Substrate mimetics-specific peptide ligases: studies on the synthetic utility of a zymogen and zymogen-like enzymes.

Article abstract of DOI:10.1021/jo026117i

Although proteases are capable of synthesizing peptide bonds, they are not proficient at peptide fragment ligation. Further manipulations are needed to shift the native enzyme activity from the cleavage to the synthesis of peptides. This account reports on the synthetic potential of nonactivatable trypsinogen and zymogen-like enzymes designed to minimize proteolytic side reactions during peptide synthesis.

Full text of DOI:10.1021/jo026117i

site. Proteases used as the ligation catalyst tolerate a
higher structural diversity at this position, although the
substrate specificity of these enzymes primarily narrows
the scope of these enzymes for synthesis.4 Substantial
improvements have been attained by the development
of substrate mimetics used as donor components that
eliminate the specificity problem.5 Nonspecificity toward
the peptide sequence of substrate mimetics is reached
by attaching a site-specific ester leaving group at the
C-terminus of the originally nonspecific peptide moiety.
This concept achieves protease-mediated peptide coupling
at nonspecific and even artificial ligation sites, regardless
of the specificity of the enzyme.6 Moreover, as nonspecific
sequences are coupled, the newly formed peptide bond
is stable against secondary cleavage. On the other hand,
the risk of undesired cleavages of sensitive peptide bonds
by the protease remains a serious limitation of this
approach.7
Su bstr a te Mim etics-Sp ecific P ep tid e
Liga ses: Stu d ies on th e Syn th etic Utility of
a Zym ogen a n d Zym ogen -Lik e En zym es
Kathrin Rall‡ and Frank Bordusa*,†,‡
Research Unit “Enzymology of Protein Folding”,
Max-Planck-Society, Weinbergweg 22,
D-06120 Halle/ Saale, Germany, and
Department of Biochemistry, University of
Leipzig, D-04103 Leipzig, Germany
bordusa@enzyme-halle.mpg.de
Received J une 28, 2002
Abstr a ct: Although proteases are capable of synthesizing
peptide bonds, they are not proficient at peptide fragment
ligation. Further manipulations are needed to shift the
native enzyme activity from the cleavage to the synthesis
of peptides. This account reports on the synthetic potential
of nonactivatable trypsinogen and zymogen-like enzymes
designed to minimize proteolytic side reactions during
peptide synthesis.
This paper reports on the synthetic utility of a nonac-
tivatable zymogen, i.e., mutant trypsinogen K15A, for the
substrate mimetics-mediated coupling of specific amino
acid-containing peptides. The study has been expanded
to a zymogen-like enzyme species (mutant trypsin D194N)
with a partly destabilized activation domain. Further
optimization of the biocatalyst yielded the mutant tryp-
sins D189S,D194N and K60E,D189S,D194N.
Chemical synthesis is the most powerful method for
assembling selectively modified proteins, providing a
freedom for protein engineering inaccessible by standard
site-directed mutagenesis. Presently, there are several
feasible strategies of chemical protein synthesis, each
with its own individual advantages and disadvantages.
Step-by-step solid-phase synthesis, although reaching a
high degree of sophistication, is restricted by the cumula-
tive effects of synthetic inefficiencies. This inevitably
results in the accumulation of low-level resin-bound by-
products that usually limits step-by-step synthesis to the
preparation of peptides of about 50 amino acids.1 Con-
vergent synthetic approaches, which apply protected pep-
tide fragments as reactants, allow for several proteins
to be assembled, but they have been mainly handicapped
by the insolubility of larger protected peptides in solvents
required for purification or subsequent ligation.2 To solve
this problem, several chemoselective ligation strategies
have been developed enabling unprotected peptides to be
coupled. Presently, “native chemical ligation” is the most
successful and most commonly used synthetic ligation
approach.3 Because of a mechanism that is based on the
reaction of a peptide thioester with an acceptor peptide
bearing an N-terminal Cys, this strategy, however, is
highly limited in the choice of residues at the ligation
Originally, zymogens were known as catalytically
inactive precursors of active enzymes. Activation of the
precursor enzyme to the active species is usually a result
of limited proteolytic cleavage. In the case of trypsinogen/
trypsin conversion, the peptide bond between Lys15 and
Ile16 of the zymogen becomes specifically cleaved either
by enterokinase or autocatalytically by trypsin itself.
Thus, the exchange of Lys15 with Ala protects trypsino-
gen against activation. Most zymogens have practically
no or only negligible proteolytic activity. In contrast,
activity vis-a`-vis ester bonds, although reduced signifi-
cantly, could be detected.8 Decrease in the zymogen’s
esterase activity was found to be mainly the result of
(4) (a) Schellenberger, V.; J akubke, H.-D. Angew. Chem., Int. Ed.
Engl. 1991, 30, 1437-1449. (b) Wong, C.-H. Science 1989, 244, 1145-
1152. (c) J ackson, D. Y.; Burnier, J .; Quan, C.; Stanley, M.; Tom, J .;
Wells, J . A. Science 1994, 266, 243-247.
(5) (a) Mitin, Y. V.; Schellenberger, V.; Schellenberger, U.; J akubke,
H.-D.; Zapevalova, N. P. In Peptides; Giralt, E.; Andreu, D., Eds.;
ESCOM: Leiden, 1990, 287-288. (b) Sekizaki, H.; Itoh, K.; Toyota,
E.; Tanizawa, K. Chem. Pharm. Bull. 1996, 44, 1577-1579. (c)
Bordusa, F.; Ullmann, D.; Elsner, C.; J akubke, H.-D. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 2473-2475. (d) Thormann, M.; Thust, S.;
Hofmann, H.-J .; Bordusa, F. Biochemistry 1999, 38, 6056-6062.
(6) (a) Sekizaki, H.; Itoh, K.; Toyota, E.; Tanizawa, K. Chem. Pharm.
Bull. 1998, 46, 846-849. (b) Wehofsky, N.; Bordusa, F. FEBS Lett.
1999, 443, 220-224. (c) Gu¨nther, R.; Stein, A.; Bordusa, F. J . Org.
Chem. In press. (d) Gu¨nther, R.; Bordusa, F. Chem. Eur. J . 2000, 6,
463-467. (e) Gu¨nther, R.; Thust, S.; Hofmann, H.-J .; Bordusa, F. Eur.
J . Biochem. 2000, 267, 3496-3501. (f) Gu¨nther, R.; Stein, A.; Bordusa,
F. J . Org. Chem. 2000, 65, 1672-1679.
(7) (a) Wehofsky, N.; Kirbach, S. W.; Haensler, M.; Wissmann, J .-
D.; Bordusa, F. Org. Lett. 2000, 2, 2027-2030. (b) Xu, S.; Rall, K.;
Bordusa, F. J . Org. Chem. 2001, 66, 1627-1632. (c) Gru¨nberg, R.;
Domgall, I.; Gu¨nther, R.; Rall, K.; Hofmann, H.-J .; Bordusa, F. Eur.
J . Biochem. 2000, 267, 7024-7230.
(8) (a) Gertler, A.; Walsh, K. A.; Neurath, H. Biochemistry, 1973,
13, 1302-1307. (b) Ullmann, D.; Salchert, K.; Bordusa, F.; Schaaf, R.;
J akubke, H.-D. In Proceedings of the 5th Akabori Conference; Wu¨nsch,
E., Ed.; R. & J . Blank: Munich, 1994; pp 70-75.
* To whom correspondence should be addressed. Tel: +49 345
5522806. Fax: +49 345 5511972.
† Max-Planck-Society.
‡ University of Leipzig.
(1) (a) Bodansky, M.; Bodansky, A. The Practice of Peptide Synthesis;
Springer-Verlag: New York, 1984. (b) Kaiser, E. T. Acc. Chem. Res.
1989, 22, 957.
(2) see review: Borgia, J . A.; Fields, G. B. TIBTECH. 2000, 18, 243-
251.
(3) (a) Kemp, D. S.; Grattan, J . A.; Reczek, J . J . Org. Chem. 1975,
40, 3465-3466. (b) Fotouhi, N.; Galakatos, N. G.; Kemp, D. S. J . Org.
Chem. 1989, 54, 2803-2809. (c) Schnalzer, M.; Kent, S. B. H. Science
1992, 256, 221-225. (d) Dawson, P. E.; Muir, T. M.; Clark-Lewis, I.;
Kent, S. B. H. Science 1994, 266, 776-779. (e) Tam, J . P.; Liu, C.-F.;
Shao, J . Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 12485-12489.
10.1021/jo026117i CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/20/2002
J . Org. Chem. 2002, 67, 9103-9106
9103

SCHEME 1. Str u ctu r es of Su bstr a te Mim eticsa
a
(1) 4-Guanidinophenyl (OGp), (2) 4-carboxamidophenyl (OCap),
(3) 4-carboxyphenyl (OCp), (4) Phenyl (OPh), (5) 4-hydroxythio-
phenyl (SPhOH), (6) picolyl (OPic), R ) Boc-alanyl.
reduced acylation rates, while the rates of deacylation
of the zymogen and the enzyme appear to be very similar.
Thus, improvements in the acylation rate could make
zymogens interesting peptide ligases without any pro-
teolytic activity. This, together with the discovery that
substrate mimetics of the 4-guanidinophenyl ester (OGp)
type act as highly efficient acylating agents for active
trypsin,5d prompted us to study the behavior of trypsi-
nogen toward these donor components. Taking into
account the differences in the substrate binding sites and,
hence, in the specificities of trypsinogen and trypsin,9
several other substrate mimetics have been included in
the study (Scheme 1). The acceptance of the esters by
the zymogen has been checked by hydrolysis studies and
peptide synthesis reactions as well. Unexpectedly, in
neither case could a conversion of the esters be detected,
indicating that trypsinogen does not accept substrate
mimetics as donor components.
To improve the catalytic power of trypsinogen, we
exchanged Asp194 with Asn keeping Lys15 intact. The
side chain carboxylate of Asp194 forms originally a salt
bridge to the NR-amino moiety of Ile16, which triggers a
conformational change in the activation domain and
creates the fully active enzyme.10 Thus, exchange of
Asp194 with Asn prevents the formation of the salt bridge
leading to a trypsinogen-like enzyme structure. Studies
in Hedstrom’s laboratory showed that this mutation
causes a ∼100-fold decrease in the specificity for cleaving
specific Arg/Lys-containing peptides, while the specificity
for hydrolyzing analogous ester bonds was decreased only
by a factor of 10.11 Hydrolysis studies with the substrate
mimetics listed in Scheme 1 revealed a well-defined
preference of trypsin D194N toward the OGp ester, while
the remaining analogues showed similar but ∼10-fold
decreased hydrolysis rates (cf. Supporting Information,
Figure S1). Compared to wild-type (wt) trypsin, a de-
crease by a factor of ∼60 is evident.5d Peptide synthesis
reactions using Bz-Gly-OGp as the donor component and
Ala-Ala-(Lys/Arg)-Ala-Gly as the acceptor peptides dis-
played a pronounced cleavage activity toward the specific
Lys and Arg moieties, a situation similar to that found
for wt-trypsin7 (cf. Supporting Information, Figure S2).
F IGURE 1. Initial rates of hydrolysis of Boc-Ala-OR esters
catalyzed by trypsin D189S,D194N (gray) and K60E,D189S,-
D194N (black). Conditions: 0.1 M Hepes buffer (pH 8.0), 0.1
M NaCl, 0.01 M CaCl2, 8% (v/v) DMF, [Boc-Ala-OR] ) 2 mM,
[enzymes] ) 1.5 × 10-5 M, X ) hydrolysis rate, R ) Boc-alanyl.
Consequently, the exchange of Asp194 with Asn, al-
though leading to an active enzyme species, does not shift
the activity of trypsin from cleavage to synthesis.
Earlier attempts to repress the proteolytic activity of
trypsin led to the enzyme variant D189S.7c The mutant
was found to have significantly reduced cleavage activity
toward trypsin-sensitive bonds, achieving the efficient
coupling of OGp esters with both Lys- and Arg-containing
peptides. Similar syntheses with Tyr-derived peptides,
however, exhibited a remarkable cleavage activity, re-
flecting the inherent chymotrypsin-like specifity of trypsin
D189S for cleaving Tyr-Xaa bonds. In contrast, trypsin
D194N does not possess chymotrypsin activity. We
expected that the combination of both mutations could
result in an enzyme variant with neither trypsin- nor
chymotrypsin-like cleavage activity. Initially, the effect
of the concurrent D194N and D189S exchange on the
acceptance of substrate mimetics was checked by hy-
drolysis studies using the esters listed in Scheme 1.
Accordingly, the additional D189S mutation, although
reducing the preference for the OGp ester, did not disturb
the activity of the enzyme toward substrate mimetics
(Figure 1). The small differences in the rate of reaction
indicate a broad specificity of trypsin D189S,D194N, a
behavior that was also found for trypsin D189S.7c The
synthetic utility of the double mutant trypsin was finally
investigated by coupling of Arg/Lys- and Phe/Tyr-
containing peptides with Bz-Gly-OGp as the donor (Fig-
ure 2a-d). Remarkably, as illustrated by the syntheses
timelines, the mutant enzyme catalyzes exclusively the
coupling of the substrate mimetic with the appropriate
peptides, remaining free of practically any proteolytic side
reactions. Only a small amount of peptide cleavage can
be detected after longer incubation times after completion
of syntheses. Obviously, the combination of the two
mutations leads to a combination of the behavior of the
appropriate single enzyme variants, resulting in an
enzyme species with reduced trypsin- and chymotrypsin-
like cleavage activity as well. From a synthetic point of
view, this offers practical, irreversible peptide bond
formation without the risk of undesired proteolytic reac-
tions at primarily specific cleavage sites within the
(9) (a) Kossiakoff, A. A.; Chambers, J . L., Kay, L. M.; Stroud, R. M.
Biochemistry 1977, 16, 654-664. (b) Fehlhammer, H., Bode, W.; Huber,
H. J . Mol. Biol. 1977, 111, 415-438.
(10) (a) Birktoft, J .; Kraut, J .; Freer, S. Biochemistry 1976, 15, 4481-
4485. (b) Huber, H.; Bode, W. Acc. Chem. Res. 1978, 11, 114-122.
(11) Hedstrom, L.; Lin, T.-Y.; Fast, W. Biochemistry 1996, 35, 4515-
4523.
9104 J . Org. Chem., Vol. 67, No. 25, 2002

F IGURE 2. Course of the trypsin mutant D189S,D194N-
catalyzed coupling of Bz-Gly-OGp with Ala-Ala-Xaa-Ala-Gly.
Xaa: (a) Lys, (b) Arg, (c) Phe, (d) Tyr. (b) Bz-Gly-OGp, (9)
Bz-Gly-OH, (1) Bz-Gly-Ala-Ala-Xaa-OH, (2) Bz-Gly-Ala-Ala-
Xaa-Ala-Gly-OH. Conditions: 0.1 M Hepes buffer (pH 8.0), 0.1
M NaCl, 0.01 M CaCl2, 8% (v/v) DMF, [Bz-Gly-OGp] ) 2 mM,
[Ala-Ala-Xaa-Ala-Gly] ) 15 mM, [enzyme] ) 1.5 × 10-4 M, X
) product yield.
F IGURE 3. Course of the trypsin mutant K60E,D189S,-
D194N-catalyzed coupling of Bz-Gly-OGp with Ala-Ala-Xaa-
Ala-Gly. Xaa: (a) Lys, (b) Arg, (c) Phe, (d) Tyr. (b) Bz-Gly-
OGp, (9) Bz-Gly-OH, (1) Bz-Gly-Ala-Ala-Xaa-OH, (2) Bz-Gly-
Ala-Ala-Xaa-Ala-Gly-OH. Conditions: 0.1 M Hepes buffer (pH
8.0), 0.1 M NaCl, 0.01 M CaCl2, 8% (v/v) DMF, [Bz-Gly-OGp]
) 2 mM, [Ala-Ala-Xaa-Ala-Gly] ) 15 mM, [enzyme] ) 6.5 ×
10-5 M, X ) product yield.
peptide reactants or products. This qualifies mutant
trypsin D189S,D194N as a useful and most suitable
biocatalyst for substrate mimetics-mediated ligation of
specific peptides. Critically, it must be noted, however,
that for the most specific OGp esters the combination of
the two mutations decreased the rate of reaction to about
10% of that of trypsin D189S and to 5% compared to the
mutant D194N, which finally corresponds to a decrease
in ligase activity of nearly 3 orders of magnitude when
wt-trypsin is taken as the standard. Further efforts to
optimize the enzyme yielded an increase in the activity
of the catalyst. Preliminary results of our own studies
suggest that the exchange of Lys60 with Glu may lead
to a general enhancement of the catalytic power, espe-
cially for mutants lacking the native Asp189 moiety.
Although the molecular basis for this behavior is com-
pletely unknown, we expected the triple mutant K60E,-
D189S,D194N to be an interesting candidate. Hydrolysis
studies with the esters listed in Scheme 1 showed 2-fold
accelerated rates of reaction, confirming the rate increas-
ing effect of the K60E mutation (cf. Figure 1). Comparison
with the data found for the parent trypsin mutant
D189S,D194N revealed that this effect appears to be
specific for the OGp, OPic, and OCap ester, while the
activity vis-a`-vis the other esters remains practically
unaffected. Because the highest reaction rate was found
for the OGp ester, we also used Bz-Gly-OGp as the donor
component for the synthesis reactions. The time courses
of syntheses are illustrated in Figure 3a-d. Despite using
less than half the amount of enzyme, the reactions
proceeded with similar rates compared to those catalyzed
by trypsin D189S,D194N. The degree of proteolytic side
reactions does not increase significantly. Interestingly,
apart from these effects, the exchange of Lys60 with Glu
also reduced the rate of competing hydrolysis of the donor
component. While the use of trypsin D189S,D194N leads
to the formation of ∼30% of Bz-Gly-OH (cf. Figure 2), only
∼20% of hydrolysis product is formed when the triple
mutant is used as the catalyst. This resulted in about
80% yield of the desired peptide products under these
nonoptimized conditions.
Finally, the synthetic scope of the triple mutant was
initially investigated by variation of the sequence and
TABLE 1. Yield s of In ta ct Hexa - a n d Octa p ep tid e P r od u cts Syn th esized by Tr yp sin Mu ta n t
K60E,D189S,D194N-Ca ta lyzed Cou p lin g of Su bstr a te Mim etics w ith Sp ecific Am in o Acid -Con ta in in g P ep tid esa
acyl donor
Bz-Leu-OGp
Bz-Leu-OGp
Bz-Leu-OGp
Bz-Leu-OGp
Bz-Phe-OGp
Bz-Phe-OGp
Bz-Phe-OGp
acyl acceptor
product
% yield
Ala-Ala-Lys-Ala-Gly
Ala-Ala-Arg-Ala-Gly
Ala-Ala-Phe-Ala-Gly
Ala-Ala-Tyr-Ala-Gly
Ala-Ala-Lys-Ala-Gly
Ala-Ala-Arg-Ala-Gly
Ala-Ala-Phe-Ala-Gly
Ala-Ala-Tyr-Ala-Gly
Ala-Ala-Lys-Ala-Gly
Ala-Ala-Arg-Ala-Gly
Ala-Ala-Phe-Ala-Gly
Ala-Ala-Tyr-Ala-Gly
Bz-Leu-Ala-Ala-Lys-Ala-Gly
Bz-Leu-Ala-Ala-Arg-Ala-Gly
Bz-Leu-Ala-Ala-Phe-Ala-Gly
Bz-Leu-Ala-Ala-Tyr-Ala-Gly
Bz-Phe-Ala-Ala-Lys-Ala-Gly
Bz-Phe-Ala-Ala-Arg-Ala-Gly
Bz-Phe-Ala-Ala-Phe-Ala-Gly
Bz-Phe-Ala-Ala-Tyr-Ala-Gly
Boc-Phe-Gly-Gly-Ala-Ala-Lys-Ala-Gly
Boc-Phe-Gly-Gly-Ala-Ala-Arg-Ala-Gly
Boc-Phe-Gly-Gly-Ala-Ala-Phe-Ala-Gly
Boc-Phe-Gly-Gly-Ala-Ala-Tyr-Ala-Gly
55.9
56.1
61.5
60.2
75.1
75.7
69.3
67.6
77.8
85.3
81.9
83.7
Bz-Phe-OGp
Boc-Phe-Gly-Gly-OGp
Boc-Phe-Gly-Gly-OGp
Boc-Phe-Gly-Gly-OGp
Boc-Phe-Gly-Gly-OGp
a
0.1 M Hepes buffer (pH 8.0), 0.1 M NaCl, 0.01 M CaCl2, 8% (v/v) DMF, [Bz-Gly-OGp] ) 2 mM, [Ala-Ala-Xaa-Ala-Gly] ) 15 mM,
[enzyme] ) 6.5 × 10-5 M.
J . Org. Chem, Vol. 67, No. 25, 2002 9105

length of the acyl donor moiety. The results obtained for
reactions with Bz-Leu-OGp, Bz-Phe-OGp, and Boc-Phe-
Gly-Gly-OGp are summarized in Table 1. As already
found for the coupling of Bz-Gly-OGp, a strong preference
of the enzyme for catalyzing the synthesis reactions
became evident. In fact, from the once formed peptide
products more than 95% could be identified as intact and
noncleaved peptide species. Accordingly, the degree of
proteolytic side reactions accounts for less than 5%,
indicating that the synthetic utility of the enzyme is
neither limited to the coupling of C-terminal glycine nor
acyl donor components containing only a single amino
acid moiety. Differences in product yields of about 30%
caused by the individual C-terminal amino acid moiety
and of less than 10% caused by the acyl acceptors used
are not typical for the triple mutant but were also found
in previous studies for reactions catalyzed by wild type
and other mutant trypsins.5d,7b
In summary, these findings imply that rational protein
engineering starting from zymogen-like enzymes can lead
to novel and powerful biocatalysts for peptide synthesis.
While trypsinogen itself appears to be inactive for
substrate mimetics-mediated peptide synthesis, the zy-
mogen-like trypsin mutant D194N catalyzes the forma-
tion of peptide bonds only ∼60-fold less efficiently than
the wt-trypsin. However, this mutation alone does not
repress the degree of proteolytic side reactions vis-a`-vis
trypsin-specific Lys- and Arg-containing peptides signifi-
cantly. Only in combination with additional mutations
in the active site of the enzyme by exchanging Asp189
with Ser is the proteolytic activity of the enzyme reduced,
achieving irreversible peptide synthesis practically with-
out undesired proteolytic reactions. This also holds true
for the coupling of lysine-containing peptides, which
remained as cleavage-sensitive reactants for our recently
reported trypsin D189E ligase.7b The accompanying
decrease of the total enzyme activity can be partly
compensated for by the exchange of Lys60 with Glu, but
remains a challenging task. Furthermore, this mutation
minimizes the competing hydrolysis of the substrate
mimetics as the second important side reaction of enzy-
matic peptide ligation. These characteristics qualify
trypsin mutant K60E,D189S,D194N as a useful and most
suitable biocatalyst for substrate mimetics-mediated
ligation of specific amino acid-containing peptides. This
designed biocatalyst should considerably extend the scope
of this semisynthetic ligation approach, especially in
combination with the solid-phase peptide synthesis of
longer peptide esters in the form of substrate mimetics.12
Studies in this direction are presently underway, paral-
leled by further attempts to increase the ligation rates
of the biocatalyst.
Ack n ow led gm en t. This work has been supported
by the Deutsche Forschungsgemeinschaft (SFB 610 A3
and BO 1770/1-1) and the Fonds der Chemischen
Industrie (Liebig-scholarship, F.B.). The authors wish
to thank Prof. L. Hedstrom (Brandeis University) and
Dr. T. Kurth (University of Leipzig) for providing
plasmids of trypsin K60E and D189S. Furthermore, we
are grateful to Prof. G. Fischer for his continuing
support and to Mr. A. Hemming for reading and
correcting the manuscript.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures of enzymatic reactions and supporting results. This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O026117I
(12) (a) Cerovsky, V.; Bordusa, F. J . Pept. Res. 2000, 55, 325-329.
(b) Cerovsky, V.; Kockskaemper; J . Glitsch, H. G.; Bordusa, F.
ChemBioChem 2000, 2, 126-129.
9106 J . Org. Chem., Vol. 67, No. 25, 2002