- Email: [email protected]
Preparation and properties of a single-sited fragment from the C-terminal domain of human transferrin
348
Biochimica et Biophysica Acta 829 (1985) 348 353
Elsevier BBA32226
Preparation and properties of a single-sited f r a g m e n t f r o m the C-terminal d o m a i n of h u m a n transferrin Olga Zak and Philip Aisen * Departments of Physiology and Biophysics, and Medicine, Albert Einstein College of Medicine, Bronx, N Y 10461 (U.S.A.)
(ReceivedJanuary 8th, 1985)
Key words: Transferrin; Metal binding; C-terminal fragment
A single-sited iron-binding fragment of human transferrin has been obtained by thermolysin cleavage of the protein, selectively loaded with iron in the C-terminal binding site, in a urea-containing buffer. The fragment contains carbohydrate, and hence derives from the C-terminal half of transferrin. Its metal-binding site accepts Fe 3+ and C u z+ with bicarbonate as accompanying anion, but only Fe 3+ with oxalate as anion. EPR spectroscopic properties of the fragment are similar to those of the corresponding site in the intact protein. However, iron-binding by the fragment is weaker than by the C-terminal site of the intact protein, particularly at low pH, suggesting that overall as well as local protein conformation influences the metal-binding functions of the site.
Introduction The transferrins comprise a class of two-sited, single-chain, iron-binding proteins widely distributed in vertebrates. Internal homologies in the amino acid sequences of transferrins differing in origin provide compelling evidence that these proteins evolved by a mechanism entailing duplication and fusion of an ancestral gene specifying a single-sited precursor molecule [1-3]. The biological advantage conferred by the doubling in size of transferrin may simply be that the larger protein i s protected from glomerular loss in the circulation [4]. Since studies based on urea-gel electrophoresis indicate that the two sites of human circulating transferrin are unequally populated [5-7], the possibility also remains that the sites differ in function, to the benefit of the host organism. Delineation of the properties of the individual sites, and the differences and similarities between them, may
* To whom correspondenceshould be addressed.
therefore have physiological as well as chemical interest. Clearly revealed differences between the two sites of native human transferrin have emerged from a variety of spectroscopic, thermodynamic and kinetic studies of human transferrin [8-10]. In seeking to determine whether these differences represent interactions dependent on the overall structure of the native protein, we have previously examined some properties of the single binding site in a proteolytically cleaved fragment from the N-terminal domain of human transferrin [11]. Now, we report on the preparation and the properties of a single-sited fragment from the C-terminal domain of the protein. Our results indicate that the metal-binding properties of this 'halftransferrin' differ from the properties of the binding site in the C-terminal domain of the intact protein.
Materials and Methods H u m a n s e r u m transferrin. Human serum transferrin was purchased from Calbiochem. Since this
0167-4838/85/$03.30 © 1985 ElsevierSciencePublishers B.V. (BiomedicalDivision)
349 commercial preparation yielded a single band on SDS-gel electrophoresis, and exhibited the spectroscopic and iron-binding properties of transferrin isolated in this laboratory from Cohn Fraction IV-7 or pooled blood bank plasma, it was used without further purification. Transferrin, and the fragments derived from it, were freed of iron and chelating agents by previously reported methods [8]. To avoid irreversible precipitation of fragment preparations at low ionic strength, dialysis against water was replaced by dialysis against 0.1 M NaNO 3. Preparation of the C-terminal fragment. The presence of iron at a binding site of transferrin renders the domain in which the site resides relatively resistant to denaturation and proteolysis [12]. Accordingly, the starting material for the preparation of the C-terminal fragment was transferrin selectively loaded at the A site in the Cterminal domain of the protein by iron presented as its complex with nitrilotriacetate [8]. Best results were obtained with 5% solutions of protein brought to 60% saturation with iron in 0.05 M Tris/0.01 M CaC12/0.001 M NaHCO 3 (pH 6.2). After making the preparation 2 M in urea to promote unfolding of the unoccupied N-terminal domain, and adding thermolysin to a ratio of 1 : 25 (w/w) with transferrin, incubation with shaking for 90 rain was carried out at 37°C. Typically, the absorbance at 470 nm of the preparation declined by 30-50% during this incubation. Proteolytic fragments were then isolated by gel-permeation chromatography on a 2.5 × 90 cm column of Sephadex G-75 SF, using 0.1 M NH4HCO 3 (pH 8.0) as buffer. The smaller, slower-moving components of the digestion mixture were further purified on a preparative electrofocusing column employing 2% ampholyte, pH 4-6.5 (Pharmacia). The fragment of interest migrated with an isoelectric point of 5.1. Final yields ranged from 5% to 10% of the starting protein, the theoretical maximum being near 50%. Physical measurements. Electron paramagnetic resonance spectra were obtained with a Varian E-9 EPR spectrometer interfaced to an Apple II + microcomputer. Optical spectra were recorded on a Cary model 14 spectrophotometer. Other procedures. SDS-gel electrophoresis followed procedures given by Maizel [13] with per-
iodic acid-Schiff staining for carbohydrate according to Segrest and Jackson [14]. Equilibrium dialysis was carried out by methods previously detailed [8]. Carbohydrate analyses, kindly performed by A. Adamany, were obtained using a Perkin-Elmer Model 910 gas chromatography with the techniques of Blumenfeld and Adamany [15]. Results and Discussion
Purification and identification of the fragment Following the initial chromatographic separation of the digestion mixture on a column of Sephadex G-75 SF, three bands were seen on SDS-gel electrophoresis, with apparent molecular weights of 38000, 42000 (the major component) and 48000. Additional minor components with mobilities similar to that of native transferrin were also present in some preparations, in which case rechromatography was carried out. Each components gave a strongly positive periodic acid-Schiff staining reaction for carbohydrate. Since both glycan chains of human serum transferrin are bound to the C-terminal half of the molecule, to asparagine residues 413 and 610 respectively [1], the fragments must have originated in this half. For contrast, the N-terminal fragment prepared by our modification of the method of Lineback-Zins and Brew [11,16] gave no periodic acid-Schiff reaction. Preparative isoelectric focusing of the lowmolecular-weight fraction from the G-75 SF column resolved two pink components, the major one migrating to pH 5.1. On SDS-gel electrophoresis, this component gave only a single prominent band which was periodic acid-Schiff-positive and had an apparent molecular weight of 42000 (Fig. 1). In some preparations a second band, with an apparent molecular weight of 47000, was observed., Because this minor band generally exhibited less than 2% of the intensity of the major band by densitometry, no further attempt at purification was made, and all additional studies were carried out on the electrofocused pH 5.1 fraction.
Characterization of the pH 5.1 fraction The absorbance at 280 nm of a 1% solution of the purified fragment was calculated as 11.55 from the previously measured absorbancies of the Nterminal fragment and native transferrin [11]. An
350
1
2
3
Mr
4
5
TABLE I C A R B O H Y D R A T E ANALYSIS OF p l 5.1 F R A G M E N T
ii - oooo-
- -
42
000
Carbohydrate
M o l / m o l of Fe-binding activity
Sialic acid Galactose Mannose N-Acetylglucosamine
2.5 2,6 3,3 4,2
--
A
The results of carbohydrate analysis are presented in Table I. From the concentrations of sugars detected, it seems likely that the bulk of the molecules in the preparation bear only a single glycan chain of the native protein.
B
Fig. 1, SDS-gel electrophoresis of transferrin and fragment eleetrofocussed at pH 5.1. (A) Coomassie blue stain; (B) periodic acid-Schiff stain. Lanes 1, 3, 5: fragment; lanes 2, 4: transferrin.
estimate of the equivalent weight for iron-binding was obtained by titration of the fragment in a 0.05 M Hepes/0.01 M HCO3- buffer (pH 7.45) with freshly prepared ferric nitrilotriacetate (1:3, pH 4.5) as titrant. A sharp end-point was reached, giving an equivalent weight near 35000 for the preparation used for most of the equilibrium dialysis studies. Discrepancies between the equivalent weights of the fragments, and the apparent molecular weights determined by SDS-gel electrophoresis are probably attributable to the insensitivity of the latter technique to carbohydrate moieties.
The strength of iron binding As with native transferrin and the N-terminal fragment, thermodynamic stability constants for iron binding by the C-terminal fragment were measured near pH 6.7 and pH 7.4 by the method of equilibrium dialysis with citrate as competing complexing agent [8,11]. A working range of Fe 3÷ activities was obtained by varying total Fe 3÷, keeping citrate concentration and ionic strength constant at each pH. The measured constants pertain to the overall reaction: Fe 3 + + fragment + H C O ~ ~ Fe 3 +-fragment-HCO 3 + 3H +-
Their values, and the experimental data from which they are calculated, are summarized in Tables II and III. At fixed pH and Pco2, effective or condi-
T A B L E II E Q U I L I B R I U M B I N D I N G O F Fe 3÷ TO C - T E R M I N A L T R A N S F E R R I N F R A G M E N T N E A R pH 7.4 Experiment
A B C
Total [Fe 3 + ] at equilibrium (/LM) protein
buffer
32.2 60.3 70.9
6.57 20.1 25.8
[ Fe 3 + bound] [fragment] a
pH
- log[Fe 3 + aq]
K
0.25 0.40 0.45
7.42 7.42 7.45
20.7 20.2 20.2
37,3 24.7 19.9 27.3±5,2 b
a Concentration of fragment near 100 # M in all experiments. b Mean 5: S.E.
351 TABLE III EQUILIBRIUM BINDING OF Fe 3+ TO C-TERMINAL TRANSFERRIN F R A G M E N T N E A R pH 6.7 Experiment
A B C D
Total [Fe 3+ ] at equilibrium ( p, M) protein
buffer
14.6 30.1 41.2 51.9
4.91 4.18 12.5 20.5
[Fragment] (/~ M)
56.8 94.4 94.2 91.3
[Fe 3+ bound] [ fragment]
0.17 0.27 0.30 0.34
pH
- l o g [ F e 3+ aq]
K
6.71 6.67 6.67 6.69
19.5 18.6 18.2 18
128 378 139 87 183___66 a
Mean_+ S.E.
tional stability constants for the apparent reaction: Fe 3 + + fragment ~ Fe 3+-fragment
may be derived. These constants, which provide more readily interpretable measures of the strength of iron binding under laboratory or physiological conditions, are given in Table IV, along with corresponding constants for the N-terminal fragment and the two sites of native transferrin. As is evident from this table, the tightness of the iron-protein bond is appreciably weaker in the C-terminal fragment than in the corresponding site of the native protein. This difference is more marked at
pH 6.7, where iron is bound 30-times more strongly to the A-site of transferrin than to the fragment, than at physiological pH where the ratio of binding strengths is 4. At pH 7.4, then, binding of iron to the A-site of transferrin is stabilized by about 10 kJ (2.4 kcal) per mol in the intact protein compared to the C-terminal fragment.
EPR spectra A compelling body of evidence that the two sites of iron-loaded transferrin have distinguisha-
A j
Fe2 Trons{errm
TABLE 1V EFFECTIVE STABILITY CONSTANTS FOR THE BIN D I N G OF IRON BY T R A N S F E R R I N A N D ITS SINGLE-SITED FRAGMENTS In air (Pco- = 3.6"10-4 atm) (M - ¢) pH 6.7 C-terminal Fragment Transferrin 'a'-Site a N-terminal Fragment b Transferrin 'b'-Sitea a From Ref. 8. b From Ref. 11.
In blood ( P c o , = 0.05 atm) (M - ~')
pH 7.4
1.1.1018
1.0.1020
1.5.10 22
2,9.1019
4.0.10 20
5.8.10 22
1.8-1017
5.7.1019
8.4" 1 0 21
<1.4.1018
6.8.1019
9.8.1021
B ~
ogment
cJ
~
A-B/IV ermlno151~eJ
,S I
I
1300
1400
I 1500 1600 GAUSS
l
I
r
1700
1800
1900
Fig. 2. EPR spectra of (A) diferric transferrin (2.2-10 -4) and
(B) C-terminal fragment (2.0-10-4). Spectrum C, representing the 'B' site in the N-terminal domain of transferrin, was obtained by computer subtraction of spectrum (B) from spectrum (A) after normalizing to protein concentration and instrumental gain. Buffer: 0.1 M Hepes/0.05 M KC1 (pH 7.4). Instrumental parameters: temperature, 77 K; microwave power, 10 mW; microwave frequency, 9.208 GHz; modulation frequency, 100 kHz; modulation amplitude, 10 (3.
352
ble EPR spectra now exists [8,11,17]. This evidence is further corroborated by the spectra in Fig. 2. The splitting between the two low-field peaks of the g' = 4.3 signal of the C-terminal fragment is 39 G. This contrasts with a splitting of 27 G in the spectrum of the single-sited fragment from the N-terminal domain of human transferrin [11]. As in our earlier study [11], the spectrum of the C-terminal fragment displays a greater amplitude than that of its N-terminal counterpart, consistent with a greater anisotropy in the spectrum of the latter. Integrated intensities of the spectra of the C-terminal fragment and diferric transferrin, each normalized to optical absorbance at 470 nm, agreed to within 1%, suggesting that the optical spectr~ of the two sites are similar. The EPR spectrum of the 65Cu2+ complex of the C-terminal fragment, with bicarbonate as the synergistic anion, was indistinguishable from the spectrum of native 65Cu-transferrin-bicarbonate. Attempts to prepare a complex of 65Cu 2+ and C-terminal fragment, with oxalate as co-anion, were unsuccessful, leading only to a mixture of nonspecific complexes and Cu-fragment-bicarbonate. In accord with earlier inferences [18], therefore, only the N-terminal site is capable of accepting oxalate when binding copper. Interestingly, and as yet inexplicably, the sites in N-terminal and C-terminal fragments alike accept oxalate when binding Fe 3+ (Fig. 3). In the preparation
from which the spectrum of this figure is obtained, there is at least lO-times as much Fe 3+ bound with oxalate as co-anion as is bound with bicarbonate, since the amplitude of a spectral line varies inversely with the square of its linewidth. This inference is corroborated by adding an excess of bicarbonate to the preparation, thereby largely abolishing the broad peaks of the oxalate complex while increasing the amplitude of the bicarbonate feature more than lO-fold. The two-sited nature of transferrin The susceptibility of the transferrins to proteolyric cleavage generating half-molecules with apparently intact metal binding sites has provided a useful tool for exploring the properties of the individual binding domains [11,19-23]. Differences between the chemical and physical properties of the sites in the intact protein have been well established [8-10,20], and many of these differences are at least grossly preserved in single-sited fragments when these fragments are sufficiently large to maintain intact binding sites [11,20,21,24]. The strength of iron binding, however, appears to be appreciably modulated by the environment offered by the intact protein, particularly at low pH. Whether this has any physiological implication is not clear. Interaction of ovotransferrin with its receptor on the chick embryo red blood cell depends on recognition regions in both halves of the transferrin molecule [23], and this may be a more important consequence of the two-domain structure of the protein. To dissect the effects of each iron-binding domain on the other remains a central goal in transferrin biochemistry. The availability of transferrin half-molecules offers unique advantage in achieving this goal [23].
Acknowledgements
600
8 0
I 1000
I 1200 GAUSS
I 1400
I 1600
2000
Fig. 3. EPR spectrum of the ternary complex of C-terminal fragment, Fe 3+ and oxalate. Fragment concentration, 1-10 -4 M. Instrumental parameters as in Fig. 2 except: microwave power, 20 mW.
This work was supported in part by Grant AM15056 from the National Institutes of Health. We are grateful to Dr. A. Adamany for carrying out the carbohydrate analysis of the purified fragment.
References 1 MacGillivray, R.T.A., Mendez, E. Shewale, J.G., Sinha, S.K., Lineback-Zins, J. and Brew, K. (1983) J. Biol. Chem. 258, 3543-3553
353 2 Mazurier, J., Metz-Boutigue, M.H., JollY, J., Spik, G., Montreuil, J. and Joll~s, P. (1983) Experientia 39, 135-141 3 Williams, J. (1982) Trends Biochem. Sci. 7, 394-397 4 Williams, J., Grace, S.A. and Williams, J.M. (1982) Biochem. J. 201,417-419 5 Leibman, A. and Aisen, P. (1979) Blood 53, 1058-1065 6 Williams, J. and Moreton, K. (1980) Biochem. J. 185, 483-488 7 Evans, R.W., Williams, J. and Moreton, K. (1982) Biochem. J. 201, 19-26 8 Aisen, P., Leibman, A. and Zweier, J. (1978) J. Biol. Chem. 253, 1930-1937 9 Chasteen, N.D., White, L.K. and Campbell, R.F. (1977) Biochemistry 16, 363-368 10 Baldwin, D.A., DeSousa, D.M.R., Von Wandruszka, R.M.A. (1982) Biochim. Biophys. Acta 719, 140-146 11 Zak, O., Leibman, A. and Aisen, P. (1983) Biochim. Biophys. Acta 742, 490-495 12 Evans, R.W. and Williams, J. (1980) Biochem. J. 541-546 13 Maizel, J.V., Jr. (1971) Methods Virol. 5, 179-246 14 Segrest, J.P. and Jackson, R.L. (1972) Methods Enzymol. 28, 54-63
15 Blumenfeld, O.O. and Adamany, A.M. (1978) Proc. Natl. Acad. Sci. USA 75, 2727-2731 16 Lineback-Zins, J. and Brew, K. (1980) J. Biol. Chem. 255, 708-713 17 Aasa, R. (1972) Biochem. Biophys. Res. Commun. 49, 806-812 18 Zweier, J.L. and Aisen, P. (1977). J. Biol. Chem. 252, 6090-6096 19 Bluard-Beconinck, J.M., Williams, J., Evans, R.W., Van Snick, J., Osinski, P.A. and Masson, P.L. (1978) Biochem. J. 171, 321-327 20 Evans, R.W. and Williams, J. (1978) Biochem. J. 173, 543-552 21 Keung, W.-M., Azari, P. and Phillips, J.L. (1982) J. Biol. Chem. 257, 1177-1183 22 Keung, W.-M. and Azari, P. (1982) J. Biol. Chem. 257, 1184-188 23 Brown-Mason, A. and Woodworth, R.C. (1984) J. Biol. Chem. 259, 1866-1873 24 Legrand, D., Mazurier, J., Metz-Boutigue, M.H., Joll6s, J., Joll6s, P., Montreuil, J. and Spik, G. (1984) Biochim. Biophys. Acta 787, 90-96
Biochimica et Biophysica Acta 829 (1985) 348 353
Elsevier BBA32226
Preparation and properties of a single-sited f r a g m e n t f r o m the C-terminal d o m a i n of h u m a n transferrin Olga Zak and Philip Aisen * Departments of Physiology and Biophysics, and Medicine, Albert Einstein College of Medicine, Bronx, N Y 10461 (U.S.A.)
(ReceivedJanuary 8th, 1985)
Key words: Transferrin; Metal binding; C-terminal fragment
A single-sited iron-binding fragment of human transferrin has been obtained by thermolysin cleavage of the protein, selectively loaded with iron in the C-terminal binding site, in a urea-containing buffer. The fragment contains carbohydrate, and hence derives from the C-terminal half of transferrin. Its metal-binding site accepts Fe 3+ and C u z+ with bicarbonate as accompanying anion, but only Fe 3+ with oxalate as anion. EPR spectroscopic properties of the fragment are similar to those of the corresponding site in the intact protein. However, iron-binding by the fragment is weaker than by the C-terminal site of the intact protein, particularly at low pH, suggesting that overall as well as local protein conformation influences the metal-binding functions of the site.
Introduction The transferrins comprise a class of two-sited, single-chain, iron-binding proteins widely distributed in vertebrates. Internal homologies in the amino acid sequences of transferrins differing in origin provide compelling evidence that these proteins evolved by a mechanism entailing duplication and fusion of an ancestral gene specifying a single-sited precursor molecule [1-3]. The biological advantage conferred by the doubling in size of transferrin may simply be that the larger protein i s protected from glomerular loss in the circulation [4]. Since studies based on urea-gel electrophoresis indicate that the two sites of human circulating transferrin are unequally populated [5-7], the possibility also remains that the sites differ in function, to the benefit of the host organism. Delineation of the properties of the individual sites, and the differences and similarities between them, may
* To whom correspondenceshould be addressed.
therefore have physiological as well as chemical interest. Clearly revealed differences between the two sites of native human transferrin have emerged from a variety of spectroscopic, thermodynamic and kinetic studies of human transferrin [8-10]. In seeking to determine whether these differences represent interactions dependent on the overall structure of the native protein, we have previously examined some properties of the single binding site in a proteolytically cleaved fragment from the N-terminal domain of human transferrin [11]. Now, we report on the preparation and the properties of a single-sited fragment from the C-terminal domain of the protein. Our results indicate that the metal-binding properties of this 'halftransferrin' differ from the properties of the binding site in the C-terminal domain of the intact protein.
Materials and Methods H u m a n s e r u m transferrin. Human serum transferrin was purchased from Calbiochem. Since this
0167-4838/85/$03.30 © 1985 ElsevierSciencePublishers B.V. (BiomedicalDivision)
349 commercial preparation yielded a single band on SDS-gel electrophoresis, and exhibited the spectroscopic and iron-binding properties of transferrin isolated in this laboratory from Cohn Fraction IV-7 or pooled blood bank plasma, it was used without further purification. Transferrin, and the fragments derived from it, were freed of iron and chelating agents by previously reported methods [8]. To avoid irreversible precipitation of fragment preparations at low ionic strength, dialysis against water was replaced by dialysis against 0.1 M NaNO 3. Preparation of the C-terminal fragment. The presence of iron at a binding site of transferrin renders the domain in which the site resides relatively resistant to denaturation and proteolysis [12]. Accordingly, the starting material for the preparation of the C-terminal fragment was transferrin selectively loaded at the A site in the Cterminal domain of the protein by iron presented as its complex with nitrilotriacetate [8]. Best results were obtained with 5% solutions of protein brought to 60% saturation with iron in 0.05 M Tris/0.01 M CaC12/0.001 M NaHCO 3 (pH 6.2). After making the preparation 2 M in urea to promote unfolding of the unoccupied N-terminal domain, and adding thermolysin to a ratio of 1 : 25 (w/w) with transferrin, incubation with shaking for 90 rain was carried out at 37°C. Typically, the absorbance at 470 nm of the preparation declined by 30-50% during this incubation. Proteolytic fragments were then isolated by gel-permeation chromatography on a 2.5 × 90 cm column of Sephadex G-75 SF, using 0.1 M NH4HCO 3 (pH 8.0) as buffer. The smaller, slower-moving components of the digestion mixture were further purified on a preparative electrofocusing column employing 2% ampholyte, pH 4-6.5 (Pharmacia). The fragment of interest migrated with an isoelectric point of 5.1. Final yields ranged from 5% to 10% of the starting protein, the theoretical maximum being near 50%. Physical measurements. Electron paramagnetic resonance spectra were obtained with a Varian E-9 EPR spectrometer interfaced to an Apple II + microcomputer. Optical spectra were recorded on a Cary model 14 spectrophotometer. Other procedures. SDS-gel electrophoresis followed procedures given by Maizel [13] with per-
iodic acid-Schiff staining for carbohydrate according to Segrest and Jackson [14]. Equilibrium dialysis was carried out by methods previously detailed [8]. Carbohydrate analyses, kindly performed by A. Adamany, were obtained using a Perkin-Elmer Model 910 gas chromatography with the techniques of Blumenfeld and Adamany [15]. Results and Discussion
Purification and identification of the fragment Following the initial chromatographic separation of the digestion mixture on a column of Sephadex G-75 SF, three bands were seen on SDS-gel electrophoresis, with apparent molecular weights of 38000, 42000 (the major component) and 48000. Additional minor components with mobilities similar to that of native transferrin were also present in some preparations, in which case rechromatography was carried out. Each components gave a strongly positive periodic acid-Schiff staining reaction for carbohydrate. Since both glycan chains of human serum transferrin are bound to the C-terminal half of the molecule, to asparagine residues 413 and 610 respectively [1], the fragments must have originated in this half. For contrast, the N-terminal fragment prepared by our modification of the method of Lineback-Zins and Brew [11,16] gave no periodic acid-Schiff reaction. Preparative isoelectric focusing of the lowmolecular-weight fraction from the G-75 SF column resolved two pink components, the major one migrating to pH 5.1. On SDS-gel electrophoresis, this component gave only a single prominent band which was periodic acid-Schiff-positive and had an apparent molecular weight of 42000 (Fig. 1). In some preparations a second band, with an apparent molecular weight of 47000, was observed., Because this minor band generally exhibited less than 2% of the intensity of the major band by densitometry, no further attempt at purification was made, and all additional studies were carried out on the electrofocused pH 5.1 fraction.
Characterization of the pH 5.1 fraction The absorbance at 280 nm of a 1% solution of the purified fragment was calculated as 11.55 from the previously measured absorbancies of the Nterminal fragment and native transferrin [11]. An
350
1
2
3
Mr
4
5
TABLE I C A R B O H Y D R A T E ANALYSIS OF p l 5.1 F R A G M E N T
ii - oooo-
- -
42
000
Carbohydrate
M o l / m o l of Fe-binding activity
Sialic acid Galactose Mannose N-Acetylglucosamine
2.5 2,6 3,3 4,2
--
A
The results of carbohydrate analysis are presented in Table I. From the concentrations of sugars detected, it seems likely that the bulk of the molecules in the preparation bear only a single glycan chain of the native protein.
B
Fig. 1, SDS-gel electrophoresis of transferrin and fragment eleetrofocussed at pH 5.1. (A) Coomassie blue stain; (B) periodic acid-Schiff stain. Lanes 1, 3, 5: fragment; lanes 2, 4: transferrin.
estimate of the equivalent weight for iron-binding was obtained by titration of the fragment in a 0.05 M Hepes/0.01 M HCO3- buffer (pH 7.45) with freshly prepared ferric nitrilotriacetate (1:3, pH 4.5) as titrant. A sharp end-point was reached, giving an equivalent weight near 35000 for the preparation used for most of the equilibrium dialysis studies. Discrepancies between the equivalent weights of the fragments, and the apparent molecular weights determined by SDS-gel electrophoresis are probably attributable to the insensitivity of the latter technique to carbohydrate moieties.
The strength of iron binding As with native transferrin and the N-terminal fragment, thermodynamic stability constants for iron binding by the C-terminal fragment were measured near pH 6.7 and pH 7.4 by the method of equilibrium dialysis with citrate as competing complexing agent [8,11]. A working range of Fe 3÷ activities was obtained by varying total Fe 3÷, keeping citrate concentration and ionic strength constant at each pH. The measured constants pertain to the overall reaction: Fe 3 + + fragment + H C O ~ ~ Fe 3 +-fragment-HCO 3 + 3H +-
Their values, and the experimental data from which they are calculated, are summarized in Tables II and III. At fixed pH and Pco2, effective or condi-
T A B L E II E Q U I L I B R I U M B I N D I N G O F Fe 3÷ TO C - T E R M I N A L T R A N S F E R R I N F R A G M E N T N E A R pH 7.4 Experiment
A B C
Total [Fe 3 + ] at equilibrium (/LM) protein
buffer
32.2 60.3 70.9
6.57 20.1 25.8
[ Fe 3 + bound] [fragment] a
pH
- log[Fe 3 + aq]
K
0.25 0.40 0.45
7.42 7.42 7.45
20.7 20.2 20.2
37,3 24.7 19.9 27.3±5,2 b
a Concentration of fragment near 100 # M in all experiments. b Mean 5: S.E.
351 TABLE III EQUILIBRIUM BINDING OF Fe 3+ TO C-TERMINAL TRANSFERRIN F R A G M E N T N E A R pH 6.7 Experiment
A B C D
Total [Fe 3+ ] at equilibrium ( p, M) protein
buffer
14.6 30.1 41.2 51.9
4.91 4.18 12.5 20.5
[Fragment] (/~ M)
56.8 94.4 94.2 91.3
[Fe 3+ bound] [ fragment]
0.17 0.27 0.30 0.34
pH
- l o g [ F e 3+ aq]
K
6.71 6.67 6.67 6.69
19.5 18.6 18.2 18
128 378 139 87 183___66 a
Mean_+ S.E.
tional stability constants for the apparent reaction: Fe 3 + + fragment ~ Fe 3+-fragment
may be derived. These constants, which provide more readily interpretable measures of the strength of iron binding under laboratory or physiological conditions, are given in Table IV, along with corresponding constants for the N-terminal fragment and the two sites of native transferrin. As is evident from this table, the tightness of the iron-protein bond is appreciably weaker in the C-terminal fragment than in the corresponding site of the native protein. This difference is more marked at
pH 6.7, where iron is bound 30-times more strongly to the A-site of transferrin than to the fragment, than at physiological pH where the ratio of binding strengths is 4. At pH 7.4, then, binding of iron to the A-site of transferrin is stabilized by about 10 kJ (2.4 kcal) per mol in the intact protein compared to the C-terminal fragment.
EPR spectra A compelling body of evidence that the two sites of iron-loaded transferrin have distinguisha-
A j
Fe2 Trons{errm
TABLE 1V EFFECTIVE STABILITY CONSTANTS FOR THE BIN D I N G OF IRON BY T R A N S F E R R I N A N D ITS SINGLE-SITED FRAGMENTS In air (Pco- = 3.6"10-4 atm) (M - ¢) pH 6.7 C-terminal Fragment Transferrin 'a'-Site a N-terminal Fragment b Transferrin 'b'-Sitea a From Ref. 8. b From Ref. 11.
In blood ( P c o , = 0.05 atm) (M - ~')
pH 7.4
1.1.1018
1.0.1020
1.5.10 22
2,9.1019
4.0.10 20
5.8.10 22
1.8-1017
5.7.1019
8.4" 1 0 21
<1.4.1018
6.8.1019
9.8.1021
B ~
ogment
cJ
~
A-B/IV ermlno151~eJ
,S I
I
1300
1400
I 1500 1600 GAUSS
l
I
r
1700
1800
1900
Fig. 2. EPR spectra of (A) diferric transferrin (2.2-10 -4) and
(B) C-terminal fragment (2.0-10-4). Spectrum C, representing the 'B' site in the N-terminal domain of transferrin, was obtained by computer subtraction of spectrum (B) from spectrum (A) after normalizing to protein concentration and instrumental gain. Buffer: 0.1 M Hepes/0.05 M KC1 (pH 7.4). Instrumental parameters: temperature, 77 K; microwave power, 10 mW; microwave frequency, 9.208 GHz; modulation frequency, 100 kHz; modulation amplitude, 10 (3.
352
ble EPR spectra now exists [8,11,17]. This evidence is further corroborated by the spectra in Fig. 2. The splitting between the two low-field peaks of the g' = 4.3 signal of the C-terminal fragment is 39 G. This contrasts with a splitting of 27 G in the spectrum of the single-sited fragment from the N-terminal domain of human transferrin [11]. As in our earlier study [11], the spectrum of the C-terminal fragment displays a greater amplitude than that of its N-terminal counterpart, consistent with a greater anisotropy in the spectrum of the latter. Integrated intensities of the spectra of the C-terminal fragment and diferric transferrin, each normalized to optical absorbance at 470 nm, agreed to within 1%, suggesting that the optical spectr~ of the two sites are similar. The EPR spectrum of the 65Cu2+ complex of the C-terminal fragment, with bicarbonate as the synergistic anion, was indistinguishable from the spectrum of native 65Cu-transferrin-bicarbonate. Attempts to prepare a complex of 65Cu 2+ and C-terminal fragment, with oxalate as co-anion, were unsuccessful, leading only to a mixture of nonspecific complexes and Cu-fragment-bicarbonate. In accord with earlier inferences [18], therefore, only the N-terminal site is capable of accepting oxalate when binding copper. Interestingly, and as yet inexplicably, the sites in N-terminal and C-terminal fragments alike accept oxalate when binding Fe 3+ (Fig. 3). In the preparation
from which the spectrum of this figure is obtained, there is at least lO-times as much Fe 3+ bound with oxalate as co-anion as is bound with bicarbonate, since the amplitude of a spectral line varies inversely with the square of its linewidth. This inference is corroborated by adding an excess of bicarbonate to the preparation, thereby largely abolishing the broad peaks of the oxalate complex while increasing the amplitude of the bicarbonate feature more than lO-fold. The two-sited nature of transferrin The susceptibility of the transferrins to proteolyric cleavage generating half-molecules with apparently intact metal binding sites has provided a useful tool for exploring the properties of the individual binding domains [11,19-23]. Differences between the chemical and physical properties of the sites in the intact protein have been well established [8-10,20], and many of these differences are at least grossly preserved in single-sited fragments when these fragments are sufficiently large to maintain intact binding sites [11,20,21,24]. The strength of iron binding, however, appears to be appreciably modulated by the environment offered by the intact protein, particularly at low pH. Whether this has any physiological implication is not clear. Interaction of ovotransferrin with its receptor on the chick embryo red blood cell depends on recognition regions in both halves of the transferrin molecule [23], and this may be a more important consequence of the two-domain structure of the protein. To dissect the effects of each iron-binding domain on the other remains a central goal in transferrin biochemistry. The availability of transferrin half-molecules offers unique advantage in achieving this goal [23].
Acknowledgements
600
8 0
I 1000
I 1200 GAUSS
I 1400
I 1600
2000
Fig. 3. EPR spectrum of the ternary complex of C-terminal fragment, Fe 3+ and oxalate. Fragment concentration, 1-10 -4 M. Instrumental parameters as in Fig. 2 except: microwave power, 20 mW.
This work was supported in part by Grant AM15056 from the National Institutes of Health. We are grateful to Dr. A. Adamany for carrying out the carbohydrate analysis of the purified fragment.
References 1 MacGillivray, R.T.A., Mendez, E. Shewale, J.G., Sinha, S.K., Lineback-Zins, J. and Brew, K. (1983) J. Biol. Chem. 258, 3543-3553
353 2 Mazurier, J., Metz-Boutigue, M.H., JollY, J., Spik, G., Montreuil, J. and Joll~s, P. (1983) Experientia 39, 135-141 3 Williams, J. (1982) Trends Biochem. Sci. 7, 394-397 4 Williams, J., Grace, S.A. and Williams, J.M. (1982) Biochem. J. 201,417-419 5 Leibman, A. and Aisen, P. (1979) Blood 53, 1058-1065 6 Williams, J. and Moreton, K. (1980) Biochem. J. 185, 483-488 7 Evans, R.W., Williams, J. and Moreton, K. (1982) Biochem. J. 201, 19-26 8 Aisen, P., Leibman, A. and Zweier, J. (1978) J. Biol. Chem. 253, 1930-1937 9 Chasteen, N.D., White, L.K. and Campbell, R.F. (1977) Biochemistry 16, 363-368 10 Baldwin, D.A., DeSousa, D.M.R., Von Wandruszka, R.M.A. (1982) Biochim. Biophys. Acta 719, 140-146 11 Zak, O., Leibman, A. and Aisen, P. (1983) Biochim. Biophys. Acta 742, 490-495 12 Evans, R.W. and Williams, J. (1980) Biochem. J. 541-546 13 Maizel, J.V., Jr. (1971) Methods Virol. 5, 179-246 14 Segrest, J.P. and Jackson, R.L. (1972) Methods Enzymol. 28, 54-63
15 Blumenfeld, O.O. and Adamany, A.M. (1978) Proc. Natl. Acad. Sci. USA 75, 2727-2731 16 Lineback-Zins, J. and Brew, K. (1980) J. Biol. Chem. 255, 708-713 17 Aasa, R. (1972) Biochem. Biophys. Res. Commun. 49, 806-812 18 Zweier, J.L. and Aisen, P. (1977). J. Biol. Chem. 252, 6090-6096 19 Bluard-Beconinck, J.M., Williams, J., Evans, R.W., Van Snick, J., Osinski, P.A. and Masson, P.L. (1978) Biochem. J. 171, 321-327 20 Evans, R.W. and Williams, J. (1978) Biochem. J. 173, 543-552 21 Keung, W.-M., Azari, P. and Phillips, J.L. (1982) J. Biol. Chem. 257, 1177-1183 22 Keung, W.-M. and Azari, P. (1982) J. Biol. Chem. 257, 1184-188 23 Brown-Mason, A. and Woodworth, R.C. (1984) J. Biol. Chem. 259, 1866-1873 24 Legrand, D., Mazurier, J., Metz-Boutigue, M.H., Joll6s, J., Joll6s, P., Montreuil, J. and Spik, G. (1984) Biochim. Biophys. Acta 787, 90-96