Báo cáo khoa học: Inhibitory activity of double-sequence analogues of trypsin inhibitor SFTI-1 from sunflower seeds: an example of peptide splicing docx

Inhibitory activity of double-sequence analogues
of trypsin inhibitor SFTI-1 from sunflower seeds:
an example of peptide splicing
Anna Łe˛ gowska
1
, Adam Lesner
1
,El
_
zbieta Bulak
1
, Anna Jas
´
kiewicz
1
, Adam Sieradzan
1
,
Marzena Cydzik
2
, Piotr Stefanowicz
2
, Zbigniew Szewczuk
2
and Krzysztof Rolka
1
1 Faculty of Chemistry, University of Gdan
´
sk, Poland
2 Faculty of Chemistry, University of Wrocław, Poland
Introduction

Trypsin inhibitor SFTI-1, the smallest member of the
family of Bowman–Birk inhibitors (BBIs), has been
found in sunflower seeds [1]. This homodetic peptide
consists of 14 amino acid residues and its structure is
stabilized by a disulfide bridge (Fig. 1). The reactive
site P
1
-P
1
¢ of this peptide is located between Lys5-Ser6.
As a result of the high sequential and structural
homology of SFTI-1 with the binding loop of the
canonical inhibitors of the BBI family, SFTI-1 forms a
complex with the target enzyme in stoichiometric ratio
of 1 : 1. As reported by Marx et al. [2], upon the incu-
bation with trypsin, the ratio of native SFTI-1 to its
acyclic permutant with hydrolyzed P
1
-P
1
¢ is approxi-
mately 9 : 1. As a result of its small size, very strong
trypsin inhibitory activity and circular backbone scaf-
fold (i.e. well defined 3D structure, displayed b-hairpin
motive), SFTI-1 has attracted interest ever since its dis-
covery. Recent studies on SFTI-1 are summarized in
three reviews [3–5]. Because small proteinaceous inhibi-
tors are of commercial interest, SFTI-1 soon became
an attractive template for the design of new protease
inhibitors with potential use as therapeutic and agro-

chemical agents.
Subsequent to the discovery of SFTI-1, several stud-
ies have shown that the presence of both cycles in
the inhibitor molecule is not essential for its activity.
A monocyclic analogue of SFTI-1, containing only the
disulfide bridge, appeared to have trypsin inhibitory
activity matching that of the wild-type SFTI-1 [6–8].
In addition, it displayed proteinase resistance similar
to that of the parent compound [9]. Another analogue,
containing only head-to-tail cyclization, ([Abu
3,11
]
Keywords
inhibitors; mass spectrometry; peptide
splicing; serine proteinases; SFTI-1
Correspondence
A. Łe˛ gowska, Faculty of Chemistry,
University of Gdan
´
sk, Sobieskiego 18,
80-952 Gdan
´
sk, Poland
Fax: +48 5852 3472
Tel: +48 5852 3359
E-mail:
(Received 4 February 2010, revised 10
March 2010, accepted 15 March 2010)
doi:10.1111/j.1742-4658.2010.07650.x
Four 28-amino acid peptides were synthesized whose sequences comprised

two molecules of trypsin inhibitor sunflower trypsin inhibitor 1 (SFTI-1)
bound through a peptide bond. The peptides in their reactive positions
(5 and 19 of the peptide chain) contain two Lys ([KK]BiSFTI-1) and two
Phe ([FF]BiSFTI-1) residues, along with a combination of the amino acid
residues named thereafter [KF]BiSFTI-1 and [FK]BiSFTI-1. Association
constants of the analogues determined with trypsin and chymotrypsin,
respectively, indicated that they were potent inhibitors of cognate protein-
ases. An MS study of the associates revealed that incubation of the com-
pounds with the proteinases resulted in cutting out a fragment of the peptide
chain to restore the native monocyclic molecule of SFTI-1 or its analogue
[Phe
5
]SFTI-1. This process, analogous to that of the DNA and protein splic-
ing, can be referred to as ‘peptide splicing’.
Abbreviations
BBI, Bowman–Birk inhibitor; ESI, electrospray ionization.
FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS 2351
SFTI-1) displayed trypsin inhibitory activity that was
only 2.5-fold lower than the wild-type inhibitor [6]. As
reported by Korsinczky et al. [8,9], solution structures
of such monocyclic SFTI-1 analogues are remarkably
similar to the solution and the crystal structures of the
wild-type SFTI-1. A higher structural flexibility of
[Abu
3,11
]SFTI-1 compared to that of SFTI-1 is com-
patible with its lower activity and higher hydrolysis
rate. In the wild-type of SFTI-1, a substrate specificity
P
1

position [1] is occupied by Lys residues. For this
reason, SFTI-1 and its monocyclic analogues with
Lys5 were demonstrated to be strong trypsin inhibitors
[6], whereas their chymotrypsin inhibitory activity was
three orders of magnitude lower when association con-
stants (K
a
) with appropriate serine proteinases were
used as a measure of their strength [7]. Monosubstitu-
tion of Lys5 by Phe reversed the SFTI-1 specificity.
Thus, [Phe
5
]SFTI-1 did not inhibit trypsin but exhib-
ited strong chymotrypsin inhibitory activity with
K
a
= 2.0 · 10
9
m
)1
[10].
When designing compounds of commercial impor-
tance, the aim is to reduce their size and simplify the
original structure of a naturally occurring compound
(e.g. protein). Bearing in mind the potential applica-
tions of BBIs, and considering the results presented
above, we designed four SFTI-1 analogues based on
the double-sequence of the wild-type inhibitor. The
primary structures of dimeric SFTI-1 analogues are
shown in Fig. 1. In all of the compounds, two

sequences are bound by a peptide bond formed
between the C-terminal a-carboxyl group of Asp in the
first molecule and the N-terminal a-amino group of
Gly in the second one. To form one disulfide bond
only, two Cys residues located in the middle of the
peptide chain (positions 11 and 17) were replaced by
their structural counterparts of a-aminobutyric acid
(Abu) residues, whereas the remaining two formed a
disulfide bond. The dimeric SFTI-1 analogues differ at
positions 5 and 19. Our synthesized analogues,
[KK]BiSFTI-1 (5) and [FF]BiSFTI-1 (6), as well as
[KF]BiSFTI-1 (7) and [FK]BiSFTI-1 (8), contain Lys
and Phe in positions 5 and 19, in addition to combina-
tions of both amino acid residues. Our intention was
to design low-molecular compounds containing two
reactive sites, with the second one located between
positions 19 and 20. We assumed that these analogues
would be able to form complexes with trypsin or chy-
motrypsin with a stoichiometry of 2 : 1 (analogues of
6 and 7), whereas the two remaining analogues would
inhibit both trypsin and chymotrypsin simultaneously
and independently. Jaulent and Leatherbarrow [11]
reported the synthesis and kinetic studies of a bicyclic
and bifunctional proteinase peptidic inhibitor consist-
ing of 16 amino acids. The inhibitor was designed by
combining two binding loops of BBI. As postulated by
Jaulent and Leatherbarrow [11], the size of the inhibi-
tor was incompatible with the simultaneous binding of
trypsin and chymotrypsin. We predicted that the size
of 28 amino acid residues peptides might be sufficient

to accommodate both enzyme molecules.
Results and Discussion
As indicated in Table 1, all four dimeric SFTI-1 per-
mutants, with the exception of 8, incubated with tryp-
sin, are potent inhibitors. The K
a
values for the
compounds are approximately one order of magnitude
lower than those for their monomeric counterparts.
Surprisingly, [FK]BiSFTI-1 (8) did not block trypsin
activity. This enzyme regained its activity (within
A
B
Fig. 1. Chemical structures of (A) SFTI-1 and (B) synthesized analogues [KK]BiSFTI-1 (5), [FF]BiSFTI-1 (6), [KF]BiSFTI-1 (7) and [FK]BiSFTI-1 (8).
An example of peptide splicing A. Łe˛ gowska et al.
2352 FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS
5 min) after incubation with 8, thus suggesting that 8
behaved as a substrate rather than an inhibitor. At the
same time, this shows that the peptide bond between
Lys19 and Ser20 (the reactive site) and the Arg
16
-
Abu
17
bond are both rapidly hydrolyzed by the
enzyme (Fig. 2) and the hydrolysis products are
quickly released from the enzyme’s substrate pocket. It
is interesting to note that, when compounds 7 and 8
were preincubated with one enzyme each and then
their inhibitory activities were checked against another

enzyme, they displayed inhibitory activity that was at
least one order of magnitude higher. On the basis of
these results, it can be assumed that compounds 7
and 8 inhibit two proteinases independently and
simultaneously. We also found that each of the
used chromogenic substrates was specific for one of
the experimental proteinases and remained intact in
the presence of the other enzyme. Consequently, the
hypothesis that the inhibitory activity of the permu-
tants in the presence of both enzymes could have been
caused by experimental conditions can be ruled out.
Indeed, we conducted the experiments under the condi-
tions recently described by Jaulent and Leatherbarrow
[11], who reported synthesis and kinetic studies on a
bicyclic and bifunctional proteinase peptidic inhibitor
consisting of 16 amino acids. The inhibitor (BiKF)
was designed by combining two building loops of
BBI and was able to inhibit both trypsin and chymo-
trypsin independently but not simultaneously. This
means that, after preincubation with one enzyme, it
completely lost its inhibitory activity against the other
one. As claimed by Jaulent and Leatherbarrow [11],
the size of the inhibitor precluded the simulta-
neous binding of both trypsin and chymotrypsin. The
28-amino acid peptides described in the present study
are remarkably bigger, although they still remain rela-
tively small compared to the proteinases, trypsin
(23 284 Da) and chymotrypsin (25 225 Da). In this sit-
uation, the hypothesis that dimeric SFTI-1 inhibitors
could form either 2 : 1 or 1 : 1 : 1 complexes still needs

further exploration. Unfortunately, based on the
kinetic studies performed, it is impossible to determine
the stoichiometry of the complexes formed by protein-
ase(s) with dimeric SFTI-1 analogues. In an attempt to
do so, we employed gel electrophoresis and HPLC
with a size-exclusion column. Unfortunately, the
results obtained in these experiments were not convinc-
ing and are not discussed here.
One of the methods of choice for studying noncova-
lent complexes formed by proteins is MS with electro-
spray ionization (ESI). An in-depth analysis of
complexes formed between bovine pancreatic trypsin
inhibitor and target proteinases was provided by
Nesatyy [12], who also emphasized that a correlation
between the solution and gas phase binding of the
complexes was not straightforward. There was a not-
oceable difference in the strength of the complexes
formed in the aqueous and gas phase, whereas their
stoichiometry was preserved.
Figures 3 and 4 represent MS spectra of trypsin and
chymotrypsin along with those recorded after their
incubation with [KK]BiSFTI-1 (5) and [FF]BiSFTI-1
(6), respectively. The ESI spectra of bovine b-trypsin
(Fig. 3A) exhibited two charge states of +9 and +10.
The molecular mass of the enzyme calculated from the
first peak was 23 322 Da, whereas the other peak cor-
responded to a trypsin molecule with a trapped cal-
cium ion. After incubation of 5 with trypsin, among
the four peaks seen in the MS spectrum, those with
m ⁄ z 2333.2753 and 2592.4874 were assigned to free

trypsin, whereas the remaining two with m ⁄ z
2486.4686 and 2762.4797 revealed the appearance of a
1 : 1 complex of trypsin with monocyclic SFTI-1
(Fig. 3B). Essentially, an identical peak pattern was
seen with an increasing incubation time of up to 20 h
(data not shown). The MS spectrum of bovine a-chy-
motrypsin (Fig. 4A) produced charge states of +10
and +11. The monoisotopic molecular mass (calculated
using the SNAP procedure in the Bruker Data Analysis
program; Bruker Daltonics, Bremen, Germany) of the
proteinase derived from those peaks was 25 225 Da.
In the spectrum of a 1 : 1 mixture of chymotrypsin
and [FF]BiSFTI-1 (6) incubated for 30 min (Fig. 4B),
two peaks were seen with charge states of +10 and
+11, both representing a 1 : 1 complex between
Table 1. Association constants (K
a
) of SFTI-1 analogues based on
the double-sequence of SFTI-1. Ch and T in parenthesis indicate
that the inhibitory activity was determined after preincubation of
the inhibitor with bovine a-chymotrypsin or bovine b-trypsin,
respectively. ND, not determined (i.e. peptide unstable under the
conditions used for K
a
determination).
Number Analogue
K
a
[M
)1

]
Trypsin Chymotrypsin
1 SFTI-1 (wild) [6,7] 1.1 · 10
10
5.2 · 10
6
2 SFTI-1
a
[6,7] 9.9 · 10
9
4.9 · 10
6
3 [Abu
3,11
]SFTI-1 [6,7] 4.6 · 10
9
1.8 · 10
6
4 [Phe
5
]SFTI-1 [10] 2.0 · 10
9
5 [KK]BiSFTI-1 (4.4 ± 0.4) · 10
8
(1.6 ± 0.2) · 10
8
6 [FF]BiSFTI-1
7 [KF]BiSFTI-1 (2.6 ± 0.2) · 10
8
(8.7 ± 0.2) · 10

8
(1.2 ± 0.2) · 10
10
(Ch) (5.3 ± 0.2) · 10
9
(T)
8 [FK]BiSFTI-1 ND (3.0 ± 0.4) · 10
8
(1.3 ± 0.3) · 10
9
(T)
a
With the exception of wild-type SFTI-1, all inhibitors are monocyclic with
one disulfide bridge only or a head-to-tail cyclization (compound 3).
A. Łe˛ gowska et al. An example of peptide splicing
FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS 2353
chymotrypsin and monocyclic (disulfide bridge only)
[Phe
5
]SFTI-1. It is worth emphasizing that the con-
ditions for the enzyme–inhibitor incubation used in
the MS study differed from those applied for the
determination of K
a
. To detect the complexes, we had
to exchange the buffer for a more volatile one (an
A
B
Fig. 2. MS spectra and results of HPLC
analysis of (A) [FK]BiSFTI-1 (8) and (B) a

mixture of b-trypsin and [FK]BiSFTI-1: peak 2,
analogue 8 without tripeptide Abu-Thr-Lys;
peak 3, analogue 8 with cleaved
Abu-Thr-Lys and Gly-Arg fragments.
An example of peptide splicing A. Łe˛ gowska et al.
2354 FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS
2333.2753
10+
2592.4874
9+
2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 m/z
0
1
2
3
4
5
Intens. x10
7
2333.2753
10+
2592.4874
9+
2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 m/
z
0
1
2
3
4

5
Intens .x10
7
2486.4686
10+
2762.4797
9+
Trypsin
SFTI-1-trypsin
complex
A
B
Fig. 3. MS spectra of (A) bovine b-trypsin and (B) a mixture of b-trypsin and [KK]BiSFTI-1 (5).
2294.1491
11+
2523.5201
10+
2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 m/z
0.0
0.2
0.4
0.6
0.8
1.0
Intens. x10
8
2435.0841
11+
2678.6160
10+

2100 2200 2300 2400 2500 2600 2700 2800 2900 m/z
0.0
0.5
1.0
1.5
2.0
Intens. x10
8
[Phe ]SFTI-1- chymotrypsin
complex
5
A
B
Fig. 4. MS spectra of (A) bovine a-chymotrypsin and (B) a mixture of a-chymotrypsin and [FF]BiSFTI-1 (6).
A. Łe˛ gowska et al. An example of peptide splicing
FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS 2355
ammonium formate buffer, pH 5.8). In a separate
experiment, we demonstrated that, under the conditions
used for the MS study, the peptides exhibited their full
inhibitory activity. In a similar way, we studied interac-
tions of the remaining inhibitors, [KF]BiSFTI-1 (7) and
[FK]BiSFTI-1 (8), by using ESI-MS. Compound 7
incubated with trypsin generated a peak corresponding
to a noncovalent complex of SFTI-1 with trypsin.
However, its formation was significantly slower com-
pared to that of a mixture of 5 with trypsin. In a mix-
ture of 8 with trypsin, only trace amounts of a complex
of [Phe
5
]SFTI-1 with the enzyme were found after 1 h

of incubation. Compound 8 incubated with chymotryp-
sin formed a noncovalent complex, although the reac-
tion was definitely slower than that with compound 6.
In a mixture of 7 with chymotrypsin, no noncovalent
complex was found after 1 h of incubation. On the
other hand, incubation of both compounds 7 and 8 in
a mixture of trypsin and chymotrypsin resulted in the
immediate high-yield formation of noncovalent com-
plexes, SFTI-1–trypsin and [Phe
5
]SFTI-1–chymotryp-
sin, respectively. Figures 5 and 6 show the results of
the MS analyses of those mixtures.
These results clearly indicate that all dimeric ana-
logues undergo proteolysis when incubated with target
enzymes. In all cases, P
1
-P
1
¢ reactive sites are located
between positions 5 ⁄ 6 and 19 ⁄ 20 to release fragments
with Ser6 and Lys19 at their N- and C-termini, respec-
tively. The cleavage is followed by resynthesis of
the peptide bond between Lys5 and Ser20 to pro-
duce monocyclic SFTI-1 or its [Phe
5
]SFTI-1 analogue.
Figure 7 shows the splicing of the permutants medi-
ated by target enzymes. These results are compatible
with the inhibitory activity of the peptides (Table 1),

with all of them being potent inhibitors of the target
enzymes.
Proteolytic susceptibility of the inhibitors was found
to be in excellent agreement with their inhibitory activ-
ity. In all cases where the dimeric species were less pro-
teolytically resistant than their monocyclic reference
compounds (i.e. SFTI-1 and [Phe
5
]SFTI-1), their inhib-
itory activities, expressed in terms of K
a
, were one
order of magnitude lower. On the other hand, in all
cases where the reference inhibitors were formed after
proteolysis, the K
a
values matched those determined
for the reference monomers.
10+
2333.2814
10+
2486.3689
10+
2523.4314
9+
2592.4929
2200 2300 2400 2500 2600 2700 2800 m/
z
0
2

4
6
8
Intens. x10
7
Trypsin
Chymotrypsin
SFTI-1-trypsin
complex
Fig. 5. MS spectrum of a mixture a-chymotrypsin, b-trypsin and [KF]BiSFTI-1 (7).
An example of peptide splicing A. Łe˛ gowska et al.
2356 FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS
Despite our expectations, the peptides did not
form 2 : 1 complexes with the enzymes, nor did they
simultaneously and independently (7 and 8) inhibit
experimental proteinases; instead, they undergo prote-
olysis. The enzymatic process involving proteolytic
cleavage, combined with resynthesis of the peptide
bond, is an intriguing finding. It may serve as a model
for the in vivo formation of cyclic peptides by enzy-
matic processing of their precursors generated by stan-
dard translation. Some bioactive cyclic peptides
Fig. 7. Splicing of the double-sequence SFTI-1 analogues mediated by target enzymes.
10+
2333.2816
11+
2435.0840
9+
2592.4965
10+

2678.6092
2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 m/z
0.0
0.2
0.4
0.6
0.8
1.0
Intens. x10
8
Trypsin
[Phe
5
]SFTI-1-chymotrypsin
complex
Fig. 6. MS spectrum of a mixture b-trypsin, a-chymotrypsin and [FK]BiSFTI-1 (8).
A. Łe˛ gowska et al. An example of peptide splicing
FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS 2357
comprising proteinogenic l-amino acids (e.g. the pep-
tides produced by Caryophyllaceae plants) are likely to
be formed by this mechanism, which can be named
‘peptide splicing’.
Materials and methods
Peptide synthesis
All peptides were synthesized manually by the solid-phase
method using standard Fmoc chemistry on 2-chlorotrityl
chloride resin (substitution of Cl 1.46 meqÆg
)1
) (Calbio-
chem-Novabiochem AG, La

¨
ufelfingen, Switzerland) apply-
ing a previously described procedure [10]. During the last
step, disulfide bridge formation was performed using 0.1 m
solution of I
2
in MeOH as described previously [13]. All
synthetic steps were monitored by HPLC analysis using an
RP Kromasil-100, C
8
,5lm column (4.6 · 250 mm)
(Knauer, Berlin, Germany). The solvent systems were 0.1%
trifluoroacetic acid (A) and 80% acetonitrile in A (B). A
linear gradient of 20–80% B for 30 min was employed with
a flow rate of 1 mlÆmin
)1
, monitored at 226 nm. Finally, all
peptides were purified on a semi-preparative HPLC column
RP Kromasil-100, C
8
,5lm column (8 · 250 mm) (Knauer)
using the same solvent system as above. A linear gradient
of 20–80% B for 30 min was employed with a flow rate of
2.5 mlÆmin
)1
, monitored at 226 nm. To confirm the correct-
ness of molecular weights of the peptides, MS analysis was
carried out on a MALDI MS (Biflex III MALDI-TOF
spectrometer; Bruker Daltonics) using a-CCA matrix.
Determination of association constants

The association constants were measured using a method
developed in the laboratory of Laskowski et al. [14,15]. The
procedure was described in detail previously [10].
The measurements were carried out at initial enzyme con-
centrations over the ranges 5.1–5.8 nm and 2.8–7.2 nm for
trypsin and chymotrypsin, respectively. To determine the
K
a
values for [KF]BiSFTI-1 (7) and [FK]BiSFTI-1 (8) with
trypsin in the presence of chymotrypsin, a two-fold molar
excess over the inhibitor of the second enzyme was added
to each of the experimental cuvettes, followed 5 min later
by the addition of appropriate volumes of the trypsin and
substrate solutions. The chymotrypsin inhibitory activity in
the presence of trypsin was determined by the same proce-
dure, using a reverse order of enzyme addition for preincu-
bation.
Proteolytic susceptibility assays
Dimeric analogues of SFTI-1 were incubated in a 100 mm
Tris-HCl buffer (pH 8.3) containing 20 mm CaCl
2
and
0.005% Triton X-100, using catalytic amounts of bovine
b-trypsin or bovine a-chymotrypsin (1 mol%) [11]. The
incubation was carried out at room temperature and
aliquots of the mixture were taken out periodically and
submitted to RP-HPLC analysis.
Analysis of enzyme–inhibitor complexes
using MS
A 1.4 · 10

)5
m solution containing a proteolytic enzyme
(trypsin or chymotrypsin) and inhibitor (1.7 · 10
)5
m)ina
20 mm aqueous ammonium formate buffer (pH 5.8) was
incubated for predetermined periods of 0.5, 1 and 20 h.
After incubation, the mixture was analysed directly by ESI-
MS spectrometry. The experiments were performed using
an FT-MS instrument (Apex-Ultra 7T; Bruker Daltonic)
equipped with a dual ESI-MADI Apollo source (Agilent
Technologies Inc., Santa Clara, CA, USA). The samples
were infused at a flow rate of 2 llÆmin
)1
. The potential
between the spray needle and the orifice was set at 4.5 kV.
Capillary temperature was 200 °C, and N
2
was used as a
nebulizing gas.
Acknowledgements
This work was supported by the Ministry of Science
and Higher Education (grant no. 2889 ⁄ H03 ⁄ 2008 ⁄ 34).
References
1 Luckett RS, Garcia JJ, Barker AV, Konarev PR,
Shewry AR, Clarke RL & Brady RL (1999) High-
resolution structure of a potent, cyclic proteinase
inhibitor from sunflower seeds. J Mol Biol 290,
525–533.
2 Marx UC, Korsinczky MLJ, Schirra HJ, Jones A,

Condie B, Otvos L Jr & Craik DJ (2003) Enzymatic
cyclization of a potent Bowman–Birk protease inhibitor,
sunflower trypsin inhibitor-1, and solution structure of
an acyclic precursor peptide. J Biol Chem 278, 21782–
21789.
3 Qi R-F & Chi C (2005) Structural features and molecu-
lar evolution of Bowman–Birk protease inhibitors and
their potential application. Acta Biochim Biophys Sin
37, 283–292.
4 McBride JD, Watson EM, Brauer ABE, Jaulent AM &
Leatherbarrow RJ (2002) Peptide mimics of the
Bowman–Birk inhibitor reactive site loop. Biopolymers
66, 79–92.
5 Korsinczky MLJ, Schirra HJ & Craik DJ (2004)
Sunflower trypsin inhibitor-1. Curr Protein Pept Sci 5,
351–364.
6 Zabłotna E, Kaz´ mierczak K, Jas
´
kiewicz A, Stawikowski
M, Kupryszewski G & Rolka K (2002) Chemical
synthesis and kinetic study of the smallest naturally
An example of peptide splicing A. Łe˛ gowska et al.
2358 FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS
occurring trypsin inhibitor SFTI-1 isolated from sun-
flower seeds and its analogues. Biochem Biophys Res
Commun 292, 855–859.
7 Zabłotna E, Kaz´ mierczak K, Jas
´
kiewicz A, Kupryszew-
ski G & Rolka K (2002) Inhibition of bovine a-chymo-

trypsin by cyclic trypsin inhibitor SFTI-1 isolated from
sunflower seeds and its two acyclic analogues. LIPS
9, 131–134.
8 Korsinczky ML, Clark RJ & Craik DJ (2005) Disulfide
bond mutagenesis and the structure and function of the
head-to-tail macrocyclic trypsin inhibitor SFTI-1.
Biochemistry 44, 1145–1153.
9 Korsinczky ML, Schirra HJ, Rosengren KJ, West J,
Condie BA, Otvos L, Anderson MA & Craik DJ (2001)
Solution structures by
1
H NMR of the novel cyclic
trypsin inhibitor SFTI-1 from sunflower seeds and an
acyclic permutant. J Mol Biol 311, 579–591.
10 Łe˛gowska A, De˛bowski D, Lesner A, Wysocka M &
Rolka K (2009) Introduction of non-natural amino acid
residues into the substrate-specific P
1
position of trypsin
inhibitor SFTI-1 yields potent chymotrypsin and
cathepsin G inhibitors. Bioorg Med Chem 17, 3302–
3307.
11 Jaulent AM & Leatherbarrow RJ (2004) Design,
synthesis and analysis of novel bicyclic and bifunc-
tional protease inhibitors. Protein Eng Des Sel 17,
681–687.
12 Nesatyy VJ (2001) Gas-phase binding of non-covalent
protein complexes between bovine pancreatic trypsin
inhibitor and its target enzymes studied by electrospray
ionization tandem mass spectrometry. J Mass Spectrom

36, 950–959.
13 Yang Y, Sweeney WV, Schneider K, Chait BT & Tam
JP (1994) Two-step selective formation of three disulfide
bridges in the synthesis of the C-terminal epidermal
growth factor-like domain in human blood coagulation
factor IX. Protein Sci 3, 1267–1275.
14 Empie MW & Laskowski M Jr (1982) Thermodynamics
and kinetics of single residue replacements in avian
ovomucoid third domains: effect on inhibitor inter-
actions with serine proteinases. Biochemistry 21, 2274–
2284.
15 Park SJ (1985) Effect of amino acid replacement in
ovomucoid third domains upon their association with
serine proteinases. PhD Thesis, Purdue University, IN,
USA.
A. Łe˛ gowska et al. An example of peptide splicing
FEBS Journal 277 (2010) 2351–2359 ª 2010 The Authors Journal compilation ª 2010 FEBS 2359