Autoproteolytic and catalytic mechanisms for the β-aminopeptidase BapA - A member of the Ntn hydrolase family

Article abstract of DOI:10.1016/j.str.2012.07.017

The β-aminopeptidase BapA from Sphingosinicella xenopeptidilytica belongs to the N-terminal nucleophile (Ntn) hydrolases of the DmpA-like family and has the unprecedented property of cleaving N-terminal β-amino acid residues from peptides. We determined the crystal structures of the native (αβ)₄ heterooctamer and of the 153 kDa precursor homotetramer at a resolution of 1.45 and 1.8 A?, respectively. These structures together with mutational analyses strongly support mechanisms for autoproteolysis and catalysis that involve residues Ser250, Ser288, and Glu290. The autoproteolytic mechanism is different from the one so far described for Ntn hydrolases. The structures together with functional data also provide insight into the discriminating features of the active site cleft that determine substrate specificity.

Full text of DOI:10.1016/j.str.2012.07.017

Structure
Article
Autoproteolytic and Catalytic Mechanisms
for the b-Aminopeptidase BapA—A Member
of the Ntn Hydrolase Family
Tobias Merz,1,4 Tobias Heck,2,5 Birgit Geueke,2 Peer R.E. Mittl,1 Christophe Briand,1 Dieter Seebach,3
Hans-Peter E. Kohler,2 and Markus G. Gru¨ tter1,
*
1Biochemistry Institute, University of Zu¨ rich, Winterthurerstrasse 190, 8057 Zu¨ rich, Switzerland
2
¨
Department of Environmental Microbiology, Swiss Federal Institute of Aquatic Science and Technology (Eawag), Uberlandstrasse 133,
8600 Du¨ bendorf, Switzerland
3
¨
Laboratory of Organic Chemistry, Department of Chemistry and Applied Biosciences, Honggerberg HCI, Wolfgang-Pauli-Strasse 10,
8093 Zu¨ rich, Switzerland
4
´
Present address: NovAliX, BioParc, bld Sebastien Brant BP 30170, F-67405 Illkirch Cedex, France
5Present address: Swiss Federal Laboratories for Materials Science and Technology (Empa), Lerchenfeldstrasse 5, 9014 St. Gallen,
Switzerland
*Correspondence: gruetter@bioc.uzh.ch
http://dx.doi.org/10.1016/j.str.2012.07.017
SUMMARY
To our knowledge, the first and so far only structure of an enzyme
with b-aminopeptidase activity is DmpA (Bompard-Gilles et al.,
The b-aminopeptidase BapA from Sphingosinicella 2000). DmpA has an abba-sandwich architecture, which is
strongly reminiscent of the fold of classical Ntn hydrolases (Bran-
nigan et al., 1995). Nevertheless, DmpA is classified as a novel
xenopeptidilytica belongs to the N-terminal nucleo-
phile (Ntn) hydrolases of the DmpA-like family
and has the unprecedented property of cleaving
N-terminal b-amino acid residues from peptides.
member of the Ntn hydrolase family due to a different con-
nectivity of the secondary structure elements (Bompard-Gilles
et al., 2000; Oinonen and Rouvinen, 2000).
Ntn hydrolases, in general, are synthesized as inactive
precursor polypeptides with or without a periplasmic leader
We determined the crystal structures of the native
(ab)4 heterooctamer and of the 153 kDa precursor
˚
homotetramer at a resolution of 1.45 and 1.8 A,
sequence. They undergo posttranslational autoproteolytic
cleavage into heterodimers. The primary nucleophile that initi-
ates this cleavage is either a serine, threonine or cysteine
residue, which, upon self-processing, becomes the essential
respectively. These structures together with muta-
tional analyses strongly support mechanisms for
autoproteolysis and catalysis that involve residues
Ser250, Ser288, and Glu290. The autoproteolytic N-terminal nucleophile for the catalytic activity. This autoproteo-
lytic activation step is a characteristic trait of all Ntn hydrolases
(Brannigan et al., 1995; Kim et al., 2003; Saarela et al., 2004; Mi-
mechanism is different from the one so far described
for Ntn hydrolases. The structures together with
functional data also provide insight into the discrim-
inating features of the active site cleft that determine
chalska et al., 2008). Ntn hydrolases occur in different oligomeric
states, such as heterodimers (e.g., penicillin G acylase; McVey
et al., 2001), heterotetramers (e.g., glutamyltranspeptidase
substrate specificity.
from Helicobacter pylori; Boanca et al., 2006), heterooctamers
(e.g., penicillin V acylase from Bacillus sphaericus; Suresh
INTRODUCTION
et al., 1999) or, exceptionally, as very large assembly in the pro-
teasome (Baumeister and Lupas, 1997). Most Ntn hydrolases
The class of b-aminopeptidases so far comprises five function- catalyze the hydrolysis of amide bonds. However, their substrate
ally characterized hydrolytic enzymes, i.e., three BapA variants specificities vary considerably. The ornithine acetyltransferase
from Sphingosinicella xenopeptidilytica 3-2W4 (Geueke et al., from the clavulanic-acid-biosynthesis gene cluster (orf6) is an
2005), S. microcystinivorans Y2 (Geueke et al., 2006), and exception. This enzyme has the same fold as DmpA, but cata-
Pseudomonas sp. MCI3434 (Komeda and Asano, 2005) as well lyzes the reversible transfer of an acetyl group from N-acetylor-
as BapF from Pseudomonas aeruginosa PAO1 (Fuchs et al., nithine to glutamate (Elkins et al., 2005). Among the Ntn hydro-
2011), and DmpA from Ochrobactrum anthropi (Fanuel et al., lases, penicillin and cephalosporin acylases as well as
1999b). These enzymes share the exceptional ability to cleave b-aminopeptidases have particularly high potential for bio-
synthetic b-peptides, which consist of backbone-elongated catalytic conversions (Bruggink et al., 1998; Barends et al.,
b-amino acid residues that are not processed by common 2004; Heck et al., 2007, 2009, 2010).
proteolytic enzymes (Frackenpohl et al., 2001; Geueke and Koh-
The enzymatic activities of b-aminopeptidases strongly de-
ler, 2007; Wiegand et al., 2002). Due to their structural properties pend on the nature of the N-terminal amino acid of the sub-
and high resistance against proteolysis, b-peptides are consid- strate peptide. While DmpA hydrolyzes D- and L-configured
ered promising building blocks for the design of novel peptido- a-amino acid amides, esters and peptides as well as peptides
mimetics (Aguilar et al., 2007; Seebach and Gardiner, 2008). carrying small N-terminal b-amino acids such as b-homoglycine
1850 Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved

Structure
Structure of the b-Aminopeptidase BapA
Figure 1. Structural Overview of Native
BapA and proBapA
(A) Cartoon representation of the (ab)4-BapA
heterooctamer. The individual subunits, each
comprising an a- and a b-polypeptide chain, are
shown in gray, magenta, cyan, and green. The
blow-up frame shows the catalytically important
residues Glu133, Ser250, Ser288, and Glu290 of
the gray subunit as sticks.
(B) Cartoon representation of one ab subunit. The
secondary structure elements contributing to the
Ntn fold are colored in red and blue.
(C) Active-site pocket of BapA before and after the
autoproteolytic cleavage: Native BapA and pro-
BapA are shown in gray. Residues 237–245 of
native BapA are depicted in yellow and residues
237 to 249 of proBapA are shown in magenta. The
catalytic nucleophile Ser250 is highlighted as
sticks. Residues Pro240 and Pro245 are located
within a helical loop conformation in proBapA,
which upon autoproteolytic processing assume
a b sheet-like conformation in the native enzyme.
The positional changes of these two residues are
indicated by arrows.
˚
high-resolution crystal structure (1.45 A)
of BapA was determined by molecular
replacement as described in Experimental
Procedures. Each heterodimer comprises
a large a-polypeptide chain (residues
and b3-homoalanine (Fanuel et al., 1999a; Heck et al., 2006), 1–249) and a small b-polypeptide chain (residues 250–373). The
BapA from S. xenopeptidilytica exhibits hydrolytic activity with molecule adopts a globular disk-like shape with subunits ar-
strict preference for variable N-terminal b-amino acids and, ranged in D2 symmetry (Figure 1A). Each ab subunit adopts an ab
remarkably, does not hydrolyze a-peptides (Geueke et al., ba-sandwich fold that is formed by two layers of four and five
2006; Heck et al., 2006). Until now, the natural substrates of mixed b sheets flanked by two helices on each side (Figure 1B).
b-aminopeptidases remain unknown and structural data about The connectivity of the secondary structure elements in BapA
exclusively b-peptide-cleaving enzymes is not available.
Here, we present the high-resolution crystal structures of the the fold of DmpA. The total interface area between all subunits
differs from the classical Ntn hydrolase fold, but it is identical to
2
˚
b-aminopeptidase BapA, which has exclusive catalytic activity comprises 7,880 A in BapA and suggests a very stable and
for b-peptide substrates. Our comprehensive structural and permanent octameric state in solution that was elucidated exper-
functional study about this novel aminopeptidase includes the imentally by gel filtration (data not shown). The model was refined
crystal structures of the active form, its precursor and of the to final R and Rfree values of 13.5% and 15.3%, respectively.
active form in complex with the serine-peptidase inhibitor 4-(2- Crystallographic data are listed in Table 1. The BapA molecule
aminethyl) benzenesulfonyl fluoride (AEBSF, trade name pefa- contains four active sites, each of which is located at the inter-
bloc SC). The analysis of these BapA structures provides insight face of three adjacent subunits forming the substrate-binding
into the mechanisms of precursor self-processing and substrate pockets. In analogy to other Ntn hydrolases, Ser250 becomes
cleavage, and shows the determinants of the enzyme’s sub- the N-terminal catalytic nucleophile of the b-polypeptide chain
strate specificity. Moreover, our mutational analysis indicates after self-processing (Figure 1C). Glu133, Ser250, Ser288, and
that BapA, DmpA, and other so far not characterized DmpA- Glu290, that we propose to participate in substrate binding and
like aminopeptidases, do not, as reported, act with a ‘‘Ser- the catalytic mechanism, are located in the same subunit (Fig-
2
only’’ active-site configuration. Instead of a single catalytically ure 1A). A water molecule (Baverage = 18.1 1.1 A ) is in close
˚
active amino acid residue, these enzymes seem to employ contact to the carboxyl group of Glu133 in each of the active
a highly conserved Glu-Ser-Ser triad for the self-processing as sites. The shape of the active sites of BapA with their buried cata-
well as for the catalytic substrate hydrolysis.
lytic residues excludes the action of associated peptidases that
could be involved in the precursor cleavage.
RESULTS
Crystal Structure of the Unprocessed BapA Precursor—
proBapA
BapA crystallized in the space group P21 and contained one (ab)4 The homotetrameric unprocessed precursor form of BapA
Crystal Structure of Native BapA
heterooctameric molecule (153 kDa) in the asymmetric unit. The (proBapA) was produced with Ser250 mutated to alanine
Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved 1851

Structure
Structure of the b-Aminopeptidase BapA
Table 1. Data Collection and Refinement Statistics—Molecular
Replacement
BapA-
proBapA
P212121
Native BapA
P21
Pefabloc SC
Data Collection
Space group
P21
Cell dimensions
˚
a, b, c (A)
99.7, 113.7, 87.4, 96.7, 101.4 87.4, 96.8,
126.0
101.3
a, b, g (ꢀ)
90, 90, 90
90, 108.2, 90
90,108.4,90
a
˚
Resolution (A)
29.4–1.8
(1.9–1.8)
48.8–1.45
48.0–1.8
(1.9–1.8)
(1.47–1.45)
Rsym or Rmerge
I / sI
7.1 (47.3)
15.6 (4.0)
99.1 (99.2)
4.40 (4.46)
4.2 (43.3)
17.0 (2.7)
98.5 (92.3)
3.24 (2.33)
10.0 (49.7)
12.9 (3.1)
Completeness (%)
Redundancy
Refinement
97.6 (94.7)
3.78 (3.70)
˚
Resolution (A)
1.8
1.45
1.8
No. reflections
Rwork / Rfree
No. atoms
Protein
131,659
15.5/18.0
278,534
13.5/15.3
144,691
16.0/19.0
10,639
49
21,935 (10,836)b 10,708
Ligand/ion
Water
83
238
1,124
1,741
1,265
B-factors
Protein
21.54
37.59
31.79
18.58 (17.31)b
30.61
19.61
27.85
29.48
Ligand/ion
Water
31.06
Rmsds
˚
Bond lengths (A)
0.009
1.138
0.011
1.445
3N2W
0.007
1.05
Bond angles (ꢀ)
PDB Accession Code 3N5I
3N33
aValues in parentheses correspond to the highest-resolution outer shell.
bWithout hydrogen atoms. One crystal was used for each data set.
Figure 2. Stereoview of the Active Sites of proBapA and Native BapA
(A) BapA precursor mutant S250A (proBapA): The important catalytic residues
Glu133, Ser288, Glu290, and Ala250 are shown as sticks and the catalytic
water molecule is shown as a red sphere. The carboxy-terminal residue of the
a chain before the autoproteolytic cleavage is shown in magenta. In the bottom
panel the same part of the structure is shown with the 2fo-fc electron density
map contoured at 1.5 s.
(S250A mutant), which prevents the primary cleavage of the
enzyme subunits into a- and b-polypeptides. This enzymatically
inactive form of BapA crystallized in the space group P212121
and contains one homotetrameric molecule in the asymmetric
˚
unit. The high-resolution crystal structure (1.8 A) of proBapA
was determined by molecular replacement as described in
Experimental Procedures. The model was refined to final R
and Rfree values of 15.5% and 18.0%, respectively. Crystallo-
graphic data are listed in Table 1. Overall, the globular shape
and the contact area between the subunits in proBapA strongly
resemble native BapA. The location of the linker peptide
(B) Native BapA: rotation of Ser250 around the C-Ca-bond decreases the
+
˚
NH3 250-OH288 distance from 3.7 to 2.8 A (distance values highlighted in red in
figures A and B). In the bottom panel the same part of the structure is shown
with the 2fo-fc electron density map contoured at 1.5 s.
connecting the designated a and b chains (residues 231–249) Ser250 peptide bond goes along with a subtle reorientation
represents the major structural difference between native of the Ser250 side chain, which represents a second function-
BapA and proBapA (Figure 1C). In proBapA, this segment partly ally highly important structural change in the active site.
covers the active site region in native BapA. The linker polypep- Comparing the conformations of Ala250 in proBapA and
tide is flexible in three subunits and in one subunit adopts Ser250 in active BapA we find that Ser250 is rotated around
a helical conformation due to crystal contacts with Pro240 its C-Ca bond upon self-processing. This rotation reduces
and Pro245. In native BapA, the same segment is moved out the distance between Ala/Ser250 and Ser288-Og from 3.67
˚
of the active site cleft and adopts a b sheet-like conformation to 2.79 A in native BapA (Figures 2A and 2B). The interresidue
with the prolines located in turns. Cleavage of the Asn249- distances between Glu133, Glu290, and Ser288 do not
1852 Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved

Structure
Structure of the b-Aminopeptidase BapA
Table 2. Mutations in BapA and Their Impact on Precursor Cleavage and Substrate Hydrolysis
Residue Essential for
Autoproteolysis
Residue
Point Mutation
wt
Precursor Cleavage (Autoproteolysis)
Enzymatic Activity (Catalysis)
Catalysis
Complete
Slow
100%a
Inactivea,b
100%a
100%a
Inactiveb
<1%b
Glu133
Lys248
Asn249
Ser250
Ser288
Glu290
E133A
K248A
N249A
S250A
S288A
E290A
No
Yes
No
Slow
No
Slow
No
No
No cleavage
No cleavage
No cleavage
Yes
Yes
Yes
Yes
Yes
Yes
24%b
aThe activities of the mutants with partially purified enzymes were measured after 192 hr of incubation under standard assay conditions (see
Experimental Procedures ). 100% correspond to the activity of partially purified native BapA that was treated under the same conditions.
bThe activities of the BapA Duet variants were measured under standard assay conditions from cell-free extracts containing the recombinantly
expressed enzymes. 100% correspond to the activity of cell-free extract containing native BapA Duet (see Experimental Procedures).
significantly differ between the precursor and the processed Mechanism of Self-Processing
enzyme.
In the commonly accepted self-processing mechanism for Ntn-
hydrolases the catalytic nucleophile attacks the carbonyl carbon
atom of the preceding peptide bond. This results in the formation
Catalytic Residues for Self-Processing and Catalysis
We exchanged six amino acid residues in the vicinity of the cata- of a tetrahedral intermediate that is decomposed with an ester
lytic nucleophile Ser250 for Ala residues, creating the BapA point bond being formed and the carbonyl carbon amide nitrogen
mutants E133A, K248A, N249A, S250A, S288A, and E290A. The bond cleaved. In a second step, the ester bond is hydrolyzed, re-
consequences of these mutations on the autoproteolysis of the sulting in an a- and a b-polypeptide chain with the free N-terminal
precursor polypeptide were judged based on the time-depen- catalytic nucleophile. This mechanism requires a conformational
dent appearance of individual a and b chains on SDS-PAGE strain on the peptide bond (Johansson et al., 2009) and, in
gels. Native BapA was already fully processed after purification, addition, includes a flip of the peptide preceding the bond to
while the BapA mutants did not self-process at all or only partially be cleaved. It was deduced from structural information of the
within 192 hr (Table 2; Figure 3). In fact, the exchange of Ser250, proform and active form of glycosylasparaginase (Xu et al.,
Ser288, or Glu290 for alanine completely abrogated precursor 1999) and cephalosporin acylase (Kim et al., 2006). This so
cleavage: neither could the individual a- and b-polypeptide called N-O shift mechanism is unlikely to occur in BapA for
chains be detected on SDS-PAGE gels (Figure 3) nor did the the following structural reasons: (1) Ser250-Og when modeled
mutant enzymes exhibit any residual enzymatic activity after in the Ala250 mutant structure would require to adopt an
192 hr of incubation (Table 2). The BapA mutants E133A, eclipsed conformation, which is not in an ideal geometry to
K248A, and N249A were clearly delayed in self-processing attack the preceding main chain carbonyl carbon of Asn249,
(SDS-PAGE gel for K248A not shown). Nevertheless, activity (2) the Asn249-Ser250 scissile peptide bond in proBapA is
measurements after 192 hr of incubation revealed that K248A completely planar and unstrained, (3) the cleavage-defective
and N249A mutants were as active as native BapA, whereas BapA mutants S288A and E290A indicate a critical involvement
the exchange of Glu133 completely abolished enzymatic activity of these residues in the processing mechanism besides
(Table 2).
Ser250, (4) a peptide flip of the scissile peptide bond between
Because the BapA mutants E133A, S250A, S288A, and E290A Asp249 and Ser250 results in an energetically unfavorable
did not undergo the complete processing step, the effects of conformation. All these features are characteristic for BapA
these mutations on substrate binding and conversion could not and DmpA-type Ntn-hydrolases and distinguish them from
be investigated. In order to study the mechanism of substrate other Ntn-hydrolases for which the N-O acyl shift mechanism
cleavage independently from precursor processing, we coex- was proposed. Based on the structural data and our mutational
pressed the individual a- and b-polypeptide chains of BapA studies, we propose the following self-processing mechanism
from a Duet expression vector and introduced the same muta- (Figure 4): Ser288-Og is the nucleophile that is in an ideal posi-
tions that were previously investigated for their effects on tion to attack the carbonyl carbon of the Asn249-Ser250 scis-
self-processing. We ensured by sequence analysis that the sile peptide bond (Figure 4A). The peptide carbonyl group of
newly introduced N-terminal methionine residue of the b-poly- Asn249 draws a proton from the side-chain OH of Ser250
peptide chain was posttranslationally removed by Escherichia leaving a partial negative charge on the Ser250-Og. The nucle-
coli (data not shown). The cell-free extracts containing the ophilic attack of Ser288 is facilitated by the transfer of a proton
respective BapA Duet mutants were then tested for their cata- from Ser288-Og to the neighboring carboxyl side chain of
lytic activity with the chromogenic substrate H-b3hAla-pNA. Glu290. The tetrahedral intermediate formed (Figure 4B) is
The mutants E133A Duet and S250A Duet are inactive, whereas stabilized by a hydrogen bond to Ser250-Og. This intermediate
the mutants S288A Duet and E290A Duet exhibit <1% and 24% collapses with the cleavage of the original Asn249-Ser250
residual activity, respectively, compared to wild-type BapA Duet peptide bond resulting in the formation of the new N-terminal
(Table 2).
Ser250 a-amino group of the b-polypeptide and the N-terminal
Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved 1853

Structure
Structure of the b-Aminopeptidase BapA
BapA, a 310-helix comprising the residues Glu120 to Arg126
creates a wide substrate-binding pocket, which is largely
occluded by loops in DmpA (Gln131 to Trp137) and PH0078
(Glu120 to Ser128). From this superposition we conclude
that Trp137, which points into the active-site cavity of
DmpA, restricts catalysis of DmpA to the removal of sterically
undemanding N-terminal amino acids from peptides. To
confirm this hypothesis, we exchanged Trp137 of DmpA for
alanine by site-directed mutagenesis in order to widen the
substrate-binding pocket of DmpA and thus make it accessible
for bulkier substrates. Hereby, it is important to mention
that the point mutation W137A negatively affected the self-
processing rate of the DmpA precursor polypeptide into the
individual a- and b-polypeptides. In contrast to the native
BapA and DmpA proteins that were fully processed after the
purification procedure, DmpA W137A was only half-processed
after purification and 3 days of incubation at pH 7.2 and 37ꢀC.
Figure 3. SDS-PAGE Analysis of Five Processing-Deficient or Slow-
Processing BapA Point Mutants
The partially purified mutant proteins (0.6 mg/ml) were incubated in 50 mM
Tris/HCl buffer (pH 8) at 37ꢀC. Samples were taken after 0 hr (0), 54 hr (1),
120 hr (2), and 192 hr (3). The mutants S250A, S288A, and E290A did not
self-process at all, while the mutants E133A and N249A partially processed
over 192 hr.
a-polypeptide, which is ester-linked to the side chain of A survey of the kinetic parameters of the conversion of the
Ser288. In a second step the water oxygen—ideally positioned b3-homoamino acid p-nitroanilides H-bhGly-pNA, H-b3hAla-
through H-bonds with Asn207-Nd and the amide nitrogen of pNA, H-D-b3hAla-pNA, H-b3hPhe-pNA, and H-D-b3hPhe-
Leu135—is the nucleophile that attacks the ester carbonyl pNA by BapA, DmpA, and DmpA W137A is shown in Table 3.
carbon atom. This leads to the hydrolysis of the ester bond While BapA converted the aromatic b3-homophenylalanine
and to the release of the a-polypeptide. Ser288-Og sub- p-nitroanilides of L- and D-configuration with kcat/KM values of
sequently binds the proton back from Glu290. This pro- 14,500 and 940
posed mechanism is clearly different from the cis-autoproteol- these bulky substrates by DmpA was negligible (kcat/KM
ysis described recently by (Buller et al., 2012): modeling of a 10
ꢁ1sꢁ1). The single amino acid exchange of Trp137 for
M , respectively, the conversion of
ꢁ1sꢁ1
<
M
cis-conformation could not be accommodated in the pre- alanine in DmpA drastically altered the substrate specificity of
sented structure, moving residues far away from their canonical DmpA W137A leading to more than 280- and 670-fold
conformation.
increased catalytic efficiencies for the conversion of the
substrates H-b3hPhe-pNA and H-D-b3hPhe-pNA, respectively.
On the other hand, DmpA W137A converted substrates with
Catalytic Mechanism of Substrate Cleavage
The substrate molecule freely diffuses into the active-site pocket small side chains, i.e., H-bhGly-pNA, H-b3hAla-pNA, and H-D-
of BapA and its N terminus makes electrostatic interactions with b3hAla-pNA, less efficiently than wild-type DmpA and dis-
the side chain carboxyl group of Glu133 (Figure 5B, Michaelis played overall
kcat/KM values similar to BapA for these
complex). Such an interaction is observed in the structure of substrates (Table 3).
BapA in complex with the inhibitor Pefabloc (Figure 5A). When
positioning a b-peptide substrate in the active site accordingly Sequence Comparison of Putative b-Aminopeptidases
the following substrate-cleavage mechanism can be deduced: and Implications on Their Substrate Specificity
Glu290 activates Ser250-Og via Ser288-Og and the a-amino In addition to the five functionally characterized b-amino-
group of Ser250 for the nucleophilic attack on the carbonyl peptidases (Geueke et al., 2006; Komeda and Asano, 2005;
carbon atom of the substrate’s scissile peptide bond. Leu135- Fuchs et al., 2011; Fanuel et al., 1999a; Heck et al., 2006),
NH and Asn207-NH2 form the oxyanion hole and stabilize the a BLAST search revealed a large number of protein sequences
negative charge of the oxyanion of the tetrahedral intermediate showing high similarity to the sequence of BapA from
(Figure 5B), which collapses to the acyl enzyme under release S. xenopeptidilytica. They mainly originated from Gram-nega-
of the C-terminal substrate moiety. The hydrolysis of the acyl tive bacteria, among them many pathogenic species such as
enzyme is subsequently catalyzed by the water molecule, which Burkholderia mallei, B. pseudomallei and various strains of
is hydrogen bonded to the N-terminal a-amino group of Ser250. Pseudomonas aeruginosa, but also from archaea (e.g., Pyro-
The structure of the inhibitor Pefabloc bound in the active site of coccus horikoshii), fungi (e.g., Aspergillus oryzae), and yeast
BapA (Figure 5A) allows to model how substrate binds in the (e.g., Yarrowia lipolytica). Strikingly, the region downstream
active site and supports the here described mechanism (Heck of the N-terminal catalytic nucleophile (ranging from Ser250
et al., 2012).
to Ser290) and the region around the salt bridge-forming
residue Glu133 are highly conserved among all compared
sequences, indicating that these regions are crucial for catal-
ysis (Figure 6). Glutamate at position 290 of BapA is less,
Comparison between the Substrate-Binding Pockets of
BapA and DmpA
The superposition of the structures of native BapA, DmpA (PDB but functionally conserved. In contrast to these highly
ID 1B65), and a so far uncharacterized protein from Pyrococ- conserved regions, the compared sequences show great vari-
cus horikoshii (PH0078, PDB ID 2DRH) show a good overlap ability over a stretch of 15–20 amino acids upstream of
of the corresponding catalytic residues. However, the sizes of Glu133. This sequence stretch corresponds to the active-site
the substrate-binding pockets differ notably (Figure 5C). In loop (310-helix in BapA), which defines the width of the
1854 Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved

Structure
Structure of the b-Aminopeptidase BapA
Figure 4. Active-Site Configuration of pro-
BapA and Proposed BapA Self-Processing
Mechanism
(A) Residues are shown as sticks and water
molecules are shown as red spheres. The
C-terminal residues of the a chain before auto-
proteolytic cleavage are shown in gray. Residues
important for autoprocessing are labeled. Dis-
tances relevant for the mechanism are indicated
on the picture.
(B) Step 1: Ser288-Og is the nucleophile that is in
an ideal position to attack the carbonyl carbon of
the Asp249-Ser250 peptide bond to be cleaved.
The tetrahedral intermediate formed is stabilized
by a hydrogen bond to the Ser250 -Og. Step 2: The
water oxygen ideally positioned through H-bonds
with Asn207-Nd and the amide nitrogen of Leu135
is the nucleophile that attacks the tetrahedral
intermediate resulting in two possible mechanism
of cleavage. Step 3: Release of the C terminus of
the a-polypeptide; Ser288-Og binds the proton
back from Glu290.
force for the large conformational change
from a helical to a b sheet-like conforma-
tion in the autoprocessing step. Similar
observations of strained conformations
of linker sequences with prolines that are
supposed to regulate the autoprocessing
of Ntn hydrolases are reported for glutaryl
7-aminocephalosporanic acid acylase
(Kim et al., 2003) and cephalosporin acy-
lase (Kim et al., 2006).
BapA differs from other classical Ntn
hydrolases by the connectivity of the
secondary structural elements, but has
the same fold as DmpA from O. anthropi
enzymes’ substrate-binding pockets and their substrate spec- (Bompard-Gilles et al., 2000). Therefore, BapA belongs to the
ificities (Table 3).
DmpA-like family. DmpA is reported to be structurally closely
related to the superfamily of Ntn hydrolases and to exhibit
a single-residue catalytic mechanism, in which the N-terminal
nucleophile (Ser250), located at the N terminus of the central
b sheet, uses its own a-amino group as Brønsted base for acti-
vation (Bompard-Gilles et al., 2000).
DISCUSSION
Reconsideration of the Single-Residue Catalytic
Mechanism of DmpA-like Proteins
In this study, we report the crystal structures of the native
BapA and DmpA share 41.8% sequence identity and are
and precursor form of BapA, a b-aminopeptidase from S. xeno- structurally and biochemically closely related. The catalytically
peptidilytica. Native BapA revealed a heterooctameric organiza- important residues of BapA overlap extremely well with corre-
tion comprising four heterodimers in a globular D2-symmetry sponding conserved amino acids of DmpA and the substrate
arrangement. Each heterodimer assumed an abba-sandwich profiles of both enzymes are similar (Heck et al., 2006). We
fold, which was very similar to the core-fold of classical therefore propose that substrate conversions by DmpA and
Ntn hydrolases (Brannigan et al., 1995; Oinonen and Rouvinen, BapA involve a similar mechanism. We showed that the resi-
2000). ProBapA revealed a homotetrameric organization with dues Ser250, Ser288, and Glu290 of BapA (corresponding to
the same globular arrangement and core-fold as the native Ser250, Ser288, and Asp290 of DmpA) have important func-
enzyme. The displacement of the linker segment between the tions for the self-processing and the catalytic mechanism
a- and b-polypeptides (residues 231–249) represented the (Table 2). The crystal structure of native BapA revealed short
+
˚
major structural difference between native BapA and proBapA. hydrogen-bond distances of 2.8 A between Ser250-NH3
˚
Although this segment contained two prolines in allowed Rama- and Ser288-OH and of 2.6 A between Ser288-OH and the
chandran regions and no further significant deviations from ideal carboxyl group of Glu290, suggesting that Glu290 and
geometry, we expect these prolines to provide the initial driving Ser288 activate the catalytic nucleophile Ser250-Og via the
Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved 1855

Structure
Structure of the b-Aminopeptidase BapA
Figure 5. Active Site Configuration and Proposed Substrate Cleavage Mechanism of Native BapA
Stereo representation of the active site configuration of native BapA in presence of the Pefabloc inhibitor (A), proposed catalytic mechanism of substrate cleavage
(B) and closeup view into the active-site pocket of native BapA and superposition of BapA with DmpA and PH0078 (C).
(A) The catalytic residues Glu133, Ser250, Ser288, and Glu290 are shown as sticks. Hydrogen bonds are shown as dashed lines. Pefabloc is colored in cyan. The
2fo-fc electron density map around the inhibitor in the active site is contoured at 1.5 s.
1856 Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved

Structure
Structure of the b-Aminopeptidase BapA
Table 3. Kinetic Analyses of the Conversion of Different b3-Homoamino-Acid p-Nitroanilides by the Purified b-Aminopeptidases BapA,
DmpA, and DmpA W137A
BapA
DmpA
DmpA (W137A)
KM
Vmax
(mmol*minꢁ1mgꢁ1
Kcat/KM
(Mꢁ1sꢁ1
KM
Vmax
(mmol*minꢁ1mgꢁ1
Kcat/KM
(Mꢁ1sꢁ1
KM
Vmax
(mmol*minꢁ1mgꢁ1
Kcat/KM
(Mꢁ1sꢁ1
)
Substrate
(mM)
)
)
(mM)
)
)
(mM)
)
H-bhGly-pNA
6.6 2.3 0.65 0.15
64 37
0.016
0.001
55.8 0.8
2,400,000
200,000
2.9 0.2 20.1 0.5
4600
400
H-b3hAla-pNA
H-D-b3hAla-pNA
H-b3hPhe-pNA
H-D-b3hPhe-pNA
1.1 0.2 16.4 0.9
9,600
1,800
<10a
0.063
0.002
109
1
1,200,000
100,000
0.50
0.02
40.0 0.3
0.017 0.002
3.5 0.1
53,800
2,100
17.0
2.7
1.3 0.2
52 15
0.38
0.13
30 13
3.5 1.3 79.3 14.5
14,500
7,800
940 3a
<10a
0.85
0.09
2,800
400
<10a
2.7 0.4 26.7 2.0
6,700
1600
aThe enzyme activities measured at pH 7.2 and 37ꢀC showed a linear dependency on the substrate concentration between 0.1 and 5 mM. The kcat/KM
values were calculated according to the equation kcat/KM = v/(Eo)*[S]), where v is the rate of the reaction, [S] the concentration of the substrate and (Eo)
the stoichiometric concentration of active centers (MBapA: 38,610, MDmpA: 40,440, MDmpA W137A: 40,330). For the calculation of all kinetic parameters,
only the processed enzyme fractions in the protein solutions were taken into account.
a-amino group. These assumptions are in accordance with our enzymes with different catalytic mechanisms. The prototypic
kinetic analysis that showed 100- and 4-fold decreased activ- Ntn hydrolase penicillin G acylase is reported to have a single
ities, when Ser288 and Glu290 were exchanged for alanines, catalytic serine residue (Duggleby et al., 1995), the g-glutamyl-
respectively. Nevertheless, the residual catalytic activity of transpeptidase from H. pylori reveals a threonine-threonine
the S288A Duet mutant indicates that Ser250 alone is capable catalytic dyad (Boanca et al., 2007) and the 20S proteasome
of being catalytically active as postulated by Bompard-Gilles from Thermoplasma acidophilum employs a catalytic tetrad
¨
et al. (2000), but exhibits only very low catalytic activity in with a threonine residue as primary nucleophile (Lowe
comparison to the native enzyme. In fact, this loss of activity et al., 1995).
is not surprising. We assume the a-amino group of Ser250
The presented substrate cleavage mechanism, which is based
to be mostly protonated (pKa > 9). Therefore, the amino group on mutational and structural data, partially resembles the previ-
is unlikely to act as a base abstracting a proton from Ser250- ously published mechanism by Tikkanen et al. (1996). However,
OH and to provide the major driving force for the activation of the proposed mechanism for BapA involves a triad of catalytic
its own nucleophile.
residues (namely, Glu 290, which activates Ser 250 via Ser 188
The main consequence arising from our study is that the for substrate cleavage), whereas the mechanism of Tikkanen
proposed ‘‘Ser-only’’ catalytic mechanism of enzymes of the et al. (1996) uses a dyad. The presented mechanism for autopro-
DmpA family has to be reconsidered. Our results strongly teolytic activation of proBapA differs from recently published
suggest that BapA and DmpA feature catalytic triads (Ser- mechanisms, such as, for example, the pantetheine hydrolase
Ser-Glu and Ser-Ser-Asp, respectively) rather than single cata- ThnT. This enzyme is cis-autoproteolysed during self-activation
lytic serine residues. This clearly extends the currently (Buller et al., 2012). For self-activation of proBapA, we propose
proposed catalytic mechanism of DmpA. The importance of a similar mechanism than for substrate cleavage involving the
these catalytic residues for BapA, DmpA and other DmpA-like Ser-Ser-Glu triad. Like already discussed, enzymes having the
proteins is further supported by their high conservation on the same Ntn-hydrolase fold do not necessarily act in the same
structural (Figure 5C) and on the sequence level (Figure 6). way during substrate cleavage and self-activation (Duggleby
¨
Interestingly, despite the close structural similarity of the et al., 1995; Lowe et al., 1995; Tikkanen et al., 1996; Boanca
abba-sandwich fold, the Ntn hydrolase superfamily comprises et al., 2007).
(B) Michaelis complex: Positioning of substrate by electrostatic interactions between the substrate’s amino terminus and Glu133. Nucleophilic attack on the
carbonyl group of the substrate by Ser250-Og, which is activated by Glu290, via Ser288 and the free amino group of Ser250. Step 1: Stabilization of the oxyanion
by the backbone amide group of Leu135 and the dNH2 group of Asn207. Step 2: Substrate cleavage and formation of the acyl enzyme. Step 3: Activation of the
catalytic water molecule by Glu133 and initialization of the acyl-enzyme hydrolysis. Step 4: Indirect protonation of the amino group of Ser250 by Glu290 via
Ser288 and formation of the second tetrahedral intermediate. Step 5: Hydrolysis of the tetrahedral intermediate, release of the cleaved product, and reconsti-
tution of the active-site residues for another catalytic cycle.
(C) Stereo representation of the active sites of DmpA from O. anthropi (PDB ID 1B65) and a noncharacterized protein form P. horikoshii (PH0078, PDB ID 2DRH),
superimposed onto the structures of BapA from S. xenopeptidilytica. The active-site loops containing Ser128 (PH0078) in purple and Trp137 (DmpA) in orange are
part of the highly variable sequence stretch of 15–20 amino acids upstream of the conserved Glu133 (BapA numbering). The two subunits forming the active site
of BapA are shown in cyan and red, the active-site loops as well as the residues of DmpA and PH0078 are shown as sticks in orange and purple, respectively. The
active-site residues Glu133, Ser250, Ser288, and Glu290 of BapA are shown as sticks in yellow. They correspond to Glu135/Glu144, Ser239/Ser250, Ser277/
Ser288, and Asp279/Asp290 of PH0078 and DmpA, respectively.
Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved 1857

Structure
Structure of the b-Aminopeptidase BapA
Figure 6. Partial Alignment of b-Aminopeptidase-like Sequences
Sequence alignment of BapA from S. xenopeptidilytica with four previously characterized b-aminopeptidases (2–5) and selected, noncharacterized protein
sequences that were retrieved from a BLAST search (6–18). Two sections of the alignment are shown (amino acids 104–145 and 250–290). The numbering of the
residues refers to BapA from S. xenopeptidilytica. Identical amino acids are highlighted in black, similar amino acids are highlighted in gray. Trp137 of DmpA from
O. anthropi, which was selected as a target for mutagenesis, and Ser128 of the noncharacterized protein PH0078 from P. horikoshii (NP_142096) are underlined
and printed in bold. The alignment was created with the program ClustalW (Thompson et al., 1994). (1) BapA from Sphingosinicella xenopeptidilytica 3-2W4
(AAX93858), (2) DmpA from Ochrobactrum anthropi (CAA66259), (3) BapA from Sphingosinicella microcystinivorans ABC59253, (4) BapA from Pseudomonas sp.
MCI3434 (BAE02664), (5) BapF from Pseudomonas aeruginosa PAO1 (NP_250177). Noncharacterized protein sequences from (6) Agrobacterium tumefaciens
C58, (7) Burkholderia mallei ATCC 23344, (8) Mycobacterium gilvum PYR-GCK, (9) Mycobacterium smegmatis MC2 155, (10) Myxococcus xanthus DK 1622, (11)
Photorhabdus asymbiotica, (12) Pseudomonas entomophila L48, (13) Pseudomonas fluorescens Pf-5, (14) Pseudomonas putida KT2440, (15) Stigmatella
aurantiaca DW4/3-1, (16) Pyrococcus horikoshii OT3, (17) Aspergillus oryzae RIB40, (18) Yarrowia lipolytica.
Structure-Based Protein Engineering of
specificities. The structure-based rational design of modified
b-aminopeptidases represents a promising approach to improve
b-Aminopeptidases
The superposition of the available crystal structures of b-amino- the properties of the available enzymes with respect to bio-
peptidases (Figure 5C) and a comparison of the biochemically catalytic applications, such as the enzyme-catalyzed syn-
characterized enzymes with related amino acid sequences (Fig- thesis of b-peptides or the production of enantiopure b-amino
ure 6) showed that despite the highly conserved positions of the acids.
catalytically important amino acid residues, the size and chemi-
cal environment of the substrate-binding pocket largely differs
among these proteins. We expect a nonconserved sequence
EXPERIMENTAL PROCEDURES
stretch, assuming different conformations within the substrate-
Generation of the BapA Duet Expression Constructs and Mutants
binding pocket, to mainly determine the substrate specificity of
proteins of the DmpA-like family. The wide substrate-binding
pocket of BapA allows the enzyme to catalyze reactions of
peptides with bulky N-terminal b-amino acid residues, while
Trp137 in a loop occludes this pocket and restricts DmpA to
hydrolyze substrates with sterically undemanding N-terminal
amino acids. However, by exchanging Trp137 for alanine,
DmpA could be rationally engineered to an altered enzyme
with catalytic properties that are more similar to BapA than to
DmpA (Table 3). Another interesting candidate to study the
influence of modifications within the substrate-binding pocket
on catalysis is the biochemically noncharacterized protein
PH0078 from P. horikoshii (PDB ID 2DRH). The active-site loop
of PH0078 assumes a similar conformation within the sub-
strate-binding pocket like the corresponding loop of DmpA but
exposes a sterically less demanding serine residue instead of
a tryptophane toward the active-site residues (Figure 5C). There-
fore, we expect PH0078 to exhibit a substrate profile, which is
similar to the one of the DmpA mutant W137A. Further specific
amino acid exchanges in this particularly variable region might
lead to the creation of b-aminopeptidases with interesting novel
The gene sequences coding for the individual a- and b-polypeptide chains of
BapA were amplified by PCR from the plasmid p3BapA containing bapA from
S. xenopeptidilytica 3-2W4 without its periplasmic signal sequence (Geueke
et al., 2006) and separately cloned into a pETDuet-1 vector (Novagen, Madi-
son, WI) using the following primers:
bapA a chain: fwd 50-GGAATTCCATGGGGCCGCGCGCTCGCGATCT-30,
bapA a chain: rev 50-GGAATTGGATCCTAATTCTTGTCCTGCGGCTTGC
CT-30,
bapA b chain: fwd 50-GGAATTCCCATATGTCGCTGCTGATCGTGATCG
CT-30,
bapA b chain: rev 50-GGAATTCTCGAGTCACCGGCGCGGAAACCGCGC
CT-30.
The designated point mutants were generated by site-directed mutagenesis
using the QuikChange multi-site-directed mutagenesis kit (Stratagene, La
Jolla, CA) with the following primers:
BapA S250A: 50-CCGCAGGACAAGAATGCGCTGCTGATCGTG-30
BapA S250C: 50-CCGCAGGACAAGAATTGCCTGCTGATCGTG-30
BapA E133A: 50-CCGGTGGTCGCCGCAACGCTCGACAAC-30
BapA N249A: 50-CAAGCCGCAGGACAAGGCTTCGCTGCTGATCGTG-30
BapA S288A: 50-GCGGGGGCGCTTGCGGGTGAGTTCGCG-30
BapA E290A: 50-GCGCTTTCGGGTGCGTTCGCGCTCGCC-30
BapA Duet S250A: 50-GGAGATATACATATGGCGCTGCTGATCGTG-30
1858 Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved

Structure
Structure of the b-Aminopeptidase BapA
BapA Duet S250C: 50-GAAGGAGATATACATATGTGCCTGCTGATCGT
GATC-30
BapA Duet E133A: 50-CCGGTGGTCGCCGCAACGCTCGACAAC-30
BapA Duet S288A: 50-GCGGGGGCGCTTGCGGGTGAGTTCGCG-30
BapA Duet E290A: 50-GCGCTTTCGGGTGCGTTCGCGCTCGCC-30.
(ε405 = 8,800 Mꢁ1cmꢁ1) for 1 min. For determination of kinetic constants, the
assay mixtures contained different molar concentrations of the substrates in
a 100 mM potassium phosphate buffer (pH 7.2) at 37ꢀC; 10% DMSO was
added as cosolvent. The standard assay contained 5 mM H-b3hAla-pNA as
substrate. The reactions were initiated by the addition of one of the purified
enzymes BapA, DmpA, or DmpA W137A, or by the addition of cell-free extract
containing recombinantly expressed enzyme. The initial rates of the enzyme-
catalyzed release of p-nitroanilide were calculated and fitted to the Michae-
lis-Menten model by nonlinear regression analysis with the VisualEnzymics
software (Softzymics, Princeton, NJ) for the program IGOR Pro (WaveMetrics,
Oswego, OR).
Expression and Purification of BapA for Biochemical Assays and
Protein Crystallization
The production and purification of the b-aminopeptidase BapA from
S. xenopeptidilytica 3-2W4 and DmpA from O. anthropi was done following es-
tablished procedures (Geueke et al., 2005, 2006; Fanuel et al., 1999b). All
mutants of DmpA, BapA and BapA Duet were purified accordingly. The cell-
free extracts and the partially purified proteins were analyzed by SDS-PAGE
using precast 10% Novex tricine gels (Invitrogen AG, Basel, Switzerland) ac-
cording to the manufacturer’s instructions. The protein gels were stained
with Coomassie Brilliant Blue and the relative intensities of the protein bands
were determined with a GS-800 calibrated imaging densitometer and the soft-
ware Quantity One (Bio-Rad, Reinach, Switzerland). The b-polypeptide chain
of partially purified BapA S250 Duet was transferred from the gel to a PVDF
membrane by electroblotting and the N terminus was sequenced at the Func-
tional Genomics Center Zu¨ rich (Switzerland).
ACCESSION NUMBERS
The structure data are deposited with the Protein Data Bank with the PDB
accession codes 3N5I, 3N2W, and 3N33.
ACKNOWLEDGMENTS
We thank the staff of beamline PX of the Swiss Synchrotron Light Source (PSI)
in Villigen, Switzerland for excellent technical assistance. This project was
funded by the Swiss National Science Foundation (grants to M.G.G.) and the
Swiss NCCR Structural Biology program. T.H. was supported by the Deutsche
Bundesstiftung Umwelt (DBU Project No. 13176-32), and B.G. was supported
by the Swiss National Science Foundation (SNF Project No. 3152A0-100770).
The b-peptidic substrates were a kind gift from D.S. H.-P.E.K. and B.G.
conceived the work. T.M. carried out crystallographic data collection, struc-
ture determination, and refinement. T.H. and B.G. carried out protein expres-
sion, purification, and crystallization experiments. T.H. carried out mutations
and kinetic analysis. D.S. and P.R.E.M. contributed to the enzyme mechanism.
M.G.G., T.M., and T.H. wrote the paper and prepared the figures with the help
of C.B. D.S., M.G.G., B.G., P.R.E.M., and H.-P.E.K. contributed to discussion,
data interpretation, and manuscript preparation. M.G.G., B.G., and H.-P.E.K.
supervised the work.
Before crystallization, the purified protein was dialyzed against a 0.5 mM
Tris/HCl buffer (pH 8) and concentrated to 15 mg/ml prior to crystallization.
Crystals were grown in 2 ml drops with a 1:1 ratio of mother liquor (1.8 M ammo-
nium sulfate and 100 mM Tris/HCl [pH 8]) to purified BapA in 24-well Cryschem
sitting-drop plates (Hampton Research, Aliso Viejo, CA) using the vapor-diffu-
sion method at 20ꢀC. Within 1 day, crystals of BapA grew as stacked layers of
thin plates. Sequential macroseeding with crystal fragments that were trans-
ferred to new wells containing 1.5 M ammonium sulfate, 100 mM Tris/HCl
(pH 8), yielded single three-dimensional, mostly plate-shaped crystals of up
to 300 3 300 3 30 mm in size. Cryoprotected crystals (mother liquor and
30% glycerol) were flash-frozen in liquid nitrogen.
Structure Determination
Data were collected at the Swiss Light Source beamline PX on a 6M Pilatus
˚
detector in a cryostream at 100 K at 1 A wavelength. The data sets were in-
dexed, integrated and scaled with XDS (Kabsch, 1993). The structure of
BapA was solved by molecular replacement with PHASER (McCoy et al.,
2007) using one subunit (a and b chain) of the heterooctameric structure of
DmpA (PDB ID 1B65). The search probe was additionally tailored at the
C terminus of the a chain (residues 224–245). The structure of proBapA was
solved by molecular replacement with PHASER using one subunit (a and
b chain) of native BapA. The C-terminal tail of the a chain (residues 231–249)
was omitted in this search probe. Unambiguous solutions of proBapA and
native BapA were obtained, including an extended region of positive difference
electron density (Fo-Fc) in the initial map (model phases of BapA and DmpA,
respectively) that corresponded to the omitted amino acid linker sequences.
Each polypeptide chain comprised 373 amino acids. The final model of native
BapA contained residues 1–245/250–371 for chain A, 1–243/250–371 for chain
B, 1–245/250–371 for chain C and 1–239/250–371 for chain D. Ramachandran
statistics: 1,375 (96.6%), 44 (3.1%), and 4 (0.3%) of the residues were in the
preferred, generously allowed and disallowed region, respectively. The final
model of proBapA contained residues 1–236/248–371 for chain A, 1–236/
248–317 for chain B, 1–238/248–371 for chain C and 1–371 for chain D. Ram-
achandran statistics: 1,348 (96.9%), 39 (2.8%), and 4 (0.3%) of the residues
were in the preferred, generously allowed and disallowed region, respectively.
The refinement of the structures was carried out through multiple cycles of
manual adjustments and rebuilding using COOT (Emsley and Cowtan, 2004),
and refinement was carried out using PHENIX (Adams et al., 2010). Interface
area calculations were performed using the Pisa server v.1.18. Superpositions
for figures were done using SSM (Krissinel and Henrick, 2004). Figures were
created with the program Pymol (DeLano, 2002).
Received: May 27, 2011
Revised: June 14, 2012
Accepted: July 15, 2012
Published online: September 13, 2012
REFERENCES
´
Adams, P.D., Afonine, P.V., Bunkoczi, G., Chen, V.B., Davis, I.W., Echols, N.,
Headd, J.J., Hung, L.-W., Kapral, G.J., Grosse-Kunstleve, R.W., et al. (2010).
PHENIX: a comprehensive Python-based system for macromolecular struc-
ture solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221.
Aguilar, M.-I., Purcell, A.W., Devi, R., Lew, R., Rossjohn, J., Smith, A.I., and
Perlmutter, P. (2007). b-amino acid-containing hybrid peptides—new opportu-
nities in peptidomimetics. Org. Biomol. Chem. 5, 2884–2890.
Barends, T.R.M., Yoshida, H., and Dijkstra, B.W. (2004). Three-dimensional
structures of enzymes useful for b-lactam antibiotic production. Curr. Opin.
Biotechnol. 15, 356–363.
Baumeister, W., and Lupas, A. (1997). The proteasome. Curr. Opin. Struct.
Biol. 7, 273–278.
Boanca, G., Sand, A., and Barycki, J.J. (2006). Uncoupling the enzymatic and
autoprocessing activities of Helicobacter pylori g-glutamyltranspeptidase.
J. Biol. Chem. 281, 19029–19037.
Boanca, G., Sand, A., Okada, T., Suzuki, H., Kumagai, H., Fukuyama, K., and
Barycki, J.J. (2007). Autoprocessing of Helicobacter pylori g-glutamyltrans-
peptidase leads to the formation of a threonine-threonine catalytic dyad.
J. Biol. Chem. 282, 534–541.
Activity Assay
The b-aminopeptidase-catalyzed conversion of the b3-homoamino-acid
p-nitroanilides H-bhGly-pNA, H-b3hAla-pNA, H-D-b3hAla-pNA, H-b3hPhe-pNA,
and H-D-b3hPhe-pNA was determined spectrophotometrically by moni-
Bompard-Gilles, C., Villeret, V., Davies, G.J., Fanuel, L., Joris, B., Fre` re, J.-M.,
and Van Beeumen, J. (2000). A new variant of the Ntn hydrolase fold revealed
by the crystal structure of L-aminopeptidase D-ala-esterase/amidase from
Ochrobactrum anthropi. Structure 8, 153–162.
toring the release of free p-nitroaniline at
a wavelength of 405 nm
Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved 1859

Structure
Structure of the b-Aminopeptidase BapA
Brannigan, J.A., Dodson, G., Duggleby, H.J., Moody, P.C.E., Smith, J.L.,
Tomchick, D.R., and Murzin, A.G. (1995). A protein catalytic framework with
an N-terminal nucleophile is capable of self-activation. Nature 378, 416–419.
complexes with b-lactam-derived inhibitors illustrate substrate specificity
and enantioselectivity of b-aminopeptidases. ChemBioChem. Published on-
line September 22, 2012. http://dx.doi.org/10.1002/cbic.201200393.
¨
Johansson, D.G., Wallin, G., Sandberg, A., Macao, B., Aqvist, J., and Hard, T.
Bruggink, A., Roos, E.C., and Vroom de, E. (1998). Penicillin acylase in the
industrial production of b-lactam antibiotics. Org. Process Res. Dev. 2,
128–133.
(2009). Protein autoproteolysis: conformational strain linked to the rate of
peptide cleavage by the pH dependence of the N —> O acyl shift reaction.
J. Am. Chem. Soc. 131, 9475–9477.
Buller, A.R., Freeman, M.F., Wright, N.T., Schildbach, J.F., and Townsend,
C.A. (2012). Insights into cis-autoproteolysis reveal a reactive state formed
through conformational rearrangement. Proc. Natl. Acad. Sci. USA 109,
2308–2313.
Kabsch, W. (1993). Automatic processing of rotation diffraction data from crys-
tals of initially unknown symmetry and cell constants. J. Appl. Cryst. 26,
795–800.
DeLano, W.L. (2002). The Pymol Molecular Graphics System (San Carlos, CA,
USA: DeLano Scientific).
Kim, J.K., Yang, I.S., Rhee, S., Dauter, Z., Lee, Y.S., Park, S.S., and Kim, K.H.
(2003). Crystal structures of glutaryl 7-aminocephalosporanic acid acylase:
insight into autoproteolytic activation. Biochemistry 42, 4084–4093.
Duggleby, H.J., Tolley, S.P., Hill, C.P., Dodson, E.J., Dodson, G., and Moody,
P.C.E. (1995). Penicillin acylase has a single-amino-acid catalytic centre.
Nature 373, 264–268.
Kim, J.K., Yang, I.S., Shin, H.J., Cho, K.J., Ryu, E.K., Kim, S.H., Park, S.S., and
Kim, K.H. (2006). Insight into autoproteolytic activation from the structure of
cephalosporin acylase: a protein with two proteolytic chemistries. Proc.
Natl. Acad. Sci. USA 103, 1732–1737.
Elkins, J.M., Kershaw, N.J., and Schofield, C.J. (2005). X-ray crystal structure
of ornithine acetyltransferase from the clavulanic acid biosynthesis gene
cluster. Biochem. J. 385, 565–573.
Komeda, H., and Asano, Y. (2005).
A DmpA-homologous protein from
Pseudomonas sp. is a dipeptidase specific for b-alanyl dipeptides. FEBS J.
272, 3075–3084.
Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular
graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.
Krissinel, E., and Henrick, K. (2004). Secondary-structure matching (SSM),
a new tool for fast protein structure alignment in three dimensions. Acta
Crystallogr. D Biol. Crystallogr. 60, 2256–2268.
Fanuel, L., Goffin, C., Cheggour, A., Devreese, B., Van Driessche, G., Joris, B.,
`
Van Beeumen, J., and Frere, J.M. (1999a). The DmpA aminopeptidase from
Ochrobactrum anthropi LMG7991 is the prototype of a new terminal nucleo-
phile hydrolase family. Biochem. J. 341, 147–155.
¨
Lowe, J., Stock, D., Jap, B., Zwickl, P., Baumeister, W., and Huber, R. (1995).
Crystal structure of the 20S proteasome from the archaeon T. acidophilum at
3.4 A resolution. Science 268, 533–539.
Fanuel, L., Thamm, I., Kostanjevecki, V., Samyn, B., Joris, B., Goffin, C.,
`
Brannigan, J., Van Beeumen, J., and Frere, J.M. (1999b). Two new aminopep-
tidases from Ochrobactrum anthropi active on D-alanyl-p-nitroanilide. Cell.
Mol. Life Sci. 55, 812–818.
McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C.,
and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Cryst. 40,
658–674.
Frackenpohl, J., Arvidsson, P.I., Schreiber, J.V., and Seebach, D. (2001). The
outstanding biological stability of b- and g-peptides toward proteolytic
enzymes: an in vitro investigation with fifteen peptidases. ChemBioChem 2,
445–455.
McVey, C.E., Walsh, M.A., Dodson, G.G., Wilson, K.S., and Brannigan, J.A.
(2001). Crystal structures of penicillin acylase enzyme-substrate complexes:
structural insights into the catalytic mechanism. J. Mol. Biol. 313, 139–150.
Fuchs, V., Jaeger, K.-E., Wilhelm, S., and Rosenau, F. (2011). The BapF protein
from Pseudomonas aeruginosa is a b-peptidyl aminopeptidase. World J.
Microbiol. Biotechnol. 27, 713–718.
Michalska, K., Hernandez-Santoyo, A., and Jaskolski, M. (2008). The mecha-
nism of autocatalytic activation of plant-type L-asparaginases. J. Biol. Chem.
283, 13388–13397.
Geueke, B., and Kohler, H.-P.E. (2007). Bacterial b-peptidyl aminopeptidases:
on the hydrolytic degradation of b-peptides. Appl. Microbiol. Biotechnol. 74,
1197–1204.
Oinonen, C., and Rouvinen, J. (2000). Structural comparison of Ntn-hydro-
lases. Protein Sci. 9, 2329–2337.
Saarela, J., Oinonen, C., Jalanko, A., Rouvinen, J., and Peltonen, L. (2004).
Autoproteolytic activation of human aspartylglucosaminidase. Biochem. J.
378, 363–371.
Geueke, B., Namoto, K., Seebach, D., and Kohler, H.-P.E. (2005). A novel
b-peptidyl aminopeptidase (BapA) from strain 3-2W4 cleaves peptide bonds
of synthetic b-tri- and b-dipeptides. J. Bacteriol. 187, 5910–5917.
Seebach, D., and Gardiner, J. (2008). b-peptidic peptidomimetics. Acc. Chem.
Res. 41, 1366–1375.
Geueke, B., Heck, T., Limbach, M., Nesatyy, V., Seebach, D., and Kohler,
H.-P.E. (2006). Bacterial b-peptidyl aminopeptidases with unique substrate
specificities for b-oligopeptides and mixed b,a-oligopeptides. FEBS J. 273,
5261–5272.
Suresh, C.G., Pundle, A.V., SivaRaman, H., Rao, K.N., Brannigan, J.A., McVey,
C.E., Verma, C.S., Dauter, Z., Dodson, E.J., and Dodson, G.G. (1999). Penicillin
V acylase crystal structure reveals new Ntn-hydrolase family members. Nat.
Struct. Biol. 6, 414–416.
Heck, T., Limbach, M., Geueke, B., Zacharias, M., Gardiner, J., Kohler,
H.-P.E., and Seebach, D. (2006). Enzymatic degradation of b- and mixed
a,b-oligopeptides. Chem. Biodivers. 3, 1325–1348.
Thompson, J.D., Higgins, D.G., and Gibson, T.J. (1994). CLUSTAL W:
improving the sensitivity of progressive multiple sequence alignment through
sequence weighting, position-specific gap penalties and weight matrix choice.
Nucleic Acids Res. 22, 4673–4680.
¨
Heck, T., Kohler, H.-P.E., Limbach, M., Flogel, O., Seebach, D., and Geueke,
B. (2007). Enzyme-catalyzed formation of b-peptides: b-peptidyl aminopepti-
dases BapA and DmpA acting as b-peptide-synthesizing enzymes. Chem.
Biodivers. 4, 2016–2030.
Tikkanen, R., Riikonen, A., Oinonen, C., Rouvinen, R., and Peltonen, L. (1996).
Functional analyses of active site residues of human lysosomal aspartylgluco-
saminidase: implications for catalytic mechanism and autocatalytic activation.
EMBO J. 15, 2954–2960.
Heck, T., Seebach, D., Osswald, S., Wielter, M.K.J., Kohler, H.-P.E., and
Geueke, B. (2009). Kinetic resolution of aliphatic b-amino acid amides by
b-aminopeptidases. ChemBioChem 10, 1558–1561.
Wiegand, H., Wirz, B., Schweitzer, A., Camenisch, G.P., Rodriguez Perez, M.I.,
Gross, G., Woessner, R., Voges, R., Arvidsson, P.I., Frackenpohl, J., and
Seebach, D. (2002). The outstanding metabolic stability of a 14C-labeled
b-nonapeptide in rats—in vitro and in vivo pharmacokinetic studies.
Biopharm. Drug Dispos. 23, 251–262.
Heck, T., Reimer, A., Seebach, D., Gardiner, J., Deniau, G., Lukaszuk, A.,
Kohler, H.-P.E., and Geueke, B. (2010). b-aminopeptidase-catalyzed biotrans-
formations of b2-dipeptides: kinetic resolution and enzymatic coupling.
ChemBioChem 11, 1129–1136.
Heck, T., Merz, T., Reimer, A., Seebach, D., Rentsch, D., Briand, C., Gru¨ tter,
M.G., Kohler, H.-P.E., and Geueke, B. (2012). Crystal structures of BapA
Xu, Q.A., Buckley, D., Guan, C.D., and Guo, H.C. (1999). Structural insights into
the mechanism of intramolecular proteolysis. Cell 98, 651–661.
1860 Structure 20, 1850–1860, November 7, 2012 ª2012 Elsevier Ltd All rights reserved