ENVIRONMENTAL MICROBIOLOGY

The Influence of Iron Availability on Human SalivaryMicrobial Community Composition

Renke Wang & Aida Kaplan & Lihong Guo &

Wenyuan Shi & Xuedong Zhou & Renate Lux &

Xuesong He

Received: 7 December 2011 /Accepted: 17 January 2012 /Published online: 9 February 2012# Springer Science+Business Media, LLC 2012

Abstract It is a well-recognized fact that the compositionof human salivary microbial community is greatly affectedby its nutritional environment. However, most studies arecurrently focused on major carbon or nitrogen sources withlimited attention to trace elements like essential mineralions. In this study, we examined the effect of iron availabil-ity on the bacterial profiles of an in vitro human salivarymicrobial community as iron is an essential trace elementfor the survival and proliferation of virtually all microorgan-isms. Analysis via a combination of PCR with denaturinggradient gel electrophoresis demonstrated a drastic changein species composition of an in vitro human salivary micro-biota when iron was scavenged from the culture medium byaddition of the iron chelator 2,2′-bipyridyl. This shift incommunity profile was prevented by the presence of exces-sive ferrous iron (Fe2+). Most interestingly, under iron defi-ciency, the in vitro grown salivary microbial communitybecame dominated by several hemolytic bacterial species,including Streptococcus spp., Gemella spp., and Granulica-tella spp. all of which have been implicated in infectiveendocarditis. These data provide evidence that iron avail-ability can modulate host-associated oral microbial commu-nities, resulting in a microbiota with potential clinicalimpact.

Introduction

Microorganisms require a variety of trace elements for theirnormal metabolic activities [1, 2]. One of the most essentialtrace elements for virtually all bacteria is iron which partic-ipates in many central biological processes, such as respira-tion, the tricarboxylic acid cycle, oxygen transport,photosynthesis, and DNA biosynthesis [1]. The bacterialuptake of iron includes its physiologically relevant species,Fe2+ (ferrous iron) and Fe3+ (ferric iron). While Fe2+ issoluble under anaerobic condition and can be transportednon-specifically by bacterial divalent metal ABC transport-ers, Fe3+ (the most predominant form of iron in naturalenvironments) is not readily bio-available due to its hydro-lysis and polymerization into insoluble forms [1, 3]. Theextremely low solubility (10−18 M at pH 7) of Fe3+ makesiron one of the growth-limiting factors within many ecolog-ical niches [1, 3–5]. This is particularly true for microbialniches within the human host where availability of iron tohost-associated microbes on mucosal surfaces is furtherrestricted due to sequestration by host iron-binding proteins[6–8]. As an effective host defensive mechanism, humanbody maintains an iron-restricted environment where mostextracellular iron is bound to iron-withholding glycopro-teins such as transferrin and lactoferrin, while intracellulariron is sequestered by heme or ferritin compounds [6, 7, 9].The concentration of free iron in living tissues has beenestimated to be only 10−18 M, far below the concentrationrequired for bacterial growth [7].

The oral cavity is home to several hundred differentspecies of bacteria [10–13]. These microorganisms forman organized multispecies community with intricate inter-actions which include competing for nutrients between thebiofilm residents [14, 15]. The sources of nutrients for oralmicrobial flora include saliva, crevicular fluid, and host diet.

Renke Wang and Aida Kaplan contributed equally to the work.

A. Kaplan : L. Guo :W. Shi : R. Lux :X. He (*)UCLA School of Dentistry,10833 Le Conte Avenue, CHS 20-118,Los Angeles, CA 90095-1668, USAe-mail: xhe@ucla.edu

R. Wang :X. ZhouState Key Laboratory of Oral Diseases,West China School of Stomatology, Sichuan University,Chengdu 610041, China

Microb Ecol (2012) 64:152–161DOI 10.1007/s00248-012-0013-2

Saliva which is considered the main nutrient source dueto its chemical composition and continuous production[16] has a very limited supply of free iron with most ofthe iron bound to lactoferrin, a host-produced transferrinprotein [17, 18]. To cope with the limited availability offree iron within the oral cavity, the resident bacteriaemploy various mechanisms to close the gap betweenFe3+ solubility and their iron requirements, by eitherreducing Fe3+ to the soluble Fe2+ via a membraneassociated ferric reductase activity [19], producing side-rophores as high-affinity ferric chelators under iron-limiting conditions [20, 21], directly acquiring and uti-lizing human transferrin [22], or acquiring iron byimporting heme via heme-binding protein, e.g., HmuRin Porphyromonas gingivalis [23]. The virtually univer-sal demands for iron and its limited bio-availability,together with the differential capabilities of microbesto acquire iron and persist under iron-deficient condi-tions, are likely to make iron one of the determiningnutritional factors in modulating the composition of oralmicrobiota. In this study, we sought to investigate theeffect of iron availability on the overall microbial pro-files of in vitro human salivary bacterial communitiesand characterize the dominant bacterial species selectedunder free iron-deficient condition.

Materials and Methods

Saliva Collection

Saliva samples were collected from six healthy subjects,age 25–35 under UCLA-IRB #09-08-068-02A. Nonehas been taking any prescription or non-prescriptionmedications or being treated for any systemic disease.Subjects were asked to refrain from any food or drink2 h before donating saliva. Five-milliliter un-stimulatedsaliva was obtained from each subject by having themspit directly into a saliva collection tube; 2.5 ml of eachindividual saliva sample was pooled together andreferred to as pooled saliva, while the rest of eachsaliva sample was referred to as individual saliva andused throughout this study.

Culturing Saliva-Derived Microbial Flora and BacterialIsolation

Two hundred microliters of pooled or individual saliva wasinoculated into 2 ml of SHI medium [24] or artificial salivasolution (ASS) defined medium [25]. Different concentra-tions of the iron chelator, 2,2′-bipyridyl (Bipy) (FisherScientific, Morris Plains, NJ, USA) and/or Fe2+ (Fe2+ solu-tion was made fresh by dissolving FeSO4 in reduced

ddH2O, followed by filter sterilization) were added to themedia prior to inoculation of the saliva samples. Cul-tures were incubated under anaerobic conditions (85%nitrogen, 5% carbon dioxide, and 10% hydrogen) for18 h at 37°C. One milliliter of bacterial sample fromeach culture was taken, and bacterial cells were collected bycentrifugation at 14,000×g for 3 min and used for total DNAextraction.

For isolation of individual bacterial species, theOD600 of the overnight culture was adjusted to 0.1,followed by serial dilution with PBS buffer and platingonto SHI medium agar plates. The plates were incubat-ed under anaerobic conditions for 3 days at 37°C toallow development of single colonies. Based on theirdifference in color and morphology, colonies werepicked and grown in SHI medium. Cells from eachisolate were collected, and stocks were made in SHImedium containing 25% glycerol and stored at −80°Cfor further studies.

PCR-DGGE Analysis

Total genomic DNA of bacterial samples was isolated usingthe MasterPure™ DNA Purification Kit (EPICENTRE), andDNA quality and quantity were measured by a spectropho-tometer at 260 and 280 nm (Spectronic Genesys™, SpectronicInstruments, Inc., Rochester, NY, USA). Amplification ofbacterial 16S ribosomal RNA genes by PCR was carried outas described previously [26]. Briefly, the universal primer set,Bac1 (5′-CGC CCG CCG CGC CCC GCG CCC GTC CCGCCG CCC CCG CCC GAC TAC GTG CCA GCA GCC-3′)[27] and Bac2 (5′-GGACTACCAGGGTATCTAATCC-3′)was used to amplify an approximately 300-bp internalfragment of the 16s ribosomal RNA gene. Each 50-μlPCR contains 100 ng purified genomic DNA, 40 pmolof each primer, 200 μM of each dNTP, 4.0 mM MgCl2,5 μl 10× PCR buffer, and 2.5 U Taq DNA polymerase(Invitrogen). Cycling conditions were 94°C for 3 min,followed by 30 cycles of 94°C for 1 min, 56°C for1 min, and 72°C for 1 min, with a final extensionperiod of 5 min at 72°C. The resulting PCR productswere evaluated by 1% agarose gel electrophoresis.

Polyacrylamide gels at an 8% concentration were preparedwith a denaturing urea/formamide gradient between 40%[containing 2.8 M urea and 16% (vol/vol) formamide] and70% [containing 4.9 M urea and 28% (vol/vol) formamide].Approximately 300 ng of the PCR product was applied perwell. The gels were submerged in 1× TAE buffer (40 mMTrisbase, 40 mM glacial acetic acid, 1 mM ethylenediaminetetra-acetic acid), and the PCR products were separated by electro-phoresis for 17 h at 58°C using a fixed voltage of 60 V in theBio-Rad DCode System (Bio-Rad Laboratories, Inc.,Hercules, CA, USA). After electrophoresis, gels were rinsed

Iron Availability Affects Human Salivary Microbiota 153

and stained for 15 min in 1× TAE buffer containing 0.5 μg/mlethidium bromide, followed by 10 min of de-staining in 1×TAE buffer. Denaturing gradient gel electrophoresis (DGGE)profile images were digitally recorded using the MolecularImager Gel Documentation system (Bio-Rad Laboratories).

Identification of Bacterial Isolates and Major BacterialSpecies Within Salivary Microbial Communitieswith Different Treatments

For identification of bacterial isolates, the universal bacterial16S rDNA primer pair, 27F (5′-AGA GTT TGATCM TGGCTC AG-3′) and 1492R (5′-TAC GGY TAC C TT GTTACG ACT T-3′), was used to generate an approximately1,500-bp amplicon. Each 50-μl PCR reaction mixturecontained 20 ng of genomic DNA, 200 μM of each dNTP,4.0 mM MgCl2, 100 nM of each primer, 5 μl of 10× PCRbuffer, and 2.5 U of Taq polymerase (Invitrogen). PCRconditions were as follows: 3 min at 94°C for initial dena-turation and 27 cycles of 94°C for 1 min, 50°C for 1 min,and 72°C for 2 min and a final chain elongation at 72°C for5 min. PCR products were purified using the QIAquick PCRpurification kit (Qiagen).

To identify major bacterial species within salivarycommunities, PCR bands were excised from the DGGEgel, eluted into 20-μl sterile dH2O as preciouslydescribed [28], and re-amplified with the Bac1/Bac2universal primers. The resulting PCR products werepurified and sequenced at the UCLA sequencing andgenotyping core facility. The obtained partial 16S rRNAgene sequences, as well as the 1,500-bp amplicons forbacterial single isolates, were used to BLAST searchagainst the HOMD (http://www.homd.org) and NCBI(http://www.ncbi.nlm.nih.gov) databases. Sequences with98% to 100% identity to those deposited in the publicdomain databases were considered to be positive identi-fication of taxa.

Hemolysis Assay

For examination of the hemolytic phenotype, salivarymicrobial communities or individual isolates grown in SHImedium supplemented with or without iron chelator wereharvested and serially diluted (10-fold) from 100 to 10−4.Ten microliters of each dilution from each sample wasspotted onto SHI medium agar plates and grown for 3 daysunder anaerobic conditions at 37°C. Hemolytic activity wasdetermined as follows: α-Hemolysis presents with darken-ing and greenish coloration of the agar under the colonies;β-hemolysis shows lightened and transparent circles underthe colonies, while lack of hemolytic activity, also known asγ-hemolysis, results in no change of the agar coloration[29].

Results

The Iron Chelating Compound 2, 2′-Bipyridyl Induceda Shift in Bacterial Profile of Salivary MicrobialCommunity Cultivated in SHI Medium

The effect of iron availability on the composition of anoral bacterial community was investigated by titratingfree iron in the bacterial growth medium via the addi-tion of the iron chelator Bipy. Human pooled saliva wasinoculated into SHI medium containing 5% sheep bloodwhich was previously described as being able to sustainthe growth of a diverse in vitro microbial communitysimilar to the original salivary microbiota [24] and apanel of increasing Bipy concentrations was added toremove free iron from the medium. PCR-DGGE analy-sis revealed a distinct Bipy-induced shift in the overallmicrobial profile of the salivary community after over-night growth (Fig. 1). In the absence of Bipy, themicrobial community was comprised of a variety ofbacterial species, including Fusobacterium periodonti-cum, Neisseria subflava, Porphyromonas spp., Campylo-bacter spp., and Streptococcus spp., while the additionof the iron chelator induced a pronounced shift in thecommunity composition at an apparent threshold valueof >0.2 mM with Streptococcus spp., Gemella spp., andGranulicatella spp. becoming the dominant species(Fig. 1).

To rule out the possibility that the observed Bipy-inducedcommunity shift was due to toxic effects of the compoundrather than lowering the free iron concentration in themedium, Bipy was added to cultures supplemented with0.1 mM free iron (Fe2+) and compared to unsupplementedcultures (Fig. 2). Consistent with our previous observation(Fig. 1), 0.3 mM Bipy alone caused a drastic communityshift. However, addition of 0.1 mM (Bipy binds Fe2+ in a3:1 stoichiometry) free iron ions (Fe2+) eliminated the Bipy-induced community shift, and the microbial profileremained similar to the original sample. The addition of0.1 mM Fe2+ alone did not have a significant effect oncommunity composition. Taken together, these results con-firmed that the observed bacterial community shift wasindeed due to Bipy-based reduction in the iron concentrationin the medium and not other effects of the chelatingcompound.

Bipy Induced Comparable Profile Shifts in SalivaryMicrobial Flora Isolated from Different HealthyHuman Subjects

Next, we investigated if the Bipy-induced shift in the sali-vary microbial community composition is a general phe-nomenon by examining its effect on the individual salivary

154 R. Wang et al.

microbiota collected from different healthy subjects. Commu-nity shifts similar to the one revealed for pooled saliva wasobserved in all individual samples (Fig. 3). More importantly,all the shifts displayed a similar trend, with Streptococcusspp., Granulicatella spp., and Gemella spp. being the mostprevalent bacteria in the presence of high concentration of ironchelator. Interestingly, for the saliva sample from subject 3, theaddition of Bipy also resulted in an increase in the populationof Abiotrophia defectiva, often known as nutritionally variantstreptococcus (NVS) or Streptococcus defectivus, a causativeagent of infective endocarditis.

The Availability of Free Iron Ions (Fe2+) Modulatedthe Microbial Profile of Salivary Bacterial Community

To further investigate the modulating effect of availability offree iron on the salivary microbial community, we generateda mixed microbial community containing bacterial popula-tions cultivated from iron-limiting as well as non-iron-limiting conditions. The resulting community with a mixedbacterial profile was then re-cultivated under either iron-depleted or repleted condition, and the modulating effectof iron accessibility on the microbial population profile wasevaluated by monitoring the microbial population shift.Specifically, we inoculated pooled saliva samples in SHImedium in the absence and presence of Bipy at a final

Fusobacterium periodonticum

Porphyromonas spp.

Porphyromonascatoniae

Neisseria subflava

Streptococcus spp.

Granulicatella spp.

Streptococcus cristatus

Granulicatellaadiacens

Bipy

Prevotella spp.

In vitro salivary communities cultivated in SHI medium

Haemophilus spp.Peptostreptococcus

stomatis

Campylobacter concisus

Streptococcus spp.

Gemella sanguinis/ G. haemolysans

0 1uM 10uM 50 uM 0.1mM 0.2 uM 0.3 mM 1 mM 5 mM 10 mM

Figure 1 PCR-DGGE analysis of microbial profiles of pooled salivacultivated in SHI medium in the presence of the iron chelator 2,2′bipyridyl (Bipy). Each lane represents pooled saliva cultured in SHImedium containing the specified amounts of 2,2′-bipyridyl. Thearrows indicate the dominant bands whose corresponding bacterialID was determined at species level when their sequenced 16S

fragments have more than 98% identity to those deposited in thedatabase. For those that have multiple hits with same identity, bothbacterial species are listed and separated with “/”, while “spp.” wasused when their sequences have less than 98% identity to that depos-ited in the database. Two biological replicates were performed for eachexperiment and a representative gel image is shown

Bipy

Fe2+

- - + +- + - +

In vitro salivary communities

Figure 2 PCR-DGGE analysis of microbial profiles of pooled salivacultivated in SHI medium supplemented with or without Bipy(0.3 mM) and/or Fe2+ (0.1 mM). Addition of Bipy or Fe2+ is indicatedby (+), while (−) represents samples without exogenously added Bipyor Fe2+. Two biological replicates were performed and a representativegel image is shown

Iron Availability Affects Human Salivary Microbiota 155

concentration of 0.3 mM which has been shown to induce asignificant shift in community composition (Fig. 1) to gen-erate culture A and culture B, respectively. After overnightincubation, bacterial cells in these two cultures were collect-ed, resuspended in fresh SHI medium to an OD600 nm of 0.2,and mixed in a 1:1 ratio to obtain co-culture (A+B). Co-cultures were supplemented either with Bipy or Fe2+ ionsand incubated further for 16 h under anaerobic condition.PCR-DGGE analysis revealed that addition of Bipy to themicrobial population present in the combined culture in-duced a shift in the community profile that resembled theBipy-treated original sample. Exogenously provided Fe2+ incontrast modulated the mixed population resulting in aprofile that was similar to the original untreated community(Fig. 4).

Bacterial Species Isolated from Medium Deficient in FreeIron Generally Displayed High Hemolytic Activity

One of the important components in SHI medium is sheepblood. It has been reported that certain oral bacterial species,such as Streptococci, are able to lyse red blood cells,retrieve, and utilize iron from heme to sustain their growth[20, 30, 31]. We speculated that the dominant speciesselected under high iron chelator concentrations in SHImedium could employ a similar strategy to obtain iron. Totest this, we cultivated saliva samples in the presence andabsence of Bipy. The overnight cultures were harvested andspotted onto blood agar plates in serial dilutions. The bac-terial community selected in the presence of the iron

chelator exhibited distinct α-hemolytic activity, while cellsharvested from cultures supplemented with iron displayedno obvious hemolytic activity, similar to those cultivated inoriginal SHI medium (Fig 5a). These results suggested that

1 2 3 4 5 6

- + - + - + - + - + - + Bipy

Fusobacteriumperiodonticum

Prevotella spp.

Neisseria subflava

Porphyromonascatoniae

Campylobacter concisus

Streptococcus spp.

Granulicatella spp.

Subject number

Abiotrophia defective (NVS)

Streptococcus cristatusStreptococcus spp.

Granulicatellaadiacens

Gemella sanguinis/ G. haemolysans

Figure 3 Effect of the iron chelator Bipy on the microbial profiles ofsalivary communities from different subjects. Individual saliva sampleswere cultivated in SHI medium in the presence and absence of exog-enously added Bipy (0.3 mM). Overnight cultures were subjected toPCR-DGGE analysis. Addition of Bipy is indicated by (+), while (−)

represents samples without exogenously added Bipy. The arrows indi-cated the dominant bands whose corresponding bacterial ID was de-termined. Two biological replicates were performed and arepresentative gel image is shown

Culture-A Co-culture (A+B)Culture-B

Bipy - + - + -Fe2+ - - - - +

1 2 3 4 5

Figure 4 The modulating effect of iron availability on the salivary mi-crobial communities. Pooled saliva samples were cultivated in the absenceand presence of Bipy (0.3 mM) overnight. The resulting cultures, culture Aand culture B, respectively, were mixed at a 1:1 ratio and further cultivatedfor 16 h with the addition of either Bipy (0.3 mM) or Fe2+ (0.1 mM).Samples were taken and subjected to PCR-DGGE analysis. Two biologicalreplicates were performed and a representative gel image is shown

156 R. Wang et al.

oral bacterial species selected under low free iron conditionsare better adapted to obtain iron from their host environ-ments. In order to further investigate this possibility, wechose three different Streptococci strains and one Gemellastrain that were able to grow in the presence of high con-centration of Bipy and tested their hemolytic activity. Allfour tested strains displayed various degree of α-hemolyticactivity (Fig. 5b, c), while species isolated from the originalSHI medium culture, such as Fusobacterium nucleatum, didnot show significant hemolytic activity under the conditionstested. More intriguingly, some of the isolated strains,including Gemella haemolysans and Streptococcus sp. oralclone C3ALM006, displayed much higher hemolytic activitywhen they were spotted onto SHI agar plates supplementedwith 0.5 mM Bipy, as indicated by the more greenish colorzone around the colonies (Fig. 5c).

Effect of Bipy on the Salivary Microbial CommunityCultivated in Chemically Defined Artificial Saliva Solution

Conditions with limited availability of free iron appear toselect for certain oral microbial species such as Streptococcicristatus and G. haemolysans that were shown to displayhemolytic activity (Fig. 5), which could potentially provideadditional iron source to sustain their growth. In addition to

their hemolytic activity, these species could also have anincreased ability to persist in iron-deficient environmentswhich would further enhance their competitiveness. To testthis hypothesis, we used ASS medium, a chemically definedmedium that was developed in our laboratory for cultivationof oral microbial flora [25]. Different concentrations of ironchelator were added to the medium, and microbial profileswere monitored by DGGE. Addition of as little as 1 μMBipy already started to cause change in community profileresulting in the loss or great reduction of Campylobacterconcisus, Veillonella atypica, and Prevotella spp. amongothers (Fig. 6). At a concentration of 50 μM, Bipy was ableto induce a drastic community shift very similar to thoseobserved in SHI medium supplemented with Bipy (Fig. 3),with G. haemolysans, Granulicatella adiacens, and Strep-tococci, including S. cristatus becoming the predominantspecies. More strikingly, at 1 mM of Bipy, a concentrationhigh enough to chelate most of the free ferrous iron(0.1 mM) in the ASS medium, several Streptococcus spp.and Gemella spp., including Streptococcus spp. oral cloneC3AKM006 and G. haemolysans which could also beselected from Bipy-supplemented SHI medium, were stillable to grow. Our data suggested that, besides being capableof lysing host red blood cells and retrieving complexed iron,these bacterial species are able to persist under iron-deficient

A

B

Bipy - + -

Fe2+ - - +

10-2

10-3

10-4

Salivary culture in SHI medium

C

SHI medium

SHI mediumwith 0.5 mM Bipy

Figure 5 Hemolytic activity of salivary communities and individualisolates under different culture conditions. a Hemolytic activity ofcultivated salivary communities: salivary cultures in SHI mediumsupplemented with Bipy (0.3 mM) or Fe2+ (0.1 mM) were dilutedand spotted onto SHI medium agar plates with 0.5% sheep blood.Plates was incubated at 37°C under anaerobic condition for 24 h.Addition of Bipy or Fe2+ is indicated by (+), while (−) representssamples without exogenously added Bipy or Fe2+. b Hemolytic activity

of salivary isolates. Salivary isolates from cultures supplemented withand without Bipy were cultivated and spotted onto solid SHI mediumwith 0.5% sheep blood. Plates were incubated at 37°C under anaerobiccondition for 24 h before hemolytic activity was recorded. c Hemolyticactivity of salivary isolates on SHI agar plates in the presence andabsence of 0.5 mM Bipy. Two biological replicates were performed foreach setup and representative images are shown

Iron Availability Affects Human Salivary Microbiota 157

conditions. It is worthwhile to note that higher concentrationof Bipy often resulted in reduced biomass (data not shown)both in SHI and ASS medium, confirming that the majorityof bacterial species require iron for growth.

Discussion

As one of the most essential trace elements for the growth ofvirtually all organisms, iron plays an important role indetermining the bio-abundance and bio-diversity within avariety of different ecosystems [32–36]. However, only alimited number of studies have evaluated the effect of ironavailability on the bacterial composition of human-associated microbiota [37]. In this study, using in vitrohuman salivary microbial communities, we investigatedwhether the accessibility of iron plays a role in modulatingthe microbial populations within the oral microbiota.

We generated culturing conditions with differential freeiron availability by adding different concentrations of Bipy,a ferrous iron (Fe2+) chelator to SHI medium containingsheep blood. Our data revealed that titration of free Fe2+

from SHI medium resulted in a drastic shift in the bacterialcomposition of an in vitro salivary microbial community,with Streptococcus spp., Gemella spp., and Granulicatellaspp. becoming the most dominant bacterial species (Fig. 1).The Bipy-induced shift in microbial composition wasobserved both in pooled and individual saliva-derived mi-crobial communities (Figs. 1 and 3). The similarity in theresultant bacterial profiles among different communities and

the seemingly abrupt shift occurring at 0.3 μM Bipy sug-gested that the addition of 0.3 μM Bipy could reduce theconcentration of free iron in the medium below a thresholdthat might be required for the optimal growth of oral com-munity. Once the free iron level is below the threshold, itcould shift the microbial population toward bacterial speciesthat might be more capable of utilizing alternative ironsources and/or more resistant to the iron starvationconditions.

Interestingly, the identified major bacterial species grow-ing under iron-limited condition are phylogenetically relatedto different branches of Streptococci [38, 39]. Gemella spp.was previously classified into the genus Streptococcus [40]and reclassified as genus Gemella later based on the nucle-otide sequence of the 16S rRNA [41]. While for Granuli-catella spp. and Abiotrophia spp. (identified from subject 3),they were originally described as “NVS,” a type of strepto-cocci exhibiting satellitism around colonies of other bacteria[42]. These bacterial species share many physiological andbiochemical properties with the viridans group streptococci,including the range of infections that they cause [43]. One ofthe common characteristics among these bacterial species istheir α-hemolytic activity [44–46], which was further con-firmed by our hemolysis assay results (Fig. 5). The SHImedium used in this study contains 5% sheep blood [24],and the abundant iron-containing heme within sheep redblood cells could potentially serve as an iron source for oralbacteria under free iron-limited condition. The α-hemolyticactivity would allow bacteria to partially lyse the red bloodcells to release iron complexes, e.g., heme and hemoglobin.

Streptococcus spp.Streptococcus spp.

Streptococcus sinensis / S. sanguinis

Haemophilus influenzae

Prevotella spp.

Veillonella atypica

In vitro salivary communities in ASS medium

Porphyromonas catoniae

Campylobacter concisus

0 1uM 50uM 100 uM 1 mM

Gemella haemolysans/ G. sanguinis

Bipy

Streptococcus spp.

Streptococcus spp.

Abiotrophia defective

Granulicatella adiacens

Streptococcus cristatus

Figure 6 PCR-DGGE analysis of microbial profiles of pooled salivacultivated in ASS medium with increasing concentrations of Bipy.Each lane represents pooled saliva cultured in ASS medium containingthe specified amounts of 2,2′-bipyridyl. The arrows indicate the

dominant bands whose corresponding bacterial ID was determinedand shown. Two biological replicates were performed and a represen-tative gel image is shown

158 R. Wang et al.

Although the mechanisms regarding heme uptake inGemella and Granulicatella spp. is currently unknown,many streptococci have been shown to employ high-affinity iron uptake systems to obtain heme [20, 47].

Intriguingly, some of the selected Streptococcus andGemella strains, including Streptococcus spp. oral cloneC3AKM006 and G. haemolysans isolated from both Bipy-supplemented ASS and SHI medium, displayed enhancedα-hemolytic activity under iron-limited condition (Fig. 5c).For many host-associated bacterial species, iron starvationcould act as an important environmental cue and inducespecific adapted response in bacteria, including the induc-tion of siderophore-mediated iron uptake systems and otherpotentially pathogenic features [1, 48]. Certain Streptococcusspp., including Streptococcus pyogenes displayed increasedhemolytic activity due to high hemolysin production whenencountering iron-limited condition [49]. Our data clearlysuggested that, during iron starvation, many oral bacterialspecies, such as G. haemolysans, G. adiacens, and certainStreptococci, including S. cristatus, might be able to sustaintheir growth by retrieving host iron complexes via hemolyticactivity.

Furthermore, when using chemically defined ASS mediumwithout the inclusion of sheep blood, the addition of Bipyresulted in a community with a profile that was very similarmicrobial to the one obtained with Bipy-supplemented SHImedium (Fig. 6). Even in the presence of Bipy at a concen-tration that would chelate most of the free Fe2+ iron within themedium, several Streptococci and Gemella species could stillbe isolated from the culture. 16s rDNA sequence analysisrevealed that their identities matched a subset of Streptococcusand Gemella strains (including Streptococcus parasanguinisand G. haemolysans) isolated from Bipy-treated SHI medium(data not shown) which had been show to possess hemolyticactivity. These results indicated that besides being capable ofacquiring host iron complexes, certain oral strains are alsohighly resistant to iron starvation conditions. Although iron isrequired for the growth of the virtually all microbes, differentbacteria might have differential iron requirement for sustain-ing their normal growth and possess different ability to persistunder iron-deficient condition [1]. Furthermore, certain bacte-rial species, including oral Lactobacilli [50], adapt their me-tabolism toward an absolute manganese requirement insteadof absolute iron requirement as a possible defense mechanismagainst endogenous superoxide during aerobic metabolism[51]. However, we did not isolate any Lactobacilli strainsfrom salivary communities cultivated using either Bipy-supplemented SHI medium or ASS medium. This could bedue to the fact that although SHI medium has been shown tosustain the growth of a variety of oral microbes, it does notseem to favor the growth of Lactobacilli [24].

The varying capability of oral bacterial species in explor-ing iron sources and acquiring iron from their environments,

as well as their differential resistance to iron starvation, arelikely to make iron, this scarce and growth-limiting traceelement, one of the modulating factors in shaping the host-associated microbial community. This is corroborated byour observation that, for a mixed community composed ofpopulations obtained from both iron-limiting and iron-containing medium, iron supplementation can restore theoriginal bacterial profile, while the addition of an iron che-lator drove the mixed culture toward a community that hadbeen shown to be specifically selected under iron-deficientcondition (Fig. 4).

The most intriguing finding of this study was that themajority of the oral bacterial species isolated under iron-deficient condition have previously been implicated in in-fectious endocarditis (IE) [52–56]. IE is an infection of thelining of the heart chambers and heart valves that is causedby bacteria or other infectious substances. Since the descrip-tion of the systemic significance of oral infection in latenineteenth century [57], every major body system has beenidentified as a potential targeting site for bacterial metastasisof oral origin [58]. The involvement of oral bacteria in thepathogenesis of endocarditis had been previously estab-lished [52, 58]. It has been estimated that 65% of IE clinicalcases were due to infection of α-hemolytic streptococci, andmajority of them were of oral origin. With the improvementof culturing technique and non-culture based detectionmethods, more and more fastidious bacteria have been iden-tified to be involved in IE, including Gemella [55, 59],Granulicatella [39, 54], as well as Abiotrophia spp. [56,60]. Normal oral functional activities like chewing, as wellas most dental procedures, such as periodontal cleaning,tooth extraction, and endodontic procedures could causetissue surface trauma and transient bacteremia [61]. In indi-viduals with heart malformations or the presence of vascularprostheses, the transient bacteremia could potentially lead toattachment of bacteria and result in IE. It is intriguing tospeculate that, for an individual with an oral microbialcommunity containing more IE-related bacterial species,such as α-hemolytic bacteria as a result of particular oralor systemic conditions, there could be a potentially higherrisk of endocarditis. Further investigation is needed to de-termine whether there is a possible association between thesaliva iron content and individual’s risk of infectiousendocarditis.

The human oral cavity represents one of the most com-plex host-associated microbial ecosystems ever identified.Like gut-associated microbial flora, the microbial composi-tion within the oral habitat could be a complex polygenictrait shaped and modulated by multiple host and environ-mental factors. A healthy and balanced oral commensalflora with stable microbial composition could play an im-portant role in maintaining the community stability andpreventing foreign/pathogenic colonization, while

Iron Availability Affects Human Salivary Microbiota 159

unbalanced population structure due to local or systemicconditions could lead to the deterioration of oral ecologicalconditions and induce adverse effect on the host. Our datasuggested that, as one of the most essential trace elements,the accessibility of free iron could play an important role inmodulating the population structure of the commensal oralmicrobiota and have potential clinical relevance in inducingcertain disease conditions.

Acknowledgments The study was supported by US National Insti-tutes of Health (NIH) Grants (DE020102 and DE021108) and Interna-tional Science and Technology Cooperation Program of China(2011DFA30940).

References

1. Andrews S, Robinson A, Rodriguez-Quinones F (2003) Bacterialiron homeostasis. FEMS Microbiol Rev 27:215–237

2. Jakubovics NS, Jenkinson HF (2001) Out of the iron age: newinsights into the critical role of manganese homeostasis in bacteria.Microbiology 147:1709–1718

3. Braun V, Killmann H (1999) Bacterial solutions to the iron-supplyproblem. Trends Biochem Sci 24:104–109

4. Braun V (2001) Iron uptake mechanisms and their regulation inpathogenic bacteria. Int J Med Microbiol 291:67–79

5. Ratledge C, Dover L (2000) Iron metabolism in pathogenic bacteria.Annu Rev Microbiol 54:881–941

6. Marx JJM (2002) Iron and infection: competition between host andmicrobes for a precious element. Best Pract Res Clin Haematol15:411–426

7. Weinberg E (1978) Iron and infection. Microbiol Rev 42:45–668. Wooldridge K, Williams P (1993) Iron uptake mechanisms of

pathogenic bacteria. FEMS Microbiol Rev 12:325–3489. Finkelstein R, Sciortino C, McIntosh M (1983) Role of iron in

microbe–host interactions. Rev Infect Dis 5:S759–S77710. Aas JA, Paster BJ, Stokes LN, Olsen I, Dewhirst FE (2005)

Defining the normal bacterial flora of the oral cavity. J Clin Micro-biol 43:5721–5732

11. Dewhirst FE, Chen T, Izard J, Paster BJ, Tanner ACR, Yu W-H,Lakshmanan A, Wade WG (2010) The human oral microbiome. JBacteriol 192:5002–5017

12. Paster BJ, Boches SK, Galvin JL, Ericson RE, Lau CN, LevanosVA, Sahasrabudhe A, Dewhirst FE (2001) Bacterial diversity inhuman subgingival plaque. J Bacteriol 183:3770–3783

13. Paster BJ, Olsen I, Aas JA, Dewhirst FE (2006) The breadth ofbacterial diversity in the human periodontal pocket and other oralsites. Periodontology 2000 42:80–87

14. Kolenbrander PE, Palmer RJ, Periasamy S, Jakubovics NS (2010)Oral multispecies biofilm development and the key role of cell–cell distance. Nat Rev Micro 8:471–480

15. Kuramitsu HK, He X, Lux R, Anderson MH, Shi W (2007)Interspecies interactions within oral microbial communities.Microbiol Mol Biol Rev 71:653–670

16. van der Hoeven JS, de Jong M, Rogers AH, Camp P (1984) Aconceptual model for the co-existence of Streptococcus spp. andActinomyces spp. in dental plaque. J Dent Res 63:389–392

17. Weinberg ED (2001) Human lactoferrin: a novel therapeutic withbroad spectrum potential. J Pharm Pharmacol 53:1303–1310

18. Lönnerdal B, Iyer S (1995) Lactoferrin: molecular structure andbiological function. Annu Rev Nutr 15:93–110

19. Evans SL, Arceneaux JE, Byers BR, Martin ME, Aranha H (1986)Ferrous iron transport in Streptococcus mutans. J Bacteriol168:1096–1099

20. Ge R, Sun X, He Q (2009) Iron acquisition by Streptococcusspecies: an updated review. Frontiers Biology China 4:392–401

21. Moelling C, Oberschlacke R, Ward P, Karijolich J, Borisova K,Bjelos N, Bergeron L (2007) Metal-dependent repression of side-rophore and biofilm formation in Actinomyces naeslundii. FEMSMicrobiol Lett 275:214–220

22. Duchesne P, Grenier D, Mayrand D (1999) Binding and utilizationof human transferrin by Prevotella nigrescens. Infect Immun67:576–580

23. Liu X, Olczak T, Guo H-C, Dixon DW, Genco CA (2006) Identi-fication of amino acid residues involved in heme binding andhemoprotein utilization in the Porphyromonas gingivalis hemereceptor HmuR. Infect Immun 74:1222–1232

24. Tian Y, He X, Torralba M, Yooseph S, Nelson KE, Lux R, McLeanJS, Yu G, Shi W (2010) Using DGGE profiling to develop a novelculture medium suitable for oral microbial communities. Mol OralMicrobiol 25:357–367

25. He X, Wu C, Yarbrough D, Sim L, Niu G, Merritt J, Shi W, Qi F(2008) The cia operon of Streptococcus mutans encodes a uniquecomponent required for calcium-mediated autoregulation. MolMicrobiol 70:112–126

26. He X, Tian Y, Guo L, Lux R, Zusman D, Shi W (2010) Oral-derived bacterial flora defends its domain by recognizing andkilling intruders—a molecular analysis using Escherichia coli asa model intestinal bacterium. Microb Ecol 60:655–664

27. Sheffield VC, Cox DR, Lerman LS, Myers RM (1989) Attachmentof a 40-base-pair G + C-rich sequence (GC-clamp) to genomicDNA fragments by the polymerase chain reaction results in im-proved detection of single-base changes. Proc Natl Acad Sci U S A86:232–236

28. He X, Tian Y, Guo L, Ano T, Lux R, Zusman D, Shi W (2010) Invitro communities derived from oral and gut microbial florasinhibit the growth of bacteria of foreign origins. Microb Ecol60:665–676

29. Ray C, Ryan K (2004) In: Ryan KJ (ed) Sherris medical microbi-ology: an introduction to infectious diseases, 4th edn. Appletonand Lange Paramount Publishing Business and ProfessionalGroup, East Norwalk, p 237

30. Zhu H, Liu M, Lei B (2008) The surface protein Shr of Strepto-coccus pyogenes binds heme and transfers it to the streptococcalheme-binding protein Shp. BMC Microbiol 8:15

31. Brown JS, Gilliland SM, Holden DW (2001) A Streptococcuspneumoniae pathogenicity island encoding an ABC transporterinvolved in iron uptake and virulence. Mol Microbiol 40:572–585

32. Hopkinson BM, Barbeau K (2008) Interactive influences of ironand light limitation on phytoplankton at subsurface chlorophyllmaxima in the eastern North Pacific. Limnol Oceanogr 53:1303–1318

33. Yang C-H, Crowley DE (2000) Rhizosphere microbial communitystructure in relation to root location and plant iron nutritionalstatus. Appl Environ Microbiol 66:345–351

34. Jin CW, Li GX, Yu XH, Zheng SJ (2010) Plant Fe status affects thecomposition of siderophore-secreting microbes in the rhizosphere.Annal Bot 105:835–841

35. Hutchins D (2001) Response of marine bacterial community com-position to iron additions in three iron-limited regimes. LimnolOceanogr 46:1535–1545

36. Rose JM, Feng Y, DiTullio GR, Dunbar RB, Hare C, Lee P, LohanM, Long M, Wo SJ, Sohst B, Tozzi S, Zhang Y, Hutchins D (2009)Synergistic effects of iron and temperature on Antarctic phyto-plankton and microzooplankton assemblages. Biogeosciences6:3131

160 R. Wang et al.

37. Zimmermann MB, Chassard C, Rohner F, N’Goran EK, Nindjin C,Dostal A, Utzinger J, Ghattas H, Lacroix C, Hurrell RF (2010) Theeffects of iron fortification on the gut microbiota in African chil-dren: a randomized controlled trial in Côte d’Ivoire. Am J ClinNutr 92:1406–1415

38. Tung SK, Teng LJ, Vaneechoutte M, Chen HM, Chang TC(2007) Identification of species of Abiotrophia, Enterococcus,Granulicatella and Streptococcus by sequence analysis of theribosomal 16S–23S intergenic spacer region. J Med Microbiol56:504–513

39. Casalta JP, Habib G, La Scola B, Drancourt M, Caus T, Raoult D(2002) Molecular diagnosis of Granulicatella elegans on the car-diac valve of a patient with culture-negative endocarditis. J ClinMicrobiol 40:1845–1847

40. Holdeman LV, Moore WEC (1974) New genus, Coprococcus,twelve new species, and emended descriptions of four previouslydescribed species of bacteria from human feces. Int J Syst Bacter-iol 24:260–277

41. Berger U (1961) A proposed new genus of gram-negative cocci:Gemella. Intern Bull Bacteriol Nomen Taxon 11:17–19

42. Frenkel A, Hirsch W (1961) Spontaneous development of L formsof streptococci requiring secretions of other bacteria or sulphydrylcompounds for normal growth. Nature 191:728–730

43. Facklam R, Elliott JA (1995) Identification, classification, andclinical relevance of catalase-negative, gram-positive cocci, ex-cluding the streptococci and enterococci. Clin Microbiol Rev8:479–495

44. Woo PCY, Lau SKP, Fung AMY, Chiu SK, Yung RWH, Yuen KY(2003) Gemella bacteraemia characterised by 16S ribosomal RNAgene sequencing. J Clin Pathol 56:690–693

45. Ruoff KL (2002) Miscellaneous catalase-negative, gram-positivecocci: emerging opportunists. J Clin Microbiol 40:1129–1133

46. Haanpera M, Jalava J, Huovinen P, Meurman O, Rantakokko-Jalava K (2007) Identification of alpha-hemolytic streptococci bypyrosequencing the 16S rRNA gene and by use of VITEK 2. J ClinMicrobiol 45:762–770

47. Bates CS, Montanez GE, Woods CR, Vincent RM, Eichenbaum Z(2003) Identification and characterization of a Streptococcus pyo-genes operon involved in binding of hemoproteins and acquisitionof Iron. Infect Immun 71:1042–1055

48. Nobles CL, Maresso AW (2011) The theft of host heme by Gram-positive pathogenic bacteria. Metallomics 3:788–796

49. Griffiths BB, McClain O (1988) The role of iron in the growth andhemolysin (streptolysin S) production in Streptococcus pyogenes. JBasic Microbio 28:427–436

50. Imbert M, Blondeau R (1998) On the iron requirement of lactoba-cilli grown in chemically defined medium. Curr Microbiol 37:64–66

51. Archibald FS, Fridovich I (1981) Manganese and defenses againstoxygen toxicity in Lactobacillus plantarum. J Bacteriol 145:442–451

52. Baddour LM, Wilson WR, Bayer AS, Fowler VG, Bolger AF,Levison ME, Ferrieri P, Gerber MA, Tani LY, Gewitz MH, TongDC, Steckelberg JM, Baltimore RS, Shulman ST, Burns JC, FalaceDA, Newburger JW, Pallasch TJ, Takahashi M, Taubert KA (2005)Infective endocarditis. Circulation 111:e394–e434

53. Bouvet A (1995) Human endocarditis due to nutritionally variantstreptococci: Streptococcus adjacens and Streptococcus defectivus.Eur Heart J 16:24–27

54. Ohara-Nemoto Y, Kishi K, Satho M, Tajika S, Sasaki M, NamiokaA, Kimura S (2005) Infective endocarditis caused by Granulica-tella elegans originating in the oral cavity. J Clin Microbiol43:1405–1407

55. Stroup JS, Bransteitter BA, Reust R (2007) Infective endocarditiscaused by Gemella species. Infect Dis Clin Pract 15:203–205

56. Yerebakan C, Westphal B, Skrabal C, Kaminski A, Ugurlucan M,Bomke A-K, Liebold A, Steinhoff G (2008) Aortic valve endocar-ditis due to abiotrophia defectiva: a rare etiology. WMW WienerMedizinische Wochenschrift 158:152–155

57. Miller W (1890) The microorganisms of the human mouth. Thelocal and general diseases which are caused by them. White,Philadelphia, pp 274–341

58. Gendron R, Grenier D, Maheu-Robert L-F (2000) The oral cavityas a reservoir of bacterial pathogens for focal infections. MicrobInfect 2:897–906

59. La Scola B, Raoult D (1998) Molecular identification of Gemellaspecies from three patients with endocarditis. J Clin Microbiol36:866–871

60. Lainscak M, Lejko-Zupanc T, Strumbelj I, Gasparac I, Mueller-Premru M, Pirs M (2005) Infective endocarditis due to Abiotrophiadefectiva: a report of two cases. J Heart Valve Dis 14:33–36

61. Parahitiyawa NB, Jin LJ, Leung WK, Yam WC, Samaranayake LP(2009) Microbiology of odontogenic bacteremia: beyond endocar-ditis. Clin Microbiol Rev 22:46–64

Iron Availability Affects Human Salivary Microbiota 161