Use of diphenylphosphorylazide for cross-linking collagen-based biornaterials

H. Petite,'.* V. Frei,'r2 A. HUC: and D. Herbage2* 'Coletica, 32 Rue Saint Jean de Dieu, 69007, Lyon, France; 21nstitut de Biologie et Chimie des Prottines (IBCP), UPR CNRS 412, 69367 Lyon Cedex 7, France

Cross-linking of collagen-based biomaterials increases their strength and persistence in vivo. Recently, we described an efficient cross-linking process via the formation of acyl azide groups on methylated carboxyl groups of collagen using hydrazine and nitrous acid (referred to here as the hydrazine method). In this report, we propose a simpler, faster way to prepare acyl azide groups and to cross-link collagen-based biomaterials, using diphenylphosphorylazide (DPPA) as a reagent. After determining the optimal conditions of cross-linking with DPPA, we compared the efficiency of this protocol with that using hydrazine and with the classical glutaraldehyde treatment. In order to validate and quantitate the extent of reaction, the degree of cross- linking was determined by the measure of the free primary amino group content of the samples.

Treatment of native bovine pericardium with 0.5% DPPA for 24 h led to efficient cross-linking, corre- sponding to a 50% decrease in the free primary

amino group content of the sample and raising its thermal stability from 62.8 up to 81.3"C. In comparison, the thermal stabilities of glutaraldehyde- or hydrazine-treated pericardium were 85 2 0.4"C and 83.4 ? 0.1"C. Similar decreases in free primary amino group content and increases in thermal stability were obtained for collagen films treated with DPPA, glu- taraldehyde, or hydrazine. These results were corrob- orated by resistance to bacterial collagenase digestion: DPPA-treated pericardium had a resistance to colla- genase digestion similar to that of glutaraldehyde- or hydrazine-treated pericardium. Residual DPPA content was measured by determining the phosphorus content: the concentration of phosphorus in tissue treated with 0.5% DPPA was not significantly different from that of untreated tissue. Treatment by DPPA thus appears to be an efficient, rapid method for cross-linking collagen- based biomaterials. 0 1994 John Wiley & Sons, Inc.

INTRODUCTION

Type I collagen is an attractive molecule for manufacturing biomaterials, owing to its biological properties.lr2 Collagen-based biomaterials are how- ever degraded by host enzymes upon implantation. The rate of biodegradation of collagen can be controlled effectively by cross-linking with glutaralde- hyde (GTA),3,4 but at the concentration required to effect adequate cross-linking GTA is cy to t~x ic .~ ,~

We proposed recently7rs a cross-linking method (re- ferred to below as the hydrazine method) which consists in transforming lateral carboxyl groups of col- lagen chains in acyl azide via formation of hydrazides, the acyl azide reacting with the amino groups of ad- jacent chains of collagen. This method led to efficient cross-linking, as demonstrated by an increase in ther- mal stability: the hydrazine method raises the thermal

*To whom correspondence should be addressed

stability of native bovine pericardium from 62.8" up to 83.4"C, in comparison with a thermal stability of 85.1 "C for GTA-treated tissue. Likewise, resistance to bacterial collagenase digestion was increased by hy- drazine treatment: fresh pericardium was entirely di- gested by collagenase within 5 days, but pericardium treated with GTA or hydrazine did not undergo diges- tion. Moreover, in pericardium samples9 and collagen sponges8 treated by the hydrazine method, the onset of calcification was delayed in comparison with GTA- treated materials when implanted subcutaneously in rats. Recent evaluation of collagen cross-linking tech- niques for the stabilization of human skin and porcine aorta showed that the acryl-azide method compared favorably with the other cross-linking agents.1°

The hydrazine method is however unwieldy to perform routinely, as it requires 5 days and extensive washing for complete removal of byproducts. An alternative is use of diphenylphosphorylazide (DPPA) to synthesize acryl azide on lateral carboxyl groups of

Journal of Biomedical Materials Research, Vol. 28, 159-165 (1994) 0 1994 John Wiley & Sons, Inc. CCC 0021-93041 941020159-07

160 PETITE ET AL.

collagen chains. DPPA has been used previously in peptide synthesis to convert carboxylic acids into acyl azide in one step." Coupling between amino acids proceeds directly, without isolation of the intermedi- ary acyl azide. Ozawa et a1.12 and Ikota et al.I3 used this method to synthesize the N-terminal sequence of secretin and intestinal motilin, respectively. Thus, DPPA can react with carboxylic groups of aspartic and glutamic acid residues on collagen lateral chains to form an acyl azide, which would then react with the lateral amino groups of collagen (chiefly lysine and hydroxylysine residues) (Fig. 1).

The aim of this study was to determine if DPPA could efficiently cross-link collagen and to define the optimal cross-linlung conditions (nature of solvent, DPPA concentration, triethylamine concentration, time). We also determined the concentrations of residual products after cross-linkage which included DPPA and dimethylformarnide (DMF).

The study was carried out with calf pericardium, which is readily available and is used frequently as a xenograft and with collagen films. The extent of cross-linking with DPPA, as compared to treatment with hydrazine or GTA, was assessed by measure of the free primary amino group content of the sample, by differential scanning calorimetry analysis of the thermal stability of collagen and by determination of its resistance to collagenase degradation.

MATERIALS AND METHODS

Materials

Boric acid, cyanogen bromide, disodium tetra- borate, glycine, hydrazine hydrate, hydrochloric acid, methanol, sodium chloride, sodium nitrite, Tris (hy- droxymethyl) aminomethane, and triethylamine were

1 OPh Coll- COO -P-OPh

&'*- (c)

Coll-CO -N3 t Coil-NH2 ( e ) I (d '

Coll-CO-NH -Col l ( f )

Figure 1. Proposed chemical pathway of collagen cross-linking by diphenylphosphorylazide (DPPA). Lat- eral carboxyl groups of aspartic and glutamic residues of collagen are transformed into acyl azide, which reacts with the lateral amino groups of collagen.

of analytical grade and were purchased from Prolabo (Paris, France). GTA was purchased from Merck (Nogent sur Marne, France). Bacterial collagenase (Clostridium histalyticum type 11), DPPA, and DMF were obtained from Sigma (La Verpilliere, France).

Pericardial sacs were obtained fresh from young calves, stripped of all visible fat and kept at 4°C in 70% ethanol until treatment. Films were prepared by air drying a bovine type I collagen solution, as described by Weadock et al.4

Determination of optimal cross-linking conditions

Standard procedure

Pericardial samples (10 mg dry weight) were im- mersed in 5 mL of a solution of DPPA in DMF at 4°C for 24 h (step l), rinsed in a borate buffer (0.04 M disodium tetraborate, 0.04 M boric acid, pH 8.9), and then incubated overnight at 4°C (step 2).

The thermal stability of the samples was deter- mined as described in Petite et al.7 The measurements were performed on a Setaram model 111 differen- tial scanning calorimeter (Lyon, France) and ther- mal changes were recorded at a constant rate of 2"C/min from 20- 100°C. Three typical temperatures were measured: the onset temperature, the tempera- ture at the peak maximum (referred to as the denatu- ration temperature, TD) and the recovery temperature. A minimum of three measurements were made. For simplification, only modifications from this standard procedure are noted.

Optimal solvent

The standard procedure was performed using ace- tone, DMF, or dichloromethane as the solvent for DPPA at a concentration of 1%.

Optimal DPPA concentration

The standard procedure was carried out using DPPA at concentration varying from 0.0125% -1.5%.

Kinetics of cross-linking

Step 1 of the standard procedure was performed for various periods of time (from 3.-48 h) with 1% DPPA. Step 2 was kept constant.

Influence of triethylamine

Experiment 1: Step 1 was performed with 1% DPPA in the presence of triethylamine at concentrations varying from 0.1 to 7%. Step 2 was kept constant. Experiment 2: Step 1 was performed with DPPA at various concentrations (0.1, 0.25, 0.5, 0.75, and 1%) in the presence of 0.5% triethylamine.

DIPHENYLPHOSPHORYLAZIDE CROSS-LINKING 161

Determination of residual products

Phosphorus content

Phosphorus content was determined by atomic ab- sorption spectrometry on untreated pericardium and on pericardium treated with 1, 0.5, 0.1, or 0% DPPA, rinsed three times in DMF, washed in a borate buffer and incubated overnight in the same buffer.

Dimethylformamide content

Samples were incubated in 2 mL of 14C DMF (276300 ? 750 dpm) containing 1% DPPA for 24 h at 4"C, rinsed three times in borate buffer (rinses R1, Rz, and R3) and incubated overnight in the same buffer (k). The tissue was then rinsed three times in 70% ethanol (Al, A2 and A3), and the radioactivity of R1, Rz, R3, Ar, Az, and A3 was measured. Radiolabel in the tissue was measured after step 1, and after rinses & and A3.

Comparison of properties of collagen films and pericardia cross-linked by GTA, DPPA or hydrazine methods

DPPA treatment

Collagen films and pericardia were cross-linked by the standard procedure, using DPPA at a concentra- tion of 0.50%.

Hydrazine and GTA treatment

Pericardium and collagen films were cross-linked by hydrazine or GTA treatment, as described by Petite et al.7 Briefly, samples were incubated in methanol containing 0.2 M hydrochloric acid for 1 week at 20"C, washed, placed in 1% aqueous hydralazine, 1 M sodium chloride overnight at 20°C, washed, treated with 0.5 M sodium nitrate, 0.3 M hydrochloric acid, and 1 M sodium chloride for 3 min at 4"C, washed, and immersed in a borate buffer (0.04 M disodium tetraborate, 0.04 M boric acid, 1 M sodium chloride) overnight at 4°C. GTA treatment was performed with a fresh solution (0.6%) for 4 days at 4°C.

Free amino group determination

The method used 2,4,6,-trinitrobenzenesulfonic acid (TNBS) which reacts specifically with free primary amino groups to give trinitrophenyl (TNP)-derivatives which can be determined spectrophotometrically.'4 The procedure was adapted to insoluble materials by Wand et al.I5 and Edwards-Levy et a1.I6 The free amino group content was expressed as pmoles per gram total amino acid. Amino acid content of the sam- ples was determined by amino acid analysis (Beckman model 6300 analyzer).

Collagenase assay

The collagenase assay was performed on fresh, GTA- or hydrazine-treated pericardium, as described previ~usly.~ Samples were incubated for 24 h in a bacterial collagenase solution (3300 U/mL), then cen- trifuged. The collagen content of the supernatant and pellet was determined. Results are expressed as a percentage of residual collagen.

RESULTS

Determination of optimal cross-linking conditions

Choice of solvent

Cross-linking of pericardium was carried out with 1% DPPA in acetone, dichloromethane, or DMF. The results are shown in Table I. The thermal stability of pericardium was increased by 15°C when the cross- linking process was carried out in DMF and by 5°C with acetone. No increase in thermal stability was observed with dichloromethane.

Optimal DPPA concentration

Between 0.0125% and about 0.5% DPPA, the ther- mal stability of pericardium increased with concentra- tion, but it remained stable at higher concentrations up to 1.5% (Fig. 2).

TABLE I Thermal Stability of Fresh Pericardium and Pericardium Treated by DPPA (1% for 24 h) in Different Solvents

Thermal stability ("C)

Pericardium To TD TR

Fresh DPPA in acetone DPPA in dimethylformamide DPPA in dichloromethane

63.6 ? 1.2 66.3 2 0.6 79.1 2 0.7 62.6 2 0.1

67.3 2 0.9 72.4 2 0.3 81.9 2 0.2 67.4 2 0.4

83.1 2 5.0 84.6 f 2.0 92.1 2 2.1 78.6 t 1.7

162 PETITE ET AL.

? c

L m 0 P

c 6

1 60 1 2

Concentration of diphenylphosphorylazide ( % )

Figure 2. Thermal stability of pericardium treated with increasing concentrations of diphenylphosphorylazide (DPPA) for 24 h.

Kinetics study

The results of the kinetics study are shown in Figure 3 for a DPPA concentration of 1% in DMF. One hour of treatment induced an increase in thermal stability of 10°C (TD = 77.6 -C 2°C); however, maximal thermal stability was obtained for 24 h of treatment (TD = 81.3 2 1°C).

Influence of triethylamine

As Ikota et all3 used triethylamine to increase pep- tide synthesis, we studied its effects in our system. To determine the optimal concentration of triethylamine, pericardial samples were cross-linked for 24 h with 1% DPPA in the presence of different concentrations of triethylamine. The thermal stability of the samples was then determined. As shown in Figure 4, cross- linking by DPPA in the presence of 0.1, 0.5, and 1% triethylamine resulted in an increase in thermal stability of 2°C (TD = 83.6 2 1°C) compared with a control without triethylamine (TD = 81.5 ? 1OC). In the presence of 4% or 7% triethylamine, thermal sta- bility was decreased by 4°C and 5"C, respectively, demonstrating an inhibitory effect of triethylamine at high concentrations.

Experiment 2 was designed to study the effect of 0.5% triethylamine on the thermal stability of pericardium cross-linked by DPPA at various con- centrations ( O . l - l % ) . The results are shown in Figure 5. Triethylamine increased thermal stability (by 2°C) only with 1% DPPA.

Determination of residual products

The phosphorus contents of samples treated with 0, 0.1, 0.5, and 1% DPPA are shown in Table 11. The level

Y E v

3

m a,

c

L

n : c

K 0 .- c

E 2 I ~

m t al n

0 1 3 4 6 2 4 4 8

Duration of reaction ( h )

Figure 3. Kinetics of cross-linking of pericardium by diphenylphosphorylazide (1 % in dimethylformamide). Maximal thermal stability was obtained after 24 h of treatment.

in pericardial samples treated with 0.1 or 0.5% DPPA was not significantly different from that of untreated or DMF-treated samples, as assessed by Student's t test; however, the phosphorus content of 1 % DPPA- treated pericardium was significantly higher than that of control samples ( P I .05).

To determine residual DMF Concentration, pericar- dial samples were treated for 24 h with 1% DPPA in labelled DMF. The radiolabel concentration was only 1/200 of its initial value after four rinses in borate buffer, and was negligible after three rinses in 70% ethanol (Table 111).

Comparison of properties of pericardiurn and collagen films cross-linked by DPPA under optimal conditions and by the GTA and hydrazine methods

The optimal conditions for the DPPA method of cross-linking were therefore concluded to be the fol- lowing: treatment for 24 h with 0.5% DPPA in DMF at 4°C. As shown in Table IV, treatment of pericardium or collagen film with GTA (0.6'%), hydrazine, or DPPA resulted in a large increase in thermal stability over that of fresh pericardium or control collagen film. As expected from the proposed chemical pathway of collagen cross-linking by DPPA (Fig. 1) and GTA, the decrease in the free amino group content of the cross-linked samples was a good marker of the extent of these reactions and was related to the increase in thermal stability of the samples (Table IV). As already described by Kakade and Liner,I7 it is worth noting that only about 60% of the lysine &-amino group, present in the control pericardium or collagen films, reacted with TNBS. GTA was the most efficient, react- ing with about 80% of the available amino groups and

DIPHENYLPHOSPHORYLAZIDE CROSS-LINKING 163

L F n

c

aJ

E, c

S 0 ._ c

? c m C aJ 0

70

Concent ra t ion of tr iethylamine ( O h )

Figure 4. Effects of different concentrations of triethyl- amine on the thermal stability of pericardium treated with diphenylphosphorylazide (1 %, 24 h). Thermal sta- bility was enhanced (+2"C) with concentrations of tri- ethylamine of 0.1-1%.

inducing an increase in thermal transition temperature of about 22°C in both pericardium and collagen films; the DPPA method led to increases in thermal tran- sitions of about 20°C, corresponding to the reaction of 50-60% of the available amino groups (Table IV). When pericardia were digested with collagenase, fresh samples were entirely digested within 24 h, but none of the treated pericardia underwent any significant degradation (Table IV).

DISCUSSION

In the first part of the present study, we deter- mined the optimal conditions for cross-linking calf pericardium by DPPA. The efficiency of the reaction was determined by the measure of the decrease in the free amino group content and of the increase in the collagen thermal stability. The effects of the solvent, of DPPA, and triethylamine concentrations and of the kinetics of cross-linking were studied. We found that cross-linking by DPPA is best carried out with DMF. This result is consistent with the data of Yokoyama et al,ll who demonstrated a greater yield with DMF (85%) than with dichloromethane (52%) for synthesis of thiol esters from carboxyl groups and thiols. Maximal increases in thermal stability were obtained with DPPA concentrations up to 0.5%. Further increases in concentration did not increase thermal stability.

Reaction with DPPA is rapid: thermal stability is increased by 70% within 1 h, and the reaction is complete by 24 h. These kinetics of tissue stabilization are similar to those of cross-linking with high concen-

0 1 0 25 0 5 0 7 5 1

Concentration of diphenylphosphorylazide ( % )

Figure 5. Effect on pericardial cross-linking of different concentrations of diphenyphosphorylazide in dimethyl- formamide in the absence (black bars) and presence (hatched bars) of 0.5% triethylamine.

trations of GTA. Woodroof5 found that when aortic valves were cross-linked with 0.1% GTA, 75% of the reaction was completed within 1 h.

The thermal stability of pericardium was increased by roughly 2°C when 1% DPPA was used in the pres- ence of 0.1 -1 % triethylamine; higher concentrations of triethylamine inhibited cross-linking. These results are consistent with those of Ikota et al.,I3 who used a ratio of DPPA: triethylamine of 0.52-0.56 to increase the yield of peptide synthesis. We found, however, that the increase in thermal stability obtained by addition of triethylamine was not large enough to justify use of this potentially toxic reagent.

In the second part of this study, we determined the residual content of DPPA and DMF after treat- ment. The DPPA content was estimated by measur- ing the phosphorus content of our samples before and after treatment with DPPA. The concentration of phosphorus in tissue treated with 0.5% DPPA and rinsed twice with DMF was not different from that of untreated tissue. A low residual DMF concentra- tion of 0.2 nmol/mg of collagen was obtained after four washings in borate buffer and 70% ethanol. This concentration, considered to be nontoxic,I8 could be decreased further by additional washings.

We therefore defined the optimal conditions of reac- tion with DPPA as follows: treatment for 24 h at 4°C with 0.5% DPPA in DMF, followed by four rinses in borate buffer and three rinses in 70% ethanol.

In the third part of our study, we compared the cross-linking efficiencies of DPPA, hydrazine, and GTA treatments by measuring free amino group content, thermal stability and resistance to bacterial collagenase digestion of pericardium and collagen films. DPPA treatment was found to be an efficient cross-linking agent, reacting with 50-60% of the available amino group and resulting in important

164 PETITE ET AL.

TABLE I1 Phosphorus Content of Pericardium Before and After Treatment by DPPA (24 h in DMF)

DPPA-'Treated Pericardium

(DPPA %) Non-Treated Pericardium DMF-Treated Pericardium 0.1% 0.5% 1%

Phosphorus content (pprn) 963 L 21

~

794 2 21 963 2 21 1034 2 50 1480 5 70

Phosphorus content was measured after washing in borate buffer and overnight incubation in the same buffer.

TABLE I11 Residual DMF in Pericardium Treated by DPPA (1% for 24 h) After four Washes

in Borate Buffer (R1*Rq) followed by three Rinses in 70% Ethanol (Al*A3) ~ ~~

Sequential washes ~ ~~

DMF concentration Initial R1 R2 R3 R4 A1 A2 A3

Radioactivity 2763002 7500 (in dpm) In washes 54500 10300 2900 1400 400 350 80 In tissue 2630 f 700 300 2 160 14 C 7

TABLE IV Thermal Stability, Bacterial Collagenase Susceptibility, and Free Amino Group Content

of Pericardium and Collagen Film with or without Treatment by Glutaraldehyde (0.6% for 4 Days) and by the Acyl Azide Methods (Hydrazine and DPPA)

Thermal Stability ("C)

Sample To T D TR Resistance to Collagenas@* Free NHz-groupst

P e ri c a r d i u m Fresh GTA-treated Hydrazine-trea ted DPPA-treated

Control GTA-treated Hydrazine-treated DPPA-treated

Film

58.6 2 0.3 62.8 2 0.2 73.6 2 1.4 81.8 It 0.6 85.1 2 0.4 93.5 t 0.7 79.2 i 0.2 83.4 C 0.1 90.9 i 1.5 78.5 i 0.3 81.4 ? 0.3 91.7 ? 1.0

44.4 2 2.9 52.0 2 1.7 61.3 2 1.2 69.6 5 0.5 74.6 2 0.5 84.6 5 0.1 65.0 C 5.0 69.9 i 0.6 77.6 t 3.0 70.3 i 0.3 72.6 ? 1.0 79.2 2 0.5

3 2 1 99 i 1 99 i 1 96 i 2

-

- -

-

232 35

120

208 49

78

-

-

*Percent total collagen nondigested. tExpressed as pmole per gram total amino acid.

increases in thermal stability (+20"C) and resistance to collagenase digestion. No significant difference was observed between the DPPA and hydrazine treatments, GTA being slightly more efficient as regards the increase in thermal stability.

Preliminary results19 obtained using a model cul- ture of chick embryo aorta demonstrate the excellent cytocompatibility of DPPA cross-linked pericardium. Thus, DPPA treatment of collagen-based biomaterials appears to be an interesting alternative to hydrazine or GTA treatments. This treatment should lead to enhanced cellular infiltration and subsequent tissue formation in vizm, as demonstrated by the preliminary results of our current experiments.

Financial support from the Groupement de Reflexion sur la Recherche Cardiovasculaire Franqais (GRRC) and the CNAMTS (grant 3-87) is gratefully acknowledged. The authors thank Dr. Eloy of Unit6 37 INSERM (Lyon) and Dr. S. Borron (Cleveland) for valuable discussions.

References 1. E. E. Sableman, "Biology, biotechnology, biocom-

patibility of collagen," in Biocompafibility of Tissue Analogs, Vol. 1, D.F. Williams (Ed.), CRC Press, Boca Raton, pp. 27-66 (1985). A. Huc, "Collagen biomaterials characteristics and applications," 1. Am. Leather Chem. Assoc., 80,

2.

195-212 (1985).

DIPHENYLPHOSPHORYLAZIDE CROSS-LINKING 165

3.

4.

5.

6.

7.

8.

9.

10.

11.

A. Woodroof, “Use of glutaraldehyde and formalde- hyde to process tissue heart valves,” 1. Bioeng., 2,

K. Weadock, R. Olson, and F. Silver, “Evaluation of collagen crosslinking techniques,” Biomat. Med. Arty. Org., 11, 293-318 (1984). D. Speer, M. Chvapil, C. Eskelton, and J. U1- reich, “Biological effects of residual glutaraldehyde in glutaraldehyde-tanned collagen biomaterials,” J . Biomed. Mater. Res., 14, 753-764 (1980). D. Niebe, J. Mergeman, G. Litalien, and W. Abbott, ”Glutaraldehyde release from vascular prostheses of biological origin,” Surgery, 104, 26-33 (1988). H. Petite, I. Rault, A. Huc, P. Menasche, and D. Herbage, “Use of the acyl azide method for cross- linking collagen-rich tissues such as pericardium,” J. Biomed. Mater. Res., 24, 179-187 (1990). K. Anselme, H. Petite, and D. Herbage, “Inhibition of calcification in vivo by acyl azide cross-linking of a collagen-glycosaminoglycan sponge,” Matrix, 12,

H. Petite, S. Rinaldi, P. Menasche, A. Piwnica, and A. Huc, “Low level of calcification and mechanical properties of the acyl azide-treated pericardium compared with glutaraldehyde-treated pericardium.” Proceeding of the International Sym- posium on Surgery for Heart Valve Disease, London, pp. 188-192, 1989. D.M. Simmons and J.N. Kearney, ”Evaluation of collagen cross-linking techniques for the stabiliza- tion of tissue matrices,” Biotechnol. A p p l . Biochem.,

Y. Yokoyama, T. Shiori, and S. Yamada, “Phospho- rus in organic synthesis XVI. Diphenylphosphory- lazide and diethylphosphocyanidate. Two reagents for the preparation of thiol esters from carboxylic

1-9 (1978).

264-273 (1992).

17, 23-29 (1993).

12.

13.

14.

15.

16.

17.

18.

19.

acids and thiols,“ Chem. Pharm. Bull., 25,2423-2429 (1977). K. Ozawa, T. Shiori, and S. Yamada, ”Application of DPPA to the synthesis of the protected N-terminal hexapeptide of secretin, “Chem. Pharm. Bull., 25,

N. Ikota, T. Shiori, S. Yamada, and S. Tachibana, ”Synthesis of porcine rotilin by the solid phase method using DPPA and diethylphosphoro- cyanidate,“ Chem. Phavm. Bull., 28,3347-3356 (1980). K. Satake, T. Okuyama, M. Ohashi, and T. Shinoda, ”The spectrophotometric determination of amine, amino acid and peptide with 2,4,6,-trinitrobenzene 1-sulfonic acid,” J. Biochem., 47, 654-660 (1960). H. Wand, M. Rudel, and H. Dautzenberg, “Die bes- timmung des aminogruppengehaltes unsloslicher tragermaterialen mittels 2,4,6-trinitrobenzolsulfon- saure,” Z. Chem., 18, 224 (1978). F. Edwards-Levy, M.-C. Andry, and M.-C. Levy, “Determination of free amino group content of serum albumin microcapsules using trinitroben- zenesulfonic acid: effect of variations in polycon- densation pH,” Int. J. Pkarm., 96, 85-90 (1993). M.L. Kakade and I.E. Liener, ”Determination of available lysine in proteins,” Anal. Biochem., 27, 272-280 (1969). G. Kennedy, ”Biological effects of acetarnine, formamide and their monomethyl and dimethyl derivates,” Crit. Rev. Toxicol., 17, 129-183 (1987). H. Petite, “Etude d’une nouvelle methode de reticulation (dite aux acyl azides) de biomateriaux A base de collagene,” Thesis, Lyon, n”23-91, 1991.

122-125 (1977).

Received April 20, 1992 Accepted August 19, 1993