Cloning and Expression of an Oligopeptidase, PepO, with Novel Specificity from Lactobacillus rhamnosus HN001 (DR20)

Abstract

Oligopeptidases of starter and nonstarter lactic acid bacteria contribute to the proteolytic events important in maturation and flavor development processes in cheese. This paper describes the molecular cloning, expression, and specificity of the oligopeptidase PepO from the probiotic nonstarter strain Lactobacillus rhamnosus HN001 (DR20). The pepO gene encodes a protein of 70.9 kDa, whose primary sequence includes the HEXXH motif present in certain classes of metallo-oligopeptidases. The pepO gene was cloned in L. rhamnosus HN001 and overexpressed in pTRKH2 from its own promoter, which was mapped by primer extension. It was further cloned in both pNZ8020 and pNZ8037 and overexpressed in Lactococcus lactis subsp. cremoris NZ9000 from the nisA promoter. The purified PepO enzyme demonstrated unique cleavage specificity for α_s1-casein fragment 1–23, hydrolyzing the bonds Pro-5-Ile-6, Lys-7-His-8, His-8-Gln-9, and Gln-9-Gly-10. The impact of this enzyme in cheese can now be assessed.

Lactic acid bacteria (LAB) produce a wide repertoire of proteolytic enzymes that are involved in metabolizing casein for cell growth in milk, and these proteolytic systems have been extensively reviewed (7, 17, 33). Essentially, the cell surface proteinase releases oligopeptides from the milk caseins (as well as a limited amount of small peptides) that are transported into the cell. Intracellular peptidase activities combine to supply the cell with critical amino acids. The activities of the proteolytic enzymes are therefore essential for the growth of LAB in milk, which itself contains only low levels of peptides and free amino acids.

In cheese, the growth of nonstarter LAB is perhaps partially facilitated by the residual proteolytic activity of the starter LAB (21), but it is still dependent on the nonstarter proteolytic capability. The activities of the various proteolytic enzymes originating from the cheese flora determine the rate of ripening of cheeses such as Cheddar, and the balance of the various enzymes elaborated also affects the flavor of the final product. The enzymes of LAB that hydrolyze the oligopeptide products of proteinase activity have been investigated both biochemically and genetically (27–30). The name PepO was originally used for a specific oligopeptidase from lactococci (38). This designation is now used for metallo-oligopeptidases from LAB with significant amino acid sequence similarity to the lactococcal enzyme. However, the name oligopeptidase O (derived from oligoendopeptidase) was recently suggested as a suitable name for general use (26). PepO has not yet received an enzyme classification number (26).

PepO from Lactococcus lactis subsp. cremoris Wg2 (39) and other oligopeptidases identified in LAB are metalloenzymes, including the oligopeptidase from Lactobacillus delbrueckii subsp. bulgaricus B14 (3), and PepF from L. lactis subsp. cremoris NCDO763 (29). This has been determined by activity studies on the purified enzymes with metal ions and chelators such as EDTA, 1,10-phenanthroline, and phosphoramidon and the identification of the Zn²⁺-binding motifs HEISH (28) and HETGH (29). Sequence data from oligopeptidase genes have identified a number of characteristic motifs present in zinc-dependent metallo-oligopeptidases (2).

Although several oligopeptidases have been purified, only a few oligopeptidase-encoding genes have been cloned and sequenced from lactococci (28, 29, 42) and lactobacilli (6). However, the manipulation of LAB peptidases is a potentially useful approach to changing cheese flavor. In a preliminary investigation, Cheddar cheese was made with a lactococcal starter overproducing different peptidases (E. Johansen, L. U. Guldfeldt, H. Behrndt, D. Williams, and P. Strøman, 6th Symposium on Lactic Acid Bacteria, Genetics, Metabolism and Applications, poster presentation, 1999). Overproduction of PepO did not affect the organoleptic properties of the cheese, but the significance of this result is unclear, as neither the expression level of PepO in the mutant strain nor the susceptibility of the strain to autolysis was reported. Autolysis is a critical factor governing both the balance of proteolytic enzymes in cheese ripening and the timing of their introduction into the cheese matrix (8).

There are in fact a number of reports that indicate an important role for oligopeptidase activity in cheese ripening. Exterkate and Alting (13) showed that PepO is active in Gouda cheese and indeed affects the peptide composition. Baankreis et al. (1) demonstrated that a neutral oligopeptidase from L. lactis is active under cheese conditions and concluded that it probably plays a crucial role in the degradation of the very bitter cheese peptide β-casein fragment from position 193 to 209 [β-casein f(193–209)]. Recent results show that in model systems with water activities, pHs, and NaCl concentrations similar to those in Cheddar cheese, PepO from L. lactis remains active and hydrolyzes several peptides known to contribute to bitterness (J. R. Reid, S. A. Harvey, and T. Coolbear, unpublished results). Together, these findings strongly suggest that the release of sufficient levels of PepO by autolysis early in the ripening process would modify the peptide composition, and therefore the organoleptic properties, of the cheese. This means that differences in the specificity of action of oligopeptidases from different sources could be important in cheese flavor development.

In this study, we focused on the PepO enzyme of Lactobacillus rhamnosus HN001 (DR20). This is a probiotic strain that was first isolated from a dairy environment (32), and its genome has been sequenced (unpublished data). The pepO gene was cloned and overexpressed in both L. rhamnosus HN001 and L. lactis subsp. cremoris NZ9000. PepO was purified to homogeneity, and the N-terminal sequence was determined. The specificity for α_s1-casein f(1–23), α_s1-casein f(1–17), and bradykinin was analyzed by reverse-phase high-pressure liquid chromatography (RP-HPLC) and mass spectroscopy and compared with previously published data for oligopeptidases in other strains. The possible role of this PepO in reducing cheese bitterness during ripening is discussed.

MATERIALS AND METHODS

Bacterial strains, plasmids, media, and culture conditions.

Strains and plasmids used in this study are listed in Table 1. Freeze-dried cultures of L. rhamnosus HN001 were inoculated into 100 ml of MRS broth (10) in 100-ml bottles and incubated at 37°C for 18 to 24 h. For selection of transformants of L. rhamnosus, 5 μg of erythromycin was added per ml of medium. L. lactis was grown in M17 with 10 g of glucose per liter (40) at 30°C, and transformants were selected by addition of 5 μg of erythromycin or 5 μg of chloramphenicol per ml of medium. Escherichia coli was grown at 37°C in LB medium (35). Transformants were plated on medium containing 200 μg of erythromycin, 40 μg of X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside), and 0.5 mM IPTG (isopropyl-β-d-galactopyranoside) per ml of medium. For solid media, agar was added to 15 g/liter.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Characteristics or further modifications	Reference
Strains
L. rhamnosus HN001 (DR20)	Wild-type industrial nonstarter	32
	L. rhamnosus HN001 with pTRKH2	This study
	L. rhamnosus HN001 with pLB001	This study
L. lactis subsp. lactis NZ9000	Strain MG1363 with pepN::nisR nisK	19
	L. lactis subsp. lactis NZ9000 with pNZ8020	This study
	L. lactis subsp. lactis NZ9000 with pLB002	This study
	L. lactis subsp. lactis NZ9000 with pLB003	This study
	L. lactis subsp. lactis NZ9000 with pNZ8037	This study
	L. lactis subsp. lactis NZ9000 with pLB004	This study
E. coli DH5α	F− φ80dlacZΔM15 Δ(lacZYA-argF′) U169 deoR recA1 endA1 hsdR17(rK−, rK+) supE44 λ−thi-1 gyrA96 relA1
Plasmids
pTRKH2	High-copy-number shuttle vector between E. coli, Lactococcus, Lactobacillus, Streptococcus, and Enterococcus, with erythromycin resistance marker	31
pLB001	pepO from L. rhamnosus HN001 with its own promoter and terminator cloned in pTRKH2	This study
pNZ8020	Transcriptional fusion vector with nisA promoter and chloramphenicol resistance marker	11
pLB002	pepO from L. rhamnosus HN001 with its own terminator and fused to nisA in pNZ8020	This study
pLB003	pepO from L. rhamnosus HN001 with its own terminator cloned in pNZ8020	This study
pNZ8037	Translational fusion vector with nisA promoter and chloramphenicol resistance marker	11
pLB004	pepO from L. rhamnosus HN001 with its own terminator cloned in pNZ8037	This study

Cloning of pepO.

Genomic sequence data for pepO were previously obtained (unpublished data). Forward primers PepO10, PepO11, and PepO12 and reverse primer PepO2 (Table 2) were constructed to amplify fragments including the ribosome binding site, the coding sequence, and two inverted repeats downstream of the coding sequence. The pepO reading frame continued 129 bp upstream of the likely translational start (inferred from sequence alignment), and forward primer PepO3 was constructed to anneal upstream of this region (Table 2). All custom primers were produced by Life Technologies (Auckland, New Zealand).

TABLE 2.

Primers

Primer	Sequence (5′-3′)a	Feature(s)
PepO2	TTTTTTCTCGAGTGGCTACATTATTACCACGACCTG	Reverse primer downstream of the two inverted repeats downstream of pepO, with recognition site for XhoI for cloning in pTRKH2, pNZ8020, and pNZ8037
PepO3	TTTTTTGTCGACCCTTGATGTTATGATCAATGTCTGC	Forward primer upstream of pepO promoter, with recognition site for SalI for cloning in pTRKH2
PepO4	GTTGATGCTGATATGCACGATG	Sequencing primer, forward
PepO5	GCTTTTGGGATTGCCAACGAG	Sequencing primer, forward
PepO6	AGCAGGCATTTGATCAGCTGG	Sequencing primer, forward
PepO7	CGAGATTTCATGACCGATCGTTG	Sequencing primer, reverse
PepO8	CATCCATGCGTGCCATTCTG	Sequencing primer, reverse
PepO9	GCTTATCCGATCACGAACTGG	Sequencing primer, reverse
PepO10	TTTTTTAGATCTTGAGATTACGACATGAAAACGGAG	Forward primer including ribosome binding site and recognition site for BglII for cloning in pNZ8020
PepO11	TTTTTTAGATCTTGAGATAACGACATGAAAACGGAG	Forward primer including ribosome binding site, recognition site for BglII, and site-directed mutagenesis to create a stop codon in nisA upstream of pepO for cloning in pNZ8020
PepO12	AAAAAGAGGCAATTACATGTCATTACCAAG	Forward primer with the translational start of pepO and site-directed mutagenesis to create a recognition site for BspLU11I for cloning in pNZ8037

^aRecognition sites for restriction endonucleases are underlined; nucleotides different from those in the pepO sequence are in bold.

Total genomic DNA of L. rhamnosus HN001 was extracted and purified as described elsewhere (32). The sequence containing pepO was amplified by PCR with Pwo DNA polymerase (Boehringer Mannheim GmbH, Mannheim, Germany) and a Perkin Elmer GenAmp 9600 thermocycler (Applied Biosystems, Foster City, Calif.). PCR products and vectors were cut with SalI, BglII, BspLU11I, XhoI, and NcoI (Table 2) according to the manufacturer’s recommendations (Boehringer Mannheim). The PepO3-PepO2 product was ligated to pTRKH2, the PepO10-PepO2 and PepO11-PepO2 products were ligated to pNZ8020, and the PepO12-PepO2 product was ligated to pNZ8037.

E. coli DH5α was transformed (35) with the pTRKH2 ligations. The recombinant plasmids were prepared with a plasmid kit (Qiagen GmbH, Hilden, Germany). L. rhamnosus HN001 was made competent by growth in the presence of glycine (43) and then transformed with pLB001 by electroporation. To verify the transformants of L. rhamnosus, plasmids were prepared essentially as described previously (25) but on a smaller scale and with a commercial kit (Concert; Life Technologies). L. lactis subsp. cremoris NZ9000 was made competent by growth in the presence of glycine (15) and then transformed with the pNZ8020 and pNZ8037 ligations by electroporation.

Sequencing of pepO clones.

Both strands of four different pepO clones were sequenced by using a BigDye terminator cycle sequencing ready-reaction kit (Applied Biosystems, Melbourne, Australia) and an MJ Research (Boston, Mass.) PTC-100TM thermal cycler and analyzed with an ABI Prism 377 automated sequencer. To assemble the data, the sequencing project in the software GeneWorks 2.5 (Oxford Molecular Group Inc., Campbell, Calif.) was used.

Northern hybridization and primer extension.

L. rhamnosus HN001 was grown in 50 ml of MRS medium in 50-ml tubes at 37°C. Total RNA was prepared as described previously (24). For primer extension, the avian myeloblastosis virus reverse transcriptase kit from Promega (Madison, Wis.) was used to end-label 10 pmol of the oligonucleotide 5′TTACTATCCGCACTCACAACAC with 30 μCi of [γ-³²P]ATP (Amersham Pharmacia Biotech UK Ltd., Little Chalfont, United Kingdom) and to perform primer extension reactions on 20 μg of total RNA. Sequencing reactions were generated by using the same ³²P-end-labeled primer with pLB001 as the template and the fmol cycle sequencing kit (Promega). The products were separated on a 7% acrylamide-7 M urea sequencing gel and visualized by autoradiography. Northern hybridization using pLB001 labeled with ³²P as the probe was done as described previously (24).

Production of PepO.

Overnight cultures of L. rhamnosus HN001 (wild type) and strains harboring the vector only or vector with the cloned pepO gene were inoculated (1% vol/vol) into fresh MRS and grown to late log phase. Overnight cultures of L. lactis subsp. cremoris NZ9000 (host strain) and strains harboring the vectors only or vectors with the cloned pepO gene were inoculated (1% vol/vol) into fresh M17 and grown to an optical density at 600 nm of 0.5. For induction of the nisin promoter, nisaplin was added to a final concentration of 0.15 μg/ml, and the cells were further cultured until the optical density at 600 nm was 1.20. As a background level control, all lactococcal cultures were also grown without addition of nisaplin. The bacteria were harvested by centrifugation at 4°C and washed in 50 mM bis-Tris propane, pH 6.4.

Purification of PepO.

The washed pellets were resuspended in the bis-Tris propane washing buffer, and the cells were disrupted by two passes through a French press. The cell debris was removed by centrifugation at 27,500 × g for 30 min at 4°C, and the supernatant was stored at −80°C until needed. Cell lysis was checked by measuring the release of an intracellular marker enzyme, tagatose 1,6-bisphosphate aldolase (EC 4.1.2.40), using the assay procedure described previously (9) and by determining the soluble protein content with the bicinchoninic assay kit (36) (Sigma Chemical Co., St. Louis, Mo.), with bovine serum albumin as the standard.

The supernatant was loaded onto a 5- by 100-mm Protein-Pak Q column (Waters New Zealand Ltd, Auckland, New Zealand) equilibrated with 20 mM bis-Tris propane buffer, pH 8.0 (buffer A), and linked to a PC-controlled fast protein liquid chromatography system (Amersham Pharmacia Biotech Europe GmbH, Uppsala, Sweden). The column was washed with 15 ml of buffer A to remove unbound material. Buffer B was 20 mM bis-Tris propane (pH 8.0) containing 1 M NaCl. Bound protein was eluted at room temperature with a linear gradient of 0 to 30% buffer B over 30 min at a flow rate of 0.7 ml per min. Fractions containing high oligopeptidase activity [i.e., hydrolyzed α_s1-casein f(1–23)] eluted at approximately 22% buffer B and were pooled. NaCl was slowly added to the pooled fraction to a final NaCl concentration of 3 M. The sample was loaded onto an HR 5/5 phenyl-Superose fast protein liquid chromatography column (Amersham Pharmacia Biotech) equilibrated with 20 mM bis-Tris propane, pH 6.8, containing 3 M NaCl (buffer C). The column was washed with 10 ml of buffer C to elute unbound material. Buffer D was 20 mM bis-Tris propane, pH 6.8. Bound protein was eluted with a linear gradient of 0 to 100% buffer D in 40 min at a flow rate of 0.5 ml per min. Fractions containing high oligopeptidase activity (eluting at approximately 2.7 M NaCl) were concentrated 10-fold with Centricon 10 concentrators (Amicon Division, WR Grace, Beverly, Mass.). Fresh 20 mM bis-Tris propane buffer, pH 6.8, was added to reduce the concentration of NaCl to 150 mM. The purified enzyme was stored at −80°C until required.

PAGE.

Samples for analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) were prepared by heating (100°C) for 2 min with an equal volume of 0.125 M Tris-HCl (pH 6.8) containing 4% (wt/vol) SDS, 20% (wt/vol) glycerol, and 10% (vol/vol) β-mercaptoethanol. SDS-PAGE was carried out according to the procedure of Laemmli (20) with 12% (wt/vol) acrylamide Ready Gel precast gels (Bio-Rad New Zealand, Auckland, New Zealand).

Determination of molecular mass.

The molecular mass of the denatured monomer was determined by SDS-PAGE with the following protein standards (Amersham Pharmacia Biotech): myosin (220 kDa), phosphorylase b (97.3 kDa), albumin (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (30 kDa), trypsin inhibitor (30 kDa), and lysozyme (14.3 kDa).

N-terminal sequence analysis.

The N-terminal amino acid sequence of the purified oligopeptidase was determined by the automated Edman method using a 476A protein sequencer (Applied Biosystems).

Enzyme assays, peptide hydrolysis, and identification of peptide products.

The chymosin-generated peptides α_s1-casein f(1–17) and α_s1-casein f(1–23), comprising the first 17 and 23 residues of α_s1-casein, respectively, and bradykinin (Sigma Chemical Co.) were used as oligopeptidase substrates for specificity studies. α_s1-Casein f(1–23) was generated and purified by the method of Exterkate and Alting (12) and used as the substrate to detect oligopeptidase activity during the enzyme purification procedure. The peptide α_s1-casein f(1–17) was extracted from an 8-week-old Cheddar cheese by the hot water extraction method of Kuchroo and Fox (18) and purified by the RP-HPLC method of Exterkate and Alting (12). The method used for the hydrolysis of all substrates was essentially identical to that described by Pritchard et al. (34). Enzyme activity was deduced from the decreasing concentrations of substrate as detected by peptide absorbance at 214 nm after separation of substrate and products by RP-HPLC. After separation, the peptide products were collected, dried under vacuum, and identified by mass spectrometry with a triple-quadrupole model API 300 mass spectrometer (Applied Biosystems).

RESULTS

Cloning of pepO.

All E. coli colonies that appeared after transformation with pTRKH2/pepO ligations were blue. Sequence analysis showed that the cloned fragment contained a promoter and translational start in the opposite orientation upstream of pepO, thus promoting read through in-frame to lacZ. We originally planned to clone pepO into pNZ8020 by using SalI or BglII, which have been reported to have unique sites in this vector (11). However, we found that both of these endonucleases cut the plasmid twice (data not shown), so the vector was restricted with BamHI instead. Attempts were made to transform E. coli with pNZ8020/pepO ligations, but all transformants screened either contained the pNZ8020 alone or contained the pNZ8020 with the pepO insert in the opposite orientation with respect to the nisA promoter. To avoid apparent instability in E. coli, the ligations were introduced directly into L. lactis subsp. cremoris NZ9000. All cloned fragments were sequenced and found to be identical to the HN001 genomic sequence of pepO. By using the forward primer PepO10, the cloned fragment in pNZ8020 was found to encode a fusion protein between nisA and pepO (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 1). Therefore, alternative primers PepO11 and PepO12 were constructed (Table 2), where primer PepO11 introduced a stop codon in nisA upstream of pepO.

FIG. 1.

Cloning strategies in pNZ8020 and pNZ8037. The nisA promoter and ribosome binding sites for nisA and pepO are in bold, primers used to amplify pepO are double underlined, and nucleotides changed by site-directed mutagenesis are in bold italics. (A) pNZ8020 was cut with BamHI and ligated to the PCR product of pepO with forward primer PepO10 cut with BglII. This generated a NisA-PepO fusion protein. (B) pNZ8020 was cut with BamHI and ligated to the PCR product of pepO with forward primer PepO11 cut with BglII. Primer PepO11 introduced a stop codon for nisA by site-directed mutagenesis. (C) pNZ8037 was cut with NcoI and ligated to the PCR product of pepO with forward primer PepO12 cut with BspLU11I. An A instead of a T was introduced by primer PepO12 by site-directed mutagenesis, resulting in a codon for serine rather than threonine as the second amino acid.

The pepO gene and gene product.

The pepO gene of L. rhamnosus HN001 was 1,896 bp, encoding a putative protein of 632 amino acids (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 2). The sequence had a GC content of 49.5%, which is close to the average overall GC content of 46 to 48% observed for several other L. rhamnosus HN001 genes (unpublished data). The molecular mass and pI were calculated from the sequence to be 70.9 kDa and 4.86, respectively. The fusion protein NisA-PepO encoded by pLB002 had a calculated mass of 76.6 kDa. The amino acid sequence of PepO was 37% identical to PepO in Lactobacillus helveticus (accession number AF019410) and 36% identical to PepO in L. lactis (accession number L18760). The pepO gene product contained the sequence HEISH, which is a characteristic HEXXH-motif present in zinc metallopeptidases (2). Northern hybridization showed that the pepO transcript was about 1.9 kilonucleotides, indicating that it is monocistronic (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 3A). Primer extension identified the start of the transcript (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 3B), and a promoter with the configuration TTGACG-N₁₇-TATGCT was inferred (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 3C).

FIG. 2.

Nucleotide and protein sequences of pepO in L. rhamnosus HN001. The promoter, ribosome binding site (RBS), and zinc motif are shown in bold. The translated protein is shown, and the full open reading frame is indicated by italics. Arrows show the inverted repeats.

FIG. 3.

L. rhamnosus HN001 pepO transcript analysis. (A) Northern hybridization of total RNA with a pepO probe. (B) Primer extension analysis of pepO. Sequence reactions (A, G, C, and T) and primer extension (PE) were done with the same primer. The arrows indicate the primer extension products. (C) Interpretation of the primer extension results. The putative pepO −35 and −10 promoter hexamers (bold), transcription start site (underlined), and putative ribosome binding site and ATG start codon (italics) are shown.

Purification of oligopeptidase PepO.

RP-HPLC was used to quantify the oligopeptidase activity obtained from cell extracts of the various cultures standardized to a known cell dry weight by culture density at 600 nm. PepO activity was twofold higher in the L. rhamnosus HN001 transformant harboring pLB001 than in wild-type HN001. However, in L. lactis subsp. cremoris NZ9000, it was estimated that PepO activity was 10- to 15-fold higher in the induced transformant harboring pLB003 and 20- to 25-fold higher in the induced transformant harboring pLB004 than in the host strain (data not shown). The increased expression level of PepO upon induction of L. lactis subsp. cremoris NZ9000 harboring pLB003 was clearly observable following SDS-PAGE analysis of the cell extract obtained before and after induction (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 4B). The plasmid encoding NisA-PepO (pLB002) also overproduced PepO, but to a lesser extent. It was estimated that the uninduced lactococcal transformants harboring either pLB003 or pLB004 expressed about twofold more PepO activity than either the uninduced host strain or the uninduced strain harboring the vectors only. This showed that the nisA promoter is turned on at a reasonable level even without nisin. Since overproduction of L. rhamnosus PepO was much higher in L. lactis subsp. cremoris NZ9000 than in L. rhamnosus HN001, the enzyme was purified from the lactococcal transformant.

FIG. 4.

SDS-PAGE showing the molecular mass and purity of PepO. (A) Cell extract of L. lactis subsp. cremoris NZ9000 with pLB003 induced with nisaplin (lane 1) and the final purified PepO (lane 2). (B) Cell extract of L. lactis subsp. cremoris NZ9000 with pLB003 obtained without induction (lane 1) and induced with nisaplin (lane 2). Lane 3, standard proteins.

Determination of the size, N-terminal amino acid sequence, and substrate specificity of PepO.

The molecular mass of the denatured monomer of the PepO from the lactococcal transformant was estimated to be 70 kDa by SDS-PAGE (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 4A), which was in agreement with the mass calculated from the nucleotide sequence. The chemically determined sequence of amino acid residues 1 to 20 of PepO purified from the clone harboring pLB003 was Thr-Leu-Pro-Arg-Ile-Gln-Asp-Asp-Leu-Tyr-Leu-Ala-Val-Asn-Gly-Glu-Trp-Gln-Ala-Lys. This amino acid sequence was in agreement with that predicted from the nucleotide sequence following the removal of the N-terminal methionine (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 2). The sequence of amino acid residues 1 to 20 of the purified PepO from the clone harboring pLB004 was identical except for the first amino acid residue, which was Ser, and confirmed the single-residue substitution necessitated by the translational fusion design (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 1). The purified PepO from both clones hydrolyzed α_s1-casein f(1–23) and bradykinin in an identical manner. The specific bonds hydrolyzed are shown in Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 5 and are compared with those hydrolyzed by other oligopeptidases from LAB. The primary sites of α_s1-casein f(1–23) hydrolysis by the cloned L. rhamnosus PepO were in the N-terminal half of the peptide at Pro-5-Ile-6, Lys-7-His-8, His-8-Gln-9, and Gln-9-Gly-10. This is in direct contrast with the specificities of all other oligopeptidases of LAB shown in Fig. 5A. The bonds Pro-5-Ile-6 and Gln-9-Gly-10 were hydrolyzed more rapidly by the L. rhamnosus PepO than were the bonds Lys-7-His-8 and His-8-Gln-9. After extended incubation times (4 h or more), all of the peptides initially produced by hydrolysis of the primary sites were themselves hydrolyzed, except for the peptide α_s1-casein f(1–5). This peptide was resistant to further hydrolysis by PepO and therefore continued to accumulate until all of the original substrate and peptide products containing the Pro-5-Ile-6 bond were completely depleted.

FIG. 5.

Bonds hydrolyzed in α_s1-casein f(1–23) and bradykinin by PepO from L. rhamnosus HN001 compared with specificities of previously described oligopeptidases: LEP II (44), Lactococcus lactis subsp. lactis MG1363 PepO (37), L. lactis neutral thermolysin-like oligoendopeptidase (NOP) (1), LEP I (45), L. lactis alkaline endopeptidase (1), L. delbrueckii subsp. bulgaricus B14 endopeptidase (3), L. lactis subsp. cremoris SK11 endopeptidase (34), L. lactis neutral neprilysin-like endopeptidase (NEP) (23), L. paracasei Lc-01 oligopeptidase (41), and L. lactis subsp. cremoris NCDO763 PepF1 (29).

The primary hydrolysis sites within α_s1-casein f(1–17) were identical to those described for α_s1-casein f(1–23). However, the relative rates of hydrolysis of the four primary sites were not the same for each substrate. When α_s1-casein f(1–17) was used, the bonds Lys-7-His-8 and Gln-9-Gly-10 were hydrolyzed more rapidly than were the bonds Pro-5-Ile-6 and His-8-Gln-9.

The cloned L. rhamnosus PepO rapidly cleaved the Pro-7-Phe-8 bond of bradykinin, thus producing the bradykinin fragments f(1–7) and f(8–9) (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 5). After prolonged digestion times, very slow cleavage of the Gly-4-Phe-5 bond of bradykinin f(1–7) was also detected (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 5).

DISCUSSION

Several oligopeptidases have been purified from lactococci and lactobacilli, but only a few oligopeptidase genes have been sequenced. Two enzymes, an oligopeptidase from L. delbrueckii subsp. bulgaricus B14 (3) and PepF1 from L. lactis subsp. cremoris NCDO763 (29), have been purified and characterized, and their respective genes have been sequenced. Here we present the characterization of the oligopeptidase PepO and its gene, pepO, from L. rhamnosus HN001 (DR20). Sequence data from all oligopeptidase genes studied to date suggest that lactobacillus genes are monocistronic (6) whereas lactococcal genes are part of operons (28, 29, 42). This is consistent with our results for pepO in L. rhamnosus HN001, shown by primer extension and Northern analysis to also be monocistronic. The sequence identity between oligopeptidase O enzymes from the strains compared is low. PepO from L. rhamnosus HN001 had only 37 and 36% amino acid identity to PepO in L. helveticus and L. lactis, respectively. All three genes encode the HEXXH Zn²⁺ binding motif (2): HEISH in L. rhamnosus and L. lactis and HEIVH in L. helveticus.

The overexpression of pepO in L. rhamnosus HN001 was very low when it was expressed from its own promoter in pTRKH2. This vector has previously been shown to be a high-copy-number vector (31) but appeared to be a low-copy-number vector in L. rhamnosus HN001, as estimated from plasmid preparations (data not shown). This may explain the relatively low overexpression of pepO in L. rhamnosus HN001. An alternative explanation might be that the pepO promoter is transcriptionally regulated, so that even with increased gene copy number, the total expression is not increased.

The production of the L. rhamnosus HN001 PepO in L. lactis subsp. cremoris NZ9000 was high because of the efficient nisin induction system (Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 4B). This suited our purpose of producing large amounts of PepO for purification. However, even the uninduced transformants possessed higher PepO activity than the host strain and the strain harboring the vectors only. This showed that the nisA promoter is partially active in M17 medium containing glucose even in the absence of nisin. This might be fatal if there is a need to precisely regulate the production of a protein. It has recently been reported that the nisA promoter is active in the absence of nisin when lactose or galactose is present (5). The production of PepO in L. lactis, which has a low average GC content of 35.4 to 37.3% (4, 16), was not hampered by the relatively high (49.5%) GC content of the L. rhamnosus pepO gene.

The purified PepO cleaved the Pro-7-Phe-8 bond in bradykinin, a specificity similar to that of a neutral oligopeptidase from L. lactis (23). In addition, PepO slowly hydrolyzed the Gly-4-Phe-5 bond of bradykinin. This has also been shown for lactic acid bacterium endopeptidase II (LEP II) from L. lactis subsp. cremoris H61 (44) and for the oligopeptidases from L. lactis subsp. cremoris SK11 (34) and L. delbrueckii subsp. bulgaricus B14 (3).

However, the specificity for α_s1-casein f(1–23) was dramatically different from that of previously described oligopeptidases from lactococci and lactobacilli. The bonds in α_s1-casein f(1–23) that were hydrolyzed by the L. rhamnosus HN001 PepO, i.e., Pro-5-Ile-6, Lys-7-His-8, His-8-Gln-9, and Gln-9-Gly-10, were all located in the N-terminal part of the substrate. In contrast, all of the bonds hydrolyzed by the other oligopeptidases studied to date (shown for comparison with L. rhamnosus HN001 PepO in Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 5) are located in the C-terminal part of the substrate, with the exception of one bond (Gln-9-Gly-10) that is hydrolyzed by LEP II (44).

The differences in the specificity of the L. rhamnosus HN001 PepO have ramifications for the flavor of the peptides produced. The peptide α_s1-casein f(1–13) has been shown to be bitter and to contribute to the bitterness of Cheddar cheese (22). α_s1-Casein f(1–14) is also likely to be bitter, since it has a very similar structure and hydrophobicity, and work by our research group has shown that α_s1-casein f(1–17) is also very bitter (unpublished data). Each of these peptides is known to accumulate during cheese ripening (13, 14).

The specificities reported for oligopeptidases from other LAB and detailed in Fig. 1A). GT198 contains a heptad leucine repeat that resembles the leucine dimerization domain (1). When displayed as an α-helical wheel, the heptad repeat of leucines is aligned at the d position and hydrophobic residues are aligned at the a positions to create the classic interacting surface found in the leucine zipper protein family (Fig. 5 indicate that if these enzymes are released into the cheese matrix during the early stages of ripening, they will hydrolyze α_s1-casein f(1–23) to produce the bitter cheese peptides discussed above and other closely related bitter peptides. Indeed, Exterkate and Alting (13) detected α_s1-casein f(1–14) in Gouda cheese and concluded that it was produced from hydrolysis of α_s1-casein f(1–23) by an intracellular endopeptidase released by early starter lysis. However, the L. rhamnosus HN001 PepO described in the present study is very active against the N-terminal region of α_s1-casein f(1–23), and it is highly likely, therefore, that this enzyme will efficiently degrade the bitter breakdown products of this casein fragment that may be produced by the action of either lactocepin or PepO from other LAB. This suggestion is strongly supported by the rapid hydrolysis of α_s1-casein f(1–17) by the L. rhamnosus HN001 PepO reported in the present study. Therefore, if a sufficient quantity of active PepO is released from L. rhamnosus HN001 by autolysis during ripening, it could play a role in controlling cheese bitterness.