The effect of trypsin on bovine transferrin and lactoferrin

Biochimica et Biophysica Acta, 446 (1976) 214-225

© Elsevier/North-HollandBiomedicalPress BBA 37430 THE EFFECT OF TRYPSIN ON BOVINE TRANSFERRIN AND LACTOFERRIN

J. H. BROCK, F. ARZABE, F. LAMPREAVE and A. PIIqEIRO Fundaci6n F. Cuenca Villoro, Gasc6n de Gotor, 4, Zaragoza-6 (Spain)

(Received March 9th, 1976) SUMMARY The iron-saturated and iron-free (apo) forms of bovine transferrin and lactoferrin were digested with trypsin and the digests analysed by column chromatography and electrophoresis. Both of the iron-saturated proteins were more resistant to proteolysis than the corresponding apoproteins, and iron-transferrin was more resistant than iron-lactoferrin. Digestion of iron-transferrin yielded two iron-binding fragments with molecular weights of 32 000 and 38 500 whereas apotransferrin yielded only the larger fragment. In digests of lactoferrin, up to five different fragments with molecular weights ranging from 25 000 to 52 700 were detected, there being no obvious qualitative difference between digests of iron-lactoferrin and apolactoferrin. The susceptibility of apolactoferrin to tryptic digestion was only slightly reduced when apolactoferrin was complexed with fl-lactoglobulin, suggesting that complex-formation is not a mechanism for protecting lactoferrin against intestinal degradation. There was no immunological cross reaction between bovine transferrin or its digestion products against anti-lactoferrin antiserum, or vice-versa.

INTRODUCTION The mammalian iron-binding proteins transferrin and lactoferrin, although similar in many respects, exhibit a number of important differences. The most fundamental is their histological distribution, transferrin being essentially a serum protein synthesised in the liver [1], whereas lactoferrin is found in external secretions and is normally absent from serum [2]. Transferrin is known to be involved in iron transport [3] but conclusive evidence of a similar role for lactoferrin is lacking. As a result of their ability to complex ferric iron, the iron-free (apo) forms of both proteins can exert an in vitro bacteriostatic effect [4-7], and in vivo experiments suggest that this mechanism may contribute to host defences against pathogenic bacteria [8, 9], thus indicating another possible physiological role of transferrin and lactoferrin. The situation thus bears a resemblance to that of the immunoglobulins, which occur in both the circulatory and secretory systems, but with a distinct class, IgA, predominating in secretions and occurring as only a minor component in serum. An important feature of IgA is its superior resistance to proteolysis in comparison with the IgG and lgM classes of serum [10], a property which may enable it to maintain its biological activity in locations such as the gastrointestinal tract. It might therefore be expected

215 by analogy that the secretory iron-binding protein lactoferrin would be more resistant to proteolysis than transferrin, its serum counterpart. Although there have been a few previous studies of proteolysis of proteins of the transferrin class [11-14] they give no indication of the relative susceptibilities of lactoferrin and transferrin, and little or no information about the biological properties of the fragments produced. In the present work the effect of trypsin (one of the most important enzymes in digestive proteolysis) on bovine transferrin and lactoferrin has been investigated. A distinctive feature of lactoferrin is its readiness to form complexes with other proteins, such as albumin and fl-lactoglobulin [15, 16]. Since this might constitute a mechanism whereby lactoferrin protects itself from proteolysis, the effect of trypsin on a lactoferrin/fl-lactoglobulin complex has also been studied. MATERIALS A N D METHODS

Transferrin Crude transferrin was prepared from bovine serum by the Rivanol (6,9-diamino-2-ethoxy acridine lactate) procedure described by Stratil and Spooner [17], or purchased from Miles Laboratories, Slough, England (Lot 3, stated purity 88 ~). No differences were observed in results obtained with materials from the two sources. Subsequent purification was effected by chromatography on DEAE-Sephadex A50 (Pharmacia, Uppsala, Sweden) equilibrated with 0.05 M Tris/0.05 M NaC1, pH 7.6, eluting first with the equilibrating buffer and then with a gradient of 0.05-0.2 M NaCI in 0.05 M Tris, pH 7.6. The transferrin thus obtained still contained a contaminant, presumed to be haemopexin, which was removed by chromatography on SP-Sephadex C50 (Pharmacia) as described by Martinez-Medellin and Schulman [18] for the purification of rabbit transferrin. The resulting transferrin gave a single band in immunoelectrophoresis at concentrations up to 5 mg/ml against antiserum to bovine serum. The SP-Sephadex step yields transferrin largely devoid of iron. Complete removal of iron was effected by dialysis against 0.02 M sodium citrate, pH 5.0, yielding apotransferrin with < 5 per cent iron saturation. The iron saturated form (Fe2-transferrin) was prepared by dialysis against 0.05 M Tris, pH 7.8, followed by addition of an excess of iron as ferric citrate and equilibrating for several hours. Both apotransferrin and Ve~-transferrin were finally dialysed against 0.05 M Tris0.02 M CaCI2, pH 7.8, and stored as frozen solutions at --20 °C. The protein content of the solutions was measured by the Lowry method [19].

Lactoferrin This was purified from bovine dry-period secretion, as described previously [20] yielding lactoferrin 20-30 ~o iron-saturated. The iron-saturated form (Fe2-1actoferrin) was prepared as described for Fe2-transferrin, and the iron-free form (apolactoferrin) by the method of Masson and Heremans [21]. Both were then dialysed, assayed for protein content and stored in the same way as the transferrin preparations.

Lactoferrin-fl-lactoglobulin complex A complex of apolactoferrin and fl-lactoglobulin, in the molar ratio 2:1 was prepared as described elsewhere [16].

216

Enzymes Crystalline trypsin (EC 3.4.21.4), activity 10 000 Kunitz units per mg, was obtained from Boehringer (Mannheim, G.F.R.). The protease inhibitor Trasylol was obtained from Bayer, S.A. (Barcelona, Spain). Neuraminidase (EC 3.2.1.18), from Clostridium perfringens, was obtained from Sigma (London).

Tryptic digestion This was carried out in 0.05 M Tris-0.02 M CaCI2, pH 7.8, at 37 °C, the concentrations of substrate and trypsin being 7.4 mg/ml and 150/~g/ml, respectively. After 3 h incubation, half the sample was removed and the trypsin neutralised with four units of Trasylol per #g of trypsin (l unit gives 50 ~o inhibition of 1/zg of trypsin). The same treatment was applied to the remainder of the sample after 24 h incubation.

Column chromatography Digests were chromatographed on a column (90 × 1.5 cm) of acrylamideagarose AcA 44 (L'Industrie Biologique Fran9aise, Genn~villiers, France) equilibrated and eluted with 0.05 M Tris, pH 7.8. This gel has a separation range of approx. 20 000-200 000 daltons [22]. To eliminate the tendency of lactoferrin to bind to the gel [16] it was necessary to raise the ionic strength of the eluting buffer by adding NaC1 to a final concentration of 0.5 M when digests of this protein were chromatographed.

Electrophoresis Electrophoresis in cellulose acetate strips ("Cellogel" Chemetron, Milan, Italy) was carried out at 200 V, either in 0.05 M veronal, pH 8.2, for 30 min, or in 0.1 M Tris/0.3 M glycine, pH 8.7, for 90 min. Protein was stained with Coomassie blue. Iron-containing proteins or protein fragments were stained specifically by the procedure described by Uriel for staining immunoelectrophoresis plates [23], but using a solution of bathophenanthroline in 2 ~o acetic acid rather than 0.02 M sodium acetate. This modification is necessary to permit detection of lactoferrin, since the acetate buffer, although satisfactory for transferrin, is not acidic enough to cause the release of iron from lactoferrin and thus allow staining. Proteins and fragments containing bound iron stained pink, the bands being most clearly seen when viewed by transmitted light after drying the strips. Proteins devoid of iron gave no reaction. Electrophoresis in sodium dod~ecylsulphate for molecular weight determination was carried out in acrylamide-agarose gels by the method of Shapiro et al. [24], and gels stained with Coomassie blue. Relative intensities of the stained bands were measured using a TLD 100 densitometer with integrator (Vitatron. Dieren, Holland). The following proteins were used as standards for molecular weight estimations: cytochrome c (M, = 12500), a-chymotrypsinogen A (Mr----25000) rabbit muscle aldolase (EC 4.1.2.13, Mr of monomeric subunit = 40 000) and ovalbumin (Mr ---45 000), all from Sigma, London, and human serum albumin (Mr = 67 000) from Behringwerke, Marburg, G.F.R. Aldolase and human serum albumin gave well defined bands corresponding to dimers which were used to give standards of Mr ---80 000 and 134 000, respectively.

Antisera Antisera to purified lactoferrin and transferrin were prepared in rabbits by

217 intradermal injection of emulsions of the proteins in Freund's complete adjuvant, according to a previously described protocol [25]. Immunoelectrophoresis This was performed by the method of Scheidegger [26] using 1 ~ agarose gels. Neuraminidase treatment Transferrin was incubated with neuraminidase for 24 h as described by Stratil and Spooner [17]. RESULTS Column chromatography Chromatography of the tryptic digests of lactoferrin and transferrin on AcA 44 normally yielded 2 major zones, though the relative sizes of the two zones varied considerably with the substrate (Fig. 1). The first zone corresponded to a heterogeneous mixture of intact protein and large fragments (henceforth referred to as the macromolecular fraction), and the second to low molecular-weight peptides. A peak of polymerised material sometimes eluted in the void volume, though it was of significant proportions only in the 24 h digest of Fez-lactoferrin (Fig. lc). It is obvious that the apoproteins were more readily digested than their iron-saturated counterparts, but no obvious differences between lactoferrin and transferrin were apparent at this stage. It was noticeable that although the same amount of protein was submitted to digestion in each case, the total absorbance of the digests of the Fez-proteins (Fig. la, lc) was greater than that of the apoproteins (Fig. lb, ld). This was probably due to the fact that the Fez-proteins have a significantly higher specific absorption at 280 nm than the apoproteins (27 and refs. therein). The fractions containing intact protein and/or large fragments, together with any containing polymeric material, were in each case pooled, concentrated by pressure dialysis and investigated by electrophoresis. Preliminary experiments showed that the peptide zones consisted of material with Mr < 10 000, and these were not further investigated. Cellulose acetate electrophoresis Electrophoresis of the macromolecular fraction of Fe2-transferrin digests in veronal buffer revealed 3 protein bands (Fig. 2). The central one corresponded to undigested transferrin, and the others to fragments with relatively fast and slow mobilities, which will be referred to as fragments F and S respectively. In 3 h digests the F and S bands were weak, most transferrin remaining undigested, whereas all 3 bands were strong in 24 h digests. In digests of apotransferrin, bands corresponding to intact transferrin and fragment S were observed, but fragment F was absent except for occasional traces in 24 h digests. These observations suggested that digestion of Feztransferrin yielded 2 distinct fragments (F and S), whereas apotransferrin normally yielded only fragment S (or a fragment of similar electrophoretic mobility) and peptides, the occasional traces of F in 24-h digests probably resulting from digestion of a small amount of Fez-transferrin in the apotransferrin preparation. All bands stained positively for iron, as shown in Fig. 2 (in the case of digests of apotransferrin sufficient iron to saturate the protein was added prior to performing electrophoresis).

218

(a) 0A0.350.3-

Io.~5(b)

0.20.15.

0.15-

I

~

t

0.1 ~

O.05 0

30

20

40

0.1.

0.05-

t

50

0

60

20

30

F r a c t i o n No.

0.25-

40 50 F r a c t i o n No.

6O

(c)

0.2-

]

E

~ 0.15-

(d)

0.15J

0.1-

o 0.1-1 ~' ~

+/ 0.0+1

~,~ ,~

'

0.05-

_

A

L,.,-'# ,'p.,-~,'.~ , 2O

30 40 Fraction No.

50

60

0

20

\ -,

30 40 F r a c t i o n No.

, 50

,'.'. 60

Fig. 1. Column chromatography o f tryptic digests o f bovine transferrin and lactoferrin on acrylamideagarose AcA 44: a, Fe2-transferrin; b, apotransferrin; c, Fe2-1actoferrin; d, apolactoferrin. - - - - , 3 h digest; -- -- - , 24 h digest. Fractions were 3 ml. The arrows indicate the position of elution o f undigested transferrin (a and b) and lactoferrin (c and d). ~

~'~ < ~i~ii'

~ii~?i ~ ~ ! {

!!!i~iiiii~i~ ¸¸

~

~i!iiiiii~i iiii ~i i iiii~iiiiiiiiii![i!i!iiiiiiii~iiii

....... ~++~ii• ~~iiii~iii~~) i~__iii~iii!ii~i~i~i!~ i ~ ~..-~ ~l~!ii~~+~ii~i~~+:~ ~+:~ii~ Iii]~- ii ii .i+]i~+~iiii

F+++i !

iii!

i

+i+!i

:!!I!iii i[iii i iiiii~.iiiiiiiiii iii i i !i i

Fig. 2. Cellulose-acetate electrophoresis in 0.05 M veronal, p H 8.2 o f (1) bovine iron-transferrin and (2) a 24 h tryptic digest of Fe2-transferrin: A, protein stain with Coomassie blue; B, specific stain for iron-containing proteins (for details see text). F and S indicate the bands corresponding to the fragments with fast and slow electrophoretic mobilities, respectively.

219 When electrophoresis was carried out in Tris/glycine buffer instead of veronal, the bands corresponding to intact transferrin and fragment S resolved into multiple bands, whereas fragment F continued to move as a single band (Fig. 3). Treatment of transferrin with neuraminidase reduced the mobility but not the number of bands, indicating that the multiple banding was not caused by differences in the number of sialic acid residues and was therefore presumably due to genetic variants of transferrin [17]. With both intact transferrin and fragment S three distinct bands were always observed, though additional weaker bands were sometimes seen, particularly in the case of intact transferrin. An additional weak unidentified band was sometimes seen between the F band and the fastest transferrin band (Fig. 3).

Fig. 3. Cellulose-acetate electrophoresis of bovine transferrin and tryptic digests of transferrin in 0.1 M tris/0.3 M glycine buffer pH 8.7: 1, Fe~-transferrin digested for 3 h; and, 2, for 24 h; 3, apotransferrin digested for 3 h; 4, undigested transferrin. F, fragment F band; ?, unidentified band; Tf, transferrin bands; S, fragment S bands. Cellulose acetate electrophoresis was less useful for analysis of digests of lactoferrin, since there was a tendency for intact lactoferrin, and probably also one of the larger fragments, to apparently interact with the support material, yielding poorlydefinied bands of variable mobility. This could be prevented by using strongly ionised buffers, but these gave virtually no separation of Lf and its fragments. Using the Tris/glycine buffer it was however, possible to distinguish 3 or 4 bands (2 strong, the others weak or very weak) in 3 h digests of apolactoferrin and 24 h digests of Fezlactoferrin, all of which stained positively for iron.

Dodecyl sulphate electrophoresis Dodecyl sulphate electrophoresis of the macromolecular fraction of Fe2transferrin digests (Fig. 4) revealed the presence of 3 bands, one of these corresponding to intact transferrin and the other two to fragments. With apotransferrin only the larger of the two fragments was generally observed, the smaller one appearing, if at all, only as a minor component of 24 h digests. These observations suggest that the

220

A~

Fig. 4. Polyacrylamide-agaroseelectrophoresis in dodecyl sulphate of trypsin-digested and undigested bovine transferrin:l, apotransferrin digested 3 h; 2, undigested transferrin; 3, Fe2-transferrin digested 24 h. Protein stained with Coomassie blue. larger and smaller fragments corresponded respectively to fragments S and F seen in cellulose acetate electrophoresis. Table I shows the molecular weight and percentage of each fragment and of intact protein and peptides in the various digests. The percentages of intact protein and of each fragment were calculated by estimating from the AcA 44 chromatograms the proportion of the original protein present in the macromolecular fraction (plus polymers, if present) and using this figure to convert the relative intensities of dodecyl sulphate electrophoresis bands (measured by densitometry) into percentages of the original protein sample. The percentages of peptides were calculated directly from the AcA 44 chromatograms. It is clear that Fe2-transferrin is much more resistant to proteolysis than apotransferrin, and that the fragments formed from Fe2-transferrin TABLEI PERCENTAGES AND MOLECULAR WEIGHTS OF INTACT TRANSFERRIN, FRAGMENTS, AND PEPTIDES IN TRYPTIC DIGESTS OF BOVINE TRANSFERRIN Molecular weights were calculated from polyacrylamide gel electrophoresis in dodecyl sulphate according to ref. 24, and percentages as described in the text. n.d., not detectable; Tf, transferrin. Percentage of original transferrin present as various components: Mr Undigested transferrin Fragment Fragment Peptides

Fe~-Tf digested for

Apo-Tf digested for

3h

24 h

3h

24 h

41 18 23 16

9 24 n.d. 64

2 13 Trace 82

72 900 81 38 500 5 32 000 3 < 10 000 5

221

in Fig. 5. Polyacrylamide-agarose electrophoresis in dodecyl sulphate of trypsin-digested and undigested bovine lactoferrin (Lf) and apolactoferrin-fl-lactoglobulin complex: 1, Fez-lactoferrin digested 3 h; 2, apolactoferrin digested 3 h; 3, apolactoferrin-fl-lactoglobulin complex, digested 3 h; 4, apolactoferrin-fl-lactoglobulin complex undigested. The fastest and slowest bands, marked flLg and Lf correspond to the fl-lactoglobulin monomer subunit (Mr 18 500) and intact lactoferrin (Mr, 91 000) respectively. The intermediate bands represent lactoferrin fragments A to E (see tables II and III). The bands corresponding to fragments C and E were generally faint and are not always visible in the photograph. Protein stained with Coomassie blue. are m o r e r e s i s t a n t to f u r t h e r p r o t e o l y s i s t h a n t h e f r a g m e n t ( s ) d e r i v e d f r o m a p o t r a n s f e r r i n w h i c h , at 24 h, is a l m o s t c o m p l e t e l y d e g r a d e d t o p e p t i d e s . D o d e c y l s u l p h a t e e l e c t r o p h o r e s i s o f l a c t o f e r r i n digests (Fig. 5 a n d T a b l e II) r e v e a l e d a m o r e c o m p l e x p a t t e r n t h a n w i t h t r a n s f e r r i n , u p to 5 f r a g m e n t s (referred to TABLE II PERCENTAGES A N D M O L E C U L A R WEIGHTS OF INTACT LACTOFERRIN, FRAGMENTS, A N D PEPTIDES IN TRYPTIC DIGESTS OF BOVINE L A C T O F E R R I N For details see Table I. n.d., not detectable; Lf, lactoferrin. Percentage of original lactoferrin present as various components: Mr

Undigested lactoferrin Fragment A Fragment B Fragment C Fragment D Fragment E Peptides

91 000 52 700" 47 300" 40 000 31 800 25 000 < I0 000

Fe2-Lf digested for

Apo-Lf digested for

3h

24h

3h

24h

6

6

4

**

36 n.d. 2 n.d. 48

** *" "* ** 89

39 5 28 Trace 12

46 n.d. 16 Trace 22

* Fragments not separable quantitatively by densitometry. "* Insufficient macromolecular material remained to quantify components.

222 as A to E in order of decreasing molecular weight) being observed, of which the two largest (A and B) were not sufficiently well separated to allow individual quantification by densitometry. In 3 h digests of Fe2-1actoferrin A predominated, changing to near equality in 24 h digests, and B predominated in 3 h digests of apolactoferrin. A smaller fragment (D; Mr ~ 31 800) was also prominent in digests of Fe2-1actoferrin, though its amount decreased from 3 to 24 h, in contrast to the larger fragment(s), which appeared to actually increase. Since little intact lactQferrin remained available for digestion even at 3 h, it seems likely that this apparent increase was due to the large amount of polymeric material in the 24 h digest, which may well have included some aggregated peptide material and thus artificially increased the apparent amounts of macromolecular material eluting from the AcA 44 column. The remaining fragments (C and E) with Mr = 40 000 and 25 000 respectively never accounted individually for more than 5 ~ of the original protein. It is clear that lactoferrin, even in the iron-saturated form, is readily cleaved by trypsin, only 6 ~ of intact protein remaining after 3 h hydrolysis as compared with over 80~o of Fe2-transferrin. However, the resulting fragments appeared to be relatively resistant to further digestion, as even at 24 h only 22 ~ of Fez-lactoferrin was in the form of peptides (Table II), a figure not greatly in excess of the 16~o for the corresponding Fe2-transferrin hydrolysate (Table I). Apolactoferrin was, like apotransferrin, rapidly digested to peptides and there was insufficient macromolecular material remaining in 24 h digests to analyse by electrophoresis.

Tryptic digestion of apolactoferrin-fl-lactoglobulin complex Column chromatography of digests of the apolactoferrin-fl-lactoglobulin complex on AcA 44 followed by dodecyl sulphate electrophoresis yielded the results shown in Table I11. Only traces of intact fl-lactoglobulin were detected, even in 3 h digests (see Fig. 5) hence in the calculations of percentages shown in Table 111 it has TABLE III PERCENTAGES AND MOLECULAR WEIGHTS OF INTACT LACTOFERR1N, FRAGMENTS, AND PEPTIDES IN TRYPTIC DIGESTS OF BOVINE APOLACTOFERRIN-flLACTOGLOBULIN COMPLEX For details and footnotes see Tables I and I1.

M~ Undigested lactoferrin Fragment A Fragment B Fragment C Fragment D Fragment E Peptides

91 000 52 700" 47 300" 40 000 31 800 25 000 < 10 000

Percentage of original lactoferrin present as various components in complex digested for 3h

24h

8

n.d.

32 Trace 12 Trace 32

22 n.d. 1 n.d. 65

223 been assumed that all the fl-lactoglobulin, which accounted for 16 ~ of the original protein, was in the peptide fraction. Comparison with Table II shows that although digestion of the complex was slighly less extensive than with uncomplexed apolactoferrin virtually no protection of intact apolactoferrin occurred. Qualitatively the pattern of fragments was similar to that obtained with uncomplexed lactoferrin, with fragments A and B predominating. As with uncomplexed lactoferrin, these could not be quantified separately, but they appeared to be present in approximately equal proportions in 3 h digests, whereas B predominated in 24 h digests.

lmmunological properties of tryptic digests lmmunoelectrophoresis of digests against homologous antisera did not reveal distinct bands which might be attributable to fragments, but showed a single arc of variable length. No cross reaction was observed with heterologous antisera, i.e. digestion of lactoferrin did not yield fragments capable of reacting with antiserum to transferrin, or vice versa. DISCUSSION Previous studies of the proteolysis of ovotransferrin [11, 28], and human transferrin [11, 12] and lactoferrin [12] have shown that in each case the apoprotein was more susceptible to proteolysis than its iron-saturated counterpart. The work reported here shows that bovine transferrin and lactoferrin clearly conform to this pattern. However, previous workers have not generally attempted to quantify the difference in susceptibilities of the apo and iron-saturated forms, an exception being Azari and Feeney [11] who made comparative estimates of the loss of chromogenic capacity of the two forms of ovotransferrin and human transferrin. However, ironbinding fragments are also chromogenic, hence this is not a reliable method for estimating the amount of intact protein present, and we have found (unpublished) that after 18 h digestion of Fe2-transferrin with trypsin the E470 had decreased by < 10%, whereas some 50% of the original transferrin had in fact been cleaved, to yield iron-binding fragments. Kaminski [14] reported that proteolysis of human Fe2-transferrin yielded coloured products which were presumably iron-binding fragments, but detailed investigations of iron-binding fragments from proteolysis of transferrins have been made only in the recent studies ofovotransferrin by Williams [28, 29]. These fragments, derived from either the C- or N-terminal half of the molecule, were obtained by proteolysis of ovotransferrin (by trypsin or chymotrypsin) specially treated to give preferential partial saturation of either the C- or N-terminal iron-binding site respectively. Under his conditions of proteolysis (which differed somewhat from those used in the present work) Williams [28] reported that Fe2-ovotransferrin was totally resistant to digestion, whereas apo-ovotransferrin was totally degraded to low molecular weight material. The present results indicate that bovine transferrin does not show such a marked difference in behaviour betwen the apo and Fez forms, iron-binding fragments being obtained from both forms. From Fe2-transferrin two fragments of M, = 32 000 and 38 500 were obtained in approximately equal amounts, suggesting that cleavage of a single bond, or elimination of a small peptide slightly to one side of the mid-point of the transferrin polypeptide chain may have occurred. The fact

224 that one fragment exhibited electrophoretic heterogeneity, probably due to genetic variants, whereas the other did not, also suggests that the two fragments were derived from different regions of the molecule. In the case of apotransferrin, only the heavier of the two fragments (or a fragment of similar molecular weight and electrophoretic mobility) was obtained, suggesting that the two iron-binding regions may differ in the degree to which their susceptibility to proteolysis increases when iron is removed. Isolation and purification of the individual fragments is currently in progress and should permit the verification or otherwise of the above suggestions. Contrary to the hypothesis advanced in the introduction to this work, it appears that bovine lactoferrin is actually more susceptible to proteolysis than transferrin. This was particularly evident with the iron-saturated forms, only some 6 ~o of intact Fe2-1actoferrin remaining after 3 h digestion as compared with over 80 ~ of Feztransferrin. However, most of the remaining lactoferrin was in the form of two large fragments which appeared to be quite resistant to further proteolysis. There was less difference between the apoproteins, both being readily digested, so that little intact apotransferrin or apolactoferrin remained after 3 h, and both were almost completely degraded to peptides at 24 h. Because of the more complex pattern of fragments produced by lactoferrin it was not possible, as with transferrin, to pinpoint any qualitative differences between the fragments of Fe2-1actoferrin and of apolactoferrin, the largest fragment(s) appearing to be the most abundant in both cases. The appearance of several bands staining positively for iron after cellulose acetate electrophoresis indicates that the major proteolysis fragments, like those from transferrin, are capable of binding iron. It has long been known that lactoferrin and transferrin do not cross react immunologically. However, the recent report by Masson et al. [30] that succinylated lactoferrin can cross-react with antiserum to transferrin suggests that the two proteins do share some common but buried antigenic determinants. Clearly, proteolysis by trypsin does not reveal any such determinants in the bovine proteins, since no crossreactions were obtained when tryptic digests were tested by immunoelectrophoresis. The lack of distinct bands corresponding to fragments in immunoelectrophoresis of digests against homologous antisera is presumably due to common antigenic determinants on the fragments and native protein, which would cause the individual immunoelectrophoresis arcs to merge into one large band, as was observed. The resistance of apolactoferrin to proteolysis was not significantly enhanced by complexing with fl-lactoglobulin, the complex still being more susceptible to digestion, in terms of stability of the fragments, than Fe:lactoferrin. It seems unlikely therefore, that complex formation with fl-lactoglobulin is a mechanism whereby lactoferrin could be protected against tryptic digestion in the intestinal tract, although complexes with other proteins, or treatment with other enzymes might yield different results. Clearly, the present results do not offer any indications of how lactoferrin can apparently exert an antibacterial effect in the gastrointestinal tract [9] without undergoing proteolysis, except, perhaps, that iron-binding fragments may also be able to mediate this effect and maintain bacteriostasis after partial degradation has occurred. Isolation and characterisation of the iron binding fragments should help to provide evidence for or against this possibility. It should also be remembered that the reported in vivo antibacterial effect of lactoferrin [9] was demonstrated in guinea pigs, whereas

225 the present work was carried out with proteins of bovine origin. There m a y be significant species variation in the susceptibilities of transferrin a n d lactoferrin to proteolysis. REFERENCES 1 Lane, R. S. (1967) Nature 215, 161-162 2 Masson, P. L., Heremans, J. F. and Dive, C. (1966) Clin. Chim. Acta 14, 735-739 3 Beam, A. G. and Parker, W. C. (1966) Glycoproteins, (Gottschalk, A., ed.), pp. 413-433, Elsevier, Amsterdam 4 Schade, A. L. and Caroline, L. (1946) Science 104, 340-341 5 Bullen, J. J. and Rogers, H. J. (1969) Nature 224, 380-382 6 Masson, P. L., Heremans, J. F., Prignot, J. J. and Wauters, G. (1966) Thorax 21,538-544 7 0 r a m , J. D. and Reiter, B. (1968) Biochim. Biophys. Acta 170, 351-365 8 Glynn, A. A. (1972) Symp. Soc. Gen. Microbiol. 22, 91-99 9 Bullen, J. J., Rogers, H. J. and Leigh, L. (1972) Br. Med. J. 1, 69-75 10 Tomasi, T. B. (1970) Annu. Rev. Med. 21,281-298 11 Azari, P. R. and Feeney, R. E. (1958) J. Biol. Chem. 232, 293-302 12 Spik, G. and Montreuil, J. (1966) C.R. Soc. Biol. 160, 93-98 13 Got, R. (1964) Protides of Biological Fluids, (Peeters, H., ed.), pp. 385-387, Elsevier, Amsterdam 14 Kaminski, M. (1962) Nature 194, 27-28 15 Hekman, A. (1971) Biochim. Biophys. Acta 251,380-387 16 Pifieiro, A., Lampreave, F., Brock, J. H. and Carbon6, R. (1976) Rev. Espafi. Fisiol., submitted for publication 17 Stratil, A. and Spooner, R. L. (1971) Biochem. Genet. 5, 347-365 18 Martinez-Medellin, J. and Schulman, H. M. (1972) Biochim. Biophys. Acta 264, 272-284 19 Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (1951) J. Biol. Chem. 193,265275 20 Reiter, B., Brock, J. H. and Steel, E. D. (1975) Immunology 28, 83-95 21 Masson, P. L. and Heremans, J. F. (1968) Eur. J. Biochem. 6, 579-584 22 Boschetti, E., Tixier, R. and Uriel, J. (1972) Biochimie 54, 439-444 23 Uriel, J. (1971) Methods in Immunology and Immunochemistry (Williams, C. A. and Chase, M.W., eds.), pp. 304-305, Vol. 3, Academic Press, New York 24 Shapiro, A. L., Vifiuela, E. and Maizel, J. V. (1967) Biochem. Biophys. Res. Comm. 28, 815-820 25 Pifieiro, A., Ortega, F. and Uriel, J. (1975) Biochim. Biophys. Acta 379, 201-206 26 Scheidegger, J. J. (1955) Int. Arch. Allergy 7, 103-110 27 Groves, M. L. (1971) Milk Proteins (McKenzie, H. A., ed.), p. 374, Vol. 2, Academic Press, New York 28 Williams, J. (1974) Biochem. J. 141,745-752 29 Williams, J. (1975) Biochem. J. 149, 237-244 30 Masson, P. L., Teuwissen, B., van Snick, J., Roberts, T. K., Heremans, J. F. and Topp, R. C. (1973) International Conference on Iron Storage and Transport Proteins, University College Hospital Medical School, London, Abstract 15