Chemical modification of silk fibroin with cyanuric chloride-activated

Aug 6, 1993 - nucleophilic e-amino group of the lysine residue and the nucleophilic imidazole group of the histidine residue in SF reacted with actPEG...
1 downloads 0 Views 722KB Size
Bioconlugete Chem. 1993, 4, 554-559

554

Chemical Modification of Silk Fibroin with Cyanuric Chloride-Activated Poly(ethy1ene glycol): Analyses of Reaction Site by 'H-NMR Spectroscopy and Conformation of the Conjugates Yohko Gotoh,' Masuhiro Tsukada, and Norihiko Minourat National Institute of Sericultural and Entomological Science, 1-2 Ohwashi, Tsukuba, Ibaraki 305, Japan, and National Institute of Materials and Chemical Research, 1-1 Higashi, Tsukuba, Ibaraki 305, Japan. Received August 6,1993' Solubilizedsilk fibroin (SF)in 0.1 M borate buffer (pH 9.4)was modified with 2-0-[methoxy(polyethylene glycol)]-4,6-dichloro-s-triazine(actPEG1) at 4 "C. The weight of the modified SF (PEG1-SF) was at least 3.2 times that of the starting material SF. Amino acid analysis of PEG1-SF suggested that the nucleophilic €-aminogroup of the lysine residue and the nucleophilic imidazole group of the histidine residue in SF reacted with actPEG1. The lH-NMR spectrum of PEG1-SF showed a downfield shift of the aromatic protons of the tyrosine residue from the corresponding protons of SF. The 'H-NMR spectrum of the SF reacted with cyanuric fluoride (CyF), whose fluorine atoms are known to react with the phenolic hydroxyl group of the tyrosine residue, also showed the downfield shift. These results suggested that the reaction site of SF with actPEGl was the phenolic hydroxyl group of the tyrosine residue in addition to the lysine and histidine residues. The conformation of PEG1-SF in a solid state was examined by means of IR and X-ray measurement. The IR spectrum of PEG1-SF revealed a change in secondary structure from random coil to @-sheetdue to the coexistence of PEG molecules. The X-ray diffraction pattern of PEG1-SF indicated that the PEG molecules covalently bonding to SF narrowed the spacing of the interchain periodicity and promoted the formation of the interchain @-sheet.

INTRODUCTION

The conjugation of synthetic or natural macromolecules to proteins has been widely developed since the 1970s (110). The purpose of the conjugation is to stabilize the protein and to add new properties to the protein. Poly(ethylene glycol) (PEG)' is a nonimmunogenic and amphipathic polymer. The consequences of protein modification with PEG are the reductions of the immunoreactivity and immunogenicity of the antigenic protein (3-6). The protein modification with PEG makes the enzyme soluble and active in organic solvents (3, 7). Therefore, many kinds of activated PEG, whose hydroxyl end groups are converted into functional groups to attach covalently to proteins, have been developed as chemical modifiers (3-9). Cyanuric chloride is one of the most effective and widely applicable coupling agents to attach synthetic or natural polymers such as polysaccharides to proteins (10). Inada et al. developed cyanuric chlorideactivated poly(ethy1eneglycol),for example 2-0- [methoxy(polyethylene glycol)l-4,6-dichloro-s-triazine(actPEG1) (3, 4 , B ) and 2,4-bis[O-[methoxypoly(ethyleneglycol)ll6-chloro-s-triazine (actPEG2) (3, 7) (Scheme I). Nevertheless, the toxicity and excessivereactivity of the cyanuric

* To whom correspondence should be addressed. + National Institute of Materials and Chemical Research.

Abstract published in Aduance ACS Abstracts, October 15, 1993. 1 Abbreviations used: SF, silk fibroin; actPEG1,2-0-[methPEG1-SF, the oxypoly(ethy1eneglycol)]-4,6-dichloro-s-triaine; silk fibroin modified with actPEG1; 1H-NMR, proton nuclear magnetic resonance; CyF, cyanuric fluoride; IR, infrared absorption; PEG, poly(ethy1ene glycol); actPEG2,2,4bis[O-[methoxypoly(ethy1eneglycol)]-6-chloro-s-triazine;PEG-OMe, methoxypoly(ethy1eneglycol); Tris, tris(hydroxymethy1)aminomethane; CyF-SF, the silk fibroin reacted with CyF; DSS, 3-(trimethylsily1)propanesulfonic acid sodium salt.

chloride have been pointed out recently (9). Although the biochemical properties of the PEGprotein conjugate have been studied extensively, the reaction site and structural analysis of the conjugate with the aid of spectroscopy have not been reported. In this paper, the PEGsilk fibroin (SF) conjugate (PEG1-SF) was prepared by reacting actPEGl with the solubilized SF in order to obtain a new functional biopolymer. We chose SF as the protein of the PEG conjugate, because SF has recently been applied to biomaterials such as an enzyme-immobilization material (11)andacellculturesubstrate (12). SF (molecularweight of 370 kDa estimated by Shimura et al. (13))is a fibrous protein consisting mainly of the sequence Gly-Ala-GlyAla-Gly-X, where X is an alanine or serine residue (14). Its partial primary structure is currently determined by sequencing the cDNA of the Bombyx mori silkworm (15). Its conformational characteristics have been studied widely (11, 14, 16).

The purpose of this paper is to modify SF with cyanuric chloride-activated PEG and to clarify the reaction site and secondary structure of the modified SF. The reaction site of SF with actPEGl was clarified by 'H-NMR spectroscopy. The secondary structure of the modified SF was revealed by IR measurement and X-ray diffractometric measurement. EXPERIMENTAL PROCEDURES

Materials. 2-0- [Methoxy(polyethylene glycol)l-4,6dichloro-s-triazine (actPEG1, MW 5000) and methoxy-

1043-1802/93/290~0554$04.00/0 0 1993 American Chemical Society

Chemical Modification of Silk Fibroin

-

o.l

r

fr16-24 ( P E G I - S T

Tube number Figure 1. Gel filtration profiles of PEG1-SF (a), SF (b), and actPEGl (c) using Sephacryl S-300column (2.6X 70 cm, each fraction = 6.0 mL). UV absorbance at 280 nm was used for the detection of each sample.

poly(ethy1ene glycol) (PEG-OMe, MW 5000) were purchased from Sigma Chemical Co. (St. Louis, MO). Sephacryl S-300 was purchased from Pharmacia Fine Chemicals (Sweden). Cyanuric fluoride (CyF) was obtained from Wako Pure Chemicals Industries Co. Ltd. (Osaka, Japan). All other chemicals were reagent-grade products obtained commercially. An ultrafiltration cell with an XM-50 membrane was purchased from the Amicon Division of W. R. Grace Co. (Danvers, MA). Preparation of the PEGl-SF Conjugate. The shells of cocoons of Bombyx mori were cut into pieces, and 15 g of cut cocoons was degummed with aqueous boiling 0.5 5% (w/v) Na2C03 solution for 30 min and rinsed with hot water. After drying in air, 11g of degummed SF fiber was obtained. SF solution was prepared by dissolving 1.25 g of the degummed fiber in 25 mL of 9 M LiBr aqueous solution at 60 "C over 30 min (11,17, 18).2The dissolved solution was dialyzed against water using a cellulose tube for a few days, and then 61 mL of 1.9% (w/v) silk fibroin solution was obtained. To ionize the residues of SF under basic conditions and make the residues reactive, a solution of 5 mL of 0.12 M sodium borate buffer (pH 9.4) (4,8) was added to 1 mL of 1.9% (w/v) SF aqueous solution. Then 6 mL of the 0.32% (w/v) SF-O.l M sodium borate aqueous solution (pH 9.4) containing 19 mg of starting material SF was prepared. To this solution was added 100 mg of actPEGl at 4 "C over 15 min. After the reaction mixture was kept at 4 "C for 1h, the solution was dialyzed against cold 0.1 M phosphate buffer (pH 7.0) for several hours and subsequently 5 M urea-0.02 M TrisoHC1 buffer (pH 8.0) overnight for gel filtration. In order to remove the unreacted modifier (actPEGl), the dialyzed solution was applied to a column of Sephacryl S-300(2.6 X 70 cm) preequilibrated with 5 M urea-0.02 M Tris.HC1 (pH 8.0) and eluted with the same buffer (13) at a flow rate of 0.3 mL/min. Each fraction of 6.0 mL of the eluate was collected by a fraction collector. Its elution pattern obtained by measuring the absorbance at 280 nm is shown in Figure 1. The PEG1-SF conjugate was eluted first. The eluate (fractions 16-24) was collected and dialyzed against water overnight. Finally, the solution of PEG1-SF was concentrated by ultrafiltration using an XM-50 membrane. After the evaporation of the concentrated solution, 61 mg of PEG1-SF product was obtained. Reaction of Silk Fibroin with CyF. To 2 mL of 2 % (w/v) SF aqueous solution was added 6 mL of 1.0 M NaHC03aqueous solution (pH 9.1). A mixture of 0.1 mL of CyF and 1mL of 1,Cdioxane was added to this 8-mL 2 The dissolution of SF in a concentrated solution of a neutral salt induced depolymerizationof SF and led to the heterogeneous molecular weight of SF from 50 to 200 kDa (18).

Bloconlugate Chem., Voi. 4,

No. 6, lQQ3 555

solution of 0.5% (w/v) SF containing 0.75 M NaHC03 (pH 9.1) at 4 "C (19,20).The reaction mixture was allowed to stand at 4 "C for 1.5h. The solution was dialyzed against cold water overnight and concentrated by ultrafiltration. The concentrated solution gave 35 mg of the SF reacted with CyF (CyF-SF). Amino Acid Analysis. Each sample (12 mg) of control SF, PEG1-SF, and CyF-SF was hydrolyzed under vacuum in 2 mL of 6 N HCl at 110 "C for 20 h. The hydrolyzed samples were dried in a rotary evaporator at 40 "C and dissolved in 10mL of 0.02 N HC1. These sample solutions were filtered through a membrane filter to remove a small amount of the residues and were applied to an amino acid analyzer. Amino acid analysis was carried out using a Hitachi L-8500 type rapid amino acid analyzer. 1H-NMR Measurement. Proton nuclear magnetic resonance ('H-NMR) spectra were observed at 25 "C using a JEOL EX-90 NMR spectrometer operating at 90 MHz. The WEFT technique utilizing the difference in relaxation times of SF and HDO was applied as follows. The pulse angles used were 90"(23 ps) and 180" (46 ps). The pulse interval between 180" and 90" was set as a variable parameter for auto stacking more than 1 s in order to weaken the peak of HDO. The IH-NMR chemical shift is represented in parts per million downfield from internal 3-(trimethylsilyl)propanesulfonicacid sodium salt (DSS). The films of SF and CyF-SF prepared according to the procedure described in the IR measurement section were soluble in water. Thus, NMR samples of CyF-SF and a mixture consisting of 71 wt % SF and 29 w t % PEG-OMe were prepared by dissolving these films in D2O. Because the PEG1-SF film was insoluble in water, the aqueous solution of PEG1-SF was exchanged with D20 by ultrafiltration, and the spectrum of PEG1-SF was measured in about 90 % D20-10 % H2O. The concentrations of these samples were about 2 % (w/v). IR Measurement. The films of PEG1-SF and SF were obtained by casting each solution (0.5-1% (w/v)) onto thin polyethylene films as a casting substrate and drying at room temperature overnight (11,17). The blended film (the weight ratio of actPEGl to SF was 2.2) was prepared by casting a mixture of aqueous SF solution and aqueous actPEGl solution. The infrared absorption (IR) spectra of these films were recorded using a Perkin-Elmer 1760X FT-IR spectrometer. X-Ray Diffractometric Measurement. X-ray diffractometric measurement was performed with a diffractometer (Rigaku Denki Co., Ltd. Ru-200) with a scintillation counter using Cu Ka (A = 1.54 A) radiation. The voltage and current of the X-ray source were 25 kV and 10mA, respectively. The diffraction intensity curves were measured with a scanningrate of 1deg/min, a time constant of 1 s, and a scanning region of 5-35", The films of SF and PEG1-SF prepared by the method described above were measured. The control specimen of SF having a P-sheet structure was prepared as follows. SF film was immersed in 50% (v/v) methanol-water mixture for 4 h at room temperature and dried in air overnight (11,17). RESULTS AND DISCUSSION

Synthesis of PEGl-SF Conjugate. Figure 1 shows that PEG1-SF eluted more rapidly than unmodified SF. This result suggests an increase in the molecular weight of PEG1-SF as a result of the modification. From 19 mg of starting material SF was prepared 61 mg of product PEG1-SF. Assuming that 100% of the starting material SF was recovered, the weight of the product PEG1-SF was 3.2 times that of the starting material SF. This means

558

Bioconjugate Chem., Vol. 4, No. 6, 1993

Table I. CyF-SF, amino acid ASP Thr Ser Glu GlY Ala Val Met Ile Leu cys Tyr

Phe Lys His

Results of Amino Acid Analyses of PEG1-SF, and SF. PEG1-SF CyF-SF SF PEG1CyF(mol % ) (mol %) (mol 7%) SF/SF SF/SF 1.45 0.88 11.1 0.93 45.5 29.6 2.29 0.13 0.63 0.46

1.51 0.89 10.9 1.18 45.6 29.3 2.17 0.11 0.60 0.45

1.43 0.85 10.7 0.96 45.6 29.8 2.37 0.11 0.62 0.55

1.0 1.0 1.0 0.97 1.0 0.99 0.97 1.2 1.1 0.84

4.99 0.72 0.22 0.11 0.41 0.55

5.21 0.67 0.24 0.15 0.45 0.53

5.00 0.70 0.32 0.18 0.44 0.48

1.0 1.0 0.69 0.61 0.93 1.1

1.1

1.0 1.0 1.2 1.0 0.98 0.92 1.0 0.97 0.82

11.

1.0 0.96 0.75 0.83 1.0

Pro 1.1 'The errors inherent in the amino acid analyses of major components were within 5 % ,but the errors inherent in the analyses of minor components were about 20% because of their very small contents.

that PEG1-SF is composed of PEG and SF in the weight ratio of 2.2:l and contains 32% SF. If the average molecular weights of the amino acid unit constituting SF and actPEGl are supposed to be 753and 5000,respectively, the mole ratio of amino acid residue to PEG is calculated to be 30 (=(1/75)/(2.2/5000)). This result implies that 3.3 mol 9% of amino acid residues in SF were modified with actPEG1. However,this value is only a minimum estimate based on the assumption of 100% recovery of SF. Thus, more than 3.3 mol % of the residues were modified. Amino Acid Analysis. The results of the amino acid analyses of the SF before and after the modification are summarized in Table I. On the basis of these results, we noticed that the contents of the lysine and histidine residues in PEG1-SF were considerably decreased by the modification. King et al. reported that the amino acid analysis of the allergen modified with cyanuric chlorideactivated PEG also showed a considerable decrease in the lysine residue, and the bond joining the e-amino group of the lysine residue with the triazine nucleus was only partially stable under hydrolysis conditions (6). Thus, our result means that nucleophilic groups such as the e-aminogroup of the lysine residue and the imidazolegroup of the histidine residue in SF reacted with actPEGl but that some linkages between these groups and actPEGl had a possibility of being broken in the hydrolysis step. A slight decrease in the content of the arginine residue in PEG1-SF is considered to be due to some errors in the analysis. The methionine and leucine residues are considered not to react with actPEG1, because these residues have no reactive groups. Since more than 3.3 mol % of the residues in SF were modified, it is estimated that not only the lysine (0.32 mol %) and histidine (0.18 mol % ) residues but also other residues reacted with actPEG1. 'H-NMR Spectrum of PEGl-SF. In order to analyze the reaction sites, the 'H-NMR spectrum of PEG1-SF The average molecular weight (W) of the amino acid unit constituting SF was calculated to be 75 by considering the amino acid composition of SF (Table I) (17,18).Namely, the average molecular weight was calculated using the following equation:

w = zwici where wi = molecular weight of each amino acid residue and ci = mole content of each amino acid residue (Table I, column 3).

1.970

p'vl

b

I

I

I

I

I

10

9

8

7

6

'

I

I

I

I

I

I

5

4

3

2

1

0

PPm

Figure 2. 1H-NMR spectra of SF in DsO (a), the mixture of SF and PEG-OMe in DzO (b), and PEG1-SF in about 10% HzO90% DzO (c).

was measured (Figure 2). Considering the amino acid composition of SF (Table I), the peaks of SF were assigned as follows (21). The y C H , and &CH resonances of the valine residue were observed at 0.88 ppm (doublet, J = 7 Hz) and 2.1 ppm (broad singlet), respectively. The ,B-CHs resonance of the alanine residue appeared at 1.39 ppm (doublet, J = 7 Hz). The broad resonance at 3.0 ppm was assigned to the &CH2 of the tyrosine, asparagine, and aspartic acid residues. The a-CH2resonance of the glycine residue and the &CH2 resonance of the serine residue appeared at 3.94 ppm as an overlapped signal. The resonances a t 4.29-4.45 ppm were assigned to the a-CH of the serine, alanine, and other residues. The shoulder at about 4.6 ppm was assigned to the a-CH of the tyrosine residue. The aromatic protons of the tyrosine residue appeared as doublet peaks (J = 7 Hz) at 6.79 and 7.09 ppm. The resonance at about 7.3 ppm was assigned to the aromatic protons of the phenylalanine residue. The spectrum of PEG1-SF showed a singlet peak a t 3.70 ppm. This peak was assigned to the methylene protons of the actPEGl bonded to SF (8) because the spectrum of the mixture composed of SF and PEG-OMe showed the same spectrum as SF except for a singlet peak at 3.70 ppm assigned to the methylene protons of PEGOMe. In the spectrum of PEG1-SF, the peaks at 6.79 and 7.09 ppm of the tyrosine residue disappeared, and a broad singlet peak at 7.25 ppm appeared. This downfield shift reveals the change in the molecular environment of the tyrosine residue caused by the modification, that is, the shift is derived from the shielding effect of the triazine ring on the tyrosine residue (22). Thus, the tyrosine residue of SF is thought to react with actPEG1. In order to prove that the downfield shift of the tyrosine residue is the result of chemical bonding of the triazine ring in actPEGl to the tyrosine residue, SF was reacted

Bioconjugate Chem., Vol. 4, No. 6, 1993 557

Chemical Modificatlon of Silk Fibroin

Scheme I11

I1

il

iEEll I All" I

11 I1

II

4"C, I h

+

borate buffer (PH9.4) actPEG1 silk fibroin O-(CH,CH2O)..CH,

\

1

1

1 0 9

1

1

1

8

7

6

1

1

1

1

1

1

5

4

3

2

1

0

PPm

Figure 3. 1H-NMR spectra of SF (a) and CyF-SF (b) in DzO. the silk fibroin modified with actPEG1

Scheme I1 OH

N - J

-HNTYr

CH--E

-

CYF

with CyF. CyF has the same triazine ring as actlPEG1, and the fluorine atoms in CyF are known to react with the phenolic hydroxyl group of the tyrosine residue of proteins (Scheme 11) (19,20). In the 1H-NMR spectrum of CyFSF, a broad singlet peak assigned to the aromatic protons of the tyrosine residue appeared at 7.20 ppm (Figure 3). The downfield shift of the tyrosine residue in CyF-SF is similar to that in PEG1-SF. This result suggests that the phenolic hydroxyl group of the tyrosine residue in SF reacted with the chlorine atom of the triazine in actPEG1. Johnson et al. reported that cyanuric chloride (2,4,6trichloro-s-triazine) reacted with nucleophilic groups such as amino, imino, and hydroxyl groups (23). Muijlwijk et al. reported that the chlorine atom of cyanuric chloride was substituted for phenolic hydroxyl groups (24). Thus, our NMR analysis results agree with their report. Although the lH-NMR spectrum of PEG1-SF indicated the modification of the tyrosine residue, amino acid analysis of PEG1-SF did not show a reduction in the tyrosine content. In order to prove that the ether linkage between the triazine of actPEGl and the hydroxyl group of the tyrosine residue was completely cleaved under hydrolysis conditions, amino acid analysis of the model compound CyF-SF was carried out. The result of the analysis of CyF-SF indicated a considerable decrease in lysine content and a slight decrease in histidine content but did not show any decrease in tyrosine content (Table I). Moreover, the 1H-NMR spectrum of the hydrolyzed product of CyF-SF for amino acid analysis showed sharp doublet peaks at 6.90 and 7.22 ppm corresponding to the aromatic protons of the tyrosine. These peaks are similar to the peaks of the aromatic protons of the unmodified tyrosine residue. These results suggest that, since the ether bond between the triazine and the hydroxyl group of the tyrosine residue was weaker than the linkages between

the triazine and the amino or imidazole group, the ether bond was completely broken during the hydrolysis. Therefore, although the result of amino acid analysis of PEG1-SF did not show a reduction in the tyrosine content, the tyrosine residue was modified with actPEG1. The aliphatic hydroxyl groups of the serine and threonine residues, which are much less reactive than the phenolic hydroxyl group of the tyrosine residue (251,could react with actPEG1. It was difficult to observe the shift in resonances of the serine and threonine residues in PEG1-SF. The peaks of the a-CH and P-CH2 of the serine residue overlapped with the big PEG peak and the peaks of other residues. The peak of the threonine residue could not be detected because of ita low content. Because the nucleophilicities of the amino group of the lysine residue and the imidazole group of the histidine residue compete with that of the phenolic hydroxyl group of the tyrosine residue (8, 25, 26), these residues should react with actPEG1. However, the peaks of these residues could not be detected by NMR measurement owing to their low contents. From the results of our amino acid analysis and NMR measurement, the reaction sites are considered to be the €-aminogroup of the lysine residue, the imidazole group of the histidine residue, and the phenolic hydroxyl group of the tyrosine residue (Scheme 111). Taking into account the lysine (0.32 mol %), histidine (0.18 mol %), and tyrosine content (5 mol %) in SF, the minimum amounts of the modified residues (3.3mol % ) obtained by the weight measurement seem to be reasonable. The content of PEG in PEG1-SF was also calculated from the integral of the NMR spectrum. The mixture composed of PEG-OMe and SF (the weight ratio of PEGOMe to SF was 0.405) served as a standard. The integral value of the y-CH3 of the valine residue and the p-CH3 of the alanine residue, which were not modified with actPEG1, in the region of 0.8-1.5 ppm was used as the integration reference. The ratio of the integral value of PEG-OMe methylene protons as 3.70 ppm to that of the valine and alanine residues in the region of 0.8-1.5 ppm was 1.97 (Figure 2). In the spectrum of PEG1-SF, the ratio of the integral value of PEG methylene protons at 3.70 ppm to that of the reference valine and alanine

558 Bkoonlugete chem.,Vd. 4,

Gotoh et el.

No. 6, 1993

10

I

1

3500

I

3M)o

I

2500

I

2ooo

I

1500

I

loo0

I

400

IS

2n

25

M

Diffraction angle (29)

I

3s

Figures. X-raydiffractionpatternsoffiilmsofSF(a),SFtreated with 50% (v/v)methanolic aqueous solution (b), and PEG1-SF (C).

Wave number (cm-' )

Figure 4. IR spectra of f i i s of SF (a), the mixture of SF and actPEGl (b), and PEG1-SF (c).

residues in the region of 0.8-1.5 ppm was 13.5. Thus, the weight ratio of PEG to SF in PEG1-SF is calculated to be 2.8 (=(13.5/1.97) X 0.405). By followingthe calculation described in the above section on synthesis of PEG1-SF conjugate, the mole ratio of the amino acid residue to PEG is calculated to be 24 (=(1/75)/(2.8/5000)),and 4.2 mol % of the residues in SF was modified with actPEG1. Therefore, the amount of the modified residues calculated from the integration is more reasonable than that obtained by the weight measurement. IRSpectrum of PEGl-SF. Figure 4 shows IR spectra of the SF films before and after the modification with actPEG1. In the spectrum of PEG1-SF, two new bands appeared at 2877 and 1100 cm-l compared with the spectrum of SF. The bands at 2877 and 1100 cm-l are assigned to -CH2- stretching (27) and C-0-C stretching (a), respectively. Therefore, these new peaks are attributed to PEG. This means that the PEG chains were successfully introduced to SF. The conformational characteristic of PEG1-SF in the solid state was examined. SF exhibited absorption bands at 1657 (amide I), 1544 (amide 11),and 662 cm-l (amide V), which are characteristic of a random-coil conformation (1 1,14,16,29),while PEG1-SF exhibited absorption bands at 1631 (amide I), 1517 (amide 111, and 690 cm-l (amide V), which are characteristic of a &sheet (11, 14, 16, 29). The E t spectrum of a mixture of SF and actPEGl was also measured in comparison and is shown in Figure 4b. The absorption bands of amides I, 11,and V appeared at 1630, 1521, and 700 cm-l, respectively. These bands are characteristic of a @-sheet. These results suggested that the coexistence with the PEG molecule in SF cawed the conformational change from random coil to @-sheet. X-ray Diffraction Pattern of PEG1-SF. The solidstate structure of PEG1-SF was studied in detail by means of X-ray diffraction. In Figure 5, the original SF film showed a broad, weak peak which is attributed to arandomcoil conformation with a low degree of molecular orientation. After immersion of the original SF film in 50% (v/v) methanolic aqueous solution, a peak appeared a t 20.2O corresponding to the 4.39-A spacing. This spacing for the 201 reflection was used as the interchain periodicity of the &sheet crystal (14, 16, 30). The PEG1-SF film showed the peak at 21.5' corresponding to the 4.13-A spacing of the interchain &sheet (Figure 5). The spacing of the interchain periodicity of PEG1-SF is narrower than that of SF. These resultsmean

that the PEG molecules covalently bonding to SF strengthened the hydrogen bondings between SF chains and promoted the formation of the interchain &sheet. Nevertheless, the peak of PEG1-SF widened compared with that of the SF having the &sheet conformation. Although the covalently bonded PEG molecules prompted the formation of the interchain &sheet of SF molecules, they disordered the molecular orientation as a whole. CONCLUSION

Solubilized SF was chemically modified with actPEG1. The reaction sites were the eamino group of the lysine residue, the imidazole group of the histidine residue, and the phenolic hydroxyl group of the tyrosine residue. The PEG molecules covalently bonding to SF narrowed the spacing of the interchain periodicity and promoted the formation of the interchain &sheet in the solid state. LITERATURE CITED (1) Lui, F.-T. and Katz, D. H.(1979) Immunological tolerance to allergenic protein determinants A therapeutic approach forselectiveinhibitionofIgEantibodyproduction,Proc.Natl. Acad. Sei. U.S.A. 76, 1430-1434. (2) Maeda, H.,Ueda,M.,Morinaga, T., and Mataumoto,T. (1986) Conjugation of poly(styrene-co-maleicacid) derivatives to the antitumor protein neocaninostathx Pronounced improvementa in pharmacologicalproperties, J.Med. Chem. 28,455461. (3) Inada,Y., Yoehimoto,T., Matauahima, A,andSaito,Y. (1986) Engineering physicochemical and biological propertiea of proteins by chemical modification, Trends Biotechnol. 4,6873. (4) Abuchowski, A., van Es,T., Palczuk, N. C., and Davis, F. F. (1977) Alteration of immunologicalproperties of bovine serum albumin by covalentattachment ofpolyethyleneglycol, J. Biol. Chem. 252,3578-3581. (5) Abuchowski, A., Kazo, G. M., Verhoest, C. R., Jr., van Ee,T., Kafkewitz, D., Nucci, M. L., Viau, A. T., and Davis, F. F. (1984) Cancer therapy with chemically modified enzymes. I. Antitumor properties of polyethylene glycol-asparaginme conjugates, Cancer Biochem. Biophys. 7, 175-186. (6) King, T. P., Kochoumian, L., and Lichtenetein, L. M. (1977) Preparation and immunochemical properties of methoxypolyethylene glycol-coupled and N-carboxymethylated derivatives of ragweed pollen allergen, antigen E, Arch. Biochem. Biophys. 178,442-450. (7) TnlmhAnhi,K.,Ajima,A.,Yoshimoto,T.,andInada,Y. (1984) Polyethylene glycol-modified catalase exhibits unexpectedly high activity in benzene, Biochem. Biophys. Res. Commun. 125,761-766.

Chemical Modification of Silk Fibroin

(8) Jackson, C.-J. C., Charlton, J. L., Kuzminski, K., h n g , G. M., and Sehon, A. H. (1987) Synthesis, isolation, and characterization of conjugates of ovalbumin with monomethoxypolyethylene glycol using cyanuric chloride as the coupling agent, Anal. Biochem. 165, 114-127. (9) Zalipsky, S., Seltzer, R., and Menon-Rudolph, S. (1992) Evaluation of a new reagent for covalent attachment of polyethylene glycol to proteins, Biotechnol. Appl. Biochem. 15, 100-114. (10) Usui, M. and Matuhasi, T. (1979)IgE-selective and antigenspecific unresponsiveness in mice I. Induction of the unresponsiveness by administration of ovalbumin-pullulan conjugate, J. Immunol. 122, 1266-1272. (11) Kuzuhara, A., Asakura, T., Tomoda, R., and Mataunaga, T. (1987)Use of silk fibroin for enzyme membrane, J.Biotechnol. 5, 199-207. (12) Minoura, N., Aiba, S., Fujiwara, Y., Taguchi, K., and Tsukada, M. (1988)Cell attachment and growth on silk protein membranes, Sen-i Gakkai (SOC. Fiber Sci. Tec., Jpn.) Symposia Preprints A 29-31. (13) Shimura, K., Kikuchi, A., Katagata, Y., and Ohotomo, K. (1982) The occurrence of small component proteins in the cocoon fibroin of Bombyx mori. J. Seric. Sci. Jpn. 51,20-26. (14) Fraser, R. D. B. and MacRae, T. P. (1973) Conformation in fibrous proteins, Conformation in fibrow proteins and related synthetic polypeptides, Chapter 13 Silks, pp 293-343, Academic Press, New York. (15) Mita, K. and Ichimura, S. (1992) The primary structure and organization of silk fibroin gene, 19th International Congress of Entomology Abstracts, pp 628. (16) Ha, W. S., Oh, S. K., Kim, J. H., and Kim, K. Y. (1987) Effect of structural changes on the surface characteristics of regenerated and graft copolymerized silk fibroin films, Sen4 Cakkaishi 43,587-594. (17) Gotoh, Y., Tsukada, M., and Minoura, N. (1992) Chemical modification of arginyl residues in silk fibroin: 1. Reaction of 1,2-~yclohexanedionein borate buffer, Int. J. Biol. Macromol. 14, 198-200. (18) Tsukada, N., Gotoh, Y., and Minoura, N. (1990) Characterization of the regenerated silk fibroin from Bombyx mori, J. Seric. Sci. Jpn. 59, 325-330.

Bioconjugate Chem., Vol. 4, No. 6, 1993 559

(19) Kurihara, K., Horinishi, H.,and Shibata, K. (1963)Reactions of cyanuric halides with proteins I. Bound tyrosine residues of insulin and lysozyme as identified with cyanuric fluoride, Biochim. Biophys. dcta 74, 678-687. (20) Aoyama, M., Kurihara, K., and Shibata, K. (1965) States of amino acid residues in proteins VII. Hydrogen bonding of tyrosine residues in the insulin molecule, Biochim. Biophys. Acta 107, 257-265. (21) Bundi, A. and Wuthrich, K. (1979) 'H-NMR parameters of the common amino acid residuesmeasured in aqueous solutions of the linear tetrapeptides H-Gly-Gly-X-L-Ala-OH, Biopolymers 18, 285-297. (22) Jackman, L. M. and Sternhell, S. (1969)Theory of chemical shift, Applications of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry 2nd ed., pp 61-113, Pergamon Press, Oxford. (23) Johnson, D., Mendicino,J.,Cormier,M., andTravis, J. (1974) Cyanuric chloride: A new coupling reagent for affinity chromatography, Fed. Proc. 33, 1502. (24) van Muijlwijk, A. W., Kieboom, A. P. G., and van Bekkum, H. (1974)Hydrogenolysisof 2,4,6-tris(aryloxy)-s-triazines over palladium: a new method for the replacement of phenolic hydroxyl groups by hydrogen, Recl. Trav. Chim. Pay-Bas 93, 204-206. (25) Vallee, B. L. and Riordan, J. F. (1969)Chemical approaches to the properties of active sites of enzymes, Annu. Rev. Biochem. 38,733-793. (26) Tanford, C. and Hauenstein, J. D. (1956) Hydrogen ion 78, 5287-5291. equilibria of ribonuclease, J. Am. Chem. SOC. (27) Nakanishi, K., Solomon, P. H., and Furutachi, N. (1990) Infrared Absorption Spectroscopy 25th ed., pp 17-20, Nankodo, Tokyo. (28) Nakanishi, K., Solomon, P. H., and Furutachi, N. (1990) Infrared Absorption Spectroscopy 25th ed., pp 35-36, Nankodo, Tokyo. (29) Miyazawa, T. and Blout, E. R. (1961) The infrared spectra of polypeptides in various conformations: Amide I and I1 83, 712-719. bands, J. Am. Chem. SOC. (30) Nagura, M. and Ishikawa, H. (1983) Molecular motion in poly(amino acid)s: 7. Molecular motion in the 8-form crystal of silk fibre, Polymer 24, 820-822.