Antonie van Leeuwenhoek 76: 217–246, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.

217

Peptidases and amino acid catabolism in lactic acid bacteria Jeffrey E. Christensen2 , Edward G. Dudley2, Jeffrey A. Pederson1 & James L. Steele1,∗ 1 Department

of Food Science, 2 Department of Bacteriology, University of Wisconsin-Madison, Madison, WI 53706-1565, USA (∗ Author for correspondence; E-mail: [email protected]) Key words: amino acid catabolism, lactic acid bacteria, peptidases, physiology

Abstract The conversion of peptides to free amino acids and their subsequent utilization is a central metabolic activity in prokaryotes. At least 16 peptidases from lactic acid bacteria (LAB) have been characterized biochemically and/or genetically. Among LAB, the peptidase systems of Lactobacillus helveticus and Lactococcus lactis have been examined in greatest detail. While there are homologous enzymes common to both systems, significant differences exist in the peptidase complement of these organisms. The characterization of single and multiple peptidase mutants indicate that these strains generally exhibit reduced specific growth rates in milk compared to the parental strains. LAB can also catabolize amino acids produced by peptide hydrolysis. While the catabolism of amino acids such as Arg, Thr, and His is well understood, few other amino acid catabolic pathways from lactic acid bacteria have been characterized in significant detail. Increasing research attention is being directed toward elucidating these pathways as well as characterizing their physiological and industrial significance.

Introduction Lactic acid bacteria (LAB) are commonly defined as Gram-positive, nonsporulating, catalase-negative, aerotolerant, acid tolerant, nutritionally fastidious, strictly fermentative organisms that lack cytochromes and produce lactic acid as the major end-product of carbohydrate metabolism (Axelsson 1998). The genera typically included are Aerococcus, Carnobacterium, Enterococcus, Lactobacillus, Lactococcus, Leuconostoc, Oenococcus, Pediococcus, Streptococcus, Tetragenococcus, Vagococcus, and Weissella (Axelsson 1998). The conversion of peptides to free amino acids and the subsequent utilization of these amino acids is a central metabolic activity in LAB. The peptidase system is involved in the hydrolysis of exogenous peptides to obtain essential amino acids for growth and hydrolysis of peptides formed by housekeeping proteinases. The amino acids formed by this system can be utilized for processes such as protein synthesis, generation of metabolic energy, and recycling of reduced cofactors. This review will cover our current

knowledge of the enzymes and metabolic pathways involved in peptide and amino acid catabolism in LAB, with a focus on LAB that are associated with food fermentations.

Peptidases of lactic acid bacteria The lactococci and lactobacilli associated with dairy environments generally have between six and fourteen amino acid auxotrophies. The ability of LAB to grow to high cell density in milk is dependent on a proteolytic system that can liberate essential amino acids from casein derived peptides. Although this review will not elaborate on the proteinase and peptide transport activities discussed in previous reviews (Kunji et al. 1996; Law & Haandrikman 1997), the relevance of recent articles concerning this function of the proteolytic system will be addressed. The PrtP hydrolysis and Opp transport of peptides derived from purified β-casein by Lc. lactis has been investigated by subtractive evaluation of the remaining media peptides and accumulation of intracellular

218 peptides and amino acids (Kunji et al. 1998). The results indicate that most of the essential amino acids can be obtained from transported peptides derived from the C-terminal end (residues 161–191) of purified βcasein. The exception was His, which was previously shown to be a required supplement for growth on purified β-casein (Kunji et al. 1995). Opp was shown to transport from ten to fourteen of the β-casein derived peptides ranging in size from five to ten residues. However, subsequent kinetic analysis of Opp indicates that peptides from four to at least eighteen residues can be transported with little specificity for particular side chains (Detmers et al. 1998). These studies begin to define the potential peptidase substrates available to Lc. lactis during growth in milk. Peptidase specificity The available substrate hydrolysis information for peptidases was tabulated (see Table 1 references) in order to obtain a more comprehensive view of the specificity of a given peptidase. The cumulative data is comprised of a variety of amino acid ρ-nitroanilide (−ρNA) and β-naphthilamide (−βNAP) substrates, ∼180 dipeptides, ∼100 tripeptides, and ∼40 oligopeptides (data not shown). The amino acids of the natural (non-chromogenic) substrates were classified according to their characteristics (i.e. hydrophilicity, charge, side chain grouping). The data was sorted with regard to the order and class of the amino acids, resulting in the grouping of peptides with common features. For example, Ala-Phe and Leu-Tyr were grouped together (hydrophobic-aromatic), as were Lys-Pro and His-Pro (basic-imino). The substrate hydrolysis generalizations that were notable are included with the descriptions of individual peptidases. While the identification and classification of peptidases according to the hydrolysis of chromogenic substrates or small peptides is convenient, this information does little to elucidate the specific role an enzyme may have in hydrolysis of exogenous peptides or general cellular physiology. Aminopeptidase C The broad specificity aminopeptidase C (PepC) has been purified and characterized from strains of Lb. delbrueckii, Lb. helveticus, Lc. lactis and S. thermophilus (see Table 1) (Neviani et al. 1989; Wohlrab & Bockelmann 1993; Chapot-Chartier et al. 1994; Mistou et al. 1994; Fernández de Palencia 1999). Significant activity of PepC on AA-βNAP substrates is reported

for residues that are basic (Arg, His, and Lys), acidic (Glu and Asp), hydrophobic/uncharged (Ala and Leu), and aromatic (Phe). A similar level of activity is found for the corresponding AA-ρNA substrates, including Met- and Gly-ρNA. No PepC activity is detectable for Pro-ρNA, Pro-βNAP, Xaa-Pro-ρNA, or Xaa-ProβNAP substrates. Hydrolysis is also reported for a variety of di- and tripeptides with uncharged or basic residues in the amino terminal position. A systematic evaluation of Lc. lactis PepC catalytic properties indicates the enzyme contains an extended binding site (Mistou & Gripon 1998). PepC was determined to have preference for a hydrophobic residue in the P0 1 (Gly-Xaa) and P0 2 (Gly-Gly-Xaa) position of the respective substrates. Analysis of various di-, tri-, and tetrapeptides revealed an increasing affinity (decreasing Km ) with increasing substrate amino acid residues. Additionally, a comparison of enzymatic parameters measured for polyglycine substrates (Glyn where n=2 through 5) demonstrated the dependence of catalytic activity on substrate length. The highest substrate affinity and hydrolysis were measured with the tetrapeptides Gly-Gly-Phe-Leu and Gly-Gly-Gly-Ala. Structural modeling of the yeast bleomycin hydrolase and Lc. lactis PepC suggest that the carboxyterminal residues are inserted into the active site toward the catalytic residues of these enzymes, permitting interaction between the enzyme α-carboxy group and the substrate α-amino group (Joshua-Tor et al. 1995; Mata et al. 1997). Characterization of a site-directed mutant enzyme, missing the C-terminal residue (PepC-1Ala435), confirmed dependence on this residue for strict aminopeptidase activity (Mata et al. 1997). Northern analysis with a pepC (∼1.3-kb ORF) specific probe identified two transcript sizes, ∼1.7 and ∼2.7-kb, from Lb. helveticus grown in MRS broth under pH controlled conditions (5.85) (Vesanto et al. 1994). Both pepC-hybridized transcripts were present at late exponential phase, but below detectable levels at stationary phase. Nucleotide sequence analysis of the regions proximal to pepC identified a downstream ORF (840-bp) that encodes a putative 30 kDa transmembrane protein with 40.4% amino acid identity to a Staphylococcus xylosus glucose uptake protein (Brückner 1998). Northern analysis with a probe specific to the downstream ORF revealed co-transcription with pepC during late exponential phase, while at stationary phase the ORF is transcribed from its own promoter resulting in a 1.0-kb monocistronic tran-

219 Table 1. Peptidases of lactic acid bacteria Strain

kDaa

Structureb

Classc

Motifsd

Lb. delbrueckii Lb. delbrueckii Lb. helveticus Lb. helveticus Lb. helveticus Lc. lactis Lc. lactis Lc. lactis S. thermophilus

54H 51N 49N 51N 50S 50H 50N 50 50N

tetramer

C C C C C C C

Wohlrab & Bockelmann (1993) PS00139e Klein et al. (1994a) PS00139, PS00639 Fern´andez et al. (1994) PS00139, PS00639 Vesanto et al. (1994) Fern´andez de Palencia (1999) Neviani et al. (1989) PS00139, PS00639 Chapot-Chartier et al. (1993) Mistou et al. (1994) PS00139, PS00639 Chapot-Chartier et al. (1994)

Lb. casei Lb. casei Lb. rhamnosus Lb. delbrueckii Lb. delbrueckii Lb. delbrueckii Lb. helveticus Lb. helveticus Lb. helveticus Lb. helveticus Lb. helveticus Lb. helveticus Lb. lactis Lb. sanfrancisco Lc. lactis Lc. lactis Lc. lactis Lc. lactis Lc. lactis Lc. lactis S. thermophilus S. thermophilus S. thermophilus S. thermophilus

95S 87S 89S 98S 95S 95N 97S 96N 97S 92S 95S 96N 78S 75S 95H 95S 95N 95S 95N 85H 92S 97S 98S 89S

monomer monomer monomer monomer

Unclassified

Lb. acidophilus Lb. delbrueckii Lb. casei Lc. lactis

38S 32S 30S 36S

tetramer octamer tetramer

M M S M

Machuga & Ives (1984) Wohlrab & Bockelmann (1994) Fern´andez de Palencia et al. (1997) Geis et al. (1985)

PepA

Lc. lactis Lc. lactis Lc. lactis Lc. lactis S. thermophilus

40S 43S 38N 41S 45S

hexamer trimer

M M

hexamer octamer

M M

Bacon et al. (1994) Exterkate & De Veer (1987) l’Anson et al. (1995) Niven (1991) Rul et al. (1995)

Lb. acidophilus Lb. casei Lb. delbrueckii Lb. delbrueckii Lb. delbrueckii

95S 79S 95S 82S 90S

dimer monomer dimer

S S S S S

Bockelmann et al. (1991) Habibi-Najafi & Lee (1994) Bockelmann et al. (1991) Atlan et al. (1990) Miyakawa et al. (1991)

Peptidase Aminopeptidase

Aminopeptidase

Aminopeptidase

XPDAP

PepC

PepN

PepX

tetramer hexamer hexamerX hexamer

monomer monomer monomer monomer monomer monomer monomer

C M M M M M M M M M M M M M M M

PS00142 PS00142

PS00142

monomer monomer

monomer monomer monomer monomer

trimer

M M M M M M M M

PS00142 PS00142

Reference

Fern´andez de Palencia et al. (1997) Arora & Lee (1992) Arora & Lee (1994) Tsakalidou et al. (1993) Bockelmann et al. (1992) Klein et al. (1993) Khalid & Marth (1990a) Christensen et al. (1995) Blanc et al. (1993) Miyakawa et al. (1992) Sasaki et al. (1996) Varmanen et al. (1994) Eggimann & Bachmann (1980) Gobbetti et al. (1996) Exterkate et al. (1992) Van Alen-Boerrigter et al. (1991) Strøman (1992) Tan & Konings (1990) Tan et al. (1992) Desmazeaud & Zevaco (1979) Tsakalidou & Kalantzopoulos (1992) Rul et al. (1994) Midwinter & Pritchard (1994) Tsakalidou et al. (1993)

220 Table 1. Continued Peptidase

Strain

kDaa

Structureb

Classc

Motifsd

Reference

88N

monomer monomer

S S S S S S S

361-GRSYLG

Lb. delbrueckii Lb. helveticus 72S Lb. helveticus 90N Lb. helveticus 87S Lb. helveticus 91N Lb. lactis 90S Lc. lactis 117H Lc. lactis 88S Lc. lactis 88N Lc. lactis 85S Lc. lactis 88S Lc. lactis 90S Lc. lactis 88N Lc. lactis 82S S. thermophilus 80S S. thermophilus 80S Dipeptidase

Dipeptidase

Pro-iminopeptidase

PepD

PepV

PepI

Pro-iminopeptidase Prolidase

Prolinase

PepQ

PepR

dimer dimer dimer

S S S

dimer dimer dimer

S S S

Meyer-Barton et al. (1993) Khalid & Marth (1990b) Yüksel & Steele (1996) Miyakawa et al. (1994) Vesanto et al. (1995) Meyer & Jordi (1987) Booth et al. (1990b) Chich et al. (1995) Nardi et al. (1991) Zevaco et al. (1990) Yan et al. (1991) Kiefer-Partsch et al. (1989) Mayo et al. (1991) Lloyd & Pritchard (1991) Tsakalidou et al. (1998) Meyer & Jordi (1987) Dudley et al. (1996) Vesanto et al. (1996)

monomer dimer dimer

346-GKSYLG

346-GKSYLG

53N 53N

octamer

C

Lb. casei

46S

monomer

M

Lb. delbrueckii 51S Lb. delbrueckii 52N Lb. helveticus 51N Lb. helveticus 50S Lb. sake 50S Lb. sanfrancisco 65S Lc. lactis 100H Lc. lactis 51N Lc. lactis 49S Lc. lactis 50H

monomer monomer

M M

33N 34S 33N 34N

Lc. lactis

361-GRSYLG

dimerX

Lb. helveticus Lb. helveticus

Lb. delbrueckii Lb. delbrueckii Lb. delbrueckii Lb. helveticus

361-GRSYLG

monomer monomer

monomer

M M M M M M M

dimer

S S S S

50S

dimer

M

Lb. casei Lb. delbrueckii Lb. delbrueckii Lb. delbrueckii Lb. helveticus Lc. lactis Lc. lactis

41S 41N 41N 41N 41N 42H,D 43H,D

monomer dimer

M

Lb. helveticus Lb. helveticus Lb. helveticus Lb. rhamnosus Lb. curvatus

35N 33S 35N 34N 32S

trimer

M M

PS00758, PS00759 PS00758, PS00759

PS00758, PS00759

105-GQSWGG 104-GQSWGG 103-GQSWGG

PS00491 PS00491 PS00491 f

109-GQSWGG S 109-GQSWGG 107-GQSWGG dimer

Atlan et al. (1994) Gilbert et al. (1994) Klein et al. (1994b) Varmanen et al. (1996a) Baankreis & Exterkate (1991)

M M

tetramer

Fern´andez-Espl´a & Mart´in-Hern´andez (1997) Wohlrab & Bockelmann (1992) Vongerichten et al. (1994) Yüksel & Steele (1997b) Tan et al. (1995) Montel et al. (1995) Gobbetti et al. (1996) Hwang et al. (1981) Hellendoorn et al. (1997) Van Boven et al. (1988) Desmazeaud & Zevaco (1977)

Fern´andez-Espl´a et al. (1997) Morel et al. (1999b) Rantanen & Palva (1997) Stucky et al. (1995) Yüksel & Steele (1997a) Booth et al. (1990a) Kaminogawa et al. (1984) Dudley & Steele (1994) Shao et al. (1997) Varmanen et al. (1996b) Varmanen et al. (1998) Magboul & McSweeney (1998)

221 Table 1. Continued Strain

kDaa

PepL

Lb. delbrueckii

35N

PepP

Lc. lactis Lc. lactis

43S 46N

monomer

Lc. lactis Lc. lactis Lc. lactis Lb. delbrueckii Lb. delbrueckii Lb. sake Lc. lactis Lc. lactis P. pentosaceus

55S 46N 52S 29S 38S 55S 23S 75H 45S

dimer

PepE

Lb. helveticus

52N

C

PS00139, PS00639

Fenster et al. (1997)

PepG

Lb. delbrueckii

50N

C

PS00139, PS00639

Klein et al. (1997)

PepO

Lb. helveticus Lc. lactis Lc. lactis Lc. lactis Lc. lactis

71N 71N 71N 70S 71N

M M M M

PS00142 PS00142 PS00142

Chen & Steele (1998) Lian et al. (1996) Mierau et al. (1993) Pritchard et al. (1994) Tynkkynen et al. (1993)

PepF1 PepF1 PepF2

Lc. lactis Lc. lactis Lc. lactis

70N 70N 70N

monomer

Unclassified

Lb. delbrueckii 68S Lb. paracasei 30S Lc. lactis 98S Lc. lactis 180 Lc. lactis 40S Lc. lactis 70S Lc. lactis 70S Lc. lactis 52S Lc. lactis 93H Lc. lactis 70S Lc. lactis 140H Lc. lactis 50H

monomer multimer monomer multimer dimer monomer monomer multimer

Peptidase

Tripeptidase

Tripeptidase

Endopeptidase

Endopeptidase

Endopeptidase

Endopeptidase

PepT

Unclassified

Structureb

Classc

Motifsd

Reference

S

105-GHSWGG

Klein et al. (1995)

M M

PS00491

Mars & Monnet (1995) Matos et al. (1998)

M PS00758, PS00759

dimer trimer dimer monomer trimer dimer

monomer

M M M M M M

PS00142 M

g g

PS00142

monomer

M M M M M M M M M M M M

Bacon et al. (1993) Mierau et al. (1994) Bosman et al. (1990) Bockelmann et al. (1995) Bockelmann et al. (1997) Sanz et al. (1998) Sahlstrøm et al. (1993) Desmazeaud & Zevaco (1979) Simitsopoulou et al. (1997)

Monnet et al. (1994) Nardi et al. (1997) Nardi et al. (1997) Bockelmann et al. (1996) Tobiassen et al. (1997) Yan et al. (1987b) Baankreis et al. (1995) Yan et al. (1987a) Baankreis et al. (1995) Stepaniak & Fox (1995) Stepaniak et al. (1998) Muset et al. (1989) Tan et al. (1991) Ohmiya & Sato (1975) Desmazeaud & Zevaco (1976)

a Molecular weight of a subunit calculated from the derived amino acid sequence (N) when nucleotide sequence was available; in other cases individual subunit size is reported from SDS-PAGE (S) or gel filtration (H), with preference given to SDS-PAGE values when both are reported. b Quaternary structure predicted from comparison of gel filtration value with either derived amino acid sequence or SDS-PAGE values; X-ray crystal analysis (X). c Catalytic class of enzyme determined biochemically or by sequence motif identification: C, cysteine-peptidase; M, metallopeptidase; or S, serine-peptidase. d Active site motif for prolyl oligopeptidase family (see Rawlings et al. 1991) (residue position-GxSxxG) or PROSITE pattern identification determination (Hofmann et al. 1999): PS00139, Thiol Protease Cys; PS00639, Thiol Protease His; PS00142, Neutral zinc metallopeptidases, zinc-binding region signature; PS00758, ArgE/dapE/ACY1/CPG2/yscS family signature 1; PS00759, ArgE/dapE/ACY1/CPG2/yscS family signature 2; and PS00491, Aminopeptidase P and proline dipeptidase signature. e The sequence of Lb. delbrueckii contains a single residue mismatch to the current PROSITE Thiol Protease His signature. f The sequence of Lb. helveticus contains a one residue mismatch to the current PROSITE Aminopeptidase P and proline dipeptidase signature. g The sequence encoded by pepF1 contains two adjacent residue mismatches to the current PROSITE Neutral zinc metallopeptidases (zinc-binding region) signature. However, the same region is coincidently overlapped by PROSITE Signal peptidases I serine active site signature.

222 script. Currently, there is no information available on the transcriptional regulation of pepC for LAB grown in milk. The physiological role of PepC in Lc. lactis has been investigated for deletion mutant derivatives as a function of specific growth rate (Mierau et al. 1996a). The growth rate of the PepC deletion mutant was reported to be the same as for the wild type strain in complex media (GM17 broth) while in milk a subtle decrease in growth rate is observed (µmax 0.60 vs 0.65 h−1 ), corresponding to a 10% longer generation time (Mierau et al. 1996a). Aminopeptidase N The broad specificity aminopeptidase N (PepN) has been purified and characterized from strains of Lb. casei, Lb. delbrueckii, Lb. helveticus, Lc. lactis and S. thermophilus (see Table 1). In general, the activity of PepN on AA-ρNA substrates indicates the highest specificity for the basic amino acids Lys and Arg, followed by the hydrophobic/uncharged residues Leu and Ala. Significant activity is also observed for Met and Phe-ρNA, while in general low to undetectable activity is reported for Asp-, Glu-, and Gly-ρNA. Niven et al. (1995) reported a general increase in activity with increase in the hydrophobicity of the carboxyterminal residue of an Arg-Xaa dipeptide. Also, a trend of increasing Km and Vmax with peptide chain length was observed for Lys-Phe-(Gly)n substrates, with the highest specific activity for the Lys-Phe(Gly)4 hexapeptide. Several tryptic digest derived oligopeptides (7–16 residues) of β-casein were at least partially hydrolyzed by purified PepN, demonstrating the capacity to liberate residues from even larger substrates (Tan et al. 1993). Additionally, a study using cell-free extracts prepared from peptidase mutants of Lb. helveticus has shown that PepN is predominant in the liberation of amino-terminal Tyr residue from βcasein f193-209 (Christensen & Steele 1998a). These results accentuate the range of peptide size and conditional specificity of PepN beyond the amino terminal residue of the substrate. Northern analyses with pepN specific probes derived from Lb. helveticus and Lc. lactis suggest that the transcripts from the genes are monocistronic (Strøman 1992; Varmanen et al. 1994). In Lb. helveticus, grown in MRS broth under pH controlled conditions (5.85), the level of transcription of pepN is reported to remain constant through exponential and stationary phase (Varmanen et al. 1994). However, evaluation

of the data provided in the publication figure indicates the ratio between pepN mRNA (dpm) and cell density nearly doubles in the two generations preceding the onset of stationary phase. Similarly, the ratio of total aminopeptidase activity to message level increases nearly six-fold in the same time frame. The fact that PepN activity remains constant 12 h into stationary phase while pepN transcript levels decrease suggests the enzyme is relatively stable and/or there is an elongated balance of turnover and expression under pH controlled conditions. In contrast, a study of PepN activity from MRS broth grown Lb. helveticus without pH control indicates a decrease in PepN activity from end of exponential phase (pH ∼4.8) and reduction by four-fold during transition to stationary phase (Christensen & Steele 1998b). The activity of Lb. helveticus PepN accounts for greater than 99% of the hydrolysis of the substrates (Lys-, Leu-, Met-, and Ala-ρNA) used to measure total aminopeptidase activity in the previously mentioned studies (Christensen & Steele 1996). The regulation of PepN expression in Lc. lactis has been demonstrated to be strain dependent (Meijer et al. 1996). Total aminopeptidase activity is equivalent or higher in cultures grown in milk compared to complex amino acid or peptide-containing media (Meijer et al. 1996). Additionally, the dipeptide Pro-Leu is reported to decrease PepN and PrtP expression in Lc. lactis MG1363 (Marugg et al. 1995). Currently, there is no information available on the transcriptional regulation of pepN for LAB grown in milk. The physiological role of PepN in Lb. helveticus and Lc. lactis has been investigated for deletion mutant derivatives as a function of specific growth rate. The growth rate and final cell densities of Lb. helveticus PepN deletion mutants are the same as for the wild type strain in complex media (MRS broth) and complete amino acid defined media (Christensen & Steele 1996). The growth rate in milk was slightly slower for the PepN deletion mutant compared to wild type (µmax 0.56 vs 0.60 h−1 ), corresponding to a 7% longer generation time, but the final cell densities were similar. Therefore, despite the Lb. helveticus auxotrophy for fourteen amino acids, the remaining peptidases are capable of liberating essential amino acids from milk derived peptides at a velocity that nearly maintains the parental growth rate. The growth rate of a Lc. lactis PepN deletion mutant was reported to be the same as for the wild type strain in complex media (GM17 broth) (Mierau et al. 1996a). However, the Lc. lactis PepN mutant dis-

223 plays a decreased growth rate in milk corresponding to a 25% longer generation time compared to wild type (µmax 0.52 vs 0.65 h−1 ) (Mierau et al. 1996a; Mierau et al. 1996b). These results suggest that PepN plays a more important role in hydrolysis of milk derived peptides in Lc. lactis than in Lb. helveticus. Aminopeptidase A Aminopeptidase A (PepA), also referred to as glutamyl aminopeptidase, has been purified and characterized from several strains of Lc. lactis (Exterkate & De Veer 1987; Niven 1991; Bacon et al. 1994) and S. thermophilus (Rul et al. 1995). PepA activity is essentially defined by the characteristic hydrolysis of N-terminal acidic amino acid residues. The enzyme hydrolyses Glu- and Asp-ρNA, but low activity was reported for Glu- and Asp-βNAP with PepA of S. thermophilus (Rul et al. 1995). Hydrolysis of Glu-Xaa and Asp-Xaa dipeptides is observed when the C-terminus amino acid is basic (-Lys), uncharged (-Gly), hydrophobic/uncharged (-Ala, -Leu), polar/uncharged (-Ser), or aromatic (-Phe, -Tyr). The dipeptides SerXaa are also hydrolyzed when the C-terminus amino acid is hydrophobic/uncharged (-Ala, -Leu) or aromatic (-Phe, -Tyr). PepA is also capable of releasing acidic residues from peptides as large as a decapeptide (Bacon et al. 1994). Significant hydrolysis is not observed for di- and tripeptides with N-terminal amino acids from any other class. The physiological role of PepA in Lc. lactis has been investigated for insertionally inactivated pepA deficient derivatives as a function of growth in milk (l’Anson et al. 1995). The hydrolytic activity for Glu-ρNA from cell extracts of three pepA derivatives was reported to be deficient relative to wild type. The insertion derivative strains displayed a statistically significant increase in lag phase, equivalent to a halfgeneration, but grew to the same final cell densities as wild type. However, the cell density and milk acidification data indicates that the pepA derivatives had the same growth and acidification rates as the parental strain. X-prolyl dipeptidyl aminopeptidase X-prolyl dipeptidyl aminopeptidase, PepX, has been purified and characterized from strains of Lb. acidophilus, Lb. delbrueckii, Lb. casei, Lb. helveticus, Lc. lactis and S. thermophilus (see Table 1). PepX liberates Xaa-Pro dipeptides from the N-terminus of peptides. In addition to peptidase activity, PepX has

been demonstrated to execute amidase and esterase reactions (Houbart et al. 1995; Yoshpe-Besançon et al. 1994). The highest activities reported for PepX on Xaa-Pro-ρNA substrates are when the N-terminal residues are uncharged (Ala-, Gly-) or basic (Arg-). While liberation of amino acids from dipeptides does not occur, PepX releases Xaa-Pro dipeptides from substrates containing from three to seven amino acid residues. No kinetics data is reported for comparison with respect to substrate size (Kiefer-Partsch et al. 1989; Zevaco et al. 1990; Booth et al. 1990b; Lloyd & Pritchard 1991; Tsakalidou et al. 1998). Liberated Xaa-Pro dipeptides contain residues that are basic (Arg-, His-, Lys-), aromatic (Phe-, Tyr-), and hydrophobic/uncharged (Ala-, Ile-, Val-, Gly-). Three Xaa-Pro dipeptides are sequentially liberated from β-casomorphin f60-66 (Tyr-Pro-Phe-Pro-GlyPro-Ile) and β-casein f176-182 (Lys-Ala-Val-Pro-TyrPro-Gln) (Kiefer-Partsch et al. 1989; Zevaco et al. 1990; Lloyd & Pritchard 1991). The hydrolysis of the latter substrate confirms the additional specificity for Xaa-Ala-(Xaa)n first identified by activity with Ala-Ala-ρNA (Meyer & Jordi 1987; Zevaco et al. 1990; Khalid & Marth 1990b). PepX is also capable of hydrolyzing Pro-Pro-(Xaa)n substrates, but little or no hydrolysis is observed for Xaa-Pro-Pro (including when Xaa is Pro) (Booth et al. 1990b; Miyakawa et al. 1991; Miyakawa et al. 1994). This is corroborated by the lack of hydrolysis of a bradykinin (Arg-Pro-ProGly-Phe-Ser-Pro-Phe-Arg) and Lys-bradykinin (LysPro-Pro-Gly-Phe-Ser-Pro-Phe-Arg), although the possibility of substrate size limitation of these nanopeptides has not been investigated (Lloyd & Pritchard 1991). Northern analysis with a pepX (∼2.4-kb ORF) specific probe identified a single transcript of ∼2.6-kb from Lb. helveticus grown in MRS broth under pH controlled conditions (5.85) (Vesanto et al. 1995). The monocistronic pepX transcript level was constant into late exponential phase and is then slowly degraded in stationary phase. Under the same conditions, the PepX activity increases four-fold in the last two generations prior to stationary phase and remains constant for at least 12 more hours. A similar PepX expression profile was reported for Lb. delbrueckii grown in MRS without pH control, but ∼50% of the activity is lost after 12 h in stationary phase (Atlan et al. 1990). The regulation of expression of PepX in Lc. lactis has been demonstrated to be strain dependent, with the activity being equivalent or higher in cultures grown in milk than from complex amino acid or peptide-containing

224 media (Meijer et al. 1996). Currently, there is no information available on the transcriptional regulation of pepX for LAB grown in milk. The physiological role of PepX in Lb. helveticus, Lb. delbrueckii, and Lc. lactis has been investigated for deletion mutants (Mayo et al. 1993; Mierau et al. 1996a; Yüksel & Steele 1996; Yüksel & Steele 1999), gene disrupted mutants (Matos et al. 1998), and chemically derived mutants (Atlan et al. 1990; Nardi et al. 1991) as a function of specific growth rate. Chemical mutagen derivatives of Lb. delbrueckii expressing no detectable PepX activity displayed elongated lag phases and lower total biomass in MRS medium, but the specific growth rate was the same as the parental strain (Atlan et al. 1990). Chemical mutagen derivatives of Lc. lactis expressing no detectable PepX activity displayed decreased growth rates in milk media corresponding to a 60% longer generation time compared to wild type (µmax 1.0 vs 1.6 h−1 ) (Nardi et al. 1991). Gene disrupted derivatives of pepX in Lc. lactis expressed no detectable PepX activity and displayed decreased growth rates in milk media corresponding to a 31% longer generation time compared to wild type (µmax 0.70 vs 0.93 h−1 ) (Matos et al. 1998). Complementation data was not reported as confirmation of physiological role for the chemically derived or gene disrupted mutants. Deletion derivatives of pepX, devoid of PepX activity, were constructed by two step recombination procedure in Lb. helveticus (Yüksel & Steele 1996). The resulting pepX deletion derivatives displayed equivalent growth profiles and final cell densities in MRS broth. However, the growth rate of the PepX negative strains in milk was decreased, corresponding to approximately two-fold longer generation times compared to wild type (µmax 0.61 vs 0.28 h−1 ) (Yüksel & Steele 1999). However, as with growth in MRS, the final cell densities were similar for parental and deletion derivative strains. Introduction of a plasmidencoded copy of pepX into a deletion derivative strain resulted in the reversion to wild type PepX activity and growth rate in milk. Mutant derivatives of pepX have also been constructed in Lc. lactis (Mayo et al. 1993; Mierau et al. 1996a). The resulting pepX deficient derivatives were reported to have acidification or growth rates similar to the wild type strains. However, the decrease in milk growth rate for one pepX deletion derivative corresponds to a 15% longer generation time (µmax 0.56 vs 0.65 h−1 ), though the final cell densities were equivalent to the parental strain (Mierau et al. 1996a).

PepD The dipeptidase, PepD, has been purified and characterized from Lb. helveticus (Vesanto et al. 1996). The information on activity of PepD for dipeptide substrates is limited but similar to that reported for the broad specificity PepV (Dudley et al. 1996; Vesanto et al. 1996). Possible exceptions include the inability of PepD to hydrolyze the Val-Xaa (-Arg, -Gly, -Leu) and Ile-Xaa (-Gln, -Val) dipeptides, though other dipeptides with hydrophobic/uncharged Nterminal residues are hydrolyzed. As with PepV, no hydrolysis of AA-ρNA substrates and no significant hydrolysis of Pro containing peptides is reported. Northern analyses with pepD (∼1.4-kb ORF) specific probes derived from Lb. helveticus and indicate that the transcript encoding PepD is monocistronic (1.5-kb) (Dudley et al. 1996; Vesanto et al. 1996). The level of pepD transcript was determined to be highest in late exponential phase for cells grown in MRS broth under pH controlled conditions. Total dipeptidase activity was measured under the same growth conditions and remained proportional to cell density. However, analysis of a PepD mutant of Lb. helveticus suggests that this enzyme does not contribute significantly to the total dipeptidase activity (Dudley et al. 1996). The physiological role of PepD in Lb. helveticus has been investigated for deletion mutant derivatives as a function of specific growth rate (Dudley et al. 1996). The growth rate and final cell densities of Lb. helveticus PepD deletion mutants are the same as for the wild type strain in complex media (MRS broth), complete amino acid defined media, and milk. Additionally, total dipeptidase activity was the same for Lb. helveticus wild type and the pepD deletion derivative with several dipeptides previously identified as PepD substrates from the original E. coli subclones. The growth rate and activity results suggest that PepD does not play a significant role in the liberation of essential amino acids from milk derived peptides. PepV The dipeptidase, PepV, has been purified and characterized from strains of Lb. casei, Lb. delbrueckii, Lb. helveticus, Lb. sake, Lb. sanfrancisco and Lc. lactis (Desmazeaud & Zevaco 1977; Hwang et al. 1981; Van Boven et al. 1988; Wohlrab & Bockelmann 1992; Vongerichten et al. 1994; Montel et al. 1995; Tan et al. 1995; Gobbetti et al. 1996; Fernández-Esplá & Martín-Hernández 1997). PepV is considered a broad

225 specificity dipeptidase. As with PepD, no hydrolysis of AA-ρNA substrates and no significant hydrolysis of Pro containing peptides is reported. A wide range of dipeptides are hydrolyzed containing basic (Arg-, His-, Lys-), hydrophobic/uncharged (Ala-, Ile-, Leu, Val-), aromatic (Phe-, Tyr-), and Met residues in the N-terminal position. In contrast, PepV generally does not hydrolyze dipeptides containing Gly as the Nterminal residue. PepV also hydrolyses several β-AlaXaa and Xaa-β-Ala substrates, including carnosine (β-Ala-His) (Vongerichten et al., 1994). Northern analyses with pepV (∼1.6-kb ORF) specific probes derived from Lc. lactis suggest that the transcript from the gene is monocistronic (1.7-kb transcript) (Hellendoorn et al. 1997). However, no further analysis or explanation was given for a second, lighter ∼3.6-kb hybridization band. While the Northern hybridization data suggests it is possible that the pepV is also encoded by a multicistronic transcript, the reported sequence data adjacent to pepV is insufficient to allow analysis for ORFs that could account for the larger transcript size. The physiological role of PepV in Lc. lactis has been investigated for insertionally inactivated derivatives as a function of specific growth rate (Hellendoorn et al. 1997). The growth rate and final cell densities of the PepV deficient strain are the same as for the wild type in complex media (M17 broth). However, the PepV deficient strain displays a decreased growth rate in milk corresponding to a 22% longer generation time compared to wild type (µmax 0.58 vs 0.71 h−1 ). Proline iminopeptidases Proline iminopeptidase activity is defined by specificity for N-terminal proline of peptides. PepI has been purified and characterized from Lb. delbrueckii and Lb. helveticus (Gilbert et al. 1994; Varmanen et al. 1996a). The Lactobacillus PepI deduced sequence contains the motif GQSWGG, which conforms to the active site consensus sequence of the prolyl oligopeptidase family (GxSxGG) (Rawlings et al. 1991) (see also description of PepR). Chemical and sitedirected mutation analysis the Lb. delbrueckii PepI has determined that the consensus region Ser-107, Asp246 and His-273 constitute the catalytic triad of this enzyme (Morel et al. 1999a). These three amino acid residues are also conserved in PepI of Lb. delbrueckii and Lb. helveticus. Hydrolytic activity has been reported for ProXaa substrates when the C-terminal residue is hydro-

phobic/uncharged (-Ala, -Gly, -Ile, -Leu, -Val), acidic (-Glu), and aromatic (-Phe, -Tyr). The PepI of Lb. delbrueckii hydrolyses a variety of other dipeptides at low rates, the majority of which contain at least one hydrophobic/uncharged residue (Gilbert et al. 1994). In addition, this enzyme liberates proline from the N-terminus of a tripeptide, but not from a limited number of the tetra- and pentapeptides analyzed. The iminopeptidase of Lc. lactis displays relatively high activity for liberation of proline from several tripeptides and Pro-Phe-Gly-Lys, but not the pentapeptides tested (Baankreis & Exterkate 1991). The Lactobacillus PepI hydrolyze Pro-Pro and show no significant activity for AA-ρNA derivatives except Pro-ρNA. In contrast, no activity was detected for these two substrates with the Lc. lactis proline iminopeptidase (Baankreis & Exterkate 1991). Differences in hydrolytic capabilities with several additional substrates, together with the metallopeptidase classification, suggests the Lc. lactis enzyme is not closely related to PepI of Lactobacillus. Transcript containing pepI was not detectable by Northern blot from pH controlled (5.6 minimum) whey broth grown culture (Varmanen et al. 1996b). However, RT-PCR using primers to pepI and the flanking ORFs resulted in identification of products consistent with the cotranscription of pepI and either of the adjacent ORFs. This evidence suggests a multicistronic transcript with a minimum size of ∼2.9-kb encoding PepI and two other ORFs. The physiological role of PepI in Lb. helveticus has been investigated for deletion mutant derivatives as a function of specific growth rate. The growth rate and final cell densities of Lb. helveticus PepI deletion mutants are the same as for the wild type strain in complex media (MRS broth) and complete amino acid defined media (Yüksel & Steele 1999). However, the growth rate in milk was slower for the PepI deletion mutant compared to wild type (µmax 0.56 vs 0.61 h−1 ), corresponding to a 9% longer generation time, but the final cell densities were similar. Prolidase (PepQ) PepQ has been purified and characterized from Lb. casei, Lb. delbrueckii, Lb. helveticus, and Lc. lactis (Kaminogawa et al. 1984; Booth et al. 1990c; Fernández-Esplá et al. 1997; Yüksel & Steele 1999; Morel et al. 1999b). Prolidase activity is defined by specificity for Xaa-Pro dipeptides. However, not all prolidases are exclusively Xaa-Pro dipeptidases, nor

226 do they hydrolyze all Xaa-Pro substrates. The predominant activity of PepQ from LAB for Xaa-Pro substrates includes N-terminal residues that are hydrophobic/uncharged (Ala-, Ile-, Leu-, Val), basic (His-), aromatic (Phe-, Tyr), and sulfur containing (Met-). Conflicting results have been obtained for approximately twenty non-proline-containing dipeptides. Not all of the characterized enzymes are able to hydrolyze Pro-Pro, but none are reported to hydrolyze Gly-Pro. Additionally, some of these enzymes appear to have activity for Pro-Ala, Pro-Pro, and Pro-Val substrates. Northern analyses with pepQ (1.1-kb ORF) specific probes derived from Lb. delbrueckii indicate that the transcript from the gene is monocistronic (∼1.1-kb transcript) (Rantanen & Palva 1997). The pepQ transcript levels reach a maximum at the transition between exponential and stationary phase during growth under pH controlled conditions (6.00) in MRS broth. No data was reported for expression of total prolidase activity for these growth conditions. The physiological role of PepQ in Lb. helveticus has been investigated for deletion mutant derivatives as a function of specific growth rate. The growth rate and final cell densities of Lb. helveticus PepQ deletion mutants are the same as for the wild type strain in complex media (MRS broth) and complete amino acid defined media (Yüksel & Steele 1999). However, the growth rate in milk was slower for the PepQ deletion mutant compared to wild type (µmax 0.54 vs 0.61 h−1 ), corresponding to a 13% longer generation time, but the final cell densities were similar. Comparison of peptidase activity from wild type and PepQ deficient strains indicated that PepQ accounts for greater than 99% of the hydrolysis of Met-Pro, Leu-Pro, and PhePro. No significant differences were observed for the hydrolysis of Met-Ala or Pro-Phe, which are known PepR and/or PepI substrates. Prolinase (PepR) A dipeptidase with prolinase activity, PepR, has been purified and characterized from Lb. helveticus (Shao et al. 1997). Prolinase activity is defined by specificity for Pro-Xaa dipeptides. The Lactobacillus PepR deduced sequences contain the motif GQSWGG (Dudley & Steele 1994; Varmanen et al. 1996b; Varmanen et al. 1998), which conforms to the active site consensus sequence of the prolyl oligopeptidase family (G×S×GG) (Rawlings et al. 1991). Replacement of consensus region serine residue (Ser-111 to Ala) res-

ulted in the loss of detectable activity, confirming the predicted active site (Shao et al. 1997). The Lactobacillus PepR displays a relatively broad dipeptidase specificity as well as limited activity for larger peptides and Pro-(ρNA, βNAP) substrates. Hydrolysis of Pro-Xaa substrates has been reported for C-terminal residues that are hydrophobic/uncharged (-Ala, -Ile, -Leu, -Val), aromatic (-Phe), and sulfur containing (-Met). Relatively high activity for several non-proline dipeptides is reported for PepR of Lb. helveticus, including Met-(Ala, Leu, Phe), Leu-(Arg, Ser), Ser-Phe, and Thr-Leu (Shao et al. 1997). Hydrolysis of Xaa-Pro dipeptides has not been reported. Northern analyses with a pepR (∼0.9-kb ORF) specific probe derived from Lb. helveticus suggest that the transcript is monocistronic (1.25-kb) (Varmanen et al. 1996b). Transcript level and prolinase activity was determined to be highest at late exponential phase for cells cultured in complex media (MRS broth) under pH controlled conditions. Northern analyses with a pepR (∼0.9-kb ORF) specific probe derived from Lb. rhamnosus identified two transcript sizes (1.0 and 1.5-kb) (Varmanen et al. 1998). The larger transcript was consistent with the expected size of a product encoding pepR and a downstream ORF of unknown function. Both transcript sizes were at their maximum level at the exponential to stationary transition phase, although no later time points were reported. The physiological role of PepR in Lb. helveticus and Lb. rhamnosus has been investigated for deletion mutant derivatives as a function of specific growth rate. The growth rate and final cell densities of Lb. helveticus and Lb. rhamnosus PepR deletion mutants are the same as for the respective wild type strains in complex media (MRS broth) and milk (Shao et al. 1997; Varmanen et al. 1998). The total hydrolytic activity from cell extracts of Lb. helveticus (grown in MRS broth) and Lb. rhamnosus (grown in whey medium) for the substrate Pro-Leu was determined to be reduced by ∼95% in the PepR deletion mutants. Additionally, PepR of Lb. helveticus accounts for significant activity from cell extracts toward Pro-Met (∼90%), Thr-Leu (∼90%), Pro-Phe (∼80%), Gly-Leu (∼55%), and Met-Ala (30%) (Shao et al. 1997). PepL A peptidase, PepL, has been cloned and characterized from Lb. delbrueckii that displays high specificity for

227 Leu-ρNA and Ala-ρNA (Klein et al. 1995). Growth tests with a Leu / Pro auxotrophic E. coli strain expressing recombinant PepL suggest hydrolysis occurs with several di- and tripeptides with N-terminal Leu residues. Growth was also obtained with Pro-Pro, Pro-Ser, and Gln-Pro as the sole source of proline. Northern analyses was conducted with a pepL (∼0.9-kb ORF) specific probe and a probe derived from a partially sequenced upstream ORF with limited C-terminal deduced sequence identity to several peptide transport related proteins, including dciAE of Bacillus subtilis (Klein et al. 1995). A 1.5-kb transcript was detected with both probes using mRNA isolated from Lb. delbrueckii grown in complex media (MRS broth). The transcript is proposed to be a product of pepL and an upstream gene (Klein et al. 1995). However, a 1.5-kb transcript (and three smaller) is also detected with the pepL probe from mRNA of E. coli CM89 with pJUK20, which contains a 6.6-kb insert with pepL and only ∼200-bp of upstream sequence, and is therefore inadequate to provide the full length transcript. Additionally, the 1.5-kb transcript is not large enough to include pepL and a homologue of the peptide transporters of B. subtilis, since the ORFs that encode the latter proteins are >1.6-kb (Mathiopoulos et al. 1991; Perego et al. 1991). PepP The aminopeptidase, PepP, has been purified and characterized from Lc. lactis (Mars & Monnet 1995). PepP liberates the N-terminal amino acid from peptides with general specificity for Xaa-Pro-Pro-(Yaa)n sequences (Mars & Monnet 1995). The rate of hydrolysis was highest for the pentapeptides Arg-Pro-ProGly-Phe (bradykinin f1-5) and Leu-Pro-Pro-Ser-Arg. Relatively high activity was observed with peptides ranging from three to nine residues, but the dipeptides tested were not hydrolyzed. Specificity was observed for the N-terminal (Xaa-) residues Arg, Met, Lys, Leu and Tyr. Increased hydrolysis was observed with the esterified derivatives of Xaa-Pro-Pro-(Yaa)n peptides containing Glu residues at the P0 3 and P0 5 position (n=4) or P0 7 and P0 9 position (n=8). These results suggest that PepP contains an extended binding site and hydrolysis may be prevented by the presence of negatively charged substrate residues beyond the P0 2 position (Mars & Monnet 1995). The physiological role of PepP in Lc. lactis has been investigated for a gene-disrupted mutant as a function of specific growth rate (Matos et al. 1998).

The growth rate of Lc. lactis PepP gene disrupted mutants are the same as for the wild type strain in complex media (M17 broth) and only slightly slower than wild type when grown in milk (µmax 0.89 vs 0.93 h−1 ). However, the assertion that PepP is not involved in hydrolysis of casein-derived prolyl peptides during growth in milk is premature without analysis of the intracellular peptide composition. PepT and other tripeptidases The tripeptidase, PepT, has been purified and characterized from Lc. lactis (Bosman et al., 1990; Bacon et al. 1993). PepT hydrolyses tripeptides with a wide range amino acids including substrates comprised of hydrophobic/uncharged, aromatic, basic, acidic, and sulfur-containing residues. Hydrolysis is observed for Pro-Gly-Gly and Leu-Xaa-Pro (-Gly-, -Ala-), but not for any Xaa-Pro-Yaa substrates tested. In accordance with the strict substrate size specificity of PepT, no activity has been reported for any di-, tetra-, or larger oligopeptides. Several other tripeptidases have been purified and characterized from Lb. delbrueckii, Lb. sake, Lc. lactis, and Pediococcus pentosaceus (Desmazeaud & Zevaco 1979; Sahlstrøm et al. 1993; Bockelmann et al. 1995; Bockelmann et al. 1997; Simitsopoulou et al. 1997; Sanz et al. 1998). Like PepT, these enzymes are generally metallopeptidases and have high activity for tripeptides comprised of hydrophobic/uncharged and aromatic residues. The physiological role of PepT in Lc. lactis has been investigated for deletion mutant derivatives as a function of specific growth rate (Mierau et al. 1996a). The growth rate and final cell densities of the PepT deficient strain are the same as for the wild type in complex media (M17 broth). However, the PepT deficient strain displays a decreased growth rate in milk corresponding to a 10% longer generation time compared to wild type (µmax 0.60 vs 0.65 h−1 ). Comparison of wild type and PepT deficient total activity with tripeptide substrates was not reported. PepE The endopeptidase, PepE, has been purified and characterized from Lb. helveticus (Fenster et al. 1997). PepE was initially identified on the basis of hydrolysis of the N-terminal and C-terminal blocked substrate N-benzoyl-Phe-Val-Arg-ρNA (Fenster et al. 1997). Purified PepE was also capable of hydrolyzing Metenkephalin (Tyr-Gly-Gly-Phe-Met) at the Gly3 -Phe4

228 bond and Bradykinin (Arg-Pro-Pro-Gly-Phe-Ser-ProPhe-Arg) at the Gly4 -Phe5 bond. No activity was reported with β-casomorphin, N-benzoyl-Val-Gly-ArgρNA, or with Arg-, Phe-, Pro-, Lys- Gly-, or Val-ρNA. PepG PepG has been purified and characterized from Lb. delbrueckii (Klein et al. 1997). Comparison of PepG with the deduced amino acid sequence of Lb. helveticus PepE and Lb. delbrueckii PepC reveals 71.2% and 44.6% identity, respectively. Activity for several AA, di-, and tripeptide-βNAP substrates was compared from E. coli CM89 expressing recombinant PepG and PepC from Lb. delbrueckii (Klein et al. 1997). Nearly all of the twenty-one substrates tested were hydrolyzed by PepC with the exception of Pro- and Gly-ProβNAP. In contrast, PepG hydrolyzed Gly-Phe-βNAP and Leu-Gly-βNAP, but not the smaller counterparts Phe-βNAP and Gly-βNAP, suggesting the enzyme may have dipeptidyl specificity. However, PepG also hydrolyzed Ala- and not Ala-Ala-βNAP, but displayed a lower rate of hydrolysis with Ala-ρNA compared to Ala-Ala-ρNA (Klein et al. 1997). Additionally, PepG had no detectable activity for several other AAρNA substrates that were hydrolyzed by PepC (Gly-, Glu-, Leu-, Lys-, Phe-, and Tyr-ρNA). Growth tests with a Leu/Pro auxotroph of Salmonella typhimerium expressing recombinant PepG suggest hydrolysis of di-/tripeptides and a tetrapeptide containing Leu, but no growth was obtained with any peptides containing Pro as the sole essential amino acid. PepO The endopeptidase, PepO, has been purified and characterized from Lc. lactis (Tan et al. 1991; Pritchard et al. 1994; Stepaniak & Fox 1995; Lian et al. 1996). PepO is reported to hydrolyze oligopeptides ranging in length from five (Met- and Leu-enkephalin) to thirtyfive residues (α S1 -casein f165-199) (Tan et al. 1991; Pritchard et al. 1994; Stepaniak & Fox 1995; Lian et al. 1996). Hydrolysis also occurs with bradykinin, angiotensin, neurotensin, and insulin β-chain. Similar to thermolysin, several points of hydrolysis by PepO occur at peptide bonds in which Leu or Phe occupy the P1N site (Pritchard et al. 1994; Stepaniak & Fox 1995). In general, the affinity and rate of hydrolysis increases for YGGF(X)n substrates as the values increase for n=(1 to 13) (Lian et al. 1996). Although PepO has activity with several large casein derived

fragments, the native caseins are not detectably hydrolyzed (Tan et al. 1991; Pritchard et al. 1994; Stepaniak & Fox 1995). Northern analysis with a pepO (∼1.9-kb ORF) specific probe identified two transcript sizes, ∼1.5 and ∼2.2-kb, from Lb. helveticus grown in MRS broth (Chen & Steele 1998). Northern analysis of total mRNA from the pepO deletion mutant also revealed two transcripts, ∼1.5 and ∼1.8-kb. The size of the pepO probe (1.5-kb) and the position of the deletion (∼1.2-kb from transcriptional start site) indicate that the 1.5-kb transcript detected in the wild type and deletion mutant strain is not a degradation product of the 2.2-kb pepO transcript. The physiological role of PepO in Lb. helveticus and Lc. lactis has been investigated for insertion disrupted and deletion mutant derivatives as a function of specific growth rate. The growth rates, acidification rates, and final cell densities of the insertion disrupted pepO mutants of Lc. lactis were reported to be indistinguishable from the parental or wild type strains when grown in milk (Mierau et al. 1993; Tynkkynen et al. 1993). The growth rate and final cell densities of Lb. helveticus PepO deletion mutants are the same as for the wild type strain in complex media (MRS broth), amino acid defined media, and milk (Chen & Steele 1998). Similar results were obtained for growth and acidification rates of wild type and PepO deletion mutants of Lc. lactis in complex media (GM17 broth) (Mierau et al. 1996a). However, the Lc. lactis PepO deletion mutant displays a decreased growth rate in milk corresponding to a 9% longer generation time compared to wild type (µmax 0.60 vs 0.65 h−1 ) (Mierau et al. 1996a). PepF The endopeptidase, PepF, has been purified and characterized from Lc. lactis (Monnet et al. 1994). The original report of PepF was from Lc. lactis ssp. lactis in which pepF1 is located chromosomally (Monnet et al. 1994). A PepF homologue (encoded by pepF2) is expressed in Lc. lactis ssp. cremoris that has 84% identity to the deduced amino acid sequence of the enzyme encoded by pepF1 located on the lactoseproteinase plasmid of this strain (Nardi et al. 1997). PepF is reported to hydrolyze oligopeptides ranging in length from seven (β-casomorphin f60-66) to seventeen residues (adrenorticotropic hormone f1-17) (Monnet et al. 1994). However, a clear preference is revealed for substrates of about eight or nine residues

229 as indicated by the rates of hydrolysis. The enzyme was shown to hydrolyze ACTH f1-17 at three positions, liberating peptides of three to seven residues. No hydrolysis was observed for ACTH f1-24, glucagon (29 residues), or insulin B chain (30 residues), suggesting the upper size limit for substrates is less than twenty-four residues. Additionally, the native caseins are not detectably hydrolyzed. Northern analyses with pepF2 (∼1.8-kb ORF) specific probes derived from Lc. lactis suggest that the transcripts from the chromosomally located pepF genes are polycistronic and include two adjacent ORFs (∼3.7-kb transcript) (Nardi et al. 1997). However, comparison of luciferase activity from lux fusions with the promoter corresponding to the 3.7-kb transcript and a second potential promoter upstream of pepF2 clearly indicated the greater strength of the latter promoter, which would result in a 2.6-kb bicistronic transcript. The physiological role of PepF in Lc. lactis has been investigated for deletion mutant derivatives as a function of specific growth rate. The growth rate of Lc. lactis PepF1 or PepF2 deletion mutants are the same as for the wild type strain in complex media (M17 broth) (Nardi et al. 1997). However, loss of either copy of PepF from Lc. lactis ssp. cremoris resulted in a decrease in growth rate in a minimal medium. Similar results were obtained with a double mutant, devoid of PepF, resulting in decrease growth rates in minimal media corresponding to a 16% increase in generation time compared to wild type (µmax 0.12 vs 0.14 h−1 ). These results suggest a potential role for PepF in protein turnover under nitrogen limiting conditions. Location of peptidases Several lines of evidence strongly suggest the intracellular location of all the peptidases identified in LAB to date. Included are the absence of recognized signal sequences or membrane anchors in any of the peptidases for which sequence is available. In addition, more recent evidence for the ability of Opp to transport several large casein derived peptides negates the previous assertions that peptidase secretion would be required for liberation of essential amino acids (Detmers et al. 1998; Kunji et al. 1998). A thorough evaluation of evidence for localization of peptidases can be found in a previous review (Kunji et al. 1996).

Physiological role of peptidases

The primary focus and evidence for physiological role of peptidases of LAB is derived from the evaluation of mutant growth rate differences in milk and complex or defined media. In addition, a few studies have evaluated the effect of multiple peptidase mutations (Christensen & Steele 1996; Mierau et al. 1996a; Matos et al. 1998; Yüksel & Steele 1999). Single and multiple peptidase mutants have been constructed in Lb. helveticus (pepC, pepD, pepI, pepN, pepO, pepQ, pepR, pepX) and Lc. lactis (pepA, pepC, pepF, pepN, pepO, pepT, pepX, pepV) (see respective sections). In general, no differences are reported for the growth parameters of single and multiple mutants compared to the wild type strain when grown in complex media (MRS or M17 broth) or complete amino acid defined media. These results suggest the mutation(s) do not effect the normal protein turnover or protein maturation requirements of the peptidase mutant strain to the degree that is detectable as a growth rate deficiency. The reduction of growth rates in milk exhibited by LAB peptidase mutants follows the generalization of greater deficiency with greater numbers of mutations (Mierau et al. 1996a; Yüksel & Steele 1999). This trend can be interpreted as reduction in rate of liberation of essential amino acids from the exogenous peptide supply, thereby limiting the availability of amino acids for new protein synthesis. While this interpretation is direct, there is also evidence for regulation of peptidases as a function of the class of intracellular nitrogen accumulated. The regulation of PepN and CEP in Lc. lactis are affected by the exogenous supply of amino acids or peptides, with specific cases reported for dipeptides such as Pro-Leu which result in decreased expression of these enzymes (Marugg et al. 1995; Meijer et al. 1996). These results suggest the pool of accumulated peptides in a peptidase mutant may regulate the activity of remaining peptidases, further reducing the hydrolytic capabilities, and indirectly causing a reduction in growth rate. Evaluation of remaining peptidase activities of single and multiple peptidase mutants indicate only minor fluctuations in levels of expression (Christensen & Steele 1996; Mierau et al. 1996a). However, these comparisons were made from cultures in complex media, conditions where no effect on growth rates were observed, and may not be representative of the expression levels during growth in milk or under conditions where intracellular peptides accumulate.

230 The observation that a peptidase is non-essential for growth in milk is clearly different from what is essential for survival in a population should the mutation arise naturally. Differences in the specific growth rates of single peptidase mutants relative to wild type are often described as insignificant, even when the deficiency corresponds to a 10% increase in the generation time. In all cases where the specific growth rates in milk were reported for peptidase mutants the mean values indicate a reduction in growth rate relative to wild type. These results indicate that the difference in growth rate is not simply a function of assay error, although the precision of some of these measurements may be low. Furthermore, while the difference may not be considered statistically significant due to a small number of replicates and inherent assay error, any measurable reduction in growth rate is clearly significant to the survival of a genotype in a population. For example, a 10% increase in generation time for a mutant strain would result in progressive reduction to 0.1% of the total biomass after competitive growth for ∼110 wild type generations starting from equivalent inoculations. Additionally, this estimation only accounts for differences observed in exponential growth and little is known about the effect of specific LAB peptidase mutations on transitional phases and survival at stationary phase. It may be considered that the evaluation of specific growth rates in milk is not the most accurate or informative means by which the importance of a given peptidase mutation can be expressed. A more sensitive analysis may include the experimental determination of the rate of loss of a genotype from a parental population in competitive growth conditions (as described above). Additionally, alternate roles may be revealed by evaluating the survival of a mutant strain under conditions that LAB are known to encounter, including carbohydrate depleted and acidic environments. Analysis such as these may in part elucidate the capacity of peptidase mutant (or over expressing) strains for survival in cheese, especially since the pathways for amino acid catabolism are clearly dependent on the availability of substrate. The physiological role of peptidases, in functions not related to the hydrolysis of exogenous peptides, may be more appropriately evaluated in minimal media as was reported for PepF mutants of Lc. lactis (see PepF section) (Nardi et al. 1997). The impact of peptidase mutations are likely to be most evident when protein turnover is required under conditions of starvation and when new protein synthesis occurs

at transitional phases (Gottesman & Maurizi 1992; Miller 1987). The precedence for this type of peptidase dependence was established in studies with E. coli and S. typhimurium multiple peptidase mutants (Yen et al. 1980a; Yen et al. 1980b; Miller & Green 1981; Miller & Green 1983). Amino acid catabolism by LAB The catabolism of amino acids by LAB has implications with regard to the quality and safety of fermented foods. Amino acid catabolism by anaerobic/fermentative microorganisms is also believed to have an important role in their ability to obtain energy in nutrient-limited environments. For example, Arg and His are known to provide energy via substrate level phosphorylation and precursor/product exchange coupled to amino acid decarboxylation, respectively. However, the catabolic pathways for many amino acids remain unknown or only partially characterized in LAB. This section of the review will summarize our current knowledge of amino acid catabolic pathways identified in LAB. The relationships between these pathways and cellular physiology, development of cheese flavor, and production of compounds with human health implications will also be discussed when possible. The catabolism of sulphur containing amino acids is not reviewed as this topic is covered elsewhere in these proceedings. Arginine catabolism Two pathways for the catabolism of arginine have been described. Strains of Lb. fermentum have been demonstrated to produce nitric oxide from arginine. Morita et al. (1997) postulated that this reaction was catalyzed by a nitric oxide synthase that could also produce citrulline as a coproduct. A more common pathway for the catabolism of arginine in LAB is via the arginine deiminase (ADI) pathway (Manca de Nadra et al. 1982; Cunin et al. 1986; Konings et al. 1995). The enzymes involved in the arginine deiminase pathway and corresponding arginine/ornithine antiporter are shown in Figure 1. This pathway results in the conversion of one mole of arginine into two moles of ammonia and one mole each of ornithine, carbon dioxide, and ATP. This is the only known LAB amino acid catabolic pathway that results in the production of ATP via substrate level phosphorylation. A number of studies have examined the occurrence of ADI pathway enzymes in LAB (Crow & Thomas

231

Figure 1. The arginine deiminase pathway and arginine/ornithine antiporter (A/O AP). Abbreviations include: M, cytoplasmic membrane; ADI, arginine deiminase; OTC, ornithine transcarboxylase; CK, carbamate kinase.

1982; Manca de Nadra et al. 1982; Liu et al. 1995). Collectively, these studies indicate that the pathway is typically present in heterofermentative lactobacilli, while it is a variable trait in Lc. lactis, Oenococcus oenos, and the Thermobacterium subgenera of Lactobacillus. Additionally, numerous LAB have been described that have an incomplete ADI pathway (Crow & Thomas 1982; Manca de Nadra et al. 1982). The gene cluster encoding the ADI pathway from Lb. sake has been characterized in detail (Zúñiga et al. 1998). The genes contained within this cluster and their order were determined to be: arcA, encoding ADI; arcB, encoding ornithine transcarboxylase; arcC, encoding carbamate kinase; arcT, encoding a noncharacterized transaminase; arcD, encoding an arginine-ornithine antiporter. The regulation of this pathway has been examined in Lb. buchneri, Lb. leichmannii, Lb. sake, Lc. lactis, and O. oenos. In general, arginine induces the expression of the ADI pathway enzymes and some carbohydrates repress their synthesis. A gene encoding ahrC, a regulator of arginine biosynthesis and catabolism, has been identified in Lc. lactis (Rallu et al. 1996); it is possible that this protein is involved in the induction of the ADI pathway enzymes by arginine. Repression by carbohydrates is dependent both on the organism examined and the carbohydrate utilized, suggesting that catabolite repression can be involved in the regulation of this pathway. A recent study by Zúñiga et al. (1998) identified two cre sequences present in the promoter region upstream of arcA; identification of these sequences strongly supports the involvement of the regulatory factor CcpA, and hence catabolite repression as a mechanism of regulation of this pathway in Lb. sake.

Both the production of ATP and ammonia from arginine have physiological implications. It has been demonstrated in Lb. buchneri (Liu & Pilone 1998), Lc. lactis (Crow & Thomas 1982), and O. oenus (Liu & Pilone 1998) that the energy derived from arginine catabolism can be coupled to growth. Additionally, survival at low pH is enhanced by the activity of the ADI pathway, most likely the result of an increase in culture pH resulting from ammonia production (Marquis et al. 1987). Arginine catabolism by LAB during wine production can result in the production of ethyl carbonate precursors. This has human health implications, as ethyl carbonate is a known animal carcinogen (Liu & Pilone 1998). Aspartate catabolism In LAB, three different Asp catabolic pathways have been identified, however the details and distribution of these pathways are still not well understood. First, an Asp aminotransferase has been purified from Lb. murinus by Rollan et al. (1988) and cloned from Lc. lactis (Dudley & Steele 1999). Secondly, an Asp decarboxylase has been reported in a nonspeciated Lactobacillus strain by Abe et al. (1996). Lastly, an aspartase, which catalyzes the reversible reaction Asp to fumarate and NH+ 3 , has been purified from Lb. murinus (Rollan et al. 1985). The gene encoding an Asp aminotransferase (aspC) has recently been characterized from Lc. lactis LM0230 (Dudley & Steele 1999). The deduced amino acid sequence indicates that the lactococcal AspC is highly related to other pyridoxal-50-phosphate dependent Asp aminotransferases from prokaryotes.

232 regulate intracellular pH. In addition, amine transport results in an electrochemical gradient that can be used by the cell to drive energy-consuming reactions such as the transport of nutrients or to generate ATP via a F1 F0 ATPase (Konings et al. 1997). Glutamate catabolism

Figure 2. A generalized schematic of the generation of metabolic energy and regulation of intracellular pH by decarboxylation and electrogenic antiport. Abbreviations include: M, cytoplasmic membrane; R-COO− , an amino acid; R-H, the corresponding amine.

Also, it was demonstrated that cell-free extracts of E. coli strains overexpressing the lactococcal aspC could catalyze Asp transamination with α-ketoglutarate, but no detectable activity with the other 19 natural amino acids was observed (Dudley & Steele 1999). The lactobacilli and lactococcal AspC, like their counterparts in other organisms, are likely involved in the last step of aspartate biosynthesis (Paulus 1993). This is supported by the observation that Lc. lactis LM0230(1aspC) cannot grow in defined medium lacking both Asp and Asn (Dudley & Steele 1999). Since Asp is believed to serve as a precursor for other cellular amino acids and nucleotides in LAB, this enzyme likely has a central role in LAB physiology. However, it is not known whether the AspC of LAB are involved in catabolism Asp in vivo. Other than the purification of the aspartase from Lb. murinus, we are not aware of any other studies concerning this enzyme in LAB. Therefore, the physiological role of this enzyme is currently unknown. Decarboxylation of Asp to Ala may be used to generate metabolic energy and regulate intracellular pH in Lactobacillus sp. (Abe et al. 1996; Konings et al. 1995) via a general mechanism common to a variety of decarboxylases. These two mechanisms are diagrammed in Figure 2. In the cytoplasm, the decarboxylation of an amino acid results in the consumption of one proton. The transport of the corresponding amine outside of the cell leads to ‘indirect proton extrusion’ (Maloney et al. 1992), which may be used to

The identified pathways of Glu catabolism in LAB are initiated by an aminotransferase, a dehydrogenase, or a decarboxylase. The first two enzymes result in the formation of α-ketoglutarate from Glu, while γ aminobutyrate (GABA) is the product of decarboxylation. As discussed below, it appears unlikely that these products are catabolized further. All purified and cloned aminotransferases identified to date from LAB are involved in catabolism of Glu, as this amino acid serves as the amino donor in these reactions. While a number of prokaryotes encode an α-ketoglutarate dehydrogenase, which converts α-ketoglutarate to succinyl-CoA, this enzyme has not been found in LAB (Morishita & Yajima 1995; Lapujade et al. 1998). Therefore, based upon our current knowledge, it appears catabolism of Glu by this pathway is restricted to the cycling of Glu and α-ketoglutarate by aminotransferases. Glu decarboxylase has been purified from Lactobacillus brevis (Ueno et al. 1997) and the genes encoding the decarboxylase (gadB), putative Glu/γ aminobutyrate antiporter (gadC), and a transcriptional activator (gadR) of these genes have been characterized from Lc. lactis (Sanders et al. 1998). The decarboxylase and antiporter are organized in an operon, and gadR is essential for gadCB transcription. Additionally, gadCB transcription is highest in complex medium around the onset of stationary phase and is positively regulated by NaCl, Glu and low pH. Glu dehydrogenase activity has been detected in cell-free extracts of Lb. fermentum and Lc. lactis (Misono 1985), however a second report was unable to detect Glu dehydrogenase in Lc. lactis (Lapujade et al. 1998). Therefore, it is not clear how widely distributed this activity is in lactococci. Higuchi et al. (1997) have demonstrated whole cells of Lactobacillus sp. incubated in the presence of Glu are capable of synthesizing ATP in a 1pH and F1 F0 ATPase dependent manner as depicted in Figure 2. Sanders et al. (1998) demonstrated that Lc. lactis cells preincubated in a defined medium with 0.3 M NaCl and 1 mM Glu had an approximately 2500–5000 fold increase in viability after a lactic acid challenge

233 (pH 3.5) when compared to cells that had been preincubated in defined medium with no additions, NaCl alone or Glu alone. This effect was dependent upon a functional gadR and gadB. The same acid sensitivity was seen with a gadC mutant, however this may be the result of a polar effect on the downstream gadB. The results of these two studies clearly implicate Glu decarboxylation in the generation of metabolic energy and regulation of intracellular pH via the mechanisms summarized in Figure 2. GABA production is known to occur in cheese, however we are not familiar with any reports suggesting that GABA has a direct or indirect impact on cheese flavor. Zoon & Allersma (1996) have correlated increased production of CO2 and GABA in cheese with an increase in the number of eyes. The authors suggest Glu decarboxylase activity of S. thermophilus and/or Lb. helveticus was responsible for the development of eyes. However, the identification of the culture(s) responsible and the construction of isogenic strains lacking Glu decarboxylase activity are needed to further support this conclusion. Histidine catabolism Histidine decarboxylases, which catalyze the reaction histidine→histamine+CO2, are broadly distributed in Escherichia, Salmonella, Clostridia, Bacillus, and Lactobacillus (Voight and Eitenmiller 1977). Most decarboxylases, including many histidine decarboxylases, require pyridoxal-50-phosphate as a coenzyme (Boecker & Snell 1972; Guirard & Snell 1964). However, there is a large subclass of amino acid decarboxylases that utilize covalently bound pyruvate rather than pyridoxal-5’-phosphate (Chang & Snell 1968; Huynh et al. 1984; Recsei et al. 1983). The best studied of these enzymes is histidine decarboxylase of Lactobacillus 30a, which is synthesized as a proenzyme containing 310 amino acids (Parks et al. 1985; Hackert et al. 1981). The proenzyme is activated via an autocatalytic event with cleavage between residues Ser81 and Ser82 (Recsei and Snell 1981). The gene encoding histidine decarboxylase (hdcA) exists in an operon with an uncharacterized ORF, designated hdcB. Transcription of this operon is induced 3-fold in the presence of histidine (Vanderslice et al. 1986; Copeland et al. 1989). The likely physiological roles for histidine decarboxylation include the regulation of intracellular pH and the generation of metabolic energy (Molenaar et al. 1993). The generalized scheme is described

in the Asp catabolism section and summarized in Figure 2. Histamine production in fermented foods is of importance for health and food safety concerns. Histamine is a biogenic amine that can result in food poisoning (tenBrink et al. 1990; Stratton et al. 1991; Santos et al. 1996). Several outbreaks of histamine poisoning have occurred subsequent to the consumption of cheese containing high levels of histamine (Doeglas et al. 1967; Kahana & Todd 1981; Taylor et al. 1982). Threonine catabolism Threonine catabolism by LAB is initiated by threonine aldolase, an enzyme widely distributed in LAB. Threonine aldolase converts Thr to acetaldehyde and Gly (Lees & Jago 1976; Marshall & Cole 1983). The enzyme has been partially purified from Lb. delbrueckii ssp. bulgaricus YOP12 and determined to have a molecular weight of 190 kDa (Manca de Nadra et al. 1987). Threonine aldolase activity in Lb. delbrueckii ssp. bulgaricus and S. thermophilus is stimulated by Thr and inhibited by Gly and Cys (Lees & Jago 1976; Manca de Nadra et al. 1987; Marranzini et al. 1989). Some LAB convert acetaldehyde to ethanol via an alcohol dehydrogenase (Marshall & Cole, 1983). Recently, a gene encoding an alcohol dehydrogenase (adhE) has been cloned and characterized from Lc. lactis (Arnau et al. 1998). It is generally believed that the physiological role of threonine aldolase is the production of Gly. The observation that Lc. lactis Z8, which lacks threonine aldolase activity, requires Gly for growth supports this hypothesis (Reiter & Oram 1962). Acetaldehyde contributes to the development of typical yogurt flavor (Law 1981; Sandine & Elliker 1970; Tamine & Seeth 1980). In addition to Thr catabolism, it is known that carbohydrate fermentation can result in the formation of acetaldehyde (Shankar 1977; Wilkins et al. 1986). Further research is required to clarify the relative contribution of carbohydrates and Thr to the production of acetaldehyde by cultures used in yogurt production. Aromatic amino acid catabolism In most LAB, aromatic amino acid catabolism is initiated by an aminotransferase (Gao et al. 1997; Yvon et al. 1997; Gummalla & Broadbent 1998). The aminotransferase initiated pathways are illustrated in Figure 3 and described below. Additionally, some LAB

234

Figure 3. Pathways for the catabolism of tryptophan (A), phenylalanine (B), and tyrosine (C) by starter, adjunct and nonstarter bacteria. Thick arrows denote enzymatic reactions. Thin arrows indicate nonenzymatic chemical reactions. Dashed arrows indicate that the mechanism of conversion is unknown. Enzymes: ATase, aminotransferase.

235 express decarboxylases that have activity on aromatic amino acids. Indole pyruvate (IpyA), the product of aminotransferase activity on Trp, can be catabolized further to indole lactic acid (ILA), indole acetic acid (IAA), indole aldehyde (IAId) and benzaldehyde (Nakazuma et al. 1977; Hummel et al. 1984; Gao et al. 1997). In certain strains of lactobacilli, IpyA is converted to ILA via IpyA dehydrogenase (Hummel et al. 1984, 1986; Gummalla & Broadbent 1998). IAA and IAld may also be produced enzymatically from IpyA. Gao et al. (1997) demonstrated that the accumulation of IAA and IAld are strain-specific in lactococci. This observation is significant because it suggests that differences between strains of lactococci may result in the accumulation of distinct Trp metabolic products. Formation of IAA is of particular importance as it can be non-enzymatically converted to skatole (Urbach 1995). In addition, Honeyfield & Carlson (1990) have demonstrated that a non-speciated strain of lactobacilli is capable of converting IAA to skatole via a decarboxylase. Aminotransferase activity on Phe results in the formation of phenylpyruvate (PPA). Under the conditions employed by Gao et al. (1997), PPA was stable and further catabolism of PPA by lactococci was not detected. However, PPA, phenyl lactate, and phenyl acetate were detected as metabolites of lactococcal Phe catabolism by Yvon et al. (1997). These differences can be attributed to the different reaction conditions employed in these two studies. Also, nonenzymatic breakdown of PPA to benzaldehyde and phenethanol has been described (Kong et al. 1996; Groot & Bont 1998). Aminotransferase activity on Tyr results in the formation of p-OH-phenyl pyruvate (HPPA). Gao et al. (1997) demonstrated that lactococci convert HPPA to 4-hydroxyl-benzaldehyde and p-OH-phenyl acetate. Certain strains of lactobacilli can decarboxylate HPPA to produce p-cresol (Yokoyama & Carlson 1981). Aromatic amino acid aminotransferases (Ar-AT) have been purified and characterized from Lc. lactis NCDO763 and S3 (Yvon et al. 1997, Gao et al. 1998). The Ar-AT from NCDO763 is an 86 kDa pyridoxal-50phosphate-dependent enzyme that has activity on aromatic amino acids, Leu and Met. In Lc. lactis S3, two oligomeric species of Ar-AT have been identified (Gao et al. 1998). The more abundant species, Ar-AT1, exists as a homodimer with a molecular mass of 84kDa. Ar-AT2 is a homotetrameric enzyme composed of the same subunits as Ar-AT1. The Ar-AT purified from Lc.

lactis NCDO763 is similar to Ar-AT1 with regard to optimum pH, subunit size, native molecular mass, and effect of NaCl and substrate specificity. However, the relative activity with aromatic amino acids for Ar-AT1 and the Ar-AT from NCDO763 are Trp>Tyr>Phe and Phe>Tyr>Trp, respectively. It remains unclear which dehydrogenase(s) is responsible for the conversion of PPA and HPPA to phenyl lactate and p-OH-phenyl lactate, respectively. However, L-hydroxyisocaproate dehydrogenase (LHic) has broad substrate specificity, including activity towards PPA (Feil et al. 1994, 1997). Additionally, some evidence suggests that the reduction of PPA is catalyzed by an NAD(H)-dependent 2-hydroxyacid dehydrogenase (D-Hic) (Hummel & Kula 1989). These enzymes are described in greater detail in the branched-chain amino acid catabolism section. The decarboxylation of Tyr and Trp results in the formation of tyramine and tryptamine, respectively. Organisms known to produce tyramine include Lb. brevis, Lb. curvatus, Lb. delbrueckii, Lb. paracasei, and Leuconostoc lactis (Straub et al. 1995; Masson et al. 1996; Roig-sagues et al. 1997). Recently, one of nine strains of Lb. casei examined was also demonstrated to produce tyramine from Tyr (Timpone & Steele, unpublished observations). Tyr decarboxylase activity has been detected in Lb. brevis (Joosten & Stadhouders 1987; Joosten & Northolt 1989) and in non-cheese related strains of Lc. lactis subsp. cremoris (Roig-Sagués et al. 1996). Trp decarboxylase activity has received less research attention. Gummalla & Broadbent (1998) reported the presence of Trp decarboxylase activity in cell-free extracts prepared from three strains of Lb. casei and two strains of Lb. helveticus. However, tryptamine was not detected in the supernatants of these cultures. The likely physiological roles for Tyr and Trp decarboxylation include the regulation of intracellular pH and the generation of metabolic energy. The generalized scheme is described in the Asp catabolism section and summarized in Figure 2. A major concern of the dairy industry is the occurrence of objectionable (unclean) flavors. Compounds believed to be formed via aromatic amino acid catabolism that have been associated with objectionable flavors include p-cresol, phenethanol, phenylacetaldehyde, indole, and skatole (Dunn & Lindsay 1985; Vangtal & Hammond 1986; Wijesundera & Urbach 1993; Milo & Reineccius 1997). Biogenic amines, such as tyramine, are frequently detected in cheese (Voight et al. 1974). Tyramine

236 is of particular concern as it is the most common cause of monoamine intoxication (MAI). MAI is characterized by an increase in blood pressure that results in palpitations, severe headaches, hypertension, nausea, vomiting, and prostration. The majority (80%) of MAIs have been associated with consumption of cheese (McCabe 1986). It is important to understand the factors that influence tyramine accumulation in cheese, so that approaches can be developed to reduce its accumulation. Branched-chain amino acid catabolism Catabolism of branched-chain amino acids (BCAAs) is most likely initiated in Lc. lactis by an aminotransferase, converting Leu, Ile and Val to α-ketoisocaproate (KIC), α-keto-β-methylvalerate (KMV) and α-ketoisovalerate (KIV), respectively. Subsequently, enzymes have been identified in Lc. lactis and other LAB that convert these α-keto acids to aldehydes, alcohols or fatty acids. The details of these pathways are described below. Aminotransferase activity on BCAAs was first reported in a set of studies interested in identifying the origin of a malt-flavor defect in milk (Jackson & Morgan 1954). Recently, a number of aminotransferases from Lc. lactis with activity on BCAAs have been described (Alting & Engels 1996; Yvon 1997; Gao & Steele 1998; Roudot-Algaron & Yvon 1998; Atiles et al. 1999). The Ar-ATs described by Yvon et al. (1997) and Gao & Steele (1998) differ significantly from the enzymes described by Alting & Engels (1996) and Atiles et al. (1999), as these latter two enzymes do not have activity with Tyr or Trp. The deduced amino acid sequence from a gene encoding a BCAA aminotransferase (ilvE) in Lc. lactis LM0230 has high identity to other BCAA aminotransferases (Atiles et al. 1999). A derivative of Lc. lactis LM0230 lacking IlvE activity [LM0230(1ilvE)] was constructed and determined to differ from LM0230 in that it was unable to grow in defined medium when either KIV or KMV was substituted for Val or Ile, respectively. However, LM0230(1ilvE) was able to grow when KIC was substituted for Leu (Atiles et al. 1999). This supports the biochemical evidence that lactococci have at least two aminotransferases with activity on Leu. Therefore, while it appears likely that lactococci initiate catabolism of BCAAs via aminotransferase activity, it is unclear how many lactococcal enzymes catalyze this reaction.

The conversion of the BCAA-derived α-ketoacids to their corresponding aldehydes is catalyzed by a decarboxylase. Tucker & Morgan (1967) identified a decarboxylase in Lc. lactis var. maltigenes strains that converts KIC to 3-methylbutanal and catalyzes the decarboxylation of KIV and KMV. Additionally, one strain of Lc. lactis, which is not part of the maltigenes biovariant, has been demonstrated to produce limited quantities of aldehydes from Leu and Val (MacLeod & Morgan 1958), suggesting this enzyme may be present in lactococci outside of the maltigenes biovariant cluster. A branched-chain α-ketoacid decarboxylase has also been identified in Lb. casei (Hickey et al. 1983). Additionally, production of the aldehydes expected from catabolism of Leu and Val by Lactobacillus maltaromicus (later classified by Collins et al. 1991, as Carnobacterium) has been described (Miller et al. 1974). The most thoroughly characterized enzymes from LAB that catalyze the reduction of BCAA-derived α-keto acids are L-2-hydroxyisocaproate dehydrogenase (L-Hic) (Schütte et al. 1984) and D-2hydroxyisocaproate dehydrogenase (D-Hic) (Hummel et al. 1985). In a screen of five strains of Bifidobacterium and 45 strains of Lactobacillus and Leuconostoc, L-Hic activity was only observed in Lactobacillus confusus (Schütte et al. 1984). The purified enzyme was NADH dependent and exhibited highest activity towards C5 and C6 acids with a preference towards branched-chain α-keto acids. The deduced amino acid sequence of the cloned gene is 30% identical to the L-lactate dehydrogenase (LLDH) gene from Lb. casei (Lerch et al. 1989). The coenzyme-binding residues and the three substrate binding amino acids conserved in all L-LDHs are also present in L-Hic (Lerch et al. 1989). These results indicate that L-Hic is structurally related to L-LDHs. In comparison to L-Hic, D-Hic activity appears to be more widely distributed among LAB. D-Hic activity has been identified in three strains of Leuconostoc, three strains of Lb. casei (Hummel et al. 1985) and in Lb. delbrueckii subsp. bulgaricus (Bernard et al. 1994). As with L-Hic, this enzyme is active on a variety of α-keto acids (Hummel et al. 1985; Kallwass 1992), and is NADH dependent. Comparison of deduced amino acid sequences of D-Hic and D-lactate dehydrogenase suggests the enzymes are related. Similar analysis suggests that D-Hic and L-Hic are genetically unrelated proteins (Lerch et al. 1989; Kochhar et al. 1992). In addition to L-Hic and D-Hic, at least two other LAB dehydrogenases with activity on BCAA-

237 derived α-keto acids are known: a D-(-)-mandelic acid dehydrogenase from Lactobacillus curvatus (Hummel et al. 1988); and an alcohol dehydrogenase from Lc. lactis var. maltigenes (Morgan et al. 1965; Harrison et al. 1969). The production of branched-chain fatty acids (BCFAs) from branched-chain amino acids by LAB does not appear to be common. However, one strain of Lc. lactis subsp. lactis var. diacetylactis and one nonspeciated strain of Lactobacillus have been shown to produce isobutyric acid (i-C4 ) from Val, and isovaleric acid (i-C5 ) or another C5 -acid from Leu and Ile (Nakae & Elliott 1965). An aminotransferase may initiate this pathway, but the other enzymes required for i-C4 and i-C5 formation remain unknown. The physiological significance of BCAA catabolism is poorly defined in LAB. In other bacteria, IlvE executes the last step of BCAA biosynthesis (Umbarger 1996). However, as dairy LAB are typically auxotrophic for BCAAs (Chopin 1993), ilvE likely has a different role in these organisms. The detection of an enzyme in Lc. lactis that catalyzes the reduction of KIC suggests BCAA catabolism may serve to regenerate NAD+ in lactococci (Gao and Steele, unpublished results). Additionally, IlvE might be involved in the recycling of Glu from α-ketoglutarate, which can then be used to catalyze the production of other amino acids synthesized by aminotransferase reactions. However, the growth of LM0230(1ilvE) in milk is indistinguishable from the growth of isogenic ilvE+ strains (Atiles et al. 1999). Therefore, a physiological role for IlvE during growth of LAB in milk is not yet apparent. Physiological roles of BCAA catabolism in other organisms include the generation of metabolic energy and carbon (Massey et al. 1976), the production of BCFAs for cell membrane biosynthesis (de Mendoza et al. 1993), and the production of α-keto acids and α-hydroxyacids that serve as siderophores in low iron environments (Kingsley et al. 1996). However, data supporting similar roles in LAB is lacking. In the production of fermented dairy products, BCAA catabolism is typically associated with the generation of compounds having a detrimental effect on product aroma or flavor. The aroma of the three aldehyde derivatives of BCAA catabolism have each been described as moderately malty (Griffith & Hammond 1989). The corresponding alcohols have similar aromas, but with approximately 50-fold higher flavor thresholds (Sheldon et al. 1971). The conversion of Leu to 3-methylbutanal by Lc. lactis var. maltigenes has been implicated in the development of malty off-

flavors in fermented milks (Jackson & Morgan 1954), as well as 2-methylpropanal and the alcohol derivatives of these two aldehydes (Morgan et al. 1965; Sheldon et al. 1971). Although 2-methylbutanal has a similar odor characteristic and flavor threshold to 3-methylbutanal, it has not been detected in significant concentrations in fermented milks. The aldehyde and alcohol products of BCAA catabolism are also known to cause defects in Cheddar cheese. Cheeses containing 18–90 ppm 3-methylbutanal and 9–45 ppm 3-methylbutanol were graded lower for flavor quality and were maltier than the control cheeses (Braun & Olson 1986). Additionally, Dunn & Lindsay (1985) correlated flavors described as harsh, dull and suppressed with elevated concentrations (>200 ppb) of 3methylbutanal and 2-methylpropanal. In Emmentaler, 3-methylbutanal can be present at levels 6–16 fold over that of the nasal or retronasal odor thresholds of this compound in sunflower oil (Preininger & Grosch 1994). However, Preininger et al. (1996) suggested that 3-methylbutanal is not one of the defining flavor or odor compounds of Emmentaler, but may play a role in suppressing the unpleasant, sweaty odor of butyric acid. The fatty acids derived from BCAAs are described as sweaty, estery, rotten fruit-like and fatty acid-like at concentrations around 5 ppm, to sweet, fruity, waxy and apple-like at higher concentrations (Brennand et al. 1989). While these acids have been detected in Cheddar (Ha & Lindsay 1991a; Yang & Min 1993), they appear to be more prevalent in other cheeses such as Swiss, Italian and blue-veined varieties (Patton 1964; Ha & Lindsay 1991a,b; Bosset et al. 1993). In Swiss, BCFAs have been correlated with flavors described as afterburn, amine and fruity (Vangtal & Hammond 1986). The study of Biede (1979) clearly implicates Propionibacterium shermanii as the primary culture involved in the production of i-C5 in Swiss cheese, however another report (Paulsen et al. 1980) suggests P. shermanii has little impact on i-C5 accumulation. Therefore, more research is required to elucidate how i-C5 is formed in Swiss cheese.

Concluding remarks A significant body of information is now available concerning many of the enzymes that comprise the peptidase enzyme systems of lactobacilli and lactococci. A complement of enzymes with comparable specificities is present in each of the systems studied.

238 Table 2. Specific growth rates of peptidase mutants in milk Peptidase Mutant

Strain

Mutation Type

Mutant vs. WT µmax in Milk

% Increase in Generation Time

Reference

PepC

Lc. lactis

deletion

0.60/0.65

10

Mierau et al. (1996)

PepN PepN

Lb. helveticus Lc. lactis

deletion deletion

0.56/0.60 0.52/0.65

7 25

Christensen et al. (1996) Mierau et al. (1996)

PepA

Lc. lactis

disrupted

No difference

PepX PepX PepX PepX PepX

Lb. helveticus Lc. lactis Lc. lactis Lc. lactis Lc. lactis

deletion deletion deletion chemical disrupted

0.28/0.61 0.56/0.65 No difference 1.0/1.6 0.70/0.93

PepD

Lb. helveticus

deletion

No difference

PepI

Lb. helveticus

deletion

0.56/0.61

9

Yüksel & Steele (1999)

PepQ

Lb. helveticus

deletion

0.54/0.61

13

Yüksel & Steele (1999)

PepR PepR

Lb. helveticus Lb. rhamnosus

deletion deletion

No difference No difference

Shao et al. (1997) Varmanen et al. (1998)

PepP

Lc. lactis

disrupted

0.89/0.93

Matos et al. (1998)

PepT

Lc. lactis

deletion

0.60/0.65

PepO PepO PepO PepO

Lb. helveticus Lc. lactis Lc. lactis Lc. lactis

deletion disrupted deletion disrupted

No difference No difference 0.60/0.65 No difference

However, there are peptidases in lactobacilli (PepD, PepE, PepI, PepL, PepR) and lactococci (PepA, PepF, PepP) for which a genetic homologue has not been reported in the converse systems. Not yet known is whether this reflects differences in the peptidase genotypic set or simply peptidase genes that have yet to be identified and sequenced. In addition, while the regulation of peptidase expression has been examined in complex media, only a few studies have examined the expression of these peptidases during growth in milk or under the conditions these organisms encounter in ripening cheese. The potential differences in lactobacilli and lactococci pertaining to the pathway of hydrolysis of casein derived peptides for

l’Anson et al. (1995) 120 15 60 31

Yüksel & Steele (1999) Mierau et al. (1996) Mayo et al. (1993) Nardi et al. (1991) Matos et al. (1998) Dudley et al. (1996)

10

9

Mierau et al. (1996) Chen & Steele (1998) Mierau et al. (1993) Mierau et al. (1996) Tynkkynen et al. (1993)

growth in milk, or as they impact the flavor and texture of fermented foods, remain to be determined. The physiological role of individual peptidases have been investigated from numerous single and multiple peptidase mutants constructed in Lb. helveticus and Lc. lactis. The growth rate of mutants have been evaluated in milk and a general decrease in growth rates has been observed (see Table 2). However, the mechanism for this reduction in growth rate has not been determined. Information on peptide accumulation in peptidase mutants grown in milk may assist in elucidating the role of specific peptidases. Additionally, little information is available regarding the potential role of these peptidases in cellular protein turnover and post-translational modification. Peptidase mutant

239 strains may be valuable for subsequent screening of activities not identified in the presence of enzymes in the wild type strains with higher and/or overlapping specificities. This may be important for identification of peptidases that have important physiological roles but have not or can not be identified using small peptides or amino acid chromogens as substrates. Relatively little information is available concerning amino acid catabolism in LAB. In the past few years there has been an increased interest in elucidating the pathways of amino acid catabolism by LAB and the production of sensory compounds and biogenic amines. Recent studies have implicated amino acid catabolism in the regulation of intracellular pH and the generation of metabolic energy. However, we do not know the significance of these pathways for cell survival in a carbohydrate-starved environment, as in ripening cheese. As more genes involved in these pathways are characterized, it will be possible to examine their regulation and construct isogenic strains lacking a particular enzyme activity. Ideally, a more thorough understanding of the peptidolytic/catabolic pathways present in LAB would focus engineering or selection for strains with controlled production of amino acid catabolites. This may include reduction of biogenic amine accumulation, a human health consideration, as well as controlled production of specific flavor precursors or compounds in fermented foods.

Acknowledgments The authors would like to thank our colleagues Jeff Broadbent and Carmen Peláez for communicating results prior to publication. The work on peptidases and amino acid catabolism conducted in our laboratory was supported in part by the College of Agricultural and Life Sciences at the University of WisconsinMadison and the Center for Dairy Research through funding from Dairy Management Inc.

References Abe K, Hayashi H & Maloney PC (1996) Exchange of aspartate and alanine. Mechanism for development of a proton-motive force in bacteria. J. Biol. Chem. 271: 3079–3084 Alting AC & Engels WJM (1996) Conversion of methionine by enzymes from Lactococcus lactis subsp. cremoris B78 during cheese ripening. poster K3, Federation of European Microbiological Societies Fifth Symposium on Lactic Acid Bacteria, 8–12 September, Veldhoven, The Netherlands

Arnau J, Jorgensen F, Madsen SM, Vrang A & Israelsen H (1998) Cloning of the Lactococcus lactis adhE gene, encoding a multifunctional alcohol dehydrogenase, by complementation of a fermentative mutant of Escherichia coli. J. Bacteriol. 180: 3049– 3055 Arora G & Lee BH (1992) Purification and characterization of aminopeptidase from Lactobacillus casei ssp. casei LLG. J. Dairy Sci. 75: 700–710 Arora G & Lee BH (1994) Purification and characterization of an aminopeptidase from Lactobacillus casei subsp. rhamnosus S93. Biotech. Appl. Biochem. 19: 179–192 Atlan D, Gilbert C, Blanc B & Portalier R (1994) Cloning, sequencing and characterization of the pepIP gene encoding a proline iminopeptidase from Lactobacillus delbrueckii subsp. bulgaricus CNRZ 397. Microbiol. 140: 527–535 Atlan D, Laloi P & Portalier R (1990) X-Prolyldipeptidyl aminopeptidase of Lactobacillus delbrueckii ssp. bulgaricus: Characterization of the enzyme and isolation of deficient mutants. Appl. Environ. Microbiol. 56: 2174–2179 Atiles MW, Dudley EG & Steele JL (1999) Characterization and inactivation of the branched-chain aminotransferase gene from Lactococcus lactis LM0230. Manuscript in preparation Axelsson L (1998) Lactic acid bacteria: classification and physiology. In: Salminen S & von Wright A (Eds) Lactic acid bacteria. Microbiology and functional aspects, (pp 1–72). Marcel Dekker, Inc., New York Baankreis R & Exterkate FA (1991) Characterisation of a peptidase from Lactococcus lactis ssp. cremoris HP that hydrolyses diand tripeptides containing proline or hydrophobic residues as the amino-terminal amino acid. Syst. Appl. Microbiol. 14: 317–323 Baankreis R, Vanschalkwijk S, Alting AC & Exterkate FA (1995) The occurrence of two intracellular oligoendopeptidases in Lactococcus lactis and their significance for peptide conversion in cheese. Appl. Microbiol. Biotech. 44: 386–392 Bacon CL, Jennings PV, Fhaolain IN & O’Cuinn G (1994) Purification and characterisation of an aminopeptidase A from cytoplasm of Lactococcus lactis subsp. cremoris AM2. Int. Dairy. J. 4: 503–519 Bacon CL, Wilkinson M, Jennings PV, Fhaolain IN & O’Cuinn G (1993) Purification and characterization of an aminotripeptidase from cytoplasm of Lactococcus lactis subsp. cremoris AM2. Int. Dairy. J. 3: 163–177 Bernard N, Johnsen K, Ferain T, Garmyn D, Hols P, Holbrook JJ & Delcour J (1994) NAD+ -dependent D-2-hydroxyisocaproate dehydrogenase of Lactobacillus delbrueckii subsp. bulgaricus. Gene cloning and enzyme characterization. Eur. J. Biochem. 224: 439–446 Biede SL, Paulsen PV, Hammond EG & Glatz BA (1979) The flavor of Swiss cheese. In: Underkofler LA (Ed) Developments in Industrial Microbiology, (pp 203–210). Society for Industrial Microbiology, Arlington, VA Blanc B, Laloi P, Atlan D, Gilbet C & Portalier R (1993) Two cellwall-associated aminopeptidases from Lactobacillus helveticus and the purification and characterization of APII from strain ITGL1. J. Gen. Microbiol. 139: 1441–1448 Bockelmann W, Beuck HP, Lick S & Heller K (1995) Purification and characterization of a new tripeptidase from Lactobacillus delbrueckii ssp. bulgaricus B14. Int. Dairy. J. 5: 493–502 Bockelmann W, Fobker M & Teuber M (1991) Purification and characterization of the X-prolyl-dipeptidyl-aminopeptidase from Lactobacillus delbrückii subsp. bulgaricus and Lactobacillus acidophilus. Int. Dairy. J. 1: 51–66

240 Bockelmann W, Gollan V & Heller KJ (1997) Purification of a second tripeptidase from Lactobacillus delbrueckii subsp. bulgaricus B14. Milchwissenschaft 52: 500–503 Bockelmann W, Hoppeseyler T & Heller KJ (1996) Purification and characterization of an endopeptidase from Lactobacillus delbrueckii subsp. bulgaricus B14. Int. Dairy. J. 6: 1167–1180 Bockelmann W, Shulz Y & Teuber M (1992) Purification and characterization of an aminopeptidase from Lactobacillus delbrückii subsp. bulgaricus. Int. Dairy. J. 2: 95–107 Boeker EA & Snell EE (1972) Amino acid decarboxylases. In PD Boyer (Ed.) The Enzymes. 6: 217–253 Booth M, Donnelly WJ, Fhaoláin IN, Jennings PV & O’Cuinn G (1990a) Proline-specific peptidases of Streptococcus cremoris AM2. J. Dairy Res. 57: 79–88 Booth M, Fhaoláin IN, Jennings PV & O’Cuinn G (1990b) Purification and characterization of a post-proline dipeptidyl aminopeptidase from Streptococcus cremoris AM2. J. Dairy Res. 57: 89–100 Booth M, Jennings V, Ni Fhaoláin I & O’Cuinn G (1990c) Prolidase activity of Lactococcus lactis ssp. cremoris AM2: Partial purification and characterization. J. Dairy Res. 57: 245–254 Bosman BW, Tan PST & Konings WN (1990) Purification and characterization of a tripeptidase from Lactococcus lactis ssp. cremoris Wg2. Appl. Environ. Microbiol. 56: 1839–1843 Bosset JO, Collomb M & Sieber R (1993) The aroma composition of Swiss Gruyère cheese IV. The acidic volatile components and their changes in content during ripening. Lebensm. Wissen. Technol. 26: 581–592 Braun SD & Olson NF (1986) Microencapsulation of cell-free extracts to demonstrate the feasibility of heterogeneous enzyme systems and cofactor recycling for development of flavor in cheese. J. Dairy Sci. 69: 1202–1208 Brennand CP, Ha JK & Lindsay RC (1989) Aroma properties and thresholds of some branched-chain and other minor volatile fatty acids occurring in milkfat and meat lipids. J. Sensory Studies 4: 105–120 Brückner R (1998) Direct sequence submission to GenBank for gltA (glucose uptake protein) from Staphylococcus xylosus (Accession Y14043). Unpublished Chang GW & Snell EE (1968) Histidine decarboxylase of Lactobacillus 30a. II. Purification, substrate specificity, and stereospecificity. Biochemistry 7: 2005–2012 Chapot-Chartier MP, Nardi M, Chopin MC, Chopin A & Gripon JC (1993) Cloning and sequencing of pepC, a cysteine aminopeptidase gene from Lactococcus lactis ssp. cremoris AM2. Appl. Environ. Microbiol. 59: 330–333 Chapot-Chartier MP, Rul F, Nardi M & Gripon JC (1994) Gene cloning and characterization of PepC, a cysteine aminopeptidase from Streptococcus thermophilus, with sequence similarity to the eukaryotic bleomycin hydrolase. Eur. J. Biochem. 224: 497–506 Chen YS & Steele JL (1998) Genetic characterization and physiological role of endopeptidase O from Lactobacillus helveticus CNRZ32. Appl. Environ. Microbiol. 64: 3411–3415 Chich J-F, Rigolet P, Nardi M, Gripon J-C, Ribadeau-Dumas B & Brunie S (1995) Purification, crystallization, and preliminary X-ray analysis of PepX, an X-Prolyl dipeptidyl aminopeptidase from Lactococcus lactis. Prot. Struct. Funct. Genet. 23: 278–281 Chopin A (1993) Organization and regulation of genes for amino acid biosynthesis in lactic acid bacteria. FEMS Microbiol. Rev. 12: 21–38 Christensen JE, Lin DL, Palva A & Steele JL (1995) Sequence analysis, distribution and expression of an aminopeptidase Nencoding gene from Lactobacillus helveticus CNRZ32. Gene 155: 89–93

Christensen JE & Steele JL (1996) Characterization of peptidasedeficient Lactobacillus helveticus CNRZ32 derivatives. Fifth Symposium on Lactic Acid Bacteria, Veldhoven, The Netherlands Christensen JE & Steele JL (1998a) Hydrolysis of casein derived peptides by peptidase-deficient Lactobacillus helveticus CNRZ32 derivatives. Joint Meeting of the American Dairy Science Association and the American Society of Animal Science. Journal of Dairy Science, Denver, Colorado, USA. Christensen JE & Steele JL (1998b) Unpublished results Collins MD, Rodrigues U, Ash C, Aguirre M, Farrow JAE, Martinez-Murcia A, Phillips BA, Williams AM & Wallbanks S (1991) Phylogenetic analysis of the genus Lactobacillus and related lactic acid bacteria as determined by reverse transcriptase sequencing of 16S rRNA. FEMS Microbiol. Lett. 77: 5–12 Copeland WC, Domena JD & Robertus JD (1989) The molecular cloning, sequence and expression of the hdcB gene from Lactobacillus 30A. Gene 85: 259–265 Crow VL & Thomas TD (1982) Arginine metabolism in lactic streptococci. J. Bacteriol. 150: 1024–1032 Cunin R, Glansdorff N, Piérard A & Stalon V (1986) Biosynthesis and metabolism of arginine in bacteria. Microbiol. Rev. 50: 314– 352 de Mendoza D, Grau R & Cronan JE Jr (1993) Biosynthesis and function of membrane lipids. In: Sonenshein AL, Hock JA & Losick R (Eds) Bacillus subtilis and other gram-positive bacteria: biochemistry, physiology, and molecular genetics, (pp 411–421). American Society for Microbiology, Washington, DC Desmazeaud MJ & Zevaco C (1976) General properties and substrate specificity of an intracellular neutral protease from Streptococcus diacetilactis. Ann. Biol. Anim. Biochim. Biophys. 16: 851–868 Desmazeaud MJ & Zevaco C (1977) General properties and substrate specificity of an intracellular soluble dipeptidase from Streptococcus diacetilactis. Ann. Biol. Anim. Biochim. Biophys. 17: 723–736 Desmazeaud MJ & Zevaco C (1979) Isolation and general properties of two intracellular amino peptidases of Streptococcus diacetylactis. Milchwissenschaft 34: 606–610 Detmers FJM, Kunji ERS, Lanfermeijer FC, Poolman B & Konings WN (1998) Kinetics and specificity of peptide uptake by the oligopeptide transport system of Lactococcus lactis. Biochem. 37: 16671–16679 Doeglas HM, Huisman J & Nater JP (1967) Histamine intoxication after cheese. Lancet 2: 1361–1372 Dudley EG, Husgen AC, He W & Steele JL (1996) Sequencing, distribution, and inactivation of the dipeptidase A gene (pepDA) from Lactobacillus helveticus CNRZ32. J. Bacteriol. 178: 701– 704 Dudley EG & Steele JL (1994) Nucleotide sequence and distribution of the pepPN gene from Lactobacillus helveticus CNRZ32. FEMS Microbiol. Lett. 119: 41–45 Dudley EG & Steele JL (1999) Characterization and inactivation of the aspartate aminotransferase gene from Lactococcus lactis LM0230. Manuscript in preparation Dunn HC & Lindsay RC (1985) Evaluation of the role of microbial Strecker-derived aroma compounds in unclean-type flavors of Cheddar cheese. J. Dairy Sci. 68: 2859–2874 Eggimann B & Bachmann M (1980) Purification and partial characterization of an aminopeptidase from Lactobacillus lactis. Appl. Environ. Microbiol. 40: 876–882 Exterkate FA, De Jong M, De Veer GJCM & Baankreis R (1992) Location and characterization of aminopeptidase N in Lacto-

241 coccus lactis ssp. cremoris HP. Appl. Microbiol. Biotech. 37: 46–54 Exterkate FA & De Veer GJCM (1987) Purification and some properties of a membrane-bound aminopeptidase A from Streptococcus cremoris. Appl. Environ. Microbiol. 53: 577–583 Feil IK, Hendle J & Schomburg D (1997) Modified substrate specificity of L-hydroxyisocaproate dehydrogenase derived from structure-based protein engineering. Protein Eng 10: 255–262 Feil IK, Lerch HP & Schomburg D (1994) Deletion variants of L-hydroxyisocaproate dehydrogenase. Probing substrate specificity. Eur. J. Biochem. 223: 857–863 Fenster KM, Parkin KL & Steele JL (1997) Characterization of a thiol-dependent endopeptidase from Lactobacillus helveticus CNRZ32. J. Bacteriol. 179: 2529–2533 Fernández de Palencia P, López de Felipe F, Requena T & Peláez C (1999) The aminopeptidase C (PepC) from Lactobacillus helveticus CNRZ32. A comparative study of PepC from dairy lactic acid bacteria. Submitted Fernández de Palencia P, Pelaez C & Martín-Hernández MC (1997) Characterization of the aminopeptidase system from Lactobacillus casei subsp. casei IFPL 731. J. Agric. Food Chem. 45: 3778–3781 Fernández L, Bhowmik T & Steele JL (1994) Characterization of the Lactobacillus helveticus CNRZ32 pepC gene. Appl. Environ. Microbiol. 60: 333–336 Fernández-Esplá MD & Martín-Hernández MC (1997) Purification and characterization of a dipeptidase from Lactobacillus casei ssp. casei IFPL 731 isolated from goat cheese made from raw milk. J. Dairy Sci. 80: 1497–1504 Fernández-Esplá MD, Martín-Hernández MC & Fox PF (1997) Purification and characterization of a prolidase from Lactobacillus casei subsp. casei IFPL 731. Appl. Environ. Microbiol. 63: 314–316 Gao S, Mooberry ES & Steele JL (1998) Use of 13C nuclear magnetic resonance and gas chromatography to examine methionine catabolism by lactococci. Appl. Environ. Microbiol. 64: 4670–4675 Gao S, Oh DH, Broadbent JR, Johnson ME, Weimer BC & Steele JL (1997) Aromatic amino acid catabolism by lactococci. Lait 77: 371–381 Gao S & Steele JL (1998) Purification and characterization of oligomeric species of an aromatic amino acid aminotransferase from Lactococcus lactis subsp. lactis S3. J. Food Biochem. 22: 197–211 Geis A, Bockelmann W & Teuber M (1985) Simultaneous extraction and purification of a cell wall-associated peptidase and beta-casein specific protease from Streptococcus cremoris AC1. Appl. Microbiol. Biotech. 23: 79–84 Gilbert C, Atlan D, Blanc B & Portalier R (1994) Proline iminopeptidase from Lactobacillus delbrueckii subsp. bulgaricus CNRZ 397: Purification and characterization. Microbiol. 140: 537–542 Gobbetti M, Smacchi E & Corsetti A (1996) The proteolytic system of Lactobacillus sanfrancisco CB1: Purification and characterization of a proteinase, a dipeptidase, and an aminopeptidase. Appl. Environ. Microbiol. 62: 3220–3226 Gottesman S & Maurizi MR (1992) Regulation by proteolysis: energy-dependent proteases and their targets. Microbiol Rev 56: 592–621 Griffith R & Hammond EG (1989) Generation of Swiss cheese flavor components by the reaction of amino acids with carbonyl compounds. J. Dairy Sci. 72: 604–613

Guirard BM & Snell EE (1964) Nutritional requirements of Lactobacillus 30a for growth and histidine decarboxylase production. J. Bacteriol. 87: 370–376 Gummalla S (1998) Tryptophan catabolism by Lactobacillus spp.: Biochemistry and implications on flavor development in reduced-fat cheddar cheese. Master of Science thesis, Utah State University Gummalla S & Broadbent JR (1996) Indole production by Lactobacillus ssp. in cheese: a possible role for tryptophanase. J. Dairy Sci. 79(Suppl. 1): 101 Guthrie BD (1993) Influence of cheese-related microflora on the production of unclean-flavored aromatic amino acid metabolites in Cheddar cheese. Ph.D thesis, University of WisconsinMadison Ha KJ & Lindsay RC (1991a) Contributions of cow, sheep, and goat milks to characterizing branched-chain fatty acid and phenolic flavors in varietal cheeses. J. Dairy Sci. 74: 3267–3274 Ha KJ & Lindsay RC (1991b) Volatile branched-chain fatty acids and phenolic compounds in aged Italian cheese flavors. J. Food Sci. 56: 1241–1247, 1250 Habibi-Najafi MB & Lee BH (1994) Purification and characterization of X-prolyl dipeptidyl peptidase from Lactobacillus casei subsp. casei LLG. Appl. Microbiol. Biotech. 42: 280–286 Hackert ML, Meador WE, Oliver RM, Salmon JB, Recsei PA & Snell EE (1981) Crystallization and subunit structure of histidine decarboxylase from Lactobacillus 30a. J. Biol. Chem. 256: 687– 690 Harrison GA, Khairallah EA & Morgan ME (1969) Alcohol dehydrogenase in Streptococcus lactis var. maltigenes: In vivo relationship to glycolysis, in vitro kinetics. poster P109. American Society for Microbiology 69th Annual Meeting, 4–9 May, Miami Beach, Fla Hellendoorn MA, Franke-Fayard BMD, Mierau I, Venema G & Kok J (1997) Cloning and analysis of the pepV dipeptidase gene of Lactococcus lactis MG1363. J. Bacteriol. 179: 3410–3415 Hickey MW, Hillier AJ & Jago GR (1983) Enzymic activities associated with lactobacilli in dairy products. Aust. J. Dairy Technol. 38: 154–157 Higuchi T, Hayashi H & Abe K (1997) Exchange of glutamate and γ -aminobutyrate in a Lactobacillus strain. J. Bacteriol. 179: 3362–3364 Hoffman K, Bucher P, Falquet L & Bairoch A (1999) The PROSITE database, its status in 1999. Nuc. Acids Res. 27: 215–219 Honeyfield DC & Carlson JR (1990) Assay for the enzymatic conversion of indoleacetic acid to 3-methylindole in a ruminal Lactobacillus species. Appl. Environ. Microbiol. 56: 724–729 Houbart V, Ribadeau-Dumas B & Chich JF (1995) Synthesis of enterostatin-amide by the Xaa-prolyl dipeptidyl aminopeptidase from Lactococcus lactis subsp. lactis NCDO 763. Biotech. Appl. Biochem. 21: 149–159 Hummel W & Kula MR (1989) Dehydrogenases for the synthesis of chiral compounds. Eur. J. Biochem. 184: 1–13 Hummel W, Schmidt E, Wandrey C & Kula MR (1986) Lphenylalanine dehydrogenase from Brevibacterium sp. for the production of L-phenylalaine by reductive amination of phenylpyruvate. Appl. Microbiol. Biotechnol. 25: 175–185 Hummel W, Schütte H & Kula M-R (1985) D-2-hydroxyisocaproate dehydrogenase from Lactobacillus casei. Appl. Microbiol. Biotechnol. 21: 7–15 Hummel W, Schütte H & Kula M-R (1988) D-(-)-mandelic acid dehydrogenase from Lactobacillus curvatus. Appl. Microbiol. Biotechnol. 28: 433–439

242 Hummel W, Weiss N & Kula MR (1984) Isolation and characterization of a bacterium possessing L-phenylalanine dehydrogenase activity. Arch. Microbiol. 137: 47–52 Huynh QK, Recsei PA, Vaaler GL & Snell EE (1984) Histidine decarboxylase of Lactobacillus 30a. Sequences of the overlapping peptides, the complete alpha chain, and prohistidine decarboxylase. J. Biol. Chem. 259: 2833–2839 Hwang IK, Kaminogawa S & Yamauchi K (1981) Purification and properties of a dipeptidase from Streptococcus cremoris. Agric. Biol. Chem. 45: 159–166 Jackson HW & Morgan ME (1954) Identity and origin of the malty aroma substance from milk cultures of Streptococcus lactis var. maltigenes. J. Dairy Sci. 37: 1316–1324 Joosten HM & Northolt MD (1989) Detection, growth, and amineproducing capacity of lactobacilli in cheese. Appl. Environ. Microbiol. 55: 2356–2359 Joosten HM & Stadhouders J (1987) Conditions allowing the formation of biogenic amines in cheese. 1. Decarboxylative properties of starter bacteria. Neth. Milk Dairy J. 41: 247–258 Joshua-Tor L, Xu HE, Johnston SA & Rees DC (1995) Crystal structure of a conserved protease that binds DNA: the bleomycin hydrolase, Gal6. Science 269: 945–950 Kahana LM & Todd E (1981) Histamine poisoning and reaction to cheese. Ann. Intern. Med. 88: 520–521 Kallwass HKW (1992) Potential of R-2-hydroxyisocaproate dehydrogenase from Lactobacillus casei for stereospecific reductions. Enzyme Microb. Technol. 14: 28–35 Kaminogawa S, Azuma N, Hwang IK, Suzuki Y & Yamauchi K (1984) Isolation and characterization of a prolidase from Streptococcus cremoris H61. Agric. Biol. Chem. 48: 3035–3040 Khalid NM & Marth EH (1990a) Partial purification and characterization of an aminopeptidase from Lactobacillus helveticus CNRZ 32. Syst. Appl. Microbiol. 13: 311–319 Khalid NM & Marth EH (1990b) Purification and partial characterization of a prolyl-dipeptidyl aminopeptidase from Lactobacillus helveticus CNRZ 32. Appl. Environ. Microbiol. 56: 381–388 Kiefer-Partsch B, Bockelmann W, Geis A & Teuber M (1989) Purification of an X-prolyldipeptidyl aminopeptidase from the cell wall proteolytic system of Lactococcus lactis ssp. cremoris. Appl. Microbiol. Biotech. 31: 75–78 Kingsley R, Rabsch W, Roberts M, Reissbrodt R & Williams PH (1996) TonB-dependent iron supply in Salmonella by αketoacids and α-hydroxyacids. FEMS Microbiol. Lett. 140: 65–70 Klein JR, Dick A, Schick J, Matern HT, Henrich B & Plapp R (1995) Molecular cloning and DNA sequence analysis of pepL, a leucyl aminopeptidase gene from Lactobacillus delbrueckii subsp. lactis DSM7290. Eur. J. Biochem. 228: 570–578 Klein JR, Henrich B & Plapp R (1994a) Cloning and nucleotide sequence analysis of the Lactobacillus delbrueckii ssp. lactis DSM7290 cysteine aminopeptidase gene pepC. FEMS Microbiol. Lett. 124: 291–299 Klein JR, Klein U, Schad M & Plapp R (1993) Cloning, DNA sequence analysis and partial characterization of pepN, a lysyl aminopeptidase from Lactobacillus delbrueckii ssp. lactis DSM7290. Eur. J. Biochem. 217: 105–114 Klein JR, Schick J, Henrich B & Plapp R (1997) Lactobacillus delbrueckii subsp. lactis DSM7290 pepG gene encodes a novel cysteine aminopeptidase. Microbiol. 143: 527–537 Klein JR, Schmidt U & Plapp R (1994b) Cloning, heterologous expression, and sequencing of a novel proline iminopeptidase gene, pepI, from Lactobacillus delbrueckii subsp. lactis DSM 7290. Microbiol. 140: 1133–1139

Kochhar S, Hunziker PE, Leong-Morgenthaler P & Hottinger H (1992) Evolutionary relationship of NAD+ -dependent D-lactate dehydrogenase: comparison of primary structure of 2-hydroxy acid dehydrogenases. Biochem. Biophys. Res. Comm. 184: 60– 66 Kong Y, Strickland M & Broadbent JR (1996) Tyrosine and phenylalanine catabolism by Lactobacillus casei flavor adjuncts: biochemistry and implications in cheese flavor. J. Dairy Sci. 79(Suppl. 1): 101 Konings WN, Lolkema JS, Bolhuis H, van Veen HW, Poolman B & Driessen AJ (1997) The role of transport processes in survival of lactic acid bacteria. Energy transduction and multidrug resistance. Antonie Van Leeuwenhoek 71: 117–128 Konings WN, Lolkema JS & Poolman B (1995) The generation of metabolic energy by solute transport. Arch. Microbiol. 164: 235– 242 Kunji ERS, Fang G, Jeronimus-Stratingh CM, Bruins AP, Poolman B & Konings WN (1998) Reconstruction of the proteolytic pathway for use of beta-casein by Lactococcus lactis. Mol. Microbiol. 27: 1107–1118 Kunji ERS, Hagting A, De Vries CJ, Juillard V, Haandrikman AJ, Poolman B & Konings WN (1995) Transport of beta-casein derived peptides by the oligopeptide transport system is a crucial step in the proteolytic pathway of Lactococcus lactis. J. Biol. Chem. 270: 1569–1574 Kunji ERS, Mierau I, Hagting A, Poolman B & Konings WN (1996) The proteolytic systems of lactic acid bacteria. Antonie van Leeuwenhoek 70: 187–221 l’Anson KJA, Movahedi S, Griffin HG, Gasson MJ & Mulholland F (1995) A non-essential glutamyl aminopeptidase is required for optimal growth of Lactococcus lactis MG1363 in milk. Microbiol. 141: 2873–2881 Lapujade P, Cocaign-Bousquet M & Loubiere P (1998) Glutamate biosynthesis in Lactococcus lactis subsp. lactis NCDO 2118. Appl. Environ. Microbiol. 64: 2485–2489 Law BA (1981) The formation of aroma and flavor compounds in fermented dairy products. Dairy Sci. Abstr. 43: 143 Law J & Haandrikman A (1997) Proteolytic enzymes of lactic acid bacteria. Int. Dairy. J. 7: 1–11 Lees GJ & Jago GR (1976) Formation of acetaldehyde from threonine by lactic acid bacteria. J. Dairy Res. 43: 75–83 Lerch H-P, Frank R & Collins J (1989) Cloning, sequencing and expression of the L-2-hydroxyisocaproate dehydrogenaseencoding gene of Lactobacillus confusus in Escherichia coli. Gene 83: 263–270 Lian W, Wu D, Konings WN, Mierau I & Hersh LB (1996) Heterologous expression and characterization of recombinant Lactococcus lactis neutral endopeptidase (Neprilysin). Arch. Biochem. Biophys. 333: 121–126 Liu S-Q & Pilone GJ (1998) A review: arginine metabolism in wine lactic acid bacteria and its practical significance. J. Appl. Microbiol. 84: 315–327 Liu S-Q, Pritchard GG, Hardman MJ & Pilone GJ (1995) Occurance of arginine deiminase pathway enzymes in arginine catabolism by wine lactic acid bacteria. Appl. Environ. Microbiol. 61: 310– 316 Lloyd RJ & Pritchard GG (1991) Characterization of Xprolyldipeptidylaminopeptidase from Lactococcus lactis ssp. lactis. J. Gen. Microbiol. 137: 49–56 Machuga EJ & Ives DH (1984) Isolation and characterization of an aminopeptidase (EC 3.4.11.1) from Lactobacillus acidophilus R-26. Biochim. Biophys. Acta 789: 26–36

243 MacLeod P & Morgan ME (1958) Differences in the ability of lactic streptococci to form aldehydes from certain amino acids. J. Dairy Sci. 41: 908–913 Magboul AAA & McSweeney PLH (1998) Purification and characterisation of an aminopeptidase from Lactobacillus curvatus DPC 2024. Aust. J. Dairy Tech. 53: 114 Manca de Nadra MC, Pese de Ruiz Holgado AA & Oliver G (1982) Arginine dihydrolase activity in lactic acid bacteria. Milchwissenschaft 37: 669–670 Manca De Nadra MC, Raya RR, Pesce De Ruiz Holgado A & Oliver G (1987) Isolation and properties of threonine aldolase of Lactobacillus bulgaricus YOP12 . Milchwissenschaft 42: 92–94 Marquis RE, Bender GR, Murray DR & Wong A (1987) Arginine deiminase system and bacterial adaptation to acid environments. Appl. Environ. Microbiol. 53: 198–200 Marranzini RM, Schmidt RH, Shireman RB, Marshall MR & Cornell JA (1989) Effect of threonine and glycine concentrations on threonine aldolase activity of yogurt microorganisms during growth in a modified milk prepared by ultrafiltration. J. Dairy Sci. 72: 1142–1148 Mars I & Monnet V (1995) An aminopeptidase P from Lactococcus lactis with original specificity. Biochim. Biophys. Acta 1243: 209–215 Marshall VM & Cole WM (1983) Threonine aldolase and alcohol dehydrogenase activities in Lactobacillus bulgaricus and Lactobacillus acidophilus and their contribution to flavour production in fermented milks. J. Dairy Res. 50: 375–379 Marugg JD, Meijer W, van Kranenburg R, Laverman P, Bruinenberg PG & de Vos WM (1995) Medium-dependent regulation of proteinase gene expression in Lactococcus lactis: control of transcription initiation by specific dipeptides. J. Bacteriol. 177: 2982–2989 Massey LK, Sokatch JR & Conrad RS (1976) Branched-chain amino acid catabolism in bacteria. Bacteriol. Rev. 40: 42–54 Masson F, Eclache L, Compte T, Talon R & Montel MC (1996) Screening of microbial strains producing amines and isolated from meat products. In Proceedings of 42nd International Congress of meat science and technology. pp 546–547, Sept. 1–6, Lillehammer, Norway Mata L, Erra-Pujada M, Gripon J-C & Mistou M-Y (1997) Experimental evidence for the essential role of the C-terminal residue in the strict aminopeptidase activity of the thiol aminopeptidase PepC, a bacterial bleomycin hydrolase. Biochem. J. 328: 343–347 Mathiopoulos C, Mueller JP, Slack FJ, Murphy CG, Patankar S, Bukusoglu G & Sonenshein AL (1991) A Bacillus subtilis dipeptide transport system expressed early during sporulation. Mol Microbiol 5: 1903–1913 Matos J, Nardi M, Kumura H & Monnet V (1998) Genetic characterization of PepP, which encodes an aminopeptidase P whose deficiency does not affect Lactococcus lactis growth in milk, unlike deficiency of the X-Prolyl dipeptidyl aminopeptidase. Appl. Environ. Microbiol. 64: 4591–4595 Mayo B, Kok J, Bockelmann W, Haandrikman A, Leenhouts KJ & Venema G (1993) Effect of X-prolyl dipeptidyl aminopeptidase deficiency on Lactococcus lactis. Appl. Environ. Microbiol. 59: 2049–2055 Mayo B, Kok J, Venema K, Bockelmann W, Teuber M, Reinke H & Venema G (1991) Molecular cloning and sequence analysis of the X-prolyldipeptidyl aminopeptidase gene from Lactococcus lactis ssp. cremoris. Appl. Environ. Microbiol. 57: 38–44 McCabe BJ (1986) Dietary tyramine and other pressor amines in MAOI regimens: A review. J. Amer. Diet. Assoc. 86: 1059–1064

Meijer W, Marugg JD & Hugenholtz J (1996) Regulation of proteolytic enzyme activity in Lactococcus lactis. Appl. Environ. Microbiol. 62: 156–161 Meyer J & Jordi R (1987) Purification and characterization of Xprolyldipeptidylaminopeptidase from Lactobacillus lactis and from Streptococcus thermophilus. J. Dairy Sci. 70: 738–745 Meyer-Barton EC, Klein JR, Imam M & Plapp R (1993) Cloning and sequence analysis of the X-prolyl-dipeptidylaminopeptidase gene (pepX) from Lactobacillus delbrueckii ssp. lactis DSM7290. Appl. Microbiol. Biotech. 40: 82–89 Midwinter RG & Pritchard GG (1994) Aminopeptidase N from Streptococcus salivarius subsp. thermophilus NCDO 573: Purification and properties. J. Appl. Bacteriol. 77: 288–295 Mierau I, Haandrikman AJ, Velterop O, Tan PST, Leenhouts KL, Konings WN, Venema G & Kok J (1994) Tripeptidase gene (pepT) of Lactococcus lactis: Molecular cloning and nucleotide sequencing of pepT and construction of a chromosomal deletion mutant. J. Bacteriol. 176: 2854–2861 Mierau I, Kunji ERS, Leenhouts KJ, Hellendoorn MA, Haandrikman AJ, Poolman B, Konings WN, Venema G & Kok J (1996a) Multiple-peptidase mutants of Lactococcus lactis are severely impaired in their ability to grow in milk. J. Bacteriol. 178: 2794–2803 Mierau I, Kunji ERS, Venema G, Poolman B & Kok J (1996b) Peptidases and growth of Lactococcus lactis in milk. Lait 76: 25–32 Mierau I, Tan PST, Haandrikman AJ, Mayo B, Kok J, Leenhouts KJ, Konings WN & Venema G (1993) Cloning and sequencing of the gene for a lactococcal endopeptidase, an enzyme with sequence similarity to mammalian enkephalinase. J. Bacteriol. 175: 2087– 2096 Miller A, Morgan ME & Libbey LM (1974) Lactobacillus maltaromicus, a new species producing a malty aroma. Int. J. System. Bacteriol. 24: 346–354 Miller CG (1987) Protein degradation and proteolytic modification. Pages 680-691 in FC Neidhardt JLI, KB Low, B Magasanik, M Schaechter & HE Umbarger (Ed.) Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. American Society for Microbiology, Washington, DC Miller CG & Green L (1981) Degradation of abnormal proteins in peptidase-deficient mutants of Salmonella typhimurium. J. Bacteriol. 147: 925–930 Miller CG & Green L (1983) Degradation of proline peptides in peptidase-deficient strains of Salmonella typhimurium. J. Bacteriol. 153: 350–356 Milo C & Reineccius GA (1997) Identification and quantification of potent odorants in regular-fat and low-fat mild Cheddar cheese. J. Agric. Food Chem. 45: 3590–3594 Misono H, Goto N & Nagasaki S (1985) Purification, crystallization and properties of NADP+ -specific glutamate dehydrogenase from Lactobacillus fermentum. Agric. Biol. Chem. 49: 117–123 Mistou M-Y & Gripon J-C (1998) Catalytic properties of the cysteine aminopeptidase PepC, a bacterial bleomycin hydrolase. Biochim. Biophys. Acta 3: 63–70 Mistou M-Y, Rigolet P, Chapot-Chartier MP, Nardi M, Gripon J-C & Brunie S (1994) Crystallization and preliminary X-ray analysis of PepC, a thiol aminopeptidase from Lactococcus lactis homologous to bleomycin hydrolase. J. Mol. Biol. 237: 160–162 Miyakawa H, Hashimoto I, Nakamura T, Ishibashi N, Shimamura S & Igoshi K (1994) Purification and characterization of an Xprolyl dipeptidyl aminopeptidase from Lactobacillus helveticus LHE-511. Milchwissenschaft 49: 670–673 Miyakawa H, Kobayashi S, Shimamura S & Tomita M (1991) Purification and characterization of an X-prolyldipeptidyl

244 aminopeptidase from Lactobacillus delbrueckii ssp. bulgaricus LBU-147. J. Dairy Sci. 74: 2375–2381 Miyakawa H, Kobayashi S, Shimamura S & Tomita M (1992) Purification and characterization of an aminopeptidase from Lactobacillus helveticus LHE-511. J. Dairy Sci. 75: 27–35 Molenaar D, Bosscher JS, Ten Brink B, Driessen AJM & Konings WN (1993) Generation of a proton motive force by histidine decarboxylation and electrogenic histidine/histamine antiport in Lactobacillus buchneri. J. Bacteriol. 175: 2864–2870 Monnet V, Nardi M, Chopin A, Chopin M-C & Gripon J-C (1994) Biochemical and genetic characterization of PepF, an oligopeptidase from Lactococcus lactis. J. Biol. Chem. 269: 32070–32076 Montel MC, Seronie MP, Talon R & Hebraud M (1995) Purification and characterization of a dipeptidase from Lactobacillus sake. Appl. Environ. Microbiol. 61: 837–839 Morel F, Frot-Coutaz J, Aubel D, Portalier R & Atlan D (1997) Direct sequence submission to GenBank for a prolidase from Lactobacillus debrueckii ssp. bulgaricus CNRZ397 (Accession 3821250). Unpublished Morel F, Gilbert C, Geourjon C, Frot-Coutaz J, Portalier R & Atlan D (1999a) The prolyl aminopeptidase from Lactobacillus delbrueckii subsp. bulgaricus belongs to the alpha/beta hydrolase fold family. Biochim. Biophys. Acta-Protein Struct. Molec. Enzym. 1429: 501–505 Morel F, Frot-Coutaz J, Aubel D, Portalier R & Atlan D (1999b) Characterization of a prolidase from Lactobacillus delbrueckii subsp. bulgaricus CNRZ 397 with an unusual regulation of biosynthesis. Microbiol. 145:437–446 Morgan ME, Lindsay RC, Libbey LM & Pereira RL (1965) Identity of additional aroma constituents in milk cultures of Streptococcus lactis var. maltigenes. J. Dairy Sci. 49: 15–18 Morishita T & Yajima M (1995) Incomplete operation of biosynthetic and bioenergetic functions of the citric acid cycle in multiple auxotrophic lactobacilli. Biosci. Biotech. Biochem. 59: 251–255 Morita H, Yoshikawa H, Sakata R, Nagata Y & Tanaka H (1997) Synthesis of nitric oxide from the two equivalent guanidino nitrogens of L-arginine by Lactobacillus fermentum. J. Bacteriol. 179: 7812–7815 Muset G, Monnet V & Gripon JC (1989) Intracellular proteinase of Lactococcus lactis ssp. lactis NCDO 763. J. Dairy Res. 56: 765–778 Nakae T & Elliott JA (1965) Production of volatile fatty acids by some lactic acid bacteria. II. Selective formation of volatile fatty acids by degradation of amino acids. J. Dairy Sci. 48: 293–299 Nardi M, Chopin MC, Chopin A, Cals MM & Gripon JC (1991) Cloning and DNA sequence analysis of an X-prolyldipeptidyl aminopeptidase gene from Lactococcus lactis ssp. lactis NCDO 763. Appl. Environ. Microbiol. 57: 45–50 Nardi M, Renault P & Monnet V (1997) Duplication of the pepF gene and shuffling of DNA fragments on the lactose plasmid of Lactococcus lactis. J. Bacteriol. 179: 4164–4171 Neviani E, Boquien CY, Monnet V, Thanh LP & Gripon JC (1989) Purification and characterization of an aminopeptidase from Lactococcus lactis ssp. cremoris AM2. Appl. Environ. Microbiol. 55: 2308–2314 Nierop Groot MN & de Bont JAM (1998) Conversion of phenylalanine to benzaldehyde initiated by an aminotransferase in Lactobacillus plantarum. Appl. Environ. Microbiol. 64: 3009–3013 Niven GW (1991) Purification and characterization of aminopeptidase A from Lactococcus lactis ssp. lactis NCDO 712. J. Gen. Microbiol. 137: 1207–1212

Ohmiya K & Sato Y (1975) Purification and properties of intracellular proteinase from Streptococcus cremoris. Applied Microbiology 30: 738–745 Parks EH, Ernst SR, Hamlin R, Xuong NH & Hackert ML (1985) Structure determination of histidine decarboxylase from Lactobacillus 30a at 3.0 A resolution. J. Mol. Biol. 182: 455–465 Patton S (1964) Volatile acids of Swiss cheese. J. Dairy Sci. 47: 817–818 Paulsen PV, Kowalewska J, Hammond EG & Glatz BA (1980) Role of microflora in production of free fatty acids and flavor in Swiss cheese. J. Dairy Sci. 63: 912–918 Paulus H (1993) Biosynthesis of the aspartate family of amino acids. In: Sonenshein AL, Hock JA & Losick R (Eds) Bacillus subtilis and other gram-positive bacteria: biochemistry, physiology, and molecular genetics, (pp 237–267). American Society for Microbiology, Washington, DC Perego M, Higgins CF, Pearce SR, Gallagher MP & Hoch JA (1991) The oligopeptide transport system of Bacillus subtilis plays a role in the initiation of sporulation. Mol Microbiol 5: 173–185 Preininger M & Grosch W (1994) Evaluation of key odorants of the neutral volatiles of Emmentaler cheese by calculation of odour activity values. Lebensm. Wiss. Technol. 27: 237–244 Preininger M, Warmke R & Grosch W (1996) Identification of the character impact flavour compounds of Swiss cheese by sensory studies of models. Z. Lebensm. Unters. Forsch. 202: 30–34 Pritchard GG, Freebairn AD & Coolbear T (1994) Purification and characterization of an endopeptidase from Lactococcus lactis subsp. cremoris SK11. Microbiol. 140: 923–930 Rallu F, Gruss A & Maguin E (1996) Lactococcus lactis and stress. Antonie van Leeuwenhoek 70: 243–251 Rantanen T & Palva A (1997) Lactobacilli carry cryptic genes encoding peptidase-related proteins: Characterization of a prolidase gene (pepQ) and a related cryptic gene (orfZ) from Lactobacillus delbrueckii subsp. bulgaricus. Microbiol. 143: 3899–3905 Rawlings ND, Polgar L & Barrett AJ (1991) A new family of serinetype peptidases related to prolyl oligopeptidase [letter]. Biochem J 279: 907-908 Recsei PA, Huynh QK & Snell EE (1983) Conversion of prohistidine decarboxylase to histidine decarboxylase: peptide chain cleavage by nonhydrolytic serinolysis. Proc. Natl. Acad. Sci. USA 80: 973–977 Rescei PA & Snell EE (1981) In: Metabolic interconversion of enzymes 1980 ed. H.Holzen, pp. 335–344. Springer-Verlag: New York Recsei PA & Snell EE (1984) Pyruvoyl enzymes. Annu. Rev. Biochem. 53: 357–387 Reiter B & Oram JD (1962) Nutritional studies on cheese starters. I. Vitamin and amino acid requirements of single strain starters. J Dairy Res. 29: 63–77 Roig-Sagues AX, Hernandez-Herrero MN, Lopez-Sabater EI, Rodriguez-Jerez JJ & Mora-Ventura MT (1997) Evaluation of three decarboxylating agar media to detect histamine and tyramine-producing bacteria in ripened sausages. Lett. Appl. Microbiol. 25: 309–312 Rollan G, de Nadra MCM, Holgado PR & Oliver G (1985) Aspartate metabolism in Lactobacillus murinus CNRS 313. I. Aspartase. J. Gen. Appl. Microbiol. 31: 403–409 Rollan G, de Nadra MCM, Holgado PR & Oliver G (1988) Aspartate aminotransferase of Lactobacillus murinus. Folia Microbiol. 33: 344–348 Roudot-Algaron F & Yvon M (1998) Le catabolisme des acides aminés aromatiques et des acides aminés à chaîne ramifiée chez Lactococcus lactis. Lait 78: 23–30

245 Rul F, Gripon J-C & Monnet V (1995) St-PepA, a Streptococcus thermophilus aminopeptidase with high specificity for acidic residues. Microbiol. 141: 2281–2287 Rul F, Monnet V & Gripon JC (1994) Purification and characterization of a general aminopeptidase (St-PepN) from Streptococcus salivarius ssp. thermophilus CNRZ 302. J. Dairy Sci. 77: 2880–2889 Sahlstrøm S, Chrzanowska J & Sorhaug T (1993) Purification and characterization of a cell wall peptidase from Lactococcus lactis ssp. cremoris IMN-C12. Appl. Environ. Microbiol. 59: 3076– 3082 Sanders JW, Leenhouts K, Burghoorn J, Brands JR, Venema G & Kok J (1998) A chloride-inducible acid resistance mechanism in Lactococcus lactis and its regulation. Mol. Microbiol. 27: 299– 310 Sandine WE & Elliker PR (1970) Microbially induced flavors in fermented dairy products. J. Agric. Food Chem. 18: 557 Santos MH (1996) Biogenic amines: their importance in foods. Int. J. Food Microbiol. 29: 213–231 Sanz Y, Mulholland F & Toldrá F (1998) Purification and characterization of a tripeptidase from Lactobacillus sake. J. Agric. Food Chem. 46: 349–353 Sasaki M, Bosman BW & Tan PST (1996) A new, broad-substratespecificity aminopeptidase from the dairy organism Lactobacillus helveticus SBT 2171. Microbiol. 142: 799–808 Schütte H, Hummel W & Kula M-R (1984) L-2-hydroxyisocaproate dehydrogenase-a new enzyme from Lactobacillus confusus for the stereospecific reduction of 2-ketocarboxylic acids. Appl. Microbiol. Biotechnol. 19: 167–176 Shankar PA (1977) Interrelationships of Streptococcus thermophilus and Lactobacillus bulgaricus in yoghurt culture. Thesis, University of Reading, Reading, UK Shao W, Yüksel GÜ, Dudley EG, Parkin KL & Steele JL (1997) Biochemical and molecular characterization of PepR, a dipeptidase, from Lactobacillus helveticus CNRZ32. Appl. Environ. Microbiol. 63: 3438–3443 Sheldon RM, Lindsay RC, Libbey LM & Morgan ME (1971) Chemical nature of malty flavor and aroma produced by Streptococcus lactis var. maltigenes. Appl. Microbiol. 22: 263–266 Simitsopoulou M, Vafopoulou A, Choli Papadopoulou T & Alichanidis E (1997) Purification and partial characterization of a tripeptidase from Pediococcus pentosaceus K9.2. Appl. Environ. Microbiol. 63: 4872–4876 Stepaniak L & Fox PF (1995) Characterization of the principal intracellular endopeptidase from Lactococcus lactis subsp. lactis MG1363. Int. Dairy. J. 5: 699–713 Stepaniak L, Gobbetti M, Pripp AH & Sorhaug T (1998a) Isolation and characterization of a 67 kDa oligopeptidase from Propionibacterium freudenreichii ATCC 9614. Ital. J Food Sci. 10: 117–125 Stepaniak L, Gobbetti M & Sorhaug T (1998b) Isolation and characterization of high molecular mass endopeptidase complex from Lactococcus lactis. Milchwissenschaft 53: 255–259 Stepaniak L, Tobiassen RO, Chukwu I, Pripp AH & Sorhaug T (1998c) Purification and characterization of a 33 kDa subunit oligopeptidase from Propionibacterium freudenreichii ATCC 9614. Int. Dairy. J. 8: 33–37 Stratton JE, Hutkins RW, Sumner SS &Taylor SL (1992) Histamine and histamine-producing bacteria in retail Swiss and low-salt cheeses. J. Food Prot. 55: 435–439 Straub BW, Kicherer M, Schilcher SM & Hammes WP (1995) The formation of biogenic amines by fermentation microorganisms. Z. Lebensm. Unters. For. 201: 79–82

Strøman P (1992) Sequence of a gene (lap) encoding a 95.3-kDa aminopeptidase from Lactococcus lactis ssp. cremoris Wg2. Gene 113: 107–112 Stucky K, Klein JR, Schueller A, Matern H, Henrich B & Plapp R (1995) Cloning and DNA sequence analysis of pepQ, a prolidase gene from Lactobacillus delbrueckii subsp. lactis DSM7290 and partial characterization of its product. Mol. Gen. Genet. 247: 494–500 Stucky K, Schick J, Klein JR, Henrich B & Plapp R (1996) Characterization of pepR1, a gene coding for a potential transcriptional regulator of Lactobacillus delbrueckii subsp. lactis DSM7290. FEMS Microbiol. Lett. 136: 63–69 Tamime AY & Deeth HC (1980) Yogurt: technology and biochemistry. J. Food Prot. 43: 939 Tan PST & Konings WN (1990) Purification and characterization of an aminopeptidase from Lactococcus lactis ssp. cremoris Wg2. Appl. Environ. Microbiol. 56: 526–532 Tan PST, Pos KM & Konings WN (1991) Purification and characterization of an endopeptidase from Lactococcus lactis ssp. cremoris Wg2. Appl. Environ. Microbiol. 57: 3593–3599 Tan PST, Sasaki M, Bosman BW & Iwasaki T (1995) Purification and characterization of a dipeptidase from Lactobacillus helveticus SBT 2171. Appl. Environ. Microbiol. 61: 3430–3435 Tan PST, Van Alen Boerrigter IJ, Poolman B, Siezen RJ, De Vos WM & Konings WN (1992) Characterization of the Lactococcus lactis pepN gene encoding an aminopeptidase homologous to mammalian aminopeptidase N. FEBS Lett. 306: 9–16 Tan PST, Van Kessel TAJM, Van De Veerdonk FLM, Zuurendonk PF, Bruins AP & Konings WN (1993) Degradation and debittering of a tryptic digest from beta-casein by aminopeptidase N from Lactococcus lactis ssp. cremoris WG2. Appl. Environ. Microbiol. 59: 1430–1436 Taylor SL, Keefe TJ, Windham ES & Howell JF (1982) Outbreak of histamine poisoning associated with consumption of Swiss cheese. J. Food Prot. 45: 455–457 ten Brink B, Damink C, Joosten HM & Huis in’t Veld JH (1990) Occurrence and formation of biologically active amines in foods. Int. J. Food Microbiol. 11: 73–84 Timpone D & Steele JL (1999) Unpublished observations Tobiassen RO, Pripp AH, Stepaniak L & Sørhaug T (1996) Purification and characterization of an endopeptidase from Propionibacterium freudenreichii. J. Dairy Sci. 79: 2129–2136 Tobiassen RO, Sørhaug T & Stephaniak L (1997) Characterization of an intracellular oligopeptidase from Lactobacillus paracasei. Appl. Environ. Microbiol. 63: 1284–1287 Tsakalidou E, Anastasiou R, Papadimitriou K, Manolopoulou E & Kalantzopoulos G (1998) Purification and characterisation of an intracellular X-prolyl-dipeptidyl aminopeptidase from Streptococcus thermophilus ACA-DC 4. J. Biotech. 59: 203–211 Tsakalidou E, Dalezios I, Georgalaki M & Kalantzopoulos G (1993) A comparative study: Aminopeptidase activities from Lactobacillus delbrueckii ssp. bulgaricus and Streptococcus thermophilus. J. Dairy Sci. 76: 2145–2151 Tsakalidou E & Kalantzopoulos G (1992) Purification and partial characterization of an intracellular aminopeptidase from Streptococcus salivarius ssp. thermophilus strain ACA-DC 114. J. Appl. Bacteriol. 72: 227–232 Tucker JS & Morgan ME (1967) Decarboxylation of α-keto acids by Streptococcus lactis var. maltigenes. Appl. Microbiol. 15: 694– 700 Tynkkynen S, Buist G, Kunji E, Kok J, Poolman B, Venema G & Haandrikman A (1993) Genetic and biochemical characterization of the oligopeptide transport system of Lactococcus lactis. J. Bacteriol. 175: 7523–7532

246 Ueno Y, Hayakawa K, Takahashi S & Oda K (1997) Purification and characterization of glutamate decarboxylase from Lactobacillus brevis IFO 12005. Biosci. Biotech. Biochem. 61: 1168–1171 Umbarger HE (1996) Biosynthesis of the branched-chain amino acids. In: Neidhardt FC (Ed) Escherichia coli and Salmonella: cellular and molecular biology, 2nd Ed, Vol 1 (pp 442–457). ASM Press, Washington DC Urbach G (1995) Contribution of lactic acid bacteria to flavor compound formation in dairy products. Int. Dairy J. 5: 877–903 Van Alen-Boerrigter IJ, Baankreis R & De Vos WM (1991) Characterization and overexpression of the Lactococcus lactis pepN gene and localization of its product, aminopeptidase N. Appl. Environ. Microbiol. 57: 2555–2561 Van Boven A, Tan PST & Konings WN (1988) Purification and characterization of a dipeptidase from Streptococcus cremoris Wg2. Appl. Environ. Microbiol. 54: 43–49 Vanderslice P, Copeland WC & Robertus JD (1986) Cloning and nucleotide sequence of wild type and a mutant histidine decarboxylase from Lactobacillus 30a. J. Biol. Chem. 261: 15186– 15191 Vangtal A & Hammond EG (1986) Correlation of the flavor characteristics of Swiss-type cheeses with chemical parameters. J. Dairy Sci. 69: 2982–2993 Varmanen P, Rantanen T & Palva A (1996a) An operon from Lactobacillus helveticus composed of a proline iminopeptidase gene (pepI) and two genes coding for putative members of the ABC transporter family of proteins. Microbiol. 142: 3459–3468 Varmanen P, Rantanen T, Palva A & Tynkkynen S (1998) Cloning and characterization of a prolinase gene (pepR) from Lactobacillus rhamnosus. Appl. Environ. Microbiol. 64: 1831–1836 Varmanen P, Steele J & Palva A (1996b) Characterization of a prolinase gene and its product and an adjacent ABC transporter gene from Lactobacillus helveticus. Microbiol. 142: 809–816 Varmanen P, Vesanto E, Steele JL & Palva A (1994) Characterization and expression of the pepN gene encoding a general aminopeptidase from Lactobacillus helveticus. FEMS Microbiol. Lett. 124: 315–320 Vesanto E, Peltoniemi K, Purtsi T, Steele JL & Palva A (1996) Molecular characterization, over-expression and purification of a novel dipeptidase from Lactobacillus helveticus. Appl. Microbiol. Biotech. 45: 638–645 Vesanto E, Savijoki K, Rantanen T, Steele JL & Palva A (1995) An X-prolyl dipeptidyl aminopeptidase (pepX) gene from Lactobacillus helveticus. Microbiol. 141: 3067–3075 Vesanto E, Varmanen P, Steele JL & Palva A (1994) Characterization and expression of the Lactobacillus helveticus pepC gene encoding a general aminopeptidase. Eur. J. Biochem. 224: 991–997 Voight MN & Eitenmiller RR (1978) Role of histidine and tyrosine decarboxylases and mono- and diamine oxidases in amine buildup in cheese. J. Food Prot. 41: 182–186 Vongerichten KF, Klein JR, Matern H & Plapp R (1994) Cloning and nucleotide sequence analysis of pepV, a carnosinase gene from Lactobacillus delbrueckii subsp. lactis DSM 7290, and partial characterization of the enzyme. Microbiol. 140: 2591–2600 Wijesundera KM & Urbach G (1993) Flavor of Cheddar cheese. Final report to the Dairy Research and Development Corporation, Project CSt66 Wilkins DW, Schmidt RH & Kennedy LB (1986) Threonine aldolase activity in yogurt bacteria as determined by headspace gas chromatography. J. Agric. Food Chem. 34: 150–152

Wohlrab Y & Bockelmann W (1992) Purification and characterization of a dipeptidase from Lactobacillus delbrueckii subsp. bulgaricus. Int. Dairy. J. 2: 345–361 Wohlrab Y & Bockelmann W (1993) Purification and characterization of a second aminopeptidase (PepC-like) from Lactobacillus delbrueckii subsp. bulgaricus B14. Int. Dairy. J. 3: 685–701 Wohlrab Y & Bockelmann W (1994) Purification and characterization of a new aminopeptidase from Lactobacillus delbrueckii subsp. bulgaricus B14. Int. Dairy. J. 4: 409–427 Yan TR, Azuma N, Kaminogawa S & Yamauchi K (1987a) Purification and characterization of a novel metalloendopeptidase from Streptococcus cremoris H61: A metalloendopeptidase that recognizes the size of its substrate. Eur. J. Biochem. 163: 259–266 Yan TR, Azuma N, Kaminogawa S & Yamauchi K (1987b) Purification and characterization of a substrate-size-recognizing metalloendopeptidase from Streptococcus cremoris H61. Appl. Environ. Microbiol. 53: 2296–2302 Yan TR, Lin MZ, Lin MJ & Sun BJ (1991) Purification and characterization of an X-prolyl-dipeptidyl aminopeptidase from Streptococcus cremoris nTR. J. Chin. Biochem. Soc. 20: 21–32 Yang TS & Min DB (1993) Dynamic headspace analysis of volatile compounds of Cheddar and Swiss cheeses during ripening. In: Charalambous (ED) Food Flavors, Ingredients and Composition, (pp 157-174), Elsevier Science Publishers BV, Amsterdam Yen C, Green L & Miller CG (1980a) Degradation of intracellular protein in Salmonella typhimurium peptidase mutants. J. Mol. Biol. 143: 21–33 Yen C, Green L & Miller CG (1980b) Peptide accumulation during growth of peptidase deficient mutants. J. Mol. Biol. 143: 35–48 Yokoyama MT & Carlson JR (1981) Production of skatole and paracresol by a rumen Lactobacillus sp. Appl. Environ. Microbiol. 41: 71–76 Yoshpe-Besançon I, Gripon J-C & Ribadeau-Dumas B (1994) XaaPro-dipeptidyl-aminopeptidase from Lactococcus lactis catalyses kinetically controlled synthesis of peptide bonds involving proline. Biotech. Appl. Biochem. 20: 131–140 Yüksel GÜ & Steele JL (1996) DNA sequence analysis, expression, distribution, and physiological role of the Xaa-prolyldipeptidyl aminopeptidase gene from Lactobacillus helveticus CNRZ32. Appl. Microbiol. Biotech. 44: 766–773 Yüksel GÜ & Steele JL (1997a) Direct sequence submission to GenBank for a prolidase (pepQ) from Lactobacillus helveticus CNRZ32 (Accession AF012084). Unpublished Yüksel GÜ & Steele JL (1997b) Direct sequence submission to GenBank for pepV from Lactobacillus helveticus CNRZ32 (Accession AF012085). Unpublished Yüksel GÜ & Steele JL (1999) Proline-specific peptidases of Lactobacillus helveticus CNRZ32: Cloning and DNA sequence analysis of pepQ, a prolidase gene, and characterization of peptidase-deficient mutants. Unpublished results Yvon M, Thirouin S, Rijnen L, Fromentier D & Gripon JC (1997) An aminotransferase from Lactococcus lactis initiates conversion of amino acids to cheese flavor compounds. Appl. Environ. Microbiol. 63: 414–419 Zevaco C, Monnet V & Gripon JC (1990) Intracellular X-prolyl dipeptidyl peptidase from Lactococcus lactis ssp. lactis: Purification and properties. J. Appl. Bacteriol. 68: 357–366 Zoon P & Allersma D (1996) Eye and crack formation in cheese by carbon dioxide from decarboxylation of glutamic acid. Neth. Milk Dairy J. 50: 309–318 Zúñiga M, Champomier-Verges M, Zagorec M & Pérez-Martínez G (1998) Structural and functional analysis of the gene cluster encoding the enzymes of the arginine deiminase pathway of Lactobacillus sake. J. Bacteriol. 180: 4154–4159