Characterization of two putative prolinases (PepR1 and PepR2) from Lactobacillus plantarum WCFS1: Occurrence of two isozymes with structural similarity and different catalytic properties

Article abstract of DOI:10.1016/j.bbapap.2014.11.003

Two putative prolinases (PepR1 and PepR2) of Lactobacillus plantarum WCSF1 share 48.5% amino acid sequence identity (55.5% at the DNA level); however, PepR1 exhibits over 80% identity at the protein level with other lactobacilli prolinases while PepR2 exhibits only 51% or less identity. In this study, the putative genes were overexpressed in Escherichia coli, purified to gel electrophoretic homogeneity, and then characterized. Purified PepR1 and PepR2 hydrolysed Pro-Xaa dipeptide substrates at similar rates, proving their nature as prolinases. Structural analyses using circular dichroism, dynamic light scattering, gel filtration, and molecular modelling revealed that the two prolinases have similar structural characteristics: high β-sheet content, homotetrameric structure, and similar folding to the PepI/PepL/PepR peptidase family. However, kinetic and thermodynamic analyses of PepR1 and PepR2 indicated differences in many aspects: optimum temperatures (25 and 30 °C, respectively), optimum pH (pH 7.5 and 8.0, respectively), substrate specificities (high stringency of PepR2), kinetic parameters, and thermal stability (29 and 48 °C, respectively). Also, these prolinases behaved differently towards inhibitor treatments, suggesting structural and/or functional differences in their active sites. Differences in the two prolinases would contribute to a diversity of catalytic activities, so that they work together cooperatively and complementarily to hydrolyse proline-containing peptides with broader specificity, working pH, working temperature, and higher efficiency, thus allowing adaptation to a wider range of environments.

Full text of DOI:10.1016/j.bbapap.2014.11.003

Biochimica et Biophysica Acta 1854 (2015) 91–100
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbapap
Characterization of two putative prolinases (PepR1 and PepR2) from
Lactobacillus plantarum WCFS1: Occurrence of two isozymes with
structural similarity and different catalytic properties
Yanyu Huang, Takuji Tanaka ⁎
Department of Food and Bioproduct Sciences, College of Agriculture and Bioresources, University of Saskatchewan, 51 Campus Dr., Saskatoon, SK S7N 5A8, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 June 2014
Received in revised form 5 November 2014
Accepted 10 November 2014
Available online 15 November 2014
Two putative prolinases (PepR1 and PepR2) of Lactobacillus plantarum WCSF1 share 48.5% amino acid sequence
identity (55.5% at the DNA level); however, PepR1 exhibits over 80% identity at the protein level with other
lactobacilli prolinases while PepR2 exhibits only 51% or less identity. In this study, the putative genes were
overexpressed in Escherichia coli, purified to gel electrophoretic homogeneity, and then characterized. Purified
PepR1 and PepR2 hydrolysed Pro-Xaa dipeptide substrates at similar rates, proving their nature as prolinases.
Structural analyses using circular dichroism, dynamic light scattering, gel filtration, and molecular modelling
revealed that the two prolinases have similar structural characteristics: high β-sheet content, homotetrameric
structure, and similar folding to the PepI/PepL/PepR peptidase family. However, kinetic and thermodynamic
analyses of PepR1 and PepR2 indicated differences in many aspects: optimum temperatures (25 and 30 °C,
respectively), optimum pH (pH 7.5 and 8.0, respectively), substrate specificities (high stringency of PepR2),
kinetic parameters, and thermal stability (29 and 48 °C, respectively). Also, these prolinases behaved differently
towards inhibitor treatments, suggesting structural and/or functional differences in their active sites. Differences
in the two prolinases would contribute to a diversity of catalytic activities, so that they work together coopera-
tively and complementarily to hydrolyse proline-containing peptides with broader specificity, working pH,
working temperature, and higher efficiency, thus allowing adaptation to a wider range of environments.
Keywords:
Bitterness
Proline
Prolidase
Molecular cloning
PepI/PepL/PepR family
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
bitterness can be alleviated with a variety of proline-specific peptidases.
Among them, two dipeptidases act in the degradation of the smallest
proline-containing peptides (Pro-Xaa and Xaa-Pro): prolinase hydroly-
ses Pro-Xaa, and prolidase hydrolyses Xaa-Pro. These enzymes are also
known for their potential significance in the clinical treatment of
prolidase deficiency, which is a rare inherited metabolic disorder caused
by low prolidase activity [4]. Although prolidase activity diminishes in
prolidase-deficient patients, prolinase activity is reported to increase
in the plasma of these patients [5] and in prolidase-deficient fibroblasts
[6]. These findings suggest that prolinase has an ability to maintain the
same level of proline in prolidase-deficient fibroblasts as in normal cells,
so prolinase may be useful in the treatment of prolidase deficiency.
Prolinase is found in many bacteria and mammals. Interestingly, the
occurrence of prolinase differs between mammals and microorganisms.
Mammals have at least two forms (isozymes) of prolinase, such as
found in human prostate [7], human skin fibroblasts [8], human eryth-
rocytes [9], human leukocytes [10], and bovine kidney [11,12]. Isozymes
of prolinase have rarely been reported or investigated in microorgan-
isms, with a few exceptions such as Lactobacillus plantarum. The
genome sequence of L. plantarum WCFS1 has two putative prolinase
genes (pepR1 and pepR2) [13]. These genes share 55.5% of DNA identity
and their deduced amino acid sequences share 48.5% identity, indicating
their evolutionary distance. Despite their differences, both prolinase
Proline is a structurally unique amino acid. Its imino ring structure
distinguishes proline from other amino acids in terms of biochemical
activities, structural behaviour, and physiological actions. For example,
the cyclic structure restricts free rotation and contributes to the rigid
structure of proline-rich proteins, such as collagen, and the rigidity
makes this amino acid act as a β-sheet/α-helix breaker [1]. Due to this
unique structure, proline residues within polypeptides act as structural
elements that limit the susceptibility of polypeptide chains to proteoly-
sis, thereby protecting peptides against enzymatic degradation [1]. This
structure also may influence the sensory perception of some fermented
foods. Ishibashi et al. [2,3] found that proline-containing peptides are
generally bitter, and their bitterness can be at the same level as bitter
organic compounds such as caffeine. In proteolysis events, such as in
fermented foods, the less susceptible proline-containing peptides will
accumulate, resulting in increased (and undesirable) bitterness. This
Abbreviations: DLS, dynamic light scattering; MAL, 2,5-pyrroledione; NEM, N-
ethylmaleimide; r-PepR1 & r-PepR2, recombinant Lactobacillus plantarum WCFS1
prolinases
⁎
Corresponding author. Tel.: +1 306 966 1697; fax: +1 306 966 8898.
E-mail address: takuji.tanaka@usask.ca (T. Tanaka).
http://dx.doi.org/10.1016/j.bbapap.2014.11.003
570-9639/© 2014 Elsevier B.V. All rights reserved.
1

92
Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
isozymes conserve high peptide sequence homology with prolinase in
other lactobacilli (PepR). L. plantarum WCFS1 PepR1 exhibits 84%, 84%,
and 83% identities with PepR from Lactobacillus zeae ATCC393 (Genbank
accession No.: BAF85818.1), Lactobacillus rhamnosus ATCC14957
2.3. Purification of recombinant prolinases with His-tag
The recombinant pET32(+)-pepR1/E. coli Rosetta was cultured in LB
media (pH 7.5) at 16 °C for three days with 1 mM IPTG induction at
an optical density at 600 nm of 0.5. Cells were disrupted with
ultrasonication (model 450, Sonifier, Branson Ultrasonics Co., CT, USA;
25 s-burst and 35 s-pause operations for 15 min in an ice bath). The
crude extracts were recovered by centrifugation (12,000 rpm, 30 min,
4 °C) in a Sorvall ss-34 rotor. The crude extract was applied to a pre-
equilibrated Ni-NTA spin column (Qiagen, Mississauga, ON, Canada);
(
Genbank accession No.: BAF85816.1), and L. zeae DSM20178 (Genbank
accession No.: BAF85817.1), respectively; PepR2 exhibits 51% identity
with PepR from the above three lactobacilli.
The presence of two putative prolinases in L. plantarum WCFS1 may
shed some light on the understanding of these proline-specific pepti-
dases and their applications in food processing and clinical treatments.
The present study aimed to clone the prolinase genes pepR1 and
pepR2 from L. plantarum WCFS1, and to characterize the enzyme
features.
2 4
equilibration buffer consisted of 50 mM NaH PO , 300 mM NaCl, and
10 mM imidazole (pH 8.0). After applying the sample, the column was
washed twice with a buffer solution (80 mM imidazole, 50 mM
NaH
PepR1) was then eluted from the column with an elution buffer
50 mM NaH PO , 300 mM NaCl, 500 mM imidazole; pH 8.0). The rH-
PepR1 fractions were dialysed against 50 mM sodium phosphate buffer
pH 8.0). Twenty microliters of the dialysed His-tagged prolinase was
2 4
PO , 300 mM NaCl; pH 8.0). The His-tagged PepR1 protein (rH-
2
. Materials and methods
(
2
4
2
.1. Materials
(
treated with 5 μL of thrombin (0.05 U/μL; New England Biolabs, MA,
USA) at 16 °C overnight. The cleaved rH-PepR1 sample was then applied
on a gel filtration column (Superdex 200 HR 10/300) equilibrated
with 50 mM sodium phosphate buffer (pH 8.0). Cleaved prolinase was
eluted with 50 mM sodium phosphate buffer (pH 8.0) at a flow rate of
Enzymes for genetic engineering were purchased from Fermentas
International (Burlington, ON, Canada) and Invitrogen (Burlington,
ON, Canada). Standard proteins were obtained from Bio-Rad Laborato-
ries (Mississauga, ON, Canada). All chemicals used in this study were
commercially available ACS grade, and were purchased from VWR In-
ternational (Edmonton, AB, Canada) and Fisher Scientific (Ottawa, ON,
Canada). PCR primers were custom-synthesized by Integrated DNA
Technologies (Coralville, IL, USA). Columns for chromatography were
purchased from GE Healthcare Bio-sciences (Mississauga, ON, Canada).
0
.2 mL/min. The eluent was collected and examined on 10% SDS-PAGE
gel to identify the fraction with prolinase. The pure cleaved prolinase
r-PepR1: His-tag removed rH-PepR1) was combined with 50% (v/v)
glycerol for storage in a freezer at −20 °C.
(
2.4. Purification of recombinant prolinase without affinity-tag
2
.2. Construction of recombinant prolinases
E. coli TOP10F′ carrying pKK223-3-pepR2 was cultured in 3 L of LB
The genome DNA of L. plantarum WCFS1 was isolated using a ge-
broth and incubated at 16 °C with vigorous aeration. When the absorp-
tion at 600 nm reached 0.5, 1 mM IPTG was added, the mixture was in-
cubated for 72 h. The harvested cells were disrupted by ultrasonication
in 20 mM Tris–HCl (pH 7.5), and the protein concentration of the crude
extract was adjusted to 10 mg/mL. r-PepR2 was separated using 20%,
40%, and 60% ammonium sulfate saturation. After dialysis, the 60%-
ammonium sulfate fraction was applied to an ion-exchange column
(DEAE-Sephacel; 3 cm ϕ × 15 cm) equilibrated with 20 mM Tris–HCl
(pH 7.5). r-PepR2 was eluted using a linear gradient of NaCl from 0 to
1 M. The r-PepR2 fractions were added with ammonium sulfate to
0.8 M and applied on a hydrophobic interaction column (Fast Flow Phe-
nyl Sepharose, 0.5 cm ϕ × 15 cm) equilibrated with 20 mM Tris–HCl/
1 M ammonium sulfate (pH 7.5). The hydrophobic interaction column
was step-wisely washed with three column volumes of wash buffer so-
lutions of ammonium sulfate (1, 0.75, 0.5, 0.25, and 0 M). The eluent
from the column was fractionated and the fractions were examined
on a 10% SDS-PAGE gel to identify the pure r-PepR2 fractions. The
pure r-PepR2 was concentrated and stored in a 50% glycerol/50 mM
phosphate buffer (pH 8.0) solution at −20 °C.
nome DNA isolation kit. The putative genes of PepR1 and PepR2 were
amplified using a pair of primers for each gene (R1-N 5′-GGTCTGCA
GATGAAACAAGGAACGACAATC-3′; R1-C 5′-GTTCAGTTAAGCTTTATTTT
TGATTAAAGC-3′; R2-N 5′-ATGAAAAACGTGACACGAATTTTAACGC-3′;
R2-C 5′-TTAGCGGCCCAATTGATCAGAAAAATAG-3′). The primer se-
quences are based on the sequence data in GenBank (pepR1, accession:
AL935263, region: 792,560–793,468; pepR2, accession: AL935263,
region: 2,597,065–2,597,961). PCR reaction mixtures (50 μL) were
prepared with 20 pmol of each primer, 20 ng of genome DNA template,
0
.04 mM of each dNTP, and 0.5 U of Pfu polymerase (Fermentas Interna-
tional), and then subjected to 50 PCR cycles (1 min at 95 °C, 30 s at 55 °C,
and 2 min at 72 °C). After PCR amplification, the mixtures were purified
with the EZ-10 Spin Column PCR Purification Kit (VWR International).
The amplified gene DNA fragment was phosphorylated and ligated
in the Sma I digested pUC18. The resultant recombinant plasmids
(
pUC18-pepR1 and pUC18-pepR2) were used as a template for further
PCR to construct expression plasmids. The pepR1 gene was amplified
using the following primers: 5′-CCGCCATGGAGTTGAAACAAGGAAC-3′
and 5′-TTAAAAGCTTTTTTTGATTAAAGCTGCCA-3′, where underlined
sections indicate Nco I (N-terminus) and Hind III (C-terminus) restric-
tion enzyme sites that flanked the ends of the open reading frame.
The pepR2 fragment was amplified with the same primers for genome
DNA PCR. The PCR reaction mixtures (total volume of 50 μL) consisted
of 20 pmol of each primer, 20 ng of recombinant pUC18-pepR1 or
pUC18-pepR2 DNA template, 0.04 mM of each dNTP, and 0.5 U of Pfu
polymerase. The PCR program was set to 25 cycles of 30 s at 95 °C for
denaturation, 30 s at 55 °C for annealing, and 2 min at 72 °C for exten-
sion. The amplified PCR fragment of pepR1 was cleaved with the desig-
nated restriction enzymes, and was ligated into Nco I–Hind III-digested
pET32b(+) plasmids. The PepR2 gene fragment was ligated into the
Sma I restriction site of pKK223-3. The recombinant plasmids were
introduced into Escherichia coli TOP10F′, and positive clones were con-
firmed using DNA sequencing. The positive recombinant pepR1 plasmid
then transformed E. coli Rosetta for overexpression.
2.5. Enzyme assay
The amount of proline released from the Pro-Xaa peptide substrates
was determined using a ninhydrin solution (15 mg/mL (w/v) ninhydrin
and 0.03 mL/mL (v/v) glacial acetic acid in n-butanol (pH 1) adjusted
with HCl), and the amount of liberated non-proline amino acids was
determined using a cadmium (Cd) ninhydrin solution (0.8 g ninhydrin
in 80 mL 90% ethanol and 10 mL glacial acetic acid supplemented with
1 g cadmium chloride [14]). Dipeptides were hydrolysed in 100 μL of
buffered reaction mixtures. The reaction was initiated by adding the en-
zyme solution. An aliquot (20 μL) was transferred from the reaction
mixture to 200 μL of the ninhydrin solution four times at 1 min intervals
(or with longer time intervals if the enzyme activities were low). When
the Cd-ninhydrin solution was used, the reaction was conducted in a
500-μL reaction mixture, and a 100-μL aliquot was mixed with 200 μL

Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
93
of Cd-ninhydrin solution at each time interval. The mixtures were heat-
ed at 95 °C (or 85 °C for the Cd-ninhydrin method) for 5 min and chilled
on ice, and then the absorption was measured at 500 nm. The increase
in absorption was converted to the amount of liberated amino acids
based on the reference made with standard amino acid solutions. All
measurements were conducted in triplicate.
molecules in a solution was used to determine the translational diffu-
sion coefficient and, subsequently, the hydrodynamic radius from the
Stokes–Einstein equation:
kT
DT ¼
6
πηRH
2
.6. pH and temperature dependency
where D
stant 1.38 × 10
viscosity, and R
is utilized to calculate the molecular weight. The calculated molecular
weight (MW) is the power to the hydrodynamic radius (R ) times the
constant of the hydrodynamic radius factor (formula: MW (kDa) =
T
is the translation diffusion coefficient, k is the Boltzmann con-
−
23
The pH dependency of recombinant prolinase activities was mea-
sured in the range of pH 3.0 to 10.0 using the following buffer systems:
0 mM sodium acetate (pH 3.0–4.5), 50 mM sodium citrate
pH 4.5–6.0), 50 mM MES (pH 5.5–6.5), 50 mM sodium phosphate
pH 6.0–7.5), 50 mM MOPS (pH 6.5–7.5), 50 mM HEPES (pH 7.0–8.0),
0 mM Tris–HCl (pH 7.5–9.0), and 50 mM sodium carbonate
, T is the temperature in kelvin (K), η is the solvent
H
is the hydrodynamic radius. The hydrodynamic radius
5
(
(
5
(
H
Power
[R
H
Factor ∗ R
H
(nm)]
). DLS data were evaluated with the Dynamics
pH 9.0–10.0). Hydrolysis reactions of the recombinant prolinases
v.5.26.60 software package (Protein Solutions Inc.).
were conducted at 37 °C and with 2 mM Pro-Gly as the substrate. Sim-
ilarly, the temperature dependency of recombinant prolinase activities
was determined at a series of temperatures (5, 15, 20, 25, 30, 37, 40,
Gel filtration chromatography for molecular mass estimation of r-
PepR1 and r-PepR2 was carried out with a Superdex 200 HR 10/300
gel filtration column. Standard proteins used in the separation were
bovine thyroglobulin (MW 670,000), bovine γ-globulin (MW
158,000), chicken ovalbumin (MW 44,000), horse myoglobin (MW
17,000), and vitamin B12 (MW 1350). The molecular weights of
4
2
5, and 55 °C) at pH 7.5 for r-PepR1 and at pH 8 for r-PepR2 with
mM Pro-Gly as the substrate. All experiments were carried out in
triplicate.
r-PepR1 and r-PepR2 were determined from the R
recombinant prolinase in the relative comparison with the R
f
values of each
values of
2
.7. Substrate specificities and enzyme kinetics
f
the standard proteins.
A variety of peptides were examined as the substrates, including
Pro-Pro, Pro-Glu, Pro-Met, Pro-Ser, Pro-Arg, Pro-Phe, Pro-Leu, Pro-Gly,
Met-Ala, Leu-Leu, Leu-Gly-Gly, and Pro-Gly-Gly. Substrate specificities
were examined for all these peptides at 2 mM substrate. Enzyme kinetic
parameters were measured for Pro-Xaa substrates under optimum con-
ditions: in 50 mM HEPES buffer (pH 7.5) and at 25 °C for r-PepR1; and in
2.10. Examination using proteinase inhibitors
The enzyme preparation was pre-incubated with 1 mM inhibitor
for 2 h in 10 mM HEPES buffer (pH 7.5) at 25 °C. The inhibitors examined
were PMSF (phenylmethylsulfonyl fluoride), EDTA, DTT (dithiothreitol),
MAL (maleimide; 2,5-pyrroledione), NEM (N-ethylmaleimide), and
TLCK (tosyllysine chloromethyl ketone hydrochloride). Residual activities
were determined under the optimized conditions (50 mM HEPES buffer
(pH 7.5), 25 °C for r-PepR1; 50 mM HEPES buffer (pH 8.0), 30 °C for r-
PepR2). For EDTA and DTT treatments, the reaction mixtures included
1 mM EDTA and 1 mM DTT, respectively. All experiments were conducted
in triplicate.
5
0 mM HEPES buffer (pH 8.0) and at 30 °C for r-PepR2. Enzyme kinetic
values K
m
and kcat were obtained using a method described by Sakoda
and Hiromi [15] for each substrate. All experiments were conducted in
triplicate.
2
.8. Secondary structures and thermal stability
Secondary structure and thermal stability of the recombinant
prolinases were measured using far UV circular dichroism (CD) spec-
troscopy (PiStar-180 CD spectrophotometer, Applied Photophysics
Ltd., Leatherhead, Surrey, UK) at the Saskatchewan Structural Sciences
Centre, University of Saskatchewan. For a secondary structure test, the
CD spectra were recorded from 190 nm to 250 nm for 300 μL of pure
recombinant prolinase solutions. Collected data were analysed using
CDNN software [16] to deduce five different secondary structure frac-
tions (α-helix, parallel and antiparallel β-sheets, β-turn, and random
coil). Thermal denaturation temperature was deduced from the change
of CD spectra over the temperature range from 20 to 85 °C. CD signal of
pure recombinant prolinase solutions (700 μL) was measured at 222 nm
over a temperature range from 25 to 90 °C. Correlation between the
rates of CD spectrum change over the temperature range was calculat-
2.11. Computational molecular modelling
The deduced protein sequences from L. plantarum WCFS1 prolinase-
coding genes pepR1 and pepR2 were submitted to the 3D-JIGSAW server
[18–20] and I-TASSER server [21–23]. The likeliness of models was
judged by indicative factors, including energy minimization of mole-
cule, C-score, TM-score, RMSD, number of decoys, and cluster density.
2.12. Phylogenic analysis
Using the PepR2 protein sequence data (GenBank accession no.:
CCC79995.1) as the search model, a BLASTP search was conducted
against non-redundant protein sequences (nr) on the NCBI server. The
resulting multiple sequence alignment was submitted to analyse the
phylogenic relations on the NCBI server. The phylogenetic data were
analysed and the phylogenic tree was drawn on Dendroscope [24].
ed, and T
M
(the midpoint temperature of the unfolding transition)
was determined as the temperature where the signal strength change
reached its maximum [17].
2
.9. Molecular mass estimation
3. Results
Molecular mass of the recombinant prolinases was determined with
3.1. Purification of recombinant prolinases
dynamic light scattering (DLS) and gel filtration chromatography. DLS
measurements were performed with a DynaPro-MS800 instrument
Recombinant prolinase PepR1 was expressed using pET32b(+) as
the expression vector, and the His-tag was removed after purification
to yield recombinant PepR1 (r-PepR1). PepR2 was also expressed
using the pET32b(+) vector, but thrombin and enterokinase treatments
failed to remove the His-tag (data not shown). Therefore, r-PepR2 was
purified with a non-tag expression (using the pKK223-3 vector). Both
(
Protein Solutions Inc., Wyatt Technology, Santa Barbara, CA, USA)
using noninvasive backscatter detection. Filtrated recombinant
prolinase protein solutions (20 μL) were placed in a quartz cuvette
and the measurements were made at 20 °C. The time-dependent fluctu-
ation of the light scattered by the recombinant prolinase protein

94
Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
r-PepR1 and r-PepR2 were successfully purified to homogeneity as
k
cat values for Pro-Gly showed the best values among the Pro-Xaa
determined by SDS-PAGE analysis.
substrates examined. The K values indicated that Pro-Leu (3.4 mM),
m
Pro-Ser (3.0 mM), and Pro-Arg (3.9 mM) also had good affinity to the
enzyme, comparable to that of Pro-Gly (3.3 mM). However, their kcat
values were 35% to 59% of the Pro-Gly value, so the overall activity
3
.2. Temperature and pH dependency
m
(kcat/K ) for these Pro-Xaa substrates was lower than that of Pro-Gly.
Meanwhile, the kinetic parameters of r-PepR2 indicated more stringen-
cy of this enzyme. The difference between the most preferred Pro-Arg
Recombinant prolinase r-PepR1 showed the highest activity at 25 °C,
whereas r-PepR2 did so at 30 °C (Fig. 1). PepR1 exhibited 33.4% of the
highest activity at 5 °C and decreased its activity to 32.1% at 40 °C, indi-
cating its psychrophilic nature. In contrast, PepR2 was more mesophilic
in nature, and exhibited only 6.3% of the highest activity at 5 °C and
retained 68.2% activity at 40 °C. In terms of pH dependency, the highest
r-PepR1 activity was observed at pH 7.5 in 50 mM HEPES buffer
Fig. 2A), whereas the highest r-PepR2 activity was obtained at pH 8.0
in 50 mM HEPES buffer (Fig. 2B). Both prolinases showed similar
profiles of pH dependency in activity within the examined pH range.
and least-favoured Pro-Glu reached 12.3-fold in a kcat/K
ison, which is similar to the kcat/K range in r-PepR1. r-PepR2 exhibited
cat/K values for the second-, third-, and fourth-preferred substrates
Pro-Ser, Pro-Gly, and Pro-Leu, respectively) at only 38%, 18%, and 16%
of the most-preferred Pro-Arg substrate. However, r-PepR1 exhibited
values at 58%, 57% and 56% of the second-, third-, and fourth-
m
value compar-
m
k
m
(
(
kcat/K
m
preferred substrates (Pro-Leu, Pro-Met, and Pro-Ser, respectively), indi-
cating higher stringency in r-PepR2. This stringency of r-PepR2 was due
to the differences in affinity (K
m
) rather than the differences in kcat
3
.3. Substrate specificity
values. For r-PepR2, the kcat values were 185.3, 170.1, 106.8, and
−
1
2
17.2 min for Pro-Arg, Pro-Ser, Pro-Gly, and Pro-Leu, respectively;
the K values increased from 2.1 mM of Pro-Arg to 4.9, 6.7, and
5.1 mM for Pro-Ser, Pro-Gly, and Pro-Leu, respectively. These observa-
Substrate preferences were examined using the activity towards
mM peptide substrates (Table 1). r-PepR1 showed the highest activity
m
2
1
−
1
−1
towards Met-Ala at 1.325 μmol·min ·mg . All Pro-Xaa substrates
were hydrolysed by r-PepR1 at 0.038 to 0.359 μmol·min ·mg
tions indicated that r-PepR1 and r-PepR2 have differences in their active
sites, which influence both affinity and catalysis. It is likely that the
stringent substrate recognition also tightens the alignment of substrates
relative to the catalytic centre in r-PepR2, whereas r-PepR1 allows some
adjustment of substrate binding for alignment of substrates at the
expense of lower overall prolinase activity.
−
1
−1
.
−
1
−1
Pro-Gly-Gly also was degraded at 0.274 μmol·min ·mg , whereas
Leu-Leu hydrolysis and Leu-Gly-Gly hydrolysis were not observed.
These results indicated that r-PepR1 is less stringent in the preference
of peptide substrates, although it clearly exhibits activity as a prolinase.
Whereas r-PepR1 hydrolysed peptides other than Pro-Xaa, r-PepR2
showed activity only towards Pro-Xaa. Among the four non-Pro-Xaa
peptides, there was subtle activity towards Leu-Gly-Gly (0.2% of Pro-
Arg) but activities towards the others were not observed. The activities
towards Pro-Xaa were 0.648 to 4.209 μmol·min ·mg , which was
generally higher than the prolinase activities of r-PepR1. The results in-
dicated that r-PepR2 also was a prolinase, confirming that L. plantarum
WCFS1 has two prolinases.
3.5. Proteolysis modes
−
1
−1
The inhibition test results (Table 3) suggest that r-PepR1 and r-
PepR2 have different catalysis modes or different structures in the active
site. In our examination, r-PepR2 was clearly identified as a serine
proteinase. Both serine proteinase inhibitors (TLCK and PMSF) inhibited
r-PepR2 to residual activities below 13% after the 2-h treatment. Cyste-
ine modifiers (MAL and NEM) also inhibited this enzyme by 27% and
3
.4. Enzyme kinetics
3
5%, respectively, suggesting the existence of cysteine residues near
Table 2 summarizes the kinetic parameters of r-PepR1 and r-PepR2.
the catalytic centre or substrate binding site. The r-PepR1 inhibition
Pro-Gly was the most preferred substrate for r-PepR1, as both K
m
and
test showed that cysteine modifiers inhibited this enzyme more than
Fig. 1. Effects of temperature on recombinant prolinase r-PepR1 and r-PepR2 activity. The highest activities of r-PepR1 and r-PepR2 were normalized to 100%. All experiments were
conducted in triplicate. Error bars indicate standard deviations.

Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
95
(
A)
(
B)
Fig. 2. Effect of pH on recombinant prolinase activity (A) r-PepR1 and (B) r-PepR2. Different pH buffers (50 mM) were utilized: sodium acetate buffer (pH 3.0–4.5), sodium citrate buffer
(
(
pH 4.5–6.0), MES buffer (pH 5.5–6.5), sodium phosphate buffer (pH 6.0–7.5), MOS buffer (pH 6.5–7.5), HEPES buffer (pH 7.0–8.0), Tris buffer (pH 7.5–9.0), and sodium carbonate buffer
pH 9.0–10.0). Hydrolysis of Pro-Gly (2 mM) was quantified by the ninhydrin method at Abs 500 nm. The highest r-PepR1 and r-PepR2 activities were normalized to 100%. All experiments
were done in triplicate.
serine proteinase inhibitors. Treatment with NEM and MAL showed the
residual activities at 7.3% and 17.1%, respectively, whereas TLCK and
PMSF showed residual activities at 47% and 82%, respectively. The
results did not clearly show the mode of proteolysis in r-PepR2. Consid-
ering that there are not many examples showing other catalytic modes
in the PepR/PepI/PepL family, it is likely that r-PepR1 also is a serine
proteinase, but the almost complete inhibition by NEM may indicate
that r-PepR1 is a cysteine proteinase. Also, we note that EDTA inhibition
3.6. Molecular mass estimation
DLS and gel filtration chromatography were employed to determine
the native molecular mass of the recombinant prolinases. The intensity
fluctuations of r-PepR1 and r-PepR2 from DLS data were processed into
diffusion coefficients, and the correlation between the diffusion coeffi-
cients and the decay times of r-PepR1 and r-PepR2 was plotted. The cor-
relation curve was an indicator to determine whether the signal was
correlated and to compare the size of the particles based on the diffusion
[25]. Both r-PepR1 and r-PepR2 exhibited typical exponential decay. The
hydrodynamic radiuses of r-PepR1 and r-PepR2 were calculated from
the Stokes–Einstein equation after outlying hydrodynamic radius data
were filtered out on the basis of SOS (sum of square) values [26]. The
native molecular mass of r-PepR1 was determined to be 156.5 kDa;
this result implied that r-PepR1 was a tetramer with a molecular mass
(
63% residual activity) and the existence of known or suggested
metallo-PepR in human kidney, swine kidney [8,11,12], and Mycobacte-
rium tuberculoses H37Rv (NCBI reference sequence: NP_217298.1) indi-
cate that r-PepR1 may be a metalloproteinase. However, we observed
that the addition of excess amounts of metal cations does not result in
the recovery of r-PepR1 activity (data not shown); therefore, we con-
cluded that r-PepR1 is not a metalloproteinase.

96
Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
Table 1
Substrate specificity of Lactobacillus plantarum WCFS1 prolinase enzymes.
Substrate
r-PepR1
r-PepR2
Specific activity
Relative activity
(%; Pro-Gly = 100%)
Specific activity
(μmol/min·mg)
Relative activity
(%; Pro-Arg = 100%)
(μmol/min·mg)
Pro-Glu
Pro-Phe
Pro-Gly
Pro-Leu
Pro-Met
Pro-Pro
Pro-Arg
Pro-Ser
Met-Ala
Pro-Gly-Gly
Leu-Gly-Gly
Leu-Leu
0.038 ± 0.002
0.064 ± 0.003
0.359 ± 0.008
0.200 ± 0.002
0.210 ± 0.002
0.061 ± 0.003
0.098 ± 0.003
0.180 ± 0.001
1.325 ± 0.067
0.274 ± 0.045
ND
10.5 ± 0.7
17.8 ± 0.9
100.0 ± 3.0
55.8 ± 1.3
58.6 ± 1.4
17.0 ± 1.0
27.4 ± 1.1
50.3 ± 1.1
369.3 ± 20.3
76.3 ± 12.6
0.648 ± 0.006
0.870 ± 0.013
1.286 ± 0.029
1.037 ± 0.032
0.745 ± 0.010
0.837 ± 0.018
4.209 ± 0.079
2.052 ± 0.041
0.002 ± 0.000
ND
15.4 ± 0.3
20.7 ± 0.5
30.6 ± 0.9
24.6 ± 0.9
17.7 ± 0.4
19.9 ± 0.6
100.0 ± 2.6
48.8 ± 1.3
0.0 ± 0.0
0.009 ± 0.002
ND
0.2 ± 0.1
ND
ND: Not detectable.
approximately four times greater than its monomer mass (39.2 kDa).
Likewise, r-PepR2 was determined to have a native molecular mass of
observation matched the temperature dependency of each enzyme,
i.e., psychrophilic and mesophilic enzymes.
1
28.2 kDa and was a tetramer, as the mass was also about four times
greater than its monomer mass (34.6 kDa).
3.8. Modelling of prolinases
In gel filtration chromatography, r-PepR1 was eluted between pro-
tein peaks of γ-globulin and ovalbumin, and the molecular mass was
calculated to be 144.7 kDa, which was approximately four times that
of its protomer. Likewise, the molecular mass of r-PepR2 was calculated
to be 124.8 kDa, which is also nearly equal to four protomers. The results
were in agreement with the estimates obtained in the DLS study.
The deduced protein sequences from L. plantarum WCFS1 prolinase-
coding genes pepR1 and pepR2 were submitted to the 3D-JIGSAW server
[18–20] and I-TASSER server [21–23]. Results from the two servers were
generally in agreement for the predicted PepR1 and PepR2 molecular
structures. PepR1 was predicted to have the highest structural similarity
with the known structure of Mycobacterium smegmatis proline
iminopeptidase (EC 3.4.11.5). PepR2 was predicted to have the highest
structural similarity with the known structure of Thermoplasma
acidophilum proline iminopeptidase (EC 3.4.11.5). The likeliness of
models is judged by indicative factors, including energy minimization
of the molecule, C-score, TM-score, RMSD, number of decoys, and clus-
ter density. The PepR1 model was based on T. acidophilum proline
iminopeptidase with minimized energy at −391.20 cal/mol, C-score
3
.7. Protein secondary structure and thermal stability
Circular dichroism (CD) spectroscopy was applied to investigate
the secondary structure of the recombinant prolinases. The CD spectra
of r-PepR1 and r-PepR2 were tested on the far-UV range of 190 to
2
60 nm. The results as summarized in Table 4 show that both r-PepR1
and r-PepR2 contained high amounts of β-sheets.
Thermal denaturation of proteins is defined as the transition of
folded to unfolded structure. In this research, the unfolding of prolinase
at 1.03, T -score at 0.85 ± 0.08, number of decoys at 9493, and cluster
M
density at 0.9091. The PepR2 model was based on M. smegmatis
was characterized for T
transition) using the change in ellipticity of the CD signal as an index
of unfolding [17]. The T of r-PepR1 was determined to be 302.30 ±
8.13 K, which approximates 29 °C, whereas the T of r-PepR2
was 321.49 ± 0.06 K, which approximates 48 °C. Thus, r-PepR1 had a
M
(the midpoint temperature of the unfolding
proline iminopeptidase with a minimized energy level of −413.04
cal/mol, C-score at 1.07, T -score at 0.86 ± 0.07, number of decoys at
M
M
10,200, and cluster density at 0.9346. The RMSD between the two
2
M
prolinases was 5.45 Å. These values for PepR1 and PepR2 indicated
that modelling errors were small. The number of decoys is the number
of structural decoys used in generating the model. The cluster density is
the number of structural decoys at a unit of space in the cluster. The
cluster densities of PepR1 and PepR2 were high, implying that the struc-
tures occurred often in the simulation trajectory and that fitted models
were high quality. The results support the premise that L. plantarum
WCFS1 prolinases (PepR1 and PepR2) belong to the proline peptidase
M
relatively low T , and r-PepR2 was more stable than r-PepR1. This
Table 2
Kinetic parameters of recombinant prolinases.
m
kcat/K )
(min−1·mM−1
Substrate
K
m
(mM)
k
cat (min−1)
r-PepR1a
Pro-Glu
Pro-Phe
Pro-Gly
Pro-Leu
Pro-Met
Pro-Pro
Pro-Arg
Pro-Ser
17.9 ± 5.8
6.9 ± 1.8
3.3 ± 0.3
3.4 ± 0.5
4.9 ± 0.9
9.6 ± 2.1
3.9 ± 0.8
3.0 ± 0.5
15.4 ± 3.7
9.7 ± 1.4
0.9 ± 0.3
1.4 ± 0.4
36.7 ± 1.6
21.7 ± 1.3
30.6 ± 2.8
15.3 ± 2.1
12.7 ± 1.2
18.3 ± 1.3
11.0 ± 1.2
6.4 ± 0.9
6.3 ± 1.3
1.6 ± 0.4
3.2 ± 0.7
6.2 ± 1.1
Table 3
Residual activity after inhibitor treatment of recombinant prolinases.
Relative residual activity (%)a
r-PepR1b
r-PepR2b
Inhibitor
EDTA
TLCK
PMSF
NEM
DTT
MAL
63.2 ± 10.8
46.8 ± 9.0
82.0 ± 6.2
7.3 ± 2.9
81.8 ± 3.7
17.1 ± 0.4
100.0 ± 2.8
98.8 ± 1.0
12.6 ± 0.0
1.2 ± 0.0
64.9 ± 0.0
99.1 ± 1.0
72.9 ± 0.0
100 ± 1.0
r-PepR2a
Pro-Glu
Pro-Phe
Pro-Gly
Pro-Leu
Pro-Met
Pro-Pro
Pro-Arg
Pro-Ser
10.8 ± 1.6
7.4 ± 1.2
6.7 ± 1.5
15.1 ± 4.6
5.4 ± 0.5
9.2 ± 1.4
2.1 ± 0.7
4.9 ± 0.8
78.7 ± 6.8
69.5 ± 5.5
106.8 ± 13.9
217.2 ± 47.0
52.8 ± 2.3
85.9 ± 7.2
185.3 ± 23.9
170.1 ± 15.0
7.3 ± 1.3
9.4 ± 1.6
15.8 ± 15.8
14.4 ± 5.3
9.9 ± 1.0
9.3 ± 1.6
89.6 ± 32.3
34.4 ± 6.6
Blank (no inhibitor)
a
The activity of controls (either r-PepR1 or r-PepR2) without inhibitors was normal-
ized to 100%.
a
Hydrolysis was conducted under the optimal condition: r-PepR1 in 50 mM HEPES
buffer (pH 7.5), 25 °C; and r-PepR2 in 50 mM HEPES buffer (pH 8.0), 30 °C.
b
Hydrolysis was conducted under the optimal condition: r-PepR1 in 50 mM HEPES
buffer (pH 7.5), 25 °C; and r-PepR2 in 50 mM HEPES buffer (pH 8.0), 30 °C.

Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
97
Table 4
which are substrates of aminopeptidases [32]. These observations
are in agreement with the results obtained for L. plantarum WCFS1
prolinases in this study. The difference in substrate specificity between
them may reflect differences in their evolution.
Estimated percentages of protein secondary structure components of r-PepR1 and r-
PepR2 from CD spectra.
α-Helix
Antiparallel
Parallel
β-Turn
Random coil
The optimum temperature and pH results suggested some interest-
ing differences between the L. plantarum WCFS1 prolinases. r-PepR1
had optimal activity at pH 7.5 and 25 °C, whereas r-PepR2 had optimal
activity at pH 8.0 and 30 °C. The optimum pH for prolinase activity is gen-
erally found in the pH range of 7.5 to 8.3 at weak alkali conditions [7], and
the optimum pH of the recombinant prolinase enzymes fell in the same
range. The denaturation temperature of r-PepR1 and r-PepR2 reflected
the differences in temperature dependency between the two prolinases.
The low optimum temperature of r-PepR1 corresponded to its thermal
r-PepR1
r-PepR2
23.1%
19.2%
27.8%
46.5%
10.1%
7.6%
19.9%
15.6%
34.0%
21.2%
family PepI/PepL/PepR, having the most similar three-dimensional
structure with PepI.
3
.9. Phylogenetic analyses
The phylogenetic tree of prolinases (Fig. 3) revealed that PepR2 be-
M
stability (T = 29 °C). The lower optimum temperature of PepR1 may
longs to a terminal branch at 7 nodes away from the common ancestor
of PepR1 and PepR2, whereas PepR1 is at 9 nodes below this ancestor.
The evolutionary distances from this ancestor to PepR1 and PepR2 are
be partly due to its low thermal stability. The denaturation temperature
of r-PepR1 was almost 20 °C lower than that of r-PepR2 (29 °C versus
48 °C). The overall energy level results of the molecular models also sup-
ported the lower thermal stability of PepR1 by showing 21.86 cal/mol
higher energy levels, suggesting lower stability. The optimum tempera-
ture and pH of reaction often reflect the optimum conditions for growth
of the microorganism from which the prolinase originated. The
observed characteristics of L. plantarum WCFS1 prolinases were not
exceptions. Interestingly, at lower temperatures (5 to 20 °C), r-PepR1
had higher activity than r-PepR2. At 15 °C, r-PepR1 showed 46.4% of
the activity at its optimum temperature, whereas r-PepR2 only showed
19.6% activity. Even at 5 °C, r-PepR1 showed 33.4% of maximum activity.
At higher temperatures, r-PepR2 had higher activity than r-PepR1.
These differences in temperature dependency suggest the advantage
of a dual-prolinase system that allows activity at a broader range of
environmental temperature. Such a system would be beneficial
for L. plantarum WCFS1, which was originally isolated from low-
temperature fermented pickled vegetables.
This study showed that PepR1 and PepR2 share similarities in their
secondary and tertiary structures, and are homo-tetrameric in their
quaternary structure. CD analyses showed that both r-PepR1 and r-
PepR2 had a higher ratio of β-sheets (r-PepR1 with 37.9% and r-PepR2
with 54.1%) over other secondary components. The high ratio of
β-sheets in r-PepR2 may contribute to the stability of the enzyme.
β-Sheets are more stable than helical regions of an enzyme at high
temperature [33]. Perczel et al. suggested that an antiparallel β-sheet
structure is more thermally stable than parallel β-sheets due to the
favourable arrangements of strands to form short, strong, and stable
hydrogen bonds [34]. Our CD analyses showed that the antiparallel
arrangement was 46.5% of the total secondary structure of r-PepR2,
whereas it made up only 27.8% in r-PepR1, which may be reflected in
the lower thermal stability of PepR1 compared with PepR2.
The strong inhibition by serine proteinase inhibitors on r-PepR2
indicated that this enzyme is a serine proteinase. The BLASTP search
sequence analysis indicated that PepR2 has a motif commonly found
in serine proteinases (GXSXGG) at residues 108 to 113, which provides
evidence of the catalytic mode of PepR2. Although the inhibition test
results (Table 3) showed that r-PepR1 was inhibited by cysteine
proteinase inhibitors more than by serine proteinase inhibitors, PepR1
also has the sequence motif at residues 107 to 112. This could be an
indication of the serine proteolytic nature of PepR1.
Amino acid sequence indicated that PepR1 and PepR2 have two and
three cysteine residues, respectively. The three-dimensional models
(Fig. 4) illustrate that Cys30 (PepR1) and Cys31 (PepR2) are conserved
between the two models; however, Cys104 and Cys191 of PepR2 and
Cys148 of PepR1 are not conserved and are situated distinctively be-
tween PepR1 and PepR2. Cys104 and Cys192 of PepR2 are fully exposed
and have little interaction with surrounding residues, whereas Cys148
of PepR1 is half buried in the structure and interacts with surrounding
residues. Also, the serine residue in the GXSXGG motif is separated by
a loop structure from Cys148 of PepR1. The inactivation of PepR1 with
cysteine modifiers may be caused by the displacement of the loop via
0
.2274 and 0.4405, respectively. The taxon of PepR2 has 12 entries on
it, and a half of the entries are L. plantarum, 4 are other lactobacilli,
and 2 are Weisella sp. In contrast, PepR1 belongs to a taxon with 35
leaves that consist of a variety of 30 lactobacilli and 5 pediococci.
4
. Discussion
Prolinase (PepR) has been categorized using BLASTP analysis, multi-
ple sequence alignments, and three-dimensional structure alignment
27]. These analyses show that prolinase belongs to the prolyl
[
oligopeptidase family (S33 family) of PepI/PepL/PepR (PepI:
iminopeptidase; PepL: aminopeptidase). The family shares a consensus
catalytic motif of GXSXGG, and this family is included in the clan of SC
(
serine carboxypeptidase), having active site residues Ser-His-Asp in
the sequence [28]. Moreover, a hierarchical and structure-based classifi-
cation of peptidases through MEROPS analyses indicates a close
relationship between PepR (prolinase from Lactobacillus) and PepI
(
bacterial iminopeptidase from various species). (MEROPS is an online
database for peptidases and their inhibitors (http://merops.sanger.ac.
uk) [29,30].) Most known Lactobacilli have one gene corresponding to
PepR, but L. plantarum WCFS1 remarkably has two putative prolinase
genes corresponding to PepR1 and PepR2. Molecular modelling of
these two putative gene products in this study supported their nature
as prolinase by showing reasonably fitted models within the PepI/
PepL/PepR family. Although the sequence-based analyses suggested
that there are two putative genes coding for prolinases, it was not prov-
en. Our recombinant prolinases both exhibited hydrolysis of Pro-Xaa di-
peptides, indicating that the two putative genes indeed code prolinases.
This result then raised other questions: what are the differences and
why does L. plantarum WCFS1 have two prolinases? Both r-PepR1 and
r-PepR2 showed a broad specificity at the substrate C-terminal, includ-
ing small residue (Gly), hydrophobic residues (Leu, Pro, Met), acid res-
idue (Glu), nucleophilic residue (Ser), basic residue (Arg), and aromatic
residue (Phe). However, r-PepR1 and r-PepR2 showed different and
unique preferences. r-PepR1 had the highest catalytic efficiency
m
towards Pro-Gly, and lower but similar kcat/K values among Pro-Met,
Pro-Ser, and Pro-Leu; Met-Ala and Pro-Gly-Gly were also good sub-
strates, indicating the broader specificity of r-PepR1. r-PepR2 showed
the highest activity with Pro-Arg, and lower activities towards other r-
PepR1-preferred dipeptides: 30.6% for Pro-Gly, 17.7% for Pro-Met,
4
m
8.8% for Pro-Ser, and 24.6% for Pro-Leu at 2 mM, and the kcat/K values
also confirmed lower activities to these Pro-Xaa substrates. Therefore,
substrate specificity is more stringent in PepR2. This stringency is most-
ly due to the catalysis itself (kcat) and not to the substrate affinity (K ).
m
The PepR of Lactobacillus helveticus hydrolyses proline-containing di-
peptides (e.g., Pro-Leu, Pro-Met, Pro-Phe) and dipeptides not containing
proline residues (e.g., Gly-Leu, Leu-Arg, Leu-Ser, Met-Ala, Met-Gly, Met-
Leu, Met-Phe, Ser-Phe, Thr-Leu, Pro-βNA) [31]; the PepR of L. rhamnosus
hydrolyses Leu-βNA (Leu-β-naphthylamide), Phe-βNA, and Pro-βNA,

98
Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
0.1
Proline_iminopeptidase__Lactobacillus_paracasei_.1
MULTISPECIES__proline_iminopeptidase__Lactobacillus_casei_group_.1
prolyl_aminopeptidase__Lactobacillus_casei_.1
prolyl_aminopeptidase__Lactobacillus_paracasei_.1
prolyl_aminopeptidase__Lactobacillus_casei_.2
proline_iminopeptidase__Lactobacillus_casei_
prolyl_aminopeptidase__Lactobacillus_kisonensis_
prolyl_aminopeptidase__Lactobacillus_coryniformis_.1
prolyl_aminopeptidase__Lactobacillus_coryniformis_.2
prolyl_aminopeptidase__Lactobacillus_paracasei_.2
prolyl_aminopeptidase__Lactobacillus_paracasei_.3
MULTISPECIES__prolyl_aminopeptidase__Lactobacillus_casei_group_
prolyl_aminopeptidase__Lactobacillus_casei_.3
prolyl_aminopeptidase__Lactobacillus_paracasei_.4
prolyl_aminopeptidase__Lactobacillus_paracasei_.5
MULTISPECIES__proline_iminopeptidase__Lactobacillus_casei_group_.2
prolyl_aminopeptidase__Lactobacillus_paracasei_subsp__paracasei_Lpp123_
prolyl_aminopeptidase__Lactobacillus_paracasei_subsp__paracasei_Lpp48_
MULTISPECIES__proline_iminopeptidase__Lactobacillus_casei_group_.3
Proline_iminopeptidase__Lactobacillus_paracasei_.2
prolinase__Lactobacillus_rhamnosus_
prolyl_aminopeptidase__Lactobacillus_rhamnosus_.1
prolyl_aminopeptidase__Lactobacillus_rhamnosus_.2
prolyl_aminopeptidase__Lactobacillus_rhamnosus_.3
prolyl_aminopeptidase__Lactobacillus_rhamnosus_.4
prolyl_aminopeptidase__Lactobacillus_casei_subsp__casei_ATCC_393_
prolyl_aminopeptidase__Lactobacillus_casei_.4
prolyl_aminopeptidase__Lactobacillus_zeae_
prolinase__Lactobacillus_zeae_
prolyl_aminopeptidase__Lactobacillus_rhamnosus_K32_
prolyl_aminopeptidase__Lactobacillus_plantarum_.1 PepR1
prolyl_aminopeptidase__Lactobacillus_plantarum_WHE_92_
prolyl_aminopeptidase__Lactobacillus_plantarum_.2
prolyl_aminopeptidase__Lactobacillus_plantarum_.3
prolyl_aminopeptidase__Lactobacillus_plantarum_.4
MULTISPECIES__prolyl_aminopeptidase__Lactobacillus_.1
prolyl_aminopeptidase__Lactobacillus_fabifermentans_.1
proline_iminopeptidase__Lactobacillus_shenzhenensis_
prolyl_aminopeptidase__Lactobacillus_harbinensis_
prolyl_aminopeptidase__Pediococcus_acidilactici_
prolyl_aminopeptidase__Pediococcus_pentosaceus_.1
prolyl_aminopeptidase__Pediococcus_pentosaceus_.2
proline_specific_peptidases_family_protein__Pediococcus_pentosaceus_
prolyl_aminopeptidase__Pediococcus_claussenii_
prolyl_aminopeptidase__Lactobacillus_suebicus_
prolyl_aminopeptidase__Lactobacillus_farciminis_
prolyl_aminopeptidase__Lactobacillus_paralimentarius_
prolyl_aminopeptidase__Lactobacillus_nodensis_
prolyl_aminopeptidase__Lactobacillus_versmoldensis_
prolyl_aminopeptidase__Lactobacillus_curvatus_
prolyl_aminopeptidase__Lactobacillus_sakei_
prolyl_aminopeptidase__Lactobacillus_fuchuensis_
prolyl_aminopeptidase__Lactobacillus_rossiae_
prolyl_aminopeptidase__Lactobacillus_gigeriorum_
prolyl_aminopeptidase__Lactobacillus_gastricus_
prolyl_aminopeptidase__Lactobacillus_reuteri_.1
prolyl_aminopeptidase__Lactobacillus_reuteri_.2
prolyl_aminopeptidase__Lactobacillus_reuteri_.3
prolyl_aminopeptidase__Lactobacillus_reuteri_.4
prolyl_aminopeptidase__Lactobacillus_reuteri_.5
prolyl_aminopeptidase__Lactobacillus_reuteri_.6
prolyl_aminopeptidase__Lactobacillus_reuteri_.7
prolyl_aminopeptidase__Lactobacillus_antri_
prolyl_aminopeptidase__Lactobacillus_oris_.1
prolyl_aminopeptidase__Lactobacillus_oris_.2
prolyl_aminopeptidase__Lactobacillus_vaginalis_
PREDICTED__uncharacterized_protein_LOC102656480__Apis_mellifera_
proline_specific_peptidase__Lactobacillus_otakiensis_
prolyl_aminopeptidase__Lactobacillus_buchneri_.1
prolyl_aminopeptidase__Lactobacillus_buchneri_.2
proline_iminopeptidase__Lactobacillus_farraginis_JCM_14108_.1
MULTISPECIES__prolyl_aminopeptidase__Lactobacillus_.2
proline_iminopeptidase_Pip__Lactobacillus_kunkeei_EFB6_
prolyl_aminopeptidase__Lactobacillus_kunkeei_
prolyl_aminopeptidase__Lactobacillus_fructivorans_
prolyl_aminopeptidase__Lactobacillus_sanfranciscensis_
prolyl_aminopeptidase__Lactobacillus_florum_
prolyl_oligopeptidase__Lactobacillus_oryzae_JCM_18671_
prolyl_aminopeptidase__Sporolactobacillus_vineae_
proline_iminopeptidase__Lactobacillus_sucicola_JCM_15457_
prolyl_aminopeptidase__Lactobacillus_vini_
prolyl_aminopeptidase__Lactobacillus_mali_
prolyl_aminopeptidase__Lactobacillus_ruminis_.1
prolyl_aminopeptidase__Lactobacillus_ruminis_.2
proline_iminopeptidase__Lactobacillus_ruminis_CAG_367_
prolyl_aminopeptidase__Lactobacillus_ruminis_.3
prolyl_aminopeptidase__Lactobacillus_pobuzihii_
prolyl_aminopeptidase__Lactobacillus_acidipiscis_
prolyl_aminopeptidase__Tetragenococcus_halophilus_
prolyl_aminopeptidase__Weissella_confusa_
prolyl_aminopeptidase__Weissella_cibaria_
proline_iminopeptidase__Lactobacillus_otakiensis_
proline_iminopeptidase__Lactobacillus_farraginis_JCM_14108_.2
prolyl_aminopeptidase__Lactobacillus_fabifermentans_.2
proline_iminopeptidase__PIP___Lactobacillus_pentosus_IG1_
prolyl_aminopeptidase__Lactobacillus_plantarum_.5
prolyl_aminopeptidase__Lactobacillus_plantarum_.6
prolyl_aminopeptidase__Lactobacillus_plantarum_.7
Proline_specific_peptidase__Lactobacillus_plantarum_
prolyl_aminopeptidase__Lactobacillus_plantarum_.8 PepR2
prolyl_aminopeptidase__Lactobacillus_plantarum_.9

Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
99
Fig. 4. Structure comparison between PepR1 and PepR2. Structure models of prolinases were constructed using homology modelling methods (A: PepR1; B: PepR2). White spheres
represent cysteine residues, and the black sphere represents the serine residue in the GXSXGG serine proteinase motif. Cys148 of PepR1 is located near the serine residue and they are
only separated by a loop. The chemical modification of Cys148 changes the bulkiness of this residue, and this change may interfere the substrate alignment to the serine catalytic centre.
PepR2 has their non-conserved cysteine residues (Cys104 and Cys192) in distant parts of the enzyme body from the serine residue, and they are exposed to the solvent, assuming smaller
structural influence upon chemical modifications. (Two-column fitting).
the bulkiness change of modified-Cys148. This speculation is supported
by the observation that the smaller MAL inhibitor inhibited PepR1 to a
lesser degree (17% residual activity) than the larger NEM (8% residual
activity). Larger NEM modification could block the substrate from
aligning to the catalytic site.
PepR1 and PepR2 with low sequence identity can be assumed to
have arisen by divergent evolution from the same ancestor. The diver-
gent evolution may have resulted in PepR1 and PepR2 becoming highly
specific for their respective substrates and their substrate-binding
pockets becoming different. L. plantarum PepR1 and PepR2 differ in
their DNA sequences (only 55.5% identity), substrate specificity, opti-
provided in this study, including optimal conditions for the enzyme
activities, substrate specificities, kinetic parameters, proteolysis
modes, and enzyme structures (secondary, tertiary and quaternary
structures), is essential for such applications. This research also showed
several interesting characteristics of L. plantarum WCFS1 prolinases
PepR1 and PepR2. Depending on the environmental conditions, proline
liberation for assimilation and recycling can be mediated through differ-
ent characteristics of the two isozymes. These isozymes work together
to hydrolyse prolyl dipeptides and to compensate prolidase-deficient
activity with broader specificity, broader pH and temperature for their
activities. The multiplicity of L. plantarum WCFS1 prolinases would
help reduce bitterness in fermented foods manufacture with higher
efficiency.
m
mum temperature and pH, enzyme kinetics with different K and kcat
values, thermal stability, and proteolysis mode; however, they are sim-
ilar in their secondary, tertiary, and quaternary structures, and have the
same enzymatic functionality (hydrolysis of Pro-Xaa). Moreover, the
evolutionary pressure of two putative prolinases not only contributes
to the diversity of catalytic activities but also to adaptation to different
pH and temperature conditions. Phylogenetic analyses (Fig. 3) suggest
an evolutionary distance between the two prolinases, which supports
the argument above. Evolutionary pressure may have resulted in two
forms of prolinase with compatible enzyme properties, as in PepR1
and PepR2. It likely contributes fine adjustments of metabolism to
meet the needs of different growth stages, and allows the use of the en-
zyme based on its environment. The presence of two enzymes would be
beneficial for L. plantarum WCFS1 to withstand the genetic problems
through encoding them in two different locus. The multiplicity of
prolinase isozymes allows them to adapt their catalytic abilities to an
extent in which they only share poor sequence homology. The prolinase
isozymes exhibited widely divergent kinetic characteristics showing
different preferences towards various substrates, which broadens the
substrate specificity as prolinase. With a wider range of optimum
temperature, pH and thermal stability, prolinase isozymes can comple-
mentarily and cooperatively contribute in maintaining the functionality
of prolinase under different environments.
Acknowledgments
This research was supported by a Discovery Grant (RGPIN 283277-
2
010) from the Natural Sciences and Engineering Research Council of
Canada. We acknowledge Sharmila Jayatilake and Dennis Marvin for
their assistance in the laboratory work, and Sylvia Yada for her contribu-
tion to the preparation of the manuscript.
References
[
[
1] D.F. Sunningham, B. O'Connor, Proline specific peptidases, Biochim. Biophys. Acta
1343 (1997) 160–186.
2] N. Ishibashi, I. Ono, K. Kato, T. Shigenaga, I. Shinada, H. Okai, S. Fukui, Role of the
hydrophobic amino acid residue in the bitterness of peptides, Agric. Biol. Chem.
49 (1987) 64–69.
[3] N. Ishibashi, T. Kubo, M. Chino, H. Fukui, I. Shinoda, E. Kikuchi, H. Okai, S. Fukui, Taste
of proline-containing peptides, Agric. Biol. Chem. 52 (1988) 95–98.
[
4] A. Milligan, R.A.C. Graham-Brown, D.A. Burns, I. Anderson, Prolidase deficiency: a
case report and literature review, Br. J. Dermatol. 121 (1989) 405–409.
[5] I. Myara, J.F. Stalder, Plasma prolidase and prolinase activity in prolidase deficiency,
Clin. Chem. 32 (1986) 562.
[
6] G. Miech, I. Myara, M. Mangeot, V. Voigtlander, A. Lemonnier, Prolinase activity in
prolidase-deficient fibroblasts, J. Inherit. Metab. Dis. 11 (1988) 266–269.
7] S. Masuda, H. Watanabe, M. Morioka, Y. Fujita, T. Ageta, H. Kodama, Characteristics
of partially purified prolidase and prolinase from the human prostate, Acta Med.
Okayama 48 (1994) 173–179.
[
5
. Conclusions
This research has revealed that putative L. plantarum WCFS1 PepR1
[8] J. Butterworth, D. Priestman, Fluorimetric assay for prolinase and partial character-
ization in cultured skin fibroblasts, Clin. Chim. Acta 122 (1982) 51–60.
and PepR2 are truly prolinase-cleaving iminodipeptides with N-
terminal proline. These isozymes may have application in fermented
foods manufacture for debittering. The fundamental information
[
9] W. Wang, G. Liu, K. Yamashita, M. Manabe, H. Kodama, Characteristics of prolinase
against various iminodipeptides in erythrocyte lysates from a normal human and
a patient with prolidase deficiency, Clin. Chem. Lab. Med. 42 (2004) 1102–1108.
Fig. 3. Phylogenetic tree of prolinases. Lactobacillus plantarum WCFS1 PepR1 and PepR2 are indicated by shaded names on the tree. The phylogenetic tree was generated on the BLASTP
server. (Two-column fitting).

100
Y. Huang, T. Tanaka / Biochimica et Biophysica Acta 1854 (2015) 91–100
[
10] H. Kodama, T. Ohhashi, C. Ohba, T. Ohno, J. Arata, I. Kubonishi, I. Miyoshi, Character-
istics and partial purification of prolidase and prolinase from leukocytes of a normal
human and a patient with prolidase deficiency, Clin. Chim. Acta 180 (1989) 65–72.
11] R.E. Neuman, E.L. Smith, Synthesis of proline and hydroxyproline peptides; their
cleavage by prolinase, J. Biol. Chem. 193 (1951) 97–111.
12] S. Sarid, A. Berger, E. Katchalski, Proline iminopeptidase II. Purification and compar-
ison with iminodipeptidase (prolinase), J. Biol. Chem. 237 (1962) 2207–2212.
13] M. Kleerebezem, J. Boekhorst, R. van Kranenburg, D. Molenaar, O.P. Kuipers, R. Leer,
R. Tarchini, S.A. Peters, H.M. Sandbrink, M.W.E.J. Fiers, W. Stiekema, R.M.K.
Lankhorst, P.A. Bron, S.M. Hoffer, M.N.N. Groot, R. Kerkhoven, M. de Vries, B.r.
Ursing, W.M. de Vos, R.J. Siezen, Complete genome sequence of Lactobacillus
plantarum WCFS1, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 1990–1995.
14] E. Doi, D. Shibata, T. Motoba, Modified colorimetric ninhydrin methods for peptidase
assay, Anal. Biochem. 118 (1981) 173–184.
15] M. Sakoda, K. Hiromi, Determination of the best-fit values of kinetic parameters of
the Michaelis–Menten equation by the method of least squares with the Taylor
expansion, J. Biochem. (Tokyo) 80 (1976) 547–555.
16] G. Bohm, R. Muhr, R. Jaenicke, Quantitative analysis of protein far UV circular dichro-
ism spectra by neutral networks, Protein Eng. 5 (1992) 191–195.
17] N.J. Greenfield, Using circular dichroism collected as a function of temperature to
determine the thermodynamics of protein unfolding and binding interactions, NIH
Public Access 1 (2006) 2527–2535.
18] P.A. Bates, M.J.E. Sternberg, Model building by comparison at CASP3: using expert
knowledge and computer automation, Proteins Struct. Funct. Genet. 3 (1999)
[22] A. Roy, A. Kucukural, Y. Zhang, I-TASSER: a unified platform for automated protein
structure and function prediction, Nat. Protoc. 5 (2010) 725–738.
[23] A. Roy, J. Yang, Y. Zhang, Cofactor: an accurate comparative algorithm for structure-
based protein function annotation, Nucleic Acids Res. 40 (2012) 471–477.
[24] D. Huson, C. Scornavacca, Dendroscope 3: an interactive tool for rooted phylogenetic
trees and networks, Syst. Biol. 61 (2012) 1061–1067.
[25] W.I. Goldburg, Dynamic light scattering, Am. J. Phys. 67 (1999) 1156–1160.
[26] R.E. Tanner, B. Herpigny, S.H. Chen, C.K. Rha, Conformational change of protein sodi-
um dodecylsulfate complexes in solution: a study of dynamic light scattering, J.
Chem. Phys. 76 (1982) 3866–3869.
[
[
[
[27] M. Liu, J.R. Bayjanov, B. Renckens, A. Nauta, R. Siezen, The proteolytic system of lactic
acid bacteria revisited:
a genomic comparison, BMC Genomics 11 (2010)
[
[
1471–2164.
[28] P. Varmanen, J. Steele, A. Palva, Characterization of a prolinase gene and its product
and an adjacent ABC transporter gene from Lactobacillus helveticus, Microbiology
142 (1996) 809–816.
[29] N.D. Rawlings, A.J. Barrett, Evolutionary families of peptidases, Biochem. J. 290
(1993) 205–218.
[30] N.D. Rawlings, F.R. Morton, C.Y. Kok, J. Kong, A.J. Barrett, MEROPS: the peptidase
database, Nucleic Acids Res. 36 (2008) 320–325.
[31] W. Shao, G.U. Yuksel, E.G. Dudley, K.L. Parkin, J.L. Steele, Biochemical and molecular
characterization of PepR, a dipeptidase, from Lactobacillus helveticus CNRZ32, Appl.
Environ. Microbiol. 63 (1997) 3438–3443.
[32] P. Varmanen, T. Rantanen, A. Palva, S. Tynkkynen, Cloning and characterization of a
prolinase gene (pepR) from Lactobacillus rhamnosus, Appl. Environ. Microbiol. 64
(1998) 1831–1836.
[33] S. Vijayakumar, S. Vishveshwara, G. Ravishanker, D.L. Beveridge, Differential stability
of β-sheets and α-helices in β-lactamase: a high temperature molecular dynamics
study of unfolding intermediates, Biophys. J. 65 (1993) 2304–2312.
[34] A. Perczel, Z. Gaspari, I.G. Csizmadia, Structure and stability of β-pleated sheets, J.
Comput. Chem. 26 (2005) 1155–1168.
[
[
[
[
47–54.
19] P.A. Bates, L.A. Kelley, R.M. MacCallum, M.J.E. Sternberg, Enhancement of protein
modeling by human intervention in applying the automatic programs 3D-JIGSAW
and 3D-PSSM, Proteins Struct. Funct. Genet. 5 (2001) 39–46.
20] P.A. Bates, B. Contreras-Moreira, Domain fishing: a first step in protein comparative
modeling, Bioinformatics 18 (2002) 1141–1142.
[
[
21] Y. Zhang, I-TASSER server for protein 3D structure prediction, BMC Bioinform. 9
(2008) 40.