Protein Thioacylation - American Chemical Society

F- 56325 Lorient, France; and Université de Caen, ISMRA, Boulevard Maréchal Juin,. F-14050 Caen, France. Received January 11, 2000; Revised Manuscri...
0 downloads 0 Views 313KB Size
Biomacromolecules 2000, 1, 400-406

400

Protein Thioacylation: 2. Reagent Stability in Aqueous Media and Thioacylation Kinetics Guy Levesque,*,† Philippe Arse` ne,‡ Vale´ rie Fanneau-Bellenger,‡ and Thi-Nha` n Pham‡ Centre de Recherche, Universite´ de Bretagne-Sud, Boulevard Flandres-Dunkerque, F- 56325 Lorient, France; and Universite´ de Caen, ISMRA, Boulevard Mare´ chal Juin, F-14050 Caen, France Received January 11, 2000; Revised Manuscript Received April 3, 2000

Several thioacylating reagents have been tested toward hydrolysis under conditions suitable for protein modifications: 20-35 °C and buffered solutions at pH 7.5-8.5. Aliphatic dithioesters are sufficiently stable in aqueous media at room temperature (or below) if protein modification reaction time does not exceed 24 h, whereas at 35 °C reaction times must be limited to a few hours. Kinetic data obtained in gelatin thioacylation at room temperature using aliphatic dithioesters and dithio acid are consistent with a second-order reaction rate with respect to amine concentration. The pH dependence of the second-order reaction rate constants indicate that dithioester reacts exclusively with the free amine form of lysine residue, whereas dithiocarboxylate ion reacts with both amine and ammonium ion, probably through a more complex mechanism. Interestingly thioacylation using dithio acids may be obtained in pH near neutrality or in slightly acidic media, thus offering protein modification possibilities at pH 5-9. Thioacylation reaction rates may be expressed as R ) -(dAt/dt) ) k[H3O+]-bAt2[thioacylating agent] in which At is the amine concentration at time t, constants k and b depending on the reagent nature. Introduction Lysine residues are the most frequently used targets in protein chemical modification, corresponding to the primary amine high nucleophilicity, particularly under basic and neutral conditions. Our preliminary reports in enzyme modification using dithioesters (thioacylation) concerned grafting reactions of either amphililic poly(ethylene oxide) oligomers or hydrophobic alkyl chains or phenyl rings.1,2 This new method for protein chemical modification is based on the known thioacylation of primary and secondary amines by dithioesters and dithio acids which proceeds rapidly at room temperature and was originally applied to esters of amino acids or dipeptides in order to promote a new protecting group.3 As the back-pathway from thioacylated products to free amino groups could not be easily performed at that time,4 the use of thioacylating reagents as protecting groups during peptide syntheses or reactions was rapidly eliminated.5 However, these older works have suggested the use of dithioesters as protein thioacylating agents. Amine thioacylation is a quite an interesting method for protein modification for the following reasons: • Primary and secondary amines react very rapidly at pH 7 and above.6 • Thioacylation reaction undergoes catalysis only by amines themselves,7 although the use of strongly basic catalysts has not been seriously considered. • Dithioester reagents are generally stable near room temperature, and their syntheses require only current chem* To whom correspondence should be addressed. † Universite ´ de Bretagne-Sud. ‡ Universite ´ de Caen, ISMRA.

istry lab equipment. However, dithio acids are sensitive toward oxidation and heat,8 and they are usually stored as ammonium salts9 at low temperatures. • Thioacylation modifies selectively aliphatic amines and converts them into thioamides which cannot be easily reversed. • Some water-soluble reagents have been previously described (mainly carboxymethyl dithioesters R-CS-SCH2-CO2H),10 and we have described the synthesis of numerous ω-functionalized reagents (see part 1). • The availability of bifunctional reagents (bis(dithioesters) and bis(dithio acids)) offers new pathways for coupling between biomacromolecules and other polymers, providing they both carry free amino groups. During thioacylation, other functional groups (such as hydroxyl, indole, imidazole) currently present in proteins remain unchanged although some exchange is theoretically possible between cysteine thiol groups and dithioesters, depending on the leaving group character of the various thiols. We have observed and used for synthetic purpose such an exchange between a carboxymethyl dithioester and N-acetylcysteine,1 but return to the free original thiol occurs immediately when some ammonia (or primary or secondary amine) is added to the mixture to destroy unreacted thioacylating reagent (Figure 1). Thioacylation reaction monitoring and protein modification evaluation are quite easy to realize as the reagents used (dithioester or dithio acid as dithiocarboxylate ion) as well as thioacylation products (thioamide-containing protein) present quite different and strong UV-visible absorptions, i.e.:

10.1021/bm000037b CCC: $19.00 © 2000 American Chemical Society Published on Web 07/08/2000

Biomacromolecules, Vol. 1, No. 3, 2000 401

Protein Thioacylation

Figure 1. Irreversible lysine thioacylation reaction and reversible exchange between cysteine residue and dithioesters (E ) enzyme or protein; R′ ) alkyl group or H). Chart 1

Figure 2. Hydrolytic behavior of dithioester 7 at room temperature and various pH (variations of optical density near 310 nm vs time).

• Aliphatic dithioesters absorb strongly near 310 nm ( g 12 000 L‚mol-1‚cm-1) and also in the visible range (bright yellow-orange color) • Aliphatic thioamides absorb at 270 nm ( g 14 500 L‚mol-1‚cm-1) but present only weak absorption in the visible range (colorless or faint yellow compounds) unless conjugation occurs with a benzene ring or a carbon-carbon double bond. Thioamides are stable in aqueous media (but mild oxidizing conditions may lead to amide formation), and modified proteins could be submitted to trinitrobenzenesulfonic acid (TNBS) alkylation11 to mesure the remaining free amine contents. In this part of our protein thioacylation study, we will describe two main features: a study of reagent stability in aqueous systems as a function of pH and temperature; the kinetics of lysine modification in a model protein (gelatin).

Figure 3. Hydrolytic behavior of dithioester 7 at 35 °C and various pH (variations of optical density near 310 nm vs time).

1. Thioacylation Reagents Stability in Aqueous Systems Thiocarbonyl compounds are thermodynamically unstable toward hydrolysis9 as the CdO double bond formation enthalpy is larger than the corresponding energy for CdS double bond (enthalpy difference ca. 180 kJ‚mol-1). However thiocarbonyl compound hydrolysis kinetics is generally slow enough to allow their isolation from aqueous media (although dithio acids isolated from Grignard synthesis followed by acid hydrolysis12 always contain small amounts of monothio acids, identified by their IR carbonyl absorption). Thus, before use as protein thioacylating reagents, it was important to test the stability of dithioesters and dithio acids toward aqueous media under normal protein modification conditions. Two dithioesters and one dithio acid were tested under experimental conditions (time, 1-100 h; temperature, e 35 °C; and pH in the range 7-9) suitable for protein modifications without major denaturation, as appears through enzymatic activity conservation in most cases.2 The tested reagents are shown in Chart 1. Results Most influent parameters on stability against water were found to be pH and temperature. The following results were

Figure 4. Hydrolytic behavior of dithioester 5 at room temperature and various pH (variations of optical density near 310 nm vs time).

obtained at 20 °C (Figures 2 and 4) and 35 °C (Figures 3 and 5) and variable (buffered) pH from 7 to 8.5, corresponding to a useful range for protein thioacylation with dithioesters. Residual dithioesters concentrations were determined through their absorbance near 310 nm ( g 12 000). This determination avoids the need to take into account conjugated unsaturated dithioesters eventually formed through secondary coupling reactions (Dieckmann condensation or analogous) as we have previously observed for aliphatic dithioesters in the presence of amines. Time Influence on Aliphatic Dithioester Hydrolysis. Residual dithioesters concentrations vary as a function of time with influence of pH at 20 °C as well at 35 °C: the reagent loss lies in the range 5-25% after 24 h near room

402

Biomacromolecules, Vol. 1, No. 3, 2000

Figure 5. Hydrolytic behavior of dithioester 5 at 35 °C and various pH (variations of optical density near 310 nm vs time).

temperature in the case of 7a which appears to be the more hydrolyzable dithioester. Compound 7d loss remains below 10% at room temperature for the same reaction time and pH range. Temperature Influence on Aliphatic Dithioester Hydrolysis. At 35 °C, a temperature often used in biochemistry, pH appears to be a less important factor on dithioester hydrolysis rates: losses of 50 ( 10% are noted over 24 h for dithioester 7d and amount to 65-80% for dithioester 7a. No attempt was made to identify hydrolysis products. However new absorption in UV and visible spectra near 330-340 and 430-440 nm are consistent with the previously observed base-catalyzed condensation of dithioesters which results in the occurrence of conjugated structures.

From these results, we may conclude aliphatic dithioesters are sufficient stable in aqueous media at room temperature (or below) if protein modification reaction time is less than 24 h. When heating to 35 °C is required, reaction times must be limited to a few hours; otherwise, the reagent would be destroyed through hydrolysis. Dithiocarboxylate Ion Hydrolysis. Similar hydrolysis tests were realized to quantify the dithiocarboxylic acid hydrolysis rate. We have used the salt CH3-CH2-CH2CH2-CS2-,+N(CH3)4 (26f) in phosphate buffer solutions: at room temperature no degradation could be measured after 24 h; degradation near ca. 15% was noted at 40 °C for the same time at pH between 6.0 and 8.0 (the usual pH range for protein modification using dithio acids). More important degradation occurred in mixed acetate-phosphate buffers, especially with bis(dithiocarboxylic acid) salts (up to 75% after 48 h). 2. Protein Thioacylation Kinetics in Buffered Aqueous Media Thioacylation of primary and secondary amines is a very rapid reaction: in a 1 M solution, using aliphatic reagents, the dithioester bright yellow-orange coloration disappears within a few seconds and solvent evaporation affords thioamide and thiol in quantitative yields. Kinetic studies in

Levesque et al.

such reactions were often considered as impossible9 until computer-assisted data acquisition allowed UV-visible absorption measurements to be made several times per second. We have previously realized such a kinetic study in various organic solvents, which allowed us to propose a coherent mechanism.7 Among these results, we retain mostly a first order with respect to dithioester and a low activation energy; i.e., practical thioacylation reaction rates are nearly independent of temperature variations. Choice of reagents. Model Protein. For this protein thioacylation kinetic study, we needed a model protein inert toward pH variations and as transparent as possible in the UV regions corresponding to thiocarbonyl intense absorptions (270 nm in thioamides and 310 nm in dithioesters). Gelatin seemed to be a good model for this purpose. Its reactive free amine content is quite sufficient (ca. 5% in weight) although it may be the sum of lysine and some hydroxylysine and ornithine units (the later resulting from guanidine hydrolysis during collagen denaturation). Moreover gelatin shows no strong UV absorption in the selected range. Thioacylating Agents. Two aliphatic carboxymethyl dithioester CH3-(CH2)xCS-S-CH2-COOH (24e, x ) 0 and 24g, x ) 6) were used as well as a quaternary ammonium dithiocarboxylate, 26h CH3(CH2)7CS2-,+N(CH3)4. Thus, we are able to detect chain length influence in dithioesters as thioacylating agents and to compare thioacylation kinetics using dithoesters or dithio acids (as dithiocarboxylate ion). All these reagents were used at different pH (8-10 for esters, 5-10 with the acid), allowing us to determine correlations existing between pH and thioacylation kinetics. Some other agents bearing quite different ω-functional groups (namely quaternary ammonium and phosphonic acid) were also tested in order to investigate the usefulness of the results obtained with the preceding neutral reagents. Experimental Section All reactions were run under a positive nitrogen pressure. Reagents were from Sigma-Alfrich. UV spectra were obtained on a Perkin-Elmer λ15 UV-visible spectrometer. Reaction cells were thermostated and stirred magnetically. Abbreviations: BAPA, R-N-benzoyl D,L-arginine-p-nitroanilide; Tris, N,N,N-tris(hydroxymethyl)aminomethane; EDTA, ethylenediaminetetraacetic acid; TNBS, 2,4,6-trinitrobenzenesulfonic acid; TNP-NH2, 2,4,6-trinitrophenylamine. Thioacylating reagents were described in the preceding paper.13 Caution! Never use thioacylating agents in a buffer containing ammonia and ammonium salts (including salts of primary or secondary amines) as they will react quickly. Hydrolytic Behavior. The tested reagent was dissolved in 0.1 M phosphate buffer at the required pH, in a 1 mm thick quartz cell and placed in a water bath either at 20 or 35 °C. UV-visible spectra were run at regular time intervals, and the specific dithioester absorbance near 310 nm was used to determine the rate of active group disappearance. Buffer pH 8-9: phosphate 0.1 M. Buffer pH 9.5: Borax-sodium hydroxide 0.1 M. Buffer pH 10: boric acid-potassium chloridesodium hydroxide 0.1 M. Thioacylation Kinetics. Gelatin (pork skin) was used as a model protein as it offers only weak UV-visible absorption in the regions

Protein Thioacylation

Biomacromolecules, Vol. 1, No. 3, 2000 403

Figure 6. Gelatin thioacylation kinetics in the presence of excess carboxymethyl dithioester (influence of pH).

where dithioesters and thioamides absorb strongly (above 280 nm), thus allowing reaction monitoring through these absorptions as well as through conventional methods. Gelatin solutions were obtained at 0.1% (g/mL): after swelling of gelatin dispersion at room temperature; dissolution occured by heating at 40 °C. (a) Lysine Residue Determination in Gelatin Solutions. Solutions used: solution A, sodium borate buffer: pH 9.5 (3.81 g of Borax and 0.40 g of NaOH in 100 mL of water; solution B, stopping solution (1.536 g of H2NaPO4‚2H2O and 0.0189 g of Na2SO3 in 100 mL of water); solution C, 1.1 M TNBS solution (0.381 g in 1 mL of water) Five test gelatin solutions were made, each containing 1 mg/ mL. Three analysis were run, and the mean result was used. In a disposable cell were introduced 0.5 mL of gelatin solution, 0.5 mL of solution A, and 0.02 mL of solution C. After 15 min reaction at room temperature, 2 mL of solution B was added at once. After 15 min, optical densities were determined against pure water as reference. (b) Thioacylation Kinetics. At time zero, 100 mL of gelatin solution and 0.375 mmol of thioacylating reagent in 0.5 mL dimethyl sulfoxide (DMSO) were mixed. 5-mL test samples were taken off at time ) 1, 5, 10, 20,30, 45, 60, 90, 120, 150, and 180 min and eventually after 12, 24, 48, or 70 h. They were poured at once into a large excess of acetone (a few drops of acetic acid were sometimes added). All reagents were soluble in water-excess acetone mixtures so that the precipitate contains only modified gelatin after acetone washings. These samples of modified gelatin were dried in vacuo and then dissolved in 5 mL of water. Protein contents in these solutions were determined according to the Lowry procedure (bicinchoninic acid and CuSO4).16 The 562 nm OD calibration curve was obtained using native gelatin. The TNBS determination of remaining free amino groups was corrected for the exact protein content. A more direct way for thioacylation kinetic evaluation may be realized through the evolution of UV spectra. Similar results were obtained, but with less accurate results as concentrations are fixed by the reagents and products molar absorbances. Typically, 53.2 mg of gelatin were dissolved in the required buffer solution (100 mL) in order to obtain ca. 2 × 10-5 mol lysine in this volume. A reagent solution was prepared, containing 2 × 10-4 mol in 100 mL of buffer solution. The sample cell was filled at time zero with 1.5 mL of each solution. The reference sample was obtained by mixing 1.5 mL of

Figure 7. Gelatin thioacylation kinetics in the presence of excess tetramethylammonium dithiononanoate (influence of pH).

reagent solution and 1.5 mL of buffer solution. The OD at 270 nm was characteristic of aliphatic thioamides.

Kinetic Results A reagent excess was constantly used so that its concentration could be assumed to be constant: in Figures 6 and 7 are reported the reacted amine content as a function of time at different pH values. It appears clearly that medium acidity or basicity is a major factor whatever the reagent used, but dithioesters react more and more slowly at pH lower than 8 where as dithiocarboxylate is able to react significantly even at pH 6 or 5, i.e., in weakly acidic media. At elevated pH however, dithioesters are more efficient thioacylating agents than dithio acids. The use of excess reagent reduces to zero the apparent rate order relative to dithioester or dithio acid. In Figures 8 and 9, we present the analytical treatment of these results assuming a pseudo-zero order in acylating reagent concentration and a second order in amine group concentration. Curves fitting to straight lines is quite correct and allows to determine the apparent second-order rate constant for the thioacylation of amine-bearing residues in gelatin at various pH. This correlation appears clearly in Figure 10. A linear dependence of log k vs PH is shown in this log-log chart, the slopes of curves relative to the two dithioesters being nearly identical, wheras the dithio acid curve presents a much lower slope. We can try to correlate these data to chemical basic notions.

404

Biomacromolecules, Vol. 1, No. 3, 2000

Levesque et al.

Figure 8. Second-order analytical treatment for gelatin thioacylation data using carboxymethyl dithioesters. Table 1. Constant Value from Thioacylation Kinetic Data thioacylating reagent CH3C(dS)CH2COOH CH3(CH2)6C(dS)S-CH2COOH CH3(CH2)7C(dS)S-N(CH3)4+ a

Figure 9. Second-order analytical treatment for gelatin thioacylation data using the tetramethylammonium salt of dithiononadecanoic acid.

Figure 10. Apparent thioacylation rate constant variation as a function of pH and reagent nature.

In aqueous media, amines are equilibrated with ammonium ions, resulting from protonation by the solvent. Then, log[amine] is increasing regularly, proportional to pH (ap-

ka 10-13

7.3 × 14.8 × 10-13 6.5 × 10-7

b 0.92 0.86 0.24

k unit: L3‚mol-3‚s-1 (assuming b ) 1).

proximation roughly valid if pH e pKa - 1, then until pH e 10 if pKa corresponds to the amine conjugate acid). It seems obvious that dithioesters react mainly with free amines (the observed slope is near 0.9 (constant b in Table 1) for both dithioesters, to be compared to a theoretical value b ) 1 in such an hypothesis). Such a difference might be correlated with the reagent slow hydrolysis or more probably to the amine group determination method in gelatin: thioacylating agents may have a different accessibility to free amines than the reagent used for residual amine concentration determination. In contrast, dithio acids, owing to their pKa value near 2,14 do not undergo significant ionization change between pH 5 and 10: they remain totally ionized over the whole pH range. However, dithiocarboxylate ions may be attracted more intensely by ammonium ions than by free amines: the data might represent the sum of two reactions of dithiocarboxylate ions respectively with free amines (at pH > 7) and protonated ammonium ions at pH < 9. However this hypothesis is unable to explain the observed linear pH dependence. Another explanation might be found in a more complex mechanism in which free amines, although not directly involved in thioacylation by dithiocarboxylate ions, would interactsfor examplesin a deprotonation step: such a mechanism might allow for the observed kinetic rate variation against pH (slope b ) 0.24). In summary, thioacylation reaction rates may be expressed as R)-

dAt ) k[H3O+]-bAt2[thioacylating agent] dt

Protein Thioacylation

Biomacromolecules, Vol. 1, No. 3, 2000 405

Figure 11. Kinetic curves for gelatin thioacylation by means of some ω-functional dithioesters.

Figure 12. Protein thioacylation kinetic data using aromatic dithioesters.

in which At is the amine concentration at time t, with constants k and b depending on the reagent nature as reported in Table 1. ω-Functional and Aromatic Dithioesters. As various reagents have been used in protein thioacylation, we have tested some of them with respect to reaction kinetics: the presence of an ω-functional group (quaternary ammonium as well as phosphonic acid) leads to kinetics data which do not fit correctly with the preceding mathematical treatment. After ca. 25 h, the ammonium-containing reagent offered a net decrease in thioacylation kinetics, which may be attributed to its slow decomposition. At pH 10, this decomposition becomes more rapid and thioacylation degrees remain low. A similar behavior was observed with the phosphonic acid-containing dithioester; however, thioacylation is going on significantly even after a 50 h reaction time, particularly at pH 8. From these results one may conclude these reagents must be best used for reaction times under 24 h and in the search of limited degrees of substitution. Some aromatic thioacylating reagents were also tested, and results are collected in Figure 12: in both cases thioacylation reactions are limited to 25-30% of the initial lysine groups. Aromatic thiocarbonyl compounds are less electrophilic than aliphatic ones as expected from the mesomery betwen the phenyl ring and functional group. Then the thioacylation rate constant must be lowered and therefore secondary reactions such as thiocarbonyl hydrolysis become more important in

the case of aromatic dithioesters. The preceding remarks on reaction time and limited DS are valid in this case as well. 3. Practical Procedure for Protein Modification Kinetics determinations were made using a large excess of thioacylating reagents (usually 10 times) in order to get an apparent zero-order reaction with respect to dithioesters or dithiocarboxylate ions. Practically the first protein (enzymes) modifications1,2 were realized at moderate temperatures (20-35 °C) and relatively basic pH (8.5-9) for several hours. The kinetic study of primary amines thioacylation realized in organic solvents8 led to the conclusion the activation energy is quite small in the temperature range 0-40 °C, and thus thioacylation reaction rates might be considered as nearly temperature independent. Thus, protein thioacylation may be obtained as well at room temperature or (best) at a temperature between 5 and 10 °C (refrigerator temperature). Comparative experiments have indicated that only a 4-fold reagent excess is sufficient to acylate more than 95% gelatin free amino groups. Reaction times are then in the range of several hours.2,3 We suggest that a very simple way for thioacylation reaction on proteins is to dissolve the required amount of the selected reagent in a buffered protein solution and to store the mixture overnight at a temperature in the range 4-20 °C. Under such conditions the reagent hydrolysis is

406

Biomacromolecules, Vol. 1, No. 3, 2000

quite slow. Excess reagent is readily destroyed in a few seconds after adding an ammonia- or amine-containing buffer solution. Several enzyme modifications have already been studied and reported in part. More detailed results will be published later. Protein cross-linking using a bifunctional reagent is also in progress. Conclusion Dithioester and dithio acid groups are slowly hydrolyzed in buffered aqueous media, more rapidly at elevated temperature and high pH. However their lifetimes under such conditions are quite sufficient to allow their use as protein thioacylating agents. Kinetic data obtained in gelatin thioacylation at room temperature using dithioester or dithio acid excess are consistent with a second-order reaction rate with respect to amine concentration. The pH dependence of second-order reaction rate constants indicates that dithioesters react mainly with the free amine form whereas dithiocarboxylate ion reacts with both forms (amine and ammonium ion) probably through a more complex mechanism. Interestingly, thioacylation using dithio acid may be obtained in pH near neutrality, even in slightly acidic media, thus offering protein modification possibilities under these quite unusual conditions. Acknowledgment. This work was financially supported by ELF (Groupe de Recherche de Lacq). Dr Jean-Louis Se´ris is gratefully acknowledged for helpful discussions.

Levesque et al.

References and Notes (1) Levesque, G.; Souppe, J.; Se´ris, J. L.; Bellenger, V. US Patent 4,870,016 . (2) Souppe, J.; Urrutigoity, M.; Levesque, G. Biochim. Biophys. Acta 1988, 957, 254. Souppe, J.; Urrutigoity, M.; Levesque, G. New J. Chem. 1989, 13, 503. (3) Kjaer, A. Acta Chem. Scand. 1950, 4, 1347. Kjaer, A. Acta Chem. Scand. 1952, 6, 327. Kjaer, A. Acta Chem. Scand. 1952, 6, 1374. (4) A good method could be: S-alkylation of thioamide to an alkylimidoester salt followed by hydrolysis or sulfhydrolysis. (5) Previero, A.; Gourdol, A.; Derancourt, J.; Coletti-Previero, M. A. FEBS Lett. 1975, 51 (1), 68-72. (6) Scheithauer, S.; Mayer, R. Thio and Dithiocarboxylic acids and their derivatives. In Topics in Sulfur Chemistry; Senning, A. Ed.; G. Thieme Verlag: Stuttgart, Germany, 1979; Vol. 4. (7) Deleˆtre, M.; Levesque, G. Macromolecules 1990, 23, 4733. (8) Levesque, G.; Mahjoub, A.; Thuillier, A. Tetrahedron Lett. 1978, 3847. Bonnans-Plaisance, C.; Mahjoub, A.; Levesque, G. Tetrahedron Lett. 1980, 1941. (9) Bonnans-Plaisance, C.; Gressier, J. C.; Levesque, G.; Mahjoub, A. Bull. Soc. Chim. Fr. 1985, 891. (10) Leon, N. H. J. Pharm. Sci. 1976, 65, 146. (11) Habeeb, A. F. S. A. Anal. Biochem. 1966, 14, 328. Fields, R. 1971, 124, 581. Snyder, S. L.; Sobocinski, P. Z. Anal. Biochem. 1975, 64, 284. (12) Beiner, J. M.; Thuillier, A. C. R. Acad. Sci. 1972, C274, 642. (13) Levesque, G.; Arse`ne, P.; Fanneau-Bellenger, V.; Pham, T. N.; Se´ris, J. L. Biomacromolecules, 2000, 1, 387. (14) Bernard, M. A.; Borel, M.M.; Dupriez, G. ReV. Chim. Miner. 1975, 12, 181. (15) Gressier, J. C.; Levesque, G. Eur. Polym. J. 1980, 16, 1101. (16) Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. J. Biol. Chem. 1951, 193, 265.

BM000037B