GLYCAN-COATED MAGNETIC NANOPARTICLES FOR FOOD SAFETY AND SECURITY By Saad Asadullah Sharief A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biosystems Engineering – Doctor of Philosophy 2023 ABSTRACT Foodborne illnesses result in many hospitalizations worldwide, and rapid detection of causative pathogens is critical for outbreak prevention. Before detection, however, enrichment of pathogens is usually required to increase the minimum bacterial count. To address this, magnetic nanoparticle (MNP)-based extraction methods have often been proposed but have limitations with expenses related to recognition ligands and cold-storage requirements. This study used glycan-coated magnetic nanoparticles (gMNP) for the concentration and extraction of pathogens from various foods. The gMNP were synthesized using a simple one-pot method, lacking expensive recognition moieties, and were found to be stable for over 3 years at room temperature. Transmission electron microscopy and confocal laser microscopy confirmed the binding of gMNP to bacterial cells in buffer solution and food matrices, respectively. The gMNP successfully captured cells in high pH environments where they displayed a net-negative zeta potential showing their binding mechanism extends beyond electrostatic interaction. The successful concentration of S. enterica and E. coli was demonstrated in cucumber, raw chicken, and lettuce samples, with natural microflora being typical among these foods. The extraction was confirmed using qPCR, demonstrating that the gMNP-qPCR system can be used in the rapid assessment of low pathogen contamination in complex food matrices. While gMNP successfully extracted bacterial cells from foods, their rapid and specific detection is critical. To achieve this, near infrared spectroscopy (NIRS) and gold nanoparticle (GNP)-based biosensor were assessed for feasibility. Among the primary obstacles with current detection methods are expenses and complexities related to sample preparation which often require technical skillset. NIRS can provide for sensitive and affordable screening, but its incapacity to identify low bacterial loads and the influence of food debris have prevented its wide application. The gMNP were used to address these drawbacks by concentrating bacterial cells and separating them from the interfering matrix. Following magnetic extraction of E. coli, NIRS was used for bacterial classification, and a classification accuracy of >90% was observed in bacterial samples extracted from a pure culture. At low bacterial loads, the processed spectra in the wavelength range of 1100 nm to 1350 nm showed a clear difference between bacterial samples with gMNP and control lacking gMNP. Serially diluted samples were used to establish a limit of detection with bacterial loads as low as 102 CFU/mL, and an accuracy of 85% was achieved using Support Vector Machine (SVM). Using gMNP allowed for bacterial concentration and resulted in an enhanced signal from the spectrophotometer; their presence did not hinder the spectral acquisition. While the gMNP-NIRS platform offers a reasonable solution for outbreak prevention related to foods contaminated with a single bacterial type, high bacterial load in natural microbiota can present problems. Alternatives that are specific to the pathogen of interest are therefore necessary. Gold nanoparticle (GNPs)-based plasmonic/colorimetric biosensors have recently gained attention owing to their remarkable surface plasmon resonance but have often required long probe- functionalization procedures or genomic amplification prior to detection. In this study, highly stable dextrin-capped GNPs (dGNP) were used for the detection of E. coli targeting the uidA gene and S. enterica targeting the invA gene. The synthesized dGNPs showed a characteristic wavelength peak at 520 nm in the visible spectrum and were <50 nm in diameter, confirmed with TEM and Dynamic Light Scattering. The dGNP biosensor demonstrated a detection limit of 10 ng/µL of DNA (p < 0.05) for E. coli while a limit of 5 ng/µL was achieved for S. enterica. The presence of target DNA was indicated by the stability of dGNPs due to their binding to the target DNA and was confirmed with a shift in peak wavelength away from 520 nm. The biosensor differentiated E. coli from S. enterica, K. pneumoniae, and Enterobacter cloacae in pure culture, followed by successful detection from lettuce and spinach. The biosensor also successfully differentiated S. enterica from E. coli, K. pneumoniae, E. cloacae, L. monocytogenes, and B. cereus in pure culture. Melons and cucumbers contaminated with S. enterica were also successfully detected. The total assay time including gMNP extraction and dGNP detection from foods was <7 h in the presence of natural microflora and food microparticles. Finally, the application of magnetic nanoparticles towards the prevention of counterfeit foods was also tested. Using DNA as an anticounterfeiting tag offers several advantages such as remarkable information density, high level of encryption, and difficulty in replication. However, its association with consumer products has some major obstacles, including the degradation of DNA in tags and time-consuming sequence determination. To address this, magnetic nanoparticles were coated with silica to provide protection to DNA-based anticounterfeiting tags. The DNA on silica-coated magnetic nanoparticles (SCMP) was found to be stable following exposure to temperatures as high as 90 °C and to UV light. The DNA/SCMP conjugate was also found to tolerate the effect of DNase while control samples readily degraded. The dGNP biosensor developed earlier was used for the successful detection of target DNA on SCMP. The DNA-based anticounterfeiting tags were readily extracted and detected from various materials including aluminum, clingwrap, and polystyrene. Copyright by SAAD ASADULLAH SHARIEF 2023 Dedicated to my parents and beloved sister, Fatima. v ACKNOWLEDGEMENTS I would first like to thank my advisor Dr. Evangelyn Alocilja for allowing me to be a part of the Nano-Biosensors Laboratory and for her support in conducting meaningful research over the last few years. I would also like to thank my research supervisory committee members, Dr. Renfu Lu, Dr. Gemma Reguera, and Dr. Prem Chahal for their input and guidance which was valuable in shaping this thesis. The members of the Nano-Biosensors Laboratory deserve credit for all that this work entails. I would like to thank Leann, Flora, and Vinni for helping me transition into the lab. My special thanks go to Oznur for helping in the design of the methods used in this work. I would also like to thank Emma and Chelsie for their helpful conversations and a positive lab ambience which made this research enjoyable. I would like to thank Alicia Withrow for her help in acquiring images with the Transmission Electron Microscope. I would also like to thank Jason Scott Smith for giving me an opportunity to work as a Teaching Assistant over the last many years so I could continue working on this research. The BAE family has been exceptionally helpful during my time here at Michigan State. I would like to thank my fellow graduate students and friends, graduate program coordinators Dr. Steven Safferman and Dr. Pouyan Nejadhashemi, and graduate secretaries Barb DeLong, Sarah Eubanks, and Tara Miller for their help and support. Finally, I would like to thank my family back home and friends across the US, UAE, and India for their unconditional support and encouragement during this long journey. vi PREFACE Chapter 3 of this document, “Magnetic extraction and qPCR detection of low-level pathogens from food”, is under review by the Public Library of Science One (PLOS One). Chapter 5 of this document, “Carbohydrate-coated magnetic and gold nanoparticles for point-of-use food contamination testing”, has been published in the Biosensors and Bioelectronics: X with minor changes, and is reprinted with permission from Elsevier (DOI: 10.1016/j.biosx.2023.100322). Chapter 6 of this document, “Carbohydrate-coated nanoparticles for PCR-less genomic detection of Salmonella from fresh produce”, has been published in Food Control, and is reprinted with permission from Elsevier (DOI: 10.1016/j.foodcont.2023.109770). Parts of Chapter 7 of this document, “DNA-based anticounterfeiting tags”, have been published in the International Journal of Pharmaceutics and has been reprinted with permission from Elsevier (DOI: 10.1016/j.ijpharm.2021.120580). vii TABLE OF CONTENTS Chapter 1. INTRODUCTION AND REVIEW OF LITERATURE ........................................................................... 1 1.1. Introduction ............................................................................................................................................ 1 1.2. Review of literature ................................................................................................................................ 3 REFERENCES ................................................................................................................................................ 18 Chapter 2. HYPOTHESES AND OBJECTIVES OF RESEARCH .......................................................................... 28 2.1. Hypotheses .......................................................................................................................................... 28 2.2. Goals and objectives ............................................................................................................................ 28 2.3. Novelty and significance of research ................................................................................................... 29 2.4. Experimental plan and methodology ................................................................................................... 32 REFERENCES ................................................................................................................................................ 44 Chapter 3. MAGNETIC EXTRACTION AND QPCR DETECTION OF LOW-LEVEL PATHOGENS FROM FOOD .. 45 3.1. Introduction ......................................................................................................................................... 45 3.2. Materials and Methods ........................................................................................................................ 48 3.3. Results and Discussion ......................................................................................................................... 51 3.4. Conclusion ............................................................................................................................................ 61 REFERENCES ................................................................................................................................................ 62 Chapter 4. RAPID CLASSIFICATION OF B. CEREUS USING MAGNETIC NANOPARTICLES, SPECTROSCOPY, AND SUPERVISED MACHINE LEARNING ...................................................................................................... 68 4.1. Introduction ......................................................................................................................................... 68 4.2. Materials and Methods ........................................................................................................................ 71 4.3. Results and discussion ......................................................................................................................... 73 4.4. Conclusion ............................................................................................................................................ 81 REFERENCES ................................................................................................................................................ 82 Chapter 5. RAPID DETECTION OF E. COLI FROM LEAFY GREENS USING CARBOHYDRATE-COATED NANOPARTICLES ......................................................................................................................................... 86 5.1. Introduction ......................................................................................................................................... 86 5.2. Materials and Methods ........................................................................................................................ 89 5.3. Results and Discussion ......................................................................................................................... 93 5.4. Conclusion .......................................................................................................................................... 103 REFERENCES .............................................................................................................................................. 104 Chapter 6. RAPID DETECTION OF SALMONELLA FROM MELONS AND CUCUMBERS USING CARBOHYDRATE-COATED NANOPARTICLES ............................................................................................. 109 6.1. Introduction ....................................................................................................................................... 109 6.2. Materials and Methods ...................................................................................................................... 111 6.3. Results ................................................................................................................................................ 115 6.4. Conclusion .......................................................................................................................................... 124 REFERENCES .............................................................................................................................................. 126 Chapter 7. DNA-BASED ANTICOUNTERFEITING TAGS .............................................................................. 131 7.1. Introduction ....................................................................................................................................... 131 7.2. Materials and Methods ...................................................................................................................... 134 7.3. Results and Discussion ....................................................................................................................... 137 REFERENCES .............................................................................................................................................. 144 Chapter 8. CONCLUSIONS AND RECOMMENDATIONS ............................................................................. 147 viii 8.1. Conclusions ........................................................................................................................................ 147 8.2. Recommendations ............................................................................................................................. 148 ix Chapter 1. INTRODUCTION AND REVIEW OF LITERATURE 1.1. Introduction With a growing world population, there is an increased need for food safety and security. As the food industry expands to satisfy growing consumer needs, foodborne illnesses have proven to be a cause of concern. The number of food-related pathogen outbreaks in North America has grown over the past years, affecting several people. In the United States, such outbreaks result in numerous hospitalizations and fatalities yearly, with a financial burden of more than $ 90 billion (Scharff, 2020). The effects of food-related illnesses are commonly seen in several other parts of the world, with children being among the primary victims. The severity of the consequences of such diseases is further multiplied in areas with a scarcity of medical resources. On a global scale, foodborne pathogens result in numerous infections and deaths each year. Health- related implications of associated diseases range from mild illnesses to hospitalizations and even death. Several pathogen types are involved in food contamination, including bacteria, viruses, and parasites. Among these, bacterial pathogens are responsible for a majority of foodborne illnesses. Infections from these pathogens arise from unsanitary production practices or packaging facilities. Some significant examples of bacterial pathogens that are involved in food-related outbreaks include multiple strains of Salmonella and Escherichia coli, Bacillus cereus, Campylobacter, and Listeria monocytogenes. Identifying pathogenic bacteria commonly associated with food products is critical for disease prevention. Traditional protocols that have successfully identified pathogens from food have often require growing bacteria for prolonged durations (X. Zhao et al., 2014). The enrichment of pathogens in the conventional protocols is necessary to ensure the presence of a minimum number of cells required for detection (Law et al., 2015; Mandal et al., 2011). Apart from being time-consuming, this step is tedious and may need specialized equipment and growth media (Hameed et al., 2018). Both traditional and rapid detection routes experience a time constraint due to the pre-enrichment of pathogens from samples (Benoit & Donahue, 2003; Paniel & Noguer, 2019). Such protocols not only contribute to an increase in time but also make use of equipment and labor that adds to the complexity of the process. It is, therefore, desirable to eliminate or shorten the culturing step by concentration for rapid pathogen identification. Conventional methods for the concentration of pathogens from food include physical processes such as centrifugation and filtration. Such techniques, however, are limited by the size of equipment, capture of food debris, or clogged filters. Other affinity-based methods are inexpensive and not applicable in low- 1 resource settings. More recently, magnetic nanoparticles (MNPs) have gained traction for food pathogen capture owing to their simplicity, rapidity, and efficiency. However, expenses related to recognition ligands, such as antibodies and phages, functionalized on the MNPs have demonstrated a need for alternatives. Coating the MNPs with carbohydrates can significantly reduce production costs while allowing prolonged storage at room temperature. Besides minimizing the time required for enrichment, effective pathogen detection methods are paramount for outbreak prevention. Recent progress in pathogen identification technologies has resulted in detection routes that are rapid, sensitive, and highly selective. However, techniques that are additionally cost-effective and simple, without a requisite scientific skillset, are urgently needed for broad applicability. This dissertation describes the use of MNPs coated with carbohydrate glycan (gMNP) for the extraction of multiple bacterial types from various foods. The primary focus of this work is towards rapid, inexpensive, and simple detection platform for food safety and security that can easily be applied in low- resource settings. For food safety, the bacteria under focus are Salmonella enterica Serovar Enteritidis (S. enterica), Escherichia coli, and Bacillus cereus. For security, the use of DNA-based anticounterfeiting tags is proposed on packaging. The following hypotheses are proposed: • The gMNP can be used for extraction and concentration of the above-mentioned bacteria from various foods followed by detection using standard methods. • The gMNP-extracted bacteria can be screened using near-infrared spectroscopy and specific detection will be achieved using dextrin-capped gold nanoparticle biosensor. • Magnetic nanoparticles can be used for protection of DNA-based anti-counterfeiting tags and dextrin-capped gold nanoparticles can be used for their rapid detection to prove product authenticity. Briefly, chapter 1 provides the introduction of the problem. Chapter 2 overviews recent progress in techniques for effective pathogen separation from complex food matrices and their rapid detection. The primary challenges with current methods and how carbohydrate (glycan)-coated nanoparticles (gMNP) can be used to address them are discussed. Chapter 3 discusses the application of gMNP towards concentration, extraction, and qPCR detection of two common foodborne pathogens- E. coli and S. enterica. The application of near infrared spectroscopy as a rapid screening tool for the detection of gMNP-extracted B.cereus is detailed in Chapter 4. Chapter 5 and 6 elaborate on the use of dextrin- 2 capped gold nanoparticles (dGNP) for detection of pathogens from food - detection of E. coli from leafy greens (Chapter 5) and detection of S. enterica from fresh produce (Chapter 6). Finally, Chapter 7 discusses the application of magnetic nanoparticles and dGNP for addressing the problem of counterfeiting of important edibles. 1.2. Review of literature Outbreaks related to contaminated food are a serious concern, making millions ill and leading to numerous deaths each year. An earlier report published by the World Health Organization indicated approximately 420,000 deaths from 600 million foodborne infections in 2010 (Havelaar et al., 2015). Some common bacteria that lead to food-related illnesses include Salmonella enterica, Listeria monocytogenes, Escherichia coli and Bacillus cereus, among others (J. Y. Huang et al., 2016; Liu et al., 2019). Identification of pathogenic bacteria commonly associated with food products is therefore critical for disease prevention. Although traditional protocols have been successful in biochemical and visual detection of foodborne pathogens, such procedures often require growing bacteria for longer durations (X. Zhao et al., 2014). Enrichment and culturing of pathogens in the recommended protocols is necessary to ensure the presence of minimum number of cells required for detection (Law et al., 2015; Mandal et al., 2011). Apart from being time consuming, this step is tedious and may require specialized equipment and growth media (Hameed et al., 2018). Both traditional and rapid detection routes experience a time constraint owing to pre-enrichment of pathogens from samples before proceeding to their detection (BENOIT & DONAHUE, 2003; Paniel & Noguer, 2019). Such protocols not only contribute to an increase in time, but also make use of equipment and labor that adds to the complexity of the process. It is therefore desirable to eliminate or shorten the culturing step for rapid pathogen identification. 1.2.1. Methods for bacterial separation 1.2.1.1. Physical methods for bacterial concentration Pre-enrichment may be eliminated by employing techniques to effectively concentrate bacteria. Centrifugation, filtration, and polymer-based partitioning are a few examples of physical methods to concentrate and isolate bacteria from food (Benoit & Donahue, 2003; Coakley et al., 2000; Kroll, 1985). While both centrifugation and polymer-based partitioning do not allow effective separation, filtration is limited by clogged filters. Variations of these methods have also been used to concentrate pathogens from food. Buoyant density gradient centrifugation for example has been used to concentrate multiple 3 pathogens from food homogenates (Fukushima et al., 2007). Similarly, cross-flow microfiltration has been used for isolation and concentration of S. enterica from chicken rinse with a 70% recovery of viable cells (X. Li et al., 2013). While these methods successfully concentrate pathogens and eliminate inhibitors to allow PCR based detection, they can prove to be expensive owing to their sophistication. Recently, microfluidic devices are gaining popularity for concentration and detection of food pathogens. However, apart from being expensive, these systems may not allow working with large sample volumes (Kant et al., 2018). Zhang et al., developed an efficient automated bacteria concentration and recovery system, their application was however limited to water contaminated samples (Y. Zhang et al., 2018). Apart from the physical methods for non-cultural separation and concentration of bacteria discussed above, biological and biochemical methods have also been used. Examples of such methods include immuno-separation (Y. Zhao et al., 2009), bacteriophage-based (Z. Wang, Wang, Kinchla, et al., 2016), and nanoparticle-based separation. These methods are discussed in further detail in the sections that follow. It is important to note that techniques that are both inexpensive and implementable in a low resource setting are highly desirable. 1.2.1.2. Chemical and biological methods for bacterial concentration The chemical and biological methods that have been used for concentration of bacterial cells from food rely on adsorption and utilize the interactions between affinity ligands and substrates on solid support. A common example of such methods is dielectrophoresis. This method makes use of polarization form a non-uniform electrical field for cell migration towards electrodes (Sarno et al., 2021). While this method has its own advantages such as rapidity, its charge-based nature presents problems. Several earlier studies have shown that food particles possess an inherent charge which results in the separation of food microparticles along with bacterial cells. Other examples include metal hydroxides such as zichronium hydroxide and titanium hydroxide. These materials have successfully captured bacterial cells by immobilizing them through a chelation process. However, centrifugation is still required to separate bacterial cells (Ogawa et al., 2021; Y. Wang & Salazar, 2016). Further, this method does not address the issue of capture of charged food microparticles. 1.2.1.3. Magnetic nanoparticles for bacterial concentration and extraction Recently, magnetic nanoparticles (MNPs) are gaining much attention as effective agents to rapidly concentrate pathogens. When functionalized, MNPs are capable of binding to pathogens and their magnetic nature allows rapid isolation. Additionally, this route does not utilize any complicated 4 equipment, resulting in simple, rapid, and effective separation. Magnetic nanoparticles may be synthesized using uncomplicated, one pot methods and can be surface functionalized after to enable bacteria binding. Various functional groups and biomolecules that have been used to surface modify MNPs include glycan, imine and amine groups, bacteriophages, and antibodies (B. Chen et al., 2019; Y.-F. Huang et al., 2010; Matta & Alocilja, 2018; Z. Wang et al., 2020). Antibody-based magnetic separation: Magnetic particles can offer significant advantages in separation and concentration of bacteria from various sources. Conventionally, particles with immunomagnetic properties have been proposed as means for isolation and concentration of bacteria. Typically, magnetic beads are conjugated with antibodies that can bind to antigens on the surface of pathogen of interest, allowing bacterial concentration (Jones, 2015). Magnetic nature of the particles provides for an easy separation of the pathogen and their application in various food matrices has been studied in literature. Further, antibody based magnetic separation, often referred to as Immunomagnetic separation (IMS), has a significant advantage of being specific for the target pathogen. Simonson et al., for instance extracted Vibrio vulnificus from homogenized seafood using IMS with binding efficiencies of about 50 % (JADEJA et al., 2010). Chen et al., successfully demonstrated approximately 90% recovery of Salmonella from milk, raw chicken rinse and egg whites, and further elaborated the effect of size of particles on recovery (J. Chen & Park, 2018). Elsewhere, immunomagnetic isolation of E. coli from ground beef and its rapid detection using electrochemical and spectrometric techniques has also been shown (Ochoa & Harrington, 2005; Xu et al., 2016). While use of IMS for pathogen separation and concentration seems to be an ideal solution, blocking of the antibody by food debris can result in prevention of capture and separation of bacteria. As Dwivedi et al., have indicated, the binding between antigen and antibody depends on the affinity between them, antigenic expression, and importantly on physico-chemical properties of the suspending sample matrix (Dwivedi & Jaykus, 2011). Moreover, antibody production can be both time consuming and expensive, limiting their application. Bacteriophage-based magnetic separation: Another method that allows targeted separation of pathogens involves coating magnetic particles with phages specific to bacteria of interest. This method is similar to antibody based magnetic separation, except that the antibodies are replaced with phages on magnetic particles (J. W. Kretzer et al., 2007). Multiple examples exist in literature where phage based magnetic particles have been used for pathogen separation from both water and foods such as lettuce, milk, cheese, and salmon, among others (J. Kretzer et al., 2018; Z. Wang, Wang, Kinchla, et al., 2016). However, coating of magnetic particles with phages may prove to be a tedious process and can 5 take long hours (Z. Wang, Wang, Chen, et al., 2016). Further, issues related to desiccation and low density of capture elements on magnetic particles are important challenges that need to be addressed (Bayat et al., 2021). Biomolecule-based magnetic separation: Since the use of antibodies and phages can present multiple obstacles, functionalizing magnetic particles with other inexpensive biomolecules may offer a desirable solution. While functionalizing MNPs with biomolecules may lead to non-specific binding, biosensor and amplification-based platforms may help in their detection with specificity. Numerous examples exist in literature where functionalized MNPs have been used to capture harmful pathogens. The antibiotic, vancomycin, has been proposed as an efficient candidate to functionalize MNPs for rapid bacterial capture. It has been shown to capture Vancomycin Resistant Enterococci and other gram-positive bacteria at ultra-low concentrations (Gu et al., 2003). Further, it is small compared to antibodies allowing each MNP to have comparatively more recognition elements. Recognizing the transpeptidase sequence, vancomycin is known to bind to gram positive bacteria, although examples exist where it was applied to capture gram negative bacteria. A difference in capture efficiencies between the two bacterial types was however noted (Chu et al., 2013; Kell et al., 2008). Although advantageous, this biomolecule is known to be bactericidal in nature (Chu et al., 2013). Further, the synthesis of Vancomycin coated MNPs can involve multiple steps that span several hours (Zhu et al., 2015), indicating a desirability of alternatives. Examples exist in literature where molecules other than vancomycin have been functionalized on MNPs for capture of pathogens. In an earlier study for instance, polyethyleneimine (PEI) was functionalized on MNPs for capture of pathogens including B. subtilis, S. aureus and E. coli suspended in phosphate buffer. At a bacterial concentration of 1 X 108 CFU/mL, capture efficiencies of >90% were achieved (Fang et al., 2016). It is important to note that implementation of magnetic extraction of pathogens from food matrices is critical. Huang et al., for example, used amine functionalized magnetic nanoparticles to successfully extract bacteria from water, green tea and grape juice achieving capture efficiencies greater than 85 % (Y.-F. Huang et al., 2010). Chen et al., used zirconia based magnetic nanoparticles to extract S. aureus and E. coli in apple juice and lettuce samples and detected them using mass-spectrometry (C.-T. Chen et al., 2019). Elsewhere, MNPs coated with carbohydrate ligands were used to capture various pathogens (Matta & Alocilja, 2018). Some examples of carbohydrate coatings that have successfully been employed to extract pathogens include cysteine-glycan, lysine-short chain glucan, mannose, galactose, and biotinylated oligosaccharides. Among these, glycan(chitosan)-coated MNPs (gMNP) have gained traction in recent times owing to the simplicity of their synthesis. For wide acceptance and applicability, it is important for 6 MNPs to provide affordable synthesis, inexpensive and long-term storage at room temperature, and implementation in large-volume samples. Table 1 below compares various surface-modified MNPs that have been used for capture, concentration, and detection of foodborne pathogens. It is important to note that concentration and magnetic extraction from a range of food types is of primary importance. Further, their combination with various rapid detection biosensor platforms and compatibility with traditional amplification-based detection techniques is also critical. 1.2.1.4. Glycan(chitosan)-coated magnetic nanoparticles for pathogen capture The glycan(chitosan)-coated MNPs (gMNP) offer several advantages including simple synthesis, long shelf life at room temperature, and lack of recognition ligands. Additionally, unlike other recognition ligand-lacking MNPs which require multiple steps for synthesis and functionalization, gMNP allow one- pot synthesis. The gMNP also offer scalable production, allowing their wide commercial application. Figure 1.1 below lists the primary advantages associated with the use of gMNP for pathogen extraction and detection. Figure 1. 1. Primary advantages with the use of gMNP for bacterial extraction and concentration. A few examples in literature have elucidated the application of gMNP for bacterial extraction from foods. Notably, the capture of bacterial cells from various complex food matrices in large volumes was recorded. In an earlier study for example, the gMNP were successful in concentrating S. enterica, E. coli O157:H7, and B. cereus from various milk types including vitamin D, 2% reduced fat, and fat-free (Matta & Alocilja, 2018). The study noted a capture efficiency between 73-90% in all combinations of milk. In another work, Salmonella, E. coli, and L. monocytogenes were successfully captured from other complex 7 matrices such as homogenized egg and apple cider (Matta et al., 2018). The extraction of pathogens from food matrices was also followed by potable biosensor detection, making the gMNP applicable in low-resource settings. Although gMNP have been used to extract and concentrate pathogens from various foods, liquid matrices have been a primary focus. Not many examples exist where gMNP have been applied to solid complex foods such as fresh produce and ready-to-eat foods. Table 1.1 below gives a brief overview of the various ligand-lacking MNPs that have been used in the past, their capture efficiency in sterile buffer, and foods, and compatibility with detection platforms. Table 1. 1. Various ligand-lacking MNPs used for pathogen capture. Surface Microorganism Capture in Tested Detection Method Ref modification Gold-coated microspheres PEI-coated PEI-coated E. coli E. coli S. typhimurium S. marcescens E. coli L. monocytogenes S. aureus B. subtilis buffer foods 65% Tap water, Raman (C. Wang et al., milk NA NA spectroscopy 2016) Plating (Z. Li et al., 2019) Optical density (B. Chen et al., 2019) >90% ~90% ~90% ~90% ~90% ~40% ~40% Protamine-coated Hepatitis A virus NA (4.9% in Milk qPCR (Wu et al., 2019) E. coli O157 food) >90% Sausage Plating (S.-M. You et al., 2021) E. coli 17-34% PBS Luciferase assay (Yosief et al., Lysine- Short Chain Glucans Biotinylated Oligosaccharides gMNP gMNP S. enterica E. coli B. cereus S. enterica E. coli gMNP E. coli O157-LVC gMNP E. coli-LVC LVC- Lage Volume Capture (100 mL) 89% 91% 89% NA NA NA Milk Plating (Matta & Alocilja, 2018) 2013) Egg, milk, Radio-frequency (Matta et al., apple cider Flour biosensor Biosensor 2018) (Dester et al., 2022) Lettuce, spinach Biiosensor (Sharief et al., 2023) 8 1.2.2. Identification of foodborne pathogens 1.2.2.1. Standard methods While traditional culture-based methods are considered the gold standard owing to their simplicity and reliability, more sophisticated methods for rapid bacteria detection are gaining popularity (Abayasekara et al., 2017; Ahmed et al., 2014). Examples of newly developed methods include DNA-based biosensors which employ techniques such as Surface Plasmon Resonance (SPR), electrochemical biosensors, and spectroscopy-based methods (Ahmed et al., 2014; Vaisocherová-Lísalová et al., 2016a). These methods along with other PCR based approaches have demonstrated excellent sensitivity in detection of specific bacteria. Although such methods have addressed the time constraint in detection, quick extraction of DNA from pathogens in complex food matrices may prove to be difficult. Using gMNP to concentrate bacteria from large volumes of samples before proceeding with detection can effectively remove unwanted food debris and provide enough target DNA for detection. Amplification based methods such as PCR and its variations, and sequencing are conventional methods used to confirm the presence of target pathogen and have extensively been applied for detection from food (McKillip & Drake, 2004; Naravaneni & Jamil, 2005; Rijpens & Herman, 2002). These apart, isothermal amplification-based methods have also been used which provide an advantage of detecting viable bacteria (Cook, 2003). While these methods are reliable, it may be difficult to implement these in the field as a specific skillset and complicated equipment is required. Evidence exists in literature where detection of bacteria was achieved directly without DNA extraction, from complex food matrices using other methods. One established technique for rapid pathogen detection is Mass Spectrometry (MS). In an earlier report, this technique was used to detect E. coli from ground beef achieving sensitivity of 2 X 106 CFU/ml. While MS by itself was not amenable, it was combined with immunomagnetic separation for concentrating the pathogen before analysis (Ochoa & Harrington, 2005). Elsewhere, E. coli and S. aureus were detected from lettuce and apple juice at a concentration as low as 3 X 103CFU/mL. MS was however combined with Electron Spray Ionization (C.-T. Chen et al., 2019). Although sensitive, MS can prove to be expensive and may not be applicable to low resource settings. Other techniques such as lateral flow assays and luminance have also been used for detecting pathogens directly from food. Kretzer et al., for example used magnetic separation and reporter phage assay in a microwell plate for detection of L. monocytogenes from cheese, shrimp, salmon, and turkey. While detection at levels of approximately 102 CFU/ml were achieved in control matrix, food samples required enrichment for 16 h (J. Kretzer et al., 2018). In another study, Cho et al., 9 achieved a similar detection limit using a Lateral Flow Assay but antibodies were used for magnetic capture of E. coli from ground beef (I. H. Cho et al., 2015). Other technologies that have been used in rapid detection of pathogens include Raman spectrophotometry, chromatography and Enzyme Linked Immuno Sorbent Assay (Jadhav et al., 2018)(I.-H. Cho & Irudayaraj, 2013). These techniques, however, can be expensive and require a skillset, similar to PCR-based approaches. Considering the drawbacks of the methods stated above which may not be implementable in a low resource setting, this work elaborates on other techniques that can offer an optimal solution. Initially, gMNP are used for concentration of pathogens to eliminate or shorten the enrichment step. The separated pathogens can be screened with Near Infrared Spectroscopy and supervised machine learning techniques. 1.2.2.2. Near Infrared Spectroscopy for pathogen classification Of the many methods that may be applied for pathogen identification is Infrared (IR) Spectroscopy and its variations. A primary advantage of IR based techniques is that they are non-invasive and do not rely on extraction of DNA before the analysis. In addition, these techniques have a short analysis time. These methods rely on the fundamental vibrations, their overtones and combination bonds in functional groups from various cell wall compounds such as polysaccharides, depending on region of IR (Treguier et al., 2019a). Fourier Transform Infrared (FTIR) Spectroscopy is perhaps the most widely applied form of spectroscopy for rapid pathogen identification. Its primary advantages include high sensitivity and simplicity in operation and has been extensively applied to food samples (Grewal et al., 2014)(Puzey et al., 2008). Notably, Ravindranath et al., combined magnetic nanoparticles with FTIR and developed a system that can rapidly detect E. coli O157:H7 from milk and spinach (Ravindranath et al., 2009). However, their limit of detection was in the range of 104-105 CFU per 2 ml. While FTIR is a widely accepted technique, Near Infrared Spectroscopy (NIRS) has not gained much attention owing to difficulty in interpreting its spectra, although weaker signals allow NIR to penetrate deeper than mid-IR. Considering the primary advantages of NIRS which include low sensitivity to water and noninvasive measurement indicate that this technique needs to be explored further for food safety applications (Tian et al., 2021). Further, NIRS coupled with fiber optic probes can allow on-line monitoring of food to ensure its safety. The spectral range of NIRS spans from 780 nm to 2526 nm, between the visible and mid-IR region of the spectrum (J.-L. Li et al., 2016). The bands resulting from NIR generally correspond to combinations and 10 overtones of C-H, N-H and O-H modes of stretching (Cozzolino, 2012; Shao et al., 2010). While more data on the use of NIRS for pathogen classification would be desirable, enough examples exist that can prove the applicability of the technique. Rodriguez-Saona et al., for instance, detected E. coli, P. aeruginosa and multiple Bacillus species from apple juice. They achieved a 100% classification efficiency between various bacterial species and found the detection limit to be 104-105 CFU/ml (Rodriguez-Saona et al., 2004). Alexandrakis et al., further achieved strain level differentiation and a limit of 102 CFU/ml and similar classification efficiency, however, an isolated system in the form of growth media was used (Alexandrakis et al., 2008). Considering the high-water content of bacterial cells, obtaining their spectra in a dry state using anodisc membrane filtered films and directly from plates has also been observed (Krepelka et al., 2020; RODRIGUEZ-SAONA et al., 2004; Treguier et al., 2019b). While this route can successfully eliminate the influence of water on spectra, it requires a high bacterial load and may not be applicable for rapid identification. Although the mentioned studies have all used NIR region of the spectrum, much variation has been found in the spectrometer and acquisition mode used and the pretreatment method applied. Both grating NIR and FT-NIR have been successfully used for identification of various food pathogens (Rodriguez-Saona et al., 2004; Slavchev et al., 2015, 2017). Traditionally, the mode of spectral acquisition used has either been transmission, trans-reflectance or diffuse reflectance (Tian et al., 2021). In addition, it has also been indicated that a second derivative of the spectra results in some additional structures which may be used for better classification efficiency. Second derivative of the spectra has often been combined with pretreatment techniques which include data centering, normalization and smoothing. As stated earlier, in comparison to FTIR, data from NIRS is relatively difficult to interpret and is therefore subjected to multivariate analysis and other machine learning tools. Slavchev et al., for example, used Principal Component Analysis (PCA) and Partial Least Square (PLS) for differentiation following smoothing and Multiplicative Scatter Correction (MSC) using grating NIR, spectral region between 1100-1850 was considered. While information on the limit of detection was not provided, growth time of 11.4-12 h was used (Slavchev et al., 2015). Elsewhere, second derivative of the spectra was used, PCA was however combined with soft independent modelling of class analogy (SIMCA) for discrimination of Pseudomonas spp., Lactococcus lactis, and Listeria innocua (Alexandrakis et al., 2008). Samples with dilutions as low as 10-6 for a 7 X 107 CFU/ml culture were used. Interestingly, 700-900 nm wavelength was considered in this study, false positives and negatives were however encountered with a classification efficiency of 77.4% for Listeria innocua. While near infrared spectroscopy-based techniques can be used for bacteria classification, surrounding food matrix and its inefficiency in low 11 bacterial loads are primary concerns. These could be addressed using gMNP which allow removal of food debris and concentrate bacterial cells providing the minimum load required for detection. NIRS can provide an excellent screening tool, although DNA sequence-based methods have their own advantages. 1.2.2.3. DNA based biosensor for pathogen detection Rapid and early identification of pathogenic microorganisms can help in prevention of food borne outbreaks. Point of care devices that are inexpensive, rapid, and applicable on the field are crucial to achieve this goal. Selectivity and sensitivity of detection of target analyte are critically important for the success of biosensor platforms. As stated earlier, gMNP can be used for capture of pathogens, which although not specific, can be detected with highly selective biosensors. This can ensure both efficient concentration of pathogens, and their rapid specific detection. With progress in various biosensor technologies, differentiation at genus, species, and strain level has proven to be possible. Electrochemical Biosensors: Among the prominent biosensor technologies is electrochemical biosensors, which are gaining popularity. These biosensors typically make use of single stranded DNA immobilized on an electrode with a sequence complementary to the DNA of interest (Drummond et al., 2003). The difference between the electrical properties of single stranded DNA and hybridized double stranded DNA help in identifying the correct DNA sequence belonging to specific pathogens. Cyclic voltammetry performed using a potentiostat is often used as a method to study such changes in electrochemical properties of the sample. Various nanomaterials have been applied for electrochemical detection of bacteria. Graphene and graphene oxide have recently gained attention due to their excellent electrical properties. These materials have also been employed for the detection of specific DNA sequences using electrochemical methods. The structural defects in graphene oxide are considered to be beneficial for electrochemical applications (Banks et al., 2005; McCreery, 2008). Tiwari et al., combined the excellent properties of graphene oxide with chitosan to detect E. coli DNA, using Electrochemical Impedance Spectroscopy (Tiwari et al., 2015). Gong et al., combined graphene with Nafion to synthesize a stable film which could detect the presence of a specific gene at concentrations of 2x10-14 M (Gong et al., 2017). Among other nanomaterials, Park et al., conjugated oligonucleotides with gold nanoparticles (GNPs) and used the changes in binding conductivities as a method to detect target DNA (Park et al., 2002). Hahm et al., utilized the P type semiconductor silicon to which oligonucleotides were conjugated and the change in conductivity was used as the sensing parameter (Hahm & Lieber, 2004). The combination of various 12 nanoparticles to detect DNA has also been observed in the literature. Chen et al., for example combined graphene with GNPs to detect specific DNA sequences (M. Chen et al., 2016). Electrochemical biosensors have been applied for detecting pathogens from various foods. Some examples include detection of S. Enteritidis from milk and orange juice reported earlier, where DNA as low as 0.1 ng/µL was successfully detected (Vetrone et al., 2012). Elsewhere, GNPs were incorporated into an electrochemical system to detect B. cereus from milk and infant formula, achieving sensitivities as low as 10 CFU/mL(Izadi et al., 2016). Electrochemical biosensors have made notable contributions as systems that can be used for rapid pathogen detection using both single stranded probes and antibodies (Z. Zhang et al., 2019). Although major advantages of this technology include availability of portable electrochemical sensor systems (Nordin et al., 2017) and cost effectiveness (Rowe et al., 2011), reduction of multiple sample preparation steps is preferable for field applicability. SPR-based biosensors: Following the demonstration of biosensing by surface plasmon resonance (SPR) by Liedberg and Nylander in 1982, this technique has been used in many applications by the scientific community (Liedberg et al., 1983). SPR makes use of diffraction due to the excitation of the surface plasmon waves (Kretschmann & Raether, 1968; Otto, 1968) and is utilized for detection of bacteria. The use of SPR for the detection of pathogens has been well established in the literature (Homola, 2003; Y. Huang et al., 2011; Rich, 2000). The technique can also be used for the detection of unique DNA sequences wherein single stranded complimentary sequences are immobilized on surface and change in plasmon resonance is observed upon hybridization. Different versions of DNA hybridization-based SPR can also be found in the literature. He et al., for example used gold-based SPR for ultrasensitive (<10 pM) detection of DNA (L. He et al., 2000). Endo et al., alternatively used a variant, Localized Surface Plasmon Resonance (LSPR) to detect even lower amounts of DNA, with a detection limit of 0.67 pM (Endo et al., 2005). Like electrochemical biosensors, SPR has been applied for detecting pathogens from various foods. Waswa et al., for example, detected E. coli O157:H7 from contaminated milk, apple juice and ground beef (Waswa et al., 2007). E. coli K12 and Shigella were used as non-targets which did not result in significant signals, confirming the specificity of the biosensor. Although other examples exist, SPR is affected by interfering sample matrix from food (Vaisocherová-Lísalová et al., 2016b). Additionally, 13 equipment complexity and cost ineffectiveness are major hindrances for the application of this technology. Colorimetric Biosensor: For a widely applicable system that can rapidly detect harmful pathogens, its ease of use, inexpensiveness and rapidity is crucial. Moreover, a biosensor that can achieve sensitive detection in a low resource setting is highly desirable. A visual system that can allow detection without the use of complicated equipment can possibly offer an ideal solution. Gold nanoparticles are characterized by excellent surface and optical properties and have been used extensively to detect DNA extracted from various pathogens. While GNPs have been used as a probe for various bench top equipment, changes in their visual appearance in response to their surrounding environment has useful applications. Xia and coworkers for instance succeeded in detecting DNA at picomolar concentrations using target DNA specific probes, GNPs, and conjugated polyelectrolyte. Stability of GNPs, which is indicated by maintenance of red appearance, was seen in the presence of target DNA. Absence of target DNA led to their aggregation indicated by their blue appearance (Xia et al., 2010). Baetsen-Young et al., followed a similar approach with fewer reagents where target DNA and specific probes were allowed to hybridize following addition of GNPs. Upon addition of sodium chloride, stability of GNPs indicated presence of target DNA and their aggregation indicated its absence, achieving a remarkable detection limit of 2.94 fM (Baetsen-Young et al., 2018). While the gold nanoparticle assay has widely been used to detect DNA extracted from pure cultures (Ahmadi et al., 2018; Bojd et al., 2017) its application for detecting pathogens from contaminated food has also been encountered. Further, GNPs have been used with variations in their surface chemistry, or without, in different works. Lin et al., detected cucumber green mottle mosaic virus (CGMMV) using unmodified GNPs from infected leaves and fruits (L. Wang et al., 2017), RT-PCR amplification was however required. Elsewhere, multiple strains of Salmonella spp. were concentrated using immunomagnetic separation and then detected from raw chicken using unmodified GNPs (Quintela et al., 2019). While a detection limit of <10 CFU/ml was obtained, the procedure involved an enrichment step and the overall procedure spanned 9 h. As an alternative to genomic DNA, PCR amplified product from food samples has also been used for detection of Salmonella spp. from food samples (Prasad et al., 2011). The use of amplification prior to detection can, however, increase costs and complexities. The dGNP used in this study, however, do not require any amplification and allow simple and rapid detection. 14 1.2.3. Selection of food matrices Several foodborne outbreaks are caused each year and are a result of contamination from various pathogens. These pathogens are often associated with a specific food-type. For instance, fresh produce has often been found to be contaminated with E. coli and Salmonella. Contamination usually arises during production, harvest, or packaging. Cross contamination during processing and washing of fresh produce has also been noted. Among other bacteria, the toxin-producing, spore-forming Bacillus cereus is usually associated with carbohydrate-rich foods and has commonly been found to contaminate rice. Taking into consideration the association of these bacteria with specific food types from past outbreaks reported by the Centers of Disease Control and Prevention, foods and bacteria tested in this study are listed in table 1.2. Table 1. 2. Food matrices used in this study. S. enterica E. coli B. cereus Melons Cucumbers Spinach Lettuce Rice Pasta 1.2.4. Counterfeit foods While food products are significantly prone to contamination with microorganisms, counterfeit foods have proven to be another major concern. Some examples of counterfeit food products include olive oil honey and saffron, among others (Moore et al., 2012)(Spink & Moyer, 2011). These apart, adulteration of milk, fish and coffee has been encountered earlier (Cozzolino, 2015). DNA can prove to be an effective anticounterfeiting tag as it is difficult to duplicate, is invisible, and can be encrypted. Various advantages and challenges with the use of DNA have been outlined in a recent work published earlier (Sharief et al., 2021). 1.2.4.1. Current progress on anti-counterfeiting The risk posed by counterfeiting of important foods has been well recognized and novel technologies are being proposed as effective measures. Examples of anti-counterfeiting technologies include the 15 traditional barcodes, quick responsive codes, holograms, watermarks and fluorescent inks to name a few (Liu et al., 2017). Recently, Radio Frequency Identification tags have been proposed to offer an efficient solution to this problem. Apart from being compact in size, RFID’s are inexpensive to manufacture (Elsherbeni, 2006; Sung-Lin Chen, 2009) thus proving to be an ideal solution to confirm authenticity of products. Unfortunately, however, it is also possible to clone these RFID tags (Mirowski, 2013), thereby hindering their application in this realm. While routes to prevent the cloning of RFID’s exist (Tuyls & Batina, n.d.), employing these may result in a direct increase in cost, defeating the purpose of their use. Another technology gaining attention with applications in anti-counterfeiting is nanotechnology with nanometer sized particles and their conjugates. Campos-Cuerva et al., used silver, gold and magnetic nanoparticles and combined them in varying proportions to be used as screen printable anti-counterfeiting ink (Campos-Cuerva et al., 2016). You et al., have proposed ink jet printing of fluorescent nanoparticles on blister packs to prevent the counterfeiting of drugs (M. You et al., 2016). Although nanoparticles have optical and spectroscopic advantages over other anti-counterfeiting technologies, a drawback lies in their encryption, hindering their application in the supply chain. 1.2.4.2. Application of DNA in anti-counterfeiting Indeed, the four individual bases of DNA can be arranged in countless ways to generate unique tags which could be tagged to valuable products to prevent their counterfeiting. The high density of information contained in DNA sequences (Goldman et al., 2013), encryption of information (Anam et al., 2010) and the ease in synthesis of specific DNA sequences make it a strong candidate for anti- counterfeiting applications. Examples exist in literature where DNA has been used for anti-counterfeiting applications both as a label and as a part of the product (Liu et al., 2017), (Puddu et al., 2014). Figure 1. 2. Primary challenges associated with the use of DNA for anti-counterfeiting applications. 16 Although the use of DNA for barcoding of products has been a topic of research in recent times, its commercial application has not yet been fully realized. Among the primary reasons that prevent the use of DNA for anti-counterfeiting applications is its instability in uncontrolled environments and rapid detection. While coating DNA with silica has been proposed as a solution (Liu et al., 2017, Paunescu, Fuhrer, et al., 2013; Paunescu, Puddu, et al., 2013; Puddu et al., 2014), the eventual extraction of DNA for further verification involves use of harmful chemicals such as Hydrogen Fluoride. Apart from using dangerous chemicals, the protocol for enveloping DNA with silica may extend to at least 2 days (Paunescu, Puddu, et al., 2013) which can prove to be a tedious process. Direct combination of printing ink with DNA and its subsequent retrieval for PCR amplification has also been reported in literature (Hashiyada, 2004), use of toxic chemicals such as phenol chloroform for DNA extraction is a primary disadvantage. Earlier, bioconjugation of plasmid DNA with silica nanoparticles to protect it from cleavage has been proposed (X. He et al., 2003), although for applications other than anticounterfeiting. An inexpensive route to protect DNA in labels without use of excessive supplemental resources is therefore desirable. Coating DNA with an external protective layer before it is made a part of the label may offer a possible solution. 17 REFERENCES Abayasekara, L. M., Perera, J., Chandrasekharan, V., Gnanam, V. S., Udunuwara, N. A., Liyanage, D. S., Bulathsinhala, N. E., Adikary, S., Aluthmuhandiram, J. V. S., Thanaseelan, C. S., Tharmakulasingam, D. P., Karunakaran, T., & Ilango, J. (2017). Detection of bacterial pathogens from clinical specimens using conventional microbial culture and 16S metagenomics: a comparative study. BMC Infectious Diseases, 17(1), 631. https://doi.org/10.1186/s12879-017-2727-8 Ahmadi, S., Kamaladini, H., Haddadi, F., & Sharifmoghadam, M. R. (2018). Thiol-Capped Gold Nanoparticle Biosensors for Rapid and Sensitive Visual Colorimetric Detection of Klebsiella pneumoniae. Journal of Fluorescence, 28(4), 987–998. https://doi.org/10.1007/s10895-018-2262-z Ahmed, A., Rushworth, J. V., Hirst, N. A., & Millner, P. A. (2014). Biosensors for Whole-Cell Bacterial Detection. Clinical Microbiology Reviews, 27(3), 631–646. https://doi.org/10.1128/CMR.00120-13 Alexandrakis, D., Downey, G., & Scannell, A. G. M. (2008). Detection and Identification of Bacteria in an Isolated System with Near-Infrared Spectroscopy and Multivariate Analysis. Journal of Agricultural and Food Chemistry, 56(10), 3431–3437. https://doi.org/10.1021/jf073407x Anam, B., Sakib, K., Hossain, Md. A., & Dahal, K. (2010). Review on the Advancements of DNA Cryptography. Baetsen-Young, A. M., Vasher, M., Matta, L. L., Colgan, P., Alocilja, E. C., & Day, B. (2018). Direct colorimetric detection of unamplified pathogen DNA by dextrin-capped gold nanoparticles. Biosensors and Bioelectronics, 101(October 2017), 29–36. https://doi.org/10.1016/j.bios.2017.10.011 Banks, C. E., Davies, T. J., Wildgoose, G. G., & Compton, R. G. (2005). Electrocatalysis at graphite and carbon nanotube modified electrodes: edge-plane sites and tube ends are the reactive sites. Chemical Communications, 7, 829. https://doi.org/10.1039/b413177k Bayat, F., Didar, T. F., & Hosseinidoust, Z. (2021). Emerging investigator series: bacteriophages as nano engineering tools for quality monitoring and pathogen detection in water and wastewater. Environmental Science: Nano, 8(2), 367–389. https://doi.org/10.1039/D0EN00962H Benoit, P. W., & Donahue, D. W. (2003). Methods for Rapid Separation and Concentration of Bacteria in Food that Bypass Time-Consuming Cultural Enrichment. Journal of Food Protection, 66(10), 1935– 1948. https://doi.org/10.4315/0362-028X-66.10.1935 Bojd, M. O., Kamaladini, H., Haddadi, F., & Vaseghi, A. (2017). Thiolated AuNP probes and multiplex PCR for molecular detection of Staphylococcus epidermidis. Molecular and Cellular Probes, 34, 30–36. https://doi.org/10.1016/j.mcp.2017.04.006 Campos-Cuerva, C., Zieba, M., Sebastian, V., Martínez, G., Sese, J., Irusta, S., Contamina, V., Arruebo, M., & Santamaria, J. (2016). Screen-printed nanoparticles as anti-counterfeiting tags. Nanotechnology, 27(9), 095702. https://doi.org/10.1088/0957-4484/27/9/095702 Chen, B., Xie, H., Zhang, A., Liu, N., Li, Q., Guo, J., & Su, B. (2019). Synthesis of PEI-Functionalized Magnetic Nanoparticles for Capturing Bacteria. Journal of Wuhan University of Technology-Mater. Sci. Ed., 34(1), 236–242. https://doi.org/10.1007/s11595-019-2041-y 18 Chen, C.-T., Yu, J.-W., & Ho, Y.-P. (2019). Identification of bacteria in juice/lettuce using magnetic nanoparticles and selected reaction monitoring mass spectrometry. Journal of Food and Drug Analysis, 27(2), 575–584. https://doi.org/10.1016/j.jfda.2018.09.006 Chen, J., & Park, B. (2018). Effect of immunomagnetic bead size on recovery of foodborne pathogenic bacteria. International Journal of Food Microbiology, 267, 1–8. https://doi.org/10.1016/j.ijfoodmicro.2017.11.022 Chen, M., Hou, C., Huo, D., Bao, J., Fa, H., & Shen, C. (2016). An electrochemical DNA biosensor based on nitrogen-doped graphene/Au nanoparticles for human multidrug resistance gene detection. Biosensors and Bioelectronics, 85, 684–691. https://doi.org/10.1016/j.bios.2016.05.051 Cho, I. H., Bhunia, A., & Irudayaraj, J. (2015). Rapid pathogen detection by lateral-flow immunochromatographic assay with gold nanoparticle-assisted enzyme signal amplification. International Journal of Food Microbiology, 206, 60–66. https://doi.org/10.1016/j.ijfoodmicro.2015.04.032 Cho, I.-H., & Irudayaraj, J. (2013). In-situ immuno-gold nanoparticle network ELISA biosensors for pathogen detection. International Journal of Food Microbiology, 164(1), 70–75. https://doi.org/10.1016/j.ijfoodmicro.2013.02.025 Chu, Y. W., Engebretson, D. A., & Carey, J. R. (2013). Bioconjugated Magnetic Nanoparticles for the Detection of Bacteria. Journal of Biomedical Nanotechnology, 9(12), 1951–1961. https://doi.org/10.1166/jbn.2013.1701 Coakley, W. T., Hawkes, J. J., Sobanski, M. A., Cousins, C. M., & Spengler, J. (2000). Analytical scale ultrasonic standing wave manipulation of cells and microparticles. Ultrasonics, 38(1–8), 638–641. https://doi.org/10.1016/S0041-624X(99)00151-1 Cook, N. (2003). The use of NASBA for the detection of microbial pathogens in food and environmental samples. Journal of Microbiological Methods, 53(2), 165–174. https://doi.org/10.1016/S0167- 7012(03)00022-8 Cozzolino, D. (2012). Recent Trends on the Use of Infrared Spectroscopy to Trace and Authenticate Natural and Agricultural Food Products. Applied Spectroscopy Reviews, 47(7), 518–530. https://doi.org/10.1080/05704928.2012.667858 Cozzolino, D. (2015). The role of vibrational spectroscopy as a tool to assess economically motivated fraud and counterfeit issues in agricultural products and foods. Analytical Methods, 7(22), 9390– 9400. https://doi.org/10.1039/C5AY01792K Dester, E., Kao, K., & Alocilja, E. C. (2022). Detection of Unamplified E. coli O157 DNA Extracted from Large Food Samples Using a Gold Nanoparticle Colorimetric Biosensor. Biosensors, 12(5), 274. https://doi.org/10.3390/bios12050274 Drummond, T. G., Hill, M. G., & Barton, J. K. (2003). Electrochemical DNA sensors. Nature Biotechnology, 21(10), 1192–1199. https://doi.org/10.1038/nbt873 Dwivedi, H. P., & Jaykus, L.-A. (2011). Detection of pathogens in foods: the current state-of-the-art and future directions. Critical Reviews in Microbiology, 37(1), 40–63. https://doi.org/10.3109/1040841X.2010.506430 Elsherbeni, A. (2006). Recent Rfid Technology and Applications. 19 Endo, T., Kerman, K., Nagatani, N., Takamura, Y., & Tamiya, E. (2005). Label-Free Detection of Peptide Nucleic Acid−DNA Hybridization Using Localized Surface Plasmon Resonance Based Optical Biosensor. Analytical Chemistry, 77(21), 6976–6984. https://doi.org/10.1021/ac0513459 Fang, W., Han, C., Zhang, H., Wei, W., Liu, R., & Shen, Y. (2016). Preparation of amino-functionalized magnetic nanoparticles for enhancement of bacterial capture efficiency. RSC Advances, 6(72), 67875–67882. https://doi.org/10.1039/C6RA13070D Fukushima, H., Katsube, K., Hata, Y., Kishi, R., & Fujiwara, S. (2007). Rapid Separation and Concentration of Food-Borne Pathogens in Food Samples Prior to Quantification by Viable-Cell Counting and Real- Time PCR. Applied and Environmental Microbiology, 73(1), 92–100. https://doi.org/10.1128/AEM.01772-06 Goldman, N., Bertone, P., Chen, S., Dessimoz, C., LeProust, E. M., Sipos, B., & Birney, E. (2013). Towards practical, high-capacity, low-maintenance information storage in synthesized DNA. Nature, 494(7435), 77–80. https://doi.org/10.1038/nature11875 Gong, Q., Wang, Y., & Yang, H. (2017). A sensitive impedimetric DNA biosensor for the determination of the HIV gene based on graphene-Nafion composite film. Biosensors and Bioelectronics, 89, 565– 569. https://doi.org/10.1016/j.bios.2016.02.045 Grewal, M. K., Jaiswal, P., & Jha, S. N. (2014). Detection of poultry meat specific bacteria using FTIR spectroscopy and chemometrics. Journal of Food Science and Technology, 52(6), 3859–3869. https://doi.org/10.1007/s13197-014-1457-9 Gu, H., Ho, P.-L., WT Tsang, K., Yu, C.-W., & Xu, B. (2003). Using biofunctional magnetic nanoparticles to capture Gram-negative bacteria at an ultra-low concentrationElectronic supplementary information (ESI) available: experimental details. See http://www.rsc.org/suppdata/cc/b3/b305421g/. Chemical Communications, 3(15), 1966. https://doi.org/10.1039/b305421g Hahm, J., & Lieber, C. M. (2004). Direct Ultrasensitive Electrical Detection of DNA and DNA Sequence Variations Using Nanowire Nanosensors. Nano Letters, 4(1), 51–54. https://doi.org/10.1021/nl034853b Hameed, S., Xie, L., & Ying, Y. (2018). Conventional and emerging detection techniques for pathogenic bacteria in food science: A review. Trends in Food Science & Technology, 81(July), 61–73. https://doi.org/10.1016/j.tifs.2018.05.020 Hashiyada, M. (2004). Development of Biometric DNA Ink for Authentication Security. The Tohoku Journal of Experimental Medicine, 204(2), 109–117. https://doi.org/10.1620/tjem.204.109 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. (2015). World Health Organization Global Estimates and Regional Comparisons of the Burden of Foodborne Disease in 2010. PLOS Medicine, 12(12), e1001923. https://doi.org/10.1371/journal.pmed.1001923 He, L., Musick, M. D., Nicewarner, S. R., Salinas, F. G., Benkovic, S. J., Natan, M. J., & Keating, C. D. (2000). Colloidal Au-Enhanced Surface Plasmon Resonance for Ultrasensitive Detection of DNA Hybridization. Journal of the American Chemical Society, 122(38), 9071–9077. https://doi.org/10.1021/ja001215b 20 He, X., Wang, K., Tan, W., Liu, B., Lin, X., He, C., Li, D., Huang, S., & Li, J. (2003). Bioconjugated Nanoparticles for DNA Protection from Cleavage. Journal of the American Chemical Society, 125(24), 7168–7169. https://doi.org/10.1021/ja034450d Homola, J. (2003). Present and future of surface plasmon resonance biosensors. Analytical and Bioanalytical Chemistry, 377(3), 528–539. https://doi.org/10.1007/s00216-003-2101-0 Huang, J. Y., Henao, O. L., Griffin, P. M., Vugia, D. J., Cronquist, A. B., Hurd, S., Tobin-D’Angelo, M., Ryan, P., Smith, K., Lathrop, S., Zansky, S., Cieslak, P. R., Dunn, J., Holt, K. G., Wolpert, B. J., & Patrick, M. E. (2016). Infection with Pathogens Transmitted Commonly Through Food and the Effect of Increasing Use of Culture-Independent Diagnostic Tests on Surveillance — Foodborne Diseases Active Surveillance Network, 10 U.S. Sites, 2012–2015. MMWR. Morbidity and Mortality Weekly Report, 65(14), 368–371. https://doi.org/10.15585/mmwr.mm6514a2 Huang, Y., Dong, X., Liu, Y., Li, L.-J., & Chen, P. (2011). Graphene-based biosensors for detection of bacteria and their metabolic activities. Journal of Materials Chemistry, 21(33), 12358. https://doi.org/10.1039/c1jm11436k Huang, Y.-F., Wang, Y.-F., & Yan, X.-P. (2010). Amine-Functionalized Magnetic Nanoparticles for Rapid Capture and Removal of Bacterial Pathogens. Environmental Science & Technology, 44(20), 7908– 7913. https://doi.org/10.1021/es102285n Izadi, Z., Sheikh-Zeinoddin, M., Ensafi, A. A., & Soleimanian-Zad, S. (2016). Fabrication of an electrochemical DNA-based biosensor for Bacillus cereus detection in milk and infant formula. Biosensors and Bioelectronics, 80, 582–589. https://doi.org/10.1016/j.bios.2016.02.032 JADEJA, R., JANES, M. E., & SIMONSON, J. G. (2010). Immunomagnetic Separation of Vibrio vulnificus with Antiflagellar Monoclonal Antibody. Journal of Food Protection, 73(7), 1288–1293. https://doi.org/10.4315/0362-028X-73.7.1288 Jadhav, S. R., Shah, R. M., Karpe, A. V., Morrison, P. D., Kouremenos, K., Beale, D. J., & Palombo, E. A. (2018). Detection of Foodborne Pathogens Using Proteomics and Metabolomics-Based Approaches. Frontiers in Microbiology, 9(December), 1–13. https://doi.org/10.3389/fmicb.2018.03132 Jones, J. L. (2015). High throughput screening of seafood for foodborne pathogens. In High Throughput Screening for Food Safety Assessment: Biosensor Technologies, Hyperspectral Imaging and Practical Applications. Elsevier Ltd. https://doi.org/10.1016/B978-0-85709-801-6.00020-4 Kant, K., Shahbazi, M.-A., Dave, V. P., Ngo, T. A., Chidambara, V. A., Than, L. Q., Bang, D. D., & Wolff, A. (2018). Microfluidic devices for sample preparation and rapid detection of foodborne pathogens. Biotechnology Advances, 36(4), 1003–1024. https://doi.org/10.1016/j.biotechadv.2018.03.002 Kell, A. J., Stewart, G., Ryan, S., Peytavi, R., Boissinot, M., Huletsky, A., Bergeron, M. G., & Simard, B. (2008). Vancomycin-Modified Nanoparticles for Efficient Targeting and Preconcentration of Gram- Positive and Gram-Negative Bacteria. ACS Nano, 2(9), 1777–1788. https://doi.org/10.1021/nn700183g Krepelka, P., Bolívar, A., & Pérez-Rodríguez, F. (2020). Two-dimensional mid and near infrared correlation spectroscopy for bacterial identification. Journal of Near Infrared Spectroscopy. https://doi.org/10.1177/0967033520974518 21 Kretschmann, E., & Raether, H. (1968). Notizen: Radiative Decay of Non-Radiative Surface Plasmons Excited by Light. Zeitschrift Für Naturforschung A, 23(12), 2135–2136. https://doi.org/10.1515/zna- 1968-1247 Kretzer, J., Schmelcher, M., & Loessner, M. (2018). Ultrasensitive and Fast Diagnostics of Viable Listeria Cells by CBD Magnetic Separation Combined with A511:luxAB Detection. Viruses, 10(11), 626. https://doi.org/10.3390/v10110626 Kretzer, J. W., Lehmann, R., Schmelcher, M., Banz, M., Kim, K.-P., Korn, C., & Loessner, M. J. (2007). Use of High-Affinity Cell Wall-Binding Domains of Bacteriophage Endolysins for Immobilization and Separation of Bacterial Cells. Applied and Environmental Microbiology, 73(6), 1992–2000. https://doi.org/10.1128/AEM.02402-06 Kroll, R. G. (1985). Electropositively charged filters for the concentration of bacteria from foods. Food Microbiology, 2(3), 183–186. https://doi.org/10.1016/0740-0020(85)90033-4 Law, J. W.-F., Ab Mutalib, N.-S., Chan, K.-G., & Lee, L.-H. (2015). Rapid methods for the detection of foodborne bacterial pathogens: principles, applications, advantages, and limitations. Frontiers in Microbiology, 5(DEC), 1–19. https://doi.org/10.3389/fmicb.2014.00770 Li, J.-L., Sun, D.-W., & Cheng, J.-H. (2016). Recent Advances in Nondestructive Analytical Techniques for Determining the Total Soluble Solids in Fruits: A Review. Comprehensive Reviews in Food Science and Food Safety, 15(5), 897–911. https://doi.org/10.1111/1541-4337.12217 Li, X., Ximenes, E., Amalaradjou, M. A. R., Vibbert, H. B., Foster, K., Jones, J., Liu, X., Bhunia, A. K., & Ladisch, M. R. (2013). Rapid Sample Processing for Detection of Food-Borne Pathogens via Cross- Flow Microfiltration. Applied and Environmental Microbiology, 79(22), 7048–7054. https://doi.org/10.1128/AEM.02587-13 Li, Z., Ma, J., Ruan, J., & Zhuang, X. (2019). Using Positively Charged Magnetic Nanoparticles to Capture Bacteria at Ultralow Concentration. Nanoscale Research Letters, 14(1), 195. https://doi.org/10.1186/s11671-019-3005-z Liedberg, B., Nylander, C., & Lunström, I. (1983). Surface plasmon resonance for gas detection and biosensing. Sensors and Actuators, 4, 299–304. https://doi.org/10.1016/0250-6874(83)85036-7 Liu, Y., Cao, Y., Wang, T., Dong, Q., Li, J., & Niu, C. (2019). Detection of 12 Common Food-Borne Bacterial Pathogens by TaqMan Real-Time PCR Using a Single Set of Reaction Conditions. Frontiers in Microbiology, 10(FEB), 1–9. https://doi.org/10.3389/fmicb.2019.00222 Liu, Y., Zhang, C., Li, X., & Wu, D. (2017). A self-monitored fluorescence DNA anti-counterfeiting system based on silica coated SYBR Green I/DNA gelatin nanoparticles. Journal of Materials Chemistry C, 5(24), 5939–5948. https://doi.org/10.1039/C7TC01380A Mandal, P. K., Biswas, A. K., Choi, K., & Pal, U. K. (2011). Methods for Rapid Detection of Foodborne Pathogens: An Overview. American Journal of Food Technology, 6(2), 87–102. https://doi.org/10.3923/ajft.2011.87.102 Matta, L. L., & Alocilja, E. C. (2018). Carbohydrate Ligands on Magnetic Nanoparticles for Centrifuge-Free Extraction of Pathogenic Contaminants in Pasteurized Milk. Journal of Food Protection, 81(12), 1941–1949. https://doi.org/10.4315/0362-028X.JFP-18-040 22 Matta, L. L., Karuppuswami, S., Chahal, P., & Alocilja, E. C. (2018). AuNP-RF sensor: An innovative application of RF technology for sensing pathogens electrically in liquids (SPEL) within the food supply chain. Biosensors and Bioelectronics, 111(March), 152–158. https://doi.org/10.1016/j.bios.2018.04.010 McCreery, R. L. (2008). Advanced Carbon Electrode Materials for Molecular Electrochemistry. Chemical Reviews, 108(7), 2646–2687. https://doi.org/10.1021/cr068076m McKillip, J. L., & Drake, M. (2004). Real-Time Nucleic Acid–Based Detection Methods for Pathogenic Bacteria in Food. Journal of Food Protection, 67(4), 823–832. https://doi.org/10.4315/0362-028X- 67.4.823 Mirowski, L. (2013). Exposing Clone RFID Tags at the Reader. 2013 12th IEEE International Conference on Trust, Security and Privacy in Computing and Communications, 1669–1674. https://doi.org/10.1109/TrustCom.2013.207 Moore, J. C., Spink, J., & Lipp, M. (2012). Development and Application of a Database of Food Ingredient Fraud and Economically Motivated Adulteration from 1980 to 2010. Journal of Food Science, 77(4), R118–R126. https://doi.org/10.1111/j.1750-3841.2012.02657.x Naravaneni, R., & Jamil, K. (2005). Rapid detection of food-borne pathogens by using molecular techniques. Journal of Medical Microbiology, 54(1), 51–54. https://doi.org/10.1099/jmm.0.45687-0 Nordin, N., Yusof, N. A., Abdullah, J., Radu, S., & Hushiarian, R. (2017). A simple, portable, electrochemical biosensor to screen shellfish for Vibrio parahaemolyticus. AMB Express, 7(1), 41. https://doi.org/10.1186/s13568-017-0339-8 Ochoa, M. L., & Harrington, P. B. (2005). Immunomagnetic Isolation of Enterohemorrhagic Escherichia c oli O157:H7 from Ground Beef and Identification by Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry and Database Searches. Analytical Chemistry, 77(16), 5258– 5267. https://doi.org/10.1021/ac0502596 Ogawa, U., Koyama, K., & Koseki, S. (2021). Rapid detection and enumeration of aerobic mesophiles in raw foods using dielectrophoresis. Journal of Microbiological Methods, 186. https://doi.org/10.1016/j.mimet.2021.106251 Otto, A. (1968). Excitation of nonradiative surface plasma waves in silver by the method of frustrated total reflection. Zeitschrift Für Physik A Hadrons and Nuclei, 216(4), 398–410. https://doi.org/10.1007/BF01391532 Paniel, & Noguer. (2019). Detection of Salmonella in Food Matrices, from Conventional Methods to Recent Aptamer-Sensing Technologies. Foods, 8(9), 371. https://doi.org/10.3390/foods8090371 Park, S.-J., Taton, T. A., & Mirkin, C. A. (2002). Array-based electrical detection of DNA with nanoparticle probes. Science (New York, N.Y.), 295(5559), 1503–1506. https://doi.org/10.1126/science.1067003 Paunescu, D., Fuhrer, R., & Grass, R. N. (2013). Protection and Deprotection of DNA-High-Temperature Stability of Nucleic Acid Barcodes for Polymer Labeling. Angewandte Chemie International Edition, 52(15), 4269–4272. https://doi.org/10.1002/anie.201208135 Paunescu, D., Puddu, M., Soellner, J. O. B., Stoessel, P. R., & Grass, R. N. (2013). Reversible DNA encapsulation in silica to produce ROS-resistant and heat-resistant synthetic DNA “fossils.” Nature Protocols, 8(12), 2440–2448. https://doi.org/10.1038/nprot.2013.154 23 Prasad, D., Shankaracharya, & Vidyarthi, A. S. (2011). Gold nanoparticles-based colorimetric assay for rapid detection of Salmonella species in food samples. World Journal of Microbiology and Biotechnology, 27(9), 2227–2230. https://doi.org/10.1007/s11274-011-0679-5 Puddu, M., Paunescu, D., Stark, W. J., & Grass, R. N. (2014). Magnetically Recoverable, Thermostable, Hydrophobic DNA/Silica Encapsulates and Their Application as Invisible Oil Tags. ACS Nano, 8(3), 2677–2685. https://doi.org/10.1021/nn4063853 Puzey, K. A., Gardner, P. J., Petrova, V. K., Donnelly, C. W., & Petrucci, G. A. (2008). Automated species and strain identification of bacteria in complex matrices using FTIR spectroscopy. In A. W. Fountain III & P. J. Gardner (Eds.), Chemical, Biological, Radiological, Nuclear, and Explosives (CBRNE) Sensing IX (Vol. 6954, Issue 1, p. 695412). https://doi.org/10.1117/12.783347 Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2019). Simultaneous Colorimetric Detection of a Variety of Salmonella spp. in Food and Environmental Samples by Optical Biosensing Using Oligonucleotide-Gold Nanoparticles. Frontiers in Microbiology, 10(MAY), 1–12. https://doi.org/10.3389/fmicb.2019.01138 Ravindranath, S. P., Mauer, L. J., Deb-Roy, C., & Irudayaraj, J. (2009). Biofunctionalized Magnetic Nanoparticle Integrated Mid-Infrared Pathogen Sensor for Food Matrixes. Analytical Chemistry, 81(8), 2840–2846. https://doi.org/10.1021/ac802158y Rich, R. (2000). Advances in surface plasmon resonance biosensor analysis. Current Opinion in Biotechnology, 11(1), 54–61. https://doi.org/10.1016/S0958-1669(99)00054-3 Rijpens, N. P., & Herman, L. M. F. (2002). Molecular Methods for Identification and Detection of Bacterial Food Pathogens. Journal of AOAC INTERNATIONAL, 85(4), 984–995. https://doi.org/10.1093/jaoac/85.4.984 Rodriguez-Saona, L. E., Khambaty, F. M., Fry, F. S., Dubois, J., & Calvey, E. M. (2004). Detection and Identification of Bacteria in a Juice Matrix with Fourier Transform–Near Infrared Spectroscopy and Multivariate Analysis. Journal of Food Protection, 67(11), 2555–2559. https://doi.org/10.4315/0362-028X-67.11.2555 Rowe, A. A., Bonham, A. J., White, R. J., Zimmer, M. P., Yadgar, R. J., Hobza, T. M., Honea, J. W., Ben- Yaacov, I., & Plaxco, K. W. (2011). CheapStat: An Open-Source, “Do-It-Yourself” Potentiostat for Analytical and Educational Applications. PLoS ONE, 6(9), e23783. https://doi.org/10.1371/journal.pone.0023783 Sarno, B., Heineck, D., Heller, M. J., & Ibsen, S. D. (2021). Dielectrophoresis: Developments and applications from 2010 to 2020. In Electrophoresis (Vol. 42, Issue 5, pp. 539–564). Wiley-VCH Verlag. https://doi.org/10.1002/elps.202000156 Scharff, R. L. (2020). Food Attribution and Economic Cost Estimates for Meat- and Poultry-Related Illnesses. Journal of Food Protection, 83(6), 959–967. https://doi.org/10.4315/JFP-19-548 Shao, X., Bian, X., Liu, J., Zhang, M., & Cai, W. (2010). Multivariate calibration methods in near infrared spectroscopic analysis. Analytical Methods, 2(11), 1662. https://doi.org/10.1039/c0ay00421a Sharief, S. A., Caliskan-Aydogan, O., & Alocilja, E. (2023). Carbohydrate-coated magnetic and gold nanoparticles for point-of-use food contamination testing. Biosensors and Bioelectronics: X, 13, 100322. https://doi.org/10.1016/j.biosx.2023.100322 24 Sharief, S. A., Chahal, P., & Alocilja, E. (2021). Application of DNA sequences in anti-counterfeiting: Current progress and challenges. International Journal of Pharmaceutics, 602(December 2020), 120580. https://doi.org/10.1016/j.ijpharm.2021.120580 Slavchev, A., Kovacs, Z., Koshiba, H., Bazar, G., Pollner, B., Krastanov, A., & Tsenkova, R. (2017). Monitoring of water spectral patterns of lactobacilli development as a tool for rapid selection of probiotic candidates. Journal of Near Infrared Spectroscopy, 25(6), 423–431. https://doi.org/10.1177/0967033517741133 Slavchev, A., Kovacs, Z., Koshiba, H., Nagai, A., Bázár, G., Krastanov, A., Kubota, Y., & Tsenkova, R. (2015). Monitoring of Water Spectral Pattern Reveals Differences in Probiotics Growth When Used for Rapid Bacteria Selection. PLOS ONE, 10(7), e0130698. https://doi.org/10.1371/journal.pone.0130698 Spink, J., & Moyer, D. C. (2011). Defining the Public Health Threat of Food Fraud. Journal of Food Science, 76(9), R157–R163. https://doi.org/10.1111/j.1750-3841.2011.02417.x Sung-Lin Chen. (2009). A Miniature RFID Tag Antenna Design for Metallic Objects Application. IEEE Antennas and Wireless Propagation Letters, 8(c), 1043–1045. https://doi.org/10.1109/LAWP.2009.2032252 Tian, Y., Gao, X., Qi, W.-L., Wang, Y., Wang, X., Zhou, J., Lu, D., & Chen, B. (2021). Advances in differentiation and identification of foodborne bacteria using near infrared spectroscopy. Analytical Methods, 13(23), 2558–2566. https://doi.org/10.1039/D1AY00124H Tiwari, I., Singh, M., Pandey, C. M., & Sumana, G. (2015). Electrochemical genosensor based on graphene oxide modified iron oxide–chitosan hybrid nanocomposite for pathogen detection. Sensors and Actuators B: Chemical, 206, 276–283. https://doi.org/10.1016/j.snb.2014.09.056 Treguier, S., Couderc, C., Tormo, H., Kleiber, D., & Levasseur-Garcia, C. (2019a). Identification of lactic acid bacteria Enterococcus and Lactococcus by near-infrared spectroscopy and multivariate classification. Journal of Microbiological Methods, 165(June), 105693. https://doi.org/10.1016/j.mimet.2019.105693 Treguier, S., Couderc, C., Tormo, H., Kleiber, D., & Levasseur-Garcia, C. (2019b). Identification of lactic acid bacteria Enterococcus and Lactococcus by near-infrared spectroscopy and multivariate classification. Journal of Microbiological Methods, 165(June), 105693. https://doi.org/10.1016/j.mimet.2019.105693 Tuyls, P., & Batina, L. (n.d.). RFID-Tags for Anti-Counterfeiting. In Cryptographers’ Track at the RSA Conference, 115–131. Vaisocherová-Lísalová, H., Víšová, I., Ermini, M. L., Špringer, T., Song, X. C., Mrázek, J., Lamačová, J., Scott Lynn, N., Šedivák, P., & Homola, J. (2016a). Low-fouling surface plasmon resonance biosensor for multi-step detection of foodborne bacterial pathogens in complex food samples. Biosensors and Bioelectronics, 80, 84–90. https://doi.org/10.1016/j.bios.2016.01.040 Vaisocherová-Lísalová, H., Víšová, I., Ermini, M. L., Špringer, T., Song, X. C., Mrázek, J., Lamačová, J., Scott Lynn, N., Šedivák, P., & Homola, J. (2016b). Low-fouling surface plasmon resonance biosensor for multi-step detection of foodborne bacterial pathogens in complex food samples. Biosensors and Bioelectronics, 80, 84–90. https://doi.org/10.1016/j.bios.2016.01.040 25 Vetrone, S. A., Huarng, M. C., & Alocilja, E. C. (2012). Detection of Non-PCR amplified S. enteritidis genomic DNA from food matrices using a gold-nanoparticle DNA biosensor: A proof-of-concept study. Sensors (Switzerland), 12(8), 10487–10499. https://doi.org/10.3390/s120810487 Wang, C., Wang, J., Li, M., Qu, X., Zhang, K., Rong, Z., Xiao, R., & Wang, S. (2016). A rapid SERS method for label-free bacteria detection using polyethylenimine-modified Au-coated magnetic microspheres and Au@Ag nanoparticles. Analyst, 141(22), 6226–6238. https://doi.org/10.1039/c6an01105e Wang, L., Liu, Z., Xia, X., Yang, C., Huang, J., & Wan, S. (2017). Colorimetric detection of Cucumber green mottle mosaic virus using unmodified gold nanoparticles as colorimetric probes. Journal of Virological Methods, 243, 113–119. https://doi.org/10.1016/j.jviromet.2017.01.010 Wang, Y., & Salazar, J. K. (2016). Culture-Independent Rapid Detection Methods for Bacterial Pathogens and Toxins in Food Matrices. Comprehensive Reviews in Food Science and Food Safety, 15(1), 183– 205. https://doi.org/10.1111/1541-4337.12175 Wang, Z., Cai, R., Gao, Z., Yuan, Y., & Yue, T. (2020). Immunomagnetic separation: An effective pretreatment technology for isolation and enrichment in food microorganisms detection. Comprehensive Reviews in Food Science and Food Safety, 19(6), 3802–3824. https://doi.org/10.1111/1541-4337.12656 Wang, Z., Wang, D., Chen, J., Sela, D. A., & Nugen, S. R. (2016). Development of a novel bacteriophage based biomagnetic separation method as an aid for sensitive detection of viable Escherichia coli. The Analyst, 141(3), 1009–1016. https://doi.org/10.1039/C5AN01769F Wang, Z., Wang, D., Kinchla, A. J., Sela, D. A., & Nugen, S. R. (2016). Rapid screening of waterborne pathogens using phage-mediated separation coupled with real-time PCR detection. Analytical and Bioanalytical Chemistry, 408(15), 4169–4178. https://doi.org/10.1007/s00216-016-9511-2 Waswa, J., Irudayaraj, J., & DebRoy, C. (2007). Direct detection of E. Coli O157:H7 in selected food systems by a surface plasmon resonance biosensor. LWT - Food Science and Technology, 40(2), 187–192. https://doi.org/10.1016/j.lwt.2005.11.001 Wu, R., Xing, X., Corredig, M., Meng, B., & Griffiths, M. W. (2019). Concentration of hepatitis A virus in milk using protamine-coated iron oxide (Fe3O4) magnetic nanoparticles. Food Microbiology, 84(May), 103236. https://doi.org/10.1016/j.fm.2019.05.020 Xia, F., Zuo, X., Yang, R., Xiao, Y., Kang, D., Vallee-Belisle, A., Gong, X., Yuen, J. D., Hsu, B. B. Y., Heeger, A. J., & Plaxco, K. W. (2010). Colorimetric detection of DNA, small molecules, proteins, and ions using unmodified gold nanoparticles and conjugated polyelectrolytes. Proceedings of the National Academy of Sciences, 107(24), 10837–10841. https://doi.org/10.1073/pnas.1005632107 Xu, M., Wang, R., & Li, Y. (2016). Rapid detection of Escherichia coli O157:H7 and Salmonella Typhimurium in foods using an electrochemical immunosensor based on screen-printed interdigitated microelectrode and immunomagnetic separation. Talanta, 148, 200–208. https://doi.org/10.1016/j.talanta.2015.10.082 Yosief, H. O., Weiss, A. A., & Iyer, S. S. (2013). Capture of uropathogenic E. coli by using synthetic glycan ligands specific for the pap-pilus. Chembiochem : A European Journal of Chemical Biology, 14(2), 251–259. https://doi.org/10.1002/cbic.201200582 26 You, M., Lin, M., Wang, S., Wang, X., Zhang, G., Hong, Y., Dong, Y., Jin, G., & Xu, F. (2016). Three- dimensional quick response code based on inkjet printing of upconversion fluorescent nanoparticles for drug anti-counterfeiting. Nanoscale, 8(19), 10096–10104. https://doi.org/10.1039/C6NR01353H You, S.-M., Jeong, K.-B., Luo, K., Park, J.-S., Park, J.-W., & Kim, Y.-R. (2021). Paper-based colorimetric detection of pathogenic bacteria in food through magnetic separation and enzyme-mediated signal amplification on paper disc. Analytica Chimica Acta, 1151, 338252. https://doi.org/10.1016/j.aca.2021.338252 Zhang, Y., Xu, C., Guo, T., & Hong, L. (2018). An automated bacterial concentration and recovery system for pre-enrichment required in rapid Escherichia coli detection. Scientific Reports, 8(1), 17808. https://doi.org/10.1038/s41598-018-35970-8 Zhang, Z., Zhou, J., & Du, X. (2019). Electrochemical Biosensors for Detection of Foodborne Pathogens. Micromachines, 10(4), 222. https://doi.org/10.3390/mi10040222 Zhao, X., Lin, C.-W., Wang, J., & Oh, D. H. (2014). Advances in Rapid Detection Methods for Foodborne Pathogens. Journal of Microbiology and Biotechnology, 24(3), 297–312. https://doi.org/10.4014/jmb.1310.10013 Zhao, Y., Ye, M., Chao, Q., Jia, N., Ge, Y., & Shen, H. (2009). Simultaneous Detection of Multifood-Borne Pathogenic Bacteria Based on Functionalized Quantum Dots Coupled with Immunomagnetic Separation in Food Samples. Journal of Agricultural and Food Chemistry, 57(2), 517–524. https://doi.org/10.1021/jf802817y Zhu, M., Liu, W., Liu, H., Liao, Y., Wei, J., Zhou, X., & Xing, D. (2015). Construction of Fe 3 O 4 /Vancomycin/PEG Magnetic Nanocarrier for Highly Efficient Pathogen Enrichment and Gene Sensing. ACS Applied Materials & Interfaces, 7(23), 12873–12881. https://doi.org/10.1021/acsami.5b02374 27 Chapter 2. HYPOTHESES AND OBJECTIVES OF RESEARCH The research hypothesis, objectives, and novelty of this work are presented below. 2.1. Hypotheses This research had four primary hypotheses. Hypothesis 1: The glycan-coated magnetic nanoparticles (gMNP) can be used for rapid extraction and concentration of pathogens from complex food matrices. Following extraction, pathogen identification can be achieved using culture-based plating and culture independent amplification assays. Hypothesis 2: The pathogens extracted using glycan-coated magnetic nanoparticles (gMNP) can be screened using near infrared spectroscopy. Hypothesis 3: The pathogens extracted using glycan-coated magnetic nanoparticles (gMNP) can be detected using a dextrin-capped gold nanoparticle-based biosensor. Hypothesis 4: The magnetic nanoparticles can be combined with DNA to function as unique anticounterfeiting tags which can be detected using dextrin-capped gold nanoparticle-based biosensor. 2.2. Goals and objectives The primary goal of this work is to successfully concentrate and detect pathogens from complex food matrices using rapid and inexpensive methods by eliminating the need for sample pre-enrichment. Glycan coated MNPs (gMNP) are proposed as means to accomplish this owing to their low-cost synthesis, proven extraction and concentration abilities, and long shelf life at room temperature. Since food pathogens are effectively destroyed at high temperatures, the focus of this work is towards ready- to-eat foods and those that do not require heating at high temperatures for longer duration. The primary objectives and associated tasks are listed in Table 2.1. Objective 1: Extract and concentrate pathogens from food matrices using gMNP The gMNP were previously found to be effective towards pathogen capture from liquid foods, including milk and apple cider. However, their efficacy in other complex solid foods has not been well documented. The bacterial species considered in this study, because of their established association with foodborne outbreaks, include E. coli, S. enterica, and B. cereus. These bacteria were chosen based on their association with various outbreaks in the past. Objective 2: Classify pathogens concentrated using gMNP 28 Rapid detection of pathogens from complex foods is imperative following their effective concentration and extraction. Objective 2 elaborates the use of near infrared spectroscopy in conjunction with machine learning to function as an effective screening tool for confirming presence of pathogens following their gMNP-based capture. Objective 3: Detect gMNP-extracted pathogens with gold nanoparticle biosensor Methods that use DNA sequence-based detection confer specificity against non-target pathogens. Dextrin-capped gold nanoparticles (dGNP) offer a cost-effective solution for sequence-specific target detection and its application for detecting pathogens from food needs attention. Objective 3, therefore, focusses on the detection of E. coli and S. enterica following their extraction using gMNP. Objective 4: Develop a DNA-based anticounterfeiting tag to secure high-value food While DNA can provide for an excellent anti-counterfeiting tag, the primary challenges with such a system include its protection in harsh environments and rapid detection. Objective 4 focuses on the use of magnetic nanoparticles for granting protection to DNA based anti-counterfeiting tags following which dextrin-capped gold nanoparticles can be used for sequence identification. Pathogen concentration from food using gMNP Rapid classification of gMNP-extracted pathogens from food Detection of gMNP-extracted pathogens using dGNP biosensor Develop DNA- based anticounterfeiting tag to secure high- value food Figure 2. 1. Primary objectives of this study. 2.3. Novelty and significance of research Conventionally used magnetic nanoparticles for pathogen capture can be expensive owing to the use of specific recognition moieties or have multi-step synthesis procedures. The novelty of this work lies in the combination of inexpensive glycan-coated magnetic nanoparticles (gMNP) with rapid and affordable detection platforms. The synthesized gMNP have demonstrated a long shelf life and can be stored at room temperature, eliminating costs related to cold storage requirements of traditional MNPs. Near infrared spectroscopy (NIRS) and dextrin-capped gold nanoparticle-based assay are proposed as 29 screening and detection tools, respectively. While NIRS allows for non-invasive screening for the presence of target pathogen, dGNP grants rapid, sensitive, and specific pathogen detection. This simple yet affordable pathogen extraction and detection system allows for their application in low-resource settings. Various carbohydrate-coated magnetic nanoparticles-based pathogen extraction and detection platforms have demonstrated pathogen identification from foods. However, these innovations are either limited to low sample volumes, have equipment or skilled labor requirements, or are limited to simple food matrices. A wide acceptance of pathogen extraction and detection systems requires its applicability to large volumes of complex food matrices. It is also imperative that the system is applicable to solid food matrices since those are primarily associated with foodborne outbreaks. Further, foods that are traditionally consumed raw or require minimal processing before consumption are at an increased risk or harboring pathogens. This research therefore focused on the extraction and detection of E. coli from leafy greens, S. enterica from fresh produce, and B. cereus from ready-to-eat rice and macaroni. The first part of this study provided for a novel detection system where pathogens were extracted using gMNP following which detection was achieved using standard plating techniques and qPCR, which are traditionally used in food pathogen testing. The gMNP successfully demonstrated their concentration and extraction abilities in various food matrices. While detection was initially achieved using qPCR, a rapid and inexpensive method to screen for the presence of pathogens without the need for expensive reagents is needed. The second innovation in this work is the combination of near infrared spectroscopy with gMNP to screen for the presence of pathogens. Since inapplicability of NIRS to low bacterial counts is a primary obstacle, gMNP significantly improved the bacterial counts by concentrating the cells. 30 Table 2. 1. Primary objectives of this study. Objectives and related Tasks Limitations of current methods Approach of this study Sources Objective 1: Extract and concentrate pathogens from food using gMNP Confirming of binding between gMNP and bacterial cells in various matrices. Factors affecting binding between gMNP and bacterial cells. Use of amplification-based methods for pathogen detection. Objective 2: Classify pathogens using gMNP Capture of bacterial cells by gMNP in sterile PBS and liquid foods has been seen earlier. However, complex solid food matrices have rarely been used. The binding between gMNP and bacterial cells has previously been confirmed but influencing factors have not been explored. Successful removal of PCR inhibitors from food with gMNP has not been documented. Use transmission electron microscopy and confocal laser microscopy to confirm gMNP-bacteria binding in various foods. Analyzing the effect of environmental pH, surface charge of gMNP and bacterial cells, bacterial concentration, and food type. Detection of bacterial cells extracted using gMNP with PCR and qPCR. (Matta, 2018) (Matta & Alocilja, 2018) (Kim & Oh, 2021) Confirming spectroscopy-based detection for gMNP-extracted bacteria. Determining the limit of detection of NIRS-based screening platform. Screening foods using gMNP-NIRS platform. MNPs have been used for pathogen detection using spectroscopy, but FTIR was used. Current methods have not used a combination of gMNP with NIRS for bacterial detection. While NIRS has been used for screening foods, overnight growth was warranted. This can be overcome using gMNP. Use NIRS for classification of gMNP-extracted bacterial samples. Pre-processing, principal component analysis, supervised machine learning can be used. Serially diluted pure bacterial cultures can be detected using gMNP-NIRS system. Foodborne bacteria can be concentrated using gMNP following which NIRS can be used for screening. Objective 3: Detect gMNP-extracted pathogens with gold nanoparticle biosensor Optimize detectable color difference between target and non- target DNA. Determine sensitivity and specificity of the plasmonic biosensor. The GNPs have a remarkable potential for foodborne pathogen detection. However, PCR-amplification of long probe conjugation methods have often been warranted. Traditionally synthesized GNPs are highly sensitive. A few GNP colorimetric biosensors have been used for detection of target DNA extracted from foods but have needed PCR amplification. Employ the dGNP biosensor to detect pathogens from various foods artificially inoculated with target bacteria. Objective 4: Develop a DNA-based anticounterfeiting tag to secure high-value food Use magnetic nanoparticles for protection and extraction of DNA- based anticounterfeiting tags Use dGNP to rapids detect DNA in the anticounterfeiting tag Current methods use sequencing for detection which is both expensive and time-consuming Extraction of DNA from tags is time-consuming and requires dangerous chemicals Optimizing the changes in color of dGNP due to presence of target DNA can allow specific bacteria detection. Coating GNPs with dextrin can improve sensitivity of the plasmonic biosensor. Measuring absorbance or target and non-target bacteria at varying concentrations for specificity. Extract DNA of gMNP-extracted bacteria from various foods and detect using dGNP biosensor. Verify the detection with PCR assay. Coat MNPs with silica to confer a net positive charge. Form a DNA and silica-MNP colloid that offers protection and easy separation The dGNP biosensor developed earlier can be used for detection DNA from anticounterfeiting tags (Yu & Irudayaraj, 2006) (Tian et al., 2021) (Dester & Alocilja, 2022) (Quintela et al., 2015) (Baetsen- Young et al., 2018) (Quintela et al., 2019) (Puddu et al., 2014) (Paunescu et al., 2013) 31 Scheme 2. 1. The novelty of this study is in the combination of affordable magnetic extraction using gMNP with various detection platforms. Removal of inhibitors in food allows amplification-based detection and effective food screening. The dGNP-based biosensor allows sensitive, specific, and inexpensive pathogen detection. Supervised machine learning was then employed to screen for the presence of target pathogens. Notably, the presence of gMNP did not interfere with spectral acquisition. Sensitivity and specificity in detection assays is of prime importance and optimal solutions have often been provided with amplification-based assays, sequencing-based approaches, or spectrometry-based whole cell detection. However, these methods require a scientific skillset or are expensive for their application in areas with scarcity of resources. A third innovation in this work was provided with a gold nanoparticles-based biosensor which allowed for colorimetric plasmonic detection and did not employ expensive reagents. Since traditionally synthesized gold nanoparticles are sensitive to their surrounding environments, biocompatible dextrin-capped GNP (dGNP) were used. The assay makes use of target- specific single-stranded DNA probe conferring specificity in detection. Scheme 2.1 provides a rendition of the novelty of this study. 2.4. Experimental plan and methodology 2.4.1. Experimental design for evaluating the efficiency of gMNP for bacterial concentration The applicability of gMNP for bacterial extraction and its compatibility with current gold standards was initially investigated. The experimental design consists of primary research questions, dependent and independent variables, and methods for data analysis for hypothesis testing. The efficiency with which gMNP can extract and concentrate bacteria under optimal conditions was tested first with high bacterial 32 loads in small volumes. The efficiency of concentration can be quantified using capture capacity which relies on optical density measurements. For wide applicability of gMNP, their ability to concentrate and extract bacterial cells from large volumes is paramount, further, factors such as bacterial loads and environmental pH can influence their efficiency. This was tested with Concentration Factor as the dependent variable since optical density measurements are limited to high bacterial loads. Finally, the application of gMNP for bacterial extraction from various foods was tested followed by detection. Since plating requires overnight incubation, detection of gMNP-extracted bacteria with amplification-based qPCR was tested which can effectively shorten the detection time. Figure 2.2 depicts the experimental plan for objective 1. 2.4.1.1. Confirming the binding of gMNP with bacterial cells using microscopy Although plating can successfully confirm the capture and concentration of bacterial cells using gMNP, visualizing their interaction can help to further understand the underlying mechanism. The binding of gMNP with bacterial cells was confirmed using Transmission Electron Microscopy (TEM) and Confocal Laser Microscopy (CLM). Initial experiments were performed in small volumes of tryptic soy broth with cultures grown overnight. Briefly, 100 µL of gMNP (5mg/mL) was added to 1mL of bacterial culture and let stand for 5 min, following which separation was achieved by placing on a magnetic rack for 5 min. Imaging with TEM (JEM-1400 Flash, Jeol, Nieuw-Vennep, Tokio, Japan) was achieved at the Center for Advanced Microscopy (CAM) at MSU. The supplies for TEM (Glutaraldehyde, cacodylate buffer, and uranyl acetate stain) were provided by the CAM, MSU. The above mentioned gMNP-bacteria mixture was magnetically separated and resuspended in 100 µL of fixative solution with 2.5% glutaraldehyde in 0.1M cacodylate buffer. Next, 5 µL of the fixed gMNP-bacterial solution was dropped onto the dark side of copper grid for 20-30 sec followed by a wash step with 5 µL water. The grid was gently dried with an edge of a filter paper after which staining was done by dropping 5 µL of 0.1% uranyl acetate and standing for 5-10 sec following which excess stain was removed with a filter paper. The grid was placed in a sample holder and images were taken in the range of 5000X-25000X. For imaging with confocal laser microscope (Keyence VK series), standard Gram staining procedure was used. First, the gMNP-bacterial mixture was magnetically separated and resuspended in 100 µL of water and 5 µL was dropped onto a glass slide. The sample was heat-fixed and 25 µL of crystal violet was added for staining and allowed to stand for 1 min. Following a wash step with distilled water, 25 µL of Grams iodine was added to the sample on slide, let stand for 1 min, and washed with distilled water. The slide was washed with 70% ethanol for decolorization and 25 µL of safranin was added followed by washing 33 with distilled water and air drying. Images of samples were obtained at 100X magnification. Color-coded height maps of images and 3-Dimensional images were created using Keyence VK-X Multifile Analyzer software. 34 Figure 2. 2. Experimental design to evaluate the efficacy of gMNP for bacterial concentration. 35 2.4.1.2. Small volume gMNP-bacterial capture Proprietary chitosan-functionalized gMNP were synthesized at the Nano-Biosensors Laboratory, MSU. The capacity of bacterial capture in small volumes of Phosphate Buffered Saline (PBS, pH 7.4) was determined using optical density at 600 nm (OD). Fresh bacterial cultures grown for 4-6 h were used for measurements, capture capacity was calculated using OD of the initial sample (ODinitial) and OD of supernatant (ODsupernatant) using the equation below. Capture Capacity (%) = OD initial−OD supernatant OD initial X 100 (2.1) Since optical density measurements require high bacterial loads, this technique could only be applied to small volumes. For large volume capture, the Concentration Factor was used to determine the efficiency of gMNP capture. 2.4.1.3. Effect of surface charge and pH on capture efficiency of gMNP To determine the effect of surface charge on capture efficiency of gMNP, zeta potential measurements were taken of both bacterial cells and gMNP (Zen3600, Malvern, Worcestershire WR14 1XZ, UK). For surface charge measurements, bacterial cells were diluted for consistency of OD (~0.5) and resuspended in 1mL deionized water. The resuspended samples were loaded into folded capillary cuvettes and placed into the instrument for measurements (Scheme 2.2). Experiments were performed in triplicates. For gMNP suspended in deionized water, measurements were taken directly. To determine the effect of pH on zeta potential, cells were finally suspended in PBS with pH adjusted to 3, 4, 5, 6, 7, 8, and 9, confirmed using a pH meter (Fisher Scientific, Waltham MA). The effect of surface charge on capture efficiency was also determined using capture capacity measurements. Scheme 2. 2. Procedure for determining the surface charge of bacterial cells. To determine the effect of bacterial load on efficiency of gMNP, fresh bacterial cultures were serially diluted until 10-6 dilution, following which 100 µL of gMNP was added for extraction. Quantification of extraction was determined using the concentration factor, given by equation (2.2), since optical density measurements are limited to high bacterial loads. Colony counts in the range of 20-200 were used. 36 Concentration Factor (CF) = CFUs in gMNP treated sample CFUs in the control sample (2.2) 2.4.1.3. Large-volume gMNP-bacterial capture in buffer solution and foods To determine the effectiveness of gMNP for concentrating bacterial cells in large volumes of PBS, CF was again used. First, 1 mL of 104 CFU/mL of bacterial cells was added to 25 mL of PBS. Then 225 mL of PBS was added, resulting in a final bacterial concentration of ~102 CFU/mL. To the cell suspension, 1 mL of gMNP was added, followed by extraction, and plating to determine CF. Serial dilutions of 10-5 and 10-6 were plated for initial bacterial counts. For experiments with various food matrices, 25 grams of the food matrix in a Whirl-Pak bag was artificially contaminated with 1 mL of ~104 CFU/mL of fresh bacterial culture. After 1 hour of acclimation at room temperature, 225 mL of PBS was added, followed by homogenization in a stomacher for 2 minutes. The liquified food matrix samples were then separated into two Whirl-Pak bags with 100 mL each, with one bag serving as a control and the other to which gMNP was added. Following gMNP extraction, the samples were resuspended in 1 mL PBS and plated on plates selective for target bacteria. The plates were incubated at 37 °C for 24-48 hours. Negative control (uninoculated food sample) was also plated to account for natural microflora. As mentioned, extraction of target bacteria was quantified using CF. Scheme 2.3 briefly describes the procedure for bacterial extraction from large volumes. Scheme 2. 3. Procedure for extraction of bacterial cells from large volumes of PBS and foods. Efficacy of extraction can be quantified with CF. 37 2.4.2. Experimental design for evaluating the effectiveness of gMNP in conjunction with NIRS as a screening platform. The primary obstacles with NIRS for bacterial classification include requirements of high bacterial loads, interference from surrounding food matrix and ineffectiveness of classifying samples with multiple bacterial types. The concentration of bacterial cells with gMNP and their magnetic nature can address the minimum bacterial load constraint and minimize the effect of food debris. Objective 2 evaluates the application with gMNP-NIRS for pathogen classification. The experimental design consists of primary research questions, dependent and independent variables, and methods for data analysis for hypothesis testing and are summarized in Figure 2.3. Initial experiments were done with fresh bacterial culture with 108 CFU/mL. Pure cultures (1 mL) of B. cereus and E. coli were magnetically extracted with gMNP and resuspended in 1 mL water, followed by spectral acquisition in 800-2500 nm range. 2.4.2.1. Limit of detection of gMNP-extracted bacteria and classification from foods Detection of low bacterial loads is important for prevention of foodborne outbreaks. Among the primary tasks for objective 2 is testing the feasibility of gMNP-NIRS system in low bacterial loads. To establish the limit of detection, overnight cultures of B. cereus and E. coli were serially diluted until 102 CFU/mL and their NIR spectra were obtained. The total sample size consisted of 54 samples each for B. cereus and E. coli. Spectral acquisition was followed by data processing and pre-processing, outlined in section 2.4.2.2. 2.4.2.2. Data analysis The acquired spectrum was processed in MATLAB and pre-processing techniques included data centering, normalization, and obtaining a second derivative. Singular Value Decomposition was then implemented to obtain principal components. The obtained principal components were then tested with three supervised machine learning algorithms – Naïve Bayes, Artificial Neural Networks, and Support Vector Machine. 2.4.3. Experimental design for the development of a plasmonic biosensor A colorimetric biosensor was designed for the detection of target bacteria, utilizing the surface plasmon resonance of dextrin-capped gold nanoparticles. The biosensor procedure developed at the Nano- Biosensors Lab was employed in this study for the detection of E. coli and S. enterica, experimental design is outlines in Figure 2.5. Prior to testing in the biosensor, extracted DNA was confirmed for its quality with absorbance ratios A260/280 and A260/230 between 1.8 and 2.2. Briefly, 10 µL of the extracted DNA was mixed with 5 µL of target specific single-stranded aminated probe and 5 µL of surface modified 38 dextrin-capped gold nanoparticles (dGNP). The dGNP were surface modified with mercaptoundecanoic acid that provided -COOH groups which can form non-covalent bonds with amine groups on the probes. This leads to almost instantaneous dGNP-probe functionalization. The samples were then placed in a thermocycler which was set at 95 °C for 5 min for denaturation and 55 °C for 10 min for annealing. Next, HCl is added to initiate aggregation of dGNP and a color change. Hybridization of the target DNA with the probe results in the protection of the dGNP under acidic environment. In the presence of a non-target DNA, however, no hybridization takes place and dGNP are readily aggregated resulting in a color change to blue or purple. Stable dGNP in 30-50 nm diameter range are wine-red in color and display a peak at 520 nm. Aggregation of dGNP results in a color change to blue or purple, and a shift in the wavelength of absorbance maxima is seen, farther away from 520 nm. A proof-of-concept of the biosensor was initially developed for the detection of bacteria. The concentration of HCl (0.1 M – 0.5 M), its volume between 5-10 µL, and response time were optimized. The experimental set up consisted of a target and non-target samples at 40 ng/ µL, and a negative control with nuclease-free water. Readings of absorbance spectra were taken following application of HCl and a shift in peak wavelength were statistically optimized. The optimized volume of HCl was then used to test analytical sensitivity and specificity of the biosensor. The DNA from S. enterica was used as a non- target for E. coli biosensor, and from E. coli was used as a non-target for S. enterica biosensor. 39 Figure 2. 3. Experimental design for evaluating the efficacy of gMNP and near infrared spectroscopy for pathogen screening. 40 Figure 2. 4. Basic procedure for dGNP biosensor, adapted from (Dester and Alocilja., 2022). 2.4.3.1. Analytical sensitivity and specificity A separate set of trials were run to determine the sensitivity of the dGNP biosensor to minimum detectable concentration of DNA. The target DNA from E. coli C3000 and a non-target DNA from S. enterica at concentrations of 20 ng/ µL, 10 ng/ µL, 5 ng/ µL, 2.5 ng/ µL, and 1 ng/ µL in a series of 9 trials were used. As done previously, shift in wavelength of absorbance maxima was used to quantify the aggregation of dGNP. Since a shift away from 520 nm is seen as a result of dGNP aggregation, a ratio of absorbance at 520 nm to 625 nm was used as a metric to quantify the aggregation. The specificity of the biosensor was tested in the presence of various non-target bacteria (Table 2.2). Experiments were conducted in a series of 9 separate trials at a DNA concentration of 40 ng/ µL. The readings of absorbance and tube images were taken following previously optimized conditions. The shift in wavelength was optimized at 95% confidence interval. Table 2. 2. Selection of non-target bacteria for E. coli and S. enterica biosensor. Non-targets for E. coli Biosensor Non-targets for S. enterica Biosensor S. enterica E. coli C3000 K. pneumoniae K. pneumoniae E. cloacae E. cloacae B. cereus L. monocytogenes 41 Figure 2. 5. A dGNP-based biosensor for detection of gMNP-extracted E. coli and S. enterica from foods. 42 2.4.3.2. Target pathogen detection from food matrices The dGNP biosensor was employed for detection of E. coli and S. enterica from various food matrices. The procedure mentioned in section 2.4.1.4 was used for extraction of pathogen from various foods. A short enrichment of 4-6 h at 37 °C was followed by DNA extraction and detection using dGNP biosensor. A series of 9 trials were run for each replica from food. The food experiments were run in triplicates for each food matric. Each food trial consisted of a target, non-target, and control, to account for natural microbiota, DNA was extracted from all samples for analysis. The results were statistically analyzed at a 95% confidence interval for a shift in peak of absorbance maxima (A520/625) mentioned earlier. The samples were also PCR amplified using primers targeting the same gene for which the probe was designed. The amplified product was run on a 2% agarose gel at an applied voltage of 140 mV. 2.4.4. Experimental design for DNA-based anticounterfeiting tags To address the issues associated with current DNA-based anticounterfeiting platforms, such as DNA extraction, stability, protection, and its rapid detection, a novel platform is proposed. Magnetic nanoparticles were coated with silica which confer a positive charge, allowing them to bind with negatively-charged DNA tags. 2.4.4.1. Silica-coated magnetic nanoparticles for DNA protection and extraction Silica-coated magnetic nanoparticles were synthesized using an alkaline synthesis method, the nanoparticles were aminated to confer a positive charge. The synthesized nanoparticles were characterized with TEM and FTIR was used to confirm the presence of silica and amine groups. This allowed for their instantaneous conjugation with DNA tags. Following conjugation, the effect of factors influencing stability of DNA tags were tested including effect of temperature (60 °C, 90 °C, and 120 °C), UV light, and DNase. 2.4.4.2. Detection of DNA-based anticounterfeiting tags A primary issue with the use of DNA for anti-counterfeiting is their rapid detection as sequencing-based techniques have often been employed adding to cost and complexity. The dGNP biosensor developed earlier was applied for target DNA detection. The presence of silica-coated magnetic nanoparticles did not prevent DNA detection using dGNP biosensor. 43 REFERENCES Baetsen-Young, A. M., Vasher, M., Matta, L. L., Colgan, P., Alocilja, E. C., & Day, B. (2018). Direct colorimetric detection of unamplified pathogen DNA by dextrin-capped gold nanoparticles. Biosensors and Bioelectronics, 101(August 2017), 29–36. https://doi.org/10.1016/j.bios.2017.10.011 Dester, E., & Alocilja, E. (2022). Current Methods for Extraction and Concentration of Foodborne Bacteria with Glycan-Coated Magnetic Nanoparticles: A Review. Biosensors, 12(2), 112. https://doi.org/10.3390/bios12020112 Kim, J.-H., & Oh, S.-W. (2021). Pretreatment methods for nucleic acid-based rapid detection of pathogens in food: A review. Food Control, 121, 107575. https://doi.org/10.1016/j.foodcont.2020.107575 Matta, L. L. (2018). BIOSENSING TOTAL BACTERIAL LOAD IN LIQUID MATRICES TO IMPROVE FOOD SUPPLY CHAIN SAFETY USING CARBOHYDRATE-FUNCTIONALIZED MAGNETIC NANOPARTICLES FOR CELL CAPTURE AND GOLD NANOPARTICLES FOR SIGNALING. Matta, L. L., & Alocilja, E. C. (2018). Carbohydrate Ligands on Magnetic Nanoparticles for Centrifuge-Free Extraction of Pathogenic Contaminants in Pasteurized Milk. Journal of Food Protection, 81(12), 1941–1949. https://doi.org/10.4315/0362-028X.JFP-18-040 Paunescu, D., Puddu, M., Soellner, J. O. B., Stoessel, P. R., & Grass, R. N. (2013). Reversible DNA encapsulation in silica to produce ROS-resistant and heat-resistant synthetic DNA “fossils.” Nature Protocols, 8(12), 2440–2448. https://doi.org/10.1038/nprot.2013.154 Puddu, M., Paunescu, D., Stark, W. J., & Grass, R. N. (2014). Magnetically Recoverable, Thermostable, Hydrophobic DNA/Silica Encapsulates and Their Application as Invisible Oil Tags. ACS Nano, 8(3), 2677–2685. https://doi.org/10.1021/nn4063853 Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2015). Simultaneous direct detection of Shiga-toxin producing Escherichia coli (STEC) strains by optical biosensing with oligonucleotide- functionalized gold nanoparticles. Nanoscale, 7(6), 2417–2426. https://doi.org/10.1039/C4NR05869K Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2019). Simultaneous Colorimetric Detection of a Variety of Salmonella spp. in Food and Environmental Samples by Optical Biosensing Using Oligonucleotide-Gold Nanoparticles. Frontiers in Microbiology, 10(MAY), 1–12. https://doi.org/10.3389/fmicb.2019.01138 Tian, Y., Gao, X., Qi, W.-L., Wang, Y., Wang, X., Zhou, J., Lu, D., & Chen, B. (2021). Advances in differentiation and identification of foodborne bacteria using near infrared spectroscopy. Analytical Methods, 13(23), 2558–2566. https://doi.org/10.1039/D1AY00124H Yu, C., & Irudayaraj, J. (2006). Identification of pathogenic bacteria in mixed cultures by FTIR spectroscopy. Transactions of the ASABE, 49(5), 1623–1632. https://doi.org/10.13031/2013.22021 44 Chapter 3. MAGNETIC EXTRACTION AND QPCR DETECTION OF LOW-LEVEL PATHOGENS FROM FOOD This chapter is under review by PLOS one 3.1. Introduction Foodborne infections lead to several hospitalizations across the globe affecting countless lives and may even result in fatalities. Various pathogens are associated with such infections contaminating food because of interaction with farm animals or workers, or due to unhygienic handling before consumption. Salmonella and E. coli are among the pathogens needing attention globally (Pires et al., 2011; Vital et al., 2017). Although these pathogens have commonly been associated with foods not customarily cooked by consumers, including fresh produce (Wadamori et al., 2017) and ready-to-eat meals (Mengistu & Tolera, 2020), their association with raw meat products is not infrequent (Moawad et al., 2017). Rapidly detecting these pathogens from foods can prevent or minimize associated health and economic consequences. However, overnight growth is required by most platforms to ensure enough bacterial count needed for detection. The concentration of pathogens from the food matrix using various techniques has often been proposed to eliminate long enrichment times (Boodoo et al., 2023; Dester & Alocilja, 2022a). Current methods for bacterial concentration include physical routes such as centrifugation and filtration, and other chemical or biological techniques (Dester & Alocilja, 2022a; Sarno et al., 2021; X. Wu et al., 2016). Recent innovations in physical methods have allowed concentration from sufficiently large volumes and automation of the process. For instance, cross-flow microfiltration was automated for the concentration of Salmonella with a 70% recovery rate following biochemical pretreatment (X. Li et al., 2013). In another study, combining large-volume filtration and rapid detection techniques decreased the sample-to-result time to <3 h (J. H. Kim & Oh, 2020). While simplicity makes these methods desirable, a series of concentration steps is usually necessary, potentially resulting in loss of bacterial recovery. Other limitations include clogging filters and capturing remnant debris, which can affect selective pathogen detection due to inhibitors in the food matrix (Vinayaka et al., 2019). Chemical or biological methods for concentration are based on adsorption and use naturally occurring interactions between affinity ligands and substrates on a solid support. Dielectrophoresis is a typical example and utilizes a nonuniform electrical field polarization to induce bacterial cell migration toward the electrodes (Sarno et al., 2021). While this technique is rapid, complex matrices present problems due to the high conductivity of food particles (Ogawa et al., 2021). Among other methods, metal 45 hydroxides using zirconium hydroxide and titanium hydroxide also successfully immobilized bacterial cells through a chelation process. However, centrifugation-based separation is still required, and the issue of charged food particles persists (Y. Wang & Salazar, 2016). Biochemical methods using MNPs coated with recognition ligands have also gained traction owing to their rapidity in bacterial concentration and simple magnetic separation. Nanomaterials can theoretically improve bacterial capture because of their large surface area-to-volume ratio for the attachment of ligands. Their small size assists in faster separation allowing multiple particles to penetrate matrix interstices and interact with bacterial cells, increasing their capture efficiency (Dester & Alocilja, 2022b; Suh et al., 2013). Some common recognition moieties on MNP include antibodies, phages, carbohydrates, and antibiotics, among which antibodies have widely been applied to foods. Immunomagnetic separation (IMS) using antibodies has gained wide acceptance owing to its large-scale production and specificity of bacterial capture. For example, polyclonal antibodies specific to outer membrane protein A of Cronobacter sakazakii were used to shorten the pre-enrichment time for its PCR-based detection from powdered infant formula (Q. Chen et al., 2017). Multiple biosensor platforms have also been combined with IMS, including electrochemical (F. Huang et al., 2021), surface plasmon resonance-based (Bhandari et al., 2019), and colorimetric (Quintela et al., 2019), among many others. Further, IMS is effective in the separation of pathogens from various complex foods, including milk, cheese (Zeinhom, Wang, Sheng, et al., 2018), yogurt, egg (Zeinhom, Wang, Song, et al., 2018), chicken (Quintela et al., 2019), and ground beef (Quintela et al., 2015). Although popular, IMS has several limitations, including blocking the antibody by food debris and reducing bacterial capture, especially in fat-rich foods (T.-H. Kim et al., 2014; Z. Wang et al., 2020). Further, the production of antibodies can be expensive with cold storage requirements (Klutz et al., 2016), affordable alternatives without specific storage requirements are therefore needed. Additionally, the use of antibodies can result in carryover of PCR-inhibiting substances presenting difficulties in rapid bacterial detection (Rådström et al., 2008). Magnetic nanoparticles modified with surface coatings other than antibodies have demonstrated bacterial capture from foods but need improvement. For example, amine-functionalized MNPs were used to extract multiple bacterial species from various liquid foods, including groundwater, grape juice, and green tea (Y.-F. Huang et al., 2010a), but functionalization procedures spanned several hours. In another study, protamine-coated MNPs successfully captured Hepatitis A virus from milk samples, although a low capture efficiency and multi-step MNP preparation were recorded (R. Wu et al., 2019). Other antibiotic-based coatings, such as vancomycin, have concerns with their bactericidal properties 46 (Dwivedi & Jaykus, 2011). Carbohydrates, alternatively, are another class of surface coatings with the advantages of simple synthesis, stability at room temperature, and cost effectiveness. Carbohydrate coating is usually achieved through ligand adsorption onto the MNP surface by covalent linking or non-covalent functionalization (Dester & Alocilja, 2022b; Fratila et al., 2016). Mannose, galactose, and cysteine-glycan are a few examples of MNP coatings (El-Boubbou et al., 2007; Matta & Alocilja, 2018; Yosief et al., 2013). Among these, glycan (chitosan)-coated MNPs (gMNP) have recently gained attention for their stability at room temperature, one-pot synthesis, scalable production, rapid capture, and biosensor compatibility (Dester & Alocilja, 2022b). Interactions of glycan-coated MNPs with bacterial cell walls mainly include van der Waals interactions and hydrogen bonds between hydroxyl and amino groups on carbohydrate and microbial protein surfaces (Dester & Alocilja, 2022a). This interaction has allowed gMNP to concentrate and extract pathogens from various foods successfully. For instance, S. enterica, E. coli O157, L. monocytogenes, and B. cereus were all successfully extracted from different milk types (Matta, Harrison, et al., 2018; Matta & Alocilja, 2018). Successful extraction from other complex food matrices, including homogenized egg, apple cider, leafy greens, and processed foods, has also been recorded (Boodoo et al., 2023; Matta, Harrison, et al., 2018; Sharief et al., 2023a). However, the application of gMNP in different complex solid food matrices and their compatibility with amplification-based detection methods needs more attention. 3.1.1. Platform novelty and applicability The use of gMNP offered several advantages such as simple one-pot synthesis allowing large-scale production, stability at room temperature, and long shelf-life. In addition, the absence of specific recognition moieties and previously proven bacterial concentration from food make gMNP a lucrative alternative to ligand-based bacterial separation. In earlier studies, plating and biosensors were used for bacterial identification following magnetic extraction but have required pre-enrichment (Dester et al., 2022; Sharief et al., 2023a, 2023b). The novelty of this work is in the effective separation of low bacterial loads from large-volumes of complex matrices using gMNP and their compatibility with amplification- based detection platforms. Site-specific binding between gMNP-bacterial cells is further elucidated with confocal laser microscopy. Here, successful extraction of S. enterica and E. coli was achieved from melons, cucumbers, chicken, and lettuce, and qPCR was used to confirm bacterial presence. The use of qPCR, in conjunction with gMNP-based separation, allowed for specific detection and elimination of the pre-enrichment step, reducing the total assay time to <4 h. This study serves as a proof-of-concept for 47 rapid gMNP-based extraction and qPCR detection of low-level pathogen contamination in complex solid food matrices. Scheme 3.1 briefly describes the underlying concept of this study. Scheme 3. 1. The gMNP-based pathogen extraction and detection platform. 3.2. Materials and Methods 3.2.1. Materials Salmonella enterica serovar Enteritidis was obtained from the Nano-Biosensors Laboratory at Michigan State University (MSU), and E. coli C-3000 (ATCC 15597) from the American Type Culture Collection (ATCC). Magnetic nanoparticles functionalized with chitosan were used as prepared. Tryptic Soy Agar (TSA) and Tryptic Soy Broth (TSB), Hydrochloric Acid (ACS reagent, 37%), and Phosphate Buffered Saline (PBS, pH 7.4) were purchased from Sigma Aldrich. CHROMagar for S. enterica and E. coli was purchased from DRG International (Springfield, NJ), and Sodium Hydroxide pellets were obtained from VWR international. For Transmission Electron Microscopy (TEM) imaging, glutaraldehyde, uranyl acetate, and cacodylate buffer were provided by the MSU Center for Advanced Microscopy. Grids for TEM (formvar/carbon 200 mesh copper) were obtained from Electron Microscopy Systems (Hatfield, PA). For confocal laser imaging, Gram staining material (Gram Iodine, Gram's Safranin Solution, ethanol, and Crystal Violet) were purchased from VWR International (Radnor, PA). Food samples were all obtained from a local seller, and Whirl-Pak bags were purchased from VWR International. Racks for magnetic separation were purchased from Spherotech (Lake Forest, IL). DNA kits were purchased from Qiagen (Hilden, Germany), and SYBR green qPCR master mix was obtained from applied biosystems (Waltham, MA). Molecular biology-grade agarose and Tris Acetate EDTA (pH 8.3) were obtained from Thermo Scientific (Waltham, MA) 3.2.2. Synthesis and characterization of gMNP The synthesis of gMNP was achieved using a procedure published previously (Bhusal et al., 2018). Briefly, ferric chloride hexahydrate (FeCl3.6H2O) was used as a precursor and added to a mixture of 48 ethylene glycol and sodium acetate, which functioned as a reducing agent and porogen, respectively. Chitosan was polymerized on the nanoparticles resulting in iron (III) oxide core and glycan (chitosan) shell. The paramagnetic nature of nanoparticles was confirmed by subjecting 1 mL of 5mg/mL gMNP solution to an external magnet. The morphology of the synthesized gMNP was analyzed using TEM (JEM-1400 Flash, Jeol, Tokyo, Japan) with a LabC6 crystal electron source operated at 100 kV. Negative staining was used for visualizing the gMNP, 5 µL of solution was dropped on a copper grid and washed with distilled water following which 5 µL of 0.1% uranyl acetate solution was added prior to imaging. The mean size of gMNP was determined using a zetasizer (Malvern, Nano-ZS series) using 2 mL of sample. 3.2.3. Confirmation of bacterial extraction The capture capacity of gMNP was determined using optical density at 600 nm (OD) measured with NanoDrop OneC (Thermo Scientific, Madison, WI). Fresh cultures (1 mL) grown in TSB at 37 °C for 4-6 hours were diluted for consistency. Growth media was removed using centrifugation at 10,000 rpm for 3 minutes (Eppendorf AG, 22331 Hamburg, German), following which, the pellet was resuspended in 1 mL of sterile PBS. To the cell suspension, 100 µL of gMNP (5 mg/mL) was added, and binding was allowed for 5 min. Following 5 min of magnetic separation, OD600 was used to determine the bacterial concentration in the supernatant, and the capture capacity was calculated using equation (3.1) where OD initial represents the OD of the initial fresh culture sample, and OD supernatant represents the OD of the sample following magnetic separation. To determine the effect of pH on the capture capacity, cells were finally resuspended in PBS with pH adjusted to 3, 4, 5, 6, 7, 8, and 9, confirmed using a pH meter. Experiments were performed in triplicates. Capture Capacity (%) = OD initial−OD supernatant OD initial X 100 (3.1) The interaction of bacterial cells with gMNP was visualized using TEM (JEM-1400 Flash, Jeol, Tokyo, Japan) at the Center for Advanced Microscopy (CAM), MSU using a standard negative staining procedure. Overnight cultures were dissolved in 0.1 mL of the fixative solution (2.5% of glutaraldehyde in 0.1M cacodylate buffer), 5 µL of which was dropped onto the black side of the copper grid for 20-30 sec. The grid was then washed with 5 µL of distilled water and gently dried before the staining. Staining was achieved using 5 µL of 0.1 % uranyl acetate stain; excess stain was removed, following which images were taken in the range of 5000-25000 X magnification. Observation of cell morphology and interaction with gMNP was also achieved using 3D laser scanning confocal microscope (VK-X1000 Series, Keyence, Osaka, Japan). For measurements, 5 µL of the 49 magnetically extracted sample was placed on a microscope slide following Gram staining. Optical images of cells were taken from multiple spots on the slide under 100 X magnification. Morphological characterization of acquired images was achieved using VK- image analyzer software. 3.2.4. Zeta potential measurements and the effect of pH The surface charge of bacterial cells and gMNP was measured in terms of zeta potential using Zetasizer (Zen3600, Malvern, Worcestershire WR14 1XZ, UK). Bacterial cells (1mL, OD ~0.5) were centrifuged at 8000 RPM for 5 min and resuspended in 1mL deionized water. The resuspended samples were loaded into folded capillary cuvettes and placed into the instrument for measurements. Experiments were performed in triplicates. For gMNP suspended in deionized water, measurements were taken directly. To determine the effect of pH on zeta potential, cells were finally suspended in water with pH adjusted to 3, 4, 5, 6, 7, 8, and 9, confirmed using a pH meter (Fisher Scientific, Waltham MA). 3.2.5 Bacterial extraction at the low concentrations The plating technique was employed to determine the effect of bacterial concentration on gMNP-based extraction. Initially, fresh bacterial cultures were serially diluted until 102 CFU/mL, following which 100 µL of gMNP was added for extraction. Quantification of extraction was determined using the Concentration Factor, given by equation (3.2). Colony counts in the range of 20-200 were used. Concentration Factor (CF) = CFUs in gMNP treated sample CFUs in the control sample (3.2) The Concentration Factor was also used to quantify extraction from large-volume samples. First, 1 mL of 104 CFU/mL of bacterial cells was added to 25 mL of PBS. Then 225 mL of PBS was added, resulting in a final bacterial concentration of ~102 CFU/mL. To the cell suspension, 1 mL of gMNP was added, followed by extraction and plating to determine CF. Serial dilutions of 10-5 and 10-6 were plated for initial bacterial counts. 3.2.5. Bacterial extraction from foods Food matrices for this study were selected based on past foodborne outbreaks. The matrices chosen include melons, romaine lettuce, cucumber, and raw chicken. The procedure of artificial contamination was adapted from the Bacteriological Analytical Manual (BAM). First, 25 grams of the food matrix in a Whirl-Pak bag was artificially contaminated with 1 mL of ~104 CFU/mL of fresh bacterial culture. After 1 hour of acclimation at room temperature, 225 mL of PBS was added, followed by homogenization 50 in a stomacher for 2 minutes. The liquified food matrix samples were then separated into two Whirl-Pak bags with 100 mL each, with one bag serving as a control and the other to which gMNP was added. Following gMNP extraction, the supernatant was discarded and separated gMNP/bacterial cells were resuspended in 1 mL PBS and plated on CHROMagar (DRG, Springfield, NJ). Negative control (uninoculated food sample) was plated on selective media and TSA to account for natural microflora. The plates were incubated at 37 °C for 24-48 hours. Extraction of target bacteria was quantified using CF. Successful binding and extraction of S. enterica and E. coli were also confirmed using qPCR (QuantStudio 5, Applied Biosystems, Waltham, MA). Following magnetic extraction, DNA from captured cells was extracted following manufacturers instructions and used with qPCR. DNA quality was confirmed, and samples with acceptable A260/A280, and A260/A230, between 1.8 and 2.2, were used. The sequence of the primers used is given in Table 1. The protocol for qPCR consisted of an initial denaturation step for 5 min at 95 °C, and 30 cycles of 1 min at 95 °C, 1 min at 56 °C, and 30 sec at 72 °C. A final extension step consisted of 7 min at 72 °C. The PCR products were also analyzed by running on a 2 % agarose gel in Tris Acetate EDTA buffer (TAE, pH 8.3) for 50 min at 120 V. Table 3. 1. Primer sequence used for qPCR. Gene Sequence (5’-3’) uidA F- GCAGTCTTACTTCCATGATTTCTTTA R- TAATGCGAGGTACGGTAGG invA F- CGGTGGTTTTAAGCGTACTCTT R- CGAATATGCTCCACAAGGTTA Reference (Srinivasan et al., 2011) (Paião et al., 2013) 3.3. Results and Discussion 3.3.1. Characterization of gMNP The synthesized gMNP were initially characterized using TEM, the micrograph in Fig 3.1 A shows multiple gMNP of varying size, all particles were roughly spherical and clumping of particles was commonly observed. The paramagnetic nature of the nanoparticles was confirmed by dissolving gMNP in ultrapure water and then subjecting the solution to an external magnet. The particles readily separated from the solvent resulting in clear water, pictured in the inset in Figure 3.1 B. The paramagnetic properties of synthesized gMNP were confirmed in an earlier study which recorded their magnetization curves. The gMNP were found to be strongly magnetized in the presence of an external 51 magnet and lost their magnetization in its absence (Matta & Alocilja, 2018). The mean particle size was also determined using a zetasizer and was found to be approximately 480 nm. Figure 3.1 B shows the size-based distribution of particles with diameter on the x-axis (nm) and intensity (%) on the y-axis. While the size obtained from zetasizer was larger than commonly observed in TEM micrographs, the random clumping of the nanoparticles could possibly have resulted in this variation. The gMNP were found to be stable for over three years and their ability to extract and concentrate bacterial cells did not diminish. Figure 3. 1. (A) TEM micrograph of synthesized gMNP and (B) results from Zetasizer to characterize particle size. 3.3.2. Capture capacity of gMNP The application of gMNP to extract E. coli and S. enterica was initially assessed using capture capacity, which quantifies the efficiency of extraction. The capture capacity was determined in PBS for both bacterial types, and the results are shown in Figure 3.2 A. Approximately 80% of E. coli cells were successfully captured, although a lower capture of 20% was observed for S. enterica. This difference in capture capacities has also been observed with other MNP types in previous studies, such as polyethyleneimine-coated and amine-functionalized MNPs (B. Chen et al., 2019; X. Huang & El-Sayed, 2010). Authors of previous studies hypothesized that this contrast in capture may have resulted from a variation in electrostatic interaction between MNPs and bacterial cell walls (Y.-F. Huang et al., 2010b). The chemical nature of the cell wall can lead to a difference in its zeta potential, a proxy for the surface charge, depending on the presence of hydrophilic or hydrophobic groups. For example, in this study, E. coli displayed a net surface charge of -54 mV, while S. enterica displayed -8 mV. However, earlier reports have also suggested the influence of pH on zeta potential, which can influence MNP-bacteria interaction 52 (B. Chen et al., 2019). Environmental pH can result in the dissociation or protonation of acidic and basic groups on the cell wall, affecting surface zeta potential (Halder et al., 2015).Therefore, the effect of pH on capture capacity and zeta potential was also determined (Figure 3.2 B and C). The capture capacity was highest at a pH of 3 for both bacterial types, although S. enterica did not show a linear trend in the pH 4-9 range (r2 of ~0.38). Contrarily, the capture capacity of E. coli followed a roughly linear trend (r2 of ~0.88) with increasing environmental pH. Since most bacterial cells have a net negative surface charge, positively-charged MNPs have widely been used for their capture. In an earlier study, for instance, positively charged MNPs successfully captured pathogens in varying pH environments. The MNPs showed a positive charge until a pH of 8, after which a net-negative surface charge was exhibited, expectedly resulting in a decline of bacterial capture (Y.-F. Huang et al., 2010b). In another study, high capture of P. aeruginosa using aminated MNPs in the pH range of 5-8 was recorded, and MNPs displayed a net positive surface charge. It was found that at pH higher than 9, the aminated MNPs lost their positive surface charge, resulting in a decline in capture efficiency (Z. Li et al., 2019). In contrast to other MNP types, our gMNP displayed a net negative charge after a pH of 5 (Figure 3.2 C). However, it is important to note that a relatively significant capture was recorded at a pH of >5, although bacterial cells and gMNP displayed a net negative charge. We hypothesize that the presence of oppositely charged chemical groups on the surface of the bacterial cell wall and gMNP allowed targeted binding between the two. Microscopy was therefore used to elaborate on this hypothesis. Figure 3. 2. Capture capacity of gMNP in PBS (A), the effect of pH on capture capacity (B), and zeta potential (C). 3.3.3. Visualization of gMNP binding The binding of gMNP to individual E. coli and S. enterica cells was initially confirmed using TEM (Figure 3.3 A and B). Roughly spherical gMNP with their small size allowed binding of multiple nanoparticles to individual cells, and their interaction with flagella was also observed. Further, binding of gMNP towards curved ends of bacterial cells was a frequent occurrence hinting at their site-specific attachment. 53 Figure 3. 3. TEM images of E. coli (A) and S. enterica (B) extracted from PBS. To further elucidate the site-specific binding of gMNP, captured cells were analyzed using confocal laser microscopy; sample images are shown in Figure 3.4. Optical, height profile and 3-dimensional images were used to analyze individual E. coli and S. enterica cells. Binding towards the edges was evident from 3-D (top right) and optical (top left) images with gMNP clusters and bacterial cells indicated by black and blue arrows, respectively. Height profiles showed their thickness to be >0.5 µm (bottom right), and surface profiles corroborated visual results (bottom left). The surface profile along the length of the bacterial cell is shown for E. coli and along the width for S. enterica. An analysis of individual cells (n=100) revealed site-specific binding in 90% of the population. Figure 3. 4. Optical micrograph (A, E), 3-D image (B, F), height profile (C, G), and surface profile (D, H) for E. coli (L) and S. enterica (R). Blue arrows show bacterial cells and black arrows indicate gMNP. Interaction between carbohydrates and the bacterial cell wall is an established phenomenon, and various surface proteins have been identified to which this attachment could be attributed. For example, the bacterial adhesin FimH, a two-domain protein, can recognize and bind to terminal 54 mannose residues by establishing a catch-bond (Sauer et al., 2016). Concanavalin A is another protein known to bind to sugar moieties in the lipopolysaccharide layer of various bacteria, such as Campylobacter jejuni (SAFINA et al., 2008). The binding to this protein has also been used for the detection of E. coli and S. aureus using electrochemical impedance spectroscopy (EIS)(Gamella et al., 2009; Yang et al., 2016). Other examples of bacterial binding to sugar moieties include H. pylori to fucose (Park et al., 2015), C-type Salmo Salar Lectin (SSL) to Pseudomonas fluorescens, and D-mannose to Pseudomonas aeruginosa (Mi et al., 2021). The site-specific, surface charge-independent attachment of bacterial cells to our gMNP indicates their possible applications for pathogen extraction from foods with varying pH environments. Since the minimum bacterial load in food for infection varies among bacteria, sampling large volumes is desirable. Therefore, the influence of bacterial load on gMNP extraction and their application for large-volume capture was studied. 3.3.4. Magnetic extraction at ultralow concentration and large-volume capture Detecting pathogens at a low concentration is important for outbreak prevention. Concentrating low bacterial loads from large volumes enables amplification-based pathogen identification. In this context, traditional ligand-based IMS focusses on selective bacterial capture, but their capture efficiency relies on high initial bacterial loads. For example, an earlier study noted that when bacterial loads of L. monocytogenes were ~109 CFU/mL, a recovery of 91% was seen but it dropped to 40% when the load reduced to 104 CFU/ml (Nexmann Jacobsen et al., 1997). Therefore, in the presence of low level of pathogens, cultural enrichment following IMS is still a necessity (J. H. Kim & Oh, 2021). The gMNP, on the contrary, allow general capture of bacterial loads following which specific qPCR-based detection can be achieved. To test the efficiency of capture at low bacterial loads, gMNP were used at varying bacterial concentrations ; results are shown in Figure 3.5 A. The CF of E. coli at lower concentrations was found to be higher, following a linear downward trend (r2 = 0.85). S. enterica, alternatively, did not show any correlation of CF with bacterial concentration and showed high CF at 102, 105, and 106 CFU/mL. Notably, negligible binding with gMNP was displayed at 103 CFU/mL, hinting at a dynamic intercellular interaction among the bacterial cells and formation of biofilms. A reduction in CF of E. coli at the same concentration was also noted, although not as low as for S. enterica. Biofilm formation by Salmonella is an established phenomenon often used by the pathogen for attachment to surfaces of fruits and vegetables (Amrutha et al., 2017). The formation of biofilms is further assisted by antibiotics (Penesyan et al., 2020) such as chitosan coated on gMNP. While not the focus of this work, variation in CF because of S. enterica concentration needs more attention. 55 The practical relevance of magnetic nanoparticles is with their application towards large sample volumes. While conventional IMS allows effective separation of target pathogens from competing microorganisms and can potentially remove interfering debris, Kim et al., have indicated its limitation to small sample volumes (J. H. Kim & Oh, 2021). Therefore, the gMNP were assessed for their extraction abilities from 100 mL samples, low concentrations (102 CFU/mL) of E. coli and S. enterica in PBS were initially used (Figure 3.5 B). Both the bacterial types were successfully concentrated, with an average CF of 5 for E. coli and 3 for S. enterica. Following bacterial concentration from large volumes of PBS, gMNP were applied to various foods. Concentration from large volumes confirmed the application of gMNP for affordable and accessible food screening with low-level contamination. Following concentration from PBS, the applicability of our gMNP in various foods was tested. Figure 3. 5. Effect of bacterial concentration on CF (A) and Concentration Factor from large volumes (B). 3.3.5. Concentration and extraction from foods followed by qPCR detection To confirm the binding of our gMNP to bacterial cells in foods- melons, cucumbers, lettuce, and raw chicken, confocal laser images were initially taken and are shown in Figure 3.6 A-D. The images confirmed that the binding of gMNP takes place in the presence of multiple food matrices with their varying physical microstructure and chemical environment. Notably, the images confirm site-specific binding of gMNP in all foods, similar to what was observed in section 3.1. To determine the effect of food matrices on the concentration abilities of gMNP, extraction from all four food types was quantified using CF. The selective plating technique was used to calculate CF; results are shown in Figure 3.6 E. Both bacterial types were successfully concentrated from all the food matrices with a CF >1, except for E. coli from melon samples. Although E. coli demonstrated higher concentration in PBS, S. enterica showed a comparatively higher CF in all foods except raw chicken. Among the foods, the highest CF of S. 56 enterica was obtained in melon samples at 5.8, followed by cucumbers at 3. For E. coli, however, the highest CF was obtained in raw chicken at 2.8, and CF from both lettuce and cucumbers was slightly higher than 1. The CF of both bacterial types from various food matrices was found to be variable, and this could be a result of multiple factors. First, the physical microstructure and chemical constituents of the tested foods significantly vary. For instance, both cucumbers and melons are water-rich, although the carbohydrate content of melons is relatively higher (Olawuyi & Lee, 2019; Perkins-Veazie et al., 2012). All three varieties of fresh produce also vary in their protein content. Chicken, alternatively, consists of significantly higher fat and protein amounts. While the chemical composition of tested foods varies considerably, their varying physical microstructure could also have affected the binding of gMNP to inoculated bacteria. Another potential barrier contributing to varying CF among foods is the presence of natural microbiota. While melon samples did not show the presence of significant natural microbiota, cucumbers, chicken, and lettuce were all naturally contaminated (Figure 3.6 E inset), and contamination levels were variable. Lettuce samples were contaminated with the highest amount of natural microbiota, followed by chicken and cucumbers. Notably, the CF of S. enterica was inversely related to contamination level in foods. However, the concentration of S. enterica by our gMNP in the presence of large amounts of commensal microorganisms was also seen. It is important to note that a generic strain of E. coli was used in this study, commonly found in natural microbiota, which could have affected its CF. However, an earlier study with E. coli O157:H7 also reported a similar CF (Boodoo et al., 2023). While the gMNP successfully concentrated bacterial cells from food, to bypass pre-enrichment, detection using sensitive techniques is necessary. The compatibility of gMNP extraction was therefore tested with qPCR-based detection. The magnetic capture of target bacteria was confirmed using qPCR and the Cq value of gMNP extracted test samples was compared with that of control (Figure 3.6 F). The Cq is inversely related to the amount of target gene present, and the ratio of Cq of test to control was used for comparing extraction from different foods. A Cq comparison value of <1, indicating successful concentration, was obtained for all the food types inoculated with S. enterica in the following ascending order- melons, cucumbers, lettuce, and chicken. Here, a lower comparison value denotes a higher bacterial concentration. For E. coli, however, a Cq comparison value of <1 was obtained only for chicken and lettuce samples. It is important to note that while a correlation of Cq comparison value with CF is expected, CF only accounts for viable 57 bacterial cells. Further, since the Cq value depends on the initial concentration of extracted DNA, possible cell loss during DNA extraction needs to be accounted. Figure 3. 6. Bacteria extracted from melon (A), cucumber (B), chicken (C), and lettuce (D) samples. Concentration Factor from various foods (E) and qPCR signal comparison for extracted bacteria from the same samples (F). Detection of gMNP-extracted bacterial cells using qPCR confirmed their compatibility with the current amplification-based gold standards. While the coupling of IMS with amplification-based detection has allowed remarkably specific detection, issues such as cost, and time for enrichment are yet to be addressed (J.-H. Kim & Oh, 2021; Wei et al., 2019). Further, the time required for conjugation of magnetic nanoparticles with antibodies can span several hours (Brandão et al., 2015). A simple alternative is provided in this study with the gMNP making use of its site-specific binding to bacterial cells and inexpensive synthesis offering affordable extraction. Bacterial extraction and concentration using our gMNP cost approximately $ 0.5 per sample, significantly lower compared to $ 10 per assay using a commercially available IMS platforms (Briceno et al., 2019; Lau et al., 2013). Combining the gMNP with qPCR allows for the specificity required in the detection of foodborne pathogens without enrichment. The food matrices used in this study were contaminated with varying amounts of natural microorganisms, but the gMNP-qPCR system was successful in rapid detection of S. enterica and E. coli. 58 The time required for pre-enrichment was eliminated and a low bacterial load of 102 CFU/ml was successfully detected in <4 h. The use of gMNP concentrated bacterial cells from complex food matrices and possibly provided for the minimum initial DNA required for amplification in qPCR (Scheme 3.2). This was particularly observed in samples with low bacterial loads such as melons and cucumbers. Notably, in multiple replicates from these foods, only gMNP-extracted test samples resulted in amplification curve while no curve was seen for control samples without gMNP extraction. To confirm this, the qPCR amplified samples were also run on a 2% agarose gel, results are also shown. As expected, only gMNP extracted samples gave a band while no band was seen for control samples without gMNP. Scheme 3. 2. Use of gMNP provides for higher bacterial concentration, allowing amplification-based detection. Amplifications products from melons and cucumber samples following qPCR. Negative control (NC), positive control (PC), gMNP-treated (T) and non-gMNP (C) samples are shown. While previous works have used MNPs devoid of antibodies, they were limited to low volumes or had complicated MNP synthesis protocols. Polyethyleneimine and gold-modified MNPs, for example, were used to capture E. coli and S. aureus from milk, but their sample volume was limited to 10 mL (C. Wang et al., 2016) In another study, E. coli O157 was extracted from sausage using Poly-L-lysine coated MNPs, but their preparation spanned several hours with multiple steps (You et al., 2021a). Rarely, ligand- lacking MNPs have also been combined with RT-PCR but afore mentioned obstacles were not addressed (R. Wu et al., 2019). The traditional IMS platforms, in contrast, have been either applied for small volume capture, or required enrichment for several hours. Table 3.2 shows various MNPs in the literature that have not used recognition biomolecules to extract pathogens from foods and how they compare with gMNP. Notably, previous works using gMNP have required a short enrichment of 4-7 h. As a comparison, a few examples from the traditional IMS-based methods are also presented. 59 Table 3. 2. Magnetic nanoparticles-based platforms for foodborne pathogen extraction. Surface Microorganism MNP Tested foods LVC Enrichment Method / Assay time Ref modification synthesis Iron/ Gold E. coli Multistep Tap water No No Raman (C. Wang et al., microspheres and milk spectroscopy/NA 2016) Protamine- Hepatitis A virus Multistep Milk No No qPCR /< 4 h (R. Wu et al., coated 2019) Lysine- Short E. coli: O157 Multistep Sausage No No Biosensor/ 2h (You et al., Chain Glucans 2021b) IMS S. Enteritidis Multistep Chicken No Yes ELISA />5h (Bai et al., Cabbage Tomato Apple juice 2023) IMS Salmonella sps. NA Chicken Yes Yes Biosensor /> 8 h (Quintela et al., IMS Salmonella Multistep IMS Salmonella Multistep E. coli O157:H7 L. monocytogenes Blueberries Milk Pork Meat 2019) No Yes qPCR / 10 h (J. Wang et al., 2018) Yes Yes qPCR/ > 6h (Fan et al., 2022) gMNP S. enterica One pot Milk Yes No Plating/> 9 h (Matta & E. coli B. cereus Alocilja, 2018) gMNP S. enterica One pot Homogenized No Yes Radio-frequency (Matta, E. coli egg, milk, apple cider biosensor/NA Karuppuswami, et al., 2018) gMNP E. coli O157 One pot Flour Yes Yes Colorimetric (Dester et al., biosensor/ > 6 h 2022) gMNP E. coli One pot Lettuce, Yes Yes Colorimetric (Sharief et al., spinach biosensor/ > 6 h 2023a) gMNP E. coli One pot Melons, Yes No Plating and qPCR / < 4 This study S. enterica cucumbers, chicken, lettuce h gMNP: Glycan-coated magnetic nanoparticles, NA: Not Available, IMS: Immunomagnetic separation, LVC: Large Volume Capture, ELISA: Enzyme-Linked Immunosorbent Assay A major obstacle with positively-charged MNPs for bacterial capture is that their electrostatic nature may favor the capture of charged food microparticles. Conversely, successful bacterial extraction using our gMNP was hypothesized to result from multiple forces, including long-range electrostatic attraction, 60 Brownian motion, and carbohydrate-protein binding (Dester & Alocilja, 2022a) Supporting this hypothesis, microscopy indicated specific binding and bacterial capture in high pH environments showed binding of bacterial cells with gMNP possessing a net-negative charge. This may be due to oppositely charged pockets on gMNPs. Allowing capture from fleshy and fat-rich foods with varying pH environments, possible interaction of our gMNP with charged food particles was not a hindrance for bacterial capture or qPCR detection. 3.4. Conclusion A rapid and cost-effective platform for bacterial extraction and direct confirmation using qPCR from large volumes of food samples is presented in this study. Glycan-coated MNPs synthesized using a simple one-pot procedure successfully extracted low concentrations (102 CFU/mL) of E. coli and S. enterica from melons, cucumbers, chicken, and lettuce. Both selective plating techniques and qPCR confirmed the extraction of target bacteria from complex food samples. Successful extraction was accomplished in the presence of natural microflora, and use of our gMNP did not inhibit amplification in qPCR. With simple, low-cost and quick extraction combined with rapid qPCR confirmation, this study demonstrates that the gMNP-qPCR system can be used for rapid determination of low pathogen contamination in complex matrices within about 4 hours without requiring an enrichment step. 61 REFERENCES Dester E, Alocilja E. Current Methods for Extraction and Concentration of Foodborne Bacteria with Glycan-Coated Magnetic Nanoparticles: A Review. Biosensors (Basel). 2022;12. doi:10.3390/bios12020112 Chen Q, Li Y, Tao T, Bie X, Lu F, Lu Z. Development, and application of a sensitive, rapid, and reliable immunomagnetic separation-PCR detection method for Cronobacter spp. J Dairy Sci. 2017;100: 961–969. doi:10.3168/jds.2016-11087 Huang, Xue L, Qi W, Cai G, Liu Y, Lin J. An ultrasensitive impedance biosensor for Salmonella detection based on rotating high gradient magnetic separation and cascade reaction signal amplification. Biosens Bioelectron. 2021;176: 112921. doi: 10.1016/j.bios.2020.112921 Bhandari D, Chen F-C, Bridgman RC. Detection of Salmonella Typhimurium in Romaine Lettuce Using a Surface Plasmon Resonance Biosensor. Biosensors (Basel). 2019;9: 94. doi:10.3390/bios9030094 Quintela IA, de los Reyes BG, Lin C-S, Wu VCH. Simultaneous Colorimetric Detection of a Variety of Salmonella spp. in Food and Environmental Samples by Optical Biosensing Using Oligonucleotide- Gold Nanoparticles. Front Microbiol. 2019;10: 1–12. doi:10.3389/fmicb.2019.01138 Zeinhom MMA, Wang Y, Sheng L, Du D, Li L, Zhu M-J, et al. Smart phone based immunosensor coupled with nanoflower signal amplification for rapid detection of Salmonella Enteritidis in milk, cheese and water. Sens Actuators B Chem. 2018;261: 75–82. doi: 10.1016/j.snb.2017.11.093 Zeinhom MMA, Wang Y, Song Y, Zhu M-J, Lin Y, Du D. A portable smart-phone device for rapid and sensitive detection of E. coli O157:H7 in Yoghurt and Egg. Biosens Bioelectron. 2018;99: 479–485. doi: 10.1016/j.bios.2017.08.002 Quintela IA, de los Reyes BG, Lin C-S, Wu VCH. Simultaneous direct detection of Shiga-toxin producing Escherichia coli (STEC) strains by optical biosensing with oligonucleotide-functionalized gold nanoparticles. Nanoscale. 2015;7: 2417–2426. doi:10.1039/C4NR05869K Wang Z, Cai R, Gao Z, Yuan Y, Yue T. Immunomagnetic separation: An effective pretreatment technology for isolation and enrichment in food microorganisms detection. Compr Rev Food Sci Food Saf. 2020;19: 3802–3824. doi:10.1111/1541-4337.12656 Kim T-H, Park J, Kim C-J, Cho Y-K. Fully Integrated Lab-on-a-Disc for Nucleic Acid Analysis of Food-Borne Pathogens. Anal Chem. 2014;86: 3841–3848. doi:10.1021/ac403971h Klutz S, Holtmann L, Lobedann M, Schembecker G. Cost evaluation of antibody production processes in different operation modes. Chem Eng Sci. 2016;141: 63–74. doi: 10.1016/j.ces.2015.10.029 Rådström P, Löfström C, Lövenklev M, Knutsson R, Wolffs P. Strategies for Overcoming PCR Inhibition. Cold Spring Harb Protoc. 2008;2008: pdb. top20. doi:10.1101/pdb. top20 Huang Y-F, Wang Y-F, Yan X-P. Amine-Functionalized Magnetic Nanoparticles for Rapid Capture and Removal of Bacterial Pathogens. Environ Sci Technol. 2010;44: 7908–7913. doi:10.1021/es102285n Wu R, Xing X, Corredig M, Meng B, Griffiths MW. Concentration of hepatitis A virus in milk using protamine-coated iron oxide (Fe3O4) magnetic nanoparticles. Food Microbiol. 2019;84: 103236. doi:10.1016/j.fm.2019.05.020 62 Dwivedi HP, Jaykus L-A. Detection of pathogens in foods: the current state-of-the-art and future directions. Crit Rev Microbiol. 2011;37: 40–63. doi:10.3109/1040841X.2010.506430 Fratila RM, Moros M, de la Fuente JM. Recent advances in biosensing using magnetic glyconanoparticles. Anal Bioanal Chem. 2016;408: 1783–1803. doi:10.1007/s00216-015-8953-2 El-Boubbou K, Gruden C, Huang X. Magnetic Glyco-nanoparticles: A Unique Tool for Rapid Pathogen Detection, Decontamination, and Strain Differentiation. J Am Chem Soc. 2007;129: 13392–13393. doi:10.1021/ja076086e Matta LL, Alocilja EC. Carbohydrate Ligands on Magnetic Nanoparticles for Centrifuge-Free Extraction of Pathogenic Contaminants in Pasteurized Milk. J Food Prot. 2018;81: 1941–1949. doi: 10.4315/0362-028X.JFP-18-040 Yosief HO, Weiss AA, Iyer SS. Capture of uropathogenic E. coli by using synthetic glycan ligands specific for the pap-pilus. Chembiochem. 2013;14: 251–9. doi:10.1002/cbic.201200582 Matta LL, Harrison J, Deol GS, Alocilja EC. Carbohydrate-Functionalized Nanobiosensor for Rapid Extraction of Pathogenic Bacteria Directly from Complex Liquids With Quick Detection Using Cyclic Voltammetry. IEEE Trans Nanotechnol. 2018;17: 1006–1013. doi:10.1109/TNANO.2018.2841320 Sharief SA, Caliskan-Aydogan O, Alocilja E. Carbohydrate-coated magnetic and gold nanoparticles for point-of-use food contamination testing. Biosens Bioelectron X. 2023;13: 100322. doi: 10.1016/j.biosx.2023.100322 Sharief SA, Caliskan-Aydogan O, Alocilja EC. Carbohydrate-coated nanoparticles for PCR-less genomic detection of Salmonella from fresh produce. Food Control. 2023;150: 109770. doi: 10.1016/j.foodcont.2023.109770 Dester E, Kao K, Alocilja EC. Detection of Unamplified E. coli O157 DNA Extracted from Large Food Samples Using a Gold Nanoparticle Colorimetric Biosensor. Biosensors (Basel). 2022;12: 274. doi:10.3390/bios12050274 Bhusal N, Shrestha S, Pote N, Alocilja E. Nanoparticle-Based Biosensing of Tuberculosis, an Affordable and Practical Alternative to Current Methods. Biosensors (Basel). 2018;9: 1. doi:10.3390/bios9010001 Srinivasan S, Aslan A, Xagoraraki I, Alocilja E, Rose JB. Escherichia coli, enterococci, and Bacteroides thetaiotaomicron qPCR signals through wastewater and septage treatment. Water Res. 2011;45: 2561–2572. doi: 10.1016/j.watres.2011.02.010 Paião FG, Arisitides LGA, Murate LS, Vilas-Bôas GT, Vilas-Boas LA, Shimokomaki M. Detection of Salmonella spp, Salmonella Enteritidis and Typhimurium in naturally infected broiler chickens by a multiplex PCR-based assay. 2013. Available: www.sbmicrobiologia.org.br Chen B, Xie H, Zhang A, Liu N, Li Q, Guo J, et al. Synthesis of PEI-Functionalized Magnetic Nanoparticles for Capturing Bacteria. Journal of Wuhan University of Technology-Mater Sci Ed. 2019;34: 236– 242. doi:10.1007/s11595-019-2041-y Huang X, El-Sayed MA. Gold nanoparticles: Optical properties and implementations in cancer diagnosis and photothermal therapy. J Adv Res. 2010;1: 13–28. doi: 10.1016/j.jare.2010.02.002 63 Huang Y-F, Wang Y-F, Yan X-P. Amine-Functionalized Magnetic Nanoparticles for Rapid Capture and Removal of Bacterial Pathogens. Environ Sci Technol. 2010;44: 7908–7913. doi:10.1021/es102285n Halder S, Yadav KK, Sarkar R, Mukherjee S, Saha P, Haldar S, et al. Alteration of Zeta potential and membrane permeability in bacteria: a study with cationic agents. Springerplus. 2015;4: 1–14. doi:10.1186/s40064-015-1476-7 Li Z, Ma J, Ruan J, Zhuang X. Using Positively Charged Magnetic Nanoparticles to Capture Bacteria at Ultralow Concentration. Nanoscale Res Lett. 2019;14: 195. doi:10.1186/s11671-019-3005-z Sauer MM, Jakob RP, Eras J, Baday S, Eriş D, Navarra G, et al. Catch-bond mechanism of the bacterial adhesin FimH. Nat Commun. 2016;7. doi:10.1038/ncomms10738 SAFINA G, VANLIER M, DANIELSSON B. Flow-injection assay of the pathogenic bacteria using lectin-based quartz crystal microbalance biosensor. Talanta. 2008;77: 468–472. doi: 10.1016/j.talanta.2008.03.033 Gamella M, Campuzano S, Parrado C, Reviejo AJ, Pingarrón JM. Microorganisms recognition and quantification by lectin adsorptive affinity impedance. Talanta. 2009;78: 1303–1309. doi:10.1016/j.talanta.2009.01.059 Yang H, Zhou H, Hao H, Gong Q, Nie K. Detection of Escherichia coli with a label-free impedimetric biosensor based on lectin functionalized mixed self-assembled monolayer. Sens Actuators B Chem. 2016;229: 297–304. doi:10.1016/j.snb.2015.08.034 Park S, Kim G-H, Park S-H, Pai J, Rathwell D, Park J-Y, et al. Probing Cell-Surface Carbohydrate Binding Proteins with Dual-Modal Glycan-Conjugated Nanoparticles. J Am Chem Soc. 2015;137: 5961– 5968. doi:10.1021/jacs.5b00592 Mi F, Guan M, Hu C, Peng F, Sun S, Wang X. Application of lectin-based biosensor technology in the detection of foodborne pathogenic bacteria: A review. Analyst. 2021;146: 429–443. doi:10.1039/d0an01459a Nexmann Jacobsen C, Fremming C, Jakobsen M. Immunomagnetic separation of Listeria monocytogenes for flow cytometric determination of viable cells in liquid. J Microbiol Methods. 1997;31: 75–81. doi:10.1016/S0167-7012(97)00084-5 Kim JH, Oh SW. Pretreatment methods for nucleic acid-based rapid detection of pathogens in food: A review. Food Control. Elsevier Ltd; 2021. doi: 10.1016/j.foodcont.2020.107575 Amrutha B, Sundar K, Shetty PH. Study on E. coli and Salmonella biofilms from fresh fruits and vegetables. J Food Sci Technol. 2017;54: 1091–1097. doi:10.1007/s13197-017-2555-2 Penesyan A, Paulsen IT, Gillings MR, Kjelleberg S, Manefield MJ. Secondary Effects of Antibiotics on Microbial Biofilms. Frontiers in Microbiology. Frontiers Media S.A.; 2020. doi:10.3389/fmicb.2020.02109 Olawuyi IF, Lee W. Influence of chitosan coating and packaging materials on the quality characteristics of fresh-cut cucumber. Korean Journal of Food Preservation. 2019;26: 371–380. doi:10.11002/kjfp.2019.26.4.371 64 Perkins-Veazie P, Beaulieu JC, Siddiq M. Watermelon, Cantaloupe and Honeydew. Tropical and Subtropical Fruits Postharvest Physiology, Processing and Packaging. 2012. Available: https://onlinelibrary.wiley.com/doi/pdf/10.1002/9781118324097#page=557 Kim J-H, Oh S-W. Pretreatment methods for nucleic acid-based rapid detection of pathogens in food: A review. Food Control. 2021;121: 107575. doi:10.1016/j.foodcont.2020.107575 Wei S, Park B-J, Kim S-H, Seo K-H, Jin Y-G, Oh D-H. Detection of Listeria monocytogenes using Dynabeads® anti-Listeria combined with real-time PCR in soybean sprouts. LWT. 2019;99: 533– 539. doi:10.1016/j.lwt.2018.10.023 Brandão D, Liébana S, Pividori MI. Multiplexed detection of foodborne pathogens based on magnetic particles. N Biotechnol. 2015;32: 511–520. doi:10.1016/j.nbt.2015.03.011 Lau HK, Clotilde LM, Lin AP, Hartman GL, Lauzon CR. Comparison of IMS Platforms for Detecting and Recovering Escherichia coli O157 and Shigella flexneri in Foods. SLAS Technol. 2013;18: 178–183. doi:10.1177/2211068212468583 Briceno RK, Sergent SR, Benites SM, Alocilja EC. Nanoparticle-Based Biosensing Assay for Universally Accessible Low-Cost TB Detection with Comparable Sensitivity as Culture. Diagnostics. 2019;9: 222. doi:10.3390/diagnostics9040222 Wang C, Wang J, Li M, Qu X, Zhang K, Rong Z, et al. A rapid SERS method for label-free bacteria detection using polyethylenimine-modified Au-coated magnetic microspheres and Au@Ag nanoparticles. Analyst. 2016;141: 6226–6238. doi:10.1039/c6an01105e You S-M, Jeong K-B, Luo K, Park J-S, Park J-W, Kim Y-R. Paper-based colorimetric detection of pathogenic bacteria in food through magnetic separation and enzyme-mediated signal amplification on paper disc. Anal Chim Acta. 2021;1151: 338252. doi:10.1016/j.aca.2021.338252 You S-M, Jeong K-B, Luo K, Park J-S, Park J-W, Kim Y-R. Paper-based colorimetric detection of pathogenic bacteria in food through magnetic separation and enzyme-mediated signal amplification on paper disc. Anal Chim Acta. 2021;1151: 338252. doi:10.1016/j.aca.2021.338252 Bai M, Wang Y, Zhang C, Wang Y, Wei J, Liao X, et al. Nanobody-based immunomagnetic separation platform for rapid isolation and detection of Salmonella enteritidis in food samples. Food Chem. 2023;424: 136416. doi:10.1016/j.foodchem.2023.136416 Wang J, Li Y, Chen J, Hua D, Li Y, Deng H, et al. Rapid detection of food-borne Salmonella contamination using IMBs-qPCR method based on pagC gene. Braz J Microbiol. 2018;49: 320–328. doi:10.1016/j.bjm.2017.09.001 Fan W, Gao X, Li H, Guo W, Li Y, Wang S. Rapid and simultaneous detection of Salmonella spp., Escherichia coli O157:H7, and Listeria monocytogenes in meat using multiplex immunomagnetic separation and multiplex real-time PCR. European Food Research and Technology. 2022;248: 869–879. doi:10.1007/s00217-021-03933-5 Matta LL, Karuppuswami S, Chahal P, Alocilja EC. AuNP-RF sensor: An innovative application of RF technology for sensing pathogens electrically in liquids (SPEL) within the food supply chain. Biosens Bioelectron. 2018;111: 152–158. doi:10.1016/j.bios.2018.04.010 65 Vital PG, Cabalies MBD, Rivera WL. Antimicrobial resistance in Escherichia coli and Salmonella spp isolates from fresh produce and impact to food safety J Environ Sci Health B. 2017;52: 683–689. doi:10. 1080/03601234.2017.1331676 M Pires ACDS, Soares N de FF, da Silva LHM, da Silva M do CH, De Almeida MV, Le Hyaric M, et al. A colorimetric biosensor for the detection of foodborne bacteria. Sens Actuators B Chem. 2011;153: 17–23. doi:10.1016/j.snb.2010.09.069 Mengistu DA, Tolera ST. Prevalence of Microorganisms of Public Health Significance in Ready-to-Eat Foods Sold in Developing Countries: Systematic Review and Meta-Analysis. International Journal of Food Science. Hindawi Limited; 2020. doi:10.1155/2020/8867250 Moawad AA, Hotzel H, Awad O, Tomaso H, Neubauer H, Hafez HM, et al. Occurrence of Salmonella enterica and Escherichia coli in raw chicken and beef meat in northern Egypt and dissemination of their antibiotic resistance markers. Gut Pathog. 2017;9: 57. doi:10.1186/s13099-017-0206-9 Boodoo C, Dester E, Asadullah Sharief S, Alocilja EC. Influence of Biological and Environmental Factors in the Extraction and Concentration of Foodborne Pathogens using Glycan-Coated Magnetic Nanoparticles J Food Prot. 2023; 100066. doi: 10.1016/j.jfp.2023.100066 Sarno B, Heineck D, Heller MJ, Ibsen SD. Dielectrophoresis: Developments and applications from 2010 to 2020 2020. Electrophoresis. Wiley-VCH Verlag; 2021. pp. 539–564. doi:10.1002/elps.202000156 Wu X, Han C, Chen J, Huang Y-W, Zhao Y. Rapid Detection of Pathogenic Bacteria from Fresh Produce by Filtration Filtration and Surface-Enhanced Raman Spectroscopy. JOM. 2016;68: 1156–1162. doi:10.1007/s11837-015-1724-x Li X, Ximenes E, Amalaradjou MAR, Vibbert HB, Foster K, Jones J, et al. Rapid sample processing for Detection of detection of food-borne pathogens via cross-flow microfiltration. Appl Environ Microbiol. 2013;79: 7048–54. doi:10.1128/AEM.02587-13 Kim JH, Oh SW. Rapid and sensitive detection of E. coli O157:H7 and S. Typhimurium in iceberg lettuce and cabbage using filtration, DNA concentration, and qPCR without enrichment. Food Chem. 2020; 327. doi: 10.1016/j.foodchem.2020.127036 Vinayaka AC, Ngo TA, Kant K, Engelsmann P, Dave VP, Shahbazi M-A, et al. Rapid detection of Salmonella enterica in food samples by a novel approach with combination of sample concentration and direct PCR. Biosens Bioelectron. 2019;129: 224–230. doi: 10.1016/j.bios.2018.09.078 Ogawa U, Koyama K, Koseki S. Rapid detection and enumeration of aerobic mesophiles in raw foods using dielectrophoresis. J Microbiol Methods. 2021;186. doi: 10.1016/j.mimet.2021.106251 Wang Y, Salazar JK. Culture-Independent Rapid Detection Methods for Bacterial Pathogens and Toxins in Food Matrices. Compr Rev Food Sci Food Saf. 2016;15: 183–205. doi:10.1111/1541- 434337.12175 Dester E, Alocilja E. Current Methods for Extraction and Concentration of Foodborne Bacteria with Glycan-Coated Magnetic Nanoparticles: A Review. Biosensors (Basel). 2022;12: 112. doi:10.3390/bios12020112 Wadamori Y, Gooneratne R, Hussain MA. Outbreaks and factors influencing microbiological contamination of fresh produce. J Sci Food Agric. 2017;97: 1396–1403. doi:10.1002/jsfa.8125 66 SH Jaykus, LA, Brehm-Stecher B. Advances in separation and concentration of microorganisms from food samples. Advances in Microbial Food Safety. Woodhead Publishing Limited; 2013 doi: 10.1533/9780857098740.3.173 67 Chapter 4. RAPID CLASSIFICATION OF B. CEREUS USING MAGNETIC NANOPARTICLES, SPECTROSCOPY, AND SUPERVISED MACHINE LEARNING 4.1. Introduction Foodborne outbreaks are a cause of growing concern resulting in many hospitalizations and deaths worldwide. These outbreaks in the United States lead to over 128,000 hospitalizations and approximately 3000 fatalities yearly(Centers for Disease Control and Prevention, 2011). Among the pathogens that cause these outbreaks is Bacillus cereus, a toxin-producing, gram-positive, rod-shaped bacterium (Bottone, 2010). The spore-forming nature of this pathogen increases its viability on dry surfaces of containers and subsequent contamination of foods (Choi & Kim, 2020). Primarily resulting in gastrointestinal illnesses, B. cereus can result in severe local and systemic infections, and its rapid detection is crucial for outbreak prevention. The recommended protocols for identifying pathogens such as B. cereus, according to the Bacteriological Analytical Manual, require selective media, morphological characterization, and biochemical confirmation (Yossa et al., 2023). These lengthy protocols can take several hours to days for identification, owing to the overnight growth required for increasing bacterial counts. To address the time concern, concentration methods such as centrifugation, filtration, and dielectrophoresis have been proposed (J. H. Kim & Oh, 2020; Sarno et al., 2021). However, heavy equipment requirements, clogging filters, and extraction of associated food debris, which can affect detection assays, are primary concerns (Vinayaka et al., 2019). Among other techniques, magnetic nanoparticles (MNPs) have gained attention for pathogen concentration and extraction due to their simplicity and rapidity (Dester & Alocilja, 2022). Nevertheless, MNPs must be coated with specific antibodies or phages for bacterial capture, which are expensive to produce and require cold storage. Further, covering antibodies with food microparticles can reduce the efficiency of bacterial capture, especially in complex matrices (T.-H. Kim et al., 2014; Wang et al., 2020). Recently, the coating of MNPs with carbohydrate-based ligands has been proposed in several studies offering advantages such as inexpensive synthesis and room-temperature storage (Dester & Alocilja, 2022). These MNPs have successfully extracted and concentrated pathogens from various complex foods. Some examples of carbohydrate-based moieties on MNPs include polylysine, cysteine-glycan, mannose, and short-chain glucans (El-Boubbou et al., 2007; Matta & Alocilja, 2018; Yosief et al., 2013; You et al., 2021). Among these, glycan-coated MNPs (gMNP) have successfully extracted pathogens from multiple solid and liquid food types. For example, S. enterica and E. coli were successfully extracted 68 from homogenized eggs, apple cider, and leafy greens (Matta et al., 2018; Sharief et al., 2023a). In another study, E. coli O157, L. monocytogenes, and S. aureus were extracted from several natural and processed foods (Boodoo et al., 2023). When applied to B. cereus, the gMNP was successful in its extraction from milk (Matta & Alocilja, 2018). Notably, the magnetic extraction of bacteria using gMNP from mentioned foods was achieved in large volumes (>100mL). In addition, previous studies have shown that gMNP does not hinder target bacteria detection (Matta et al., 2018). While the gMNP are not specific to bacterial types, they can be combined with selective detection techniques. For example, the gMNP-extracted pathogens from food were successfully detected using biosensor platforms and nucleic acid amplification-based assays (Sharief et al., 2023b). For preventing food-related outbreaks, minimizing the time for pathogen detection is paramount. While the concentration of low-level bacterial contaminants may be achieved using an MNP-based approach, their rapid detection is critical to address the time constraint. The conventional methods for detecting B. cereus have employed lengthy culture-based biochemical techniques or nucleic acid amplification-based assays (Meng et al., 2022), which require expensive equipment and skilled personnel. Whole genome sequencing has also been used for B. cereus identification providing accurate results, but associated costs and analyses limit its use (Nguyen & Tallent, 2019). Although mass spectrometry-based methods offer reliability, they have high initial costs and database requirements (Cheng et al., 2016). Rapid, simple-to-implement, and affordable techniques are highly desirable for preventing outbreaks. Various spectroscopy-based methods have also been used to identify B. cereus, enabling the advantage of whole-cell bacteria detection. For example, Raman spectroscopy has successfully differentiated B. cereus using silver nanoparticles and applying multivariate analysis (Cozar et al., 2019). This technique has also successfully identified sporulated forms of Bacillus cereus from dairy products (Vidic et al., 2020). However, a primary disadvantage of this technique is the requirement of expensive substrates and initial equipment installation costs (Jiang et al., 2023). Another method, Fourier Transform Infrared Spectroscopy (FTIR), successfully identified various genetically close strains of the Bacillus cereus group. Combined with artificial neural network (ANN)-based machine learning, strains were successfully differentiated in foods such as rice, milk, and soil samples (Bağcıoğlu et al., 2019). The method has also been extensively used for pathogen detection from various foods such as raw meat, apple juice, and milk, to name a few (Al-Qadiri et al., 2006; Ellis et al., 2004; Nicolaou & Goodacre, 2008; Ziyaina et al., 2020). An FTIR-based strain-typing instrument has recently been made available commercially owing to 69 its simplicity, inexpensiveness, and reliability (Tian et al., 2021). While FTIR is an established and well- accepted technique for bacterial strain identification, near-infrared spectroscopy offers some unique advantages, including further reduced costs and low sensitivity to water. NIRS is also more affordable owing to simple device construction when compared to other spectroscopic techniques and allows non- invasive measurements. Further, recent advancements in micro-electromechanical system technology have significantly reduced the size of the NIR spectrophotometer, paving the way for its enhanced field applicability. While the technique has not commonly been used for the identification of B. cereus from foods, some examples in literature that have demonstrated the use of NIRS for the detection of other food pathogens include short-wave NIR for cabbage contamination (Matulaprungsan et al., 2019), detection of L. monocytogenes, E. coli, and S. aureus from contaminated meats and cheese (Daskalov & Atanasova, 2010), and identification of lactic acid bacteria in milk samples (Treguier et al., 2019). 4.1.1. Platform novelty Although offering additional advantages such as bacteria-specific absorption signature and relatively deeper penetration of NIR light, certain aspects of NIRS need attention. For example, the requirement of bacteria separation and enrichment before obtaining the NIR spectra has been suggested as a significant obstacle (Tian et al., 2021). Further, the inability of NIRS to classify mixed strains of bacterial culture without an isolation step is another primary concern. Among these obstacles, separation and enrichment can be significantly addressed with gMNP, which have successfully concentrated and isolated bacteria from complex matrices. For instance, gMNP reduced the time for enrichment to <5 h for the isolation of S. enterica from cucumbers and melons and E. coli from leafy green vegetables, followed by genomic detection (Sharief et al., 2023a, 2023b). In the present study, B. cereus was concentrated from buffer solution and foods, following which NIRS was used for its classification. The presence of gMNP did not prevent the spectral acquisition and differentiation of B. cereus. To assess the application of this technique with complex matrices and in large volumes, the concentration of B. cereus was attempted in various carbohydrate-rich foods, including white rice, brown rice, and macaroni. Scheme 4.1 briefly describes the underlying concept of this study. While the study primarily focused on B. cereus, E. coli was also used as a model for the extraction and classification of gram-negative bacteria. 70 Scheme 4. 1. The gMNP-based bacterial extraction and detection system. 4.2. Materials and Methods 4.2.1. Materials Frozen stock culture of Escherichia coli (ATCC 15597) was obtained from the American Type Culture Collection and Bacillus cereus from the Nano-Biosensors Laboratory at Michigan State University (MSU). A UV-Visible-Near Infrared Spectrophotometer (Shimadzu) was used for spectral measurements. Magnetic nanoparticles functionalized with chitosan (glycan), proprietary to the nano-Biosensors Laboratory, MSU, were used without further modification. Food matrices, including white rice, ready-to- eat brown rice, and macaroni, were purchased from a local seller. Whirl-Pak bags for magnetic extraction were purchased from VWR (Radnor, PA, USA). Racks for magnetic separation were obtained from Spherotech Inc (Lake Forest, IL). Media for bacterial growth, Tryptic Soy Agar and Broth (TSA and TSB), and Phosphate Buffer Solution (PBS) were obtained from Sigma Aldrich (St. Louis, MO). CHROMagar was purchased from DRG International for the selective growth of E. coli and B. cereus (Springfield, NJ). Copper grids (formvar/carbon 200 mesh) for visualizing bacterial cell interaction with gMNP were obtained from Electron Microscopy Systems (Hatfield, PA) and performed at the Center for Advanced Microscopy at MSU. 4.2.2. Bacterial culture and small-volume magnetic capture Frozen bacterial cultures were revived by plating on TSA and incubated at 37 °C for 24-48 h. A single colony from the plate was transferred to 9 ml of TSB and grown fresh before each experiment of magnetic extraction. 71 The capture of bacterial cells in small volumes was quantified using capture capacity, which uses optical density at 600 nm of pure culture (OD initial) and supernatant of magnetically separated sample (OD treated). Capture Capacity (%) = OD initial−OD treated OD initial X 100 (4.1) 4.2.3. Large-volume magnetic separation of bacteria from foods Carbohydrate-rich food matrices chosen for this study include cooked white rice, ready-to-eat brown rice, and macaroni. Artificial inoculation was achieved by transferring 1 ml of 10-3 dilution of fresh bacterial culture (103 CFU/mL) to 25 gm of food matrix in Whirl-Pak bags. Following artificial inoculation, the bags were incubated at room temperature for 1 h, after which 225 ml of PBS was added and homogenized in a stomacher. The liquefied food matrix was separated into 2 bags with 100 ml each, one serving as a control and the other as a test to which gMNP (1 ml, 5 mg/ml) were added. Both test and control samples were plated to determine the effectiveness of capture, given by Concentration Factor. Concentration Factor = # CFU in test sample # CFU in control sample (4.2) 4.2.4. Visualization of binding between bacterial cells and gMNP The binding between bacterial cells and gMNP was initially confirmed using Transmission Electron Microscopy (TEM, JEM-1400 Flash, Joel, Tokyo, Japan). The standard negative staining procedure was used to visualize the bacteria where the magnetically separated bacterial cells were first dissolved in a fixative solution of 2.5 % glutaraldehyde in 0.1 M cacodylate buffer, following which 5 µl of suspension was dropped onto the copper grid for about 20 seconds and washed with distilled water. Following drying, the sample was stained with 0.1% uranyl acetate and air dried before image acquisition, which was done in the range of 5000-25000 X magnification. 4.2.5. Acquisition of near-infrared spectrum The spectra of pure culture and magnetically extracted samples were obtained from 800 nm to 1500 nm, and 1100-1350 nm was used for classification. Data pre-processing and classification were accomplished in MATLAB (MathWorks, R2022a). Spectra of magnetically extracted cells, control samples, and pure gMNP were obtained. The spectrometer was set up in transmittance mode and each sample was blanked with gMNP solution suspended in water. The sample sizes for classification of samples from pure culture was 56. Data pre-processing involved normalization of the acquired spectra followed by data centering and smoothing using Savitzky-Golay, a second derivative of the spectra was 72 used. Data pre-processing was followed by principal component analysis, achieved using singular value decomposition, and implemented with an in-built function in MATLAB. Finally, the obtained spectra were classified using an Artificial Neural Network, Support Vector Machine, and Naïve Bayes. 4.2.6. Limit of detection To establish the limit of detection, overnight cultures of B. cereus and E. coli were serially diluted until 10-6 and their NIR spectra were obtained. As mentioned in the previous section, the spectra were pre- processed before classification. The sample size used for establishing the limit of detection was 108 of which 72 were used for calibration and 36 were used for testing. 4.3. Results and discussion 4.3.1. Visualizing the binding between bacterial cells and gMNP The binding of gMNP to B. cereus and E. coli was initially visualized using TEM, and micrographs are shown in Figure 4.1 (A and B). Individual gMNP are seen binding to bacterial cells towards their curved ends for both bacterial types. This binding towards the curved ends confirm earlier reports of gMNP binding in other gram-positive and gram-negative bacteria (Matta et al., 2018). It has therefore been hypothesized that along with electrostatic forces and random Brownian motion, attachment is further assisted with site-specific interaction of bacterial cell wall with chitosan functionalized on the gMNP (Dester & Alocilja, 2022). While not seen in these images, gMNP have also been found to interact with the flagella, possibly due to their charged nature (Matta & Alocilja, 2018). Further, multiple nanoparticles binding to individual bacterial cells was commonly observed, which could have enhanced their concentration and extraction. Following the confirmation of binding between bacterial cells and gMNP with imaging, the capture capacity was determined using optical density measurements (Figure 4.1 C). These measurements were taken with bacterial cells suspended in PBS (pH 7.4) to provide for a simple matrix with a negligible influence of physical and chemical factors. A difference in capture capacity was seen between the two bacteria: higher at 80% for E. coli and 70% for B. cereus. Compared to previously published reports which used electrostatic binding, such differences among bacterial species have commonly been observed. For example, polyethyleneimine-coated MNPs successfully captured multiple bacterial strains, including Salmonella, E. coli, L. monocytogenes, and B. subtilis with variable capture capacities, from 40% to 90% depending on bacterial type (Chen et al., 2019). Oligosaccharide-coated magnetic beads showed a strain-dependent capture of 17-34% (Yosief et al., 2013). A disparity was also seen with 73 amine-functionalized MNPs, where their positively charged nature was used for bacterial capture. A lower capture of 55% and 66% was seen for Salmonella and S. aureus, respectively, although E. coli and B. subtilis showed a capture of 97% (Huang et al., 2010). This difference in the capture of E. coli and B. cereus using gMNP may be attributed to multiple factors. The polarity of the bacterial cell surface is one example, with the bacterial cell wall displaying a net electronegative charge. With numerous earlier studies suggesting a difference in the surface charge of cell wall depending on bacterial strains and types, this could have contributed towards a differing capture capacity. Additionally, the chain-forming properties of cells and flagellar arrangements are other factors that need further investigation. Apart from the effect of cell-related properties, the surface properties of gMNP could also have influenced capture since multiple surface properties can influence bacterial adhesion. For instance, surface topography, wettability, charge, and roughness have all been proposed as possible factors affecting bacterial adhesion (Zheng et al., 2021). Further, since bacterial surfaces consist of carbohydrate-binding sites, the distribution of these sites could also have played a role in its capture. After confirming the successful capture of B. cereus and E. coli in small volumes, their application towards large volumes was assessed. Concentration and extraction from large volumes of PBS and water were implemented first, followed by various food matrices; results are discussed in the following section. Figure 4. 1. Transmission Electron micrographs show the binding of gMNP with (A) B. cereus and (B) E. coli. The capture capacity of gMNP in PBS is shown in (C). 4.3.2. Concentration Factor in water, PBS, and foods Among the primary obstacles to MNP-based bacterial extraction from food is their application to large sample volumes. While multiple MNP types in previous studies have shown promise in bacterial capture, their application to large volumes of liquid and solid foods is uncommon. To test the applicability of our 74 gMNP towards large-scale magnetic capture in liquids, initial capture efficiency experiments were done in 100 ml of water and PBS. Since large sample volumes prevent the use of optical density measurements to quantify extraction efficiency, which typically needs bacterial loads > 107 CFU/ml, concentration factor (CF) was used. With CF>1, a successful concentration of both bacterial types was seen in water and PBS, although the concentration in water was lower than that in PBS; results are shown in Figure 4.2 A. This higher concentration in PBS could be attributed to the isotonic nature of the buffer, which allows the cells to maintain their integrity. In both liquids, a relatively higher concentration of B. cereus was achieved with a CF of 6 and 10 in water and PBS, respectively, compared to 2 and 5 for E. coli. The osmolarity and ionic composition of PBS allow the preservation of cell wall structure and prevent osmotic stress on the cells, possibly resulting in a higher capture. Previously, promising results using similar gMNP were seen to extract B. cereus, E. coli, and Salmonella from various milk sample types (Matta & Alocilja, 2018). Our results demonstrate the applicability of gMNP for bacterial extraction from large volumes of liquids. Following this, the concentration factor of gMNP from various solid foods was tested. The concentration of B. cereus and E. coli from various solid complex matrices, including freshly prepared white rice, ready-to-eat brown rice, and macaroni, was attempted; results are shown in Figure 4.2 B. B. cereus was successfully concentrated from all the foods, although a variation among the matrices was seen. The CF of B. cereus was found to be highest for white rice, followed by brown rice and macaroni. The absence of additional ingredients in fresh white rice could explain a comparatively high CF of 8. In contrast, the ready-to-eat brown rice showed a lower CF of 6, which could be attributed to additional contents and a further complicated microstructure of these foods. While freshly prepared white rice may have provided for a relatively simple matrix, ready-to-eat brown rice also consisted of canola or safflower oil. Furthermore, brown rice has relatively higher protein, fat, and fiber content, possibly influencing capture (Saleh et al., 2019). For ready-to-eat macaroni, the presence of fats and proteins and its glutinous texture, owing to the cheese sauce mix, may have resulted in low CF. While additional contents may also have contributed to the variation in CF for brown rice and macaroni as indicated by the error bars, the successful concentration of B. cereus from such complex matrices by gMNP is noteworthy. Although plating-based CF confirmed successful extraction, bacterial separation in the presence of a food matrix was also analyzed using laser microscopy, and the micrographs are shown in Figure 4.2 (C and D). 75 With a CF of 1, no significant concentration of E. coli from any of the matrices was observed, which could be due to multiple factors. For example, the chain-forming nature of individual B. cereus cells, the influence of food matrix on surface properties of bacterial cells, and the interaction of gMNP with food microparticles may have resulted in this difference in concentration. Previously published work has elaborated on the charged nature of food microparticles (Cano-Sarmiento et al., 2018). Since the bacterial attachment to gMNP is hypothesized to rely on multiple forces, including electrostatic, binding to charged food microparticles is inevitable (Boodoo et al., 2023). Figure 4. 2. (A) Concentration factor of B. cereus and E. coli in sterile matrices and (B) in carbohydrate- rich foods. Confocal laser micrographs confirmed successful bacterial capture in foods; images from rice and macaroni are shown in (C) and (D). Applied to foods, other MNP coatings such as gold, protamine, and amine have also been used. Among these, aminated MNPs with a net positive charge successfully captured multiple bacterial species from various beverages, including grape juice, green tea, and groundwater (Huang et al., 2010). While these aminated MNPs achieved high capture efficiencies with liquid foods, their application to complex solid 76 foods was not attempted. Short-chain glucans are among the relatively few examples of coatings for bacterial capture from solid foods. The short-chain glucan-coated MNPs successfully captured E.coli O157 from sausage (You et al., 2021), but their preparation time spanned several hours with multiple steps. The preparation of gMNP, on the contrary, is achieved via a one-pot method and does not require additional surface functionalization steps. Notably, while gMNP were earlier successful in capturing E. coli and Salmonella from various solid foods (Sharief et al., 2023a, 2023b), extraction of B. cereus from complex solid foods was not implemented. Following the confirmation of concentration and extraction of B. cereus using gMNP from PBS, water, and various foods, NIRS was used for its classification. The section that follows discusses the rapid classification of magnetically separated B. cereus. Initial experiments were done following the extraction of bacterial cells from culture media. Following this, the extraction and classification of serially diluted bacterial samples were attempted to establish a detection limit. 4.3.3. Classification of B. cereus and E. coli Near-infrared spectroscopy has been used earlier to identify the contamination of foods with various bacteria (Veleva-Doneva et al., 2010). However, overnight bacterial enrichment has commonly been observed to ensure minimum bacterial counts (Tian et al., 2021). Therefore, a primary obstacle with the use of NIRS for pathogen classification is its inefficiency in low bacterial loads and overnight enrichment requirements. Our study addressed this issue using gMNP for bacterial cell concentration to omit or shorten the enrichment step, following which NIRS can be used for classification. To assess the effect of gMNP, the raw spectra of extracted samples were initially analyzed. Figure 4.3 A shows a sample of the normalized raw spectra of E. coli and B. cereus extracted using gMNP, spectra of non-gMNP samples, and that of pure gMNP. The sample spectra are shown in the wavelength range of 1100-1350 nm; the same range was used to classify B. cereus from E. coli. The wavelength range used for classification is variable in the literature. For instance, the whole NIR spectrum was used to determine pharmaceutical contamination levels (Quintelas et al., 2015) and monitor the contamination of Atlantic salmon (Tito et al., 2012). Alternatively, a wavelength range of 1445-2348 nm was used for detecting contamination in cheese (Daskalov & Atanasova, 2010), and 600-1100 nm was used for beef contamination (Daskalov et al., 2011). Short-wave NIR in 700-1100 nm range, in another study, was used to monitor spoilage of shredded cabbage (Matulaprungsan et al., 2019). The gMNP extracted samples showed two dominant peaks at 1180 and 1220 nm. These peaks correspond to N-H frequency doubling and second overtone of C-H stretching, possibly arising due to 77 the presence of lipids in the bacterial cell wall (Wilson et al., 2015; Zhao et al., 2021). While the influence of chitosan on the gMNP cannot be ignored, peaks at similar locations are seen for samples from bacterial cultures without gMNP. However, the peaks for non-gMNP samples are not as strong as those for gMNP-extracted test samples, resulting in them being suppressed in the figure. Figure 4. 3. (A) Processed spectra of gMNP-extracted samples, non-gMNP controls, and pure gMNP. Results from the principal component analysis are shown in (B). Following these, a second derivate of the spectra for gMNP-extracted samples was obtained and is shown in the inset in Figure 4.3 A. As expected, the pre-processing did not result in significant differences in the two bacterial types. Therefore, principal component analysis was performed by treating the pre-processed spectral data as a matrix and using singular value decomposition for its factorization. The first 3 principal components accounted for the maximum variation in the data and successfully differentiated gMNP-extracted B. cereus from E. coli (Figure 4.3 B). Although a complete delineation of the two samples was not seen, most of the samples from each bacterial type clustered together. Since the detection of low-level bacterial contamination is of prime importance, the limit of detection of the classification was also assessed. Identifying low bacterial load is critical for preventing food-related outbreaks. Therefore, a limit of detection using NIRS was established using spectra of serially diluted samples. Overnight cultures of B. cereus and E. coli were serially diluted until 102 CFU/mL. 4.3.4. Limit of detection and classification from foods The principal components were obtained following gMNP-based magnetic separation, spectral acquisition, and data processing. As seen previously, the first 3 principal components successfully differentiated the two bacterial types, seen with two clusters in Figure 4.4. The results from the principal 78 component analysis were also fed to three supervised machine learning-based classification techniques, shown in Table 4.1. Support Vector Machine provided the highest classification accuracy at 85%, followed by Artificial Neural Networks at 80% and Naïve Bayes at 79%. Figure 4. 4. Results from principal component analysis to establish a limit of detection. While a classification accuracy of 85% was obtained using Support Vector Machine in cells extracted from pure culture, a lower accuracy of 70% was seen in bacterial samples extracted from foods. There could be multiple reasons for the misclassification. First, the charge-based electrostatic nature of the gMNP may have resulted in food microparticles adhering to them and influencing the spectra. Previous studies have shown how charged food microparticles affected bacterial separation from foods using dielectrophoresis, the polarization-based separation technique (Boodoo et al., 2023; Cano-Sarmiento et al., 2018). Further, the presence of oils and other ingredients in brown rice and macaroni must have presented complications, although the efficiency of gMNP in their presence is noteworthy. Secondly, while no colonies were seen upon plating the control with no inoculation, the presence of natural microflora cannot be dismissed, as bacterial growth depends on the media type. Third, while high CF values obtained for B. cereus samples for all food types confirm their effective concentration and separation, the possibility of adherence of bacterial cells to the food matrix cannot be disregarded. Previous studies using NIRS for bacterial classification from food have used centrifuge-based bacterial extraction and have commonly relied on long pre-enrichment times prior to bacteria detection for higher classification accuracy. For example, contamination in cheese samples by L. monocytogenes was 79 successfully detected using NIR but required an enrichment, bringing the total detection time to >24 h (Daskalov & Atanasova, 2010). Contamination detection in beef with various types of bacteria, including L. monocytogenes, S. aureus, and E. coli, was achieved with an accuracy of 98%, but the total detection time was 48 h (Daskalov et al., 2011). Increasing the enrichment time can possibly enhance the classification accuracy using gMNP. Table 4. 1. Classification accuracy of the techniques used. Classification Classification method Accuracy (%) 82.6 Artificial Neural Network Support Vector 85.2 Machine Naïve Bayes 79 Tian et al., indicated 3 possible reasons preventing the use of NIR for food pathogen detection: its inapplicability with low bacterial counts, the effect of the surrounding matrix, and the inability to classify samples with multiple strains (Tian et al., 2021). Using our gMNP addressed the first two concerns by concentrating bacterial cells and, to an extent, separating them from the surrounding matrix. The concentration of cells possibly resulted in greater bacterial counts and thereby contributed towards an enhanced signal from the NIR spectrophotometer. Scheme 4.2 below shows our rendition of the beneficial effect of gMNP. Previously, MNPs have been used to extract and detect. E. coli O157 and S. typhimurium from food using FTIR, although antibodies were used for their capture (Ravindranath et al., 2009). Specific recognition moieties, such as antibodies and phages, add costs and require cold storage. The gMNP, alternatively, were produced in-house, cost < $ 0.5 per extraction (Briceno et al., 2019), and have a long shelf life at room temperature. While the gMNP do not have recognition molecules, their combination with specific detection techniques such as NIR allows rapid bacterial classification. 80 Scheme 4. 2. Our hypothesis showing that low bacterial loads do not give a strong signal using NIRS (a); the signal is amplified using gMNP, which concentrate bacterial cells. The bacterial extraction of cells using gMNP combined with a NIRS-based approach can provide for quick screening of food samples, following which other methods can be used for analysis. Previous studies have confirmed that gMNP did not hinder bacterial detection from traditional techniques such as PCR or qPCR. Additionally, gMNP-based bacterial extraction from various foods has successfully been combined with electrochemical and gold-nanoparticles-based biosensors. 4.4. Conclusion Magnetic extraction and concentration and extraction of Bacillus cereus and E. coli using gMNP were combined with near-infrared spectroscopy for rapid classification. Successful concentration of B. cereus was observed in artificially contaminated rice and pasta samples with high concentration factors of 7.8, 6.1, and 3.8, respectively. Upon using near infrared spectroscopy for classification, the presence of gMNP did not hinder spectral acquisition, and successful differentiation of B. cereus from E. coli was observed. The concentration of bacterial cells using gMNP allowed for a stronger NIR signal. Using support vector machine, a classification accuracy of 85% was achieved at bacterial loads as low as 102 CFU/mL. This concentration and classification system can enable rapid screening of food samples in low- resource settings. 81 REFERENCES Al-Qadiri, H. M., Lin, M., Cavinato, A. G., & Rasco, B. A. (2006). Fourier transform infrared spectroscopy, detection and identification of Escherichia coli O157:H7 and Alicyclobacillus strains in apple juice. International Journal of Food Microbiology, 111(1), 73–80. https://doi.org/10.1016/j.ijfoodmicro.2006.05.004 Bağcıoğlu, M., Fricker, M., Johler, S., & Ehling-Schulz, M. (2019). Detection and Identification of Bacillus cereus, Bacillus cytotoxicus, Bacillus thuringiensis, Bacillus mycoides and Bacillus weihenstephanensis via Machine Learning Based FTIR Spectroscopy. Frontiers in Microbiology, 10(APR). https://doi.org/10.3389/fmicb.2019.00902 Boodoo, C., Dester, E., Asadullah Sharief, S., & Alocilja, E. C. (2023). Influence of Biological and Environmental Factors in the Extraction and Concentration of Foodborne Pathogens using Glycan- Coated Magnetic Nanoparticles. Journal of Food Protection, 100066. https://doi.org/10.1016/j.jfp.2023.100066 Bottone, E. J. (2010). Bacillus cereus , a Volatile Human Pathogen. Clinical Microbiology Reviews, 23(2), 382–398. https://doi.org/10.1128/CMR.00073-09 Briceno, R. K., Sergent, S. R., Benites, S. M., & Alocilja, E. C. (2019). Nanoparticle-Based Biosensing Assay for Universally Accessible Low-Cost TB Detection with Comparable Sensitivity as Culture. Diagnostics, 9(4), 222. https://doi.org/10.3390/diagnostics9040222 Cano-Sarmiento, C., Téllez-Medina, D. I., Viveros-Contreras, R., Cornejo-Mazón, M., Figueroa-Hernández, C. Y., García-Armenta, E., Alamilla-Beltrán, L., García, H. S., & Gutiérrez-López, G. F. (2018). Zeta Potential of Food Matrices. Food Engineering Reviews, 10(3), 113–138. https://doi.org/10.1007/s12393-018-9176-z Centers for Disease Control and Prevention. (2011). CDC Estimates of Foodborne Illness in the United States. CDC 2011 Estimates. http://www.cdc.gov/foodborneburden/2011-foodborne- estimates.html Chen, B., Xie, H., Zhang, A., Liu, N., Li, Q., Guo, J., & Su, B. (2019). Synthesis of PEI-Functionalized Magnetic Nanoparticles for Capturing Bacteria. Journal of Wuhan University of Technology-Mater. Sci. Ed., 34(1), 236–242. https://doi.org/10.1007/s11595-019-2041-y Cheng, K., Chui, H., Domish, L., Hernandez, D., & Wang, G. (2016). Recent development of mass spectrometry and proteomics applications in identification and typing of bacteria. PROTEOMICS – Clinical Applications, 10(4), 346–357. https://doi.org/10.1002/prca.201500086 Choi, W., & Kim, S.-S. (2020). Outbreaks, Germination, and Inactivation of Bacillus cereus in Food Products: A Review. Journal of Food Protection, 83(9), 1480–1487. https://doi.org/10.4315/0362- 028X.JFP-19-429 Cozar, I. B., Colniţă, A., Szöke-Nagy, T., Gherman, A. M. R., & Dina, N. E. (2019). Label-Free Detection of Bacteria Using Surface-Enhanced Raman Scattering and Principal Component Analysis. Analytical Letters, 52(1), 177–189. https://doi.org/10.1080/00032719.2018.1445747 Daskalov, H., & Atanasova, S. (2010). Microbiological and Near Infrared Spectroscopy Analysis of Bulgarian Yellow Cheese, Conta-Minated With Listeria. Bulgarian Journal of Veterinary Medicine, 31–37. http://tru.uni-sz.bg/bjvm/BJVM March 2010 p.31-37.pdf 82 Daskalov, H., Atanassova, S., Stoyanchev, T., & Santo, R. (2011). Application of near infrared spectroscopy for rapid noninvasive detection of Listeria monocytogenes, Escherichia coli and staphylococcus aureus growth in foods. Bulgarian Journal of Veterinary Medicine, 14(3), 150–157. Dester, E., & Alocilja, E. (2022). Current Methods for Extraction and Concentration of Foodborne Bacteria with Glycan-Coated Magnetic Nanoparticles: A Review. Biosensors, 12(2), 112. https://doi.org/10.3390/bios12020112 El-Boubbou, K., Gruden, C., & Huang, X. (2007). Magnetic Glyco-nanoparticles: A Unique Tool for Rapid Pathogen Detection, Decontamination, and Strain Differentiation. Journal of the American Chemical Society, 129(44), 13392–13393. https://doi.org/10.1021/ja076086e Ellis, D. I., Broadhurst, D., & Goodacre, R. (2004). Rapid and quantitative detection of the microbial spoilage of beef by Fourier transform infrared spectroscopy and machine learning. Analytica Chimica Acta, 514(2), 193–201. https://doi.org/10.1016/j.aca.2004.03.060 Huang, Wang, Y.-F., & Yan, X.-P. (2010). Amine-functionalized magnetic nanoparticles for rapid capture and removal of bacterial pathogens. Environmental Science & Technology, 44(20), 7908–7913. https://doi.org/10.1021/es102285n Jiang, G., Xu, W., Huang, X., Dai, Z., Zhou, C., Hong, P., & Li, C. (2023). Detection of Bacillus cereus Spore Biomarkers Using SERS-Based Cuttlebone-Derived Organic Matrix/Silver Nanoparticles. ACS Sustainable Chemistry & Engineering, 11(10), 4145–4154. https://doi.org/10.1021/acssuschemeng.2c06626 Kim, J. H., & Oh, S. W. (2020). Rapid and sensitive detection of E. coli O157:H7 and S. Typhimurium in iceberg lettuce and cabbage using filtration, DNA concentration, and qPCR without enrichment. Food Chemistry, 327. https://doi.org/10.1016/j.foodchem.2020.127036 Kim, T.-H., Park, J., Kim, C.-J., & Cho, Y.-K. (2014). Fully Integrated Lab-on-a-Disc for Nucleic Acid Analysis of Food-Borne Pathogens. Analytical Chemistry, 86(8), 3841–3848. https://doi.org/10.1021/ac403971h Matta, L. L., & Alocilja, E. C. (2018). Carbohydrate Ligands on Magnetic Nanoparticles for Centrifuge-Free Extraction of Pathogenic Contaminants in Pasteurized Milk. Journal of Food Protection, 81(12), 1941–1949. https://doi.org/10.4315/0362-028X.JFP-18-040 Matta, L. L., Harrison, J., Deol, G. S., & Alocilja, E. C. (2018). Carbohydrate-Functionalized Nanobiosensor for Rapid Extraction of Pathogenic Bacteria Directly From Complex Liquids With Quick Detection Using Cyclic Voltammetry. IEEE Transactions on Nanotechnology, 17(5), 1006–1013. https://doi.org/10.1109/TNANO.2018.2841320 Matulaprungsan, B., Wongs-Aree, C., Penchaiya, P., Maniwara, P., Kanlayanarat, S., Ohashi, S., & Nakano, K. (2019). Feasibility of Determination of Foodborne Microbe Contamination of Fresh-Cut Shredded Cabbage Using SW-NIR. AgriEngineering, 1(2), 246–256. https://doi.org/10.3390/agriengineering1020018 Meng, J.-N., Liu, Y.-J., Shen, X., Wang, J., Xu, Z.-K., Ding, Y., Beier, R. C., Luo, L., Lei, H.-T., & Xu, Z.-L. (2022). Detection of emetic Bacillus cereus and the emetic toxin cereulide in food matrices: Progress and perspectives. Trends in Food Science & Technology, 123, 322–333. https://doi.org/10.1016/j.tifs.2022.03.023 83 Nguyen, A. T., & Tallent, S. M. (2019). Screening food for Bacillus cereus toxins using whole genome sequencing. Food Microbiology, 78, 164–170. https://doi.org/10.1016/j.fm.2018.10.008 Nicolaou, N., & Goodacre, R. (2008). Rapid and quantitative detection of the microbial spoilage in milk using Fourier transform infrared spectroscopy and chemometrics. The Analyst, 133(10), 1424. https://doi.org/10.1039/b804439b Ravindranath, S. P., Mauer, L. J., Deb-Roy, C., & Irudayaraj, J. (2009). Biofunctionalized Magnetic Nanoparticle Integrated Mid-Infrared Pathogen Sensor for Food Matrixes. Analytical Chemistry, 81(8), 2840–2846. https://doi.org/10.1021/ac802158y Saleh, A. S. M., Wang, P., Wang, N., Yang, L., & Xiao, Z. (2019). Brown Rice Versus White Rice: Nutritional Quality, Potential Health Benefits, Development of Food Products, and Preservation Technologies. In Comprehensive Reviews in Food Science and Food Safety. Blackwell Publishing Inc. https://doi.org/10.1111/1541-4337.12449 Sarno, B., Heineck, D., Heller, M. J., & Ibsen, S. D. (2021). Dielectrophoresis: Developments and applications from 2010 to 2020. In Electrophoresis (Vol. 42, Issue 5, pp. 539–564). Wiley-VCH Verlag. https://doi.org/10.1002/elps.202000156 Sharief, S. A., Caliskan-Aydogan, O., & Alocilja, E. (2023a). Carbohydrate-coated magnetic and gold nanoparticles for point-of-use food contamination testing. Biosensors and Bioelectronics: X, 13, 100322. https://doi.org/10.1016/j.biosx.2023.100322 Sharief, S. A., Caliskan-Aydogan, O., & Alocilja, E. C. (2023b). Carbohydrate-coated nanoparticles for PCR- less genomic detection of Salmonella from fresh produce. Food Control, 150, 109770. https://doi.org/10.1016/j.foodcont.2023.109770 Tian, Y., Gao, X., Qi, W.-L., Wang, Y., Wang, X., Zhou, J., Lu, D., & Chen, B. (2021). Advances in differentiation and identification of foodborne bacteria using near infrared spectroscopy. Analytical Methods, 13(23), 2558–2566. https://doi.org/10.1039/D1AY00124H Tito, N. B., Rodemann, T., & Powell, S. M. (2012). Use of near infrared spectroscopy to predict microbial numbers on Atlantic salmon. Food Microbiology, 32(2), 431–436. https://doi.org/10.1016/j.fm.2012.07.009 Treguier, S., Couderc, C., Tormo, H., Kleiber, D., & Levasseur-Garcia, C. (2019). Identification of lactic acid bacteria Enterococcus and Lactococcus by near-infrared spectroscopy and multivariate classification. Journal of Microbiological Methods, 165(June), 105693. https://doi.org/10.1016/j.mimet.2019.105693 Veleva-Doneva, P., Draganova, T., Atanassova, S., & Tsenkova, R. (2010). Detection of bacterial contamination in milk using NIR spectroscopy and two classification methods - SIMCA and Neuro – Fuzzy classifier. IFAC Proceedings Volumes, 43(26), 225–229. https://doi.org/10.3182/20101206-3- JP-3009.00039 Vidic, J., Chaix, C., Manzano, M., & Heyndrickx, M. (2020). Food Sensing: Detection of Bacillus cereus Spores in Dairy Products. Biosensors, 10(3), 15. https://doi.org/10.3390/bios10030015 Vinayaka, A. C., Ngo, T. A., Kant, K., Engelsmann, P., Dave, V. P., Shahbazi, M.-A., Wolff, A., & Bang, D. D. (2019). Rapid detection of Salmonella enterica in food samples by a novel approach with combination of sample concentration and direct PCR. Biosensors and Bioelectronics, 129, 224–230. https://doi.org/10.1016/j.bios.2018.09.078 84 Wang, Z., Cai, R., Gao, Z., Yuan, Y., & Yue, T. (2020). Immunomagnetic separation: An effective pretreatment technology for isolation and enrichment in food microorganisms detection. Comprehensive Reviews in Food Science and Food Safety, 19(6), 3802–3824. https://doi.org/10.1111/1541-4337.12656 Wilson, R. H., Nadeau, K. P., Jaworski, F. B., Tromberg, B. J., & Durkin, A. J. (2015). Review of short-wave infrared spectroscopy and imaging methods for biological tissue characterization. Journal of Biomedical Optics, 20(3), 030901. https://doi.org/10.1117/1.JBO.20.3.030901 Yosief, H. O., Weiss, A. A., & Iyer, S. S. (2013). Capture of uropathogenic E. coli by using synthetic glycan ligands specific for the pap-pilus. Chembiochem : A European Journal of Chemical Biology, 14(2), 251–259. https://doi.org/10.1002/cbic.201200582 Yossa, N., Huang, S., Canida, T., Binet, R., Macarisin, D., Bell, R., Tallent, S., Brown, E., & Hammack, T. (2023). qPCR detection of viable Bacillus cereus group cells in cosmetic products. Scientific Reports, 13(1), 4477. https://doi.org/10.1038/s41598-023-31128-3 You, S.-M., Jeong, K.-B., Luo, K., Park, J.-S., Park, J.-W., & Kim, Y.-R. (2021). Paper-based colorimetric detection of pathogenic bacteria in food through magnetic separation and enzyme-mediated signal amplification on paper disc. Analytica Chimica Acta, 1151, 338252. https://doi.org/10.1016/j.aca.2021.338252 Zhao, Y., Fang, S., Ye, Y., & Yu, K. (2021). Chemometric development using portable molecular vibrational spectrometers for rapid evaluation of AVC (Valsa mali Miyabe et Yamada) infection of apple trees. Vibrational Spectroscopy, 114, 103231. https://doi.org/10.1016/j.vibspec.2021.103231 Zheng, S., Bawazir, M., Dhall, A., Kim, H. E., He, L., Heo, J., & Hwang, G. (2021). Implication of Surface Properties, Bacterial Motility, and Hydrodynamic Conditions on Bacterial Surface Sensing and Their Initial Adhesion. In Frontiers in Bioengineering and Biotechnology (Vol. 9). Frontiers Media S.A. https://doi.org/10.3389/fbioe.2021.643722 Ziyaina, M., Rasco, B., & Sablani, S. S. (2020). Rapid methods of microbial detection in dairy products. Food Control, 110, 107008. https://doi.org/10.1016/j.foodcont.2019.107008 85 Chapter 5. RAPID DETECTION OF E. COLI FROM LEAFY GREENS USING CARBOHYDRATE-COATED NANOPARTICLES This chapter has been published in Biosensors and Bioelectronics: X (10.1016/j.biosx.2023.100322) with minor changes and is reprinted with permission. Copyright 2023 Elsevier Ltd. 5.1. Introduction Outbreaks related to contaminated food are a serious concern, making millions ill and leading to numerous deaths each year. An earlier report by the World Health Organization indicated approximately 420,000 deaths from 600 million foodborne infections annually (Havelaar et al., 2015). Some common bacteria that lead to food-related illnesses include Salmonella enterica, Listeria monocytogenes, Escherichia coli, and Bacillus cereus (J. Y. Huang et al., 2016; Liu et al., 2019). According to the Centers for Disease Control, E. coli has commonly been linked with outbreaks related to leafy greens such as romaine lettuce, baby spinach, packaged salads, and ready-to-eat salads, among others (Centers for Disease Control and Prevention, 2022). Thus, rapid detection of pathogenic bacteria frequently associated with food is critical for controlling and preventing outbreaks. Traditionally, culture-based methods followed by biochemical or PCR-based identification have been used for pathogen detection (Law et al., 2015; Zhao et al., 2014). For these techniques, however, overnight enrichment of pathogens is necessary to ensure the presence of a minimum number of cells required for detection (Law et al., 2015), (Mandal et al., 2011). The pre-enrichment step may be eliminated or shortened by employing techniques to separate bacteria effectively. Centrifugation, filtration, and polymer-based partitioning are a few common physical methods used to achieve this objective (Paniel & Noguer, 2019). Although these methods enable the physical enrichment of cells, clogging filters and isolating undesirable food debris are common concerns. Other recently developed bacterial concentration systems, such as syringe-based pathogen enrichment and pump-based automated bacterial concentration, were unsuccessful in addressing previously mentioned drawbacks (Hahm et al., 2015; Zhang et al., 2018). Therefore, rapid and efficient concentration techniques are highly desirable to minimize detection time. Among primary alternatives, magnetic nanoparticles (MNPs) based methods have gained attention owing to their simplicity. Immunomagnetic nanoparticles, which are conjugated with antibodies, have conventionally been used to concentrate and extract target bacteria. However, blocking of the antibody by food debris can result in the prevention of capture and separation (Dwivedi & Jaykus, 2011). Moreover, antibody production can be time-consuming and expensive, limiting their application. 86 Glycan-coated non-specific MNPs (gMNP) offer cost-effective extraction of bacteria with longer shelf life, robust storage conditions, and efficient bacterial capture. These MNPs have been shown to bind to both gram-positive and gram-negative bacteria in complex food matrices (E. Dester & Alocilja, 2022). For example, an earlier study used gMNP to concentrate pathogens from milk. Significant concentrations of S. Enteritidis, E. coli, and B. cereus were shown to be captured from various types of milk- Vitamin D, 2 % reduced fat, and fat-free (Matta & Alocilja, 2018). Although non-specific, these gMNP may be combined with detection platforms that can guarantee both sensitivity and specificity, as shown earlier (Matta, Karuppuswami, et al., 2018). This study used gMNP to isolate E. coli from artificially contaminated spinach and lettuce samples. For food pathogen identification in low-resource settings, simplicity, inexpensiveness, and rapidity of detection platforms are of primary importance. While traditional culture-based methods are considered the gold standard owing to their low cost and reliability, more sophisticated methods for rapid bacteria detection are gaining popularity (Abayasekara et al., 2017; Ahmed et al., 2014a). Examples include liquid and gas chromatography, mass spectrometry-based approaches, and spectroscopy-based methods (Ahmed et al., 2014a; Vaisocherová-Lísalová et al., 2016). These methods can detect whole cells and multiple variations of these techniques have been developed allowing rapid label free identification of various gram-positive and gram-negative bacteria. These methods can detect whole cells, but the size and cost of the equipment required for their implementation make them inapplicable to low-resource environments. As with PCR-based approaches, portable DNA-based electrochemical and surface plasmon resonance biosensors have demonstrated excellent sensitivity for pathogen detection (Malhotra et al., 2022; Prabowo et al., 2018). However, expenses associated with these instruments are still a significant concern. A desirable solution may be provided by colorimetric methods such as gold nanoparticles-based, where equipment use is minimal, allowing rapid visual detection of pathogens. Recently, gold nanoparticles (GNPs) have gained popularity owing to their size-dependent optical properties and ability to be functionalized with biomolecules (Baetsen-Young et al., 2018a; Shafiqa et al., 2018). The surface plasmon resonance of GNPs has been employed in the past to rapidly detect pathogens using both whole-cell and nucleic acid-based approaches (Ahmed et al., 2014b; Peng & Chen, 2019). While GNPs may be combined with various detection platforms such as electrochemical, surface plasmon, and spectroscopy-based (Lee et al., 2018; Mohammed et al., 2014), colorimetric pathogen detection using GNPs alone minimizes the use of heavy equipment. These detection assays use the stability of GNPs in unfavorable acidic or basic environments to detect DNA from target bacteria. 87 Hybridization between single-stranded probes immobilized on GNPs and target DNA contributes towards their stability, while the presence of non-target DNA results in their aggregation. The stability of GNPs is indicated by their red appearance; aggregation of nanoparticles changes their color to blue. The use of GNPs for DNA detection from pure bacterial cultures has commonly been seen in the past (Andreadou et al., 2014; Bakthavathsalam et al., 2012; Quintela et al., 2019); however, their application towards detection from complex matrices such as food needs attention. GNPs can offer a practical solution for on-site pathogen detection from complex matrices. However, traditional GNP synthesis methods are not economical and result in highly sensitive nanoparticles. Considering that molar range of reaction conditions using traditional GNPs is lower than that of many biological salt solutions, robust alternatives are preferable. Carbohydrate-coated GNPs have been proposed earlier for their relatively high stability, biocompatibility, and reduced environmental toxicity (Baetsen-Young et al., 2018b). Low molecular weight carbohydrates such as dextrin have been used for GNP synthesis at the Nano-Biosensors Laboratory, Michigan State University, via a one-pot synthesis route and offer longer shelf life. Dextrin-coated GNPs have been used earlier to rapidly detect unamplified DNA from E. coli O157:H7 and Pseudoperonospora cubensis (Baetsen-Young et al., 2018b; E. Dester et al., 2022). However, their application for detecting foodborne pathogens needs to be further explored. This study used these GNPs to detect E. coli, followed by their extraction from spinach and lettuce, using gMNP. Scheme 5.1 briefly describes the generalized procedure used in this study. 88 Scheme 5. 1. Generalized procedure for artificial contamination, bacterial extraction using gMNP, and detection using dextrin-capped GNPs. 5.2. Materials and Methods 5.2.1. Materials Frozen stock culture of Escherichia coli C-3000 (15597) was obtained from the American Type Culture Collection (ATCC), Salmonella enterica serovar Enteritidis, Klebsiella pneumoniae, and Enterobacter cloacae were obtained from the Nano-Biosensors Laboratory at Michigan State University. DNA extraction kits were purchased from Qiagen (Germantown, MD, USA). NanoDrop one (Thermo Scientific) was used to quantify DNA samples and absorption spectra (Waltham, MA, USA). Proprietary chitosan (glycan) functionalized magnetic nanoparticles synthesized at Nano-Biosensors Laboratory, Michigan State University, were used without modification. Whirl-Pak bags for bacterial extraction (12 oz. and 18 oz.) were purchased from VWR International (Radnor, PA, USA). Magnetic separation racks were purchased from Spherotech Inc (Lake Forest, IL, USA). Tryptic Soy Agar (TSA) and Tryptic Soy Broth (TSB) were purchased from Sigma Aldrich (St. Louis, MO, USA), and CHROMagar for E. coli was purchased from DRG International (Springfield, NJ) and were prepared according to the manufacturer’s instructions. Hydrochloric acid (HCl), gold (III) chloride (HAuCl4), sodium carbonate (Na2Co3), 11-mercaptoundecanoic acid (MUDA HS(CH2)10CO2H), sodium dodecyl sulfate (SDS, C12H25NaO4S), dextrin from potato starch, molecular biology grade agarose, Tris Acetate EDTA buffer, and Phosphate Buffer Solution (PBS), pH 7.4 were also purchased from Sigma Aldrich (St. Louis, MO, USA) and prepared according to manufacturer’s instructions. Grids (formvar/carbon 200 mesh copper) 89 for TEM were purchased from Electron Microscopy Systems, Hatfield, PA. Glutaraldehyde, cacodylate buffer and uranyl acetate stain used for TEM were provided by the Center for Advanced Microscopy (MSU CAM), Michigan State University. 5.2.2. Bacterial culture Stock cultures of each bacterial species were stored at -80 °C and were revived by plating on TSA plates at 37 °C for 24-48 h. Fresh bacterial cultures were grown overnight by transferring a single colony from TSA plates to 9 ml of TSB. 5.2.3. Isolation of bacterial cells from food matrices The food matrices chosen for this study include baby spinach and romaine lettuce. The leafy greens were purchased from a local seller and stored at 4 °C before use. The overall gMNP-extraction method from large-volume samples was earlier illustrated in Scheme 1. Overnight bacterial cultures were grown for 4 h, then diluted for inoculation in PBS and food samples. Bacteria inoculation was based on procedures from Bacteriological Analytical Manual (U.S. Food & Drug Administration). Briefly, 25 grams of spinach or lettuce in filter Whirl-Pak bags were contaminated with 1 mL of bacterial culture serially diluted until 10-4. The bags were incubated at room temperature for 1 h for acclimation. Serial dilutions of 10-5 and 10-6 were plated to confirm initial counts. Following incubation, 225 mL of PBS was added and homogenized in a stomacher. The liquified food matrix was separated into 2 bags with 100 mL each, one designated as a control to determine the initial bacterial concentration and the other to which gMNP was added (test). Addition of gMNP was followed by 5 min incubation and 5 min magnetic separation. After resuspending in 1 mL of PBS, both test and control samples were plated to confirm bacterial concentration; binding of bacterial cells with MNPs was also confirmed with TEM. Each food trial consisted of food matrix inoculated with target bacteria (E. coli), non-target bacteria (S. enterica), and a control with no inoculation to account for natural microflora. 5.2.4. Visualization of gMNP-bacteria binding Following the standard negative staining procedure, the gMNP-bacterial cell-binding was observed by a transmission electron microscope (TEM) (JEM-1400 Flash, Jeol, Nieuw-Vennep, Tokyo, Japan). The gMNP-bacteria solution obtained from PBS and food samples was first dissolved in 0.1 mL of the fixative solution (2.5% of glutaraldehyde in 0.1M cacodylate buffer). Then, 5 µL of the fixed bacteria was dropped onto the black side of a grid for 20-30 seconds before washing with 5 µL distilled water. Staining following drying was achieved using 5 µL of 0.1% uranyl acetate stain. The excess stain was 90 removed after 5-10 seconds and air-dried before being loaded into the TEM specimen holder. Images were taken in the range of 5000-25000x magnification. 5.2.5. DNA extraction DNA extraction from pure bacterial cultures grown overnight was achieved following the manufacturer’s instructions. For bacterial cells extracted from food matrices, 500 µL of gMNP/bacteria mixture was added to 4.5 mL of TSB and grown for 4-5 h at 37 °C. DNA quality and quantity were determined using NanoDrop one. Only samples with acceptable A260/A280 and A260/A230 ratios, between 1.8 and 2.2, were used for biosensor assay and PCR amplification. 5.2.6. Probe design and PCR confirmation Oligonucleotide probes targeting specifically the uid A gene with amination in the 5’ end and a poly-A tail were used. The probe sequence was obtained from a previous study (Srinivasan et al., 2011). Probe and PCR primers targeting the same gene were purchased from Integrated DNA Technologies (Coralville, IA, USA), and PCR protocol was adapted from (Srinivasan et al., 2011) and listed in a table below. PCR was conducted on DNA from pure E. coli culture and DNA samples extracted from leafy green samples to confirm biosensor results. Gel electrophoresis of PCR amplification product was performed in a 2% agarose gel in Tris Acetate EDTA (TAE) buffer at an applied voltage of 140 mV. Table 5. 1. Probe and primers used in this study. Gene Assay Sequence (5’-3’) Reference uid A Biosensor CAATGGTGATGTCAGCGTT (Srinivasan et al., 2011) uid A PCR F- GCAGTCTTACTTCCATGATTTCTTTA (Srinivasan et al., 2011) R- TAATGCGAGGTACGGTAGG 5.2.7. Synthesis of GNPs and Surface Modification Dextrin-capped GNPs used in this study were synthesized using the alkaline synthesis method previously described (Anderson et al., 2011). Briefly, gold (III) chloride trihydrate was added to water and neutralized with sodium carbonate. Next, dextrin was added under continuous stirring conditions at a temperature of 150 °C. The formation of gold nanoparticles was confirmed by the evolution of a wine- red color, following which the reaction was stopped. 91 The successful synthesis of GNPs was confirmed by their dark red appearance, and characterization was performed by obtaining their spectra using NanoDrop one, determining the wavelength of maximum absorbance. TEM was also used to determine the size of the synthesized GNPs. GNPs (5 µL) were dropped onto the black side of a grid for 20-30 seconds before washing with 5 µL distilled water. No staining was required for imaging. Synthesized GNPs were thiol coated using 25 µM MUDA and suspended in 0.1 M borate buffer. Batches of surface-modified GNPs were stored at 4 °C until further use. 5.2.8. GNP Biosensor Design Each biosensor trial included the addition of 5 µL of 25 µM DNA probe, 5 µL of GNPs, and 10 µL of DNA in a single tube before placing them in a thermocycler for probe hybridization. A cycle in the thermocycler consisted of subjecting the tubes to 95 °C for 5 min for denaturation, followed by 10 min at 55 °C for annealing, and finally cooling to room temperature. HCl (0.1 M) was then added to the tubes for aggregation of GNPs, turning them purple or gray in the absence of target DNA. Presence of target DNA results in the GNPs maintaining their red color. The change in color of GNPs was determined visually and by measuring their absorbance using a spectrophotometer in a wavelength range of 400 nm to 800 nm. The procedure for biosensor design was adapted from an earlier study (E. Dester et al., 2022). 5.2.9. Analytical Sensitivity and Specificity of GNP Biosensor The specificity of the biosensor was determined in 9 trials with 40 ng/µL of DNA. Each trial consisted of DNA from Escherichia coli (target) and 3 non-targets, including S. enterica, K. pneumoniae, and E. cloacae. Negative control with water was also included in each trial. DNA was diluted in elution buffer if required. Images of tubes and their absorption spectra were obtained within 5 min after the addition of 10 µL of 0.1 M HCl. Since a shift in the peak of wavelength at maximum absorption as a result of GNP aggregation is expected, results were analyzed by determining the ratio of absorbance at 625 nm and 520 nm. The sensitivity of the biosensor was also determined in 9 separate trials with DNA concentrations including 20 ng/µL, 10 ng/µL, 5 ng/µL, and 2.5 ng/µL. Each trail included DNA from the target and non- target at the same concentration and a negative control. As previously mentioned, the ratio of absorbance at 625 nm and 520 nm was determined for analyzing the results. 92 5.2.10. Detection of E. coli From Food Samples Since each food trial consisted of a matrix artificially contaminated with target and non-target bacteria, and a control with no contamination, DNA was extracted from each sample and used in the biosensor. While a difference in concentration of DNA extracted from different samples was observed, samples were diluted for consistency. The DNA samples were then used in the biosensor assay for detection, and results were analyzed by determining the ratio of absorbance at 625 nm and 520 nm. 5.2.11. Statistical Analysis All experiments in this study were run in triplicate, and data were displayed with averages and standard deviations. The ratio of absorbance at 625 nm and 520 nm was compared between target and non- target results at a 95% confidence interval. Comparison of multiple groups for specificity, sensitivity, and food testing was performed using One-way Analysis of Variance (ANOVA) and Tukey’s HSD (Honestly significant difference) test (p < 0.05) by statistical analysis software, Minitab (Version 16, Minitab Inc, State College, PA). 5.3. Results and Discussion 5.3.1. Confirmation of E. coli/gMNP binding in PBS and foods Successful binding between gMNP and E. coli was confirmed with TEM, as seen in Figure 5.1. The micrograph shows individual E. coli cells bound to gMNP; multiple gMNP binding to individual bacterial cells can be seen, represented by blue and black arrows, respectively. Attachment to curved ends of bacterial cells was found to be a common occurrence and confirmed earlier reports with S. Enteritidis, Listeria monocytogenes, and E. coli O157:H7 (E. Dester & Alocilja, 2022; Matta, Harrison, et al., 2018) . TEM was also used to confirm the binding of bacterial cells to gMNP in food matrices; sample images can be observed in Figure 5.1 (b) & (c). Although samples such as lettuce and spinach are relatively more complex, in comparison to PBS, gMNP was successful in binding and extraction of bacterial cells. The micrographs show individual cells to which multiple gMNP are bound. As seen with binding in PBS, attachment to curved ends of bacterial cells was observed. The use of non-specific gMNP for successful bacterial capture has been encountered in the past. For example, these were earlier employed to rapidly concentrate S. Enteritidis, B. cereus, and E. coli O157:H7 from PBS and milk samples (E. Dester & Alocilja, 2022; Matta & Alocilja, 2018). Elsewhere, the same gMNP was successful in extracting S. aureus and L. monocytogenes from milk, sausage, deli ham, and romaine lettuce. In the same study, E. coli O157:H7 was also successfully extracted from romaine 93 lettuce, spinach, chicken salad, and flour (E. F. Dester, 2022). Similar glycan-coated MNPs were used elsewhere to rapidly extract S. Enteritidis, Listeria monocytogenes, and E. coli O 157:H7 from homogenized egg, milk, and apple cider. The pathogens were further detected using an electrochemical biosensor (Matta, Harrison, et al., 2018). Separation of bacterial cells using non-specific aminated MNPs has also been seen earlier in literature. Positively charged MNPs, for example, were used to concentrate bacteria at ultra-low concentrations from buffer solution (Li et al., 2019). In another study, Huang et al., used non-specific aminated MNPs to capture E. coli, Pseudomonas aeruginosa, Staphylococcus aureus, and Staphylococcus lutea from buffer solution achieving a capture efficiency of > 90%. Following this, they successfully isolated bacteria from lake and river water, green tea, and grape juice (Y.-F. Huang et al., 2010); however, the efficiency of the MNPs in other carbohydrate-rich foods was not explored. The use of gMNP in a range of carbohydrate and fat-rich foods suggest their wide applicability with various food types. Rapid extraction with gMNP can then be combined with novel biosensor platforms for rapid pathogen detection. Successful binding between gMNP and E. coli from lettuce, and spinach samples was followed by its detection using GNP biosensor. Figure 5. 1. TEM micrographs showing binding between gMNP and E. coli in PBS (a) and successful extraction of bacterial cells from spinach (b) and lettuce (c). 5.3.2. GNP synthesis and Proof-of-concept of biosensor Synthesis of dextrin-capped GNPs for the biosensor was initially confirmed by the wine-red appearance of the product (Anderson et al., 2011); the inset in Figure 5.2 (a) shows a sample of as-synthesized GNPs in a tube. TEM was used to characterize the nanoparticles, and a micrograph can be seen in Figure 5.2 (a). The GNPs were found to be roughly spherical in shape and were less than 50 nm in diameter. Confirming earlier reports, the red color of synthesized GNPs indicates their size ranging from 10 to 50 nm in diameter (Ghosh & Pal, 2007; Ragavan et al., 2013). Absorption spectra of the samples were also 94 obtained in the wavelength range of 200 - 700 nm and can be observed in Figure 5.2 (b). The plot compares spectra of as-synthesized GNPs and surface-modified GNPs used for biosensor application; a peak at approximately 520 nm confirmed their small size. Surface-modified GNPs did not show a shift in wavelength of maximum absorption, indicating that modification did not affect the size of nanoparticles. Figure 5. 2. TEM micrograph of as-synthesized GNPs, inset shows GNPs in a tube. The absorption spectrum of GNPs, before and after surface modification, is shown in (b). A proof-of-concept of the biosensor can be seen in (c) with negative control (N.C.), target (T), and non-target (N.T.). A proof-of-concept for the biosensor was initially developed at a DNA concentration of 40 ng/µl. A typical biosensor assay involved the addition of aminated probe, surface-modified GNPs, and DNA into a tube, followed by its placement in a thermocycler for probe hybridization. HCl was then added to determine the stability of GNPs. Figure 5.2 (c) shows the tubes and their spectra following the addition of HCl with negative control (N.C.), target (T) DNA from E. coli, and non-target (N.T.) DNA from K. pneumoniae. As observed, tubes with target DNA remained stable, while those of non-target and negative control displayed aggregation. Spectra in the wavelength range of 400 to 850 nm of the same samples confirmed visual results, as seen in the figure. The spectrum of the samples from target showed 95 a sharp peak with a maximum absorption wavelength near 520 nm, indicating minimal GNP aggregation. The spectrum of sample from non-target and control indicated GNP aggregation, as was evident from a broader peak for the former and 2 broad peaks for the latter. This aggregation of samples from non- target and control was expected and confirmed the successful detection of target DNA by the biosensor. We hypothesize that the hybridization between GNP-probe and target DNA results in the protection of GNPs under the influence of HCl. While similar biosensor concepts have been seen in literature, longer reaction times and complicated surface modification routes have been primary limitations (Ahmadi et al., 2018; Andreadou et al., 2014; Bojd et al., 2017). Dextrin-capped GNPs used in this study provided for quicker assays and simplified surface-modification procedure for biosensor application. 5.3.3. Sensitivity and Specificity of GNP biosensor The specificity of the E. coli biosensor was confirmed against DNA from three non-targets- S. enterica, K. pneumoniae, and E. cloacae at a DNA concentration of 20 ng/µL. Figure 5.3 (a) shows the absorbance spectra of the control and samples from the target and the non-targets with images of tubes following HCl addition in the inset. As seen from the inset of the figure, the target sample remained red, while others showed an aggregation of GNPs. Spectra from the target sample show a sharp peak, indicating minimal aggregation, while the non-targets displayed aggregation at varying levels. Both NC and NT3 (E. cloacae) showed the highest aggregation among the samples displaying a broad spectrum and two peaks. Results from the biosensor were also analyzed by comparing the absorbance ratio at 625 nm and 520 nm and can be seen in Figure 5.3 (b) with standard deviation. As expected, control and NT3 showed the highest ratio, indicating maximum aggregation among the samples, followed by NT1 (S. enterica). Target samples showed the least ratio indicating minimal aggregation of GNPs. Specifically, the target group has a significantly lower ratio than the control and all non-targets. Although the ratio of samples from NT2 (K. pneumoniae) is seen to be closer to target samples, significant differences between target and NT2 were statistically confirmed by ANOVA followed by Tukey’s method (p < 0.05). Additionally, non-targets NT1 and NT2 are significantly different from each other and from control and NT3, which seem to have similar ratios. Overall, it was determined that the target could be significantly differentiated from non-targets and the control group. 96 Figure 5. 3. (a) Absorption spectrum of GNPs showing the specificity of the GNP biosensor for negative control (N.C.), target (T, E. coli), and non-targets (NT1, S. enterica; NT2, K. pneumoniae; NT3, E. cloacae). The ratio of absorbance at 625 nm and 520 nm for the same samples can be seen in (b). The sensitivity of the biosensor using K. pneumoniae was also determined at various DNA concentrations (c). Following the successful detection of target DNA, the biosensor sensitivity was determined with K. pnuemoniae as the non-target due to its closer ratio with the target (E. coli). DNA with concentrations of 40 ng/µL, 20 ng/µL, 10 ng/µL, and 5 ng/µL of both target and non-target were used for the analysis. The ratio of absorbance at 625 nm and 520 nm was determined and can be seen in Figure 5.3 (c). The difference in these ratios between target and non-target samples at 40 ng/µL and 20 ng/µL was higher, followed by 10 ng/µL. At 5 ng/µL, the least difference between the target and non-target was observed. A significant difference between target and non-target samples for each concentration was confirmed using ANOVA and Tukey’s test. The targets at 40 ng/µL, 20 ng/µL, and 10 ng/µL differed significantly from their non-target groups (p < 0.05). However, the target at 5 ng/µL was similar to its non-target group (p > 0.05). In addition, the target concentration at 20 ng/µL and 40 ng/µL had a similar ratio with no significant difference (p > 0.05). Based on the statistical difference, the detection limit was found to 97 be 10 ng/µL. This detection limit was lower than earlier studies for detecting Staphylococcus epidermis (Bojd et al., 2017) and Leishmania sps. (Andreadou et al., 2014), where 20 ng/µL and 11.5 ng/µL of target DNA were detected, respectively. Further, while detection of S. epidermis was achieved in 2-3 h, Leishmania detection involved the use of multiple probes (Andreadou et al., 2014). An earlier study using dextrin-capped GNPs was successful in obtaining a detection limit of 2.5 ng/µL with E. coli O157:H7 DNA, however, a longer probe was used. Detection of E. coli DNA from pure cultures was followed by detection from food matrices. 5.3.4. Detection of E. coli from food Detection of pathogens directly from food is critical for preventing foodborne outbreaks. Various techniques have been proposed to achieve this goal. Traditional biosensor platforms that have successfully detected pathogens from food matrices include surface plasmon resonance (SPR)-based (Waswa et al., 2007) (Zhou et al., 2018), electrochemical-based (Xu et al., 2016), spectroscopy (Petersen et al., 2021), spectrometry (Jadhav et al., 2018), and sequencing-based methods (Townsend et al., 2020). While these biosensor platforms were successful with various matrices, equipment used in these techniques is often not portable, reducing accessibility. Although portable versions of these techniques may be available, they are associated with increased costs. Other alternatives applicable in a low- resource setting include paper and smartphone-based detection platforms (Jung et al., 2020; Pang et al., 2018). Such systems, however, use antibodies that have drawbacks, including lower shelf-life. GNP- based colorimetric biosensor discussed previously can, to a significant extent, address concerns related to cost and accessibility in low-resource settings. The sections that follow discuss detection of E. coli directly from food matrices including spinach and lettuce. 5.3.5. Detection of E. coli from artificially contaminated Spinach Magnetic extraction of bacterial cells from spinach artificially contaminated with E. coli and S. enterica was followed by growth in TSB for 4-5 hours and extraction of DNA. As mentioned, a control with no inoculation was also included in the experiment to account for natural microflora. Once extracted, DNA was tested with the GNP biosensor for detection of E. coli, absorption spectra of the samples can be seen in Figure 5.4 (a). The plot compares the absorption spectra of the control, target (E. coli), non- target (S. enterica), and natural microflora (N.F.). As seen from the spectra and images of the tubes, minimum aggregation was observed for the target. Both natural microflora and negative control displayed significant aggregation of GNPs indicated by their broad spectra. The ratio of absorbance at 625 nm and 520 nm was again used to analyze the aggregation of GNPs, as seen in Figure 5.4 (b). The 98 figure compares the average ratio for samples from negative control, target, non-target, and natural flora with standard deviation. The lowest ratio was seen for the samples from the target, confirming successful detection. The target (E. coli) from spinach was significantly different (p < 0.05) from the non- target and natural microflora that were similar (p > 0.05). To confirm that the stability of GNPs was indeed due to the presence of DNA from E. coli, extracted DNA was PCR amplified, and the amplification product was run on a 2% agarose gel. Figure 5.4 (c) shows an amplification product for E. coli from pure culture as a positive control and samples from the target, non-target, and natural microflora with a 100 bp ladder in the first lane. Amplification of the sample in the lane for positive control was observed as expected. Amplification in the lane for the target confirmed the presence of E. coli DNA in the samples. No bands were observed in the lanes for non- target and natural microflora, indicating the absence of E. coli DNA in those samples. The results from gel electrophoresis confirmed those obtained using the GNP biosensor. Figure 5. 4. Detection of E. coli from spinach samples, (a) shows spectra of the target (T, E. coli), non- target (N.T., S.enterica), natural microflora (N.F.), and negative control (N.C.). The ratio of absorbance at 520 nm and 625 nm can be seen in (b). Gel electrophoresis following PCR amplification for the same samples confirmed amplification only in target samples (c). 99 5.3.6. Detection of E. coli from artificially contaminated Lettuce Following the successful detection of E. coli from spinach, its extraction and detection from lettuce were attempted. As with spinach, lettuce samples were inoculated with E. coli and S. enterica as target and non-target pathogens, followed by the growth of extracted cells in TSB. Additionally, a lettuce sample with no inoculation was included to account for natural microflora. DNA extracted from the samples was tested in the biosensor; images of tubes following the addition of HCl can be seen in the inset in Figure 5.5 (a). The figure shows the average absorption spectra of corresponding tubes, stability of the target samples is clearly seen with a single sharp peak with maximum absorption at approximately 525 nm. Figure 5. 5. Detection of E. coli from lettuce samples, (a) shows spectra of the target (T, E. coli), non- target (N.T., S. enterica), natural microflora (N.F.), and negative control (N.C.). The ratio of absorbance at 520 nm and 625 nm can be seen in (b). Gel electrophoresis following PCR amplification for the same samples confirmed amplification only in target samples (c). 100 Samples from both non-target and natural microflora showed aggregation of GNPs along with control, as expected. The ratio of absorption at 625 nm and 520 nm was then used to quantify the aggregation of GNPs (Figure 5.5 (b)). As can be seen from the figure, the lowest ratio was seen for the sample from the target, confirming visual results. Ratios of samples from control, non-target, and from natural microflora showed significantly higher ratios indicating GNP aggregation. Statistically significant differences between samples from the target and others were confirmed using ANOVA and Tukey’s test (p < 0.05). The target (E. coli) from lettuce was significantly (p < 0.05) different from the non-target and natural microflora that were found to be similar. To confirm that the stability of GNPs was indeed due to the presence of DNA from E. coli, extracted DNA was PCR amplified with primers targeting the uidA gene. Amplified products from the target, non-target, and natural microflora samples can be seen in Figure 5.5 (c), with a 100 bp ladder in lane 1. PCR product of DNA extracted from pure culture (positive control) and target sample showed amplification, while no amplification was seen for non-target and natural microflora samples. Negative control, as expected, did not show any amplification. Results from gel electrophoresis confirmed that the stability of GNPs was indeed due to the presence of E. coli DNA. Successful detection of E. coli from spinach and lettuce using gMNP and dextrin-capped GNPs confirmed a recent report where E. coli O157:H7 was detected from flour samples (E. Dester et al., 2022). PCR amplification of magnetically extracted samples from both spinach and lettuce confirmed results from the biosensor. Target samples from both the matrices tested positive using GNP biosensor and showed amplification following PCR. The data supports our initial hypothesis that presence of target DNA contributes towards protection of GNPs under the influence of HCl. Scheme 2 (a) illustrates our rendition of how the genomic structure of E. coli is protects the GNPs. In the absence of target DNA, no hybridization of probe-GNP takes place, resulting in GNP aggregation (Scheme 5.2 (b)). With the advantage of long-term storage at room temperature, gMNP can rapidly concentrate pathogens. Spinach and lettuce are both commonly associated with a significant amount of natural microflora, all readily captured by gMNP (data not shown). 101 Scheme 5. 2. Hypothetical illustration of protection of dextrin-capped GNPs by target DNA, due to hybridization with probe on GNPs (a) and aggregation of GNPs in the absence of target DNA (b). Plating on selective media was also conducted to confirm the successful capture of E. coli cells. Average colony counts of target E. coli cells were then used to determine the limit of detection from food samples and found to be approximately 6.3 X 102 CFU/mL and 5.2 X 102 CFU/mL from lettuce and spinach samples, respectively. An earlier study using gMNP allowed extraction of E. coli O157:H7 from flour samples at an approximate limit of 103 CFU/ml. Following incubation, the extracted pathogen was successfully detected using GNP biosensor (E. Dester et al., 2022). Isolation of bacteria from food using MNPs and detection using GNP biosensor has been observed previously in the literature. In an earlier study, for example, 19 Salmonella strains were detected using a similar GNP biosensor (Quintela et al., 2019). However, antibody-based immunomagnetic separation was used. The same group successfully detected multiple STEC strains of E. coli from pure culture, ground beef, and blueberries (Quintela et al., 2015). The authors achieved a low detection limit with multiple STEC strains. Their approach, however, involved DNA amplification before detection. Non- functionalized GNPs were used in another study to detect Salmonella species with PCR-amplified products from various food samples in less than 8 h (Prasad et al., 2011). Elsewhere, Wang et al., used a similar biosensor with unmodified GNPs to detect cucumber green mottle mosaic virus from infected leaves and fruits. Again, while a detection limit of 30 pg/ µL was obtained, DNA amplification was used for detection (Wang et al., 2017). While these studies were successful in pathogen detection from food, 102 use of PCR amplification can contribute towards increased costs. The use of carbohydrate coated nanoparticles allowed detection of unamplified DNA from food samples. 5.4. Conclusion In conclusion, a rapid and cost-effective platform for isolating and detecting E. coli from buffer solutions and leafy greens was successfully developed. The time required for pre-enrichment was reduced to approximately 4-5 hours, and bacterial detection was achieved in about 30 min following DNA extraction. This approach did not involve DNA amplification or antibodies, making it applicable in a low- resource setting. The applicability of this approach for detecting pathogens from other complex matrices needs to be explored. 103 REFERENCES Abayasekara, L. M., Perera, J., Chandrasekharan, V., Gnanam, V. S., Udunuwara, N. A., Liyanage, D. S., Bulathsinhala, N. E., Adikary, S., Aluthmuhandiram, J. V. S., Thanaseelan, C. S., Tharmakulasingam, D. P., Karunakaran, T., & Ilango, J. (2017). Detection of bacterial pathogens from clinical specimens using conventional microbial culture and 16S metagenomics: a comparative study. BMC Infectious Diseases, 17(1), 631. https://doi.org/10.1186/s12879-017-2727-8 Ahmadi, S., Kamaladini, H., Haddadi, F., & Sharifmoghadam, M. R. (2018). Thiol-Capped Gold Nanoparticle Biosensors for Rapid and Sensitive Visual Colorimetric Detection of Klebsiella pneumoniae. Journal of Fluorescence, 28(4), 987–998. https://doi.org/10.1007/s10895-018-2262-z Ahmed, A., Rushworth, J. V., Hirst, N. A., & Millner, P. A. (2014a). Biosensors for Whole-Cell Bacterial Detection. Clinical Microbiology Reviews, 27(3), 631–646. https://doi.org/10.1128/CMR.00120-13 Ahmed, A., Rushworth, J. V., Hirst, N. A., & Millner, P. A. (2014b). Biosensors for Whole-Cell Bacterial Detection. Clinical Microbiology Reviews, 27(3), 631–646. https://doi.org/10.1128/CMR.00120-13 Anderson, M. J., Torres-Chavolla, E., Castro, B. A., & Alocilja, E. C. (2011). One step alkaline synthesis of biocompatible gold nanoparticles using dextrin as capping agent. Journal of Nanoparticle Research, 13(7), 2843–2851. https://doi.org/10.1007/s11051-010-0172-3 Andreadou, M., Liandris, E., Gazouli, M., Taka, S., Antoniou, M., Theodoropoulos, G., Tachtsidis, I., Goutas, N., Vlachodimitropoulos, D., Kasampalidis, I., & Ikonomopoulos, J. (2014). A novel non- amplification assay for the detection of Leishmania spp. in clinical samples using gold nanoparticles. Journal of Microbiological Methods, 96(1), 56–61. https://doi.org/10.1016/j.mimet.2013.10.011 Baetsen-Young, A. M., Vasher, M., Matta, L. L., Colgan, P., Alocilja, E. C., & Day, B. (2018a). Direct colorimetric detection of unamplified pathogen DNA by dextrin-capped gold nanoparticles. Biosensors and Bioelectronics, 101(October 2017), 29–36. https://doi.org/10.1016/j.bios.2017.10.011 Baetsen-Young, A. M., Vasher, M., Matta, L. L., Colgan, P., Alocilja, E. C., & Day, B. (2018b). Direct colorimetric detection of unamplified pathogen DNA by dextrin-capped gold nanoparticles. Biosensors and Bioelectronics, 101(August 2017), 29–36. https://doi.org/10.1016/j.bios.2017.10.011 Bakthavathsalam, P., Rajendran, V. K., & Baquir Mohammed, J. A. (2012). A direct detection of Escherichia coli genomic DNA using gold nanoprobes. Journal of Nanobiotechnology, 10(1), 8. https://doi.org/10.1186/1477-3155-10-8 Bojd, M. O., Kamaladini, H., Haddadi, F., & Vaseghi, A. (2017). Thiolated AuNP probes and multiplex PCR for molecular detection of Staphylococcus epidermidis. Molecular and Cellular Probes, 34, 30–36. https://doi.org/10.1016/j.mcp.2017.04.006 Centers for Disease Control and Prevention. (2022). Reports of Selected E. coli Outbreak Investigations. https://www.cdc.gov/ecoli/outbreaks.html Dester, E., & Alocilja, E. (2022). Current Methods for Extraction and Concentration of Foodborne Bacteria with Glycan-Coated Magnetic Nanoparticles: A Review. Biosensors, 12(2), 112. https://doi.org/10.3390/bios12020112 104 Dester, E. F. (2022). Extraction, concentration, and detection of foodborne pathogens using glycan- coated magnetic nanoparticles and a gold nanoparticle colorimetric biosensor [Michigan State University]. https://www.proquest.com/pagepdf/2676659293?accountid=12598 Dester, E., Kao, K., & Alocilja, E. C. (2022). Detection of Unamplified E. coli O157 DNA Extracted from Large Food Samples Using a Gold Nanoparticle Colorimetric Biosensor. Biosensors, 12(5), 274. https://doi.org/10.3390/bios12050274 Dwivedi, H. P., & Jaykus, L.-A. (2011). Detection of pathogens in foods: the current state-of-the-art and future directions. Critical Reviews in Microbiology, 37(1), 40–63. https://doi.org/10.3109/1040841X.2010.506430 Ghosh, S. K., & Pal, T. (2007). Interparticle Coupling Effect on the Surface Plasmon Resonance of Gold Nanoparticles: From Theory to Applications. Chemical Reviews, 107(11), 4797–4862. https://doi.org/10.1021/cr0680282 Hahm, B.-K., Kim, H., Singh, A. K., & Bhunia, A. K. (2015). Pathogen enrichment device (PED) enables one-step growth, enrichment and separation of pathogen from food matrices for detection using bioanalytical platforms. Journal of Microbiological Methods, 117, 64–73. https://doi.org/10.1016/j.mimet.2015.07.016 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. (2015). World Health Organization Global Estimates and Regional Comparisons of the Burden of Foodborne Disease in 2010. PLOS Medicine, 12(12), e1001923. https://doi.org/10.1371/journal.pmed.1001923 Huang, J. Y., Henao, O. L., Griffin, P. M., Vugia, D. J., Cronquist, A. B., Hurd, S., Tobin-D’Angelo, M., Ryan, P., Smith, K., Lathrop, S., Zansky, S., Cieslak, P. R., Dunn, J., Holt, K. G., Wolpert, B. J., & Patrick, M. E. (2016). Infection with Pathogens Transmitted Commonly Through Food and the Effect of Increasing Use of Culture-Independent Diagnostic Tests on Surveillance — Foodborne Diseases Active Surveillance Network, 10 U.S. Sites, 2012–2015. MMWR. Morbidity and Mortality Weekly Report, 65(14), 368–371. https://doi.org/10.15585/mmwr.mm6514a2 Huang, Y.-F., Wang, Y.-F., & Yan, X.-P. (2010). Amine-Functionalized Magnetic Nanoparticles for Rapid Capture and Removal of Bacterial Pathogens. Environmental Science & Technology, 44(20), 7908– 7913. https://doi.org/10.1021/es102285n Jadhav, S. R., Shah, R. M., Karpe, A. V., Morrison, P. D., Kouremenos, K., Beale, D. J., & Palombo, E. A. (2018). Detection of Foodborne Pathogens Using Proteomics and Metabolomics-Based Approaches. Frontiers in Microbiology, 9(December), 1–13. https://doi.org/10.3389/fmicb.2018.03132 Jung, Y., Heo, Y., Lee, J. J., Deering, A., & Bae, E. (2020). Smartphone-based lateral flow imaging system for detection of food-borne bacteria E.coli O157:H7. Journal of Microbiological Methods, 168(October 2019), 105800. https://doi.org/10.1016/j.mimet.2019.105800 Law, J. W.-F., Ab Mutalib, N.-S., Chan, K.-G., & Lee, L.-H. (2015). Rapid methods for the detection of foodborne bacterial pathogens: principles, applications, advantages and limitations. Frontiers in Microbiology, 5(DEC), 1–19. https://doi.org/10.3389/fmicb.2014.00770 105 Lee, J.-H., Cho, H.-Y., Choi, H., Lee, J.-Y., & Choi, J.-W. (2018). Application of Gold Nanoparticle to Plasmonic Biosensors. International Journal of Molecular Sciences, 19(7), 2021. https://doi.org/10.3390/ijms19072021 Li, Z., Ma, J., Ruan, J., & Zhuang, X. (2019). Using Positively Charged Magnetic Nanoparticles to Capture Bacteria at Ultralow Concentration. Nanoscale Research Letters, 14(1), 195. https://doi.org/10.1186/s11671-019-3005-z Liu, Y., Cao, Y., Wang, T., Dong, Q., Li, J., & Niu, C. (2019). Detection of 12 Common Food-Borne Bacterial Pathogens by TaqMan Real-Time PCR Using a Single Set of Reaction Conditions. Frontiers in Microbiology, 10(FEB), 1–9. https://doi.org/10.3389/fmicb.2019.00222 Malhotra, S., Pham, D. S., Lau, M. P. H., Nguyen, A. H., & Cao, H. (2022). A Low-Cost, 3D-Printed Biosensor for Rapid Detection of Escherichia coli. Sensors, 22(6), 2382. https://doi.org/10.3390/s22062382 Mandal, P. K., Biswas, A. K., Choi, K., & Pal, U. K. (2011). Methods for Rapid Detection of Foodborne Pathogens: An Overview. American Journal of Food Technology, 6(2), 87–102. https://doi.org/10.3923/ajft.2011.87.102 Matta, L. L., & Alocilja, E. C. (2018). Carbohydrate Ligands on Magnetic Nanoparticles for Centrifuge-Free Extraction of Pathogenic Contaminants in Pasteurized Milk. Journal of Food Protection, 81(12), 1941–1949. https://doi.org/10.4315/0362-028X.JFP-18-040 Matta, L. L., Harrison, J., Deol, G. S., & Alocilja, E. C. (2018). Carbohydrate-Functionalized Nanobiosensor for Rapid Extraction of Pathogenic Bacteria Directly from Complex Liquids with Quick Detection Using Cyclic Voltammetry. IEEE Transactions on Nanotechnology, 17(5), 1006–1013. https://doi.org/10.1109/TNANO.2018.2841320 Matta, L. L., Karuppuswami, S., Chahal, P., & Alocilja, E. C. (2018). AuNP-RF sensor: An innovative application of RF technology for sensing pathogens electrically in liquids (SPEL) within the food supply chain. Biosensors and Bioelectronics, 111(March), 152–158. https://doi.org/10.1016/j.bios.2018.04.010 Mohammed, A. M., Rahim, R. A., Ibraheem, I. J., Loong, F. K., Hisham, H., Hashim, U., & Al-Douri, Y. (2014). Application of Gold Nanoparticles for Electrochemical DNA Biosensor. Journal of Nanomaterials, 2014, 1–7. https://doi.org/10.1155/2014/683460 Pang, B., Zhao, C., Li, L., Song, X., Xu, K., Wang, J., Liu, Y., Fu, K., Bao, H., Song, D., Meng, X., Qu, X., Zhang, Z., & Li, J. (2018). Development of a low-cost paper-based ELISA method for rapid Escherichia coli O157:H7 detection. Analytical Biochemistry, 542(February 2017), 58–62. https://doi.org/10.1016/j.ab.2017.11.010 Paniel, & Noguer. (2019). Detection of Salmonella in Food Matrices, from Conventional Methods to Recent Aptamer-Sensing Technologies. Foods, 8(9), 371. https://doi.org/10.3390/foods8090371 Peng, H., & Chen, I. A. (2019). Rapid Colorimetric Detection of Bacterial Species through the Capture of Gold Nanoparticles by Chimeric Phages. ACS Nano, 13(2), 1244–1252. https://doi.org/10.1021/acsnano.8b06395 Petersen, M., Yu, Z., & Lu, X. (2021). Application of Raman Spectroscopic Methods in Food Safety: A Review. Biosensors, 11(6), 187. https://doi.org/10.3390/bios11060187 106 Prabowo, B. A., Chang, Y.-F., Lai, H.-C., Alom, A., Pal, P., Lee, Y.-Y., Chiu, N.-F., Hatanaka, K., Su, L.-C., & Liu, K.-C. (2018). Rapid screening of Mycobacterium tuberculosis complex (MTBC) in clinical samples by a modular portable biosensor. Sensors and Actuators B: Chemical, 254, 742–748. https://doi.org/10.1016/j.snb.2017.07.102 Prasad, D., Shankaracharya, & Vidyarthi, A. S. (2011). Gold nanoparticles-based colorimetric assay for rapid detection of Salmonella species in food samples. World Journal of Microbiology and Biotechnology, 27(9), 2227–2230. https://doi.org/10.1007/s11274-011-0679-5 Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2015). Simultaneous direct detection of Shiga-toxin producing Escherichia coli (STEC) strains by optical biosensing with oligonucleotide- functionalized gold nanoparticles. Nanoscale, 7(6), 2417–2426. https://doi.org/10.1039/C4NR05869K Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2019). Simultaneous Colorimetric Detection of a Variety of Salmonella spp. in Food and Environmental Samples by Optical Biosensing Using Oligonucleotide-Gold Nanoparticles. Frontiers in Microbiology, 10(MAY), 1–12. https://doi.org/10.3389/fmicb.2019.01138 Ragavan, K. V., Selvakumar, L. S., & Thakur, M. S. (2013). Functionalized aptamers as nano-bioprobes for ultrasensitive detection of bisphenol-A. Chemical Communications, 49(53), 5960. https://doi.org/10.1039/c3cc42002g Shafiqa, A. R., Abdul Aziz, A., & Mehrdel, B. (2018). Nanoparticle Optical Properties: Size Dependence of a Single Gold Spherical Nanoparticle. Journal of Physics: Conference Series, 1083(1), 012040. https://doi.org/10.1088/1742-6596/1083/1/012040 Srinivasan, S., Aslan, A., Xagoraraki, I., Alocilja, E., & Rose, J. B. (2011). Escherichia coli, enterococci, and Bacteroides thetaiotaomicron qPCR signals through wastewater and septage treatment. Water Research, 45(8), 2561–2572. https://doi.org/10.1016/j.watres.2011.02.010 Townsend, A., Li, S., Mann, D. A., & Deng, X. (2020). A quasimetagenomics method for concerted detection and subtyping of Salmonella enterica and E. coli O157:H7 from romaine lettuce. Food Microbiology, 92(May), 103575. https://doi.org/10.1016/j.fm.2020.103575 U.S. Food & Drug Administration. (, February). Bacteriological Analytical Manual. https://www.fda.gov/food/laboratory-methods-food/bacteriological-analytical-manual-bam Vaisocherová-Lísalová, H., Víšová, I., Ermini, M. L., Špringer, T., Song, X. C., Mrázek, J., Lamačová, J., Scott Lynn, N., Šedivák, P., & Homola, J. (2016). Low-fouling surface plasmon resonance biosensor for multi-step detection of foodborne bacterial pathogens in complex food samples. Biosensors and Bioelectronics, 80, 84–90. https://doi.org/10.1016/j.bios.2016.01.040 Wang, L., Liu, Z., Xia, X., Yang, C., Huang, J., & Wan, S. (2017). Colorimetric detection of Cucumber green mottle mosaic virus using unmodified gold nanoparticles as colorimetric probes. Journal of Virological Methods, 243, 113–119. https://doi.org/10.1016/j.jviromet.2017.01.010 Waswa, J., Irudayaraj, J., & DebRoy, C. (2007). Direct detection of E. Coli O157:H7 in selected food systems by a surface plasmon resonance biosensor. LWT - Food Science and Technology, 40(2), 187–192. https://doi.org/10.1016/j.lwt.2005.11.001 Xu, M., Wang, R., & Li, Y. (2016). An electrochemical biosensor for rapid detection of E. coli O157:H7 with highly efficient bi-functional glucose oxidase-polydopamine nanocomposites and Prussian 107 blue modified screen-printed interdigitated electrodes. The Analyst, 141(18), 5441–5449. https://doi.org/10.1039/C6AN00873A Zhang, Y., Xu, C., Guo, T., & Hong, L. (2018). An automated bacterial concentration and recovery system for pre-enrichment required in rapid Escherichia coli detection. Scientific Reports, 8(1), 17808. https://doi.org/10.1038/s41598-018-35970-8 Zhao, X., Lin, C.-W., Wang, J., & Oh, D. H. (2014). Advances in Rapid Detection Methods for Foodborne Pathogens. Journal of Microbiology and Biotechnology, 24(3), 297–312. https://doi.org/10.4014/jmb.1310.10013 Zhou, C., Zou, H., Li, M., Sun, C., Ren, D., & Li, Y. (2018). Fiber optic surface plasmon resonance sensor for detection of E. coli O157:H7 based on antimicrobial peptides and AgNPs-rGO. Biosensors and Bioelectronics, 117(March), 347–353. https://doi.org/10.1016/j.bios.2018.06.005 108 Chapter 6. RAPID DETECTION OF SALMONELLA FROM MELONS ANDCUCUMBERS USING CARBOHYDRATE-COATED NANOPARTICLES This chapter has been published in Food Control (10.1016/j.foodcont.2023.109770) and is reprinted with permission. Copyright 2022 Elsevier Ltd. 6.1. Introduction Food-related outbreaks result in severe health and economic consequences across the globe. According to estimates, approximately 128,000 hospitalizations occur annually in the United States alone due to illnesses linked to food and beverages (Centers for Disease Control and Prevention, 2011). Escherichia coli and Salmonella are among the common pathogens linked with such outbreaks (J. Y. Huang et al., 2016; Liu et al., 2019), and their association with fresh produce needs attention. Salmonella can lead to gastrointestinal illnesses and salmonellosis. It is commonly spread through inadequate sanitization procedures or the consumption of raw or undercooked food (Kim et al., 2015) An earlier published report indicated that in North America, Salmonella was responsible for approximately half of the recorded outbreaks in the past millennium (Aiyedun et al., 2021) Common foods associated with Salmonella contamination include dairy and poultry products, meats, fresh vegetables, and fruits (Silva et al., 2018). Foods traditionally consumed raw and undergo minimal processing before consumption pose a significant risk. Some examples of fresh produce involved in Salmonella-related outbreaks include cucumbers, melons, and papayas, to name a few ("Reports of Selected Salmonella Outbreak Investigations, List of Selected Outbreak Investigations Linked to Food, by Year," 2020). Rapid isolation and detection of Salmonella from such foods can play an essential role in outbreak prevention. Isolation of pathogens from food matrices has commonly been achieved using techniques such as filtration and centrifugation. Although these methods have successfully separated pathogens, interference from the food matrix, ineffective separation, and long enrichment are primary concerns (Fukushima et al., 2007). A reliable alternative may be provided by magnetic nanoparticles (MNPs), which allow rapid concentration and extraction of pathogens (E. Dester & Alocilja, 2022) However, MNPs have traditionally been labeled with antibodies that have specific storage conditions and can be expensive (Ruigrok et al., 2011). Alternatively, carbohydrate-coated MNPs allow storage at room temperature and are cost-effective. For example, carbohydrate-based MNPs have been found to achieve extraction at 25% of the cost compared to antibody-based methods [9]. Some examples of carbohydrates that have been used to coat MNPs include mannose (El-Boubbou et al., 2007), short- 109 chain glucans (You et al., 2021a), biotinylated oligosaccharides (Yosief et al., 2013), and cysteine-glycans (Matta & Alocilja, 2018b) The cysteine-glycan-based gMNP has been used with various food types, including flour, milk, apple cider, and homogenized eggs (E. Dester et al., 2022; Matta, Harrison, et al., 2018; Matta & Alocilja, 2018a). In this study, gMNP was used to isolate S. enterica from cucumbers and melons. The gMNP-based bacterial extraction can be combined with selective detection platforms for pathogen identification. Traditional techniques that have been used for the identification of pathogens from foods include polymerase chain reaction-based assays (McKillip & Drake, 2004), pulse field gel electrophoresis (Neoh et al., 2019), and multi-locus variable number tandem repeat analysis (Muvhali et al., 2017), among others. While these methods are considered the gold standard for food pathogen detection, the procedures are cumbersome and can take significantly longer times for detection (Aiyedun et al., 2021). Whole genome sequencing-based methods have been developed recently, allowing rapid and effective pathogen detection but require expensive equipment unsuitable for low-resource settings [16] Among other methods, label-free detection of whole bacterial cells with point-of-need applications has also been achieved with spectroscopy-based techniques. Infrared spectroscopy, for example, is successful in strain-level differentiation of pathogens (Tian et al., 2021), and Raman spectroscopy has been shown to detect bacterial counts as low as 10 CFU/mL (Li et al., 2020). However, multivariate analysis of acquired data and effect of natural microbiota on detection are primary concerns with these techniques. Various biosensor platforms have also gained traction recently owing to their rapidity, sensitivity, and specificity, with electrochemical and surface plasmon-based being common examples (Castle et al., 2021; Ravindran et al., 2023). Nevertheless, reducing expenses associated with these and related platforms is crucial for their broad applicability. Optical biosensors that provide a visual output and do not need complicated equipment, such as those using gold nanoparticles (GNPs), can offer a reliable alternative. The surface plasmon resonance of GNPs has been widely used earlier to detect DNA from target bacteria and viruses rapidly (Draz & Shafiee, 2018; Marin et al., 2021). Although combining GNPs with multiple biosensor platforms has been commonly observed (Zhou et al., 2015), using GNPs alone for colorimetric DNA detection offers excellent prospects in areas with a scarcity of resources. This technique makes use of the stability of GNPs under unfavorable environments for DNA detection. Multiple variations of the procedure have been recorded where hybridization between single-stranded probe functionalized GNPs and target DNA results in their stability, allowing them to maintain their red color. The absence of DNA results in the aggregation of GNPs, leading to a color change to violet (Andreadou et al., 2014; 110 Bakthavathsalam et al., 2012). Previous works utilizing this detection method have successfully detected bacterial loads as low as 9 CFU/g, but lengthy probe-functionalization procedures and DNA amplification were needed (Quintela et al., 2019). In our study, no probe-functionalization or DNA amplification was required for detection. 6.1.1. Platform novelty and applicability Capping gMNP and GNPs with carbohydrates allows for room-temperature storage, affordable synthesis, and long shelf-life. The gMNP does not require specific recognition ligands for pathogen extraction, and dGNP are stable at high salt concentrations. These nanoparticles developed at Nano- Biosensors Laboratory, Michigan State University, were used to rapidly detect S. enterica from foods without PCR amplification, following magnetic extraction. Scheme 6.1 depicts a brief overview of the isolation and detection platform. Scheme 6. 1. Procedure overview showing pathogen extraction using gMNP and detection using dGNP. 6.2. Materials and Methods 6.2.1. Materials Frozen stock cultures of Salmonella enterica serovar Enteritidis, Klebsiella pneumoniae, and Enterobacter cloacae were obtained from the Nano-Biosensors Laboratory at Michigan State University (MSU). Escherichia coli C-3000 (ATCC 15597, strain derived from E. coli K-12) was purchased from American Type Culture Collection (ATCC). Kits for DNA extraction were purchased from Qiagen (Germantown, MD, USA). NanoDrop one from Thermo Scientific (Waltham, MA, USA) was used for DNA quantification and determination of absorption spectra. TEM (JEM-1400 Flash, Jeol, Nieuw-Vennep, Tokio, Japan) and supplies (Glutaraldehyde, cacodylate buffer, and uranyl acetate stain) were provided by the Center for Advanced Microscopy (CAM) at MSU. 111 Chitosan functionalized magnetic nanoparticles, synthesized at the Nano-Biosensors Lab, MSU, were used without modifications. For extraction of bacterial cells from food, Whirl-Pak bags (92 oz. and 18 oz.) were purchased from VWR International (Radnor, PA, USA), and racks for magnetic extraction were purchased from Spherotech Inc (Lake Forest, IL, USA). Phosphate Buffer Saline, Tryptic Soy Agar (TSA), and Broth (TSB) were purchased from Sigma Aldrich (St. Louis, MO, USA) and prepared following the manufacturer's instructions. Salmonella selective agar was purchased from CHROMAgar. Dextrin from potato starch, Hydrochloric acid, gold (III) chloride (HAuCl4), Tris Acetate EDTA buffer (40 mM Tris, 40 mM Acetate, 1 mM EDTA, pH 8.3) , sodium dodecyl sulfate (SDS, C12H25NaO4S), sodium carbonate (Na2Co3), 11-mercaptoundecanoic acid (MUDA HS(CH2)10CO2H), and molecular biology grade agarose were also purchased from Sigma Aldrich (St. Louis, MO, USA). Food samples were purchased from the local market. 6.2.2. Bacterial culture Stock cultures of each bacterial species stored at -80 °C were revived by plating and incubation at 37 °C for 24-48 h. Fresh bacterial cultures were grown overnight, at 37 °C and agitation at 500 rpm, by transferring a single colony from Tryptic Soy Agar plates to 9 mL of Tryptic Soy Broth and growing at 6.2.3. Artificial contamination of foods and their rapid concentration using gMNP For artificial contamination, overnight bacterial cultures were grown for 4 h, followed by serial dilution and inoculation in foods. Cucumbers and melons (25 grams) were separately contaminated with 1 mL of bacterial culture serially diluted until 104 CFU/mL, followed by incubation at room temperature for one hour for acclimation. Serial dilutions of 10-5 and 10-6 were plated to confirm initial counts. 225 mL of PBS was then added before grinding in a stomacher for each food sample bag. Filter in the Whirl-Pak bags allowed for rapid separation of liquified food samples. The samples were then separated into 2 bags with 100 mL each, and one was used as a control for determining initial bacterial concentration; the other served as a test to which 1 mL of gMNP (5 mg/mL) was added. The addition of gMNP was followed by incubation for 5 min and 5 min of magnetic separation. The concentration of bacterial cells was confirmed by plating both test and control samples. TEM was also used, following magnetic separation, to ensure binding between bacterial cells and gMNP. Each food trial consisted of a food matrix inoculated with target bacteria (S. enterica), non-target bacteria (E. coli), and a control with no inoculation to account for natural microbiota. 112 For TEM imaging from both pure culture and food matrices, 5 µL of the gMNP-bacteria solution was dropped onto the black side of a grid for 20-30 seconds before washing with 5 µL distilled water. Staining of bacterial cells was achieved using 0.1% uranyl acetate. Filter paper was used to remove the excess stain after 5-10 seconds, followed by air drying. The grids were loaded into the specimen holder of the TEM, and images were taken in the range of 5000-25000 X magnification. 6.2.4. DNA extraction Extraction of DNA from pure bacterial cultures was achieved using kits, following manufacturer instructions. For DNA extraction from gMNP-bound bacterial cells, 500 µL of gMNP/bacteria mixture was inoculated in 4.5 mL of TSB, followed by growth for 4-5 h. The total volume (5 mL) was used for DNA extraction following manufacturer instructions with modifications to allow the removal of gMNP. NanoDrop one was used to determine DNA quality and quantity. Samples with acceptable A260/A280 and A260/A230 ratios, between 1.8 and 2.2, were used for biosensor assay and PCR amplification. 6.2.5. Probe design, qPCR assay, and PCR amplification Single-stranded oligonucleotide probes used for target DNA detection from S. enterica were designed to target the inv A gene (Gene Accession Number: NP_461817.1) utilizing the design tools from NCBI BLAST (National Center for Biotechnology Information Basic Location Alignment Search Tool); E-values were checked to indicate that the gene sequence is specific. The probe was aminated in the 5' end along with a poly-A tail. For PCR amplification and qPCR, primers targeting the same gene were used. Both probe and primers were obtained from Integrated DNA Technologies (Coralville, IA, USA). Table 6.1 below shows the sequence of the probe and primers used. Following PCR amplification, the amplified product was run on a 2% agarose gel with a 100 bp ladder to confirm amplicon size (~500 bp). Gel electrophoresis was performed in Tris-Acetate-EDTA buffer at a voltage of 140 mV. Table 6. 1. Oligonucleotide probe and primer sequences used in this study. Gene Assay Sequence (5’-3’) Ref invA Biosensor Am- AAAAAAAAAAAATGAAGCCGATGCCGGTGAAATTATCGCCAC This study invA QPCR/ PCR F- CGGTGGTTTTAAGCGTACTCTT R- CGAATATGCTCCACAAGGTTA (Paião et al., 2013) 113 6.2.6 Synthesis of dGNP and surface modification Synthesis of dGNP using the alkaline synthesis method and their surface modification was achieved as described previously (Anderson et al., 2011). The successful synthesis of dGNP was confirmed by their dark red appearance and obtaining their spectra in the visible range. The size of the synthesized product was further characterized using Zetasizer (Malvern, nano-zs). Synthesized dGNP were thiol coated using 25 µM MUDA and suspended in 0.1 M borate buffer. Batches of surface-modified dGNP were stored at 4 °C until further use. Spectrophotometric analysis allowed use of small volumes. For the detection of DNA, 5 µL of 25 µM DNA probe and 5 µL of GNPs in a tube were mixed with 10 µL of DNA and placed in a thermocycler to allow hybridization of the probe. Each thermocycler cycle involved heating the tubes to 95 °C for 5 min to denature the DNA, followed by 10 min at 55 °C for annealing. Tubes were then cooled to room temperature and maintained at 4 °C. Probe hybridization was followed by HCl addition to allow GNP aggregation, turning them purple or gray in the absence of target DNA. The presence of target DNA resulted in the dGNP maintaining their red color. Both visual and spectrophotometric analysis was done to confirm a change in the color of dGNP. The absorbance of samples in a wavelength range of 400 nm to 800 nm was determined. The procedure for biosensor design was adopted from an earlier study (E. Dester et al., 2022). 6.2.7. Sensitivity and specificity of biosensor Following the successful working of the GNP biosensor, its sensitivity and specificity were determined. A series of 9 trials with 40 ng/µL of DNA were used to confirm the specificity of the biosensor, each trial consisting of DNA from S. enterica (target) and 3 non-targets including E. coli, K. pneumoniae, and Enterobacter cloacae. Each trial also consisted of a control with water. Images of tubes and their absorption spectra were obtained 5 min after adding 4 µL of 0.5 M HCl. Since a shift in the peak of wavelength at maximum absorption as a result of dGNP aggregation is expected, results were analyzed by determining the shift in wavelength of maximum absorption. Quantification of dGNP aggregation was achieved by determining ratios of absorbance at 625 nm and 520 nm. As with specificity, the sensitivity of the biosensor was also determined in 9 trials with DNA concentrations of target and non-target (E. coli) at 20 ng/µL, 10 ng/µL, 5 ng/µL, and 2.5 ng/µL. A control with water was also used in each trial. Absorbance ratios at 625 and 520 nm were again used to quantify dGNP aggregation. 114 6.2.8. DNA detection from food samples Each food trial consisted of a matrix artificially contaminated with the target (S. enterica) and non-target (E. coli) bacteria, along with a control with no inoculation. DNA was extracted from each sample for use in the biosensor; samples were diluted for consistency. The DNA samples were then used in the biosensor assay for detection, and results were analyzed visually and by determining spectra of dGNP following HCl addition. 6.2.9. Statistical Analysis All experiments in this study were conducted in triplicate. The differences in absorbance ratio at 625 nm and 520 nm among target and non-target samples were analyzed at a 95% confidence interval. The data were first checked in terms of normality and equality of variance, then the appropriate statistical test (parametric or non-parametric) was applied to compare the groups. For sensitivity and specificity results, the differences in the absorbance ratio among target and non-target groups were analyzed using One-way Analysis of Variance (ANOVA) followed by Tukey's HSD (Honestly significant difference) test (p < 0.05). However, assumptions of normality and equality of variances of data from the food samples were not met for the parametric ANOVA test. Thus, the Kruskal-Wallis H test followed by post-hoc Dunn's test was used to analyze differences in the absorbance ratio between target and non-target samples from food (p < 0.05). 6.3. Results 6.3.1. S. enterica can be successfully extracted from fresh produce The successful binding between gMNP and S. enterica in PBS was initially confirmed using TEM by visualizing magnetically separated cells. Figure 6.1(a) shows 2 individual cells to which gMNP are bound, indicated by blue and black arrows, respectively; individual gMNP was observed to be roughly spherical. The binding of cells with gMNP was commonly seen along the curved ends of bacterial cells, confirming earlier reports with S. enterica, E. coli O157:H7, and L. monocytogenes (Matta, Harrison, et al., 2018). Previously published literature has suggested that this binding with bacterial cells may result from multiple forces. The random Brownian motion of cells, long-range electrostatic interactions, and glycan- protein-based site-specific electrostatic attachment were hypothesized to synergistically aid the attachment of bacterial cells to gMNP (E. Dester & Alocilja, 2022). Although recognizing antibodies and phages conjugated with MNPs have been widely seen in the literature for rapid bacterial capture, non-specific MNPs, similar to gMNP, have also been encountered. 115 Among common examples of such MNPs are those surface-functionalized with amine groups that are efficient at capturing bacteria at concentrations as low as 10 CFU/mL (Li et al., 2019) Huang et al., in a different study, used similar aminated MNPs to rapidly capture bacteria from various beverages, including water, grape juice, and green tea. E. coli, P. aeruginosa, and S. aureus were successfully captured at an efficiency of > 90% (Y.-F. Huang et al., 2010), although no solid food matrices were included in their study. may also have contributed to differences in bacterial extraction (Cano-Sarmiento et al., 2018). Figure 6. 1. TEM showing binding between gMNP and S. enterica in PBS with blue and black arrows indicating bacterial cells and gMNP, respectively (a), bacterial cells extracted from cucumbers and melons can be seen in (b) and (c). Mean Ct following qPCR was determined, shown in (d) along with amplification plots (e). Confirmation of the successful isolation of bacterial cells from PBS was followed by their extraction from artificially contaminated cucumbers and melons, visualized in Figure 1 (b & c). Although the microstructure of the tested foods is significantly more complex than PBS, it did not prevent the gMNP binding. Amplification using PCR was initially done to confirm the extraction, gel electrophoresis results 116 are shown in Fig 1 (d). The figure shows samples from target (T), non-target (NT), and natural microbiota (NF) along with 100 bp ladder, negative control with no DNA (NC), and positive control (PC) with DNA from pure culture. As observed, samples from only positive control and target amplified while no amplification is seen for other samples. Quantitative PCR (qPCR) was used to confirm the successful concentration of S. enterica; the average cycle of threshold (Ct) for gMNP-extracted samples and non- gMNP control is shown in Figure 1 (d). The Ct, which is inversely related to the concentration of target DNA present, was lower for gMNP-extracted samples. A difference in the Ct value for the two food matrices could be attributed to their surface properties and chemical composition. Both cucumbers and melons are rich in water, but melons are characterized by relatively high carbohydrate content (Olawuyi & Lee, 2019; Perkins-Veazie et al., 2012). Further, while the pH of the tested foods does not vary significantly, electrostatic binding between gMNP and charged food particles is a possibility. The gMNP were earlier successful in the capture of E. coli O157:H7, S. Enteritidis, and B. cereus from various milk types, including Vitamin D, 2% reduced fat, and fat-free (Matta & Alocilja, 2018a). Similar gMNPs were also used to rapidly concentrate E. coli O157:H7, S. Enteritidis, and L. monocytogenes from apple cider and homogenized eggs (Matta, Harrison, et al., 2018). The extraction of S. enterica from solid foods used in this study confirmed their application in a wide range of edibles. While gMNP contributes towards the simple extraction of bacterial cells in low-resource settings, their rapid and affordable detection is equally important. The following sections discuss the detection of S. enterica DNA using a dGNP biosensor. Biosensor sensitivity, specificity, and its application for detecting S. enterica extracted from cucumber and melons are detailed. 6.3.2. Synthesis of dGNP and Biosensor Proof-of-concept The biosensor was initially used to detect S. enterica DNA from a pure culture. The dGNP were synthesized using an alkaline synthesis route, and successful synthesis was confirmed by their wine-red appearance (Anderson et al., 2011) The inset in Figure 6.2 (a) shows an image of as-synthesized dGNP, with their wine-red appearance indicating a size range between 10 – 50 nm (Ghosh & Pal, 2007; Ragavan et al., 2013) The size of dGNP was confirmed using dynamic light scattering; Figure 6.2 (a) shows a plot with signal intensity (%) on the y-axis and nanoparticle size (nm) on the x-axis. The size distribution of as-synthesized dGNP followed a roughly normal distribution with an average size of approximately 30 nm. The nanoparticles were initially surface-modified with mercaptoundecanoic acid to prepare for binding with target-specific single-stranded probes. Neither surface-modification nor binding with probes resulted in a change in dGNP size, confirmed spectrophotometrically by determining their 117 absorption maxima (Figure 6.2 (b)) (X. Huang & El-Sayed, 2010). The dGNP were subjected to multiple wash steps to eliminate unbound probes, and no shift in peak wavelength was seen. The successful binding of probes to dGNP was confirmed with a peak at approximately 260 nm, indicated by a black arrow in the figure. Figure 6. 2. Size distribution of synthesized dGNP, inset shows dGNP in a tube (a). The absorption spectrum of dGNP before and after probe functionalization is shown in (b). A proof-of-concept of the biosensor can be seen in (c) with negative control (NC), target (T), and non-target (NT). A proof-of-concept for the biosensor was first tested with target DNA from S. enterica and non-target DNA from E. coli both at a concentration of 40 ng/µL and negative control with water. As mentioned previously, a typical assay included mixing the aminated probe, surface-modified dGNP, and sample DNA, followed by placing the mixture in a thermocycler for probe hybridization. The stability of dGNP was then determined under the influence of HCl. Samples were analyzed visually and by obtaining their absorption spectra in the visible region (Figure 6.2(c)). As observed, the sample with target DNA from S. enterica (T) showed minimum aggregation of dGNP, followed by non-target (NT) and negative control (NC). Absorbance spectra showed that samples with non-target DNA and negative control displayed a right shift in absorption maxima away from 520 nm. Aggregation of samples from non-target and control was expected and confirmed the contribution of target DNA towards the stability of dGNP. Biosensor concepts making use of the stability of GNPs under unfavorable environments have been proposed before but have one or more drawbacks. Acinetobacter baumannii, for example, was detected from clinical samples using unmodified GNPs, and a detection limit of 0.8125 ng/µL was achieved (Khalil et al., 2014). However, the DNA used in the assay was PCR-amplified. Another similar study successfully detected Mycobacterium tuberculosis from clinical samples within 2-4 h but utilized nested-PCR before detection (Soo et al., 2009). In our study, however, the target DNA was not amplified, allowing for detection in <30 min. 118 6.3.3. Analytical Sensitivity and Specificity of dGNP biosensor The specificity of the S. enterica biosensor was confirmed using DNA from 3 non-target bacteria, including E. coli C-3000, K. pneumoniae, and E. cloacae. The average spectra in the wavelength range of 400 to 750 nm are shown in Figure 6.3 (a); tube images are shown in the inset. As can be observed, the target sample (T) indeed showed the lowest aggregation of dGNP, with a minimal shift from 520 nm. In contrast, the non-target samples (NT1, NT2, and NT3) showed relatively higher dGNP aggregation. The spectra confirmed the visual results of the tubes where the target sample showed a red color and non- target samples changed to blue or gray. The average ratio of absorbance at 625 nm and 520 nm (A625/520) was also obtained to quantify the aggregation of dGNP (Figure 6.3 (b)). The target sample showed the lowest ratio, followed by K. pneumoniae (NT2), E. coli (NT1), and E. cloacae (NT3), in that order. Differences in the A625/520 among the negative control, target, and non-target samples were analyzed using ANOVA followed by Tukey's test (p < 0.05). The target samples were significantly different from the negative control and non-target samples (p < 0.05). Thus, the visual results from the target group, confirmed by A625/520, were found to be significantly different from non-targets and the control group. The successful differentiation of S. enterica DNA from other non-targets was followed by determining biosensor sensitivity, using E. coli as the non-target. DNA concentrations of 20 ng/µL, 10 ng/µL, 5 ng/µL, and 2.5 ng/µL were used to ascertain the limit of detection. The dGNP aggregation was quantified by determining the A625/520 (Figure 6.3 (c)). The difference in the average ratios of target and non-target samples was higher at 20 ng/µL and 10 ng/µL followed by 5 ng/µl. At 2.5 ng/µl, the least difference between the target and non-target samples was observed. Significant differences between target and non-target samples for each concentration were assessed using ANOVA and Tukey's test. The target samples at 20 ng/µL, 10 ng/µL, and 5 ng/µL were significantly different from their non-target samples (p < 0.05). Based on the statistical differences, the detection limit was found to be 5 ng/µL. This limit of unamplified DNA detection was lower than similar assays developed earlier. For instance, an assay for detecting S. epidermis followed a similar approach, although using thiolated GNPs, achieving a limit of 20 ng/µL (Bojd et al., 2017). Elsewhere, Leishmania sps. were detected using multiple probes with DNA as low as 11.5 ng/µL from blood samples (Andreadou et al., 2014). In other examples where a significantly lower detection limit was observed, amplification of DNA was necessary (Khalil et al., 2014; Soo et al., 2009). Following target DNA detection from pure culture, our biosensor was used to detect S. enterica extracted from foods. 119 Figure 6. 3. (a) Absorption spectrum of dGNP showing the specificity of the biosensor with negative control (NC), target (T, S. enterica), and non-targets (NT1, E. coli; NT2, K. pneumoniae; NT3, E. cloacae, NT4, B.cereus; NT5, L. monocytogenes). The ratio of absorbance at 625 nm and 520 nm for the same samples can be seen in (b). The sensitivity of the biosensor using E. coli was also determined at various DNA concentrations (c). 6.3.4. Detection of Salmonella from artificially contaminated fresh produce. Identifying pathogens directly from food matrices with minimal sample pre-enrichment is crucial for outbreak prevention. Following the isolation of bacterial cells from food samples using gMNP, DNA extracted from captured cells was detected with dGNP biosensor. Cucumbers and melons were artificially contaminated with S. enterica and E. coli to serve as target and non-target, respectively. A control with no inoculation was also included to account for natural microbiota. A 4-5 h pre-enrichment preceded the DNA extraction from all the samples. The insets in Figure 6.4 (a and b) show tubes after adding HCl for samples from cucumbers and melons, respectively. The target samples (T) maintained their red color, while samples from both non-target (NT) and natural microbiota (NF) from both food matrices displayed dGNP aggregation. The aggregation was again quantified by comparing A625/520 (Figure S6.1 and S6.2); target samples had the lowest ratio compared to non-target and natural microbiota samples. The differences between target and non-target samples were separately analyzed using the Kruskal-Wallis and Dunn's tests (p < 0.05). The target samples (S. enterica) were all significantly (p < 0.05) different from the non-target samples (E. coli) and natural microbiota from both food matrices. 120 Figure 6. 4. Absorbance spectra following HCl addition for samples from cucumber (a) and melons (b). Results from PCR amplification and gel electrophoresis (Figure 6.1 (d)) confirmed that the stability of dGNP was indeed due to the presence of S. enterica DNA. Only the target samples showed PCR amplification of the invA gene while no amplification was seen for either natural microbiota or non- target samples, in agreement with biosensor results. Using MNPs to concentrate and extract pathogens from food and their rapid detection has been widely encountered (Table 6.2). Traditional biosensor techniques, including SPR-based, electrochemical, and impedimetric platforms, have all successfully detected pathogens from food following MNP-based extraction, with low detection limits. An impedance biosensor, for example, was used to detect Salmonella in 2 h from spiked chicken samples using the oxidation of glucose oxidase (F. Huang et al., 2021) Other technologies have also been used, such as nuclear magnetic resonance (T. Li et al., 2020) and electrical response-based biosensors (Hou et al., 2019). Apart from employing antibody-based magnetic extraction, these methods have separate equipment requirements, adding towards costs and limiting field applicability. Lateral flow immunoassay platforms developed previously eliminated expensive equipment but still used antibodies for magnetic extraction and detection from foods. Although these studies successfully detected whole Salmonella cells, reducing expenses associated with the techniques is highly desirable and could be addressed using a GNP biosensor. The use of GNPs for Salmonella detection from food as a low-cost approach, similar to this study, has been recorded earlier, where 19 strains of Salmonella were simultaneously identified following IMS. Blueberries and chicken meat were the matrices of focus (Quintela et al., 2019). In another related study, multiple strains of E. coli were detected from ground beef and blueberries (Quintela et al., 2015). Elsewhere, a similar approach was applied to detect cucumber green mottle mosaic virus from infected 121 leaves (Wang et al., 2017). It is noteworthy, however, that all these studies required amplification of target DNA before being detected by a GNP biosensor, while no amplification was needed in our study. Although kits were used in this study, use of crude methods for DNA extraction can be more beneficial. 122 1 Table 6. 2. Biosensor platforms used to extract and detect the foodborne pathogens from food matrices. Biosensor Target bacteria Food matrix Separation method Detection limit Detection Equipment Ref. platform time/ Total requirement Assay time Electrochemical S. Typhimurium Chicken SPR S. Typhimurium Lettuce IMS IMS 102 CFU/mL 2 h 0.9 log CFU/g NA / >24 h Lateral flow S. Typhimurium Milk Aptamer-based 102 CFU/mL NA/ > 1.5 h Smartphone E. coli O157:H7 Yogurt, egg Magnetic beads/ 10 CFU/mL 2 h fluorescence antibody-based Colorimetric E. coli O157:H7 Sausage Lysine-SCG-based 30.8 CFU/mL <1 h enzyme-based Colorimetric Salmonella Chicken Colorimetric GNP- STEC Ground beef based and blueberries IMS IMS 10 CFU/mL >9 h <9 CFU/g 9 h following PCR Yes Yes No No No No No (Wang et al., 2020) (Bhandari et al., 2019) (Gao et al., 2021) (Zeinhom, Wang, Song, et al., 2018) (S.-M. You et al., 2021) (Quintela et al., 2019) (Quintela et al., 2015) Colorimetric GNP- Cucumber green Cucumber NA 9 pg/µL > 60 min No (Wang et al., based mottle mosaic virus seeds and Colorimetric GNP- E. coli O157:H7 based leaves Flour following RT- PCR gMNP-based 103 CFU/mL <6 h Plasmonic/Colorim S. enterica Cucumbers gMNP-based 102 CFU/mL <7 h etric GNP-based and melons 2017) ( Dester, 2022) This study No No 123 The gMNP and dGNP used in our work previously detected E. coli O157: H7 from flour samples with a detection limit of 103 CFU/mL (E. Dester et al., 2022) Notably, the gMNP are not S. enterica specific, and their charged nature makes capturing non-target bacterial cells and food microparticles inevitable. The presence of natural microbiota was confirmed by plating magnetically separated samples on Salmonella selective plates, and variability in their counts was observed based on the food matrix type (Fig. S6.3). However, specific detection as low as 4 X 102 CFU/g in the presence of natural microbiota was achieved using our biosensor; the presence of non-target bacteria did not significantly affect the assay. Although contaminated with non-target DNA, the hybridization between target DNA and probes conjugated with dGNP possibly resulted in their protection. Scheme 6.2 shows our underlying hypothesis regarding the protection of dGNP in the presence of HCl. Scheme 6. 2. The underlying hypothesis for the protection of dGNP due to hybridization between the target and probe on dGNP. The presence of non-target DNA in the mixture did not affect the stability of dGNP. The absence of target DNA resulted in the aggregation of dGNP. Allowing rapid visual detection in < 30 min without PCR amplification, the colorimetric nature of our platform is noteworthy. The dGNP were prepared using a one-pot method and did not require additional steps for probe-conjugation. Additionally, gMNP-based magnetic extraction costs significantly less than antibody-based approaches. Neither gMNP nor dGNP have specific storage requirements. Previous studies determined the cost of detection to be as low as $0.50 per test (Briceno et al., 2019), enabling the application of this platform on a commercial scale. 6.4. Conclusion Rapid and cost-effective detection of pathogens from food can aid in preventing outbreaks. Carbohydrate-coated nanoparticles were used in this study to extract and detect S. enterica from fresh 124 produce without needing expensive recognition moieties. The time required from sample processing through gMNP isolation to DNA extraction and dGNP-based detection was approximately 7 h. No DNA amplification or cold storage for the nanoparticles was needed, making this platform applicable in low- resource settings. Future work may include detecting other pathogens from complex food matrices. 125 REFERENCES Aiyedun, S. O., Onarinde, B. A., Swainson, M., & Dixon, R. A. (2021). Foodborne outbreaks of microbial infection from fresh produce in Europe and North America: a systematic review of data from this millennium. International Journal of Food Science & Technology, 56(5), 2215–2223. https://doi.org/10.1111/ijfs.14884 Alleweldt, F., Kara, Ş., Best, K., Aarestrup, F. M., Beer, M., Bestebroer, T. M., Campos, J., Casadei, G., Chinen, I., Van Domselaar, G., Dominguez, C., Everett, H. E., Fouchier, R. A., Grant, K., Green, J., Höper, D., Johnston, J., Koopmans, M. P., Oude Munnink, B. B., … Wylezich, C. (2021). Economic evaluation of whole genome sequencing for pathogen identification and surveillance – results of case studies in Europe and the Americas 2016 to 2019. Eurosurveillance, 26(9). https://doi.org/10.2807/1560-7917.ES.2021.26.9.1900606 Anderson, M. J., Torres-Chavolla, E., Castro, B. A., & Alocilja, E. C. (2011). One step alkaline synthesis of biocompatible gold nanoparticles using dextrin as capping agent. Journal of Nanoparticle Research, 13(7), 2843–2851. https://doi.org/10.1007/s11051-010-0172-3 Andreadou, M., Liandris, E., Gazouli, M., Taka, S., Antoniou, M., Theodoropoulos, G., Tachtsidis, I., Goutas, N., Vlachodimitropoulos, D., Kasampalidis, I., & Ikonomopoulos, J. (2014). A novel non- amplification assay for the detection of Leishmania spp. in clinical samples using gold nanoparticles. Journal of Microbiological Methods, 96(1), 56–61. https://doi.org/10.1016/j.mimet.2013.10.011 Bakthavathsalam, P., Rajendran, V. K., & Baquir Mohammed, J. A. (2012). A direct detection of Escherichia coli genomic DNA using gold nanoprobes. Journal of Nanobiotechnology, 10(1), 8. https://doi.org/10.1186/1477-3155-10-8 Bhandari, D., Chen, F.-C., & Bridgman, R. C. (2019). Detection of Salmonella Typhimurium in Romaine Lettuce Using a Surface Plasmon Resonance Biosensor. Biosensors, 9(3), 94. https://doi.org/10.3390/bios9030094 Bojd, M. O., Kamaladini, H., Haddadi, F., & Vaseghi, A. (2017). Thiolated AuNP probes and multiplex PCR for molecular detection of Staphylococcus epidermidis. Molecular and Cellular Probes, 34, 30–36. https://doi.org/10.1016/j.mcp.2017.04.006 Briceno, R. K., Sergent, S. R., Benites, S. M., & Alocilja, E. C. (2019). Nanoparticle-Based Biosensing Assay for Universally Accessible Low-Cost TB Detection with Comparable Sensitivity as Culture. Diagnostics, 9(4), 222. https://doi.org/10.3390/diagnostics9040222 Cano-Sarmiento, C., Téllez-Medina, D. I., Viveros-Contreras, R., Cornejo-Mazón, M., Figueroa-Hernández, C. Y., García-Armenta, E., Alamilla-Beltrán, L., García, H. S., & Gutiérrez-López, G. F. (2018). Zeta Potential of Food Matrices. Food Engineering Reviews, 10(3), 113–138. https://doi.org/10.1007/s12393-018-9176-z Castle, L. M., Schuh, D. A., Reynolds, E. E., & Furst, A. L. (2021). Electrochemical Sensors to Detect Bacterial Foodborne Pathogens. ACS Sensors, 6(5), 1717–1730. https://doi.org/10.1021/acssensors.1c00481 Centers for Disease Control and Prevention. (2011). CDC Estimates of Foodborne Illness in the United States. CDC 2011 Estimates. http://www.cdc.gov/foodborneburden/2011-foodborne- estimates.html 126 Centers for Disease Control and Prevention. (2020). Reports of Selected Salmonella Outbreak Investigations, List of Selected Outbreak Investigations Linked to Food, by Year. Centers for Disease Control and Prevention. https://www.cdc.gov/salmonella/outbreaks.html Dester, E., & Alocilja, E. (2022). Current Methods for Extraction and Concentration of Foodborne Bacteria with Glycan-Coated Magnetic Nanoparticles: A Review. Biosensors, 12(2), 112. https://doi.org/10.3390/bios12020112 Dester, E. F. (2022). Extraction, concentration, and detection of foodborne pathogens using glycan- coated magnetic nanoparticles and a gold nanoparticle colorimetric biosensor [Michigan State University]. https://www.proquest.com/pagepdf/2676659293?accountid=12598 Dester, E., Kao, K., & Alocilja, E. C. (2022). Detection of Unamplified E. coli O157 DNA Extracted from Large Food Samples Using a Gold Nanoparticle Colorimetric Biosensor. Biosensors, 12(5), 274. https://doi.org/10.3390/bios12050274 Draz, M. S., & Shafiee, H. (2018). Applications of gold nanoparticles in virus detection. Theranostics, 8(7), 1985–2017. https://doi.org/10.7150/thno.23856 El-Boubbou, K., Gruden, C., & Huang, X. (2007). Magnetic Glyco-nanoparticles: A Unique Tool for Rapid Pathogen Detection, Decontamination, and Strain Differentiation. Journal of the American Chemical Society, 129(44), 13392–13393. https://doi.org/10.1021/ja076086e Fukushima, H., Katsube, K., Hata, Y., Kishi, R., & Fujiwara, S. (2007). Rapid Separation and Concentration of Food-Borne Pathogens in Food Samples Prior to Quantification by Viable-Cell Counting and Real- Time PCR. Applied and Environmental Microbiology, 73(1), 92–100. https://doi.org/10.1128/AEM.01772-06 Gao, P., Wang, L., He, Y., Wang, Y., Yang, X., Fu, S., Qin, X., Chen, Q., Man, C., & Jiang, Y. (2021). An Enhanced Lateral Flow Assay Based on Aptamer–Magnetic Separation and Multifold AuNPs for Ultrasensitive Detection of Salmonella Typhimurium in Milk. Foods, 10(7), 1605. https://doi.org/10.3390/foods10071605 Ghosh, S. K., & Pal, T. (2007). Interparticle Coupling Effect on the Surface Plasmon Resonance of Gold Nanoparticles: From Theory to Applications. Chemical Reviews, 107(11), 4797–4862. https://doi.org/10.1021/cr0680282 Hou, Y., Cai, G., Zheng, L., & Lin, J. (2019). A microfluidic signal-off biosensor for rapid and sensitive detection of Salmonella using magnetic separation and enzymatic catalysis. Food Control, 103(April), 186–193. https://doi.org/10.1016/j.foodcont.2019.04.008 Huang, F., Xue, L., Qi, W., Cai, G., Liu, Y., & Lin, J. (2021). An ultrasensitive impedance biosensor for Salmonella detection based on rotating high gradient magnetic separation and cascade reaction signal amplification. Biosensors and Bioelectronics, 176, 112921. https://doi.org/10.1016/j.bios.2020.112921 Huang, J. Y., Henao, O. L., Griffin, P. M., Vugia, D. J., Cronquist, A. B., Hurd, S., Tobin-D’Angelo, M., Ryan, P., Smith, K., Lathrop, S., Zansky, S., Cieslak, P. R., Dunn, J., Holt, K. G., Wolpert, B. J., & Patrick, M. E. (2016). Infection with Pathogens Transmitted Commonly Through Food and the Effect of Increasing Use of Culture-Independent Diagnostic Tests on Surveillance — Foodborne Diseases Active Surveillance Network, 10 U.S. Sites, 2012–2015. MMWR. Morbidity and Mortality Weekly Report, 65(14), 368–371. https://doi.org/10.15585/mmwr.mm6514a2 127 Huang, X., & El-Sayed, M. A. (2010). Gold nanoparticles: Optical properties and implementations in cancer diagnosis and photothermal therapy. Journal of Advanced Research, 1(1), 13–28. https://doi.org/10.1016/j.jare.2010.02.002 Huang, Y.-F., Wang, Y.-F., & Yan, X.-P. (2010). Amine-Functionalized Magnetic Nanoparticles for Rapid Capture and Removal of Bacterial Pathogens. Environmental Science & Technology, 44(20), 7908– 7913. https://doi.org/10.1021/es102285n Khalil, M. A. F., Azzazy, H. M. E., Attia, A. S., & Hashem, A. G. M. (2014). A sensitive colorimetric assay for identification of Acinetobacter baumannii using unmodified gold nanoparticles. Journal of Applied Microbiology, 117(2), 465–471. https://doi.org/10.1111/jam.12546 Kim, G., Moon, J.-H., Moh, C.-Y., & Lim, J. (2015). A microfluidic nano-biosensor for the detection of pathogenic Salmonella. Biosensors and Bioelectronics, 67, 243–247. https://doi.org/10.1016/j.bios.2014.08.023 Li, Guo, Y., Ye, B., Zhuang, Z., Lan, P., Zhang, Y., Zhong, H., Liu, H., Guo, Z., & Liu, Z. (2020). Rapid label- free SERS detection of foodborne pathogenic bacteria based on hafnium ditelluride-Au nanocomposites. Journal of Innovative Optical Health Sciences, 13(05). https://doi.org/10.1142/S1793545820410047 Li, Z., Ma, J., Ruan, J., & Zhuang, X. (2019). Using Positively Charged Magnetic Nanoparticles to Capture Bacteria at Ultralow Concentration. Nanoscale Research Letters, 14(1), 195. https://doi.org/10.1186/s11671-019-3005-z Liu, Y., Cao, Y., Wang, T., Dong, Q., Li, J., & Niu, C. (2019). Detection of 12 Common Food-Borne Bacterial Pathogens by TaqMan Real-Time PCR Using a Single Set of Reaction Conditions. Frontiers in Microbiology, 10(FEB), 1–9. https://doi.org/10.3389/fmicb.2019.00222 Marin, M., Nikolic, M. V., & Vidic, J. (2021). Rapid point‐of‐need detection of bacteria and their toxins in food using gold nanoparticles. Comprehensive Reviews in Food Science and Food Safety, 20(6), 5880–5900. https://doi.org/10.1111/1541-4337.12839 Matta, L. L., & Alocilja, E. C. (2018a). Carbohydrate Ligands on Magnetic Nanoparticles for Centrifuge- Free Extraction of Pathogenic Contaminants in Pasteurized Milk. Journal of Food Protection, 81(12), 1941–1949. https://doi.org/10.4315/0362-028X.JFP-18-040 Matta, L. L., & Alocilja, E. C. (2018b). Emerging nano-biosensing with suspended MNP microbial extraction and EANP labeling. Biosensors and Bioelectronics, 117(July), 781–793. https://doi.org/10.1016/j.bios.2018.07.007 Matta, L. L., Harrison, J., Deol, G. S., & Alocilja, E. C. (2018). Carbohydrate-Functionalized Nanobiosensor for Rapid Extraction of Pathogenic Bacteria Directly from Complex Liquids with Quick Detection Using Cyclic Voltammetry. IEEE Transactions on Nanotechnology, 17(5), 1006–1013. https://doi.org/10.1109/TNANO.2018.2841320 Matta, L. L., Karuppuswami, S., Chahal, P., & Alocilja, E. C. (2018). AuNP-RF sensor: An innovative application of RF technology for sensing pathogens electrically in liquids (SPEL) within the food supply chain. Biosensors and Bioelectronics, 111(March), 152–158. https://doi.org/10.1016/j.bios.2018.04.010 128 Mckillip, J. L., & Drake, M. (2004). Real-Time Nucleic Acid–Based Detection Methods for Pathogenic Bacteria in Food. Journal of Food Protection, 67(4), 823–832. https://doi.org/10.4315/0362-028X- 67.4.823 Muvhali, M., Smith, A. M., Rakgantso, A. M., & Keddy, K. H. (2017). Investigation of Salmonella Enteritidis outbreaks in South Africa using multi-locus variable-number tandem-repeats analysis, 2013-2015. BMC Infectious Diseases, 17(1), 661. https://doi.org/10.1186/s12879-017-2751-8 Neoh, H., Tan, X.-E., Sapri, H. F., & Tan, T. L. (2019). Pulsed-field gel electrophoresis (PFGE): A review of the “gold standard” for bacteria typing and current alternatives. Infection, Genetics and Evolution, 74(March), 103935. https://doi.org/10.1016/j.meegid.2019.103935 Olawuyi, I. F., & Lee, W. (2019). Influence of chitosan coating and packaging materials on the quality characteristics of fresh-cut cucumber. Korean Journal of Food Preservation, 26(4), 371–380. https://doi.org/10.11002/kjfp.2019.26.4.371 Paião, F. G., Arisitides, L. G. A., Murate, L. S., Vilas-Bôas, G. T., Vilas-Boas, L. A., & Shimokomaki, M. (2013). Detection of Salmonella spp, Salmonella Enteritidis and Typhimurium in naturally infected broiler chickens by a multiplex PCR-based assay. www.sbmicrobiologia.org.br Perkins-Veazie, P., Beaulieu, J. C., & Siddiq, M. (2012). Watermelon, Cantaloupe and Honeydew. In Tropical and Subtropical Fruits Postharvest Physiology, Processing and Packaging. https://onlinelibrary.wiley.com/doi/pdf/10.1002/9781118324097#page=557 Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2015). Simultaneous direct detection of Shiga-toxin producing Escherichia coli (STEC) strains by optical biosensing with oligonucleotide- functionalized gold nanoparticles. Nanoscale, 7(6), 2417–2426. https://doi.org/10.1039/C4NR05869K Quintela, I. A., de los Reyes, B. G., Lin, C.-S., & Wu, V. C. H. (2019). Simultaneous Colorimetric Detection of a Variety of Salmonella spp. in Food and Environmental Samples by Optical Biosensing Using Oligonucleotide-Gold Nanoparticles. Frontiers in Microbiology, 10(MAY), 1–12. https://doi.org/10.3389/fmicb.2019.01138 Ragavan, K. V., Selvakumar, L. S., & Thakur, M. S. (2013). Functionalized aptamers as nano-bioprobes for ultrasensitive detection of bisphenol-A. Chemical Communications, 49(53), 5960. https://doi.org/10.1039/c3cc42002g Ravindran, N., Kumar, S., M, Y., S, R., C A, M., Thirunavookarasu S, N., & C K, S. (2023). Recent advances in Surface Plasmon Resonance (SPR) biosensors for food analysis: a review. Critical Reviews in Food Science and Nutrition, 63(8), 1055–1077. https://doi.org/10.1080/10408398.2021.1958745 Ruigrok, V. J. B., Levisson, M., Eppink, M. H. M., Smidt, H., & van der Oost, J. (2011). Alternative affinity tools: more attractive than antibodies? Biochemical Journal, 436(1), 1–13. https://doi.org/10.1042/BJ20101860 Silva, N. F. D., Magalhães, J. M. C. S., Freire, C., & Delerue-Matos, C. (2018). Electrochemical biosensors for Salmonella: State of the art and challenges in food safety assessment. Biosensors and Bioelectronics, 99(June 2017), 667–682. https://doi.org/10.1016/j.bios.2017.08.019 Soo, P.-C., Horng, Y.-T., Chang, K.-C., Wang, J.-Y., Hsueh, P.-R., Chuang, C.-Y., Lu, C.-C., & Lai, H.-C. (2009). A simple gold nanoparticle probes assay for identification of Mycobacterium tuberculosis and 129 Mycobacterium tuberculosis complex from clinical specimens. Molecular and Cellular Probes, 23(5), 240–246. https://doi.org/10.1016/j.mcp.2009.04.006 Tian, Y., Gao, X., Qi, W.-L., Wang, Y., Wang, X., Zhou, J., Lu, D., & Chen, B. (2021). Advances in differentiation and identification of foodborne bacteria using near infrared spectroscopy. Analytical Methods, 13(23), 2558–2566. https://doi.org/10.1039/D1AY00124H Wang, L., Liu, Z., Xia, X., Yang, C., Huang, J., & Wan, S. (2017). Colorimetric detection of Cucumber green mottle mosaic virus using unmodified gold nanoparticles as colorimetric probes. Journal of Virological Methods, 243, 113–119. https://doi.org/10.1016/j.jviromet.2017.01.010 Wang, L., Xue, L., Guo, R., Zheng, L., Wang, S., Yao, L., Huo, X., Liu, N., Liao, M., Li, Y., & Lin, J. (2020). Combining impedance biosensor with immunomagnetic separation for rapid screening of Salmonella in poultry supply chains. Poultry Science, 99(3), 1606–1614. https://doi.org/10.1016/j.psj.2019.12.007 Yosief, H. O., Weiss, A. A., & Iyer, S. S. (2013). Capture of uropathogenic E. coli by using synthetic glycan ligands specific for the pap-pilus. Chembiochem : A European Journal of Chemical Biology, 14(2), 251–259. https://doi.org/10.1002/cbic.201200582 You, S.-M., Jeong, K.-B., Luo, K., Park, J.-S., Park, J.-W., & Kim, Y.-R. (2021a). Paper-based colorimetric detection of pathogenic bacteria in food through magnetic separation and enzyme-mediated signal amplification on paper disc. Analytica Chimica Acta, 1151, 338252. https://doi.org/10.1016/j.aca.2021.338252 You, S.-M., Jeong, K.-B., Luo, K., Park, J.-S., Park, J.-W., & Kim, Y.-R. (2021b). Paper-based colorimetric detection of pathogenic bacteria in food through magnetic separation and enzyme-mediated signal amplification on paper disc. Analytica Chimica Acta, 1151, 338252. https://doi.org/10.1016/j.aca.2021.338252 Zeinhom, M. M. A., Wang, Y., Sheng, L., Du, D., Li, L., Zhu, M.-J., & Lin, Y. (2018). Smart phone based immunosensor coupled with nanoflower signal amplification for rapid detection of Salmonella Enteritidis in milk, cheese and water. Sensors and Actuators B: Chemical, 261, 75–82. https://doi.org/10.1016/j.snb.2017.11.093 Zeinhom, M. M. A., Wang, Y., Song, Y., Zhu, M.-J., Lin, Y., & Du, D. (2018). A portable smart-phone device for rapid and sensitive detection of E. coli O157:H7 in Yoghurt and Egg. Biosensors and Bioelectronics, 99, 479–485. https://doi.org/10.1016/j.bios.2017.08.002 Zhou, W., Gao, X., Liu, D., & Chen, X. (2015). Gold Nanoparticles for In Vitro Diagnostics. Chemical Reviews, 115(19), 10575–10636. https://doi.org/10.1021/acs.chemrev.5b00100 130 Chapter 7. DNA-BASED ANTICOUNTERFEITING TAGS Parts of this chapter were published in the International Journal of Pharmaceutics (/10.1016/j.ijpharm.2021.120580) and reprinted with permission from Elsevier. 7.1. Introduction Advancements in electronic commerce and transportation routes have resulted in increased trade of important goods and products across the globe. An undesirable consequence of the ever-increasing merchandizing of products is the predation of the supply chain with counterfeit consumer goods. In the United States alone, counterfeit products are estimated to lead to losses worth more than 600 Billion USD (Schneider, 2011). Virtually every industry has and is being affected by the counterfeiting of its products. Important consumables that are counterfeited include, but are not limited to, pharmaceutical drugs, medical equipment, food products, jewelry and electronics (Cozzolino, 2015; Fernandez et al., 2008; K. Huang et al., 2013; Shrivastava et al., 2014). These counterfeited products are usually of low quality often lacking authentic ingredients, leading not just to financial losses but have also proven to be fatal in many cases (Deisingh, 2005). The risk posed by counterfeiting of important goods has been well recognized and novel technologies are being proposed as effective measures. Examples of anticounterfeiting technologies include the traditional barcodes, quick responsive codes, holograms, watermarks and fluorescent inks to name a few (Liu et al., 2017). Recently, Radio Frequency Identification tags have been proposed to offer an efficient solution to this problem. Apart from being compact in size, RFID’s are inexpensive to manufacture (Elsherbeni, 2006; Sung-Lin Chen, 2009) thus proving to be an ideal solution to confirm authenticity of products. Unfortunately, however, it is also possible to clone these RFID tags (Mirowski, 2013), thereby hindering their application in this realm. While routes to prevent the cloning of RFID’s exist (Tuyls & Batina, n.d.), employing these may result in a direct increase in cost, defeating the purpose of their use. Another technology gaining attention with applications in anti-counterfeiting is nanoparticles and their conjugates. Campus-Cuerva et al. used silver, gold and magnetic nanoparticles and combined them in varying proportions to be used as screen printable anti-counterfeiting ink (Z. Huang et al., 2015). You et al. have proposed ink jet printing of fluorescent nanoparticles on blister packs to prevent the counterfeiting of drugs (You et al., 2016). Although nanoparticles have optical and spectroscopic advantages over other anti-counterfeiting technologies, a drawback lies in their encryption, hindering their application in the supply chain. The remarkable properties of nanoparticles may however be combined with unique DNA sequences to build an efficient anti-counterfeiting system. 131 Here, we have combined silica coated magnetic particles with genomic DNA to function as unique identification tags. The magnetic nature of these tags allows for their easy extraction under the influence of an external magnetic field for further verification. Indeed, the four individual bases of DNA molecules can be arranged in countless ways to generate unique tags which could be associated with valuable products to prevent their counterfeiting. The high density of information contained in DNA sequences (Goldman et al., 2013), encryption of information in these sequences (Anam et al., 2010) and the ease in synthesis of specific DNA sequences make it a strong candidate for anticounterfeiting applications. Examples exist in literature where DNA has been used for anti-counterfeiting applications both as a label and also as a part of the product (Liu et al., 2017), (Puddu et al., 2014). Table 7. 1. Comparison of various anticounterfeiting technologies. Technology Bar code/QR code RFID Nanoparticles PUF DNA Time for verification Low (seconds) Low (seconds) Low (seconds) Low (seconds) High (hours) Information Density Low- Moderate High Replicability Easy Level of Implementation Commercial Level of Encryption Low Difficult Laboratory/Commercial High Low Difficult Laboratory/Commercial Low Very High Unobtainable Laboratory/Commercial Very High Very High Very Difficult Laboratory/Commercial High Although the use of DNA for barcoding of products has been a topic of research in recent times, its commercial application has not yet been fully realized. Among the primary reasons that prevent the use of DNA for anticounterfeiting applications is its instability in uncontrolled environments. While coating DNA with silica has been proposed as a solution (Liu et al., 2017), (Paunescu, Fuhrer, et al., 2013; Paunescu, Puddu, et al., 2013; Puddu et al., 2014), the eventual extraction of DNA for further verification involves the use of harmful chemicals such as Hydrogen Fluoride. Apart from using dangerous chemicals, the protocol for enveloping DNA with silica may extend to at least 2 days (Paunescu, Puddu, et al., 2013) which can prove to be a tedious process. He et al. have proposed the bioconjugation of plasmid DNA with silica nanoparticles to protect it from cleavage (He et al., 2003a). Unlike encapsulation of DNA within silica particles, the authors attempted immobilization of DNA onto 132 amino-modified silica nanoparticles and demonstrated successful protection of DNA for drug delivery. Here, DNA was immobilized on silica coated magnetic particles which can be sprayed onto products to function as verification tags that can later be extracted and authenticated. For widespread application of DNA based anticounterfeiting system, its compatibility with the current state of technology, preservation of DNA in harsh environments and its rapid and simple detection is essential. Direct combination of printing ink with DNA and its subsequent retrieval for PCR amplification has been reported in literature (Hashiyada, 2004). A major disadvantage of such methods however is the use of toxic chemicals such as phenol chloroform for DNA extraction. In this study, silica coated magnetic particles (SCMP’s) were synthesized followed by their conjugation with DNA from Escherichia coli. The conjugates could then be sprayed on products or product labels. The magnetic nature of SCMP’s allowed for an easy separation of DNA for authentication without the use of expensive equipment. The presence of silica contributed towards the protection of DNA against harmful environments, confirming the results from literature (He et al., 2003a). Traditional methods for confirming the presence of DNA sequence involve PCR based approaches which can be both expensive and time consuming. Therefore, for simple and rapid detection of DNA, a gold nanoparticle based colorimetric biosensor was developed. Under the influence of acidic environment, retention of red color by gold nanoparticles confirms the presence of correct DNA sequence in product labels, thereby proving product authenticity. The absence of DNA or presence of incorrect DNA in product labels results in agglomeration of gold nanoparticles. Finally, the SCMP/DNA conjugate was sprayed on various materials such as aluminum, polymers- plastic petri plates and clingwrap and on paper-traditional printing paper and cellulose paper. The sprayed DNA was then magnetically extracted and detected using the developed biosensor. Scheme 7.1 (b) describes an overview of the proposed anticounterfeiting system. DNA sequence used as an anticounterfeiting tag is initially combined with Silica Coated Magnetic Particles (SCMP) to function as unique tags. The conjugate containing unique DNA sequence can be sprayed onto important products. For verification purposes, the SCMP/DNA conjugate can be extracted from the product, magnetically separated and tested for authenticity using a gold nanoparticle-based biosensor. 133 Scheme 7. 1. Primary advantages with the use of DNA for anticounterfeiing and (b) an overview of the proposed anticounterfeiting system with steps 1 and 2 describing the synthesis of Smart Ink and steps 3 and 4 describing its rapid detection. 7.2. Materials and Methods FeCl2.6H2O, glycerol, HAuCl4, sodium bicarbonate, sodium dodecyl sulphate, ethylene, ethyl(dimethylaminopropyl) carbodiimide (EDC) and mercaptoundecanoic acid were all purchased from Sigma Aldrich (St. Louis, Missouri). FeCl3.4H2O, N-Hydroxysuccinimide and (3-Aminopropyl) triethoxysilane (APTS) were purchased from Honeywell Fluka (Buchs, Switzerland). KBr was purchased from JT Baker (Phillipsburg, New Jersey). Agarose was purchased from Invitrogen (Carlsbad, California), NH4OH was purchased from EMD Millipore (Gibbstown, New Jersey) and ethanol was purchased from Koptec (King of Prussia, Pennsylvania). Single stranded DNA probes and PCR primers were purchased from Integrated DNA Technologies (Coralville, Iowa). dsGreen stain was purchased from Lumiprobe (Cockeysville, Maryland), DNA ladder, DNase enzyme and gel loading dye were all purchased from New England Biolabs (Ipswich, Massachusetts). DNA extraction kits, PCR reagents were purchased from Qiagen (Hilden, Germany) and media to grow the cells was purchased from Sigma Aldrich (St. Louis, Missouri). 7.2.1. Synthesis of silica coated magnetic particles Magnetic particles were synthesized using alkaline synthesis route. Briefly, 2 grams of FeCl2.4H2O and 5.4 grams of FeCl3.6H2O were dissolved in 50 mL of distilled water. The solution was added dropwise to 250 mL of 1 M NH4OH under vigorous stirring conditions. Silica layer was coated on the synthesized magnetic particles using a sol-gel approach. Briefly, 100 mg of magnetic particles were dissolved in a solution of 164 mL ethanol, 34 mL distilled water and 2 ml of 1M 134 NH4OH and the mixture was sonicated (Branson) for 40 minutes at a high setting. To the obtained mixture, 4 mL of tetraethyl orthosilicate was added dropwise under vigorous stirring conditions. Following 6 hours of stirring, the product was separated magnetically and washed at least 3 times with distilled water. The final product was obtained by drying in a vacuum oven (Fischer Scientific) at 75 °C for 1 hour. The synthesized silica coated magnetic particles were characterized for their morphology using Transmission Electron Microscope (JOEL 1400 Flash) at a magnification of 60000. The sample was prepared by dropping a small volume of the solution containing the particles onto a copper mesh, no staining was required. 7.2.2. Amination of silica coated magnetic particles and their conjugation with DNA For amination of silica coated magnetic particles, 100 mg of particles were dissolved in pure ethanol (5 mL) following which 25 µL of (3-Aminopropyl) triethoxysilane (APTS) was added. The mixture was sonicated for 1 hour followed by addition of 5 mL distilled water and sonication for another 30 min. The product was magnetically separated and washed 3 times with ethanol and drying in a vacuum oven at 50 °C. The chemical characterization of the aminated particles was achieved using Fourier Transform Infrared Spectroscopy (JASCO, FT/IR 4000) using Transmittance mode and the sample was prepared by dropping 5 µL of sample on a potassium bromide pellet followed by drying. DNA used in the experiments was extracted by growing an overnight culture of E. coli at 37 °C under constant shaking conditions. DNA extraction kits were used for the isolation of genomic DNA using manufacturer’s instructions. The quality of the extracted DNA was checked using Nanodrop Onec (Thermo Scientific) by measuring the absorbance at 260 nm. The binding of silica coated magnetic particles to genomic DNA was achieved by mixing 5 µL of the particle solution at a concentration of 0.02 mg/mL with 20 µL of DNA and incubated at room temperature for 30 min. Following magnetic separation, the mixture was run on a 1% agarose gel at 120 Volts for 20 min for DNA visualization. The protective effect of silica particles on DNA was tested by exposing the silica coated magnetic particles/DNA conjugate to deoxyribonuclease enzyme following manufacturer’s instructions. Briefly, 3 µg of sample containing silica coated magnetic particles/DNA conjugate was tested for DNA degradation 135 along with pure DNA as a control. The results were analyzed by running the samples on a 1% agarose gel at 120 volts for 20 min. 7.2.3. Synthesis of SCMP/DNA conjugates and subsequent DNA recovery The extracted DNA (~70 ng/ µL, >30 µL) was mixed with silica coated magnetic particles (0.02 mg/mL, 5 µL) and incubated for 30 min for binding to take place. The DNA from the spray was recovered by placing the microcentrifuge tubes containing the conjugates on a magnetic rack. The supernatant was discarded, and the particles suspended in nuclease free water. The presence of DNA was confirmed using the gold nanoparticle-based biosensor discussed below. 7.2.4. Synthesis and Characterization of Gold Nanoparticles Dextrin capped gold nanoparticles were synthesized using a protocol developed earlier (Anderson et al., 2011)(Yrad et al., 2019). Briefly, Gold (III) Chloride Trihydrate was added to water followed by neutralization with sodium carbonate. Next dextrin was added under continuous stirring conditions at a temperature of 150 °C. The formation of gold nanoparticles was confirmed by the evolution of wine-red color following which the reaction was stopped. The characterization of the gold nanoparticles was first done by measuring the absorbance of the sample at 520 nm using Nanodrop Onec (Thermo Scientific). The samples were also characterized using Transmission Electron Microscopy (JOEL) by magnifying up to 140,000X using an acceleration voltage of 100,000 Volts. 7.2.5. Gold nanoparticle-based detector probe and PCR experiments Detector probes were made by conjugation of gold nanoparticles with the aminated probe using ethyl(dimethylaminopropyl) carbodiimide / N-Hydroxysuccinimide chemistry. Briefly, gold nanoparticles (920 µL) synthesized above were mixed with sodium dodecyl sulphate (40 µL, 0.025 M). Following incubation for 30 min, 11-mercaptoundecanoic acid (25 µM) was added and further incubated with shaking for 30 min to prepare the gold nanoparticles for the reaction. The gold nanoparticles were washed with distilled water and purified using centrifugation at 10,000 rpm (Eppendorf) for 15 min followed by adding a mixture of EDC/NHS. Additional wash step was followed by the addition of an aminated probe (Integrated DNA Technologies) left to react at room temperature for 1 hour. Multiple wash steps were further performed to obtain the final product. The UV/visible spectrum of the sample was used as a confirmation of binding of the probe to the gold nanoparticles. 136 DNA detection was achieved by mixing the SCMP/DNA conjugates with gold nanoparticles labelled with detector probes. Typically, 15 µL of gold nanoparticles were mixed with 15 µL of SCMP/DNA conjugates and placed in a thermocycler (Eppendorf) to allow binding of probe to DNA. Each cycle included heating of sample at 94°C for 30 seconds to allow strand separation, followed by 65°C for 30 seconds for DNA/probe binding and 72°C for 1 min. Five such cycles were repeated. PCR amplification was done using thermocycler (Eppendorf) for 30 cycles with each cycle consisting of a heating step at 94°C for 30 seconds followed by 53°C for 30 seconds and 72°C for 1 min. Then, the samples were separated and run on a 1% agarose gel at 120 Volts for 20 min. The sequence of primers and probe used are given in Table 7.1. Table 7. 2. Sequences for primers and probes. Gene Assay Sequence (5’-3’) Reference uid A Biosensor CAATGGTGATGTCAGCGTT (Srinivasan et al., 2011) uid A PCR F- GCAGTCTTACTTCCATGATTTCTTTA (Srinivasan et al., 2011) R- TAATGCGAGGTACGGTAGG 7.3. Results and Discussion 7.3.1 Characterization of silica coated magnetic particles To confirm the synthesis of Fe2O3 particles and their coating with silica, the samples were analyzed using Fourier Transform Infrared Spectroscopy (FTIR). Both Fe2O3 particles and SCMP’s were confirmed to be magnetic under the influence of an external magnet. The results of the FTIR can be seen in Figure 7.1 (a) with the Wavenumber (1/cm) on the X axis and Transmittance (A.U) on the Y axis. The spectra of both as synthesized Fe2O3 particles and SCMP showed a peak at ~443 cm-1 which can be assigned to Fe-O vibration (Sodipo & Azlan, 2015). Additionally, a broad band at ~3400 cm -1 and a short peak at 1640 cm - 1 was observed in both the samples and can be attributed to O-H stretching vibration (Farahmandjou & Soflaee, 2015) (Birsan et al., 2007), indicating the synthesis of iron oxide. The SCMP displayed peaks at ~1110 cm -1 , 1230 cm -1 and 459 cm -1 which can all be assigned to Si-O-Si vibration (Akl et al., 2013; Hu, 1980; Sodipo & Azlan, 2015). These peaks confirmed the binding between Fe2O3 and silica particles. Moreover, a blunt peak at ~790 cm-1 can also be observed which can be assigned to NH2 vibration (Bruce 137 & Sen, 2005). This peak was expected as the SCMP’s were treated with APTS and indicates successful amination of the particles for DNA binding. Figure 7. 1. FTIR spectra of SCMP and as synthesized Iron Oxide particles (a) and TEM micrograph of silica coated magnetic particles (b). The SCMP’s were also characterized using a Transmission Electron Microscope (TEM) for their morphology and micrographs obtained are displayed in Figure 7.1 (b). As can be observed from the figure, the particles were found to be over 100 nm in diameter and roughly circular in shape. The presence of silica was confirmed by the FTIR and their presence may be indicated by the lightly shaded core of the particles in figure. The figure shows 4 individual particles clumped together each with the darker Fe2O3 magnetic particles embedded inside a silica core. The silica core appears to have a lighter color compared to the magnetic particles and is indicated by arrows in the figure. 7.3.2. Conjugation of Silica Coated Magnetic Nanoparticles with DNA Following the successful coating of Fe2O3 particles with silica, their binding with genomic DNA was attempted. The SCMP’s were aminated to prepare them for binding to genomic DNA. The samples were purified to remove any unbound DNA and were visualized on an agarose gel (Figure 7.2 (a)). Figure 7.2 (a) shows the result from gel electrophoresis with Lane 1 consisting of a 1 kilobase ladder, Lane 2 consisting of the SCMP’s without DNA (control) and Lanes 3,4 with SCMP/DNA conjugate in duplicates. As expected, the control sample does not fluoresce under UV light indicating that the silica coated magnetic particles do not exhibit fluorescence. The samples in Lane 3 fluoresces indicating the presence of DNA. Indeed, the samples display a limited fluorescence compared to the ladder from Lane 1 indicating that the presence of magnetic particles and silica may contribute towards quenching of fluorescence from the DNA. Although the quenching of fluorescence due to the presence of silica was confirmed with an earlier study by He and coworkers (He et al., 2003b), we hypothesize that the 138 presence of magnetic particles may also play a role in the same. It can also be observed that the samples from Lane 3 does not migrate under the influence of electric potential which could be due to large size of the particles to which the DNA is bound and also due to their contribution towards the polarity of the DNA molecules. Figure 7. 2. (a) Gel electrophoresis results following magnetic separation. Lane 1 is a 1 Kb ladder, Lane 2 is SCMP particles without DNA, Lanes 3 and 4 are samples in duplicates and (b) shows the effect of DNase on SCMP/DNA conjugates (Lane 2) and pure DNA in Lane 3. The effect of temperature on stability is shown in (c) with pure DNA and SCMP/DNA at 60 °C (Lanes 2,3) at 90 °C (Lanes 4, 5) and at 120 °C (Lanes 6, 7). The effect of UV after 30 min exposure of pure DNA and SCMP/DNA (Lanes 2, 3) and 60 min exposure (Lanes 4, 5). The primary reason for the use of silica in the SCMP was to protect the DNA against harmful environments since the stability of DNA in anticounterfeiting applications is important. To test the beneficial effect of silica on the stability of DNA under unfavorable conditions, the SCMP/DNA conjugates were subjected to the activity of DNase. Figure 7.2 (b) shows the effect of DNase on the conjugate (Lane 2) and the control (Lane 3). As can be observed, the control which lacked silica coated magnetic particles as support, displayed a smear indicating the degradation of DNA. Although the conjugate shows less fluorescence, the absence of a smear confirmed the beneficial effect of silica on the stability of DNA. The SCMP/DNA particles were found to be stable at temperatures of 60 °C, 90 °C, 139 and 120 °C (Fig 7.2 c, lanes 3, 5, and 7) and after exposure to UV light for 30 and 60 min respectively (Fig 7.2 d, lanes 3 and 5). DNA without SCMP readily degraded. Following a successful binding, the SCMP/DNA conjugates were tested on the biosensor described below. The conjugates were sprayed on various materials to check for the stability of DNA. 7.3.3. DNA recovery and detection An important factor governing the success of DNA as an anti-counterfeiting agent is its rapid detection in a low resource setting. Hence, we developed a rapid colorimetric method to confirm the presence of correct DNA sequence. The method utilizes optical properties of gold nanoparticles as a means of biosensing, specifically their stability under the influence of acidic environments following successful binding with DNA. The gold nanoparticles were initially conjugated with a detector probe with a sequence specific to the DNA used in Smart Ink. 7.3.4. Characterization of reporter probe Gold nanoparticles conjugated with oligonucleotide probes were used to confirm the presence of DNA. The dextrin capped gold nanoparticles exhibited a dark red appearance and were characterized using TEM and by measuring the absorbance at 520 nano meters. Figure 3 (a) shows a transmission electron micrograph of the as-synthesized gold nanoparticles. The figure shows a cluster of gold nanoparticles which exhibited spherical morphology. The inset in the figure shows a single gold nanoparticle with an approximate diameter of 33 nm. The particle size distribution was measured using a Zetasizer (not shown, Supplementary Information) which revealed that majority of the particles were between 25 to 30 nm in diameter, confirming the results of the TEM. The nanoparticles were also characterized by measuring their absorbance at 520 nm and the result can be observed in Figure 7.3 (b). Results from the absorbance spectrum are shown with Wavelength (nm) on the X axis and the absorbance (A.U) on the Y axis. As can be observed from the figure, the characteristic peak for gold nanoparticles at 520 nm confirmed their successful synthesis. Synthesis of gold nanoparticles was followed by their conjugation with aminated oligonucleotides with the conjugate functioning as a detector probe. To confirm the conjugation, the characterization using spectroscopy was done and can be observed from Figure 7.3 (b) (red). As expected, along with a peak at around 520 nm for gold nanoparticles, a blunt peak at about 260 nm for the oligonucleotide can be observed which indicates the successful conjugation between the gold nanoparticles and 140 oligonucleotide probes. It can also be observed, however, that following conjugation with oligonucleotide probe, the exhibited a broadened peak and an additional peak between 600 and 700 nm indicating a change in size (Zuber et al., 2016). Figure 7. 3. Transmission electron micrograph of the synthesized gold nanoparticles (GNP) (a) and absorbance of GNP and GNP based detector probe (b). 7.3.5. Detection using Biosensor The SCMP/DNA conjugates were initially subjected to magnetic separation for the isolation of SCMP/DNA conjugates. Following multiple wash steps, gold nanoparticle-based detector probe was mixed with the SCMP/DNA conjugates and placed in a thermocycler which allowed strand separation and binding of detector probe to the genomic DNA. The supernatant was subjected to acidic environment by adding 0.1 M HCL and change in color of the sample was observed (Figure 7.4 (b)). Figure 7.4 (a) shows a picture of sample tubes with no DNA (Left), correct DNA (Centre) and incorrect DNA (Right). As can be observed from the color of the tubes, the absence of DNA and the presence of incorrect DNA sequence resulted in aggregation of gold nanoparticles while the presence of correct DNA contributed towards the stability of gold nanoparticles which appear red in color. The spectrum of the samples was also obtained and can be observed in Figure 7.4 (b). As expected, a sharp peak at around 520 nm for sample with correct DNA was observed while broad peaks for both samples without DNA and with incorrect DNA were seen indicating their aggregation. Although both incorrect DNA and no DNA resulted in aggregation of gold nanoparticles, the level of aggregation was found to be different for the two. 141 It is important to note that acidic treatment was applied to the supernatant rather than to the magnetically purified SCMP’s. We hypothesized the separation of DNA from the SCMP’s following its exposure to varying temperatures in the thermocycler. (c) (a) Figure 7. 4. Biosensor to detect the presence of specific DNA sequence (a), absorption spectrum of the same samples (b) and PCR product of amplicon from sample (Lane 2), sample with no DNA(Lane 3) and incorrect DNA (Lane 4). 1Kb ladder was used for reference. To test the release of DNA from the SCMP’s, the supernatant was amplified using PCR, a ~1000 bp amplicon was selected for the reaction. Following amplification, the samples were run on an agarose gel and the results can be observed from Figure 7.4 ©. Lane 1 in the figure consists of a 100 base pair ladder and Lane 2 consists of the sample. Lane 3 consists of PCR amplified SCMP’s following magnetic separation and Lane 4 consists of pure SCMP’s. A distinct band of approximately 500 base pairs can be clearly seen for the sample (Lane 2). The presence of a distinct band confirmed the presence of DNA in the supernatant, thereby proving its release from the SCMP’s. No band however was observed for PCR amplified SCMP’s confirming the release of DNA. Although the exact cause of the release of DNA is not known, we hypothesize that subjecting the SCMP/DNA conjugates to high temperature variations results in the separation of DNA from the SCMP’s. 142 The separated DNA molecules then bind to detector probes attached to gold nanoparticles. A gold nanoparticles/ genomic DNA conjugate therefore results where the genomic DNA contributes towards the protection of gold nanoparticles against aggregation under the influence of acidic environment. 7.3.6. Stability of SCMP/DNA conjugates on various substrates To test the stability of the SCMP/DNA conjugates, they were sprayed onto the following materials- aluminum, polymers- plastic petri plates and clingwrap and paper-traditional printing paper and cellulose paper. The conjugate was extracted magnetically following which the biosensor was used to test their stability, maintenance of red color by the gold nanoparticles indicated stability of the conjugates and successful DNA detection (+). Agglomeration of gold nanoparticles indicated the absence of DNA (-). Table 7. 3. Stability of DNA/SCMP conjugates tested on various materials. Material Aluminum Petri plate Clingwrap Cellulose paper Printing paper DNA Detection + + + + - As can be observed from Table 7.2, the stability of the conjugates was confirmed for aluminum, polymers and cellulose paper. The conjugate was however not found to be stable on traditional printing paper. We hypothesize the absorption of DNA onto printing paper may be a possible reason behind this. 143 REFERENCES Akl, M. A., Aly, H. F., Soliman, H. M. A., Abd ElRahman, A. M. E., & Abd-Elhamid, A. I. (2013). Preparation and characterization of silica nanoparticles by wet mechanical attrition of white and yellow sand. Journal of Nanomedicine and Nanotechnology, 4(6). https://doi.org/10.4172/2157-7439.1000183 Anam, B., Sakib, K., Hossain, Md. A., & Dahal, K. (2010). Review on the Advancements of DNA Cryptography. Anderson, M. J., Torres-Chavolla, E., Castro, B. A., & Alocilja, E. C. (2011). One step alkaline synthesis of biocompatible gold nanoparticles using dextrin as capping agent. Journal of Nanoparticle Research, 13(7), 2843–2851. https://doi.org/10.1007/s11051-010-0172-3 Birsan, C., Predoi, D., & Andronescu, E. (2007). IR and thermal studies of iron oxide nanoparticles in a bioceramic matrix. Journal of Optoelectronics and Advanced Materials, 9(6), 1821–1824. Bruce, I. J., & Sen, T. (2005). Surface Modification of Magnetic Nanoparticles with Alkoxysilanes and Their Application in Magnetic Bioseparations. Langmuir, 21(15), 7029–7035. https://doi.org/10.1021/la050553t Cozzolino, D. (2015). The role of vibrational spectroscopy as a tool to assess economically motivated fraud and counterfeit issues in agricultural products and foods. Analytical Methods, 7(22), 9390– 9400. https://doi.org/10.1039/C5AY01792K Deisingh, A. K. (2005). Pharmaceutical counterfeiting. The Analyst, 130(3), 271. https://doi.org/10.1039/b407759h Elsherbeni, A. (2006). Recent Rfid Technology and Applications. Farahmandjou, M., & Soflaee, F. (2015). Synthesis and characterization of α-Fe2O3 nanoparticles by simple co-precipitation method. Physical Chemistry Research, 3(3), 191–196. https://doi.org/10.22036/pcr.2015.9193 Fernandez, F. M., Green, M. D., & Newton, P. N. (2008). Prevalence and Detection of Counterfeit Pharmaceuticals: A Mini Review. Industrial & Engineering Chemistry Research, 47(3), 585–590. https://doi.org/10.1021/ie0703787 Goldman, N., Bertone, P., Chen, S., Dessimoz, C., LeProust, E. M., Sipos, B., & Birney, E. (2013). Towards practical, high-capacity, low-maintenance information storage in synthesized DNA. Nature, 494(7435), 77–80. https://doi.org/10.1038/nature11875 Hashiyada, M. (2004). Development of Biometric DNA Ink for Authentication Security. The Tohoku Journal of Experimental Medicine, 204(2), 109–117. https://doi.org/10.1620/tjem.204.109 He, X., Wang, K., Tan, W., Liu, B., Lin, X., He, C., Li, D., Huang, S., & Li, J. (2003a). Bioconjugated Nanoparticles for DNA Protection from Cleavage. Journal of the American Chemical Society, 125(24), 7168–7169. https://doi.org/10.1021/ja034450d He, X., Wang, K., Tan, W., Liu, B., Lin, X., He, C., Li, D., Huang, S., & Li, J. (2003b). Bioconjugated Nanoparticles for DNA Protection from Cleavage. Journal of the American Chemical Society, 125(24), 7168–7169. https://doi.org/10.1021/ja034450d 144 Hu, S. M. (1980). Infrared absorption spectra of SiO 2 precipitates of various shapes in silicon: calculated and experimental. Journal of Applied Physics, 51(11), 5945–5948. https://doi.org/10.1063/1.327512 Huang, K., Carulli, J. M., & Makris, Y. (2013). Counterfeit electronics: A rising threat in the semiconductor manufacturing industry. 2013 IEEE International Test Conference (ITC), 1–4. https://doi.org/10.1109/TEST.2013.6651880 Huang, Z., Wang, S., Wang, J., Yu, Y., Wen, J., & Li, R. (2015). Exfoliation-restacking synthesis of coal- layered double hydroxide nanosheets/reduced graphene oxide composite for high performance supercapacitors. Electrochimica Acta, 152, 117–125. https://doi.org/10.1016/j.electacta.2014.11.085 Liu, Y., Zhang, C., Li, X., & Wu, D. (2017). A self-monitored fluorescence DNA anti-counterfeiting system based on silica coated SYBR Green I/DNA gelatin nanoparticles. Journal of Materials Chemistry C, 5(24), 5939–5948. https://doi.org/10.1039/C7TC01380A Mirowski, L. (2013). Exposing Clone RFID Tags at the Reader. 2013 12th IEEE International Conference on Trust, Security and Privacy in Computing and Communications, 1669–1674. https://doi.org/10.1109/TrustCom.2013.207 Paunescu, D., Fuhrer, R., & Grass, R. N. (2013). Protection and Deprotection of DNA-High-Temperature Stability of Nucleic Acid Barcodes for Polymer Labeling. Angewandte Chemie International Edition, 52(15), 4269–4272. https://doi.org/10.1002/anie.201208135 Paunescu, D., Puddu, M., Soellner, J. O. B., Stoessel, P. R., & Grass, R. N. (2013). Reversible DNA encapsulation in silica to produce ROS-resistant and heat-resistant synthetic DNA “fossils.” Nature Protocols, 8(12), 2440–2448. https://doi.org/10.1038/nprot.2013.154 Puddu, M., Paunescu, D., Stark, W. J., & Grass, R. N. (2014). Magnetically Recoverable, Thermostable, Hydrophobic DNA/Silica Encapsulates and Their Application as Invisible Oil Tags. ACS Nano, 8(3), 2677–2685. https://doi.org/10.1021/nn4063853 Schneider, F. (2011). The Financial Flows of the Transnational Crime: Some Preliminary Empirical Results. Shrivastava, A., Azarian, M. H., Morillo, C., Sood, B., & Pecht, M. (2014). Detection and Reliability Risks of Counterfeit Electrolytic Capacitors. IEEE Transactions on Reliability, 63(2), 468–479. https://doi.org/10.1109/TR.2014.2315914 Sodipo, B. K., & Azlan, A. A. (2015). Superparamagnetic iron oxide nanoparticles incorporated into silica nanoparticles by inelastic collision via ultrasonic field: Role of colloidal stability. AIP Conference Proceedings, 1657(November), 100002. https://doi.org/10.1063/1.4915209 Srinivasan, S., Aslan, A., Xagoraraki, I., Alocilja, E., & Rose, J. B. (2011). Escherichia coli, enterococci, and Bacteroides thetaiotaomicron qPCR signals through wastewater and septage treatment. Water Research, 45(8), 2561–2572. https://doi.org/10.1016/j.watres.2011.02.010 Sung-Lin Chen. (2009). A Miniature RFID Tag Antenna Design for Metallic Objects Application. IEEE Antennas and Wireless Propagation Letters, 8(c), 1043–1045. https://doi.org/10.1109/LAWP.2009.2032252 Tuyls, P., & Batina, L. (n.d.). RFID-Tags for Anti-Counterfeiting. In Cryptographers’ Track at the RSA Conference, 115–131. 145 You, M., Lin, M., Wang, S., Wang, X., Zhang, G., Hong, Y., Dong, Y., Jin, G., & Xu, F. (2016). Three- dimensional quick response code based on inkjet printing of upconversion fluorescent nanoparticles for drug anti-counterfeiting. Nanoscale, 8(19), 10096–10104. https://doi.org/10.1039/C6NR01353H Yrad, F. M., Castañares, J. M., & Alocilja, E. C. (2019). Visual Detection of Dengue-1 RNA Using Gold Nanoparticle-Based Lateral Flow Biosensor. Diagnostics, 9(3), 74. https://doi.org/10.3390/diagnostics9030074 Zuber, A., Purdey, M., Schartner, E., Forbes, C., van der Hoek, B., Giles, D., Abell, A., Monro, T., & Ebendorff-Heidepriem, H. (2016). Detection of gold nanoparticles with different sizes using absorption and fluorescence based method. Sensors and Actuators B: Chemical, 227, 117–127. https://doi.org/10.1016/j.snb.2015.12.044 146 Chapter 8. CONCLUSIONS AND RECOMMENDATIONS 8.1. Conclusions This work elaborated on the use of glycan-coated magnetic nanoparticles for foodborne outbreak prevention. The remarkable properties of gMNP were assessed for extraction of E. coli, S. enterica, and B. cereus from various complex food matrices. Concentration and extraction of the bacteria from leafy greens, fresh produce, and ready-to-eat meals demonstrated the applicability of gMNP in a range of solid foods. Notably, rapid concentration was achieved in the presence of food microparticles and natural microbiota. The concentration factor was found to vary among the food matrices tested. The magnetically extracted pathogens were also detected using amplification-based methods such as PCR and qPCR proving that gMNP can effectively remove PCR inhibitors commonly associated with food matrices. While using qPCR for pathogen detection effectively eliminated the need for any pre- enrichment, expensive reagents and equipment warrant alternatives. The gMNP-based extraction was combined with NIRS and, using supervised machine, a high classification accuracy of 90% was achieved with pure culture. From foods, the classification efficiency reduced to 70%, which was expected. Considering the complex nature of rice and macaroni, this accuracy is reasonable. The gMNP were successful in concentrating bacterial cells and improving the signal from NIR and, to a significant extent, separated the food matrix. However, the inability of NIRS to classify food matrices with high bacterial loads of natural microbiota is a primary obstacle that needs to be addressed. Notably, the dGNP-based biosensor was successful in detection of target pathogens in the presence of natural microbiota. The dGNP-based plasmonic biosensor provided a robust platform for rapid detection of S. enterica and E. coli from various complex matrices. The simple procedure of the biosensor with 3 simple steps allowed pathogen detection in under 30 min after DNA extraction. Initial bacterial loads as low as 103 CFU/mL were successfully detected. In-house synthesis of dGNP, absence of probe conjugation procedures, and no requirement for amplification contributed towards reducing costs associated with this assay. The dGNP were prepared in batches and stored for more than a year with no degradation of nanoparticles. Among the techniques used in this work, and those currently being used for pathogen detection from foods, the dGNP biosensor is reasonably simple, affordable, and well-suited for low- resource settings. 147 8.2. Recommendations While the use of carbohydrate-coated nanoparticles provided an affordable route for pathogen detection from foods, platform accessibility can be further improved. Initial food screening may be implemented using the NIRS platform to indicate contamination of food. For specific pathogen identification, the plasmonic biosensor can be used. Some recommendations for the gMNP-based extraction and detection platform are discussed below. The gMNP offer an affordable alternative to recognition ligand-based methods, however, additional studies on their reusability can further help reduce costs. To achieve this, acid treatment-based methods can be explored to eliminate remnant bacterial debris. The gMNP-based magnetic extraction from food demonstrated relatively lower CF, compared to that in pure culture, which could be due to multiple factors. The efficiency of concentration from foods may further be improved by aminating the gMNP to increase their positive surface charge. Additionally, the effect of natural microbiota in food can be minimized by using species-specific glycan which can selectively capture the target bacterial species. While the NIRS platform can be used for foods with low levels of natural microbiota, plasmonic biosensor is recommended in other cases owing to the specificity of probes. For plasmonic biosensor, the DNA was extracted using traditional kits which can prove to be expensive and difficult to implement in areas with scarcity of resources. This obstacle may be addressed using inexpensive methods for DNA extraction, such as heat treatment-based, which are easy to implement and do not require complex experimental setup or equipment. The heat treatment of magnetically separated bacterial cells results in the release of genomic DNA, and cell debris can be magnetically removed. The method can easily be implemented using a water bath. While no amplification was required for target DNA detection, the biosensor was implemented in a thermocycler which can add to costs for detection. However, since static temperature conditions are needed for the biosensor, 95 °C for strand separation and 55 °C for probe hybridization, a water bath can easily replace the thermocycler. For field application of the plasmonic biosensor platform, large-volume batches of dGNP and probe can be prepared and transported to points of need. Lastly, while the use of nanodrop for quantification of dGNP agglomeration provided accurate results, its field implementation can be expensive. Alternatively, a smartphone-based application can successfully differentiate between aggregated and well-dispersed dGNP, eliminating the need for a spectrophotometer. 148