Forming Bacteria in Foods Ανάπτυξη Και Εφαρμογή Μαθηματικών Μοντέλων Πρόβλεψης Της Συμπεριφοράς Θερμόφιλων Σποριογόνων Βακτηρίων Στα Τρόφιμα

ARISTOTLE UNIVERSITY OF THESSALONIKI FACULTY OF AGRICULTURE, FORESTRY AND NATURAL ENVIRONMENT SCHOOL OF AGRICULTURE ΑΡΙΣΤΟΤΕΛΕΙΟ ΠΑΝΕΠΙΣΤΗΜΙΟ ΘΕΣΣΑΛΟΝΙΚΗΣ ΣΧΟΛΗ ΓΕΩΠΟΝΙΑΣ, ΔΑΣΟΛΟΓΙΑΣ ΚΑΙ ΦΥΣΙΚΟΥ ΠΕΡΙΒΑΛΛΟΝΤΟΣ ΤΜΗΜΑ ΓΕΩΠΟΝΙΑΣ

MYRSINI N. KAKAGIANNI Biologist, MSc

ΜΥΡΣΙΝΗ Ν. ΚΑΚΑΓΙΑΝΝΗ Πτυχιούχος Βιολόγος, MSc

Development and application of predictive models for the growth of thermophilic spore - forming bacteria in foods

Ανάπτυξη και εφαρμογή μαθηματικών μοντέλων πρόβλεψης της συμπεριφοράς θερμόφιλων σποριογόνων βακτηρίων στα τρόφιμα

PhD THESIS ΔΙΔΑΚΤΟΡΙΚΗ ΔΙΑΤΡΙΒΗ

THESSALONIKI / 2018 ΘΕΣΣΑΛΟΝΙΚΗ / 2018

MYRSINI N. KAKAGIANNI Biologist, MSc

Development and application of predictive models for the growth of thermophilic spore - forming bacteria in foods

Ph.D. THESIS

Submitted at School of Agriculture of Aristotle University of Thessaloniki

DATE OF ORAL PRESENTATION: Friday 13 July 2018

THESIS COMMITTEE (Decision ΓΣΕΣ 122/01-06-2018):

FULL NAME LEVEL ROLE INSTITUTION Konstantinos P. Aristotle University of Professor Supervisor Koutsoumanis Thessaloniki Member of Three – Aristotle University of Konstantinos Biliaderis Professor member Committee Thessaloniki Assistant Member of Three – Aristotle University of Eugenios Katsanidis Professor member Committee Thessaloniki Agricultural University of George – John Nychas Professor Member Athens Associate Aristotle University of Apostolos S. Angelidis Member Professor Thessaloniki Associate Agricultural University of Efstathios Panagou Member Professor Athens Assistant Aristotle University of Athanasia Goula Member Professor Thessaloniki

ΜΥΡΣΙΝΗ Ν. ΚΑΚΑΓΙΑΝΝΗ Πτυχιούχος Βιολόγος, MSc

ΔΙΔΑΚΤΟΡΙΚΗ ΔΙΑΤΡΙΒΗ

Υποβλήθηκε στο Τμήμα Γεωπονίας του Αριστοτελείου Πανεπιστημίου Θεσσαλονίκης

ΗΜΕΡΟΜΗΝΙΑ ΥΠΟΣΤΗΡΙΞΗΣ: Παρασκευή 13 Ιουλίου 2018

ΕΞΕΤΑΣΤΙΚΗ ΕΠΙΤΡΟΠΗ (Απόφαση ΓΣΕΣ 122/01-06-2018):

ΟΝΟΜΑΤΕΠΩΝΥΜΟ ΒΑΘΜΙΔΑ ΙΔΙΟΤΗΤΑ ΙΔΡΥΜΑ Κωνσταντίνος Π. Αριστοτέλειο Πανεπιστήμιο Καθηγητής Επιβλέπων Κουτσουμανής Θεσσαλονίκης Μέλος Κωνσταντίνος Αριστοτέλειο Πανεπιστήμιο Καθηγητής Τριμελούς Μπιλιαδέρης Θεσσαλονίκης Επιτροπής Μέλος Επίκουρος Αριστοτέλειο Πανεπιστήμιο Ευγένιος Κατσανίδης Τριμελούς Καθηγητής Θεσσαλονίκης Επιτροπής Γεώργιος – Ιωάννης Γεωπονικό Πανεπιστήμιο Καθηγητής Μέλος Νυχάς Αθηνών Αναπληρωτής Αριστοτέλειο Πανεπιστήμιο Απόστολος Σ. Αγγελίδης Μέλος Καθηγητής Θεσσαλονίκης Αναπληρωτής Γεωπονικό Πανεπιστήμιο Ευστάθιος Πανάγου Μέλος Καθηγητής Αθηνών Επίκουρος Αριστοτέλειο Πανεπιστήμιο Αθανασία Γούλα Μέλος Καθηγήτρια Θεσσαλονίκης

© Myrsini N. Kakagianni, 2018 © PhD thesis, Aristotle University of Thessaloniki, Thessaloniki, Greece, 2018 Development and application of predictive models for the growth of thermophilic spore - forming bacteria in foods Ανάπτυξη και εφαρμογή μαθηματικών μοντέλων πρόβλεψης της συμπεριφοράς θερμόφιλων σποριογόνων βακτηρίων στα τρόφιμα

ISBN

«Η έγκριση της παρούσης διδακτορικής διατριβής από το Τμήμα Γεωπονίας δεν υποδηλώνει αποδοχή των γνωμών και απόψεων του συγγραφέως» (Ν. 5343/1932, άρθρο 202, παρ. 2)

Μέρος της παρούσας διατριβής υλοποιήθηκε στο πλαίσιο της Δράσης «Συνεργασία» ΕΣΠΑ 2007 - 2013 και του προγράμματος Erasmus+ και συγχρηματοδοτήθηκε από: -Ευρωπαϊκή ΄Ενωση - Ευρωπαϊκό Κοινωνικό Ταμείο -Γενική Γραμματεία Έρευνας και Τεχνολογίας

στον Αποστόλη και το Νικόλα

Ευχαριστίες

Αρχικά, θα ήθελα να εκφράσω τη βαθύτατη ευγνωμοσύνη μου στον κ. Κωνσταντίνο Κουτσουμανή για την πολύτιμη καθοδήγηση και τη συνεχή στήριξή του κατά τη διάρκεια όλων αυτών των ετών. Τον ευχαριστώ θερμά για το χρόνο, τις ιδέες και την εμπιστοσύνη που έδειξε στο πρόσωπό μου δίνοντάς μου την ευκαιρία να συνεχίσω τις πανεπιστημιακές μου σπουδές με την εκπόνηση της διδακτορικής αυτής μελέτης και να εξελιχθώ ως ερευνήτρια. Τον ευχαριστώ, επίσης, που μου έμαθε τη σημασία της συνεργασίας των επιστημών αλλά και τη μετάδοση της γνώσης από γενιά σε γενιά προκειμένου να είμαστε σε θέση να προβούμε σε αξιόλογα αποτελέσματα. Θεωρώ τον εαυτό μου τυχερό που ήμουν ένα από τα μέλη της ερευνητικής του ομάδας καθώς μου μετέδωσε πολλές και πλούσιες γνώσεις στο χώρο της Ποσοτικής Μικροβιολογίας.

Επίσης, θα ήθελα να ευχαριστήσω όλα τα μέλη της εξεταστικής μου επιτροπής, τους καθηγητές κ. Κωνσταντίνο Μπιλιαδέρη, κ. Ευγένιο Κατσανίδη, κ. Γεώργιο – Ιωάννη Νυχά, κ. Ευστάθιο Πανάγου, κ. Απόστολο Αγγελίδη και κα. Αθανασία Γούλα, που αξιολόγησαν αυτή τη διατριβή και ενήργησαν ως «αντίπαλοι» κατά την υπεράσπισή της.

Θα ήθελα να εκφράσω τις θερμές μου ευχαριστίες στο Βασίλη Βαλδραμίδη, Αν. Καθηγητή του Πανεπιστημίου της Μάλτας, για την εμπιστοσύνη του και την ευκαιρία που μου έδωσε να υλοποιήσω ένα μικρό μέρος της διδακτορικής αυτής μελέτης, υπό την επίβλεψή του, στο εργαστήριό του. Εκτιμώ βαθιά την εξαίρετη συνεργασία μας, την πολύτιμη καθοδήγησή του και την ανιδιοτελή μετάδοση των γνώσεών του στο αντικείμενο της ποσοτικής μικροβιολογίας.

Ιδιαίτερες ευχαριστίες θα ήθελα να εκφράσω στο δάσκαλό μου κ. Μηνά Γιάγκου για την εμπιστοσύνη που έδειξε στο πρόσωπό μου και την ευκαιρία που μου έδωσε να γνωρίσω τον κόσμο της Ανοσοβιολογίας. Έπαιξε καθοριστικό ρόλο στη ζωή μου καθώς ήταν ο άνθρωπος που μετά από συζητήσεις με παρότρυνε να συνεχίσω τις μεταπτυχιακές μου σπουδές εξερευνώντας το χώρο των τροφίμων.

Ένα μεγάλο ευχαριστώ οφείλω στους συναδέλφους, συνεργάτες και συμφοιτητές, συμπεριλαμβανομένων και των προπτυχιακών φοιτητών (Φωτεινή Μπόμπολα, Φανή Τσίνα, Αγγελική Χωραΐτη, Χριστίνα Νταμπούδη, Όλια Χαρισμιάδου, Μαρία Ρέπα, Μαίρη Τσιλφίδου και Μαρία Παπακρασά), για τη

βοήθεια και ουσιαστική συμβολή τους καθ’ όλη τη διάρκεια της διδακτορικής διατριβής. Την ιδιαίτερη ευγνωμοσύνη και αγάπη μου θα ήθελα να εκφράσω στη Μαίρη Γουγουλή, για την άριστη συνεργασία και την αμέριστη βοήθεια οποτεδήποτε της ζήτησα, και τη Δάφνη Δημακοπούλου – Παπάζογλου, για τη σημαντική συμβολή της σε παρατηρήσεις και διορθώσεις της παρούσας διδακτορικής διατριβής, τη φιλία και συμπαράστασή της όλα αυτά τα χρόνια, ενώ παράλληλα μαζί με την Αλεξία Λιανού, Ζάφη Ασπρίδου, και Ζαχαρένια Ζαμπουλάκη ζήσαμε αξέχαστες στιγμές, όμορφες και δύσκολες, δημιουργώντας ένα ιδανικό κλίμα συνεργασίας, δημιουργικότητας και σκληρής δουλειάς. Είστε πλέον αναπόσπαστο κομμάτι της καρδιάς μου.

Ιδιαιτέρως θα ήθελα να ευχαριστήσω τους φίλους μου Δημήτρη Βασιλείου για την καθοριστική βοήθειά του όσον αφορά τη συλλογή των θερμοκρασιακών δεδομένων καθώς και την Ελένη Νάστα για την επιμέλεια του εξωφύλλου της συγκεκριμένης διατριβής.

Τέλος, ένα μεγάλο ευχαριστώ θα ήθελα να εκφράσω στον Αποστόλη μου για τις διορθώσεις του στην παρούσα διδακτορική διατριβή, την αγάπη, την υποστήριξη και την υπομονή του όλα αυτά τα χρόνια, ενώ πάρα πολλές φορές με βοήθησε να επαναπροσδιορίσω τις προτεραιότητές μου. Είμαι παντοτινά ευγνώμων στους γονείς μου, Μαίρη και Νίκο, αλλά και την αδερφή μου, Βίκυ, για την αγάπη τους, τη συνεχή και ανεξάντλητη συμπαράσταση και υποστήριξη σε κάθε μου απόφαση όλα τα χρόνια των σπουδών μου.

Μυρσίνη

Contents

List of Tables 13

List of Figures 15

List of Abbreviations 19

Abstract 25

Περίληψη 29

Chapter 1 Literature Review and Thesis Outline 33

Chapter 2 Development and application of Geobacillus 97 stearothermophilus growth model for predicting spoilage of evaporated milk

Chapter 3 Development and validation of predictive models for the 117 effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

Chapter 4 Effect of storage temperature on the lag time of Geobacillus 141 stearothermophilus individual spores

Chapter 5 Mapping the risk of evaporated milk spoilage in the 159 Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

Chapter 6 General Discussion and potential applications of the 174 developed models to the food industry

References 186

List of Publications 233

List of Tables

Table 2.1. Estimated values and statistics for the parameters of the cardinal 106 model with inflection (Equation (2.1)) describing the effect of temperature on the maximum specific growth rate (μmax) of Geobacillus stearothermophilus ATCC 7953 in tryptone soy broth.

Table 2.2. Comparison between observed and predicted spoilage time of the 115 evaporated milk, stored under nonisothermal conditions, by Geobacillus stearothermophilus ATCC 7953.

Table 3.1. Estimated values and fitting statistics for the parameters of the 129 logistic polynomial regression model for the combined temperature and pH limits of Alicyclobacillus acidoterrestris growth in K broth.

Table 3.2. Probability of Alicyclobacillus acidoterrestris growth predicted by the 129 growth/no growth boundaries model for different combinations of temperature and pH.

Table 3.3. Effect of pH on the maximum specific growth rate (μmax) of 131 Alicyclobacillus acidoterrestris in K broth at 48 °C. Points represent the observed μmax values, the solid line corresponds to the fitting on the cardinal pH model to the data, and the dotted lines depict the 95% confidence and prediction intervals.

Table 3.4. Comparison between observed behavior of Alicyclobacillus 133 acidoterrestris and probability of growth predicted by the developed growth/no growth interface model in commercial pasteurized fruit drinks tested in validation studies

Table 4.1. Statistics of the individual spore lag time (λ) of Geobacillus 149 stearothermophilus.

Table 4.2. Parameter estimation of the Gamma distribution fitted to the 149 individual spore lag times of Geobacillus stearothermophilus. The probability density function of the Gamma distribution in the shape-rate parametrization is: for x>0 and α, β>0.

Table 4.3. Estimated values and statistics for the parameters of the Cardinal 151 Model with Inflection (Equations (4.3) and (4.4)) describing the effect of temperature on the reciprocal of the mean lag time (λ) of individual spores of Geobacillus stearothermophilus.

List of Figures

Figure 1.1. Schematic structure of a spore of Bacillus or Clostridium species. 37 Some species contain an additional layer, known as an exosporium (adapted from Setlow, 2012).

Figure 1.2. Stages of sporulation. 42

Figure 1.3. Schematic outline of germination of spores of Bacillus species 43 (adapted from Setlow, 2003).

Figure 1.4. Overview of the research themes addressed in this thesis 95 chapters.

Figure 2.1. Effect of temperature on the maximum specific growth rate (μmax) 106 of Geobacillus stearothermophilus ATCC 7953 in tryptone soy broth, fitted in the Cardinal Model with Inflection (solid line) (Equation (2.1)). Points represent observed values of the μmax. The dotted and the discontinuous lines depict the 95% confidence and the prediction limits, respectively, of the effect of temperature on the maximum specific growth rate.

Figure 2.2. Growth kinetics of Geobacillus stearothermophilus ATCC 7953 108 cells, derived from spores, ( ) in evaporated milk and pH evolution ( ) during storage at optimum growth temperature (62 °C). The black solid line ( ) depicts the fitting of the Baranyi and Roberts model (Equation (2.3)) to the growth data. The point ( ) is showing the observed time of evaporated milk coagulation. Each point is a mean of eight values. Vertical and horizontal bars indicate the standard deviation.

Figure 2.3. Comparison between observed (points) and predicted (solid line) 111 growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature condition 1. Discontinuous lines indicate milk pH (- - -) and temperature changes (-----).

Figure 2.4. Comparison between observed (points) and predicted (solid line) 111 growth of Geobacillus stearothermophilus ATCC 7953 in evaporated milk stored under periodically changing temperature condition 2. Discontinuous lines indicate milk pH (- - -) and temperature changes (-----).

Figure 2.5. Comparison between observed (points) and predicted (solid line) 112 growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature (24 h at 37 °C, 12 h at 42 °C and 24 h at 45 °C). Discontinuous lines indicate milk pH (- - -) and temperature changes (-----).

Figure 2.6. Comparison between observed (points) and predicted (solid line) 112 growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature (78 h at 20 °C, 59 h at 25 °C and 163 h at 40 °C). Discontinuous lines indicate milk pH (- - -) and temperature changes (-----).

Figure 2.7. Comparison between observed (points) and predicted (solid line) 113 growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature (6 h at 50 °C, 12 h at 30 °C and 24 h at 42 °C). Discontinuous lines indicate milk pH (- - -) and temperature changes (-----).

Figure 3.1. Growth/no growth interface of Alicyclobacillus acidoterrestris in K 128 broth with respect to temperature and pH predicted by the model (lines) compared to the data used to generate the model (points). Black symbols: growth in all replicates; white symbols: no growth in all replicates; grey symbols: growth in some replicates; lower line: predicted boundary P=0.1; middle line: predicted boundary P=0.5; upper line: predicted boundary P=0.9.

Figure 3.2. Effect of temperature on the maximum specific growth rate (μmax) 131 of Alicyclobacillus acidoterrestris in K broth of pH 4.5. Points represent the observed μmax values, the solid line corresponds to the fitting of the cardinal temperature model with inflection to the data, and the dotted lines depict the 95% confidence and prediction intervals.

Figure 3.3. Effect of pH on the maximum specific growth rate (μmax) of 132 Alicyclobacillus acidoterrestris in K broth at 48 °C. Points represent the observed μmax values, the solid line corresponds to the fitting of the cardinal pH model to the data, and the dotted lines depict the 95% confidence and prediction intervals.

Figure 3.4. Comparison between predicted growth boundaries (lines) and 135 observed behavior (points) of Alicyclobacillus acidoterrestris in fruit drinks. Black symbols: growth; white symbols: no growth lower line: predicted

boundary P=0.1; middle line: predicted boundary P=0.5; upper line: predicted boundary P=0.9.

Figure 3.5. Comparison between observed (points) and predicted (solid line) 138 growth of Alicyclobacillus acidoterrestris ATCC 49025 in: (a) K broth (pH=4.5) at 35 °C, (b) apple - orange - carrot drink (pH=3.52) at 30 °C, (c) peach drink (pH=3.34) at 35 °C, (d) 9 fruits & 10 vitamins drink (pH=3.71) stored at 5 °C for 2 days followed by storage at 30 °C, (e) apple - orange - carrot drink (pH=3.52) stored at 5 °C for 5 days followed by storage at 30 °C, (f) apple - orange - carrot drink (pH=3.52) stored at dynamic temperature conditions (6 h at 25 °C, 12 h at 35 °C and 6 h at 45 °C), (g) apple - orange - carrot drink (pH=3.52) stored at room temperature during the summer period in Greece, and (h) peach drink (pH=3.34) stored at stored at room temperature during summer period in Greece. Discontinuous lines indicate medium temperature during storage. For growth prediction the parameters ymax (maximum population density) and h0 (physiological state parameter) of the primary model were fixed to 106.2 CFU/ml and 4.0, respectively.

Figure 4.1. OD growth curves during growth of 199 single spores of 147 Geobacillus stearothermophilus in tryptone soy broth at 45 °C.

Figure 4.2. Cumulative distributions of Geobacillus stearothermophilus single 148 spore lag times (λ) at different growth temperature conditions. The number of data was 224, 199, 227, 199 and 193 for 45, 47.5, 50, 55 and 59 °C, respectively.

Figure 4.3. Comparison between observed and fitted quantiles of the Gamma 150 distribution for Geobacillus stearothermophilus single spore lag times (λ) at 45 °C (a), 47.5 °C (b), 50 °C (c), 55 °C (d) and 59 °C (e).

Figure 4.4. Effect of temperature on the reciprocal of the mean lag times of 152 Geobacillus stearothermophilus single spores fitted to the Cardinal Model with Inflection (solid line) (Equations (4.3) and (4.4)). Points ( ) represent observed values.

Figure 4.5. Simulations of Geobacillus stearothermophilus growth at 45 °C for 154 various initial contamination levels: 1 spore (a), 10 spores (b) and 100 spores (c). Growth is predicted by the stochastic model (Equation (4.5)) using Monte Carlo simulation with 10,000 iterations and a Uniform distribution for time t [t ~ Uniform (0, 80)].

Figure 5.1. Representative examples of historical hourly temperature data for 167 Tunis for the years 2012 (a), 2013 (b), 2014 (c), 2015 (d) and 2016 (e) obtained from the Weather Underground database. Time = 0 corresponds to January 1st.

Figure 5.2. Representative examples of Geobacillus stearothermophilus 169 growth prediction (red lines) in evaporated milk with a shelf life of one year in the supply chain of 3 Mediterranean capitals based on the historical hourly temperature data (blue lines) obtained from the Weather Underground database. a: Athens, (data from 2012); b: Tunis, (data from 2014); c: Damascus, (data from 2015)

Figure 5.3. Box - plot representations of the total predicted growth of 170 Geobacillus stearothermophilus in evaporated milk with a shelf life of one year for the supply chain of 23 Mediterranean capitals for five years (2012 - 2016). The bottom and top of the box are the 25th and 75th percentiles (Q1 and Q3, respectively), the band is the median and the dots correspond to the minimum and maximum value.

Figure 5.4. Geographical risk assessment for evaporated milk spoilage in the 172 Mediterranean region. Risk is assessed based on the predicted growth of Geobacillus stearothermophilus in evaporated milk with a shelf life of one year for the supply chain of 23 Mediterranean capitals for five years (2012 - 2016). Red: High risk, Orange: Moderate Risk, Yellow: Low Risk, Green: Very Low Risk

Figure 5.5. Predicted growth of Geobacillus stearothermophilus (red lines) for 173 scenarios in which evaporated milk with a shelf life of six months is exported to Damascus for the periods October - March (a), January - June (b) and March - August (c) based on the hourly temperature data (blue lines) of the years 2014 - 2015.

Figure 6.1. Excel application for the use of the developed predictive models 179 in the quality control system of fruit drinks production.

Figure 6.2. Sources of variability affecting microbial growth and exposure 182 assessment (adapted from Koutsoumanis et al., 2016).

List of Abbreviations

aw Water activity

API CHB kits Microbial identification kits bioMerieux

ATCC American Type Culture Collection

BAM Bacillus acidocaldarius medium

Ca2+ Calcium cation

CFU Colony - Forming Units

CIP Cleaning - In - Place

CMI Cardinal Model with Inflection

CO2 Carbon dioxide

CwlJ Spore coat protein, cell wall hydrolase

D - sugars Sugars with hydroxyl group is to the right

2,6 - DBP 2,6 - Dibromophenol

2,6 - DCP 2,6 - Dichlorophenol

DEFT Direct Epifluorescent Filter Technique

DnaK Molecular chaperone protein (also known as Hsp70 - heat shock protein 70 kD)

DNA Deoxyribonucleic Acid

DPA Dipicolinic Acid

DSI Direct Steam Injection

DSM Leibniz Institute German Collection of Microorganisms and Cell Cultures

ELISA Enzyme - linked Immunosorbent Assay

EOs Essential Oils

FDA Food and Drug Administration

FIFO First - In - First - Out approach

FT - IRS Fourier Transform Infrared Spectroscopy

G Generation time

G + C Guanine and Cytosine content

GC - MS Gas Chromatography - Mass Spectrometry ger operons Germination operons encoding Ger proteins

GHPs Good Hygiene Practices

GMPs Good Manufacturing Practices

GRAS Generally Recognized as Safe gyrB Gene encoding the different topological forms h0 Physiological state parameter of cells/spores

H+ Hydrogen cation

H2 Dihydrogen

HACCP Hazard Analysis and Critical Control Points

HGYE Hiraishi Glucose Yeast Extract

HL Hosmer - Lemeshow statistic

HRMA High Resolution Melting Analysis

H2S Hydrogen Sulfide

HV Hypervariable

IFU International Fruit and Vegetable Juice Association

IM Inner Membrane

ISR Intergenic Spacer Regions

ITS Internal Transcribed Spacer

JFJA Japan Fruit Juice Association

K+ Potassium cation

λ Lag times derived from single spores

λopt Value of lag at T = Topt

L+ Right - turning (L+) lactic acid

LACF Low - Acid Canned Foods

L - amino acids “Left - handed” amino acids

μmax Maximum specific growth rate

Maximum specific growth rate of G. stearothermophilus in the evaporated milk stored at 62 °C

μopt Optimum value for maximum specific growth rate corresponding to optimum growth conditions

M - CGH Microarray - based Comparative Genome Hybridization

MICs Minimum Inhibitory Concentrations

Mg2+ Magnesium cation

MK - 7 Menaquinone K7

MLST Multilocus Sequence Typing

MLVA Multiple - Locus Variable number tandem repeat Analysis

MLV ‐ HRMA Multilocus variable number tandem repeat combined with high - resolution melting analysis

Mn2+ Mangenese cation

MPCA Milk Plate Count Agar mRNA Messenger Ribonucleic Acid

N0 Average initial number of spores initiating growth

Ndet Population level at the chosen Tdet

Nmax Maximum population level

NA Nutrient Agar

NB Nutrient Broth nGRs Nutrient Germinant Receptors

OD Optical Density

ODzero Average of OD values recorded in all 100 wells at time - zero

OM Outer Membrane

OSA Orange Serum Agar

PATP Pressure - Assisted Thermal Processing

PCR Polymerase Chain Reaction

PDA Potato Dextrose Agar

PFGE Pulsed - Field Gel Electrophoresis

PG Peptidoglycan pHmax Theoretical maximum pH for bacterial growth pHmin Theoretical minimum pH for bacterial growth pHopt Theoretical optimum pH for bacterial growth

PM Predictive Microbiology

QMRA Quantitative Microbial Risk Assessment

R2 Coefficient of determination

RAPD - PCR Random Amplification of Polymorphic DNA - Polymerase Chain Reaction recA Gene for the repair and maintenance of DNA recN Gene for repair of double strand breaks in the chromosome when these breaks occur at two or more locations, but not when there is a single break in the chromosome

RFID Radio - Frequency Identification

RFLP Restriction Fragment Length Polymorphism

RMSE Root Mean Square Error

RNA Ribonucleic Acid rpoB Gene encodes the β subunit of bacterial RNA polymerase

16S rRNA 16S ribosomal Ribonucleic Acid

RT ‐ PCR Reverse Transcription Polymerase Chain Reaction

σ Sigma factors

SASPs Small, Acid - soluble Spore Proteins sch squalene - hopene cyclase - encoding gene

SleB Spore cortex - lytic enzyme

SPA Staphylococcal Protein A spo0 Stage 0 sporulation gene

Tdet Absorbance time to detection

Tmax Theoretical maximum temperature for bacterial growth

Tmin Theoretical minimum temperature for bacterial growth tNmax Sum of the time required for the microorganism to multiply from the initial to the maximum level

Topt Theoretical optimum temperature for bacterial growth

TaqMan Hydrolysis probes designed to increase the specificity of quantitative PCR

TEM Transmission Electron Microscopy

TPC Thermophilic Plate Count tRNAala Transfer Ribonucleic Acid molecule that binds L - alanine tRNAile Transfer Ribonucleic Acid molecule that binds L - isoleucine

TSA Tryptone Soy Agar

TSB Tryptone Soy Broth

TSC Thermophilic Spore Count

TTI Time Temperature Integrators tts Time - to - spoilage

UF Ultra - Filtration

UHT Ultra - High Temperature

UV Ultraviolet

VNTR Variable Number Tandem Repeat

WGS Whole Genome Sequencing

WPC 80 Whey Protein Concentrate 80%

YSG agar Yeast - Starch - Glucose agar

Abstract The emergence of thermophilic endospore – forming bacteria is especially troublesome regarding the quality of thermally processed non – refrigerated foods. The elimination and control of their spores in the final product is extremely difficult as they are characterized by high heat resistance and as a result they are able to survive the heat treatments applied in the food industry. Therefore, thermophilic spore - forming bacteria are considered as one of the most important quality problems for non – refrigerated foods especially those that are distributed and stored for a long – term in hot climate countries. In addition, although these products are considered ‘microbiologically stable’ for the supply chain of most countries with temperate climate, the increase of temperature, due to global warming, may affect their stability and increase the risk of spoilage. A typical example supporting this hypothesis is the recent recall of pasteurized natural fruit juices in Greece, after consumer complaints for spoilage (off – flavour) and the detection of high populations of Alicyclobacillus spp. in the product. A structured quality assurance system for controlling thermophilic spore - forming bacteria in foods should be based on thorough risk analysis and prevention through monitoring, recording and controlling of critical parameters throughout the entire product’s shelf - life. Predictive microbiology (PM) has been established itself as a scientific discipline that uses mathematical equations to summarize and provide quantitative information on the microbial responses in various foods under different conditions. The development of models to predict survival, growth or inactivation of microorganisms in foods has been a highly active research area within food microbiology during the last 25 years. However, although a considerable number of predictive models and predictive software tools have been developed for numerous pathogenic and spoilage bacteria, the available models for the growth of thermophilic spore - forming bacteria are limited. Therefore, the objective of this PhD thesis was to develop mathematical models for thermophilic spoilers and apply them for predicting spoilage of non - refrigerated foods, such as evaporated milk and pasteurized fruit drinks. The first part of this research was undertaken to provide an approach for modeling the effect of temperature on Geobacillus stearothermophilus ATCC 7953 growth and in predicting spoilage of evaporated milk. The growth of G. stearothermophilus was monitored in tryptone soy broth at isothermal conditions (35 - 67 °C). The data generated were used to model the effect of temperature on G. stearothermophilus growth with a cardinal type model. The cardinal values of the model for the maximum specific growth rate were Tmin=33.76°C, Tmax=68.14°C, Topt=61.82°C and μopt=2.068/h. The growth of G. stearothermophilus was assessed in evaporated milk at Topt in order to adjust the model to milk. The efficiency of the model in predicting G. stearothermophilus growth under non - isothermal conditions was evaluated by comparing predictions with observed growth kinetics under dynamic conditions and the results showed a good compliance with the model. The model was further used to predict the time-to-spoilage

(tts) of evaporated milk. The spoilage of this product was caused by acid coagulation when the pH approached a level around 5.2, eight generations after G. stearothermophilus reached the maximum population density (Nmax). Based on the above, the tts was predicted from the growth model as the sum of the time required for the microorganism to multiply from the initial to the maximum level ( ), plus the time required after the to complete eight generations. The observed tts was very close to the predicted one, indicating that the model is able to describe satisfactorily the growth of G. stearothermophilus and to provide realistic predictions for evaporated milk spoilage. Aiming at providing quantitative tools for predicting the behavior of the spoilage bacterium Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks, a growth/no growth interface model was developed, predicting the probability of growth as a function of temperature and pH. For this purpose, the growth ability of A. acidoterrestris ATCC 49025 was studied at different combinations of temperature (15 - 45 °C) and pH (2.02 - 5.05). The minimum pH and temperature where growth was observed was 2.52 (at 35 and 45 oC) and 25 oC (at pH ≥ 3.32), respectively. A logistic polynomial regression model was then fitted to the binary data (0: no growth, 1: growth) and, based on the concordance index (98.8%) and the Hosmer - Lemeshow statistic (6.226, P = 0.622), a satisfactory goodness – of - fit was demonstrated. In the second part of the study, the effects of temperature (25 - 55 °C) and pH (3.03 - 5.53) on A. acidoterrestris ATCC 49025 growth rate were investigated and quantitatively described using the cardinal temperature model with inflection and the cardinal pH model, respectively. The estimated values for the cardinal parameters Tmin, Tmax, Topt and pHmin, pHmax, pHopt were 18.11, 55.68, 48.60 °C and 2.93, 5.90, 4.22, respectively. The developed models were validated against growth data of A. acidoterrestris ATCC 49025 obtained in eight commercial pasteurized fruit drinks. The validation results showed a good performance of both models. In all cases where the growth/no growth interface model predicted a probability lower than 0.5, A. acidoterrestris ATCC 49025 was, indeed, not able to grow in the tested fruit drinks. Similarly, when the model predicted a probability above 0.9, growth was observed in all cases. A good agreement was also observed between growth predicted by the kinetic model and the observed kinetics of A. acidoterrestris ATCC 49025 in fruit drinks at both static and dynamic temperature conditions. The above studies were conducted at the population level and then the research was extended at the single spore level in order to evaluate the growth response of more realistic concentrations and to take variability into account. Hence, the lag times (λ) of G. stearothermophilus single spores were studied at different storage temperatures ranging from 45 to 59 °C using the Bioscreen C method. A significant variability of λ was observed among individual spores at all temperatures tested. The storage temperature affected both the position and the spread of the λ distributions. The minimum mean value of λ (i.e. 10.87 h) was observed at 55 °C, while moving away from this temperature resulted in an increase for both the mean and standard deviation of λ. A Cardinal Model with Inflection (CMI) was fitted to the reverse mean λ, and the estimated values for the cardinal parameters Tmin, Tmax, Topt and the optimum mean λ of G. stearothermophilus

were found to be 38.1, 64.2, 53.6 °C and 10.3 h, respectively. To interpret the observations, a probabilistic growth model for G. stearothermophilus individual spores, taking into account the λ variability, was developed. The model describes the growth of a population over time, initially consisting of N0 spores, as the sum of cells in each of the N0 imminent subpopulations originating from a single spore. Growth simulations for different initial contamination levels showed that for low N0 the number of cells in the population at any time is highly variable. An increase in N0 to levels exceeding 100 spores results in a significant decrease of the above variability and a shorter λ of the population. Considering for that the number of G. stearothermophilus surviving spores in the final product is usually very low, the data provided in this work can be used to evaluate the probability distribution of the time-to-spoilage and enable decision-making based on the “acceptable level of risk”. The predictive model for the effect of storage temperature on the growth of G. stearothermophilus was also applied in order to assess the risk of evaporated milk spoilage in the markets of the Mediterranean region. The growth of G. stearothermophilus in evaporated milk was evaluated during a shelf life of one year based on historical temperature profiles (hourly) covering 23 Mediterranean capitals for five years over the period 2012 – 2016, obtained from the Weather Underground database (http://www.wunderground.com/). In total, 115 scenarios were tested, simulating the distribution and storage conditions of evaporated milk in the Mediterranean region. The highest growth of G. stearothermophilus was predicted for Marrakech, Damascus and Cairo over the period 2012 – 2016, with mean values of 7.2, 7.4 and 5.5 log CFU/ml, respectively, followed by Tunis, Podgorica and Tripoli, with mean growth of 2.8, 2.4 and 2.3 log CFU/ml, respectively. For the rest 17 capitals, the mean growth of the spoiler was <1.5 log CFU/ml. The capitals Podgorica, Cairo, Tunis and Ankara showed the highest variability in the growth with standard deviation values for growth of 2.01, 1.79, 1.77 and 1.25 log CFU/ml, respectively. The predicted extent and the variability of growth during the shelf life were used to assess the risk of spoilage which was visualised in a geographical risk map. The growth model of G. stearothermophilus was also used to evaluate adjustments of the evaporated milk expiration date which can reduce the risk of spoilage. The quantitative data provided in the present study can assist the food industry to effectively evaluate the microbiological stability of these products throughout distribution and storage at a reduced cost (by reducing sampling quality control) and assess whether and under which conditions (e.g. expiration date) it may be possible to export a product to a country without spoilage problems. This decision support may lead to a significant benefit for both the competitiveness of the food industry and the consumer. The quantitative findings of the present PhD thesis are considered to be important in relation to the development of models for predicting food spoilage that are readily applicable in the food industry. Moreover, the quantitative description of individual spore heterogeneity can be used in stochastic approaches for evaluating the spoilage potential and enable decision – making, based on the “acceptable level of risk”. Future applications of the developed models may include the evaluation of the impact of global

warming on the quality of thermally processed non - refrigerated foods, which will help the food industry to design effective quality management and logistic systems.

Περίληψη

Τα θερμόφιλα σπορογόνα βακτήρια αποτελούν ένα από τα σημαντικότερα προβλήματα ποιότητας σε θερμικά επεξεργασμένα τρόφιμα που διακινούνται και αποθηκεύονται εκτός ψυγείου. Ο έλεγχος των σπορίων τους στο τελικό προϊόν είναι εξαιρετικά δύσκολος, καθώς χαρακτηρίζονται από υψηλή θερμοανθεκτικότητα και ως εκ τούτου είναι σε θέση να επιβιώνουν ακόμα και μετά από ισχυρή θερμική επεξεργασία. Συνεπώς, ο έλεγχος των συγκεκριμένων μικροοργανισμών θεωρείται ως σημαντική πρόκληση για τους υπευθύνους ποιότητας της βιομηχανίας τροφίμων, ειδικά όσον αφορά σε προϊόντα που πρόκειται να διακινηθούν και να αποθηκευτούν για μεγάλο χρονικό διάστημα σε χώρες με θερμό κλίμα. Επιπρόσθετα, για πολλά θερμικά επεξεργασμένα τρόφιμα που θεωρούνται ως «μικροβιολογικώς σταθερά» σε χώρες με εύκρατο κλίμα, η αύξηση της θερμοκρασίας, λόγω του φαινομένου της υπερθέρμανσης του πλανήτη, μπορεί να μεταβάλλει το δυναμικό μικροβιολογικής τους αλλοίωσης και να επιδράσει σημαντικά στη σταθερότητά τους. Χαρακτηριστικό παράδειγμα που ενισχύει αυτήν την υπόθεση είναι η πρόσφατη ανάκληση παστεριωμένων χυμών φρούτων στην Ελλάδα, μετά από καταγγελίες των καταναλωτών σχετικά με την αλλοίωσή τους (δυσάρεστη γεύση) και την ανίχνευση υψηλών πληθυσμών του Alicyclobacillus acidoterrestris στο προϊόν. Ένα δραστικό σύστημα διασφάλισης της ποιότητας και ελέγχου των θερμόφιλων σπορογόνων μικροοργανισμών στα τρόφιμα θα πρέπει να βασίζεται στην αρχή της ανάλυσης επικινδυνότητας και την πρόληψη μέσω της παρακολούθησης, της καταγραφής και του ελέγχου των κρίσιμων παραμέτρων καθόλη τη διάρκεια ζωής των προϊόντων. Η Ποσοτική Μικροβιολογία έχει καθιερωθεί ως μία επιστημονική προσέγγιση που καλύπτει τα παραπάνω προαπαιτούμενα μέσω της χρήσης μαθηματικών μοντέλων πρόβλεψης, που παρέχουν πληροφορίες σχετικά με τη μικροβιακή συμπεριφορά σε σχέση με τα χαρακτηριστικά των τροφίμων και τις συνθήκες επεξεργασίας και συντήρησής τους. Τα τελευταία 25 χρόνια η ανάπτυξη μαθηματικών μοντέλων πρόβλεψης της συμπεριφοράς των μικροοργανισμών αποτελεί ένα από τα περισσότερο ταχέως αναπτυσσόμενα πεδία της Μικροβιολογίας Τροφίμων. Ωστόσο, παρά την ανάπτυξη ενός μεγάλου αριθμού μοντέλων και λογισμικών για διάφορους παθογόνους και αλλοιογόνους μικροοργανισμούς, τα διαθέσιμα μοντέλα πρόβλεψης της συμπεριφοράς θερμόφιλων σπορογόνων μικροοργανισμών στα τρόφιμα είναι πολύ περιορισμένα. Ως εκ τούτου, ο στόχος της παρούσας διδακτορικής διατριβής ήταν η ανάπτυξη και αξιολόγηση μαθηματικών μοντέλων πρόβλεψης της συμπεριφοράς θερμόφιλων σπορογόνων μικροοργανισμών και η εφαρμογή τους για την εκτίμηση του χρόνου αλλοίωσης των θερμικά επεξεργασμένων τροφίμων που διακινούνται και αποθηκεύονται εκτός ψυγείου. Το πρώτο μέρος αυτής της έρευνας εστίασε στη μαθηματική περιγραφή της επίδρασης της θερμοκρασίας στην ανάπτυξη του Geobacillus stearothermophilus ATCC 7953 και την πρόβλεψη της αλλοίωσης του συμπυκνωμένου – αποστειρωμένου

γάλακτος. Η ανάπτυξη του G. stearothermophilus μελετήθηκε σε tryptone soy broth υπό σταθερές συνθήκες θερμοκρασίας (35 - 67 °C). Τα δεδομένα που προέκυψαν χρησιμοποιήθηκαν για να περιγραφεί μαθηματικά η επίδραση της θερμοκρασίας στην ανάπτυξη του G. stearothermophilus με ένα μοντέλο θεμελιωδών παραμέτρων. Οι θεμελιώδεις τιμές του μοντέλου για το μέγιστο ειδικό ρυθμό ανάπτυξης ήταν Tmin=33.76 °C, Tmax=68.14 °C, Topt=61.82 °C και μopt=2.068/h. Η ανάπτυξη του G. stearothermophilus μελετήθηκε επίσης σε συμπυκνωμένο - αποστειρωμένο γάλα στην

Topt προκειμένου το μοντέλο να προσαρμοστεί στο συγκεκριμένο προϊόν. Η αποτελεσματικότητα του μοντέλου στην πρόβλεψη της ανάπτυξης του G. stearothermophilus σε δυναμικές συνθήκες θερμοκρασίας αξιολογήθηκε συγκρίνοντας τις προβλέψεις με την παρατηρούμενη ανάπτυξη και τα αποτελέσματα έδειξαν πολύ καλές επιδόσεις του μοντέλου. Το μοντέλο χρησιμοποιήθηκε περαιτέρω για την πρόβλεψη του χρόνου αλλοίωσης (tts) του συμπυκνωμένου - αποστειρωμένου γάλακτος. Τα αποτελέσματα έδειξαν ότι η αλλοίωση αυτού του προϊόντος προκαλείται από όξινη πήξη όταν το pH πλησίασει στο 5,2 περίπου, οκτώ γενεές όταν ο G. stearothermophilus φτάσει στο μέγιστο επίπεδο πληθυσμού (Nmax). Με βάση τα παραπάνω, ο tts προβλέφθηκε από το μοντέλο ανάπτυξης ως το άθροισμα του χρόνου που απαιτείται για τον πολλαπλασιασμό του μικροοργανισμού από το αρχικό στο

μέγιστο επίπεδο ( ), συν το χρόνο που απαιτείται για να ολοκληρώσει οκτώ γενεές. Ο παρατηρούμενος tts ήταν πολύ κοντά στον προβλεπόμενο, γεγονός που δείχνει ότι το μοντέλο είναι σε θέση να περιγράψει ικανοποιητικά την ανάπτυξη του G. stearothermophilus και να δώσει ρεαλιστικές προβλέψεις για την αλλοίωση του συμπυκνωμένου – αποστειρωμένου γάλακτος. Με στόχο την παροχή ποσοτικών εργαλείων για την πρόβλεψη της συμπεριφοράς του αλλοιωγόνου βακτηρίου Alicyclobacillus acidoterrestris ATCC 49025 σε ποτά φρούτων, στο δεύτερο μέρος της διατριβής αναπτύχθηκε ένα μοντέλο πρόβλεψης της πιθανότητας ανάπτυξης ως συνάρτηση της θερμοκρασίας και του pΗ. Για το σκοπό αυτό, μελετήθηκε η ικανότητα ανάπτυξης του A. acidoterrestris ATCC 49025 σε διαφορετικούς συνδυασμούς θερμοκρασίας (15 - 45 °C) και pΗ (2,02 - 5,05). Το ελάχιστο pΗ και θερμοκρασία όπου παρατηρήθηκε ανάπτυξη ήταν 2,52 (στους 35 και 45 °C) και 25 °C (σε pH ≥ 3,32), αντίστοιχα. Ακολούθως, ένα μοντέλο λογιστικής πολυωνυμικής παλινδρόμησης προσαρμόστηκε στα δυαδικά δεδομένα (0: καμία ανάπτυξη, 1: ανάπτυξη) και με βάση το δείκτη σύγκρισης (98,8%) και το δείκτη Hosmer - Lemeshow (6,226, P = 0,622), αποδείχθηκε η ικανοποιητική προσαρμογή του μοντέλου. Επιπλέον, πραγματοποιήθηκε ποσοτική περιγραφή της συνδυαστικής επίδρασης της θερμοκρασίας (25 - 55 °C) και του pH (3,03 - 5,53) στο ρυθμό ανάπτυξης του Α. acidoterrestris ATCC 49025 χρησιμοποιώντας το θεμελιώδες μοντέλο θερμοκρασίας και pH, αντίστοιχα. Οι εκτιμώμενες τιμές για τις θεμελιώδεις

παραμέτρους Tmin, Tmax, Topt και pHmin, pHmax, pHopt ήταν 18,11, 55,68, 48,60 °C και 2,93, 5,90, 4,22, αντίστοιχα. Τα μοντέλα στη συνέχεια αξιολογήθηκαν έναντι των δεδομένων ανάπτυξης του Α. acidoterrestris ATCC 49025 που ελήφθησαν σε οκτώ παστεριωμένα ποτά φρούτων που κυκλοφορούν στην Ελληνική αγορά. Τα αποτελέσματα της αξιολόγησης έδειξαν πολύ καλές επιδόσεις των μοντέλων. Σε όλες τις περιπτώσεις όπου

το μοντέλο ικανότητας ανάπτυξης προέβλεπε πιθανότητα μικρότερη από 0,5, o Α. acidoterrestris ATCC 49025 δεν ήταν πράγματι ικανός να αναπτυχθεί στα υπό μελέτη ποτά φρούτων. Ομοίως, όταν το μοντέλο προέβλεπε πιθανότητα άνω του 0,9, παρατηρήθηκε ανάπτυξη σε όλες τις περιπτώσεις. Καλή συμφωνία παρατηρήθηκε επίσης μεταξύ της προβλεπόμενης ανάπτυξης από το κινητικό μοντέλο και της παρατηρούμενης κινητικής του A. acidoterrestris ATCC 49025 σε ποτά φρούτων τόσο σε στατικές όσο και σε δυναμικές θερμοκρασιακές συνθήκες συντήρησης. Μετά τις παραπάνω μελέτες που διεξήχθησαν σε επίπεδο πληθυσμού, στο τρίτο μέρος της διατριβής η έρευνα επεκτάθηκε σε επίπεδο μεμονωμένων σπορίων προκειμένου να αξιολογηθεί η απόκριση της ανάπτυξης πιο ρεαλιστικών επιπέδων αρχικής επιμόλυνσης και να ληφθεί υπόψη η μεταβλητότητα στην ατομική συμπεριφορά των σπορίων. Οι χρόνοι υστέρησης (λ) των μεμονωμένων σπορίων του G. stearothermophilus μελετήθηκαν σε διαφορετικές θερμοκρασίες συντήρησης μεταξύ 45 και 59 °C χρησιμοποιώντας τη μέθοδο Bioscreen C. Τα αποτελέσματα έδειξαν μία σημαντική μεταβλητότητα του λ μεταξύ των μεμονωμένων σπορίων σε όλες τις υπό μελέτη θερμοκρασίες. Η θερμοκρασία αποθήκευσης επηρέασε τόσο τη θέση όσο και τη διασπορά των λ κατανομών. Η ελάχιστη μέση τιμή λ (10,87 h) παρατηρήθηκε στους 55 °C, ενώ η απομάκρυνση από αυτή τη θερμοκρασία είχε ως αποτέλεσμα την αύξηση τόσο για τη μέση τιμή όσο και για την τυπική απόκλιση του λ. Ένα μοντέλο θεμελιωδών παραμέτρων προσαρμόστηκε στην αντίστροφη μέση τιμή του λ και οι εκτιμώμενες

τιμές για τις θεμελιώδεις παραμέτρους Tmin, Tmax, Topt και η άριστη μέση τιμή λ του G. stearothermophilus βρέθηκαν να είναι 38,1, 64,2, 53,6 °C και 10,3 h, αντίστοιχα. Για την ερμηνεία των παρατηρήσεων, αναπτύχθηκε ένα στοχαστικό μοντέλο ανάπτυξης για μεμονωμένα σπόρια του G. stearothermophilus, λαμβάνοντας υπόψη τη μεταβλητότητα του λ. Το μοντέλο περιγράφει την ανάπτυξη ενός πληθυσμού, αρχικά αποτελούμενου από Ν0 σπόρια, συναρτήσει του χρόνου, ως το άθροισμα των κυττάρων σε καθένα από τους Ν0 επικείμενους υποπληθυσμούς προερχόμενους από ένα μόνο σπόριο. Οι προσομοιώσεις ανάπτυξης για διαφορετικά αρχικά επίπεδα επιμόλυνσης

έδειξαν ότι για χαμηλό Ν0 ο αριθμός των κυττάρων στον πληθυσμό ανά πάσα χρονική στιγμή παρουσιάζει πολύ μεγάλη μεταβλητότητα. Η αύξηση του Ν0 σε επίπεδα που υπερβαίνουν τα 100 σπόρια έχει ως αποτέλεσμα τη σημαντική μείωση της παραπάνω μεταβλητότητας καθώς και τη μείωση του χρόνου υστέρησης λ του πληθυσμού. Λαμβάνοντας υπόψη ότι ο αριθμός των σπορίων του G. stearothermophilus που επιβιώνουν μετά τη θερμική επεξεργασία στο τελικό προϊόν είναι συνήθως πολύ χαμηλός, τα δεδομένα που παρέχονται μπορούν να χρησιμοποιηθούν για να αξιολογηθεί η κατανομή πιθανότητας του χρόνου αλλοίωσης και να επιτραπεί η λήψη αποφάσεων με βάση το "αποδεκτό επίπεδο επικινδυνότητας". Στο τελευταίο μέρος της διατριβής το μοντέλο πρόβλεψης για την επίδραση της θερμοκρασίας στην ανάπτυξη του G. stearothermophilus χρησιμοποιήθηκε, προκειμένου να εκτιμηθεί η επικινδυνότητα αλλοίωσης του συμπυκνωμένου – αποστειρωμένου γάλακτος στις αγορές της περιοχής της Μεσογείου. Η ανάπτυξη του G. stearothermophilus στο γάλα εκτιμήθηκε κατά τη διάρκεια ζωής ενός έτους με βάση τα ιστορικά προφίλ θερμοκρασίας (ωριαία) από 23 Μεσογειακές πρωτεύουσες για πέντε

έτη κατά την περίοδο 2012 – 2016, όπως ελήφθησαν από τη βάση δεδομένων Weather Underground (http://www.wunderground.com/). Συνολικά, ελέγχθηκαν 115 σενάρια προσομοιάζοντας τις συνθήκες διανομής και συντήρησης του γάλακτος στην περιοχή της Μεσογείου. Η μεγαλύτερη ανάπτυξη του G. stearothermophilus προβλέφθηκε για το Μαρακές, τη Δαμασκό και το Κάιρο για την περίοδο 2012 – 2016 με μέσες τιμές 7,2, 7,4 και 5,5 log CFU/ml, αντίστοιχα, ακολουθούμενες από την Τύνιδα, την Ποντγκόριτσα και την Τρίπολη με μέση αύξηση 2,8, 2,4 και 2,3 log CFU/ml, αντίστοιχα. Για τις υπόλοιπες 17 πρωτεύουσες, η μέση αύξηση του αλλοιωγόνου ήταν <1,5 log CFU/ml. Οι πρωτεύουσες Ποντγκόριτσα, Κάιρο, Τύνιδα και Άγκυρα έδειξαν τη μεγαλύτερη μεταβλητότητα στην ανάπτυξη κατά τη διάρκεια των 5 ετών που εξετάστηκαν, με τιμές τυπικής απόκλισης της ανάπτυξης 2,01, 1,79, 1,77 και 1,25 log CFU/ml, αντίστοιχα. Η προβλεπόμενη έκταση και η μεταβλητότητα της ανάπτυξης κατά τη διάρκεια ζωής χρησιμοποιήθηκαν για να εκτιμηθεί η επικινδυνότητα αλλοίωσης που απεικονίστηκε σε ένα γεωγραφικό χάρτη επικινδυνότητας. Το μοντέλο ανάπτυξης του G. stearothermophilus χρησιμοποιήθηκε επίσης για να αξιολογήσει την απαιτούμενη προσαρμογή της ημερομηνίας λήξης του γάλακτος που μπορεί να μειώσει την επικινδυνότητα αλλοίωσης σε αποδεκτά επίπεδα. Τα ποσοτικά στοιχεία που παρέχονται στην παρούσα μελέτη μπορούν να βοηθήσουν τη βιομηχανία τροφίμων να αξιολογήσει αποτελεσματικά τη μικροβιολογική σταθερότητα αυτών των προϊόντων σε όλη τη διάρκεια της διανομής και αποθήκευσής τους με μειωμένο κόστος (μειώνοντας σημαντικά το δειγματοληπτικό ποιοτικό έλεγχο) και να αξιολογήσει εάν και υπό ποιες συνθήκες (π.χ. ημερομηνία λήξης) θα είναι σε θέση να εξάγει ένα προϊόν σε μία χώρα ελαχιστοποιώντας τα προβλήματα αλλοίωσης. Αυτή η υποστηρικτική πληροφορία μπορεί να οδηγήσει σε σημαντικό όφελος τόσο για την ανταγωνιστικότητα της βιομηχανίας τροφίμων όσο και για τον καταναλωτή. Τα ευρήματα της παρούσας διδακτορικής διατριβής σχετικά με την ανάπτυξη των μοντέλων πρόβλεψης είναι άμεσα εφαρμόσιμα στη βιομηχανία τροφίμων και μπορούν να προσφέρουν πολύτιμες πληροφορίες για τη βελτίωση του συστήματος διασφάλισης της ποιότητας των τροφίμων. Επιπλέον, οι στοχαστικές προσεγγίσεις που προτείνονται σε σχέση με την ετερογένεια της ατομικής συμπεριφοράς των σπορίων μπορούν να υποστηρίξουν τη λήψη αποφάσεων με βάση το "αποδεκτό επίπεδο επικινδυνότητας". Τέλος τα ποσοτικά εργαλεία που αναπτύχθηκαν μπορούν να αποτελέσουν τη βάση για τη μελέτη των επιπτώσεων της υπερθέρμανσης του πλανήτη στην ποιότητα - σταθερότητα των τροφίμων που διακινούνται και αποθηκεύονται εκτός ψυγείου και να βοηθήσουν τη βιομηχανία τροφίμων να αντιμετωπίσει το πρόβλημα σχεδιάζοντας αποτελεσματικά τα συστήματα διασφάλισης της ποιότητας και διαχείρισης της αντίστοιχης εφοδιαστικής αλυσίδας.

Chapter 1 Literature Review and Thesis Outline

Chapter 1

Most of the thermophilic endospore - formers can spoil a wide range of thermally processed, non - refrigerated foods, including acid products (Sperber & Doyle, 2009), leading to product downgrades and revenue losses for food manufacturers. Failures in products quality is frequently caused by thermostable spores of members of the Bacillaceae family, which show a wide spectrum of resistance to cleaning and preservation treatments. Hence, the presence of thermophilic bacteria is especially troublesome regarding the commercial viability of thermally processed non – refrigerated foods, which are stored for extended periods of time, both in local and international markets with hot climates, at temperatures frequently exceeding 30 - 35 °C. The contamination of the final product with viable spores of thermophiles cannot be easily prevented. In addition, contamination usually involves very low number of spores and cannot be detected until the food is exposed to high ambient temperatures, allowing germination and growth (Hill & Smythe, 2012). Τhe key, therefore, to preventing spoilage by thermophilic endospore – formers, is to apply a structured quality assurance system based on thorough risk analysis and prevention through monitoring, recording and controlling of critical parameters through the entire products shelf - life. Moreover, the climate of the planet is changing with rapid pace and one of the most important and immediate expressions of such climate change is the increase of the planet’s temperature (IPCC, 2013; Morice, Kennedy, Rayner, & Jones, 2012; Räisänen et al., 2004; UNFCCC, 2014). Elevated temperatures correlate with changes in survival, acceleration of the replication cycles of foodborne microorganisms or the transmission potential of pathogens in the environment, food and feed (Miraglia et al., 2009), while extended summer seasons may also increase the chance of food mishandling (ECDC, 2012; Hellberg & Chu, 2015). From a food safety perspective, there is a need to control microbiological risks at any stage of the food chain, along the primary production, processing, storage and transport chain (Chakraborty & Newton, 2011; Miraglia et al., 2009) and develop adaptation strategies to cope with food spoilage implications due to climate change in the future (Jacxsens, Luning, Van der Vorst, Devlieghere, Leemans, & Uyttendaele, 2010; Tirado et al., 2010). Unlike the number of studies on food sustainability and food safety, the studies regarding the association of global warming with the shelf - life of foods are, in general, very limited (Medina-Martínez, Allende, Barberá & Gil, 2015). However, further warming is likely to bring about significant changes in the growth and the spoilage potential of the thermophilic endospore - forming bacteria. Particularly, climate change may lead to a long exposure of foods at high ambient temperatures, allowing microbial

Literature review and thesis outline growth up to the spoilage level. As a result, the “microbiologically stable’’ food products, which are distributed and stored without refrigeration, will be affected by the abovementioned bacteria due to global warming. In addition, the food industry, through Predictive Microbiology tools, will be able to assess whether and under which conditions (e.g. expiration date) it will be possible to export a product to a country, in particular with warm climate (i.e. Mediterranean region), while minimizing spoilage problems. Τhe ultimate goal of this PhD thesis was to develop stochastic predictive models for the growth of thermophilic endospore – forming bacteria in thermally processed non – refrigerated food products, by incorporating various factors characterized by variability, their further application by the food industry under the influence of the climate change and eventually redefining the shelf - life of canned food products so that the food industry is able to export high - quality foods to other countries.

1. The class Bacilli

The genus Bacillus, part of the Bacillaceae family, includes a widely diverse taxonomic group of aerobic or facultative anaerobic, spore – forming, Gram - positive (or Gram – variable), rod – shaped bacteria that utilize a wide range of carbon sources for heterotrophic or, seldom, autotrophic growth. The Bacillus group is phenotypically and genotypically heterogeneous and versatile (Ash, Farrow, Wallbanks, & Collins, 1991; Priest, 1993), which is reflected in their great physiological diversity, including a wide variety of physiological specializations associated with survival and colonization of various ecological niches. Thus, the presence or absence of certain genes involved in metabolism is pivotal in suggesting the niche of a strain. Among these features are the degradation of most substrates derived from plant and animal sources (cellulose, starch, pectin, proteins, agar, carbohydrates and others), the production of antimicrobial substances (antibiotics and bacteriocins), nitrification, denitrification, nitrogen fixation, facultative anaerobiosis, facultative lithotrophy, autotrophy, heterotrophy, acidophily (growth down to pH 1.5), neutrophily, alkaliphily (growth up to pH 11), psychrophily (growth down to - 5 °C), thermophily (growth up to 78 °C), halophily and parasitism. Due to this heterogeneity in phenotypic properties, Bacillus spp. exhibits an extremely wide range of nutritional requirements, growth conditions, metabolic diversity and DNA base composition. However, the main characteristics that define this genus are the Gram - positive cell wall and the capacity to form endospores under various

Chapter 1 unfavorable and stressful conditions (Abriouel, Benomar, Huch, Franz, & Gálvez, 2014; Ash et al., 1991).

2. The class Clostridia

Clostridium species are Gram - positive, obligate anaerobes, usually catalase negative, rod – shaped, endospore – forming and not osmotolerant. They are typically involved in the spoilage of foods that have a highly negative oxidation - reduction potential, such as canned or vacuum - packaged foods (Sperber & Doyle, 2009). Spoilage by clostridia is caused via saccharolytic or proteolytic activity and is manifested by changes in product pH (caused by organic acid production), gas production and production of foul odours (e.g. through volatile acids). While Clostridia pose a problem, the spoilage Bacilli pose an even greater challenge to microbial food stability, often due to their extreme levels of thermal resistance.

3. Bacterial Endospores 3.1. Spore structure and resistance

The key element of the adaptive responses of any bacterium is the ability to sense and respond to environmental changes by modulating the pattern of gene expression in a coordinated manner. In the harsh and ever - changing environment (soil, animal gut, food processing facilities, processed foods), bacteria have evolved specific adaptive strategies, including the formation of dormant, highly resistant entities called endospores, to ensure their individual survival and, even more importantly, the survival of the population (Eijlander, Abee, & Kuipers, 2011). The contamination of foods with bacterial spores is well - recognized in almost any type of food in many food industries around the world as a major issue affecting food safety and quality (Faille et al., 2014; Faille, Fontaine, & Bénézech, 2001). This is due to the innate resistance of spores to many of the preservation treatments, such as thermal processing and the addition of antimicrobial compounds, employed by food manufacturers to inactivate microorganisms and increase the microbial stability of foods (Chandler et al., 2001; Cortezzo & Setlow, 2005; Jones, Padula, & Setlow, 2005; Scheldeman, Herman, Foster, & Heyndrickx, 2006). Several spore structural properties contribute to this resistance.

Literature review and thesis outline

Spore structure plays a major role in spore resistance due to a number of spore layers. Each plays a specific role in resistance. From the outside inwards, the various spore layers include the exosporium, coat, outer membrane (OM), cortex, germ cell wall, inner membrane (IM), and core (Fig. 1.1) (Setlow, 2014).

Figure 1.1. Schematic structure of a spore of Bacillus or Clostridium species. Some species contain a layer, known as an exosporium (adapted from Setlow, 2012).

The outermost structure (exosporium) is not present in spores of all species and is absent in B. subtilis spores. Transmission Electron Microscopy (TEM) of thin sections of Geobacillus spp. spores isolated from milk powder production lines has revealed an exosporium (Seale, Bremer, Flint, & McQuillan, 2010). However, the role of this structure for Geobacillus is unknown. In spores of B. anthracis the exosporium may act as a permeability barrier, restricting access of antibodies to antigens present in the spore coat and provide a hydrophobic surface that enhances the spore’s adhesive properties (Henriques & Moran, 2007). Other than this, little is known about the function of the exosporium in spore resistance. The spore coat contains a large fraction of the spore’s total protein and acts as a permeability barrier restricting access of large molecules, such as enzymes, to potential sensitive targets located further within the spore’s inner layers (Setlow, 2014). Consequently, the spore coat is responsible for protection against hydrolytic enzymes, such as lysozyme that degrade peptidoglycan (PG), and thus protects spores against predation by protozoa (Klobutcher, Ragkousi, & Setlow, 2006;

Chapter 1

Laaberki & Dworkin, 2008). The coat is also important in spore protection against a variety of biocidal chemicals, probably by reacting nonspecifically with and detoxifying such chemicals before they attain sensitive targets further within the spore (Henriques & Moran, 2007; Setlow, 2006, 2012), protecting against UV irradiation through pigments that absorb strongly in the UV region (Khaneja et al., 2010) and mechanical disruption. Underneath the coat is the OM, the role of which in spore resistance remains unknown. The OM can also contain pigments, generally carotenoids that may play a role in spore UV resistance as noted above (Khaneja et al., 2010). Although this membrane plays an essential role in spore formation, the possible role of the OM as a permeability barrier is not completely clear. Underneath the OM are two PG layers of slightly different structure, the first one called spore cortex and the second is the thinner germ cell wall. While both layers are essential for spore viability, and the cortex undoubtedly is essential for some of the newly discovered properties of the core (see below), these two layers are not known to play any direct active role in spore resistance (Setlow, 2014). Beneath the germ cell wall is the spore’s IM. While the lipid composition of the IM is not particularly unusual (Griffiths & Setlow, 2009), the IM itself has some newly discovered properties (Setlow, 2006). In particular, (i) fluorescent lipid molecules in the IM are largely immobile; (ii) the IM has a much higher viscosity than the germinated spore’s plasma membrane; and (iii) the IM’s passive permeability to small hydrophilic and hydrophobic molecules is extremely low, including methylamine and perhaps even water (Cortezzo & Setlow, 2005; Setlow, 2006). While the reasons for these properties of the IM are largely unknown, they, especially the IM’s relative impermeability, seem likely to be important in spore resistance to some biocidal hydrophilic chemicals, by restricting their access to targets in the spore’s interior. Indeed, damage to the IM appears to be the mechanism by which several oxidizing agents kill spores, although the nature of this damage is unknown (Cortezzo & Setlow, 2005; Setlow, 2006, 2012). The newly discovered properties of the spore’s IM are lost when spores complete germination. The final spore layer is the central core, otherwise known as protoplast, which maintains the genomic DNA and has a few newly discovered features that appear to play many roles in spore resistance (Setlow, 2006, 2007, 2012). These include (i) the core’s low water content (25 to 55% of wet weight), depending on the species, important in spore wet heat resistance; (ii) the accumulation of high levels of pyridine - 2,6 - dicarboxylic acid (dipicolinic acid [DPA]) in a 1:1 chelate with various

Literature review and thesis outline divalent cations, generally mostly Ca2+ (Ca - DPA), during the late stages of sporulation, important in spore resistance to some DNA - damaging agents and in maintenance of spore dormancy; and (iii) high levels of a spore – specific group of newly discovered proteins, the α/β - type small, acid - soluble spore proteins (SASPs) that saturate spore DNA and protect it from damage, by binding to GC ‐ rich regions of spore DNA and forming a tightly packed assembly, caused by desiccation, dry - and wet - heat, toxic chemicals and enzymes, UV and γ - radiation (Lee, Bumbaca, Kosman, Setlow, & Jedrzejas, 2008; Melly et al., 2002; Setlow, 2006). The a/β - type SASPs, synthesized late in sporulation within developing spores, are degraded early in spore outgrowth and the ultimately generated amino acids are important for spore metabolism and protein synthesis at this stage (Setlow, 2007). Finally, it should be noted that there is increasing evidence that spore resistance is not completely static, but can change as spores “mature” (Setlow, 2014). The resistance properties of spores are influenced by environmental conditions (such as temperature, pH and media composition in terms of mineral ions, in particular) during sporulation (Palop, Manas, & Conton, 1999), the physiological state of the microorganism, the composition of the heating medium (e.g. pH) and the recovery conditions for enumeration of heated bacterial spores (Scheldeman et al., 2006). The effect that different sporulation conditions can have on spore heat - resistance is an important element in the validation of thermal processes. Spore heat - resistance is strongly affected by thermal adaptation, core mineralization and core dehydration, though the latter has been identified as the defining property necessary for resistance (Beaman & Gerhardt, 1986). The distinctively low water content in the core is established during sporulation, in parallel with the uptake of divalent cations (e.g., Ca2+, Mn2+ and Mg2+), known as mineralization (Setlow, 1994). The type of divalent cation present in the spore core has been shown to influence heat - resistance. The degree of resistance acquired is subject to a combination of inherent and extrinsic influences. Thermophilic species produce spores that are inherently more resistant than spores of mesophiles, which in turn are inherently more resistant than spores of psychrophiles. This genetically inherited thermotolerance occurs due to the higher intrinsic thermostability of their proteins independently of core dehydration and/or mineralization (Beaman & Gerhardt, 1986; Palop, Manas, et al., 1999; Setlow, 2006). Thermal adaptation can also occur via extrinsic control of temperature during the sporulation process, where an increase in temperature subsequently increases spore heat - resistance (Atrih & Foster, 2001; Beaman & Gerhardt, 1986; Garcia, van der Voort, & Abee, 2010;

Chapter 1

González, Lopez, Martınez, Bernardo, & González, 1999; Melly et al., 2002; Palop, Raso, Pagán, Condón, & Sala, 1999). Similarly, the inclusion of a supplemental source of divalent cations in sporulation media, especially calcium, can increase resistance (Atrih & Foster, 2001; Cazemier, Wagenaars, & Ter Steeg, 2001). Higher incubation temperatures or the availability of a source of minerals to sporulating cells have both been shown to correlate with a decrease in the amount of water in the core and an increase in heat resistance (Atrih & Foster, 2001; Cazemier et al., 2001; Melly et al., 2002). In G. stearothermophilus ATCC 7953 spores, the impact of sporulation temperature, core dehydration and mineralization can add to the intrinsic wet - heat resistance to render the spore especially resistant (Leguérinel, Couvert, & Mafart, 2007; Marquis, Sim, & Shin, 1994).

3.2. Sporulation

Many Gram - positive bacteria, including members of the Bacilli and Clostridia, produce endospores according to a cellular program like that of B. subtilis, which has been characterized as the best - studied endospore - forming bacterium in terms of cell differentiation and development (Holt & Leadbetter, 1969; Piggot & Coote, 1976). When confronted by modification in the composition of the growth medium (nutrient depletion of carbon, nitrogen or, in some circumstances a phosphorus source for Bacillus sp., acidification for clostridia), high cell density or DNA damage, these normally rod - shaped organisms produce an oval, dormant cell called endospore, in a complicated process called sporulation. The spore is entirely distinct from the vegetative cell, possessing several molecules and cellular structures seen nowhere else in nature (Carlin, 2011; Murrell, 1967; Warth, Ohye, & Murrell, 1963). These unique components contribute to the spore’s most striking characteristics: it is metabolically dormant but actively interacting with its environment, highly resilient to environmental assault, viable and stable for extreme periods of time. Consequently, they can survive long distance transport by climatic or biological agents, leading to a higher dispersion rate than non - endospore - forming organisms (Martiny et al., 2006; Roberts & Cohan, 1995). In addition, the broad metabolic diversity found within this group is probably another factor that explains their ubiquity (Mandic-Mulec & Prosser, 2011). Sporulation begins well before the first morphological changes are detected by light microscopy. Before they can commit to sporulation, cells sense and measure a

Literature review and thesis outline complex variety of parameters, only part of which are nutrient levels. Sporulation is characterized by an asymmetric cell division that yields two compartments of unequal size, during which a copy of the genome is partitioned into each of the sister cells (Fig. 1.2). After the cell commits to sporulation, it builds a specialized spore septum. Unlike the septum that appears during vegetative cell division, this one is positioned off to one side, generating a smaller forespore, which carries its own genome and contributes to the differentiation process, and a larger mother cell. Although at this stage the two cells contain identical copies of DNA, they are programmed to express quite different subsets of genes that ultimately shape their respective fates. As sporulation proceeds, a phagocytosis - like process takes place, where the mother cell engulfs the forespore, insulating it within a double - membrane system derived from the septum, and contributes to the transformation of the forespore into mature spore; ultimately, the mother cell undergoes lysis (programmed cell death) releasing the endospore into the surrounding medium (Errington, 2003; Piggot & Hilbert, 2004). Following the completion of engulfment, two peptidoglycan structures which are layered between IM and OM, surround the forespore. The first, called germ cell wall, serves as the primordial wall of the newly formed vegetative cell, and the second the cortex, is essential for the attainment and maintenance of the dehydrated state of the spore core, for spore mineralization and for spore dormancy (Adam Driks, 1999; Henriques & Moran, 2007; Piggot & Coote, 1976). The next structure to be built is a thick protein shell that encases the forespore, called the coat. The final period of spore development, termed maturation, occurs with little obvious change in morphology, but during this period the characteristic properties of resistance, dormancy and germinability appear in sequence (Errington, 2003). Afterwards, the release of the mature spore by lysis of the mother cell occurs. The initiation of the process is kept under the strict control of an expanded two - component signal transduction system, called a phosphorelay. The phosphorelay integrates multiple stimuli to ensure that the cell sporulates only when all other survival strategies have been exhausted. Genes whose products are required for the initiation of sporulation and the proper formation of the asymmetrically positioned septum are denoted as spo0 (Driks, 2002; Piggot & Hilbert, 2004). The other key positive regulator of sporulation is a specific set of sigma factors, σ, which interact with core RNA polymerase and direct it to initiate transcription from at least 49 promoters, controlling 87 or more genes (Driks, 1999).

Chapter 1

Figure 1.2 Stages of sporulation (adapted from Errington, 2003).

3.3 Germination

In spite of its inert state, the spore can sense the reappearance of even minute amounts of nutrients in the environment and respond by converting back to a vegetatively growing cell (Abdelmadjid Atrih & Foster, 1999). This process, called germination, is reviewed by Moir (2006) followed by outgrowth. As soon as the spores are exposed to proper stimuli for growth (nutrients, temperature), a lag phase is always observed. It is defined as the time needed for a spore to return to its vegetative state, passing through the dormant phase, through germination after an irreversible cascade of events, followed by elongation to the size of a mature cell and, eventually, as a metabolically active cell to divide (Fig. 1.3) (Moir, Corfe, & Behravan, 2002; Setlow, 2003). Spore germination plays a key role in the first stage of food spoilage and food - borne infection, as it initiates the transformation of a dormant spore to the vegetative cell. Molecules in the food product, possibly in combination with the thermal treatment, act as a heat - activation step and may initiate spore germination (Abee et al., 2011). Specifically, as discussed in the

Literature review and thesis outline literature, the presence and the concentrations of specific compounds in milk may exert different behavioral responses of G. stearothermophilus. Ljunger (1970) and Vinter (1969) reported that the existence of ions, such as divalent cations (calcium, magnesium, potassium) in milk, can contribute to the outgrowth of mature spores and may be involved in the activation of sporulation.

Figure 1.3. Schematic outline of germination of spores of Bacillus species (adapted from Setlow, 2003).

It is generally accepted that spores must be activated before germination can occur. Heat, chemicals and a decrease in pH to 2 - 3 can activate spores (Sonali Ghosh, Zhang, Li, & Setlow, 2009; Iciek, Papiewska, & Molska, 2006). In the dairy industry, heat is the most likely mechanism of thermophilic spore activation, due to its extensive use in preservation technologies. Following activation, the germination process is induced by specific nutrients and some non - nutrient agents. The nutrient germinants (L - amino acids, D - sugars and purine nucleosides) bind to specific germinant receptors (nGRs), encoded by ger operons and residing in the spore’s inner membrane, triggering the release of spore small molecules, most notably monovalent cations (H+, K+) and pyridine - 2,6 - dicarboxylic acid (dipicolinic acid (DPA)) chelated with Ca2+ (CaDPA) (Moir, 2006; Setlow, 2003) and replacing them by water. The next step involves hydrolysis of the peptidoglycan in the cortex and further swelling of the spore core due to the ingress of water, expansion of the germ cell wall and the resumption of the spore metabolism and macromolecule synthesis that converts the germinated spore into a growing cell (Moir, 2006).

Chapter 1

Observations of individual spores showed that, in general, phenotypically different variable states with respect to spore germination performance are generated (Billon, McKirgan, McClure, & Adair, 1997; Chen, Huang, & Li, 2006; Sonali Ghosh & Setlow, 2009; Ghosh & Setlow, 2010; Kong, Zhang, Setlow, & Li, 2010; Leuschner & Lillford, 1999; Peng, Chen, Setlow, & Li, 2009; Stringer, Webb, George, Pin, & Peck, 2005; Webb, Pin, Peck, & Stringer, 2007; Zhang et al., 2010). This population variability is reflected by a spread of values for any property (i.e. individual spore lag time) of individual spores within the population. The reasons for this heterogeneity is possibly related to the stochastic effects on receptor – encoding gene expression and the differences in culture environmental conditions during sporulation (Elowitz, Levine, Siggia, & Swain, 2002; Setlow, Liu, & Faeder, 2012). In fact, differences in germination are the result of the genetic makeup of sporeformers and the conditions during sporulation. Specifically, the spore germination diversity is probable due to a variable lag phase between the time when the spores are mixed with a nutrient germinant and the time at which the rapid release of DPA starts (Chen et al., 2006; Kong et al., 2010; Setlow, 2013; Stringer et al., 2005; Zhang et al., 2010). Interestingly, the differences in the levels of cortex germination - specific lytic enzymes in individual spores, such as CwlJ and SleB, which initiate the spore’s peptidoglycan cortex degradation by released CaDPA from every spore during germination (Ghosh & Setlow, 2009; Moir, 2006; Peng et al., 2009; Setlow, 2003) may also contribute to the germination efficiency. Nevertheless, the stochastic fluctuations in the number of nutrient germinant receptors (nGRs) per spore due to epigenetic variations among individual spores, appear to be one of the main reasons of the significant heterogeneity in spore germination capacity at the population level (Ghosh & Setlow, 2009, 2010; Setlow, 2003). The rate of germination of the spores, and subsequently their lag time, is influenced by factors such as pH, water activity (aw), temperature, differences in temperature requirements for heat activation prior to germination among individual spores in populations, the concentration of a nutrient germinant, the chemical composition of the medium (Hornstra, Ter Beek, Smelt, Kallemeijn, & Brul, 2009; Moir et al., 2002), and the communication among the spores during germination (quorum sensing). A spore would be more likely to be affected by signals of an adjacent spore, as a signal of a molecule that diffuses into the growth medium would be more successful to reach the neighboring spores than the distant ones. Accordingly, the spores in close proximity tend to have a greater possibility to synchronize their germination (Caipo, Duffy, Zhao, & Schaffner, 2002; Webb, Stringer, Le Marc, Baranyi, & Peck, 2012;

Literature review and thesis outline

Zhang et al., 2010). However, it is argued that the heterogeneity in germinated spores is considered as an adaptive response of the bacterium to increase the probability of its population’s survival, as the germinated spores and subsequent growing cells are more vulnerable to environmental hazards. Therefore, the simultaneous germination of spores within a population could be disadvantageous, if the environment was actually not favorable for growth, thus endangering the fate of the whole population (Ghosh & Setlow, 2009; Setlow et al., 2012).

4. Spoilage thermophilic endospore – forming bacteria

Thermophilic bacteria are defined as aerobic, facultatively anaerobic and anaerobic bacteria capable of growing at temperatures between 35 and 70 °C. Their optimum growth temperature is usually between 45 – 65 °C, but it varies among species and strains. The pH range for growth varies with the different species of organisms. In general, most of the thermophilic endospore - formers can spoil a wide range of thermally processed products, including acidic foods (Sperber & Doyle, 2009). Failure of food preservation is frequently caused by thermostable spores of members of the Bacillaceae family, which show a wide spectrum of resistance to cleaning and preservation treatments. Thus, they are considered a nuisance rather than a help in agricultural and industrial practice, where elevated temperatures (35 - 65 °C) prevail during the manufacturing process or when a product is stored, causing great economic loss if the proper controls are not in place. These industries include paper mills, canning, juice pasteurization, sugar refining, gelatin production, dehydrated vegetable manufacture and dairy product manufacture (Chen et al., 2006; De Clerck, Marina Rodriguez-Diaz, Gillian Forsyth, et al., 2004; De Clerck, Marina Rodriguez-Diaz, Tom Vanhoutte, et al., 2004; De Clerck et al., 2004; Scott, Brooks, Rakonjac, Walker, & Flint, 2007; Tai, Lin, Kuo, & Liu, 2004). Although thermophilic endospore - formers are not described as pathogenic, they are difficult to eliminate and are among the principal causes of spoilage of canned foods, since not only their vegetative cells are adapted to life at elevated temperatures, but also their spores are unusually highly resistant to environmental insults. Their spores may be able to survive intense heat treatments, including commercial food sterilization, processes used for canning and ultra – high temperature processes applied in the manufacturing of liquid foods (André, Zuber, & Remize, 2013; Burgess, Lindsay, & Flint, 2010).

Chapter 1

Interestingly, it has been observed that thermophilic microorganisms, such as Geobacillus stearothermophilus, can grow adequately at temperatures which are sublethal or lethal for the majority of other microorganisms. There is strong evidence that the ability of thermophilic microorganisms to grow at high temperatures is based on maintaining their membrane fluidity (homeoviscous adaptation) (Sinensky, 1974). Particularly, the correct membrane function could be achieved due to the higher ratio of longer straight – chained saturated fatty acids in membrane lipids (Martins, Jurado, & Madeira, 1990; Russell & Fukunaga, 1990; Suutari & Laakso, 1994; Zeikus, 1979). Besides the above theory, another factor, that could be responsible for the growth ability of thermophiles, is the production of sufficient amounts of thermostable gene products under elevated temperature conditions. Particularly, thermophiles have a tendency to increase the purine levels at the codon positions within their genome compared to mesophiles, something which may correlate with mRNA thermostability (Cate, Gooding, Podell, & Zhou, 1996; Wang & Hickey, 2002). Likewise, the trend that cytosine is preferred over thymine in many codons could play a crucial role in the greater thermostability, maybe due to the increased number of potential formed hydrogen bonds (Querol, Perez-Pons, & Mozo-Villarias, 1996; Sadeghi, Naderi-Manesh, Zarrabi, & Ranjbar, 2006; Singer & Hickey, 2003). In addition to that, at the protein level, the increased frequency of hydrophobic and/or charged amino acids (e.g., glutamic acid, isoleucine, valine) and, at the same time the decreased frequency or removal of glutamine, which is a thermolabile amino acid, has been found to have a great effect on thermostability of the encoded proteins, probably because it reduces the possibility of the thermal unfolding process (Lynn, Singer, & Hickey, 2002; Singer & Hickey, 2003). Hence, since contamination with viable spores of thermophiles cannot be easily detected until the food is exposed to high ambient temperatures, the presence of thermophilic bacteria is especially troublesome regarding the commercial viability of canned products which are stored for extended periods of time both in local and international markets with tropical climates at temperatures frequently exceeding 30 - 35 °C (Hill & Smythe, 2012). Τhe key, therefore, to prevent spoilage by thermophilic endospore - formers is to quickly cool thermally processed products to below 40 - 43 °C and store them below 35 °C.

Literature review and thesis outline

4.1. The genus Bacillus and related genera 4.1.1. Bacillus licheniformis

Bacillus licheniformis is a motile, facultatively anaerobic, moderately thermophilic endospore - forming bacterium, capable of growing at a temperature range of 35 - 55 °C and belongs to the B. subtilis group. It is an ubiquitous microorganism in soil that is commonly isolated from spoiled canned food and from a range of dairy products (Boor & Murphy, 2002; Rückert et al., 2004; Reginensi et al., 2011; Scheldeman, Pil, Herman, De Vos, & Heyndrickx, 2005), but at low to moderate numbers due to its inability to form biofilms (Cook & Sandeman, 2000; Crielly, Logan, & Anderton, 1994; Ronimus et al., 2003). However, a recent research has proven that B. licheniformis may contaminate whey protein concentrate containing 80% protein (WPC 80) through biofilm formation on a protein - fouled surface under the conditions of a dairy manufacturing plant, causing a quality problem for the dairy industry (Zain, Flint, Bennett, & Tay, 2016). It is one of the most common Bacillus sp., prevalent in the US and Chinese dairy industries (Boor & Murphy, 2002; Buehner, Anand, & Garcia, 2014; Yuan et al., 2012). The high prevalence of B. licheniformis in powdered milk samples is potentially attributed to its widespread distribution in the environment and across the dairy farms (e.g. feed concentrate, fecal matter, soiled udders and teats, raw milk as well as from factory – derived contamination) (Rückert et al., 2004; Reginensi et al., 2011; Scheldeman et al., 2005). It is generally considered to be non - pathogenic, but some strains are reported as toxigenic and have been linked to food poisoning outbreaks associated with raw milk and industrially processed baby foods (Brown, 2000; Salkinoja-Salonen et al., 1999; Sorokulova et al., 2003). Certain strains have the ability to produce spoilage enzymes showing high proteolytic, lypolytic and glycolytic activity (Carlin, 2011; De Jonghe et al., 2010; Salkinoja-Salonen et al., 1999), while other strains are capable of producing a slimy extracellular substance that can affect the quality of pasteurized milk and cream. It is also well - established that the different stages of milk powder processing may create conditions that differentially select for different contaminants (Reginensi et al., 2011). Interestingly, the microorganisms that prefer higher temperatures (and high aw) would be found in the first evaporator effects and those that prefer lower growth temperatures tend to be found in the lower, further downstream evaporators (Ronimus et al., 2003). Thus, highly resistant endospores of B.

Chapter 1 licheniformis can survive pasteurization and they are known to contaminate products at final stages of industrial processes; despite that, they are not viewed as so much of a concern as spores from Geobacillus spp. or Anoxybacillus flavithermus (De Clerck & De Vos, 2002; Heyndrickx & Scheldeman, 2002).

4.1.2. Bacillus coagulans

Bacillus coagulans is a facultative anaerobe, motile, rod - shaped, moderately thermophilic (temperature range of 30 - 60 °C) and acid - tolerant bacterium, with an optimum pH of 5.5 - 6.5. Its cells can be found singly, in pairs or in chains of variable length and rarely as filaments. This species differs from other bacilli by the position of their ellipsoidal spores, which are located terminally in one of the cellular poles (De Clerck, Marina Rodriguez-Diaz, Gillian Forsyth, et al., 2004). The etymology of coagulans refers to the curdling or coagulating capacity of this species due to its ability to ferment lactose and produce large amounts of lactic acid (L+) without gas formation, resulting in low acidity spoilage (Nakamura, Blumenstock, & Claus, 1988; Palop, Raso, et al., 1999; Vercammen, Vivijs, Lurquin, & Michiels, 2012; Wang et al., 2009). Hence it is implicated in spoilage of UHT and canned dairy products (Robinson, 2005), moderately - acid canned vegetables (like tomato products) and other canned vegetables (Oomes et al., 2007), and fruits (Anderson, 1984; Cosentino, Mulargia, Pisano, Tuveri, & Palmas, 1997; De Clerck et al., 2004). In low - pH foods, B. coagulans is also able to alter the food pH value so that surviving Clostridium botulinum spores can germinate. However, lactic acid and some other of its metabolites like the antimicrobial peptide coagulin and thermostable enzymes (β - galactosidase, chitosanase, etc.) are exploited industrially (Batra, Singh, Banerjee, Patnaik, & Sobti, 2002; Payot, Chemaly, & Fick, 1999; Yoon et al., 2002). In addition, due to its ability to form spores, some strains of B. coagulans resist high temperature and the harsh human gastric environment. According to some recent studies, this has made these strains an area of interest for probiotic benefits delivering in human body (Majeed et al., 2016), alone or combined with lactobacilli or bifidobacteria, vitamins (particularly B complex), minerals, hormones and prebiotics or in animal feed, as an alternative to antibiotics (De Clerck, Marina Rodriguez-Diaz, Gillian Forsyth, et al., 2004).

Literature review and thesis outline

4.1.3. Bacillus subtilis

Bacillus subtilis is an aerobe, moderately thermophilic (temperature range of 5 - 55 °C) and acid - tolerant (pH of 5.5 - 8.5) bacterium. It characteristicaly produces an endospore per cell, that is located centrally in a non - swollen sporangium. Its heat - resistant spores often pose a challenge to the thermal efficacy of heat processes, resulting in reduced shelf - life of many processed foods. B. subtilis is associated with the breakdown of protein, thus resulting in the production of nonacid (sweet) curd, browning and bitter taste in yogurt due to proteinase activity (Mistry, 2001). It has also been related to the ropiness defect in raw and pasteurized milk, as well as with the spoilage of UHT and canned milk products (Heyndrickx & Scheldeman, 2002).

4.1.4. Bacillus sporothermodurans

Bacillus sporothermodurans is a mesophilic, strictly aerobic endospore - former which produces highly heat - resistant spores (Pettersson, Lembke, Hammer, Stackebrandt, & Priest, 1996) causing non – sterility problems in canned foods (Oomes et al., 2007). It is frequently found indirectly responsible for defects in UHT - treated milk products, processed milk and affected milk products ranged from whole, skimmed, evaporated, or reconstituted UHT milk to UHT cream, chocolate milk, and milk powders (Pettersson et al., 1996). The spores survive the heat process and then multiply to a maximum of about 105/ml of milk during incubation at 30 °C for 5 days, but generally cause no noticeable spoilage and are non - pathogenic (Brown, 2000; Pettersson et al., 1996; Scheldeman et al., 2006; Scheldeman et al., 2005). Although no major sensory changes occur during their growth in milk, contamination levels, particularly in UHT - treated milk, frequently exceed the criterion of the EU regulations (Scheldeman et al., 2006), resulting in considerable economic losses. These spores germinate during storage in UHT products, causing instability due to their proteolytic activities and thereby reducing shelf - life and consumer acceptability. Spoilage due to B. sporothermodurans growth can be in the form of a slight change in the color (pink) of milk, off - flavors and destabilization of casein micelles. If such heat resistant spores would enter the manufacturing process in large numbers through raw milk, a very severe thermal process would be required to achieve microbiological stability (Oomes et al., 2007). This, however, would lead to protein denaturation, Maillard reactions and lactose isomerization.

Chapter 1

4.1.5. The genus Geobacillus

The genus Geobacillus was established by Nazina et al. (2001) and derived from 16S rRNA genetic Group 5 of the genus Bacillus, as defined in the study of Ash et al. (1991). The genus Geobacillus includes rod - shaped, chemo - organotrophic, aerobic or facultatively anaerobic using oxygen as the terminal electron acceptor, replaced by nitrate in some species, neutrophilic with a growth range of pH 6.0 - 8.5 (optimum at pH 6.2 - 7.5), Gram - positive, endospore - forming bacteria that are widely distributed in nature. They can be found either singly or in short chains, are motile by means of peritrichous flagella, and may produce pigments on certain media (Nazina et al., 2001). They produce catalase and are usually oxidase - negative. They are obligately thermophilic (Manachini et al., 2000), with a growth temperature range of 37 - 75 °C and an optimum growth temperature of 55 - 65 °C. Geobacillus spp. is characterized by the production of a single ellipsoidal or cylindrical endospore per cell, located terminally or subterminally in a slightly swollen or non - swollen sporangium. They produce several metabolites that have been patented for commercial use, like various enzymes (catalase, acetate kinase, α - amylase, DNA polymerase), ethanol and others. Geobacillus strains also produce amylases, which may degrade starches used as thickening agents for sauces. Some species (mainly G. stearothermophilus) are known to be among the most common contaminants of milk powders and may cause thermophilic flat - sour spoilage of canned foods stored at temperatures above 43 °C. The mol % G+C content in the DNA of Geobacillus ranges between 48.2 and 58%. All species within the genus are very closely related, with similarity levels of 16S rDNA sequences in the range 96.0 – 99.4% (Coorevits et al., 2012; Nazina et al., 2001), but G. stearothermophilus is the type species isolated from milk powder (Flint, Ward, & Walker, 2001; Rückert, Ronimus, & Morgan, 2004; Ronimus et al., 2003). The Geobacillus genus forms a phenotypically and phylogenetically coherent group of thermophilic bacilli that were reclassified as a new genus according to DNA - DNA reassociation studies, 16S rDNA gene sequence analysis and physiological characteristics (Nazina et al., 2001). The Geobacillus genus includes various species formerly included in the genus Bacillus, such as G. stearothermophilus (Nazina et al., 2001) isolated from milk powder (Caspers et al., 2016), G. kaustophilus (Nazina et al., 2001), G. thermocatenulatus (Nazina et al., 2001), G. thermodenitrificans (Nazina et al., 2001), G. thermoglucosidans (Nazina et al., 2001) obtained from processing lines and milk powder in the Netherlands (Caspers et al., 2016; Zhao, Caspers, Abee, Siezen, &

Literature review and thesis outline

Kort, 2012) and G. thermoleovorans (Nazina et al., 2001). Isolates of Geobacillus spp. have been also obtained from temperate areas, hot springs, oilfields, deep sea sediments, sugar beet juice and dairy products (Banat, Marchant, & Rahman, 2004; Nazina et al., 2001; Ronimus et al., 2003; Scott et al., 2007; Tai et al., 2004; Zeigler, 2014).

4.1.5.1. Geobacillus stearothermophilus

Geobacillus stearothermophilus, which includes strains known, at various times, as B. calidolactis and B stearothermophilus var. calidolactis, has been associated with dairy products since at least the 1950s, when it was found to cause contamination issues with ultra ‐ high ‐ temperature (UHT) ‐ treated (134 - 145 °C for 1 - 10 s) dairy products. The optimal growth temperature of strains of G. stearothermophilus isolated from milk powder is approximately 63 °C (Knight, unpublished). Particular strains of Geobacillus spp. can produce extremely heat ‐ resistant spores that can survive UHT treatment and retorting. Due to the strong heat - resistance of G. stearothermophilus endospores, they are frequently used as an indicator for the verification of heat sterilization processes (Iciek et al., 2006; Viedma, Abriouel, Omar, López, & Gálvez, 2010). G. stearothermophilus is responsible for ca. 35% of canned food spoilage during incubation at 55 °C (André et al., 2013). It has been also identified in deteriorated canned vegetables (green beans, green peas, sweet corn), ready - made meals containing meat, fruit preparations, dehydrated ingredients, gelatin extracts and processing facilities (André et al., 2013; Coorevits et al., 2012; Postollec et al., 2012). For instance, up to 50% milk powders may contain the bacterium, which may represent more than 10% of thermophilic isolates (Rückert et al., 2004) and G. stearothermophilus spores were detected in 8.6% and 2.1% of raw carrots and green bean samples, respectively. Hence, G. stearothermophilus is a proven problematic endospore - forming bacterium encountered in whole and skim milk powders produced across the world, as it can cause long term persistent contamination of dairy processing facilities, due to its ability to form biofilms on stainless steel surfaces of processing equipment (Burgess, Flint, & Lindsay, 2014; Flint, Bremer, & Brooks, 1997; Rückert et al., 2004). However, G. stearothermophilus has been typically linked to ‘flat - sour’ spoilage in relation to low - acid foods (pH > 4.5), mainly of evaporated milk, due to acid coagulation at a pH level of around 5.2 (Boor & Murphy, 2002; Kakagianni, Gougouli, & Koutsoumanis, 2016; Kalogridou-Vassiliadou, 1992) in tropical climates where

Chapter 1 ambient temperatures may allow G. stearothermophilus growth. Significant spoilage causes include under - processing, inadequate cooling, contamination of the product resulting from leakage through seams, and pre - process spoilage (Jay, 2012). According to the studies of Hill and Smythe (2012) and Yoo et al. (2006), G. stearothermophilus cells produce acid without gas formation from saccharides, thus enhancing the formation of protein aggregates. This is related to the unfolding and gelation of β - lactoglobulin which has been found to be pH and temperature dependent. There are several studies (Ashton & Busta, 1968; Cleverdon, Pelczar Jr, & Doetsch, 1949; Stahl & Ljunger, 1976) supporting that the presence of divalent cations (such as calcium, magnesium and iron(II)) and vitamins (like niacin and biotin) in milk have considerable effect on G. stearothermophilus’ growth and further on spoilage of evaporated milk. Arancia et al. (1980) reported that the presence of calcium cations stimulated Escherichia coli growth and reduced lag periods, while later findings of Jurado et al. (1987) confirmed the above case for G. stearothermophilus and also demonstrated that magnesium cations above a critical concentration exert an inhibitory effect on the growth of the microorganism. Strains of G. stearothermophilus associated with dairy products cannot be differentiated from the type strain of G. stearothermophilus (ATCC 12980 = DSM 22) based on partial 16S rDNA sequencing (Burgess et al., 2014), but they can be differentiated when analyzed by molecular biology ‐ based techniques, such as random amplified polymorphic DNA PCR (RAPD - PCR) profiling, restriction fragment length polymorphism (RFLP) and the internal transcribed spacer (ITS) region, and by phenotypic characterization using biochemical testing. Notable phenotypic properties of dairy strains of G. stearothermophilus include the ability to utilize lactose and to grow under anaerobic conditions (Flint et al., 2001; Ronimus et al., 2003). Burgess et al. (2014) and Flint et al. (2001) also demonstrated, by molecular and biochemical results, that dairy isolates of G. stearothermophilus can vary in their ability to form biofilms, both in terms of numbers and morphology of the biofilm. However, Flint et al. (2001) did not determine whether strains of G. stearothermophilus could be differentiated by other characteristics or grouped based on their ability to form biofilms using a typing method.

4.1.6. The genus Alicyclobacillus

There are many factors such as microbial development, phenol, pectin, protein and starch etc. that can make a juice turbid or cause it to precipitate. Although the

Literature review and thesis outline low pH of acidic foods and beverages, such as fruit products and fruit juice, serves as a natural control measure against spoilage, there are a few microorganisms that can survive in the acidic environment, thus contribute to juice spoilage. The Alicyclobacillus genus (Wisotzkey, Jurtshuk JR, Fox, Deinhard, & Poralla, 1992) composes of a Gram - positive group of rod – shaped, thermo – acidophilic, strictly aerobic, heterotrophic, nonpathogenic and endospore - forming bacteria, that originally consisted of 20 species, 2 subspecies and 2 genomic species (Nakano, Takahashi, Tanaka, & Okada, 2015). The type species is Alicyclobacillus acidocaldarius (Wisotzkey et al., 1992). Alicyclobacillus spp. can grow in a pH range from 2.0 to 6.0 as well as at a temperature range from 23 °C to 70 °C, with the optimum growth temperature being 45 – 50 °C (Yokota, Fujii, & Goto, 2008). The inherent heat - resistant and acid - tolerant properties of Alicyclobacillus spp. allow their spores to survive in fruit juices even after the conventional hot - fill - hold pasteurization process (86 - 96 °C, 2 minutes), then germinate and increase to high cell concentrations at ≥ 20 °C, having the potential to produce offensive flavors and spoil the shelf stable products (Huang, Yuan, Guo, Gekas, & Yue, 2015; Maldonado, Belfiore, & Navarro, 2008; Smit, Cameron, Venter, & Witthuhn, 2011). Two species of this genus, A. acidoterrestris and A. acidocaldarius, have been reported to cause juice spoilage, manifested as an off - flavor and light cloudiness (Cerny, Hennlich, & Poralla, 1984; Pettipher, Osmundson, & Murphy, 1997; Walls & Chuyate, 2000). However, there may be new species of Alicyclobacillus that can spoil juice. All Alicyclobacillus species form one phylogenetic cluster (exclusive of the species Sulfobacillus disulfidooxidans) differed from any other Bacillus species based on their distinct 16S rRNA gene sequence analysis, while they possess unique ω - alicyclic fatty acids (ω - cyclohexane or ω - cycloheptane fatty acids) as the major cellular membrane lipid component (Wisotzkey et al., 1992). Some researchers have suggested that the unique ω - alicyclic fatty acids contribute to the heat resistance and thermo - acidophilic nature of these microbes, but it is not an indispensable condition for survival. It has been found that the presence of the cyclohexane ring increased the acyl chain density, leading to a denser packing of the lipids in the membrane core, structural stabilization of the membrane, lower membrane fluidity and reduced permeability via strong hydrophobic bonds. This may contribute to the maintenance of the membrane’s barrier function, protecting the microbes against acidic conditions and high temperatures (Chang & Kang, 2004; Smit et al., 2011). Furthermore, a number of species also contain hopanoids in their membranes (Cerny et al., 1984; Hippchen, Röll, & Poralla, 1981; Poralla, Kannenberg, & Blume,

Chapter 1

1980). The hopane ring is structurally similar to cholesterol, which is known to affect membrane lipid organization. It has been shown that the hopane glycolipids have a condensing effect on the cell membrane due to a decrease of the acyl chains lipids’ mobility and the membrane stabilization. This condensing action is also advantageous at low pH, since it hinders the passive diffusion of protons through the membrane, thereby facilitating the establishment of an approximately neutral cytoplasmic pH (Poralla et al., 1980). The Alicyclobacillus species have been isolated from natural sources such as hot springs and soil (Albuquerque et al., 2000; Nicolaus et al., 1998; Tsuruoka et al., 2003), as well as secondary spoiled fruit - based beverages (Duong & Jensen, 2000; Keiichi Goto et al., 2002; Matsubara et al., 2002; Osopale, Witthuhn, Albertyn, & Oguntoyinbo, 2016; Yamazaki, Teduka, & Shinano, 1996; Zhang, Yue, & Yuan, 2013). The most likely route of contamination is that fruits are contaminated during harvest and then subjected to manufacturing processes without proper cleaning. Another possibility is that soil can be carried into the manufacturing facilities by employees. In addition to garden soil, water has also been proposed as another important vector of contamination in the processing environment (Chen et al., 2006; Groenewald, Gouws, & Witthuhn, 2009; Zhang et al., 2013), especially when recycled water is used in the production of fruit juices and concentrates (Jensen, 2000; Walls & Chuyate, 2000). Spoilage caused by Alicyclobacillus spp. is classified as flat sour spoilage, which has no typical gas production with or without cloudiness that can swell the container and appears to be slow and often unnoticed (Pettipher et al., 1997; Yamazaki et al., 1996). Consequently, juice processors do not notice spoilage occurring until consumer complaints are received, which leads to the damage of company image and loss (Gobbi et al., 2010). Factors known to affect spoilage include Alicyclobacillus cell concentration, incubation temperature, heat - shock treatment, growth medium, and oxygen availability (Chang & Kang, 2004). Spoiled fruit juices exude an undesirable off - flavor or - odor, described as medicinal, antiseptic, phenolic, hammy, or smoky. In fact, the off - flavor or - odor is mainly imputed to the chemical compounds guaiacol (2 - methoxyphenol), 2,6 - dibromophenol (2,6 - DBP), and 2,6 - dichlorophenol (2,6 - DCP) (Gocmen, Elston, Williams, Parish, & Rouseff, 2005; Orr, Shewfelt, Huang, Tefera, & Beuchat, 2000; Walker & Phillips, 2005). Therefore, it is essential for juice processors to detect these bacteria or their metabolites as early as possible (Huang et al., 2015). Pettipher et al. (1997) and Gocmen et al. (2005) reported that when the cell concentration was above 105 CFU/ml, plenty of guaiacol

Literature review and thesis outline was produced to cause detectable off - flavor, at a level where the spoilage could not be visibly detected. The most widely accepted cell concentration to produce detectable taint is 104 or 105 CFU/ml (Bahçeci, Gökmen, & Acar, 2005). The production rate of taint tends to increase as incubation temperature increases and time elapses (Bahçeci et al., 2005; Jensen, 2000; Pettipher et al., 1997; Siegmund & Pöllinger-Zierler, 2007), but spore germination and outgrowth of Alicyclobacillus spp. is inhibited below 20 °C (Chang & Kang, 2004; Jensen & Whitfield, 2003). Appropriate heat - shock treatment can activate and encourage the germination of dormant endospores, allowing the generation of vegetative cells and production of taints. Various heat - shock treatments have been suggested for the activation of Alicyclobacillus spp. endospores, including 70 °C for 20 min (Eiroa, Junqueira, & Schmidt, 1999), 80 °C for 10 min (Walls & Chuyate, 2000) and 80 °C for 20 min (Terano, Takahashi, & Sakakibara, 2005). On the other hand, different fruit juices contain different types and amounts of taint precursors so that the production of taints varies among different types of juices. It is worth noting that some juice products contain growth inhibitors such as high sugar content, polyphenols and ethanol (Spllttstoesser, Churey, & Lee, 1994; Walls & Chuyate, 2000) which could restrain the germination of spores and the growth of vegetative cells. Furthermore, most species of Alicyclobacillus are aerobic so that a limited oxygen supply could slow down the growth rate of vegetative cells; however, taint production could not be completely suppressed at low oxygen concentration and even exceed those with free oxygen supply (Siegmund & Pöllinger-Zierler, 2007).

4.1.6.1. Alicyclobacillus acidoterrestris

Endospore - forming microbes were traditionally thought not to be of concern in the spoilage of fruit juices, as the majority of endospore - formers cannot survive in the acidic environment after endospore germination. Because of this, traditionally, pasteurized acidic fruit juices with pH values below 4.0 have been considered unlikely to support the germination and growth of spore - forming bacteria leading to spoilage (Walker & Phillips, 2005). However, in 1982, a large - scale commercial spoilage event in shelf - stable, aseptically packaged apple juice was detected by a thermo - acidophilic endospore - forming bacterium during distribution and storage in an unusually long warm season in Germany (Cerny et al., 1984). Since then, manufacturers and processors in the fruit industry have recognized the substantial economic and spoiling potential of A. acidoterrestris

Chapter 1 worldwide (Duong & Jensen, 2000; Jensen & Whitfield, 2003; Spllttstoesser et al., 1994; Yamazaki et al., 1996). The responsible microbe was ultimately identified as a strain of A. acidoterrestris of the genus Alicyclobacillus that could grow over the pH range from 2.5 to 5.8 (optimum: 4 – 5) and at temperatures of 20 to 70 °C (optimum: 36 – 53 °C) (Hippchen et al., 1981; Wisotzkey et al., 1992; Yamazaki et al., 1996). A. acidoterrestris living in hot environments with lower pH must have possessed an effective defense and adaptation mechanism involved in these stress responses, in which the molecular chaperone DnaK1 is a key factor (Jiao, Ran, Xu, & Wang, 2015). Alongside, ω – cyclohexyl - fatty acids are the major lipid components of A. acidoterrestris membranes and are associated with the exceptional resistance of the organism to acidic conditions and high temperatures (Hippchen et al., 1981). Among all Alicyclobacillus spp., A. acidoterrestris is the most frequently detected species responsible for spoilage and has been the center of study for a long time (Huang et al., 2015). Though spoilage by A. acidoterrestris was previously regarded as sporadic, the survey by the National Food Processors Association (NFPA) in 1998 showed the large scale of fruit juice spoilage potential associated with this bacterium. Spoilage may not be widespread, but when it appears, it can result in a variety of undesirable reactions within the product that alter the organoleptic characteristics and thus cause consumer rejection of high - acid, shelf - stable fruit products and a great loss to the manufacturer (Silva & Gibbs, 2004; Silva, Tan, & Farid, 2012; Walker & Phillips, 2005). Especially, spoilage by A. acidoterrestris is difficult to be visually detected before consumption, since it is not associated with gas or acid production, although the flat - sour spoiled products may show discoloration or increased cloudiness and sediment formation at the bottom of the package (Bevilacqua, Sinigaglia, & Corbo, 2009; Walker & Phillips, 2005); however, most of the times, the spoiled juice appears normal or has only a light sediment (Gocmen et al., 2005; Walker & Phillips, 2005). The only evidence of an occurring spoilage is an offensive smell described as ‘smoky’, ‘medicinal’, ‘phenolic’, ‘disinfectant’ or ‘antiseptic’ taint in acidic beverages attributed to production, predominantly of the chemical metabolite guaiacol possibly, during ferulic acid2

1 DnaK, also known as Hsp70 (heat shock protein 70 kD), is involved in protein refolding of the domain bacteria that have been improperly assembled or denatured and its expression levels are up - regulated by elevated temperatures and acid and salt stress.

2 Ferulic acid is a ubiquitous aromatic phenolic compound in nature and a common compound of the structural plant cell wall polymer (lignin).

Literature review and thesis outline metabolism via vanillin (Bahçeci et al., 2005; Bevilacqua et al., 2009; Gocmen et al., 2005; Pettipher et al., 1997). However, Cai et al. (2015) proposed an alternative formation pathway of guaiacol from vanillic acid and vanillin, present in the juice as derivatives of lignin or of ferulic acid degradation, based on resting cell studies and enzyme assays. According to their results, the four potential precursor substrates, including catechol, ferulic acid, tyrosine and phenylalanine, could not be metabolized to guaiacol by all of the 30 guaiacol - producing A. acidoterrestris strains tested. Owing to these facts, the precursors of guaiacol, vanillin and vanillic acid are prerequisites to produce guaiacol by Alicyclobacillus spp. (Van der Merwe, 2011). The lower limit of guaiacol detection in fruit juices by a trained sensory panel is 2 μg/l (2 ppb). Detectable taint production in fruit juices is generally reported when the levels of A. acidoterrestris reach about 104 - 105 CFU/ml (Bahçeci et al., 2005; Komitopoulou, Boziaris, Davies, Delves‐Broughton, & Adams, 1999; Pettipher et al., 1997; Sinigaglia et al., 2003). This is very detrimental for producers because it often brings about product recall from the market, with consequent economic and image damages for the company (Gobbi et al., 2010). A. acidoterrestris has been detected in a wide range of commercially pasteurized fruit juices, bottled tea, isotonic drinks and other low - pH, shelf - stable products, as well as processing facilities, where it enters most likely via fruit surfaces contaminated with soil during production and harvesting (Eiroa et al., 1999; Merle, 2012). The heat resistance of the spores is such that pasteurization will not guarantee the organism’s elimination. A. acidoterrestris spores survive pasteurization regimes typically employed in the juice industry, and, given their ability to grow at low pH and its moderate thermophilic nature, they then germinate, grow and spoil the acidic products stored at warm temperatures (Alpas, Alma, & Bozoglu, 2003; Bevilacqua et al., 2009; Pettipher et al., 1997; Silva et al., 2012). For this reason, this microorganism should be designated as a possible quality control target for the design and optimization of adequate thermal treatments of pasteurization in acidic beverages (Murakami, Tedzuka, & Yamazaki, 1998; Silva, Gibbs, Vieira, & Silva, 1999; Silva & Gibbs, 2001; Smit et al., 2011; Vieira, Teixeira, Silva, Gaspar, & Silva, 2002), while it is also important for manufacturers to control growth of A. acidoterrestris in the processing environment of acidic beverages. Thorough washing of the fruit prior to processing, optimization of the traditional thermal processing and the use of chemicals or natural antimicrobials are the control methods traditionally used to inhibit this microorganism at present (Vieira et al., 2002). However, the growth of A. acidoterrestris in fruit juices can be directly or

Chapter 1 indirectly influenced by storage temperature, pH and available oxygen. Apart from the above, other factors that can impact growth are the nutrient composition, their availability and naturally occurring antimicrobial compounds present in the fruit juice (Orr et al., 2000; Walker & Phillips, 2005; Walker & Phillips, 2008). Nevertheless, the fruit juice pH and the temperature during storage and distribution (Giannakourou, Koutsoumanis, Nychas, & Taoukis, 2005; Koutsoumanis, 2001; Koutsoumanis, Pavlis, Nychas, & Xanthiakos, 2010; Koutsoumanis, Stamatiou, Skandamis, & Nychas, 2006) are among factors that deeply influence the initiation of growth of A. acidoterrestris spores. Research data support that germination, outgrowth and subsequent vegetative growth of A. acidoterrestris spores would not be expected to occur when pasteurized fruit products with naturally low pH are stored below 20 °C, wherein the formation of guaiacol is suppressed (Bahçeci & Acar, 2007; Bahçeci et al., 2005; Chang & Kang, 2004; Jensen & Whitfield, 2003; Spinelli, Sant'Ana, Rodrigues-Junior, & Massaguer, 2009). However, the conditions prevailing in the supply chain of the pasteurized fruit juices are out of the direct control of the manufacturer and often deviate from specifications. In particular, the storage of fruit juices for long periods at improper and changing temperature conditions may provoke the germination of spores, if these are present, the outgrowth and the subsequent growth of the vegetative cells of the organism (Orr et al., 2000). These conditions prevail in warehouses, delivery trucks, retail display, storage rooms and home storage (Bahçeci et al., 2005; Chang & Kang, 2004; Heyndrickx, 2011; Pettipher et al., 1997), especially during warmer months or in tropical and semitropical regions (Roig-Sagues, Asto, Engers, & Hernández-Herrero, 2015), where the temperature is relatively high (above 30 – 40 °C). Refrigeration temperatures for shelf - stable products traditionally stored at ambient temperatures would probably be effective (Siegmund & Pöllinger-Zierler, 2007), but the institution of such a control measure would present a major additional cost factor in production and distribution. Obviously, A. acidoterrestris is a concern for the quality of fruit juice and the estimation of the risk of spoilage constitutes a major target of the quality managers of the food industry, especially for the products that are going to be distributed in hot climate countries. On the other hand, there is no evidence that the organism is pathogenic and the consumption of products containing Alicyclobacillus may not pose a health or safety risk (Walls & Chuyate, 2000), and since it does not change the pH of the fruit, there should be no risk of secondary growth of endospore - forming pathogens such as C. botulinum (Brown, 2000).

Literature review and thesis outline

4.1.7. The genus Anoxybacillus 4.1.7.1. Anoxybacillus flavithermus

Anoxybacillus flavithermus (formerly B. flavothermus) was firstly isolated from a hot spring in New Zealand and was described as a Gram ‐ positive, facultative anaerobic, motile, thermophilic, rod ‐ shaped, endospore ‐ forming bacterium producing terminal endospores (Flint et al., 2001; Heinen, Lauwers, & Mulders, 1982). The G + C content ranges between 41.6 and 61 mol %, the temperature range for growth is between 30 and 70 °C and the optimum growth temperature, under aerobic conditions, is 50 - 62 °C (Flint et al., 2001; Heinen et al., 1982; Pikuta et al., 2000). The pH range for growth is 5.5 – 9.0, with an optimum at pH 7 (Flint et al., 2001). This species was transferred to the genus Anoxybacillus (and its epithet corrected to flavithermus), alongside the newly described species, A. pushchinoensis (Pikuta et al., 2000). Strains of A. flavithermus have subsequently been isolated from food gelatin batches, heat - processed dairy products (milk powder) and biofilms from food processing environments resulting in factory - derived contamination (Burgess, Brooks, Rakonjac, Walker, & Flint, 2009; Caspers, Boekhorst, Abee, Siezen, & Kort, 2013; Caspers et al., 2016; De Clerck et al., 2004; Flint et al., 2001; Postollec et al., 2012; Ronimus et al., 2003; Rueckert, Ronimus, & Morgan, 2005; Scott et al., 2007). Isolates from milk powder have an optimum growth temperature ranging between 50 and 65 °C (Caspers et al., 2013; Ronimus et al., 2003). In addition, A. flavithermus was not reported in raw milk or other sources of contamination of the dairy environment. As a source of contamination unlikely to be derived from dairy farms, this species has predominantly been found in hot spring habitats (Heinen et al., 1982).

4.1.8. The genus Brevibacillus 4.1.8.1. Brevibacillus bortelensis

Low levels of B. borstelensis were detected in whole milk and skim milk powders in a Chinese study looking at whole, skim and infant milk powders (Yuan et al., 2012). Hence, Brevibacillus spp. is often encountered in milk and milk products globally, albeit at low levels (Gopal et al., 2015).

Chapter 1

4.1.9. The genus Paenibacillus

The Paenibacillus genus is part of the Paenibacillaceae family and currently comprises of 154 named species. The type species is Paenibacillus polymyxa. P. polymyxa, together with P. macerans, is one of the sporeformers capable of germination and growth at low pH (between pH values of 3.8 and 3.9) and routinely found in heat - treated acidic foods, low - acid canned products or dairy products in particular. As highlighted in a review by Heyndrickx and Scheldeman (2002), low numbers of Paenibacillus spores can be found in both raw and pasteurized milk. Although there had been no previous reports of these spores surviving industrial sterilization or UHT processing of milk, P. lactis has been isolated directly from raw and UHT milk as well as from the dairy farm environment (Scheldeman et al., 2004). In that study, P. lactis was isolated in conjunction with B. sporothermodurans (Heyndrickx et al., 2012). The simultaneous isolation of these two species indicates that Paenibacillus spp. is also capable of resisting UHT and go on to have an impact on food safety and quality (Scheldeman et al., 2004). The significance of these bacteria lies not only in their spoilage potential, but also in the fact that the growth of certain strains in acid foods may lead to an increase in pH, thus allowing less acid - tolerant species, such as clostridia, to grow (Casadei, Ingram, Skinner, & Gaze, 2000).

4.2. The genus Clostridium and related genera 4.2.1. Clostridium thermopalmarium/thermobutyricum

Clostridium thermopalmarium, originally derived from palm wine in Senegal (Soh et al., 1991), is moderately thermophilic, anaerobic, rod - shaped with an optimal growth temperature of 50 - 55 °C and pH of about 6.6, which produces butyric acid as the main end product of sugar metabolism. It is genetically close to C. thermobutyricum, but with distinct physiological traits (Wiegel, Braun, & Gottschalk, 1981). This species was identified exclusively in products containing fatty duck. It represented the main bacterium involved in the spoilage of this food category and was never isolated in other food categories.

Literature review and thesis outline

4.2.2. Clostridium thermosaccharolyticum

The most heat - resistant spores are those of C. thermosaccharolyticum. D values as high as 195 min at 121 °C have been recorded. A spoilage event in canned mushrooms was caused by heat - resistant spores of C. thermosaccharolyticum that had grown in the composted forest bark used on the mushroom beds. Spoilage from this organism manifests itself by blown or burst packs with a strong butyric or cheesy odor. The spores survive thermal processing to germinate and grow when the product is stored at elevated temperatures around 30 - 60 °C (e.g. in pallets of inadequately cooled cans). Spoilage by C. thermosaccharolyticum is not uncommon (Brown, 2000).

4.2.3. The genus Moorella

The Carlier and Bedora - Faure (2006) study was the first to report Moorella species isolated from spoiled cans. The genus Moorella was described by Collins et al. (1994) in the mid ‐ 1990s to accommodate two thermophilic organisms formerly classified as Clostridium. This was based on descriptions of Clostridium thermoaceticum (Fontaine, Peterson, McCoy, Johnson, & Ritter, 1942) and Clostridium thermoautotrophicum (Wiegel et al., 1981). C. thermoautotrophicum was originally differentiated from C. thermoaceticum by its ability to grow chemolithotrophically or with methanol, and by the sporulation frequency as well as the location of spores. C. thermoaceticum, unlike C. thermoautotrophicum, was non - motile and did not use arabinose. However, these two species seem to be closely related as evidenced by: (i) DNA - DNA similarity (Wiegel et al., 1981); (ii) presence of the same menaquinone (MK - 7) molecule in the membranes of both bacteria (Das, Hugenholtz, Van Halbeek, & Ljungdahl, 1989); (iii) cellular fatty acid composition; and (iv) the same diaminopimelic acid isomer type in the peptidoglycan (Yamamoto, Murakami, & Takamura, 1998). Moreover, it was shown that most strains of M. thermoacetica (originally isolated as C. thermoaceticum) are able to grow chemo - lithotrophically or with methanol (Collins et al., 1994; Daniel, Hsu, Dean, & Drake, 1990). More recently, two strains with physiological differences were classified as M. thermoacetica based on their 16S rDNA sequence similarity (99.8%) and sporulation frequency, although their sequence similarity with M. thermoautotrophica species was also very high 99.5% (Byrer et al., 2000). From a genetic point of view the food -

Chapter 1 borne Moorella strains and the type strains of M. thermoacetica and M. thermoautotrophica appeared to be closely related or even identical. The unique discrepancy with the above genetic data was found with the 16S - 23S rDNA ISR patterns. Indeed, both type strains M. thermoacetica and M. thermoautotrophica were clearly separated by two distinct PCR profiles. This hypothesis is supported by the high similarity of the tRNAile and tRNAala genes found between the strains (Carlier & Bedora-Faure, 2006).

4.2.3.1. Moorella thermoacetica

M. thermoacetica is an anaerobic endospore - former described as highly resistant to heat (Wagner & Wiegel, 2008). Its growth in canned food is reported to result in strong acidification, abnormal odors and colors, pH decrease, separation and coagulation of milk components and can swelling (Matsuda, Masuda, Komaki, & Matsumoto, 1982). Its optimal growth temperature is 55 - 60 °C, it produces acetate as the main fermentation product, and as such, is considered as a model acetogen and used for biotechnological applications (Drake & Daniel, 2004; Fontaine et al., 1942). M. thermoacetica was detected in the 1970s in Japan in hot vending machines serving canned coffee with milk, soup, and “shiruko” (sweet beverage made from azuki bean powder), all of which were kept hot at 55 to 60 °C for a prolonged storage time (Matsuda et al., 1982). M. thermoacetica has occasionally been isolated from canned vegetables (Carlier & Bedora-Faure, 2006), ready – made meals (André et al., 2013), or from specific spoiled food products such as canned coffee and “shiruko” (Matsuda et al., 1982).

4.2.4. The genus Thermoanaerobacterium

Thermoanaerobacterium spp. are thermophilic, obligately anaerobic bacteria that produce ethanol, CO2, H2 and organic acids (L - lactic acid, acetic acid) by carbohydrate and polysaccharide fermentation (Shaw, Hogsett, & Lynd, 2010). Thermoanaerobacterium spp. are detected in in spoiled canned food and they are one of the main spoilage genera in products containing fatty duck (André et al., 2013). According to Andre et al. (2013), Thermoanaerobacterium spp. were present in 8% of spoiled canned food samples collected over a period of ten years from French canners. They share similar physiological characteristics, such as anaerobic growth and an optimal growth temperature of around 63 °C, to M. thermoacetica. T.

Literature review and thesis outline thermosaccharolyticum (formerly Clostridium thermosaccharolyticum) is a thermophilic anaerobic, non - H2S gas – producing, spore former important to the spoilage of thermally processed foods. It was the dominating taxon within the genus and was recognized as a spoiler of great interest in canned food industries due to its acid tolerant nature (Dotzauer, Ehrmann, & Vogel, 2002; Mtimet et al., 2016). This species has been called “the swelling canned food spoiler” due to the abundant gas production, mainly carbon dioxide and hydrogen, along with acetic acid, butyric acid, lactic acid, succinic acid, and ethanol, during growth (Mtimet et al., 2016). According to Mtimet et al. (2016) the type strain DSM 571 is able to grow at pH ranging from 4.13 to 7.61. Except for the species T. thermosaccharolyticum, the other species were not previously described in foods.

4.2.5. The genus Thermoanaerobacter

The genus Thermoanaerobacter was firstly described by Wiegel and Ljungdahl (1981). Nowadays, this genus consists of 13 species and subspecies (Garrity, Bell, & Lilburn, 2004), which are Gram - variable, strict anaerobic, thermophilic bacteria that are capable of carbohydrate and polysaccharide fermentation producing primarily ethanol, CO2, H2 and organic acids (L - lactic acid, acetic acid). Several species form spores (Shaw et al., 2010). T. thermohydrosulfuricus has been isolated from sugar - beet juice (Carlier, Bonne, & Bedora-Faure, 2006). During quality control analysis of food processing, sometimes the laboratories of canning factories isolate anaerobic thermophilic strains. The biochemical characteristics and phylogenetic relationship of Thermoanaerobacter isolates from canning factories were investigated by Carlier and Bedora - Faure (2006) who proposed a new subspecies, Thermoanaerobacter mathranii subsp. alimentarius. While some isolates were identified as T. thermohydrosulfuricus, others were phylogenetically closely related to T. mathranii as demonstrated by 16S rRNA gene similarity values > 99%, although their phenotypes were quite distinct (Carlier et al., 2006).

4.2.6. The genus Desulfotomaculum

The genus Desulfutomaculum was described as a spoilage anaerobic thermophilic (growth temperature: 30 - 70 °C with minimum pH: 5.6), endospore -

Chapter 1 forming bacterium of canned foods, such as canned milk - containing ‘shiruko’ (a soft drink made from red beans and cane sugar produced for retail in hot - vending machines at temperatures above 50 °C), coffee or low - acid canned vegetables, such as sweet corn and peas (Matsuda et al., 1982; Sperber & Doyle, 2009). The spores of D. nigrificans are highly heat resistant and can survive typical commercial sterilization processes. D. nigrificans (formerly known as Clostridium nigrificans) causes 'sulphur stinker' spoilage often resulting in blackened product with its characteristic rotten egg odor and no gas production, when the steel in containers reacts with the dissolved H2S produced. D. nigrificans is the only sulfur - reducing organism that has been associated with spoilage of thermally processed foods.

5. Thermophilic endospore – forming spoilage bacteria and their importance in food products 5.1. Low acid foods

For many years the canning industry has produced Low Acid Canned Foods (LACF according to Codex Alimentarius (1979) by applying thermal processes which are safe and shelf - stable at ambient temperatures for several years. The thermal process not only determines the organoleptic properties of the product but is also necessary to eliminate all vegetative microorganisms and partially or totally inactivate spores. Practical interest for products produced for the European market is mainly restricted to mesophilic strains. In general, the ambient temperatures in these countries are not high enough for thermophilic spores to germinate. Good Manufacturing Hygienic Practices include stability tests at room temperatures for managing the pathogen risk related to surviving mesophilic bacterial spores in moderate climates. Spores, which are formed in products stuck onto high temperature processing lines, are of extra concern, since they will likely have a higher heat resistance. LACF are also often submitted to additional incubation conditions, typically 55 °C for 7 days, to monitor spoilage by thermophiles which are not able to grow out under ambient storage conditions. In addition, LACF’s non - stability after a prolonged 55 °C incubation, reflects insufficient control of hygiene during the end - to - end food processing chain, mainly due to: i) insufficient heat treatment and/or ii) the presence of highly heat - resistant spores on processing lines and raw materials, even at low concentrations. Therefore, global hygiene management on industrial - line processes essentially relies on surveys of the

Literature review and thesis outline thermophilic spores that contaminate food before the can sterilization step, as highlighted by Burgess (Burgess et al., 2010). Consequently, the canning industry needs better knowledge of thermophilic endospore - forming bacteria, specifically regarding the heat resistance parameters of the various strains, and their origin on processing lines, in order to ensure better process control of hygiene conditions. For this purpose, it is necessary to determine precisely the process steps where these species proliferate and to collect data on the genetic and physiologic diversity of strains so as to link raw material, multiplication sites and spoiled products (André et al., 2013).

5.1.1 Dairy products

Aerobic thermophilic endospore ‐ forming bacilli are particular contaminants of dairy products, that enter processing plants from farm environments via milk, and in some cases multiply within processing stages where conditions are suitable for bacterial growth. Thus, they can significantly affect food quality and safety. They have been detected throughout the dairy processing environments, including dairy farm environments, storage and transportation tanks, and dairy processing plants (Postollec et al., 2012). Thermophilic bacteria are defined in the dairy industry as those which are capable of growing on aerobic plate count or milk plate count agar (MPCA) during incubation at 55 °C for 48 hours. Although non pathogenic, many thermophilic endospore ‐ forming bacilli produce extracellular heat - stable hydrolytic enzymes, such as proteinases and lipases, exhibit optimum activity at temperatures between 60 and 75 °C, which, if allowed to form, produce acid during growth and can consequently have a negative impact on the sensory qualities of the final product during storage (Chen, Coolbear, & Daniel, 2004; Cosentino et al., 1997). Many species of thermophilic bacteria, belonging to the genus Bacillus and related genera, also produce endospores and biofilms, which is the main source of bacterial contamination of the final products under inadequate pasteurization/storage conditions, leading to product downgrades and losses in revenue for dairy manufacturers. Due to these reasons, thermophilic bacilli are used as hygiene indicators to control the microbial quality of milk and dairy products and provide wholesome dairy processed products to the consumers. Thermophilic bacteria can grow in different manufacturing processes, but the dairy products in which they are a most significant concern are milk powders. Milk powders constitute one of the main dehydrated ingredients of the dairy industry with a prolonged shelf - life, that

Chapter 1 are used in the manufacturing of UHT - or retort - treated products (Seale et al., 2015). Considering that especially dehydrated ingredients are traded worldwide and originate from regions with different climatic conditions, new endospore - formers that produce highly thermo - resistant spores may emerge and enter the dairy processing environments, thereby challenging the dairy manufacturers (Stoeckel, Lücking, Ehling-Schulz, Atamer, & Hinrichs, 2016). Due to their ubiquitous nature, obligately thermophilic endospore - forming bacilli are present at very low levels in raw milk (e.g. < 10 CFU/ml) used in dairy processing (Buehner et al., 2014; McGuiggan, McCleery, Hannan, & Gilmour, 2002; Scheldeman et al., 2005). However, their spores have the potential to withstand cleaning procedures and industrial heating processes, such as pasteurization and commercial sterilization treatments, due to their ability to adhere to stainless steel surfaces within the walls of processing facilities or to fouled milk deposits. Due to hydrophobicity (provided by the exosporium mainly) or physicochemical interactions (Lifshitz – van der Waal’s forces, acid/base interactions, electrostatic charge) between their surfaces and the substratum, some of the spores present germinate when and where the conditions are suitable and vegetative cells then reproduce, forming biofilms3. As the thermophilic biofilm develops, new highly heat - resistant spores may be generated from vegetative cells at the same time. Therefore, the biofilm becomes a reservoir of spores, that can result in cross contamination of milk and subsequently dairy products and get carried over to the final product with unacceptably high load of spores, thus downgrading the quality and impacting the shelf life of that product, especially when products are stored at ambient temperatures above 37 °C (Burgess et al., 2009; Burgess et al., 2010; Faille et al., 2014; Hassan, Anand, & Avadhanula, 2010; Murphy, Lynch, & Kelly, 1999; Scheldeman et al., 2006; Scott et al., 2007; Seale et al., 2012; van Loosdrecht, Lyklema, Norde, & Zehnder, 1989). Canned milk is also subject to occasional episodes of coagulation by thermophilic bacteria. Anoxybacillus flavithermus and Geobacillus stearothermophilus are the predominant obligately thermophilic endospore ‐ forming bacilli found in milk - powder products, milk powders manufacturing plants, UHT milk, and retorted products worldwide as temperature parameters for processing milk powders are high, thus allowing growth

3 Biofilms may be defined as surface associated communities of microbial cells which are adhered onto surfaces, biotic or abiotic, and are enclosed within an extracellular polymeric matrix produced by the bacteria themselves providing protection to vegetative cells and their spores from the biocidal effects of cleaning and sanitizing agents (Branda et al., 2005; Wong & O’Toole, 2011).

Literature review and thesis outline of thermophiles in evaporators and pre - heaters (Burgess et al., 2014; Burgess et al., 2010; Flint et al., 2001; Rückert et al., 2004; Ronimus et al., 2003; Scott et al., 2007; Seale et al., 2012; Trmčić, Martin, Boor, & Wiedmann, 2015; Watterson, Kent, Boor, Wiedmann, & Martin, 2014; Yuan et al., 2012). Other facultative spore ‐ forming bacilli (B. pumilus, B. licheniformis, B. coagulans, B. sporothermodurans and B. subtilis), that are able to grow at both mesophilic and thermophilic temperatures, are also found in the above products (Ronimus et al., 2003; Watterson et al., 2014; Yuan et al., 2012). Cellular adaptations, such as heat ‐ stable DNA, proteins and membranes, enable thermophilic endospore ‐ forming bacilli to grow and survive in hot environments, but also enable these microorganisms to grow, colonize and survive in all post pasteurization process steps within heated dairy processing equipment (regenerative sections of pasteurizing plate heat - exchangers, cream separators, evaporator preheaters, evaporators) (Murphy et al., 1999; Scott et al., 2007) in which the prevailing temperature is suitable for their growth and ranges between 45 and 75 °C. This is believed to occur in the form of a biofilm. However, the development of a biofilm of thermophilic bacilli within a dairy processing plant is still not well understood. Under favorable conditions, biofilms of thermophilic endospore ‐ forming bacilli will develop and both vegetative cells and spores will detach from surfaces, enter the product stream, multiply and contaminate the final product up to a level of 106 CFU/ml (Flint et al., 1997; Seale et al., 2015) jeopardizing its quality and safety. Unfortunately, so far, they are difficult to eradicate entirely by cleaning and disinfection, exceeding by far the typical specifications and thus resulting in lower values out - of - specification products (Seale, Flint, McQuillan, & Bremer, 2008) that fail to meet customer requirements. The contamination can go up when these end - products enter the recirculation loop as rework or in the manufacturing of secondary food products. Contamination of milk products by thermophilic endospore ‐ forming bacilli is not new, as they have been isolated from milk powder recovered from supplies used in an Antarctic expedition in 1907 (Ronimus, Rueckert, & Morgan, 2006). In reality, spore - mediated defects of commercially sterilized milk products are not commonly reported as customer complaints. Even so, the potential for survival after the applied heat treatment and incidents that do occasionally occur mean that the thermophilic spores cannot be overlooked as possible agents of spoilage of commercially sterilized products (Hill & Smythe, 2012). However, effective control of these bacteria in milk products and the processing environment is still a difficult

Chapter 1 task, since knowledge about their origin and food quality related characteristics, such as thermo - resistance or spoilage and toxic potential, is generally limited (Lücking, Stoeckel, Atamer, Hinrichs, & Ehling-Schulz, 2013).

5.1.2. Canned vegetables

Sealed and sterilized canned vegetables remain microbiologically stable for years at ambient temperature as the heat process inactivates mesophilic microorganisms. Spoilage in low - acid (pH > 4.5) canned vegetables occurs mainly at high incubation temperatures (> 40 °C) and it is caused by the survival and further multiplication of thermophilic spore - forming bacteria (Durand et al., 2015). Geobacillus stearothermophilus, Moorella thermoacetica/thermoautotrophica and Thermoanaerobacterium spp. are regularly identified as the most common causes of spoilage in low - acid canned foods, including canned vegetables, representing up to 75% of the species responsible for non - stability at 55 °C (André et al., 2013; Carlier & Bedora - Faure, 2006; Carlier et al., 2006). The process of canning vegetables combines several operations which result in contamination with microorganisms. Freshly - harvested vegetables are washed, trimmed and cut, then blanched by steaming or dipping in hot water for a few minutes at temperatures close to 100 °C, filled into cans, covered with hot cover brine, and finally sterilized. However, between blanching and sterilization, vegetables remain at relatively high temperatures that may allow thermophilic bacteria to grow and eventually sporulate. The range of temperatures supporting the growth of thermophilic endospore - formers, including G. stearothermophilus and M. thermoacetica (Coorevits et al., 2012; Fontaine et al., 1942), suggests that the multiplication and ultimate sporulation of thermophiles may only occur in a few niches along the processing line. In the study of Durand et al. (2015), these niches were probably out - of - flow peas with prolonged residence times, recycled brine, peas in cans, alone or mixed with carrots, with or without brine, and probably also process - line surfaces, as the temperatures were in the range 41 - 69 °C. In addition, all these niches are encountered after blanching and the blanching step may induce germination of spores. Moreover, spore adhesion to industrial materials and spore resistance to cleaning and disinfection favor spore persistence along the chain (Parkar, Flint, Palmer, & Brooks, 2001). A delay in spore release from debris and surfaces may occur, and cross - contamination may emerge long after initial contamination (Seale et al., 2008).

Literature review and thesis outline

5.1.3. Bread

Bread is a principal food for a balanced healthy diet for all consumer groups of people, including children, the elderly and patients with various illnesses, and is considered unsafe for such people when bread with a high count of Bacillus strains with unknown properties is consumed. A microbiological concern relevant to the bakery industry the rope spoilage, mainly associated to the presence of spores of Bacillus species in raw materials (Pepe, Blaiotta, Moschetti, Greco, & Villani, 2003) that can decrease the shelf - life of bread and may result in substantial monetary losses for producers. Bacillus spp. spores contaminate flours, especially wheat flour, because of soil contamination, cultivation and processing methods and may also be present in the bakery environment (atmosphere, surfaces of processing equipment) and in other raw materials such as bran, whole natural grains or seeds which could be added, brewer's yeast and bread improvers. Heat - resistant spores can survive the baking process in the center of the loaf where the temperature reaches up to 97 - 101 °C for a few minutes (Thompson, Dodd, & Waites, 1993). Under favorable conditions, (temperature above 25 °C, water activity ≥ 0.95, pH > 5) spore germination occurs and vegetative cells cause a deterioration process of bread texture, the rope spoilage. Initially, this spoilage is characterized by a distinctive sweet fruity and unpleasant odor similar to that of over - ripe melons or pineapples. Later, the loaf becomes discolored, soft and sticky to touch, making the bread inedible and, in advanced stages, can be almost liquefied due to slime formation as a result of the combined effect of proteolytic and amylolytic bacterial enzymes (Rosenkvist & Hansen, 1995; Sorokulova et al., 2003; Thompson et al., 1993). Therefore, ropiness can develop very rapidly under warm and humid conditions, becoming a remarkable monetary problem in the warm (25 - 30 °C) climates of Mediterranean countries, Africa and Australia, particularly during summer (Bailey & Von Holy, 1993). Also, a low level (≈102 spores/g) of Bacillus spp. spores in flour can lead to 107 CFU/g in bread crumb within 2 days, causing bread spoilage (Rosenkvist & Hansen, 1995). An exact quantification of economic losses is difficult to obtain because the ropy spoilage is often confused with changes in the structure of bread caused by insufficient baking or unleavened dough. Some previous studies have focused on the identification of Bacillus species commonly contaminating raw materials and bakery products and being responsible for rope spoilage. These studies led to the identification mainly of B. subtilis and B.

Chapter 1 licheniformis, probably due to their higher resistance to heat, and to a lesser extent of B. pumilus, B. amyloliquefaciens, B. megaterium and B. cereus (Bailey & Von Holy, 1993; Rosenkvist & Hansen, 1995; Thompson, Waites, & Dodd, 1998). The need for controlling survival and growth of B. subtilis and B. licheniformis in bread is obvious in order to avoid the spoilage of bread and the concomitant economic losses for the industry. The usual recommended control procedures such as sanitation of bakery equipment, stringent temperature control during baking and testing of raw materials (Bailey & Von Holy, 1993) can reduce the initial spore counts in dough, but not prevent germination and growth in bread.

6. Methods for characterization and identification of thermophilic bacilli

The taxonomy and especially the identification of thermophilic, endospore - forming bacteria have generated considerable interest over the past decades, which continuously increases due to their role in thermostable biotechnologically important products. These bacteria are associated with heat - treated foods, and, although they are not pathogenic, they can cause food spoilage by producing acids and thermostable enzymes. Thus, the reliable rapid identification and typing of thermophilic endospore - forming bacilli is very important in investigating and understanding the microbial sources of contamination during the different manufacturing processes and in developing strategies for controlling their effective control. Characterization of thermophilic bacilli can be achieved by using traditional culturing and biochemical taxonomic approaches or by modern molecular biology ‐ based approaches. Although culture - based methods have improved greatly and are usually simple to operate and cost - effective, they require multiple culture steps to achieve isolation, thus making them time - consuming and unreliable to repeat (Baumgart, 2000). Moreover, followed by a series of morphological and biochemical tests for confirmation, their detection efficiency varies greatly. Hence, rapid and robust detection methods are essential for manufacturers in order to screen raw material and products quickly, allowing early steps to be taken to eliminate and prevent contamination promptly. Several studies have been done to compare different isolation media for Alicyclobacillus spp. (Chang & Kang, 2005; Chang, Park, & Kang, 2013; Henczka, Djas, & Filipek, 2013; Huang et al., 2015; Murray, Gurtler, Ryu, Harrison, & Beuchat, 2007;

Literature review and thesis outline

Orr et al., 2000; Pettipher & Osmundson, 2000; Pettipher et al., 1997). In addition to choosing a suitable medium type, pH, incubation temperature, and some other procedures such as heat - shock treatment, membrane filtration, and pre - enrichment can also lower the detection limits. After commercial sterilization, Alicyclobacillus spp. could remain in juice products as dormant endospores. An appropriate heat - shock treatment can kill vegetative cells as well as activate spore germination. The most frequently used heat - shock treatments are either at 80 °C for 10 minutes (Walls & Chuyate, 2000) or 70 °C for 20 minutes (Eiroa et al., 1999), which are recommended by the International Federation of Fruit Juice Producers (IFU) and the Japan Fruit Juice Association (JFJA), respectively (Baumgart, 2000; IFU, 2007; Yokota et al., 2008). After the heat - shock treatment, a pre - enrichment is proposed by keeping concentrated fruit products incubated at 40 - 50 °C for 48 hours, so that low levels of endospores are easier to detect (Walker & Phillips, 2008). Besides, membrane filtration can retain and collect Alicyclobacillus spp. from fruit juice or beverages in a short time, thus enabling manufacturers to test samples of large volume and low contamination levels (Chang & Kang, 2004; Lee, Chang, Shin, & Kang, 2007). Filters containing spores are then directly spread placed onto agar plates containing the growth medium and are incubated at optimal temperatures for ≥ 18 hours to allow the spores to germinate and grow until colonies are visible. Filtration is more sensitive and has a lower detection limit than conventional spread plating, as larger samples can be passed through the filter (Chang & Kang, 2004; Henczka et al., 2013; Lee et al., 2007). Referring to direct plate technology, spread plates are superior to pour plates because the former can present colonies that are much larger and more obvious and therefore easier to observe (IFU, 2007; Pettipher et al., 1997). Biochemical test kits (e.g. Gram stains, API CHB kits) are commonly employed to characterize Gram ‐ positive endospore ‐ forming bacilli. Such kits are useful for many mesophilic species, such as B. subtilis and B. cereus, but are not very reliable for the identification of thermophilic bacilli commonly found in dairy products. Despite this, phenotypic testing, using biochemical test kits and traditional test methods, is still valuable in characterizing thermophilic bacilli. Such testing can identify characters that can be used to easily differentiate between species and can reveal properties that are relevant to growth in dairy processing equipment, such as lactose utilization and growth under anaerobic conditions (which is important, as the oxygen content of milk during evaporation is low). Such testing can also reveal the

Chapter 1 spoilage potential of strains or species, including the ability to produce extracellular enzymes such as amylases, proteases and lipases (Seale et al., 2015). Bianchi et al. (2010) have reported the application of the rapid and reliable analytical Gas Chromatography – Mass spectrometry (GC - MS) of volatile compound profile approach for the early detection of Alicyclobacillus acidoterrestris in spoiled orange juice (104 CFU/ml). The comparison of the volatile profile of uncontaminated samples with that of contaminated ones allows the detection of significant differences amongst samples. In fact, it has been assessed that the volatile profile of food, depending both on the nature and the relative amount of volatile compounds, represents a fingerprint of the product that can be used not only to assess food quality but also to detect adulteration, including microbial spoilage. Another type of spectroscopic technique, the Fourier Transform Infrared Spectroscopy (FT - IRS) can be used to correctly identify pure as well as mixed cultures of several spoilage - causing Alicyclobacillus spp. and human pathogenic Escherichia coli microbes in fruit juice samples. It is based on the unique spectral features of various components of the microbial cells with a sensitivity range of 103 - 104 CFU/ml (Al-Qadiri, Lin, Cavinato, & Rasco, 2006; Al‐Holy, Lin, Alhaj, & Abu‐Goush, 2015). FT - IRS is a detection method, used for classification and identification of bacteria with minimal sample preparation. It captures subtle differences in biochemical characteristics of various components of the cell wall, membrane (phospholipid bilayer, peptidoglycan and lipopolysaccharides) and cytoplasm (water, fatty acids, proteins, polysaccharides and nucleic acids) to present distinct spectral features at 400 - 4000 cm−1. Hence, subtle differences in the chemical composition of microbial cells contribute to a typical and unique IR spectral fingerprint (Al‐Holy et al., 2015). Despite that the FT - IRS method is rapid and reliable, it is expensive, is still in the development phase, and requires a comprehensive spectral reference database for the identification of other unclassified Alicyclobacillus strains (Huang et al., 2015). Enzyme - linked immunosorbent assay (ELISA) is a popular biochemical assay that utilizes antibodies and color change to detect and identify a substance, basically an antigen, in a liquid or wet sample. ELISA has been widely employed as a sensitive, simple, rapid, cost - effective and reliable diagnostic tool in medicine and plant pathology as well as foodborne microorganisms. Wang et al. (2012) obtained a specific polyclonal anti - Alicyclobacillus antibody from immunized New Zealand white rabbits and established a sensitive indirect ELISA for the detection of Alicyclobacillus spp. in apple juice. The results indicated that the established ELISA

Literature review and thesis outline was a potentially useful analytical method for detection of Alicyclobacillus spp. in apple juice. A novel staphylococcal protein A (SPA) - ELISA assay was also developed to rapidly detect A. acidoterrestris in apple juice based on the polyclonal antibodies from Japanese white rabbit against A. acidoterrestris (Li, Xia, & Yu, 2013). The results showed that the SPA - ELISA has no cross - reaction with five common food - borne microorganisms, such as Listeria spp., Salmonella typhimurium and Escherichia coli, and exhibited high sensitivity and excellent agreement with isolation by K medium. Flow cytometry is a laser - based, biophysical technology, allowing simultaneous multiparametric analysis of both physical and chemical characteristics through fluorescent dyes. Flow cytometry can be rapid and sensitive compared with conventional culture - based methods, but it can only examine fluid samples and cannot achieve the quick scanning of solid raw materials or fruit products. The apparatus can scatter samples to obtain corresponding light patterns of DNA density and cell size. Borlinghaus and Engel (1997) introduced flow cytometry to the detection of Alicyclobacillus spp. in fruit juice concentrates. Another method for the identification of Alicyclobacillus spp. is the microscopic method (Pettipher & Osmundson, 2000). This test uses a direct epifluorescent filter technique (DEFT), which is a combination of membrane filtration with a nucleopore polycarbonate membrane with a pore size of 0.6 μm, a fluorescent dye, such as acridine orange, and epifluorescence microscopy. With this technique it is possible to see the rod - shaped bacteria if they are present in the sample tested.

6.1. Molecular ‐ based typing methods

Over the past decades, molecular methods have been introduced as alternatives to culturing methods to detect or enumerate bacteria. The typing of microorganisms can provide information on the sources of contamination within a dairy manufacturing plant. Sequence - based techniques provide fast and detailed identification of spoilage and pathogenic sporeformers, which are highly diverse. Implementation at a routine level in the food industry requires the development of rapid, cheap methods that are easily performed and interpreted. However, given the usually very low concentrations of spores of interest in milk and dairy products, sample pre - treatment, enrichment/concentration and cultivation steps will continue to be important, thus forming a rate limiting step in detection speed (Wells-Bennik, Driehuis, & van Hijum, 2016).

Chapter 1

A number of different molecular methods are now available for the routine identification of bacteria, including pulsed field gel electrophoresis (PFGE), restriction fragment length polymorphism (RFLP), multilocus sequence typing (MLST), randomly amplified polymorphic DNA analysis (RAPD), 16S - 23S rRNA internal transcribed spacer - PCR (ITS - PCR), multiparametric TaqMan RT ‐ PCR and multilocus variable - number of tandem repeat (VNTR) analysis (MLVA). The current recommendations for the characterization of new bacterial species include PCR - based identification techniques obtaining 16S rRNA gene sequence data, performing DNA – DNA hybridization with closely related bacteria and determining phenotypic and chemotaxonomic characteristics (Stackebrandt et al., 2002). These classic genotyping techniques based on sequence variability of single or multi locus PCR amplified genes, often lack discriminating power at the level of individual isolates within the same species or need laborious and extensive sequencing (Caspers et al., 2011). All species of the genus Geobacillus are very phylogenetically closely related. Thus, the distinction between species of Geobacillus based on 16S rDNA sequence data is not always clear and accurate. For example, the intragenic similarities for the 16S rRNA gene sequence of G. kaustophilus, G. thermoleovorans and G. stearothermophilus are more than 99% (Weng, Chiou, Lin, & Yang, 2009). Alternatives to 16S rDNA sequencing, and in particular DNA – DNA hybridization, for description of species are currently being investigated. Many of these approaches are based around phylogenies created from sequencing of housekeeping genes. Recently, a 16S rDNA method was developed by Chauhan et al. (2013) that can rapidly identify a number of dairy bacilli, including Geobacillus spp., A. flavithermus and B. licheniformis. This method uses primers to amplify two separate variable regions within the 16S rDNA gene, using PCR. The products then undergo a high ‐ resolution melt analysis (HRMA) for differentiating polymorphic PCR products using a double stranded DNA binding fluorescent dye and a PCR machine which is highly precise to increases in temperature. Amplicons with different DNA sequences, G + C contents, and lengths are differentiated on the basis of their melt curves (Wittwer, Reed, Gundry, Vandersteen, & Pryor, 2003). While this method is excellent for identifying a wide range of bacilli from different dairy products, it is unable to differentiate species of Geobacillus, due to their similarity in the 16S rDNA sequence. However, Caspers et al. (2011) developed a microarray containing genetic information of 34 different strains from the Bacillus and Geobacillus genera allowing high - resolution genotyping by comparative genome hybridizations. They showed

Literature review and thesis outline that this microarray - based comparative genome hybridization (M - CGH) is a powerful tool for the rapid and unambiguous discrimination of genetically distant, as well as closely related isolates of endospore - forming bacilli in the food chain. In addition, the hypervariable (HV) region of 16S rRNA gene is highly conserved within Alicyclobacillus species (>98.8% sequence similarity); thus it can be utilized to identify and differentiate species of the genus Alicyclobacillus (Goto et al., 2002). A method for identifying dairy Bacillus spp. and Paenibacillus spp. has been suggested based on gene sequence analysis of the less conserved rpoB “housekeeping gene”4 (Durak, Fromm, Huck, Zadoks, & Boor, 2006; Tellez & Antonio, 2016). However, this may not be a suitable genotyping method for Geobacillus spp. as the rpoB gene is highly conserved in this genus compared with other bacilli. Studies have shown that sequencing of the variable regions within the rpoB gene could replace 16S rRNA gene sequencing in Geobacillus spp. as a species ‐ level identification method (Meintanis et al., 2008; Weng et al., 2009). Other gene targets for the typing of Geobacillus spp. include recA (Weng et al., 2009), spo0A (Kuisiene, Raugalas, & Chitavichius, 2009), which is the master regulator gene of the sporulation, recN (Zeigler, 2005) and the 16S ‐ 23S rRNA gene internal transcribed spacer (ITS) region profiling (Flint et al., 2001; Kuisiene, Raugalas, Stuknyte, & Chitavichius, 2007). Specifically, 16S - 23S rRNA gene internal transcribed spacer (ITS) separates 16S and 23S rRNA genes and may contain tRNA genes. The sequence of ITS exhibits greater variations than that of the 16S rRNA structural gene. This variation can occur between species in both the length and the sequence of this region. Hence, ITS sequences are more useful for the genus - and species - specific primer design than 16S rRNA gene (Kuisiene et al., 2007). While the above targets are effective in discriminating between different thermophilic bacilli, there has been limited success with Geobacillus spp. (Coorevits et al., 2012). Furthermore, Yamazaki, Teduka, Inoue and Shinano (1996) used the reverse transcription polymerase chain reaction (RT - PCR) to detect A. acidoterrestris and A. acidocaldarius on the basis of squalene – hopene cyclase – encoding gene (sch - Src homology 2 domain containing), which encodes a key enzyme in the biosynthesis of humanoids, vital membrane components of Alicyclobacillus spp. The gyrB is another gene encoding the different topological forms of DNA and is widely distributed among living organisms. Regarding the nucleotide sequence variability, gyrB

4“Housekeeping genes” are referred to the genes whose expression is stable during different growth stages and usually they are involved in processes vital to the cell.

Chapter 1

(average similarity within Alicyclobacillus species: 77.2 %) is less conserved than 16S rDNA (average similarity within Alicyclobacillus species: 94.5 %); thus, it enables a more refined differentiation among the DNA sequences. Just like 16S rDNA, it can be used to determine the nucleotide sequence and identify and differentiate the species. Recently a method using an easier and more specific primer has been developed (Goto et al., 2006) using the chromosomal DNA extracted from colonies as a template, performing a PCR reaction using a specific primer for the genus Alicyclobacillus or A. acidoterrestris, and detecting the difference between the reaction products by gel electrophoresis. Positive aspects of this method are the utilization of a low cost, common apparatus in the laboratory and a required time of approximately 5 hours to get the results. Pyroprinting is a new, easy, rapid, reproducible and automatic typing tool that requires bacterial isolation (but no further confirmation before pyroprinting), a PCR step directly from bacterial culture, and then pyrosequencing. The PCR primers used for the study of VanderKelen et al. (2016) were designed to recognize both Bacillus and Geobacillus species from endospores isolated from raw milk or milk powder, and pyroprinting incorporates information from at least 10 genomic sites, giving this method both breadth in isolate detection and depth in discriminatory power. The derived results showed that pyroprinting can fingerprint isolates quickly with a high degree of reproducibility, group isolates at the subspecies level from a wide variety of raw milk and milk powder isolates, and cluster isolates into taxonomically relevant groups. Pyroprinting presumptively identified groups within several Bacillaceae genera, and even groups within a species. Thus, pyroprinting may also provide presumptive identification of a broad range of contaminants with a specificity that allows source tracking of Bacillaceae groups from raw milk to powdered milk. RAPD ‐ PCR profile analysis (Williams, Kubelik, Livak, Rafalski, & Tingey, 1990) has previously been used to identify and genotype G. stearothermophilus, A. flavithermus, B. licheniformis and B. subtilis isolates obtained from milk powders produced in New Zealand (Flint et al., 2001; Ronimus et al., 2003) and around the world (Rückert et al., 2004; Reginensi et al., 2011; Sadiq et al., 2016). This assay has also been applied to distinguish A. acidoterrestris from the other thermoacidophilic bacteria within 6 hours. The results of RAPD were identical to conventional, biochemical, morphological and physiological inspection methods (Yamazaki, Okubo, Inoue, & Shinano, 1997). The RAPD - PCR method can be also used to produce unique genetic fingerprints of isolated Bacillus strains in bread manufacturing

Literature review and thesis outline

(Sorokulova et al., 2003). This technique employs PCR amplification of random regions of the polymorphic genome using short arbitrary primers. The PCR products are then run through gel electrophoresis, resulting in different banding patterns between strains creating a fingerprint, which can be compared between known controls and samples. RAPD ‐ PCR profiling requires no genome sequence information, can detect differences between closely related species enhancing the sub - species differentiation of the isolates, is quick and easy. However, this technique is not an effective tool in providing a wide range of genetic heterogeneity among strains of one species, has poor reproducibility between different laboratories and difficult interpretation of the banding patterns due to weak bands in an isolate’s profile. The above result from varying efficiency of the PCR reaction and mismatches between the primer and the DNA template, thus the technique is rendered questionable (Seale et al., 2012). Recently, a multiparametric, TaqMan probe - based, RT ‐ PCR assay has been developed. It can discriminate between 38 different species of spore ‐ forming bacilli, including psychrotrophic, mesophilic and thermophilic aerobic bacilli, as well as members of the genus Clostridia. Sensitivity is high if a pre ‐ enrichment step is used and the detection limit is 1 spore of B. cereus in a 25 g food sample (Postollec et al., 2010; Postollec et al., 2012). Unfortunately, this method has a relatively low sample throughput (3 samples at a time). Fernandez‐No et al. (2011) reported a quantitative TaqMan ‐ probe assay for B. cereus, B. licheniformis and B. subtilis directly from foods without a pre ‐ enrichment step. This method, however, is unable to differentiate between the three species. Luo et al. (2004) established a TaqMan real - time PCR - based method, targeting the sch gene to detect A. acidoterrestris and A. acidocaldarius, obtaining a sensitivity of less than 100 cells within 3 - 5 hours. Although there was no cross - reactivity with other common foodborne microorganisms, the assay could only be applied to these two species. Consequently, Connor et al. (2005) developed an expanded assay targeting the 16S rRNA gene sequence to detect cells of A. acidoterrestris, A. acidocaldarius and A. cycloheptanicus. Rueckert et al. (2005b) also developed a TaqMan - based realtime - PCR assay using a small amplicon of the ribosomal 16S rRNA gene for the selective and quantitative detection of thermophilic bacilli in milk powders in less than 1 hour. Recently, a new typing method has been developed, based on Multiple - Locus Variable number tandem repeat Analysis (MLVA). Length polymorphisms arise in a variable number of tandem ‐ repeat (VNTR) loci due to the variable copy number of tandem repeats found within genes or noncoding regions of a genome. They can be

Chapter 1 analyzed using agarose gel electrophoresis or combined with high ‐ resolution melt analysis (MLV ‐ HRMA) (Keim et al., 2004; Reyes & Tanaka, 2010; Vogler et al., 2007). MLV ‐ HRMA techniques have been developed for the typing of Geobacillus spp. and B. licheniformis isolates obtained from milk powders manufactured in Australia (Dhakal et al., 2013; Seale et al., 2012). Seale et al. (2012) demonstrated that three different genotypes of Geobacillus spp. could coexist in a milk powder manufacturing plant during a single processing run, and that specific types were associated with high ‐ spore ‐ count powders. However, there was no correlation of specific types with particular dairy manufacturing plants and retailed samples. Another interesting finding was that isolates obtained from powder processing plants in the same region of Australia in 1995 were of the same type as those obtained in 2012, indicating that the same types remained prominent over 17 years. The study by Dhakal et al. (2013) showed that isolates of B. licheniformis were more heterogeneous across multiple product runs and milk powders than previously thought, and no correlation could be drawn between prominent types and specific dairy manufacturing plants. These studies showed that both MLVA and MLV ‐ HRMA were more reproducible and discriminatory, distinguishing subgroups within the clusters defined by RAPD - PCR and rpoB gene sequencing methods previously used to type thermophilic bacilli. A microarray ‐ based genotyping method using 130 genomic markers has shown to discriminate 34 different strains from 6 Bacillus species and 4 species of Geobacillus genus isolated from a variety of food products (Caspers et al., 2011). This method detects differences between core and accessory genome markers across Bacillus and related genera. The majority of the core genome markers do not hybridize between species, thus resulting in discrimination at the species level, while the accessory genome markers can result in high ‐ resolution discrimination between individual isolates of a single species. A rapid detection method based on DNA probe technology, called VIT® Alicyclobacillus, was developed by Vermicon AG. Gene probes that are complimentary to Alicyclobacillus spp. gene sequences are combined with fluorescent dyes and are allowed to anneal to Alicyclobacillus DNA. When viewed under a fluorescent microscope the dye becomes visible where it has bound, indicating the presence of Alicyclobacillus spp. The method is able to distinguish bacteria belonging to the genus Alicyclobacillus from other species (fluorescence green) and can also distinguish A. acidoterrestris from other Alicyclobacillus species (fluorescence red). Further advantages include the fact that only viable cells are

Literature review and thesis outline detected, the results can be obtained rapidly (after a pre - enrichment procedure of 2 d the complete analysis can be performed within 3 h), the staff do not need to be trained in molecular biology and the equipment required is routinely found in most laboratories (Thelen, Snaidr, & Beimfohr, 2003). Future developments in the typing of thermophilic endospore ‐ forming bacilli will arise as Whole Genome Sequencing (WGS) becomes more readily available and more economical. A number of thermophilic bacilli isolated from dairy products and dairy manufacturing plants have had their genomes sequenced, including isolates of G. thermoglucosidans (Zhao et al., 2012), A. flavithermus (Caspers et al., 2013) and B. licheniformis (Dhakal, Seale, Deeth, Craven, & Turner, 2014). These genomes will provide targets that might serve as the basis of typing techniques and might provide some insight into how these microorganisms persist within dairy manufacturing environments.

7. Enumeration of thermophilic bacilli

The enumeration of vegetative cells and spores of thermophilic bacilli is very important in monitoring food product manufacture and for grading the quality of food products. Traditional viable plate ‐ counting techniques, used to determine both thermophiles and thermophilic spore counts, can take a long time to obtain a result, which can delay the release of dairy products. As a result, the focus is on developing novel rapid enumeration methods that enable more rapid product release. Overall, molecular methods are rapid, sensitive, specific, time - and labor - saving. They are particularly suitable for detailed species and strain identification purposes, but are not used routinely in the food industry for enumeration purposes.

7.1. Viable plate counts

Currently there is no standard enumeration technique for either vegetative cells or spores of thermophilic bacilli. The two traditional methods used are the total thermophilic plate count (TPC) and the thermophilic spore count (TSC). The TPC method involves reconstitution of product, followed by the transfer of 1 ml of sample and decimal dilutions into separate Petri dishes containing milk plate count agar (MPCA) and incubation at 55 °C for 48 h. In TSC, the sample is first heat treated for 30 minutes at 100 °C to eliminate vegetative cells and activate spores enabling

Chapter 1 them to germinate. The sample and decimal dilutions are then pour plated with MPCA supplemented with 0.2% starch and incubated at 55 °C for 48 h. The starch is added as aid for spore germination. This heat treatment is more intense than in previously published methods for thermophilic spores, reaching 80 °C for 10 minutes (Coorevits et al., 2008), 80 °C for 20 minutes (McGuiggan et al., 2002) or 100 °C for 10 minutes (Rückert et al., 2004; Rueckert et al., 2005a). This higher temperature over a longer of period of time ensures that the method selects spores that may survive the higher processing temperatures used during dairy manufacturing. Recently, a new method has been developed to enumerate highly heat ‐ resistant spores in milk powder that are to be further processed for UHT or retort treatment. This method involves heat treating at 106 °C for 30 minutes; this higher temperature specifically selects spores of Geobacillus spp. and destroys spores of A. flavithermus (Hill and Smythe, 2004). On the other hand, not all media are able to support the growth of A. acidoterrestris, including nutrient agar, tryptone soy agar, brain heart infusion agar, standard plate count agar and veal infusion agar, even when these media are acidified to pH 3.5 (Pettipher et al., 1997; Spllttstoesser et al., 1994). The inability of Alicyclobacillus spp. to grow on these media may be due to the presence of inhibitory substances, such as peptones. Beginning with acidified potato dextrose agar, which is commonly used for yeast and mold detection, a variety of media have been used and new media have been developed for the isolation and enumeration of Alicyclobacillus spp. Commonly used isolation media include Bacillus acidocaldarius medium (BAM), orange serum agar (OSA), potato dextrose agar (PDA), Yeast – Starch - Glucose agar (YSG agar), Hiraishi glucose yeast extract (HGYE) agar, and K agar (Bevilacqua, Sinigaglia, & Corbo, 2008; Chang & Kang, 2005; Chang et al., 2013; Henczka et al., 2013; IFU, 2007; Murray et al., 2007; Pettipher & Osmundson, 2000; Pettipher et al., 1997; Silva & Gibbs, 2001; Spllttstoesser et al., 1994; Walls & Chuyate, 2000; Yamazaki et al., 1996). Typical incubation conditions are 43 °C for 2 - 3 days, although incubation temperatures vary from 40 to 55 °C (Murray et al., 2007).

7.2. Alternative rapid methods

Rapid methods have the potential to increase the control over the process and reduce both labor costs and the time required to obtain results for thermophile and thermophilic spore counts. Unfortunately, rapid methods generally require access to

Literature review and thesis outline expensive equipment and reagents, as well as specialized technical training. Two rapid methods have been developed recently for the enumeration of thermophilic bacteria in milk powder, one using flow cytometry and the other using real ‐ time polymerase chain reaction (RT ‐ PCR). A rapid method of counting viable mesophilic bacterial cell numbers in milk powder (equivalent to a standard plate count) was developed by Flint et al. (2006) using a BactiFlowTM flow cytometer based on a fluorescent substrate that can detect esterase activity of viable cells with results available in 2 h. This method showed promise and was modified to enumerate thermophilic bacteria in milk powder by including a 55 °C incubation step (Flint, Walker, Waters, & Crawford, 2007). It showed good correlation with TPC data during the development phase, but it did not always correlate well during routine use in a manufacturing context. In addition, the detection limit was low for some milk powders. A RT ‐ PCR assay was developed by Rueckert et al. (2005a, b) to enumerate total viable vegetative cells and spores of A. flavithermus, B. licheniformis and mesophilic B. megaterium in milk powder. However, Geobacillus strains were not included in the study. The assay targeted the 16S rDNA gene, which can have a variable copy number impacting on the results. A SYBR® Green - based quantitative RT ‐ PCR assay was then developed which targeted the spo0A sporulation gene (Rueckert, Ronimus, & Morgan, 2006); this assay amplified DNA from a variety of the thermophilic bacilli (the targeted strains), as well as a number of non-targeted strains, including B. cereus and B. smithii. The assay was rapid and provided a result within 1 hour. Although RT - PCR assays are costly to perform, require technical expertise and are not sensitive or specific enough, they provide the manufacturer with the ability to monitor the extent of thermophilic contamination during milk powder manufacturing 60 - 90 min after sampling and include shorter production times between cleaning reducing thermophilic bacilli contamination (Rueckert et al., 2005a). In addition, RT ‐ PCR assays require incorporation of a reverse transcriptase step in order to target viable bacterial cells. In summary, while there has been progress in the development of rapid methods for the detection and enumeration of thermophilic bacilli, these methods require specialized equipment and training. Further research is also required to improve the sensitivity, specificity and robustness of these techniques before they can be applied to the food industry.

Chapter 1

8. Prevention and control measures for thermophilic bacilli in dairy, beverage and processed – food production plants

Current practices that have been employed by the food industry for reducing thermophilic bacilli contamination include shorter production times between cleaning, an increased cleaning frequency and the use of sanitizers (disinfectants). Recently, focus has turned to the development of novel control techniques, such as the alteration of temperature cycling, the reduction of the surface area of equipment in the optimal temperature growth zone and the duplication of equipment (Burgess et al., 2010).

8.1. Cleaning ‐ in ‐ place

At the end of every production run, cleaning-in-place (CIP) processes are applied in order to maintain a clean and hygienic environment, including piping and fitting systems (Bayoumi, Kamal, El Aal, & Awad, 2012). Using a CIP regime suggests performing cleaning without having to dismantle the processing equipment. A typical CIP regime consists of the following steps: a warm water rinse, an alkaline wash, a water rinse, a nitric acid wash and a second cold water rinse. The caustic wash is designed to solubilize organic matter (fats and milk proteins), while the nitric acid is a strong oxidizer and removes inorganic material (calcium phosphate and other salt deposits). In some cases, a sanitizer may be applied at the end of CIP, to inactivate any planktonic cell, that might remain on equipment surfaces (Burgess et al., 2010; Marchand et al., 2012; Seale et al., 2015), which are quite different from the biofilm cells due to their altered physiological status. Therefore, the surviving bacteria can potentially attach to piping and fitting surfaces, where they could promote the development of biofilm structures that enable protection against high temperatures and chemical compounds applied during pasteurization and sanitization procedures. Bacteria within biofilms may attach to tools and equipment at other positions of the plant, thus persist a longer time in the dairy environment (Brooks & Flint, 2008; Marchand et al., 2012). Additionally, Shemesh et al. (2014) found that the reduction of external pH is an environmental cue for the behavioral switch from surface motility to biofilm formation of A. acidoterrestris. Gaining insight into the multicellular behavior, based on external pH, that shelters A. acidoterrestris in food contact surfaces from external harsh conditions may

Literature review and thesis outline contribute to the development of novel antimicrobial means to prevent cross – contamination caused by this bacterium. Thereto, inactivation and removal of bacterial cells capable of forming biofilms deserves much more attention. Bacteria embedded in a biofilm as well as bacterial spores have been considered as more resistant to heat, chemicals, irradiation and desiccation than the corresponding planktonic cells; however, resistance can vary widely among species. The ability of CIP procedures and the benefit of using sanitizers to remove biofilms and spores from the processing equipment are still subject to debate. Viable bacterial cells may remain attached to manufacturing surfaces after a CIP, even though they appear visibly clean (Marchand et al., 2012; Watkinson, 2008). Parkar et al. (2003) demonstrated that the sequential application of a 2% sodium hydroxide solution (75 °C for 30 minutes) and a 1.8% nitric acid solution (75 °C for 30 minutes) removed biofilms of A. flavithermus from stainless steel surfaces. However, changing the temperatures and/or the concentrations of the sodium hydroxide and nitric acid solutions reduced the ability of the cleaning procedure to remove biofilm cells. The sodium hydroxide and nitric acid treatments employed were sporicidal (Seale, Flint, James McQuillan, & Bremer, 2011; Knight & Weeks, 2008). It is important to monitor and control the chemical concentration of cleaning solutions and the temperature employed during cleaning, as both affect the sporicidal activity of cleaning solutions (Knight & Weeks, 2008). The most widely used disinfectant is probably chlorine, but it is only a slow sporicidal and is readily absorbed by organic matter, although there are organic chlorine release agents which are more effective in the presence of soiling (Brown, 2000). Chlorous acid can be used as an alternative sanitizer of chlorine to control a major A. acidoterrestris contamination source in juice processing plants (Lee, Ryu, & Kang, 2010). Lindsay et al. (2000) were able to isolate viable spores of Bacillus spp., in particular B. cereus, from alkaline cleaning solutions that had been used for dairy CIP procedures, while Brent Seale et al. (2011) demonstrated that exposure to a 1% sodium hydroxide solution enhanced the ability of surviving spores of Geobacillus spp. to attach to stainless steel surfaces. These findings suggest that the practice of recycling sodium hydroxide cleaning solutions during CIP could potentially spread viable spores around the surfaces of the processing equipment if not properly controlled. This has also some serious implications to the practice of reusing sodium hydroxide cleaning solutions to clean other sections of the plant. Hence, it becomes very important to design a cleaning regime to ensure that spores are removed from the surfaces of

Chapter 1 equipment and that the spores which have been suspended in cleaning solutions are inactivated (Burgess et al., 2010). Hsiao and Siebert (1999) investigated the effectiveness of several organic acids towards vegetative cells of A. acidoterrestris. According to the minimum inhibitory concentrations (MICs), they concluded to a possible hierarchy of their effectiveness: benzoic > butyric - caprylic > acetic >> citric-malic-lactic-tartaric acids. Although various studies have demonstrated the effectiveness of benzoate, or sorbate to inhibit Alicyclobacillus, the exact mode of action is still undefined; however, Bevilacqua et al. (2008) suggested that the competitive inhibition with the germinants is a promising way for future researches. They also proposed that sorbate could act as a competitive inhibitor of some germinants, such as L - alanine and L - cysteine.

8.2. Other prevention and control methods

The growth of thermophilic bacteria in dairy processing equipment essentially comes down to a time - temperature relationship. To counter this problem, control can be achieved by either limiting the production cycles, which actually limits the time available for the growth of thermophilic bacteria, or by altering the temperatures prevailing on the operating processing equipment at which the growth rates of thermophilic bacteria are reduced. It is very common for dairy manufacturers to reduce production runtimes of milk treatment to a 6-8 hours timespan using centrifugal cream separators, heat exchangers, evaporator preheaters and evaporators, in order to limit the growth of thermophilic bacteria. Similarly, production lengths for the manufacture of milk powder can be limited to between 18 and 24 hours, or up to 10 hours when manufacturing ‘high - spec’ milk powders, which have very strict limits on thermophilic spore counts. It is also common to reduce the operating temperature (e.g. to between 15 and 30 °C) of processing equipment, such as centrifugal separators and ultrafiltration (UF) plants, to prevent thermophilic growth (Burgess et al., 2010; Seale et al., 2015). Another approach that uses temperature to control biofilm development is the implementation of temperature step changes, which have been shown to control the development of biofilms of thermo - resistant streptococci in cheese ‐ milk pasteurization equipment (Knight, Nicol, & McMeekin, 2004). A modification of this method may be feasible as a way of controlling the development of biofilms and

Literature review and thesis outline sporulation by thermophilic bacilli˙ Burgess et al. (2009) demonstrated that lowering the temperature of the growth environment from 55 to 48 °C prevented the formation of spores in biofilms of A. flavithermus. Reducing the preheating surface area of processing equipment in the optimized temperature growth zone of thermophiles may also reduce the thermophilic growth in milk powder manufacturing plants (Refstrup, 2000). This can be achieved by using a direct ‐ contact heating system, such as a direct steam injection (DSI) unit. Heating in a DSI unit is achieved by injecting steam directly into the milk, rapidly increasing the temperature of the milk and avoiding the optimum growth temperature of thermophilic bacteria. Such a system is more expensive to operate than an indirect heating system, such as a plate heat exchanger, due to the requirement for additional steam. The use of a DSI unit also results in dilution of the milk, due to the added water (as steam), so flash evaporation is required downstream (Burgess et al., 2010). Ιt is also possible to use a dual preheating system, in which milk is directed from one preheater to another after 8 - 12 hours of processing. This allows the first preheater to undergo a CIP procedure without disrupting the manufacturing cycle (Refstrup, 2000). On the other side, due to increasing consumers’ and retailers’ demand for minimally processed (natural and artificial chemical additive - free) food to be safe and with a lesser impact on the environment, various kinds of natural compounds (bacteriocins, chitosan, lysozyme, essential oils - EOs, tea polyphenols) have been employed as potential inhibitors against Alicyclobacillus spp. and G. stearothermophilus (Bevilacqua, Ciuffreda, Sinigaglia, & Corbo, 2014; Grande et al., 2005; Huertas, Esteban, Antolinos, & Palop, 2014; Pei, Yue, & Yuan, 2014; Viedma et al., 2009). For example, it has been proven that nisin produced by Lactococcus lactis subsp. lactis (Delves-Broughton, Blackburn, Evans, & Hugenholtz, 1996) could prevent the pre - emergent swelling of spore formers at post - germination stages of spore development, thus suppressing the outgrowth, the formation of vegetative cells and the microbial growth (Komitopoulou et al., 1999). The effect of nisin against spores is more likely to be sporostatic than sporicidal (Delves-Broughton et al., 1996). Moreover, nisin exhibits excellent stability and antibacterial activity at low pH (3.4) even when treated at high temperatures, thus indicating its potential use as an inhibitor of Alicyclobacillus in the processing of different fruit juices (de Oliveira Junior, de Araújo Couto, Barbosa, Carnelossi, & de Moura, 2015; Huertas et al., 2014; Jiangbo, Xinghua, Tianli, & Yahong, 2016; Komitopoulou et al., 1999; Pena, De Massaguer, Zuniga, & Saraiva, 2011; Peña & Rodríguez, 2010; Peña & de Massaguer,

Chapter 1

2006; Peña, Massaguer, & Teixeira, 2009; Yamazaki, Murakami, Kawai, Inoue, & Matsuda, 2000). Its use, however, is limited, due to the relatively high cost of extraction and purification and also because of the reduction in bactericidal activity in complex food substrates (Ross, Griffiths, Mittal, & Deeth, 2003). Apart from nisin, enterocin EJ97 produced by Enterococcus faecalis EJ97 has been investigated for its effectiveness against G. stearothermophilus vegetative cells and endospores in different types of canned fruit juice and vegetable foods during storage and under temperature abuse conditions (Viedma et al., 2010). Results from this study strengthened the potential of enterocin EJ97 for biopreservation, because of its antimicrobial activity against G. stearothermophilus in canned vegetable foods and drinks. Aside from the above, according to some reports (Bevilacqua, Campaniello, Speranza, Sinigaglia, & Corbo, 2013; Bevilacqua, Corbo, & Sinigaglia, 2008b; Bevilacqua, Corbo, & Sinigaglia, 2010; Bevilacqua et al., 2009; Huertas et al., 2014) EOs (eugenol, cinnamaldehyde and limonene) and their active compounds alone or conjugated with a mild thermal treatment have significant antimicrobial activity against Αlicyclobacillus spp. spores and their control during storage of fruit juices (Bevilacqua, Corbo, & Sinigaglia, 2008a; A Bevilacqua et al., 2008b; Bevilacqua et al., 2009). The antimicrobial activity of EOs against the Alicyclobacillus spp. spores was strain - dependent and likely to be affected by the secondary alkyl groups, as it was possible to point out a hierarchy of effectiveness for the active compound: cinnamaldehyde > eugenol >limonene. At low concentrations these chemicals seemed to exert a reversible stress, as their effect was time - dependent, and act at the outgrowth step (Bevilacqua et al., 2008b). Their proper use as flavoring and antimicrobial agents in fruit juice would be an alternative to prevent microbial proliferation with minimum change of sensory properties, while these compounds have been recognized as GRAS (Generally Recognized as Safe) by the FDA (Food and Drug Administration) and their use is permitted in the E.U. (Bevilacqua et al., 2008a; Bevilacqua et al., 2010). In addition, Sakanaka et al. (2000) found that tea polyphenols, which are constituents of green tea extract, exhibit bactericidal activity against G. stearothermophilus and decreased its spores heat resistance, possibly by interacting with it, and the adsorption of tea polyphenols to proteins on the spore surface. Tea polyphenols would be, therefore, useful in preventing the spoilage of canned drinks containing high carbohydrate content. Conventional thermal - processing approaches may cause undesirable non - enzymatic browning and loss of vitamins and volatile flavor compounds that can lead to degradation of the nutritional and sensory quality of juices. In consideration of

Literature review and thesis outline that, alternative mild (non - thermal) technologies (high hydrostatic pressure, high - pressure homogenization, ultra - high - pressure homogenization, irradiation, ultrasound, microwaves, high - intensity pulsed electric field, pressure - assisted thermal processing - PATP) for sterilization and preservation have been intensively investigated to prevent fruit juice from Alicyclobacilli spoilage (Alpas et al., 2003; Baysal, Molva, & Unluturk, 2013; Bevilacqua, Cibelli, Corbo, & Sinigaglia, 2007; Ferrario, Alzamora, & Guerrero, 2015; Lee, Dougherty, & Kang, 2002; Roig-Sagues et al., 2015; Silva, 2016; Silva et al., 2012; Sokołowska et al., 2013; Uemura, Kobayashi, & Inoue, 2009; Vercammen et al., 2012; Wang, Hu, & Wang, 2010; Wang, Zhao, Liao, & Hu, 2010) and low – acid foods from G. stearothermophilus spoilage (Ahn, Lee, & Balasubramaniam, 2015; Ananta, Heinz, Schlüter, & Knorr, 2001; Espejo, Hernández- Herrero, Juan, & Trujillo, 2014; Gao, Ju, & Jiang, 2006; Gao, 2010; Georget et al., 2014; Georget et al., 2014; Khanal, Anand, & Muthukumarappan, 2014; Olivier et al., 2011; Sevenich et al., 2014; Shibeshi & Farid, 2010, 2011) to assure their quality. Particularly, ultrasound is one of the simplest and versatile methods for cellular disruption and food extract production. Yuan et al. (2009) evaluated the effect of ultrasonic treatments on A. acidoterrestris in apple juice and the changes of the juice quality after ultrasonic treatments. In general, inactivation of the cells was more pronounced at an elevated power level and as the processing time increased. Changes of sugar content, acidity, haze and juice browning were noted after ultrasonic treatments but did not adversely alter the juice quality.It has also been proven that ultrasound treatment is more effective when combined with other processes such as pressure and/or heat (Silva, 2016; Yuan, Hu, Yue, Chen, & Lo, 2009). However, adverse effects of ultrasound on the organoleptic properties of fruit juice should not be ignored in terms of customer demand and commercial interest (Yuan et al., 2009). Although several non - thermal physical techniques are emerging to improve the microbiological safety and quality of fruit juices, some of them have been considered too energy - expensive or costly to be practical for use in food processing.

9. Predictive modeling for growth of endospore – forming bacteria

For the evaluation of safety and quality throughout the food supply chain, the traditional bacterial enumeration techniques are time consuming. Therefore,

Chapter 1 mathematical models are being developed for predicting microbial behavior in order to optimize, control and improve food quality and safety (Juneja et al., 2007). It is of economic interest for the food industry to be able to predict (without the need for time consuming and expensive challenging test processes) the potential for growth of specific microorganisms in a food or to establish product formulations excluding growth (Khanipour et al., 2016). That said, the estimation of growth characteristics by mathematical modeling could be a very useful tool for manufacturers, allowing them to identify the important variables that determine the shelf - life of their product. Consequently, in recent years, predictive microbiology has been established as an important tool for the food industry by developing models used to describe and predict the behavior of pathogenic and spoilage bacteria during processing and storage (Bovill, Bew, & Baranyi, 2001; Koseki & Isobe, 2005; McMeekin et al., 1997; McMeekin, Olley, Ratkowsky, Corkrey, & Ross, 2013) by combining experimental data, microbiological knowledge and mathematical techniques. The aim of predictive modeling is to use mathematical equations to describe a quantitative evolution (growth, survival and inactivation) of foodborne pathogenic and spoilage bacteria in foods under the influence of several environmental factors, such as temperature, pH and water activity (McMeekin et al., 1997; McMeekin et al., 2013; Membré et al., 2005). The ability of spores to germinate and the ability of vegetative cells to grow is clearly affected by the composition of the medium in addition to changes in pH and incubation temperature (Carlin et al., 2000). Therefore, an important issue in predictive food microbiology is the extent to which the results that are obtained under controlled laboratory conditions can be also applied to a less controlled environment of a product distribution process or an industrial process (Koseki & Isobe, 2005; Koutsoumanis, Lianou & Gougouli, 2016). Particularly, predictive microbiology describes and quantifies microbial population evolution, based upon the premise that the responses of microorganisms to environmental factors are reproducible. A primary model aims at accurately describing the changes in growth as a function of time under constant environmental conditions. Sigmoidal type models, such as the modified Gompertz, and logistic models are often used for fitting observed microbial growth data and predicting growth kinetic parameters (that is, initial concentration, lag time, maximum specific growth rate and maximum population density). At present, the Baranyi model is getting more popular among researchers and several studies have reported that the Baranyi model performs better than the modified Gompertz and

Literature review and thesis outline the logistic models (Juneja et al., 2007). Though several models predict the microbial growth accurately under certain isothermal conditions, they cannot estimate the growth under dynamic (varying environmental) conditions (Huang, 2003). To predict the microbial growth under nonisothermal conditions, a secondary model that describes the effect of one or more environmental factors (temperature, pH, aw) on one or more primary growth parameters is required (Membré et al., 2005). In a next step, a tertiary model is developed that applies both primary and secondary models to predict the growth of endospore – forming bacteria under different conditions (Huang & Vinyard, 2016). Thus, the performances of both primary and secondary models affect the overall accuracy of the tertiary model (Membré et al., 2005). Temperature is usually one of the major environmental factors affecting bacterial growth during its passage through the food chain. Temperatures may fluctuate wildly even during food processing, distribution, storage, retail or consumption at home and it is desirable that the predictive models are able to cope with these changes (Bovill et al., 2001). Therefore, certain dynamic models have been developed as a result of the combination of primary and secondary models, and a temperature profile, to predict the growth of bacteria in foods under varying temperature conditions. Arrhenius – type models, polynomial models, cardinal models, square - root type models and artificial neural networks have been developed to predict the growth of bacteria in foods under changing temperature conditions (Juneja et al., 2007). Such models can be validated using experimental data obtained under realistic time - temperature conditions (Valdramidis, Geeraerd, Bernaerts, & Van Impe, 2006). Generally, diseases and spoilage caused by endospore – formers are associated with thermally processed foods, as heat kills the vegetative cells but allows survival and subsequent growth of endospore - forming microorganisms. Thus, their presence poses a big challenge to the food industry. Spore stress - resistance is, in part, due to their multi - layered structure and the ability to germinate under favorable environmental conditions. The lack of uniformity of the stress response between different species and strains or even within a population of spores limits the applicability of predictive modeling (Ananta, Heinz, Schlüter, & Knorr, 2001). In addition, the germination process itself is heterogeneous in the sense that not all spores in a population may germinate at the same time. To build models and systems it becomes necessary to understand the heterogeneous properties of spores as well as to gain knowledge about the entire sporulation and germination pathways (Abhyankar, Stelder, de Koning, de Koster, & Brul, 2017). Thereto, a

Chapter 1 deeper understanding of the mechanisms involved in spore resistance, adaptation and killing (and heterogeneity there - in) may lead to improved models for spore behavior prediction (Eijlander, Abee, & Kuipers, 2011). Modeling inactivation of foodborne pathogens has been one of the first achievements in predictive microbiology. Inactivation models were initially focused on the thermal destruction of Clostridium botulinum type A spores in low acid canned foods (Esty & Meyer, 1922; Jay, 2012). Nevertheless, due to their practical usefulness the principles on which these models were grounded were also used for the generation of predictive models of microbial stability for the less hazardous but (much) more thermal - resistant spoilage endospore - formers. It may come as no surprise that, due to their conservative nature, such models for these (generally Bacillus) species still leave significant room for improvement in terms of the tradeoff between providing the desired product quality and the necessary microbial product stability (Abhyankar et al., 2017). Indeed, it has been found that the modeling of growth and inactivation of endospore - formers requires quantitative predictive models for spore germination and sporulation; with spores the situation becomes even more complicated because of the possibility of heat - induced spore activation (Graham, Mason, & Peck, 1996; van Boekel, 2002). The goodness - of - fit of any model can be adversely affected by using spore inocula, which introduces the additional variable of time needed for germination, and also by including experiments near the limits of growth, where growth tends to be less reproducible (Graham et al., 1996). Within the domain of predictive microbiology, the supporting documentation for food spoilage endospore – forming bacteria growth kinetics is still very limited (Antolinos et al., 2012; Baril et al., 2012; Kakagianni, Gougouli, & Koutsoumanis, 2016; Kakagianni et al., 2018; Llaudes, Zhao, Duffy, & Schaffner, 2001; Mtimet et al., 2015; Ng, Viard, Caipo, Duffy, & Schaffner, 2002; Ng & Schaffner, 1997). As inactivation models usually describe the fate of spores whereas growth models describe the fate of the total viable concentrations of vegetative cells and spores (CFU/ml) as a function of time, one should be extremely careful with linking growth and inactivation models without considering the state of the cells (Nauta, 2002; Nauta, Litman, Barker, & Carlin, 2003).

9.1. Modeling individual cell/spore behavior

In the context of a risk analysis framework, currently applied at the European and international levels for food safety management, the importance of variability of

Literature review and thesis outline biological and natural phenomena is widely recognized (Codex Alimentarius Commission, 1999). The importance of describing variability in microbial models used in risk assessment was stressed by Nauta (2002), who illustrated the effect of ignoring variability in management decisions. There are numerous sources of variability in microbial behavior that should be taken into account in risk - based approaches. One of the latest advances in predictive microbiology is the development of stochastic models with a probabilistic structure that consider the variability of factors that affect the microbial behavior and food safety or quality. A stochastic process analysis brings another dimension to the prediction of microbial growth (Huang & Vinyard, 2016). Stochastic modeling can take into account all possible circumstances with their associated probabilities to quantitatively assess the probability (risk) of a food being unsafe or spoiled. Once the mathematical model that describes the relationship between environmental factors and microbial behavior is constructed and the variability of the input factors is defined, methods like Monte Carlo simulation can be used to assess the distribution of the output for different values of the input factors taken from their respective probability distributions (Couvert et al., 2010; Koutsoumanis & Angelidis, 2007; Ross, Rasmussen, Fazil, Paoli & Sumner, 2009). The later probabilistic approach enables decision - making based on the ‘acceptable level of risk’ and can provide structured information on the effect of potential interventions, allowing decision - makers of public health authorities or the food industry to compare various interventions and identify those that can lead to effective and economic reduction of safety or quality risks. Several stochastic predictive modeling approaches, aiming at quantifying and integrating different types of variability, have been described the last decade (Augustin et al., 2011; Delignette-Muller, Cornu, Pouillot & Denis, 2006; Koutsoumanis, Pavlis, Nychas & Xanthiakos, 2010; Membré et al., 2005; Pouillot, Albert, Cornu & Denis, 2003). Among the various variability sources, the so - called ‘biological variability’, which includes the inherent differences in microbial behavior among identically treated strains of the same species as well as the heterogeneity in individual cell behavior, has received particular attention (Aguirre & Koutsoumanis, 2016; Aspridou & Koutsoumanis; 2015; Koutsoumanis, 2008; Koutsoumanis & Lianou, 2013; Lianou & Koutsoumanis, 2011). The recent technological, methodological and computational developments allowed the monitoring of biological variability and its quantitative description with stochastic models. The use of the developed methods, based on microscopy, image

Chapter 1 analysis and automated time - lapse microscopy, to monitor the microbial behavior at the single cell level has shown a high heterogeneity among individual cells (Elfwing, LeMarc, Baranyi & Ballagi, 2004; Koutsoumanis & Lianou; 2013; Siegal- Gaskins & Crosson, 2008; Wakamoto, Ramsden & Yasuda, 2005). In the case of food contamination by spores, the number of spores surviving the heat treatment is usually very low and is characterized by heterogeneity in terms of lag time (Baranyi, 1998; Barker, Malakar, & Peck, 2005; Billon, McKirgan, McClure, & Adair, 1997; Caipo, Duffy, Zhao, & Schaffner, 2002; Coote et al., 1995; Kakagianni, Aguirre, Lianou & Koutsoumanis, 2017; Llaudes et al., 2001; Pin & Baranyi, 2006; S. Stringer, Webb, & Peck, 2011; Whiting & Strobaugh, 1998; Whiting & Oriente, 1997). Therefore, an increased understanding of the causes of heterogeneity in spore - resistance mechanisms, germination and outgrowth will aid in a better understanding of the major physiological processes of cell differentiation in the microbial world, in the prediction of spore properties and behavior and may eventually lead to improved predictive models for food preservation (Eijlander et al., 2011; Trunet, Carlin, & Coroller, 2017). Depending on the use of the model, predictions based on high initial contamination levels can be on the “fail - safe” or the “fail - dangerous” side. For example, when the model is used to predict the time-to-spoilage, the above predictions represent a worst - case scenario. This is because the early germinated and outgrown spores within the population begin to multiply up to a microbial concentration sufficient to spoil the food leading to shorter estimations of the time- to-spoilage compared to low initial contamination levels (Baranyi, 1998; Baranyi, George, & Kutalik, 2009; Pin & Baranyi, 2006). These ‘worst case’ predictions may be useful to identify important non - safe processing steps, but they are not suited for quantifying risks as they may wrongly proved to be ‘fail - safe’ without giving an indication of the probability of an undesired situation (Nauta, 2002; Nauta et al., 2003). Therefore, attention needs to be given to relatively slow - growing spore populations, when the shelf life of a product is extended by control of rapidly growing spoilage organisms (McMeekin et al., 1997). In contrast, when the objective is to evaluate the duration of a quality control test in which products are stored at the optimum temperature for growth for a certain time and the percentage of spoiled items is evaluated, the above predictions can lead to “fail - dangerous” estimates. Clearly modelers have moved beyond predictive models developed using high initial bacterial concentrations to models that use a range of contamination spore

Literature review and thesis outline levels. Most of the effort in this area has focused on models for endospore - forming organisms, specifically C. botulinum (Whiting & Strobaugh, 1998; Whiting & Oriente, 1997). The model developed by Whiting and Oriente (1997) had spore numbers as a variable, recognizing that at low spore numbers and at a less favorable environment the probability distribution of spore germination and outgrowth becomes an important factor in estimating the likelihood of growth and toxin production. Indeed, the modeling of lag time is a more complex task for endospore - formers than for vegetative cells, requiring separate modeling of the germination and outgrowth processes. Furthermore, the probability of germination and outgrowth must be modeled independently from the time to growth (Augustin, 2011). A limited amount of work has also been done on other endospore – forming organisms and vegetative cells. It appears that the inoculum size has the most dramatic effect on the lag time (for vegetative cells) or the germination, outgrowth, and lag time (for spores) (Kakagianni et al., 2017; Schaffner, 2003). Often, the data show that the timing of germination and outgrowth of low numbers of Bacillus spores becomes difficult to be predicted with certainty (Ter Beek et al., 2011). Due to growth, products that are contaminated at a low level (possibly below the detection limit) and therefore are considered safe at a certain moment of time may become unsafe at some later time (Nauta, 2002). Therefore, by ignoring the increased lag phase heterogeneity and the biological explanation of the underlying complex physiological and molecular mechanisms behind adaptation of cells to environmental stress conditions (D'Arrigo et al., 2006), current deterministic mathematical growth models fail to characterize the lag phase appropriately (Baranyi, 2002; McKellar, 1997). Thus, for an effective risk - based quality control of food products, the growth model needs to be combined with the history of the cells/spores before they contaminate food and the quantitative data on the variability of individual spore behavior (Aguirre, Rodríguez, & de Fernando, 2011; Baranyi et al., 2009; Huang & Vinyard, 2016; Métris, George, Mackey, & Baranyi, 2008). The ideal model for endospore - forming microorganisms would include the probability of spore germination/outgrowth with an estimated value and variation for the growth rate or doubling time (Huang & Vinyard, 2016; Whiting & Oriente, 1997).

Chapter 1

Thesis Outline

This PhD thesis consists of the following structure (Fig. 1.4.): (i) an introductory part which serves to give a review of literature of the subjects under study (Chapter 1), (ii) a presentation of four research chapters, generating from the obtained results, dedicated to quantitative studies on the growth dynamics of two spoilage thermophilic endospore – forming bacilli in constant and dynamic environments at population as well as at single spore’s level and (iii) a final summarizing discussion and future applications for the food industry. The introductory theory constitutes the basis of the scientific chapters. Therefore, in Chapter 2, the research presents the development of an evaporated milk specific model for the effect of storage temperature on growth kinetics of G. stearothermophilus ATCC 7953. Furthermore, the validation of the developed model in predicting spoilage of evaporated milk at dynamic temperature scenarios simulating distribution and storage of the product is investigated. Chapter 3 concerns the development of a growth/no growth interface model to predict the probability of growth of A. acidoterrestris ATCC 49025 and a growth kinetic model as a function of temperature and pH. Moreover, the developed models were further validated against the observed growth behavior of A. acidoterrestris in 8 commercial pasteurized fruit drinks. Given that in practice the spoilage defects of evaporated milks are derived from low bacterial spore numbers, in Chapter 4 the focus moves from predicting population growth to the prediction of G. stearothermophilus single spore’s lag time (λ) under different isothermal conditions and the inclusion of heterogeneity, hence applying a stochastic modelling approach. The research conducted in Chapter 5 focuses on investigating the effect of storage temperature on the growth of G. stearothermophilus using the previously developed predictive model in order to assess the risk of spoilage for evaporated milk exported to the markets of the Mediterranean region. In each chapter, the findings of the individual studies are discussed and related to existing knowledge. The final, summarizing discussion facilitates the extension of the discussions from the chapters or supplements the discussion of new project - related issues. Finally, thoughts about areas deserving further investigation, interesting topics and potential applications for the food industry are outlined in Chapter 6.

Literature review and thesis outline

10 7.045 7,0 pH temperature Chapter 2 40

8 6.5 6,5 C) 35 6 6.0 6,0 Development of an evaporated milk specific model for

CFU/ml 30 pH

10 4 5.5 5,5

25 Log

Temperature ( Temperature the effect of storage temperature on growth kinetics of 2 5.0 5,0 20

0 4.515 4,5 0 50 100 150 200 250 300 G. stearothermophilus ATCC 7953 (population level) Time (h)

Chapter 3 Development of a growth/no growth interface model and a growth kinetic model for the effect of temperature and pH on growth behavior of A.

acidoterrestris ATCC 49025 (population level)

1 Chapter 4 0.9 0.8 0.7 59°C 0.6 55°C 0.5 Modeling the effect of temperature on Geobacillus 50°C 0.4 47.5°C

0.3 Cumulative Probability 0.2 45°C 0.1 stearothermophilus individual spore’s lag time (λ) 0 0 20 40 60 80 100 120 Lag time (h) (single spore level)

Chapter 5 Mapping the risk of spoilage for evaporated milk exported to the markets of Mediterranean region based on the growth model of G. stearothermophilus

Chapter 6 General discussion and potential applications of the developed models to the food industry

Figure 1.4. Overview of the research themes addressed in this thesis chapters.

Chapter 2 Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk

Myrsini Kakagianni, Maria Gougouli, Konstantinos P. Koutsoumanis

Published in Food Microbiology 57, 28-35 (2016)

Chapter 2

Abstract

The presence of Geobacillus stearothermophilus spores in evaporated milk constitutes an important quality problem for the milk industry. This study was undertaken to provide an approach in modeling the effect of temperature on G. stearothermophilus ATCC 7953 growth and in predicting spoilage of evaporated milk. The growth of G. stearothermophilus was monitored in tryptone soy broth at isothermal conditions (35 - 67 °C). The data derived were used to model the effect of temperature on G. stearothermophilus growth with a cardinal type model. The cardinal values of the model for the maximum specific growth rate were Tmin = 33.76

°C, Tmax = 68.14 °C, Topt = 61.82 °C and μopt = 2.068/h. The growth of G. stearothermophilus was assessed in evaporated milk at Topt in order to adjust the model to milk. The efficiency of the model in predicting G. stearothermophilus growth at non - isothermal conditions was evaluated by comparing predictions with observed growth under dynamic conditions and the results showed a good performance of the model. The model was further used to predict the time – to - spoilage (tts) of evaporated milk. The spoilage of this product caused by acid coagulation when the pH approached a level around 5.2, eight generations after G. stearothermophilus reached the maximum population density (Nmax). Based on the above, the tts was predicted from the growth model as the sum of the time required for the microorganism to multiply from the initial to the maximum level (tNmax), plus the time required after the tNmax to complete eight generations. The observed tts was very close to the predicted one indicating that the model is able to describe satisfactorily the growth of G. stearothermophilus and to provide realistic predictions for evaporated milk spoilage.

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk 1. Introduction

The thermal processing of evaporated milk cannot guarantee the sterility of this product, since it is not able to eliminate some spores of bacteria, such as Geobacillus stearothermophilus, which are extremely heat resistant (Membré & Van Zuijlen, 2011). Insufficient thermal treatment, high initial load of the spore - forming microorganism or spores, adhesive characteristics of spores that enhance their persistence in industrial plants or harsh conditions encountered in food ingredients processing and packaging technologies, as well as milk composition, are among the major factors explaining the emergence of thermophilic sporeformers, such as G. stearothermophilus, in thermally processed foods (André, Zuber, & Remize, 2013; Postollec et al., 2012; Simmonds, Mossel, Intaraphan, & Deeth, 2003; Yoo, Hardin, & Chen, 2006). Although a substantial effort in assessing inactivation kinetics of spores of G. stearothermophilus has been made (Ananta, Heinz, Schlüter, & Knorr, 2001; Georget et al., 2014; Iciek, Papiewska, & Molska, 2006; Watanabe et al., 2003), the presence of spores in the final product reflects a persistent quality problem for the canned food industry (André et al., 2013; Rigaux Ancelet, Carlin, Nguyen‐thé, & Albert, 2013). As soon as the spores are exposed to conditions suitable for growth (nutrients, temperature), they germinate, outgrow and further grow, after an irreversible cascade of events. Τhe metabolic active cells then proceed to cell division up to a critical level, which may cause significant spoilage defects to thermally processed foods leading to significant economic losses for the dairy industry. Research results showed that the growth of this species in heat - treated milk and other products caused physicochemical changes like acidification (from saccharides) without gas production (Fields, 1970; Nazina et al., 2001), which in some cases led in coagulation (Burgess, Lindsay, & Flint, 2010). Rigaux et al. (2014) reported that the time to spoilage of canned green beans corresponds to a population of 107 CFU/g, while Laudes et al. (2001) showed that color change of a laboratory medium is observed when G. stearothermophilus reaches a level of 108 CFU/ml. However, limited information is available for the population concentration of the microorganism at which these changes occur in evaporated milk. One critical condition for the initiation of growth of G. stearothermophilus is the storage temperature. Research data support that germination, outgrowth and subsequent vegetative growth of G. stearothermophilus spores does not occur below 35 – 40 °C (Burgess et al., 2010; Hill & Smythe, 2012; Llaudes, Zhao, Duffy, &

Chapter 2

Schaffner, 2001; Ng & Schaffner, 1997; Oomes et al., 2007). However, the conditions prevailing in the supply chain of the evaporated milk are out of direct control of the manufacturer and often deviate from specifications. In particular, the storage of evaporated milk for long periods at improper and changing temperature conditions, like the ones existing in summer months in some countries, where the temperature is higher than 35 °C, or in tropical and semitropical regions, may provoke the germination of spores, if these are present, the outgrowth and the subsequent growth of the vegetative cells of the organism. Obviously, G. stearothermophilus is a particular concern for the quality of evaporated milk and the estimation of the risk of spoilage constitutes a major target of the quality managers of the dairy industry, especially for the products that are going to be distributed in hot climate countries. For the development of a risk assessment of evaporated milk spoilage from G. stearothermophilus, a growth kinetic model is required that is able to predict the microbial behavior for both static and dynamic temperature conditions. However, within the domain of predictive microbiology, the supporting documentation for G. stearothermophilus growth kinetics is still very limited. Laudes et al. (2001) quantified the effect of inoculum size of G. stearothermophilus spores on spoilage time (change in color) in a laboratory medium, while Ng and Schaffner (1997) developed a model for the effect of pH (5.5 to 7.0), temperature (45 to 60 °C) and NaCl concentrations (0 to 1%) on growth of G. stearothermophilus in a laboratory medium (salty carrot medium). Later Ng et al. (2002) expanded the existing model (NaCl concentrations 0 to 1.5%) and validated it in tryptone soy broth and eight military ready - to - eat meals under constant temperature conditions. Nevertheless, information on the biokinetic range for growth considering the storage temperature and the impact of the food composition, which predominantly determine the behavior of the organism, is still very limited (Mtimet et al., 2015). The objective of the present study was to develop a predictive model for the effect of temperature on growth of G. stearothermophilus and validate it in predicting spoilage of evaporated milk at dynamic temperature conditions simulating distribution and storage of the product. Such a product - specific model can be used for the development of a risk assessment approach for ensuring evaporated milk quality.

100

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk 2. Materials and Methods

2.1. Bacterial strain The type strain G. stearothermophilus ATCC 7953 was used for all experiments in the present study. The stock culture of the strain was stored frozen (- 70 °C) onto MicrobankTM porous beads (Pro - Lab Diagnostics, Ontario, Canada). The working culture was stored refrigerated (5 °C) on nutrient agar (NA; Lab M Limited, Lancashire, United Kingdom) slants and was renewed bimonthly. The microorganism was activated by transferring a loopful from the Nutrient Agar (NA; Lab M Limited) slants into 10 ml nutrient broth (NB; Lab M Limited) and incubating at 55 °C for 24 h. The initial concentration of the inoculum was determined by surface plating on NA.

2.2. Growth experiments in TSB The growth kinetic behavior of the G. stearothermophilus was evaluated in tryptone soy broth (TSB; Lab M Limited) at temperatures of 35, 37.5, 40, 42.5, 45, 50, 52.5, 55, 57, 59, 64, 65, 66 and 67 °C. The above mentioned temperatures were selected in an attempt to cover the growth region of the species to the greatest possible extent, based on preliminary experiments.

Maximum specific growth rate (μmax) values corresponding to each temperature were estimated by means of absorbance detection times of serially decimally diluted cultures using the automated turbidimetric system Bioscreen C (Oy Growth Curves Ab Ltd., Raisio, Finland) as described in the study of Lianou and Koutsoumanis (2011). The difference with the above study was that the 24 - h culture of the microorganism was decimally diluted in TSB to a concentration of approximately 108 CFU/ml, while the range of initial concentrations obtained in the microtiter plates was approximately 106 - 102 CFU/well. For the temperatures from 35 to 59 °C the microtiter plates were placed in the Bioscreen C, whilst for temperatures from 64 to 67 °C, given the temperature limitations of the instrument, the microtiter plates were placed in high - precision incubators (model MIR 153, Sanyo Electric Co., Ora - Gun, Gunma, Japan), and the temperatures were monitored during incubation using electronic temperature - monitoring devices (Cox Tracer data logger; Cox Technologies, Belmont, NC, USA). Afterwards, optical density (OD) measurements were taken at 15-min and 20-min intervals for the temperatures from 35 to 59 °C and from 64 to 67 °C, respectively, using the wideband ﬁlter (420 - 580 nm) of the instrument, for a total time period such that a considerable OD change was

101

Chapter 2 observed, if possible, for all ﬁve decimally diluted cultures. The microtiter plates were agitated for 15 s at medium amplitude prior to the OD measurements. The detection times (h) of five serial decimal dilutions of the bacterial culture were plotted against the natural logarithm of their initial concentrations, and μmax values were determined by linear regression (Dalgaard & Koutsoumanis, 2001). One experiment was conducted for each temperature and five samples (e.g., quintuple wells of five serially diluted cultures) were analyzed (n = 5).

2.3. Growth experiments in evaporated milk For the experiments conducted in evaporated milk, G. stearothermophilus spores were used. The 24 - h cultures in NB of the strain were heat shocked at 80 °C for 10 min (Dogan, Weidendorfer, Müller - Merbach, Lembke, & Hinrichs, 2009), before inoculation into the product. The heat shock treatment was applied to G. stearothermophilus cultures in order to eliminate vegetative cells of Geobacillus endospores (Antolinos et al., 2012; Yuan et al., 2012). Then, the heat shocked cultures were centrifuged (6000 rpm for 20 min) in a refrigerated centrifuge (4 °C) (model PK120R, ThermoElectron Corporation, Waltham, MA). The pellet was resuspended with 5 ml of quarter - strength Ringer’s solution (Lab M, Limited) and used for inoculation. The evaporated milk used for inoculation was a commercial evaporated milk (ingredients: skimmed milk, corn syrup, vegetable oils, milk fat, prebiotic fibers, soy lecithin, vitamins - C, PP, E, calcium pantothenate, A, B6, B1, D3, B2, folic acid, K1, D - biotin, B12 -, minerals - potassium carbonate, ferrous sulphate, calcium citrate, zinc sulphate, copper sulphate, potassium iodide, sodium selenide). For this product, the initial aw and pH values were measured at 25 °C using an Aqualab Series 3 water activity determination device (Decagon Devices Inc., Pullman, WA, United States) and a pH meter with a glass electrode (pH 211 Microprocessor, Hanna Instruments

BV, Ijsselstein, the Netherlands), respectively. The pH (mean ± st.dev.) and aw (mean ± st.dev.) of evaporated milk were 6.16 (± 0.03) and 0.994 (± 0.03), respectively. Portions (200 ml) of the evaporated milk were dispensed in 200 ml Duran bottles and were inoculated with the appropriate dilution of the inoculum in order to obtain an initial concentration of ca. 103 CFU/ml. The artificially contaminated samples were then submerged in a preheated (62 ± 0.1 °C) water bath (NB 9, 20, Nüve Sanayi Malzemeleri Imalat Ve Ticaret A.Ş., Ankara, Turkey), where the milk temperature reached 62 °C. The above temperature was selected as optimum for growth based on the results derived from experiments conducted previously in TSB.

102

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk

For model validation, the behaviour of G. stearothermophilus spores in the evaporated milk was also studied under five different changing temperature scenarios designed in the laboratory simulating distribution and storage of the product in hot climate countries. For these experiments, high - precision programmable incubators (model MIR 153) were used. The fluctuating time – temperature protocols examined in this context were electronically recorded using cox tracer data loggers with the internal and external sensors monitoring temperature of the incubator and milk, respectively (with a time interval of 10 min). During incubation of the evaporated milk at 62 °C or at dynamic temperature conditions, the inoculated samples were examined at appropriate time intervals in order i) to allow for an efficient kinetic analysis of microbial growth, ii) to monitor pH, and iii) to observe if there is any macroscopic change in the structure of milk (e.g., coagulation). Appropriate serial decimal dilutions of samples in Ringer’s solution were surface plated on NA plates for the enumeration of G. stearothermophilus population. Colonies were counted after incubation of plates at 55 °C for 24 h. Four independent experiments were conducted with two replicates for the optimum temperature (n = 8) and two independent experiments were conducted with two replicates for the non - isothermal scenarios (n = 4).

2.4. Data analysis

The effect of temperature on μmax, derived from the experiments conducted in TSB, was modelled using the Cardinal Model with Inflection (CMI) of Rosso et al. (1993):

(2.1)

(2.2) where Tmin, Topt and Tmax are the theoretical minimum, optimum and maximum temperature (°C) for growth, respectively, and is the optimum value for the maximum specific growth rate (1/h) (when T = Topt). In order to stabilise the variance a square root transformation of was used.

103

Chapter 2

The values of Tmin, Topt and Tmax as well as the confidence and the predictions limits were determined by fitting the estimated μmax values for the tested microorganism to the above model using the Excel v4 format of the curve - fitting program TableCurve 2D (Systat Software Inc., San Jose, CA, USA). The adequacy of the developed models to fit data was evaluated graphically and also by the coefficient of determination R2 and the Root Mean Square Error (RMSE) (Ratkowsky, McKellar, & Lu, 2004). The growth data (log CFU/ml) in the evaporated milk stored under isothermal temperature (62 °C) were fitted to the primary model of Baranyi and Roberts (1994) using the program DMFit, in order to estimate the kinetic parameter for growth, maximum specific growth rate, μmax, in evaporated milk and the physiological state

(ho) of the spores. The original dynamic model has an explicit solution for static situations (when the model parameters do not depend on time), which describes the natural logarithm of the cell concentration, y(t) = lnx(t), by the equation:

(2.3) where μmax is the maximum specific growth rate of the cell population; ymax is the natural logarithm of the maximum population’s concentration; y0, the natural logarithm of the initial cell concentration; m is a curvature parameter characterizing the transition from the exponential to the stationary phase of growth and A(t) is a gradually delayed time variable described by the equation:

(2.4)

where ho is a parameter characterizing the ‘adaptation work’ required by the cells, which in our case are spores, to adjust to the new environment (Baranyi & Roberts, 1994).

The prediction of growth under dynamic temperature was based on the assumption that after a temperature shift, the growth rate is adopted instantaneously to the new temperature environment. Equations (2.3) and (2.4) were used for the prediction of growth at dynamic temperature conditions based on the “momentary” which was calculated from Equation (2.1) and (2.2). The

“momentary” was also used for the estimation of the number of generations

(G(T)) at dynamic temperature conditions using the following equation:

(2.5)

104

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk 3. Results and Discussion

3.1. Growth experiments in TSB In the first part of this work the effect of temperature on G. stearothermophilus growth rate was investigated using the Bioscreen C method. All experiments were carried out with G. stearothermophilus cells in a laboratory medium (TSB) under isothermal storage conditions (35 °C to 67 °C). The average (± st.dev.) μmax increased from 0.293 (± 0.016)/h at 37.5 °C to 1.449 (± 0.0004)/h at 64 °C, while at temperatures > 64 °C a gradual decrease of μmax was observed. In the next step, the above experimental data (μmax) were modelled as a function of temperature using a

CMI (Equation (2.1)), provided that this model incorporates parameters (Tmin, Topt,

Tmax) which are regarded as biologically interpretable (Cuppers, Oomes, & Brul, 1997; Ratkowsky et al., 2004). The R2 and RMSE values (Table 2.1), as well as the graphical evaluation from the fitting curve (Fig. 2.1), indicated the satisfactory performance of the CMI in describing the effect of temperature on G. stearothermophilus μmax. The estimated values for the cardinal parameters Tmin, Tmax,

Topt and the optimum maximum specific growth rate ( ) of G. stearothermophilus were found to be 33.76, 68.14, 61.82 °C and 2.068 1/h, respectively (Table 2.1). It should be mentioned that the Tmin value constitutes the theoretical minimum temperature for growth, considering that at 35 °C no growth was observed (data not shown). The above results are similar with those obtained in the study of Mtimet et al. (2015), in which it was found that the Tmin, Tmax and Topt (mean ± st.dev.) were 38.52 ± 3.22, 68.02 ± 5.62 and 57.59 ± 1.75, respectively, although these data generated with a different strain of G. stearothermophilus with the surface plating technique.

105

Chapter 2

Table 2.1. Estimated values and statistics for the parameters of the Cardinal Model with Inflection (Equation (2.1)) describing the effect of temperature on the maximum specific growth rate (μmax) of Geobacillus stearothermophilus ATCC 7953 in tryptone soy broth. Parameter Estimated Valuea Lower 95% CLb Upper 95% CLb RMSEc R2d

(1/h) 2.068±0.036 1.996 2.140 0.0033 0.977

Tmax 68.14±0.15 67.83 68.44

Tmin 33.76±0.36 33.03 34.48

Topt 61.82±0.20 61.43 62.21 a ±: Standard Error b CL: Confidence Limits c RMSE: Root Mean Square Error d R2: Coefficient of determination

2.5

2.0 )

h 1.5

( max

μ 1.0

0.5

0.0 30 40 50 60 70 80 Temperature ( C)

Figure 2.1. Effect of temperature on the maximum specific growth rate (μmax) of Geobacillus stearothermophilus ATCC 7953 in tryptone soy broth, fitted in the Cardinal Model with Inflection (solid line) (Equation (2.1)). Points represent observed values of the μmax. The dotted and the discontinuous lines depict the 95% confidence and the prediction limits, respectively, of the effect of temperature on the maximum specific growth rate.

106

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk

Interestingly, it was observed that G. stearothermophilus can grow adequately at temperatures which are sublethal or lethal for the majority of the microorganisms. There is strong evidence that the growth ability of thermophilic microorganisms, such as G. stearothermophilus, at high temperatures, is based on keeping their membrane fluidity constant (homeoviscous adaptation) (Sinensky, 1974). Particularly, the correct membrane function could be achieved due to the higher ratio of longer straight - chained saturated fatty acids in membrane lipids (Martins, Jurado, & Madeira, 1990; Russell & Fukunaga, 1990; Suutari & Laakso, 1994; Zeikus, 1979). Except for the above theory, another factor, that could be responsible for the growth ability of thermophiles, is the production of sufficient amounts of thermostable gene products under elevated temperature conditions. Particularly, it has been observed a tendency of the purines levels to increase at the codon positions within the genome of thermophiles, compared to mesophiles, something which may correlate with mRNA thermostability (Cate et al., 1996; Wang & Hickey, 2002). Likewise, the trend that cytosine is preferred over thymine in many codons could play a crucial role in the greater thermostability, maybe due to the increased number of potential formed hydrogen bonds (Querol, Perez-Pons, & Mozo-Villarias, 1996; Sadeghi, Naderi-Manesh, Zarrabi, & Ranjbar, 2006; Singer & Hickey, 2003). In addition to that, at the protein level, the increased frequency of hydrophobic and/or charged amino acids (e.g., glutamic acid, isoleucine, valine) and the simultaneously decreased frequency or removal of glutamine, which is a thermolabile amino acid, has been found that it has a great effect on thermostability of the encoded proteins probably because it reduces the possibility of the thermal unfolding process (Lynn, Singer, & Hickey, 2002; Singer & Hickey, 2003).

3.2. Validation of the growth model for evaporated milk stored under dynamic temperature conditions By examining in quantitative terms the effect of temperature on G. stearothermophilus growth, a first picture of the biokinetic growth region of the microorganism is being provided (Fig. 2.1). However, in order to evaluate the performance of the developed model in predicting the growth behaviour of G. stearothermophilus spores in evaporated milk additional experiments were performed. More specifically, growth trials with the artificially contaminated evaporated milk with spores were conducted at a reference temperature of 62 °C (Fig. 2.2), which was found to be the optimum temperature for the microorganism’s

107

Chapter 2 growth (Table 2.1). The obtained growth data were further expressed as a function of time, and via the use of the Baranyi and Roberts model (Equations (2.3) and (2.4)) the growth kinetic parameters were determined. The maximum specific growth rate

(mean ± st.dev.), ° , of G. stearothermophilus in the evaporated milk stored at 62 °C was 2.083 ± 0.288 1/h, which was not different to that observed in TSB (2.068 ± 0.036 1/h; Table 2.1), considering the standard deviation. Given the above similarity, for predictive modelling purposes the growth rate derived from TSB was selected for growth prediction at dynamic temperature conditions.

8 6.5 7 6 6.0 5

4 5.5

CFU/ml pH

10 3

Log 2 5.0 1 0 4.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Time (h) Figure 2.2. Growth kinetics of Geobacillus stearothermophilus ATCC 7953 vegetative cells derived from spores ( ) in evaporated milk and pH changes ( ) during storage at optimum growth temperature (62 °C). The black solid line (▬) depicts the fitting of the Baranyi and Roberts model (Equation (2.3)) to the growth data. The point ( ) is showing the observed time of evaporated milk coagulation. Each point is a mean of eight values. Vertical and horizontal bars indicate the standard deviation.

At this point, it should be noted that the selected value of growth rate, used for the construction of the model (Equation (2.1)), may not be valid to other evaporated milk products with different composition than the examined one (Abee et al., 2011; Østergaard, Eklöw, & Dalgaard, 2014). As discussed in the literature, the presence and the concentrations of specific compounds in milk may exert different behavioural responses of the microorganism. Ljunger (1970) and Vinter (1969) reported that the existence of ions in milk, such as divalent cations (calcium,

108

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk magnesium, potassium), can contribute to the outgrowth of mature spores and may be involved in the activation of sporulation. Moreover, there are several studies (Ashton & Busta, 1968; Cleverdon, Pelczar Jr, & Doetsch, 1949; Stahl & Ljunger, 1976) supporting that the presence of divalent cations, such as calcium, magnesium and iron(II), and vitamins, like niacin and biotin, in milk have considerable effect on G. stearothermophilus growth and further on spoilage of evaporated milk. Arancia et al. (1980) reported that the presence of calcium cations stimulated Escherichia coli growth and reduced lag periods, while later the findings of Jurado et al. (1987) confirmed the above case for G. stearothermophilus but also demonstrated that magnesium cations above a critical concentration exert an inhibitory effect on microorganism’s growth. Given the above, it is obvious that in any case of use of the developed model, the appropriate corrections that correspond to specific products should be made. The developed model was further validated at dynamic temperature conditions. Prediction of growth at dynamic temperature conditions was based on the combination of the secondary model (Equation (2.1)) with the differential equations of the Baranyi and Roberts primary model (Equations (2.3) and (2.4)), which were numerically integrated with respect to time (Gougouli, Angelidis, & Koutsoumanis, 2008; Koutsoumanis, 2001; Xanthiakos, Simos, Angelidis, Nychas, & Koutsoumanis,

2006). However, for predicting growth of G. stearothermophilus a selection of a ho value is required.

In bacterial growth, ho represents the amount of “work” that a cell has to perform to adapt to its new environment. The “work” for adaptation is determined by the product of μmax and λ (lag phase) that is also called “physiological state” of the cells (Baranyi & Roberts, 1994). Several studies have reported a relation between

μmax and λ with their product μmax*λ being constant at different storage temperatures when the pre - inoculation history of the cells culture was the same (Baranyi & Roberts, 1994, 1995; Gougouli et al., 2008; Koutsoumanis, Stama ou, Skandamis, & Nychas, 2006; Pin, Garcıa de Fernando, Ordóñez, & Baranyi, 2002). In the experiments conducted in this study, the inoculum was constituted from spores which have been produced under a well - defined environment. Based on the above, we assumed that the ho, which in this case refers to spore germination and outgrowth, is not affected by the storage temperature and we set its value to the one determined from the growth of G. stearothemophilus spores in evaporated milk under constant conditions 62 °C, which was found to be 3.787. The Nmax was also set at 7.4 log CFU/ml based on the observation at 62 °C.

109

Chapter 2

The predicted growth was compared to observed growth data derived from five experiments at changing temperatures (Figs. 2.3 – 2.7) simulating conditions of temperature abuses during distribution and storage of the product in regions with hot climate and/or during warm summer months (Weather Underground database, http://www.wunderground.com/). Abrupt temperature upshifts and downshifts were included in the tested profiles in order to evaluate model’s assumptions (i.e. growth rate is adopted instantaneously to the new temperature) at extreme conditions representing a worst - case scenario for the performance of the model. In general, at all temperature scenarios tested, the model adequately predicted the growth of G. stearothermophilus in evaporated milk, suggesting that the assumptions made for growth prediction were valid. Accurate predictions were obtained in the cases of temperature shifts inside the growth region of the microorganisms (Figs. 2.3 – 2.5) as well as in scenarios including temperatures lower from the Tmin (Figs. 2.6 – 2.7). For the last scenarios (Figs. 2.6 – 2.7) the results showed that the bacterium adapts instantaneously to the new environment without presenting any additional lag phase and grow with the expected μmax. Even after a storage period of about 140 h at temperatures below Tmin, G. stearothermophilus was able to initiate growth when temperature increased to levels within the biokinetic range with a lag phase and a growth rate very close to those predicted by the model (Fig. 2.6).

110

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk

10 607.0 7.0 Temperature ( C) pH pH 8 6.5 6.5

50 C)

6 6.0 6.0

CFU/ml 40

pH pH

10 4 5.5 5.5 Log

30 ( Temperature 2 5.0 5.0

0 204.5 4.5 0 10 20 30 40 50

Time (h) Figure 2.3. Comparison between observed (points) and predicted (solid line) growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature condition 1. Discontinuous lines indicate milk pH (- - - ) and temperature changes (------).

10 7.060 7.0 Temperature ( C) pH pH 8 6.5 6.5

50 C)

6 6.0 6.0

CFU/ml 40 pH

10 4 5.5 5.5 Log

30 ( Temperature 2 5.0 5.0

0 4.520 4.5 0 10 20 30 40 Time (h) Figure 2.4. Comparison between observed (points) and predicted (solid line) growth of Geobacillus stearothermophilus ATCC 7953 in evaporated milk stored under periodically changing temperature condition 2. Discontinuous lines indicate milk pH (- - -) and temperature changes (------).

111

Chapter 2

10 7.050 7.0 TemperaturepH ( C) pH

8 6.545 6.5

6 6.040 6.0

CFU/ml pH

10 4 5.535 5.5

Log Temperature ( Temperature 2 5.030 5.0

0 4.525 4.5 0 20 40 60 80 100 120 Time (h) Figure 2.5. Comparison between observed (points) and predicted (solid line) growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature (24 h at 37 °C, 12 h at 42 °C and 24 h at 45 °C). Discontinuous lines indicate milk pH (- - -) and temperature changes (------).

10 457.0 7.0 Temperature ( C) pH pH 40

8 6.5 6.5 C) 35 6 6.0 6.0

CFU/ml 30 pH

10 4 5.5 5.5

25 Log

2 5.0 ( Temperature 5.0 20

0 154.5 4.5 0 50 100 150 200 250 300 Time (h) Figure 2.6. Comparison between observed (points) and predicted (solid line) growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature (78 h at 20 °C, 59 h at 25 °C and 163 h at 40 °C). Discontinuous lines indicate milk pH (- - -) and temperature changes (------).

112

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk

10 7.060 7.0 TemperaturepH ( C) pH 8 6.5 6.5

( 6 6.0 6.0

CFU/ml 40

pH pH

10 4 5.5 5.5 Log

30 Temperature 2 5.0 5.0

0 4.520 4.5 0 10 20 30 40 50 60 70 80 Time (h) Figure 2.7. Comparison between observed (points) and predicted (solid line) growth of Geobacillus stearothermophilus ATCC 7953 in the evaporated milk stored under periodically changing temperature (6 h at 50 °C, 12 h at 30 °C and 24 h at 42 °C). Discontinuous lines indicate milk pH (- - -) and temperature changes (------).

3.3. Prediction of the time – to - spoilage of the evaporated milk The results of the experiment with evaporated milk inoculated with G. stearothermophilus spores and stored at 62 °C showed that spoilage of this product is due to acid coagulation observed when the pH approaches a level around 5.2 (Fig. 2.2). This is in agreement with previous studies (Hill & Smythe, 2012; Yoo et al., 2006) which have demonstrated that G. stearothermophilus cells are producing acid, enhancing in this way the formation of protein aggregates, something that is related with the unfolding and gelation of β - lactoglobulin which has been found to be pH and temperature dependent. The decrease of milk pH to the spoilage level was observed at a certain time (ts) after G. stearothermophilus reached the maximum 7.4 population density (Nmax = 10 CFU/ml). Considering that each generation time (G) can be calculated under constant conditions as G = μmax/ln(2) the time ts corresponded to an average of eight generations. Based on the above findings, the time – to - spoilage (ttspred) was predicted from the growth model as following: (2.6)

113

Chapter 2

In Equation (2.6) the ttspred derives from the sum of the time ( ) that is required for the microorganism to multiply from the initial level to the maximum 7.4 level (Nmax = 10 CFU/ml), which can be determined from the growth model, and the time that is required from the cells after the to complete eight

generations under the existing temperature conditions. Generation Time (G(T)) under dynamic temperature conditions was estimated from Equation (2.5). The applicability of the model to predict spoilage of evaporated milk was evaluated by comparing the predicted time to spoilage (ttspred) from Equation (2.6) with the time at which coagulation was observed (ttsobs) for the five dynamic temperature experiments. A numerical comparison between the ttsobs and ttspred is presented in Table 2.2. The ttsobs ranged from 29 to 223 h for the various temperature scenarios. As shown in Figs. 2.3 – 2.7, milk coagulation coincided with a pH decrease to levels around 5.2 confirming the findings at static temperature conditions (Fig. 2.2). For all the temperature scenarios examined the observed time to spoilage was very close to the predicted one. The approach exploited in this research for predicting the spoilage time of evaporated milk did not show any specific trend of overestimation or underestimation considering the percent relative errors, which were ranging from -8.7 to 4.5 (Table 2.2). The variation in the initial pH of evaporated milk was in general limited. In particular, the initial pH (mean ± st.dev.) of 4 milk batches tested at isothermal conditions and 5 milk batches tested at dynamic temperature conditions was 6.16 ± 0.03 and 6.08 ± 0.13, respectively. The validation of the model indicated that the pH within the above ranges did not significantly affect the performance of the model. However, application of the model for milk with initial pH outside these ranges requires further validation studies.

114

Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk

Table 2.2. Comparison between observed and predicted spoilage time of the evaporated milk, stored under nonisothermal conditions, by Geobacillus stearothermophilus ATCC 7953. a b c d Temperature profile (figure) (h) (h) %RE 1 (Fig. 2.3) 35.0 36.75 -4.8 2 (Fig. 2.4) 29.0 28.0 3.6 3 (Fig. 2.5) 88.0 91.5 -3.8 4 (Fig. 2.6) 223.0 244.33 -8.7 5 (Fig.2.7) 78.0 74.67 4.5 aEach figure corresponds to the indicated temperature profile b , observed spoilage time c , predicted spoilage time based on the Equation (2.6). dRE: Relative Error =

4. Conclusions

In conclusion, the model developed in the present study is able to describe satisfactorily the effect of storage temperature on the growth of G. stearothermophilus in evaporated milk and to provide realistic predictions for the rejection time of the product due to spoilage. Beside the current value of this approach for the prediction of evaporated milk’s quality, the developed model can be the basis for the construction of a quantitative microbial risk assessment (QMRA) model for spoilage of evaporated milk from G.stearothermophilus. However, as frequently commented by various researchers, the strain variability may have an important impact on the microbial risk assessment outcomes, and, for that reason it should be assessed and taken into consideration for such approaches (Coleman, Tamplin, Phillips, & Marmer, 2003; Delignette-Muller & Rosso, 2000; Lianou & Koutsoumanis, 2013; Pouillot & Lubran, 2011). So far, strain - depended differences in growth behaviour of G. stearothermophilus have not been documented and the precision of a QMRA model would be inaccurate. Thus, for moving from deterministic to stochastic modelling approaches further research on strain variability is required.

115

Chapter 3 Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

Myrsini Kakagianni, Kelly Kalantzi, Evangelos Beletsiotis, Dimitrios Ghikas, Alexandra Lianou, Konstantinos P. Koutsoumanis

Published in Food Microbiology 74, 40-49 (2018)

Chapter 3

Abstract

This study was undertaken to provide quantitative tools for predicting the behavior of the spoilage bacterium Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks. In the first part of the study, a growth/no growth interface model was developed, predicting the probability of growth as a function of temperature and pH. For this purpose, the growth ability of A. acidoterrestris was studied at different combinations of temperature (15 - 45 °C) and pH (2.02 - 5.05). The minimum pH and temperature where growth was observed was 2.52 (at 35 and 45 °C) and 25 °C (at pH ≥ 3.32), respectively. Then a logistic polynomial regression model was fitted to the binary data (0: no growth, 1: growth) and, based on the concordance index (98.8%) and the Hosmer - Lemeshow statistic (6.226, P = 0.622), a satisfactory goodness of fit was demonstrated. In the second part of the study, the effects of temperature (25 - 55 °C) and pH (3.03 - 5.53) on A. acidoterrestris growth rate were investigated and quantitatively described using the cardinal temperature model with inflection and the cardinal pH model, respectively. The estimated values for the cardinal parameters Tmin, Tmax, Topt and pHmin, pHmax, pHopt were 18.11, 55.68, 48.60 °C and 2.93, 5.90, 4.22, respectively. The developed models were validated against growth data of A. acidoterrestris obtained in eight commercial pasteurized fruit drinks. The validation results showed a good performance of both models. In all cases where the growth/no growth interface model predicted a probability lower than 0.5, A. acidoterrestris was, indeed, not able to grow in the tested fruit drinks; similarly, when the model predicted a probability above 0.9, growth was observed in all cases. A good agreement was also observed between growth predicted by the kinetic model and the observed kinetics of A. acidoterrestris in fruit drinks at both static and dynamic temperature conditions.

118

For many years, heat processed fruit drinks were considered as microbiologically stable foods mainly due to their low pH (<4.0). During the 80’s however, Alicyclobacillus acidoterrestris was identified as the causative agent of a large spoilage incident of apple juice in Germany (Cerny, Hennlich, & Poralla, 1984). Since then, this spore - forming bacterium has been recognized as a major quality problem by manufacturers and processors in the fruit industry (Huang, Yuan, Guo, Gekas & Yue, 2015; Steyn, Cameron, & Witthuhn, 2011; Vieira, Teixeira, Silva, Gaspar, & Silva, 2002; Wang et al., 2014). The main characteristics of Alicyclobacillus spp. are the heat resistance of its spores and their ability to germinate and outgrow at acidic environments. After spore germination and outgrowth, the metabolically active cells can multiply up to critical cell concentrations and produce spoilage taint compounds leading to organoleptic rejection of the products with consequent large economic and credibility losses for the food company (Gobbi et al., 2010). Undesirable effects on the sensory attributes of fruit juices and drinks are mainly attributed to the production of the metabolic product guaiacol which causes “smoky, medicinal, phenolic, antiseptic disinfectant” off - flavors and/or off - odors (Bahçeci, Gökmen, & Acar, 2005; Bevilacqua, Sinigaglia, & Corbo, 2009; Gocmen, Elston, Williams, Parish, & Rouseff, 2005; Jensen, 2000), with normal or light sediment (Gocmen et al., 2005; Walker & Phillips, 2005). The lower limit of guaiacol detection in fruit juices by a trained sensory panel is 2 μg/l (2 ppb) while detectable off-odors in fruit juices and drinks are generally reported when the levels of A. acidoterrestris reach about 104 - 105 CFU/ml (Bahçeci et al., 2005; Sinigaglia et al., 2003). Due to its high spoilage potential, A. acidoterrestris has been suggested as a possible target microbe in the design of pasteurization processes for acidic products such as fruit drinks (Vieira et al., 2002). However, standard pasteurization processes applied in the case of fruit drinks are not effective against Alicyclobacillus spores, while processing at higher temperatures is not feasible due to the negative effect on the quality of these products (Palop, Alvarez, Raso, & Condon, 2000). As a result, control of Alicyclobacillus growth during distribution and storage is a key factor for the efficient risk management of fruit drinks’ spoilage. The fruit drink pH and the temperature during storage and distribution are the most important factors affecting the growth of A. acidoterrestris. Research data have provided evidence that germination, outgrowth and subsequent vegetative

119

Chapter 3 growth of A. acidoterrestris spores would not be expected to occur when pasteurized fruit products of naturally low pH (< 4.0) are stored below 20 °C (Bahçeci & Acar, 2007; Bahçeci et al., 2005; Spinelli, Sant'Ana, Rodrigues-Junior, & Massaguer, 2009). However, the conditions prevailing in the supply chain of pasteurized fruit drinks are out of the manufacturers’ direct control and often deviate from specifications (Bahçeci et al., 2005; Heyndrickx, 2011), especially during the warmer months or in tropical and semitropical regions (Roig - Sagues, Asto, Engers, & Hernández-Herrero, 2015). Thus, the estimation of the risk of spoilage constitutes a major target of quality managers, especially for products that are going to be distributed in hot climate countries. For assessing the risk of fruit drink spoilage caused by A. acidoterrestris, a growth model is required that is able to predict the microbial behavior during distribution and storage. However, within the domain of predictive microbiology literature, models for A. acidoterrestris growth kinetics are not available. The objective of the present study was the development of predictive mathematical models for the description of the effects of temperature and pH on the growth of A. acidoterrestris and the evaluation of their performance in predicting growth in fruit drinks under isothermal and non - isothermal conditions simulating transportation, distribution and storage of the products before delivery to the consumer. Such validated models could be used for an effective risk management of fruit drinks spoilage.

2. Materials and Methods

2.1. Bacterial strain The type strain Alicyclobacillus acidoterrestris ATCC 49025 was used for all experiments in the present study. The stock culture of the strain was stored frozen (- 70 °C) onto MicrobankTM porous beads (Pro - Lab Diagnostics, Ontario, Canada). The working culture was stored refrigerated (5 °C) on K Agar (2.5 g/l yeast extract; 5.0 g/l peptone; 1.0 g/l glucose; 1.0 g/l tween 80; 15 g/l agar) plates (IFU, 2007; Walls & Chuyate, 2000) and was renewed biweekly. After sterilization, the pH of the medium (K agar) was adjusted to 4.0 with filtered 25% (w/v) citric acid (Merck, Darmstadt, Germany) using a digital pH meter with an epoxy refillable pH probe (Orion 3 - Star pH Benchtop; Thermo Electron Corporation, Beverly, MA, USA). The microorganism was activated by transferring a loopful from the K Agar plates into 10 ml Κ broth (2.5

120

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks g/l yeast extract; 5 g/l peptone; 1 g/l glucose; 1 g/l tween 80), adjusted to pH = 4.0 with filtered 25% (w/v) citric acid, and incubating at 45 °C for 48 h. The 48 - h culture of the strain was then heat shocked at 80 °C for 10 min (IFU, 2007; Murray, Gurtler, Ryu, Harrison, & Beuchat, 2007; Walls & Chuyate, 2000). The heat shock treatment was applied to A. acidoterrestris cultures in order to eliminate any vegetative cell and obtain uniform activation and germination of dormant endospores (Goto et al., 2008). Then, the heat shocked cultures were centrifuged (6000 rpm for 20 min) in a refrigerated centrifuge (4 °C) (model PK120R, ThermoElectron Corporation, Waltham, MA). The pellet was resuspended with 5 ml of quarter - strength Ringer’s solution (Lab M, Limited, Lancashire, United Kingdom) and used for inoculation. The initial concentration of the inoculum was determined by surface plating on K Agar.

2.2. Development of the growth/no growth interface model 2.2.1. Experimental design K broth was used as the basal medium for all experiments, while all experiments were performed using spores of A. acidoterrestris ATCC 49025 obtained as described in section 2.1. The growth ability of the spoilage microorganism was tested at different combinations of temperature (15, 18, 20, 25, 30, 35, 40 and 45 °C) and pH (2.02, 2.31, 2.52, 2.72, 3.05, 3.32, 3.60, 3.87, 4.22, 4.62 and 5.05) with five replicates for each combination. These values were selected based on pH measurements of eight different industrial fruit drinks used during this study and the biokinetic range of the microorganism’s growth. For all conditions, the pH of the medium was adjusted to the appropriate values with filtered 25% (w/v) citric acid, and was measured both before and after autoclaving (prior to inoculation) using a digital pH meter with an epoxy refillable pH probe. The abovementioned pH values were the ones measured after autoclaving and used for the purpose of model development. Portions of 180 μl of the modified K broth (K broth with modified pH) for each treatment were pipetted into wells of 100-well microtiter plates and 20 μl of the appropriate dilution of the inoculum were added to each well, with the initial inoculation spore level being approximately 104 CFU/well. In order to verify the exact inoculum density, immediately after inoculation 100 μl from each of the five wells were surface plated on K Agar (pH=4.0) and colonies were counted after incubation at 45 °C for 48 h. The microtiter plates were then sealed with Parafilm (Parafilm ‘M’; American National Can, Greenwich, CT, USA) to avoid evaporation, and stored in high - precision (± 0.2 °C) programmable incubators (model MIR 153, Sanyo Electric Co., Ora-Gun, Gunma, Japan) for 35 days. The temperature during

121

Chapter 3 storage of non - inoculated microplates was recorded using electronic temperature - monitoring devices (Cox Tracer data logger; Cox Technologies, Belmont, NC, USA).

2.2.2. Assessment of growth During the 35 – day storage, the microtiter plate wells were measured for growth on a weekly basis by recording the optical density (OD) of the medium using the automated turbidimetric system Bioscreen C (Oy Growth Curves Ab Ltd., Raisio, Finland) set to read at the wideband ﬁlter (420 - 580 nm). Prior to each measurement, the microtiter plates were shaken for 15 s. Five wells containing 200 μl of sterile K broth (pH = 6.73) served as negative controls. In order to define growth, the OD of a well was compared to the ODzero which is the average of OD values recorded in all 100 wells at time - zero. A given well (corresponding to a certain temperature - pH combination) was considered as positive for growth, if the difference between its OD and the ODzero was three times higher than the standard ® deviation of the ODzero (Daelman et al., 2013). Data were processed using Microsoft Excel (Microsoft, Redmond, Virginia, USA).

2.2.3 Data analysis For each replicate response of A. acidoterrestris, growth or no growth were scored as values of 1 or 0, respectively. Then, a logistic polynomial regression model was fitted to the obtained binary data using the commercial software SAS version 8 (SAS Inst. Cary, NC, USA) based on the approach described by Ratkowsky and Ross (1995). The combined effect of storage temperature and pH on the probability of growth of the organism was described using a polynomial model:

(3.1) where P is the probability of growth (in the range of 0 - 1), Logit (P) is an abbreviation of ln[P/(1-P)], ai are coefficients to be estimated, pH is the pH of the medium and T (°C) is temperature.

The automatic variable selection option with a stepwise selection method was used to choose the most significant effects (P < 0.05). The concordance index (c – value) and the Hosmer - Lemeshow (HL) statistic were used as measures of the goodness of fit of the developed model. In order to compare observed growth/no growth data with model predictions, estimated probabilities (P) at 0.1, 0.5 and 0.9 levels were calculated using Microsoft® Excel Solver.

122

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

2.3. Development of the kinetic model 2.3.1 Experimental design The effect of temperature and pH on the growth kinetic behavior of A. acidoterrestris was assessed: (i) in K broth with pH = 4.5 (optimum growth pH) adjusted with filtered 25% (w/v) citric acid, at incubation temperatures of 25, 27, 30, 35, 40, 45, 48, 50, 53, 55 °C, and (ii) in K broth with pH 3.03, 3.32, 3.6, 3.78, 3.99, 4.32, 4.52, 4.8, 5.04, 5.29, 5.53, adjusted with filtered 25% (w/v) citric acid, at an incubation temperature of 48 °C (optimum growth temperature). The above growth conditions were selected in an attempt to cover the growth region of the strain to the greatest possible extent.

The maximum specific growth rate (μmax) values corresponding to each growth condition were estimated by means of absorbance detection times of serially decimally diluted cultures using the automated turbidimetric system Bioscreen C as described previously (Kakagianni et al., 2016; Lianou & Koutsoumanis, 2011). The 48- h culture of the microorganism was decimally diluted in K broth, and an appropriate dilution was used to inoculate microtiter plates so that the range of initial bacterial attained was approximately 101-105 CFU/well. Optical density measurements were taken at 15 - min intervals using the wideband ﬁlter (420 - 580 nm) of the instrument, for a total time period such that a considerable OD change was observed, if possible, for all ﬁve decimally diluted cultures. The microtiter plates were agitated for 15 s at medium amplitude prior to the OD measurements. The detection times (h) of five serial decimal dilutions of the bacterial culture were plotted against the natural logarithm of their initial concentrations, and μmax values were determined by linear regression (Dalgaard & Koutsoumanis, 2001). One experiment was conducted at each growth condition and five samples (e.g., quintuple wells) were analysed (n = 5).

2.3.2 Data analysis

The effect of temperature on μmax, derived from the experiments conducted in K broth (pH = 4.5), was modelled using the cardinal model with inflection of Rosso et al. (1993):

(3.2)

123

Chapter 3

(3.3)

where Tmin, Topt and Tmax are the theoretical minimum, optimum and maximum temperature (°C) for growth, respectively, and is the optimum value for the maximum specific growth rate (i.e when T = Topt).

The effect of pH on μmax, derived from the experiments conducted in K broth, was modelled using the cardinal type model of Rosso (Rosso et al., 1995):

(3.4)

(3.5)

where pHmin, pHopt and pHmax are the theoretical minimum, optimum and maximum pH, respectively, for growth and μopt is the optimum value for the maximum specific growth rate (i.e when pH = pHopt).

The values of Tmin, Topt, Tmax, pHmin, pHopt and pHmax as well as the confidence and the prediction limits were determined by fitting the estimated μmax values to the above models to using the Excel v4 format of the curve - fitting program TableCurve 2D (Systat Software Inc., San Jose, CA, USA). The adequacy of the developed models to fit data was evaluated both graphically numerically based on the values of the coefficient of determination R2 and the Root Mean Square Error (RMSE) (D. A. Ratkowsky, McKellar, & Lu, 2004). Making the well - established assumption that discrete environmental conditions exert independent effects on microbial growth (Buchanan et al., 1993), a multiplicative without interaction - type model, combining the above models for temperature and pH, was used to describe the combined effect of these two environmental parameters on μmax (Rosso, Lobry, Bajard, & Flandrois, 1995):

124

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

max  opt  ( pH)  () (3.6)

where μopt is the maximum specific growth rate corresponding to optimum growth conditions.

2.4. Model validation in fruit drinks 2.4.1 Validation of the growth/no growth interface model The growth/no growth interface model was validated against the observed growth behavior of A. acidoterrestris in eight commercial pasteurized fruit drinks stored at constant temperatures from 21 to 48 °C for 35 days. The pH values of the tested drinks were measured prior to inoculation as described previously (sections

2.1 and 2.2). The water activity (aw) of the fruit drinks was measured with an AquaLab water activity meter (Model series 3; Decagon Devices, Inc., Pullman, WA, USA). The soluble solids content was measured with an Atago Digital Abbe Refractometer (Atago, Tokyo, Japan). Aliquots (200 ml) from each of the fruit drinks were dispensed in 500 - ml Duran bottles and were inoculated with the appropriate dilution of the inoculum in order to obtain an initial concentration of approximately 103 CFU/ml. The artificially contaminated samples were stored under controlled isothermal conditions in high – precision programmable incubators (model MIR 153, Sanyo Electric Co.). The temperature of the incubator and fruit drink samples was recorded using electronic temperature - monitoring devices (Cox Tracer data logger; Cox Technologies). During incubation, the inoculated samples were examined at appropriate time intervals in order to allow for an efficient kinetic analysis of microbial growth. Appropriate serial decimal dilutions of samples in Ringer’s solution were surface plated on K Agar plates (pH = 4.0) for the enumeration of A. acidoterrestris populations. Colonies were counted after incubation of plates at 45 °C for 48 h. Two independent experiments were conducted with two replicates (n=4). Growth was considered when a total increase of 1 log CFU/ml was observed during the storage period or between two sampling time intervals. The observed growth/no growth behavior in fruit drinks was compared to the probability of growth predicted by the model based on the pH of the drink and the storage temperature. Uninoculated samples tested as negative controls did not show any presence of indigenous A. acidoterrestris.

125

Chapter 3

2.4.2 Validation of the kinetic model For validation of the kinetic model, the growth of A. acidoterrestris in fruit drinks was studied under static and dynamic temperature scenarios designed in the laboratory to simulate distribution and storage conditions likely to be encountered in the supply chain. For these experiments fruit drink samples were stored in high - precision programmable incubators or at room temperature. The temperatures of the incubator/environment and fruit drinks were recorded at 10 - min intervals using electronic temperature - monitoring devices as described previously. The inoculation and sampling procedure was the same with that described in the previous section 2.4.1. The observed growth of A. acidoterrestris in fruit drinks was compared graphically with the growth predicted by the model. For the purpose of growth prediction, the primary model of Baranyi and Roberts (1994) was used:

(3.7)

where μmax is the maximum specific growth rate of the cell population; ymax is the natural logarithm of the maximum population’s concentration; y0, the natural logarithm of the initial cell concentration; m is a curvature parameter characterizing the transition from the exponential to the stationary phase of growth and A(t) is a gradually delayed time variable described by the equation:

(3.8)

where ho is a parameter characterizing the ‘adaptation work’ required by the cells, which in our case are spores, to adjust to the new environment (Baranyi & Roberts, 1994).

Prediction of growth under dynamic temperature was based on the above equations which were numerically integrated with respect to time and the assumption that after a temperature shift, the growth rate is adopted instantaneously to the new temperature environment. The “momentary” was calculated using the abovementioned multiplicative secondary model (Equation (3.6)).

126

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks 3. Results and Discussion 3.1. Model development 3.1.1 Growth/no growth interface model The growth ability of A. acidoterrestris was studied at different combinations of temperature (15, 18, 20, 25, 30, 35, 40 and 45 °C) and pH (2.02, 2.31, 2.52, 2.72, 3.05, 3.32, 3.60, 3.87, 4.22, 4.62 and 5.05). In total, 440 samples were tested (88 combinations of temperature and pH with five replicates in each combination) and growth was observed in 218 samples. Among the 88 combination treatments, growth of A. acidoterrestris was observed in 40 conditions, no growth in 41 conditions while in seven conditions growth occurred in some (but not all) of the five replicates (Fig. 3.1). The minimum pH and temperature where growth was observed were 2.52 (at 35 and 45 oC) and 25 oC (at pH ≥ 3.32), respectively. These results are in agreement with previous studies reporting that the temperature and pH for growth of Alicyclobacillus spp. ranges from < 20 to 55 - 65 oC and from 2 - 2.5 to 6.0 - 6.5, respectively (Jiang et al., 2008; Smit, Cameron, Venter, & Witthuhn, 2011). Slightly higher pH limits (2.9 - 3.8) of A. acidoterrestris growth in apple and orange drink compared to the present work have been reported by other researchers (Pena, De Massaguer, Zuniga, & Saraiva, 2011; Peña & Rodríguez, 2010) which can be attributed to the different strain(s), media and acidifier used.

127

Chapter 3

Figure 3.1. Growth/no growth interface of Alicyclobacillus acidoterrestris in K broth with respect to temperature and pH predicted by the model (lines) compared to the data used to generate the model (points). Black symbols: growth in all replicates; white symbols: no growth in all replicates; grey symbols: growth in some replicates; lower line: predicted boundary P = 0.1; middle line: predicted boundary P = 0.5; upper line: predicted boundary P = 0.9.

For the development of the growth/no growth interface model, the collected OD measurements, corresponding to different environmental conditions, were converted into binary data (0: no growth, 1: growth). A logistic polynomial regression model was then fitted to these binary data (Equation (3.1)), and the estimated model parameters are summarized in Table 3.1. The parameters with no significant effect (P ≥ 0.05) were removed from the model. The concordance index and the Hosmer - Lemeshow statistic were used as measures of the goodness of fit of the developed model. As demonstrated by the concordance index, the degree of agreement between the predicted probabilities and the observations was 98.8%. The Hosmer - Lemeshow goodness - of - fit statistic was 6.226 (χ2, df 8; P = 0.622). The goodness - of - fit was also evaluated graphically by comparing the model predictions at probabilities of 0.1, 0.5 and 0.9 with the corresponding observed data (Fig. 3.1). The developed model can be used to predict both probabilities of growth and temperature - pH limits at a certain probability level. The predicted probabilities of A. acidoterrestris growth for representative combinations of temperature and pH

128

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks are presented in Table 3.2. The application of the model to foods however, requires validation studies in order to evaluate the effect of factors such as food composition, strain variability, microbial interactions, physiological state of the cells etc. on its performance.

Table 3.1. Estimated values and fitting statistics for the parameters of the logistic polynomial regression model for the combined temperature and pH limits of Alicyclobacillus acidoterrestris growth in K broth. Coefficients Estimate Standard Error P value

a0 (Constant) 25.9564 9.99059 0.009

a1 (Temperature, (T)) -15.8016 3.24411 0.000

a2 (pH) -1.63928 0.549044 0.003

a3 (pHxT) 0.950885 0.173144 0.000 2 a4(T ) -0.0122219 0.0044819 0.006

Table 3.2. Probability of Alicyclobacillus acidoterrestris growth predicted by the growth/no growth boundaries model for different combinations of temperature and pH. Temperature (oC)

20 25 30 35 40 45

pH Predicted Probability of Growth 2.5 0.025 0.061 0.082 0.063 0.027 0.006 3.0 0.112 0.777 0.981 0.998 0.999 1.000 3.5 0.387 0.995 1.000 1.000 1.000 1.000 4.0 0.759 1.000 1.000 1.000 1.000 1.000 4.5 0.940 1.000 1.000 1.000 1.000 1.000

3.1.2 Kinetic model The effect of temperature and pH on A. acidoterrestris growth rate was investigated using the Bioscreen C method. Experiments were carried in (i) K broth, adjusted to pH = 4.5 with citric acid, at isothermal storage temperature of 25, 27, 30, 35, 40, 45, 48, 50, 53, 55 °C, and (ii) in K broth with pH 3.03, 3.32, 3.6, 3.78, 3.99, 4.32, 4.52, 4.8, 5.04, 5.29, 5.53, adjusted with citric acid, at an incubation temperature of 48 °C. The average (± standard deviation) μmax (1/h) increased from 0.098 (± 0.008) at 25 °C to a maximum value of 1.031 (± 0.008) at 48 °C, while at temperatures > 48 °C a gradual decrease of μmax was observed. Regarding the pH effect, the average (± standard deviation) μmax (1/h) increased from 0.179 (± 0.030)

129

Chapter 3 at pH 3.03 to maximum value of 1.036 (± 0.014) at pH 3.99, while at higher pH values

μmax gradually decreased. The above experimental data (μmax) were modelled as a function of temperature and pH using cardinal parameter models (Equations (3.2) and (3.4), respectively), for the estimation of Tmin, Topt, Tmax, pHmin, pHopt, pHmax. The R2 and RMSE values (Table 3.3), as well as the graphical evaluation of the fitting curves (Figs. 3.2 and 3.3), indicated a satisfactory performance of the model in describing the effect of temperature and pH on the μmax of A. acidoterrestris. The estimated values for the cardinal parameters Tmin, Tmax, Topt and the optimum maximum specific growth rate (μopt) of A. acidoterrestris were found to be 18.11, 55.68, 48.60 °C and 0.980 1/h, respectively (Table 3.3). For the effect of pH, the estimated values for pHmin, pHmax, pHopt and the of A. acidoterrestris were found to be 2.93, 5.90, 4.22 and 1.090 1/h, respectively (Table 3.3). Despite the fact that the above cardinal parameters for the effect of temperature and pH are estimated based on different datasets the values were very close. Although, to our knowledge, this is the first model on the effect of temperature and pH on Alicyclobacillus spp growth rate, previous qualitative studies have shown a similar effect. Goto (2000) reported a temperature range for A. acidoterrestris growth between 20 and 55 °C. Ikegami et al. (1996) studied the effect of pH on the growth of A. acidocaldarius and reported a similar trend, with a pH range of growth from 2.5 to 6.5 and an optimum pH between 4.5 and 5.5.

130

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

Table 3.3. Estimated values and fitting statistics for the parameters of the cardinal parameter models describing the effect of temperature and pH on the maximum specific growth rate (μmax) of Alicyclobacillus acidoterrestris in K broth. Parameter Estimated Valuea 95% Confidence Limits RMSEb R2c Lower Upper Temperature model

(1/h) 0.980±0.013 0.954 1.006 0.0010 0.986

Tmax 55.68±0.07 55.54 55.82

Tmin 18.11±0.40 17.31 18.91

Topt 48.60±0.22 48.16 49.03 pH model

(1/h) 1.090±0.014 1.062 1.118 0.0015 0.958

pHmax 5.90±0.04 5.83 5.97

pHmin 2.93±0.01 2.90 2.95

pHopt 4.22±0.03 4.17 4.27 a Values are means ± standard errors b RMSE: Root Mean Square Error c R2: Coefficient of determination

1.2 Observed Fitted 1.0 95th Confidence Interval

0.8 95th Prediction Interval

( 1/h) 0.6

max μ 0.4

0.2

0.0 15 20 25 30 35 40 45 50 55 60

Temperature ( C)

Figure 3.2. Effect of temperature on the maximum specific growth rate (μmax) of

Alicyclobacillus acidoterrestris in K broth of pH 4.5. Points represent the observed μmax values, the solid line corresponds to the fitting on the cardinal parameter model with inflection to the data, and the dotted lines depict the 95% confidence and prediction intervals.

131

Chapter 3

Observed 1.2 Fitted 95th Confidence Interval 1.0 95th Prediction Interval

0.8 ( 1/h)

0.6

max μ 0.4

0.2

0.0 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7

Figure 3.3. Effect of pH on the maximum specific growth rate (μmax) of Alicyclobacillus acidoterrestris in K broth at 48 °C. Points represent the observed μmax values, the solid line corresponds to the fitting on the cardinal pH model to the data, and the dotted lines depict the 95% confidence and prediction intervals.

3.2. Validation of the models in fruit drinks 3.2.1 Validation of growth/no growth interface model The growth/no growth interface model was further validated against observed growth data of A. acidoterrestris in fruit drinks. Validation experiments were performed with eight different commercial fruit drink products with a pH ranging from 2.59 to 3.83 at different storage temperatures from 21 to 48 oC (Table 3.4). The o aw of all tested drinks was very high (> 0.990), while the soluble solids contents ( Brix) were similar ranging from 10.75 to 11.5. A comparison between the probability of growth predicted by the growth/no growth interface model and the observed behavior of A. acidoterrestris is presented in Table 3.4. The validation results showed a very good performance of the model. In all cases where the probability of growth predicted by the model was lower than 0.5, A. acidoterrestris was, indeed, not able to grow in the fruit drinks. Similarly, when the model predicted a probability above 0.9, growth was observed in all cases.

132

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

Table 3.4 Comparison between observed behavior of Alicyclobacillus acidoterrestris and probability of growth predicted by the developed growth/no growth interface model in commercial pasteurized fruit drinks tested in validation studies.

Tested Storage Predicted Observed Products Ingredients pH a oBrix Temperature probability w behavior (oC) of growth Natural juices of apple, orange and carrot from concentrated juices 21 No growth 0.692 Apple- (50%), water, sugar, 25 Growth 0.995 orange- concentrated natural 3.52 0.995 11.25 30 Growth 1.000 carrot juice of lemon, 45 Growth 1.000

flavourings, 48 Growth 1.000 antioxidant: ascorbic acid Natural juices of apple, grape, peach, pineapple, orange, apricot, grapefruit, 9 Fruits & passion fruit and 21 No growth 0.832 10 mango from 3.71 0.994 11.50 25 Growth 0.999 Vitamins concentrated juices and 48 Growth 1.000 purée (99.9%), flavourings, vitamins: C, E, B1, B2, B6, B12, niacin, folic acid, pantothenic acid, biotin Natural orange juice 21 No growth 0.891 from concentrated Orange 3.83 0.992 11.00 25 Growth 1.000 juice, orange cells 48 Growth 1.000 (1.4%) Water, natural juices of apricot, apple and orange from Apricot- concentrated juices and 21 No growth 0.525 apple- apricot purée (40%), 3.35 0.995 11.20 25 Growth 0.983 orange sugar, acidity regulator: 48 Growth 1.000

citric acid, flavourings, antioxidant: ascorbic acid

133

Chapter 3

Water, natural purée 21 No growth 0.515 from peach (30%), 25 Growth 0.981 Peach sugar, acidity regulator: 3.34 0.994 11.25 35 Growth 1.000 citric acid, flavourings, 45 Growth 1.000 antioxidant: ascorbic 48 Growth 1.000 acid Water, natural juice of mandarin and blood orange from concentrated juice Mandarin (20%), sugar, acidity 21 No growth 0.198 and Blood regulator: citric acid, 2.99 0.993 11.00 25 No growth 0.763 Orange colorant: concentrated 48 Growth 1.000 natural elderberry juice, natural aroma of mandarin, antioxidant: ascorbic acid Water, natural juices of white and red grape from concentrated juices, sugar, cranberry and raspberry and Cranberry blueberry purée (7.5%), – natural juice of acerola 21 No growth 0.093 2.78 0.993 10.75 Raspberry from concentrated 25 No growth 0.376 - Blueberry juices, acidity regulator: citric acid, antioxidant: ascorbic acid, flavourings. Total juice and fruit purée content: 20%. Water, natural lemon juice from 21 No growth 0.045 concentrated juice Lemonade 2.59 0.992 11.50 25 No growth 0.117 (13.3%), sugar, lemon 48 No growth 0.027 cells (0.5%), natural flavouring

134

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks

The observed behavior of A. acidoterrestris in fruit drinks and the growth boundaries predicted by the model based on their pH and storage temperature are illustrated in Figure 3.4. As shown in this figure, in six cases growth was not observed although the predicted probability by the model was higher than 0.5 (Fig. 3.4). This slight over - prediction by the model of the growth ability of the organism could be attributed to the potential presence of natural antimicrobial compounds in the fruit drinks, the effect of which has not been taken into account in model development (McNamara, Wiebe & Gomez, 2011; Yokota, Fujii, & Goto, 2008). However, the overall effect of the above factors (or other factors not included in the model) on its performance was not found to be significant. Indeed, Figure 3.4 clearly shows that the 90% probability limits predicted by the model can be successfully used for describing the critical combination of temperature and pH which inhibits the growth of A. acidoterrestris in fruit drinks.

5.5 Predicted (p=0.1)

Predicted (p=0.5) 5.0 Predicted (p=0.9)

4.5 Observed (No growth)

Observed (Growth) 4.0

3.5 pH

3.0

2.5

2.0

1.5 15 20 25 30 35 40 45 50 55 Temperature ( C) Figure 3.4. Comparison between predicted growth boundaries (lines) and observed behavior (points) of Alicyclobacillus acidoterrestris in fruit drinks. Black symbols: growth; white symbols: no growth; lower line: predicted boundary P=0.1; middle line: predicted boundary P=0.5; upper line: predicted boundary P=0.9.

135

Chapter 3

3.2.2 Validation of kinetic model The developed kinetic model was validated against observed growth data of A. acidoterrestris in fruit drinks at both static and dynamic storage temperature conditions. Prediction of growth was based on the combination of the secondary model (Equation (3.6)) with the differential equations of the Baranyi and Roberts primary model (Equations (3.7) and (3.8)), which were numerically integrated with respect to time. The parameter μopt used in the secondary model (Equation (3.6)) was the average of μopt values estimated from the models describing the distinct effects of temperature and pH on μmax (Equation (3.2) and (3.4)). For the purpose of growth prediction, the parameters ymax (maximum population density) and h0 (physiological state parameter) of the primary model (Equations (3.7) and (3.8)) were fixed to 106.2 CFU/ml and 4.0, respectively, based on the observed growth of A. acidoterrestris in K broth at static temperature conditions (Fig. 3.5a). In bacterial growth, ho represents the amount of “work” that a cell has to perform to adapt to its new environment. The parameter ho, also referred to as “work to be done” is estimated as the product of μmax and λ (lag phase) (Baranyi & Roberts, 1994). Several studies have reported a relation between μmax and λ, with their product μmax*λ being constant at different storage temperatures when the pre - inoculation history of the cells culture was the same (Gougouli et al., 2008; K Koutsoumanis, Stamatiou, Skandamis, & Nychas, 2006; Pin, de Fernando, Ordóñez, & Baranyi, 2002). In the experiments conducted in this study, the inoculum was constituted of spores which were produced under a well - defined environment. Based on the above, we assumed that ho, which in this case refers to spore germination and outgrowth, is not affected by the storage temperature. The comparison between observed and predicted growth of A. acidoterrestris in apple – orange - carrot drink at 30 oC and in peach drink at 35 oC is illustrated in Figure 3.5b and 3.5c, respectively. In both cases, the model showed a good performance with a difference between predicted and observed growth being less than 1 log CFU/ml. Pasteurized fruit drink products are commonly stored at the retail level at room temperature. Nonetheless, in the absence of disparate distribution systems for refrigerated and non – refrigerated products, the current practice of the fruit drink industry is to store and distribute these products under refrigeration before deliver to retail. In this context, the data presented in Figures 3.5d and 3.5e simulate the latter scenario of distribution and storage. Specifically, Figure 3.5d presents the validation of the model against observed growth of A. acidoterrestris in “9 fruits & 10 vitamins” drink (pH = 3.71) stored at 5 oC for 2 days followed by storage at 30 oC. The model validation results against observed growth in apple – orange - carrot drink (pH = 3.52) stored at a similar temperature profile

136

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks with the refrigeration storage being extended to 5 days, are shown in Figure 3.5e. A good agreement between predicted and observed growth of the spoilage organism was observed for both products and dynamic temperature conditions. In the later scenario, however, the observed lag time was longer that the predicted, indicating that the prolonged exposure to low temperature may result in a physiological stress and an additional lag time. Nevertheless, the prediction of the model at the exponential phase was very close to the observed growth (Fig. 3.5e). A periodic temperature profile involving 6 h at 25 oC, 12 h at 35 oC and 6 h at 45 oC was tested for storage of apple - orange - carrot drink (Fig. 3.5f). At these temperature conditions, A. acidoterrestris exhibited a total growth of 4 log CFU/ml within 60 hours which was satisfactorily predicted by the growth model. Figures 3.5g and 5h present the growth of the spoilage bacterium in apple - orange - carrot drink and peach drink respectively, stored at room temperature during the summer period in Greece. The illustrated temperature fluctuations reflect the temperature differences between day and night time. Again, the growth predicted by the model, based on the pH of the products and the recorded temperature profile, described very well the observed microbial behavior.

137

Chapter 3

Figure 3.5. Comparison between observed (points) and predicted (solid line) growth of Alicyclobacillus acidoterrestris ATCC 49025 in: (a) K broth (pH=4.5) at 35 °C, (b) apple - orange - carrot drink (pH=3.52) at 30 °C, (c) peach drink (pH=3.34) at 35 °C, (d) 9 fruits & 10 vitamins drink (pH=3.71) stored at 5 °C for 2 days followed by storage at 30 °C, (e)

138

Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks apple - orange - carrot drink (pH=3.52) stored at 5 °C for 5 days followed by storage at 30 °C, (f) apple - orange - carrot drink (pH=3.52) stored at dynamic temperature conditions (6 h at 25 °C, 12 h at 35 °C and 6 h at 45 °C), (g) apple - orange - carrot drink (pH=3.52) stored at room temperature during the summer period in Greece, and (h) peach drink (pH=3.34) stored at stored at room temperature during summer period in Greece. Discontinuous lines indicate medium temperature during storage. For growth prediction the parameters ymax (maximum population density) and h0 (physiological state parameter) of the primary model were fixed to 106.2 CFU/ml and 4.0, respectively.

4. Conclusions In conclusion, the results of the present study demonstrate that the effect of the environment on growth/no growth interface and the growth kinetics of A. acidoterrestris can be quantitatively expressed using mathematical models. Extensive validation studies showed that the models developed in this study can be used to predict the behavior of this spoilage microorganism in fruit drinks. These models could bring benefits for the industry by identifying the conditions that should be applied to processing, distribution and storage in order to minimize the risk of A. acidoterrestris growth. Additional research data on the intra - species differences in the growth kinetics of the A. acidoterrestris strains are certainly expected to improve the model by incorporating strain variability in its prediction (Lianou & Koutsoumanis, 2013; Pouillot & Lubran, 2011). Furthermore, given that in practice, fruit drinks are contaminated with low bacterial spore numbers, studies on the effect of processing and storage conditions on the variability of individual spore’s lag time will increase the precision and credibility of the developed model (Kakagianni, Aguirre, Lianou, & Koutsoumanis, 2017; Stringer, Webb, & Peck, 2011). Finally, further research on the quantification of the effect of the environmental parameters on guaiacol production in relation to A. acidoterrestris growth will enhance the model’s value and applicability, allowing its utilization in spoilage predictions and shelf life assessment of fruit drinks.

139

140

Chapter 4 Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

Myrsini Kakagianni, Juan S. Aguirre, Alexandra Lianou, Konstantinos P. Koutsoumanis

Published in Food Microbiology 67, 76-84 (2017)

Chapter 4

Abstract

The lag times (λ) of Geobacillus stearothermophilus single spores were studied at different storage temperatures ranging from 45 to 59 °C using the Bioscreen C method. A significant variability of λ was observed among individual spores at all temperature tested. The storage temperature affected both the position and the spread of the λ distributions. The minimum mean value of λ (i.e. 10.87 h) was observed at 55 °C, while moving away from this temperature resulted in an increase for both the mean and standard deviation of λ. A Cardinal Model with Inflection (CMI) was fitted to the reverse mean λ, and the estimated values for the cardinal parameters Tmin, Tmax, Topt and the optimum mean λ of G. stearothermophilus were found to be 38.1, 64.2, 53.6 °C and 10.3 h, respectively. To interpret the observations, a probabilistic growth model for G. stearothermophilus individual spores, taking into account λ variability, was developed. The model describes the growth of a population, initially consisting of N0 spores, over time as the sum of cells in each of the N0 imminent subpopulations originating from a single spore. Growth simulations for different initial contamination levels showed that for low N0 the number of cells in the population at any time is highly variable. An increase in N0 to levels exceeding 100 spores results in a significant decrease of the above variability and a shorter λ of the population. Considering that the number of G. stearothermophilus surviving spores in the final product is usually very low, the data provided in this work can be used to evaluate the probability distribution of the time - to - spoilage and enable decision - making based on the “acceptable level of risk”.

142

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

1. Introduction

The presence of Geobacillus stearothermophilus spores in thermally processed foods constitutes one of the most important quality problems for the food industry. The high prevalence and concentration of spores in the raw materials, the adhesive characteristics of spores that enhance their persistence in industrial environments and, most importantly, the extreme heat resistance of the spores are among the major factors explaining the importance of this endospore-former (André, Zuber, & Remize, 2013; Postollec et al., 2012; Yoo, Hardin, & Chen, 2006). Thus, the estimation of the risk of G. stearothermophilus growth to spoilage levels constitutes a major target of the quality managers. Recently, Kakagianni et al. (2016) developed a predictive model for the effect of storage temperature on the growth of G. stearothermophilus, which can be used to predict the time - to - spoilage as the time required by the contaminating spores to germinate and multiply from the initial to a spoilage level. The model was developed and successfully validated for evaporated milk based on data from experiments with high initial contamination levels. In practice, however, the number of spores surviving the heat treatment is usually very low. In this case, the early germinated and outgrown spores within the population begin to multiply up to a microbial concentration sufficient to spoil the food (Baranyi, 1998, 2002; Baranyi, George, & Kutalik, 2009; Pin & Baranyi, 2006) generating an extreme scenario in assessing spoilage. Depending on the use of the model, the latter extreme estimate can be on the “fail - safe” or the “fail - dangerous” side. For example, when the model is used to predict the time - to - spoilage, the above predictions represent a worst - case scenario. In contrast, when the objective is to evaluate the duration of a quality control test in which products are stored at the optimum temperature for growth for a certain time and the percentage of spoiled items is evaluated, the above extreme can lead to “fail - dangerous” estimates. Thus, for an effective risk assessment of spoilage the growth model needs to be combined with quantitative data on the variability of individual spore behavior. Lag time is much more uncertain and difficult to predict compared to growth rate (Baranyi, 2002). However, the credibility of a mathematical model in predicting the conditions that lead to microbiological spoilage depends highly on its ability to describe the effect of the environment on lag time. This implies that a better understanding of the lag phase is of great importance for the effective application of

143

Chapter 4 predictive models in food quality management. For a complete understanding of lag time, the behaviour of single cells or spores has to be taken into account through the development of stochastic models which are able to deal with more “realistic” low contamination events (Koutsoumanis & Lianou, 2013). Available studies on single spore lag time focus on either Bacillus (Aguirre et al., 2015; Aguirre, Ordóñez, & de Fernando, 2012; Pandey et al., 2013; Pandey et al., 2015; van Melis, Den Besten, Groot, & Abee, 2014; Warda, den Besten, Sha, Abee, & Groot, 2015; Zhou, Dong, Setlow, & Li, 2013) or Clostridium (Lenz, Schnabel, & Vogel, 2014; Smelt, Stringer, & Brul, 2013; Stringer, Webb, George, Pin, & Peck, 2005; Stringer, Webb, & Peck, 2009; Wang et al., 2012; Webb, Pin, Peck, & Stringer, 2007) species, while data on G. stearothermophilus are very limited. Zhou, et al. (2013) studied the kinetics of germination of individual spores of G. stearothermophilus measured by Raman spectroscopy and differential interference contrast microscopy. Previous studies, however, demonstrated that germination and lag time are independent (Stringer et al., 2005). The variability of post - germination stages such as outgrowth and doubling is relatively large, and as a result, the germination time cannot be used to predict the overall lag time. The objective of the present work was to assess the impact of different storage temperature conditions on the lag time of individual G. stearothermophilus spores, and to demonstrate the role of individual spore heterogeneity in growth predictions using a stochastic modeling approach.

2. Materials and Methods

2.1. Bacterial strain and inoculum preparation The type strain G. stearothermophilus ATCC 7953 was used for all experiments in the present study. The stock culture of the strain was stored frozen (- 70 °C) onto MicrobankTM porous beads (Pro - Lab Diagnostics, Ontario, Canada). The working culture was stored refrigerated (5 °C) on nutrient agar (NA; Lab M Limited, Lancashire, United Kingdom) slants and was renewed bimonthly. Prior to each experiment, the microorganism was activated by transferring a loopful from the NA slants into 10 ml nutrient broth (NB; Lab M Limited) and incubating at 55 °C for 24 h. The 24 - h G. stearothermophilus cultures were subjected to a heat shock treatment in order to eliminate vegetative cells and enhance the homogeneous activation and germination of dormant endospores of the organism (Antolinos et al., 2012; Yuan et

144

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores al., 2012). Then, the heat - shocked cultures were centrifuged (6000 rpm for 20 min) in a refrigerated centrifuge (4 °C) (model PK120R, ThermoElectron Corporation, Waltham, MA). The harvested cells were washed with 20 ml of quarter - strength Ringer’s solution (Lab M Limited) through centrifugation under the same conditions. The pellet was finally resuspended in 10 ml of tryptone soy broth (TSB; Lab M Limited) in order to obtain an initial concentration of ca. 106 spores/ml.

2.2. Growth kinetic experiments 2.2.1 Maximum specific growth rate

The maximum specific growth rate (μmax) of G. stearothermophilus was evaluated in TSB at 45, 47.5, 50, 55 and 59 °C. The μmax values corresponding to each storage temperature were estimated by means of absorbance detection times of serially decimally diluted cultures using the automated turbidimetric system Bioscreen C (Oy Growth Curves Ab Ltd., Raisio, Finland) as described previously (Kakagianni, Gougouli, & Koutsoumanis, 2016; Lianou & Koutsoumanis, 2011). In this study, the range of initial concentrations obtained in the microtiter plates was approximately 106 - 102 cfu/well. Optical density (OD) measurements in the Bioscreen C were measured at 15-min intervals using the wideband ﬁlter (420 - 580 nm) for a total time period such that a considerable OD change was observed, if possible, for all ﬁve decimally diluted cultures. The microtiter plates were agitated for 15 s at a medium amplitude prior to the OD measurements. The absorbance time to detection (Tdet) was defined as the time required for the OD to reach 0.2 units. Specifically, proper dilutions of the content of each well of a microtiter plate (100 wells per plate) were surface plated onto tryptone soy agar (TSA; Lab M Limited) plates during the exponential growth phase of the cultures at Tdet (ODdet = 0.2 A420-

580). Colonies were enumerated after incubation of the plates at 55 °C for 24 h allowing the calculation of the bacterial concentration at the chosen Tdet (Ndet). The detection times (h) of five serial decimal dilutions of the bacterial culture were plotted against the corresponding natural logarithm of their initial concentrations, and μmax values were determined by linear regression as the negative reciprocal of the slope of the regression line (Dalgaard & Koutsoumanis, 2001). For each temperature, five samples (i.e. quintuple wells of five serially diluted cultures) were analysed (n = 5).

145

Chapter 4

2.2.2 Lag time of individual spores

Assuming an exponential bacterial growth at a constant μmax, after the lag period and until the Tdet, the single - spore lag time (λ) values of the tested microorganism were estimated based on the generated turbidity growth curves and according to the Bioscreen C methodology (Aguirre, Rodríguez, & de Fernando, 2011; Baranyi et al., 2009; Baranyi & Pin, 1999). The culture of the tested strain was decimally and binary diluted in TSB to obtain a target concentration of ca. 1 spore/well. Two microtiter plates (100 wells per plate) were inoculated with 350 μl of the aforementioned diluted organism’s culture in each well, incubated in the Bioscreen C at different temperatures (45, 47.5, 50, 55 and 59 °C), and the OD was recorded at 420 - 580 nm at 15-min intervals. Microtiter plates were shaken at medium amplitude for 15 s before measuring OD. The average initial number of spores per well (m) was estimated by assuming that the probability of having one spore per well is described by the Poisson distribution (Baranyi et al., 2009), as a result of the dilution process, using the following equation: m = -LnP0 (4.1)

where P0 is the probability of having no growth in a well (estimated as the percentage of wells without detected growth). To ensure that most of the positive wells contained on average one spore, when microtiter plates showed a growth/no growth ratio less than 0.2 or higher than 0.7, were discarded and the experiment was repeated.

The individual spore λ values were calculated using the following equation proposed by Baranyi and Pin (1999):

(4.2)

where Tdet is the time (h) required for the OD at 420 - 580 nm to reach 0.2 units, μmax is the maximum specific growth rate (1/h) determined under the experimental conditions as described in Section 2.2.1, LnNdet is the natural logarithm of the cell population (ln cfu/well) at Tdet, LnN0 is the natural logarithm of the average initial number of spores initiating growth (ln cfu/well) in the considered well estimated according to Eq. (4.1). For each tested condition, the experiments were repeated as many times needed to obtain about 200 values of λ.

146

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

2.2.3 Statistical analysis and growth simulation The cumulative distributions of individual spore λ values were fitted to various theoretical distributions using @RISK 6.0 Professional Edition (Palisade Corporation, Newfield, New York, USA). The goodness of fit was evaluated using RMSE values.

Monte Carlo simulation was performed to predict growth for different N0 (1, 10 and 100 spores) taking into account the variability in individual spore λ using the best fitted distribution. The simulation was performed using 10,000 iterations.

3. Results

The differences in the Tdet from OD curves of 199 individual spores growth at 45

°C are illustrated in Fig. 4.1. The observed Tdet, μmax and Ndet values were used to evaluate the variability of single G. stearothermophilus spores λ based on equation

3.2. The estimated μmax values (mean ± standard deviation) were 0.452 ± 0.022, 0.563 ± 0.051, 0.883 ± 0.047, 1.666 ± 0.047 and 1.955 ± 0.091 h-1 for 45, 47.5, 50, 55 and 59 °C, respectively. The bacterial concentration parameter Ndet was found to be unaffected by temperature (P ≥ 0.05) with a mean (± standard deviation) value of 6.99 (± 0.36) log cfu/well.

0,50

0,45

0,40

0,35 580nm)

- 0,30

0,25 OD OD (420 0,20

0,15

0,10 0 10 20 30 40 50 60 70 Time (h) Figure 4.1. OD growth curves during growth of 199 single spores of Geobacillus stearothermophilus in tryptone soy broth at 45 °C.

147

Chapter 4

Figure 4.2 and Table 4.1 present the cumulative probability distributions and the statistics of the individual spore λ at the different tested temperatures. As shown, λ varied significantly among individual spores at all temperature tested. The minimum mean value of λ was 10.87 h and it was observed at 55 °C. Moving away from the later temperature resulted in an increase for both the mean and the standard deviation of λ. The Gamma distribution was further fitted to the λ data, following the theory of Baranyi and Pin (2001). The estimated parameters and the RMSE values of the fitting are presented in Table 4.2. A comparison between the observed and fitted quantiles (Fig. 4.3) clearly revealed that the Gamma distribution is suitable for describing the individual spore λ for all the tested temperatures.

1 0.9 0.8 0.7 59°C 0.6 55°C 0.5 50°C 0.4 47.5°C

0.3 Cumulative Probability 0.2 45°C 0.1 0 0 20 40 60 80 100 120 Lag time (h)

Figure 4.2. Cumulative distributions of Geobacillus stearothermophilus single spore lag times (λ) at different growth temperature conditions. The number of data wa 224, 199, 227, 199 and 193 for 45, 47.5, 50, 55 and 59 °C, respectively.

148

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

Table 4.1. Statistics of the individual spore lag time (λ) of Geobacillus stearothermophilus. Storage λα (h) temperature Mea St. 5th 95th Nb Median Skewness Kurtosis (°C) n Deviation Percentile Percentile 45 224 37.9 18.6 33.3 16.4 72.6 1.39 2.62 47.5 199 21.4 17.7 15.3 1.26 56.9 1.03 0.102 50 227 11.7 11.8 6.46 1.50 37.8 1.66 2.06 55 199 10.9 9.2 7.16 0.705 29.9 1.42 1.85 59 193 17.9 15.0 13.6 0.760 47.7 1.00 0.305 a Individual spore lag time. b Number of spores tested.

Table 4.2. Parameter estimation of the Gamma distribution fitted to the individual spore lag times of Geobacillus stearothermophilus. The probability density function of the Gamma distribution in the shape - rate parametrization is:

for x > 0 and α, β > 0.

Temperature (oC) a β RMSEa 45 4.69 7.84 0.0186 47.5 1.30 16.54 0.0316 50 1.28 7.81 0.0576 55 1.40 4.72 0.0603 59 1.10 17.06 0.0200 a Root Mean Square Error

149

Chapter 4

Fitted quantiles

Observed quantiles

Figure 4.3. Comparison between observed and fitted quantiles of the Gamma distribution for Geobacillus stearothermophilus single spore lag times (λ) at 45 °C (a), 47.5 °C (b), 50 °C (c), 55 °C (d) and 59 °C (e).

150

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

In order to quantify the temperature dependence of individual spore λ, the following Cardinal Model with Inflection (CMI) (Rosso, Lobry, & Flandrois, 1993) was fitted to the reverse mean λ:

(4.3)

(4.4)

where Tmin and Tmax are the theoretical minimum and maximum temperature for growth (λ becomes infinite), respectively, Topt is the theoretical optimum temperature for growth (λ reaches a minimum value), and λopt is the value of lag at T

= Topt.

In order to stabilise the variance, a square root transformation was used. Both the R2 (Table 4.3) and the graphical evaluation of the fitting (Fig. 4.4) indicated a satisfactory performance of the CMI in describing the effect of temperature on the mean λ of G. stearothermophilus individual spores. The estimated values for the cardinal parameters Tmin, Tmax and Topt were 38.2, 64.2 and 53.6 °C, respectively, while the mean λopt of G. stearothermophilus was estimated 10.3 h (Table 4.3).

Table 4.3. Estimated values and statistics for the parameters of the Cardinal Model with Inflection (Equations 4.3 and 4.4) describing the effect of temperature on the reciprocal of the mean lag time (λ) of individual spores of Geobacillus stearothermophilus. Parameter Estimated Value Standard Error R2a

λopt (h) 10.3 0.5 0.977

Tmax (°C) 64.2 2.3

Tmin (°C) 38.2 5.1

Topt (°C) 53.6 0.8 a R2: coefficient of determination

151

Chapter 4

0.40

0.35

0.30

)

h 0.25 1/

(

0.20 1/Lag 0.15

0.10

0.05

0.00 40 45 50 55 60 65 Temperature (°C)

Figure 4.4. Effect of temperature on the reciprocal of the mean lag times of Geobacillus stearothermophilus single spores fitted to the Cardinal Model with Inflection (solid line) (Equations 4.3 and 4.4). Points ( ) represent observed values.

To interpret the observations, a stochastic model for G. stearothermophilus spores’ growth was developed by introducing the distribution of individual spores λ values into a simple exponential growth with lag model (Koutsoumanis & Lianou, 2013) as follows:

N 1 for t   N  0 i t 1  max(ti ) (4.5)  e for t  i where Nt is the total number of cells or spores in a population at time t, N0 is the initial number of spores at t = 0, μmax is the maximum specific growth rate and λi is the lag time of individual spores following the gamma distribution.

The model describes the growth of a bacterial population, initially consisting of

N0 spores, over time as the sum of cells in each of the N0 imminent subpopulations originating from a single spore. The above approach allows for taking into account the heterogeneity in the growth dynamics of single spores by introducing λ variability in the model using Monte Carlo simulation.

152

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

Figure 4.5 presents the output of the model (Εq. (4.5)) for spore levels N0 equal to 1 (Fig. 4.5a), 10 (Fig. 4.5b) and 100 spores (Fig. 4.5c) at 45 °C based on a Monte Carlo simulation with 10,000 iterations and with a Uniform distribution for t [t ~

Uniform (0, 80)]. As shown, the output of the model for N0 = 1 spore is a stochastic growth curve in which the number of cells in the population at any time is a distribution. An increase in N0 to 100 spores resulted in a significant decrease of the above variability as well as in a shorter lag phase of the population. The model can be used to evaluate the probability distribution of the time - to - spoilage, with the latter being defined as the time required by the initial contaminating spores to germinate and multiply to a spoilage level (Kakagianni et al., 2016).

153

Chapter 4

Figure 4.5. Simulations of Geobacillus stearothermophilus growth at 45 °C for various initial contamination levels: 1 spore (a), 10 spores (b) and 100 spores (c). Growth is predicted by the stochastic model (Equation (4.5)) using Monte Carlo simulation with 10,000 iterations and a Uniform distribution for time t [t ~ Uniform (0, 80)].

154

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores

4. Discussion

The Bioscreen C method, applied in the present work to study the λ of single G. stearothermophilus spores, has been extensively used for monitoring single cell and spore kinetics (Aguirre et al., 2015; Blana, Lianou, & Nychas, 2015; Dupont & Augustin, 2009; George, Metris, & Stringer, 2008; Métris, George, Mackey, & Baranyi, 2008; Miled et al., 2011). One of the main assupmtions of the method is that growth is initiated from a single spore in a well. The number of spores in each well can be estimated based on the ratio between positive wells (wells showing growth) and negative wells (wells showing no growth) according to the Poisson distribution. In all experiments carried out in this study, the percentage of positive wells was always between 20 and 70% providing an average number of spores in a well between 0.22 and 1.2. According to the Poisson distribution, the above values correspond to a percentage of wells with one spore ranging from 51.4% to 98.1%. Previous studies have shown that the above range allows for the estimation of the distributions of single cell parameters (Métris, George, Peck, & Baranyi, 2003; Stringer et al., 2005). Indeed, Stringer et al. (2005) studied the lag times of Clostridium botulinum individual spores utilizing both the Bioscreen C and a microscopic method, and reported that the two methodologies produced distributions with very similar estimates of both lag mean value and variability. The findings of this study highlight a considerable heterogeneity in the λ of G. stearothermophilus single spores. In the case of spores, the lag time is mainly determined by the germination and outgrowth processes (Smelt, Otten, & Bos, 2002; Stringer et al., 2005). Thus, the observed λ variability can be considered as a reflection of the differences in the rates of the above processes among spores. The stochastic fluctuations in the number of the nutrient germinant receptors (nGRs) per spore due to epigenetic variations among individual spores, appears to be one of the main reasons of the significant heterogeneity in spore germination capacity at the population level (Chen, Huang, & Li, 2006; Elowitz, Levine, Siggia, & Swain, 2002; Ghosh & Setlow, 2009; Kong, Zhang, Setlow, & Li, 2010; Peng, Chen, Setlow, & Li, 2009; Stringer et al., 2005; Webb et al., 2007; Zhang et al., 2010). Moreover, the stochastic inactivation of nGRs during heat treatment can contribute to the differences in the germination rates among individual spores (Smelt, Bos, Kort, & Brul, 2008). Finally, the differences in the levels of specific lytic enzymes such as CwlJ and SleB, which initiate the spore’s peptidoglycan cortex degradation during

155

Chapter 4 germination can be an additional variability source (Ghosh & Setlow, 2009; Moir, 2006; Peng et al., 2009; Setlow, 2003). The storage temperature affected both the position and the spread of the λ distributions of G. stearothermophilus spores. In agreement to previous single cell studies (Guillier, Pardon, & Augustin, 2005, 2006; Métris et al., 2003), the mean and the standard deviation of the λ values estimated in this study followed a similar trend in relation to temperature, with less optimum temperatures leading to longer and more scattered λ. Moreover, at all tested temperatures, the λ distributions of G. stearothermophilus spores showed a positive skewness (Table 4.1) indicating an asymmetry with a tail extending towards longer times. This confirmed available microscopic and/or turbidimetric findings on other bacilli (Leuschner & Lillford, 1999; Pandey et al., 2013; Smelt et al., 2008) and clostridia (Billon, McKirgan, McClure, & Adair, 1997; Stringer, Webb, & Peck, 2011; Stringer et al., 2009), but also on non - sporeforming bacteria such as Listeria spp. and Escherichia coli O157:H7 (Laurent Guillier & Augustin, 2006; Li, Odumeru, Griffiths, & McKellar, 2006; McKellar & Lu, 2005; McKellar & Hawke, 2006; Métris et al., 2003). The above skewed shape was sucessfully described with a Gamma distribution which has been previously used for individual cell λ values (Métris et al., 2003). Information about the shape of λ distributions are of great importance in predicting microbial growth at low contamination levels due to the stochastic nature of microbial lag phase (Baranyi, 1998, 2002). The temperature dependence of the mean G. stearothermophilus λ was considerably different from that reported for the growth rate of the same strain by Kakagianni et al. (2016). In the latter study, the authors reported an optimum temperature (mean value ± standard deviation) for growth Topt = 61.82 ± 0.20 °C, which is about 9°C higher than the corresponding parameter for the mean lag time estimated in the present study (Table 4.3). This difference was confirmed by the μmax values estimated in this study which showed a continuous increase as temperature increased from 47.5 to 59.0 oC, in contrast to what was the case for the mean lag time which showed a minimum value at 55.0 oC. These results indicate that the physiological state theorem (Baranyi, 1998) applied for vegetative cells, considering the lag as inversely proportional to the maximum specific growth rate, is not necessarily valid in the case of spores. This can be attributed to the different processes during the lag time of spores and vegetative cells. As described previously, in the case of spores the lag time is mainly determined by the germination and outgrowth process, the rate of which can be considerable different than the growth

156

Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores rate of vegetative cells. Indeed, Stringer et al. (2005) reported that the time intervals for germination, emergence, outgrowth and doubling of C. botulinum spores were all independent of each other. Consequently, although in the case of vegetative cells with known physiological state, growth at different temperatures can be predicted based on a model for the growth rate (Koutsoumanis, Pavlis, Nychas, & Xanthiakos, 2010), predicting growth of spores may require models for both lag time and growth rate since the effect of temperature on these two growth kinetic parameters can be substantially different. G. stearothermophilus is an important spoiler in various heat - treated foods. Considering that the number of surviving spores in the final product is usually very low, knowledge of their lag time variability is important for predicting growth and spoilage in a probabilistic way (Fig. 4.5). Probabilistic modeling approaches that take into account the variability of the factors affecting microbial behavior can provide more realistic estimation of food quality. Most importantly, probabilistic models enable decision - making based on the “acceptable level of risk” and can provide structured information allowing decision - makers to compare various actions and identify those that can lead to effective and economic reduction of quality risks.

157

Chapter 4

158

Chapter 5 Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

Myrsini Kakagianni, Konstantinos P. Koutsoumanis

Published in Food Research International 111, 104-110 (2018)

Chapter 5

Abstract

A predictive model for the effect of storage temperature on the growth of Geobacillus stearothermophilus was applied in order to assess the risk of evaporated milk spoilage in the markets of the Mediterranean region. The growth of G. stearothermophilus in evaporated milk was evaluated during a shelf life of one year based on historical temperature profiles (hourly) covering 23 Mediterranean capitals for five years over the period 2012 - 2016 obtained from the Weather Underground database (http://www.wunderground.com/). In total, 115 scenarios were tested simulating the distribution and storage conditions of evaporated milk in the Mediterranean region. The highest growth of G. stearothermophilus was predicted over the period 2012 – 2016 for Marrakech, Damascus and Cairo with mean values of 7.2, 7.4 and 5.5 log CFU/ml, respectively, followed by Tunis, Podgorica and Tripoli with mean growth of 2.8, 2.4 and 2.3 log CFU/ml, respectively. For the rest 17 capitals the mean growth of the spoiler was <1.5 log CFU/ml. The capitals Podgorica, Cairo, Tunis and Ankara showed the highest variability in the growth during the 5 years examined with standard deviation values for growth of 2.01, 1.79, 1.77 and 1.25 log CFU/ml, respectively. The predicted extent and the variability of growth during the shelf life were used to assess the risk of spoilage which was visualised in a geographical risk map. The growth model of G. stearothermophilus was also used to evaluate adjustments of the evaporated milk expiration date which can reduce the risk of spoilage. The quantitative data provided in the present study can assist the food industry to effectively evaluate the microbiological stability of these products throughout distribution and storage at a reduced cost (by reducing sampling quality control) and assess whether and under which conditions (e.g. expiration date) will be able to export a product to a country without spoilage problems. This decision support may lead to a significant benefit for both the competitiveness of the food industry and the consumer.

160

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth 1. Introduction

Shelf - stable foods have undergone preservation processes that have produced foods that are considered to be ‘commercially sterile’, i.e., they will not spoil or cause disease under normal conditions of handling and distribution (Zottola, 2003). Shelf - stable foods include foods of a type that, because of their composition (low moisture, high acidity, high salt or sugar content), do not require any special storage conditions as they generally do not provide conditions suitable for the growth of microorganisms, or foods which have been processed so that they can be safely stored at room temperature for a long time (NZFSA, 2016). Regarding the second case, the microbiological stability of these, non – perishable, food products is based on the intense thermal processing that kills the vegetative cells of microorganisms. Evaporated milk is a product of high consumption belonging to the latter category. Although the heat processing of evaporated milk eliminates all vegetative bacterial cells, the presence of bacterial spores in the final product is possible (Chen, Coolbear, & Daniel, 2004; Cosentino, Mulargia, Pisano, Tuveri, & Palmas, 1997). Indeed, the spore - forming bacterium Geobacillus stearothermophilus constitutes one of the most important quality problems for evaporated milk products. The high prevalence and concentration of spores in the raw material, the adhesive characteristics of spores that enhance their persistence in industrial environments and, most importantly, the extreme heat resistance of the spores are among the major factors explaining the importance of this bacterium (André, Zuber, & Remize, 2013; Postollec et al., 2012; Yoo, Hardin, & Chen, 2006). At favorable storage conditions, G. stearothermophilus spores that survived the heat process may subsequently germinate and grow out leading to sensory rejection. The resulting spoilage is characterised by the coagulation of milk caused by acid production due to the metabolic activity of the cells (Boor & Murphy, 2002; Hill & Smythe, 2012; Kakagianni, Gougouli, & Koutsoumanis, 2016; Kalogridou-Vassiliadou, 1992; Yoo et al., 2006). The microbiological stability of evaporated milk is mainly based on the thermophilic nature of G. stearothermophilus spores that may be present in the final product which does not allow their germination and growth at temperatures below 33 - 40 °C (Burgess, Lindsay, & Flint, 2010; Hill & Smythe, 2012; Kakagianni et al., 2016; Llaudes, Zhao, Duffy, & Schaffner, 2001; Ng & Schaffner, 1997; Oomes et al., 2007). Unlike other parameters affecting food quality (e.g. pH, water activity, redox

161

Chapter 5 potential) however, the temperature along the supply chain of non - refrigerated products, including transportation and retail storage, is out of the direct control of the manufacturer and often deviates from the specifications. Moreover, temperature control is completely lacking during transportation from the retail stores to consumers’ shelf and during domestic storage until the time of preparation and consumption (Koutsoumanis & Gougouli, 2015). Temperature abuses during any stage of the supply chain may result in an unexpected loss of quality. Thus, recording the temperature conditions of a supply chain and evaluating their effect on the microbiological stability of the products are of paramount importance in a food quality management system especially for countries with hot climate (Koutsoumanis, Taoukis, & Nychas, 2005; Nychas, Panagou, & Mohareb, 2016; Nychas, Skandamis, Tassou, & Koutsoumanis, 2008). The objective of the present study was to apply predictive microbiology tools for mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of storage temperature on G. stearothermophilus growth. Predictive microbiology models can assist the food industry to effectively evaluate the microbiological stability of these products throughout distribution and storage at a reduced cost (by reducing quality control sampling) and assess whether and under which conditions (e.g. expiration date) will be able to export a product to a country without spoilage problems. This decision support may lead to a significant benefit for both the competitiveness of the food industry and the consumer.

2. Materials and Methods 2.1. Prediction of Geobacillus stearothermophilus growth in evaporated milk The predictive model of Kakagianni et al. (2016) describing the effect of temperature on the growth of Geobacillus stearothermophilus ATCC 7953 was applied to assess the risk of spoilage of evaporated milk during distribution and storage in the market of 23 Mediterranean capitals. The prediction of growth was based on the combination of the secondary Cardinal Model with Inflection (Rosso, Lobry, & Flandrois, 1993) with the differential equations of the Baranyi and Roberts primary model (Baranyi & Roberts, 1994). The effect of temperature on the maximum specific growth rate (μmax) was predicted by the secondary model as follows:

(5.1)

162

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

(5.2)

where Tmin = 33.76 °C, Topt = 61.82 °C and Tmax = 68.14 °C are the theoretical minimum, optimum and maximum temperature for growth, respectively, and =

2.068/h is the optimum value for the maximum specific growth rate (when T=Topt).

The prediction of growth under dynamic storage temperature conditions was based on the time - temperature profile of evaporated milk storage, T(t), in conjunction with the secondary model (Eq. 5.1) for the estimation of the

“momentary” maximum specific growth rate (max) and the differential equations of Baranyi and Roberts model (Eq. 5.3 and 5.4), which were numerically integrated with respect to time:

(5.3)

(5.4)

where μmax (T) is the maximum specific growth rate and xmax is the maximum population density.

The parameter q denotes the concentration of a substance critical to growth and is related to the physiological parameter a0 as follows:

The model was based on the assumption that after a temperature shift, the growth rate is adopted instantaneously to the new temperature environment. Hourly temperature data from Mediterranean capitals for five years over the period 2012 - 2016 were used as model data input serving to predict growth and impeding

163

Chapter 5 spoilage of the evaporated milk. For the growth prediction the maximum population density was fixed at 107.4 CFU/ml based on the value reported by Kakagianni et al.

(2016). The a0 was set to 1 representing a worst case scenario with the spoilage organism growing with no lag phase.

2.2. Collection of temperature data Historical temperature data covering 23 Mediterranean capitals for five years over the period 2012 - 2016 was obtained from the Weather Underground database (http://www.wunderground.com/). Data was collected from all recording airport weather stations. For the capitals of Slovenia, Spain, Cyprus and Israel, no hourly temperature data was available in the database and therefore temperature data was collected from the next highly populated city. The temperature data was collected at 1, 3 or 6 h intervals, depending on the station. Τhe extracted information was in the form of time – temperature data in Excel worksheets.

2.3 Mapping the risk of evaporated milk spoilage in Mediterranean capitals The temperature data collected from 23 Mediterranean capitals was introduced into the developed model, according to the Section 2.1, to predict the growth of G. stearothermophilus under storage conditions corresponding to five years over the period 2012 – 2016 for each capital. The total growth was predicted for a storage period of one year representing the expiration date currently applied to evaporated milk by a Greek dairy industry. Box - plots were used to present the growth data, providing the principal measures of central tendency and dispersion. In the Box - plot representation (Fig. 5.3), the bottom and top of the box are the 25th and 75th percentiles (Q1 and Q3, respectively), the band is the median and the dots correspond to the minimum and maximum value. The total growth was further used to evaluate the risk that a product exported to a certain Mediterranean capital will be spoiled before it expires. For this, a maximum initial contamination level of 101 spores/ml observed by a Greek dairy industry and a spoilage level of 107.4 CFU/ml reported by Kakagianni et al. (2016) for the products of the same industry were considered. Based on the above, the time - to - spoilage can be predicted as the time required for G. stearothermophilus population to increase by 6.4 log CFU/ml. Taking into account the variability of the total growth (during 1 - year shelf life) over the five years tested, the risk of spoilage was categorized in the levels of high, moderate, low and very low as described in the Section 3 (Results and Discussion section).

164

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth 3. Results and Discussion

The main objective of the present study was to provide information to the dairy industry on the risk of spoilage of evaporated milk exported to Mediterranean countries. In a previous study (Kakagianni et al., 2016) we have shown that spoilage of this product is due to acid coagulation observed when the pH of evaporated milk decreases to a level around value of 5.2 due to the growth and metabolic activity of G. stearothermophilus. Indeed, G. stearothermophilus growth in evaporated milk results in acid production enhancing the formation of protein aggregates which is mainly related to unfolding and gelation of β - lactoglobulin (Hill & Smythe, 2012; Yoo et al., 2006). Considering the high prevalence of G. stearothermophilus in raw milk and the fact that its spores are extremely heat resistant (André et al., 2013; Membré and van Zuijlen, 2011; Postollec et al., 2012; Simmonds, Mossel, Intaraphan & Deeth, 2003; Yoo et al., 2006), the spoilage of evaporated milk depends almost exclusively on the time - temperature conditions at which the product is exposed to during distribution and storage. Research data support that germination, outgrowth and subsequent growth of G. stearothermophilus spores do not occur below 33 - 40 °C (Burgess et al., 2010; Hill & Smythe, 2012; Kakagianni et al., 2016; Llaudes et al., 2001; Ng & Schaffner, 1997; Oomes et al., 2007). However, the conditions prevailing during the evaporated milk supply chain in some Mediterranean countries may exceed the above minimum temperature range for growth. The important question for the dairy industry is whether the climate of a Mediterranean capital allows sufficient growth of G. stearothermophilus resulting in a high (unacceptable) risk of spoilage, or in other words, the problem is based on the identification of the capitals which allow exporting of evaporated milk with a low (acceptable) risk of spoilage. Predictive microbiology can be used as an effective tool in answering the above questions. A predictive model for G. stearothermophilus growth (Kakagianni et al., 2016) was applied to assess the risk of spoilage of evaporated milk, with 1 - year shelf - life produced by a Greek dairy industry, when exported to 23 Mediterranean capitals. Since evaporated milk is a non - refrigerated product, predictions of growth during a one year of storage were based on historical temperature data obtained from a meteorological database (http://www.wunderground.com/). In order to take into account the year - to - year variability of the temperature conditions, data for five years, over the period 2012 – 2016, was collected and used. Fig. 5.1 shows the annual temperature profiles of a representative capital (Tunis) for the five-year

165

Chapter 5 period. As expected, temperature data is characterized by a high daily and seasonable variance. Differences in the annual temperature profiles among the years are also observed. For the five years examined, the minimum, maximum, mean and standard deviation of the annual temperature in the latter country ranges from 2.0 to 4.0, 40.0 to 42.0, 19.3 to 20.8 and 6.3 to 7.6oC, respectively.

166

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

50 (a) 45 40

C) 35 ° 30 25

20 Temperature ( Temperature 15 10

0 50 (b) 45 40

C) 35

° 30 25

20 Temperature ( Temperature 15 10 5 0 50 (c) 45 40

C) 35 ° 30 25

20 Temperature ( Temperature 15

10 5 0 50 (d) 45 40

C) 35 ° 30 25

20 Temperature ( Temperature 15 10 5 0 50 (e) 45 40

C) 35

° 30 25

20 Temperature ( Temperature 15 10 5 0 0 30 60 90 120 150 180 210 240 270 300 330 360 Time (days) Figure 5.1. Representative examples of historical hourly temperature data for Tunis for the years 2012 (a), 2013 (b), 2014 (c), 2015 (d) and 2016 (e) obtained from the Weather Underground database. Time = 0 corresponds to January 1st.

167

Chapter 5

The predictive model for G. stearothermophilus can be used to translate the temperature data to total growth of the spoiler during a storage period of one year (shelf - life). It needs to be noted that the efficiency of the model used in predicting G. stearothermophilus growth at dynamic temperature conditions has been previously evaluated with extensive validation experiments. Indeed, Kakagianni et al. (2016) compared the predictions of the model with observed growth under several dynamic temperature profiles which included temperatures inside and outside the growth region of the microorganism. For all the temperature scenarios tested in the latter study, the model adequately predicted the growth of G. stearothermophilus in evaporated milk, suggesting a good performance of the model. In Fig. 5.2, examples of the predicted growth of G. stearothermophilus in evaporated milk are illustrated for three capitals of the Mediterranean region (Athens, Tunis and Damascus) for the years 2012, 2014 and 2015, respectively. As shown, depending mainly on the climate of each city but also on the year, the growth of the spoiler may range from negligible (Athens, 2012) to significant exceeding the spoilage threshold of 107.4 CFU/ml (Damascus, 2015).

168

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

8.0 (a) 50 45 7.0 40 6.0 35 30 5.0

25 C) ° 4.0 20 (cfu/ml) 3.0 15 10 10 Log 2.0

5 Temperature ( Temperature 1.0 0 -5 0.0 -10 -1.0 -15 0 45 90 135 180 225 270 315 360 Time (days)

8.0 (b) 50 45 7.0 40 6.0 35 30 5.0

25 C) °

4.0 20 (cfu/ml)

3.0 15 10 10 Log 2.0 5 ( Temperature 1.0 0 -5 0.0 -10 -1.0 -15 0 45 90 135 180 225 270 315 360 Time (days)

8.0 50 (c) 45 7.0 40 6.0 35 30 5.0

25 C) ° 4.0 20

(cfu/ml) 15

10 3.0 10 Log 2.0 5

( Temperature 1.0 0 -5 0.0 -10 -1.0 -15 0 45 90 135 180 225 270 315 360 Time (days) Figure 5.2. Representative examples of Geobacillus stearothermophilus growth prediction (red lines) in evaporated milk with a shelf - life of one year in the supply chain of 3 Mediterranean capitals based on the historical hourly temperature data (blue lines) obtained from the Weather Underground database. a: Athens, (data from 2012); b: Tunis, (data from 2014); c: Damascus, (data from 2015)

169

Chapter 5

The box - plots displaying the 25th, 50th and 75th percentiles and the extreme values of growth for the five years are plotted in Fig. 5.3 for the 23 Mediterranean capitals examined. The box - plots illustrate both the extend of G. stearothermophilus growth in evaporated milk during one year of storage but also the variability in the growth among the five years examined. The highest growth of G. stearothermophilus was predicted over the period 2012 - 2016 for Marrakech, Damascus and Cairo with mean values of 7.4, 7.2 and 5.5 log CFU/ml, respectively, followed by Tunis, Podgorica and Tripoli with mean growth of 2.8, 2.4 and 2.3 log CFU/ml, respectively. For the rest 17 capitals the mean growth of the spoiler was <1.5 log CFU/ml. The capitals with the highest variability in the growth among the 5 years examined were Podgorica, Cairo, Tunis and Ankara with standard deviation values for growth of 2.01, 1.79, 1.77 and 1.25 log CFU/ml, respectively.

Ankara Tunis Damascus Barcelona Maribor Belgrade Lisbon Marrakech Podgorica Valletta Tripoli Beirut Amman Rome Tel Aviv Athens Paris Cairo Larnaca Zagreb Sarajevo Algiers Tirana

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

Log10 (CFU/ml)

Figure 5.3. Box - plot representations of the total predicted growth of Geobacillus stearothermophilus in evaporated milk with a shelf life of one year for the supply chain of 23 Mediterranean capitals for five years (2012 - 2016). The bottom and top of the box are the 25th and 75th percentiles (Q1 and Q3, respectively), the band is the median and the dots correspond to the minimum and maximum value.

170

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

The predicted growth of G. stearothermophilus was further used as the basis for ranking the 23 Mediterranean capitals in relation to the risk of spoilage of evaporated milk products distributed and stored within their region. Based on maximum initial contamination level of 101 spores/ml observed by a Greek dairy industry and a spoilage level of 107.4 CFU/ml reported by Kakagianni et al. (2016). The time - to - spoilage can be predicted as the time required for G. stearothermophilus population to increase by 6.4 log CFU/ml. However, apart from the extent of growth, it is important to take into account the variability of growth among the years in assessing the risk. Given the considerations of a Greek dairy industry about the acceptable (low) level of risk, the following risk categorization was set based on both the mean growth (Mg) during one year of storage and the standard deviation of growth (SDg) among the 5 years examined:

 High Risk: Mg > 4 logs CFU/ml and Mg + 1*SDg > 6 logs CFU/ml  Moderate Risk: Mg < 4 logs CFU/ml and Mg + 2*SDg > 6 logs CFU/ml  Low Risk: Mg < 3 logs CFU/ml and Mg + 3*SDg > 4 logs CFU/ml  Very Low Risk: Mg < 2 logs CFU/ml and Mg + 3*SDg < 4 logs CFU/ml

Fig. 5.4 illustrates the geographical risk assessment for evaporated milk spoilage in the Mediterranean region. As shown Marrakech, Damascus and Cairo present a high risk with Mg + 1*SDg of G. stearothermophilus growth for the five years exceeding 7 log CFU/ml. Tunis and Podgorica present a moderate risk Mg + 2*SDg exceeding 6 logs CFU/ml. Tripoli, Ankara and Amman present a low risk (Mg + 3*SDg > 4.5 log CFU/ml) while all the rest capitals present a very low risk of spoilage (Mg + 3*SDg < 3.5 logs CFU/ml). It needs to be noted the above categorization of risk is also a risk management decision and should be based on the relation between the information provided by the model and the quality requirements of the food industry. In addition, the risk depends on the initial contamination level of G. stearothermophilus. In this study, a maximum initial contamination level of 101 spores/ml observed by a Greek dairy industry was considered but any changes to this level should be taken into account in risk evaluation.

171

Chapter 5

Figure 5.4. Geographical risk assessment for evaporated milk spoilage in the Mediterranean region. Risk is assessed based on the predicted growth of G. stearothermophilus in evaporated milk with a shelf - life of one year for the supply chain of 23 Mediterranean capitals for five years (2012 - 2016). Red: High risk, Orange: Moderate Risk, Yellow: Low Risk, Green: Very Low Risk

The growth model of G. stearothermophilus was also used to evaluate adjustments of the evaporated milk expiration date which can reduce the risk of spoilage. Fig. 5.5 shows the growth of G. stearothermophilus for scenarios in which products with a shelf - life of six months are exported to Damascus for the periods October - March, January - June and March - August based on the temperature data of the years 2014 - 2015. It can be readily seen that for the first two scenarios, which do not include the summer period, the risk of spoilage is significantly reduced since the growth of G. stearothermophilus during distribution and storage of evaporated milk is negligible. However, as shown in Fig. 5.5c, even with a reduced shelf - life, products which are distributed and stored during the summer period present a high risk of spoilage. Based on the above prediction, the dairy industry can manage the shelf - life of the product in combination of the season of exporting in order to reduce the risk of spoilage.

172

Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth

8.0 50 (a) 45 7.0 40 6.0 35 . C) 30 ° 5.0 25

(cfu/ml) 4.0 20 10

3.0 15 Log

10 ( Temperature 2.0 5 1.0 0 -5 0.0 -10 -1.0 -15 0 30 60 90 120 150 180 Time (days)

8.0 50 (b) 45 7.0 40

6.0 35 C)

30 ° 5.0 25

(cfu/ml) 4.0 20 10

3.0 15 Log

10 ( Temperature 2.0 5 1.0 0 -5 0.0 -10 -1.0 -15 0 30 60 90 120 150 180 Time (days)

8.0 50 (c) 45 7.0 40 6.0 35

30 ° 5.0 25

(cfu/ml)

4.0 20 10

3.0 15 Log

10 ( Temperature 2.0 5 1.0 0 -5 0.0 -10 -1.0 -15 0 30 60 90 120 150 180 Time (days) Figure 5.5. Predicted growth of Geobacillus stearothermophilus (red lines) for scenarios in which evaporated milk with a shelf - life of six months is exported to Damascus for the periods October - March (a), January - June (b) and March - August (c) based on the hourly temperature data (blue lines) of the years 2014 - 2015.

173

Chapter 5

4. Conclusions

In conclusion, the present study focuses on the prediction of G. stearothermophilus growth in evaporated milk as a basis for assessing the risk of spoilage for products exported to Mediterranean capitals. The quantitative data provided is useful for the identification of the capitals which allow exporting of evaporated milk with a low (acceptable) risk of spoilage and support quality management in the dairy industry. The results of the study showed that based on the temperature data of the last years, the storage and distribution of evaporated milk in most Mediterranean capitals present a low risk of spoilage.

174

Chapter 6 General Discussion and potential applications of the developed models to the food industry

Chapter 6

Most of the thermophilic endospore - formers can spoil a wide range of thermally processed, non - refrigerated foods, including acidic products (Sperber & Doyle, 2009), leading to product downgrades and losses in revenue for food manufacturers. Failures in product quality are frequently caused by thermostable spores of members of the Bacillaceae family, which show a wide spectrum of resistance to cleaning and preservation treatments. Hence, the presence of thermophilic bacteria is especially troublesome regarding the commercial viability of thermally processed, non - refrigerated foods, which are stored for extended periods of time both at temperatures frequently exceeding 30 - 35 °C in local and international markets with hot climates. The contamination of the final product with viable spores of thermophiles cannot be easily prevented. In addition, contamination usually involves very low number of spores and cannot be detected until the food is exposed to high ambient temperatures, allowing germination and growth (Hill & Smythe, 2012). Τhe key, therefore, to preventing spoilage by thermophilic endospore - formers is to apply a structured quality assurance system based on thorough risk analysis and prevention through monitoring, recording and controlling of critical parameters throughout the entire product’s shelf - life. G. stearothermophilus is a proven quality problem for whole and skim milk powder produced across the world as it can cause long term persistent contamination of dairy processing facilities, due to its ability to form biofilms on stainless steel surfaces of processing equipment (Burgess et al., 2014; Flint et al., 1997; Rückert et al., 2004). G. stearothermophilus has been typically linked to ‘flat - sour’ spoilage, mainly of evaporated milk, due to acid coagulation at a pH level of around 5.2 (Boor & Murphy, 2002; Kakagianni, Gougouli, & Koutsoumanis, 2016; Kalogridou - Vassiliadou, 1992) in climates where ambient temperatures may allow G. stearothermophilus growth. The significant causes for spoilage include under - processing, inadequate cooling, contamination of the product, resulting from leakage through seams, pre - process spoilage and mainly temperature abuse during distribution and storage (Jay, 2012). On the other hand, since the first large - scale commercial spoilage event (Cerny et al., 1984), manufacturers and processors in the fruit industry have recognized the substantial economic and spoiling potential of A. acidoterrestris worldwide (Duong & Jensen, 2000; Jensen & Whitfield, 2003; Spllttstoesser et al., 1994; Yamazaki et al., 1996). Particularly, it has been detected in a wide range of commercially pasteurized fruit juices, bottled tea, isotonic drinks and other low pH, shelf - stable products, as

176

General Discussion and potential applications of the developed models to the food industry well as processing facilities, where it enters most likely via fruit surfaces contaminated from soil during production and harvesting (Eiroa et al., 1999; Merle, 2012). In addition to soil, water has also been proposed to be another important source of contamination in the processing environment, especially when recycled water is used in the production of fruit juices and concentrates (Chen et al., 2006; Jensen, 2000; Walls & Chuyate, 2000). The storage and distribution of fruit juices for long periods at improper and temperature - abuse conditions, like the ones existing in warehouses, delivery trucks, retail display, storage rooms and home storage (Bahçeci et al., 2005; Chang & Kang, 2004; Heyndrickx, 2011; Pettipher et al., 1997), especially during the warmer months or in tropical and semitropical regions (Roig - Sagues et al., 2015), may provoke the germination of spores, if these are present, the outgrowth and the subsequent growth of the vegetative cells of the organism (Orr et al., 2000). Aiming at plugging the gap of knowledge of spoilage thermophilic endospore - formers’ behavior in food products, this PhD research thesis was designed to study the effect of storage temperature and pH on the growth dynamics of two thermophilic endospore - forming species of bacilli at the population as well as at single spore level. Beyond the scientific interest in understanding thermophilic endospore - formers’ behavior, the development of effective hygiene control measures and sampling procedures for these bacillus species are of great importance for the quality assurance of evaporated milk and fruit drink products. Therefore, the concept of predictive microbiology can be used as an effective tool for improving the quality control systems in the mentioned food production through the improvement of shelf - life decision - making and optimization of control measures (e.g., evaporated milk or fruit drinks testing for acid coagulation or guaiacol production, respectively). In particular, predictive microbiology describes and quantifies microbial population evolution (growth, survival, inactivation) in food products as a function of various environmental (extrinsic and intrinsic) factors, such as temperature, pH and water activity, during processing and storage, combining experimental data, microbiological knowledge, and mathematical techniques (Bovill et al., 2001; Koseki & Isobe, 2005; McMeekin et al., 1997; McMeekin et al., 2013; Membré et al., 2005). The major output of the present PhD thesis is the development of predictive models for the growth of G. stearothermophilus and A. acidoterrestris, two major spoilage bacteria for heat - processed, non - refrigerated foods. Extensive validation studies showed a high performance of the models in predicting microbial growth and spoilage of foods products such as evaporated milk and fruit drinks. The

177

Chapter 6 developed models can provide the information and data required for effective shelf - life management and scientifically - based design of quality control tests, which can lead to a successful decision - making for improving quality. In the following paragraphs the potential applications of the models as well as the challenges in their use by the food industry are presented in detail. Controlling the contamination of heat - processed, non - refrigerated food with thermophilic spores is the first important step in a food quality assurance system. Apart form the prerequisite programs of the HACCP system (Good Manufacturing Practices – GMPs and Good Hygiene Practices - GHPs), an additional effective control measure for the food industry is the application of challenge tests. A challenge test for the presence of thermophilic spores is based on the incubation of a certain proportion of packages from each batch at an optimum temperature for thermophilic bacterial growth and the subsequent detection of spoiled packages after a certain time. Based on the results of this test, the decision about the release of the batch to the market can be made. The effectiveness of such a test depends on the appropriate selection of the time - temperature conditions which should allow for all potential contaminating spores to cause spoilage at the chosen test period. It needs to be noted that this period should be as short as possible, since it is related to the production cost. The optimization of the above test conditions is not an easy task and requires detailed knowledge of the thermophilic bacterial growth kinetics. The selection of the challenge test conditions without a scientific background may lead to a wrong decision with significant economic cost for the industry. The mathematical models developed in the present PhD thesis can be used as the basis for the selection of the appropriate conditions of the challenge tests for thermophilic endospore – former detection in evaporated milk and fruit drinks, leading to a successful decision - making for improving quality (Gougouli, Kalantzi, Beletsiotis & Koutsoumanis, 2011). In order to support the above option, the models developed in the thesis were incorporated in a user - friendly excel application (Fig. 6.1), which allows their use by the food industry without the need of a detailed mathematical background. With this application the time - to - spoilage (tts) of procuct can be easily estimated at various storage temperatures. Based on the estimated tts, the appropriate time - temperature conditions of the challenge test can be selected and the results can be translated to risk of spoilage providing support for the decisions of the food industry.

178

General Discussion and potential applications of the developed models to the food industry

Figure 6.1. Excel application for the use of the developed predictive models in the quality control system of fruit drinks production.

Based on the results from the first parts of the study (Chapter 2 and 3), temperature is considered as the most important environmental factor determining the germination, outgrowth and subsequent growth of thermophilic spores and, consequently, the time - to - spoilage of food products. Temperature abuses that frequently prevai in the supply chain may lead to the rejection of the product before the end of its shelf - life. Thus, the mathematical description of their effect on food spoilage is of great importance. The models developed in this thesis can be used in combination with temperature recording technologies (i.e data loggers, Time Temperature Integrators - TTI, Radio - Frequency Identification - RFID) as the basis for development of an effective supply chain management system as an alternative to the classical First – In – First - Out (FIFO) approach leading to optimization of food quality at consumer’s end (Giannakourou, Koutsoumanis, Nychas & Taoukis, 2001, 2005; Koutsoumanis, Taoukis & Nychas, 2005). In such a system, a decision - making routine at a specified control point of the supply chain is based on the growth of the spoiler that has potentially occurred within the period between production and arrival of the product at the control point. Microbial growth is estimated based on the product’s characteristics and the time - temperature history of the product using

179

Chapter 6 the appropriate predictive model. The above elements can form the programme core of an integrated software that allows the calculation of microbial growth and spoilage risk of individual product units (e.g. small pallets, 5 - 10 kg boxes or even single packs) at strategic control points of the chill chain. Based on the probability distribution of the microbial growth, it is possible to make decisions for optimal handling, shipping destination and stock rotation, aiming to obtain a narrow distribution of quality at the point of consumption. For example, at a certain point of the chill chain, e.g. at a distribution center, product from the same initial shipment is split in half and is forwarded to two different retail markets, a close and a distant one that requires long transportation. The split could be random according to conventional, currently used FIFO practice or it can be based on the actual risk of spoilage of the product units and the developed decision system leading to an optimization of product’s quality. Another potential application of the developed predictive microbiology tools on mapping the risk of non - refrigerated food products spoilage in the Mediterranean region is presented in Chapter 5. The latter application, which was actually the starting point of this PhD thesis, is related to the question from the food industry on whether and under which conditions (e.g. expiration date) will be able to export a product to a country without spoilage problems. In this case the predictive model for the effect of storage temperature on the growth of Geobacillus stearothermophilus was applied in order to assess the risk of evaporated milk spoilage in the markets of the Mediterranean region. The growth of G. stearothermophilus in evaporated milk was evaluated during a shelf life of one year based on historical temperature profiles (hourly) covering 23 Mediterranean capitals for five years over the period 2012 - 2016 obtained from the Weather Underground database (http://www.wunderground.com/). In total, 115 scenarios were tested simulating the distribution and storage conditions of evaporated milk in the Mediterranean region. The predicted extent and the variability of growth during the shelf - life were used to assess the risk of spoilage which was visualised in a geographical risk map. The growth model of G. stearothermophilus was also used to evaluate adjustments of the evaporated milk expiration date which can reduce the risk of spoilage. The quantitative data provided in the present study can assist the food industry to effectively evaluate the microbiological stability of these products throughout distribution and storage at a reduced cost (by reducing sampling quality control) leading to a significant benefit for both the competitiveness of the food industry and the consumer.

180

General Discussion and potential applications of the developed models to the food industry

In recent years, the need for including the actual variability in the microbial models used in risk assessment led to the development of stochastic models in order to provide an estimate of all possible circumstances with their associated probability, based on the variability of factors affecting microbial behavior, and to quantitatively assess the probability (risk) that a food is unsafe or spoiled (Koutsoumanis, Lianou & Gougouli, 2016; Koutsoumanis et al., 2010; Nauta, 2002). The importance of describing variability for a risk - based quality and safety control of foods was highlighted by Nauta (2002), who illustrated the effect of ignoring variability in management decisions. Variability is an inherent part of food and microorganisms. Therefore, there are various variability sources in microbial behavior that should be taken into account in risk - based approaches providing more realistic estimation of food quality. Figure 6.2 presents an example of how the variations in various factors related to microbial growth can affect exposure assessment. In particular, the behavioral differences observed among identically treated contaminating strains of the same microbial species (Lianou & Koutsoumanis, 2011a, b), the initial physiological state of cells, the initial contamination level, the individual cell heterogeneity (Aguirre & Koutsoumanis, 2016; Aspridou & Koutsoumanis, 2015; Koutsoumanis, 2008; Koutsoumanis &

Lianou, 2013), the intrinsic factors of a food product (pH, aw etc.), the extrinsic factors e.g. storage temperature and the storage time (Koutsoumanis et al., 2016) are considered as major sources of variability.

181

Chapter 6

Figure 6.2. Sources of variability affecting microbial growth and exposure assessment (adapted from Koutsoumanis et al., 2016).

Thus, moving from deterministic to stochastic modelling approaches further research on the between - spore variability is required. So in order to built a better understanding of spore’s behavior as affected by the storage temperature and given that, in practice, the spoilage defects of evaporated milks are derived from low bacterial spore numbers and these individual spores are characterised with high heterogeneity in terms of lag time duration (Baranyi, 1998; Barker et al., 2005; Pin & Baranyi, 2006; Stringer et al., 2011), further research in this thesis was conducted at single spore level (Chapter 4). Considering that the probability of growth of endospore - formers on thermally - processed, acidic products, or generally on foods, during their shelf life is strongly affected by the lag phase, the research of Chapter 4 focused on investigating the impact of different storage temperature conditions on the lag time (λ) of individual G. stearothermophilus spores. In addition, a stochastic modeling approach was applied by including the individual spore heterogeneity in growth predictions and increasing the precision and credibility of the model. A considerable heterogeneity in the cumulative distributions of the individual spore λ at the different tested temperatures was obtained, which was more profound at conditions far from the optimum ones or, in other words, it was

182

General Discussion and potential applications of the developed models to the food industry demonstrated that suboptimal temperatures led to more skewed distributions for lag time. Given the above, it can be concluded that the impact of the growth environment on the observed parameter’s variability may be considerable. We also elaborated on the effect of storage temperature that characterises the individual spore λ of G. stearothermophilus with a secondary model (CMI), and based on this model, we described quantitatively the temperature dependence of the mean individual spore λ of this bacterium. To interpret the observations, a stochastic growth model for G. stearothermophilus spores was developed, by introducing the distribution of individual spores λ values into a simple exponential growth with lag model (Koutsoumanis & Lianou, 2013). The stochastic model describes the growth of a population, initially consisting of N0 spores, over time as the sum of spores in each of the N0 imminent subpopulations originating from a single spore. The above modeling procedure allows for taking into account the heterogeneity in the growth dynamics of single spores by introducing λ variability in the model as a probability distribution using Monte Carlo simulation. As a matter of fact, variability in population growth gradually decreases with increasing initial population size (N0) to 100 spores and a shorter λ of the population is observed. For bacterial populations with N0 > 100 spores, the variability is almost eliminated and the system seems to behave deterministically, even though the underlying law is stochastic. Hence, considering that the number of G. stearothermophilus surviving spores in the final product is usually very low, the quantification of variability in the single spore behavior can provide a stochastic approach to evaluate the probability distribution of the time - to - spoilage in a more realistic and accurate way, enable decision - making based on the “acceptable level of risk” and design risk - based quality management systems for evaporated milk products. Future applications of the developed predictive microbiology tools may include the evaluation of the effect of climate change on food quality. The results of Chapter 5 showed that, based on the temperature data of the last years, the storage and distribution of evaporated milk in most Mediterranean capitals present a low risk of spoilage. Temperature, however, is expected to increase significantly due to the global warming phenomenon. Recent studies presented climate change projections over the Mediterranean region based on the latest and most advanced sets of global and regional climate model simulations. These simulations give a collective picture of a substantial warming ranging from 1 to 7 °C (Giorgi & Lionello, 2008; IPCC, 2014; Räisänen et al., 2004), depending on future Greenhouse emissions and the characteristics of the predictive model (IPCC, 2014; Morice et al., 2012). However, in

183

Chapter 6 contrary to the “undeniable global warming”, as described in the report at the Intergovernmental Panel on Climate Change (IPCC, 2013), there is uncertainty about its effects (Schröter et al., 2005). Among the anticipated effects of climate change, those concerning the agricultural and food sector have been extensively highlighted (Kleter & Marvin, 2009) and documented (Miraglia et al., 2009; Tirado et al., 2010; van der Spiegel, van der Fels-Klerx & Marvin, 2012). In 2014, the Intergovernmental Committee for the Climate Change presented a comprehensive report on the effects of climate change in food safety issues with predictions on the incidence and prevalence of foodborne diseases by pathogenic parasites, bacteria and viruses (FAO, 2016; McMichael, 2013; Smith et al., 2014) but also chemical hazards, such as mycotoxins, pesticides and algae toxins (Porter et al., 2014). Elevated temperatures are correlated with changes in survival, acceleration of the replication cycles of foodborne microorganisms or transmission potential of pathogens in the environment, food and feed (Miraglia et al., 2009), while extended summer seasons may increase the chance of food - handling errors (ECDC, 2012; Hellberg & Chu, 2015). From a food safety perspective, it should be stressed the need to control microbiological risks at any stage of the food chain, along the primary production, processing, storage and transport (Chakraborty & Newton, 2011; Miraglia et al., 2009) and develop adaptation strategies to cope with food spoilage implications of climate change in the future (Jacxsens et al., 2010; Tirado et al., 2010). Unlike the number of studies on food sustainability and food safety, the studies regarding the association of global warming with the shelf - life of foods are, in general, very limited (Medina-Martínez et al., 2015). The fact that the last decade has been characterized as the warmest of the last 150 years (Climate Central, 2016) combined with reports from the food industry about increased sporadic cases of microbiological spoilage of foods that have consistently been characterized as “microbiologically stable”, can be taken as an indication of the impact of climate change on the spoilage activity of thermophilic endopore - forming bacteria. The microbiological stability of the shelf - stable, non - perishable food products (including fruit and vegetable juices, sauces, evaporated milk, highly pasteurized milk, beverages) is based on the intense thermal processing that destroys the vegetative forms of foodborne microorganisms (i.e. Clostridium botulinum). Although the presence of the extremely heat - resistant bacterial spores surviving in the final product is possible (Bevilacqua, Sinigaglia & Corbo, 2009; Gocmen, Elston, Williams, Parish, & Rouseff, 2005), their thermophilic nature does

184

General Discussion and potential applications of the developed models to the food industry not allow their potential germination and their extended growth during long - term storage at the usual distribution and storage temperature conditions. Recently, the temperature data observed in Greece demonstrate that this situation is critical. Further warming is likely to bring about significant changes in the growth and the spoilage potential of the particular bacteria. Specificallly, the climate change may lead to a long exposure at high ambient temperatures allowing microbial growth at the spoilage level. The basic research question is whether the possible scenarios for global warming, as predicted by the models for the climate change, can affect the shelf - life of food products with high consumption, which have hitherto been considered as “microbiologically stable” and distributed without refrigeration. Consequently, with regard to the studies described in Chapter 2 and 3 and the approach presented in Chapter 5, the developed models could be extended to also evaluate the consequences of global warming on the quality of non - refrigerated foods and help the food industry be properly prepared to address the impacts of climate change through the development of adaptation strategies in processing, distribution, storage and quality control and to design effective quality management and logistic systems.

185

Chapter 6

186

References

Abee, T., Groot, M. N., Tempelaars, M., Zwietering, M., Moezelaar, R., & van der Voort, M. (2011). Germination and outgrowth of spores of Bacillus cereus group members: diversity and role of germinant receptors. Food Microbiology, 28(2), 199-208. Abriouel, H., Benomar, N., Huch, M., Franz, C. M., & Gálvez, A. (2014). The genera Bacillus, Geobacillus and Halobacillus. Lactic Acid Bacteria: Biodiversity and Taxonomy, 555-570. Aguirre, J. S., de Fernando, G. G., Hierro, E., Hospital, X. F., Ordóñez, J. A., & Fernández, M. (2015). Estimation of the growth kinetic parameters of Bacillus cereus spores as affected by pulsed light treatment. International journal of food microbiology, 202, 20-26. Aguirre, J. S., & Koutsoumanis, K. P. (2016). Towards lag phase of microbial populations at growth-limiting conditions: The role of the variability in the growth limits of individual cells. International journal of food microbiology, 224(Supplement C), 1-6. doi: https://doi.org/10.1016/j.ijfoodmicro.2016.01.021 Aguirre, J. S., Ordóñez, J. A., & de Fernando, G. D. G. (2012). A comparison of the effects of E-beam irradiation and heat treatment on the variability of Bacillus cereus inactivation and lag phase duration of surviving cells. International journal of food microbiology, 153(3), 444-452. Aguirre, J. S., Rodríguez, M. R., & de Fernando, G. D. G. (2011). Effects of electron beam irradiation on the variability in survivor number and duration of lag phase of four food-borne organisms. International journal of food microbiology, 149(3), 236-246. Ahn, J., Lee, H. Y., & Balasubramaniam, V. (2015). Inactivation of Geobacillus stearothermophilus spores in low‐acid foods by pressure‐assisted thermal processing. Journal of the Science of Food and Agriculture, 95(1), 174-178. Al-Qadiri, H. M., Lin, M., Cavinato, A. G., & Rasco, B. A. (2006). Fourier transform infrared spectroscopy, detection and identification of Escherichia coli O157: H7 and Alicyclobacillus strains in apple juice. International journal of food microbiology, 111(1), 73-80. Al‐Holy, M. A., Lin, M., Alhaj, O. A., & Abu‐Goush, M. H. (2015). Discrimination between Bacillus and Alicyclobacillus isolates in apple juice by Fourier Transform Infrared Spectroscopy and multivariate analysis. Journal of food science, 80(2), M399-M404. Albuquerque, L., Rainey, F., Chung, A., Sunna, A., Nobre, M., Grote, R., . . . Da Costa, M. (2000). Alicyclobacillus hesperidum sp. nov. and a related genomic species from

solfataric soils of São Miguel in the Azores. International Journal of Systematic and Evolutionary Microbiology, 50(2), 451-457. Alpas, H., Alma, L., & Bozoglu, F. (2003). Inactivation of Alicyclobacillus acidoterrestris vegetative cells in model system, apple, orange and tomato juices by high hydrostatic pressure. World Journal of Microbiology and Biotechnology, 19(6), 619- 623. Ananta, E., Heinz, V., Schlüter, O., & Knorr, D. (2001). Kinetic studies on high-pressure inactivation of Bacillus stearothermophilus spores suspended in food matrices. Innovative Food Science & Emerging Technologies, 2(4), 261-272. Anderson, R. E. (1984). Growth and corresponding elevation of tomato juice pH by Bacillus coagulans. Journal of food science, 49(2), 647-647. André, S., Zuber, F., & Remize, F. (2013). Thermophilic spore-forming bacteria isolated from spoiled canned food and their heat resistance. Results of a French ten-year survey. International journal of food microbiology, 165(2), 134-143. Antolinos, V., Muñoz-Cuevas, M., Ros-Chumillas, M., Periago, P., Fernández, P., & Le Marc, Y. (2012). Modelling the effects of temperature and osmotic shifts on the growth kinetics of Bacillus weihenstephanensis in broth and food products. International journal of food microbiology, 158(1), 36-41. Arancia, G., Belli, S., Donelli, G., & Trovalusci, P. (1980). Ultrastructural changes in Escherichia coli grown in divalent cation-deficient medium. Microbiology, 119, 155- 164. Ash, C., Farrow, J., Wallbanks, S., & Collins, M. (1991). Phylogenetic heterogeneity of the genus Bacillus revealed by comparative analysis of small‐subunit‐ribosomal RNA sequences. Letters in applied microbiology, 13(4), 202-206. Ashton, D., & Busta, F. (1968). Milk Components Inhibitory to Bacillus stearothermophilus 1, 2. Journal of dairy science, 51(6), 842-847. Aspridou, Z., & Koutsoumanis, K. P. (2015). Individual cell heterogeneity as variability source in population dynamics of microbial inactivation. Food Microbiology, 45(Part B), 216-221. doi: https://doi.org/10.1016/j.fm.2014.04.008 Atrih, A., & Foster, S. (2001). Analysis of the role of bacterial endospore cortex structure in resistance properties and demonstration of its conservation amongst species. Journal of Applied Microbiology, 91(2), 364-372.

189

Atrih, A., & Foster, S. J. (1999). The role of peptidoglycan structure and structural dynamics during endospore dormancy and germination. Antonie van Leeuwenhoek, 75(4), 299-307. Augustin, J.-C., Bergis, H., Midelet-Bourdin, G., Cornu, M., Couvert, O., Denis, C., . . . Stahl, V. (2011). Design of challenge testing experiments to assess the variability of Listeria monocytogenes growth in foods. Food Microbiology, 28(4), 746-754. doi: https://doi.org/10.1016/j.fm.2010.05.028 Bahçeci, K. S., & Acar, J. (2007). Determination of guaiacol produced by Alicyclobacillus acidoterrestris in apple juice by using HPLC and spectrophotometric methods, and mathematical modeling of guaiacol production. European Food Research and Technology, 225(5-6), 873-878. Bahçeci, K. S., Gökmen, V., & Acar, J. (2005). Formation of guaiacol from vanillin by Alicyclobacillus acidoterrestris in apple juice: a model study. European Food Research and Technology, 220(2), 196-199. Bailey, C., & Von Holy, A. (1993). Bacillus spore contamination associated with commercial bread manufacture. Food Microbiology, 10(4), 287-294. Banat, I. M., Marchant, R., & Rahman, T. J. (2004). Geobacillus debilis sp. nov., a novel obligately thermophilic bacterium isolated from a cool soil environment, and reassignment of Bacillus pallidus to Geobacillus pallidus comb. nov. International Journal of Systematic and Evolutionary Microbiology, 54(6), 2197-2201. Baranyi, J. (1998). Comparison of stochastic and deterministic concepts of bacterial lag. Journal of Theoretical Biology, 192(3), 403-408. Baranyi, J. (2002). Stochastic modelling of bacterial lag phase. International journal of food microbiology, 73(2), 203-206. Baranyi, J., George, S. M., & Kutalik, Z. (2009). Parameter estimation for the distribution of single cell lag times. Journal of Theoretical Biology, 259(1), 24-30. Baranyi, J., & Pin, C. (1999). Estimating bacterial growth parameters by means of detection times. Applied and Environmental Microbiology, 65(2), 732-736. Baranyi, J., & Pin, C. (2001). A Parallel Study on Bacterial Growth and Inactivation. Journal of Theoretical Biology, 210(3), 327-336. doi: https://doi.org/10.1006/jtbi.2001.2312

190

Baranyi, J., & Pin, C. (2004). Modeling the history effect on microbial growth and survival: deterministic and stochastic approaches Modeling microbial responses in food (pp. 285-302): CRC Press Inc. Boca Raton, FL. Baranyi, J., & Roberts, T. A. (1994). A dynamic approach to predicting bacterial growth in food. International journal of food microbiology, 23(3-4), 277-294. Baranyi, J., & Roberts, T. A. (1995). Mathematics of predictive food microbiology. International journal of food microbiology, 26(2), 199-218. Barker, G., Malakar, P., & Peck, M. (2005). Germination and growth from spores: variability and uncertainty in the assessment of food borne hazards. International journal of food microbiology, 100(1), 67-76. Batra, N., Singh, J., Banerjee, U. C., Patnaik, P. R., & Sobti, R. C. (2002). Production and characterization of a thermostable β‐galactosidase from Bacillus coagulans RCS3. Biotechnology and applied biochemistry, 36(1), 1-6. Baumgart, J. (2000). The impact of Alicyclobacillus acidoterrestris on the quality of juices and soft drinks. Fruit processing: journal for the fruit processing and juice producing european and overseas industry, 10(7), 251-254. Bayoumi, M. A., Kamal, R. M., El Aal, S. F. A., & Awad, E. I. (2012). Assessment of a regulatory sanitization process in Egyptian dairy plants in regard to the adherence of some food-borne pathogens and their biofilms. International journal of food microbiology, 158(3), 225-231. Baysal, A. H., Molva, C., & Unluturk, S. (2013). UV-C light inactivation and modeling kinetics of Alicyclobacillus acidoterrestris spores in white grape and apple juices. International journal of food microbiology, 166(3), 494-498. Beaman, T. C., & Gerhardt, P. (1986). Heat resistance of bacterial spores correlated with protoplast dehydration, mineralization, and thermal adaptation. Applied and Environmental Microbiology, 52(6), 1242-1246. Bevilacqua, A., Campaniello, D., Speranza, B., Sinigaglia, M., & Corbo, M. R. (2013). Control of Alicyclobacillus acidoterrestris in apple juice by citrus extracts and a mild heat-treatment. Food control, 31(2), 553-559. Bevilacqua, A., Cibelli, F., Corbo, M., & Sinigaglia, M. (2007). Effects of high‐pressure homogenization on the survival of Alicyclobacillus acidoterrestris in a laboratory medium. Letters in applied microbiology, 45(4), 382-386.

191

Bevilacqua, A., Ciuffreda, E., Sinigaglia, M., & Corbo, M. R. (2014). Effects of lysozyme on Alicyclobacillus acidoterrestris under laboratory conditions. International journal of food science & technology, 49(1), 224-229. Bevilacqua, A., Corbo, M., & Sinigaglia, M. (2008a). Combined effects of low pH and cinnamaldehyde on the inhibition of Alicyclobacillus acidoterrestris spores in a laboratory medium. Journal of Food Processing and Preservation, 32(5), 839-852. Bevilacqua, A., Corbo, M., & Sinigaglia, M. (2008b). Inhibition of Alicyclobacillus acidoterrestris spores by natural compounds. International journal of food science & technology, 43(7), 1271-1275. Bevilacqua, A., Corbo, M. R., & Sinigaglia, M. (2010). Combining eugenol and cinnamaldehyde to control the growth of Alicyclobacillus acidoterrestris. Food control, 21(2), 172-177. Bevilacqua, A., Sinigaglia, M., & Corbo, M. (2008). Alicyclobacillus acidoterrestris: new methods for inhibiting spore germination. International journal of food microbiology, 125(2), 103-110. Bevilacqua, A., Sinigaglia, M., & Corbo, M. R. (2009). Effects of pH, cinnamaldehyde and heat‐treatment time on spore viability of Alicyclobacillus acidoterrestris. International journal of food science & technology, 44(2), 380-385. Bianchi, F., Careri, M., Mangia, A., Mattarozzi, M., Musci, M., Concina, I., & Gobbi, E. (2010). Characterisation of the volatile profile of orange juice contaminated with Alicyclobacillus acidoterrestris. Food chemistry, 123, 653-658. Billon, C. P., McKirgan, C., McClure, P., & Adair, C. (1997). The effect of temperature on the germination of single spores of Clostridium botulinum 62A. Journal of Applied Microbiology, 82(1), 48-56. Blana, V. A., Lianou, A., & Nychas, G.-J. E. (2015). Assessment of the effect of a Salmonella enterica ser. Typhimurium culture supernatant on the single-cell lag time of foodborne pathogens. International journal of food microbiology, 215, 143- 148. Boor, K. J., & Murphy, S. C. (2002). Microbiology of market milks Dairy Microbiology Handbook: The Microbiology of Milk and Milk Products (pp. 91-122): John Wiley & Sons Inc., Hoboken, NJ. Borlinghaus, A., & Engel, R. (1997). Alicyclobacillus incidence in commercial apple juice concentrate (AJC) supplies - method development and validation. Fruit processing, 7, 262-266.

192

Bovill, R. A., Bew, J., & Baranyi, J. (2001). Measurements and predictions of growth for Listeria monocytogenes and Salmonella during fluctuating temperature: II. Rapidly changing temperatures. International journal of food microbiology, 67(1-2), 131- 137. Branda, S. S., Vik, Å., Friedman, L., & Kolter, R. (2005). Biofilms: the matrix revisited. Trends in microbiology, 13(1), 20-26. Brent Seale, R., Flint, S., James McQuillan, A., & Bremer, P. (2011). Effect of NaOH (caustic wash) on the viability, surface characteristics and adhesion of spores of a Geobacillus sp. isolated from a milk powder production line. Letters in applied microbiology, 52(2), 104-108. Brooks, J. D., & Flint, S. H. (2008). Biofilms in the food industry: problems and potential solutions. International journal of food science & technology, 43(12), 2163-2176. Brown, K. (2000). Control of bacterial spores. British Medical Bulletin, 56(1), 158-171. Buchanan, R., Smith, J., McColgan, C., Marmer, B., Golden, M., & Dell, B. (1993). Response surface models for the effects of temperature, pH, sodium chloride, and sodium nitrite on the aerobic and anaerobic growth of Staphylococcus aureus 196E. Journal of Food Safety, 13, 159-175. Buehner, K. P., Anand, S., & Garcia, A. (2014). Prevalence of thermoduric bacteria and spores on 10 Midwest dairy farms. Journal of dairy science, 97(11), 6777-6784. Burgess, S., Brooks, J., Rakonjac, J., Walker, K., & Flint, S. (2009). The formation of spores in biofilms of Anoxybacillus flavithermus. Journal of Applied Microbiology, 107(3), 1012-1018. Burgess, S., Flint, S., & Lindsay, D. (2014). Characterization of thermophilic bacilli from a milk powder processing plant. Journal of Applied Microbiology, 116(2), 350-359. Burgess, S. A., Lindsay, D., & Flint, S. H. (2010). Thermophilic bacilli and their importance in dairy processing. International journal of food microbiology, 144(2), 215-225. doi: http://dx.doi.org/10.1016/j.ijfoodmicro.2010.09.027 Byrer, D. E., Rainey, F. A., & Wiegel, J. (2000). Novel strains of Moorella thermoacetica form unusually heat-resistant spores. Archives of Microbiology, 174(5), 334-339. Cai, R., Yuan, Y., Wang, Z., Guo, C., Liu, B., Liu, L., Wang, Y., & Yue, T. (2015). Precursors and metabolic pathway for guaiacol production by Alicyclobacillus acidoterrestris. International journal of food microbiology, 214, 48-53.

193

Caipo, M., Duffy, S., Zhao, L., & Schaffner, D. (2002). Bacillus megaterium spore germination is influenced by inoculum size. Journal of Applied Microbiology, 92(5), 879-884. Carlier, J.-P., & Bedora-Faure, M. (2006). Phenotypic and genotypic characterization of some Moorella sp. strains isolated from canned foods. Systematic and Applied Microbiology, 29(7), 581-588. Carlier, J.-P., Bonne, I., & Bedora-Faure, M. (2006). Isolation from canned foods of a novel Thermoanaerobacter species phylogenetically related to Thermoanaerobacter mathranii (Larsen 1997): Emendation of the species description and proposal of Thermoanaerobacter mathranii subsp. Alimentarius subsp. Nov. Anaerobe, 12(3), 153-159. Carlin, F. (2011). Origin of bacterial spores contaminating foods. Food Microbiology, 28(2), 177-182. Casadei, M., Ingram, R., Skinner, R., & Gaze, J. (2000). Heat resistance of Paenibacillus polymyxa in relation to pH and acidulants. Journal of Applied Microbiology, 89(5), 801-806. Caspers, M. P., Boekhorst, J., Abee, T., Siezen, R. J., & Kort, R. (2013). Complete genome sequence of Anoxybacillus flavithermus TNO-09.006, a thermophilic sporeformer associated with a dairy-processing environment. Genome announcements, 1(1), e00010-00013. Caspers, M. P., Boekhorst, J., de Jong, A., Kort, R., Groot, M. N., & Abee, T. (2016). Draft genome sequences of four thermophilic spore formers isolated from a dairy- processing environment. Genome announcements, 4(4), e00757-00716. Caspers, M. P., Schuren, F. H., van Zuijlen, A. C., Brul, S., Montijn, R. C., Abee, T., & Kort, R. (2011). A mixed-species microarray for identification of food spoilage bacilli. Food Microbiology, 28(2), 245-251. Cate, J. H., Gooding, A. R., Podell, E., & Zhou, K. (1996). RNA tertiary structure mediation by adenosine platforms. Science, 273(5282), 1696. Cazemier, A., Wagenaars, S., & Ter Steeg, P. (2001). Effect of sporulation and recovery medium on the heat resistance and amount of injury of spores from spoilage bacilli. Journal of Applied Microbiology, 90(5), 761-770. Cerny, G., Hennlich, W., & Poralla, K. (1984). Fruchtsaftverderb durch Bacillen: isolierung und charakterisierung des verderbserregers. Zeitschrift fuer Lebensmittel- Untersuchung und Forschung, 179(3), 224-227.

194

Chakraborty, S., & Newton, A. C. (2011). Climate change, plant diseases and food security: an overview. Plant Pathology, 60(1), 2-14. Chandler, D. P., Brown, J., Bruckner-Lea, C. J., Olson, L., Posakony, G. J., Stults, J. R., . . . Bond, L. J. (2001). Continuous spore disruption using radially focused, high- frequency ultrasound. Analytical chemistry, 73(15), 3784-3789. Chang, S.-S., & Kang, D.-H. (2004). Alicyclobacillus spp. in the fruit juice industry: history, characteristics, and current isolation/detection procedures. Critical reviews in microbiology, 30(2), 55-74. Chang, S., & Kang, D. H. (2005). Development of novel Alicyclobacillus spp. isolation medium. Journal of Applied Microbiology, 99(5), 1051-1060. Chang, S., Park, S., & Kang, D. (2013). Development of novel agar media for isolating guaiacol producing Alicyclobacillus spp. International journal of food microbiology, 164(1), 1-6. Chauhan, K., Dhakal, R., Seale, R.B., Deeth, H.C., Pillidge, C.J., Powell, I.B., Craven, H., & Turner, M.S. (2013). Rapid identification of dairy mesophilic and thermophilic sporeforming bacteria using DNA high resolution melt analysis of variable 16S rDNA regions. International journal of food microbiology, 165, 175-183. Chen, D., Huang, S.-s., & Li, Y.-q. (2006). Real-time detection of kinetic germination and heterogeneity of single Bacillus spores by laser tweezers Raman spectroscopy. Analytical chemistry, 78(19), 6936-6941. Chen, L., Coolbear, T., & Daniel, R. M. (2004). Characteristics of proteinases and lipases produced by seven Bacillus sp. isolated from milk powder production lines. International dairy journal, 14(6), 495-504. Chen, S., Tang, Q., Zhang, X., Zhao, G., Hu, X., Liao, X., . . . Xiang, H. (2006). Isolation and characterization of thermo-acidophilic endospore-forming bacteria from the concentrated apple juice-processing environment. Food Microbiology, 23(5), 439- 445. Cleverdon, R. C., Pelczar Jr, M. J., & Doetsch, R. N. (1949). The vitamin requirements of stenothermophilic aerobic sporogenous Bacilli. Journal of bacteriology, 58(4), 523. Climate Central [WWW Document], (2016). URL http://www.climatecentral.org/gallery/graphics/the-10-hottest-years-on-record Codex alimentarius (1979). Code of hygienic practice for low acid and acidified low acid canned foods, CAC/RCP, vol. 23. Codex alimentarius (rev 2, 1993).

195

Codex Alimentarius Commission (1999). Draft principles and guidelines for the conduct of microbiological risk assessment (at step 8 of the procedure). ALINORM: Appendix 2. http://www.fao.org/docrep/meeting/005/X0798e.X0798e01.htm#bm21.2. Coleman, M. E., Tamplin, M. L., Phillips, J. G., & Marmer, B. S. (2003). Influence of agitation, inoculum density, pH, and strain on the growth parameters of Escherichia coli O157:H7—relevance to risk assessment. International journal of food microbiology, 83(2), 147-160. doi: https://doi.org/10.1016/S0168-1605(02)00367-7 Collins, M., Lawson, P., Willems, A., Cordoba, J., Fernandez-Garayzabal, J., Garcia, P., Cai, J., Hippe, H., & Farrow, J. (1994). The phylogeny of the genus Clostridium: proposal of five new genera and eleven new species combinations. International Journal of Systematic and Evolutionary Microbiology, 44, 812-826. Connor, C.J., Luo, H., Gardener, B.B.M., & Wang, H.H. (2005). Development of a real- time PCR-based system targeting the 16S rRNA gene sequence for rapid detection of Alicyclobacillus spp. in juice products. International journal of food microbiology, 99, 229-235. Cook, G., & Sandeman, R. (2000). Sources and characterisation of spore-forming bacteria in raw milk. Australian Journal of Dairy Technology, 55(3), 119. Coorevits, A., De Jonghe, V., Vandroemme, J., Reekmans, R., Heyrman, J., Messens, W., . . . Heyndrickx, M. (2008). Comparative analysis of the diversity of aerobic spore- forming bacteria in raw milk from organic and conventional dairy farms. Systematic and Applied Microbiology, 31(2), 126-140. Coorevits, A., Dinsdale, A. E., Halket, G., Lebbe, L., De Vos, P., Van Landschoot, A., & Logan, N. A. (2012). Taxonomic revision of the genus Geobacillus: emendation of Geobacillus, G. stearothermophilus, G. jurassicus, G. toebii, G. thermodenitrificans and G. thermoglucosidans (nom. corrig., formerly ‘thermoglucosidasius’); transfer of Bacillus thermantarcticus to the genus as G. thermantarcticus comb. nov.; proposal of Caldibacillus debilis gen. nov., comb. nov.; transfer of G. tepidamans to Anoxybacillus as A. tepidamans comb. nov.; and proposal of Anoxybacillus caldiproteolyticus sp. nov. International Journal of Systematic and Evolutionary Microbiology, 62(7), 1470-1485. Cortezzo, D., & Setlow, P. (2005). Analysis of factors that influence the sensitivity of spores of Bacillus subtilis to DNA damaging chemicals. Journal of Applied Microbiology, 98(3), 606-617.

196

Cosentino, S., Mulargia, A., Pisano, B., Tuveri, P., & Palmas, F. (1997). Incidence and biochemical characteristics of Bacillus flora in Sardinian dairy products. International journal of food microbiology, 38(2), 235-238. Couvert, O., Pinon, A., Bergis, H., Bourdichon, F., Carlin, F., Cornu, M., . . . Augustin, J.-C. (2010). Validation of a stochastic modelling approach for Listeria monocytogenes growth in refrigerated foods. International journal of food microbiology, 144(2), 236-242. doi: https://doi.org/10.1016/j.ijfoodmicro.2010.09.024 Crielly, E., Logan, N., & Anderton, A. (1994). Studies on the Bacillus flora of milk and milk products. Journal of Applied Bacteriology, 77(3), 256-263. Cuppers, H. G., Oomes, S., & Brul, S. (1997). A model for the combined effects of temperature and salt concentration on growth rate of food spoilage molds. Applied and Environmental Microbiology, 63(10), 3764-3769. Daelman, J., Vermeulen, A., Willemyns, T., Ongenaert, R., Jacxsens, L., Uyttendaele, M., & Devlieghere, F. (2013). Growth/no growth models for heat-treated psychrotrophic Bacillus cereus spores under cold storage. International journal of food microbiology, 161(1), 7-15. Dalgaard, P., & Koutsoumanis, K. (2001). Comparison of maximum specific growth rates and lag times estimated from absorbance and viable count data by different mathematical models. Journal of microbiological methods, 43(3), 183-196.

Daniel, S. L., Hsu, T., Dean, S., & Drake, H. (1990). Characterization of the H2-and CO- dependent chemolithotrophic potentials of the acetogens Clostridium thermoaceticum and Acetogenium kivui. Journal of bacteriology, 172(8), 4464-4471. Das, A., Hugenholtz, J., Van Halbeek, H., & Ljungdahl, L. (1989). Structure and function of a menaquinone involved in electron transport in membranes of Clostridium thermoautotrophicum and Clostridium thermoaceticum. Journal of bacteriology, 171(11), 5823-5829. De Clerck, E., & De Vos, P. (2002). Study of the bacterial load in a gelatine production process focussed on Bacillus and related endosporeforming genera. Systematic and Applied Microbiology, 25(4), 611-617. De Clerck, E., Rodriguez-Diaz, M., Forsyth, G., Lebbe, L., Logan, N. A., & DeVos, P. (2004). Polyphasic characterization of Bacillus coagulans strains, illustrating heterogeneity within this species, and emended description of the species. Systematic and Applied Microbiology, 27(1), 50-60.

197

De Clerck, E., Rodriguez-Diaz, M., Vanhoutte, T., Heyrman, J., Logan, N. A., & De Vos, P. (2004). Anoxybacillus contaminans sp. nov. and Bacillus gelatini sp. nov., isolated from contaminated gelatin batches. International Journal of Systematic and Evolutionary Microbiology, 54(3), 941-946. De Clerck, E., Vanhoutte, T., Hebb, T., Geerinck, J., Devos, J., & De Vos, P. (2004). Isolation, characterization, and identification of bacterial contaminants in semifinal gelatin extracts. Applied and Environmental Microbiology, 70(6), 3664-3672. De Jonghe, V., Coorevits, A., De Block, J., Van Coillie, E., Grijspeerdt, K., Herman, L., . . . Heyndrickx, M. (2010). Toxinogenic and spoilage potential of aerobic spore-formers isolated from raw milk. International journal of food microbiology, 136(3), 318-325. de Oliveira Junior, A. A., de Araújo Couto, H. G. S., Barbosa, A. A. T., Carnelossi, M. A. G., & de Moura, T. R. (2015). Stability, antimicrobial activity, and effect of nisin on the physico-chemical properties of fruit juices. International journal of food microbiology, 211, 38-43. Delignette-Muller, M. L., Cornu, M., Pouillot, R., & Denis, J. B. (2006). Use of Bayesian modelling in risk assessment: Application to growth of Listeria monocytogenes and food flora in cold-smoked salmon. International journal of food microbiology, 106(2), 195-208. doi: https://doi.org/10.1016/j.ijfoodmicro.2005.06.021 Delignette-Muller, M. L., & Rosso, L. (2000). Biological variability and exposure assessment. International journal of food microbiology, 58(3), 203-212. doi: https://doi.org/10.1016/S0168-1605(00)00274-9 Delves-Broughton, J., Blackburn, P., Evans, R., & Hugenholtz, J. (1996). Applications of the bacteriocin, nisin. Antonie van Leeuwenhoek, 69(2), 193-202. Dhakal, R., Chauhan, K., Seale, R. B., Deeth, H. C., Pillidge, C. J., Powell, I. B., . . . Turner, M. S. (2013). Genotyping of dairy Bacillus licheniformis isolates by high resolution melt analysis of multiple variable number tandem repeat loci. Food Microbiology, 34(2), 344-351. Dhakal, R., Seale, R. B., Deeth, H. C., Craven, H., & Turner, M. S. (2014). Draft genome comparison of representatives of the three dominant genotype groups of dairy Bacillus licheniformis strains. Applied and Environmental Microbiology, 80(11), 3453-3462. Dogan, Z., Weidendorfer, K., Müller-Merbach, M., Lembke, F., & Hinrichs, J. (2009). Inactivation kinetics of Bacillus spores in batch- and continuous-heating systems.

198

LWT - Food Science and Technology, 42(1), 81-86. doi: https://doi.org/10.1016/j.lwt.2008.05.005 Dotzauer, C., Ehrmann, M. A., & Vogel, R. F. (2002). Occurrence and detection of Thermoanaerobacterium and Thermoanaerobacter in canned food. Food Technology and Biotechnology, 40(1), 21-26. Drake, H. L., & Daniel, S. L. (2004). Physiology of the thermophilic acetogen Moorella thermoacetica. Research in microbiology, 155(10), 869-883. Driks, A. (1999). Bacillus subtilis spore coat. Microbiology and Molecular Biology Reviews, 63(1), 1-20. Driks, A. (2002). Overview: development in bacteria: spore formation in Bacillus subtilis. Cellular and Molecular Life Sciences CMLS, 59(3), 389-391. Duong, H., & Jensen, N. (2000). Spoilage of iced tea by Alicyclobacillus. Food Australia, 52(7), 292. Dupont, C., & Augustin, J.-C. (2009). Influence of stress on single-cell lag time and growth probability for Listeria monocytogenes in half Fraser broth. Applied and Environmental Microbiology, 75(10), 3069-3076. Durak, M. Z., Fromm, H. I., Huck, J. R., Zadoks, R. N., & Boor, K. J. (2006). Development of molecular typing methods for Bacillus spp. and Paenibacillus spp. isolated from fluid milk products. Journal of food science, 71(2), M50-M56. Durand, L., Planchon, S., Guinebretiere, M.-H., André, S., Carlin, F., & Remize, F. (2015). Contamination pathways of spore-forming bacteria in a vegetable cannery. International journal of food microbiology, 202, 10-19. European Centre for Disease Prevention and Control (ECDC) (2012). Assessing the potential impacts of climate change on food- and waterborne diseases in Europe. Stockholm (available at http://ecdc.europa.eu/en/publications/Publications/1203- TER-Potential-impacts-climate-change-food-water-borne-diseases.pdf). Eijlander, R. T., Abee, T., & Kuipers, O. P. (2011). Bacterial spores in food: how phenotypic variability complicates prediction of spore properties and bacterial behavior. Current opinion in biotechnology, 22(2), 180-186. Eiroa, M. N. U., Junqueira, V. C. A., & Schmidt, F. L. (1999). Alicyclobacillus in orange juice: occurrence and heat resistance of spores. Journal of Food Protection®, 62(8), 883-886.

199

Elfwing, A., LeMarc, Y., Baranyi, J., & Ballagi, A. (2004). Observing growth and division of large numbers of individual bacteria by image analysis. Applied and Environmental Microbiology, 70(2), 675-678. Elowitz, M. B., Levine, A. J., Siggia, E. D., & Swain, P. S. (2002). Stochastic gene expression in a single cell. Science, 297(5584), 1183-1186. Errington, J. (2003). Regulation of endospore formation in Bacillus subtilis. Nature Reviews Microbiology, 1(2), 117-126. Espejo, G. G. A., Hernández-Herrero, M., Juan, B., & Trujillo, A. (2014). Inactivation of Bacillus spores inoculated in milk by ultra high pressure homogenization. Food Microbiology, 44, 204-210. Faille, C., Bénézech, T., Midelet-Bourdin, G., Lequette, Y., Clarisse, M., Ronse, G., . . . Slomianny, C. (2014). Sporulation of Bacillus spp. within biofilms: a potential source of contamination in food processing environments. Food Microbiology, 40, 64-74. Faille, C., Fontaine, F., & Bénézech, T. (2001). Potential occurrence of adhering living Bacillus spores in milk product processing lines. Journal of Applied Microbiology, 90(6), 892-900. FAO (2016). Climate change and food security: risks and responses. Rome (available at http://www.fao.org/3/a-i5188e.pdf). Ferrario, M., Alzamora, S. M., & Guerrero, S. (2015). Study of the inactivation of spoilage microorganisms in apple juice by pulsed light and ultrasound. Food Microbiology, 46, 635-642. Fernández-No, I., Guarddon, M., Böhme, K., Cepeda, A., Calo-Mata, P., & Barros- Velázquez, J. (2011). Detection and quantification of spoilage and pathogenic Bacillus cereus, Bacillus subtilis and Bacillus licheniformis by real-time PCR. Food Microbiology, 28, 605-610. Fields, M. L. (1970). The flat sour bacteria. Adv. Fd. Res., 18, 163-217. Flint, S., Bremer, P., & Brooks, J. (1997). Biofilms in dairy manufacturing plant‐description, current concerns and methods of control. Biofouling, 11(1), 81- 97. Flint, S., Drocourt, J.-L., Walker, K., Stevenson, B., Dwyer, M., Clarke, I., & McGill, D. (2006). A rapid, two-hour method for the enumeration of total viable bacteria in samples from commercial milk powder and whey protein concentrate powder manufacturing plants. International dairy journal, 16, 379-384.

200

Flint, S., Walker, K., Waters, B., & Crawford, R. (2007). Description and validation of a rapid (1 h) flow cytometry test for enumerating thermophilic bacteria in milk powders. Journal of Applied Microbiology, 102(4), 909-915. Flint, S. H., Ward, L. J., & Walker, K. M. (2001). Functional grouping of thermophilic Bacillus strains using amplification profiles of the 16S–23S internal spacer region. Systematic and Applied Microbiology, 24(4), 539-548. Fontaine, F., Peterson, W., McCoy, E., Johnson, M. J., & Ritter, G. J. (1942). A new type of glucose fermentation by Clostridium thermoaceticum. Journal of bacteriology, 43(6), 701. Gao, Y.-l., Ju, X.-R., & Jiang, H.-H. (2006). Analysis of reduction of Geobacillus stearothermophilus spores treated with high hydrostatic pressure and mild heat in milk buffer. Journal of biotechnology, 125(3), 351-360. Gao, Y. L. (2010). Efficacy of high hydrostatic pressure and mild heat to reduce Geobacillus stearothermophilus as 1.1923 spores in model food systems. Journal of Food Safety, 30(1), 124-140. Garcia, D., van der Voort, M., & Abee, T. (2010). Comparative analysis of Bacillus weihenstephanensis KBAB4 spores obtained at different temperatures. International journal of food microbiology, 140(2), 146-153. Garrity, G. M., Bell, J. A., & Lilburn, T. G. (2004). Taxonomic outline of the prokaryotes. Bergey's manual of systematic bacteriology. Springer, New York, Berlin, Heidelberg. George, S. M., Metris, A., & Stringer, S. C. (2008). Physiological state of single cells of Listeria innocua in organic acids. International journal of food microbiology, 124(2), 204-210. Georget, E., Kapoor, S., Winter, R., Reineke, K., Song, Y., Callanan, M., . . . Mathys, A. (2014). In situ investigation of Geobacillus stearothermophilus spore germination and inactivation mechanisms under moderate high pressure. Food Microbiology, 41, 8-18. Georget, E., Miller, B., Aganovic, K., Callanan, M., Heinz, V., & Mathys, A. (2014). Bacterial spore inactivation by ultra-high pressure homogenization. Innovative Food Science & Emerging Technologies, 26, 116-123. Ghosh, S., & Setlow, P. (2009). Isolation and characterization of superdormant spores of Bacillus species. Journal of bacteriology, 191(6), 1787-1797.

201

Ghosh, S., & Setlow, P. (2010). The preparation, germination properties and stability of superdormant spores of Bacillus cereus. Journal of Applied Microbiology, 108(2), 582-590. Ghosh, S., Zhang, P., Li, Y.-q., & Setlow, P. (2009). Superdormant spores of Bacillus species have elevated wet-heat resistance and temperature requirements for heat activation. Journal of bacteriology, 191(18), 5584-5591. Giannakourou, M., Koutsoumanis, K., Nychas, G., & Taoukis, P. (2001). Development and assessment of an intelligent shelf life decision system for quality optimization of the food chill chain. Journal of Food Protection, 64(7), 1051-1057. Giannakourou, M., Koutsoumanis, K., Nychas, G., & Taoukis, P. (2005). Field evaluation of the application of time temperature integrators for monitoring fish quality in the chill chain. International journal of food microbiology, 102(3), 323-336. Giorgi, F., & Lionello, P. (2008). Climate change projections for the Mediterranean region. Global and Planetary Change, 63(2-3), 90-104. Gobbi, E., Falasconi, M., Concina, I., Mantero, G., Bianchi, F., Mattarozzi, M., . . . Sberveglieri, G. (2010). Electronic nose and Alicyclobacillus spp. spoilage of fruit juices: an emerging diagnostic tool. Food control, 21(10), 1374-1382. Gocmen, D., Elston, A., Williams, T., Parish, M., & Rouseff, R. (2005). Identification of medicinal off‐flavours generated by Alicyclobacillus species in orange juice using GC–olfactometry and GC–MS. Letters in applied microbiology, 40(3), 172-177. González, I., Lopez, M., Martınez, S., Bernardo, A., & González, J. (1999). Thermal inactivation of Bacillus cereus spores formed at different temperatures. International journal of food microbiology, 51(1), 81-84. Gopal, N., Hill, C., Ross, P. R., Beresford, T. P., Fenelon, M. A., & Cotter, P. D. (2015). The Prevalence and Control of Bacillus and Related Spore-Forming Bacteria in the Dairy Industry. Frontiers in microbiology, 6. Goto, K. (2000). Thermoacidophilic spore-forming bacteria: Alicyclobacillus. Journal Antibacterial and Antifungal Agents Japan, 28, 13-38. Goto, K., Matsubara, H., Mochida, K., Matsumura, T., Hara, Y., Niwa, M., & Yamasato, K. (2002). Alicyclobacillus herbarius sp. nov., a novel bacterium containing omega- cycloheptane fatty acids, isolated from herbal tea. International Journal of Systematic and Evolutionary Microbiology, 52(1), 109-113.

202

Goto, K., Mochida, K., Kato, Y., Asahara, M., Ozawa, C., Kasai, H., & Yokota, A. (2006). Diversity of Alicyclobacillus isolated from fruit juices and their raw materials, and emended description of Alicyclobacillus acidocaldarius. Microbiology and Culture Collections (Japan). Goto, K., Nishibori, A., Wasada, Y., Furuhata, K., Fukuyama, M., & Hara, M. (2008). Identification of thermo‐acidophilic bacteria isolated from the soil of several Japanese fruit orchards. Letters in applied microbiology, 46(3), 289-294. Gougouli, M., Angelidis, A. S., & Koutsoumanis, K. (2008). A Study on the Kinetic Behavior of Listeria monocytogenes in Ice Cream Stored Under Static and Dynamic Chilling and Freezing Conditions. Journal of dairy science, 91(2), 523-530. doi: 10.3168/jds.2007-0255 Gougouli, M., Kalantzi, K., Beletsiotis, E., & Koutsoumanis, K. P. (2011). Development and application of predictive models for fungal growth as tools to improve quality control in yogurt production. Food Microbiology, 28(8), 1453-1462. Grande, M. J., Lucas, R., Abriouel, H., Omar, N. B., Maqueda, M., Martínez-Bueno, M., . . . Gálvez, A. (2005). Control of Alicyclobacillus acidoterrestris in fruit juices by enterocin AS-48. International journal of food microbiology, 104(3), 289-297. Griffiths, K., & Setlow, P. (2009). Effects of modification of membrane lipid composition on Bacillus subtilis sporulation and spore properties. Journal of Applied Microbiology, 106(6), 2064-2078. Groenewald, W. H., Gouws, P. A., & Witthuhn, R. C. (2009). Isolation, identification and typification of Alicyclobacillus acidoterrestris and Alicyclobacillus acidocaldarius strains from orchard soil and the fruit processing environment in South Africa. Food Microbiology, 26(1), 71-76. Guillier, L., & Augustin, J.-C. (2006). Modelling the individual cell lag time distributions of Listeria monocytogenes as a function of the physiological state and the growth conditions. International journal of food microbiology, 111(3), 241-251. Guillier, L., Pardon, P., & Augustin, J.-C. (2005). Influence of stress on individual lag time distributions of Listeria monocytogenes. Applied and Environmental Microbiology, 71(6), 2940-2948. Guillier, L., Pardon, P., & Augustin, J.-C. (2006). Automated image analysis of bacterial colony growth as a tool to study individual lag time distributions of immobilized cells. Journal of microbiological methods, 65(2), 324-334.

203

Hassan, A., Anand, S., & Avadhanula, M. (2010). Microscopic observation of multispecies biofilm of various structures on whey concentration membranes. Journal of dairy science, 93(6), 2321-2329. Heinen, W., Lauwers, A., & Mulders, J. (1982). Bacillus flavothermus, a newly isolated facultative thermophile. Antonie van Leeuwenhoek, 48(3), 265-272. Hellberg, R. S., & Chu, E. (2015). Effects of climate change on the persistence and dispersal of foodborne bacterial pathogens in the outdoor environment: A review. Critical reviews in microbiology, 1-25. Henczka, M., Djas, M., & Filipek, K. (2013). Optimisation of a direct plating method for the detection and enumeration of Alicyclobacillus acidoterrestris spores. Journal of microbiological methods, 92(1), 1-8. Henriques, A. O., & Moran, J., Charles P. (2007). Structure, assembly, and function of the spore surface layers. Annu. Rev. Microbiol., 61, 555-588. Heyndrickx, M. (2011). The importance of endospore-forming bacteria originating from soil for contamination of industrial food processing. Applied and Environmental Soil Science, 2011. Heyndrickx, M., Coorevits, A., Scheldeman, P., Lebbe, L., Schumann, P., Rodríguez-Diaz, M., Forsyth, G., Dinsdale, A., Heyrman, J., & Logan, N.A. (2012). Emended descriptions of Bacillus sporothermodurans and Bacillus oleronius with the inclusion of dairy farm isolates of both species. International Journal of Systematic and Evolutionary Microbiology, 62, 307-314. Heyndrickx, M., & Scheldeman, P. (2002). Bacilli associated with spoilage in dairy products and other food. Applications and systematics of Bacillus and relatives, 64- 82. Hill, B., & Smythe, B. (2004). Thermophilic spores in milk powder destined for use in UHT and retort sterilization processes. Paper presented at the Proceedings of New Zealand Microbiological Society Conference. Palmerston North, New Zealand: New Zealand Microbiological Society. Hill, B. M., & Smythe, B. W. (2012). Endospores of thermophilic bacteria in ingredient milk powders and their significance to the manufacture of sterilized milk products: an industrial perspective. Food Reviews International, 28(3), 299-312. Hippchen, B., Röll, A., & Poralla, K. (1981). Occurrence in soil of thermo-acidophilic bacilli possessing ω-cyclohexane fatty acids and hopanoids. Archives of Microbiology, 129(1), 53-55.

204

Holt, S., & Leadbetter, E. (1969). Comparative ultrastructure of selected aerobic spore- forming bacteria: a freeze-etching study. Bacteriological reviews, 33(2), 346. Hornstra, L. M., Ter Beek, A., Smelt, J. P., Kallemeijn, W. W., & Brul, S. (2009). On the origin of heterogeneity in (preservation) resistance of Bacillus spores: input for a ‘systems’ analysis approach of bacterial spore outgrowth. International journal of food microbiology, 134(1), 9-15. Hsiao, C.-P., & Siebert, K.J. (1999). Modeling the inhibitory effects of organic acids on bacteria. International journal of food microbiology, 47, 189-201. Huang, X.-C., Yuan, Y.-H., Guo, C.-F., Gekas, V., & Yue, T.-L. (2015). Alicyclobacillus in the Fruit Juice Industry: Spoilage, Detection, and Prevention/Control. Food Reviews International, 31(2), 91-124. Huertas, J.-P., Esteban, M.-D., Antolinos, V., & Palop, A. (2014). Combined effect of natural antimicrobials and thermal treatments on Alicyclobacillus acidoterrestris spores. Food control, 35(1), 73-78. Iciek, J., Papiewska, A., & Molska, M. (2006). Inactivation of Bacillus stearothermophilus spores during thermal processing. Journal of Food Engineering, 77(3), 406-410. IFU (International Federation of Fruit Juice Producers) (2007). Method on the Detection of Taint Producing Alicyclobacillus in Fruit Juices. IFU Method No. 12. IFU, Paris, 1- 11. Ikegami Y, Enda M, Matsui C, & Nakagawa K. (1996). Barenshia orenji jusu chu ni okeru kousansei kin no zoushoku. Toyo Shokuhin Kogyo College/Toyo Shokuhin Laboratories, Kenkyu Hokokusho, 21, 117-122 (in Japanese). IPCC (2013). Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. doi:10.1017/CBO9781107415324.022 IPCC (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Jacxsens, L., Luning, P., Van der Vorst, J., Devlieghere, F., Leemans, R., & Uyttendaele, M. (2010). Simulation modelling and risk assessment as tools to identify the impact of climate change on microbiological food safety–The case study of fresh produce supply chain. Food Research International, 43(7), 1925-1935.

205

Jay, J. M. (2012). Modern food microbiology: Springer Science & Business Media. Jensen, N. (2000). Alicyclobacillus in Australia [Paper based on a presentation at the 10th World Congress of Food Science and Technology (1999: Sydney)]. Food Australia, 52(7), 282. Jensen, N., & Whitfield, F. (2003). Role of Alicyclobacillus acidoterrestris in the development of a disinfectant taint in shelf‐stable fruit juice. Letters in applied microbiology, 36(1), 9-14. Jiang, C.-Y., Liu, Y., Liu, Y.-Y., You, X.-Y., Guo, X., & Liu, S.-J. (2008). Alicyclobacillus ferrooxydans sp. nov., a ferrous-oxidizing bacterium from solfataric soil. International Journal of Systematic and Evolutionary Microbiology, 58(12), 2898- 2903. doi:10.1099/ijs.0.2008/000562-0 Jiangbo, Z., Xinghua, L., Tianli, Y., & Yahong, Y. (2016). The Effect of Nisin on the Growth and Heat Resistance of Alicyclobacillus acidoterrestris, Alicyclobacillus herbarius and Alicyclobacillus contaminans in Kiwi Juice. Journal of Food Processing and Preservation. Jiao, L., Ran, J., Xu, X., & Wang, J. (2015). Heat, acid and cold stresses enhance the expression of DnaK gene in Alicyclobacillus acidoterrestris. Food Research International, 67, 183-192. Jones, C., Padula, N., & Setlow, P. (2005). Effect of mechanical abrasion on the viability, disruption and germination of spores of Bacillus subtilis. Journal of Applied Microbiology, 99(6), 1484-1494. Jurado, A.S., Santana, A.C., Da Costa, M.S., & Madeira, V.M. (1987). Influence of divalent cations on the growth and morphology of Bacillus stearothermophilus. Microbiology, 133, 507-513. Kakagianni, M., Aguirre, J.S., Lianou, A., & Koutsoumanis, K.P. (2017). Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores. Food Microbiology, 67, 76-84. Kakagianni, M., Gougouli, M., & Koutsoumanis, K. P. (2016). Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk. Food Microbiology, 57, 28-35. doi: http://dx.doi.org/10.1016/j.fm.2016.01.001 Kakagianni, M., Kalantzi, K., Beletsiotis, E., Ghikas, D., Lianou, A., & Koutsoumanis, K.P. (2018). Development and validation of predictive models for the effect of storage

206

temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris ATCC 49025 in fruit drinks. Food Microbiology, 74, 40-49. Kalogridou-Vassiliadou, D. (1992). Biochemical activities of Bacillus species isolated from flat sour evaporated milk. Journal of dairy science, 75(10), 2681-2686. Keim, P., Van Ert, M. N., Pearson, T., Vogler, A. J., Huynh, L. Y., & Wagner, D. M. (2004). Anthrax molecular epidemiology and forensics: using the appropriate marker for different evolutionary scales. Infection, Genetics and Evolution, 4(3), 205-213. Khanal, S. N., Anand, S., & Muthukumarappan, K. (2014). Evaluation of high-intensity ultrasonication for the inactivation of endospores of 3 bacilli species in nonfat milk. Journal of dairy science, 97(10), 5952-5963. Khaneja, R., Perez‐Fons, L., Fakhry, S., Baccigalupi, L., Steiger, S., To, E., . . . Fraser, P. (2010). Carotenoids found in Bacillus. Journal of Applied Microbiology, 108(6), 1889-1902. Kleter, G. A., & Marvin, H. J. P. (2009). Indicators of emerging hazards and risks to food safety. Food and chemical toxicology, 47(5), 1022-1039. doi: https://doi.org/10.1016/j.fct.2008.07.028 Klobutcher, L. A., Ragkousi, K., & Setlow, P. (2006). The Bacillus subtilis spore coat provides “eat resistance” during phagocytic predation by the protozoan Tetrahymena thermophila. Proceedings of the National Academy of Sciences of the United States of America, 103(1), 165-170. Knight, G., Nicol, R., & McMeekin, T. (2004). Temperature step changes: a novel approach to control biofilms of Streptococcus thermophilus in a pilot plant-scale cheese-milk pasteurisation plant. International journal of food microbiology, 93(3), 305-318. Knight, G., & Weeks, M. (2008). Conditions for inactivation of thermophilic spores in NaOH solutions for reuse applications. Australian Journal of Dairy Technology, 63(3), 82. Komitopoulou, E., Boziaris, I. S., Davies, E. A., Delves‐Broughton, J., & Adams, M. R. (1999). Alicyclobacillus acidoterrestris in fruit juices and its control by nisin. International journal of food science & technology, 34(1), 81-85. Kong, L., Zhang, P., Setlow, P., & Li, Y.-q. (2010). Characterization of bacterial spore germination using integrated phase contrast microscopy, Raman spectroscopy, and optical tweezers. Analytical chemistry, 82(9), 3840-3847.

207

Koseki, S., & Isobe, S. (2005). Prediction of pathogen growth on iceberg lettuce under real temperature history during distribution from farm to table. International journal of food microbiology, 104(3), 239-248. Koutsoumanis, K. (2008). A study on the variability in the growth limits of individual cells and its effect on the behavior of microbial populations. International journal of food microbiology, 128(1), 116-121. doi: https://doi.org/10.1016/j.ijfoodmicro.2008.07.013 Koutsoumanis K., & Angelidis. A.S. (2007). Probabilistic modeling approach for evaluating the compliance of ready-to-eat foods with new European Union safety criteria for Listeria monocytogenes. Applied and Environmental Microbiology, 73, 4996-5004. Koutsoumanis, K. P., Gougouli, M. (2015). Use of Time Temperature Integrators in food safety management. Trends in Food Science & Technology, 43(2), 236-244. doi: https://doi.org/10.1016/j.tifs.2015.02.008 Koutsoumanis, K. P., & Lianou, A. (2013). Stochasticity in colonial growth dynamics of individual bacterial cells. Applied and Environmental Microbiology, 79(7), 2294- 2301. Koutsoumanis, K. P., Lianou, A., & Gougouli, M. (2016). Latest developments in foodborne pathogens modeling. Current Opinion in Food Science, 8(Supplement C), 89-98. doi: https://doi.org/10.1016/j.cofs.2016.04.006 Koutsoumanis, K., Pavlis, A., Nychas, G.-J. E., & Xanthiakos, K. (2010). Probabilistic model for Listeria monocytogenes growth during distribution, retail storage, and domestic storage of pasteurized milk. Applied and Environmental Microbiology, 76(7), 2181- 2191. Koutsoumanis, K., Stamatiou, A., Skandamis, P., & Nychas, G.-J. (2006). Development of a microbial model for the combined effect of temperature and pH on spoilage of ground meat, and validation of the model under dynamic temperature conditions. Applied and Environmental Microbiology, 72(1), 124-134. Koutsoumanis, K., Taoukis, P., & Nychas, G. (2005). Development of a safety monitoring and assurance system for chilled food products. International Journal of Food Microbiology, 100(1), 253-260. Kuisiene, N., Raugalas, J., & Chitavichius, D. (2009). Phylogenetic, inter, and intraspecific sequence analysis of spo0A gene of the genus Geobacillus. Current Microbiology, 58(6), 547-553.

208

Kuisiene, N., Raugalas, J., Stuknyte, M., & Chitavichius, D. (2007). Identification of the genus Geobacillus using genus-specific primers, based on the 16S–23S rRNA gene internal transcribed spacer. FEMS microbiology letters, 277(2), 165-172. Lücking, G., Stoeckel, M., Atamer, Z., Hinrichs, J., & Ehling-Schulz, M. (2013). Characterization of aerobic spore-forming bacteria associated with industrial dairy processing environments and product spoilage. International journal of food microbiology, 166(2), 270-279. Laaberki, M.-H., & Dworkin, J. (2008). Role of spore coat proteins in the resistance of Bacillus subtilis spores to Caenorhabditis elegans predation. Journal of bacteriology, 190(18), 6197-6203. Lee, K. S., Bumbaca, D., Kosman, J., Setlow, P., & Jedrzejas, M. J. (2008). Structure of a protein–DNA complex essential for DNA protection in spores of Bacillus species. Proceedings of the National Academy of Sciences, 105(8), 2806-2811. Lee, S.-Y., Dougherty, R. H., & Kang, D.-H. (2002). Inhibitory effects of high pressure and heat on Alicyclobacillus acidoterrestris spores in apple juice. Applied and Environmental Microbiology, 68(8), 4158-4161. Lee, S. Y., Chang, S. S., Shin, J. H., & Kang, D. H. (2007). Membrane filtration method for enumeration and isolation of Alicyclobacillus spp. from apple juice. Letters in applied microbiology, 45(5), 540-546. Lee, S. Y., Ryu, S. R., & Kang, D. H. (2010). Treatment with chlorous acid to inhibit spores of Alicyclobacillus acidoterrestris in aqueous suspension and on apples. Letters in applied microbiology, 51(2), 164-169. Leguérinel, I., Couvert, O., & Mafart, P. (2007). Modelling the influence of the sporulation temperature upon the bacterial spore heat resistance, application to heating process calculation. International journal of food microbiology, 114(1), 100- 104. Lenz, C. A., Schnabel, J., & Vogel, R. F. (2014). The variability of times to detect growth from individual Clostridium botulinum type E endospores is differentially affected by high pressure treatments. High Pressure Research, 34(4), 412-418. Leuschner, R., & Lillford, P. (1999). Effects of temperature and heat activation on germination of individual spores of Bacillus subtilis. Letters in applied microbiology, 29(4), 228-232. Li, J., Xia, K., & Yu, C. (2013). Detection of Alicyclobacillus acidoterrestris in apple juice concentrate by enzyme-linked immunosorbent assay. Food control, 30(1), 251-254.

209

Li, Y., Odumeru, J. A., Griffiths, M., & McKellar, R. C. (2006). Effect of environmental stresses on the mean and distribution of individual cell lag times of Escherichia coli O157: H7. International journal of food microbiology, 110(3), 278-285. Lianou, A., & Koutsoumanis, K. P. (2011a). Effect of the growth environment on the strain variability of Salmonella enterica kinetic behavior. Food Microbiology, 28(4), 828-837. Lianou, A., & Koutsoumanis, K. P. (2011b). A stochastic approach for integrating strain variability in modeling Salmonella enterica growth as a function of pH and water activity. International journal of food microbiology, 149(3), 254-261. Lianou, A., & Koutsoumanis, K. P. (2013). Strain variability of the behavior of foodborne bacterial pathogens: A review. International journal of food microbiology, 167(3), 310-321. doi: https://doi.org/10.1016/j.ijfoodmicro.2013.09.016 Lindsay, D., Brözel, V., Mostert, J., & Von Holy, A. (2000). Physiology of dairy-associated Bacillus spp. over a wide pH range. International journal of food microbiology, 54, 49-62. Ljunger, C. (1970). On the nature of the heat resistance of thermophilic bacteria. Physiologia plantarum, 23, 351-364. Llaudes, M. K., Zhao, L., Duffy, S., & Schaffner, D. W. (2001). Simulation and modelling of the effect of small inoculum size on time to spoilage by Bacillus stearothermophilus. Food Microbiology, 18(4), 395-405. Luo, H., Yousef, A., & Wang, H. (2004). A real‐time polymerase chain reaction‐based method for rapid and specific detection of spoilage Alicyclobacillus spp. in apple juice. Letters in applied microbiology, 39, 376-382. Lynn, D. J., Singer, G. A., & Hickey, D. A. (2002). Synonymous codon usage is subject to selection in thermophilic bacteria. Nucleic acids research, 30(19), 4272-4277. Majeed, M., Majeed, S., Nagabhushanam, K., Natarajan, S., Sivakumar, A., & Ali, F. (2016). Evaluation of the stability of Bacillus coagulans MTCC 5856 during processing and storage of functional foods. International journal of food science & technology. Maldonado, M. C., Belfiore, C., & Navarro, A. R. (2008). Temperature, soluble solids and pH effect on Alicyclobacillus acidoterrestris viability in lemon juice concentrate. Journal of industrial microbiology & biotechnology, 35(2), 141-144.

210

Manachini, P. L., Mora, D., Nicastro, G., Parini, C., Stackebrandt, E., Pukall, R., & Fortina, M. (2000). Bacillus thermodenitrificans sp. nov., nom. rev. International Journal of Systematic and Evolutionary Microbiology, 50(3), 1331-1337. Mandic-Mulec, I., & Prosser, J. I. (2011). Diversity of endospore-forming bacteria in soil: characterization and driving mechanisms Endospore-forming Soil Bacteria (pp. 31- 59): Springer. Marchand, S., De Block, J., De Jonghe, V., Coorevits, A., Heyndrickx, M., & Herman, L. (2012). Biofilm formation in milk production and processing environments; influence on milk quality and safety. Comprehensive Reviews in Food Science and Food Safety, 11(2), 133-147. Marquis, R., Sim, J., & Shin, S. (1994). Molecular mechanisms of resistance to heat and oxidative damage. Journal of Applied Bacteriology, 76(S23). Martins, L. O., Jurado, A. S., & Madeira, V. M. (1990). Composition of polar lipid acyl chains of Bacillus stearothermophilus as affected by temperature and calcium. Biochimica et Biophysica Acta (BBA)-Lipids and Lipid Metabolism, 1045(1), 17-20. Martiny, J. B. H., Bohannan, B. J., Brown, J. H., Colwell, R. K., Fuhrman, J. A., Green, J. L., . . . Kuske, C. R. (2006). Microbial biogeography: putting microorganisms on the map. Nature Reviews Microbiology, 4(2), 102-112. Matsubara, H., Goto, K., Matsumura, T., Mochida, K., Iwaki, M., Niwa, M., & Yamasato, K. (2002). Alicyclobacillus acidiphilus sp. nov., a novel thermo-acidophilic, omega- alicyclic fatty acid-containing bacterium isolated from acidic beverages. International Journal of Systematic and Evolutionary Microbiology, 52(5), 1681- 1685. Matsuda, N., Masuda, H., Komaki, M., & Matsumoto, N. (1982). Thermophilic, spore- forming, strict anaerobes isolated from spoiled canned" Shiruko" and coffee containing milk. J. FOOD HYG. SOC. JAP., 23(6), 480-486. McGuiggan, J., McCleery, D. R., Hannan, A., & Gilmour, A. (2002). Aerobic spore‐forming bacteria in bulk raw milk: factors influencing the numbers of psychrotrophic, mesophilic and thermophilic Bacillus spores. International Journal of Dairy Technology, 55(2), 100-107. McKellar, R., & Lu, X. (2005). Development of a global stochastic model relating the distribution of individual cell and population physiological states. International journal of food microbiology, 100(1), 33-40.

211

McKellar, R. C., & Hawke, A. (2006). Assessment of distributions for fitting lag times of individual cells in bacterial populations. International journal of food microbiology, 106(2), 169-175. McMeekin, T., Brown, J., Krist, K., Miles, D., Neumeyer, K., Nichols, D., . . . Ross, T. (1997). Quantitative microbiology: a basis for food safety. Emerging infectious diseases, 3(4), 541. McMeekin, T., Olley, J., Ratkowsky, D., Corkrey, R., & Ross, T. (2013). Predictive microbiology theory and application: Is it all about rates? Food control, 29(2), 290- 299. doi: https://doi.org/10.1016/j.foodcont.2012.06.001 McMichael, A. J. (2013). Globalization, climate change, and human health. New England Journal of Medicine, 368(14), 1335-1343. McNamara, C. J., Wiebe, D., & Gomez, M. (2011). Recovery of Alicyclobacillus from Inhibitory Fruit Juice Concentrates. Journal of Food Protection, 74(8), 1370-1373. doi: 10.4315/0362-028x.jfp-11-027 Medina-Martínez, M. S., Allende, A., Barberá, G. G., & Gil, M. I. (2015). Climatic variations influence the dynamic of epiphyte bacteria of baby lettuce. Food Research International, 68, 54-61. Meintanis, C., Chalkou, K., Kormas, K. A., Lymperopoulou, D., Katsifas, E., Hatzinikolaou, D., & Karagouni, A. (2008). Application of rpoB sequence similarity analysis, REP‐PCR and BOX‐PCR for the differentiation of species within the genus Geobacillus. Letters in applied microbiology, 46(3), 395-401. Melly, E., Genest, P., Gilmore, M., Little, S., Popham, D., Driks, A., & Setlow, P. (2002). Analysis of the properties of spores of Bacillus subtilis prepared at different temperatures. Journal of Applied Microbiology, 92(6), 1105-1115. Membré, J.-M., Leporq, B., Vialette, M., Mettler, E., Perrier, L., Thuault, D., & Zwietering, M. (2005). Temperature effect on bacterial growth rate: quantitative microbiology approach including cardinal values and variability estimates to perform growth simulations on/in food. International journal of food microbiology, 100(1), 179-186. doi: https://doi.org/10.1016/j.ijfoodmicro.2004.10.015 Membré, J.-M., & Van Zuijlen, A. (2011). A probabilistic approach to determine thermal process setting parameters: application for commercial sterility of products. International journal of food microbiology, 144(3), 413-420. Merle, J. M. (2012). Growth behavior of Alicyclobacillus acidoterrestris as a function of strain and culture conditions. Rutgers University-Graduate School-New Brunswick.

212

Métris, A., George, S., Mackey, B., & Baranyi, J. (2008). Modeling the variability of single- cell lag times for Listeria innocua populations after sublethal and lethal heat treatments. Applied and Environmental Microbiology, 74(22), 6949-6955. Métris, A., George, S. M., Peck, M. W., & Baranyi, J. (2003). Distribution of turbidity detection times produced by single cell-generated bacterial populations. Journal of microbiological methods, 55(3), 821-827. Miled, R. B., Guillier, L., Neves, S., Augustin, J.-C., Colin, P., & Besse, N. G. (2011). Individual cell lag time distributions of Cronobacter (Enterobacter sakazakii) and impact of pooling samples on its detection in powdered infant formula. Food Microbiology, 28(4), 648-655. Miraglia, M., Marvin, H., Kleter, G., Battilani, P., Brera, C., Coni, E., . . . Dekkers, S. (2009). Climate change and food safety: an emerging issue with special focus on Europe. Food and chemical toxicology, 47(5), 1009-1021. Mistry, V. V. (2001). Fermented milks and cream. FOOD SCIENCE AND TECHNOLOGY- NEW YORK-MARCEL DEKKER-, 301-326. Moir, A. (2006). How do spores germinate? Journal of Applied Microbiology, 101(3), 526- 530. Moir, A., Corfe, B., & Behravan, J. (2002). Spore germination. Cellular and Molecular Life Sciences CMLS, 59(3), 403-409. Morice, C. P., Kennedy, J. J., Rayner, N. A., & Jones, P. D. (2012). Quantifying uncertainties in global and regional temperature change using an ensemble of observational estimates: The HadCRUT4 data set. Journal of Geophysical Research: Atmospheres, 117(D8), n/a-n/a. doi: 10.1029/2011JD017187 Mtimet, N., Guégan, S., Durand, L., Mathot, A.-G., Venaille, L., Leguérinel, I., . . . Couvert, O. (2016). Effect of pH on Thermoanaerobacterium thermosaccharolyticum DSM 571 growth, spore heat resistance and recovery. Food Microbiology, 55, 64-72. Mtimet, N., Trunet, C., Mathot, A.-G., Venaille, L., Leguérinel, I., Coroller, L., & Couvert, O. (2015). Modeling the behavior of Geobacillus stearothermophilus ATCC 12980 throughout its life cycle as vegetative cells or spores using growth boundaries. Food Microbiology, 48, 153-162. Murakami, M., Tedzuka, H., & Yamazaki, K. (1998). Thermal resistance of Alicyclobacillus acidoterrestris spores in different buffers and pH. Food Microbiology, 15(6), 577- 582.

213

Murphy, P. M., Lynch, D., & Kelly, P. M. (1999). Growth of thermophilic spore forming bacilli in milk during the manufacture of low heat powders. International Journal of Dairy Technology, 52(2), 45-50. Murray, M. B., Gurtler, J. B., Ryu, J.-H., Harrison, M. A., & Beuchat, L. R. (2007). Evaluation of direct plating methods to enumerate Alicyclobacillus in beverages. International journal of food microbiology, 115(1), 59-69. Murrell, W. (1967). The biochemistry of the bacterial endospore. Advances in microbial physiology, 1, 133-251. Nakamura, L. K., Blumenstock, I., & Claus, D. (1988). Taxonomic Study of Bacillus coagulans Hammer 1915 with a Proposal for Bacillus smithii sp. nov. International Journal of Systematic and Evolutionary Microbiology, 38(1), 63-73. Nakano, C., Takahashi, N., Tanaka, N., & Okada, S. (2015). Alicyclobacillus dauci sp. nov., a slightly thermophilic, acidophilic bacterium isolated from a spoiled mixed vegetable and fruit juice product. International Journal of Systematic and Evolutionary Microbiology, 65(2), 716-722. Nauta, M. J. (2002). Modelling bacterial growth in quantitative microbiological risk assessment: is it possible? International journal of food microbiology, 73(2), 297- 304. doi: https://doi.org/10.1016/S0168-1605(01)00664-X Nauta, M. J., Litman, S., Barker, G. C., & Carlin, F. (2003). A retail and consumer phase model for exposure assessment of Bacillus cereus. International journal of food microbiology, 83(2), 205-218. doi: https://doi.org/10.1016/S0168-1605(02)00374-4 Nazina, T., Tourova, T., Poltaraus, A., Novikova, E., Grigoryan, A., Ivanova, A., . . . Belyaev, S. (2001). Taxonomic study of aerobic thermophilic bacilli: descriptions of Geobacillus subterraneus gen. nov., sp. nov. and Geobacillus uzenensis sp. nov. from petroleum reservoirs and transfer of Bacillus stearothermophilus, Bacillus thermocatenulatus, Bacillus thermoleovorans, Bacillus kaustophilus, Bacillus thermodenitrificans to Geobacillus as the new combinations G. stearothermophilus, G. th. International Journal of Systematic and Evolutionary Microbiology, 51(2), 433- 446. Ng, T. M., & Schaffner, D. W. (1997). Mathematical Models for the Effects of pH, Temperature, and Sodium Chloride on the Growth of Bacillus stearothermophilus in Salty Carrots. Applied and Environmental Microbiology, 63(4), 1237-1243.

214

Ng, T., Viard, E., Caipo, M., Duffy, S., & Schaffner, D. (2002). Expansion and validation of a predictive model for the growth of Bacillus stearothermophilus in military rations. Journal of food science, 67(5), 1872-1878. Nicolaus, B., Improta, R., Manca, M. C., Lama, L., Esposito, E., & Gambacorta, A. (1998). Alicyclobacilli from an unexplored geothermal soil in Antarctica: Mount Rittmann. Polar Biology, 19(2), 133-141. Nychas, G.-J. E., Panagou, E. Z., & Mohareb, F. (2016). Novel approaches for food safety management and communication. Current Opinion in Food Science, 12, 13-20. Nychas, G.-J. E., Skandamis, P. N., Tassou, C. C., & Koutsoumanis, K. P. (2008). Meat spoilage during distribution. Meat Science, 78(1), 77-89. doi: https://doi.org/10.1016/j.meatsci.2007.06.020 NZFSA (New Zealand Food Safety Authority) (2016). A guide to calculating shelf life in foods. Available online at http://www.foodsafety.govt.nz/elibrary/industry/determine-shelf-life-of-food/.

Olivier, S., Bull, M., Stone, G., Van Diepenbeek, R., Kormelink, F., Jacops, L., & Chapman, B. (2011). Strong and consistently synergistic inactivation of spores of spoilage- associated Bacillus and Geobacillus spp. by high pressure and heat compared with inactivation by heat alone. Applied and Environmental Microbiology, 77(7), 2317- 2324. Oomes, S., Van Zuijlen, A., Hehenkamp, J., Witsenboer, H., Van der Vossen, J., & Brul, S. (2007). The characterisation of Bacillus spores occurring in the manufacturing of (low acid) canned products. International journal of food microbiology, 120(1), 85- 94. Orr, R. V., Shewfelt, R. L., Huang, C., Tefera, S., & Beuchat, L. R. (2000). Detection of guaiacol produced by Alicyclobacillus acidoterrestris in apple juice by sensory and chromatographic analyses, and comparison with spore and vegetative cell populations. Journal of Food Protection®, 63(11), 1517-1522. Osopale, B. A., Witthuhn, C. R., Albertyn, J., & Oguntoyinbo, F. A. (2016). Culture dependent and independent genomic identification of Alicyclobacillus species in contaminated commercial fruit juices. Food Microbiology, 56, 21-28. Østergaard, N. B., Eklöw, A., & Dalgaard, P. (2014). Modelling the effect of lactic acid bacteria from starter-and aroma culture on growth of Listeria monocytogenes in cottage cheese. International journal of food microbiology, 188, 15-25.

215

Pachauri, R. K., Allen, M. R., Barros, V. R., Broome, J., Cramer, W., Christ, R., . . . Dasgupta, P. (2014). Climate change 2014: synthesis report. Contribution of Working Groups I, II and III to the fifth assessment report of the Intergovernmental Panel on Climate Change: IPCC. Palop, A., Álvarez, I., Raso, J., & Condon, S. (2000). Heat Resistance of Alicyclobacillus acidocaldarius in Water, Various Buffers, and Orange Juice. Journal of Food Protection, 63(10), 1377-1380. doi: 10.4315/0362-028x-63.10.1377 Palop, A., Manas, P., & Conton, S. (1999). Sporulation temperature and heat resistance of Bacillus spores: a review. Journal of Food Safety, 19(1), 57-72. Palop, A., Raso, J., Pagán, R., Condón, S., & Sala, F. J. (1999). Influence of pH on heat resistance of spores of Bacillus coagulans in buffer and homogenized foods. International journal of food microbiology, 46(3), 243-249. Pandey, R., Ter Beek, A., Vischer, N. O., Smelt, J. P., Brul, S., & Manders, E. M. (2013). Live cell imaging of germination and outgrowth of individual Bacillus subtilis spores; the effect of heat stress quantitatively analyzed with SporeTracker. PloS one, 8(3), e58972. Pandey, R., Ter Beek, A., Vischer, N. O., Smelt, J. P., Kemperman, R., Manders, E. M., & Brul, S. (2015). Quantitative analysis of the effect of specific tea compounds on germination and outgrowth of Bacillus subtilis spores at single cell resolution. Food Microbiology, 45, 63-70. Parkar, S., Flint, S., & Brooks, J. (2003). Physiology of biofilms of thermophilic bacilli— potential consequences for cleaning. Journal of Industrial Microbiology and Biotechnology, 30, 553-560. Parkar, S., Flint, S., Palmer, J., & Brooks, J. (2001). Factors influencing attachment of thermophilic bacilli to stainless steel. Journal of Applied Microbiology, 90(6), 901- 908. Payot, T., Chemaly, Z., & Fick, M. (1999). Lactic acid production by Bacillus coagulans— kinetic studies and optimization of culture medium for batch and continuous fermentations. Enzyme and Microbial Technology, 24(3), 191-199. Pei, J., Yue, T., & Yuan, Y. (2014). Control of Alicyclobacillus acidoterrestris in fruit juices by a newly discovered bacteriocin. World Journal of Microbiology and Biotechnology, 30(3), 855-863. Pena, W., De Massaguer, P., Zuniga, A., & Saraiva, S. (2011). Modeling the growth limit of Alicyclobacillus acidoterrestris CRA7152 in apple juice: effect of pH, Brix,

216

temperature and nisin concentration. Journal of Food Processing and Preservation, 35(4), 509-517. Peña, W., & Rodríguez, P. (2010). Modeling the growth/no growth interface of Alicyclobacillus acidoterrestris CRA 7152 in orange juice as a function of pH, temperature, Brix and nisin concentration. Scientia Agropecuaria, 1(1), 47-61. Peña, W. E. L., & de Massaguer, P. R. (2006). Microbial modeling of Alicyclobacillus acidoterrestris CRA 7152 growth in orange juice with nisin added. Journal of Food Protection®, 69(8), 1904-1912. Peña, W. E. L., Massaguer, P. R. d., & Teixeira, L. Q. (2009). Microbial modeling of thermal resistance of Alicyclobacillus acidoterrestris CRA7152 spores in concentrated orange juice with nisin addition. Brazilian Journal of Microbiology, 40(3), 601-611. Peng, L., Chen, D., Setlow, P., & Li, Y.-q. (2009). Elastic and inelastic light scattering from single bacterial spores in an optical trap allows the monitoring of spore germination dynamics. Analytical chemistry, 81(10), 4035-4042. Pepe, O., Blaiotta, G., Moschetti, G., Greco, T., & Villani, F. (2003). Rope-producing strains of Bacillus spp. from wheat bread and strategy for their control by lactic acid bacteria. Applied and Environmental Microbiology, 69(4), 2321-2329. Pettersson, B., Lembke, F., Hammer, P., Stackebrandt, E., & Priest, F. G. (1996). Bacillus sporothermodurans, a new species producing highly heat-resistant endospores. International Journal of Systematic and Evolutionary Microbiology, 46(3), 759-764. Pettipher, G., & Osmundson, M. (2000). Methods for the detection, enumeration and identification of Alicyclobacillus acidoterrestris: Alicyclobacillus in the food industry. Food Australia, 52(7), 293-294. Pettipher, G., Osmundson, M., & Murphy, J. (1997). Methods for the detection and enumeration of Alicyclobacillus acidoterrestris and investigation of growth and production of taint in fruit juice and fruit juice‐containing drinks. Letters in applied microbiology, 24(3), 185-189. Piggot, P., & Coote, J. (1976). Genetic aspects of bacterial endospore formation. Bacteriological reviews, 40(4), 908. Piggot, P. J., & Hilbert, D. W. (2004). Sporulation of Bacillus subtilis. Current opinion in microbiology, 7(6), 579-586.

217

Pikuta, E., Lysenko, A., Chuvilskaya, N., Mendrock, U., Hippe, H., Suzina, N., . . . Laurinavichius, K. (2000). Anoxybacillus pushchinensis gen. nov., sp. nov., a novel anaerobic, alkaliphilic, moderately thermophilic bacterium from manure, and description of Anoxybacillus flavitherms comb. nov. International Journal of Systematic and Evolutionary Microbiology, 50(6), 2109-2117. Pin, C., & Baranyi, J. (2006). Kinetics of single cells: observation and modeling of a stochastic process. Applied and Environmental Microbiology, 72(3), 2163-2169. Pin, C., Garcıa de Fernando, G. D., Ordóñez, J. A., & Baranyi, J. (2002). Analysing the lag– growth rate relationship of Yersinia enterocolitica. International journal of food microbiology, 73(2), 197-201. doi: https://doi.org/10.1016/S0168-1605(01)00649-3 Poralla, K., Kannenberg, E., & Blume, A. (1980). A glycolipid containing hopane isolated from the acidophilic, thermophilic Bacillus acidocaldarius, has a cholesterol-like function in membranes. FEBS letters, 113(1), 107-110. Porter, , J.R., L. Xie, A.J. Challinor, K. Cochrane, S.M. Howden, M.M. Iqbal, D.B. Lobell, & M.I.T. (2014). Food security and food production systems, in: Field, C.B., V.R. Barros, D.J. Dokken, K.J. Mach, M.D.M., T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S.M., P.R. Mastrandrea, and L.L.W. (Eds.), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Postollec, F., Bonilla, S., Baron, F., Jan, S., Gautier, M., Mathot, A. G., . . . Sohier, D. (2010). A multiparametric PCR-based tool for fast detection and identification of spore-forming bacteria in food. International journal of food microbiology, 142(1), 78-88. Postollec, F., Mathot, A.-G., Bernard, M., Divanac'h, M.-L., Pavan, S., & Sohier, D. (2012). Tracking spore-forming bacteria in food: from natural biodiversity to selection by processes. International journal of food microbiology, 158(1), 1-8. Pouillot, R., Albert, I., Cornu, M., & Denis, J.-B. (2003). Estimation of uncertainty and variability in bacterial growth using Bayesian inference. Application to Listeria monocytogenes. International journal of food microbiology, 81(2), 87-104. doi: https://doi.org/10.1016/S0168-1605(02)00192-7 Pouillot, R., & Lubran, M. B. (2011). Predictive microbiology models vs. modeling microbial growth within Listeria monocytogenes risk assessment: What parameters

218

matter and why. Food Microbiology, 28(4), 720-726. doi: https://doi.org/10.1016/j.fm.2010.06.002 Priest, F. G. (1993). Systematics and ecology of Bacillus subtilis and other Gram-positive bacteria (pp. 3-16): American Society of Microbiology. Querol, E., Perez-Pons, J. A., & Mozo-Villarias, A. (1996). Analysis of protein conformational characteristics related to thermostability. Protein Engineering, 9(3), 265-271. Räisänen, J., Hansson, U., Ullerstig, A., Döscher, R., Graham, L. P., Jones, C., . . . Willén, U. (2004). European climate in the late twenty-first century: regional simulations with two driving global models and two forcing scenarios. Climate Dynamics, 22(1), 13- 31. doi: 10.1007/s00382-003-0365-x Ratkowsky, D. A., McKellar, R., & Lu, X. (2004). Model fitting and uncertainty. In R. C. McKellar, Lu, X. (Ed.), Modeling microbial responses in food (pp. 151-196). Boca Raton, Florida: CRC Press. Ratkowsky, D.A., & Ross, T. (1995). Modelling the bacterial growth/no growth interface. Letters in Applied Microbiology, 20, 29-33. Rückert, A., Ronimus, R. S., & Morgan, H. W. (2004). A RAPD-based survey of thermophilic bacilli in milk powders from different countries. International journal of food microbiology, 96(3), 263-272. Refstrup, E. (2000). Evaporation and drying technology developments. International Journal of Dairy Technology, 53(4), 163-167. Reginensi, S. M., Gonzalez, M. J., Olivera, J. A., Sosa, M., Juliano, P., & Bermudez, J. (2011). RAPD-based screening for spore-forming bacterial populations in Uruguayan commercial powdered milk. International journal of food microbiology, 148(1), 36- 41. Reyes, J. F., & Tanaka, M. M. (2010). Mutation rates of spoligotypes and variable numbers of tandem repeat loci in Mycobacterium tuberculosis. Infection, Genetics and Evolution, 10(7), 1046-1051. Rigaux Ancelet, C. S., Carlin, F., Nguyen‐thé, C., & Albert, I. (2013). Inferring an augmented Bayesian network to confront a complex quantitative microbial risk assessment model with durability studies: application to Bacillus cereus on a courgette purée production chain. Risk Analysis, 33(5), 877-892.

219

Rigaux, C., André, S., Albert, I., & Carlin, F. (2014). Quantitative assessment of the risk of microbial spoilage in foods. Prediction of non-stability at 55°C caused by Geobacillus stearothermophilus in canned green beans. International journal of food microbiology, 171(Supplement C), 119-128. doi: https://doi.org/10.1016/j.ijfoodmicro.2013.11.014 Roberts, M. S., & Cohan, F. M. (1995). Recombination and migration rates in natural populations of Bacillus subtilis and Bacillus mojavensis. Evolution, 1081-1094. Robinson, R. K. (2005). Dairy microbiology handbook: the microbiology of milk and milk products: John Wiley & Sons. Roig-Sagues, A., Asto, E., Engers, I., & Hernández-Herrero, M. (2015). Improving the efficiency of ultra-high pressure homogenization treatments to inactivate spores of Alicyclobacillus spp. in orange juice controlling the inlet temperature. LWT-Food Science and Technology, 63(2), 866-871. Ronimus, R. S., Parker, L. E., Turner, N., Poudel, S., Rückert, A., & Morgan, H. W. (2003). A RAPD-based comparison of thermophilic bacilli from milk powders. International journal of food microbiology, 85(1), 45-61. Ronimus, R. S., Rueckert, A., & Morgan, H. W. (2006). Survival of thermophilic spore- forming bacteria in a 90+ year old milk powder from Ernest Shackelton's Cape Royds Hut in Antarctica. Journal of dairy research, 73(02), 235-243. Rosenkvist, H., & Hansen, Å. (1995). Contamination profiles and characterisation of Bacillus species in wheat bread and raw materials for bread production. International journal of food microbiology, 26(3), 353-363. Ross, A. I., Griffiths, M. W., Mittal, G. S., & Deeth, H. C. (2003). Combining nonthermal technologies to control foodborne microorganisms. International journal of food microbiology, 89(2), 125-138. Ross, T., Rasmussen, S., Fazil, A., Paoli, G., & Sumner, J. (2009). Quantitative risk assessment of Listeria monocytogenes in ready-to-eat meats in Australia. International journal of food microbiology, 131(2), 128-137. doi: https://doi.org/10.1016/j.ijfoodmicro.2009.02.007 Rosso, L., Lobry, J., Bajard, S. & Flandrois, J. (1995). Convenient model to describe the combined effects of temperature and pH on microbial growth. Applied and Environmental Microbiology, 61, 610-616.

220

Rosso, L., Lobry, J., & Flandrois, J. (1993). An unexpected correlation between cardinal temperatures of microbial growth highlighted by a new model. Journal of Theoretical Biology, 162(4), 447-463. Rueckert, A., Ronimus, R., & Morgan, H. (2005a). Rapid differentiation and enumeration of the total, viable vegetative cell and spore content of thermophilic bacilli in milk powders with reference to Anoxybacillus flavithermus. Journal of Applied Microbiology, 99(5), 1246-1255. Rueckert, A., Ronimus, R. S., & Morgan, H. W. (2005b). Development of a rapid detection and enumeration method for thermophilic bacilli in milk powders. Journal of microbiological methods, 60(2), 155-167. Rueckert, A., Ronimus, R. S., & Morgan, H. W. (2006). Development of a real-time PCR assay targeting the sporulation gene, spo0A, for the enumeration of thermophilic bacilli in milk powder. Food Microbiology, 23(3), 220-230. Russell, N., & Fukunaga, N. (1990). A comparison of thermal adaptation of membrane lipids in psychrophilic and thermophilic bacteria. FEMS microbiology letters, 75(2-3), 171-182. Sadeghi, M., Naderi-Manesh, H., Zarrabi, M., & Ranjbar, B. (2006). Effective factors in thermostability of thermophilic proteins. Biophysical chemistry, 119(3), 256-270. Sadiq, F. A., Li, Y., Liu, T., Flint, S., Zhang, G., & He, G. (2016). A RAPD based study revealing a previously unreported wide range of mesophilic and thermophilic spore formers associated with milk powders in China. International journal of food microbiology, 217, 200-208. Sakanaka, S., Juneja, L.R., & Taniguchi, M. (2000). Antimicrobial effects of green tea polyphenols on thermophilic spore-forming bacteria. Journal of Bioscience and Bioengineering, 90, 81-85. Salkinoja-Salonen, M., Vuorio, R., Andersson, M., Kämpfer, P., Andersson, M., Honkanen-Buzalski, T., & Scoging, A. (1999). Toxigenic strains of Bacillus licheniformis related to food poisoning. Applied and Environmental Microbiology, 65(10), 4637-4645. Scheldeman, P., Goossens, K., Rodriguez-Diaz, M., Pil, A., Goris, J., Herman, L., . . . Heyndrickx, M. (2004). Paenibacillus lactis sp. nov., isolated from raw and heat- treated milk. International Journal of Systematic and Evolutionary Microbiology, 54(3), 885-891.

221

Scheldeman, P., Herman, L., Foster, S., & Heyndrickx, M. (2006). Bacillus sporothermodurans and other highly heat‐resistant spore formers in milk. Journal of Applied Microbiology, 101(3), 542-555. Scheldeman, P., Pil, A., Herman, L., De Vos, P., & Heyndrickx, M. (2005). Incidence and diversity of potentially highly heat-resistant spores isolated at dairy farms. Applied and Environmental Microbiology, 71(3), 1480-1494. Schröter, D., Cramer, W., Leemans, R., Prentice, I. C., Araújo, M. B., Arnell, N. W., . . . Zierl, B. (2005). Ecosystem Service Supply and Vulnerability to Global Change in Europe. Science, 310(5752), 1333-1337. doi: 10.1126/science.1115233 Scott, S. A., Brooks, J. D., Rakonjac, J., Walker, K. M., & Flint, S. H. (2007). The formation of thermophilic spores during the manufacture of whole milk powder. International Journal of Dairy Technology, 60(2), 109-117. Seale, B., Burgess, S., Flint, S., Brooks, J., Bremer, P., & Parkar, S. (2015). Thermophilic Spore‐Forming Bacilli in the Dairy Industry. Biofilms in the Dairy Industry, 112-137. Seale, R., Bremer, P., Flint, S., & McQuillan, A. (2010). Characterization of spore surfaces from a Geobacillus sp. isolate by pH dependence of surface charge and infrared spectra. Journal of Applied Microbiology, 109(4), 1339-1348. Seale, R., Flint, S., McQuillan, A., & Bremer, P. (2008). Recovery of spores from thermophilic dairy bacilli and effects of their surface characteristics on attachment to different surfaces. Applied and Environmental Microbiology, 74(3), 731-737. Seale, R. B., Dhakal, R., Chauhan, K., Craven, H. M., Deeth, H. C., Pillidge, C. J., . . . Turner, M. S. (2012). Genotyping of present-day and historical Geobacillus species isolates from milk powders by high-resolution melt analysis of multiple variable-number tandem-repeat loci. Applied and Environmental Microbiology, 78(19), 7090-7097. Setlow, P. (1994). Mechanisms which contribute to the long‐term survival of spores of Bacillus species. Journal of Applied Bacteriology, 76(S23). Setlow, P. (2003). Spore germination. Current opinion in microbiology, 6(6), 550-556. Setlow, P. (2006). Spores of Bacillus subtilis: their resistance to and killing by radiation, heat and chemicals. Journal of Applied Microbiology, 101(3), 514-525. Setlow, P. (2007). I will survive: DNA protection in bacterial spores. Trends in microbiology, 15(4), 172-180.

222

Setlow, P. (2012). Resistance of bacterial spores to chemical agents. Russell, Hugo & Ayliffe's: Principles and Practice of Disinfection, Preservation and Sterilization, 5th Edition, 121-130. Setlow, P. (2013). Summer meeting 2013–when the sleepers wake: the germination of spores of Bacillus species. Journal of Applied Microbiology, 115(6), 1251-1268. Setlow, P. (2014). Spore resistance properties. Microbiology spectrum, 2(5). Setlow, P., Liu, J., & Faeder, J. R. (2012). Heterogeneity in bacterial spore populations. In E. Abel-Santos (Ed.), Bacterial spores: current research and applications (pp. 201- 216). Norwich, United Kingdom: Horizon Scientific Press. Sevenich, R., Kleinstueck, E., Crews, C., Anderson, W., Pye, C., Riddellova, K., . . . Knorr, D. (2014). High‐Pressure Thermal Sterilization: Food Safety and Food Quality of Baby Food Puree. Journal of food science, 79(2), M230-M237. Shaw, A. J., Hogsett, D. A., & Lynd, L. R. (2010). Natural competence in Thermoanaerobacter and Thermoanaerobacterium species. Applied and Environmental Microbiology, 76(14), 4713-4719. Shemesh, M., Pasvolsky, R., & Zakin, V. (2014). External pH is a cue for the behavioral switch that determines surface motility and biofilm formation of Alicyclobacillus acidoterrestris. Journal of Food Protection®, 77, 1418-1423. Shibeshi, K., & Farid, M. M. (2010). Pressure-assisted thermal sterilization of soup. High Pressure Research, 30(4), 530-537. Shibeshi, K., & Farid, M. M. (2011). Scale-up unit of a unique moderately high pressure unit to enhance microbial inactivation. Journal of Food Engineering, 105(3), 522- 529. Siegal-Gaskins, D., & Crosson, S. (2008). Tightly Regulated and Heritable Division Control in Single Bacterial Cells. Biophysical Journal, 95(4), 2063-2072. doi: https://doi.org/10.1529/biophysj.108.128785 Siegmund, B., & Pöllinger-Zierler, B. (2007). Growth behavior of off-flavor-forming microorganisms in apple juice. Journal of agricultural and Food Chemistry, 55(16), 6692-6699. Silva, F. M., Gibbs, P., Vieira, M. C., & Silva, C. L. (1999). Thermal inactivation of Alicyclobacillus acidoterrestris spores under different temperature, soluble solids and pH conditions for the design of fruit processes. International journal of food microbiology, 51(2), 95-103.

223

Silva, F. V. (2016). High pressure processing pretreatment enhanced the thermosonication inactivation of Alicyclobacillus acidoterrestris spores in orange juice. Food control, 62, 365-372. Silva, F. V., & Gibbs, P. (2001). Alicyclobacillus acidoterrestris spores in fruit products and design of pasteurization processes. Trends in Food Science & Technology, 12(2), 68- 74. Silva, F. V., & Gibbs, P. (2004). Target selection in designing pasteurization processes for shelf-stable high-acid fruit products. Critical reviews in food science and nutrition, 44(5), 353-360. Silva, F. V., Tan, E. K., & Farid, M. (2012). Bacterial spore inactivation at 45–65 °C using high pressure processing: study of Alicyclobacillus acidoterrestris in orange juice. Food Microbiology, 32(1), 206-211. Simmonds, P., Mossel, B. L., Intaraphan, T., & Deeth, H. C. (2003). Heat Resistance of Bacillus Spores When Adhered to Stainless Steel and Its Relationship to Spore Hydrophobicity. Journal of Food Protection, 66(11), 2070-2075. doi: 10.4315/0362- 028x-66.11.2070 Sinensky, M. (1974). Homeoviscous adaptation—a homeostatic process that regulates the viscosity of membrane lipids in Escherichia coli. Proceedings of the National Academy of Sciences, 71(2), 522-525. Singer, G. A., & Hickey, D. A. (2003). Thermophilic prokaryotes have characteristic patterns of codon usage, amino acid composition and nucleotide content. Gene, 317, 39-47. Sinigaglia, M., Corbo, M. R., Altieri, C., Campaniello, D., D'Amato, D., & Bevilacqua, A. (2003). Combined effects of temperature, water activity, and pH on Alicyclobacillus acidoterrestris spores. Journal of Food Protection®, 66(12), 2216-2221. Smelt, J., Bos, A., Kort, R., & Brul, S. (2008). Modelling the effect of sub (lethal) heat treatment of Bacillus subtilis spores on germination rate and outgrowth to exponentially growing vegetative cells. International journal of food microbiology, 128(1), 34-40. Smelt, J., Stringer, S., & Brul, S. (2013). Behaviour of individual spores of non proteolytic Clostridium botulinum as an element in quantitative risk assessment. Food control, 29(2), 358-363.

224

Smelt, J. P., Otten, G. D., & Bos, A. P. (2002). Modelling the effect of sublethal injury on the distribution of the lag times of individual cells of Lactobacillus plantarum. International journal of food microbiology, 73(2), 207-212. Smit, Y., Cameron, M., Venter, P., & Witthuhn, R. C. (2011). Alicyclobacillus spoilage and isolation–A review. Food Microbiology, 28(3), 331-349. Smith,K.R., A. Woodward, D. Campbell-Lendrum, D.D. Chadee, Y. Honda, Q. Liu, J.M. Olwoch, B. Revich, A., & Sauerbor, R. (2014). Human health: impacts, adaptation, and co-benefits, in: Field, C.B., V.R.B., D.J. Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C.G., B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L.W. (Eds.), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Soh, A. L. A., Ralambotiana, H., Ollivier, B., Prensier, G., Tine, E., & Garcia, J.-L. (1991). Clostridium thermopalmarium sp. nov., a moderately thermophilic butyrate- producing bacterium isolated from palm wine in Senegal. Systematic and Applied Microbiology, 14(2), 135-139. Sokołowska, B., Skąpska, S., Fonberg-Broczek, M., Niezgoda, J., Chotkiewicz, M., Dekowska, A., & Rzoska, S. (2013). Factors influencing the inactivation of Alicyclobacillus acidoterrestris spores exposed to high hydrostatic pressure in apple juice. High Pressure Research, 33(1), 73-82. Sorokulova, I., Reva, O., Smirnov, V., Pinchuk, I., Lapa, S., & Urdaci, M. (2003). Genetic diversity and involvement in bread spoilage of Bacillus strains isolated from flour and ropy bread. Letters in applied microbiology, 37(2), 169-173. Sperber, W. H., & Doyle, M. P. (2009). Compendium of the microbiological spoilage of food and beverages. New York: Springer. Spinelli, A. C. N., Sant'Ana, A. S., Rodrigues-Junior, S., & Massaguer, P. R. (2009). Influence of different filling, cooling, and storage conditions on the growth of Alicyclobacillus acidoterrestris CRA7152 in orange juice. Applied and Environmental Microbiology, 75(23), 7409-7416. Spllttstoesser, D., Churey, J., & Lee, C. (1994). Growth characteristics of aciduric sporeforming bacilli isolated from fruit juices. Journal of Food Protection®, 57(12), 1080-1083.

225

Stackebrandt, E., Frederiksen, W., Garrity, G. M., Grimont, P. A., Kämpfer, P., Maiden, M. C., . . . Trüper, H. G. (2002). Report of the ad hoc committee for the re-evaluation of the species definition in bacteriology. International Journal of Systematic and Evolutionary Microbiology, 52(3), 1043-1047. Ståhl, S., & Ljunger, C. (1976). Calcium uptake by Bacillus stearothermophilus: a requirement for thermophilic growth. FEBS Letters, 63 (1), 184-187. Steyn, C. E., Cameron, M., & Witthuhn, R. C. (2011). Occurrence of Alicyclobacillus in the fruit processing environment—a review. International journal of food microbiology, 147(1), 1-11. Stoeckel, M., Lücking, G., Ehling-Schulz, M., Atamer, Z., & Hinrichs, J. (2016). Bacterial spores isolated from ingredients, intermediate and final products obtained from dairies: thermal resistance in milk. Dairy Science & Technology, 1-9. Stringer, S., Webb, M., & Peck, M. (2011). Lag time variability in individual spores of Clostridium botulinum. Food Microbiology, 28(2), 228-235. Stringer, S. C., Webb, M. D., George, S. M., Pin, C., & Peck, M. W. (2005). Heterogeneity of times required for germination and outgrowth from single spores of nonproteolytic Clostridium botulinum. Applied and Environmental Microbiology, 71(9), 4998-5003. Stringer, S. C., Webb, M. D., & Peck, M. W. (2009). Contrasting effects of heat treatment and incubation temperature on germination and outgrowth of individual spores of nonproteolytic Clostridium botulinum bacteria. Applied and Environmental Microbiology, 75(9), 2712-2719. Suutari, M., & Laakso, S. (1994). Microbial fatty acids and thermal adaptation. Critical reviews in microbiology, 20(4), 285-328. Tai, S.-K., Lin, H.-P. P., Kuo, J., & Liu, J.-K. (2004). Isolation and characterization of a cellulolytic Geobacillus thermoleovorans T4 strain from sugar refinery wastewater. Extremophiles, 8(5), 345-349. Tellez, M., & Antonio, B. (2016). Identification of sporeforming bacteria in the milk chain and its potential sources in farm environment. Terano, H., Takahashi, K., & Sakakibara, Y. (2005). Characterization of spore germination of a thermoacidophilic spore-forming bacterium, Alicyclobacillus acidoterrestris. Bioscience, biotechnology, and biochemistry, 69(6), 1217-1220.

226

Thelen, K., Snaidr, J., & Beimfohr, C. (2003). Specific rapid detection of Alicyclobacilus by fluorescently-labeled gene probes in fruit juices. Fruit processing: journal for the fruit processing and juice producing european and overseas industry, 13(6), 416- 418. Thompson, J., Waites, W., & Dodd, C. (1998). Detection of rope spoilage in bread caused by Bacillus species. Journal of Applied Microbiology, 85(3), 481-486. Thompson, J. M., Dodd, C. E., & Waites, W. M. (1993). Spoilage of bread by Bacillus. International Biodeterioration & Biodegradation, 32(1), 55-66. Tirado, M., Clarke, R., Jaykus, L.-A., McQuatters-Gollop, A., & Frank, J. (2010). Climate change and food safety: A review. Food Research International, 43(7), 1745-1765. Trmčić, A., Martin, N., Boor, K., & Wiedmann, M. (2015). A standard bacterial isolate set for research on contemporary dairy spoilage. Journal of dairy science, 98(8), 5806- 5817. Tsuruoka, N., Isono, Y., Shida, O., Hemmi, H., Nakayama, T., & Nishino, T. (2003). Alicyclobacillus sendaiensis sp. nov., a novel acidophilic, slightly thermophilic species isolated from soil in Sendai, Japan. International Journal of Systematic and Evolutionary Microbiology, 53(4), 1081-1084. Uemura, K., Kobayashi, I., & Inoue, T. (2009). Inactivation of Alicyclobacillus acidoterrestris in orange juice by high electric field alternating current. Food science and technology research, 15(3), 211-216. UNFCCC, 2014. United Nations Framework Convention on Climate Change. VanderKelen, J.J., Mitchell, R.D., Laubscher, A., Black, M.W., Goodman, A.L., Montana, A.K., Dekhtyar, A.M., Jimenez-Flores, R., & Kitts, C.L. (2016). Short communication: Typing and tracking Bacillaceae in raw milk and milk powder using pyroprinting. Journal of dairy science, 99, 146-151. Van der Merwe, E. (2011). Identification of precursors present in fruit juice that lead to the production of guaiacol by Alicyclobacillus acidoterrestris. Stellenbosch University. van der Spiegel, M., van der Fels-Klerx, H. J., & Marvin, H. J. P. (2012). Effects of climate change on food safety hazards in the dairy production chain. Food Research International, 46(1), 201-208. doi: https://doi.org/10.1016/j.foodres.2011.12.011 van Loosdrecht, M. C., Lyklema, J., Norde, W., & Zehnder, A. J. (1989). Bacterial adhesion: a physicochemical approach. Microbial Ecology, 17(1), 1-15.

227

van Melis, C., Den Besten, H., Groot, M. N., & Abee, T. (2014). Quantification of the impact of single and multiple mild stresses on outgrowth heterogeneity of Bacillus cereus spores. International journal of food microbiology, 177, 57-62. Vercammen, A., Vivijs, B., Lurquin, I., & Michiels, C. W. (2012). Germination and inactivation of Bacillus coagulans and Alicyclobacillus acidoterrestris spores by high hydrostatic pressure treatment in buffer and tomato sauce. International journal of food microbiology, 152(3), 162-167. Viedma, P. M., Abriouel, H., Omar, N. B., López, R. L., & Gálvez, A. (2010). Effect of enterocin EJ97 against Geobacillus stearothermophilus vegetative cells and endospores in canned foods and beverages. European Food Research and Technology, 230(3), 513-519. Viedma, P. M., Abriouel, H., Omar, N. B., López, R. L., Valdivia, E., & Gálvez, A. (2009). Inactivation of Geobacillus stearothermophilus in canned food and coconut milk samples by addition of enterocin AS-48. Food Microbiology, 26(3), 289-293. Vieira, M. C., Teixeira, A. A., Silva, F. M., Gaspar, N., & Silva, C. L. (2002). Alicyclobacillus acidoterrestris spores as a target for Cupuaçu (Theobroma grandiflorum) nectar thermal processing: kinetic parameters and experimental methods. International journal of food microbiology, 77(1), 71-81. Vinter, V. (1969). Physiology and biochemistry of sporulation. The bacterial spore, 1, 73- 123. Vogler, A. J., Keys, C. E., Allender, C., Bailey, I., Girard, J., Pearson, T., . . . Keim, P. (2007). Mutations, mutation rates, and evolution at the hypervariable VNTR loci of Yersinia pestis. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 616(1), 145-158. Wagner, I. D., & Wiegel, J. (2008). Diversity of thermophilic anaerobes. Annals of the New York Academy of Sciences, 1125(1), 1-43. Wakamoto, Y., Ramsden, J., & Yasuda, K. (2005). Single-cell growth and division dynamics showing epigenetic correlations. Analyst, 130(3), 311-317. doi: 10.1039/B409860A Walker, M., & Phillips, C. A. (2005). The effect of intermittent shaking, headspace and temperature on the growth of Alicyclobacillus acidoterrestris in stored apple juice. International journal of food science & technology, 40(5), 557-562.

228

Walker, M., & Phillips, C. A. (2008). Alicyclobacillus acidoterrestris: an increasing threat to the fruit juice industry? International journal of food science & technology, 43(2), 250-260. Walls, I., & Chuyate, R. (2000). Isolation of Alicyclobacillus acidoterrestris from fruit juices. Journal of AOAC International, 83(5), 1115-1120. Wang, B. S., Li, B. S., Zeng, Q. X., Huang, J., Ruan, Z., ZHU, Z. W., & Li, L. (2009). Inactivation kinetics and reduction of Bacillus coagulans spore by the combination of high pressure and moderate heat. Journal of food process engineering, 32(5), 692-708. Wang, G., Paredes‐Sabja, D., Sarker, M., Green, C., Setlow, P., & Li, Y. q. (2012). Effects of wet heat treatment on the germination of individual spores of Clostridium perfringens. Journal of Applied Microbiology, 113(4), 824-836. Wang, H.-c., & Hickey, D. A. (2002). Evidence for strong selective constraint acting on the nucleotide composition of 16S ribosomal RNA genes. Nucleic acids research, 30(11), 2501-2507. Wang, J., Hu, X., & Wang, Z. (2010). Kinetics models for the inactivation of Alicyclobacillus acidiphilus DSM14558 T and Alicyclobacillus acidoterrestris DSM 3922 T in apple juice by ultrasound. International journal of food microbiology, 139(3), 177-181. Wang, J., Zhao, G., Liao, X., & Hu, X. (2010). Effects of microwave and ultrasonic wave treatment on inactivation of Alicyclobacillus. International journal of food science & technology, 45(3), 459-465. Wang, Z., Cai, R., Yuan, Y., Niu, C., Hu, Z., & Yue, T. (2014). An immunomagnetic separation-real-time PCR system for the detection of Alicyclobacillus acidoterrestris in fruit products. International journal of food microbiology, 175, 30-35. Wang, Z., Yue, T., Yuan, Y., Cai, R., Guo, C., Wang, X., & Niu, C. (2012). Development of polyclonal antibody‐based indirect enzyme‐linked immunosorbent assay for the detection of Alicyclobacillus strains in apple juice. Journal of food science, 77, M643- M649. Warda, A. K., den Besten, H. M., Sha, N., Abee, T., & Groot, M. N. N. (2015). Influence of food matrix on outgrowth heterogeneity of heat damaged Bacillus cereus spores. International journal of food microbiology, 201, 27-34. Warth, A., Ohye, D., & Murrell, W. (1963). The composition and structure of bacterial spores. The Journal of cell biology, 16(3), 579-592.

229

Watanabe, T., Furukawa, S., Hirata, J., Koyama, T., Ogihara, H., & Yamasaki, M. (2003). Inactivation of Geobacillus stearothermophilus spores by high-pressure carbon dioxide treatment. Applied and Environmental Microbiology, 69(12), 7124-7129. Watkinson, W. (2008). Chemistry of detergents and disinfectants. Cleaning-in-Place: Dairy, Food and Beverage Operations, 56-79. Watterson, M., Kent, D., Boor, K., Wiedmann, M., & Martin, N. (2014). Evaluation of dairy powder products implicates thermophilic sporeformers as the primary organisms of interest. Journal of dairy science, 97(4), 2487-2497. Webb, M. D., Pin, C., Peck, M. W., & Stringer, S. C. (2007). Historical and contemporary NaCl concentrations affect the duration and distribution of lag times from individual spores of nonproteolytic Clostridium botulinum. Applied and Environmental Microbiology, 73(7), 2118-2127. Webb, M. D., Stringer, S. C., Le Marc, Y., Baranyi, J., & Peck, M. W. (2012). Does proximity to neighbours affect germination of spores of non-proteolytic Clostridium botulinum? Food Microbiology, 32(1), 104-109. Wells-Bennik, M. H., Driehuis, F., & van Hijum, S. A. (2016). Prospects for improved control of dairy-relevant sporeformers using-omics technologies. Current Opinion in Food Science, 10, 38-44. Weng, F., Chiou, C., Lin, P., & Yang, S. (2009). Application of recA and rpoB sequence analysis on phylogeny and molecular identification of Geobacillus species. Journal of Applied Microbiology, 107(2), 452-464. Wiegel, J., Braun, M., & Gottschalk, G. (1981). Clostridium thermoautotrophicum species novum, a thermophile producing acetate from molecular hydrogen and carbon dioxide. Current Microbiology, 5(4), 255-260. Williams, J. G., Kubelik, A. R., Livak, K. J., Rafalski, J. A., & Tingey, S. V. (1990). DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic acids research, 18(22), 6531-6535. Wisotzkey, J. D., Jurtshuk JR, P., Fox, G. E., Deinhard, G., & Poralla, K. (1992). Comparative Sequence Analyses on the 16S rRNA (rDNA) of Bacillus acidocaldarius, Bacillus acidoterrestris, and Bacillus cycloheptanicus and Proposal for Creation of a New Genus, Alicyclobacillus gen. nov. International Journal of Systematic and Evolutionary Microbiology, 42(2), 263-269.

230

Wittwer, C. T., Reed, G. H., Gundry, C. N., Vandersteen, J. G., & Pryor, R. J. (2003). High- resolution genotyping by amplicon melting analysis using LCGreen. Clinical chemistry, 49(6), 853-860. Wong, G. C., & O’Toole, G. A. (2011). All together now: Integrating biofilm research across disciplines. Mrs Bull, 36(5), 339-342. Xanthiakos, K., Simos, D., Angelidis, A., Nychas, G. E., & Koutsoumanis, K. (2006). Dynamic modeling of Listeria monocytogenes growth in pasteurized milk. Journal of Applied Microbiology, 100(6), 1289-1298. Yamamoto, K., Murakami, R., & Takamura, Y. (1998). Isoprenoid quinone, cellular fatty acid composition and diaminopimelic acid isomers of newly classified thermophilic anaerobic Gram-positive bacteria. FEMS microbiology letters, 161(2), 351-358. Yamazaki, K., Murakami, M., Kawai, Y., Inoue, N., & Matsuda, T. (2000). Use of nisin for inhibition of Alicyclobacillus acidoterrestris in acidic drinks. Food Microbiology, 17(3), 315-320. Yamazaki, K., Okubo, T., Inoue, N., & Shinano, H. (1997). Randomly amplified polymorphic DNA (RAPD) for rapid identification of the spoilage bacterium Alicyclobacillus acidoterrestris. Bioscience, biotechnology, and biochemistry, 61(6), 1016-1018. Yamazaki, K., Teduka, H., Inoue, N., & Shinano, H. (1996). Specific primers for detection of Alicyclobacillus acidoterrestris by RT‐PCR. Letters in applied microbiology, 23(5), 350-354. Yamazaki, K., Teduka, H., & Shinano, H. (1996). Isolation and identification of Alicyclobacillus acidoterrestris from acidic beverages. Bioscience, biotechnology, and biochemistry, 60(3), 543-545. Yokota, A., Fujii, T., & Goto, K. (2008). Alicyclobacillus: thermophilic acidophilic Bacilli: Springer Science & Business Media. Yoo, J.-A., Hardin, M.T., & Chen, X.D. (2006). The influence of milk composition on the growth of Bacillus stearothermophilus. Journal of Food Engineering, 77, 96-102. Yoon, H.-G., Lee, K.-H., Kim, H.-Y., Kim, H.-K., Shin, D.-H., Hong, B.-S., & Cho, H.-Y. (2002). Gene cloning and biochemical analysis of thermostable chitosanase (TCH-2) from Bacillus coagulans CK108. Bioscience, biotechnology, and biochemistry, 66(5), 986- 995.

231

Yuan, D.-D., Liu, G.-C., Ren, D.-Y., Zhang, D., Zhao, L., Kan, C.-P., . . . Zhang, L.-B. (2012). A survey on occurrence of thermophilic bacilli in commercial milk powders in China. Food control, 25(2), 752-757. Yuan, Y., Hu, Y., Yue, T., Chen, T., & Lo, Y. (2009). Effect of ultrasonic treatments on thermoacidophilic Alicyclobacillus acidoterrestris in apple juice. Journal of Food Processing and Preservation, 33(3), 370-383. Zain, S. N. M., Flint, S. H., Bennett, R., & Tay, H.-S. (2016). Characterisation and biofilm screening of the predominant bacteria isolated from whey protein concentrate 80. Dairy Science & Technology, 96(3), 285-295. Zeigler, D. R. (2005). Application of a recN sequence similarity analysis to the identification of species within the bacterial genus Geobacillus. International Journal of Systematic and Evolutionary Microbiology, 55(3), 1171-1179. Zeigler, D. R. (2014). The Geobacillus paradox: why is a thermophilic bacterial genus so prevalent on a mesophilic planet? Microbiology, 160(1), 1-11. Zeikus, J. (1979). Thermophilic bacteria: ecology, physiology and technology. Enzyme and Microbial Technology, 1(4), 243-252. Zhang, J., Yue, T., & Yuan, Y. (2013). Alicyclobacillus contamination in the production line of kiwi products in China. PloS one, 8(7), e67704. Zhang, P., Garner, W., Yi, X., Yu, J., Li, Y.-q., & Setlow, P. (2010). Factors affecting variability in time between addition of nutrient germinants and rapid dipicolinic acid release during germination of spores of Bacillus species. Journal of bacteriology, 192(14), 3608-3619. Zhao, Y., Caspers, M. P., Abee, T., Siezen, R. J., & Kort, R. (2012). Complete genome sequence of Geobacillus thermoglucosidans TNO-09.020, a thermophilic sporeformer associated with a dairy-processing environment. Journal of bacteriology, 194(15), 4118-4118. Zhou, T., Dong, Z., Setlow, P., & Li, Y.-q. (2013). Kinetics of germination of individual spores of Geobacillus stearothermophilus as measured by Raman spectroscopy and differential interference contrast microscopy. PloS one, 8(9), e74987. Zottola, E. A. (2003). Bacterial Spoilage. In B. Caballero, P. M. Finglas, & F. Toldra (Eds.), Encyclopedia of Food Sciences and Nutrition (pp. 5506-5510), Academic Press.

232

List of Publications Peer reviewed articles Kakagianni, M., Gougouli M., Koutsoumanis, K.P., 2016. Development and application of Geobacillus stearothermophilus growth model for predicting spoilage of evaporated milk. Food Microbiology 57, 28-35. Kakagianni, M., Aguirre, J.S., Lianou, A., Koutsoumanis, K.P., 2017. Effect of storage temperature on the lag time of Geobacillus stearothermophilus individual spores. Food Microbiology 67, 76-84. Kakagianni, M., Kalantzi, K., Beletsiotis, E., Ghikas, D., Lianou, A., Koutsoumanis, K.P., 2018. Development and validation of predictive models for the effect of storage temperature and pH on the growth boundaries and kinetics of Alicyclobacillus acidoterrestris in fruit juices. Food Microbiology 74, 40-49. Kakagianni, M., Koutsoumanis, K.P., 2018. Mapping the risk of evaporated milk spoilage in the Mediterranean region based on the effect of temperature conditions on Geobacillus stearothermophilus growth. Food Research International 111, 104-110.

Awards Participation, after proposal submission and selection, in the στην 6th campaign “Spin your Thesis! – 2015’’ of the European Space Agency (07-18/09/15), held at the European Space Research and Technology Center (ESTEC) in Noordwijk (Netherlands). Team MAH (Microbiology And Hygiene Group) (Aspridou Zafeiro, Kakagianni Myrsini, Dimakopoulou-Papazoglou Dafni), under the supervision of Assoc. Prof. Konstantinos Koutsoumanis entitled “Effect of hypergravity on microbial heat resistance”

(Award of Excellence from Aristotle University of Thessaloniki)

International conferences M. Kakagianni and K.P. Koutsoumanis (2017). Development and application of Geobacillus stearothermophilus predictive growth model as a tool to assess risk of evaporated milk spoilage. Q-Safe International conference of Predictive Modelling, Quantitative Risk Assessment and Life Cycle Analysis in Food Science and Biosciences, April 10-12, Syros Island, Greece (Oral Presentation).

233

M. Kakagianni, Koutsoumanis K. and Valdramidis V. (2016). Effect of ultrasound on recovery kinetics of Alicyclobacillus acidoterrestris spores. 9th biennial FOODSIM’ 2016, April 03-07, Ghent, Belgium (Oral presentation). Ζ. Aspridou, Μ. Kakagianni, D. Dimakopoulou-Papazoglou and K. Koutsoumanis (2015). Effect of Hypergravity on bacterial motility and heat resistance. 1st Symposium on Space Educational Activities, December 09-12, Padova, Italy (Oral presentation).

M. Kakagianni and K.P. Koutsoumanis (2015). Modelling the effect of temperature and pH on Alicyclobacillus acidoterrestris growth as a tool to assess the risk of spoilage in fruit juices. 29th EFFoST International Conference Food Science Research and Innovation: Delivering sustainable solutions to the global economy and society, November 10-12, Athens, Greece (Poster presentation, P1.152, Ref. No. 0319). M. Kakagianni and K.P. Koutsoumanis (2015). A predictive model for Alicyclobacillus acidoterrestris growth as a tool to assess risk of fruit juice spoilage. EFSA’s 2nd Scientific Conference: Shaping the Future of Food Safety, Together, October 14- 16, Milan, Italy (Poster presentation, P. 179). M. Kakagianni and K.P. Koutsoumanis (2015). Development and application of a predictive model for Alicyclobacillus acidoterrestris growth as a tool to assess risk of fruit juice spoilage. ICPMF 9th International Conference on Predictive Modelling in Food, September 08-12, Rio de Janeiro, Brazil (Poster presentation, P.066). Kakagianni, M., Gougouli, M. and Koutsoumanis K. (2014). Development and application of a predictive model for Geobacillus stearothermophilus growth as a tool to assess risk of evaporated milk spoilage. 24th International ICFMH Conference, Food Micro 2014, September 01-04, Nantes, France (Poster presentation, pp. 285).

National conferences Μ. Κακαγιάννη, Κ. Κουτσουμανής (2016). Πρόβλεψη της αλλοίωσης του συμπυκνωμένου γάλακτος κατά την εξαγωγή του σε τρίτες χώρες με θερμό κλίμα. 4η Επιστημονική Ημερίδα Γαλακτοκομίας & Τυροκομίας, Οκτώβριος 01, Αθήνα (Προφορική Παρουσίαση μετά από πρόσκληση). Kakagianni M., Koutsoumanis K.P. (2012). Study of the kinetic behavior of Alicyclobacillus acidoterrestris in broth. 5ο Συνέδριο Μικροβιόκοσμου, Δεκέμβριος 13-15, Αθήνα (Αναρτημένη Ανακοίνωση, p. 231).

234

Κακαγιάννη Μ., Χαρισμιάδου Ο., Κουτσουμανής Κ.Π. (2011). Μελέτη της κινητικής συμπεριφοράς του Geobacillus stearothermophilus και της επίδρασης του στην αλλοίωση γάλακτος εβαπορέ. 4ο Πανελλήνιο Συνέδριο Τροφίμων της Ελληνικής Κτηνιατρικής Εταιρίας, Σύγχρονη προσέγγιση στην Υγιεινή και Ασφάλεια Τροφίμων, Νοέμβριος 11-13, Θεσσαλονίκη (Αναρτημένη ανακοίνωση, P08).

235

Cover design: Eleni Nasta

236