Antimicrobial Properties of Human Lysozyme Transgenic Mouse Milk

ELIZABETH A. MAGA,l GARY B. ANDERSON,! JAMES S. CULLOR,2 WAYNE SMITH,2 AND JAMES D. MURRAY1,2*

lDepartment of Animal Science and 2Department of Population Health and Reproduction, School of Veterinary Medicine, University of California-Davis, Davis, California 95616, USA

MS 97-20: Received 5 February 1997/Accepted 23 April1997

ABSTRACT Downloaded from http://meridian.allenpress.com/jfp/article-pdf/61/1/52/1669839/0362-028x-61_1_52.pdf by guest on 23 September 2021

The antimicrobial properties of standard human lysozyme and the milk of transgenic mice expressing human lysozyme were investigated using bacterial strains important to the dairy industry. Standard human lysozyme was found to be effective at significantly slowing the growth of the milk cold-spoilage organism Pseudomonas fragi (P < 0.001). of a clinical isolate of the mastitis-causing organism Staphylococcus aureus (P < 0.005), and of a nonpathogenic strain of E. coli (P < 0.05). Milk from transgenic mice secreting human lysozyme in their milk at an average concentration of 0.38 mg/ml was found to be bacteriostatic against the cold-spoilage organisms Pseudomonas fragi and Lactobacillus viscous and a mastitis-causing strain of Staphylococcus aureus, but not against a pathogenic strain of E. coli. These results demonstrate that transgenic animals producing human lysozyme in their milk can affect the microbial nature of milk.

Transgenic animals are a useful experimental model milk, which is high when compared with that in the milk of because they are an indicator of the in vivo performance of a cows (0.13 flg/ml), goats (0.25 flg/ml), sheep (0.10 flg/ml), foreign gene product in a biological system. Expression of and rodents (<0.00 flg/ml) (3). the gene of interest is controlled by DNA promoter elements Due to its inherent positive charge and antimicrobial that direct where and when the gene will be expressed in the nature, if human lysozyme were present in bovine milk it animal. Expression of a trans gene in the mammary gland of could affect the physical and functional properties of the an animal requires the use of a promoter and regulatory milk protein system and the microbial status of the milk. For regions of a milk protein gene, i.e., sequences that direct these reasons, transgenic mice as model systems have been gene expression only in the mammary gland and only during generated that produce active human lysozyme in their milk lactation. at concentrations ranging from 0.25 to 0.71 flg/fll (14, 15). Most of the research done with targeting trans gene The presence of human lysozyme in mouse milk signifi- expression to the mammary gland of an animal has been cantly altered several of the physical and functional proper- directed at producing biologically important and active ties of the milk protein system, including decreasing rennet proteins (such as pharmaceuticals) in the milk of a trans- clotting time and increasing curd strength (15). These genic animal with the intent of recovering the protein of findings if extrapolated to dairy cattle could lead to a quicker interest from the milk (1, 24). It is now also possible to use cheese-making process and a firmer cheese, both of which transgenic animals to alter the properties and composition of could be of benefit to the dairy industry. However, the the milk itself (16). In a transgenic system, new genes, or antimicrobial effect of human lysozyme remains to be modified milk protein genes, can be introduced and ex- characterized in these animals. If milk containing human pressed in the mammary gland in order to benefit the milk lysozyme is able to affect the microbial nature of milk, for protein system. More specifically, we are examining the instance if active lysozyme produced in milk is able to kill or transgenic addition of human lysozyme to bovine milk in to slow the growth of mastitis-causing bacteria or cold- order to affect the functional and physical properties of the spoilage organisms, the end result could be a safer supply of milk protein system and thereby alter the manufacturing milk. The knowledge of these potential effects could be of applications of milk and enhance the mammary gland host great use to the diary and milk-packaging industry. defense capability. Lysozyme is an antimicrobial protein that specifically Lysozyme is normally found in mammalian secretions, cleaves the 1,4-r3-D-linkage between N-acetylmuramic acid such as tears, saliva, and milk (13), and is the primary and N-acetylglucosamine found in the peptidoglycan layer product of macrophages (8). Among their known functions, of bacterial cell walls (18). Cleavage of the protective lysozymes serve as a defense mechanism against bacterial peptidoglycan layer causes leakage of the cell's interior infection (7) or, in ruminants, for the digestion of intestinal components and results in cell lysis. Thus, because of its bacteria (5). Lysozyme is present at 400 flg/ml in human inherent antimicrobial action, the expression of increased lysozyme could help to reduce the overall level of bacteria in *Author for correspondence. Tel: (916) 752-3179; Fax: (916) 752-0175; the udder of the cow, in milk, or in milk products. E-mail: [email protected] Several microorganisms are susceptible to human and J. Food Prot., Vol. 61, No. 1 ANTIMICROBIAL PROPERTIES OF LYSOZYME TRANSGENIC MILK 53 bovine milk lysozymes and it has been suggested that human lysozyme was obtained from Sigma Chemical Company lysozyme is part of the inherent antimicrobial activity of (St. Louis, MO) and reconstituted in sterile water. milk (23). Lysozyme has been shown to be effective against Antibacterial assays. Four to five single colonies of bacteria bacteria causing food-borne disease and spoilage at concen- were inoculated in 50 m1 of nutrient broth (Ditco Laboratories, trations of 10 to 200 mg/liter (12). Lysozyme isolated from Detroit, MI) (cold spoilage) or Trypticase soy broth (TSB) (Media rainbow trout kidney was active against several bacterial Room; School of Veterinary Medicine, University of California, strains that cause mastitis, including Staphylococcus aureus, Davis) (mastitis causing) and grown overnight at 37°C. Of the E. coli, and two strains of Streptococcus (9). Human milk overnight culture, 1 ml was then inoculated in 50 ml of fresh has been shown to be antibacterial, with human milk being medium and grown to log phase. The culture was then centrifuged more effective at stopping the growth of Bordetella pertussis (2,000 X g, 10 min, 22°C) and the resulting bacterial pellet was than was bovine milk (19). Together, lysozyme and lactofer- washed two times (10 roM sodium phosphate buffer, pH 7.4) and resuspended in buffer. Ten microliters of bacterial suspension rin were inhibitory at concentrations normally occurring in adjusted to 1 X 106 CPU/ml were added to either 0, 10, or 50 Ilg of human milk (2) and the antibacterial activity of these two standard human lysozyme or 50 Jl! of 1:10 diluted skim or whole proteins are considered to be the most important factors in milk from both transgenic and nontransgenic mice. The amount of Downloaded from http://meridian.allenpress.com/jfp/article-pdf/61/1/52/1669839/0362-028x-61_1_52.pdf by guest on 23 September 2021 the nonspecific immunity of milk (11). Overall, it may be transgenic milk added corresponded to an average of 1.911g of total concluded that lysozyme can be active as an additive to human lysozyme present in the assay. The final reaction volume foods containing susceptible bacteria and act to lower the was adjusted to 100 111with buffer. The samples were incubated at levels of the bacteria or inhibit their initial growth. 37°C for 2 h. Cold-spoilage organisms were then serially diluted in Transgenic mice producing human lysozyme in their buffer and plated on nutrient agar. To construct a growth curve, milk would be good models to enable the study of the in vivo mastitis-causing organisms were diluted 1:100 in buffer at times 0, antimicrobial effect of lysozyme in milk. The direct addition 30, 60, 90, and 120 min after the start of incubation and plated on of human lysozyme to milk may not have the same activity Trypticase soy agar with a spiral plater. Plates were incubated (24 to 48 h, 37°C) and characteristic colonies enumerated. All assays or action as does lysozyme produced de novo in the were conducted in duplicate and repeated at least six times. mammary gland. The extent of the antimicrobial action of lysozyme on microorganisms important to the dairy industry Statistical analysis. The differences between the means of (dairy-related bacteria are defined here as those involved in the number of colonies (cold spoilage) or CPU per milliliter mastitis and the cold spoilage of milk), with the exception of (mastitis causing) obtained for standard human lysozyme (10 or 50 S. aureus, is not well known. Moreover, most studies have I1g) versus a control (0 I1g), transgenic skim milk versus nontransgenic control skim milk, and transgenic whole milk versus been conducted using chicken egg white or other nonhuman nontransgenic control whole milk were analyzed by using a lysozyme (9, 12, 17). The properties and behavior of chicken standard Student's t-test. lysozyme may not be completely comparable to those of human lysozyme, since the lysozymes may act (or not act) to RESULTS different extents on the same organism as lysozymes from different sources vary in their level of activity (13). Cold-spoilage organisms. Standard human lysozyme. The action of human lysozyme standard as well as a at levels of 10 and 50 Ilg slowed the growth of the specific lysozyme in milk has not been well studied with cold-spoilage organism P.fragi but not of L. viscous (Table respect to dairy-related bacteria. Here we report the effective- 1). The mean number of colonies formed on lysozyme- ness of human lysozyme standard and transgenic mouse containing plates was significantly lower than on control milk producing active human lysozyme on milk-important plates (P < 0.02, P < 0.001, respectively). The skim frac- bacterial strains. tion of transgenic mouse milk was also shown to have a significantly lower mean number of colonies than nontrans- MATERIALS AND METHODS genic skim milk for both microorganisms (P < 0.05). Trans- genic whole milk from the same mice did not significantly Transgenic mice and milk. Bovine usl-casein-human lysozyme DNA constructs, generation of transgenic mice, and mRNA decrease the number of colonies when compared with and protein analyses have been reported previously (14, 15). nontransgenic whole milk. Human lysozyme was found to Transgenic mice were from line HZ5, which is homozygous for the be more bacteriostatic against P.fragi than L. viscous (Table trans gene insert and expresses on average 0.38 mg of human 2). lysozyme per ml of milk. Control mice were nontransgenic mice Mastitis-causing organisms. Standard human lyso- from the same genetic background (JU) (6). Mice was collected and fractionated as described previously (15). All milk was stored zyme at levels of 10 and 50 Ilg slowed the growth of a at -70°C prior to use. All animals were housed in AAALAC- clinical isolate of S. aureus (Tables 2 and 3). The mean count approved facilities. (CFU per milliliter) was significantly lower than that for the control after 120 min of incubation. The whole, but not skim, Bacteria and human lysozyme standard Bacterial strains fraction of transgenic milk significantly decreased the num- were chosen on the basis of their relevance to the dairy industry. ber of CFU per milliliter compared with nontransgenic The organisms included a nonpathogenic E. coli J5, two cold- spoilage-causing microorganisms, Pseudomonas fragi (ATCC 4973) control milk after 90 min of incubation. Growth curves and Lactobacillus viscous (ATCC 337), and two clinical isolates of showed that transgenic milk was affecting the growth of S. mastitis-causing organisms, including Staphylococcus aureus and aureus after 90 min of incubation, while standard human an E. coli, isolated and typed by the California Dairy Foods Safety lysozyme acted after 30 min (Fig. 1). Research Laboratory, University of California, Davis. Standard None of the samples tested, either standard human 54 MAGA, ANDERSON, CULLOR, SMITH, AND MURRAY 1. Food Prot., Vol. 61, No.1

TABLE 1. Effect of standard human lysozyme and transgenic milk a) on cold-spoilage bacteria 1.4 Bacterial strain .LCCFU/ml Mean CFU per plate ± SD (n) *L10CFU/ml BL50CFU/ml E Lactobacillus Pseudomonas :3 viscous fragi u. () Sample" (ATCC 337) (ATCC4973) 0.4 hLZO/lg 88.38 ::'::59.7 (13) 69.00 ::'::22.9 (7) 0.2 hLZ 10 J.lg NDb 26.00 ::'::7.21 *c (3) o 60 90 120 hLZ 50 J.lg 103.20 ::'::61.8 (5) 24.40 ::'::6.6***c (5) o 30 Time (min) JU skim 71.22 ::'::16.6 (9) 158.72 ::'::43.3 (11) HZ5 skim 50.54 ::'::23.8**c (13) 112.36::':: 37.8**c (11) JU whole 84.08 ::'::38.7 (12) 137.20::':: 40.4 (10) Downloaded from http://meridian.allenpress.com/jfp/article-pdf/61/1/52/1669839/0362-028x-61_1_52.pdf by guest on 23 September 2021 HZ5 whole 79.50 ::'::33.7 (12) 116.09::':: 32.1 (11) b) a hLZ, standard human lysozyme; JU, non transgenic control mice; 3.5 HZ5, hLZ transgenic mice. 3 .•. Hw CFU/ml b ND, not determined. 2.5 BJwCFU/ml c * Standard hLZ was statistically different from control (P < 0.02); ** transgenic mouse milk was statistically different from nontrans- :3E 2 genic control milk (P < 0.05); *** standard hLZ was statistically ~ 1.5 different from control (P < 0.001). 1 0.5 TABLE 2. Comparative bacteriostatic activities of human lyso- o zyme and transgenic milk o 30 60 90 120 Time (min) Log reduction in CFU of organisms by: FIGURE 1. Growth curves of the effect of a) human lysozyme standard: LC, control, no added human lysozyme; LID, 10 flg of Organisms LlD L5D HZs HZw human lysozyme; L50, 50 flg of human lysozyme; b) human Cold spoilage lysozyme transgenic milk on the growth of Staphylococcus aureus: Lactobacillus viscous NDb -.101 .218 -.044 Hw, whole milk from HZ5 transgenic mouse; lw, whole milk from Pseudomonas fragi .378 .455 .146 -.013 nontransgenic control mouse. Mastitis causing Staphylococcus au reus .825 1.15 .242 .381 lysozyme or transgenic milk, had a significant effect on the Pathogenic E. coli .045 .276 .088 .089 Nonpathogenic E. coli .181 .245 growth of a mastitis-causing clinical isolate of E. coli (Tabte 3). Standard human lysozyme was found to be effective at a LIO, 10 J.lg human lysozyme standard; L50, 50 /lg human slowing the growth of nonpathogenic E. coli 15 at the 50 flg, lysozyme standard; HZ5s, skim transgenic mouse milk; HZ5w, but not the 10 flg level (Table 3). Milk was not tested on this whole transgenic mouse milk; log CFU reduction = log (CFU organism. Human lysozyme was more bacteriostatic against control T1201CFU lysozyme T120). b ND, not determined. S. aureus than it was against either E. coli strain (Table 2).

TABLE 3. Effect of human lysozyme and transgenic milk on mastitis-causing bacteria

Bacterial strain, mean CFU/ml ± SD (n)

Samplea Staphylococcus aureusb Pathogenic E. colib Nonpathogenic E. coli J5

hLZO/lg (4.33" ::'::3.05) X 1QS(7) (2.02 ::'::1.86) X 106 (6) (7.35 ::'::2.97) X 105 (6) hLZ 10 J.lg (6.13 ::'::2.35) X 104 **e (7) (1.82 ::'::1.42) X 106 (6) (4.84 ::'::2.03) X 1QS(6) hLZ 50 /lg (2.87 ::'::2.34) X 104 ***e (7) (1.07 ::'::.75) X 106 (7) (4.18 ::'::1.21) X 105 *e (6) JU skim (2.01 ::'::1.12) X 106 (7) (1.42 ::'::.55) X IQ7 (5) NDI HZ5 skim (1.1 ::'::.76) X 106 (7) (1.16 ::'::.52) X IQ7(6) ND JU whole (3.09 ::'::2.24) X 106 (7) (1.5 ::'::.78) X IQ7(5) ND HZ5 whole (1.29d::,:: .54) X 106 *e (7) (1.37 ::'::.66) X IQ7(5) ND

a hLZ, human lysozyme; JU, nontransgenic control mice; HZ5, high-expressing hLZ homozygous transgenic mice. b Clinical isolates. c After l20-min incubation. dAfter 90-min incubation. e * Standard hLZ was statistically different from control, and transgenic milk was different from control milk (P < 0.05); ** standard hLZ was statistically different from control (P < 0.01); *** standard hLZ was statistically different from control (P < 0.005). IND, not determined. J. Food Prot., Vol. 61, No. 1 ANTIMICROBIAL PROPERTIES OF LYSOZYME TRANSGENIC MILK 55

DISCUSSION zyme secreted in the transgenic mouse milk had a significant effect on the growth of L. viscous, while standard human We report the effects of human lysozyme standard and lysozyme did not show any effect. human lysozyme-containing transgenic mouse milk on sev- With human lysozyme-containing transgenic mouse eral milk-related strains of bacteria. Standard human lyso- milk at an average level of 1.9 Jlg, we have observed the zyme alone was effective at slowing the growth of the slowed growth of two cold-spoilage organisms and a cold-spoilage organism P. fragi but not of L. viscous. mastitis-causing S. aureus. P.fragi and L. viscous are two of Lysozymes in general tend to be more effective against the bacteria most commonly responsible for the cold spoil- gram-positive bacteria, as the outer membrane found in age of milk after pasteurization due to the fact that they are gram-negative bacteria is not present and lysozyme has free proteolytic and psychrotrophic. S. aureus is one of the most access to the peptidoglycan layer of the cell (4). However, in common and important pathogens causing mastitis (22). this case, human lysozyme was more effective against the Cure rates of S. aureus mastitis are low (50%) for new gram-negative bacteria. Standard human lysozyme was also intramammary infections (lMI), and even lower for chronic

able to slow the growth of S. aureus and, to a lesser extent, of Downloaded from http://meridian.allenpress.com/jfp/article-pdf/61/1/52/1669839/0362-028x-61_1_52.pdf by guest on 23 September 2021 IMI, due to the poor penetration of antibiotics in the nonpathogenic E. coli. This too may have been due to mammary gland (20). If transgenic cattle were generated differences in the gram-positive (S. aureus) versus gram- that produce human lysozyme in their milk, the microbial negative (E. coli) nature of these bacteria. nature of the milk and cow could be altered, which might The effect of lysozyme on S. aureus has been reported result in an increase in the shelf life of dairy products and previously with mixed results. Chicken lysozyme is effec- decrease in infection in the cow. This and previous results tive against S. aureus (17) while several other investigators (10, 14, 15, 16) indicate that transgenic animals are a viable have noted that S. aureus, as well as some food-spoilage and option for the manipulation of important milk production food-borne disease-causing bacteria, are susceptible to lyso- traits that can address both milk quality and food safety zyme only with the addition of EDTA (9, 12). We observed issues. that human lysozyme is effective against S. aureus. The differences in the source of lysozyme may account for its ACKNOWLEDGMENTS activity in these experiments. Transgenic milk containing human lysozyme was also We kindly thank Jennifer Diamond for her technical assistance and Dairy Management Inc. for support of the research through the California effective against several bacterial strains studied. Skim but Dairy Research Foundation and the California Dairy Foods Research not whole transgenic milk was capable of slowing the Center. growth of both cold-spoilage organisms studied. It has also REFERENCES been observed that the lack of effect of lysozyme added to whole milk may be due to the formation of a protective 1. Bawden, W. S., R. J. Passey, and A. G. Mackinlay. 1994. The genes encoding the major milk specific proteins and their use in transgenic capsule around the bacteria induced by the milk (9). studies and protein engineering. Biotech. Gen. Eng. Rev. 12:89-137. Lysozyme can be inhibited by surface-active agents such as 2. Carlsson, A., L. Bjorck, and K. Persson. 1989. Lactoferrin and sodium dodecyl sulfate (SDS), alcohols, and fatty acids (21). lysozyme in milk during acute mastitis and their inhibitory effect in Delvotest P. J. Dairy Sci. 72:3166-3175. Whole mouse milk contains approximately 13% fat com- 3. Chandan, R. C., R. M. Parry, and K. M. Shahani. 1968. Lysozyme, pared to 3 to 4% in cows' milk. Therefore, it is likely that the lipase, and ribonuclease in milk of various species. J. Dairy Sci. presence of the fat inhibited the action of lysozyme, since it 51:60(H)07. was active in the skim fraction of milk that contained no fat. 4. Costerton, J. W., J. M. Ingram, and K. J. Chang. 1974. Structure and function of the cell envelope of G bacteria. Bacteriol. Rev. 38:87-110. However, only the whole fraction of transgenic milk was 5. Dobson, D. E., E. M. Prager, and A. C. Wilson. 1984. Stomach effective against S. aureus. This result could have been due lysozyme of ruminants, I. Distribution and catalytic properties. J. to a bacterial difference between S. aureus and the cold- BioI. Chern. 259:11607-11616. 6. Eklund, J., and G.E. Bradford. 1976. A note on short term selection spoilage organisms or to the fact that S. aureus was able to for birth weight in mice. Anim. Prod. 22: 127-130. grow better in whole control milk. It has been suggested that 7. Flemming, A. 1922. On a remarkable bacteriolytic element found in iflysozyme were present in the udder, it might be effective at tissues and secretions. Proc. R. Soc. Lond Ser. B BioI. Sci. 93: the start of infection before large numbers of bacteria can 306-317. 8. Gordon, S., J. Todd, and Z. A. Cohn. 1974. In vitro synthesis and accumulate (9). It should also be noted that the average secretion of lysozyme by mononuclear phagocytes. J. Exp. Med. concentration of human lysozyme present in the transgenic 139:1228-1248. milk was on average 0.38 mg/ml. This corresponds to the 9. Grinde, B. 1989. A lysozyme isolated from rainbow trout acts on mastitis pathogens. FEMS Microbiol. Lett. 60:179-182. equivalent of 1.9 Jlg of total human lysozyme being present 10. Gutierrez, A., E. A. Maga, H. M. Meade, C. E Shoemaker, J. E in each bacterial analysis. This is a much lower level than the Medrano, G. B. Anderson, and J. D. Murray. 1996. Alteration of 10 and 50 Jlg of standard human lysozyme that were used for physical characteristics of milk from bovine kappa-casein transgenic these experiments, and yet the transgenic milk still had a mice. J. Dairy Sci. 79:791-799. 11. Hennart, P. E, D. J. Brasseur, J. B. Delogne-Desnoeck, M. M. slight effect on some of the bacteria analyzed. This observa- Dramaix, and C. E. Robyn. 1991. Lysozyme, lactoferrin, and secretion suggests that other factors endogenously present in milk tory immunoglobulin A content in breast milk: influence of duration of may contribute to the antimicrobial effect of lysozyme and lactation, nutritional status, prolactin status, and parity of mother. Am. that, if human lysozyme were present in milk at higher J. Clin. Nutr. 53:32-39. 12. Hughey, V. L., and E. A. Johnson. 1987. Antimicrobial activity of levels, more of an effect might be observed. This suggestion lysozyme against bacteria involved in food spoilage and food borne is further reenforced by the observation that human lyso- disease. Appl. Environ. Microbiol. 53:2165-2170. 56 MAGA, ANDERSON, CULLOR, SMITH, AND MURRAY J. Food Prot., Vol. 61, No.1

13. Jolles, P., and J. Jolles. 1984. What's new in lysozyme research. Mol. human milk on Bordetella pertussis. FEMS Microbiol. Lett. 70: Cell. Biochem. 63:165-189. 179-182. 14. Maga, E. A., G. B. Anderson, M. C. Huang, and J. D. Murray. 1994. 20. Sanchez, M. S., C. W. Ford, and R J. Yancy, Jr. 1994. Effect of tumor Expression of human lysozyme mRNA in the mammary gland of necrosis factor-a, interleukin-I13, and antibiotics on the killing of transgenic mice. Transgen. Res. 3:36--42. intracellular Staphylococcus au reus. J. Dairy Sci. 77: 1251-1258. 1'5. Maga, E. A., G. B. Anderson, and J. D. Murray. 1995. The effect of 21. Smith, G. N., and C. Stroker. 1949. Inhibition of crystalline lysozyme. mammary gland expression of human lysozyme on the properties of Arch. Biochem. 21:383-394. milk in transgenic mice. J. Dairy Sci. 78:2645-2652. 22. Trinidad, P., S. C. Nickerson, and D. G. Luther. 1990. Antimicrobial 16. Maga, E. A., and J. D. Murray. 1995. Mammary gland expression of susceptibilities of staphylococcal species isolated from mammary trans genes and the potential for altering the properties of milk. glands of unbred and primigravid dairy heifers. J. Dairy Sci. Biotrechnology 13:1452-1457. 73:357-362. 17. Pellegrini, A., U. Thomas, R. von Fellenberg, and P. Wild. 1992. Bactericidal activities of lysozyme and aprotinin against gram- 23. Vakil, J. R, R C. Chandan, R. M. Parry, and K. M. Shahani. 1970. negative and gram-positive bacteria related to their basic character. J. Susceptibility of several microorganisms to milk lysozymes. J. Dairy Appl. Bacteriol. 72:180-187. Sci. 52:1192-1197. 18. Phillips, D. C. 1966. The three-dimensional structure of an enzyme 24. Wilmut, I., A. L. Archibald, M. McClenaghan, J. P. Simons, C. B. A. molecule. Sci. Am. 215:78-90. Whitelaw, and A. J. Clark. 1991. Production of pharmaceutical Downloaded from http://meridian.allenpress.com/jfp/article-pdf/61/1/52/1669839/0362-028x-61_1_52.pdf by guest on 23 September 2021 19. Redhead, K., T. Hill, and B. Mulloy. 1990. Antimicrobial effect of proteins in milk. Experientia 47:905-912.