Impaired Growth Rates in Milk of Lactobacillus helveticus Peptidase Mutants Can Be Overcome by Use of Amino Acid Supplements

Abstract

To evaluate the contribution of intracellular peptidases to the growth of the 14-amino-acid (aa) auxotroph Lactobacillus helveticus CNRZ32, single- and multiple-peptidase-deletion mutants were constructed. Two broad-specificity aminopeptidases (PepC and PepN) and X-prolyl dipeptidyl aminopeptidase (PepX) were inactivated through successive cycles of chromosomal gene replacement mutagenesis. The inactivation of all three peptidases in JLS247 (ΔpepC ΔpepN ΔpepX) did not affect the growth rate in amino acid-defined medium. However, the peptidase mutants generally had decreased specific growth rates when acquisition of amino acids required hydrolysis of the proteins in milk, the most significant result being a 73% increase in generation time for JLS247. The growth rate deficiencies in milk were overcome by amino acid supplements with some specificity to each of the peptidase mutants. For example, milk supplementation with Pro resulted in the most significant growth rate increase for ΔpepX strains and a 7-aa supplement (Asn, Cys, Ile, Pro, Ser, Thr, and Val) resulted in a JLS247 growth rate indistinguishable from that of the wild type. Our results show that characterization of the activities of the broad-specificity aminopeptidases had little predictive value regarding the amino acid supplements found to enhance the milk growth rates of the peptidase mutant strains. These results represent the first determination of the physiological roles with respect to specific amino acid requirements for peptidase mutants grown in milk.

Lactobacillus helveticus belongs to a diverse group of organisms known as lactic acid bacteria (LAB), which are defined by the production of lactic acid as a major product of carbohydrate fermentation. These bacteria are ubiquitous in the environment and propagate in niches such as plant surfaces, decaying plant material, diseased tissues, the oral cavity, and the gastrointestinal tract of many animals (4). They also have significant commercial value due to their involvement in the production of wine, lambic beers, sour mash whiskey, sourdough bread, pickled vegetables, cured meats, sausages, fermented milks, and cheeses.

Bos taurus milk is used routinely for the study of the proteolytic system and physiology of L. helveticus, providing a consistent medium with a well-characterized set of proteins and adaptive significance for dairy-related LAB (10). L. helveticus CNRZ32 has 14 amino acid auxotrophies and thus is dependent on amino acid transport and/or transport and hydrolysis of exogenous peptides to satisfy these nutritional requirements. The nonprotein nitrogen fraction of milk (defined as free amino acids and <3,000-Da peptides) is estimated to contain approximately 1% (wt/vol) of the total nitrogen component, while proteins (mainly caseins) are the predominant form of nitrogen in milk. Approximately 1/10 of the final cell density of Lactococcus lactis growth in milk is supported by nonprotein nitrogen (17), though no similar evaluation has been reported for L. helveticus or any other LAB prior to this work.

Acquisition of amino acids through the hydrolysis of caseins (the preferentially hydrolyzed milk proteins) is accomplished by a complex proteolytic system comprised of proteinase(s), endopeptidase(s), aminopeptidase(s), tripeptidase(s), dipeptidase(s), and peptide transport systems (6, 18, 27). Since all of the identified peptidases of L. helveticus are believed to be intracellular (18), the acquisition of amino acids is dependent on the activity of at least one extracellular proteinase to supply transportable peptides. A thorough review of the known specificities of LAB peptidases can be found in Christensen et al. (6). Broad-specificity aminopeptidases remove the N-terminal amino acids from a peptide (X↓Y-Z…), with specificity dependent on the amino-terminal and penultimate amino acids as well as the peptide length. For example, PepC and PepN both hydrolyze N-terminal lysine or arginine with approximately 100-fold-greater efficiency than that with which they hydrolyze N-terminal proline or glycine. Although the common chromogenic substrates resemble dipeptides, analysis of natural substrates indicates that both PepC and PepN hydrolyze substrates >6 amino acids (aa) in length. PepX has specificity for removal of proline-containing dipeptides (X-Pro↓Y…) from the N termini of peptides but is also dependent on the amino-terminal and tertiary amino acids of the peptide. Endopeptidases hydrolyze internal peptide bonds (…U-V-W↓X-Y-Z…) relatively independently of the N-terminal amino acids and with differing specificity for substrate size.

Optimal growth of LAB in milk depends on the cooperative action of several peptidases with different specificities for the release of amino acids from milk protein-derived peptides. The growth rates of L. lactis multiple-peptidase mutant derivatives are generally reduced (22), suggesting limited liberation of one (or more) essential amino acid. The general lack of milk growth rate differences observed for lactococcal single-peptidase mutants is attributed to compensation of activity by peptidases with overlapping specificities. However, the burden for acquisition of essential amino acids by the L. lactis multiple-peptidase mutant is less severe than would be predicted for L. helveticus CNRZ32 (8-aa auxotroph and 14-aa auxotroph, respectively).

In this study, we characterized peptidase mutants of L. helveticus CNRZ32 for their effect on growth rates and peptidase activities. We evaluated the peptidase mutants to determine which supplemented amino acids could compensate for the limited milk growth rates. To our knowledge, the present work is the first report detailing the physiological requirement to attain specific amino acids from milk protein as well as the role of specific peptidases in acquiring these amino acids from milk protein-derived peptides.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The strains and plasmids used in this study are listed in Table 1. L. helveticus CNRZ32 cultures were grown in MRS broth (Difco Laboratories, Detroit, Mich.) or on MRS broth solidified with 1.5% granulated agar or pour plated in MRS broth solidified with 0.75% granulated agar. Skim milk for growth analysis was obtained in pasteurized form (Babcock Dairy, Madison, Wis.) and then steamed twice for 20 min, with a 1-h, 42°C incubation between treatments. Amino acid-defined medium was prepared with the components listed in Table 2 and autoclaved for sterilization. Plated cultures of L. helveticus were incubated under anaerobic conditions with a BBL GasPak system (Becton Dickinson, Sparks, Md.). Incubations were routinely performed at 42°C except as described for transformation and construction of peptidase mutant strains. Erythromycin (ERY) was added at 10 ng, 50 ng, 500 ng, or 1,000 ng per ml, depending on the stage of mutant construction and the need for a screen or selection. Escherichia coli strains DH5α and SURE were grown in Luria-Bertani medium with shaking at 37°C or on Luria-Bertani medium solidified with 1.5% granulated agar; ampicillin, chloramphenicol, ERY, and tetracycline were added at 60 μg, 100 μg, 500 μg, and 12 μg per ml, respectively.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant feature(s)	Source or reference
Strains
L. helveticus
CNRZ32	Wild type; strictly auxotrophic for 13 aa (Arg, Glu, His, Ile, Leu, Lys, Met, Phe, Pro, Thr, Trp, Tyr, and Val); requires either Asp or Asn (but not both); does not require Ala, Cys, Gln, Gly, and Ser	Laboratory strain
JLS241	ΔpepC derivative of CNRZ32	This work
JLS242	ΔpepN derivative of CNRZ32	This work
JLS243	ΔpepX derivative of CNRZ32	This work
JLS244	ΔpepC ΔpepN (ΔpepN derivative of JLS241)	This work
JLS245	ΔpepC ΔpepX (ΔpepX derivative of JLS241)	This work
JLS246	ΔpepX ΔpepN (ΔpepN derivative of JLS243)	This work
JLS247	ΔpepC ΔpepN ΔpepX (ΔpepN derivative of JLS245)	This work
E. coli
DH5α	Used for plasmid construction and maintenance	Bethesda Research Laboratories
SURE	Used for plasmid construction and maintenance	Stratagene
Plasmids
pJDC9	6.95 kb; lacZ Emr	5
pUC19	2.69 kb; Ampr	Bethesda Research Laboratories
pSUW80	2.0-kb KpnI-SphI fragment containing pepC ligated into pJDC9	11
pSUW33	3.8-kb XbaI fragment containing pepN ligated into pUC19	7
pSUW2	3.2-kb fragment containing pepX ligated into pJDC9	25
pSA3	10.2-kb shuttle vector; Cmr Emr Tcr; replication deficient at 44°C in L. helveticus	8
pSUW241	3.4-kb SalI-BamHI fragment containing 366-bp internal deletion in pepC (bases 438-803) ligated into the Tcr marker of pSA3	This work
pSUW242	3.5-kb BamHI-SphI fragment containing 272-bp internal deletion in pepN (bases 1397-1668) ligated into the Tcr marker of pSA3	This work
pSUW202	3.1-kb SmaI-SphI fragment containing 350-bp internal deletion in pepX (bases 1582-1931) ligated into the Tcr marker of pSA3	2
pTRK-L2	6.4-kb shuttle vector; lacZ Emr	26
pTRK-LC	1.7-kb SspI fragment from pSUW80 containing pepC ligated into pTRK-L2	This work
pTRK-LN	3.8-kb SmaI-SalI fragment from pSUW33 containing pepN ligated into pTRK-L2	This work
pTRK-LX	3.2-kb SmaI-SalI fragment from pSUW2 containing pepX ligated into pTRK-L2	This work

TABLE 2.

Defined medium components

Componenta	Addition per liter
Salts
Sodium acetate (trihydrate)	5.0 g
Sodium citrate (dihydrate)	2.0 g
Potassium phosphate (monobasic)	1.0 g
Potassium phosphate (dibasic)	1.0 g
Sodium chloride	200 mg
Calcium chloride (dihydrate)	200 mg
Magnesium sulfate	200 mg
Manganese sulfate (monohydrate)	50 mg
Components dissolved in 1 M HCl (5 ml)b
dl-Aspartate	200 mg
l-Glutamate	400 mg
l-Tryptophan	25 mg
Carbohydrate
d-(+)-Glucose (anhydrous)	20 g
Components dissolved in 1 M NaOH (6 ml)b
l-Phenylalanine	20 mg
l-Tyrosine	70 mg
Adenine, guanine, uracil, xanthine	25 mg each
Other amino acidsb
l-Alanine	400 mg
l-Arginine	200 mg
l-Asparagine	400 mg
l-Cysteine	80 mg
l-Glutamine	400 mg
Glycine	400 mg
l-Histidine	200 mg
l-Isoleucine	80 mg
l-Leucine	80 mg
l-Lysine	400 mg
dl-Methionine	80 mg
l-Proline	800 mg
dl-Serine	400 mg
dl-Threonine	400 mg
l-Valine	400 mg
Miscellaneous components
Tween 80 (polyoxyethylenesorbitan monooleate)	1 ml
Tween 20 (polyoxyethylenesorbitan monolaurate)	1 ml
Glycerol	1 ml
dl-Mevalonic acid lactone	10 μl
Pyridoxal HCl	10 mg
RPMI 1640 vitamin solution (100×)c	20 ml

^aComponents were dissolved in the order listed to maintain relatively moderate pH during addition of the amino acid stocks. After dissolving all components (except the vitamin solution), the medium was adjusted to pH 6.50, autoclaved (121°C, 15 lb/in², 15 min), and then stored at ambient temperature.

^bComponents were combined, stored as dry stocks, and dissolved just before addition.

^cSigma R-7256 (supplied sterile) was added to the cooled medium after autoclaving just prior to inoculation.

General DNA techniques and sequencing.

Molecular cloning was done essentially as described by Sambrook et al. (28). Restriction enzymes, Klenow enzyme, T4 DNA ligase, Platinum Taq polymerase, deoxynucleotides, and custom PCR primers were obtained from Gibco-BRL (Gibco-BRL Life Technologies, Inc., Gaithersburg, Md.) and used per the supplier's protocols. Purification of chromosomal DNA, plasmids, and PCR products was generally accomplished using the DNeasy tissue kit and QIAquick PCR purification kit (Qiagen, Inc., Valencia, Calif.). Plasmids and PCR products were routinely separated in 0.6 to 1.5% agarose gels and stained with ethidium bromide. Photodocumentation was done with a Foto/Analyst Archiver system (Fotodyne Inc., Hartland, Wis.). Cycle sequence reactions were prepared with BigDye Terminator mix (PE Applied Biosystems, Foster City, Calif.), and sequence determination was done on a PE Biosystems 377XL automated DNA sequencing instrument (DNA Sequence Laboratory, University of Wisconsin Biotechnology Center). DNA sequences were analyzed using Lasergene99 software (DNASTAR Inc., Madison, Wis.). Predicted protein identity searches were done using the BLAST network service (1).

Construction of peptidase deletion vectors.

The deletion derivative of a pepC subclone in pTRK-L2 was constructed by PCR amplification using divergent 5′ phosphorylated primers for sites internal to the gene. The amplification product was purified and ligated. The ligation mix was used to transform E. coli cells. Plasmids obtained from transformed E. coli isolates were screened according to the sizes of PCR amplification products from primers convergent on the deletion site. In addition to the deletion copy of pepC, additional DNA segments derived from inverse PCR amplification of flanking chromosomal DNA were subcloned into pSA3 to construct the deletion vector pSUW241. The pSA3 deletion derivatives of pepN and pepX were constructed using the Erase-A-Base kit (Promega Corp., Madison, Wis.) per the supplier's protocol.

Transformation methods.

The transformation of E. coli strains was accomplished with a Gene Pulser per the protocol supplied by the manufacturer (Bio-Rad Laboratories, Richmond, Calif.). The transformation of L. helveticus was accomplished by the following method. A late-exponential-phase MRS culture (200 ml; optical density at 600 nanometers [OD₆₀₀] of ∼1.5 as calculated from a 1:10 dilution) was chilled to 4°C and harvested by centrifugation (5,000 × g, 8 min, 4°C). The cells were washed four times by vigorous shaking with 150 ml of 4°C sterile distilled water. After the last wash, cells were resuspended in 10 ml of sterile distilled water. The OD₆₀₀ was determined, and a volume of cell suspension was calculated that, when the suspension was concentrated, would result in an OD₆₀₀ of 50 in a 0.8-ml volume. The calculated volume of cell suspension was centrifuged (14,000 × g, 15 s), the supernatant was discarded, and the cells were resuspended in 0.8 ml of 0.5 M sucrose-50 mM l-proline (unadjusted pH). Plasmid (1 to 2 μg) was added, and the entire volume (0.8 ml) of cell suspension was pulsed at 2.5 kV, 25 μF, and 200 Ω in a 0.4-cm-path-length cuvette. A 0.5-ml volume of the pulsed cell suspension was transferred to 10 ml of MRS broth-5 mM CaCl₂ and incubated for 3 h at 37°C. Following this incubation, 200-μl samples were added to 10 ml of 50°C tempered MRS broth-5 mM CaCl₂-0.75% agar containing 500 ng of ERY/ml and pour plated. Upon solidification, the plates were incubated at 37°C (for pSA3 and derived vectors) or 42°C under anaerobic conditions. Transformant CFU were visible after 24 to 48 h of incubation.

Construction of L. helveticus peptidase deletion strains.

For each cycle of mutant construction, a minimum of eight independent L. helveticus transformants with pSA3 or of one of the peptidase deletion derivative vectors (pSUW202, pSUW241, or pSUW242) was picked and resuspended in 1 ml of MRS broth. Screening for isolates with the peptidase deletion derivative vector integrated into the chromosome was done by plating the cell suspension on MRS agar containing 50 ng of ERY/ml and incubation for 24 h at 44°C (a replication-inhibitory temperature for pSA3). Screening was utilized because higher (selective) concentrations of ERY resulted in an increased incidence of chromosomal amplification of pSA3 and the integrated genes. Nonintegrant isolates appeared as pinpoint-sized CFU and were not viable when propagation in MRS broth with 50 ng of ERY/ml at 37°C was attempted. Integrant isolates appeared as normally sized CFU (2- to 3-mm diameter in 24 h). Confirmation of integration was done by PCR using one of two primer pairs, specific to pSA3 and the undeleted copy of the respective peptidase gene, for both possible integration orientations (data not shown). An integrant isolate containing a single copy of pSA3 and the wild-type gene and the deletion copy of the peptidase gene in the chromosome resulted in a DNA band in only one of the two reactions, depending on the orientation of the wild-type gene with pSA3.

Integrant isolates were propagated without antibiotics at the pSA3 replication-permissive temperature of 37°C, allowing recombinatorial excision and curing of pSA3 and either the wild-type or deletion copy of the peptidase gene (24). Stationary-phase cultures were plated (to prevent acid inhibition; ≤100 CFU/plate) on MRS agar containing 10 ng of ERY/ml and incubated at 44°C under anaerobic conditions. MRS agar containing 10 ng of ERY/ml was determined to be inhibitory, but nonlethal and nonselective, to Em^s isolates. During a window period of the incubation (between 16 and 20 h), colonies of Em^s isolates were distinctly smaller (<0.5-mm diameter) than those of the Em^r isolates (2- to 3-mm diameter). The Em^s isolates were confirmed by spot plating (spot inoculation of ∼10³ cells) on MRS agar with or without 50 ng of ERY/ml. No less than 20% of Em^s isolates from each cycle of mutant construction screened were of the deletion genotype. The peptidase genotype of the isolates was determined by PCR using primer pairs designed to distinguish the wild-type and deleted genes by size (Fig. 1).

FIG. 1.

Genotyping of Lactobacillus helveticus CNRZ32 and its peptidase mutants by PCR from chromosomal DNA preparations. Expected PCR product sizes as determined from sequence data were 759 bp (pepC), 575 bp (pepN), 490 bp (pepX), 393 bp (ΔpepC), 303 bp (ΔpepN), and 140 bp (ΔpepX). Lanes 1 and 10 contain a 1-kb DNA ladder (Gibco). Lanes 2 to 9 contain the three products derived from a chromosomal template with the primer pairs described below. Chromosomal templates were as follows: lane 2, wild type; lane 3, strain JLS241 (ΔpepC); lane 4, JLS242 (ΔpepN); lane 5, JLS243 (ΔpepX); lane 6, JLS244 (ΔpepC ΔpepN); lane 7, JLS245 (ΔpepC ΔpepX); lane 8, JLS246 (ΔpepX ΔpepN); and lane 9, JLS247 (ΔpepC ΔpepN ΔpepX). The primer pairs for chromosomal genotype screening were as follows: for pepC, 5′-AGCAATCAGGTCGTTGCTG-3′ and 5′-CAAACCATACAACTTCGCC-3′; for pepN, 5′-CACCATAAGTTTGGCAATGC-3′ and 5′-TCTGGTTCAACGAATTGACG-3′; and for pepX, 5′-TGCCAATGAAGCATCACC-3′ and 5′-ATCTCACCCTTTGGCAGAC-3′. The expected PCR product sizes determined from sequence data were 759 bp (pepC), 575 bp (pepN), 490 bp (pepX), 393 bp (ΔpepC), 303 bp (ΔpepN), and 140 bp (ΔpepX).

The peptidase mutant phenotype of isolates from MRS cultures was confirmed by enzyme assays with Triton X-100-permeated cells (data not shown). The substrate GlyPro-ρNA was used to confirm the loss of PepX activity. The substrate Lys-ρNA was used to confirm the loss of PepN activity and of PepC activity from ΔpepN parental strains. Stationary-phase MRS broth cultures of strains JLS242 (ΔpepN) and JLS246 (ΔpepX ΔpepN) had fourfold higher activity for Lys-ρNA than those of strains lacking PepC and PepN activity (JLS244 and JLS247). Due to the disproportionately high activity of PepN, the loss of PepC activity was not discernible by a permeability assay for strain JLS241 (ΔpepC) relative to the parental strain (wild type).

From each set of eight independent transformants picked for each cycle of mutant construction, at least three independent mutant and parental type recombinants were isolated. Prior to use for subsequent deletion construction, the milk acidification rates of the parent strain and each of the parental type recombinants and peptidase mutants were evaluated to confirm that the phenotypes were similar for peptidase genotypically equivalent strains. One peptidase mutant isolate for each peptidase (JLS241, JLS242, and JLS243) was then chosen randomly for additional cycles of peptidase deletion for construction of the multiple-peptidase mutants (Table 1).

Preparation of cells for direct PCR.

A 50-μl sample of full-growth culture was washed with 1 ml of sterile water, resuspended in 200 μl of 12.5% sucrose (filter sterile), and incubated for 15 min at 25°C. A total of 3 μl of the cell suspension was mixed with 27 μl of complete PCR stock solution containing primers, 2 mM MgCl₂, and Platinum Taq polymerase (Life Technologies, Rockville, Md.) per the manufacturer's protocol. Thermal cycle reactions were done in a Perkin-Elmer model 480 thermal cycler (Perkin-Elmer Corp., Norwalk, Conn.) programmed as follows: 10 min at 94°C and 35 cycles of 45 s at 94°C, 30 s at 55°C, 60 s per expected product kilobase at 72°C, and 2 min at 72°C, followed by 4°C until analysis.

UFMP.

The ultrafiltered milk permeate (UFMP) was prepared from pasteurized skim milk in an ultrafiltration (UF) system consisting of two polyethersulfone UF membranes in parallel (model AES30X; Ladish Co., Kenosha, Wis.), providing a nominal cutoff of 10,000 kDa and 5.57 square meters of membrane surface area. Removal of milk protein from UFMP was confirmed by comparison of samples against known concentrations of bovine milk protein standards (α-, β-, and κ-casein; Sigma, St. Louis, Mo.) and skim milk samples electrophoresed in a 16.5% Tris-tricine gel (Bio-Rad Laboratories) and stained with Coomassie G-250 (data not shown). Evaluation of UFMP samples (concentrated 20-fold) for >3-kDa peptides in a Centricon-3 spin cartridge (Amicon, Beverly, Mass.) indicated that the UF process achieved a minimum 1,000-fold reduction of native milk protein, corresponding to a protein concentration of <30 μg/ml.

Quantification of free amino acids and small peptides in UFMP against Gly standard curves was accomplished using the Cd-ninhydrin assay (method D; 9) with modifications (12). The concentration of the unhydrolyzed UFMP sample was 580 μM (44 μg/ml) Gly standard equivalents. Together, the gel electrophoresis and Cd-ninhydrin assay data indicate that the predominant source of nitrogen in the UFMP was small peptides.

Milk medium.

The milk medium for cultures used to determine peptidase activities was prepared by addition of a 10% volume of heat-treated skim milk (see above) to UFMP medium. The UFMP medium was prepared by supplementation of UFMP with the following components of the defined medium (Table 2). Sodium acetate (5.0 g/liter), sodium citrate (2.0 g/liter), potassium phosphate monobasic (1.0 g/liter), and potassium phosphate dibasic (1.0 g/liter) were added to achieve a buffering capacity and titration profile similar to those of skim milk. Polyoxyethylenesorbitan monooleate (Sigma) was added (1.0 ml/liter) to replace essential fatty acids lost from the milk ultrafiltration process. The pH was adjusted to 6.60, filtered for sterility, and stored at 4°C for up to 3 days. Immediately prior to inoculation, the medium was supplemented with RPMI 1640 vitamin stock (20 μl/ml). Growth of cultures in milk medium circumvented the need for the clarification process required for cell density measurements from skim milk cultures, which was determined to be detrimental to the measurement of aminopeptidase activity. Cells were harvested by decanting ∼90% of the culture medium from above the precipitated milk protein.

Preparation of cell-free extracts (CFE).

Cells were harvested and washed twice in 0.1 M bis-Tris-2 mM dithiothreitol (DTT) buffer (pH 6.5) by centrifugation (10,000 × g, 8 min, 4°C). The cell pellets were resuspended to a calculated OD₆₀₀ of ∼25 with 15 ml of 0.1 M bis-Tris-2 mM DTT buffer (pH 6.5). The cell suspension was transferred to 40-ml centrifuge tubes containing 15 g of glass beads (Sigma G-9018; 150 to 212 μm) and held in an ice-water bath. Cells were disrupted by 3 cycles of shaking (using a Red Devil model 5410 paint shaker [Red Devil, Union, N.J.]) for 5 min each cycle interspersed with 15-min incubations in an ice-water bath. An additional 15 ml of 0.1 M bis-Tris-2 mM DTT buffer (pH 6.5) was added to the tubes, and the cell debris and glass beads were pelleted by centrifugation (13,000 × g, 10 min, 4°C). The supernatants were transferred to clean 40-ml centrifuge bottles and stored at 4°C for <24 h before determination of enzyme activities.

Peptidase specific activities.

Hydrolysis of amino acid-ρ-nitroanilide (amino acid-ρ-NA) substrates at 42°C was done as a discontinuous assay with appropriate dilutions of CFE in 0.1 M bis-Tris-2 mM DTT buffer (pH 6.5). The reactions commenced by addition of the amino acid-ρ-NA substrate (2 mM initial concentration) to 3 ml of a CFE dilution preincubated at 42°C for 2 min. Samples of 0.8 ml were periodically removed and mixed with 0.2 ml of 30% (vol/vol) glacial acetic acid (sample pH ∼3.0) over a 5- to 30-min reaction period. No more than 10% of the initial substrate was hydrolyzed in the samples used for determination of initial velocity. The samples were centrifuged (14,000 × g, 3 min), and the supernatant was transferred to a cuvette for A₄₁₀ determination.

Measurement of protein concentration.

The protein concentration of the CFE and bovine serum albumin standards was determined using a bicinchoninic acid assay kit (Sigma) (30). A modification of the bicinchoninic acid assay by addition of iodoacetamide was used for samples containing thiol agents (16).

Peptidase activity from permeated cells.

An appropriate volume of culture was harvested to obtain an OD₆₀₀ of 5.0 upon resuspension in a 1.0-ml volume. Cells were washed twice, centrifuged (14,000 × g, 1 min), and resuspended in 1.0 ml of 0.1 M bis-Tris-2 mM DTT buffer (pH 6.5). For each reaction, 200 μl of the cell suspension was diluted with 500 μl of 0.1 M bis-Tris-2 mM DTT buffer (pH 6.5) containing 0.1% Triton X-100, mixed, and preincubated for 5 min. The reactions commenced by addition of the amino acid-ρ-NA substrate (2 mM initial concentration). The reactions were stopped by addition and mixing with 300 μl of 30% (vol/vol) glacial acetic acid. The samples were centrifuged (14,000 × g, 3 min), and the supernatant was transferred to a cuvette for A₄₁₀ determination.

Growth rate determinations.

Growth rates were determined in skim milk (pasteurized, double steamed) and amino acid-defined medium (Table 2). Cultures propagated in MRS at 42°C to late exponential phase were washed and resuspended in 0.85% NaCl to reduce the addition of complex medium components in the inoculations. Inoculations were made at 1 × 10⁶ cells/ml into milk and 5 × 10⁶ cells/ml into amino acid-defined medium followed by incubation at 42°C. Samples for pH and OD were taken at 1-h intervals. The pH was determined using a model 410A pH meter (Orion Research, Boston, Mass.) with an Ingold puncture-type pH probe (Mettler-Toledo, Greifensee, Switzerland). The cell density in amino acid-defined medium was determined using a Beckman DU-65 spectrophotometer to monitor OD₆₀₀. Cell suspensions were diluted into the linear range as necessary (OD600, 0.030 to 0.300).

The cell density in skim milk was determined by monitoring the OD₆₀₀ of clarified samples. Briefly, 0.5 ml of skim milk culture was mixed and incubated with 0.5 ml of 2 M borate-200 mM EDTA (pH 8.0) at 55°C for 10 min. The cells were then harvested by centrifugation and washed once with 1.0 ml of 2 M borate-200 mM EDTA (pH 8.0). The cell pellet was washed twice and resuspended with 1-ml samples of 100 mM bis-Tris buffer (pH 6.5). Linear regression analysis of log OD₆₀₀ values versus time was performed to determine slope values with a minimum correlation coefficient (r²) of 0.95 from four duplicate time points. OD₆₀₀ slopes were used to calculate specific growth rates.

Amino acid-supplemented milk.

Supplementation of milk with amino acids was done from both individual and combined stocks. Each of the amino acids listed in Table 2 and used for supplementation was prepared as a 100× stock relative to its concentration in the defined medium. Asp was not included in the supplements, as L. helveticus requires either Asn or Asp for growth. A complete amino acid stock was prepared by equal-volume pooling of all included individual stocks. Amino acid stocks for analysis of growth in the absence of a single amino acid were prepared by pooling all but one of the individual stocks. The mixed stocks were pH adjusted (if necessary) to 6.40 to 6.75 and sterile filtered prior to addition to milk. The amino acid supplements were added to milk to achieve individual concentrations of 0.5× relative to amino acid-defined medium and resulted in amino acid-supplemented milk with an initial pH of 6.55 to 6.63.

RESULTS

Characterization of peptidase deletion vectors.

The sizes of the peptidase gene internal deletions of ΔpepC (366 bp), ΔpepN (272 bp), and ΔpepX (350 bp) were determined by sequencing across the deletion junctions. To confirm the loss of peptidase activity from the deletion derivative vectors, aminopeptidase activities was compared from Triton X-100-permeated cells of E. coli transformed with control plasmids and with subclones containing wild-type or deletion derivatives. The level of aminopeptidase (PepC or PepN) and PepX activity were measured with Lys-ρ-NA and GlyPro-ρ-NA, respectively. The strains with plasmids carrying pepC, pepN, and pepX expressed respective activities 20-, 19-, and 55-fold higher than those of the strains carrying vector alone or carrying a derivative lacking a peptidase gene (data not shown).

Peptidase specific activities of L. helveticus CNRZ32 and its peptidase deletion derivatives.

To quantify and compare the peptidase activities of L. helveticus CNRZ32 and its peptidase deletion derivatives, specific activities were determined from CFE prepared from late logarithmic cultures in amino acid-defined medium (pH 4.7 to 5.0) and in milk medium (pH 4.5 to 4.8). The specific activities for several amino acid-ρ-NA (Lys-ρ-NA, Met-ρ-NA, and Pro-ρ-NA) substrates and a dipeptide ρ-NA (GlyPro-ρ-NA) substrate are shown in Fig. 2. Several other amino acid-ρ-NA substrates (Ala-ρ-NA, Glu-ρ-NA, and Phe-ρ-NA) were also tested (data not shown), and hydrolysis was found to be at intermediate or insignificant levels relative to those of the reported substrates.

FIG. 2.

Specific activities from CFE of Lactobacillus helveticus CNRZ32 and its peptidase mutants cultured in amino acid-defined medium or milk medium. The specific activities are reported as nanomoles of ρ-nitroaniline released per milligram of protein per min at 42°C in 100 mM bis-Tris buffer-2 mM DTT (pH 6.5). All reported values are the averages from two different CFE preparations and analyses. The mean variation for specific activities was ±12% from amino acid-defined medium cultures and ±11% from milk medium cultures. The quantifiable limit was determined to be 1.0 U of activity.

The highest activity was observed with GlyPro-ρ-NA from strains carrying PepX, while the activity for this substrate from ΔpepX strains was below quantifiable limits. In addition, PepX-expressing strains showed two- to threefold-higher activity for GlyPro-ρ-NA when grown in amino acid-defined medium compared to that seen in milk medium.

The highest levels of broad-specificity aminopeptidase activity were observed with Lys-ρ-NA from strains carrying PepN; the activity levels for the ΔpepN strains were reduced to <1% of the highest levels. The aminopeptidase activities for Lys, Met, Ala, and Phe each increased by 30 to 40% in strains JLS243 (ΔpepX) and JLS245 (ΔpepC ΔpepX) relative to that of the wild type when cultured in milk medium. The ΔpepC strains had broad-specificity aminopeptidase activities for Lys-, Met-, Ala-, and Phe-ρ-NA that were indistinguishable from those of their PepC-carrying genotypic parental strains (i.e., JLS241 versus wild type, JLS244 versus JLS242, etc.).

The activities for Glu-ρ-NA were essentially at or below the quantifiable limit (1.0 U) for all strains cultured in both media. The activities for Pro-ρ-NA were very similar for all strains cultured in amino acid-defined medium, but a general decline in activity was measured for multiple mutants from milk medium cultures.

Specific growth rates in amino acid-defined medium and milk.

To examine any effects of peptidase mutations on growth in medium not requiring hydrolysis of exogenous peptides to obtain amino acids, the cell densities and culture pH of L. helveticus and peptidase mutant derivatives were monitored during growth in amino acid-defined medium. The growth rates were very similar for all strains in defined medium (Table 3). The maximum viable counts were ∼5.0 × 10⁸ CFU/ml (OD₆₀₀, ∼2.4 to 2.8; final pH ∼3.6).

TABLE 3.

Growth rates of L. helveticus CNRZ32 and peptidase mutants in milk, amino acid-supplemented milk, and amino acid-defined medium

Medium and strain	Plasmid	Genotypea			Amino acid supplementb	Growth rate (μmax)		No. of trials	Gen. time (min)	% Difference vs. wild type
		pepC	pepN	pepX		Mean	SD
Milk
Wild type	None	+	+	+	None	0.605	0.034	7	69	0
JLS241	None	−	+	+	None	0.583	0.010	3	71	4
JLS242	None	+	−	+	None	0.580	0.038	5	72	4
JLS243	None	+	+	−	None	0.491	0.020	5	85	23
JLS244	None	−	−	+	None	0.426	0.021	6	98	42
JLS245	None	−	+	−	None	0.493	0.020	3	84	22
JLS246	None	+	−	−	None	0.405	0.011	3	103	49
JLS247	None	−	−	−	None	0.350	0.012	6	119	73
JLS244	pTRK-LN	−	(+)	+	None	0.632	0.022	2	66	−4
JLS244	pTRK-LC	(+)	−	+	None	0.610	0.018	2	68	−1
JLS247	pTRK-LX	−	−	(+)	None	0.451	0.018	2	92	34
JLS243	None	+	+	−	I S T P	0.568	0.004	2	73	6
JLS244	None	−	−	+	I S T C	0.604	0.011	2	69	0
JLS247	None	−	−	−	I S T P C N V	0.583	0.047	2	72	4
JLS247	None	−	−	−	I S T Pc	0.408	0.033	2	102	48
JLS247	None	−	−	−	I S T Cc	0.397	0.006	2	105	52
Amino acid-defined medium
Wild type and mutants (JLS241-JLS247)d	None	See above					Complete	0.038	0.005	16	110	60

^aPlasmid-borne peptidase genes are indicated with parentheses.

^bI, Ile; S, Ser; T, Thr; P, Pro; C, Cys; N, Asn; V, Val.

^cAmino acid supplements for JLS243 and JLS244 added to milk for growth analysis to emphasize the effect of the JLS247 supplement on growth rate.

^dGrowth rate, generation time, and percent difference data are means of combined data from all strains tested.

To examine any effects of peptidase mutations on growth in medium requiring hydrolysis of exogenous peptides to obtain amino acids, the cell densities and culture pH of L. helveticus and peptidase mutant derivatives were monitored during growth in skim milk. While the final cell densities were similar for all strains (OD₆₀₀, ∼4.5 to 5.5, 1.6 × 10⁹ to 2.8 × 10⁹ CFU/ml), the specific growth rates differed significantly (Table 3). The most significant decrease in growth rate of the single-peptidase mutant strains was observed with JLS243 (ΔpepX), corresponding to a 16-min (23%) increase in generation time compared to the wild type. The greatest total decrease in growth rate was observed with JLS247 (ΔpepC ΔpepN ΔpepX), corresponding to a 50-min (73%) increase in generation time compared to that for the wild type.

To address the possibility that the lower milk growth rates of the peptidase mutant strains were due to unintended secondary mutations, each of the peptidase genes was reintroduced into either strain JLS244 or JLS247 on pTRK-L2-derived vectors. MRS broth cultures of JLS244(pTRK-LC) and JLS244(pTRK-LN) each expressed activity for Lys-ρ-NA at two- to threefold-higher levels than those of the respective genotypically equivalent strains, JLS242 and JLS241 (data not shown). An MRS broth culture of JLS247(pTRK-LX) expressed activity for GlyPro-ρ-NA at an approximately threefold-higher level than that of the genotypically equivalent strain JLS244. Milk growth rates of the JLS244(pTRK-LC), JLS244(pTRK-LN), and JLS247(pTRK-LX) strains were essentially indistinguishable from those of their genotypic equivalents (JLS242, JLS241, and JLS244, respectively), indicating that the growth rate differences were dependent on the expression of the respective peptidases (Table 3). Additionally, JLS247(pTRK-LN) had milk acidification rates that were indistinguishable from its respective genotypically equivalent strain JLS245 (data not shown).

Determination of amino acids relevant to impaired peptidase mutant growth rates in milk.

To screen for which amino acid is relevant to impaired peptidase mutant growth rates in milk culture, amino acid supplements were added to milk cultures and the acidification rates were determined for L. helveticus strains CNRZ32, JLS243 (ΔpepX), JLS244 (ΔpepC ΔpepN), and JLS247 (ΔpepC ΔpepN ΔpepX). The generation times were calculated using acidification data for amino acid-supplemented milk and the following equation derived from Fig. 3: generation time = [ln2 × (−0.8367) × 60]/(pH slope) − 0.0841. The screening for potentially growth rate-limiting amino acids was done with supplements containing 17 of the 18 total aa used (not including Asp and Glu), containing all 18 aa, or containing no supplements (Fig. 4). Each growth set was performed with L. helveticus CNRZ32 and one of the peptidase mutants.

FIG. 3.

Regression analysis of specific growth rates and acidification slopes in milk for Lactobacillus helveticus CNRZ32 and its peptidase mutants (not inclusive of all data used for growth rate calculations in Table 3). Strain designations are as follows: WT, L. helveticus strain CNRZ32; C, JLS241 (ΔpepC); N, JLS242 (ΔpepN); X, JLS243 (ΔpepX); CN, JLS244 (ΔpepC ΔpepN); CX, JLS245 (ΔpepC ΔpepX); XN, JLS246 (ΔpepX ΔpepN); CNX, JLS247 (ΔpepC ΔpepX ΔpepN).

FIG. 4.

Generation times for Lactobacillus helveticus CNRZ32 and peptidase mutants in milk with amino acid supplements. Columns are labeled with the name of the residue missing from each supplement (and also excluding Asp and Glu; see text). Columns labeled “Comp” were supplemented with 18 aa, not including Asp and Glu. Columns labeled “None” contained no amino acid supplement. Generation times to the right of the dashed vertical lines are ≥5% longer than those for the same strain grown in milk with the complete amino acid supplement.

The generation times for the wild type ranged from 63 to 69 min. In contrast, the generation times for strain JLS247 (ΔpepC ΔpepN ΔpepX) ranged from 74 to 117 min. The most significant effects due to the absence of a single amino acid from the supplement were observed with JLS243 (ΔpepX) lacking Pro and with JLS247 (ΔpepC ΔpepN ΔpepX) lacking Pro or Ile.

Growth of peptidase mutants in milk with minimized amino acid supplements.

The amino acids absent from the supplements resulting in the lowest growth rates (Fig. 4) were added to milk individually and in combinations (in each possible grouping: two-way, three-way, etc.), and the specific growth rates were determined (data not shown). Supplements were evaluated for the accumulated ability to overcome the growth rate limitation compared to complete supplement (18 aa) for the same strain. When a supplement of a given amino acid did not measurably contribute to an increase in growth rate compared to a supplement lacking the same amino acid, it was eliminated from the final minimal amino acid supplement. Since incremental changes in growth rate occurred near the rates obtained for complete supplements, no additional amino acids were included in the test supplements once the generation times were within 5% of those of the same strain with the complete amino acid supplement. Similarly, no additional amino acids were evaluated when they could not reasonably be chosen based on marginal differences in generation times from the initial screening (Fig. 4). In this manner, a minimal amino acid supplement was chosen for each of the strains (Table 3).

The growth rate-enhancing amino acid supplement that met the above criteria for strain JLS243 (ΔpepX) was determined to be that including Ile, Ser, Thr, and Pro (Table 3). Likewise, the minimal supplement for JLS244 (ΔpepC ΔpepN) was determined to be that containing Ile, Ser, Thr, and Cys. The milk growth rate of JLS247 (ΔpepC ΔpepN ΔpepX) was determined with the above amino acid combinations and a supplement of Ile, Ser, Thr, Pro, Cys, Asn, and Val. The minimal amino acid supplements in milk for each of the mutants resulted in growth rates similar to that of the wild type without supplements.

DISCUSSION

L. helveticus CNRZ32 is auxotrophic for multiple amino acids and therefore depends on their acquisition from its environment, either as free amino acids or though the transport of peptides and subsequent hydrolysis. The present study demonstrated that the loss of PepC, PepN, and PepX activities resulted in significant impairment of growth rate in milk but that growth rates for mutants in defined amino acid medium are unaffected relative to those for the wild type. This indicates that these peptidases are important in the pathway of hydrolysis of peptides derived from milk proteins. In addition, our data indicate that supplementation of milk with a limited subset of amino acids is adequate to overcome the impaired growth rates of the mutants.

Measurement of peptidase activities confirmed the expected phenotypic changes due to the peptidase mutations and indicated that some were affected by the form of nitrogen (free amino acids versus peptides) available from the medium. As measured by hydrolysis of GlyPro-ρ-NA, PepX activity was found to be two- to threefold higher in amino acid-defined medium than in milk medium for L. helveticus strain CNRZ32 (Fig. 2C). This result suggests that PepX expression in L. helveticus CNRZ32 is regulated by the concentration of free amino acids or peptides. Studies of peptidase activity in Lactococcus have shown PepX and PepN expression to be affected by the source of nitrogen (20, 21). PepX expression was found to differ among strains and was growth rate dependent in chemostatic culture, though all variations were within a twofold difference. In addition, the peptide Pro-Leu, previously reported to regulate PrtP expression (19), was also shown to affect PepX and PepN expression (20).

The majority of aminopeptidase activity measured with Lys-, Ala-, Met-, and Phe-ρ-NA in L. helveticus strain CNRZ32 is attributable to PepN (data not shown for Ala- and Phe-ρ-NA). Pro-ρ-NA has been shown to be efficiently hydrolyzed by PepI from L. helveticus and L. delbrueckii (13, 31; this work). Significant activity was measured for Met-ρ-NA in mutant strains lacking both PepN and PepC (JLS244 and JLS247), which indicates the presence of at least one additional aminopeptidase. All the strains had similar activity for Pro-ρ-NA when grown in amino acid-defined medium, indicating that PepN and PepC do not contribute significantly to its hydrolysis. Aminopeptidase activity attributable to PepC was near or below the limit of detection for most assay conditions and substrates. Transformants of E. coli expressing L. helveticus PepC indicated this enzyme has relatively high activity for Lys-ρ-NA compared to those of other chromogenic amino acid substrates (25, 32; this work). Despite this specificity for Lys-ρ-NA, only a threefold difference in aminopeptidase activity was measured between JLS242 (ΔpepN) and JLS244 (ΔpepC ΔpepN) near the limit of detection from permeated cells of early-stationary-phase culture in MRS broth (data not shown). While this confirmed the loss of PepC activity from the respective mutants, PepC activity was not discernible from activities measured in CFE of L. helveticus (Fig. 2). These results indicate that PepC of L. helveticus is expressed at a low level compared to PepN, and at least one other aminopeptidase, under the growth conditions employed.

The growth rates of the peptidase mutants in amino acid-defined medium were evaluated to determine whether any of the peptidases had roles in hydrolysis of endogenous proteins (maturation and/or general turnover) that were essential. Relative to the wild type, no growth impairment in amino acid-defined medium was observed for any of the peptidase mutant strains (Table 3). This indicates that PepC, PepN, and PepX do not have critical functions individually or cooperatively when L. helveticus CNRZ32 is grown in medium rich in free amino acids. These results are similar to those reported for multiple-peptidase mutants of Lactococcus grown in complex medium (22). Relatively slow growth rates in amino acid-defined medium are most likely due to amino acid transport requirements. The uptake and subsequent hydrolysis of peptides are more energetically favorable than the transport of free amino acids (especially Glu/Gln, Asp/Asn, and Pro). This is apparent in the same defined medium supplemented with peptides, which results in growth rates of the wild type similar to those observed in milk (data not shown).

The growth rates of the peptidase mutants in milk were evaluated to determine whether any of the peptidases have crucial roles in the hydrolysis of peptides derived from exogenous proteins. The deletion of pepX resulted in the most significant decrease in growth rate for a single-peptidase mutant, suggesting that access to one or more amino acids is limited in the absence of PepX activity. PepX has previously been suggested to have a key role in the hydrolysis of β-casein-derived peptides, which are relatively high in Pro residues (3, 29, 33, 34). The generation times of strains JLS243 (ΔpepX) and JLS247 (ΔpepC ΔpepN ΔpepX) in milk supplemented with all of the amino acids except Pro (Fig. 4) suggest that this amino acid is limiting for growth, while the absence of Pro from the amino acid supplement to milk did not affect the generation times of the wild type or JLS244 (ΔpepC ΔpepN). In addition, the growth rate for JLS243 in milk with the addition of a specific amino acid supplement (Ile, Ser, Thr, or Pro; Fig. 4) was increased but the individual addition of Pro was responsible for the most significant increase during the screening (data not shown). This result is somewhat surprising, as PepX has generally been viewed as representing a means for exposure of the remaining peptide to hydrolysis by aminopeptidases through removal of the Xaa-Pro dipeptide, whereas our results suggest that PepX is important for acquisition of Pro for growth. This implies a key role in the proteolytic pathway for the L. helveticus CNRZ32 prolidase (PepQ), which would subsequently hydrolyze the Xaa-Pro dipeptides to liberate Pro (23; G. Ü. Yüksel, J. E. Christensen, and J. L. Steele, unpublished data). This is consistent with results indicating that a ΔpepQ strain of L. helveticus CNRZ32 has growth rate reduction in milk representing a 13% increase in generation time (Yüksel et al., unpublished). In addition, reintroduction of pepN to JLS247 on the multicopy vector pTRK-LN resulted in broad-specificity aminopeptidase activity two- to threefold higher than that seen for the wild type, but the milk growth rate corresponded to that of JLS245 (ΔpepC ΔpepX). This indicates that an elevated level of PepN activity is insufficient to compensate for the loss of PepX activity.

A previous study of Lactococcus evaluated growth rates in milk and accumulated intracellular amino acids in the wild type and the multiple-peptidase mutants (22). The accumulated intracellular amino acid data for the multiple-peptidase mutants lacking PepX indicated that the most significant decrease in concentration was for Pro (8- to 18-fold reduction in mutants). However, since the critical concentration of intracellular amino acids for optimal growth is not known, it is not possible to know which of the pool reductions were significant for growth rate limitation. Also, since the analysis was only performed on multiple mutants (no fewer than four deleted peptidases), it is not possible to attribute the amino acid pool decrease to a particular peptidase or peptidases.

The milk growth rates of strains JLS241 (ΔpepC) and JLS242 (ΔpepN) did not differ significantly from that of L. helveticus CNRZ32. The minor growth rate impairment observed with JLS242 (ΔpepN) contrasts with the high level of PepN activity (Fig. 2). However, the combined loss of these aminopeptidases in JLS244 (ΔpepC ΔpepN) significantly reduced the growth rate (Table 3). This result suggests that PepC and PepN compensate for each other for hydrolysis of peptides derived from milk proteins.

Growth of JLS244 in milk with a specific amino acid supplement (Ile, Ser, Cys, and Thr) resulted in generation times similar to that for the wild type (Table 3). Additionally, the inverse supplement, containing the 14 aa of least apparent need (i.e., not including Ile, Ser, Cys, and Thr), did not result in an increased milk growth rate relative to that of unsupplemented milk (data not shown). The specific amino acid supplement for strain JLS247 included Asn and Val in addition to the specific amino acid supplements for JLS243 (ΔpepX) and JLS244 (ΔpepC ΔpepN). Evaluation of the specific supplements for JLS243 (Ile, Ser, Thr, and Pro) and JLS244 (Ile, Ser, Thr, and Cys) indicates that neither of these additions alone was adequate to completely overcome the milk growth rate deficiency of JLS247 (Table 3).

The increased growth rate of a peptidase mutant supplemented with the nonessential Cys and Ser may be the result of a decreased energetic requirement for amino acid biosynthesis. In addition, biosynthesis would require the catabolism of other amino acids, diverting the pool of other amino acids that may already be near a limiting concentration. The availability of Ser from caseins may be complicated by the frequency with which these residues are phosphorylated. To our knowledge, the capability of L. helveticus to hydrolyze and/or de-phosphorylate Ser residues has not been investigated. An additional draw on available Ser may also occur through the required biosynthesis of Gly and Cys, both of which are present at low levels in casein. The possibility that supplemented Cys is beneficial due to a reducing effect in the medium cannot be completely discounted, although the absence of Cys from supplements did not result in a significant growth rate decrease for JLS243 and JLS247 (Fig. 4).

Six of the essential amino acid (Thr, Ile, Val, Met, Leu, and Phe) concentrations were determined to be too low to support growth of L. helveticus CNRZ32 in UFMP (data not shown). This observation accentuates the dependence on the proteolytic system for obtaining specific essential amino acids from milk proteins. Decreased growth rates of the peptidase mutants may be due to inadequate “decapping” of peptides to allow subsequent hydrolysis and liberation of required amino acids. In the absence of a given peptidase, other aminopeptidases may accomplish the required hydrolysis or more-extensive hydrolysis of peptides may occur by endopeptidases with subsequent hydrolysis by di- and tripeptidases. Specific activities of PepN and PepC for amino acid-ρ-NA substrates were found to have little predictive value for amino acid supplements that were determined to enhance growth rates of the respective peptidase mutants in milk.

While the complexity and diversity of proteolytic systems in LAB are well established (6, 18), the potential for compensatory regulation that affects growth rate in milk (or any protein source) has not been thoroughly investigated (22). However, the transcription of several genes of the lactococcal proteolytic system (prtP, opp-pepO1, pepN, and pepC) has been shown to be negatively regulated by dipeptides in the growth medium (14). In addition, another report indicates that a lactococcal CodY homologue represses expression of opp-pepO1 in the presence of elevated levels of branched-chain amino acids (Ile, Leu, and Val), suggesting that CodY is a pleiotropic repressor (15).