Comparative studies on the chemical and immunochemical properties of human milk, human pancreatic juice and bovine milk lactoferrin

Comp. Biochem. Physiol. Vol. 78B, No. 3, pp. 575-580, 1984 Printed in Great Britain

0305-0491/84 $3.00+0.00 © 1984PergamonPress Ltd

COMPARATIVE STUDIES ON THE CHEMICAL A N D IMMUNOCHEMICAL PROPERTIES OF H U M A N MILK, H U M A N PANCREATIC JUICE A N D BOVINE MILK LACTOFERRIN Crn-Stm WANO,*t WAI-YE~.CnAN,:~ and HANS U. KLOER~f tLaboratory of Lipid and Lipoprotein Studies, Oklahoma Medical Research Foundation and :~Departmentof Pediatrics, Oklahoma Children's Memorial Hospital, 825 N.E. 13th. Oklahoma City, OK 73104, USA (Received 4 January 1984)

Abstract--l. Lactoferrin from human milk, pancreatic juice and bovine milk were purified by heparinSepharose affinity chromatography procedure. 2. Urea-sodium dodecylsulfate-polyacrylamide gel electrophoresis of these lactoferrin preparations indicated similar molecular weights (80,000). 3. Metal analyses showed that lactoferrin of bovine milk contained the highest amount of iron while lactoferrin of human milk and human pancreatic juice were similar in content of iron, approximately four-fold lower than bovine milk. All these lactoferrin preparations were also found to contain minor amounts of copper and manganese. 4. Double diffusion analyses indicated that lactoferrin of human milk was immunochemicallyidentical to lactoferrin of human pancreatic juice. On the other hand, immunochemically,bovine milk lactoferrin was not identical to human milk lactoferrin. 5. Sequenceanalyses of human milk and pancreatic juice lactoferrin indicated that they shared the same N-terminal sequence for the 16 residues analyzed. Although human milk and bovine milk lactoferrin had sequencehomologies, bovine milk lactoferrin had a closer homology to ovotransferrin than to human milk lactoferrin.

functional and nutritional role of milk proteins, we have modified their procedure for a more efficient purification of lactoferrin from human skim milk. In our studies, we have observed that the procedure is not only useful for the purification of human milk lactoferrin, it is also applicable for the purification of lactoferrin from human pancreatic juice and bovine skim milk. In this paper, we wish to describe the comparative studies on the chemical and immunochemical properties of these three lactoferrin preparations.

INTRODUCTION

Lactoferrin has been isolated from human milk (Johansson, 1960), bovine milk (Groves, 1960), and has been detected in human pancreatic juice (Colomb et al., 1974; Multigner et al., 1981). Partial amino acid sequence analyses have revealed the homology between human milk lactoferrin and transferrin (JollOs et al., 1976). The study by Metz-Boutigue et al. (1981) had also revealed a structural homology between human milk lactoferrin and human ceruloplasmin. Thus, these metalloproteins are not only functionally but structurally related to each other. Experiments in vitro indicated that lactoferrin is an important component of the bacteriostatic system present in human milk (Oram and Reiter, 1968; Kirkpatrick et al., 1971). Lactoferrin, by binding iron, prevents growth of iron requiring bacteria in the milk. It has also been reported that lactoferrin acts as an inhibitor of iron and ascorbic acid induced lipid peroxidation (Gutteridge et al., 1981). Recently, Bl/ickberg and Herneli (1980) described a simple procedure using heparin-Sepharose column chromatography for the purification of lactoferrin from human whey. Because of our continued interest in the

MATERIALS AND METHODS

Materials Unless otherwise stated, all chemicals including human serum transferrin and ovalbumin were purchased from Sigma Chemical Company. Cross-linked Sepharose 4B was obtained from Pharmacia. Pig intestinal mucosa heparin was obtained from ICN Pharmaceuticals, Inc. Commercial antiserum against human milk lactoferrin was obtained from Behring Diagnostics. Nitric acid (Ultrex grade) was obtained from Baker.

*Correspondence to be addressed to: Chi-Sun Wang, Ph.D., Associate Member, Laboratory of Lipid and Lipoprotein Studies, Oklahoma Medical Research Foundation, 825 N.E. 13th Street, Oklahoma City, OK 73104, USA (Tel: 405-271-7284). 575

Preparation of skim milk Human milk obtained in the second month post-partum was centrifuged at 25,000rpm at 4°C for 30min in a Beckman SW 27 swinging bucket rotor. The infranatant fraction was stored in 200ml aliquots at -20°C. Unpasteurized cow's milk was obtained from a dairy farm. The milk (4.5 liters) was centrifuged at 4000rpm at 4°C for 30 min in a Beckman J-6 swinging bucket centrifuge. The infranatant fraction (skim milk) was stored at -20°C.

576

CHI-SuN WAYG et al.

Human pancreatic juice Human pancreatic juice was obtained from a 60-year-old patient with chronic pancreatitis and a pancreatic fistula. Prior to surgical revision of the fistula, 15 portions of pancreatic juice (100-200ml) were collected by frequent emptying of the collection bag for the juice and were immediately frozen and stored at -20°C. The total volume of pancreatic juice thus obtained was 2 liters with an amylase and lipase activity of 12,000 i.u./1 and 5,000i.u./l, respectively.

RESULTS

Purification o f lactoferrin .[?om human milk, human pancreatic .juice and bovine milk

The electrophoresis of the lactoferrin preparations was performed in 7~o acrylamide gels in the presence of 0.1~, sodium dodecyl sulfate and 8 M urea according to a previously described procedure (Carey et al., 1975).

The procedure for the purification of lactoferrin from human milk, human pancreatic juice and bovine milk was identical and was performed at 4'~C. Human skim milk, bovine skim milk or human pancreatic juice (150-200ml) was thawed at 37"C and filtered through a Buchner funnel using Whatman No. I paper for the removal of the insoluble material. The soluble fraction was then applied onto a heparinSepharose column (2 × 10cm). After applying the sample, the column was washed with 300ml of N H 4 O H - H C 1 buffer, pH 8.5 (referred to as ammonium buffer) followed by 300 ml of the same buffer containing 0.3 M NaCI. Lactoferrin was eluted from the column with the a m m o n i u m buffer containing 0.72 M NaCI. The elution of lactoferrin was monitored for the absorbance at 280 nm. The combined peak fraction (80-100 ml) was dialyzed against distilled water (2 x 14 liters) and lyophilized. The heparin-Sepharose was regenerated by first washing with 500 ml of a m m o n i u m buffer containing 2 M NaCI followed by equilibrating the column with 1 liter of a m m o n i u m buffer. Approximately 60rag, 5 mg and 15 mg of lactoferrin could be isolated from each 100 ml of human skim milk, human pancreatic juice and bovine skim milk, respectively.

Amino acid analyses

Characterization of lactoJ~,rrin

Protein samples for the amino acid analyses were hydrolyzed with 6 M HCI in evacuated, sealed tubes at 110°C for 24, 48 and 72 hr. Amino acid analyses and performic acid oxidation were carried out as previously described (Lee and Alaupovic, 1970). Tryptophan was determined according to the procedure of Liu and Chang (1971).

By urea-sodium dodecylsulfate polyacrylamide gel electrophoresis (Fig. 1), they had similar electrophoretic mobilities with an apparent molecular weight of 80,000. The amino acid compositions of these three lactoferrin preparations are similar as shown in Table 1. Each preparation had an unusually high half-cystine content (3.2 4,5mol 0o). Metal analyses indicated the presence of iron in all lactoferrin preparations. Bovine milk lactoferrin contained the highest amount of iron, while lactoferrin of human milk and human pancreatic juice contained similar amounts of iron, approximately four-fold lower than bovine milk lactoferrin (Table 2). In all lactoferrin preparations, minor amounts of copper and manganese were detected.

Immunochemical methods Antisera to lactoferrin of human milk, bovine milk and human pancreatic juice were prepared by immunizing white New Zealand rabbits with the purified lactoferrin preparation at weekly intervals for 4 weeks. One ml of the purified lactoferrin preparations (0.2-0.5mg/ml) in physiological saline was mixed with an equal volume of complete Freund's adjuvant (Difco Lab, Detroit, MI, USA) and injected intraperitoneally. The rabbits were bled by cardiac puncture. Precipitating antibodies were produced 4-6 weeks after the initial injection. Double immunodiffusion was performed on glass slides (25 × 75 mm) coated with 1~o agarose using barbital buffer, pH 8.6, ionic strength 0.1 Plates were allowed to develop for 24-30 hr at room temperature.

Polyacrylamide gel electrophoresis

Metal analysis Lyophilized protein (l mg) was dissolved in 1 ml of 0.1 N nitric acid. Each sample was analyzed in triplicate for each element using a flameless atomic adsorption spectrophotometer model 703 equipped with an HGA 500 graphite furnace and an AS-1 autosampler. The following programs were used. For the copper determination wavelength, 324.7 nm; slit width, 0.7 nm; drying at l l0°C for 40sec; ashing at 1000°C for 40sec; and atomization at 2600°C for 5 sec. A background correction was used. For iron: wavelength, 248.3 nm; slit width, 0.2 nm; drying at II0°C for 20sec; ashing at 1000°C for 20sec and atomization at 2600°C for 5 sec. For manganese: wavelength, 279.5 nm; slit width, 0.7 nm; drying at 120°C for 60sec; ashing at 110ff'C for 20 sec; and atomization at 2600' C for 6 sec.

Protein analysis Samples for protein concentrations were determined according to Wang and Smith (1975).

N-terminal analysis The N-terminal sequence analysis was performed by a Beckman Sequencer Model 890C. A dilute Quadrol program (Brauer et al., 1975) was used with 5 mg polybrene as carrier. The conversion of anilinothiazolinone to PTHamino acids was accomplished with methanolic-HCl (Tarr, 1975) in a Sequemet P-6 Autoconverter. The PTH-amino acids were identified using a Waters Associates HPLC with a reversed phase 5 micron C-18 column (radial compression) as described by Takahashi et al. (1983) and independently identified by thin-layer chromatography.

Immunochemical studies Immunochemical analyses of the antisera specific for lactoferrin of human milk and human pancreatic juice as well as the commercial antiserum to lactoferrin of human milk showed a reaction of identity when reacted against purified lactoferrin from human milk (Fig. 2A) and human pancreatic juice (Fig. 2B). Similarly, the lactoferrin of human milk and human pancreatic juice also gave a reaction of identity when tested against antisera specific for lactoferrin of human milk or human pancreatic juice (pattern not shown). These results suggest that lactoferrin of human milk and human pancreatic juice are immunochemically identical. In contrast to the immunochemical identity between lactoferrin of human milk and pancreatic juice, there was no immunochemical precipitin line formation between bovine milk lactoferrin and the antiserum to human milk lactoferrin or between

Lactoferrin

a

b

577

c

Fig. 1. Urea-sodium dodecylsulfate polyacrylamide gel electrophoresis pattern of (a) human milk lactoferrin, (b) human pancreatic juice lactoferrin and (c) bovine milk lactoferrin. human milk lactoferrin and the antiserum to bovine milk lactoferrin (Fig. 2C). Both lactoferrin of human milk (Fig. 2D) and pancreatic juice (Fig. 2E) gave a Table 1. Amino acid composition of human milk lactoferrin (HML), human pancreaticjuice lactoferrin (HPL) and bovine milk lactoferrin (BML) I~L

HPL

BHL

reaction of non-identity when compared with lactoferrin of bovine milk. Thus, the results indicated that human lactoferrin is immunochemically distinct from bovine lactoferrin.

N-terminal sequence analyses Lactoferrin of human milk, human pancreatic juice and bovine milk were analyzed for their N-terminal sequence by autosequencer. The 16 residue N-

mole % Lysine

6.44

5.98

7.75

HistidEne

1.19

1.88

1.38

Arginine AsparCic acid Threonine Serine Glutamic

acid

Pre!ine Glycine Alanine Half-cystine ValEne Hethionine Iaoleucine Leucine Tyrosine Phenytatanine Tryptophan

5.49

6.59

5.82

10.84

10.96

9.72

5.32 6.70

4.65 5.32

5,22 5.02

10.41

9.96

9.84

4.30 7.71 8.88 4.49

5.00 8.27 8.44 3.17

4.46 7.43 9.79 3.84

7.28 0.2i 2,78 8,23 2.79 3,96 2.94

6.85 0.29 2,96 9.37 2.80 4.67 2.90

6.64 0,38 2,62 10,58 3.46 4.32 1.67

Table 2. Metal compositiona of human milk lactoferrin (HML), human pancreatic juice lactoferrin (HPL) and bovine milk lactoferrin (BML) H~L

HPL

BML

ng/mg Iron

243

~ 25

377 ± 18

1146 ± 34

Copper

17 t 2

58 t 14

37 t 13

Manganese

I0 t 6

13 t

29 ~ 19

6

~The data represent the analyses of purified lactoferrin from two separate preparations, each were determined in triplicates.

578

CHt-SuN WANGet al.

(A)

(C)

(B)

I

C

j

2

U

3

2

b

3 (D) 4

1 U

(E) 2 Cl

÷

C

4

1

Fig. 2. Double diffusion analyses of lactoferrin of human milk, human pancreatic juice and bovine milk. Lactoferrin preparations were dissolved in 50 mM barbital buffer, pH 8.6 at a concentration of 1mg/ml. The troughs in (A), (B), (C), (D), (E) are as follows: 1. antiserum against human milk lactoferrin; 2. antiserum against human pancreatic juice lactoferrin; 3. commercial antiserum against human milk lactoferrin; 4. antiserum against bovine milk tactoferrin; a. purified human milk lactoferrin; b. purified human pancreatic juice lactoferrin; c. purified bovine milk lactoferrin.

terminal sequence determined for lactoferrin of human milk and human pancreatic juice (Table 3) was identical to that reported by Joll6s et al. (1976) for human milk lactoferrin. The N-terminal sequence of bovine lactoferrin (Table 3) was found to have a closer homology to ovotransferrin (Williams et al., 1982) than to human milk lactoferrin. Heparin-Sepharose binding spec![icio' Despite the sequence homology between human milk lactoferrin and human serum transferrin as well as the homology between bovine milk lactoferrin and ovotransferrin, both transferrin and ovotransferrin were found not to be retained by heparin-Sepharose in contrast to that observed for human and bovine milk lactoferrin.

DISCUSSION

Because of the simplicity of the column chromatography procedure, it was possible not only to purify human milk lactoferrin in large quantities, it was also possible to isolate the less abundant lactoferrin from bovine milk and human pancreatic juice. Especially in the latter case, the presence of proteolytic enzymes in pancreatic juice could represent a major obstacle in the purification of lactoferrin by multiple step purification procedures. In the present study, we utilized step-wise elution rather than a salt gradient elution (Blfickberg and Hernell, 1980) which provides a well-defined separation of lactoferrin from those proteins having a lower affinity to heparin-Sepharose. In addition, our finding indicated

Table 3. Homology of the amino-terminal regions of human milk and pancreatic juice lactoferrin (this study), human serum transferrin (Joll+s et al., 1976), hen ovotransferrin (Williams et al., 1982) and bovine milk lactoferrin (this study) Human milk and human pancreatic juice lactoferrin a'b

Ar NH2-Gly-Arg-Arg-Ar ~ ~er-Va1-G|n-Trp- X -A[a-Val-Ser-Gln-Pro-Glu-

Human serum transferrin

NH2-Val-Pro-Asp-Lys-Thr-Val-Arg-Trp-Cys-Ala-Val-Ser-Glu-His-Clu-

Bovine milk [actoferrln a

NH2-Ala-Pro-Arg-Lys-Asn-Val-Arg-Tr p- X -Thr-lle-Ser-Gln-Pro-Glu-

Ovotransferrlm

NH2-Ala-Pro-Pro-Lys-Ser-V~l ~rg-Trp-Cys-Thr-lle-Ser-Ser-Pro-GluIle

~The protein was not carboxymethylated, therefore cysteine residue appeared as a missing residue by autosequence analyses. The residue was denoted as X bThe sequence of human milk and pancreatic juice lactoferrin and determined to be the same and identical to that reported by Joll~s et al. (1976).

Lactoferrin the use of skim milk as a starting material for the purification of lactoferrin is satisfactory and the preliminary removal of casein by isoelectric precipitation was not necessary. The comparative studies indicated that lactoferrin of human milk and human pancreatic juice had a similar apparent molecular weight, amino acid composition and an identical immunochemical reactivity. The N-terminal sequence of the 16 terminal residues were also identical. These data suggest that lactoferrin of human milk and human pancreatic juice are similar molecular entities. The sequence data of human milk lactoferrin is also identical to that reported by Joll6s et al. (1976). In contrast, despite the similarity of apparent molecular and amino acid composition of bovine milk iactoferrin and human lactoferrin, it was demonstrated that the lactoferrin of bovine milk could be differentiated from human lactoferrin by immunochemical analyses. Metal analyses indicated that human milk and pancreatic juice had a similar content of iron which was approximately four-fold less than bovine lactoferrin. Since the bacteriostatic effect of lactoferrin is inversely related to their extent of saturation with iron binding (Oram and Reiter, 1968; Kirkpatrick et al., 1971), it can be concluded that human milk and pancreatic juice lactoferrin have higher potential activity as anti-infectious reagent than bovine milk lactoferrin. All these lactoferrin preparations were found to contain minor amounts of copper and manganese, probably representing a minor binding specificity. The slight difference of copper content of human milk and human pancreatic juice lactoferrin might be due to the difference in tissue origin. Although the complete sequence data of human milk lactoferrin and human transferrin are not available at present, a great deal of information already has been gained from partial sequence data about the relationship of various iron-binding proteins (Joll6s et al., 1976; Metz-Boutigue et al., 1981) with respect to their sequence homology and evolutionary relationship. Recently, the complete nucleotide sequence of the chicken ovotransferrin cDNA has been deduced (Jeltsch and Chambon, 1982). Of the 705 positions of the whole protein, 605 can be matched by the peptide sequences and the possible discrepancies between the two methods has been pointed out (Williams et al., 1982). It is not unexpected to observe that the various iron binding proteins have sequence homologies; however, it was unexpected to find that bovine milk lactoferrin had a sequence homology closer to ovotransferrin than to human milk lactoferrin. Human lactoferrin had a 4-consecutive arginine sequence (residue 2-residue 5); such an unusual sequence feature has been found in nuclear protein and protamine-type molecules (Ando et al., 1967; Coelingh et al., 1972). The physiological significance of this unusual sequence feature of human lactoferrin is unknown. Since heparin represents a major component of intestinal mucosa (Jacques, 1980) as evidenced from the fact that this tissue represents a major source for the commercial heparin isolation, it is conceivable that lactoferrin-heparin interaction might be occurring in vivo by the direct interaction of lactoferrin with intestinal mucosa. We have envisioned that the

579

immobilization of lactoferrin by such interaction could have such advantages as the following: (1) the immobilization of lactoferrin might be coupled with membrane transport of iron; (2) the immobilization of lactoferrin might inhibit the attachment of bacteria to the mucosa layer; (3) the immobilization of lactoferrin might prevent or retard its degradation by proteolytic enzymes and (4) the retardation of the flow of this protein during the intestinal fluid movement and could result in a high concentration of lactoferrin in the intestine where it exerts its physiological function. These possible roles of the in vivo function of the lactoferrin-heparin interaction remains to be examined in future studies. Recently Cox et al. (1979) have demonstrated that the presence of a specific lactoferrin receptor in the human intestine; it is likely that the receptor might be heparin or heparin-like glycosaminoglycan. In contrast to lactoferrin, human transferrin and ovotransferrin were not found to have binding affinity to heparin-Sepharose. This may be related to the fact that the site of action of these two proteins are in the blood compartment rather than in the intestinal lumen. Acknowledgements--This work was supported by Grant

HD-14104 from the National Institute of Health and the

resources of the Oklahoma Medical Research Foundation. We thank Dr. W. J. McConathy for his criticism and suggestions in the preparation of this manuscript. We want to thank Dr. J. Tang and Ms. Azar Fesmire for their assistance in the sequence analyses. REFERENCES

Ando T. and Suzuki K. (1967) The amino acid sequence of the third component of clupeine. Biochim. biophys. Acta 140, 375-377. Bl~ickberg L. and Hernell O. (1980) Isolation of lactoferrin from human whey by a single chromatographic step. FEBS Lett. 109, 180-184. Brauer A. W., Margolies M. N. and Haber E. (1975) The application of 0.1 M Quardrol to the microsequence of proteins and the sequence of tryptic peptides. Biochemistry, N.Y. 14, 3029-3040. Carey C., Wang C.-S. and Alaupovic P. (1975) A critical evaluation of the proposal that serum apolipoproteins are the major constituents of the human erythrocyte membrane. Biochim. biophys. Acta 401, 6-14. Coelingh J. P., Monfoort C. H., Rozijn T. H., Leuven J. A. G., Schiphof R., Steyn-Parve E. P., Braunitzer G., Schrank B. and Ruhfus A. (1972) The complete amino acid sequence of the basic nuclear protein of bull spermatozoa. Biochim. biophys. Acta 285, 1-14. Colomb E., Estevenon J. P., Figarella C., Guy O. and Sarles H. (1974) Characterization of an additional protein in pancreatic juice of men with chronic calcifying pancreatitis. Identification to lactoferrin. Biochim. biophys. Acta 342, 306-312. Cox T. M., Mazurier J., Spik G., Montreuil J. and Peters T. J. (1979) Iron binding proteins and influx of iron across the duodenal brush border. Evidence for specific lactoferrin receptors in the human intestine. Biochim. biophys. Acta 588, 120-128. Groves M. L. (1960) The isolation of a red protein from milk. J. Am. Chem. Soc. 82, 3345-3350. Gutteridge J. M. C., Patterson S. K., Segal A. W. and Halliwell B. (1981 ) Inhibition of lipid peroxidation by the iron-binding protein lactoferrin. Biochem. J. 199, 259-261. Jacques L. B. (1980) Heparin-anionic polyelectrolyte drugs. Pharmac. Rev. 31, 99-166.

580

CHI-SUN WANG et al.

Jeltsch J.-M. and Chambon P. (1982) The complete nucleotide sequence of the chicken ovotransferrin mRNA. Eur. J. Biochem. 122, 291 295. Johanson B. (1960) Isolation of an iron-containing red protein from human milk. Acta chem. scand. 14, 510 512. Joll6s J., Mazurier J., Metz-Boutigue M.-H., Spik G., Montreuil J. and Joll6s P. (1976) FEBS Lett. 69, 27-31. Kirkpatrick C. H., Green I., Rich R. R. and Schade A. L. (1971) Inhibition of growth of Candida albicans by ironunsaturated lactoferrin: relation to host-defense mechanisms in chronic mucocutaneous candidiasis. J. Inject. Dis. 124, 539 544. Lee D. and Alaupovic P. (1970) Studies of the composition and structure of plasma lipoproteins. Isolation, composition, and immunochemical characterization of low density lipoprotein subfractions of human plasma. Biochemistry, N.Y. 9, 2244-2252. Liu T.-Y. and Chang Y. H. (1971) Hydrolysis of proteins with P-toluene-sulfonic acid. Determination of tryptophan. J. biol. Chem. 246, 2842 2848. Metz-Boutigue M.-H., Mazurier J., Joll~s J., Spik G., Montreuil J. and Joll+s P. (1981) The present state of the human lactoferrin sequence. Study and alignment of the

cyanogen bromide fragments and characterization of Nand C-terminal domains. Biochim. biophys. Acta 670, 243 254. Multigner L., Figarella C. and Sarles H. (1981) Diagnosis chronic pancreatitis by measurement of lactoferrin in duodenal juice. Gut 22, 350-354. Oram J. D. and Reiter B. (1968) Inhibition of bacteria by lactoferrin and other iron-chelating agents. Biochim. biophys. Acta 170, 351-365. Takahashi T., Schmidt P. G. and Tang J. (1983) Oligosaccharide units of liposomal cathepsin D from porcine spleen. Amino acid sequence and carbohydrate structure of the glycopeptide. J. biol. Chem. 258, 2819-2830. Tarr G. E. (1975) A general procedure for the manual sequencing of small quantities of peptides. Ana(vt. Biochem. 63, 361-370. Wang C.-S. and Smith R. L. (1975) Lowry determination of protein in the presence of Triton X-100. Analyt. Biochem. 63, 414-417. Williams J., Elleman T. C., Kingston 1. B., Wilkins A. G. and Kuhn K. A. (1982) The primary structure of hen ovotransferrin. Eur. J. Biochem. 122, 297 303.