The Pennsylvania State University

The Graduate School

Department of Agricultural and Biological Engineering

ENHANCED NISIN PRODUCTION IN A BIOFILM REACTOR

AND SEPARATION OF NISIN

A Thesis in

Agricultural and Biological Engineering

by

Thunyarat Pongtharangkul

© 2006 Thunyarat Pongtharangkul

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

May 2006 The thesis of Thunyarat Pongtharangkul was reviewed and approved* by the following:

Ali Demirci Associate Professor of Agricultural Engineering Thesis Advisor Chair of Committee

Virendra M. Puri Professor of Agricultural Engineering

Emine Koc Assistant Professor of Biochemistry and Molecular Biology

Roy E. Young Professor of Agricultural Engineering

Joseph Irudayaraj Associate Professor of Agricultural and Biological Engineering Purdue University Special Member

Roy E. Young Head of Department of Agricultural and Biological Engineering

* Signatures are on file in the Graduate School.

ii ABSTRACT

Nisin, a 34-amino acid polypeptide produced by Lactococcus lactis subsp. lactis during the fermentation, is the only FDA approved bacteriocin for food applications.

Improving nisin production through optimization of fermentation parameters would make nisin more cost-effective for various applications.

In this study, a biofilm reactor with Plastic Composite Support (PCS) was evaluated for nisin production using L. lactis NIZO 22186. The high-biomass density of the biofilm reactor contributed to a significantly shorter lag time of nisin production relative to a suspended-cell reactor. The medium evaluation study suggested a complex media for nisin production in the biofilm reactor as: 4% (w/v) sucrose, 0.02% (w/v)

MgSO 4•7H 2O, 1% (w/v) KH 2PO 4, 0.2% (w/v) NaCl, 1% (w/v) peptone, and 1% (w/v) yeast extract. Nisin production in the biofilm reactor was significantly increased by 3.8- fold (2,208 IU/ml) with the suggested complex medium.

Nisin production in biofilm reactor was highly affected by the pH profiles during fermentation. In batch fermentation, profile Const4, in which pH was allowed to drop freely via autoacidification after 4 h, yielded almost 1.9-time higher nisin (3,553 IU/ml) than profile Const12 (1,898 IU/ml), possibly as a result of less adsorption of nisin onto producer cells. In fed-batch fermentation, profile Const12 enhanced nisin production in suspended-cell (4,188 IU/ml) and biofilm (4,314 IU/ml) reactors while the pH profiles that include periods of autoacidification (e.g. Const4) resulted in a significantly lower nisin production due to toxicity of lactic acid in acidic environment.

iii Silicic acid has been successfully used to remove nisin from the fermentation broth. The maximum recovery (47% total recovery) can be achieved when the adsorption was carried out at pH 6.8 and 1 M NaCl with 20% ethanol was used as an eluent.

Using the results from suspended-cell batch fermentation, the re-modified logistic and Gompertz models proposed in this study adequately described the growth and the validation showed that they could be used for the prediction accurately (slope = 1.01, R- square = 0.99). As for nisin production, the Luedeking-Piret model fit considerably well and could be successfully use to predict nisin production from the value of biomass

(slope = 1.04, R-square = 0.98).

iv Table of Contents

Page

List of Figures ------x List of Tables ------xiii Acknowledgements ------xiv Technical Acknowledgement------xv

Chapter 1. Introduction ------1

Chapter 2. Literature Review ------5 2.1. General background ------5 2.2. Nisin ------6 2.2.1. Structure, solubility and stability of nisin ------6 2.2.2. Biosynthesis of nisin ------9 2.2.3. Antimicrobial activity of nisin------10 2.2.4. Nisin activity and quantification techniques------12 2.2.5. Applications of nisin ------15 2.3. Nisin production ------18 2.3.1. Producer strain ------18 2.3.2. Nutrients ------19 2.3.2.1. Carbon source------19 2.3.2.2. Nitrogen source------21 2.3.2.3. Mineral source ------22 2.3.3. pH and temperature------23 2.3.4. Agitation and aeration ------24 2.3.5. Types of fermentation------25 2.4. Biofilm reactor ------28 2.4.1. Biofilm formation ------29 2.4.2. Types of biofilm reactors ------32 2.4.3. Solid support for biofilm ------33

v 2.4.4. Plastic Composite Support (PCS) ------34 2.5. Adsorption and degradation of nisin ------36 2.6. Inhibitions in nisin production ------37 2.7. Recovery and purification of nisin ------38 2.8. Models of nisin production ------41 2.9. Summary of literature review ------43

Chapter 3. Evaluation of Agar Diffusion Bioassay for Nisin Quantification ------45 Abstract ------45 Introduction ------46 Materials and methods------49 Microorganisms and media------49 Nisin standards ------49 Nisin bioassay ------50 Verification ------52 Statistical analysis------52 Results ------52 Discussions------57 References------59

Chapter 4. Evaluation of Culture Medium for Nisin Production in Repeated-batch Biofilm Reactor------62 Abstract ------62 Introduction ------63 Materials and methods------67 Microorganisms and media------67 Plastic Composite Support (PCS)------68 Test-tube fermentations for selection of PCS ------69 Nisin fermentation in bioreactor ------70 Bioreactor experimental design ------72 Analysis ------73

vi Statistical analysis------75 Results and discussions------76 PCS selection ------76 Fermentation profiles of L. lactis in PCS biofilm reactor ------78 Effect of different sugars on production of nisin Z and lactic acid --81 Evaluation of medium for nisin Z production in PCS biofilm reactor------83 Conclusions ------87 References------89

Chapter 5. Effect of pH Profiles on Nisin Production in Biofilm Reactor ------95 Abstract ------95 Introduction ------96 Materials and methods------98 Microorganisms and media------98 Plastic Composite Support (PCS)------98 Nisin fermentation in bioreactor ------99 Experimental design ------100 Analysis ------102 Statistical analysis------103 Results ------104 Discussions------108 References------111

Chapter 6. Fed-batch Fermentation for Nisin Production in a Biofilm Reactor and Effects of pH Profiles ------114 Abstract ------114 Introduction ------115 Materials and methods------117 Microorganisms and media------117 Experimental design ------118

vii Plastic Composite Support (PCS)------120 Nisin fermentation in bioreactor ------120 Analysis ------121 Statistical analysis------123 Results ------123 Discussions------129 References------131

Chapter 7. Recovery of Nisin from Fermentation Broth Produced in a Biofilm Reactor ------135 Abstract ------135 Introduction ------136 Materials and methods------138 Microorganisms and media------138 Adsorption and desorption of nisin------139 Statistical analysis------140 Results and discussions------141 Conclusion ------143 References------144

Chapter 8. Modeling of Nisin Production by Lactococcus lactis during Batch Fermentation------146 Abstract ------146 Introduction ------147 Materials and methods------148 Microorganisms and media------148 Batch nisin fermentation ------149 Analysis ------149 Model development------151 Model for biomass production ------151 Model for lactic acid production ------159

viii Model for sucrose consumption------159 Model for nisin production ------161 Conclusions ------163 References------164

Chapter 9. Conclusions and scope for future research ------166

References------170 Appendix A1. Properties of Commercial Nisin Powder (Nisaplin)------186 Appendix A2. Effects of Various Protease Inhibitors on an Inactivation of Nisin --187 Introduction ------187 Materials and methods------188 Microorganisms and medium ------188 Preparation of cell-free fermentation broth------188 Study on nisin inactivation ------189 Study on effects of various protease inhibitors ------189 Results and discussions------191 Study on nisin inactivation ------191 Study on effects of various protease inhibitors ------192 Conclusions ------193 References------194 Vita

ix List of Figures Page

Figure 2.1. Structure of natural nisin variants; (a) nisin A, (b) nisin Z, and (c) nisin Q ------7 Figure 2.2. Formation of biofilm ------30 Figure 2.3. Factors affecting biofilm formation and structure------32 Figure 2.4. Dimensions of the Plastic Composite Support or PCS ------35 Figure 2.5. PCS biofilm reactors------35 Figure 3.1. Diameter of inhibition zone obtained from assays using different nisin sensitive microorganisms with method 2 ------53 Figure 3.2. A linear regression of actual and predicted nisin concentration obtained from an assay using L. sakei with method 2 and an assay using M. luteus with method 1 (conventional procedure)------57 Figure 4.1. Diagram of PCS biofilm reactor ------71 Figure 4.2. Effects of different PCS blends on numbers of viable attached cells on the PCS in terst-tube systems without pH control after 24 h (n = 3). (S, soybean hulls; F, soybean flour; Y, yeast extract; R, dried bovine RBC; B, dried bovine albumin; +, mineral salts) ---77 Figure 4.3. Effects of different PCS blends on nisin production in terst-tube systems without pH control after 24 h (n = 3). (S, soybean hulls; F, soybean flour; Y, yeast extract; R, dried bovine RBC; B, dried bovine albumin; +, mineral salts) ------78 Figure 4.4. Biomass (g/l) in liquid medium when using glucose and sucrose as a carbon source in suspended-cell reactor (SC) and biofilm reactor (BF) ------79 Figure 4.5. Nisin production (IU/ml) when using glucose and sucrose as a carbon source in suspension cell reactor (SC) and biofilm reactor (BF) ------79 Figure 4.6. Effects of sucrose, magnesium, and phosphate levels on maximum nisin production rate (IU/ml/h) in PCS biofilm reactor ------84

x Figure 4.7. Effects of sucrose, magnesium, and phosphate levels on final lactic acid concentration (g/l) in PCS biofilm reactor ------85 Figure 5.1. Actual pH profiles of nisin fermentation by L. lactis when using (a) profile Const12; (b) profile Const4; and (c) profile Stepwise ----101 Figure 5.2. Growth of L. lactis (in liquid broth) using the PCS biofilm reactor with different pH profiles ------105 Figure 5.3. Nisin production of L. lactis using the PCS biofilm reactor (BF) and the suspended-cell reactor (SC) when using different pH profiles ------105 Figure 5.4. Lactic acid production of L. lactis from the PCS biofilm reactor when using different pH profiles------106 Figure 5.5. Sucrose consumption of L. lactis from the PCS biofilm reactor when using different pH profiles------106 Figure 6.1. Actual pH profiles of nisin fermentation by L. lacits when using (a) profile Const12; (b) profile Const4; and (c) profile Const4-Stepwise------119 Figure 6.2. Suspended biomass of L. lactis in different culture types (BF: Biofilm; SC: Suspended-cell) ------124 Figure 6.3. Sucrose consumption of L. lactis in different culture types (BF: Biofilm; SC: Suspended-cell) ------124 Figure 6.4. Nisin and lactic acid production of L. lactis when using different culture types and pH profiles (BF: Biofilm; SC: Suspended-cell) ------126 Figure 6.5. Nisin production of L. lactis in the biofilm reactor using fed-batch Fermentation with various pH profiles (BF: Biofilm) ------127 Figure 6.6. Lactic acid production of L. lactis in the biofilm reactor using fed-batch Fermentation with various pH profiles (BF: Biofilm) ------128 Figure 7.1. Adsorption and desorption of nisin from silicic acid using Eluent 1 (DI water), Eluent 2 (20% ethanol), Eluent 3 (1 M NaCl), and Eluent 4 (1 M NaCl + 20% ethanol) when the adsorption was carried out at pH 6.8 ------142

xi Figure 7.2. Adsorption and desorption of nisin from silicic acid using Eluent 1 (DI water), Eluent 2 (20% ethanol), Eluent 3 (1 M NaCl), and Eluent 4 (1 M NaCl + 20% ethanol) when the adsorption was carried out at pH 3.0 ------142 Figure 8.1. Growth curve of L. lactis fitted with the modified logistic and re-modified logistic models------153 Figure 8.2. Growth curve of L. lactis fitted with the modified logistic and re-modified Gompertz models ------155 Figure 8.3. Regression of experimental and predicted biomass densities of L. lactis from validation with an independent set of data when fitted with the re-modified logistic model ------157 Figure 8.4. Regression of experimental and predicted biomass densities of L. lactis from validation with an independent set of data when fitted with the re-modified Gompertz model------157 Figure 8.5. Experimental and predicted lactic acid production during batch fermentation ------158 Figure 8.6. Experimental and predicted sucrose consumption during batch fermentation ------160 Figure 8.7. Experimental and predicted nisin production during batch fermentation ------160 Figure 8.8. Validation of nisin production model with an independent set of data------162 Figure 8.9. Validation of nisin production model with an independent set of data ------163 Figure 9.1. Summary of improvements in nisin production achieved in this study ------166 Figure A2.1. Inactivation of nisin in control (heated) and non-heated cell-free fermentation broth ------191 Figure A2.2. Effects of various protease inhibitors on inactivation of nisin in fermentation broth (C: Control; PI: Sample with protease inhibitor)192

xii List of Tables Page

Table 2.1. Examples of effective use of nisin in food ------16 Table 2.2. Applications of nisin as a part of hurdle technology------17 Table 2.3. Summary of nisin production in various types of fermentation ------26 Table 3.1. Effects of nisin sensitive microorganisms and incubation methods on slopes of regression equations------53 Table 3.2. Effect of nisin sensitive microorganisms and incubation methods on correlation coefficients of regression equations ------55 Table 3.3. Effects of nisin sensitive microorganisms and incubation methods on diameters of inhibition zones obtained from nisin standard solution at 5 IU/ml ------56 Table 4.1. List of PCS ingredients------69 Table 4.2. Effects of fermentation types and sugars on nisin and lactic acid productions------81 Table 4.3. Effect of different nutrient compositions on nisin and lactic acid productions------84 Table 8.1. Estimated model parameters of growth curve, root mean square (RMS), correlation coefficient (R-sq), and slope from the regression of experimental and predicted value when fitted with various models ------156 Table 8.2. Root mean square (RMS), correlation coefficient (R-square), and slope of the regression for validations of various growth models----158 Table A1.1. Properties of commercial nisin powder (Nisaplin) ------186 Table A2.1. Specificity of protease inhibitor tested in this study ------190 Table A2.2. Recommended concentrations for protease inhibitors tested in this study ------190

xiii ACKNOWLEDGEMENTS

I believe that one becomes what his experiences have shaped him and everyone who comes into our lives has taught us a valuable lesson or inspire us in some way. During the past four years, I have learned to develop myself professionally and personally here in Penn

State. All of these could not happen with out kind helps, concerns, and sacrifices of many people, some nearby and some half-world away.

I would like to express my sincere appreciation and deep gratitude to my advisor, Dr.

Ali Demirci, for always giving me his guidance and encouragements during the past four years. Thank you for being a great role model of a teacher and a mentor. I believe that I cannot wish for a better advisor than you.

I sincerely appreciate Dr. Virendra Puri, Dr. Joseph Irudayaraj, Dr. Emine Koc, and

Dr. Roy E. Young who always gave me their kind attentions and helpful guidance as the committee members. It was my pleasure having a chance to work with you all.

Many warm wishes and thanks to all my beloved friends here at Penn State,

Kathiravan Krishnamurthy, Deniz Cekmecelioglu, Mark A. Bechara, Stephen Walker, Ratna

R. Sharma, Katherine L. Bialka, Todd J. Miserendino, Yen Ching Huang, Ebru Ersay, Ozlem

Zabitgil, Ozlem Ozden, Matala J. Gupta, Christina Torres, Zahra Lotfi, Thitinan

Sumranwanich, Suchada Butnark, Nisanart Krasaechol, and Wallop (Worapas) Promsen.

Last but not least, I’m sincerely indebted to my family for their sacrifices and supports. I cannot be here today without them. I would have to write one more page of this acknowledgement to thank you for all the things you’ve done for me and how you always support me. ☺ A short version of it would be….I love you all. Thank you so very much.

xiv TECHNICAL ACKNOWLEDGEMENTS

I wish to thank Dr. Anthony L. Pometto III for kindly supplying the PCS tubes used in this research. Funding for this project was provided by the Pennsylvania

Agricultural Experiment Station and by a scholarship from the Royal Thai Government.

xv CHAPTER 1

INTRODUCTION

Nisin, a small antimicrobial polypeptide produced by food-grade lactic acid bacteria (LAB) Lactoccocus lactis subsp. lactis , not only is the most extensively studied bacteriocin, but also is the only bacteriocin approved for applications in food (FDA,

1988). Due to its antimicrobial activity against a wide range of Gram-positive spoilage and pathogenic bacteria, nisin is widely accepted and used in many countries worldwide

(O’Keeffe and Hill, 2000). Its suitability as a food preservative is a result of various extraordinary characteristics including its nontoxicity, stability under extreme heat treatment at low pH; and absence of color or flavor (Davies and Delves-Broughton,

2000). The principal commercial applications of nisin are in foods and beverages that are pasteurized, but not fully sterilized; such as cheese, milk, and dairy desserts. (O’Keeffe and Hill, 2000). Although it is mainly used as a food preservative, its applications for medical purposes are promising.

The commercial product Nisaplin® is prepared from the milk-based medium in batch fermentation followed by recovery processes of frothing, dewatering, and drying process that involves spraying onto a rotating heated drum (Hirsch, 1950). Development of new applications for nisin is still limited, mainly by its low production rate during fermentation and its tedious recovery processes (Kelly et al., 2000), which consequently contributes to high price of partially- and fully-purified nisin. The product, which contains only 2.5% nisin, is sold for approximately $250-350 per kg. Improving nisin

1 production would make nisin more cost-effective for various applications. Therefore, it is essential to develop production processes that yield both high nisin production and volumetric productivity.

This research attempts to enhance nisin production rate by utilization of a biofilm reactor with plastic composite supports (PCS) in both repeated-batch and fed-batch fermentations. PCS, an extrusion product of polyprolylene and agricultural products mixture, was developed at Iowa State University and was later awarded with a U.S.

Patent (U.S. Patent Number: 5,595,893). The polypropylene in PCS acts as a matrix to integrate the mixture of agricultural products, which provide essential nutrients to sustain cell growth. PCS has been shown to improve the growth of other LAB and to increase ethanol production in fed-batch and continuous fermentations (Demirci et al., 1995;

Demirci et al ., 1997; Ho et al ., 1997a; 1997b). De Vuyst and Vandamme (1994) and

Parente and Ricciardi (1999) reported that nisin production was growth-associated and that a high cell concentration was required for high nisin titers, therefore it is most likely that nisin production can be improved in the biofilm reactor where high biomass density can be maintained.

There have been many studies on utilization of cell immobilization for production of nisin. However, nisin production reported with these systems was either very low

(Wan et al ., 1995) or 2.5 to10-fold lower than that obtained with pH controlled, suspended-cell, batch cultures (Sonomoto et al ., 2000; Desjardins et al ., 2002). A recent study (Bertrand et al ., 2001), in contrary, showed that pH-controlled, repeated batch cultures with cell immobilization can greatly increase nisin production compared with continuous cultures using free and immobilized cells. Bertrand et al. (2001) suggested

2 that evolutionary conditions produced during batch cultures are required for high nisin production. So far, the only study on utilization of biofilm system in nisin fermentation was from Bober and Demirci (2004). Although the results showed no significant increase in nisin production or productivity rates using the PCS biofilm reactor in batch and fed- batch fermentation, an inappropriate culture medium and sampling method as well as inadequate mixing in the reactor were suspected as limitations.

In this study, several objectives have been fulfilled in order to improve the nisin production in PCS biofilm reactor. First, the appropriate bioassay procedure for monitoring of nisin concentration during fermentation process has been established

(Chapter 3). The culture medium for production of nisin in PCS biofilm reactor has been optimized, and the nisin production from the biofilm reactor has been compared with a result from suspended-cell reactor (Chapter 4). Then, effects of various pH control profiles on nisin production in both biofilm and suspended-cell reactors were studied in both repeated-batch (Chapter 5) and fed-batch (Chapter 6) fermentations.

Significant loss of nisin during fermentation was known to be a result both adsorption of nisin onto the producer cells (Kim et al., 1997) and enzymatic degradation of nisin by proteases (De Vuyst and Vandamme, 1992). In order to minimize this loss and prevent the effect of product inhibition by nisin itself (Kim et al., 1997), integrated adsorbent unit can be coupled with the fermentation process. In this research, silicic acid has been evaluated for recovery of nisin Z from the fermentation broth obtained from the reactor (Chapter 7). Effects of pH and eluents on adsorption and desorption of nisin have been studied. It was expected that the obtained recovery method can be refined and integrated into the existing biofilm system.

3 Last, in order to gain insight information on kinetic-metabolic nature of nisin production, the results obtained from suspended-cell system were used to construct model of nisin production (Chapter 8). The obtained models could also be used for monitoring and prediction of nisin production during fermentation.

4 CHAPTER 2

LITERATURE REVIEW

2.1. General background

Nisin is an antimicrobial peptide that is accepted worldwide as a natural and safe food preservative. It also has many other potential industrial applications varying from therapeutic uses in both animal and human to medical implants and food packaging. The commercial nisin is produced by Lactococcus lactis subsp. lactis using a suspended-cell batch fermentation with milk-based culture medium (Delves-Broughton and Friis, 1998).

Purification steps are commercially sensitive, but are suspected to include foam separation (frothing), sodium chloride precipitation, centrifugation, and spray drying

(Hirsch, 1950; Kelly et al., 2000). Both low production rate during fermentation and tedious separation processes, which involves the use of large tank volumes, eventually result in high costs of partially- and highly-purified nisin, limiting the development of potential applications. It is, therefore, essential to develop more effective production and separation methods for nisin production.

In this research, a biofilm reactor with Plastic Composite Support (PCS) was successfully developed for nisin production and several fermentation conditions have been optimized in order to improve the nisin production rate. Moreover, a recovery process using silicic acid and mathematical models that describes nisin production have been studied. This literature review not only presents background information on nisin

5 production and separation, but also focuses on related aspects such as biofilm reactor and modeling of nisin production as well.

2.2. Nisin

Nisin was first discovered in 1928 via failures of starter cultures ( Lactobacillus bulgaricus ) to clot milk during cheese making (Rogers, 1928; Rogers and Whittier,

1928). Later, it was realized that storage of milk had allowed the contaminant lactic streptococci (at present recognized as lactococci) to grow and produce the proteinaceous inhibitor (Hunter and Whitehead, 1944). The protein was characterized and called ‘nisin’ in order to coin with its serogroup, Group N Inhibitory Substance, determined by the

Lancefield serotyping scheme for streptococci (Mattick and Hirsch, 1947).

Although an initial interest in nisin focused on human and veterinary medicine, its first commercial application by Aplin and Barrett, Ltd in 1953 was to prevent Clostridial spoilage of processed cheese (Thomas et al., 2000). The potential of nisin as a natural preservative was soon realized and before long it was recognized as a safe and legal biological food preservative (FAO/WHO, 1969; US FDA, 1988). Up until now, nisin is still the only purified bacteriocin commercially available and is accepted worldwide as a food preservative.

2.2.1. Structure, solubility and stability of nisin

So far, only three natural variants of nisin were discovered; nisin A, nisin Z, and nisin Q (Fig. 2.1). The structure of nisin A, first proposed by Gross and Morell (1971), comprises of 34 amino acids with five intramolecular sulfide bridges (ring A-E). Nisin Z,

6 found to be widely distributed: 14 out of 26 nisin producers (De Vos et al., 1993), contains asparagine instead of histidine at position 27 (Mulders et al., 1991).

Interestingly, nisin Z was shown to have antimicrobial activity (Mulders et al., 1991), membrane insertion (Demel et al., 1996), and pore-formation ability (Breukink et al.,

1997) similar to nisin A. On the other hand, nisin Q was discovered recently in only one producer. It differs from nisin A in four amino acids (Ala15Val, Met21Leu, His27Asn, and Ile30Val substitution) in the mature peptide and two in the leader sequence (Pro-2Thr and Lys-8Thr substitution) (Zendo et al., 2003). Properties of nisin Q are yet to be studied.

(a)

(b)

(c)

Figure 2.1. Structure of natural nisin variants; (a) nisin A (Gross and Morell, 1971), (b) nisin Z (Mulders et al., 1991), and (c) nisin Q (Zendo et al., 2003).

7 Nisin contains four unusual amino acids; lanthionine (Lan), β-methyllanthionine

(MeLan), 2,3-dehydroalanine (Dha), and 2,3-dehydrobutyrine (Dhb), which are the results of post-translation processes. According to a scheme proposed by Ingram (1970), during post-translation of nisin, serines and threonines are dehydrated to give Dha and

Dhb, respectively. Some of the dehydrated residues then react with the thiol (--SH) group of nearby cysteine residues, forming Lan (from Dha) and MeLan (from Dhb) rings.

Totally, mature nisin has two Dha, one Dhb, one Lan, and four MeLan residues.

The presence of lanthionine rings contribute to nisin’s high level of hydrophobicity, rigid structure and thermal resistance. The N-terminal of nisin contains a high ratio of hydrophobic residues and the C-terminal is more hydrophilic (van de Ven et al., 1991). A high proportion of basic amino acids gives nisin a net positive charge

(Breukink et al., 2000). It should be noted that nisin does not contain any aromatic amino acids (Bailey and Hurst, 1971). The monomer of nisin has a molecular mass of 3,353

Daltons (Jung, 1991); however nisin normally occurs as the more stable dimer (Jarvis et al., 1968).

The solubility and stability of nisin are highly dependent on the pH of the solution. In aqueous solution, it is most soluble and stable at pH 3.2-3.3 (Davies et al.,

1998; Kelly et al., 2000) and both solubility and stability of nisin decrease at neutrality, whereas an alkali environment inactivates nisin (Hurst, 1978). The nisin produced by L. lactis subsp. lactis NIZO 22186, a nisin-producer used in this research, was reported as nisin Z (Kuipers et al ., 1991; Mulders et al ., 1991). Nisin Z has improved solubility at high pH values, compared to nisin A, due to more hydrophilic nature of asparagine compared to the deprotonated histidine (Kuipers et al ., 1991). Nisin remains stable after

8 autoclaving at 115.6 °C at pH 2.0, but loses 40% of its activity at pH 5.0 and more than

90% at pH 6.8 (Tramer, 1966). Stability of nisin also depends on several other factors such as presence of other chemical and the protective effect of proteins. Refrigerated storage of nisin for months usually gives no detectable chemical or biological changes

(Motlagh et al ., 1991).

2.2.2. Biosynthesis of nisin

Nisin is initially synthesized by ribosomes as a pre-cursor peptide, which then post-translationally modified and proteolytically cleaved to generate the mature nisin

(Schnell et al., 1988). Although the specific role of nisin for the producer strains remains undefined, it possibly provides producer strains a competitive advantage towards other lactococci (Hawley, 1955; Gross, 1977).

The complete gene cluster for nisin biosynthesis in L. lactis is located on a 70-kb conjugative transposon, which also contains gene for sucrose fermenting ability and reduced bacteriophage sensitivity (Rauch et al., 1994). The cluster contains genes encoding for nisin structure ( nisA or nisZ ), dehydration of serine and threonine ( nisB ), formation of lanthionine rings ( nisC ), secretion of nisin ( nisT ), cleavage of the leader peptide ( nisP ), nisin immunity ( nisI, nisE, nisF, and nisG ), and quorum sensing regulatory system ( nisK and nisR ). Details of organization and transcription of the nisin gene cluster have been reviewed extensively elsewhere (Rauch et al., 1994; Kleerebezem,

2004; Cheigh and Pyun, 2005).

Although a nisA -specific transcript is already present in the early exponential phase, detectable active nisin starts to appear in mid-exponential growth phase (when

9 biomass reached 50% of maximum biomass) and increased to reach its maximum at the beginning of the stationary phase (Hirsch, 1951; Buchman et al., 1988). The late appearance of active nisin was suggested to be a result of the delayed expression of the maturation enzymes (De Vuyst and Vandamme, 1992).

In general, nisin production is growth-associated and shows a cell-density- dependent characteristic (Kuipers et al., 1995). However, maximization of growth does not necessarily result in maximization of nisin production since the specific nisin production rate is not directly proportional to the specific growth rate (Kim et al ., 1998;

Cabo et al ., 2001; Bober and Demirci, 2004). For instance, Li et al . (2002) studied optimization of complex medium using response surface methodology and reported that for L. lactis ATCC 11454 higher biomass concentration did not necessarily result in higher nisin concentration. It was suggested that the non-linear relationship between nisin production and growth is probably a result of the complexity of nisin biosynthesis and regulation (Parente and Ricciardi, 1999).

2.2.3. Antimicrobial activity of nisin

Nisin exhibits antimicrobial activity against a wide range of Gram-positive bacteria, including strains or species of streptococci, staphylococci, lactobacilli, micrococci, Listeria , and most spore-forming species of Clostridium and Bacillus .

However, it shows little or no activity against Gram-negative bacteria, yeasts or molds

(O’Keeffe and Hill, 2000). Generally, Gram-negative cells are protected from nisin because of the presence of an outer membrane in cell wall, but when the outer membrane

10 is weakened by chelating agents (such as EDTA) the cell then becomes sensitive to nisin

(Stevens et al ., 1991).

Interestingly, nisin effectively kill bacteria in nanomolar concentrations, while many non-lantibiotic pore-forming antimicrobial peptides only work in the micromolar range. This remarkable potency of nisin was suggested as a result of its dual killing mechanism; targeted pore formation and inhibition of cell wall synthesis (Brotz and Sahl,

2000; Wiedemann et al., 2001).

A specific target of nisin for pore formation is an integral component of the eubacterial cytoplasmis membrane, called Lipid II. According to the model proposed by

Hasper et al. (2004), nisin first form a 1:1 complex with Lipid II before one more nisin molecule binds to Lipid II, giving a 2:1 nisin:Lipid II complex. Finally, a staple pore complex is formed by the insertion of nisin molecules into a perpendicular orientation with respect to the membrane surface, giving final pore complex of 8 nisin and 4 Lipid II molecules with a diameter of 2-2.5 nm (Wiedemann et al., 2004). Binding of nisin onto

Lipid II not only causes collapses of the proton motive force and membrane integrity via pore formation, but also interferes with cell-wall synthesis by blocking lipid II from incorporation into peptidoglycan (Wiedemann et al., 2001).

In case of bacterial spore, effect of nisin is known to be different from that against vegetative cell and is mostly sporostatic. Nisin affects the post-germination stages of spore development by inhibiting pre-emergent swelling, the out-growth and formation of vegetative cells (Hitchins et al., 1963; Gould, 1964).

11 2.2.4. Nisin activity and quantification techniques

The activity of nisin is expressed in terms of International Units (IU): 1 g of pure nisin is usually equivalent to 40x10 6 IU and 1 g of Nisaplin®, commercial nisin reference by Aplin and Barrett Ltd, is equivalent to 1x10 6 IU. Another unit used to express the activity of nisin is an Activity Unit (AU), which is defined as the reciprocal of the highest dilution giving growth inhibition of the indicator cells. Techniques used for quantification of nisin can be divided into four main categories;

1) Techniques based on nisin’s inhibitory activity to a test organism, such as

Turbidimetry assay (Berridge and Barrett, 1952; Cabo et al., 1999); Microtitration method (Turcotte et al., 2004); Agar well diffusion assay (Tramer and Fowler, 1964;

Wolf and Gibbons, 1996); Critical dilution techniques (Hirsch, 1950); ATP- bioluminescence-based method (Valat et al., 2003); and Flow cytometry (Budde and

Rasch, 2001).

These techniques directly measure the antimicrobial activity of nisin on the nisin- sensitive strains. The nisin activity is highly affected by many factors, e.g. indicator microorganism (strain and growth phase), compositions of culture medium and sample

(salt, fat, protein, agar, surfactant, etc.), pH and temperature; thus, all these factors need to be strictly controlled when using these techniques. At the same concentration, nisin Z resulted in a larger inhibition zone in the agar well diffusion method due to its higher hydrophilic nature. In order to use the commercial nisin standard (nisin A from Aplin &

Barret Ltd or Sigma) for quantification of nisin Z, the concentration of nisin in the tested sample and standard should not exceed 100 IU/ml (de Vos et al., 1993). The interference from sample compositions (e.g. organic acid, salt, etc.) can be tested by the use of nisin-

12 inactivated sample (heat at 60 °C at pH 11) and then compromised by including those components into the diluent for preparing standard (Tramer and Fowler, 1964).

Agar diffusion assay (Tramer and Fowler, 1964) is the most widely used nisin quantification assay because of its simplicity, high sensitivity, and cost effectiveness. The method involves the inoculation of nisin-sensitive strain, used as an indicator organism, into the bioassay agar. The sample is placed in a well that is made on the agar. After 24 h incubation, an inhibition zone is created by the antimicrobial effect of nisin. The diameter of the zone is measured and correlated with a standard curve plotted between diameter of the zone and log 10 of nisin concentration in standard. Tramer and Fowler (1964) used

Tween 20 to enhance nisin diffusion through the agar with Micrococcus flavus (at present recognized as Micrococcus luteus ) as an indicator organism. Then, Wolf and Gibbon

(1996) improved the method further by reducing agar level from 1.5% to 0.75% and obtained an increase in sensitivity by 21%. Roger and Montville (1991) reported that the assay sensitivity (increase in zone size/increase in nisin concentration) was greatly increased when using Lactobacillus sakei as the indicator organism. The same study also revealed that prediffusion of plates at 3 °C for 24 h increased sensitivity further and resulted in maximum assay reproducibility.

The agar well diffusion assay is without a doubt very sensitive (0.5 IU/ml or

0.0125 µg/ml) considering its simplicity and cost-effectiveness. However, the method is time-consuming (require at least 2-3 days to complete) and laborious. Turbidimetry assay and microtitration method require less time (4-6 h), but may require specific equipment such as microplate reader. Moreover, both techniques are not appropriate for analysis of opaque food extracted sample. The more recent techniques like ATP-bioluminescence-

13 based method (Valat et al., 2003), based on the release of adenylic-nucleotides by a sensitive strain under the action of nisin, gave a higher detection limit (20 IU/ml or 0.5

µg/ml). Flow cytometry method (Budde and Rasch, 2001), based on the leakage of carboxyfluorescein from the indicator strain caused by exposure to nisin, gave an even higher detection limit of more than 200 IU/ml.

2) Immunological techniques, such as sandwich-type ELISA based on sheep polyclonal antibodies (pAb) for nisin A (Falahee et al., 1990); competitive direct ELISA using mice pAb (Saurez et al., 1996a) and monoclonal antibodies (mAb) (Saurez et al.,

1996b) against nisin A; immunodot detection using rabbit-pAb (Bouksaim et al., 1998); competitive direct ELISA based on mAb against nisin Z (Daoudi et al., 2001); and direct

ELISA with rabbit pAb for nisin A (Leung et al., 2002).

Although immunological techniques are more rapid (requires 5-6 h) and highly sensitive - detection limit of at least 3 IU/ml or 0.078 µg/ml (Daoudi et al., 2001) and in some case as low as 0.12 IU/ml or 0.003 µg/ml (Bouksaim et al., 1998) - compared to the standard methods like agar well diffusion assay, the obtained data is not directly related with nisin’s antimicrobial activity. Previous study showed that the antibodies could not distinguish between nisin A and nisin Z (Bouksaim et al., 1998) and even cross-reacted with a related compound, such as subtilin (Falahee and Adams, 1992) and degraded or inactive nisin (Daoudi et al., 2001). While the cross-reaction with other variant of nisin can be solved by carefully developed specific mAb, the cross-reaction with inactive nisin was proved to be more challenging.

14 3) Techniques based on the autoinducibility of the nisin promoter, such as methods in which nisin promoter P nisF was fused with bioluminescence luciferase gene

(Wahlstrom and Saris, 1999) and green fluorescent protein (GFP) gene (Reunanen and

Saris, 2003). Both techniques gave a comparably low detection limit as immunological techniques, but nisin A seemed to exhibit higher inducing activity than its variant.

Although the luciferase assay was the most sensitive quantification reported (0.0005

IU/ml), it demanded an exact time for each step of analysis and limited the number of the samples that can be analyzed each time. Both techniques require a development and maintenance of cloned indicator strain.

4) Techniques based on nisin physical properties, such as Reverse Phase-HPLC

(Matsusaki et al., 1996a; Pfeiffer and Orben, 1997) and capillary zonal electrophoresis

(Rossano et al., 1998). Both techniques are fast, but active and inactive nisin cannot be distinguished.

2.2.5. Applications of nisin

Nisin was confirmed as Generally Recognized as Safe (GRAS) by the U.S. Food and Drug Administration (FDA, 1988) and is allowed for using as food preservative in about fifty countries around the world (Delves-Broughton, 1990). The suitability of nisin as a food preservative arises from many good characteristics such as non-toxicity and heat stable at low pH values. Moreover, nisin is not used clinically, digested by enzyme

α-chymotrypsin, and it does not add color or flavor to the food.

The principal commercial application of nisin are in foods and beverages which are pasteurized but not fully sterilized, such as processed cheese, milk, clotted cream,

15 dairy desserts, ice cream mixes, liquid egg, and hot-baked flour products such as crumpets and potato cakes. It is also used in canned products to prevent spoilage by thermophilic, heat-resistant spore formers. Another application includes alcoholic beverages, such as beer and wine, in which it is used to control spoilage due to lactic acid bacteria (Davies and Delves-Broughton, 2000). In meat products, nisin acted effectively in inhibiting Brochothrix thermosphacta when incorporated in a cold meat-binding system (Cutter and Siragusa, 1998) and was more effective to inhibit Gram-negative bacteria when used together with lactic acid (Ariyapitipun et al ., 1999; 2000). Examples of effective use of nisin in food systems are presented in Table 2.1.

Table 2.1. Examples of effective use of nisin in food.

Effective nisin Food Products Target organisms Concentration References (IU/ml or IU/g) Canned beans C. thermosaccharolyticum 200 Gillespy, 1957 in tomato sauce Beer Lactobacilli and Pediococci 100 Ogden, 1986 Fermenting L. plantarum 1,400 Henning et al ., 1986 apple mashes Wine Leuconostoc and Pediococcus 100 – 1,000 Radler, 1990a; 1990b Cottage cheese L. monocytogenes 2,000 Ferreira and Lund, 1996 Ricotta cheese L. monocytogenes 100 Davies et al ., 1997 Skim milk B. cereus spores 4,000 Wandling et al ., 1999 Cooked ham C. sporogenes 3,000 Rayman et al ., 1981 Bologna sausage L. sakei and L. curvatus 1,000 Davies et al ., 1999 Lean beef B. thermospacta 400 Cutter and Siragusa, 1998 Kimchi Lactobacilli 100 Choi and Park, 2000

16 In hurdle technology, combining of different preservation methods to inhibit microbial growth more effectively, nisin is used together with other treatments to create the most effective combination of treatments (Cleveland et al ., 2001). For example, nisin has been used with the application of pulsed electric field (PEF), which increases the permeability of cell membranes (Pol et al ., 2000; Terebiznik et al ., 2000). It was used in combination with metal chelators such as EDTA, which disrupts the outer membrane and allows the penetration of nisin, to control the growth of Gram-negative pathogens such as

E. coli O157:H7 and Salmonella (Zhang and Mustapha, 1999). More examples of application of nisin as a part of hurdle technology are presented in Table 2.2.

Table 2.2. Applications of nisin as a part of hurdle technology.

Other factors combined Target organisms References with nisin Nitrite C. sporogenes Rayman et al ., 1981 Nitrite C. perfringens Caserio et al ., 1979a; b Nitrite C. botulinum spore Taylor et al ., 1984

N2, CO 2, low temperature L. monocytogenes Szabo and Cahill, 1998 Milk lactoperoxidase (LP) and L. monocytogenes Rodriguez et al ., 1997 Low temperature L. mocytogenes, B. cereus, Sucrose fatty acid esters Thomas et al ., 1998 L. plantarum and S. aureus Carbon dioxide L. monocytogenes Nilsson et al ., 2000 Pulsed electric field (PEF) B. cereus Pol et al ., 2000 Pulsed electric field (PEF) E. coli Terebiznik et al ., 2000 Modified atmosphere packaging L. monocytogenes Fang and Lin, 1994

Apart from the most important application as a food preservative, nisin is now widely used for prevention of bovine mastitis (Sears et al ., 1991). Moreover, nisin was found to be very effective in preventing oral plaque and gingivitis. Thus, nisin shows a potential

17 for many dental applications (Howell et al ., 1993). With all these applications, demand for nisin keeps increasing and thus there is a need for economical production of nisin.

2.3. Nisin production

It is known that nisin production is influenced by many cultural factors such as producer strain, compositions of the nutrient broth, pH, temperature, and aeration

(Parente and Ricciardi, 1999). Moreover, nisin production is also affected and limited by its challenging characteristics such as substrate inhibition, adsorption of nisin onto the producer cells, and enzymatic degradation (De Vuyst, 1992; Yang et al., 1992). Mattick and Hirsch (1947) first described a method for large-scale production of nisin using medium based on glucose and yeast extract, and they obtained a yield of only 80 IU/ml.

Since then, many aspects of nisin production have been studied extensively in order to improve the production rate and productivity.

2.3.1. Producer strain

In 1994, De Vuyst and Vandamme screened 21 nisin-producing and 6 non- producing strains of L. lactis for nisin production and immunity. They found that differences between strains were attributed to expression level and activity of maturing enzymes and nisin immunity, but not to the number of nisin structural genes ( nisA and nisZ ) or their transcription or translation levels. In the same study, L. lactis strain NIZO

22186 was reported as the best nisin producer (1,886 IU/ml) among the tested strain.

Kim et al. (1997b) showed that the maximum level of nisin produced seemed to be host-dependent. Later, Kim et al. (1998) found that different nisin producers produced

18 nisin to different ceiling concentrations and once the host-specific ceiling concentration of nisin was reached then the nisin production stopped, even if the producing strain continued to grow well. On the other hand, when the same host carried transposon from different donors the concentrations of nisin produced were similar. The end-product inhibition was believed to be the reason that the strains could not produce nisin beyond their ceiling concentrations since increase in nisin resistance by introduction of plasmids containing nisin immunity determinants has been shown to result in higher nisin production and faster growth rate.

2.3.2. Nutrients

Lactococci are nutritionally fastidious microorganisms and they grow well in milk or in the usual complex organic medium (De Vuyst and Vandamme, 1994). Therefore, nisin production is highly affected by type and level of carbon, nitrogen and phosphate sources, cations, and inhibitors.

2.3.2.1. Carbon source

Both type and concentration of carbon sources play an important role in nisin production. Although nisin can be produced from many types of carbon sources, different producer strains showed different preferences on sugar consumption, for example, glucose was reported as an optimal carbon source for nisin production in L. lactis strain

ATCC 11454 (Chandrapati and O’Sullivan, 1998), while sucrose, xylose and lactose were reported as the most efficient carbon sources in strain LM 0230 (Yu et al ., 2002),

19 strain JCM 7638 (also known as IO-1) (Chinachoti et al., 1997a) and strain A 164

(Cheigh et al., 2005), respectively.

Hengstenberg (1977) suggested that sucrose was rapidly utilized by the nisin producer strain because of its highly efficient phosphoenolpyruvate-dependent phosphotransferase system (PTS) for sucrose uptake, transport, and metabolism. It was shown that three following sucrose-specific proteins were induced when cells were grown on sucrose: a sucrose-specific uptake protein Enzyme II, a sucrose 6-phosphate hydrolase and a fructokinase (Thompson and Chassy, 1981; Thompson et al., 1991). On the other hand, in case of strain A 164, lactose was shown to induce transcription of the nisZ pre-peptide gene, thus resulting in the highest nisin production among the tested carbon sources, including sucrose (Cheigh et al., 2005).

Carbon source regulation has been reported to have a profound effect on both growth of L. lactis and nisin production. De Vuyst and Vandamme (1992) reported a maximum nisin titer of 3,267 IU/ml at the initial sucrose concentration of 40 g/l. Even if more biomass were produced, nisin yield (Y p/x ) decreased as sucrose concentration increased from 10 to 40 g/l. Nisin production clearly showed a substrate inhibition at high levels of sugar as carbon source. A sucrose concentration higher than 40 g/l resulted in decrease of both biomass and maximum nisin titre (De Vuyst and Vandamme, 1992;

Matsusaki et al., 1996b; Cheigh et al., 2002). Therefore, fed-batch culture was proposed as a better alternative.

Although the information of nisin production in fed-batch culture is still scarce, a few studies have revealed the advantage of fed-batch fermentation on nisin production in suspended-cell reactor. Lv et al. (2004) reported a significant improvement on nisin

20 production of L. lactis ATCC 11454 when a fed-batch fermentation with a slow feeding of sucrose was used (3,887 IU/ml, 64% higher than a batch fermentation). Further improvement could be achieved when the sucrose level was maintained at a lower concentration, for instance, obtained nisin titer were 4,961 IU/ml (for 2 g/l) and 3,498

(for 20 g/l).

Though the results from De Vuyst and Vandamme’s study suggested that the higher initial concentration of sucrose helped maintaining the nisin titer during longer fermentation runs, the study of fed-batch fermentation in biofilm reactor (Bober and

Demirci, 2004) clearly showed that the decrease of nisin titer in stationary phase was not a result of carbon source depletion.

2.3.2.2. Nitrogen source

Growth and nisin production are often limited by organic nitrogen sources rather than by the carbon substrate. De Vuyst and Vandamme (1993) studied the effect of organic nitrogen sources at 10 g/l on nisin production by L. lactis NIZO 22186 in a complex medium. They found that the highest nisin titer (2,500 IU/ml) was obtained with cotton-seed meal, but high nisin yields (more than 2,000 IU/ml) were also obtained with yeast extract and fish meal. They suggested that slow-metabolisable nutrients, such as cotton-seed meal, resulted in a slow metabolism and a low specific growth rate, which seemed to favor nisin biosynthesis. Thus, a slow feeding rate of the nitrogen source could also prolong the nisin production phase. Guerra and Pastrana (2001) found that the supplementation of whey with yeast extract and casitone increased nisin production, whereas ammonium chloride and glycine did not improve nisin production. Kim et al .

21 (1997a) reported that maximum nisin concentration increased with increasing organic acid content in the medium.

2.3.2.3. Mineral source

Both anions (phosphate) and cations (Mg 2+ and Ca 2+ ) affect nisin production, but their effects may vary between strains. In batch fermentation using L. lactis NIZO 22186 at pH 6.8, KH 2PO 4 at concentration of 50 g/l was shown to yield the best result of 3,500

IU/ml (De Vuyst and Vandamme, 1993). However, the result showed a comparably high nisin concentration (3,000 IU/ml) and a 4 h shorter lag phase when using 30 g/l KH 2PO 4.

The fact that nisin formation is not negatively influenced by high phosphate concentrations is an indication of the primary metabolite behavior of nisin. Kozlova et al.

(1979) suggested that K 2HPO 4 was essential not only as a buffering agent, but also as a source of phosphate ions, which were required for growth. However, K 2HPO 4 did not improve nisin production for strain IO-1 (Matsusaki et al ., 1996b). Recently, Li et al .

(2002) reported that the concentration of K 2HPO 4 strongly affected nisin production of L. lactis ATCC 11454 and a higher level of K 2HPO 4 (15 g/l) resulted in higher nisin concentration than a lower level (5 g/l).

Meghrous et al . (1992) reported that addition of MgSO 4•7H 2O at 0.25 g/l level improved nisin production (from 157 to 238 AU/ml) and reduced cell-adhered nisin

(from 76 to 31%) in the strain ATCC 11454. However, addition of magnesium ion did not affect nisin production in strain IO-1 (Matsusaki et al ., 1996b). However, in strain IO-

1, addition of 0.1 M CaCl 2 resulted in an increase of maximum nisin Z concentration

(from 2,100 to 3,150 IU/ml), even though it did not affect growth on glucose and xylose

22 in batch fermentations with controlled pH. Matsusaki et al. (1996b) suggested that calcium ion showed a stimulating effect on nisin Z production as it may stimulate the synthesis of the prepeptide or the activation of the prepeptide maturation enzymes, and transport of nisin out of the cell. They also suggested that calcium ion might activate the leader peptidase, raise the immunity of producer cells by protecting the lipid membrane or displace nisin adsorbed on cell wall of producer cells.

2.3.3. pH and temperature

The optimal pH for nisin production is usually at 5.5-6.0 (Meghrous et al ., 1992; and Matsusaki et al ., 1996b), which is lower than optimal pH for growth. Optimal pH may also be affected by the culture medium; for instance, nisin Z production by strain IO-

1 was optimal at pH 6.0 in xylose medium (Chinachoti et al ., 1997a) and at pH 5.5 in glucose medium (Matsusaki et al ., 1996b).

Generally, dropping of pH below the optimal value for nisin production due to lactic acid production resulted in lower level of maximum nisin titer in uncontrolled pH fermentation. De Vuyst and Vandamme (1992) reported that batch fermentations of L. lactis subsp. lactis NIZO 22186 with 1% sucrose gave the maximum nisin titer of 1500 and 1800 IU/ml for uncontrolled-pH and controlled pH at 6.8, respectively.

Only recently that studies indicated a positive effect of pH-drop gradient. A study from Cabo et al . (2001) on the effects of pH on nisin production by L. lactis strain IIM

Lb.1.13 showed that a pH-drop gradient enhanced nisin production approximately four- fold when pH of fermentation broth was adjusted back to 7.0 every 6 h. Moreover, the pH-drop gradient increased the efficiency of nutrient consumption of the microorganism.

23 A reduction of re-alkalization period from 6 to 3 h resulted in a doubling in nisin production (110 AU/ml). A model developed in the same study enabled nisin to be characterized as a primary metabolite, which tended to move to a secondary nature when the productive period of the culture was prolonged by imposing a stepwise-pH profile.

Although the optimal temperature for growth of L. lactis was reported at both

30 °C, for strain ATCC 11454 (Meghrous et al ., 1992), and 37 °C, for strain IO-1

(Matsusake et al ., 1996) and strain A164 (Cheigh et al , 2002), the optimal temperature for nisin production was reported only at 30 °C.

2.3.4. Agitation and aeration

Nisin production typically does not require aeration and agitation. A slow agitation is needed to achieve a homogeneous suspension (De Vuyst and Vandamme,

1994). However, many studies also reported significant effects of agitation and aeration on nisin production. In fermentations of L.lactis strain IO-1 using glucose-based medium at pH 5.5, maximum nisin Z concentration (3,940 IU/ml) was obtained at 320 rpm and only a small decrease of nisin (3,410 IU/ml) was obtained at 1,000 rpm (Chinachoti et al.,

1997c). On the other hand, agitation beyond 540 rpm resulted in inhibition of growth and nisin production in xylose media for the same strain (Chinachoti et al ., 1997a). Whereas some studies suggested no aeration and moderate agitation (De Vuyst and Vandamme,

1994) or even an aerobic condition (Hirsch, 1951) because of decreasing nisin production during aeration, some others found that an oxygen-enriched atmosphere enhanced nisin production. Amiali et al . (1998) reported that an initial dissolved oxygen level of 60% air saturation resulted in higher nisin Z production (4,132 IU/ml) than at 0% (548 IU/ml) in

24 strain UL719. Another study from Cabo et al . (2001) showed that when the oxygen saturation percentage increased from 50 to 100% nisin production increased 4-fold in L. lactis strain IIM Lb. 1.13. It was suggested that the discrepancy was due to variations among different strains of nisin producers (Cabo et al., 2001).

2.3.5. Types of fermentation

As nisin production is both growth-associated and biomass density-dependent

(Buchman et al., 1988; De Vuyst and Vandamme, 1994; Parente and Ricciardi, 1999), several systems using cell recycle or cell-immobilized reactor in order to increase cell density in the reactor have been tested. Nisin production rate in various types of fermentation are summarized in Table 2.3.

Continuous fermentation coupled with cell recycle with L. lactis IFO 12007 resulted in an increase of nisin titer and volumetric nisin productivity compared to batch fermentations (Taniguchi et al ., 1994). Chinachoti et al . (1997b) used continuous culture with a ceramic membrane to produce nisin Z at high dilution rates and found a slight improvement in productivity compared to batch or continuous fermentation without cell- recycle. Wan et al . (1995) studied nisin production with suspended cells and Ca-alginate immobilized cells. They found that in repeated batch fermentation the immobilized cells produced less nisin than the suspended cells, perhaps because of diffusional limitation in alginate gels. However, the same study also showed that, in continuous fermentation at a

25 Table 2.3. Summary of nisin production in various types of fermentation.

Strain of Process Nisin References L. lactis Concentraion NIZO 22186 SC: Batch, pH = 6.8 3,267 IU/ml De Vuyst and Vandamme (1992) BF: Batch, pH = 6.8 948 IU/ml Bober and Demirci (2004) ATCC 11454 SC: Continuous, D = 0.1 8 x 10 3 AU/l Meghrous et al . (1992) SC: Continuous, D = 0.25 18 x 10 3 AU/l SC: Continuous, D = 0.4 6 x 10 3 AU/l SC: Batch 1,590 IU/ml Van’t Hul and Gibbons (1997) SC: Batch + Lactate removal column 1,500 IU/ml SC: Batch, pH = 6 2,320 IU/ml Shimizu et al. (1999) SC: Batch + Mixed culture with 3,920 IU/ml Kluyveromyces marxianus , pH = 6 SC: Batch, Uncontrolled pH 603 IU/ml Li et al . (2000) SC: Two-phase culture, Uncontrolled pH 803 IU/ml SC: Batch, pH = 6.8 2,371 IU/ml Lv et al. (2004) SC: Fed-batch, pH = 6.8 3,887 IU/ml IFO 12007 SC: Batch 0.12 x 10 6 AU/l Taniguchi et al . (1994) MF: Continuous, D = 0.5 0.15 x 10 6 AU/l AFISC 2011 SC: Batch 125 IU/ml Wan et al . (1995) IC (Ca-alginate): Continuous, D = 0.1 IO-1 SC: Batch, pH = 5.5 3,150 IU/ml Matsusaki et al . (1996b) SC: Continuous , D = 0.1 2.81 x 10 6 AU/l Chinachoti et al . (1997b) IC (ENTG-3800): Continuous, D = 0.1 2.16 x 10 6 AU/l IC (ENTG-3800): Continuous, D = 0.3 1.30 x 10 6 AU/l IIM Lb. 1.13 SC: Batch 0.30 x 10 5 AU/l Cabo et al . (2001) SC: Batch + Re-alkalization (pH stepwise 0.66 x 10 5 AU/l profile) UL719 SC: Batch, non-aerated 4,230 IU/ml Amiali et al . (1998) SC: Batch, aerated 20,480 IU/ml IC: Repeated-batch (1-h cycles) 8,190 IU/ml Bertrand et al . (2001) IC: Repeated-batch (2-h cycles) 10,240 IU/ml SC: Continuous, non-aerated, D = 0.15 1,493 IU/ml Desjardins et al . (2002) SC: Continuous, non-aerated, D = 0.25 1,152 IU/ml SC: Continuous, aerated, D = 0.15 2,560 IU/ml SC: Continuous, aerated, D = 0.25 1,536 IU/ml IC: Continuous, non-aerated, D = 0.25 1,088 IU/ml IC: Continuous, aerated, D = 0.25 2,432 IU/ml D = dilution rate (1/h); SC = Suspended-cell; BF = Biofilm; MC = Microfiltration; IC = Immobilized-cell

26 high dilution rate (3 1/h), the nisin titer from immobilized cells was equal to that obtained from batch fermentation.

Apart from increasing cell density in the fermentor, immobilized cells are used to carry out stable continuous cultures at high dilution rates with a high nisin volumetric productivity (Sonomoto et al ., 2000; Bertrand et al ., 2001; and Desjardins et al ., 2002).

However, maximum nisin production using immobilized cell reactor with continuous fermentation was 5-10 folds lower than that of suspended-cell reactor with batch fermentation for the same strain, pH, temperature and medium (Amiali et al ., 1998;

Sonomoto et al ., 2000; and Desjardins et al ., 2002). Desjardins et al . (2002) suggested that nisin production in steady state condition of continuous culture was limited by biosynthesis steps of the mature peptides for post-translational reactions, transport, or maturation. This idea is supported by results from a study of nisin Z production in pH- controlled, repeated-batch fermentation (RCB) using L. lactis strain UL719 immobilized in κ-carrageenan/locust bean gum gel beads in supplemented whey permeate (Bertrand et al ., 2001). It was reported that RCB resulted in higher nisin productivity (5,730 IU/ml/h) compared to suspended-cell batch culture (850 IU/ml/h), suspended-cell continuous culture (460 IU/ml/h), and immobilized-cell continuous culture (1,760 IU/ml/h), using the same strain and fermentation conditions. Bertrand et al. (2001) suggested that the evolutionary conditions produced during batch cultures are required for high nisin production.

So far, there is only one study on nisin production using a biofilm reactor (Bober and Demirci, 2004). The results indicated no significant difference in nisin production or productivity rate when comparing to a suspended-cell reactor. However, an inadequate

27 mixing in the biofilm reactor was suspected. Bober and Demirci (2004) mentioned that slow rate of agitation (75 rpm) used in PCS biofilm reactor, in order to prevent disruption of PCS by the magnetic impellor, might not be adequate to provide a thorough dispersion of nisin produced. Inadequate mixing resulted in unfavorable microenvironments surrounding the biofilm. Moreover, the collected samples might not represent real concentrations of nisin in the system. Nevertheless, this problem may be overcome by the use of a mechanically driven impellor and a better orientation of PCS, such as a grid-like orientation of PCS tubes suggested by Cotton et al. (2001) which allows the use of higher agitation rate in the reactor.

2.4. Biofilm reactor

One way to increase the productivity of fermentation is to increase biomass in the reactor by using techniques such as cell-recycle reactors, hollow-fiber reactors, and cell immobilization. Immobilized-cell reactors like biofilm reactors are excellent examples of high-biomass density systems with lesser tendencies to develop membrane fouling and lower required capital costs.

In general, cell immobilization techniques can be divided into two main categories: 1) active or artificial immobilization, and 2) passive or natural immobilization

(Fukuda, 1995). Active immobilizations, including covalent bonding to surfaces using various coupling agents and entrapments in polymer matrix, can be achieved by chemical agents, while passive immobilization occurs via natural adsorptions and colonization of microbial films around or within the solid support materials. Several disadvantages of active immobilization include 1) toxicity of coupling agents and cross-linking agents on

28 cell viability and activity; 2) instability of the polymer matrix (e.g. calcium alginate gel) with various anions including phosphate, citrate, EDTA, and lactate; 3) cell leakage from the gel matrix; 4) limited mass transfer across the beads; 5) poor operational stability; and

6) high cost of the carrier.

Biofilm reactors show many advantages over suspended cell reactors, especially in their higher biomass density and operation stability. Biofilm reactors are able to retain

5 or 10 times more biomass per unit volume of reactor, increasing production rates, reducing the risk of washing out when operating at high dilution rates during continuous fermentation, and eliminating need for re-inoculation during repeated-batch fermentation

(Fukuda, 1995). Moreover, the structure of the biofilm matrix contributes to high resistance of microorganisms to extreme conditions of pH and temperature, contaminations, hydraulic shocks, antibiotics, and toxic substances (White, 1984;

Norwood and Gilmour, 2000). Finally, products from biofilm reactor can be easily recovered (in comparison to product from suspended cell systems, for instance), resulting in more efficient downstream processes.

2.4.1. Biofilm formation

Biofilm is an accumulation of cells embedded in an organic polymer matrix of microbial origin (Characklis and Marshall, 1990). More than 90% of wet biofilm mass is water, while the extracellular polymeric substances (EPS), containing polysaccharides and glycoproteins, corresponds to ≥70% of the dry biofilm mass (Melo and Oliveira,

2001). Biofilm thickness can vary from a few microns to even a few centimeters

29 depending on several factors such as microbial species, biofilm age, available nutrients, and liquid shear stresses.

A formation of biofilm takes several steps (Bryers, 2000; Busscher and van der

Mei, 2000) as shown in Fig. 2.2. First, the conditioning of the substratum is formed either by macromolecules in bulk liquid or intentionally coated material (step1). Then, microorganisms suspended in the liquid are transported to the surface by diffusion, convection, or self-motility (step 2) before forming weak reversible adhesions with the solid surface (steps 3 and 4). Later, irreversible adhesions take place as a result of formations of polymer bridges between the conditioning layer (an adsorbed layer of macromolecules on the solid surface) and the EPS excreted by the microbes (step 5).

Details of the adhesion mechanism of microbial cells onto the support surface were reviewed in many literatures (Busscher et al., 1995; Azeredo et al., 1999; Teixeira and

Oliveira, 1999; Bryers, 2000).

Figure 2.2. Formation of biofilm (Bryers, 2000).

30 At this step, a coadhesion -- a process in which the secondary colonizers coadhere with organisms already adhering to the surface -- occurs simultaneously with the initial adhesion. Coaggregates of organisms may also form in bulk liquid and then adhere to the biofilm surface as well. After these initial steps, the growth of microorganisms (step 6), which depends on substrate availability, plays a more important role in biofilm build up than the transport and adhesion of microorganisms to the biofilm (Bott, 1995).

Detachment processes (step 7), erosion or sloughing off, occur simultaneously and in response to fluid shear forces, weak internal cohesion, and depletion of nutrients or oxygen supplies in the biofilm. Erosion, a continuous process resulted by liquid shear forces, is the removal of small portions of biofilm. Sloughing, in contrast, is the random detachment of large portions of the biofilm as a result of rapid change or depletion of nutrients (Heukelelian and Crosby, 1956; Howell and Atkinson, 1976). Finally, when growth balances with detachment, the maximum average thickness of biofilm is reached and the system is considered as pseudo-steady state. Although the condition inside the biofilm may be different from the liquid outside, biofilm formation appears to be favored when the temperature and pH of the outside liquid approach the optimum values for microbial growth (Melo and Oliveira, 2001). Overall, biofilm formation and structure is affected by several factors as summarized in Fig. 2.3.

31

Figure 2.3. Factors affecting biofilm formation and structure.

2.4.2. Types of biofilm reactors

In general, biofilm reactors can be categorized into two categories: fixed bed and expanded bed reactors. Fixed bed reactors, which include all processes in which the biofilms develop on static media, can be divided into 1) submerged beds in which the biofilm particles are completely immersed in the liquid; 2) trickling filters in which the liquid flows downward through the biofilm bed, while the gas flows upward; 3) rotating disk reactors in which the biofilm develops on the surface of a vertical disk that is partially submerged and rotates within the liquid; and 4) membrane biofilm reactors which the microbial layer is attached to a porous gas-permeable membrane.

32 Expanded bed reactors, by contrast, which include all biofilm processes with continuously moving media maintained by high air or liquid velocity or by mechanical stirring, can be divided into 1) fluidized beds in which particles move up and down within the expanded bed in the well defined zone of the reactor; and 2) moving beds in which the whole expanded bed circulates throughout the reactors such as air-lift reactor and circulating bed reactors. In expanded beds, the bed is usually expanded by the upward flow of liquid and gas bubbles. Detailed descriptions and comparative analysis of the advantages and disadvantages of various types of biofilm reactors can be found in many publications (Cabral and Tramper, 1994; Fukuda, 1995; Willaert et al., 1996;

Lazarova and Manem, 2000; Melo and Oliveira, 2001).

2.4.3. Solid support for biofilm

In a selection of solid supports for biofilm reactors, the support should meet several requirements; 1) favorable to microorganism adhesion; 2) high mechanical resistance; 3) inexpensive; and 4) widely available. Properties of solid support, such as surface charge, hydrophobicity, porosity, roughness, particle diameter, and density, dramatically affect the adhesion of microorganisms. A great variety of solid supports have been developed and designed in order to increase the specific surface area per volume of reactor so that the reactor achieves higher efficiency and compactness.

The interaction between microorganism cells and solid support is a result of a balance between the van der Waals forces of attraction and repulsive forces (Oliveira,

1992). Generally, the bacterial cell surfaces and most of the existing solid support materials display a net negative charge when immersed in aqueous solution with pH near

33 neutrality (Melo and Oliveira, 2001). Studies on selections of support materials for different types of biofilm reactors and microorganisms suggested that the higher degree of hydrophobicity of solid surfaces strongly enhances adhesions of microorganisms

(Sousa et al., 1997; Teixeira and Oliveira, 1999; Pereira et al., 2000). However, Ho et al.

(1997a) reported that Lactobacillus casei , which was considered to be hydrophilic, was more readily attached to the less hydrophobic support materials. The leaching of nutrients can compensate for the hydrophobic nature of solid supports in some cases, resulting in a higher biofilm population in more hydrophobic support materials (Ho et al., 1997b).

To achieve a larger surface area on solid support, a smaller diameter of carrier particles and rough and/or porous surface materials are used. Porous matrix of materials also provides niches sheltered from hydraulic shear forces (Bryers, 1987; Massol-Deya et al., 1995; Ho et al., 1997a). In the porous particle, the biofilm is formed not only on the surface, but also within the pores. Moreover, the problems of deficient nutrient diffusion to the inner area and accumulation of gaseous metabolites inside porous carriers, resulting in carrier washout, can be overcome by using materials with adequately large pores and internal porous volume (Melo and Oliveira, 2001).

2.4.4. Plastic Composite Support (PCS)

Plastic Composite Support (PCS), developed at Iowa State University (U.S.

Patent Number: 5,535,893), is an extrusion solid support made from polypropylene and several agricultural products. Polypropylene acts as a matrix to integrate the mixture of agricultural products, which provide essential nutrients to sustain cell growth. Therefore,

PCS not only provides ideal physical structure for biofilm formation, but also slowly

34 releases nutrients for the microorganism. Furthermore, nutrients compositions of PCS can be customarily adjusted to meet the specific requirement of the desire microorganism.

PCS is manufactured in a form of an extrusion tube with a wall thickness of 3.5 mm and an outer diameter of 10.5 mm (Fig. 2.4). It can be cut into desired dimension, e.g. 10-cm length tubes (PCS tubes) or 3-mm slices (PCS rings) (Ho et al., 1997a). The

PCS rings can be used as solid supports for biofilm formation in a packed-bed reactor

(Ho et al., 1997) while PCS tubes can be attached to the agitator shaft of the reactor in a grid-like fashion (Fig. 2.5) in order to obtain a better mixing (Cotton et al., 2001).

Figure 2.4. Dimensions of the Plastic Composite Support or PCS (Not to scale).

Agitator shaft

Packed-bed reactor Biofilm reactor with PCS rings with PCS tubes

Figure 2.5. PCS biofilm reactors.

35 PCS was used to enhance ethanol and lactic acid production in repeated-batch and continuous fermentations while lowering the nitrogen requirement of the medium

(Demirci et al ., 1997; Ho et al., 1997c; Cotton et al., 2001). Other advantages of using

PCS in fermentation are its longevity and durability for long-run fermentation due to its strength and slow nutrient release characteristics (Ho et al ., 1997b).

2.5. Adsorption and degradation of nisin

In nisin fermentation, a decrease in nisin level after reaching the peak value might come from proteolytic degradation or adsorption of nisin by producer cells. The loss of nisin activity after the active growth phase was suggested as a result of non-specific proteases released during cell lysis since most nisin producers are lysogenic (Kozak et al .,

1973). A specific nisinase released from the producer cells was also suspected (De Vuyst and Vandamme, 1992).

Adsorption of nisin by the producer cells seems to be dependent on the pH of the culture broth (Hurst and Dring, 1968; Lee and Kim, 1985). Hurst and Dring (1968) found that at pH 6.80 (controlled fermentation) more than 80% of the nisin synthesized was bound to the cells, while at a pH below 6.0 more than 80% of the nisin was in the culture fluid. Yang et al . (1992) showed that the optimal pH for bacteriocin adsorption to cells ranges from pH 5.5 to 6.5, which is in accordance with a dramatic decrease of nisin level observed in controlled-pH fermentation compared to uncontrolled-pH fermentation (De

Vuyst and Vandamme, 1992).

36 2.6. Inhibitions in Nisin Production

Nisin production was autoregulated by its own product since the mature peptide nisin was the external signal. Hurst and Kruse (1972) found that when nisin was added to fresh medium and the producer organism was inoculated afterwards, growth was slightly delayed. The organisms continued to grow at the same rate as the nisin-free control after a short period. On the other hand, if the organism was first grown and nisin was added

(up to 2 h after inoculation), the growth was completely inhibited. Thus, removing nisin from the system may reduce product inhibition and eventually enhance the production rate.

To overcome the inhibition of nisin production by its high concentration and enzymatic degradation of nisin, continuous removal of nisin from fermentation medium should be used. For this attempt, Hirsch (1951) used various adsorbents to remove nisin from the system. The study showed that Deacidite E and Amberlite IRC-50 were unsuitable for on-line recovery, because essential nutrients were removed and the broth that passed through the column no longer supported growth. On the other hand, Hirsch

(1951) could not find a suitable solvent to elute nisin adsorbed on a Zeocarb 216 column.

Chinachoti et al .(1997b) separated nisin Z from fermentation broth using various kinds of adsorbents including Amberlite IR-120B, CM Sephadex C-25, Celite, and Sep-Pak tC 18 ,

C18 , C 8 and tC 3 cartridges. Sep-Pak C 8 , a moderate reversed-phase column, was chosen by Chinachoti et al. (1997b) because it showed a moderate adsorption-desorption ability for nisin Z, even though it showed a lower nisin Z adsorption than tC 18 column. When the

Sep-Pak C 8 was incorporated into the batch fermentation, higher cell growth and a 1.7 times greater nisin Z production rate were obtained. Kim (1997), using a two-phase batch

37 culture in which some of the nisin produced was removed into the solvent phase, reported a 24 % higher nisin production compared to suspended-cell fermentation.

Due to its toxicity, lactic acid accumulated during growth of L. lactis was suspected to hinder both growth and nisin production. In order to eliminate toxicity of lactic acid, several techniques have been tested and used successfully. For example, the use of Kluyveromyces marxianus to consume the produced lactic acid in a mixed culture system (Shimizu et al., 1999) resulted in a significantly higher nisin titer (3,920 IU/ml) than that of a pure culture system with alkali pH-adjustment (2,320 IU/ml). Li et al.

(2000) also reported an increase in nisin production by 33% (803 IU/ml) when the two- phase system was used. Recently, a study using the reactor equipped with the anionic- exchange resin Amberlite IRA-67 also confirmed the benefit of lactate removal as it effectively enhanced nisin production by at least two-fold (Yu et al., 2002).

2.7. Recovery and purification of nisin

Purification steps of large-scale nisin production are commercially sensitive, but are suspected to include foam precipitation (frothing), sodium chloride precipitation, centrifugation or ultrafiltration, and spray or drum drying (Hirsch, 1950; Kelly et al.,

2000). The resulting nisin is centrifuged and dried before being standardized with finely ground salt (NaCl) in order to achieve an activity of 1x10 6 IU/g (in accordant with 2.5% nisin). On the other hand, laboratory-scale purification of nisin includes an ammonium sulphate precipication step, followed by various combinations of ion-exchange and hydrophobic interaction chromatography, with a final Reverse Phase-High Pressure

Liquid Chromatography (RP-HPLC) purification step (Parente and Ricciardi, 1999).

38 Several immunological-based techniques have been developed and tested. Suarez et al.

(1997) reported a one-step purification of nisin A using immunoaffinity purification with the specific monoclonal antibody against nisin A. The procedure was highly specific and gave a very high yield of 73%. A technique using a specific monoclonal antibiody-coated magnetic bead was developed for recovery of nisin Z (Prioult et al., 2000). A high yield of 62% was obtained, however when the beads were re-used the recovery decreased to

27% only after three cycles. Even though these procedures provide excellent results in term of yield and purification (Cintas et al ., 1998), they are not suitable for large-scale recovery and purification due to requirement for specific monoclonal antibodies.

For large-scale recovery and purification, several methods based on adsorption/desorption or on phase partitioning have been developed. Yang et al . (1992) developed a method to recover nisin and other bacteriocins by adsorption on producer cells at pH 6.0-6.5, followed by cell separation and desorption at pH 2.0 and 0.1 M NaCl.

For large-scale purification, Van’t Hul and Gibbon (1996) suggested that a vortex flow filtration system might replace centrifugation for large-scale recovery. However, the ability of cells to absorb nisin can be exceeded when nisin concentration is very high

(Parente and Ricciardi, 1999). A large-scale solvent extraction using toluene was developed, resulting in 40-61% purity with 40-72% recovery (Kelly et al., 2000).

Boyaval et al . (1998) developed a two-step purification system based on detergent Triton

X-114, phase partitioning and adsorption/ desorption on a cation exchange resin. They claimed the process was effective in the recovery of nisin and mesenterocin Y105.

However, Triton X-114 is not suitable for a product that requires food-grade standard like nisin.

39 Isoelectric point of nisin falls in the alkaline range of pH 8.8 (estimated using a program ‘Compute pI/MW’ obtained from http://aegis.ateneo.net/nrojas/abi/tools.htm), thus it posses a positive charge at a pH value near or below neutrality. Recovery of nisin is usually conducted using a cationic adsorbant, which facilitates an electrostatic attraction between the adsorbent and nisin molecule. C3 silica gel with granular size of

0.15-0.25 mm was selected as the best adsorbent for isolating of nisin from the cell-free fermentation broth (Baranova et al., 1989). The pH values from 4.2 to 6.2 were found to be most effective (44% nisin recovery) while the lower pH of 2.3 greatly reduce the sorptive capacity (26% nisin recovery). Various Amberlite XAD polymeric resins have been used for nisin adsorption. While Amberlite XAD-1180 (with a less polar characteristic than XAD-4 and XAD-16) was selected for adsorption of epidermin, a 21- amino acid lantibiotic with strongly cationic nature like nisin (Horner et al., 1989), both

Micro-Cel E (a synthetic calcium silicate from diatomaceous earth) and Hi-Sil HOA (a synthetic ultrafine silicon dioxide) were successfully used for adsorption of nisin (Wan et al., 1996). The results showed that 128-fold purification was achieved for nisin.

However, the nisin adsorped on Micro-Cel E can be eluted out only by 1% sodium dodecyl sulfate (SDS) and thus effective SDS removal technique was required (Coventry et al., 1996).

Silicic acid was successfully used to recover 84% of nisin from fermentation broth, however only 3-fold purification was achieved (Janes et al., 1998). Recently, a one-step purification of nisin Z using expanded bed ion exchange chromatography was developed for large-scale purification (Cheigh et al., 2004). A very high recovery of 90%

40 yield was achieved with the optimal pH for the adsorption at 3-4, possibly as a result of higher positive net charge of nisin Z at low pH.

2.8. Models of nisin production

Only a few mathematical models have been developed to gain information about the kinetic-metabolic nature of nisin. Cabo et al . (2001) proposed a model describing an effect of pH drop gradient (VpH) on growth and nisin production in L. lactis IIM Lb.

1.13. The growth was formulated using the logistic equation with an assumption that

VpH had an effect on the growth rate (r x) as follow:

Rx = r x (1 + b (VpH)) (Eqn. 2.1)

When Rx: growth rate induced by pH variation

rx: growth rate of species x at constant pH = dx/dt

b: constant ratio obtain from the experiment

VpH: decrease in pH per unit of time = ∆pH/ ∆t

Finally, the biomass at variable pH (X R) is obtained by numerical integration of

Rx . On the other hand, the accumulation of nisin was formulated from the Luedeking -

Piret model with the obtained values of Rx and XR as follow:

BT = ( Σ α)(R x) + ( β)(XR) (Eqn. 2.2) where BT: bacteriocin

α and β: growth-associated and non-growth-associated constants

41 The results indicated that nisin production, which was a primary metabolite, only depends on the growth rate ( β = 0); when pH drops freely, it tended to be a mixed metabolite, and the production rate depended on the growth rate and the amount of biomass present in the medium ( α ≠ 0, β ≠ 0), when a stepwise-pH profile was used. This trend was more noticeable with higher pH drop gradients.

In 2003, Guerra and Pastrana developed a kinetic model for describing the specific effect of pH drop on nisin production of L. lactis strain CECT 539 in whey.

However, their results contradicted with the study of Cabo et al . (2001) since a null value of β was obtained in all treatments with different pH drop. This model characterized nisin as a primary metabolite. It is only recently that the models of nisin production in batch and fed-batch fermentations were developed using optimization program for non-linear regression (Lv et al., 2005). Instead of Luedeking-Piret model, nisin production was described using the following equations.

dN dx dx = K when ≥ 0 (Eqn. 2.3) dt N dt dt

dN dx = −K x when < 0 (Eqn. 2.4) dt D dt where N: nisin titer (mg/l)

x: biomass (g/l)

KN: specific nisin production rate (mg/g-cell dry weight)

KD: rate of nisin inactivation (mg/g-cell dry weight/h)

42 2.9. Summary of literature review

Nisin is an antimicrobial peptide that is accepted worldwide as a natural and safe food preservative. It also shows many other potential industrial applications varying from therapeutic uses to packaging and cosmetic products. Both low production rate during fermentation and tedious separation processes eventually result in high costs of partially- and highly-purified nisin (e.g. $250-350 per kg of product containing 2.5% nisin), limiting the development of potential applications. Therefore, it is essential to develop more effective production and separation methods for nisin production.

Many aspects of nisin production have been well studied and reported, ranging from producer strains (De Vuyst, 1994; Kim et al., 1997), media formula (Cabo et al.,

2001a; Li et al., 2002), and cultural conditions such as pH, aeration, and temperature

(Shimizu et al., 1999; Desjardins et al., 2001; Cabo et al., 2001b). As nisin production was shown to be cell-density-dependent, several systems using mostly active immobilized-cell reactor was developed and evaluated (Wan et al., 1995; Chinachoti et al., 1997b; Bertrand et al., 2001; Desjardins et al., 2002). Thus far, only one study on nisin production using a passive immobilized-cell reactor such as biofilm reactor has been reported (Bober and Demirci, 2004). Although no improvement in nisin production was found, there were several aspects of the system that can be improved. Therefore, this study has been undertaken to evaluate many of these factors.

Production of nisin is affected by several cultural factors, such as producer strain, nutrients composition of media, pH, temperature, agitation and aeration, as well as the unique characteristic of nisin production like substrate and product inhibition, adsorption of nisin onto the producer cells, and enzymatic degradation. Many studies showed that

43 removal of nisin from culture broth during fermentation significantly enhanced nisin production (Kim, 1997; Chinachoti et al ., 1997b). Removal of nisin not only eliminates effect of product inhibition, but may also reduce amount of nisin degraded by proteases during fermentation. Thus, incorporation of on-line adsorption of nisin from fermentation broth by silica adsorbent, which was shown to provide good recovery for nisin (Baranova et al ., 1989; Coventry et al ., 1996; and Wan et al ., 1995) in PCS biofilm fermentation, may enhance nisin production.

The goal of this study was to improve the nisin production in a biofilm reactor. In order to achieve the goal, several specific objectives have been fulfilled. First, the appropriate bioassay procedure for monitoring of nisin concentration during fermentation process has been established for further use throughout the study (Chapter 3). The ingredents of PCS and culture medium for production of nisin in PCS biofilm reactor has been optimized, and the nisin production from the biofilm reactor was compared with a result from suspended-cell reactor (Chapter 4). Effects of various pH control profiles on nisin production in both biofilm and suspended-cell reactors were studied in both repeated-batch (Chapter 5) and fed-batch (Chapter 6) fermentations. Then, silicic acid has been evaluated for recovery of nisin Z from the fermentation broth obtained from the reactor (Chapter 7). Effects of pH and eluents on adsorption and desorption of nisin have been studied. The obtained result can be used in a development of an on-line recovery unit, which can be integrated into the existing biofilm reactor. Finally, in order to gain insight information on kinetic-metabolic nature of nisin production, the results obtained from suspended-cell system were used to construct model of nisin production (Chapter

8).

44 CHAPTER 3

EVALUATION OF AGAR DIFFUSION BIOASSAY

FOR NISIN QUANTIFICATION *

ABSTRACT

Agar diffusion bioassay is the most widely used method for quantification of nisin due to its high sensitivity, simplicity, and cost effectiveness. This method is based on the measurement of inhibition zone produced in nisin-sensitive microorganisms. The size of the zones is affected by many factors such as nisin-sensitive strain, amount of added agar and surfactant, and pre-diffusion step. This research aims to evaluate effects of nisin- sensitive strains and pre-diffusion on accuracy and precision of nisin quantification.

Three strains of nisin-sensitive microorganisms ( Micrococcus luteus, Lactobacillus sakei, and Brochothrix thermosphacta ) were tested along with three different incubation processes. The best combination was the method using L. sakei as an indicator strain with pre-diffusion at 4 °C for 24 h. Compared to M. luteus and B. thermosphacta , L. sakei gave more accurate and reproducible results. Moreover, pre-diffusion step resulted in larger inhibition zones and more precise results. Finally, the best combination was validated and compared with the method that is usually used and the result showed that the method using L. sakei with pre-diffusion gave more accurate and precise results.

* This chapter has been published in Journal of Applied Microbiology and Biotechnology, Vol. 65, pp. 268- 272, 2004.

45 INTRODUCTION

Increasing concerns of government agencies, food manufacturers, and consumers on food safety and a recently increasing demand for more “natural”, “safe”, and

“minimally processed” food has brought great attention to applications of natural antimicrobial polypeptides produced by bacteria, called bacteriocin. At present, nisin is the only bacteriocin approved by U.S. Food and Drug Administration for using as a food preservative (FDA, 1988) and is allowed in about fifty countries around the world

(Delves-Broughton, 1990).

Nisin is a small polypeptide (34 amino acids) produced by Gram-positive bacteria

Lactococcus lactis subsp. lactis. It shows antimicrobial activity against a wide range of gram-positive bacteria such as the foodborne pathogens Listeria monocytogenes,

Staphylococcus aureus, Clostridium botulinum and its spore (McMullen and Stiles,

1996). It is now proposed that the antimicrobial effect of nisin is due to the incorporation of nisin into the membrane leading to membrane disruption and collapse of the proton motive force. In case of bacterial spore, nisin affects the post-germination stages of spore development by inhibiting pre-emergent swelling, the out-growth and formation of vegetative cells (Davies and Delves-Broughton, 2000). Nisin has many advantages over other food preservatives such as its non-toxicity, immediate digestibility by enzyme α- chymotrypsin, heat-stability at low pH, and absence of color and flavor.

Nisin production is closely related to growth since its biosynthesis occurs during the exponential growth phase and completely stops when the cells enter the stationary growth phase (Hirsch, 1951; De Vuyst and Vandamme, 1992; Kim et al ., 1997; Cheigh et

46 al., 2002). In attempt to monitor rate of nisin production during fermentation, precise and accurate quantification method of nisin is necessary.

There were many techniques on quantification of nisin such as tube dilution techniques (Hirsch, 1950), turbidity assays (Berridge and Berrett, 1952), agar diffusion bioassay (Rogers and Montville, 1991; Tramer and Fowler, 1964; Wolf and Gibbons,

1996), photometric (Parente et al., 1993), chemiluminescence (Bouksaim et al., 1998), capillary zonal electrophoresis (Rossana et al., 1998), flow cytometry (Budde and Rasch,

2001), bioluminescence (Wahlstrom and Saris, 1999), and ELISA (Falahee et al., 1990;

Suarez et al., 1996; Bouksaim et al., 1999; Leung et al., 2002). Although the agar diffusion bioassay has many limitations such as lower sensitivity and longer time requirement, it is the most widely used method for everyday trials due to its high sensitivity, simplicity, and cost effectiveness.

The activity of nisin is expressed in terms of International Units (IU) when 1 g of pure nisin is usually equivalent to 40x10 6 IU and 1 g of Nisaplin®, commercial nisin reference by Aplin & Barrett Ltd (UK), is equivalent to 1x10 6 IU (Davies and Delves-

Broughton, 2000). Agar diffusion bioassay is based on measurement of the zones of inhibition in indicator-seeded agar plates. As the antimicrobial agent diffuses through the agar to inhibit growth of the indicator organism, it was found that there was a linear relationship between the zone size and the log 10 of concentrations and the size of inhibition zone was a result of diffusion of the antimicrobial agent and the growth rate of the indicator organism (Linton, 1983). The greater the concentration is, the greater the zone diameter. For nisin quantification of unknown sample, the diameters of the zones are measured and correlated with a standard curve plotted between diameter of the zone

47 and log 10 of nisin concentration in IU/ml. Due to its relatively large molecular size, nisin is more difficult to diffuse in agar medium compared to molecule of antibiotics. Thus,

Mocquot and Lefebvre (1956) used surfactant, Tween 80, to enhance nisin diffusion through the agar. Later, Tramer and Fowler (1964) introduced an assay medium using

Tween 20 with Micrococcus flavus , presently called Micrococcus luteus , as an indicator microorganism. They concluded that Tween 20 gave sufficient advantage over Tween 80 so that overnight refrigeration (pre-diffusion) was unnecessary. Their proposed procedure was accepted and used as standard procedures. However, Rogers and Montville (1991) showed that pre-diffusion of plates at 3 °C increased both assay sensitivity by increasing size of inhibition zone, and assay reproducibility by giving less variability between readings on separate days. Also, the effect of storing the seeded plate at low temperature, which was reported to cause slower growth of test organism and consequently gave larger inhibition zone in antibiotic quantification (Piddock, 1990), has not been evaluated with test organisms using in nisin quantification.

As for the indicator microorganism, Wolf and Gibbons (1995) compared M. luteus , Lactobacillus brevis , and Staphylococcus aureus and chose M. luteus over other two strains due to its higher sensitivity, accuracy, and faster growth. However, the study of Rogers and Montville (1991), comparing M. luteus and Lactobacillus sakei , indicated that L. sakei exhibited a higher sensitivity to nisin. In attempt to find the most resistant organism to nisin for an evaluation of nisin spray treatment on beef carcass, Cutter and

Siragusa (1994) reported compared qualitatively Brochothrix thermosphacta ,

Carnobacterium divergens , and Listeria innocua as nisin sensitive microorganisms and reported that B. thermosphacta as the most sensitive strain among tested microorganisms.

48 Therefore, in this research all three nisin-sensitive microorganisms which have been used as indicator organisms ( M. luteus , L. sakei , and B. thermosphacta ) were tested with three different assay procedures to evaluate the effect of pre-diffusion and storing of seeded plate at low temperature on the assay in order to find the best combination of indicator strain and the assay procedure which gives more accurate and precise nisin quantification.

MATERIALS AND METHODS

Microorganisms and media

The nisin sensitive microorganisms evaluated are Micrococcus luteus (ATCC

10420), Lactobacillus sakei (ATCC 15521), and Brochothrix thermosphacta (ATCC

11509), which were obtained from the American Type Culture Collection, Rockville,

MD. All stock cultures were maintained at -80 °C in 20% v/v glycerol. M. luteus , L. sakei , and B. thermosphacta were grown in nutrient broth (NB, Difco, Detroit, MI), de Man,

Rogosa and Sharpe (MRS, Difco), and triptic soy broth (TSB, Difco), respectively. The working cultures were maintained on agar slants of the same media at 4 °C.

Nisin standards

A stock nisin solution (1,000 IU/ml) was prepared by adding 0.025 g of commercial nisin10 6 IU/g (Sigma Chemical Co., St. Louis, MO) into 25 ml of sterile diluent solution of 9:1 volumetric ratio of 0.02 N HCl acid: Lactococcus lactis fermentation medium, which consists of 80 g/l of glucose (Staley, Decatur, IL), 10 g/l of

49 peptone (Amber Ferm 4015G, Universal Flavors, Milwaukee, WI), 10 g/l of yeast extract

(Ardamine Z, Sensient Bionutrient, Indianapolis, IN), 10 g/l of KH 2PO 4, 2 g/l of NaCl, and 0.2 g/l of MgSO 4•7H 2O. Standard nisin solutions of 500, 400, 300, 200, 100, 50, 25,

10, 5, and 0 IU/ml were prepared by using the 1,000 IU/ml nisin stock solution and diluent solution, and utilized to construct the standard curve.

Nisin bioassay

The bioassay agar plates were prepared for M. luteus, L. sakei, and B. thermosphacta using the optimal media for each organism, NB, MRS, and TSB, respectively. In each media, 0.75% w/v Bacto agar (Difco) and 1% v/v of Tween 20 (J.T.

Baker, Phillipsburg, NJ) were added before bringing to boiling and sterilizing. After autoclaving, the agar medium was cooled to 40 °C, and inoculated with 1% v/v of 24-h corresponding culture of the nisin sensitive microorganism. To ensure that an equivalent number of each microorganism was inoculated into the agar medium each time, the inoculum size was adjusted to reflect a 1.7 optical cell density reading at 600 nm using a proportion. Corresponding diluted media of each microorganism was used as blank for optical density measurement. The final population of the microorganism was approximately 10 8 CFU/ml of agar medium. The bioassay agar (25 ml) was aseptically poured into sterile Petri dishes (100 x 15 mm) and allowed to solidify for 3 h. On each plate, four or five holes were bored using a 7-mm outer diameter stainless steel borer with slight suction applied. Three different assay procedures were evaluated for introduction of nisin solution. i) A 100 µl of each standard nisin solution was placed into a well and the bioassay agar plates were incubated right away at 30 °C for 24 h, which is a

50 commonly used conventional method (method 1); ii) after placing 100 µl of each nisin solution into a well, the agar plates were stored at 4 °C for 24 h to allow pre-diffusion of nisin then incubated at 30 °C for another 24 h (method 2); iii) bioassay agar plates were stored first at 4 °C for 24 h, then 100 µl of nisin standard solution was placed into a well and incubated at 30 °C for 24 h (method 3). Three wells were used for each nisin standard concentration. Diameter of each inhibition zone around the well was measured horizontally and vertically using a digital caliper (Digimatic caliper, Mitutoyo,

Kanagawa, Japan) to the nearest 0.01 mm and averaged. Each nisin sensitive microorganism was evaluated using all three methods and all treatments were replicated three times.

Diameters of inhibition zones versus log 10 nisin concentrations were plotted in order to obtain standard curve. Line regression equation was determined for each standard curve. The slopes, correlation coefficients of the regression lines, and diameters of inhibition zone at 5 IU/ml were compared for accuracy, precision, and sensitivity of the bioassay. Accuracy was inferred from slope of the regression line since greater slope represented larger change in the diameter of inhibition zone for a given nisin concentration, which resulted in greater accuracy of the bioassay. Precision was related with the correlation coefficient (R-square) of each regression equation. Sensitivity of the bioassay was correlated with the diameter of inhibition zone obtained from the nisin standard solution at 5 IU/ml. A set of data (one value of slope, R-sq, and diameter at 5

IU/ml) was obtained from each replicate and finally three sets of data were obtained from each combination of indicator organisms and methods.

51 Verification

The selected bioassay procedure and nisin-sensitive microorganism was validated and compared with the conventional method proposed by Tramer and Fowler (1964)

(method 1 with M. luteus ) with an experiment in which two separate sets of data

(diameters of inhibition zones) were obtained from the same set of nisin standard solutions. For both conventional and selected methods, one set of data was used to create a standard curve while another set of data was used to obtain a predicted concentration of nisin. A linear correlation was performed between the actual and predicted concentration of nisin.

Statistical analysis

The results (slope, R-sq, and diameter at 5 IU/ml) were analyzed using MINITAB

(Minitab Inc., State College, PA) for analysis of variance (ANOVA) and determined significant and non-significant difference with 95% confidence level ( α = 0.05). Then, multiple comparison were performed using Tukey’s test and the best combination of nisin sensitive microorganism and incubation method was chosen based on accuracy, precision, and sensitivity.

RESULTS

In the regression equations of log 10 nisin concentration versus diameter of inhibition zone, a higher slope represents a larger change in the size of inhibition zone over the same increase in nisin concentration, therefore conferring higher accuracy to the

52 bioassay. However, there was no significant difference in the slopes between each strain when using the same method, thus types of nisin sensitive microorganisms evaluated in this research did not affect the accuracy of the bioassay (Table 3.1 and Fig. 3.1). On the other hand the pre-diffusion of bioassay agar plates for every microorganism demonstrated significant increase in the slopes, allowing a more accurate bioassay.

Table 3.1. Effects of nisin sensitive microorganisms and incubation methods on slopes of regression equations.

Slope M. luteus L. sakei B. thermosphacta (mm/IU/ml) a a a Method 1 3.68 ± 0.92 4.46 ± 0.43 4.04 ± 1.32

bc bd bef Method 2 8.19 ± 1.05 8.29 ± 0.31 7.73 ± 1.41

a ae acdf Method 3 4.49 ± 0.46 5.13 ± 0.22 5.85 ± 1.30 Values not followed by the same letter are significantly different (P ≤ 0.05).

40

30

20

10 Diameter of Inhibition zone (mm) 0 0.5 1.0 1.5 2.0 2.5 3.0

Log 10 of nisin concentration (IU/ml)

M.luteus L.sakei B.thermospacta

Figure 3.1. Diameter of inhibition zone obtained from assays using different nisin sensitive microorganisms with method 2.

53 For every nisin-sensitive microorganism evaluated in this research, pre-diffusion demonstrated significant increase in the slopes. It was agreed with the previous investigation by Rogers and Montville (1991) that the pre-diffusion step increased size of inhibition zone and slope of the standard curve, enabling a more sensitive and accurate assay. However, an advantage of pre-diffusion on reproducibility of the assay as previously reported was not found in this research as indicated by indifference between the standard deviation values of slopes obtained from method 1 and method 2 (Table 3.1).

Although the slopes of each strain when using method 2 were not significantly different, lower standard variation of the results obtained from L. sakei was denoted. In general, the difference between standard deviation values of the slopes obtained from different indicator strains using the same method, for example the value of 0.31 from L. sakei using method 2 and 1.05 from M. luteus using method 2, clearly indicated that high reproducibility of the bioassay could be obtained by using L. sakei as a nisin sensitive strain. This advantage has not been discovered before for L. sakei .

It was proved in this research that storage of agar plates at 4°C for 24 h (Method

3) did not affect both the size of inhibition zone and the slope of the standard curve for the all nisin-sensitive microorganisms evaluated. Therefore, the seeded plates can be stored at 4 °C for 24 h if needed. This will provide more convenience to the bioassay.

Moreover, this also confirmed that the larger size of inhibition zone obtained from method 2 was a consequence of pre-diffusion, not the delayed growth of the microorganism.

B. thermosphacta demonstrated the lowest correlation coefficients with the highest standard deviation, which means this strain does not always demonstrate a good

54 linear response to nisin concentration (Table 3.2). Therefore B. thermosphacta is not appropriate for using as an indicator strain since an acceptable standard curve should have a correlation coefficient (R-sq) at least 0.98. On the other hand, standard curves obtained from L. sakei showed consistency in correlation coefficient, enabling a more precise assay. Although the slopes and the correlation coefficients from L. sakei and M. luteus when using method 2 were equally high, L. sakei resulted in significantly higher sensitivity as previously described.

Table 3.2. Effect of nisin sensitive microorganisms and incubation methods on correlation coefficients of regression equations.

R-sq M. luteus L. sakei B. thermosphacta

Method 1 0.954 ± 0.030 0.988 ± 0.006 0.905 ± 0.085

Method 2 0.983 ± 0.018 0.987 ± 0.003 0.967 ± 0.003

Method 3 0.975 ± 0.018 0.984 ± 0.011 0.935 ± 0.051

As a larger size of inhibition zone indicates that the test is more sensitive to nisin, the diameters of inhibition zone obtained from the lowest nisin standard concentration (5

IU/ml) were compared to show the sensitivity of each combination. The results demonstrated that B. thermosphacta was significantly more sensitive to nisin than M. luteus and L. sakei when using the same methods (Table 3.3). Storage of bioassay agar plates at 4°C for 24 h (method 3) did not affect sensitivity of the bioassay. However, pre- diffusion step (method 2) enhanced sensitivity in the bioassays using L. sakei and B. thermosphacta. Therefore, the combination of L. sakei and method 2 was selected as the best combination since it gave large size of inhibition zone, steep standard curve, and

55 consistent high correlation coefficient of regression equation, conferring a sensitive, accurate, and precise bioassay.

Table 3.3. Effects of nisin sensitive microorganisms and incubation methods on diameters of inhibition zones obtained from nisin standard solution at 5 IU/ml.

Diameter (mm) M. luteus L. sakei B. thermosphacta

ac ad be Method 1 9.25 ± 0.51 9.95 ± 0.50 14.65 ± 1.76

a cde f Method 2 9.09 ± 0.57 12.29 ± 0.98 19.16 ± 2.29

a a be Method 3 9.23 ± 0.28 9.26 ± 0.24 14.03 ± 0.18 Values not followed by the same letter are significantly different (P ≤ 0.05).

Finally, the selected combination ( L. sakei with method 2) was validated and compared with the conventional procedure ( M. luteus with method 1). According to the slopes (1.01 versus 0.68) and correlation coefficients (0.92 versus 0.94) obtained from linear regressions of actual and predicted nisin concentrations (Fig. 3.2), although the correlation coefficients of both methods were not different, it was shown that the bioassay using L. sakei with method 2 was more accurate and precise as it gave the slope value which was closer to unity. The plot also showed that the appropriate range of nisin concentration used for creating standard curve should not exceed 300 IU/ml since above that concentration the predicted nisin concentrations did not exhibit a good linear relationship with the actual nisin concentrations. When a linear regression was performed using nisin standard solution from 5 to 300 IU/ml, the regression equation of actual and predicted nisin concentrations became y = 1.02x + 2.93 with a correlation coefficient of

0.99.

56

600

500 y = 1.01x + 7.67 R-sq = 0.92 400

300

200

y = 0.68x + 19.61

Predicted concentration (IU/ml) concentration Predicted 100 R-sq = 0.93

0 0 100 200 300 400 500 600 Actual concentration (IU/ml)

L. sakei with method 2 M. luteus with method 1

Figure 3.2. A linear regression of actual and predicted nisin concentration obtained from an assay using L. sakei with method 2 and an assay using M. luteus with method 1 (conventional procedure).

DISCUSSIONS

Immense industrial uses and research studies of nisin in various applications (e.g. food preservative, human ulcer therapy, and mastitis control) require accurate and precise method for quantification of nisin in food samples as well as in fermentation broth and other samples. The agar diffusion bioassay is the most widely used method due to its high sensitivity, measuring as low as 0.0125 µg/ml, simplicity and cost effectiveness. Our study indicated that the traditional agar diffusion bioassay could be improved further

57 using more sensitive indicator microorganism, pre-diffusion, and appropriate range of nisin concentration in sample.

In this study, L. sakei not only exhibited higher consistency in slope and correlation coefficient of the standard curve, but when combining with pre-diffusion of agar plates at 4 °C for 24 h, it resulted in a more sensitive, accurate, and precise bioassay compared with other combinations including the conventional procedure suggested by previous studies. Further study on effect of culture time and type of media for bioassay agar on the assay should be investigated. Moreover, for application of nisin quantification in food materials the accuracy and precision of the assay when using nisin extracted from food should be investigated to confirm the superiority of the assay using L. sakei with pre-diffusion over the traditional procedure. On the other hand, our finding suggested that there was no significant effect of storage of seeded-agar plate at 4 °C for 24 h on the sensitivity and accuracy of the bioassay not only confirmed the positive effect of pre- diffusion, but also resulted in more flexible agar diffusion bioassay procedure.

58 REFERENCES

Berridge, N. J., and J. Barrett. 1952. A rapid method for the turbidimetric assay of antibiotics. J. Gen. Microbiol. 6: 14-20. Bouksaim, M., I. Fliss, J. Meghrous, R. E. Simard, and C. Lacroix. 1998. Immunodot detection of nisin Z in milk and whey using enhanced chemiluminescence. J. Appl. Microbiol. 84: 176-184. Bouksaim, M., C. Lacroix, R. Bazin, and R. E. Simard. 1999. Production and utilization of polyclonal antibodies against nisin in an ELISA and for immuno-location of nisin in producing and sensitive bacterial strains. J. Appl. Microbiol. 87: 500-510. Budde, B. B., and M. Rasch. 2001. A comparative study on the use of flow cytometry and colony forming units for assessment of the antibacterial effect of bacteriocins. Int. J. Food Microbiol. 63: 65-72. Cheigh, C. I., H. J. Choi, H. Park, S. B. Kim, M. C. Kook, T. S. Kim, J. K. Hwang, and Y. R. Pyun. 2002. Influence of growth conditions on the production of a nisin-like bacteriocin by Lactococcus lactis subsp. lactis A164 isolated from kimchi. J. Biotechnol. 95: 225-235. Cutter, C. N., and G. R. Siragusa. 1994. Decontamination of beef carcass tissue with nisin using a pilot scale model carcass washer. Food Microbiol. 11: 481-489. Davies, E. A., and J. Delves-Broughton. 2000. Nisin. In Encyclopedia of Food Microbiology, eds. C. A. Batt, and P. D. Patel. 191-198. Academic Press. Delves-Broughton, J. 1990. Nisin and its uses as a food preservative. Food Technol. 11: 110-117. De Vuyst, L., and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. Falahee, M. B., M. R. Adams, J. W. Dale, and B. A. Morris. 1990. An enzyme immunoassay for nisin. Int. J. Food Sci. and Tech. 25: 590-595. Hirsch, A. 1950. The assay of the antibiotic nisin. J. Gen. Microbiol. 4: 70-74.

59 Hirsch, A. 1951. Growth and nisin production of a strain of Streptococcus lactis . J. Gen. Microbiol. 5: 208-221. Kim, W. S., R. J. Hall, and N. W. Dunn. 1997. The effect of nisin concentration and nutrient depletion on nisin production of Lactococcus lactis. Appl. Microbiol. Biotechnol. 48: 449-453. Leung, P. P., M. Khadre, T. H. Shellhammer, and A. E. Yousef. 2002. Immunoassay method for quantitative determination of nisin in solution and on polymeric films. Lett. Appl. Microbiol. 34: 199-204. McMullen, L. M., and M. E. Stiles. 1996. Potential for use of bacteriocin-producing lactic acid bacteria in the preservation of meats. J. Food Prot. 64-71. Mocquot, G., and E. Lefebvre. 1956. A simple procedure to detect nisin in cheese. J. Appl. Bact. 19: 322. Parente, E., A. Ricciardi, and V. Villani. 1993. Evaluation of two methods for the measurement of bacteriocin activity. In Biotechnology and Molecular Biology of Lactic Acid Bacteria for the Improvement of Foods and Feeds Quality , eds. A. Zamorani, P. L. Manachini, V. Bottazzi, and S. Coppola. 318-327. Instituto Poligrafico e Zecca dello Stato, Rome, Italy. Piddock, L. J. V. 1990. Techniques used for the determination of antimicrobial resistance and sensitivity in bacteria. J. Appl. Bacteriol. 68: 307-318. Rogers, A. M., and T. J. Montville. 1991. Improved agar diffusion assay for nisin quantification. Food Biotechnol. 5(2): 161-168. Rossana, R., A. Del Fiore, A. D’Elia, G. Pesole, E. Parente, and P. Riccio. 1998. New procedure for the determination of nisin in milk. Biotechnol. Techniques 12(10): 783-786. Saurez, A. M., J. M. P. Rodriguez, P. Morales, P. E. Hernandez, and J. I. Azcona-Olivera. 1996. Development of monoclonal antibodies to the lantibiotic nisin A. J. Agric.Food Chem. 44: 2936-2940. Tramer, J., and G. G. Fowler. 1964. Estimation of nisin in foods. J. Sci. Fd Agric. 15: 522-528. U.S. Food and Drug Administration. 1988. Nisin preparation: Affirmation of GRAS status as direct human food ingredient. Federal Register. 53, April 6.

60 Wahlstrom, G., and P. E. J. Saris. 1999. A nisin bioassay based on bioluminescence. Appl. Environ. Microbiol. 65(8): 3742-3745. Wolf, C.E., and W. R. Gibbons. 1996. Improved method for quantification of the bacteriocin nisin. J. Appl. Bacteriol. 80: 453-457.

61 CHAPTER 4

EVALUATION OF CULTURE MEDIUM FOR NISIN PRODUCTION

IN REPEATED-BATCH BIOFILM REACTOR *

ABSTRACT

In this study, a biofilm reactor with Plastic Composite Support (PCS), made by high temperature extrusion of agricultural products and polypropylene, was evaluated for nisin production using L. lactis strain NIZO 22186. The high-biomass density of the biofilm reactor was found to contribute to a significantly shorter lag time of nisin production relative to a suspension cell reactor. In comparison to glucose (579 IU/ml), sucrose significantly increased the nisin production rate by 1.4-fold (1,100 IU/ml).

However, results revealed that high level of sucrose (8% w/v) had a suppressing effect on nisin production and a stimulating effect on lactic acid production. A high concentration of MgSO 4•7H 2O at 0.04% w/v was found to reduce the nisin production, while concentrations of KH 2PO 4 of up to 3% w/v did not have any significant effect on growth or nisin production. The best of the tested complex media for nisin production using the

PCS biofilm reactor consisted of 4% w/v sucrose, 0.02% w/v MgSO 4•7H 2O, 1% w/v

KH 2PO 4, 0.2% w/v NaCl, 1% peptone, and 1% yeast extract. Nisin production rate in the biofilm reactor was significantly increased by 3.8-fold (2,208 IU/ml) when using the best complex medium tested.

* This chapter has been published in Biotechnology Progress, Vol. 22, No. 1, pp. 217-224, 2006.

62 INTRODUCTION

Nisin is an antimicrobial peptide that exhibits a broad antibacterial activity against

Gram-positive spoilage and pathogenic bacteria, thus it is widely used in food industry as a natural preservative. So far, nisin is the only bacteriocin approved by FDA to use as food additive (FDA, 1988). Production of nisin by bacteria Lactococcus lactis subsp. lactis appeared to be regulated by a quorum sensing at the transcriptional level in a cell- density-dependent manner, with nisin itself acting as an inducer molecule, or a so called peptide pheromone (Kuipers et al., 1995). It was shown that detectable production of nisin was initiated at mid-logarithmic growth phase (when biomass reached 50% of maximum biomass) and increased to reach its maximum at the beginning of the stationary phase (Hirsch, 1951; Buchman et al., 1988), a characteristic of Type I fermentation. Kleerebezem (2004) suggested that the high cell-density-dependent regulation was a mechanism to ensure that the concentration of the antimicrobial peptide in the environment rapidly reached an inhibition level, in order to avoid a development of a defense mechanism in target microorganisms. High biomass density proved to be beneficial for nisin production in many cases (Hirsch, 1951; De Vuyst and Vandamme,

1993; Yang and Ray, 1994), thus it is likely that nisin production can be enhanced in a system with high biomass density such as biofilm reactor.

High density of biomass can be achieved by several alternatives such as cell- recycle reactors, hollow-fiber reactors, and cell immobilization. Hollow-fiber reactors and cell-recycle reactors have their limitations due to their high capital and operation cost as well as in the potential for membrane fouling during fermentation. Biofilm reactors, on the other hand, are excellent examples of high-biomass density systems with lower

63 required capital costs. Biofilm reactors have been shown to result in very high volumetric productivity of batch or continuous cultures compared with suspension cell (SC) cultures in many systems due to the high cell density maintained in the reactor (Demirci et al,

1997; Wang, 2000; Cotton et al., 2001; Velazquez et al., 2001). It was reported that cell immobilization can also change the cell growth rate, morphology, and even metabolic activities (Teixeira et al., 1994; Krishnan et al., 2001; Doleyres et al., 2004).

There have been many studies on application of cell immobilization and continuous cultures in nisin fermentation, including a Ca-alginate immobilized L. lactis in a continuous packed-bed reactor (Wan et al., 1995); an adsorption on porous chitosan bead/photo-crosslinked resin gel beads and an entrapment in photo-crosslinkable resin prepolymers (Sonomoto et al., 2000); a hydrophobic hollow fiber membrane (Chinachoti et al., 1997); and immobilized cells on carrageenan/locust bean gum gel beads (Bertrand et al., 2001). However, only one study reported the utilization of a biofilm reactor in nisin fermentation (Bober and Demirci, 2004). Although the results of their study showed that there was no increase in nisin production or productivity rates using the biofilm reactor in batch fermentation, inadequate mixing in the reactor was suspected as the cause of these limitations.

Plastic composite support (PCS) is an extrusion product of a mixture between polyprolylene and agricultural products (Demirci et al., 1997) and was later awarded with a U.S. Patent (U.S. Patent Number: 5,595,893) (Pometto et al., 1997). The polypropylene acts as a matrix to integrate the mixture of agricultural products, which provide essential nutrients to sustain cell growth. Thus, PCS not only provides ideal physical structure for biofilm formation, but also slowly releases nutrients during fermentation. Moreover,

64 added nutrients can be adjusted in order to meet specific requirements of the desired microorganism. Many studies showed that biofilm reactors using PCS enhanced ethanol and organic acid production in both repeated-batch and continuous fermentation (Demirci et al., 1995; 1997; Kunduru and Pometto, 1996a; 1996b; Ho et al., 1997a; 1997b; 1997c;

Cotton et al., 2001; Velazquez et al., 2001; Urbance et al., 2003). Therefore, a PCS biofilm reactor system was evaluated for nisin production using L. lactis subsp. lactis

NIZO 22186 in this study.

Lactococci are nutritionally fastidious microorganisms which grow well in milk or complex organic media. Nisin production is greatly affected by many factors such as producer strains, type and level of carbon, nitrogen and phosphate sources, cations, surfactants, pH, temperatures, accumulated lactic acid and aeration. Culture media formulation for large-scale nisin production was first developed by Mattick and Hirsch

(1947). However their media, based on glucose and yeast extract, resulted in nisin concentration of only 80 IU/ml. Increasing initial glucose concentration (from 1% to

2.5% w/v) and neutralization of pH (using either buffer solution or concentrated base) proved to be beneficial for nisin production (Hirsch, 1951). Nisin concentrations of about

1,500-2,200 IU/ml were obtained using a medium based on glucose, peptone and/or yeast extract, potassium dihydrogen phosphate, magnesium sulphate, and sodium chloride

(Hirsch, 1951; Egorov et al., 1971; Kozlova et al., 1972).

Nisin can be produced from many types of carbon sources. Different producer strains showed different preferences on sugar consumption (Chandrapati and O’Sullivan,

1998; Li et al., 2002a; Yu et al., 2002). Nisin-producing L. lactis subsp. lactis rapidly utilizes sucrose due to an efficient sucrose-specific phosphoenolpyruvate-dependent

65 phosphotransferase system (PTS) it posses. A study on the expression of sucrose catabolic genes in the presence of different carbon source indicated a form of glucose repression (Luesink et al., 1999). Phenotypic correlation and genetic linkage between sucrose fermentation capacity and nisin production were previously reported (Hirsch and

Grinsted, 1951; De Vuyst, 1994). Carbon source regulation has been reported to have a profound effect on both growth of L. lactis and nisin production. High initial sugar concentration was reported to cause a catabolic suppression, leading to lower growth and nisin production, in many L. lactis strains (De Vuyst and Vandamme, 1992; Matsusaki et al., 1996). A sucrose concentration higher than 4% resulted in lower biomass and maximum nisin titre (De Vuyst and Vandamme, 1992). Overall, optimal initial sugar concentrations for nisin production reported so far were between 3 to 4% (De Vuyst and

Vandamme, 1992; Matsusaki et al., 1996; Cheigh et al., 2002).

Calcium, sodium, and potassium ions were shown to play an important role in metabolism of lactococci (Kozlova et al., 1979). KH 2PO 4 (at 1-5% w/v) was previously selected as the best phosphate source for growth and nisin production of L. lactis subsp. lactis (De Vuyst and Vandamme, 1993; Kozlova et al., 1979) . Low concentration of

KH 2PO 4 caused a 30-35% reduction in nisin production even when the pH was maintained near neutrality with sodium hydroxide (Reiter and Oram, 1962). Thus,

KH 2PO 4 not only served as a buffering agent, but also as a source of potassium and phosphorus ions for growth. However, since the consumption of phosphorus from

KH 2PO 4 was extremely low in L. lactis (Kozlova et al., 1972) and the study from Reiter and Oram (1962) suggested that a large amount of potassium ions was required for

66 growth of L. lactis , the presence of KH 2PO 4 at high level in the medium might serve the producer requirement of potassium ions rather than phosphorus ions.

Increasing KH 2PO 4 concentration was found to prolong the time of maximum nisin titre, although a concentration higher than 6% caused cell lysis and a drastic decrease of nisin production in L. lactis NIZO 22186 (De Vuyst and Vandamme, 1993).

However, the fact that a similar phosphate concentration did not affect nisin production in

L. lactis strain IO-1 (Matsusaki et al., 1996) suggested that the effect of KH 2PO 4 might be different from strain to strain, the same phenomenon as observed for MgSO 4•7H 2O

(Egorov et al., 1971; Matsusaki et al., 1996; Meghrous et al., 1992).

The objective of this study was to investigate the potential of biofilm reactor for nisin production using L. lactis strain NIZO 22186 comparing to the conventional suspension cell culture. Several parameters; i) PCS types, ii) types of carbon source

(glucose and sucrose), and iii) different levels of selected medium compositions (sucrose, magnesium, and phosphate); are tested in order to optimize the nisin production when using PCS biofilm reactor.

MATERIALS AND METHODS

Microorganisms and media

L. lactis subsp. lactis (NIZO 22186) was used in this study as the nisin Z- producing strain. Lactobacillus sakei (ATCC 15521) was used as the nisin sensitive test organism. L. lactis and L. sakei were grown at 30 °C for 14 h in a complex medium and

Lactobacillus MRS broth (Difco Laboratories, Detroit, MI), respectively, unless stated

67 otherwise. The medium consisted of 40 g of sucrose; 10 g of peptone (Amber Ferm

4015G, Universal Flavors, Milwaukee, WI); 10 g of yeast extract (Ardamine Z, Sensient

Bionutrient, Indianapolis, IN); 10 g of KH 2PO 4; 2 g of NaCl; and 0.2 g of MgSO 4•7H 2O per liter of deionized (DI) water. The initial pH of the medium was adjusted to 6.8 using

4N NaOH. For a long-term storage, all stock cultures were maintained at -80°C in 20% v/v glycerol in deionized water.

Plastic Composite Support (PCS)

Eleven types of PCS tubes (Table 4.1) were manufactured at Iowa State

University using a twin-screw co-rotating Brabender PL2000 extruder (model CTSE-V;

C.W. Brabender Instruments, Inc., South Hackensack, NJ) as described by Ho et al.

(1997a). Polypropylene and other ingredients of PCS were mixed together before being extruded at 13 rpm through a medium pipe die with barrel temperatures of 200, 220, and

200 °C and a die temperature of 165 °C. To randomize manufacturing effects 500 g of the mixture were extruded twice for each blend in a random order to ensure reproducibility, producing 1 kg PCS from each blend. The resulting tubes had a wall thickness of 2.5 mm and an outer diameter of 10.5 mm. For bioreactor trials, these PCS tubes were cut into

7.5-cm length with both ends cut at a 45 degree angle to allow better flow of medium inside the tubes. For testing and selection of PCS in test-tube fermentations, the PCS tubes were cut into small discs with a thickness of 3 mm using a band saw.

68 Table 4.1. List of PCS ingredients. % Dry Weight Support PP a Sb Bc Fd Ye Rf MS g S 50 50 SB+ 50 45 5 + SF+ 50 45 5 + SFB 50 40 5 5 SFR 50 40 5 5 SFY 50 40 5 5 SFYB+ 50 35 5 5 5 + SFYR+ 50 35 5 5 5 + SR+ 50 45 5 + SY+ 50 45 5 + SYB 50 40 5 5 aPP, Polypropylene; Quantum USI Division, Cincinnati, OH bS, Dried ground (20-mesh) soybean hulls; Cargill Soy Processing Plant, Iowa Falls, IA. cB, Dried bovine albumin; Proliant Corp., Ames, IA. dF, Defatted soy bean flour; Archer Daniels Midland, Decatur, IL. eY, Yeast extract Ardamine Z; Red Star Bioproducts, Juneau, WI. fR, Dried bovine red blood cell; Proliant Corp., Ames, IA. gMS, Mineral salts; per kilogram PCS blend, 2.72 g of sodium acetate, 0.004 g of MgCl 2 •6H 2O, and 0.02 g of NaCl.

Test-tube fermentations for selection of PCS

Eleven types of PCS with different compositions were evaluated for both biofilm formation and nisin production using test-tube fermentation with three replicates. For each replicate, three grams of PCS discs were sterilized dry in capped 50 ml culture tubes for 1 h at 121 °C. Ten milliliters of sterilized medium with 8% w/v glucose as a carbon source was added aseptically to the sterile PCS before being incubated at 30°C overnight in order to hydrate the PCS. The medium was then aseptically decanted and fresh sterile

69 medium was added with 1% v/v inoculum of 14-h L. lactis before incubating at 30 °C for

24 h. Five repeated-batch fermentation runs were performed by decanting broth and adding fresh sterile fermentation medium each time in order to establish good biofilm formation on PCS discs. Each replicate was tested under three consecutive repeated-batch fermentations, thus the value reported for each PCS type was an average of nine obtained values, except the biofilm formation data which could be determined only after three consecutive fermentation runs. The best PCS type was selected according to the biofilm formation on PCS (CFU/g-PCS), nisin production (IU/ml), nitrogen content (% w/w), and nitrogen leaching rate from PCS (% w/w of total nitrogen after the first repeated-batch fermentation). Results on biofilm formation and nisin production, obtained from this study, were evaluated using Analysis of Variance (ANOVA) and Tukey’s HSD multiple comparison in the MINITAB Statistical Software package (Release 13.30) (State College,

PA), where as the results on nitrogen content and nitrogen leaching rate were obtained from the study of Ho et al. (1997a).

Nisin fermentation in bioreactor

Nisin fermentations were conducted in a 1.25 l-Bioflo II fermentor (New

Brunswick Scientific, Edison, NJ) at 30 °C with a working volume of 1 liter and the agitation rate of 100 rpm. For the biofilm reactor, twelve PCS tubes were bounded to the agitator shaft in a grid-like fashion, with six rows of two parallel tubes (Fig. 4.1). The reactor vessel with PCS was sterilized with water at 121 °C. Sugar (glucose or sucrose) and nitrogenous components with mineral salts were sterilized separately, and added to the reactor aseptically after draining the water from the reactor as recommended by Ho et

70 al. (1997c). After inoculation with a 14-h culture of L. lactis (1% v/v), at least five preliminary repeated-batch fermentation runs were performed in order to establish biofilm formation on PCS supports. The pH was controlled at 6.8 by adding 4N NaOH.

Once a stable biofilm was formed, repeated-batch fermentations were carried out in the biofilm reactor. Each repeated-batch fermentation was re-initialized by pumping out the previous fermentation broth and adding fresh sterile medium without a reinoculation.

After each experimental run, the biofilm reactor was rinsed with the medium that was used in the next experimental run. Suspension cell fermentation was conducted using the same procedure except that the reactor containing medium without PCS was sterilized and inoculated with 14-h culture of L. lactis (1% v/v) at the beginning of each batch.

Each treatment was duplicated.

Agitator shaft

PCS

SIDE VIEW TOP VIEW

Figure 4.1. Diagram of PCS biofilm reactor.

71 Bioreactor experimental design

Experimental trials were divided into two groups. The first group was a factorial design to determine effects of biofilm reactor (compared to suspension cell fermentation) and type of sugar (glucose versus sucrose at 8% w/v) on nisin production. The second group was a factorial design to identify the important nutrients and interactions between them. There were six ingredients in the complex medium, but only three major ingredients (sucrose, MgSO 4•7H 2O, and KH 2PO 4) were evaluated in order to reduce the number of experiments. Sucrose was selected as a better carbon source based on the results from the first group. Each ingredient was tested using two different levels as follow: sucrose (4% and 8% w/v); MgSO 4•7H 2O (0.02% and 0.04% w/v), and KH 2PO 4

(1% and 3% w/v). Selected levels of each composition tested were chosen according to the results of previous studies, mostly from suspension cell culture systems (De Vuyst and Vandamme, 1992; 1993; Matsusaki et al., 1996; Cheigh et al., 2002; Meghrous et al.,

1992). For a full 2 3 factorial design, 8 treatments were required. Each treatment was performed in duplicate. Therefore, 16 repeated-batch fermentations were performed.

Before the data was collected in each treatment, the biofilm reactor was run using the desired medium for at least two batches in order to remove the possible remaining effects from the previous treatment.

72 Analysis

For test-tube fermentation, after 24 h of incubation, decanted broth was analyzed for biomass density in broth and nisin concentration (as described below) whereas the

PCS discs were analyzed for biofilm formation using the stripping sand method as described by Ho et al. (1997c). Nisin production was evaluated for three consecutive days and the experiment was repeated in triplicate. For nisin fermentation in bioreactors, samples were collected every two hours and analyzed for suspended biomass density, sugar consumption, nisin and lactic acid production.

Suspended biomass. Suspended-cell density was estimated by optical density at

600 nm using a spectrophotometry DU series 500 (Beckman, Fullerton, CA). Sterile fermentation medium was used as a blank. Optical density values were converted into biomass concentration by using a standard curve (y = 0.1663x – 0.0199; y: biomass (g/l) and x: optical density at 600 nm).

Nisin. Sample of fermentation broth was immediately adjusted to pH 3.0 using concentrated HCl and 0.1% v/v of Tween 20 (J.T. Baker, Phillipsburg, NJ) was added in order to avoid any non-specific adsorption of nisin on the container’s surfaces. Then, the sample was heated at 90 °C for 5 min to eliminate the protease activity in the broth. After centrifugation at 2,000 x g, 4 °C for 10 min (National Labnet Co., mini centrifuge Model

C-1200), clear supernatant was collected and kept frozen at -20 °C until analysis. Nisin quantifications were performed by an agar well diffusion bioassay, using L. sakei and pre-diffusion at 4 °C, as described in Chapter 3.

Briefly, a stock standard nisin solution (100 IU/ml) was prepared by adding 0.01 g of commercial nisin (10 6 IU/g, Sigma Chemical Co., St. Louis, MO) into 100 ml of

73 sterile deionized water which the pH was adjusted to 3.0 using concentrated HCl.

Standard nisin solution of 100, 50, 25, 15, 10, 5, and 0 IU/ml were then prepared using the 100 IU/ml nisin stock solution and sterile deionized water with pH 3.0 (Diluent). The sample from fermentation broth was diluted (x40) using the same diluent solution. The bioassay agar was prepared using MRS (Difco) with 0.75% w/v Bacto agar (Difco) and

1% v/v Tween 20. Then, the sterile agar was inoculated with 1% of the 14-h culture of L. sakei before plating. To ensure that an equivalent number of cells were inoculated into the agar medium each time, the inoculum size was adjusted to reflect an optical cell density reading of 1.7 at 600 nm (approx. 10 8 CFU/ml of agar medium). Then, the inoculated bioassay agar (25ml) was aseptically transferred into sterile Petri dishes

(100x15 mm) and allowed to solidify for 1 h. On each plate, five holes were bored using a 7-mm outer diameter stainless steel borer with a slight suction applied. After 100 µl aliquot of sample was placed into each well, the agar plate was stored at 4 °C for 24 h

(pre-diffusion step) and then incubated at 30 °C for another 24 h. Each sample was performed in triplicate.

The diameter of the inhibition zone around each well was measured horizontally and vertically using a digital caliper (Digimatic caliper, Mitutoyo, Kanagawa, Japan) to the nearest 0.01 mm and then averaged. Diameters of inhibition zones from each nisin standard solution and log 10 nisin concentrations were plotted in order to create a standard curve. The concentration of nisin in the fermentation sample was calculated using the equation of the best fitted linear line in the standard curve, which should have the correlation coefficient of at least 0.98.

74 Sucrose. The fermentation broth was centrifuged at 2,000 x g for 10 min

(National Labnet Co., mini centrifuge Model C-1200) and clear supernatant was collected for analysis. For initial acid hydrolysis, 100 µl of concentrated HCl was added to 5 ml of the sample. After allowing the hydrolysis to proceed at 90 °C for 5 min, 0.25 ml of 5N

KOH solution was added to neutralize the acid. The hydrolyzed samples were then analyzed using the DNS Method as described by Miller (1959).

Glucose and lactic acid. Glucose and lactic acid concentration were determined using a high performance liquid chromatograph (HPLC) equipped with a refractive index detector (Waters, Franklin, MA). Components were separated on a Bio-Rad Aminex

HPX-87H column (300 x 7.8 mm) (Bio-Rad, Richmond, CA) with 0.012 N sulfuric acid as a mobile phase at a flow rate of 0.8 ml/min with an injection volume of 20 µl and a column temperature of 65 °C. Before injection, the samples were centrifuged (2,000 x g,

4°C) for 10 min and filtered through 0.22 micron filters (13 mm diameter disc filters,

Millipore, Bedford, MA).

Statistical analysis

All treatments were replicated three times. The significant difference of the results was evaluated using the Generalized Linear Model (GLM) (with p < 0.05) and Tukey’s

HSD multiple comparison module of the MINITAB Statistical Software package

(Release 13.30) (State College, PA).

75 RESULTS AND DISCUSSIONS

This study was designed not only to evaluate the potential of a PCS biofilm reactor for nisin production, but also to optimize the nisin production through medium optimization. The obtained data revealed different effects of the evaluated medium components and was used for further optimization of other culture conditions (e.g. pH profile) for nisin production using PCS biofilm reactor.

PCS selection

Biofilm formation (cfu/g-PCS) and nisin production (IU/ml) from three replicates of each PCS types are presented in Figures 4.2 and 4.3. The results showed no specific relationship between viable cell attached on PCS, biomass in liquid medium, and nisin production as the correlation coefficients between these values were extremely low; R- square value of 0.10 for biomass vs nisin production and of 0.15 for biofilm formation vs nisin production. Interestingly, the same lack of pattern was also observed in PCS biofilm fermentation of succinic acid (Urbance et al., 2003). Comparable levels of cell mass were attached to every PCS, ranging from 1 to 7x10 6 cells/g-PCS. Overall, considerably low nisin concentrations (106-372 IU/ml) were obtained as a result of uncontrolled-pH condition in test-tube fermentation.

In attempt to select the most appropriate PCS, the means of nisin produced

(IU/ml) were comparatively ranked according to the results from Tukey’s multiple comparisons: SFYR+ > SB+, SFB, SFYB+, SY+, SYB > SF+, SFR, SR+ > S, SFY.

However, after considering other factors (such as nitrogen content per gram of PCS,

76 nitrogen leaching rate, and amounts of viable cells attached on the PCS), SFYB+ was selected for long-term biofilm nisin fermentation by L. lactis due to its highest nitrogen content, moderate nitrogen-leaching rate, and sufficient viable cells attached on the PCS.

Both PCS with low nitrogen content and PCS with high nitrogen leaching rate were considered as inappropriate for long-term nisin production due to depletion of nitrogen after several fermentation runs. It should be noted that SFYB+ was also recommended for production of lactic acid (Ho et al., 1997b; 1997c) and succinic acid (Urbance et al.,

2003) as well.

9 8 7 6 5 per g-PCS) per 6 4 3 2 CFU(x10 1 0 S SF+ SY+ SB+ SFY SFB SR+ SYB SFR SFYB+ SFYR+ PCS Blend

Figure 4.2. Effects of different PCS blends on numbers of viable attached cells on the PCS in terst-tube systems without pH control after 24 h (n = 3). (S, soybean hulls; F, soybean flour; Y, yeast extract; R, dried bovine RBC; B, dried bovine albumin; +, mineral salts).

77 450

400

350

300

250

200

Nisin (IU/ml) Nisin 150

100

50

0

+ + + + + + Y B S B R

F Y B F B F R R Y F

S Y S S S S Y S S S

F F

S S PCS Blend

Figure 4.3. Effects of different PCS blends on nisin production in test-tube systems without pH control after 24 h (n = 3). (S, soybean hulls; F, soybean flour; Y, yeast extract; R, dried bovine RBC; B, dried bovine albumin; +, mineral salts).

Fermentation profiles of L. lactis in PCS biofilm reactor

Compared to the suspension cell reactor, cultivation in PCS biofilm reactor resulted in 2 to 4 h shorter lag time in growth with glucose as carbon source (Fig. 4.4), and in nisin production with sucrose as carbon source (Fig. 4.5), possibly as a result of higher biomass density maintained in the biofilm reactor and the cell-density-dependent regulation mechanism of nisin production previously described. It was noted that the shorter lag time observed in PCS biofilm reactor was also observed in other types of immobilized system (Chinachoti et al., 1997; Sonomoto et al., 2000). Although the amount of biomass in between the liquid broth and the biofilm have not been compared in this study, biofilm reactor was generally reported to retain 5 to 10 times more biomass per unit volume of the reactor (Melo and Oliveira, 2001).

78 1.2

1.0

0.8

0.6

Biomass (g/l) 0.4

0.2

0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Glucose-SC Glucose-BF Sucrose-SC Sucrose-BF

Figure 4.4. Biomass (g/l) in liquid medium when using glucose and sucrose as a carbon source in suspended-cell reactor (SC) and biofilm reactor (BF).

1600

1400 1200

1000

800 600 Nisin (IU/ml) 400 200

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Glucose-SC Glucose-BF Sucrose-SC Sucrose-BF

Figure 4.5. Nisin production (IU/ml) when using glucose and sucrose as a carbon source in suspension cell reactor (SC) and biofilm reactor (BF).

79 Apart from that, both suspension cell and biofilm fermentation followed a typical exponential growth phase with simultaneous production of nisin (Fig. 4.4 and 4.5). In a suspension cell reactor the higher maximum biomass obtained when using sucrose indicated the superiority of sucrose as a carbon source for growth of L. lactis strain NIZO

22186 (Fig. 4.4), while the constantly lower suspended biomass in biofilm reactor, on the other hand, suggested a likelihood that more producer cells were embedded in the biofilm rather than suspended in the liquid broth.

The fermentation pattern observed in this study agreed with previous reports

(Yang and Ray, 1994; De Vuyst and Vandamme, 1992) as nisin production is triggered by high biomass concentration, and follows primary metabolite kinetics. The nisin titer reached its maximum at the end of the exponential growth phase, generally between 4 to

6 h, then dramatically decreased and was almost depleted after 24 h. The striking reduction of nisin level observed in this study was in accordance with the result obtained from a study of De Vuyst and Vandamme (1993) using the same producer strain.

However, this phenomenon was not always observed during nisin production using other

L. lactis strains (Van’t Hul and Gibbons, 1996; Chinachoti et al., 1997; Amiali et al.,

1998; Cabo et al., 2001). Decreasing of nisin level was suggested to be a result of either active protease (sometimes referred as nisinase) of the producer strain or an adsorption of nisin to the producer cells at a pH higher than 6.0 (De Vuyst and Vandamme, 1992; Hurst and Dring, 1968). Using a nisin-producing L. lactis strain with less pronounced protease activity will not only improve the nisin titre of fermentation product, but also facilitate the downstream recovery processes.

80 Effect of different sugars on production of nisin Z and lactic acid

The effects of fermentation reactor and sugar type on nisin and lactic acid production are shown in Table 4.2. There was no interaction between these factors on either nisin or lactic acid production (P ≤ 0.05). Nisin production was highly enhanced when using sucrose as a carbon source. When replacing glucose with sucrose as a carbon source, the maximum nisin concentrations (IU/ml) were shown to increase by factors of

2.2 (from 697 ± 5 to 1,566 ± 156 IU/ml) and 1.9 (from 579 ± 62 to 1,100 ± 49 IU/ml) in suspension cell and biofilm reactors, respectively (Table 4.2). The sole significant difference in the sugar consumption pattern between suspension cell reactor and biofilm reactor was a shorter lag time in biofilm reactor, possibly as a result of higher biomass density in the system. At 80 g/l concentration, 50-60% of both glucose and sucrose were consumed in the first 10 h and the sugars were almost depleted after 24 h.

Table 4.2. Effects of fermentation types and sugars on nisin and lactic acid productions. Carbon Sources* Parameters Bioreactors Glucose Sucrose a c Suspension cell 697 ± 5 1566 ± 156 Max nisin (IU/ml) a b Biofilm 579 ± 62 1100 ± 49 a c Suspension cell 155 ± 17 274 ± 53 Max nisin production rate (IU/ml/h) ab bc Biofilm 184 ±18 264 ± 2 a ab Suspension cell 50 ± 0 54 ± 1 Final lactic acid concentration (g/l) b b Biofilm 56 ± 2 58 ± 0

* Carbon source at 8% w/v. Fermentations were conducted using bioreactor with pH control (n = 3). Values (in the same row) not marked by the same alphabet are significantly different (P ≤ 0.05).

81 A genetic linkage between sucrose utilization and nisin production has been reported (Gasson, 1984; James and Larry, 1986) with the effect of sucrose on nisin production in L. lactis varying from strain to strain. For example, glucose was reported as an optimal carbon source for nisin production in strain ATCC 11454 (Chandrapati and

O’Sullivan, 1998), while xylose was reported as the most efficient carbon source in strain

JCM 7638 (also called as IO-1) (Chinachoti et al., 1997). In this study, we chose to test only glucose and sucrose, the most cost-effective and available sugars for commercial production. According to the results, sucrose proved to be the better carbon source for biomass (Fig. 4.4) and nisin production (Fig. 4.5) for strain NIZO 22186. Sucrose also slightly enhanced the production of lactic acid (Table 4.2), which in this case is being considered as a by-product. The outcome that sucrose resulted in higher biomass and nisin concentration than glucose did not change the fact that in the presence of glucose, L. lactis would consumes glucose before sucrose due to the repressive effect of glucose on sucrose catabolic genes (Luesink et al., 1999).

The influence of accumulated lactic acid on nisin production is not well understood. According to a nisin production study of Van’t Hul and Gibbons (1997) using strain ATCC 11454, removal of lactic acid from the system encouraged cells to direct metabolism towards lactic acid production at the expense of nisin production. On the other hand, several recent attempts to remove lactic acid from the system indicated a suppressive effect of lactic acid on growth and nisin production (Kim, 1997; Li et al.,

2002b; Yu et al., 2002).

82 On the basis of nisin production alone, the biofilm reactor was only comparable or slightly worse than the suspension cell reactor. However, after considering the biofilm reactor’s advantages of a shorter lag time for nisin production and a less viscous fermentation broth with lower amount of biomass, which would be easier to process during downstream recovery, biofilm reactor still exhibits a potential for nisin production. Therefore, we investigated further improvements of nisin production in the biofilm reactor performance through varying of the medium compositions.

Evaluation of medium for nisin Z production in PCS biofilm reactor

The full-factorial experiment was designed to evaluate the effects of initial concentrations of sucrose, magnesium, and phosphate on nisin production in the PCS biofilm reactor. In all treatments the maximum nisin concentration correlated with the maximum nisin production rates, thus it is noted that both parameters were interchangeable representations of nisin production in this case. The results showed no interaction between factors and indicated that only initial sucrose concentration and magnesium sulfate exhibited significant effects on nisin production (Fig. 4.6). The medium compositions containing the lower level of sucrose and magnesium phosphate,

Medium 1 and No. 5, yielded the highest nisin production rates of 551 and 515 IU/ml/h, respectively (Table 4.3). The nisin obtained from the best performing medium (Medium

1) was approximately 2-time higher (2,208 IU/ml) than the value obtained previously

(1,100 IU/ml).

83

Main Effects Plot - Max Nisin Production Rate

Sucrose Magnesium Phosphate

380

350

320

290

260 Max NisinMax ProductionRate (IU/ml/hr) % % 2 4 4% 8% 0 0 1% 3% 0. 0.

Figure 4.6. Effects of sucrose, magnesium, and phosphate levels on maximum nisin production rate (IU/ml/h) in PCS biofilm reactor.

Table 4.3. Effect of different nutrient compositions on nisin and lactic acid productions.

Percent (% w/v) Average Final Nisin Production Medium Max Nisin Lactic Acid Rate (IU/ml/h) X1 X2 X3 (IU/ml) (g/l) a a a 1 4 0.02 1 2208 551 32 b b b 2 8 0.02 1 1057 264 58 b b a 3 4 0.04 1 1483 280 32 b b b 4 8 0.04 1 761 242 56 a a a 5 4 0.02 3 2073 515 30 b b c 6 8 0.02 3 825 192 47 b b a 7 4 0.04 3 1249 252 31 b b b 8 8 0.04 3 1006 305 54

• Fermentations were conducted using bioreactor with pH control (n = 3). Values not followed by the same letter are significantly different (P ≤ 0.05). • X1, X 2, and X 3 stand for sucrose, MgSO 4•7H 2O, and KH 2PO 4, respectively.

84 The suppressive effect of high initial sucrose concentration observed in this study was consistent with the catabolic repression reported in many studies (De Vuyst and

Vandamme, 1992; Matsusaki et al., 1996). This study demonstrated that high sucrose concentrations decreased both maximum nisin concentration and nisin production rate

(Table 4.3). For example, doubling the sucrose concentration in Medium 2 to 8% resulted in half the maximum nisin concentration and nisin production rate (1,057 IU/ml; 264

IU/ml/h) compared to Medium 1 (2,208 IU/ml; 551 IU/ml/h). Moreover, high initial sucrose concentration not only suppressed nisin production, but also significantly enhanced the production of lactic acid (Fig. 4.7). Compared to the medium with 4% sucrose concentration (Medium 1), Medium 2 (8% sucrose) yielded 1.8 times higher lactic acid production (Table 4.3).

Main Effects Plot - Final Lactic Acid

Sucrose Magnesium Phosphate

52

47

42

37

FinalLactic Acid (g/l)

32

% % 4% 8% 2 4 1% 3% .0 .0 0 0

Figure 4.7. Effects of sucrose, magnesium, and phosphate levels on final lactic acid concentration (g/l) in PCS biofilm reactor.

85 Although both MgSO 4•7H 2O, and KH 2PO 4 had been reported to have profound effects on nisin production with ATCC 11454 (Meghrous et al., 1992; Lv et al., 2004), this study demonstrated that only magnesium sulfate had a significant effect on nisin production with strain NIZO 22186 (Fig. 4.6). MgSO 4•7H 2O at the concentration of

0.04% w/v dramatically reduced the nisin production from that of 0.02% MgSO 4•7H 2O, contrasting with the optimal concentration of up to 0.057% reported for ATCC 11454

(Liu et al., 2003). However, many reports indicate that the effect of magnesium ion varies from strain to strain. While addition of magnesium sulphate did not have a considerable effect on nisin production with L. lactis strain MSU (Egorov et al., 1971) and strain IO-1

(Matsusaki et al., 1996), Meghrous et al. (1992) reported that addition of magnesium sulfate at 0.025% improved nisin production (from 157 AU/ml to 238 AU/ml) and reduced cell-adhered nisin (to 31%) in the strain ATCC 11454. To our knowledge this study was so far the first one to report the effect of magnesium ion on strain NIZO 22186.

As for KH 2PO 4, initial concentrations higher than 1% w/v had been reported to enhance nisin production, but they also prolonged the lag phase of strain NIZO 22186

(De Vuyst and Vandamme, 1993). This was contrary to results of this study, which indicated that increasing KH 2PO 4 up to 3% did not significantly affect the nisin production of strain NIZO 22186 in the complex medium. For example, Medium 5 (3%

KH 2PO 4) yielded the same level of nisin production as Medium 1 (1% KH 2PO 4), and the lag phase was not prolonged in our study. However, these could be a result of differences in pH control pattern and type of fermentation in the two studies. In De Vuyst and

Vandamme’s study, the culture was grown in uncontrolled-pH, suspension cell fermentation in which the stimulating effect of KH 2PO 4 might be more pronounced.

86 Compared to the optimal concentration of KH 2PO 4 reported (about 3%) for L. lactis strain

ATCC 11454 and MSU (Egorov et al., 1971; Li et al., 2002a), the suggested concentration in this study was much lower. This might indicate lower requirements of potassium ion for strain NIZO 22186 compared to other strains of L. lactis . Whether this lower requirement is a characteristic of strain NIZO 22186 or a characteristic of biofilm adaptation still needs to be investigated.

CONCLUSIONS

A biofilm reactor system using PCS proved to be a promising tool for nisin

production with L. lactis . Although its high-biomass density characteristic did not

contribute to a higher production rate, a significantly shorter lag time during nisin

production was observed. In this study, sucrose was selected as a better carbon source for

nisin production with L. lactis strain NIZO 22186. High levels of initial sucrose not only

suppressed the production of nisin, but also enhanced the production of by-product, lactic

acid. A high concentration of MgSO 4•7H 2O at 0.04% w/v resulted in a reduction of nisin

production, whereas KH 2PO 4 concentrations of 1 to 3% w/v did not have any significant

effect on growth or nisin production. The best performing complex media for nisin

production using PCS biofilm reactor consisted of 4% w/v sucrose, 0.02% w/v

MgSO 4•7H 2O, 1% w/v KH 2PO 4, 0.2% w/v NaCl, 1% peptone, and 1% yeast extract.

Overall, relative to the medium tested in the suspension cell reactor comparison

(8% glucose, 0.02% MgSO 4•7H 2O, and 1% KH 2PO 4), nisin production in the biofilm

reactor was enhanced 1.9 times when using sucrose as a carbon source and 3.8 times

87 when using the best performing medium. The time that nisin should be harvested varied

from 4 h in PCS biofilm reactor to 8-10 h in suspended cell reactor. In this study, the PCS

biofilm reactor was proved to be stable even after more than 1 year of operation, however

the maximum possible runs achieved shall be evaluated further.

A PCS biofilm reactor system appears to have several advantages for nisin production, such as a shorter lag phase and easier-to-process fermentation broth. Apart from their well-known benefits of being highly stable system with lower capital cost, biofilm reactors also hold great potential to be developed into a more productive process, such as continuous fermentation, without causing a wash-out. The suppressive effect of initial sucrose concentration observed in this study suggested that fed-batch fermentation with low initial sucrose concentration and continuous feeding of sucrose would be a better approach for nisin production. Fed-batch fermentation using biofilm reactor system and optimization of other culture conditions which can affect nisin production, such as the pH profile during fermentation, need to be investigated further.

88 REFERENCES

Amiali, M. N., C. Lacroix, and R. E. Simard. 1998. High nisin Z production by Lactococcus lactis UL719 in whey permeate with aeration. World J. Microbiol. Biotechnol. 14: 887-894. Baranova, I. P., and N. S. Egorov. 1969. Prikl. Biokhim. Mikrobiol. 5: 175. Bertrand, N., I. Fliss, and C. Lacroix. 2001. High nisin-Z production during repeated- cycle batch cultures in supplemented whey permeate using immobilized Lactococcus lactis UL719. Int. Dairy J. 11: 953-960. Bober, J, and A. Demirci. 2004. Nisin fermentation by Lactoccocus lactis subsp. lactis using plastic composite supports in biofilm reactors. Ag Eng Int: the CIGR J. Sci. Res. Dev. Manuscript FP 04-001 Vol. VI 15 pp. Buchman, G. W., S. Banerjee, and J. N. Hansen. 1988. Structure, expression, and evolution of a gene encoding the precursor of nisin, a small protein antibiotic. J. Biol. Chem. 263(31): 16260-16266. Cabo, M. L., M. A. Murado, M. P. Gonzalez, and P. Pastoriza. 2001. Effects of aeration and pH gradient on nisin production. A mathematical model. Enz. Microbial. Technol. 29: 264-273. Chandrapati, S., and D. J. O’Sullivan. 1998. Procedure for quantifiable assessment of nutritional parameters influencing nisin production by Lactococcus lactis subsp. lactis . J. Biotechnol. 63: 229-233. Cheigh, C. I., H. J. Choi, H. Park, S. B. Kim, M. C. Kook, T. S. Kim, J. K. Hwang, and Y. R. Pyun. 2002. Influence of growth conditions on the production of a nisin-like bacteriocin by Lactococcus lactis subsp. lactis A164 isolated from kimchi. J. Biotechnol. 95: 225-235. Chinachoti, N., N. Endo, K. Sonomoto, and A. Ishizaki. 1997. Bioreactor systems for efficient production and separation of nisin Z using Lactococcus lactis IO-1. J. Fac. Agric. Kyushu Univ. 43: 421-436.

89 Cotton, J. C., A. L. Pometto III, and J. Gvocdenovic-Jeremic. 2001. Continuous lactic acid fermentation using a plastic composite support biofilm reactor. Appl. Microbiol. Biotechnol. 57: 626-630. Demirci, A., A. L. Pometto III, K-L. G. Ho. 1997. Ethanol production by Saccharomyces cerevisiae in biofilm reactors. J. Ind. Microbiol. 19: 299-304. Demirci, A., A. L. Pometto III, K-L. G. Ho. 1995. Continuous ethanol production in biofilm reactors containing plastic composite rings and disks. Proc.-Biomass Conference of the Americas: Energy, Environment, Agricultural and Industry, 2 nd , Aug, 21-24, Portland, Oregon, National Renewable Energy Laboratory, Golden, Colorado. De Vuyst, L. 1994. Nisin production variability between natural Lactococcus lactis subsp. lactis strains. Biotechnol. Lett. 16: 287-292. De Vuyst, L., and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. De Vuyst, L., and E. J. Vandamme. 1993. Influence of the phosphorus and nitrogen source on nisin production in Lactococcus lactis supsp. lactis batch fermentations using a complex medium. Appl. Microbiol. Biotechnol. 40: 17-22. Doleyres, Y., I. Fliss, and C. Lacroix. 2004. Increased stress tolerance of Bifidobacterium longum and Lactococcus lactis produced during continuous mixed-strain immobilized-cell fermentation. J. Appl. Microbiol. 97(3): 527-539. Egorov, N. S., I. P. Baranova, and Y. I. Kozlova. 1971. Optimization of nutrient medium composition for the production of the antibiotic nisin by Streptococcus lactis . Mikrobiologiia 40(6): 993-998. Gasson, M. J. 1984. Transfer of sucrose fermenting ability, nisin resistance and nisin production into Streptococcus lactis 712. FEMS Microbiol. Lett. 21: 7-10. Hirsch, A. 1951. Growth and nisin production of a strain of Streptococcus lactis . J. Gen. Microbiol. 5: 208-221. Hirsch, A., and E. Grinsted. 1951. The differentiation of the lactic streptococci and their antibiotics. J. Dairy Res. 18: 198-204.

90 Ho, K-L. G., A. L. Pometto III, and P. N. Hinz. 1997a. Ingredient selection for plastic composite supports for L-(+)-lactic acid biofilm fermentation by Lactobacillus casei subsp. Rhamnosus . Appl. Environ. Microbiol. 63: 2516-2523. Ho, K-L. G., A. L. Pometto III, P. N. Hinz, and A. Demirci. 1997b. Nutrient leaching and end product accumulation in plastic composite supports for L-(+)-lactic acid biofilm fermentation. Appl. Environ. Microbiol. 63: 2524-2532. Ho, K-L. G., A. L. Pometto III, and P. N. Hinz. 1997c. Optimization of L-(+)-lactic acid production by ring and disc plastic composite supports through repeated-batch biofilm fermentation. Appl. Environ. Microbiol. 63: 2533-2542. Hurst, A., and G. J. Dring. 1968. The relation of the length of lag phase of growth to the synthesis of nisin and other basic proteins by Streptococcus lactis grown under different conditions. J. Gen. Microbiol. 50: 383-390. James, L. S., and L. M. Larry. 1986. Partial characterization of the genetic basis for sucrose metabolism and nisin production in S treptococcus lactis . Appl. Environ. Microbiol. 51: 57-64. Kim, W. S. 1997. Nisin production by Lactococcus lactis using two-phase batch culture. Lett. Appl. Microbiol. 25: 169-171. Kleerebezem, M. 2004. Quorum sensing control of lantibiotic production: nisin and subtilin autoregulate their own biosynthesis. Peptides 25(9): 1405-1414. Kozlova, Y. I., N. S. Egorov, I. P. Baranova, and V. N. Maksimov. 1972. Kinetics of Streptococcus lactis metabolism on initial and optimized media. Microbiol. 41: 893-897. Kozlova, Y. I., T. I. Golikova, I. P. Baranova, and N. S. Egorov. 1979. Investigation of

the influence of K 2HPO 4 on the growth of Streptococcus lactis and nisin synthesis at constant pH values of the medium. Mikrobiologiya 48: 443-336. Krishnan, S., L. R. Gowda, M. C. Misra, and N. G. Karanth. 2001. Physiological and morphological changes in immobilized L. plantarum NCIM 2084 cells during repeated bach fermentation for production of lactic acid. Fd. Biotechnol. 15: 193- 202.

91 Kuipers, O. P., M. M. Beerthuyzen, P. G. de Ruyter, E. J. Luesink, and W. M. de Vos. 1995. Autoregulation of nisin biosynthesis in Lactococcus lactis by signal- transduction. J. Biol. Chem. 270: 27299-27304. Kunduru, R. M., and A. L. Pometto III. 1996a. Evaluation of plastic composite supports for enhanced ethanol production in biofilm reactors. J. Ind. Microbiol. 16: 241-248. Kunduru, R. M., and A. L. Pometto III. 1996b. Continuous ethanol production by Zymomonas mobilis and Saccharomyces cerevisiae in biofilm reactors. J. Ind. Microbiol. 16: 249-256. Li, C., J. Bai, J., Z. Cai, Z., and F. Ouyang. 2002a. Optimization of a culture medium for bacteriocin production by Lactococcus lactis using response surface methodology. J. Biotechnol. 93: 27-34. Li, C., O. Y. Fan, and J. H. Bai. 2002b. Extractive cultivation of Lactococcus lactis using

a polyethylene glycol/MgSO 4.7H 2O aqueous two-phase system to produce nisin. Biotech. Lett. 22(10): 843-847. Liu, C., Y. Liu, W. Liao, Z. Wen, and S. Chen. 2003. Application of statistically-based experimental designs for the optimization of nisin production from whey. Biotechnol. Lett. 25: 877-882. Lv, W., W. Cong, and Z. Cai. 2004. Nisin production by Lactococcus lactis subsp. lactis under nutritional limitation in fed-batch culture. Biotechnol. Lett. 26: 235-238. Luesink, E. J., J. D. Marugg, O. P. Kuipers, and W. M. de Vos. 1999. Characterization of the divergent sac BK and sac AR operons, involved in sucrose utilization by Lactococcus lactis. J. Bacteriol. 181(6): 1924-1926. Matsusaki, H., N. Endo, K. Sonomoto, and A. Ishizaki. 1996. Lantibiotic nisin Z fermentative production by Lactococcus lactis IO-1: relationship between production of the lantibiotic and lactate and cell growth. Appl. Microbiol. Biotechnol. 45: 36-40. Mattick, A. T. R., and A. Hirsch. 1947. Further observations on an inhibitory substance (nisin) from lactic streptococci. Lancet 2: 5-12. Meghrous, J., E. Huot, M. Quittelier, and H. Petitdemange. 1992. Regulation of nisin biosynthesis by continuous cultures and by resting cells of Lactococcus lactis supsp. lactis. Res. Microbiol. 143: 897-890.

92 Melo, L. F., and R. Oliveira. 2001. Biofilm reactors. In Multiphase bioreactor design, eds. J. M. S. Cabral, M. Mota, and J. Tramper. 271-308. Taylor & Francis Inc, New York. Miller, G. L. 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31: 426. Pometto III, A. L., A. Demirci, and K. E. Johnson. 1997. Immobilization of microorganisms on a support made of synthetic polymer and plant material. US Patent No 5595893. Reiter, B., and D. Oram. 1962. Nutritional studies on cheese starters. J. Dairy Res. 29: 63-77. Sonomoto, K., N. Chinachoti, N. Endo, and A. Ishizaki. 2000. Biosynthetic production of nisin Z by immobilized Lactococcus lactis IO-1. J. Mol. Catalys. B Enz. 10: 325- 334. Teixeira de Mattos, M. J., J. P. de Boer, G. R. Zoutberg, and O. M. Neijssel. 1994. Metabolic shift analysis at high cell densities. FEMS Microbiol. Rev. 14: 21-28. Urbance, S. E., A. L. Pometto III, A. A. DiSpirito, and A. Demirci. 2003. Medium evaluation and plastic composite support ingredient selection for biofilm formation and succinic acid production by Actinobacillus succinogenes . Fd. Biotechnol. 17: 53-65. U. S. Food and Drug Administration. 1988. Nisin preparation: Affirmation of GRAS status as direct human food ingredient. Federal Register, 53, April 6. Van’t Hul, J. S., and W. R. Gibbons. 1996. Concentration and recovery of the bacteriocin nisin from Lactococcus lactis subsp. lactis. Biotechnol. Appl. Biochem. 24: 251- 256. Van’t Hul, J. S., and W. R. Gibbons. 1997. Neutralization/recovery of lactic acid from Lactococcus lactis : effects on biomass, lactic acid, and nisin production. World J. Microbiol. Biotechnol. 13: 527-532. Velazquez, A. C., A. L. Pometto III, K-L. G. Ho, and A. Demirci. 2001. Evaluation of plastic-composite supports in repeated fed-batch biofilm lactic acid fermentation by Lactobacillus casei. Appl. Microbiol. Biotechnol. 55: 434-441.

93 Wan, J., M. W. Hickey, and M. J. Coventry. 1995. Continuous production of bacteriocins, brevicin, nisin and pediocin, using calcium alginate-immobilized bacteria. J. Appl. Bacteriol. 79: 671-676. Wang, J. 2000. Production of citric acid by immobilized Aspergillus niger using a rotating biological contactor (RBC). Biores. Tech. 75: 245-247. Yang, R., and B. Ray. 1994. Factors influencing production of bacteriocins by lactic-acid bacteria. Fd. Micro. 11(4): 281-291. Yu, P. L., N. W. Dunn, and W. S. Kim. 2002. Lactate removal by anionic-exchange resin improves nisin production by Lactococcus lactis. Biotechnol. Lett. 24: 59-64.

94 CHAPTER 5

EFFECTS OF pH PROFILES ON NISIN PRODUCTION

IN BIOFILM REACTOR *

ABSTRACT

Apart from its widely accepted commercial applications as a food preservative, nisin emerges as a promising alternative in medical applications for bacterial infection in both humans and livestock. Improving nisin production through optimization of fermentation parameters would make nisin more cost-effective for various applications.

Since nisin production by Lactococcus lactis NIZO 22186 was highly influenced by the pH profile employed during fermentation, three different pH profiles were evaluated in this study; i) a constant pH profile at 6.8 (profile Const12), ii) a constant pH profile with autoacidification at 4 h (profile Const4), and iii) a step-wise pH profile with pH adjustment every 2 h (profile Stepwise). The results demonstrated that the low pH-stress exerted during the first 4 h of fermentation in profile Stepwise detrimentally affected nisin production, resulting in a very low maximum nisin concentration (593 IU/ml). On the other hand, growth and lactic acid production were only slightly delayed, indicating that the loss in nisin production was not a result of lower growth or shifting of metabolic activity toward lactic acid production. profile Const4, in which pH was allowed to drop freely via autoacidification after 4 h of fermentation, was found to yield almost 1.9-time higher nisin (3,553 IU/ml) than profile Const12 (1,898 IU/ml), possibly as a result of less

* This chapter was accepted to be published in Journal of Applied Microbiology and Biotechnology, pp. 1- 8, Dec. 6, 2005.

95 adsorption of nisin onto producer cells. Therefore, a combination of constant pH and autoacidification period (profile Const4) was recommended as the pH profile during nisin production in a biofilm reactor.

INTRODUCTION

Nisin is the most well studied and widely used bacteriocin produced by

Lactococcus lactis during fermentation processes. Biofilm reactors, like other immobilized reactors, not only yield higher production rates, but also possess greater potential compared with suspended cell reactors for development into a more efficient continuous culture. Combining exceptional stability and lower nutrient requirements, biofilm reactors exhibit high potential for applications in the large-scale production of many products such as alcohols, organic acids, antibiotics, and enzymes.

As for nisin production, our previous study suggested that a high-biomass density characteristic of the biofilm reactor resulted in its advantages over a free-cell reactor as follows; i) significantly shorter lag time on nisin production, and ii) easier-to-process fermentation broth with lower viscosity and biomass content (Chapter 4). Plastic

Composite Support (PCS) was an extrusion product of a mixture of polyprolylene and various agricultural products (Demirci et al., 1997). The polypropylene acted as a matrix to integrate the mixture of agricultural products, which provided essential nutrients to sustain cell growth. Thus, PCS not only provided an ideal physical structure for biofilm formation, but also supplied a slow release of nutrients, which was reported to benefit the nisin production (De Vuyst and Vandamme, 1993).

96 The pH was widely accepted as one of the most significant factors on nisin production. The optimal pH for nisin production was found to vary between 5.5 and 6.8 depending on the strains of nisin-producer (De Vuyst and Vandamme, 1992; Meghrous et al., 1992; Matsusaki et al., 1996). Controlling pH via a periodical realkalization of medium was shown to enhance both growth and nisin production of L. lactis in many cases (Hirsch, 1951; De Vuyst and Vandamme, 1992; Bober and Demirci, 2004). Thus, nisin fermentation was generally conducted using a constant pH profile controlled by an addition of bases. It was only recently that the pH-drop gradient has been reported to have a stimulating effect on nutrient consumption and nisin production (Cabo et al.,

2001; Guerra and Pastrana, 2003a; 2003b).

The adsorption of nisin by the producer cells, as a result of attraction between cationic nisin molecule and anionic bacterial cell surface, was highly significant at the pH used for nisin production (Hurst and Dring, 1968; Grushina et al., 1980; Yang et al.,

1992). For example, Yang et al. (1992) showed that more than 95% of nisin was adsorbed on the producer cells at pH 6.8, but a complete loss of adsorption was found at pH 3.

Thus, a major loss of nisin during fermentation was suspected as a result of a constant pH profile approach.

Since it was clearly proven by Yang et al. (1992) that 93% of adsorbed nisin could be recovered via pH adjustment, changing the pH profile could be an option to prevent the loss. The most suitable pH profile for nisin production is still inconclusive and more options need to be explored, especially in biofilm system in which the effect of pH has never been evaluated. In several instances, biofilm system has been shown to respond differently to the cultural conditions than suspended-cell system (Chung and

97 Park, 1983; Norwood and Gilmour, 2000; Velazquez et al., 2001). Thus, in this study, effects of various pH profiles on nisin production have been evaluated for the PCS biofilm reactor in attempt to develop a more effective nisin production system.

MATERIALS AND METHODS

Microorganisms and media

L. lactis subsp. lactis (NIZO 22186) was used in this study as the nisin Z- producing strain. Lactobacillus sakei (ATCC 15521) was used as the nisin sensitive test organism. L. lactis and L. sakei were grown at 30 °C for 14 h in a complex medium and

Lactobacillus MRS broth (Difco Laboratories, Detroit, MI), respectively, unless stated otherwise. The medium consisted of 40 g of sucrose; 10 g of peptone (Amber Ferm

4015G, Universal Flavors, Milwaukee, WI); 10 g of yeast extract (Ardamine Z, Sensient

Bionutrient, Indianapolis, IN); 10 g of KH 2PO 4; 2 g of NaCl; and 0.2 g of MgSO 4•7H 2O per liter of deionized (DI) water. The initial pH of the medium was adjusted to 6.8 using

4N NaOH. For a long-term storage, all stock cultures were maintained at -80°C in 20% v/v glycerol in deionized water.

Plastic Composite Support (PCS)

PCS tubes were manufactured at Iowa State University using a twin-screw co- rotating Brabender PL2000 extruder (model CTSE-V; C.W. Brabender Instruments, Inc.,

South Hackensack, NJ) as described by Ho et al. (1997). The composition of PCS used in this study was chosen according to biofilm formation, nisin production, nitrogen leaching

98 rate, and % nitrogen content (Chapter 4). Polypropylene (50%) and other ingredients

(35% Soybean hulls, 5% Soybean flour, 5% Yeast extract, 5% Dried bovine albumin, and

0.272% of sodium acetate, 0.0004% of MgCl 2.6H 2O, and 0.002% of NaCl) were mixed together before being extruded at 13 rpm through a medium pipe die with barrel temperatures of 200, 220, and 200 °C and a die temperature of 165 °C. The extruded tubes with a wall thickness of 2.5 mm and an outer diameter of 10.5 mm were cut into 7.5-cm length tubes with both ends cut at a 45 degree angle to allow better flow of medium inside the tubes.

Nisin fermentation in bioreactor

Nisin fermentations were conducted in a 1.25-l Bioflo II fermentor (New

Brunswick Scientific, Edison, NJ) at 30 °C with agitation at 100 rpm. For the biofilm reactor, twelve PCS tubes were attached to the agitator shaft in a grid-like fashion, six rows of two parallel tubes (Fig. 4.1). The reactor vessel with PCS was sterilized with water at 121 °C, while sucrose and nitrogenous components with mineral salts were sterilized separately and added to the reactor aseptically. After inoculation with a 14-h culture of L. lactis (1% v/v), at least five fermentation runs were performed in order to establish biofilm formation on PCS supports. The pH was controlled at 6.8 by adding 4 N

NaOH. Once a stable biofilm was formed, repeated-batch fermentations were carried out in the biofilm reactor by pumping out the previous fermentation broth before aseptically adding a sterile fresh medium. Suspended-cell fermentation was conducted using the same procedure except that the medium and reactor without PCS were sterilized and inoculated with 14-h culture of L. lactis at the beginning of each batch.

99 Experimental design

Three different pH-control profiles as described follow were evaluated in (Fig.

5.1); i) pH was constantly controlled at 6.8 during the fermentation period (profile

Const12), ii) pH was controlled at 6.8 from 0 to 4 h, and then allowed to drop freely via autoacidification from 4 to 12 h before being maintained at constant pH of 6.8 until 24 h

(in order to avoid the sloughing off of the biofilm in acidic condition) (profile Const4), and iii) pH was controlled using a step-wise profile, in which the pH was allowed to drop freely and was adjusted back to 6.8 every 2 h for 12 h before being maintained at 6.8 until

24 h (profile Stepwise). In this study, the commonly used constant pH profile (profile

Const12) was served as a control for comparison. In profile Const4, the pH drop was introduced after 4 h of fermentation, a period in which there was sufficient autoacidification while the nisin concentration approached the maximum value. An interval period of 2 h was chosen for stepwise pH profile (profile Stepwise) according to the results of Cabo et al. (2000).

In order to ensure the consistency of biofilm in the reactor, the medium was change on a regular basis of 24 h. Before the data was collected for each tested pH profile, the biofilm reactor was run using the tested pH profile for at least two batches

(acclimation period) in order to allow the biofilm reactor to adjust to the new cultural conditions. Then, after each treatment the biofilm was maintained using the constant pH profile with regular changes of medium. The thickness of biofilm on PCS was controlled via the use of a constant agitation rate (100 rpm) and the previously described maintenance procedure through out the entire experiment. For each pH profile, three repeated batches were performed for data collections.

100 7.0

6.5

6.0 pH 5.5

5.0 (a)

4.5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

7.0

6.5

6.0 pH 5.5

5.0 (b) 4.5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

7.0

6.5

6.0 pH 5.5

5.0 (c) 4.5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 5.1. Actual pH profiles of nisin fermentation by L. lactis when using (a) profile Const12; (b) profile Const4; and (c) profile Stepwise.

101 Finally, the pH profile, which gave the highest nisin production, was tested in the suspended-cell system using the same medium and culture conditions. The obtained results from the PCS biofilm system and the suspended-cell system were then compared.

Analysis

Samples were collected every 1 or 2 h and analyzed for cell density, sugar consumption, nisin production, and lactic acid production.

Suspended biomass. Suspended-cell density was estimated by optical density at

600 nm using a spectrophotometry DU series 500 (Beckman, Fullerton, CA). Sterile fermentation medium was used as a blank. Optical density values were converted into biomass concentration by using a standard curve (y = 0.1663x – 0.0199; y: biomass (g/l) and x: optical density at 600 nm).

Nisin. Sample of fermentation broth was immediately adjusted to pH 3.0 using concentrated HCl and 0.1% v/v of Tween 20 (J.T. Baker, Phillipsburg, NJ) was added in order to avoid any non-specific adsorption of nisin on the container’s surfaces. Then, the sample was heated at 90 °C for 5 min to eliminate the protease activity in the broth. After centrifugation at 2,000 x g, 4 °C for 10 min (National Labnet Co., mini centrifuge Model

C-1200), clear supernatant was collected and kept frozen at -20 °C until analysis. Nisin quantifications were performed by an agar well diffusion bioassay, using L. sakei and pre-diffusion at 4 °C, as described in Chapter 3.

Sucrose. The fermentation broth was centrifuged at 2,000 x g for 10 min

(National Labnet Co., mini centrifuge Model C-1200) and clear supernatant was collected for analysis. For initial acid hydrolysis, 100 µl of concentrated HCl was added to 5 ml of

102 the sample. After allowing the hydrolysis to proceed at 90 °C for 5 min, 0.25 ml of 5N

KOH solution was added to neutralize the acid. The hydrolyzed samples were then analyzed using the DNS Method as described by Miller (1959).

Glucose and lactic acid. Glucose and lactic acid concentration were determined using a high performance liquid chromatograph (HPLC) equipped with a refractive index detector (Waters, Franklin, MA). Components were separated on a Bio-Rad Aminex

HPX-87H column (300 x 7.8 mm) (Bio-Rad, Richmond, CA) with 0.012 N sulfuric acid as a mobile phase at a flow rate of 0.8 ml/min with an injection volume of 20 µl and a column temperature of 65 °C. Before injection, the samples were centrifuged (2,000 x g,

4°C) for 10 min and filtered through 0.22 micron filters (13 mm diameter disc filters,

Millipore, Bedford, MA).

Statistical analysis

All treatments were replicated three times. The significant difference of the results was evaluated using the Generalized Linear Model (GLM) (with p < 0.05) and Tukey’s

HSD multiple comparison module of the MINITAB Statistical Software package

(Release 13.30) (State College, PA).

103 RESULTS

During fermentation, L. lactis (characterized as moderately acid-tolerant lactic cocci) autoacidifies the environment mainly through production of lactic acid. Fig. 5.1 shows the actual pH profiles during fermentation when using all tested profiles. In profile

Const4, as soon as the addition of NaOH was stopped at 4 h, the pH of the fermentation medium sharply decreased by 1.0 pH unit in 1 h and then slowly decreased further until reaching the minimum pH of 5.0 at 12 h (Fig. 5.1b). The period of 4 h before autoacidification was selected because it allowed the system to reach a sufficient amount of growth (Fig. 5.2), maximum nisin production (Fig. 5.3), and adequate active lactic acid production to acidify the fermentation medium (Fig. 5.4). The longer period of 6 h also has been evaluated, but the results were unsuccessful due to an insufficient autoacidification. The actual pH profile of profile Stepwise (Fig. 5.1c) indicated a reduction in the system’s ability to autoacidify the medium over time as a result of carbon source depletion. It was noted that the autoacidification in profile Stepwise always slowed down between 2 and 3 h, possibly as a result of lower growth rate caused by the pH drop during the first two hours (Fig. 5.2).

Profile Const4, in which the pH was allowed to drop freely after the nisin production reached its peak, was designed to release the nisin absorbed on the producer cells back to the liquid broth. The obtained results seemed to support the idea as the nisin concentration obtained when using profile Const4 (3,553 IU/ml) was significantly higher than the value obtained from profile Const12 (1,898 IU/ml) (Fig. 5.3).

104 0.8

0.7 0.6

0.5

0.4 0.3

Biomass (g/l) Biomass BF, Profile Const12

0.2 BF, Profile Const4

0.1 BF, Profile Stepwise 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24

Time (h)

Figure 5.2. Growth of L. lactis (in liquid broth) using the PCS biofilm reactor (BF) with different pH profiles.

5000

4500 BF, Profile Const12 4000 BF, Profile Const4

3500 BF, Profile Stepwise

3000 SC, Profile Const4

2500

2000 Nisin (IU/ml)

1500

1000

500

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 5.3. Nisin production of L. lactis using the PCS biofilm reactor (BF) and the suspended-cell reactor (SC) when using different pH profiles.

105 50

40

30

20

(g/l) acid Lactic BF, Profile Const12 BF, Profile Const4 10 BF, Profile Stepwise

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 5.4. Lactic acid production of L. lactis from the PCS biofilm reactor (BF) when using different pH profiles.

60

BF, Profile Const12 50

BF, Profile Const4

40 BF, Profile Stepwise

30

(g/l) Sucrose 20

10

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 5.5. Sucrose consumption of L. lactis from the PCS biofilm reactor (BF) when using different pH profiles.

106 Furthermore, the results demonstrated no significant difference in either sucrose consumption or lactic acid production from both profiles, possibly because at 4 h most of the sucrose was consumed and the lactic acid production was already slow down (Fig. 5.4 and 5.5). Introduction of profile Const4 to the suspended-cell system resulted in lower nisin production (2,018 IU/ml), indicating that higher biomass density in PCS biofilm system not only shortened the lag phase of nisin production, but also increased the amount of nisin produced by the system.

Difference in nisin production from 0-4 h when using profile Const12 and Const4 revealed the remaining effects of different culture conditions, which were previously introduced to the system during the acclimation period. Interestingly, the remaining effect from exposing the system with profile Const4 positively affected the nisin production of the PCS biofilm system. Furthermore, in profile Const4, decreasing nisin concentration between 6 to 12 h of fermentation, when the pH was maintained at 5.0-5.3, suggested an active activity of nisin-degrading enzyme from the producer cells even at low pH.

The results clearly indicated that the step-wise profile, in which the producer cells had to experience higher level of pH-stress, resulted in a drastic decrease of nisin production (Fig. 5.3). However, the fact that growth, sugar consumption, and lactic acid production from profile stepwise (Figs. 5.2, 5.4, and 5.5) were only slow down and finally reached the same levels as other pH profiles indicated an uncoupling of growth and nisin production. Thus, lower nisin production in profile Stepwise was not simply a result of lower growth rate or shifting of metabolic activities toward lactic acid production. Compared to other profiles, the nisin production obtained from step-wise pH profile was extremely low (593 IU/ml). Nevertheless, consistent results of sucrose

107 consumption and productions of nisin and lactic acid (Figures 5.3, 5.4, and 5.5) indicated a consistence of the PCS biofilm system between different batches.

DISCUSSIONS

Nisin production, which is normally conducted under a constant optimal pH value near neutrality (pH 6.0-6.8), shows considerable loss due to the adsorption of nisin onto the producer cells. Owing to the fact that adsorption of nisin on the cells is pH-dependent, reversible mechanism (Yang et al., 1992), in this study we attempted to recover the adsorbed nisin through a process of autoacidification of L. lactis (in profile Const4).

Compared to the constant pH profile (profile Const12), profile Const4 allowed us to recover almost 1.9-time higher nisin just by allowing the pH to drop freely after 4 h of fermentation. However, like other bacteriocins, the final pH of the culture not only influenced the adsorption of nisin onto the producer cells, but also likely affected the post-translational and transport processes of nisin (Biswas et al., 1991; Guerra and

Pastrana, 2003a). Thus, the higher nisin obtained from pH profile Const4 might not simply be a result of nisin being released from the producer cells, but a result of more favorable condition for post-translational and transport processes of nisin. Nevertheless, profile Const4 is very practical and can be applied in large-scale production with lower requirement of the amount of base used for pH control.

Although bacteria can survive dramatic changes in cytosolic composition, the cytoplasmic pH should not be too far from neutrality, which is optimal for most metabolisms. During growth of acid-tolerant fermentative bacteria like L. lactis , the

108 intracellular pH (pH i ) decreases as the extracellular pH (pH ex ) decreased in order to maintain a constant pH gradient rather than a constant pH i (Nannen and Hutkins, 1991;

Cook and Russell, 1994). By allowing the pH ex to drop freely via autoacidification, a dynamic intensification of stress on the cells can be created (Even et al., 2002). For L. lactis, as the pH ex decreases, the pH in of both exponential and stationary-phase cells also drops accordingly in about 5-15 min in order to maintain the ∆pH values of 0.5 to 0.8 pH unit (Siegumfeldt et al., 2000). Thus, lowering of pH ex through autoacidification not only affected the adsorption of nisin on the producer cells, but also affected the pH in which directly associated with metabolic activities, including nisin production and transportation. It should be noted that the 2 h period used in the step-wise pH profile was sufficient for the producer cells to experience the internal pH drop. At lower pH, various enzymes of central metabolism were inhibited. Growth was also stopped as the energy derived from sugar catabolism was reoriented toward cytoplasmic alkalization via

ATPase instead of biomass synthesis (Even et al., 2002). Guerra and Pastrana (2002) clearly showed that both growth and bacteriocin production of L. lactis subsp. lactis stop when pH ex fell below 5. In this study, both profile Const4 and Stepwise created a pH stress on the cells, but profile Stepwise introduced the unfavorable environment earlier in the growth cycle, which was accountable for synthesis of genetic materials and enzymes.

As a result, nisin yield obtained from profile Stepwise was considerably lower compared to the other two profiles.

The stress exerted by the extracellular pH variation seemed to be extremely unfavorable for nisin production by L. lactis NIZO 22186, in contrast to the previous report by Cabo et al. (2001) in which a pH-drop gradient enhanced both nisin production

109 (approximately 4-fold when pH was re-alkalized every 3 h) and nutrient consumption.

However, it was noted that the time required to acidify the medium in the first cycle in

Cabo’s study was at least 3-4 h longer than that was observed in this study, possibly as a result of lower biomass density in the free-cell bioreactor. This lag time might provide the cells with a favorable environment to adjust and synthesize the genetic materials and enzymes required for growth and nisin production. Moreover, it should be noted that the

L. lactis strains used in those studies seemed to yield better growth and nisin production in uncontrolled pH condition compared to other producer strains.

Overall, results of this study have shown that higher nisin can be recovered by using the autoacidification ability of L. lactis . The results from the suspended-cell system with profile Const4 (Fig. 5.3) revealed two significant advantages of the PCS biofilm system over the suspended-cell system; 1) the biofilm system could fasten the nisin production by at least 4 h, and 2) when combining with profile Const4, higher amount of nisin was produced and recovered from the biofilm system. On the other hand, periodically re-alkalizing of the culture (a step-wise pH profile) had detrimental effects on culture of L. lactis NIZO 22186, especially on nisin production. Therefore, these results confirmed that nisin production was very sensitive to pH treatments and even a minor adjustment could result in a considerable change in nisin yield. Combining effects of various pH profiles and fed-batch culture on nisin production using a biofilm reactor as well as the nisin-degrading activity of the producer cells need to be investigated further.

110 REFERENCES

Biswas, S. R., P. Ray, M. C. Johnson, and B. Ray. 1991. Influence of growth conditions on the production of a bacteriocin, pediocin AcH by Pediococcus acidilactici H. Appl. Environ. Microbiol. 57: 1265-1267. Bober, J., and A. Demirci. 2004. Nisin fermentation by Lactoccocus lactis subsp. lactis using plastic composite supports in biofilm reactors. Agricultural Engineering International: the CIGR J Sci Res Dev Manuscript FP 04-001 Vol. VI 15 pp. Cabo, M. L., M. A. Murado, M. P. Gonzalez, and L. Pastoriza. 2001. Effects of aeration and pH gradient on nisin production. A mathematical model. Enz. Microbial. Technol. 29: 264-273. Chung, I. J., and Y. S. Park. 1983. Ethanol fermentation by S. cerevisiae in a bioreactor packed vertically with ceramic rods. Proc-Pac Chem Eng Congr 3rd, Korean Inst Chem Eng, Seoul, Korea, 4: 174-179. Cook, G. M., and J. B. Russell. 1994. The effect of extracellular pH and lactic acid on pH homeostasis in Lactococcus lactis and Streptococcus bovis. Curr. Microbiol. 28: 165-168. Demirci, A., A. L. Pometto III, and K-LG Ho. 1997. Ethanol production by Saccharomyces cerevisiae in biofilm reactors. J. Ind. Microbiol. 19: 299-304. De Vuyst, L., and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. De Vuyst, L., and E. J. Vandamme. 1993. Influence of the phosphorus and nitrogen source on nisin production in Lactococcus lactis supsp. lactis batch fermentations using a complex medium. Appl. Microbiol. Biotechnol. 40: 17-22. Even, S., N. D. Lindley, P. Loubiere, and M. Cocaign-Bousquet. 2002. Dynamic response of catabolic pathways to autoacidification in Lactococcus lactis : transcript profiling and stability in relation to metabolic and energetic constraints. Mol. Microbiol. 45: 1143-1152.

111 Grushina, V. A., I. P. Baranova, and N. S. Egorov. 1980. Adsorption of the antibiotic, nisin, to Streptococcus lactis cells. Antibiotiki 25: 495-499. Guerra, N. P., and L. Pastrana. 2002. Modelling the influence of pH on the kinetics of both nisin and pediocin production and characterization of their functional properties. Proc. Biochem. 37: 1005-1015. Guerra, N. P., and L. Pastrana. 2003a. Influence of pH drop on both nisin and pediocin production by Lactococcus lactis and Pediococcus acidilactici. Lett. Appl. Microbiol. 37: 51-55. Guerra, N. P., and L. Pastrana. 2003b. Enhancement of nisin production by Lactococcus lactis in periodically re-alkalized cultures. Biotechnol. Appl. Biochem. 38: 157-167. Hirsch, A. 1951. Growth and nisin production of a strain of Streptococcus lactis . J. Gen. Microbiol. 5: 208-221. Ho, K-L. G., A. L. Pometto III, and P. N. Hinz. 1997c. Optimization of L-(+)-lactic acid production by ring and disc plastic composite supports through repeated-batch biofilm fermentation. Appl. Environ. Microbiol. 63: 2533-2542. Hurst, A., and G. J. Dring. 1968. The relation of the length of lag phase of growth to the synthesis of nisin and other basic proteins by Streptococcus lactis grown under different conditions. J. Gen. Microbiol. 50: 383-390. Matsusaki, H., N. Endo, K. Sonomoto, and A. Ishizaki. 1996. Lantibiotic nisin Z fermentative production by Lactococcus lactis IO-1: relationship between production of the lantibiotic and lactate and cell growth. Appl. Microbiol. Biotechnol. 45: 36-40. Meghrous, J., E. Huot, M. Quittelier, and H. Petitdemange. 1992. Regulation of nisin biosynthesis by continuous cultures and by resting cells of Lactococcus lactis supsp. lactis. Res. Microbiol. 143: 897-890. Miller, G. L. 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31: 426. Nannen, N. L., and R. W. Hutkins. 1991. Intracellular pH effects in lactic acid bacteria. J. Dairy Sci. 74: 741-746.

112 Norwood, D. E., and A. Gilmour. 2000. The growth and resistance to sodium hypochlorite of Listeria monocytogenes in a steady-state multispecies biofilm. J. Appl. Microbiol. 88: 512-520. Siegumfeldt, H., K. B. Rechinger, and M. Jakobsen. 2000. Dynamic changes of intracellular pH in individual lactic acid bacterium cells in response to a rapid drop in extracellular pH. Appl. Environ. Microbiol. 66: 2330-2335. Velazquez, A. C., A. L. Pometto III, K-L. G. Ho, and A. Demirci. 2001. Evaluation of plastic-composite supports in repeated fed-batch biofilm lactic acid fermentation by Lactobacillus casei. Appl. Microbiol. Biotechnol. 55: 434-441. Yang, R., M. C. Johnson, and B. Ray. 1992. Novel method to extract large amount of bacteriocins from lactic acid bacterial. Appl. Environ. Microbiol. 58: 3355-3359.

113 CHAPTER 6

FED-BATCH FERMENTATION FOR NISIN PRODUCTION IN A BIOFILM

REACTOR AND EFFECTS OF pH PROFILES *

ABSTRACT

A biofilm reactor not only shortens the lag phase of nisin production, but also enhances nisin production when combined with an appropriate pH profile. Due to the substrate inhibition that takes place at high levels of carbon source, fed-batch fermentation was proposed as a better alternative for nisin production. In this study, the combined effects of fed-batch fermentation and various pH profiles on nisin production in a biofilm reactor were evaluated. The tested pH profiles include i) a constant pH profile at 6.8 (profile Const12), ii) a constant pH profile with an autoacidification after 4 h (profile Const4), and iii) a profile with constant pH at 6.8 until 4 h follow by a step- wise pH adjustment every 2 h (profile Const4-Stepwise). When profile Const12 was applied, fed-batch fermentation enhanced nisin production of suspended-cell (4,188

IU/ml) and biofilm (4,314 IU/ml) reactors, yielded 1.8 and 2.3-fold higher nisin titer than their respective batch fermentation. On the other hand, pH profiles that include periods of autoacidification (profiles Const4 and Const4-Stepwise) resulted in a significantly lower nisin production in fed-batch fermentation (2,494 and 1,861 IU/ml for biofilm reactor using profile Const4 and Const4-Stepwise, respectively) due to toxicity of excess lactic acid produced during the fermentation. Overall, this study suggested that fed-batch

* This chapter was accepted to be published in Journal of Applied Microbiology and Biotechnology, 2006.

114 fermentation can be successfully used to enhance nisin production for both suspended- cell and biofilm reactors.

INTRODUCTION

Nisin is a 34-amino acid, antimicrobial peptide produced by Lactococcus lactis during its fermentation process . It is widely used in the food industry as a natural preservative due to its broad antibacterial activity against Gram-positive spoilage and pathogenic bacteria. Many aspects of nisin production have been well studied and reported, ranging from producer strains (De Vuyst, 1994; Kim et al., 1997), media formula (Cabo et al., 2001a; Li et al., 2002), and cultural conditions such as pH, aeration, and temperature (Shimizu et al., 1999; Desjardins et al., 2001; Cabo et al., 2001b).

It is known that nisin production is affected and limited by several factors, for example, substrate inhibition, adsorption of nisin onto the producer cells, and enzymatic degradation of nisin (De Vuyst, 1992; Yang et al., 1992). Previous studies clearly indicated a substrate inhibition in nisin production at high levels of sucrose (De Vuyst and Vandamme, 1992; Lv et al., 2004a). Based on these results, fed-batch fermentation has been proposed as a better alternative for production of nisin, because the substrate can be added to the reactor intermittently or continuously as needed. This type of approach has been used successfully for many types of fermentation processes, such as lactic acid production (Liu et al., 2005), biomass production (Demirci et al., 1999), antibiotic (Colombie et al., 2005) and recombinant proteins (Choi and Park, 2006).

Although information on nisin production in fed-batch cultures is still scarce, a few

115 studies have revealed advantages of fed-batch fermentation on nisin production in suspended-cell reactors (Lv et al., 2004a; 2004b; 2004c; Lv et al. 2005).

In this study, fed-batch fermentation was evaluated using a biofilm reactor constructed with a custom-made solid support called Plastic Composite Support (PCS).

PCS, an extruded mixture of polypropylene and several agricultural products such as soybean hulls and flour, not only provides an ideal physical structure for biofilm development, but also slowly releases nutrients to sustain the growth of biofilms (Ho et al. 1997a; 1997b; 1997c). A PCS biofilm reactor was shown to enhance production of several valued-added products such as lactic acid, succinic acid, and ethanol (Demirci et al., 1995; Cotton et al., 2001; Valazquez et al., 2001; Urbance et al., 2003). As for nisin, our previous studies showed that PCS biofilm reactors can shorten the lag phase of nisin production by 4 h (Chapter 4).

Nisin fermentation is commonly carried out using a constant pH profile at the optimal pH for nisin production, which varies from pH 5.5 to 6.8 depending on the producer strain (De Vuyst and Vandamme, 1992; Meghrous et al., 1992; Matsusaki et al.,

1996). More recent studies indicate a step-wise pH profile (repeated re-alkalizations after periodic autoacidications) enhances nisin production and nutrient consumption of L. lactis in a suspended-cell reactor (Cabo et al., 2001; Guerra and Pastrana, 2003a; 2003b).

However, nisin production of strain NIZO 22186 in a biofilm reactor using batch fermentation with the step-wise pH profile was very low (593 IU/ml), possibly as a result of pH stress exerted on the producer cells (Chapter 5).

According to our previous study, nisin production in biofilm reactors was highly affected by pH profiles applied during fermentation (Chapter 5). When a proper pH

116 profile is employed, the biofilm reactor significantly enhances nisin production in batch fermentation. The profile in which pH was allowed to drop freely via autoacidification after 4 h of fermentation yielded approximately 1.9-time higher nisin (3,553 IU/ml) than with a constant pH at 6.8 (1,898 IU/ml), possibly as a result of less adsorption of nisin onto producer cells. Therefore, autoacidification after 4 h of fermentation was recommended as an appropriate pH profile for batch fermentation of nisin production in a biofilm reactor.

The previous results also indicated a limitation of sucrose, which was completely depleted after 4 and 8 h in batch fermentations of biofilm and suspended-cell reactors, respectively (Chapter 5). In attempt to avoid the effects of substrate limitation and inhibition, fed-batch fermentation in which the sucrose is maintained at low concentration was employed in this study. Since the proper pH profiles may be different for batch and fed-batch fermentations, combining effects of fed-batch fermentation and various pH profiles on nisin production in the biofilm reactor were investigated as well.

MATERIALS AND METHODS

Microorganisms and media

The nisin Z-producing strain used in this study was Lactococcus lactis subsp. lactis (NIZO 22186). Nisin activity was quantified using Lactobacillus sakei (ATCC

15521) as the nisin-sensitive test organism. L. lactis and L. sakei were grown at 30 °C for

14 h in a complex medium (CM) and Lactobacillus MRS broth (Difco Laboratories,

Detroit, MI), respectively. The CM consisted of 40 g of sucrose, 10 g of peptone (Amber

117 Ferm 4015G, Universal Flavors, Milwaukee, WI), 10 g of yeast extract (Ardamine Z,

Sensient Bionutrient, Indianapolis, IN), 10 g of KH 2PO 4, 2 g of NaCl, and 0.2 g of

MgSO4·7H2O per liter of deionized (DI) water. The initial pH of CM was adjusted to 6.8 using 4 N NaOH. For a long-term storage, all stock cultures were maintained at -80 °C in

20% glycerol.

Experimental design

Three different pH-control profiles were evaluated in a biofilm reactor with PCS support using fed-batch fermentation (Fig. 6.1). These profiles were: i) pH controlled at

6.8 during the entire fermentation period (profile Const12), ii) pH controlled at 6.8 from

0 to 4 h and then allowed to drop freely via autoacidification from 4 to 12 h before being maintained at constant pH of 6.8 until 24 h (in order to avoid the sloughing off of the biofilm in acidic condition) (profile Const4), and iii) pH controlled using a step-wise profile, in which the pH was maintained at 6.8 until 4 h and then allowed to drop freely and adjusted back to 6.8 every 2 h until 12 h of fermentation, then held constant at 6.8 until 24 h (profile Const4-Stepwise). Before the data was collected in each treatment, the biofilm reactor was run using the desired treatment for a few runs (acclimation period) in order to allow the biofilm reactor to adjust to the new cultural conditions.

118 7.0

6.5

6.0 pH 5.5

5.0

4.5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

7.0

6.5

6.0 pH 5.5

5.0

4.5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

7.0

6.5

6.0 pH 5.5

5.0

4.5 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Figure 6.1. Actual pH profiles of nisin fermentation by L. lactis when using (a) profile Const12; (b) profile Const4; and (c) profile Const4-Stepwise.

119 For comparison, batch and fed-batch fermentations in the suspended-cell reactor were carried out using the pH profile that performed best for fed-batch fermentation in the biofilm reactor. Previously reported results of batch fermentations in biofilm reactors using various pH profiles (Chapter 5) were occasionally mentioned for the benefit of comparison.

Plastic Composite Support (PCS)

PCS tubes were manufactured at Iowa State University using a twin-screw co- rotating Brabender PL2000 extruder (model CTSE-V; C.W. Brabender Instruments, Inc.,

South Hackensack, NJ) as described by Ho et al. (1997). The composition of PCS used in this study was chosen according to biofilm formation, nisin production, nitrogen leaching rate, and % nitrogen content (Chapter 4). Polypropylene (50%) and other ingredients

(35% Soybean hulls, 5% Soybean flour, 5% Yeast extract, 5% Dried bovine albumin, and

0.272% of sodium acetate, 0.0004% of MgCl 2.6H 2O, and 0.002% of NaCl) were mixed together before being extruded at 13 rpm through a medium pipe die with barrel temperatures of 200, 220, and 200 °C and a die temperature of 165 °C. The extruded tubes with a wall thickness of 2.5 mm and an outer diameter of 10.5 mm were cut into 7.5-cm length tubes with both ends cut at a 45 degree angle to allow better flow of medium inside the tubes.

Nisin fermentation in biofilm reactor

Nisin fermentations were conducted in a 1.25-l Bioflo II fermentor (New

Brunswick Scientific, Edison, NJ) at 30 °C with agitation at 100 rpm. For the biofilm

120 reactor trials, twelve PCS tubes were bounded to the agitator shaft in a grid-like fashion, six rows of two parallel tubes (Chapter 4). The reactor vessel with PCS was sterilized with water at 121 °C, while sucrose and nitrogenous components with mineral salts were sterilized separately and added to the reactor aseptically. After inoculation with a 14-h culture of L. lactis (1% v/v), at least five fermentation runs were performed in order to establish biofilm formation on PCS supports. The pH was controlled at 6.8 by adding 4 N

NaOH. Once a stable biofilm was formed, repeated-batch fermentations were carried out in the biofilm reactor by pumping out the previous fermentation broth before aseptically adding a sterile fresh medium. Suspended-cell fermentation trials were conducted without

PCS by inoculating with the 14-h culture of L. lactis (1% v/v) at the beginning of each fermentation. For the biofilm reactor, fed-batch fermentation was performed by the continuous feeding of 50% sterile sucrose solution at the rate of 10 g-sucrose/h from 0-2 h and 5 g-sucrose/h from 2-12 h. For the suspended-cell reactor, the sucrose was fed at rate of 10 g-sucrose/h from 2-4 h and 5 g-sucrose/h from 4-12 h due to its longer lag phase.

Analysis

Samples were collected every 1 or 2 h and analyzed for cell density, sugar consumption, nisin production, and lactic acid production.

Suspended biomass. Suspended-cell density was estimated by optical density at

600 nm using a spectrophotometry DU series 500 (Beckman, Fullerton, CA). Sterile fermentation medium was used as a blank. Optical density values were converted into

121 biomass concentration by using a standard curve (y = 0.1663x – 0.0199; y: biomass (g/l) and x: optical density at 600 nm).

Nisin. Sample of fermentation broth was immediately adjusted to pH 3.0 using concentrated HCl and 0.1% v/v of Tween 20 (J.T. Baker, Phillipsburg, NJ) was added in order to avoid any non-specific adsorption of nisin on the container’s surfaces. Then, the sample was heated at 90 °C for 5 min to eliminate the protease activity in the broth. After centrifugation at 2,000 x g, 4 °C for 10 min (National Labnet Co., mini centrifuge Model

C-1200), clear supernatant was collected and kept frozen at -20 °C until analysis. Nisin quantifications were performed by an agar well diffusion bioassay, using L. sakei and pre-diffusion at 4 °C, as described in Chapter 3.

Sucrose. The fermentation broth was centrifuged at 2,000 x g for 10 min

(National Labnet Co., mini centrifuge Model C-1200) and clear supernatant was collected for analysis. For initial acid hydrolysis, 100 µl of concentrated HCl was added to 5 ml of the sample. After allowing the hydrolysis to proceed at 90 °C for 5 min, 0.25 ml of 5N

KOH solution was added to neutralize the acid. The hydrolyzed samples were then analyzed using the DNS Method as described by Miller (1959).

Glucose and lactic acid. Glucose and lactic acid concentration were determined using a high performance liquid chromatograph (HPLC) equipped with a refractive index detector (Waters, Franklin, MA). Components were separated on a Bio-Rad Aminex

HPX-87H column (300 x 7.8 mm) (Bio-Rad, Richmond, CA) with 0.012 N sulfuric acid as a mobile phase at a flow rate of 0.8 ml/min with an injection volume of 20 µl and a column temperature of 65 °C. Before injection, the samples were centrifuged (2,000 x g,

122 4°C) for 10 min and filtered through 0.22 micron filters (13 mm diameter disc filters,

Millipore, Bedford, MA).

Statistical Analysis

All treatments were replicated three times. The significant difference of the results was evaluated using the Generalized Linear Model (GLM) (with p < 0.05) and Tukey’s

HSD multiple comparison module of the MINITAB Statistical Software package

(Release 13.30) (State College, PA).

RESULTS

In this study, fed-batch fermentation was evaluated along with various pH profiles in the biofilm system. Figure 6.2 shows the amount of suspended biomass (g/l) in the fermentation broth of both suspended-cell (SC) and biofilm (BF) reactors with profile

Const12 (constant pH at 6.8). The results clearly indicated a shorter lag time of growth in the biofilm reactor, however it should be noted that the observed biomass only represented growth in the broth and not growth in the biofilm. In the biofilm reactor, there was no significant difference in term of biomass observed between batch and fed- batch cultures or between fermentations with different pH profiles (data not shown).

When compared to the suspended-cell system (1.35 g/l), maximum suspended biomass of the biofilm system (0.65 g/l) was consistently lower, possibly because a greater part of the biomass had accumulated in the biofilm. A similar trend was also observed in batch fermentations (Chapter 4).

123

1.6

1.4

1.2 1.0

0.8

0.6 (g/l) Biomass 0.4

0.2

0.0 0 2 4 6 8 10 12 14 16 18 20 22 24

Time (h)

BF-Fedbatch-Profile Const12 SC-Fedbatch-Profile Const12

Figure 6.2. Suspended biomass of L. lactis in different culture types (BF: Biofilm; SC: Suspended-cell).

60

50

40

30

20

Residual sucrose (g/l) 10

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

BF-Fedbatch-Profile Const12 SC-Batch-Profile Const12 SC-Fedbatch-Profile Const12

Figure 6.3. Sucrose consumptions of L. lactis when using different culture types (BF: Biofilm; SC: Suspended-cell).

124 In batch fermentation, sucrose was depleted after only 4 h in the biofilm system

(Chapter 5) and after 8 h in the suspended-cell system (Fig. 6.3). It should be noted that

the nisin and lactic acid concentrations also reached their maximum levels synchronously

at the same time as the depletion of sucrose (Fig. 6.4). This evidence of substrate

limitation as well as the substrate inhibition present at high sucrose level indicated the

appropriateness of fed-batch culture for nisin production. Continuous feeding of sucrose

in fed-batch culture prevented the depletion of sucrose as well as eliminated the substrate

inhibition by maintaining sucrose at a low level (Fig. 6.3). Due to the longer lag phase in

suspended-cell fermentation, the feeding of sucrose was prolonged for 2 h in order to

achieve a comparable feeding rate to the biofilm fermentation. However, the lower rate of

sucrose consumption in the suspended-cell fermentation resulted in higher level of

sucrose residual compared to the level in the biofilm fermentation (Fig. 6.3).

When profile Const12 was used, fed-batch fermentation highly enhanced nisin production in both suspended-cell (Fig. 6.4) and biofilm (Fig. 6.5) reactors. For suspended-cell reactor, nisin titer reached its maximum at 2,394 IU/ml in batch fermentation and 4,188 IU/ml in fed-batch fermentation. For the biofilm reactor, nisin titer reached its maximum at 1,898 IU/ml in batch fermentation (Chapter 5) and 4,314

IU/ml in fed-batch fermentation (Fig. 6.5). When profile Const12 was used, both biomass

(Fig. 6.2) and lactic acid concentration (Fig. 6.4) were comparable between batch and fed-batch fermentations up until the time that nisin reached its maximum. This is true for both biofilm and suspended-cell fermentations. Therefore, the enhanced nisin production was not a result of higher growth or shifting of metabolism away from lactic acid

125 5000 70.0

60.0 4000 50.0

3000 40.0

2000 30.0 Nisin (IU/ml)

20.0 Lactic acid (g/l) 1000 10.0

0 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

Nisin-SC-Batch-Profile Const12 Nisin-SC-Fedbatch-Profile Const12

Lactic acid-SC-Batch-Profile Const12 Lactic acid-SC-Fedbatch-Profile Const12 Figure 6.4. Nisin and lactic acid production of L. lactis when using different culture types and pH profiles. (BF: Biofilm; SC: Suspended-cell).

production, but rather an effect of the fed-batch strategy with its continuous supplemention of sucrose.

Profile Const4, which allows the pH to drop freely via autoacidification after 4 h, was designed to help prevent adsorption of nisin onto the producer cells, as the adsorption was less pronounced in an acidic environment (Yang et al., 1992). According to our earlier study (Chapter 5), profile Const4 significantly enhanced production of nisin during batch fermentation in the biofilm reactor. However, for fed-batch fermentation in the biofilm reactor, profile const4 (2,494 ± 417 IU/ml) resulted in a significantly lower nisin production than the level achieved using profile Const12 (4,314 ± 485 IU/ml) (Fig.

6.5). The suppression is suspected to be a result of the higher toxicity of lactic acid in an

126 acidic environment. Significantly lower lactic acid production was observed when profile

Const4 was used (Fig. 6.6).

5000

4000

3000

2000 Nisin (IU/ml)

1000

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

BF-Fedbatch-Profile Const12 BF-Fedbatch-Profile Const4 BF-Fedbatch-Profile Const4-Stepwise

Figure 6.5. Nisin production of L. lactis in the biofilm reactor using fed-batch fermentation with various pH profiles (BF: Biofilm).

Profile Const4-Stepwise is modified from a step-wise pH profile, which was reported to enhance nisin production in other strains of nisin producer (Cabo et al.,

2001b; Guerra and Pastrana, 2003). Since the drop of pH in the early period of fermentation seemed to detrimentally affect growth and nisin production (Chapter 5), profile Const4-Stepwise was designed so that the pH was maintained at near neutrality

(pH = 6.8) up to 4 h before freely dropping via autoacidification. In this way, profile

127 Const4-Stepwise is parallel to profile Const4 and is able to reveal the effect of step-wise pH on nisin production. The results in Figure 6.6 indicated that a step-wise pH profile highly promoted lactic acid production (44 ± 6 g/ml at 8 h), correlating with a much lower nisin production (1,861 ± 140 IU/ml) observed in Figure 6.5. This suggests a combined effect of high lactic acid concentration and low pH introduced during the autoacidification period.

70

60

50

40

30 Lactic acid (g/l) acid Lactic 20

10

0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (h)

BF-Fedbatch-Profile Const12 BF-Fedbatch-Profile Const4 BF-Fedbatch-Profile Const4-Stepwise

Figure 6.6. Lactic acid production of L. lactis in the biofilm reactor using fed-batch fermentation with various pH profiles (BF: Biofilm).

128 DISCUSSIONS

The results in this study confirmed the positive influence of the fed-batch culture on nisin production in both suspended-cell and biofilm reactors. Continuous feeding of sucrose in the fed-batch fermentation prolonged the production period of not only nisin, but also of lactic acid. In fed-batch fermentation, nisin yield was tremendously improved in both suspended-cell (1.8-fold, 1.58 mg-nisin per g-sucrose consumed, max. nisin at 12 h) and biofilm (2.3-fold, 1.3 mg-nisin per g-sucrose consumed, max. nisin at 8 h) reactors when pH was maintained at near neutrality (profile Const12). However, nisin yield indicated that batch fermentation of a biofilm reactor with pH controlled according to profile Const4 (Chapter 5) yielded the most efficient nisin production (1.92 mg-nisin per g-sucrose consumed), with a shorter fermentation period required (max. nisin at 6 h).

Further supplementation of sucrose after the nisin level reached its maximum did not improve nisin production further, but promoted lactic acid production instead.

Although profile Const4 and the profile Const4-Stepwise were reported to enhance nisin production of batch fermentation in both biofilm reactors (Chapter 5) and suspended-cell reactors (Cabo et al., 2001b), toxicity of excess accumulated lactic acid in the acidic environment created in both profiles made these profiles highly inappropriate for the fed-batch cultures. Higher acidity increases concentration of the non-dissociated form of lactic acid, which can diffuse through the cell membrane. Once inside the cell, the non-dissociated form is converted into the dissociated form due to the near-neutrality environment of the cytoplasm (Kashket, 1987), causing the acidification of the cytoplasm, which can lead to inhibition of growth and metabolism (Mercade et al., 2000).

129 Thus, even if the lactic acid concentration of the fed-batch culture with profile Const4 were slightly lower at the time nisin reached its maximum (39 ± 1 and 34 ± 3 g/ml for profiles Const12 and Const4, respectively), the lower pH (approx. pH = 5) in profile

Const4 severely changed the cellular environment, resulting in significantly lower nisin production (4,314 ± 485 and 2,494 ± 417 IU/ml for profile Const12 and Const4, respectively). This effect was even more pronounced in the fed-batch fermentation with profile Const4-Stepwise due to its higher accumulated lactic acid (Fig. 6.6). As a result, profile Const4-Stepwise yielded the lowest nisin yield observed in this study (0.49 mg- nisin per g-sucrose consumed).

In order to eliminate the toxicity of lactic acid, both separation of lactic acid by adsorbent (Yu et al., 2002) and consumption of lactic acid by Kluyveromyces marxianus in mixed-culture (Shimizu et al., 1999) have been shown to effectively enhance production of nisin. Since addition of an adsorbent unit may not be a cost-effective alternative, application of a flocculating strain of K. marxianus (e.g. ATCC 10022) may be more appropriate for incorporation into the biofilm system. To evaluate that second option, the kinetics of growth, nisin production, and lactate consumption of a mixed culture of L. lactis and K. marxianus need to be investigated in the actual biofilm reactor.

In conclusion, this study suggested that fed-batch fermentation can be successfully used to enhance nisin production for both suspended-cell and biofilm reactors, with a constant pH profile as the most suitable pH profile.

130 REFERENCES

Cabo, M. L., M. A. Murado, M. P. Gonzalez, J. A. Vazquez, and L. Pastoriza. 2001a. An empirical model for describing the effects of nitrogen sources on nisin production. Lett. Appl. Microbiol. 33: 425-429. Cabo, M. L., M. A. Murado, M. P. Gonzalez, and L. Pastoriza. 2001b. Effects of aeration and pH gradient on nisin production. A mathematical model. Enz. Microbial. Technol. 29: 264-273. Choi, D. B., and E. Y. Park. 2006. Enhanced production of mouse alpha-amylase by feeding combined nitrogen and carbon sources in fed-batch culture of recombinant Pichia pastoris . Proc. Biochem. 41(2): 390-397. Colombie, V., C. Bideaux, G. Goma, and J. L. Uribelarrea. 2005. Effects of glucose limitation on biomass and spiramycin production by Streptomyces ambofaciens . Bioproc. Biosys. Eng. 28(1): 55-61. Cotton, J. C., A. L. Pometto III, and J. Gvocdenovic-Jeremic. 2001. Continuous lactic acid fermentation using a plastic composite support biofilm reactor. Appl. Microbiol. Biotechnol. 57: 626-630. Demirci, A., A. L. Pometto III, K-L. G. Ho. 1995. Continuous ethanol production in biofilm reactors containing plastic composite rings and disks. Proc.-Biomass Conference of the Americas: Energy, Environment, Agricultural and Industry, 2 nd , Aug 21-24, Portland, Oregon, National Renewable Energy Laboratory, Golden, Colorado. Demirci, A., A. L. Pometto III, and D. J. Cox. 1999. Enhanced organically bound selenium yeast production by fed-batch fermentation. J. Ag. Food Chem. 47(6): 2496-2500. Desjardins, P., J. Meghrous, and C. Lacroix. 2001. Effect of aeration and dilution rate on nisin Z production during continuous fermentation with free and immobilized Lactococcus lactis UL719 in supplemented whey permeate. Int. Dairy J. 11: 943- 951

131 De Vuyst, L., and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. De Vuyst, L. 1994. Nisin production variability between natural Lactococcus lactis subsp. lactis strains. Biotech. Lett. 16(3): 287-292. Guerra, N. P., and L. Pastrana. 2003a. Influence of pH drop on both nisin and pediocin production by Lactococcus lactis and Pediococcus acidilactici. Lett. Appl. Microbiol. 37: 51-55. Guerra, N. P., and L. Pastrana. 2003b. Enhancement of nisin production by Lactococcus lactis in periodically re-alkalized cultures. Biotechnol. Appl. Biochem. 38: 157-167. Ho, K-L. G., A. L. Pometto III, and P. N. Hinz. 1997a. Ingredient selection for plastic composite supports for L-(+)-lactic acid biofilm fermentation by Lactobacillus casei subsp. Rhamnosus . Appl. Environ. Microbiol. 63: 2516-2523. Ho, K-L. G., A. L. Pometto III, P. N. Hinz, and A. Demirci. 1997b. Nutrient leaching and end product accumulation in plastic composite supports for L-(+)-lactic acid biofilm fermentation. Appl. Environ. Microbiol. 63: 2524-2532. Ho, K-L. G., A. L. Pometto III, and P. N. Hinz. 1997c. Optimization of L-(+)-lactic acid production by ring and disc plastic composite supports through repeated-batch biofilm fermentation. Appl. Environ. Microbiol . 63: 2533-2542. Kashket, E. R. 1987. Bioenergetics of lactic acid bacteria: cytoplasmic pH and osmotolerance. FEMS Microbiol. Rev. 46: 233-244. Kim, W. S., R. J. Hall, and N. W. Dunn. 1997. Host specificity of nisin production by Lactococcus lactis . Biotech. Lett. 19(12): 1235-1238. Li, C., J. Bai, Z. Cai, and F. Ouyang. 2002. Optimization of a cultural medium for bacteriocin production by Lactococcus lactis using response surface methodology. J. Biotechnol. 93: 27-34. Liu, T. J., S. Miura, T. Arimura, M. Y. Tei, E. Y. Park, and M. Okabe. 2005. Evaluation of L-lactic acid production in batch, fed-batch, and continuous cultures of Rhizopus sp MK-96 1196 using an airlift bioreactor. Biotechnol. Bioproc. Eng. 10(6): 522- 527.

132 Lv, W., W. Cong, and Z. Cai. 2004a. Effect of sucrose on nisin production in batch and fed-batch culture by Lactococcus lactis . J. Chem. Technol. Biotechnol . 80: 511-514. Lv, W., W. Cong, and Z. Cai. 2004b. Nisin production by Lactococcus lactis subsp. lactis under nutritional limitation in fed-batch culture. Biotechnol. Lett. 26: 235-238. Lv, W., W. Cong, and Z. Cai. 2004c. Improvement of nisin production in pH feed-back controlled, fed-batch culture by Lactococcus lactis subsp. lactis . Biotechnol. Lett. 26: 1713-1716. Lv, W., Z. Zhang, and W. Cong. 2005. Modelling the production of nisin by Lactococcus lactis in fed-batch culture. Appl. Microbiol. Biotechnol. 68: 322-326. Matsusaki, H., N. Endo, K. Sonomoto, and A. Ishizaki. 1996. Lantibiotic nisin Z fermentative production by Lactococcus lactis IO-1: relationship between production of the lantibiotic and lactate and cell growth. Appl. Microbiol. Biotechnol. 45: 36-40. Meghrous, J., E. Huot, M. Quittelier, and H. Petitdemange. 1992. Regulation of nisin biosynthesis by continuous cultures and by resting cells of Lactococcus lactis supsp. lactis. Res. Microbiol. 143: 897-890. Mercade. M., N. D. Lindley, and P. Loubiere. 2000. Metabolism of Lactococcus lactis subsp. cremoris MG 1363 in acid stress conditions. Int. J. Food Microbiol. 55: 161- 165. Miller, G. L. 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31: 426. Shimizu, H., T. Mizuguchi, E. Tanaka, and S. Shioya. 1999. Nisin production by a mixed-culture system consisting of Lactococcus lactis and Kluyveromyces marxianus . Appl. Environ. Microbiol. 65(7): 3134-3141. Urbance, S. E., A. L. Pometto III, A. A. DiSpirito, and A. Demirci. 2003. Medium evaluation and plastic composite support ingredient selection for biofilm formation and succinic acid production by Actinobacillus succinogenes . Food Biotechnol. 17: 53-65. Velazquez, A. C., A. L. Pometto III, K-L. G. Ho, and A. Demirci. 2001. Evaluation of plastic-composite supports in repeated fed-batch biofilm lactic acid fermentation by Lactobacillus casei. Appl. Microbiol. Biotechnol. 55: 434-441.

133 Yang, R. , M. C. Johnson, and B. Ray. 1992. Novel method to extract large amount of bacteriocins from lactic acid bacterial. Appl. Environ. Microbiol. 58: 3355-3359. Yu, P. L., N. W. Dunn, and W. S. Kim. 2002. Lactate removal by anionic-exchange resin improves nisin production by Lactococcus lactis . Biotechnol. Lett . 24: 59-64.

134 CHAPTER 7

RECOVERY OF NISIN FROM FERMENTATION BROTH

PRODUCED IN A BIOFILM REACTOR

ABSTRACT

Nisin is a 34-amino acid polypeptide with an antimicrobial activity against wide range of Gram-positive bacteria. While it is mainly used as a natural food preservative, other applications including medical purposes are also promising. It is known that nisin production is limited by many factors, including adsorption of nisin onto producer cells, enzymatic degradation by protease, and even product inhibition from the nisin itself.

Continuous removal of nisin from the broth during fermentation can be used to minimize these detrimental effects. In this study, silicic acid was successfully used to recover nisin from fermentation broth of L. lactis subsp. lactis NIZO 22186. The effect of pH (at 6.8 and 3.0) during adsorption process and several eluents (DI water, 20% ethanol, 1 M

NaCl, and 1 M NaCl + 20% ethanol) for desorption were evaluated. Higher nisin adsorption onto silicic acid can be achieved when the adsorption was carried out at pH

6.8 (67% adsorption) than at pH 3.0 (54% adsorption). The maximum recovery (47% of nisin was harvested) was achieved when the adsorption was carried out at pH 6.8 and 1

M NaCl + 20% ethanol was used as an eluent for desorption.

135 INTRODUCTION

Nisin is a 34-amino acid antimicrobial polypeptide produced during a fermentation of Lactococcus lactis subsp. lactis. Due to its antimicrobial activity against a wide range of Gram-positive bacteria, including several major food-borne pathogens like Clostridium and Listeria , nisin has been used extensively in the food industry as a natural food preservative. It is, by far, the only bacteriocin that is approved for food applications (FDA, 1988) and produced commercially.

In our previous studies, nisin production was improved significantly when using a biofilm reactor in combination with an optimized complex medium (Chapter 4), pH profiles (Chapter 5), and fed-batch culture (Chapter 6). Nisin production is known to be affected by several cultural factors, such as producer strain, nutrients composition of media, pH, temperature, agitation and aeration, as well as the unique characteristic of nisin production like substrate and product inhibition, adsorption of nisin onto the producer cells, and enzymatic degradation (Parente and Ricciardi, 1999). Although nisin production was autoregulated with nisin itself acting as an inducer molecule or peptide pheromone (Kuipers et al., 1995), a few studies have suggested the presence of product inhibition caused by nisin and showed that removal of nisin from culture broth during fermentation significantly enhanced nisin production (Kim, 1997; Chinachoti et al ., 1997;

Tolonen et al., 2004).

A suitable silica adsorbent for on-line recovery not only has to exhibit excellent nisin adsorption and desorption properties, but also has to be compatible to the growth of the producer (being non-toxic and does not remove necessary nutrients from the medium)

136 once it was incorporated into the bioreactor system. Several silica adsorbents, usually a cationic adsorbent that facilitates an electrostatic attraction between the adsorbent and nisin molecule, have been tested either in batch experiment or incorperated into the bioreactor system (Chinachoti et al., 1997; Wan et al., 1997, Tolonen et al., 2004).

Chinachoti et al . (1997) separated nisin Z from fermentation broth using various types of adsorbents including Amberlite IR-120B, CM Sephadex C-25, Celite, and Sep-

Pak tC 18 , C 18 , C 8 and tC 3 cartridges. Sep-Pak C 8 , a moderate reversed-phase column, was selected, because it exhibited a moderate adsorption-desorption ability for nisin Z, even though it yielded a lower nisin Z adsorption than tC 18 column. When the Sep-Pak C 8 was incorporated into the batch fermentation, higher cell growth and a 1.7-time greater nisin Z production rate were obtained. Interestingly, the results obtained by Kim (1997) also indicated higher biomass production of the producer when nisin was continuously removed from the system. These results agreed with a study from Hurst and Kruse

(1972), which indicated toxicity that nisin exert on the producer itself. However, the two- phase system used in the study of Kim (1997) also removed lactic acid from the system, thus higher biomass production obtained can also be a result of less toxicity from accumulated lactic acid.

Tolonen et al. (2004) reported approximately 3-time higher nisin Z production when a column of Amberlite XAD-4 was integrated into the reactor system. The results showed that the column could remove approximately 80% of nisin from the fermentation broth, but only 57% of nisin could be harvested. Both Micro-Cel E (a synthetic calcium silicate from diatomaceous earth) and Hi-Sil HOA (a synthetic ultrafine silicon dioxide) were successfully used for adsorption of nisin (Wan et al., 1996). However the nisin

137 adsorped on Micro-Cel E can be eluted out only by 1% sodium dodecyl sulfate (SDS) and thus an effective SDS removal technique was required (Coventry et al., 1996).

Silicic acid was successfully used to recover 84% of the nisin from fermentation broth. Although only a 3-fold purification was achieved (Janes et al., 1998), the method was preferable, as it did not require the use of toxic solvent. In this study, silicic acid was used to remove nisin Z from the fermentation broth obtained from the biofilm reactor and optimal conditions for adsorption and desorption of nisin were evaluated in a batch recovery approach. The adsorption was tested at two different pH values (3.0 and 6.8) while four types of diluents were tested for desorption of nisin Z. The obtained result can be used to construct an on-line recovery unit that will be integrated into the existing biofilm reactor.

MATERIALS AND METHODS

Microorganisms and media

The nisin Z-producing strain used in this study was Lactococcus lactis subsp. lactis (NIZO 22186). Nisin activity was quantified using Lactobacillus sakei (ATCC

15521) as the nisin-sensitive test organism. L. lactis and L. sakei were grown at 30 °C for

14 h in a complex medium (CM) and Lactobacillus MRS broth (Difco Laboratories,

Detroit, MI), respectively. The CM consisted of 40 g of sucrose, 10 g of peptone (Amber

Ferm 4015G, Universal Flavors, Milwaukee, WI), 10 g of yeast extract (Ardamine Z,

Sensient Bionutrient, Indianapolis, IN), 10 g of KH 2PO 4, 2 g of NaCl, and 0.2 g of

MgSO4·7H2O per liter of deionized (DI) water. The initial pH of CM was adjusted to 6.8

138 using 4 N NaOH. For a long-term storage, all stock cultures were maintained at -80 °C in

20% glycerol.

Adsorption and desorption of nisin

Nisin fermentations were conducted in a 1.25-l Bioflo II fermentor (New

Brunswick Scientific, Edison, NJ) at 30 °C with agitation at 100 rpm using a constant pH profile at 6.8. The biofilm reactor was set up as described previously (Chapter 4).

Fermentation broth, which was removed from the biofilm reactor after 6 h of fermentation, was filtered using a tangential flow filter unit with pore size of 0.45 µm and filtration area of 50 cm 2 (Pellicon ® XL Durapore HVMP, Millipore, Bedford, MA) and immediately adjusted to pH 3.0 and 6.8 using concentrated HCl or 4 N NaOH. In order to create the condition that represent the environment in an on-line adsorption, the filtered fermentation broth was not heated to destroy the activity of protease.

After pH adjustment, 5% (w/v) silicic acid (mesh size 100, Sigma, St. Louis, MO) was added into 10 ml of pH-adjusted, filtered broth. The samples were stirred at 25 °C for

1 h. Then, the silicic acid was separated and washed with sterile DI water (pH 3.0) before resuspending in 10 ml of eluent. Four different eluents tested were; 1) DI water, 2) DI water with 20% ethanol, 3) 1 M NaCl, and 4) 1 M NaCl with 20% ethanol. All eluents were adjusted to pH 3.0 with concentrated HCl. The samples were then stirred for 30 min at 25 °C and heated for 10 min at 90 °C. The nisin activity of all fractions collected at different stages of adsorption and desorption was assayed using an agar well diffusion method as described previously (Chapter 4). In order to take into account the loss of nisin due to the adsorption onto the container’s surfaces and even degradation during the

139 adsorption process, controls (without silicic acid) have been carried out along with the samples.

Percent adsorption and percent desorption were calculated as follow:

% Adsorption = (N c – N a) x 100/N c

% Desorption = N d x 100/(N c - N a)

% Harvest = N d x 100/N c where % Adsorption = percent of nisin that adsorbed onto the silicic acid

% Desorption = percent of nisin that can be eluted from the silicic acid

based on the amount of adsorbed nisin

% Harvest = percent of nisin that can be harvested based on the

amount of nisin in the control

Nc = nisin activity (IU/ml) in control

Na = nisin activity (IU/ml) in sample after the adsorption

Nd = nisin activity (IU/ml) released into the diluent after the desorption

Statistical Analysis

All treatments were replicated three times. The significant difference of the results was evaluated using the Generalized Linear Model (GLM) (with p < 0.05) and Tukey’s

HSD multiple comparison module of the MINITAB Statistical Software package

(Release 13.30) (State College, PA).

140 RESULTS AND DISCUSSIONS

The adsorption of nisin onto silicic acid, unlike the adsorption of nisin onto producer cells which relies on the cationic nature of nisin, was proposed as a result of a multilayer of protein-protein weak and intermediate hydrogen bonds influenced by electrostatic interactions (Janes et al., 1998). Adsorption of nisin Z onto silicic acid was significantly higher in neutral environment (Fig. 7.1 and 7.2), indicated by higher % adsorption at pH 6.8 (68.8%) as compared to pH 3.0 (54.0%). The same trend was also reported for adsorption of nisin A onto silicic acid (Janes et al., 1998). Neutral environment, in which nisin possesses less positive net charge, favors the adsorption process because interactions between polypeptide are generally favored under conditions that reduce the net charge on the molecules.

Heat treatment during the desorption processes was reported as an effective measure to brake the hydrogen bonds, and thus to enhance the desorption of nisin from silicic acid (Janes et al., 1998). Therefore, all eluents were adjusted to pH 3.0, which is the pH where nisin is most stable (Davies et al., 1998). Although eluent 1 (DI water) was successfully used to elute nisin A that adsorbed on silicic acid (91% desorption) (Janes et al., 1998), it failed to elute adsorbed nisin Z in this study because only 0.6% of adsorbed nisin was desorbed when the adsorption was conducted at pH 6.8 (Fig. 7.1). This may be the result of lower net positive charge of nisin Z molecule, +2 instead of +3 charges due to the substitution of histidine by asparagine.

141 80 68.8 65.6 70 60 46.5 50 40

Percent 30

20 13.5 9.8 10 0.6 0.4 0.6 0.4 0 % % % % % Harvest % Harvest % Harvest % Harvest Desorption Desorption Desorption Desorption % Eluent 1 Eluent 2 Eluent 3 Eluent 4 Adsorption

Figure 7.1. Adsorption and desorption of nisin from silicic acid using Eluent 1 (DI water), Eluent 2 (20% ethanol), Eluent 3 (1 M NaCl), and Eluent 4 (1 M NaCl + 20% ethanol) when the adsorption was carried out at pH 6.8.

80 70 54 52.8 60 48.1 50 41.9 40 29.9 27.4 Percent 30 19.6 16.3 20 9.1 10 0 % % % % % Harvest % Harvest % Harvest % Harvest Desorption Desorption Desorption Desorption % Eluent 1 Eluent 2 Eluent 3 Eluent 4 Adsorption

Figure 7.2. Adsorption and desorption of nisin from silicic acid using Eluent 1 (DI water), Eluent 2 (20% ethanol), Eluent 3 (1 M NaCl), and Eluent 4 (1 M NaCl + 20% ethanol) when the adsorption was carried out at pH 3.0. .

142 When the adsorption was carried out at pH 6.8, addition of NaCl and ethanol separately, as in eluent 2 (0.6% desorption) and 3 (13.5% desorption), did not improve the desorption of nisin as much as when both of them combined together in eluent 4

(65.6% desorption). On the other hand, although the adsorption of nisin onto silicic acid was lower when the adsorption process was conducted at pH 3.0, the adsorbed nisin seemed to be easier to elute (Fig. 7.2) as indicated by significantly higher % desorption values when eluent 1 (41.9% desorption) and eluent 2 (52.8% desorption) were used.

NaCl and ethanol were used in this study because of their non-toxic nature and ability to weaken or destroy the hydrogen bonds (Chandra, 2000). Overall, eluent 4 (1 M NaCl +

20% ethanol) with adsorption at pH 6.8 yielded the highest desorption (65.6% desorption) and overall recovery (46.5% harvest) among all the tested treatments.

CONCLUSION

Since an adsorption at pH 6.8 is highly practical and considerably easy to incorporate into the bioreactor system, which is also carried out at pH 6.8, the adsorption at 6.8 and desorption using eluent 4 was recommended for further application. Based on the total nisin activity in fermentation broth, 31% of nisin did not bind to the silicic acid,

47% of nisin was harvested, and 22% was unelutable from the silicic acid (this portion included both potentially degraded and irreversibly bounded nisin). For an on-line recovery, the existing procedure may need to be refined so that the adsorption and desorption can be processed continuously.

143 REFERENCES

Chandra, A. 2000. Effects of ion atmosphere on hydrogen-bond dynamics in aqueous electrolyte solutions. Phys. Rev. Lett. 85(4): 768-771. Chinachoti, N., N. Endo, K. Sonomoto, and A. Ishizaki. 1997. Bioreactor systems for efficient production and separation of nisin Z using Lactococcus lactis IO-1. J. Fac. Agric. Kyushu Univ. 43: 421-436. Coventry, M.J., J.B. Gordon, M. Alexander, M.W. Hickey, and J. Wan. 1996. A food- grade process for isolation and partial purification of bacteriocins of lactic acid bacteria that used diatomite calcium silicate. Appl. Environ. Microbiol. 62(5): 1764- 1769. Davies, E. A., H. E. Bevis, R. Potter, J. Harris, G. C. Williams, and J. Delves-Broughton. 1998. The effect of pH on the stability of nisin solution during autoclaving. Lett. Appl. Microbiol. 27: 186-187. Hurst, A., and H. Kruse. 1972. Effect of secondary metabolites on the organisms producing them: Effect of nisin on Streptococcus lactis and enterotoxin B on Staphylococcus aureus. Antimicrobial Agents and Chemotherapy 1: 277-279. Janes, M. E., R. Nannapaneni, A. Proctor, and M. G. Johnson. 1998. Rice hull ash and silicic acid as adsorbents for concentration of bacteriocins. Appl. Environ. Microbiol. 64(11): 4403-4409. Kim, W. S. 1997. Nisin production by Lactococcus lactis using two-phase batch culture. Lett. Appl. Microbiol. 25: 169-171. Kuipers, O. P., M. M. Beerthuyzen, P. G. de Ruyter, E. J. Luesink, and W. M. de Vos. 1995. Autoregulation of nisin biosynthesis in Lactococcus lactis by signal- transduction. J. Biol. Chem. 270: 27299-27304. Parente, E., and A. Ricciardi. 1999. Production, recovery and purification of bacteriocins from lactic acid bacteria. Appl. Microbiol. Biotechnol. 52: 628-638. Tolonen, M., P. E. J. Saris, and M. Siika-aho. 2004. Production of nisin with continuous adsorption to Amberlite XAD-4 resin using Lactococcus lactis N8 and L. lactis LAC48. Appl. Microbiol. Biotechnol. 63: 659-665.

144 U. S. Food and Drug Administration. 1988. Nisin preparation: Affirmation of GRAS status as direct human food ingredient. Federal Register, 53, April 6. Wan, J., M. W. Hickey, and M. J. Coventry. 1995. Continuous production of bacteriocins, brevicin, nisin and pediocin, using calcium alginate-immobilized bacteria. J. Appl. Bacteriol. 79: 671-676.

145 CHAPTER 8

Modeling of Nisin Production by Lactococcus lactis

During Batch Fermentation

ABSTRACT

Nisin, the only FDA approved bacteriocin, is widely used as a natural food preservative against Gram-positive bacteria. It is produced by lactic acid bacteria

Lactococcus lactis subsp. lactis during fermentation . Mathematical models of biomass and nisin production during fermentation not only provide information about the kinetic- metabolic nature of nisin, but also facilitate the control and optimization of nisin production. In this study, several models were developed to describe biomass, nisin, and lactic acid production as well as sucrose consumption during nisin fermentation by L. lactis in suspended-cell batch culture. All models were designed so that they contained biologically relevant parameters. The re-modified logistic and Gompertz models proposed in this study adequately described the growth and the validation showed that they could be used for the prediction accurately (slope = 1.01, R-square = 0.99). As for nisin production, the Luedeking-Piret model fit considerably well and could be successfully use to predict nisin production from the value of biomass (slope = 1.04, R- square = 0.98).

146 INTRODUCTION

Nisin, one of the most well studied bacteriocins, is a 34 amino acid antimicrobial peptide that is produced by Lactococcus lactis subsp. lactis during fermentation . It is widely used in the food industry as a natural preservative due to its broad antibacterial activity against Gram-positive spoilage and pathogenic bacteria. Many aspects of nisin production have been well studied and reported, ranging from producer strains (De Vuyst

1994; Kim et al. 1997), media formulation (Cabo et al. 2001a; Li et al. 2002), and cultural conditions such as pH, aeration, and temperature (Shimizu et al. 1999; Desjardins et al. 2001; Cabo et al. 2001b).

Mathematical models of biomass and nisin production not only provide information about the kinetic-metabolic nature of nisin, but also facilitate the control and optimization of nisin production during fermentation. Predictions of nisin production based on biomass production would be practically valuable because the biomass can be measured much faster or even in a continuous manner. Thus far, only a few attempts have been made to develop a model that describes growth of nisin-producers and nisin production (Cabo et al, 2001b; Guerra et al., 2001; Lv et al., 2005).

The logistic function model (Pearl and Reed, 1920) is commonly used to describe growth of microbial cells, including L. lactis . The model describes the growth of a microbial population as a function of maximum population density, lag time, specific growth rate, and time. Since the logistic equation is symmetrical around time t m, the the

Gompertz function is usually chosen for generating an asymmetrical growth curve.

Although both functions were shown to sufficiently describe the growth of microbial

147 populations in most cases, both of them failed to capture the decrease of biomass during the death phase. In this study, both functions were modified to describe the growth of L. lactis during nisin fermentation.

Mathematical models describing nisin production were developed with either the

Leudeking-Piret model (Cabo et al., 2001b; Guerra et al., 2001) or with a term for growth-associated nisin production and a term for nisin degradation or adsorption (Lv et al., 2005). Nisin is usually reported as a primary metabolite, however, it can acquire a secondary nature when the productive period of the culture is prolonged by imposing a stepwise-pH profile (Cabo et al., 2001b). In this study, the model for nisin production was developed and validated in an attempt to predict nisin production in batch culture using biomass density observed during fermentation. The results provide valuable information essential for development of nisin production.

MATERIALS AND METHODS

Microorganisms and media

Lactococcus lactis subsp. lactis (NIZO 22186), nisin Z-producing strain, was used in this study. Nisin activity was quantified using Lactobacillus sakei (ATCC 15521) as the nisin-sensitive test organism. L. lactis and L. sakei were grown at 30 °C for 14 h in a complex medium (CM) and Lactobacillus MRS broth (Difco Laboratories, Detroit, MI), respectively. The CM consisted of 40 g of sucrose; 10 g of peptone (Amber Ferm 4015G,

Universal Flavors, Milwaukee, WI); 10 g of yeast extract (Ardamine Z, Sensient

Bionutrient, Indianapolis, IN); 10 g of KH 2PO 4; 2 g of NaCl; and 0.2 g of MgSO 4·7H 2O

148 per liter of deionized water. The initial pH of the CM medium was adjusted to 6.8 using 4

N NaOH. For long-term storage, all stock cultures were maintained at -80 °C in 20% glycerol.

Batch nisin fermentation

Nisin fermentations were conducted in a 1.25-l Bioflo II fermentor (New

Brunswick Scientific, Edison, NJ) at 30 °C with agitation at 100 rpm. The reactor vessel was sterilized with nitrogenous components and mineral salts of CM while sucrose was sterilized separately and added to the reactor aseptically. The medium was inoculated with 1% (v/v) of 14-h culture of L. lactis at the beginning of each trial. During fermentation, the pH was controlled at 6.8 by addition of 4 N NaOH. The experiment was replicated three times.

Analysis

Samples were collected every hour for 12 h and analyzed for biomass density, nisin, sucrose, and lactic acid concentrations.

Biomass. Cell density was estimated by optical density at 600 nm using a spectrophotometry DU series 500 (Beckman, Fullerton, CA). Sterile fermentation medium was used as a blank. Optical density values were converted into biomass concentration by using a standard curve (y = 0.1663x – 0.0199; y: biomass (g/l) and x: optical density at 600 nm).

Nisin. The sample of fermentation broth was immediately adjusted to pH 3.0 using concentrated HCl, and 0.01% v/v of Tween 20 (J.T. Baker, Phillipsburg, NJ) was

149 added in order to avoid any nonspecific adsorption of nisin on the container’s surfaces.

The sample was then heated at 90 °C for 5 min to eliminate the protease activity in the broth. After centrifugation at 3,800 x g, 4 °C for 20 min (Sorvall Super T 21, ST-H750,

Kendro Lab Products, Newton, CT), clear supernatant was collected and kept frozen at -

20 °C until analysis. Nisin quantifications were performed by the agar well diffusion bioassay, using Lactobacillus sakei and pre-diffusion at 4 °C, as described in our previous study (Pongtharangkul and Demirci, 2004).

Sucrose. The fermentation broth was centrifuged at 3,800 x g, 4 °C for 15 min and analyzed with the Dinitrosalicylic Colorimetric (DNS) Method. For initial acid hydrolysis, 100 µl of concentrated HCl was added to 5 ml of the sample. After allowing the hydrolysis to proceed at 90 °C for 5 min, 0.25 ml of 5 N KOH solution was added to neutralize the acid. Then the hydrolyzed samples were analyzed using the DNS Method as described by Miller (1959).

Lactic acid. Lactic acid concentrations were determined using a high performance liquid chromatograph (HPLC) equipped with a refractive index detector (Waters,

Franklin, MA). Components were separated on a Bio-Rad Aminex HPX-87H column

(300 x 7.8 mm) (Bio-Rad, Richmond, CA) with 0.012 N sulfuric acid as a mobile phase at a flow rate of 0.8 ml/min with an injection volume of 20 µl and a column temperature of 65 °C. Before injection, the samples were centrifuged (3,800 x g, 4 °C) for 15 min and filtered through 0.22 µm filters (13 mm diameter disc filters, Millipore, Bedford, MA).

150 Model development

Primary modeling of biomass production, sugar consumption, and the production of nisin and lactic acid were developed using an average of two different experimental replicates. Out of three experimental data sets, two data sets were selected so that the model yielded neither the highest nor the lowest RMS values. The third data set was used for a validation of the constructed model. All models were designed so that they contained biologically relevant parameters. In order to achieve the best fit of the experimental data, model fitting and parameters estimation was performed by minimizing the sum square error between experimental and model-predicted values, using the non- linear least-squares (quasi-Newton) method provided by the macro ‘Solver’ of Microsoft

Excel 2000. Regression through origin (RTO) was performed using MINITAB (release

13.30, State College, PA) and the R-square value was calculated from the sum square of regression (SSR) divided by the sum square of total (SST).

Model for biomass production

The growth curve of bacteria, in which the biomass is plotted against time, usually exhibits a sigmoid pattern with a lag phase followed by an exponential phase and then a stationary phase. Modified sigmoid functions (logistic and Gompertz) were previously modified to contain parameters with biological meaning instead of mathematical parameters and successfully used to describe growth of several bacteria

(Zwietering et al., 1990).

The generic logistic equation (Pearl and Reed, 1920) can be written as:

a x = Eqn. (8.1) []1+ exp ()b − ct

151 The modified logistic equation of Zwietering et al. (1990):

x x = max Eqn. (8.2)  4µ  1+ exp  m ()λ − t + 2   xmax  where x: biomass density (g/l)

xmax : maximum biomass density (g/l)

t: time (h)

µm: specific growth rate (1/h)

λ: lag time of growth or t-axis intercept of the tangent through the

inflection point (h)

a, b, and c: model parameters

Autolysis is common in lactococcus strains, as it appears to be a crucial step in the release of intracytoplasmic enzymes such as peptidases that produce free amino acids for growth. L. lactis NIZO 22186, used in this study, as previously reported by De Vuyst and

Vandamme (1992), exhibits a death phase, indicated by a decrease of biomass density, as a result of autolysis. Thus, the modified logistic equation (Eqn. 8.2), which can only describe the sigmoid growth curve with constant upper asymptote, failed to describe the biomass production of L. lactis NIZO 22186 (Fig. 8.1).

In order to capture the decrease of biomass density toward the end of fermentation, Eqn. (8.2) has been re-modified in this study. An additional term,

(t − t ) max , which represents the death of cells, has been added, and includes two more τ parameters; the time when biomass reaches its maximum (t max ) and the time since t max that biomass reaches zero ( τ), as shown in Eqn. (8.3). This term was designed so that it is

152 negative only after the biomass reaches its maximum value, i.e., t > t max . The value of t max was calculated by iteration method using a time step of 0.1 h.

The re-modified Logistic Equation developed in this study:

 (t − tmax ) xmax 1−   τ  x = Eqn. (8.3)  4µ  1+ exp  m ()λ − t + 2   xmax 

dx where tmax : time (h) when ≈ 0 dt

τ: time (h) since t max for biomass to reach zero

1.6

1.4

1.2

1.0

0.8

0.6 Experimental

Biomass, x (g/l) 0.4 Modified logistic (Eq. 8.2)

0.2 Re-modified logistic (Eq. 8.3)

0.0 0 1 2 3 4 5 6 7 8 9 10111213

Time, t (h)

Figure 8.1. Growth curve of L. lactis fitted with the modified logistic and re-modified logistic models.

153 A more flexible Gompertz function was also evaluated and modified using the same approach as follows.

The Gompertz Equation (Gompertz, 1825; Winsor, 1932):

x = a ⋅ exp[− exp (b − ct )] Eqn. (8.4)

The modified Gompertz Equation (Zwietering et al., 1990):

   µ m ⋅ e x = xmax ⋅ exp − exp  ()λ − t +1 Eqn. (8.5)   xmax 

The re-modified Gompertz Equation developed in this study:

   µm ⋅ e  (t − tmax ) x = xmax ⋅ exp − exp  ()λ − t +1⋅ 1−  Eqn. (8.6)   xmax   τ  where x: biomass density (g/l)

xmax : maximum biomass density (g/l)

t: time (h)

µm: specific growth rate (1/h)

λ: lag time of growth or t-axis intercept of the tangent through the

inflection point (h)

a, b, and c: model parameters

dx tmax : time (h) when ≈ 0 dt

τ: time (h) since t max for biomass to reach zero

Subjective comparisons of the actual growth curves generated with modified and re-modified logistic and Gompertz functions are given by plotting both the actual data and the predicted values from the model (Figs. 8.1 and 8.2). Further comparison was

154 achieved via regression through origin of actual experimental data and predicted values.

If the model can describe the data, the resulting R-square should be high and the slope should fall near unity.

1.6

1.4

1.2

1.0

0.8

0.6

Biomass, x (g/l) Experimental Value 0.4

Modified Gompertz, Eq. (8.5) 0.2 Re-modified Gompertz, Eq. (8.6)

0.0 0 1 2 3 4 5 6 7 8 9 10111213 Time, t (h) Figure 8.2. Growth curve of L. lactis fitted with the modified Gompertz and re-modified

Gompertz models.

Overall, the modifications made in Eqns. (8.3) and (8.6) allow the models to describe the data better, indicated by lower RMS values, at a cost of having two more parameters in the models (Table 1). For example, compared to Eqn. (8.2) (RMS = 0.117),

Eqn. (8.3) gave a much lower RMS (0.072). The same trend was also observed between

Eqns. (8.5) and (8.6) as shown in Table 1. Interestingly, Eqn. (8.6) represents the data better overall (Fig. 8.2), but Eqn. (8.3) yields a more accurate predicted biomass density around the maximum value (Fig. 8.1). Therefore, Eqn. (8.3) would be more appropriate

155 for applications in which the accurate prediction of maximum biomass density (x max) is required.

Table 8.1. Estimated model parameters of growth curve, root mean square (RMS), correlation coefficient (R-square), and slope from the regression of experimental and predicted value when fitted with various models.

Models RMS µm λ Kd tmax Modified logistic* 0.117 0.281 2.039 N/A** N/A Modified Gompertz* 0.110 0.265 1.777 N/A N/A Re-modified logistic 0.072 0.267 2.426 18.537 8.2 Re-modified Gompertz 0.074 0.259 2.158 19.350 8.5 *Modified Equations according to Zwietering et al. (1990) ** Not available for the model

The re-modified logistic and Gompertz models were validated with an independent set of experimental data and a regression of experimental and predicted values were conducted as shown in Figs. 8.3 and 8.4, respectively. In general, both Eqns.

(8.3) and (8.6) can be used successfully to predict the growth of L. lactis NIZO 22186 in batch culture (Table 8.2). Estimated parameters (Table 8.1) yielded sufficiently accurate predictions for both re-modified logistic (slope = 1.01, R-square = 0.99) and re-modified

Gompertz models (slope = 1.01, R-square = 0.99).

Although NIZO 22186 yields a significantly lower x max compared to the results obtained from batch culture of L. lactis strain ATCC 11454 (Lv et al., 2004), both Eqns.

(8.3) and (8.6) could be successfully used to describe the growth of ATCC 11454

(experimental data obtained from Lv et al., 2004) as shown in Fig. 8.5. Therefore, the re- modified logistic and Gompertz models suggested in this study are applicable for other strains of nisin producers in addition to NIZO 22186.

156 1.6

1.4

1.2

1.0

0.8

0.6

0.4 y = 1.01 x Predicted Biomass, x (g/l) x Biomass, Predicted 0.2 R-square = 0.99

0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Experimental Biomass, x (g/l)

Figure 8.3. Regression of experimental and predicted biomass densities of L. lactis from validation with an independent set of data when fitted with the re-modified logistic model.

1.6

1.4

1.2

1.0

0.8

0.6

0.4

Predicted Biomass, x (g/l) x Biomass, Predicted y = 1.01 x 0.2 R-square = 0.99

0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Experimental Biomass, x (g/l)

Figure 8.4. Regression of experimental and predicted biomass densities of L. lactis from validation with an independent set of data when fitted with the re-modified Gompertz model.

157 Table 8.2. Root mean square (RMS), correlation coefficient (R-square), and slope of the regression for validations of various growth models.

Models RMS R-square Slope Modified logistic* 0.146 0.98 1.09 Modified Gompertz* 0.130 0.98 1.06 Re-modified logistic 0.094 0.99 1.01 Re-modified Gompertz 0.091 0.99 1.01 *Modified Equations according to Zwietering et al. (1990)

4.5

4.0

3.5

3.0

2.5

2.0

Biomass, x (g/l) 1.5 Experimental Value 1.0 Re-modified Logistic, Eq. (8.3)

0.5 Re-modified Gompertz, Eq. (8.6)

0.0 0 1 2 3 4 5 6 7 8 9 10111213 Time, t (h)

Figure 8.5. Growth curve of L. lactis ATCC 11454 fitted with the re-modified logistic and re-modified Gompertz models (experimental data obtained from Lv et al., 2004).

158 Model for lactic acid production

Lactic acid production can be described using the Luedeking-Piret model

(Luedeking and Piret, 1959) as follows:

dP dx L = α + βx Eqn. (8.7) dt dt where PL: lactic acid concentration (g/l)

α: growth-associated constant (g-lactic acid/g-biomass)

β: non-growth-associated constant (g-lactic acid/g-biomass/h)

x: biomass density (g/l)

The model obtained is shown in Fig. 8.6. According to the result, lactic acid production can be described as related to both primary and secondary metabolites ( α =

8.422 g-lactic acid/g-biomass and β = 2.211 g-lactic acid/g-biomass/h). The obtained lactic acid yield (Y P/S ) of 0.72 g-lactic acid/g-sucrose is comparable to the value reported for other nisin-producer strains, such as ATCC 11454 (Lv et al., 2005) and A164 (Cheigh et al., 2002), when the same level of sugar is used.

Model for sucrose consumption

Due to the homofermentative nature of lactococci, sucrose is mainly converted into lactic acid and can be described by the following equation.

dS 1 dP L − = + ms x Eqn. (8.8) dt YP / S dt

159 35 dP dx 30 L = .8 422 + .2 211 x dt dt 25 (g/l) L 20 RMS = 1.492 15

10 Lactic acid, P Experimental 5 Predicted

0 0 1 2 3 4 5 6 7 8 9 10111213 Time, t (h)

Figure 8.6. Experimental and predicted lactic acid production during batch fermentation.

45 40 dS 1 dL − = + 43.0 x 35 dt 0.717 dt 30 RMS = 4.022 25

20

15 Sucrose, S (g/l) S Sucrose, Experimental 10 Predicted 5

0 0 1 2 3 4 5 6 7 8 9 10111213 Time, t (h)

Figure 8.7. Experimental and predicted sucrose consumption during batch fermentation.

160 where S: sucrose concentration (g/l)

YP/S : lactic acid yield (g-lactic acid/g-sucrose)

ms: maintenance constant (g-sucrose/g-biomass/h)

x: biomass density (g/l)

The results obtained suggest further modification because the model does not fit well enough when observed visually (Fig. 8.7). Since the predicted values are lower than the experimental values, there may be an underlying factors, which may contributes to sucrose consumption other than lactic acid production and cell maintenance.

Model for nisin production

Several models have been proposed to describe nisin production, including the

Luedeking-Piret model shown below:

dN dx = α + β x Eqn. (8.9) dt N dt N where N: nisin concentration (mg-nisin/l)

αN: growth-associated constant (mg-nisin/g-biomass)

βN: non-growth-associated constant (mg-nisin/g-biomass/h)

x: biomass density (g/l)

It should be noted that Eqn. (8.9) could describe nisin production only up to the time that the maximum nisin titer was reached, approximately 7 h for this study (Fig.

8.8). Interestingly, the obtained parameters ( αN = -9.273 mg-nisin/g-biomass and βN =

18.295 mg-nisin/g-biomass/h) indicated that a non-growth-associated factor seemed to

161 play a more important role in nisin production observed in this study. This result contrasts with previous reports in which the value of αN was usually higher than βN, for example, αN = 39.71 and βN = 28.20 (Cabo et al., 2001b). In other words, these results indicated that nisin production seemed to depend more on biomass density more than growth (change of biomass over time). When the obtained model was validated using an independent set of experimental data (Fig. 8.9), the regression indicated that the model could be successfully used to predict the nisin production (slope = 1.02 and R-square =

0.98). Although high correlations between experimental and predicted values have been reported (Cabo et al., 2001b), it should be noted that none of the models proposed for nisin production have been properly validated with a separate set of experimental data.

70

60 dN dx = − .9 273 +18 .295 x 50 dt dt

40 RMS = 2.780

30

Nisin, N (mg/l) 20 Experimental 10 Predicted

0 0 1 2 3 4 5 6 7 8 Time, t (h)

Figure 8.8. Experimental and predicted nisin production during batch fermentation.

162 70

60

50

40

30

20 Predicted Nisin (mg/l) Predicted Nisin y = 1.04 x 10 R-square = 0.98

0 0 10 20 30 40 50 60 70 Experimental Nisin (mg/l)

Figure 8.9. Validation of nisin production model with an independent set of data.

CONCLUSIONS

In this study, several models have been proposed for describing growth, sucrose consumption, nisin and lactic acid production of L. lactis NIZO 22186 during batch fermentation. Sigmoidal functions, such as the logistic and Gompertz equations, have been modified in order to capture the decrease of biomass from autolysis of the producer.

The re-modified logistic and Gompertz models proposed in this study adequately described growth and the validation showed that they could be used for accurate prediction (slope = 1.01, R-square = 0.99). As for nisin production, the Luedeking-Piret model fit well and could be successfully used to predict nisin production from the value of biomass (slope = 1.04, R-square = 0.98). This information would be valuable for further development of nisin production in batch culture. However, further adjustments will be needed if the conditions used in fermentation are changed.

163 REFERENCES

Cabo, M. L., M. A. Murado, M. P. Gonzalez, J. A. Vazquez, and L. Pastoriza. 2001a. An empirical model for describing the effects of nitrogen sources on nisin production. Lett. Appl. Microbiol. 33: 425-429. Cabo, M. L., M. A. Murado, M. P. Gonzalez, and L. Pastoriza. 2001b. Effects of aeration and pH gradient on nisin production. A mathematical model. Enz. Microbial. Technol. 29: 264-273. Cheigh, C. I., H. J. Choi, H. Park, S. B. Kim, M. C. Kook, T. S. Kim, J. K. Hwang, and Y. R. Pyun. 2002. Influence of growth conditions on the production of a nisin-like bacteriocin by Lactococcus lactis subsp. lactis A164 isolated from kimchi. J. Biotechnol. 95: 225-235. Desjardins, P., J. Meghrous, and C. Lacroix. 2001. Effect of aeration and dilution rate on nisin Z production during continuous fermentation with free and immobilized Lactococcus lactis UL719 in supplemented whey permeate. Int. Dairy. J. 11: 943- 951. De Vuyst, L., and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. De Vuyst, L. 1994. Nisin production variability between natural Lactococcus lactis subsp. lactis strains. Biotech. Lett. 16(3): 287-292. Gompertz, B. 1825. On the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of life contingencies. Philosophical Transactions of the Royal Society 182: 513-585. Guerra, N. P., M. L. Rua, and L. Pastrana. 2001. Nutritional factors affecting the production of two bacteriocins from lactic acid bacteria on whey. Int. J. Food Microbiol. 70: 267-281. Kim, W. S., R. J. Hall, and N. W. Dunn. 1997. Host specificity of nisin production by Lactococcus lactis . Biotech. Lett. 19(12): 1235-1238.

164 Li, C., J. Bai, Z. Cai, and F. Ouyang. 2002. Optimization of a cultural medium for bacteriocin production by Lactococcus lactis using response surface methodology. J. Biotechnol. 93: 27-34. Luedeking, R., and E. L. Piret. 1959. Transient and steady states in continuous fermentation. Theory and experiment. J. Biochem. Microbiol. Technol. Eng. 1: 431-459. Lv, W., W. Cong, and Z. Cai. 2004. Nisin production by Lactococcus lactis subsp. lactis under nutritional limitation in fed-batch culture. Biotechnol. Lett. 26: 235-238. Lv, W., X. Zhang, and W. Cong. 2005. Modelling the production of nisin by Lactococcus lactis in fed-batch culture. Appl. Microbiol. Biotechnol. 68: 322-326. Miller, G. L. 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31: 426. Pearl, R., and L. J. Reed. 1920. On the rate of growth of the population of the United States since 1970 and its mathematical representation. Proc. Nat. Acad. Sci. 6: 275-288. Shimizu, H., T. Mizuguchi, E. Tanaka, and S. Shioya. 1999. Nisin production by a mixed-culture system consisting of Lactococcus lactis and Kluyveromyces marxianus . Appl. Environ. Microbiol. 65(7): 3134-3141. Swietering, M. H., I. Jongenburger, F. M. Rombouts, and K. Van’t Riet. 1990. Modeling of the bacterial growth curve. Appl. Environ. Microbiol. 56(6): 1875-1881. Winsor, C. P. 1932. The Gompertz curve as a growth curve. Proc. Nat. Acad. Sci. 18: 1- 8.

165 CHAPTER 9

CONCLUSIONS AND SCOPE FOR FUTURE RESEARCH

In this study, a biofilm reactor with Plastic Composite Support (PCS) was evaluated for nisin production using L. lactis strain NIZO 22186. Gradual improvement of nisin production was achieved in each step throughout the study (Fig. 9.1).

5000 4314

4000 3553

3000 2208

2000 1566 Nisin (IU/ml) Nisin 1100 1000 579 697

0 BF with 8% FC with BF with 8% FC with BF with BF with pH BF with Glucose 8% Sucrose 8% Optimized Profile fed-batch Glucose Sucrose Medium Const4 and pH Profile Const12

Figure 9.1. Summary of improvements in nisin production achieved in this study

First, an agar well diffusion technique for accurate assay of nisin was established and used throughout the research. The method using L. sakei with pre-diffusion gave the most accurate and precise results among tested procedures. Various blends of PCS were tested, and the most appropriate PCS was chosen for construction of a biofilm reactor.

166 Among tested PCS blends, SFYB+ was selected for long-term biofilm nisin fermentation by L. lactis due to its highest nitrogen content, moderate nitrogen-leaching rate, and sufficient viable cells attached on the PCS.

The high-biomass density of the biofilm reactor was found to contribute to a significantly shorter lag time of nisin production relative to a suspension cell reactor. In comparison to glucose (579 IU/ml), sucrose significantly increased the nisin production rate by 1.4-fold (1,100 IU/ml) in the biofilm reactor. However, results showed that the higher level of sucrose (8% w/v) had a suppressing effect on nisin production and a stimulating effect on lactic acid production. The medium evaluation study suggested a complex media for nisin production in the biofilm reactor as follows: 4% (w/v) sucrose,

0.02% (w/v) MgSO 4•7H 2O, and 1% (w/v) KH 2PO 4, 0.2% (w/v) NaCl, 1% (w/v) peptone, and 1% (w/v) yeast extract. Nisin production rate in the biofilm reactor was significantly increased by 3.8-fold (2,208 IU/ml) with this suggested complex medium.

Nisin production in a biofilm reactor was highly affected by the pH profiles during fermentation. Profile Const4, in which pH was allowed to drop freely via autoacidification after 4 h of fermentation, was found to yield almost 1.9-time higher nisin (3,553 IU/ml) than profile Const12 (1,898 IU/ml), possibly as a result of less adsorption of nisin onto producer cells. Therefore, a combination of constant pH and autoacidification period (profile Const4) was recommended as the pH profile during nisin production in a biofilm reactor.

When profile Const12 was applied, fed-batch fermentation enhanced nisin production of suspended-cell (4,188 IU/ml) and biofilm (4,314 IU/ml) reactors, yielded

1.8 and 2.3-fold higher nisin titer than their respective batch fermentation. On the other

167 hand, pH profiles that include periods of autoacidification (profiles Const4 and Const4-

Stepwise) resulted in significantly lower nisin production in fed-batch fermentation

(2,494 and 1,861 IU/ml for biofilm reactor using profile Const4 and Const4-Stepwise, respectively) due to higher toxicity of produced lactic acid in acidic environment.

Silicic acid has been successfully used to remove nisin from the fermentation broth and higher nisin adsorption onto silicic acid can be achieved when the adsorption was carried out at pH 6.8 (67% adsorption) than at pH 3.0 (54% adsorption). The maximum recovery (47% of nisin was harvested) can be achieved when the adsorption was carried out at pH 6.8 and 1 M NaCl with 20% ethanol was used as an eluent.

Mathematic models of biomass and nisin production during fermentation not only provide information about the kinetic-metabolic nature of nisin, but also facilitate the control and optimization of nisin production. The re-modified logistic and Gompertz models proposed in this study adequately described the growth and the validation showed that they could be use for the prediction accurately (slope = 1.01, R-square = 0.99). As for nisin production, the Luedeking-Piret model fit considerably well and could be successfully use to predict nisin production from the value of biomass (slope = 1.04, R- square = 0.98).

Several approaches used in this study have significantly improved nisin production in the biofilm reactor. However, commercial application of the biofilm reactor may require a new design of PCS orientation in the reactor along with a further up- scaling study. Since L. lactis NIZO 22186 used in this research showed a high nisin degrading ability, higher nisin production could be suspected if a producer strain with lower ability to degrade nisin were used.

168 In this study, fed-batch fermentation greatly enhanced nisin production in both suspended-cell and biofilm reactors. A better control over residual sucrose concentration should be attempted and a certain optimum level of residual sucrose concentration should be established in order to get the most benefit from a fed-batch culture. Moreover, since the biofilm reactor has great potential for being developed into a continuous system, a further development on continuous biofilm system is recommended.

The recovery procedure proposed in this study should be refined and used to construct an on-line adsorption unit for the existing biofilm reactor. The approach has to take into account the fact that the adsorption and desorption of nisin was highly affected by the pH of fermentation broth during adsorption process. In the recovery unit will be used along with the pH profiles other than a constant pH profile (profile Const12), the effect of pH on adsorption and desorption of nisin need to be studied in more detail. Last, further development of mathematical models of nisin production under a continuous fermentation and/or on-line recovery would give valuable information for optimizing and controlling nisin production.

169 REFERENCES

Amiali, M. N., C. Lacroix, and R. E. Simard. 1998. High nisin Z production by Lactococcus lactis UL719 in whey permeate with aeration. World J. Microbiol. Biotechnol. 14: 887-894. Ariyapitipun, T., A. Mustapha, and A. D. Clarke. 1999. Microbial shelf life determination of vacuum-packaged fresh beef treated with polylactic acid, lactic acid, and nisin solutions. J. Food Protection 62(8): 913-920. Ariyapitipun, T., A. Mustapha, and A. D. Clarke. 2000. Survival of Listeria monocytogenes Scott A on vacuum-packaged raw beef treated with polylactic acid, lactic acid, and nisin. J. Food Protection 63(1): 131-136. Azeredo, J., J. Visser, and R. Oliveira. 1999. Exopolymers in bacterial adhesion: interpretation in terms of DLVO and XDLVO theories. Colloids and Surfaces B: Biointerfaces , 14: 141-148. Bailey, F. J., and A. Hurst. 1971. Preparation of a highly active form of nisin from Streptococcus lactis. Can. J. Microbiol. 17: 61-67. Baranova, I. P., N. S. Egorov, D. M. Isakova, M. N. Khodzhaev, A. Y. Popov, and Y. I. Kozlova. 1989. Employment of silica adsorbents produced in the USSR for isolating nisin from the native solution. Biotekhnologiya 5: 588-593. Berridge, N. J., and J. Barrett. 1952. A rapid method for the turbidimetric assay of antibiotics. J. Gen. Microbiol. 6: 14-20. Bertrand, N., I. Fliss, and C. Lacroix. 2001. High nisin-Z production during repeated- cycle batch cultures in supplemented whey permeate using immobilized Lactococcus lactis UL719. International Dairy Journal 11: 953-960. Bober, J., and A. Demirci. 2004. Nisin fermentation by Lactoccocus lactis subsp. lactis using plastic composite supports in biofilm reactors. Agricultural Engineering International: the CIGR J Sci Res Dev Manuscript FP 04-001 Vol. VI 15 pp. Bott, T. R. 1995. Biological growth on heat exchanger surfaces. In Bott, T.R. Fouling of Heat Exchangers , chapt. 12, pp. 223-267, Elsevier, Amsterdam.

170 Bouksaim, M., I. Fliss, J. Meghrous, R. E. Simard, and C. 1998. Immunodot detection of nisin Z in milk and whey using enhanced chemiluminescence. J. Appl. Microbiol. 84: 176-184. Boyaval, P., P. Bhugaloo-Vial, F. Duffes, A. Metivier, X. Dousset and D. Marion. 1998. Reacteurs a haute densites cellulaires pur la production de solutions concentrees de bacteriocines. Lait 78: 129-133. cited in E. Parente and A. Ricciardi. 1999. Production, recovery and purification of bacteriocins from lactic acid bacteria. Appl. Microbiol. Biotechnol. 52: 628-638. Breukink, E., P. Ganz, B. de Kruijff, and J. Seelig. 2000. Binding of nisin Z to biolayer vesicles as determined with isothermal titration calorimetry. Biochem. 36: 10247- 10254. Breukink, E., C. van Kraaij, R. A. Demel, R. J. Siezen, O. P. Kuipers, and B. de Kruijff. 1997. The C-terminal region of nisin is responsible for the initial interaction of nisin with the target membrane. Biochem. 36: 6968-6976. Brotz, H., and H. G. Sahl. 2000. New insights into the mechanism of action of lantibiotics—diverse biological effects by binding to the same molecular target. J Antimicrob Chemother 46(1): 1-6. Bryers, J. D. 2000. Biofilm formation and persistence. In Biofilms II: Process Analysis and Applications , edited by Bryers, J.D., pp. 45-88, Wiley-Liss, Inc., New York. Bryers, J. D. 1987. Biologically active surfaces: processes governing the formation and persistence of biofilms. Biotechnology Progress, 3: 57-68. Buchman, G. W., S. Banerjee, and J. N. Hansen. 1988. Structure, expression, and evolution of a gene encoding the precursor of nisin, a small protein antibiotic. J. Biol. Chem. 263: 9508-9514. Budde, B. B., and M. Rasch. 2001. A comparative study on the use of flow cytometry and colony forming units for assessment of the antibacterial effect of bacteriocins. Int. J. Food Microbiol. 63: 65-72. Busscher, H. J., R. Bos, and H. C. van der Mei. 1995. Initial microbial adhesion is determinant for the strength of biofilm adhesion. Hypothesis. FEMS Microbiology Letters, 128: 229-234.

171 Busscher, H. J., and H. C. van der Mei. 2000. Initial microbial adhesion events: mechanisms and implications. In Community structure and co-operation in biofilms , edited by Allison, D., Gilbert, P., Lappin-Scott, H., and Wilson, M., pp. 25-36, Cambridge University Press, Cambridge. Cabo, M. L., M. A. Murado, M. P. Gonzalez, and L. Pastoriza. 1999. A method for bacteriocin quantification. J. Appl. Microbiol. 87: 907-914. Cabo, M.L., M.A. Murado, M.P. Gonzalez and L. Pastoriza. 2001. Effects of aeration and pH gradient on nisin production. A mathematical model. Enzyme Microbial Technol. 29: 264-273. Cabral, J.M.S., and J. Tramper. 1994. Bioreactor design. In Applied Biocatalysis , edited by Cabral, J.M.S., Best, D., Boross, L., and Tramper, J., pp. 333-370, Harwood Academic Publishers, Switzerland. Caserio, G., A. Ciampella, M. Gennari, and A. M. Barliezizi. 1979a. Utilization of nisin in cooked sausages and other cured meat products. Ind. Aliment. , 18: 1-12, 19. Caserio, G., M. Stecchini, M. Pastore, and M. Gennari. 1979b. The individual and combined effects of nisin and nitrite on the spore germination of Clostridium perfringens in meat mixtures subjected to fermentation. Ind. Aliment. 18: 894-898. Chan, L., F. Ouyang, and J. Bai. 2000. Extractive cultivation of Lactococcus lactis using

a polyethylene glycol/MgSO 4.7H 2O aqueous two-phase system to produce nisin. Biotechnol. Lett. 22: 843-847. Chandrapati, S., and D. J. O’Sullivan. 1998. Procedure for quantifiable assessment of nutritional parameters influencing nisin production by Lactococcus lactis subsp. lactis . J. Biotechnol. 63: 229-233. Characklis, W. G., and K. C. Marshall. 1990. Biofilms , John Wiley and Sons, New York. Cheigh, C. I., H. J. Choi, H. Park, S. B. Kim, M. C. Kook, T. S. Kim, J. K. Hwang, and Y. R. Pyun. 2002. Influence of growth conditions on the production of a nisin-like bacteriocin by Lactococcus lactis subsp. lactis A164 isolated from kimchi. J. Biotechnol. 95: 225-235. Cheigh, C. I., M. C. Kook, S. B. Kim, Y. H. Hong, and Y. R. Pyun. 2004. Simple ond- step purification of nisin Z from unclarified culture broth of Lactococcus lactis

172 subsp. lactis A164 using expanded bed ion exchange chromatography. Biotechnol. Lett. 26: 1341-1345. Cheigh, C. I., H. Park, H. J. Choi, and Y. R. Pyun. 2005. Enhanced nisin production by increasing genes involved in nisin Z biosynthesis in Lactococcus lactis subsp. lactis A164. Biotechnol. Lett. 27: 155-160. Cheigh, C. I., and Y-R. Pyun. 2005. Nisin biosynthesis and its properties. Biotechnol. Lett. 27: 1641-1648. Chinachoti, N., H. Matsusaki, K. Sonomoto, and A. Ishizaki. 1997a. Utilization of xylose as an alternative carbon source for nisin Z production by Lactococcus lactis IO-1. J. Fac. Agric. Kyushu Univ. 43: 421-436. Chinachoti, N., N. Endo, K. Sonomoto, and A. Ishizaki. 1997b. Bioreactor systems for efficient production and separation of nisin Z using Lactococcus lactis IO-1. J. Fac. Agric. Kyushu. Univ. 43: 421-436. Choi, M. H., and Y. H. Park. 2000. Selective control of lactobacilli in kimchi with nisin. Lett. Appl. Microbiol. 30: 173-177. Cintas, L. M., P. Casaus, M. F. Fernandez and P. E. Hernandez. 1998. Comparative antimicrobial activity of enterocin L50, pediocin PA-1, nisin A and lactocin S against spoilage and foodborne pathogenic bacteria. Food Microbiol. 15: 289-298. Cleveland, J., T. J. Montville, I. F. Nes, and M. L. Chikindas. 2001. Bacteriocins: safe, natural antimicrobials for food preservation. Int. J. Food Microbiol. 71: 1-20. Cotton, J. C., A. L. Pometto III, and J. Gvocdenovic-Jeremic. 2001. Continuous lactic acid fermentation using a plastic composite support biofilm reactor. Appl Microbiol. Biotechnol. 57: 626-630. Coventry, M. J., J. B. Gordon, M. Alexander, M. W. Hickey, and J. Wan. 1996. A food- grade process for isolation and partial purification of bacteriocins of lactic acid bacteria that used diatomite calcium silicate. Appl. Environ. Microbiol. 62(5): 1764- 1769. Cutter, C. N., and G. R. Siragusa. 1998. Incorporation of nisin into a meat binding system to inhibit bacteria on beef surfaces. Lett. Appl. Microbiol. 27: 19-23. Daoudi, L., C. Turcotte, C. Lacroix, and I. Fliss. 2001. Production and characterization of anti-nisin Z monoclonal antibodies: suitability for distinguishing acive from

173 inactive forms through a competitive enzyme immunoassay. Appl. Microbiol. Biotechnol. 56: 114-119. Davies, E. A., H. E. Bevis, and J. Delves-Broughton. 1997. The use of the bacteriocin, nisin, as a preservative in ricotta-type cheeses to control the food-borne pathogen Listeria monocytogenes. Lett. Appl. Microbiol. 24: 343-346. Davies, E. A., H. E. Bevis, R. Potter, J. Harris, G. C. Williams, and J. Delves-Broughton. 1998. Effect of pH on the stability of nisin solution during autoclaving. Letts. Appl. Microbiol. 27: 186-187. Davies, E. A., C. F. Milne, H. E. Bevis, R. W. Potter, J. M. Harris, G. C. Williams, L. V. Thomas, and J. Delves-Broughton. 1999. Effective use of nisin to control lactic acid bacteria spoilage in vacuum-packed bologna-type sausage. J. Food Prot. 62: 1004- 1010. Davies, E. A., and J. Delves-Broughton. 2000. Nisin. In Encyclopedia of Food Microbiology , eds. C.A. Batt and P.D. Patel, 191-198. Academic Press. Delves-Broughton, J. 1990. Nisin and its uses as a food preservative. Food Technol. 11: 110-117. Delves-Broughton, J., and M. Friis. 1998. Nisin preparation – production, specifications and assay procedures. Bull. Int. Dairy Fed. 329: 18-19. Demel, R. A., T. Peelen, R. J. Siezen, B. de Kruijff, and O. P. Kuipers. 1996. Nisin Z, mutant nisin Z and lacticin 481 interactions with anionic lipids correlate with antimicrobial activity. A monolayer study. Eur. J. Biochem. 235: 267-274. Demirci, A., A. L. Pometto III, K-L. G. Ho. 1995. Continuous ethanol production in biofilm reactors containing plastic composite rings and disks. Proc.-Biomass Conference of the Americas: Energy, Environment, Agricultural and Industry, 2 nd , Aug, 21-24, Portland, Oregon, National Renewable Energy Laboratory, Golden, Colorado. Demirci, A., A. L. Pometto III, K-L. G. Ho. 1997. Ethanol production by Saccharomyces cerevisiae in biofilm reactors. J. Ind. Microbiol. 19: 299-304. de Vos, W. M., J. W. M. Mulders, J. Hugenholtz, R. J. Siezen, and O. P. Kuipers. 1993. Properties of nisin Z and the distribution of its gene, nisZ , in Lactococcus lactis. Appl. Environ. Microbiol. 59: 213-218.

174 De Vuyst, L. 1994. Nisin production variability between natural Lactococcus lactis subsp. lactis strains. Biotechnol. Lett. 16: 287-292. De Vuyst, L. and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. De Vuyst, L. and E. J. Vandamme. 1993. Influence of the phosphorus and nitrogen source on nisin production in Lactococcus lactis supsp. lactis batch fermentations using a complex medium. Appl. Microbiol. Biotechnol. 40: 17-22. De Vuyst, L. and E. J. Vandamme. 1994. Nisin, a lantibiotic produced by Lactococcus lactis subsp. lactis : biosynthesis, fermentations and applications. In Bacteriocins of Lactic Acid Bacteria: Microbiology, Genetics and Applications, eds. L. De Vuyst and E.J. Vandamme, 151-221. Blackie Academic and Professional. London, England. Desjardins, P., J. Meghrous, and C. Lacroix. 2002. Effect of aeration and dilution rate on nisin Z production during continuous fermentation with free and immobilized Lactococcus lactis spp. lactis biovar. d iacetylactis UL719 in supplemented whey permeate. Int. Dairy J. 11: 943-951. Falahee, M. B., and M. R. Adams. 1992. Cross-reactivity of bacteriocins from lactic acid bacteria and lantibiotics in a nisin bioassay and ELISA. Lett. Appl. Microbiol. 15: 214-216. Falahee, M. B., M. R. Adams, J. W. Dale, and B. A. Morrs. 1990. An enzyme immunoassay for nisin. Int. J. Food Sci. Technol. 25: 590-595. Fang, T. J., and L. W. Lin. 1994. Growth of Listeria monocytogenes and Pseudomonas fragi on cooked pork in a modified atmosphere packaging/nisin combination. J. Food Prot. 57: 479-485. FAO/WHO Expert Committee on Food Additives. 1969. Specifications for identity and purity of some antibiotics. 12 th Report. WHO Technical Report Series, No. 430. Ferreira, M. A., and B. M. Lund. 1996. The effect of nisin on Listeria monocytogenes in culture medium and long-life cottage cheese. Lett. Appl. Microbiol. 22: 433-438. Fukuda, H. 1995. Immobilized microorganism bioreactors. In Bioreactor System Design , edited by Asenjo, J.A., and Merchuk, J.C., pp. 339-375, Marcel Dekker, New York.

175 Gillespy, T. G. 1957. Nisin trials. Fruit and Veg. Cann. Quick Freez. Res. Assoc., Leaflet No. 3, Chipping Campden, Glos, England. Gould, G. W. 1964. Effect of food preservatives on the growth of bacteria from spores. In: Microbial Inhibitors in Foods. Ed. M. Molin, pp. 17-24. Almqvist and Wiksell, Stockholm. Gross, E. 1977. α,β-unsaturated and related amino acids in peptides and proteins. In: Protein Cross-Linking. Ed. M. Friedmann, pp. 131-153. Plenum Press, New York. Gross, E., and J. L. Morell. 1971. The structure of nisin. J. Am. Chem. Soc. 93: 4634- 4635. Guerra, N. P. and L. Pastrana. 2001. Enhanced nisin and pediocin production on whey supplemented with different nitrogen sources. Biotechnol. Lett. 23: 609-612. Guerra, N. P. and L. Pastrana. 2003. Influence of pH drop on both nisin and pediocin production by Lactococcus lactis and Pediococcus acidilactici.Lett. Appl. Microbiol. 37: 51-55. Hasper, H. E., B. de kruijff, and E. Breukink. 2004. Assembly and stability of nisin-Lipid II pores. Biochem. 43: 11567-11575. Hawley, H. B. 1955. The development and use of nisin. J. Appl. Bacteriol. 18: 388-395. Hengstenberg, W. 1977. Enzymology of carbohydrate transport in bacteria. Curr. Top. Microbiol. Immunol. 77: 97-126. Henning, S., R. Metz, and W. P. Hammes. 1986. New aspects for the application of nisin to food products based on its mode of action. Int. J. Food Microbiol. 3: 135-141. Heukelelian, H. and Crosby, E. S. 1956. Slime formation in polluted waters. Sewage Industrial Wastes , 28: 78. Hirsch, A. 1950. The assay of the antibiotic nisin. J. Gen. Microbiol. 4: 70-74. Hirsch, A. 1951. Growth and nisin production of a strain of Streptococcus lactis . J. Gen. Microbiol. 5: 208-221. Hitchens, A. D., G. W. Gould, and A. Hurst. 1963. The swelling of bacterial spores during germination and outgrowth. J. Gen. Microbiol. 30: 445-453. Ho, K. L. G., A. L. Pometto III and P. N. Hinz. 1997a. Optimization of L-(+)-lactic acid production by ring and disc plastic composite supports through repeated-batch biofilm fermentation. Appl. Environ. Microbiol. 63: 2533-2542.

176 Ho, K. L. G., A. L. Pometto III, P. N. Hinz and A. Demirci. 1997b. Nutrient leaching and end product accumulation in plastic composite supports for L-(+)-lactic acid biofilm fermentation. Appl. Environ. Microbiol. 63: 2524-2532. Horner, T., H. Zahner, R. Kellner, and G. Jung. 1989. Fermentation and isolation of epidermin, a lanthionine containing polypeptide antibiotic from Staphylococcus epidermidis. Appl. Microbiol. Biotechnol. 30: 219-225. Howell, J.A., and B. Atkinson. 1976. Sloughing of microbial film in trickling filters. Water Res. 10: 307. Howell, T. H., J. P. Fiorellini, P. Blackburn, S. J. Projan, J. de la Harpe, and R. C. Williams. 1993. The effect of a mouthrise based on nisin, a bacteriocin, on developing plaque and gingivitis in beagle dogs. J. Clin. Periodontol. 20: 335-339. Hunter, G. J. E., and H. R. Whitehead. 1944. J. Dairy Res. 13: 123. Hurst, A. 1978. Nisin: its preservative effect and function in the growth cycle of the producer organism. In Streptococci , eds. F.A. Skinner and L.B. Quesnel, 297-314. Academic Press. London. Hurst, A., and G. J. Dring. 1968. The relation of the length of lag phase of growth to the synthesis of nisin and other basic proteins by Streptococcus lactis grown under different conditions. J. Gen. Microbiol. 50: 383-390. Hurst, A., and H. Kruse. 1972. Effect of secondary metabolites on the organisms producing them: Effect of nisin on Streptococcus lactis and enterotoxin B on Staphylococcus aureus. Antimicrobial Agents and Chemotherapy. 1: 277-279. Ingram, L. 1970. A ribosomal mechanism for synthesis of peptides related to nisin. Biochem. Biophys. Acta. 224: 263-365. James, L. S. and L. M. Larry. 1986. Partial characterization of the genetic basis for sucrose metabolism and nisin production in Streptococcus lactis . Appl. Environ. Microbiol. 51: 57-64. Janes, M. E., R. Nannapaneni, A. Proctor, and M. G. Johnson. 1998. Rice hull ash and silicic acid as adsorbents for concentration of bacteriocins. Appl. Environ. Microbiol. 64(11): 4403-4409.

177 Jarvis, B., J. Jeffcoat, and G. C. Cheeseman. 1968. Molecular weight distribution of nisin. Biochim. Biophys. Acta. 168: 153-155. Jung, G. 1991. Lantibiotics: a survey. In Nisin and Novel Lantibiotics, ed. G. Jung and H.-G. Sahl, 1-34. Leiden:ESCOM Science Publishers B.V. Kelly, N. A., B. G. Reuben, J. Rhoades, and S. Roller. 2000. Solvent extraction of bacteriocins from model solutions and fermentation broths. J. Chem. Technol. Biotechnol. 75: 777-784. Kim, W. S. 1997. Nisin production by Lactococcus lactis using two-phase batch culture. Lett. Appl. Microbiol. 25: 169-171. Kim, W. S., R. J. Hall and N. W. Dunn. 1997a. The effect of nisin concentration and nutrient depletion on nisin production of Lactococcus lactis. Appl. Microbiol. Biotechnol. 48: 449-453. Kim, W. S., R. J. Hall and N. W. Dunn. 1997b. Host specificity of nisin production by Lactococcus lactis. Biotechnol. Lett . 19: 1235-1238. Kim, W. S., R. J. Hall and N. W. Dunn. 1998. Improving nisin production by increasing nisin immunity/resistance genes in the producer organism Lactococcus lactis . Appl. Microbiol. Biotechnol. 50: 429-433. Kleerebezem, M. 2004. Quorum sensing control of lantibiotic production: nisin and subtilin autoregulate their own biosynthesis. Peptides 25(9): 1405-1414. Kozak, W., M. Rajchert-Trzpil, J. Zajdel and W. T. Dobrzanski. 1973. Lysogeny in lactic streptococci producing and not producing nisin. Appl. Microbiol. 25: 305-308. Kozlova, Y. I., T. I. Golikova, I. P. Baranova, and N.S. Egorov. 1979. Investigation of

the influence of K2HPO 4 on the growth of Streptococcus lactis and nisin synthesis at constant pH values of the medium. Mikrobiologiya 48: 443-336. Kuipers, O. P., M. M. Beerthuyzen, P. G. de Ruyter, E. J. Luesink, and W. M. de Vos. 1995. Autoregulation of nisin biosynthesis in Lactococcus lactis by signal- transduction. J. Biol. Chem. 270: 27299-27304. Kuipers, O. P., W. M. G. J. Yap, H. S. Rollema, M. M. Beerthuyzen, R. J. Siezen and W. M. De Vos. 1991. Expression of wild-type and mutant nisin genes in Lactococcus lactis. In Nisin and Novel Lantibiotics, eds. G. Jung and H. G. Sahl, 1-34. ESCOM Science Publishers B.V. Leiden.

178 Lazarova, V., and J. Manem. 2000. Innovative biofilm treatment technologies for water and wastewater treatment. In Biofilms II: Process Analysis and Applications , edited by Bryers, J.D., pp. 159-206, Wiley-Liss, Inc., New York. Lee, S. H., and H. U. Kim. 1985. Studies on lantibiotic nisin produced by Streptococcus lactis IFO 12007. Korean Journal of Animal Science . 27: 476-479. Leung, P. P., M. Khadre, T. H. Shellhammer, and A. E. Yousef. 2002. Immunoassay method for quantitative determination of nisin in solution and on polymeric films. Lett. Appl. Microbiol. 34: 199-204. Li, C., O. Y. Fan, and J. H. Bai. 2000. Extractive cultivation of Lactococcus lactis using a

polyethylene glycol/MgSO4 •7H 2O aqueous two-phase system to produce nisin. Biotech. Lett. 22(10): 843-847. Li, C., J. Bai, Z. Cai and F. Ouyang. 2002. Optimization of a culture medium for bacteriocin production by Lactococcus lactis using response surface methodology. J. Biotechnol. 93: 27-34. Lv, W., W. Cong, and Z. Cai. 2004. Nisin production by Lactococcus lactis subsp. lactis under nutritional limitation in fed-batch culture. Biotechnol. Lett. 26: 235-238. Lv, W., X. Zhang, and W. Cong. 2005. Modelling the production of nisin by Lactococcus lactis in fed-batch culture. Appl. Microbiol. Biotechnol. 68: 322-326. Massol-Deya, A. A., J. Whallon, R. F. Hickey, and J. M. Tiedje. 1995. Channel structures in aerobic biofilms of fixed-film reactors treating contaminated groundwater. Appl. Environ. Microbiol. 61: 79-777. Matsusaki, H., N. Endo, K. Sonomoto, and A. Ishizaki. 1996a. Development of purification method and identification of a peptide antibiotic produced by Lactococcus lactis IO-1. Food Sci. Technol. Int. 2: 157-162. Matsusaki, H., N. Endo, K. Sonomoto and A. Ishizaki. 1996b. Lantibiotic nisin Z fermentative production by Lactococcus lactis IO-1: relationship between production of the lantibiotic and lactate and cell growth. Appl. Microbiol. Biotechnol. 45: 36-40. Mattick, A. T. R., and A. Hirsch. 1947. Further observations on an inhibitory substance (nisin) from lactic streptococci. Lancet 2: 5-7.

179 Meghrous, J., E. Huot, M. Quittelier, and H. Petitdemange. 1992. Regulation of nisin biosynthesis by continuous cultures and by resting cells of Lactococcus lactis supsp. lactis. Res. Microbiol. 143: 897-890. Melo, L. F., and R. Oliveira. 2001. Biofilm reactors. In Multiphase Bioreactor Design , edited by J. M. S. Cabral, M. Mota, and J. Tramper, pp. 271-308, Taylor & Francis Inc, New York. Motlagh, A. M., M. C. Johnson, and B. Ray. 1991. Viability loss of foodborne pathogens by starter culture metabolites. J. Food Prot. , 54: 873-878,884. Mulders, J. W. M., I. J. Boerrigter, H. S. Rollema, R. J. Siezen, and W. M. De Vos. 1991. Identification and characterization of the lantibiotic nisin Z, a natural nisin variant. Eur. J. Biochem. 201: 581-584. Nilsson, L., Y. Chen, M. L. Chikindas, H. H. Huss, L. Gram, and T. J. Montville. 2000. Carbon dioxide and nisin act synergistically on Listeria monocytogenes. Appl. Environ. Microbiol. 66: 769-774. Norwood, D. E., and Gilmour, A. 2000. The growth and resistance to sodium hypochlorite of Listeria monocytogenes in a steady-state multispecies biofilm. J. Appl. Microbiol. 88(3): 512-520. Ogden, K. 1986. Nisin: a bacteriocin with a potential use in brewing. J. Inst. Brew. 92: 379-383. Oliveira, R. 1992. Physico-chemical aspects of adhesion. In Biofilms-Science and Technology, edited by L. F. Melo, T. R. Bott, M. Fletcher, and B. Capdeville. pp. 45-58. Dordrecht, Boston, London: Kluwer Academic Publishers. O’Keeffe, T., and C. Hill. 2000. Bacteriocins. In Encyclopedia of Food Microbiology , eds. C.A. Batt and P.D. Patel, 191-198. Academic Press. Parente, E., and A. Ricciardi. 1999. Production, recovery and purification of bacteriocins from lactic acid bacteria. Appl. Microbiol. Biotechnol. 52: 628-638. Pereira, M. A., M. M. Alves, J. Azeredo, M. Mota, and R. Oliveira. 2000. Physico- chemical properties of porous microcarriers in relation with the adhesion of an anaerobic consortium. J. Ind. Microbiol. Biotechnol. 24(3): 181-186. Pfeiffer, P., and C. Orben. 1997. Deutsche Lebensmittel-Rundschau 93: 47-49.

180 Pol, I. E., H. C. Mastwijk, P. V. Bartels, and E. J. Smid. 2000. Pulsed-electric field treatment enhances the bactericidal action of nisin against Bacillus cereus . Appl. Environ. Microbiol. 66: 428-430. Prioult, G., C. Turcotte, L. Labarre, C. Lacroix, and I. Fliss. 2000. Rapid purification of nisin Z using specific monoclonal antibody-coated magnetic beads. Int. Dairy J. 10: 627-633. Radler, F. 1990a. Possible us of nisin in winemaking. I. Action of nisin against lactic acid bacteria and wine yeasts in solid and liquid media. Am. J. Enol. Vitic. 41: 1-6. Radler, F. 1990b. Possible use of nisin in winemaking. II. Experiments to control lactic acid bacteria in the production of wine. Am. J. Enol. Vitic. 41: 7-11. Rauch, P. J. G., O. P. Kuipers, R. J. Siezen, and W. M. de Vos. 1994. Genetics and protein engineering of nisin. In: Bacteriocins of Lactic Acid Bacteria: Microbiology, Genetics and Applications , eds. L. de Vuyst and E. J. Vandamme, pp. 223-249. Blackie Academic & Professional, London. Rayman, K., B. Aris, and A. Hurst. 1981. Nisin: a possible alternative or adjunct to nitrite in the preservation of meats. Appl. Environ. Microbiol. 41: 375-380. Reunanen, J., and P. E. J. Saris. 2003. Microplate bioassay for nisin in foods, based on nisin-induced green fluorescent protein fluorescence. Appl. Environ. Microbiol. 69(7): 4214-4218. Rodriguez, E., J. Tomillo, M. Nunez, and M. Medina. 1997. Combined effect of bacteriocin-producing lactic acid bacteria and lactoperoxidase system activation on Listeria monocytogenes in refrigerated raw milk. J. Appl. Microbiol. 83: 389-395. Rogers, L. A. 1928. The inhibitory effect of Streptococcus lactis on Lactobacillus bulgaricus. J. Bacteriol. 16: 321-325. Rogers, L. A., and E. O. Whittier. 1928. Limiting factors in lactic fermentation. J. Bacteriol. 16: 211-229. Rogers, A. M. and T. J. Montville. 1991. Improved agar diffusion assay for nisin quantification. Food Biotechnol. 5(2): 161-168. Rossano, R., A. Del Fiore, A. D’Elia, G. Pesole, E. Parente, and P. Riccio. 1998. New procedure for the determination of nisin in milk. Biotechnol. Techniq. 12(10): 783- 786.

181 Schnell, N., K. D. Entian, U. Schneider, F. Gotz, H. Zahner, R. Kellner, and G. Jung. 1988. Prepeptide sequence of epidermin, a ribosomally synthesized antibiotic with four sulphide rings. Nature 333: 276-278. Sears, P. M., B. S. Smith, S. D. Rubino, E. Kulisek, S. Gusik, and P. Blackburn. 1991. Non-antibiotic approach to treatment of mastitis in the lactating dairy cow. J. Dairy Sci. 74: 203. Shimizu, H., T. Mizuguchi, E. Tanaka, and S. Shioya. 1999. Nisin production by a mixed-culture system consisting of Lactococcus lactis and Kluyveromyces marxianus. Appl. Environ. Microbiol. 65(7): 3134-3141. Sonomoto, K., N. Chinachoti, N. Endo, and A. Ishizaki. 2000. Biosynthetic production of nisin Z by immobilized Lactococcus lactis IO-1. Journal of Molecular Catalysis B. Enzymatic, 10: 325-334. Sousa, M., J. Azeredo, J. Feijo, and R. Oliveira. 1997. Polymeric supports for the adhesion of a consortium of autotrophic nitrifying bacteria. Biotechnol. Techniques , 11: 751-754. Stevens, K. A., B. W. Sheldon, N. A. Klapes, and T. R. Klaenhammer. 1991. Nisin treatment for inactivation of Salmonella species and other gram-negative bacteria. Appl. Environ. Microbiol. 57: 3613-3615. Suarez, A. M., J. M. Rodriguez, P. E. Hernandez, and J. I. Azcona-Olivera. 1996a. Generation of polyclonal antibodies against nisin: Immunisation strategies and immunoassay development. Appl. Environ. Microbiol. 62: 2117-2121. Suarez, A. M., J. M. Rodriguez, P. Morales, P. E. Hernandez, and J. I. Azcona-Olivera. 1996b. Development of monoclonal antibodies to the lantibiotic nisin A. J. Agric. Food Chem. 44: 2936-2940. Suarez, A. M., J. I. Azcona, J. M. Rodriquez, B. Sanz, and P. Hernandez. 1997. One-step purification of nisin A by immunoaffinity chromatography. Appl. Environ. Microbiol. 63: 4990-4992. Szabo, E. A., and M. E. Cahill. 1998. The combined affects of modified atmosphere, temperature, nisin and ALTA 2341 on the growth of Listeria monocytogenes. Int. J. Food Microbiol. 43: 21-31.

182 Taniguchi, M., K. Hoshino, H. Urasaki and M. Fuji. 1994. Continuous production of an antibiotic polypeptide (nisin) by Lactococcus lactis using a bioreactor coupled to a microfiltration module. J. Ferm. Bioeng. 77: 704-708. Taylor, S. L., E. B. Somers, and L.A. Krueger. 1984. Antibotulinal effectiveness of nisin- nitrite combinations in culture medium and chicken frankfurter emulsions. J. Food Prot. 48: 234-239. Teixeira, P., and R. Oliveira. 1999. Influence of surface characteristics on the adhesion of Alcaligenes denitrificans to polymeric substrates. J. of Adhesion Science and Technology , 13: 1287-1294. Terebiznik, M. R., R. J. Jagus, P. Cerutti, M. S. de Huergo, and A. M. Pilosof. 2000. Combined effect of nisin and pulsed electric fields on the inactivation of Escherichia coli. J. Food Prot. 63: 741-746. Thomas, L.V., M. R. Clarkson, and J. Delves-Broughton. 2000. Nisin. In: Natural Food Antimicrobial Systems , ed. A. S. Naidu, pp. 463-523, CRC Press, New York. Thompson, J., and B. M. Chassy. 1981. Uptake and metabolism of sucrose by Streptococcus lactis. J. Bacteriol. 147: 543-551. Thompson, J., N. Y. Nguyen, D. L. Sackett, and J. A. Donkersloot. 1991. Transposon- encoded sucrose metabolism in Lactococcus lactis. Purification of sucrose-6- phosphate hydrolase and genetic linkage to N5-(L-1-carboxyethyl)-L-ornithine synthase in strain K1. J. Biol. Chem. 266: 14573-14579. Tramer, J. 1966. Nisin in food preservation. Chem. Ind. , 11: 446-450. Tramer, J. and G. G. Fowler. 1964. Estimation of nisin in foods. J. Sci. Food Agric. 15: 522-528. Turcotte, C., C. Lacroix, E. Kheadr, L. Grignon, and I. Fliss. 2004. A rapid turbidometric microplate bioassay for accurate quantification of lactic acid bacteria bacteriocins. Int. J. Food Microbiol. 90: 283-293. U.S. Food and Drug Administration. 1988. Nisin Preparation: Affirmation of GRAS status as direct human food ingredient. Federal Register. 53, April 6. Valat, C., D. Champiat, T. T. T. N’Guyen, G. Loiseau, M. Raimbault, and D. Montet. 2003. Use of ATP bioluminescence to determine the bacterial sensitivity threshold to a bacteriocin. Luminescence 18: 254-258.

183 van de Ven, F. J. M., H. W. van den Hooven, R. N. H. Konings, and C. W. Hilbers. 1991. NMR studies of lantibiotics. The structure of nisin in aqueous solution. Eur. J. Biochem. 202: 1181-1188. Van’t Hul, J. S., and W. R. Gibbons. 1996. Concentration and recovery of the bacteriocin nisin from Lactococcus lactis subsp. lactis. Biotechnol. Appl. Biochem. 24: 251- 256. Van’t Hul, J. S., and W. R. Gibbons. 1997. Neutralization/recovery of lactic acid from Lactococcus lactis : effects on biomass, lactic acid, and nisin production. World J. Microbiol. Biotechnol. 13: 527-532. Wahlstrom, G., and P. E. J. Saris. 1999. A nisin bioassay based on bioluminescence. Appl. Environ. Microbiol. 65(8): 3742-3745. Wan, J, J. Gordon, M. W. Hickey, R. F. Mawson, and M. J. Coventry. 1996. Adsorption of bacteriocins by ingestible silica compounds. J. Appl. Bacteriol. 81: 167-173. Wan, J., M. W. Hickey, and M. J. Coventry. 1995. Continuous production of bacteriocins, brevicin, nisin and pediocin, using calcium alginate-immobilized bacteria. J. Appl. Bacteriol. 79: 671-676. Wandling, L. R., B. W. Sheldon, and P. M. Foegeding. 1999. Nisin in milk sensitizes Bacillus spores to heat and prevents recovery of survivors. J. Food Prot. 62: 492- 498. White, D. C. 1984. Chemical characterization of films. Life Sciences Research Report, 31: 159-176. Wiedemann, I., E. Breukink, C. van Kraaij, O. P. Kuipers, G. Bierbaum, B. de Kruijff, and H. G. Sahl. 2001. Specific binding of nisin to the peptidoglycan precursor Lipid II combines pore formation and inhibition of cell wall biosynthesis for potent antibiotic activity. J. Biol. Chem. 276(1): 1772-1779. Willaert, R. G., Baron, G. V., and de Backer, L. 1996. Immobilised Living Cell Systems: Modelling and Experimental Methods , Chapter 3, John Wiley and Sons, Chichester, England. Wolf, C. E. and W. R. Gibbons. 1996. Improved method for quantification of the bacteriocin nisin. J. Appl. Bacteriol. 80: 453-457.

184 Yang, R., M. C. Johnson and B. Ray. 1992. Novel method to extract large amount of bacteriocins from lactic acid bacterial. Appl. Environ. Microbiol. 58:3355-3359. Yu, P. L., N. W. Dunn, and W. S. Kim. 2002. Lactate removal by anionic-exchange resin improves nisin production by Lactococcus lactis. Biotechnol. Lett. 24: 59-64. Zendo, T., M. Fukao, K. Ueda, T. Higuchi, J. Nakayama, and K. Sonomoto. 2003. Identification of the lantibiotic, misin Q, a new natural nisin variant produced by Lactococcus lactis 61-14 isolated from a river in Japan. Biosci. Biotechnol. Biochem. 67(7): 1616-1619. Zhang, S. S., and A. Mustapha. 1999. Reduction of Listeria monocytogenes and Escherichia coli O157 : H7 numbers on vacuum-packaged fresh beef treated with nisin or nisin combined with EDTA. J. Food Protection 62(7): 1123-1127.

185 APPENDIX A1

PROPERTIES OF COMMERCIAL NISIN POWDER (Nisaplin) *

Table A1.1. Properties of commercial nisin powder (Nisaplin)

General specifications

Color Off white

Nisin content Minimum of 900 IU nisin/mg (WHO)

Moisture Not more than 3%

Shelf life 2 years from date of packing if stored below 25 °C in dry conditions away from direct sunlight

Microbiological Specifications

Total viable count <10 2

Yeasts and molds <10 2

Coliforms <10 2

Salmonellae Absent in 10 g

Coagulase-positive Staphylococci Absent in 10 g

Heavy metals

Arsenic Not more than 1 mg/kg

Lead Not more than 2 mg/kg

Copper Not more than 50 mg/kg

Zinc Not more than 50 mg/kg

Mercury Not more than 1 mg/kg

Heavy metals as lead Not more than 10 mg/kg

* Adapted from Delves-Broughton and Friis (1998)

186 APPENDIX A2

EFFECTS OF VARIOUS PROTEASE INHIBITORS

ON AN INACTIVATION OF NISIN

INTRODUCTION

The inability of LAB to synthesize many of the amino acids required for protein synthesis necessitates its effective proteolytic system. Several studies have shown that L. lactis contains a complex proteolytic system, involving several intracellular peptidases and cell surface-anchored proteases (Kunji et al., 1996). Although cell-free extracellular proteases were reported as a product of L. lactis in some studies, assessments of cell lysis or leakage were not applied and the reliability of the results was questioned (Law and

Kolstad, 1983). Nisin-inactivating protease (nisinase) was reported in several bacteria, including Staphylococcus aureus (Carlson and Bauer, 1957), Streptococcus thermophilus

(Alifax and Chevalier, 1962), and Bacillus cereus (Jarvis and Farr, 1971).

Drastic decrease of nisin activity during fermentation of Lactococcus lactis subsp. lactis has been reported in several producer strains, including NIZO 22186 (De Vuyst and

Vandamme, 1992), IFO 12007 (Taniguchi et al., 1994), IO-1 (Matsusaki et al., 1996), and LM0230 (Yu et al., 2002). Although enzymatic degradation by proteases has been suggested, it has not at all been systematically studied. In this study, various types of protease inhibitors were tested on the inactivation of nisin in the fermentation broth of L. lactis NIZO 22186.

187 MATERIALS AND METHODS

Microorganisms and medium

Lactococcus lactis subsp. lactis (NIZO 22186), nisin Z-producing strain, was used in this study. Nisin activity was quantified using Lactobacillus sakei (ATCC 15521) as the nisin-sensitive test organism. L. lactis and L. sakei were grown at 30 °C for 14 h in a complex medium (CM) and Lactobacillus MRS broth (Difco Laboratories, Detroit, MI), respectively. The CM consisted of 40 g of sucrose; 10 g of peptone (Amber Ferm 4015G,

Universal Flavors, Milwaukee, WI); 10 g of yeast extract (Ardamine Z, Sensient

Bionutrient, Indianapolis, IN); 10 g of KH 2PO 4; 2 g of NaCl; and 0.2 g of MgSO 4•7H 2O per liter of deionized water. The initial pH of CM medium was adjusted to 6.8 using 4 N

NaOH. For long-term storage, all stock cultures were maintained at -80 °C in 20% glycerol.

Preparation of cell-free fermentation broth

Nisin fermentations were carried out in a 1.25-l Bioflo II fermentor (New

Brunswick Scientific, Edison, NJ) at 30 °C with agitation at 100 rpm using a constant pH of 6.8 by adding 4N NaOH. The biofilm reactor was set up as described previously

(Chapter 4). After 6 h of fermentation, fermentation broth was harvested and centrifuged at 3,800 x g, 4 °C for 20 min (Sorvall Super T 21, ST-H750, Kendro Lab Products,

Newton, CT) and the supernatant was collected and filtered using 0.22 micron filters (25 mm Acrodisc ® syringe filter with HT Tuffryn, Pall, NY).

188 Study on nisin inactivation

The prepared cell-free broth containing active proteases was used in order to study the rate of inactivation over time. A portion of cell-free fermentation broth was heated at 90 °C for 10 min and cooled down to 25 °C in order to eliminate the activity of protease before being used as a control. The heat-treated or non-heat-treated fermentation broth (20 ml) was placed in a sterile polyethylene bottle (volume 250 ml, diameter 61.2 mm, NALGENE, OH) and incubated at 30 °C. Tween 20 (0.1% v/v) was added to prevent a non-specific adsorption of nisin onto the container’s surfaces. Samples (1 ml) were taken after 0 (immediately after heating and cooling for control), 2, 4, 16, and 24 h of incubation and analyzed for nisin activity (IU/ml) by an agar well diffusion bioassay, using L. sakei and pre-diffusion at 4 °C, as described in Chapter 3. The experiment was duplicated and the average values were used for comparison.

Study on effects of various protease inhibitors

Six protease inhibitors with different specificity (Table A2.1) were chosen for this study: 1) M222 (Pro-Pure TM Protease Inhibitor Cocktail with EDTA, Amresco, Solon,

OH), 2) Complete TM , Mini (Roche, Nutley, NJ), 3) Complete TM , Mini EDTA-free

(Roche), 4) Benzamidine (Amresco), 5) Bestatin (Amresco), and 6) 1,10-Phenanthroline

(Amresco). All protease inhibitors were prepared (as stock solution) and diluted using their specific diluents (Table A2.2). For each protease inhibitor, an amount of stock solution added to the cell-free broth was based on the recommended working concentration from manufacturer and the total volume of 1 ml. For a control, equal amount of diluent was added instead of the protease inhibitor. Tween 20 (0.1% v/v) was

189 added to the sample in order to eliminate a non-specific adsorption of nisin onto the container’s surfaces. All samples were incubated at 30 °C. An enzymatic reaction was stopped at a specific time (0, 3, and 6 h) by heating the sample at 90 °C for 10 min. The sample was then analyzed for nisin activity (IU/ml) by an agar well diffusion bioassay, using L. sakei and pre-diffusion at 4 °C, as described in Chapter 3. Each treatment was duplicated and the average values were used for comparison.

Table A2.1. Specificity of protease inhibitor tested in this study.

Protease Components Specificity Inhibitor AEBSF Serine proteases (chymotrypsin, trypsin, kallikrein, plasmin, thrombin)

Aprotinin Serine proteases (trypsin, plasmin, thrombin, factor X a) M222 Bestatin Amino peptidases Cocktail E64 Cysteine proteases Leupeptin Serine and Cysteine proteases (trypsin, papain, plasmin, cathepsin B) Complete TM , Mini* Serine, Cysteine, and Metallo-proteases Complete TM , Mini* Serine and Cysteine, but not Metallo-proteases EDTA-free Benzamidine Trypsin-like serine proteases Bestatin Amino peptidases 1,10-Phenanthroline Metallo-proteinases and metal activated proteinases * Components of Complete TM , Mini are trade secret.

Table A2.2. Recommended concentrations for protease inhibitors tested in this study.

Protease Inhibitor Diluent Stock Concentration Working Concentration M222 Cocktail DI water 1 ml of cocktail/ 100 ml of sample Complete TM , Mini DI water 1 tablet of cocktail/ 50 ml of sample Complete TM , Mini DI water 1 tablet of cocktail/ 50 ml of sample EDTA-free Benzamidine DI water 10 mM 1 mM Bestatin Methanol 2-5 mg/ml 40 µg/ml 1,10-Phenanthroline DMSO 200 mM 10 mM

190 RESULTS AND DISCUSSIONS

Study on nisin inactivation

In this study, inactivation of nisin was clearly observed in the non-heated, cell- free fermentation broth removed from the biofilm reactor, but not in the control with heat treatment (Fig. A2.1). An initial nisin activity (at 0 h) of control was significantly lower than that of non-heated sample as a result of the heat treatment, which was applied (at pH

6.8) in order to eliminate the activity of protease. This result indicated a presence of nisin-inactivating enzyme in the cell-free fermentation broth unless it is inactivated by heating. However, since assessment of cell lysis or leakage has not been applied, the location of the enzyme (extracellular, cell-bound, or intracellular) was still inconclusive.

1400 Control 1200 Non-heated 1000

800

600 Nisin (IU/ml) Nisin 400

200

0 0 2 4 16 24 Time (h)

Figure A2.1. Inactivation of nisin in control (heated) and non-heated cell-free fermentation broth

191 Study on effects of various protease inhibitors

In order to eliminate influences of different recommended diluents and working concentrations between tested protease inhibitors, different control has been used for each protease inhibitor. For a better comparison, the value of nisin activity has been converted into a percent relative nisin activity based on the activity at 0 h. Among tested protease inhibitors, only Complete TM , Mini and 1,10-Phenanthroline lessened the inactivation of nisin (Fig. A2.2). Addition of Complete TM , Mini (21% inactivation of nisin after 6 h) and

1,10-Phenanthroline (25% inactivation of nisin after 6 h) remarkably reduced the inactivation of nisin when compared to other protease inhibitors tested (52 – 63% inactivation of nisin after 6 h).

120 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100 93% 86%

79% 73% 80 69% 75% 70% 61% 63% 61% 62% 60 48% 56% 54% 41% 47% 39% 40% 42% 37% 40 Relative Nisin ActivityRelative (%) Nisin 20

0 C PI C PI PI C PI C PI C PI

W222 Complete Complete Benzamidine Bestatin Phenanthroline EDTA- free Protease Inhibitors

0 Hour 3 Hour 6 Hour

Figure A2.2. Effects of various protease inhibitors on inactivation of nisin in fermentation broth (C: Control; PI: Sample with protease inhibitor).

192 Considering that only Complete TM , Mini and 1,10-Phenanthroline - the only two inhibitors that can inhibit metallo-proteases among tested inhibitors - can inhibit the enzyme while the other inhibitors failed to do so, the nisin-inactivating enzyme detected in this study was likely to be a metallo-protease. Decrease in nisin activity observed in the sample with Complete TM , Mini and 1,10-Phenanthroline might be a result of incomplete inhibition.

CONCLUSIONS

Although enzymatic degradation of nisin in several nisin producers has long been observed, it has never been systematically studied. Preliminary results obtained in this study confirmed the presence of nisin-inactivating enzyme in the cell-free fermentation broth of L. lactis NIZO 22186. Although the exact location of the enzyme was still inconclusive, its activity could be effectively eliminated by a heat treatment at 90 °C for

10 min. The inhibition profiles obtained from various protease inhibitors indicated that the enzyme of interest is likely to be a metallo-protease.

In order to further identify and characterize the nisin-inactivating enzyme, the enzyme should be purified and the properties (e.g. molecular weight, optimum pH and temperature) should be studied. The effects of metal chelator, such as EDTA and 1,10- phenanthroline, on growth and nisin production of the producer should be studied further in case the chelator is going to be applied during fermentation. It is expected that the information on nisin-inactivating enzyme will provide an alternative to enhance nisin production further.

193 REFERENCES

Alifax, P. R., and R. Chevalier. 1962. Etude de la nisinase produite par Streptococcus thermophilus. J. Dairy Res. 29: 233-240. Carlson, S., and H. M. Bauer. 1957. Arch. Hug., Berl. 141: 445. De Vuyst, L. and E. J. Vandamme. 1992. Influence of the carbon source on nisin production in Lactococcus lactis supsp. lactis batch fermentations. J. Gen. Microbiol. 138: 571-578. Jarvis, B., and J. Farr. 1971. Partial purification, specificity and mechanism of action of the nisin-inactivating enzyme from Bacillus cereus. Biochim. Biophys. Acta 227: 232-240. Kunji, E. R. S., I. Mierau, A. Hagting, B. Poolman, and W. N. Konings. 1996. The proteolytic systems of lactic acid bacteria. Antonie van Leeuwenhoek 70: 187-221. Law, B. A., and J. Kolstad. 1983. Proteolytic systems in lactic acid bacteria. Antonie van Leeuwenhoek 49: 225-245. Matsusaki, H., N. Endo, K. Sonomoto and A. Ishizaki. 1996. Lantibiotic nisin Z fermentative production by Lactococcus lactis IO-1: relationship between production of the lantibiotic and lactate and cell growth. Appl. Microbiol. Biotechnol. 45: 36-40. Taniguchi, M., K. Hoshino, H. Urasaki and M. Fuji. 1994. Continuous production of an antibiotic polypeptide (nisin) by Lactococcus lactis using a bioreactor coupled to a microfiltration module. J. Ferm. Bioeng. 77: 704-708. Yu, P. L., N. W. Dunn, and W. S. Kim. 2002. Lactate removal by anionic-exchange resin improves nisin production by Lactococcus lactis. Biotechnol. Lett. 24: 59-64.

194 VITA

THUNYARAT PONGTHARANGKUL

Education Doctor of Philosophy, May 2006. Department of Agricultural and Biological Engineering, The Pennsylvania State University, University Park, PA. Master of Science, May 2001. Biotechnology, Chulalongkorn University, Bangkok, Thailand. Bachelor of Science, May 1997. Food Science and Technology, Chulalongkorn University, Bangkok, Thailand. Work Experiences • Teaching Assistant, January – December, 2004-2005. Department of Agricultural and Biological Engineering, The Pennsylvania State University, University Park, PA. • Summer Trainee, April 1996. Surathip Wiangping, Chiang Mai, Thailand. Hornors and Awards • First Place in Graduate Student Competition (Physical Sciences and Engineering), College of Agricultural Sciences, Gamma Sigma Delta, Annual Research Expo, The Pennsylvania State University, March, 2006. • First Place in Graduate Student Competition, Northeast Agricultural and Biological Engineering Conference (NABEC), August, 2005. • First Place in Graduate Poster Presentation, Allegheny Branch American Society for Microbiology (ABASM), November, 2004. Publications Pongtharangkul, T. and A. Demirci. 2004. Evaluation of agar well diffusion method for nisin quantification. Journal of Applied Microbiology and Biotechnology, 65:268-272. Demirci, A., T. Pongtharangkul, and A. L. Pometto III. 2004. Applications of biofilm reactors for production of value_added products by microbial fermentation. In Biofilms in the Food Environment. H. P. Blaschek, H. Wang, and M. Agle, Eds. Blackwell Publishing, Ames, Iowa. Pongtharangkul, T. and A. Demirci. 2005. Effects of pH profiles on nisin production in biofilm reactor. Journal of Applied Microbiology and Biotechnology (In press). Pongtharangkul, T. and A. Demirci. 2006a. Evaluation of medium for nisin production in repeated-batch biofilm reactor. Biotechnology Progress, 22(1): 217-224. Pongtharangkul, T. and A. Demirci. 2006b. Effects of fed-batch fermentation and pH profiles on nisin production in suspended-cell and biofilm reactors. Journal of Applied Microbiology and Biotechnology (In press). Pongtharangkul, T. , A. Demirci, and V. M. Puri. 2006c. Modeling of nisin production by Lactococcus lactis during batch fermentation. Biotechnology Progress (In preparation).