Acyl carrier protein.

A protected linear polypeptide of 74 amino acids with the sequence of Escherichia coli 1 to 74 apo-acyl carrier protein (ACP) was synthesized by the automated solid phase method. The polypeptide was removed from the solid support and partially deprotected by treatment of the peptide-resin with hydrogen bromide and trifluoroacetic acid, and the product was purified by gel titration. The removal of protecting groups was completed by hydrogenation, and the prosthetic group, 4’-phosphopantetheine, was introduced enzymatically with holo-ACP synthetase. Ion exchange chromatography of the product yielded a preparation in which 55 % of the protein in the purified fraction contained the prosthetic group. This synthetic 1 to 74 holo-ACP was as active as native holo-ACP in the malonyl pantetheine-CO2 exchange reaction which is dependent upon malonyl-coenzyme A-ACP transacylase and /3-ketoacyl-ACP synthetase. The purified synthetic 1 to ‘74 holo-ACP preparation was found to be homogeneous and similar to native 1 to 74 holoACP as judged by co-chromatography on DEAE-cellulose and Sephadex G-50 and by sodium dodecyl sulfate disc gel electrophoresis. In addition the synthetic and native proteins were similar with respect to their ultraviolet spectra, amino acid compositions, and their immunological activity with antiserum prepared against native ACP.


Biophysics,
St. Louis, Missouri 63'110 SUMMARY A protected linear polypeptide of 74 amino acids with the sequence of Escherichia coli 1 to 74 apo-acyl carrier protein (ACP) was synthesized by the automated solid phase method. The polypeptide was removed from the solid support and partially deprotected by treatment of the peptide-resin with hydrogen bromide and trifluoroacetic acid, and the product was purified by gel titration.
The removal of protecting groups was completed by hydrogenation, and the prosthetic group, 4'-phosphopantetheine, was introduced enzymatically with holo-ACP synthetase.
Ion exchange chromatography of the product yielded a preparation in which 55 % of the protein in the purified fraction contained the prosthetic group. This synthetic 1 to 74 holo-ACP was as active as native holo-ACP in the malonyl pantetheine-CO2 exchange reaction which is dependent upon malonyl-coenzyme A-ACP transacylase and /3-ketoacyl-ACP synthetase. The purified synthetic 1 to '74 holo-ACP preparation was found to be homogeneous and similar to native 1 to 74 holo-ACP as judged by co-chromatography on DEAE-cellulose and Sephadex G-50 and by sodium dodecyl sulfate disc gel electrophoresis.
In addition the synthetic and native proteins were similar with respect to their ultraviolet spectra, amino acid compositions, and their immunological activity with antiserum prepared against native ACP.
It has the prosthetic group, 4'.phosphopantetheine, which is attached to the protein through a phosphodiester linkage to a serine residue (13,14). The sulfhydryl group of the 4'-* For the preceding report of this series see Reference 1. This investigation was supported in part by Grants 5-ROl-HL 10406 and AM-13025 from theunited StatesPublic Health Service, Grant GB-5142X from the National Science Foundation, and the George Murray Scholarship (University of Adelaide, Australia). phosphopantetheine is the substrate binding site of ACP,' and the intermediates of fatty acid biosynthesis are bound as thioesters.
The enzymes that catalyze reactions involving ACP have a high degree of specificit.y for the protein component of this coenzyme, and little perturbation of its structure is tolerated (15)(16)(17)(18)(19)(20).
It has been shown, for example, that replacement of the ACP of a plant fatty acid-synthesizing system with Escherichia coli ACP causes a change in the spectrum of fatty acids produced (19). Structure-function studies have been initiated through investigations of ACP peptides (15-18) and investigations of the effects of chemical modifications of specific amino acid residues (16,20). Both of these procedures, however, are limited in their ability to define the role of specific residues. It is only by chemical synthesis that specific amino acid substitutions can be systematically prepared, and for this reason the synthesis of ACP was undertaken as a necessary prelude to such a study.
The synthesis of even a small protein, such as ACP containing 77 residues, presents a challenge since at the present time the syntheses of ody a very limited number of proteins have been attempted, e.g. bovine insulin (21,22), ferredoxin (23)) ribonuclease A (24, 25), fragment Ps of Staphylococcus aUreUs nuclease T (26)) cytochrome c (27)) a protein with growth hormone activity (28), and soybean trypsin inhibitor (29). It was decided to use the solid phase synthetic procedure of Merrifield because this method has been found to be much faster and to give a higher yield than the classical approach (30).
Several important prerequisites for the synthesis of ACP have been accomplished.
The complete amino acid sequence of E. coli ACP was determined by Vanaman et al. (31), while Tagaki and Tanford (32) showed that denatured ACP can be readily renatured.
Holo-ACP synthetase, which catalyzes the synthesis of holo-ACP from CoA and apo-ACP, as shown in Reaction 1 (see "Experimental Procedures") was purified from extracts of E. coli and characterized (18,33). Thus if apo-ACP could be synthesized by the solid phase method, the enzyme could be utilized to add the prosthetic group to the deprotected product.
E. coli 1 to 74 ACP, formed by digestion of native 1 to 77 ACP with carboxypeptidase A, was shown to possess full biological activity with ACP synthetase (18) and in fatty acid biosynthesis (15). The synthesis of 1 to 74 apo-ACP would have an important advantage over that of 1 to 77 apo-ACP since ACP 1 to 74 does not contain histidine; histidine residues tend to complicate peptide syntheses (34).
The pentapeptide 33 to 37, around the active site of E. coli ACP, has been synthesized previously by Miura and Sato (35), while the manual synthesis of 1 to 74 ACP has been the subject of a previous communication (36). In this paper the preparation and purification of a protein with ACP activity is described, together with a comparison of the synthetic product with native ACP.
The following side chain blocking groups were used: lysine, benzyloxycarbonyl; arginine, N'-nitro; aspartic acid, P-benzyl ester; threonine, serine and tyrosine, benzyl ethers; glutamic acid, y-benzyl ester, while the t-butoxycarbonyl group was used for a-amino protection. R. coli B, harvested in early exponential growth, was purchased from the Grain Processing Corporation, Muscatine, Iowa. All other chemicals of reagent grade or better were purchased from common sources.
Unlabeled h'. coli ACP and B. coli ACP and CoA labeled with 3H-or %-labeled /%alanine in the prosthetic group were isolated as described previously (37,38). Purified Fraction A and holo-ACP synthetase were prepared according to the method of Elovson and Vagelos (33) and desalted immediately before use.
The holo-ACP thus formed was measured by the holo-ACP-dependent incorporation of 14C02 into malonyl pantetheine that involves the reversibility of Reactions 2 to 4. Reaction 2 is catalyzed by malonyl-CoA-ACP transacylase, and Reactions 3 and 4 are catalyzed by P-ketoacyl-ACP synthetase (39). Malonyl-CoA-ACP transacylase and P-ketoacyl-ACP synthetase are present in Fraction A, a crude enzyme preparation from E. coli (33). In the first stage of the assay the standard reaction mixtures contained 2 pmoles of Tris-HCl, pH 8.0, 0.2 Imole of dithiothreitol, 1 to 30 pmoles of apo-ACP, 24 nmoles of reduced CoA, 0.8 pmole of MgC12, and 0.001 unit of holo-ACP synthetase in a total volume of 0.04 ml. After 10 min at 33" the holo-ACP that was synthesized was assayed in the second stage of the assay by addition of 0.05 ml of a mixture containing 85 nmoles of malonyl pantetheine, 17 nmoles of caproyl pantetheine, 8.5 pmoles of imidazole-HCI, pH 6.2, 2 pmoles of EDTA, 4.5 pmoles of KH14C03 (200 PCi per mmole), and 0.5 mg of Fraction A. The addition of EDTA immediately terminated the holo-ACP synthetase reaction, which requires Mg2+. After 15 min at 33" the exchange reaction was stopped by the addition of 0.01 ml of 4 N HCl, the reaction mixtures were transferred to liquid scintillation counting vials and dried at 100" for 15 min. The samples were counted after the addition of 1 ml of water and 10 ml of Bray's Folution (40). Native apo-ACP residues 1 to 77 and 1 to 74 gave identical values when assayed in this system. Antibody against pure E'. co& native ACP was prepared by Dr. D. A. K. Roncari from rabbits receiving subcutaneous injections of ACP mixed with Freund's complete adjuvant as described previously (41)  counter. Yields for coupling reactions in the synthesis were based on the limiting reactant, the peptide, and were calculated from the amount of the first amino acid, glycine, attached to the resin. Peptide-resin (2 mg) was hydrolyzed with 2 ml of 12 N HCl and propionic acid (1: 1) for 2 hours at 130" according to the method of Scotchler et al. (45). Free peptides were hydrolyzed with 6 N HCl in sealed, evacuated tubes for 24 hours at 110". Amino acid compositions of peptides were determined on a Beckman-Spinco amino acid analyzer.
Solid Phase Synthetic Procedure-A Schwarz automated synthesizer was used for the automated syntheses. The synthesis was based on the stepwise addition of protected amino acids to the carboxyl-terminal amino acid, glycine, which was esterified via a benzyl ester to a polystyrene resin. Most of the procedures common to the solid phase method (46) were used, although some significant modifications were introduced. The rationale for these changes will be described in a separate publication.2 After the desired sequence of 74 amino acids had been assembled (Fig. l), the peptide was cleaved from the resin, deprotected, and isolated as described below.
In a typical synthesis, 0.66 mmole of t-but.oxycarbonyl-glycine esterified to 2 g of a 1% cross-linked polystyrene resin support was used. The sequence of washes utilized in the addition of a single amino acid residue is outlined in Table I. Since the number of washes was increased greatly in our modified procedure, a single manual synthesis of ACP required 3 months, while automation of the procedure decreased this time to 3 weeks. The results of the automated synthesis of ACP are described in this paper, while the manual synthesis was the subject of a previous communication (36). The coupling steps were carried out with 2 W. S. Hancock, I). J. Prescott, P. R. Vagelos, and G. R.   3  2  3  3  3  2  3  3  3  3  3  3  3   1  3  3  3  3  3  3  3   1  3  3  3  3  3  3  3  3   -Time (each  application) a Before the addition of dicyclohexylcarbodiimide, the resin was shaken with the amino acid for 5 min.
a 4-fold excess of the appropriate amino acid, 0.15 M, with dicyclohexylcarbodiimide, 0.15 M, as the coupling reagent except for glutamine and asparagine. These two amino acids were added as the p-nitrophenyl esters. t-Butoxycarbonyl groups were removed by two treatments of the resin with 50% (v/v) trifluoroacetic acid in dichloromethane (CHK&). Cleavage of Peptide from Resin with Hydrogen Fluor&-The reaction was carried out in a polypropylene apparatus, which was based on the design of Pourchot and Johnson (47) and the procedure of Sakakibara et ai. (48). The peptide-resin (0.2 g) was dried in a vacuum overnight, transferred into a polypropylene reaction vessel, and 0.2 ml of anisole and 3 mg of methionine were added to protect the product. The reaction and reservoir vessels were placed on the HF line and nitrogen gas was passed through the apparatus for 1 hour.
HF (10 ml) was distilled into the reservoir vessel and then redistilled into the reaction vessel. The reaction mixture was stirred for 1 hour while the temperature was maintained at 10". The HF was removed under a stream of nitrogen (3 hours), then the last traces were removed with the use of an oil pump.
Residual anisole and its derivatives were removed by ether washes, and the product was dissolved in 5 ml of 0.01 M Tris-HCl, pH 7.3.
Cleavage of Peptidej?om Resin with Hydrogen Bromide-A 0.2-g sample of the peptide-resin was cleaved from the solid support by passing a stream of hydrogen bromide gas through a suspension of the resin in a mixture of 10 ml of trifluoroacetic acid, 2 mg of methionine, and 0.2 ml of anisole. The apparatus described by Stewart and Young (49) was used for the cleavage reaction.
After f$ hour at 25", the cleaved peptides were removed by filtration, the resin was washed with 20 ml of trifluoroacetic acid, and the combined filtrates were evaporated to dryness under reduced pressure.
The product was dissolved in 5 ml of 0.01 M Tris-HCl, pH 7.3 (pH adjusted with base). The cleavage was then repeated on the peptide-resin under exactly the same conditions as before, except that the time of reaction was increased to 1 hour. Deprotection of Cleaved Peptides by Hydrogenation-m'hen HBr was used to cleave the peptide from the support, it was necessary to remove the protecting nitro group on Arge by hydrogenation.
A 20-mg sample of the cleaved product was dissolved in a mixture of 5 ml of acetic acid and 5 ml of water, and 22 mg of 5% Pd-barium sulfate catalyst was added. The sample was hydrogenated at atmospheric pressure for 2 hours. The catalyst was then removed by centrifugation at 2000 X g for 10 min, and the pellet was washed three times with lo-ml portions of an acetic acid-water mixture (1: 1). The supernatants were combined, the solvent was removed under reduced pressure, and the residue was dissolved in 5 ml of 0.01 M Tris-HCl, pH 7.3.

Synthesis of Protected 74 Amino Acid Residue Polypeptide
Chain of 1 to 74 Acyl Carrier Protein-The course of the synthesis was followed by the removal of samples at steps in the synthesis where the first of a given amino acid was added. The amino acid ratios were determined on the acid hydrolysate of the peptide-resin. Even with our modifications to the synthetic procedure, the yield of the growing peptide dropped considerably from G1yT4 to Va165 (Table II).
However, from that point, the yield of the synthesis remained constant at 40 to 500/,, which indicated that t.he peptide chains that had become unavailable for coupling ea,rly in the synthesis did not reinitiate growth at a later stage.
The final weight of the fully protected 1 to 74 ACP-resin was by guest on March 23, 2020 http://www.jbc.org/ Downloaded from 3.6 g, while the product contained 1 NOz-, 31 benzyl-, and 4 benzyloxycarbonyl groups and had a calculated molecular weight of 11,685. The product contained 0.065 mmole of protein per g of protected ACP-resin as estimated by amino acid analysis and by the amount of peptide obtained from the HBr-trifluoroacetic acid cleavage reaction.
This corresponds to 20% of the initial value for the amount of glycine esterified to the resin.

Stability
of Apo-ACP in Conditions Necessary for Cleavage of Peptide-Resin-Before the removal of the protected peptide from the resin could be attempted, it was necessary to determine the stability of native apo-ACP in the conditions necessary for the cleavage of the benzyl ester. Treatment of apo-ACP with I-IF as described in "Experimental Procedures" led to complete inactivation of the protein (Fig. 2). When equal volumes of HF and trifluoroacetic acid were utilized under similar conditions, the recovered apo-ACP was 46% active. The best retention of biological activity (72yJ was obt,ained when apo-ACP was treated with HBr and trifluoroacetic acid. Therefore, this procedure appeared to be suitable for use with the synthetic product.
As HF could not be used to cleave the synthetic apo-ACP from the resin, it was necessary to investigate the use of a second deprotection step, hydrogenation, which would remove the nitro group protecting Arge. Apo-ACP was hydrogenated using the conditions described in "Experimental Procedures." Samples (10%) were removed at different time intervals during the hydrogenation, and the activity of the peptide was measured in the two stage assay for apo-ACP (Fig. 3). In a parallel experiment, the efficiency of the palladium on barium sulfate cat'alyst was tested in the reduction of N-nitro-t-butoxycarbonyl-arginine. Although N-nitro-arginine re:\;idues in some proteins have been found to be resistant to hydrogenation (49), presumably due to steric hindrance, this study would at least give the minimum time necessary for complete removal of the nitro group. A sample of N-nitro-t-butoxycarbonyl-arginine was hydrogenated under conditions identical with those used for apo-ACP deprotection; the reduction was rapid (Fig. 3) as no further release of ammonia was observed after 10 min. The preceding experiments indicated that native apo-ACP could not be quantitatively recovered with any of the cleavage or deprotection procedures.   e All values corrected for the formation of methionine sulfoxide and sulfone.
It appeared that HBr-trifluoroacetic acid treatment for cleavage of the peptide from the resin, followed by hydrogenation to remove the N-nitro protecting group, might yield the best results.
Isolation and PuriJication of Synthetic Product-Samples of the peptide-resin were subjected to the I-IF, HF-trifluoroacetic acid and HBr-trifluoroacetic acid cleavage procedures in an attempt to verify the optimal procedure for cleavage of the peptide from the resin. The yield and conditions of cleavage are described in Table III. The general procedure used for the transesterification studies was based on the work of Beyerman et al. (51), but even with the most vigorous conditions only a small portion of the peptide was cleaved from the resin with methanol-EtaN or benzyl alcohol-EtsN (Table III). From studies of the stability of apo-ACP (Figs. 2 and 3) and the extent of cleavage of the peptide-resin (Table III), it was decided to use HBr-trifluoroacetic acid as the method of cleavage and deprotection.
The amino acid composition of the cleaved product was found to be very similar to that of the peptide-resin (Table IV).
The ratio of the values for tyrosine (residue 71) to phenylalanine (residues 28 and 50), which should be 0.5, was 1.3 in the crude cleaved peptide, indicating the presence of small incomplete peptides in the product.
For comparison, the amino acid analysis of the peptide-resin obtained in the manual synthesis (36) is also shown in Table IV. The crude, partially deprotected material was purified by gel filtration on Sephadex G-25. The elution profile of the column is shown in Fig. 4. Peak I, which had the same elution volume as native ACP, contained 57 '% of the protein (Table V) . Only material from Peak I had any biological activity in the assay with ACP synthetase (see below).
The rest of the material presumably consisted of ACP fragments formed by incomplete coupling reactions.
The amino acid analysis for Peak I is shown in to the crude peptide mixture, can be attributed to the removal of short peptides, e.g. the ratio of tyrosine to phenylalanine decreased to 0.7. For comparison the elution profile obtained with peptide from the manual synthesis (Fig. 4) and the peptide yields in the different peaks of the chromatogram (Table V)   a A total of 4 mg was chromatographed on the column, but the optical densities were increased proportionally in Fig. 4 to match the other sample.
6 A total of 20 mg was chromatographed on the column. c The elution volume for native ACP was found t,o be 72 ml. by removal of the N-vitro group protecting arginine as described under "Experimental Procedures." The recovery of protein from this reaction was 88% and the amino acid analysis of the product was not significantly different from that obtained with material of Peak I (under the acidic conditions required to hydrolyze the sample for amino acid analysis most of the nitroarginine was converted to arginine).
An attempt was then made to convert the synthetic 1 to 74 apoprotein to 1 to 74 holo-ACP by reacting it with holo-ACP synthetase and CoA. To facilitate identification of the holo-ACP, [pantetheine-3H]CoA was used so that the product would be radioactive, and the reaction mixture was chromatographed on DEAE-cellulose-52 (Fig. 5). Peak I was readily identified as the synthetic ACP, as the holo-XCP activity in the malonyl pantetheine-CO2 exchange reaction was catalyzed by fractions in this peak. The fractions of Peak I were pooled, concentrated, and desalted on a Sephadex G-25 column.
The desalted material contained 6 X 104 dpm of 3H radioactivity, which corresponded to 1.5 pmoles of synthetic 1 to 74 holo-ACP. Lowry protein determination and amino acid analysis indicated that the sample contained 22 mg of protein (2.75 pmoles).
Thus 55% of the protein in this sample cont'ained the radioactive 4'phosphopantetheine.
This material was used without further purification in the following studies which attempted to establish the identity of the synthetic product.
The amino acid analysis of the material from Peak I gave a composition which agreed closely with the expected values for 1 to 74 ACP (Table IV).
In addition, ,8-alanine, a constituent of the prosthetic group, could be identified in the analysis, and the value of 0.7 residue per molecule was in close agreement with the above finding that 55% of the protein contained 4'-phosphopantetheine.
Because of the small difference in properties of 1 to 74 ACP and Cob on DEAE-cellulose, a sample of Peak I (3H radioactivity), unlabeled CoA, and native 1 to 77 ACP (W radioactivity) were chromatographed on DEAE-cellulose-52, and the peak of 3H radioactivity was shown to be well resolved from the peak of CoA (Fig. 6). It is of interest that 1 to 77 ACP, which contains three additional residues including a histidine residue, was eluted at a much higher salt concentration than 1 to 74 ACP. 5. Chromatography of products formed by holo-ACP synthetase from synthetic 1 to 74 apo-ACP and CoA. The incubation mixture contained Tris-HCI, pH 8.5,0.2 mM; dithiothreitol, 0.1 mM; [pantetheineJH]CoA, 5 pmoles (2.0 X lo6 dpm) ; synthetic apoprotein, 3.3 rmoles; and ACP synthetase, 30 pg (1.0 enzyme unit) in a total volume of 20 ml. The reaction mixture was incubated at 37" for 20 hours. The mixture was then diluted p-ith water to a conductivity of 0.5 mmho and applied to a DEAEcellulose-52 column (2 X 10 cm), which had been equilibrated with 0.01 M Tris-HCl, pH 7.3,O.OOl M dithiothreitol.
The column was eluted with a linear gradient composed of 250 ml of 0.01 M Tris-HCl, pH 7.3, 0.001 M dithiothreitol, and 250 ml of the same buffer containing 0.3 M LiCl (18.2 mmhos). Fractions (3 ml) were collected and were assayed for 3H radioactivity (O-O), absorbance at 280 nm (O---0), conductivity (A--A), and holo-ACP activity, l4C radioactivity (a). Holo-ACP activity was measured in the malonyl pantetheine-CO2 exchange reaction, described in "Experiment,al Procedures" as the second stage of the two-stage assay for apo-ACP.
The column was eluted with a linear gradient composed of 100 ml of 0.01 M Tris-HCl, pH 7. chromatograph on a DEAE-cellulose column that was developed with a shallow lithium chloride gradient (36), indicating that the synthetic and native products had similar charge properties.
Synthetic 1  A, 80 pg of synthetic 1 to 74 ACP (0.6 X lo3 dpm of 3H radio-t&e&e-W]ACP were chromatographed on a Sephadex G-50 column, and fractions were assayed for 3H and l4C radioactivity. The 3I-I and l4C radioactivity were found to be coincident (Fig. 7)) indicating that the synthetic and native 1 to 74 ACP were of similar size.
Sodium dodecyl sulfate disc gel electrophoresis (Fig. 8) of synthetic 1 to 74 [pantetheine-3H]holo-ACP and native 1 to 77 [panletheiw-SHjholo-ACP indicated that both samples had identical electrophoretic mobility at pH 8.9. A single protein peak, obtained with the synthetic ACP preparation, contained all of the 3H radioactivity, suggesting that the synthetic ACP was homogeneous as measured by this particular technique.
As 1 to 74 ACP contains only 1 tyrosine, 2 phenylalanine, and no histidine or tryptophan residues, there is little absorption in the ultraviolet spectrum of native ACP, and one could expect that the ultraviolet spectrum of synthetic ACP would be significantly different in the presence of protecting groups such as benzyloxycarbonyl and benzyl derivatives. It was found that the ultraviolet spectra of synthetic and native 1 to 74 ACP ( Fig.  9) were completely superimposable at all wave lengths, which indicated that all of the protecting groups had been removed by the procedures used for deprotection.
This was in distinct contrast to the crude material from the HF and trifluoroacetic acid cleavage which had a much greater ultraviolet absorption than the native protein.
Synthetic 1 to 74 holo-ACP was examined for activity in the activity). B, 50 pg of native 1 to 77 ACP (19 X lo3 dpm of 3H radioactivity).
The gels were scanned for absorbance at 280 nm (O--O) and were then sliced. The slices were digested in 0.5 ml of 33% hydrogen peroxide by heating at 100" for 10 min. Bray's solution (10 ml) was added and the 3H radioactivity (bar graphs) was measured. OD, optical density. FIG. 9 (right).
based upon the activities of malonyl-CoA-ACP transacylase and P-ketoacyl-ACP synthetase (39). As seen in Fig. 10, synthetic 1 to 74 holo-ACP was as active as native 1 to 77 holo-ACP in this assay. Furthermore the activity of synthetic 1 to 74 holo-ACP in the exchange reaction (Table VI) was dependent upon malonyl pantetheine-CO2 exchange reaction, a coupled assay The synthesis of ACP was not monitored, except by amino acid analysis, because none of the available analytical methods was found to be satisfactory for use with a rapid automated synthesis. Fortunately, the first residue of each amino acid is scattered throughout the ACP sequence (Table II), and the yield of the addition of these residues gave an estimate of the amount of peptide that was growing at any stage of the synthesis.
The analyses indicated that the amino acid additions occurred in good yield even toward the end of the synthesis, so that an increase in size of the peptide chain did not cause additional steric hindrance for the coupling reactions.
To ensure an adequate yield, however, it was found necessary to repeat all coupling and deprotection reactions and to include a much larger number of washes into the synthetic scheme; for example, extensive use was made of a 3-butanol wash to shrink the resin. Thus the resin was subjected to a swell-shrink-swell cycle before a deprotection or coupling reaction was repeated (Table I). Such a wash scheme has been shown to expose buried functional groups so that repetition of a given step will lead to further reaction with newly exposed terminal residues.2 The extra steps, however, greatly increased the time that was required for a single synthesis, which now involved 6216 washes of the resin. If the synthesis of analogues was to be attempted, it was obvious that the process had to be automated so that the protein could be prepared with a reasonable expenditure of effort. The automation of the synthesis has the further advantage of much greater reproducibility as human error becomes a significant factor in long syntheses.
A commercially available peptide synthesizer was adapted so that the synthesis of ACP could be carried out continuously for 24 hours a day. The automated synthesis gave a product, in a similar yield, indistinguishable from that prepared by the manual method (see "Appendix").
As both coupling and deprotection reactions were repeated, the time of each reaction was decreased to 2 hours and 10 mm, respectively, except for active esters which were still coupled for 6 hours. There is considerable evidence that extended dicyclohexylcarbodiimide-mediated coupling reactions are not useful because of a rapid side reaction, the formation of N-acyl ureas, which consumes the reactants.
One could expect, therefore, that two 2hour couplings would be much more effective than one 4.hour coupling.
As is common with other syntheses, the molar excess of the reagents increased during the synthesis, since the same quantity of reagents was used even though the number of available peptide chains decreased.
The side chain protecting groups were chosen so that cleavage of the benzyl ester linkage of the peptide chain to the resin would remove all of the protecting groups.
This goal could not be achieved because of the instability of ACP to anhydrous HF, and this necessitated the use of HBr and trifluoroacetic acid as the cleavage procedure, with the consequent retention of N-nitro protecting groups. Although benzyloxycarbonyl groups were used to block the e-amino groups of lysine, careful sizing of the cleaved product on Sephades G-50 indicated that little growth had occurred on side chains because of premature removal of this derivative during the synt'hesis. Methionine was used without protection of the thioether, and amino acid analysis of the cleaved peptides ( We wish to thank Mrs. L. Warren and 1Llrs. K. Woodin for expert technical assistance.

APPENDIX
The manual solid phase synthesis of ACP, which was reported previously (36), utilized several modifications which were not used in the subsequent automated syntheses. As is shown in Table VII, the yield of the automated synthesis was significantly better than that of the manual synthesis, although it has been shown that automation of the procedure is not responsible for this increase in the yield.3 It was of interest, therefore, to examine these modifications in the manual synthesis procedure and to determine the reasons for this lower yield.
In the manual synthesis all coupling reactions were repeated, but t,he second coupling reaction was done with 1.5 M urea included in the solvent.
The use of this reagent has been reported to increase the yield of difficult coupling reactions (54). It can be noted in Table IV that the amino acid analysis of the peptide-resin at the end of the manual synthesis was closer to the expected values for ACP than in the automated synthesis, which would suggest that the urea treatment increased the yield of the synthesis. Unfortunately, the properties of the resin after the urea treatment were quite different from the normal polystyrene resin; it had a "shriveled" appearance when dry, as if the urea t,reatment caused cross-linking of the resin. This was reflected in the difficulty experienced in the cleavage and extraction of the product (Table VIII); in fact, only 12% of the product could be estracted from the resin with a wide variety of solvents.
A disadvantage of the use of urea is the presence of isocyanate impurities, which can cause undesirable side reactions such as chain termination by reaction with free amino groups to form thioureas (55).
Acetylation was also used in the first manual synthesis to block partially complete sequences after the addition of residues 2, 10, 20, 47, 62, and 70 (36). Subsequent studies, however, have shown that acetylation is often not effective in blocking unreactive amino groups2 (56, 57) and, in fact, the product from the manual synthesis contained less higher molecular weight material (Tables V and VII and Fig. 4). The use of acetic anhydride has been found to cause side reactions (58), and if such reactions occurred in the manual synthesis, they would explain the greater amount of short peptides in the product.
It was clear that these modifications were harmful in the synthesis of ACP, and, therefore, the use of urea and acetic anhydride were abandoned.