Oligosaccharides Released from Milk Glycoproteins Are SelectiveGrowth Substrates for Infant-Associated Bifidobacteria

Sercan Karav,a Annabelle Le Parc,a Juliana Maria Leite Nobrega de Moura Bell,a Steven A. Frese,a,b* Nina Kirmiz,a David E. Block,c,d

Daniela Barile,a,b David A. Millsa,b,c

Department of Food Science and Technology, University of California, Davis, California, USAa; Foods for Health Institute, University of California, Davis, California, USAb;Department of Viticulture & Enology, University of California, Davis, California, USAc; Department of Chemical Engineering and Materials Science, University of California,Davis, California, USAd

ABSTRACT

Milk, in addition to nourishing the neonate, provides a range of complex glycans whose construction ensures a specific enrich-ment of key members of the gut microbiota in the nursing infant, a consortium known as the milk-oriented microbiome. Milkglycoproteins are thought to function similarly, as specific growth substrates for bifidobacteria common to the breast-fed infantgut. Recently, a cell wall-associated endo-�-N-acetylglucosaminidase (EndoBI-1) found in various infant-borne bifidobacteriawas shown to remove a range of intact N-linked glycans. We hypothesized that these released oligosaccharide structures canserve as a sole source for the selective growth of bifidobacteria. We demonstrated that EndoBI-1 released N-glycans from concen-trated bovine colostrum at the pilot scale. EndoBI-1-released N-glycans supported the rapid growth of Bifidobacterium longumsubsp. infantis (B. infantis), a species that grows well on human milk oligosaccharides, but did not support growth of Bifidobac-terium animalis subsp. lactis (B. lactis), a species which does not. Conversely, B. infantis ATCC 15697 did not grow on the degly-cosylated milk protein fraction, clearly demonstrating that the glycan portion of milk glycoproteins provided the key substratefor growth. Mass spectrometry-based profiling revealed that B. infantis consumed 73% of neutral and 92% of sialylated N-gly-cans, while B. lactis degraded only 11% of neutral and virtually no (<1%) sialylated N-glycans. These results provide mechanis-tic support that N-linked glycoproteins from milk serve as selective substrates for the enrichment of infant-associated bifidobac-teria capable of carrying out the initial deglycosylation. Moreover, released N-glycans were better growth substrates than theintact milk glycoproteins, suggesting that EndoBI-1 cleavage is a key initial step in consumption of glycoproteins. Finally, thevariety of N-glycans released from bovine milk glycoproteins suggests that they may serve as novel prebiotic substrates with se-lective properties similar to those of human milk oligosaccharides.

IMPORTANCE

It has been previously shown that glycoproteins serve as growth substrates for bifidobacteria. However, which part of a glyco-protein (glycans or polypeptides) is responsible for this function was not known. In this study, we used a novel enzyme to cleaveconjugated N-glycans from milk glycoproteins and tested their consumption by various bifidobacteria. The results showed thatthe glycans selectively stimulated the growth of B. infantis, which is a key infant gut microbe. The selectivity of consumption ofindividual N-glycans was determined using advanced mass spectrometry (nano-liquid chromatography chip– quadrupole timeof flight mass spectrometry [nano-LC-Chip-Q-TOF MS]) to reveal that B. infantis can consume the range of glycan structuresreleased from whey protein concentrate.

Protein glycosylation is a common modification that adds amajor dimension of complexity to protein structure and has

been linked to significant roles in protein functions, such as pro-tein folding, biological recognition, and enzymatic protection (1,2). In eukaryotes, N-glycosylation and O-glycosylation are the twomajor types of glycosylation, whereby N-linked glycans are linkedto a polypeptide at a specific asparagine, while O-linked glycansoccur at a serine/threonine residue. N-Linked glycoproteins areclassified into the following three groups based on composition:high-mannose, complex, and hybrid (combination of high-man-nose and complex types) glycoproteins (3). The nearly infinitecombinations of linkages, compositions, and structures providedby complex glycan synthesis can create a similarly diverse array ofpossible glycoprotein structures (4).

While the structure of glycoproteins may be critical for theirfunction, for a variety of purposes, certain bacteria have the abilityto deglycosylate and degrade glycoproteins (5–9). This activity hasbeen described primarily for pathogens deglycosylating various

host defense glycoproteins, including human IgG (10), RNase B(7), and lactoferrin (11). It has also been shown that released gly-cans from N-linked glycoproteins serve as a carbon source forpathogen growth. Endo D, an endo-�-N-acetylglucosaminidasefrom Streptococcus pneumoniae, acts to release asparagine-linked

Received 18 February 2016 Accepted 26 March 2016

Accepted manuscript posted online 15 April 2016

Citation Karav S, Le Parc A, Leite Nobrega de Moura Bell JM, Frese SA, Kirmiz N,Block DE, Barile D, Mills DA. 2016. Oligosaccharides released from milkglycoproteins are selective growth substrates for infant-associated bifidobacteria.Appl Environ Microbiol 82:3622–3630. doi:10.1128/AEM.00547-16.

Editor: C. M. Dozois, INRS—Institut Armand-Frappier

Address correspondence to David A. Mills, damills@ucdavis.edu.

* Present address: Steven A. Frese, Evolve Biosystems, Inc., Davis, California, USA.

S.K. and A.L.P. contributed equally to this article.

Copyright © 2016, American Society for Microbiology. All Rights Reserved.

crossmark

3622 aem.asm.org June 2016 Volume 82 Number 12Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

oligosaccharides (6); however, additional degradation of the re-leased oligosaccharides by exoglycosidases is required for subse-quent growth (5, 9, 12, 13). Commensal Bacteroides species canrelease and consume the O-linked glycans in mucin, providing afitness advantage in vivo (14, 15). These Bacteroides species alsoconsume the structurally similar oligosaccharides found in milk(16). Curiously, there is little information regarding whether thereported bifidogenic effect of milk oligosaccharides is restricted tofree oligosaccharides or if the structurally similar glycans foundconjugated to proteins and lipids may serve a similar function(17–19).

Milk from many species contains an abundant amount of com-plex oligosaccharides (20–22). These can be found free, such ashuman milk oligosaccharides (HMOs), or as conjugated glycans,such as those found attached to proteins (glycoproteins) or lipids(glycolipids) (22). HMOs possess many diverse structures (23),the result of an array of configurations of the component mono-saccharides: glucose, galactose, N-acetylglucosamine, fucose, andN-acetylneuraminic acid. HMOs are clearly a factor that influ-ences populations of beneficial microorganisms, such as Bifido-bacterium, in the infant gut (22, 24). Several Bifidobacterium spe-cies are known to grow to high cell densities during in vitrogrowth on HMOs as the sole carbon source (25–28), whichmirrors observations of breast-fed infants consuming humanmilk supplemented with Bifidobacterium longum subsp. infantis(B. infantis) (29). Of these species of Bifidobacterium, only B.infantis possesses a large cassette of genes associated with HMOconsumption. The possession of this cluster of genes uniquelyenables this subspecies to grow to high cell densities on a broadarray of HMOs as sole carbon sources, while many other bifido-bacteria do not have this ability (30) or are able to consume onlystructures with limited diversity (31). In addition, several differentstrategies are employed by groups of Bifidobacterium to utilizeHMOs. B. infantis ATCC 15697 is believed to import HMOs intactand subsequently to employ intracellular glycoside hydrolases,such as �-galactosidases, N-acetyl-�-D-hexosaminidases, �-fuco-sidases, and �-sialidases, to deconstruct HMOs (31–35). In con-trast, other species of Bifidobacterium (e.g., B. bifidum) employextracellular glycosidases to sequentially degrade these structuresand to transport mono-, di-, or oligosaccharides rather than thecomplex, compositionally diverse, and often branched structureswhich comprise HMOs (31).

Similar to the diverse structures of HMOs, many differenttypes of conjugated glycans are found in human milk. WhileHMOs have long been known to serve as prebiotic substrates forBifidobacterium, it is unclear whether glycoproteins, owing totheir structural similarity to HMOs, perform a similar function.Some glycoconjugates, such as lactoferrin and the �-casein-de-rived glycomacropeptide, have previously been associated with abifidogenic effect (8, 17, 18, 36, 37), and strains of Bifidobacteriumhave even been shown to grow on a structurally distant yeast man-noprotein as the sole carbon source (8). In the case of glycoproteinenrichment of bifidobacteria, it remained unclear to what extentthe individual glycan and protein components are responsible forthe enrichment (24). The ability to deglycosylate or degrade gly-coproteins has been demonstrated for select species of Bifidobac-terium (8, 38, 39). We recently showed that B. infantis ATCC15697 contains an endoglycosidase, EndoBI-1 (glycosyl hydrolasefamily 18), that has activity on major types of N-linked glycans

found in glycoproteins and can release glycans from human lac-toferrin and immunoglobulins (8).

Bovine milk glycans have similarities to human milk glycans(40–43). We recently determined the optimal conditions and ki-netic parameters of N-glycan release from concentrated bovinewhey for large-scale N-glycan production using EndoBI-1 (13, 44,45). In this work, EndoBI-1 treatment of milk glycoproteins dem-onstrated that freed glycans, as opposed to deglycosylated milkprotein, are a good substrate for rapid and selective growth, sim-ilar to HMOs. Moreover, the selectivity of consumption of indi-vidual N-glycans was determined using advanced mass spectrom-etry (nano-liquid chromatography chip– quadrupole time offlight mass spectrometry [nano-LC-Chip-Q-TOF MS]) to revealthat B. infantis can consume the range of glycan structures re-leased from whey protein concentrate.

MATERIALS AND METHODSBacteria and media. Bifidobacterium longum subsp. infantis (B. infantis)ATCC 15697 and Bifidobacterium animalis subsp. lactis (B. lactis)UCD316 cultures were routinely propagated in deMan-Rogosa-Sharpe(MRS) broth (Becton Dickinson, Franklin Lakes, NJ) supplemented with0.05% (wt/vol) L-cysteine at 37°C under anaerobic conditions. Escherichiacoli strains Top10 (GeneTarget Inc., San Diego, CA) and BL21* (Invitro-gen, Carlsbad, CA), used for gene cloning and expression, were propa-gated in Luria broth (LB) under selective conditions (100 �g/ml carben-icillin).

Gene cloning and protein expression and purification. Gene ampli-fication and protein expression in E. coli BL21* were performed as de-scribed by Sela et al. (31). A pEco-T7-cHis Eco cloning kit (GenTargetInc., San Diego, CA) was used for gene cloning. Blon_2468 (EndoBI-1)was PCR amplified from B. infantis ATCC 15697 genomic DNA by use ofa high-fidelity polymerase with the following primers: 5=-TTTGTACAAAAAAGCAGGCACCATGAATGCGGACGCCGTTTCTCCGAC-3= and5=-TTTGTACAAGAAAGCTGGGTTGCCGGTCGCACTCAGTTGCTTCGG-3=. Transmembrane domains and the signal peptide were not am-plified to facilitate protein expression and purification, and a C-terminalpolyhistidine tag was added to facilitate purification. E. coli BL21* con-taining the vector was grown for 3 h at 37°C to reach an optical density at600 nm (OD600) of �0.6. Protein expression was induced by the additionof IPTG (isopropyl-�-D-thiogalactopyranoside) to a final concentrationof 0.5 mM, and cells were grown at 37°C for an additional 6 h. Bacteriawere collected by centrifugation at 4,000 rpm (model A-4-81 rotor; Ep-pendorf) for 20 min at 4°C. All subsequent steps for bacterial lysis wereachieved at 4°C. All solutions were supplemented with a protease inhibi-tor cocktail (Roche, San Francisco, CA). Cell pellets were incubated in 100ml of BugBuster (Novagen, Billerica, MA) for 10 min at 24°C. DNase I(200 �l) (Roche, San Francisco, CA) and 100 �l of lysozyme (100 mg/ml)were added to the sample, and the mixture was placed on ice for 30 min.Cell lysates were centrifuged at 13,000 rpm (model F45-24-11 rotor; Ep-pendorf) for 30 min to remove cell debris.

Protein was purified by affinity chromatography using 5-ml pre-packed Ni-charged columns (Bio-Rad, Hercules, CA). All chromato-graphic steps were performed using a model EP-1 Econo pump (Bio-Rad)and a model 2110 fraction collector (Bio-Rad) at a flow rate of 5 ml/min.The column was equilibrated with 25 ml of a solution containing 300 mMKCl–50 mM KH2PO4, and 5 mM imidazole buffer (pH 8). Fifty millilitersof sample was loaded into the column. The flowthrough was collected,and the column was washed successively with 30 ml of 300 mM KCl–50mM KH2PO4–5 mM imidazole buffer (pH 8) and then 20 ml of 300 mMKCl–50 mM KH2PO4–10 mM imidazole buffer (pH 8). The bound pro-tein was eluted with a stepwise gradient, using imidazole concentrationsranging from 100 to 300 mM. The purity of the EndoBI-1 fractions wasevaluated by SDS-PAGE. Purified protein was concentrated using a 15-ml, 30-kDa-cutoff centrifugal filter device (Amicon, Millipore, Billerica,

Selective Growth of Bifidobacteria on Milk N-Glycans

June 2016 Volume 82 Number 12 aem.asm.org 3623Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

MA), and buffer was exchanged for 1� saline sodium citrate (SSC; 1�SSC is 0.15 M NaCl plus 0.015 M sodium citrate), using Bio-Gel P-30 inSSC buffer columns (Bio-Rad). The protein concentration was deter-mined by using a Qubit protein assay kit (Life Technologies, Grand Island,NY), and the purified enzyme was stored at �80°C.

Gene expression analysis. Biological triplicate cultures of B. infantisATCC 15697 were grown to an OD600 of 0.6 (early mid-log phase) in MRSmedium containing lactose (2%) as the sole carbon source, harvested bycentrifugation, and washed in prewarmed (37°C) MRS containing nocarbohydrates. Cells were pelleted and resuspended in the original volumeof prewarmed medium containing lactose (2% [wt/vol]; positive control),treated whey protein (10% [wt/vol]), or untreated whey protein (10%[wt/vol]) as the sole carbon source. After 2 h of anaerobic incubation at37°C, 2-ml aliquots of culture were harvested by centrifugation and im-mediately resuspended in an equal volume of RNALater (Ambion/LifeTechnologies, Grand Island, NY) and frozen at �80°C for later use. RNAwas extracted using an RNAqueous kit according to the manufacturer’sinstructions, except that cell lysis was preceded by a 15-min incubationwith lysozyme (50 mg/ml) and mutanolysin (120 U) at 37°C and a 5-minincubation at 42°C with 10 �l of proteinase K (Qiagen, Valencia, CA).RNA was subjected to a DNase treatment using TURBOfree DNase (Am-bion/Life Technologies, Grand Island, NY) before reverse transcription-PCR (RT-PCR).

Two microliters of DNase-treated RNA was used with 20 �l ofSYBR master mix (Applied Biosystems/Life Technologies, Grand Is-land, NY) in a reaction mixture containing 0.1 �l of RNase inhibitor(Applied Biosystems/Life Technologies, Grand Island, NY), 10 pmolof each primer, and 0.1 �l of MultiScribe reverse transcriptase (Ap-plied Biosystems/Life Technologies, Grand Island, NY). Blon_2468expression was measured via primers Blon_2468F (5=-ACCGGCAAGATCTACACAGC-3=) and Blon_2468R (5=-GCACTCAGTTGCTTCGGTTG-3=) and compared to expression of a tRNA synthase gene byuse of primers Bl_0301F (5=-CAACCGCCGCGATCTTC-3=) andBl_0301R (5=-CCAGCTGTGAAAGCAACGTGTT-3=), previously de-scribed for B. longum and used in studies of B. infantis expression (46) butmodified to account for a single-base mismatch to B. infantis ATCC15697. RT-PCR was preceded by a synthesis step (48°C for 30 min and95°C for 10 min), followed by 40 cycles of 95°C for 15 s and 60°C for 60 s.A subsequent melting curve step was conducted to ensure amplificationproduct specificity. Normalized expression was calculated as previouslydescribed (31), with outliers removed by Grubb’s test, and the results werecompared to those for cells incubated on lactose by a ratio-paired t test,using GraphPad Prism 6.

Pilot-scale concentration of bovine milk glycoproteins. Protein con-centration from bovine colostrum whey was carried out using a pilot-scalecross-flow membrane system (model L; GEA Filtration, Hudson, WI).The system was composed of a 2.5-in.-diameter spiral membrane housing(1 to 2 m2), a 95-liter jacketed stainless steel feed tank, a Proline Promass80 E flowmeter (EndressHauser, Reinach, Switzerland), a heat ex-changer, and a 7.0-horsepower feed pump (model D10EKSGSNECF; Hy-dra-Cell, Minneapolis, MN). After upstream lactose hydrolysis (0.1%Aspergillus oryzae beta-galactosidase [EC 3.2.1.23], 30 min, 40 to 43°C), 74liters of bovine colostrum whey was ultrafiltered in a single batch with a10-kDa-cutoff polyethersulfone spiral-wound membrane (effective areaof 1.86 m2) to a concentration factor of 5.4 (concentration factor vol-ume of feed/volume of retentate). Whey protein concentration was per-formed at a constant temperature of 40 to 43°C, a transmembrane pres-sure of 300 kPa, and a recirculation flow rate of 10 liters/min. After aconcentration factor of 5.4 was achieved, the protein-rich retentate wasdiluted back to its original volume with water. Simple sugars (e.g., residuallactose and the free monosaccharides derived from its hydrolysis) lackselective prebiotic activity and can be a confounding factor in all func-tional studies, so they were removed by diafiltration. Additionally, in or-der to test only the prebiotic activity of the N-glycans newly released fromglycoproteins, we performed two discontinuous diafiltrations (by volume

reduction) to increase the removal of free oligosaccharides from the ul-trafiltration retentate.

N-Glycan release from milk proteins. One hundred milliliters of con-centrated bovine colostrum whey was incubated for 18 h at 37°C with 20mg of EndoBI-1 in 20 mM Na2HPO4, pH 5. Seven hundred milliliters ofpure cold ethanol was added to the mixture to precipitate proteins. Themixture was incubated for 2 h at �20°C. The mixture was centrifuged at4,000 rpm (model A-4-81 rotor; Eppendorf) for 10 min at 4°C. The re-sulting supernatant, containing the soluble released N-glycans, was driedin a rotary evaporator (Heidolph 36000130 Hei-Vap value collegiate ro-tary; Fisher Scientific, Waltham, MA). Samples were rehydrated in 100 �lof water, vortexed, and sonicated for glycan quantification and purifica-tion.

Glycan quantification. A microplate colorimetric carbohydrate assay(Biovision, Milpitas, CA) was used to quantify the purified glycans. Acommercial mannose standard (0, 2, 4, 6, 8, and 10 �l of a 2-mg/mlsolution) was used to create a standard curve. The volume of each samplewas adjusted to 30 �l per well with water. A sample and 150 �l of concen-trated sulfuric acid (98%) were added to each well. Samples were mixedon a shaker for �1 min and then incubated at 85°C for 15 min. Afterincubation, 30 �l of developer (provided by the manufacturer) was addedto each well. Samples were again mixed on the shaker for 5 min. TheOD490 of each sample was measured. The OD490 was applied to the man-nose standard curve linear function to calculate the quantity of carbohy-drate in the sample.

Bacterial fermentation of N-glycans. B. infantis ATCC 15697 and B.lactis UCD316 colonies were cultured in MRS broth and incubated over-night at 37°C in an anaerobic growth chamber (Coy Laboratory Prod-ucts). The resultant cultures were inoculated at 1% (vol/vol) into 100 �l ofreconstituted MRS broth supplemented with 2% N-glycans (wt/vol), 10%deglycosylated bovine whey (wt/vol), or 10% bovine whey (MRSN-glycan)as the sole carbohydrate and then overlaid with 25 �l of sterile mineral oilin a 96-well microtiter plate to prevent evaporation. Cell growth was mon-itored in real time by assessing the OD600 by use of a BioTek PowerWave340 plate reader (BioTek, Winooski, VT) every 30 min, preceded by 15 s ofshaking at variable speed. Three biological replicates were performed foreach sample. Once harvested, culture supernatants were centrifuged at3,000 � g for 15 min and filtered through a 0.22-�m-pore-size membrane(Millipore, Billerica, MA) prior to storage at �80°C. An inoculated sam-ple (control) with no carbon source was performed in parallel as a platecontrol.

Glycan purification for mass spectrometry. Recovered N-glycansfrom the bacterial supernatants and control samples were loaded on a C18

plate (Glygen, Columbia, MD). The plate was conditioned three timeswith 100 �l of 80% acetonitrile (ACN) containing 0.1% trifluoroaceticacid (TFA) in water, followed by three times with 100 �l of water. Sampleswere loaded, and N-glycans were eluted with 3 volumes of water. TheN-glycan solution was loaded on a PGC SPE (porous graphitic carbon,solid-phase extraction) plate (Glygen, Columbia, MD) previously opti-mized for the purification of free glycans. The plate was conditioned asdescribed previously (47). After sample loading, wells were washed sixtimes with 200 �l of water, and N-glycans were eluted thrice with 200 �l of40% ACN containing 0.1% TFA in water. The enriched N-glycan frac-tions were dried overnight by vacuum drying. Samples were rehydrated in50 �l of water, vortexed, sonicated, and diluted 50 times prior to massspectrometry analysis.

Nano-LC-Chip-Q-TOF MS. N-Glycans were analyzed using an Agi-lent 6520 accurate-mass Q-TOF LC-MS with a microfluidic nano-electro-spray chip (Agilent Technologies, Santa Clara, CA). N-Glycans were sep-arated using a high-pressure liquid chromatography (HPLC) chip with a40-nl enrichment column and a 43-mm by 75-�m analytical column,both packed with 5-�m porous graphitized carbon (PGC). The systemwas composed of a capillary and a nanoflow pump, and both used binarysolvents consisting of solvent A (3% [vol/vol] ACN, 0.1% [vol/vol] formicacid in water) and solvent B (90% [vol/vol] ACN, 0.1% [vol/vol] formic

Karav et al.

3624 aem.asm.org June 2016 Volume 82 Number 12Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

acid in water). Two microliters of sample was loaded with solvent A at acapillary pump flow rate of 4 �l/min. N-Glycan separation was performedwith a 65-min gradient delivered by the nanopump at a flow rate of 0.3�l/min. The 65-min gradient used the following program: 0% B (0.0 to 2.5min), 0 to 16% B (2.5 to 20.0 min), 16 to 44% B (20.0 to 30.0 min), 44 to100% B (30.0 to 35.0 min), and 100% B (35.0 to 45.0 min). The gradientwas followed by equilibration at 0% B (45.0 to 65.0 min). Data wereacquired within the mass range of 450 to 3,000 m/z for N-glycans in thepositive ionization mode, with an acquisition rate of 1 spectrum/s forN-glycans. An internal calibrator ion of 922.010 m/z from the tuningmix (ESI-TOF tuning mix G1969-85000; Agilent Technologies) wasused for continual mass calibration. For tandem MS analysis, N-gly-cans were fragmented with nitrogen as the collision gas. Spectra wereacquired within the mass range of 100 to 3,000 m/z. The collisionenergies corresponded to voltages (Vcollision) based on the followingequation: Vcollision m/z (1.5/100 Da) V � 3.6 V; the slope and offsetof the voltages were set at 1.5/100 Da and �3.6, respectively. Acquisi-tion was controlled by MassHunter workstation data acquisition soft-ware (Agilent Technologies).

N-Glycan identification. All compounds in the chromatogramswere identified with MassHunter qualitative analysis software (versionB.06.00 SP2; Agilent Technologies). Compounds were extracted usingthe molecular feature extractor algorithm. The software generated ex-tracted compound chromatograms in the range of 400 to 3,000 m/z,with an ion count cutoff of 1,000, allowed charge states of 1 to 3,retention times of 5 to 40 min, and a typical isotopic distribution ofsmall biological molecules. The resulting compounds were matched toa bovine milk N-glycan library (41), using a mass error tolerance of 20ppm. The N-glycans from the library were composed of hexose (Hex),HexNAc, fucose, NeuAc, and N-glycolylneuraminic acid (NeuGc).The assignment of N-glycans was confirmed by tandem mass spec-trometry.

Calculation of the relative abundances of N-glycans was performed byMassHunter Profinder software, using the batch targeted feature extrac-tion algorithm. A database was built to contain the molecular formulas,masses, and retention times of identified N-glycans from the control sam-ple. This library was used in combination with the batch targeted featureextraction algorithm. A minimum abundance of 1,000 counts was used tofilter out low-abundance compounds. Compounds were extracted by us-ing allowed charge states of 1 to 3, a mass error tolerance of 20 ppm,and a retention time tolerance of 1 min.

N-Glycan consumption was calculated with respect to that of the uni-noculated control by normalizing the summed abundance of neutral andsialylated N-glycans in the ion count for the bacterial supernatant to thatfor the control by using the following equation:

N-glycan consumption � �1 � ��i�1

n

API for bacterial sample

�i�1

n

API for control sample ��� 100

where API is absolute peak intensity and n is the number of identifiedN-glycans. The relative amount was expressed as a percentage of the totalconsumption.

RESULTSRelease and purification of N-glycans from concentrated wheyproteins. Release and purification of N-glycans were accom-plished using EndoBI-1 as described in Fig. 1. Briefly, EndoBI-1was incubated with concentrated bovine colostrum whey underconditions previously determined to be optimal for enzyme activ-ity (pH 5, overnight, 37°C) (44). Cleavage was performed undernondenaturing conditions because the goal of this study was toproduce the N-glycans in a native state and to preserve biolog-ical function. Using this method, a total of 80 mg of N-glycanswas released from 40 ml of concentrated bovine colostrum

whey. Mass spectrometry-based characterization of the re-leased glycans identified 18 species, including 6 neutral and 12sialylated moieties (Table 1). Interestingly, four of the compo-sitions contained fucosylated oligosaccharides, and their detec-tion extends the previous identification of fucosylated oligo-saccharides within the bovine milk glycome to N-linkedglycoproteins (48).

Growth of bifidobacteria on whey, deglycosylated whey pro-teins, and N-glycans released from whey. Previous work showeda bifidogenic effect upon ingestion of milk glycoproteins (17, 18,49). Since milk glycoprotein-derived peptides have been impli-cated in bifidogenic responses (50), it was unclear if this effect is aresult of the catabolism of the protein itself or the associated gly-can (or both) of these glycoconjugates. To test this, we examinedthe growth of B. infantis ATCC 15697 and B. lactis UCD316 on10% whey glycoprotein concentrate, in which the total glycanconcentration matches that used in a previous study (i.e., 2%)(13). B. infantis ATCC 15697 grows well on HMOs (25, 26), whileB. lactis UCD316 does not (29). B. infantis readily grew to a highcell density on released N-glycans, while no growth occurred on10% deglycosylated whey protein concentrate (Fig. 2). Intactwhey protein concentrate supported moderate growth of B. infan-tis; however, none of the substrates tested supported vigorousgrowth of B. lactis UCD316.

To determine if EndoBI-1 expression is related to substrateavailability, given the growth of B. infantis on whey protein con-centrate, we measured the relative expression of Blon_2468, thegene encoding the EndoBI-1 endoglycosidase (8), during growthon this substrate. Blon_2468 was significantly upregulated duringincubation on untreated whey but not on whey that had beendeglycosylated by pretreatment with EndoBI-1 prior to incuba-tion, and not in the presence of lactose (Fig. 3), suggesting that theBlon_2468 induction signaling mechanism is based on the glycanportion of the glycoprotein. Given that free milk oligosaccharidesdo not induce Blon_2468 expression (51), this suggests that B.

FIG 1 Methods involved in release and purification of N-glycans by use ofpurified EndoBI-1. EndoBI-1 was cloned and expressed in E. coli. PurifiedEndoBI-1 was incubated with bovine colostrum whey to release N-glycans.

Selective Growth of Bifidobacteria on Milk N-Glycans

June 2016 Volume 82 Number 12 aem.asm.org 3625Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

infantis senses intact glycoprotein and induces Blon_2468 in re-sponse. The specific mechanism for this signal remains to be de-termined.

Glycoprofiling of N-glycan consumption by B. infantis andB. lactis. To determine the potential preferential consumption ofspecific N-glycan compositions, mass spectrometry-based glyco-profiling was performed on culture supernatants. Supernatantswere collected after 96 h of growth, and residual N-glycans werepurified by solid-phase extraction and profiled by nano-HPLC-chip-TOF MS. The separation of N-glycans was performed on amicrochip packed with porous graphitized carbon, which enabledisomer-level separation with reproducible retention times. Tan-dem MS analysis generated specific fragment ions common to allN-glycans and allowed confirmation of N-glycan compositions,including those of both neutral and sialylated N-glycans. For thisstudy, we specifically profiled the bifidobacterial consumption of6 neutral and 12 sialylated N-glycans (Table 1). Figure 4 shows theextracted compound chromatograms for N-glycans obtainedfrom the B. infantis and B. lactis growth supernatants as well as anuninoculated control. This analysis revealed the presence of 18,16, and 6 N-glycans remaining in the uninoculated, B. lactis, andB. infantis samples, respectively. The majority of the N-glycans inthe uninoculated control were also present in the B. lactis sample,with similar peak areas, suggesting a minimal consumption ofN-glycans by B. lactis. Conversely, B. infantis consumed the ma-jority of the N-glycans present.

The data revealed that B. infantis consumed 73% of neutral and92% of sialylated N-glycans, whereas B. lactis degraded 11% ofneutral and only 0.9% of sialylated N-glycans. The majority ofreleased N-glycans were consumed by B. infantis, except for theneutral N-glycan 3Hex-3HexNAc (Fig. 5). Notably, three N-glycans (3Hex-5HexNAc, 4Hex-3HexNAc-1Fuc, and 5Hex-3HexNAc-2NeuGc) could not be detected after incubationwith B. lactis.

FIG 2 Representative growth curves for B. infantis (A) and B. lactis (used as anegative control) (B) on released N-glycans ( ), whey ( ), or deglycosy-lated whey ( ). The growth responses of each sample were measured by deter-mining the OD600. Growth curves were performed in biological triplicates.

TABLE 1 Details of released N-glycansa

Mass

Composition

Glycan type

Presence/absence of released N-glycan

Hexose HexNAc Fucose NeuAc NeuGc Control B. lactis UCD316 B. infantis ATCC 15697

1,113.408 3 3 0 0 0 Neutral N-glycan 1,275.460 4 3 0 0 0 Neutral N-glycan 1,421.547 4 3 1 0 0 Neutral N-glycan �1,437.525 5 3 0 0 0 Neutral N-glycan 1,519.573 3 5 0 0 0 Neutral N-glycan � �1,665.632 3 5 1 0 0 Neutral N-glycan �1,404.509 3 3 0 1 0 Sialylated N-glycan 1,525.543 5 2 0 1 0 Sialylated N-glycan �1,566.563 4 3 0 1 0 Sialylated N-glycan 1,728.627 5 3 0 1 0 Sialylated N-glycan 1,744.620 5 3 0 0 1 Sialylated N-glycan �1,769.668 4 4 0 1 0 Sialylated N-glycan �1,810.660 3 5 0 1 0 Sialylated N-glycan �1,874.687 5 3 1 1 0 Sialylated N-glycan �1,890.686 5 3 1 0 1 Sialylated N-glycan �1,931.713 5 4 0 1 0 Sialylated N-glycan �2,035.737 5 3 0 1 1 Sialylated N-glycan � �2,051.725 5 3 0 0 2 Sialylated N-glycan � �a Neutral mass, monosaccharide composition, glycan type, and whether or not the composition remained in the supernatants of the control, B. lactis UCD316, and B. infantis ATCC15697 samples. HexNAc, N-acetylglucosamine; NeuAc, N-acetylneuraminic acid; NeuGc, N-glycolylneuraminic acid; , presence of the released glycan; �, absence of the releasedglycan.

Karav et al.

3626 aem.asm.org June 2016 Volume 82 Number 12Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

DISCUSSION

Previous work identified EndoBI-1 as a possible mechanism bywhich B. infantis utilizes glycan moieties from N-linked glycopro-teins in milk as growth substrates (52); however, it remained un-clear if the released oligosaccharides actually serve as a bettergrowth substrate than the glycoconjugate. To examine this, puri-fied EndoBI-1 was used to release a range of N-glycans from wheyglycoproteins at the laboratory scale. B. infantis clearly showed an

ability to consume these released N-glycans as the sole carbonsource, growing to a high cell density mimicking the stronggrowth of B. infantis on complex HMOs. In contrast, B. lactis didnot readily consume released N-glycans as the sole carbonsource, which is not surprising considering that the strain isunable to consume structurally analogous HMOs and lacks ahomolog to EndoBI-1 (52) or other genes associated withHMO catabolism. Furthermore, released N-glycans enabledmore rapid growth than that seen with the conjugated form,reflecting both a preference for free glycans (similar to HMOs)and the comparative difficulty in accessing conjugated glycans,where the requirement for B. infantis EndoBI-1 expression cre-ates a rate-limiting step prior to consumption.

Only a select assortment of infant-associated strains of Bifido-bacterium, such as B. longum subsp. longum, B. infantis, B. breve,and B. bifidum, have the ability to grow to high cell densities onHMOs as the sole carbon source (25–28). B. bifidum and B. infan-tis represent the two models for HMO degradation. B. bifidum, aspecies well known to grow on the glycoprotein mucin (32, 53),deploys extracellular glycosyl hydrolases to degrade complexHMOs, followed by importation and consumption of select com-ponents. This “external” degradation phenotype results in the po-tential of released sugars cross-feeding other bacterial clades (54,55), or even pathogens, as has been witnessed for Bacteroides spe-cies (56). In contrast, B. infantis and, to a lesser extent, B. brevehave an “internal” degradation phenotype. B. infantis selectivelyimports HMOs via an array of specific family 1 solute bindingproteins that bind the oligosaccharides prior to transport by asso-ciated ABC transporters (32). Upon import, B. infantis deploys a

FIG 3 Normalized gene expression levels show a significant increase inBlon_2468 (EndoBI-1) expression relative to that in lactose-grown cells (P 0.0145) for cells incubated on whey protein concentrate but not for cells incu-bated on the same whey protein concentrate pretreated with EndoBI-1 toremove complex glycans from the underlying protein scaffold. Gene expres-sion was normalized using cysteinyl-tRNA synthetase gene expression.

FIG 4 Nano-LC-Chip-Q-TOF MS extracted compound chromatograms for free N-glycans from bovine colostrum whey concentrate remaining in the mediumafter bacterial fermentation (blue, uninoculated control; green, B. lactis; and red, B. infantis). Freed N-glycans from whey concentrate were produced as describedin Materials and Methods. Green circles, yellow circles, blue squares, red triangles, purple diamonds, and gray diamonds represent mannose, galactose, HexNAc,fucose, NeuAc, and NeuGc residues, respectively.

Selective Growth of Bifidobacteria on Milk N-Glycans

June 2016 Volume 82 Number 12 aem.asm.org 3627Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

suite of glycosyl hydrolases (�-fucosidases, �-sialidases, �-galac-tosidases, and N-acetyl-�-D-hexosaminidases), whose activity onHMOs has been described previously (31, 33–35). The latter, “in-ternal” degradation phenotype is likely a competitive advantageby removing the opportunity for cross-feeding.

Compared to the role of HMOs in shaping the distal gut mi-crobiota of infants, which has been described extensively (57), therole of glycoconjugates has been relatively unexplored. Glycopro-tein degradation strategies have been examined among severalbifidobacterial species (8, 58), and EndoBI-1 homologs have beenidentified and characterized for strains of B. longum subsp.longum, B. breve, and others (8). All are deployed on the cell wall,which facilitates the release of intact milk oligosaccharides that,like HMOs, are resistant to degradation by other gut bacteria butare readily transported and consumed by the EndoBI-1-produc-ing strain or neighboring bacteria.

The glycoproteins in milk thus provide an additional energysource for a select group of infant-associated gut bacteria (48). Inthis work, we demonstrate that the oligosaccharide portions of

bovine milk protein glycoconjugates, once released by an endogly-cosidase, can serve as selective growth substrates for infant-asso-ciated taxa, such as B. infantis. The nature of the specific release ofthese larger milk oligosaccharides from glycoconjugates instead ofthe release of sugar monomers would cross-feed only similarlycoevolved species, such as B. infantis, over less specific colonizersthat are unable to internalize and degrade these released glycans (8).

Finally, the production of synthetic HMO-like structures ischallenging due to the complexity of HMOs (22), and despitedifferences between the oligosaccharide compositions of bovineand human milks (22, 42), we demonstrate here that selective,bioactive milk oligosaccharides that mimic the specificity ofHMOs can be produced from widely available and low-cost bo-vine dairy sources. While bovine milk is generally considered to bea poor source of complex milk oligosaccharides (42), we showhere that bovine colostrum whey can be treated to release a signif-icant amount of complex milk glycans whose biological activityand enrichment specificity are intact. Future strategies that utilizeexisting products with demonstrable safety and low substrate cost

FIG 5 Glycoprofiling of consumption of freed N-glycans extracted from the medium after fermentation by two different Bifidobacterium species. (A) Con-sumption of freed neutral N-glycans from whey concentrate by B. lactis or B. infantis in comparison to uninoculated medium (control). (B) Consumption offreed sialylated N-glycans from whey concentrate by B. lactis or B. infantis in comparison to uninoculated medium (control). The x axis labels depict themonosaccharidic compositions of the N-glycans (Hex-HexNAc-Fuc-NeuAc-NeuGc); for example, the 4_3_0_1_0 composition is 4Hex-3HexNAc-0Fuc-1NeuAc-0NeuGc. Inset graphs show enlarged versions of the lower-abundance glycan types depicted within the indicated portion of each graph.

Karav et al.

3628 aem.asm.org June 2016 Volume 82 Number 12Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

may be an attractive alternative to chemical synthesis of complexHMO mimics.

ACKNOWLEDGMENTS

This work was supported by the UC Davis RISE program, National Insti-tutes of Health (NIH) awards AT008759 and AT007079, and the Peter J.Shields Endowed Chair in Dairy Food Science. S.K. was supported in partby the Ministry of Education, Turkey, and S.A.F. was supported in part byNIH grant F32AT008533.

D.A.M. and D.B. are cofounders of, and S.A.F. is an employee of,Evolve Biosystems, a company focused on diet-based manipulation of thegut microbiota.

FUNDING INFORMATIONThis work was funded by National Institutes of Health grants AT008759and AT007079 (D.A.M.) and F32AT008533 (S.A.F.), the Ministry of Ed-ucation, Turkey (S.K.), the UC Davis RISE program, and the Peter J.Shields Endowed Chair in Dairy Food Science (D.A.M.).

REFERENCES1. Moremen KW, Tiemeyer M, Nairn AV. 2012. Vertebrate protein glyco-

sylation: diversity, synthesis and function. Nat Rev Mol Cell Biol 13:448 –462. http://dx.doi.org/10.1038/nrm3383.

2. Van Berkel P, Geerts M, Van Veen H, Kooiman P, Pieper F, De Boer H,Nuijens J. 1995. Glycosylated and unglycosylated human lactoferrinsboth bind iron and show identical affinities towards human lysozyme andbacterial lipopolysaccharide, but differ in their susceptibilities towardstryptic proteolysis. Biochem J 312:107–114. http://dx.doi.org/10.1042/bj3120107.

3. Stanley P, Cummings RD. 2009. Structures common to different glycans,p 23–36. In Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P,Bertozzi CR, Hart GW, Etzler ME (ed), Essentials of glycobiology, 2nd ed.Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

4. Robinson LN, Artpradit C, Raman R, Shriver ZH, Ruchirawat M,Sasisekharan R. 2012. Harnessing glycomics technologies: integratingstructure with function for glycan characterization. Electrophoresis 33:797– 814. http://dx.doi.org/10.1002/elps.201100231.

5. Collin M, Olsén A. 2001. EndoS, a novel secreted protein from Strepto-coccus pyogenes with endoglycosidase activity on human IgG. EMBO J20:3046 –3055. http://dx.doi.org/10.1093/emboj/20.12.3046.

6. Muramatsu H, Tachikui H, Ushida H, Song X-J, Qiu Y, Yamamoto S,Muramatsu T. 2001. Molecular cloning and expression of endo-�-N-acetylglucosaminidase D, which acts on the core structure of complex typeasparagine-linked oligosaccharides. J Biochem 129:923–928. http://dx.doi.org/10.1093/oxfordjournals.jbchem.a002938.

7. Collin M, Fischetti VA. 2004. A novel secreted endoglycosidase fromEnterococcus faecalis with activity on human immunoglobulin G and ribo-nuclease B. J Biol Chem 279:22558 –22570. http://dx.doi.org/10.1074/jbc.M402156200.

8. Garrido D, Nwosu C, Ruiz-Moyano S, Aldredge D, German JB, LebrillaCB, Mills DA. 2012. Endo-beta-N-acetylglucosaminidases from infantgut-associated bifidobacteria release complex N-glycans from humanmilk glycoproteins. Mol Cell Proteomics 11:775–785. http://dx.doi.org/10.1074/mcp.M112.018119.

9. Byers HL, Tarelli E, Homer KA, Beighton D. 1999. Sequential deglyco-sylation and utilization of the N-linked, complex-type glycans of humana1-acid glycoprotein mediates growth of Streptococcus oralis. Glycobiol-ogy 9:469 – 479. http://dx.doi.org/10.1093/glycob/9.5.469.

10. Renzi F, Manfredi P, Mally M, Moes S, Jenö P, Cornelis GR. 2011. TheN-glycan glycoprotein deglycosylation complex (Gpd) from Capnocy-tophaga canimorsus deglycosylates human IgG. PLoS Pathog 7:e1002118.http://dx.doi.org/10.1371/journal.ppat.1002118.

11. Garbe J, Sjögren J, Cosgrave EF, Struwe WB, Bober M, Olin AI,Rudd PM, Collin M. 2014. EndoE from Enterococcus faecalis hydrolyzesthe glycans of the biofilm inhibiting protein lactoferrin and mediatesgrowth. PLoS One 9:e91035. http://dx.doi.org/10.1371/journal.pone.0091035.

12. Burnaugh AM, Frantz LJ, King SJ. 2008. Growth of Streptococcus pneu-moniae on human glycoconjugates is dependent upon the sequential ac-

tivity of bacterial exoglycosidases. J Bacteriol 190:221–230. http://dx.doi.org/10.1128/JB.01251-07.

13. Karav S, Parc AL, Moura Bell JM, Rouquie C, Mills DA, Barile D, BlockDE. 2015. Kinetic characterization of a novel endo-beta-N-acetylglucosaminidase on concentrated bovine colostrum whey to releasebioactive glycans. Enzyme Microb Technol 77:46 –53. http://dx.doi.org/10.1016/j.enzmictec.2015.05.007.

14. Martens EC, Roth R, Heuser JE, Gordon JI. 2009. Coordinate regulationof glycan degradation and polysaccharide capsule biosynthesis by a prom-inent human gut symbiont. J Biol Chem 284:18445–18457. http://dx.doi.org/10.1074/jbc.M109.008094.

15. Sonnenburg JL, Xu J, Leip DD, Chen C-H, Westover BP, WeatherfordJ, Buhler JD, Gordon JI. 2005. Glycan foraging in vivo by an intestine-adapted bacterial symbiont. Science 307:1955–1959. http://dx.doi.org/10.1126/science.1109051.

16. Crost EH, Tailford LE, Le Gall G, Fons M, Henrissat B, Juge N. 2013.Utilisation of mucin glycans by the human gut symbiont Ruminococcusgnavus is strain-dependent. PLoS One 8:e76341. http://dx.doi.org/10.1371/journal.pone.0076341.

17. Kim W-S, Ohashi M, Tanaka T, Kumura H, Kim G-Y, Kwon I-K, GohJ-S, Shimazaki K-I. 2004. Growth-promoting effects of lactoferrin on L.acidophilus and Bifidobacterium spp. Biometals 17:279 –283. http://dx.doi.org/10.1023/B:BIOM.0000027705.57430.f1.

18. Hernandez-Hernandez O, Sanz ML, Kolida S, Rastall RA, Moreno FJ.2011. In vitro fermentation by human gut bacteria of proteolytically di-gested caseinomacropeptide nonenzymatically glycosylated with prebi-otic carbohydrates. J Agric Food Chem 59:11949 –11955. http://dx.doi.org/10.1021/jf203576g.

19. Petschow B, Talbott R, Batema R. 1999. Ability of lactoferrin to promotethe growth of Bifidobacterium spp. in vitro is independent of receptorbinding capacity and iron saturation level. J Med Microbiol 48:541–549.http://dx.doi.org/10.1099/00222615-48-6-541.

20. Wu S, Tao N, German J, Grimm R, Lebrilla C. 2010. Development of anannotated library of neutral human milk oligosaccharides. J Proteome Res9:4138 – 4151. http://dx.doi.org/10.1021/pr100362f.

21. Wu S, Grimm R, German JB, Lebrilla CB. 2011. Annotation and struc-tural analysis of sialylated human milk oligosaccharides. J Proteome Res10:856 – 868. http://dx.doi.org/10.1021/pr101006u.

22. Garrido D, Dallas DC, Mills DA. 2013. Consumption of human milkglycoconjugates by infant-associated bifidobacteria: mechanisms and im-plications. Microbiology 159:649 – 664. http://dx.doi.org/10.1099/mic.0.064113-0.

23. Kunz C, Rudloff S, Baier W, Klein N, Strobel S. 2000. Oligosaccharidesin human milk: structural, functional, and metabolic aspects. Annu RevNutr 20:699 –722. http://dx.doi.org/10.1146/annurev.nutr.20.1.699.

24. Smilowitz JT, Lebrilla CB, Mills DA, German JB, Freeman SL. 2014.Breast milk oligosaccharides: structure-function relationships in the neo-nate. Annu Rev Nutr 34:143. http://dx.doi.org/10.1146/annurev-nutr-071813-105721.

25. Ward RE, Niñonuevo M, Mills DA, Lebrilla CB, German JB. 2007. Invitro fermentability of human milk oligosaccharides by several strains ofbifidobacteria. Mol Nutr Food Res 51:1398 –1405. http://dx.doi.org/10.1002/mnfr.200700150.

26. LoCascio RG, Niñonuevo MR, Kronewitter SR, Freeman SL, GermanJB, Lebrilla CB, Mills DA. 2009. A versatile and scalable strategy forglycoprofiling bifidobacterial consumption of human milk oligosaccha-rides. Microb Biotechnol 2:333–342. http://dx.doi.org/10.1111/j.1751-7915.2008.00072.x.

27. Asakuma S, Hatakeyama E, Urashima T, Yoshida E, Katayama T,Yamamoto K, Kumagai H, Ashida H, Hirose J, Kitaoka M. 2011.Physiology of consumption of human milk oligosaccharides by infantgut-associated bifidobacteria. J Biol Chem 286:34583–34592. http://dx.doi.org/10.1074/jbc.M111.248138.

28. Ruiz-Moyano S, Totten SM, Garrido DA, Smilowitz JT, German JB,Lebrilla CB, Mills DA. 2013. Variation in consumption of human milkoligosaccharides by infant gut-associated strains of Bifidobacterium breve.Appl Environ Microbiol 79:6040 – 6049. http://dx.doi.org/10.1128/AEM.01843-13.

29. Underwood MA, Kalanetra KM, Bokulich NA, Lewis ZT, Mirmiran M,Tancredi DJ, Mills DA. 2013. A comparison of two probiotic strains ofbifidobacteria in premature infants. J Pediatr 163:1585.e9 –1591.e9. http://dx.doi.org/10.1016/j.jpeds.2013.07.017.

30. Marcobal A, Barboza M, Froehlich J, Block D, German J, Lebrilla C,

Selective Growth of Bifidobacteria on Milk N-Glycans

June 2016 Volume 82 Number 12 aem.asm.org 3629Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Mills D. 2010. Consumption of human milk oligosaccharides by gut-related microbes. J Agric Food Chem 58:5334 –5340. http://dx.doi.org/10.1021/jf9044205.

31. Sela DA, Garrido D, Lerno L, Wu S, Tan K, Eom H-J, Joachimiak A,Lebrilla CB, Mills DA. 2012. Bifidobacterium longum subsp. infantisATCC 15697 �-fucosidases are active on fucosylated human milk oligo-saccharides. Appl Environ Microbiol 78:795– 803. http://dx.doi.org/10.1128/AEM.06762-11.

32. Garrido D, Kim JH, German JB, Raybould HE, Mills DA. 2011. Oligo-saccharide binding proteins from Bifidobacterium longum subsp. infantisreveal a preference for host glycans. PLoS One 6:e17315. http://dx.doi.org/10.1371/journal.pone.0017315.

33. Yoshida E, Sakurama H, Kiyohara M, Nakajima M, Kitaoka M,Ashida H, Hirose J, Katayama T, Yamamoto K, Kumagai H. 2012.Bifidobacterium longum subsp. infantis uses two different �-galactosi-dases for selectively degrading type-1 and type-2 human milk oligosac-charides. Glycobiology 22:361–368. http://dx.doi.org/10.1093/glycob/cwr116.

34. Sela DA, Li Y, Lerno L, Wu S, Marcobal AM, German JB, Chen X,Lebrilla CB, Mills DA. 2011. An infant-associated bacterial commensalutilizes breast milk sialyloligosaccharides. J Biol Chem 286:11909 –11918.http://dx.doi.org/10.1074/jbc.M110.193359.

35. Garrido D, Ruiz-Moyano S, Mills DA. 2012. Release and utilization ofN-acetyl-D-glucosamine from human milk oligosaccharides by Bifidobac-terium longum subsp. infantis. Anaerobe 18:430 – 435. http://dx.doi.org/10.1016/j.anaerobe.2012.04.012.

36. Azuma N, Yamauchi K, Mitsuoka T. 1984. Bifidus growth-promotingactivity of a glycomacropeptide derived from human k-casein. Agric BiolChem 48:2159 –2162.

37. Lee H, Garrido D, Mills DA, Barile D. 2014. Hydrolysis of milk ganglio-sides by infant-gut associated bifidobacteria determined by microfluidicchips and high-resolution mass spectrometry. Electrophoresis 35:1742–1750. http://dx.doi.org/10.1002/elps.201300653.

38. Ruas-Madiedo P, Gueimonde M, Fernández-García M, Clara G, Mar-golles A. 2008. Mucin degradation by Bifidobacterium strains isolatedfrom the human intestinal microbiota. Appl Environ Microbiol 74:1936 –1940. http://dx.doi.org/10.1128/AEM.02509-07.

39. Kiyohara M, Nakatomi T, Kurihara S, Fushinobu S, Suzuki H, TanakaT, Shoda S-I, Kitaoka M, Katayama T, Yamamoto K. 2012. �-N-Acetylgalactosaminidase from infant-associated bifidobacteria belongingto novel glycoside hydrolase family 129 is implicated in alternative mucindegradation pathway. J Biol Chem 287:693–700. http://dx.doi.org/10.1074/jbc.M111.277384.

40. Dallas DC, Martin WF, Strum JS, Zivkovic AM, Smilowitz JT,Underwood MA, Affolter M, Lebrilla CB, German JB. 2011. N-linkedglycan profiling of mature human milk by high-performance micro-fluidic chip liquid chromatography time-of-flight tandem mass spec-trometry. J Agric Food Chem 59:4255– 4263. http://dx.doi.org/10.1021/jf104681p.

41. Nwosu CC, Aldredge DL, Lee H, Lerno LA, Zivkovic AM, German JB,Lebrilla CB. 2012. Comparison of the human and bovine milk N-glycomevia high-performance microfluidic chip liquid chromatography and tan-dem mass spectrometry. J Proteome Res 11:2912–2924. http://dx.doi.org/10.1021/pr300008u.

42. Tao N, DePeters E, Freeman S, German J, Grimm R, Lebrilla C. 2008.Bovine milk glycome. J Dairy Sci 91:3768 –3778. http://dx.doi.org/10.3168/jds.2008-1305.

43. Barile D, Marotta M, Chu C, Mehra R, Grimm R, Lebrilla C, GermanJ. 2010. Neutral and acidic oligosaccharides in Holstein-Friesian colos-trum during the first 3 days of lactation measured by high performanceliquid chromatography on a microfluidic chip and time-of-flight massspectrometry. J Dairy Sci 93:3940 –3949. http://dx.doi.org/10.3168/jds.2010-3156.

44. Karav S, Bell JM, Le Parc A, Liu Y, Mills DA, Block DE, Barile D. 2015.Characterizing the release of bioactive N-glycans from dairy products by a

novel endo-beta-N-acetylglucosaminidase. Biotechnol Prog 31:1331–1339. http://dx.doi.org/10.1002/btpr.2135.

45. Le Parc A, Karav S, Bell JM, Frese SA, Liu Y, Mills DA, Block DE, BarileD. 2015. A novel endo-beta-N-acetylglucosaminidase releases specific N-glycans depending on different reaction conditions. Biotechnol Prog 31:1323–1330. http://dx.doi.org/10.1002/btpr.2133.

46. Parche S, Beleut M, Rezzonico E, Jacobs D, Arigoni F, Titgemeyer F,Jankovic I. 2006. Lactose-over-glucose preference in Bifidobacteriumlongum NCC2705: glcP, encoding a glucose transporter, is subject to lac-tose repression. J Bacteriol 188:1260 –1265. http://dx.doi.org/10.1128/JB.188.4.1260-1265.2006.

47. Le Parc A, Dallas DC, Duaut S, Leonil J, Martin P, Barile D. 2014.Characterization of goat milk lactoferrin N-glycans and comparison withthe N-glycomes of human and bovine milk. Electrophoresis 35:1560 –1570. http://dx.doi.org/10.1002/elps.201300619.

48. Zivkovic AM, Barile D. 2011. Bovine milk as a source of functionaloligosaccharides for improving human health. Adv Nutr 2:284 –289. http://dx.doi.org/10.3945/an.111.000455.

49. Petschow B, Talbott R, Batema R. 1999. Ability of lactoferrin to promotethe growth of Bifidobacterium spp. in vitro is independent of receptorbinding capacity and iron saturation level J Med Microbiol 48:541. http://dx.doi.org/10.1099/00222615-48-6-541.

50. Liepke C, Adermann K, Raida M, Mägert HJ, Forssmann WG, ZuchtHD. 2002. Human milk provides peptides highly stimulating the growthof bifidobacteria. Eur J Biochem 269:712–718. http://dx.doi.org/10.1046/j.0014-2956.2001.02712.x.

51. Garrido D, Ruiz-Moyano S, Lemay DG, Sela DA, German JB, Mills DA.2015. Comparative transcriptomics reveals key differences in the responseto milk oligosaccharides of infant gut-associated bifidobacteria. Sci Rep5:13517. http://dx.doi.org/10.1038/srep13517. (Erratum, 5:15311, http://dx.doi.org/10.1038/srep15311.)

52. LoCascio RG, Ninonuevo MR, Freeman SL, Sela DA, Grimm R, LebrillaCB, Mills DA, German JB. 2007. Glycoprofiling of bifidobacterial con-sumption of human milk oligosaccharides demonstrates strain specific,preferential consumption of small chain glycans secreted in early humanlactation. J Agric Food Chem 55:8914 – 8919. http://dx.doi.org/10.1021/jf0710480.

53. Turroni F, Bottacini F, Foroni E, Mulder I, Kim J-H, Zomer A, SánchezB, Bidossi A, Ferrarini A, Giubellini V. 2010. Genome analysis of Bifi-dobacterium bifidum PRL2010 reveals metabolic pathways for host-derived glycan foraging. Proc Natl Acad Sci U S A 107:19514 –19519. http://dx.doi.org/10.1073/pnas.1011100107.

54. Turroni F, Ozcan E, Milani C, Mancabelli L, Viappiani A, van SinderenD, Sela DA, Ventura M. 2015. Glycan cross-feeding activities betweenbifidobacteria under in vitro conditions. Front Microbiol 6:1030. http://dx.doi.org/10.3389/fmicb.2015.01030.

55. Egan M, Motherway MOC, Kilcoyne M, Kane M, Joshi L, Ventura M,van Sinderen D. 2014. Cross-feeding by Bifidobacterium breve UCC2003during co-cultivation with Bifidobacterium bifidum PRL2010 in a mucin-based medium. BMC Microbiol 14:1. http://dx.doi.org/10.1186/1471-2180-14-1.

56. Ng KM, Ferreyra JA, Higginbottom SK, Lynch JB, Kashyap PC, GopinathS, Naidu N, Choudhury B, Weimer BC, Monack DM. 2013. Microbiota-liberated host sugars facilitate post-antibiotic expansion of enteric pathogens.Nature 502:96–99. http://dx.doi.org/10.1038/nature12503.

57. De Leoz MLA, Kalanetra KM, Bokulich NA, Strum JS, Underwood MA,German JB, Mills DA, Lebrilla CB. 2015. Human milk glycomics and gutmicrobial genomics in infant feces show a correlation between humanmilk oligosaccharides and gut microbiota: a proof-of-concept study. JProteome Res 14:491–502. http://dx.doi.org/10.1021/pr500759e.

58. Sakurama H, Kiyohara M, Wada J, Honda Y, Yamaguchi M, Fukiya S,Yokota A, Ashida H, Kumagai H, Kitaoka M. 2013. Lacto-N-biosidaseencoded by a novel gene of Bifidobacterium longum subspecies longumshows unique substrate specificity and requires a designated chaperone forits active expression. J Biol Chem 288:25194 –25206. http://dx.doi.org/10.1074/jbc.M113.484733.

Karav et al.

3630 aem.asm.org June 2016 Volume 82 Number 12Applied and Environmental Microbiology

on Septem

ber 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from