A Lactotransferrin Single Nucleotide Polymorphism Demonstrates ...

2 downloads 209 Views 888KB Size Report
Nov 8, 2012 - Groenink J, Walgreen-Weterings E, Nazmi K, Bloscher JG, Veerman. EC, van Winkelhoff AJ, Nieuw Amerongen AV. 1999. Salivary lactofer-.
A Lactotransferrin Single Nucleotide Polymorphism Demonstrates Biological Activity That Can Reduce Susceptibility to Caries Daniel H. Fine,a Gokce A. Toruner,b Kabilan Velliyagounder,a Vandana Sampathkumar,a Dipti Godboley,a David Furganga Department of Oral Biology, New Jersey Dental School, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, USAa; Department of Pediatrics, Institute of Genetic Medicine, New Jersey Medical School, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, USAb

Streptococcus mutans is prominently linked to dental caries. Saliva’s influence on caries is incompletely understood. Our goal was to identify a salivary protein with anti-S. mutans activity, characterize its genotype, and determine genotypic variants associated with S. mutans activity and reduced caries. An S. mutans affinity column was used to isolate active moieties from saliva obtained from a subject with minimal caries. The bound and eluted protein was identified as lactotransferrin (LTF) by matrixassisted laser desorption ionization–time of flight (MALDI-TOF) analysis and confirmed by Western blotting with LTF antibody. A single nucleotide polymorphism (SNP) that produced a shift from arginine (R) to lysine (K) at amino acid position 47 in the LTF antimicrobial region (rs: 1126478) killed S. mutans in vitro. Saliva from a subject with moderate caries and with the LTF “wild-type” R form at position 47 had no such activity. A pilot genetic study (n ⴝ 30) showed that KK subjects were more likely to have anti-S. mutans activity than RR subjects (P ⴝ 0.001; relative risk ⴝ 3.6; 95% confidence interval [95% CI] ⴝ 1.5 to 11.13). Pretreatment of KK saliva with antibody to LTF reduced S. mutans killing in a dose-dependent manner (P ⴝ 0.02). KK subjects were less likely to have caries (P ⴝ 0.02). A synthetic 11-mer LTF/K peptide killed S. mutans and other caries-related bacteria, while the LTF/R peptide had no effect (P ⴝ 0.01). Our results provide functional evidence that the LTF/K variant results in both anti-S. mutans activity and reduced decay. We suggest that the LTF/K variant can influence oral microbial ecology in general and caries-provoking microbes specifically.

C

aries, a pandemic infection provoked by acid-producing Gram-positive bacteria such as Streptococcus mutans, is reported to affect approximately 70% of the population and is among the most common diseases of mankind, being more common than asthma and as prevalent as the common cold (1, 2). It has long been speculated that saliva, due to its intimate contact with oral microbes, could have a profound influence on oral diseases, particularly one as prominent as dental caries (3). This speculation is also based on the fact that saliva bathes all oral mucosal and dentate surfaces, has functions that include but are not limited to buffering capacity, washing, antimicrobial activity, and healing, and, as such, should affect both microbial and host responsiveness (4, 5). However, while it has been proposed that several salivary glycoproteins, singly or in combination, can be linked to oral disease, in vivo results have been contradictory (6, 7). Along these lines, lactotransferrin (LTF) is proposed to be associated with a multitude of oral diseases, although in vivo substantiation is lacking (8, 9). LTF, a principal salivary glycoprotein, has been shown to demonstrate in vitro activity against S. mutans, hence its link to dental caries (9), and in vitro activity against Aggregatibacter actinomycetemcomitans and Porphyromonas gingivalis, suggesting its link to periodontal disease (8, 10). Several years ago, our group identified a single nucleotide polymorphism (SNP) (rs: 1126478) in the LTF gene of one individual who had an aggressive form of periodontal disease and no proximal decay (11). This SNP produced a shift from arginine (R) to lysine (K) at amino acid position 47 in the antimicrobial region of the LTF protein (11). As a result of this finding, we created two recombinant LTF proteins differing in one amino acid, i.e., lysine as opposed to arginine at position 47 in the N terminus. Two full-length LTF proteins were created, expressed in an insect vector, and purified for testing against S. mutans and A. actinomycetemcomitans. The lysine variant had antimicrobial activity against

1596

iai.asm.org

Infection and Immunity

S. mutans, while the “wild-type” arginine form had no such activity (11). However, proof that these recombinant proteins act similarly in vivo in human saliva is still lacking. Independently, but based on the genetic findings discussed above, caries levels were surveyed in a group of 12-year-old children from Brazil (12). In that study, subjects were assessed for caries, and genetic studies were done to determine their LTF variant genotypes. It was shown that subjects with the lysine LTF variant had significantly less caries, and it was suggested that the lysine variant had a protective effect against caries in this genetic association study (12). While the data were loosely connected, neither of the above-cited studies (11, 12) conclusively proved a direct functional link between the lysine variant in human saliva and (i) a preferential activity of this saliva against S. mutans and other acid-producing/caries-associated bacteria or (ii) reduced caries levels in humans homozygous for this lysine variant. This is the first report to isolate, identify, and characterize LTF in human saliva as a major protein with anti- S. mutans activity and to demonstrate that saliva containing this lysine variant protein displays antimicrobial activity against S. mutans and a host of other acid-producing oral bacteria. In addition, this report is the

Received 2 October 2012 Returned for modification 8 November 2012 Accepted 20 February 2013 Published ahead of print 4 March 2013 Editor: J. B. Bliska Address correspondence to Daniel H. Fine, [email protected]. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /IAI.01063-12. Copyright © 2013, American Society for Microbiology. All Rights Reserved. doi:10.1128/IAI.01063-12

p. 1596 –1605

May 2013 Volume 81 Number 5

Lactotransferrin Variant Reduces Caries Susceptibility

first to indicate that individuals homozygous for the lysine variant exhibit both anti-S. mutans activity and reduced proximal and smooth surface decay. With this evidence, we suggest that the LTF lysine variant can provide a plausible biological explanation for the functional role of LTF in the ecology of dental caries and can serve as a useful genetic marker for prediction of reduced susceptibility to decay. MATERIALS AND METHODS Selection of subjects for contribution of saliva and identification of anti-S. mutans activity. The Institutional Review Board of the University of Medicine and Dentistry of New Jersey (UMDNJ) approved the study. In the initial portion of the study, subjects were selected in an effort to identify one that had a specific salivary protein(s) with anti-S. mutans activity. Two subjects were selected for screening for suspected anti-S. mutans activity, and two were selected as inactive controls. Based on prior evidence, two subjects susceptible to localized aggressive periodontitis (LAP) with minimal proximal decay were selected because previous research indicated that they might possess saliva with anti-S. mutans activity (13). In addition, two subjects who were caries susceptible and harbored elevated levels of S. mutans were selected as screening controls, since previous research suggested that they were unlikely to possess the anti-S. mutans factor in their saliva (13). All subjects selected for screening were medically healthy, willing to be subjected to an oral examination, and willing to contribute plaque and saliva for testing for anti-S. mutans activity. Signed consent was required prior to participation in the study. Collection of saliva for testing. The two subjects susceptible to LAP (the test group) and the two individuals who had elevated levels of S. mutans and were susceptible to caries (the control group) were screened and contributed 1 to 2 ml of saliva for determination of anti-S. mutans activity. After the screening results were obtained, one subject whose saliva showed antimicrobial activity against S. mutans (from the test group) was asked to contribute 100 ml of whole saliva for isolation and characterization of the anti-S. mutans salivary protein (see below). In addition, one subject from the control group was also asked to volunteer for collection of 100 ml of saliva. Testing for antimicrobial activity of saliva. To accurately test the anti-S. mutans activity of the saliva of interest, we created a rifampin (rif)-resistant strain of S. mutans for in vitro testing. In this manner, S. mutans already existing in the subject’s saliva would not interfere with our ability to assess anti-S. mutans activity in whole saliva, because native S. mutans could not grow on rifampin-containing selective medium. A rifampin-resistant strain of S. mutans ATCC 33402 was created for testing by growing the parent strain in mitis salivarius (MS) agar containing 35 ␮g/ml rif (13). Resistant colonies that grew on the agar were restreaked on MS agar containing 70 ␮g/ml rif to obtain S. mutans 33402 rif, which was prepared for testing of salivary activity against S. mutans. To identify saliva that had anti-S. mutans activity, we developed a disk diffusion assay as described previously (13). Briefly, 200 ␮l of clarified saliva, at a protein concentration of 40 to 50 ␮g/ml, was applied to the surface of a sterile 22- by 22-mm filter paper disk. The disk was placed in an incubator set at 37°C for 1 h so that the saliva was fully incorporated into the dried disk (13). For testing, S. mutans strain 33402 rif, obtained from a frozen stock, was grown under aerobic conditions overnight at 37°C in Trypticase soy broth (TSB) containing 70 ␮g/ml rif. Cells were collected by centrifugation of the broth culture at 8,000 ⫻ g for 10 min. After the supernatant was discarded, the pellet containing the densely packed cells was resuspended in phosphate-buffered saline (PBS) to an optical density at 560 nm (OD560) of 0.8, equivalent to 1 ⫻ 108 cells/ml. A culture of S. mutans 33402 rif obtained from this washed and PBSsuspended pellet was streaked to confluence on an MS agar plate containing rif. The filter paper disk containing the saliva to be tested was placed over the center of the confluent S. mutans rif culture and incubated for 24 h at 37°C. The zone of inhibition encircling the disk was evaluated and compared to those for both positive and negative controls. The negative

May 2013 Volume 81 Number 5

control was a filter paper disk soaked in PBS, and the positive control was a disk soaked in the antimicrobial chlorhexidine at 10 ␮g/ml. Disks containing saliva from the control subjects were compared to disks containing saliva obtained from the test group. The saliva samples were diluted 2-fold to 10-fold and tested in the diffusion assay. In the initial tests, we assessed the minimal concentration of saliva that demonstrated a perceptible clear zone that indicated that S. mutans growth was impaired. In order to calculate the MIC of saliva, saliva was diluted sequentially and added to TSB containing 1 ⫻ 108 S. mutans rif cells/ml. In this manner, we could assess the minimal concentration of salivary protein required to completely eliminate the activity of S. mutans in the saliva being tested. Anti-S. mutans activity was defined as the salivary protein concentration required to reduce S. mutans rif growth by more than 70% compared to 2-hour growth in enriched medium (our definition of MIC for these tests). Saliva was defined as inactive if it reduced S. mutans rif growth by less than 20% compared to 2-hour growth in TSB. Saliva samples from subjects were classified as either active (having anti-S. mutans activity) or inactive (not having anti-S. mutans activity). Isolation of salivary factor by affinity column chromatography. For testing, a minimum of 100 ml of saliva was required from a test subject with anti-S. mutans activity. Collection was done over ice over a period of several days, and saliva was stored at 20°C until a total of 100 ml was collected. Prior to use, saliva was subjected to centrifugation at 10,000 ⫻ g for 30 min, after which the supernatant was decanted, yielding a clarified saliva which was used for testing. A second subject with no anti-S. mutans activity contributed saliva in a similar manner for control purposes. For isolation of the anti-S. mutans salivary protein, S. mutans strain 33402 was grown overnight at 37°C in TSB to obtain a cell density of 4 ⫻ 1010 cells/ml. In this case, rifampin labeling was not necessary because we were intent on isolating a salivary protein that bound to our immobilized S. mutans target. After checking for purity, S. mutans cells were heat fixed in a water bath at 60°C for 30 min, washed in PBS, subjected to centrifugation at 5,000 rpm for 5 min, and then fixed in 1% formalin in PBS for 2 h at room temperature (RT). Cells were washed again in a solution of 0.5% formalized saline, slurried in acetone, and then packed by centrifugation at 5,000 rpm for 10 min. The packed cell volume was measured, and cells were washed with 0.01 M Tris-HCl– 0.154 M NaCl (pH 7.4; standard Tris buffer [STB]) or 0.01 M Tris-acetate buffer (STAB). DEAE-cellulose (Sigma, St. Louis, MO) in its chloride form was prepared by mixing 15 g of DEAE-cellulose suspended in STB with 5 ml of packed S. mutans cells (14). Columns with a 5-ml bed volume were poured and equilibrated several times with STB. The eluate was monitored at 280 nm. After equilibration and assurance that bound S. mutans was retained, clarified saliva was loaded onto the column. The column was washed with STB, and nonabsorbed proteins were eluted with the equilibrating buffer (STB). When it was clear that equilibration was complete, glycine-HCl (pH 2.3) was added to elute bound salivary proteins. The column was reequilibrated with STB, and then potassium thiocyanate (KSCN) was used to elute any other proteins still bound to the column (14). Fractions eluted with either glycine-HCl or KSCN after neutralization were tested for anti-S. mutans activity in the disk diffusion assay. Identification of salivary fraction with antimicrobial activity. For SDS-PAGE, the anti-S. mutans fraction was separated in precast 10% polyacrylamide gels with 5% stacking gels (Bio-Rad), using the buffer system of Laemmli (15). Protein bands were stained with either a silverXpress kit (Invitrogen, Carlsbad, CA) or CYPRO Ruby (Bio-Rad Laboratories, Hercules, CA) per the manufacturer’s instructions. The SDS gel containing fraction 19, the active fraction, only occasionally contained more than one visible band. As a result, samples from the active fraction were concentrated by use of a YM-3 centrifugal filter (Millipore, Billerica, MA) and electrophoresed in an SDS-PAGE gel. Protein bands were excised from the stained gel and stored at 4°C until mass spectroscopy could be performed. Since only a few well-separated bands were seen, we excised each band and subjected each excised band to trypsin digestion. Protein identification of peptides thus obtained was performed by matrix-assisted laser desorption

iai.asm.org 1597

Fine et al.

TABLE 1 Demographics of subjects by genotype No. of subjects Genotype

Gender

Mean age ⫾ SD (yr)

Caucasian

Hispanic

AfricanAmerican

KK

Male Female

33.3 ⫾ 10.9 28.8 ⫾ 10.8

3 2

1 1

0 1

RK

Male Female

37.6 ⫾ 9.1 29.0 ⫾ 6.5

2 1

3 2

2 2

RR

Male Female

31.6 ⫾ 8.3 38.6 ⫾ 19.7

1 1

3 1

2 2

Total

Male Female

34.6 ⫾ 9.1 32.4 ⫾ 13.5

6 4

7 4

4 5

ionization–time of flight (MALDI-TOF) mass spectroscopy using an Applied Biosystems Voyager DE-STR mass spectrometer. Database searching based on the masses of identified peptides was performed by the Center for Advanced Proteomics research at UMDNJ-NJMS. The major band identified by MALDI-TOF analysis had a molecular weight of 78,000, which the report stated likely corresponded to LTF. As a result, we ran another SDS gel with the affinity column-eluted proteins and used commercially available LTF in a separate lane as a reference against which the eluted proteins could be compared. Western blotting was performed on the eluted proteins as well as the commercial LTF reference protein. Proteins were transferred from the SDS gel to Immobilon-P membranes (Millipore). Rabbit anti-human LTF IgG (Sigma) at a 1:1,000 dilution was used to interrogate the transferred proteins for the presence of LTF. Goat anti-rabbit IgG at a 1:7,500 dilution was used as the secondary antibody. The goat antibody was incubated with alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma) and visualized with a p-nitrophenylphosphate colorimetric reagent per the manufacturer’s instructions. This allowed us to compare the eluted protein to the commercially available control LTF on the Western blot in an effort to confirm the identification of LTF by MALDI-TOF. Genotyping assay to classify individuals homozygous for KK or RR. Previous work in our lab had shown that a specific insect-expressed recombinant genotype of LTF, the lysine variant, had anti-S. mutans activity, while the “wild-type” form with arginine at position 47 had no such activity (11). The lysine variant with defined activity against S. mutans had never been shown in human saliva (11). With evidence that the active anti-S. mutans fraction that was eluted from saliva was LTF, we decided to screen subjects for their K or R genotype and to test the saliva of all volunteers for anti-S. mutans activity. To accomplish this screening exercise, we made the assumption that identification of individuals homozygous for the lysine genotype would provide us with saliva that exhibited the pure lysine phenotype. In contrast, we assumed that identification of subjects homozygous for the “wild-type” arginine genotype would provide us with saliva that exhibited the pure arginine phenotype for assessment. We also decided to determine the clinical caries status of those homozygous for each genotype (KK versus RR). We performed the same evaluations on a group that was heterozygous (KR). From the 200 subjects in our clinical panel, we selected a group of subjects who were willing to be subjected to genetic screening for the LTF lysine and arginine genotypes. Within this group, we selected a subset of 30 subjects who were also willing to allow for clinical caries evaluation, salivary testing, and repeat buccal epithelial cell (BUC)/DNA extraction and genotyping. We were able to enroll a total of 30 volunteers, among which 8 were homozygous for KK, 10 were homozygous for RR, and 12 were heterozygous (Table 1). Genomic DNA was isolated from BEC samples by use of a DNeasy tissue kit (Qiagen, Valencia, CA). A total of 100 ng of genomic DNA was used to amplify LTF by PCR, using 50 pmol each of

1598

iai.asm.org

primers 5=-GAACCACAGACCTCTAGCCAAT-3= and 5=-CTTTTGGAG GATTTCCCTCTCTCT-3= in a 100-␮l reaction mixture. Analysis of the allelic gene frequency among subjects was performed using PCR-restriction fragment length polymorphism (PCR-RFLP) analysis. Briefly, 30 PCR cycles of 94°C, 55°C, and 72°C (1 min, 1 min, and 20 s, respectively) were performed in a TC-412 DNA thermocycler (Techne, Duxford, United Kingdom). The PCR product was purified with a QIAquick PCR purification kit (Qiagen) and eluted with 20 ␮l of Tris-EDTA buffer according to the manufacturer’s instructions. PCR-RFLP analysis was carried out by amplifying a 720-bp region of spanning polymorphic sites from DNAs isolated from test subjects and controls, as described previously (11). The 720-bp amplified product was digested with the EarI enzyme (New England Biolabs, Ipswich, MA) and electrophoresed in a 3% agarose gel. The presence of the restriction site resulted in two fragments, of 441 bp and 306 bp, which was indicative of the G allele encoding the K amino acid at codon 47. The PCR-RFLP products were also sequenced using Applied Biosystems 3130 XL DNA sequencing. Determination of anti-S. mutans activity of KK, RR, and RK saliva and further pretreatment of saliva with lactotransferrin antibody. One hundred microliters of adjusted cell suspension of S. mutans 33402 rif (A620 ⫽ 0.8) was pelleted and mixed with 100 ␮l of one of the following: PBS, clarified KK saliva (n ⫽ 8), clarified RR saliva (n ⫽ 10), or clarified heterozygous RK saliva (n ⫽ 12). Tubes were incubated at 37°C with constant agitation for 2 h. One hundred microliters of the mixture was removed from each tube and then serially diluted in PBS, plated on mitis salivarius rif agar, and incubated for 48 h, after which the CFU per ml was enumerated. All saliva data were normalized on a per-mg-of-protein basis. To determine whether the antimicrobial activity in the tested saliva could be attributed to LTF or something other than LTF, saliva was pretreated with antibody to LTF. In these experiments, saliva from each individual was pretreated with doubling dilutions (from 1:1 to 1:32) of anti-human LTF antibody (Sigma Chemical Co.) to obtain an endpoint for LTF antibody inhibition of anti-S. mutans activity. Thus, pretreatment of saliva with doubling dilutions of LTF antibody in relationship to S. mutans growth and survival was evaluated for each of the subjects to rule out the contribution of factors other than LTF to salivary anti-S. mutans activity. Caries status. The clinical condition with respect to caries status as well as salivary anti-S. mutans activity was recorded for each of the subjects (Table 2). Caries determination for proximal and smooth surfaces was done by means of radiographic interpretation. Subjects were considered to be caries positive if they demonstrated more than one tooth that had proximal or smooth surface decay as determined by a radiolucency extending from the outer enamel surface into the dentino-enamel junction. Each proximal and smooth surface was evaluated. Those surfaces with restorations were also recorded as decayed. The number of decayed proximal surfaces was also calculated for each subject in each group. Two examiners were calibrated for caries determination and then examined each set of X-rays in a random manner, using coded numbers for X-rays so as to blind the investigators and reduce bias. Assessments of differences between groups (KK versus KR versus RR) were performed using nonparametric (chi-square) statistics. Creation of synthetic LTF peptides and testing the effects of the variants on a broad spectrum of microbes. Our goal in this study was to test each variant against a broad spectrum of oral bacteria, including several acid-producing bacteria (5). To accomplish this goal, we had synthetic peptides produced to represent the minimal region of activity encompassing both the lysine and arginine variants. It was our feeling that this testing could provide us with a clearer picture of the way in which a purified variant might influence a broad spectrum of oral plaque bacteria. Thus, two synthetic peptides were made, one for each variant, to determine the spectrum of antimicrobial activity of each variant (11). The two peptide variants were designed as 11-mer peptides of amino acids 39 to 49 of the N-terminal region of LTF, as follows: the lysine peptide sequence was FQWQRNMRKVR, with lysine at position 47 (in italics), and the arginine

Infection and Immunity

Lactotransferrin Variant Reduces Caries Susceptibility

TABLE 2 Genotype comparisons of LTF rs 1126478 between individuals with and without anti-S. mutans activity and with and without carries No. of subjects Genotype KK RK KK ⫹ RK RR KK RK KK ⫹ RK RR a

Anti-S. mutans activity

No anti-S. mutans activity

8 8 16 2

0 4 4 8

No caries

7 8 15 3

Caries

1 4 5 7

Fisher’s exact test P value

RR (95% CI)a

0.0011 0.0427 .0041

3.6 (1.5–11.13) 3.33 (1.09–14.63) 4.00 (1.62–14.10)

0.0248 0.1984 0.0104

5.58 (1.17–100) 2.10 (0.84–7.14) 2.80 (1.17–6.66)

Ninety-five percent confidence intervals for relative risk (RR) were calculated using Cytel StatXact.

peptide sequence was FQWQRNMRRVR, with arginine at position 47 (also in italics). Peptides were synthesized using Fmoc chemistry. The addition of amino acids was done in a protected environment, and the reagents and by-products were removed by filtration and recrystallization. The growing peptide chain was protected by the insoluble solid phase. Resultant peptides are known to be identical to samples prepared by the standard p-nitrophenyl ester procedure. Peptides were prepared in this manner by Sigma and by our Center for Advanced Proteomics and then tested against a number of Gram-positive and Gram-negative bacteria to determine their spectra of activity (Table 3). Typically, 100 ␮g/ml of either peptide (K or R form) was mixed with a standard overnight culture of several Gram-positive bacteria, including S. mutans, Actinomyces viscosus ATCC 43146, S. sanguinis ATCC 10556, S. mitis ATCC 15914, S. gordonii ATCC 10558, etc. (Table 4 gives a complete list of the bacteria tested; see Table S1 in the supplemental material for information on colony counts), each at a concentration of 1 ⫻ 108 cells/ml. Prior to testing, the optical density and purity of the culture were confirmed. After overnight incubation, 100 ␮l of the culture plus its associated peptide treatment was removed and then plated on TSB agar for colony counting (CFU/ml), and the colony count was log10 transformed for analysis. When the activities of the two peptides were determined, both the LTF/K peptide and the LTF/R peptide were used in add-back experiments to determine if addition of the peptide to the inactive saliva or PBS would now increase its antimicrobial activity. One hundred microliters of adjusted cell suspension of S. mutans (A620 ⫽ 0.8) was pelleted and mixed with 100 ␮l of one of the following: PBS, clarified RR saliva, clarified RR saliva with peptide LTF/K (100 ␮g/ ml) added, or a series of clarified RR saliva samples with peptide added and doubling dilutions from 1:1 to 1:32 of anti-human LTF antibody (Sigma Chemical Co.). Tubes were incubated at 37°C with constant agitation for 2 h. One hundred microliters of the mixture was removed from each tube and then serially diluted in PBS, plated on mitis salivarius agar, and incubated for 48 h, after which the CFU/ml was enumerated and then log10 transformed for statistical testing. Data analysis. CFU/ml data were determined and log transformed. All statistical analyses were performed using JMP 9.0 and Cytel StatXact software. Determinations of significant differences (P ⱕ 0.05) between salivary treatments and LTF antibody impediment of S. mutans activity were made by analysis of variance (ANOVA). For chi-square analysis and determination of relative risk for clinical caries levels of subjects, the data were transformed into nominal data. For caries scores, a score of 0 or 1 was considered to represent no caries, and scores of 2 or higher were consid-

TABLE 3 Lactotransferrin peptide variants and their propertiesa

ered to be positive for caries. Anti-S. mutans activity was considered present (⫹) if the drop in CFU/ml after salivary treatment was 70% or greater and absent if the activity was 20% or lower than the baseline (0 h) value. Chi-square analysis was performed on the nominal data to determine if significant differences in the patterns of caries or anti-S. mutans activity versus the kind of salivary treatment (RR, RK, or KK) existed. Pearson scores were calculated to show the level of significance.

RESULTS

Isolation of salivary anti-S. mutans factor by affinity column chromatography. The antimicrobial activity found in saliva obtained from a subject with LAP is shown in Fig. 1. The elution profile indicated that the greatest anti-S. mutans activity was found in fraction 19 (Fig. 1A and B). This fraction (19) was sub-

TABLE 4 Killing of various streptococci and other bacteria by peptides Change in log CFU/ml with: Bacterial strain

Lysine peptide

Arginine peptide

PBS control

A. viscosus 43146

⫺0.40

⫺0.07

⫺0.08

S. sanguinis 903

⫺1.05

⫺0.04

⫺0.06

S. mitis 15914

⫺0.39

⫺0.04

⫺0.03

F. nucleatum 25586 C. sputigena 33612 V. parvula 10790

⫺0.46

⫺0.45

⫺0.02

⫺0.59

⫺0.60

⫺0.02

⫺0.58

⫺0.47

⫺0.06

S. sanguinis 10556 S. gordonii 10558

⫺0.56

⫺0.15

⫺0.12

⫺0.52

⫺0.06

⫺0.10

S. mutans 25175

⫺1.13

⫺0.26

⫺0.20

S. mutans 31383

⫺1.08

⫺0.16

⫺0.12

S. mutans 33535

⫺1.07

⫺0.15

⫺0.09

Peptide

Sequence

Mol wt

S. mutans 31341

⫺1.03

⫺0.14

⫺0.16

LTF/K LTF/R

FQWQRNMRKVR FQWQRNMRRVR

1,548.85 1,576.9

S. mutans 33534

⫺1.10

⫺0.16

⫺0.14

a

Both peptides had 11 amino acids and a charge of ⫹4.

May 2013 Volume 81 Number 5

Growth conditions ATCC medium 260, aerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 1053, anaerobic, 37°C ATCC medium 1490, anaerobic, 37°C ATCC medium 1252, anaerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 260, aerobic, 37°C ATCC medium 1169, aerobic, 37°C ATCC medium 260, aerobic, 37°C

iai.asm.org 1599

Fine et al.

FIG 1 Isolation of anti-S. mutans activity from saliva from an LAP subject and its associated activity. (A) Elution profile of fractions collected from the S. mutans affinity column, with the protein peak at fraction 19. (B) Activity against S. mutans derived from fraction 19. (C) Disk diffusion assay. Anti-S. mutans activity is shown as a zone of inhibition at the edge of the filter disk. The zone is seen in the middle panel (fraction 19 added to PBS) and the right panel (addition of fraction 19 to normal, nonactive saliva).

sequently added to either PBS or saliva from a healthy non-periodontally diseased subject with no anti-S. mutans activity. This new enriched saliva or PBS was then absorbed onto filter paper, which was then placed over a confluent culture of S. mutans. A zone of anti-S. mutans activity was now seen in the PBS mixed with fraction 19 and in the initially nonactive “normal” saliva now treated with fraction 19 (Fig. 1C; the panel on the extreme left shows no zone of antibacterial activity with untreated “normal” saliva). Elution fraction 19 was run in an SDS-PAGE gel and demonstrated one major band (Fig. 2A and B). MALDI-TOF spectroscopy indicated that the peptide fingerprint removed from the major band on the SDS gel was related to LTF. Furthermore, since MALDI-TOF suggested that LTF and our initial gels showed a band at approximately 78 kDa, we ran another SDS gel with LTF as a reference (Fig. 2A, lanes 1 and 2). The gel with both the eluted fraction and the reference LTF was then transferred for Western blotting (Fig. 2A, lanes 3 and 4). As shown, fraction 19, the active fraction, had a major band which migrated an identical distance compared to the LTF reference band (Fig. 2, lanes 1 and 2). The Western blot was interrogated with a commercially available rabbit antibody to LTF (Sigma) and indicated that the band at 78 kDa reacted with antiserum to LTF (Fig. 2A, lanes 3 and 4). A faint lower-molecular-weight band was found in the SDS gel in some cases, but this band(s) did not react with the commercial LTF antiserum. Both the SDS gel and the Western blot suggested that LTF was the major protein band in fraction 19. Furthermore, pretreatment of salivary fraction 19 with antibody to LTF decreased the anti-S. mutans activity of fraction 19 by almost 2 log, whereas pretreatment of fraction 19 with bovine serum albumin (BSA) at the same protein concentration as antibody to LTF had no effect (Fig. 2B). These results suggested that LTF was the major band

1600

iai.asm.org

contained within fraction 19. Taken together, these results suggested that LTF was likely to be an important component of the antimicrobial fraction isolated from the saliva of this LAP subject. Anti-S. mutans activity is associated with the lysine (K) variant and with minimal decay. Since we had previously shown that a recombinant human lysine variant of LTF produced in an insect vector killed S. mutans, we examined the saliva of subjects who were homozygous for the lysine variant of LTF to determine if anti-S. mutans activity could be shown (11). Saliva samples from 30 subjects, obtained from a panel of over 200 subjects on file in our clinical research center, were tested for antimicrobial activity against S. mutans. In addition, genomic DNAs obtained from buccal epithelial cells from all 30 subjects were genotyped for the LTF R47K polymorphism. The demographics of the subjects are shown in Table 1. Overall, saliva from KK subjects reduced the growth of S. mutans anywhere from 1.0 to 1.4 log compared to saliva from RR subjects (Fig. 3A). Those individuals who were heterozygous (RK) had intermediate antimicrobial activity that was variable. Furthermore, this activity was dose dependent, and thus, as saliva was diluted, the activity decreased (Fig. 3B). On an individual basis, saliva samples from all 8 subjects possessing the KK genotype had significant anti-S. mutans activity (P ⫽ 0.001), those from 2 of 10 RR individuals had modest anti-S. mutans activity, and those from 8 of 12 RK subjects had intermediate anti-S. mutans activity. These results were statistically significant and biologically meaningful. Compared to RR subjects, the likelihood of having anti-S. mutans activity was 3.6 for KK subjects (95% confidence interval [95% CI], 1.5 to 11.13; P ⫽ 0.001) and 3.33 for RK subjects (95% CI, 1.09 to 14.63; P ⫽ 0.043) (Table 2). Relationship of genotype to proximal decay. Among the 30 subjects available for clinical evaluation, 1 of 8 subjects in the KK

Infection and Immunity

Lactotransferrin Variant Reduces Caries Susceptibility

FIG 2 Identification of LTF as a salivary “factor” with anti-S. mutans activity. (A) SDS-PAGE gel (lanes 1 and 2) and Western blot (lanes 3 and 4). Lane 1 contains the LTF standard, while lane 2 contains fraction 19. Lanes 3 and 4 show that the 78-kDa bands transferred from lanes 1 and 2 reacted with antibody to LTF. (B) Graph showing the 2-log reduction in S. mutans growth with salivary fraction 19 and the increase in S. mutans growth when this fraction was pretreated with a commercially available antibody to LTF (compare the two middle bars). Pretreatment of fraction 19 with BSA had no effect.

group had proximal decay, with a total of 4 carious lesions, while 4 of 12 subjects in the RK group had proximal decay, with a total of 15 proximal lesions. In the RR group, 7 of 10 subjects had proximal decay, with a total of 31 proximal lesions and 4 smooth surface lesions. Comparison of caries in the three groups showed the following order: RR ⬎ RK ⬎ KK (Table 2). Thus, KK subjects were less likely than RR subjects to have proximal caries (P ⫽ 0.02; relative risk ⫽ 5.58; 95% CI ⫽ 1.17 to 100), while the relative risk for the KK and RK groups was 2.8 compared to the RR group (95% CI ⫽ 1.17 to 6.66; P ⱕ 0.01) (Table 2). Reduction of anti-S. mutans activity in saliva by pretreatment with lactotransferrin antibody. To further determine whether the activity seen in the saliva obtained from KK, RK, and RR subjects could be attributed to LTF, we pretreated saliva with commercially available antiserum to LTF (Sigma). The anti-S. mutans activity of saliva from KK subjects was compared to that of KK saliva after pretreatment with preimmune serum and/or commercially available antiserum to LTF (Fig. 4). Activity of the KK saliva against S. mutans was not affected by pretreatment with preimmune serum but was affected in a dose-dependent manner by pretreatment with commercially available antiserum to LTF (data not shown). The reduction of activity produced by the LTF antibody was significant for the RK and KK subjects. These reductions were significantly different from the activity found in treated saliva from RR subjects (no change) or in untreated saliva from KK subjects (Fig. 4). In the case of the RR subjects, saliva from 8 of 10 subjects had no effect on S. mutans viability, and antibody pretreatment had no effect. Two of the RR subjects showed some anti-S. mutans activity (RR-X); however, pretreatment of their saliva with antibody to LTF showed no reduction in activity, suggesting that the antimicrobial activity was derived from some

May 2013 Volume 81 Number 5

source other than LTF (Fig. 4). The experiments indicated that saliva from 18 homozygous subjects (8 KK subjects and 10 RR subjects) expressed the antimicrobial activity pattern expected and that the many “extraneous” agglutinating or heterocomplexing factors or proteases known to be present in saliva did not interfere with the antimicrobial activity observed. Those individuals who were heterozygous (RK) had intermediate antimicrobial activity that was variable, and this activity was reduced by pretreatment with LTF antibody (Fig. 4). Creation of synthetic peptides, testing of their antimicrobial activity, and use of antibody to block the antimicrobial activity. Two 11-mer (amino acids 39 to 49) synthetic peptides were created from the LTF N-terminal region (Table 3). The peptides contained either the K or R amino acid at position 47 but were otherwise identical. Each peptide was tested against a battery of Gram-positive and Gram-negative oral bacteria. A 100-␮g/ml concentration of the K or R peptide in PBS was added to overnight cultures of each bacterium for 2 h with rotation. Controls consisted of PBS alone added to overnight cultures. For example, the K peptide reduced S. mitis, S. sanguinis, S. gordonii, and A. viscosus growth by about 1/2 log (Table 4). The five S. mutans strains tested with the K peptide were reduced by about 1 log, and little to no reduction was seen with the R peptide. In all cases, differences between the K and R peptides were statistically significant as measured by ANOVA (P ⬍ 0.01). In assessments of the activity of the R peptide relative to the Gram-positive bacteria tested, we concluded that the R peptide had no effect (Table 4). The effects of the two peptides on three prominent Gram-negative plaque bacteria, i.e., Fusobacterium nucleatum ATCC 25586, Capnocytophaga sputigena ATCC 33612, and Veillonella parvula ATCC 10790, were in

iai.asm.org 1601

Fine et al.

FIG 3 Antibacterial activity of saliva from subjects with LTF genotypes. (A) The leftmost bar shows the anti-S. mutans activity of saliva from subjects homozygous for lysine (KK). The next bar shows results for heterozygous (RK) subjects, the following bar shows anti-S. mutans activity of saliva from subjects homozygous for arginine (RR), and the last bar shows the PBS control. Those homozygous for lysine (n ⫽ 8) showed an approximately 1.3-log reduction (⫺1.3 ⫾ 0.10) in survival of S. mutans. Those with a heterozygous genotype had about a 1.1-log reduction, while those homozygous for arginine (n ⫽ 10) had no effect (P ⬍ 0.01; ANOVA). (B) Dose-dependent reduction observed by diluting saliva from KK subjects. S, undiluted saliva; S1, 1/10 dilution of saliva; S4, 1/2 dilution of saliva. Saliva for dilution testing was derived from 5 KK subjects. Assays were performed in triplicate.

the range of about 1/2 log, but differences between the two peptides were not evident (Table 4). DISCUSSION

Our studies are unique in that they demonstrate the isolation, definitive identification, and characterization of LTF as a specific salivary glycoprotein with antimicrobial activity against S. mutans. Furthermore, our new data show that the LTF lysine variant in whole saliva from individuals homozygous for this lysine variant has salivary activity against S. mutans. In contrast, individuals homozygous for the arginine variant have saliva with minimal to no antimicrobial activity against S. mutans. In this manner, our

1602

iai.asm.org

data reinforce the finding that a human recombinant lysine variant has preferential activity against S. mutans compared to the “wild-type” arginine form of LTF (11). The functional salivary data were supported by a reduction in caries prevalence in the lysine group as opposed to the arginine group. Furthermore, the creation of a lysine variant peptide allowed us to demonstrate the broad spectrum of antimicrobial activity that this variant can have on a range of Gram-positive, acid-producing/caries-associated bacteria. Taken together, these cumulative data suggest that the LTF lysine variant can exert a profound influence on the caries process. For the most part, the failure to conclusively prove an associa-

Infection and Immunity

Lactotransferrin Variant Reduces Caries Susceptibility

FIG 4 Anti-S. mutans activity in saliva derived from subjects with KK, RK, and RR genotypes, and effect of pretreatment of this saliva with antibodies to LTF. The figure shows the anti-S. mutans activity of 10 RR, 12 RK, and 8 KK individuals and the effect of pretreatment of saliva with antibody to LTF. The KK bars show results for 8 individuals: the left bar shows growth of S. mutans after treatment with antibody, while the second bar shows a significant reduction in S. mutans growth in untreated saliva with active LTF. While 8 of 10 subjects in the RR group had no anti-S. mutans activity, 2 RR subjects had some activity. It is particularly informative that the 2 individuals with activity (RR-X) were not affected by pretreatment of saliva with LTF antibody, suggesting that the anti-S. mutans activity seen was not related to lactotransferrin. The RK bars show a significant reduction in anti-S. mutans activity when saliva was pretreated with antibody to LTF (ANOVA; P ⱕ 0.01).

tion between a specific salivary protein such as LTF and its biological utility is attributed to both the multifunctionality and redundancy of salivary proteins, suggesting that one protein can be replaced by the function of another that has similar and overlapping biological activity (4, 16). Moreover, salivary proteins form heterotypic complexes, making it difficult to assign a unique function to one particular protein (17, 18). In the current study, the first step and major strength of the study were the successful isolation, identification, and characterization of LTF from human saliva and proof of the linkage of this protein to caries reduction. This success is attributed to two experimental choices: (i) selection of salivary donors drawn from a specific group (LAP) and (ii) use of an affinity column with S. mutans as the target microorganism to secure the salivary protein(s) of interest (14). First, by selecting a unique population that has minimal decay, we enriched our chances of choosing an individual who might possess a salivary protein with significant activity against caries-provoking bacteria. Second, by using an affinity column with S. mutans as the target organism, we enriched our chances of capturing the salivary protein of interest, based on the assumption that a protein with activity against a particular bacterium, i.e., S. mutans, will bind to that bacterium. This strategy proved successful, and MALDI-TOF spectroscopy and Western blotting definitively identified the affinity-captured protein as LTF. It is our feeling that our success was due to these two interconnected approaches to isolation and identification of LTF as a major salivary anti-S. mutans factor. Several years ago, we suggested that the relationship between

May 2013 Volume 81 Number 5

saliva and the two most prominent dental diseases, caries and periodontitis, might be clarified by studying localized aggressive periodontitis. This suggestion was based on the fact that children with this form of periodontitis had almost no proximal decay (19). Typically, dental caries is pandemic for children of this age, while periodontitis is rare. This uncommon pattern of interproximal dental disease, e.g., minimal proximal decay compared to aggressive periodontitis (20), suggested to us that the prevalences of these two distinctly different diseases could be mediated by saliva, a pervasive environmental determinant in the oral cavity (4, 18, 19). In the current study, after definitively identifying LTF as the active salivary factor, and with the knowledge that the recombinant lysine variant killed S. mutans (11), we explored the relationship of the lysine and arginine SNPs to salivary anti-S. mutans activity and proximal caries levels. Thirty subjects were genotyped for LTF/K and LTF/R and examined for their antimicrobial activity and clinical caries status. We found that saliva from subjects homozygous for the LTF/K variant had significantly more anti-S. mutans activity and significantly less proximal decay. Our conclusion that the anti-S. mutans activity was due to the LTF/K variant was based on two assumptions: (i) those homozygous for K produce pure K saliva, while those homozygous for R produce pure R saliva; and (ii) pretreatment of LTF saliva with LTF antibody will reduce the anti-S. mutans activity in K saliva but have no effect on R saliva. These assumptions were proven to be correct and are supported by data indicating that an antibody specific to LTF blocks anti-S. mutans activity in whole saliva, in a dose-dependent

iai.asm.org 1603

Fine et al.

manner, in subjects homozygous or heterozygous for the lysine form of LTF. Since those homozygous for the “wild-type” R form of LTF had no S. mutans activity, blocking with LTF antibody had no effect. While 2 of 10 RR subjects did demonstrate anti-S. mutans activity, these two subjects maintained this activity despite pretreatment of their saliva with antibody to LTF, indicating that this salivary anti-S. mutans activity was due to something other than LTF. To increase our assuredness that the activity is based on the LTF/K variant, in future studies we will need to increase the number of subjects we genotype and remove the LTF/K protein from saliva, as opposed to blocking its activity by pretreatment with LTF antibody as we did in this study. In the current study, to further substantiate the specificity of the LTF lysine variant for S. mutans and to explore the range of antibacterial activity, we developed two synthetic 11-mer peptides based on the LTF antimicrobial sequence (21). These peptides were tested against S. mutans and other Gram-positive, acid-producing bacteria in vitro (Tables 3 and 4). The results of these experiments confirmed the S. mutans activity of the lysine variant and also showed that LTF/K has activity against other Gram-positive, acid-producing microorganisms, while the LTF/R peptide has no activity against these organisms. The peptide data for a wider range of bacteria are consistent with those for other defensin-like agents that demonstrate a broad spectrum of antimicrobial activity (22, 23) and also support our hypothesis that the lysine variant can reduce the viability of caries-provoking bacteria. The peptide activity is also consistent with activity against acidproducing caries-associated bacteria such as A. viscosus and Lactobacillus acidophilus, as seen in whole saliva derived from subjects with aggressive periodontitis (24). Lactotransferrin’s antimicrobial effect is not seen exclusively in the oral cavity. The effect of LTF derived from breast milk and its influence on Haemophilus influenzae infection in children have been reported previously (25). Moreover, LTF has also been shown to affect Escherichia coli (26, 27), Salmonella enterica serovar Typhimurium (26), Vibrio cholerae (8), and Legionella pneumophila (28). Our data differ in that they indicate that a particular form of LTF, that is, the lysine form, can play a prominent role in the host defense against caries, a specific oral infection. While we cannot be certain that LTF acts alone, our research with LTF antibody pretreatment indicates that LTF accounted for at least 60 to 80% of the anti-S. mutans activity in the saliva tested. It is interesting that the recombinant lysine variant of LTF expressed in an insect vector also induced an upregulation of epithelial production of human beta defensin 2 (hBD-2) (11). hBD-2 has been shown to have significant antibacterial activity against S. mutans and S. sobrinus, two cariogenic microbes (29). Conversely, hBD-2 has been shown to have little to no effect on several Gramnegative periodontal pathogens (30). While hBD-2 clearly was not the dominant factor in the saliva tested in this study, it is possible that the caries reduction seen in the subjects we surveyed could have been due to the combined effect of the direct antimicrobial activity of the LTF lysine variant in concert with the upregulated mucosal defensin hBD-2, resulting in a coordinated assault on cariogenic microorganisms (31). This speculation needs further confirmation. In summary, we showed that an LTF variant in the saliva of a subset of subjects has the capability of killing S. mutans and other early plaque-forming, Gram-positive, acid-producing bacteria. In addition, we have developed a genotyping assay that differentiates

1604

iai.asm.org

between individuals who are homozygous for the LTF/K variant and the LTF/R “wild-type” form and that demonstrates that saliva samples obtained from subjects who are homozygous for each genotype show differing spectra of antimicrobial activity. Furthermore, we demonstrated that peptides designed to reflect the LTF/K and LTF/R phenotypes, and thus differing in one amino acid in the N-terminal region of LTF, show differing antimicrobial activities that are comparable to those seen for their salivary LTF counterparts. Moreover, pretreatment of active saliva containing the K variant with LTF antibody was shown to reduce its anti-S. mutans activity in a dose-dependent manner. Furthermore, subjects who have this lysine variant have minimal proximal and smooth surface decay compared to subjects who have the arginine form, who also show significantly greater decay. These results lead us to conclude that subjects who possess saliva with the LTF lysine variant genotype produce anti-S. mutans activity that coincides with the minimal proximal and smooth surface caries seen in these subjects. These findings provide insight into the role of salivaderived innate host factors such as LTF in relationship to microbial ecology and its association with susceptibility to caries, a pervasive oral infectious disease. ACKNOWLEDGMENTS We thank Kenneth Markowitz, Helen Schreiner, Marie McKiernan, and Gary Yue for their help throughout the study. We also acknowledge the help and suggestions received from G. Diamond in his review of the paper. We acknowledge support from the National Institutes of Dental and Craniofacial Research for providing financial support in the form of grants DE-016474 and DE-017968.

REFERENCES 1. Mandel ID. 2002. Oral infections: impact on human health, well-being, and health-care costs. Compend. Contin. Educ. Dent. 23:403– 406. 2. Satcher D. 2000. Oral health in America: a report of the Surgeon General, p 616 – 623. Department of Health and Human Services, NIH, Rockville, MD. 3. Mandel ID. 1987. The functions of saliva. J. Dent. Res. 66:623– 627. 4. Scannapieco FA. 1994. Saliva-bacterial interactions in oral microbial ecology. Crit. Rev. Oral Biol. Med. 5:203–248. 5. Loesche WJ. 1986. Role of Streptococcus mutans in human dental decay. Microbiol. Rev. 50:353–380. 6. Groenink J, Walgreen-Weterings E, Nazmi K, Bloscher JG, Veerman EC, van Winkelhoff AJ, Nieuw Amerongen AV. 1999. Salivary lactoferrin and low-Mr mucin MG2 in Actinobacillus actinomycetemcomitansassociated periodontitis. J. Clin. Periodontol. 26:269 –275. 7. Liu B, Rayment SA, Oppenheim FG, Offner GD, Fives-Taylor P, Troxler RF. 2002. Interaction of salivary mucin MG2, its recombinant N-terminal region and synthetic peptide with Actinobacillus actinomycetemcomitans. J. Periodontal Res. 37:416 – 424. 8. Arnold RR, Brewer M, Gauthier JJ. 1980. Bactericidal activity of human lactoferrin: sensitivity of a variety of organisms. Infect. Immun. 28:893– 898. 9. Lassiter MO, Newsome AL, Sams LD, Arnold RR. 1987. Characterization of lactoferrin interaction with Streptococcus mutans. J. Dent. Res. 66:480 – 485. 10. Kalmar JR, Arnold RR. 1988. Killing of Actinobacillus actinomycetemcomitans by human lactoferrin. Infect. Immun. 56:2552–2557. 11. Velliyagounder K, Kaplan JB, Furgang D, Legarda D, Diamond G, Parkin RE, Fine DH. 2003. One of two human lactoferrin variants exhibits increased antibacterial and transcriptional activation activities and is associated with localized juvenile periodontitis. Infect. Immun. 71:6141– 6147. 12. Azevedo LF, Pecharki GD, Brancher JA, Cordeiro CA, Jr, Medeiros KG, Antunes AA, Arruda ES, Werneck RI, de Azevedo LR, Mazur RF, Moyses ST, Faucz FR, Trevilatto PC. 2010. Analysis of the association between lactotransferrin (LTF) gene polymorphism and dental caries. J. Appl. Oral Sci. 18:166 –170.

Infection and Immunity

Lactotransferrin Variant Reduces Caries Susceptibility

13. Fine DH, Furgang D, Goldman D. 2007. Saliva from subjects harboring Actinobacillus actinomycetemcomitans kills Streptococcus mutans in vitro. J. Periodontol. 78:518 –526. 14. Fine DH, Tabak L, Stevens R. 1977. Affinity chromatography of antiserum to a gram negative organism. J. Immunol. Methods 16:91–96. 15. Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680 – 685. 16. Helmerhorst EJ, Oppenheim FG. 2007. Saliva: a dynamic proteome. J. Dent. Res. 86:680 – 693. 17. Rudney JD. 1989. Relationships between human parotid saliva lysozyme lactoferrin, salivary peroxidase and secretory immunoglobulin A in a large sample population. Arch. Oral Biol. 34:499 –506. 18. Rudney JD. 1995. Does the variability in salivary protein concentrations influence oral microbial ecology and oral health? Crit. Rev. Oral Biol. Med. 6:343–367. 19. Fine DH, Goldberg D, Karol R. 1984. Caries levels in patients with juvenile periodontitis. J. Periodontol. 55:242–246. 20. Loe H, Brown LJ. 1991. Early onset periodontitis in the United States of America. J. Periodontol. 62:608 – 616. 21. Hancock RE, Chapple DS. 1999. Peptide antibiotics. Antimicrob. Agents Chemother. 43:1317–1323. 22. Brogden KA. 2005. Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria? Nat. Rev. 3:238 –250. 23. Zasloff M. 2002. Antimicrobial peptides of multicellular organisms. Nature 415:389 –395. 24. Fine DH, Furgang D, McKiernan M, Rubin M. 2012. Can salivary activity predict periodontal breakdown in A. actinomycetemcomitans

May 2013 Volume 81 Number 5

25.

26.

27.

28. 29.

30.

31.

infected adolescents? Arch. Oral Biol. http://dx.doi.org/10.1016/j .archoralbio.2012.10.009. Plaut AG, Qui J, Grundy F, Wright A. 1992. Growth of Haemophilus influenzae in human milk: synthesis, distribution, and activity of IgA protease as determined by study of IgA⫹ and mutant IgA cells. J. Infect. Dis. 166:43–52. Naidu SS, Svensson U, Kishore AR, Naidu AS. 1993. Relationship between antibacterial activity and porin binding of lactoferrin in Escherichia coli and Salmonella typhimurium. Antimicrob. Agents Chemother. 37:240 –245. Chapple DS, Hussain R, Joannou CL, Hancock RE, Odell E, Evans RW, Siligardi G. 2004. Structure and association of human lactoferrin peptides with Escherichia coli lipopolysaccharide. Antimicrob. Agents Chemother. 48:2190 –2198. Bortner CA, Miller RD, Arnold RR. 1986. Bactericidal effect of lactoferrin on Legionella pneumophila. Infect. Immun. 51:373–377. Nishimura E, Eto A, Kato M, Hashizume S, Imai S, Nisizawa T, Hanada N. 2004. Oral streptococci exhibit diverse susceptibility to human b defensin-2: antimicrobial effects of hBD-2 on oral streptococci. Curr. Microbiol. 48:85– 87. Joly S, Maze C, McCray PB, Guthmiller JM. 2004. Human b-defensins 2 and 3 demonstrate strain-selective activity against oral microorganisms. J. Clin. Microbiol. 42:1024 –1029. Singh PK, Parsek MR, Greenberg PE, Welsh MJ. 2002. A component of innate immunity prevents bacterial biofilm development. Nature 417: 552–555.

iai.asm.org 1605