Characterization of Lactoferrin Interaction with Streptococcus mutans

M.O. LASSITER, A.L. NEWSOME, L.D. SAMS, and R.R. ARNOLD

Department of Oral Biology, Dental Research Center, Emory University School of Dentistry, Atlanta, Georgia 30322

Lactoferrin (LF) is an iron-binding glycoprotein common to exocrine secretions and the specific granules of neutrophils. Each molecule is capable of high-affinity coordinate-binding of two ferric ions with two bicarbonate or carbonic anions. The initial aspect of the present study was directed at determining the nature of anion involvement in LF bactericidal activity. It was found that selective anions were capable of inhibiting the expression of bactericidal activity by LF on S. mutans 10449. The ability to block LF expression was directly related to the capacity of the anion to serve as a coordinate ion in iron-binding by the transferrin molecules. These data support the hypothesis that the LE target site on the bacterial surface is anionic. There has been controversy in the literature regarding LF involvement in hydroxy radical generation. The second phase of these studies indicated that treatment of S. mutans with LF under anaerobic conditions abrogated the bactericidal effect of this molecule. LF-killing could be enhanced by the presence of thiocyanate and inhibited by catalase and lactoperoxidase; however, bovine serum albumin was equally eeffctive as an inhibitor. The apparent requirement for oxygen in LF bactericidal effect on S. mutans is not inconsistent with a hydroxy radical mechanism.

are single-chain, two-sited glycoproteins that characteristically have metal-binding activity linked to an anion-binding requirement (Aisen et al., 1967; Aisen and Leibman, 1973). This coordinate-binding of metal and anion appears to be unique to this family of proteins. Previous investigations concerning coordinate-binding have been aimed primarily at the specific ability to bind iron, although other multivalent cations may be capable of occupying the binding sites (Bates and Schlabach, 1973). The specific binding is usually characterized by: (1) no more than two metal ions bound per molecule, (2) the absence of binding to iron-saturated glycoprotein, and (3) one bicarbonate, or other acceptable anion, bound per each metal ion. The bicarbonate or carbonate anion is preferred, but other anions, such as EDTA, oxalate, or nitrilotriacetate, can serve as linking anions to promote the specific binding of iron (Bates and Schlabach, 1973; Aisen et al., 1967; Aisen and Leibman,

1973).
Previous studies from this laboratory have indicated that LF that is iron-saturated in the presence of appropriate anions is capable of neither killing S. mutans nor binding to its surface (Arnold et al., 1977). In apparent contrast, if bacteria are treated with apo LF in the presence of iron, killing proceeds, suggesting that iron is incapable of competitively inhibiting LFkilling (Arnold et al., 1982). The possibility exists that the surfaces of S. mutans and other target bacteria provide both anionic and cationic sites in proximity to each other, such that these sites could act to occupy the specific binding sites on LF. Such a necessary, initial event would help account for the inability of iron-saturated LF to bind to and, subsequently, kill S. mutans (Arnold et al., 1977, 1981). If LF anion-binding at the bacteria site precedes bactericidal activity, then the addition of exogenous anions to LF could serve to compete with a bacteria-anionic site. This would, in turn, inhibit LF-mediated killing of S. mutans. Furthermore, it would hold that those anions, whose molecular configuration would best serve as a synergistic anion to promote specific metal-binding, would be most efficient in blocking the bactericidal activity of LF. Several investigations have examined the role of iron-saturated LF in hydroxyl radical (OH-) generation. Ambruso and Johnston (1981) found that purified PMN LF and milk LF from a commercial source were capable of generating OH when iron-saturated, but not in its unsaturated form. Bannister et al. (1982) confirmed these results, suggesting that OH was generated by a Fenton-type, Haber-Weiss reaction. Winterbourne (1983), in contrast to these studies, found iron-saturated LF to be a poor catalyst of OH generation, and suggested that the complexing agents such as EDTA or NTA may influence radical production. Although there is not universal agreement as to the involvement of LF in OH' production, the potent properties of this radical would be attractive as an explanation for the bactericidal activity observed with LF. The first purpose of the current studies was to determine the involvement of anion in the targeting of LF activity by a series of competitive inhibition studies, using selective anions based on their ability or inability to serve as coordinate ion with iron. The second aspect of these studies was directed at determining the influence of oxygen on LF killing of S. mutans and involvement of oxidative species necessary for OH generation in the killing mechanism.

J Dent Res 66(2):480-485, February, 1987

Introduction.
Lactoferrin (LF) is an iron-binding glycoprotein of the transferrin family that is common to exocrine secretions of mammals and the specific granules of polymorphonuclear leukocytes of humans (Massons and Heremans, 1966; Massons et al., 1969). Each molecule of LF is capable of high-affinity coordinate-binding of two ferric ions with two bicarbonate or carbonic anions (Mazurier and Spik, 1980). In situ LF is found primarily in its iron-free state (Massons et al., 1969). In addition to bacteriostasis through sequestering of iron (Weinberg, 1978), previous studies from this laboratory have described a bactericidal activity of human iron-free (apo) LF against a variety of bacteria (Arnold et al., 1977, 1980). In subsequent investigations with Streptococcus mutans NCTC 10449 used as a model, it was determined that certain environmental factors can influence the expression of apo LF bactericidal activity (Arnold et al., 1981, 1982). Washed, exponential-phase bacteria begin to die within 15 min at 370C in the presence of 4 [Lmol/L apo LF but not of iron-saturated LF. This killing is optimal between pH 5 and 6 and is enhanced by temperature increase to 390C and eliminated by temperature reduction to 20C. Late-exponential/early-stationary-phase bacteria are more resistant to LF-killing. Iron-free LF is capable of binding to the surfaces of bacteria, and permeability barriers such as capsules, glucans, and outer membranes increase resistance to LF (Arnold et al., 1981). Addition of certain buffers such as glycine or sodium phosphate reduces killing of S. mutans by LF (Arnold et al., 198 1). The transferring, including LF, transferrin, and conalbumin,
Presented at the 10th International Conference on Oral BiologySaliva and Salivary Glands, June 22-24, 1986, RAI Conference Centre, Amsterdam, The Netherlands This research was supported by U.S. Public Health Service Grant DE 06869 from the National Institute of Dental Research. A.L.N. was supported by National Research Service Award F32-DE05409. 480

Vol. 66 No. 2

LACTOFERRIN EFFECTS ON S. mutans

Materials and methods.

Preparation of LF. - LF was purified from pooled human colostrum by heparin affinity chromatography according to Blackberg and Hernell (1980). There were no detectable contaminants, as determined by sodium dodecyl sulfate (SDS)polyacrylamide gel electrophoresis and by immunoelectrophoresis against rabbit anti-human colostrum. Apo LF was prepared by dialysis against an acetic acid/sodium acetate buffer containing 40 mmol/L EDTA and 0.2 mol/L sodium phosphate (pH 4), followed by exhaustive dialysis against de-ionized distilled water (Mazurier and Spik, 1980). Protein concentration of the purified LF preparation was determined by absorbance at 280 nm (El' - 11 0) and by the BioRad protein assay. Cultivation and enumeration of S. mutans. - Streptococcus mutans NCTC 10449 (serotype c) was grown in Todd-Hewitt broth at 370C to early-exponential phase (optical density between 0.2 and 0.3 at 660 nm) in an atmosphere of 95% N2 and 5% H2. Bacteria were harvested and washed twice by centrifugation at 1000 g, and re-suspended in de-ionized distilled water (pH 5.5) to a concentration of 107 CFU S. mutans per mL. Colony-forming units were quantitated on trypticase soy agar containing 1g% sucrose (TSA-sucrose) by means of the Spiral Plater System (Spiral Systems, Inc., Cincinnati, OH). All samples were plated in duplicate, and all assays were repeated a minimum of three times. Competitive assay for anion inhibition. All reagents were dissolved in de-ionized, distilled water. The pH of each anionic solution was adjusted to 5.5 with 0.01 N HC1 or NaOH and filter-sterilized. The solutions of the sodium salts of oxalate, phosphate, sulfate, nalidixic acid, and EDTA were prepared immediately before use. To evaluate the effect of carbon dioxide, we incubated de-ionized water overnight in an atmosphere containing 5% C02; this water was added to the reaction tubes, and the tubes incubated in the C02-enriched atmosphere. Apo LF (5-10 gmol/L) was added to S. mutans (106 CFU per mL) in de-ionized water containing an appropriate concentration of the test solution of anion. Positive controls contained LF and S. mutans in the absence of iron, and negative controls omitted the LF or employed iron- and anion-saturated LF. The percent inhibition of LF bactericidal activity was calculated by dividing the difference between the test CFU and the positive control CFU by the negative control CFU and multiplying by 100. Neither the anion solutions nor the saturated LF killed the bacteria during the test period. Anaerobic experiments. - These experiments were conducted in an anaerobic glove box (Anaerobic Systems Model 1024, Forma Sci., Marietta, OH). LF preparations (10.4 vmol/ L) were equilibrated in 95% N2 and 5% H2 in aliquots of 0.3 mL for one or two hours prior to being incubated with bacteria. In addition, plates (TSA-sucrose) and dilution tubes with distilled water were equilibrated in the appropriate atmosphere for at least 24 hr prior to the experiment. Bacterial cultures in early-exponential phase were removed from the chamber for centrifugation in carefully sealed vessels to minimize exposure to oxygen. One sample was kept outside the glove box and was re-suspended in 10 mL of sterile distilled water equilibrated in air and treated with non-reduced LF. CFU were quantitated by aerobic plating of serial 10-fold dilutions at 0, 1, and 2 hr after LF addition. The other sample was returned to the box and treated, plated, and incubated anaerobically in the same fashion. Reaction tubes were maintained at 370C in dry heating blocks obtained from Fisher Scientific Company (Fair Lawn, NJ). Quantitation of colony-forming units was accomplished with TSA sucrose. All plates in these experiments were incubated under the N2-H2 atmosphere for 48 hr.

0
U-

us

Effect of nalidixic acid on LF bactericidal activity against S. mutans: * S. mutans only; 0 LF + 50 pLmol/L nalidixic acid; 0 LF + 20 iLmol/L nalidixic acid; X LF + 10 pmol/L nalidixic acid; A LF only and S. mutans. According to the text.

.u

0)

T2100' 90' )
a
4

UD80' 70' o 60' *50 f 40' E 30' U 20'

0Q' I

481

70

6.0

5.0'

*=ZZJ

4.0*
3.0
IN < 2.41
-

-3

Time(hrs)
-

Fig. 1

&3
I

DUX.
/~~~~~~~~

05 1020304050

pJM Concentration

100

1.0 10

50 Iloo mM Concentration

Modification of LF bacteria killing (after 2 hrs) with different anionic solutions. Left panel: O-nalidixic acid; A-citrate;-LCEDTA;*--oxalate. Right panel: j phosphate; *-glycine;--U-sulfate; * sodium chloride. According to the text.

Fig. 2

Results.
Anion effects on bactericidal activity of LF. - In this series of experiments, the sodium salts of oxalate, phosphate, sulfate, EDTA, and nalidixic acid were added to various concentrations to the reaction mixture of S. mutans and apo LF at a suboptimal concentration (2 pmol/L). Recoverable CFU of these experimental groups were compared with those of LF treatment in distilled water or in 100 mmol/L NaCl (pH adjusted to 5.5). Viability was monitored through 3 hr at 370C. Fig. 1 presents the data for these experiments with nalidixic acid. As can be seen, 50 jpmol/L nalidixic acid totally inhibited LF bactericidal activity over the entire incubation period. Concentrations of 20 jxrmol/L and 10 pimol/L both resulted in significant inhibition of LF-killing, although killing was apparent at the second and third hours with both concentrations. Fig. 2 presents the percent inhibition of LF bactericidal activity, after 2 hr of

482

LASSITER et aDl. J Dent Res

FebruarY

1987

E
0
-

I& To

TIME (hrs)

3 ~~~2
TIME(HRS)
2

Fig. 3- Effect of carbon dioxide on LF bactericidal activity against S. mutaos: 0 S. mnutans only: * S. mutans only + CO, treatment; a 3.0 pmol/L LF; L 3.0 Vjmol/L LF + CO2 treatment; A 1.5 Ftmol/L LF: A 1.5 Rmol/L LF + CO2 treatment; 0 0.75 FLmol/L LF: * 0.75 pLmol/ L LF + CO,. According to the text.

Fig. 4 -Effect of 95% N,25% H2 on LF bactericidal activity against S. mutans: A S. mutans in air; 0 S. mutans in N2/H2; A S. nutans +
LF in air; * S. mutans + LF in N,/H,. According to the text.

incubation, for each of the test anions. As can be seen, oxalate, EDTA, and citrate as well as naladixic acid had significant effects on LF bactericidal activity at Rmol/L concentrations. In contrast, phosphate, glycine, and sulfate required a 1000fold greater concentration to give comparable inhibition of LFkilling. In Effects of carbon dioxide on LF bactericidal activity.
this
series

of experiments, the numbers of bacteria and the

CO2 concentration were kept constant, and the concentration of apo LF was varied. As can be seen in Fig. 3, an atmosphere of 5% CO2 resulted in a marked diminution of LF bactericidal activity against S. mutans. This inhibition was most apparent at reduced LF concentrations. When the concentration of LF was lowered to 750 nmol/L and incubated in normal atmosphere. no CFU could be recovered after 3 hr (> 3.5 log reduction); however, with CO2 treatment, fewer than I log of bacteria were killed with the same LF concentration. Effects of anaerobiosis on LF-killing. In this series of experiments, apo LF resulted in > 99.9% reduction in viability of S. mutans in 2 hr at 37TC in air. However, when the same bacteria were treated with LF under an atmosphere of 95% N25% H2 (Fig. 4), the bactericidal effect of LF was totally abolished. This anaerobic gas mixture was free of CO2 to avoid the CO2 phenomenon observed in the previous experiment. Involvement of endogenous hydrogen peroxide in LF-killing. The previous experiments suggested that oxygen was involved in LF-mediated killing. If the mechanism involved generation of OH by LF-associated iron via a Fenton-type Haber-

Weiss reaction, then both hydrogen peroxide and superoxide would be predicted to be involved. To test this hypothesis, we added bovine catalase (E.C. 1. 11. 1.6; 20,000 units per mg), bovine lactoperoxidase (E.C. 1. 11. 1.7; 93 units per mg), and bovine serum albumin (all obtained from Sigma Chemical Co. St. Louis, MO) at concentrations of 1 pumol/L to S. mutans concurrently with apo LF (650 nmol/L). As can be seen in Fig. 5, catalase inhibited LF-mediated killing of S. mutans. Heat inactivation of the catalase resulted in an elimination of this inhibition. Lactoperoxidase also resulted in a partial reduction of LF-mediated killing. The enzymatic involvement of these proteins may be questioned by the observation that bovine serum albumin (BSA) was also capable of inhibiting the bactericidal effect. However, such inhibition may be explained by scavenging reactive oxygen species by BSA. None of the enzymes affected the viability of S. mutans in the absence of LF. Influence of thiocyanate (SCN-) on LF-mediated killing. In the preceding experiments, when NaSCN was added with the lactoperoxidase, there was enhancement of bactericidal activity in the presence of LF (data not shown), in contrast to the partial inhibition that was observed by lactoperoxidase alone. This series of experiments was designed to examine this apparent enhancement of LF activity, and for these studies, various concentrations of NaSCN were added concurrently with LF (in the absence of lactoperoxidase) to washed S. mutans. As can be seen in Fig. 6, 5 mmol/L SCN- resulted in enhancement of LF-killing at 3 hr. 10 mmol/L SCN- consistently demonstrated a partial inhibition of LF-killing; whereas I mmol/

Vol. 66 No. 2

LACTOFERRIN EFFECTS ON S. mutans

483

Discussion.
More than 30 years ago, Schade et al. (1949) originally described a role for carbonate in the binding of iron to transferrin. Since then, the role of anion-binding in the biological activities of the transferring has been the subject of additional investigations (Aisen et al., 1967; Aisen and Leibman, 1973). It is apparent from these studies that specific and high-affinity binding of iron does not occur unless a suitable anion is bound concomitantly (reviewed in Aisen, 1980). At the anion-binding site, several different anions may be acceptable. Bicarbonate (or possibly carbonate) is apparently a preferred anion for formation of the ternary complex; however, studies have shown that EDTA and oxalate may act as coordinate anions to promote the specific binding of iron (Aisen et al., 1967; Aisen and Leibman, 1973). Conversely, other anions, such as SCN acetate, and sulfate, may completely lack or exhibit only poor synergistic activity (Aisen et al., 1967). Data from the current investigations indicate that anionic solutions can impede LF bactericidal activity against S. mutans. Within the group of anions investigated, there was more than a 1000-fold difference in the molar concentrations of the anionic solutions necessary to inhibit bactericidal activity. The most effective anions were those that were previously demonstrated to be effective in occupying the anion-binding sites of the transferrin family of molecules. If the ability of anionic solutions or buffers to inhibit bactericidal activity were merely a reflection of ionic strength, it might be expected that anionic buffers would be relatively equivalent. This was not supported by these data. It is likely that the variable effect of the anions is a reflection of their ability to occupy the specific anion-binding sites on the LF molecule. This further suggests that S. mutans offers anionic sites that may serve as a point of attachment or target site for apo LF. The importance of a vacant anion-binding site on LF for antistreptococcal activity is indicated by our previous observation (Arnold et al., 1982). Occupation of this site, as a result of coordinate binding with iron, provides a LF molecule that does not attach to S. mutans, and thus, bacteria are not killed (Arnold et al., 1982). The relative importance of the anion target versus an iron target is suggested by the inability of exogenously added iron to inhibit bactericidal activity

TIME(HRS)
Fig. 5 Effects of catalase, lactoperoxidase, and BSA on LF bactericidal activity against S. mutans: A S. mutans only; A S. mutans + LF; o S. mutans + LF in catalase; * S. mutans + LF in lactoperoxidase; O S. mutans + LF in BSA. According to the text.

> 1.

competitively (Arnold et al., 1982). Although some anions showed blocking activity initially, the number of CFU was reduced with increasing incubation time (data not shown). In some instances, as with CO2 exposure (Fig. 3), the numbers of bacteria killed eventually reached the
same level as the LF-treated bacteria in air. These data may indicate that bactericidal activity is delayed rather than blocked by the carbonic anion. Two explanations are available to account for this observation. It is possible that the anion-binding site is initially occupied with the addition of exogenous anions, but it is subsequently displaced by the higher-affinity binding site on the bacteria. It is likely that, if bacteria have both exposed anionic and cationic sites in proximity to each other and available for high-affinity coordinate binding, the affinity of LF for these sites would exceed the LF affinity for exogenous anion. Another interpretation is that the anions do serve to inactivate a portion of the LF, and that the lag before detection of killing is a result of decreased availability of active LF. We have previously shown that a decrease in LF concentration results in an increase in the lag rather than a reduction in the rate of killing (Arnold et al., 1981, 1984). If it is hypothesized that anion-binding by the LF molecules occurs at the bacterial surface, followed by the bactericidal event, then the most likely explanation for our observation of CO2 inhibition of killing involves carbonic anion interacting with the LF anion-binding site, and thus, competitively inhib-

TIME(HRS)

Titration of SCN- effect on LF-killing: A S. mutans in 10 mmol/L NaSCN; A S. mutans + LF; O S. mutans + LF in tO mmol/L SCN-; E S. mutans + LF in 5 mmol/L SCN-; H S. mutans + LF in t mmol/L SCN -; 0 S. mutans + LF in 0.1 mmol/L SCN According to the text.

Fig. 6

L and 0.1 mmol/L SCN- consistently resulted in a small but detectable increase in LF killing.

484

LASSITER

et af.

J Dent Res Februorv 1987

iting attachment to bacterial anionic sites. The ability of compounds to serve as synergistic anions (at least with transferrin) is a result of strategically located carboxilic acid groups with a stereochemical carbonate-like configuration. For example, oxalate, EDTA, and citrate each have exposed carboxyl groups, and it is likely that the efficient neutralization of LF activity with the antibiotic nalidixic acid is related to an exposed carboxyl group which occurs on this molecule. It has been shown that, for some enzymatic sites, arginyl residues serve as positively charged recognition sites in binding carboxyl groups on substrates. Since the terminal sequence at the NH2 end of LF is GLY-ARG-ARG-ARG-ARG (Metz-Boutique et al., 1984), the arginine group could function in an analogous manner for LF-carboxy-anion-binding at a bacterial site. The potential for OH' involvement in LF-mediated killing is especially attractive. The OH is highly reactive, and as such, a mechanism for focusing the site of action would likely be essential for efficient killing of bacteria. Based on the anion target data and the requirement for coordinate iron-binding, it could be hypothesized that the iron prosthetic group necessary for the Haber-Weiss reaction may be targeted by the coordinate association with anion on the bacterial surface. This would require generation of hydrogen peroxide and possibly, though not necessarily, superoxide at the site of LF-binding. The necessity for oxygen involvement in LF-killing lends support for this or a similar mechanism of bactericidal activity. Catalase effectively blocked LF-killing of S. mutans. Superoxide dismutase-, on the other hand, had no effect on LF-killing. Heat inactivation of catalase eliminated its ability to block LF-killing. Implication of hydrogen peroxide involvement in the LF mechanism, apparently, was compromised by the observation that bovine serum albumin was equally effective at reducing the bactericidal activity. However, the inhibitory effect on LFkilling may be explained if LF has ' pseudoperoxidase" activity (suggested by SCN- enhancement). The bovine serum albumin has one sulfhydryl group per mole that may be oxidized by peroxidation catalyzed by LF. In addition, enzymatic inactivation of LPO with inhibitory concentrations of hydrogen peroxide did not reverse its inhibitory effect (data not shown). These data suggested that it is more likely that this is a nonspecific protein effect rather than an enzymatic effect. The apparent enhancement of LF-killing by SCN-, on the other hand, may lend support to radical involvement. These experiments were originally designed to consume hydrogen peroxide. SCN- and LPO were added to S. mutans in tandem with LF, with the expectation that the peroxidase system would consume available hydrogen peroxide and, in addition, statically inhibit the bacteria. At higher concentrations of SCN-, bactericidal activity was reduced, while lower concentrations enhanced killing. These results suggest that SCN- enhancement does not involve its chaotropic properties. The efficacy of OH as a toxic product relies on generation at a specific target site due to its rapid reactivity. SCN- may, therefore, be able to scavenge the OH and direct it as a more specific vehicle, the thiocyanate radical (Hoogendoorn et al., 1977). The inability of radical scavengers to inhibit bactericidal activity does not eliminate the possible involvement of OH in LF killing. Radical damage may occur at a site which would not be physically accessible to these scavengers. The requirement for iron in OH generation proposed by Ambruso and Johnston (1981) and Bannister et al. (1982) was an apparent contradiction to our previous results (Arnold et al., 1977), which indicated that the coordinate association of iron eliminated the bactericidal effect of LF. High-affinity association versus low-affinity association of the bound iron may partially explain the differences seen. Low-affinity association
of iron with transferrin in the absence of anion has been re-

ricidal assay

ported (Bates and Schlabach, 1973). In the S. mutans bactesystem, both ferrous iron and hydrogen peroxide would potentially be available from the bacteria. Repine et al. (1981) demonstrated that Staphylococcus aureus grown with FeSO4 were more sensitive to exogenous hydrogen-peroxidekilling than controls grown without additional iron. S. aureus grown under iron-depleted conditions were found to be even less sensitive to killing by exogenous hydrogen peroxide. Iron on the surfaces of the bacteria was reported to be in the ferrous state, and iron associated with S. mutans would, presumably, also be in the ferrous state. Arnold et al. (1981, 1984) have previously demonstrated that iron starvation of S. mutans renders these cells more resistant to LF-killing. A role for bacterially-associated iron in the bactericidal process would be supported by this finding. The availability of iron, suitable anion, and hydrogen peroxide would appear to be essential to LF bactericidal activity in this mechanistic model. Anion and ferrous iron associated with the bacterial cell surface would provide specific target sites for LF attachment. The capacity for competitive blocking of LF bactericidal activity by anionic buffers indicates the important role of anion. In contrast, ferrous and ferric ions, ferritin and hematin did not competitively block killing (Arnold et al., 1982). Certain factors point to a requirement for bacterial metabolism in LF-killing of S. mutans. Under normal growth conditions, S. mutans 10449 has been shown to be a poor producer of hydrogen peroxide (Carlsson et al., 1983). However, a requirement for metabolism by the bacteria and for increased hydrogen peroxide production may explain why LF-killing was found to be decreased at temperatures below 370C. The current data would support a need for bacterial metabolism for generation of hydrogen peroxide and possibly superoxide anion. According to our current hypothesis, the bacteria would actively participate in their own demise through the production of metabolic byproducts associated with aerobic metabolism.

Acknowledgments.
The authors wish to thank C.A. Zinney for technical assistance, and M.H. Hall and B.A. Couch for their secretarial contributions. This work was in partial fulfillment of the MS degree of M.O.L. in Microbiology and Immunology at the University of Louisville.

REFERENCES

AISEN, P. (1980): The Transferrins. In: Iron in Biochemistry and Medicine II, A. Jacobs and M. Warwood, Eds., New York: Academic Press, pp. 87-129. AISEN, P.; AASA, R.; MALMSTROM, B.G.; and VANNGARD, T. (1967): Bicarbonate and the Binding of Iron to Transferrin, J Biol Chem 242:2484-2490. AISEN, P. and LEIBMAN, A. (1973): The Role of Anion-binding Site of Transferrin in its Interaction with the Reticulocyte, Biochim Biophys Acta 304: 797-804. AMBRUSO, D.R. and JOHNSTON, R.B. (1981): Lactoferrin Enhances Hydroxyl Radical Production by Human Neutrophil Particulate Fractions, and an Enzymatic Generating System, J Clin Invest 67:352-360. ARNOLD, R.R.; BREWER, M.; and GAUTHIER, J.J. (1980): Bactericidal Activity of Human Lactoferrin: Sensitivity of a Variety of Microorganisms, Infect Immun 28:893-898. ARNOLD, R.R.; COLE, M.F.; and McGHEE, J.R. (1977): A Bactericidal Effect for Human Lactoferrin, Science 197:263-265. ARNOLD, R.R.; RUSSELL, J.E.: CHAMPION, W.J.: BREWER,

Vol. 66 No. 2

LACTOFERRIN EFFECTS ON S. mutans

485

M.; and GAUTHIER, J.J. (1981): Bactericidal Activity of Human Lactoferrin: Influence of Physical Conditions and Metabolic State of Target Microorganisms, Infect Immun 32:655-660. ARNOLD, R.R.; RUSSELL, J.E.; CHAMPION, W.J.; BREWER, M.; and GAUTHIER, J.J. (1982): Bactericidal Activity of Human Lactoferrin: Differentiation from the Status of Iron Deprivation, Infect Immun 35:792-799.

ARNOLD, R.R.; RUSSELL, J.E.; DEVINE, S.M.; ADAMSON, M.; and PRUITT, K.M. (1984): Antimicrobial Activity of the Secretory Innate Defense Factors Lactoferrin, Lactoperoxidase, and Lysozyme. In: Cariology Today, B. Guggenheim, Ed., Basel: Karger
Press, pp. 75-88. BANNISTER, J.; BANNISTER, R.W.; HILL, H.; and THORNALLEY, R. (1982): Enhanced Production of Hydroxyl Radicals by the Xanthine-Xanthine Oxidase System in the Presence of Lactoferrin, Biochim Biophys Acta 715:118-120. BATES, G.W. and SCHLABACH, M.R. (1973): A Study of the Anion Binding Site of Transferrin, FEBS Lett 33:289-292. BLACKBERG, L. and HERNELL, 0. (1980): Isolation of Lactoferrin from Human Whey by a Single Chromatographic Step, FEBS Lett 109:180-184. CARLSSON, J.; IWAMI, Y.; and YAMADA, T. (1983): Hydrogen Peroxide Excretion by Oral Streptococci and Effect of Lactoperoxidase-Thiocyanate-Hydrogen Peroxide, Infect Immun 40:70-80. HOOGENDOORN, H.; PIESSENS, J.P.; SCHOLTES, W.; and STODDARD, L.A. (1977): Hypothiocyanite Ion; the Inhibition

Formed by the System Lactoperoxidase-Thiocyanate-Hydrogen Peroxide, Caries Res 11:77-84. MASSONS, P. and HEREMANS, J. (1966): Studies on Lactoferrin, the Iron Binding Protein of Secretions, Prot Biol Fluids 14:115124. MASSONS, P.; HEREMANS, J.; and SCHONNE, E. (1969): Lactoferrin, an Iron-binding Protein in Neutrophilic Leukocytes, J Exp Med 130:643-658. MAZURIER, J. and SPIK, G. (1980): Comparative Study of the Ironbinding Properties of Human Transferrins, Biochim Biophys Acta 629:399-408. METZ-BOUTIQUE, M.-H.; JOLLES, J.; MAZURIER, J.; SCHOENTGEN, F.; LEGRAND, D.; SPIK, G.; MONTREUIL, J.; and JOLLES, P. (1984): Human Lactoferrin: Amino Acid Sequence and Structural Comparisons With Other Transferrins, Eur J Biochem 145:659-676. REPINE, J.E.; FOX, R.B.; and BERGER, E.M. (1981): Hydrogen Peroxide Kills S. aureus by Reacting with Staphylococcal Iron to Form Hydroxyl Radical, J Biol Chem 256:7094-7096. SCHADE, A.L.; REINHART, R.W.; and LEVY, H. (1949): Carbon Dioxide and Oxygen in Complex Formation with Iron and Siderophilin, the Iron-Binding Component of Human Plasma, Arch Biochem 20:170-172. WEINBERG, E. (1978): Iron and Infection, Microbiol Rev 42:45-66. WINTERBOURNE, C.C. (1983): Lactoferrin-catalyzed Hydroxyl Radical Production. Additional Requirement for a Chelating Agent, Biochem J 210:15-19.