Characterization of the Interaction Between

Home , Chymotrypsin, Lactobacillus helveticus, Lactocepin, Lactococcus lactis, Propionibacterium, Trypsin

CHARACTERIZATION OF THE INTERACTION

BETWEEN LACTOBACILLUS HELVETICUS

AND PROPIONIBACTERIUM IN SWISS CHEESE

DISSERTATION

Presented in Partial Fulfillment of the Requirement for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Patcharee Limpisathian, M.A.

* * * * *

The Ohio State University 2005

Dissertation Committee:

Professor W. James Harper, Advisor Approved by

Assoc. Professor Polly D. Courtney ______

Professor David B. Min Advisor

Professor Ahmed Yousef Food Science and Nutrition Graduate Program

Abstract

ABSTRACT

Swiss cheese makers report that some combinations of Lactobacillus helveticus and

Propionibacterium freudenreichii cultures result in poor eye and flavor formation, whereas other combinations perform well. The objectives were to develop a rapid method to predict successful strain pairings and to evaluate peptide metabolism by P. freudenreichii.

A cheese model system was developed to help predict the compatibility of L. helveticus and P. freudenreichii culture pairings. Cheese curds were prepared aseptically from

UHT milk using S. thermophilus alone (control) or S. thermophilus plus one of four L. helveticus strains. Curds were homogenized into a slurry with a final pH of 5.3, 60% moisture and 1% salt in moisture. The slurries were centrifuged, and the resulting serum was sterile filtered.

Each sterile slurry serum was added in the ratio 1:1 to 2× chemically defined medium (CDM) lacking amino acids. One of five P. freudenreichii strains was inoculated into CDM containing each slurry serum, and growth was monitored spectrophotometrically. Maximum specific

growth rate (µmax) and time to reach µmax were calculated using the Richards model.

Propionibacterium freudenreichii growth was not observed in CDM lacking slurry serum. CDM supplemented with slurry sera, including the control serum prepared without L. helveticus, supported the growth of four of the five P. freudenreichii strains to different extents.

Propionibacterium freudenreichii P764M1 did not grow with any slurry serum. L. helveticus L350

ii slurry serum stimulated the µmax and reduced time to reach µmax of four P. freudenreichii strains beyond that observed with the control serum. In contrast, L. helveticus L346 slurry serum

increased the time to reach µmax of these four strains in comparison to the control serum;

however, this serum also increased µmax of P859 and P873 despite in the delay in growth onset. L887 slurry serum delayed P. freudenreichii ATCC9614 growth, but had no effect compared to the control on any other P. freudenreichii strain. L856 slurry serum also slightly

delayed ATCC9614 growth compared with the control serum, but increased the µmax of

P859.

Slurry sera produced with different L. helveticus strains differed in their peptide profiles. A total of 126 peptide peaks were identified in all slurry sera. Twenty of these were not found in the control serum, but were found in one or more of the sera prepared with L. helveticus. There were significant differences (P<0.05) in peptide content between the control slurry and those prepared with L. helveticus strains. The peptide contents of slurries prepared with L. helveticus strains L350 and L 856 produce more peptides that were significant different from the control than L. helveticus strains L346 and L887. Most peaks (97 of 126) were found both in the control serum and in the sera prepared with L. helveticus strains, but the concentrations of the peptides decreased or increased in some cases when L. helveticus was included. Approximately one third of the peaks decreased in concentration in the presence of L. helveticus. Differences in peptide profiles may contribute to the observed differences in P. freudenreichii growth in the slurry. The slurry serum model has potential for screening P. freudenreichii and L. helveticus pairings for Swiss cheese manufacturing.

Propionibacterium strains have intracellular peptidases to breakdown peptides inside the cell, but little information exists concerning peptide transport in propionibacteria. It is iii unclear what type(s) of transporters are used to import the peptides and what their specificities are. The goal of this study is to evaluate specific peptide transport of ten P. freudenreichii strains by using commercially available peptides. The peptide concentration before and after incubation of each peptide with each strain was determined by high performance liquid chromatography to reveal the amount of peptide transported. All ten P. freudenreichii strains were able to transport the Ala-Ala di-peptide 30.89 to 100% peptide transport among the strains. Five strains were the highest and were not significant different

(p<0.01) among them. Percent peptide transport of Ala-Ala-Ala ranged from 0 to 100.

Three strains had no or very low transport of the Ala-Ala-Ala tri-peptide, but transported

Ala-Ala di-peptide well. Percent peptide transport of Glu-Ala ranged from 0 to 100.

Percent peptide transport of Glu-Glu ranged from 0 to 100. Only three strains transported the Glu-Glu di-peptide. Most Propionibacterium strains transported the Ala-Ala di-peptide better than other peptides, and two of the ten strains transported all four peptides tested.

There are significant differences in peptide transport among P. freudenreichii strains (P<0.01).

Some peptides are more readily transported and degraded by propionibacteria than others, thus the peptide composition of cheese may impact Propionibacterium growth.

iv as substrates: pro-ρNA, lys-ρNA, ala-pro-ρNA, and glu-ρNA. Using each chromagenic substrate, the liberation of ρ-nitroanilide was monitored spectrophotometrically at 410 nm in each cell free extract in 20 mM Tris-HCl buffer, pH 7.0, at 30°C. The Bio-Rad protein assay established the protein concentration in each cell-free extract which was used to calculate the enzyme specific activity. Significant difference in enzyme activity (P<0.01) was found among the Propionibacterium strains. Proline iminopeptidase (PepI) activity on the pro-ρNA substrate ranged from 0 to158 µmol/min/mg protein. Six strains had no or very low PepI activity. Aminopeptidase activity on the lys-ρNA and the glu-ρNA substrates ranged from 0 to 79.7 and from 0 to 3.9 µmol/min/mg protein, respectively. X-prolyl dipeptidyl aminopeptidase (PepX) activity on the ala-pro-ρNA substrate ranged from 0 to73.3

µmol/min/mg protein. Most strains, 27 of 34 strains, were higher in PepI activity than other activities. One strain with no PepI activity had the highest PepX and aminopeptidase activities. The intracellular peptidase activities of dairy Propionibacterium strains varied in specificity and activity levels. These differences may contribute to flavor and growth rate variations observed in cheeses manufactured with different Propionibacterium strains.

Different peptide transport and peptidase profiles in Propionibacterium will affect their performance with different L. helveticus strains. Peptides produced by L. helveticus and peptide transport and peptidase activity of P. freudenreichii are all strain-specific characteristics. These strain-to-strain differences may contribute to the growth rate and flavor variation of P. freudenreichii strains when paired with particular L. helveticus strains. This study describes several factors that may contribute to interactions between Swiss cheese starter cultures causing some culture combinations to perform better than others.

v Dedication

Dedicated to my family members and my mother.

vi Acknowledgement

ACKNOWLEDGMENT

I greatly thank my advisor, Assoc. Professor Polly D. Courtney for her intellectual advice, support, encouragement, and enthusiasm, which made this dissertation complete, and for her patience in correcting my English, stylistic, and also the scientific errors.

I am very thankful the committee, Professor W. James Harper, Professor David Min, and Professor Armed Yousef for their encouragement and support.

I acknowledge financial support from the Swiss Cheese Consortium, the Ohio

Agricultural Research and Development Center, and the Royal Thai government.

I wish to thank all of my colleagues for their friendships, understanding, and supports.

vii Vita

VITA

May 15, 1958………………………………. Born – Songkla, Thailand

1981……………………………………….. B.S. (Agro Industry) Prince of Songkla University, Songkla, Thailand

1981-1983…………………………………. Quality control, Chaopraya and Akenos Frozen Seafood Co., Ltd., Bangkok, Thailand

1990 - 1992……………………………….. Graduate Research Associate, School of Home Economics and Family Ecology, The University of Akron, Akron, Ohio, USA

1992……………………………………….. M.A. Food Science, The University of Akron, Akron, Ohio, USA

1984 - present……………………………… Food Scientist, the Royal Thai Government Department of Agriculture, Bangkok, Thailand

2001 - present……………………….……… PhD. Candidate, The Ohio State University, Columbus, Ohio, USA

viii PUBLICATIONS

Research Publication

1. Limpisathian, P, Courtney, P.D., and Harper, W.J. 2005. Peptide utilization by

Propionibacterium. Swiss Cheese Consortium Research Update, 5, p 16-17.

2. Limpisathian, P, Courtney, P.D., and Harper, W.J. 2005. Effect of lactobacilli cheese curd serum on growth of various Propionibacterium strains. Swiss Cheese Consortium

Research Update, 5, p 17-18.

3. Limpisathian, P. and Courtney, P.D. 2005. Peptidase specificity and activity of thirty- four dairy Propionibacterium strains. Abstract 71A-30. Institute of food technologists annual meeting. New Orlean. LU.

4. Limpisathian, P, Harper, W.J., and Courtney, P.D. 2005. Propionibacterium freudenreichii growth is differentially affected by the serum of Swiss cheese slurries prepared with different

Lactobacillus helveticus strains. Abstract W55The Joint ADSA-ASAS-CSAS Meeting. Cincinati.

OH.

5. Limpisathian, P., and Courtney, P. 2003. Effect of Lactobacillus strain on

Propionibacterrium growth. Swiss Cheese Consortium Research Update, 4, p 16-17.

6. Limpisathian, P., and Courtney, P. 2002. Effect of Lactobacillus strain on

Propionibacterrium growth. Swiss Cheese Consortium Research Update, 3, p 8-9.

FIELD OF STUDY

Major Field: Food Science and Nutrition

TABLE OF CONTENTS

Page

Abstract...... ii

Dedication ...... vi

Acknowledgement...... vii

Vita…...... viii

List of Tables ...... xiv

List of Figures...... xv

1. LITERATURE REVIEW...... 1

1.1 Swiss Cheese Making ...... 1

1.2 Proteolysis of Casein...... 2

1.3 Cheese Model Systems...... 4

1.4 Autolysis of Dairy Starter cultures...... 5

1.5 Bacterial Proteinase Enzymes...... 6

1.6 Bacterial Peptidase Enzymes...... 8

1.7 Peptide Transport...... 15

1.8 Proteolytic Activity of L. helveticus...... 16

1.9 Peptide and Amino Acid Utilization by Propionibacteria...... 17

x 1.10 Effect of Lactobacillus on Propionibacterium Growth ...... 20

1.11 Figures...... 23

1.12 References...... 27

2. GROWTH OF PROPIONIBACTERIUM FREUDENREICHII IN CHEESE

SLURRY SERUM PRODUCED WITH DIFFERENT LACTOBACILLUS

HELVETICUS STRAINS ...... 31

2.1 Abstract...... 31

2.2 Introduction ...... 33

2.3 Materials and Methods...... 36

2.3.1 Swiss cheese slurry serum preparation with different Lactobacillus helveticus

strains...... 36

2.3.2 Propionibacterium freudenreichii strains and growth conditions ...... 38

2.3.3 Evaluation of peptide profiles in slurry sera by HPLC...... 40

2.4 Results...... 41

2.4.1 Propionibacteium growth in CDM + slurry serum ...... 41

2.4.2 Comparison of peptides generated by four different L. helveticus strains...... 42

2.5 Discussion...... 43

2.5.1 Propionibacteium growth in CDM+ slurry serum...... 44

2.5.2 Comparison of slurry model used in this study to others utilized...... 44

2.5.3 Comparison of peptides generated by four different L. helveticus strains...... 45

2.6 Tables…...... 47

2.7 Figures...... 53

2.8 References...... 63

xi 3. PEPTIDE TRANSPORT IN TEN PROPIONIBACTERIUM FREUDENREICHII

STRAINS ...... 66

3.1 Abstract...... 66

3.2 Introduction ...... 67

3.3 Materials and Method ...... 68

3.3.1 Chemicals...... 68

3.3.2 HPLC Analysis...... 69

3.3.3 Bacterial strains and transport assay ...... 70

3.3.4 Statistical Analysis...... 71

3.4 Results and Discussion ...... 71

3.5 Tables…...... 76

3.6 Figures...... 78

3.7 References...... 83

4. PEPTIDASE ACTIVITIES OF THIRTY-FOUR DAIRY PROPIONIBACTERIUM

FREUDENREICHII STRAINS...... 84

4.1 Abstract...... 84

4.2 Introduction ...... 85

4.3 Materials and Methods...... 87

4.4 Results ...... 89

4.4.1 Proline iminopeptidase activity of Propionibacterium strains...... 89

4.4.2 X-prolyl dipeptidyl aminopeptidase activity of Propionibacterium strains ...... 90

4.4.3 Aminopeptidase activity of Propionibacterium strains...... 90

4.5 Discussion...... 91

xii 4.6 Table…...... 95

4.7 Figures...... 97

4.8 References...... 101

5. CONCLUSION ...... 103

List of References...... 106

APPENDIX...... 112

xiii

LIST OF TABLES

List of Tables

Table Page

2.1 Maximum specific growth rates (d-1) of five Propionibacterium strains in chemically defined medium (0.5% NaCl) with added slurry serum prepared with one of four Lactobacillus helveticus strains…………………………………………………………47

2.2 Lag phase (days) of five Propionibacterium strains in chemically defined medium (0.5% NaCl) with added slurry serum prepared with one of four Lactobacillus helveticus strains……………………………………………………………………………...48

2.3 The peptide content for each slurry prepared with the different L. helveticus strains and the control S. thermophilus…………………………………………………………...49

3.1 Preparation of peptides standard 0.3, 0.25, 0.2, 0.15, 0.1, 0.05, 0.01, 0.005, 0.001 mM with 3 mM and 0.3 mM peptide solution…………………………………………...76

3.2 Percent of total peptide transported by ten P. freudenreichii strains from a 3 mM solution …………………………………………………………………………………...... 77

4.1 Peptidase specificity activity (µmol/min/mg protein) of 34 dairy Propionibacterium strains………………………………………………………………………..…....95

xiv

LIST OF FIGURES

List of Figures

Figure Page

1.1 The primary structure of αS1–casein………………………………………………... 23

1.2 The primary structure of αS2–casein………………………………………………...24

1.3 The primary structure of β–casein…………………………………………………. 25

1.4 The primary structure of κ–casein………………………………………………….26

2.1 Cheese-like model system (slurry serum) for evaluating L. helveticus and P. freudenreichii strain pairings……………………………………………………………………....41

2.2 Growth of P. freudenreichii strain P859 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains …………………………….53

2.3 Growth of P. freudenreichii strain P873 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains ………………………….....54

2.4 Growth of P. freudenreichii strain P812 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains …………………………….55

2.5 Growth of P. freudenreichii strain P764M1 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains ……..………….……...56

2.6 Growth of P. freudenreichii strain ATCC9614 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains ………………....57

2.7 The peptide profile of the slurry serum prepared with L. helveticus strain L350………58

2.8 The peptide profile of the slurry serum prepared with L. helveticus strain L346………59

2.9 The peptide profile of the slurry serum prepared with L. helveticus strain L856………60 xv

2.10 The peptide profile of the slurry serum prepared with L. helveticus strain L887………61

2.11 The peptide profile of the control slurry serum lacking L. helveticus………………….62

3.1 Percent peptide transport of ala-ala of ten Propionibacterium freudenreichii strains……...78

3.2 Percent peptide transport of ala-ala-ala of ten Propionibacterium freudenreichii strains...... 79

3.3 Percent peptide transport of glu-ala of ten Propionibacterium freudenreichii strains……...80

3.4 Percent peptide transport of glu-glu of ten Propionibacterium freudenreichii strains….…..81

3.5 Percent peptide transport of ala-ala, ala-ala-ala, glu-ala, and glu-glu of ten Propionibacterium freudenreichii strains……………………………………………….....82

4.1 Proline iminopeptidase (PepI) specific activity (µmol/min/mg protein) of thirty-four dairy Propionibacterium strains………………………………………………………..97

4.2 X-prolyldipeptidyl aminopeptidase (PepX) specific activity (µmol/min/mg protein) of thirty-four dairy Propionibacterium strains………………………………………….98

4.3 Aminopeptidase specific activity (µmol/min/mg protein) on lys-ρNA of thirty-four dairy Propionibacterium strains…………………………………………….99

4.4 Aminopeptidase specific activity (µmol/min/mg protein) on glu-ρNA of thirty-four dairy Propionibacterium strains…………………………………………...100

A.1 The growth P. freudenreichii strain P859 in slurry sera prepared with different L. helveticus strains (pH 5.3)…………………………………………………………………...125

A.2 The growth of P. freudenreichii strain P873 in slurry sera prepared with different L. helveticus strains (pH 5.3)…………………………………………………………..126

A.3 The growth of P. freudenreichii strain P812 in slurry sera prepared with different L. helveticus strains (pH 5.3).…………………………………………………………..127

A.4 The growth P. freudenreichii strain P764M1 in slurry sera prepared with different L. helveticus strains (pH 5.3)…………………………………………………………..128

A.5 The growth P. freudenreichii strain ATCC9614 in slurry sera prepared with different L. helveticus strains (pH 5.3)…………………………………………………………...129

A.6 The different strains of P. freudenreichii growth in Propionibacterium media within 96 hours ……………………………………………………………………………..130

xvi A.7 The growth P. freudenreichii strain P859 in slurry sera prepared with different L. helveticus strains within 96 hours……………………………………………………………131

A.8 The growth P. freudenreichii strain P873 in slurry sera prepared with different L. helveticus strains within 96 hours …………………………………………………………..132

A.9 The growth P. freudenreichii strain P812 in slurry sera prepared with different L. helveticus strains within 96 hours …………………………………………………………..133

A.10 The growth P. freudenreichii strain P764M1 in slurry sera prepared with different L. helveticus strains within 96 hours ………………………………………………….134

A.11 The growth P. freudenreichii strain ATCC9614 in slurry sera prepared with different L. helveticus strains within 96 hours ………………………………………………….135

A.12 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L856……………………………………………………………………….136

A.13 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L887………………………………………………………………………..137

A.14 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L887………………………………………………………………………..138

A.15 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L346M1……………………………………………………………………139

A.16 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L346M2……………………………………………………………………140

A.17 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L367……………………………………………………………………….141

A.18 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L350……………………………………………………………………….142

A.19 The growth of twelve P. freudenreichii in the slurry serum prepared with L. helveticus strain L774M1……………………………………………………………………143

xvii 1. LITERATURE REVIEW

CHAPTER 1

LITERATURE REVIEW

1.1 Swiss Cheese Making

The initial steps in Swiss cheese making involve coagulation of the casein proteins in the milk using rennet, a proteolytic enzyme, followed by cutting and cooking the coagulated curd to remove whey. During these steps, the Streptococcus thermophilus and Lactobacillus helveticus starter cultures metabolize the lactose in the milk to lactic acid. After cooking, the curd particles are pressed together, and then the cheese block is soaked in a brine solution to add salt. The cheese is then cooled at ~8°C for 4-14 days to allow the salt to equilibrate throughout the cheese. The S. thermophilus and L. helveticus cultures are inhibited by the salt and cool temperature and the population will begin to slowly lyse and die. Though the cells are dying, the enzymes present in the bacteria will be released as the cells lyse and continue to degrade cheese components during cheese ripening.

Swiss cheese must be ripened to achieve its flavor and appearance. After the salt equilibration step, the first phase of ripening, the “warm room” ripening begins. In the warm room, the cheese is held at 20 - 24°C for about three weeks. This temperature

1 encourages the growth of the Propionibacterium freudenreichii starter culture that produces

propionic and acetic acids for flavor and CO2 gas for eye formation. During this time, the proteolytic enzymes from L. helveticus degrade some of the casein proteins into peptides and amino acids. These contribute to flavor, but may also stimulate P. freudenreichii growth.

When the desired eye size has been achieved, the cheeses proceed to the second phase of ripening, “cold room” ripening, in which the cheeses are held at 4 - 10°C for 1 – 3 months.

Flavor continues to develop during this time as more proteolysis of the casein occurs along with other chemical reactions.

1.2 Proteolysis of Casein

The proteolysis of casein, including αS1, αS2, β, and κ caseins, occurs via the action of native milk proteinases, chymosin and other rennet proteases, and enzyme from the Swiss cheese culture bacteria. Plasmin is the most active milk enzyme in Swiss cheese and degrades caseins in the order of β-casein (figure 1.3), αS2-casein (figure 1.2), and αS1-casein (figure

1.1), but κ-casein (figure 1.4) is resistant to this proteinase. The αS1-casein protein consists of 199 amino acids, including 2 half cysteines, 11 phosphate groups, and 10 prolines. The C- terminal (f160-199) is mostly hydrophobic, and the N-terminal (f1-68) is mostly hydrophilic.

It is cleaved at lysyl-X bond by enzyme plasmin. The αS2-casein protein consists of 207 amino acids, is high in phosphate groups, and is sensitive to Ca++. It is precipitated by less

++ than 2 mM Ca at 37ºC. Plasmin hydrolyzed αs2-casein at Lys21-Gln22, Lys24-Asn25,

Lys152-Leu153, Lys188-Asp189, and Lys199-Val200 (Gagnaire and others 2001). The β- casein protein precipitates in the presence of 8 to 15 mM Ca++. It consists of 209 amino

2 acids, including 5 phosphate groups, and 53 prolines. Proline comprises 16% of of the amino acid residues. It is cleaved by enzyme plasmin at f(108-209), f(114-209), f(29-209),

and f(106-209). Additionally, plasmin releases γ1 -Cn [β-CN-IP (f29-209)], γ2 -Cn [β-CN

(f106-209)], and γ3 -Cn [β-CN (f108-209)] including proteose peptone component 5-[β-CN-

5P(f1-105/107)], proteose peptone component 8 slow-[β-CN-1P(f29-107)], and proteose peptone component 8 fast-[β-CN-4P(f1-28), (Farrell 2004; and Imafidon 2001).

Chymosin is the enzyme used to coagulate casein in cheese process, but it is mostly denatured during cooking at 55 ºC. Therefore, its contribution to proteolysis in a ripening

Swiss cheese is minimal. Imafidon (2001) has reviewed the activities of chymosin. It cleaves

αS1 at Phe23-Phe24 and Leu101–Lys102 resulting in small peptide release: f(1-23) including

αS1–I(f24/24-199), αS1-II(24/25-169), III, and IV (f24/25-149/150), Vf(29/33-199), VI(f56-

179), and VII/VIII (f56-179). αs2-Casein is more resistant to chymosin than αS1-casein.

Chymosin is active in the regions of residue 88 to 98 and 164 to 180 (Farrell 2004). From β-

Casein, it releases β-I (f1-199/192), β-II (f1-165/167), β-IIIa (f1-139), and III-b (f1-140). κ–

Casein consists of 169 amino acids, 2 half cystine, 1 phosphate group and 20 proline.

Hydrophobic region is N-terminal at f(1-100) and hydrophilic is C-terminal at f(105-169). It has no clusters of phosphorylated serine like αs1, αs2, and β-casein. It contains only 1 phosphate group. Chymosin cleaves at Phe 105-Met 106 to release a glycomacropeptide

(GMP). It is this action that causes the coagulation of the casein micelles in the early phases of cheese ripening. The GMP is glycosylated reaction and insensitive to Ca2+. κ-Casein can form disulfide linkage with two cysteine residues resulting in increase rennet clotting time and reduce curd strength. Para κ–casein does not hydrolyze in cheese.

3 Enzymes in lactic acid bacteria (LAB) can hydrolyze casein as well. Most species have a cell envelope-associated proteinases and numerous cytoplasmic peptidase enzymes. The proteinase degrades intermediate size peptides produced by the chymosin or plasmin. The peptidase enzymes play an important role in cheese proteolysis to produce free amino acids and small peptides that contribute to cheese flavor. Oligopeptidase and tripeptidase use various peptide substrates. Aminopeptidases, Pep N and Pep C, have also broad specific substrate peptides. Pep A is specific on Glu/Asp residues, and Pep L releases pyroglutamic from N-terminal of peptides. Pep X, Pep I, Pep P, Pep R and Pep Q release peptides containing proline. Pep X releases X-Pro dipeptides from N-terminal peptides, then Pep I

releases free proline. Pep P removes X-Pro-Pro-(X)n sequence, Pep R cleaves Pro-X dipeptide and Pep Q cleaves X-Pro dipeptides (Christensen and others 1999).

1.3 Cheese Model Systems

Cheese experiments are expensive and time consuming. Some studies utilize a model system to reduce time and cost of experimentation. Some researchers suggested that the model should be close to cheese properties and the biochemical reactions close to normal cheese during ripening (Shakeel-Ur-Rehman and others 2001). Some models were reported for microbiological studies of cheese ripening with a short period of time. They described a protocol of the miniature model under controlled microbiological conditions. The dry matter and the salt in moisture were not significant difference from old cheeses (Hynes 2000). Dully

(1976) conducted an experimental model by using 5% NaCl and a preservative (potassium sorbate) in a closed container at 30ºC. The cheese model developed flavor in a short period of

4 time. However, the cheese slurry models used for studies of the flavor development during cheese ripening were unclear and the slurries developed off flavors (Davies and Law 1984).

1.4 Autolysis of Dairy Starter cultures.

Peptides and amino acids produced by autolysis of Lactobacillus impact the growth of

Propionibacterium. Valence and others (1998) have studied autolysis of Lactobacillus. They found that Lactobacillus strain ITG LH77 lyses 99.9% (<100 cfu/g cheese) at the beginning of cold room ripening. Intracellular dipeptidase (Pep D) was released through autolysis after galactose depletion. He pointed out that the rate of autolysis was strain specific. After 22 days of ripening at pH 5.5 and 24°C, lactate was depleted during the middle of warm room treatment, resulting in induction of the autolysis condition. The authors concluded that the autolysis of Lactobacillus depended on cooking time and temperature, brining process, the amount of lactose and acidity of cheese during ripening.

Husson-Kao and others (1999) studied autolysis in S. thermophilus DN 001065 by detecting Pep X activity. The data were shown in large extent of autolysis at pH 5 and 2%

NaCl. The heat shock between 42°C to 50°C or 55°C and cold shock between 42°C to 4°C or 8°C increased the autolysis. In addition, a low level of lactose associated together with acidity could increase the autolysis to 75%. At 1% lactose level, the cells lysed between pH 5 to 7. A pH less than 5 and more than 7 could reduce autolysis. The study showed that Pep

X activity was constant at 42°C, pH 4.9 and increased 4.8 fold after cell autolysis. The minimal pH at which lysis could occur was strain specific. For the strain DN 001065, the lysis was inhibited at pH below 5. Deutsch and others’s study (2002) also found strain 5 differences in that the amount of free amino acids at the end of the ripening could vary 1.5 fold indicating strain variation in the intracellular peptidase activity released. In addition, the authors noted that the lysis of S. thermophilus depended on the lysis of Lactobacillus strains.

After the autolysis, lactate dehydrogenase was detected. Most of them were lysed at day 1 and two strains were found late lysis between days 20 and 36 with galactose depletion.

Lamee (1995) studied autolysis of P. freudenreichii strain CNRZ 725. He showed that the autolysis increased at higher temperature. An increase in temperature from 30°C to

43°C during exponential growth phase resulted in the autolysis occurring immediately. The depletion of lactate at pH 6.2 and 30°C caused the strain of Propionibacterium to induce autolysis whereas the production of propionic acid and acetic acid stopped. N- acetylglucosaminidase activity was found in this strain associates with cell wall neutral polysaccharides.

1.5 Bacterial Proteinase Enzymes

The proteolytic enzymes play an important role in Swiss cheese flavor and may also be involved in P. freudenreichii growth stimulation. Oberg (2002) studied proteinase activity of

two specific types; PI and PIII were found in L. helveticus strains. Substrates of PI and PIII were

β–CN and αs1-CN, β–CN, and κ–CN. They demonstrated that PI type proteinase cleaved at f1-16+f17-23 for L. helveticus strain 29; f1-9+f10-23 and f1-13+f14-23 for L. helveticus strain

10, 12, and 36; f1-16+f17-23 for L. helveticus strain 9 and 11; f1-8+f9-23 for L. helveticus strain 3, 37,and 41. The finding of Oberg agrees with Peterson and others (1999) that L.

helveticus expressed at least one group of proteinase. They studied αs1-CN(f1-23) with L. 6 helveticus CNRZ 32 which found two different surface proteinases. The first proteinase

produced fragment αs1-CN f1-16, f17-23, f1-17, and f18-23. The second proteinase

produced fragments αs1-CNf1-9, f10-23, f1-6, and f9-23.

Proteinase activity in L. helveticus CRL 974 (ATCC15009) and L. helveticus CRL 1062 was studied (Deusch and others 2000). The research showed that the proteinase activity was regulated by low molecular mass (less than 3000 Da) of casamino acids and the dipeptide leucyl proline. The addition of casamino acids caused a 25% decrease in proteinase activity.

The proteinase activity of L. helveticus CRL 1062 could hydrolyze α and β–casein, whereas there was much less activity with κ–casein. The peptides produced from α, β–casein were

20 to 27 KDa and 24, 15, and 6.5 KDa, respectively. The highest activity was found in exponential growth phase.. Additionally, this research suggested that proteinase (Prt H) activity was found to decrease in the rich-peptide medium and maintained when increased amino acids 10-fold in the medium.

Gagnaire and others (2001) reported proteinase activities of cathepsin D and the cell envelope proteinase (CEP) of LAB involved in producing peptides and amino acids. The

cathepsin D cleaved at Phe24-Val25, Phe32-Gly33 of αs1-casein and Phe52-Ala53, Pro81-

Val82 of β-casein. Peptides releases were αs1-CN (25-32), β-CN (32-52), β-CN (82-88), and

β-CN (11/12/13/15/16-28). Plasmin hydrolyzed αs2-casein at Lys21-Gln22, Lys24-Asn25,

Lys152-Leu153, Lys188-Asp189, and Lys199-Val200. Additionally, only dephosphorylated peptides; β-CN (12-28)(-1P), β-CN (13-25)(-1P), β-CN (15-25)(-1P) of phosphoserine residue that loss N-terminal amino acid residues to form β-CN(1-25) and β-CN (18-25) were found on Ser15, not including Ser cluster; Ser17-Ser18-Ser19. The author concluded that S.

7 thermophilus were possibly involved in some peptides that split bond at Gly10-Glu11, Pro-63-

Gly64, Asn73-Val74, Gln79-Thr80, Leu88-Gln89, Gly94-Val95, Met102-Ala103, Glu14-

Val15, Asn19-Leu20, Glu30-Val31, and Asn38-Glu39 from αs1-casein.

1.6 Bacterial Peptidase Enzymes

The peptides can be imported into the cell via specific peptide transporters and then degraded to amino acids by intracellular peptidase enzyme. Sousa and others (2001) have reviewed the studies of the proteolysis during cheese ripening. They stated that Lactobacillus and S. thermophilus contained proteolytic enzymes: the extracellular proteinase, intracellular oligoendopeptidases, glutamyl aminopeptidase, pyrolidone carboxylyl peptidase, leucyl aminopeptidase, X-prolyldipeptidyl aminopeptidase, proline iminopeptidase, aminopeptidase

P, prolinase, prolidase, general dipeptidase, and general tripeptidase. Propionibacterium contain amino peptidase, proline iminopeptidase, and X-prolyldipeptidyl amino peptidase. The proteinase hydrolyses oligopeptides to the small peptides whereas the peptidases contribute to the free amino acids (Sousa and others 2001).

Some studies have isolated and identified enzyme activities of Propionibacterium. Chaia and others (1990) has studied peptidase activity. They found that the peptidase were more active on proline-ρ-nitroanilide than on leucine-ρNA. Panon (1990) purified proline iminopeptidase from P. freudenreichii 13673 with molecular mass of 61 KDa and determined its activity at 40 °C, pH 8.0. They found that proline iminopeptidase cleaved dipeptide and peptide containing proline residue (Pro-X). The activity increased at pH 5.6 to 5.8 during

8 ripening at temperature between 10 to 24°C. The X-Pro peptide was resistance to enzymatic hydrolysis.

Fernandy-Espla and Fox (1996) studied the enzymes that detected in strain P. freudenreichii subsp. shermanii strain NCDO 853 were prolidase, iminopeptidase, and X-prolyl dipeptidyl aminopeptidase. The iminopeptidase from P. freudenreichii subsp. shermanii strain

NCDO 853 was proline specific. The iminopeptidase hydrolyzed Pro-ρNA, dipeptidases containing proline (Pro-Met, Pro-Leu, Pro-Gly). They found that the X-prolyl dipeptidyl aminopeptidase (Pep X) had high activity between pH 5.5 to 7.5 at 40°C. The enzyme activity was highest for Ala-Pro-ρNA and Gly-Pro-ρNA and unable to hydrolyze Leu-Pro or simple amino acids, or ρ-nitroanilides, Pro- ρNA, and Ala-ρNA.

Stepaniak and others (1998) reported “the oligopeptidase from Propionibacterium

ATCC 9614 did not hydrolyze pentapeptide, methionine enkephalin, and had very low activity on peptides containing 17 or more amino acid residues.” They reported the activity of oligopeptidase on brandykynin as a substrate. The enzyme hydrolyzed the N-terminal of

Arg-Pro. It was active at pH 6.7-7.5 and at 40-50°C and retained 20% of activity at 7°C and pH 5.5. They concluded that Propionibacterium oligopeptidase perhaps was active in cheese during ripening of cold temperature.

Quelen and others (1995) stated that all strains of Propionibacterium possessed amino peptidase activity. He studied X-prolyl-dipeptidyl, aminopeptidase, proline iminopeptidase, and proline endopeptidase. They found strong proline iminopeptidase activity in the intracellular fraction of P. freudenreichii subspecies shermanii by using pro-βNa as substrate.

X-pro DPAP, prolinase and prolidase activities were found in this study.

9 Langsrud and others (1995) have reviewed protein degradation by Propionibacterium.

They reported peptidase enzyme in different strains and species. Seven to 8 strain specific peptidase bands were found. Serine proteinase activity was associated with the cell wall and membrane. They pointed out that Propionibacterium contained proteinases and peptidases.

One was cell wall associated and another was intracellular enzyme. They agreed with Sousa and others (2001) that Propionibacterium contained peptidases, such as aminopeptidases, proline imidopeptidase, X-prolyl-dipeptidyl-amino-peptidase. They stated that X-prolyldipeptidyl-aminopeptidase was the main intercellular peptidase and claimed that

Propionibacterium contains endopeptidase, proline imidopeptidase and praline imidopeptidase.

They concluded that aspartic acid, alanine, serine, and glycine were degraded by

Propionibacterium, and that the utilization was strain specific.

On the other hand, Valence and others (2000) reported that proteolysis increased mainly in the warm room (22-25°C) by Lactobacillus. Thus, the peptides were hydrolyzed during this ripening time. The peptide hydrolysis of L. helveticus strain ITGLH1 was greater than ITGLH77 on the basis of free amino group and free amino acids. The amino groups remained constant during cold storage. The amino acids found were mainly Proline,

Leucine, Lysine, Valine, and Glycine.

Langsrud and others (1995) emphasized that lactobacilli have more proteolytic activity than Propionibacterium. They found that proline iminopeptidase activity of Lactobacillus higher than Lactococcus. Peptidase activity has been studied in L. helveticus CNRZ 32. The research found that the highest activity with Gly-Pro-ρNA was for strains carrying PepX and the highest aminopeptidase activity for Lys-ρNA was with strains carrying pepN. For aminopeptidase activities, Lys, Met, Ala, and Phe increased 30-40% in some strains; JLS 243

10 (Pep X) and JLS 245 (Pep C and Pep X). The research showed that Pep C activity expressed at low level compared to Pep N.

Deutsch and others (2000) studied peptidase activity by using β-casein. They found that phosphorylated residues (Phe33-Lys48) were resistant to all peptidase. They also suggested that the phosphorylated residues are possibly involved with bacteria phosphatase.

The carboxypeptidase in L. helveticus and S. thermophilus was found to be involved in removal of C-terminal amino acids. Free proline was released by prolidase (Pep Q) specific for X- pro dipeptidase in L. helveticus, but not found in S. thermophilus resulting from the absence of

Pep IP and a weak Pep Q activity. The Pep IP found in L. helveticus and L. delbrueckii removed proline in N-terminal position and it also was active on tetrapeptides. The high activity of Pep IP produced di- and tripeptides; (Ile74-Pro76) and (Leu150-Pro152). Short peptides were found in samples that undigested in β-casein hydrolysate. The peptides containing less than 12 amino acids from N and C-terminal regions were completely hydrolyzed for all species. Interestingly, research revealed that hydrolysis of peptides located in central region of the β-casein sequence (Gly94-Lys97), (Tyr114-Pro119), and (Thr126-

Leu133) depended on the strain used. The long peptides (more than 12 amino acids) containing proline sequence; (Pro75-Pro76), (Pro85-Pro86), (Pro152-Pro153), and (Pro158-

Pro159) were hydrolyzed except (Pro136-Pro137) which hydrolyzed by prolyl aminopeptidase including (Thr126-Leu39), (Thr126-Trp143), (His134-Leu139), and also active on di and tripeptides. Thus, endopeptidase and aminopeptidase involved in forming long peptide sequence. The researchers also reported that carboxypeptidase involved with a single amino acid in C-terminal; (Gly64-Pro69), (Gly64-Asn68), and (Gly64-Ser67) in L. helveticus and (Asp129-Glu131) in S. thermophilus. Juillard and others (1998) studied peptide

11 utilization by Lactococcus lactis strain MG1363. Peptides were separated into acidic, neutral and basic. They reported that Lactococcus lactis utilized hydrophobic basic peptide (600-1100

Da.). The large peptide and acidic peptide were not utilized. No significance cell lysis was found. They concluded that the growth was inhibited when Leu and Met were lacking. In addition, they pointed out that the free amino acid themselves have the synergism for the growth.

Peptidase activities have been studied in L helveticus. Mostly, Pep N, Pep C and Pep

X play important role in producing peptides. Pep N hydrolyzed AA-ρNA substrates and with high activity for basic amino acids (Lys, Arg) and hydrophobic uncharged (Leu, Ala).

The significant bond is Met-Phe ρNA. Pep N has low activity on Asp-, Glu-, and Gly-ρNA

(Christensen and others 1999). PepC is reported to have significant activity on AA-βNAP substrates and is highly active on basic amino acids (Arg, His, Lys) along with Gly, Asp, Ala,

Leu and Phe. As with PepN, AA-ρNA is substrate for Pep C, including Met- and Gly-ρNA.

PepC has been reported to be inactive on Pro-ρNA, Pro-βNAP, Xaa-Pro-ρNA and Xaa-

Pro-βNap. Hydrolysis of hydrophobic P1 Gly-Xaa, P2- Gly-Gly-Xaa and di, tri, and tetra peptides was observed (Christensen and others 1999). The highest activity was Gly2-5 substrates including tetrapeptides (Gly-Gly-Phe-Leu, and Gly-Gly-Gly-Ala). PepX is another important peptidase in L. helveticus. PepX hydrolyzed Xaa-Pro dipeptides from the

N-terminal with 3-7 amino acids residues. PepX has high activity on Xaa- Pro dipeptides when the N-terminal residues are uncharged hydrophobic (Ala Ile, Val, Gly), basic amino acids (Arg, His, Lys), or aromatic amino acids residues (Phe, Tyr). Pro-Pro-Xaa was found to be a substrate of PepX. PepX activity is strain specific, and activity is higher in milk than in media (William 1998). 12 Christensen and others (1999) reported that PepD activity was similar to PepV.

PepD is limited in liberation of amino acid from milk peptides. As PepD, PepV hydrolyzed basic amino acid residues (Arg, His, Lys) and hydrophobic uncharged (Ala, Ile, Leu, Val) including aromatic amino acid residues (Phe, Tyr) and Met in the N-terminal position. PepV has activity on β-Ala-Xaa and Xaa-β-Ala. Proline iminopeptidase hydrolyzes specific N- terminal proline and is active on sequence of prolyl oligopeptides in region Ser-107, Asp-

246, and His-273 in L. delbrueckii subsp. bulgaricus. Pro-Xaa was a substrate for C-terminal hydrophobic uncharged residue (Ala, Gly, Ile, Leu, Val), acidic residues (Glu) and aromatic residues (Phe, Try). Pep V also liberates proline from N-terminal tripeptides, but it was reported not to be active on tetra/or pentapeptides. Pro-Pro and Pro-pNA are substrate for

Pep I.

William and others (1998) reported serine prolyl aminopeptidase for a specific substrate for some strains. Gly-Pro 7-amido-4-methyl-coumarin as substrate has been reported as an inhibitor of serine dipeptidyl peptidase and metalloprotease as an inhibitor for dipeptidyl peptidase. Leucyl aminopeptidase activity was higher than prolyl (proline- iminopeptidase), proline (aminopeptidase P) and aspartyl (glutamyl) aminopeptidase like enzyme. High leucyl aminopeptidase activity is associated with high content of smaller peptides and amino acids (Frohlich-Wyder 2002).

PepQ (prolidase) hydrolyzed specific Xaa-Pro dipeptides at N-terminal residues of hydrophobic uncharged (Ala, Ile, Leu, Val), basic residues (His), and aromatic (Phe, Try) including surface containing residue (Met). This enzyme hydrolyzed Pro-Ala, Pro-Pro, and

Pro-Val substrates to some degree. PepQ activity was 99% of Met-Pro, Leu-Pro, and Phe-

Pro (William and others 1998; and Christensen and others 1999).

13 PepR has high activity on Pro-Xaa dipeptides and limit activity for larger peptides and Pro-(ρNA, βNAP) substrates. PepR hydrolyzed C-terminal residues that are hydrophobic/uncharged (-Ala, -Ile, -Leu, -Val), aromatic (-Phe), and sulfur containing (-

Met). PepR also hydrolyzed Xaa-Pro dipeptides and significant activity has been shown on

Pro-Met, Thr-Leu, Pro-Phe, Gly-Leu and Met-Ala (William and others 1998; and

Christensen and others 1999).

PepE is an endopeptidase enzyme which is active on N-terminal and C-terminal residues. PepE hydrolyzed Met-enkephalin (Try-Gly-Gly-Phe-Met) at the Gly3-Phe4 and brankinin (Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg) at Gly4-Phe5 position.

PepO has been reported to hydrolyze oligopeptides from 3 to 35 residues (Met and

Leu-enkephalin and α-s1-casein f(165-169). PepO has activity on large peptides i.e. bradykinin, angiotansin, neurotansin, and isulin-β-chain (William and others 1998; and

Christensen and others 1999).

Deutsch and others (2000) studied peptidase activity of L. helveticus, Lactococcus lactis and

S. thermophilus on hydrolysis of β-casein. The procedure was carried out by using lyophilized β- casein hydrolyzate as a substrate. Trypsin and chymotrypsin were used in this study. After three hours of incubation, the β-casein was converted to peptides. Then, heating at 80°C for

20 minutes inactivated enzyme. The peptides were separated by reverse phase high-pressure liquid chromatography. The masses of the peptides were determined by HPLC electro spray ionization (ESI) MS analysis. The results contained 34 amino acids and three phosphorylated peptides; (Arg1-Arg25), (Phe33-Lys48), and (Phe33-Phe52). The bonds were hydrolyzed at

(Leu125-Thr126) and (Met144-His145). The peptidase activity rate depended on species and strain. Lactobacillus helveticus showed the most peptidase activity (0.12 mM/h) compared to S. 14 thermophilus strain ITG LL51, L. delbrueckii subsp. lactis strain ITG LL14, and L. delbrueckii subsp. bulgaricus strain CNRZ397.

1.7 Peptide Transport

Peptides can be imported into the cell via specific peptide transporters. To our knowledge, no reports exist on peptide transport in Propionibacterium. It is unclear what type(s) of transporters are used to import the peptides and what their specificities are.

Most of the research has pointed out the peptide transport of Lactococcus lactis species. The peptide transport system supports essential amino acids for their growth in milk. The oligopeptide transport system, and two di-tripeptide transport system; a proton motive force-driven di-tripeptide carrier (DtpT) and an ATP-driven di-tripeptide transporter (DtpP) has been reported for Lactococcus lactis. Helinck and others (2003) investigated the competition of casein-derived peptide to be transported by Lactococcus lactis. They found that casein-derived peptides competitively inhibited the transport of a reporter peptide by the oligopeptide transport system. The authors demonstrated that the presence of a negatively charged residue decreased the competition of the peptides. In contrast, the presence of uncharged (Ser, Thr, and Tyr) or a hydrophobic N-terminal residue (Met), or positive charged residue (Arg and Lys) did not effect the competition. The researchers concluded that the more increase in proteolytic activity the more peptide release from the casein resulted in the impairment of the oligopeptide transport function and then inhibited the growth rate. Garault and others (2002) studied the oligopeptide transport of S. thermophilus.

They indicated that S. thermophilus transported oligopeptides containing from 3 to 23 amino

15 acids with preference for hydrophobic oligopeptides. Charbonnel and others (2003) studied the peptide utilization by different Lactococcus lactis strains. They suggested that the peptides were consumed differently by Lactococcus lactis. Strain Wg2 grew poorly in the presence of the basic heptapeptide ISQRYQK as the source of Gln or Ile, whereas it grew at maximum rate in the presence of the basic heptapeptide YPFPGPI as the source of Ile or TVYQHQK as the source of Gln. They indicated that peptide utilization related to both the mass and the charge of the peptides, and also related to the specific strain of Lactococcus lactis. They emphasized that a periplasmic solute-binding protein (OppA) captured oligopeptides and initiated the transport process. Hagting and others (1994) has studied the di and tripeptide transport protein of Lactococcus lactis. They concluded that only a proton motive force- dependent carrier (DtpT) was the transport protein in Lactococcus lactis for hydrophilic di- and tripeptides. Kunji and others (1995) indicated that extracellular peptides formed by the action of proteinase on β-casein disappeared only when the cell express an active oligopeptide transport system in Lactococcus lactis.

1.8 Proteolytic Activity of L. helveticus

The proteolytic enzymes of L. helveticus play an important role in Swiss cheese flavor and may also be involved in P. freudenreichii growth stimulation. The proteolytic system consists of a cell wall-bound extracellular protease, peptide transporters, and intracellular peptidases (Sousa and others 2001). The extracellular protease breaks down large proteins into peptides outside of the cell. The peptides can be imported into the cell via specific peptide transporters and then degraded to amino acids by intracellular peptidases. The

16 amino acids generated are used for cell growth. Upon cell lysis and death during cheese ripening, these peptidases are released into the cheese where they can degrade peptides directly. Most L. helveticus lysis and related proteolysis occurs during the warm room ripening of Swiss cheese, though the lysis rate varies depending on the strain (Valence and others

1998 and 2000). The specificities of the extracellular protease and the released peptidases will dictate the specific peptides present in the cheese. The proteases and peptidases of each

L. helveticus strain may have slightly different specificities, meaning that different peptides would be formed depending on the strain. Caira and others (2003) compared the proteolytic capacity, including protease and peptidase activities, of four L. helveticus strains against synthetic peptides. They found that the rates of hydrolysis and specificities varied markedly among the strains. This conclusion is supported by the findings of Oberg and others (2002)

who observed four different αS1-casein (f 1-23) digestion patterns with nine L. helveticus strains indicating a variety of protease specificities in this species. These studies indicate that the peptide composition of cheese will vary depending on the L. helveticus strain used as a starter.

1.9 Peptide and Amino Acid Utilization by Propionibacteria

Propionibacterium species are weakly proteolytic, but strongly peptidolytic. Lemee and others (1998) evaluated proteolytic activity of the cell free extracts of 37 Propionibacterium strains using a tryptic/chymotryptic β-casein hydrolysate as a substrate. They found that the highest hydrolysis after 24 hours was for P. freudenreichii subsp. shermanii strain CIPO 103027

(1.86 µmol eq. Met/ml). In addition, they found high variation of hydrolysis among

17 Propionibacterium strains. Interestingly, new peptide peaks were produced during the hydrolysis. From this study, 24% of total amino acid content was generated by P. freudenreichii subsp. shermanii strain CIP 103027. Asn was found with P. freudenreichii subsp. freudenreichii strain H, K, and CIP 5932. Gly was absent with P. freudenreichii subsp. shermanii strain CIP 103027. Ile was absent with P. freudenreichii subsp. freudenreichii strain CIP 5932, K and subsp. shermanii strain CNRZ 725. The authors classified the total free amino acid content into three (3) groups. The first group where the highest amount was over 1,100

µmol/L, contained five P. freudenreichii subsp. freudenreichii strains. The second group, with a total free amino acid content between 900-1,000 µmol/L. contained three strains. The last group with less than 800 µmol/L contained two strains. The ratio of each amino acid released upon the hydrolysis was calculated in this study. The result showed that Tyr was the highest at about 57% with strain CIP 103027 and proline was low about 4% for strain

CNRZ725 and 15% for strain CIP 103027, and phenylalamine was about 14% for strain

CNAZ 725 and 31% for strain CIP 103027. The finding of Lamee and others (1998) agreed with Panon (1990), Perez Chaia and others (1990), Quelen and others (1995), and Fernandez and others (1997) in that the samples were low in proline due to proline iminopeptidase,

prolinase, prolidase, and X-prolyl dipeptidyl aminopeptidase enzyme activities.

Langsrud and others (1995) reviewed the rapid degradation of aspartic acid, serine, glycine and alanine in Propionibacterium. They agreed with Piveteau and others (1995) that aspartate is the only amino acid that was metabolized during lactate fermentation.

3 aspartate + propionate 3 succinate + acetate + CO2 + NH3

18 They noted that alanine was only metabolized after lactate was depleted due to high concentration of pyruvate from the lactate fermentation.

3 alanine 2 propionate + acetate + CO2 + NH3

The intracellular peptidases of P. freudenreichii are especially active against proline- containing peptides (Sousa and others 2001). This suggests that though they are unable to generate peptides outside of the cell, they are well equipped to breakdown peptides inside the cell. Therefore, the peptides generated by L. helveticus may promote the growth of P. freudenreichii by providing an amino acid source.

Conflicting reports in the literature make it unclear whether propionibacteria prefer free amino acids or peptides as their amino acid source. Baer (1995) measured the growth of

Propionibacterium and suggested that high concentration of amino acids inhibited their growth.

The author found that peptides stimulated the growth twice as much as amino acids.

However, the subsequent study of Baer and Ryba (1999) presented conflicting results concluding that the addition of the pure amino acids did not cause inhibition and that

Propionibacterium preferred amino acids to peptides for growth. The Baer and Ryba (1999) study conflicted with Thierry and others (1999) who concluded that Propionibacterium prefers peptides to amino acids for growth. Moreover, they found that the addition of 10 g/l amino acids to medium containing peptides inhibited the growth. These differences may be due to study design or to different strains tested.

Though it is unclear if propionibacteria prefer amino acids or peptides, the previous studies establish that these bacteria can utilize peptides. This is not surprising since their

19 intracellular peptidase activity is well known (Sahlstrom and others 1989). However, no reports exist concerning peptide transport in propionibacteria, so it is unclear what type(s) of transporters are used to import the peptides and what their specificities are. It is possible that some peptides are more readily transported and degraded by propionibacteria than others, thus the peptide composition of cheese may impact Propionibacterium growth.

1.10 Effect of Lactobacillus on Propionibacterium Growth

Piveteau and others (1995) found that three strains of L. helveticus stimulated the growth of three strains of P. freudenreichii. Furthermore, the interaction with Lactobacillus may provide some essential peptides to Propionibacterium. Each strain of Lactobacillus potentially hydrolyzes different peptides and/or utilizes different peptides and amino acids for their growth. The peptides and amino acids produced by L. helveticus may affect the growth of P. freudenreichii. Each strain of Propionibacterium possibly possesses different peptide transport systems and intracellular peptidases. Sousa and others (2001) emphasized that

Propionibacterium are weakly proteolytic bacteria. They, however, are strongly peptidolytic ones, especially in proline containing peptides. They also discovered that Pep N, Pep I, Pep

X, and endopeptidase have been found in P. freudenreichii. The finding of Langsrud and others (1995) also suggested that Lactobacillus are more proteolytic than Propionibacterium.

However, Propionibacterium had much higher proline iminopeptidase activities. Finally, they pointed out that amino acids, such as Asp, Ser, Gly, and Ala were rapid utilized by P. freudenreichii. Valence and others (2000) has reported that proteolysis increased mainly in the warm room (22-25°C) by Lactobacillus. Thus, the peptides were hydrolyzed during this

20 ripening time. The peptide hydrolysis of L. helveticus strain ITGLH1 was greater than strain

ITGLH77 by comparing free amino groups and free amino acids. The amino group remained constant during cold storage. The amino acids found mainly were Proline,

Leucine, Lysine, Valine, and Glycine.

Several researchers have explored the relationship between Lactobacillus and P. freudenreichii (Baer 1995; Baer and Ryba 1999; Condon and others 2001; Jimeno and others

1995; Fröhlich-Wyder and others 2002; Parker and Moon 1982; Pérez Chaia and others

1995; Thierry 1999). Piveteau and others (1995) observed the interactions between three strains of P. freudenreichii (T, KM, and H) and three strains of L. helveticus (RR, 303, and 3321).

They found that all three strains of L. helveticus stimulated the growth of the three strains of

P. freudenreichii. The stimulation of the KM strain by RR strain increased the growth rate and conversion of the lactate to propionate and acetate. The authors also found that among five mainly metabolized amino acids, which were aspartate, serine, glycine, alanine and glutamic acid, the aspartate was the only one that stimulated the growth of the KM strain. However, they concluded that it is possible that the aspartate was in form of a peptide with molecular masses around 3000. Their finding agrees with the study of Thierry and others (1999) that peptides stimulate the growth over amino acids, but it opposes Baer and Ryba’s work (1999).

Codon and Cogan (2001) emphasized that many peptides can be involved in stimulation activity. They studied the growth of P. freudenreichii in a whey model and concluded that peptone, tryptone, casein hydrolyzed by crude proteinase of L helveticus strain DPC4571, and by pre-growth of Lactobacillus in milk stimulated P. freudenreichii strain DPC3801 growth.

Pérez Chaia and others (1995) also studied the compound production of these two bacteria.

They found that slow acid production from Lactobacillus strains stimulated the growth rate of

21 Propionibacterium strains and that rapid acid production inhibited them. This finding is an advantage for the cheese making process because it helps researchers understand and control the acid production of the cultures in the cheese process. Beresford and others (2001) reviewed interactions between some Propionibacterium and Lactobacillus strains and pointed out that the consequences of stimulation and inhibition of these bacteria need to be considered in more detail. The purpose of the proposed study is to determine if peptides produced by specific strains of L. helveticus stimulate the growth of specific strains of P. freudenreichii.

22 1.11 Figures

H-arg-pro-lys-his-pro-ile-lys-his-gln-gly-leu-pro-gln-glu-val-leu-asn-glu-asn-leu- leu-arg-phe-phe-val-ala-pro-phe-pro-glu-val-phe-gly-lys-glu-lys-val-asn-glu-leu- ser-lys-asp-ile-gly-sep-glu-sep-thr-glu-asp-gln-ala-met-glu-asp-ile-lys-gln-met- glu-ala-glu-sep-ile-sep-sep-sep-glu-glu-ile-val-pro-asn-sep-val-glu-gln-lys-his- ile-gln-lys-glu-asp-val-pro-ser-glu-arg-tyr-leu-gly-tyr-leu-glu-gln-leu-leu-arg- leu-lys-lys-tyr-lys-val-pro-gln-leu-glu-ile-val-pro-asn-sep-ala-glu-glu-arg-leu- his-ser-met-lys-glu-gly-ile-his-ala-gln-gln-lys-glu-pro-met-ile-gly-val-asn-gln- glu-leu-ala-tyr-phe-tyr-pro-glu-leu-phe-arg-gln-phe-tyr-gln-leu-asp-ala-tyr-pro- ser-gly-ala-trp-tyr-tyr-val-pro-leu-gly-thr-gln-tyr-thr-asp-ala-pro-ser-phe-ser- asp-ile-pro-asn-pro-ile-gly-ser-glu-asn-ser-glu-lys-thr-thr-met-pro-leu-trp-OH

Figure 1.1: The primary structure of αS1–casein (Farrel and others 2004)

H-lys-asn-thr-met-glu-his-val-sep-sep-sep-glu-glu-ser-ile-ile-sep-gln-glu-thr-tyr- lys-gln-glu-lys-asn-met-ala-ile-asn-pro-sep-lys-glu-asn-leu-cys-ser-thr-phe-cys- lys-glu-val-val-arg-asn-ala-asn-glu-glu-glu-tyr-ser-ile-gly-sep-sep-sep-glu-glu- sep-ala-glu-val-ala-thr-glu-glu-val-lys-ile-thr-val-asp-asp-lys-his-tyr-gln-lys- ala-leu-asn-glu-ile-asn-gln-phe-tyr-gln-lys-phe-pro-gln-tyr-leu-gln-tyr-leu-tyr- gln-gly-pro-ile-val-leu-asn-pro-trp-asp-gln-val-lys-arg-asn-ala-val-pro-ile-thr- pro-thr-leu-asn-arg-glu-gln-leu-sep-thr-sep-glu-glu-asn-ser-lys-lys-thr-val-asp- met-glu-sep-thr-glu-val-phe-thr-lys-lys-thr-lys-leu-thr-glu-glu-glu-lys-asn-arg- leu-asn-phe-leu-lys-lys-ile-ser-gln-arg-tyr-gln-lys-phe-ala-leu-pro-gln-tyr-leu- lys-thr-val-tyr-gln-his-gln-lys-ala-met-lys-pro-trp-ile-gln-pro-lys-thr-lys-val- ile-pro-tyr-val-arg-tyr-leu-OH

Figure 1.2: The primary structure of αS2–casein (Farrel and others 2004)

H-arg-glu-leu-glu-glu-leu-asn-val-pro-gly-glu-ile-val-glu-sep-leu-sep-sep-sep-glu- glu-ser-ile-thr-arg-ile-asn-lys-lys-ile-glu-lys-phe-gln-sep-glu-glu-gln-gln-gln- thr-glu-asp-glu-leu-gln-asp-lys-ile-his-pro-phe-ala-gln-thr-gln-ser-leu-val-tyr- pro-phe-pro-gly-pro-ile-pro-asn-ser-leu-pro-gln-asn-ile-pro-pro-leu-thr-gln-thr- pro-val-val-val-pro-pro-phe-leu-gln-pro-glu-val-met-gly-val-ser-lys-val-lys-glu- ala-met-ala-pro-lys-his-lys-glu-met-pro-phe-pro-lys-tyr-pro-val-glu-pro-leu-leu- glu-ser-gln-ser-leu-thr-leu-thr-asp-val-glu-asn-leu-his-leu-pro-leu-pro-leu-leu- gln-ser-trp-met-his-gln-pro-his-gln-pro-leu-pro-pro-thr-val-met-phe-pro-pro-gln- ser-val-leu-ser-leu-ser-gln-ser-lys-val-leu-pro-val-pro-gln-lys-ala-val-pro-tyr- pro-gln-arg-asp-met-pro-ile-gln-ala-phe-leu-leu-try-gln-glu-pro-val-leu-gly-pro- val-arg-gly-pro-phe-pro-ile-ile-val-OH

Figure 1.3: The primary structure of β–casein (Farrel and others 2004)

H-glu-glu-gln-asn-gln-glu-gln-pro-ile-arg-cys-glu-lys-asp-glu-arg-phe-phe-ser-asp- lys-ile-ala-lys-tyr-ile-pro-ile-gln-tyr-val-leu-ser-arg-tyr-pro-ser-tyr-gly-leu- asn-tyr-tyr-gln-gln-lys-pro-val-ala-leu-ile-asn-asn-gln-phe-leu-pro-tyr-pro-tyr- tyr-ala-lys-pro-ala-ala-val-arg-ser-pro-ala-gln-ile-leu-gln-trp-gln-val-leu-ser- asn-thr-val-pro-ala-lys-ser-cys-gln-ala-gln-pro-thr-thr-met-ala-arg-his-pro-his- pro-his-leu-ser-phe-met-ala-ile-pro-pro-lys-lys-asn-gln-asp-lys-thr-glu-ile-pro- thr-ile-asn-thr-ile-ala-ser-gly-glu-pro-thr-ser-thr-pro-thr-thr-glu-ala-val-glu- ser-thr-val-ala-thr-leu-glu-asp-sep-pro-glu-val-ile-glu-ser-pro-pro-glu-ile-asn- thr-val-gln-val-thr-ser-thr-ala-val-OH

Figure 1.4: The primary structure of κ–casein (Farrel and others 2004)

26 1.12 References

Baer, A. (1995). Influence of casein proteolysis by starter bacteria, rennet and plasmin on the growth of Propionibacteria in Swiss-type cheese. Lait, 75, 391-400.

Baer, A., and Ryba, I. (1999). Interactions between propionic acid bacteria and thermophilic lactic acid bacteria. Lait, 79, 79-92.

Beresford, T. P., Fitzsimons, N. A., Brennan, N.L., and Cogan, T.M. (2001). Resent advances in cheese microbiology. International Dairy Journal, 11, 259 - 274.

Caira, S., Ferranti, Ferranti, P., Gatti, M., Fornasari, M.E., Barone, F, Lilla, S., Mucchetti, G., Picariello, G., Chianese, L., Neviani, E., and Addeo, F. (2003). Synthetic peptides as substrate for assaying the proteolytic activity of Lactobacillus helveticus. Journal of Dairy Research, 70, 315-325.

Chaia, A.P., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

Chaia, A.P., Strasser de Saad, A.M., de Ruiz Holgado, A.P., and Oliver, G. (1995). Short- chain fatty acids modulate growth of lactobacilli in mixed culture fermentations with propionibacteria. International Journal of Food Microbiology, 26, 365-374.

Charbonnel, P., Lamarques, M., Piard, J.C., Gilbert, C., Juillard, V., and Atlan, D. (2003). Diversity of oligopeptide transport specificity in Lactococcus lactis species. The Journal of Biological Chemistry, 278, 14832-14840.

Christensen, J. E., Dudley, E.G., Pederson, J. A., and Steele, J.L. (1999). Peptidases and amino acid catabolism in lactic acid bacteria. Antonie van Leeuwenhoek, 76, 217-246.

Condon, S., Cogan, and T.M. (2001). Stimulation of propionic acid bacteria by lactic acid bacteria in cheese manufacture. Dairy Product Research Centre, Cork, Ireland, Project report; ISBN 1841701866 DPRC No.32.

Deutsch, S-M., Ferain, T., Delcour, J., and Lortal, S. (2002). Lysis of lysogenic strains of Lactobacillus helveticus in Swiss cheeses and first evidence of concomitant Streptococcus thermophilus lysis, International Dairy Journal, 12, 591-600.

Deutsch, S-M., Molle, D., Gagnaire, V., Plot, M., Atlan, D., and Lortal S. (2000). Hydrolysis of sequenced β-casein peptides provides new insight into peptidase activity from thermophilic lactic acid bacteria and highlights intrinsic resistance of phosphopeptides. Applied and Environmental Microbiology, 66, 5360-5367.

27 Farrell, H.M., Jimenez-Flores, R., Bleck, G.T., Brown, E.M., Butler, J.E., Creamer, L.K., Hicks, C.L., Hollar, C.M., Ng-Kwai-Hang, K.F., and Swaisgood, H.E. (2004). Nomenclature of the proteins of cows’ milk-sixth revision. Journal of Dairy Science, 87, 1641-1674.

Fernandez-Espla, M. D., and Fox, P. F. (1997). Purification and characterization of X-prolyl dipeptidyl aminopeptidase from Propionibacterium shermanii NCDO 853, International Dairy Journal, 7, 23-29.

Frohlich-Wyder, M. T., Bachmann, H.-P., and Casey, M. G. (2002). Interaction between propionibacteria and starter/ non-starter lactic acid bacteria in Swiss-type cheeses, Lait, 82, 1-15.

Gagnaire, V., Boutrou, R., and Leonil, J. (2001). How can the peptides produced from Emmental cheese give some insights on the structural features of the paracasein matrix?, Elsevier, 11, 449-454.

Gagnaire, V., Molle, D., Herrouin, M., and Leonil, J. (2001). Peptides identified during Emmental ripening: Origin and proteolytic systems involved, Journal of Agricultural Food Chemistry, 49, 4402-4413.

Garault, P., Bars, D.L., Besset, C., and Monnet, V. (2002). Three oligopeptide-binding proteins are involved in the oligopeptide transport of Streptococcus thermophilus, The Journal of Biological Chemistry, 277, 32-39.

Guimont, Ch. (2002). Change of free amino acids in M17 medium after growth of Streptococcus thermophilus and identification of a glutamine transport ATP-binding protein. International Dairy Journal, 12, 729-736.

Hagting, A., Kunji, E.R.S., Leenhouts, K.J., Poolman, B., and Konings, W.N. (1994). The di- and tripeptide transport protein of Lactococcus lactis. The journal of Biochemical Chemistry, 269, 11391-11399.

Helinck, S., Charbonnel, P., Foucaud-Scheunemann, C., Piard, J.-C., and Juillard, V. (2003). Charged casein-derived oligopeptides competitively inhibit the transport of a reporter oligopeptide by Lactococcus lactis, Journal of Applied Microbiology, 94, 900-907.

Husson-Kao, C., Mengaud, J. Gripon, J.- C., Benbadis, L., and Chapot-Chartier, M.- P. (1999). The autolysis of Streptococcus thermophilus DN-001065 is triggered by several food-grade environmental signals, International Dairy Journal, 9, 715-723.

Hynes, E., Ogier, Jean-Claude, Delacroix-Buchet, A. (2000). Protocol for the manufacture of miniature washed-curd cheeses under controlled microbiological conditions. International Dairy Journal, 10, 733-737.

28 Jimeno, J., Lazaro, M. J., and Sollberger, H. (1995). Antagonistic interaction between propionic acid bacteria and non-starter lactic acid bacteria. Lait, 75, 401-413.

Juillard, V., Guillot, A, Bars, D., and Gripon, J-C. (1998). Specificity of milk peptide utilization by Lactococcus lactis Applied and Environmental Microbiology, 64, 1230-1236.

Kunji, E.R.S., Hagting, A., Vries, C.J.D., Juillard, V., Haandrikman, A.J., Poolman, B., and Konings, W.K. (1995). Transport of β-casein-derived peptides by the oligopeptide transport system is crucial step in the proteolytic transport system is a crucial step in the proteolytic pathway of Lactococcus lactis. The Journal of Biological chemistry, 270, 1569-1574.

Langsrud, T., Sorhaug, T., and Vegarud, GE. (1995). Protein degradation and amino acid metabolism by propionibacteria. Lait, 75, 325-330.

Lemee, R., Gagnaire, V., and Maubois, J-L. (1998). Strain variability of the cell-free proteolytic activity of dairy propionibacteria toward β-casein peptides. Lait, 78, 227- 240.

Lemee, R., Lortal, S., and Heijenoort, J.V. (1995). Autolysis of dairy propionibacteria: isolation and renaturing gel electrophoresis of the autolysins of Propionibacterium freudenreichii CNRZ 725. Lait, 75, 345-365.

Oberg, C.J., Broadbent, J.R., Strickland, M. and McMahon, D.J. (2002). Diversity in specificity of the extracellular proteinases in Lactobacillus helveticus and Lactobacillus delbrueckii subsp. bulgaricus. Letters in Applied Microbiology, 34, 455-460.

Panon, G. (1990). Purification and characterization of a proline iminopeptidase from Propionibacterium shermanii 13673, Le Lait, 70, 439-452.

Parker, J.A., and Moon, N.J. (1982). Interactions of Lactobacillus and Propionibacterium in mixed culture. Journal of Food Protection., 45, 326-330.

Perez Chaia, A., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

Pivetean, PG., Condon, S., and Cogan, TM. (1995). Interactions between lactic and propionic acid bacteria. Lait, 75, 331-343.

Quelen, L., Dupuis, C., and Boyaval, P. (1995). Proline-specific activities of Propionibacterium freudenreichii subsp. shermanii. Journal of Dairy Research, 62, 661-666.

Sahlstrom, S., Espinosa, C., Langsrud, T., Sorhaug, T. (1989). Cell Wall, membrane, and intracellular peptides activities of Propionibacterium shemanii. Journal of Dairy Science, 72, 342-350.

29 Shakeel-Ur-Rehman, Fox, P.F., McSweeney P.L.H, Madkor, S.A. and Farkye, N.Y. (2001). Alternatives to Pilot Plant Experiments in Cheese-Ripening Studies. International Journal of Dairy Technology, 54, 121-126.

Sousa, M.J., Ardo, Y., and McSweeney, P.L.H. (2001). Advances in the study of proteolysis during cheese ripening. International Dairy Journal, 11, 327-345.

Stepaniak, L., Tobiassen, R.O., Chuckwu, I., Pripp, A.H., and Sorhaug, T. (1998). Purification and characterization of a 33 Kda Subunit Oligopeptidase from Propionibacterium frendenreichii ATCC 9614. International Dairy Journal, 8, 33-37.

Thierry, A., Salvat-Brunaud, D., and Maubols, Jean-Louis. (1999). Influence of thermophilic lactic acid bacteria strains on propionibacteria growth and lactate consumption in an Emmental juice like medium. Journal of Dairy Research, 66, 105-113.

Valence, F., Dentsch, S.M., Richoux, R., Gagnaire, V. and Lortal, S. (2000). Autolysis and related proteolysis in Swiss cheese from two Lactobacillus helveticus strains. Journal of Dairy Research, 67, 261-271.

Valence, F., Richoux, R., Thierry, A., Palva, A., and Lortal, S. (1998). Autolysis of Lactobacillus helveticus and Propionibacterium freudenreichii in Swiss cheese: first evidence by using species-specific lysis markers. Journal of Dairy Research, 65, 609-620.

Williams, A.G., Felipe, X., and Banks, J.M. (1998). Aminopeptidase and dipeptidyl peptidase activity of Lactobacillus spp. and non-starter lactic acid bacteria (NSLAB) isolated from cheddar cheese. International Dairy Journal, 8, 255-266.

30 2. GROWTH OF PROPIONIBACTERIUM FREUDENREICHII IN CHEESE SLURRY SERUM PRODUCED

WITH DIFFERENT LACTOBACILLUS HELVETICUS STRAINS

CHAPTER 2

GROWTH OF PROPIONIBACTERIUM FREUDENREICHII IN

CHEESE SLURRY SERUM PRODUCED WITH DIFFERENT

LACTOBACILLUS HELVETICUS STRAINS

2.1 Abstract

Swiss cheese makers report that some combinations of L. helveticus and P. freudenreichii cultures result in poor eye and flavor formation, whereas other combinations perform well.

The objective was to develop a rapid method to predict successful strain pairings. Cheese curds were prepared aseptically from UHT milk using S. thermophilus alone (control) or S. thermophilus plus one of four L. helveticus strains. Curds were homogenized into a slurry with a final pH of 5.3, 60% moisture and 1% salt in moisture. The slurries were centrifuged, and the resulting serum was sterile filtered. Each sterile slurry serum was added in the ratio 1:1 to

2× chemically defined medium (CDM) lacking amino acids. One of five P. freudenreichii strains was inoculated into CDM containing each slurry serum, and growth was monitored

spectrophotometrically. Maximum specific growth rate (µmax) and time to reach µmax were calculated using the Richards model. Propionibacterium freudenreichii growth was not observed in

31 CDM lacking slurry serum. CDM supplemented with slurry sera, including the control serum prepared without L. helveticus, supported the growth of four of the five P. freudenreichii strains to different extents. Propionibacterium freudenreichii P764M1 did not grow with any slurry

serum. L. helveticus L350 slurry serum stimulated the µmax and reduced time to reach µmax of four P. freudenreichii strains beyond that observed with the control serum. In contrast, L.

helveticus L346 slurry serum increased the time to reach µmax of these four strains in

comparison to the control serum; however, this serum also increased µmax of P859 and P873 despite in the delay in growth onset. L887 slurry serum delayed P. freudenreichii ATCC9614 growth, but had no effect compared to the control on any other P. freudenreichii strain. L856 slurry serum also slightly delayed ATCC9614 growth compared with the control serum, but

increased the µmax of P859.

32 differences in P. freudenreichii growth in the slurry. The slurry serum model has potential for screening P. freudenreichii and L. helveticus pairings for Swiss cheese manufacturing.

2.2 Introduction

Three bacterial species are used to manufacture Swiss cheese in the U.S.: Streptococcus thermophilus, Lactobacillus helveticus and Propionibacterium freudenreichii. Streptococcus thermophilus produces lactic acid from lactose during the first stages of cheese production. Lactoabacillus helveticus also produces lactic acid and plays another important role by contributing to the proteolysis of the casein proteins during cheese ripening. Proteolytic enzymes (a protease and several peptidases) from L. helveticus help to breakdown casein into peptides and amino acids important to cheese flavor. Propionibacterium freudenreichii degrades the lactic acid produced by the other starter cultures and produces propionic and acetic acids, which contribute to cheese flavor, and carbon dioxide gas, which creates the holes called “eyes” in

Swiss cheese.

Many different strains of each bacterial species are available commercially. Thus, the cheese maker has many options concerning the combination of the three cultures to use.

Cheese makers have reported that some combinations of L. helveticus and P. freudenreichii result in poor eye and flavor formation during ripening.

Cheese experiments are expensive and time consuming. Most experiments in cheese studies used a model to reduce time and cost. Some researchers suggested that pilot plant scale cheese models were closer to commercial cheese properties and the biochemical reaction closes to normal cheese during ripening (Shakeel-Ur-Rehman 2001). While pilot scale studies

33 reduce the cost of the experiment, the ripening time required is the same. Some models were reported for microbiological studies of cheese ripening with a short period of time. The experiment was conducted under sterile conditions. Hynes and others (2000) described a protocol of the miniature model under controlled microbiological conditions. The dry matter and the salt in moisture were not significant difference from aged cheeses as for large-scale cheeses. However, the cheese slurry models used for studies of the flavor development during cheese ripening were unclear and the slurries developed off flavors (Davies and Law 1984).

The results of a preliminary study (Appendix A) indicated that the slurry sera prepared with different L. helveticus strains supported different growth patterns of different P. freudenreichii strains. The nutrient requirements of Propionibacterium differ from strain to strain. Glatz (1988) developed a minimal medium to support the Propionibacterium growth. This medium is composed of vitamins, adenosine, inorganic salts and amino acids. The Propionibacterium strains did not grow as well as in complex medium. The strains that synthesized their own nutrients grow slower than in the rich medium. Some strains required amino acids such as cysteine or methionine for their growth and some strains were reported to be inhibited by cysteine or did not require any amino acid for the growth. All strains, however, used lactate as a carbon source

and produced propionate, acetate, and CO2 in cheese. McCarthy and Courtney (2004) also reported strain-dependent differences in amino acid requirements of Propionibacterium.

The casein in milk can provide amino acids source through the proteolysis process.

The peptides and amino acids produced by Streptococcus and Lactobacillus may help support or stimulate the growth of Propionibacterium. Some research indicated that Propionibacterium growth depend on free amino acids (Lemee and others 1998; Piveteau and others, 1995). Aspartate as free amino acid in the Propionibacterium metabolic pathway was reported in several studies to

34 release succinate, acetate, CO2, and NH3 in a strain dependent manner. In addition, alanine was metabolized after exhaustion of lactate to propionate, acetate, CO2, and NH3. However, variation in growth and maximum autolysis was reported to be strain specific (Langsrud and others, 1995; Piveteau and others, 1995; Valence and others, 1998).

Propionibacterium species are weakly proteolytic, but strongly peptidolytic. Their intracellular peptidases are especially active against proline-containing peptides (Sousa and others. 2001). This suggests that though they have low ability to generate peptides outside of the cell, they are well equipped to breakdown peptides inside the cell. Therefore, the peptides generated by L. helveticus during cheese manufacturing may promote the growth of

P. freudenreichii by providing an amino acid source. Pairing of a L. helveticus strain providing the necessary peptides and amino acids for the P. freudenreichii strain selected may be important for proper flavor and eye development of Swiss cheese.

The goals of the present study were to better understand the relationship between L. helveticus and P. freudenreichii and to develop a model to predict which strains will work well together. The approach was to use the serum from cheese slurries prepared with different L. helveticus strains as a growth medium additive for studying Propionibacterium growth. The slurry sera are proposed to serve as a peptide and amino acid source to support Propionibacterium growth in chemically defined medium (CDM) lacking amino acids.

Hypothesis:

1. Different L. helveticus strains differentially affect the growth of P. freudenreichii in a

cheese-like model system.

2. Different L. helveticus strains produce different peptides in a cheese-like model

system (slurry serum).

Objectives:

1. Develop a cheese-like model system for rapid evaluation of strain interactions.

2. Evaluate the growth of P. freudenreichii strains in the model system prepared

with different L. helveticus strains.

3. Compare the peptide content of slurry sera prepared with different L. helveticus

strains.

2.3 Materials and Methods

2.3.1 Swiss cheese slurry serum preparation with different Lactobacillus helveticus strains

A model aseptic cheese slurry system representing conditions similar to a ripening

Swiss cheese was developed (Jenkins, Limpisathian and Courtney, unpublished results) and is described below. The model has two advantages over real Swiss cheese for the present study: (1) it can be performed under aseptic conditions, allowing study of specific microorganisms and (2) it requires less time and raw material than cheese.

Swiss cheese slurries were prepared aseptically in miniature cheese make vats designed and made according to Roberts and others (1995). All vats and equipment were sterilized by autoclaving prior to making curds. To prepare cheese slurries, one liter of sterile fat-free UHT milk (Parmalat USA) was aseptically dispensed into a sterile 1 L

container. The pH was measured then adjusted to 6.0 using L-lactic acid, and sterile CaCl2

36 was added to a final concentration of 0.018%. The milk was warmed to 33.3°C for 30 minutes in a circulating water bath before adding Streptococcus thermophilus (S787) and one of the L. helveticus strains. Or, for the control, no L. helveticus strain was added. Four strains of

L. helveticus were used. Approximately 150 mg (wet weight) of each culture were added. The

S. thermophilus strain was the same for all preparations. A 2.5% solution of chymosin

(ChyMax Extra, Chr. Hansen, Milwaukee, WI) (5.5 ml) was added and the milk was left undisturbed at 33.3°C for 40 minutes to produce the coagulated curd. The curd was cut into small cubes with a sterile wire-mesh cutting device. The curds and whey were cooked to

50°C slowly in water bath. A “dummy” container was used to monitor the temperature.

The pH near the end of cooking was measured and curds were held at 50 °C until the pH reached 5.8. When the pH was 5.8, the whey was drained through sterile cheesecloth. The remaining curd was cooked further until the pH of the whey reached 5.6-5.5. After further whey draining, the curd moisture content was measured using the microwave moisture analyzer method (LabWave 9000 Moisture/Solids Analyzer, CEM, Metthews, NC). Sterile

22% NaCl and sterile distilled water were added to the curds to yield a final moisture content of 60%, salt in moisture of 1% at pH 5.3. The mixture was homogenized for 5 minutes to form a thick slurry.

Each slurry was centrifuged at 10 krpm for 10 minutes and the supernatant was recovered. The supernatants were filtered through a 0.45 µm filter and frozen at -80°C until use. This sterile slurry supernatant is referred to as the slurry serum. Part of each sterile slurry serum was used as the growth medium for P. freudenreichii in the preliminary experiments described in Appendix A. The rest was used a growth medium additive in the experiments described below and for peptide analysis by HPLC. Triplicate slurry sera were 37 prepared from each of the four L. helveticus strains. Triplicate control slurries lacking L. helveticus also were prepared.

2.3.2 Propionibacterium freudenreichii strains and growth conditions

Five different P. freudenreichii strains from commercial starter cultures or the American

Type Culture Collection were selected for analysis: P859, P873, P812, P764M1, and

ATCC9614. The pure stock cultures were stored at -80 ºC and inoculated into Propionibacterium medium (20 g/L casein-digestive tryptone, 5 g/L yeast extract, 5 g/L glucose). Cultures were grown in an anaerobe chamber (85% nitrogen, 10% hydrogen, 5% carbon dioxide) at 28 to

30° C to stationary phase before used to initiate an experiment. Slurry sera prepared with different L. helveticus strains; L346, L350, L856, and L887 including a control lacking L. helveticus were kept at -80 ºC before use in growth assays. Single-strength (1×) chemically defined medium (CDM) (Glatz and Anderson 1988) with the following composition was prepared and

filter-sterilized: 1.74 g/L K2HPO4, 1.36 g/L KH2PO4, 1.6 g/L (NH4)2NO3, 100 mg/L

MgSO4·7H2O, 1.0 mg/L FeSO4·7H2O, 3.0 mg/L MnSO4·4H2O, 1.0 mg/L CoCl3, 1.7 mg/L thiamine HCl, 2.5 mg/L biotin, 2.5 mg/L Ca-pantothenate, 133.5 mg/L adenosine, and 12 g/L Na lactate. Double-strength CDM (2×) was prepared by doubling the concentrations of the nutrients listed. Single-strength CDM containing amino acids was prepared by adding amino acids at the following concentrations (Glatz and Anderson,1988): 42 mg/L Alanine, 104 mg/L arginine, 42 mg/L asparagine, 40 mg/L aspartic acid, 36 mg/L cysteine, 735 mg/L glutamic acid, 730 mg/L glutamine, 10 mg/L glycine, 16 mg/L histidine, 40 mg/L isoleucine,

55 mg/L lysine, 40 mg/L leucine, 45 mg/L methionine, 50 mg/L phenylalanine, 230 mg/L

38 proline, 420 mg/L serine, and 36 mg/L threonine. A final concentration of 0.5% NaCl (w/v) was also added to 1× CDM. All CDM were filter-sterilized through a 0.2 µm filter after all components were added.

After growth of each of the five P. freudenreichii strains in complex Propionibacterium medium to stationary phase, strains were inoculated into 1× CDM (0.5% NaCl), 1× CDM

(0.5% NaCl) + amino acids, or 1× CDM + slurry serum (figure 2.1). Single strength CDM + slurry serum was prepared by combining 2× CDM and sterile slurry serum in a 1:1 volumetric ratio. Because the slurry sera contain 1% NaCl, no NaCl was added to the 2× CDM, giving a final growth medium with 0.5% NaCl. Five different 1× CDM + slurry serum media were prepared, one for each slurry serum (control + 4 L. helveticus strains). 200 µl of each medium were pipetted to a well of a sterile 96-well microtiter plate. Four µl of each stationary phase P. freudenreichii culture were inoculated into each of the seven media. An uninoculated blank was left as the absorbance blank. The microtiter plate was incubated in an anaerobe chamber for 12 days at 28-30°C. Absorbance (600 nm) was measured at 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and

12 days in a microtiter plate reader. The experiment was replicated three times for each P. freudenreichii strain.

Growth curves were plotted ,and the maximum specific growth rate (µmax) and time

to reach µmax were determined by fitting the growth curves to the Richards model with the m parameter fixed to 2.0 as described by Dalgaard and Koutsoumanis (2001) using the

SigmaPlot 9.0 software package (Systat Software, Inc., Point Richmond, CA). Curves that

did not reach an A600 of 0.2 within 45 days could not be fit to the Richards model and were

labeled as NG (no growth). Mean values for µmax and time to reach µmax from triplicate

39 growth curves were compared by one-way analysis of variance with Tukey’s post-hoc test

(SigmaStat 3.1, SysStat Software, Inc.). Values with P < 0.05 were considered to be significantly different.

2.3.3 Evaluation of peptide profiles in slurry sera by HPLC

The remainder of the slurry sera described above were stored at -80°C until analysis by HPLC. One hundred micro liters of each slurry serum were diluted with 200 µl 8.75 M urea and adjusted to pH 2.0 with 10% TFA. The volume of TFA added to each sample was recorded to calculate the total volume of each sample. The samples were then filtered through a 0.22 µm membrane and again kept at -80ºC before HPLC analysis. Fifty µl of the slurry serum were analyzed by injecting to the HPLC. Solvent A was 0.106% TFA (v/v) in

HPLC water. Solvent B was 0.1% TFA (v/v) and 80% acetonitrile (v/v) in HPLC water.

The peptide content of the original slurry serum was observed by analytical reverse phase

HPLC on reversed phase Vydac C18 column with 4.6 i.d, 250 mm length, 300 Å pore size, 5

µm particle size and column volume 4.2 ml using UV detection at 214 nm as described by

Gagnaire and others. (2001). A linear gradient from 2 to 70% of solvent B and solvent A were applied for the elution at 0.8 ml/min at 40ºC. The peptide peaks were reported as peak area per volume of sample (Area/µl).

Two separate, replicate slurry sera were evaluated for each data point. Peptides generated by the four L. helveticus strains were compared to each other and to the control lacking L. helveticus. Differences in peptide content between the samples were evaluated by comparing quantitatively as concentration using one-way analysis of variance followed by

40 Tukey’s test. The peaks that were different from the control lacking L. helveticus were considered to compare the peptide content.

Propionibacterium Peptide profile growth evaluation evaluation by HPLC

2× chemically defined Slurry serum medium lacking amino (1% NaCl) acids and peptides

1:1

Monitored culture Inoculated with one of five absorbance (600 nm). P. freudenreichii strains at Growth curves were 2% (vol:vol) plotted and modeled. CDM + serum (0.5% NaCl)

Figure 2.1: Cheese-like model system (slurry serum) for evaluating L. helveticus and P. freudenreichii strain pairings

2.4 Results

2.4.1 Propionibacteium growth in CDM + slurry serum

Propionibacterium freudenreichii growth was not observed in 1× CDM (0.5% NaCl) lacking slurry serum or amino acids; whereas, the positive control medium, 1× CDM (0.5%

NaCl) + amino acids did support growth (data not shown). CDM supplemented with slurry sera, including the control serum prepared without L. helveticus, supported the growth of four

41 of the five P. freudenreichii strains to different extents (tables 2.1 and 2.2). Propionibacterium freudenreichii P764M1 did not grow with any slurry serum (figure 2.5). L. helveticus L350 slurry

serum stimulated growth rate (table 2.1) and reduced the time to reach µmax of four P. freudenreichii strains beyond that observed with the control serum (table 2.2). In contrast, L.

helveticus L346 slurry serum increased the time to reach µmax of these four strains in

comparison to the control serum; however, this serum also increased the µmax of P859 and

P873 despite in the delay in growth onset. L887 slurry serum delayed P. freudenreichii

ATCC9614 growth, had no effect compared to the control on any other P. freudenreichii strain. L856 slurry serum also slightly delayed ATCC9614 growth compared with the control serum, but increased the growth rate of P859.

2.4.2 Comparison of peptides generated by four different L. helveticus strains

The results indicated that 54 of 126 peaks were significantly different (p<0.05) in peptide content between the control containing only S. thermophilus and those slurry sera prepared with L. helveticus strains (table 2.3). Although, ninety-seven peptide peaks were found both in the control lacking L. helveticus strains and in L. helveticus strains. Among those, most of them were increased quantitatively as concentration; 26, 24, 15, and 5 peaks in L. helveticus strain L350, L856, L887, and L346, respectively (table 2.3). Four to ten peaks were decreased for all L. helveticus strains. As compared to the control, the extracellular protease of L. helveticus likely plays a role in reducing the concentration of some peptides. However, there were seven to fifteen peaks lost as compared to the control for all L. helveticus strains.

The peaks were not found in the control lacking L. helveticus, but found in L. helveticus strains

42 were considered. These may be smaller peptides resulting from protease activity. Among them, thirteen and ten additional peptide peaks were found in L. helveticus L350 and L856 strain and only half, 3 and 6 peaks, were found in L. helveticus L346 and L887 strains, respectively. Lactobacillus helveticus strain L350 and L856 produced more peptide peaks, 36 and

32 peaks that were significantly different in peptide content from the control lacking L. helveticus than L. helveticus strains L346 and L887, 9 and 20 peaks, respectively. Lactobacillus helveticus, thus, produced peptides in a strain specific manner.

2.5 Discussion

Slurry sera from different L. helveticus strains differed in their peptide profiles (2.4.2), which may contribute to the observed differences in P. freudenreichii growth depending on L. helveticus strain. Specific peptides produced by the extracellular protease of each L. helveticus strain may play a role in these differences since P. freudenreichii strains have different amino acid requirements (McCarthy and Courtney, unpublished data) and have different abilities to import (Chapter 3) and utilize peptides (Chapter 4). The screening system developed provides a rapid method for evaluating potential culture combinations for Swiss cheese manufacturing. It will also be useful in further research on the relationship between these species.

43 2.5.1 Propionibacteium growth in CDM+ slurry serum

The different growth pattern of P. freudenreichii found in this study has been observed in other researchers. Parker (1982) studied the interaction of Lactobacillus and

Propionibacterium in mixed culture by using a defined synthetic medium with glucose as an energy source and casamino acids as a nitrogen source. Fifteen of 16 paired cultures have been less grown, acid and carbon dioxide produced than the pure culture controls. Some paired cultures were inhibited whereas some paired cultures only Propionibacterium was inhibited. The study of these interactions in whey (Piveteau, 1995) showed that the growth of 3 P. freudenreichii strains were stimulated by 3 strains of L. helveticus and S. thermophilus, but no inhibition of growth was found. Baer (1995) studied the growth of Propionibacterium in

UHT milk containing calcium carbonate to maintain the pH constant. The author reported that the peptides produced by plasmin were used by Propionibacterium. The rennet may have the significant effect on the growth of Propionibacterium and too high concentration of free amino acids in cheese may inhibit growth of Propionibacterium.

2.5.2 Comparison of slurry model used in this study to others utilized

The moisture of previously reported slurry models were between 57-73% whereas the moisture of the slurry model in this study was 60%. Most of the models were high in 3-

4% NaCl, but the slurry model used in this study was 1% NaCl. Many studies investigated the cheese slurry model to evaluate the effect of cultures. Roberts (1995) developed an aseptic cheese curd slurry system containing 73% (w/w) moisture and 3% salt. This model

44 was used to investigate the starter bacteria on the development of cheese flavor. Farkye

(1995) modified the cheese slurry of Kristoffersen, 1967 to give a slurry containing 57-70% moisture, 18-23% protein, 3-4% NaCl with a pH of 4.85 - 5.32. The model was used to evaluate the lactic cultures and enzymes to cheese ripening and flavor development. In the present study, the serum was recovered from the slurry and sterilized for further analyses.

To our knowledge, this is the first study utilizing cheese slurry serum to evaluation culture combinations.

2.5.3 Comparison of peptides generated by four different L. helveticus strains.

Most peptides produced by L. helveticus strain L346 and L887 were not significantly different from the control (117 and 106 peaks, respectively). The result from the study of P. freudenreichii growth in CDM+slurry serum (2.5.1) showed that L. helveticus strain L346 slurry serum slowed the onset of growth of P. freudenreichii strains P859, P873, P812, and

ATCC9614. In addition, L. helveticus strain L887 did not stimulate or inhibit the growth of any Propionibacterium strain (2.5.1). In contrast, L. helveticus strain L350 produced more peptides that were significant different from the control than L346 and L887. It is interesting that L. helveticus strain L350 consistently promoted the growth of all of the P. freudenreichii strains (P859, P873, P812, and ATCC9614) to the greatest degree compared to other L. helveticus strains. It implied that peptides produced by Lactobacillus strains stimulate

Propionibacterium growth. Baer (1995) found propionibacteria grow well in the presence of

LAB (starters in Emmental manufacture) alone. Thierry and others (1999) studied influence of LAB by mixed cultures of S. thermophilus with each 4 strains L. helveticus (ITG LH77, ITG

45 LH56, KFL72 and KFL79) on five P. freudenreichii strains (CIP 103026, CIP 10327, TL173, and ITG 14). The authors found that propionibacteria growth was influenced by LAB.

They were stimulated by high peptide level, low level of free amino acids and NaCl. Condon and Cogan (2000) reported the stimulation of seven P. freudenreichii strains by eight L. helveticus in a whey model. The intensity of the stimulation of P. freudenreichii strains varied with the L. helveticus strains used. The authors emphasized that the peptides produced by L. helveticus strains stimulated P. freudenreichii growth. Gagnaire and others (2001) studied peptides in Emmental cheese and observed P. freudenreichii consumed peptides more than produced them and the peptide profile was changed quantitatively but not qualitatively during warm room ripening. Lemee and other (1998) studied proteolytic activity of P. freudenreichii. They showed that the activity varied significantly among the strains in term of peptides and free amino acids release.

Lactobacillus helveticus L856 produced more peptides than strains L346 and L887 and promoted the growth of P. freudenreichii only one strain (P859). It is possible that the peptides produced and the stimulation effect are strain specific, thus only certain L. helveticus

+ P. freudenreichii pairs result in sufficient P. freudenreichii growth. Piveteau and others (1995) reported that peptides produced by L. helveticus strain RR were utilized by P. freudenreichii strain KM. Some peptides were higher in concentration and some peptides were hydrolyzed to smaller ones during subsequent growth of P. freudenreichii.

The slurry serum model has potential for rapid screening P. freudenreichii and L. helveticus pairings for Swiss cheese manufacturing. The peptides produced and the stimulation effect observed are strain-specific, thus only certain L. helveticus + P. freudenreichii pairs result in sufficient P. freudenreichii growth. In addition, the slurry serum model has potential for rapid

46 screening of L. helveticus strains which yield different peptide profiles. Future work recommendation is to validate the model by making Swiss cheese with selected strain pairings and further analysis of peptide profiles may reveal the identities of specific peptides utilized by

P. freudenreichii strains.

2.6 Tables

Maximum specific growth rate (d-1) Propionibacterium Lactobacillus helveticus strain used to prepare slurry serum freudenreichii strain Control L350 L346 L856 L887 (no L. helveticus) P859 0.58 c 2.88 a 1.05 b 0.96 b 0.54 c P873 0.63 c 3.00 a 1.27 b 0.73 b,c 0.63 b P812 1.10 b 3.00 a 0.69 b 0.65 b 0.71 b P764M1 NG NG NG NG NG ATCC9614 1.03 b 2.62 a 0.67 b 0.73 b 0.89 b Means with the same superscript letter within a row are not significantly different (P < 0.05), n=3.

Table 2.1: Maximum specific growth rates (d-1) of five Propionibacterium strains in chemically defined medium (0.5% NaCl) with added slurry serum prepared with one of four Lactobacillus helveticus strains.

Lag phase (days) Propionibacterium Lactobacillus helveticus strain used to prepare slurry serum freudenreichii strain Control L350 L346 L856 L887 (no L. helveticus) P859 7.37 b 1.99 c 10.34 a 7.27 b 6.76 b P873 6.61 b 2.13 c 9.93 a 6.82 b 5.84 b P812 4.47 b,c 2.03 d 7.36 a 3.91 c 5.16 b P764M1 NG NG NG NG NG ATCC9614 2.98 d 2.03 e 5.52 a 3.72 c 4.43 b Means with the same superscript letter within a row are not significantly different (P < 0.05), n=3.

Table 2.2: Lag phase (days) of five Propionibacterium strains in chemically defined medium (0.5% NaCl) with added slurry serum prepared with one of four Lactobacillus helveticus strains.

Peak Time Control L346 L350 L856 L887 No. Min Area/µlArea/µlArea/µlArea/µl Area/µl 1 5.11 284.4 258.7 264.5 242.8 267.0 2 5.20 194.3 265.5 283.0 249.6 315.0 3 5.56 519.9 510.9 545.8 554.2 551.5 4 5.97 52.7 54.7 71.0 33.8 66.4 5 6.31 124.3 77.2 84.9 58.3 61.8 6 6.68 7.8 16.3 16.0 24.9 26.0 7 6.82 7.2 0.8 20.0 10.7 0.0 8 7.14 9.7 16.9 22.4 18.3 18.4 9 7.38 9.0 11.9 21.7 14.9 15.0 10 7.62 9.5 11.3 20.3 21.3 12.9 11 7.88 7.1 8.1 16.8 9.0 0.0 12 8.13 77.9 66.1 85.7 91.1 40.3 13 8.49 7.1 0.4 5.3 22.4 1.7 14 9.86 18.0 22.5 34.5 65.1 31.1 15 12.04 5.4 4.7 5.9 12.0 6.2 16 13.35 6.3 6.0 13.6 19.3 3.4 17 13.68 0.0 0.0 3.9 8.3 0.0 18 14.29 3.8 4.3 5.2 13.1 4.2 19 14.68 0.0 1.7 4.6 6.4 1.7 20 15.34 58.1 72.7 59.1 83.3 85.0 21 15.97 3.2 7.9 8.5 11.3 6.1 22 16.56 8.34b 6.7b 12.5b 33.2a 7.6b 23 16.93 1.4 4.9 13.4 12.9 4.3 24 17.32 15.9b 22.4b 88.2a 39.2b 22.3b 25 17.79 2.6b 4.1b 57.8a 47.8a 3.3b 26 18.15 49.9a 40.7a 32.2a 7.5b 35.5a 27 18.60 0.0b 0.0b 21.6a 43.3a 1.8b 28 19.27 0.8 4.1 10.1 15.0 7.7 29 19.99 1.9 21.1 6.2 6.2 4.1 30 20.71 7.4 1.2 7.8 11.9 0.0

Continued

Table 2.3: The peptide content for each slurry prepared with the different L. helveticus strains and the control S. thermophilus

49 Table 2.3 continued

Peak Time Control L346 L350 L856 L887 No. Min Area/µlArea/µlArea/µlArea/µl Area/µl 31 21.47 18.5 17.1 18.9 21.9 7.9 32 21.79 0.0b 0.7b 39.2a 32.8a 0.2b 33 22.64 3.9c 6.4c 96.2a 128.8a 28.7b 34 22.85 5.7b 8.0b 38.9a 39.6a 8.7b 35 23.67 0.0b 0.0b 16.4a 14.9a 4.2b 36 24.09 0.0 0.0 0.0 6.2 0.0 37 24.57 2.7b 0.0b 29.6a 26.6a 0.0b 38 25.09 14.9 8.2 18.6 21.4 11.4 39 25.84 6.3 6.7 21.8 16.0 15.6 40 26.37 7.3b 9.6b 18.5ab 28.6a 19.0ab 41 26.86 6.3 6.0 0.0 8.6 11.9 42 27.60 0.0 0.7 6.8 7.9 6.2 43 28.89 6.8b 6.3b 12.0b 16.0ab 29.7a 44 29.18 5.4b 0.0b 0.0b 15.7a 12.7a 45 29.55 18.0 5.2 16.8 7.6 0.0 46 29.95 0.0b 0.0b 7.5a 14.4a 0.0b 47 30.15 279.3a 91.1b 86.0b 52.7b 113.4b 48 30.49 1.6b 2.2b 5.4b 3.3b 25.4a 49 30.73 18.9ab 10.8b 0.0b 35.3a 17.7ab 50 31.00 24.4 14.3 40.5 58.3 36.7 51 31.15 0.0b 0.0b 27.8a 0.0b 4.0b 52 31.36 0.0b 9.2ab 0.0b 1.0b 18.3a 53 31.50 17.9 12.8 3.9 4.6 0.0 54 32.19 2.5 6.4 16.0 13.1 2.6 55 32.42 6.8a 9.2a 0.0b 0.0b 2.8a 56 32.91 5.5b 2.2b 33.5a 10.9b 6.6b 57 33.30 0.0c 5.0b 79.5a 72.7a 16.9b 58 34.13 12.7b 9.6b 58.0a 43.2a 39.3a 59 34.37 2.4b 0.0b 39.0a 71.1a 17.2ab 60 34.63 25.3a 6.3b 3.5b 21.2a 48.2a 61 35.24 399.1a 459.5a 32.0b 46.5b 480.1a 62 35.30 0.0b 0.0b 55.7a 14.7b 0.0b 63 35.95 6.8 4.8 0.0 1.7 3.4 64 36.38 2.2 3.1 6.1 13.7 13.0 65 36.74 10.4 7.6 8.1 12.2 9.9

Continued

50 Table 2.3 continued

Peak Time Control L346 L350 L856 L887 No. Min Area/µlArea/µlArea/µlArea/µl Area/µl 66 37.10 8.8 3.9 20.4 19.2 10.4 67 37.50 1.2 1.6 0.0 8.3 5.0 68 37.66 6.0 1.0 5.5 4.3 5.2 69 37.93 7.7 0.5 7.9 20.9 5.6 70 38.33 220.0 164.7 165.6 117.9 187.8 71 38.85 92.0a 110.2a 8.8b 26.9b 141.8a 72 39.24 0.0b 1.6b 11.3a 10.5a 0.0b 73 39.57 4.1 7.3 17.6 19.4 10.9 74 39.91 5.4 0.0 2.6 15.9 6.4 75 40.31 13.1 14.4 22.6 26.8 17.8 76 40.89 5.1 3.5 2.1 15.1 0.0 77 41.56 4.7c 0.0c 14.4b 44.2a 18.6b 78 41.79 4.6 13.6 3.5 2.9 20.4 79 41.96 13.5 12.7 6.3 18.9 7.0 80 42.28 0.0b 0.0b 0.0b 0.0b 9.6a 81 42.30 2.4 10.5 4.7 0.0 7.7 82 42.60 15.6 6.2 11.5 21.2 10.5 83 42.87 6.1 8.5 6.5 17.3 10.8 84 43.22 42.1a 9.4b 60.5a 37.0a 15.7b 85 43.60 0.0 10.4 12.4 20.0 6.8 86 44.03 0.0 4.4 0.0 10.8 13.6 87 44.22 0.0b 0.0b 16.7ab 35.5a 9.1ab 88 44.46 25.6 48.6 52.1 58.6 51.4 89 44.93 62.6ab 91.6a 28.9b 34.6b 0.0c 90 45.33 42.8b 0.0b 0.0b 0.0 116.7a 91 45.63 48.2 49.4 67.0 78.2 66.6 92 46.00 42.9 45.9 39.5 34.7 63.5 93 46.71 9.5 13.5 16.0 18.5 24.1 94 47.18 0.0 21.1 8.1 13.2 0.0 95 47.31 13.6 21.8 36.1 10.9 63.0 96 47.65 46.8 48.7 6.5 7.3 57.5 97 48.00 27.9a 0.0b 19.6a 28.0a 0.0b 98 48.14 0.0b 0.0b 20.2a 16.9a 7.8ab 99 48.46 4.8b 24.1a 29.8a 25.9a 21.5a 100 48.95 13.9a 16.1a 0.0b 18.4a 27.3a

Continued

51 Table 2.3 continued

Peak Time Control L346 L350 L856 L887 No. Min Area/µlArea/µlArea/µlArea/µl Area/µl 101 49.25 7.7a 11.9a 0.0b 22.6a 15.0a 102 49.65 5.4 19.1 16.4 24.6 7.3 103 49.98 6.4 0.0 5.7 0.0 24.2 104 50.17 0.0b 0.0b 21.7a 0.0b 0.0b 105 50.28 34.6a 31.4a 0.0b 27.6a 19.5a 106 50.63 13.5 11.0 28.1 9.8 33.1 107 51.00 10.5 22.2 16.4 16.8 25.0 108 51.15 0.0b 7.7b 33.0a 27.2a 5.9b 109 51.74 16.4 17.4 17.3 27.6 13.9 110 52.07 14.5b 30.0a 0.0c 0.0c 34.8a 111 52.50 20.4b 37.7b 143.2a 120.7a 22.9b 112 52.76 13.7b 30.3ab 0.0b 0.0b 48.4a 113 52.88 42.0ab 78.6a 19.4b 16.1b 92.7a 114 53.16 79.3 55.6 24.9 21.6 53.0 115 53.79 4.6 17.9 24.9 6.1 15.4 116 54.21 51.3 53.7 49.0 30.5 70.4 117 54.52 259.9a 252.5a 0.0c 53.6b 277.4a 118 54.91 0.0b 0.0b 0.0b 53.2a 0.0b 119 55.27 0.0c 0.0c 68.4a 11.6b 1.7c 120 55.56 0.0 0.0 15.4 1.3 12.4 121 55.81 0.0 0.0 2.0 8.0 4.4 122 56.32 0.0b 0.0b 21.1a 26.3a 0.0b 123 56.69 0.0b 3.3b 6.8b 2.0b 44.6a 124 56.82 12.0 21.5 14.7 29.4 27.7 125 57.60 0.0b 11.4a 26.6a 0.0b 17.4a 126 57.89 0.0b 18.6a 0.0b 33.2a 29.3a

52 2.7 Figures

1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4 Absorbance at 600 nm

0.2

0.0

0 2 4 6 8 10 12 14 Day

L346+CDM L350+CDM L856+CDM L887+CDM Control+CDM

Figure 2.2: Growth of P. freudenreichii strain P859 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains.

53 1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4 Absorbance at 600 nm

0.2

0.0

0 2 4 6 8 101214 Day

L346+CDM L350+CDM L856+CDM L887+CDM Control+CDM

Figure 2.3: Growth of P. freudenreichii strain P873 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains.

54 1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4 Absorbance at 600 nm

0.2

0.0

0 2 4 6 8 10 12 14 Day

L346+CDM L350+CDM L856+CDM L887+CDM Control+CDM

Figure 2.4: Growth of P. freudenreichii strain P812 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains.

55 1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4 Absorbance at 660 nm

0.2

0.0

0 2 4 6 8 10 12 14 Day

L346+CDM L350+CDM L856+CDM L887+CDM Control+CDM

Figure 2.5: Growth of P. freudenreichii strain P764M1 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains.

56 1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4 Absorbance at 600 nm

0.2

0.0

0 2 4 6 8 10 12 14 Day

L346+CDM L350+CDM L856+CDM L887+CDM Control+CDM

Figure 2.6: Growth of P. freudenreichii strain ATCC9614 in chemically defined medium (0.5% NaCl) plus slurry serum prepared with different L. helveticus strains.

600 mAU Height (mAU)

Retention time (Min) 65

Figure 2.7: The peptide profile of the slurry serum prepared with L. helveticus strain L350

600 mAU Height (mAU) Height

Retention time (Min) 65

Figure 2.8: The peptide profile of the slurry serum prepared with L. helveticus strain L346

600 mAU Height (mAu) (mAu) Height

65 Retention time (Min)

Figure 2.9: The peptide profile of the slurry serum prepared with L. helveticus strain L856

600 mAU Height (mAU) Height

65 Retention time (Min)

Figure 2.10: The peptide profile of the slurry serum prepared with L. helveticus strain L887

600 Height (mAU) Height

65 Retention time (Min)

Figure 2.11: The peptide profile of the control slurry serum lacking L. helveticus

62 2.8 References

Baer, A. (1995). Influence of casein proteolysis by starter bacteria, rennet and plasmin on the growth of Propionibacteria in Swiss-type cheese. Lait, 75, 391-400.

Baer, A., and Ryba, I. (1999). Interactions between propionic acid bacteria and thermophilic lactic acid bacteria. Lait, 79, 79-92.

Dalgaard, P., and Koutsoumanis, K. (2001). Comparison of maximum specific growth rates and lag times estimated from absorbance and viable count data by different mathematical models. Journal of Microbiological Methods, 43, 183-196.

Farkye, N.Y. and Fox, P.F. (1990). Observations on plasmin activity in cheese. Journal of Dairy Research. 57, 412-418.

Frohlich-Wyder, M. T., Bachmann, H.-P., and Casey, M. G. (2002). Interaction between propionibacteria and starter/ non-starter lactic acid bacteria in Swiss-type cheeses, Lait, 82, 1-15.

Gagnaire, V., Boutrou, R., and Leonil, J. (2001). How can the peptides produced from Emmental cheese give some insights on the structural features of the paracasein matrix?, Elsevier, 11, 449-454.

Glatz, B.A. and Anderson, K.T. (1988). Isolation and characterization of mutants of Propionibacterium strains. Journal of Dairy Science, 71, 1769-1776.

Langsrud, T., Sorhaug, T., and Vegarud, GE. (1995). Protein degradation and amino acid metabolism by propionibacteria. Lait, 75, 325-330. 63

Lemee, R., Gagnaire, V., and Maubois, J-L. (1998). Strain variability of the cell-free proteolytic activity of dairy propionibacteria toward β-casein peptides. Lait, 78, 227- 240.

McCarthy, R.J., and Courtney, P.D. 2004. Amino acid requirements of dairy Propionibacterium strains. Abstract 17A-10. Institute of Food Technologiests Annual Meeting, Las Vegas, NV.

Parker, J.A., and Moon, N.J. (1982). Interactions of Lactobacillus and Propionibacterium in mixed culture. Journal of Food Protection., 45, 326-330.

Perez Chaia, A., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

Piuri, M., Sanchez-Rivas, C., and Ruzal, S.M. (2003). Adaptation to high salt in Lactobacillus: role of peptides and proteolytic enzymes. Journal of Applied Microbiology, 95, 372-379.

Pivetean, PG., Condon, S., and Cogan, TM. (1995). Interactions between lactic and propionic acid bacteria. Lait, 75, 331-343.

Roberts, M., Wijesundera, C., Bruinenberg, P.G., and Limsowtin, G.K.Y. 1995. Development of an aseptic cheese curd slurry system for cheese ripening studies. Australian Journal of Dairy Technology. 50: 66-69.

Sahlstrom, S., Espinosa, C., Langsrud, T., Sorhaug, T. (1989). Cell Wall, membrane, and intracellular peptides activities of Propionibacterium shemanii. Journal of Dairy Science, 72, 342-350.

Shakeel-Ur-Rehman, Fox, P.F., McSweeney P.L.H, Madkor, S.A. and Farkye, N.Y. (2001). Alternatives to Pilot Plant Experiments in Cheese-Ripening Studies. International Journal of Dairy Technology, 54, 121-126.

Sousa, M.J., Ardo, Y., and McSweeney, P.L.H. (2001). Advances in the study of proteolysis during cheese ripening. International Dairy Journal, 11, 327-345.

Valence, F., Dentsch, S.M., Richoux, R., Gagauaire, V. and Lortal, S. (2000). Autolysis and related proteolysis in Swiss cheese from two Lactobacillus helveticus strains. Journal of Dairy Research, 67, 261-271.

64 Valence, F., Richoux, R., Thierry, A., Palva, A., and Lortal, S. (1998). Autolysis of Lactobacillus helueticus and Propionibacterium freudenreichii in Swiss cheese: first evidence by using species- specific lysis markers. Journal of Dairy Research, 65, 609-620.

Zwietering, M.H., Jongenburger, Il, Rombouts, F.M. and van’t Reit, K. 1990. Modeling of the bacterial growth curve. Applied and Environmental Microbiology. 56: 1875-1881.

65 3. PEPTIDE TRANSPORT IN TEN PROPIONIBACTERIUM FREUDENREICHII STRAINS

CHAPTER 3

PEPTIDE TRANSPORT IN TEN PROPIONIBACTERIUM

FREUDENREICHII STRAINS

3.1 Abstract

Propionibacterium strains have intracellular peptidases to breakdown peptides inside the cell, but little information exists concerning peptide transport in propionibacteria. It is unclear what type(s) of transporters are used to import the peptides and what their specificities are. The goal of this study is to evaluate specific peptide transport of ten P. freudenreichii strains by using commercially available peptides. The peptide concentration before and after incubation of each peptide with each strain was determined by high performance liquid chromatography to reveal the amount of peptide transported. All ten P. freudenreichii strains were able to transport the Ala-Ala di-peptide 30.89 to 100% peptide transport among the strains. Five strains were the highest and were not significant different

(p<0.01) among them. Percent peptide transport of Ala-Ala-Ala ranged from 0 to 100.

Three strains had no or very low transport of the Ala-Ala-Ala tri-peptide, but transported

Ala-Ala di-peptide well. Percent peptide transport of Glu-Ala ranged from 0 to 100.

Percent peptide transport of Glu-Glu ranged from 0 to 100. Only three strains transported 66 the Glu-Glu di-peptide. Most Propionibacterium strains transported the Ala-Ala di-peptide better than other peptides, and two of the ten strains transported all four peptides tested.

There are significant differences in peptide transport among P. freudenreichii strains (P<0.01).

Some peptides are more readily transported and degraded by propionibacteria than others, thus the peptide composition of cheese may impact Propionibacterium growth.

3.2 Introduction

Propionibacterium species are weakly proteolytic, but strongly peptidolytic. Their intracellular peptidases are especially active against proline-containing peptides (Sousa and others 2001). This suggests that though they are unable to generate peptides outside of the cell, they are well equipped to breakdown peptides inside the cell. Therefore, the peptides generated by L. helveticus during cheese ripening may promote the growth of P. freudenreichii by providing an amino acid source.

Conflicting reports in the literature make it unclear whether propionibacteria prefer free amino acids or peptides as their amino acid source. Baer (1995) measured the growth of

Propionibacterium and suggested that high concentration of amino acids inhibited their growth.

The author found that peptides stimulated the growth twice as much as amino acids.

However, the subsequent study of Baer and Ryba (1999) presented conflicting results concluding that the addition of the pure amino acids did not cause inhibition and that

67 acids to medium containing peptides inhibited the growth. These differences may be due to study design or to different strains tested.

Though it is unclear if propionibacteria prefer amino acids or peptides, the previous studies establish that these bacteria can utilize peptides. This is not surprising since their intracellular peptidase activity has been somewhat characterized (Sahlstrom and others 1989).

However, no reports exist concerning peptide transport in propionibacteria, so it is unclear what type(s) of transporters are used to import the peptides and what their specificities are.

It is possible that some peptides are more readily transported and degraded by propionibacteria than others, thus the peptide composition of cheese may impact

Propionibacterium growth.

Hypothesis:

Propionibacterium freudenreichii strains differ in their ability to transport peptides.

Objective:

Evaluate the peptide transport capabilities of P. freudenreichii strains in vitro.

3.3 Materials and Method

3.3.1 Chemicals

The peptides, Ala-Glu, Glu-Glu, Ala-Ala, and Ala-Ala-Ala, were purchased from

Sigma Chemical Company (Saint Louis, Missouri). Alanine-Alanine (Ala-Ala) di peptide occurs in primary structure of αs1, αs2, β, and κ-caseins 0, 0, 0, and 1 times, respectively.

Glutamic acid-Alanine (Glu-Ala) peptide occurs in primary structure of αs1, αs2, β, and κ- caseins 1, 0, 0, and 1 times, respectively. Glutamic Acid-Glutamic Acid (Glu-Glu) di peptide 68 occurs in primary structure of αs1, αs2, β, and κ-caseins 2, 8, 2, and 1 times, respectively

(Farrell and others 2004). Alanine-Alanine-Alanine (Ala-Ala-Ala) tri-peptide does not occur in primary structure of αs1, αs2, β, and κ-caseins. Standard concentrations of each peptide were prepared to 3 mM and 0.3 mM before dilution to 0.3, 0.25, 0.2, 0.15, 0.1, 0.05, 0.01,

0.005, 0.001 mM with buffer and buffer + glucose as shown in table 3.1. Buffer (50 mM potassium phosphate buffer, pH 6.5 in 5 mM magnesium sulfate) was prepared by adding 50 mM monopotassium phosphate to 50 mM dipotassium phosphate to pH 6.5 and then adding magnesium sulfate to 1.2324 g per liter to obtain 5 mM magnesium sulfate. Buffer + glucose was prepared by adding 0.99 g dextrose to 250 ml buffer.

Dansyl chloride solution was prepared by dissolving 5 mg dansyl chloride in 1 ml acetonitrile. Methylamine was prepared by dissolving 0.135 g methylamine hydrochloride in

10 ml deionized water. All chemicals were purchased from Sigma Chemical Company.

3.3.2 HPLC Analysis

An Agilent (Vydac 218TP54 (C18) PLC (Agilent Technologies, 363 Vintage Park

Drive, Foster City, CA 94404) was used to analyze peptides. The sample was filtered through a 0.2 µm membrane before injection into the HPLC. The method of Verheul et al.

(1995) was used. Solvent A was prepared by using 30 mM sodium phosphate in HPLC water and solvent B was prepared by using 55:45 acetonitrile in 30 mM sodium phosphate.

The solvents were filtered through a 0.22 µm membrane. A linear gradient from 55 to 75 % of solvent B was applied at 0.8 ml/min at 40 ºC. The column was an octadecyl (C18) silica reverse phase column with a selectivity for peptides up to 4-5000 daltons with 4.6 i.d 250

69 mm at 300 ºA pore size, 5 µm particle size and column volume of 4.2 ml ( BioBasic-18 ),

(Thermoelectron corporation, 355 River Oaks Parkway San Jose, CA 95134-1991 United

States). Peptides were detected by UV detection at 214 nm A standard curve relating peak area to peptide concentration was constructed for each peptide.

3.3.3 Bacterial strains and transport assay

Propionibacterium freudenreichii strains P572, P873, P568, ATCC9614, E15, P835,

P745M1, P107, P812, and P859 ,which were used in Chapters 1 and 4 and were used for the peptide transport assay in this chapter. Each P. freudenreichii strain was grown to stationary phase in Propionibacterium medium , a complex medium containing bacto tryptone (20 g L-1), yeast extract (5 g L-1), and glucose (5 g L-1) (Difco Laboratories, Detroit, MI). The stationary phase culture was inoculated into 20 ml Propionibacterium medium and grown to mid-log

phase (A600 0.5 to 0.8) at 28-30°C anaerobically. The culture was centrifuged at 6000 rpm for

10 minutes, the cell pellet was washed twice with 10 ml of sterile buffer, and then the cell pellet was resuspended in 6 ml buffer + glucose at 30 ºC. A blank without Propionibacterium cells was performed to confirm the peptide concentration before transport. The suspension was aliquotted into four microfuge tubes (0.9 ml each), then incubated for 5 minutes at 30

ºC to energized the cells. One of each of the four peptide solutions (0.1 ml of a 3 mM peptide solution) was added to each tube at 30 ºC. The sample was mixed and incubated at

30ºC for 30 minutes and then immediately centrifuged at max speed for 30 seconds. The supernatant was immediately filtered through a 0.45 µm membrane to remove all bacterial cells and recover only the extracellular peptides. The supernatant was dansylated with 5

70 mg/ml dansyl chloride by adding 55.5 µl to give a final concentration of 0.5 mg/ml. The sample was incubated for 2 hours at 37 ºC in water bath, then the reaction was stopped by adding 61.5 µl of 200 mM methylamine to give a final concentration of 20 mM. The sample was filtered through a 0.22 µm membrane and analyzed by HPLC as described above. The extracellular peptide concentrations before and after incubation with the P. freudenreichii strain were used to calculate % peptide transport into the cell.

3.3.4 Statistical Analysis

The means values from duplicate assays were compared by one-way analysis of variance (PROC GLM, SAS, Cary, NC) with Duncan’s test. Values with P < 0.01 were considered to be significantly different.

3.4 Results and Discussion

The peptide concentration before and after incubation of each peptide with each strain was determined by high performance liquid chromatography to reveal the amount of peptide transported. All P. freudenreichii strains transported the Ala-Ala di-peptide (figure 3.1) with Percent peptide transport ranging from 30.89 to 100. Five P. freudenreichii strains, P107,

P568, P745M1, P812, and P873 were the highest and were not significantly different

(P<0.01) among them. The percent peptide transport of Ala-Ala-Ala ranged from 0 to 100.

Three strains, P873, P745, and P835 had no or very low transport of the Ala-Ala-Ala tripeptide (figure 3.2), but transported the Ala-Ala di-peptide well. The percent peptide

71 transport of Glu-Ala ranged from 0 to 100. Propionibacterium freudenreichii strain P859 and E15 had no or very low Glu-Ala transport (figure 3.3). Percent peptide transport of Glu-Glu ranged from 0 to 100. Only three strains, P107, P568, and P812 transported the Glu-Glu di- peptide. (table 3.2) Most Propionibacterium strains transported the di-peptide Ala-Ala better than any other peptide tested. Thus, peptide transport is high for the hydrophobic/uncharged di-peptide (Ala-Ala) and low for the acidic di-peptide (Glu-Glu).

Propionibacterium freudenreichii strain P107 and P568 transported all four peptides tested. There are significant differences in peptide transport among P. freudenreichii strains (P<0.01). Some peptides are more readily transported and degraded by propionibacteria than others, thus the peptide composition of cheese may impact Propionibacterium growth.

To our knowledge, this is the first report of peptide transport in dairy

Propionibacterium, so it is unclear what type(s) of transporters are used to import the peptides and what their specificities are. The Ala-Ala neutral di-peptide was found to be transported better than Ala-Ala-Ala tri-peptide for most P. freudenreichii strains in this study, suggesting that smaller peptides are preferred. The addition of one or more negatively charged groups decreased the likelihood of the transport since seven strains did not transport the Glu-Glu di-peptide and the three that did transport this peptided did so at a low level. The peptide transport system of Lactococcus lactis has been studied extensively. The peptide transport system supports essential amino acids for their growth in milk. Helinck and others (2003) reported that negatively charged peptides impaired the oligopeptide transport of Lactococcus lactis. They investigated the competition of casein-derived peptides to be transported by

Lactococcus lactis. They found that casein-derived peptides competitively inhibit the transport of a reporter peptide by the oligopeptide transport system. The authors demonstrated that

72 the presence of a negatively charged residue impaired the oligopeptide transport function..

In contrast, the presence of uncharged (Ser, Thr, and Tyr) or a hydrophobic N-terminal residue (Met), or positive charged residue (Arg and Lys) did not effect the competition. The researchers concluded that the more increase in proteolytic activity the more peptide release from the casein resulted in the impairment of the oligopeptide transport function and then inhibited the growth rate.

Di- and tri-peptides were reported to be utilized at a higherrate than oligopeptides in

Lactococcus lactis strain MG1363 (Juillard and others 1998). Hagting and others (1994) studied the di- and tripeptide transport protein of Lactococcus lactis. They concluded that only a proton motive force-dependent carrier (DtpT) was the transport protein in Lactococcus lactis for hydrophilic di- and tripeptides. Some studies were reported that most milk peptides are neutral (Juillard and others 1998 and Helinck and others 2003). The research could not detect the amount of low molecular weight acidic peptide, whereas neutral and basic peptides were 96% and 4%. However, Juillard and others (1998) noted that the basic peptides with molecular masses ranging between 600 and 1100 Da were preferentially used by L. lactis and acidic peptides were poorly used. Hydrophobic peptides were utilized prior to hydrophilic. Garault and others (2002) studied the oligopeptide transport of S. thermophilus. They indicated that S. thermophilus transported oligopeptides containing from 3 to 23 amino acids with preference for hydrophobic oligopeptides. Juillard and others (1998) reported that oligopeptides containing 11 residues were consumed at the higher rate than tetrapeptides and suggested that the size of the peptide was not the only factor to determine the rate of utilization. They demonstrated that large, acidic, and phosphopeptides were not

73 utilized. Thus, the translocation of oligopeptides depended on not only the net charge but also its molecular weight of the peptides.

The oligopeptide transport system which transports peptides containing 2-5 amino acid residues is essential for growth of Lactococcus lactis. Juillard and others (1995) reported that one-fifth of oligopeptides generated from β-casein by proteinase of Lactococcus lactis can be transpoted and supplied their growth with amino acids. Kunji and others (1995) studied the peptides produced from β-casein by proteinase of Lactococcus lactis disappeared only in cells expressing an active oligopeptide transport system. Charbonnel and others (2003) studied the consumption of peptides during growth in chemically defined medium and the ability to transport the peptides. They reported the ability of L. lactis to transport peptides in a strain specific manner and provided evidence for variability in the ability of L. lactis to consume peptides. Strain L. lactis Wg2 grew poorly in the presence of the basic heptapeptide

ISQRYQK as the source of Gln or Ile, whereas it grew at maximum rate in the presence of the basic heptapeptide YPFPGPI as the source of Ile or TVYQHQK as the source of Gln.

They indicated that peptide utilization related to both the mass and the charge of the peptides, and also related to the specific strain of L. lactis. They emphasized that a periplasmic solute-binding protein (OppA) captured oligopeptides and initiated the transport process. Lactococcus lactis strain MG1363 was unable to grow using the tetrapeptide VGDE as the source of Valine whereas other strains utilized this peptide. This strain was unable to utilized acidic peptides with low molecular mass. They emphasized that the strain was atyptical compared with gram-negative bacteria such as E. coli and S. typhimurium.

Amino acid and peptide transport by Propionibacterium are poorly understood. Strain specific differences in peptide transport capability may affect the growth of particular P.

74 freudenreichii strains in cheese. Since the difference of L. helveticus strains affect the peptide profiles of cheese (Chapter 2), P. freudenreichii peptide transport may play a role in the interactions between these organisms in Swiss cheese. The differences in percent peptide transport of P. freudenreichii strains found in this study will be significant for culture suppliers and cheese manufacturers to select Propionibacterium strains with matching L. helveticus strain.

Future work recommendation is further analysis of peptide transport using peptides isolated from Swiss cheese prior to warm room ripening and conducting transport assays under Swiss cheese like conditions will give a more realistic assessment of the impact of these differences in cheese.

75 3.5 Tables

Standard 3mM peptide 0.3 mM peptide Buffer + Buffer solution solution solution glucose

Blank 0 0 0.9 ml 100 µl

0.001 mM 0 3.3 µl 0.9 ml 96.7 µl

0.005 mM 0 16.7 µl 0.9 ml 83.3 µl

0.01 mM 0 33.3 µl 0.9 ml 66.7 µl

0.05 mM 16.7 µl 0 0.9 ml 83.3 µl

0.1 mM 33.3 µl 0 0.9 ml 66.7 µl

0.15 mM 50 µl 0 0.9 ml 50 µl

0.2 mM 66.7 µl 0 0.9 ml 33.3 µl

0.25 mM 83.3 µl 0 0.9 ml 16.7 µl

0.3 mM 100 µl 0 0.9 ml 0

Table 3.1: Preparation of peptides standard 0.3, 0.25, 0.2, 0.15, 0.1, 0.05, 0.01, 0.005, 0.001 mM with 3 mM and 0.3 mM peptide solution

Strain Ala-Ala Ala-Ala-Ala Glu-Ala Glu-Glu

E15 46.073cd 33.065b 0.739g 0c

P107 100a 100a 100a 100a

P568 100a 100a 100a 100a

P572 74.72b 29.137bc 78.348b 0c

P745M1 93.642a 3.473ef 46.468d 0c

P812 97.345a 29.016bc 57.918c 56.508b

P835 51.122c 4.923ef 18.306f 0c

P859 30.89d 20.26cd 0c 0c

P873 97.008a 0c 59.534c 0c

ATCC9614 72.625b 10.794de 32.757e 0c

Means of the same letter in the same column are not significant different (p<0.01)

Table 3.2: Percent of total peptide transported by ten P. freudenreichii strains from a 3 mM solution

77 3.6 Figures

120

100

40 Percent peptide transport

0 E15 P107 P568 P572 P745M1 P812 P835 P859 P873 P9614 P. freudenreichii strain

Ala-Ala

Figure 3.1: Percent peptide transport of Ala-Ala of ten Propionibacterium freudenreichii strains

78 120

100

40 Percent peptide transport

0 E15 P107 P568 P572 P745M1 P812 P835 P859 P873 P9614 P. freudenreichii strain

Ala-Ala-Ala

Figure 3.2: Percent peptide transport of Ala-Ala-Ala of ten Propionibacterium freudenreichii strains

79 120

100

40 Percent peptide transport

0 E15 P107 P568 P572 P745M1 P812 P835 P859 P873 P9614 P. freudenreichii strain

Glu-Ala

Figure 3.3: Percent peptide transport of Glu-Ala of ten Propionibacterium freudenreichii strains

80 120

100

40 Percent peptide transport

0 E15 P107 P568 P572 P745M1 P812 P835 P859 P873 P9614 P. freudenreichii strain

Glu-Glu

Figure 3.4: Percent peptide transport of Glu-Glu of ten Propionibacterium freudenreichii strains

81 120

100

40 Percent Peptide Transport 20

0 E15 P107 P568 P572 P745M1 P812 P835 P859 P873 P9614 P. freudenreichii strains

Ala-Ala Ala-Ala-Ala Glu-Ala Glu-Glu

Figure 3.5: Percent peptide transport of Ala-Ala, Ala-Ala-Ala, Glu-Ala, and Glu-Glu of ten Propionibacterium freudenreichii strains

82 3.7 References

Farrell, H.M., Jimenez-Flores, R., Bleck, G.T., Brown, E.M., Butler, J.E., Creamer, L.K., Hicks, C.L., Hollar, C.M., Ng-Kwai-Hang, K.F., and Swaisgood, H.E. (2004). Nomenclature of the proteins of cows’ milk-sixth revision. Journal of Dairy Science, 87, 1641-1674.

Juillard, V., Guillot, A, Bars, D., and Gripon, J-C. (1998). Specificity of milk peptide utilization by Lactococcus lactis. Applied and Environmental Microbiology, 64, 1230-1236.

Strickland, M., Johnson, M. E., and Broadbent, J. R. (2001). Qualitative and quantitative analysis of proteins and peptides in milk products by capillary electrophoresis, Electrophoresis, 22, 1510-1517.

Tapuhi, Y., Schmidt, D.E., Lindner, W., and Karger, B.L. 1981. Dansylation of amino acids for high-performance liquid chromatography analysis. Analytical Biochemistry. 115, 123- 129.

83 4. PEPTIDASE ACTIVITIES OF THIRTY-FOUR DAIRY PROPIONIBACTERIUM FREUDENREICHII STRAINS

CHAPTER 4

PEPTIDASE ACTIVITIES OF THIRTY-FOUR DAIRY

PROPIONIBACTERIUM FREUDENREICHII STRAINS

4.1 Abstract

Dairy Propionibacterium strains have proteolytic activities capable of hydrolyzing casein-derived peptides to small peptides and amino acids which contribute to cell growth and cheese flavor. The diversity of peptidase activities among dairy Propionibacterium strains is not well characterized. The objective was to evaluate the intracellular peptidase activities of dairy Propionibacterium strains. Thirty-four different P. freudenreichii strains from commercial starter cultures and cheeses were selected for analysis. Cells were grown to mid-exponential phase, and cell-free extracts were analyzed using the following ρ-nitroanilide labeled peptides as substrates: pro-ρNA, lys-ρNA, ala-pro-ρNA, and glu-ρNA. Using each chromagenic substrate, the liberation of ρ-nitroanilide was monitored spectrophotometrically at 410 nm in each cell free extract in 20 mM Tris-HCl buffer, pH 7.0, at 30°C. The Bio-Rad protein assay established the protein concentration in each cell-free extract which was used to calculate the enzyme specific activity. Significant differences in enzyme activity (P<0.01) were found

84 among the Propionibacterium strains. Proline iminopeptidase (PepI) activity on the pro-ρNA substrate ranged from 0 to 158 µmol/min/mg protein. Six strains had no or very low PepI activity. Aminopeptidase activity on the lys-ρNA and the glu-ρNA substrates ranged from 0 to 79.7 and from 0 to 3.9 µmol/min/mg protein, respectively. X-prolyl dipeptidyl aminopeptidase (PepX) activity on the ala-pro-ρNA substrate ranged from 0 to73.3

4.2 Introduction

Proteolysis is an important process in cheese ripening to produce cheese flavor and the overall quality of cheese (Law and Haandrikman 1997). Proteinase and peptidase are enzymes involve in proteolysis. In many cheese-associated bacteria, there are a cell envelope-associated proteinase and several cytoplasmic peptidase enzymes. The proteinase, which usually acts extracellularly, degrades intermediate size peptides produced by chymosin or plasmin. These peptides may then be transported inside the cell where intracellular peptidase enzymes further degrade the peptide into smaller peptides and amino acids. Both starter and nonstarter bacteria may have these activities. In Swiss cheese, the Lactobacillus helveticus culture is primarily responsible for the proteolysis of casein. However, the proteolytic activities of the

85 Propionibacterium freudenreichii culture may also be an important factor in the growth of this culture and may also contribute to flavor.

Many different peptidases have been identified in dairy starter cultures due to their important role in cheese flavor and texture generation during ripening. Peptidases differ in their specificities in terms of amino acid chain length and composition. Oligopeptidase and tripeptidase vary in the chain length of their specificity. The aminopeptidases, PepN and

PepC, also have broad specificity, cleaving N-terminal amino acids from peptides of varying lengths (McSweeney 2004). PepA is specific for N-terminal glu or asp residues, and PepL releases pyroglutamic acid from N-terminal of peptides. PepX, PepI, PepP, PepR and PepQ release peptides containing proline from the N-terminus (Christensen 1999). PepX releases

Xaa-Pro dipeptide from substrates containing from three to seven amino acid residues. PepX has high activity on Xaa-Pro dipeptides when the N-terminal residues are uncharged hydrophobic (Ala, Ile, Val, Gly) or basic amino acids (Arg, His, Lys) including aromatic amino acids residues (Phe, Try). Pro-Pro-Xaa was also found to be a substrate of PepX. PepX activity is strain specific in L. helveticus and shows higher activity in milk than on peptides in

laboratory growth medium (William 1998). PepP removes X-pro-pro-(X)n sequence, PepR cleaves pro-X dipeptide and PepQ cleaves X-pro dipeptides. PepI releases free proline from the N-terminal of di- and tripeptides. Proline is an amino acid found in large quantities in

Swiss cheese and may contribute to the characteristic flavor of this cheese (Langsrud and others 1995; Sousa and others 2001).

The peptidase enzymes of Propionibacterium strains have been studied. Langsrud and others (1995) reviewed peptidase enzyme in 6 different strains and species. Proteinase and peptidase activities were strain specific, especially serine proteinase associated with cell wall

86 and cell membrane. The research pointed out that Propionibacterium contained proteinases and peptidases. The proteinase was cell wall associated and the peptidase activity was intracellular. Several reports agree that Propionibacterium contained mostly peptidase activity, such as aminopeptidases, proline imidopeptidase, X-prolyl-dipeptidyl-aminopeptidase

(Panon 1990; Perez Chaia and others 1990; Chaia and others 1990; Ostlie 1995). Some reported that X-prolyl-dipeptidyl-aminopeptidase was the main intercellular peptidase and claim that Propionibacterium contains endopeptidase, and proline imidopeptidase. However, diversity in the peptidase activity of the P. freudenreichii strains has been observed, but only a few of strains have been investigated in previous study. Therefore, the objective of this study is to evaluate the peptidase activity of PepI, aminopeptidase, and PepX of 34 P. freudenreichii strains.

Hypothesis:

Propionibacterium freudenreichii strains differ in their intracellular peptidase activities.

Objective:

Evaluate the peptidase activities P. freudenreichii strains in vitro.

4.3 Materials and Methods

Thirty-four different Propionibacterium strains from commercial starter cultures and cheeses were selected for analysis; D2, P745M1, P745M2, D47, P318, P 835, C44, P196,

F46, P134, F1, F2, E16, F3, F22, F33, E18, F47, P568, E15, E36, P728M1, ATCC9614,

P873, P812, P764M1, G1, A50, P891,G13, G50, P859, P572, P107. Cells were grown to mid-exponential phase with 100 ml Propionibacterium broth, a complex medium containing

87 bacto tryptone (20 g L-1), yeast extract (5 g L-1), and glucose (5 g L-1) (Difco Laboratories,

Detroit, MI). The cells were then collected by centrifugation for 10 minutes at 6000 rpm.

The cells were washed washing twice with 10 ml sterile 20 mM Tris-HCl buffer, pH 7.0, and then resuspended with 2 ml of buffer before pipetting 0.4 ml each into a screw cap tube containing a small amount of 0.1 mm glass beads. Cells were pulverized in a Bead-Beater-8

(BioSpec Products, INC, Bartlesville, OK) for 30 seconds, three times, with incubation on ice between each session to prevent excessive heating. The supernatants were collected by centrifugation for five minutes at 13,000 rpm. 200 µl of the supernatant were used for a peptidase assay, and 50 µl were used to determine the protein concentration in the cell-free extract. Cell-free extracts were kept on ice until immediately before the enzyme assay.

The cell-free extracts were analyzed using the following ρ-nitroanilide labeled peptides as substrates: pro-ρNA, lys-ρNA, ala-pro-ρNA, and glu-ρNA (Bachem Califonia

Inc. 3132 Kashiwa Street Torrance, CA 90505 The 2.5 mM solutions of each substrate were prepared using 20 mM Tris-HCl buffer, pH 7.0, as the solvent and diluted to 0.25 mM at

30°C. The cell-free extracts were pre-warmed to 30°C for 1 minute in water bath before beginning the assay. Using 800 µl of 0.25 mM chromagenic substrate and 200 µl of cell-free extract, the liberation of ρ-nitroanilide was monitored spectrophotometrically at 410 nm in each cell-free extract in 20 mM Tris-HCl buffer, pH 7.0, at 30°C. For the maximum activity of each sample, the region of maximum absorbance change per minute was monitored. The blank containing the 800 µl buffer and 200 µl for each substrate was used for adjustment to the zero absorbance. The Bio-Rad protein assay established the protein concentration in each cell-free extract which was used to calculate the enzyme specific activity. A standard

88 curve was prepared using bovine serum albumin (BSA) at the following concentrations: 0,

0.1, 0.2, 0.3, 0.4, and 0.5 mg/ml in the Tris-HCL buffer. Bio-Rad Dye reagent was diluted 1 part dye to 4 parts deionized water and filtered through Whatman filter paper number 1 before measuring. Ten µl of each cell-free extract and each standard and blank (buffer) were mixed in a micro titer plate well with 200 µl of the dye reagent for 5 minutes and measured at 600 nm. The blank was subtracted from all values. A standard curve using the average of the duplicate standard samples was used to calculate protein concentration. Using the equation of the line, the protein concentrations of the samples were determined. The molar extinction coefficient of 8800 M-1cm-1 was used to calculate the specific enzyme activity as

µmol ρ-NA/min/mg protein.

The means of the specific activities from duplicate independent peptidase assays were compared by one-way analysis of variance (PROC GLM, SAS) with Tukey’s test.

Values with P < 0.01 were considered to be significantly different.

4.4 Results

4.4.1 Proline iminopeptidase activity of Propionibacterium strains

Significant differences in enzyme activity (P<0.01) were found among the

Propionibacterium strains. Proline iminopeptidase (PepI) activity on the pro-ρNA substrate ranged from 0 to158 µmol/min/mg protein (Table 4.1, Fig. 4.1). Propionibacterium freudenreichii strain P745M1 had the highest activity. Six strains, E16, F3, P568, ATCC9614,

89 P859, and P107, had no or very low PepI activity. Their activities ranged from 0 to 2.966

µmol/min/mg protein.

4.4.2 X-prolyl dipeptidyl aminopeptidase activity of Propionibacterium strains

X-prolyl dipeptidyl aminopeptidase (PepX) activity on the ala-pro-ρNA substrate ranged from 0 to 73.3 µmol/min/mg protein (Table 4.1, Fig. 4.2). There were significant differences in Pep X activities among Propionibacterium strains. Propionibacterium freudenreichii strain P568 had the highest activity (73.3 µmol/min/mg protein). This peptidase releases

Xaa-Pro dipeptide that may be hydrolyzed by aminopeptidase or PepI to free proline.

However, strain P568 was found to be low in PepI activity, but high in aminopeptidase on lys-ρNA substrate.

4.4.3 Aminopeptidase activity of Propionibacterium strains

Aminopeptidase activity on the lys-ρNA and the glu-ρNA substrates ranged from 0 to 79.7 and from 0 to 3.9 µmol/min/mg protein, respectively (Table 4.1, Figures 4.3 and

4.4). There were significant differences (P<0.01) among the strains in aminopeptidase activity on both the lys-ρNA and the glu-ρNA substrates. Propionibacterium freudenreichii strains P568, E16 and P859 had the highest activities on the lys-ρNA substrate. The activities were 79.7, 73.6 and 68.2 µmol/min/mg protein, respectively. Statistical analysis

90 showed no significant different between these three values. The aminopeptidase activity on the glu-ρNA substrate ranged from 0 to 3.85 µmol/min/mg protein. Propionibacterium freudenreichii strains ATCC9614 and P745M1 were the highest with activities of 3.9 and 2.8

µmol/min/mg protein.

4.5 Discussion

The intracellular peptidase activities of dairy Propionibacterium strains varied in specificity and activity levels. These differences rate may contribute to flavor and growth variations observed in cheeses manufactured with different Propionibacterium strains. Most strains, 27 of 34 strains, were higher in PepI activity than other activities. One strain, P568, with no PepI activity had the highest PepX and aminopeptidase activities (Table 4.1). The

Propionibacterium strain P745M1 was high in PepI activity (Fig. 4.1), but was in the middle of the range of PepX activity (figure 4.2) and aminopeptidase activity on Lys-ρNA substrate

(figure 4.3) among Propionibacterium strains and low in aminopeptidase activity on Glu-ρNA substrate (figure 4.4). The same as Propionibacterium strains A50 was high in PepI activity and in the middle of PepX activity and aminopeptidase on Lys-ρNA substrate, but low in aminopeptidase activity on Glu-ρNA substrate. On the other hand, the Propionibacterium strain P568 was extremely high in PepX activity and aminopeptidase on Lys-ρNA substrate, but no PepI activity and low activity on aminopeptidase on Glu-ρNA substrate. In contrast with the Propionibacterium strain E16 was high only aminopeptidase on Lys-ρNA substrate

91 (figure 4.3), but no PepI and PepX activity and very low in aminopeptidase activity on Glu-

ρNA substrate.

Fernandy-Espla and Fox (1997), Langsrud and others (1995), Quelen and others

(1995), Sausa and others (2001), and Tobiassen and others (1997) reported that

Propionibacterium possess PepI, PepX and aminopeptidase activities. Fernandy-Espla and Fox

(1996) studied the enzymes in strain P. freudenreichii subsp. shermanii strain NCDO 853 and found prolidase, iminopeptidase, and X-prolyl dipeptidyl aminopeptidase activities. The iminopeptidase from P. freudenreichii subsp. shermanii strain NCDO 853 was proline specific, hydrolyzing Pro-ρNA, dipeptidases containing praline. Langsrud and others (1995) have reviewed protein degradation by Propionibacterium. They reported peptidase enzyme in 6 different strains and species. Seven to 8 strain specific peptidase bands were found. Serine proteinase activity was associated with the cell wall and membrane. Stepaniak (2000) showed that PepI had 100% activity on pro-ρNA substrate when compared with other substrates.

Chaia and others (1990) found that the peptidase were more active on proline-ρ-nitroanilide than on leucine-ρNA. Panon (1990) purified proline iminopeptidase from P. freudenreichii

13673 with molecular mass of 61 KDa and determined its activity at 40°C, pH 8.0. They found that proline iminopeptidase cleaved dipeptide and peptide containing proline residue.

The PepX activity on L-Ala-L-Pro peptide substrate was reported to be 86 nmol/ml/min in P. freudenreichii strain ISU P-59 (Østile and others 1995). This is within the range observed in the present study. PepX and PepI activities were reported in P. freudenreichii strain INF-α and strain ATCC 9614 that were low activity, 2.49 and 3.76

µmol/min on Gly-Pro-ρNA substrate, respectively and the PepI on Pro-ρNA substrate was

92 14.47 and 10.04 µmol/min, respectively (Tobiassen and others 1997). This study support the previous study that the PepI activity in the same substrate for strain ATCC 9614 was found no activity and low activity for PepX in the different substrate (Ala-Pro-ρNA substrate). The different in PepI activity is possibly due to the different in method used.

Aminopeptidase activity on Leu-ρNA substrate was low, 4.31 and 9.36 µmol/min for the strain INF- α and strain ATCC 961, respectively. Aminopeptidase was also found to be low in the present study for the strain ATCC9614 (on the Lys--ρNA substrate and Glu--ρNA substrate). The aminopeptidase activity of Lactobacillus species was reported to be the highest

(333 to 666 U mg-1) on Lys-ρNA substrate than Glu-ρNA substrate (77 to 315 U mg-1) and

Pro-ρNA substrate (89 to 142 U mg-1), the same substrates used in this study. These results showed that the Lactobacillus species was high in aminopeptidase activity than other activities and different from P. freudenreichii in this study which were the highest in Pep I activity. The

PepI activity was found stronger than other activity in the study of Quelen and others

(1995).

Several researchers (Langsrud and others 1995, Sousa and others 2001) have suggested that proline is an important contributor to ‘nutty’ flavor in Swiss cheese. Since most strains of P. freudenreichii such as P745M1, A50, P745M2, and D2 found in this study have high in PepI activity which cleaves N-terminal proline and P. freudenreichii strains (P568) have high in PepX activity cleave N-terminal Xaa-Pro dipeptide, they produce proline and contribute to the sweet and nutty flavor in Swiss cheese. The rate of proline production was shown to relate of growth and autolysis of Propionibacterium (Langsrud and others 1978), so it affects the eye in Swiss cheese.

93 This study describes an extensive survey of P. freudenreichii peptidase activity.

Diversity in the peptidase activity of the P. freudenreichii strains has been previously reported, but only a few strains have been investigated in the previous studies. Therefore, this study provides the new information of peptidase profile of 34 P. freudenreichii strains and may contribute to cheese manufacture by providing information that will help in culture selection. The differences in intracellular peptidase activities of dairy Propionibacterium strains may contribute to flavor and growth rate variations observed in cheeses manufactured with different Propionibacterium strains. Future work recommendation is evaluation of peptidase activity under cheese-like conditions and using peptides isolated from cheese will give a more realistic assessment of the impact of these differences in cheese.

94 4.6 Table

Sample Pro-ρNA Lys-ρNA Ala-pro-ρNA Glu-ρNA 1 D2 100.013c 13.748defghi 8.253fgh 0.000h 2 P745M1 158.781a 20.463d 20.459bc 2.821ab 3 P745M2 102.894bc 13.587defghij 14.925de 0.000h 4 D47 60.199e 7.982efghijk 9.005fgh 1.033efgh 5 P318 25.319jk 2.188hijk 3.202jki 0.000h 6 P835 25.124jk 1.791jk 0.000k 0.000h 7 C44 50.611efg 10.477defghijk 12.035ef 1.432cdefg 8 P196 51.967efg 13.865defgh 23.439b 2.764ab 9 F46 23.315jk 2.653ghijk 0.000k 0.000h 10 P134 40.941hig 7.091efghijk 3.515jki 0.468gh 11 F1 45.533fg 12.623defghij 0.000k 0.000h 12 F2 18.809klm 2.546ghijk 0.000k 0.000h 13 E16 0.000n 73.571ab 0.000k 1.767bcdef 14 F3 2.966n 63.081bc 0.000k 1.891bcde 15 F22 19.152klm 4.913fghijk 0.000k 0.000h 16 F33 30.359jk 1.947ijk 0.000k 0.000h 17 E18 18.639klm 2.646ghijk 0.000k 0.000h 18 F47 18.198klm 2.503ghijk 0.000k 0.000h 19 P568 0.000n 79.731a 73.285a 2.303bcd 20 E15 21.417jkl 0.000k 0.000k 0.000h

Continued

Table 4.1: Peptidase specificity activity (µmol/min/mg protein) of 34 dairy Propionibacterium strains

95 Table 4.1 Continued

Sample Pro-ρNA Lys-ρNA Ala-pro-ρNA Glu-ρNA 21 E36 24.984jk 3.549ghijk 0.000k 0.000h 22 P728M1 81.567d 14.134defg 6.140ghi 0.000h 23 ATCC9614 0.000n 5.169fghijk 6.257ghi 3.850a 24 P873 9.330lmn 2.156hijk 0.000k 0.000h 25 P812 51.901efg 9.929defghijk 5.436hi 0.000h 26 P764M1 9.612lmn 15.666def 0.675jk 1.096efgh 27 G1 32.831hij 9.792defghijk 3.031jki 0.453gh 28 A50 112.668b 17.940de 18.160cd 0.674fgh 29 P891 8.196mn 3.811ghijk 0.486k 0.380gh 30 G13 54.242ef 8.436efghijk 6.948ghi 0.000h 31 G50 95.284c 10.500defghijk 5.292hij 1.383defg 32 P859 2.153n 68.182abc 0.383k 0.864efgh 33 P572 44.598fgh 7.511efghijk 10.665efg 0.000h 34 P107 0.000n 58.166c 0.000k 2.624bc Means with the same letter for each enzyme substrate are not significant different (P<0.01).

96 4.7 Figures

200

180

160

140

120

100

Specific activity (µmol/min/mg protein) 20

0 F1 F2 F3 D2 G1 F46 F22 F33 F47 A50 E16 E18 E15 E36 D47 C44 G13 G50 P891 P859 P572 P107 P318 P835 P196 P134 P568 P873 P812 P9614 P745M1 P745M2 P764M1 P728M1 P. freudenreichii strains

Figure 4.1: Proline iminopeptidase (PepI) specific activity (µmol/min/mg protein) of thirty-four dairy Propionibacterium strains. Error bars represent plus one standard deviation, n=2.

100

20 Specific activity (µmol/min/mg protein)

0 F1 F2 F3 D2 G1 F46 F22 F33 F47 E16 E18 E15 E36 A50 D47 C44 G13 G50 P318 P835 P196 P134 P568 P891 P859 P572 P107 P873 P812 P9614 P745M1 P745M2 P728M1 P764M1 P. freudenreichii strains

Figure 4.2: X-prolyldipeptidyl aminopeptidase (PepX) specific activity (µmol/min/mg protein) of thirty-four dairy Propionibacterium strains. Error bars represent plus one standard deviation, n=2.

120

100

20 Specific activity (µmol/min/mg protein)

0 F1 F2 F3 D2 G1 F46 F22 F33 F47 E16 E18 E15 E36 A50 D47 C44 G13 G50 P318 P835 P196 P134 P568 P891 P859 P572 P107 P873 P812 P9614 P745M1 P745M2 P728M1 P764M1 P. freudenreichii strain

Figure 4.3: Aminopeptidase specific activity (µmol/min/mg protein) on lys-ρNA of thirty-four dairy Propionibacterium strains. Error bars represent plus one standard deviation, n=2.

1 Specific activity(µmol/min/mg protein)

0 F1 F2 F3 D2 G1 F46 F22 F33 F47 E15 E36 A50 E16 E18 D47 C44 G13 G50 P318 P835 P196 P134 P568 P873 P812 P891 P859 P572 P107 P9614 P728M1 P764M1 P745M1 P745M2 P.freudenreichii strain

Figure 4.4: Aminopeptidase specific activity (µmol/min/mg protein) on glu-ρNA of thirty-four dairy Propionibacterium strains. Error bars represent plus one standard deviation, n=2.

100 4.8 References

Chaia, A.P., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

Christensen, J. E., Dudley, E.G., Pederson, J. A., and Steele, J.L. (1999). Peptidases and amino acid catabolism in lactic acid bacteria. Antonie van Leeuwenhoek, 76, 217-246.

Fernandez-Espla, M. D., and Fox, P. F. (1997). Purification and characterization of X-prolyl dipeptidyl aminopeptidase from Propionibacterium shermanii NCDO 853, International Dairy Journal, 7, 23-29

Law, J. and Haandrikman, A. (1997). Proteolytic enzymes of lactic acid bacteria. International Dairy Journal, 7, 1-11.

Lungsrud, T., Reinbold, G.W., and Hammond, E.G. (1978). Free proline production by strains of Propionibacteria. Journal of Dairy Science, 61, 303-308.

Langsrud, T., Sorhaug, T., and Vegarud, GE. (1995). Protein degradation and amino acid metabolism by propionibacteria. Lait, 75, 325-330.

Malone, A.S., Wick, C., Shellhammer, T.H., and Courtney, P.D. (2003). High pressure effects on proteolytic and glycolytic enzymes involved in cheese manufacturing. Journal of Dairy Science, 86, 1139-1146.

McSweeney, P. L. (2004). Biochemistry of cheese ripening. International Journal of Dairy Technology, 57, 127-144.

Ostile, H., Floberghagen, V., Reinbold, G., Hammond, E.G., Vegarud, G., and Langsrud, T. (1995). Autolysis of dairy propionibacteria: growth studies, peptidase activities, and proline production. Journal of Dairy Science, 78, 1224-1237.

Panon, G. (1990). Purification and characterization of a proline iminopeptidase from Propionibacterium shermanii 13673, Le Lait, 70, 439-452.

Perez Chaia, A., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

Quelen, L., Dupuis, C., and Boyaval, P. (1995). Proline-specific activities of Propionibacterium freudenreichii subsp. shermanii. Journal of Dairy Research, 62, 661-666.

Stepaniak, L. (2000). Isolation and characterization of proline iminopeptidase from Propionibacterium frendenreichii ATCC 9614. Nahrung, 44, 102-106.

101 Sousa, M.J., Ardo, Y., and McSweeney, P.L.H. (2001). Advances in the study of proteolysis during cheese ripening. International Dairy Journal, 11, 327-345.

Tobiassen, R. O., Steoaniak, L., and Sorhaug, T. (1997). Screening for difference in the proteolytic systems of Lactococcus, Lactobacillus and Propionibacterium. Z Lebensm Unters Forsch A, 204, 273-278.

102

CHAPTER 5

5. CONCLUSION CONCLUSION

Slurry sera produced with different L. helveticus strains differed in their peptide profiles.

There were significant differences (P<0.05) in peptide content between the control slurry and those prepared with L. helveticus strains. The peptide contents of slurries prepared with L. helveticus strains L350 and L856 produce more peptides that were significant different from the control than L. helveticus strains L346 and L887. Most peaks (97 of 126) were found both in the control serum and in the sera prepared with L. helveticus strains, but the concentrations of the peptides decreased or increased in some cases when L. helveticus was included.

Approximately one third of the peaks decreased in concentration in the presence of L. helveticus. Differences in peptide profiles may contribute to the observed differences in P. freudenreichii growth in the slurry. Different L. helveticus strains differentially affect the growth of

P. freudenreichii in a cheese-like model system.

A slurry serum model was developed and has potential for rapid screening of P. freudenreichii and L. helveticus pairings for Swiss cheese manufacturing. The peptides produced and the stimulation effect observed are strain-specific, thus only certain L. helveticus + P. freudenreichii pairs result in sufficient P. freudenreichii growth. In addition, the slurry serum model

103 has potential for rapid screening of L. helveticus strains which yield different peptide profiles.

Future work recommendation is to validate the model by making Swiss cheese with selected strain pairings and further analysis of peptide profiles may reveal the identities of specific peptides utilized by P. freudenreichii strains.

Amino acid and peptide transport by Propionibacterium are poorly understood. This study is the first report for peptide transport in P. freudenreichii (10 strains). Strain specific differences in peptide transport capability may affect the growth of particular P. freudenreichii strains in cheese. Since different L. helveticus strains affect the peptide profiles of cheese

(Chapter 2), P. freudenreichii peptide transport specificity may play a role in the interactions between these organisms in Swiss cheese. The differences in percent peptide transport of P. freudenreichii strains found in this study will be significant for culture suppliers and cheese manufacturers to select Propionibacterium strains with matching L. helveticus strain. Future work recommendation is further analysis of peptide transport using peptides isolated from

Swiss cheese prior to warm room ripening and conducting transport assays under Swiss cheese like conditions will give a more realistic assessment of the impact of these differences in cheese.

This study showed the most extensive survey of P. freudenreichii activity. Diversity in the peptidase activity of the P. freudenreichii strains has been observed, but only a few of strains has been investigated in previous study. Therefore, this study provides the new information of peptidase profile of 34 strains P. freudenreichii and contributes to cheese manufacture to combine Lactobacillus strains and Propionibacterium strain to work well together to obtain flavor in Swiss cheese. The differences in intracellular peptidase activities of dairy

Propionibacterium strains may contribute to flavor and growth rate variations observed in

104 cheeses manufactured with different Propionibacterium strains. Future work recommendation is evaluation of peptidase activity under cheese-like conditions and using peptides isolated from cheese will give a more realistic assessment of the impact of these differences in cheese.

The different peptide transport and peptidase profiles observed in different

Propionibacterium strains will likely affect their performance with different L. helveticus strains which generate different peptides during cheese ripening. This study describes several factors that may contribute to interactions between Swiss cheese starter cultures causing some culture combinations to perform better than others.

105 List of References

LIST OF REFERENCES

Addeo, F., Chianese, L., Salzano, A., Sacchi, R., Cappuccio, U., Ferranti, P., and Malorni, A. (1992). Characterization of the 12% trichloroacetic acid soluble oligopeptides of Parmigiano-Reggiauo cheese. Journal of Dairy Research, 59, 401-411.

Addeo, F., Garro, G., Intorcina, N., Pellegrino, L., Resmini, P., and Chianese, L. (1995). Gel electrophoresis and immunoblotting for the detection of casein proteolysis immunoblotting for the detection of casein proteolysis in cheese. Journal of Dairy Research, 62, 297-309.

Baer, A. (1995). Influence of casein proteolysis by starter bacteria, rennet and plasmin on the growth of Propionibacteria in Swiss-type cheese. Lait, 75, 391-400.

Baer, A., and Ryba, I. (1999). Interactions between propionic acid bacteria and thermophilic lactic acid bacteria. Lait, 79, 79-92.

Beresford, T. P., Fitzsimons, N. A., Brennan, N.L., and Cogan, T.M. (2001). Resent advances in cheese microbiology. International Dairy Journal, 11, 259 - 274.

Califano, A.N., and Bevilacqua, A.E. (2000). Multivariate analysis of the organic acids content of Gouda type cheese during ripening. Journal of Food Composition and Analysis, 13, 949-960.

Chaia, A.P., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

106 Charbonnel, P., Lamarques, M., Piard, J.C., Gilbert, C., Juillard, V., and Atlan, D. (2003). Diversity of oligopeptide transport specificity in Lactococcus lactis species. The Journal of Biological Chemistry, 278, 14832-14840.

Christensen, J. E., Dudley, E.G., Pederson, J. A., and Steele, J.L. (1999). Peptidases and amino acid catabolism in lactic acid bacteria. Antonie van Leeuwenhoek, 76, 217-246.

Dalgaard, P., and Koutsoumanis, K. 2001. Comparison of maximum specific growth rates and lag times estimated from absorbance and viable count data by different mathematical models. Journal of Microbiological Methods, 43, 183-196.

Farkye, N.Y. and Fox P.F. (1990) Observations on plasmin activity in cheese. Journal of Diary Research, 57, 412-418

Frohlich-Wyder, M. T., Bachmann, H.-P., and Casey, M. G. (2002). Interaction between propionibacteria and starter/ non-starter lactic acid bacteria in Swiss-type cheeses, Lait, 82, 1-15.

Gagnaire, V., Boutrou, R., and Leonil, J. (2001). How can the peptides produced from Emmental cheese give some insights on the structural features of the paracasein matrix? International Dairy Journal, 11, 449-454.

107 Gagnaire, V., Molle, D., Herrouin, M., and Leonil, J. (2001). Peptides identified during Emmental ripening: Origin and proteolytic systems involved, Journal of Agricultural Food Chemistry, 49, 4402-4413.

Glatz, B.A. and Anderson, K.T. (1988). Isolation and characterization of mutants of Propionibacterium strains. Journal of Dairy Science, 71, 1769-1776.

Gobbetti, M., Lanciotti, R., Angelis, M.D., Corbo, M.R., Massini, R., and Fox, P.F.(1999). Study of the effects of temperature, pH and NaCl on the peptidase activities of non- starter lactic acid bacteria (NSLAB) by quadratic response surface methodology. International Dairy Journal, 9, 865-875.

Jimeno, J., Lazaro, M. J., and Sollberger, H. (1995). Antagonistic interaction between propionic acid bacteria and non-starter lactic acid bacteria. Lait, 75, 401-413.

Juillard, V., Guillot, A, Bars, D., and Gripon, J-C. (1998). Specificity of milk peptide utilization by Lactococcus lactis. Applied and Environmental Microbiology, 64, 1230-1236.

Kasicka, V. (2001) Recent advances in capillary electrophoresis of peptides. Electrophoresis, 22, 4139-4162.

108 Kunji, E.R.S., Hagting, A., Vries, C.J.D., Juillard, V., Haandrikman, A.J., Poolman, B., and Konings, W.K. (1995). Transport of β-casein-derived peptides by the oligopeptide transport system is crucial step in the proteolytic transport system is a crucial step in the proteolytic pathway of Lactococcus lactis. The Journal of Biological chemistry, 270, 1569- 1574.

Lungsrud, T., Reinbold, G.W., and Hammond, E.G. (1978). Free proline production by strains of Propionibacteria. Journal of Dairy Science, 61, 303-308.

Langsrud, T., Sorhaug, T., and Vegarud, GE. (1995). Protein degradation and amino acid metabolism by propionibacteria. Lait, 75, 325-330.

Law, J. and Haandrikman, A. (1997). Proteolytic enzymes of lactic acid bacteria. International Dairy Journal, 7, 1-11.

Lemee, R., Gagnaire, V., and Maubois, J-L. (1998). Strain variability of the cell-free proteolytic activity of dairy propionibacteria toward β-casein peptides. Lait, 78, 227- 240.

McCarthy, R.J., and Courtney, P.D. 2004. Amino acid requirements of dairy Propionibacterium strains. Abstract 17A-10. Institute of Food Technologiests Annual Meeting, Las Vegas, NV.

McSweeney, P. L. (2004). Biochemistry of cheese repening. International Journal of Dairy Technology, 57, 127-144.

Noel, Y., Boyaval, P.,Thierry, A.,Gagnaire, V., and Grappin, R. (1999). Eye formation and Swiss-type cheese, Technology of Cheesemaking. Sheffiled Academic Press, 235.

Panon, G. (1990). Purification and characterization of a proline iminopeptidase from Propionibacterium shermanii 13673, Le Lait, 70, 439-452. 109

Parker, J.A., and Moon, N.J. (1982). Interactions of Lactobacillus and Propionibacterium in mixed culture. Journal of Food Protection., 45, 326-330.

Perez Chaia, A., Pasce de Ruiz Holgado, A., and Oliver, G. (1990). Peptide hydrolases of propionibacteria: effect of pH and temperature. Journal of Food Protection, 3, 227-240.

Pivetean, PG., Condon, S., and Cogan, TM. (1995). Interactions between lactic and propionic acid bacteria. Lait, 75, 331-343.

Piuri, M., Sanchez-Rivas, C., and Ruzal, S.M. (2003). Adaptation to high salt in Lactobacillus: role of peptides and proteolytic enzymes. Journal of Applied Microbiology, 95, 372-379.

Quelen, L., Dupuis, C., and Boyaval, P. (1995). Proline-specific activities of Propionibacterium freudenreichii subsp. shermanii. Journal of Dairy Research, 62, 661-666.

Reid, J. R., Coolbear, T., Ayers, J. S., and Coolbear, K. P. (1997). The action of chymosin on k-casein and its macropeptide: effect of pH and analysis of products of secondary hydrolysis, International Dairy Journal, 7, 559-569.

Sahlstrom, S., Espinosa, C., Langsrud, T., Sorhaug, T. (1989). Cell Wall, membrane, and intracellular peptides activities of Propionibacterium shemanii. Journal of Dairy Science, 72, 342-350.

Sousa, M.J., Ardo, Y., and McSweeney, P.L.H. (2001). Advances in the study of proteolysis during cheese ripening. International Dairy Journal, 11, 327-345.

Stepaniak, L. (2000). Isolation and characterization of proline iminopeptidase from Propionibacterium frendenreichii ATCC 9614. Nahrung, 44, 102-106.

Tapuhi, Y., Schmidt, D.E., Lindner, W., and Karger, B.L. 1981. Dansylation of amino acids for high-performance liquid chromatography analysis. Analytical Biochemistry. 115, 123- 129.

110 Thierry, A., Salvat-Brunaud, D., and Maubols, Jean-Louis. (1999). Influence of thermophilic lactic acid bacteria strains on propionibacteria growth and lactate consumption in an Emmental juice like medium. Journal of Dairy Research, 66, 105-113.

Valence, F., Richoux, R., Thierry, A., Palva, A., and Lortal, S. (1998). Autolysis of Lactobacillus helueticus and Propionibacterium freudenreichii in Swiss cheese: first evidence by using species-specific lysis markers. Journal of Dairy Research, 65, 609-620.

Verheul, A., Hagting, A., Amezaga, M-R., Booth, I. R., Rombouts, F. M., and Abee, T. (1995). A di- and tripeptide transport system can supply Listeria monocytogenes scott A with amino acids essential for growth. Applied and Environmental Microbiology, 226-233. Williams, A.G., Felipe, X., and Banks, J.M. (1998). Aminopeptidase and dipeptidyl peptidase activity of Lactobacillus spp. and non-starter lactic acid bacteria (NSLAB) isolated from cheddar cheese. International Dairy Journal, 8, 255-266.

Zwietering, M.H., Jongenburger, Il, Rombouts, F.M. and van’t Reit, K. 1990. Modeling of the bacterial growth curve. Applied and Environmental Microbiology, 56, 1875-1881.

111

APPENDIX

112

PRELIMINARY STUDY TO EVALUATE THE GROWTH OF

PROPIONIBACTERIUM FREUDENREICHII IN CHEESE

SLURRY SERUM PRODUCED WITH DIFFERENT

LACTOBACILLUS HELVETICUS STRAINS

1. ABSTRACT

Three bacterial species are used to manufacture Swiss cheese: Streptococcus thermophilus,

Lactobacillus helveticus and Propionibacterium freudenreichii. Swiss cheese manufacturers report that some P. freudenreichii strains do not grow well in cheese in the presence of particular L. helveticus strains, resulting in poor eye and flavor formation. The goal of the study was to establish a rapid screening system for successful culture combinations, so that strains may be paired systematically. Cheese curds prepared aseptically with one S. thermophilus strain and one of three different L. helveticus strains were homogenized with NaCl and water to prepare a cheese slurry. A control slurry was prepared from curds containing only S. thermophilus.

The aqueous phase (serum) was collected when the slurry pH reached 5.1-5.3. The serum was filter-sterilized and used as the growth medium for five P. freudenreichii strains.

Propionibacterium freudenreichii growth depended on the L. helveticus strain. The P. freudenreichii strains P859, P764M1, and ATCC9614 grew in the slurry sera prepared with all L. helveticus strains, though better with L. helveticus strain L856 and L887. On the other hand, P. 113 freudenreichii P873 grew in the L. helveticus 856 serum, but did not grow well or possible grew slowly in the sera prepared with L. helveticus strains. The opposite was observed for P. freudenreichii strain P812, which did not grow in any slurry serum. P. freudenreichii strain P859,

P764M1, and ATCC9614 could grow slowly in the control serum prepared without any L. helveticus. The results indicated that the slurry supported the Propionibacterium growth.

Specific peptides produced by the extracellular protease of each L. helveticus strain may play a role in these differences since P. freudenreichii strains have different amino acid requirements and may have different abilities to import peptides. The screening system developed provides a rapid method for evaluating potential culture combinations for Swiss cheese manufacturing. It will also be useful in further research on the relationship between these species.

2. INTRODUCTION

Three bacterial species are used to manufacture Swiss cheese in the U.S.: S. thermophilus, L. helveticus and P. freudenreichii. S. thermophilus produces lactic acid from lactose during the first stages of cheese production. L. helveticus also produces lactic acid and plays another important role by contributing to the proteolysis of the casein proteins during cheese ripening. Proteolytic enzymes (a protease and several peptidases) from L. helveticus help to breakdown casein into peptides and amino acids important to cheese flavor. P. freudenreichii degrades the lactic acid produced by the other starter cultures and produces propionic and acetic acids, which contribute to cheese flavor, and carbon dioxide gas, which creates the holes called “eyes” in Swiss cheese.

114 Many different strains of each bacterial species are available commercially. Thus, the cheese maker has many options concerning the combination of the three cultures to use.

Cheese makers have reported that some combinations of L. helveticus and P. freudenreichii result in poor eye and flavor formation during ripening. Successful culture combinations are usually found by trial and error or in consultation with the culture supplier. The goal of the project is to better understand the relationship between L. helveticus and P. freudenreichii, to be able to predict which strains works well together.

The initial steps in Swiss cheese making involve coagulation of the casein proteins in the milk using rennet, a proteolytic enzyme, followed by cutting and cooking the coagulated curd to remove whey. During these steps, the S. thermophilus and L. helveticus starter cultures metabolize the lactose in the milk to lactic acid. After cooking, the curd particles are pressed together, and then the cheese block is soaked in a brine solution to add salt. The cheese is then cooled 8°C for 4-14 days to allow the salt to equilibrate throughout the cheese. The S. thermophilus and L. helveticus strains are inhibited by the salt and cool temperature and the population will begin to slowly lyse and die. Though the cells are dying, the enzymes present in the bacteria will be released as the cells lyse and continue to work on cheese components during cheese ripening.

flavor and CO2 gas for eye formation. During this time, the proteolytic enzymes from L. helveticus degrade some of the casein proteins into peptides and amino acids. These

115 contribute to flavor, but may also stimulate P. freudenreichii growth. When the desired eye size has been achieved, the cheeses proceed to the second phase of ripening, “cold room” ripening, in which the cheeses are held at 4 - 10°C for 1 – 3 months. Flavor continues to develop during this time as more proteolysis of the casein occurs along with other chemical reactions.

Cheese experiments are expensive and consume time up to two years for ripening to develop cheese flavor during biochemical change such as proteolysis, lipolysis and glycolysis.

Most experiments in cheese studies used a model to reduce time consuming and cost. Some researchers suggested the models used for pilot plant experiments the models were closer to cheese properties and the biochemical reaction closes to normal cheese during ripening. Some models were reported for microbiological studies of cheese ripening with a short period of time. The experiment was conducted under the sterilization condition. They described a protocol of the miniature model under controlled microbiological conditions. The dry matter and the salt in moisture were not significant difference from old cheeses. However, the cheese slurry models used for studies of the flavor development during cheese ripening were unclear and the slurries effected the off flavor development. This study, thus, developed the cheese slurry to understand the relationship between L. helveticus and P. freudenreichii in Swiss cheese.

The goal was to prepare Swiss cheese slurry models with different L. helveticus strains to study the growth of Propionibacterium.

116 3. MATERIALS AND METHOD

3.1 Prepare model Swiss cheese slurry serum with Lactobacillus helveticus strains

A model cheese slurry system representing conditions similar to a ripening Swiss cheese was developed (Jenkins, Limpisathian and Courtney, unpublished results) and is described below. The model has two advantages over real Swiss cheese for the present study: (1) it can be performed under aseptic conditions, allowing study of specific microorganisms and (2) it requires less time and raw material than cheese.

To prepare cheese slurries, one liter of sterile fat-free UHT milk was aseptically dispensed into a sterile 1 L container. The pH was measured then adjusted to 6.0 using L-

lactic acid and sterile CaCl2 was added to a final concentration of 0.018%. The milk was warmed to 33.3°C for 30 minutes in a circulating water bath before adding 1 ml of S. thermophilus (S787) and one of the Lactobacillus strains. Eight strains of the L. helveticus were used (see materials and method). Approximately 150 mg (wet weight) of cells were added for each culture. The S. thermophilus strain was the same for all preparations. A 2.5% solution (5.5 ml) of chymosin was added and the milk was left undisturbed at 33.3°C for 40 minutes to produce the coagulated curd. The curd was cut into small cubes with a sterile wire-mesh cutting device. The curds and whey were cooked to 50°C slowly in water bath. A

“dummy” container was used to monitor the temperature. The pH near the end of cooking is measured and curds are held at 50 °C until the pH reaches 5.8. When the pH is 5.8, the whey is drained through sterile cheesecloth. The remaining curd is cooked further until the 117 pH of the whey reaches 5.6-5.5. After further whey draining, the curd moisture content was measured using the microwave moisture analyzer method (LabWave 9000 Moisture/Solids

Analyzer, Metthews, NC). Sterile 22% salt and sterile distilled water were added to the curds to yield a final moisture content of 60%, salt in moisture of 1% at pH 5.3. The mixture was homogenized for 5 minutes to form a thick slurry.

Each slurry was centrifuged at 10 krpm for 10 minutes and the supernatant was recovered. The supernatants were filtered through a 0.45 µm filter and frozen at -80°C until use. This sterile slurry supernatant is referred to as the slurry serum. Part of each sterile slurry serum was used as the growth medium for P. freudenreichii in the preliminary experiments described above. Triplicate slurry sera were prepared from each of the four L. helveticus strains. Triplicate control slurries lacking L. helveticus also were prepared.

3.2 Screening Propionibacterium freudenreichii and Lactobacillus helveticus strains

Propionibacterium medium were prepared by using 20 g casein-digestive tryptone, 5 g yeast extract, 5 g dextrose. Propionibacterium agar were prepared by using 20 g casein-digestive tryptone, 5 g yeast extract, 5 g dextrose, and 5 g agar. Twelve different P. freudenreichii strains from commercial starter cultures and cheeses were selected for analysis; P843, P891, P859,

P835, P873, P812, P745M1, P745M2, P764M1, P728M1, P318, and ATCC9614. The cultures were kept at -80ºC and inoculated to Propionibacterium medium, then incubated in anaerobe chamber at 28 to 30° C for 2 days before used. L. helveticus strain L856, L887M1,

118 L879, L346M1, L346M2, L367, L350, and L774M1 were kept at -80ºC and inoculated to

MRS broth, then incubated in anaerobe chamber at 28 to 30° C for 1 days before used.

To prepare the slurry, fat free UHT Milk (Parmalat), 20% L-lactic acid (Sigma, St.

Louis, MO) diluted with sterile water, 10% sterile calcium chloride, and 2.5% rennet solution; a standardized solution of 100% fermentation product chymosin, a milk clotting enzyme (ChyMax Extra, Chr. Hansen, Milwaukee, WI) was prepared fresh, and 22% NaCl were used. The objective one has been evaluated by preparing the cheese slurry with eight

(8) strains of L. helveticus: L856, L887M, L879, L346M, L346M2, L367, L350, L774M. After the supernatants were recovered as described above, the whey samples were inoculated with

P. freudenreichii: P843, P891, P859, P873, P812, P745M1, P745M2, P764M1, P728M1, P318,

ATCC9614

Twelve strains of the P. freudenreichii and 8 strains of the L. helveticus were used to screen in preliminary study. Each Propionibacterium strain was grown in each slurry serum made with one Lactobacillus strain. The results were shown the different growth curve at 0, 3,

5, and 9 days for each Propionibacterium (The figures were shown in the appendix A).

Propionibacterium with the high, medium, and low growth curve were identified. Five (5) strains of Propionibacterium: P859, P873, P812, P764M1, and ATCC9614 and four (4) strains of L. helveticus: L346M1, L350, L856, and L887M1, were selected to study.

Two hundred µl of each of five (5) strains of Propionibacterium: P859, P873, P812,

P764M1, and ATCC9614 are inoculated in 10 ml Propionibacterium medium and incubated for

2 days at 28 to 30 °C in anaerobe chamber. Then each 1000 µl slurry was inoculated with each 20 µl P. freudenreichii strain incubated at 28 to 30 °C anaerobic chamber at 0, 3, 5 and 9 days. After completing the days, the samples are spread on selective Propionibacterium agars. 119 The cells of Propionibacterium in cfu/ml will be counted to evaluate the growth of P. freudenreichii in cheese slurry serum produced with different L. helveticus . All results from the analysis will be identified.

3.3 Evaluate the growth rate of Propionibacterium freudenreichii in cheese slurry serum produced with different Lactobacillus helveticus in 96 hours

Frozen five strains of P. freudenreichii; P843, P745M1, P764M1, P318,and ATCC9614 were grown in the Propionibacterium medium for 3 days in anaerobe chamber and then the culture, 200 µl was inoculated in 10 ml the same medium for another 3 days. The cultures were diluted to 1:100, except strain P812 diluted to 1:10, with 0.1% peptone solution before

80 µl of each P. freudenreichii strain; was inoculated to 4 ml of Propionibacterium medium and the four slurry sera. The blank sample, Propionibacterium medium alone and each four slurry sera alone without P. freudenreichii were performed. All samples were incubated in the anaerobe chamber at 28 to 30 ºC and collected the sample at 0, 6, 12, 18, 24, 30, 36, 42, 48,

72, 96 hours. After completing the hours, the samples are spread on selective

Propionibacterium agars. The cells of Propionibacterium in cfu/ml will be counted to evaluate the growth of P. freudenreichii in cheese slurry serum and Propionibacterium medium produced with different L. helveticus.

120 4. RESULTS AND DISCUSSION

4.1 Screening Propionibacterium freudenreichii and Lactobacillus helveticus strains

Using a Swiss cheese model prepared with S. thermophilus and L. helveticus as a growth medium for P. freudenreichii strains, we were able to see differences in P. freudenreichii growth depending on the L. helveticus strains used. Eight L. helveticus and eleven P. freudenreichii strains were screened (results showed in Appendix A). Three different growth pattern was observed:

(1) L. helveticus strain 367 did not stimulate any P. freudenreichii growth, (2) L. helveticus strains

L887, L346M1, and L774M1 stimulated the growth of most P. freudenreichii strains and reduced the growth of the strain P812 and (3) L. helveticus strain L879 stimulated all P. freudenreichii growth (Figures shown in Appendix A).

The results from the growth of Propionibacterium 12 strains showed that the growth of

8 9 them were high (10 -10 cfu/ml); with L. helveticus strain L887M1, except Propionibacterium strain P812 and P835 were low (105-106 cfu/ml). The other Propionibacterium strains grew different within each strain of L. helveticus. L. helveticus strain L879 has the most effective to

7 9 9 Propionibacterium growth all strains (10 -10 cfu/ml), especially strain P764M1 (10 cfu/ml) whereas Propionibacterium strains; P891, P318, ATCC9614 were lower (107 cfu/ml). Among

Propionibacterium growth, L. helveticus strain L346M1 was the only the less effective to the growth rate of Propionibacterium for all strains (104-107 cfu/ml), except Propionibacterium strain

9 P764M1 that showed the high growth rate (10 cfu/ml). On the other hand, different L. helveticus strains affect different strains of Propionibacterium growth. The growth rate of

121 Propionibacterium strain P812 were the lowest (104-105 cfu/ml) among L. helveticus strain L856,

L887M, L346M1, L345M2, L350 and L774M1, except strain L879 and L367; the growth rate of Propionibacterium strain P812 were 108 cfu/ml and 107 cfu/ml, respectively. In general, the growth rates of Propionibacterium strain P843 were high (108 cfu/ml) among L. helveticus strain

L887M1, L879, L346M2, L350 and L774M1. Furthermore, Propionibacterium strain P764M1

9 was the highest growth rate (10 cfu/ml) among L. helveticus strain L346M1, L346M2, and

L367.

The preliminary study showed that L. helveticus strains interact with P. freudenreichii growth. In addition, the growth of P. freudenreichii were different within each strain of L. helveticus. Some Propionibacterium grew in high rate (109 cfu/ml) and some had slow in the growth rate. This preliminary study, therefore, provided an evident that the strains of these bacteria were involved in the growth interaction of these bacteria. This finding agreed with

Pivetean and others (1995). They found that three strains of L. helveticus stimulated the growth of three strains of P. freudenreichii. Furthermore, the interaction may involve with Lactobacillus provides some essential peptide to Propionibacterium. Each strain of Lactobacillus potentially hydrolyzes different peptides and/or utilizes different peptides and amino acids for their growth.

The proteolytic system; peptides, amino acid transport system such as proteinase

(CEP, lactocepin, PrtP) intercellular oligoendopeptidase (Pep O and Pep F), aminopeptidase

(Pep N, Pep C, and Pep G) has been involved. Each strains of Propionibacterium possibly involve in different peptide transport systems. The finding that Propionibacterium are strongly peptidolytic ones, especially in proline containing peptides and much higher in proline iminopeptidase activities (Langsrud and others 1995). Pep N, Pep I, Pep X, and

122 endopeptidase have been found in P. freudenreichii (Sousa and others 2001). In addition, amino acid, such as aspartic acid, serine, glycine, and alanine were rapid degraded by P. freudenreichii. Due to the previous investigation, the growth of Propionibacterium possibly increases by change and size of peptides. The neutral and base peptides may probably support the growth. It may inhibit by acidic peptides. Each strain of Lactobacillus will hydrolyzed different in size peptides and change peptides results in several amino acids in the whey samples. All finding can be involved in the interaction between Lactobacillus and

Propionibacterium.

4.2 Effect of slurry incubation time on Propionibacterium growth

Most of P. freudenreichii strain did not grow well in the slurry incubated for 4 and 10 days. Only P. freudenreichii strain P859 grew in slurry prepared with L. helveticus strain 346 at day

10. The pH at day 4 and day 10 of the slurry sera after incubated at 22 ºC dropped between

4.8 and 4.1. They grew better in the slurry sera after finished the process pH 5.3 to 5.4 (figure shown in appendix B). The Propionibacterium growth is dependent on Lactobacillus strain. Thus,

Acidity produced by both Lactobacillus and Streptococcus influenced the growth.

4.3 Propionibacterium growth with four selected Lactobacillus helveticus strains

The strains selected for further study are shown in Figures . P. freudenreichii strain P859 grew well in the cheese sera prepared with all four L. helveticus strains, though better with L. helveticus

123 strain L350 and L856 (Figure 2.1). P. freudenreichii strain P812, on the other hand, did not grew in any the serum from L. helveticus. In fact, L. helveticus strain L350 and L887M1 (Figures 2.3) reduced the viable population of P. freudenreichii strain P812. Different growth stimulation patterns can also be observed for P. freudenreichii strain P764M1 and ATCC9614 using the four different L. helveticus cheese sera. Strain P873 did not grow well in the cheese sera from L. helveticus strain L350 and L346M1. None of P. freudenreichii grew well in the control lacking L. helveticus.

Peptides produced by the protease of each L. helveticus strain are a possible factor that account for these differences since P. freudenreichii strains may have different abilities to import peptides and different amino acid requirements. Our laboratory has demonstrated that dairy propionibacteria grew slowly without amino acids, but were stimulated by certain amino acids.

The specific amino acids needed for maximal growth vary from strain-to-strain (McCarthy and

Courtney, unpublished results).

4.4 The growth rate of Propionibacterium freudenreichii in cheese slurry serum produced with different Lactobacillus helveticus in 96 hours

The results of the growth curve indicated that P. freudenreichii grew well to above 108 cfu/ml for all strains, except the strain P764M1 grew to 107 cfu/ml within 96 hours in

Propionibacterium medium (shown in Figure A.1). All P. freudenreichii strains grew slowly within

96 hours in all slurry sera compare to Propionibacterium medium. P. freudenreichii strain P859

(Figure A.1) and ATCC9614 (Figure A.5) grew faster than other strains to 106 in all L. helveticus strains. P. freudenreichii strain P873 (Figure A.2) and P764M1 (Figure A.4) did not grew well

124 within 96 hours in all slurry sera. The result indicated that all P. freudenreichii strains grew slowly in the slurry sera.

6 Log cfu/ml Log

2 0246810 Day

L346 L350 L856 L887 Control

Figure A.1: The growth P. freudenreichii strain P859 in slurry sera prepared with different L. helveticus strains (pH 5.3)

125 10

6 Log cfu/ml Log

2 0246810 Day

L346 L350 L856 L887 Control

Figure A.2: The growth of P. freudenreichii strain P873 in slurry sera prepared with different L. helveticus strains (pH 5.3).

126 10

6 Log cfu/ml Log

2 0246810 Day

L346 L350 L856 L887 Control

Figure A.3: The growth of P. freudenreichii strain P812 in slurry sera prepared with different L. helveticus strains (pH 5.3).

127

6 Log cfu/ml Log

2 0246810 Day

L346 L350 L856 L887 Control

Figure A.4: The growth of P. freudenreichii strain P764M1 in slurry sera prepared with different L. helveticus strains (pH 5.3).

128 10

6 Log cfu/ml Log

2 0246810 Day

L346 L350 L856 L887 Control

Figure A.5: The growth of P. freudenreichii strain ATCC9614 in slurry sera prepared with different L. helveticus strains (pH 5.3).

129

7 Log cfu/ml Log 6

0 20 40 60 80 100 120 Hour

P859 P873 P812 P764M1 P9614

Figure A.6: The different strains of P. freudenreichii growth in Propionibacterium media within 96 hours

130 6.5

6.0

5.5 Log cfu/ml

5.0

0 20 40 60 80 100 120 Hour

L346 L350 L856 L887 Control

Figure A.7: The growth of P. freudenreichii strain P859 in slurry sera prepared with different L. helveticus strains within 96 hours.

131 6.5

6.0

5.5 Log cfu/ml

5.0

0 20 40 60 80 100 120 Hour

L346 L350 L856 L887 Control

Figure A.8: The growth of P. freudenreichii strain P873 in slurry sera prepared with different L. helveticus strains within 96 hours.

132 6.5

6.0

5.5 Log cfu/ml

5.0

0 20 40 60 80 100 120 Hour

L346 L350 L856 L887 Control

Figure A.9: The growth of P. freudenreichii strain P812 in slurry sera prepared with different L. helveticus strains within 96 hours.

133 6.5

6.0

5.5 Log cfu/ml

5.0

0 20 40 60 80 100 120 Hour

L346 L350 L856 L887 Control

Figure A.10: The growth of P. freudenreichii strain P764M1 in slurry sera prepared with different L. helveticus strains within 96 hours.

134 6.5

6.0

5.5 Log cfu/ml

5.0

0 20 40 60 80 100 120 Hour

L346 L350 L856 L887 Control

Figure A.11: The growth of P. freudenreichii strain ATCC9614 in slurry sera prepared with different L. helveticus strains within 96 hours.

135

6 Log cfu/ml Log

02468101214 Day