSACCHARIDES

Materials. Gamanase 1.5L was obtained from Novo Nordisk Bioindustrial Inc. (Danbury, CT). Depol 670L was kindly provided by Biocatalysts Ltd. (Mid Glamorgan, UK), logen HS 70 was obtained from logen Bio-Products, while Newlase II were from Amano Enzyme (USA). Pectinex Ultra SPL and Viscozyme L were from Novozymes, (Ca) and endo- polygalacturonase, endo-1 ,4-p-galactanase, endo-1 ,5-a-arabinanase, soybean RG and potato pectic galactan were from Megazyme (Wicklow, Ir). Arabinan, polygalacturonic acid from oranges, KOH, sodium acetate, acetic acid, dextran standards, Gal, arabinose (Ara), rhamnose (Rha), Glc, xylose (Xyl), mannose (Man), galacturonic acid (GalA), trifluoroacetic acid, thermostable a-amylase from Bacillus licheniformis and amyloglucosidase from Aspergillus niger, were purchased from Sigma Chemical Co. (St-Louis, MO, USA). 1- Kestose, nystose and 1 F-fructofuranosyl- nystose were obtained from Wako Pure Chemical (Japan). All salts were obtained from Fisher Scientific (Fair Lawn, NJ).

Preparation of RG I Polysaccharides from Potato Cell Wall Material. Prior to the extraction of pectic polysaccharides, protein and starch were removed from potato pulp as described previously by Khodaei & Karboune, 2013. Microwave-assisted alkaline extraction method was used for the isolation of galactan-rich RG l-type pectic polysaccharides from potato cell wall using the previous identified optimized conditions (Khodaei et a/., 2015). Isolated cell wall material was suspended in 1.5 M potassium hydroxide solution to yield solid/liquid ratio of 2.9 % (w/v). The suspension was heated for 2 min under power of 36 W in a cylindrical Pyrex tube in a Prolabo Synthewave 402 Microwave Reactor. Solubilized polysaccharides were recovered upon filtration through Glass microfiber filter grade GF/D with pore size of 2.7 μηι (Whatman®). Filtrate was neutralized using hydrochloric acid, dialyzed against water at 4°C using tubes with a 6-8 kDa cut-off and freeze dried. Extracted polysaccharides were analyzed for their MW distribution by high-performance size exclusion chromatography (HPSEC).

Assessment of Glycosyl-Hydrolase Activities of Selected Multi-Enzymatic Products. Selected multi-enzymatic preparations (Depol 670L, Gamanase 1 ,5L, logen HS 70, Newlase II, Pectinex Ultra SPL and Viscozyme L) were assessed for their pectinase, rhamnogalacturonase, galactanase and arabinanase activities. 0.025 mL of multi-enzymatic product was added to 0.475 mL of substrate solution of 0.25% polygalacturonic acid, 0.55% soybean RG, 0.55% potato pectic galactan or 0.55% arabinan for pectinase, rhamnogalacturonase, galactanase and arabinanase activity, respectively. Substrate solutions were prepared in 50 mM sodium acetate buffer, pH 5.0 at concentration of 0.5% (w/v), and the reaction mixtures were incubated at 40°C for 20 min. Two blank tests were also performed without substrate or without enzyme. The reducing ends were quantified using dinitrosalicylic acid (DNS) assay and the enzymatic activity unit (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction.

Enzymatic Hydrolysis of Galactan-rich RG l-Type Pectic Polysaccharides. The time course for the hydrolysis of galactan-rich RG l-type pectic polysaccharides was carried out over a reaction time of 24 h. Pectic polysaccharides suspensions in sodium acetate buffer (50 mM, pH 5.0) at selected concentrations of 0.2, 0.5 and 1 % (w/v) were prepared. The multi- enzymatic products (Depol 670L, Gamanase 1 ,5L, logen HS 70, Newlase II, Pectinex Ultra SPL and Viscozyme L) or mono-component enzymes (endo-1 ,4-B-galactanase and endo- 1 ,5-a-arabinanase) were added to the pectic polysaccharides suspensions to yield final enzyme/substrate ratio of 0.1 , 0.2 and 0.4 galactanase U/mg substrate. After incubation at 40°C, the reaction mixtures were boiled for 5 min in order to inactivate enzyme. Unhydrolyzed polysaccharides were precipitated by adding equal volume of ethanol and removed by centrifugation (10000 ^χ g, 10 min). The progress of the hydrolysis was monitored by measuring the release of reducing sugars using DNS assay in which 0.75 mL of DNS reagent (1 %, w/v DNS in 1 .6%, w/v sodium hydroxide) was added to 0.5 mL of reaction samples. After boiling the mixtures for 5 min, 0.25 mL of potassium sodium tartrate (50%, w/v) was added, and the absorbance was read at 540 nm. Standard curve was plotted using selected concentrations of Gal. Oligosaccharides (DP of 2 to 12) and oligomers (DP of 3 to 70) were also quantified by HPSEC. Yield was defined as the amount of generated oligosaccharides/oligomers over the initial amount of potato RG l-type pectic polysaccharides. The MW distribution of generated oligosaccharides was also determined by HPSEC, and their monosaccharide profile was determined by high performance anionic exchange chromatography (HPAEC).

Characterization of Structural Properties. The MW distribution of carbohydrates was carried out by HPSEC using Waters HPLC system equipped with refractive index detector and Breeze software for the data analysis. For the analysis of the MW distribution of pectic polysaccharides, three columns in series (TSK G3000 PWXL, TSK G4000 PWXL and TSK G5000 PWXL (Tosoh Bioscience, Montgomeryville, PA) were used. Isocratic elution with 0.1 M sodium chloride was applied at flow rate of 0.5 mL/min; dextrans with MW of 50, 150, 270, 410 and 670 kDa were used as standards. To determine the yield and the MW distributions of oligosaccharides, hydrolysates were separated on TSKgel G-Oligo-PW column (Tosoh Bioscience, Montgomeryville, PA). Isocratic elution with 0.1 M sodium chloride was applied at flow rate of 0.5 mL/min; and Gal, sucrose, 1-kestose, nystose and 1 F-fructofuranosyl- nystose were used as standards for calibration.

To determine the monosaccharide profile, the carbohydrates (oligosaccharides, oligomers, polysaccharides) were hydrolyzed using a two-step procedure, according to the method of Khodaei & Karboune, 2013. They were first hydrolyzed with HCI/methanol mixture (1 :4, v/v) at 60°C for 24 h, and then trifluoroacetic acid (TFA) was added at a ratio of 1 :8 (v/v). After 1 h boiling, the monosaccharide profile of hydrolyzed samples was measured by HPAEC using Dionex system equipped with a pulsed amperometric detector (PAD) and a CarboPac PA20 column (3 ^χ 150 mm). Mobile phase of 5 mM sodium hydroxide was used at an isocratic flow of 0.5 mL/min. L-Rha, L-Ara, D-Gal, D-GIc, D-Xyl and D-Man were used as standards. GalA content was measured using sulphamate/m-hydroxydiphenyl colorimetric assay (Blumenkrantz & Asboe-Hansen, 1973).

The profile of generated oligosaccharides was also analyzed by HPAEC-PAD on CarboPac PA200 column. For the elution of oligosaccharides, gradient of 0 to 100% of 200 mM sodium acetate in 100 mM NaOH for 20 min was used as a mobile phase. Sucrose, raffinose, 1 - kestose, nystose and 1 F-fructofuranosyl- nystose were used as internal standards to identify peaks.

Structural Properties of Extracted Galactan-Rich RG l-Type Pectic Polysaccharides. Previously, the optimum conditions for microwave-assisted alkaline extraction of galactan- rich RG l-type pectic polysaccharides were identified to be KOH concentration of 1 .5 M, extraction time of 2.0 min, microwave power of 36.0 W and solid/liquid of 2.9 % (w/v). Table 4 shows the monosaccharide profile and the MW distribution of extracted polysaccharides. The results show that 71 % of monosaccharides in the extracted polysaccharides are those originating from galactan-rich RG I (Gal, Ara and Rha) with Gal being the major monosaccharide (63.1 %, mole/mole). In previous studies, polysaccharide extracts from potato pulp were reported to be composed of 77.2% and 69.0% of RG I. The experimental findings also indicate that extracted polysaccharides contained low amounts of Glc (13.2%), Xyl (4.9%) and Man (3.0%) originating from hemicellulosic polysaccharides. In addition, low content of GalA in the extract (7.9%) reveals the high hydrolysis rate of HG by β-elimination during the microwave-assisted alkaline extraction. The MW distribution indicates the presence of two polysaccharide populations (>600 kDa; <600 kDa) with the high MW fraction being the most abundant one (80.4 %, w/w). Since the polysaccharide extract was dialyzed using 6-8 kDa cut-off, no oligosaccharide was detected.

Table 4. Sugar composition and molecular weight of extracted Rhamnogalacturonan I (RG I) using microwave assisted-alkaline extraction method.

RG properties Values

>600 kDa (%) 78.0

<600 kDa (%) 19.1

Polysaccharide (%) 100.0

Gal ^a 4802.5(63.1)^b

Ara 434.8 (5.7)

Rha 170.0 (2.2)

Glc 1000.9 (13.2)

Xyl 373.5(4.9)

Man 226.7 (3.0)

GalA 597.1 (7.9)

Monosaccharide composition is expressed

in μηιοΙ per g of extract.

bRelative molar proportion of monosaccharides.

Characterization of Activity Profile of Selected Multi-Enzymatic Preparations. The efficiency of the synergistic actions of glycosyl-hydrolase enzymes present in a multi-enzymatic preparations was reported to be higher than that of their individual action (Karboune et al., 2009). Considering the structural characteristics of potato RG l-type pectic polysaccharides, pectinase (EC. 3.2.1.15), rhamnogalacturonase (EC 3.2.1.171), endo-1 ,4-p galactanase (EC 3.2.1.89) and endo-1 ,5-a arabinanase (EC 3.2.1.99) were identified as the appropriate glycosyl-hydrolases for the tailored hydrolysis of potato pectic polysaccharides into oligosaccharides and oligomers. Indeed, up to 80% of rhamnose residues of potato RG I are substituted at 0-4 with galactan, arabinan and arabinogalactan. The levels of the selected glycosyl hydrolase activities were assessed in six multi-enzymatic preparations: Depol 670L from Trichoderma reesei, Gamanase 1 ,5L from A. niger, logen HS 70 and Newlase II from Rhizopus niveus and Pectinex Ultra SPL and Viscozyme L from A. aculeatus. Different glycosyl-hydrolase activities were reported to be expressed in the selected multi-enzymatic preparations: Depol 670L (β-glucanase, arabinanase, endoglucanase, xylanase, polygalacturonase), Gamanase 1 ,5L (β-1-4 mannanase, xylaanase), logen HS70 (xylanase), Pectinex Ultra SPL (polygalacturonase, cellulose) and Viscozyme L (β-glucanase, xylanase, cellulase, hemicellulase). Mono-component enzymes (endo-1 ,4-p-galactanase and endo-1 ,5- a-arabinanase) were also investigated in order to compare their efficiency for the generation of oligosaccharides and oligomers with those of multi-enzymatic preparations. The specific activity of mono-component enzymes were 506.0 and 15.1 U/mg for endo-1 ,4-p-galactanase and endo-1 ,5-a-arabinanase from A niger, respectively. As shown in Table 5, some of the investigated multi-enzymatic preparations showed higher enzymatic activities toward side chains of RG I (galactan and arabinan), while other exhibited higher levels of pectin backbone-hydrolyzing activities (RG I and HG regions). Depol 670L expressed the highest level of galactanase activity (6751.1 U/mL, 230.9 U/mg protein) and the highest enzymatic ratio of galactanase+arabinanase/rhamnogalacturonase (121 .6), revealing its high hydrolyzing effectiveness toward side chains of RG I rather than its backbone. Other multi- enzymatic preparations expressing higher enzymatic activity toward galactan side chains include Gamanase 1 .5L and logen HS 70 with a galactanase+arabinanase/rhamnogalacturonase ratio of 33.0. These results also reveal the high relative proportion of the rhamnogalacturonase enzymatic units in Gamanase 1 .5L and logen HS 70 as compared to Depol 670 L. However, both Gamanase 1 .5L and logen HS showed a low catalytic efficiency for hydrolyzing arabinan side chains with an arabinanase activity of 0.1 U/mg protein and a galactanase/arabinanase ratio of 32.0-166.0. No arabinanase and rhamnogalacturonase activities were detected in the Newlase II, while the levels of pectinase and galactanase present in this multi-enzymatic preparation were similar (10.8 U/ mg protein). Viscozyme L showed the highest level of arabinanase activity (200.0 U/mL, 1 1 .2 U/mg protein) and the lowest galactanase/arabinanase ratio (13.2). As expected Pectinex Ultra SPL had the highest pectinase activity level (1600.0 U/mL, 226.2 U/mg protein) and relatively high arabinanase one (63.0 U/mL, 8.9 U/mg protein).

Table 5. Pectin-hydrolyzing enzymatic activities expressed in multi-enzymatic preparations³

Multi-enzymatic Enzymatic Activity (U ^{b c}/ml) Enzymatic Activity Ratio (U/U)

preparations

Pectinase Rhamogalacturo- Galactanase Arabina- Galactanase/ (Galactanase+Arabina Pectinase/

nase se Arabinase se)/Rhamnogalacturo- Galactananase se

Depol 670L 0.0 (0.0) ^d 56.0 (1 .9) 6751 .0 57.0 (1 .9) 1 18.4 121 .6 0.0

(230.9)

Gamanase 66.0 (4.4) 5.0 (0.3) 166.0 (1 1 .0) 1 .0 (0.1) 166.0 33.4 0.4

1 ,5L

logen HS 70 2.0 (0.2) 1 .0 (0.1) 32.0 (2.8) 1 .0 (0.1) 32.0 33.0 0.1

Newlase II ^e 34.0 (10.8) 0.0 (0.0) 34.0 (10.8) 0.0 (0.0) - - 1.0

Pectinex Ultra 1600.0 90.0 (12.7) 1210.0 63.0 (8.9) 19.2 14.1 1 .3

SPL (226.2) (171 .1)

Viscozyme L 2996.0 218.0 (12.2) 2636.0 200.0 13.2 13.0 1.1

(167.5) (147.4) (1 1 .2) ^a All measurements were run in triplicates and the relative standard deviations were less than 10%. ^b Each unit of enzymatic activity (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction. ^c Pectinase, rhamogalacturonase, galactanase and arabinanase activities were assessed using polygalacturonic acid, soybean RG, potato pectic galactan and arabinan as substrates, respectively. ^d Specific activity per milligram of total protein (expressed in U/mg protein). ^e Enzymatic activity was expressed in U/mg of enzyme

Enzymatic Generation of Potato Oligo-RG I. The hydrolysis of potato RG l-type pectic polysaccharides by selected multi-enzymatic preparations was investigated, over a 24 h time course using well-defined galactanase units (1 -2 U/g substrate). The results (data not shown) showed that the maximum hydrolysis yields were achieved after 7 h of reaction, and no significant change was observed thereafter. In addition, the initial RG l-type pectic polysaccharide concentration of 0.5% (w/v) resulted in the highest hydrolysis rate as compared to 0.2 and 1 %. The low hydrolysis rate at 1 % may be attributed to the presence of polysaccharide/polysaccharide and/or polysaccharide/enzyme interactions limiting the access of the enzyme's active site to the substrate. Therefore, the reaction times of 0.5, 4 and 7 h as well as a substrate concentration of 0.5% (w/v) were subsequently used. These reaction times and substrate concentration fall within the same range as those reported in the literature for the enzymatic generation of oligosaccharides. Others have used reaction times of 0.25, 1 .0 and 4.5 h for the hydrolysis of galactan and potato pulp polysaccharides with endo-1 ,4-p-galactanase. Some have investigated the hydrolysis of sugar beet pectin into rhamno-oligosaccharides by a mixture of pectin-methylesterase, rhamnogalacturonase, galactanase and arabinanase activities using a substrate concentration of 0.8% and a reaction time of 1 .5 h. Moreover, others have isolated homogalacturonides and rhamnogalacturonides from sugar beet pectin by the sequential use of mono-component enzymes at substrate concentrations of 0.75-3% and reaction times of 2-16 h. However, higher substrate concentration of 8.3% and similar reaction times of 4-16 h have been used for investigating the hydrolysis of orange peel waste and sugar beet pulp by Celluclast 1 .5L from Trichoderma reesei and Viscozyme L from A. aculeatus, respectively.

Table 6 summarizes the yield of oligo-RG I (oligosaccharides, oligomers) and their MW distribution as well as the yield of monosaccharides. Because monosaccharides can be faster metabolized as an energy source and increase the caloric content of final product, their presence in prebiotic products is not favorable. In addition, it has been shown that hydrolyzed galactans and solubilized potato polysaccharides showed better prebiotic properties than pure Gal. Commercially available galacto-oligosaccharides were reported to be composed of higher amounts of monosaccharide (Glc, 20%) and lactose (~20%) with an oligosaccharide content of more than 55%. Table 6. Effect of biocatalyst type and reaction time on the yield of generated hydrolysates (oligo-RG I) and their corresponding composition.³

0.1 U/mg substrate^b 0.2 U/mg substrate

Yield (%,w/w) ⁰ MW distribution (%, w/w) Yield (%,w/w) MW distribution (%, w/w) Multi-enzymatic preparation

Time Oliao RG I ^{d Monosaccnaride} Oligosaccharide Oligomer „,. _,„, „, , . , Oligosaccharide Oligomer

Ohgo-RG Monosaccharide

9^" (DP 2-12) (DP 13-70) ^a (DP 2-12) (DP 13-70)

Depol 670L 0.5 3.1 0.0 90.2 9.8 3.4 0.0 89.8 10.2

4 16.7 0.2 90.0 10.0 93.9 2.3 89.7 10.3

7 60.2 4.0 97.5 2.5 82.3 14.0 79.8 20.2

Gamanase 1 ,5L 0.5 1.7 0.0 93.2 6.8 1.0 0.0 92.6 7.4

4 21 .9 0.2 92.0 8.0 43.4 1 .5 92.1 7.9

7 24.6 0.5 91.7 8.3 41 .9 4.9 93.7 6.3 logen HS 70 0.5 43.9 5.0 97.9 2.1 48.7 2.9 99.9 0.1

4 26.4 22.7 93.7 6.3 26.3 26.6 100.0 0.0

7 23.5 25.8 100.0 0.0 27.6 29.4 100.0 0.0

Newlase II 0.5 3.2 0.1 97.0 3.0 26.7 0.9 98.7 1 .3

4 3.0 0.3 100.0 0.0 17.3 10.4 99.7 0.3

7 3.2 0.3 96.6 3.4 22.4 13.7 98.9 1 .1

Pectinex Ultra SPL 0.5 5.2 0.1 94.1 5.9 54.5 5.7 94.4 5.6

4 24.8 0.6 95.0 5.0 30.1 30.8 88.1 1 1.9

7 39.7 1 .2 95.0 5.0 27.3 34.2 89.0 1 1.0

Viscozyme L 0.5 60.8 0.7 94.0 6.0 65.6 4.8 93.2 6.8

4 41 .2 21.2 95.3 4.7 61 .9 10.3 91 .9 8.1

7 40.6 23.4 93.2 6.8 52.3 20.6 99.3 0.7

0.5 17.6 0.4 94.0 6.0 24.7 1 .5 93.7 6.3

Endo-1 ,4-p-galactanase

4 56.6 7.9 93.6 6.4 66.2 9.8 93.2 6.8

7 52.5 15.8 92.7 7.3 45.4 32.2 90.9 9.1

0.5 10.2 1 .7 99.0 1 .0 27.8 5.1 98.8 1 .2

Endo-1 ,5-a-arabinanase 4 16.5 6.9 98.7 1 .3 23.0 10.5 99.4 0.6

+ endo-1 ,4-p-galactanase ^e

7 20.4 10.6 99.1 0.9 13.5 20.5 100.0 0.0

^a All measurements were run as triplicate and the relative standard deviations are less than 12%.

b Each unit of enzymatic activity (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction.

c Yield was expressed as weight (g) of generated compounds per 100 g of substrate.

5 ^d Oligo-RG I includes oligosaccharides (DP 2-12) and oligomers (13-70).

e Endo-1 ,5-a-arabinanase and endo-1 ,4^-galactanase was used at 1 : 1 (U:U) ratio.

As expected, the yield (1.0 - 93.9 %) and the MW distribution of oligo-RG I were dependent on the enzyme activity profile of multi-enzymatic preparations, the number of enzymatic units and the reaction time. The main hydrolysis products were oligosaccharides with a DP of 2-12 (79.8 - 100%), while the oligomers with a DP of 13-70 comprised a smaller proportion of final oligo-RG I products (0.0 - 20.2 %). The use of endo-1 ,4^-galactanase as a mono-component enzyme resulted in a high yield of oligo-RG I of 66.2% (0.2 U/g substrate, 4 h) with relatively low yield of monosaccharides of 9.8%. Only 6.0 to 9.1 % of the oligo-RG I generated by endo-1 ,4^- galactanase were oligomers with high DP. These results reveal the significance of endo-1 ,4^- galactanase for the hydrolysis of potato RG l-type pectic polysaccharides into oligosaccharides. Indeed, potato RG I consists of 67% of galactan side chains. As expected, the rate of oligo-RG I generation by endo-1 ,4^-galactanase was higher at enzyme concentration of 0.2 U/mg substrate than at 0.1 U/mg substrate; however, increasing the reaction time beyond 4 h led to a significant decrease in the yield of oligo-RG I to 45.4% at 0.2 U/mg substrate. This decrease was accompanied by an increase in the yield of monosaccharides, revealing the partial hydrolysis of oligosaccharides upon extended enzymatic treatment of potato RG I with endo- 1 ,4^-galactanase.

Table 6 also indicates that the highest oligo-RG I yield of 93.9 % was achieved upon the treatment of potato RG l-type pectic polysaccharides with Depol 670L at an enzyme concentration of 0.2 U/g substrate for 4 h. This result reveals the higher catalytic efficiency of the synergistic actions of pectin-hydrolyzing enzymes present in Depol 670L multi-enzymatic preparation as compared to the endo-1 ,4^-galactanase mono-component. Indeed, Depol 670L exhibited the highest galactanase+arabinanase/ rhamnogalacturonase ratio of 121.6 and no pectinase activity, revealing its higher hydrolyzing activity toward side chains of RG I as compared to its backbone. In addition, the monosaccharide yields (0.0-14.0% w/w) generated in the presence of Depol 670L was lower as compared to that obtained with endo-1 ,4^- galactanase mono-component (0.4-32.2%); while the oligomer content (2.5-20.2%, w/w) was higher with Depol 670L. These results reveal that the pectin-hydrolyzing enzymes present in Depol 670L may have displayed an endo-hydrolyzing mode of action for the synergistic hydrolysis of potato RG I. On the other hand, Gamanase 1.5L, which exhibited galactanase+arabinanase /rhamnogalacturonase ratio of 33.4, resulted in a relatively moderate maximum yield of oligo-RG I of 24.6 and 43.4% at enzyme concentration of 0.1 and 0.2 U/mg substrate, respectively; only low amount of monosaccharides (0.0-4.9%) was released revealing the endo-mode action of its enzymes on potato RG-I. However, Gamanase 1.5L had the highest galactanase/arabinanase ratio of 166.0, revealing the low proportion of its arabinanase activity as compared to the galactanase one. The moderate maximum yields of oligo-RG I obtained upon enzymatic treatment with Gamanase 1.5L may be due to the enzyme inhibition and/or to the steric hindrance of its endo-1 ,4^-galactanase by the non-hydrolyzed arabinan side chains. Indeed, the arabinan side chains in RG I consist of a main chain of a-1 ,5-linked L- arabinan with 2- and 3-linked arabinose, which can also be linked to galactosyl residues of β-1 ,4-linked galactan side chains at 0-3. It has been reported that the hydrolysis of arabinan can improve the access to galactan side chains. Nevertheless, using a mixture of endo-1 ,4^-galactanase and endo-1 ,5-a-arabinanaseat a ratio of 1 : 1 (U:U) as mono-component enzymes resulted in a low maximum oligo-RG I yields of 20.4 and 27.8% and a high maximum monosaccharide content of 10.6 and 20.5% at enzyme concentration of 0.1 and 0.2 U/mg substrate, respectively. In addition, the low content of oligomers (1.0 - 1.3%) confirms the high hydrolysis rate of the branches of RG I using this mixture of mono-component enzymes. The modulation of galactanase/arabinanase ratio seems to be determinant for the efficient generation of oligo-RG I and the limitation of the release of monosaccharides.

The results also show that Newlase II, which is lacking arabinanase and rhamanoglacturonase activities, hydrolyzed the potato RG l-type pectic polysaccharides with the lowest extents. Indeed, the maximum yields of oligo-RG I of 3.2 and 26.7% were obtained upon the enzymatic treatment with Newlase II for a short reaction time of 0.5 h at an enzyme concentration of 0.1 and 0.2 U/mg, respectively. Increasing the reaction time did not improve the yields, and the release of monosaccharides (0.1-0.9%) and the oligomer contents (1.3-3.0%) were not significant at 0.5 h of incubation. These results may be due to the low accessibility of galactanase expressed in Newlase II to its substrate. Although logen HS and Newlase II showed more or less similar galactanase activity, logen HS was more efficient as a multi- enzymatic preparation for the generation of oligo-RG I with a maximum yield of 43.9 and 48.7 % at 0.1 and 0.2 U/g substrate, respectively, and 0.5 h. Increasing the reaction time to 4 h led to a decrease in the oligo-RG I yield to 26.4 and 26.3% with a concomitant increase in the monosaccharide yields from 5.0 and 2.9 to 22.7 and 26.6%; beyond this reaction time of 4 h, the yields and the product profiles remained constant. The high hydrolysis rate of the potato RG I by logen HS may be attributed (a) to the appropriate enzymatic profile of its pectin-hydrolyzing enzymes with galactanase/arabinanase ratio of 32.0 and galactanase+arabinanase/rhamnogalacturonase ratio of 33.0 and/or (b) to its high xylanase activity, which may have helped the hydrolysis of linkages between RG I and xyloglucan/ xylogalacturonan, allowing easiest access of pectin-hydrolyzing enzymes to their substrates. In addition, the results also indicate the exo-hydrolyzing action of logen HS at extended reaction times and the importance of controlling its catalytic action in order to limit the subsequent hydrolysis of oligosaccharides.

Pectinex Ultra SPL and Viscozyme L, which exhibited similar activity profile of pectin- hydrolyzing enzymes, showed more or less similar trends for the hydrolysis of potato RG l-type pectic polysaccharides, at high enzyme concentration of 0.2 U/ mg substrate, in which the generation of oligo-RG I achieved higher yields (yield of 54.5 and 65.6% for Pectinex Ultra SPL and Viscozyme L, respectively) upon 0.5 h of reaction and decreased thereafter to 27.3 and 52.3% at 7 h as a result of their partial hydrolysis into monosaccharides (34.2 and 20.6%). As compared to Depol 670 L, the presence of branches- and backbone-hydrolyzing enzymes at low ratios in Pectinex Ultra SPL and Viscozyme L (galactanase+arabinanase/ rhamnogalacturonase ratio of 13.0-14.1) may have favored an extensive hydrolysis of potato RG l-type pectic polysaccharides into oligo-RG I at higher rate. As a result, the maximum oligomer contents of oligo-RG I, generated upon Pectinex Ultra SPL (1 1.9%) and Viscozyme L (8.1 %) treatments, were lower as compared to that (20.2%) obtained with Depol 670 L.

It was previously reported that a lower yield of 26.7% for the generation of pectic oligosaccharides from sugar beet pulp using Celluclast 1.5L from Trichoderma reesei and Viscozyme L from A. aculeatus. Higher pectic oligosaccharide yield of 95% was reported upon the conversion of low and high meyhoxyl pectin from citrus and apple by endo- polygalacturonase from Aspergillus pulverulentus. Others have used a mixture of β- galactosidase from A. niger, β-galactosidase from Kluyveromyces lactis, galactanase, arabinofuranosidase and arabinanase (0.2-4.7% enzyme/ substrate) to hydrolyze side chains of sugar beet pectin, have converted 38% and 44% of arabinan and galactan chains, respectively, to hydrolysis product with MW of <3kDa. Furthermore, some have investigated the use of Celluclast 1.5L from Trichoderma reesei and Viscozyme L from A. aculeatus for the conversion of orange peel waste into mono- and oligosaccharides; these authors have reported the solubilization of 46.5-73.8% of galactan to galactose and galacto-oligosaccharide and 70.5- 100% of arabinan to arabinose and arabino-oligosaccharides.

MW Distribution and Monosaccharide Profile of Generated Oligo-RG I. The MW distribution of generated oligosaccharides, being the major products, and the monosaccharide profile of the whole oligo-RG I (oligosaccharides, oligomers) were characterized. HPAEC-PAD elution profiles (Figure 3) indicate the heterogeneous nature of generated oligosaccharides. The elution profiles of the reaction components of Pectinex and Viscozyme-catalyzed hydrolysis reactions were similar with four abundant peaks # 2,5,6 and 7 at retention times of 7.2, 8.6, 10.7 and 11.9 min, respectively. Indeed, Pectinex Ultra SPL and Viscozyme L exhibited similar activity profile of pectin-hydrolyzing enzymes. These results reveal the importance of the enzyme activity profile and the substrate specificity of the multi-enzymatic preparation rather than their microbial sources. Depol 670L preparation showed similar end-product profiles of low (DP of 3-4) to medium (DP of 5-6) MW oligosaccharides (Peaks#2-5 and 6-7, respectively) than that of Gamanase 1.5L; however, the molecular species detected beyond a retention time of 12 min were different for both multi-enzymatic preparations. As compared to the multi-enzymatic preparations, the elution profile of endo-1 ,4^-galactanase mono-component showed relatively one abundant peak#2 with retention time of 7.2 min. The limited heterogeneity of reaction components of endo-1 ,4^-galactanase may be due its substrate specificity towards galactan. The limited number of peaks detected in the elution profiles of the reaction components of logen HS (peaks # 1 and 3) and NewLase ll-catalyzed hydrolysis (peak # 5) may be attributed to their low galactanase/arabinanase enzymatic ratio.

The percentage distribution of the selected heterogeneous pools (DP of 2-3; 4-5; 6-10) of generated oligosaccharides were determined and reported in Figure 4. The results show that the MW distribution of oligosaccharides (42.0-54.8% of 2-3 DP; 45.2-58.0% of 4-5 DP) generated by endo-1 ,4^-galactanase was not significantly affected by the enzyme amount nor by the reaction time, expect at high enzyme concentration of 0.2 U/mg and longer reaction time of 7 h (19.9% of 2-3 DP; 80.1 % of 4-5 DP). These results can be explained by the mono- component behaviors of endo-1 ,4^-galactanase acting on the same glycoside sites of RG I. Supplementing the endo-1 ,4^-galactanase with endo-1 ,5-a-arabinanase (1 : 1 , U:U) resulted in the release of mainly low MW oligosaccharides (92-100% of 2-3 DP), confirming the extensive hydrolysis of RG I branches as depicted by the monosaccharide content (Table 5). 0bro et al., 2004 have investigated the synergy between endo-1 ,4^-galactanase and endo-1 ,5-a- arabinanase for the hydrolysis of side chains of potato RG I and reported that this synergistic action was more efficient towards the degradation of arabinans than galactans. Figure 4 also shows a shift in the MW distribution of oligosaccharides generated by Viscozyme L towards low MW one (2-3 DP) upon the increase of the enzyme concentration and the reaction time to reach 100% of di/trisaccharides at 0.2 U/mg substrate and 7 h. In the presence of Pectinex Ultra SPL, this shift was less pronounced resulting in only 22.1 and 40.0% of low MW oligosaccharides (2- 3 DP) at 0.1 and 0.2 U/mg substrate, respectively. However, as compared to Viscozyme L, Pectinex Ultra SPL led to a rapid release of monosaccharides within the investigated reaction time course, revealing its pronounced exo-mode action (Table 5). logen HS 70 generated mainly low MW oligosaccharides (39.7-100.0% of 2-3 DP) over the reaction course. While the most abundant oligosaccharides generated by Gamanase 1 ,5L are tetra/pentaoligosaccharides (4-5 DP) representing 65.0-100.0%. The high MW oligosaccharides (6-10 DP) were released by Depol 670L (49.4%) and Newlase II (33.6%) at a high enzyme concentration of 0.2 U/mg substrate after 4 and 7 h reaction, respectively. From these results, it is suggested that an enzymatic profile of multi-enzymatic preparation with relatively low or no arabinanase activity or high galactanase to rhamnogalacturonase ratio can favor the production of high MW oligosaccharides.

The reaction conditions that can lead to the generation of high proportion of high-MW oligosaccharides and/or to high yield of oligo-RG I were identified to be : Depol 670L (0.2U/mg, 4h-7h), Gamanase 1 ,5L (0.2U/mg, 4h), Pectinex Ultra SPL (0.2U/mg, 0.5h), Viscozyme L (0.2 U/mg, 0.5h), endo-1 ,4^-galactanase (0.1-0.2 U/mg, 4h), logen HS 70 (0.1-0.2U/mg, 0.5h), Newlase II (0.2U/mg, 0.5h-7h). Monosaccharide profile of oligo-RG I obtained using the identified conditions was characterised (Table 7). Gal was the major monosaccharide for all generated oligo-RG I (58.9-91.2%, w/w). Ara content of oligo-RG I varied between 0.0 and 12.1 % depending on the type of biocatalysts and the reaction time. Newlase II, expressing no arabinanase and rhamnogalacturonase activities, released oligo-RG I with no Ara and Rha residues. As an overall, the reaction time and the enzyme concentration had more significant effect on the Ara content than Gal one of oligo-RG I. This can be attributed to the low enzymatic units of arabinanase present in the multi-enzymatic products. The generated oligo-RG I end- products with high Ara content were also characterized by high Gal Content. These results reveal the importance of synergism for the efficient hydrolysis of arabinan and galactan side chains of RG I and the presence of Ara substitutions in the galacto-oligosaccharides. Indeed, despite the low arabinanase activity of Gamanase 1.5L, it generated oligo-RG I with high Ara (12.0%) and Gal (77.8%) contents. 0bro et ai, 2004 have also reported that the treatment of potato RG I with endo-galactanase released mainly galacto-oligosaccharides with containing small amount of Ara (Gal/Ara ratio of ratio of 20/1).

Table 7. Monosaccharide composition of generated oligosaccharides/oligomers from potato rhamnogalacturonan I (g/100g of Oligo-RG l)^a

Enzyme/incubation condition Rha Ara Gal Glc Xyl Man GalA

Depol 0.2U ^b /mg, 4h 0.9 8.5 78.0 4.7 0.0 0.5 6.6

Depol 0.2U/mg, 7h 1.9 12.1 78.1 5.6 0.0 0.0 0.9

Gamanase 0.2U/mg, 4h 0.9 12.0 77.8 3.2 1 .9 2.3 1 .9 logen 0.1 U/mg, 0.5h 0.0 0.0 73.0 3.7 2.7 0.0 20.6 logen 0.2U/mg, 0.5h 0.0 0.2 74.3 3.7 3.2 0.0 18.6

Newlase 0.4U/mg, 0.5h 0.0 0.0 63.8 3.8 1 .0 0.0 31.4

Newlase 0.4U/mg, 7h 0.0 0.0 58.9 4.7 1 .9 0.0 34.6

Pectinex 0.2U/mg, 0.5h 0.9 6.6 77.4 3.8 1 .9 0.0 9.4

Viscozyme 0.2U/mg, 0.5h 1 .1 8.7 76.1 5.4 2.2 1 .0 6.5

Viscozyme 0.2U/mg, 7h 1 .9 10.2 72.4 5.3 1 .9 1.0 8.3

Endo-1 ,4-β-galactanase

0.0 2.9 91 .2 2.1 0.9 0.0 2.9

0.1 U/mg, 4h

Endo-1 ,4- β-galactanase

0.0 3.8 88.0 3.3 1.1 0.0 3.8

0.2U/mg, 4h

a All measurements were run as triplicate and the relative standard deviations are less than 10%.

Rha content of pectic polysaccharides was 2.2%, while its content in most generated oligosaccharides/oligomers was 0.0-1.1 %. Most of RG I backbone remained therefore intact upon enzymatic treatment, logen HS 70 and Newlase I I resulted in the highest content of GalA in the oligo-RG I (18.6-34.6%) indicating that these multi-enzymatic preparations hydrolyzed more efficiently the HG region rather than RG I region. About 21.1 % of potato RG l-type pectic polysaccharides consisted of hemicellulosic polysaccharides as indicated by the content of Glc + Xyl + Man. However, the proportion of monosaccharides originating from hemicellulosic polysaccharides decreased in the generated oligo-RG I to 3.0-8.6%, indicating their low release. The highest amounts of Glc + Xyl + Man was obtained through treatment with Viscozyme L (8.6%), Gamanase 1 .5L (7.4%) and logen HS 70 (6.9%). Others used galactan-rich potato pulp polysaccharides to produce two fractions of <10 kDa (yield of 20.6%) and >10 kDa (yield of 65.0%) through treatment with endo-1 ,4^-galactanase from Emericella nidulans. The <10 kDa fraction had very similar monosaccharide profile to sample obtained with endo-1 ,4^-galactanasetreatment in the present example (Rha (0%), Ara (3.6%), Gal (92.9%), Glc (0.3%), Man (0.0%) and GalA (3.2%)) and the >10 kDa fraction had lower amount of Gal (53.9%) and higher amount of Rha, Ara, GalA, Glc and Man.

The efficiency of multi-enzymatic preparations, exhibiting high hydrolysis activity toward side chains and backbone of RG I, for the hydrolysis of RG I into oligosaccharides/oligomers was demonstrated. Generated oligosaccharides/oligomers were mainly composed of Gal. The developed enzymatic approach will, therefore, allow the production of galacto-oligosaccharides and their corresponding oligomers, from an abundant source, as a potential functional food ingredient.

EXAMPLE III - OPTIMIZATION OF ENZYMATIC PRODUCTION OF PREBIOTIC OLIGOSACCHARIDES AND OLIGOMERS FROM POTATO RHAMNOGALACTURONAN I Materials. Gamanase 1.5L from A. niger was obtained from Novo Nordisk Bioindustrial, Inc. (Danbury, CT), and Depol 670L from T. reesei was from Biocatalysts Ltd. (Mid Glamorgan, UK). Soybean RG and potato pectic galactan were from Megazyme (Wicklow, Ireland). Arabinan, polygalacturonic acid from oranges, KOH, sodium acetate, acetic acid, dextran standards, Gal, Ara, Rha, Glc, Xyl, Man, GalA, trifluoroacetic acid, thermostable a-amylase from B. licheniformis and amyloglucosidase from A. niger, were purchased from Sigma Chemical Co. (St-Louis, MO, USA). All salts were obtained from Fisher Scientific (Fair Lawn, NJ). 1-kestose, nystose and 1 F- fructofuranosyl- nystose were obtained from Wako Pure Chemical (Japan).

Extraction of RG l-rich pectic polysaccharides from potato pulp. Prior to the extraction of pectic polysaccharides, protein and starch were removed from potato pulp as described previously (Khodaei & Karboune, 2013). A microwave-assisted alkaline extraction method was used for the extraction of potato RG l-rich pectic polysaccharides (Khodaei et al., 2015d). The cell wall material was dissolved in 1.5M KOH at solid/liquid of 2.9% w/v and heated using a power/suspension ratio of 1.2 W/mL for 2 min in a Panasonic 1 200 W domestic microwave. After heating, the insoluble fraction was removed by centrifugation at 8000 ^χ g for 25 min and the supernatant was neutralized with HCI. The recovered solution was concentrated to 25% of its original volume using ultrafiltration membrane (EMD Millipore Prep/Scale spiral-wound, TFF- 2, cut-off 10 kDa) connected to an EMD Millipore Pellicon™ easy-load peristaltic pump. The high-MW polysaccharides were recovered by mixing the solution with ethanol (1 : 1 , v:v) and centrifugation at 8 000 ^χ g for 25 min, followed by freeze drying of the pellets.

Assays of glycosyl-hydrolase activities. Pectinase, rhamnogalacturonase, galactanase and arabinanase activities expressed in Depol 670L and Gamanase 1 ,5L were assessed using 0.5% (w/v) substrate solutions of polygalacturonic acid, soybean RG, potato pectic galactan and arabinan respectively, prepared in 50 mM sodium acetate buffer, pH 5.0. The activity was measured by adding 0.025 mL of multi-enzymatic product to 0.475 mL of substrate solution, incubated at 40 °C for 20 min. Two blank tests were also performed without substrate or without enzyme and the reducing ends were quantified using a dinitrosalyciclate (DNS) assay. The enzymatic activity was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction per mL or g of multi-enzymatic products. DNS assay was done by adding 0.75 mL of DNS reagent (1 % (w/v) DNS in 1.6% (w/v) sodium hydroxide) to 0.5 mL of samples. Reaction tubes were placed in boiling water for 5 min and then 0.25 mL of potassium sodium tartrate (50%, w/v) was added and absorbance of generated yellow color was measured at 540 nm. Standard curve was plotted with Gal and the enzymatic activity unit (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction.

Generation of oligosaccharides/oligomers from potato RG l-rich pectic polysaccharides. Extracted RG l-rich potato polysaccharides were suspended in 50 mM sodium acetate buffer, pH 5.0. A mixture of Depol 670L and Gamanase 1 ,5L at selected ratios was added to the polysaccharide suspension. Total enzymatic unit (expressed in galactanase unit) was kept constant at 0.2 U/mg RG l-rich pectic polysaccharides. Incubation of reaction mixture was performed at 40 °C with an agitation at 200 rpm. After selected reaction times, the enzymes were inactivated by boiling the reaction mixture for 5 min. In order to remove unhydrolyzed polysaccharides, reaction mixtures were mixed with ethanol (1 : 1 , v:v) and were centrifuged at 12000 xg for 10 min after 30 min of incubation at 5 °C. Yield and MW distribution of oligosaccharides was measured using High Performance Size Exclusion Chromatography (HPSEC), and the monosaccharide profile of oligosaccharides was measured using High- Performance Anion-Exchange Chromatography (HPAEC).

Experimental design and statistical analysis. The generation of oligosaccharides from potato RG l-rich pectic polysaccharides was optimized using response surface methodology (RSM) and a Central Composite Rotatable Design (CCRD). The investigated parameters and their levels were: incubation time (0.5, 2, 3.5, 5 and 6.5 h), Depol 670L/Gamanase 1 ,5L ratio (37.5/62.5, 50/50, 62.5/37.5, 75/25, and 87.5/12.5%, U/U) and substrate concentration (0.3, 0.88, 1.44, 2.0, and 2.6%, w/v). A total of 34 experiments were performed based on the CCRD with three variables at five levels. The complete design included 16 factorial points (levels ±1), 12 axial points (levels ± a), and 6 replicates in center point and experiments were done in random order.

ANOVA of obtained data was done with the Design-Expert™ 8.0.2 software (Stat-Ease, Inc., Minneapolis, MN, USA). Experimental data were fitted to a quadratic polynomial model, which shows behavior of responses (yield of the oligosaccharides with DP of 2-6 and 7-12 and oligomers with DP of 13-70) when independent variables (incubation conditions) change. The general form of the quadratic polynomial model was as follows (Eq. II 1.1):

The coefficients of the polynomial equations were represented by β (constant term), β, (linear coefficient), β,; (quadratic coefficient), and β,₇ (interaction coefficient), / represents the response and x, and x, represent independent variables. Counter plots were developed for the most significant interaction terms.

MW distribution. Degree of polymerization (DP) of the generated oligosaccharides/oligomers was measured using Waters HPLC system equipped with TSKgel G-Oligo-PW column (Tosoh Bioscience, Montgomeryville, PA), refractive index detector and Breeze software for the data analysis. Isocratic elution was applied with 0.1 M NaCI at flow rate of 0.5 mL/min. The overall the concentration of generated oligosaccharides/oligomers was quantified from the standard curves constructed using Gal, sucrose, trisaccharide (1-kestose), tetrasaccharide (nystose) and pentasaccharide (1 F-fructofuranosyl- nystose) as standards. MW distribution of the polysaccharides was investigated using three columns in series (TSK G3000 PWXL, TSK G4000 PWXL and TSK G5000 PWXL, Tosoh Bioscience, Montgomeryville, PA) and 0.1 M NaCI as a mobile phase. Dextrans with MW of 50, 150, 270, 410 and 670 kDa were used as standards for calibration.

Monosaccharide profile. Generated oligosaccharides/oligomers were analyzed for their monosaccharide profile as described previously (Khodaei & Karboune, 2013). Upon the hydrolysis, the monosaccharides were analyzed using HPAEC-PAD (Dionex), and a CarboPac PA20 column (3 ^χ 150 mm). 5 mM NaOH with isocratic flow rate of 0.5 mL/min was used and Rha, Ara, Gal, Glc, Xyl and Man were used as standards. GalA content was measured using sulfamate/m-hydroxydiphenyl colorimetric assay (Blumenkrantz & Asboe-Hansen, 1973).

Effect of reaction conditions on the generation of oligosaccharides and oligomers. A previous study (Khodaei et al., 2015) has shown that 71 % of polysaccharides, extracted by microwave assisted-alkaline method, corresponded to the galactan-rich RG I with a monosaccharide profile composed of Gal, Ara and Rha at a molar proportion of 63.1 , 5.7, and 2.2%, respectively (data not shown). Only low amount of hemicellulosic polysaccharides were isolated, as it was indicated by the molar proportion of Glc (13.2%), Xyl (4.9%) and Man (3.0%). The MW distribution of potato RG l-rich pectic polysaccharides was characterized by a high population of high-MW polysaccharides (>600 kDa, 80.4 %, w/w), and the absence of oligosaccharides/oligomers.

The hydrolysis of potato RG l-rich pectic polysaccharides into low and high-MW oligosaccharides (DP of 2-6 and 7-12) and oligomers (DP of 13-70) was investigated using Depol 670L/Gamanase 1 ,5L bi-enzymatic system. In order to better understand the effects of reaction parameters and their interactions on the efficiency of bi-enzymatic system and the end- product profile, RSM has been used. Table 8 shows the experimental conditions and the selected responses. Yields of oligosaccharides (DP of 2-6 and 7-12) and oligomers (DP of 13- 70) were selected as responses since the prebiotic property can be affected by the MW distribution. Indeed, low-MW oligosaccharides are fermented by Bifidobacterium at a faster rate. On the other hand, higher-MW oligosaccharides and oligomers are slowly fermented, but can reach the distal region of colon and provide a more prolonged prebiotic effect in the intestine. Therefore, producing tailored mix of oligosaccharides and oligomers with well-defined MW distribution can provide the ability to modulate the intestinal health more efficiently.

Table 8 shows that the yields of low-MW (DP of 2-6) and high-MW (DP of 7-12) oligosaccharides and oligomers (DP of 13-70) varied between 0.1-13.9, 0.0-37.5 and 0.0-75.7% (g/ 100 g substrate), respectively. The highest yield of low-MW oligosaccharides was obtained when the substrate concentration was at its lowest levels of -a and -1 , corresponding to 0.3%, w/v (treatments n° 17 and 18; 13.9%) and 0.9% (treatment n° 3; 10.3%), respectively. Similarly, the highest yields of high-MW oligosaccharides were achieved upon the use of low substrate concentrations of 0.3 and 0.9%, w/v (treatment n° 2, 17 and 18; 26.2, 37.5 and 37.5%, respectively). On the other hand, when the substrate concentration of 2.0 and 2.6% w/v (+1 and +2 level) was used the yield of both low (0.9-5.3%) and high-MW (0.0-8.0%) oligosaccharides decreased. Table 8. Central composite rotatable design of experiments and obtained responses.

Oligosaccharide and oligomer yield epol x₃-Substrate Monosaccharide

(g/ 100 g substrate)

manase

concentration (%, (g/ 100 g DP DP DP Total order (h) 1.5L (U/U) ^{a b} w/v) substrate) 2-6 7-12 13-70 oligosaccharide/oligomer

1 -1 (2.0)^c -1 (50.0/50.0) -1 (0.9) 3.1 4.6 19.3 35.2 59.1

2 -1 (2.0) -1 (50.0/50.0) -1 (0.9) 3.0 3.7 26.2 44.2 74.1

3 1 (5.0) -1 (50.0/50.0) -1 (0.9) 10.3 10.3 7.7 43.9 62.0

4 1 (5.0) -1 (50.0/50.0) -1 (0.9) 8.3 8.5 7.7 41.1 57.3

5 -1 (2.0) 1 (75.0/25.0) -1 (0.9) 6.3 7.3 17.0 0.0 24.2

6 -1 (2.0) 1 (75.0/25.0) -1 (0.9) 5.4 6.2 17.0 3.7 26.8

7 1 (5.0) 1 (75.0/25.0) -1 (0.9) 7.2 7.4 10.8 42.5 60.6

8 1 (5.0) 1 (75.0/25.0) -1 (0.9) 7.4 7.3 6.8 23.4 37.5

9 -1 (2.0) -1 (50.0/50.0) 1 (2.0) 3.1 3.8 0.0 34.6 38.3

10 -1 (2.0) -1 (50.0/50.0) 1 (2.0) 3.0 3.8 0.6 34.6 38.9

1 1 1 (5.0) -1 (50.0/50.0) 1 (2.0) 2.1 2.2 0.0 65.5 67.7

12 1 (5.0) -1 (50.0/50.0) 1 (2.0) 5.0 5.2 0.0 59.7 64.9

13 -1 (2.0) 1 (75.0/25.0) 1 (2.0) 4.3 5.2 0.0 0.0 5.2

14 -1 (2.0) 1 (75.0/25.0) 1 (2.0) 2.1 2.6 0.0 0.0 2.6

15 1 (5.0) 1 (75.0/25.0) 1 (2.0) 1.0 0.9 0.0 27.3 28.2

16 1 (5.0) 1 (75.0/25.0) 1 (2.0) 1.6 1.6 0.0 47.6 49.1

17 0 (3.5) 0 (62.5/37.5) -2 (0.3) 17.6 13.9 37.5 0.0 51.4

18 0 (3.5) 0 (62.5/37.5) -2 (0.3) 15.9 13.9 37.5 0.0 51.4

19 0 (3.5) 0 (62.5/37.5) +2 (2.6) 4.7 5.3 8.0 0.0 13.3

20 0 (3.5) 0 (62.5/37.5) +2 (2.6) 4.9 5.3 6.6 0.0 1 1.9

21 0 (3.5) -2 (37.5/62.5) 0 (1.4) 2.6 3.1 0.0 69.4 72.5

22 0 (3.5) -2 (37.5/62.5) 0 (1.4) 4.5 3.7 0.0 72.9 76.6

23 0 (3.5) +2 (87.5/12.5) 0 (1.4) 0.4 0.4 0.0 0.0 0.4

24 0 (3.5) +2 (87.5/12.5) 0 (1.4) 0.1 0.1 0.0 0.0 0.1

25 -2 (0.5) 0 (62.5/37.5) 0 (1.4) 0.1 0.3 5.0 1 1.9 17.2

26 -2 (0.5) 0 (62.5/37.5) 0 (1.4) 0.1 0.3 0.0 12.8 13.1

27 +2 (6.5) 0 (62.5/37.5) 0 (1.4) 1.5 1.0 0.0 75.7 76.7

28 +2 (6.5) 0 (62.5/37.5) 0 (1.4) 2.3 2.1 0.0 74.6 76.6

29 0 (3.5) 0 (62.5/37.5) 0 (1.4) 4.4 5.1 14.1 35.1 54.3

Oligosaccharide and oligomer yield x-i-Time x₂- Depol x₃-Substrate Monosaccharide

Std (g/ 100 g substrate)

670L/Gamanase

30 0 (3.5) 0 (62.5/37.5) 0 (1.4) 4.3 5.1 7.0 41.3 53.4

31 0 (3.5) 0 (62.5/37.5) 0 (1.4) 3.9 5.1 7.0 36.3 48.4

32 0 (3.5) 0 (62.5/37.5) 0 (1.4) 4.3 5.1 7.0 37.4 49.5

33 0 (3.5) 0 (62.5/37.5) 0 (1.4) 4.1 4.5 8.2 35.4 48.1

34 0 (3.5) 0 (62.5/37.5) 0 (1.4) 4.1 4.5 8.2 32.6 45.3 ^a The enzymatic activity unit (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction.

b Total enzymatic unit (expressed in galactanase unit) was kept constant at 0.2 U/mg RG l-rich pectic polysaccharides.

c Number in parenthesis represent actual experimental amounts.

These results indicate that higher rate and hydrolysis extent were achieved, when lower substrate concentration was used. No oligomer was accumulated upon the use of the lowest substrate concentration of 0.3%, w/v (treatments n° 17/18) and the highest one of 2.6% (treatments n° 19/20) using Depol 670L/Gamanase 1 .5L ratio of 62.5/37.5 (U/U) and reaction time of 3.5 h. The hydrolysis of oligomers into oligosaccharides may have happened at low substrate concentration, limiting their accumulation; while the mass transfer limitations may have hindered the access of enzymes to their action sites at high substrate concentration. Comparing the end-product profiles at treatments n° 2 and 3 reveals that increasing reaction time from 2 to 5 h, at a Depol 670L/Gamanase 1 ,5L ratio of 50/50 U/U, increased the yield of low-MW oligosaccharides from 3.7 to 10.3%, while that of high-MW oligosaccharides decreased from 26.2 to 7.7%. Highest yields of oligomers (69.4-75.7%) were obtained upon treatments n° 21/22 (3.5 h; 37.5/62.5 U/U of Depol 670L/Gamanase 1 .5L ratio) and 27/28 (6.5 h; 62.5/37.5 U/U of Depol 670L/Gamanase 1 .5L ratio), in which a similar substrate concentration of 1 .4% (w/v) was used. These treatments (21/22; 27/28) resulted in low yields of low-MW (1 .0-3.7%) and high-MW (0.0%) oligosaccharides. Treatments n° 22, 27, 28, 2, 21 , and 1 1 led to the highest yields (67.7-76.6%) of total hydrolysates. All of these treatments resulted in higher yields of oligomers (44.2-75.7%).

Similar or lower yields have been reported in the literature when a combination of mono- component enzymes or commercial multi-enzymatic preparations were used. Others have studied the effect of three variables including polygalacturonase to solid ratio (10-15 U/g), Celluclast 1 .5L (cellulases) to Viscozyme 1 .5L (endo-polygalacturonase) ratio (0-1 U/U) and reaction time (4-16 h) on the production of neutral and acidic oligosaccharides from sugar beet pulp; they have reported that 3.9-30.0, 8.2-31 .2, 29.2-63.5 and 19.9-42.2% of cellulose, galactan, arabinan and HG were converted into oligosaccharides, respectively. Moreover, Some have investigated the hydrolysis of side chains of sugar beet pectin by a mixture of enzymes, including β-galactosidase from A. niger, β-galactosidase from K. lactis, β- galactanase from A. niger and a-arabinofuranosidase, and a-arabinanase from A. aculeatus using similar substrate concentration of 1 .4% (w/v) and longer incubation time of 16 h as compared to our study. Under these conditions, lower conversion yields of 38% and 44% were reported for the arabinan and galactan chains, respectively. On the other hand, others have studied the hydrolysis of galactan by endo-1 ,4-p-galactanase from E. nidulans and obtained yield of 28.7% and 67.0% for generation of end-products with MW of <3 kDa (60 min; enzyme/substrate ratio of 0.03% v/w) and <10 kDa (15 min; enzyme/substrate ratio of 0.3% v/w), respectively. These authors have also reported that the hydrolysis of potato pulp polysaccharides (15 min; enzyme/substrate ratio of 0.3% v/w) resulted in the release of two fractions of <10 kDa with a yield of 21 % and >10 kDa with the yield of 65%. However, higher yields of 67-99.7% were reported when ultrafiltration dead-end membrane reactor to convert low and high methoxyl pectin from citrus and apple to pectic oligosaccharides using endo- polygalacturonase from A pulverulentus were used. Substrate concentration of 1 % to 5% w/v, enzyme concentration of 90 to 2700 U/l, and residence time of 40 to120 min were used 5 by these authors.

Regression analysis. The multiple regression analysis was carried out using Design-Expert software version 8.0.2. , and the best fitting models were selected based on F and P values (Table 9). The reduced quadratic model was statistically significant for all three responses, including the yield of low-MW (F-value of 57.4 and P valueof < 0.0001) and high-MW (F-

10 value of 56.6 and P valueof < 0.0001) oligosaccharides as well as that of oligomers (F-value

of 57.7 and P value of < 0.0001). The coefficient of determination (R²) of all fitted models was 0.96, indicating the appropriateness of all models to predict the responses. The lack of fit was not significant relative to pure error with F values of 1 .3-2.7 and P values of 0.052-0.299; these results also indicate a good quality of the fit. Table 9 also shows F and P values for

15 each coefficient. Model coefficients with F-value higher than 1 and P value less than 0.05

were considered significant.

Substrate concentration (%, w/v) was the most significant linear term in the low-MW (x₃, F value of 188.1 ; P value of <0.0001) and high-MW (x₃, F value of 267.6; P value of <0.0001) oligosaccharide yield predictive models. The linear term with the largest effect on the

20 oligomer yield was Depol 670L/Gamanase 1 ,5L ratio (x₂, F value of 209.2; P value of

<0.0001). This demonstrates the significance of the synergistic interaction between Depol 670L and Gamanase 1 ,5L for the hydrolysis of potato RG l-rich pectic polysaccharides. In terms of quadratic terms, substrate concentration (%, w/v, x ) was the most significant in all predictive models (F value of 61 .0- 88.1 and P value of <0.0001). Among all interactive

25 effects, the most important one for the low and high-MW oligosaccharide models was the

interaction between reaction time and substrate concentration (x_? x₃) (F value of 20.7-26.7; P value of <0.0001).

Table 9. The analysis of variance for response surface model of oligosaccharides yield with DP of 2-6, 7-12 and 13-70.

Response DP 2-6 ^b DP 7-12 ^b DP 13-70 ^b

F Value P Value F Value P Value F Value P Value

Model 57.4 < 0.0001 56.6 < 0.0001 57.7 < 0.0001

XrTime (h) 5.7 0.0258 16.1 0.0005 170.0 < 0.0001 x₂- Depol 670L/Gamanase 1 .5L (U/U) ^a 1 1 .9 0.0022 0.5 0.491 1 209.2 < 0.0001 x₃-Substrate concentration (%, w/v) 188.1 < 0.0001 267.6 < 0.0001 1 .0 0.3174 Response DP 2-6 ^b DP 7-12 b DP 13-70 b

F Value P Value F Value P Value F Value P Value

X₁ x₂ 19.1 0.0002 2.0 0.1664 9.5 0.0052

Xi X3 26.7 < 0.0001 20.9 0.0001 6.7 0.0166

X2 X3 2.9 0.102 0.8 0.388 1 .0 0.3249

X1² 37.5 < 0.0001 22.2 < 0.0001 5.4 0.03

X2² 20.2 0.0002 29.7 < 0.0001 0.1 0.7532

X3² 88.1 < 0.0001 61 .0 < 0.0001 69.9 < 0.0001

Residual

Lack of Fit 1 .3 0.2992 2.7 0.052 2.7 0.0544 ^a R² = 0.96 for all three responses.

b The enzymatic activity unit (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction.

P < 0.05 indicates statistical significance.

5 Therefore, the time courses for the release of low and high-MW oligosaccharides seem to be

more dependent on the substrate concentration than Depol 670L/Gamanase 1 ,5L ratio. On the other hand, the significant time/enzyme ratio interactive effect (F value of 9.5; P value of 0.0052) on the yield of oligomers reveals the dependence of their time course on Depol 670L/ Gamanase 1 ,5L G ratio.

10 The following equations (2, 3, and 4) show the quadratic models for the yields of low-MW and

high-MW oligosaccharides as well as for that of oligomers in terms of coded factors. Model terms that were not statistically significant (p > 0.05) were removed.

Low- M oligosaccliarid.e3 yield, g/lQQg) = 4,8 -t- 0,43-_! - Q.5x₃ - 2,Gs,₃ - $ ι ₂ - Ll-tiSta - Q.¾i²- Q.&¾² -t- 1.3¾²

(111.2)

High - M a.igo3acc ha rides (yield, g/lQQg) = 8.6 - LS^ - 0.3¾ - 7.3¾ 4- - -15 Qx^- 2.3¾² 4- 3.3x₃ ²

(111.3)

O]igomera yield,g/100§) = 3&,4 + 1 .1s,_! - 15.&3.₂ +^■ 1-1*3 +^■ 4.7¾_!¾ +^■ ΐ,^χ^ +^■

2.Ζχ!² - 8.72.₃ ²

(111.4)

Effects of reaction parameters on the yield of oligosaccharides and oligomers. The 20 relationships between the reaction parameters and the yields of produced oligosaccharides

and oligomers can be better understood by studying the contour plots generated from the predicted models. The contour plots (not shown) illustrate the interaction effect of reaction time and substrate concentration on the predicted yields of low-MW and high-MW oligosaccharides, respectively, at constant Depol 670L/ Gamanase 1 ,5L ratios of 50/50, 62.5/37.5, and 75/25 (-1 , 0, and +1 levels). The highest yields of high-MW oligosaccharides were achieved at the initial stage of reaction (< 3 h) upon the use of low substrate concentrations (<0.5 %, w/v); while the low-MW oligosaccharides were dominant at the last stage of reaction (>3 h) at low substrate concentrations (<0.5 %, w/v). At the low range of substrate concentration (<1.7%, w/v), the yield of low-MW oligosaccharides increases with the increase of reaction time up to 5 h and with the decrease of substrate concentration to 0.3%.

These results reveal the adverse interactive effect of reaction time and substrate concentration on low-MW oligosaccharides; however, this interaction was not demonstrated at prolonged reaction times. Contrary to the low-MW oligosaccharides, the reaction time exhibited higher negative effect on the yield of high-MW oligosaccharides at prolonged reaction times (>3 h); and no or little significant positive effect was observed at the early stage. These contour plot trends may reveal the further hydrolysis of high-MW oligosaccharides, accumulated at the early stage of the reaction, into low-MW ones at the last stage. The increase in the substrate concentration had a significant negative effect on the predicted yields of both low-MW and high-MW oligosaccharides; this can be attributed to (a) the enzyme inhibition by the substrate excess, (b) the occurrence of substrate diffusional limitations and/or (c) to the presence of substrate/substrate interactions hindering their access to the enzyme's active site. Increasing Depol 670L/Gamanase 1 ,5L ratio from 50/50 to 75/25 (U/U) decreased the space region corresponding to the maximum yield of low-MW oligosaccharides; these results can be due to the excessive hydrolysis of low-MW oligosaccharides in the presence of higher ratio of Depol 670L as demonstrated by the released monosaccharides (data not shown). Contrary to low-MW oligosaccharides, increasing Depol 670L/Gamanase 1 ,5L ratio increased the space region corresponding to maximum yield of high-MW oligosaccharides. These results reveal that an excess of Depol 670L over that of Gamanase 1 ,5L can favor the release of high-MW oligosaccharides. Both reaction time and enzyme ratio can be controlled in order to modulate the end-product profiles of oligosaccharides.

The oligomers contour plots (not shown) indicate that the interaction between the enzyme ratio and the reaction time was more significant at the middle investigated ranges. As an overall, increasing Depol 670L/ Gamanase 1 ,5L ratio resulted in a decrease in the yield of the oligomers; this effect of enzyme ratio was more significant at the early stage of reaction. These results may be explained by the occurrence of antagonistic interaction in the presence of an excess of Depol 670L and/or by the high efficiency of glycosyl-hydrolases expressed in Gamanase 1 ,5L to initiate the hydrolysis of the steric hindered regions of potato RG l-rich pectic polysaccharides. The effect of reaction time on the predicted oligomers yield was more pronounced at the last stage (> 2.9) in which an increase of the yield was observed. The hydrolysis of polysaccharide regions was enhanced at prolonged reaction times than at shorter ones. Moreover, increasing substrate concentration up to 1 .44 % (w/v) increased the yield of the oligomers. These results show that due to substrate inhibition, the further conversion of oligomers into oligosaccharides was limited resulting in the accumulation of oligomers. However further increases of the substrate caused slight decrease in the space region corresponding to maximum yield, which shows that substrate inhibition hinders the conversion of polysaccharide to oligomers. Substrate inhibition has been reported as high substrate concentration had negative effect on product conversion when all other factors where kept constant.

Optimum conditions. Using Design-Expert 8.0.2 software, the developed models were used to determine the optimum conditions for the production of each of low-MW and high-MW oligosaccharides as well as oligomers. Table 10 shows the yield of low and high-MW oligosaccharides and oligomers as well as the monosaccharide profile of hydrolysates produced under the identified optimal conditions. Predicted values obtained for the yields of oligosaccharides and oligomers were no significantly different (P <0.05) from the experimental data; and all experimental findings are within the statistically significant ranges of the estimated optimum values with 95% prediction intervals, confirming the adequacy of the models (data not shown).

At all three identified optimum conditions, Depol 670L/Gamanase 1 .5L ratio was more or less at equal amount (52.0/48.0; 57.7/42.3; 50.0/50.0, U/U), revealing their synergistic interaction. In contrast, treatments no 23 and 24, in which Depol 670L was the dominant biocatalyst (87.5%), resulted in a very low yield of hydrolysates (0.1 -0.4%). Moreover, treatments no 21 and 22 with Gamanase 1 .5L being the dominant biocatalyst (62.5%) resulted in very low yields of both low and high-MW oligosaccharides (0.0-3.7%) and high yield of oligomers (69.4-72.9%). As expected, the highest amount of low-MW oligosaccharides (1 1 .2%) was obtained when longer incubation time (5.0 h) and smaller substrate concentration (0.9%, w/v) was used (Opt 1). Higher yield of high-MW oligosaccharides was obtained upon the decrease of incubation time to 2 h (Opt 2, yield of 21 .9%); while the highest yield of oligomers was obtained when the substrate concentration was increased to 1 .7 % (Opt 3, 67%), indicating the reduced hydrolysis rate.

Table 10 also shows the monosaccharide profile of the hydrolysates, obtained using the optimum conditions. The highest proportions of Rha, Ara and GalA (2.3, 1 1 .3 and 8.1 %, Table 10. Optimum conditions resulting in highest yield of oligosaccharides with DP of 2-6 (Opt 1), 7-12 (Opt 2) and 13-70 (Opt 3), corresponding responses and their monosaccharide profile.

Sample RG^a Opt 1 Opt 2 Opt 3

Time (h) 5 2 5

Incubation condition Depol 670L/Gamanase 1.5L (U/U) ^{b c} 52.0/48.0 57.7/42.3 50.0/50.0

Substrate concentration (%,w/v) 0.9 0.9 1.7

DP2-6 1 1.2 3.7 2.9

Yield of oligosaccharides and oligomers (%) DP7-12 7.3 21.9 0.7

DP13-70 38.3 37.9 67.0

Total oligosaccharide/oligomer (%) 56.8 63.5 70.6

Rha 2.2 2.3 0.0 1.7

Ara 5.7 1 1.3 7.2 7.1

Gal 63.1 72.9 77.0 75.0

Sugar composition (%) Glc 13.2 3.2 7.9 6.5

Xyl 4.9 0.5 0.0 0.9

Man 3.0 1.8 0.0 1.9

GalA 7.9 8.1 7.9 6.9 a RG was extracted using microwave assisted -alkaline extraction method (KOH concentration: 1.5M, time: 2.0 min, power: 36.0 W, solid/liquid: 2.9 %).

bThe enzymatic activity unit (U) was defined as the amount of enzyme that produces 1 μηιοΐβ of reducing ends per min of reaction. c Total enzymatic unit (expressed in galactanase unit) was kept constant at 0.2 U/mg RG l-rich pectic polysaccharides.

respectively) were obtained at Opt 1 condition identified for the optimal generation of low-MW oligosaccharides. These results reveal that higher proportion of low-MW oligosaccharides is originated from HG region and arabinan side chains as compared to high-MW oligosaccharides and oligomers. Moreover, higher proportion of Rha at Opt 1 condition indicates more breakdown of RG I backbone, when longer reaction tine is used. The highest proportions of Gal (75.0-77.0%) and Glc (6.5-7.9%) were obtained at Opt 2 and Opt 3 conditions corresponding to optimum ones for the release of high-MW oligosaccharides and oligomers, respectively. Opt 3 condition resulted in the highest total yield (oligosaccharides/oligomers); these results confirm the efficient hydrolysis of galactan-rich RG I in the presence of equal amount of Depol 670L/Gamanase 1 .5L (50/50, U/U) and upon longer reaction time. As compared to potato RG l-rich pectic polysaccharides, oligosaccharides and oligomers were enriched with Ara and Gal, revealing the efficient hydrolysis of side chains of RG I. However, the proportion of hemicellulosic monosaccharides in hydrolysates was lower than that of potato RG l-rich pectic polysaccharides (21 .1 %), indicating the limited hydrolysis of the hemicellulosic polysaccharides.

A bi-enzymatic system was studied for the generation of Gal-rich oligosaccharides/oligomers with defined DP. The effects of the reaction conditions, including reaction time, Depol 670L/Gamanase 1 .5L ratio and substrate concentration, on the yield, the monosaccharide profile and the DP of generated oligosaccharides/oligomers were investigated. The results showed that the highest yields were obtained for both oligosaccharides and oligomers when equal enzymatic unit of Depol 670L and Gamanase 1 .5L was used in combination, confirming their synergistic interactions. The highest amount of oligosaccharides with DP of 2-6 was obtained when longer incubation time and smaller substrate concentration were used. Decreasing incubation time resulted in the highest yield of oligosaccharides with DP of 7-12 while increasing substrate concentration led to the highest yield of oligomers with of DP of 13-70.

EXAMPLE IV - DIGESTIBILITY AND PREBIOTIC PROPERTIES OF POTATO

RHAMNOGALACTURONAN AND ITS OLIGOSACCHARIDES

Materials. KOH, sodium acetate, acetic acid, peptone, L-cysteine HCI, Tris-EDTA, dextran standards, Gal, Ara, Rha, Glc, Xyl, Man, GalA, gellan and xanthan gums, thermostable a- amylase from B. licheniformis, amyloglucosidase from A. niger, lysozyme, mutanolysin, proteinase K, pepsin from porcine gastric mucosa, porcine bile extract and pancreatin (4xUSP, P1750, from porcine pancreas) were purchased from Sigma-Aldrich. All other salts were obtained from Fisher Scientific (Fair Lawn, NJ). Lipase (150 units/mg, from Rhizopus oryzae DF 15) was obtained from Amano Enzyme USA Co. and human saliva was from Lee BioSolutions (991 -05-P). Endo-galactanase (A niger) was from Megazyme and Depol 670L (7. reesei) was from Biocatalysts Ltd. (Mid Glamorgan, UK). Propidium monoazide was obtained from Biotium, (Hayward, CA) and fast SYBR Green mastermix was from Life technologies (Carlsbad, CA). 1 -kestose, nystose and 1 F-fructofuranosyl-nystose were obtained from Wako Pure Chemical (Japan) and FOS was from Orafti (P95, Belgium).

Preparation of potato RG l-type polysaccharide extract. Potato RG l-type polysaccharides were prepared by microwave-assisted alkaline treatment of potato pulp (Lyckeby Starch AB), according to our modified method (Khodaei et al. , 2015). The used optimum conditions were: potassium hydroxide concentration of 1 .5 M, microwave treatment time of 2 min, power/suspension ratio of 1 .2 W/mL and solid/liquid ratio of 2.9% (w/v). The recovered supernatants after centrifugation of treated suspensions (8000 ^χ g, 25 min) were neutralized. In order to remove starch from the recovered extract, thermostable a-amylase from B. licheniformis and amyloglucosidase from A. niger (>300 units/mL,) were added to yield a concentration of 1 .0 and 2.2 mL per g of extract, respectively. After incubation at 40 °C under 150 rpm agitation for 16 h, the monosaccharides and hydrolysed products were removed by ultrafiltration through a membrane with 10 kDa cut-off (EMD Millipore Prep/Scale spiral- wound, TFF-2,) using a EMD Millipore Pellicon™ easy-load peristaltic system. The high molecular weight (MW) polysaccharides in the retentate were recovered by ethanol precipitation (1 : 1 , v:v), followed by centrifugation at 8,000 ^χ g for 25 min.

Generation of oligosaccharides from potato RG l-type polysaccharides. Potato RG l-type polysaccharide suspension (3%, w/v) in sodium acetate buffer (50 mM, pH 5.0) was subjected to enzymatic treatment (0.2 U/mg substrate) with endo-galactanase as mono- component enzyme or Depol 670L as a multi-enzymatic preparation. After incubation at 40°C for 4 h, the unhydrolyzed polysaccharides were precipitated by ethanol (1 : 1 , v:v). The hydrolysates, recovered upon centrifugation (8,000 ^χ g, 25 min) were analyzed for their MW profile and sugar composition.

Monosaccharide profile analysis. The monosaccharide profile of potato RG l-type polysaccharides and their hydrolysates was analyzed as described by Khodaei & Karboune, 2013. GalA content was measured using sulphamate/m-hydroxydiphenyl assay (Blumenkrantz & Asboe-Hansen, 1973).

Determination of MW distribution. MW distribution of potato RG l-type polysaccharides and their hydrolysates were analyzed using Waters HPLC system equipped with refractive index detector. For the polysaccharide fractions, three columns in series (TSK G3000 PWXL, TSK G4000 PWXL and TSK G5000 PWXL) were used; while the oligosaccharides were analyzed on a TSKgel G-Oligo-PW column. All columns were from Tosoh Bioscience (Montgomeryville, PA). Isocratic elution with 0.1 M sodium chloride was applied and dextrans with MW of 50, 150, 270, 410 and 670 kDa for polysaccharide analysis; Gal, sucrose, 1 - kestose, nystose and 1 F-fructofuranosyl-nystose were used as standards for oligosaccharide analysis.

Evaluation of digestibility of RG I and generated oligosaccharides. The TIM-1 system (TNO, Nutrition and Food Research Institute, Zeist, Nehterland) was used in order to assess the digestibility of RG l-type polysaccharides and Depol-based RG I oligomers under intestinal conditions (Minekus et a/. , 1995) (Figure 5). The model consists of four compartments simulating stomach, duodenum, jejunum and ileum working under controlled temperature of 37°C. Initial pH in stomach was 5.5, which was dropped gradually to 2.7, 2.2 and 1 .8, by injecting 1 M hydrochloric acid, after 40, 60 and 90 min of digestion, respectively, and remained constant at 1 .7 from 120 to 300 min. The pH in duodenum, jejunum and ileum were maintained at 6.5, 6.8 and 7.2, respectively, using 1 M sodium bicarbonate solution. Gastric secretions consisted of pepsin (0.19 g/l, >3200 units/mg solid) from porcine gastric mucosa and lipase (0.25 g/l, 150 units/mg protein), both in an electrolyte solution (NaCI, 4.8 g/l; KCI, 2.2 g/l; CaCI₂, 0.3 g/l; NaHC0₃, 1 .25 g/l; pH 4.0) delivered at flow rates of 0.25 mL/min. Duodenal secretions were composed of 21 % (w/w) pancreatin solution in distilled water and 4% (0-30 min) or 2 % (30-300 min) (w/w) porcine bile extract as well as small intestine electrolyte solution (NaCI, 5.0 g/l; KCI, 0.60 g/l; CaCI₂, 0.30 g/l; pH 7.0). Contents of jejunum and ileum were dialyzed through hollow fiber membranes (cut-off 1 1 .8 kDa) against small intestinal electrolyte solution and the compounds that passed through membrane were collected for further analysis (jejunum and ileum dialysates). The residue that exited from ileum (effluent) and solution remaining in duodenal, jejunal and ileal compartments at the end of the digestion (chyme) were collected and considered to be indigestible. The chymes and effluent residues were heated (75°C/5 min) immediately following removal from TIM-1 and all collected samples after 300 min of digestion were freeze dried and analyzed for their MW and sugar composition.

Immobilization of fecal inoculum in gel beads. Human faeces was collected from one volunteer, who has not been on antibiotics for at least 3 months and had no history of gastrointestinal disorders. Fecal sample was suspended in peptone water (0.1 %, w/v) containing 0.05% (w/v) L-cysteine HCI at 37°C. The fecal inoculum was entrapped into gum- based beads according to the dispersion method of Cinquin et a/. , 2004. Briefly, the fecal inoculum (2%, w/v) was added to sterilized gum solution, made of gellan (2.5%, w/v) and xanthan gum (2.5% w/v), in sodium citrate solution (0.2%, w/v). Then the inoculated gum suspension was added to a hydrophobic phase (commercial canola oil) at a ratio of 1 :2 (v/v). Beads with diameter of 1 -2 mm were separated by sieving Fecal batch culture fermentation. A stirred glass bioreactor (BIOSTAT® Qplus, Sartorius AG, Goettingen, Germany) was used to investigate the fermentability of potato RG l-type polysaccharides and their corresponding hydrolysates by immobilized fecal inoculum (62.5g of gel beads for 187.5 mL of broth). The colonic broth developed by Macfarlane et a/. , 1998, supplemented with vitamin solution (0.5 mL/l), 0.05% L-cysteine-HCI (16.0 mL/l) and bile salt (2.0 g/l), was continuously fed into the bioreactor with a flow rate of 0.35 mL/min resulting in a retention time of 12 h. The vitamin solution was composed of pyridoxine HCI (20 mg/l), p- aminobenzoic acid (10 mg/l), nicotinic acid (10 mg/l), biotin (4 mg/l), folic acid (4 mg/l), vitamin B12 (1 mg/l); thiamine 8 mg/l ; riboflavin 10 mg/l; menadione 2 mg/l; vitamin K1 0.005 mg/l; pantothenate 20 mg/l (Gibson & Wang, 1994). Culture temperature and pH were maintained at 37 °C and 6.2, respectively. Anaerobic conditions were kept by constant flow of oxygen free carbon dioxide gas in the reactor. After 16 days of stabilization, 75% of carbohydrate source of culture medium were substituted with potato RG l-type polysaccharides, Depol-based hydrolysates, galactanase-based hydrolysates and fructo- oligosaccharides as standard reference. A fermentation with only 25% carbohydrate was run as a negative control. The experimental setup design is shown in Figure 6.

Bacterial enumeration - Bacterial strains and growth conditions of reference strain. E. coli ATCC 25922 and E. faecalis ATCC 27275 were grown aerobically at 37°C for 24 h in brain- heart infusion broth (Difco Laboratories) and tryptic soy broth supplemented with 0.6 % yeast extract, respectively. Bifidobacterium adolescentis ATCC 15703, Bacteroides thetaiotaomicron ATCC 29741 , Clostridium leptum group ATCC 29065 and Blautia cocoides ATCC 29236 were cultured at 37°C aerobically (Forma scientific anaerobic system Model 1025; Forma Scientific, Marietta, OH, USA) using the MRS broth supplemented with 0.05 % L-cysteine-HCI (sigma) for Bifidobacterium, brain-heart infusion broth supplemented with 0.05 % L-cysteine-HCI for Bacteroides and Blautia cocoides and a modified chopped meat medium with maltose (ATCC medium 2751) for Clostidium leptum. All these strains were obtained from the American Type Culture Collection (Rockville, MD, USA), and sub-cultured at least three times before use.

Bacterial enumeration - Sample collection and DNA extraction. The enumeration of the target fecal bacterial groups was performed by quantitative polymerase chain reaction coupling with propidium monoazide treatment (PMA-qPCR). Fermentation samples were collected every day from bioreactor and 2.5 μί of propidium monoazide solution (20 mM in 20 % dimethylsulfoxide) was added to 1 mL of diluted sample. Samples were incubated for 5 min in the dark followed by incubation for 5 min under halogen light (Fujimoto et a/. , 201 1). DNA was extracted as described by Ahlroos & Tynkkynen, 2009 using genomic DNA Purification Kit (Promega, Madison, Wl, USA) with some modifications. The cells were washed three times with Tris-EDTA buffer (20 mM Tris-HCI, 2 mM EDTA), pellets obtained after centrifugation were suspended in 200 μΙ Tris-EDTA buffer containing 40 mg/mL of lysozyme, 200 U/mL of mutanolysin and 4 μg/mL of proteinase K. After 1 h incubation at 37°C, the DNA was recovered. 5 μί of diluted DNA (10x) was added to the reaction mixture (20 μί) 5 consisted of 12.5 μΙ of SYBR™ green, 1 μΙ_ of each forward and reverse primer at 5 μΜ and 5.5 μΙ_ of DNase free water. List of the specific primers used for the quantification of bacteria is presented in Table 1 1 .

Table 1 1 . Specific primers used for quantification of bacteria.

Species Gene Direction sequence 5'-3' (SEQ ID NO:) Amplicon (pb)

Bacteroides /

Prevottela / 16S F GGTGTCGG CTT AAG TG C CAT (1) 140

Porphiromonas

R CGGA(C/T)GTAAGGGCCGTGC (2)

Clostridium leptum

16S F GCACAAGCAGTGGAGT (3) 208 group

R CTTCCTCCG I I I TGTCAA (4)

Bifidobacterium

16S F TCGCGTC(C/T)GGTGTGAAAG (5) 243 spp.

R CCACATCCAGC(A/G)TCCAC (6)

Blautia cocoides 16S F AAATCACGGTACCTGACTAA (7) 440

R CTTTG AGTTCATTCTTG CG AA (8)

Lactobacillus spp. 16S F TGGATGCCTTGGCACTAGGA (9) 92

R AAATCTCCGGATCAAAGCTTACTTAT (10)

Enterobacteriaceae 16S F CAT GCC GCG TGT ATG AAG AA (1 1) 96

R CGG GTA ACG TCA ATG AGC AAA (12)

Enterococcus spp. 16S F CCCTTATTGTTAGTTGCCATCATT (13) 144

R ACTCGTTGTACTTCCCATTGT (14)

Analysis was performed on an ABI 7500 fast real-time PCR system (Applied Biosystems, Streetsville, ON, Canada) in 96-well plates. Prebiotic index (PI) was estimated as follow (Manderson et a/. , 2005):

(d logic Bifidobacter rti) x— tQ (Δ logig Lactobacillvsrytx— tQ

15 Where t₀ is last the day of stabilization (day 16) and t_x is last day of each fermentation period (days 20, 24, 28, 32, 36, 38, 40, 44, 48, 52 and 56).

Short chain fatty acids (SCFA) analysis. Fermentation samples recovered at selected fermentation period (16-56 days) were centrifuged (10,000 ^χ g, 10 min, 4°C) and were injected to a HPLC system (Waters, Milford, MA, USA) equipped with an ICSep-lon-300 (300 x 7.8 mm) column (Transgenomics, San Jose, CA, USA) and a differential refractometer detector (Model 2142, LKB, Bromma, Sweden). The eluent was 0.0065 N (or 3.25 mM) sulfuric acid.

Properties of RG I and generated oligo-RG I. The structural properties of potato RG I and its hydrolysates (Dep-based oligo-RG I; Gal-based oligo-RG I) are reported in Table 12. Because of the efficient removal of hemicellulosic fragments by ultrafiltration, the RG I extract exhibited lower molar proportions of hemicellulosic (Glc, Xyl and Man) monosaccharides (5.3%) as compared to that obtained in a previous study (21.1 %) (Khodaei and Karboune, 2015b).

Table 12. Molecular weight distribution and sugar composition of RG l^a, Dep^b and Gal-based oligo-RG l^c

Molecular Weight Distribution (% w/w)

Samples Oligosaccharides -DP Oligosaccharide Oligomers

1 2 3 4 5 6 7 8 (DP 2-12) (DP of 13-70) Polysacchar

RG I 2.7 - 2.9 - 0.8 0 - 1.4 5.1 1.4 90.8

Dep-based oligo-RG I 6.3 - - - 26.3 24.9 - 0.0 51.2 42.5 -

Gal-based oligo-RG I 6.1 1.1 19 - 10.6 - - 12.6 43.3 50.6 -

Sugar Composition (mg/g)

Rha Ara Gal Glc Xyl Man GalA

RG I 18.6 (1.9) ^d 55.2 (5.5) 800.2 (79.9) 34.4 (3.4) 7.4 (0.7) 8.5 (0.9) 76.7 (7.7)

Dep-based oligo-RG I 32.5 (1.6) 176.9 (8.8) 1465.2 (72.7) 120.8 (6.0) 34.2 (1.7) 10 (0.5) 176.9 (8.8)

Gal-based oligo-RG I 5.5 (1.6) 33.3 (9.8) 237.2 (70.0) 41.1 (12.1) 5.4 (1.6) 3.3 (1.0) 12.8 (3.8) ^a Extracted polysaccharides from potato pulp using microwave-assisted alkaline treatment.

b Generated oligosaccharide from potato RG I upon treatment with Depol 670L.

c Generated oligosaccharides from potato RG I upon treatment with endo-p-1 ,4-galactanase.

d Monosaccharide composition of RG I, Dep and Gal-based oligo-RG I expressed in weight percentage.

The results show that the monosaccharide profiles of RG I, Dep-based oligo-RG I and Gal- based oligo-RG I were more or less similar with Gal being the main monosaccharide (70.0- 79.9%). Slightly higher proportion of Gal in RG I as compared to oligo-RG I may be attributed to the incomplete hydrolysis of galactan side chains. Indeed, endo-galactanase catalyzes the cleavage of p-(1 -4)-D-Gal linkages of galactan branches with less steric hindrance. The results also reveal the slightly higher proportion of Ara, in the oligo-RG I extracts (8.8-9.8%) as compared to RG I (5.5%) which shows efficient hydrolysis of arabinan side chains. Since the galactanase mono-component enzyme exhibited low pectinase activity, the resulted oligo-RG I had low GalA proportion (3.8%) as compared to RG I (7.7%).

Table 12 shows MW distribution of RG I, Dep-based oligo-RG I and Gal-based oligo-RG I. As expected major part of the RG I sample were polysaccharides (90.8%). Dep-based oligo-RG I and Gal-based oligo-RG I did not contain any polysaccharide, but they were composed of 42.5 and 50.6% of oligomers (DP of 13-70), respectively. Major oligosaccharides (DP of 2- 12) in Dep-based extract were those with DP of 5 (26.3%) and 6 (24.9%); while Gal-based oligosaccharide proportion was mainly composed of oligosaccharides with DP of 3 (19.0%), 5 (10.6%) and 8 (12.6%). Others have investigated the production of prebiotic compounds, by endo-1 ,4-p-galactanase from Emericella nidulans, from enzymatically solubilized β-1 ,4- galactan-rich potato pulp polysaccharides (>100 kDa). They have reported the recovery of two main fractions of less (20.6%, w/w) and more than 10 kDa (65.0%, w/w) characterized by a Gal molar proportion of 92.9 and 53.9 %.

Assessment of digestibly of RG I polysaccharides and generated oligosaccharides. In order to determine the colonic persistence, the digestibility of a mixture of RG I and Dep-based oligo-RG I (1 : 1) was assessed using the TIM-1 model. The MW distributions and the monosaccharides composition of the initial fed (stomach) and digestas are reported in Table 13. The mixture characterized at the stomach level corresponds to the properties of initial products. The initial mixture contained 28.2% and 22.0% of oligosaccharides (DP of 2-10) and oligomers, respectively (mostly originating from Dep-based oligo-RG I) and 45.4% of polysaccharides (originating from RG I). As expected, the monosaccharide composition of the mixture shows the abundance of Gal (75.3 %) followed with lower amount of GalA (8.3 %) and Ara (7.7 %). 86% of the initial amount (31 .3 g) fed to TIM-1 system (stomach) were recovered in the digesta, including: jejunum/ileum dialysates, chyme (remaining solution in duodenal/jejunal/ileal compartments) and effluent; the remaining amount (14%) may have been lost during handling and/or left in the tubing. Considering the 1 1 .8 kDa cut-off in hollow fiber membranes, oligosaccharides part of the mixture would enter jejunum and ileum dialysates regardless of their hydrolysis rate. However, the low recovery extent of larger oligomers (DP of 13-70) and polysaccharides in chyme and effluent can provide an indication of their hydrolysis. The initial carbohydrate mixture contained 6.9 and 14.2 g of oligomers (DP of 13-70) and polysaccharides, respectively, corresponding to 22.0 and 45.4% of initially fed carbohydrate. While the total amount of oligomers (DP of 13-70) and polysaccharides recovered in chyme and effluent was 4.8 and 9.7 g, respectively, representing 81 .6 and 79.3% of their initial amount. These results reveal that most of RG I and oligo-RG I remained undigested. The partial observed hydrolysis of oligomers and polysaccharides may be due to low pH of stomach. Similarly, it has been reported that when inulin or oligofructoses were fed to ileostomy subjects, 86-89% of the original compounds were recovered at the terminal ileum. Moreover, others who have studied the ileal digestibility of two galacto- oligosaccharides derived from lactulose (DP >2, 14.0% trisaccharides) and lactose (DP >3, 35.1 % trisaccharides) in growing rats, have reported that 12.5% and 52.9% of the trisaccharide fraction of lactulose and lactose-derived oligosaccharides were hydrolyzed, and the disaccharide fraction of lactulose derived oligosaccharides was fully resistant to the hydrolysis of digestive tract.

Table 13. Molecular weight distribution and sugar composition of carbohydrates (stomach: a mix of RG l^a and Dep-based oligo-RG l^b, 1 : 1) fed to the TIM-1 system and their recovered digesta (jejunum and ileum dialysate, chyme^c, and effluent^d).

Molecular Weight Distribution (expressed in g)

Samples Oligosaccharides - DP Oligosaccharides Oligomers

1 2 3 4 5 6 7 8 9 10 (DP 2-12) (DP of 13-70)

Stomach 0.4 - 0.5 - 3.8 3.9 - 0.2 - - 8.8 (28.2)^e 6.9 (22.0) 14.2 (45.4

Chyme 0.1 0.1 0.1 0.1 0.4 (1 .4)' 2.2 (8.2) 2.6 (9.7)

Small Jejunum dialysate 0.4 4.3 - 0.1 - - 0.4 - 0.3 5.1 (19.2)

Intestine Ileum dialysate 5.6 5.7 (21 .1 )

Effluent 0.1 0.2 0.1 0.2 0.5 (1.8) 2.6 (9.7) 7.1 (26.3

Sugar Composition (expressed in g)

Rha Ara Gal Glc Xyl Man GalA

Total fed 0.5 (1 .8)⁹ 2.2 (7.2) 23.9 (76.3) 1.5 (4.7) 0.4 (1 .2) 0.2 (0.7) 2.6 (8.3)

Polysaccharide and

Stomach 0.4 (1 .8) 1.4 (6.6) 16.4 (77.5) 0.9 (4.2) 0.2 (1.0) 0.2 (0.8) 1 .7 (8.1 )

oligomer fraction

Oligosaccharide

0.2 (1 .6) 0.9 (8.8) 7.4 (72.7) 0.6 (6.0) 0.2 (1.7) 0.1 (0.5) 0.9 (8.8)

fraction

Chyme 0.0 (0.0) 0.0 (0.0) 4.6 (88.5) 0.3 (5.8) 0.0 (0.0) 0.0 (0.0) 0.3 (5.8)

Ileum dialysate 0.1 (1 .8) 0.2 (3.6) 4.4 (78.6) 0.3 (5.4) 0.1 (1.8) 0.0 (0.0) 0.5 (8.9)

Small

Intestine Jejunum dialysate 0.1 (1 .7) 0.2 (3.4) 4.5 (77.6) 0.3 (5.2) 0.1 (1.7) 0.1 (1 .7) 0.5 (8.6)

Effluent 0.3 (2.9) 2.0 (19.4) 6.7 (65.0) 0.4 (3.9) 0.1 (1 .0) 0.1 (1 .0) 0.7 (6.8)

a Extracted polysaccharides from potato pulp using microwave-assisted alkaline treatment.

Generated oligosaccharides from potato RG I upon treatment with Depol 670L.

c Solution remaining in duodenal, jejunal and ileal compartments at the end of the digestion.

d The residue exited from ileum.

e Molecular weight distribution of the sample fed to the TIM-1 expressed in weight percentage of initial amount (stomach, 31.3 g).

f Molecular weight distribution of digestas expressed as weight percentage of total recovered digesta (chyme + ileum dialysate + jejunum dialysate + effluent, 26.8 g).

ffJVlonosaccharide composition of the samples expressed in weight percentage.

Table 13 also indicates that the amount of monosaccharides in the total recovered carbohydrates (0.63 g, 2.4%) did not increase significantly as compared to the initial amount (0.4 g, 1.3%); these results reveal that major part of hydrolyzed polysaccharides and oligomers were converted into oligosaccharides, which themselves have prebiotic properties. Total amount of oligosaccharides (DP of 2-12) in the initial mixture was 8.8 g (28.1 %) and was increased to 1 1.7 g (43.7%) in the total recovered carbohydrate. Due to large cut-off in membrane of TIM 1 (1 1.8 kDa), these oligosaccharides were absorbed through jejunum and ileum dialysates; however, in human intestine, all carbohydrates such as sucrose, lactose, and starch must be hydrolyzed to monosaccharides before they can be absorbed in the small intestine. Disaccharides constituted the major part of the oligosaccharides collected from jejunum (4.3 g, 16%) and ileum (5.6 g, 20.9%) dialysates.

The sugar composition of the digesta shows that the amounts of Rha and Ara recovered in the ileum/jejunum dialysates (0.2 and 0.4 g) were similar or lower than the initial fed amounts corresponding to the oligosaccharide fraction (0.2 and 0.9 g); while their amounts in the chyme/effluent (0.3 and 2.0 g) were similar or higher than those present in the polysaccharide and oligomer fractions of the initial fed (0.4 and 1.4 g). These results reveal the low hydrolysis extent of RG I backbone and arabinan side chains during digestion. The decrease in the amount of Ara may be due incomplete transition of the components from membrane and/or their entrapment in the digested materials. In contrast, 7.4 g of Gal in the fed was originated from the oligosaccharides, which slightly increased to 8.9 g in the ileum/jejunum dialysates, indicating a mild hydrolysis of galactan side chains. As a consequence of this hydrolysis, the amount of Gal in the chyme/effluent (1 1.3 g) was lower than the initial amount present in the polysaccharide and oligomer fractions of the initial fed (16.4 g). The amounts of hemicellulosic (Glc, Xyl, Man) monosaccharides (0.6, 0.2, 0.1 g) recovered at the ileum/jejunum dialysates were similar to the initial amounts present in the oligosaccharide fraction of the fed; this confirms the resistance of these polysaccharides to digestion. Therefore, the amounts of these monosaccharides in the chyme/effluent (0.7, 0.1 and 0.1 g) did not show significant difference with those present in the polysaccharide and oligomer fractions of the initial fed (0.9, 0.2 and 0.2 g). The amount of GalA recovered in the ileum/jejunum dialysates (1.0 g) was higher than its corresponding initial amount originating from the oligosaccharide fraction of the fed (0.9 g), while its amount in the chyme/effluent (1.0 g) was lower than its initial amount in the polysaccharide/oligomer fraction (1.7 g), indicating the hydrolysis of HG region during digestion. Since an increase in the amount of Rha was not observed in parallel with GalA, it can be concluded that increased amount of GalA in the ileum/jejunum dialysates was originated from HG and not from RG I backbone. Measurement of prebiotic properties of RG I and generated oligo-RG I. The prebiotic activities of potato pectic-type RG I, Dep-based oligo-RG I and Gal-based oligo-RG I were assessed by measuring their abilities to support the growth of mixed bacterial populations. Figure 6 shows the scheme of the continuous fermentations in which 75% of carbohydrate source of the medium was substituted with RG I, oligo-RG I and FOSs. Changes in bacterial population (human faecal inocula) were daily determined using PMA-qPCR. Bifidobacterium and Lactobacillus consist of 4% and less than 2% of the total intestinal microbiota in adults, respectively; the effect of increasing their numbers, as a result of consuming prebiotics, on the intestinal health is well documented. Well-established prebiotic FOS was selected as positive control because it is well known to increase growth rate of Bifidobacterium and Lactobacillus and decrease that of Clostridium.

Table 14 shows the bacterial counts at the last day of fermentation and the changes occurring as result of the substitution by each of the investigated carbohydrate sources. Changes smaller than 0.5 log was not considered significant. Looking at the beneficial bacteria, the populations of both Bifidobacterium and Lactobacillus increased during the fermentation of all investigated carbohydrate sources. Indeed, Bifidobacterium and Lactobacillus spp. were reported to be able to express β-galactosidase (EC 3.2.1 .23), β-1 ,3- galactosidase (EC 3.2.1 .145), and endo-β-1 ,4-galactanase (EC 3.2.1 .89) activities (Thomassen et al. 201 1). As an overall, the increase in the bacterial population of Lactobacillus (0.8-2.2 log) was higher as compared to that of Bifidobacterium (0.6-1 .5 log). For Bifidobacterium, the highest increase was achieved with Dep-based oligo-RG I (1 .4 log) and FOSs (1 .5 log) followed by Gal-based oligo-RG I (1.3 log). These results reveal that oligo-RG I can be considered bifidogenic under the investigated conditions. On the other hand, Gal-based oligo-RG I resulted in higher increase in Lactobacillus (2.2 log) as compared to Dep-based oligo-RG I (2.0 log) and FOS (2.0 log). However, the use of non-hydrolyzed RG I resulted in an increase of Bifidobacterium and Lactobacillus populations with a smaller extent (0.6 log) as compared to their corresponding hydrolysates. Differences observed between RG I and its oligosaccharides/oligomers may be due to the complex structure and lower solubility of RG I (Van Loo 2004). It has been reported that low-MW oligosaccharides were more selectively fermented by Bifidobacterium than their parent high-MW polysaccharides (sugar beet arabinan and fructans). In addition, RG I backbone may be less fermentable as compared to RG I side chains. Oligosaccharides derived from side chains of RG I ((arabino)galacto-oligosaccharides and arabino-oligosaccharides) are more fermented by Bifidobacterium than those from backbone of RG I (rhamnogalacturono-oligosaccharides and galacturono-oligosaccharides). Table 14. Bacterial populations at the last day of fermentation for each sample (Log cells/ml) and change in bacterial count during fermentation (difference between last day of each sample's fermentation and stabilization period).

Bacterial populations (Log cells/ml) Change in Bacterial Count

Bacterial species

Dep-based Gal-based Dep-based Gal-based

Fecal sample Stabilization RG l^a FOS^d RG I FOS oligo-RG l^b oligo-RG l^c oligo-RG I oligo-RG I

Bacteroides /

Prevottela / 10.02 ± 0.08 9.36 ± 0.04 9.7 ± 0.07 9.4 ± 0.13 9.3 ± 0.07 9.4 ± 0.17 0.3 0.1 0.0 0.0

Porphiromonas

Clostridium leptum 9.57 ± 0.07 8.06 ± 0.07 6.9 ± 0.12 7.3 ± 0.12 7.3 ± 0.16 7.4 ± 0.05 -1.2 -0.7 -0.7 -0.6 group

Bifidobacterium spp. 9.12 ± 0.07 5.54 ± 0.05 6.2 ± 0.18 7.0 ± 0.13 6.8 ± 0.07 7.0 ± 0.14 0.6 1.4 1.3 1.5

Blautia cocoides 9.19 ± 0.1 6.67 ± 0.07 7.3 ± 0.17 7.3 ± 0.03 7.3 ± 0.05 7.3 ± 0.13 0.7 0.6 0.6 0.6

Lactobacillus spp. 8.34 ± 0.06 4.80 ± 0.1 5.6 ± 0.14 6.8 ± 0.13 7.0 ± 0.12 6.8 ± 0.04 0.8 2.0 2.2 2.0

Enterobacteriaceae 7.93 ± 0.07 7.55 ± 0.07 8.3 ± 0.04 7.7 ± 0.14 7.9 ± 0.12 7.6 ± 0.07 0.8 0.2 0.3 0.1

Enterococcus spp. 5.94 ± 0.05 4.78 ± 0.09 6.1 ± 0.05 6.7 ± 0.02 6.8 ± 0.05 6.3 ± 0.11 1.3 1.9 2.0 1.5

Prebiotic index (PI) 0.3 0.5 0.5 0.5 a Extracted polysaccharides from potato pulp using microwave-assisted alkaline treatment.

b Generated oligosaccharide from potato RG I upon treatment with Depol 670L.

d Fructo-oligosaccharide.

The results of bacterial populations (Table 14) also show that Enterococcus exhibited a significant increase in its population, with 1.9 to 2.0 log unit increment, over the fermentation process of the investigated hydrolysates as carbohydrate sources. For Dep- and Gal-based oligo-RG I, the increase in the Enterococcus was higher than that of Bifidobacterium and lower than that of Lactobacillus. On the hand, the fermentation of FOS led to similar increase (1 .5 log) of Enterococcus population as compared to Bifidobacterium (1 .5 log) and to lower increase as compared to Lactobacillus (2.0 log). Potato RG I resulted in a higher increase of Enterococcus (1 .3 log) population as compared to Bifidobacterium and Lactobacillus. Table 14 indicates that Blautia cocoides population increased to equal extent when RG I (0.7 log), Dep- and Gal-based oligo-RG I and FOSs (0.6 log) were used. The MW of the investigated substrates seems not to affect the growth of Blautia. Enterobacteriaceae number increased when RG I was used (0.8 log), while they didn't show significant changes in their populations during the fermentation of oligo-RG I and FOS I (0.1 -0.3 log).

The results also show that the proliferation of Bacteroidetes and Clostridium leptum group was not supported by the investigated carbohydrates, and their populations either remained constant or decreased. More specifically, Bacteroides/Prevottela/Porphiromonas did not show significant changes in their populations except for RG I with a small increase of 0.3 log. Similar results have been reported in the literature demonstrating that Bacteroides are able to use polysaccharides, while Bifidobacterium prefer low-MW carbohydrates. Moreover, it was found that Bacteroides are able to ferment rhamnogalacturono-oligosaccharides, arabinogalactans and galacto-oligosaccharides and concluded that Bacteroides can express a wide variety of glycanases and glycosidases. These results are supported by the present example as the number of Bacteroides spp. didn't decrease during fermentation. Indeed, Bacteroides comprise one of the major phylum in the human gut flora (approximately 25%) and the balance between Bacteroidetes and Firmicutes phyla can affect obesity and intestinal inflammation. It has been reported that relative proportion of Bacteroidetes is decreased in obese people by comparison with Firmicutes. On the other hand, the results (Table 14) reveal that C. leptum population decreased significantly with more than one log in the presence of RG I and with 0.6-0.7 log with oligo-RG I. The low affinity of Clostridium spp. for rhamnogalacturono-oligosaccharides has also been reported by Van Laere et al. (2000). However, these authors have found that Clostridium spp. were able to utilize galacto- oligosaccharides as a carbohydrate source in a single culture fermentation. The decrease in the number of Clostridium population in our study may be the result of growth and production of secondary metabolite by other species. Gibson & Wang (1994) who have studied the fermentation of oligofructose by Bifidobacterium infantis, E. coli, and C. perfringens reported that the growth of Bifidobacterium inhibited that of Escherichia coli and Clostridium perfringens. Decrease in number of Clostridium while maintaining the number of Bacteroides/Prevottela/Porphiromonas can be advantageous as it has been shown that decrease in Clostridium and increase in Bacteroides number is associated with weight loss in obese adolescents.

Despite differences in the DP between Gal- and Dep-based oligo-RG I , their effects on the selective bacterial growth were more or less similar with less than 0.2 log unit difference. The results also show that there is no significant difference (<0.5 log) in the bacterial counts between FOSs and oligo-RG I. In contrast, others have reported that galactose-containing oligosaccharides, namely lactulose, galacto-oligosaccharides and soybean oligosaccharides, increased the number of Bifidobacterium more than fructose-containing inulin and FOS. Moreover, the investigation of the fermentability of selected oligosaccharides obtained from apple pomace by faecal microbiota indicated that gluco-oligosaccharide, galacto- oligosacchairdes and xylo-oligosaccharide were first fermented followed by arabino- oligosaccharides and oligogalacturonides.

Selective fermentation of potato derived polysaccharides and oligosaccharides by Bifidobacterium and Lactobacillus have been reported previously. For instance, others have reported that hydrolyzed galactan (<3, 3-10 and 10-100 kDa) and potato pulp polysaccharides (<10 and >10 kDa) promoted the growth of probiotic strains of Bifidobacterium longum and Lactobacillus acidophilus and generally did not support the propagation of Clostridium perfringens in single culture fermentations. Some also reported that stimulation of Bifidobacterium longum growth in the presence of hydrolyzed galactan was significantly higher than with FOS, galactose and unhydrolyzed galactan. Moreover, Some have obtained an increase in the numbers of Bifidobacterium spp. and Lactobacillus spp, upon the fermentation of high-MW polysaccharides (up to 400 kDa) from potato pulp by faecal microbiota. Others have reported that the increase in number of the Bifidobacterium spp. was higher when RG l-rich and HG-rich fractions were used as carbohydrate sources than FOS (DP 2-8 Orafti®P95), and RG l-rich fraction had better bifidogenic property than HG-rich one.

Table 14 also shows PI of investigated carbohydrates. Gal- and Dep-based oligo-RG I and FOS exhibited more or less similar PI of 0.46-0.5, while RG I resulted in lower PI of 0.2. Some have reported that orange peel pectic oligosaccharides, characterized by a monosaccharide profile of Glc, Ara, Gal, GalA, Xyl and Rha of 48.1 , 31 .2, 9.6, 6.3, 2.4 and 2.1 %, respectively, resulted in a lower PI than FOS. In contrast, others have shown that bergamot oligosaccharides obtained from alcohol insoluble residue, with a monosaccharide profile of GalA, Ara, Glc, Gal, Man, Rha and Xyl of 48.2, 10.5, 5.8, 5.7, 2.1 , 1 .9 and 1 .3%, respectively, exhibited a greater PI value than FOS. Mandalari 2007 considered number of eubacteria in their formula but, in this statement this does not matter because we compare PI of FOS and generated oligosaccharides from the same study and not to other study.

Although RG I generally resulted in lower prebiotic effects, it has been reported that high-MW carbohydrates are fermented more slowly and hence can reach the distal region of colon. Therefore, mixing RG I and oligo-RG I may provide the combination of both advantages of selective fermentation by lactic acid bacteria and the extended effect of prebiotic activity through proximal and distal regions of colon.

Measurement of SCFA. The major end products of prebiotic metabolism are SCFAs (acetate, formate, propionate and butyrate), the gases (hydrogen, carbon dioxide), and bacterial cell mass. Increased amount of total SCFA is desirable since low pH and high amount of SCFA suppress pathogenic bacteria. More specifically, propionate and formate were reported to reduce the activity of E. coll and Salmonella at pH 5. Table 15 summarizes the profile of SCFAs produced during fermentation of RG I, oligo-RG I and FOSs as control. The concentration of the SCFAs was measured at the end of the four day fermentation period for each carbohydrate source.

Table 15. Short chain fatty acids (SCFAs) released at the end of the fermentation of selected carbohydrates.

Dep-based oligo-RG I Gal-based oligo-RG

SCFA (mM) RG l^a b FOS^d 25% CH^e l^c

Acetate 79.8 104.3 100.7 69.2 68.1

Propionate 46.8 36.0 30.4 35.8 27.4

Butyrate 31 .3 26.5 30.6 51 .9 16.9

Valerate 6.5 6.3 6.1 6.7 10.8

Total SCFA 164.4 173.1 167.8 163.6 123.2 ^a Extracted polysaccharides from potato pulp using microwave-assisted alkaline treatment. ^b Generated oligosaccharide from potato RG I upon treatment with Depol 670L.

c Generated oligosaccharides from potato RG I upon treatment with endo-p-1 ,4-galactanase. ^d Fructo-oligosaccharide.

e Fermentation with 25% of carbohydrate source in the medium as a negative control.

The results show that the concentration of total SCFAs produced at the end of Oligo-RG I fermentations (167.8-173.1 mM) was slightly higher than those produced in the presence of RG I and FOSs (163.6-164.4 mM). Similarly, it was reported reported that after 24 h fermentation, pectic oligosaccharides from orange peel (Glc 48.12, Ara 31.19, Gal 9.59, GalA 6.29%) resulted in higher total SCFA as compared to FOS and orange albedo. Moreover, it was also reported that parent polysaccharide (apple pectin) resulted in less total SCFA as compared generated oligomers (5000 Da) and parent pectin produced less acetate and more butyrate than pectic oligosaccharides, which is similar to results obtained in the current study.

For all tested carbohydrates, acetate production was the dominant one as compared to other SCFAs. However, the fermentation of oligo-RG I (Gal- and Dep-based ones) induced more production of acetate (100.7-104.4 mM) than unhydrolyzed RG I (79.8 mM) and FOSs (69.2 mM). Other studies on the prebiotic properties of pectic polysaccharides, their oligosaccharides and FOS have also reported the production of acetate as a dominant SCFA. For instance, the fermentation of oligomeric saccharides derived from apple pectin have resulted in the production of mainly acetate and butyrate. Similarly, it was shown that FOS and galacto-oligosaccharides induced the acetate and butyrate formation after 12 h of fermentation.

The high acetate production upon all fermentations is consistent with Bifidobacterium and Lactobacillus population increases. Acetate can be formed by many anaerobic bacteria from the human gut; however, acetate is typically generated via bifidus pathway, and more specifically it is a major end product of the fermentation by Bifidobacterium. This can explain the higher production of acetate upon the fermentation of oligo-RG I as compared to the parent RG I polysaccharides stimulating less increase in the Bifidobacterium and Lactobacillus populations. Although the fermentation of FOSs supported the growth of Bifidobacterium and Lactobacillus populations with the same extents as oligo-RG I, it induced less production of acetate. These results may indicate the higher activity of Bifidobacterium in the presence of the generated oligo-RG I. Others have studied the effect of Bifidobacterium on promoting host defense against infection and showed that increased production of acetate can protect mice from enterohaemorrhagic E. coli 0157:1-17 and improve defense function of epithelial cell in intestine. In addition, acetate and propionate have shown to have anti- cholesterolemic effect. Acetate may also be used by microbial populations to produce propionate or butyrate through metabolic cross-feeding between Bifidobacterium and butyrate-forming bacteria.

The results (Table 15) also show that the amount of valerate was small and similar for both FOS and RG I derived samples. Oligo-RG I maintained more or less the production of similar concentrations of butyrate (26.5-30.6 mM) and propionate (30.4-36.0 mM); while the fermentation of FOSs as a control led to higher production of butyrate (51.9 mM) followed by propionate (35.8 mM). On the other hand, the fermentation of unhydrolysed RG I resulted in higher concentration of propionate (46.8 mM) as compared to butyrate (31.3 mM). Higher concentration of propionate resulted by fermentation of RG I is consistent with the larger increase in the number of Bacteroides/Prevottela/Porphiromonas upon fermentation of RG I since Bacteroides/Prevotella are known for their ability to produce propionate. The butyrigenicity of FOSs has been previously reported. Butyrate is the preferred energy source for colonocytes and plays a role in maintaining normal colonocyte population, participating in energy metabolism and protection against colorectal disease and cancer have found that FOS resulted in higher acetate and lower propionate production as compared to arabinogalactan. On the other hand, those who have investigated the fermentation properties of arabino-oligosaccharides with different DP, have reported similar trends as RG I derived samples in our study with acetate being the predominant SCFA followed by propionate and butyrate.

The prebiotic properties of RG I, Dep and Gal-based oligo-RG I were evaluated. The in vitro digestibility study revealed that 81.6 and 79.3% of the initial amount of RG I polysaccharides and their corresponding oligomers remained unhydrolyzed, respectively. Both potato RG I polysaccharides and their corresponding oligosaccharides/oligomers were fermented by Bifidobacterium and Lactobacillus; however, the oligosaccharides and the oligomers exhibited higher selective fermentability and resulted in stronger prebiotic effect. On the other hand, the addition of potato RG I polysaccharides or its corresponding oligosaccharides/oligomers did not have a negative effect on Bacteroidetes and seems to reduce the population of the Clostridium leptum group. This could be of great interest since a higher ratio Firmicute/Bacteroides was found in obese people compared with lean people.

Despite some differences in sugar composition and DP between Dep and Gal-based oligo- RG I, their effects on the bacterial flora were similar. Furthermore, the difference in bacterial count change between FOS and RG l-generated oligosaccharides was not significant. In addition, Dep and Gal-based oligo-RG I and FOS had similar PI; while RG I resulted in lower PI. Total amounts of SCFAs generated upon the fermentation of Dep and Gal-based oligo- RG I were higher than those obtained with RG I and FOS. This study shows that the RG I polysaccharides and their corresponding oligomers extract from potato had a good prebiotic effect.

REFERENCES Ahlroos, T., & Tynkkynen, S. (2009). Quantitative strain-specific detection of Lactobacillus rhamnosus GG in human faecal samples by real-time PCR. Journal of applied microbiology, 106(2), 506-514.

Blumenkrantz, N.; Asboe-Hansen, G. New method for quantitative determination of uronic acids. Analytical Biochemistry 1973, 54(2), 484-489.

Cinquin, C, Le Blay, G., Fliss, I. , & Lacroix, C. (2004). Immobilization of infant fecal microbiota and utilization in an in vitro colonic fermentation model. Microbial ecology, 48(1), 128-138.

Fujimoto, J., Tanigawa, K., Kudo, Y., Makino, H., & Watanabe, K. (201 1). Identification and quantification of viable Bifidobacterium breve strain Yakult in human faeces by using strain-specific primers and propidium monoazide. Journal of applied microbiology, 1 10(1), 209-217.

Gibson, G. R., & Wang, X. (1994). Enrichment of bifidobacteria from human gut contents by oligofructose using continuous culture. FEMS Microbiology Letters, 1 18(1 -2), 121 -127. Karboune, S.; L'Hocine, L.; Anthoni, J.; Geraert, P. A.; Kermasha, S. Properties of selected hemicellulases of a multi-enzymatic system from Penicillium funiculosum. Bioscience, Biotechnology, and Biochemistry 2009, 73(6), 1286-1292.

Kelly GS. Larch arabinogalactan: clinical relevance of a novel immune-enhancing polysaccharide. Altern Med Rev. 1999 Apr;4(2):96-103. Khodaei, N., & Karboune, S. (2013). Extraction and structural characterisation of rhamnogalacturonan l-type pectic polysaccharides from potato cell wall. Food Chemistry, 139(1), 617-623.

Khodaei, N., & Karboune, S. (2014). Enzymatic extraction of galactan-rich rhamnogalacturonan I from potato cell wall by-product. LWT-Food Science and Technology, 57(1), 207-216.

Khodaei, N.; Karboune, S.; Orsat, V. (2015d) Microwave-alkaline extraction of galactan-rich rhamnogalacturonan I from potato cell wall by-product. Food Chemistry 190, 495-505. Macfarlane, G., Macfarlane, S., & Gibson, G. (1997). Validation of a three-stage compound continuous culture system for investigating the effect of retention time on the ecology and metabolism of bacteria in the human colon. Microbial ecology, 35(2), 180-187.

Manderson, K., Pinart, M., Tuohy, K., Grace, W., Hotchkiss, A., Widmer, W., Rastall, R. (2005). In vitro determination of prebiotic properties of oligosaccharides derived from an orange juice manufacturing by-product stream. Applied and environmental microbiology, 71 (12), 8383-8389.

Minekus, M. , Marteau, P., & Havenaar, R. (1995). Multicompartmental dynamic computer- controlled model simulating the stomach and small intestine. Alternatives to laboratory animals: ATLA.

Nangia-Makker P, Hogan V, Honjo Y, Baccarini S, Tait L, Bresalier R, Raz A. Inhibition of human cancer cell growth and metastasis in nude mice by oral intake of modified citrus pectin. J Natl Cancer Inst. 2002 Dec 18;94(24): 1854-62.

0Bro J, Harholt J, Scheller HV, Orfila C. Rhamnogalacturonan I in Solanum tuberosum tubers contains complex arabinogalactan structures. Phytochemistry. 2004 May;65(10): 1429- 38.

Olesen M, Gudmand-Hoyer E, Norsker M, Kofod L, Adler-Nissen J. Fermentability of an enzymatically modified solubilised potato polysaccharide (SPP). Eur J Clin Nutr. 1998 Feb;52(2): 1 10-4. Oomen RJ, Doeswijk-Voragen CH, Bush MS, Vincken JP, Borkhardt B, van den Broek LA, Corsar J, Ulvskov P, Voragen AG, McCann MC, Visser RG. In muro fragmentation of the rhamnogalacturonan I backbone in potato (Solanum tuberosum L.) results in a reduction and altered location of the galactan and arabinan side-chains and abnormal periderm development. Plant J. 2002 May;30(4):403-13. Thomassen LV, Larsen DM, Mikkelsen JD, Meyer AS. Definition and characterization of enzymes for maximal biocatalytic solubilization of prebiotic polysaccharides from potato pulp. Enzyme Microb Technol. 201 1 Aug 10;49(3):289-97

Van Laere KM, Hartemink R, Bosveld M, Schols HA, Voragen AG. Fermentation of plant cell wall derived polysaccharides and their corresponding oligosaccharides by intestinal bacteria. J Agric Food Chem. 2000 May;48(5): 1644-52.

Claims

While the invention has been described in connection with specific embodiments thereof, it will be understood that it is capable of further modifications and this application is intended to cover any variations, uses, or adaptations of the invention following, including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains and as may be applied to the essential features hereinbefore set forth, and as follows in the scope of the appended claims. WHAT IS CLAIMED IS:

1 . A process for making a non-digestible mixture of oligosaccharides and oligomers from a potato pulp, said process comprising: (a) providing a source of potato-derived rhamnogalacturonan I (RG I), wherein the source of potato-derived RG I is the potato pulp or a RG l-enriched fraction obtained from the potato pulp;

(b) hydrolyzing the source of potato-derived RG I with a multi-enzymatic mixture to yield an hydrolyzed mixture; and (c) isolating the non-digestible mixture from the hydrolyzed mixture, wherein the non-digestible mixture comprises:

- oligosaccharides having a degree of polymerization between 2 and 12; and

- oligomers having a degree of polymerization between 13 and 70.

2. The process of claim 1 , wherein the potato pulp is substantially free of starch or proteins.

3. The process of claim 2, wherein said potato pulp consists essentially of potato cell wall materials.

4. The process of any one of claims 1 to 3, further comprising, prior to step (a), treating the potato pulp with an endo-polygalacturonase to obtain the RG l-enriched fraction.

5. The process of claim 4, wherein the endo-polygalacturonase is from Aspergillus niger.

6. The process of any one of claims 1 to 3, further comprising, prior to step (a), submitting the potato pulp to a microwave-assisted alkaline extraction to obtain the RG l-enriched fraction.

7. The process of claim 6, wherein the microwave-assisted alkaline extraction comprises suspending the potato pulp in an alkaline solution to obtain an alkaline potato pulp mixture and submitting the alkaline potato pulp mixture to a microwave treatment.

8. The process of claim 7, wherein the alkaline solution is a solution of KOH.

9. The process of claim 7 or 8, wherein the power of the microwave treatment is between 20 and 220 W/g of the alkaline potato pulp mixture.

10. The process of any one of claims 1 to 9, wherein step (c) further comprises precipitating or filtering the hydrolyzed mixture to isolate the non-digestible mixture.

1 1 . The process of claim 10, wherein said filtering is ultrafiltration.

12. The process of any one of claims 1 to 1 1 , wherein the multi-enzymatic mixture of step (b) has arabinanase enzymatic activity and galactanase enzymatic activity.

13. The process of claim 12, wherein the multi-enzymatic mixture comprises at least one of Gamanase and Depol.

14. The process of claim 12, wherein the multi-enzymatic mixture comprises Gamanase and Depol.

15. The process of claim 14, wherein the ratio (U/U) of Gamanase and Depol enzymatic activity is about 1 : 1.

16. A non-digestible mixture of oligosaccharides and oligomers obtained by the process of any one of claims 1 to 15.

17. A prebiotic composition comprising the non-digestible mixture of claim 16 and a carrier.

18. A food additive comprising the non-digestible mixture of claim 16 and a carrier.