Differential Utilization of Basic Proline-Rich Glycoproteins duringGrowth of Oral Bacteria in Saliva

Yuan Zhou,a* Jinghua Yang,a* Luxia Zhang,a Xuedong Zhou,b John O. Cisar,a Robert J. Palmer, Jr.a

Microbial Receptors Section, Laboratory of Cell and Developmental Biology, National Institutes of Health, Bethesda, Maryland, USAa; State Key Laboratory of Oral Diseases,West China Hospital of Stomatology, Sichuan University, Chengdu, Chinab

ABSTRACT

Although saliva is widely recognized as a primary source of carbon and nitrogen for growth of the dental plaque biofilmcommunity, little is known about how different oral bacteria utilize specific salivary components. To address this question,32 strains representing 16 genera commonly isolated from early plaque biofilms were compared for growth over two trans-fers in stimulated (by chewing Parafilm) whole saliva that was stabilized by heat treatment and dialysis. The cell densities,measured by quantitative PCR (qPCR), ranged from �1 � 106 to 1 � 107/ml for strains of Streptococcus gordonii, Strepto-coccus oralis, and Streptococcus mitis and one strain of Streptococcus sanguinis. Strains of Streptococcus mutans, Gemella hae-molysans, and Granulicatella adiacens reached �1 � 105 to 1 � 106/ml. In contrast, little or no growth was noted for threeother strains of S. sanguinis, as well as for strains of Streptococcus parasanguinis, Streptococcus salivarius, Streptococcus ves-tibularis, Streptococcus sobrinus, Actinomyces spp., Abiotrophia defectiva, and Rothia dentocariosa. SDS-PAGE, lectin blotting,and two-dimensional gel electrophoresis of saliva from cultures of S. gordonii, S. oralis, and S. mitis revealed species-spe-cific differences in the degradation of basic proline-rich glycoproteins (PRG). In contrast, saliva from cultures of otherbacteria was indistinguishable from control saliva. Species-dependent differences in the utilization of individual host sug-ars were minor. Thus, differences in salivary glycan foraging between oral species may be important to cross-feeding andcooperation between organisms in dental plaque biofilm development.

IMPORTANCE

Bacteria in the mouth use saliva for nutrition. How each of the many types of bacteria uses saliva is not clear. We show that amajor protein in saliva, called PRG, is an important nutrition source for certain bacteria but not for others. PRG has many sugarmolecules linked in chains, but the sugar is not available for bacteria until the chains are degraded. The bacteria that can grow bydigesting this protein break the sugar chains into parts which not only support their own growth but could also be available tosupport the growth of those bacteria that cannot use the intact protein.

Dental plaque biofilm development involves an array of spe-cific interactions between bacteria and saliva, beginning with

the attachment of early colonizers, including Streptococcus sangui-nis, Streptococcus oralis, Streptococcus gordonii, and Streptococcusmitis, through binding to salivary components in the acquiredenamel pellicle (1, 2). While interbacterial adhesion (i.e., coaggre-gation) between these pioneers and other bacteria enhances spe-cies diversity in the nascent biofilm (3), community maturationdepends primarily on the growth of different closely associatedbacteria (4, 5). Growth of bacteria, measured as an increase inoptical density (OD), was noted in saliva inoculated with dentalplaque (6) but not in saliva inoculated with monocultures ofStreptococcus spp. and Actinomyces spp. unless glucose was addedas a carbon source (7). However, growth of S. gordonii DL1 to celldensities below those measurable by OD (i.e., �107 CFU/ml) hasbeen observed in 25% saliva without added glucose (8). Thisgrowth, which presumably depends on salivary glycan foraging,has not been correlated to catabolism of specific substrates in sa-liva. In this regard, metabolic activity of bacteria was observedduring incubation of fresh dental plaque in the presence ofMUC5B but not in similar experiments performed with individ-ual plaque isolates of S. oralis or S. gordonii (9).

Several oral streptococci have been found to grow on 1% piggastric mucin (10), and a few are known to grow on 0.5% humanserum �1-acid glycoprotein (11). Observations made in studies

with these model glycoproteins include elevated cell surface gly-cosidase and sugar phosphotransferase activities in Streptococcusintermedius (12), as well as in consortia (13), when grown onglycoprotein rather than on glucose. Sialidase-dependent se-quential removal of sugars from N-linked glycan chains by S.oralis (14) and cooperative degradation of mucin by mixed cul-tures of different oral species (15) have also been reported. Thesefindings suggest a link between salivary glycan foraging and cross-feeding between species, and they emphasize the need for studies

Received 29 April 2016 Accepted 13 June 2016

Accepted manuscript posted online 17 June 2016

Citation Zhou Y, Yang J, Zhang L, Zhou X, Cisar JO, Palmer RJ, Jr. 2016. Differentialutilization of basic proline-rich glycoproteins during growth of oral bacteria insaliva. Appl Environ Microbiol 82:5249 –5258. doi:10.1128/AEM.01111-16.

Editor: A. M. Spormann, Stanford University

Address correspondence to Robert J. Palmer, Jr., rjpalmer@dir.nidcr.nih.gov.

* Present address: Yuan Zhou, State Key Laboratory of Oral Diseases, West ChinaHospital of Stomatology, Sichuan University, Chengdu, China; Jinghua Yang, StateKey Laboratory of Mycology, Institute of Microbiology, Chinese Academy ofSciences, Beijing, China.

For a companion article on this topic, see doi:10.1128/AEM.01291-16.

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

crossmark

September 2016 Volume 82 Number 17 aem.asm.org 5249Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

that compare the ability of individual oral bacteria to grow insaliva.

In the present study, we combined quantitative PCR(qPCR) with electrophoretic methods to assess the growth andglycan foraging of oral bacteria in stimulated whole saliva thathad been stabilized by heat treatment and subsequent dialysis.We compared 32 strains from the early biofilm genera Strepto-coccus, Actinomyces, Gemella, Granulicatella, Abiotrophia, andRothia. Growth occurred for only roughly one-third of the strains,most notably those of S. oralis, S. gordonii, and S. mitis, and growthwas associated with species-dependent differences in the degrada-tion of basic proline-rich glycoproteins (PRG).

MATERIALS AND METHODSBacterial strains and culture conditions. The bacterial strains used in thisstudy are listed in Table 1. Streptococci and actinomyces were grown in

Todd-Hewitt broth (THB; BD, Franklin Lakes, NJ). Granulicatella spp.and Abiotrophia defectiva were grown in brain heart infusion (BHI)broth (Difco; BD, Franklin Lakes, NJ) supplemented with 10 mg/literpyridoxal-HCl (Sigma-Aldrich, St. Louis, MO). Gemella haemolysanswas grown in BHI broth supplemented with pyridoxal and 10% (wt/vol) horse serum (Remel, Lenexa, KS). All bacteria were grown over-night at 37°C in a 5% CO2 incubator as static tube cultures (10 ml).

Growth of bacteria in saliva. All work with saliva was performed usingplastic tubes and beakers. Stimulated (by chewing Parafilm) whole salivafrom two healthy individuals was collected on ice. Volumes of fifty milli-liters of each saliva sample were combined, reduced with 1.25 mM (finalconcentration) dithiothreitol (Sigma-Aldrich) for 10 min with gentlestirring on ice, and then centrifuged at 15,000 � g (4°C) for 30 min.The supernatants were combined and sterile filtered (Express PLUS,0.22-�m pore size; EMD Millipore, Billerica, MA), incubated for 30min at 70°C, and then dialyzed (12- to 14-kDa-cutoff Spectra/Portubing; Spectrum Labs, Rancho Dominguez, CA) against two over-night changes (2 liters each) of a saliva-salts buffer (Table 2) based onthe composition of synthetic saliva (22). Dialyzed saliva was centri-fuged, filter sterilized again, and stored at 4°C for no longer than 24 hprior to use in growth studies.

Growth experiments were initiated with bacteria from complexmedium that were washed twice by centrifugation in saliva, suspendedto the original culture volume in saliva, and then diluted 1,000-foldinto 2-ml aliquots of saliva (i.e., the primary saliva cultures). Follow-ing overnight incubation at 37°C in 5% CO2, each primary saliva cul-ture was diluted 1,000-fold into three 6-ml aliquots of fresh saliva (i.e.,the secondary saliva cultures) in plastic tubes, which were then incu-bated for 20 h at 37°C in 5% CO2. Uninoculated saliva was incubatedas the control. DNA templates for qPCR were prepared by pelleting(14,000 � g for 10 min) bacteria from 1-ml aliquots of each secondarysaliva culture taken immediately after inoculation from the primaryculture and after overnight incubation. Pellets were suspended in 180�l lysis buffer consisting of 1,670 U/ml mutanolysin (Sigma-Aldrich),20 mg/ml lysozyme (Sigma-Aldrich), 20 mM Tris HCl, pH 8.0, 2 mMEDTA, and 1.2% Triton X-100 (Sigma-Aldrich) and then incubatedfor 1 h at 37°C. DNA was extracted from lysed cell suspensions using aDNeasy blood and tissue kit (Qiagen, Valencia, CA) following themanufacturer’s protocol for Gram-positive bacteria. Species-specificprimers for 16S rRNA genes (Table 3) were designed using Vector NTI11.5 software (Invitrogen, Carlsbad, CA). Although nonspecific prim-ers would have sufficed for the experiments reported on in detail here,

TABLE 1 Bacteria used in this study

Species StrainSource orreference(s)

Abiotrophia defectiva 49176 ATCCa

Actinomyces oris T14V 1627044 ATCC49339 ATCCMG1 16

Actinomyces naeslundii 12104 ATCCGemella haemolysans 10379 ATCCGranulicatella adiacens 49175 ATCCGranulicatella elegans 700633 ATCCRothia dentocariosa 17931 ATCC

Streptococcus gordonii DL1 16, 1738 16, 1710558 ATCC

Streptococcus mitis SK137 17, 18SK142 17, 18

Streptococcus mutans UA519 1925175 ATCC

Streptococcus oralis H1 16, 17SK23 17, 18C104 16, 1734 16, 17Uo5 20

Streptococcus parasanguinis FW213 2115912 ATCC

Streptococcus salivarius 27945 ATCC

Streptococcus sanguinis 10556 ATCCSK36 17, 18SK112 17, 18SK45 17, 18

Streptococcus sobrinus 27351 ATCC33478 ATCC

Streptococcus vestibularis 49124 ATCCa ATCC, American Type Culture Collection.

TABLE 2 Saliva-salts buffera

Constituent Concn [mg/liter (mM)]

Ammonium chloride 233Calcium chloride 55.5 (0.5)Magnesium chloride 20.2Potassium chloride 1,162Potassium dihydrogen phosphate 1,062 (7.8)Sodium thiocyanate 185Sodium citrate 15Sodium hydrogen carbonate 535Disodium hydrogen phosphate 1,732 (12.2)Urea 173Uric acid 10.5Creatinine 0.1Choline 17Casamino Acids 41Yeast extract 0.8a The concentrations of constituents are similar to those in synthetic saliva (16), exceptfor calcium chloride, which was reduced from 1.4 mM to 0.5 mM, and phosphatebuffer, which was increased from 5.2 mM to 20 mM. In addition, Casamino Acids andyeast extract were used in place of specific amino acid and vitamin mixtures.

Zhou et al.

5250 aem.asm.org September 2016 Volume 82 Number 17Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

we used the specific primers for all these experiments with an eyetoward their use in the mixed-species experiments noted in Discus-sion. Each 25-�l qPCR mixture contained 5 �l of DNA template, 12.5�l of iQ SYBR green supermix (Bio-Rad, Hercules CA), 2.5 �l of 10�ROX passive reference dye (Bio-Rad), and 2.5 �l of each forward andreverse primer at a 10� concentration of 3.33 �M. The qPCRs wereperformed with a Stratagene MX3500P thermal cycler programmedfor 3 min at 95°C, followed by 40 cycles of 15 s at 95°C and 30 s at 60°C.Dissociation curves were generated by incubating the reaction prod-ucts at 95°C for 20 s and 60°C for 20 s, followed by 95°C for 20 s.Fluorescence data were collected at the end of the 60°C primer-anneal-ing step for 40 amplification cycles and throughout the dissociationcurve analysis. Melt curve analysis showed a single sharp peak from allprimer sets. The cell concentration (cells/ml) was calculated based onstandard curves from 10-fold serial dilutions of bacterial genomicDNA (23).

Salivary protein and glycoprotein analyses. Saliva from control andbacterial cultures was filtered (0.22 �m) prior to all biochemical assays.Samples were assayed in triplicate for protein and carbohydrate, using thePierce Micro BCA protein assay kit (Thermo Fisher, Grand Island NY) forprotein, with bovine serum albumin as the standard, and the phenol-sulfuric acid method for carbohydrate (24), with glucose as the standard.Samples (5 �g salivary protein) for SDS-PAGE were heated at 100°C for 5min in loading dye (Invitrogen) with 5% �-mercaptoethanol (MP Bio-medicals, Santa Ana CA) and then loaded on 4 to 12% Bis-Tris gels (In-vitrogen, Grand Island, NY). Gels were run with morpholinepropanesul-

fonic acid (MOPS) buffer and silver stained using the SilverQuest kit(Invitrogen). For lectin blotting, samples (10 �g salivary protein) of SDS-PAGE-separated saliva were transferred to nitrocellulose membranes us-ing a Trans-Blot Turbo transfer system (Bio-Rad, Hercules, CA).Transfers were blocked overnight in Tris-buffered saline, pH 7.4, con-taining 2% polyvinyl alcohol (type II, low molecular weight; Sigma-Aldrich), 0.1% Tween 20 (Sigma-Aldrich), 1 mM CaCl2, 1 mM MnCl2,and 1 mM MgCl2 and then incubated for 2 h with fluorescein-labeledlectins at final concentrations that ranged from 5 to 25 �g/ml in block-ing buffer. Lectin probes included Sambucus nigra agglutinin (SNA),Erythrina cristagalli agglutinin (ECA), succinylated wheat germ agglu-tinin (SWGA), and Aleuria aurantia lectin (AAL), all from Vector Labo-ratories (Burlingame, CA). Galanthus nivalis agglutinin (GNA) was ob-tained from EY Laboratories (San Mateo, CA). Transfers were washedwith phosphate-buffered saline containing 0.1% Tween 20 (Sigma-Aldrich) to remove unbound lectin and then documented (excitationat 460 nm, emission at 500 to 520 nm, standard resolution, and 500-msexposure) with an ImageQuant LAS4000 image reader (GE HealthcareLife Sciences, Pittsburgh, PA).

For two-dimensional (2-D) gel electrophoresis, salivary proteinswere precipitated with 2 volumes of ice-cold acetone for 30 minat �20°C prior to centrifugation at 14,000 � g for 30 min at 4°C.Protein pellets were rinsed with ice-cold acetone, air dried, and thenresuspended in DeStreak rehydration solution (GE Healthcare). Iso-electric focusing (IEF) strips (11 cm, pH 3 to 10 or 6 to 11, linear; GEHealthcare) were rehydrated overnight with 200 �l of sample contain-ing 50 �g of protein and 0.5% IPG buffer (GE Healthcare). Focusingwas performed using a Protean i12 system (Bio-Rad). Proteins sepa-rated by isoelectric focusing were then reduced with dithiothreitol andalkylated with iodoacetamide prior to SDS-PAGE and subsequent sil-ver staining.

Growth of bacteria on host sugars. Experiments were initiatedwith bacteria from complex medium that were washed and suspendedin minimal medium (0.2% yeast extract, 0.2% Casamino Acids inphosphate-buffered saline) to the original culture volume and thendiluted 1:20 into minimal medium supplemented with 0.5% (wt/vol)filter-sterilized D-glucose, D-galactose, D-mannose, L-fucose, N-acetyl-D-galactosamine, or N-acetyl-D-glucosamine (all from Sigma-Aldrich)or N-acetylneuraminic acid (Zymetx, Inc., Oklahoma City, OK). Mea-surements of overnight growth at 37°C were made with a Klett-Sum-merson colorimeter (red filter) and were corrected for the smallamount of turbidity in control cultures that contained minimal me-dium alone.

RESULTSGrowth of bacteria in whole saliva. In initial studies, we foundthat overnight incubation of control saliva cultures at 37°C wasaccompanied by the appearance of mineralized precipitate and apH shift from neutral to alkaline (�8). These changes, which re-flect the metastable nature of whole human saliva (25), were notobserved in saliva that was first dialyzed against saliva-salts buffer(Table 2) to reduce the concentration of calcium ions (from �1.4mM to 0.5 mM) and to increase the concentration of phosphatebuffer (from �5 mM to 20 mM). Another early obstacle arosefrom changes in SDS-PAGE and lectin-blotting profiles of controlsaliva cultures that obscured concomitant changes associated withbacterial growth. Thus, overnight incubation of untreated saliva at37°C resulted in loss of silver-stained (Fig. 1A) and lectin-binding(Fig. 1B and C) salivary components (Fig. 1A to C, compare lanesU4 with lanes U37). While the initial protein profiles and lectinbinding of treated saliva varied somewhat from those of untreatedsaliva (Fig. 1A to C, compare lanes U4 with lanes T4), loss of silverstain and of bound lectin in a region around �70 kDa was signif-icantly reduced when saliva was heat treated (70°C for 30 min)

TABLE 3 16S ribosomal primers used in this study

Primer Sequence

Ab. defectiva-F GGGATGTCAAGACCTGGTAAb. defectiva-R CGCAAGGCTGAAACTCAAA. oris-F TTGCTGGTTCTGGATGAGTGA. oris-R CAGGCACAGAATATCCCGTATTAA. naeslundii-F GGCTGCGATACCGTGAGGA. naeslundii-R TCTGCGATTACTAGCGACTCCG. haemolysans-F GTGGAGGGTCATTGGAAACTG. haemolysans-R CCACTGGTGTTCCTCCTAATCG. adiacens-F CCTTGACGGTATCTAACCAGAAAG. adiacens-R CGCTTTACGCCCAATAAATCG. elegans-F CCTTGACGGTATCTAACCAGAAAG. elegans-R CGCTTTACGCCCAATAAATCR. dentocariosa-F CGTGAGTGACCTACCTTTGACR. dentocariosa-R CATAACGCTTTCCACCAACS. gordonii-F GCTTGCTACACCATAGACTS. gordonii-R CCGTTACCTCACCTACTAGS. mitis-F GCAGGCGGTTAGATAAGTS. mitis-R CTCAGCGTCAGTTACAASK137-F CACACCCGTTCTTCTCTTACASK137-R TAGGGAATCTTCGGCAATGGS. mutans-F GATAATTGATTGAAAGATGCAAGCS. mutans-R ATTCCCTACTGCTGCCTCCCS. oralis-F CCGCATAAGAGTAGATGTTGS. oralis-R TATGTATCGTTGCCTTGGTS. parasanguinis-F GTAGGGAATCTTCGGCAATGS. parasanguinis-R CACTCTCACACTCGTTCTTCTCS. salivarius-F GTGACGGTAGCTTACCAGAAAS. salivarius-R CGCTTTACGCCCAATAAATCS. sanguinis-F AGTTGCCATCATTGAGTTGS. sanguinis-R GTACCAGCCATTGTAACACS. sobrinus-F TGCTCCAGTGTTACTAATGAS. sobrinus-R TAACTCCTCTTATGCGGTATTS. vestibularis-F GTGACGGTAGCTTACCAGAAAGS. vestibularis-R CGCTTTACGCCCAATAAATC

Growth of Oral Bacteria in Saliva

September 2016 Volume 82 Number 17 aem.asm.org 5251Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

immediately after collection (Fig. 1A to C, compare lanes T4 withlanes T37). Thus, silver staining and prominent lectin binding ofthis region occurred in treated and untreated saliva alike (Fig. 1Ato C, lanes U4 and T4), but the reactivity of the region was com-pletely lost unless saliva was heat treated (Fig. 1A to C, lanes U37and T37).

The growth of oral bacteria in stabilized saliva was measured byqPCR. Individual experiments were performed three times for allstreptococci, for A. oris strain T14V, and for the two nonstrepto-cocci that grew. Otherwise, if an organism failed to grow, no fur-ther experiments were conducted, but subsequent biochemicalanalysis of saliva was performed in every case. Increases in celldensity are shown in Fig. 2 and are from one representative exper-iment with each strain. Upon inoculation, the secondary salivacultures generally contained from 103 to 104 cells/ml. After over-night growth, the cell densities of S. gordonii, S. oralis, and S. mitiscultures and also of S. sanguinis ATCC 10556 cultures increased100- to 1,000-fold (final densities of 106 to 107 cells/ml). Lessgrowth was noted for strains of Streptococcus mutans, Granulica-tella elegans, and Gemella haemolysans, all of which increased 10-to 100-fold (final densities of 105 to 106 cells/ml). In contrast,strains of Streptococcus parasanguinis, Streptococcus sobrinus,Streptococcus vestibularis, Streptococcus salivarius, Actinomycesspp., Rothia dentocariosa, Abiotrophia defectiva, and the remainingthree strains of S. sanguinis either failed to grow or reached at most�8 � 104 cells/ml. Thus, only certain streptococci grew well insaliva, whereas an intermediate level of growth was seen with somestreptococci, as well as with Gemella haemolysans and one Granu-licatella strain. All other strains, including most nonstreptococci,failed to grow.

Growth-related changes in saliva. Significant reductions intotal protein (as measured by the Micro BCA assay) and in totalcarbohydrate (as measured by the phenol sulfuric acid assay) weregenerally not seen in saliva cultures of different oral species (re-sults not shown). However, clear differences were noted betweensilver-stained SDS-PAGE profiles of saliva from uninoculatedcontrols and secondary cultures of all S. gordonii, S. oralis, and S.mitis strains (Fig. 3). Saliva from these streptococcal cultures wasdevoid of silver-stained protein in the �65- to 75-kDa region.Additional changes included an absence of bands in the 20- to25-kDa region for S. gordonii cultures and the appearance of bandsin the 35- to 55-kDa region for S. oralis cultures. Minor bands inthe 50-kDa region also appeared in S. mitis cultures. Saliva fromcultures of other bacteria, regardless of whether they were inter-mediate or poor growers, was indistinguishable from control sa-liva; G. haemolysans, S. mutans, and A. oris are shown as examplesin Fig. 3.

Similar to the results of one-dimensional (1-D) electrophore-sis, 2-D electrophoresis of control saliva and saliva from secondarycultures of S. gordonii DL1, S. oralis strain C104, and S. mitis strainSK137 revealed changes in a relatively small number of compo-nents, most notably in a region from control saliva that migratedat �70 kDa that was presumptively identified as basic proline-richglycoprotein (PRG) (26) by its high pI (Fig. 4). The apparent mo-lecular mass of PRG in saliva cultures of each streptococcus wasshifted downward to the �55-kDa region, and much weakerstaining of PRG was seen in saliva from S. gordonii DL1 cultures.Importantly, the novel bands seen in 1-D gels of saliva from cul-tures of S. oralis C104 and S. mitis SK137 were also shown tomigrate at high pI. The growth of strain DL1 (but not of strain

FIG 1 (A to C) Untreated saliva (U) and heat-treated saliva (T) were incubated overnight at 4°C (U4 and T4) or 37°C (U37 and T37). SDS-PAGE gels weresilver stained (A) or transferred to nitrocellulose membranes and incubated with an FITC-conjugated lectin, i.e., either sialic acid-specific SNA (B) orgalactose-specific ECA (C). Heat-treatment reduced degradation during incubation at 37°C, and a lectin-binding region at �70 kDa was retained only intreated saliva.

Zhou et al.

5252 aem.asm.org September 2016 Volume 82 Number 17Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

C104 or SK137) was also associated with the loss of an acidicproline-rich phosphoprotein that migrated at �25 kDa (26), aswell as minor basic components that appeared below 20 kDa. Fur-ther resolution of all these key spots was achieved by narrow-rangeIEF (pH 6 to 11) (Fig. 5).

Saliva samples from control and secondary cultures of theseselected strains were separated by SDS-PAGE, transferred tonitrocellulose membranes, and probed with fluorescein iso-thiocyanate (FITC)-labeled lectins (Fig. 6). The diffuse �70-kDa band seen in control saliva (as in Fig. 3) bound sialicacid-specific SNA, galactose-specific ECA, N-acetylgluco-samine-specific SWGA, mannose-specific GNA and L-fucose-specific AAL. Binding of ECA was abolished in saliva fromcultures of S. gordonii, S. oralis, and S. mitis, while binding ofother lectins was shifted downward in accordance with thedownward migration of basic PRG (as in Fig. 4). Relativelystrong binding of lectins other than ECA occurred in the �60-kDa region for saliva from cultures of S. mitis, but weak bindingwas seen to saliva from cultures of S. gordonii. In contrast, lectin

binding to saliva from S. oralis cultures was limited to the novelbands in the 35- to 65-kDa region (Fig. 3). Differences in theglycan compositions of these bands were suggested by the re-action of all lectins (except ECA) with bands in the 50- to60-kDa region but of only SWGA and AAL with the most rap-idly migrating band (�35 kDa). In contrast with these changes,lectin blots of saliva from cultures of all other bacteria, includ-ing those that grew to intermediate levels, such as G. haemoly-sans and S. mutans, and those that failed to grow, such as A. oris,were like those of control saliva.

Growth of bacteria on host sugars. To complement the stud-ies of growth in saliva, we determined the growth profiles of thestrains in minimal medium supplemented with different host-derived sugars (Fig. 7). All strains grew on glucose, and many alsogrew on galactose, N-acetylglucosamine, and mannose. However,G. haemolysans, G. elegans, G. adiacens, and R. dentocariosa failedto grow on galactose, S. sobrinus, S. salivarius, S. vestibularis, andR. dentocariosa failed to grow on N-acetylglucosamine, and S. ora-lis H1, S. mutans, A. oris, and S. sobrinus failed to grow on man-

FIG 2 Growth of oral bacteria in secondary cultures of whole saliva. Cell concentrations (cells/ml) in secondary saliva cultures were measured by qPCRimmediately after inoculation (hatched bars) and after 20 h of incubation at 37°C (open bars). Error bars show the �standard deviation (SD) (n 3). Strains ofS. gordonii, S. mitis, and S. oralis grew well; the growth of other species was classified as either intermediate or poor. *, NCTC 10449.

Growth of Oral Bacteria in Saliva

September 2016 Volume 82 Number 17 aem.asm.org 5253Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

nose. Interestingly, only Ab. defectiva and S. sanguinis SK36 grewon N-acetylgalactosamine, and only strains of Actinomyces spp.grew on L-fucose. Actinomyces spp. also grew on N-acetyl-neuraminic acid, as did G. haemolysans and certain streptococci,including S. mitis SK137 and three of four S. sanguinis strains.Thus, considerable variation exists in the ability of oral bacteria togrow on different host-derived sugars.

DISCUSSION

The present study was initiated to determine whether members ofthe dental plaque biofilm community differ in their ability to growas monocultures in undiluted whole saliva. We stabilized saliva byusing heat treatment to inactivate endogenous hydrolytic en-zymes. The saliva was then dialyzed to promote pH stability and toeliminate mineral precipitation (as well as to remove potentiallow-molecular-weight growth substrates), and then it was filtersterilized. Bacteria were transferred from standard growth me-dium into saliva, incubated overnight to activate salivary-sub-strate-specific enzymes, and then transferred to fresh saliva at alow density to allow an accurate assessment of growth evenwhen cultures grew poorly. Measurement by PCR rather thanCFU reduces potential error arising from clumped cells, andwhile it is possible that some intrinsic bacterial DNA remainedafter saliva preparation, the high level of growth seen for manystrains suggests the concentration of any such template to below. Furthermore, species-specific primers reduce the genera-tion of PCR product that might have arisen from amplificationof residual DNA template had a general eubacterial primerbeen used.

After 1 day of growth, the cell densities for strains of S. gordonii,S. oralis, and S. mitis ranged from 106 to 107 cells per ml (Fig. 2).Strains of G. haemolysans, S. mutans, and G. elegans reached 105 to

106 cells per ml. While significant growth of one S. parasanguinisand one S. sanguinis strain was also noted, other strains of thesespecies, as well as strains of S. salivarius, S. vestibularis, S. sobri-nus, Actinomyces spp., R. dentocariosa, and Ab. defectiva, grew ei-ther poorly or not at all. Changes in the SDS-PAGE profiles ofsaliva from cultures of these bacteria were limited to strains ofS. gordonii, S. oralis, and S. mitis (Fig. 3) and were associatedalmost exclusively with the degradation of silver-stained pro-tein in the region around 70 kDa. When examined using 2-Delectrophoresis (Fig. 4 and 5), this region was shown to behighly basic, which implied it to be PRG. Conclusive identifi-cation by liquid chromatography-tandem mass spectrometry(LC-MS/MS) was obtained in a subsequent study that exam-ined the role of specific cell surface hydrolases in the growth ofS. gordonii in saliva (27).

Deglycosylation that corresponded to changes in PRG mi-gration was also evident from the patterns of labeling on lectinblots (Fig. 6). The binding of ECA, which reacts well with ter-minal Gal�1-4GlcNAc and Fuc�1-2Gal�1-4GlcNAc (28, 29),was abolished in cultures of S. gordonii, S. oralis, and S. mitis,while the binding of other lectins to different features of N-linked glycan chains shifted from the �70-kDa region in con-trol saliva down to the region from �35 to 60 kDa, whichincludes the regions (35 to 46 kDa) expected for PRB3 andPRB4 protein chains (30). N-Acetylglucosamine with an �-1-6-linked fucose branch (31) is the innermost (peptide-linked)monosaccharide of PRG. Although structural data are neededto draw an unambiguous conclusion, the binding of only fu-cose-specific AAL and N-acetylglucosamine-specific SWGA tothe 35-kDa band suggests it may be PRG that bears primarilythis single disaccharide. Considered together, these findings

FIG 3 Silver-stained SDS-PAGE profiles of saliva from uninoculated (control) and secondary cultures of selected strains. Protein that migrated in the�70-kDa region (solid box) was reduced in or absent from saliva cultures of all S. gordonii, S. oralis, and S. mitis strains. In addition, bands in the �25-kDaregion of S. gordonii saliva cultures (dotted box) were reduced or absent, whereas novel bands appeared in the 35- to 55-kDa region of S. oralis salivacultures (dashed box). Saliva from cultures of all other bacteria was indistinguishable from control saliva regardless of the bacterium’s ability to grow.Shown here as examples are G. haemolysans (G.h; intermediate growth), S. mutans (S.m; intermediate growth), and A. oris (A.o; did not grow). The numbersto the left of the gel are kDa.

Zhou et al.

5254 aem.asm.org September 2016 Volume 82 Number 17Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

suggest that species-specific N-glycan foraging of PRG occursduring the growth of oral streptococci and that the abundantprotein PRG is an important nutrient in whole stimulated sa-liva that is biologically available to single bacterial species.

While PRG are readily identified from parotid saliva (32),this group of closely related proteins (33) has generally notbeen identified as a prominent component of resting wholesaliva (26, 34). In the present study, PRG in stimulated wholesaliva were stable only when the saliva was heat treated imme-

diately after collection (Fig. 1), which reinforces the conceptthat degradation of these and other proteins occurs readilythrough the action of hydrolytic enzymes from oral bacteria.This degradation, like that seen for different mucins (35, 36), islikewise consistent with an important role for these PRG assubstrates for bacterial growth. While the present findings as-sociate growth of S. gordonii, S. oralis, and S. mitis in saliva withglycan foraging of PRG, it is important to note that growth tointermediate cell densities of G. haemolysans, S. mutans, G. el-egans, and one strain of S. sanguinis (Fig. 2) was not associatedwith detectable changes in PRG or other salivary components.

FIG 4 2-D electrophoresis of uninoculated (Con; control) saliva and salivafrom secondary cultures of S. gordonii DL1, S. oralis C104, or S. mitis SK137.Bacterial growth altered the migration and staining of the highly basic PRGproteins (regions boxed with solid lines, with the horizontal line indicatingmolecular mass). Bands seen in 1-D electrophoresis in the 35- to 55-kDa re-gion of S. oralis and S. mitis saliva cultures in Fig. 3 are focused in this high-pIregion. Loss of 25-kDa (dotted boxes) and 15-kDa (dashed circles) proteinswas seen in cultures of S. gordonii DL1 but not for other species. The num-bers to the left of the gel are kDa.

FIG 5 Silver-stained, narrow-range 2-D electrophoretic profiles of controlsaliva and of saliva incubated with selected bacteria. PRG isoforms at pI 9 to10 in saliva incubated with S. oralis (box) are in the same molecular-massrange as bands and spots in boxed regions of Fig. 3 and 4. They are absentin saliva incubated with S. gordonii but present to a lesser degree in salivaincubated with S. mitis. The spot in the ellipses is degraded only by S.gordonii (cf. dashed circle in Fig. 4). Degradation of acidic proline-richprotein by S. gordonii is seen here as a loss of the distinct ovoid spot at thearrowheads (cf. dotted boxes in Fig. 3 and 4). The numbers to the left of thegel are kDa.

Growth of Oral Bacteria in Saliva

September 2016 Volume 82 Number 17 aem.asm.org 5255Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

Further studies on these oral species are essential to an under-standing of salivary growth substrates, including the mucins.

In view of the generally similar sugar utilization patternsbetween different streptococci (Fig. 7), we wondered whethergrowth in saliva might depend on specific cell wall-anchoredglycoside hydrolases (GHs) that act on host glycans. To assessthis possibility, we used the Carbohydrate-Active Enzyme Da-tabase (CAZy; http://www.cazy.org/) (37) to compare the GHprofiles of relevant oral species and PROSITE (http://prosite.expasy.org) to identify the LPXTG cell wall-anchor motif inGH amino acid sequences. Interestingly, cell wall-anchored�-galactosidase in GH family 2 (GH2), �-N-acetylhexosamini-dase (GH20), and endo-�-N-acetylglucosaminidase (GH85)are predicted for S. gordonii strains Challis and KCOM 1506. In

addition to these GHs, cell wall-anchored sialidase (GH33),N-acetyl �-glucosaminidase (GH84), endo-�-N-acetylgalac-tosaminidase (GH101), and two �-L-fucosidases (GH29 andGH95) are predicted for S. oralis strains Uo5 and ATCC 35037.In our study, all S. gordonii and S. oralis strains grew well in saliva(Fig. 2). S. mitis strains also grew well, and cell wall-anchoredGH33, GH84, and GH85 are predicted for S. mitis strains B6 andKCOM 1350. Furthermore, strain B6 also has cell wall-an-chored GH2, GH20, and GH101. In addition to the GHs pres-ent in B6 and KCOM 1350, S. mitis strain SVGS_061 has fourmore cell wall-anchored GHs. In contrast, no cell wall-an-chored GHs are predicted for S. sobrinus strain TCI-107, S. ves-tibularis strain F0396 (B. Henrissatt, personal communication),Actinomyces oris strain T14v, Rothia dentocariosa ATCC 17931,and numerous S. salivarius strains. These species all grew poorlyin our experiments. In recent studies to assess the role of cellwall-anchored GHs in salivary glycan foraging (27), we deletedgenes for the three cell wall-anchored GHs of S. gordonii DL1,which are homologous to pneumococcal �-galactosidase(BgaA), �-N-acetylglucosaminidase (StrH), and endohexo-saminidase D (EndoD). Importantly, the loss of these enzymesresulted in a 20-fold reduction in the growth of DL1 in saliva andabolished the degradation of PRG.

However, cell wall-anchored GHs are not predicted for sev-eral strains of S. mutans and for Gemella haemolysans strain 10379(B. Henrissatt, personal communication), two species that grew tointermediate cell densities in our study. From the opposite per-spective, cell wall-anchored GH20, GH29, and GH85 are pre-dicted for S. parasanguinis strains ATCC 15912 and FW213, andthe latter strain also has cell wall-anchored GH2. However, thesestrains grew poorly in saliva. The only S. sanguinis strain in theCAZy database is SK36; it lacks cell wall-anchored GHs and it grewpoorly, as did S. sanguinis strains SK112 and SK45. Yet, quiteunexpectedly, S. sanguinis strain 10556 grew well. Clearly, addi-tional studies are needed to define the roles of different cell surfaceGHs in bacterial growth on saliva. Variation in GH complementsmay be important for the fitness of specific strains (38) withinnascent biofilm communities through cross-feeding by closely as-sociated cells.

The failure of Actinomyces spp. to grow in saliva (Fig. 2) isconsistent with the possibility that the growth of these bacteria indental plaque depends on interactions with other species. Indeed,enhanced growth of both A. naeslundii strain T14V and S. oralisstrain 34 was noted when these strains were cocultured as a biofilmin 25% saliva (8). In the present study, a similar enhancement ofgrowth was not observed in planktonic cocultures of thesestrains or of several other strain pairs (results not shown). De-spite these results, certain observations made in the presentstudy are compatible with the concept that salivary glycan for-aging is linked with cross-feeding between these species. TheActinomyces spp. do not encode putative �-L-fucosidases, but theygrow on L-fucose (Fig. 7), whereas S. oralis produces two cellwall-anchored �-L-fucosidases but does not grow on L-fucose.Moreover, S. oralis lacks genes for L-fucose utilization like thoseidentified in S. pneumoniae (39). Thus, cleavage of L-fucosefrom salivary glycoproteins by S. oralis may well promote thegrowth of A. oris when these organisms are closely associated indental plaque biofilms.

FIG 6 Lectin blots of control saliva (Con) and of saliva from secondary cul-tures of selected strains (abbreviations are defined in the legend to Fig. 3)probed with the following FITC-labeled lectins: N-acetylneuraminic acid-spe-cific SNA, galactose-specific ECA, N-acetyl-D-glucosamine-specific SWGA,mannose-specific GNA, and L-fucose-specific AAL. The numbers to the left ofthe gel are kDa.

Zhou et al.

5256 aem.asm.org September 2016 Volume 82 Number 17Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

ACKNOWLEDGMENT

We thank Bernard Henrissat for the identification of putative GHs from S.vestibularis F0396 and G. haemolysans ATCC 10379.

FUNDING INFORMATIONThis work was supported by the Intramural Research Program of theNIDCR, NIH. Y.Z. received support from the State Key Laboratory ofOral Diseases.

REFERENCES1. Scannapieco FA. 1994. Saliva-bacterium interactions in oral microbial

ecology. Crit Rev Oral Biol Med 5:203–248.2. Yao Y, Berg EA, Costello CE, Troxler RF, Oppenheim FG. 2003.

Identification of protein components in human acquired enamel pellicleand whole saliva using novel proteomics approaches. J Biol Chem 278:5300 –5308. http://dx.doi.org/10.1074/jbc.M206333200.

3. Kolenbrander PE, Ganeshkumar N, Cassels FJ, Hughes CV. 1993.Coaggregation: specific adherence among human oral plaque bacteria.FASEB J 7:406 – 413.

4. Brecx M, Theilade J, Attström R. 1983. An ultrastructural quantitativestudy of the significance of microbial multiplication during early dentalplaque growth. J Periodont Res 18:177–186. http://dx.doi.org/10.1111/j.1600-0765.1983.tb00351.x.

5. Jakubovics NS. 2015. Saliva as the sole nutritional source in the develop-ment of multispecies communities in dental plaque. Microbiol Spectr 3:1–11. http://dx.doi.org/10.1128/microbiolspec.MBP-0013-2014.

6. De Jong MH, Van der Hoeven JS. 1987. The growth of oral bacteria on saliva. JDent Res 66:498–505. http://dx.doi.org/10.1177/00220345870660021901.

7. de Jong MH, van der Hoeven JS, van Os JH, Olijve JH. 1984. Growth oforal Streptococcus species and Actinomyces viscosus in human saliva. ApplEnviron Microbiol 47:901–904.

8. Palmer RJ, Jr, Kazmerzak K, Hansen MC, Kolenbrander PE. 2001.Mutualism versus independence: strategies of mixed-species oral biofilmsin vitro using saliva as the sole nutrient source. Infect Immun 69:5794 –5804. http://dx.doi.org/10.1128/IAI.69.9.5794-5804.2001.

9. Wickström C, Svensäter G. 2008. Salivary gel-forming mucinMUC5B—a nutrient for dental plaque bacteria. Oral Microbiol Immunol23:177–182. http://dx.doi.org/10.1111/j.1399-302X.2007.00407.x.

10. van der Hoeven JS, van den Kieboom CW, Camp PJ. 1990. Utilizationof mucin by oral Streptococcus species. Antonie Van Leeuwenhoek 57:165–172. http://dx.doi.org/10.1007/BF00403951.

11. Byers HL, Tarelli E, Homer KA, Hambley H, Beighton D. 1999. Growthof viridans streptococci on human serum alpha1-acid glycoprotein. J DentRes 78:1370 –1380. http://dx.doi.org/10.1177/00220345990780071201.

12. Homer KA, Whiley RA, Beighton D. 1994. Production of specific glyco-sidase activities by Streptococcus intermedius strain UNS35 grown in thepresence of mucin. J Med Microbiol 41:184 –190. http://dx.doi.org/10.1099/00222615-41-3-184.

13. Beighton D, Smith K, Glenister DA, Salamon K, Keevil CW. 1988.Increased degradative enzyme production by dental plaque bacteria inmucin-limited continuous culture. Microb Ecol Health Dis 1:85–94. http://dx.doi.org/10.3109/08910608809140186.

14. Byers HL, Tarelli E, Homer KA, Beighton D. 1999. Sequential deglyco-sylation and utilization of the N-linked, complex-type glycans of humanalpha1-acid glycoprotein mediates growth of Streptococcus oralis. Glyco-biology 9:469 – 479. http://dx.doi.org/10.1093/glycob/9.5.469.

15. Bradshaw DJ, Homer KA, Marsh PD, Beighton D. 1994. Metaboliccooperation in oral microbial communities during growth on mucin. Mi-crobiology 140(Pt 12):3407–3412. http://dx.doi.org/10.1099/13500872-140-12-3407.

16. Cisar JO, Kolenbrander PE, McIntire FC. 1979. Specificity of coaggre-gation reactions between human oral streptococci and strains of Actino-myces viscosus or Actinomyces naeslundii. Infect Immun 24:742–752.

FIG 7 Growth of oral bacteria in minimal medium supplemented with different host sugars. Color-coded bars show control-corrected optical densities (in Klettunits) of overnight cultures supplemented with different individual sugars. In this stacked bar chart, the length of each colored bar from left to right is the Klettunit value for growth on that sugar; i.e., only the bars for glucose begin on the value 0. All other bar lengths can be compared with that of the glucose bar, or, foran absolute value, each bar must be transposed to the zero value on the x axis.

Growth of Oral Bacteria in Saliva

September 2016 Volume 82 Number 17 aem.asm.org 5257Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from

17. Hsu SD, Cisar JO, Sandberg AL, Kilian M. 1994. Adhesive properties ofviridans group streptococcal species. Microb Ecol Health Dis 7:125–137.http://dx.doi.org/10.3109/08910609409141342.

18. Kilian M, Mikkelsen L, Henrichsen J. 1989. Taxonomic study of viridansstreptococci: description of Streptococcus gordonii sp. nov. and emendeddescriptions of Streptococcus sanguis (White and Niven, 1946), Streptococ-cus oralis (Bridge and Sneath, 1982), Streptococcus mitis (Andrewes andHorder, 1906). Int J Syst Bacteriol 39:471– 484. http://dx.doi.org/10.1099/00207713-39-4-471.

19. Koo H, Xiao J, Klein MI, Jeon JG. 2010. Exopolysaccharides produced byStreptococcus mutans glucosyltransferases modulate the establishment ofmicrocolonies within multispecies biofilms. J Bacteriol 192:3024 –3032.http://dx.doi.org/10.1128/JB.01649-09.

20. Reichmann P, Nuhn M, Denapaite D, Bruckner R, Henrich B, MaurerP, Rieger M, Klages S, Reinhard R, Hakenbeck R. 2011. Genome ofStreptococcus oralis strain Uo5. J Bacteriol 193:2888 –2889. http://dx.doi.org/10.1128/JB.00321-11.

21. Fives-Taylor PM, Thompson DW. 1985. Surface properties of Strepto-coccus sanguis FW213 mutants nonadherent to saliva-coated hydroxyap-atite. Infect Immun 47:752–759.

22. Shellis RP. 1978. A synthetic saliva for cultural studies of dental plaque. ArchOral Biol 23:485–489. http://dx.doi.org/10.1016/0003-9969(78)90081-X.

23. Chalmers NI, Palmer RJ, Jr, Cisar JO, Kolenbrander PE. 2008. Char-acterization of a Streptococcus sp.-Veillonella sp. community microma-nipulated from dental plaque. J Bacteriol 190:8145– 8154. http://dx.doi.org/10.1128/JB.00983-08.

24. Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. 1956. Color-imetric method for determination of sugars and related substances. AnalChem 28:350 –356. http://dx.doi.org/10.1021/ac60111a017.

25. Schipper RG, Silletti E, Vingerhoeds MH. 2007. Saliva as research ma-terial: biochemical, physicochemical and practical aspects. Arch Oral Biol52:1114 –1135. http://dx.doi.org/10.1016/j.archoralbio.2007.06.009.

26. Walz A, Stühler K, Wattenberg A, Hawranke E, Meyer HE, Schmalz G,Blüggel M, Ruhl S. 2006. Proteome analysis of glandular parotid andsubmandibular-sublingual saliva in comparison to whole human saliva bytwo-dimensional gel electrophoresis. Proteomics 6:1631–1639. http://dx.doi.org/10.1002/pmic.200500125.

27. Yang J, Zhou Y, Zhang L, Shah N, Jin C, Palmer RJ, Jr, Cisar JO. 2016.Cell surface glycoside hydrolases of Streptococcus gordonii promote growthin saliva. Appl Environ Microbiol 82:5278 –5286. http://dx.doi.org/10.1128/AEM.01291-16.

28. Blixt O, Head S, Mondala T, Scanlan C, Huflejt ME, Alvarez R, BryanMC, Fazio F, Calarese D, Stevens J, Razi N, Stevens DJ, Skehel JJ, vanDie I, Burton DR, Wilson IA, Cummings R, Bovin N, Wong CH,Paulson JC. 2004. Printed covalent glycan array for ligand profiling of

diverse glycan binding proteins. Proc Natl Acad Sci U S A 101:17033–17038. http://dx.doi.org/10.1073/pnas.0407902101.

29. Teneberg S, Angström J, Jovall PA, Karlsson KA. 1994. Characterizationof binding of Gal beta 4GlcNAc-specific lectins from Erythrina cristagalliand Erythrina corallodendron to glycosphinogolipids. Detection, isolation,and characterization of a novel glycosphinglipid of bovine buttermilk. JBiol Chem 269:8554 – 8563.

30. Robinson R, Kauffman DL, Waye MM, Blum M, Bennick A, Keller PJ.1989. Primary structure and possible origin of the non-glycosylated basicproline-rich protein of human submandibular/sublingual saliva. BiochemJ 263:497–503. http://dx.doi.org/10.1042/bj2630497.

31. Gillece-Castro BL, Prakobphol A, Burlingame AL, Leffler H, Fisher SJ.1991. Structure and bacterial receptor activity of a human salivary proline-rich glycoprotein. J Biol Chem 266:17358 –17368.

32. Hardt M, Thomas LR, Dixon SE, Newport G, Agabian N, Prakob-phol A, Hall SC, Witkowska HE, Fisher SJ. 2005. Toward defining thehuman parotid gland salivary proteome and peptidome: identificationand characterization using 2D SDS-PAGE, ultrafiltration, HPLC, andmass spectrometry. Biochemistry 44:2885–2899. http://dx.doi.org/10.1021/bi048176r.

33. Kim HS, Lyons KM, Saitoh E, Azen EA, Smithies O, Maeda N. 1993. Thestructure and evolution of the human salivary proline-rich protein gene fam-ily. Mamm Genome 4:3–14. http://dx.doi.org/10.1007/BF00364656.

34. Hu S, Xie Y, Ramachandran P, Ogorzalek Loo RR, Li Y, Loo JA, WongDT. 2005. Large-scale identification of proteins in human salivary pro-teome by liquid chromatography/mass spectrometry and two-dimensional gel electrophoresis-mass spectrometry. Proteomics 5:1714 –1728. http://dx.doi.org/10.1002/pmic.200401037.

35. Takehara S, Yanagishita M, Podyma-Inoue KA, Kawaguchi Y. 2013.Degradation of MUC7 and MUC5B in human saliva. PLoS One 8:e69059.http://dx.doi.org/10.1371/journal.pone.0069059.

36. Wickström C, Herzberg MC, Beighton D, Svensäter G. 2009. Proteolyticdegradation of human salivary MUC5B by dental biofilms. Microbiology155:2866 –2872. http://dx.doi.org/10.1099/mic.0.030536-0.

37. Lombard V, Golaconda Ramulu H, Drula E, Coutinho PM, HenrissatB. 2014. The carbohydrate-active enzymes database (CAZy) in 2013. Nu-cleic Acids Res 42:D490 –D495. http://dx.doi.org/10.1093/nar/gkt1178.

38. Rakoff-Nahoum S, Foster KR, Comstock LE. 2016. The evolution ofcooperation within the gut microbiota. Nature 533:255–259. http://dx.doi.org/10.1038/nature17626.

39. Higgins MA, Whitworth GE, El Warry N, Randriantsoa M, Samain E,Burke RD, Vocadlo DJ, Boraston AB. 2009. Differential recognition andhydrolysis of host carbohydrate antigens by Streptococcus pneumoniaefamily 98 glycoside hydrolases. J Biol Chem 284:26161–26173. http://dx.doi.org/10.1074/jbc.M109.024067.

Zhou et al.

5258 aem.asm.org September 2016 Volume 82 Number 17Applied and Environmental Microbiology

on February 10, 2021 by guest

http://aem.asm

.org/D

ownloaded from