Structure, mechanism, and substrate profile for Sco3058: The closest bacterial homologue to human renal dipeptidase

Article abstract of DOI:10.1021/bi901935y

Human renal dipeptidase, an enzyme associated with glutathione metabolism and the hydrolysis of β-lactams, is similar in sequence to a cluster of ～400 microbial proteins currently annotated as nonspecific dipeptidases within the amidohydrolase superfamily. The closest homologue to the human renal dipeptidase from a fully sequenced microbe is Sco3058 from Streptomyces coelicolor. Dipeptide substrates of Sco3058 were identified by screening a comprehensive series of L-Xaa-L-Xaa, L-Xaa-D-Xaa, and D-Xaa-L-Xaa dipeptide libraries. The substrate specificity profile shows that Sco3058 hydrolyzes a broad range of dipeptides with a marked preference for an L-amino acid at the N-terminus and a D-amino acid at the C-terminus. The best substrate identified was L-Arg-D-Asp (k_cat/K_m = 7.6 x 10⁵ M ^-1 s^-1). The three-dimensional structure of Sco3058 was determined in the absence and presence of the inhibitors citrate and a phosphinate mimic of L-Ala-D-Asp. The enzyme folds as a (β/α)8 barrel, and two zinc ions are bound in the active site. Site-directed mutagenesis was used to probe the importance of specific residues that have direct interactions with the substrate analogues in the active site (Asp-22, His-150, Arg-223, and Asp-320). The solvent viscosity and kinetic effects of D₂O indicate that substrate binding is relatively sticky and that proton transfers do not occurr during the rate-limiting step. A bell-shaped pH-rate profile for k_cat and k_cat/K_m indicated that one group needs to be deprotonated and a second group must be protonated for optimal turnover. Computational docking of high-energy intermediate forms of L/D-Ala-L/D-Ala to the three-dimensional structure of Sco3058 identified the structural determinants for the stereochemical preferences for substrate binding and turnover.

Full text of DOI:10.1021/bi901935y

Biochemistry 2010, 49, 611–622 611
DOI: 10.1021/bi901935y
Structure, Mechanism, and Substrate Profile for Sco3058: The Closest Bacterial
†,‡
Homologue to Human Renal Dipeptidase
§
§
^
§
Jennifer A. Cummings, Tinh T. Nguyen, Alexander A. Fedorov, Peter Kolb, Chengfu Xu, Elena V. Fedorov,
^
§
,§
Brian K. Shoichet, David P. Barondeau, Steven C. Almo, and Frank M. Raushel*
§
Department of Chemistry, P.O. Box 30012, Texas A&M University, College Station, Texas 77843, Albert Einstein College
of Medicine, 1300 Morris Park Avenue, Bronx, New York 10461, and Department of Pharmaceutical Chemistry,
^
th
University of California, 1700 4 Street, San Francisco, California 94158-2330
Received November 10, 2009; Revised Manuscript Received December 11, 2009
ABSTRACT: Human renal dipeptidase, an enzyme associated with glutathione metabolism and the hydrolysis
of β-lactams, is similar in sequence to a cluster of ∼400 microbial proteins currently annotated as nonspecific
dipeptidases within the amidohydrolase superfamily. The closest homologue to the human renal dipeptidase
from a fully sequenced microbe is Sco3058 from Streptomyces coelicolor. Dipeptide substrates of Sco3058
were identified by screening a comprehensive series of
dipeptide libraries. The substrate specificity profile shows that Sco3058 hydrolyzes a broad range of
dipeptides with a marked preference for an -amino acid at the N-terminus and a -amino acid at the
C-terminus. The best substrate identified was
dimensional structure of Sco3058 was determined in the absence and presence of the inhibitors citrate and a
phosphinate mimic of -Ala-
L
-Xaa-
L
-Xaa,
L-Xaa-
D
-Xaa, and
D
-Xaa-
L
-Xaa
L
D
5
-1 -1
L
-Arg-
D
-Asp (kcat/Km = 7.6 ꢀ 10 M
s ). The three-
L
D
-Asp. The enzyme folds as a (β/R) barrel, and two zinc ions are bound in the
8
active site. Site-directed mutagenesis was used to probe the importance of specific residues that have direct
interactions with the substrate analogues in the active site (Asp-22, His-150, Arg-223, and Asp-320). The
solvent viscosity and kinetic effects of D O indicate that substrate binding is relatively sticky and that proton
2
transfers do not occurr during the rate-limiting step. A bell-shaped pH-rate profile for k and kcat/Km
cat
indicated that one group needs to be deprotonated and a second group must be protonated for optimal
turnover. Computational docking of high-energy intermediate forms of
L
/
D-Ala-
L
/ -Ala to the three-
D
dimensional structure of Sco3058 identified the structural determinants for the stereochemical preferences
for substrate binding and turnover.
The mammalian renal dipeptidase catalyzes the general reac-
dases and some of their microbial homologues have been
assayed with various substrates, but a complete substrate
specificity profile for these types of enzymes has not been
determined (2-4, 6, 10-12).
tion outlined in Scheme 1 (1-6). This enzyme is also able to
catalyze the hydrolysis of leukotriene D (2, 3, 7) as well as the
4
oxidized degradation product of glutathione, -Cys-Gly (2-4).
L
1
Human renal dipeptidase (hRDP) is the only enzyme in humans
that has been shown to catalyze the hydrolysis of β-lactams such
as SCH 29482, imipenem, meropenem, and DA-1131 (8, 9).
However, dipeptides are consistently better substrates for hRDP
than are β-lactams (2, 7, 10). Glycine is the preferred amino acid
at the C-terminus, relative to an amino acid in the L-configura-
tion, and glycyldehydrophenylalanine is a better substrate for
hRDP than glycylphenylalanine (10). With rat renal dipeptidase,
The crystal structure of hRDP has been determined with the
inhibitorcilastatinboundinthe active site(PDBentry1ITU) (13).
Cilastatin is considered to be a dipeptide analogue, but it does not
contain a free amino group at the N-terminus and is not
hydrolyzed by hRDP. The metal center in the active site consists
of two Zn ions that are bridged by a molecule from solvent (water
or hydroxide) and the side chain carboxylate of Glu-125. The
R-Zn is additionally ligated by His-20 and Asp-22, whereas the
β-Zn is coordinated to His-198 and His-219. It has been suggested
that the water that bridges the two metal ions attacks the
carbonyl of the peptide bond to form a tetrahedral intermediate
that is stabilized by an electrostatic interaction with His-152. This
L-Ala-Gly is hydrolyzed faster than
L-Ala-L-Ala and Gly-D-Ala is
hydrolyzed faster than Gly-L-Ala (3). Mammalian renal dipepti-
†
This work was supported in part by the National Institutes of Health
GM71790), the Robert A. Welch Foundation (A-840), and the Hackerman
(
residue is found at the end of β-strand 4 of the (β/R) barrel
8
Advanced Research Program (010366-0034-2007). P.K. was supported by
the Swiss National Science Foundation for a Fellowship for Prospective
structure and has been shown to be important for the binding of
the cilastatin inhibitor (13, 14).
Researchers (PBZHA-118815).
‡
The X-ray coordinates and structure factors for Sco3058 have been
The structure of hRDP confirms that this enzyme is a member
of the amidohydrolase superfamily (AHS). This is a rather large
enzyme superfamily that has been shown to catalyze the hydro-
lysis of ester and amide bonds found within carbohydrate-,
peptide-, and nucleic acid-containing substrates (15). The hRDP
is contained within a large cluster of proteins designated as
cog2355E. This cluster of proteins is typified by an HxD motif at
deposited in the Protein Data Bank as entries 3ID7, 3ISI, and 3K5X.
*To whom correspondence should be addressed. Telephone: (979)
8
45-3373. Fax: (979) 845-9452. E-mail: raushel@tamu.edu.
Abbreviations: hRDP, human renal dipeptidase; AHS, amidohydro-
1
lase superfamily; IPTG, isopropyl β-D-thiogalactopyranoside; PEG, poly-
ethylene glycol; PDB, Protein Data Bank; SDS-PAGE, sodium dodecyl
sulfate-polyacrylamide gel electrophoresis; rmsd, root-mean-square
deviation.
r 2009 American Chemical Society
Published on Web 12/15/2009
pubs.acs.org/Biochemistry

6
12 Biochemistry, Vol. 49, No. 3, 2010
Cummings et al.
Scheme 1
Scheme 2
the end of β-strand 1 that is used to coordinate one of the two
metal ions in the active site, whereas most of the other enzymes in
this superfamily have an HxH motif. In hRDP, the invariant
aspartate at the end of β-strand 8 is not coordinated to the
R-metal, whereas in nearly all of the other members of the AHS
that have been structurally characterized, this residue is a direct
ligand to the R-metal ion. The proposed chemical mechanism for
the hydrolysis of dipeptides by hRDP is at variance with those
mechanisms of amide bond hydrolysis proposed previously for
other members of the AHS (16-18). In hRDP, the carbonyl
group of the amide bond about to be hydrolyzed is proposed to
be polarized by an electrostatic interaction with His-152, whereas
in other members of the AHS, this function is contributed by a
direct interaction with the β-metal ion (13, 16-18). The renal
dipeptidase-like enzyme BCS-1 from Brevibacillus borstelensis
has a tryptophan in place of the residue equivalent to His-152.
N-Fmoc-D-Xaa-OH amino acids were substituted for N-Fmoc-L-
Xaa-OH amino acids. These compound libraries contain a single
amino acid at the N-terminus and all of the common amino acids
at the C-terminus. However, amino acids L-Cys, D-Cys, and D-Ile
were not incorporated into any of the dipeptide libraries. A
sample of these libraries, expected to contain ∼2 nmol of each
library component, was submitted for amino acid analysis
(
AAA) to quantify the amount of every amino acid in these
libraries except for tryptophan. The amount of tryptophan in
these libraries was determined by the absorbance at 280 nm.
Synthesis of Phosphinate Pseudodipeptides. Three race-
mic phosphinate pseudodipeptide mimics of the putative tetra-
hedral intermediates for the hydrolysis of Ala-Asp (1), Phe-Asp
(
2), and Tyr-Asp (3) were created as potential inhibitors of
Sco3058. These compounds were synthesized according to pre-
5
BCS-1 is reported to have respectable activity (k /K = 2 ꢀ 10
cat
m
viously described protocols (23-26). The structures of these
-
1 -1
M
s
), which supports the conclusion that either His-152 does
31
compounds are presented in Scheme 2. Compound 1:
P
not play a crucial role in the chemical mechanism or the
mechanism for substrate hydrolysis of enzymes such as BCS-1
varies from that of hRDP (11).
The putative dipeptidase Sco3058 from Streptomyces coelicolor
is the closest bacterial homologue to the human enzyme from a
completely sequenced microbial organism. The level of sequence
identity is ∼41%. S. coelicolor is a member of a family of soil
bacteria known for their versatile metabolism and production of
antibiotics and pharmaceutical agents (19). In this paper, we
unravel the substrate specificity of Sco3058 by determining the
NMR (121.4 MHz, D O) δ 36.24, 36.20; MS (ESI negative
2
-
mode) m/z 238.0129 (M - H) ; MS (ESI positive mode) m/z
þ
2
40.0977 (M þ H) ; C H O NP requires m/z 239.0559. Com-
7
14
6
31
pound 2: P NMR (121.4 MHz, D O) δ 35.01; MS (ESI negative
2
-
mode) m/z 314.1116 (M - H) ; MS (ESI positive mode) m/z
þ
3
16.0786 (M þ H) ; C H O NP requires m/z 315.0872. Com-
13
18 6
31
pound 3: P NMR (121.4 MHz, D O) δ 34.71; MS (ESI negative
2
-
mode) m/z 330.0984 (M - H) ; MS (ESI positive mode) m/z
þ
3
32.0828 (M þ H) ; C H O NP requires m/z 331.0821.
1
3
18 7
Cloning and Expression of Sco3058. The gene for Sco3058
catalytic activity with 55
L-Xaa-L-Xaa, L-Xaa-D-Xaa, and D-Xaa-
(
gi|21221500) was amplified from S. coelicolor A3(2) genomic
L-Xaa dipeptide libraries. The results demonstrate that this
0
DNA with the primers 5 -GCAGGAGCCATATGACAT-
CGCTGGAGAAGGCCCGCGAGCTGCTGCGC-3 and 5 -
CGCGGAATTCAGCCTTCCGGCTGCTCGGCGGCCGT-
enzyme is a rather promiscuous dipeptidase that is able to
hydrolyze a large fraction of the ∼1000 compounds screened.
Nevertheless, Sco3058 exhibited some degree of specificity, and
0
0
0
GCCG-3 . The Pfx platinum polymerase (Invitrogen), 0.05 unit/μL
the best substrate identified was
L
-Arg-
D
-Asp (kcat/Km = 7.6 ꢀ
PCR mixture, was used according to the manufacturer’s instruc-
tions except the annealing step was omitted. The PCR product
was digested with NdeI and EcoRI and ligated into the pET30a-
5
-1 -1
10 M
s
). The structure of Sco3058 has been determined to a
˚
resolution of 1.3 A in the presence and absence of dipeptide
mimics. The chemical mechanism has been addressed through the
utilization of pH-rate profiles, changes in solvent viscosity, and
the utilization of site-directed mutants.
(
þ) vector (Novagen) with T4 ligase (New England Biolabs).
Escherichia coli BL21 Rosetta 2 (DE3) cells were electro-trans-
formed with the pET30a(þ)-Sco3058 plasmid. Cells harboring
the pET30a(þ)-Sco3058 plasmid were grown at 30 ꢀC in Terrific
Broth containing 25 and 50 μg/mL chloramphenicol and kana-
mycin, respectively. When the optical density at 600 nm reached
MATERIALS AND METHODS
Materials. All chemicals were obtained from Sigma Aldrich,
unless otherwise stated. The coupling enzymes, glutamate-
oxaloacetate transaminase and malate dehydrogenase, were
purchased from Sigma Aldrich and Calbiochem, respectively.
∼0.6, 2.5 mM Zn(acetate) was added and the expression of
2
Sco3058 was induced with 1.0 mM isopropyl β-thiogalactoside
(IPTG). The cells were grown overnight at room temperature.
Site-directed mutagenesis was conducted on the gene for Sco3058
at residues Asp-22, His-150, Arg-223, and Asp-320 using the
QuikChange kit from Stratagene.
Glycyl-D-Leu was purchased from TCI America. ICP standards
were obtained from Inorganic Ventures Inc. Nitrocefin was
purchased from Calbiochem. Metal analyses were conducted
using inductively coupled plasma mass spectrometry (ICP-MS)
as previously described (20). Oligonucleotide syntheses and DNA
sequencing were performed by the Gene Technologies Lab of
Texas A&M University.
Purification of Sco3058. The cells containing Sco3058 were
suspended in a purification buffer [20 mM Tris with 100 μg/mL
phenylmethanesulfonyl fluoride (pH 7.5)] and lysed via sonica-
tion at 0 ꢀC. The cell lysate was obtained by centrifugation and
subsequently treated with protamine sulfate [2% (w/w) of cell
mass] and removed by centrifugation. The protein was precipi-
tated at 60% ammonium sulfate (371 g/L at 0 ꢀC), isolated by
centrifugation, resuspended in 20 mM Tris (pH 7.5), and purified
Synthesis of Dipeptide Libraries. The
-Xaa- -Xaa dipeptide libraries were synthesized as described
previously (21, 22). The 17 -Xaa- -Xaa libraries were synthe-
sized in the same manner as the -Xaa- -Xaa libraries except that
L-Xaa-D-Xaa and
L
L
D
L
L
L

Article
Biochemistry, Vol. 49, No. 3, 2010 613
by size exclusion chromatography using a 3 L column of AcA34
in 20 mM Tris (pH 7.5) at a flow rate of 1.0 mL/min. The
fractions containing the protein of expected size, as assessed by
SDS-PAGE, were combined. Sco3058 was further purified by
ion exchange chromatography with a Resource Q column
absorbance at 507 nm (Q) was plotted as a function of Sco3058
concentration (E ) and fit to eq 1. In the case of the D-Xaa-L-Xaa
t
libraries, the free amino acids were quantified by measuring the
absorbance at 507 nm and the linear rates of hydrolysis as
a function of enzyme concentration were compared for each
(
Pharmacia) using a gradient of NaCl in 50 mM Tris (pH 7.5).
A final gel filtration step, with a column of Superdex 200
Amersham Biosciences) in 50 mM HEPES (pH 7.5), was utilized
D-Xaa-L-Xaa library.
-kE
t
(
Q ¼ Að1 -e
Þ
ð1Þ
to remove the salt.
Ninhydrin-Based Enzyme Assays. The measurement of
dipeptidase activity was conducted in 50 mM HEPES (pH 7.5).
For measurement of the inhibitory properties of the phosphinate
C-Terminal Substrate Specificity of Sco3058. Enzyme-
course assays for six dipeptide libraries (
L
-Arg- -Trp-
L
-Xaa,
L
L
-
-
Xaa, -Ala- -Xaa, -Met- -Xaa, -Arg-
L
L
L
L
L
D
-Xaa, and D-Leu-
L
pseudodipeptides, 0.40 mM
strate and the inhibitor concentrations were varied from 0 to
00 μM. All assays were conducted at 30 ꢀC. The formation of
free amino acids was quantified using a ninhydrin-based assay as
previously described (27). One volume of the enzymatic reaction
was mixed with 2 volumes of ice-cold Cd-ninhydrin reagent
L
-Leu-
D
-Ala was used as the sub-
Xaa) were analyzed by HPLC for determination of the
C-terminal specificity. With the exception of the -Leu- -Xaa
library, each library was incubated at 30 ꢀC for 5 h, with and
without enzyme, in 25 mM (NH )HCO (pH 8.0). The -Leu-
Xaa library was incubated under the same conditions for 18 h.
The concentrations of enzyme used were 0-1000 nM ( -Ala-
Xaa), 0-200 nM ( -Arg- -Xaa), 0-250 nM ( -Arg- -Xaa,
-Trp- -Xaa, and -Met- -Xaa), or 20-4000 nM ( -Leu- -Xaa).
D
L
1
D
L-
4
3
L
L-
[
0.9% (w/v) ninhydrin dissolved in a 1:10:80 1 g/mL CdCl /acetic
L
D
L
L
2
acid/EtOH mixture], heated at 80 ꢀC for 5 min, and cooled prior
to the absorbance being read at 507 nm. The correlation between
the absorbance at 507 nm and the concentration of free amino
acids was determined for each amino acid.
L
L
L
L
D
L
The sample preparation and data analysis were conducted as
previously described (21, 22).
Data Analysis. The kinetic parameters, kcat, K , and k /K ,
m
cat
m
β-Lactamase Assay. Sco3058 was tested for its ability to
hydrolyze the β-lactam nitrocefin. The hydrolysis of nitro-
cefin was monitored by the increase in absorbance at 486 nm
were determined by fitting the initial velocity data to eq 2
v=E
t
¼ ðkcat½AꢁÞ=ðK
m
þ ½AꢁÞ
ð2Þ
-
1
-1
(
ε = 20500 M cm ) in the presence of 100 mM potassium
phosphate (pH 7.0). The kcat/K was estimated by fitting the data
where v is the initial velocity, E is the total enzyme concentration,
m
t
to eq 2.
kcat is the turnover number, [A] is the substrate concentration,
Coupled Enzyme Assays. The hydrolysis of
L-Arg-L-Asp
and K is the Michaelis constant. The profiles for the variation of
m
was monitored by coupling the formation of -aspartate to the
L
kcat and kcat/Km with pH were fit to eq 3
oxidation of NADH in a system that includes glutamate-oxa-
loacetate transaminase and malate dehydrogenase. The change in
the concentration of NADH was measured spectrophotometri-
cally using a SPECTRAmax-340 plate reader (Molecular Devices
Inc.) by following the decrease in absorbance at 340 nm. The
standard assay conditions included 100 mM Tris (pH 7.5),
log y ¼ log½c=ð1 þ H=K
a
þ K
=HÞꢁ
ð3Þ
b
where c is the pH-independent value of y, K and K are the
a
b
dissociation constants of the ionizable groups, and H is the
proton concentration. The kinetic constants for the inhibition of
Sco3058 by the pseudodipeptide mimics were obtained by a fit of
the data to eq 4
varying concentrations of L-Arg-L-Asp, 0.36 mM NADH, 7 units
of glutamate-oxaloacetate transaminase, 1.0 unit of malate
dehydrogenase, 3.7 mM R-ketoglutarate, 100 mM KCl, and
Sco3058 in a final volume of 250 μL at 30 ꢀC. This assay was used
v=E
t
¼ ðkcat½AꢁÞ=½K
m
ð1 þ I=K
i
Þ þ Aꢁ
ð4Þ
to determine the kinetic parameters of
L-Arg-L-Asp and L-Ala-L-
Asp and the K of citrate versus the -Arg-
L
L-Asp substrate.
where K is the competitive inhibition constant.
i
i
The pH dependence of the kinetic parameters, kcat and kcat/Km,
was measured over the pH range of 5.5-10.0 at 0.25 pH unit
intervals for the zinc-substituted enzyme. The buffers used for the
pH-rate profiles were MES, Bis-Tris, Tris, and CABS. The pH
values were recorded after the completion of the assays. The
effects of solvent viscosity on the activity of Sco3058 were
determined at pH 7.5 using sucrose as the microviscogen at
Crystallization and Data Collection. Crystals of Sco3058
were grown by hanging drop vapor diffusion at room tempera-
ture using 0.7-0.9 M citrate, 0.1 M imidazole (pH 8.0), and
12.0 mg/mL enzyme. Crystals were observed within 2 weeks and
exhibited diffraction consistent with space group P3 21, with one
1
molecule of Sco3058 per asymmetric unit and 62% solvent. Prior
to data collection, the citrate-grown crystals were transferred to a
cryoprotectant solution composed of 80% mother liquor and
20% glycerol and flash-cooled and stored in liquid nitrogen.
Data for the glycerol citrate Sco3058 complex were collected to
2
5 ꢀC. The concentrations of sucrose were 0, 10, 14, 20, 24, and
32% (w/w), with corresponding η values of 1.0, 1.3, 1.5, 1.9, 2.2,
rel
and 3.2, respectively. The solvent isotope effects on the kinetic
3
3
˚
parameters of the wild-type Sco3058 were measured in 99% D O
a maximum resolution of 1.7 A using the single-wavelength
SSRL 7-1 beamline (Stanford Synchrotron Radiation Light-
source), and intensities were integrated and scaled with DENZO
and SCALEPACK (28).
2
at pD 7.90.
N-Terminal Substrate Specificity of Sco3058. Dipeptide
library screens were conducted in 50 mM HEPES (pH 8.0) at
2þ
3
0 ꢀC. Enzyme-course studies (variable enzyme concentrations
for a fixed time period) were performed by incubating 19 -Xaa-
-Xaa, 19 -Xaa- -Xaa, and 17 -Xaa- -Xaa libraries with
-2000 nM Sco3058 for 6.5, 6.5, and 18 h, respectively.
For the -Xaa- -Xaa and -Xaa- -Xaa dipeptide libraries, the
The crystals of wild-type Sco3058 complexed with Zn were
grown by vapor diffusion at room temperature using the follow-
ing crystallization conditions. The protein solution contained
wild-type Sco3058 (14.0 mg/mL) in 50 mM HEPES (pH 7.5);
the precipitant contained 22% poly(acrylic acid), 0.1 M HEPES
L
L
L
D
D
L
2
L
L
L
D

6
14 Biochemistry, Vol. 49, No. 3, 2010
Cummings et al.
Table 1: Crystallographic Statistics for Sco3058
Sco3058
Sco3058 citrate
Sco3058 inhibitor 1
3
3
Data Collection
beamline
NSLS X4A
0.97915
SSRL 7-1
NSLS X4A
0.97915
˚
wavelength (A)
space group
P3
1
21
P3
1
21
1
P3 21
no. of molecules per asymmetric unit
cell dimensions
1
1
1
˚
a, b, c (A)
96.75, 96.75, 104.72
90.0, 90.0, 120.0
25-1.3 (1.35-1.30)
133376 (9677)
0.067 (0.449)
97.29, 97.29, 104.41
90.0, 90.0, 120.0
1.7
96.91, 96.91, 104.19
90.0, 90.0, 120.0
25-1.4 (1.45-1.40)
106229 (7419)
0.072 (0.430)
R, β, γ (deg)
resolution (A)
a
˚
a
no. of unique reflections
62182
a
R
merge
a
0.071
I/σ
35.6 (4.2)
29.7 (4.5)
98.3 (95.7)
28.8 (3.8)
a
completeness (%)
95.9 (74.1)
95.3 (70.9)
Refinement
a
˚
resolution (A)
25.0-1.3
0.184
0.198
2980
20.0-1.7
16.7
25.0-1.4
0.193
0.206
2980
R
R
cryst
free
20.3
no. of protein atoms
no. of waters
3055
390
504
417
˚
rmsd, bond lengths (A)
rmsd, bond angles (deg)
bound inhibitor
no. of inhibitor atoms
no. of bound ions
PDB entry
0.004
1.3
0.018
0.004
1.3
2.4
citrate, glycerol
19
L-Ala-D-Asp pseudodipeptide
15
2þ
2þ
2þ
2 Zn
3ID7
2 Zn
2 Zn
3ITC
3K5X
a
Numbers in parentheses indicate values for the highest-resolution shell.
(
pH 7.5), and 20 mM ZnCl . Crystals appeared in 3 days and
with TOM (36), and refinement with CNS (31) were performed.
2
˚
exhibited diffraction consistent with space group P3 21, with one
The model was refined at 1.3 A with an R
of 0.184 and an Rfree
1
cryst
molecule of Sco3058 per asymmetric unit. Prior to data collec-
tion, the crystals were transferred to cryoprotectant solutions
composed of mother liquors supplemented with 20% glycerol
and flash-cooled in a nitrogen stream. Data for Sco3058 with Zn
of 0.198. The final structure contained protein residues 1-391,
2þ
390 water molecules, and two well-defined Zn ions. Structure
2þ
determination and refinement for the Sco3058 Zn pseudo-
3
3
dipeptide ternary complex were analogous to the procedures
2
þ
˚
were collected to 1.3 A resolution at the NSLS X4A beamline
used for the Sco3058 Zn
complex. The model of the
˚
3
2þ
(
Brookhaven National Laboratory) on an ADSC CCD detector.
Sco3058 Zn pseudodipeptide complex was refined at 1.4 A
of 0.193 and an R of 0.206. The final structure
includes protein residues 1-391, 417 water molecules, two Zn
3
3
Intensities were integrated and scaled with DENZO and SCA-
with an R
cryst
free
2þ
LEPACK (28).
2þ
The crystals containing the Zn -bound enzyme complex were
soaked for 4 h in cryo-buffer containing 22% poly(acrylic acid),
ions, and one well-defined pseudodipeptide inhibitor bound in
the active site. Final crystallographic refinement statistics for all
complexes are listed in Table 1.
0.1 M HEPES (pH 7.5), 0.5 mM ZnCl , 30% ethylene glycol, and
2
5
0
mM pseudodipeptide inhibitor 1. Crystals of the
Docking Calculations. In parallel with the experimental
assays, a comprehensive library of dipeptides and capped amino
acids was generated computationally. This library was composed
2þ
Sco3058 Zn pseudodipeptide ternary complex were flash-
cooled in a nitrogen stream. Data collection and reduction for
the ternary complex were analogous to the procedures used for
the Sco3058 Zn complex. The data collection statistics for all
crystals are listed in Table 1.
3
3
of all
N-succinylated, N-formylated, N-carbamoylated, N-formimi-
noylated, and hydantoin variants of individual - and -amino
L-L, L-D, D-L, and D-D dipeptides as well as N-acetylated,
2þ
3
L
D
Structure Determination and Model Refinement. The
Sco3058 glycerol citrate structure was determined by molecular
replacement using AMoRe with the renal dipeptidase from
humans (PDB entry 1ITQ) as the template (29, 30). Manual
model building and solvent building were performed with Xfit,
acids. All molecules were transformed to their high-energy
intermediate (HEI) forms by in silico reaction with a hydroxide
anion. The HEI form of a potential substrate mimics the
tetrahedral state of the molecule immediately after the attack
of water or hydroxide (37, 38). This state of a substrate should
show higher complementarity toward the enzyme, as it is the
form that the enzyme is preconfigured to stabilize (39). Since the
reaction center is tetrahedral, two enantiomers of each HEI exist,
bringing the total number of diastereomers of any given dipeptide
to eight. This HEI library was docked into the binding site of
Sco3058 with DOCK 3.5.54 as described previously (40). We note
that the score assigned to each pose of a docked dipeptide is based
on its fit to the structure (the score consists of van der Waals and
3
3
˚
and the structure was refined to 1.7 A with CNS (31-33).
2þ
Electron density corresponding to the bound citrate and Zn
ions was evident in the active site.
2þ
The structure of wild-type Sco3058 complexed with Zn was
similarly determined by molecular replacement with PHA-
SER (34), using the coordinates of human renal dipeptidase
(
PDB entry 1ITQ) as the search model. Iterative cycles of
automatic model rebuilding with ARP (35), manual rebuilding

Article
Biochemistry, Vol. 49, No. 3, 2010 615
F
IGURE 1: Typical enzyme-course plots for the hydrolysis of dipep-
tide libraries. (A) L-Arg-L-Xaa treated with various concentrations
of Sco3058. (B) D-Val-L-Xaa treated with various concentrations
of Sco3058. Additional details can be found in the text.
electrostatic terms and is corrected for desolvation) and does not
contain any terms reflecting differences in reactivity. For a library
in which the reactive center is consistent, the latter term can be
assumed to be approximately constant, allowing conclusions to
be drawn from complentarity alone. To investigate the structural
determinants for binding in more detail, the four diastereomers of
the Ala-Ala dipeptide (L-L, L-D, D-L, and D-D) were analyzed
separately from the rest and directly compared with the experi-
mentally determined turnover rates.
RESULTS
Cloning, Expression, and Purification of Sco3058. The
gene for Sco3058 was successfully cloned into the pET30a(þ)
vector. The protein was expressed at moderate levels and purified
to homogeneity. The identity of Sco3058 was confirmed by
sequencing the first five amino acids from the N-terminus of
the purified protein. The amino acid sequence (90% MTSLE and
F
IGURE 2: N-Terminal specificity for 55 dipeptide libraries. (A)
Relative rates of hydrolysis for 19 L-Xaa-L-Xaa dipeptides. (B)
Relative rates of hydrolysis for 19 L-Xaa-D-Xaa dipeptide libraries.
(C) Relative rates of hydrolysis for 17 D-Xaa-L-Xaa dipeptide
libraries.
1
0% TSLEK) matched that of MTSLEK predicted from the
and glutamine. For the
preference at the N-terminus is for arginine, glutamine, methio-
nine, and leucine. With the series of -Xaa- -Xaa dipeptide
L-Xaa-D-Xaa series of dipeptides, the
DNA sequence for this enzyme. ICP-MS confirmed the presence
of 2.0 equiv of Zn per subunit.
D
L
N-Terminal Substrate Specificity of Sco3058. The sub-
strate specificity of Sco3058 for the amino acid at the N-terminus
of the various dipeptide libraries was determined by measuring
the composite rate of dipeptide hydrolysis in these libraries. A
libraries, the preferred amino acids at the N-terminus are
methionine, leucine, arginine, and lysine.
C-Terminal Substrate Specificity for Sco3058. The ob-
served rates for the liberation of specific amino acids from the
typical enzyme-course plot for the hydrolysis of the
library is presented in Figure 1A. An enzyme-course plot for the
slower hydrolysis of the -Xaa- -Xaa dipeptide libraries using
-Tyr- -Xaa as an example is presented in Figure 1B. The relative
hydrolysis rates for the -Xaa- -Xaa, -Xaa- -Xaa, and -Xaa-
L
-Arg-
L
-Xaa
C-terminus of the dipeptides contained within
Arg- -Xaa, -Ala- -Xaa, -Trp- -Xaa, -Met- -Xaa, and
-Xaa are presented in Figure 3A-F, respectively. The relative
rates were obtained from fits of the data to eq 1. Within each
library, dipeptides that terminated with -Asp were hydrolyzed the
L
-Arg-
L
-Xaa,
L-
D
L
L
L
L
L
L
D-Leu-
D
L
L
D
L
L
L
L
D
D
L-
L
Xaa dipeptide libraries are presented in Figure 2A-C. These
libraries contain a fixed amino acid at the N-terminus and 17-19
fastest. Proline was not detected in any of the hydrolysis products,
and thus, dipeptides that terminate in a proline are not hydrolyzed.
different amino acids at the C-terminus. For the
L-Xaa-L-Xaa
Dipeptides with
were consistently hydrolyzed more slowly than the 14 remaining
-amino acids given the same amino acid at the N-terminus.
L-Val, L-Thr, L-Arg, or L-Lys at the C-terminus
dipeptide libraries, the N-terminal preference was for a leucine
residue with slightly lower turnovers for methionine, arginine,
L

6
16 Biochemistry, Vol. 49, No. 3, 2010
Cummings et al.
-1
F
IGURE 3: C-Terminal specificity for Sco3058 with six dipeptide libraries. kobs (nM ) for the hydrolysis of each dipeptide within the (A) L-Arg-L-
Xaa, (B) L-Arg-D-Xaa, (C) L-Ala-L-Xaa, (D) L-Trp-L-Xaa, (E) L-Met-L-Xaa, and (F) D-Leu-L-Xaa dipeptide libraries. All data for L-amino acids
are shaded in gray except for those of L-Asp (red). Data for the D-amino acids are shaded in yellow except for those of D-Asp (blue).
Kinetic Constants for Selected Substrates. The kinetic
constants, kcat, K , and k /K , for 25 dipeptide substrates are
a
Table 2: Kinetic Constants for Select Substrates of Sco3058
m
cat
m
-1
-1 -1
substrate
k
cat (s
)
K
m
(mM)
k
cat/K
m
(M
s
)
presented in Table 2. Of the compounds tested,
L
-Ala-
-Arg- -Asp has the
). For the four
enantiomers of the Ala-Ala dipeptide, the relative rates of
hydrolysis are in the following order: -Ala- -Ala > -Ala-
Ala > -Ala- -Ala > -Ala- -Ala. The rates of hydrolysis of
N-formyl- -Glu, N-acetyl- -Glu, and N-propionyl- -Leu at a
concentration of 2.0 mM were less than 2 ꢀ 10
was slowly hydrolyzed by Sco3058, but the value of kcat/Km was
L-Asp has
-
1
the highest value of kcat (1400 s ) and
L
D
5
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
-Arg- -Asp
D
107 ( 3.2
340 ( 4
194 ( 4
135 ( 2
160 ( 3
400 ( 10
750 ( 25
24 ( 0.2
310 ( 6
288 ( 10
1400 ( 60
140 ( 6
160 ( 11
132 ( 2
21 ( 0.3
0.14 ( 0.02
0.89 ( 0.04
0.56 ( 0.03
0.55 ( 0.03
0.75 ( 0.04
2.0 ( 0.2
4.4 ( 0.3
0.17 ( 0.01
2.3 ( 0.1
2.4 ( 0.2
13.8 ( 0.9
1.4 ( 0.1
2.0 ( 0.3
1.8 ( 0.1
(7.6 ( 0.9) ꢀ 10
(3.8 ( 0.2) ꢀ 10
(3.5 ( 0.3) ꢀ 10
(2.5 ( 0.1) ꢀ 10
(2.1 ( 0.1) ꢀ 10
(2.0 ( 0.1) ꢀ 10
(1.7 ( 0.1) ꢀ 10
5
-1 -1
highest value of kcat/Km (7.6 ꢀ 10
M
s
5
5
5
5
5
5
-Met-
D
-Glu
-Leu
-Met-
D
L
D
L
L-
-Leu-
-Leu-
-Leu-
D
-Glu
-Ala
-Ser
D
D
D
D
D
L
D
D
D
-Ala-
D
-Ala
-Leu
-
3 -1
s
. Nitrocefin
5
-Tyr-
D
(1.5 ( 0.07) ꢀ 10
5
-Met-
L
-Leu
-Glu
-Asp
(1.4 ( 0.1) ꢀ 10
-1
-1
5
less than 0.1 M
s
. The phosphinate pseudodipeptide mimics
-Asn-
D
(1.2 ( 0.1) ꢀ 10
5
-Ala-
-Leu-
-Leu-
-Arg-
-Leu-
L
(1.0 ( 0.1) ꢀ 10
1, 2, and 3 were found to be competitive inhibitors of Sco3058
5
L
-Leu
-Asp
-Asp
(1.0 ( 0.1) ꢀ 10
with K values of 390 ( 21 nM, 2.3 ( 0.3 μM, and 1.1 ( 0.2 μM,
i
4
L
L
(8.3 ( 1.4) ꢀ 10
respectively. The K of citrate was 25 ( 15 mM.
4
i
(7.2 ( 0.4) ꢀ 10
pH-Rate Profiles. The kinetic constants for the hydrolysis
4
D
-Leu
-Leu
-Ala- -Glu
-Arg-Gly
0.37 ( 0.03
(5.6 ( 0.4) ꢀ 10
4
of the
L-Arg-L-Asp dipeptide were obtained as a function of pH.
Gly-
D
(3.3 ( 0.1) ꢀ 10
4
The pH-rate profiles for kcat and kcat/K are presented in panels
L
L
L
L
L
L
L
220 ( 14
140 ( 17
9.5 ( 1.0
16 ( 3.3
(2.4 ( 0.3) ꢀ 10
m
4
(2.2 ( 0.1) ꢀ 10
A and B of Figure 4, respectively. The pH profiles are bell-shaped
and are consistent with a single functional group that must be
unprotonated for activity and another group that must be
4
2
-Ala-
-Ala-
-Ala-
-Ala-
L-Ala
L-Gln
L-His
L-Arg
(1.4 ( 0.04) ꢀ 10
3
(8.4 ( 1.9) ꢀ 10
2
(4.5 ( 0.1) ꢀ 10
2
protonated for catalytic activity. The pK values for the effect
(1.8 ( 0.1) ꢀ 10
a
of pH on kcat (Figure 4A) were 6.2 ( 0.1 and 8.4 ( 0.1. In
Figure 4B, the two ionizations observed for Sco3058 are less than
D
-Ala-
D
-Ala
-Lys
-Ala
(1.3 ( 0.05) ꢀ 10
(7.2 ( 0.2) ꢀ 10
2.2 ( 0.05
L-Ala-
L
D
-Ala-L
2
pK units apart, and therefore, only an average of the two pK
a a
a
From fits of the data to eq 2. For those entries without values of kcat and
, saturation was not achieved at concentrations up to 5 mM.
values of 7.6 ( 0.4 can be determined (41).
K
m
Site-Directed Mutagenesis. Site-directed mutagenesis was
utilized to identify residues important for enzymatic activity.
Several conserved residues in the active site, including Asp-22,
His-150, Arg-223, and Asp-320, were altered, and the kinetic
constants for the purified mutants are listed in Table 3. When
Asp-22 from the HxD motif at the end of β-strand 1 was changed
to histidine, no activity was detected even though the purified
protein contained 2 equiv of zinc. His-150 was mutated to
asparagine and alanine; both of these mutants exhibited a
significant decrease in kcat (200- and 20-fold, respectively), and
the value of Km more than doubled. The conserved Arg-223 was
changed to lysine and methionine. The lysine mutant was reduced
in activity by 3 orders of magnitude relative to the wild-type
enzyme. No hydrolytic activity was detected with the methionine

Article
Biochemistry, Vol. 49, No. 3, 2010 617
F
IGURE 5: Ribbon diagram of the structure of Sco3058. The two
subunits are colored green and blue. The active site metals are colored
gray.
˚
determined to 1.3, 1.7, and 1.4 A, respectively. A ribbon
representation of the dimeric (β/R) barrel structure is presented
8
in Figure 5. The zinc in the R-site is coordinated to His-20 and
Asp-22 from the HxD motif that originates from the end of
β-strand 1. The zinc in the β-site is ligated to the two conserved
histidines (His-191 and His-212) from the ends of β-strands 5 and
6
, respectively. The two metals are bridged by Glu-123 from
β-strand 3 and a hydroxide/water molecule. The next water
˚
molecule nearest to either of the two metal ions is 2.4 A from
M . The conserved aspartate (Asp-320) from the end of β-strand
β
˚
8
is 3.6 A from M and thus does not coordinate this metal.
R
F
using L-Arg-L-Asp as the substrate. (A) Effect of pH on kcat, (B)
IGURE 4: pH-rate profiles for the zinc-bound forms of Sco3058
However, this residue interacts with the bridging solvent mole-
cule at a hydrogen bonding distance of 3.0 A. The binuclear metal
˚
m
Effect of pH on kcat/K . Additional details can be found in the text.
center is shown in Figure 6A.
The active site of Sco3058 in the presence of glycerol and
citrate is shown in Figures 6B and 7A. In the glycerol- and citrate-
bound structure, the R-metal is tetrahedral in coordination while
the β-metal exhibits a distorted trigonal bipyramidal geometry.
In addition to the four ligands described for the native structure,
the β-metal is additionally coordinated to the C3 carboxylate of
citrate. This carboxylate is also ion-paired with Arg-223 and
a
Table 3: Kinetic Parameters for Mutants of Sco3058
-1
-1 -1
enzyme
k
cat (s
)
K
m
(mM)
k
cat/K
m
(M
s
)
4
wild type
D22H
H150N
H150A
R223K
R223M
D320N
D320A
a
132 ( 2
<0.03
1.8 ( 0.1
(7.2 ( 0.4) ꢀ 10
2
3
0.7 ( 0.01
7.2 ( 0.4
0.4 ( 0.04
<0.03
4.9 ( 2.1
4.3 ( 0.5
8.8 ( 1.4
(1.4 ( 0.4) ꢀ 10
(1.7 ( 0.2) ꢀ 10
46 ( 9
˚
occupies the same position as the water molecule, which is 2.4 A
from M in the native structure. The carboxylate of citrate from
β
<0.03
the pro-S arm is ion-paired to His-150, while the carboxylate
from the pro-R arm is interacting with Thr-324. The hydroxyl
<0.03
˚
group attached to C3 of citrate is 2.8 A from Asp-320 and is also
These constants were obtained by a fit of the data to eq 2.
hydrogen bonded with the C2 hydroxyl group of glycerol. In
addition, the C1 hydroxyl group of glycerol is within hydrogen
mutant. The conserved Asp-320 from β-strand 8 was modified to
asparagine and alanine, and no activity was observed with either
mutant.
Solvent Isotope and Viscosity Effects. Solvent isotope
effects were utilized to determine the effect of proton transfers
during substrate turnover. The kinetic parameters were obtained
˚
˚
bonding distance (3.0 A) of Asp-22 and 2.4 A from the R-metal.
The F - F electron density map that illustrates the orienta-
o
c
tion of the pseudodipeptide inhibitor 1 in the active site of
Sco3058 is presented in Figure 8. The two chiral centers that
mimic the C stereocenters of the N- and C-terminal ends of a
R
for the wild-type enzyme in H O and D O with
L
D O
2
-Arg-
L
-Asp
kcat and
kcat/K ) are 1.30 ( 0.02 and 1.10 ( 0.05, respectively (data
dipeptide are both in the R-configuration. This would correspond
to the dipeptide L-Ala-D-Asp. In this complex, the β-zinc is
2
2
as the substrate. The solvent isotope effects for
2
D O
(
coordinated by five ligands in a distorted trigonal bipyramidal
arrangement. One of these interactions originates from the
phosphinate group, and another is from the carboxylate that
would correspond to the C-terminal carboxylate of a dipep-
tide substrate. The R-metal is in a distorted square pyrimidal
m
not shown). Alterations in solvent viscosity using sucrose were
utilized to probe the degree of rate limitation for the binding and
dissociation of products and substrates on the kinetic constants
o
η
of Sco3058 (42, 43). The plot of k / k versus the relative
cat
cat
˚
solvent viscosity for the wild-type enzyme exhibits a slope of
arrangement, and the R-amino group of the inhibitor is 2.3 A
0
0
.05 ( 0.03, and the plot for the effect of viscosity on kcat/Km is
.50 ( 0.05.
from this metal ion. The C-terminal carboxylate of the inhi-
bitor is ion-paired with the guanidino group of Arg-223. His-150
makes polar interactions with the phosphinate moiety at
Structure of Sco3058. The crystal structures of Sco3058 and
˚
two additional complexes with either glycerol and citrate or the
pseudodipeptide Ala-Asp inhibitor (1) in the active site were
a distance of 2.7 A. The side chain carboxylate of the inhibitor
˚
interacts with Thr-324 at a distance of 2.7 A. The orientation

6
18 Biochemistry, Vol. 49, No. 3, 2010
Cummings et al.
F
IGURE 7: Interactions of citrate and the pseudodipeptide inhibitor (1)
bound in the active site of Sco3058. (A) Polar contacts of citrate (salmon
carbons) bound in the active site of Sco3058 (gray carbons). (B) Polar
contacts between the inhibitor (1) (gold carbons and orange phos-
phorus) and Sco3058 (gray carbons). Polar contacts are indicated with
dashed lines with distances given in angstroms. The zincs and bound
water molecule are represented as green and red spheres, respectively.
F
IGURE 6: Binuclear metal center of Sco3058. (A) Contacts between
the zinc ions (green spheres) and water (red spheres) or active site
residuesare indicated withdashedlines. (B) Structureofthe binuclear
metal center with citrate and glycerol in the active site. (C) Structure
of the binuclear metal center with the phosphinate pseudodipeptide
(
1) bound in the active site.
of inhibitor 1 in the active site of Sco3058 is presented in
Figures 6C and 7B.
Docking of Dipeptides. The docking results were postpro-
cessed to exclude all poses in which the high-energy intermediate
˚
portion of the molecule was more than 4 A from one of the two
metal ions. The top 100 of the remaining poses were inspected
visually to reject molecules with intramolecular clashes and
geometries incompatible with the reaction mechanism. This left
F
IGURE 8: Representative electron density for the L-Ala-D-Asp pseu-
dodipeptide (1) bound in active site of Sco3058. The electron density
was obtained from the F - F omit map contoured at 4σ. The ligand
o
c
63 molecules for further consideration. Among these, there is no
was omitted from the model, and the remainder of the unit cell was
subjected to a cycle of simulated annealing with CNS at 4000 ꢀC. The
details of the interactions between the pseudodipeptide and the active
site are described in the text.
clear preference for a specific residue at the N-terminus; in fact,
most of the 63 molecules are capped (i.e., N-acetylated, N-
formylated, etc.). At the C-terminus, 41% of the amino acids
are in an
L-configuration, and 59% are
D
(no glycine residues
types in more detail, 11% are
L-glutamate, 29% are
D-glutamate,
were observed at this position). Upon examination of the residue
14% are -aspartate, 21% are
L
D-aspartate, and 3% each are

Article
Biochemistry, Vol. 49, No. 3, 2010 619
FIGURE 10: Overlay of the active sites of the L-Ala-D-Asp pseudodi-
peptide-bound Sco3058 (dark gray carbons, pink Zn, and gold
inhibitor carbons) and the docked L-Ala-D-Ala Sco3058 complex
(
3
light gray carbons, teal Zn, and green substrate carbons).
tetrahedral HEI portion of the dipeptide, and the C-terminal
carboxy moieties both interact with Arg-223. The two backbones
are lined up to a high degree, with the exception of the N-terminal
amino groups, which are rotated by 90ꢀ with respect to each
other.
DISCUSSION
Substrate Specificity. The substrate specificity of Sco3058
was determined using an array of dipeptide libraries. It is evident
from the data presented in Table 2 that this enzyme has the
ability to hydrolyze a rather broad range of dipeptides having
amino acids of either the
N-terminus, amino acids of the
preferred over those of the -configuration. For example, the
values of kcat/Km for -Ala- -Ala and -Ala- -Ala are 2-3
orders of magnitude greater than those of their -Ala- -Ala
and -Ala- -Ala counterparts. Conversely, at the C-terminus,
amino acids of the -configuration are preferred over those of the
-configuration. Thus, the values of kcat/K for -Arg- -Asp and
-Ala are ∼1 order of magnitude greater than those of
-Asp and -Ala- -Ala, respectively. At the N-terminus,
D
- or
L
-configuration. However, at the
L-configuration are greatly
D
F
(
IGURE 9: Cartoon of the computed binding modes of L-Ala-D-Ala
A) and D-Ala-L-Ala (B). In both panels, the protein is depicted with
L
D
L
L
D
D
gray carbons and the dipeptide with green carbons. The Zn ions,
nitrogens, and oxygens are colored purple (spheres), blue, and red,
˚
respectively. All contacts between 2.1 and 3.7 A are indicated as
dashed lines.
D
L
D
L
L
L
L
D
m
-Ala-
-Arg-
D
L-asparagine,
D-asparagine,
D-serine, and
L-glutamine. These
L
L
L
computed specificities, at the C-terminus, may be compared with
those determined experimentally (Figure 3).
there is a preference for arginine, methionine, glutamine, and
leucine, whereas at the C-terminus, there is a clear preference for
glutamate and aspartate. The very best substrates have kcat/Km
The computationally determined poses of the eight diastereo-
mers of the Ala-Ala dipeptide fall into two categories based on
the stereochemistry at the HEI center. These would include
molecules with an HEI center that would originate from an
attack of the water molecule on the re face of the amide bond and
those that would originate from an attack on the si face. In all
5
-1 -1
values that exceed 10 M
s . N-Acyl derivatives of D-amino
acids are hydrolyzed approximately 5 orders of magnitude more
slowly than are the corresponding dipeptide substrates, and thus,
this enzyme appears to be a true dipeptidase rather than a
carboxypeptidase.
cases, except for
D
-Ala-
D
-Ala, the intermediates corresponding to
Structural Determinants of Substrate Specificity. The
three-dimensional structure of Sco3058 was determined in the
absence of ligands and also in the presence of citrate and a
pseudodipeptide analogue bound in the active site. The analogue
of the proposed tetrahedral intermediate was synthesized without
regard for the stereochemistry at the two chiral centers. The
structure of the inhibitor bound in the active site is a mimic of the
an attack from the si face docked in unfavorable binding modes,
either forming intramolecular clashes or pointing the negatively
charged oxygen (corresponding to the attacking hydroxyl) away
from the catalytic zinc ions. The C-terminal carboxy moiety is
always in contact with Arg-223, whereas the N-terminal amino
group can form polar interactions with either the backbone
carbonyl of Gly-323, the side chain of Tyr-68, or Asp-22. The
L-Ala-D-Asp dipeptide. As a structural mimic of the tetrahedral
proposed orientations of
active site of Sco3058 are presented in panels A and B of Figure 9,
respectively.
L
-Ala-
D
-Ala and
D
-Ala-
L
-Ala in the
addition complex, this analogue is compromised since the amide
nitrogen has been substituted with a methylene group to maintain
the hydrolytic stability of the inhibitor. Nevertheless, the struc-
ture of this complex indicates that the free R-amino group at the
The overlay of the docking-derived pose of L-Ala-D-Ala with
the crystallographic pose of the pseudodipeptide 1 (Figure 10)
shows very good agreement in the parts that are similar. The
phosphinate moiety of 1 overlaps almost perfectly with the
N-terminus of the pseudodipeptide interacts with M and the side
R
chain carboxylate of Asp-22. Conversely, the free R-carboxylate
at the C-terminus interacts with M and the side chain guanidino
β

6
20 Biochemistry, Vol. 49, No. 3, 2010
Cummings et al.
Scheme 3
group of Arg-223. The side chain carboxylate of the inhibitor
interacts with Thr-324 and is in the proximity of the ε-amino
group of Lys-247. The phosphinate group of the inhibitor
A similar rationale can be employed at the N-terminus. The
most likely interactions formed by a positively charged N-
terminal amino group of a dipeptide substrate are with the
backbone carbonyl of Gly-323 and the side chain of Asp-22.
Under these circumstances, the side chain of an Ala in the
interacts with M and the side chain of His-150.
β
A similar set of interactions is apparent in the binding of citrate
in the active site of Sco3058. In many ways, the structure of citrate
mimics the pseudodipeptide inhibitor. Thus, the carboxylate at
D-configuration would come too close to itself, forming an
intramolecular clash (Figure 9B). However, the structural pre-
ference at the N-terminus is more subtle, and it is likely not the
only factor contributing to the observed differences in the turn-
C3 interacts with M and Arg-223 in much the same way as the
β
free R-carboxylate of the inhibitor. The carboxylate from the
pro-R arm of citrate interacts with Thr-324 in a manner similar to
how the side chain carboxylate of the inhibitor interacts with this
residue. The carboxylate from the pro-S arm of citrate interacts
with His-150 in a fashion that duplicates the interactions with the
phosphinate moiety with this same residue. The closest residue
to the hydroxyl group of citrate is the side chain carboxylate of
Asp-320. However, there are certain interactions that are missing,
and this may explain why the pseudodipeptide inhibitor binds
with a dissociation constant that is ∼4 orders of magnitude
tighter than that of citrate.
over rates. This is also apparent from the overlay of the L-Ala-D-
Ala dipeptide with the pseudodipeptide (1) in the X-ray structure
where the most similar parts overlay well but the N-terminus
shows the highest divergence.
Mechanism of Action. One can utilize the structure of
Sco3058 in the presence of the pseudodipeptide inhibitor to
propose a mode of binding for typical dipeptide substrates and
a working model for the chemical mechanism of substrate
activation and hydrolysis. In this model, the N-terminal R-amino
group of dipeptide substrates interacts with Asp-22 and M while
R
Docking of Dipeptides. The docking of the full spectrum of
the C-terminal R-carboxylate group interacts with M and Arg-
β
dipeptides reproduced the observed preference for
D-enantiomers
223. The carbonyl group of the amide bond about to be
hydrolyzed is polarized by Mβ with assistance from His-150.
Since the mutation of His-150 to an alanine reduces kcat by a
factor of only 20, the presence of this histidine in the active site is
not critical for substrate turnover. His-150 from Sco3058 aligns
with His-152 in the human renal dipeptidase. The hydro-
lytic water is the bridging water between the two metals (see
of the C-terminal amino acids found experimentally. Moreover,
the statistics of the top 100 poses show a dominance of acidic
residues at the C-terminus, consistent with experiment. There
were also discrepancies. Experimentally, aspartate is preferred
over glutamate, whereas the reverse is true in the docking; in
neither case were the differences large. The other residues that
appear more than once in the list of 63 molecules (
L
-asparagine,
Figure 6A). The binding of the R-carboxylate group to M occurs
β
D
-asparagine, -serine, and -glutamine) are “true positives” in
D
L
with the displacement of the nonbridging water molecule to this
metal ion. The phosphinate moiety of the inhibitor mimics the
addition of hydroxide to the carbonyl carbon. After attack by the
bridging hydroxide on the carbonyl carbon, a proton is trans-
ferred to the side chain carboxylate of Asp-320 and another
proton is transferred to the leaving group amine. Asp-320 is
hydrogen bonded to the bridging hydroxide and is the closest
residue to the methylene group of the pseudodipeptide inhibitor
and the hydroxyl at C3 of citrate. This residue is the only one in
the active site that is positioned to facilitate the transfer of a
proton from the bridging hydroxide to the leaving group amine.
This mechanism is illustrated in Scheme 3, and it is very similar to
those previously proposed for other members of the amidohy-
drolase superfamily (16-18).
the sense that these residues are also observed with significant
turnover rates in the experiments. The docking misses some of the
other residues at the C-terminus that show turnover in vitro, such
as leucine and methionine.
The interactions in the docking poses also help explain the
differences in the enzyme specificities, as measured by kcat/K , for
m
the four Ala-Ala diastereomers. The docked poses show that if
the C-terminal Ala is the
L-enantiomer, its methyl group points
toward Lys-247. Conversely, when this Ala is in the
D-config-
uration, the methyl is positioned more favorably in a hydro-
phobic, and thus more favorable, pocket formed by Val-245 and
Phe-248 (Figure 9A). This is consistent with the observed
preference for a
and is also borne out in the docking scores of both the
Ala-Ala pairs. The variants with the C-terminal amino
acid in the -configuration receive a more favorable score.
D-configuration of the C-terminal amino acid
L-L/L-D
and
Rate-Limiting Steps. The reaction mechanism was also
D-L/D-D
addressed with the measurement of pH-rate profiles and the
D
effects of solvent viscosity and D O. The changes in the kinetic
2

Article
Biochemistry, Vol. 49, No. 3, 2010 621
constants, kcat and kcat/K , show a bell-shaped profile. The loss
those that are in the vicinity of the active site (e.g., Tyr-68,
Val-245, and Phe-248) are not conserved among hRDP and the
cluster of renal dipeptidase-like sequences. Lys-247 is actually
quite rare within cog2355E. All of these residues are at the ends of
m
of activity at low pH is consistent with the protonation of the
bridging hydroxide or the carboxylate of Asp-320 that is hydro-
gen bonded to the bridging hydroxide. The loss of activity at high
pH is more difficult to assign, but it is consistent with the
deprotonation of the R-amino group of the substrate. When
or immediately follow the β-strands of the (β/R) barrel, which is
8
a common location for substrate-contacting residues of AHS
the reactions were conducted in D O, there was a rather modest
enzymes (16, 21, 22).
2
reduction in the value of the kinetic constants. Thus, it appears
that proton transfers are not occurring during the rate-limiting
steps. There is, however, an effect of solvent viscosity on the value
Given the lack of conservation of putative substrate recogni-
tion residues within the subset of human renal dipeptidase and
its microbial homologues, there is an opportunity to identify
enzymes with varying substrate specificities and define the struc-
ture-function relationship of these enzymes. Therefore, the sub-
strate specificity profiles and the three-dimensional structures of
several microbial renal dipeptidase-like proteins are currently under
investigation using the same methods applied to Sco3058.
of kcat/K . A change in solvent viscosity can modulate and
m
diminish those steps that involve the binding and/or dissociation
of substrates and products from the enzyme active site but not the
chemical step (43). The value of kcat/Km is affected by the rate
constants of all processes up through the first irreversible
step (44). Since the value of kcat/Km is affected by solvent
viscosity, it appears that the rate of dissociation of substrate
from the EA complex is similar in magnitude to the net rate
REFERENCES
1
. Adachi, H., Tawaragi, Y., Inuzuka, C., Kubota, I., Tsujimoto, M.,
Nishihara, T., and Nakazato, H. (1990) Primary Structure of Human
Microsomal Dipeptidase Deduced from Molecular Cloning. J. Biol.
Chem. 265, 3992–3995.
constant for the conversion of EA to the enzyme product
3
complex, EP (43). The relatively small effect of viscosity on k
cat
suggests that product release is not rate-limiting.
2
. Adachi, H., Kubota, I., Okamura, N., Iwata, H., Tsujimoto, M.,
Nakazato, H., Nishihara, T., and Noguchi, T. (1989) Purification and
Characterization of Human Microsomal Dipeptidase. J. Biochem.
105, 957–961.
Genome Context for Sco3058 and cog2355E. The genome
context of Sco3058 would not have been useful in predicting the
general reaction of this protein or the substrate specificity. The
neighboring protein in the genome, Sco3057, is ∼40% identical in
amino acid sequence to Sco3058, but Sco3057 does not have a
single metal-binding residue in common with Sco3058. The
Sco3058 gene is also adjacent to PurE (Sco3059) and PurK
3
4
5
. Kozak, E. M., and Tate, S. S. (1982) Glutathione-degrading Enzymes
of Microvillus Membranes. J. Biol. Chem. 257, 6322–6327.
. McIntyre, T., and Curthoys, N. P. (1982) Renal Catabolism of
Glutathione. J. Biol. Chem. 257, 11915–11921.
. Satoh, S., Keida, Y., Konta, Y., Maeda, M., Matsumoto, Y., Niwa,
M., and Kohsaka, M. (1993) Purification and molecular cloning of
mouse renal dipeptidase. Biochim. Biophys. Acta 1163, 234–242.
. Watanabe, T., Kera, Y., Matsumoto, T., and Yamada, R.-H. (1996)
(
Sco3060) which encode proteins annotated as phosphoribosyl-
aminoimidazole carboxylase catalytic and ATPase subunits,
respectively. It is conceivable that Sco3058 could either hydrolyze
a molecule such as 4-[(N-succinylamino)carbonyl]-5-aminoimida-
zole ribonucleotide (SAICAR) to liberate aspartate or supply
aspartate for the SAICAR synthetase reaction through the
6
Purification and kinetic properties of a D-amino-acid peptide hydro-
lyzing enzyme from pig kidney cortex and its tentative identification
with renal membrane dipeptidase. Biochim. Biophys. Acta 1298, 109–
1
18.
7. Habib, G. M., Shi, Z.-Z., Cuevas, A. A., Guo, Q., Matzuk, M. M., and
Lieberman, M. W. (1998) Leukotriene D and cystinyl-bis-glycine
cleavage of an Xaa-L-Asp dipeptide (45). However, the immediate
4
metabolism in membrane-bound dipeptidase-deficient mice. Proc.
Natl. Acad. Sci. U.S.A. 95, 4859–4863.
proximity of PurE and PurK to genes that encode renal dipepti-
dase-like proteins is not a common feature of the genome contexts
of these proteins. To the best of our knowledge, the protein en-
coded by the SirJ gene from Leptosphaeria maculans (gi|46403052)
is the only renal dipeptidase-like protein thought to be involved in
a biosynthetic pathway (46). It is unclear whether there is any other
specific function of the microbial renal dipeptidase-like proteins
beyond their ability to serve as true dipeptidases.
8
9
. Campbell, B. J., Forrester, L. J., Zahler, W. L., and Burks, M. (1984)
β-Lactamase Activity of Purified and Partially Characterized Human
Renal Dipeptidase. J. Biol. Chem. 259, 14586–14590.
. Park, S. W., We, J. S., Kim, G. W., Choi, S. H., and Park, H. S. (2002)
Stability of New Carbapenem DA-1131 to Renal Dipeptidase
(Dehydropeptidase I). Antimicrob. Agents Chemother. 46, 575–577.
1
1
0. Campbell, B. J., Shih, Y. D., Forrester, L. J., and Zahler, W. L. (1988)
Specificity and inhibition studies of human renal dipeptidase. Bio-
chim. Biophys. Acta 956, 110–118.
1. Baek, D. H., Song, J. J., Kwon, S.-J., Park, C., Jung, C.-M., and Sung,
M. H. (2004) Characteristics of a New Enantioselective Thermostable
Dipeptidase from Brevibacillus borstelensis BCS-1 and Its Application
Conservation of Residues within cog2355E. There are
∼400 amino acid sequences from the completely sequenced
bacterial genomes that are categorized as part of cog2355E.
Approximately two-thirds of these sequences can be classified as
renal dipeptidase-like because of the conservation of the complete
set of metal ligands for the binuclear active site (13). Arginine 223
recognizes the R-carboxylate of the dipeptide, and the residue
equivalent to Arg-223 is conserved in more than 95% of the
sequences found in cog2355E. Within the renal dipeptidase
subset of cog2355E, the residue equivalent to His-150 in
Sco3058 is semiconserved with approximately 30% of the
sequences having a histidine and 70% a tryptophan at this
position. In the case of the renal dipeptidase-like protein from
Rhodobacter sphaeroides (Rsp0802; PDB entry 3FDG; gi|
to Synthesis of a
Microbiol. 70, 1570–1575.
12. Adachi, H., and Tsujimoto, M. (1995) Cloning and Expression of
Dipeptidase from Acinetobacter calcoaceticus ATCC 23055. J. Bio-
chem. 118, 555–561.
3. Nitanai, Y., Satow, Y., Adachi, H., and Tsujimoto, M. (2002) Crystal
Structure of Human Renal Dipeptidase Involved in β-Lactam Hydro-
lysis. J. Mol. Biol. 321, 177–184.
4. Keynan, S., Hooper, N. M., and Turner, A. J. (1997) Identification by
site-directed mutagenesis of three essential histidine residues in mem-
brane dipeptidase, a novel mammalian zinc peptidase. Biochem. J.
326, 47–51.
D-Amino-Acid-Containing Dipeptide. Appl. Environ.
1
1
1
5. Seibert, C. M., and Raushel, F. M. (2005) Structural and Catalytic
Diversity within the Amidohydrolase Superfamily. Biochemistry 44,
6383–6391.
2
18766909), this tryptophan residue is structurally equivalent
16. Thoden, J. B., Phillips, G. N., Jr., Neal, T. M., Raushel, F. M., and
Holden, H. M. (2001) Molecular Structure of Dihydroorotase: A
Paradigm for Catalysis through the Use of a Binuclear Metal Center.
Biochemistry 40, 6989–6997.
to His-150 from Sco3058 with the nitrogen of the tryptophan side
chain being in the same position as the N atom of the histidine.
The other residues that contact or are close to citrate or the
phosphinate inhibitor (i.e., Thr-324, Asn-151, and Lys-247) and
ε
17. Marti-Arbona, R., Fresquet, V., Thoden, J. B., Davis, M. L., Holden,
H. M., and Raushel, F. M. (2005) Mechanism of the Reaction

6
22 Biochemistry, Vol. 49, No. 3, 2010
Cummings et al.
Catalyzed by Isoaspartyl Dipeptidase from Escherichia coli. Biochem-
istry 44, 7115–7124.
Macromolecular Structure Determination. Acta Crystallogr. D54,
905–921.
1
1
8. Marti-Arbona, R., Thoden, J. B., Holden, H. M., and Raushel, F. M.
2005) Functional Significance of Glu-77 and Tyr-137 within the
32. Brunger, A. T. (2007) Version 1.2 of the Crystallography and NMR
System. Nat. Protoc. 2, 2728–2733.
33. McRee, D. E. (1999) XtalView/Xfit: A Versatile Program for Manip-
ulating Atomic Coordinates and Electron Density. J. Struct. Biol. 125,
156–165.
34. McCoy, A. J., Grosse-Kunstleve, R. W., Storoni, L. C., and Read, R.
J. (2005) Likelihood-enhanced fast translation functions. Acta Crys-
tallogr. D61, 458–464.
35. Perrakis, A., Morris, R., and Lamzin, V. S. (1999) Automated protein
model building combined with iterative structure refinement. Nat.
Struct. Biol. 6, 458–463.
(
active site of isoaspartyl dipeptidase. Bioorg. Chem. 33, 448–458.
9. Bentley, S. D., Chater, K. F., Cerdeno-Tarraga, A. M., Challis, G. L.,
Thomson, N. R., James, K. D., Harris, D. E., Quail, M. A., Kieser, H.,
Harper, D., Bateman, A., Brown, S., Chandra, G., Chen, C. W.,
Collins, M., Cronin, A., Fraser, A., Goble, A., Hidalgo, J., Hornsby,
T., Howarth, S., Huang, C. H., Kieser, T., Larke, L., Murphy, L.,
Oliver, K., O’Neil, S., Rabbinowitsch, E., Rajandream, M. A.,
Rutherford, K., Rutter, S., Seeger, K., Saunders, D., Sharp, S.,
Squares, R., Squares, S., Taylor, K., Warren, T., Wietzorrek, A.,
Woodward, J., Barrell, B. G., Parkhill, J., and Hopwood, D. A. (2002)
Complete genome sequence of the model actinomycete Streptomyces
36. Jones, T. A. (1985) Interactive computer graphics: FRODO. Methods
Enzymol. 115, 157–171.
coelicolor A3
0. Hall, R. S., Xiang, D. F., and Raushel, F. M. (2007) N-Acetyl-D-
glucosamine-6-phosphate Deacetylase: Substrate Activation via a
Single Divalent Metal Ion. Biochemistry 46, 7942–7952.
2
. Nature 417, 141–147.
37. Hermann, J. C., Ghanem, E., Li, Y., Raushel, F. M., Irwin, J. J., and
Shoichet, B. K. (2006) Predicting Substrates by Docking High-Energy
Intermediates to Enzyme Structures. J. Am. Chem. Soc. 128, 15882–
15891.
2
2
1. Cummings, J. A., Fedorov, A. A., Xu, C., Brown, S., Fedorov, E.,
Babbitt, P. C., Almo, S. C., and Raushel, F. M. (2009) Annotating
38. Nowlan, C., Li, Y., Hermann, J. C., Evans, T., Carpenter, J.,
Ghanem, E., Shoichet, B. K., and Raushel, F. M. (2006) Resolution
of Chiral Phosphate, Phosphonate, and Phosphinate Esters by an
Enantioselective Enzyme Library. J. Am. Chem. Soc. 128, 15892–
15902.
Enzymes of Uncertain Function: The Deacylation of
by Members of the Amidohydrolase Superfamily. Biochemistry 48,
469–6481.
D-Amino Acids
6
2
2. Xiang, D. F., Patskovsky, Y., Xu, C., Meyer, A. J., Sauder, J. M.,
Burley, S. K., Almo, S. C., and Raushel, F. M. (2009) Functional
Identification of Incorrectly Annotated Prolidases from the Amido-
hydrolase Superfamily of Enzymes. Biochemistry 48, 3730–3742.
3. Buchardt, J., and Meldal, M. (2000) Novel methodology for the solid-
phase synthesis of phosphinic peptides. J. Chem. Soc., Perkin Trans. 1,
39. Hermann, J. C., Marti-Arbona, R., Fedorov, A. A., Fedorov, E.,
Fedorov, E., Almo, S. C., Shoichet, B. K., and Raushel, F. M. (2007)
Structure-based activity prediction for an enzyme of unknown func-
tion. Nature 448, 775–779.
40. Xiang, D. F., Kolb, P., Fedorov, A. A., Meier, M. M., Fedorov, L. V.,
Nguyen, T. T., Sterner, R., Almo, S. C., Shoichet, B. K., and Raushel,
F. M. (2009) Functional Annotation and Three-Dimensional Struc-
ture of Dr0930 from Deinococcus radiodurans, a Close Relative of
Phosphotriesterase in the Amidohydrolase Superfamily. Biochemistry
48, 2237–2247.
2
2
3
306–3310.
4. Georgiadis, D., Matziari, M., and Yiotakis, A. (2001) A highly
efficient method for the preparation of phosphinic pseudodipeptidic
blocks suitably protected for solid-phase peptide synthesis. Tetra-
hedron 57, 3471–3478.
5. Baylis, E. K., Campbell, C. D., and Dingwall, J. G. (1984) 1-
Aminoalkylphosphonous Acids. Part 1. Isosteres of the Protein
Amino Acids. J. Chem. Soc., Perkin Trans. 1, 2845–2853.
41. Cleland, W. W., and Cook, P. F. (2007) pH Dependence of kinetic
parameters and isotope effects. In Enzyme Kinetics and Mechanism
(Rogers, R. L., Ed.) pp 326-366, Taylor & Francis Group, New
York.
2
2
6. Liboska, R., Picha, J., Hanclova, I., Budesinsky, M., Sanda, M., and
Jiracek, J. (2008) Synthesis of methionine- and norleucine-derived
phosphinopeptides. Tetrahedron Lett. 49, 5629–5631.
42. Brouwer, A. C., and Kirsch, J. F. (1982) Investigation of Diffusion-
Limited Rates of Chymotrypsin Reactions by Viscosity Variation.
Biochemistry 21, 1302–1307.
2
7. Doi, E., Shibata, D., and Matoba, T. (1981) Modified Colorimetric
Ninhydrin Methods for Peptidase Assay. Anal. Biochem. 118, 173–184.
8. Otwinowski, Z., and Minor, W. (1997) Processing of X-ray Diffrac-
tion Data Collected in Oscillation Mode. Methods Enzymol. 276, 307–
43. Cleland, W. W., and Cook, P. F. (2007) Isotopic Probes of Kinetic
Mechanism. In Enzyme Kinetics and Mechanism (Rogers, R. L., Ed.)
pp 249-251, Taylor & Francis Group, New York.
44. Cleland, W. W. (1975) What Limits the Rate of an Enzyme-Catalyzed
Reaction? Acc. Chem. Res. 8, 145–151.
2
3
26.
2
3
3
9. Baily, S. (1994) The CCP4 Suite: Programs for Protein Crystal-
lography. Acta Crystallogr. D50, 760–763.
0. Navaza, J. (1994) AMoRe: An Automated Package for Molecular
Replacement. Acta Crytallogr. A50, 157–163.
1. Br u€ nger, 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., Read, R. J., Rice, L. M., Simonson, T., and Warren, G.
L. (1998) Crystallography & NMR system: A New Software Suite for
45. Firestine, S. M., Poon, S.-W., Mueller, E. J., Stubbe, J., and Davisson,
V. J. (1994) Reactions Catalyzed by 5-Aminoimidazole Ribonucleo-
tide Carboxylase from Escherichia coli and Gallus gallus: A Case for
Divergent Catalytic Mechanisms? Biochemistry 33, 11927–11934.
46. Gardiner, D. M., Cozijnsen, A. J., Wilson, L. M., Soledade, M.,
Pedras, C., and Howlett, B. J. (2004) The sirodesmin biosynthetic gene
cluster of the plant pathogenic fungus Leptosphaeria maculans. Mol.
Microbiol. 53, 1307–1318.