An improved synthesis of crystalline mammalian ... - Wiley Online Library

E u r J Biochem 145.601 -605 (1984) (

FEBS 1084

An improved synthesis of crystalline mammalian glucagon Svetlana MOJSOV and R. B. MERRIFIELD The Rockefeller University. New York City, New York (Received May 8iAugust 22, 1984) - EJB 84 0480

Mammalian glucagon was synthesized by the stepwise solid-phase method using several improvements developed in recent years. Peptide was assembled on a 4-(oxymethyl)phenylacetamidomethyl-copoly(styrene-divinyl benzene) resin support with N"-t-butoxycarbonyl and benzyl-based side-chain protection for most of the trifunctional amino acids. Crude synthetic glucagon was obtained in 75% yield by deprotection and cleavage from the resin with a new modified H F procedure. Pure material was isolated in 48% overall yield by a one-step purification on preparative C I 8 reverse-phase chromatography. It was crystallized from aqueous solution at pH 9.2. The synthetic glucagon activated adenylate cyclase in rat liver membranes in the same manner as natural glucagon, with both achieving half-maximum activation at a concentration of 7 nM. Glucagon, a 29-residue peptide hormone that regulates glycogenolysis and gluconeogenesis, has in recent years been the subject of numerous studies to determine the relation of the structure of the molecule to its function [l - 51. Our efforts have been directed toward developing a rapid and efficient stepwise solid-phase synthesis for the preparation of glucagon analogues for structure - function studies. We reported [6] the first stepwise solid-phase synthesis of glucagon that produced crystalline material in reasonable yield. In that synthesis, an alkoxybenzyl alcohol resin was used in combination with acidlabile N"-biphenylisopropyloxycarboxyl amino acid derivatives and irri-butyl-based side-chain protection. These amino acid derivatives are quite unstable during synthetic manipulations and, in general, are not readily available. Because rapid syntheses of glucagon analogues will be facilitated by commercially available reagents that can be used without any purification, we decided to synthesize glucagon by a different synthetic scheme. Most of the problems previously encountered in the synthesis of glucagon were circumvented by the new approach which was based on N"-t-butoxycarbonyl and benzyl-based side-chain protection [7] and relied on use of the recently developed 4-(oxymethy1)phenylacetamidomethyl-copoly(styrene-divenyl benzene) resin support (-OCH,-Pam-resin) with improved acidolytic stability [8], the cyclohexyl-protecting group for the /$carboxyl of aspartic acid [9] that minimizes aspartimide formation, and new modified H F conditions for deprotection and cleavage of peptide from the resin [lo]. In addition to benzyl-based protecting groups for the side chains of the nine hydroxyl-containing residues and one lysyl residue, the guanidine groups of the two arginine residues and the imidazole ring of the N-terminal histidine were protected by the p-toluenesulfonyl group. The indole ring of tryptophan -~

Ahhrcviu[ioizs. HPLC, high-pressure liquid chromatography; -OCH2-Pam-resin, 4-(oxymethyl)phenylacetamidomethyl-copoly(styrene-divinyl benzene) resin; abbreviations for amino acids and peptides mainly follow IUPAC- IUB Rccommcndations [see Eur. J . Biochoii. 138.9 - 37 (1 984)] and include the following: Boc, t-butoxycarbonyl; HOBt. 1-hydroxybenzotriazole; cHex, cyclohexyl; Aoc, t-amyloxycarbonyl; T0s.p-toluenesulfonyl; For, formyl; OBzl, benzyl ester; Br-Z. p-bromobenzyloxycarbonyl; Cl-Z. p-chlorobenzyloxycarbonyl.

was protected with the formyl group against alkylation and oxidation. whereas the thioether group of methionine was left unprotected. The protection scheme is shown in Fig. 1. Amino acid derivatives were coupled by activation as anhydrides or active esters and the reactions were monitored by a quantitative ninhydrin test. The synthetic strategy allowed removal of all protecting groups and cleavage of the peptide from the resin to be accomplished in a single step. Homogeneous glucagon was obtained by reverse-phase chromatography in high overall yield. The product was crystallized and was fully active in the adenylate cyclase assay.

EXPERIMENTAL PROCEDURES Materials Commercial reagents were as follows: 1-butyloxycarbonyl (Boc) amino acid derivatives and Aoc-Arg(Tos) (Pennisula) Boc-Asp-OBzl (Vega), copoly(styrene-1 %-divinylbenzene) resin beads (Bio-Rad), 4-(bromomethyl)phenylacetic acid phenacyl ester (RSA Corp, Ardsely, NY, USA), trifluoroacetic acid (Halocarbon Products), dicyclohexylcarbodiimide (Pierce), 1-hydroxybenzotriazole (Aldrich) recrystallized from 70% ethanol, dichloromethane (Fisher) distilled from N a 2 C 0 3 , p-cresol (Aldrich), p-thiocresol, dimethylsulfide, and ethanedithiol (Fluka), 4 M methanesulfonic acid containing 0.2% 3-(2-aminoethyl)indole and toluenesulfonic acid (Pierce), [3H]leucine and [ ''C]glycine (New England Nuclear, Boston, MA, USA). Preparative reverse-phase C18resin (50100 pm) was from Waters Associates. Natural glucagon was purchased from Sigma. Mrthods The purity of the commercial amino acid derivatives was confirmed by melting points and thin-layer chromatography. B ~ c - [ ~ H l L e(specific u activity 3.1 x 10' dpm/mmol) and Boc[14C]Gly (specific activity 1.1 x lo9 dpm/mmol) were synthesized with the help of t-butoxycarbonyl anhydride (Fluka). Boc-Asp(0cHex) was prepared from Boc-Asp-OBzl according to Tam et al. [9].

602 Boc-His(Tos)-Ser(Bzl)-Gln-[

14C]Gly-Thr(Bzl)-Phe-Thr(Bzl)-Ser(Bzl)-Asp(g~hex)-

Tyr(Br-Z)-Ser(Bzl)-Lys(Cl-Z)-Tyr(Br-Z)-Leu-Asp(OcHex)-Ser(Bzl)-Arg(Tos)-

Arg(Tos )-Ala-Gln-Asp(gcHex) -Phe-Val-Gln-Trp(For)-

3H]Leu-Met-Asn-Thr(Bzl )-OCHz-

CgHq-CH2CONH-CH2-CgHq- resin

Fig. 1 . Protcc,tion scheme f o r the steppnise solid-phase synthesis ofglucagon

Mojsov et al. [I91 with 4 equivalents of reagents, and (c) AocArg (Tos) at positions 18 and 17 was coupled both times with dicyclohexylcarbodiimide (Merrifield, unpublished results) in fourfold excess. Special precautions were taken in the deprotection of and subsequent couplings to the three glutamine residues [20]. After incorporation of the methionine residue at position 27, ethanedithiol (0.03% v/v) was added to the 50%) F3CCOOH/CHzCI2 as an alkyl scavenger. The N'-formyl group was found to be stable to this dilute solution 01' ethanedithiol in acid. For most residues, the monitoring data indicated that the second coupling was necessary for quantitative incorporation. In some cases, such as the glutamine residue at position 24, aspartic acid at 21, glutamine at 20, arginine at 17, and lysinc at 12, the ninhydrin test indicated that > 1Yo of uncoupled amino groups remained after the second coupling. A third coupling was then performed, also for 60 min, with HOBt,' dicyclohexylcarbodiimide for the glutamine residues, dicyclohexylcarbodiimide alone for arginine, and preformed symmetrical anhydride in dimethylformamide 1211for aspartic acid and lysine. The ninhydrin values were then satisfactory. RESULTS After deprotection of the glutamine residue at position 24 Assenihly qf the amino acid sequence of giucagon and phenylalanine at 22, the trifluoroacetic acid/CH2CI2(1 : I ) Boc-Thr(Bz1)-OCH2-Pam-resin was synthesized as pre- solution and subsequent CH2C12washes were collected, the viously described [ 141. Substitution of C-terminal Boc- volume of solution was reduced (10 ml) and an aliquot (25 pl) Thr(Bz1) on the resin was determined to be 0.175 mmol/g by was counted. From radioactivity recovered in solution, it was quantitative ninhydrin reaction [I61 and picric acid titration calculated that there was about a 0.13% loss of chains per [I71 on small deprotected samples. The two results were in deprotection step. Amino acid analysis of the hydrolysate after residues 26 good agreement. For the synthesis 1.O g was used. One synthetic cycle consisted of the following steps: (1) and 4 showed that the amount of growing peptide chains CHzClz wash and pre-swell ( 3 x 3 5 ml for 1 min); (2) remained the same at the beginning (residue 26) and toward deprotection with 50% F3CCOOH/CH2CI2(I x 35 ml I-min the end (residue 4) of the synthesis, indicating that no detectprewash + 1 x 35 ml for 20 min); (3) CH2Clzwash (6 x 35 ml able chain termination occurred during the assembly of the for 1 niin); (4) neutralization with 5 % iPr2EtN/CH2CI2 peptide. (3 x 35 ml for 1 min; (5) CHzC12wash (6 x 35 ml for 1 min); (6) preformed symmetrical anhydride coupling in CH2C12(4 Isolation and purifi'cation ofsynthetic glucrigon equivalents, 0.1 M) for 60 min; (7) CH2Cl2 wash (6 x 35 ml Before final deprotection and cleavage of the peptide-resin for 1 min); (8) neutralization with 5 % iPr2EtN/CHzCI2 (3 x 35 rnl for 1 min); (9) CH2C12 wash (6 x 35 ml for 1 min); with HF, an aliquot (0.39 g ; 0.07 mmol) was treated with 50%) (10) repeat steps (6) and (7). For monitoring of the extent of F3CCOOH/CH2CI2(10 ml) containing ethanedithiol (0.03%) coupling reaction, after the repeat of step (7), the entire v/v) for 30 min to remove the N-terminal t-butoxycarbonyl peptide-resin was neutralized with 5% iPr2EtN/CHzC12 group, followed by CH2C12 washes (6 x 10 ml for 1 min), (1 x 35 ml for 1 min), followed by CHzC12washes (6 x 35 ml neutralization with 5 % iPr2 EtN/CH,Cl, (3 x 10 ml for 1 min) for 1 min) and removal of an aliquot (3-6 mg) of peptide and CH2C12washes (6 x 10 ml for 1 min). Preparation qf',fully deprotected glucrrgon by deprotection resin. Aliquots (3-6 mg) were also removed after the first and cleavage qf'protected glucagon-OCH2 Pam-resin with new coupling and neutralization (after step 9). The symmetrical anhydrides of amino acids were prepared H F procedure. Protected glucagon-OCH,Pam-resin (0.17 g; at 0 C for 20 min with 8 equivalents of t-butoxycarbonyl 0.03 mmol) was placed in the Teflon vessel of the H F appaamino acid derivative and 4 equivalents of dicyclohexyl- ratus, with addition of the following reagents: melted p-cresol carbodiimide in CH2C12.Dicyclohexyl urea was removed by (0.25 ml), p-thiocresol (0.25 ml), dimethylsulfide (3.25 ml) filtration before addition of the anhydride to the peptide- [lo]. For H F treatment of the protected glucagon-OCH2Painresin. Exceptions to the above coupling protocol were: (a) resin, methionine was added (0.27 g, 1.7 mmol, 50-fold molar Boc-["HILeu and Boc-[ 14]Gly in 1.6 equivalents and 1.2 equiv- excess over peptide). The vessel was cooled to -78 'C; H F alen ts respectively, were coupled by dicyclohexylcarbodiimide was condensed up to the 5-ml mark and the mixture was kept activation; (b) Boc-Asn and Boc-Gln were coupled with the for 2 h at - 5°C. The solution was then evaporated to remove HOBt,'dicyclohexylcarbodiimide method [I 81 as described by H F and dimethylsulfide, and H F was condensed again to the Hydrolysates of peptide resins were obtained in 12 M HCl/phenol/AcOH (2: 1 : 1) at llO°C, 24 h, as described by Gutte and Merrifield [ I l l . Hydrolysates of natural and synthetic glucagon were obtained in 3 M toluenesulfonic acid [I21 and in 4 M methanesulfonic acid containing 0.2% of 3-(2-aminoethyl)indole for determination of tryptophan [13]. The synthesis was performed with a manual shaker in a reaction vessel described by Merrifield et al. [14]. Cleavage with H F was in a Diaflon H F line from Toho Co. (Osaka, Japan). Analytical HPLC was on the thermostated reversephase pBondapak C 8 column (4 mm x 30 cm, Waters Associates). The chromatograms were monitored by a Schoeffel variable-wavelength ultraviolet photometer at 225 nm and 280 nm. Peptide samples were counted on a Beckman liquid scintillation counter, with 30% efficiency for tritium and 64% efficiency for I4C. Adenylate cyclase assays were performed according to the method of Solomon et al. [I51 with the addition of G T P in the assay system.

603

'O1

2co

Id0

300

Effluent Volume (ml)

Fig. 2. lon-e.diunge chromatogruphy of crude synthetic glucugon. The sample (1.6 pmol) directly after deprotection and clcavage by modificd H F proccdure; column: CM-cellulose (CM-52, 1.6 x 28 cm). Conditions: linear gradient of 0.01 M ammonium acetate in 6 M urea. pH 4.5 (250 ml). and 0.1 M ammonium acetate in 6 M urea, pH 5.4 (250 ml); flow rate: 18 ml/h

5-ml mark. Treatment with the more concentrated H F was carried out for 1 h at -8°C. This lower temperature was maintained to minimize aspartimide formation. After evaporation of the HF, the resin was extracted three times with ethyl acetate (1 5 ml) to remove the organic reagents and then three times with 10% acetic acid (15 ml). The aqueous phase was lyophilized. Methionine and the remaining reagents were removed by dialysis against 10% acetic acid and crude glucagon was recovered by lyophihzdtion. The cleavage yield, determined by amino acid analysis of a hydrolysate of peptideresin remaining after H F cleavage, was 75%. An ultraviolet spectrum of crude glucagon (0.3 mg/ml in 10% acetic acid) showed the maximum absorbance at 278 nm, indicating a quantitative deprotection of the formyl group from the tryptophan residue by the modified H F treatment [22]. CM-c-ellulose ion-exchange chromatography. Crude synthetic peptide (5.1 mg; 1.6 pmol) in 1 ml of 0.01 M ammonium acetate buffer in 6 M urea, pH 4.5 (buffer 1) was applied manually on a CM-cellulose (CM-52) column (1.6 x 28 cm). The sample was eluted with a linear gradient of buffer 1 (250 ml) and 0.1 M ammonium acetate in 6 M urea, pH 5.4 (250 ml). The elution was monitored by radioactivity. As shown in Fig. 2, more than 90% of the material was eluted in a single sharp peak. In a separate chromatographic run with natural glucagon the same elutim position of the main peak was detected by ultraviolet absorption at 280 nm. Combined f'ractions were lyophilized. Urea was eliminated by dialysis against 10% acetic acid, and the synthetic glucagon was recovered by lyophilization. The homogeneity of synthetic glucagon purified on ion-exchange chromatography was checked by running an aliquot (20 pg) on the analytical reverse-phase C column (4 mm x 30 cm) with a linear gradient of 30%) to 80% solution B in solution A (details of the solvent systems are given in the legend to Fig. 3). The chromatogram showed that between 85% and 90% of the material eluted with the retention time of the natural material (21.5 min). In addition two small peaks were also present. When a sample of crude synthetic glucagon (without any purification) was applied on the column described above, the elution profile looked similar to the one shown in Fig. 3. Purifktrtioti of'crudesynthetic glucugon by preparative C1 rewrse-phusr c,hronicitogruphy. Crude synthetic glucagon (10.5 mg, 3 pmol) obtained from the modified H F deprotection and cleavage (chromatogram shown in Fig. 3) was purified on a preparative reverse-phase C I 8column (2.6 x 28 cm).

:0.1,

o j

x

::

-40

Q

o

10 20 30 Retention Time (min)

40

Fig. 3 . HPLC' on anulytical reverse-phase pBondapuk C', column ( 4 mm x 30 cm) of 20 pg qf synthetic glucugon purified by C'Mcellulose ion-exchange chromatography. Solvent systems were as follows: buffer A, 90% phosphoric acid (0.1 %) and 10% CH3CN; buffer B, 50% phosphoric acid (0.1%) and 50% CH3CN. The sample was eluted with linear gradient of 30% B to 80% B obtained in 45 min at flow rate of 2 ml/min. Elution profile was monitored at 225 nm

A shallow linear gradient of 30% solution B in A (200 ml) to 50% B in A (200 ml) was applied to obtain a greater separation of the two by-products from the major desired peak. The flow rate was 54ml/h; the eluant was monitored by radioactivity. The early by-product eluted at a volume of 180 ml (45% B in A), and synthetic glucagon at 234 ml (47% B in A). To obtain the by-product eluting after the synthetic glucagon peak at the end of the gradient, the column was washed with 50% B in A (100 ml), and the material was eluted at 30-ml volume. Pooled fractions for each of the three peaks were neutralized with 1 M NaOH, acetonitrile was removed by evaporation under vacuum and the material was recovered by lyophilization. The peptides were desalted on Sephadex G-10 (2.5 x 23 cm) using 10% acetic acid as the eluant. The material from the second peak of the reverse-phase column, which corresponded to synthetic glucagon, was isolated after desalting in 5 mg yield (48% overall yield). Chromatography of a n aliquot (20 pg) of purified synthetic glucagon on the analytical HPLC reverse-phase C I 8 column showed only a sharp single peak eluting at 21.5 min. Preparation of N'+rmyI glucagon by deprotection uncl cleuvuge of protected glucugon-OCH, Pam-resin with HF/pcresol ( 9 :I )

An aliquot of protected glucagon-OCH,Pam-resin (15 mg; 2.6 pmol) was placed in the Teflon vessel of the H F

604

apparatus, and p-cresol (0.5 ml) and methionine (0.1 g, 0.67 mmol) were added. After cooling to - 78"C, H F (4.5 ml) was distilled into the vessel, and the reaction proceeded with stirring for 1 h at - 5 ' C . Work-up of the sample was the same as that described earlier for the sample treated with the modified HF. Absorption spectra of this sample (0.3 mg in 1 ml 10% acetic acid) had maxima at 251 nm, 286 nm, and 304 nm, indicating the stability of N'-formyl group to the HF/ p-cresol (9: 1 ) reagent. An aliquot (20 pl) of the crude synthetic N-formyl glucagon was eluted from the analytical HPLC reverse-phase C, column under the same conditions as those described o n at 21 min, confirming in Fig. 3 . [ T r p ( F ~ r ) ~ ~ ] g l u c a geluted previous observation that peptides containing Trp and Trp( For) are not well separated in this system. An additional peak, eluting at 16 min and representing 46% of the material applied was also present. Under the same chromatographic conditions, a sample of authentic synthetic [methionine s~lfoxide~~]glucagon eluted with a retention time of 16 min. Treatment of this crude material with the solvent mixture o f 2 5 % HF. 65% (CH3)2S, 10% p-cresol that contained 50fold molar excess of methionine quantitatively converted it to the desired synthetic glucagon. Since no sulfoxide was obtained during the work up following the new H F procedure, we concluded that oxidation occurred during the synthesis of the peptide chain rather than during the work-up.

I L.-

li

- ~ - , -~

-

7-~--7

10

20

30

40

Reieniion time ( m i d

Fig. 4. HPLC unulysis of nufurul (20 p g ) und syntketic g l u c ~ ~ g o n (20 pg) u n d u i,onditionsdescribed in Fig. 3 3001

Clicircic.teri_utioncfsynthrtic glucagon Aniim cicid rinalysis. Samples of purified synthetic, as well as natural. glucagon wcrc hydrolyzed with 3 M toluenesulfonic acid for 24 h at 130 C and in 4 M methanesulfonic acid containing 0.2% 3-(2-aminoethyl)indole also for 24 h at 110 C. The ratio of amino acids in the synthetic material was: Glucogon [ M I Lys 0.99 (1.00). His 0.98 (1.00), Arg 2.10 (2.00), Trp 0.92 (1.00), Asp + Asn 3.94 (4.00), Thr 2.86 (3.00), Ser 3.90 (4.00), Fig. 5. Acthution of'odc~nylutec~cluscin rat 1ivc.r membrunes. Synthetic Glu(G1n) 2.96 (3.00), Gly 1.02 (1 .OO). Ala 1 .OO (1 .OO), Val 0.98 ( 0 )and natural ( 0 )glucagon were assayed simultaneously (1.00). Met 0.94 (1.00), Leu 1.97 (2.00), Tyr 1.90 (2.00), Phe I .96 (2.00). The values for serine and threonine are not corrected. DISCUSSION Coi.hsotiiutographp of' synthetic and natural glucagon. This stepwise solid-phase synthesis of glucagon, which was Chromatographic analysis of synthetic and natural glucagon on the analytical reverse-phase C I BHPLC is shown in Fig. 4. based on differential acid stability of protecting groups and Commercial natural glucagon eluted at 21 .5 min and con- the ester linkage of the peptide to the solid support, produced tained a small impurity eluting in front of the glucagon peak. material of high homogeneity and good yield. The design of When cochromatographed, a single major compound eluting the synthetic protocol incorporated several recent imat 21.5 min was obtained in addition to the minor impurity. provements in the methodology [8- 10, 16,20,22]; the results Therefore, identity of synthetic and natural glucagon in this obtained showed that the following were most important of system was demonstrated. these. (a) The use of Pam-resin minimized the acidolytic loss C'i.!.siirlIiiution. Synthetic glucagon was crystallized with of peptide chains from the solid support (about 3.6% total). slight modification of the conditions described [6]. Pure leading to increased yield of peptide on the resin at the end synthetic material (1.5 mg) was dissolved in 1.5 ml of 0.2 M of the synthesis. (b) Quantitative ninhydrin monitoring sodium phosphate buffer, pH 9.2, with slow heating to 50"C, facilitated the achievement of almost complete incorporation and the solution was allowed to cool to room temperature. of each residue ( > 99.5%) into the peptide chain, increasing After 48 h it was transferred to the cold room (4°C). Typical the homogeneity of the synthetic material. (c) The modified hexagonal crystals appeared after several days. They were H F procedure deprotected all of the protecting groups in one seen under polarized microscope. step, avoiding the need for the usual base treatment for the Biologicd cictivitj*. Biological activity of synthetic removal of the formyl group of tryptophan. Absorption glucagon was tested for its activation of the adenylate cyclase spectra of the synthetic material showed no evidence ofoxidaof the rat liver membranes. Both the concentration (1 pM) tion or alkylation of the indole ring. In addition, the applicarequired to achieve the maximum activation in the system and tion of the new HF conditions was essential in reducing the dose-related concentration for this activation were the quantitatively methionine sulfoxide that was formed during same as those observed with the natural hormone (Fig. 5). the assembly of the peptide chain to the desired methionine Half-maximum activation was at a concentration of 7 nM. residue.

605

The crude cleaved material contained 85 -90% glucagon. From this material pure synthetic glucagon was isolated in a single purification step by preparative reverse-phase chromatography. The high-resolution properties of the reverse-phase chromatography were superior to those of CMcellulose (CM-52) ion-exchange chromatography in separating glucagon from its by-product. The amount of this byproduct was much more pronounced (about 50%) in the crude synthetic glucagon obtained from a stepwise solid-phase synthesis in which methionine sulfoxide was incorporated into the peptide chain (our unpublished results). We did not detect the presence of methionine sulfonium salt [23] by either ionexchange or HPLC. Presumably the sulfonium salt formed during the synthesis but it was effectively decomposed to methionine during the work-up, which included dialysis in acetic acid. The overall 48% isolated yield of homogeneous glucagon based on the starting C-terminal residues is much higher than the yield obtained in our earlier stepwise solid-phase synthesis of glucagon in which more-acid-labile protecting groups were used [6]. I t is also higher than the yield reported for a solid-phase fragment synthesis or for synthesis by solution methods (see [24]). Since the synthetic glucagon activated adenyl cyclase of rat liver membranes in the same concentration range as natural glucagon, we concluded that our synthesis material had the biological properties expected from the hormone. Considering the ease of purification of synthetic glucagon and the high yield obtained in the synthesis, it seems to be feasible for the first time to approach structure-function studies of the glucagon molecule through the total synthesis of selected analogues. We wish to thank Dr V. lwanij for adenylate cyclase assays, Miss Z. Batarags for excellent technical assistance and Louise Barker Fred for cditorial assistance. This work was supported by grant AM24039 from the United States Public Health Service.

REFERENCES 1. Cote. T. E. & Epand. R. M. (1979) Biochiin. Biophys. Acra 582, 295 -- 306. 2. Bregman, M. D., Trivedi, D. & Hruby, V. J. (1980) J . Riol. Chem. -755. 11 725- 11 731.

S. Mojsov. Endocrine Unit, Massachusetts General Hospital. Fruit Street. Boston. Massachusetts, USA 021 14 R. B. Merrilicld, Rockefeller University, 1230 York Avenue, New York City, New York, USA 10021

3. Coolican, S. A.. Jones, B. N., England. R. D.. Flanders, K. C.. Coudit, J. D. & Curd, R. S. (1982) Bioc/icw?i.strj~21. 49744981. 4. Coy, D. H., Suerras-Diaz, J., Murphy, W. A . & Lance, V. (1983) in Peptides 1983: Proc. 8rh Atnericun Peptide Sjinp. (Hruby, V. J. & Rich, D. € I . , eds) pp. 369-373, Pierce Chemical Co, Rockford IL. 5. Mojsov, S., Lu, G. S., Iwanij, V. & Merrifield, R. B. (1983) in Peprides 1983: Proc. 8th American Peptide Syinp. (Hruby. V. J. & Rich, D. H.. eds) pp. 373-377, Pierce Chemical Co, Rockford, IL. 6. Mojsov, S. & Merrifield. R. B. (1981) Bioclzenfistvj. 20. 29502956. 7. Barany, G. & Merrifield, R. B. (1979) in The Peprides (Gross, E. & Meienhofer, J., eds) vol 2A, pp. 1 -284, Academic Press. New York. 8. Mitchell, A. R., Kent, S . B. H., Engelhard, M. & Merrifield, R. B. (1978) J . O V ~C. ~ W43, I . 2845-2852. 9. Tam. J. P., Wong, T. W., Riemen, M. W.. Tjoeng, F. S. & Merrifield. R. €3. (1979) Tetrahedron L ~ t t .4033-4036. , 10. Tam, J. P., Heath, W. F. & Merrifield, R. B. (1982) nJtralzedron LPtr. 23, 4435-4438. 11. Gutte, B. & Merrifield, R . B. (1971) J . Biol. Clicni. 246, 19221941. 12. Liu, T. Y. & Chang, Y. J. (1971)J. Biol. C/zivn.246, 2842-2848. 13. Simpson, R. J., Neuberger, M. R. & Liu, T. Y. (1976) J . B i d . Chefif.251, 1936- 1940. 14. Merrifield, R. B., Vizioli, L. D. & Boman, H. G. (1982) B i o c h r ~ i s f r y 21, 5020- 5031. 15. Solomon, Y., Londos. C. & Rodbell. M. (1974) A n d . Bioclitwf. 58, 541 -548. 16. Sarin, V., Kent, S. B. H., Tam, J. P. & Merrifield. R . B. (1981) Anal. Biochem. 117. 147 - 157. 17. Gisin, B. F. (1972) Anal. Chim. Actu 58. 248-249. 18. Konig, W. & Geiger, R. (1970) C'hem. Ber. 103, 788-798. 19. Mojsov, S . , Mitchell. A. R. & Merrifield, R. B. (1980) J . Org. Chem. 45, 555 - 560. 20. DiMarchi, R. D., Tam, J. P., Kent, S. B. H. & Merrifield, R. B. (1982) In/. J . Peptide Protein Rex 19. 88-93. 21. Kent, S . B. H. & Mcrrifield, R. B. (1981) in Peprides 1980: Proc 16th Europeun Peptide Symp. (Brunfeld, K., ed.) pp. 328 - 333, Scriptor, Copenhagen. 22. Heath, W., Tam, J. P. & Merrifield, R. B. (1982) J . C. S. Chenn. Conzmun., 896 - 897. 23. Noble, R. L.. Yamashiro. D. & Li, C. H. (1976) J . Am. Clienz. SOC.98. 2324-2328. 24. Merrifield, R. B. & Mojsov, S . (1983) in Handbook o f Experirnental Phurmacolngy (Lefebvre, P. J., ed.) vol. 66/1, pp. 23 - 35, Springer-Verlag, Berlin.