- Email: [email protected]
Unusual stability properties of a reptilian ceruloplasmin
ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 279, No. 1, May 15, pp. S-13,1990
Unusual Stability Properties of a Reptilian Ceruloplasmin’ Giovanni Musci,2,* Marina Carbonaro,? Angelo Adriani,? Antonio Galtieri,$ and Lilia Calabrese”f
Amalia Lania,$
*Center of Molecular Biology of C.N.R., University of Rome “La Sapienza”, piazzale A. Moro 5-0018.5 Rome, Italy; TDepartment of Biochemical Sciences, University of Rome “La Sapienza”, piazzale A. Moro 5-00185 Rome, Italy; and $Department of Organic and Biological Chemistry, University of Messina, Salita Sperone 31, S. Agata-98166 Messina, Italy
Received September 8,1989, and in revised form December 26,1989
Ceruloplasmin from the turtle Caretta caretta was isolated to purity by using the single-step procedure recently developed by us to purify sheep and chicken ceruloplasmin. It has a M, of ca. 145,000 and a total copper content of 5.1 f 0.2 atoms of copper per molecule, 50% of which are detectable by EPR. The spectroscopic features include an absorption maximum at 603 nm in the electronic spectrum and the total absence of any resonance attributable to Type 2 copper in the EPR spectrum. Turtle ceruloplasmin was found to be unusually resistant to aging and proteolysis, when compared to ceruloplasmins isolated from other species. p-Phenylendiamine oxidase activity measurements revealed an unusually low catalytic efficiency, while the kinetic parameters of Fe(I1) oxidation were consistent with those (Cl 1990 reported for other species of ceruloplasmin. Academic
Press,
Inc.
Ceruloplasmin is a multifunctional glycoprotein present in large amounts in the vertebrate plasma. It is a metalloprotein containing several copper atoms per molecule, which can be distinguished, on the basis of their spectroscopical attributes, into three classes (1). Type 1 copper is responsible for the unusually strong electronic absorption around 600 nm and is paramagnetic. Type 2 copper is essentially silent in the optical spectrum, but contributes to the EPR3 spectrum with a lineshape quite ’ This work was partially supported by the CNR Special Project “Chimica Fine.” * To whom correspondence should be addressed at Center of Molecular Biology of CNR % Department of Biochemical Sciences, University of Rome “La Sapienza,” piazzale A. Moro 5-00185 Rome, Italy. ’ Abbreviations used: DTPA, diethylenetriaminepentaacetic acid; EDTA, ethylenediaminetetraacetic acid; EPR, electron paramagnetic resonance; NADH, dihydronicotinamide adenine dinucleotide; PPD, p-phenylenediamine; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.
typical of regularly coordinated tetragonal complexes. Type 3 copper consists of a pair of metal ions, antiferromagnetically coupled to give an EPR-silent species. It absorbs in the near-uv region of the electronic spectrum, giving a shoulder around 330 nm. The molecular parameters (molecular weight, total copper content, percentage of paramagnetic copper), as well as the spectroscopic parameters (including optical and EPR features) of this protein have been a matter of debate for a long time. In particular, ceruloplasmin has been long considered a multisubunit protein, as it was usually purified as a mixture of several components (24). It was later shown that addition of a protease inhibitor to the plasma permits the purification of a singlechain protein (5,6). The M, of human ceruloplasmin was then fixed at 132,000, after completion of the amino acid sequence (7). Furthermore, the total copper content and the relative ratios between the three types of copper inside the molecule have never been definitively clarified, although a stoichiometry of 6 or 7 atoms of copper per molecule is generally accepted, at least for the human protein (8). We recently purified with a single-step method and characterized ceruloplasmin from chicken plasma and found it significantly different from all mammalian proteins so far studied (9). In particular, we measured a copper content of only 5 copper atoms per molecule, which, on the basis of their spectroscopic features, would be distributed as two Type 1 ions and a trinuclear cluster with unusual magnetic properties (9). The same conclusions were successively drawn in a reinvestigation of the spectroscopic and functional properties of a mammalian ceruloplasmin (17), leading to the hypothesis that the native structure of the protein is only maintained when a nondestructive purification protocol is adopted. In a continuing effort to characterize ceruloplasmin from other classes of Vertebrates, we isolated and partially characterized a reptilian ceruloplasmin, from the turtle Caretta caretta.
8 All
OOO3-9861/90 $3.00 Copyright 8 1990 by Academic Press, Inc. rights of reproduction in any form reserved.
PROTEOLYTICALLY MATERIALS
AND
STABLE
METHODS
All reagents were of analytical grade and were purchased from Merck (Darmstadt, Germany). Sepharose 4B was obtained from Pharmacia-LKB Biotechnology Inc. (Uppsala). Molecular weight markers were from Sigma Chemical Co. (St. Louis). The plasma from C. car&a was a generous gift of Museo Civic0 di Storia Naturale (Calimera di Lecce, Italy) that worked within the Italian Project on turtles, in collaboration with WWF Italia. The plasma had been taken from animals collected after traumatic collisions with boat propellers and for whom any attempts at life saving were in vain. Heparin was used to prevent blood coagulation. The plasma was kept frozen at -20°C until use. Turtle ceruloplasmin was isolated by a single-step procedure similar to that used for chicken ceruloplasmin (9). Plasma (300 ml) was passed down on a column packed with 100 ml of derivatized Sepharose 4B (10) and the resin was then sequentially washed with increasing concentrations of phosphate buffer, pH 7.4 (3, 10, 20, 40, and 60 mM). The protein was finally eluted with a 80 mM solution of the same buffer and concentrated on Amicon cells. The whole procedure was carried out at 4°C and took less than 4 h. Mammalian, sheep, and human ceruloplasmins and chicken ceruloplasmin were purified according to Calabrese cl al. (10) and (9), respectively. Protein concentration was determined according to Lowry et al. (11). Polyacrylamide gels were stained for oxidase activity by 1-h incubation with a saturated aqueous solution of benzidine as substrate. SDSgel electrophoresis was performed according to Weber and Osborn (12). The following proteins of known molecular weight were (M,, used: cr,-macroglobulin (M,, 770,000), human ceruloplasmin 132,000), phosphorylase b (M,, 97,400), bovine serum albumin (M,, 66,000), and carbonic anhydrase (M,, 29,000). The determinations were carried out both in the presence and in the absence of 1% & mercapthoethanol. Copper content was measured chemically by the method of Brumby and Massey (13). The amount of paramagnetic copper was estimated by double integration of the EPR signal versus a standard solution of Cu(II)-EDTA (lo-20 mW, 100 K). Enzyme activity as PPD oxidase was assayed by the coupled assay of Lovstad and Frieden (14), on the basis of NADH consumption by the first oxidation product of PPD. The incubation mixture contained 0.25 mM DTPA, 0.25 mh~ NADH, 1.3 WM ceruloplasmin, and enough NaCl to keep the Cl concentration constant (9.6 mM) in 0.1 M phosphate buffer at pH 6.3. Ferroxidase activity was assayed according to Osaki et al. (15) on a Gilson 5/G oxygraph at 25°C and pH 6.3, in 0.1 M phosphate buffer. Iron stock solutions were freshly made by dissolving Fe(I1) ammonium sulfate in 10 ’ M HCI to prevent metal autoxidation. Incubation mixtures contained 2 pM ceruloplasmin and Fe(I1) ranging from 0.04 to 0.5 rnM. A blank without ceruloplasmin was run for each assay to subtract the contribution due to Fe(II) autoxidation to the observed 0, consumption. Optical spectra were recorded on a Perkin-Elmer 330 spectrometer equipped with a Haake Mod. G temperature control unit. EPR spectra were measured on a Varian E-9 spectrometer operating at 9.15 GHz equipped with a Stelar variable temperature unit and interfaced to a Stelar Prometheus Data System for computer analysis and handling of the spectra. Computer simulations were run with a FORTRAN program for IBM kindly provided by Dr. Bencini, Iiniversity of Florence.
RESULTS
A typical elution profile of turtle ceruloplasmin after chromatography on derivatized Sepharose 4B is depicted in Fig. 1. The protein elutes as a single, symmetric peak with 80 mM phosphate buffer at pH 7.4. Only the fractions with a ABOO/APKOratio higher than 0.04 (indicated by a bar in Fig. 1) were collected, with a final yield of approximately 30 mg of ceruloplasmin per liter of plasma.
TURTLE
CERULOPLASMIN
9
A standard polyacrylamide electrophoresis analysis of the purified protein revealed a single component by both protein- and oxidase activity-stained gels (Fig. 1, inset). The substantial homogeneity of the protein was indicated by the presence of a single band on SDS-electrophoresis gels both in the presence and in the absence of P-mercapthoethanol (see lane C of Fig. 3). The molecular weight of turtle ceruloplasmin was determined as M,, 145,000 ?Z5000 in a separate set of experiments, by comparison of the relative mobility on SDS-PAGE of turtle ceruloplasmin vs a series of standard molecular weight markers, including human ceruloplasmin, whose molecular weight has been unequivocally determined from the primary structure (7). Chemical analysis of the copper content reproducibly gave a value of 5.1 * 0.2 atoms of copper per molecule. The amount of EPR-detectable copper, as measured by double integration of the EPR signal (Fig. 2A), was 50% of the total copper content, corresponding to ca. 2.5 copper atoms. The EPR spectrum (Fig. 2A, upper curve) was totally lacking any signal with spectral parameters typical of the Type 2 copper of multinuclear blue oxidases (1). The lineshape was unchanged in the temperature range lOO200 K, and did not show any effect from microwave power saturation between 2 and 20 mW. A computer simulation of the EPR spectrum is also shown in Fig. 2A (lower curve). The experimental lineshape could be accounted for by the presence of two copper ions with slightly different magnetic parameters (Q = 2.217-2.221; AlI = 68.4-94.8 G). These parameters are typical of Type 1 copper atoms (1). To best fit the experimental curve, however, a small contribution (ca. 0.3 spins per molecule) of a broad, ca. 250 G, unstructured signal centered at g = 2.106 had t,o be added. This is consistent with the fact that the amount of paramagnetic copper always exceeded 2 copper atoms per molecule in our determinations. The absorption peak due to the Type 1 copper in the optical spectrum (Fig. 2B) was centered at 603 nm, a value identical to that found for chicken ceruloplasmin (9). The molar extinction coefficient at 603 nm was 10,000 * 400 M ml cm-‘, which is one of the highest so far reported for ceruloplasmin and did not change upon treatment with H202, which has been shown to cause an -15% increase in the intensity of this band in mammalian, but not chicken, ceruloplasmin (9, 16). The shoulder at 330 nm was well resolved, with a ratio between the optical densities at 603 and 330 nm of 2.7. It has been reported in the past (8) that both the optical and the EPR spectra of ceruloplasmin undergo severe and irreversible modifications upon aging, consisting primarily in the appearance of a signal typical of Type 2 copper in the EPR spectrum and in an increase of the optical density at 330 nm in the electronic spectrum (9, 17). Unexpectedly, there was no appreciable
10
MlJSCl
i
A260 k-----l
3
ET AI,
i
3
A 603 [-----)
0.E
0.4
1.04
0.2
1.02
ELUTION
VOLUME
(ml)
FIG. 1. Elution profile of turtle ceruloplasmin. The resin was sequentially washed with 3 mM (l), 20 mM (2), 40 mM (3), 60 mM (4), and 80 mu (5) phosphate huff’er, pH 7.4. Inset: Polyacrylamide gel electrophoresis of turtle ceruloplasmin stained for activity (A) and proteins (B).
modification of the spectroscopic properties of turtle ceruloplasmin after standing at -20°C for over three months. In particular, both the optical densities at 603 and 330 nm were unaffected, and the shape and intensity of the EPR spectrum were unchanged. Since ceruloplasmin is known to be extremely susceptible to the action of serine proteases (8), we tested the resistance of turtle ceruloplasmin toward proteolytic attacks. Figure 3 shows the SDS electrophoretic pattern of 4 PM chicken and turtle ceruloplasmin before and after treatment with 0.2 PM trypsin at 25°C and pH 7.4 (trypsin:ceruloplasmin ratio, 1:20). Remarkably, the turtle protein was essentially unaltered by incubation with trypsin, with only a faint band at higher mobility appearing after 2 h (lanes C and D). The results were identical when the trypsin concentration was raised up to 10~’ M (trypsin:ceruloplasmin ratio, 1:4). In contrast, chicken ceruloplasmin was totally proteolyzed after analogous treatment with trypsin (lanes A and B), as manifested by the total disappearance of the band corresponding to the native protein. Fragments similar to those previously reported for the human protein subjected to the same treatment (5) were generated in this case. The turnover kinetic parameters of turtle ceruloplasmin for either p-phenylendiamineor ferro-oxidase ac-
tivity are reported in Table I in comparison with those of chicken and sheep ceruloplasmins. While only minor differences among the three species were observed in the kinetic parameters of the enzymatic Fe(I1) + Fe(II1) conversion, turtle ceruloplasmin showed the lowest K,,, and V,,, values when examined as PPD oxidase. The results obtained with the sheep and the chicken enzyme are in line with previous reports (14,18). To test whether the differences in the catalytic properties of turtle vs other ceruloplasmins were due to anomalous electron transfer between the various copper sites of the protein, the redox behavior of turtle ceruloplasmin was tested by reacting the protein with ascorbate. Under our experimental conditions (phosphate buffer 0.1 M, pH 7.4, 25-C), reduction was very fast, as monitored by the total disappearance of the optical bands at 330 nm and 603 nm in ca. 15 min. The 330-nm band disappeared at a rate faster than that of the 603nm band (Fig. 4) and was paralleled by the appearance of a Type 2 copper transient EPR signal, as previously observed for chicken (9) and sheep ceruloplasmin (17). All the EPR features of Type 1 copper decreased homogeneously during the reduction course, indicating that the two blue ions have similar rates of reduction by ascorbate, in spite of their different EPR spectral parameters. When reduced turtle ceruloplasmin was reex-
PROTEOLYTICALLY
I
,
1
I
I
400
500
600
700
WAVELENGTH
STABLE
TURTLE
11
CERCJLOPLASMIN
I
(nm I
FIG. 2. (A) upper curve: low temperature EPR spectrum of 0.125 mM turtle ceruloplasmin in 0.1 M phosphate buffer, pH 7.4. Experi mental settings: microwave power, 20 mW; frequency, 9.178 GHz; modulation amplitude, 10 G; tield set, :X)00 G; scan range, 1000 (:; temperature, 100 K. Lower curve: computer simulation of the spectrum. Three contributions were summed to reproduce the experimental lineshape: a copper ion with g, = 2.055, 8,) = 2.221, A, = 8 G, Ai, 68.4 G. LW, = 30 G. LW), = 30 (+; a copper ion with gl = 2.045, ~1) 7 2.217, A. ~ 8 G, AlI = 94.8 (i, LW;, = 40 G, LW), = 55 G; and an isotropic signal (0.3 spins per molecule) centered at g = 2.106 with LW y 2.50 G. The hyperfines are shown at 10 times higher gain. (B) Optical spectrum of 0.084 mM turtle ceruloplasmin in 0.1 M phosphate buffer, pH 7.4.
FIG. 3. SDS gel electrophoresis of (A) chicken ceruloplasmin; (H) chicken ceruloplasmin incubated 2 h with 10~“M trypsin at pH 7.4; (C) turtle ceruloplasmin; (D) turtle ceruloplasmin incubated 2 h with 10 ‘M trypsin at pH 7.4.
the same single-step procedure (9,17). We cannot completely rule out the possibility that the measured value (5.1) is the result of heterogeneity of the preparation, although there is no evidence in our electrophoretic analyses, both in denaturing and in nondenaturing conditions, for such a heterogeneity. One would also expect in this case a greater variation than that actually observed (k0.2). It is worth remembering, on the other hand, that a structural and functional model based on a stoichiometry of 5 copper atoms per molecule has been
posed to air, a complete recovery of all optical (Fig. 4) and EPR native spectroscopic parameters was achieved within a few minutes.
TABLE Turnover
DISCUSSION
The presence of ceruloplasmin in the blood of several reptiles has already been suggested on the basis of oxidase activity assays (19, 20). This is, however, the first report on a purified reptilian ceruloplasmin. There are no gross differences in the chemico-physical parameters of turtle ceruloplasmin as compared to other ceruloplasmins. However, turtle ceruloplasmin has a stoichiometry of ca. 5 copper atoms per molecule, as recently reported for other ceruloplasmins purified with
Kinetic Parameters Species, with PPD
I
of Ceruloplasmin from or Fe(I1) as Substrate
wr)
Fe(H) Ceruloplasmin Turtle Chicken Sheep
K,, (mM) 0.1% 0.13 O.lri
Various
vln,xn
K,, (mM)
V”,,,,”
120 150 82
0.019 0.085 0.22
0.63 2.50 2.10
’ Expressed as pM(& )/min/mg ceruloplasmin. ‘Expressed as p~(NADH)/min/mg ceruloplasmin.
12
MUSCI
MINUTES FIG. 4. Time courses of the optical absorbances at 330 nm (0) and at 603 nm (a) during anaerobic reduction ofO.1 mM turtle ceruloplasmin with 1 mM ascorbate in 0.1 M phosphate buffer, pH 7.4 (left) and after reexposition to air (right). Changes are expressed as percentages of the difference between the optical densities of the fully oxidized and the fully reduced protein.
recently proposed for ceruloplasmin (9, 17). The model, which implies the presence of a trinuclear copper cluster, has found indirect support from an X-ray study of ascorbic oxidase, a multinuclear blue oxidase found in plants, where a structural trinuclear unit has been revealed (21). The optical features of the turtle protein are very close to those reported for the chicken protein, i.e., the absorption maximum of the blue band is centered at 603 nm, at variance with that always observed for all mammalian proteins (610 nm). The basis of the blue-shift of this charge-transfer band is uncertain, but it might indicate differences in the region of the Type 1 copper binding site, such as the number of hydrogen bonds between the thiolate sulfur liganding the copper and NH groups on the peptide chain (22). Turtle ceruloplasmin preparations very reproducibly lack the EPR signal typical of Type 2 copper of multinuclear blue oxidases, and only show EPR resonances attributable to two nonequivalent Type 1 copper ions. It is interesting to remember that only recently this property has been established as a diagnostic feature of native ceruloplasmin, which is only maintained when a suitable nondestructive purification protocol starting from fresh plasma is adopted. This is the first case reported so far of a ceruloplasmin prepared from frozen plasma that lacks the EPR signal of Type 2 copper (17). The nonequivalence of the two blue copper ions is not surprising as it has been reported several times in the past for ceruloplasmins from other species (1). On the other hand, the computer simulation reported in Fig. 2A reveals that a broad signal centered at g = 2.106 underlies the features of Type 1 copper. The origin of this signal, which accounts for ca. 0.3 spins per molecule, has not yet been established to certainty. In a recent report on chicken ceruloplasmin (9), we postulated that it arises from the trinuclear unit which constitutes the oxygen-binding
ET AL.
site of the protein. The presence of this broad resonance, anyhow, can also explain why the amount of EPR-detectable copper in turtle ceruloplasmin is always slightly greater than two copper atoms per molecule. The redox behavior of turtle ceruloplasmin shows the decrease in the absorption at 330 nm was faster than that at 603 nm. This phenomenon has been recently demonstrated to reflect the structural integrity of the protein (17). The sole, significant difference between turtle and other ceruloplasmins, as far as the redox reactivity is concerned, is in the reduction times: turtle ceruloplasmin turned out to be reduced very quickly by ascorbate (ca. 15 min) when compared with sheep or chicken ceruloplasmins under the same experimental conditions (ca. 2 h) (9, 17). On the other hand, reoxidation by oxygen occurred on a comparable time scale, at least with respect to the chicken protein. A faster reduction rate, as monitored by the disappearance of the optical band at 603 nm, implies a more efficient electron transfer from the substrate to Type 1 copper. This fact can be the basis also for the lower Km value of turtle ceruloplasmin with PPD (Table I), but does not necessarily imply a higher catalytic efficiency. It is generally accepted that the rate-limiting step in the catalytic mechanism of ceruloplasmin is the reaction between oxygen and the reduced enzyme (23). In fact, this may explain the low V,,,,, value (Table I). On the other hand, the efficiency of the enzymatic oxidation of the ferrous ion by turtle ceruloplasmin is comparable to other proteins examined. This can be interpreted as further evidence that different substrates possibly use different electron transfer pathways and even bind to different electron accepting sites on the protein. A very peculiar property of turtle ceruloplasmin is its unusual resistance to both aging and proteolytic stresses. The stability to prolonged storage can be used to rationalize the observation that it retains all spectroscopic features typical of the intact protein even when prepared from frozen plasma. On the other hand, this is also the first ceruloplasmin reported so far to be unaffected by a proteolytic attack (5, 24). It is interesting to note that, due to the extreme susceptibility to proteolytic attacks, human ceruloplasmin has been long considered a multisubunit protein, as it was normally purified as a mixture of several components (2-4). While we cannot give, at this juncture, a rigorous explanation at the molecular level, this might indicate that in the turtle protein the key bond(s), the rupture of which is essential to further fragmentation, is (are) shielded from the destructive action of trypsin. Although no data are available yet on the sugar content of turtle ceruloplasmin, the most likely candidate as shielding factor is the carbohydrate moiety, as it has been implied in the different susceptibility to proteolysis between the two forms (I and II) of human ceruloplasmin, as well as between homologous domains of the same polypeptide chain (25).
I’ROTEOLYTICALLY
STABLE
TURTLE
In conclusion, the present report on turtle ceruloplasmin demonstrates that, while ceruloplasmins isolated from different species mostly retain the same spectroscopic parameters, indicative of an essentially invariant molecular architecture, subtle differences, reflected in different kinetic parameters and stability toward aging and proteolysis, could substantially modify the functional valence of the protein.
10. Calabrese, L., Mateescu,
REFERENCES
15. Osaki, S., Johnson,
1. Fee, ,J. A. (1975) in Structure and Bonding (Dunitz, J. D., Hemmerick, P., Helm, R. H., Ibers, ,J. A., ,Jorgensen, C. K., Neilands, ,J. B., Keinen. D., and Williams R. d. P., Eds.) Vol. 23, pp. l-60, Springer-Verlag, Berlin. 2. Poulik, M. D. (1962) Nature (London) 194,842. 3. Kasper, C. B. and Deutsch, H. F. (196X) J. Riol. Chem. 238,23252x17. 4. Poillon, W. N., and Bearn, A. (:. (1966) B&him. Biophys. Acta
127,407. 5. Ryden, L. (1971) ~EHS Lett. 18, 321-325 6. Ryden, I,. (1972) Eur. J. Biochem. 26,380~384 7. Takahashi, N., Ortel, T. L., and Putnam, F. W. (1984) Proc. Natl. Acnd.
Sci. l!SA
81,390~394.
8. Ryden, I,. (1984) in Copper Proteins and Copper Enzymes (Lontie. L.. Ed.) Vol. 3. pp. 37-100, CRC Press Inc., Boca Raton, Florida. 9. (‘alabrese, I,.. Carhonaro, M., and Musci, G. (1988) J. Biol. (‘hem. 263, 6480~64833.
13
CERULOPLASMIN M. A., Carhonaro,
M., and Mondovi’,
B.
(1988)Biochem.Znt.16,199-208. 11. Lowry, 0. H., Rosebrough, N. ,J., Farr, A. L., and Randall, (1951) J. Hiol. Chcm. 193,265-275.
R. ,J.
12. Weber, K., and Oshorn, M. (1969) J. Riol. Chem. 244,4406-4412. 13. Brumby, P. E., and Massey, V. (1967) in Methods in Enzymology (Estahrook, R. W., and Pullman, M. E., Eds.) Vol. 10, pp. 473 474, Academic Press, San Diego. 14. Lovstad, R. A., andFrieden, E. (1973) Acta (‘hem. Stand. 27,121l 126. D. A., and Frieden,
E. (1966) J. Hiol. C’hcm.
241,2746-2751. 16. Calabrese, L., and Leuzzi, U. (1984) Biochem.
Int. 8, 35-39.
17. Calabrese, L., Carbonaro, M., and Musci, G. (1989) J. Biol. Chem. 264,6183-6187. 18. Frieden, E., and Hsieh, H. S. (1976) Adu. Enzymol. 44,187-196. 19. Seal, Ii. S. (1965) Camp. Biochpm. Physiol. 13, 14%159. 20. Perrier, H., Delcroix, ,J. P., Perrier, C., and Gras, ,J. (1974) romp. Biochrm.
Physiol.
49B, 679-685.
21. Messerschmidt, A., Rossi, A., Ladenstein, R., Huber, R., Bolognesi, M., Gatti, G., Marchesini, A., Petruzzelli, R., and FinazziAgro, A. (1989) J. Mol. Biol. 206, ,513. 22. Baker, E. N. (1988) J. Mol. Biol. 203,1071-1095. 23. Gunnarson, P., Nylen, V., and Pettersson, G. (1973) Eur. J. Biothem.
37,47-50.
24. Kingston,
I. B., Kingston,
Nntl.
Sci. lJ.SA 74,
Acad.
25. Ortel, T. L., Takahashi, Acnd.
Sci. lJ.SA
B. L., and Putnam, F. W. (1977) I’roc. 5377-5381. N., and Putnam, F. W. (1984) I’roc. Nat/.
81.4761-4765.
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 279, No. 1, May 15, pp. S-13,1990
Unusual Stability Properties of a Reptilian Ceruloplasmin’ Giovanni Musci,2,* Marina Carbonaro,? Angelo Adriani,? Antonio Galtieri,$ and Lilia Calabrese”f
Amalia Lania,$
*Center of Molecular Biology of C.N.R., University of Rome “La Sapienza”, piazzale A. Moro 5-0018.5 Rome, Italy; TDepartment of Biochemical Sciences, University of Rome “La Sapienza”, piazzale A. Moro 5-00185 Rome, Italy; and $Department of Organic and Biological Chemistry, University of Messina, Salita Sperone 31, S. Agata-98166 Messina, Italy
Received September 8,1989, and in revised form December 26,1989
Ceruloplasmin from the turtle Caretta caretta was isolated to purity by using the single-step procedure recently developed by us to purify sheep and chicken ceruloplasmin. It has a M, of ca. 145,000 and a total copper content of 5.1 f 0.2 atoms of copper per molecule, 50% of which are detectable by EPR. The spectroscopic features include an absorption maximum at 603 nm in the electronic spectrum and the total absence of any resonance attributable to Type 2 copper in the EPR spectrum. Turtle ceruloplasmin was found to be unusually resistant to aging and proteolysis, when compared to ceruloplasmins isolated from other species. p-Phenylendiamine oxidase activity measurements revealed an unusually low catalytic efficiency, while the kinetic parameters of Fe(I1) oxidation were consistent with those (Cl 1990 reported for other species of ceruloplasmin. Academic
Press,
Inc.
Ceruloplasmin is a multifunctional glycoprotein present in large amounts in the vertebrate plasma. It is a metalloprotein containing several copper atoms per molecule, which can be distinguished, on the basis of their spectroscopical attributes, into three classes (1). Type 1 copper is responsible for the unusually strong electronic absorption around 600 nm and is paramagnetic. Type 2 copper is essentially silent in the optical spectrum, but contributes to the EPR3 spectrum with a lineshape quite ’ This work was partially supported by the CNR Special Project “Chimica Fine.” * To whom correspondence should be addressed at Center of Molecular Biology of CNR % Department of Biochemical Sciences, University of Rome “La Sapienza,” piazzale A. Moro 5-00185 Rome, Italy. ’ Abbreviations used: DTPA, diethylenetriaminepentaacetic acid; EDTA, ethylenediaminetetraacetic acid; EPR, electron paramagnetic resonance; NADH, dihydronicotinamide adenine dinucleotide; PPD, p-phenylenediamine; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.
typical of regularly coordinated tetragonal complexes. Type 3 copper consists of a pair of metal ions, antiferromagnetically coupled to give an EPR-silent species. It absorbs in the near-uv region of the electronic spectrum, giving a shoulder around 330 nm. The molecular parameters (molecular weight, total copper content, percentage of paramagnetic copper), as well as the spectroscopic parameters (including optical and EPR features) of this protein have been a matter of debate for a long time. In particular, ceruloplasmin has been long considered a multisubunit protein, as it was usually purified as a mixture of several components (24). It was later shown that addition of a protease inhibitor to the plasma permits the purification of a singlechain protein (5,6). The M, of human ceruloplasmin was then fixed at 132,000, after completion of the amino acid sequence (7). Furthermore, the total copper content and the relative ratios between the three types of copper inside the molecule have never been definitively clarified, although a stoichiometry of 6 or 7 atoms of copper per molecule is generally accepted, at least for the human protein (8). We recently purified with a single-step method and characterized ceruloplasmin from chicken plasma and found it significantly different from all mammalian proteins so far studied (9). In particular, we measured a copper content of only 5 copper atoms per molecule, which, on the basis of their spectroscopic features, would be distributed as two Type 1 ions and a trinuclear cluster with unusual magnetic properties (9). The same conclusions were successively drawn in a reinvestigation of the spectroscopic and functional properties of a mammalian ceruloplasmin (17), leading to the hypothesis that the native structure of the protein is only maintained when a nondestructive purification protocol is adopted. In a continuing effort to characterize ceruloplasmin from other classes of Vertebrates, we isolated and partially characterized a reptilian ceruloplasmin, from the turtle Caretta caretta.
8 All
OOO3-9861/90 $3.00 Copyright 8 1990 by Academic Press, Inc. rights of reproduction in any form reserved.
PROTEOLYTICALLY MATERIALS
AND
STABLE
METHODS
All reagents were of analytical grade and were purchased from Merck (Darmstadt, Germany). Sepharose 4B was obtained from Pharmacia-LKB Biotechnology Inc. (Uppsala). Molecular weight markers were from Sigma Chemical Co. (St. Louis). The plasma from C. car&a was a generous gift of Museo Civic0 di Storia Naturale (Calimera di Lecce, Italy) that worked within the Italian Project on turtles, in collaboration with WWF Italia. The plasma had been taken from animals collected after traumatic collisions with boat propellers and for whom any attempts at life saving were in vain. Heparin was used to prevent blood coagulation. The plasma was kept frozen at -20°C until use. Turtle ceruloplasmin was isolated by a single-step procedure similar to that used for chicken ceruloplasmin (9). Plasma (300 ml) was passed down on a column packed with 100 ml of derivatized Sepharose 4B (10) and the resin was then sequentially washed with increasing concentrations of phosphate buffer, pH 7.4 (3, 10, 20, 40, and 60 mM). The protein was finally eluted with a 80 mM solution of the same buffer and concentrated on Amicon cells. The whole procedure was carried out at 4°C and took less than 4 h. Mammalian, sheep, and human ceruloplasmins and chicken ceruloplasmin were purified according to Calabrese cl al. (10) and (9), respectively. Protein concentration was determined according to Lowry et al. (11). Polyacrylamide gels were stained for oxidase activity by 1-h incubation with a saturated aqueous solution of benzidine as substrate. SDSgel electrophoresis was performed according to Weber and Osborn (12). The following proteins of known molecular weight were (M,, used: cr,-macroglobulin (M,, 770,000), human ceruloplasmin 132,000), phosphorylase b (M,, 97,400), bovine serum albumin (M,, 66,000), and carbonic anhydrase (M,, 29,000). The determinations were carried out both in the presence and in the absence of 1% & mercapthoethanol. Copper content was measured chemically by the method of Brumby and Massey (13). The amount of paramagnetic copper was estimated by double integration of the EPR signal versus a standard solution of Cu(II)-EDTA (lo-20 mW, 100 K). Enzyme activity as PPD oxidase was assayed by the coupled assay of Lovstad and Frieden (14), on the basis of NADH consumption by the first oxidation product of PPD. The incubation mixture contained 0.25 mM DTPA, 0.25 mh~ NADH, 1.3 WM ceruloplasmin, and enough NaCl to keep the Cl concentration constant (9.6 mM) in 0.1 M phosphate buffer at pH 6.3. Ferroxidase activity was assayed according to Osaki et al. (15) on a Gilson 5/G oxygraph at 25°C and pH 6.3, in 0.1 M phosphate buffer. Iron stock solutions were freshly made by dissolving Fe(I1) ammonium sulfate in 10 ’ M HCI to prevent metal autoxidation. Incubation mixtures contained 2 pM ceruloplasmin and Fe(I1) ranging from 0.04 to 0.5 rnM. A blank without ceruloplasmin was run for each assay to subtract the contribution due to Fe(II) autoxidation to the observed 0, consumption. Optical spectra were recorded on a Perkin-Elmer 330 spectrometer equipped with a Haake Mod. G temperature control unit. EPR spectra were measured on a Varian E-9 spectrometer operating at 9.15 GHz equipped with a Stelar variable temperature unit and interfaced to a Stelar Prometheus Data System for computer analysis and handling of the spectra. Computer simulations were run with a FORTRAN program for IBM kindly provided by Dr. Bencini, Iiniversity of Florence.
RESULTS
A typical elution profile of turtle ceruloplasmin after chromatography on derivatized Sepharose 4B is depicted in Fig. 1. The protein elutes as a single, symmetric peak with 80 mM phosphate buffer at pH 7.4. Only the fractions with a ABOO/APKOratio higher than 0.04 (indicated by a bar in Fig. 1) were collected, with a final yield of approximately 30 mg of ceruloplasmin per liter of plasma.
TURTLE
CERULOPLASMIN
9
A standard polyacrylamide electrophoresis analysis of the purified protein revealed a single component by both protein- and oxidase activity-stained gels (Fig. 1, inset). The substantial homogeneity of the protein was indicated by the presence of a single band on SDS-electrophoresis gels both in the presence and in the absence of P-mercapthoethanol (see lane C of Fig. 3). The molecular weight of turtle ceruloplasmin was determined as M,, 145,000 ?Z5000 in a separate set of experiments, by comparison of the relative mobility on SDS-PAGE of turtle ceruloplasmin vs a series of standard molecular weight markers, including human ceruloplasmin, whose molecular weight has been unequivocally determined from the primary structure (7). Chemical analysis of the copper content reproducibly gave a value of 5.1 * 0.2 atoms of copper per molecule. The amount of EPR-detectable copper, as measured by double integration of the EPR signal (Fig. 2A), was 50% of the total copper content, corresponding to ca. 2.5 copper atoms. The EPR spectrum (Fig. 2A, upper curve) was totally lacking any signal with spectral parameters typical of the Type 2 copper of multinuclear blue oxidases (1). The lineshape was unchanged in the temperature range lOO200 K, and did not show any effect from microwave power saturation between 2 and 20 mW. A computer simulation of the EPR spectrum is also shown in Fig. 2A (lower curve). The experimental lineshape could be accounted for by the presence of two copper ions with slightly different magnetic parameters (Q = 2.217-2.221; AlI = 68.4-94.8 G). These parameters are typical of Type 1 copper atoms (1). To best fit the experimental curve, however, a small contribution (ca. 0.3 spins per molecule) of a broad, ca. 250 G, unstructured signal centered at g = 2.106 had t,o be added. This is consistent with the fact that the amount of paramagnetic copper always exceeded 2 copper atoms per molecule in our determinations. The absorption peak due to the Type 1 copper in the optical spectrum (Fig. 2B) was centered at 603 nm, a value identical to that found for chicken ceruloplasmin (9). The molar extinction coefficient at 603 nm was 10,000 * 400 M ml cm-‘, which is one of the highest so far reported for ceruloplasmin and did not change upon treatment with H202, which has been shown to cause an -15% increase in the intensity of this band in mammalian, but not chicken, ceruloplasmin (9, 16). The shoulder at 330 nm was well resolved, with a ratio between the optical densities at 603 and 330 nm of 2.7. It has been reported in the past (8) that both the optical and the EPR spectra of ceruloplasmin undergo severe and irreversible modifications upon aging, consisting primarily in the appearance of a signal typical of Type 2 copper in the EPR spectrum and in an increase of the optical density at 330 nm in the electronic spectrum (9, 17). Unexpectedly, there was no appreciable
10
MlJSCl
i
A260 k-----l
3
ET AI,
i
3
A 603 [-----)
0.E
0.4
1.04
0.2
1.02
ELUTION
VOLUME
(ml)
FIG. 1. Elution profile of turtle ceruloplasmin. The resin was sequentially washed with 3 mM (l), 20 mM (2), 40 mM (3), 60 mM (4), and 80 mu (5) phosphate huff’er, pH 7.4. Inset: Polyacrylamide gel electrophoresis of turtle ceruloplasmin stained for activity (A) and proteins (B).
modification of the spectroscopic properties of turtle ceruloplasmin after standing at -20°C for over three months. In particular, both the optical densities at 603 and 330 nm were unaffected, and the shape and intensity of the EPR spectrum were unchanged. Since ceruloplasmin is known to be extremely susceptible to the action of serine proteases (8), we tested the resistance of turtle ceruloplasmin toward proteolytic attacks. Figure 3 shows the SDS electrophoretic pattern of 4 PM chicken and turtle ceruloplasmin before and after treatment with 0.2 PM trypsin at 25°C and pH 7.4 (trypsin:ceruloplasmin ratio, 1:20). Remarkably, the turtle protein was essentially unaltered by incubation with trypsin, with only a faint band at higher mobility appearing after 2 h (lanes C and D). The results were identical when the trypsin concentration was raised up to 10~’ M (trypsin:ceruloplasmin ratio, 1:4). In contrast, chicken ceruloplasmin was totally proteolyzed after analogous treatment with trypsin (lanes A and B), as manifested by the total disappearance of the band corresponding to the native protein. Fragments similar to those previously reported for the human protein subjected to the same treatment (5) were generated in this case. The turnover kinetic parameters of turtle ceruloplasmin for either p-phenylendiamineor ferro-oxidase ac-
tivity are reported in Table I in comparison with those of chicken and sheep ceruloplasmins. While only minor differences among the three species were observed in the kinetic parameters of the enzymatic Fe(I1) + Fe(II1) conversion, turtle ceruloplasmin showed the lowest K,,, and V,,, values when examined as PPD oxidase. The results obtained with the sheep and the chicken enzyme are in line with previous reports (14,18). To test whether the differences in the catalytic properties of turtle vs other ceruloplasmins were due to anomalous electron transfer between the various copper sites of the protein, the redox behavior of turtle ceruloplasmin was tested by reacting the protein with ascorbate. Under our experimental conditions (phosphate buffer 0.1 M, pH 7.4, 25-C), reduction was very fast, as monitored by the total disappearance of the optical bands at 330 nm and 603 nm in ca. 15 min. The 330-nm band disappeared at a rate faster than that of the 603nm band (Fig. 4) and was paralleled by the appearance of a Type 2 copper transient EPR signal, as previously observed for chicken (9) and sheep ceruloplasmin (17). All the EPR features of Type 1 copper decreased homogeneously during the reduction course, indicating that the two blue ions have similar rates of reduction by ascorbate, in spite of their different EPR spectral parameters. When reduced turtle ceruloplasmin was reex-
PROTEOLYTICALLY
I
,
1
I
I
400
500
600
700
WAVELENGTH
STABLE
TURTLE
11
CERCJLOPLASMIN
I
(nm I
FIG. 2. (A) upper curve: low temperature EPR spectrum of 0.125 mM turtle ceruloplasmin in 0.1 M phosphate buffer, pH 7.4. Experi mental settings: microwave power, 20 mW; frequency, 9.178 GHz; modulation amplitude, 10 G; tield set, :X)00 G; scan range, 1000 (:; temperature, 100 K. Lower curve: computer simulation of the spectrum. Three contributions were summed to reproduce the experimental lineshape: a copper ion with g, = 2.055, 8,) = 2.221, A, = 8 G, Ai, 68.4 G. LW, = 30 G. LW), = 30 (+; a copper ion with gl = 2.045, ~1) 7 2.217, A. ~ 8 G, AlI = 94.8 (i, LW;, = 40 G, LW), = 55 G; and an isotropic signal (0.3 spins per molecule) centered at g = 2.106 with LW y 2.50 G. The hyperfines are shown at 10 times higher gain. (B) Optical spectrum of 0.084 mM turtle ceruloplasmin in 0.1 M phosphate buffer, pH 7.4.
FIG. 3. SDS gel electrophoresis of (A) chicken ceruloplasmin; (H) chicken ceruloplasmin incubated 2 h with 10~“M trypsin at pH 7.4; (C) turtle ceruloplasmin; (D) turtle ceruloplasmin incubated 2 h with 10 ‘M trypsin at pH 7.4.
the same single-step procedure (9,17). We cannot completely rule out the possibility that the measured value (5.1) is the result of heterogeneity of the preparation, although there is no evidence in our electrophoretic analyses, both in denaturing and in nondenaturing conditions, for such a heterogeneity. One would also expect in this case a greater variation than that actually observed (k0.2). It is worth remembering, on the other hand, that a structural and functional model based on a stoichiometry of 5 copper atoms per molecule has been
posed to air, a complete recovery of all optical (Fig. 4) and EPR native spectroscopic parameters was achieved within a few minutes.
TABLE Turnover
DISCUSSION
The presence of ceruloplasmin in the blood of several reptiles has already been suggested on the basis of oxidase activity assays (19, 20). This is, however, the first report on a purified reptilian ceruloplasmin. There are no gross differences in the chemico-physical parameters of turtle ceruloplasmin as compared to other ceruloplasmins. However, turtle ceruloplasmin has a stoichiometry of ca. 5 copper atoms per molecule, as recently reported for other ceruloplasmins purified with
Kinetic Parameters Species, with PPD
I
of Ceruloplasmin from or Fe(I1) as Substrate
wr)
Fe(H) Ceruloplasmin Turtle Chicken Sheep
K,, (mM) 0.1% 0.13 O.lri
Various
vln,xn
K,, (mM)
V”,,,,”
120 150 82
0.019 0.085 0.22
0.63 2.50 2.10
’ Expressed as pM(& )/min/mg ceruloplasmin. ‘Expressed as p~(NADH)/min/mg ceruloplasmin.
12
MUSCI
MINUTES FIG. 4. Time courses of the optical absorbances at 330 nm (0) and at 603 nm (a) during anaerobic reduction ofO.1 mM turtle ceruloplasmin with 1 mM ascorbate in 0.1 M phosphate buffer, pH 7.4 (left) and after reexposition to air (right). Changes are expressed as percentages of the difference between the optical densities of the fully oxidized and the fully reduced protein.
recently proposed for ceruloplasmin (9, 17). The model, which implies the presence of a trinuclear copper cluster, has found indirect support from an X-ray study of ascorbic oxidase, a multinuclear blue oxidase found in plants, where a structural trinuclear unit has been revealed (21). The optical features of the turtle protein are very close to those reported for the chicken protein, i.e., the absorption maximum of the blue band is centered at 603 nm, at variance with that always observed for all mammalian proteins (610 nm). The basis of the blue-shift of this charge-transfer band is uncertain, but it might indicate differences in the region of the Type 1 copper binding site, such as the number of hydrogen bonds between the thiolate sulfur liganding the copper and NH groups on the peptide chain (22). Turtle ceruloplasmin preparations very reproducibly lack the EPR signal typical of Type 2 copper of multinuclear blue oxidases, and only show EPR resonances attributable to two nonequivalent Type 1 copper ions. It is interesting to remember that only recently this property has been established as a diagnostic feature of native ceruloplasmin, which is only maintained when a suitable nondestructive purification protocol starting from fresh plasma is adopted. This is the first case reported so far of a ceruloplasmin prepared from frozen plasma that lacks the EPR signal of Type 2 copper (17). The nonequivalence of the two blue copper ions is not surprising as it has been reported several times in the past for ceruloplasmins from other species (1). On the other hand, the computer simulation reported in Fig. 2A reveals that a broad signal centered at g = 2.106 underlies the features of Type 1 copper. The origin of this signal, which accounts for ca. 0.3 spins per molecule, has not yet been established to certainty. In a recent report on chicken ceruloplasmin (9), we postulated that it arises from the trinuclear unit which constitutes the oxygen-binding
ET AL.
site of the protein. The presence of this broad resonance, anyhow, can also explain why the amount of EPR-detectable copper in turtle ceruloplasmin is always slightly greater than two copper atoms per molecule. The redox behavior of turtle ceruloplasmin shows the decrease in the absorption at 330 nm was faster than that at 603 nm. This phenomenon has been recently demonstrated to reflect the structural integrity of the protein (17). The sole, significant difference between turtle and other ceruloplasmins, as far as the redox reactivity is concerned, is in the reduction times: turtle ceruloplasmin turned out to be reduced very quickly by ascorbate (ca. 15 min) when compared with sheep or chicken ceruloplasmins under the same experimental conditions (ca. 2 h) (9, 17). On the other hand, reoxidation by oxygen occurred on a comparable time scale, at least with respect to the chicken protein. A faster reduction rate, as monitored by the disappearance of the optical band at 603 nm, implies a more efficient electron transfer from the substrate to Type 1 copper. This fact can be the basis also for the lower Km value of turtle ceruloplasmin with PPD (Table I), but does not necessarily imply a higher catalytic efficiency. It is generally accepted that the rate-limiting step in the catalytic mechanism of ceruloplasmin is the reaction between oxygen and the reduced enzyme (23). In fact, this may explain the low V,,,,, value (Table I). On the other hand, the efficiency of the enzymatic oxidation of the ferrous ion by turtle ceruloplasmin is comparable to other proteins examined. This can be interpreted as further evidence that different substrates possibly use different electron transfer pathways and even bind to different electron accepting sites on the protein. A very peculiar property of turtle ceruloplasmin is its unusual resistance to both aging and proteolytic stresses. The stability to prolonged storage can be used to rationalize the observation that it retains all spectroscopic features typical of the intact protein even when prepared from frozen plasma. On the other hand, this is also the first ceruloplasmin reported so far to be unaffected by a proteolytic attack (5, 24). It is interesting to note that, due to the extreme susceptibility to proteolytic attacks, human ceruloplasmin has been long considered a multisubunit protein, as it was normally purified as a mixture of several components (2-4). While we cannot give, at this juncture, a rigorous explanation at the molecular level, this might indicate that in the turtle protein the key bond(s), the rupture of which is essential to further fragmentation, is (are) shielded from the destructive action of trypsin. Although no data are available yet on the sugar content of turtle ceruloplasmin, the most likely candidate as shielding factor is the carbohydrate moiety, as it has been implied in the different susceptibility to proteolysis between the two forms (I and II) of human ceruloplasmin, as well as between homologous domains of the same polypeptide chain (25).
I’ROTEOLYTICALLY
STABLE
TURTLE
In conclusion, the present report on turtle ceruloplasmin demonstrates that, while ceruloplasmins isolated from different species mostly retain the same spectroscopic parameters, indicative of an essentially invariant molecular architecture, subtle differences, reflected in different kinetic parameters and stability toward aging and proteolysis, could substantially modify the functional valence of the protein.
10. Calabrese, L., Mateescu,
REFERENCES
15. Osaki, S., Johnson,
1. Fee, ,J. A. (1975) in Structure and Bonding (Dunitz, J. D., Hemmerick, P., Helm, R. H., Ibers, ,J. A., ,Jorgensen, C. K., Neilands, ,J. B., Keinen. D., and Williams R. d. P., Eds.) Vol. 23, pp. l-60, Springer-Verlag, Berlin. 2. Poulik, M. D. (1962) Nature (London) 194,842. 3. Kasper, C. B. and Deutsch, H. F. (196X) J. Riol. Chem. 238,23252x17. 4. Poillon, W. N., and Bearn, A. (:. (1966) B&him. Biophys. Acta
127,407. 5. Ryden, L. (1971) ~EHS Lett. 18, 321-325 6. Ryden, I,. (1972) Eur. J. Biochem. 26,380~384 7. Takahashi, N., Ortel, T. L., and Putnam, F. W. (1984) Proc. Natl. Acnd.
Sci. l!SA
81,390~394.
8. Ryden, I,. (1984) in Copper Proteins and Copper Enzymes (Lontie. L.. Ed.) Vol. 3. pp. 37-100, CRC Press Inc., Boca Raton, Florida. 9. (‘alabrese, I,.. Carhonaro, M., and Musci, G. (1988) J. Biol. (‘hem. 263, 6480~64833.
13
CERULOPLASMIN M. A., Carhonaro,
M., and Mondovi’,
B.
(1988)Biochem.Znt.16,199-208. 11. Lowry, 0. H., Rosebrough, N. ,J., Farr, A. L., and Randall, (1951) J. Hiol. Chcm. 193,265-275.
R. ,J.
12. Weber, K., and Oshorn, M. (1969) J. Riol. Chem. 244,4406-4412. 13. Brumby, P. E., and Massey, V. (1967) in Methods in Enzymology (Estahrook, R. W., and Pullman, M. E., Eds.) Vol. 10, pp. 473 474, Academic Press, San Diego. 14. Lovstad, R. A., andFrieden, E. (1973) Acta (‘hem. Stand. 27,121l 126. D. A., and Frieden,
E. (1966) J. Hiol. C’hcm.
241,2746-2751. 16. Calabrese, L., and Leuzzi, U. (1984) Biochem.
Int. 8, 35-39.
17. Calabrese, L., Carbonaro, M., and Musci, G. (1989) J. Biol. Chem. 264,6183-6187. 18. Frieden, E., and Hsieh, H. S. (1976) Adu. Enzymol. 44,187-196. 19. Seal, Ii. S. (1965) Camp. Biochpm. Physiol. 13, 14%159. 20. Perrier, H., Delcroix, ,J. P., Perrier, C., and Gras, ,J. (1974) romp. Biochrm.
Physiol.
49B, 679-685.
21. Messerschmidt, A., Rossi, A., Ladenstein, R., Huber, R., Bolognesi, M., Gatti, G., Marchesini, A., Petruzzelli, R., and FinazziAgro, A. (1989) J. Mol. Biol. 206, ,513. 22. Baker, E. N. (1988) J. Mol. Biol. 203,1071-1095. 23. Gunnarson, P., Nylen, V., and Pettersson, G. (1973) Eur. J. Biothem.
37,47-50.
24. Kingston,
I. B., Kingston,
Nntl.
Sci. lJ.SA 74,
Acad.
25. Ortel, T. L., Takahashi, Acnd.
Sci. lJ.SA
B. L., and Putnam, F. W. (1977) I’roc. 5377-5381. N., and Putnam, F. W. (1984) I’roc. Nat/.
81.4761-4765.