A Novel Glutamyl (Aspartyl)-Specific Aminopeptidase A from Lactobacillus delbrueckii with Promising Properties for Application

Lactic acid bacteria (LAB) are auxotrophic for a number of amino acids. Thus, LAB have one of the strongest proteolytic systems to acquit their amino acid requirements. One of the intracellular exopeptidases present in LAB is the glutamyl (aspartyl) specific aminopeptidase (PepA; EC 3.4.11.7). Most of the PepA enzymes characterized yet, belonged to Lactococcus lactis sp., but no PepA from a Lactobacillus sp. has been characterized so far. In this study, we cloned a putative pepA gene from Lb. delbrueckii ssp. lactis DSM 20072 and characterized it after purification. For comparison, we also cloned, purified and characterized PepA from Lc. lactis ssp. lactis DSM 20481. Due to the low homology between both enzymes (30%), differences between the biochemical characteristics were very likely. This was confirmed, for example, by the more acidic optimum pH value of 6.0 for Lb-PepA compared to pH 8.0 for Lc-PepA. In addition, although the optimum temperature is quite similar for both enzymes (Lb-PepA: 60°C; Lc-PepA: 65°C), the temperature stability after three days, 20°C below the optimum temperature, was higher for Lb-PepA (60% residual activity) than for Lc-PepA (2% residual activity). EDTA inhibited both enzymes and the strongest activation was found for CoCl2, indicating that both enzymes are metallopeptidases. In contrast to Lc-PepA, disulfide bond-reducing agents such as dithiothreitol did not inhibit Lb-PepA. Finally, Lb-PepA was not product-inhibited by L-Glu, whereas Lc-PepA showed an inhibition.


Introduction
Lactic acid bacteria (LAB) are a heterogeneous group of microorganisms which have a common metabolic property: the production of lactic acid as the majority end-product from the fermentation of carbohydrates [1,2]. The LAB are commonly Gram-positive, aerobic to facultative anaerobic, asporogenous roods and cocci, which are oxidase, catalase and benzidine Lc. lactis ssp. cremoris MG1316 (UniProt ID: Q48677; reviewed) and Lc. lactis ssp. lactis IL1403 (UniProt ID: Q9CIH3; unreviewed), respectively. Due to the low homology between the known PepA from Lc. lactis sp. and the heretofore unknown PepA from Lb. delbrueckii, it is very likely that the biochemical characteristics are quite different. Thus, the aim of the current study was the cloning and heterologous recombinant production of the PepA from Lb. delbrueckii ssp. lactis DSM 20072 and its biochemical characterization. Furthermore, the characteristics were compared directly with the characteristics of PepA from Lc. lactis ssp. lactis DSM 20481, which was also heterologously recombinantly produced in Escherichia coli BL21(DE3).

Materials and Methods
Chemicals, enzymes, kits, materials and devices All chemicals were of analytical grade and purchased from Sigma Aldrich (Taufkirchen, Germany), Carl Roth GmbH (Karlsruhe, Germany) or Applichem (Darmstadt, Germany). Hexokinase/glucose-6-phosphate dehydrogenase was purchased from Megazyme International Ireland (Wicklow, Irland) and was used for the glucose concentration determination assay based on the commercial D-glucose/D-fructose test kit from R-Biopharm AG (Darmstadt, Germany; product code 10 139 106 035). Chromogenic peptides were obtained from Bachem AG (Bubendorf, Switzerland). Molecular weight markers were bought from New England Biolabs (NEB; Frankfurt, Germany) and GE Healthcare (München, Germany). The enzymes required for molecular biological work were purchased from NEB (Frankfurt, Germany), Qiagen (Hilden, Germany), Thermo Scientific (Schwerte, Germany) or Roche Applied Science (Penzberg, Germany). Kits for molecular biological work were obtained from Thermo Scientific (Schwerte, Germany) or Qiagen (Hilden, Germany). Agarose was bought from SERVA Electrophoresis GmbH (Heidelberg, Germany). PD-10 columns were obtained from GE Healthcare (München, Germany). The bioreactor cultivation was realized using the Multifors system (Infors AG, Bottmingen/Basel, Switzerland). The MINI-PROTEAN system (Bio-Rad Laboratories GmbH, München, Germany) was used for polyacrylamide gel electrophoresis. The ÄKTA-FPLC system (GE Healthcare, München, Germany) equipped with a Ni-NTA column (Cube Biotech GmbH, Monheim, Germany) was used for protein purification.

Bacterial strains
Lactobacillus delbrueckii ssp. lactis DSM 20072 was cultivated in de Man, Rogosa and Sharpe (MRS) medium [23] with constant shaking at 37°C. Lactococcus lactis ssp. lactis DSM 20481 was cultivated as described previously [10]. Escherichia coli XL1 Blue (Merck, Darmstadt, Germany) and E. coli BL21(DE3) (Novagen, Madison, USA) were used as the hosts for the cloned polymerase chain reaction (PCR) products and T7 expression work, respectively. Standard protocols were employed for the preparation and transformation of competent E. coli cells with plasmid DNA via heat shock [24]. The E. coli cells were cultivated as described previously [8,10].

Cloning, construction of expression vectors and sequencing
Total genomic DNA from either Lb. delbrueckii or Lc. lactis was extracted using an identical method to that described previously [10]. The PCR was performed using HotStar HiFidelity polymerase (Qiagen), according to the manufacturer's instructions. All primers used in this study were synthesized by biomers.net GmbH (Ulm, Germany). The primers Lb_pepA_for (5´-GTGA CGAACATATGGAAAAAGCCGCTGAAATTC-3´; NdeI restriction site is underlined) and Lb_pepA_rev (5´-CGGAGCTCGAGATTAAAGCTTTTAAAGGATTCCAGCTTTTC-3´; XhoI restriction site is underlined) were used for the amplification of the pepA gene from Lb. delbrueckii ssp. lactis DSM 20072 (UniProt ID: F0HXE4; EMBL: EGD26747). The primers Lc_pepA_for (5´-GCCGCCGCATATGGAACTATTCGACAAAG-3´; NdeI restriction site is underlined) and Lc_pepA_rev (5´-CGGAGCCGCTCGAGATAGTTTTTAATTTCAGCTA C-3´; XhoI restriction site is underlined) were used for the amplification of the pepA gene from Lc. lactis ssp. lactis DSM 20481. The last two primers were designed based on the nucleotide sequence of the pepA gene from Lc. lactis ssp. lactis (strain IL1403; UniProt ID: Q9CIH3; EMBL: AAK04485).
The PCR products obtained (about 1100 bp) of the Lb-pepA gene (1086 bp) and the Lc-pepA gene (1068 bp) were purified using the QIAquick PCR Purification kit (Qiagen) according to the manufacturer's instructions. The purified PCR products and the vector pET20b(+) (Novagen) were digested using the restriction enzymes NdeI and XhoI. T4-DNA-ligase was used for ligation of the digested PCR products and vector and resulted in the plasmids pET20b (+)_Lb-pepA and pET20b(+)_Lc-pepA, respectively. Both vectors were individually transformed into competent E. coli XL1 Blue cells via heat shock and plated on LB amp agar plates (tryptone: 10 g L -1 , yeast extract: 5 g L -1 , NaCl: 5 g L -1 , ampicillin: 100 μg mL -1 ). After cultivation overnight at 37°C, single colonies were picked and cultivated in 5 mL LB amp medium overnight at 37°C. The plasmids were isolated using the GeneJET Plasmid Miniprep kit (Fermentas), according to the manufacturer's instructions. The plasmids obtained were used for sequencing (SRD-Scientific Research and Development GmbH; Bad Homburg, Germany). Database searches and alignments were performed online with the programs blastn and blastp provided by the BLAST server [25,26]. All parameters were set at their default values.
Expression of recombinant PepA in E. coli BL21(DE3) Transformed E. coli BL21(DE3) strains were grown in 2xYT medium (tryptone: 16 g L -1 , yeast extract: 10 g L -1 , NaCl: 5 g L -1 ) that contained glucose (10 g L -1 ) supplemented with ampicillin (100 μg mL -1 ). Pre-cultures were incubated at 37°C on a rotary shaker. The first (5 mL) and the second pre-culture (50 mL) were each cultivated for 15 h. The main cultures (600 mL) were grown in a parallel bioreactor system (Multifors) and inoculated with 10% (v/v) of the particular pre-culture. The pH value of the bioreactor cultivation was kept at pH 7.0 by using 2 M NaOH and 2 M H 3 PO 4 . The O 2 concentration (pO 2 ) dissolved in the medium was maintained above 30% saturation by regulation of the stirrer speed (500-1000 rpm). The aeration rate was 1 vvm. Samples were taken during the cultivation to analyze the optical density (OD 600nm ), the bio dry mass (BDM) and the glucose concentration, as described previously [10]. When the OD 600nm value reached 5, the temperature was maintained at 30°C to minimize the formation of inclusion bodies, and recombinant protein expression was induced by the addition of 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). The cultures were harvested after 11 h of cultivation, as described previously [10] and then stored at -20°C.
Samples (10 mL) were taken at various time points during the cultivations, and the PepA activity (see below) was determined from the cell-free extract after cell disruption, centrifugation (8000 x g, 10 min, 4°C) and filtration (0.45 μm). Consequently, the samples were centrifuged (see above) and the cell-pellets were suspended in a 0.9% (w/v) NaCl-solution (5 mL). After centrifugation (see above), the cell-pellets were suspended in Na/K-phosphate (50 mM; pH 6.0 for Lb-PepA or pH 8.0 for Lc-PepA). Subsequently, the cell disruption was realized by sonification (UP200S ultrasonic processor, Dr. Hielscher, Berlin, Germany; 20 cycles containing 1 min disruption, 1 min break) on ice.

Polyacrylamide gel electrophoresis (PAGE)
Samples were divided into soluble and insoluble fractions after cell disruption. These fractions and purified PepA (also soluble) were analyzed by sodium dodecyl sulfate (SDS) PAGE (12.5% gel) [27]. An amount of 5 μg protein [28] or 7.5 μg protein [29] was applied to gel in the case of the soluble or insoluble fractions, respectively. Bovine serum albumin was used as a standard for both protein determination methods [28,29]. A commercial molecular weight protein mixture was used as a reference for molecular weight estimation (NEB; 2-212 kDa). Gels were stained with Coomassie Brilliant Blue for protein detection.

Size exclusion chromatography (SEC)
The native molecular mass determination of the purified enzymes by SEC was realized using an ÄKTA-FPLC system (GE Healthcare) equipped with a Superdex TM 200 10/300 GL column (GE Healthcare). The injection volume was 30 μL (1 mg Protein mL -1 ) and the flow rate was 0.75 mL min -1 using Na/K-phosphate buffer (10 mM; pH 6.0 for Lb-PepA or pH 8.0, for Lc-PepA) containing NaCl (150 mM) as the eluent. Eluted protein was detected at 280 nm and fractions (0.5 mL) were taken. Standard proteins (Gel Filtration HMW and LMW Calibration Kit, GE Healthcare) were used as references for molecular mass determination.

Standard PepA enzyme activity assay
PepA activity in the standard assay was determined with H-Asp-pNA as a substrate. The standard assay was performed as follows: Initially, 25 μL enzyme solution was added to 192.5 μL Na/K-phosphate buffer (50 mM; pH 6.0 for Lb-PepA or pH 8.0 for Lc-PepA). Additionally, 10 μL of a CoCl 2 stock solution (15 mM for Lb-PepA or 7.5 mM for Lc-PepA) was added. After incubation for 10 min at the required temperature (60°C for Lb-PepA or 65°C for Lc-PepA), 12.5 μL of the substrate solution (5 mg mL DMSO -1 ) was added to the reaction mixture. The reaction was terminated by adding 50 μL acetic acid (50% (v/v) to the sample. After centrifugation (8000 x g, 5 min, 4°C), 240 μL of the supernatant was transferred into a microtiter plate and the absorption was measured (Multiskan FC, Thermo Scientific, Braunschweig, Germany) at 405 nm. One katal (kat) of PepA activity was defined as the release of 1 mol p-nitroanilin per s. The specific activity of a particular sample was determined by dividing the volumetric activity by the corresponding protein content [28]. Thus, the specific activity during the bioreactor cultivation is referred to the protein content in the supernatant after cell disruption. All other specific activity values are related to the enzymes after purification.

Characterization of Lb-PepA and Lc-PepA
The purified Lb-PepA and Lc-PepA were characterized. The standard assay with H-Asp-pNA as a substrate was used unless stated otherwise.
Influence of temperature and pH on the initial PepA activity. In contrast to the standard assay, the temperature varied between 10 and 80°C. The pH was varied in the range between pH 4.5 and 10.0 (depending on the enzyme tested) for determination of the pHdependent effect. All buffers had a concentration of 50 mM. The following buffers were used for Lb-PepA: Na-acetate/acetic acid (pH 5.0-6.0), Na/K-phosphate (pH 5.5-8.0) and Bis-Trispropane/HCl (pH 6.0-7.0). The following buffers were tested for Lc-PepA: Na/K-phosphate (pH 6.0-8.0), Tris/HCl (pH 7.5-8.5) and Glycine/NaOH (pH 8.5-10.0).
Temperature stability of PepA. The enzyme preparations were incubated at 0, 40, 50 and 60°C (Lb-PepA) or at 0, 45, 55 and 65°C (Lc-PepA) for up to three days for the temperature stability and samples were taken several times. Sodium azide (0.1% (w/v)) was added to prevent microbial growth.
Storage stability of PepA. The storage stability of PepA was determined for four different storage conditions: (i) aliquotes (20 μL) of the enzyme solutions were stored at -80°C; (ii) aliquotes (20 μL) were lyophilized and stored in a desiccator at 20°C; (iii) aliquotes (20 μL) were lyophilized and stored at -80°C; and (iv) aliquotes (300 μL) were stored at -80°C, but were frozen again after thawing. Several samples were taken during the storage time (two months) and the PepA activity was measured.
Influence of metal ions on the PepA activity. The influence of different metal ions (CoCl 2 , MnCl 2 , ZnCl 2 ) on the activity of purified PepA (1.2 ± 0.3 mg Protein mL -1 ) was analyzed. In addition, apo-PepA was prepared by treating purified PepA with 20 mM ethylenediaminetetraacetic acid (EDTA), followed by dialysis in Na/K-phosphate buffer (50 mM; pH 6.0 for Lb-PepA or pH 8.0 for Lc-PepA). Subsequently, the reactivation of apo-PepA was tested for CoCl 2 , MnCl 2 , ZnCl 2 , CaCl 2 and MgCl 2 . In contrast to the standard assay, the concentration of these substances in the final assay varied between 0.04 and 5 mM.
Influence of organic solvents, inhibitors and other reagents. The substances tested were dissolved in H 2 O dd , DMSO, acetone or ethanol, depending on the substance. The assay conditions were identical to the standard assay, except that 10 μL of the test substance and 182.5 μL buffer were used, instead of 192.5 μL buffer. The concentration of the inhibitors, metal chelators, reducing agents and other substances in the final assay varied between 0.001 and 10 mM.
Determination of the substrate specificity of PepA. The substrate specificity of PepA was determined with different chromogenic substrates. The following pNA-derivates were used in a concentration of 5 mg mL DMSO Determination of product inhibition of PepA. The product inhibition of PepA was tested for the single amino acids L-Asp and L-Glu in a final concentration of 0.1, 1 and 10 mM. The assay conditions were identical to the standard assay, except that 10 μL of the particular amino acid solutions and 182.5 μL buffer were used, instead of 192.5 μL buffer.
Determination of apparent kinetic parameters of PepA. The apparent kinetic parameters of PepA were determined using H-Asp-pNA and H-Glu-pNA as a substrate. Standard PepA activity assay conditions were used in which the final substrate concentration ranged from 0.06-18 mM, depending on the particular enzyme and substrate. In addition, the apparent kinetic parameters of reactivated apo-Lb-PepA were investigated. In contrast to the standard assay, a particular metal salt stock solution was added to apo-Lb-PepA to gain the optimal metal concentration, as determined previously. The results were plotted according to Michaelis-Menten and the apparent kinetic parameters were calculated by nonlinear regression fitting using SigmaPlot 12.5 (Systat Software, Inc., San Jose, CA).

Statistical analysis
Standard deviation was used for data evaluation and calculated with Excel (Microsoft, Redmond, USA). All experiments were conducted at least in duplicate, with three independent measurements. The standard deviation was always below 5%.

Results
In this study, the heretofore unknown PepA from Lb. delbrueckii ssp. lactis DSM 20072 was produced recombinantly and compared to the biochemical characteristics of PepA from Lc. lactis ssp. lactis DSM 20481. Due to the low homology between the two enzymes (see below), it is very likely that the biochemical characteristics will be different.

Cloning and sequencing of Lb-pepA and Lc-pepA
The Lb-PepA expression vector (pET20b(+)_Lb-pepA) and the Lc-PepA expression vector (pET20b(+)_Lc-pepA) were constructed, sequenced and used for individual expression in the E. coli BL21(DE3) host strain. Due to the cloning strategy chosen, the proteins produced contained a C-terminal His 6 -tag. The nucleotide sequence of both the Lb-pepA gene (source: Lb. delbrueckii ssp. lactis DSM 20072) and the Lc-pepA gene (source: Lc. lactis ssp. lactis DSM 20481) obtained in this study exhibited 100% identity to sequences of pepA genes from Lb. delbrueckii ssp. lactis DSM 20072 (UniProt ID: F0HXE4; EMBL: EGD26747) and Lc. lactis ssp. lactis CV56 (UniProt ID: F2HIS5; EMBL: ADZ63034), respectively, deposited previously. By comparison, an identity of about 30% was ascertained for the amino acid sequences of both PepA enzymes used in this study.
Heterologous expression of Lb-PepA and Lc-PepA in E. coli Both Lb-PepA and Lc-PepA were individually expressed in soluble form using the expression host E. coli BL21(DE3) under identical cultivation conditions (Fig 1). During the cultivation of the recombinant E. coli for Lb-PepA production, the glucose was completely consumed after approximately 9 h when the cells entered the stationary growth phase (Fig 1A). The OD 600 nm value increased up to 36 during cultivation (corresponding to a cell dry weight of 10.3 g L -1 ), and the maximum volumetric Lb-PepA activity was achieved after 8 h of cultivation with about 90 μkat H-Asp-pNA L Culture -1 (specific Lb-PepA activity: 92 nkat H-Asp-pNA mg Protein -1 ). The Lb-PepA activity decreased during the stationary growth phase, which was caused by degradation of the enzyme, as seen on the SDS-PAGE analysis (data not shown).
In the case of Lc-PepA production, a comparable maximum OD 600 nm value of 33 was achieved with the E. coli expression host (Fig 1B). This corresponds to a cell dry weight of 9.4 g L -1 . Again, the glucose was consumed after approximately 9 h by entering the stationary growth phase. The maximum volumetric Lc-PepA activity (260 μkat H-Asp-pNA L Culture Purification of Lb-PepA and Lc-PepA and molecular mass determination The His 6 -tagged enzymes Lb-PepA and Lc-PepA were both individually purified using a FPLC procedure based on Ni 2+ -NTA chromatography resin. An enzymatic activity yield of approximately 30% was achieved for each enzyme, whereas a purification factor of 2.8 and 7.1 was determined for Lb-PepA and Lc-PepA, respectively. The purity and the molecular mass of the monomers were determined by SDS-PAGE (Fig 2). The molecular mass of the Lb-PepA monomer was determined at approximately 41 kDa (Fig 2, lane 1 and 2). This is in accordance with the theoretical molecular mass of 41.1 kDa (based on the amino acid sequence including the His 6 -tag). A molecular mass of approximately 40 kDa was determined for Lc-PepA (Fig 2, lane  3 and 4). The theoretical molecular mass of Lc-PepA (including the His 6 -tag) is 39.4 kDa. Size exclusion chromatography experiments were performed to analyze the native mass of the enzymes, and a molecular mass of 508 kDa and 470 kDa was determined for Lb-PepA and Lc-PepA, respectively. By taking the molecular mass of the monomers into account, it is suggested that both enzymes are homo dodecamers.
Effect of temperature and pH on the initial PepA activity At first, the influence of the temperature on the initial PepA activity was determined. As shown in Fig 3A, the optimum temperature for Lb-PepA was determined at 60°C (100% = 144 nkat H-Asp-pNA mL -1 ). At a higher temperature (65°C), 97% of the maximum activity was achieved. Almost no activity (0.8%) was detected at 75°C. The highest activity for Lc-PepA was determined at 65°C (100% = 1345 nkat H-Asp-pNA mL -1 ; Fig 3B). A minor lower activity (98%) was detected at 60°C, and a residual activity of 9% was measured at 75°C. Thus, the direct comparison of both enzymes showed a similar optimum temperature for the initial enzyme activity, whereas for Lc-PepA it was slightly higher.
Secondly, the optimum pH of the two different PepA enzymes was determined using different buffers (all 50 mM; Fig 3C and 3D). The highest activity for Lb-PepA was detected using Na/K-phosphate buffer (pH 6; 100% = 73.2 nkat H-Asp-pNA mL -1 ; Fig 3C). At the same pH value, but using Na-acetate/acetic acid buffer or Bis-Tris-propane/HCl buffer, the activity was 85% and 40%, respectively. Almost no Lb-PepA activity (2%) was determined at a pH value of 8.0. These results differ from the pH profile of Lc-PepA. The highest activity for Lc-PepA was determined at pH 8.0 using Na/K-phosphate buffer (100% = 1691 nkat H-Asp-pNA mL -1 ; Fig 3D). At a pH value of 6.0, the optimum of Lb-PepA, Lc-PepA had a residual activity of 50%. At a pH value of 10.0, using glycine/NaOH buffer, the residual activity of Lc-PepA was 1%.  The temperature stability of Lb-PepA and Lc-PepA was determined at 0°C, as well as at the optimum temperature and 10°C and 20°C below the optimum temperature (Fig 3E and 3F). At 0°C (on ice), both enzymes were stable over the analysis time (72 h) with a residual activity of 80% and 100% for Lb-PepA and Lc-PepA, respectively. At their optimum temperatures, the stability of both enzymes was quite low and almost no activity was detectable after 24 h. In the case of Lb-PepA, the temperature stability at 10°C and 20°C below the optimum temperature was better compared to Lc-PepA. At 10°C below the optimum temperature, Lb-PepA had a residual activity of 14% after 72 h, whereas Lc-PepA showed almost no activity (0.3%). At 20°C below the optimum temperatures, the difference in stability between both enzymes was higher. After 72 h, Lb-PepA showed a residual activity of 60%, whereas Lc-PepA had only 2% of activity remaining. In addition, the temperature stability of Lb-PepA and Lc-PepA was analyzed in the presence of the optimal concentration of CoCl 2 (see below) during the incubation at the particular temperatures. As a result, no stabilizing effect of the metal ions was determined (data not shown).
The storage stability for both PepA enzymes was tested under four different conditions ((i)-(iv), see Material and Methods for details). The storage at -80°C showed good results and a residual activity of 99% and 90% was determined for Lb-PepA and Lc-PepA, respectively, after two months. The storage stability after lyophilization and subsequent storage at 20°C in a desiccator was the lowest of all four storage conditions tested. The residual activity after two months was 62% and 20% for Lb-PepA and Lc-PepA, respectively. By contrast, the storage of the lyophilized enzyme preparations at -80°C resulted in a residual activity for Lb-PepA and Lc-PepA of 100%. Finally, the influence of freezing and thawing on the particular PepA activity was investigated. The enzyme solutions were frozen and thawed six times during the storage time of two months and showed a residual activity of 100% and 75% for Lb-PepA and Lc-PepA, respectively, at the end of the period of time.

Influence of metal ions on the PepA activity
The influence of different metal ions on the particular PepA activity was investigated (Fig 4A  and 4B). All metal salts were used as chlorides to prevent an influence of the anion. Without any added metal salt, the activity of Lb-PepA was 0.2 nkat H-Asp-pNA mL -1 (0.12% of the overall maximum activity; Fig 4A). The highest Lb-PepA activity (100% = 165 nkat H-Asp-pNA mL -1 ) overall was obtained with 0.625 mM CoCl 2 added. In the case of ZnCl 2 , an addition of 0.625 mM also resulted in the highest Lb-PepA activity, but the activity was only 1 nkat H-Asp-pNA mL -1 . The highest activity (5.36 nkat H-Asp-pNA mL -1 ) for MnCl 2 was determined for an added concentration of 1.25 mM. The Lb-PepA activity decreased for all three metal salts tested with concentrations above the optimum concentration determined. The activity of Lc-PepA without any added metal salt was 24.1 nkat H-Asp-pNA mL -1 (1.35% of the overall maximum activity; Fig  4B). The overall highest Lc-PepA activity (100% = 1785 nkat H-Asp-pNA mL -1 ) was obtained with 0.3125 mM CoCl 2 added. The highest Lc-PepA activity (47.4 nkat H-Asp-pNA mL -1 ) for MnCl 2 was also determined for an added concentration of 0.3125 mM. By contrast, the highest Lc-PepA activity (55.1 nkat H-Asp-pNA mL -1 ) was obtained with an added ZnCl 2 concentration of 5 mM. In contrast to Lb-PepA activity, the Lc-PepA activity decreased only slightly or was almost constant with metal salt concentrations above the optimum concentrations determined.
In addition, the reactivating effect of apo-PepA by five different metal salts was investigated. After treating Lb-PepA and Lc-PepA with 20 mM EDTA (ethylenediaminetetraacetic acid) and subsequent dialysis, no PepA activity was determined. CoCl 2 , ZnCl 2 , MnCl 2 , CaCl 2 and MgCl 2 were then added individually up to final concentrations of 5 mM each. A reactivating effect of Lb-PepA and Lc-PepA was determined for CoCl 2 , ZnCl 2 and MnCl 2 , but none was determined for CaCl 2 and MgCl 2 .

Influence of organic solvents, inhibitors and other reagents on the PepA activity
The substrate and some of the reagents tested were dissolved in organic solvents prior to their addition to the assay due to their limited solubility in pure water ( Table 1). The pNA standard assay contained 5.2% (v/v) DMSO (dimethyl sulfoxide). The activity was measured in the presence of an additional 4.2% (v/v) of the particular solvent to determine the influence of each organic solvent on the PepA activity. The activity value after the addition of 4.2% (v/v) H 2 O dd was used as a reference (100%). Both the activity of Lb-PepA and Lc-PepA were reduced by additional DMSO (about 85% residual activity). However, all other solvents tested (ethanol, acetone, dimethylformamide (DMF)) reduced the particular PepA activity stronger than DMSO. Thus, DMSO is the most suitable organic solvent for substrates that are not soluble in water.
The PepA activity values, which were measured in the presence of additional 4.2% (v/v) water, acetone, DMSO or ethanol, were considered as 100% for the inhibition studies of the different substances ( Table 1). The addition of the cysteine peptidase inhibitor E64 showed no effect on Lb-PepA and a negligible effect on Lc-PepA. The same was observed for the carboxy peptidase inhibitor pepstatin A and the serine peptidase inhibitor PMSF (phenylmethylsulfonyl fluoride). The metallopeptidase inhibitor 1,10-phenantroline and the metal chelating reagent EDTA had a strong inactivating effect on both PepA enzymes. This indicates that Lb-PepA and Lc-PepA belong to the group of metallopeptidases. A difference between both enzymes was observed concerning the disulfide bond-reducing agents DTT (dithiothreitol) and β-mercaptoethanol. They had no effect on the Lb-PepA activity but inactivated Lc-PepA completely. This indicates that potential disulfide bonds are essential for the activity of Lc-PepA, but not for Lb-PepA.

Determination of the substrate specificity and product inhibition of PepA
The substrate specificity of Lb-PepA and Lc-PepA was analyzed using twelve single amino acid-pNA substrates (see Material and Methods for details). Both PepA enzymes exhibited activity only for the substrates H-Asp-pNA and H-Glu-pNA. In the case of Lb-PepA, the highest activity (100% = 118 nkat mL -1 ) was determined using H-Asp-pNA as a substrate. The activity was 2.33% using H-Glu-pNA as a substrate compared to the activity determined with H-Asp-pNA as a substrate. The highest activity (100% = 1218 nkat mL -1 ) for Lc-PepA was also Presented are the means of three independent measurements and the standard deviation was < 5%. doi:10.1371/journal.pone.0152139.t001 Novel PepA from Lactobacillus delbrueckii determined with H-Asp-pNA as a substrate. In contrast to Lb-PepA, the activity of Lc-PepA with H-Glu-pNA as a substrate was higher with 45.5%. The product inhibition of Lb-PepA and Lc-PepA was tested for the single amino acids L-Asp and L-Glu using either H-Asp-pNA or H-Glu-pNA as a substrate ( Table 2). None of the enzymes was inhibited by L-Asp up to a tested concentration of 1 mM independent of the substrate used for the PepA activity determination. Both enzymes were inhibited at an L-Asp concentration of 10 mM and showed a residual activity between 61% and 78%. There was a difference between both enzymes using the product L-Glu. The Lb-PepA was not inhibited in all cases, whereas Lc-PepA was inhibited in all combinations tested. The strongest inhibition of Lc-PepA (69% residual activity) was determined for a L-Glu concentration of 10 mM and H-Glu-pNA as a substrate.

Determination of apparent kinetic parameters of PepA
The apparent kinetic parameters of Lb-PepA and Lc-PepA (V max , K M and K IS ) were determined using either H-Asp-pNA or H-Glu-pNA as a substrate. The particular specific activities were plotted according to Michaelis-Menten (Fig 5) and the kinetic parameters (Table 3) were calculated by nonlinear regression fitting using SigmaPlot 12.5. A strong substrate inhibition was observed for Lb-PepA using the substrate H-Asp-pNA, as seen in Fig 5A. In the case of using H-Glu-pNA as a substrate (Fig 5C), no substrate inhibition was determined up to 7.3 mM. A higher concentration was not determinable due to a reduced solubility of the substrate. Based on the K M and V max values determined, it was shown clearly that the preferable substrate for Lb-PepA is H-Asp-pNA. A substrate inhibition was also determined for H-Asp-pNA as a substrate for Lc-PepA (Fig 5B). However, this substrate inhibition was less severe compared to the substrate inhibition of H-Asp-pNA for Lb-PepA. In contrast to Lb-PepA, a substrate inhibition was determined for Lc-PepA using H-Glu-pNA as a substrate (Fig 5D). This substrate inhibition was stronger for Lc-PepA than that using H-Asp-pNA as a substrate. The most preferable substrate for Lc-PepA was also H-Asp-pNA due to the comparable K M values and the higher V max and K IS values. Presented are the means of three independent measurements and the standard deviation was < 5%. doi:10.1371/journal.pone.0152139.t002 Finally, the apparent kinetic parameters of reactivated apo-Lb-PepA were determined using H-Asp-pNA and H-Glu-pNA as a substrate (Table 4). No differences were observed concerning the K M and K IS values between Lb-PepA (see above) and apo-Lb-PepA reactivated with CoCl 2 . The K M and K IS values determined for each substrate used were in the same range, independent of the metal salt used for reactivation of the apo-Lb-PepA. A notable difference was observed concerning the V max values. The highest activity was always determined using CoCl 2 for reactivation, followed by MnCl 2 and ZnCl 2 . In the case of using ZnCl 2 for reactivation apo-Lb-PepA and H-Glu-pNA as a substrate, the Lb-PepA activity determined was too low to confidently evaluate the kinetic parameters.

Structural comparison of PepA
The metal binding site of PepA from Streptococcus pneumonia R6 (Sc-PepA) was described by Kim et al. [22]. The authors stated in this article that Sc-PepA belongs to the M42 family of peptidases and exhibits a dodecameric structure. Although the presence of the zinc ions in the crystal was not confirmed, they modeled two zinc ions in the active site based on the electron densities of its crystal diffraction. His66 and Asp236 coordinated one of the zinc ions, whereas Glu214 and His318 coordinated the second zinc ion. Asp181 coordinated both zinc ions. The substrate binding pocket itself was constructed by Asp236, Ser238, Leu255, Arg257, Thr309, and Gly311. The authors stated that Arg257 is notable because the position of the Arg257 side chain creates a positive patch in the S1 pocket and, therefore, the positive patch of Sc-PepA appears to be responsible for its specificity towards acidic amino acids in the S1 position.
Due to the well conserved metal coordinating residues in the active site [22], novel PepA enzymes can be found, and a comparison to the known PepA enzymes from Lc. lactis sp. and the novel PepA from Lb. delbrueckii is shown in Table 5. Although the gene and protein homology is quite low compared to other PepA, the position of the catalytically important residues appears similar. This indicates that the different enzymes show the same catalytic activity, but, due to the low homology, other enzyme characteristics, such as optimum conditions and stability, could differ among the PepA enzymes.

Comparison of biochemical characteristics of different PepA
In 1987, Exterkate and De Veer [11] purified and characterized the first PepA of a LAB, called Streptococcus cremoris HP, which is now called Lc. lactis ssp. cremoris HP. Four years later, Niven [19] described the characteristics of PepA from Lc. lactis ssp. lactis NCDO 712, followed by Bacon et al. [20], who examined the PepA from Lc. lactis ssp. cremoris AM2. All these PepA enzymes belonged to microorganisms of the genus Lactococcus. In the current study, the first PepA from a microorganism of the genus Lactobacillus is described. Some selected characteristics of PepA from different microorganisms are shown in Table 6. The optimum temperature of all PepA described varied between 50 and 65°C. The optimum pH for all lactococcal PepA was between 8.0 and 8.3, and only the novel PepA from Lb. delbrueckii had an acidic pH The pepA gene/PepA protein from Lc. lactis ssp. cremoris MG1363 was used as a reference. 2 The pepA gene is identical to the pepA gene from Lc. lactis ssp. lactis DSM 20481 used in this study. 3 Automated UniRule annotation. 4 Experimental evidence [22]. 5 By similarity. doi:10.1371/journal.pone.0152139.t005 Novel PepA from Lactobacillus delbrueckii optimum (pH 6.0). In agreement with our results, the lactococcal PepA was inhibited by DTT, but not the PepA from Lb. delbrueckii. To the best of our knowledge, no crystal structures are available yet, neither for Lb-PepA nor Lc-PepA. Thus, further research is needed to explain the difference concerning the inhibition by disulfide bond-reducing agents.
Potential application of PepA for food protein hydrolysis using an enzyme membrane reactor As mentioned in the introduction, PepA can probably be used for the production of flavoring hydrolysates out of glutamyl/aspartyl-rich food proteins. Nowadays, the industrial biotransformation of proteins is mainly performed in discontinuous batch processes [30,31]. A promising alternative to the batch processes are continuous biotransformations using an enzyme membrane reactor [30,32]. Using this process approach, the enzymes are located in a reaction space, entrapped by a membrane and, thus, can be reused, which is a remarkable economic benefit [30,33]. Due to the high molecular mass of the PepA enzymes (about 480 kDa), they will not penetrate the membrane used (normally 1-10 kDa). The sufficient temperature stability, especially of Lb-PepA, also makes this enzyme interesting for application in an EMR.
In conclusion, the gene sequence for the PepA from Lb. delbrueckii ssp. lactis DSM 20072, heretofore described as putative, was heterologously expressed in E. coli and the recombinant protein showed the enzyme activity desired. Thus, the gene/function relationship was proven and the Lb-PepA produced was characterized biochemically for the first time. Most of the characteristics determined were different to the PepA enzymes from Lactococcus sp. described heretofore. The more acid optimum pH value of Lb-PepA (pH 6.0) compared to Lc-PepA (pH 8.0) makes the Lb-PepA especially interesting for food protein hydrolysis, because food generally has an acid or neutral pH value. In addition, Lb-PepA showed higher temperature stability and was not product-inhibited by L-Glu. Due to the fact that the Lb-PepA was characterized using synthetic pNA-substrates, as is common for this enzyme class in the literature, a closing statement about its activity against original peptide substrates cannot be made at this point and was not in the focus of this fundamental study. The activity against original peptide substrates and the application of Lb-PepA in food protein hydrolysis will be presented in a further study.