THE JOURNAL OF BIOLOGICAL CHEMISTRY

Printed m U.S.A. Val. 256, No. 3, Issue of February 10, pp. 1301-1304, 1981

Protease-mediated Peptide Bond Formation ON SOME UNEXPECTED OUTCOMES DURING ENZYMATIC SYNTHESIS OF LEU-ENKEPHALIN*

(Received for publication, August 6, 1980)

Willi Kullmann From the Max-Planck-Znstitut fur Biophysikalische Chemie, 0.3400 Gottingen, Am Fussberg. Federal Republic of Germany

In the course of a study in protease-controlled peptide synthesis, several promising pathways to synthetic en- kephalins had to be discarded or modified because they failed to give the required products. Partial deprotec- tion of peptide fragments prior to chain elongation resulted in an enhanced susceptibility of scissile bonds to proteolysis. The use of proteases, the specificity of which was not confined solely to the bond to be synthe- sized, thus led to a priori cleavage of pre-existing pep- tide bonds. Subsequently, enzyme-mediated synthesis of new peptide bonds between initial reactants and nascent degradation products furnished undesired, of- ten truncated peptides.

A detailed characterization of the undesired products gave information as to whetber or not the peptide bonds were susceptible to proteolytic cleavage or ac- cessible via enzymatic synthesis. In this way, the un- expected outcome of protease catalysis led to working predictions regarding enzyme-mediated peptide bond formation, and thus, finally contributed to the success- ful synthesis of the target peptides.

Peptide bond formation on a preparative scale by the agency of proteases has become a rapidly developing field (1- 10). In this approach, the enzymatic specificity suppresses the formation of undesired side-products often found in the course of conventional synthesis. In particular, it provides for an outstanding way of maintaining chiral integrity because of the stereoselectivity of the proteases. Furthermore, since the products are often largely insoluble in the reaction media, this method facilitates the purification procedure.

Despite these promising features, the enzymatic procedure has not yet reached general applicability. Much of this draw- back is due to the specificity, which is often not sharp enough to permit predictions whether and to what extent the desired compounds are going to be formed. On the other hand, it cannot be reliably determined whether the possible products formed will be sufficiently insoluble to escape from a poster- iori proteolysis of the newly synthesized peptide bonds.

Notwithstanding the above-mentioned objections, the fact that the lack of unique specificity of some proteolytic enzymes allows for the formation of various kinds of peptide bonds should not be ignored. However, it implies a further risk. The inherent capacity of these proteases to act at a variety of peptidic substrates jeopardizes pre-existing, scissile bonds when dealing with fragment condensation. If hydrolysis oc- curs, the resultant cleavage products may nevertheless act as substrates and give rise to undesired compounds. Conse-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adc~ertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

quently, these imponderables keep the outcome of protease- catalyzed reactions in suspense, thus making methodical de- sign of a synthesis more difficult or even rendering it a trial and error procedure.

Recently, two alternative routes to sequential synthesis of Leu- and Met-enkephalin were reported (10). The peptide bonds were catalytically formed either by papain or n-chy- motrypsin. In the progress of this work, several pathways had to be discontinued because they proved to be dead ends. In this paper, some unsuccessful strategies for the preparation of opioid peptides are described. They are depicted in Fig. 1 together with the synthetic routes that do lead to the desired products.

EXPERIMENTAL PROCEDURES

All chemicals and solvents were reagent grade. Bocl-amino acids

synthesis of Boc-Tyr-OEt (I) and Boc-Phe-OEt (XXI) followed the were purchased from Fluka and were of the L-configuration. The

method reported by Yonemitsu et al. (11). Boc-Gly-NLH2Ph (II), Boc- Phe-NZHsPh (V), and Boc-Leu-NLHLPh (VIII) were prepared by papain-catalyzed condensation in the presence of 2-mercaptoethanol, referring to methods described by Milne and Carpenter (12). Prior to use, their purity was confirmed by thin laver chromatography and elementary analysis.

Deacylation of Boc-amino acid or peptide phenylhydrazides was achieved as reported previously (10).

Boc-Gly-Phe-OH was obtained from Boc-Gly-Phe-NLHLPh (VI) after removal of the phenylhydrazide group by FeCIl treatment, as described for the preparation of Boc-Tyr(Bz1)-Gly-OH (IV) (10).

Papain (EC 3.4.22.2) and thermolysin (EC 3.4.24.4) were obtained from Sigma and Bohringer, respectively.

Prepacked Silica Gel 60 columns were purchased from Merck (Darmstadt). Thin layer chromatograms were developed on precoated silica gel plates (Merck) using the following solvent systems (v/v): System A, chloroform/methanol (3:l); System B. chloroform/meth- anol/acetic acid (45:4:1).

Enzymatic Synthesis (Fig. I)

Boc-Gly-Gly-N2H2Ph (XVZZZ)-Boc-Gly-OH (2.28 g, 13 mniol) and H-Gly-N,HLPh.CF.lCOOH (3.63 g, 13 mmol) were dissolved in 150 ml of 3 M sodium acetate buffer (pH 6.25) which was 2.7 M in potassium chloride. 2-Mercaptoethanol (1.4 ml) was added following which the solution was kept a t 40°C in the presence of papain (2 g) for 40 h. The reaction mixture was extensively extracted with ethyl acetate. The organic layer was washed in succession with 0.1 N HCI, water, 59; aqueous ammonia, water, and dried over NaZSOt. On evaporation, an oil was obtained which was further purified bv means of HI’LC on a prepacked Silica Gel 60 column (44 X 3.7 cm). The pure dipeptide (XVIII) was eluted from the column using chloroform/methanol (50: 1) (v/v). Cr~ystallization from methanol with ether yielded 2.58 g (8.85 mmol, 68%). Chromatographically homogeneous in Systems A and 8; m.p. 175-177°C. ~~~~~ ~~ ~~~~ ~ ~~~~~~~~~~ ~ .~~

I The abbreviations used are: Boc, t-butyloxycarbonyl; Bpoc. 2-(4- biphenyl)-2-propyloxycarbon,~l; Bz. benzoyl; Rzl, benzyl; I’h, phenyl; OEt, ethyl ester; OTMB, 2,4,fi,-trimethylbenzylester; HPLC. high performance liquid chromatography.

1301

1302 Side-reactions during Protease-catalyzed Peptide Synthesis

m,< ; , , i r , , ,,', '.r."-p . . I , ,"Y P&PA *. - qor j , , G , . a . r." ., . . * . .

- -.,, 0 8 , 1 YC,.,

- -I.; 3,m.v- ,. &~-~yrl&!~-~-?+Jf,FMk%'lJ

FIG. 1. Failing and successful pathways to enzymatically prepared Leu-enkephalin. Dashed lines indicate targeted hut missed products. The actual, undesired products are shaded.

Cl~~HrZNt0.t Calculated: C 55.88. H 6.87. N 17.37

Found: C 56.26, H 7.21, N 17.21

Roc-Tyr-Gly-[;Iy-N?H,Ph (XIX)-Boc-Tyr-OEt ( I ) (1.24 g. 4 mmol) and H-Cly-Cly-N,H,l'h.CFI COOH (2.69 g, 8 mmol) were suspended in 40 ml of dimethylformamide. 0.2 M carbonate buffer ( 1 2 ) (v/v) (pH 9.9). After addition of 145 mg of a-chymotrypsin the reaction pro- ceeded under vigorous stirring at room temperature for 10 min and was then stopped by acidification to pH 2.8 using 1 N HCI. The resulting precipitate was removed by filtration and successively washed with water, 0.5 M NaHCO.,. water, and dried in ~ ( t c u o over NaOH. Further purification via HPLC as described for Compound XVIII gave the pure tripeptide (XIX) which was recrystallized from hot ethyl acetate. Homogeneous in Systems A and B. Yield: 1.03 g (2.12 mmol, 53%); m.p. 125-128OC.

Cz.tH:uNr,O,; Calculated: C 59.37, H 6.43 N 14.42

Found: C 59.50. H 6.27 N 14.39

Boc-Tyr(Bzl).GIv.CIy-OH(XX)-Boc-Tyr-CIy-Cly-NZHZl'h (XIX) (970 mg. 2 mmol) was treated in succession with FeCI:,.HzO (6.4 g) and benzyl bromide (0.03 ml, 2.8 mmol) under conditions identical with those previously described for the preparation of Roc-Tyr (Bzl)- Cly-OH ( IV) (10) to give 426 mg of compound XX (0.88 mmol, 44'7). Compound XX was crystallized from methanol/ether and was ho- mogeneous in Systems A and R; m.p. 105-106°C

CzsH.,tN:10: Calculated: C 61.85, H 6.43. N 8.65

Found: C 62.01. H 6.29. N 8.55

Boc-Phe-Leu-NLH2Ph (XXII)-a-Chymotrypsin (30 mg) was added to a suspension of Roc-Phe-OEt (XXI) (293 mg, 1 mmol) and H-Leu- NzH?I'h.CF., COOH (335 mg. 1 mmol) in 9 ml of dimethylformamide. 0.2 M carbonate buffer (1:2) (v/v) (pH 10.1). The reaction proceeded under vigorous stirring for 10 min at room temperature and was terminated by acidification to pH 3.0 using 1 N HCI. The precipitate which had been formed was worked up as depicted for the tripeptide (XIX). A final fractionation step involving HPLC and using chloro- form as eluant provided the dipeptide (XXII). which was homogene- ous in Systems A and B. Recrystallization from ethanol/water yielded 380 mg of Cornpound XXII (0.81 mmol. 817); m.p. 186-188°C.

CzcH.,N,Ot Calculated: C 66.65, H 7.74. N 11.96

Found: C 67.15, H 7.69, N 12.06

Boc-Tyr(Bzl)-Cly-Phe-Leu-N~H~Ph(XYIIII)-Boc-Tyr(Bzl)-Gly- Cly-OH (XX) (194 mg, 0.4 mmol) and H-Phe-Leu-NIHZPh. CF,COOH (193 mg. 0.4 mmol) were dissolved in 6 ml of ethanol/ McIlvain buffer (2:3) (v/v) (pH 6.0). 2-Mercaptoethanol (0.016 ml) was added following which the solution was incubated at 38°C in the presence of papain (20 mg). After 5 h, the resulting precipitate was collected by filtration. washed successively with 5'; citric acid, water. 5% aqueous ammonia, water. and dried in lwcuo over NaOH. The tetrapeptide (XXIII) was eluted from a prepacked silica gel column with chloroform. lkuvstallization from methanol/ether gave chro- matographicallv homogeneous XXIII (Systems A and B). Yield: T2.5 mg (0.29 mmol. V i ) ; m . p . 216-219°C.

CIIH;tN+,O: Calculated: C 67.8:%. H 6.99. N 10.79

Found: C 68.05. H 6.75. N 10.87

Amino acid analysis: Cly. 1 .OO( 1 ); Leu. 0.97 ( 1 ) : Tyr. 0 . 9 1 ( 1 ); I'he 1.04 ( 1 ) . Boc.Tvr-N.H?Ph (S.YIV)-2-Mercal)toethanol (0.14 ml) and pa-

pain ( 1 0 0 mg) were added to a solution of Roc-'l'vr-OH (140 mg. 0.5 mmol) and H-Cly-N,H~l'h.CFICOOH (279 mg. 1 nlmol) i n 10 nll of methanol, 3 M acetate buffer (1:5) (v/v) (pH 4.9). After incubation at 38'C for 50 h, the reaction mixture was worked up as described for Compound XVIII. Crystallization from chloroform gave 30 nlg of Compound XXIV (0.08 mmol. 167 based on Roc-Tyr-OH). Homo- geneous in Systems A and €3; m.p. 178-180OC.

C,,H,:,N Calculated: C 64.64. H 6.78, N 1 1 . 3 6

Found: C 64.32, H 6.M. N 11.52

Amino acid analysis: Cly, negative. Tyr. positive. Boc-GIwLeu.N?HJ'h (SS~')-Boc-<;ly-I'he-OH (130 mg. 0.4

mmol) and H-Leu-N,H?l'h.CF,COOH ( 134 mg. 0.4 mmol) were dis- solved in 5 ml of ethanol/McIlvain buffer ( 1 2 ) (v/v) (pH 6.1). The solution was reacted with 2-mercaptoethanol (0.032 ml) and papain (40 mg) at 38°C for 36 h. The reaction mixture was then repeatedly extracted with ethyl acetate and worked u p as reported ahove for the dipeptide (XVIII). After fractionation via HI'LC, the protected di- peptide (XXV) was eluted with chloroform/methanoI (8O:I) (v/v) following which Boc-Clv-N!H.l'h (11) was washed from the column using chloroform/methanol (40:l) (v/v) as eluant. The products were homogeneous in Systems A and R. Compound XXV crystallized from ethanol/water to yield 72 mg (0.19 mmol. 48';) n1.p. 138-139°C.

Ct*+HwNtOa Calculated: C W.:30. H 7.111). N 14.80

Found: C 60.23, H 7.96. N 14.78

Amino acid analysis: Cly, l.OO(1); Leu, 0.96( 1); I'he, negative. Boc- Gly-NyH?l'h i l I ) was crystallized from ethvl acetate with petroleum ether to give I L 6 mg (0.04 mmol, 10%). m.p. 122-124°C.

CI ,HI*,N IO., Calculated: C 58.84. H 722. N 15.84

Found: C 58.71, H 7.1 1, N 15.80

Amino acid analysis: Gly. positive; Leu, negative; I'ht., negative. Boc-Tyr(Bzl)-Phe-NzH?Ph (XXVI)-Boc-l'yr(Bzl)-Glv-Cly-OH

(XX) (145 mg. 0.3 mmol) and H-l'he-NZH,I'h.CF,COOH (1 10 mg, 0.3 mmol) were dissolved in 6 ml of methanol/Veronal buffer ( 1 2 ) (v/v) (pH 7.5). After addition of thermolysin (25 mg), the solution was kept a t 38OC for 18 h. The resulting precipitate was removed by filtration and worked up as described for Compound XXII. Homogeneous in Systems A and B. Recrystallization from ethyl acetate/petroleum ether gave 91 mg of XXVI (0.15 mmol, 5 0 % ) . m.p. 218-220°C.

C.W,H.~,N.IO~, Calculated: C 71.03, H 6.62. N 9.20

Found: C 71.06. H 6.37, N 9.43

Amino acid analysis Gly. negative; Tyr. O.W( 1); Phe. I.OO(1).

RESULTS AND DISCUSSION

The fmt design of a protease-catalyzed synthesis of Leu- enkephalin implied a final fragment condensation of Boc- Tyr(Bz1)-Gly-Gly-OH (XX) and H-Phe-Leu-N2HzPh with pa-

Side-reactions during Protease-catalyzed Peptide Synthesis 1303

pain (Fig. 1). The synthesis of the protected tripeptide (xx) started with the preparation of Boc-Gly-Gly-NzHJ'h (XVIII), which was readily available in 68% yield after 40 h in a solution of high ionic strength (13) from Boc-Gly-OH and H-Gly- NnH2Ph in the presence of papain. Boc-Tyr-OEt (1) and 2 eq of the deacylated dipeptide (XVIII) were treated with a- chymotrypsin for 10 min to furnish Boc-Tyr-Gly-Gly-NzH2Ph (XIX) (53%). The phenylhydrazide group of the tripeptide (XIX) was removed by means of ferric chloride followed by benzylation of the phenol function of tyrosine with benzyl bromide, to provide Boc-Tyr(Bz1)-Gly-Gly-OH (XX) (44%).

The modification of the tyrosine residue had become indis- pensable because the tripeptide Boc-Tyr-Gly-Gly-OH did not react to a significant extent with H-Phe-Leu-NJHzPh in the presence of papain. This finding, which is in accordance with the failure of Boc-Tyr-Gly-OH to couple with H-Gly-Phe-Leu- NzH2Ph (lo), may be explained by the assumption that the diminished solubility of the resultant products brought about by the additional, bulky benzyl group favored peptide bond formation.

Boc-Phe-Leu-NeHnPh (XXII) was prepared via a-chymo- trypsin-mediated coupling of Boc-Phe-OEt (XXI) and H-Leu- N2H2Ph. The reaction was stopped after 10 min to give 81% of XXII. Two alternative routes also led to the synthesis of Boc- Phe-Leu-N2H2Ph. Starting from Boc-Phe-OH and H-Leu- NrH2Ph the desired dipeptide (XXII) was accessible both by papain and thermolysin catalysis after 24 h at 38°C in 71% and 58% yield, respectively.

Boc-Tyr(Bz1)-Gly-Gly-OH (XX) and H-Phe-Leu-NpHzPh were reacted in the presence of papain for 4 h to produce the tetrapeptide Boc-Tyr(Bzl)-Gly-Phe-Leu-NnHzPh (XXIII) in 72% yield. The structure of Compound XXIII was confirmed by amino acid and elemental analysis. Obviously the COOH- terminal glycine residue of the tripeptide (XX) had been proteolytically removed followed by a papain-induced peptide bond formation between the resulting dipeptide Boc-Tyr(Bz1)- Gly-OH and H-Phe-Leu-N2H2Ph. Evidence for this suggestion is supplied by the emergence of free glycine, which was detectable on thin layer chromatograms. Further support for this view is given by a report of Schechter and Berger (14), who pointed out that peptides containing a phenylalanine residue preceding the residue which contributes the carboxyl group to the bond to be cleaved are suitable substrates for papain catalysis. This conclusion applies also for peptides containing a benzylated tyrosine instead of a phenylalanine residue a t this position.

Additional attempts to prepare slightly modified Leu-en- kephalin derivatives failed as well and produced the corre- sponding tetrapeptides. These experiments were carried out under conditions identical with those described for the syn- thesis of the tetrapeptide (XXIII), with the exception of the amino component, the phenylhydrazide moiety of which was replaced by a 2,4,6-trirnethylbenzyl ester and a t-butyl ester.

Despite the discouraging outcome, the unexpected reactions revealed a successful pathway to the synthesis of the required pentapeptide (XIV). The interpretation of the foregoing re- sults indicated that the Gly-Gly bond of an enkephalin deriv- ative could be accessible to papain catalysis whether for hydrolysis or synthesis, and that both the Tyr-Gly and the Phe-Leu bond would remain unaffected. A promising ap- proach appeared one in which Boc-Tyr(Bz1)-Gly-OH (IV) and H-Gly-Phe-Leu-N?H?Ph were coupled in the presence of pa- pain, although it risked the cleavage of the Gly-Phe bond of the amino component (see below). As recently described (10) and illustrated in Fig. 1, this strategy led to the Leu-enkeph- alin derivative (XIV).

The dipeptide Roc-Tyr-Gly-N2H,Ph (HI) , a precursor of

the above carboxyl component (IV), was readily obtained by a-chymotrypsin catalysis (10). As an alternative approach to the synthesis of this dipeptide, an attempt to condense Boc- Tyr-OH and H-Gly-N2H2Ph using papain as catalyst was made. The desired product was not obtained, but a different compound was formed in 16% yield, which was proved to be Boc-Tyr-N2H2Ph (XXIV). The tyrosyl derivative had presum- ably formed via transamidation in analogy to a papain-aided reaction between Bz-Phe-OH and H-Gly-NH-Ph that, accord- ing to Janssen et al. (15), gave Bz-Phe-NH-Ph. These results are in agreement with the observed inertness of Tyr-Gly bonds in the presence of papain (see above), thus indicating that the protease hardly affects this peptide bond whether for hydro- lytic or for synthetic purposes.

The tripeptide X, the deacylated form of which is men- tioned above in the context of the synthesis of the protected pentapeptide (XIV), was prepared by an a-chymotrypsin-me- diated coupling of Boc-Gly-Phe-OEt (VII) and H-Leu-N?H?Ph (10). A previous attempt to obtain this tripeptide via papain catalysis using Boc-Gly-Phe-OH and H-Leu-NsHnPh as reac- tants failed due to an apriori proteolysis of the Gly-Phe bond of the carboxyl component. The cleavage products, Boc-Gly- OH and H-Phe-OH, were easily detected by thin layer chro- matography. The major, unexpected product (48%) of this enzymatic reaction was the truncated, with respect to the intended tripeptide (X), dipeptide Boc-Gly-Leu-N,H2Ph (XXV) (Fig. I), which had obviously formed from the released Boc-Gly-OH and H-Leu-NzHnPh. In addition, Boc-Gly- N2HzPh (11) was formed in 10% yield via transamidation, suggesting a partial removal of the phenylhydrazide group from H-Leu-NzH2Ph.

In a further experiment we attempted to prepare Bpoc-Gly- Phe-Leu-OTMB from Bpoc-Gly-Phe-OH and H-Leu-OTMB by papain catalysis. The intention was that the drastically diminished solubility of the potential product would accelerate the synthetic reaction so as to exceed the rate of hydrolysis of the Gly-Phe bond. But again, the desired tripeptide could not be obtained; the same cleavage pattern as above led to the synthesis of Bpoc-Gly-Leu-OTMB.

An alternative concept for the preparation of Leu-enkeph- alin presupposed a final a-chymotrypsin-assisted coupling step between Boc- Tyr(Bz1)-Gly-Gly-Phe-OEt and H-Leu-NeH2Ph (10). In order to synthesize Boc-Tyr(Bz1)-Gly-Gly-Phe- N,H,Ph (XII), a precursor of the foregoing ethyl ester, we proposed reaction of Boc-Tyr(Bz1)-Gly-Gly-OH (XX) and H- Phe-N2H2Ph. Bearing in mind the undesired results obtained with papain on the protected tripeptide (XX), we decided to use thermolysin as catalyst. As reported by Morihara and Tsuzuki (16), who explored the specificity of thermolysin using synthetic benzyloxycarbonyl-dipeptidyl amides as substrates, the proteinase exhibits a considerable preference to cleave Gly-Phe, only a small tendency to split Tyr-Gly, and a negli- gible capacity to hydrolyze Gly-Gly bonds. Consequently, the thermolysin-mediated synthesis appeared to possess a realistic chance to provide the required tetrapeptide derivative (XII). In fact, the thermolysin-catalyzed reaction produced the di- peptide Boc-Tyr(Bzl)-Phe-NeHz (XXVI) (Fig. 1) in 50(1 yield. According to thin layer chromatographic data, the Tyr-Gly bond of tripeptide (XX) was hydrolyzed at the very beginning to release Boc-Tyr(Bz1)-OH and H-Gly-Gly-OH. The former cleavage product reacted with H-Phe-N2H2Ph to furnish Boc- Tyr(Bzl)-Phe-NnHZPh (XXVI). There was no evidence for further degradation of the dipeptide glycylglycine.

During this work, the enzymatic elongation of peptide chains required the partial deprotection of one or even two fully protected intermediates. The newly formed reactants exhibited enhanced solubility in aqueous media and thereby

1304 Side-reactions during Protease-catalyzed Peptide Synthesis

increased the susceptibility to proteolysis of their scissile bonds. The drive toward hydrolytic action was additionally strengthened by the imminent tendency of partially depro- tected substrates to furnish energetically favored dipolar ions on peptide bond cleavage. Non-reactivity of these ions proteo- lytically split off prior to enzymatic synthesis led to the truncated peptides characterized above. A different kind of undesired products resulted from reutilization of formerly excised free amino acids during protease-mediated peptide synthesis. The random insertion of surplus amino acids or their enzymatically derived oligomers led to fully protected peptides which exceeded the required chain length.

This phenomenon was described by Janssen et al. (15) for the papain-aided reaction of Bz-Phe-OH and H-Gly-NH-Ph that, in addition to the above-mentioned Bz-Phe-NH-Ph, yielded the tripeptide Bz-Phe-Gly-Gly-NH-Ph. In the course of the present study, an attempt to prepare via papain catal- ysis the tripeptide Boc-Gly-Gly-Phe-N2H2Ph starting with either Boc-Gly-OH and H-Gly-Phe-N,H2Ph or Boc-Gly-Gly- OH and H-Phe-N2H2Ph resulted in the formation of Boc- peptidyl phenylhydrazides the amino acid composition of which was determined to be (approximately): Gly(l), Phe(2); and Gly(6), Phe(1): and Gly(3), Phe(l), Gly(l), Phe(2), re- spectively. These findings suggest proteolytic cleavage of both the peptide bonds of Boc-Gly-Gly-OH and H-Gly-Phe- N2H2Ph and the partial removal of the phenylhydrazide moi- ety of the amino components followed by papain-induced introduction of a varying number of free amino acids. As proposed by Anderson and Luisi for the oligomerization of amino acids esters by papain (17), the enzymatic reaction may need an initiator like Boc-Gly-OH to begin with. The elonga- tion of the peptide chain then may be terminated by intro- ducing an amino acyl phenylhydrazide, which probably favors the precipitation of the products. The chain length was not determined and cannot be derived from amino acid analyses, which served solely to reveal the failure to synthesize the desired peptides.

It is self-evident that a generally valid method of suppress- ing the undesired properties of the proteases does not exist. However, the synthetic pathways which proved to be dead ends can be by-passed by reasonable detours. This requires the selection of suitable peptide fragments to serve as sub- strates. This purpose is facilitated by a detailed interpretation of the results of undesired enzymatic reactions with particular emphasis on the site where preferential splitting of peptide bonds occurs. By taking into account the fact that the prote- olytic and synthetic specificities of the proteases are fre- quently indistinguishable, the enzymes can be beaten at their own game. Furthermore, the endangered reactants can be rendered more resistant to a priori proteolysis by increasing their hydrophobicity. Consequently, they are removed more rapidly from the solution as integral parts of the newly formed

insoluble products. This procedure proved suitable when ap- plied to the synthesis of the pentapeptide (XIV). An additional benzyl group enabled the preparation of the desired enkeph- alin derivative in spite of the presence of two peptide bonds susceptible to protease action. Unfortunately, this principle is not generally applicable, as demonstrated both by the failure to obtain the appropriate pentapeptide derivatives using Boc- Tyr(Bz1)-Gly-Gly-OH (XX) and varying phenylalanylleucine peptides displaying increasingly hydrophobic properties (see above) and by the inability of reacting Bpoc-Gly-Phe-OH and H-Leu-OTMB to yield the corresponding tripeptide. In both cases proteolytic cleavage of preexisting peptide bonds pre- vented the synthesis of the required products. Last but not least, modification of synthetic concepts to enable the use of more than one protease with noninterfering specificities can provide for pathways leading to the target peptides.

In summary, application in peptide synthesis of proteases exhibiting a broad specificity cannot ensure an unambiguously predictable outcome. To exploit the multifunctional poten- tialities of these enzymes, awareness is needed of the limita- tions. Consequently, a critical evaluation of the nature of the resultant peptides is indispensable to avoid serious confusions.

1.

2. 3. 4.

5.

6.

7.

8.

9.

10. 11.

12.

13.

14.

15.

16.

17.

REFERENCES

Oka, T., and Morihara, K. (1977) J. Biochem. (Tokyo) 82, 1055-

Morihara, K., and Oka, T. (1977) Biochem. J. 163, 531-542 Pellegrini, A,, and Luisi, P. L. (1978) Biopolymers 17, 2573-2580 Isowa, Y., and Ichikawa, T. (1978) Bull. Chem. Soc. Jpn. 52, 796-

800 Inouye, S., Watanage, K., Morihara, K., Tochino, Y., Kanaya, T.,

Emura, J., and Sakakibara, S . (1979) J. Am. Chem. Soc. 101, 751-752

Kullmann, W. (1979) Biochem. Biophys. Res. Comrnun. 91, 693- 698

Homandberg, G. A,, and Laskowski, M., Jr. (1979) Biochemistry 18, 586-592

Kuhl, P., Konnecke, A,, Doring, G., Daumer, H., and Jakubke, H.-D. (1980) Tetrahedron Lett. 21,893-896

Morihara, K., Oka, T., Tsuzuki, H., Tochino, Y., and Kanaya, T . (1980) Biochem. Biophys. Res. Commun. 92,396-402

Kullmann, W. (1980) J. Biol. Chern. 255,8234-8238 Yonemitsu, O., Hamada, T., and Kanaoka, Y. (1969) Tetrahedron

Milne, H. B., and Carpenter, F. H. (1968) J. Org. Chem. 33,4476-

Milne, H. B., and Most, C . F., Jr. (1968) J . Org. Chem. 33, 169-

Schechter, I., and Berger, A. (1968) Biochem. Biophys. Res.

Janssen, F., Winitz, M., and Fox, S. W. (1953) J. Am. Chem. Soc.

Morihara, K., and Tsuzuki, H. (1970) Eur. J. Biochem. 15, 374-

Anderson, G., and Luisi, P. L. (1979) Helu. Chim. Acta 62, 488-

1062

Lett. 23, 1819-1820

4479

175

Commun. 32,898-902

75,704-707

380

494