Báo cáo khoa học: Arg143 and Lys192 of the human mast cell chymase mediate the preference for acidic amino acids in position P2¢ of substrates pdf

Arg143 and Lys192 of the human mast cell chymase
mediate the preference for acidic amino acids in
position P2¢ of substrates
Mattias K. Andersson, Michael Thorpe and Lars Hellman
Department of Cell and Molecular Biology, Uppsala University, The Biomedical Center, Sweden
Introduction
Mast cells (MCs) are resident tissue cells that are dis-
tributed along the surfaces of the body. They are fre-
quently found in the mucosa of the airways and
intestine, in connective tissue of the skin, and around
blood vessels and nerves. Upon activation, MCs are
able to rapidly exocytose their cytoplasmic granules,
resulting in the release of prestored physiologically
active inflammatory mediators. The majority of pro-
teins found in these granules are serine proteases, and
one subfamily of these proteases comprises the
chymases. Chymases cleave substrates after aromatic
amino acids, and are therefore chymotrypsin-like.
Phylogenetic analyses of the chymases have led to
the identification of two distinct subfamilies, the a-chy-
mases and the b-chymases [1]. The a-chymases are
encoded by a single gene in all species investigated,
except for ruminants, where two very similar a-chym-
ase genes have been identified [2]. The b-chymases
Keywords
chymase; cleavage specificity; human
chymase; mast cell; site-directed
mutagenesis
Correspondence
L. Hellman, Department of Cell and
Molecular Biology, Uppsala University, The

Biomedical Center, Box 596, SE-751 24
Uppsala, Sweden
Fax: +46 0 18 471 4382
Tel: +46 0 18 471 4532
E-mail:
Website: />(Received 29 December 2009, revised 2
March 2010, accepted 4 March 2010)
doi:10.1111/j.1742-4658.2010.07642.x
Chymases are chymotrypsin-like serine proteases that are found in large
amounts in mast cell granules. So far, the extended cleavage specificities of
eight such chymases have been determined, and four of these were shown
to have a strong preference for acidic amino acids at position P2¢. These
enzymes have basic amino acids in positions 143 and 192 (Arg and Lys,
respectively). We therefore hypothesized that Arg143 and Lys192 of human
chymase mediate the preference for acidic amino acids at position P2¢ of
substrates. In order to address this question, we performed site-directed
mutagenesis of these two positions in human chymase. Analysis of the
extended cleavage specificities of two single mutants (Arg143 fi Gln and
Lys192 fi Met) and the combined double mutant revealed an altered
specificity for P2¢ amino acids, whereas all other positions were essentially
unaffected. A weakened preference for acidic amino acids at position P2¢
was observed for the two single mutants, whereas the double mutant lacked
this preference. Therefore, we conclude that positions 143 and 192 in
human chymase contribute to the strong preference for negatively charged
amino acids at position P2¢. This is the first time that a similar combined
effect has been shown to influence the cleavage specificity, apart from posi-
tion P1, among the chymases. Furthermore, the conservation of the prefer-
ence for acidic P2¢ amino acids for several mast cell chymases clearly
indicates that other substrates than angiotensin I may be major in vivo
targets for these enzymes.

Abbreviations
Ang, angiotensin; DC, dog chymase; EK, enterokinase; HC, human chymase; IPTG, isopropyl thio-b-
D-galactoside; MC, mast cell; mMCP,
mouse mast cell protease; OC, opossum chymase; rMCP, rat mast cell protease.
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2255
have only been identified in rodents. Interestingly, the
rodent a-chymases mouse MC protease (mMCP)-5
and rat MC protease (rMCP)-5 have changed their pri-
mary cleavage specificity from aromatic amino acids
(chymotrysin-like) to aliphatic amino acids (elastase-
like).
A large number of in vitro substrates have been
identified for the chymases. However, the absolute
majority of these have never been shown to also be
substrates in vivo [3]. Therefore, the true functions of
the chymases most likely remain to be identified. In
order to increase our understanding of these enzymes,
a necessary step is to determine the most important
feature of an enzyme, the specificity-determining inter-
actions with its substrates.
In a previous study, we determined the cleavage
specificity in seven positions from positions P4 to P3¢
for human chymase (HC) [4]. The cleavage of the pep-
tide bond occurs between positions P1 and P1¢, where
the amino acids N-terminal of this bond are designated
P1, P2, P3, P4 Pn and those C-terminal P1¢,P2¢,
P3¢ Pn¢ [5]. The strongest preference observed,
besides the primary specificity for P1 Phe or Tyr, was
the preference for negatively charged (acidic) amino
acids at position P2¢. An evaluation of natural sub-

strates for HC showed that many of these also have
acidic amino acids at position P2¢ [4,6]. These observa-
tions suggest an important role for negatively charged
amino acids at position P2¢ during substrate discrimi-
nation by HC. The structure of HC has been exten-
sively investigated, and also compared with those of
MC chymases in other species. These studies have pro-
vided insights into important enzyme–substrate interac-
tions. For example, molecular modeling of HC
interacting with angiotensin (Ang) I has led to conclu-
sions regarding the S2¢ binding site of HC [7–9]. These
studies have shown that Lys40, Arg143 and Lys192
are located close to the S2¢ binding site, which may
favor negatively charged P2¢ side chains of substrates.
However, these data are based on structural studies of
HC in complex with inhibitors, where the interaction
of an acidic amino acid with the S2¢ subsite cannot be
determined. In addition, Ang I that was modeled into
the active cleft of HC did not bring an acidic amino
acid to this position. Therefore, it is still uncertain
which of the side chains of Lys40, Arg143 or Lys192
are able to contact the acidic side chain of a negatively
charged amino acid in position P2¢ of a substrate.
The extended cleavage specificities of several related
MC chymases have recently been determined. The
a-chymases opossum chymase (OC) guinea pig chymase
and rMCP-5 and the b-chymase mMCP-4 have also
been found to prefer acidic P2¢ amino acids (Table 1)
[6,10,11,17]. In contrast, the dog a-chymase and the
b-chymases mMCP-1, rMCP-1 and rMCP-4 were found

to prefer other amino acids than Asp or Glu in this posi-
tion [6,12,13] (submitted manuscript Gallwitz et al.,
2009). When the amino acids in positions 40, 143 and
192 of the above chymases are compared, the five chy-
mases with a specificity for acidic P2¢ amino acids all
have Arg143 and Lys192. However, they differ at posi-
tion 40 (Table 1). Furthermore, none of the four
chymases that lack the acidic P2¢ specificity has an Arg
at position 143. However, three of them have Lys at
position 192. On the basis of these observations, we
hypothesized that Arg143 alone or in cooperation with
Lys192 mediates the preference for acidic amino acids
at position P2¢. In the present study, we tested the
roles of Arg143 and Lys192 as P2¢ specificity-determin-
ing residues. By in vitro mutagenesis, the HC coding
region was modified so that Arg143 was replaced by
Gln and Lys192 was replaced by Met, which are
amino acids found in the same positions of chymases
that lack acidic P2¢ specificity. Our results clearly show
that positions 143 and 192 have an effect in mediating
the acidic P2¢
specificity. Arg143 and Lys192 are essen-
tial in conferring a strong preference for acidic amino
acids at position P2¢ of the substrates.
Table 1. P2¢ specificity and amino acids found in positions 40, 143 and 192 of nine different chymases.
Chymase P2¢ specificity Residue 40 Residue 143 Residue 192 Reference
Human a-chymase Asp ⁄ Glu Lys Arg Lys 4
Opossum a-chymase Asp His Arg Lys 10
rMCP-5 (a-chymase) Glu > Leu Ser Arg Lys 11
mMCP-4 (b-chymase) Glu ⁄ Asp Ala Arg Lys 6

Guinea pig chymase Glu ⁄ Asp > Gln > Ala Lys Arg Lys 17
Dog a-chymase Ser ⁄ Leu > Glu ⁄ Asp Ala Lys Lys (Gallwitz et al. 2009,
unpublished results)
rMCP-1 (b-chymase) Ser Ala Gln Lys 6
rMCP-4 (b-chymase) Gly > Ser ⁄ Leu Thr Gln Lys 13
mMCP-1 (b-chymase) Leu > Val > Ala Asp Lys Met 12
P2¢ specificity of the human mast cell chymase M. K. Andersson et al.
2256 FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS
Results
Production and purification of recombinant HC
mutants
Two single mutants of HC, Arg143 fi Gln and Lys192
fi Met, and a double mutant, Arg143 fi Gln +
Lys192 fi Met, were produced by in vitro mutagenesis.
Following control sequencing of the full coding
regions, the three different pCEP-Pu2 vector constructs
were transfected into HEK 293 EBNA cells for protein
production. Recombinant protein was purified from
conditioned media on Ni
2+
–nitrilotriacetic acid aga-
rose, by binding through the N-terminal His6-tag. The
protein yield was 100–150 lg recombinant protein
from 1 L of medium for all three mutants.
Activation and further purification of the
recombinant HC mutants
Following the initial Ni
2+
–nitrilotriacetic acid agarose
purification, the three different recombinant HC

mutants were activated by removal of the His6-tag by
proteolytic cleavage with enterokinase (EK). Approxi-
mately 30 lg of each mutant was treated with EK for
5 h at 37 °C. Samples of inactive and activated prote-
ases were separated on SDS ⁄ PAGE gels, in order to
ensure successful removal of the His6-tag and the
EK-susceptible cleavage site (Fig. 1). Like the wild-
type enzyme, the mutated inactive proteases migrated
as 35 kDa bands, and the EK-digested enzymes
migrated as 33 kDa bands (Fig. 1). This is somewhat
over the theoretical value of 25 kDa for wild-type HC,
which indicates glycosylation at two sites of these
proteases. To purify the activated proteases from
contaminating serum and cellular proteins, imidazole,
and EK, they were purified over a heparin–Sepharose
column. The heparin–Sepharose-purified fractions were
separated on SDS ⁄ PAGE gels, and no contaminating
bands could be detected (Fig. 1). The proteolytic activ-
ities of the eluted fractions of the three mutated HCs
were analyzed by cleavage of the chymotrypsin-sensi-
tive chromogenic substrate S-2586 (MeO-Suc-Arg-
Ala-Tyr-pNA, Chromogenix, Mo
¨
lndal, Sweden, data
not shown).
Determination of the extended cleavage
specificity of the three HC mutants by phage
display technology
The phage library used to determine the extended
cleavage specificity of the HC mutants contains

 5 · 10
7
phage clones. Each phage clone expresses a
unique sequence of nine random amino acids, followed
by a His6-tag in the C-terminus of capsid protein 10.
Thereby, the phages display a random nonamer on
their surface, and by interactions of the His6-tag the
phages can be immobilized on Ni
2+
–nitrilotriacetic
acid agarose beads. The three HC mutants were used
to screen the phage library for peptides susceptible to
cleavage. After the first selection step (biopanning), the
phages, released by digestion of nonapeptides, were
amplified in Escherichia coli and subjected to addi-
tional biopannings. Selection of nonamers susceptible
to cleavage by the Lys192 fi Met and Arg143
fi Gln + Lys192 fi Met HC mutants was per-
formed over five biopannings, after which they induced
the release of 47 and 46 times more phages, respec-
tively, than an NaCl ⁄ P
i
control (Fig. 2). Peptides sensi-
tive to cleavage by the Arg143 fi Gln mutant were
selected over six biopannings, which resulted in an
81-fold greater release of phages as compared with an
NaCl ⁄ P
i
control (Fig. 2).
After the last biopanning, 44 individual phage

clones were isolated for each of the three HC
mutants, and the sequences encoding the randomly
synthesized nonapeptides were determined. The nucle-
otide sequences were then translated into nonapep-
tides, which were aligned on the basis of similarities
to the cleavage specificity of wild-type HC [4]. For
the Arg143 fi Gln mutant, 41 sequences were deter-
mined in total, two of which were obvious back-
ground sequences and therefore not included in the
–EK +EK Hep –EK +EK Hep –EK +EK Hep
R143Q K192M K192M
R143Q +
97
66
45
30
20
14
kDa
Fig. 1. Purification and activation of recombinant HC mutants. Three
different recombinant HC mutants were expressed with an N-termi-
nal His6-tag followed by an EK-susceptible sequence replacing the
signal peptide. These proenzymes were first purified on Ni
2+
–nitrilo-
triacetic acid beads ()EK), and then activated by removal of the
His6-tag by EK digestion (+EK). Following activation, the enzymes
were further purified on heparin–Sepharose columns (Hep). Proen-
zymes and activated enzymes before and after heparin–Sepharose
purification were analyzed by separation on SDS ⁄ PAGE gels and

visualized with Coomassie Brilliant Blue staining.
M. K. Andersson et al. P2¢ specificity of the human mast cell chymase
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2257
alignment. The 39 remaining sequences were aligned
(Fig. 3A). Twelve of these sequences were derived from
the same phage clone (Trp-Trp-Ala-Ile-Glu-Met-Phe-
Asp-Met), four sequences from the clone Trp-Phe-Val-
Thr-Phe-Tyr-Asp-Ser-Leu, and two from the clone
Val-Val-Ser-Tyr-Gly-Gly-Val-Leu-Glu. From the 44
phage clones isolated for the Lys192 fi Met mutant,
40 sequences were determined, of which three were
background phages. The remaining 37 sequences were
aligned. Among these sequences, one of the phage
clones (Pro-Met-Leu-Tyr-Ser-Leu-Asn-Asp-Ser) was
found twice (Fig. 3B). Of the phage clones from the
double mutant, 36 clones were analyzed. Three of
these were background phages and the remaining 33
were aligned (Fig. 3C). Two of the aligned sequences
were derived from the same clone (Thr-Leu-Phe-Tyr-
Trp-Gly-Ala-Thr-Gly).
On the basis of the alignments, the distribution of
amino acids in positions P4–P3¢ were calculated for
each HC mutant (Fig. 4). In order to normalize for the
uneven occurrence of individual phage clones in the
alignment, all clones that were found more than once
were calculated as one. As expected, the three HC
mutants showed very similar preferences for amino
acids flanking the cleaved peptide bond. In the posi-
tions N-terminal of the cleaved bond, positions P2–P4,
a clear overrepresentation of aliphatic amino acids,

particularly Val and Leu, was observed. In
position P1, Phe and Tyr were more frequently seen
than Trp. The three mutants also shared preferences in
the positions on the C-terminal side of the scissile
bond. The aliphatic Gly and Ala, and the hydrophilic
Ser, dominate in position P1¢. Similar preferences were
identified in position P3¢, where aliphatic amino acids
were generally preferred. However, in this position, the
hydrophilic amino acids Ser and Thr were also fre-
quently found. All of the positions mentioned above fit
very well with the cleavage specificity of wild-type HC,
as previously determined by our group [4]. The only
position where we detected an altered specificity
between the wild-type HC and the HC mutants was in
position P2¢. The strong preference for acidic amino
acids found in wild-type HC had disappeared in the
Arg143 fi Gln and Lys192 fi Met mutants and the
double mutant, and instead we observed a preference
for aliphatic amino acids. A comparison between the
wild-type HC and the HC mutants regarding the speci-
ficity for acidic amino acids at position P2¢ is shown in
Fig. 5. In a previous study, where the same phage-dis-
played nonapeptide library was used to determine the
cleavage specificity of wild-type HC, a negatively
charged amino acid at position P2¢ was seen in 58% of
the sequences. For the Arg143 fi Gln mutant, this
figure was reduced to 25%, and for the Lys192 fi Met
mutant, 19% of the sequences were aligned with an
acidic amino acid at position P2¢. When these muta-
tions were combined in the double mutant, we could

only observe acidic amino acids at position P2¢ in 6%
of the sequences. Thus, Arg143 and Lys192 contribute
almost equally to the P2¢ specificity of HC. Mutating
either of the two basic amino acids in these positions
partially disrupts the acidic P2¢ preference, and mutat-
ing both of them totally removes this preference. The
negatively charged amino acids Asp and Glu are
roughly equally distributed in this position in the wild-
type HC and the three HC mutants. Therefore,
Arg143 and Lys192 do not distinguish between Asp or
Glu, but attract these residues equally well, probably
solely on the basis on the negative charge of their side
chains.
Verifying the consensus sequence by the use of a
new type of recombinant protein substrate
In order to verify the results from the phage display
analysis, a new type of recombinant substrate was
developed. The consensus sequence obtained from the
phage display analysis was inserted in the linker region
between two E. coli thioredoxin molecules by ligating
a double-stranded oligonucleotide encoding the actual
sequence into a BamHI and a SalI site of the vector con-
struct (Fig. 6A). For purification purposes, a His6-tag
R143Q
K192M
R143Q + K192M
Biopannings
Cleaved/control ratio
0
20

40
60
80
100
0123456
Fig. 2. Amount of released T7 phages after digestion with HC
mutants, as compared with an NaCl ⁄ P
i
control. A library of ran-
domly synthesized nonamers expressed at the C-terminus of T7
phages were subjected to cleavage by HC mutants. Selection for
nonamers susceptible to cleavage by the HC mutants was per-
formed in five or six rounds of selection (biopannings). After each
biopanning, the amount of released phages was determined and
compared to an NaCl ⁄ P
i
control. The ratio of phages released by
enzyme digestion over the NaCl ⁄ P
i
control for each biopanning is
shown. The Arg143 fi Gln mutant is indicated by a dashed line,
the Lys192 fi Met mutant by a dotted line, and the
Arg143 fi Gln + Lys192 fi Met double mutant by an unbroken
line.
P2¢ specificity of the human mast cell chymase M. K. Andersson et al.
2258 FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS
was added to the C-terminus of this protein (Fig. 6A).
A number of related and unrelated substrate sequences
were also produced with this system, by ligating the
corresponding oligonuclotides into the BamHI ⁄ SalI

sites of the vector. All of these substrates were
expressed as soluble proteins in a bacterial host,
E. coli, and purified on immobilized metal affinity
chromatography columns to obtain a protein with a
purity of 90–95%. These recombinant proteins were
then used to study the preference of HC and the dou-
ble mutant for these different sequences (Fig. 6B–E).
The results showed that HC very efficiently cleaved the
HC consensus sequence (VVLFSEVL) [4]. When the
Glu in position P2¢ of the HC consensus sequence was
replaced by a Gly (VVLFSGVL), the efficiency of
cleavage by HC dropped by a factor of 4–5 (Fig. 6B).
In contrast, the HC double mutant preferred this sub-
strate over the HC consensus site by a factor of 2–5
(Fig. 6C and data not shown). This latter experiment
shows that HC has a marked preference for negatively
charged amino acids in position P2¢ and that the dou-
ble mutant has lost this preference, instead preferring
aliphatic amino acids in this position.
The consensus site for dog chymase (DC) has
recently been determined (submitted manuscript
Gallwitz et al., 2009). This site (VVRFLSLL) shows
similarities with the preferred site for the HC double
mutant, and neither sequence contains an acidic P2¢
residue. Consequently, HC was found to cleave the
DC consensus sequence five-fold to seven-fold less
Y
S
A
R R

V
G
V
M
F
S S
V
A A
V
GP
F
S SM M
A
L L
E
Y
V
G
I
K
G
S
IL
F
G G
S
V
A
R
L L

Y
G G
V
A
M
LV
Q
Y
G G
LV
S S
L V
W
H HT
GG A
R
S
G
S
Y
S
A A
L L
MM M
Y
A A
L
R
LL LV
FF

T
L
M
G
S
L
R
F
LV
T
W
G
T
P A
FF
LV
A G
D
G
V
F
T
LV
E
W
A A
I
FF
V
A G G

L
R R
FF
R
A G
E
S T
A
Y
L L
S
F
A
V
R
A
F F
C C
G
L L
P
T
Y W
M C
LVV
AG
F F
VV
G A
R

L
R R
F W
V
S
R
V
A
I
P
F
A
LV
F
G
E
S
A
F F
G
S
V
G
T
L
R
H
W WW
RR
T

GG
D
Y WW
R
E
V L
S
G
F F Y
G
V
A
M
G
T T
F WW
V
T
A
V
G G
F F W
VV
A P G
V
W WF
LV
P
E
T

A
F Y Y F
G
LI
GA P
2
Y WF
L
T
G A
T
G
W
H H H
L L
S S
G
E
KR
P4 P3 P2 P1 P1′P2′ P3′
R143Q + K192M
R
I
S
F
A G
E
V
F
Y

E
M
V L
MM
V
P
R
Y
A A G
C
A
V
A
Y
A
L
E E
A
MM
L
Y
A A
T
V
D
G
V
A
F
T

D
M
KR
V
G
W
PP G G
Y
I
G
R
L
W
R
L
Y
S
L
E
L
T
V
F
K
W
E
K
I
P
W

G
S S
W
T
I
P
L
Y
A
R
S
FF
V
S S
L
S
V
A
FF
G
Q
L L
R
A
L
F
A G
S
LILV
W

FF
S
L
T
L
A
L
M
V
T
L
F W
G
T
P A
Y
S
A
S
V V
W
G
E
S
F
S H H
L V
P
Y
Q

I
FF
AA
V L
E
L V
L L
W
R
F
L L
C C
Y
S
L
M
V VV
W
D
F
H H
L
SQ
G
R
L
R
W
F
E

A
Y
S
Y
M
L
R
F
A
L
P
F
A
W
G
L
Y
G
D
L
F
V
T
A
W
F
E
A
L
R

FW
I
A
F
S
G
F
L V
P
V
W
A
Y
S
L
F
A
F
V
T
F
S
I
F
G
F
T
L L
VV V
S

F
A
WW
G
H
Y
E
A
F
LL
W
A
V
FF
S
W
V
A
V
S
P
F
A
L
H
L
W
P
YW
A

Y
S
D
E
S
W
R
FW
F
E
A
WYF
V
T
E
W
R
V V
A
Y
L
G
D
2
Y
S
L
N
D
S

L
M
P
P4 P3 P2 P1 P1′P2′ P3′
K192M
Negatively charged
Positively charged
Aromatic
Large aliphatic
Small aliphatic
P G G
W
S
R
M
L LV
R
P
VV L
F
S T
G
E
S
E
MT
I
T
V
Y

E
S H
R
Q
R R
F
P
V V
S
G
L
F
L
S S
G
R
E
D
V L
Y
A
F
GG
T
V
Y F
I
S
L
S

A
D
I
G
M
P
Y
S T
L
M
F
G
W
A
S
VV
F
G
T T
Y F
MM S
G
VL V
V
Y
V
M
G
S
Y

I
G
F
R
S
L V
T
V
E
W
W
E
T T
V
F
L
A
T
YY
GG
E
V
P
V
AG
V
FF
A
T
P A

D
I
W
E
W
P
L
F
L
A A
W
VV
S
F W
V
A
D
T
W W
R
D
F
S T N
G
W
A A
Y F
L
D
F

T
V
W
D
F FF
G
E
H
W
E
I VV
W
L
F
G
12
M
W W
A
I
E
F
D
M H
W F
V
F
T S
L
Y

D
4
2
Y
GG
E
S
VV LV
P4 P3 P2 P1 P1′P2′ P3′
R143Q
ABC
Fig. 3. Phage-displayed nonamers susceptible to cleavage by HC mutants after five or six biopannings. After the last selection step, phages
released by proteolytic cleavage of the HC mutants were isolated, and the sequences encoding the nonamers were determined. The general
sequence of the T7 phage capsid proteins is PGG(X)
9
HHHHHH, where (X)
9
indicates the randomized nonamers. The protein sequences were
aligned into a P4–P3¢ consensus, where cleavage occurs between positions P1 and P1¢. If the sequence was found more than once, this is
indicated by the corresponding number to the left of the sequence. The amino acids are color coded according to the side chain properties
as indicated in the key. For the Arg143 fi Gln mutant (A), 24 unique sequences were aligned; for the Lys192 fi Met mutant (B), 36 unique
sequences were aligned; and for the Arg143 fi Gln + Lys192 fi Met double mutant (C), 32 unique sequences were aligned. The
sequences with one aromatic amino acid (potential cleavage site) are placed on the top, followed by sequences containing two, three or four
aromatic amino acids.
M. K. Andersson et al. P2¢ specificity of the human mast cell chymase
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2259
0
10
20
30

40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
WT
R143Q
K192M
R143Q+
K192M
P2
0
10
20
30

40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
WT
R143Q
K192M
R143Q+
K192M
P3
0
10
20
30

40
0
10
20
30
40
0
10
20
30
40
WT
R143Q
K192M
R143Q+
K192M
P4
FYWGAVL I PSTCMNQHKRDE
FYWGAVL I PSTCMNQHKRDE
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30

40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
WT
R143Q
K192M
R143Q+
K192M
P3´
0
10
20
30

40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
WT
R143Q
K192M
R143Q+
K192M
P2´
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30

40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
0
10
20
30
40
FYWGAVL I PSTCMNQHKRDE
WT
R143Q
K192M
R143Q+
K192M
P1´
0
10
20
30

40
50
60
70
FYWGAVL I PSTCMNQHKRDE
R143Q+
K192M
P1
0
10
20
30
40
50
60
70
FYWGAVL I PSTCMNQHKRDE
R143Q
P1
P1
0
10
20
30
40
50
60
70
FYWGAVL I PSTCMNQHKRDE
WT

0
10
20
30
40
50
60
70
FYWGAVL I PSTCMNQHKRDE
K192M
P1
Occurrence (%)
Occurrence (%)Occurrence (%)
Occurrence (%)Occurrence (%)Occurrence (%)Occurrence (%)Occurrence (%)Occurrence (%)Occurrence (%)
Fig. 4. Distribution of amino acids at positions P4–P3¢ in phage-displayed nonamers cleaved by wild-type (WT) HC or HC mutants after five
or six biopannings. On the basis of the alignment in Fig. 3 and previously published data on wild-type HC, the percentage of each amino acid
present in each position, P4 to P3¢, was calculated. The amino acids are ordered from left to right: aromatic, aliphatic, hydrophilic, basic (pos-
itively charged), and acidic (negatively charged).
P2¢ specificity of the human mast cell chymase M. K. Andersson et al.
2260 FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS
efficiently than the HC consensus sequence, whereas
the HC double mutant cleaved this substrate almost as
efficiently as its own consensus site (Fig. 6B,C).
A few additional substrates were also included in
this study. The optimal sequence for cleavage by OC
has recently been determined [10]. As compared with
HC, this enzyme was found to have a preference for
Trp over Phe and Tyr at position P1. When we ana-
lyzed the cleavage of this sequence (VGLWLDRV), we
observed that HC cleaved this sequence  50-fold less

efficiently than the HC consensus sequence (Fig. 6B
and data not shown). Similarly, the HC double mutant
cleaved this sequence with a very low cleavage rate
(Fig. 6C).
We also tested three additional sequences, the
human thrombin consensus (LTPRGVRL), which we
recently determined by phage display analysis, and the
rat granzyme B (LIETDSGL) [14] and EK consensus
sequences (LDDDDKGL). Neither the granzyme B,
the EK nor the thrombin substrate was cleaved at all
by HC and the HC double mutant, even after 150 min
at room temperature (Fig. 6D,E).
OC was included here as reference and to compare
its preferences for the different sequences used to
study HC and the HC double mutant (Fig. 6F). OC
was found to cleave the OC consensus sequence five-
fold to eight-fold more efficiently than the HC or the
DC consensus sites. This verifies its preference for
Trp over Phe and Tyr at position P1 and the accu-
racy of the information obtained by the phage display
analysis.
Discussion
Site-directed mutagenesis has previously been used to
study the effect on cleavage specificity of changing a
single amino acid in an enzyme. As an example, the
primary specificity of the serine protease mouse gran-
zyme B was altered so that it cleaved after an aromatic
amino acid instead of Asp by a single Arg226 fi Gly
mutation [15]. A preference for basic amino acids was
seen when Arg226 was replaced by Glu in the same

enzyme [16]. In the present study, we used the same
strategy to investigate the effect of two amino acids on
the extended substrate interactions of HC. We showed
that we could change the cleavage specificity of HC
for position P2¢ of substrates by mutating posi-
tions 143 and 192 in the enzyme. By replacing Arg143
by Gln or Lys192 by Met, which are amino acids com-
monly found in these positions in related rodent chy-
mases, the preference for acidic amino acids at
position P2¢ was markedly reduced, whereas the cleav-
age specificity for all other positions was essentially
unaffected. The basic amino acids at positions 143 and
192 attract and stabilize the interaction of acidic amino
acid side chains at position P2¢. When either of these
positively charged residues was replaced by an amino
acid with an uncharged side chain, a preference for
mainly aliphatic amino acids was observed. However,
a weak preference for acidic P2¢ amino acids still
remained for the two single mutants. In the double
mutant, the preference for negatively charged P2¢
amino acids was lost completely.
In order to put our hypothesis to a stringent test, we
aligned the enzyme-selected peptides with acidic P2¢
amino acids, when an aromatic amino acid could be
aligned at position P1. We then aligned the sequences
according to the cleavage specificity of the wild-type
enzyme, considering the remaining positions. We could
thus be certain that we were not overestimating the
effect of the mutations. However, a large fraction of
the acidic amino acids aligned at position P2¢ belong

to sequences with three or four possible aromatic P1
amino acids. In these cases, it is difficult to know
where the cleavage actually occurred or if they were
selected owing to multiple cleavage sites. Therefore,
the effects of the two single mutants may be slightly
underestimated. On the other hand, when the enzyme-
selected sequences are examined overall, fewer nega-
tively charged amino acids are found among the
sequences cleaved by the HC double mutant than
among those cleaved by the single mutants (nine, 21,
and 21, respectively). This is an indication that the
single mutants are more tolerant of acidic amino acids
in cleavable substrates. More importantly, we can
Acidic amino acids in P2′ (%)
Asp
Glu
HC wt R143Q K192M R143Q +
K192M
0
10
20
30
40
50
60
70
Fig. 5. Distribution of acidic amino acids in position P2¢ of wild-type
(wt) HC and HC mutants. The occurrence of acidic amino acids
aligned in position P2¢ was compared between wild-type HC and
the Arg143 fi Gln mutant, the Lys192 fi Met mutant, and the

Arg143 fi Gln + Lys192 fi Met double mutant. Glu residues in
this position are depicted as open bars, and Asp residues as filled
bars. The occurrence of acidic amino acids in position P2¢ of wild-
type HC was determined in a previous study [4].
M. K. Andersson et al. P2¢ specificity of the human mast cell chymase
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2261
A
B
C
D
E
F
P2¢ specificity of the human mast cell chymase M. K. Andersson et al.
2262 FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS
conclude from our data that the preference of HC for
acidic amino acids at position P2¢ is mediated by the
combined effects of both Arg143 and Lys192. A simi-
lar combined effect has, to our knowledge, never been
identified for a serine protease.
By the use of a new type of recombinant substrate,
we were also able to verify this marked change in
preference for a negatively charged amino acid at
position P2¢ by mutating Arg143 and Lys192. Wild-
type HC was found to cleave the consensus substrate
four-fold to five-fold more efficiently than the substrate
in which the amino acid at position P2¢ had been
exchanged for a Gly. In contrast, the HC double
mutant showed a two-fold to five-fold greater prefer-
ence for the substrate in which position P2¢ was not
negatively charged. These results clearly show that the

double mutant had lost its preference for negatively
charged amino acids, and instead preferred aliphatic or
noncharged amino acids at this position.
The identification of Arg143 and Lys192 as P2¢ pref-
erence-determining amino acids facilitates predictions
about other related chymases. As stated earlier, we
have identified three other chymases with the HC-like
preference for acidic amino acids at position P2¢,
namely the opossum a-chymase, rMCP-5, and mMCP-4
[6,10,11]. All of these chymases have Arg143 and
Lys192. Other chymases that have Arg143 and Lys192
are the a-chymases from the macaque and the baboon,
the sheep MC protease, mMCP-5, and hamster chym-
ase-2. Furthermore, the b-chymases rMCP-3, hamster
chymase-1 and gerbil chymase-1 also have Arg143 and
Lys192. The guinea pig a-chymase was recently cloned
and shown to have Arg143 and Lys192 [17]. In agree-
ment with our prediction, screening of a combinatorial
library indicated a preference for acidic P2¢ amino
acids [17]. The preference for acidic P2¢ amino acids
seems to be highly preserved among the a-chymases.
However, there may be minor exceptions. The dog
a-chymase has Lys143 and Lys192, and according to
our analysis this chymase has only a weak preference
for acidic amino acids at position P2¢ (submitted
manuscript Gallwitz et al., 2009). Apparently, a minor
change from the positively charged Arg to a slightly
smaller but still positively charged side chain of a Lys
at position 143 is enough to partially affect the prefer-
ence for acidic P2¢ amino acids. The gerbil chymase-2

also has Lys143 and Lys192, and may therefore
also have a lower preference for acidic amino acids at
position P2¢.
HC efficiently converts Ang I to Ang II (cleavage of
the Phe8-His9 bond of Ang I), and the structural
requirements for this specificity have been addressed in
several studies. Synergistic interactions of posi-
tions P4–P1, together with the dipeptidyl leaving group
of Ang I, were found to be important for efficient con-
version by HC [18,19]. However, the side chains on the
leaving group of Ang I do not seem to be important
for the selectivity of HC in converting Ang I [19].
Instead, the negatively charged C-terminal carboxyl
group of Ang I probably interacts with the Lys40 side
chain of HC to stabilize the substrate [8]. Lys40 and
Arg143 of HC have previously been analyzed for their
role in the selective Ang I conversion by HC. Analysis
of Lys40 fi Ala and Arg143 fi Gln mutants of HC
showed that Lys40 but not Arg143 contributed to the
high specificity of HC in converting Ang I to Ang II
[20]. The Arg143 fi Gln mutant was actually shown to
be more active than the wild type in converting Ang I,
indicating a minor role or no role at all of Arg143 in
Ang I conversion. The lack of function for Arg143
and Lys192 in Ang I conversion is also substantiated
by the fact that mMCP-1, which has Lys143 and
Met192, and has a P2¢ specificity for Leu, is a good
Ang I converter [21]. Furthermore, the rat vascular
chymase, which has Arg143 and Thr192 and thus does
not have a predicted P2¢ specificity for acidic amino

acids, also has a very good Ang I conversion capability
[22]. However, our results clearly show that Arg143
and Lys192 of HC are of major importance in mediat-
ing the specificity for acidic P2¢ side chains of sub-
strates, but not when the negative charge is situated at
a C-terminal carboxyl group of a P2¢ amino acid of
Fig. 6. Analysis of cleavage specificity by the use of new types of recombinant protein substrate. (A) The overall structure of the recombi-
nant protein substrates used for analysis of the efficiency of cleavage by HC and the HC Arg143 fi Gln + Lys192 fi Met double mutant.
In these substrates, two thioredoxin molecules are positioned in tandem, and the proteins have His6-tags positioned in their C-termini. The
different cleavable sequences are inserted in the linker region between the two thioredoxin molecules by the use of two unique restriction
sites, one BamHI site and one SalI site, which are indicated at the bottom of (A). (B, D) The cleavage of a number of substrates by HC. The
names and sequences of the different substrates are indicated above the pictures of the gels. The times of cleavage in minutes are also
indicated above the corresponding lanes of the different gels. The uncleaved substrates have a molecular mass of  25 kDa, and the
cleaved substrates appear as two closely located bands with a size of 12–13 kDa. (C, E) The cleavage of the same substrates as for HC in
(B) and (D), but now after cleavage with the HC double mutant. (F) The cleavage of a number of the above selected substrates with OC. On
the right side of the figure, the scanned and quantified protein bands are summarized in individual diagrams. The quantification was per-
formed with
IMAGEJ.
M. K. Andersson et al. P2¢ specificity of the human mast cell chymase
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2263
the substrate. The marked difference in the importance
of Lys40, Arg143 and Lys192 in determining substrate
specificity between peptides and long substrates is
striking. This clearly shows the importance of analyz-
ing a broad range of different substrates, with different
biochemical characteristics, when looking for the natu-
ral in vivo substrates. The high degree of conservation
of the preference for negatively charged amino acids at
position P2¢ is also a strong indicator that substrates
other than Ang I are evolutionarily conserved targets

for the MC a-chymases or their rodent counterpart,
the b-chymase mMCP-4.
The search for potential in vivo substrates is being
performed using bioinformatic screening. However, the
identification of these substrates may be challenging,
mainly because of the ability of HC to interact with
different amino acid side chains in each subsite of the
enzyme, which leads to a very large number of poten-
tial substrates during the screening of the full human
proteome. This highlights the importance of factors
other than the extended cleavage specificity in deter-
mining whether a protein will be a biologically signifi-
cant substrate for HC. For example, the local
concentration of the protease and availability of the
potential substrate in the immediate environment are
significant factors. The extended cleavage specificity
predicts those sequences (and hence substrates) that
would be preferentially cleaved within a shorter time
frame. These substrates can theoretically be cleaved
before the protease encounters a protease inhibitor.
The influence of the cleavage sequence position in the
protein is also important. It is likely that surface-
exposed, flexible regions would be cleaved efficiently.
Conversely, nonexposed, more rigid regions would
remain uncleaved, regardless of whether the preferred
sequence for the protease was present or not. In the
event of substrate cleavage, whether this leads to a bio-
logical effect also needs to be determined. With knowl-
edge of the extended cleavage specificity and other
parameters in hand, the search for biologically signifi-

cant substrates for HC is more conceivable.
Experimental procedures
In vitro mutagenesis of HC
The HC wild-type sequence has previously been cloned and
inserted into the pCEP-Pu2 vector for expression in mam-
malian cells [4,23,24]. This vector construct was used for
in vitro mutagenesis of HC. In this construct, the 5¢-end of
the coding region contains a signal sequence followed by a
sequence encoding a His6-tag and an EK-susceptible site
(Asp-Asp-Asp-Asp-Lys). The N-terminal His6-tag and EK
site in the translated protein facilitated purification and
activation of the enzyme.
Mutagenesis of HC was performed using the Quik-
Change II XL Site-Directed Mutagenesis Kit (Stratagene,
La Jolla, CA, USA). The Arg143 fi Gln mutant was pro-
duced by cycling the wild-type sequence in pCEP-Pu2,
according to the manufacturer’s recommendations, using
the following primers: sense primer, 5¢-CGGGTGGCTGG
CTGGGGA
CAGACAGGTGTGTTGAAGC-3¢; and anti-
sense primer, 5¢-GCTTCAACACACCTGT
CTGTCCCCA
GCCAGCCACCCG-3¢. These primers had a melting tem-
perature of 79.5 °C, and the mismatches resulting in
replacement of the Arg codon with a Gln codon are
underlined. To produce the Lys192 fi Met mutant, the
following primers were used: sense primer, 5¢-CAGGAA
GACAAAATCTGCATTTA
TGGGAGACTCTGG-3¢; and
antisense primer, 5¢-CCAGAGTCTCCC

ATAAATGCAGA
TTTTGTCTTCCTG-3¢. These primers had a melting tem-
perature of 78 °C, and the mismatches resulting in replace-
ment of the Lys codon with a Met codon are underlined.
The Arg143 fi Gln + Lys192 fi Met double mutant was
produced by introducing the Lys192 fi Met mutation into
the Arg143 fi Gln mutant. All primers were purchased
from Sigma-Aldrich (Steinheim an Albuch, Germany) in a
PAGE-purified form. Thermal cycling was performed with
PfuUltra high-fidelity DNA polymerase (provided by the
manufacturer). After this PCR step, nonmutated parental
DNA constructs were digested with Dpn1 endonuclease for
1 h at 37 °C. The remaining nondigested and mutated DNA
vector constructs were ethanol precipitated. The salt and
ethanol concentration during precipitation was 75 mm
NaAc (pH 5.2) in 75% ethanol. After precipitation, the
DNA was resuspended in 15 lL of double-distilled H
2
O.
This DNA was then used to transform XL10-Gold ultra-
competent E. coli cells (provided by the manufacturer). All
mutants were sequenced to confirm the inserted mutations
and the absence of unintended additional mutations, by
using an ABI PRISM 3730 DNA Analyzer (Applied Biosys-
tems, Foster City, CA, USA) and vector-specific primers.
Production and purification of recombinant HC
mutants
The vector constructs encoding the HC mutants were trans-
fected into a human embryonic kidney cell line (HEK 293
EBNA) at  80% confluence, using Lipofectamine (Invitro-

gen, Carlsbad, CA, USA) as previously described [23,24].
Selection of transfected cells was initiated by the addition
of 1.5 lgÆmL
)1
puromycin to the cell culture medium
(DMEM supplemented with 5% fetal bovine serum,
50 lgÆmL
)1
gentamicin, and 5 lgÆ mL
)1
heparin). The level
of puromycin was decreased to 0.5 lgÆmL
)1
after  7 days
of selection.
Conditioned medium was collected and centrifuged at
1600 g to remove cell debris, and this was followed by the
P2¢ specificity of the human mast cell chymase M. K. Andersson et al.
2264 FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS
addition of 300 lLofNi
2+
–nitrilotriacetic acid agarose
beads (Qiagen, GmbH, Hilden, Germany) per liter of con-
ditioned medium. After 1 h of incubation with gentle agita-
tion at 4 °C, the beads were pelleted by centrifugation at
135 g and transferred to 1.5 mL reaction tubes (Trefflab,
Degersheim, Switzerland). The collected Ni
2+
–nitrilotriace-
tic acid beads were washed five times with washing buffer

(1 m NaCl, 0.2% Tween in NaCl ⁄ P
i
). Bound protein was
then eluted with elution buffer (100 mm imidazole, 0.2%
Triton X-100 in NaCl ⁄ P
i
). Protein purity and concentration
was estimated by separation on 12.5% SDS ⁄ PAGE gels.
Protein samples were mixed with sample buffer, and b-mer-
captoethanol was added to a final concentration of 5%. To
visualize the protein bands, the gel was stained with Coo-
massie Brilliant Blue.
Activation and further purification of
recombinant HC variants
Approximately 30 lg of each HC mutant was diluted 1 : 2
in double-distilled H
2
O and digested for 5 h at 37 °C with
EKMax EK (Invitrogen), using one unit per 10 lgof
recombinant protease.
In order to remove EK and other impurities, the
EK-digested HC mutants were purified by affinity chroma-
tography on heparin–Sepharose columns as described pre-
viously [13]. PolyPrep Chromatography columns
containing 0.2 mL of heparin–Sepharose beads (Sigma-
Aldrich) were equilibrated with NaCl ⁄ P
i
(pH 7.2). Each
EK-cleaved HC mutant was applied to a column, and this
was followed by washing with 0.3 m NaCl in NaCl ⁄ P

i
and
subsequent elution with 1 m NaCl in NaCl ⁄ P
i
. Enzymatic
activity towards the chromogenic substrate S-2586 (MeO-
Suc-Arg-Ala-Tyr-pNA) (Chromogenix) was measured. Mea-
surements were performed in 96-well microtiter plates with
a substrate concentration of 0.18 mm in 200 lLof
NaCl ⁄ P
i
. S-2586 hydrolysis was monitored spectrophoto-
metrically at 405 nm in a Versamax microplate reader
(Molecular Devices, Sunnyvale, CA, USA). The protein
content of flow through, wash and eluted fractions was
analyzed on SDS ⁄ PAGE gels.
Determination of cleavage specificity with a
phage-displayed nonapeptide library
A library of 5 · 10
7
unique phage-displayed nonameric
peptides was used to determine the cleavage specificity of
the HC mutants, as previously described [6,11,13]. In these
T7 phages, the C-terminus of capsid protein 10 was modi-
fied to contain a nine amino acid random peptide followed
by a His6-tag [13]. An aliquot of the amplified phages
( 10
9
plaque-forming units) was bound to 100 lLof
Ni

2+
–nitrilotriacetic acid beads by their His6-tags for 1 h
at 4 °C under gentle agitation. Unbound phages were
removed by washing 10 times in 1.5 mL of 1 m NaCl and
0.1% Tween-20 in NaCl ⁄ P
i
(pH 7.2), and two subsequent
washes with 1.5 mL of NaCl ⁄ P
i
. The beads were finally
resuspended in 1 mL of NaCl ⁄ P
i
. Activated and heparin–
Sepharose-purified HC mutant ( 0.1 l g) was added to the
resuspended beads and left to digest susceptible phage no-
napeptides under gentle agitation at room temperature
overnight. NaCl ⁄ P
i
without protease was used as control.
Phages with a random peptide that was susceptible to pro-
tease cleavage were released from the Ni
2+
–nitrilotriacetic
acid matrix, and the supernatant containing these phages
was recovered. To ensure that all of the released phages
were recovered, the beads were resuspended in 100 lLof
NaCl ⁄ P
i
(pH 7.2) and the supernatant, after mixing and
centrifugation at 10 000 g, was added to the first superna-

tant. To ensure that the His6-tags had been hydrolyzed on
all phages recovered after protease digestion, 15 l L of fresh
Ni
2+
–nitrilotriacetic acid agarose beads was added to the
combined phage supernatant, and the mixture was agitated
for 15 min and then centrifuged at 135 g. A control elution
of the phages still bound to the beads, using 100 lLof
100 mm imidazole, showed that at least 1 · 10
8
phages were
attached to the matrix during each selection. Ten microli-
ters of the supernatant containing the released phages was
used to determine the amount of phage detached in each
round of selection. Dilutions of the supernatant were plated
in 2.5 mL of 0.6% top agarose containing 300 lLofE. coli
(BLT5615), 100 lL of diluted supernatant, and 100 lLof
100 mm isopropyl thio-b-d-galactoside (IPTG). The remain-
ing volume of the supernatant was added to a 10 mL cul-
ture of BLT5615 (D  0.6). Thirty minutes prior to phage
addition IPTG was added to the bacterial culture to a final
concentration of 1mm to induce production of the native
T7 phage capsid protein. The bacteria lysed  75 min after
phage addition. The lysate was centrifuged at 10 000 g to
remove cell debris, and 500 lL of the phage sublibrary was
added to 100 lL of fresh Ni
2+
–nitrilotriacetic acid beads to
start the next round of selection. After binding of the subli-
brary for 1 h at 4 °C under gentle agitation, the Ni

2+
–
nitrilotriacetic acid beads were washed 15 times in 1.5 mL
of 1 m NaCl and 0.1% Tween-20 in NaCl ⁄ P
i
(pH 7.2), and
then twice in 1.5 mL of NaCl ⁄ P
i
.
Following five or six rounds of selection, 44 plaques for
each HC mutant were isolated from LB plates after plating
in top agarose. Each phage plaque, corresponding to a
phage clone, was dissolved in phage extraction buffer
(100 mm NaCl and 6 mm MgSO
4
in 20 mm Tris ⁄ HCl,
pH 8.0) and vigorously shaken for 30 min in order to
extract the phages from the agarose. The phage DNA was
then amplified by PCR, using primers flanking the variable
region of the gene encoding the modified T7 phage capsid
protein. After amplification, PCR fragments were purified
using the E.Z.N.A Micro Elute Cycle-Pure kit (Omega bio-
tek, Doraville, GA, USA). Purified PCR fragments were
then sequenced (Macrogen Inc., Seoul, Korea) using an
ABI PRISM 3730 DNA Analyzer (Applied Biosystems).
M. K. Andersson et al. P2¢ specificity of the human mast cell chymase
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2265
Generation of a consensus sequence from
sequenced phage inserts
Phage insert sequences were aligned by hand, assuming a

preference for aromatic amino acids at position P1.
Sequences with only one aromatic amino acid were aligned
first, and sequences with more than one possible cleavage
site were then aligned to fit this pattern. Amino acids with
similar characteristics were grouped together as follows:
aromatic amino acids (Phe, Tyr, and Trp); negatively
charged amino acids (Asp and Glu); positively charged
amino acids (Lys and Arg); small aliphatic amino acids
(Gly and Ala); larger aliphatic amino acids (Val, Leu, Ile,
and Pro); and hydrophilic amino acids (Ser, Thr, His, Asn,
Gln, Cys, and Met). The nomenclature of Schechter and
Berger [5] was adopted to designate the amino acids in the
substrate cleavage region, where P1–P1¢ corresponds to the
scissile bond.
Generation of recombinant substrates for
analysis of cleavage specificity
A new type of substrate was developed to verify the results
obtained from the phage display analysis. Two copies of
the E. coli thioredoxin gene were inserted in tandem into
the pET21 vector for bacterial expression (Fig. 6A). In the
C-terminal end, a His6-tag was inserted for purification on
Ni
2+
immobilized metal affinity chromatography columns.
In the linker region, between the two thioredoxin mole-
cules, the different substrate sequences were inserted by
ligating double-stranded oligonucleotides into two unique
restriction sites, one BamHI site and one SalI site (Fig. 6A).
The sequences of the individual clones were verified after
cloning by sequencing of both DNA strains. The plasmids

were then transformed into the E. coli Rosetta gami strain
for protein expression (Novagen; Merck, Darmstadt,
Germany). A 10 mL overnight culture of the bacteria har-
boring the plasmid was diluted 10 times in LB + ampicil-
lin, and grown at 37 °C for 1–2 h until D
600 nm
reached 0.5.
IPTG was then added to a final concentration of 1 mm.
The culture was then grown at 37 °C for an additional 3 h
with vigorous shaking, after which the bacteria were pel-
leted by centrifugation at 1600 g for 12 min. The pellet was
washed once with 25 mL of NaCl ⁄ P
i
+ 0.05% Tween-20.
The pellet was then dissolved in 2 mL of NaCl ⁄ P
i
and soni-
cated six times for 30 s each to open the cells. The lysate
was centrifuged at 10 000 g for 10 min, and the supernatant
was transferred to a new tube. Five hundred microliters of
Ni
2+
–nitrilotriacetic acid slurry (50 : 50) (Qiagen, Hilden,
Germany) was added, and the sample was slowly rotated
for 45 min at room temperature. The sample was then
transferred to a 2 mL column, and the supernatant was
allowed to slowly pass through the filter, leaving the Ni
2+
–
nitrilotriacetic acid beads with the bound protein in the col-

umn. The column was then washed four times with 1 mL
of washing buffer (NaCl ⁄ P
i
, 0.05% Tween-20, 10 mm imid-
azole, and 1 m NaCl). Elution of the protein was performed
by adding 150 lL of elution buffer followed by five 300 lL
fractions of elution buffer (NaCl ⁄ P
i
, 0.05% Tween-20, and
100 mm imidazole). Each fraction was collected individu-
ally. Ten microliters from each of the eluted fractions was
then mixed with one volume of 2 · sample buffer and 1 lL
of b-mercaptoethanol, and heated for 3 min at 80 °C. The
samples were analyzed on an SDS Bis ⁄ Tris 4–12% PAGE
gel, and the second and third fractions, which contained
the most protein, were pooled. The protein concentration
of the combined fractions was determined with a Bio-Rad
DC Protein assay kit (Bio-Rad Laboratories Hercules, CA,
USA). Approximately 60 lg of recombinant protein was
added to each 120 lL of cleavage reaction (in NaCl ⁄ P
i
).
Twenty microliters from this tube was removed before addi-
tion of the enzyme (the 0 min time point). The active
enzyme was then added ( 35 ng of HC or the HC double
mutant), and the reaction was kept at room temperature
for the entire experiment. Twenty-microliter samples were
removed at the indicated time points (15, 45, and 150 min),
and prevented from reacting further by addition of one vol-
ume of 2 · sample buffer. One microliter of b-mercaptoeth-

anol was then added to each sample, and this was followed
by heating for 3 min at 80 °C. Twenty microliters from
each of these samples was then analyzed on 4–12% precast
SDS ⁄ PAGE gels (Invitrogen). The gels were stained over-
night in colloidal Coomassie staining solution, and
destained for several hours according to previously
described procedures [25]. The intensities of the individual
bands on the gel were determined from scanned high-reso-
lution pictures by densitometric scanning of the gels and
using imagej (rsb.info.nih.gov ⁄ nih-image ⁄ ). In order to
obtain good estimates of the differences in activity towards
different substrates, different concentrations of the enzyme
were used in several individual experiments. The combined
results from these different gels were then used to obtain a
good estimate of the differences in activity against the vari-
ous substrates.
Acknowledgement
This study was supported by grants from the Swedish
National Research Council (VR-NT).
References
1 Chandrasekharan UM, Sanker S, Glynias MJ, Karnik
SS & Husain A (1996) Angiotensin II-forming activity
in a reconstructed ancestral chymase. Science 271, 502–
505.
2 Gallwitz M, Reimer JM & Hellman L (2006) Expansion
of the mast cell chymase locus over the past 200 million
years of mammalian evolution. Immunogenetics 58, 655–
669.
P2¢ specificity of the human mast cell chymase M. K. Andersson et al.
2266 FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS

3 Pejler G, Abrink M, Ringvall M & Wernersson S
(2007) Mast cell proteases. Adv Immunol 95 , 167–255.
4 Andersson MK, Enoksson M, Gallwitz M & Hellman
L (2009) The extended substrate specificity of the
human mast cell chymase reveals a serine protease with
well-defined substrate recognition profile. Int Immunol
21, 95–104.
5 Schechter I & Berger A (1967) On the size of the active
site in proteases. I. Papain. Biochem Biophys Res
Commun 27, 157–162.
6 Andersson MK, Karlson U & Hellman L (2008) The
extended cleavage specificity of the rodent beta-chymas-
es rMCP-1 and mMCP-4 reveal major functional simi-
larities to the human mast cell chymase. Mol Immunol
45, 766–775.
7 McGrath ME, Mirzadegan T & Schmidt BF (1997)
Crystal structure of phenylmethanesulfonyl fluoride-
treated human chymase at 1.9 A. Biochemistry 36,
14318–14324.
8 Pereira PJ, Wang ZM, Rubin H, Huber R, Bode W,
Schechter NM & Strobl S (1999) The 2.2 A crystal
structure of human chymase in complex with succinyl-
Ala-Ala-Pro-Phe-chloromethylketone: structural expla-
nation for its dipeptidyl carboxypeptidase specificity.
J Mol Biol 286, 163–173.
9 Yamamoto D, Shiota N, Takai S, Ishida T, Okunishi H
& Miyazaki M (1998) Three-dimensional molecular
modeling explains why catalytic function for angioten-
sin-I is different between human and rat chymases.
Biochem Biophys Res Commun 242, 158–163.

10 Reimer JM, Enoksson M, Samollow PB & Hellman L
(2008) Extended substrate specificity of opossum chym-
ase – implications for the origin of mast cell chymases.
Mol Immunol 45, 2116–2125.
11 Karlson U, Pejler G, Tomasini-Johansson B &
Hellman L (2003) Extended substrate specificity of rat
mast cell protease 5, a rodent alpha-chymase with
elastase-like primary specificity. J Biol Chem 278,
39625–39631.
12 Andersson MK, Pemberton AD, Miller HR & Hellman
L (2008) Extended cleavage specificity of mMCP-1, the
major mucosal mast cell protease in mouse – high speci-
ficity indicates high substrate selectivity. Mol Immunol
45, 2548–2558.
13 Karlson U, Pejler G, Froman G & Hellman L (2002)
Rat mast cell protease 4 is a beta-chymase with unusu-
ally stringent substrate recognition profile. J Biol Chem
277, 18579–18585.
14 Harris JL, Peterson EP, Hudig D, Thornberry NA &
Craik CS (1998) Definition and redesign of the extended
substrate specificity of granzyme B. J Biol Chem 273,
27364–27373.
15 Caputo A, James MN, Powers JC, Hudig D & Bleack-
ley RC (1994) Conversion of the substrate specificity of
mouse proteinase granzyme B. Nat Struct Biol 1, 364–
367.
16 Caputo A, Parrish JC, James MN, Powers JC &
Bleackley RC (1999) Electrostatic reversal of serine
proteinase substrate specificity. Proteins 35, 415–424.
17 Caughey GH, Beauchamp J, Schlatter D, Raymond

WW, Trivedi NN, Banner D, Mauser H & Fingerle J
(2008) Guinea pig chymase is leucine-specific: a novel
example of functional plasticity in the chymase ⁄ gran-
zyme family of serine peptidases. J Biol Chem 283,
13943–13951.
18 Kinoshita A, Urata H, Bumpus FM & Husain A (1991)
Multiple determinants for the high substrate specificity
of an angiotensin II-forming chymase from the human
heart. J Biol Chem 266, 19192–19197.
19 Sanker S, Chandrasekharan UM, Wilk D, Glynias MJ,
Karnik SS & Husain A (1997) Distinct multisite syner-
gistic interactions determine substrate specificities of
human chymase and rat chymase-1 for angiotensin II
formation and degradation. J Biol Chem 272, 2963–
2968.
20 Muilenburg DJ, Raymond WW, Wolters PJ & Caughey
GH (2002) Lys40 but not Arg143 influences selectivity
of angiotensin conversion by human alpha-chymase.
Biochim Biophys Acta
1596, 346–356.
21 Saito K, Muto T, Tomimori Y, Imajo S, Maruoka H,
Tanaka T, Yamashiro K & Fukuda Y (2003) Mouse
mast cell protease-1 cleaves angiotensin I to form angio-
tensin II. Biochem Biophys Res Commun 302, 773–777.
22 Guo C, Ju H, Leung D, Massaeli H, Shi M & Rabinov-
itch M (2001) A novel vascular smooth muscle chymase
is upregulated in hypertensive rats. J Clin Invest 107,
703–715.
23 Vernersson M, Ledin A, Johansson J & Hellman L
(2002) Generation of therapeutic antibody responses

against IgE through vaccination. FASEB J 16, 875–877.
24 Hallgren J, Karlson U, Poorafshar M, Hellman L &
Pejler G (2000) Mechanism for activation of mouse
mast cell tryptase: dependence on heparin and acidic
pH for formation of active tetramers of mouse mast cell
protease 6. Biochemistry 39, 13068–13077.
25 Neuhoff V, Arold N, Taube D & Ehrhardt W (1988)
Improved staining of proteins in polyacrylamide gels
including isoelectric focusing gels with clear background
at nanogram sensitivity using Coomassie Brilliant Blue
G-250 and R-250. Electrophoresis 9, 255–262.
M. K. Andersson et al. P2¢ specificity of the human mast cell chymase
FEBS Journal 277 (2010) 2255–2267 ª 2010 The Authors Journal compilation ª 2010 FEBS 2267