The molecular structure and catalytic mechanism of a novel carboxyl peptidase from Scytalidium lignicolum

Abstract

The molecular structure of the pepstatin-insensitive carboxyl peptidase from Scytalidium lignicolum, formerly known as scytalidopepsin B, was solved by multiple isomorphous replacement phasing methods and refined to an R factor of 0.230 (R(free) = 0.246) at 2.1-A resolution. In addition to the structure of the unbound peptidase, the structure of a product complex of cleaved angiotensin II bound in the active site of the enzyme was also determined. We propose the name scytalidocarboxyl peptidase B (SCP-B) for this enzyme. On the basis of conserved, catalytic residues identified at the active site, we suggest the name Eqolisin for the enzyme family. The previously uninvestigated SCP-B fold is that of a beta-sandwich; each sheet has seven antiparallel strands. A tripeptide product, Ala-Ile-His, bound in the active site of SCP-B has allowed for identification of the catalytic residues and the residues in subsites S1, S2, and S3, which are important for substrate binding. The most likely hydrolytic mechanism involves nucleophilic attack of a general base (Glu-136)-activated water (OH(-)) on the si-face of the scissile peptide carbonylcarbon atom to form a tetrahedral intermediate. Electrophilic assistance and oxyanion stabilization is provided by the side-chain amide of Gln-53. Protonation of the leaving-group nitrogen is accomplished by the general acid function of the protonated carboxyl group of Glu-136.

Full text

The molecular structure and catalytic mechanism of a
novel carboxyl peptidase from
Scytalidium lignicolum
Masao Fujinaga*, Maia M. Cherney*, Hiroshi Oyama
†
, Kohei Oda
†
, and Michael N. G. James*
‡
*Canadian Institutes of Health Research Group in Protein Structure and Function, Department of Biochemistry, University of Alberta, Edmonton, Alberta,
Canada T6G 2H7; and
†
Department of Applied Biology, Faculty of Textile Science, Kyoto Institute of Technology, Sakyo-ku, Kyoto 6068585, Japan
Communicated by David R. Davies, National Institutes of Health, Bethesda, MD, January 22, 2004 (received for review December 1, 2003)
The molecular structure of the pepstatin-insensitive carboxyl pepti-
dase from Scytalidium lignicolum, formerly known as scytalidopepsin
B, was solved by multiple isomorphous replacement phasing methods
and refined to an R factor of 0.230 (R
free
ⴝ 0.246) at 2.1-Å resolution.
In addition to the structure of the unbound peptidase, the structure
of a product complex of cleaved angiotensin II bound in the active site
of the enzyme was also determined. We propose the name scytali-
docarboxyl peptidase B (SCP-B) for this enzyme. On the basis of
conserved, catalytic residues identified at the active site, we suggest
the name Eqolisin for the enzyme family. The previously uninvesti-
gated SCP-B fold is that of a
␤
-sandwich; each sheet has seven
antiparallel strands. A tripeptide product, Ala-Ile-His, bound in the
active site of SCP-B has allowed for identification of the catalytic
residues and the residues in subsites S1, S2, and S3, which are
important for substrate binding. The most likely hydrolytic mecha-
nism involves nucleophilic attack of a general base (Glu-136)-acti-
vated water (OH
ⴚ
)onthesi-face of the scissile peptide carbonyl-
carbon atom to form a tetrahedral intermediate. Electrophilic
assistance and oxyanion stabilization is provided by the side-chain
amide of Gln-53. Protonation of the leaving-group nitrogen is accom-
plished by the general acid function of the protonated carboxyl group
of Glu-136.
P
epstatin-insensitive carboxyl peptidases (1–8) originally found
by Murao and Oda can be classified into two groups, bacterial
and fungal carboxyl peptidases. The 3D structures of the two
members of the bacterial carboxyl peptidases have recently been
determined (9, 10). From the tertiary folds, it was clear that these
enzymes are members of the subtilisin family. Each enzyme has a
glutamate residue acting as the general base instead of the histidine
present in the subtilisins. The environments of the carboxylates
contribute to the low pH optima of these enzymes.
The fungal pepstatin-insensitive carboxyl peptidases (5) are
represented by several enzymes, among them the enzyme that is the
subject of this article. Formerly, this enzyme has been called
Scytalidopepsin-B because of its low pH optimum and preponder-
ance of acidic residues (8). It is isolated from Scytalidium lignicolum
ATCC24568. We have chosen to rename this enzyme scytalido-
carboxyl peptidase-B (SCP-B) to reflect its fungal source and
active-site carboxyl groups. SCP-B is synthesized as a precursor
consisting of two regions: an amino-terminal preprosegment of 54
amino acid residues and a mature enzyme consisting of 206 amino
acid residues (11). The mature enzyme has no sequence similarity
to the well known pepsin-like or retroviral aspartic peptidases, but
it has significant similarity to the other fungal pepstatin-insensitive
carboxyl peptidases (Fig. 1).
SCP-B and the other carboxyl peptidases described here have
been classified in family A4 of the aspartic endopeptidases in the
MEROPS (http:兾兾merops.sanger.ac.uk). database. On the basis of
our structural results, however, we prefer to name this family the
Eqolisins (pronounced ‘‘echo’’-lisin) derived from the active-site
residues, glutamic acid (E) and glutamine (Q). We suggest that the
eqolisins constitute a new peptidase family that is not part of the
aspartic peptidases.
SCP-B works best in acidic conditions, having an optimal pH of
2.0 with casein as substrate, suggesting that carboxyl groups are
intimately involved in the catalytic function. SCP-B is not inhibited
by pepstatin, acetyl-pepstatin, nor by diazoacetyl-
DL-norleucine
methyl ester, but it is inhibited by 1,2-epoxy-3-(p-nitrophenoxy)pro-
pane. SCP-B cleaves the oxidized insulin B chain at Tyr-26–Thr-27,
Phe-24–Phe-25, and other positions after Tyr and Glu. It has also
been shown to cleave the His-6–Pro-7 bond of angiotensin II and,
to a lesser extent, the bond between Tyr-4 and Ile-5 of this
octapeptide (15).
Several studies have been carried out to identify the catalytic
residues of the eqolisins. Tsuru and colleagues (16, 17) reported
that Glu-53 and Asp-98, residues in SCP-B, were candidates for its
catalytic residues. The nucleotide sequence of the SCP-B gene,
elucidated after these studies, however, revealed that Glu-53 was in
fact a glutamine residue (18). Furthermore, it turned out that
Asp-98 is not conserved in any of the other members of this
peptidase family. Takahashi and colleagues (19) reported from
mutational data that Glu-219 and Asp-123 (sequence numbers in
the mature enzyme, Glu-149 and Asp-53) in AnCPA were involved
in the catalytic function. These residues are conserved in the SCP-B
molecule as Glu-136 and Asp-43, respectively. Thus, without struc-
tural data, the catalytic mechanism for this unique family of
peptidases is still shrouded in mystery.
We have solved the crystal structures of SCP-B in the unbound
form and in a form complexed with hydrolytic products of angio-
tensin II. On the basis of these 3D structures and of sequence and
biochemical data, it was determined that SCP-B has a catalytic dyad
consisting of residues Gln-53 and Glu-136. These two residues and
the surrounding segments of polypeptide chain are highly con-
served in all fungal pepstatin-insensitive carboxyl peptidases of the
eqolisin family (Fig. 1). From the results derived here, the eqolisins
should occupy a more distinct position in the overall classification
of peptidases (20).
Methods
Crystallization and Data Collection. Lyophilized SCP-B was dissolved
in water at 20 mg兾ml. Crystals were grown at room temperature by
the hanging-drop method from 42% saturated ammonium sulfate兾
0.1 M sodium acetate buffer (pH 4.0)兾10% (vol/vol) ethylene
glycol. Before data collection, crystals of SCP-B were transferred
for several seconds into a cryosolution containing 30% glycerol,
45% saturated ammonium sulfate, and 0.1 M sodium acetate (pH
4.0) then frozen in the cryostream. Heavy-atom derivative crystals
were obtained by soaking native crystals in 1.5 mM mersalyl for 4 h
or in 3 mM uranyl acetate overnight before freezing and data
collection. For soaking a potential substrate into the crystals, a drop
consisting of 3
␮
l of 10 mM angiotensin II in water and 3
␮
lofthe
crystallization solution was equilibrated for several hours and an
SCP-B crystal was soaked in it overnight. The diffraction data were
collected the next day from a flash-frozen crystal.
Abbreviations: SCP-B, scytalidocarboxyl peptidase B; AnCPA, Aspergillus niger carboxyl
peptidase A.
Data deposition: The atomic coordinates have been deposited in the Protein Data Bank,
www.pdb.org (PDB ID codes 1S2B and 1S2K).
‡
To whom correspondence should be addressed. E-mail: michael.james@ualberta.ca.
© 2004 by The National Academy of Sciences of the USA
3364–3369
兩
PNAS
兩
March 9, 2004
兩
vol. 101
兩
no. 10 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0400246101

All data except for the native set were collected on a Rigaku
RAXIS-IV
⫹⫹
area detector with CuK
␣
radiation generated by a
Rigaku RU-H3R rotating anode generator (Table 1). The native
data were collected on beam line 9-1 at the Stanford Synchrotron
Radiation Laboratory. All data were processed with the
HKL suite
of programs (21).
Structure Determination and Refinement. The structure was solved
by the multiple isomorphous replacement method from data col-
lected from the mersalyl and uranyl heavy-atom derivatives by using
the program
SOLVE (22). SOLVE assigned one unique heavy-atom
site for each of the derivatives. The phases derived from the
multiple isomorphous replacement procedure resulted in a figure-
of-merit (FOM) of 0.35. Subsequently, density modification carried
out by the program
RESOLVE (23, 24) increased the FOM to 0.68.
Examination of the electron density computed from the phase set
at this stage showed clearly interpretable molecular features. Thus,
these phases and the corresponding electron density map were used
for automated chain tracing by the program
WARPNTRACE (25). The
program was able to fit 168 residues in nine chain segments.
Substitution of the correct sequence and additional model building
was done manually with
XTALVIEW (26). The program CNS (27) was
used for refinement of both the native and product-bound struc-
tures (Table 1). The refined coordinates have been deposited in the
Protein Data Bank with ID codes 1S2B and 1S2K.
The crystal quality of SCP-B was variable. Despite the relatively
high resolution recorded for the native data set, the quality of the
data in the high-resolution shells was poor. Overall the mean I兾
␴
(I)
is 8.1; it is much worse (1.3) for the data in the range from 2.1 Å
to 1.9 Å. The value of R
merge
is also large for the same range of
data; overall, the data are only 89% complete. For these reasons,
the data were truncated to 2.1 Å for use in the refinement. Very
likely the poor data quality is the cause for the disappointingly high
R factors for the unbound enzyme (Table 1). The data for the
product-bound form of SCP-B was not truncated in resolution, as
the I兾
␴
(I) was much larger in the high-resolution shell (2.07 to 2.0
Å) than that of the native data set.
Results
SCP-B is the founding member of the eqolisins, a new family of
proteolytic enzymes (Fig. 1). The overall structure of the enzyme is
Table 1. Summary of crystallographic data
Native Uranyl Mercury Angiotensin II
Crystals
Space group P6
3
22 P6
3
22 P6
3
22 P6
3
22
Cell parameters
a, b 108.553 108.601 108.648 109.369
c 114.149 113.880 114.312 113.830
X-ray SSRL BL 9-1,*
␭
⫽ 0.979 Å
Cu K
␣
Cu K
␣
Cu K
␣
Resolution, Å 20–1.9 (1.97–1.90) 47–2.4 (2.49–2.40) 36–2.4 (2.49–2.40) 27–2.0 (2.07–2.0)
R
merge
†
5.8 (48.2) 9.3 (42.6) 9.3 (43.7) 5.0 (41.3)
Completeness, % 96.5 (88.8) 93.9 (89.3) 99.9 (100) 99.9 (100)
Mean I兾
␴
(I) 8.1 14.2 13.6 19.2
Refinement
Resolution, Å 20–2.1 27–2.0
No. of protein atoms 1,450 1,450
No. of water molecules 167 121
No. of ligand atoms — 37
R factor
†
(R
free
)*
§
22.97 (24.57) 23.35 (25.58)
No. of reflections 18,539 (1,032) 26,254 (1,422)
rms deviation bond, Å 0.006 0.006
rms deviation angle, ° 1.4 1.4
*SSRL BL 9-1, Stanford Synchrotron Radiation Laboratory (Stanford University, Stanford, CA) beam line 9-1.
†
R
merge
⫽ 100 兺 兩I
h
⫺具I典兩兾兺I
h
.
‡
R factor ⫽ 100 兺 储F
obs
兩⫺兩 F
calc
储兾兺 兩F
obs
兩 calculated for all observed data.
§
R
free
⫽ R factor for ⬇5% of the randomly chosen unique reflections not used in the refinement.
Fig. 1. CLUSTALW multiple-sequence alignment of the eqolisins. The sequence
numbering for each mature enzyme is indicated. The enzymes in this align-
ment are SCP-B (11), Aspergillus niger carboxyl peptidase A (AnCPA) (12),
Sclerotinia sclerotiorum carboxyl peptidase (SSCP) (13), Cryphonectria para-
sitica peptidases B and C (EAPB and EAPC) (14), and Talaromyces emersonii
carboxyl peptidase (TECP) (GenBank accession no. AF439998). The
␤
-strands
constituting the secondary structure of SCP-B determined herein are indicated
by numbered arrows above the sequence of SCP-B. Amino acid color-coding is
as follows: green, neutral hydrophilic side chains; red, hydrophobic side
chains; blue, negatively charged side chains; and magenta, positively charged
side chains. The catalytic residues Gln-53 and Glu-136 are highlighted in black.
Asterisks denote residues that are identical in all six sequences; colons or
periods denote highly conserved residues. The 11 residues in the box in the
AnCPA sequence are removed in the mature enzyme (12).
Fujinaga et al. PNAS
兩
March 9, 2004
兩
vol. 101
兩
no. 10
兩
3365
BIOCHEMISTRY

that of a
␤
-sandwich formed from two antiparallel
␤
-sheets; each
sheet consists of seven strands (Fig. 2 a and b). A
BLAST search with
the SCP-B sequence has identified five additional members of the
eqolisins (Fig. 1). Each of these enzymes is synthesized as a
preproprecursor protein. The preprosegments are ⬇55 amino acids
in length, and the prosegments are rich in positively charged
residues. This observation is very similar to the charge distribution
in the zymogens of the aspartic peptidases (28). The mature
eqolisins have a preponderance of negatively charged residues (e.g.,
SCP-B has 32 Asp plus Glu residues) over positively charged
residues (SCP-B has only 2 Lys and 1 Arg). Clearly, this distribution
is one of the determining factors for the low pI of SCP-B (3.0).
The mature enzymes have several highly conserved segments of
polypeptide chain (in SCP-B numbering: residues 4–10, 38–46,
50–58, 63–73, 81–89, 130–140, and 150–157; Fig. 1). From Fig. 2c,
it can be seen that many of these regions of conserved residues are
clustered on the upper
␤
-sheet surrounding the active site and the
substrate-binding site. The high level of sequence identity and the
clustering of the conserved regions on the upper
␤
-sheet that
comprises the active site and substrate-binding region confirm that
these enzymes are all related members of the eqolisin family.
SCP-B has three disulfide bridges, not all of which are conserved
in the other members of the eqolisins. The most highly conserved
disulfide is that between Cys-47 and Cys-127 (absent in the C
enzyme from C. parasitica). In SCP-B this disulfide surrounds the
buried Asp-43, a residue that is highly conserved in all members of
the family. Another disulfide is unique to SCP-B (Cys-141 to
Cys-148). It is at the base of the hairpin loop (green arrows in Fig.
2b) that is present only in SCP-B. This loop has been deleted in the
other carboxyl peptidases of the eqolisin family (residues Asn-144
to Glu-149) (see Figs. 1 and 2c).
The isoelectric point of SCP-B is in the acidic range (pI ⫽ 3.0).
Most of the carboxylate groups in SCP-B are solvent-exposed, so
they would not contribute to such a low pI. However, several
carboxylate residues have depressed pKa values either as a result of
being involved in carboxyl–carboxylate interactions (29) (i.e., the
sharing of one proton by two carboxyl groups) or as a result of being
buried in the core of the enzyme. The carboxyl groups of Asp-57
and Asp-65 share a hydrogen bond (OOO distance of 2.5 Å) and
likely have one of the two pKa values of ⬇1.5. The carboxylate of
Asp-43 is buried and negatively charged, because it is the recipient
of four hydrogen bonds from the neighboring main-chain NH
groups from residues Cys-127 to Asn-130, inclusive. The environ-
ments of these two carboxyl groups would contribute strongly to
the low pI value of SCP-B.
The Unbound Active Site of SCP-B. From the results of the biochem-
ical, kinetic, and mutagenic studies that have been carried out on
SCP-B and AnCPA, it is most likely that residues Glu-136 and
Gln-53 are key catalytic residues. These residues are located on the
concave surface of the upper
␤
-sheet (Fig. 2a) and are separated by
a closest approach of 5.03 Å between atoms Glu-136 O␧
2
and
Gln-53 O␧
1
. A water molecule, HOH169, forms a hydrogen bond
to each of these oxygen atoms in the native enzyme (Table 2 and
Fig. 3). By analogy with the aspartic peptidases’ active sites (30), we
propose that this water is the nucleophilic water in the hydrolytic
mechanism of the eqolisins. For the side-chain carboxylate of
Glu-136 to act as a general base, it must be negatively charged. The
negative charge on the carboxyl group of Glu-136 is likely, because
it is the recipient of three hydrogen bonds (Fig. 3b), one from
HOH169 (3.0 Å), one from the N␧
1
of Trp-39 (2.9 Å), and one from
the N
␦
2
atom of Asn-5 (3.0 Å). Glu-136 is only partially solvent-
accessible. Therefore, Glu-136 is unlikely to enter into a covalent
acyl-enzyme intermediate (an acid anhydride) by direct nucleo-
philic attack on the carbonyl-carbon of the scissile peptide. As with
the aspartic peptidases, no space exists for such a close substrate
approach.
The side chain of Gln-53 is wedged between the indole ring of
Fig. 2. (a) Ribbon representation of SCP-B. The secondary structure assign-
ment was determined by Database of Secondary Structure of Proteins (31).
This drawing and Figs. 2c,3–5, and 6b were produced by the program
PYMOL
(32). The active-site residues, Gln-53 and Glu-136, are shown in stick repre-
sentation. The upper sheet is slate, and the lower sheet is red. (b) A topological
diagram representing the secondary structure in SCP-B. The upper sheet
harboring the active site is blue, and the lower sheet is red. The loop unique
to SCP-B that is adjacent to the active site is shown in green. Seven loops cross
from the upper to the lower sheet or from the lower to the upper sheet in this
topography. (c) A representation of the regions of SCP-B that are highly
conserved among the six members of the eqolisin family. Those segments of
the chain corresponding to these highly conserved regions of SCP-B (asterisks
in Fig. 1) are green.
3366
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0400246101 Fujinaga et al.

Trp-67 and the side chain of Ile-51 (Fig. 3b). The amide N␧
2
of
Gln-53 donates a hydrogen bond to the O␧
1
of Glu-69 (2.9 Å). The
O␧
1
carbonyl oxygen of Gln-53 receives a hydrogen bond from the
nucleophilic water, HOH169 (2.9 Å). In addition, the amide
nitrogen N␧
2
of Gln-53 makes a hydrogen bond to another water,
HOH304 (2.7 Å). This latter site is postulated as the binding site for
the carbonyl-oxygen atom of the scissile peptide bond of a substrate,
thereby dictating the
␺
conformational angle of the P1 residue of
a bound substrate.
The carboxyl–carboxylate interaction between Asp-65 and
Asp-57 is at the distal end of the S1 pocket; it provides a net negative
charge for positively charged P1 side chains (Fig. 3b). The side chain
O
␦
1
of Asp-57 receives two other hydrogen bonds, one from the
O
␥
1
atom of Thr-37 (2.8 Å) and the other from the N␧
1
atom of
Trp-67 (2.8 Å). All the above-mentioned hydrogen bonds are
between residues that are conserved in all six known members of
the eqolisin family (Fig. 1).
The Binding of Hydrolytic Products of Angiotensin II to SCP-B. To
identify the active site of SCP-B and to visualize the key residues
that are involved in catalysis, angiotensin II (Asp-Arg-Val-Tyr-Ile-
His-Pro-Phe) was soaked into pregrown crystals of SCP-B; diffrac-
tion data were collected to 2.0-Å resolution (Table 1). A difference
electron density map (Fig. 4a) showed two separate regions of
positive density in the substrate-binding cleft of SCP-B. One was
interpreted as a tripeptide, Ala-Ile-His (Fig. 4), and the other was
interpreted as a single tyrosine (not shown). One would have
expected a hexapeptide to remain as a product complex in the
enzymes’ active site; however, the interpretable regions of differ-
ence density corresponded to a tripeptide Tyr-Ile-His with the Tyr
side chain of this tripeptide not visible beyond C
␤
. It is not at all
clear from where the single tyrosine derived, but a tyrosine is in
angiotensin II that with appropriate enzyme-catalyzed cleavages
could be the source.
The atomic model of the refined product complex included the
tripeptide, Ala-Ile-His, the single tyrosine, and 200 residues of
SCP-B (six of the residues in the segment Tyr-71 to Gly-80 are
disordered). The resulting electron density for the P1 His of the
product is very clear (Fig. 4a), but the side chain of the Ile in the
P2 position is in weak electron density and the atoms have B factors
in the range of 60 Å
2
. The tripeptide binds in an extended
conformation with the NH and CO of the P2 Ile, making antipa-
rallel hydrogen-bonding interactions with the CO and NH of
Glu-139 (Fig. 4b). The OXT atom of the P1 His carboxyl group at
the C terminus of the tripeptide forms a hydrogen bond with the
side-chain carboxyl group of Glu-136 (2.6 Å). The distance from
Gln-53 O␧
1
to P1 His OXT is 3.4 Å and is likely a van der Waals
contact rather than a hydrogen bond. Thus, the simplest interpre-
tation for these interactions would have the side-chain carboxyl
group of Glu-136 protonated and a second hydrogen bond to the
P1 His O donated by Gln-53 N␧
2
(3.18 Å). The side-chain imida-
zolium group of the P1 His packs between the phenyl side chain of
Phe-138 and the indole ring of Trp-67 (Fig. 4b). Both residues are
highly conserved among the six known members of this family. In
addition, the N␧
2
atom of the P1 His imidazolium group forms a
hydrogen bond to the carboxylate of Asp-57 (3.1 Å).
The protein inhibitors of the serine peptidases [e.g., the third
Table 2. Hydrogen bonding in the active site of SCP-B
Atoms
Distance,
Å Atoms
Distance,
Å
Glu136 O
␧2
ѧHOH169 3.0 Trp6 N
␧1
ѧHOH212 3.0
Gln53 O
␧1
ѧHOH169 2.9 Asp57 O
␦
2
ѧAsp65 O
␦
2
2.6
Glu136 O
␧2
ѧTrp39 N
␧1
2.9 Asp57 O
␦
1
ѧThr37 O
␥
1
2.8
Glu136 O
␧1
ѧAsn5 N
␦
2
3.0 Trp67 N
␧1
ѧAsp57 O
␦
1
2.8
Gln53 N
␧2
ѧGlu69 O
␧1
2.9 Asp65 O
␦
1
ѧTyr59 OH 2.7
Gln53 N
␧2
ѧHOH304 2.7 HOH169ѧHOH304 2.8
Fig. 3. (a) Electron density associated with residues that constitute the active
site of SCP-B. The electron density was calculated by using coefficients
(2m兩F
o
兩⫺D兩F
c
兩), and calculated phases (
␣
c
) (33); the map is contoured at 1
␴
within 3.5 Å of the active-site residues and solvent molecules. (b) Residues of
the active site of the unbound form of SCP-B are represented in stereo; oxygen
atoms are red, nitrogen atoms are blue, and carbon atoms are slate. The water
molecules are shown as red spheres; hydrogen bonds between polar side-
chain atoms and to waters are shown as dotted lines. The hydrogen bond
distances are given in Table 2.
Fig. 4. (a) Electron density associated with the product Ala-Ile-His-COO
⫺
and
the two catalytic residues Gln-53 and Glu-136. The details of the contour levels
are as given in Fig. 3a.(b) The environment of the angiotensin II hydrolysis
product Ala-Ile-His (orange) is shown. The C terminus of the tripeptide cor-
responds to a cleavage site in Angiotensin II at the His–Pro bond, and the
tripeptide indicates the binding mode of three residues of the N-terminal
segment of the substrate. Hydrogen bonds between groups on the enzyme
and on the tripeptide are indicated by dotted lines.
Fujinaga et al. PNAS
兩
March 9, 2004
兩
vol. 101
兩
no. 10
兩
3367
BIOCHEMISTRY

domains of avian ovomucoids (34, 35)] bind to their cognate
enzymes in the manner of good substrates. The P1 residues of these
inhibitors have average main-chain conformational angles of
␾
⬇
⫺118°,
␺
⬇⫹35°, and side chain
␹
1
⬇⫺65°. The nucleophilic attack
of the serine O
␥
on the carbonyl-carbon atom of the substrate takes
place on the re-face of the scissile peptide in those enzymes. The
resulting tetrahedral intermediate would have a
␺
angle of ⬇⫹60°,
a staggered conformation. In the eqolisins, the direction of nucleo-
philic attack of water HOH169 is on the si-face of the scissile
peptide (see below). In the product complex with SCP-B, the
conformational angles for the P1 His are
␾
⫽⫺112°,
␺
(pseudo) ⫽
⫺6°, and side chains
␹
1
⫽⫺49° and
␹
2
⫽⫹77°. The values of
␹
1
and
␹
2
are well within the range of values observed for histidine side
chains.
From a series of fluorescence resonance energy transfer
experiments on oligopeptide substrates, the preference for the
P1 residue was determined to be tyrosine followed closely by
histidine (K.O. and K. Takada, unpublished results). In addition,
these substrates showed preferences at the P2 and P3 positions
for tyrosine and lysine or arginine, respectively. The preference
for positively charged residues at P2 and P3 can be understood
from the overall net negative charge on SCP-B due to the
preponderance of Asp and Glu residues in the molecule. The
most favored substrate was cleaved at the Tyr–Ala bond in the
tetrapeptide sequence, Arg-Tyr-Tyr-Ala.
Two main reasons exist for why pepstatin is a poor inhibitor of
the eqolisins. Modeling a leucine side chain at the P1 position
indicates that Leu would not fit well sterically or electrostatically.
Also, the 3-OH group of a statine residue in pepstatin has the wrong
stereochemical configuration.
Mutational data from the closely related aspergillopepsin II
(A. niger carboxyl proteinase-A or AnCPA) suggested that
Asp-123 and Glu-219 (mature numbering, Asp-53 and Glu-149)
are the catalytic residues (20, 36). The suggestion that Asp-53
(Asp-43 in SCP-B) is a catalytic residue is better interpreted as
a disruption of the folding pathway of the enzyme caused by the
mutation of Asp-53. As pointed out, in SCP-B, Asp-43 is in a
highly conserved region (Fig. 1), but it is distant from the active
site and buried in the core of the enzyme. In addition, the
carboxylate of Asp-43 is most likely negatively charged, because
it is the recipient of four hydrogen bonds from successive
main-chain NH groups in the segment of Cys-127 to Asn-130,
inclusive (Fig. 5). Any mutation of Asp-53 (Asp-43) would
therefore have a severely detrimental effect on the folding of
AnCPA or any of the members of this family.
Proposed Hydrolytic Mechanism for the Eqolisins. We have devel-
oped a plausible catalytic mechanism (Fig. 6a) for SCP-B that is
based on the experimentally determined structures of the native
unbound SCP-B (Fig. 3b) and the bound tripeptide product
complexed to SCP-B (Fig. 4b). This mechanism should be
applicable to all the members of the eqolisin family. The
mechanism proposed in Fig. 6a has similarities to those of the
aspartic and metallopeptidases in that it has a general base-
activated water molecule as the nucleophile, but it is unique in
many its features.
Fig. 5. The carboxylate group of Asp-43 in SCP-B is buried and is the recipient
of four H bonds from the main-chain NH groups from residues Cys-127 to
Asn-130, inclusively.
Fig. 6. (a) The proposed catalytic mechanism of SCP-B. The water molecule
hydrogen-bonded to both Glu-136 and Gln-53 is the nucleophile. The general
base is the carboxylate of Glu-136. The side-chain amide of Gln-53 assists in the
nucleophilic attack and stabilizes the tetrahedral intermediate by hydrogen
bonding. (b) A model of a substrate Ala-Ile-His-Pro bound in a productive
mode in the active site of SCP-B. The nucleophilic attack by the OH
⫺
ion (blue
sphere) is on the si-face of the scissile peptide. The surface of SCP-B is repre-
sented and colored according to the underlying atoms (slate, carbon; blue,
nitrogen; red, oxygen).
3368
兩
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0400246101 Fujinaga et al.

In our mechanistic proposal for the eqolisins, the water
(HOH169) bound between Glu-136 and Gln-53 (Fig. 3b) acts as the
nucleophile that is activated to a hydroxide ion by the general base
functionality of the carboxylate of Glu-136 (Fig. 6a). Electrophilic
assistance by polarizing the carbonyl bond of the scissile peptide is
provided by the side-chain amide of Gln-53. The hydrogen-bonded
interaction from Gln-53 N␧
2
to the substrate carbonyl oxygen is
bifunctional; it assists in the formation of the tetrahedral interme-
diate, and it provides stabilization of the oxyanion that results in
that intermediate.
We propose that a productive enzyme–substrate complex would
have P1 conformational angles of
␾
⬇⫺112° and
␺
⬇ 90° as
indicated in Fig. 6. As the nucleophilic attack takes place (the
formation of the covalent bond from the OH
⫺
ion to the carbonyl-
carbon atom of the scissile peptide), the carbonyl-carbon atom
becomes tetrahedral in character and the
␺
angle decreases to ⬇60°,
thus adopting a fully staggered conformation for the tetrahedral
intermediate (Fig. 6a). Gln-53 provides for stabilization of the
tetrahedral intermediate by donating a hydrogen bond from Gln-53
N␧
2
to the oxyanion of the tetrahedral intermediate and by receiv-
ing a hydrogen bond (to Gln-53 O␧
1
) from the OH of the attacking
water molecule. The nucleophilic attack takes place on the si-face
of the scissile peptide. This direction of attack is seen only in the
papain-like cysteine peptidase family; it is the re-face attack that
takes place in the serine, aspartic, and metallopeptidases.
Protonation of the leaving-group nitrogen is an obligatory step in
any hydrolytic mechanism for an amide. In our mechanism, it is
likely that the proton donor will be the protonated Glu-136 (Fig. 6a)
by analogy with the aspartic and metallopeptidases. The tetrahedral
character of the nitrogen atom in the intermediate will place the
lone-pair electrons on the nitrogen in an appropriate position to
abstract the proton from the carboxyl group of Glu-136. The
distance from the proton-donor oxygen atom of the Glu-136
carboxyl group to the leaving-group nitrogen atom of the tetrahe-
dral intermediate is ⬇3.2 Å (estimated from the equivalent distance
in the metalloenzymes).
Conclusions
Several reasons exist to place the eqolisins into a new family of
peptidases. First, they have a unique fold not previously observed
for proteolytic enzymes. No other known peptidase has a
␤
-sand-
wich as a tertiary fold. Second, they have a unique catalytic dyad of
a glutamate and a glutamine to activate the nucleophilic water and
to stabilize the tetrahedral intermediate on the hydrolytic pathway,
respectively. Third, the nucleophilic attack on the carbonyl-carbon
atom is from the si-face of the scissile peptide. This feature is in
common with the papain-like peptidases but distinct from the
serine, aspartic, and metalloproteinases. Last, the conformational
angle
␺
of the P1 residue of ecolisin substrates is unique among
peptidase families. In the eqolisins, substrates approach the cata-
lytic residues with the P1 residue having
␾
,
␺
⬇⫺115°, ⫹90°. The
serine and aspartic peptidases bind substrates with the main-chain
conformational angles of the P1residuesof
␾
,
␺
⬇⫺120°, ⫹35°,
respectively; the
␺
angle for the P1 residues of substrates of the
metallopeptidases is ⬇150° [estimated from the coordinates, 5TMN
(37)]. The cysteine peptidases of the papain family have
␾
,
␺
angles
of ⬇⫺120°, ⫺30° for their substrate P1 residues. Nature has used
common features of proteolysis but has mixed these with unusual
and different folds and different modes of substrate binding. The
chemistry, however, that of hydrolysis of a peptide bond, is common
to all peptidolytic enzymes.
We thank Ken Ng and Ernst Bergmann for collecting the native data set
of SCP-B on the synchrotron at SSRL; M.N.G.J. thanks Daniel Bur for
some important discussions at a recent conference ICAPI-03 in Japan;
and we thank Jason Maynes and Perry d’Obrenan for help with the
figures. This research was supported by grants from the Canadian
Institutes of Health Research to the Canadian Institutes of Health
Research Group in Protein Structure and Function, the Alberta Syn-
chrotron Institute, and the Alberta Heritage Foundation for Medical
Research.
1. Murao, S., Oda, K. & Matsushita, Y. (1973) Agric. Biol. Chem. 37, 1417–1421.
2. Oda, K. & Murao, S. (1974) Agric. Biol. Chem. 38, 2435–2444.
3. Oda, K. & Murao, S. (1976) Agric. Biol. Chem. 40, 1221–1225.
4. Majima, E., Murao, S., Oda, K. & Ichishima, E. (1988) Agric. Biol. Chem. 52,
787–793.
5. Murao, S. & Oda, K. (1985) in Aspartic Proteinases and their Inhibitors, ed.
Kostka, V. (Walter de Gruyter, Berlin), pp. 379–399.
6. Oda, K. & Murao, S. (1991) in Structure and Function of the Aspartic
Proteinases: Genetics, Structures, and Mechanisms, ed. Dunn, B. M. (Plenum,
New York), pp. 185–201.
7. Oda, K., Takahashi, S., Shin, T. & Murao, S. (1995) in Aspartic Proteinases:
Structure, Function, Biology, and Biochemical Implications, ed. Takahashi, K.
(Plenum, New York), pp. 529–542.
8. Oda, K., Takahashi, S., Ito, M. & Dunn, B. M. (1998) in Aspartic Proteinases:
Retroviral and Cellular Enzymes, ed. James, M. N. G. (Plenum, New York), pp.
349–353.
9. Wlodawer, A., Li, M., Dauter, Z., Gustchina, A., Uchida, K., Oyama, H., Dunn,
B. M. & Oda, K. (2001) Nat. Struct. Biol. 8, 442–446.
10. Comellas-Bigler, M., Fuentes-Prior, P., Maskos, K., Huber, R., Oyama, H., Uchida
K., Dunn, B. M., Oda, K. & Bode, W. (2002) Structure (London) 10, 865–876.
11. Oda, N., Gotoh, Y., Oyama, H., Murao, S., Oda, K. & Tsuru, D. (1998) Biosci.
Biotechnol. Biochem. 62, 1637–1639.
12. Inoue, H., Kimura, T., Makabe, O. & Takahashi, K. (1991) J. Biol. Chem. 266,
19484–19489.
13. Jara, P., Gilbert, S., Delmas, P., Guillemot, J. C., Jaghad, M., Ferrara, P. &
Loison, G. (1996) Mol. Gen. Genet. 250, 97–105.
14. Poussereau, N., Creton, S., Billon-Grant, G., Rascle, C. & Fevre, M. (2001)
Microbiology 147, 717–726.
15. Oda, K. (1998) in Handbook of Proteolytic Enzymes, eds. Barrett, A. J.,
Rawlings, N. D. & Woessner, J. F. (Academic, San Diego), pp. 969–970.
16. Tsuru, D., Shimada, S., Maruta, S., Yoshimoto, T., Oda, K., Murao, S., Miyata,
T. & Iwanaga, S. (1986) J. Biochem. (Tokyo) 99, 1537–1539.
17. Tsuru, D., Kobayashi, R., Nakagawa, N. & Yoshimoto, T. (1989) Agric. Biol.
Chem. 53, 1305–1312.
18. Kakimori, T., Yoshimoto, T., Oyama, H., Oda, N., Gotoh, Y., Oda, K., Murao,
S. & Tsuru, D. (1996) Biosci. Biotechnol. Biochem. 60, 1210–1211.
19. Huang, X. P., Kagami, N., Inoue, H., Kojima, M., Kimura, T., Makabe, O.,
Suzuki, K. & Takahashi, K. (2000) J. Biol. Chem. 275, 26607–26614.
20. Rawling, N. D. (1998) in Handbook of Proteolytic Enzymes, eds. Barrett, A. J.,
Rawlings, N. D. & Woessner, J. F. (Academic, San Diego), pp. 963–967.
21. Otwinowski, Z. & Minor, W. (1997) Methods Enzymol. 276, 307–326.
22. Terwilliger, T. C. & Berendzen, J. (1999) Acta Crystallogr. D 55, 849–861.
23. Terwilliger, T. C. (1999) Acta Crystallogr. D 55, 1863–1871.
24. Terwilliger, T. C. (2000) Acta Crystallogr. D 56, 965–972.
25. Perrakis, A., Morris, R. & Lamzin, V. S. (1999) Nat. Struct. Biol. 6, 458–463.
26. McRee, D. E. (1999) J. Struc. Biol. 125, 156–165.
27. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P.,
Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S.,
et al. (1998) Acta Crystallogr. D 54, 905–921.
28. James, M. N. G. & Sielecki, A. R. (1986) Nature 319, 33–38.
29. Sawyer, L. & James, M. N. G. (1982) Nature 295, 79–80.
30. James, M. N. G., Sielecki, A. R., Hayakawa, K. & Gelb, M. H. (1992)
Biochemistry 31, 3872–3886.
31. Kabsch, W. & Sander, C. (1983) Biopolymers 22, 2577–2637.
32. Delano, W. (2002) The
PYMOL Molecular Graphics System (Delano Scientific,
San Carlos, CA).
33. Read, R. J. (1986) Acta Crystallogr. A 42, 140–149.
34. Laskowski, M., Jr., & Qasim M. A. (2000) Biochim. Biophys. Acta 1477, 324–327.
35. Bode, W. & Huber, R. (1992) Eur. J. Biochem. 204, 433–451.
36. Takahashi, K., Kagami, N., Huang, X.-P., Kojima, M. & Inoue, H. (1998) in
Aspartic Proteinases: Retroviral and Cellular Enzymes, ed. James, M. N. G.
(Plenum, New York), pp. 275–282.
37. Holden, H. M., Tronrud, D. E., Monzingo, A. F., Weaver, L. H. & Matthews,
B. W. (1987) Biochemistry 26, 8542–8553.
Fujinaga et al. PNAS
兩
March 9, 2004
兩
vol. 101
兩
no. 10
兩
3369
BIOCHEMISTRY