Bioconjugate Chem. 1995, 6, 269-277
269
Synthesis and Use of a New Bromoacetyl-Derivatized Heterotrifunctional Amino Acid for Conjugation of Cyclic RGD-ContainingPeptides Derived from Human Bone Sialoprotein Boris Ivanov,*,+Wojciech Grzesik,* a n d F r a n k A. Robey' Peptide and Immunochemistry Unit, Laboratory of Cell Development and Oncology, and Bone Research Branch, The National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20892. Received October 31, 1994@
A new amino acid derivative, Na-(tert-butyloxycarbonyl)-N~-(bromoacetyl)diaminopropionicacid (BBDap), has been synthesized as a reagent for introducing side-chain bromoacetyl groups into any position of a peptide sequence during solid-phase peptide synthesis. By using minor modifications to the protocol of the automated peptide synthesizer and a two-step in situ neutralization procedure, the syntheses of (bromoacety1)diaminopropionic acid (BDap) in Arg-Gly-Asp-containing peptides from human bone sialoprotein were optimized and completed. Following HPLC purification, the BDapderivatized peptides were cyclized orland conjugated to carrier protein or to glass cover slips. In addition, a new procedure for site-specific conjugation of cyclic peptides to protein carriers or to glass was developed. The cell attachment activity of the peptide derivatives and conjugates was tested in cell adhesion assays with human osteoblasts, and the specificity of the binding was confirmed by competition with linear andor cyclic forms of GRGDS. The results show that conjugates containing the linear and cyclic derivatives of the peptide EPRGDNYR supported cell attachment and spreading in a dose-dependent manner when the peptides were immobilized as described. Cell attachment to the intact bone sialoprotein and to conjugates containing the linear peptides was abolished by competition with linear and cyclic RGD-containing peptides, whereas the attachment to conjugates containing the cyclic peptide was inhibited only partially, and the cell spreading was preserved even in the presence of RGD-peptides.
INTRODUCTION
Conformationally constrained analogs of biologically active peptides have proved to be useful starting points for the rational design of biomaterials containing peptides that possess unique biological activities. It has been shown in a number of reports that cyclizing or polymerizing peptides can influence their receptor selectivity (1), metabolic stability (2),and antigenidimmunogenic properties ( 3 , 4 ) . However, despite all the progress that has been made in the field of peptide chemistry, there still is a paucity of reliable synthetic techniques for making conjugates and polymeric forms of these conformationally constrained peptides. Recently, N-terminal chloroacetylation and bromoacetylation were shown to be very efficient tools for the conjugation of synthetic peptides to proteins and for cyclizing or polymerizing peptides (5- 7). Subsequently, the synthesis of Na-(tert-butyloxycarbonyl)-N'-(N-(bromoacetyl)-/3-alanyl)lysine(BBAL),l a heterotrifunctional spacer which could be placed a t any desirable position in a peptide chain, was reported (8). Although BBAL is quite versatile, there are certain limitations to its use in syntheses of constrained and especially cyclic analogues. These limitations are due to the length (12 carbon bonds) of the side chain bearing the bromoacetyl moiety. Now, we report on the design, preparation, and use of a new reagent, Na-(tert-butyloxycarbonyl)-NB-(bromoacety1)diaminopropionic acid (BBDap),
* Author to whom correspondence should be addressed. Tel: (301) 496-2616. Fax: (301) 402-0823. ' Peptide and Immunochemistry Unit. Bone Research Branch. Abstract published in Advance ACS Abstracts, April 15, 1995.
*
@
which is suitable for automated peptide synthesis and has a much shorter (5 bonds) side chain than BBAL. In this report, we demonstrate the use of BBDap to construct several conformationally constrained peptide analogues designed to mimic the cell attachment site of human bone sialoprotein (BSP), which is a member of a family of RGD-containing proteins that resemble vitronectin, a cell attachment protein in which RGD may be found in quasicyclic conformation (IO). Since cyclic RGD peptides are 100-fold more active toward binding to the vitronectin receptor than binding to the fibronectin receptor (11, it was of interest to us to extend the previous vitronectin findings and study the effect of conformationally constrained RGD-peptides from BSP on osteoblast RGD receptors. The information obtained from such a study, together with novel methods to conjugate the conformationally constrained peptides, could be useful in developing osteoblast-specific therapeutics. EXPERIMENTAL PROCEDURES
Materials and Methods. GRGDS was purchased from Calbiochem Corp. (San Diego, CAI, and ITS+ (insulin, transferrin, and selenium BSA) was obtained from Collaborative Research, Inc. (Bedford, MA). Boc-
+
Abbreviations: AcOH, acetic acid; BBDap, Na-(tert-butyloxycarbonyl)-N4bromoacetyl)diaminopropionicacid; BDap, NP(bromoacety1)diaminopropionic acid; BBAL, Na-(tert-butyloxycarbonyl)-Nc-(N-(bromoacetyl)-B-alanyl)lysine; BSA, bovine serum albumin; BSP, bone sialoprotein; CEC, S-(1-carboxyethy1)cysteine; CMC, S-(carboxymethy1)cysteine; Dap, diaminopropionic acid; DCC, dicyclohexylcarbodiimide;DCM, dichloromethane; DIEA, diisopropylethylamine; DIC, diisopropylcarbodiimide; NMP, 1-methylpyrrolidone;TCEP, tris(carboxyethy1)phosphine; TEA, triethylamine; TFA, trifluoroacetic acid; chl, chloroform; MeOH, methanol.
Not subject to U S . Copyright. Published 1995 by American Chemical Society
lvanov et al.
270 Bioconjugafe Chem., Vol. 6,No. 3, 1995
NMP
DIC
1
1 DISSOLVE
CARTRIDGE
REACTION VESSEL
V I 1 WASH
TFA/DCM
I
TI
TFA
I
DIEA
I
ACTIVATE
RINSE/DRY
I I
PRE-NEUTRALIZE WASH COUPLE/ /RINSE/DRAINI IMIDAZOLE/NMP NEUTRALIZE NMP/DCM
I
I
I
I
I
I 1
IMIDAZOLE
Figure 1. Structure of the the AB1 430A synthetic cycle for BoJDIC chemistry using two-step in situ neutralization. The activator and reaction vessel cycle procedures are aligned. The time scale is not presented.
L-a,P-diaminopropionicacid (Boc-Dap)and Boc-@-alanine were obtained from Bachem Bioscience Inc. (Philadelphia, PA). Bromoacetic acid, 2-bromopropionic acid, 2-(bromomethyl)propionic acid, 3-bromopropionic acid, m-cresole, TEA, DIC, and imidazole were from Aldrich Chemical Co. (Milwaukee, WI). Radioactive tC1J4C1 bromoacetic acid (58.0 mCi/mmol) was from DuPont (Boston, MA). Tris-(2-carboxyethyl)phosphinehydrochloride (TCEP) was supplied by Molecular Probes, Inc. (Eugene, OR). BSA and S-(carboxymethy1)cysteinewere purchased from Sigma (St. Louis, MO). All chemicals for peptide synthesis with the exception of BDap, DIC, and imidazole were purchased from Applied Biosystems, Inc. (Foster City, CAI. Reagent grade methanol, ethyl acetate, diethyl ether, chloroform, and pentane were ordered from Mallinckrodt (Paris, KN). HPLC grade water and acetonitrile were from J. T. Baker Inc. (Phillipsburg, NJ). Analytical and preparative HPLC were performed on a Vydac C18 (4.6 x 250 mm) reversed-phase column (Separation Group, Hesperia, CA) using a Waters 840 HPLC system (Millipore Corp., Millford, MA) or on a Vydac C18 (25 x 250 mm) column with a Waters 600E system, correspondingly. Silica Gel 60 F254 plates (EM Science, Gibbstown, NJ) were used for TLC. Picotag amino acid analysis developed by Waters Associates was performed as described by the manufacturers but with the use of an H P 1090 LC module (Hewlett-Packard, Inc., Gaithersburg, MD). Melting points were determined on a Kofler apparatus and are uncorrected. Elemental analyses of the samples were performed by Galbraith Laboratories, Inc. (Knoxville, TN). Normal human bone cells were obtained as described (11). Briefly, cells were obtained by outgrowth from small ( < 1mm in diameter) collagenase-treated fragments of human trabecular bone. Cells were grown in low (0.2 mM) Ca2+medium (DMEM/Ham's F12, Biofluids Inc., Rockville, MD) supplemented with 2 mM glutamine and 100 units/mL penicillinhtreptomycin (Biofluids Inc.), 50 mg/mL ascorbate, and 10% fetal bovine serum (Gibco BRL, Gaithersburg, MD) until confluent (usually 4-7 weeks). Synthesis of BBDap. A precooled 0.5 M solution of DCC in DCM (120 mL, 60 mmol) was added to a stirred solution of bromoacetic acid (16.56 g, 120 mmol) in DCM (50 mL) a t 0 "C. The reaction mixture was stirred for 30 min a t 0 "C and filtered to remove the dicyclohexylurea that had formed, and the filtrate was evaporated on a Buchi rotary evaporator at 20 "C. The residue was dissolved in 20 mL of acetonitrile, and the solution was immediately added to a solution of Boc-L-Dap (8.32 g, 40 mmol) and TEA (5.6 mL, 40 mmol) in 40 mL of 50% aq.
acetonitrile. The reaction mixture was stirred a t 0 "C for 10 min, a n additional 5.6 mL of TEA (40 mMol) was added, and stirring was continued for 60 min more a t room temperature. The acetonitrile was evaporated, and EtOAc (300 mL) was added to the residual solution. The mixture was successively washed with 0.1 N sulfuric acid (200 mL x 3) and saturated sodium chloride (200 mL x 2), dried over sodium sulfate, and evaporated. The residual oil was dissolved in dry ether (50 mL), and petroleum ether (200 mL) was added. A gum-like precipitate appeared after the mixture stood a t 4 "C for 24 h, and white crystals (9.98 g, 30.7 mmol, 77% yield) were later obtained, which were washed with dry ether (50 mL x2) and pentane (100 mL), filtered, and air-dried. Recrystallization of the product from warm chloroform (15 mL) was performed by adding dry diethyl ether (50 mL). Six and one-half grams (20 mmol, 50%)of the final product that was pure by TLC (ch1:MeOH:AcOH(80:18: 2), chlorine/toluidine staining, Rf0.5) were obtained: mp 138 "C (decomp). Anal. Calcd for C1oH17N2O5Br: C, 36.92; H, 5.27; 0, 24.61; N, 8.62; Br, 24.58. Found: C, 36.80; H, 5.11; 0, 24.88; N, 8.60; Br, 24.61; C1, < 0.5. Synthesis of BDap-Containing Peptides. A twostep neutralization procedure for the AB1 430A peptide synthesizer using a t-Boc synthesis protocol is briefly outlined in Figure 1. Typically the peptides were synthesized on a 0.4-0.45 mmol scale. The peptide synthesis program was created by modifymg the standard NMP/ HOBt "rbocll" program of the instrument (version 1.41, ABI). The additions of DMSO and DIEA to the end of the synthesis cycle as well as the "end capping" procedure were omitted from the reaction vessel cycle. DIEA in reagent bottle 1 was replaced with a 1 M solution of imidazole in NMP which provided mild preneutralization of the peptidyl resins after the removal of the t-Boc group. Because of the high viscosity of NMP, the delivery time of the solution of DIEA in the original rboc 11 program was increased to 11s, providing delivery of 1 mL of the solution. DIC (1M solution in NMP) was used for activation of the amino acids and placed in reagent bottle 8. Reagent bottle 7 was used for delivery of a precise volume of a DIEA solution to the reaction vessel, and programs were written for the activator and concentrator programs to allow delivery from reagent bottle 7 (0.5 M DIEA). Two millimoles of Boc-amino acids, bromocarboxylic acids or BBDap was dissolved in the cartridges in 3 mL of NMP as a part of the "optll" routine (except for Boc-Asn which was dissolved in 2 mL of 1M HOBUNMP before synthesis). Two milliliters of 1 M DIC/NMP was transferred to the cartridges from bottle 8 with the assistance of a
Heterotrifunctional Amino Acid Conjugation
user-defined function, and the activations were allowed to proceed in the cartridges for 30 min with periodic agitation. Then the activated reagents were directly transferred from the cartridges to the reaction vessel, bypassing the activator and concentrator vessels. Then 1 mL of 0.5 M DIENNMP solution was transferred to the reaction vessel uia the cartridge. All of the Boc-amino acids and the bromo-containing reagents were coupled using the same synthetic cycle. Reagent bottles 4 and 5 were not used in the syntheses. Deprotection and Purification. The peptides were cleaved from the resin by treatment with HFlm-cresol (9:1, v/v, -4 "C, 1h). HF was evaporated under reduced pressure, and the crude peptides were precipitated and washed twice with dry ether. The crude peptides were extracted in ice-cold 10% acetic acid (200 mL per gram of the peptidyl resin), purified by preparative HPLC using a gradient of acetonitrile in 0.1% TFA, and lyophilized. Peptide Cyclization. Typical reaction conditions were the following: the peptide to be cyclized (50 pMol) was dissolved in 25-50 mL of 0.1% TFA, and the pH of the solution was brought to 8-9 by the dropwise addition of TEA. Usually, after 5 min a t room temperature, the Ellman test for the presence of free thiols (12) in the reaction mixture was negative. Monitoring the reaction by HPLC typically showed complete conversion of the starting peptide to the cyclic form; the cyclic form elutes 1-3 min earlier on a reversed-phase column than the linear peptide. The yields after HPLC purification ranged from 70 to 85%. Conjugation to BSA. TCEP (28 mg, 100 pmol) was dissolved in PBS (0.5 mL), neutralized with 4 M sodium hydroxide (100 pL, 400 pmol), and immediately mixed with a solution of BSA (34 mg, 0.5 pmol) in PBS (2 mL). After being stirred at room temperature for 30 min, the reaction mixture was loaded onto a column of Toyopearl HW-40f (25 x 500 mm) in degassed PBS. The fractions containing the high molecular weight components were collected, and the BSA concentration was determined by absorbance a t 280 nm. The Ellman test showed 20-22 pmol of free SH groups per 1pmol of the reduced protein. Two to 20 pmol of bromoacetylated (LBP, CBA) or 2-bromopropionylated peptide (CBP, Ccp) was dissolved in the solution of the reduced BSA, and the pH of the reaction mixture was adjusted to 8-9 with 4 M NaOH solution. Monitoring of the reaction by HPLC and the Ellman test showed that the reaction was completed within 2 h a t room temperature where bromoacetylated peptides were conjugated and overnight when 2-bromopropionylated peptides were conjugated. Dialysis of the reaction mixture against water (3 x 5000 mL x 12 h) followed by lyophylization yielded peptide-protein conjugates which contained 10-25% (w/w)peptide as was determined by PicoTag amino acid analysis (see Table 2). Peptide Conjugation to Glass. Microscope cover glasses (22 x 22 mm, No. 1)were soaked in 0.2 M sodium hydroxide solution overnight and rinsed in water until neutral; the glasses then were refluxed in 50% aq. NMP containing 1%(3-mercaptopropyl)trimethoxysilane(Aldrich) for 2 h and successively washed with 50% aq. methanol (3x) and 0.1% acetic acid. Then the glasses were immediately immersed in coupling buffer (degased 0.05 M phosphate, pH 8) containing 50 nmol/mL bromoacetylated peptide and incubated overnight at room temperature. The glasses were transferred to coupling buffer containing 5 mg/mL iodoacetamide, incubated for 2 h, extensively washed with water, and air-dried. For cell attachment assays the glasses next were placed in Petri dishes and washed with serum-free medium.
Biocon/ugate Chem., Vol. 6,No. 3, 1995 271
To estimate the number of free SH groups that (3mercaptopropy1)trimethoxysilane coupled to the glass surface, the thiolated glasses were incubated overnight in coupling buffer containing 10-50 pmol/mL [C1J4C1bromoacetic acid and washed with water (5x). The glasses were placed in scintillation vials with 10 mL of CytoScint (ICN Biomedicals, Inc., Irvine, CAI, and the retained radioactivity was counted with a Beckman LS1801 counter. All mesurements were done in triplicates. Cell Attachment Assay. Bacteriological Petri dishes were coated with 10 mL of substrate solution in PBS containing 1 mM CaC12, resulting in a sample-coated ("dot") area of approximately 0.12 cm2. After 16 h of incubation a t 4 "C, the fluid was aspirated and 10 pL of ice-cold 60% aq. methanol was added to each dot followed by 2 h of incubation a t 4 "C. Methanol was then aspirated, and the plates or prepared glasses were washed for 30 min a t 4 "C with washing buffer (50 mM Tris-HC1, pH 7.8, 110 mM NaC1, 5 mM CaC12,0.1 mM PMSF, 1%BSA, and 0.001% sodium azide). This was followed by washing three times with serum-free DMEM/ Ham's F12 (1:l) medium supplemented with 2 mM glutamine, 100 units /mL penicillidstreptomycin, 0.5% ITS+, and 50 mg/mL ascorbate. Bone cells obtained as described above were trypsinized, centrifuged, resuspended in serum-free medium, and incubated for 30 min a t 37 "C in order to allow recovery after trypsinization. Next, cells were seeded a t a density of 10 000/cm2onto prepared dishes, and simultaneously with plating, several studied compounds (GRGDS, CBA, LBP) were added to the medium. After 24 h of incubation a t 37 "C, the plates were washed twice in serum-free medium to remove nonadherent cells and k e d with 80% aq. methanol at -20 "C for 20 min. Next, cells were stained with 0.1% Amido Black, and the attached cells were counted using the Optomax H V image analyzer. The results were expressed as percent of the number of the cells attached to plastic coated with a 0.2 pM solution of BSP. RESULTS
Nu-(tert-Butyloxycarbonyl)-No-(bromoacetyl)-L-diaminopropionic acid (BBDap) was synthesized from commercial Na-(tert-butyloxycarbonyl)-~-diaminopropionic acid (Boc-Dap) and the symmetric anhydride of bromoacetic acid as shown in Scheme 1. The product was judged to be approximately 95% pure as determined by TLC and HPLC after crystallization from ethyl acetate/ether and recrystallization from chloroformlether. Although some exchange of Br with C1 in the bromoacetyl moiety has been reported when C1-containing reagents have been used for treatment of BBAL (a),in this study we did not observe the phenomenon widely using a saturated solution of NaCl for extractions and chloroform for crystallization. BBDap was obtained in crystalline form and was stable after being allowed to stand as a powder a t room temperature for a t least 6 months. Studies that we have performed on the stability of bromoacetyl moieties in solid-phase peptide synthesis revealed that the bromoacetyl moiety appears to be stable under acidic conditions but unstable in the presence of repeat exposure to a base like TEA or DIEA (data not shown). We found that two or three standard NMP/ HOBt cycles completely destroyed the bromoacetyl moiety introduced in the peptide chain. To minimize the exposure of BDap to a n excess of base during the neutralization step, a n in situ neutralization procedure (13)was slightly modified and programmed into the AB1 430A synthesizer.
lvanov et al.
272 BioconjugafeChem., Vol. 6,No. 3, 1995 Scheme 1
BBDap
Scheme 2
0
I -C
50% A)
BrAc-pAla-BDap-EPRGDNYR-CysNH2
C H 2 1 BrAc-pAla-Dap-EPRGDNYR-CysNH2
pH8-81r-
100%
0
I
R2
R,=Me R p H , Me
The primary feature of the modification is an initial step of preneutralization of resin with a mild base. Briefly, after the TFA deprotection of the t-Boc group, peptidyl resins were preneutralized by a wash solution of imidazole in NMP to remove excess TFA absorbed on the resin (see Figure 1). Transfer of the activated amino acids to the preneutralized resin was followed by the addition of a stoichiometric amount of DIEA to the reaction vessel for completion of the neutralization and coupling steps. The preneutralization step, which neutralized the TFA bound to the resin without deprotonating the Nu-amine on the resin, increased the coupling efficiency by allowing the use of a minimal amount of DIEA for neutralization of the resin. Activations of all amino acids except of Boc-Asn were performed without HOBt to eliminate its possible reaction with the bromoacetyl moiety. Activations of Bocamino acids were carried out with DIC in NMP for 30 min using a 4-fold excess of Boc-amino acids over the amount of the amino groups of the resin. Activated amino acids were then added to the preneutralized resins, and after that, a precise amount of DIEA was added to the reaction vessel to complete the neutralization in situ. A ninhydrin test (14)showed that the coupling yields were usually higher than 99.5% after 15 min a t room temperature. This method yielded peptides containing BDap a t the C-terminus or in a central position of the peptide chain in high yield and with good quality (data not shown). After the crude linear peptides were purified by reversed-phase HPLC, cyclizing the peptides was carried out by raising the pH of 0.1-5 mg/ml peptide solution in 0.1% TFA (see Scheme 2). The absence of free thiols was noted to be complete after 5 min of the reaction a t room temperature. HPLC analysis of the reaction mixtures revealed quantitative conversion of the starting material to the cyclic analogue with a n elution time on reversedphased HPLC that was earlier than that obtained for the
linear peptide (Figure 2). Picotag amino acid analysis of the cyclic peptides showed the presence of stoichiometric amounts of CMC in the cyclized peptide’s hydrolysate, thereby proving that a thioether bond had formed between Cys and BDap (Figure 3a). Peptides with T w o Bromoacetyl Moieties for Conjugating Cyclic Peptides. The designations and structures of the peptides used in this study are presented in Table 1. The introduction of a second bromoacetyl moiety into peptides (CBA, cyclic bromoacetylated RGD peptide; see Table 1)led to the formation of two cyclic products during the cyclization due to alternative attack of the sulfhydryl group by equally reactive bromoacetyl moieties (Scheme 2a). HPLC analysis of the reaction mixture showed the formation of two products with close retention times, and the two products had identical amino acid contents (data not shown). In an attempt to increase the selectivity of the cyclization, we tested the activity of various other bromoacetyl derivatives toward thiols using the described conditions. Model studies showed that a-bromoisobutyryl and P-bromopropionyl moieties of CBB and PCBP (cyclic a-bromoisobutirylated and cyclic ,!?-bromopropionylatedRGD peptides) peptides did not react with N-acetyl cysteine or cysteine-containing peptides under the conditions applied, while an a-bromopropionyl moiety reacted with thiols a t a rate that was much slower than the rate a t which the bromoacetyl moiety would react (data not shown). The difference in the reactivity between bromoacetyl and a-bromopropionyl moieties allows for the selective formation of a thioether bond between Cys and the bromoacetyl group, thereby leaving the slower reacting a-bromopropionyl moiety intact and available for further reaction with other thiol-containing materials under more rigorous reaction conditions (Scheme 2). Peptides CBP and CCP (cyclic a-bromopropionylated RGD peptide and cyclic control a-bromopropionylated non-RGD peptide) bearing Cys a t their C-termini, BDap
Heterotrifunctional Amino Acid Conjugation
Bioconjugafe Chem., Vol. 6,No. 3, 1995 273
0.0
20.0 Retention time (min)
10.0
30.0
Figure 2. Overlaid HPLC profiles of linear precursors of CBP (dashed line) and the product resulting from cyclizing the peptide using the chemistry described herein (solid line). The peptides were eluted from a Vydac C18 (4.6x 250 mm) column with an elution gradient of 0 to 70% acetonitrile in 0.1% TFA and a flow rate of 1 m u m i n .
60:
1
40
U
20
A
I L \-
I k
-
A -
r
2
4 Time
6 Cmin)
8
h
J k . 10
Figure 3. Picotag amino acid analyses of the acid hydrolysate of 1 mg of the cyclic peptides: (A) CNB; (B) CBP. The peptides were obtained from the linear analogues by cyclizing via the bromoacetyl moiety of BDap and Cys, producing CMC (RT 2.06 min), which is stable in the conditions of the acid hydrolysis and clearly resolved in the Picotag system.
in a central position, and a-bromopropionyl moiety a t their N-termini were synthesized. Cyclizing the peptides resulted in the rapid (ea. 5 min) formation of the thioether bond between Cys and BDap. No products formed by cyclizing via the N-terminally branched bromoacetyl moieties were detected by HPLC or amino acid analysis (Figure 3b). Peptides containing bromoacetyl moieties can be conjugated easily to carriers bearing available sulfhydryl groups (6). Two different strategies of peptide anchoring were employed either via the C-terminal bromoacetyl moiety of the linear peptide LBP or via a n additional
N-terminal a-bromopropionic moiety of cyclic peptides CBP and CCP. The linear and cyclic peptides were coupled to thiolcontaining carriers as shown in Scheme 3. In this study, we used thiol-derivatized glass or BSA which had been reduced with the water-soluble phosphine TCEP (15). The peptide-protein conjugates were analyzed by amino acid analysis after acidic hydrolysis of the conjugates. The hydrolyses liberate thioalkylated cysteine derivatives which are formed during the conjugation step (Scheme 3) and can be easily quantitated with amino acid analysis (7). The presence of the cysteine derivatives in the
274 Bioconjugate Chem., Vol. 6,No. 3, 1995
lvanov et al.
Table 1. Designated Abbreviations and Structures of Linear and Cyclic PeDtides Used in the Study
the thioether bond formation and finally, unreacted thiols, which would remain on the glass after the peptide conjugation, were blocked with an excess of iodoacetapeptide peptide structurea mide. designation The number of bromoacetyl-reactive sites on the glass LBP EPRGDNYR-BDap-NH2 surface was estimated with the use of radioactive broLCPb EPRGENYR-BDap-NH2 moacetic acid [C 1J4C1. Treatment of the thiolylated glass CNB Dap-EPRGDNYR-CysNHz LCOCHzwith 10 and 50 nmoUmL bromoacetic acid solutions CBA BrCHzCO-PAla-Dap-EPRGDNYR-C ysNH2 resulted in the same amount of radioactivity retained by LCOCHzthe glass surface, 11f 3 pmoUcmz. This finding indicates CBP BrCH(CH3)CO-/3Ala-Dap-EPRGDNYR-CysNH2 that the number of coupling sites on the modified glass LCOCHzsurface is limited by the amount of free thiol groups PCBP BrCH2CH2CO-BAla-Dap-EPRGDNYR-CysNHz available. Our data correlate with the value of the LCOCHzCCPb BrCH(CH3)C0-/3Ala-Dap-EPRGENYR-CysNHz saturated surface concentration of peptides grafted to tresyl-activated glass, which was 12 pmol/cm2as reported LCOCHzby Massia and Hubbel(16). Since the amount of peptide CBB BrC(CH&CO-/?Ala-Dap-EPRGDNYR-CysNH2 -COCHzin the coupling buffer greatly exceeded the amount sufficient to reach saturation of the glass surface, the a BDap, P-bromoacetyl-a,/3-diaminopropionic acid, single letter peptide-grafted glasses were assumed to have the maxicodes were used for all amino acids except cisteinamide, /3-alanine, mum surface concentration of the peptides. and a,b-diaminopropionic acid. The designations of the control RGE-containing peptides are underlined. Osteoblast Cell Attachment. Finally, the cell attachment activity of the peptides conjugated to BSA Table 2. Compositionaof the Conjugates (Table 2) or glass was tested in the cell attachment assay peptide/ peptide (17). All synthetic peptides conjugated to BSA mediated CMCb/ protein ratio content in cell attachment and spreading in a dose-dependent conjugates Val ratio (mol/mol) conjugates (pmol/mg) manner. BSA-LBP 0.36 13 80 Osteoblast attachment to immobilized intact BSP and BSA-LCP 0.62 22 210 linear peptide-BSA conjugates was abolished by the BSA-CBA 1.03 18 220 addition of linear and cyclic RGD-containing peptides into BSA-CBP 0.09 3 30 the medium. Bone cell attachment to the conjugate BSA-CCP 0.16 6 50 containing the cyclic RGD peptide was inhibited only Data were obtained by Picotag amino acid analysis. CEC in slightly by the addition of a single conformation of the case of BSA-CBP and BSA-CCP. competing peptide (either linear or cyclic) and full blocking of the attachment was observed only when both linear analysis confirms the covalent bond between peptide and and cyclic peptides were simultaneously added into the BSA and allows calculation of the peptidelprotein ratio medium (Figure 5). in the conjugates under study (Table 2). The conjugation To examine whether the short peptides would support of peptides LBP, Lcp (linear bromoacetylated RGD the attachment of the bone cells to a n artificial surface peptide and linear control bromoacetylated non-RGD without a large proteinaceous carrier, we tested the peptide), and CBA via a bromoacetyl moiety resulted in activity of the peptides covalently bound to the glass S-carboxymethylcysteine (CMC) formation, while the microscopic slides. The number of cells attached to the conjugation via racemic 2-bromopropionyl moiety (CBP, peptides covalently linked to the glass proved difficult CCP peptides) formed a pair of diastereomeric S-(lto estimate due to the high nonspecific background carboxyethy1)cysteines(CEC) which were clearly resolved attachment of the cells to the underivatized glass. in the analysis as well (Figure 4). During the course of However, inclusion of cycloheximide ( a protein synthesis this study we learned that there is no benefit to using inhibitor) to the cell growth medium resulted in differoptically pure 2-(+)-bromopropionic acid, which also ences in the cell spreading. Cells that were plated on forms a pair of diastereomeric products because of the glass-CBA (cyclic RGD peptide) surface displayed racemization occurring in the reaction with thiols. spreading, while those plated on glass alone, glass-= A three-step chemical procedure was employed for the (cyclic non-RGD peptide), or glass-LBP (linear RGDconjugation of bromoacetylated peptides to glass. In the peptide) did not spread upon attachment (Figure 6). first step, glass was silanized by a thiol- containing reagent, (3-mercaptopropyl)trimethoxysilane, to present DISCUSSION free thiols as components of the glass surface. Next, the modified glass was treated with bromoacetylated or The coupling of BBDap to any position in a synthetic 2-bromopropionylated peptides under conditions favoring peptide was most successful when the peptide synthesizer Scheme 3
?!
R1 0 C C H 2 7 + Br-i-+3Ala-Dap-EPRGDNYR-CysNH2
5 /
R2
I
R1=R2=Me
xi
Heterotrifunctional Amino Acid Conjugation
Bioconjugate Chem., Vol. 6,No. 3, 1995 275
100 80 60 40
u
20
( u 0 0
100
00
80
0
60
0
40
0
20
0
0
2
6 8 10 (min) Figure 4. Picotag amino acid analyses of the acid hydrolysate of 1 mg of the conjugates of the cyclic peptides with BSA: (A) BSACBA conjugate; (B) BSA-CBP conjugate. These conjugates release S-(carboxymethy1)cysteine or/and S-(1-carboxyethy1)cysteineduring the hydrolysis. 4 Tlms
was modified to minimize undesirable side reactions between the bromoacetyl moiety on BBDap and the Naamine of the peptide, HOBt, and/or the carboxylate on residual TFA. In developing the optimized conditions for the syntheses of BDap-containing peptides, we followed the procedure of in situ neutralization established by Schnolzer et al. (13). However, we found it necessary to modify that procedure even further; side reactions involving the bromoacetyl moieties were most noticeable when an excess of DIEA was present during the reaction with the Na-amine on the resin. To circumvent these side reactions, we minimized the amount of time the bromoacetyl moiety was exposed to the free Na-amine on the resin. In this “preneutralization” step, we used imidazole to neutralize the free TFA that was adsorbed, presumably hydrophobically, to the resin and to the walls of the reaction vessel. Imidazole is basic enough to neutralize free TFA but, as we have learned, not to dissociate the TFA from the Na-amine. Second, following the preneutralization step, we added the bromoacetyl-containing amino acid in its activated form together with a n amount of DIEA that would be needed to deprotonate the Na-amine on the resin. In developing this method, we learned that 2,6-lutidine could substitute for imidazole in the preneutralization step; however, because of the offensive odor of lutidine, we prefer to use imidazole, which is odorless. Conjugates of synthetic peptides are being employed increasingly in biomedical research and biotechnology. Applications of peptide conjugates include implant materials, protein antigens, and immunogens. Recent uses of haloacetyl peptide chemistry generally have been limited to Na-amine derivatization, and past uses of this technology have included the syntheses of cyclic RGD peptides as antithrombotics (181,the preparation of a well-defined sugar-peptide conjugate vaccine candidate against Neisseria meningitidis (19) and the synthesis of a backbone-engineered HIV protease constructed by dovetailing unprotected synthetic peptides (20). This report adds to the growing list of uses for haloacetyl peptide chemistry with the introduction of RGD protein and glass conjugates used as osteoblast adhesion support matrices. A review of RGD peptides has been published (21).
BBDap, a s well as its predecessor, BBAL (8),were designed and synthesized for the purpose of providing additional chemical tools for the growing repertoire of cross-linking agents used to perform controlled interand/or intramolecular peptide conjugation reactions via haloacetyl-derivatized peptide chemistry. In conjunction with the well-known carbodiimide coupling chemistry routinely used to link amino acids, due to the stability of bromoacetyl and chloroacetyl groups in HF, we now are able to place reactive leaving groups a t any position in a peptide, and the applications will include countless new discoveries of conformationally constrained peptidebased derivatives having controlled and specific biological activities, The cyclization reaction of the peptide described in this paper went to completion in under 5 min (see Experimental Procedures). Although peptide cyclization via the formation of disulfide bonds appears to be the preferred method for cyclizing peptides, we have found that, for the amino acid sequences described here, the sulfhydryl oxidation is slow, often taking several days, and often the reaction does not proceed to completion (unpublished data). Products formed using sulfhydryl oxidation also are susceptible to being covalently coupled to thiolcontaining materials via disulfide-exchange reactions. Such potentially permanent modifications of proteins in vivo could lead to autoimmune concerns should the newly modified conjugates be immunogenic in vivo. In addition, cyclizing peptides by using the reaction of a thiol with a haloacetyl moiety present in the peptide provides a very strong thioether linkage in the cyclic peptide that can be oxidized further to add rigidity to the degrees of freedom of the peptide (18). As a n alternative approach to incorporating haloacetyl groups into a peptide prior to HF deprotection, Wetzel et al. (22) developed a method for incorporating a n iodoacetyl moiety a t the amino terminus of a peptide after HF deprotection. The iodoacetic acid anhydride reacts preferentially with the Na-amine of a peptide a t a pH of 6.0, and this allows for the specific tagging of the a-amine with a good leaving group. Template-assembled synthetic proteins (TASPs) have been produced over the last decade for creatively constructing new proteins and enzymes (23, 24). The
lvanov et al.
276 Bioconjugate Chem., Vol. 6,No. 3, 1995
Intact bone sialoprotein
BSA-CBA
BSA-LBP
JL B
iRSP
RSA-CRA
MA-LRP
0 control #!
CRA (cyclic) I1 mM1
Q GRGDS (linear) l l m M l H CRA (cyclic) 10.5 mMI + (;R(;DS (linear) 10.5 mM1
Figure 5. (A) Normal human osteoblastic cells on intact BSP, BSA-CBA, and BSA-LBP conjugates after 24 h of incubation in the absence (control) and presence of linear GRGDS peptide in the medium. Original magnification, 200x. (R) Effect of various blocking peptides on osteoblastic cell attachment to intact BSP, BSA-CBA, and BSA-LRP peptide conjugates. Data present means of triplicates; the dashed line shows the background attachment to BSA. glass
CCP
LBP
CBA ff
t
Figure 6. Normal human osteoblastic cells on glass and CCP, LBP, and CBA peptides conjugated to glass: 24 h of incubation in the presence of cycloheximide; original magnification, 200 x .
approach provided here can complement the TASPs, especially as interest in the syntheses of conformationally constrained a-helices increases. Dawson and Kent (23) recently reported using bromoacetyl-modified peptides in the synthesis of a 4-helix TASP product, and more applications using haloacetyl peptide chemistry to make
conformationally constrained synthetic peptides will be forthcoming from this laboratory. Immobilization of bioactive molecules, e.g., cell adhesion molecules, to control cellular interactions is a wellestablished way to enhance biocompatibility of artificial materials. Covalent attachment of chemically defined
Heterotrifunctional Amino Acid Conjugation peptide ligands is preferred to surface adsorption of proteins or protein conjugates. A method to covalently modify glass has been reported in which the surface hydroxyl groups were activated with a sulfonyl chloride, thus forming a surface-coupled sulfonyl ester (16, 25). This group is then displaced by primary amine and/or thiol groups. It appears that the only obvious limitation t o this process would be the possible inability of the procedure to protect W-amines of lysines from reacting with the glass surface. For many proteins and peptides, the most reactive Nf-amines could be needed to play critical roles in the biological function of the protein or peptide. The methods described here for synthesis, cyclization, and conjugation of peptides appear to be useful tools for studying biological functions of specific sequences within proteins. All BSP-derived peptides analyzed here showed expected biological activity (evaluated by cell attachment assay), and in addition, we were able to detect differences in their action depending on the spatial conformation of the peptide (cyclic vs linear). LITERATURE CITED (1) Pierschbacher, M. D., and Ruoslahti, E. (1987) Influence of stereochemistry of the sequence Arg-Gly-Asp-Xaa on binding specificity in cell adhesion. J. Biol. Chem. 262, 17294- 17298. (2) Szewczuk, Z., Gibbs, B. F., Yue, S. Y., Purisima, E. O., and Konishi, Y. (1992) Conformationally restricted thrombin inhibitors resistant to proteolytic digestion. Biochemistry 31, 9132-9140. (3) Christodoulides, M., McGunness, B. T., and Heckels, J. E. (1993) Immunization with synthetic peptides containing epitopes of the class I outer-membrane protein of Neisseria meningitidis: production of bactericidal antibodies on immunization with a cyclic peptide. J . Gen. Virol. 139, 17291738. (4) Dorow, D. S., Shi, P.-t., Carbone, F. R., Minasian, R., Todd, P. E. E., and Leach, S.J. (1985) Two large immunogenic and antigenic myoglobin peptides and the effect of cyclization. Mol. Zmmunol. 22, 1255-1264. (5) Lindner, W., and Robey, F. A. (1987) Automated synthesis and use of N-chloroacetyl-modified peptides for the preparation of synthetic peptide polymers and peptide-protein immunogens. Znt. J . Pept. Protein Res. 30, 794-800. (6) Robey, F. A,, and Fields, R. L. (1989) Automated synthesis of N-bromoacetyl-modified peptides for the preparation of synthetic peptide polymers, peptide-protein conjugates, and cyclic peptides. Anal. Biochem. 177, 373-377. (7) Kolodny, N., and Robey, F. A. (1990) Conjugation of synthetic peptides to proteins: quantitation from S-carboxymethylcysteine released upon acid hydrolysis. Anal. Biochem. 187, i36-140. (8) Inman, J. K., Highet, P. F., Kolodny, N., and Robey, F. A. (1991) Svnthesis of Na-(tert-butoxvcarbonvl)-Nf4N(bromoacetyl)-~-alanyll-~-lysine: its use in peptide synthesis for placing a bromoacetyl cross-linking function at any desired sequence position. Bioconjugate Chem. 2 , 458-463. (9) Fisher, L. W., McBride, 0. W., Termine, J. D., and Young, M. F. (1990) Human bone sialoprotein: deduced protein sequence and chromosomal localization. J. Biol. Chem. 265, 2347-2351.
Biucunjugafe Chem., Vol. 6,No. 3, 1995 277 (10) Oldberg, A., Franzen, A., and Heinegard, D. (1988) The primary structure of a cell-binding bone sialoprotein. J . Bio2. Chem. 263, 19430-19432. (11) Gehron Robey, P., and Termine, J. D. (1985) Human bone cell in vitro. Calcif. Tissue Int. 37, 453-460. (12) Riddles, P. W., Blakeley, R. L., and Zerner, B. (1983) Reassessment of Ellman’s reagent. Methods Enzymol. 91,4960. (13) Schnolzer, M., Alewood, P., Jones, A., Alewood, D., and Kent, S. B. H. (1992) Zn situ neutralization in Boc-chemistry solid phase peptide synthesis. Rapid, high yield assembly of difficult sequences. Znt. J . Pept. Protein Res. 40, 180-193. (14) Sarin, V. K., Kent, S. B., Tam, J. P., and Merrifield, R. B. (1981) Quantitative monitoring of solid-phase peptide synthesis by the ninhydrin reaction. Anal. Biochem. 117, 147157. (15) Burns, J. A., Butler, J. C., Moran, J., and Whitesides, G. M (1991) Selective reduction of disulfides by tris(2-carboxyethy1)phosphine. J . Org. Chem. 56, 2648-2650. (16) Massia, S. P., and Hubbel, J. A. (1990) Covalent surface immoblization of Arg-Gly-Asp- and Tyr-Ile-Gly-Ser-Argcontaining peptides to obtain well-defined cell-adhesive substrates. Anal. Biochem. 187, 292-301. (17) Grzesik, W. J., and Robey, P. G. (1994) Bone matrix RGD glycoproteins: immunolocalization and interaction with human primary osteoblastic bone cells in vitro. J . Bone Miner. Res. 9, 487-496. (18) Barker, P. L., Bullens, S., Bunting, S., Burdick, D. J., Chan, K. S., Deisher, T., Eigenbrot, C., Gadek, T. R., Gantzos, R., Lipari, M. T., Muir, C. D., Napier, M. A., Pitti, R. M., Padua, A., Quan, C., Stanley, M., Struble, M., Tom, J. Y. K., and Brunier, J . P. (1992) Cyclic RGD peptide analogues as antiplatelet antithrombotics. J . Med. Chem. 35, 2040-2048. (19) Boons, G. J.,Hoozerhout, P., Poolman, J. T., van der Marel, G. A. and van Boom, J. H. (1991) Preparation of a well-defined sugar-peptide conjugate: A possible approach to a synthetic vaccine against Neisseria meningitidis. Bioorg. Med. Chem. Lett. 1, 303-308. (20) Schnolzer, M., and Kent, S. B. (1992) Constructing proteins by dovetailing unprotected synthetic peptides: backboneengineered HIV protease. Science 256, 221-225. (21) Robey, F. A. (1993) Biology and chemistry of extracellular matrix cell attachment proteins, in Biologically Active Peptides: Design, Syntheses and Utilization, (W. V. Williams, and D. B. Weiner, Eds.) Vol. 1,pp 307-324, Technomic Publishing Co., Lancaster, PA. (22) Wetzel, R., Halualani, R., Stults, J. T., and Quan, C. (1990) A general method for highly selective cross-linking of unprotected polypeptides via pH-controlled modification of Nterminal a-amino groups. Bioconjugate Chem. 1 , 114-122. (23) Dawson, P. E., and Kent, S. B. H. (1993) Convenient total synthesis of a 4-helix TASP molecule by chemoselective ligation. J. A m . Chem. SOC.115, 7263-7266. (24) Mutter, M. (1980) In Peptides-Chemistry and Biology, Proceedings of the 10th American Peptide Symposium (G. R. Marshall, Ed.) pp 349-535, Escom, Leiden. (25) Massia, S. P., and Hubbell, J. A. (1991) Human endothelial cell interactions with surface-coupled adhesion peptides on a nonadhesive glass substrate and two polymeric biomaterials. J . Biomed. Mater. Res. 25, 223-242. BC950017K
