HONEY MICROBIOME AND METABOLOME: A VAST RESERVOIR OF NATURAL ANTIMICROBIALS A Dissertation Presented to the Faculty of the Graduate School of Cornell University In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy by Zirui Ray Xiong August 2022 © 2022 Zirui Ray Xiong HONEY MICROBIOME AND METABOLOME: A VAST RESERVOIR OF NATURAL ANTIMICROBIALS Zirui Ray Xiong, Ph. D. Cornell University 2022 Raw honeys contain diverse microbial communities, which have the potential to produce antimicrobial secondary metabolites. These naturally occurring antimicrobials are highly valuable for industry application. Investigating honey microbiome and metabolome can provide important information on factors influencing the microbial communities and antimicrobials produced. In our study, amplicon metagenomics was used to analyze the composition of microorganisms in raw honey and investigate environmental and physicochemical variables that are associated with different microbial communities. The analyzed honey samples had relatively similar bacterial communities but more distinct and diverse fungal communities. Honey type was determined as a significant factor influencing alpha and beta diversity metrics of bacterial and fungal communities. Important bacterial and fungal amplicon sequence variants (ASVs) that influenced the overall community were identified. To obtain novel antimicrobials from natural sources, bacteria from raw honey were isolated and their antifungal-producing potential was evaluated. Naturally occurring antifungal secondary metabolites from these bacteria were further purified and identified. Using mass spectrometry and whole-genome sequence data, the main antifungal compound produced by two Bacillus velezensis isolates was determined as iturin A, a lipopeptide exhibiting broad spectrum antifungal activity. Results from this study provide important insights into the microbial communities associated with different types of raw honey and their antifungal metabolites. This research could improve our understanding of microbial dynamics in beehives, improve honey production, and prevent honeybee disease. Currently, there is a high demand for natural, broad- spectrum, and eco-friendly bio-fungicides in the food industry. Naturally occurring antifungal products from food-isolated bacteria are ideal candidates for agricultural and food applications. BIOGRAPHICAL SKETCH Zirui Ray Xiong was born in Wuhan, China. He received his Bachelor of Science degree in Biotechnology from University of Science and Technology of China. Following the guidance of a mentor, he developed a strong interest in Food Science and Microbiology. He decided to pursue a master’s degree in Food Science at Cornell University in 2016. After receiving the master of professional sciences degree, he worked as a laboratory technician at the Cornell AgriTech HPP Validation Center. Under the influence of great mentors, he started his PhD journey in 2018 spring. Over the past six years at Cornell University, he has worked on several research projects related to food safety, food quality, fermentation, protein chemistry, and bioinformatics. After graduating in summer 2022, he plans to work in the industry as a research scientist. iii To my parents, Wang Ping and Xiong Zhizhong, for their love, sacrifice, and dedication. iv ACKNOWLEDGMENTS First, I would like to thank Randy for believing in me since day one. Looking back now, it was tough being in a new country, away from all my families and friends, studying a new subject which I knew very little before. Randy has been a great mentor, showering me with trust, support, and encouragement. Not only did I learn from all his knowledge and experience, but more importantly, I learned to push myself out of my comfort zone and explore the unknown territory in science fearlessly. I would like to thank everyone from the Worobo lab, who have provided me with immense support during the past six years, especially to Jonathan Sogin, Ann Charles Vegdahl, Mario Cobo, John Churey, and Abby Snyder. I would also like to thank Dr. Martin Wiedmann and everyone from the FSL/MQIP family, who have welcomed me and provided me with resources and invaluable knowledge. This big science family has never hesitated whenever I needed help. I am confident that all the knowledge and skills I’ve learned from them for the past six years will benefit me for a lifetime. To all my friends, who stood by me all along this wonderful journey. I am grateful every day for having so many kind, generous, compassionate, brilliant, and loving people by my side. Last but not least, I would like to thank my parents. They have been supporting me unconditionally for the past 27 years. Even though they have no higher education or any science background, they know the value of education and they trust in science, which had an influence on me ever since I was a little kid. They have taught me to believe in myself, to overcome obstacles with persistent determination, and always be patient. They are and will always be my beacon in the night and the anchor to my soul. v TABLE OF CONTENTS BIOGRAPHICAL SKETCH iii DEDICATION iv ACKNOWLEDGMENTS v TABLE OF CONTENTS vi LIST OF FIGURES ix LIST OF TABLES xi CHAPTER 1 LITERATURE REVIEW: THE IMMENSE VALUE OF RAW HONEY AND ITS MICROBIOME 1 Part I. Honey 1 Physical properties, chemical composition, and biological activities 1 Honey history as traditional medicine 4 Honey microbiota 6 Part II. Food microbiome and foodomics studies 8 Definition, previous investigation, and significance 8 Identifying food microbiota with 16S rRNA gene 10 Whole-genome metagenomic sequencing 12 Other foodomic approaches 13 Part III. Controlling microbial contamination in food systems 15 Overview 15 Bacteriocins 17 Nonribosomal peptides (NRPs) and polyketides 19 vi Enzymes and other compounds 21 REFERENCES 23 CHAPTER 2 MICROBIOME ANALYSIS OF RAW HONEY REVEALS IMPORTANT FACTORS INFLUENCING THE BACTERIAL AND FUNGAL COMMUNITIES 48 Abstract 48 Introduction 50 Materials and methods 55 Results 59 Discussion 78 REFERENCES 87 CHAPTER 3 PURIFICATION AND CHARACTERIZATION OF ANTIFUNGAL LIPOPEPTIDE PRODUCED BY BACILLUS VELEZENSIS ISOLATED FROM RAW HONEY 102 Abstract 102 Introduction 103 Materials and methods 107 Results 115 Discussion 135 REFERENCES 143 CHAPTER 4 vii LOOKING AHEAD: UNLOCK THE FULL POTENTIAL OF RAW HONEY 155 REFERENCES 160 viii LIST OF FIGURES Figure 2.1 Honey bacterial composition plot. 65 Figure 2.2 Honey fungal composition plot. 66 Figure 2.3 Venn diagrams for bacterial and fungal ASVs of four honey types. 67 Figure 2.4 Alpha diversity metrics of honey bacterial community. 69 Figure 2.5 Alpha diversity metrics of honey fungal community. 70 Figure 2.6 Heatmaps of Bray-Curtis distances between honey bacterial and fungal community. 73 Figure 2.7 Non-metric multidimensional scaling (NMDS) ordination for bacterial community structure based on the relative abundance of 16S ASVs. 74 Figure 2.8 Non-metric multidimensional scaling (NMDS) ordination for fungal community structure based on the relative abundance of ITS ASVs. 75 Figure 2.9 Top coefficient amplicon sequence variants (ASVs) for beta diversity. 77 Figure 3.1 Deferred inhibition assay of purified products from Bacillus velezensis ix WRB-ZX-001 and WRB-ZX-002 against food-isolated Aspergillus fumigatus. 120 Figure 3.2 Reversed-phase HPLC of purified products of Bacillus velezensis WRB-ZX-001 and WRB-ZX-002. 121 Figure 3.3 Mass spectra for purified antifungal compounds produced by Bacillus velezensis WRB-ZX-001 and WRB-ZX-002. 122 Figure 3.4 Core genome phylogeny of 43 Bacillus amyloliquefaciens group isolates. Maximum likelihood tree was constructed with core genome SNPs identified by kSNP. 126 Figure 3.5 Genome comparison of Bacillus velezensis WRB-ZX-001 and WRB- ZX-002 against closely related Bacillus type strains. 128 Figure 3.6 Growth curve and antifungal activity curve for Bacillus velezensis WRB-ZX-001 (A) and WRB-ZX-002 (B). 133 x LIST OF TABLES Table 2.1 Physicochemical properties of honey. 61 Table 3.1 Food-isolated fungal strains used in this study as indicators. 116 Table 3.2 Summary of identity, source, and cross-reactivity against food- associated fungal indicators of honey bacterial isolates. 117 Table 3.3 Antifungal activity of purification products of Bacillus velezensis isolates against food-isolated Aspergillus fumigatus. Antifungal activity unit (AU/mL) is defined as the reciprocal of the highest dilution showing a clear inhibition zone. 118 Table 3.4 Potential secondary metabolite synthesis gene clusters identified in Bacillus velezensis WRB-ZX-001 and WRB-ZX-002 by antiSMASH. 129 Table 3.5 Antifungal activity of heat-treated and protease-treated purified products of Bacillus velezensis isolates against food-isolated Aspergillus fumigatus. 134 xi CHAPTER 1 LITERATURE REVIEW: THE IMMENSE VALUE OF RAW HONEY AND ITS MICROBIOME Part I. Honey Physical properties, chemical composition, and biological activities Honey is produced by honey bees, Apis melifera, which collect and transform nectar, plant excretions, and bee secretions into a natural sweet substance (Olaitan et al., 2007). Excess water evaporates during the process, which takes several days. Bees produce invertase, an enzyme that digests sucrose, the predominant sugar in nectar, into glucose and fructose. Honey is a nutritious food that has been consumed by humans for thousands of years. It is viscous, aromatic, sweet, and commonly used as a sweetener for food. Honey has a variety of nutritional and health benefits, making it popular among consumers. Sugar accounts for about 85% of the total solids in honey. Fructose and glucose are the two major simple sugars, accounting for 85-95% of total sugar (Olaitan et al., 2007). Other monosaccharides, disaccharides, and oligosaccharides are present, including galactose, maltose (7%), isomaltose, and sucrose (1%), accounting for 5-15% of honey (Lazaridou et al., 2004; Ouchemoukh et al., 2010; Ruiz-Matute et al., 2010; Val et al., 1998). Honey crystalizes when monohydrate glucose crystallizes. The crystallization process is influenced by several factors including water and glucose content, micro-chemical composition, and storage conditions. As a supersaturated solution of sugar, honey has a high viscosity and hygroscopicity. Honey with water content less than 18.8% will absorb moisture from 1 the environment with a relative humidity of 60% or higher (Olaitan et al., 2007). As a product containing mostly sugar, honey has a low enough water activity to suppress the growth of most bacteria and fungi (Costa et al., 2013; Machado De-Melo et al., 2018). The mean water activity of honey ranges between 0.49 and 0.65 (Cavia et al., 2004; Costa et al., 2013). The growth of most bacteria is suppressed in foods when the water activity is below 0.91, while most yeast and molds are suppressed when the water activity is below 0.7 (Machado De-Melo et al., 2018). The threshold for microorganism growth in foods is 0.6 (Fontana Jr., 2007). For the nitrogen compounds in honey, total protein ranges from 0.2% to 1.6%, most of which come from bee glands and plant pollen (Chua et al., 2013; Won et al., 2009). Some amino acids in free or bound forms are also detected in honey (Paramás et al., 2006). Invertase (alpha-glucosidase), glucose oxidase, and diastase are the three most common and important enzymes in honey. Invertase and glucose oxidase are produced by the hypopharyngeal glands of honeybees, which help the transition of nectar to honey (Machado De-Melo et al., 2018). Diastase (amylase) originates from either bees or plants, and is used as a measure of honey freshness and an indicator of adulteration (Machado De-Melo et al., 2018). Due to the presence of reducing sugar and amino acids, the Maillard reaction may occur depending on storage and processing conditions, which may yield undesirable compounds like 5- hydroxymethylfurfural (5-HMF) (Iglesias et al., 2006). Glucose is oxidized by glucose oxidase to produce gluconic acid and hydrogen peroxide. As a slow-release antiseptic, hydrogen peroxide can inhibit the growth of microorganisms in honey, especially prior to sufficient water evaporation to reach low 2 water activity (al Somal et al., 1994). Honey has a pH between 3.2 and 4.5. Due to the buffering capacity of honey, its pH is not correlated to its acidity. Gluconic acid contributes to the acidity and characteristic taste of honey (Olaitan et al., 2007), and comprises 70-90% of organic acids. More than 30 non-aromatic organic acids are found in different types of honey, which originate from plants or enzymatic reactions and contribute to the color and flavor of honey (Mato et al., 2003; Mato et al., 2006). Acidity in honey can act as a hurdle to the growth of spoilage and pathogenic microorganisms, and free acidity is used for honey quality control (Leistner & Gorris, 1995; Ojeda de Rodrı́guez et al., 2004; Terrab et al., 2002). Honey contains flavonoids and phenolic acids, mainly originating from plant nectar, honeydew, and pollen (Ferreres et al., 1992). Phenolic compounds constitute approximately 56-500 mg per kg of honey (Al-Mamary et al., 2002). The most abundant phenolic compounds in honey are myricetin, quercetin, luteolin, protocatechuic acid, and p-hydroxybenzoic acid (Olas, 2020). Phenolic compounds contribute to the functional properties of honey as a healthy sweetener. These bioactive antioxidants can scavenge free radicals and reactive oxygen species (ROS), preventing their damage to cell membrane, enzymes, lipids, and DNA (Samarghandian et al., 2017). Flavonoids, phenolic acids and derivatives, and other organic acids are stable during honey storage and dilution, contributing to the antimicrobial, anti- inflammatory, and immunomodulatory properties of honey (Ben Sghaier et al., 2011). Additionally, honey contains different vitamins and minerals originating from plants, which are important for human diet (Madejczyk & Baralkiewicz, 2008). Some examples are potassium, calcium, copper, iron, manganese, phosphorus, vitamin C, 3 thiamine, riboflavin, nicotinic acid, pantothenic acid and so on (Machado De-Melo et al., 2018; Olaitan et al., 2007). In terms of absolute value, the content of these minerals and vitamins is low. But compared to sugar, honey is a healthier sweetener (Bogdanov et al., 2008). Liquid honey color ranges from clear and colorless to dark amber or black (Olaitan et al., 2007). Some factors influencing honey color include botanical origin, honey age, storage conditions, suspended particles (pollen), and enzymatic reactions (Maillard reaction, lipid oxidation). In terms of chemical composition, sugar, carotenoids, xanthophylls, anthocyanins, phenolic compounds, minerals can have an influence on honey color (de Almeida-Muradian et al., 2014). Overall, several environmental factors determine the chemical composition and physicochemical properties of honey. These factors include botanical source, geographic origin, soil composition, climate, harvest season, and nectar flux intensity (Ojeda de Rodrı́guez et al., 2004). Maturation degree (ripeness) of honey and human factors, like extraction, processing, and storage will also influence honey chemical and biological composition (Machado De-Melo et al., 2018). Honeys from different floral sources, season, and location have varied levels of antimicrobial activity (Grabowski & Klein, 2017). Honey history as traditional medicine Honey has been used for its medicinal properties by different cultures for millennia (Allsop & Miller, 1996). The first record of humans collecting honey from wild bees dates to 6000 BC, while the first written record of using honey as a drug and 4 ointment dates back to 2000 BC (Crane, 1975). It was used as a remedy for infection long before bacteria were identified as the causative agents. Early records by Aristotle (c. 350 BC) mentioned honey as salve for wounds and sore eyes. Dioscorides from c. 50 AD described honey as “good for sunburn and spots on the face” and “good for all rotten and hollow ulcers” (Molan, 1999). More recently, honey was described to effectively clear infection and promote healing (Subrahmanyam, 2005). For wounds without infection, honey can reduce inflammation by quenching free radicals and reducing inflammatory mediators (Al-Waili & Boni, 2003; Bilsel et al., 2002; Postmes, 2001). For infected wounds, honey is a broad-spectrum antimicrobial agent that is effective against both bacteria and fungi (Molan et al., 1988; Radwan et al., 1984). Researchers found that honey can clear the infection of Pseudomonas in patients suffering from wounds (Cavanagh et al., 1970). Honey was found to be inhibitory towards a variety of bacterial and fungal pathogens, including Pseudomonas, Acinetobacter, Staphylococcus, Streptococcus, Salmonella, E. coli, Vibrio, Yersinia, Plesiomonas, Shigella, Clostridium, Aspergillus, Penicillium, Candida, and others (Efem, 1993; Molan, 1992; Mundo et al., 2004; Obaseiki-Ebor & Afonya, 1984; Obi et al., 1994). In addition to its bacteriostatic and bactericidal effects, honey has also been demonstrated as antiviral and antiparasitic (Maddocks & Jenkins, 2013). Honey is antimicrobial due to its osmotic effect, high acidity, presence of hydrogen peroxide and phytochemicals (Olaitan et al., 2007). Hydrogen peroxide produced by glucose oxidase was determined as the main antimicrobial component of honey (Adcock, 1962; Molan, 1992; White et al., 1963). Non-peroxide antimicrobial activity of honey comes from antioxidants like phenolic compounds, antimicrobial 5 peptides, and proteinaceous substances (Lee et al., 2008a, 2008b; Molan, 1992; Mundo et al., 2004; Truchado et al., 2009). The factors that may influence the antimicrobial activity of honey include botanical source, bee metabolism, seasonality, geographic location, climate, and other environmental factors (Basualdo et al., 2007). Honey microbiota Most microorganisms in honey are dormant and cannot grow or reproduce (Olaitan et al., 2007). These microorganisms originate from pollen, flowers, air, dust, dirt, and the honeybee digestive tract. Spore-forming bacteria, yeasts and molds can be introduced through various sources and survive in honey. Microbial contamination can introduce yeasts and bacteria into honey during human processing. The two most prevalent bacterial families are considered the core microbiota of honey: Bacillaceae and Lactobacillaceae. Other prevalent bacterial families include Enterobacteraceae, Acetobacteraceae, Microbacteriaceae, and Bifidobacteriaceae (Brudzynski, 2021). Lactic acid bacteria are frequently found in plant nectar and honey, including Lactobacillus and Fructobacillus. Strains of these genera possess antibacterial, antifungal, and anti-biofilm potentials (Berríos et al., 2018; Olofsson & Vásquez, 2008; Ramos et al., 2020). Honey contains probiotic bacteria like Bifidobacterium and Lactobacillus, originating from honeybee digestive tract. These probiotic bacteria are beneficial to human gut (Olofsson & Vásquez, 2008). Honey is also considered prebiotic due to the presence of oligosaccharides, which increase the population of lactobacilli and bifidobacteria in human gut microbiome (Sanz et al., 1995; Ustunol & Gandhi, 2001; Yun, 1996). 6 The potential presence of Clostridium botulinum spores is a risk associated with raw honey (Brown, 2000). Infants should not consume raw honey because these spores can germinate and grow inside their stomach and produce toxins, resulting in infant botulism (Machado De-Melo et al., 2018). Yeasts and molds that are osmotolerant, xerotolerant, and acidotolerant are also found in honey, including Bettsia, Ascosphaera, Metschnikowia, Pichia, Saccharomyces, Zygosaccharomyces (Kačániová et al., 2012; Rodríguez-Andrade et al., 2019). Many of these fungi are considered contaminants and may spoil the honey when the honey moisture content is above 18% (Chaven, 2014). Studying the honey microbiome can reveal bacterial and fungal communities originally from bee digestive tract, flower, and pathogens introduced during harvesting, processing, and storage (Bovo et al., 2020; Kňazovická et al., 2020; Wen et al., 2017). These studies further contribute to our understanding of the nutritional value and health benefits of this highly valuable food products. 7 Part II. Food microbiome and foodomics studies Definition, previous investigation, and significance Microbiome is defined as the total population of all microorganisms and their genomes inhabiting a particular environment, while microbiota is defined as all microorganisms of an ecosystem or a specific niche (Berg et al., 2020). Under this definition, microbiome is the sum of microbiota in an ecological niche and their activities, including structural elements, microbial metabolites, and surrounding environmental conditions. Microorganisms in foods play important roles, including fermentation, contamination, and spoilage. The rapid development of easy-to-use next-generation sequencing (NGS) technologies along with the lowering of costs to access those technologies has allowed more researchers than ever before to explore the microbiome of foods. Using high-throughput and high-resolution genomics, transcriptomics, proteomics, and metabolomics technologies to study food microbiomes is an emerging research field called “foodomics” (García-Cañas et al., 2012; Herrero et al., 2012). Food microbiomes are important to human health and food production. In-depth sequencing of food microbiomes can provide knowledge of community composition, functional potential, microbial activities and interactions in the environment, which can contribute to fermentation control, food safety and quality improvement, food adulteration prevention, identification of bioactive compounds in complex food systems, and elucidation of agricultural and economic values of these food products (Kafantaris et al., 2021). Foodomics research connects food components, diet, individual health, and diseases (Capozzi & Bordoni, 2013). Some specific applications 8 include early, rapid, reliable detection of pathogens, antimicrobial resistance genes, toxins, allergens, and other adulterants in foods (Andjelković et al., 2017). Foodomics research increases our understanding of biochemical, molecular, and cellular mechanisms of food microbiomes and their implications on the human microbiome and human health. Fermented foods were the first type of food products subjected to genomic analysis for microbiome investigation, including cheese, sausages, and kimchi (Ahn et al., 2014; Jung et al., 2013; Lessard et al., 2014; Połka et al., 2015). These studies revealed the dynamics of different microorganisms in food during the fermentation process over time, provided important information on starter cultures and spoilage organisms, and offered guidance on fermentation control. Additionally, the expression of metabolite genes was analyzed to elucidate the competition and survival strategies of these microorganisms as well as their contribution to flavor, nutrition, and human health (Jung et al., 2013; Lessard et al., 2014). Food microbial community structure and dynamics are associated with physicochemical properties of food products (De Filippis et al., 2017). One specific example is the survey of 60 Irish cheeses to evaluate their bacterial diversity (Quigley et al., 2012). Microbial composition was influenced by cheese type, milk origin, ingredients, salt content, and processing conditions. The distribution of microorganisms is spatially heterogeneous, with certain bacteria dominating rind, crust, and core of cheese. Studies on the cheese microbiome provided important information on the roles the microbiota play in cheese ripening, flavor, preservation, spoilage, and ecological dynamics (Ercolini, 2013). 9 Identifying food microbiota with 16S rRNA gene The “gold standard” to elucidate the food microbiome is to isolate and identify individual strains via culture-based methods, which are of low efficiency and biased. Some researchers estimate that only 0.1% of the microbial community can be identified with culture-based methods (Cao et al., 2017). The shift from using traditional culture-based methods to NGS technologies to characterize microbial communities in ecological systems is evolutionary, providing insights into diverse and dynamic systems that were previously uncharacterized or only partially characterized. At present, the most common high-throughput sequencing technology used in food-related research is amplicon sequencing. Marker genes, like the 16S rRNA gene (for bacteria identification), are amplified with primers using metagenomic DNA as templates and sequenced on massively parallel high-throughput platforms.16S rRNA gene sequencing can be used to characterize food microbiota composition and analyze the relative abundance and taxonomy of microorganisms. The 16S rRNA gene is comprised of 9 hypervariable regions flanked by conserved sequences (Neefs et al., 1993). This region is ideal for designing primers for DNA amplification of hypervariable regions and bacterial taxonomic classification. Traditional Sanger sequencing has been used extensively to amplify the 16S rRNA gene region and to investigate the food microbiota when combined with culture methods. However, this method is of low throughput and misses microbial population of low abundance. Using NGS platforms to amplify 16S rRNA gene region can significantly increase the sequencing capacity and thoroughly identify microbial population with a reasonable cost. The requirement for input DNA template is relatively low, making it possible to 10 use on foods with low bacterial abundance (Cao et al., 2017). However, this method has its limitations. The taxonomic and phylogenetic resolution from 16S rRNA gene sequencing is relatively low, and it cannot be used to classify taxonomy beyond species level. Moreover, the primer selection is complicated. There are 9 hypervariable regions for the 16S rRNA gene, and these regions do not perform equally well for amplicon sequencing. Some studies showed that V4/V5 region performs better than standard V3/V4 region in terms of sequencing efficiency and reducing amplification biases (Claesson et al., 2010). Regardless of the primer pair selection, longer amplicon fragments, longer read length and higher coverage will likely yield better classification results, but the amplification biases persist (Cao et al., 2017). 16S rRNA amplicon sequencing has been used by researchers to evaluate food microbiome composition and provide evidence on the impact of environment on food microbiome. Researchers evaluated the core microbiome of raw milk with 16S rRNA gene sequencing (Rodrigues et al., 2017). Spoilage organisms and pathogens were found in raw milk, including Acinetobacter, Thermoanaerobacterium, Enterobacteriaceae, and Streptococcus. A cheese microbiome study using 16S rRNA gene amplicon sequencing found that environmental microbiota from the cheese production site dominated the cheese samples, confirming that processing environment influenced food microbial community and may shape site-specific product characteristics (Bokulich et al., 2018). The microbiome analysis of powdered infant formula with high throughput 16S rRNA gene sequencing found that the most prevalent genera were Pseudomonas, Acinetobacter, and Streptococcus (Anvarian et 11 al., 2016). Microbiota with the highest diversity were found in areas with low care, and most microorganisms were associated with soil. The microbiota of ready-to-eat fruits and vegetables were found to be influenced by season, irrigation water, and soil using high throughput 16S rRNA gene sequencing (Telias et al., 2011; Williams et al., 2013). These studies provided important insights into the origin of spoilage and pathogenic organisms in fresh produce, contributing to the efforts on spoilage prevention and outbreak investigation (van Dyk et al., 2016). Whole-genome metagenomic sequencing To avoid primer and amplification biases, metagenomic sequencing (also called shotgun metagenomic sequencing) can be used to evaluate the microbiome composition by sequencing the entire DNA content in the sample without PCR amplification (Ercolini, 2013). Metagenomic sequencing obtains genetic information of all members within the sample community and provides in-depth taxonomic classification beyond the species level. The metagenomic data can be used to elucidate evolutionary history, community structure, metabolism, and function capabilities. For example, metagenomic analysis of food fermentation process can allow for monitoring of microbiota on strain-level and identifying key enzymes and metabolic activities that facilitate the fermentation process, like sugar and amino acid metabolism and production of flavor compounds (Scholz et al., 2016; Siezen et al., 2008). Furthermore, industrial strains with desirable traits, like high stress tolerance and high metabolism efficiency, can be selected based on metagenomic data and used to produce high quality products (Hao et al., 2011). However, metagenomic sequencing 12 is substantially more expensive than amplicon sequencing, and produces large amount of data that include virus, bacteria, archaea, fungi, protozoa, algae, and other DNAs present in the sample, making it computationally intensive to analyze and interpret (De Filippis et al., 2017). Other foodomic approaches In terms of transcriptomics, high-throughput microarray and RNA-seq have been used in foods to evaluate the presence, growth, and metabolism of foodborne pathogens (Lamas et al., 2019). Other applications include food authentication, detection of adulteration and genetically modified ingredients, and analysis of herbal food metabolites (Kafantaris et al., 2021; Ko et al., 2018; Lancova et al., 2011; Roy et al., 2018). Metatranscriptomic analysis can help us characterize the complex interactions between different microbial communities within a sample. For example, amino acid metabolism during cheese ripening is an important indicator of the roles that fungi and bacteria play in the flavor development and cheese maturation process (Dugat-Bony et al., 2015; Lessard et al., 2014; Monnet et al., 2016). Transcriptomic data provide insights into the metabolic activities of complex microbial communities and shed light into their interaction and dynamics. This information can help us control and manage the fermentation process and improve product quality. High-throughput proteomic and metabolomic tools can also be used on food products (Andjelković et al., 2017). Some applications include detecting adulteration of pathogens, toxins, and allergens; ensuring proper ingredient composition; studying potential biomarkers for authentication/traceability and bioactive compounds (Bordoni 13 & Capozzi, 2015; D’Alessandro & Zolla, 2012; Rešetar et al., 2015; Rezzi et al., 2007). Previous studies analyzed the proteome of honey using 2D electrophoresis with mass spectrometry (Borutinskaite et al., 2018; Rossano et al., 2012; Zhang et al., 2019). Glucose oxidase, alpha-glucosidase, and other antimicrobial peptides were identified in honey. Moreover, the proteomic profile of honey could be used as an identifier to differentiate honeys (Azevedo et al., 2017). For the metabolomic profile of honey, researchers used NMR spectroscopy to classify different types of honey (Schievano et al., 2010; Schievano et al., 2012). Metabolite fingerprinting with specific chemical markers can be used to distinguish honey from different geographic origins, floral source, and composition (Boffo et al., 2012; Razali et al., 2018). However, effectively analyzing the proteomic and metabolomic data and combining these analyses with transcriptomic and genomic data is computationally extremely difficult. With the advancement of NGS and the continuous expansion of the genome database of microorganisms in food, we can develop standardized global surveillance of foodborne pathogens, virulence factors, and antimicrobial resistance genes, making it possible to quickly identify and track contaminated food products and reduce adverse impacts on human health (Schlundt et al., 2020). These tools provide us access to minimize food safety and security problems and improve global public health. 14 Part III. Controlling microbial contamination in food systems Overview Microbial contamination of food products can cause serious health problems, food security issues, and environmental impacts. Food loss and waste generated by foodborne pathogens and spoilage organisms pose risks to a safe and efficient food system. Food waste at the retail and consumer levels was estimated to be 133 billion pounds in 2010, with an estimated value of $162 billion (Thakali & MacRae, 2021). Spoilage and pathogenic microorganisms are ubiquitous in the environment, and can be introduced to food products during production, processing, transportation, retail, and consumer stages (Thakali & MacRae, 2021). Spoilage microorganisms produce enzymes and by-products that cause the deterioration of odor, appearance, and taste of foods, making products undesirable and unacceptable for consumption, leading to economic loss and food waste (Nychas & Panagou, 2011). Some common spoilage organisms include nonspore-forming lactic acid bacteria, spore-forming Bacillus and Clostridium species, yeasts and molds (Lorenzo et al., 2018). Ingestion of food contaminated with pathogens causes foodborne illnesses. Some of the leading bacterial pathogens that cause illnesses, outbreaks, hospitalizations, and deaths in the US include pathogenic E. coli (ETEC), Shigella spp., Campylobacter spp., and Salmonella spp.. Foodborne diseases have significant economic and social costs. There are more than 600 million cases of foodborne illnesses and 420,000 deaths caused by 31 foodborne pathogens annually as estimated by WHO (Havelaar et al., 2015). Global burden of foodborne disease is estimated by 15 the disability adjusted life year (DALY) metric established by WHO, and it was 33 million DALYs in 2010 (Havelaar et al., 2015). Food safety and spoilage issues caused by pathogens and spoilage organisms are interrelated from ecological and microbiologic perspectives (Petruzzi et al., 2017). Chemical preservatives, such as sodium benzoates, sodium propionates, potassium sorbates, sodium nitrites and nitrates, sulfur dioxide, and other organic acids are commonly used in food products to control microorganism growth and extend shelf life. However, these synthetic chemical preservatives have low consumer acceptance, and may have adverse health effects, especially after long-term exposure (Sharma, 2015; Trasande et al., 2018; Zhong et al., 2018). As an alternative, natural antimicrobial agents produced by plants, animals, mushrooms, bacteria, and other natural sources are highly desirable (Villalobos-Delgado et al., 2019). These bioactive compounds can be incorporated into food formulation, films, coatings, and packaging to increase product shelf life. Honey is an ecological reservoir of antimicrobial compounds produced by microorganisms originating from plants and honeybees (Brudzynski, 2021). These antimicrobial compounds are considered secondary metabolites, defined as auxiliary metabolites not required for growth and survival of microorganisms, including antibiotics, pigments, hormones, and others (Singh et al., 2019). Some researchers consider honey as a stable colloidal system that preserves the structure and function of bioactive compounds, while releasing them upon honey dilution (Brudzynski & Sjaarda, 2021). Antimicrobial compounds produced by microorganisms in honey are mostly nonselective, partially explaining honey’s broad-spectrum inhibition of 16 bacteria and fungi. Some common antimicrobial compounds produced by bacterial strains include bacteriocins, surfactants, siderophores, and secreted enzymes. Studying these secondary metabolites can lead to discovery of new antimicrobial compounds in honey, with potential application in agricultural and medical fields. Here is a discussion of natural antimicrobials produced by bacteria with potential food applications. Bacteriocins Bacteriocins are ribosomally synthesized antimicrobial proteins or peptides produced by bacteria like Lactobacillus, Lactococcus, and Pediococcus. Some of these bacteriocins have a highly specific target, while others are broad-spectrum and effective against a variety of bacteria. Class I bacteriocins are small antimicrobial proteins with 19 to 38 amino acids. All class I bacteriocins undergo post-translational modifications to include uncommon amino acids and structures, such as lantibiotics with inter-residual thioester bonds (Alvarez-Sieiro et al., 2016). Some examples of class I bacteriocins are cyclized peptide enterocin AS-48, negatively charged circular sactipeptide subtilosin A, linear azole/azoline-containing peptides (LAPs) streptolysin S, glycocins, and lasso peptides. One of the most studied bacteriocins is nisin, a class IA lantibiotic produced by Lactococcus lactis subsp. lactis (Dodd et al., 1990). Nisin is a broad-spectrum bacteriocin that is active against a variety of Gram-positive bacteria including Staphylococcus, Listeria, Lactobacillus, Bacillus etc. It contains thioether amino acids lanthionine and methyllanthionine. Nisin is cationic because of the N-terminal 17 lanthionine ring. It can bind to the anionic phosphate group on the lipid II of Gram- positive bacterial cell wall, forming peptide-lipid II complex and initiating pore formation (Breukink et al., 1999). Subsequently, the C-terminal peptide is inserted into the cytoplasmic membrane to form transmembrane pore, causing ion efflux, collapse of proton motive force, cell permeabilization, and rapid cell death (Moll et al., 1999; van Heusden et al., 2002). Additionally, nisin can inhibit cell wall biosynthesis by binding to lipid II and disrupting the formation of peptidoglycan chain (Moll et al., 1999; Wiedemann et al., 2001). Nisin is used in food products like milk, cream, yogurt, cheese, canned vegetables, bakery products, cured meat, and others (Delves- Broughton et al., 1996; Silva et al., 2018). However, due to the presence of an outer membrane for Gram-negative bacteria, nisin and other lantibiotics are not able to penetrate and access lipid II on the cytoplasmic membrane and are generally not effective against Gram-negative bacteria, unless at extremely high concentration or combined with chelating agents to compromise the outer membrane (Stevens et al., 1991). Nisin is also limited to products with pH lower than 7 since it loses its activity at high pH (de Arauz et al., 2009). Class II bacteriocins are small, linear proteins that contain unmodified peptides. These bacteriocins are heat and pH stable (Abriouel et al., 2011). Pediocin is a class IIA bacteriocin produced by Pediococcus spp.. Pediocin is broad-spectrum antimicrobial and effective against both Gram-positive and Gram-negative bacteria, including Listeria monocytogenes, Staphylococcus aureus, Clostridium perfringens, Pseudomonas and E. coli (Silva et al., 2018). Pediocin functions by binding to the receptor of sugar transporter mannose phosphotransferase system and inserting into 18 target cell cytoplasmic membrane, which leads to pore formation and cell lysis (Diep Dzung et al., 2007). Pediocin is generally recognized as safe (GRAS) and widely used in milk and dairy products, like cream, cottage cheese, and cheese sauce to extend their shelf life (Pucci et al., 1988). Some other examples of class II bacteriocins include two-peptide bacteriocin lactococcin G, leaderless plasmid-encoded two- peptide enterocin L50, and single linear peptide lactococcin A. Class III bacteriocins are large heat-labile antimicrobial proteins with a molecular weight larger than 10 kDa. These bacteriocins usually have phospholipase activity. Examples include enterolysin A (34.5 kDa), zoocin A (29.2 kDa), and megacin A-216 (66 kDa). Both enterolysin A and zoocin A target the bacterial cell wall by cleaving the peptidoglycan and disrupting the cell wall structure (Khan et al., 2013; Simmonds et al., 1996). Megacin A-216 has a narrow antibacterial spectrum. It functions like phospholipase, converting phospholipids to lysophospholipids and impairing cell membrane integrity (Kiss et al., 2008). Since these bacteriocins are large and heat sensitive, they have yet to be used for food applications. Nonribosomal peptides (NRPs) and polyketides Nonribosomal peptides (NRPs) are peptide secondary metabolites synthesized by multidomain mega-enzymes called nonribosomal peptide synthetases (NRPSs). Their synthesis is independent of ribosomes and messenger RNAs (Evans et al., 2011). Bacteria and fungi synthesize these NRPs naturally. The peptide chain of NRPs are usually 3-15 amino acids in linear, cyclic, or branched forms (Mootz et al., 2002). Lipopeptides and siderophores are examples of thiotemplate NRPs. Lipopeptides have 19 a hydrophilic peptide moiety and a hydrophobic alkyl chain, forming a linear or cyclic structure. These lipopeptides are amphiphilic and can disrupt target cell membranes. Some examples include surfactin, fengycin, and iturin. Surfactins are mainly antibacterial and antiviral, while fengycin and iturin are antifungal. Polymyxin is one of the most used and well-studied cyclic lipopeptide produced by Paenibacillus polymyxa and P. alvei. Polymyxin can bind to the lipid A of lipopolysaccharide on the outer membrane of Gram-negative bacteria and destabilize the membrane, resulting in membrane permeabilization and cell lysis (Abriouel et al., 2011). In addition to their antimicrobial activities, lipopeptides have also been demonstrated to disrupt biofilm formation, cell motility, virulence expression, and other functions associated with plant defense and root colonization (Raaijmakers et al., 2010). On the other hand, siderophores function by sequestering and depleting iron from the environment and inhibit the proper function of other microorganisms in the niche. Bacillibactin is a common siderophore produced by Bacillus spp. and can efficiently chelate ferric iron and reduce its bioavailability, suppressing surrounding microorganisms (Caulier et al., 2019). Polyketides are a group of natural bioactive secondary metabolites with diverse structure and function. Polyketides are synthesized by multi-domain enzymes polyketide synthases (PKSs). PKSs are categorized into three groups: large and highly modular type I PKS, monofunctional type II PKS, and type III PKS with no acyl carrier protein domains (Ridley et al., 2008). Polyketides are produced by bacteria, fungi, and plants. They act as antibacterial and antifungal agents. The three types of antimicrobial polyketides produced by Bacillus spp. are bacillaene, difficidin, and 20 macrolactin (Caulier et al., 2019). Bacillaene is a polyene polyketide and exhibits inhibition against a variety of bacteria and fungi, including E. coli, B. thuringiensis, S. aureus and Fusarium. Difficidin is also a polyene polyketide and is active against bacterial pathogens including E. coli and C. perfringens. Macrolactin is both antibacterial and antifungal, and is active against E. coli, B. subtilis, S. aureus, Fusarium and others. The biosynthesis of NRPs and polyketides is highly similar, and hybrid NRPS- PKS gene clusters are widespread in bacterial and fungal genome (Wang et al., 2014). Peptide-polyketide hybrids have great structural diversity. Examples include antibiotic bacillaene, mycotoxin fusarin C, cyclic antifungal lipopeptide mycosubtilin, and paenilamicin (Aleti et al., 2015; Fisch, 2013; Van Lanen & Shen, 2006). These natural metabolites have a broad range of biological activities and enormous pharmaceutical potentials. Enzymes and other compounds Some other proteins synthesized by bacteria through ribosomes that exhibit antibacterial activities are cell wall degrading enzymes, like cellulase, glucanases, proteases and chitinases, and quorum quenching enzymes that disrupt bacterial quorum sensing, like lactonase, decarboxylase, acylase, and deaminase (Caulier et al., 2019). There are other small antimicrobial secondary metabolites produced by bacteria. Volatile organic compounds (VOCs) that containing sulfur or nitrogen, fatty acids and their derivatives, like benzenoid, terpene, and isoprenoid, are all important for the antimicrobial activities of these bacteria. 21 Food contamination is a constant threat to the public health and social- economic development around the world, and specific measures must be taken to improve food safety and quality throughout supply chain (Havelaar et al., 2015). Natural antimicrobials produced by food-grade microorganisms, like bacteriocins produced by lactic acid bacteria, are potentially safe for human consumption. These natural antimicrobials can be used as food additives to meet the consumer’s expectation of healthy, natural, and safe food products. Currently, nisin and pediocin PA-1 have been used extensively in the food industry to increase product shelf life. Many other enzymes, bacteriocins, nonribosomal peptides, and polyketides have the potential to be used as food preservatives. With the increase of bacterial resistance to conventional antibiotics and the growing consumer demands of natural foods with minimal processing and no added chemical preservatives, antibiotic-producing bacterial strains isolated from food sources and purified antimicrobials produced by these strains have great potentials for food industry applications. Continuous search for novel, safe, broad-spectrum antimicrobials from natural sources with stabilities across a wide range of pH, temperature, and other environmental conditions is necessary. 22 23 REFERENCES Abriouel, H., Franz, C. M. A. P., Omar, N. B., & Gálvez, A. (2011). Diversity and applications of Bacillus bacteriocins. FEMS Microbiology Reviews, 35(1), 201- 232. https://doi.org/10.1111/j.1574-6976.2010.00244.x Adcock, D. (1962). The Effect of Catalase on the Inhibine and Peroxide Values of Various Honeys. Journal of Apicultural Research, 1(1), 38-40. https://doi.org/10.1080/00218839.1962.11100047 Ahn, J.-Y., Min, J., Lee, S.-H., Jang, A., Park, C.-K., Kwon, S.-D., Park, S.-K., Lee, K., & Kim, Y.-H. (2014). Metagenomic analysis for identifying Kimchi sp. during the industrial-scale batch fermentation. Toxicology and Environmental Health Sciences, 6(1), 8-15. https://doi.org/10.1007/s13530-014-0182-0 al Somal, N., Coley, K. E., Molan, P. C., & Hancock, B. M. (1994). Susceptibility of Helicobacter pylori to the antibacterial activity of manuka honey. Journal of the Royal Society of Medicine, 87(1), 9-12. https://pubmed.ncbi.nlm.nih.gov/8308841 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1294271/ Al-Mamary, M., Al-Meeri, A., & Al-Habori, M. (2002). Antioxidant activities and total phenolics of different types of honey. Nutrition Research, 22(9), 1041- 1047. https://doi.org/https://doi.org/10.1016/S0271-5317(02)00406-2 Al-Waili, N. S., & Boni, N. S. (2003). Natural Honey Lowers Plasma Prostaglandin Concentrations in Normal Individuals. Journal of medicinal food, 6(2), 129- 133. https://doi.org/10.1089/109662003322233530 24 Aleti, G., Sessitsch, A., & Brader, G. (2015). Genome mining: Prediction of lipopeptides and polyketides from Bacillus and related Firmicutes. Computational and structural biotechnology journal, 13, 192-203. https://doi.org/10.1016/j.csbj.2015.03.003 Allsop, K. A., & Miller, J. B. (1996). Honey revisited: a reappraisal of honey in pre- industrial diets. British Journal of Nutrition, 75(4), 513-520. https://doi.org/10.1079/BJN19960155 Alvarez-Sieiro, P., Montalban-Lopez, M., Mu, D., & Kuipers, O. P. (2016). Bacteriocins of lactic acid bacteria: extending the family. Appl Microbiol Biotechnol, 100(7), 2939-2951. https://doi.org/10.1007/s00253-016-7343-9 Andjelković, U., Šrajer Gajdošik, M., Gašo-Sokač, D., Martinović, T., & Josić, D. (2017). Foodomics and Food Safety: Where We Are. Food Technol Biotechnol, 55(3), 290-307. https://doi.org/10.17113/ftb.55.03.17.5044 Anvarian, A. H. P., Cao, Y., Srikumar, S., Fanning, S., & Jordan, K. (2016). Flow Cytometric and 16S Sequencing Methodologies for Monitoring the Physiological Status of the Microbiome in Powdered Infant Formula Production [Methods]. Frontiers in microbiology, 7. https://doi.org/10.3389/fmicb.2016.00968 Azevedo, M. S., Valentim-Neto, P. A., Seraglio, S. K. T., da Luz, C. F. P., Arisi, A. C. M., & Costa, A. C. O. (2017). Proteome comparison for discrimination between honeydew and floral honeys from botanical species Mimosa scabrella Bentham by principal component analysis [https://doi.org/10.1002/jsfa.8317]. 25 Journal of the Science of Food and Agriculture, 97(13), 4515-4519. https://doi.org/https://doi.org/10.1002/jsfa.8317 Basualdo, C., Sgroy, V., Finola, M. S., & Marioli, J. M. (2007). Comparison of the antibacterial activity of honey from different provenance against bacteria usually isolated from skin wounds. Veterinary Microbiology, 124(3), 375-381. https://doi.org/https://doi.org/10.1016/j.vetmic.2007.04.039 Ben Sghaier, M., Skandrani, I., Nasr, N., Franca, M.-G. D., Chekir-Ghedira, L., & Ghedira, K. (2011). Flavonoids and sesquiterpenes from Tecurium ramosissimum promote antiproliferation of human cancer cells and enhance antioxidant activity: A structure–activity relationship study. Environmental Toxicology and Pharmacology, 32(3), 336-348. https://doi.org/https://doi.org/10.1016/j.etap.2011.07.003 Berg, G., Rybakova, D., Fischer, D., Cernava, T., Vergès, M.-C. C., Charles, T., Chen, X., Cocolin, L., Eversole, K., Corral, G. H., Kazou, M., Kinkel, L., Lange, L., Lima, N., Loy, A., Macklin, J. A., Maguin, E., Mauchline, T., McClure, R., . . . Schloter, M. (2020). Microbiome definition re-visited: old concepts and new challenges. Microbiome, 8(1), 103. https://doi.org/10.1186/s40168-020-00875- 0 Berríos, P., Fuentes, J. A., Salas, D., Carreño, A., Aldea, P., Fernández, F., & Trombert, A. N. (2018). Inhibitory effect of biofilm-forming Lactobacillus kunkeei strains against virulent Pseudomonas aeruginosa in vitro and in honeycomb moth (Galleria mellonella) infection model. Benef Microbes, 9(2), 257-268. https://doi.org/10.3920/bm2017.0048 26 Bilsel, Y., Bugra, D., Yamaner, S., Bulut, T., Cevikbas, U., & Turkoglu, U. (2002). Could Honey Have a Place in Colitis Therapy? Effects of Honey, Prednisolone, and Disulfiram on Inflammation, Nitric Oxide, and Free Radical Formation. Digestive Surgery, 19(4), 306-312. https://doi.org/10.1159/000064580 Boffo, E. F., Tavares, L. A., Tobias, A. C. T., Ferreira, M. M. C., & Ferreira, A. G. (2012). Identification of components of Brazilian honey by 1H NMR and classification of its botanical origin by chemometric methods. LWT, 49(1), 55- 63. https://doi.org/https://doi.org/10.1016/j.lwt.2012.04.024 Bogdanov, S., Jurendic, T., Sieber, R., & Gallmann, P. (2008). Honey for Nutrition and Health: A Review. Journal of the American College of Nutrition, 27(6), 677-689. https://doi.org/10.1080/07315724.2008.10719745 Bokulich, N. A., Kaehler, B. D., Rideout, J. R., Dillon, M., Bolyen, E., Knight, R., Huttley, G. A., & Gregory Caporaso, J. (2018). Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2's q2-feature- classifier plugin. Microbiome, 6(1), 90. https://doi.org/10.1186/s40168-018- 0470-z Bordoni, A., & Capozzi, F. (2015). The foodomics approach for discovering biomarkers of food consumption in nutrition studies. Current Opinion in Food Science, 4, 124-128. https://doi.org/https://doi.org/10.1016/j.cofs.2015.07.005 Borutinskaite, V., Treigyte, G., Čeksteryte, V., Kurtinaitiene, B., & Navakauskienė, R. (2018). Proteomic identification and enzymatic activity of buckwheat 27 (Fagopyrum esculentum) honey based on different assays. Journal of Food and Nutrition Research, 57. Bovo, S., Utzeri, V. J., Ribani, A., Cabbri, R., & Fontanesi, L. (2020). Shotgun sequencing of honey DNA can describe honey bee derived environmental signatures and the honey bee hologenome complexity. Sci Rep, 10(1), 9279. https://doi.org/10.1038/s41598-020-66127-1 Breukink, E., Wiedemann, I., Kraaij, C. v., Kuipers, O. P., Sahl, H. G., & de Kruijff, B. (1999). Use of the Cell Wall Precursor Lipid II by a Pore-Forming Peptide Antibiotic. Science, 286(5448), 2361-2364. https://doi.org/10.1126/science.286.5448.2361 Brown, K. L. (2000). Control of bacterial spores. British Medical Bulletin, 56(1), 158- 171. https://doi.org/10.1258/0007142001902860 Brudzynski, K. (2021). Honey as an Ecological Reservoir of Antibacterial Compounds Produced by Antagonistic Microbial Interactions in Plant Nectars, Honey and Honey Bee. Antibiotics (Basel), 10(5). https://doi.org/10.3390/antibiotics10050551 Brudzynski, K., & Sjaarda, C. P. (2021). Colloidal structure of honey and its influence on antibacterial activity [https://doi.org/10.1111/1541-4337.12720]. Comprehensive Reviews in Food Science and Food Safety, 20(2), 2063-2080. https://doi.org/https://doi.org/10.1111/1541-4337.12720 Cao, Y., Fanning, S., Proos, S., Jordan, K., & Srikumar, S. (2017). A Review on the Applications of Next Generation Sequencing Technologies as Applied to 28 Food-Related Microbiome Studies [Review]. Frontiers in microbiology, 8. https://doi.org/10.3389/fmicb.2017.01829 Capozzi, F., & Bordoni, A. (2013). Foodomics: a new comprehensive approach to food and nutrition. Genes & Nutrition, 8(1), 1-4. https://doi.org/10.1007/s12263-012-0310-x Caulier, S., Nannan, C., Gillis, A., Licciardi, F., Bragard, C., & Mahillon, J. (2019). Overview of the Antimicrobial Compounds Produced by Members of the Bacillus subtilis Group. Front Microbiol, 10, 302. https://doi.org/10.3389/fmicb.2019.00302 Cavanagh, D., Beazley, J., & Ostapowicz, F. (1970). Radical operation for carcinoma of the vulva. Journal of Obstetrics and Gynaecology of the British Commonwealth, 77(11), 1037-1040. Cavia, M. M., Fernáez-Muiño, M. A., Huidobro, J. F., & Sancho, M. T. (2004). Correlation between Moisture and Water Activity of Honeys Harvested in Different Years. Journal of Food Science, 69(5), C368-C370. https://doi.org/https://doi.org/10.1111/j.1365-2621.2004.tb10699.x Chaven, S. (2014). Chapter 11 - Honey, Confectionery and Bakery Products. In Y. Motarjemi & H. Lelieveld (Eds.), Food Safety Management (pp. 283-299). Academic Press. https://doi.org/https://doi.org/10.1016/B978-0-12-381504- 0.00011-1 Chua, L. S., Lee, J. Y., & Chan, G. F. (2013). Honey protein extraction and determination by mass spectrometry. Analytical and Bioanalytical Chemistry, 405(10), 3063-3074. https://doi.org/10.1007/s00216-012-6630-2 29 Claesson, M. J., Wang, Q., O'Sullivan, O., Greene-Diniz, R., Cole, J. R., Ross, R. P., & O'Toole, P. W. (2010). Comparison of two next-generation sequencing technologies for resolving highly complex microbiota composition using tandem variable 16S rRNA gene regions. Nucleic Acids Research, 38(22), e200-e200. https://doi.org/10.1093/nar/gkq873 Costa, P. A., Moraes, I. C. F., Bittante, A. M. Q., do Amaral Sobral, P. J., Gomide, C. A., & Carrer, C. C. (2013). Physical properties of honeys produced in the Northeast of Brazil. International Journal of Food Studies, 2(1). Crane, E. (1975). Honey: a comprehensive survey. D’Alessandro, A., & Zolla, L. (2012). We Are What We Eat: Food Safety and Proteomics. Journal of Proteome Research, 11(1), 26-36. https://doi.org/10.1021/pr2008829 de Almeida-Muradian, L. B., Stramm, K. M., & Estevinho, L. M. (2014). Efficiency of the FT-IR ATR spectrometry for the prediction of the physicochemical characteristics of Melipona subnitida honey and study of the temperature's effect on those properties. International Journal of Food Science & Technology, 49(1), 188-195. https://doi.org/https://doi.org/10.1111/ijfs.12297 de Arauz, L. J., Jozala, A. F., Mazzola, P. G., & Vessoni Penna, T. C. (2009). Nisin biotechnological production and application: a review. Trends in Food Science & Technology, 20(3), 146-154. https://doi.org/https://doi.org/10.1016/j.tifs.2009.01.056 De Filippis, F., Parente, E., & Ercolini, D. (2017). Metagenomics insights into food fermentations [https://doi.org/10.1111/1751-7915.12421]. Microbial 30 Biotechnology, 10(1), 91-102. https://doi.org/https://doi.org/10.1111/1751- 7915.12421 Delves-Broughton, J., Blackburn, P., Evans, R. J., & Hugenholtz, J. (1996). Applications of the bacteriocin, nisin. Antonie Van Leeuwenhoek, 69(2), 193- 202. https://doi.org/10.1007/bf00399424 Diep Dzung, B., Skaugen, M., Salehian, Z., Holo, H., & Nes Ingolf, F. (2007). Common mechanisms of target cell recognition and immunity for class II bacteriocins. Proceedings of the National Academy of Sciences, 104(7), 2384- 2389. https://doi.org/10.1073/pnas.0608775104 Dodd, H. M., Horn, N., & Gasson, M. J. (1990). Analysis of the Genetic Determinant for Production of the Peptide Antibiotic Nisin. Microbiology, 136(3), 555-556. https://doi.org/https://doi.org/10.1099/00221287-136-3-555 Dugat-Bony, E., Straub, C., Teissandier, A., Onésime, D., Loux, V., Monnet, C., Irlinger, F., Landaud, S., Leclercq-Perlat, M. N., Bento, P., Fraud, S., Gibrat, J. F., Aubert, J., Fer, F., Guédon, E., Pons, N., Kennedy, S., Beckerich, J. M., Swennen, D., & Bonnarme, P. (2015). Overview of a surface-ripened cheese community functioning by meta-omics analyses. PLoS One, 10(4), e0124360. https://doi.org/10.1371/journal.pone.0124360 Efem, S. E. (1993). Recent advances in the management of Fournier's gangrene: preliminary observations. Surgery, 113(2), 200-204. Ercolini, D. (2013). High-Throughput Sequencing and Metagenomics: Moving Forward in the Culture-Independent Analysis of Food Microbial Ecology. 31 Applied and Environmental Microbiology, 79(10), 3148-3155. https://doi.org/doi:10.1128/AEM.00256-13 Evans, B. S., Chen, Y., Metcalf, W. W., Zhao, H., & Kelleher, N. L. (2011). Directed evolution of the nonribosomal peptide synthetase AdmK generates new andrimid derivatives in vivo. Chemistry & biology, 18(5), 601-607. https://doi.org/10.1016/j.chembiol.2011.03.008 Ferreres, F., Ortiz, A., Silva, C., Garcia-Viguera, C., Tomás-Barberán, F. A., & Tomás-Lorente, F. (1992). Flavonoids of “La Alcarria” honey A study of their botanical origin. Zeitschrift für Lebensmittel-Untersuchung und Forschung, 194(2), 139-143. https://doi.org/10.1007/BF01190185 Fisch, K. M. (2013). Biosynthesis of natural products by microbial iterative hybrid PKS–NRPS [10.1039/C3RA42661K]. RSC Advances, 3(40), 18228-18247. https://doi.org/10.1039/C3RA42661K Fontana Jr., A. J. (2007). Appendix D: Minimum Water Activity Limits for Growth of Microorganisms. In Water Activity in Foods (pp. 405-405). https://doi.org/https://doi.org/10.1002/9780470376454.app4 García-Cañas, V., Simó, C., Herrero, M., Ibáñez, E., & Cifuentes, A. (2012). Present and Future Challenges in Food Analysis: Foodomics. Analytical Chemistry, 84(23), 10150-10159. https://doi.org/10.1021/ac301680q Grabowski, N. T., & Klein, G. (2017). Microbiology and foodborne pathogens in honey. Crit Rev Food Sci Nutr, 57(9), 1852-1862. https://doi.org/10.1080/10408398.2015.1029041 32 Hao, P., Zheng, H., Yu, Y., Ding, G., Gu, W., Chen, S., Yu, Z., Ren, S., Oda, M., Konno, T., Wang, S., Li, X., Ji, Z. S., & Zhao, G. (2011). Complete sequencing and pan-genomic analysis of Lactobacillus delbrueckii subsp. bulgaricus reveal its genetic basis for industrial yogurt production. PLoS One, 6(1), e15964. https://doi.org/10.1371/journal.pone.0015964 Havelaar, A. H., Kirk, M. D., Torgerson, P. R., Gibb, H. J., Hald, T., Lake, R. J., Praet, N., Bellinger, D. C., de Silva, N. R., Gargouri, N., Speybroeck, N., Cawthorne, A., Mathers, C., Stein, C., Angulo, F. J., Devleesschauwer, B., & World Health Organization Foodborne Disease Burden Epidemiology Reference, G. (2015). World Health Organization Global Estimates and Regional Comparisons of the Burden of Foodborne Disease in 2010. PLoS medicine, 12(12), e1001923-e1001923. https://doi.org/10.1371/journal.pmed.1001923 Herrero, M., Simó, C., García-Cañas, V., Ibáñez, E., & Cifuentes, A. (2012). Foodomics: MS-based strategies in modern food science and nutrition [https://doi.org/10.1002/mas.20335]. Mass Spectrometry Reviews, 31(1), 49- 69. https://doi.org/https://doi.org/10.1002/mas.20335 Iglesias, M. T., Martín-Álvarez, P. J., Polo, M. C., de Lorenzo, C., González, M., & Pueyo, E. (2006). Changes in the Free Amino Acid Contents of Honeys During Storage at Ambient Temperature. Journal of agricultural and food chemistry, 54(24), 9099-9104. https://doi.org/10.1021/jf061712x Jung, J. Y., Lee, S. H., Jin, H. M., Hahn, Y., Madsen, E. L., & Jeon, C. O. (2013). Metatranscriptomic analysis of lactic acid bacterial gene expression during 33 kimchi fermentation. International Journal of Food Microbiology, 163(2), 171-179. https://doi.org/https://doi.org/10.1016/j.ijfoodmicro.2013.02.022 Kačániová, M., Kňazovická, V., Felšöciová, S., & Rovná, K. (2012). Microscopic fungi recovered from honey and their toxinogenity. J Environ Sci Health A Tox Hazard Subst Environ Eng, 47(11), 1659-1664. https://doi.org/10.1080/10934529.2012.687242 Kafantaris, I., Amoutzias, G. D., & Mossialos, D. (2021). Foodomics in bee product research: a systematic literature review. European Food Research and Technology, 247(2), 309-331. https://doi.org/10.1007/s00217-020-03634-5 Khan, H., Flint, S. H., & Yu, P. L. (2013). Determination of the mode of action of enterolysin A, produced by Enterococcus faecalis B9510 [https://doi.org/10.1111/jam.12240]. Journal of Applied Microbiology, 115(2), 484-494. https://doi.org/https://doi.org/10.1111/jam.12240 Kiss, A., Balikó, G., Csorba, A., Chuluunbaatar, T., Medzihradszky, K. F., & Alföldi, L. (2008). Cloning and characterization of the DNA region responsible for Megacin A-216 production in Bacillus megaterium 216. Journal of Bacteriology, 190(19), 6448-6457. https://doi.org/10.1128/JB.00557-08 Kňazovická, V., Gábor, M., Miluchová, M., Bobko, M., & Medo, J. (2020). Diversity of bacteria in Slovak and foreign honey, with assessment of its physico- chemical quality and counts of cultivable microorganisms. Journal of Microbiology, Biotechnology and Food Sciences, 9(6), 414-421. Ko, D. K., Nadakuduti, S. S., Douches, D. S., & Buell, C. R. (2018). Transcriptome profiling of transgenic potato plants provides insights into variability caused by 34 plant transformation. PLoS One, 13(11), e0206055. https://doi.org/10.1371/journal.pone.0206055 Lamas, A., Regal, P., Vázquez, B., Miranda, J. M., Franco, C. M., & Cepeda, A. (2019). Transcriptomics: A powerful tool to evaluate the behavior of foodborne pathogens in the food production chain. Food Res Int, 125, 108543. https://doi.org/10.1016/j.foodres.2019.108543 Lancova, K., Dip, R., Antignac, J.-P., Bizec, B. L., Elliott, C. T., & Naegeli, H. (2011). Detection of hazardous food contaminants by transcriptomics fingerprinting. TrAC Trends in Analytical Chemistry, 30(2), 181-191. https://doi.org/https://doi.org/10.1016/j.trac.2010.10.013 Lazaridou, A., Biliaderis, C. G., Bacandritsos, N., & Sabatini, A. G. (2004). Composition, thermal and rheological behaviour of selected Greek honeys. Journal of Food Engineering, 64(1), 9-21. https://doi.org/https://doi.org/10.1016/j.jfoodeng.2003.09.007 Lee, H., Churey, J. J., & Worobo, R. W. (2008a). Antimicrobial activity of bacterial isolates from different floral sources of honey. Int J Food Microbiol, 126(1-2), 240-244. https://doi.org/10.1016/j.ijfoodmicro.2008.04.030 Lee, H., Churey, J. J., & Worobo, R. W. (2008b). Purification and structural characterization of bacillomycin F produced by a bacterial honey isolate active against Byssochlamys fulva H25. J Appl Microbiol, 105(3), 663-673. https://doi.org/10.1111/j.1365-2672.2008.03797.x 35 Leistner, L., & Gorris, L. G. M. (1995). Food preservation by hurdle technology. Trends in Food Science & Technology, 6(2), 41-46. https://doi.org/https://doi.org/10.1016/S0924-2244(00)88941-4 Lessard, M.-H., Viel, C., Boyle, B., St-Gelais, D., & Labrie, S. (2014). Metatranscriptome analysis of fungal strains Penicillium camemberti and Geotrichum candidumreveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese. BMC genomics, 15(1), 235. https://doi.org/10.1186/1471-2164-15-235 Lorenzo, J. M., Munekata, P. E., Dominguez, R., Pateiro, M., Saraiva, J. A., & Franco, D. (2018). Main Groups of Microorganisms of Relevance for Food Safety and Stability: General Aspects and Overall Description. Innovative Technologies for Food Preservation, 53-107. https://doi.org/10.1016/B978-0-12-811031- 7.00003-0 Machado De-Melo, A. A., Almeida-Muradian, L. B. d., Sancho, M. T., & Pascual- Maté, A. (2018). Composition and properties of Apis mellifera honey: A review. Journal of Apicultural Research, 57(1), 5-37. https://doi.org/10.1080/00218839.2017.1338444 Maddocks, S. E., & Jenkins, R. E. (2013). Honey: a sweet solution to the growing problem of antimicrobial resistance? Future Microbiology, 8(11), 1419-1429. https://doi.org/10.2217/fmb.13.105 Madejczyk, M., & Baralkiewicz, D. (2008). Characterization of Polish rape and honeydew honey according to their mineral contents using ICP-MS and F- 36 AAS/AES. Analytica Chimica Acta, 617(1), 11-17. https://doi.org/https://doi.org/10.1016/j.aca.2008.01.038 Mato, I., Huidobro, J. F., Simal-Lozano, J., & Sancho, M. T. (2003). Significance of nonaromatic organic acids in honey. J Food Prot, 66(12), 2371-2376. https://doi.org/10.4315/0362-028x-66.12.2371 Mato, I., Huidobro, J. F., Simal-Lozano, J., & Sancho, M. T. (2006). Rapid Determination of Nonaromatic Organic Acids in Honey by Capillary Zone Electrophoresis with Direct Ultraviolet Detection. Journal of agricultural and food chemistry, 54(5), 1541-1550. https://doi.org/10.1021/jf051757i Molan, P. C. (1992). The antibacterial activity of honey: 1. The nature of the antibacterial activity. Bee World, 73(1), 5-28. Molan, P. C. (1999). Why honey is effective as a medicine. 1. Its use in modern medicine. Bee World, 80(2), 80-92. https://doi.org/10.1080/0005772X.1999.11099430 Molan, P. C., Smith, I. M., & Reid, G. M. (1988). A Comparison of the Antibacterial Activities of Some new Zealand Honeys. Journal of Apicultural Research, 27(4), 252-256. https://doi.org/10.1080/00218839.1988.11100811 Moll, G. N., Konings, W. N., & Driessen, A. J. (1999). Bacteriocins: mechanism of membrane insertion and pore formation. Antonie Van Leeuwenhoek, 76(1-4), 185-198. Monnet, C., Dugat-Bony, E., Swennen, D., Beckerich, J. M., Irlinger, F., Fraud, S., & Bonnarme, P. (2016). Investigation of the Activity of the Microorganisms in a 37 Reblochon-Style Cheese by Metatranscriptomic Analysis. Front Microbiol, 7, 536. https://doi.org/10.3389/fmicb.2016.00536 Mootz, H. D., Schwarzer, D., & Marahiel, M. A. (2002). Ways of Assembling Complex Natural Products on Modular Nonribosomal Peptide Synthetases [https://doi.org/10.1002/1439-7633(20020603)3:6<490::AID- CBIC490>3.0.CO;2-N]. ChemBioChem, 3(6), 490-504. https://doi.org/https://doi.org/10.1002/1439-7633(20020603)3:6<490::AID- CBIC490>3.0.CO;2-N Mundo, M. A., Padilla-Zakour, O. I., & Worobo, R. W. (2004). Growth inhibition of foodborne pathogens and food spoilage organisms by select raw honeys. Int J Food Microbiol, 97(1), 1-8. https://doi.org/10.1016/j.ijfoodmicro.2004.03.025 Neefs, J.-M., Van de Peer, Y., De Rijk, P., Chapelle, S., & De Wachter, R. (1993). Compilation of small ribosomal subunit RNA structures. Nucleic Acids Research, 21(13), 3025-3049. https://doi.org/10.1093/nar/21.13.3025 Nychas, G. J. E., & Panagou, E. (2011). 1 - Microbiological spoilage of foods and beverages. In D. Kilcast & P. Subramaniam (Eds.), Food and Beverage Stability and Shelf Life (pp. 3-28). Woodhead Publishing. https://doi.org/https://doi.org/10.1533/9780857092540.1.3 Obaseiki-Ebor, E. E., & Afonya, T. C. (1984). In-vitro evaluation of the anticandidiasis activity of honey distillate (HY-1) compared with that of some antimycotic agents. J Pharm Pharmacol, 36(4), 283-284. https://doi.org/10.1111/j.2042-7158.1984.tb04373.x 38 Obi, C. L., Ugoji, E. O., Edun, S. A., Lawal, S. F., & Anyiwo, C. E. (1994). The antibacterial effect of honey on diarrhoea causing bacterial agents isolated in Lagos, Nigeria. Afr J Med Med Sci, 23(3), 257-260. Ojeda de Rodrı́guez, G., Sulbarán de Ferrer, B., Ferrer, A., & Rodrı́guez, B. (2004). Characterization of honey produced in Venezuela. Food Chemistry, 84(4), 499-502. https://doi.org/https://doi.org/10.1016/S0308-8146(02)00517-4 Olaitan, P. B., Adeleke, O. E., & Ola, I. O. (2007). Honey: a reservoir for microorganisms and an inhibitory agent for microbes. Afr Health Sci, 7(3), 159-165. https://doi.org/10.5555/afhs.2007.7.3.159 Olas, B. (2020). Honey and Its Phenolic Compounds as an Effective Natural Medicine for Cardiovascular Diseases in Humans? Nutrients, 12(2), 283. https://doi.org/10.3390/nu12020283 Olofsson, T. C., & Vásquez, A. (2008). Detection and identification of a novel lactic acid bacterial flora within the honey stomach of the honeybee Apis mellifera. Curr Microbiol, 57(4), 356-363. https://doi.org/10.1007/s00284-008-9202-0 Ouchemoukh, S., Schweitzer, P., Bachir Bey, M., Djoudad-Kadji, H., & Louaileche, H. (2010). HPLC sugar profiles of Algerian honeys. Food Chemistry, 121(2), 561-568. https://doi.org/https://doi.org/10.1016/j.foodchem.2009.12.047 Paramás, A. M. G., Bárez, J. A. G., Marcos, C. C., García-Villanova, R. J., & Sánchez, J. S. (2006). HPLC-fluorimetric method for analysis of amino acids in products of the hive (honey and bee-pollen). Food Chemistry, 95(1), 148- 156. https://doi.org/https://doi.org/10.1016/j.foodchem.2005.02.008 39 Petruzzi, L., Corbo, M. R., Sinigaglia, M., & Bevilacqua, A. (2017). Chapter 1 - Microbial Spoilage of Foods: Fundamentals. In A. Bevilacqua, M. R. Corbo, & M. Sinigaglia (Eds.), The Microbiological Quality of Food (pp. 1-21). Woodhead Publishing. https://doi.org/https://doi.org/10.1016/B978-0-08- 100502-6.00002-9 Połka, J., Rebecchi, A., Pisacane, V., Morelli, L., & Puglisi, E. (2015). Bacterial diversity in typical Italian salami at different ripening stages as revealed by high-throughput sequencing of 16S rRNA amplicons. Food Microbiology, 46, 342-356. https://doi.org/https://doi.org/10.1016/j.fm.2014.08.023 Postmes, T. (2001). The treatment of burns and other wounds with honey. Honey and healing, ed. P. Munn and R Jones. International Bee Research Association, Cardiff, UK. Pucci, M. J., Vedamuthu, E. R., Kunka, B. S., & Vandenbergh, P. A. (1988). Inhibition of Listeria monocytogenes by using bacteriocin PA-1 produced by Pediococcus acidilactici PAC 1.0. Appl Environ Microbiol, 54(10), 2349-2353. https://doi.org/10.1128/aem.54.10.2349-2353.1988 Quigley, L., O'Sullivan, O., Beresford, T. P., Ross, R. P., Fitzgerald, G. F., & Cotter, P. D. (2012). High-throughput sequencing for detection of subpopulations of bacteria not previously associated with artisanal cheeses. Appl Environ Microbiol, 78(16), 5717-5723. https://doi.org/10.1128/aem.00918-12 Raaijmakers, J. M., De Bruijn, I., Nybroe, O., & Ongena, M. (2010). Natural functions of lipopeptides from Bacillus and Pseudomonas: more than surfactants and 40 antibiotics. FEMS Microbiol Rev, 34(6), 1037-1062. https://doi.org/10.1111/j.1574-6976.2010.00221.x Radwan, S. S., El-Essawy, A. A., & Sarhan, M. M. (1984). Experimental evidence for the occurrence in honey of specific substances active against microorganisms. Zentralbl Mikrobiol, 139(4), 249-255. Ramos, O. Y., Basualdo, M., Libonatti, C., & Vega, M. F. (2020). Current status and application of lactic acid bacteria in animal production systems with a focus on bacteria from honey bee colonies [https://doi.org/10.1111/jam.14469]. Journal of Applied Microbiology, 128(5), 1248-1260. https://doi.org/https://doi.org/10.1111/jam.14469 Razali, M. T. A., Zainal, Z. A., Maulidiani, M., Shaari, K., Zamri, Z., Mohd Idrus, M. Z., Khatib, A., Abas, F., Ling, Y. S., Rui, L. L., & Ismail, I. S. (2018). Classification of Raw Stingless Bee Honeys by Bee Species Origins Using the NMR- and LC-MS-Based Metabolomics Approach. Molecules, 23(9), 2160. https://www.mdpi.com/1420-3049/23/9/2160 Rešetar, D., Pavelić, S. K., & Josić, D. (2015). Foodomics for investigations of food toxins. Current Opinion in Food Science, 4, 86-91. https://doi.org/https://doi.org/10.1016/j.cofs.2015.05.004 Rezzi, S., Ramadan, Z., Fay, L. B., & Kochhar, S. (2007). Nutritional Metabonomics:  Applications and Perspectives. Journal of Proteome Research, 6(2), 513-525. https://doi.org/10.1021/pr060522z 41 Ridley, C. P., Lee, H. Y., & Khosla, C. (2008). Evolution of polyketide synthases in bacteria. Proceedings of the National Academy of Sciences of the United States of America, 105(12), 4595-4600. https://doi.org/10.1073/pnas.0710107105 Rodrigues, M. X., Lima, S. F., Canniatti-Brazaca, S. G., & Bicalho, R. C. (2017). The microbiome of bulk tank milk: Characterization and associations with somatic cell count and bacterial count. Journal of Dairy Science, 100(4), 2536-2552. https://doi.org/10.3168/jds.2016-11540 Rodríguez-Andrade, E., Stchigel, A. M., Terrab, A., Guarro, J., & Cano-Lira, J. F. (2019). Diversity of xerotolerant and xerophilic fungi in honey. IMA fungus, 10(1), 20. https://doi.org/10.1186/s43008-019-0021-7 Rossano, R., Larocca, M., Polito, T., Perna, A. M., Padula, M. C., Martelli, G., & Riccio, P. (2012). What Are the Proteolytic Enzymes of Honey and What They Do Tell Us? A Fingerprint Analysis by 2-D Zymography of Unifloral Honeys. PLoS One, 7(11), e49164. https://doi.org/10.1371/journal.pone.0049164 Roy, N. S., Kim, J.-A., Choi, A.-Y., Ban, Y.-W., Park, N.-I., Park, K.-C., Yang, H.-S., Choi, I.-Y., & Kim, S. (2018). RNA-Seq De Novo Assembly and Differential Transcriptome Analysis of Korean Medicinal Herb Cirsium japonicum var. spinossimum. Genomics & informatics, 16(4), e34-e34. https://doi.org/10.5808/GI.2018.16.4.e34 Ruiz-Matute, A. I., Brokl, M., Soria, A. C., Sanz, M. L., & Martínez-Castro, I. (2010). Gas chromatographic–mass spectrometric characterisation of tri- and tetrasaccharides in honey. Food Chemistry, 120(2), 637-642. https://doi.org/https://doi.org/10.1016/j.foodchem.2009.10.050 42 Samarghandian, S., Farkhondeh, T., & Samini, F. (2017). Honey and Health: A Review of Recent Clinical Research. Pharmacognosy research, 9(2), 121-127. https://doi.org/10.4103/0974-8490.204647 Sanz, S., Gradillas, G., Jimeno, F., Perez, C., & Juan, T. (1995). Fermentation Problem in Spanish North-Coast Honey. J Food Prot, 58(5), 515-518. https://doi.org/10.4315/0362-028x-58.5.515 Schievano, E., Peggion, E., & Mammi, S. (2010). 1H Nuclear Magnetic Resonance Spectra of Chloroform Extracts of Honey for Chemometric Determination of Its Botanical Origin. Journal of agricultural and food chemistry, 58(1), 57-65. https://doi.org/10.1021/jf9022977 Schievano, E., Stocchero, M., Morelato, E., Facchin, C., & Mammi, S. (2012). An NMR-based metabolomic approach to identify the botanical origin of honey. Metabolomics, 8(4), 679-690. https://doi.org/10.1007/s11306-011-0362-8 Schlundt, J., Tay, M. Y. F., Chengcheng, H., & Liwei, C. (2020). Food Security: Microbiological and Chemical Risks. In A. J. Masys, R. Izurieta, & M. Reina Ortiz (Eds.), Global Health Security: Recognizing Vulnerabilities, Creating Opportunities (pp. 231-274). Springer International Publishing. https://doi.org/10.1007/978-3-030-23491-1_11 Scholz, M., Ward, D. V., Pasolli, E., Tolio, T., Zolfo, M., Asnicar, F., Truong, D. T., Tett, A., Morrow, A. L., & Segata, N. (2016). Strain-level microbial epidemiology and population genomics from shotgun metagenomics. Nat Methods, 13(5), 435-438. https://doi.org/10.1038/nmeth.3802 43 Sharma, S. (2015). Food preservatives and their harmful effects. International journal of scientific and research publications, 5(4), 1-2. Siezen, R. J., Starrenburg, M. J., Boekhorst, J., Renckens, B., Molenaar, D., & van Hylckama Vlieg, J. E. (2008). Genome-scale genotype-phenotype matching of two Lactococcus lactis isolates from plants identifies mechanisms of adaptation to the plant niche. Appl Environ Microbiol, 74(2), 424-436. https://doi.org/10.1128/aem.01850-07 Silva, C. C. G., Silva, S. P. M., & Ribeiro, S. C. (2018). Application of Bacteriocins and Protective Cultures in Dairy Food Preservation [Review]. Frontiers in microbiology, 9. https://doi.org/10.3389/fmicb.2018.00594 Simmonds, R. S., Pearson, L., Kennedy, R. C., & Tagg, J. R. (1996). Mode of action of a lysostaphin-like bacteriolytic agent produced by Streptococcus zooepidemicus 4881. Applied and Environmental Microbiology, 62(12), 4536- 4541. https://doi.org/doi:10.1128/aem.62.12.4536-4541.1996 Singh, B. P., Rateb, M. E., Rodriguez-Couto, S., Polizeli, M. d. L. T. d. M., & Li, W.- J. (2019). Editorial: Microbial Secondary Metabolites: Recent Developments and Technological Challenges [Editorial]. Frontiers in microbiology, 10. https://doi.org/10.3389/fmicb.2019.00914 Stevens, K. A., Sheldon, B. W., Klapes, N. A., & Klaenhammer, T. R. (1991). Nisin treatment for inactivation of Salmonella species and other gram-negative bacteria. Applied and Environmental Microbiology, 57(12), 3613-3615. https://doi.org/doi:10.1128/aem.57.12.3613-3615.1991 44 Subrahmanyam, M. (2005). Topical application of honey in treatment of burns. British Journal of Surgery, 78(4), 497-498. https://doi.org/10.1002/bjs.1800780435 Telias, A., White, J. R., Pahl, D. M., Ottesen, A. R., & Walsh, C. S. (2011). Bacterial community diversity and variation in spray water sources and the tomato fruit surface. BMC microbiology, 11(1), 81. https://doi.org/10.1186/1471-2180-11- 81 Terrab, A., Dı́ez, M. J., & Heredia, F. J. (2002). Characterisation of Moroccan unifloral honeys by their physicochemical characteristics. Food Chemistry, 79(3), 373-379. https://doi.org/https://doi.org/10.1016/S0308-8146(02)00189-9 Thakali, A., & MacRae, J. D. (2021). A review of chemical and microbial contamination in food: What are the threats to a circular food system? Environmental Research, 194, 110635. https://doi.org/https://doi.org/10.1016/j.envres.2020.110635 Trasande, L., Shaffer, R. M., Sathyanarayana, S., & Council On Environmental, H. (2018). Food Additives and Child Health. Pediatrics, 142(2), e20181410. https://doi.org/10.1542/peds.2018-1410 Truchado, P., López-Gálvez, F., Gil, M. I., Tomás-Barberán, F. A., & Allende, A. (2009). Quorum sensing inhibitory and antimicrobial activities of honeys and the relationship with individual phenolics. Food Chemistry, 115(4), 1337- 1344. https://doi.org/https://doi.org/10.1016/j.foodchem.2009.01.065 Ustunol, Z., & Gandhi, H. (2001). Growth and viability of commercial Bifidobacterium spp. in honey-sweetened skim milk. J Food Prot, 64(11), 1775-1779. https://doi.org/10.4315/0362-028x-64.11.1775 45 Val, A., Huidobro, J. F., Sánchez, M. P., Muniategui, S., Fernández-Muiño, M. A., & Sancho, M. T. (1998). Enzymatic Determination of Galactose and Lactose in Honey. Journal of agricultural and food chemistry, 46(4), 1381-1385. https://doi.org/10.1021/jf970483z van Dyk, B. N., de Bruin, W., du Plessis, E. M., & Korsten, L. (2016). Microbiological Food Safety Status of Commercially Produced Tomatoes from Production to Marketing. Journal of Food Protection, 79(3), 392-406. https://doi.org/10.4315/0362-028X.JFP-15-300 van Heusden, H. E., de Kruijff, B., & Breukink, E. (2002). Lipid II Induces a Transmembrane Orientation of the Pore-Forming Peptide Lantibiotic Nisin. Biochemistry, 41(40), 12171-12178. https://doi.org/10.1021/bi026090x Van Lanen, S. G., & Shen, B. (2006). Progress in combinatorial biosynthesis for drug discovery. Drug Discovery Today: Technologies, 3(3), 285-292. https://doi.org/https://doi.org/10.1016/j.ddtec.2006.09.014 Villalobos-Delgado, L. H., Nevárez-Moorillon, G. V., Caro, I., Quinto, E. J., & Mateo, J. (2019). 4 - Natural antimicrobial agents to improve foods shelf life. In C. M. Galanakis (Ed.), Food Quality and Shelf Life (pp. 125-157). Academic Press. https://doi.org/https://doi.org/10.1016/B978-0-12-817190-5.00004-5 Wang, H., Fewer, D. P., Holm, L., Rouhiainen, L., & Sivonen, K. (2014). Atlas of nonribosomal peptide and polyketide biosynthetic pathways reveals common occurrence of nonmodular enzymes. Proceedings of the National Academy of Sciences, 111(25), 9259-9264. https://doi.org/doi:10.1073/pnas.1401734111 46 Wen, Y., Wang, L., Jin, Y., Zhang, J., Su, L., Zhang, X., Zhou, J., & Li, Y. (2017). The Microbial Community Dynamics during the Vitex Honey Ripening Process in the Honeycomb [Original Research]. Frontiers in microbiology, 8. https://doi.org/10.3389/fmicb.2017.01649 White, J. W., Subers, M. H., & Schepartz, A. I. (1963). The identification of inhibine, the antibacterial factor in honey, as hydrogen peroxide and its origin in a honey glucose-oxidase system. Biochimica et Biophysica Acta (BBA) - Specialized Section on Enzymological Subjects, 73(1), 57-70. https://doi.org/https://doi.org/10.1016/0926-6569(63)90108-1 Wiedemann, I., Breukink, E., van Kraaij, C., Kuipers, O. P., Bierbaum, G., de Kruijff, B., & Sahl, H.-G. (2001). Specific Binding of Nisin to the Peptidoglycan Precursor Lipid II Combines Pore Formation and Inhibition of Cell Wall Biosynthesis for Potent Antibiotic Activity *. Journal of Biological Chemistry, 276(3), 1772-1779. https://doi.org/10.1074/jbc.M006770200 Williams, T. R., Moyne, A.-L., Harris, L. J., & Marco, M. L. (2013). Season, Irrigation, Leaf Age, and Escherichia coli Inoculation Influence the Bacterial Diversity in the Lettuce Phyllosphere. PLoS One, 8(7), e68642. https://doi.org/10.1371/journal.pone.0068642 Won, S.-R., Li, C.-Y., Kim, J.-W., & Rhee, H.-I. (2009). Immunological characterization of honey major protein and its application. Food Chemistry, 113(4), 1334-1338. https://doi.org/https://doi.org/10.1016/j.foodchem.2008.08.082 47 Yun, J. W. (1996). Fructooligosaccharides—Occurrence, preparation, and application. Enzyme and Microbial Technology, 19(2), 107-117. https://doi.org/https://doi.org/10.1016/0141-0229(95)00188-3 Zhang, Y.-Z., Chen, Y.-F., Wu, Y.-Q., Si, J.-J., Zhang, C.-P., Zheng, H.-Q., & Hu, F.- L. (2019). Discrimination of the entomological origin of honey according to the secretions of the bee (Apis cerana or Apis mellifera). Food Research International, 116, 362-369. https://doi.org/https://doi.org/10.1016/j.foodres.2018.08.049 Zhong, Y., Wu, L., Chen, X., Huang, Z., & Hu, W. (2018). Effects of Food-Additive- Information on Consumers' Willingness to Accept Food with Additives. International journal of environmental research and public health, 15(11), 2394. https://doi.org/10.3390/ijerph15112394 48 CHAPTER 2 MICROBIOME ANALYSIS OF RAW HONEY REVEALS IMPORTANT FACTORS INFLUENCING THE BACTERIAL AND FUNGAL COMMUNITIES Abstract Raw honeys contain diverse microbial communities. Previous studies have focused on isolating bacteria and fungi that are culturable, while missing a large proportion of the microbial community due to culture-based constraints. This study utilized next- generation sequencing (NGS) to analyze the composition of microorganisms in raw honey; these data can reveal environmental and physicochemical variables that are associated with different microbial communities. To examine the microbial composition (bacteria and fungi) of raw honey and analyze its association with physicochemical properties, four types of honey (monofloral, wildflower, manuka, and feral; ntotal = 36) were analyzed via amplicon metagenomics. The analyzed honey samples had relatively similar bacterial communities but more distinct and diverse fungal communities. Honey type was determined as a significant factor influencing alpha and beta diversity metrics of bacterial and fungal communities. For the bacterial communities, titratable acidity (TA) was associated with community richness and diversity. For the fungal communities, Brix, TA, and color were associated with community richness, while water activity and color were associated with community diversity. Additionally, important bacterial and fungal amplicon sequence variants (ASVs) that influenced the overall community were identified. Results from this study provide important insights into the microbial communities associated with different types of raw honey, which could improve our understanding of microbial dynamics in 4 9 beehives, improve honey production, and prevent honeybee disease. 50 Introduction Honey has a diverse microbiome, most of which originates from pollen, flowers, soil, air, dust, and the honeybee digestive tract (Snowdon & Cliver, 1996). Additionally, some secondary microbial contaminants may be introduced into honey during human processing (Snowdon & Cliver, 1996). Honey has a water activity between 0.50 - 0.65. It is generally acidic, with pH ranging from 3 - 5 due to the presence of organic acids like gluconic acid (Balzan et al., 2020; Olaitan et al., 2007). The physicochemical properties of honey have an influence on the microbial communities. The low water activity, low pH, and antimicrobial components (including hydrogen peroxide, antioxidants, and antimicrobial peptides) of honey inhibit the growth of vegetative bacterial cells (Olaitan et al., 2007). Few organisms can survive the osmotic stress of honey; those that do are mainly spore-forming bacteria and yeasts. Previous studies found osmotolerant bacteria that were transmitted to honey from flower nectar through bee pollination (Álvarez-Pérez et al., 2012; Fridman et al., 2012). Honey-associated microorganisms can be grouped into three types based on origin and ecological niche: bee gut microorganisms, bee pathogens, and plant- associated microorganisms (Bovo et al., 2018). Lactobacillus and Bifidobacterium lactic acid bacteria (LAB) are major components of the bee gut microbiome and are relatively conserved in honeybee digestive tracts globally (Anderson et al., 2013; Raymann & Moran, 2018). These genera have been found in bee-collected nectar and honey (Olofsson & Vásquez, 2008). Up to 108 CFU per gram of viable LAB have been found in different honey samples (Vásquez et al., 2012). A few other bacterial 51 genera are frequently, though not ubiquitously, found in honeybee digestive tracts; these include Apibacter, Acetobacter, and Asaia. Some rarer bacteria that can cause disease in and death of honeybees may also be found in honeybee digestive tracts; these include Enterobacter, Klebsiella, Citrobacter, and Serratia (Raymann & Moran, 2018). Fungal genera found in honeybee digestive tracts include Saccharomyces, Zygosaccharomyces, and Candida (Yun et al., 2018). Overall honeybee health can be threatened by bacterial and fungal pathogens, which may contribute to colony collapses (Schwarz et al., 2015). Common bacterial pathogens include Melissococcus, Paenibacillus, and Spiroplasma. Fungal pathogens for honeybees include Ascosphaera, Aspergillus, and Nosema. As a common mold found in the environment, some Aspergillus spp. are opportunistic pathogens that can infect honeybee larvae and cause stonebrood disease. The common chalkbrood disease is caused by Ascosphaera apis, while nosema disease is caused by spore-forming fungi Nosema apis and Nosema ceranae (Jensen et al., 2013; Schwarz et al., 2015). Bacteria and fungi that are commonly found in plants and soil can be transmitted to beehives through pollination. These plant-associated microorganisms are present in honey and other bee products like bee bread (a fermented mixture of pollen and nectar used as food for bees), and some of these microorganisms are beneficial to the bee colonies (Kurek-Górecka et al., 2020). One example is Actinobacteria. Even though some Actinobacteria spp. are plant pathogens, many of them are protective microbes for honeybees and other insects because they produce secondary metabolites which prevent fungal growth and spoilage (Anderson et al., 2013; Barke et al., 2010; Mohr & Tebbe, 2006). A variety of Enterobacteriaceae and Firmicutes were found in flowers 52 including Lactobacillus, Bacillus, and Weissella spp., many of which are present in honeybee digestive tracts and honey products. Lactobacillus kuneei has been found in flowers, honeybee gut, and bee bread (Anderson et al., 2013). The ubiquitous presence of LAB across different bee species is the result of horizontal transmission between beehives and environment. The high similarity between Firmicutes found in flower nectar and those isolated from honeybee hives is further indication that horizontal transmission of these bacteria happens through honeybee pollination (Vásquez et al., 2012). The plant-associated bacteria Paenibacillus spp. are commonly found in soil. Some species of Paenibacillus are bee pathogens: P. larvae is the causative agent for American foulbrood disease; P. alvei is commonly found as a secondary invader of European foulbrood disease caused by Melissococcus plutonius (Genersch, 2010). As for the common plant-associated fungi found in honey, Cladosporium is a filamentous fungus that is common in the environment, and some species are potential plant pathogens (Bensch et al., 2012). It was proposed that Cladosporium could cohabit with bees and transmit from plant or bees to persist in bee products (E. O. Martinson et al., 2012). Other filamentous fungi that are commonly found in plant pollen include Botrytis, Penicillium, and Mucor, which are transmitted to honeybees and frequently found in bee bread (Disayathanoowat et al., 2020). Some common genera of yeast that were isolated from pollen and bee bread include Candida, Cryptococcus, Kloeckera, Metschnikowia, and Rhodotorula (Gilliam et al., 1974). Flower-derived microorganisms are subjected to environmental changes, which in turn contribute to the variation, growth, and secondary metabolite production of other environment- derived microorganisms in honeybees (Vásquez et al., 2012). 53 As previous studies suggest, microbial and honeybee DNA present in honey reflect the hive microbiome and honeybee hologenome; these data may reveal the bee pathosphere and indicate overall bee colony health (Bovo et al., 2020). Analyzing the honey microbiome can potentially help in the understanding of microbial hive dynamics, which may improve honey production and prevent honeybee diseases. However, most previous honey microbiome studies use traditional culture-based methods to isolate and identify microorganisms in honey, which is subject to culture biases (Anderson et al., 2013). Using culture-independent methods to investigate the microbiome of honey avoids biases induced by researcher-selected growth conditions. Recent studies have used next-generation sequencing (NGS) methods to study the microbiome of honeybee gastrointestinal tracts, pollen, and bee bread, while metagenomic analyses of honey are limited (Disayathanoowat et al., 2020; Engel et al., 2012; Jones et al., 2018; Moran et al., 2012; Powell et al., 2014; Yun et al., 2018). In our study, we used 16S and ITS metabarcoding method to evaluate and compare the microbiomes of raw honey derived from different sources. We selected monofloral honey and multifloral honey from central NY region. To compare the differences between different honey types, we also chose to include two special types, manuka honey and feral honey, that have not been studied previously and could potentially have distinct and interesting microbial communities. Manuka honey is a highly valuable New Zealand monofloral honey with antimicrobial and antioxidant capabilities (Niaz et al., 2017). The high content of antioxidants could be produced by certain microorganisms in honey, which would deter the growth of other microorganisms in the environmental niche (Brudzynski, 2021). Feral honey is 54 produced by domesticated western honey bees Apis mellifera that swarmed and established wild colonies (Hinshaw et al., 2021). Feral colonies are able to survive in the wild without human management and develop mechanisms to defend against varroa mites and other pathogens (Youngsteadt et al., 2015). The microbiome of feral honeybees is potentially associated with the strong immune systems and mite survival strategies of these bee colonies, which could potentially be reflected in the honey. Previous studies have reported the association between honey microbiomes and parameters like moisture, electrical conductivity, and botanical origin (Balzan et al., 2020; Kňazovická et al., 2020; Wen et al., 2017). To further evaluate different physicochemical parameters of raw honey and their association with the microbiome of different types of honey, we measured honey pH, water activity, Brix, titratable acidity, color, and evaluated their association with microbial community diversity. We found that the bacterial communities among honey samples were relatively conservative, while fungal communities were more diverse. Some physicochemical properties of honey, including titratable acidity, water activity, and color, were associated with microbiome composition. To the best of our knowledge, this is the first article assessing the microbiome of manuka honey and feral honey via amplicon metagenomics. 55 Materials and methods Honey sample collection In this study, we performed physicochemical and microbiome analysis on four types of honey: monofloral, wildflower, manuka, and feral. Monofloral honey was purchased from two local honey shops (Ithaca, NY). Different floral sources were selected, including basswood, bamboo, buckwheat, orange blossom, goldenrod, and black locust. Wildflower honey was purchased from three honey shops in the central NY region. For the New Zealand manuka honey, three different brands were purchased online. Three feral honey samples were provided by a local beekeeper (Utica, NY), where the honeys were collected from swarmed honeybees. Honey samples were stored at room temperature until processing, due to honey’s shelf-stable nature. A total number of 36 honey samples were analyzed in this study. Physicochemical analysis of raw honey All honey samples were subject to physicochemical analysis. pH, titratable acidity, and Brix were measured using pH meter (pHi 470, Beckman Coulter, Brea, CA), automatic titrator (Ti-Note EasyPlus Titrators AP002, Mettler Toledo, Columbus, OH), and pocket digital refractometer (Sper Scientific, Scottsdale, AZ). Water activity was measured with water activity meter (AQUALAB 4TE, METER Group, Pullman, WA) and color was measured with Chroma Meter (Konica Minolta CR-400, Tokyo, Japan) using CIELAB scale. All measurements were performed in triplicate. DNA extraction, library preparation, and Illumina amplicon sequencing Honey was dissolved in phosphate-buffered saline (PBS) and treated with 1500 56 U/mL catalase to remove hydrogen peroxide that could be produced during dilution (Brudzynski et al., 2011; Chen et al., 2012). The 50% (w/w) honey solution was incubated at room temperature for 2 hours and centrifuged at 10,000 rpm, 4 C for 15 min. The pellet was resuspended in 10 mL PBS and centrifuged at 10,000 rpm, 4 C for another 15 min. DNA was extracted from this pellet with the DNeasy PowerSoil Pro Kit according to the manufacturer’s recommendation. Illumina MiSeq library preparation for 16S rRNA and ITS gene amplicon was performed. The 16S V3-V4 region was amplified with primers IL_Bakt341F (CCTACGGGNGGCWGCAG) and IL_Bakt805R (GACTACHVGGGTATCTAATCC) (Herlemann et al., 2011; Klindworth et al., 2013). A 0-4 bp heterogeneity spacer between Illumina index sequence and the 16S locus-specific primer was included to improve sequencing quality on the flow cell (Fadrosh et al., 2014). The ITS 5.8S-ITS2 region was amplified with primer IL_5.8SFungF (AACTTTYRRCAAYGGATCWCT) and IL_ITS4FungR (AGCCTCCGCTTATTGATATGCTTAART) (Taylor et al., 2016). Similarly, a 0-4 bp heterogeneity space was added between Illumina index sequence and the ITS primer. A two-step library preparation was adapted from a previous study by Holm et al. (2019). Successful target amplification from the first PCR was verified by gel electrophoresis and samples were then submitted to the Cornell Biotechnology Resource Center, Cornell Institute of Biotechnology (Ithaca, NY), for indexing and sequencing. Samples were quantified by Qubit 4 Fluorometer and then normalized prior to performing unique dual indexing. After dual indexing, samples were pooled and the library was cleaned using AMPure XP beads. Quality control with fragment analysis confirmed the correct distribution of fragment lengths. An Illumina MiSeq 2 x 57 250 bp (V2 chemistry) reagent kit was used to sequence the library. Two PCR negative controls and 4 extraction negative controls were included in this study. A total number of 84 amplicon samples were sequenced. Data Analysis QIIME 2 2021.11.0 was used to process and analyze the demultiplexed 16S and ITS amplicon sequencing data (Bolyen et al., 2019). Primers were trimmed from raw reads of 16S and ITS sequences using q2-cutadapt plugin. To achieve more accurate fungal taxonomic classification, demultiplexed ITS sequences were trimmed and conserved regions were removed using the q2-ITSxpress plugin (Rivers et al., 2018). DADA2 was used to filter, denoise, and merge trimmed reads to identify all observed amplicon sequence variants (ASVs) (Callahan et al., 2016). The chimeric sequences identified by DADA2 were removed. Taxonomy assignment was performed using a precomputed naïve Bayesian classifier (SILVA version 138 reference alignment for 16S rDNA sequences and UNITED version 8.3 database for ITS sequences) using q2-feature-classifier (Bokulich et al., 2018). Downstream analyses and visualization, including diversity analysis, statistical testing, and microbial community composition were performed in R (version 4.1.1). For 16S and ITS sequences, ASVs identified as mitochondria or chloroplast by the classifier were treated as contaminants and removed. Unknown ASVs at the phylum level were removed; these were typically unassigned mitochondria or chloroplast sequences (data not shown). Further decontamination of the sequences was based on extraction control and PCR control using “decontam” package in R (Davis et al., 2018). Sequences were rarefied and normalized with “phyloseq” package in R by 5 8 resampling the abundance values to achieve parity between samples (McMurdie & Holmes, 2013). The most abundant bacterial and fungal ASVs in honey samples were visualized with “ggpubr” package (version 0.4.0). Alpha diversity metrics, including Shannon diversity, Simpson and inverse Simpson diversity, Pielou’s evenness, ACE and Chao richness metrics, were calculated with “vegan” R package (version 2.5-7). For normally distributed alpha diversity metrics, ANOVA with Tukey’s honest significance test was used to perform pairwise comparisons between groups of categorical variables. General linear model with normal distribution was used to fit alpha diversity metrics to continuous variables. For non-normally distributed alpha diversity metrics, Wilcoxon rank sum test with false discovery rate (FDR) corrections for multiple comparisons. Kruskal-Wallis tests were performed on categorical variables. General linear model with quasipoisson distribution was used to fit continuous variables. For beta diversity, Bray-Curtis dissimilarity, Jaccard distance, and phylogeny-based UniFrac (weighted and unweighted) metrics were calculated (Beals, 1984; Jaccard, 1912; Lozupone & Knight, 2005; Lozupone et al., 2007). To visualize the differences in microbiome composition, beta diversity metrics were plotted with non-metric multidimensional scaling (NMDS). The multivariate homogeneity of group dispersion was tested by beta dispersion and the community composition was compared with permutation analysis of variance (PERMANOVA) using the “adonis” function in the “vegan” package with 1000 permutations. Additional visualizations, including heat maps and Venn diagrams, were created in R. 5 9 Results A total of 36 honey samples were subjected to physicochemical analysis and genome extraction for amplicon sequencing and microbiome analysis. The physicochemical data, including Brix, pH, titratable acidity, color, and water activity, were summarized in Table 2.1. Two feral honey samples, FF1 and FF2, were missing physicochemical data due to limited sample quantity. These two samples were subjected to 16S and ITS amplicon sequencing but were removed when performing diversity analyses. A total of 2,040,648 raw 16S amplicons and 4,084,874 raw ITS amplicons were sequence for 42 samples, including 4 extraction controls and 2 PCR controls. After trimming adapter sequences and primers, filtering low quality reads, denoising, and removing chimeric sequences, a total of 1,317,356 16S and 2,308,930 ITS reads remained for downstream analyses. After removing sequences unidentified at phylum level and decontamination using PCR and extraction controls, a total number of 1,285,423 16S reads remained. For ITS sequence, there were 2,305,376 reads retained after removing sequences unidentified at class level. After careful consideration, we removed two extraction contaminants identified as Yarrowia lipolytica and only used 830,110 reads of ITS sequence for downstream analysis. Before rarefaction and normalization, samples with low reads were removed: 16S sequences of 2 monofloral honey (HRG2 and HRO2) and 1 wildflower honey (W21S2), ITS sequences of 1 monofloral honey (HRO2) and 2 feral honeys (FF1 and FF2). Rarefaction curves for 16S and ITS sequences were visualized in Supplemental Figure 2.1. All samples reached plateau after resampling, indicating that the sequencing depth was sufficient 60 to capture microbial community diversity of the samples. To visualize the most abundant genera for bacterial composition, bacterial ASVs were agglomerated to the genus level, and genera with abundance higher than 0.05% were selected for honey bacterial composition plot (Figure 2.1). Similarly, fungal ASVs were agglomerated to the species level, and species with relative abundance higher than 0.3% were selected for honey fungal composition plot to illustrate the most abundant fungal species (Figure 2.2). The bacterial community showed less variability compared to fungal community, and it was dominated by Lactococcus lactis. Some other common genera of bacteria include Citrobacter, Pseudomonas, Serratia, and Cedecea. Specific fungal species were dominant in certain samples. Yarrowia lipolytica and Bettsia alvei were dominant in some of the monofloral, manuka, and wildflower honey, while feral honey was dominated by Zygosaccharomyces mellis. Some other fungal species were found in particular types of wildflower honey, such as Skoua sp., Zygosaccharomyces rouxii, Ascosphaera celerrima, and Saccharomyces sp. The core honey microbiome can be represented by shared taxa among different types of honey. As shown in the Venn diagrams (Figure 2.3), 66 bacterial ASVs were shared amongst all four types of honey, while there was 0 fungal ASV shared by all four types of honey. Based on our result, we can presume that there is a core bacterial microbiome for honey. Additionally, 167 fungal ASVs were present only in wildflower honey and 80 fungal ASVs were found only in monofloral honey, which further demonstrated that honey has a diverse and distinct fungal community. 6 1 Table 2.1. Physicochemical properties of honey. Each value is the mean of three measurements. 62 Table 2.1. Honey Sample ID Type Brix pH TA Color_L* Color_a* Color_b* Aw S1 Manuka 77.33 3.91 0.03427 29.67 2.26 10.62 0.6087 NZ1 Manuka 77.93 4.08 0.02843 30.45 0.94 11.64 0.5985 A1 Manuka 76.03 3.74 0.03327 30.72 2.23 12.42 0.6002 HRB1 Monofloral 81.77 4.06 0.03222 32.77 0.16 9.25 0.5433 HRO1 Monofloral 79.2 3.69 0.03465 31.56 0.37 11.85 0.5285 HRG1 Monofloral 78.3 4.07 0.03442 30.79 1.06 12.34 0.5425 WBB1 Monofloral 79.73 4.05 0.01967 29.1 3.59 12.75 0.5465 WBW1 Monofloral 80.37 4.2 0.03327 32.8 -0.01 8.51 0.5487 WBU1 Monofloral 79.63 3.79 0.03756 22.65 3.91 3.8 0.5602 WL1 Monofloral 78.73 4.11 0.02223 32.8 0.09 8.84 0.5226 W20F1 Wildflower 81.8 4.23 0.01302 29.04 2.78 12.22 0.561 W21S1 Wildflower 82.27 4.3 0.01152 30.84 0.45 10.87 0.5378 W21F1 Wildflower 80.03 3.93 0.01317 32.86 -1 7.38 0.5761 JS1 Wildflower 77.77 3.75 0.01245 41.05 -1.04 7.84 0.5651 KH1 Wildflower 82.8 4.05 0.01352 32.14 0.84 11.35 0.5287 FR1 Feral 77.4 4.29 0.0119 29.6 0.16 3.21 0.6337 63 FF1 Feral NA* NA NA NA NA NA NA FH1 Feral 77.03 3.84 0.0208 35.64 -0.62 8.1 0.6168 S2 Manuka 77.33 3.91 0.03427 29.67 2.26 10.62 0.6087 NZ2 Manuka 77.93 4.08 0.02843 30.45 0.94 11.64 0.5985 A2 Manuka 76.03 3.74 0.03327 30.72 2.23 12.42 0.6002 HRB2 Monofloral 81.77 4.06 0.03222 32.77 0.16 9.25 0.5433 HRO2 Monofloral 79.2 3.69 0.03465 31.56 0.37 11.85 0.5285 HRG2 Monofloral 78.3 4.07 0.03442 30.79 1.06 12.34 0.5425 WBB2 Monofloral 79.73 4.05 0.01967 29.1 3.59 12.75 0.5465 WBW2 Monofloral 80.37 4.2 0.03327 32.8 -0.01 8.51 0.5487 WBU2 Monofloral 79.63 3.79 0.03756 22.65 3.91 3.8 0.5602 WL2 Monofloral 78.73 4.11 0.02223 32.8 0.09 8.84 0.5226 W20F2 Wildflower 81.8 4.23 0.01302 29.04 2.78 12.22 0.561 W21S2 Wildflower 82.27 4.3 0.01152 30.84 0.45 10.87 0.5378 W21F2 Wildflower 80.03 3.93 0.01317 32.86 -1 7.38 0.5761 JS2 Wildflower 77.77 3.75 0.01245 41.05 -1.04 7.84 0.5651 KH2 Wildflower 82.8 4.05 0.01352 32.14 0.84 11.35 0.5287 FR2 Feral 77.4 4.29 0.0119 29.6 0.16 3.21 0.6337 FF2 Feral NA NA NA NA NA NA NA FH2 Feral 77.03 3.84 0.0208 35.64 -0.62 8.1 0.6168 6 4 *Physicochemical data were not available for FF1 and FF2 due to limited amount of honey samples collected. 65 Monofloral Honey Manuka Honey 1.00 1.00 0.75 0.75 0.50 0.50 0.25 0.25 0.00 0.00 WBB1 WBB2 WBW1 WBW2 WBU1 WBU2 WL1 WL2 HRB1 HRB2 HRG1 HRO1 NZ1 NZ2 S1 S2 A1 A2 Wildflower Honey Feral Honey 1.00 1.00 0.75 0.75 0.50 0.50 0.25 0.25 0.00 0.00 W20F1 W20F2 W21S1 W21F1 W21F2 JS1 JS2 KH1 KH2 FR1 FR2 FF1 FF2 FH1 FH2 Lactococcus Cedecea Lactobacillus Raoultella other Citrobacter Providencia Enterobacter Tyzzerella Genus Serratia Asaia Bombella unresolved Yersiniaceae Pseudomonas Proteus Bacillus Vagococcus unresolved Enterobacteriaceae Morganella Delftia Stenotrophomonas Figure 2.1. Honey bacterial composition plot. 16S ASVs were agglomerated to the genus level for each honey sample. Genera with relative abundance higher than 0.05% across samples were selected and plotted. Genera with less than 0.05% were agglomerated as “others”. 66 Monofloral Honey Manuka Honey 1.00 1.00 0.75 0.75 0.50 0.50 0.25 0.25 0.00 0.00 WBB1 WBB2 WBW1WBW2 WBU1 WBU2 WL1 WL2 HRB1 HRB2 HRG1 HRG2 HRO1 A1 A2 NZ1 NZ2 S1 S2 Wildflower Honey Feral Honey 1.00 1.00 0.75 0.75 0.50 0.50 0.25 0.25 0.00 0.00 W20F1 W20F2 W21S1 W21S2 W21F1 W21F2 JS1 JS2 KH1 KH2 FR1 FR2 FH1 FH2 Yarrowia lipolytica Skoua sp. Ascosphaera celerrima Robbauera albescens other Zygosaccharomyces mellis Wickerhamomyces anomalus Clavispora lusitaniae Candida santamariae Species Bettsia alvei unresolved Dipodascaceae Starmerella etchellsii Trichomonascus apis Aspergillus cibarius Saccharomyces sp. Hypoxylon submonticulosum Metschnikowia sp. Zygosaccharomyces rouxii Metschnikowia cibodasensis Aureobasidium pullulans Phallus rugulosus Figure 2.2. Honey fungal composition plot. ITS ASVs were agglomerated to the species level for each honey sample. Species with relative abundance higher than 0.3% across samples were selected and plotted. Species with less than 0.3% were agglomerated as “others”. 67 Figure 2.3. Venn diagrams for bacterial and fungal ASVs of four honey types. Bacterial and fungal ASVs were grouped based on four types of honey: monofloral, manuka, wildflower, and feral. ASVs shared between different types of honey were labeled in the overlapping area in the diagram. Left: bacterial. Right: fungal. To evaluate the species diversity within each honey type, alpha diversity of bacterial composition was assessed with Shannon diversity, inverse Simpson diversity, Chao richness and ACE richness indices, while alpha diversity of the fungal composition was measured with Shannon diversity, inverse Simpson diversity, Chao richness and Pielou’s evenness indices. Bar plots of alpha diversity indices grouped by honey types were visualized in Figure 2.4 and 2.5. ANOVA analysis reported p-value below 0.05 for all 4 metrics of bacterial alpha diversity and 1 metric of fungal alpha diversity (Chao richness), indicating that there were differences in the mean of these 68 indices between honey types. Pairwise comparisons using Tukey’s HSD test showed that there were significant differences between the bacterial community richness of monofloral and wildflower honey as measured by Chao and ACE indices (p < 0.05). In terms of the bacterial community diversity, there were significant differences between wildflower and monofloral honey as estimated by Shannon diversity metric (p = 0.0144) and between wildflower and manuka honey as measured with inverse Simpson diversity (p = 0.0416). Considering that Shannon diversity and inverse Simpson diversity metrics were not normally distributed, we thus performed Kruskal- Wallis rank sum test on these two metrics and found that they also differed by honey types (p < 0.01). Pairwise comparison was performed using Wilcoxon rank sum exact test and p-value was adjusted with false discovery rate (FDR) correction. Monofloral honey was found to be significantly different from both feral and wildflower honey as estimated by Shannon and inverse Simpson diversity metrics (p < 0.01). Fungal diversity metrics (Shannon and inverse Simpson) were relatively similar between different honey types. For the fungal community richness, only wildflower honey showed significant difference from the other three types of honey as estimated by Chao richness (p < 0.05). Similarly, Kruskal-Wallis rank sum test was used for non- normally distributed Chao richness metric and significant differences were found between groups (p < 0.05). Wildflower honey had a significantly different richness compared to the other three types of honey (p < 0.05) using Wilcoxon rank sum exact test with FDR adjustment. 6 9 a a ab ab ab a b b 4.0 3.5 10 3.0 2.5 5 2.0 Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Monofloral Honey Manuka Honey Wildflower Honey Feral Honey b ab a ab b ab a ab 80 80 70 70 60 60 50 Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Figure 2.4. Alpha diversity metrics of honey bacterial community. Diversity was measured with Shannon and inverse Simpson indices. Richness was measured with Chao and ACE indices. Comparison between honey sample types was performed with ANOVA and Tukey’s honest significance test. Letters above the bar plots represented shared significance groups (p-value cutoff is 0.05). 70 Chao richness Shannon's diversity Ace richness Inverse Simpson diversity a a a a a a a a 8 3 6 2 4 1 2 Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Monofloral Honey Manuka Honey Wildflower Honey Feral Honey 2.0 b b a b a a a a 50 40 1.5 30 1.0 20 10 0.5 0 Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Figure 2.5. Alpha diversity metrics of honey fungal community. Diversity was measured with Shannon and inverse Simpson indices. Richness was measured with Chao index. Evenness was measured with Pielou’s evenness index. Comparison between honey sample types was performed with ANOVA and Tukey’s honest significance test. Letters above the bar plots represented shared significance groups (p- value cutoff is 0.05). Physicochemical properties were tested for their correlations to alpha diversity metrics. For the bacterial community, titratable acidity (TA) was found to be associated with ACE richness by fitting the data to a general linear model (p-value = 0.0445). For the non-normal diversity metrics, TA was also found to be significantly 71 Chao1 richness Shannon's diversity Evenness Inverse Simpson diversity correlated with Shannon (t value = 4.025, Pr(>|t|) = 0.000414) and inverse Simpson (t value = 3.860, Pr(>|t|) = 0.000641) metrics using quasipoisson distribution (p-value < 0.001). For the fungal community composition, Brix, TA, and color (L* and a*) were found to be significantly associated with Chao richness estimator using quasipoisson distribution (p < 0.05). However, Pielou’s evenness index and diversity metrics were not correlated with any physicochemical properties. To evaluate the degree of differentiation among microbial communities of different honey types, Bray-Curtis and Jaccard beta diversity indices were calculated for both bacterial and fungal community of each sample. Weighted and unweighted UniFrac distance metrics were calculated only for the bacterial community but not for the fungal community, because ITS sequences cannot be used to inform evolutionary distances among distantly related species (Lücking et al., 2020; Schoch et al., 2012). The differences of Bray-Curtis index between samples were visualized with heatmap for both bacterial and fungal community (Figure 2.6). Honey samples were separated into three clusters based on bacterial composition, while fungal composition was divided into 12 clusters. Overall, honey samples used in this study had similar bacterial composition, but the fungal composition was more diverse. Varying degrees of overlap can be observed for clusters of each honey type in NMDS plots, especially for the fungal community of monofloral, wildflower, and manuka honey (Figure 2.7 and 2.8). PERMANOVA analysis showed significant differences in microbial community composition for different honey types using Bray-Curtis dissimilarity (pseudo F = 4.2385, R2 = 0.33714, p = 0.001998 for bacterial community, pseudo F = 2.7998, R2 = 0.22459, p = 0.000999 for fungal community). Pairwise comparison 72 between honey types was performed to further evaluate the differences. Results showed significant differences between the bacterial community of monofloral and wildflower honey using Bray-Curtis index (pseudo F = 9.0657543, R2 = 0.32301837, p = 0.001, adjusted p-value = 0.006). The differences of bacterial community between monofloral and feral honey was also significant (pseudo F = 4.3954513, R2 = 0.23894229, p = 0.008, adjusted p-value = 0.048). Other distance metrics for bacterial community, including Jaccard, weighted UniFrac and unweighted UniFrac showed similar results. As for the fungal community, pairwise comparison between different honey types showed significant differences between monofloral honey and feral honey using Bray-Curtis index (pseudo F = 4.8119958, R2 = 0.24288294, p = 0.001, adjusted p-value = 0.006). The fungal composition differences between wildflower and feral honey were also significant (pseudo F = 7.1475886, R2 = 0.37328923, p = 0.002, adjusted p-value = 0.012). Although the adjusted p-value was higher than 0.05 for the pairwise PERMANOVA analysis between manuka honey and feral honey using Bray- Curtis distance metric (pseudo F = 7.0767424, R2 = 0.46938140, p = 0.009, adjusted p-value = 0.054), the difference of Jaccard distance metric for these two honey types was significant, with an adjusted p-value lower than 0.05 (p = 0.005, adjusted p-value = 0.030). 73 Figure 2.6. Heatmaps of Bray-Curtis distances between honey bacterial and fungal community. Left: bacterial community. Right: fungal community. Each line and column represented a honey sample. The degree of similarity based on Bray- Curtis distances was represented by the color and dendrogram. Color red represented high similarity while light yellow represented low similarity. Samples grouped together in the dendrogram were highly similar. Color key and histogram in the top left corner of the heatmap represented the distribution of Bray-Curtis distances. 74 Bray−Curtis Jaccard 0.50 Aw pH 0.25 0.25 Color_L. 0.00 0.00 −0.25 Brix −0.25 −0.50 TA −0.50 CoCloorlo_rb_.a. −0.6 −0.3 0.0 0.3 −1.0 −0.5 0.0 0.5 NMDS1 NMDS1 Weighted UniFrac Unweighted UniFrac 0.4 pH 0.4 0.2 Color_b. Color_L. 0.2 Brix 0.0 Color_a. 0.0 TA −0.2 −0.2 −0.4 Aw −0.6 −0.4 −0.2 0.0 0.2 −0.25 0.00 0.25 0.50 NMDS1 NMDS1 Product Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Figure 2.7. Non-metric multidimensional scaling (NMDS) ordination for bacterial community structure based on the relative abundance of 16S ASVs. Community dissimilarity was evaluated with four metrics: Bray-Curtis, Jaccard, weighted UniFrac, and unweighted UniFrac. Arrowed lines (vectors) showing correlation between physicochemical properties and community dissimilarity were plotted for Bray-Curtis and weighted UniFrac metrics. The vectors represented mean direction and strength of correlation. Ellipses indicating confidence intervals of 95% for all honey types were plotted for Jaccard and unweighted UniFrac metrics. 75 NMDS2 NMDS2 NMDS2 NMDS2 Bray−Curtis Jaccard 0.4 Color_b. pH 0.5 0.2 Brix Color_L. 0.0 0.0 Color_a. Aw −0.2 −0.5 TA −0.5 0.0 0.5 1.0 −0.6 −0.3 0.0 0.3 0.6 NMDS1 NMDS1 Product Monofloral Honey Manuka Honey Wildflower Honey Feral Honey Figure 2.8. Non-metric multidimensional scaling (NMDS) ordination for fungal community structure based on the relative abundance of ITS ASVs. Community dissimilarity was evaluated with Bray-Curtis and Jaccard metrics. Ellipses indicating confidence intervals of 95% for all honey types were plotted for Bray-Curtis dissimilarity. Arrowed lines (vectors) showing correlation between physicochemical properties and Jaccard dissimilarity were plotted. The vectors represented mean direction and strength of correlation. Based on the visualization of the beta diversity metrics and the beta dispersion test, these differences in beta diversity can be attributed to the non-homogeneous distribution of each honey group. Permutation test for homogeneity of multivariate dispersions showed that the group distances of bacterial Bray-Curtis index were significant (F = 6.4887, Pr(>F) = 0.002997). Pairwise comparison further 7 6 NMDS2 NMDS2 demonstrated that the dispersion of wildflower and feral honey was significantly different from manuka and monofloral honey (p < 0.05). Similarly, the group distances of fungal Bray-Curtis index were significant (F = 6.6999, Pr(>F) = 0.002997). Further pairwise comparison showed that the beta dispersion of feral honey was significantly different from the other 3 types of honey (p < 0.05). To elucidate the relationship between physicochemical properties and the microbial community, all physicochemical parameters were treated as continuous variables and fitted to the bacterial Bray-Curtis metric, bacterial weighted Unifrac, and fungal Jaccard metric. Vectors of Brix, pH, TA, water activity and CIELAB color were visualized in NMDS plots for bacterial Bray-Curtis, bacterial weighted Unifrac, and fungal Jaccard indices (Figure 2.7 and 2.8). PERMANOVA analysis was performed on physicochemical data to evaluate the correlation between these variables and the microbial composition. Titratable acidity was determined as a factor that was significant for bacterial Bray-Curtis dissimilarity (pseudo F = 7.1182, R2 = 0.20863, p = 0.001998) and weighted UniFrac (pseudo F = 13.242, R2 = 0.32906, p = 0.000999). The fungal community measured by Jaccard distance was determined to be significantly associated with water activity (pseudo F = 2.6309, R2 = 0.07823, p = 0.01199) and color (L*: p = 0.03497, a*: p = 0.007992, b*: p = 0.005994). The differences in beta diversity of different types of honey can be attributed to the taxonomic composition of the microbiota, and the taxa with the highest coefficient values were visualized in Figure 9. For the bacterial community measured with Bray- Curtis metric, the top 5 ASVs with the largest effects on PERMANOVA coefficient were in the genera of Lactococcus, Serratia, Citrobacter, Serratia, and Pseudomonas. 77 For the fungal community, the top 5 ASVs with the largest effects on PERMANOVA coefficient were under the species of Zygosaccharomyces mellis, Yarrowia lipolytica, Bettsia alvei, Zygosaccharomyces mellis, and Skoua sp. Figure 2.9. Top coefficient amplicon sequence variants (ASVs) for beta diversity. Top bacterial and fungal ASVs that were associated with community differences between samples as estimated by Bray-Curtis dissimilarity were plotted. Color of each bar represented the genera for 16S ASVs and species for ITS ASVs. The top 20 ASVs with the highest PERMANOVA coefficient values were plotted. 78 Discussion Honey microbiota is a complex matrix that contains ecological information regarding the host microenvironment, the hive pathosphere, and the honeybee hologenome (Schwarz et al., 2015). Some bioindicators, including the agricultural and urban landscape, microbial environment that honeybees are exposed to, and the chemical pollutants in the foraging routes, can be reflected in the honey microbiota (Bargańska et al., 2016; Lambert et al., 2012; Rissato et al., 2007). NGS methods, including metabarcoding, can elucidate the complicated mutualism and symbiotic ecological relationships between honeybees and the environment (Bovo et al., 2018; Bovo et al., 2020). The information we obtained from next-generation sequencing can provide taxonomic classification of honey microbiota and potentially be used as an indicator for the overall beehive health and honey origin (Bovo et al., 2020). Most of the bacterial species identified in honey were osmotolerant, xerotolerant, and acidotolerant, considering that honey has a relatively high sugar content, low water activity, and low pH (Brudzynski, 2021). One of the most abundant bacterial species we found in our honey samples is Lactococcus lactis, which is consistently present in all honey samples we sequenced (Fig. 1). Lactococcus is a member of the lactic acid bacteria (LAB), which are able to ferment carbohydrates in honey (fructose and glucose) and produce lactic acid. As a ubiquitous group of bacteria that are commonly found in plant materials, LAB have been isolated from honeybee hives and bee products in previous research (Kňazovická et al., 2020; Sinacori et al., 2014). Some secondary metabolites produced by LAB strains can inhibit spoilage organisms and pathogens and contribute to the overall beehive health. 79 One example is Lactobacillus kunkeei, which is beneficial to the bee colony by protecting the hive from potential pathogens like Paenibacillus larvae and Nosema ceranae (Arredondo et al., 2018). As a ubiquitous species that is commonly found in flower, fruits, and soil, L. kunkeei is commonly associated with honeybee hive environment and bee products. L. kunkeei was found in honey bee bread using both culture-dependent and culture-independent method (Anderson et al., 2013). For honey samples in our study, L. kunkeei is one of the fructophiles that can be found in some but not all honey samples (Fig. 1). Comparatively, another study on the microbiome of stingless bee honey revealed that the most abundant species is Lactobacillus malefermentans, and the top 7 OTUs in this study were all members of the genus Lactobacillus (Rosli et al., 2020). One of the possible reasons that Lactobacillus is missing in some of our honey samples is that Lactobacillus disappears below moisture content of 18% during honey ripening process (Ruiz-Argueso & Rodriguez-Navarro, 1975; Wen et al., 2017). Other studies also suggested that the presence of L. kunkeei is sporadic and its detection is dependent on the factors like floral source and season (Vásquez et al., 2012). Distinct differences can be seen when comparing our bacterial profile with the bacterial profile of vitex honey during ripening, which was dominated by Bacillus spp. (Wen et al., 2017). However, some bacteria with high abundance in vitex honey can be found in our honey samples, including Lactococcus and Pseudomonas. Some unresolved Enterobacteriaceae were present in our honey samples, which are likely from the pollination environment since they are frequently isolated from crops of forager bees (Corby-Harris et al., 2014). Even though gut microbiota could be a source of microbial community members in honey, many gut 80 bacteria are considered gut-specific and do not survive well in other environments. Only L. kunkeei and Acetobacteraceae (Asaia spp.) were found in extreme conditions like honey and royal jelly (Anderson et al., 2013; V. G. Martinson et al., 2012; Vojvodic et al., 2013). Serratia is one of the most abundant genera found in our honey samples, which is consistent with a previous microbial metabarcoding study on three polyfloral honeys from Italy, where Serratia symbiotica was the fourth most abundant bacteria accounting for 4.8% of the bacteria reads (Bovo et al., 2020). The origin of Serratia is somewhat puzzling, since it is commonly associated with aphids as a secondary endosymbiont. It is possible that Serratia originated from honeydew produced by aphids, which was then fed to honeybees to produce honey (Bovo et al., 2020). For the fungal communities, diverse profiles can be observed across different types of honey. The most abundant fungal genera in our honey samples were Bettsia, Yarrowia, Skoua, Zygosaccharomyces, and Metschnikowia. Similar to our study, the fungal profile of vitex honey is also heterogeneous, with Waitea, Phoma, Metschnikowia, Cryptococcus being the most predominant genera. Metschnikowia was found to be relatively stable in mature vitex honey and dominant in vitex flower. We propose that Metschnikowia and other yeasts in our honey samples originated from nectar, which can be transmitted from flower and fruits to honeybee products (Hong et al., 2001; Lievens et al., 2015). The absence of Waitea and Cryptococcus in our honey samples could be due to flower origin, since these two genera were found to be dominant in vitex flower (Wen et al., 2017). Culture-based methods identified yeasts like Zygosaccharomyces and Debaryomyces as the most prevalent genera in honey 81 (Sinacori et al., 2014). In a culture-independent study with ITS2 metabarcoding, Zygosaccharomyces was the only species shared among almost all honey samples (Balzan et al., 2020). Filamentous fungi like Aspergillus are considered environmental contaminants for honey (Kacániová et al., 2009). A shotgun metagenomic study found that the second most represented fungus in polyfloral Italian honey was Aspergillus flavus (Bovo et al., 2020). Aspergillus flavus is a potential honeybee pathogen that could cause stonebrood disease, and was found to be abundant in some of our monofloral, wildflower, and manuka honey samples (Fig. 2). Similarly, Ascosphaera apis is the causative agent for chalkbrood disease (Vojvodic et al., 2011). Ascosphaera sp. was found to be prevalent in some of the wildflower honey samples in our study (Fig. 2). However, the presence of pathogenic fungi does not necessarily mean that the beehives are infected. Indeed, as shown in the study by Bovo et al, none of the sampled colonies that contained pathogenic fungi DNA in metagenomic analysis displayed any of these symptoms over two years (Bovo et al., 2020). The onset of these diseases requires specific environment factors, and most of the pathogenic fungi only survive in honey as dormant spores. In our study, we chose to not perform culture-based isolation methods due to culture biases. Performing bacterial and fungal culture isolation could not give us a whole picture of the microbiota, nor could it provide proof of the absence of certain species. As previous studies shown, species from genera Bacillus and Paenibacillus were considered dominant when evaluating the honey bacterial composition with culture-based method because aerobic plate counts were usually dominated by fast- growing bacteria like Bacillus spp., Staphylococcus spp. and Paenibacillus spp., while 82 the dominant bacteria identified using amplicon sequencing were under-represented in culture-based methods due to various factors, like injured cells, persister cells, improper culture environment, or failing to compete with other organisms in culture (Balzan et al., 2020; Iurlina & Fritz, 2005; Sinacori et al., 2014). Moreover, plate count methods overestimated the bacteria abundance in honeybee stomach by over one order of magnitude, and core crop bacteria identified using culture-based method were inconsistent and occurred at low frequency when using qRT-PCR or NGS methods (Corby-Harris et al., 2014). In our opinion, using culture-independent methods to investigate the microbiome of honey avoids the growth condition and culture biases, and culture-based methods should not be performed as a complement to culture- independent amplicon sequencing or metagenomic studies. Alternatively, designing strain-specific primers and performing qRT-PCR is the proper way to confirm the presence/absence of certain species identified by amplicon sequencing. The physicochemical properties of different honey types in this study were highly comparable, especially for pH and Brix (Table 1). Color is one of the parameters that can be used to distinguish different honeys. Ecological diversity indices can be assessed based on the ASVs in different honey products, and certain hypotheses can be drawn based on statistical analysis. In our study, titratable acidity was found to be correlated with bacterial alpha diversity metrics, including ACE richness, Shannon diversity, and inverse Simpson diversity. A few physicochemical factors were also found to be correlated with fungal Chao richness metric, including TA, Brix, and color. Furthermore, based on beta diversity metric correlation analysis, we determined that TA was a significant factor associated with the differences in 83 bacterial communities, while water activity and color were associated with the differences in fungal communities. Previous studies showed that honey pH and acidity were independent of geographic origins but associated with nectar composition and botanical source (da Silva et al., 2016; Scholz et al., 2020). Honey age, moisture, and purchase source were considered as relevant factors for the microbial community in raw honey, while botanical origin only affected the fungal composition (Balzan et al., 2020). pH, water activity, and country of origin were considered as minor factors. In our study, moisture was not a significant factor shaping the bacterial or fungal community. Conversely, several previous studies showed that honey microbial profile was associated with its moisture. In the study by Wen et al. (2017), the fungal community of vitex honey was correlated with moisture. Honey with high moisture content is more likely to ferment and spoil. However, the moisture content variation in our honey samples was relatively small, which may be the reason that the moisture content was not a significant factor influencing the microbial community of our honey samples. Another group of researchers evaluated physicochemical parameters including pH, water content, free acidity and electrical conductivity and determined that only electrical conductivity was associated with bacterial community of honey based on RDA analysis and permutation test (Kňazovická et al., 2020). In the study by Rosli et al, the authors considered that the microbiome of stingless bee honey was associated with physicochemical factors including pH, acidity, and moisture content (Rosli et al., 2020). The marginal effects of limited sample size may contribute to the discrepancy among different studies. Some other authors also mentioned the geographic region may be an important factor influencing the microbial community in 84 honeybee products (Disayathanoowat et al., 2020). We only included two geographic regions in our study, which is why we cannot draw any conclusions on its effect on the microbial community. Future metagenomic studies should take geographic location into consideration when evaluating factors that may impact the microbial community of honey. To fully understand the effects of geographic location and other relevant variables on the microbiome diversity, samples collected in different regions of US or world need to be included, with specific details on the geographic distribution, local flowering plants diversity, and the honeybee genetic background. Next-generation sequencing tools provide a higher level of resolution of the community composition compared to traditional culture methods. Species with low abundance can be detected with in depth sequencing, which enables us to evaluate the microbiome composition more precisely (Claesson et al., 2010; Gupta et al., 2019). Some studies have been performed to evaluate the floral source of honey using DNA metabarcoding for authentication, and the digestive tract microbiome of honeybees with metagenomic tools (de Vere et al., 2017; Graystock et al., 2017; Jones et al., 2018; Utzeri et al., 2018; Yun et al., 2018). Many of these studies used 16S rRNA amplicon sequencing, which is what we chose to use in our study to evaluate the composition of bacterial community in honey. Metabarcoding methods have high coverage, high sequencing depth, and is non-selective (Cao et al., 2017). However, the disadvantage is that most of the sequences are assigned to the taxon with high abundance, which may neglect some of the less common species in a community with high complexity (Clooney et al., 2016). The bacterial classification is also identified at genus level or above (Claesson et al., 2010). To avoid using a pre-defined percentage 85 threshold to determine variants, we chose to use amplicon sequence variants (ASVs), which considers amplicon abundance and error rates to discard spurious sequences and retain biologically meaningful sequences (Callahan et al., 2017). This method has a finer resolution and identifies microorganisms at phenotypic levels (Rognes et al., 2016). Using ASVs to represent original sequences is considered a step forward compared to previous studies using operational taxonomic units (OTUs) to construct consensus sequences with 97% similarity, which inevitably loses some taxonomic information (Strube, 2021). In our study, we chose the 16S V3-V4 region and 5.8S- ITS2 regions considering the limited read length of Illumina MiSeq. The potential sequencing biases from Illumina MiSeq is also the reason we condensed ASVs to the genus level for 16S amplicons and species level for ITS amplicons instead of using ASVs as individual units (Strube, 2021). Future studies should use the full 16S V1-V9 region and full-length ITS1-5.8S-ITS2 region to get better resolution of the honey bacterial and fungal population. Using sequencing platforms with higher read length and choosing proper primers for multiple barcode sequences will yield results with higher resolution. This study contributes to the knowledge of environmental effects on microbial biodiversity and ecosystem associated with different types of honey. Investigation on the microbiome of honey and other bee products could shed light into Colony Collapse Disorder (CCD), a common disease in honeybee colonies that causes significant ecological and economic damage (Cox-Foster et al., 2007). By comparing the microbiome of honey produced by different bee colonies, we can investigate the correlation between the microbiome and honeybee diseases. Even though the presence 86 of pathogenic microbial DNA may not directly correlate to honeybee diseases, using metagenomic tools to determine the relative abundance of these pathogens can provide information on possible hive diseases and overall beehive health. Future studies should focus on the shift of certain bacterial and fungal species in honey and other bee products to decipher the implication of honey microbiome on honeybee diseases and overall bee colony health. 8 7 REFERENCES Álvarez-Pérez, S., Herrera, C. M., & de Vega, C. (2012). Zooming-in on floral nectar: a first exploration of nectar-associated bacteria in wild plant communities. FEMS Microbiology Ecology, 80(3), 591-602. https://doi.org/10.1111/j.1574- 6941.2012.01329.x Anderson, K. E., Sheehan, T. H., Mott, B. M., Maes, P., Snyder, L., Schwan, M. R., Walton, A., Jones, B. M., & Corby-Harris, V. (2013). Microbial ecology of the hive and pollination landscape: bacterial associates from floral nectar, the alimentary tract and stored food of honey bees (Apis mellifera). PLoS One, 8(12), e83125-e83125. https://doi.org/10.1371/journal.pone.0083125 Arredondo, D., Castelli, L., Porrini, M. P., Garrido, P. M., Eguaras, M. J., Zunino, P., & Antúnez, K. (2018). Lactobacillus kunkeei strains decreased the infection by honey bee pathogens Paenibacillus larvae and Nosema ceranae. Benef Microbes, 9(2), 279-290. https://doi.org/10.3920/bm2017.0075 Balzan, S., Carraro, L., Merlanti, R., Lucatello, L., Capolongo, F., Fontana, F., Novelli, E., Larini, I., Vitulo, N., & Cardazzo, B. (2020). Microbial metabarcoding highlights different bacterial and fungal populations in honey samples from local beekeepers and market in north-eastern Italy. International Journal of Food Microbiology, 334, 108806. https://doi.org/https://doi.org/10.1016/j.ijfoodmicro.2020.108806 Bargańska, Ż., Ślebioda, M., & Namieśnik, J. (2016). Honey bees and their products: Bioindicators of environmental contamination. Critical Reviews in 88 Environmental Science and Technology, 46(3), 235-248. https://doi.org/10.1080/10643389.2015.1078220 Barke, J., Seipke, R. F., Grüschow, S., Heavens, D., Drou, N., Bibb, M. J., Goss, R. J., Yu, D. W., & Hutchings, M. I. (2010). A mixed community of actinomycetes produce multiple antibiotics for the fungus farming ant Acromyrmex octospinosus. BMC Biol, 8, 109. https://doi.org/10.1186/1741-7007-8-109 Beals, E. W. (1984). Bray-Curtis Ordination: An Effective Strategy for Analysis of Multivariate Ecological Data. In A. MacFadyen & E. D. Ford (Eds.), Advances in Ecological Research (Vol. 14, pp. 1-55). Academic Press. https://doi.org/https://doi.org/10.1016/S0065-2504(08)60168-3 Bensch, K., Braun, U., Groenewald, J. Z., & Crous, P. W. (2012). The genus Cladosporium. Studies in mycology, 72, 1-401. https://doi.org/https://doi.org/10.3114/sim0003 Bokulich, N. A., Kaehler, B. D., Rideout, J. R., Dillon, M., Bolyen, E., Knight, R., Huttley, G. A., & Gregory Caporaso, J. (2018). Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2's q2-feature- classifier plugin. Microbiome, 6(1), 90. https://doi.org/10.1186/s40168-018- 0470-z Bolyen, E., Rideout, J. R., Dillon, M. R., Bokulich, N. A., Abnet, C. C., Al-Ghalith, G. A., Alexander, H., Alm, E. J., Arumugam, M., Asnicar, F., Bai, Y., Bisanz, J. E., Bittinger, K., Brejnrod, A., Brislawn, C. J., Brown, C. T., Callahan, B. J., Caraballo-Rodríguez, A. M., Chase, J., . . . Caporaso, J. G. (2019). Reproducible, interactive, scalable and extensible microbiome data science 8 9 using QIIME 2. Nature biotechnology, 37(8), 852-857. https://doi.org/10.1038/s41587-019-0209-9 Bovo, S., Ribani, A., Utzeri, V. J., Schiavo, G., Bertolini, F., & Fontanesi, L. (2018). Shotgun metagenomics of honey DNA: Evaluation of a methodological approach to describe a multi-kingdom honey bee derived environmental DNA signature. PLoS One, 13(10), e0205575. https://doi.org/10.1371/journal.pone.0205575 Bovo, S., Utzeri, V. J., Ribani, A., Cabbri, R., & Fontanesi, L. (2020). Shotgun sequencing of honey DNA can describe honey bee derived environmental signatures and the honey bee hologenome complexity. Sci Rep, 10(1), 9279. https://doi.org/10.1038/s41598-020-66127-1 Brudzynski, K. (2021). Honey as an Ecological Reservoir of Antibacterial Compounds Produced by Antagonistic Microbial Interactions in Plant Nectars, Honey and Honey Bee. Antibiotics (Basel), 10(5). https://doi.org/10.3390/antibiotics10050551 Brudzynski, K., Abubaker, K., St-Martin, L., & Castle, A. (2011). Re-examining the role of hydrogen peroxide in bacteriostatic and bactericidal activities of honey. Front Microbiol, 2, 213. https://doi.org/10.3389/fmicb.2011.00213 Callahan, B. J., McMurdie, P. J., & Holmes, S. P. (2017). Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. ISME J, 11(12), 2639-2643. https://doi.org/10.1038/ismej.2017.119 Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J., & Holmes, S. P. (2016). DADA2: High-resolution sample inference from Illumina 90 amplicon data. Nat Methods, 13(7), 581-583. https://doi.org/10.1038/nmeth.3869 Cao, Y., Fanning, S., Proos, S., Jordan, K., & Srikumar, S. (2017). A Review on the Applications of Next Generation Sequencing Technologies as Applied to Food-Related Microbiome Studies [Review]. Frontiers in microbiology, 8. https://doi.org/10.3389/fmicb.2017.01829 Chen, C., Campbell, L., Blair, S., & Carter, D. (2012). The effect of standard heat and filtration processing procedures on antimicrobial activity and hydrogen peroxide levels in honey [Original Research]. Frontiers in microbiology, 3. https://doi.org/10.3389/fmicb.2012.00265 Claesson, M. J., Wang, Q., O'Sullivan, O., Greene-Diniz, R., Cole, J. R., Ross, R. P., & O'Toole, P. W. (2010). Comparison of two next-generation sequencing technologies for resolving highly complex microbiota composition using tandem variable 16S rRNA gene regions. Nucleic Acids Research, 38(22), e200-e200. https://doi.org/10.1093/nar/gkq873 Clooney, A. G., Fouhy, F., Sleator, R. D., A, O. D., Stanton, C., Cotter, P. D., & Claesson, M. J. (2016). Comparing Apples and Oranges?: Next Generation Sequencing and Its Impact on Microbiome Analysis. PLoS One, 11(2), e0148028. https://doi.org/10.1371/journal.pone.0148028 Corby-Harris, V., Maes, P., & Anderson, K. E. (2014). The bacterial communities associated with honey bee (Apis mellifera) foragers. PLoS One, 9(4), e95056. https://doi.org/10.1371/journal.pone.0095056 91 Cox-Foster, D. L., Conlan, S., Holmes, E. C., Palacios, G., Evans, J. D., Moran, N. A., Quan, P. L., Briese, T., Hornig, M., Geiser, D. M., Martinson, V., vanEngelsdorp, D., Kalkstein, A. L., Drysdale, A., Hui, J., Zhai, J., Cui, L., Hutchison, S. K., Simons, J. F., . . . Lipkin, W. I. (2007). A metagenomic survey of microbes in honey bee colony collapse disorder. Science, 318(5848), 283-287. https://doi.org/10.1126/science.1146498 da Silva, P. M., Gauche, C., Gonzaga, L. V., Costa, A. C. O., & Fett, R. (2016). Honey: Chemical composition, stability and authenticity. Food Chemistry, 196, 309-323. https://doi.org/https://doi.org/10.1016/j.foodchem.2015.09.051 Davis, N. M., Proctor, D. M., Holmes, S. P., Relman, D. A., & Callahan, B. J. (2018). Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome, 6(1), 226. https://doi.org/10.1186/s40168-018-0605-2 de Vere, N., Jones, L. E., Gilmore, T., Moscrop, J., Lowe, A., Smith, D., Hegarty, M. J., Creer, S., & Ford, C. R. (2017). Using DNA metabarcoding to investigate honey bee foraging reveals limited flower use despite high floral availability. Sci Rep, 7, 42838. https://doi.org/10.1038/srep42838 Disayathanoowat, T., Li, H., Supapimon, N., Suwannarach, N., Lumyong, S., Chantawannakul, P., & Guo, J. (2020). Different Dynamics of Bacterial and Fungal Communities in Hive-Stored Bee Bread and Their Possible Roles: A Case Study from Two Commercial Honey Bees in China. Microorganisms, 8(2), 264. https://www.mdpi.com/2076-2607/8/2/264 9 2 Engel, P., Martinson, V. G., & Moran, N. A. (2012). Functional diversity within the simple gut microbiota of the honey bee. Proceedings of the National Academy of Sciences, 109(27), 11002-11007. https://doi.org/doi:10.1073/pnas.1202970109 Fadrosh, D. W., Ma, B., Gajer, P., Sengamalay, N., Ott, S., Brotman, R. M., & Ravel, J. (2014). An improved dual-indexing approach for multiplexed 16S rRNA gene sequencing on the Illumina MiSeq platform. Microbiome, 2(1), 6. https://doi.org/10.1186/2049-2618-2-6 Fridman, S., Izhaki, I., Gerchman, Y., & Halpern, M. (2012). Bacterial communities in floral nectar. Environmental Microbiology Reports, 4(1), 97-104. https://doi.org/https://doi.org/10.1111/j.1758-2229.2011.00309.x Genersch, E. (2010). American Foulbrood in honeybees and its causative agent, Paenibacillus larvae. Journal of Invertebrate Pathology, 103, S10-S19. https://doi.org/https://doi.org/10.1016/j.jip.2009.06.015 Gilliam, M., Wickerham, L. J., Morton, H. L., & Martin, R. D. (1974). Yeasts isolated from honey bees, Apis mellifera, fed 2,4-D and antibiotics. Journal of Invertebrate Pathology, 24(3), 349-356. https://doi.org/https://doi.org/10.1016/0022-2011(74)90143-8 Graystock, P., Rehan, S. M., & McFrederick, Q. S. (2017). Hunting for healthy microbiomes: determining the core microbiomes of Ceratina, Megalopta, and Apis bees and how they associate with microbes in bee collected pollen. Conservation Genetics, 18(3), 701-711. https://doi.org/10.1007/s10592-017- 0937-7 93 Gupta, S., Mortensen, M. S., Schjørring, S., Trivedi, U., Vestergaard, G., Stokholm, J., Bisgaard, H., Krogfelt, K. A., & Sørensen, S. J. (2019). Amplicon sequencing provides more accurate microbiome information in healthy children compared to culturing. Communications Biology, 2(1), 291. https://doi.org/10.1038/s42003-019-0540-1 Herlemann, D. P., Labrenz, M., Jürgens, K., Bertilsson, S., Waniek, J. J., & Andersson, A. F. (2011). Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J, 5(10), 1571-1579. https://doi.org/10.1038/ismej.2011.41 Hinshaw, C., Evans, K. C., Rosa, C., & López-Uribe, M. M. (2021). The Role of Pathogen Dynamics and Immune Gene Expression in the Survival of Feral Honey Bees [Original Research]. Frontiers in Ecology and Evolution, 8. https://doi.org/10.3389/fevo.2020.594263 Holm, J. B., Humphrys, M. S., Robinson, C. K., Settles, M. L., Ott, S., Fu, L., Yang, H., Gajer, P., He, X., McComb, E., Gravitt, P. E., Ghanem, K. G., Brotman, R. M., Ravel, J., & Neufeld, J. D. (2019). Ultrahigh-Throughput Multiplexing and Sequencing of >500-Base-Pair Amplicon Regions on the Illumina HiSeq 2500 Platform. mSystems, 4(1), e00029-00019. https://doi.org/doi:10.1128/mSystems.00029-19 Hong, S. G., Chun, J., Oh, H. W., & Bae, K. S. (2001). Metschnikowia koreensis sp. nov., a novel yeast species isolated from flowers in Korea. International Journal of Systematic and Evolutionary Microbiology, 51(5), 1927-1931. https://doi.org/https://doi.org/10.1099/00207713-51-5-1927 9 4 Iurlina, M. O., & Fritz, R. (2005). Characterization of microorganisms in Argentinean honeys from different sources. International Journal of Food Microbiology, 105(3), 297-304. https://doi.org/https://doi.org/10.1016/j.ijfoodmicro.2005.03.017 Jaccard, P. (1912). THE DISTRIBUTION OF THE FLORA IN THE ALPINE ZONE.1. New Phytologist, 11(2), 37-50. https://doi.org/https://doi.org/10.1111/j.1469-8137.1912.tb05611.x Jensen, A. B., Aronstein, K., Flores, J. M., Vojvodic, S., Palacio, M. A., & Spivak, M. (2013). Standard methods for fungal brood disease research. Journal of Apicultural Research, 52(1), 1-20. https://doi.org/10.3896/IBRA.1.52.1.13 Jones, J. C., Fruciano, C., Hildebrand, F., Al Toufalilia, H., Balfour, N. J., Bork, P., Engel, P., Ratnieks, F. L., & Hughes, W. O. (2018). Gut microbiota composition is associated with environmental landscape in honey bees. Ecol Evol, 8(1), 441-451. https://doi.org/10.1002/ece3.3597 Kacániová, M., Pavlicová, S., Hascík, P., Kociubinski, G., Kńazovická, V., Sudzina, M., Sudzinová, J., & Fikselová, M. (2009). Microbial communities in bees, pollen and honey from Slovakia. Acta Microbiol Immunol Hung, 56(3), 285- 295. https://doi.org/10.1556/AMicr.56.2009.3.7 Klindworth, A., Pruesse, E., Schweer, T., Peplies, J., Quast, C., Horn, M., & Glöckner, F. O. (2013). Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res, 41(1), e1. https://doi.org/10.1093/nar/gks808 95 Kňazovická, V., Gábor, M., Miluchová, M., Bobko, M., & Medo, J. (2020). Diversity of bacteria in Slovak and foreign honey, with assessment of its physico- chemical quality and counts of cultivable microorganisms. Journal of Microbiology, Biotechnology and Food Sciences, 9(6), 414-421. Kurek-Górecka, A., Górecki, M., Rzepecka-Stojko, A., Balwierz, R., & Stojko, J. (2020). Bee Products in Dermatology and Skin Care. Molecules (Basel, Switzerland), 25(3), 556. https://doi.org/10.3390/molecules25030556 Lambert, O., Piroux, M., Puyo, S., Thorin, C., Larhantec, M., Delbac, F., & Pouliquen, H. (2012). Bees, honey and pollen as sentinels for lead environmental contamination. Environ Pollut, 170, 254-259. https://doi.org/10.1016/j.envpol.2012.07.012 Lievens, B., Hallsworth, J. E., Pozo, M. I., Belgacem, Z. B., Stevenson, A., Willems, K. A., & Jacquemyn, H. (2015). Microbiology of sugar-rich environments: diversity, ecology and system constraints. Environmental Microbiology, 17(2), 278-298. https://doi.org/https://doi.org/10.1111/1462-2920.12570 Lozupone, C., & Knight, R. (2005). UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol, 71(12), 8228- 8235. https://doi.org/10.1128/aem.71.12.8228-8235.2005 Lozupone, C. A., Hamady, M., Kelley, S. T., & Knight, R. (2007). Quantitative and qualitative beta diversity measures lead to different insights into factors that structure microbial communities. Appl Environ Microbiol, 73(5), 1576-1585. https://doi.org/10.1128/aem.01996-06 96 Lücking, R., Aime, M. C., Robbertse, B., Miller, A. N., Ariyawansa, H. A., Aoki, T., Cardinali, G., Crous, P. W., Druzhinina, I. S., Geiser, D. M., Hawksworth, D. L., Hyde, K. D., Irinyi, L., Jeewon, R., Johnston, P. R., Kirk, P. M., Malosso, E., May, T. W., Meyer, W., . . . Schoch, C. L. (2020). Unambiguous identification of fungi: where do we stand and how accurate and precise is fungal DNA barcoding? IMA fungus, 11, 14-14. https://doi.org/10.1186/s43008-020-00033-z Martinson, E. O., Herre, E. A., Machado, C. A., & Arnold, A. E. (2012). Culture-free survey reveals diverse and distinctive fungal communities associated with developing figs (Ficus spp.) in Panama. Microb Ecol, 64(4), 1073-1084. https://doi.org/10.1007/s00248-012-0079-x Martinson, V. G., Moy, J., & Moran, N. A. (2012). Establishment of characteristic gut bacteria during development of the honeybee worker. Appl Environ Microbiol, 78(8), 2830-2840. https://doi.org/10.1128/aem.07810-11 McMurdie, P. J., & Holmes, S. (2013). phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. PLoS One, 8(4), e61217. https://doi.org/10.1371/journal.pone.0061217 Mohr, K. I., & Tebbe, C. C. (2006). Diversity and phylotype consistency of bacteria in the guts of three bee species (Apoidea) at an oilseed rape field. Environmental Microbiology, 8(2), 258-272. https://doi.org/https://doi.org/10.1111/j.1462- 2920.2005.00893.x Moran, N. A., Hansen, A. K., Powell, J. E., & Sabree, Z. L. (2012). Distinctive gut microbiota of honey bees assessed using deep sampling from individual 9 7 worker bees. PLoS One, 7(4), e36393. https://doi.org/10.1371/journal.pone.0036393 Niaz, K., Maqbool, F., Bahadar, H., & Abdollahi, M. (2017). Health Benefits of Manuka Honey as an Essential Constituent for Tissue Regeneration. Curr Drug Metab, 18(10), 881-892. https://doi.org/10.2174/1389200218666170911152240 Olaitan, P. B., Adeleke, O. E., & Ola, I. O. (2007). Honey: a reservoir for microorganisms and an inhibitory agent for microbes. Afr Health Sci, 7(3), 159-165. https://doi.org/10.5555/afhs.2007.7.3.159 Olofsson, T. C., & Vásquez, A. (2008). Detection and identification of a novel lactic acid bacterial flora within the honey stomach of the honeybee Apis mellifera. Curr Microbiol, 57(4), 356-363. https://doi.org/10.1007/s00284-008-9202-0 Powell, J. E., Martinson, V. G., Urban-Mead, K., Moran, N. A., & Goodrich-Blair, H. (2014). Routes of Acquisition of the Gut Microbiota of the Honey Bee Apis mellifera. Applied and Environmental Microbiology, 80(23), 7378-7387. https://doi.org/doi:10.1128/AEM.01861-14 Raymann, K., & Moran, N. A. (2018). The role of the gut microbiome in health and disease of adult honey bee workers. Curr Opin Insect Sci, 26, 97-104. https://doi.org/10.1016/j.cois.2018.02.012 Rissato, S. R., Galhiane, M. S., de Almeida, M. V., Gerenutti, M., & Apon, B. M. (2007). Multiresidue determination of pesticides in honey samples by gas chromatography–mass spectrometry and application in environmental 98 contamination. Food Chemistry, 101(4), 1719-1726. https://doi.org/https://doi.org/10.1016/j.foodchem.2005.10.034 Rivers, A., Weber, K., Gardner, T., Liu, S., & Armstrong, S. (2018). ITSxpress: Software to rapidly trim internally transcribed spacer sequences with quality scores for marker gene analysis [version 1; peer review: 2 approved]. F1000Research, 7(1418). https://doi.org/10.12688/f1000research.15704.1 Rognes, T., Flouri, T., Nichols, B., Quince, C., & Mahé, F. (2016). VSEARCH: a versatile open source tool for metagenomics. PeerJ, 4, e2584. https://doi.org/10.7717/peerj.2584 Rosli, F. N., Hazemi, M. H. F., Akbar, M. A., Basir, S., Kassim, H., & Bunawan, H. (2020). Stingless Bee Honey: Evaluating Its Antibacterial Activity and Bacterial Diversity. Insects, 11(8), 500. https://www.mdpi.com/2075- 4450/11/8/500 Ruiz-Argueso, T., & Rodriguez-Navarro, A. (1975). Microbiology of ripening honey. Appl Microbiol, 30(6), 893-896. https://doi.org/10.1128/am.30.6.893-896.1975 Schoch, C. L., Seifert, K. A., Huhndorf, S., Robert, V., Spouge, J. L., Levesque, C. A., Chen, W., Fungal Barcoding, C., & Fungal Barcoding Consortium Author, L. (2012). Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Proceedings of the National Academy of Sciences of the United States of America, 109(16), 6241-6246. https://doi.org/10.1073/pnas.1117018109 Scholz, M. B. d. S., Quinhone Júnior, A., Delamuta, B. H., Nakamura, J. M., Baudraz, M. C., Reis, M. O., Kato, T., Pedrão, M. R., Dias, L. F., dos Santos, D. T. R., 9 9 Kitzberger, C. S. G., & Bianchini, F. P. (2020). Indication of the geographical origin of honey using its physicochemical characteristics and multivariate analysis. Journal of Food Science and Technology, 57(5), 1896-1903. https://doi.org/10.1007/s13197-019-04225-3 Schwarz, R. S., Huang, Q., & Evans, J. D. (2015). Hologenome theory and the honey bee pathosphere. Current Opinion in Insect Science, 10, 1-7. https://doi.org/https://doi.org/10.1016/j.cois.2015.04.006 Sinacori, M., Francesca, N., Alfonzo, A., Cruciata, M., Sannino, C., Settanni, L., & Moschetti, G. (2014). Cultivable microorganisms associated with honeys of different geographical and botanical origin. Food Microbiol, 38, 284-294. https://doi.org/10.1016/j.fm.2013.07.013 Snowdon, J. A., & Cliver, D. O. (1996). Microorganisms in honey. International Journal of Food Microbiology, 31(1), 1-26. https://doi.org/https://doi.org/10.1016/0168-1605(96)00970-1 Strube, M. L. (2021). RibDif: can individual species be differentiated by 16S sequencing? Bioinformatics Advances, 1(1). https://doi.org/10.1093/bioadv/vbab020 Taylor, D. L., Walters, W. A., Lennon, N. J., Bochicchio, J., Krohn, A., Caporaso, J. G., Pennanen, T., & Cullen, D. (2016). Accurate Estimation of Fungal Diversity and Abundance through Improved Lineage-Specific Primers Optimized for Illumina Amplicon Sequencing. Applied and Environmental Microbiology, 82(24), 7217-7226. https://doi.org/doi:10.1128/AEM.02576-16 10 0 Utzeri, V. J., Ribani, A., Schiavo, G., Bertolini, F., Bovo, S., & Fontanesi, L. (2018). Application of next generation semiconductor based sequencing to detect the botanical composition of monofloral, polyfloral and honeydew honey. Food Control, 86, 342-349. https://doi.org/https://doi.org/10.1016/j.foodcont.2017.11.033 Vásquez, A., Forsgren, E., Fries, I., Paxton, R. J., Flaberg, E., Szekely, L., & Olofsson, T. C. (2012). Symbionts as major modulators of insect health: lactic acid bacteria and honeybees. PLoS One, 7(3), e33188. https://doi.org/10.1371/journal.pone.0033188 Vojvodic, S., Jensen, A. B., James, R. R., Boomsma, J. J., & Eilenberg, J. (2011). Temperature dependent virulence of obligate and facultative fungal pathogens of honeybee brood. Veterinary Microbiology, 149(1), 200-205. https://doi.org/https://doi.org/10.1016/j.vetmic.2010.10.001 Vojvodic, S., Rehan, S. M., & Anderson, K. E. (2013). Microbial gut diversity of Africanized and European honey bee larval instars. PLoS One, 8(8), e72106. https://doi.org/10.1371/journal.pone.0072106 Wen, Y., Wang, L., Jin, Y., Zhang, J., Su, L., Zhang, X., Zhou, J., & Li, Y. (2017). The Microbial Community Dynamics during the Vitex Honey Ripening Process in the Honeycomb [Original Research]. Frontiers in microbiology, 8. https://doi.org/10.3389/fmicb.2017.01649 Youngsteadt, E., Appler, R. H., López-Uribe, M. M., Tarpy, D. R., & Frank, S. D. (2015). Urbanization Increases Pathogen Pressure on Feral and Managed 10 1 Honey Bees. PLoS One, 10(11), e0142031. https://doi.org/10.1371/journal.pone.0142031 Yun, J. H., Jung, M. J., Kim, P. S., & Bae, J. W. (2018). Social status shapes the bacterial and fungal gut communities of the honey bee. Sci Rep, 8(1), 2019. https://doi.org/10.1038/s41598-018-19860-7 10 2 CHAPTER 3 PURIFICATION AND CHARACTERIZATION OF ANTIFUNGAL LIPOPEPTIDE PRODUCED BY BACILLUS VELEZENSIS ISOLATED FROM RAW HONEY Abstract Raw honey contains a diverse microbiota originating from honeybees, plants, and soil. Some gram-positive bacteria isolated from raw honey are known for their ability to produce secondary metabolites that have the potential to be exploited as antimicrobial agents. Currently, there is a high demand for natural, broad-spectrum, and eco-friendly bio-fungicides in the food industry. Naturally occurring antifungal products from food-isolated bacteria are ideal candidates for agricultural applications. To obtain novel antifungals from natural sources, we isolated bacteria from raw clover and orange blossom honey to evaluate their antifungal-producing potential. Two Bacillus velezensis isolates showed strong antifungal activity against food-isolated fungal strains. Antifungal compound production was optimized by adjusting the growth conditions of these bacterial isolates. Extracellular proteinaceous compounds were purified via ammonium sulfate precipitation, solid phase extraction, and RP-HPLC. Antifungal activity of purified products was confirmed by deferred overlay inhibition assay. Mass spectrometry (MS) was performed to determine the molecular weight of the isolated compounds. Whole genome sequencing (WGS) was conducted to predict secondary metabolite gene clusters encoded by the two antifungal-producing strains. Using MS and WGS data, we determined that the main antifungal compound produced 103 by these two Bacillus velezensis isolates was iturin A, a lipopeptide exhibiting broad spectrum antifungal activity. Citation: Xiong, Z. R., Cobo, M., Whittal, R. M., Snyder, A. B., & Worobo, R. W. (2022). Purification and characterization of antifungal lipopeptide produced by Bacillus velezensis isolated from raw honey. PLoS One, 17(4), e0266470. 104 Introduction Antifungal resistance in medically and agriculturally relevant fungi is increasing globally, straining the limited selection of safe and effective antifungal agents. The development of novel antifungal agents is much slower than the spread of antifungal resistant strains, which presents a serious human health and food security problem (Fisher, Hawkins, Sanglard, & Gurr, 2018). In the medical field, fungal infections are extremely difficult to treat. Fungicides that are broad spectrum, effective, and safe to use, are limited. Furthermore, the prevalence of multi-drug resistant fungal pathogens has been increasing in hospitals and nursing homes (Slifka, Kabbani, & Stone, 2020). For the widely deployed azole family, resistance has been observed in common fungal pathogens (Fisher, Hawkins, Sanglard, & Gurr, 2018). For example, multi-azole-resistant strains of the opportunistic pathogen Aspergillus fumigatus have been isolated from patients with invasive aspergillosis (van Paassen, Russcher, In 't Veld-van Wingerden, Verweij, & Kuijper, 2016). Fluconazole-resistant Candida glabrata with increased resistance to the other first-line antifungal drug echinocandin was also observed, which further limited the available options to treat this infection. (Alexander, Johnson, Pfeiffer, Jimenez-Ortigosa, Catania, Booker, Castanheira, Messer, Perlin, & Pfaller, 2013). Additionally, multidrug-resistant Candida auris, first isolated in 2009, has invasively infected patients worldwide through hospital-acquired transmission (Chowdhary, Sharma, & Meis, 2017). In the agricultural field, fungal plant pathogens have also acquired resistance against antifungal agents. Even though more fungicides are available for field application, the rapid rate of antifungal resistance development is alarming. A classic example of an 105 organism with high risk of antifungal resistance development is Botrytis cinerea, which is able to adapt to new fungicide classes. Multidrug-resistant B. cinerea strains have been isolated in strawberry fields around the world (Hahn, 2014). The predominant class of chemicals used for antifungal treatment of crops is azoles. Scientists have urged to restrict the use of azoles in agriculture, as resistant fungal strains are being continuously isolated from environmental and clinical settings at an increasing rate (Denning & Bromley, 2015). However, due to the lack of alternatives, it is still being widely used in economically important crops to avoid crop losses. In contemporary food systems, spoilage caused by fungi is no less serious. Food loss due to fungal spoilage was estimated to account for 5-10% of the world food supply, and post-harvest microbial spoilage was estimated to contribute to 25% of global food waste (Cook & Johnson, 2009; Gram, Ravn, Rasch, Bruhn, Christensen, & Givskov, 2002). In a survey of 51 juice manufacturers, 92% reported experiencing yeast or mold spoilage in their finished product and 89% reported previous occurrences of yeast or mold spoilage of their ingredients (Abigail B. Snyder & Randy W. Worobo, 2018). Spoilage fungi are difficult to control due to their ability to survive extreme conditions, like low water activity, limited nutrients, high acidity, and extreme heat treatment. Moreover, the trade-off of common fungal-controlling approaches in the food industry is the negative environmental impact, such as food waste, unsustainable packaging, and environmental damage by synthesized chemicals (A. B. Snyder & R. W. Worobo, 2018). Natural bio-fungicide could be a beneficial addition to traditional fungal-controlling approaches and mitigate the environmental impact. The urgent need for natural, novel, safe, and potent antifungal compounds lead 106 us to seek solutions from natural products, like honey. Raw honey is inhibitory to fungi, partially due to its high sugar content and low water activity (Molan, 2015). However, a survey comparing the antifungal effects of raw monofloral honey with synthetic honey demonstrated that heather and lavender honey exhibited higher antifungal activity than sugar-based synthetic honey (Estevinho, Afonso, & Feas, 2011; Feas & Estevinho, 2011). Other than osmotic inhibition, some chemical components in raw honey are also antifungal: hydrogen peroxide, flavonoids, phenolic acids, lysozymes, and other antioxidant compounds (Wahdan, 1998). Additionally, antifungal bacteria are present in raw honey. In previous studies, Bacillus spp. strains isolated from raw honey were able to produce a variety of secondary metabolites to inhibit the growth of other microorganisms and gain survival advantages. B. subtilis H215 was isolated from raw honey and it was inhibitory to Byssochlamys fulva H25 (H. Lee, Churey, & Worobo, 2008b). Another isolate found in US domestic honey, B. thuringiensis SF361, showed broad spectrum antifungal activity against Aspergillus, Penicillium, Byssochlamys, and Candida albicans (H. Lee, Churey, & Worobo, 2008a; Manns, Churey, & Worobo, 2012). Additionally, lactic acid bacteria isolated from honey samples including Lactobacillus plantarum, Lactobacillus curvatus, Pediococcus acidilactici, and Pediococcus pentosaceus showed inhibition against pathogenic Candida species (Bulgasem, Lani, Hassan, Wan Yusoff, & Fnaish, 2016). Both lactic acid bacteria and Bacillus spp. produce a variety of antifungal secondary metabolites including organic acids, volatile compounds, ribosomally synthesized peptides, and nonribosomal peptides (Caulier, Nannan, Gillis, Licciardi, Bragard, & Mahillon, 2019; Reis, Paula, Casarotti, & Penna, 2012; Schnürer & Magnusson, 107 2005). The potential application of these microbial natural products in the food industry, agricultural and medical field is promising. One example is nisin, a bacteriocin isolated from Lactococcus lactis subsp. lactis strain and exhibits broad- spectrum antibacterial activity (Reis, Paula, Casarotti, & Penna, 2012). Nisin is used in dairy and meat products as a biopreservative compound to inhibit foodborne pathogen Listeria monocytogenes (Martinez & Rodriguez, 2005). Additionally, several strains of B. subtilis, B. thuringiensis, and B. amyloliquefaciens were approved as commercial biopesticides by the Environmental Protection Agency (EPA) (U.S. Environmental Protection Agency, 2021). Lipopeptides secreted by these Bacillus species were used commercially as antifungal agents to control plant diseases caused by phytopathogens (Fira, Dimkic, Beric, Lozo, & Stankovic, 2018). In an effort to isolate novel antifungal compounds as candidates for medical and/or agricultural applications, we designed this study to isolate, purify, and characterize antifungal proteinaceous compounds from raw honey. Several Bacillus strains were isolated from raw clover and orange blossom honey. Extracellular antifungal compounds were purified via ammonium sulfate precipitation, solid phase extraction (SPE), and reversed-phase high performance liquid chromatography (RP- HPLC). Whole genome sequencing was performed on two antifungal producing strains identified as B. velezensis. Using a combination of genome secondary metabolite gene cluster analysis and mass spectrometry (MS), we determined that the antifungal compound belonged to the iturin family. Materials and methods Antifungal isolates selection 108 Raw clover honey and orange blossom honey were purchased from a local honey shop (Dundee, NY). Honey samples were diluted with 0.1% peptone water, and 100 μL of 10-1 and 10-2 dilutions were spread plated on tryptic soy agar (TSA) (BD Difco, Franklin Lakes, NJ). Plates were incubated at 30 °C for 24 hours. Visually distinct colonies were selected to test their antifungal activity. Eight fungal strains isolated from commercially processed food products were used as antifungal activity indicators (Snyder, Churey, & Worobo, 2019). Food-isolated fungal strains were incubated at ambient temperature on potato dextrose agar (PDA, BD Difco, Franklin Lakes, NJ) for at least 4 weeks prior to harvest. Fungal spores were harvested by flooding the surface of fully grown plates with 10 mL 0.1% Tween 80 (Sigma, St. Lois, MO). Spore suspension was filtered with several layers of sterile cheese cloth to remove debris and stored at -80 °C. Antifungal assays Antifungal activities of bacterial isolates were determined by deferred overlay inhibition assay: fungal spore suspensions were mixed with 10 mL 0.75% soft TSA and overlaid on PDA plates. Bacterial isolates were spotted with sterile toothpicks on the surface of solidified soft agar with fungal indicators. Plates were incubated at ambient temperature for 48 to 72 hours and inhibition zones were recorded. Bacterial colonies that showed antifungal properties were selected for further analysis. Bacterial isolates were stored in 20% glycerol at -80 °C. Bacterial classification through 16S rRNA gene sequencing Bacterial isolates exhibiting strong inhibition toward fungal indicators were initially identified by 16S rRNA gene sequencing. DNA was obtained using the 109 Genomic DNA extraction kit (Qiagen, Germantown, MD) and 16S rRNA genes were amplified through polymerase chain reaction (PCR). A set of primers (IDT, Coralville, IA) were used to amplify the conserved region in bacteria. 16S forward primer sequence: 5’-AGAGTTTGATCCTGGCTCAG-3’. 16S reverse primer sequence: 5’- AAGGAGGTGATCCAGCC-3’. PCR procedures were as follows: 3 μL DNA template, 1 μL forward primer and 1 μL reverse primer, 0.3 μL GoTaq Flexi DNA polymerase (Promega, Madison, WI), 10 μL 5X Colorless GoTaq Flexi buffer (Promega, Madison, WI), 4 μL 25 mM MgCl2 (Promega, Madison, WI), 2 μL 10 mM dNTP (New England Biolabs, Ipswich, MA), 29 μL dH2O. Total volume was 50 μL per PCR tube. Thermal cycling conditions were as follows: 1 cycle of 94 °C for 5 minutes, 35 cycles of 94 °C for 30 seconds, 50 °C for 1 minute, 72 °C for 2 minutes, 1 cycle of 72 °C for 10 minutes. PCR products were purified by QIAquick PCR purification kit (Qiagen, Germantown, MD). Purified DNA products were sent to Cornell University Biotechnology Resource Center (Ithaca, NY) for Sanger sequencing. The sequencing data were analyzed using NCBI Nucleotide Blast homology search to determine the species of those antifungal bacterial isolates (Altschul, Gish, Miller, Myers, & Lipman, 1990). Optimized production of antifungal compounds Different growth conditions were tested to optimize antifungal production by the honey isolates. Four media were selected for growth optimization: tryptic soy broth (TSB) (BD Difco, Franklin Lakes, NJ), brain-heart infusion (BHI) (BD Difco, Franklin Lakes, NJ) broth, 1.5% casamino acids (CAA) (BD Difco, Franklin Lakes, NJ) with 0.5% yeast extract (BD Difco, Franklin Lakes, NJ) broth, and potato dextrose 110 broth (PDB) (BD Difco, Franklin Lakes, NJ). Selected growth times were 24 hours or 48 hours, and selected incubation temperature and shaking speed combinations were 37 °C at 250 rpm or 30 °C at 150 rpm. Following the growth of each strain under each condition, the cell-free supernatant was tested for antifungal activity. Cultivated media was first centrifuged at 4 °C, 13000 x g for 10 minutes. Supernatant was then filtered through a 0.22 μm polyethersulfone (PES) bottle top filter (250 mL, Celltreat, Pepperell, MA). The cell-free filtrate was tested for antifungal activity using a well diffusion overlay inhibition assay. Wells were made on 25 mL PDA plates using the wide end of sterile 1000 μL pipette tips (diameter: 8.8 mm). A total volume of 600 μL filtrate was added to each well and dried in a biosafety cabinet. Fungal spores were suspended and mixed with 10 mL 0.75% soft TSA and poured onto PDA plates. Plates were incubated at ambient temperature for 48-72 hours, until the complete growth of fungi or the inhibition zone could be visualized. Clear inhibition zones were observed and recorded. Purification of antifungal proteinaceous compounds Two bacterial isolates WRB-ZX-001 and WRB-ZX-002 that showed the ability to excrete antifungal compounds into the broth media were selected for purification. Supernatant of the cell culture grown at optimized condition was treated with ammonium sulfate to precipitate proteins. Solid ammonium sulfate was added to the supernatant at 4 °C to reach saturation of 20%, 40%, 60%, 80% and 100%. Ammonium sulfate precipitates of each percentage saturation were collected separately by centrifugation at 13000 x g, 4 °C for 20 min and re-dissolved in sterile Milli-Q H2O. Precipitates were tested against fungal indicator strain A. fumigatus and 111 fractions showed antifungal activity were further purified by reversed-phase solid phase extraction (SPE) using a C18 sorbent cartridge (Sep-Pak Classic, Waters, Milford, MA) with acetonitrile as solvent. Acetonitrile with gradient concentrations from 0% to 100% with an increment of 10% was added to eluate the antifungal compounds. All fractions were tested against fungal indicator strain A. fumigatus through the well diffusion overlay inhibition assay as described before. Antifungal fractions from SPE were purified via high-performance liquid chromatography (HPLC, Agilent 1200 Series Gradient System, Santa Clara, CA). The following HPLC elution condition was used: 0–10 min mobile phase A (0.05% TFA in dH2O); 10–40 min a gradient of 0–100% mobile phase B (0.05% TFA in acetonitrile); and 40–50 min mobile phase B, with a flow rate of 1 mL/min. Fraction collection from HPLC was performed every 1.5 min. The active fractions were re-injected onto HPLC with the same elution condition to confirm its purity. The antifungal activity of HPLC collected fractions was determined by well diffusion overlay inhibition assay as mentioned previously. Antifungal activity units (AU/mL) of active ammonium sulfate precipitate, SPE fractions and HPLC collected fractions, defined as the reciprocal of the highest dilution yielding a clear inhibition zone, were calculated. Growth curve and antifungal production The growth curves of two selected bacterial isolates, WRB-ZX-001 and WRB- ZX-002, and their antifungal production over time were determined. These two isolates were pre-grown in 5 mL BHI broth at 30 °C, 150 rpm for 12 hours. Pre- growth cell culture (500 µL) was inoculated into 50 mL BHI broth. Samples were taken every two hours from 0 h to 96 h for cell density and antifungal activity 112 measurement. The absorbance of the samples was measured at 600 nm using a spectrophotometer (Spectronic 20D+, Thermo Scientific, Waltham, MA); absorbance values were used to plot growth curves for the two isolates. Antifungal activity was tested by well diffusion overlay inhibition assay of sterile-filtered supernatant against fungal indicator strain A. fumigatus. Cell-free supernatants were diluted two-fold and antifungal activity units were calculated as the reciprocal of the highest dilution showing a clear inhibition zone. Biological duplicates were performed. Data was analyzed and visualized in R version 4.0.2. R package growthcurver 0.3.0 was used to fit the microbial growth data to a standard form of logistic equation (Sprouffske & Wagner, 2016). Heat stability and protease stability test To measure the heat stability and protease stability of the antifungal compounds produced by WRB-ZX-001 and WRB-ZX-002, active antifungal fractions of ammonium sulfate precipitate were selected for testing. For heat stability, samples were treated by steam sterilization at 121 °C for 15 min in an autoclave. Antifungal activity was measured by deferred overlay inhibition assay of 10 µL 2-fold diluted heat-treated samples. The protease stability was tested by incubating the samples individually with 100 µg of pronase E (10 mg/mL, Sigma, St. Lois, MO), α- chymotrypsin (25 mg/mL, Sigma, St. Lois, MO), pepsin (20 mg/mL, Sigma, St. Lois, MO), and trypsin (2.5%, Sigma, St. Lois, MO) at 37 °C for 30 min. Antifungal activity was measured by deferred overlay inhibition assay of 10 µL 2-fold diluted protease- treated samples. Antifungal activity units of heat-treated and protease-treated samples were calculated. 113 Protein molecular weight determination via mass spectrometry Active fractions from SPE were analyzed with direct-infusion mass spectrometry (DIMS) to determine the molecular weight of the antifungal compounds. DIMS was performed on a Triversa Nanomate nanospray direct infusion robot (Advion, Ithaca, NY) attached to a Orbitrap Fusion Lumos Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA). Samples were diluted in 50 mM ammonium formate followed by centrifugation prior to direct infusion. Spectra were acquired in positive ion mode with a resolution setting of 500,000 (at m/z 200). Active fractions collected from HPLC were analyzed by liquid chromatography-mass spectrometry (LC-MS) to measure accurate mass of intact protein. Each sample was diluted with 0.1% formic acid and analyzed by LC-MS with a Dionex RSLCnano HPLC coupled to an OrbiTrap Fusion Lumos (Thermo Fisher Scientific, Waltham, MA) mass spectrometer using a 60 min gradient (2-90% acetonitrile). Sample was resolved using a 75 µm x 150 cm PepMap C4 column (Thermo Scientific, Waltham, MA). MS spectra of protein ions of different charge-states were acquired in positive ion mode with a resolution setting of 120,000 (at m/z 200) and accurate mass was deconvoluted using Xcalibur (Thermo Scientific, Waltham, MA). DIMS and LC-MS analyses were performed at Donald Danforth Plant Science Center, Proteomics & Mass Spectrometry Facility (St. Louis, MO). Whole genome sequencing and genome analysis Cell pellets from overnight BHI culture of the isolates were treated with lysozyme (20 mg/mL, Millipore Sigma, St. Lois, MO) and RNase A (Qiagen, Germantown, MD). Genomic DNA was extracted using QiaAMP DNA Minikit 114 (Qiagen, Germantown, MD). Library preparation, quality control, and sequencing were conducted by Cornell University Biotechnology Resource Center (Ithaca, NY) using Nextera XT DNA library preparation and indexing kits (Illumina, San Diego, CA). Illumina MiSeq (Illumina, San Diego, CA) was used to obtain 2 × 250 bp paired- end reads. Reads were trimmed using Trimmomatic (version 0.39) and de novo assembled with SPAdes (version 3.13.1) using the default k-mer settings for bacterial genome assembly (Bankevich, Nurk, Antipov, Gurevich, Dvorkin, Kulikov, Lesin, Nikolenko, Pham, Prjibelski, Pyshkin, Sirotkin, Vyahhi, Tesler, Alekseyev, & Pevzner, 2012; Bolger, Lohse, & Usadel, 2014). Scaffolds less than 500 bp were trimmed and assembly quality was assessed using QUAST (version 4.0) (Gurevich, Saveliev, Vyahhi, & Tesler, 2013). Average genome coverage was determined using BBmap (version 38.45) and SAMtools (version 1.11) (Li, Handsaker, Wysoker, Fennell, Ruan, Homer, Marth, Abecasis, Durbin, & Genome Project Data Processing, 2009). Genome assemblies of B. amyloliquefaciens group type strains were downloaded from the National Center for Biotechnology Information (NCBI) assembly database and average nucleotide identity (ANI) analysis of the isolates was conducted via the OrthoANI method using OAT (version 1.40) with BLAST+ (version 2.9.0) (I. Lee, Ouk Kim, Park, & Chun, 2016). The draft genomes of B. velezensis WRB-ZX-001 and WRB-ZX-002 sequenced in this study, the complete genome of B. subtilis 168 as an outgroup, and other 42 genomes of B. amyloliquefaciens group extracted from NCBI were used to construct a SNP-based phylogeny. The program kSNP v3.0 was used with a kmer size of 19 as determined by Kchooser (Gardner, Slezak, & Hall, 2015). The core SNPs were used to build the maximum likelihood 115 phylogeny in RAxML v8.2.12 under general time-reversible model with gamma distributed sites (GTRGAMMA) and 1000 bootstrap repetitions (Stamatakis, 2014). The phylogenetic tree was edited in FigTree v1.4.4 and deposited on Figshare (https://doi.org/10.6084/m9.figshare.16688839). The absolute core SNP distance matrix was calculated using Geneious v2020.2.4. Rapid annotation of the genomes was performed using prokka v1.12 (Seemann, 2014). Functional annotation of the predicted proteins was performed with BLAST2GO v1.4.4 (Conesa, Gotz, Garcia- Gomez, Terol, Talon, & Robles, 2005). Additionally, genome annotation was performed by the NCBI using the Prokaryotic Genome Annotation Pipeline (PGAP) database (Tatusova, DiCuccio, Badretdin, Chetvernin, Nawrocki, Zaslavsky, Lomsadze, Pruitt, Borodovsky, & Ostell, 2016). Putative bacteriocin genes were identified using BAGEL4 (van Heel, de Jong, Song, Viel, Kok, & Kuipers, 2018). Secondary metabolite genome mining pipeline (antiSMASH) was used to identify potential secondary metabolite synthesis gene clusters (Blin, Shaw, Steinke, Villebro, Ziemert, Lee, Medema, & Weber, 2019). Genome alignment between our isolates and the most closely related type strains was performed using BRIG (version 0.95) (Alikhan, Petty, Ben Zakour, & Beatson, 2011). Assembled genomes of B. velezensis WRB-ZX-001 and WRB-ZX-002 were submitted to Sequence Read Archive (SRA) and GenBank under the BioProject ID PRJNA580475 and PRJNA596478. SRA accession numbers are SRR10397796 and SRR10729003. Results Four of 15 bacterial isolates from clover honey and 8 of 23 isolates from 116 orange blossom honey yielded an inhibition zone when spotted on at least one fungal indicator. The 16S rRNA gene sequence of these 12 isolated strains showed highest identity to that of several Bacillus spp. To evaluate the antifungal potential of honey isolates, food-isolated fungal strains were selected as indicators for antifungal assay (Table 1). Cross reactivity of the honey bacterial isolates against these fungal strains and BLAST identification results were summarized in Table 2. Isolates that showed antifungal activity against at least three fungal indicators were selected for antifungal production in liquid broth. The production conditions, including the medium type, incubation temperature, and shaking speed, were optimized. As the only two isolates showing the ability to excrete antifungal compounds, isolate Co-29 and Co-30 were selected and renamed as WRB-ZX-001 and WRB-ZX-002 for the following experiments. These two isolates were grown in BHI broth at 30 °C, 150 rpm for 24 hours and 48 hours, and cell-free supernatant showed clear inhibition zones against fungal indicators. The antifungal compounds produced by the isolates were further purified and the isolates were whole genome sequenced. Table 3.1. Food-isolated fungal strains used in this study as indicators (adapted from Snyder, Churey, and Worobo (2019)). 117 Organism Strain ID Food source Syncephalastrum S11-0015 Raw sprouted almonds Aspergillus S11-0016 Nut mix Aspergillus S11-0033 Oatmeal A. fumigatus S11-0039 Kombucha A. niger S11-0054 Pomegranate juice Rhodotorula S11-0057 Red hot sauce P. glabrum S11-0071 Hard-boiled egg Cladosporium S11-0111 Juice beverage 118 Table 3.2. Summary of identity, source, and cross-reactivity against food-associated fungal indicators of honey bacterial isolates. Cross reactivityb Honey Aspergillus Aspergillus A. P. Isolates BLAST ID a Source Syncephalastrum S11-0016 S11-0033 A. fumigatus niger Rhodotorula glabrum Cladosporium Co-1 B. toyonensis Clover - - - + - - - - Co-5 B. toyonensis Clover - - + ++ ++ - ++ - Co-6 B. toyonensis Clover + - + ++ ++ + ++ - Co-10 B. aerius Clover \ \ \ + - \ + + Orange Co-17 B. cereus blossom + - + ++ + - + - Orange Co-18 B. megaterium blossom \ \ \ - - \ - - B. Orange Co-20 amyloliquefaciens blossom + \ - + + + - ++ Orange Co-21 B. cereus blossom + - + + - - - - B. Orange Co-26 amyloliquefaciens blossom + \ \ + - - + + B. Orange Co-29 amyloliquefaciens blossom + \ \ + + - - ++ B. Orange Co-30 amyloliquefaciens blossom + \ \ + + - - ++ Orange Co-33 B. aryabhattai blossom \ \ \ - - \ - - a BLAST ID was determined based on 16S rRNA gene homology search using NCBI Nucleotide BLAST tools. The species with the highest BLAST score were reported. 119 b Cross reactivity was determined using deferred overlay inhibition assay. The inhibition level against the fungal indicators was defined based on visual observation. “+”: low inhibition level. “++”: strong inhibition level. “-”: no observed inhibition. “\”: inconclusive result. Table 3.3. Antifungal activity of purification products of Bacillus velezensis isolates against food-isolated Aspergillus fumigatus. Antifungal activity unit (AU/mL) is defined as the reciprocal of the highest dilution showing a clear inhibition zone. Purification procedure Bacillus velezensis WRB-ZX-001 Bacillus velezensis WRB-ZX-002 Cell-free filtrate 20 AU/mL 40 AU/mL Ammonium sulfate precipitant 800 AU/mL 800 AU/mL Solid phase extraction eluate 800 AU/mL 1600 AU/mL HPLC fraction 200 AU/mL 200 AU/mL 120 Antifungal compounds produced by isolates WRB-ZX-001 and WRB-ZX-002 were first purified by ammonium sulfate precipitation of the cell-free culture supernatant. Precipitate from 60% ammonium sulfate showed the highest antifungal activity (Fig 1, Table 3). Ammonium sulfate precipitates were further purified by solid phase extraction with C18 columns and acetonitrile. The eluants for the optimal recovery of antifungal compounds were 50% and 60% acetonitrile. The SPE eluates were loaded onto HPLC, and fractions were collected to test for antifungal activity. Two major peaks were observed in the HPLC spectra and fractions with elution times between 28.5 min to 30 min for both isolates showed highest antifungal activity (200 AU/mL). These fractions were loaded once more onto HPLC to confirm their purity, and single peak was observed for both samples (Fig 2). SPE eluates (50% acetonitrile) and HPLC fraction collection samples (28.5 min to 30 min) for isolates WRB-ZX-001 and WRB-ZX-002 were analyzed with DIMS and LC-MS, respectively. The HPLC samples analyzed with LC-MS showed major peaks with m/z value of 1057.57 (Fig 3), which was also present in SPE sample WRB-ZX-002 (results not shown). Another compound with singly charged m/z value of 1043.55 and doubly charged m/z value of 522.28 was present in both SPE eluates and HPLC collected samples (S1 Fig.). Based on results from previous studies, we presumed that the ions with m/z value of 1043.55 and 1057.57 were C14 and C15 iturin A [M+H]+, respectively (Pathak & Keharia, 2014; Price, Rooney, Swezey, Perry, & Cohan, 2007). The molecular formula of C14 and C15 iturin A is C48H74N12O14 and C49H76N12O14, respectively (Peypoux, Guinand, Michel, Delcambe, Das, & Lederer, 1978). 121 Figure 3.1. Deferred inhibition assay of purified products from Bacillus velezensis WRB-ZX-001 and WRB-ZX-002 against food-isolated Aspergillus fumigatus. Precipitates of WRB-ZX-001 and WRB-ZX-002 from 60% ammonium sulfate were shown in A and B. Solid phase extraction eluates of 60% acetonitrile for WRB-ZX- 001 and WRB-ZX-002 were shown in C and D. Two-fold serial dilution was performed for all samples to determine the antifungal activity units. 122 Figure 3.2. Reversed-phase HPLC of purified products of Bacillus velezensis WRB-ZX-001 and WRB-ZX-002. Purification process included ammonium sulfate precipitation, solid phase extraction, and HPLC fraction collection. Single peaks shown in A and C were from isolate WRB-ZX-001 and WRB-ZX-002, respectively, and both have shown inhibition against fungal indicator strain Aspergillus fumigatus as shown in B and D. 123 Figure 3.3. Mass spectra for purified antifungal compounds produced by Bacillus velezensis WRB-ZX-001 and WRB-ZX-002. A and B are LC-MS spectra for HPLC collected active fraction of WRB-ZX-001 and WRB-ZX-002. Ion with m/z value of 1057.57 was assigned to C15 iturin A [M+H]+. Ion with m/z value of 1079.55 was assigned to C15 iturin A [M+Na]+. For isolate WRB-ZX-001, the 4,183,488 bp genome was assembled to 15 contigs with an average coverage of 104x and N50 of 685,546 bp. For isolate WRB- ZX-002, the genome size is 4,185,188 bp, and the genome was assembled to 15 contigs with an average coverage of 128x and N50 of 1,001,971 bp. Both isolates have the same GC content of 45.97%. Isolate WRB-ZX-001 contains an estimated 4,165 genes and 4,003 coding sequences (CDSs), while isolate WRB-ZX-002 contains an estimated 4,167 genes and 4,004 CDSs. To obtain functional labels, protein BLAST hits were mapped against the curated Gene Ontology (GO) database and GO terms were assigned to the query sequences, with 3,261 annotated sequences for isolate WRB-ZX-001 and 3,263 annotated sequences for isolate WRB-ZX-002. Based on BLAST2GO genome annotation results, the predicted CDSs were assigned to three 124 principal categories: biological process, cellular component, and molecular function. For isolates WRB-ZX-001 and WRB-ZX-002, the most abundant groups in the category of biological process were cellular process (38%), metabolic process (36%), and biological regulation (8%). In the category of cellular component, the most dominant terms were integral component of membrane (55%), cytoplasm (26%), and plasma membrane (16%). In the category of molecular function, the most representative terms were hydrolase activity (32%), oxidoreductase activity (17%), metal ion binding (13%), transmembrane transporter activity (11%), DNA binding (10%), and ATP binding (9%). Detailed GO annotation and node score distribution for Bacillus velezensis WRB-ZX-001 and WRB-ZX-002 was reported in S3 Table. To calculate average nucleotide identity (ANI) and classify the two isolates at species level, orthoANI analysis was performed. The type strain that isolates WRB-ZX-001 and WRB-ZX-002 were most closely related to was B. velezensis FZB42, with orthoANI values of 98.96% and 98.93% respectively. Based on the proposed species boundary of 95-96% orthoANI value, we concluded that both WRB-ZX-001 and WRB-ZX-002 should be classified as B. velezensis species (Goris, Konstantinidis, Klappenbach, Coenye, Vandamme, & Tiedje, 2007; I. Lee, Ouk Kim, Park, & Chun, 2016; Richter & Rossello-Mora, 2009). To elucidate the phylogenetic relationships between our two isolates and the closely related B. amyloliquefaciens group, a total of 42 reference genomes were obtained from the NCBI database. Forty-one B. amyloliquefaciens group isolates and one B. subtilis subsp. subtilis str. 168 were included in the phylogenetic analysis. The phylogenetic tree based on 4,035 core genome SNPs revealed close relatedness of the two isolates from this study with type 125 strains B. velezensis FZB42 and B. velezensis KACC18228 (Fig 4). Additional genome comparison of B. velezensis type strains FZB42 and CBMB205, B. amyloliquefaciens type strain DSM7, and isolates WRB-ZX-001 and WRB-ZX-002 was visualized with BRIG version 0.95 (Fig 5). Gaps in the circular chromosome represented regions with no homology to the reference strain B. velezensis FZB42. Gaps for the two isolates from our study were consistent due to high levels of nucleotide homology. Several gaps were present when comparing two isolates from this study with the closely related type strain B. velezensis FZB42, indicating the potential presence of novel gene products. To evaluate the secondary metabolite synthesis potential, genomes of WRB-ZX-001 and WRB-ZX-002 were annotated using NCBI Prokaryotic Genome Annotation Pipeline (PGAP) database. BAGEL4 was used to predict open reading frames (ORFs) for ribosomally synthesized proteins and peptides, including bacteriocins, ribosomally synthesized and post-translationally modified peptides (RiPPs). Five putative gene clusters of interest were identified by BAGEL4 in the genomes of B. velezensis WRB-ZX-001 and WRB-ZX-002. Both strains contained 3 contigs with genes related to the production of secondary metabolites, including antimicrobial peptide LCI and thiopeptide, bacteriocin amylocyclicin, linear azole/azoline-containing peptide (LAP), and lantibiotic cerecidin. Additionally, antiSMASH was used to identify secondary metabolite biosynthetic gene clusters (BGCs) including nonribosomal peptide synthetases (NRPSs), polyketide synthases (PKSs), RiPPs, and other antimicrobial synthases. A total of 16 putative BGCs were identified in both genomes, including 5 NRPSs for bacillibactin, fengycin, bacillomycin D, iturin and surfactin, three trans-acyl- 126 transferase polyketide synthases (transAT-PKS) for macrolactin H, bacillaene and difficidin, one type III PKS, three RiPP clusters for thiopeptide, lanthipeptide, amylocyclin, and others (Table 4). According to the results of antiSMASH analysis, both isolates contained a gene cluster with 88% similarity to iturin synthetase, and the predicted peptide sequence of the nonribosomal peptide is L‐Asn‐D‐Tyr‐D‐Asn‐L‐Gln‐L‐Pro‐D‐Asn‐L‐Ser. To further confirm the presence of iturin gene cluster, BLAST analysis was performed on both genomes. Four iturin genes (ituD, ituA, ituB, ituC) were detected in the genome of both isolates, with a similarity of 98.60% to itu operon complete CDS from the reference strains B. subtilis ZK0 (NCBI accession number: KT781920.1) and B. subtilis subsp. krictiensis str. ATCC 55079 (NCBI accession number: KU170613.1). The presence of iturin gene clusters in the genome further validated the MS data, indicating the production of C14- iturin (m/z of [M+H]+ 1043.55) and C +15-iturin (m/z of [M+H] 1057.57). 127 Figure 3.4. Core genome phylogeny of 43 Bacillus amyloliquefaciens group isolates. Maximum likelihood tree was constructed with core genome SNPs identified by kSNP. 41 reference genomes of Bacillus amyloliquefaciens group isolates were obtained from NCBI genome database. The core genome of Bacillus subtilis 168 was used as outgroup. Phylogeny was inferred by RAxML under time- reversible model with gamma distributed substitution sites and 1000 bootstrap repetitions. Bar represents 0.2 substitution per site. Isolates from this study were 128 labeled with red circle. 129 Figure 3.5. Genome comparison of Bacillus velezensis WRB-ZX-001 and WRB- ZX-002 against closely related Bacillus type strains. Bacillus velezensis FZB42 was used as the reference strain. The circular ring map was constructed by BLAST Ring Image Generator (BRIG, version 0.95). From inner to outer ring: 1) GC content; 2) Bacillus velezensis FZB42 nucleotide sequence; 3) GC Skew; 4) Bacillus velezensis WRB-ZX-001 nucleotide sequence; 5) Bacillus velezensis WRB-ZX-002 nucleotide sequence; 6) Bacillus velezensis CBMB205 nucleotide sequence; 7) Bacillus amyloliquefaciens DSM7 nucleotide sequence. 130 Table 3.4. Potential secondary metabolite synthesis gene clusters identified in Bacillus velezensis WRB-ZX-001 and WRB-ZX-002 by antiSMASH. 131 Strain Cluster Type Froma Toa Secondary metabolite Similarityb (%) 1 Other 298857 354273 Bacilysin 100 1 Other 500231 554177 Teichuronic acid 100 1 NRPS 876498 928287 Bacillibactin 100 1 RiPP 876498 928287 Amylocyclicin 100 2 PKS-like 65404 106648 \ \ 2 Terpene 189448 210188 \ \ 2 TransAT-PKS 557837 646070 Macrolactin H 100 3 T3PKS 212154 250873 \ \ 3 Terpene 314561 336444 \ \ 3 NRPS 360563 498152 Fengycin/Plipastatin 100 3 NRPS 360563 498152 Bacillomycin D 100 3 NRPS 360563 498152 Iturin 88 3 TransAT-PKS 560507 670621 Bacillaene 100 4 TransAT-PKS 85797 191987 Difficidin 100 5 Thiopeptide/LAP 114108 143862 \ \ 5 NRPS 154585 219992 Surfactin 91 Bacillus velezensis WRB-ZX-001 8 Class II lanthipeptide 47789 66288 \ \ Bacillus velezensis WRB-ZX-002 1 TransAT-PKS 550815 656586 Difficidin 100 132 1 T3PKS 954998 993717 \ \ 1 Terpene 1057405 1079288 \ \ 1 NRPS 1103407 1240996 Fengycin/Plipastatin 100 1 NRPS 1103407 1240996 Bacillomycin D 100 1 NRPS 1103407 1240996 Iturin 88 1 TransAT-PKS 1303351 1413465 Bacillaene 100 2 NRPS 73679 125468 Bacillibactin 100 2 Other 447789 501741 Teichuronic acid 100 2 Other 647699 703115 Bacilysin 100 3 PKS-like 65404 106648 \ \ 3 Terpene 189448 210188 \ \ 3 TransAT-PKS 557837 646070 Macrolactin H 100 4 Thiopeptide/LAP 114001 143734 \ \ 4 NRPS 154457 219864 Surfactin 91 6 Class II lanthipeptide 1 18500 \ \ a Location of gene clusters in the Bacillus velezensis genome. b Similarity based on BLAST analysis against known gene clusters. 133 Absorption at OD 600nm was used to plot the growth curve for isolate WRB- ZX-001 and WRB-ZX-002. Antifungal activity against fungal indicator A. fumigatus was calculated and plotted with the growth curve (Fig 6). The antifungal production started at 24 hours and 18 hours for WRB-ZX-001 and WRB-ZX-002, respectively. The highest antifungal production occurred after cells reached late stationary phase. The antifungal activity was quantified by serial dilution, which was the reason why the antifungal activity fluctuated before reaching maximum production. To optimize the production of antifungal compounds, bacterial cells were collected at 48 hours for the following experiments. The heat stability and protease stability for the antifungal compounds produced by WRB-ZX-001 and WRB-ZX-002 were tested and results were summarized in Table 5. After heat treatment using a 15 min, 121 °C cycle in the autoclave, a 2-fold decrease in antifungal activity for WRB-ZX-001 was observed while sample WRB-ZX-002 had no decrease. For the protease stability test, antifungal compounds produced by WRB-ZX-001 and WRB-ZX-002 showed resistance to pronase E, α-chymotrypsin, and trypsin, with no change in their antifungal activity compared to control. Only sample WRB-ZX-002 showed a 2-fold decrease in antifungal activity after treatment with pepsin. Based on these results, we concluded that antifungal compounds produced by WRB-ZX-001 and WRB-ZX-002 were heat- resistant and protease-resistant. 134 Figure 3.6. Growth curve and antifungal activity curve for Bacillus velezensis WRB-ZX-001 (A) and WRB-ZX-002 (B). Growth curve was plotted by measuring absorption at OD600nm every 30 min and a standard form of logistic equation was used to fit the absorption data (red line). Antifungal activity was measured by well diffusion overlay inhibition assay of serially diluted cell-free supernatant every two hours against fungal indicator strain Aspergillus fumigatus and data were shown in bar plots. 135 Table 3.5. Antifungal activity of heat-treated and protease-treated purified products of Bacillus velezensis isolates against food- isolated Aspergillus fumigatus. Treatment Bacillus velezensis WRB-ZX-001 Bacillus velezensis WRB-ZX-002 Control 800 AU/mL 800 AU/mL 121 °C, 15 min 400 AU/mL 800 AU/mL Pronase E 800 AU/mL 800 AU/mL Chymotrypsin 800 AU/mL 800 AU/mL Pepsin 800 AU/mL 400 AU/mL Trypsin 800 AU/mL 800 AU/mL 136 Discussion In general, bacterial spores are abundant in raw honey, many of which have the potential to exhibit antifungal properties (Bulgasem, Lani, Hassan, Wan Yusoff, & Fnaish, 2016). Previous studies have isolated Bacillus spp., Clostridium spp., Lactobacillus spp. and other lactic acid bacteria (LAB) from raw honey (Grabowski & Klein, 2017). Many members from LAB and Bacillus species have been shown to be antifungal, including Lactobacillus casei, Lactobacillus plantarum, B. subtilis and B. velezensis (Schnürer & Magnusson, 2005). Bioactive compounds, like ribosomally synthesized bacteriocins and non-ribosomally synthesized small peptides, can be produced by these bacteria, which could potentially be exploited for industrial and medical applications. In this study, our two antifungal B. velezensis isolates from raw honey are inhibitory against various food-isolated fungi (Table 2). Bacillus species devote a large portion of their genome to secondary metabolism, potentially due to competition they face in the environment (Chen, Koumoutsi, Scholz, Eisenreich, Schneider, Heinemeyer, Morgenstern, Voss, Hess, Reva, Junge, Voigt, Jungblut, Vater, Sussmuth, Liesegang, Strittmatter, Gottschalk, & Borriss, 2007). Bacillus species are ubiquitous in soil and the ocean, which often have complex microbial communities. By producing secondary metabolites that can inhibit closely related species and other microorganisms in the ecological niche, Bacillus species have gained significant survival advantages (Harwood, Mouillon, Pohl, & Arnau, 2018). Previous researchers have isolated a variety of secondary metabolites with antibacterial and antifungal properties from Bacillus species, some of which are nonribosomal peptides (NRPs) (Harwood, Mouillon, Pohl, & Arnau, 2018). NRPs are 137 synthesized by nonribosomal peptide synthetases (NRPSs) and independent of messenger RNA. NRPs usually go through extensive modifications, including glycosylation, acylation, and hydroxylation. Due to these modifications, some NRPs are amphiphilic and able to insert into cell membrane to form pores, like gramicidin, surfactin, fengycin, iturin, and other lipopeptides. Pore formation in cell membrane will lead to ion leakage and cell death (Maget-Dana, Harnois, & Ptak, 1989; Maget- Dana, Ptak, Peypoux, & Michel, 1985). Some NRPs target closely related cells while others have broad spectrum. Taking account of the results from LC-MS (Fig 3) and secondary metabolite genome mining pipeline (BAGEL4 and antiSMASH) (Table 4), we determined that the major broad-spectrum antifungal compound produced by our B. velezensis isolates was a nonribosomal lipopeptide, iturin A. The iturin A operon was demonstrated to contain four open reading frames (ORFs): ituD, ituA, ituB, and ituC. ituD encodes a putative malomyl coenzyme A transacylase, while ituA, ituB, and ituC encode iturin synthetases (Tsuge, Akiyama, & Shoda, 2001). Iturin A production is regulated by the promoter on the upstream of ituD (Tsuge, Akiyama, & Shoda, 2001). All four ORFs as well as the promoter Pitu were present in the genome of our two B. velezensis isolates based on BLAST search, with an identity of 98.6% to itu operon complete CDS. The iturin family is a group of cyclic lipopeptides with hydrophilic C-terminal heptapeptides and characteristic hydrophobic N-terminal β-amino fatty acids. The aliphatic chain of iturin contains between 14 to 17 carbons and the peptide chain has a chiral sequence of LDDLLDL (Penha, Vandenberghe, Faulds, Soccol, & Soccol, 2020). The iturin family primarily has broad-spectrum antifungal activity, with limited antibacterial activity (Cochrane & 138 Vederas, 2016). In our study, iturin-producing B. velezensis strains showed broad spectrum antifungal activities, with antagonistic ability against Aspergillus, Cladosporium, Syncephalastrum (Table 2), and Candida albicans (results not shown). The proposed antifungal mechanism for the iturin family is that they can interact with sterol components on the surface of fungal membrane and increase potassium permeability (Maget-Dana & Peypoux, 1994). Previous studies showed that iturin A can form ion-conducting pores on bimolecular lipid membranes and cholesterol can facilitate the pore-formation by expanding the open-state lifespan (Grau, Ortiz, de Godos, & Gomez-Fernandez, 2000; Maget-Dana, Harnois, & Ptak, 1989; Maget-Dana, Ptak, Peypoux, & Michel, 1985). Additionally, iturin is able to self-associate and interact with lipid membranes by forming a stoichiometric complex with cholesterol on the membrane surface (Maget-Dana & Peypoux, 1994). Furthermore, iturins with longer acyl chains have stronger antifungal properties due to their ability to form oligomers and insert deeply into target membranes to form ion-conducting pores (Malina & Shai, 2005). The pore-forming and membrane permeabilizing abilities of iturin A is concentration dependent. At high concentrations, iturin A showed higher antagonistic activity against fungal cells and higher hemolytic activity (Ines & Dhouha, 2015). In previous studies, Bacillus species have been demonstrated to be able to produce antifungal lipopeptides including members from iturin family. In a study by Pathak and Keharia (2014), iturin isomers and surfactin families were isolated from crude extract of B. subtilis. Iturin A2 and Iturin A3/A4/A5 were found to have broad spectrum antifungal activities against Aspergillus, Fusarium, Chrysosporium, Candida 139 albicans, Trichosporium, Alternaria, and Cladosporium (Pathak & Keharia, 2014). One of the iturin A homologues in their study had a mass of 1057.5, the same as the iturin isolated from our study. Similar to the results from our research, Gong et al. (2006) identified antifungal lipopeptides from B. subtilis strain PY-1 that was temperature stable and protease resistant. By using ESI-TOF MS, FAB-MS/MS CID spectrometry and NMR, they identified the antifungal compounds as iturin A isomers and determined that the (M+H)+ ions at m/z 1057 were iturin A3 and A4 (C17 aliphatic chain) (M. Gong, Wang, Zhang, Yang, Lu, Pei, & Cheng, 2006). Moreover, another group of researchers isolated B. amyloliquefaciens S76-3 from wheat spikes, which produced antifungal lipopeptides active against Fusarium graminearum. These lipopeptides were identified through RP-HPLC and ESI-MS, with iturin A and plipastatin A being the most abundant molecules. The m/z value of iturin A with C-14 acyl acid chain was 1043.35. Fluorescence microscopy analyses and transmission electron microscopy (TEM) analyses of lipopeptide-treated Fusarium graminearum conidia and hyphae showed damages to cell wall and plasma membrane, which was consistent with the proposed antifungal mechanism of iturin family (A. D. Gong, Li, Yuan, Song, Yao, He, Zhang, & Liao, 2015). Overall, itu operon is common in B. subtilis group and B. amyloliquefaciens group, and our B. velezensis isolates were demonstrated to possess itu operon and produce C14-15 iturin A. The production of iturin and other lipopeptides by Bacillus species is dependent on the environment factors, including temperature, pH, carbon source, and oxygen availability. Iturin is mainly produced at temperature between 25 °C and 37 °C under aerobic conditions (Jacques, 2011). In our study, the optimum temperature for 140 the production of iturin A by B. velezensis strains was 30 °C. A neutral pH is generally favorable for the production of lipopeptides (Ines & Dhouha, 2015). In our study, iturin production was optimized by adjusting the pH of BHI broth to 7.4. In a recent study by Dang et al (2019), the optimal condition for the iturin A production by B. amyloliquefaciens LL3 derivative strain was thoroughly investigated using single factor optimization and response surface methodology. It was determined that inulin was the best carbon source and L-sodium glutamate was the best nitrogen source. The optimal production condition was determined to be pH 7.0 and 27 °C with 7, 15 and 0.5 g/L of inulin, L-sodium glutamate and MgSO4 (Dang, Zhao, Liu, Fan, Huang, Gao, Wang, & Yang, 2019). In our future studies, this condition will be validated to optimize the production of iturin by our B. velezensis isolates. Compared to conventional synthetic fungicides, which raise concerns regarding chemical residues and antibiotic resistance, biocontrol agents synthesized by living organisms are relatively more environmentally friendly for agricultural applications (Meena & Kanwar, 2015; Ongena & Jacques, 2008). Lipopeptides, like iturin, are considered safe, biodegradable, and eco-friendly. Some previous studies have demonstrated their potential application. Lipopeptides produced by B. subtilis RB14, containing iturin A and surfactin, were effective at suppressing the damping-off of tomato seedings cause by Rhizoctonia solani. A mutant of B. subtilis RB14 that cannot produce iturin A or surfactin failed to inhibit R. solani. Restoration of the gene successfully reinstated the suppressibility toward the fungal disease (Asaka & Shoda, 1996). Another study constructed a mutant of B. subtilis ATCC6633 by replacing native promoter with constitutive promoter to increase the production of mycosubtilin. 141 The mutant strain was able to reduce Pythium infection in tomato seedlings and increase germination rate (Leclere, Bechet, Adam, Guez, Wathelet, Ongena, Thonart, Gancel, Chollet-Imbert, & Jacques, 2005). Romero et al (2007) showed in their study that direct application of lipopeptide-producing B. subtilis cells or cell-free filtrate to leaf surface can prevent powdery mildew caused by Podosphaera fusca. Furthermore, by using site-directed mutagenesis, they demonstrated that bacterial mutants that lost the ability to produce bacillomycin, fengycin or iturin A were not able to control the powdery mildew disease (Romero, de Vicente, Rakotoaly, Dufour, Veening, Arrebola, Cazorla, Kuipers, Paquot, & Perez-Garcia, 2007). Antifungal lipopeptides produced by Bacillus species could have additional applications. These lipopeptides possess the ability to change biofilm formation, motility, and virulence gene expression of various microorganisms. It is also associated with plant root colonization, plant defense, and plant growth promotion (Raaijmakers, De Bruijn, Nybroe, & Ongena, 2010). With the increased need for biopesticides that have high specificity, low environmental persistence, and low toxicity, industrial exploitation of these chemicals or compounds derived from natural products as food preservatives and crop protection agents is continuously expanding (Seiber, Coats, Duke, & Gross, 2014). Iturin A, a biopesticide produced by food-isolated Bacillus spp. and naturally present in food systems, can be exploited for industrial applications. The safety of the producer strains and their products need to be evaluated to achieve broader application of iturin A produced by Bacillus species. One of the major producer strains for iturin A, B. subtilis, has Qualified Presumption of Safety (QPS) status according to European Food Standards Authority (EFSA), which indicates that 142 this strain does not harbor acquired antimicrobial resistance (AMR) genes or exhibit toxigenic activity (Harwood, Mouillon, Pohl, & Arnau, 2018). B. velezensis FZB42 type strain, which is closely related to our isolates, was also evaluated by US Environmental Protection Agency (EPA) and considered not toxic, pathogenic, or infective. Therefore, a tolerance exemption for residues of B. velezensis FZB42 in food commodities was established. As for the surface-active agents produced by these strains, including surfactin, iturin and other detergents, they can penetrate cell membrane but are not necessarily cytotoxic. Toxicity assays need to be developed to determine their cytotoxicity specifically. For now, current safety measures taken by the industry, including historical safety data and routine testing of the strains and products, are sufficient to ensure the safe usage of strains from the B. subtilis and B. amyloliquefaciens group as enzyme production workhorse (Harwood, Mouillon, Pohl, & Arnau, 2018). Regarding the safety of iturin A, the acute and subacute toxicity was previously evaluated in mouse models. Preliminary toxicology study showed that iturin A can induce hepatotoxicity and was deposited in liver, lung, and spleen. However, organ-specific toxicity of iturin A was reversible after discontinuation of treatment, which indicated that medical application is still possible (Dey, Bharti, Banerjee, Das, Das, Das, Jena, Misra, Sen, & Mandal, 2016). On the other hand, in a study by Zhao et al (2018), iturin produced by B. subtilis was intragastrically administrated to mouse models. In acute (7-day) and subacute (28-day) toxicity tests under concentration of 5000 mg/kg and 2000 mg/kg, respectively, iturin was deemed safe and non-toxic, with no significant damage to liver, kidney, or small intestines (Zhao, Li, Zhang, Lei, Zhao, Shao, Jiang, Shi, & Sun, 2018). Overall, rigorous and 143 large scale in vivo and clinical studies are still needed to fully understand the potential toxicity of iturin A. A large portion of recent publications on antifungal lipopeptides produced by Bacillus species and other gram-positive bacteria focused on partially purified mixtures with varying antifungal activities. The chemical identities of these semi- purified compounds remain uncharacterized, and the biological implications of these studies remain unclear, which poses barriers to future studies. Moreover, with the increased availability and popularity of next-generation sequencing (NGS) and genome mining tools, more recent studies are using these tools to evaluate the secondary metabolites produced by Bacillus species. However, chemical confirmation of those potential metabolites is falling behind. Future studies need to combine genetic and genomic methods with traditional chemical identification methods to properly identify, classify, and characterize these potential secondary metabolites. As for the future of iturin family, additional studies are necessary to improve the production efficiency and evaluate the resistance development in fungal model systems. 144 REFERENCES Alexander, B. D., Johnson, M. D., Pfeiffer, C. D., Jimenez-Ortigosa, C., Catania, J., Booker, R., . . . Pfaller, M. A. (2013). Increasing echinocandin resistance in Candida glabrata: clinical failure correlates with presence of FKS mutations and elevated minimum inhibitory concentrations. Clinical Infectious Diseases, 56(12), 1724-1732. https://doi.org/10.1093/cid/cit136 Alikhan, N. F., Petty, N. K., Ben Zakour, N. L., & Beatson, S. A. (2011). BLAST Ring Image Generator (BRIG): simple prokaryote genome comparisons. BMC Genomics, 12(1), 402. https://doi.org/10.1186/1471-2164-12-402 Altschul, S. F., Gish, W., Miller, W., Myers, E. W., & Lipman, D. J. (1990). Basic local alignment search tool. Journal of Molecular Biology, 215(3), 403-410. https://doi.org/10.1016/S0022-2836(05)80360-2 Asaka, O., & Shoda, M. (1996). Biocontrol of Rhizoctonia solani Damping-Off of Tomato with Bacillus subtilis RB14. Applied and Environmental Microbiology, 62(11), 4081-4085. https://doi.org/10.1128/aem.62.11.4081- 4085.1996 Bankevich, A., Nurk, S., Antipov, D., Gurevich, A. A., Dvorkin, M., Kulikov, A. S., . . . Pevzner, P. A. (2012). SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. Journal of Computational Biology, 19(5), 455-477. https://doi.org/10.1089/cmb.2012.0021 Blin, K., Shaw, S., Steinke, K., Villebro, R., Ziemert, N., Lee, S. Y., . . . Weber, T. (2019). antiSMASH 5.0: updates to the secondary metabolite genome mining 145 pipeline. Nucleic Acids Research, 47(W1), W81-W87. https://doi.org/10.1093/nar/gkz310 Bolger, A. M., Lohse, M., & Usadel, B. (2014). Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics, 30(15), 2114-2120. https://doi.org/10.1093/bioinformatics/btu170 Bulgasem, B. Y., Lani, M. N., Hassan, Z., Wan Yusoff, W. M., & Fnaish, S. G. (2016). Antifungal Activity of Lactic Acid Bacteria Strains Isolated from Natural Honey against Pathogenic Candida Species. Mycobiology, 44(4), 302- 309. https://doi.org/10.5941/MYCO.2016.44.4.302 Caulier, S., Nannan, C., Gillis, A., Licciardi, F., Bragard, C., & Mahillon, J. (2019). Overview of the Antimicrobial Compounds Produced by Members of the Bacillus subtilis Group. Front Microbiol, 10, 302. https://doi.org/10.3389/fmicb.2019.00302 Chen, X. H., Koumoutsi, A., Scholz, R., Eisenreich, A., Schneider, K., Heinemeyer, I., . . . Borriss, R. (2007). Comparative analysis of the complete genome sequence of the plant growth-promoting bacterium Bacillus amyloliquefaciens FZB42. Nature Biotechnology, 25(9), 1007-1014. https://doi.org/10.1038/nbt1325 Chowdhary, A., Sharma, C., & Meis, J. F. (2017). Candida auris: A rapidly emerging cause of hospital-acquired multidrug-resistant fungal infections globally. PLoS Pathog, 13(5), e1006290. https://doi.org/10.1371/journal.ppat.1006290 Cochrane, S. A., & Vederas, J. C. (2016). Lipopeptides from Bacillus and Paenibacillus spp.: A Gold Mine of Antibiotic Candidates. Med Res Rev, 36(1), 4-31. https://doi.org/10.1002/med.21321 146 Conesa, A., Gotz, S., Garcia-Gomez, J. M., Terol, J., Talon, M., & Robles, M. (2005). Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics, 21(18), 3674-3676. https://doi.org/10.1093/bioinformatics/bti610 Cook, F. K., & Johnson, B. L. (2009). Microbiological spoilage of cereal products. In Compendium of the microbiological spoilage of foods and beverages (pp. 223- 244). Springer. Dang, Y., Zhao, F., Liu, X., Fan, X., Huang, R., Gao, W., . . . Yang, C. (2019). Enhanced production of antifungal lipopeptide iturin A by Bacillus amyloliquefaciens LL3 through metabolic engineering and culture conditions optimization. Microb Cell Fact, 18(1), 68. https://doi.org/10.1186/s12934-019- 1121-1 Denning, D. W., & Bromley, M. J. (2015). Infectious Disease. How to bolster the antifungal pipeline. Science, 347(6229), 1414-1416. https://doi.org/10.1126/science.aaa6097 Dey, G., Bharti, R., Banerjee, I., Das, A. K., Das, C. K., Das, S., . . . Mandal, M. (2016). Pre-clinical risk assessment and therapeutic potential of antitumor lipopeptide ‘Iturin A’ in an in vivo and in vitro model. RSC Advances, 6(75), 71612-71623. https://doi.org/10.1039/c6ra13476a Estevinho, M. L., Afonso, S. E., & Feas, X. (2011). Antifungal effect of lavender honey against Candida albicans , Candida krusei and Cryptococcus neoformans. J Food Sci Technol, 48(5), 640-643. https://doi.org/10.1007/s13197-011-0243-1 147 Feas, X., & Estevinho, M. L. (2011). A survey of the in vitro antifungal activity of heather (Erica sp.) organic honey. Journal of Medicinal Food, 14(10), 1284- 1288. https://doi.org/10.1089/jmf.2010.0211 Fira, D., Dimkic, I., Beric, T., Lozo, J., & Stankovic, S. (2018). Biological control of plant pathogens by Bacillus species. Journal of Biotechnology, 285, 44-55. https://doi.org/10.1016/j.jbiotec.2018.07.044 Fisher, M. C., Hawkins, N. J., Sanglard, D., & Gurr, S. J. (2018). Worldwide emergence of resistance to antifungal drugs challenges human health and food security. Science, 360(6390), 739-742. https://doi.org/10.1126/science.aap7999 Gardner, S. N., Slezak, T., & Hall, B. G. (2015). kSNP3.0: SNP detection and phylogenetic analysis of genomes without genome alignment or reference genome. Bioinformatics, 31(17), 2877-2878. https://doi.org/10.1093/bioinformatics/btv271 Gong, A. D., Li, H. P., Yuan, Q. S., Song, X. S., Yao, W., He, W. J., . . . Liao, Y. C. (2015). Antagonistic mechanism of iturin A and plipastatin A from Bacillus amyloliquefaciens S76-3 from wheat spikes against Fusarium graminearum. PLoS One, 10(2), e0116871. https://doi.org/10.1371/journal.pone.0116871 Gong, M., Wang, J. D., Zhang, J., Yang, H., Lu, X. F., Pei, Y., & Cheng, J. Q. (2006). Study of the antifungal ability of Bacillus subtilis strain PY-1 in vitro and identification of its antifungal substance (iturin A). Acta Biochim Biophys Sin (Shanghai), 38(4), 233-240. https://doi.org/10.1111/j.1745-7270.2006.00157.x Goris, J., Konstantinidis, K. T., Klappenbach, J. A., Coenye, T., Vandamme, P., & Tiedje, J. M. (2007). DNA-DNA hybridization values and their relationship to 148 whole-genome sequence similarities. International Journal of Systematic and Evolutionary Microbiology, 57(Pt 1), 81-91. https://doi.org/10.1099/ijs.0.64483-0 Grabowski, N. T., & Klein, G. (2017). Microbiology and foodborne pathogens in honey. Critical Reviews in Food Science and Nutrition, 57(9), 1852-1862. https://doi.org/10.1080/10408398.2015.1029041 Gram, L., Ravn, L., Rasch, M., Bruhn, J. B., Christensen, A. B., & Givskov, M. (2002). Food spoilage—interactions between food spoilage bacteria. International Journal of Food Microbiology, 78(1-2), 79-97. https://doi.org/10.1016/s0168-1605(02)00233-7 Grau, A., Ortiz, A., de Godos, A., & Gomez-Fernandez, J. C. (2000). A biophysical study of the interaction of the lipopeptide antibiotic iturin A with aqueous phospholipid bilayers. Arch Biochem Biophys, 377(2), 315-323. https://doi.org/10.1006/abbi.2000.1791 Gurevich, A., Saveliev, V., Vyahhi, N., & Tesler, G. (2013). QUAST: quality assessment tool for genome assemblies. Bioinformatics, 29(8), 1072-1075. https://doi.org/10.1093/bioinformatics/btt086 Hahn, M. (2014). The rising threat of fungicide resistance in plant pathogenic fungi: Botrytis as a case study. Journal of chemical biology, 7(4), 133-141. Harwood, C. R., Mouillon, J. M., Pohl, S., & Arnau, J. (2018). Secondary metabolite production and the safety of industrially important members of the Bacillus subtilis group. FEMS Microbiology Reviews, 42(6), 721-738. https://doi.org/10.1093/femsre/fuy028 149 Ines, M., & Dhouha, G. (2015). Lipopeptide surfactants: Production, recovery and pore forming capacity. Peptides, 71, 100-112. https://doi.org/10.1016/j.peptides.2015.07.006 Jacques, P. (2011). Surfactin and Other Lipopeptides from Bacillus spp. In G. Soberón-Chávez (Ed.), Biosurfactants (pp. 57-91). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-642-14490-5_3 Leclere, V., Bechet, M., Adam, A., Guez, J. S., Wathelet, B., Ongena, M., . . . Jacques, P. (2005). Mycosubtilin overproduction by Bacillus subtilis BBG100 enhances the organism's antagonistic and biocontrol activities. Applied and Environmental Microbiology, 71(8), 4577-4584. https://doi.org/10.1128/AEM.71.8.4577-4584.2005 Lee, H., Churey, J. J., & Worobo, R. W. (2008a). Antimicrobial activity of bacterial isolates from different floral sources of honey. International Journal of Food Microbiology, 126(1-2), 240-244. https://doi.org/10.1016/j.ijfoodmicro.2008.04.030 Lee, H., Churey, J. J., & Worobo, R. W. (2008b). Purification and structural characterization of bacillomycin F produced by a bacterial honey isolate active against Byssochlamys fulva H25. Journal of Applied Microbiology, 105(3), 663-673. https://doi.org/10.1111/j.1365-2672.2008.03797.x Lee, I., Ouk Kim, Y., Park, S. C., & Chun, J. (2016). OrthoANI: An improved algorithm and software for calculating average nucleotide identity. International Journal of Systematic and Evolutionary Microbiology, 66(2), 1100-1103. https://doi.org/10.1099/ijsem.0.000760 150 Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., . . . Genome Project Data Processing, S. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25(16), 2078-2079. https://doi.org/10.1093/bioinformatics/btp352 Maget-Dana, R., Harnois, I., & Ptak, M. (1989). Interactions of the lipopeptide antifungal iturin A with lipids in mixed monolayers. Biochim Biophys Acta, 981(2), 309-314. https://doi.org/10.1016/0005-2736(89)90042-4 Maget-Dana, R., & Peypoux, F. (1994). Iturins, a special class of pore-forming lipopeptides: biological and physicochemical properties. Toxicology, 87(1-3), 151-174. https://doi.org/10.1016/0300-483x(94)90159-7 Maget-Dana, R., Ptak, M., Peypoux, F., & Michel, G. (1985). Pore-forming properties of iturin A, a lipopeptide antibiotic. Biochim Biophys Acta, 815(3), 405-409. https://doi.org/10.1016/0005-2736(85)90367-0 Malina, A., & Shai, Y. (2005). Conjugation of fatty acids with different lengths modulates the antibacterial and antifungal activity of a cationic biologically inactive peptide. Biochemical Journal, 390(Pt 3), 695-702. https://doi.org/10.1042/BJ20050520 Manns, D. C., Churey, J. J., & Worobo, R. W. (2012). Functional assignment of YvgO, a novel set of purified and chemically characterized proteinaceous antifungal variants produced by Bacillus thuringiensis SF361. Applied and Environmental Microbiology, 78(8), 2543-2552. https://doi.org/10.1128/AEM.07727-11 151 Martinez, B., & Rodriguez, A. (2005). Antimicrobial susceptibility of nisin resistant Listeria monocytogenes of dairy origin. FEMS Microbiology Letters, 252(1), 67-72. https://doi.org/10.1016/j.femsle.2005.08.025 Meena, K. R., & Kanwar, S. S. (2015). Lipopeptides as the antifungal and antibacterial agents: applications in food safety and therapeutics. Biomed Res Int, 2015, 473050. https://doi.org/10.1155/2015/473050 Molan, P. C. (2015). The Antibacterial Activity of Honey. Bee World, 73(1), 5-28. https://doi.org/10.1080/0005772x.1992.11099109 Ongena, M., & Jacques, P. (2008). Bacillus lipopeptides: versatile weapons for plant disease biocontrol. Trends in Microbiology, 16(3), 115-125. https://doi.org/10.1016/j.tim.2007.12.009 Pathak, K. V., & Keharia, H. (2014). Identification of surfactins and iturins produced by potent fungal antagonist, Bacillus subtilis K1 isolated from aerial roots of banyan (Ficus benghalensis) tree using mass spectrometry. 3 Biotech, 4(3), 283-295. https://doi.org/10.1007/s13205-013-0151-3 Penha, R. O., Vandenberghe, L. P. S., Faulds, C., Soccol, V. T., & Soccol, C. R. (2020). Bacillus lipopeptides as powerful pest control agents for a more sustainable and healthy agriculture: recent studies and innovations. Planta, 251(3), 70. https://doi.org/10.1007/s00425-020-03357-7 Peypoux, F., Guinand, M., Michel, G., Delcambe, L., Das, B. C., & Lederer, E. (1978). Structure of iturine A, a peptidolipid antibiotic from Bacillus subtilis. Biochemistry, 17(19), 3992-3996. https://doi.org/10.1021/bi00612a018 152 Price, N. P. J., Rooney, A. P., Swezey, J. L., Perry, E., & Cohan, F. M. (2007). Mass spectrometric analysis of lipopeptides from Bacillus strains isolated from diverse geographical locations. FEMS Microbiology Letters, 271(1), 83-89. https://doi.org/10.1111/j.1574-6968.2007.00702.x Raaijmakers, J. M., De Bruijn, I., Nybroe, O., & Ongena, M. (2010). Natural functions of lipopeptides from Bacillus and Pseudomonas: more than surfactants and antibiotics. FEMS Microbiology Reviews, 34(6), 1037-1062. https://doi.org/10.1111/j.1574-6976.2010.00221.x Reis, J., Paula, A., Casarotti, S., & Penna, A. (2012). Lactic acid bacteria antimicrobial compounds: characteristics and applications. Food Engineering Reviews, 4(2), 124-140. Richter, M., & Rossello-Mora, R. (2009). Shifting the genomic gold standard for the prokaryotic species definition. Proc Natl Acad Sci U S A, 106(45), 19126- 19131. https://doi.org/10.1073/pnas.0906412106 Romero, D., de Vicente, A., Rakotoaly, R. H., Dufour, S. E., Veening, J. W., Arrebola, E., . . . Perez-Garcia, A. (2007). The iturin and fengycin families of lipopeptides are key factors in antagonism of Bacillus subtilis toward Podosphaera fusca. Mol Plant Microbe Interact, 20(4), 430-440. https://doi.org/10.1094/MPMI-20-4-0430 Schnürer, J., & Magnusson, J. (2005). Antifungal lactic acid bacteria as biopreservatives. Trends in Food Science & Technology, 16(1-3), 70-78. https://doi.org/10.1016/j.tifs.2004.02.014 153 Seemann, T. (2014). Prokka: rapid prokaryotic genome annotation. Bioinformatics, 30(14), 2068-2069. https://doi.org/10.1093/bioinformatics/btu153 Seiber, J. N., Coats, J., Duke, S. O., & Gross, A. D. (2014). Biopesticides: state of the art and future opportunities. Journal of Agricultural and Food Chemistry, 62(48), 11613-11619. https://doi.org/10.1021/jf504252n Slifka, K. M. J., Kabbani, S., & Stone, N. D. (2020). Prioritizing prevention to combat multidrug resistance in nursing homes: A call to action. Journal of the American Medical Directors Association, 21(1), 5-7. Snyder, A. B., Churey, J. J., & Worobo, R. W. (2019). Association of fungal genera from spoiled processed foods with physicochemical food properties and processing conditions. Food Microbiology, 83, 211-218. https://doi.org/10.1016/j.fm.2019.05.012 Snyder, A. B., & Worobo, R. W. (2018). Fungal Spoilage in Food Processing. Journal of Food Protection, 81(6), 1035-1040. https://doi.org/10.4315/0362-028X.JFP- 18-031 Snyder, A. B., & Worobo, R. W. (2018). The incidence and impact of microbial spoilage in the production of fruit and vegetable juices as reported by juice manufacturers. Food Control, 85, 144-150. https://doi.org/10.1016/j.foodcont.2017.09.025 Sprouffske, K., & Wagner, A. (2016). Growthcurver: an R package for obtaining interpretable metrics from microbial growth curves. BMC Bioinformatics, 17, 172. https://doi.org/10.1186/s12859-016-1016-7 154 Stamatakis, A. (2014). RAxML version 8: a tool for phylogenetic analysis and post- analysis of large phylogenies. Bioinformatics, 30(9), 1312-1313. https://doi.org/10.1093/bioinformatics/btu033 Tatusova, T., DiCuccio, M., Badretdin, A., Chetvernin, V., Nawrocki, E. P., Zaslavsky, L., . . . Ostell, J. (2016). NCBI prokaryotic genome annotation pipeline. Nucleic Acids Research, 44(14), 6614-6624. https://doi.org/10.1093/nar/gkw569 Tsuge, K., Akiyama, T., & Shoda, M. (2001). Cloning, sequencing, and characterization of the iturin A operon. Journal of Bacteriology, 183(21), 6265-6273. https://doi.org/10.1128/JB.183.21.6265-6273.2001 U.S. Environmental Protection Agency. (2021). Biopesticides. Retrieved from https://www.epa.gov/pesticides/biopesticides on January 04, 2021. Retrieved January 04, 2021 from https://www.epa.gov/pesticides/biopesticides van Heel, A. J., de Jong, A., Song, C., Viel, J. H., Kok, J., & Kuipers, O. P. (2018). BAGEL4: a user-friendly web server to thoroughly mine RiPPs and bacteriocins. Nucleic Acids Research, 46(W1), W278-W281. https://doi.org/10.1093/nar/gky383 van Paassen, J., Russcher, A., In 't Veld-van Wingerden, A. W., Verweij, P. E., & Kuijper, E. J. (2016). Emerging aspergillosis by azole-resistant Aspergillus fumigatus at an intensive care unit in the Netherlands, 2010 to 2013. Euro Surveill, 21(30), 30300. https://doi.org/10.2807/1560- 7917.ES.2016.21.30.30300 155 Wahdan, H. A. (1998). Causes of the antimicrobial activity of honey. Infection, 26(1), 26-31. https://doi.org/10.1007/BF02768748 Zhao, H., Li, J., Zhang, Y., Lei, S., Zhao, X., Shao, D., . . . Sun, H. (2018). Potential of iturins as functional agents: safe, probiotic, and cytotoxic to cancer cells. Food Funct, 9(11), 5580-5587. https://doi.org/10.1039/c8fo01523f 156 CHAPTER 4 LOOKING AHEAD: UNLOCK THE FULL POTENTIAL OF RAW HONEY Honey is a valuable food with antimicrobial properties. It is widely used in folk medicines since ancient times for wound and burn care. We attributed part of the antimicrobial activities of raw honey to its microbiota and antimicrobial metabolites in our studies. The honey microbiome is a great reservoir of natural antimicrobials with potential industrial applications. Environmental factors influence the physicochemical properties as well as the microbiome composition of raw honey. Honeybee foraging behavior is affected by several factors, including bee morphology, geographic origin, season, temperature, time of the day, and floral phenology, which in turn affects the collected pollen and honey production (Alqarni, 2006; Joshi & Joshi, 2010). Honeybees with longer wingspan and body size have longer foraging distances and prefer certain types of flowers, influencing wildflower honey microbiota (Mostajeran et al., 2006; Oldroyd et al., 1992). Nectar and pollen collected from different flowers have different physicochemical properties (pH, acidity), chemical and microbial composition, contributing different properties to beehives and honey (Egorova, 1971; Lenaerts et al., 2016; Loper et al., 1980). Traditional culture methods to profile and analyze food microbiota are well- established and robust. But these methods are semi-quantitative, time-consuming, and labor-intensive. The need for identifying microorganisms in low concentration or in persister states, and the need for information on metabolites and genomic potentials of these microorganisms further validate the necessity of using genomic methods to 157 analyze food microbiome accurately and efficiently (Andjelković et al., 2017; Giacometti et al., 2013; Senoh et al., 2012). In our study, we chose to use high- throughput amplicon sequencing methods to characterize the microbial composition of raw honey. Additionally, we combined the genomic approach with traditional culture isolation methods to identify two bacterial strains from honey with antifungal potentials. High-throughput analysis of honey microbiomes can facilitate traditional culture-based screening methods to discover novel antimicrobials. Genome mining of microorganisms in the food of interest can reveal the vast repertoire of antimicrobials that are encoded in the genome, revealing those putative metabolites that may be overlooked by traditional culture-based methods, which are limited by physicochemical and environmental conditions. Future application of the broad range of antimicrobials in honey and other natural foods is unlimited. The next step of investigating the honey microbiome is using foodomic approaches. Shotgun metagenomic sequencing will provide vast amounts of data on all fragmented DNA in the sample without amplification. Shotgun sequences can be assembled and used for functional characterization, providing prediction of potential metabolites. Additionally, metatranscriptomic and metabolomic approaches will provide gene expression profiles and functional information, contributing to our understanding of microbial diversity in honey and improvement of food safety and quality. However, there are still some potential problems for these genomic, transcriptomic, proteomic, and metabolomic techniques. One issue is false-positive identification and lack of reproducibility, especially for detection of low-abundance 158 components (Gallo & Ferranti, 2016; Martinović et al., 2016). The complex food matrix may interfere with sequencing sample preparation, yielding low quality results and inaccurate interpretation (Andjelković et al., 2017). Metal ions, lipids, fat, and proteins are likely to inhibit DNA purification and PCR reactions (Bickley et al., 1996; Rossen et al., 1992). One solution is to use proper sample treatment to remove any components in the food matrix that may interact with target molecules, which should be validated during experimental design. Some other common issues encountered in microbiome data interpretation include limitations of the sequencing platform (limited read length, high error rate), and limitations of selected primers (primer biases) (Claesson et al., 2010). Better platforms with high-quality, longer reads can provide more coverage to elucidate complex, diverse microbial systems in food. Additionally, sampling and storage methods, DNA extraction and amplification methods, and sequence analysis pipelines are all variables that may compromise the reproducibility and comparability of the sequencing results, leading to dubious diversity analysis (Gihring et al., 2012; Salonen et al., 2010; Sinclair et al., 2015). For example, improper conditions of sample transportation and storage may alter the microbial composition, leading to unrepresentative amplified sequences and thwarting the correct interpretation of foodomic sequencing data. Similarly, biases can originate from the different sensitivities of microorganisms to cell lysis agents and DNA extraction methods, different levels of DNA amplification for selected primers, and reference databases selected for sequence analysis, all of which can contribute to unrepresentative microbial abundance and gene expression profiles (Engelbrektson et al., 2010). Moreover, proper internal controls should be included in foodomic studies. 159 For the positive controls, DNA sequences from known bacterial mock communities should be sequenced in parallel to provide an estimate of the sequencing errors for downstream analysis (Kozich et al., 2013). This is especially important for sequencing error evaluation of 16S rRNA amplicon studies. Negative controls should also be included to evaluate any trace amount of contamination that may be present in sequencing reagents (Salter et al., 2014). Contamination may be unavoidable, considering the sensitivity of the sequencing platform, but downstream analysis can be performed to remove background noise from contamination. Honey microbiome research, like other microbiome studies, should be driven by hypotheses and concepts, not methods (Brüssow, 2020). Due to the rapid development of highly sophisticated sequencing technologies and enormous amount of complex data, researchers may easily lose sight of research hypotheses and pursue studies only based on technical developments. For future food microbiome studies, appropriate experimental, methodological, and statistical design is necessary to perform high quality research and obtain spatial, temporal, and community dynamic information (Berg et al., 2020). Ideally, for a dynamic system like honey, the microbial interaction in a space-time continuum is more meaningful for the interpretation of microbiome function and evolutionary dynamics than a snapshot of a particular time and space. Future studies of honey microbiome should evaluate and compare honey samples over time during different stages, like honeybee collection, maturation in hive, human processing, storage and transportation. In terms of food applications of antimicrobial secondary metabolites from 160 natural sources, these natural compounds are highly desirable because of their stability and broad inhibition spectrum. However, considering the complexity of food matrices, these natural compounds may not function properly when used as food additives due to the presence of proteins, lipids, and other inhibitory components in food. Other challenges include the undesirable taste or smell, processing conditions that may reduce or inactivate the antimicrobial properties of these compounds. Using processing and packaging technologies to safely deliver these antimicrobials and preserve their antimicrobial properties at appropriate concentrations while not changing the sensory characteristics of food products is the essential issue we have to solve for industrial application of these food biopreservatives. 161 REFERENCES Abriouel, H., Franz, C. M. A. P., Omar, N. B., & Gálvez, A. (2011). Diversity and applications of Bacillus bacteriocins. FEMS Microbiology Reviews, 35(1), 201- 232. https://doi.org/10.1111/j.1574-6976.2010.00244.x Adcock, D. (1962). The Effect of Catalase on the Inhibine and Peroxide Values of Various Honeys. Journal of Apicultural Research, 1(1), 38-40. https://doi.org/10.1080/00218839.1962.11100047 Ahn, J.-Y., Min, J., Lee, S.-H., Jang, A., Park, C.-K., Kwon, S.-D., Park, S.-K., Lee, K., & Kim, Y.-H. (2014). Metagenomic analysis for identifying Kimchi sp. during the industrial-scale batch fermentation. Toxicology and Environmental Health Sciences, 6(1), 8-15. https://doi.org/10.1007/s13530-014-0182-0 al Somal, N., Coley, K. E., Molan, P. C., & Hancock, B. M. (1994). Susceptibility of Helicobacter pylori to the antibacterial activity of manuka honey. Journal of the Royal Society of Medicine, 87(1), 9-12. https://pubmed.ncbi.nlm.nih.gov/8308841 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1294271/ Al-Mamary, M., Al-Meeri, A., & Al-Habori, M. (2002). Antioxidant activities and total phenolics of different types of honey. Nutrition Research, 22(9), 1041- 1047. https://doi.org/https://doi.org/10.1016/S0271-5317(02)00406-2 Al-Waili, N. S., & Boni, N. S. (2003). Natural Honey Lowers Plasma Prostaglandin Concentrations in Normal Individuals. Journal of medicinal food, 6(2), 129- 133. https://doi.org/10.1089/109662003322233530 162 Aleti, G., Sessitsch, A., & Brader, G. (2015). Genome mining: Prediction of lipopeptides and polyketides from Bacillus and related Firmicutes. Computational and structural biotechnology journal, 13, 192-203. https://doi.org/10.1016/j.csbj.2015.03.003 Allsop, K. A., & Miller, J. B. (1996). Honey revisited: a reappraisal of honey in pre- industrial diets. British Journal of Nutrition, 75(4), 513-520. https://doi.org/10.1079/BJN19960155 Alqarni, A. (2006). Tolerance of Summer Temperature in Imported and Indigenous Honeybee Apis mellifera L. Races in Central Saudi Arabia. Saudi Journal of Biological Sciences, 13. Alvarez-Sieiro, P., Montalban-Lopez, M., Mu, D., & Kuipers, O. P. (2016). Bacteriocins of lactic acid bacteria: extending the family. Appl Microbiol Biotechnol, 100(7), 2939-2951. https://doi.org/10.1007/s00253-016-7343-9 Andjelković, U., Šrajer Gajdošik, M., Gašo-Sokač, D., Martinović, T., & Josić, D. (2017). Foodomics and Food Safety: Where We Are. Food Technol Biotechnol, 55(3), 290-307. https://doi.org/10.17113/ftb.55.03.17.5044 Anvarian, A. H. P., Cao, Y., Srikumar, S., Fanning, S., & Jordan, K. (2016). Flow Cytometric and 16S Sequencing Methodologies for Monitoring the Physiological Status of the Microbiome in Powdered Infant Formula Production [Methods]. Frontiers in microbiology, 7. https://doi.org/10.3389/fmicb.2016.00968 Azevedo, M. S., Valentim-Neto, P. A., Seraglio, S. K. T., da Luz, C. F. P., Arisi, A. C. M., & Costa, A. C. O. (2017). Proteome comparison for discrimination 16 3 between honeydew and floral honeys from botanical species Mimosa scabrella Bentham by principal component analysis [https://doi.org/10.1002/jsfa.8317]. Journal of the Science of Food and Agriculture, 97(13), 4515-4519. https://doi.org/https://doi.org/10.1002/jsfa.8317 Basualdo, C., Sgroy, V., Finola, M. S., & Marioli, J. M. (2007). Comparison of the antibacterial activity of honey from different provenance against bacteria usually isolated from skin wounds. Veterinary Microbiology, 124(3), 375-381. https://doi.org/https://doi.org/10.1016/j.vetmic.2007.04.039 Ben Sghaier, M., Skandrani, I., Nasr, N., Franca, M.-G. D., Chekir-Ghedira, L., & Ghedira, K. (2011). Flavonoids and sesquiterpenes from Tecurium ramosissimum promote antiproliferation of human cancer cells and enhance antioxidant activity: A structure–activity relationship study. Environmental Toxicology and Pharmacology, 32(3), 336-348. https://doi.org/https://doi.org/10.1016/j.etap.2011.07.003 Berg, G., Rybakova, D., Fischer, D., Cernava, T., Vergès, M.-C. C., Charles, T., Chen, X., Cocolin, L., Eversole, K., Corral, G. H., Kazou, M., Kinkel, L., Lange, L., Lima, N., Loy, A., Macklin, J. A., Maguin, E., Mauchline, T., McClure, R., . . . Schloter, M. (2020). Microbiome definition re-visited: old concepts and new challenges. Microbiome, 8(1), 103. https://doi.org/10.1186/s40168-020-00875- 0 Berríos, P., Fuentes, J. A., Salas, D., Carreño, A., Aldea, P., Fernández, F., & Trombert, A. N. (2018). Inhibitory effect of biofilm-forming Lactobacillus kunkeei strains against virulent Pseudomonas aeruginosa in vitro and in 16 4 honeycomb moth (Galleria mellonella) infection model. Benef Microbes, 9(2), 257-268. https://doi.org/10.3920/bm2017.0048 Bickley, J., Short, J. K., McDowell, D. G., & Parkes, H. C. (1996). Polymerase chain reaction (PCR) detection of Listeria monocytogenes in diluted milk and reversal of PCR inhibition caused by calcium ions [https://doi.org/10.1111/j.1472-765X.1996.tb01131.x]. Letters in Applied Microbiology, 22(2), 153-158. https://doi.org/https://doi.org/10.1111/j.1472- 765X.1996.tb01131.x Bilsel, Y., Bugra, D., Yamaner, S., Bulut, T., Cevikbas, U., & Turkoglu, U. (2002). Could Honey Have a Place in Colitis Therapy? Effects of Honey, Prednisolone, and Disulfiram on Inflammation, Nitric Oxide, and Free Radical Formation. Digestive Surgery, 19(4), 306-312. https://doi.org/10.1159/000064580 Boffo, E. F., Tavares, L. A., Tobias, A. C. T., Ferreira, M. M. C., & Ferreira, A. G. (2012). Identification of components of Brazilian honey by 1H NMR and classification of its botanical origin by chemometric methods. LWT, 49(1), 55- 63. https://doi.org/https://doi.org/10.1016/j.lwt.2012.04.024 Bogdanov, S., Jurendic, T., Sieber, R., & Gallmann, P. (2008). Honey for Nutrition and Health: A Review. Journal of the American College of Nutrition, 27(6), 677-689. https://doi.org/10.1080/07315724.2008.10719745 Bokulich, N. A., Kaehler, B. D., Rideout, J. R., Dillon, M., Bolyen, E., Knight, R., Huttley, G. A., & Gregory Caporaso, J. (2018). Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2's q2-feature- 16 5 classifier plugin. Microbiome, 6(1), 90. https://doi.org/10.1186/s40168-018- 0470-z Bordoni, A., & Capozzi, F. (2015). The foodomics approach for discovering biomarkers of food consumption in nutrition studies. Current Opinion in Food Science, 4, 124-128. https://doi.org/https://doi.org/10.1016/j.cofs.2015.07.005 Borutinskaite, V., Treigyte, G., Čeksteryte, V., Kurtinaitiene, B., & Navakauskienė, R. (2018). Proteomic identification and enzymatic activity of buckwheat (Fagopyrum esculentum) honey based on different assays. Journal of Food and Nutrition Research, 57. Bovo, S., Utzeri, V. J., Ribani, A., Cabbri, R., & Fontanesi, L. (2020). Shotgun sequencing of honey DNA can describe honey bee derived environmental signatures and the honey bee hologenome complexity. Sci Rep, 10(1), 9279. https://doi.org/10.1038/s41598-020-66127-1 Breukink, E., Wiedemann, I., Kraaij, C. v., Kuipers, O. P., Sahl, H. G., & de Kruijff, B. (1999). Use of the Cell Wall Precursor Lipid II by a Pore-Forming Peptide Antibiotic. Science, 286(5448), 2361-2364. https://doi.org/10.1126/science.286.5448.2361 Brown, K. L. (2000). Control of bacterial spores. British Medical Bulletin, 56(1), 158- 171. https://doi.org/10.1258/0007142001902860 Brudzynski, K. (2021). Honey as an Ecological Reservoir of Antibacterial Compounds Produced by Antagonistic Microbial Interactions in Plant Nectars, Honey and Honey Bee. Antibiotics (Basel), 10(5). https://doi.org/10.3390/antibiotics10050551 16 6 Brudzynski, K., & Sjaarda, C. P. (2021). Colloidal structure of honey and its influence on antibacterial activity [https://doi.org/10.1111/1541-4337.12720]. Comprehensive Reviews in Food Science and Food Safety, 20(2), 2063-2080. https://doi.org/https://doi.org/10.1111/1541-4337.12720 Brüssow, H. (2020). Problems with the concept of gut microbiota dysbiosis [https://doi.org/10.1111/1751-7915.13479]. Microbial Biotechnology, 13(2), 423-434. https://doi.org/https://doi.org/10.1111/1751-7915.13479 Cao, Y., Fanning, S., Proos, S., Jordan, K., & Srikumar, S. (2017). A Review on the Applications of Next Generation Sequencing Technologies as Applied to Food-Related Microbiome Studies [Review]. Frontiers in microbiology, 8. https://doi.org/10.3389/fmicb.2017.01829 Capozzi, F., & Bordoni, A. (2013). Foodomics: a new comprehensive approach to food and nutrition. Genes & Nutrition, 8(1), 1-4. https://doi.org/10.1007/s12263-012-0310-x Caulier, S., Nannan, C., Gillis, A., Licciardi, F., Bragard, C., & Mahillon, J. (2019). Overview of the Antimicrobial Compounds Produced by Members of the Bacillus subtilis Group. Front Microbiol, 10, 302. https://doi.org/10.3389/fmicb.2019.00302 Cavanagh, D., Beazley, J., & Ostapowicz, F. (1970). Radical operation for carcinoma of the vulva. Journal of Obstetrics and Gynaecology of the British Commonwealth, 77(11), 1037-1040. Cavia, M. M., Fernáez-Muiño, M. A., Huidobro, J. F., & Sancho, M. T. (2004). Correlation between Moisture and Water Activity of Honeys Harvested in 16 7 Different Years. Journal of Food Science, 69(5), C368-C370. https://doi.org/https://doi.org/10.1111/j.1365-2621.2004.tb10699.x Chaven, S. (2014). Chapter 11 - Honey, Confectionery and Bakery Products. In Y. Motarjemi & H. Lelieveld (Eds.), Food Safety Management (pp. 283-299). Academic Press. https://doi.org/https://doi.org/10.1016/B978-0-12-381504- 0.00011-1 Chua, L. S., Lee, J. Y., & Chan, G. F. (2013). Honey protein extraction and determination by mass spectrometry. Analytical and Bioanalytical Chemistry, 405(10), 3063-3074. https://doi.org/10.1007/s00216-012-6630-2 Claesson, M. J., Wang, Q., O'Sullivan, O., Greene-Diniz, R., Cole, J. R., Ross, R. P., & O'Toole, P. W. (2010). Comparison of two next-generation sequencing technologies for resolving highly complex microbiota composition using tandem variable 16S rRNA gene regions. Nucleic Acids Research, 38(22), e200-e200. https://doi.org/10.1093/nar/gkq873 Costa, P. A., Moraes, I. C. F., Bittante, A. M. Q., do Amaral Sobral, P. J., Gomide, C. A., & Carrer, C. C. (2013). Physical properties of honeys produced in the Northeast of Brazil. International Journal of Food Studies, 2(1). Crane, E. (1975). Honey: a comprehensive survey. D’Alessandro, A., & Zolla, L. (2012). We Are What We Eat: Food Safety and Proteomics. Journal of Proteome Research, 11(1), 26-36. https://doi.org/10.1021/pr2008829 de Almeida-Muradian, L. B., Stramm, K. M., & Estevinho, L. M. (2014). Efficiency of the FT-IR ATR spectrometry for the prediction of the physicochemical 16 8 characteristics of Melipona subnitida honey and study of the temperature's effect on those properties. International Journal of Food Science & Technology, 49(1), 188-195. https://doi.org/https://doi.org/10.1111/ijfs.12297 de Arauz, L. J., Jozala, A. F., Mazzola, P. G., & Vessoni Penna, T. C. (2009). Nisin biotechnological production and application: a review. Trends in Food Science & Technology, 20(3), 146-154. https://doi.org/https://doi.org/10.1016/j.tifs.2009.01.056 De Filippis, F., Parente, E., & Ercolini, D. (2017). Metagenomics insights into food fermentations [https://doi.org/10.1111/1751-7915.12421]. Microbial Biotechnology, 10(1), 91-102. https://doi.org/https://doi.org/10.1111/1751- 7915.12421 Delves-Broughton, J., Blackburn, P., Evans, R. J., & Hugenholtz, J. (1996). Applications of the bacteriocin, nisin. Antonie Van Leeuwenhoek, 69(2), 193- 202. https://doi.org/10.1007/bf00399424 Diep Dzung, B., Skaugen, M., Salehian, Z., Holo, H., & Nes Ingolf, F. (2007). Common mechanisms of target cell recognition and immunity for class II bacteriocins. Proceedings of the National Academy of Sciences, 104(7), 2384- 2389. https://doi.org/10.1073/pnas.0608775104 Dodd, H. M., Horn, N., & Gasson, M. J. (1990). Analysis of the Genetic Determinant for Production of the Peptide Antibiotic Nisin. Microbiology, 136(3), 555-556. https://doi.org/https://doi.org/10.1099/00221287-136-3-555 Dugat-Bony, E., Straub, C., Teissandier, A., Onésime, D., Loux, V., Monnet, C., Irlinger, F., Landaud, S., Leclercq-Perlat, M. N., Bento, P., Fraud, S., Gibrat, J. 16 9 F., Aubert, J., Fer, F., Guédon, E., Pons, N., Kennedy, S., Beckerich, J. M., Swennen, D., & Bonnarme, P. (2015). Overview of a surface-ripened cheese community functioning by meta-omics analyses. PLoS One, 10(4), e0124360. https://doi.org/10.1371/journal.pone.0124360 Efem, S. E. (1993). Recent advances in the management of Fournier's gangrene: preliminary observations. Surgery, 113(2), 200-204. Egorova, A. (1971). Preservative microflora in stored pollen. Veterinariya, 8, 40-41. Engelbrektson, A., Kunin, V., Wrighton, K. C., Zvenigorodsky, N., Chen, F., Ochman, H., & Hugenholtz, P. (2010). Experimental factors affecting PCR-based estimates of microbial species richness and evenness. The ISME Journal, 4(5), 642-647. https://doi.org/10.1038/ismej.2009.153 Ercolini, D. (2013). High-Throughput Sequencing and Metagenomics: Moving Forward in the Culture-Independent Analysis of Food Microbial Ecology. Applied and Environmental Microbiology, 79(10), 3148-3155. https://doi.org/doi:10.1128/AEM.00256-13 Evans, B. S., Chen, Y., Metcalf, W. W., Zhao, H., & Kelleher, N. L. (2011). Directed evolution of the nonribosomal peptide synthetase AdmK generates new andrimid derivatives in vivo. Chemistry & biology, 18(5), 601-607. https://doi.org/10.1016/j.chembiol.2011.03.008 Ferreres, F., Ortiz, A., Silva, C., Garcia-Viguera, C., Tomás-Barberán, F. A., & Tomás-Lorente, F. (1992). Flavonoids of “La Alcarria” honey A study of their botanical origin. Zeitschrift für Lebensmittel-Untersuchung und Forschung, 194(2), 139-143. https://doi.org/10.1007/BF01190185 17 0 Fisch, K. M. (2013). Biosynthesis of natural products by microbial iterative hybrid PKS–NRPS [10.1039/C3RA42661K]. RSC Advances, 3(40), 18228-18247. https://doi.org/10.1039/C3RA42661K Fontana Jr., A. J. (2007). Appendix D: Minimum Water Activity Limits for Growth of Microorganisms. In Water Activity in Foods (pp. 405-405). https://doi.org/https://doi.org/10.1002/9780470376454.app4 Gallo, M., & Ferranti, P. (2016). The evolution of analytical chemistry methods in foodomics. Journal of Chromatography A, 1428, 3-15. https://doi.org/https://doi.org/10.1016/j.chroma.2015.09.007 García-Cañas, V., Simó, C., Herrero, M., Ibáñez, E., & Cifuentes, A. (2012). Present and Future Challenges in Food Analysis: Foodomics. Analytical Chemistry, 84(23), 10150-10159. https://doi.org/10.1021/ac301680q Giacometti, J., Tomljanović, A. B., & Josić, D. (2013). Application of proteomics and metabolomics for investigation of food toxins. Food Research International, 54(1), 1042-1051. https://doi.org/https://doi.org/10.1016/j.foodres.2012.10.019 Gihring, T. M., Green, S. J., & Schadt, C. W. (2012). Massively parallel rRNA gene sequencing exacerbates the potential for biased community diversity comparisons due to variable library sizes [https://doi.org/10.1111/j.1462- 2920.2011.02550.x]. Environmental Microbiology, 14(2), 285-290. https://doi.org/https://doi.org/10.1111/j.1462-2920.2011.02550.x Grabowski, N. T., & Klein, G. (2017). Microbiology and foodborne pathogens in honey. Crit Rev Food Sci Nutr, 57(9), 1852-1862. https://doi.org/10.1080/10408398.2015.1029041 17 1 Hao, P., Zheng, H., Yu, Y., Ding, G., Gu, W., Chen, S., Yu, Z., Ren, S., Oda, M., Konno, T., Wang, S., Li, X., Ji, Z. S., & Zhao, G. (2011). Complete sequencing and pan-genomic analysis of Lactobacillus delbrueckii subsp. bulgaricus reveal its genetic basis for industrial yogurt production. PLoS One, 6(1), e15964. https://doi.org/10.1371/journal.pone.0015964 Havelaar, A. H., Kirk, M. D., Torgerson, P. R., Gibb, H. J., Hald, T., Lake, R. J., Praet, N., Bellinger, D. C., de Silva, N. R., Gargouri, N., Speybroeck, N., Cawthorne, A., Mathers, C., Stein, C., Angulo, F. J., Devleesschauwer, B., & World Health Organization Foodborne Disease Burden Epidemiology Reference, G. (2015). World Health Organization Global Estimates and Regional Comparisons of the Burden of Foodborne Disease in 2010. PLoS medicine, 12(12), e1001923-e1001923. https://doi.org/10.1371/journal.pmed.1001923 Herrero, M., Simó, C., García-Cañas, V., Ibáñez, E., & Cifuentes, A. (2012). Foodomics: MS-based strategies in modern food science and nutrition [https://doi.org/10.1002/mas.20335]. Mass Spectrometry Reviews, 31(1), 49- 69. https://doi.org/https://doi.org/10.1002/mas.20335 Iglesias, M. T., Martín-Álvarez, P. J., Polo, M. C., de Lorenzo, C., González, M., & Pueyo, E. (2006). Changes in the Free Amino Acid Contents of Honeys During Storage at Ambient Temperature. Journal of agricultural and food chemistry, 54(24), 9099-9104. https://doi.org/10.1021/jf061712x Joshi, N., & Joshi, P. (2010). Foraging Behaviour of Apis Spp. on Apple Flowers in a Subtropical Environment. New York Science Journal, 3. 17 2 Jung, J. Y., Lee, S. H., Jin, H. M., Hahn, Y., Madsen, E. L., & Jeon, C. O. (2013). Metatranscriptomic analysis of lactic acid bacterial gene expression during kimchi fermentation. International Journal of Food Microbiology, 163(2), 171-179. https://doi.org/https://doi.org/10.1016/j.ijfoodmicro.2013.02.022 Kačániová, M., Kňazovická, V., Felšöciová, S., & Rovná, K. (2012). Microscopic fungi recovered from honey and their toxinogenity. J Environ Sci Health A Tox Hazard Subst Environ Eng, 47(11), 1659-1664. https://doi.org/10.1080/10934529.2012.687242 Kafantaris, I., Amoutzias, G. D., & Mossialos, D. (2021). Foodomics in bee product research: a systematic literature review. European Food Research and Technology, 247(2), 309-331. https://doi.org/10.1007/s00217-020-03634-5 Khan, H., Flint, S. H., & Yu, P. L. (2013). Determination of the mode of action of enterolysin A, produced by Enterococcus faecalis B9510 [https://doi.org/10.1111/jam.12240]. Journal of Applied Microbiology, 115(2), 484-494. https://doi.org/https://doi.org/10.1111/jam.12240 Kiss, A., Balikó, G., Csorba, A., Chuluunbaatar, T., Medzihradszky, K. F., & Alföldi, L. (2008). Cloning and characterization of the DNA region responsible for Megacin A-216 production in Bacillus megaterium 216. Journal of Bacteriology, 190(19), 6448-6457. https://doi.org/10.1128/JB.00557-08 Kňazovická, V., Gábor, M., Miluchová, M., Bobko, M., & Medo, J. (2020). Diversity of bacteria in Slovak and foreign honey, with assessment of its physico- chemical quality and counts of cultivable microorganisms. Journal of Microbiology, Biotechnology and Food Sciences, 9(6), 414-421. 17 3 Ko, D. K., Nadakuduti, S. S., Douches, D. S., & Buell, C. R. (2018). Transcriptome profiling of transgenic potato plants provides insights into variability caused by plant transformation. PLoS One, 13(11), e0206055. https://doi.org/10.1371/journal.pone.0206055 Kozich, J. J., Westcott, S. L., Baxter, N. T., Highlander, S. K., & Schloss, P. D. (2013). Development of a Dual-Index Sequencing Strategy and Curation Pipeline for Analyzing Amplicon Sequence Data on the MiSeq Illumina Sequencing Platform. Applied and Environmental Microbiology, 79(17), 5112- 5120. https://doi.org/doi:10.1128/AEM.01043-13 Lamas, A., Regal, P., Vázquez, B., Miranda, J. M., Franco, C. M., & Cepeda, A. (2019). Transcriptomics: A powerful tool to evaluate the behavior of foodborne pathogens in the food production chain. Food Res Int, 125, 108543. https://doi.org/10.1016/j.foodres.2019.108543 Lancova, K., Dip, R., Antignac, J.-P., Bizec, B. L., Elliott, C. T., & Naegeli, H. (2011). Detection of hazardous food contaminants by transcriptomics fingerprinting. TrAC Trends in Analytical Chemistry, 30(2), 181-191. https://doi.org/https://doi.org/10.1016/j.trac.2010.10.013 Lazaridou, A., Biliaderis, C. G., Bacandritsos, N., & Sabatini, A. G. (2004). Composition, thermal and rheological behaviour of selected Greek honeys. Journal of Food Engineering, 64(1), 9-21. https://doi.org/https://doi.org/10.1016/j.jfoodeng.2003.09.007 17 4 Lee, H., Churey, J. J., & Worobo, R. W. (2008a). Antimicrobial activity of bacterial isolates from different floral sources of honey. Int J Food Microbiol, 126(1-2), 240-244. https://doi.org/10.1016/j.ijfoodmicro.2008.04.030 Lee, H., Churey, J. J., & Worobo, R. W. (2008b). Purification and structural characterization of bacillomycin F produced by a bacterial honey isolate active against Byssochlamys fulva H25. J Appl Microbiol, 105(3), 663-673. https://doi.org/10.1111/j.1365-2672.2008.03797.x Leistner, L., & Gorris, L. G. M. (1995). Food preservation by hurdle technology. Trends in Food Science & Technology, 6(2), 41-46. https://doi.org/https://doi.org/10.1016/S0924-2244(00)88941-4 Lenaerts, M., Pozo, M. I., Wäckers, F., Van den Ende, W., Jacquemyn, H., & Lievens, B. (2016). Impact of microbial communities on floral nectar chemistry: Potential implications for biological control of pest insects. Basic and Applied Ecology, 17(3), 189-198. https://doi.org/https://doi.org/10.1016/j.baae.2015.10.001 Lessard, M.-H., Viel, C., Boyle, B., St-Gelais, D., & Labrie, S. (2014). Metatranscriptome analysis of fungal strains Penicillium camemberti and Geotrichum candidumreveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese. BMC genomics, 15(1), 235. https://doi.org/10.1186/1471-2164-15-235 Loper, G., Standifer, L., Thompson, M., & Gilliam, M. (1980). Biochemistry and microbiology of bee-collected almond (Prunus dulcis) pollen and bee bread. I- Fatty Acids, Sterols, Vitamins and Minerals. Apidologie, 11(1), 63-73. 17 5 Lorenzo, J. M., Munekata, P. E., Dominguez, R., Pateiro, M., Saraiva, J. A., & Franco, D. (2018). Main Groups of Microorganisms of Relevance for Food Safety and Stability: General Aspects and Overall Description. Innovative Technologies for Food Preservation, 53-107. https://doi.org/10.1016/B978-0-12-811031- 7.00003-0 Machado De-Melo, A. A., Almeida-Muradian, L. B. d., Sancho, M. T., & Pascual- Maté, A. (2018). Composition and properties of Apis mellifera honey: A review. Journal of Apicultural Research, 57(1), 5-37. https://doi.org/10.1080/00218839.2017.1338444 Maddocks, S. E., & Jenkins, R. E. (2013). Honey: a sweet solution to the growing problem of antimicrobial resistance? Future Microbiology, 8(11), 1419-1429. https://doi.org/10.2217/fmb.13.105 Madejczyk, M., & Baralkiewicz, D. (2008). Characterization of Polish rape and honeydew honey according to their mineral contents using ICP-MS and F- AAS/AES. Analytica Chimica Acta, 617(1), 11-17. https://doi.org/https://doi.org/10.1016/j.aca.2008.01.038 Martinović, T., Andjelković, U., Gajdošik, M., Rešetar, D., & Josić, D. (2016). Foodborne pathogens and their toxins. J Proteomics, 147, 226-235. https://doi.org/10.1016/j.jprot.2016.04.029 Mato, I., Huidobro, J. F., Simal-Lozano, J., & Sancho, M. T. (2003). Significance of nonaromatic organic acids in honey. J Food Prot, 66(12), 2371-2376. https://doi.org/10.4315/0362-028x-66.12.2371 17 6 Mato, I., Huidobro, J. F., Simal-Lozano, J., & Sancho, M. T. (2006). Rapid Determination of Nonaromatic Organic Acids in Honey by Capillary Zone Electrophoresis with Direct Ultraviolet Detection. Journal of agricultural and food chemistry, 54(5), 1541-1550. https://doi.org/10.1021/jf051757i Molan, P. C. (1992). The antibacterial activity of honey: 1. The nature of the antibacterial activity. Bee World, 73(1), 5-28. Molan, P. C. (1999). Why honey is effective as a medicine. 1. Its use in modern medicine. Bee World, 80(2), 80-92. https://doi.org/10.1080/0005772X.1999.11099430 Molan, P. C., Smith, I. M., & Reid, G. M. (1988). A Comparison of the Antibacterial Activities of Some new Zealand Honeys. Journal of Apicultural Research, 27(4), 252-256. https://doi.org/10.1080/00218839.1988.11100811 Moll, G. N., Konings, W. N., & Driessen, A. J. (1999). Bacteriocins: mechanism of membrane insertion and pore formation. Antonie Van Leeuwenhoek, 76(1-4), 185-198. Monnet, C., Dugat-Bony, E., Swennen, D., Beckerich, J. M., Irlinger, F., Fraud, S., & Bonnarme, P. (2016). Investigation of the Activity of the Microorganisms in a Reblochon-Style Cheese by Metatranscriptomic Analysis. Front Microbiol, 7, 536. https://doi.org/10.3389/fmicb.2016.00536 Mootz, H. D., Schwarzer, D., & Marahiel, M. A. (2002). Ways of Assembling Complex Natural Products on Modular Nonribosomal Peptide Synthetases [https://doi.org/10.1002/1439-7633(20020603)3:6<490::AID- CBIC490>3.0.CO;2-N]. ChemBioChem, 3(6), 490-504. 17 7 https://doi.org/https://doi.org/10.1002/1439-7633(20020603)3:6<490::AID- CBIC490>3.0.CO;2-N Mostajeran, M., Edriss, M. A., & Basiri, M. R. (2006). Analysis of colony and morphological characters in honey bees (Apis mellifera meda). Pak. J. Biol. Sci, 9(14), 2685-2688. Mundo, M. A., Padilla-Zakour, O. I., & Worobo, R. W. (2004). Growth inhibition of foodborne pathogens and food spoilage organisms by select raw honeys. Int J Food Microbiol, 97(1), 1-8. https://doi.org/10.1016/j.ijfoodmicro.2004.03.025 Neefs, J.-M., Van de Peer, Y., De Rijk, P., Chapelle, S., & De Wachter, R. (1993). Compilation of small ribosomal subunit RNA structures. Nucleic Acids Research, 21(13), 3025-3049. https://doi.org/10.1093/nar/21.13.3025 Nychas, G. J. E., & Panagou, E. (2011). 1 - Microbiological spoilage of foods and beverages. In D. Kilcast & P. Subramaniam (Eds.), Food and Beverage Stability and Shelf Life (pp. 3-28). Woodhead Publishing. https://doi.org/https://doi.org/10.1533/9780857092540.1.3 Obaseiki-Ebor, E. E., & Afonya, T. C. (1984). In-vitro evaluation of the anticandidiasis activity of honey distillate (HY-1) compared with that of some antimycotic agents. J Pharm Pharmacol, 36(4), 283-284. https://doi.org/10.1111/j.2042-7158.1984.tb04373.x Obi, C. L., Ugoji, E. O., Edun, S. A., Lawal, S. F., & Anyiwo, C. E. (1994). The antibacterial effect of honey on diarrhoea causing bacterial agents isolated in Lagos, Nigeria. Afr J Med Med Sci, 23(3), 257-260. 17 8 Ojeda de Rodrı́guez, G., Sulbarán de Ferrer, B., Ferrer, A., & Rodrı́guez, B. (2004). Characterization of honey produced in Venezuela. Food Chemistry, 84(4), 499-502. https://doi.org/https://doi.org/10.1016/S0308-8146(02)00517-4 Olaitan, P. B., Adeleke, O. E., & Ola, I. O. (2007). Honey: a reservoir for microorganisms and an inhibitory agent for microbes. Afr Health Sci, 7(3), 159-165. https://doi.org/10.5555/afhs.2007.7.3.159 Olas, B. (2020). Honey and Its Phenolic Compounds as an Effective Natural Medicine for Cardiovascular Diseases in Humans? Nutrients, 12(2), 283. https://doi.org/10.3390/nu12020283 Oldroyd, B., Rinderer, T., & Wongsiri, S. (1992). Pollen resource partitioning by Apis dorsata, A. cerana, A. andreniformis and A. florea in Thailand. Journal of Apicultural Research, 31(1), 3-7. https://doi.org/10.1080/00218839.1992.11101253 Olofsson, T. C., & Vásquez, A. (2008). Detection and identification of a novel lactic acid bacterial flora within the honey stomach of the honeybee Apis mellifera. Curr Microbiol, 57(4), 356-363. https://doi.org/10.1007/s00284-008-9202-0 Ouchemoukh, S., Schweitzer, P., Bachir Bey, M., Djoudad-Kadji, H., & Louaileche, H. (2010). HPLC sugar profiles of Algerian honeys. Food Chemistry, 121(2), 561-568. https://doi.org/https://doi.org/10.1016/j.foodchem.2009.12.047 Paramás, A. M. G., Bárez, J. A. G., Marcos, C. C., García-Villanova, R. J., & Sánchez, J. S. (2006). HPLC-fluorimetric method for analysis of amino acids in products of the hive (honey and bee-pollen). Food Chemistry, 95(1), 148- 156. https://doi.org/https://doi.org/10.1016/j.foodchem.2005.02.008 17 9 Petruzzi, L., Corbo, M. R., Sinigaglia, M., & Bevilacqua, A. (2017). Chapter 1 - Microbial Spoilage of Foods: Fundamentals. In A. Bevilacqua, M. R. Corbo, & M. Sinigaglia (Eds.), The Microbiological Quality of Food (pp. 1-21). Woodhead Publishing. https://doi.org/https://doi.org/10.1016/B978-0-08- 100502-6.00002-9 Połka, J., Rebecchi, A., Pisacane, V., Morelli, L., & Puglisi, E. (2015). Bacterial diversity in typical Italian salami at different ripening stages as revealed by high-throughput sequencing of 16S rRNA amplicons. Food Microbiology, 46, 342-356. https://doi.org/https://doi.org/10.1016/j.fm.2014.08.023 Postmes, T. (2001). The treatment of burns and other wounds with honey. Honey and healing, ed. P. Munn and R Jones. International Bee Research Association, Cardiff, UK. Pucci, M. J., Vedamuthu, E. R., Kunka, B. S., & Vandenbergh, P. A. (1988). Inhibition of Listeria monocytogenes by using bacteriocin PA-1 produced by Pediococcus acidilactici PAC 1.0. Appl Environ Microbiol, 54(10), 2349-2353. https://doi.org/10.1128/aem.54.10.2349-2353.1988 Quigley, L., O'Sullivan, O., Beresford, T. P., Ross, R. P., Fitzgerald, G. F., & Cotter, P. D. (2012). High-throughput sequencing for detection of subpopulations of bacteria not previously associated with artisanal cheeses. Appl Environ Microbiol, 78(16), 5717-5723. https://doi.org/10.1128/aem.00918-12 Raaijmakers, J. M., De Bruijn, I., Nybroe, O., & Ongena, M. (2010). Natural functions of lipopeptides from Bacillus and Pseudomonas: more than surfactants and 18 0 antibiotics. FEMS Microbiol Rev, 34(6), 1037-1062. https://doi.org/10.1111/j.1574-6976.2010.00221.x Radwan, S. S., El-Essawy, A. A., & Sarhan, M. M. (1984). Experimental evidence for the occurrence in honey of specific substances active against microorganisms. Zentralbl Mikrobiol, 139(4), 249-255. Ramos, O. Y., Basualdo, M., Libonatti, C., & Vega, M. F. (2020). Current status and application of lactic acid bacteria in animal production systems with a focus on bacteria from honey bee colonies [https://doi.org/10.1111/jam.14469]. Journal of Applied Microbiology, 128(5), 1248-1260. https://doi.org/https://doi.org/10.1111/jam.14469 Razali, M. T. A., Zainal, Z. A., Maulidiani, M., Shaari, K., Zamri, Z., Mohd Idrus, M. Z., Khatib, A., Abas, F., Ling, Y. S., Rui, L. L., & Ismail, I. S. (2018). Classification of Raw Stingless Bee Honeys by Bee Species Origins Using the NMR- and LC-MS-Based Metabolomics Approach. Molecules, 23(9), 2160. https://www.mdpi.com/1420-3049/23/9/2160 Rešetar, D., Pavelić, S. K., & Josić, D. (2015). Foodomics for investigations of food toxins. Current Opinion in Food Science, 4, 86-91. https://doi.org/https://doi.org/10.1016/j.cofs.2015.05.004 Rezzi, S., Ramadan, Z., Fay, L. B., & Kochhar, S. (2007). Nutritional Metabonomics:  Applications and Perspectives. Journal of Proteome Research, 6(2), 513-525. https://doi.org/10.1021/pr060522z 18 1 Ridley, C. P., Lee, H. Y., & Khosla, C. (2008). Evolution of polyketide synthases in bacteria. Proceedings of the National Academy of Sciences of the United States of America, 105(12), 4595-4600. https://doi.org/10.1073/pnas.0710107105 Rodrigues, M. X., Lima, S. F., Canniatti-Brazaca, S. G., & Bicalho, R. C. (2017). The microbiome of bulk tank milk: Characterization and associations with somatic cell count and bacterial count. Journal of Dairy Science, 100(4), 2536-2552. https://doi.org/10.3168/jds.2016-11540 Rodríguez-Andrade, E., Stchigel, A. M., Terrab, A., Guarro, J., & Cano-Lira, J. F. (2019). Diversity of xerotolerant and xerophilic fungi in honey. IMA fungus, 10(1), 20. https://doi.org/10.1186/s43008-019-0021-7 Rossano, R., Larocca, M., Polito, T., Perna, A. M., Padula, M. C., Martelli, G., & Riccio, P. (2012). What Are the Proteolytic Enzymes of Honey and What They Do Tell Us? A Fingerprint Analysis by 2-D Zymography of Unifloral Honeys. PLoS One, 7(11), e49164. https://doi.org/10.1371/journal.pone.0049164 Rossen, L., Nørskov, P., Holmstrøm, K., & Rasmussen, O. F. (1992). Inhibition of PCR by components of food samples, microbial diagnostic assays and DNA- extraction solutions. International Journal of Food Microbiology, 17(1), 37- 45. https://doi.org/https://doi.org/10.1016/0168-1605(92)90017-W Roy, N. S., Kim, J.-A., Choi, A.-Y., Ban, Y.-W., Park, N.-I., Park, K.-C., Yang, H.-S., Choi, I.-Y., & Kim, S. (2018). RNA-Seq De Novo Assembly and Differential Transcriptome Analysis of Korean Medicinal Herb Cirsium japonicum var. spinossimum. Genomics & informatics, 16(4), e34-e34. https://doi.org/10.5808/GI.2018.16.4.e34 18 2 Ruiz-Matute, A. I., Brokl, M., Soria, A. C., Sanz, M. L., & Martínez-Castro, I. (2010). Gas chromatographic–mass spectrometric characterisation of tri- and tetrasaccharides in honey. Food Chemistry, 120(2), 637-642. https://doi.org/https://doi.org/10.1016/j.foodchem.2009.10.050 Salonen, A., Nikkilä, J., Jalanka-Tuovinen, J., Immonen, O., Rajilić-Stojanović, M., Kekkonen, R. A., Palva, A., & de Vos, W. M. (2010). Comparative analysis of fecal DNA extraction methods with phylogenetic microarray: Effective recovery of bacterial and archaeal DNA using mechanical cell lysis. Journal of Microbiological Methods, 81(2), 127-134. https://doi.org/https://doi.org/10.1016/j.mimet.2010.02.007 Salter, S. J., Cox, M. J., Turek, E. M., Calus, S. T., Cookson, W. O., Moffatt, M. F., Turner, P., Parkhill, J., Loman, N. J., & Walker, A. W. (2014). Reagent and laboratory contamination can critically impact sequence-based microbiome analyses. BMC biology, 12(1), 87. Samarghandian, S., Farkhondeh, T., & Samini, F. (2017). Honey and Health: A Review of Recent Clinical Research. Pharmacognosy research, 9(2), 121-127. https://doi.org/10.4103/0974-8490.204647 Sanz, S., Gradillas, G., Jimeno, F., Perez, C., & Juan, T. (1995). Fermentation Problem in Spanish North-Coast Honey. J Food Prot, 58(5), 515-518. https://doi.org/10.4315/0362-028x-58.5.515 Schievano, E., Peggion, E., & Mammi, S. (2010). 1H Nuclear Magnetic Resonance Spectra of Chloroform Extracts of Honey for Chemometric Determination of 18 3 Its Botanical Origin. Journal of agricultural and food chemistry, 58(1), 57-65. https://doi.org/10.1021/jf9022977 Schievano, E., Stocchero, M., Morelato, E., Facchin, C., & Mammi, S. (2012). An NMR-based metabolomic approach to identify the botanical origin of honey. Metabolomics, 8(4), 679-690. https://doi.org/10.1007/s11306-011-0362-8 Schlundt, J., Tay, M. Y. F., Chengcheng, H., & Liwei, C. (2020). Food Security: Microbiological and Chemical Risks. In A. J. Masys, R. Izurieta, & M. Reina Ortiz (Eds.), Global Health Security: Recognizing Vulnerabilities, Creating Opportunities (pp. 231-274). Springer International Publishing. https://doi.org/10.1007/978-3-030-23491-1_11 Scholz, M., Ward, D. V., Pasolli, E., Tolio, T., Zolfo, M., Asnicar, F., Truong, D. T., Tett, A., Morrow, A. L., & Segata, N. (2016). Strain-level microbial epidemiology and population genomics from shotgun metagenomics. Nat Methods, 13(5), 435-438. https://doi.org/10.1038/nmeth.3802 Senoh, M., Ghosh-Banerjee, J., Ramamurthy, T., Colwell, R. R., Miyoshi, S.-i., Nair, G. B., & Takeda, Y. (2012). Conversion of viable but nonculturable enteric bacteria
 to culturable by co-culture with eukaryotic cells [https://doi.org/10.1111/j.1348-0421.2012.00440.x]. Microbiology and Immunology, 56(5), 342-345. https://doi.org/https://doi.org/10.1111/j.1348- 0421.2012.00440.x Sharma, S. (2015). Food preservatives and their harmful effects. International journal of scientific and research publications, 5(4), 1-2. 18 4 Siezen, R. J., Starrenburg, M. J., Boekhorst, J., Renckens, B., Molenaar, D., & van Hylckama Vlieg, J. E. (2008). Genome-scale genotype-phenotype matching of two Lactococcus lactis isolates from plants identifies mechanisms of adaptation to the plant niche. Appl Environ Microbiol, 74(2), 424-436. https://doi.org/10.1128/aem.01850-07 Silva, C. C. G., Silva, S. P. M., & Ribeiro, S. C. (2018). Application of Bacteriocins and Protective Cultures in Dairy Food Preservation [Review]. Frontiers in microbiology, 9. https://doi.org/10.3389/fmicb.2018.00594 Simmonds, R. S., Pearson, L., Kennedy, R. C., & Tagg, J. R. (1996). Mode of action of a lysostaphin-like bacteriolytic agent produced by Streptococcus zooepidemicus 4881. Applied and Environmental Microbiology, 62(12), 4536- 4541. https://doi.org/doi:10.1128/aem.62.12.4536-4541.1996 Sinclair, L., Osman, O. A., Bertilsson, S., & Eiler, A. (2015). Microbial Community Composition and Diversity via 16S rRNA Gene Amplicons: Evaluating the Illumina Platform. PLoS One, 10(2), e0116955. https://doi.org/10.1371/journal.pone.0116955 Singh, B. P., Rateb, M. E., Rodriguez-Couto, S., Polizeli, M. d. L. T. d. M., & Li, W.- J. (2019). Editorial: Microbial Secondary Metabolites: Recent Developments and Technological Challenges [Editorial]. Frontiers in microbiology, 10. https://doi.org/10.3389/fmicb.2019.00914 Stevens, K. A., Sheldon, B. W., Klapes, N. A., & Klaenhammer, T. R. (1991). Nisin treatment for inactivation of Salmonella species and other gram-negative 18 5 bacteria. Applied and Environmental Microbiology, 57(12), 3613-3615. https://doi.org/doi:10.1128/aem.57.12.3613-3615.1991 Subrahmanyam, M. (2005). Topical application of honey in treatment of burns. British Journal of Surgery, 78(4), 497-498. https://doi.org/10.1002/bjs.1800780435 Telias, A., White, J. R., Pahl, D. M., Ottesen, A. R., & Walsh, C. S. (2011). Bacterial community diversity and variation in spray water sources and the tomato fruit surface. BMC microbiology, 11(1), 81. https://doi.org/10.1186/1471-2180-11- 81 Terrab, A., Dı́ez, M. J., & Heredia, F. J. (2002). Characterisation of Moroccan unifloral honeys by their physicochemical characteristics. Food Chemistry, 79(3), 373-379. https://doi.org/https://doi.org/10.1016/S0308-8146(02)00189-9 Thakali, A., & MacRae, J. D. (2021). A review of chemical and microbial contamination in food: What are the threats to a circular food system? Environmental Research, 194, 110635. https://doi.org/https://doi.org/10.1016/j.envres.2020.110635 Trasande, L., Shaffer, R. M., Sathyanarayana, S., & Council On Environmental, H. (2018). Food Additives and Child Health. Pediatrics, 142(2), e20181410. https://doi.org/10.1542/peds.2018-1410 Truchado, P., López-Gálvez, F., Gil, M. I., Tomás-Barberán, F. A., & Allende, A. (2009). Quorum sensing inhibitory and antimicrobial activities of honeys and the relationship with individual phenolics. Food Chemistry, 115(4), 1337- 1344. https://doi.org/https://doi.org/10.1016/j.foodchem.2009.01.065 18 6 Ustunol, Z., & Gandhi, H. (2001). Growth and viability of commercial Bifidobacterium spp. in honey-sweetened skim milk. J Food Prot, 64(11), 1775-1779. https://doi.org/10.4315/0362-028x-64.11.1775 Val, A., Huidobro, J. F., Sánchez, M. P., Muniategui, S., Fernández-Muiño, M. A., & Sancho, M. T. (1998). Enzymatic Determination of Galactose and Lactose in Honey. Journal of agricultural and food chemistry, 46(4), 1381-1385. https://doi.org/10.1021/jf970483z van Dyk, B. N., de Bruin, W., du Plessis, E. M., & Korsten, L. (2016). Microbiological Food Safety Status of Commercially Produced Tomatoes from Production to Marketing. Journal of Food Protection, 79(3), 392-406. https://doi.org/10.4315/0362-028X.JFP-15-300 van Heusden, H. E., de Kruijff, B., & Breukink, E. (2002). Lipid II Induces a Transmembrane Orientation of the Pore-Forming Peptide Lantibiotic Nisin. Biochemistry, 41(40), 12171-12178. https://doi.org/10.1021/bi026090x Van Lanen, S. G., & Shen, B. (2006). Progress in combinatorial biosynthesis for drug discovery. Drug Discovery Today: Technologies, 3(3), 285-292. https://doi.org/https://doi.org/10.1016/j.ddtec.2006.09.014 Villalobos-Delgado, L. H., Nevárez-Moorillon, G. V., Caro, I., Quinto, E. J., & Mateo, J. (2019). 4 - Natural antimicrobial agents to improve foods shelf life. In C. M. Galanakis (Ed.), Food Quality and Shelf Life (pp. 125-157). Academic Press. https://doi.org/https://doi.org/10.1016/B978-0-12-817190-5.00004-5 Wang, H., Fewer, D. P., Holm, L., Rouhiainen, L., & Sivonen, K. (2014). Atlas of nonribosomal peptide and polyketide biosynthetic pathways reveals common 18 7 occurrence of nonmodular enzymes. Proceedings of the National Academy of Sciences, 111(25), 9259-9264. https://doi.org/doi:10.1073/pnas.1401734111 Wen, Y., Wang, L., Jin, Y., Zhang, J., Su, L., Zhang, X., Zhou, J., & Li, Y. (2017). The Microbial Community Dynamics during the Vitex Honey Ripening Process in the Honeycomb [Original Research]. Frontiers in microbiology, 8. https://doi.org/10.3389/fmicb.2017.01649 White, J. W., Subers, M. H., & Schepartz, A. I. (1963). The identification of inhibine, the antibacterial factor in honey, as hydrogen peroxide and its origin in a honey glucose-oxidase system. Biochimica et Biophysica Acta (BBA) - Specialized Section on Enzymological Subjects, 73(1), 57-70. https://doi.org/https://doi.org/10.1016/0926-6569(63)90108-1 Wiedemann, I., Breukink, E., van Kraaij, C., Kuipers, O. P., Bierbaum, G., de Kruijff, B., & Sahl, H.-G. (2001). Specific Binding of Nisin to the Peptidoglycan Precursor Lipid II Combines Pore Formation and Inhibition of Cell Wall Biosynthesis for Potent Antibiotic Activity *. Journal of Biological Chemistry, 276(3), 1772-1779. https://doi.org/10.1074/jbc.M006770200 Williams, T. R., Moyne, A.-L., Harris, L. J., & Marco, M. L. (2013). Season, Irrigation, Leaf Age, and Escherichia coli Inoculation Influence the Bacterial Diversity in the Lettuce Phyllosphere. PLoS One, 8(7), e68642. https://doi.org/10.1371/journal.pone.0068642 Won, S.-R., Li, C.-Y., Kim, J.-W., & Rhee, H.-I. (2009). Immunological characterization of honey major protein and its application. Food Chemistry, 18 8 113(4), 1334-1338. https://doi.org/https://doi.org/10.1016/j.foodchem.2008.08.082 Yun, J. W. (1996). Fructooligosaccharides—Occurrence, preparation, and application. Enzyme and Microbial Technology, 19(2), 107-117. https://doi.org/https://doi.org/10.1016/0141-0229(95)00188-3 Zhang, Y.-Z., Chen, Y.-F., Wu, Y.-Q., Si, J.-J., Zhang, C.-P., Zheng, H.-Q., & Hu, F.- L. (2019). Discrimination of the entomological origin of honey according to the secretions of the bee (Apis cerana or Apis mellifera). Food Research International, 116, 362-369. https://doi.org/https://doi.org/10.1016/j.foodres.2018.08.049 Zhong, Y., Wu, L., Chen, X., Huang, Z., & Hu, W. (2018). Effects of Food-Additive- Information on Consumers' Willingness to Accept Food with Additives. International journal of environmental research and public health, 15(11), 2394. https://doi.org/10.3390/ijerph15112394 18 9 APPENDICES Supplemental Figure 2.1. Rarefaction curves of bacterial and fungal ASV diversity for each honey sample based on the number of obtained reads in the sequencing libraries and identified species. 190 Supplemental Figure 3.1. LC-MS spectrum for singly charged m/z 1043.5 and doubly charged 522.3 of C14 iturin A. Spectrum was extracted from LC-MS for purified antifungal compounds produced by Bacillus velezensis WRB-ZX-001. 19 1 Supplemental Table 3.1. List of publicly available Bacillus spp. genome assembly included in this study. 19 2 Isolate Assembly Accession Number B_amyloliquefaciens_ATCC_23350_strain_DSM7 GCA_000196735.1 B_amyloliquefaciens_strain_CC178 GCA_000494835.1 B_amyloliquefaciens_strain_DC-12 GCA_000330805.1 B_amyloliquefaciens_strain_EBL11 GCA_000559145.1 B_amyloliquefaciens_strain_EGD-AQ14 GCA_000465655.1 B_amyloliquefaciens_strain_HB-26 GCA_000784675.1 B_amyloliquefaciens_strain_IT-45 GCA_000242855.2 B_amyloliquefaciens_strain_KHG19 GCA_000835145.1 B_amyloliquefaciens_strain_LFB112 GCA_000508265.1 B_amyloliquefaciens_strain_LL3 GCA_000204275.1 B_amyloliquefaciens_strain_Lx-11 GCA_001077735.1 B_amyloliquefaciens_strain_TA208 GCA_000195515.1 B_amyloliquefaciens_strain_UASWS_BA1 GCA_000469015.2 B_amyloliquefaciens_strain_UMAF6614 GCA_001593785.1 B_amyloliquefaciens_strain_UMAF6639 GCA_001593765.1 B_amyloliquefaciens_strain_XH7 GCA_000221645.1 B_amyloliquefaciens_strain_Y2 GCA_000262385.1 B_siamensis_strain_7551 GCA_002271775.1 B_siamensis_strain_JJC33M GCA_000798615.1 B_siamensis_strain_KCTC_13613 GCA_000262045.1 B_siamensis_strain_SCSIO_05746 GCA_002850535.1 B_siamensis_strain_SRCM100169 GCA_001662915.1 B_siamensis_strain_XY18 GCA_000966575.1 B_subtilis_strain_168 GCA_000009045.1 B_velezensis_strain_AS43_3 GCA_000319475.1 B_velezensis_strain_CAU_B946 GCA_000283695.1 B_velezensis_strain_CBMB205 GCA_002117165.1 B_velezensis_strain_FZB42 GCA_000015785.2 B_velezensis_strain_KACC13105 GCA_000960265.2 B_velezensis_strain_KACC18228 GCA_001461835.1 B_velezensis_strain_KCTC13012 GCA_001267695.1 B_velezensis_strain_M27 GCA_000299615.1 B_velezensis_strain_NAU-B3 GCA_000493375.1 B_velezensis_strain_NJN-6 GCA_000973585.1 B_velezensis_strain_NRRL_B-41580 GCA_001461825.1 B_velezensis_strain_SK19_001 GCA_000513755.1 B_velezensis_strain_SQR9 GCA_000685725.1 B_velezensis_strain_TrigoCor1448 GCA_000583065.1 B_velezensis_strain_UCMB5033 GCA_000455565.1 B_velezensis_strain_UCMB5036 GCA_000341875.1 B_velezensis_strain_UCMB5113 GCA_000455585.1 193 B_velezensis_strain_YAU_B9601-Y2 GCA_000284395.1 19 4 Supplemental Table 3.2. Gene Ontology (GO) annotation and node score distribution for Bacillus velezensis WRB-ZX-001 and WRB-ZX-002. GO term Bacillus velezensis Bacillus velezensis WRB-ZX-001 WRB-ZX-002 Biological Process (BP) Number of 1953 1955 Sequences Transport 11% 11% Carbohydrate derivative metabolic process 10% 10% Regulation of gene expression 9% 9% Transcription, DNA-templated 9% 9% Carbohydrate metabolic process 8% 8% Alpha-amino acid metabolic process 8% 8% Organonitrogen compound biosynthetic 7% 7% process Nucleobase-containing small molecule 5% 5% metabolic process Phosphate-containing compound metabolic 5% 5% process Organophosphate metabolic process 4% 4% Monocarboxylic acid metabolic process 4% 4% Regulation of cellular macromolecule 4% 4% biosynthetic process Cellular Component (CC) Number of 1870 1870 Sequences Integral component of membrane 55% 55% Cytoplasm 26% 26% Plasma membrane 16% 16% Catalytic complex 3% 3% Molecular Function (MF) Number of 2660 2662 Sequences Hydrolase activity 32% 32% Oxidoreductase activity 17% 17% Metal ion binding 13% 13% Transmembrane transporter activity 11% 11% DNA binding 10% 10% ATP binding 9% 9% Transferase activity; transferring 8% 8% phosphorus-containing groups 195