Exploitation of subtilisin BPN' as catalyst for the synthesis of peptides containing noncoded amino acids, peptide mimetics and peptide conjugates

Article abstract of DOI:10.1021/ja964399w

The ability of the serine protease subtilisin BPN' to catalyze peptide bond formation between fragments containing noncoded amino acids, peptide mimetics, and peptide conjugates in a kinetic approach was explored. It was found that the enzyme accepts nume

Full text of DOI:10.1021/ja964399w

3942
J. Am. Chem. Soc. 1997, 119, 3942-3947
Exploitation of Subtilisin BPN′ as Catalyst for the Synthesis of
Peptides Containing Noncoded Amino Acids, Peptide Mimetics
and Peptide Conjugates
Wilna J. Moree, Pamela Sears, Katsuhiro Kawashiro, Krista Witte, and
Chi-Huey Wong*
Contribution from the Department of Chemistry and The Skaggs Institute of Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
ReceiVed December 23, 1996X
Abstract: The ability of the serine protease subtilisin BPN′ to catalyze peptide bond formation between fragments
containing noncoded amino acids, peptide mimetics, and peptide conjugates in a kinetic approach was explored. It
was found that the enzyme accepts numerous of these types of compounds both as acyl donor and acyl acceptor.
The results together with specificity studies reported by others provide an active site model as a guideline in the
design of enzymatic synthesis of biologically important compounds.
Introduction
acyl donor and various noncoded amino acids, peptide mimetics,
and peptide conjugates as acyl acceptors (Figure 1). (Although
N-acetylated amino acids are preferred by the enzyme, as
detailed in Table 6, Z-protected amino acids are more useful
for synthetic purposes as they afford easy deprotection.) We
expanded this by using several peptide conjugates as the acyl
donors with H-Gly-NH2 as the acyl acceptor.
Protease-catalyzed peptide bond formation has received a lot
of attention for a number of years.1 Enzymatic peptide coupling
is considered to be an attractive alternative for solution and solid
phase peptide synthesis, since a high degree of regio- and
stereoselectivity can be reached, the reaction conditions are mild,
minimal side protection is required, and couplings are generally
racemization free. However, a general methodology for enzy-
matic synthesis of peptides has not yet been achieved, the major
limitation being the hydrolytic activity of the protease and the
enzyme specificity which often limits the residues between
which a peptide bond can be synthesized.
Results and Discussion
Before coupling Z-Phe-OBn in the presence of subtilisin BPN′
to several noncoded amino acids we used a standard amino acid
(glycinamide 1) as a nucleophile (Table 1) for a point of
reference (80% yield). The noncoded (but naturally occurring)
â-alanine amide (2) as a P1′ residue was almost an equally good
substrate as 1 (70% yield). The R-branched amino acids 3 and
4, useful for restriction of the conformation of a peptide6 and
of biological importance,8 and 5, employed as a galactose
mimetic in the development of sialyl Lewis X mimetics,9
showed little peptide bond formation. In these cases mainly
competitive hydrolysis of the benzyl ester was observed.
However, when the R-branched amino acids were positioned
as P2′ residues with Gly as the P1′ residue, the desired peptides
17 and 18 were obtained in high yields (89% and 75%,
respectively), confirming the statement that the specificity of
the S2′ subsite is more flexible than of the S1′ subsite.5
Allylamine (8) and the aminoethanol (derivatives) (9-11) which
can be considered as amino acid precursors and have been
used in an approach toward the enzymatic synthesis of cy-
closporin A analogs,10 all gave good yields, although the
yields were considerably influenced by the O-protecting group
(Table 1).
Although numerous papers appeared on the enzymatic
synthesis of various peptides,1 only a few report the enzymatic
coupling of noncoded amino acids,2 peptide mimetics,3 and
peptide conjugates.4 Since peptide isosteres have become very
important for the development of protease inhibitors, new
efficient and environmentally friendly methods are required for
the synthesis of these types of compounds. We have investi-
gated whether enzymatic coupling, using the commercially
available protease subtilisin BPN′ in a kinetic approach, could
be potentially useful for this purpose. Subtilisin BPN′ and
several variants are well described5 and frequently applied to
peptide synthesis. This serine protease is very specific for
L-amino acids in the P1 position (acyl donor), whereas the P1′
residue (nucleophile) is more flexible.6 For this reason we
started our study of P1′ specificity, using Z-Phe-OBn as the
X Abstract published in AdVance ACS Abstracts, April 15, 1997.
(1) Wong, C.-H.; Whitesides, G. M. Enzymes in Synthetic Organic
Chemistry; Pergamon: Oxford, 1994; pp 41-130 and references cited
therein.
(2) (a) Widmer, F.; Breddam, K.; Johansen, J. T. Carlsberg Res.
Commun. 1981, 46, 97-106. (b) Cerovsky, V.; Jakubke, H.-D. Int. J.
Peptide Protein Res. 1994, 44, 466-471. (c) Margolin, A. L.; Tai, D.-F.;
Klibanov, A. M. J. Am. Chem. Soc. 1987, 109, 7885-7887.
(3) (a) Schuster, M.; Munoz, B.; Yuan, W.; Wong, C.-H. Tetrahedron
Lett. 1993, 34, 1247-1250. (b) Schuster, M.; Wang, P.; Paulson, J. C.;
Wong, C.-H. J. Am. Chem. Soc. 1994, 116, 1135-1136.
(4) Wong, C.-H.; Schuster, M.; Wang, P.; Sears, P. J. Am. Chem. Soc.
1993, 115, 5893-5901.
(5) (a) Sears, P.; Schuster, M.; Wang, P.; Witte, K.; Wong, C.-H. J. Am.
Chem. Soc. 1994, 116, 6521-6530. (b) Jackson, D. Y., Burnier, J.; Quan,
C.; Stanley, M.; Tom, J.; Wells, J. A. Science 1994, 266, 243.
(6) Nomenclature according to Schechter, I.; Berger, A. Biochem.
Biophys. Res. Commun. 1967, 27, 157-162.
(7) (a) Toniolo, C. Biopolymers 1989, 28, 247-257. (b) Santini, A.;
Barone, V.; Pedone, C.; Crisma, M.; Bonora, G. M.; Toniolo, C. Int. J.
Biol. Macro. 1988, 10, 292-299.
(8) E.g.: (a) Bodo, B.; Rebuffat, S.; El Hajji, M.; Davoust, D. J. Am.
Chem. Soc. 1985, 107, 6011-6017. (b) Rebuffat, S.; Goulard, C.; Bodo,
B. J. Chem. Soc., Perkin Trans. I 1995, 1849-1855.
(9) (a) Wu, S-H.; Shimazaki, M.; Lin, C.-C.; Qiao, L.; Moree, W. J.;
Weitz-Schmidt, G.; Wong, C.-H. Angew. Chem., Int. Ed. Engl. 1996, 35,
88-90. (b) Lin, C.-C.; Shimazaki, M.; Heck, M.-P.; Aoki, S.; Wang, R.;
Kimura, T.; Ritze`n, H.; Takayama, S.; Wu, S. H.; Weitz-Schmidt, G.; Wong,
C.-H. J. Am. Chem. Soc. 1996, 118, 6826-6840. (c) Cappi, M. W.; Moree,
W. J.; Qiao, L.; Marron, T. G.; Weitz-Schmidt, G.; Wong, C.-H. Bioorg.
Med. Chem., in press.
(10) Sears, P. Ph.D. Thesis, The Scripps Research Institute, 1996.
S0002-7863(96)04399-5 CCC: $14.00 © 1997 American Chemical Society

Exploitation of Subtilisin BPN′
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 3943
Table 2. Subtilisin BPN′ Catalyzed Coupling of Peptide
Fragments Containing the Statine-Type Isostere
Figure 1. Kinetic approach (aminolysis) to enzymatic amidation
catalyzed by subtilisin BPN′.
Table 1. Subtilisin BPN′ Catalyzed Coupling of Peptide
Fragments Containing Noncoded Amino Acids and Amino Acid
Precursors
a The isosteres (23-25) were obtained by coupling of NH3, MeNH2,
or PheNH2 to vinylacetic acid, followed by epoxidation, azide opening,
and reduction. Statine and AHPPA derivatives (26-28) were synthe-
sized by isopropylidene protection of Boc-Statine and Boc-AHPPA,
followed by standard peptide chemistry. b See Table 1.
product (Table 2). By changing the amide to a methylamide
(24), the yield of formed peptide (37) dropped dramatically
(30%). Replacing the amide with a Phe amide (25) gave again
upon enzymatic coupling a good yield (78%). Statine amide
(26) itself was not a good substrate, and only a trace amount of
the desired product was obtained. Placing statine or an analog
(4-amino-3-hydroxy-5-phenylpentanoic acid, AHPPA) between
the P2′ and P3′ residue with Gly as the P1′ residue resulted
in peptides 40 and 41 in good yields (74% and 70%, respec-
tively).
The hydroxyketo and diketo isosteres have been applied with
success in the development of HIV and FIV protease inhibi-
tors.13 The hydroxyketoisostere of Gly (35) placed between
P1′ and P2′ residue gave upon coupling to Z-Phe-OBn a
reasonable yield of 42 (46%), but the same isostere of Phe, alias
norstatine (36) did not give any of the desired product (43)
(Table 3). However, once this isostere was placed between the
P2′ and P3′ residue, with Gly as P1′ residue, a moderate yield
of the peptide (44) was obtained (36%). The hydroxyethylamine
isostere, one of the most frequently used isosteres for the
development of HIV protease inhibitors,14 is an interesting
nucleophile for enzymatic coupling, especially due to the
presence of the various functionalities. The isostere mimicking
Gly-Phe (38) gave a reasonable yield (Table 3), which was
lowered when the Phe at P2′ was replaced with the unfavorable
Pro (39). The yield dropped to 0% when the isostere of Gly
was replaced by the isostere of Phe (40), again reflecting the
preference of BPN′ for Gly over Phe as P1′ residue. However,
the same trend as described previously was observed. By
positioning this unit as P2′-P3′ with a Gly as the P1′ residue,
a All substrates were prepared using standard peptide chemistry,
employing the EDCI/HOBt and the mixed anhydride coupling methods
in combination with Boc or Z protection and removal by TFA or HCl
and hydrogenolysis, respectively.33 Hydoxylmethylserine amide (5) was
synthesized from serine according to Otani and Winitz.34 The O-
protected amino alcohols (10) and (11) were prepared by alkylation of
the corresponding Boc amino alcohol, followed by removal of the Boc
group. b The reaction was performed in DMF/water (8/3, v/v), using
0.18 M acyl donor and 0.55M acyl acceptor in the presence of BPN′
(9 mg/mL). Under the same conditions without addition of the protease
no hydrolysis nor aminolysis took place.
Statine and derivatives (23-28) have a long history as peptide
bond isosteres and have been frequently used in the development
of protease inhibitors of, for example, renin11 and HIV pro-
tease.12 Although these isosteres are potentially inhibitors of
subtilisin BPN′, we reasoned that some peptide bond formation
might occur if the isostere was placed at or especially after the
P1′ residue in the acyl acceptor. Indeed, the statine analog of
GlyGly (23) was a good acyl acceptor, yielding 83% of coupled
(11) E.g.: (a) Umezawa, H.; Aoyagi, T.; Morishima, H.; Matsuzaki, M.;
Hamada, M.; Takeuchi, T. J. Antibiotics 1970, 23, 259-262. (b) Boger, J.;
Lohr, N. S.; Ulm, E. H.; Poe, M.; Blaine, E. H.; Fanelli, G. M.; Lin, T.-Y.;
Payne, L. S.; Schorn, T. W.; LaMont, B. I.; Vassil, T. C.; Stabilito, I. I.;
Veber, D. F.; Rich, D. H.; Bopari, A. S. Nature 1983, 303, 81-84 (c) Boger,
J.; Payne, L. S.; Perlow, D. S.; Lohr, N. S.; Poe, M.; Blaine, E.H.; Ulm, E.
H.; Schorn, T. W.; LaMont, B. I.; Lin, T.-Y.; Kawai, M.; Rich, D. H.;
Veber, D. F. J. Med. Chem. 1985, 28, 1779-1790.
(12) E.g.: (a) Dreyer, G. B.; Metcalf, B. W.; Tomaszek, Jr., T. A.; Carr,
T. J.; Chandler, A. C. III, Hyland, L.; Fakhoury, S. A.; Magaard, V. W.;
Moore, M. L.; Strickler, J. E.; Debouck, C.; Meek, T. D. Proc. Natl. Acad.
Sci. U.S.A. 1989, 86, 9752-9756. (b) Sakurai, M.; Sugano, M.; Handa, H.;
Komai, T.; Yagi, R.; Nishigaki, T.; Yabe, Y. Chem. Pharm. Bull. 1993,
41, 1369-1377. (c) Scholz, D.; Billich, A.; Charpiot, B.; Ettmayer, P.; Lehr,
P.; Rosenwirth, B.; Schreiner, E.; Gstach, H. J. Med. Chem. 1994, 37, 3079-
3089. (d) Hui, K. Y.; Hermann, R. B.; Mannetta, J. V.; Gygi, T.; Angleton,
E.L. FEBS Lett. 1993, 327, 355-360. (d) Fehrentz, J. A.; Chromier, B.;
Bignon, E.; Venaud, S.; Chermann, J.C.; Nisato, D. Biochem. Biophys. Res.
Commun. 1992, 188, 873-878.
(13) (a) Munoz, B.; Giam, C.-Z.; Wong, C.-H. Bioorg. Med. Chem. 1994,
2, 1085-1090. (b) Slee, D. H.; Laslo, K. L.; Elder, J. H.; Ollmann, I. R.;
Gustchina, A.; Kervinen, J.; Zdanov, A.; Wlodawer, A.; Wong, C.-H. J.
Am. Chem. Soc. 1995, 117, 11867-11878. (c) Mimoto, T.; Imai, J.; Tanaka,
S.; Hattori, N.; Takahashi, O.; Kisanuki, S.; Nagano, Y.; Shintani, M.;
Hayashi, H.; Sakikawa, H.; Akaji, K.; Kiso, Y. Chem. Pharm. Bull. 1991,
39, 2465-2467.
(14) See, e.g.: (a) Rich, D. H.; Green, J.; Toth, M. V.; Marshall, G. R.;
Kent, S. B. H. J. Med. Chem. 1990, 33, 1285-1288. (b) Rich, D. H.; Sun,
C.-Q.; Vara Prasad, J. V. N.; Pathiasseril, A.; Toth, M. V.; Marshall, G.
R.; Clare, M.; Mueller, R. A.; Houseman, K. J. Med. Chem. 1991, 34, 1222-
1225. (c) Alewood, P. F.; Brinkworth, R. I.; Dancer, R. J.; Garnham, B.;
Jones, A.; Kent, S. B. H. Tetrahedron Lett. 1992, 33, 977-980. (d) Krohn,
A.; Redshaw, S.; Ritchie, J. C.; Graves, B. J.; Hatada, M. H. J. Med. Chem.
1991, 34, 3340-3342.

3944 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Moree et al.
Table 3. Subtilisin BPN′ Catalyzed Coupling of Peptide
Fragments Containing Peptide Bond Isosteres and Peptide Mimetics
couple phosphonamidate (51) enzymatically failed, with neither
aminolysis nor hydrolysis of the benzyl ester taking place. It is
not clear why this coupling failed, since glycine and phenyla-
lanine (the residues mimicked by this isostere) are preferred in
the S1′ and S2′ sites, respectively. The fact that no hydrolysis
is observed indicates that enzymatic activity has been inhibited,
so perhaps the phosphonamidate is lying at the position of the
scissile bond, despite the fact that glycine is not favorable in
the S1 site. In close resemblance to the phosphonamidate is
the sulfonamide transition-state isostere. Taurine18 and taurine
peptides contain a sulfonamide isostere and have often been
employed as potential pharmacophores for the development of
protease inhibitors.19,20 Coupling of two taurine containing
peptides (52) and (53) to the acyl donor gave the desired
products (57 and 58 respectively) in moderate yields (Table 4).
Simple peptide conjugates, such as p-nitroanilides (pNA),
have been used extensively as substrates for proteases in
chromogenic assays, though synthesis of these base-sensitive
substrates is not trivial. A previous enzymatic approach toward
the synthesis of a pNA derivative of a laminin fragment21
prompted us to investigate small pNA fragments (66) and (67)
as acyl acceptors (Table 5). Good yields of the corresponding
products 71 and 72 (64% and 69%, respectively) were obtained.
Protein conjugates containing glycosyl- and phosphoamino
acids are involved in various specific biological processes, and
in order to elucidate these processes, continued effort has been
devoted to find synthetic routes for these types of compounds.
Only a limited number of enzymatic methods have been
described, and coupling of peptide conjugates catalyzed by
subtilisin BPN′ has only been reported for N- and O-glycosy-
lamino acids and glycopeptide fragments.22 Tyrosine phospho-
rylation appears to be a major mechanism of cellular signal
transduction,23 and since there is still a lack of synthetic
approaches to these phosphoconjugates, it is of considerable
interest to develop enzymatic routes for them. Enzymatic
phosphorylation of tyrosine-containing peptides was achieved
by use of glutamine synthetase adenylyltransferase.24 We
exploited an alternative method to enzymatically couple phos-
phorylated Tyr containing peptides either as part of an acyl
acceptor to Z-Phe-OBn or as part of an acyl donor to glycina-
mide (1). The use of the diethyl ester of phosphotyrosine amide
(69) as a nucleophile gave a moderate yield (26%) of dipeptide
upon coupling. The ethyl ester was chosen as the protecting
group of the phosphate (69 and 70) since it is small enough to
fit in the hydrophobic S2′ subsite and will not be hydrolyzed
by subtilisin (Vide infra). Surprisingly, placing the Tyr dieth-
a The hydroxyketoisostere (35) was prepared by addition of TMSCN
to N-benyzloxycarbonyl glycinaldehyde, followed by hydrolysis under
basic conditions and deprotection. The hydroxyketoisosteres (36) and
(37) were prepared by aminolysis of (2R,3S)-N-(benzyloxycarbonyl)(3-
amino-2-hydroxy-4-phenylbutanoic acid) ethyl ester,13 followed by
standard peptide chemistry. The hydroxyethyleneamine isosteres (38)
and (39) were obtained by opening of the corresponding Z protected
epoxide with an amino acid followed by deprotection, whereas the
hydroxyethyleneamine isosteres (40) and (41) were synthesized by
opening of the epoxide derived from Z-Phe-OH with proline amide
followed by standard peptide chemistry. b See Table 1. c No enanti-
oselectivity was observed; the peptide was obtained as a diastereomeric
mixture in a ratio of approximately 1/1.
Table 4. Subtilisin BPN′ Catalyzed Coupling of Peptide
Fragments Containing Peptide Bond Isosteres
(15) Jacobsen, N. E.; Bartlett, P. A. J. Am. Chem. Soc. 1981, 103, 654-
657
(16) (a) Bartlett, P. A.; Marlowe, C. K. Science 1987, 235, 569-571.
(b) McLeod, D. A.; Brinkworth, R. I.; Ashley, J. A.; Janda, K. D.; Wirsching,
P. Bioorg. Med. Chem. Lett. 1991, 1, 653-658. (c) Camp, N. P.; Hawkins,
P. C. D.; Hitchcock, P. B.; Gani, D. Bioorg. Med. Chem. Lett. 1992, 2,
1047-1052.
a The phosphonates (49) and (50) and phosphonamidate (51) were
obtained starting from diethyl(phthaloylaminomethyl)phosphonate ac-
cording to the procedures described by Jacobsen and Bartlett.15 The
taurine derivatives (52) and (53) were synthesized using the methodol-
ogy described previously.20 b See Table 1.
(17) Janda, K. D.; Schloeder, D.; Benkovic, S. J.; Lerner, R. A. Science
1988, 241, 1188-1191.
(18) Taurine (Tau) is the abbreviation for aminoethane sulfonic acid.
(19) (a) Sturman, J. A. Physiol. ReV. 1993, 73, 119-148. (b) Sebestve´n,
F.; Furka, A.; Feuer, L.; Gulvas, J.; Szokan, G. Int. J. Peptide Protein Res.
1980, 16, 245-247.
the yield increased considerably (47%). Interestingly, both
isomers of the isostere substrates are accepted by the enzyme.
The dimethyl ester of amino methane phosphonic acid (49),
an analog of Gly, can be coupled to the acyl donor in 54%
yield without any hydrolysis of the phosphate ester (Table 4).
The results were even better when this phosphonic ester
containing mimetic was placed as the P2′ residue with Gly as
the P1′ residue (87%). The phosphonamidate is one of the
earliest described transition-state isosteres of the hydrolysis of
an amide bond15 and frequently applied to the development of
protease inhibitors16 and catalytic antibodies.17 An attempt to
(20) (a) Moree, W. J.; van Gent, L. C.; van der Marel, G. A.; Liskamp,
R. M. J. Tetrahedron 1993, 49, 1133-1150. (b) Moree, W. J.; van der
Marel, G. A.; Liskamp, R. M. J. J. Org. Chem. 1995, 60, 5157-5169.
(21) Ternt’eva, E. Y.; Voyushina, T. L.; Stepanov, V. M. Bioorg. Med.
Chem. 1995, 5, 2523-2526.
(22) Wong, C.-H.; Schuster, M.; Wang, P.; Sears, P. J. Am. Chem. Soc.
1993, 115, 5893-5901.
(23) (a) Boyer, P. D.; Krebs, E. G. The Enzymes; 1986; Vol. XVII, pp
192-237. (b) Sugimoto, Y.; Whitman, M.; Cantley, L. C.; Erikson, R. L.
Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 2117-2121.
(24) Gibson, B. W.; Hines, W.; Yu, Z.; Kenyon, G. L.; McNemar, L.;
Villafranca, J. J. J. Am. Chem. Soc. 1990, 112, 8523-8528.

Exploitation of Subtilisin BPN′
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 3945
Table 5. Subtilisin BPN′ Catalyzed Coupling of Peptide Fragments Containing Peptide Conjugates
acyl donor
Z-Phe-OBn
Z-Phe-OBn
Z-Phe-OBn
Z-Phe-OBn
acyl acceptora
productb
yield (%)
HCl.H-Gly-pNA (66)
HCl.H-Gly-Phe-pNA (67)
TFA. H-Gly-NHNHZ (68)
H-Tyr(PO3Et2)NH2 (69)
H-Gly-Tyr(PO3Et2)NH2 (70)
HCl.H-Gly-NH2
HCl.H-Gly-NH2
HCl.H-Gly-NH2
HCl.H-Gly-NH2
HCl.H-Gly-NH2
Z-Phe-Gly-pNA (71)
Z-Phe-Gly-Phe-pNA (72)
Z-Phe-Gly-NHNHZ (73)
64
69
60
26
28
0
26
79
0
Z-Phe-Tyr(PO3Et2)NH2 (74)
Z-Phe-Gly-Tyr(PO3Et2)NH2 (75)
Z-Tyr(PO3H2)-Gly-NH2 (76)
Z-Tyr(PO3Me2)-Gly-NH2 (77)
Z-Tyr(PO3Et2)-Gly-NH2 (78)
Z-Tyr(PO3Bn2)-Gly-NH2 (79)
Z-Tyr(PO3tBu2)-Gly-NH2 (80)
Z-Tyr(PO3tBu2)-Ala-Gly-NH2 (81)
Z-Tyr(PO3Et2)-Ala-Gly-NH2 (82)
Z-Phe-OBn
Z-Tyr(PO3H2)ONb (59)
Z-Tyr(PO3Me2)OMe (60)
Z-Tyr(PO3Et2)OMe (61)
Z-Tyr(PO3Bn2)ONb (62)
Z-Tyr(PO3tBu2)OMe (63)
Z-Tyr(PO3tBu2)-Ala-OMe (64)
Z-Tyr(PO3Et2)-Ala-OMe (65)
0
17
40
HCl.H-Gly-NH2
HCl.H-Gly-NH2
a The p-nitroanilide containing fragments (66) and (67) were prepared using standard peptide chemistry; the hydrazine (68) was purchased from
Bachem. The phosphate methyl esters and phosphate ethyl esters of the tyrosine derivatives (69, 70 and 60, 61, and 65) were obtained by
phosphorylation using the phosphite/iodine method,35 whereas the tert-butyl and benzyl esters (62, 63, and 64) and the phosphate 59 were prepared
using phosphonamidite followed by oxidation.36 b See Table 1.
ylphosphate ester in the P2′ position, hardly improved the yield
(28%) (Table 5).
Since Tyr is favorable in the P1 position, we envisioned that
phosphorylated Tyr might be accepted as an acyl donor.
Various phosphate esters (as shown in Table 5) were tested for
coupling to glycinamide (1). The benzyl ester (62) and tert-
butyl ester (63) were not accepted as substrates. These groups
are probably too large for the hydrophobic pocket. The
unprotected phosphate (59) was not a substrate either. The
charge on the phosphate most likely causes unfavorable interac-
tions with the Glu 156 residue in the S1 subsite. The methyl
ester (60) and the ethyl ester (61) however, were very good
substrates, giving rise to coupling yields of 26% and 79%,
respectively. The low yield of the methyl ester (77) was due
to hydrolysis of the phosphate methyl ester, which under the
Figure 2. Connolly surface of the active site of subtilisin BPN′
coupling conditions did not occur for the ethyl ester. Finally,
when we placed the tert-butyl protected phosphorylated Tyr in
the P2 position (64), some peptide formation took place (17%),
indicating that the preference of the P2 position is more flexible.
The results were slightly better with the phosphate ethyl ester
(65, 40%), although at this position the hydrolysis of the methyl
ester was very fast.
containing a portion of the chymotrypsin inhibitor 2 and showing the
positions of four subtsites (S1, S2, S1′, and S2′). Amino acids are color-
coded by type: acids, red; bases, blue; polar residues, green; and
hydrophobic residues, white. The chymotrypsin inhibitor is shown in
yellow. The black arrow shows the position of the amide bond
corresponding to the scissile bond of a subtilisin substrate.
of the substrate.25 Only a portion of the inhibitor is shown, in
yellow. Both steric and electrostatic factors are indicated to
provide a general picture about how substrates are positioned
in the active site. Figure 3 indicates the active site binding
model based on the results of this and previous studies,26 and
Table 6 tabulates the trends compiled from the available kinetic
data regarding the subsite preferences of subtilisin BPN′. When
planning a kinetically controlled synthesis in large concentrations
of cosolvent, however, one should treat the data in Table 6 with
care. The trends in the table, other than those directly related
to this work, were taken from studies performed in buffer (with
no cosolvent), since these were most amenable to comparison:
actual kinetic constants were available and typically were
obtained under similar conditions of ionic strength and pH. The
crystal structure of subtilisin 8397/Lys256fTyr (a stabilized
BPN′ variant, which shows only minor changes in structure
when compared to BPN′ 27a) in 50% DMF27b demonstrates that
addition of solvent to 50% has little change on the structure of
the subsites; one should bear in mind, though, that the addition
of cosolvent may have some effects on substrate partitioning
Conclusions
The results of this study, based on a limited number of
noncoded amino acids, peptide isosteres, and peptide conjugates
chosen as potential substrates for subtilisin BPN′ indicate that
the enzyme accepts numerous unnatural units both as acyl
donors and acyl acceptors. Preference was observed for small
amine fragments closely related to Gly, positioned as the P1′
residue. Isosteres mimicking the Gly-XX bond were generally
well accepted as the P1′-P2′ residues, giving moderate to good
yields of coupled product, depending on XX. Isosteres mimick-
ing the Phe-XX bond as the P1′-P2′ residues gave only trace
amounts of products. The yields of the latter considerably
improved, once placed as the P2′-P3′ residues with a Gly as
P1′ residue. Furthermore it was found that phosphorylated Tyr
was a good substrate both as the P1 and P2 residue if the
phosphate was protected as a diethyl ester.
To gain further insights into the specificity of subtilisin BPN′
in order to provide a general guideline for the design of
substrates, we consider the results of this study together with
the great number of specificity studies reported previously.
Figure 2 illustrates the active site complexed with potent
inhibitor (chymotrypsin inhibitor 2) which interacts with the
S4-S4′ region of the enzyme in a manner similar to the binding
(25) Eder, J.; Rheinnecker, M.; Fersht, A. R. FEBS Lett. 1993, 335, 349.
(26) The model is an extension of the model proposed by J. Bryan Jones
et al. outlined in the following: Bonneau, P. R.; Graycar, T. P.; Estell, D.
A.; Jones, J. B. J. Am. Chem. Soc. 1991, 113, 1026.
(27) (a) Kidd, R. D.; Yennawar, H. P; Sears, P.; Wong, C.-H.; Farber,
G. K. J. Am. Chem. Soc. 1996, 118, 1645-1650. (b) Kidd et al., unpublished
results.

3946 J. Am. Chem. Soc., Vol. 119, No. 17, 1997
Moree et al.
natural components are properly positioned. Further improve-
ment perhaps can be achieved with the use of subtilisin variants
designed to prevent hydrolysis, to perform reactions in DMF,
or to have altered substrate specificity.
Experimental Section
General Methods. Subtilisin BPN′ was obtained from Sigma
(Protease Type XXVII, 7.8 units/mg solid). All protected amino
acids as well as (3S,4S)-Boc-4-amino-3-hydroxy-6-methylhep-
tanoic acid (Boc-Statine or Boc-AHMHA) were obtained from
Bachem. (3S,4S)-4-Amino-3-hydroxy-5-phenylpentanoic acid
(AHPPA) was purchased form Novabiochem. Isobutylchloro-
formate was distilled under Ar atmosphere. 1H NMR spectra
were recorded on a Bruker AC 250 (250 MHz) or AMX 400
(400 MHz) spectrometer, and chemical shifts are given in ppm
(δ) relative to TMS as internal standard. 13C NMR spectra were
measured with a Bruker AC 250 (62 MHz) or AMX 400 (100
MHz) spectrometer and chemical shifts are given relative to
CDCl3 or MeOD as internal standard. 31P NMR spectra were
obtained with a Brucker AMX 400 spectrometer using 85% H3-
PO4 as an external standard.
Thin layer chromatography analysis was performed on
precoated Merck Silica gel (60 F-254) plates (0.25 mm); the
solvent system used CH2Cl2/MeOH (9/1, v/v) unless indicated
otherwise. Spots were visualized with UV light or ninhydrin
(after treatment with HCl). Column chromatography was
carried out on Merck Kieselgel 60 (230-400 Mesh, ASTM).
High resolution mass spectra (HRMS) were recorded on a VG
ZAB-ZSE instrument under fast atom bombardment (FAB)
conditions.
General Procedure for Enzymatic Peptide Coupling. To
a solution of Z-Phe-OBn (40 mg, 0.10 mmol) in DMF (0.4 mL)
and water (0.15 mL), a free amine (0.30 mmol) or the
hydrochloride of an amine (0.30 mmol) in combination with
Et3N (0.30 mmol) were added, followed by subtilisin BPN′ (5
mg). The mixture was stirred at room temperature until the
benzyl ester disappeared (as indicated by TLC). The mixture
was concentrated in Vacuo, taken up in EtOAC, and washed
with 1 N KHSO4, 1 M Na2CO3 (3×), and brine. The organic
layers were dried (MgSO4), filtered, and concentrated in Vacuo.
Silica gel column chromatography afforded the coupling
product.
General Procedure for Aminolysis of a Methyl Ester. The
methyl ester was dissolved in anhydrous MeOH and cooled to
0 °C. NH3(g) was bubbled through until the solution was
saturated. The mixture was stirred at room temperature until
TLC indicated disappearance of the methyl ester.
Figure 3. Schematic drawing of the general features of the subtilisin
active site. The large S1 pocket is primarily hydrophobic but contains
a carboxylic acid (Glu 156) at the bottom, which explains its affinity
for basic and polar residues but its poor affinity for acids, including
unprotected phosphotyrosine. The S2 site is a somewhat smaller cleft
lying behind Asp 99 (the red bulge in the lower left corner of Figure
1). It is lined with hydrophobic and acidic residues, explaining the
poorness of subtilisin at cleaving peptides with acids in the P2 position
(unprotected phosphotyrosine is not accepted). It accepts proline quite
well, presumably since the substrate naturally “kinks” at this site
anyway, and though it prefers small hydrophobes, it can also accept
very large residues, such as butyl-protected phosphotyrosine, to a certain
extent. The S3 site is not much of a pocket; the substrate side chain
projects out of the active site. As a result, the specificity is broad. The
narrow “neck” of the subsite probably precludes bulky groups at the â
position and may explain the L-stereoselectivity at this site. The S4
site is an extremely large hydrophobic pocket. The S1′ site is narrow
and hydrophobic, backed by Met222, and capped by Tyr217. Small
residues are preferred, and R-branched amino acids are accepted poorly.
Isosteres are accepted in this site as long as the mimicked residue is
small. (Note that in Figures 1 and 2, the P3′ residue of the chymotrypsin
inhibitor 2 has bent around to sit primarily in the S1′ pocket, and the
P1′ residue, a glutamic acid, has bent out of the plane of the page.)
There is strong stereoselectivity at this site, and R-branched amino acids
are poorly acccepted, as are bulky side chains. Although subtilisin can
acetylate sugars and nucleotides, it is not clear how they are binding.
The S1′ site clearly cannot hold them entirely. The S2′ site is not so
much a cleft as a hydrophobic surface, flanked by phenylalanine. It
prefers large hydrophobes and can accept a variety of isosteres, as long
as the P1′ residue is small (esp. glycine).
into the active site. In general, the longer the peptide chain
that is used as a substrate, the better. Addition of an extra
glycine to AcGlyTyrOMe to form the tripeptide, AcGlyGly-
TyrOMe, for example, improves the rate of hydrolysis by 50-
100-fold,28 and similar results have been obtained with other
amino acids at nearly every position.29,30 The specificity of
subtilisin can often be broadened by using activated esters, such
as p-nitrophenyl esters, which do not appear to require both
base and then acid catalysis (only base catalysis) and thus can
have “floppier” transition states.31 Finally, substrates are
accepted best when protected both at the N- and C-termini. This
is true from the S4 site to at least the S3′32 site.
General Procedure for Removal of the Z Group via
Hydrogenation. The Z protected compound was dissolved in
a mixture of tBuOH/water (4/1, v/v) and a catalytic amount of
10% Pd on C was added. The mixture was stirred at 1 atm of
H2 until the Z group was cleaved, filtered over Celite, and
concentrated in Vacuo to afford the free amine.
(32) Morihara, K.; Tsuzuki, H.; Oka, T. Biochem. Biophys. Res. Commun.
1971, 42, 1000.
(33) Bodanszky, M.; Bodanszky, A. The practice of Peptide Chemistry
- ReactiVity and Structure. Concepts in Organic Chemistry; Hafner, K.,
Rees, C. W., Trost, B. M., Lehn, J.-M., Von Rague Schleyer, P., Zahradnik,
R., Eds.; Springer-Verlag: 1984.
In general, the results suggest that subtilisin BPN′ is useful
for the synthesis of peptide isosteres and analogs if the non-
(34) Otani, T. T.; Winitz, M. Arch. Biochem. Biophys. 1960, 90, 254-
259.
(28) Karasaki, Y.; Ohno, M. J. Biochem. 1979, 86, 563-567.
(29) Morihara, K.; Oka, T.; Tsuzuki, H. Arch. Biochem. Biophys. 1974,
165, 72-79.
(35) Stowell, J. K.; Widlanski, T. S. Tetrahedron Lett. 1995, 36, 1825-
1826.
(36) (a) Perich, J. W.; Johns, R. B. Tetrahedron Lett. 1988, 29, 2369-
2372. (b) Perich, J. W.; Johns, R. B. J. Org. Chem. 1989, 54, 1750-1752.
(c) Perich, J. W.; Alewood, P. F.; Johns, R. B. Aust. J. Chem. 1991, 44,
253-263.
(30) Bromme, D.; Peters, K.; Fink, S.; Fittkau, S. Arch. Biochem. Biophys.
1986, 244, 439.
(31) Polgar, L.; Fejes, J. Eur. J. Biochem. 1979, 102, 531.

Exploitation of Subtilisin BPN′
J. Am. Chem. Soc., Vol. 119, No. 17, 1997 3947
Table 6. Subsite Preferences of Subtilisin BPN′
subsite
preference
S4
S3
•broad specificity; generally prefers hydrophobic residuesa
•has a strong preference for L- vs D-amino acidsb,c
•prefers Bz over CBzd
•broad range of accepted residuesa
•D- and L-leucine not differentiated very welld but L-alanine preferred over D-Alab,e
•order of preference: Phe > Ala > Gly > D-Alab,c
•Asn(â-GlcNAc) and even some disaccharide moieties acceptedf
•accepts ethyl-protected phosphotyrosine and even butyl-protected but less efficiently (Table 5)
•prefers small hydrophobes
S2
•L-Ala strongly preferred over D-Alag
•acetyl > CBz > benzoyle,h
•Ala > Pro, Val > Leu > Gly > His, Tyr > D-Alac,d,e
•Ala > Pro . Lys > Arg > Aspi
•Gly > Leu with Phe in P1 sitej
•both unprotected and peracetylated, O- and N-linked glycosyl amino acids acceptedf
•accepts ethyl-, methyl-protected phosphotyrosine but not butyl- or benzyl-(Table 5)
•prefers large hydrophobes with polar or positively charged group at the endk
•Small amino acids such as glycine are much poorer than the aromatics. One should avoid Gly, Pro, Thr, Val, Ile.a
•Very large hydrophobic residues can be accepted, including dansylated aminobenzene derivatives,l γ-diphenyl residues,m
and various polycyclic aromatics.n
S1
•Tyr > Phe, Trp, Met > Lys > Argi,j,o
•Proline not accepted,a but other R-branched (locked) substrates are accepted.m
•His > Glup
•vinyl acetate accepted; used for acetylating a variety of substrates, including 6′-hydroxyls of sugarsq and 5′-hydroxyls of
nucleosidesr
•can accept peracetylated or nonprotected O-linked xylosyl amino acids but not Asn(â-GlcNAc)f
•A variety of peptide isosteres are accepted (statine-type, hydroxyketo, hydroxyethylamine, phosphonate, and sulfonamide)
if the residue mimicked is small (glycine), not large and hydrophobic (leucine, phenylalanine).
•very small, narrow, restricted site; avoid R-branching (Table 1), though paradoxically will accept (ethyl) protected
phosphotyrosine to a moderate degree (Table 5)
S1′
•avoid Ile, Pro, Asp, Glua
•Gly, Ala . Leu . Pros
•Gly > Ser, Lys, Ala > Arg > Gln, Thr, Val, Met, Asn > His > Ile, Trp > Glu > Asp, Leu . Prou
•can accept D-Leu, D-Arg to some extentV
•does not accept glycosyl amino acidsf
S2′
S3′
•methylated N not favorable between P1′ and P2′ positions (Table 2, 24)
•accepts (ethyl) protected phosphotyrosine to a moderate extent (Table 5)
•most peptide isosteres accepted, if residue in P1′ position is small (glycine)
•Leu > Gly, D-Leu > Pros
•poor differentiation of D-Ala from L-Alae
•Phe > Gly, Ala > D-Alac
•likes large hydrophobics; avoid Pro, Asp, Glu, Glya
•deprotected N- and O-linked glycosyl amino acids acceptedf
•minimal differentiation of D- and L-Alac
a Abrahmsen, L.; Tom, J.; Burnier, J.; Butcher, K.; Kossiakoff, A.; Wells, J. A. Biochemistry 1991, 30, 4151. b Morihara, K.; Tsuzuki, H.; Oka,
T. Biochem. Biophys. Res. Commun. 1971, 42, 1000. c Morihara, K.; Oka, T.; Tsuzuki, H. Arch. Biochem. Biophys. 1970, 138, 515. d Pozsgay, M.;
Gaspar, R.; Elodi, P.; Bajusz, S. FEBS Lett. 1977, 74, 67-70. e Morihara, K.; Oka, T.; Tsuzuki, H. Biochem. Biophys. Res. Commun. 1969, 35,
210. f Wong, C.-H.; Schuster, M.; Wang, P.; Sears, P. J. Am. Chem. Soc. 1993, 115, 5893. g Morihara, K.; Oka, T.; Tsuzuki, H. Arch. Biochem.
Biophys. 1974, 165, 72-79. h Philipp, M.; Polgar, L. Molec. Cell. Biochemistry 1983, 51, 5. i Ballinger, M. D.; Tom, J.; Wells, J. A. Biochemistry
1995, 34, 13312. j Powers, J. C.; Lively, M. O.; Tippett, J. T. Biochim. Biophys. Acta 1977, 480, 246-261. k Karasaki, Y.; Ohno, M. J. Biochemistry
1978, 84, 531-538. l Philipp, M; Maripuri, S. FEBS Lett. 1981, 133, 36-38. m Pattabiraman, T. N.; Lawson, W. B. Biochem. J. 1972, 126, 659-
665. n Bosshard, H. R.; Berger, A. Biochemistry 1974, 13, 266-277. o Carter, P.; Nilsson, B.; Burnier, J. P.; Burdick, D.; Wells, J. A. Proteins
1989, 6, 240. p Carter, P.; Abrahmsen, L.; Wells, J. A. Biochemistry 1991, 30, 6142. q Wong, C.-H.; Chen, S.-T.; Hennen, W. J.; Bibbs, J. A.;
Wang, Y.-F.; Liu, J. L-C.; Pantoliano, M. W.; Whitlow, M.; Bryan, P. N. J. Am. Chem. Soc. 1990, 112, 945. r Zhong, Z.; Liu, J. L.-C.; Dinterman,
L. M.; Finkelman, M. A. J.; Mueller, W. T.; Rollence, M. L.; Whitlow, M.; Wong, C.-H. J. Am. Chem. Soc. 1991, 113, 683. s Morihara, K.;
Tatsushi, O. Arch. Biochem. Biophys. 1977, 178, 188-194. t Barel, A. O.; Glazer, A. N. J. Biol. Chem. 1968, 243, 1344-1348. u Cerovsky, V.;
Jakubke, H.-D. Biotech. Bioeng. 1996, 49, 553. V Wong, C.-H.; Chen, S.-T.; Hennen, W. J.; Bibbs, J. A.; Wang, Y.-F.; Liu, J. L-C.; Pantoliano, M.
W.; Whitlow, M.; Bryan, P. N. J. Am. Chem. Soc. 1990, 112, 945.
General Procedure for EDCI/HOBt Peptide Coupling.
The carboxylic acid (1 equiv) and the amine (1.05 equiv) were
dissolved in dry CH2Cl2 (if necessary in combination with
DMF), and HOBt (1.05 equiv) was added. The mixture was
cooled to 0 °C, EDCI (1.1 equiv) was added, and the apparent
pH (as indicated by wet pH paper) of the mixture was adjusted
to 8 with NMM. After stirring for 0.5 h to 1 h at 0 °C, stirring
was continued at room temperature until starting materials were
gone. The mixture was diluted with CH2Cl2, washed with 1 N
KHSO4 (1×), 1 M Na2CO3 (1-3×), brine, dried (MgSO4),
filtered, and concentrated in Vacuo. The coupling product was
purified by silica gel column chromatography.
Acknowledgment. W.J.M. gratefully acknowledges support
from a NATO-Science fellowship and we thank Dr. D. H. Slee
for supplying the precursor molecule for the synthesis of 36
and 37, K. L. Laslo for the precursor molecule of 40 and 41,
and Dr. R. M. J. Liskamp for the precursors of 52 and 53.
Supporting Information Available: 1H NMR spectra for
compounds including the substrates and the products and
detailed experimental on the synthesis of the individual com-
pounds (49 pages). See any current masthead page for ordering
and Internet access instructions.
JA964399W