Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of innovative biosensors for the determination of melamine in milk Hu, Yaxi 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_september_hu_yaxi.pdf [ 2.38MB ]
Metadata
JSON: 24-1.0166414.json
JSON-LD: 24-1.0166414-ld.json
RDF/XML (Pretty): 24-1.0166414-rdf.xml
RDF/JSON: 24-1.0166414-rdf.json
Turtle: 24-1.0166414-turtle.txt
N-Triples: 24-1.0166414-rdf-ntriples.txt
Original Record: 24-1.0166414-source.json
Full Text
24-1.0166414-fulltext.txt
Citation
24-1.0166414.ris

Full Text

DEVELOPMENT OF INNOVATIVE BIOSENSORS FOR THE DETERMINATION OF MELAMINE IN MILK  by  Yaxi Hu  B.Sc., China Agricultural University, 2013  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Food Science)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  July 2015  © Yaxi Hu, 2015 ii  Abstract  After discovering the potential detriment of melamine to humans, the society calls for novel techniques to accomplish accurate, rapid, high-throughput, and on-line or in-field detection of melamine in foods, as required by the food industry and government laboratories. The aim of this study was to investigate different innovative biosensors combining antibodies or molecularly imprinted polymers (MIPs) with surface enhanced Raman spectroscopy (SERS) to determine melamine in a representative dairy product (i.e. milk).   A “two-step” antibody-SERS biosensor was developed to detect melamine in whole milk. The anti-melamine antibody, produced by immunizing New Zealand white rabbits with melamine hapten-ovalbumin immunogen, was used to extract melamine from whole milk exclusively. After releasing melamine from the antibody, the eluents were deposited onto silver dendrite SERS-active substrate for SERS spectral collection. The limit of detection (LOD) calculated by the principal component analysis (PCA) model was lower than 0.79×10-3 mmol/L. The overall analysis was completed in 20 min.  The MIP for the “two-step” MIPs-SERS biosensor was synthesized by bulk polymerization of melamine, methacrylic acid, ethylene glycol dimethacrylate and 2,2’-azobisisobutyronitrile. After confirming the specific affinity of the MIP towards melamine by adsorption capacity tests, MIP was used as sorbent for solid phase extraction (SPE) to extract melamine from whole milk. SERS spectra were collected by depositing the eluents from MISPE onto silver dendrite. The LOD and limit of quantification (LOQ) calculated by the linear regression model correlating iii  relative intensity of melamine SERS band at 703 cm-1 and melamine concentration in whole milk were 0.012 mmol L-1 and 0.039 mmol L-1, respectively, and the full analysis was accomplished in 18 min.  “One-step” MIPs-SERS biosensor incorporated silver nanoparticles (AgNPs) into MIPs synthesized by bulk polymerization. Adsorption capacity tests verified the specific affinity of MIPs-AgNPs to melamine, and PCA model resulted in the LOD between 0.01 and 0.017 mmol L-1 melamine in skim milk. The time required to detect melamine in skim milk was 25 min.   The low LOD and LOQ, as well as rapid detection confirm the potential of applying these three types of biosensors for accurate and high-throughput detection of melamine in dairy products. iv  Preface  The work presented in Chapter 3 has been published in Journal of Food Science (2015, 80, C1196-C1201) under the title “Rapid detection of melamine in milk using immunological separation and surface enhanced Raman spectroscopy” (with the exception of information with footnotes). I, as one of the three first authors, conducted all the data analysis and wrote the result and discussion section of the manuscript. This study was designed with the help of my supervisor Dr. Xiaonan Lu. He provided feedback on the manuscript and made editorial changes. The other two co-first authors conducted the experiments, while all the other co-authors helped review the manuscript and provided feedback.  A version of Chapter 4 was published in Food Chemistry (2015, 176, 123-129) under the title “Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy”. Additional data were added for this thesis, and are marked with footnotes. Dr. Xiaonan Lu and I designed the research; I carried out the experiments, conducted data analysis, and wrote the manuscript. Dr. Xiaonan Lu critically reviewed the manuscript, provided feedback, and made editorial changes. The second and third authors trained me at the beginning of this study. All the other co-authors were involved in reviewing the manuscript and provided feedback.  A version of Chapter 5 was submitted to related journal under the title “Rapid detection of melamine in tap water and milk using conjugated ‘one-step’ molecularly imprinted polymers-surface enhanced Raman spectroscopic biosensor”. Information about the application in tap water was removed for this thesis to assure the consistency and adequate flow of the information. v  Additional data were added into this thesis, and are marked with footnotes. Dr. Xiaonan Lu and I designed the research; I conducted the experiments and data analysis, as well as wrote the manuscript. Dr. Xiaonan Lu critically reviewed the manuscript, provided feedback, and made editorial changes.  My committee member, Dr. Christine H. Scaman and Dr. Edward Grant provided additional advice and support in this research.  vi  Table of Contents  Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ......................................................................................................................... vi List of Tables ..................................................................................................................................x List of Figures ............................................................................................................................... xi List of Abbreviations ...................................................................................................................xv Acknowledgements ................................................................................................................... xvii Dedication ................................................................................................................................... xix Chapter 1: Introduction ............................................................................................................... 1  Melamine issues in the food industry ......................................................................... 1 1.1 Traditional techniques for melamine determination ................................................... 1 1.2 Innovative techniques for melamine determination .................................................... 3 1.31.3.1 Surface enhanced Raman spectroscopy (SERS) ................................................. 3 1.3.2 Techniques for analyte extraction ....................................................................... 4 1.3.2.1 Antibody ......................................................................................................... 4 1.3.2.2 Molecularly imprinted polymers (MIPs) ........................................................ 7 1.3.3 Antibody or MIPs coupled SERS biosensors ................................................... 10 Chapter 2: Hypothesis and experimental overview .................................................................. 12  Rationale ................................................................................................................... 12 2.1 Objectives ................................................................................................................. 13 2.2 Hypothesis................................................................................................................. 14 2.3vii   Experimental overview ............................................................................................. 14 2.4Chapter 3: Rapid detection of melamine in whole milk using a “two-step” antibody-surface enhanced Raman spectroscopic biosensor ................................................................................ 17  Introduction ............................................................................................................... 17 3.1 Materials ................................................................................................................... 17 3.2 Methods..................................................................................................................... 18 3.33.3.1 Preparation of immunogen ................................................................................ 18 3.3.2 Polyclonal antibody production ........................................................................ 19 3.3.3 Whole milk sample pretreatment ...................................................................... 19 3.3.4 Immunological separation of melamine in whole milk .................................... 20 3.3.5 Synthesis of silver dendrite SERS substrate ..................................................... 21 3.3.6 Raman spectroscopic instrumentation .............................................................. 22 3.3.7 Spectral analysis and chemometric models ...................................................... 22  Results and discussion .............................................................................................. 23 3.43.4.1 Specificity test ................................................................................................... 23 3.4.2 Sensitivity test ................................................................................................... 28  Conclusion ................................................................................................................ 30 3.5Chapter 4: Rapid detection of melamine in whole milk using a “two-step” MIPs-SERS biosensor ................................................................................................................................... 32  Introduction ............................................................................................................... 32 4.1 Materials ................................................................................................................... 33 4.2 Methods..................................................................................................................... 33 4.34.3.1 Synthesis of molecularly imprinted polymers .................................................. 33 viii  4.3.2 Adsorption capacity tests .................................................................................. 34 4.3.3 Whole milk sample pretreatment ...................................................................... 35 4.3.4 High performance liquid chromatography conditions ...................................... 35 4.3.5 Molecularly imprinted solid phase extraction (MISPE) ................................... 35 4.3.6 Synthesis of silver dendrite SERS substrate ..................................................... 36 4.3.7 Raman spectroscopic instrumentation .............................................................. 36 4.3.8 Spectral analysis and chemometric models ...................................................... 36  Results and discussion .............................................................................................. 37 4.44.4.1 Synthesis and characterization of MIPs ............................................................ 37 4.4.2 MISPE for melamine spiked whole milk .......................................................... 41 4.4.3 Determination of melamine in whole milk by SERS........................................ 43  Conclusion ................................................................................................................ 51 4.5Chapter 5: Rapid detection of melamine in whole milk using a “one-step” MIPs-SERS biosensor ................................................................................................................................... 52  Introduction ............................................................................................................... 52 5.1 Materials ................................................................................................................... 53 5.2 Methods..................................................................................................................... 54 5.35.3.1 Synthesis of molecularly imprinted polymers-silver nanoparticles .................. 54 5.3.2 Adsorption capacity tests .................................................................................. 55 5.3.3 Skim milk sample pretreatment ........................................................................ 56 5.3.4 High performance liquid chromatography conditions ...................................... 56 5.3.5 MIPs-AgNPs extraction .................................................................................... 56 5.3.6 Raman spectroscopic instrumentation .............................................................. 57 ix  5.3.7 Spectral analysis and chemometric models ...................................................... 57  Results and discussion .............................................................................................. 57 5.45.4.1 Synthesis of MIPs-AgNPs ................................................................................ 57 5.4.2 Characterization of MIPs-AgNPs ..................................................................... 62 5.4.3 Extraction and determination of melamine in water and skim milk by MIPs-AgNPs ........................................................................................................................... 65  Conclusion ................................................................................................................ 70 5.5Chapter 6: Conclusion............................................................................................................... 72  Main findings ............................................................................................................ 72 6.1 Future research directions ......................................................................................... 75 6.2 Conclusion ................................................................................................................ 77 6.3References .....................................................................................................................................79 Appendices ....................................................................................................................................95  x  List of Tables  Table 4.1 Recoveries of melamine in whole milk by molecularly imprinted solid phase extraction (MISPE) and non-imprinted solid phase extraction (NISPE). ...................................................... 43 Table 6.1 Summary of the decision for hypotheses developed in this study. ............................... 73 Table 6.2 Summary of the results of three types of innovative SERS biosensors. ....................... 78  xi  List of Figures  Figure 3.1 Chemical structures of melamine and corresponding hapten. ..................................... 19 Figure 3.2 Schematic illustration of immunological separation integrated with surface enhanced Raman spectroscopy to determine melamine in milk. .................................................................. 21 Figure 3.3 Representative normalized SERS spectral features of eluent from milk spiked with melamine (0.79 μmol L-1) by immunological separation using water, milk, and melamine (0.16 mmol L-1 in water) as negative and positive controls, and elution buffer as background: a) silver dendrite SERS-active substrate, b) elution buffer, c) water negative control, d) milk negative control, e) melamine eluent from milk without the antibody, f) melamine eluent from milk with the antibody, and g) melamine in methanol (positive control). .................................................... 24 Figure 3.4 Representative principal component analysis for negative controls (i.e. milk and water), elution buffer, and milk sample spiked with 0.79 μmol L-1 melamine (N=6). Different marks correspond to different samples: circle: water negative control; diamond: milk negative control; plus sign: elution buffer; asterisk: milk samples spiked with melamine. Cluster with dash line is the milk samples spiked with melamine, while cluster with the solid line contains the negative controls and elution buffer. ............................................................................................ 26 Figure 3.5 Loading plot of the representative principal component analysis model. Principal component (PC) 1 represents 85.9% variances. ............................................................................ 27 Figure 3.6 Representative principal component analysis for milk samples spiked with different concentrations of melamine (N=8). Different markers represent different melamine concentrations: asterisk, 0.79×10-3 mmol L-1 melamine; circle, 4.0×10-3 mmol L-1; plus sign, 7.9×10-3 mmol L-1; cross sign, 0.040 mmol L-1, diamond, 0.079 mmol L-1; hexagram, 0.16 mmol xii  L-1; and pentagram, non-spiked samples. Cluster with dash line is the non-spiked milk samples, while cluster with the solid line contains the samples spiked at different melamine concentrations. .............................................................................................................................. 29 Figure 3.7 Loading plot of the representative principal component analysis model. Principal component (PC) 1 represents 82.7% variances. ............................................................................ 30 Figure 4.1 Schematic illustration of MIPs-SERS biosensor for the detection of melamine in whole milk. CCD: charge coupled device; MIPs: molecularly imprinted polymers; SERS: surface enhanced-Raman spectroscopy. .................................................................................................... 38 Figure 4.2 Static binding isotherm of molecularly imprinted polymers (MIPs) and non-imprinted polymers (NIPs) for melamine. Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3). ...................................................................................... 40 Figure 4.3 Kinetic binding isotherm of MIPs and NIPs (Ci: 6.3 mg L-1). Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3). ..................... 41 Figure 4.4 Representative spectral features of melamine with (a) MIPs-SERS of whole milk sample spiked with 6.3 mg L-1 melamine, (b) SERS of standard melamine solution in methanol (63 mg L-1), and (c) normal Raman of melamine crystal. ............................................................ 45 Figure 4.5 Representative MIPs-SERS spectra of whole milk samples spiked with different concentrations of melamine: (a) 6.3 mg L-1, (b) 3.15 mg L-1, (c) 1.26 mg L-1, (d) 0.63 mg L-1, (e) 0.126 mg L-1, and (f) 0 mg L-1. ..................................................................................................... 46 Figure 4.6 Representative two-dimensional principal component analysis for whole milk samples spiked with different contents of melamine. Principal component (PC) 1 represents 70.4% variances and PC2 represents 16.3% variances. Letters from A to F denote whole milk samples xiii  spiked with melamine at the concentrations of 6.3, 3.15, 1.26, 0.63, 0.126 and 0 mg L-1, respectively (N=7). ....................................................................................................................... 47 Figure 4.7 Loading plot of representative principal component analysis model. Principal component (PC) 1 represents 70.4% variances and PC2 represents 16.3% variances. ................ 48 Figure 4.8 Linear relationship between the intensity of SERS band at 703 cm-1 and the spiked levels of melamine in whole milk samples (N=6). ....................................................................... 49 Figure 5.1 Schematic illustration of one-step MIPs-SERS biosensor for the detection of melamine in dairy samples. CCD: charge coupled device; MIPs: molecularly imprinted polymers; SERS: surface enhanced Raman spectroscopy. ........................................................... 58 Figure 5.2 Representative Raman spectra of A) MIPs-AgNPs and B) NIPs-AgNPs before and after the reduction of Ag+ and after Soxhlet extraction. a-f: MIPs-AgNPs before reduction, MIPs-AgNPs after reduction, MIPs-AgNPs after Soxhlet extraction, NIPs-AgNPs before reduction, NIPs-AgNPs after reduction, and NIPs-AgNPs after Soxhlet extraction. .................................... 60 Figure 5.3 Representative UV-Vis spectra of “two-step” MIPs, MIPs-AgNPs before and after the reduction of Ag+. Solid line: two-step MIPs; dotted line: MIPs-AgNPs before the reduction of Ag+; dashed line: MIPs-Ag-NPs after the reduction of Ag+. ........................................................ 61 Figure 5.4 Static binding isotherm of molecularly imprinted polymers-silver nanoparticles (MIPs-AgNPs) and non-imprinted polymers-silver nanoparticles (NIPs-AgNPs) for melamine. Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3). ............................................................................................................................................ 63 Figure 5.5 Comparison of the adsorption capacity of one-step and two-step MIPs and NIPs (Ci: 6.3 mg L-1). Means with different letters are significantly different (one-tail paired t-test, α=0.05). Q: binding capacity; Ci: initial concentration of melamine. ......................................................... 64 xiv  Figure 5.6 Kinetic binding isotherm of MIPs and NIPs (Ci: 12.6 mg L-1). Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3). ..................... 65 Figure 5.7 Representative SERS spectral features of MIPs-AgNPs with 0.1 mmol L-1 melamine in methanol as positive control (a), 0.017 mmol L-1 melamine in skim milk (b) and non-spiked skim milk as negative control (c). ................................................................................................. 67 Figure 5.8 Representative principal component analysis for skim milk samples spiked with different concentrations of melamine (N=9). Different markers represent different melamine concentrations: cross sign, 0.1 mmol L-1; pentagram, 0.025 mmol L-1, diamond, 0.017 mmol L-1; asterisk, 0.01 mmol L-1; and dot, non-spiked samples. Cluster with dashed line is the skim milk samples with melamine concentration lower than 0.01 mmol L-1, while cluster with the solid line contains the samples with melamine concentration higher than 0.01 mmol L-1. .......................... 68 Figure 5.9 Loading plot of the representative principal component analysis model. Principal component (PC) 1 represents 45.69% variances and PC2 represents 25.11% variances. ............ 69  xv  List of Abbreviations  AgNPs silver nanoparticles AIBN 2,2’-azobis(isobutyronitrile)  CCD charge coupled device DAD photodiode array detector DCC N,N'-dicyclohexylcarbodiimide  DMF dimethylformamide EGDMA ethylene glycol dimethacrylate  ELISA enzyme-linked immunosorbent assay EM field electromagnetic field GC gas chromatography H chain heavy chain HILIC hydrophilic interaction chromatography HPLC high performance liquid chromatography L chain light chain LC liquid chromatography LOD limit of detection LOQ limit of quantification MAA methacrylic acid MIPs molecularly imprinted polymers MIPs-AgNPs molecularly imprinted polymers-silver nanoparticles xvi  MISPE molecularly imprinted solid phase extraction MS mass spectrometry MW molecular weight NHS N-hydroxysuccinimide  NIPs non-imprinted polymers NIPs-AgNPs Non-imprinted polymers-silver nanoparticles NISPE non-imprinted solid phase extraction OVA ovalbumin PBS phosphate buffer saline PC  principal component PCA principal component analysis PPM part per million Q value adsorption capacity value R2 coefficient of determination SERS surface enhanced Raman spectroscopy SPE solid phase extraction SPR surface plasmon resonance UV ultraviolet  xvii  Acknowledgements  Two years, time really flies, but I have already owed a great debt of gratitude to people around me.   I would like to offer my sincere gratitude to my supervisor, Dr. Xiaonan Lu, for giving me the opportunity to study food science in your Food Safety Engineering Laboratory in the University of British Columbia. Thank you for all your support and encouragement throughout my master’s study. You are like a mentor who has not only taught me the skills and revelations to make my projects go smoothly, but also helped me in many other ways. Thank you for willing to spend several more years supervising my Ph.D. study.   My special thanks also go to my supervisory committee members, Dr. Christine H. Scaman and Dr. Edward Grant, for your time commitment and invaluable comments and suggestions throughout my study.  I am extremely thankful for the financial support from National Sciences and Engineering Research Council of Canada, Mitacs Canada, and UBC Food Nutrition and Health Vitamin Research Fund, as well as scholarships provided by UBC and our faculty – Faculty of Land and Food Systems.   I thank all the Lu lab members for your help and support for all the years we spent together. I treasure all the laugher we have shared.  xviii   Lastly, very special thanks are owed to my friends, especially Dan Peng, Qian Zhang and Xiaoyue Liu, and my beloved parents, Xiaoxiang Hu and Fengling Li, for your love, care, and support, both mortal and financial, as well as bearing all my bad tempers. You made my journey of graduate study full of courage and happiness. xix  Dedication  I dedicate this thesis to my family for their unconditional love, encouragement and support. 1  Chapter 1: Introduction   Melamine issues in the food industry 1.1Melamine, a triazine compound, is widely used for the industrial production of melamine resins. However, it was added deliberately to protein-rich foods for economically motivated adulteration to increase the amount of apparent protein content determined by Kjeldahl method (1). Protein contains 16% nitrogen on average while melamine contains 67% nitrogen by weight. Although melamine itself is of low toxicity, it is metabolized to cyanuric acid in the body and form low soluble crystals, resulting in renal failure (2). This caused thousands of injuries and deaths of babies and pets due to the consumption of infant formula and pet foods adulterated by melamine (3, 4). Reports published in 2014 indicated that milk candies tainted with melamine were discovered in a few Asian countries, including China and Malaysia. Although in principle melamine can be transferred into food systems from packaging or other industrial sources, and trace levels are allowed in foods, it is universally banned to be used as a substance for intentional adulteration. According to World Health Organization, the maximum melamine residue allowed in dried infant formula and other dairy products are 1 parts per million (ppm) and 2.5 ppm (5), respectively. Assured conformance with these standards requires an accurate and highly efficient approach for the detection and quantification of trace level of melamine in food systems.   Traditional techniques for melamine determination 1.2To determine melamine in dairy products, an approach combining a sample pretreatment method, a separation technique and a detection tool is required. Before applying liquid chromatography (LC) or gas chromatography (GC) to separate targeted molecule (i.e. melamine), sample 2  pretreatment methods, such as liquid-liquid extraction and solid phase extraction (SPE), are required to extensively remove dairy matrices (e.g. lipids and proteins).  Then a detection tool, such as mass spectrometry (MS), photodiode array detector (DAD) and/or ultraviolet (UV) detector, is used to identify melamine by detecting the retention time on LC or GC columns. LC or GC based approaches are adopted by most of the government laboratories because of the high reproducibility and sensitivity (6-8), resulting in a limit of detection (LOD) and limit of quantification (LOQ) in liquid milk as low as 0.3 μg L-1 (9).   However, these approaches lack specificity towards targeted molecules. The materials used in sample pretreatment and separation techniques may not fully recognize the analyte by its comprehensive chemical and physical attributes. Some detection tools, such as DAD and UV spectrometers, may not be able to differentiate structural analogues. These inherent properties of LC and GC hyphenated methods make these analytical approaches labor-intensive and time-consuming. In addition, the complex instrumentations demand highly trained personnel to perform the overall analysis, which contradicts with high-throughput and accurate determination required by the food industry and government laboratories. Development of simple, rapid and accurate techniques for the determination of trace level melamine in agricultural and food products is therefore of significant importance.    3   Innovative techniques for melamine determination 1.3 1.3.1 Surface enhanced Raman spectroscopy (SERS) Raman spectroscopy has been recognized as a rapid and non- or less-destructive technique for the detection and quantification of chemicals and microbes (10-12). Raman spectra demonstrate the vibrational modes of functional groups of molecules in the sample. As a Raman-active molecule, melamine has been determined by identifying its distinct Raman bands at wavenumbers of 674 cm-1 and 981 cm-1, which are associated with the triazine ring-breathing mode II of in-plane deformation and triazine ring-breathing mode I, respectively (13, 14). However, due to small inelastic Raman scattering cross-section (15), normal Raman spectroscopy resulted in a high LOD of 0.13% (w/w) (corresponding to 1300 ppm) melamine in dried milk powder (14), which largely exceeded the government regulation (i.e. 2.5 ppm) and makes normal Raman spectroscopy inapplicable for the determination of trace-level melamine in dairy products.   Surface enhanced Raman spectroscopy (SERS) has been widely investigated owing to its significant enhancement of the intensity of faint Raman scattering signals. Noble metallic (e.g., silver and gold) nanostructures, serving as SERS-active substrates, generate a localized surface plasmon resonance (SPR) that induces an electromagnetic (EM) field on the surface of substrates (16). Analytes approaching to the EM field present significantly enhanced Raman scattering cross-sections, which lowers LOD to ppb or even single molecule levels (15). This unique feature has contributed to fast development and application of SERS in food safety inspection, such as detection of foodborne pathogens (17), chemical contaminants (18), and food adulterants 4  (19). Numerous studies have been conducted to evaluate the potential of using SERS to determine melamine with a variety of metallic nanostructures (20-24), and the LOD were lower than 2 mg L-1 (i.e. 2 ppm) in milk, conforming the regulation for melamine in dairy products (i.e. 2.5 ppm).  1.3.2 Techniques for analyte extraction  1.3.2.1 Antibody Immunoglobulins (Igs), produced by adaptive immune system of animals, can be divided into two groups: secreted Igs and membrane-bound Igs. Secreted Igs (the so-called antibodies) bind, react, and damage antigens by effectively covering the surface of antigens thus blocking their biological functions, or stimulating other innate immune system (e.g. complement system). Another group of Igs locates on the membrane of B cells, recognizes and binds antigens, followed by evoking the production of antibodies through the secondary immune response (25). The basic structure of antibodies contains two identical heavy (H) chains and light (L) chains arranging in Y-shaped H-L pairs by disulfide bonds. H chains of the antibodies are usually composed of three or four constant domains and one variable domain, while L chains have one constant domain and one variable domain. The base of the Y-shaped antibody (i.e. Fc or fragment crystallizable region) is consisted of two H chains contributing two or three constant domains, imparting biological functions (e.g. interact with cell receptors and differentiate various antibody classes), whereas a pair of L chains and the rest domains of the two H chains (i.e. one constant and one variable domains) make up the arm of the Y-shape, providing specific binding affinity towards antigen (i.e. Fab or fragment antigen-binding region).  5   To produce the antibody for a specific molecule, designing an appropriate immunogen that can trigger animals’ immune responses is necessary. Molecules with strong immunogenicity require the following attributes (26): to be recognized as “foreign”, large molecular weight (MW) (i.e. at least larger than 1000), and a certain extent of structural complexity. Molecules lacking any of the aforementioned properties, such as antibiotics and pesticides with small MW, can still elicit the immune response to produce the corresponding antibody by conjugating with a suitable protein carrier and forming the immunogen (27, 28). Then, animals are vaccinated with an adequate amount of immunogen to trigger the immune response and produce sufficient antibodies. The immunization routes, amounts of immunogen injected, and immunization schedules are affected by different animal species, immunogens used, and the quality and quantity of the antibody required (25, 29, 30).   Based on the specific affinity towards target molecules, antibodies are not only applied in biological or medical research for biomarker recognition (31), isolation (32), and detection (33), but also extensively used in immunological biosensors for the determination of biological and chemical hazards in food matrices (34-37). Enzyme-linked immunosorbent assay (ELISA), which uses enzyme-linked antibodies, reported the appearance of the target molecule (e.g. ochratoxin A, aflatoxin M1 and erythromycin) by the changes in the color of substrate after being metabolized by the enzyme (27, 38, 39). Besides transducing chemical signals into colorimetric signals, other forms of transducers were also broadly used in food safety detection. For examples, fiber optic immunosensors transducing chemical interaction energy to fluorescent signals determine triazine and Listeria monocytogenes (40, 41). Surface plasmon resonance 6  immunosensors were used to determine the changes in the mass concentration after forming analyte-antibody complex, resulting in accurate detection of 2,4-dichlorophenoxyacetic acid (i.e. a toxin) and Salmonella enterica serovar Typhimurium (42, 43). Electrochemical sensors converted antibody-antigen interactions to the changes in potential or current to detect aflatoxin M1 and aflatoxin B1 (44, 45). Changes in frequencies of acoustic waves were collected in piezoelectric immunosensors for analyzing Escherichia coli O157:H7 and chloramphenicol after the binding between analytes and antibodies coated on the quartz crystal microbalance (46, 47). In addition, metallic nanostructures can be combined with antibodies, generating SERS signals, for the detection of ovalbumin, ricin and clenbuterol in different food matrices (48-50).   Approaches based on immunoassay have been extensively investigated for the detection of melamine in dairy products and pet food after melamine adulteration incidents (51-56). Monoclonal antibodies and polyclonal antibodies were produced by immunizing experimental animals (e.g. New Zealand rabbit or Balb/C mouse) with melamine hapten-carrier protein complex (51, 57). Melamine haptens were conventionally derived from melamine by using glutaraldehyde method or from 2-chloro-4,6-diamino-1,3,5-triazine  by adding 3-mercaptopropionic acid, and the carrier proteins used were ovalbumin (OVA) or bovine serum albumin (57). Detection tools involved in melamine immunosensors included SPR (53), colorimetric probe using gold colloid (58), chemiluminescent detection (59) and fluorescence analysis (54).  7  1.3.2.2 Molecularly imprinted polymers (MIPs) Molecularly imprinted polymers (MIPs), known as “artificial antibodies” (60), are of increasing interests to researchers in a variety of fields, because of their specific affinity to targeted analytes and high stability compared to antibodies at extreme conditions (e.g. extreme temperature or pH). The synthesis of MIPs includes the following procedures: 1) interaction of template molecule (i.e. analyte) and functional monomer; 2) copolymerization of functional monomer and cross-linking agent; 3) removal of template molecule. With the existence of template molecule, functional monomer and cross-linking agent form complex around template molecule (61). Upon the removal of template molecules by appropriate solvent of choice, cavities with chemical and physical specificities to template molecule are therefore exposed. These cavities act as “locks” to exclusively recognize and bind analyte that share the same properties as template molecule, which serves as the “keys”. By employing these cavities, separation and enrichment of analyte from complex biological matrices can be achieved.   MIPs are normally categorized by the types of bonds involved between the template and the functional monomer, as well as the polymerization mechanism of the polymers. MIPs involving covalent bonds between the template and the functional monomer show higher affinity to template molecule, and more homogeneous binding sites. However, the utilization of this type of MIPs is restricted due to the strict conditions required for the removal and rebinding of MIPs and analyte (62). In contrast, non-covalent MIPs bind template molecule through hydrogen bonds and van der Waals forces. Non-covalent bonds are easier to break and regenerate, resulting in more common use in real applications. Nevertheless, non-covalent MIPs has less specificity to target (i.e., template molecule) and binding sites distribution lacks uniformity. The conventional 8  approaches used to synthesize MIPs include bulk polymerization, precipitation polymerization, and several others for producing MIPs films. Generally, each type has its advantages and drawbacks. Bulk polymerization generates monolith MIPs, followed by grounding and sieving to obtain fine particles. Although the synthesis of MIPs is easy, this approach gives MIPs with irregular sizes and shapes, and damaged binding sites (63). Precipitation polymerization provides MIPs with regular size of fine particles. With large amount of organic solvent, the particles become more insoluble along with the growth of its polymeric chains. The drawback for precipitation polymerization is the large amount of template and organic solvent requirement (64). MIPs films synthesized by sol-gel polymerization (65), electrochemical polymerization (66) and in-situ polymerization (67) generate porous polymer films with controllable thickness, provide more binding sites and exhibit more rapid equilibration rate. However, the sophisticated instrumentation and specific materials required by these techniques limit their practical applications (68).   Applications of MIPs include stationary phase for chromatography (69), drug delivery (70), clean-up material for sensors (71), and sorbents for SPE (72). In food safety analysis, MIPs are generally used as materials with specific affinity to remove food matrices, and extract and concentrate trace-level analytes. Compared to the traditional materials used as sorbents in SPE columns, the specific affinity of MIPs towards analytes results in higher extraction and enrichment effectiveness. Efficient and effective sample matrices clean-up and analyte extraction can be achieved by MISPE columns. Extracts from MISPE can be detected by conventional and advanced analytical methods, such as high performance liquid chromatography (HPLC) coupled with mass spectrometry (MS) (73) or Raman spectroscopy (74). Besides, MIPs can be 9  synthesized by in-situ polymerization and applied as stationary phase in HPLC columns, thus a conventional detection tool (e.g. DAD or UV) coupled with MIPs-HPLC can be adopted for the detection. Moreover, other detection techniques were also investigated to couple with different forms of MIPs. These included quantum dot coated MIPs films fluorescence sensors to determine ractopamine and carbaryl (75, 76), quartz crystal based MIPs gravimetric sensors for the determination of caffeine and atrazine (77, 78), and graphite electrodes modified with MIPs electrochemical voltammetric sensors to sense theophylline and artemisinin (67, 79).   MIPs with specific affinity to melamine have been successfully developed by different research groups. MIPs, synthesized by bulk polymerization using melamine as template or cyromazine as pseudo template molecule, have been used as the sorbent in SPE to extract melamine from liquid milk and infant formula powder (73, 80, 81). A glass stir bar coated with bulk MIPs was used for stir bar sorptive extraction of melamine from animal feed and milk samples. The thickness of the polymers was controlled by the size of polytetrafluoroethylene (PTFE) mold used to hold the glass stir bar (82). By increasing the volume of porogen used during polymerization, MIPs microspheres synthesized by precipitation or suspension polymerization were also investigated to recognize and bind melamine in milk (83, 84). In addition, sol-gel polymerization and electrochemical polymerization were employed to fabricate MIPs films to determine melamine in milk by combining a chemiluminescent or a piezomicrogravimetric sensor, respectively (85, 86).  10  1.3.3 Antibody or MIPs coupled SERS biosensors Based on the specific affinity to target molecules, antibodies and MIPs have been incorporated in many biosensors for the detection of various biological and chemical hazards in foods by applying different detection techniques as stated in Section 1.3.2.1 and 1.3.2.2. However, most of these techniques require additional chemicals (e.g. enzyme-substrate and fluorophore) to transduce the molecule interaction signals to other types of signals (e.g. optic signals) that can be collected by detectors, which could significantly increase the time for analysis. Moreover, some techniques demand complex, time-consuming and sophisticated coating or modification procedures to immobilize antibodies or MIPs on the surface of probes to exclusively bind analyte in samples (e.g. coating quartz crystal microbalance and electrode for gravimetric and electrochemical sensors). These aforementioned attributes make some techniques lack the potential for accurate, low-cost and rapid detection required by the food industry and government laboratories. Therefore, biosensors coupling antibodies or MIPs with a simple, user-friendly, rapid, and sensitive detector are highly demanded.  SERS, as described in Section 1.3.1, is a simple, rapid, and sensitive detection technique. With the development of portable and handheld Raman spectrometers, determination of analytes by SERS is becoming more affordable for high-throughput analysis needed by the food industry and government laboratories. Based on the principle of SERS, collecting vibrational signals associated with numerous functional groups in molecules (16), both targeted analytes and interferences derived from food components (e.g., proteins, lipids, polysaccharides, nucleic acids) contribute to SERS spectral features. Extensive sample pretreatment is therefore required to eliminate the interferences and obtain accurate and reliable sensing results. Common methods 11  for sample clean-up and analyte extraction have no specificity to analytes and require large amounts of solvent with laborious extraction procedures (e.g., liquid-liquid extraction). New approaches such as antibody-based separations and MIPs-based separations have been applied to separate and enrich food chemical and microbiological hazards, such as ovalbumin (48), E. coli (87) , α-tocopherol (74), and chloramphenicol (88), enabling the subsequent use of SERS for detection.   Applications of coupling antibodies or MIPs with SERS can be classified into two categories based on the combined approaches used to connect the separation and detection techniques. “Two-step” biosensors have two sensing elements. Briefly, antibodies or MIPs were used in immunological separation or SPE to specifically separate targeted analyte from complex sample matrices. Then the second element, SERS-active substrates were applied to enhance the Raman scattering signals of chemicals appearing in the eluent samples. Studies applying “two-step” antibody-SERS or MIPs-SERS biosensors have been conducted to determine α-tocopherol (74), chloramphenicol (88), SudanⅠ(89), ricin (49), and ovalbumin (90). In contrast, “one-step” biosensors conjugate the two elements into a complex possessing properties for specific recognition and Raman signal enhancement, such as antibody-SERS biosensors for the cocaine metabolite benzoylecgonine (91), E. coli (87), and bovine leukemia virus (92), as well as MIPs-SERS biosensors for theophylline (71), S-propranolol (93), 2,4,6-trinitrotoluene (94), and bisphenol A (95).  12  Chapter 2: Hypothesis and experimental overview    Rationale 2.1Because of the potential detriment of melamine to human body and complexity of sample pretreatment and instrumentation for the determination of melamine, the development of rapid and accurate detection and quantification techniques are highly demanded.   Raman spectroscopy is generally regarded as a non- or less-destructive detection technique that requires fewer sample pretreatment compared to conventional detection techniques such as MS and DAD (10-12, 96). As a Raman active molecule, melamine has an intense and specific Raman feature band at the wavenumber of 682 cm-1 associated with ring-breathing mode II involving the in-plane deformation of the triazine ring (22). Therefore, Raman spectroscopy has the potential to be used to determine melamine. However, due to the weak inelastic Raman scattering (90), normal Raman spectroscopy is not applicable to determine trace level of melamine residue. By using noble metallic nanostructures, weak Raman scattering signals can be tremendously enhanced by plasmon resonance on the surface of metallic nanostructures, achieving single molecule detection (15) – the so-called “SERS”. Many studies have been conducted using SERS to determine melamine in a wide range of food matrices (6, 20, 97, 98).   However, the accuracy of SERS can be influenced by food components/residues in the sample. Based on the principle of SERS, SERS-active molecules contribute to the inelastic scattering signals. Therefore, sample pretreatment or analyte extraction is still necessary before illuminating the sample for Raman spectral collection. To simplify sample pretreatment and 13  make melamine extraction more efficient, separation elements with specific affinity to melamine are required. Antibodies have high sensitivity and specificity towards analytes, which can be used for specific extraction of melamine from dairy products. MIPs, known as “artificial antibodies”, also exhibit specific affinity towards analyte. Although MIPs have less sensitivity and specificity to analyte than antibodies, the synthesis of MIPs is simpler and MIPs are more stable in the harsh condition (e.g. extreme pH and temperature).  Based on the attributes of melamine and advantages of the detection technique (i.e. SERS) and two separation methods (i.e. antibodies and MIPs), biosensors combining antibodies and SERS or MIPs and SERS were investigated for rapid and accurate determination of melamine in dairy products as my master’s thesis. Dairy products (i.e. whole milk and skim milk) were selected as the food matrices because the adulterations in milk by melamine were mostly discovered and World Health Organization has specific regulations for this category of food (5). Studies of using SERS for the detection or antibodies or MIPs for the separation of melamine have been published, but no research has been conducted integrating these techniques to determine melamine in dairy products.   Objectives 2.2Objective 1: To develop a “two-step” antibody-SERS biosensor for rapid detection of melamine in whole milk by using an antibody for immunological separation and silver (Ag) dendrite (i.e. SERS-active substrate) for detection.  14  Objective 2: To develop a “two-step” MIPs-SERS biosensor for rapid detection of melamine in whole milk by using MIPs for SPE and Ag dendrite (i.e. SERS-active substrate) for detection.  Objective 3: To develop a “one-step” MIPs-SERS biosensor for rapid detection of melamine in whole milk by using silver nanoparticles (AgNPs) incorporated MIPs for simultaneous extraction and detection.   Hypothesis 2.3Hypothesis 1: The antibody produced using melamine hapten and MIPs synthesized by bulk polymerization have the specific affinity to melamine.  Hypothesis 2: The antibody and MIPs can extract and concentrate melamine from dairy products.  Hypothesis 3: Ag dendrite and AgNPs incorporated in MIPs can serve as the SERS-active substrates to sense trace level of melamine in dairy products.   Experimental overview 2.4Three types of biosensors were developed to match the objectives of this study and the experimental design was based upon the verification of each hypothesis.   The melamine antibody (i.e. anti-melamine) was produced according to the previous study (53). Anti-melamine formed a complex with protein G Sepharose 4 Fast Flow and was used to capture 15  melamine from whole milk samples. After eluting melamine from the complex, the eluents were deposited onto silver dendrite nanostructure. SERS spectra of the eluents were collected to validate the specific affinity of the antibody towards melamine and the efficiency of this immunological separation in the aspect of analyte extraction and food matrix removal. The LOD of this biosensor was determined by principal component analysis (PCA).    MIPs for the “two-step” MIPs-SERS biosensor were synthesized by bulk polymerization and applied as the sorbent in SPE following the study by Yang and others (73). The specificity of MIPs to melamine was confirmed by comparing the adsorption capacity at various melamine concentrations with NIPs (i.e. non-imprinted polymers). NIPs were synthesized without the addition of melamine during the polymerization. The recoveries of melamine from milk by applying MISPE were analyzed using HPLC-DAD. Finally, the eluents from MISPE were deposited onto silver dendrite for SERS spectral collection. PCA was used to determine LOD and linear regression model was constructed to reveal the relationship between SERS signal intensities and melamine concentrations in milk samples.  “One-step” MIPs-SERS biosensor was fabricated with addition of AgNPs precursor (i.e. AgNO3) during the polymerization based on the study by Liu and others (71). Theoretically, AgNPs located close to cavities specifically for melamine in MIPs provide the EM field to generate significantly enhanced melamine Raman signals because of the binding between melamine and silver ions during the polymerization step. Thus, after confirming the desired specific affinity of MIPs-AgNPs towards melamine by adsorption capacity tests, MIPs-AgNPs were applied to 16  extract melamine in milk and illuminated by Raman incident laser directly. SERS spectra were collected to verify the successful extraction of melamine from milk and calculate the LOD. 17  Chapter 3: Rapid detection of melamine in whole milk using a “two-step” antibody-surface enhanced Raman spectroscopic biosensor   Introduction 3.1As indicated in Section 2.1, a rapid and accurate detection technique is highly demanded for the determination of melamine in dairy products. To shorten the analysis time, more efficient separation technique and more rapid detection technique are required. Immunological separation using antibodies for the recognition and binding of the target molecule (i.e. analyte) has high efficiency for analyte extraction, due to high specificity and sensitivity of the antibody towards analyte. With the assistance of metallic nanostructure, faint Raman signals can be enhanced significantly, resulting in the detection of trace level chemicals, which is ideal for the applications required in food safety detection. Recent studies have confirmed the feasibility to integrate immunological separation and SERS for the determination of ricin (49) and foreign proteins (48, 90) in milk.  The objective of this study was to develop a “two-step” antibody-SERS biosensor for rapid detection of melamine in whole milk by using an antibody for immunological separation and Ag dendrite as SERS-active substrate for detection.   Materials 3.2Melamine, disodium phosphate, monosodium phosphate, sodium chloride, ovalbumin (OVA), glycine, hydrochloric acid, silver nitrate, and zinc plate were purchased from Sigma-Aldrich (St 18  Louis, MO, USA). Protein G Sepharose 4 Fast Flow and Protein A Sepharose 4B were purchased from GE Healthcare Life Sciences (Piscataway, NJ, USA). N-hydroxysuccinimide (NHS), N,N-dimethylformamide (DMF), and N,N'-dicyclohexylcarbodiimide (DCC) were obtained from General Electric Co. (Tianjin, China). Whole milk (3.25% fat) samples were purchased from local food markets (Tianjin, China).   Methods 3.3 3.3.1 Preparation of immunogen The synthesis of melamine hapten (i.e. 3-(4,6-diamino-1,3,5-triazin-2-methylthio) propanoic acid) was conducted following the protocol described in our recent work (53). The chemical structures of melamine and its corresponding hapten are shown in Figure 3.1. Hapten (4 mg) and NHS (3 mg) were dissolved in 200 µl DMF, followed by addition of 5.4 mg DCC. After 20 h of stirring at 4°C, the mixture was centrifuged at 4,500 ×g for 10 min at 4°C and the supernatant was collected. The supernatant was then mixed with 13.5 mg ovalbumin (OVA) in 1.8 mL phosphate buffer saline (PBS, pH 7.2) on ice bath. The reaction mixture was kept stirring at 4°C overnight and then dialyzed in PBS (pH 7.2) for three days. The formation of melamine hapten-OVA immunogen was confirmed by UV spectrometer (Appendix A.1)1 and stored at -20°C.                                                  1 Appendix A.1 was not presented in the published paper (Rapid detection of melamine in milk using immunological separation and surface enhanced Raman spectroscopy) and was added for this thesis. 19  NNNNH2NH2H2N     NNNSHOOH2N NH2 Melamine           Hapten Figure 3.1 Chemical structures of melamine and corresponding hapten.  3.3.2 Polyclonal antibody production Female New Zealand white rabbits were immunized by intradermal and intramuscular injections. The initial immunizing dose consisted of 2 mL melamine hapten-OVA immunogen (1 mg/mL) and 2 ml Freund’s complete emulsion. Subsequent booster injections were performed 2, 4, 6 and 8 weeks later. Blood was collected from the marginal ear vein 8-10 days after each injection in the 4th, 6th and 8th week. The plasma was collected by centrifugation at 4,500 ×g for 10 min at 4°C. Antibody titer and affinity were measured by indirect ELISA and antiserum was then purified using affinity chromatography on Protein A Sepharose 4B. The detailed procedures can be referred to our recent work (53).  The concentration of the purified antibody was determined by the ultraviolet absorbance at 280 nm in wavelength and then stored at 4°C. Six rabbits were immunized with melamine hapten-OVA immunogen and the polyclonal antibodies (i.e., anti-melamine) were collected from the 4th bleed of individual rabbits.  3.3.3 Whole milk sample pretreatment Whole milk samples were spiked with different concentrations of melamine ranging from 0.79×10-3 mmol/L to 0.16 mmol/L. An aliquot of 1 mL spiked milk samples were centrifuged at 20  12,000 ×g for 5 min at 22°C. Three layers (i.e. lipid layer, aqueous layer and protein pellet) were formed in the samples and the middle layer (aqueous layer) was collected for further analysis.  3.3.4 Immunological separation of melamine in whole milk Immunological separation procedure was illustrated in Figure 3.2. In brief, pre-swollen protein G Sepharose 4 Fast Flow was rinsed with the antibody binding buffer [(0.2 M Na2HPO4 (30.5 mL) mixed with 0.2 M NaH2PO4 (19.5 mL), pH 7.0] to remove 20% ethanol. Then 50 µL protein G Sepharose 4 Fast Flow was mixed with 50 µL anti-melamine in 400 µL antibody binding buffer and incubated for 30 min using shaker at 110 rpm, 22°C, forming the converted protein G-anti-melamine complex. Whole milk samples spiked with melamine (0.5 mL) was mixed with the converted protein G-anti-melamine complex and incubated for 10 min at 22°C with shaking at 110 rpm. After washing protein G-anti-melamine-melamine complex three times with PBS (pH 7.2) followed by centrifugation at 4000 ×g for 1 min at 4°C, the supernatant was discarded and 1 mL elution buffer (0.375 g glycine in 50 mL H2O, adjusted pH to 2.7 by 4 M HCl) was added, incubated for 5 min with shaking at 110rpm, 22°C, followed by centrifugation at 4000 ×g for 1 min at 4°C. The supernatant was the final immunological separation eluent and was deposited onto SERS-active substrate for Raman spectral collection. All experiments for different concentrations were conducted in triplicate. 21   Figure 3.2 Schematic illustration of immunological separation integrated with surface enhanced Raman spectroscopy to determine melamine in milk.  3.3.5 Synthesis of silver dendrite SERS substrate Silver dendrite SERS substrate was synthesized by replacement reaction following the protocol by Feng and others (74). The zinc foil (99.99% purity) was immersed into 0.02 mol L-1 HCl and ultrasonicated overnight to remove the oxidative products and contaminations on the surface. After rinsing with distilled water, the pretreated zinc foil was soaked and reacted with 200 mmol L-1 AgNO3 solution at room temperature for 1 min. Then, the silver dendrites formed on the surface of zinc foil were gently peeled off and rinsed with distilled water to remove excess reagents. Homogeneous nanostructure was formed after 20 min ultrasonication to break the silver dendrites. Finally, these silver dendrite products were stored in a glass vial at room temperature and it can be used as SERS substrate for more than 6 months (74). For Raman spectral collection, 3 mg of silver dendrite was deposited onto a gold-coated microarray chip 22  (Thermo Electron, Waltham, MA, USA) and dried with the stream of nitrogen, followed by loading 2 μL samples on it.  3.3.6 Raman spectroscopic instrumentation Raman spectra were collected using a confocal micro-Raman spectroscopic system coupled with a 785 nm near-infrared laser (0.25 mW incident laser power). The spectrometer (Renishaw, Gloucestershire, UK) has an entrance aperture of 50 μm, a focal length of 300 mm, and is equipped with a 1200 line mm-1 grating. The Raman signals were recorded by a 578- by 385-pixel charge coupled device (CCD) array detector, with a pixel size of 22 μm.   After mounting the SERS substrate onto the stage of Leica microscope (Leica Biosystems, Wetzlar, Germany), Raman spectra were collected using a 50× Nikon objective (NA = 0.75, WD = 0.37) with an exposure time of 10 s. At least five spectra were collected from each replicate of the experiments.  3.3.7 Spectral analysis and chemometric models Vancouver Raman Algorithm software (BC Cancer Agency & University of British Columbia, Vancouver, BC, Canada) was applied for six-order of polynomial fit baseline correction and smoothing (11-Size Boxcar algorithm) of collected Raman spectra. These spectral preprocessing can reduce spectral noise and increase spectral signal-to-noise ratio (74). SERS band at 900 cm-1 is derived from glycine in elution buffer and it can be used as internal standard for spectral normalization before chemometric analysis (99). PCA was applied to differentiate milk samples 23  with different concentrations of melamine using Delight version 3.2.1 (Textron Systems, Wilmington, MA, USA) software.   Results and discussion 3.4 3.4.1 Specificity test In the current study, a polyclonal antibody was used to capture melamine from milk. Although polyclonal antibodies can be used for various applications to extract different antigens, they have less specificity compared to monoclonal antibodies (100). Therefore, the evaluation of antibody specificity towards the target molecule is necessary. Representative SERS spectra for silver dendrite (SERS-active substrate), eluents of two negative controls (i.e. milk and water without melamine), elution buffer, the eluent from milk sample spiked with 0.79×10-3 mmol L-1 melamine with and without additional antibody, and the positive control (i.e. 0.16 mmol L-1 melamine) are shown in Figure 3.3. For the immunological separation, elution buffer was added to separate anti-melamine and melamine from protein G. Thus, elution buffer was shown in every eluted sample and contributed to the final SERS spectra. The SERS band at 900 cm-1 is associated with the -NH2 twisting and -CH2 twisting of glycine (101) in elution buffer, and was employed as the internal standard to normalize SERS spectra and eliminate the experimental and systematic errors. 24   Figure 3.3 Representative normalized SERS spectral features of eluent from milk spiked with melamine (0.79 μmol L-1) by immunological separation using water, milk, and melamine (0.16 mmol L-1 in water) as negative and positive controls, and elution buffer as background: a) silver dendrite SERS-active substrate, b) elution buffer, c) water negative control, d) milk negative control, e) melamine eluent from milk without the antibody, f) melamine eluent from milk with the antibody, and g) melamine in methanol (positive control).  Without the existence of antibodies, melamine cannot bind to protein G and remain in the pellet after washing, resulting in the disappearance of melamine feature band in Figure 3.3. Water and milk without the addition of melamine were two negative controls and they exhibited similar SERS spectral features, indicating extremely low non-specific binding between milk components and anti-melamine. Besides, the high similarity between negative controls and elution buffer demonstrated that few interference signals were contributed by anti-melamine. The band at 703 cm-1 in the spectrum of 0.16 mmol L-1 melamine is the featured SERS band of melamine, attributed to the ring-breathing mode II involving the in-plane deformation of the triazine ring 25  (20), while the band at 1072 cm-1 is derived from nitrate residues on silver dendrite (88). The disappearance of the SERS band at 1072 cm-1 in other spectra can be associated with the competitive adsorption between glycine in the elution buffer and nitrate on silver dendrite. Moreover, the high proton concentration of elution buffer (pH 2.7) resulted in the extended aggregation of silver nanostructure (102), leading to the shift of melamine featured SERS band from 703 cm-1 in the positive control spectrum to 680 cm-1 in the spectrum of eluent from milk. This can be attributed to the SERS effect (16). Specifically, the extended aggregation of SERS-active substrate modifies the interaction between target molecule and SERS substrate, contributing to the changes in the polarity of target molecule and consequently influencing the location of SERS bands. However, the melamine featured SERS band in the spectrum of melamine eluent represents the successful extraction of melamine from milk, while the similarity of other bands in the spectrum of melamine eluents from milk to the spectrum of elution buffer demonstrates the thorough removal of food matrices.   Unsupervised PCA was conducted and the cluster model further confirmed this result of specificity test (Figure 3.4). The first principal component (PC1) explained 85.9% of the total variances, while PC2 only contributed to 7.9% variances. Therefore, the variations in PC1 were of more significance compared to PC2 in terms of the discrimination power. These apparent differences in PC1 scores between melamine eluents from milk (ranging from 2 to 4) and negative controls and elution buffer (ranging from -2 to -0.5) exhibited the identity of two negative controls and elution buffer, and the distinction from the melamine eluents from milk. The loading plot of PC1 (Figure 3.5) demonstrates that the differences between spectra were 26  primarily contributed by the band at the wavenumber of 680 cm-1, associating with featured melamine SERS band 2.    Figure 3.4 Representative principal component analysis for negative controls (i.e. milk and water), elution buffer, and milk sample spiked with 0.79 μmol L-1 melamine (N=6). Different marks correspond to different samples: circle: water negative control; diamond: milk negative control; plus sign: elution buffer; asterisk: milk samples spiked with melamine. Cluster with dash line is the milk samples spiked with melamine, while cluster with the solid line contains the negative controls and elution buffer.                                                  2 Data of the PCA loading plot and the corresponding discussion were not presented in the published paper (Rapid detection of melamine in milk using immunological separation and surface enhanced Raman spectroscopy) and were added for this thesis. 27   Figure 3.5 Loading plot of the representative principal component analysis model. Principal component (PC) 1 represents 85.9% variances.  Extraction capacity of the produced anti-melamine towards melamine analogue (i.e. cyanuric acid) was not conducted because false positive results could not be acquired by applying SERS for detection 3. SERS reveals the chemical structure of analytes and provides fingerprinting spectral features to identify chemicals. The wavenumbers of featured SERS bands of melamine and cyanuric acid are different (23), resulting in differentiation of detection of melamine and cyanuric acid. Therefore, even if anti-melamine has partial capacity to recognize and bind cyanuric acid during immunological separation, SERS spectral features can be used to further distinguish cyanuric acid from melamine, avoiding false positive results. Besides, as a metabolite                                                  3 This paragraph, the discussion about testing structural analogues, was not presented in the published paper (Rapid detection of melamine in milk using immunological separation and surface enhanced Raman spectroscopy) and was added for this thesis. 28  of melamine, cyanuric acid also possesses potential detriment to human (103). The false positive results associated with cyanuric acid are also of great significance to assure food safety.  3.4.2 Sensitivity test Milk samples spiked with 0.79×10-3, 4.0×10-3, 7.9×10-3, 0.040, 0.079 and 0.16 mmol L-1 melamine were tested using immunological separation-SERS method. PCA, a multivariate statistical analysis method, was performed to differentiate milk samples containing different contents of melamine. PCA cluster model constructed by SERS spectra ranging of 600 cm-1 to 1000 cm-1 in wavenumbers showed that the proposed immunological separation-SERS method is feasible to distinguish melamine spiked milk samples from non-spiked milk samples (Figure 3.6). The loading plot of PC1 (Figure 3.7) demonstrates that the differentiation was mostly based upon the featured melamine SERS band at wavenumber of 680 cm-1 4. However, all the milk samples spiked with different melamine concentrations were overlapped, indicating that this methodology is not appropriate for the differentiation of different melamine concentrations in milk.                                                   4 Data of the PCA loading plot and the corresponding discussion were not presented in the published paper (Rapid detection of melamine in milk using immunological separation and surface enhanced Raman spectroscopy) and were added for this thesis. 29   Figure 3.6 Representative principal component analysis for milk samples spiked with different concentrations of melamine (N=8). Different markers represent different melamine concentrations: asterisk, 0.79×10-3 mmol L-1 melamine; circle, 4.0×10-3 mmol L-1; plus sign, 7.9×10-3 mmol L-1; cross sign, 0.040 mmol L-1, diamond, 0.079 mmol L-1; hexagram, 0.16 mmol L-1; and pentagram, non-spiked samples. Cluster with dash line is the non-spiked milk samples, while cluster with the solid line contains the samples spiked at different melamine concentrations. 30   Figure 3.7 Loading plot of the representative principal component analysis model. Principal component (PC) 1 represents 82.7% variances.  The LOD determined by PCA model was estimated to be lower than 0.79×10-3 mmol L-1 (i.e. 0.1 ppm) melamine in milk. The maximum residue level regulations for melamine are 1 ppm for infant formula and 2.5 ppm for other dairy products, including liquid milk (104). Therefore, our developed immunological separation-SERS method can satisfy the requirement for the detection of melamine in milk.   Conclusion 3.5In conclusion, our current study validated that immunological separation integrated with SERS could separate and detect trace level of melamine (i.e. lower than 0.79×10-3 mmol L-1) in milk within 20 min. A large amount of milk samples could be screened using this method in a high-throughput manner. Once a sample shows positive melamine signal, quantification analysis with 31  other techniques, such as liquid chromatography-mass spectroscopy, can be applied to further determine the content of adulteration. Further study is proposed to establish a quantitative analysis using this immunological separation-SERS approach and conduct a portable Raman spectrometer for on-site detection in a milk processing plant. The methodology of combining antibody for the separation and SERS for the detection can be generalized to a variety of chemical hazards in foods to fulfill the high throughput and sensitive detection demanded by the food industry and government laboratories.  32  Chapter 4: Rapid detection of melamine in whole milk using a “two-step” MIPs-SERS biosensor   Introduction 4.1Based on the results in Section 3.4, the “two-step” antibody-SERS was not applicable for the quantification of melamine in dairy products. Moreover, the production of antibodies is complicated as described in Section 3.3.1 and 3.3.2, and antibodies denatured easily in harsh conditions (e.g. extreme pH or temperature), making antibodies not ideal for on-site or in-field analysis.   MIPs, recognized as the “artificial antibody”, is a polymer based material that also has specific affinity to analyte. MIPs specifically for melamine have been synthesized successfully by several research groups (73, 83). The synthesis of MIPs can be much simpler than that of antibodies, and the synthesized polymers have higher stableness, giving MIPs a good potential for the real-world application. Recent publications in our group validated the feasibility of using MIPs as the sorbent in SPE followed by using SERS for rapid and accurate detection of α-tocopherol in vegetable oils (74), chloramphenicol in honey and milk (88) and Sudan I in paprika(89).   The objective of this study was to develop a “two-step” MIPs-SERS biosensor for rapid detection of melamine in whole milk by using MIPs as the sorbent in SPE and Ag dendrite as SERS-active substrate for the detection.  33   Materials 4.2Methacrylic acid (MAA), ethylene glycol dimethacrylate (EGDMA), 2,2’-azobis(isobutyronitrile) (AIBN), melamine, silver nitrate, and zinc plate were purchased from Sigma-Aldrich (St. Louis, MO, USA). Methanol (HPLC grade), acetonitrile (HPLC grade), ethanol (HPLC grade), ammonium hydroxide and acetic acid were purchased from Thermo Fisher Scientific (Toronto, ON, Canada). Whole milk (3.25% fat) samples were obtained from local grocery stores. Melamine stock solution (1 mmol L-1) was prepared in methanol, stored at room temperature, and diluted by methanol for further use as working solution.   Methods 4.3 4.3.1 Synthesis of molecularly imprinted polymers The synthesis of MIPs for melamine followed the procedures by Yang and others (73) with modifications. Briefly, 0.5 mmol of melamine (i.e. template) was dissolved in 5 mL of ethanol and 1.5 mL of water (i.e. porogen) using 50 mL round-bottom flask. After adding 6 mmol (509 mL) of MAA (i.e. functional monomer), the mixture was magnetically stirred for 10 min at 155 rpm. Then 30 mmol (5663 mL) of EGDMA (i.e. cross-linker) and 0.25 mmol (0.041 g) of AIBN (i.e. initiator) were added, followed by purging with nitrogen for 10 min to remove oxygen. The solution was sealed and incubated for 24 h at 60°C in oil bath with magnetic stir (155 rpm). The monolithic product was ground and sieved through a 200 mesh (74 microns) steel sieve to obtain the homogeneous fine particles. Soxhlet extraction was conducted to remove the template using 250 mL of methanol/acetic acid (8:2, v/v) for 24 h. Another 250 mL of methanol was used for further extraction of template and removal of acetic acid. Finally, the MIPs were dried for 12 h in 34  a vacuum drying oven at 50°C. The non-imprinted polymers (NIPs) was synthesized following the same process without the addition of template (i.e., melamine). The synthesis of MIPs and NIPs were repeated twice.  4.3.2 Adsorption capacity tests The static adsorption capabilities for MIPs and NIPs were determined by the following procedure. Ten milligrams of MIPs or NIPs was mixed with 2 mL melamine standard solutions at selective concentrations ranging from 0.01 mmol L-1 to 0.06 mmol L-1. The mixtures were twistedly shaken at 200 rpm for 18 h at room temperature. After centrifugation at 12,000 ×g for 3 min, the supernatant was filtered through a 0.22-μm nylon syringe filter (Thermo Scientific, Rockwood, TN, the United states), and the final concentration of melamine was determined by HPLC-DAD set at a wavelength of 240 nm. Two batches of MIPs and NIPs were determined individually for three times and merged together after being confirmed of having the same static adsorption binding capacity.   The kinetic adsorption test was conducted by mixing 10 mg MIPs or NIPs with 2 mL melamine standard solution (0.05 mmol L-1). The mixtures were twistedly shaken for different time intervals ranging from 5 min to 240 min. After the centrifugation at 12,000 ×g for 3 min, the supernatant was treated and determined following the same procedures as for static adsorption. The merged polymers products were tested for three times.  35  4.3.3 Whole milk sample pretreatment Whole milk was spiked with melamine powder for different concentrations ranging from 0.001 mmol L-1 to 0.05 mmol L-1. Melamine was recovered following the procedures by Tran and others (105). Briefly, 0.5 mL of milk was mixed with 4.5 mL of methanol. After vortex shaking for 30 s, the mixture was ultrasonicated for 5 min and then centrifuged at 8000 ×g for 5 min. The supernatant was collected for further analysis.  4.3.4 High performance liquid chromatography conditions HPLC analysis was conducted on an Agilent 1100 series HPLC system with DAD set at a wavelength of 240 nm. Samples (2 μL) were injected into a hydrophilic interaction liquid chromatography column (HILIC) (Waters Atlantis HILIC silica, 5 μm, 2.1 mm × 150 mm, Milford, MA, USA) at 30°C. Each sample was tested in triplicate. The mobile phase was 0.01 mol L-1 ammonium acetate – acetonitrile (10:90, v/v) with the flow rate of 0.2 mL min-1. The total running time for melamine standard solution was 10 min and 22 min for milk samples.  4.3.5 Molecularly imprinted solid phase extraction (MISPE) A portion of 300 mg merged MIPs or NIPs were packed into a 6 mL SPE cartridge with one PTFE frit at each side (Agilent, Santa Clara, CA, USA). Two MISPE and two NISPE columns were used to examine the effectiveness and efficiency of extracting and enriching melamine from samples with various concentrations. The columns were conditioned with 2 mL water and then 2 mL methanol. Each sample (5 mL) was loaded onto the columns, followed by washing with 2 mL water and then 2 mL methanol. After purging all washing solvent out of the column, 3 mL methanol containing 5% ammonium hydroxide was used to elute the analyte (i.e., melamine) 36  adsorbed on the sorbent. The flow rate was 1.5 mL min-1 for each step. An aliquot of eluent (i.e. 2 μL) was dried under stream of nitrogen, redissolved in 2 μL of 90% methanol, and then directly deposited onto silver dendrite SERS-active substrate for spectral collection. .  HPLC-DAD was conducted to determine the recovery of melamine. The eluted solution was evaporated to dryness and redissolved in 1 mL methanol for melamine standard solutions or 0.5 mL methanol for milk samples, followed by HPLC-DAD determination.  4.3.6 Synthesis of silver dendrite SERS substrate Procedures for the synthesis of silver dendrite have been stated in Section 3.3.5.  4.3.7 Raman spectroscopic instrumentation Please refer to the Section 3.3.6 for detailed instrumentation parameters.  4.3.8 Spectral analysis and chemometric models Raw SERS spectra were analyzed by OMNIC software version 7.0 (Thermo-Nicolet, Madison, WI, USA). Automatic baseline correction and spectral smoothing (11-point Savitzky-Golay algorithm) were employed to reduce the spectral noise. Raman band at 1072 cm-1, which was attributed to nitrate on silver dendrite SERS substrate, was used as internal standard to normalize SERS spectra (48).   By using Matlab, unsupervised PCA was conducted to cluster and differentiate whole milk samples with different concentrations of melamine. Linear regression model was constructed to 37  reveal the correlation between spiking melamine content and the relative intensity of featured SERS band at 703 cm-1, which is derived from melamine.   Results and discussion 4.4 4.4.1 Synthesis and characterization of MIPs The schematic illustration of the fabrication and application of MIPs-SERS biosensor is shown in Figure 4.1. MIPs for melamine were prepared by bulk polymerization. Previous studies have confirmed the feasibility of MIPs synthesized for the separation and enrichment of melamine, using techniques that include bulk polymerization (106), precipitation polymerization (83), and polymer films (107). Among these methods, bulk polymerization has successfully yielded a variety of templates (108). It requires no sophisticated instruments and the synthesis procedure is fast and simple (108).  38   Figure 4.1 Schematic illustration of MIPs-SERS biosensor for the detection of melamine in whole milk. CCD: charge coupled device; MIPs: molecularly imprinted polymers; SERS: surface enhanced-Raman spectroscopy.  To evaluate the rebinding properties of the synthesized MIPs, both static adsorption and kinetic adsorption were tested on MIPs and NIPs. The selectivity of MIPs toward analyte molecules is derived from complementary cavities and specific bonds. The cavities for template molecules are produced during polymerization process and remain unchanged, but specific binding sites may fail to interact with the analyte in some circumstances because the solvent used in the rebinding experiments can influence the binding of between analyte and polymers. In general, the solvent used as porogen for the synthesis of MIPs provides the highest specificity in rebinding 39  experiments (109). However, in the current study, when melamine was dissolved in ethanol-water (10:3, v/v), no difference in the adsorption capacity (Q) values was observed between MIPs and NIPs. This result may be attributed to the fact that water in the solvent increases the overall polarity of the mixture and consequently interferes with the formation of hydrogen bonds between the MIPs and the analyte.   In accordance with the study by Yang and coworkers (73), we selected methanol as the solvent for adsorption tests. Figure 4.2 illustrates the static adsorption isotherm of MIPs and NIPs. The Q value was calculated by the following formula: Q = ((Ci - Cf)×V/W) (110), where Ci and Cf represent the initial and final concentration of melamine in the solution, respectively. V is the volume of the solution, and W is the mass of the polymer. Regardless of the initial concentration of melamine, MIPs yield higher Q values than NIPs. QMIPs and QNIPs increase with the increase of melamine concentration, and reach 0.27 mg g-1 and 0.15 mg g-1, respectively, at a melamine concentration of 0.06 mmol L-1. Due to the small size and simple structure of melamine molecule, it fits better to the tiny cavities or gaps on rough materials than larger and more complex molecules, thus adsorbs easier by the polymers through non-specific binding (high NIPs adsorption capacity).5 Therefore, the differences in Q values between MIPs and NIPs are not as significant for melamine as some other targeted molecules (88).                                                  5 This discussion was not presented in the published paper (Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy) and was added for this thesis.  40   Figure 4.2 Static binding isotherm of molecularly imprinted polymers (MIPs) and non-imprinted polymers (NIPs) for melamine. Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3).  We conducted kinetic adsorption tests to assess the rate of development of specific bonds between analyte and MIPs. If the time required to reach equilibrium is short, specific bonds can be formed rapidly, indicating that MIPs are ideal for rapid separation of targeted analyte. Figure 4.3 shows the equilibration isotherm of 10 mg of MIPs and NIPs with 2 mL of melamine (0.05 mmol L-1) for different time intervals at room temperature. MIPs reached to binding equilibrium within 20 min, which was faster than that of NIPs, confirming the potential of using MIPs for fast and specific separation and enrichment of melamine in complicated food matrices. 00.050.10.150.20.250.31.26 2.1 3.15 4.2 6.3 7.56Q (mg g-1) Ci (mg L-1) MIPsNIPs41   Figure 4.3 Kinetic binding isotherm of MIPs and NIPs (Ci: 6.3 mg L-1). Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3).  Q value of the synthesized MIPs towards melamine analogue (i.e. cyanuric acid) was not analyzed. The reason can be referred to Section 3.4.1 6.  4.4.2 MISPE for melamine spiked whole milk The recoveries of melamine in samples after MISPE and NISPE columns were determined by HPLC-DAD. In the preliminary experiment, the flow rate of MISPE procedure was optimized using a melamine standard solution. At a slow flow rate, more melamine was flushed out of the NIPs column before eluting, resulting in a lower recovery from NIPs (Appendix B.1). More weak bonds between melamine and NIPs could be destroyed, with longer interaction time                                                  6 This paragraph, the discussion about testing structural analogues, was not presented in the published paper (Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy) and was added for this thesis. 00.050.10.150.20.250.30.350 50 100 150 200 250 300Q (mg g-1) Time (min) MIPsNIPs42  between the washing solvent and molecules adsorbed on the sorbent, and thus more melamine was flushed away.7 However, one of the objectives in the current study was to create a biosensor with high efficiency. To reduce operation time, the flow rate was increased, leading to an increase in the recovery of NIPs. By setting the flow rate at 1.5 mL min-1, the time required for SPE operation and the differences in recoveries of MIPs and NIPs reached a desirable balance. The recovery of melamine from MIPs using a standard melamine solution of 0.05 mmol L-1 was 106%, while the corresponding recovery from NIPs was 51%.   After the optimization of MISPE procedures, we evaluated the recovery of melamine from whole milk samples. Before loading onto the MISPE column, these samples required a simple pretreatment to remove a variety of macromolecules, including proteins and fats. After spiking with melamine at concentrations from 0.001 mmol L-1 to 0.05 mmol L-1, milk samples were deproteinized and defatted with methanol. The pretreated milk samples were then directly loaded onto the MISPE column following the same procedure as the melamine standard solution. The recovery of melamine in whole milk from MISPE and NISPE ranged from 96% to 108% and 45% to 64%, respectively, as shown in Table 4.1.                                                     7 The Appendix A.1 and the discussion about the recovery were not presented in the published paper (Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy) and were added for this thesis. 43  Table 4.1 Recoveries of melamine in whole milk by molecularly imprinted solid phase extraction (MISPE) and non-imprinted solid phase extraction (NISPE). Concentration (mmol L-1) MISPE Recoveries (%) NISPE Recoveries (%) 0.05 100 49 0.025 107 45 0.01 96 64 0.005 108 62 0.001 N/A* N/A * The recovered concentration of melamine spiked at 0.001 mmol L-1 was too low to be detected by HPLC-DAD.  Traditional methods for extraction and recovery of melamine from milk samples take more than an hour (Health Canada, 2008c). Hydrochloric acid and dichloromethane were used to remove proteins and fats, followed by loading the supernatant onto silica-based SPE column for further separation of melamine from remaining milk residues. Pretreatment of milk samples is not only time-consuming. It involves a large amount of organic solvent, which is not environmentally friendly. A previous study applied trichloroacetic acid to extract melamine from milk. After neutralizing pH, the supernatant was percolated through the MISPE column. The time required for sample pretreatment was over 30 min and the flow rate for MISPE was 1 mL min-1 (111). In the current study, the time for sample pretreatment was only 10 min and the flow rate was higher (i.e. 1.5 mL min-1). The complete recovery of melamine from MISPE procedures validated the feasibility to use methanol for the extraction of melamine and MISPE for sample clean up.  4.4.3 Determination of melamine in whole milk by SERS We have selected silver dendrite as a SERS-active substrate for its easy fabrication and reliable SERS enhancement factor (~104) (74). Figure 4.4 shows SERS spectra collected after depositing 2 μL milk samples eluted from MISPE onto silver dendrite. These average spectra (N=8) show 44  features of melamine crystal determined by normal Raman, standard melamine methanol solution by SERS, and melamine in milk by MIPs-SERS. Each band in the Raman spectrum represents a specific vibrational mode derived from the molecules present, and thus reflect the structural features of the analytes (96) and interferents. We assign the distinct bands at wavenumbers of 674 cm-1 and 981 cm-1 in the normal Raman spectrum to ring-breathing mode II, involving the in-plane deformation of the triazine ring and the triazine ring-breathing mode I, respectively (112). In the SERS spectra of standard melamine methanol solution and melamine in milk after MISPE, the band that occurs at 674 cm-1 in the normal Raman spectrum shifts to 703 cm-1, which can be attributed to a SERS effect (16). When target molecules adsorbed onto the surface of SERS-active substrate, some molecules interact with the noble-metallic nanostructures, resulting in the changes in dipole of the molecules and subsequent shifts in the location of SERS spectral bands. In the MIPs-SERS spectra, the band at 703 cm-1 shows a minor blue shift compared to SERS spectrum of melamine in a standard solution. This blue shift may arise from the presence of Ca2+, Fe2+ and Zn2+ in whole milk (113). We use the band at 1072 cm-1, derived from nitrate residues on silver dendrite, as an internal standard. Although a SERS-active substrate generally acts to significantly increase Raman signal intensity, the band observed at 981 cm-1 in the normal Raman spectrum does not appear in SERS and MIPs-SERS spectra because the surface of the silver dendrite orients the polarization tensor of the melamine molecule orthogonal to polarization of the light field. The uniformity of the bands in SERS and MIPs-SERS demonstrate the effectiveness of MIPs for the separation and enrichment of melamine from whole milk. MIPs-SERS spectra of whole milk spiked with different levels of melamine were collected, and the raw spectra were normalized based upon the band at 1072 cm-1 and are shown in Figure 4.5. The intensity of band at 703 cm-1 decreases with decreasing 45  melamine concentration in whole milk samples. The band intensity for whole milk samples spiked with melamine at the level of 0.001 mmol L-1 (equivalent to 0.126 mg L-1) shows no features that can be assigned to melamine.   Figure 4.4 Representative spectral features of melamine with (a) MIPs-SERS of whole milk sample spiked with 6.3 mg L-1 melamine, (b) SERS of standard melamine solution in methanol (63 mg L-1), and (c) normal Raman of melamine crystal. 46   Figure 4.5 Representative MIPs-SERS spectra of whole milk samples spiked with different concentrations of melamine: (a) 6.3 mg L-1, (b) 3.15 mg L-1, (c) 1.26 mg L-1, (d) 0.63 mg L-1, (e) 0.126 mg L-1, and (f) 0 mg L-1.  Figure 4.6 represents a two-dimensional PCA model that differentiates whole milk samples spiked with different levels of melamine using features at wavenumbers of 650 cm-1 to 750 cm-1. As shown in the loading plot (Figure 4.7), the differentiation of the different melamine concentration is mostly based upon the melamine feature band at 703 cm-1.8 Most of the samples were tightly clustered and clearly separated on the basis of different spiked levels of melamine.                                                  8 The PCA loading plot and the corresponding discussion were not presented in the published paper (Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy) and were added for this thesis. 47  However, cluster overlap appeared between groups with a melamine spike of 0.001 mmol L-1 and samples with no added melamine, indicating that the LOD of our MIPs-SERS biosensor based on the PCA model is 0.005 mmol L-1 (equivalent to 0.63 mg L-1 or 0.63 ppm).   Figure 4.6 Representative two-dimensional principal component analysis for whole milk samples spiked with different contents of melamine. Principal component (PC) 1 represents 70.4% variances and PC2 represents 16.3% variances. Letters from A to F denote whole milk samples spiked with melamine at the concentrations of 6.3, 3.15, 1.26, 0.63, 0.126 and 0 mg L-1, respectively (N=7). 48   Figure 4.7 Loading plot of representative principal component analysis model. Principal component (PC) 1 represents 70.4% variances and PC2 represents 16.3% variances.  Partial least squares regression models were generally employed to construct multivariate calibration model for the quantitative analysis of analyte determined by SERS (114), because the changes in SERS spectral intensity at many wavenumbers follows the same trend, which can be represented by fewer latent variables and correlate better with the analyte concentration measured by a standard method 9. However, based on representative PCA loading plot (Figure 4.7), the band at 703 cm-1 was the major parameter that contributed to the differentiation of milks samples spiked with various melamine concentrations. Therefore, only the standardized SERS band intensity at 703 cm-1 was used to correlate with melamine concentrations for the calibration                                                  9 The discussion on the partial least squares regression was not presented in the published paper (Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy) and was added for this thesis. 49  model. Figure 4.8 shows a linear regression model constructed to correlate the concentration of melamine with the standardized Raman intensity of the SERS band at 703 cm-1 (N = 6). The model shows a good linear relationship for concentration of melamine in whole milk between 0.005 mmol L-1 and 0.05 mmol L-1. The coefficient of determination (R2) of this regression model is 0.93. However, R2 decreased to 0.89 with the range of melamine content extended from 0.001 to 0.05 mmol L-1 (data not shown). We thus conclude that this regression model is applicable to accurately predict the content of melamine in an unknown whole milk sample. Taken together, our MIPs-SERS biosensor can be used for quantification when the concentration of melamine in whole milk sample falls between 0.005 mmol L-1 and 0.05 mmol L-1.  Figure 4.8 Linear relationship between the intensity of SERS band at 703 cm-1 and the spiked levels of melamine in whole milk samples (N=6).  The LOD (0.012 mmol L-1) and LOQ (0.039 mmol L-1) calculated by the three times and ten times of standard deviation of the standardized Raman intensity at 703 cm-1 of non-spiked y = 0.0869x + 0.3915 R² = 0.9293 0.000.200.400.600.801.001.200.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00Raman Intensity (AU) Melamine Concentration (mg L-1) 50  samples (20) are higher compared to those calculated by PCA (0.005 mmol L-1) and linear regression models (0.005 mmol L-1). However, because of the dilution during the sample pretreatment and MISPE, the real concentrations of melamine deposited onto the SERS-active substrate (from 0.00017 mmol L-1 to 0.0083 mmol L-1) were lower than the spiked amount. The corresponding LOD and LOQ are 0.0021 mmol L-1 and 0.0069 mmol L-1, demonstrating high sensitivity of the SERS-active substrate employed in the current study. These results validated that if the procedures for sample pretreatment and MISPE could be optimized, lower LOD and LOQ could be achieved.  HPLC-DAD conducted in the current study required 22 min for the detection of melamine in whole milk, not to mention a long time for sample pretreatment. In comparison, Raman spectral collection significantly shortened the detection time (51 s per spectrum). The overall analysis time for our MIPs-SERS biosensor is about 18 min, including sample pre-treatment. A recent study demonstrated that the LOD and LOQ of melamine in milk was 0.17 mg L-1 and 0.57 mg L-1 (equivalent to 0.0013 mmol L-1 and 0.0045 mmol L-1) by using gold nanoparticles as SERS-active substrate, and the total analysis time was 30 min (20). Owing to the efficient separation and enrichment of melamine by the incorporation of MIPs, the LOD and LOQ of this MIPs-SERS biosensor meets the detection requirements for melamine in dairy products and infant formula demanded by Health Canada regulations (5), which are 2.5 ppm (equivalent to 0.02 mmol L-1 in whole milk) and 0.5 ppm (equivalent to 0.005 mmol L-1 when 12.6 g formula redissolved in 10 mL water), respectively.  51   Conclusion 4.5This “two-step” MIPs-SERS biosensor offers a means for efficient separation, enrichment and accurate detection and quantification of melamine in whole milk. Used as a sorbent in SPE, MIPs achieved effective clean-up of whole milk samples rapidly. SERS, applied for rapid and precise detection of melamine, provides LOD and LOQ values of 0.012 and 0.039 mmol L-1 melamine in whole milk, respectively. In summary, this innovative biosensor shows great promise for the use both in government and food industry laboratories where high-throughput and trace level detection of food chemical hazards is necessary.   52  Chapter 5: Rapid detection of melamine in whole milk using a “one-step” MIPs-SERS biosensor   Introduction 5.1The results in Section 4.4 show the promise of applying MIPs and SERS for the separation and determination of melamine in whole milk. However, by depositing samples onto SERS-active substrate (i.e. Ag dendrite), the accuracy of SERS detection suffers from the poor reproducibility of Ag dendrite, and complicated spectra preprocessing methods (e.g. different combinations of baseline correction, smoothing, binning, second derivatives, and normalization with an internal or external standard applying various chemometric algorithms) are necessary to improve the signal to noise ratio. According to the principle of SERS, only molecules located onto or close to the EM field generated by localized SPR can have significantly enhanced Raman signals. However, variations exist in the synthesis of Ag dendrite, leading to variances in the shape and intensity of the EM field. Thus, the Raman scattering signals of molecules varies from time to time. In addition, by depositing the liquid samples onto Ag dendrite, the final location of molecules on the substrate is random due to the Brownian movement.   “One-step” MIPs-SERS biosensors for theophylline (71), TNT (94), S-propranolol (115) and other molecules have been successfully fabricated by conjugating MIPs with different SERS-active substrates. Through fixing the location of SERS-active substrates on MIPs, distances between the cavities for analyte in MIPs and SERS-active substrates can be controlled, thus analyte adsorbed onto MIPs also located closely to SERS-active substrates (< 10 nm), resulting 53  in improved reproducibility of SERS spectra. Amongst the previous applications, the methodology proposed by Liu and others (71) integrated AgNPs with MIPs by adding silver nitrate as AgNPs precursor that bind to theophylline during bulk polymerization. Melamine contains a triazine ring, which is similar to the imidazole structure in theophylline that interacted with AgNPs precursor during polymerization. In addition, MIPs for melamine have been synthesized successfully by bulk polymerization in Chapter 4:. Therefore, the methodology proposed by Liu and coauthors was adopted in the current research work.  The objective of this study was to develop a “one-step” MIPs-SERS biosensor for rapid detection of melamine in whole milk by using AgNPs incorporated MIPs for simultaneous extraction and detection.   Materials 5.2MAA, EGDMA, AIBN, melamine, silver nitrate (AgNO3), sodium borohydride (NaBH4) and zinc plate were purchased from Sigma-Aldrich (St. Louis, MO, USA). Methanol (HPLC grade), acetonitrile (HPLC grade), ethanol (HPLC grade), ammonium hydroxide and acetic acid were purchased from Thermo Fisher Scientific (Toronto, ON, Canada). Skim milk samples were obtained from local grocery stores. Melamine stock solution (1 mmol L-1) was prepared in methanol, stored at room temperature, and diluted by methanol for further use as working solution.   54   Methods 5.3 5.3.1 Synthesis of molecularly imprinted polymers-silver nanoparticles The synthesis of MIPs-AgNPs for melamine followed the procedures indicated in Section 4.3.1 with modifications. Briefly, 0.5 mmol of melamine (template) was dissolved in 5 mL of ethanol and 1.5 mL of water (porogen) using 50 mL round-bottom flask. After adding 6 mmol (509 mL) of MAA (i.e. functional monomer), the mixture was magnetically stirred for 10 min at 155 rpm. Then 30 mmol (5663 mL) of EGDMA (i.e. cross-linker) was added, followed by incubation at 50°C in oil bath overnight. Subsequently, 500 mg AgNO3 (i.e. AgNPs precursor) and 0.25 mmol (0.041 g) of AIBN were added, and the mixture was purged with nitrogen for 10 min to remove oxygen. The solution was sealed and incubated for 24 h at 60°C in oil bath with magnetic stir (155 rpm). The monolithic product was ground and sieved through a 200 mesh (74 microns) steel sieve to obtain the homogeneous fine particles. Before applying Soxhlet extraction to remove template, excessive NaBH4 (0.1 mol L-1) was added to reduce MIPs-AgNO3 into MIPs-AgNPs. Soxhlet extraction was conducted using 250 mL of methanol/acetic acid (8:2, v/v) for 24 h, and then another 250 mL of methanol was used for further extraction. Finally, the MIPs were dried for 12 h in a vacuum drying oven at 50°C. The NIPs-AgNPs was synthesized following the same procedure without addition of template (i.e., melamine). The synthesis of MIPs-AgNPs and NIPs-AgNPs were repeated twice.    Raman spectra of MIPs-AgNPs and NIPs-AgNPs were collected to confirm the success of AgNPs formation and the complete removal of template molecule. MIPs-AgNPs before and after 55  the reduction were also scanned by UV-Vis spectrometer to confirm the formation of AgNPs. MIPs synthesized for two-step MIPs-SERS biosensor (Chapter 4:) was used as no-AgNPs control.  5.3.2 Adsorption capacity tests Static and kinetic adsorption tests were conducted following the procedures described in Section 4.3.2 with modifications. For static adsorption test, 10 mg of MIPs or NIPs was mixed with 2 mL melamine standard solutions at selected concentrations ranging from 0.01 mmol L-1 to 0.2 mmol L-1. The mixtures were twistedly shaken for 2 h at room temperature. After centrifugation at 12,000 ×g for 3 min, the supernatant was filtered through a 0.22-μm nylon syringe filter (Thermo Scientific, Rockwood, TN, the United states), and the final concentration of melamine was determined by HPLC with DAD set at a wavelength of 240 nm. The pellet (i.e. MIPs-AgNPs and NIPs-AgNPs) was deposited onto gold-coated microarray for Raman spectral collection. . Two batches of MIPs and NIPs were determined individually for three times and merged together after being confirmed of having the same static adsorption binding capacity.    Kinetic adsorption test was conducted by mixing 10 mg MIPs or NIPs with 2 mL melamine solution (0.1 mmol L-1). The mixtures were twistedly shaken for different time intervals ranging from 5 min to 90 min. After the centrifugation at 12,000 ×g for 3 min, the supernatant and pellet were treated and determined following the same procedures as static adsorption aforementioned. The merged polymers products were tested for three times.  56   5.3.3 Skim milk sample pretreatment Skim milk was spiked with melamine powder at different concentrations ranging from 0.01 mmol L-1 to 0.1 mmol L-1. Non-specific sample pretreatment was conducted by using Phenomenex Strate Melamine SPE cartridge (silica-based sorbent) to partially remove milk matrices. Briefly, 1 mL of milk was mixed with 3 mL of acetonitrile. After shaken for 10 s, the mixture was centrifuged at 8000 ×g for 5 min. Two milliliter of the supernatant was loaded onto SPE cartridge. After washing with 1 mL 50% acetonitrile and 0.5 mL methanol, melamine was eluted out by 1 mL methanol with 5% ammonium hydroxide. The eluent was evaporated to dryness before redissolved in 0.2 mL methanol. The recovery of melamine was evaluated by HPLC-DAD. This pretreatment process took ca. 15 min in total.  5.3.4 High performance liquid chromatography conditions Please refer to the Section 4.3.4 for detailed procedures.  5.3.5 MIPs-AgNPs extraction An aliquot of 2.5 mg of MIPs-AgNPs was mixed with 0.5 mL melamine water solution or pretreated skim milk samples. After twistedly shaking for 5 min, the mixture was centrifuged at 12500 ×g for 1 min. The supernatant was discarded and the pellet was deposited onto the gold-coated microarray chip and dried for Raman spectral collection.    57  5.3.6 Raman spectroscopic instrumentation Please refer to Section 3.3.6 for detailed instrumentation parameters.  5.3.7 Spectral analysis and chemometric models Vancouver Raman Algorithm software (BC Cancer Agency & University of British Columbia, Vancouver, BC, Canada) was applied for six-order of polynomial fit baseline correction and smoothing (11-Size Boxcar algorithm) of collected SERS spectra. These spectral preprocessing reduced spectral noise and increased spectral signal-to-noise ratio (74). Unsupervised PCA was applied to differentiate skim milk samples with different concentrations of melamine using MATLAB R2014a (the MathWorks, Inc., Natick, MA, USA) software.   Results and discussion 5.4 5.4.1 Synthesis of MIPs-AgNPs Schematic illustration of the synthesis of MIPs-AgNPs is shown in Figure 5.1. Based on the results in Section 4.4.1, MIPs with specific affinity to melamine had been successfully synthesized by bulk polymerization. Previous research conducted by P. Liu and others (71) validated the feasibility of incorporating MIPs and SERS-active substrate by addition of AgNO3 as the AgNPs precursor during the synthesis of MIPs. The integrated MIPs-AgNPs was then fabricated and used to bind melamine, followed by illumination using Raman spectrometer to detect trace level of melamine.  58   Figure 5.1 Schematic illustration of one-step MIPs-SERS biosensor for the detection of melamine in dairy samples. CCD: charge coupled device; MIPs: molecularly imprinted polymers; SERS: surface enhanced Raman spectroscopy.  Figure 5.2 shows the representative Raman spectra of MIPs-AgNPs (A) and NIPs-AgNPs (B) before and after the reduction of Ag+ (i.e. AgNPs precursor) and Soxhlet extraction, respectively. Before adding NaBH4 for the reduction of Ag+, the Raman spectral intensities were relatively weak because of the faint Raman inelastic scattering. Thus, no melamine feature band was found. However, after the formation of AgNPs, the featured melamine band at 703 cm-1 (i.e. ring-breathing mode II of triazine ring) appeared in the spectra of MIPs-AgNPs. According to the principle of SERS, Raman scattering can be enhanced significantly only when the molecule approaches to the surface of SERS-active substrate, because EM field generated by the localized SPR shows distance dependency (116). With the increasing distance between molecule and SERS-active substrate, the enhancement effect decreases with the squares of distance. Generally, 59  only molecules located within 10 nm to the surface of SERS-active substrates have significantly enhanced Raman scattering signals. Therefore, the appearance of melamine feature band in the spectrum of MIPs-AgNPs after the reduction validated the adjacent location of AgNPs and cavities for melamine. Moreover, the disappearance of melamine feature band in the spectrum of MIPs-AgNPs after Soxhlet extraction demonstrated the thorough removal of template molecule (i.e. melamine). Variations found between the spectra can be attributed to the inhomogeneity of the location of AgNPs in MIPs-AgNPs, leading to the differences in chemical structures being enhanced by SERS effect. Except for some minor variations in the intensity of some bands of NIPs-AgNPs spectra, no obvious change can be found for NIPs-AgNPs spectra before and after the reduction and Soxhlet extraction, indicating that Raman scattering signals contributed by the polymers themselves had no negative effect for the detection of melamine.   60    Figure 5.2 Representative Raman spectra of A) MIPs-AgNPs and B) NIPs-AgNPs before and after the reduction of Ag+ and after Soxhlet extraction. a-f: MIPs-AgNPs before reduction, MIPs-AgNPs after reduction, MIPs-AgNPs after Soxhlet extraction, NIPs-AgNPs before reduction, NIPs-AgNPs after reduction, and NIPs-AgNPs after Soxhlet extraction.  61  The formation of AgNPs can be further validated by the band at 400 nm in UV-Vis spectra, which is contributed by the SPR (71). Figure 5.3 shows the representative UV-Vis spectra of MIPs synthesized for the “two-step” biosensor and MIPs-AgNPs before and after the reduction. MIPs-AgNPs before the reduction of Ag+ shows stronger absorbance at 400 nm compared to the “two-step” MIPs without the addition of AgNPs precursor, demonstrating the partial formation of AgNPs during the polymerization. However, after the reduction, the absorbance at 400 nm further increased and the band became sharper, due to the enhanced SPR by the formation of more AgNPs.  Figure 5.3 Representative UV-Vis spectra of “two-step” MIPs, MIPs-AgNPs before and after the reduction of Ag+. Solid line: two-step MIPs; dotted line: MIPs-AgNPs before the reduction of Ag+; dashed line: MIPs-Ag-NPs after the reduction of Ag+.  62  5.4.2 Characterization of MIPs-AgNPs After the complete removal of template molecule, static and kinetic adsorption capacity tests were conducted for the validation of the specific affinity of MIPs-AgNPs towards melamine and the evaluation of the equilibration rate, respectively. The binding capacity of NIPs-AgNPs was contributed by non-specific bindings that adsorb the analyte without recognizing its exact structure, while MIPs-AgNPs bond analyte by both specific and non-specific bindings. Therefore, MIPs-AgNPs possesses specific affinity towards analyte if its Q value is higher than NIPs-AgNPs. The results of the static adsorption capacity test (Figure 5.4) shows that regardless of the initial melamine concentration, MIPs-AgNPs exhibited higher Q value than NIPs-AgNPs, demonstrating the successful imprint of the chemical and physical properties of melamine in MIPs-AgNPs. Comparing with the “two-step” MIPs and NIPs, these one-step MIPs-AgNPs and NIP-AgNPs showed similar adsorption capacities at the initial melamine concentration of 6.3 mg L-1 (Figure 5.5). 63   Figure 5.4 Static binding isotherm of molecularly imprinted polymers-silver nanoparticles (MIPs-AgNPs) and non-imprinted polymers-silver nanoparticles (NIPs-AgNPs) for melamine. Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3).  00.10.20.30.40.50.60.70 5 10 15 20 25 30Q mg g-1 Ci mg L-1 MIPsNIPs64   Figure 5.5 Comparison of the adsorption capacity of one-step and two-step MIPs and NIPs (Ci: 6.3 mg L-1). Means with different letters are significantly different (one-tail paired t-test, α=0.05). Q: binding capacity; Ci: initial concentration of melamine.  Figure 5.6 shows the results of kinetic adsorption capacity test. Both MIPs-AgNPs and NIPs-AgNPs reached the equilibrium within 5 min and remained stable for longer incubation time. However, NIPs-AgNPs exhibited lower Q value compared to MIPs-AgNPs, further validating the specific affinity of MIPs-AgNPs towards melamine. The short equilibration time indicates the potential for rapid recognition and extraction of melamine from dairy products.  65   Figure 5.6 Kinetic binding isotherm of MIPs and NIPs (Ci: 12.6 mg L-1). Ci: initial concentration of melamine; Q: binding capacity; error bars indicated standard deviation (N=3).  Q value of the synthesized MIPs towards melamine analogue (i.e. cyanuric acid) was not tested. The reason can be referred to Section 4.4.1 10.  5.4.3 Extraction and determination of melamine in water and skim milk by MIPs-AgNPs In the preliminary experiment, whole milk samples spiked with melamine were pretreated with methanol or acetonitrile to remove fat and protein. Syringe filters (2 μm) were also used to eliminate the influence from small protein particles. However, no melamine feature band can be found even after being incubated with MIPs-AgNPs together for 2 h. SPE was used to further remove the food matrices, but no difference was found between the spectra of MIPs-AgNPs after                                                  10 This discussion was not presented in the submitted paper (Rapid detection of melamine in tap water and milk using conjugated “one-step” molecularly imprinted polymers-surface enhanced Raman spectroscopic biosensor) and was added for this thesis. 00.10.20.30.40.50.60.70 20 40 60 80 100Q (mg g-1) Time (min)  MIPsNIPs66  mixing with melamine spiked whole milk and non-spiked whole milk. The specific affinity of MIPs-AgNPs was hindered by whole milk food matrices. This phenomenon could be attributed to the concealing of the specific cavities or the masking of AgNPs by some large molecules. Therefore, skim milk was selected as the food matrix of choice for this study.   After spiking with melamine at different concentrations, skim milk samples were firstly pretreated with the commercial SPE cartridges. Then the eluents were mixed with MIPs-AgNPs for 5 min followed by Raman spectral collection. The overall analysis can be accomplished in 25 min. Representative SERS spectra for 0.1 mmol L-1 melamine in methanol (i.e. positive control), 0.017 mmol L-1 melamine in skim milk and non-spiked skim milk (i.e. negative control) are shown in Figure 5.7. The featured melamine band in the positive control was identical to the band in MIPs-AgNPs after the reduction as shown in Figure 5.2. The intense band at 684 cm-1 in the spectrum of 0.017 mmol L-1 melamine spiked skim milk is also assigned to the ring-breathing mode II of triazine ring in melamine. The shift of the melamine feature band was due to SERS effect (16) and changes of melamine structure in different pH conditions. Melamine spiked in methanol had pH of 5.04, while pH of melamine spiked in skim milk was 6.71. As a base itself, melamine acquires more positive charges in lower pH conditions, thus the dipole of melamine changed accordingly, leading to different Raman scattering behaviors. Moreover, the changes in melamine structure could also change the interaction between melamine and AgNPs, resulting in shift of the wavenumber of melamine feature band. The obvious band in the spectrum of 0.017 mmol L-1 melamine spiked skim milk compared to the non-spiked samples validated the feasibility of using this MIPs-AgNPs to capture and determine melamine in skim milk.  67   Figure 5.7 Representative SERS spectral features of MIPs-AgNPs with 0.1 mmol L-1 melamine in methanol as positive control (a), 0.017 mmol L-1 melamine in skim milk (b) and non-spiked skim milk as negative control (c).  Representative two-dimensional PCA model was constructed to differentiate skim milk samples with various melamine concentrations using spectral feature between wavenumbers of 650 to 750 cm-1 (Figure 5.8). Loading plot (Figure 5.9) demonstrates that Raman intensities of band at 683 cm-1 contributed most to the overall variation of the spectra, which is associated with the featured melamine band. Forty-five spectra for five concentrations were clearly separated into two clusters. Samples with melamine concentrations lower than 0.01 mmol L-1 were grouped together, while samples with higher concentrations have similar PC scores and located closely to each other. Therefore, the LOD based on this PCA model is between 0.01 and 0.017 mmol L-1.  68   Figure 5.8 Representative principal component analysis for skim milk samples spiked with different concentrations of melamine (N=9). Different markers represent different melamine concentrations: cross sign, 0.1 mmol L-1; pentagram, 0.025 mmol L-1, diamond, 0.017 mmol L-1; asterisk, 0.01 mmol L-1; and dot, non-spiked samples. Cluster with dashed line is the skim milk samples with melamine concentration lower than 0.01 mmol L-1, while cluster with the solid line contains the samples with melamine concentration higher than 0.01 mmol L-1.  69   Figure 5.9 Loading plot of the representative principal component analysis model. Principal component (PC) 1 represents 45.69% variances and PC2 represents 25.11% variances.  Although Liu and others (71) reported the success of applying MIPs-AgNPs to quantify theophylline in tea drinks, the current MIPs-AgNPs for melamine cannot be used for quantitative analysis for melamine in skim milk.11 Skim milk samples with melamine concentrations higher than 0.01 mmol L-1 were overlapped in PCA model (Figure 5.8). The reason could be attributed to simple components in tea drinks that contain almost no fat or protein, resulting in little interference appeared in samples. In addition, the intensity of band at 2935 cm-1, derived by MIPs, was used as internal standard to normalize featured theophylline band at 567 cm-1 (71).                                                  11 This paragraph was not presented in the published paper (Rapid detection of melamine in tap water and milk using conjugated “one-step” molecularly imprinted polymers-surface enhanced Raman spectroscopic biosensor) and was added for this thesis. 70  However, due to the limitation of the Raman system in our laboratory, the intensity of the band at 2935 cm-1 was too weak to be used as internal standard. Therefore, with a more advanced Raman system, quantitative analysis of melamine in skim milk might be achieved by using the current MIPs-AgNPs.   SPE was involved in sample pretreatment because of the inferences by the complex food matrix, leading to increase in time required for the overall analysis (i.e. 25 min) compared to the “two-step” antibody-SERS (i.e. 20 min) and MIPs-SERS (i.e. 18 min) biosensors. However, by decorating the specific cavities in MIPs with SERS-active substrate (i.e. AgNPs) the reproducibility of SERS spectra significantly increased. In the “two-step” biosensors, eluents with melamine were deposited onto Ag dendrite nanostructure for SERS measurement. The location of melamine on SERS-active substrate is random, and not all of them are located in the EM field. Thus, spots of the enhanced melamine signals require extensive time to identify. However, with the integrated MIPs-AgNPs, melamine was adsorbed in the cavities specifically for it. Meanwhile, adjacent AgNPs provide EM field to enhance the Raman signals. By modifying the synthesis of this conjugated MIPs-AgNPs, or by integrating more effective SERS-active substrates, the sample pretreatment procedures could be further simplified, making this technique more appealing for the food industry and government laboratories.   Conclusion 5.5In conclusion, this “one-step” MIPs-SERS conjugating MIPs and AgNPs during the polymerization cannot be employed for the determination of melamine in whole milk, but showed feasibility to be applied for the specific separation and detection of trace level melamine 71  in skim milk samples, with the LOD between 0.01 to 0.017 mmol L-1. Although the overall analysis required longer time (i.e. 25 min) compared to the previous two biosensors (i.e. 20 min and 18 min), the detection was still rapid and the reproducibility of SERS spectra was increased by controlling the distances between melamine and SERS-active substrate, resulting in more efficient and accurate detection. Thus, this “one-step” MIPs-SERS biosensor meets the demands of high-throughput and trace level detection of melamine required by the food industry and government laboratories. 72  Chapter 6: Conclusion   Main findings 6.1After the widely publicized melamine adulteration in infant formula and pet food, many attempts have been made to investigate accurate, rapid and high-throughput techniques to determine melamine in food matrices. LC and GC hyphenated detection approaches are adopted by most of the government laboratories based on their high reproducibility and sensitivity (6-8). However, these techniques require complex sample pretreatment, sophisticated instruments, and highly trained personnel, failing to meet rapid, high-throughput, and on-line or in-field detection demanded by the food industry. To achieve this requirement, studies have been performed applying antibodies or MIPs to extract melamine from food matrices more efficiently and effectively (56, 58, 59, 73, 80, 111). Although antibody and MIPs, both with the specific affinity to melamine, shorten the sample pretreatment and analyte extraction time, the detection techniques still involve complex instrumentation and labor-intensive procedures, which interfere with the application in high-throughput analysis. As a non- or less-destructive detection technique, Raman spectroscopy was applied to sense melamine in food matrices by collecting Raman scattering signals associated with ring-breathing mode I and mode II involving the triazine ring structure (13, 14). However, traditional Raman spectroscopy lacks the sensitivity to trace level chemicals because of the faint inelastic Raman scattering. With the assistance of the EM field on the surface of noble metallic nanostructures (e.g. silver or gold nanoparticles or colloids), surface enhanced Raman spectroscopy can enhance Raman scattering signals tremendously and results in even single molecule detection (15). These inherent properties of Raman spectroscopy and SERS make it possible to apply SERS for the determination of trace-73  level melamine in food matrices. Based on the principle of SERS, both analyte compounds and interferents contribute to spectral features. Therefore, moderate sample pretreatment is still necessary before applying SERS. However, until now no attempt to combine antibody or MIPs with SERS for accurate, rapid and high-throughput determination of melamine in whole milk has been described.   Accordingly, three objectives were developed to investigate three types of SERS-based biosensors coupling with antibody or MIPs to determine melamine in whole milk, and Chapter 3, Chapter 4 and Chapter 5 each described the experimental design and results corresponding to one of the objectives. Three hypotheses were ascertained for each objective, and the results are summarized in Table 6.1.   Table 6.1 Summary of the decision for hypotheses developed in this study.  Chapter 3 Objective 1 Chapter 4 Objective 2 Chapter 5 Objective 3 Hypothesis 1 Accepted Accepted Accepted Hypothesis 2 Accepted Accepted Rejected Hypothesis 3 Accepted Accepted Rejected Objective 1: To develop a “two-step” antibody-SERS biosensor for rapid detection of melamine in whole milk by using antibody for immunological separation and Ag dendrite (i.e. SERS-active substrate) for detection. Objective 2: To develop a “two-step” MIPs-SERS biosensor for rapid detection of melamine in whole milk by using MIPs for SPE and Ag dendrite (i.e. SERS-active substrate) for detection. 74  Objective 3: To develop a “one-step” MIPs-SERS biosensor for rapid detection of melamine in whole milk by using AgNPs incorporated MIPs for simultaneous extraction and detection. Hypothesis 1: Antibody produced by using melamine hapten and MIPs synthesized by bulk polymerization have the specific affinity to melamine. Hypothesis 2: Antibody and MIPs can extract and concentrate melamine from dairy products. Hypothesis 3: Ag dendrite and AgNPs incorporated in MIPs can serve as the SERS-active substrates to sense trace level of melamine in dairy products.  For the first objective, Figure 3.3 and 3.4 demonstrated the result that hypothesis 1 and 2 can be accepted, while Figure 3.6 resulted in a LOD of less than < 0.79×10-3 mmol L-1, which supported hypothesis 3. Therefore, the first objective, to develop a “two-step” antibody-SERS biosensor for melamine in whole milk, has been achieved successfully.  The three hypotheses for the second objective were all supported by the results shown in Figure 4.2 and 4.3, Table 4.1, and Figure 4.4, 4.6 and 4.8, respectively. Based on the regression model (Figure 4.8), the LOD was 0.012 mmol L-1 and LOQ was 0.039 mmol L-1, representing the accomplishment of the second objective.  Nevertheless, not all the hypotheses for the last objective were accepted. The first hypothesis was accepted based on the results in Figure 5.4 and 5.6. However, “one-step” MIPs-AgNPs failed to extract melamine from whole milk or to generate enhanced melamine Raman scattering signals after binding melamine in milk. Therefore, the original hypothesis 2 and 3 were rejected. Two alternatives that used substituted food matrix (i.e. skim milk) were developed and were accepted 75  based upon the results in Figure 5.7 and 5.8. The LOD determined based upon PCA model was 0.01~0.017 mmol L-1. By changing the original objective 3 to “to develop a ‘one-step’ MIPs-SERS biosensor for rapid detection of melamine in skim milk by using AgNPs incorporated MIPs for simultaneous extraction and detection”, the last objective was also achieved.  In summary, antibody for melamine produced using 3-(4,6-diamino-1,3,5-triazin-2-methylthio) propanoic acid as melamine hapten and MIPs synthesized by bulk polymerization was able to give specific affinity towards melamine and extract melamine exclusively from whole milk samples. By applying Ag dendrite nanostructure as SERS-active substrate, the LOD of “two-step” antibody-SERS biosensor and MIPs-SERS biosensor were less than < 0.79×10-3 mmol L-1 and 0.012 mmol L-1 melamine in whole milk, respectively, while LOQ of “two-step” MIPs-SERS was 0.039 mmol L-1. MIPs-AgNPs synthesized by addition of silver nitrate during bulk polymerization for the “one-step” MIPs-SERS biosensor was validated to have specific affinity towards melamine. Although the “one-step” MIPs-SERS biosensor failed to extract melamine from whole milk or to generate SERS signals for melamine in whole milk, it could extract melamine exclusively from skim milk and resulted in a LOD of 0.01~0.017 mmol L-1 melamine in skim milk.   Future research directions 6.2In the current study, “two-step” antibody-SERS biosensor and “one-step” MIPs-SERS biosensor were not applicable for quantitative analysis of melamine in milk. Further investigation should be conducted to optimize the experimental design, such as lowering melamine concentrations in milk to avoid saturation of specific binding sites on antibody or MIPs and “hot spot” on SERS-76  active substrate. Besides, due to the limitation of the Raman system used in this study, no internal standard SERS band was used to further remove the “one-step” MIPs-SERS spectral variations, resulting in the loss of spectral reproducibility. Therefore, with a more advanced Raman system, quantitative analysis of melamine in skim milk might be achieved by using the current “one-step” MIPs-AgNPs.  It is of great significance to optimize sample pretreatment procedures to achieve detection of melamine in whole milk by this current “one-step” MIPs-SERS biosensor, because whole milk is more widely consumed than skim milk. Although silica based SPE cartridge was used to remove whole milk food matrices, milk compounds remained in the pretreated samples due to the non-specific affinity of the sorbent towards melamine. Therefore, with more suitable sample pretreatment procedures, this “one-step” MIPs-SERS biosensor might be applied to sense trace level of melamine in whole milk. Moreover, different forms of MIPs and SERS-active substrates can be conjugated in various ways (93, 94, 115, 117), which have been elaborated in Section 1.3.3. Consequently, there is an opportunity to explore different combinations of MIPs and SERS-active substrates to obtain optimized distances between specific binding cavities and SERS-active substrates. Thus, improved binding capacity and specificity might be achieved, and the interference from food matrices could be reduced, resulting in the potential to be used to determine melamine in whole milk.  Based upon the results that all the three types of biosensors have low LOD and can finish the whole analysis within 25 min, these biosensors have the potential to be used for accurate, rapid and high-throughput detection required by the food industry and government laboratories. To be 77  applied for on-site or in-field analysis, a more affordable portable or handheld Raman spectrometer can be employed as the alternative to bench-top Raman system used in this study.   After acquiring antibodies or MIPs for different compounds, methodologies proposed in this study of coupling antibody or MIPs and SERS could be generalized to determine other food chemical hazards, and even biological hazards. These studies are of significant importance to the food industry and government laboratories to develop rapid and accurate detection techniques to assure conformance of food safety.   Conclusion 6.3By employing antibody or MIPs for separation and SERS for detection, three types of innovative SERS-based biosensors were developed successfully for rapid detection of melamine in milk, and the results are summarized and compared in Table 6.2. However, not all the original objectives were achieved in this study. The first two types of biosensors, “two-step” antibody-SERS and MIPs-SERS, could be used to detect trace level of melamine in whole milk, which met the original objectives. However, the “one-step” MIPs-SERS biosensor failed to meet the original objective, but achieved the detection of trace level of melamine in skim milk. Among all the three types of biosensors, “two-step” antibody-SERS had the highest sensitivity with the lowest LOD of melamine in whole milk. “Two-step” MIPs-SERS biosensor had the shortest analysis time and was also suitable for quantitative analysis of melamine in whole milk. Although “one-step” MIPs-SERS biosensor required longer time for analysis and was only suitable for the detection but not quantification of melamine in skim milk, this biosensor proposed a potentially good solution to overcome the problem of poor reproducibility of SERS-78  active substrate by fixing the distances between cavities for melamine and AgNPs. Compared to the protein-based antibody, polymer-based MIPs showed much higher stability at extreme conditions, including at high temperature and low pH environment, which are always the case for on-site and in-field analysis. Moreover, the production of anti-melamine was complicated and time consuming compared to the synthesis of MIPs.   Therefore, MIPs-based biosensors have a good potential to be applied for rapid and high-throughput detection of melamine in dairy products.   Table 6.2 Summary of the results of three types of innovative SERS biosensors.  “two-step” antibody-SERS “two-step” MIPs-SERS “one-step” MIPs-SERS Production/synthesis complicated simple simple Stability poor good good Food matrices tested whole milk whole milk skim milk Analysis time (min) 20 18 25 LOD (mmol L-1) 0.79×10-3 0.012 0.01~0.017 LOQ (mmol L-1) not applicable 0.039 not applicable Spectra reproducibility moderately good moderately good good   79  References  1. Jung, S.; Rickert, D. A.; Deak, N. A.; Aldin, E. D.; Recknor, J.; Johnson, L. A.; Murphy, P. A., Comparison of kjeldahl and dumas methods for determining protein contents of soybean products. Journal of American Oil Chemists' Society 2003, 80, 1169-1173. 2. Li, X.; Feng, S.; Hu, Y.; Sheng, W.; Zhang, Y.; Yuan, S.; Zeng, H.; Wang, S.; Lu, X., Rapid detection of melamine in milk using immunological separation and surface enhanced Raman spectroscopy. Journal of Food Science 2015, 80, C1196-C1201. 3. Brown, C. A.; Jeong, K.-S.; Poppenga, R. H.; Puschner, B.; Miller, D. M.; Ellis, A. E.; Kang, K.-I.; Sum, S.; Cistola, A. M.; Brown, S. A., Outbreaks of renal failure associated with melamine and cyanuric acid in dogs and cats in 2004 and 2007. Journal of Veterinary Diagnostic Investigation 2007, 19, 525-531. 4. Wu, Y.; Zhao, Y.; Li, J., A survey on occurrence of melamine and its analogues in tainted infant formula in China. Biomedical and Environmental Sciences 2009, 22, 95-99. 5. Ingelfinger, J. R., Melamine and the global implications of food contamination. New England Journal of Medicine 2008, 359, 2745-2748. 6. Lin, M.; He, L.; Awika, J.; Yang, L.; Ledoux, D. R.; Li, H.; Mustapha, A., Detection of melamine in gluten, chicken feed, and processed foods using surface enhanced Raman spectroscopy and HPLC. Journal of Food Science 2008, 73, T129-T134. 7. Sun, H.; Wang, L.; Ai, L.; Liang, S.; Wu, H., A sensitive and validated method for determination of melamine residue in liquid milk by reversed phase high-performance liquid chromatography with solid-phase extraction. Food Control 2010, 21, 686-691. 80  8. Chao, Y.-Y.; Lee, C.-T.; Wei, Y.-T.; Kou, H.-S.; Huang, Y.-L., Using an on-line microdialysis/HPLC system for the simultaneous determination of melamine and cyanuric acid in non-dairy creamer. Analytica Chimica Acta 2011, 702, 56-61. 9. Li, J.; Qi, H.; Shi, Y., Determination of melamine residues in milk products by zirconia hollow fiber sorptive microextraction and gas chromatography–mass spectrometry. Journal of Chromatography A 2009, 1216, 5467-5471. 10. Choo-Smith, L. P.; Edwards, H. G. M.; Endtz, H. P.; Kros, J. M.; Heule, F.; Barr, H.; Robinson, J. S.; Bruining, H. A.; Puppels, G. J., Medical applications of Raman spectroscopy: From proof of principle to clinical implementation. Biopolymers 2002, 67, 1-9. 11. Li-Chan, E. C. Y., The applications of Raman spectroscopy in food science. Trends in Food Science and Technology 1996, 7, 361-370. 12. Williams, A. C.; Edwards, H. G. M., Fourier transform Raman spectroscopy of bacterial cell walls. Journal of Raman Spectroscopy 1994, 25, 673-677. 13. Okazaki, S.; Hiramatsu, M.; Gonmori, K.; Suzuki, O.; Tu, A., Rapid nondestructive screening for melamine in dried milk by Raman spectroscopy. Forensic Toxicology 2009, 27, 94-97. 14. Cheng, Y.; Dong, Y.; Wu, J.; Yang, X.; Bai, H.; Zheng, H.; Ren, D.; Zou, Y.; Li, M., Screening melamine adulterant in milk powder with laser Raman spectrometry. Journal of Food Composition and Analysis 2010, 23, 199-202. 15. Nie, S.; Emory, S. R., Probing single molecules and single nanoparticles by surface-enhanced Raman scattering. Science 1997, 275, 1102-1106. 16. Haynes, C. L.; McFarland, A. D.; Duyne, R. P. V., Surface-enhanced Raman spectroscopy. Analytical Chemistry 2005, 77, 338 A-346 A. 81  17. Lu, X.; Huang, Q.; Miller, W. G.; Aston, D. E.; Xu, J.; Xue, F.; Zhang, H.; Rasco, B. A.; Wang, S.; Konkel, M. E., Comprehensive detection and discrimination of Campylobacter species by use of confocal micro-Raman spectroscopy and multilocus sequence typing. Journal of Clinical Microbiology 2012, 50, 2932-2946. 18. Costa, J. C. S.; Ando, R. m. A.; Camargo, P. H. C.; Corio, P., Understanding the effect of adsorption geometry over substrate selectivity in the surface-enhanced Raman scattering spectra of simazine and atrazine. The Journal of Physical Chemistry C 2011, 115, 4184-4190. 19. Cheung, W.; Shadi, I. T.; Xu, Y.; Goodacre, R., Quantitative analysis of the banned food dye Sudan-1 using surface enhanced Raman scattering with multivariate chemometrics. The Journal of Physical Chemistry C 2010, 114, 7285-7290. 20. Giovannozzi, A. M.; Rolle, F.; Sega, M.; Abete, M. C.; Marchis, D.; Rossi, A. M., Rapid and sensitive detection of melamine in milk with gold nanoparticles by surface enhanced Raman scattering. Food Chemistry 2014, 159, 250-256. 21. Yazgan, N.; Boyacı, İ.; Topcu, A.; Tamer, U., Detection of melamine in milk by surface-enhanced Raman spectroscopy coupled with magnetic and Raman-labeled nanoparticles. Analytical and Bioanalytical Chemistry 2012, 403, 2009-2017. 22. Liu, B.; Lin, M.; Li, H., Potential of SERS for rapid detection of melamine and cyanuric acid extracted from milk. Sensing and Instrumentation for Food Quality and Safety. 2010, 4, 13-19. 23. He, L.; Liu, Y.; Lin, M.; Awika, J.; Ledoux, D.; Li, H.; Mustapha, A., A new approach to measure melamine, cyanuric acid, and melamine cyanurate using surface enhanced Raman spectroscopy coupled with gold nanosubstrates. Sensing and Instrumentation for Food Quality and Safety. 2008, 2, 66-71. 82  24. Kim, A.; Barcelo, S. J.; Williams, R. S.; Li, Z., Melamine sensing in milk products by using surface enhanced Raman scattering. Analytical Chemistry 2012, 84, 9303-9309. 25. Hanly, W. C.; Artwohl, J. E.; Bennett, B. T., Review of polyclonal antibody production procedures in mammals and poultry. Institute for Laboratory Animal Research Journal 1995, 37, 93-118. 26. Zhang, H.; Wang, S., Review on enzyme-linked immunosorbent assays for sulfonamide residues in edible animal products. Journal of Immunological Methods 2009, 350, 1-13. 27. Wang, Z.; Mi, T.; Beier, R. C.; Zhang, H.; Sheng, Y.; Shi, W.; Zhang, S.; Shen, J., Hapten synthesis, monoclonal antibody production and development of a competitive indirect enzyme-linked immunosorbent assay for erythromycin in milk. Food Chemistry 2015, 171, 98-107. 28. Esteve-Turrillas, F. A.; Abad-Somovilla, A.; Quiñones-Reyes, G.; Agulló, C.; Mercader, J. V.; Abad-Fuentes, A., Monoclonal antibody-based immunoassays for cyprodinil residue analysis in QuEChERS-based fruit extracts. Food Chemistry 2015, 187, 530-536. 29. Hu, J.; Yokoyama, T.; Kitagawa, T., Studies on the optimal immunization schedule of experimental animals. V. The effects of the route of injection, the content of Mycobacteria in Freund's adjuvant and the emulsifying antigen. Chemical and Pharmaceutical Bulletin 1990, 38, 1961-1965. 30. Hu, J.; Ide, A.; Yokoyama, T.; Kitagawa, T., Studies on the optimal immunization schedule of the mouse as an experimental animal. The effect of antigen dose and adjuvant type. Chemical and Pharmaceutical Bulletin 1989, 37, 3042-3046. 83  31. Maderna, A.; Leverett, C. A., Recent advances in the development of new auristatins: Structural modifications and application in antibody drug conjugates. Molecular Pharmaceutics 2015. 32. Robertson, G.; Hirst, M.; Bainbridge, M.; Bilenky, M.; Zhao, Y.; Zeng, T.; Euskirchen, G.; Bernier, B.; Varhol, R.; Delaney, A., Genome-wide profiles of STAT1 DNA association using chromatin immunoprecipitation and massively parallel sequencing. Nature Methods 2007, 4, 651-657. 33. Barr, T. P.; Kornberg, D.; Montmayeur, J.-P.; Long, M.; Reichheld, S.; Strichartz, G. R., Validation of endothelin B receptor antibodies reveals two distinct receptor-related bands on Western blot. Analytical Biochemistry 2015, 468, 28-33. 34. Yu, H.; Bruno, J. G., Immunomagnetic-electrochemiluminescent detection of Escherichia coli O157 and Salmonella typhimurium in foods and environmental water samples. Applied and Environmental Microbiology 1996, 62, 587-592. 35. Skjerve, E.; Rørvik, L.; Olsvik, O., Detection of Listeria monocytogenes in foods by immunomagnetic separation. Applied and Environmental Microbiology 1990, 56, 3478-3481. 36. Su, M.; Venkatachalam, M.; Gradziel, T. M.; Liu, C.; Zhang, Y.; Roux, K. H.; Sathe, S. K., Application of mouse monoclonal antibody (mAb) 4C10-based enzyme-linked immunosorbent assay (ELISA) for amandin detection in almond (Prunus dulcis L.) genotypes and hybrids. LWT - Food Science and Technology 2015, 60, 535-543. 37. Suri, C. R.; Raje, M.; Varshney, G. C., Immunosensors for pesticide analysis: Antibody production and sensor development. Critical Reviews in Biotechnology 2002, 22, 15-32. 38. Sarımehmetoglu, B.; Kuplulu, O.; Celik, T. H., Detection of aflatoxin M 1 in cheese samples by ELISA. Food Control 2004, 15, 45-49. 84  39. Thirumala-Devi, K.; Mayo, M. A.; Gopal, R.; Reddy, S. V.; Delfosse, P.; Reddy, D. V. R., Production of polyclonal antibodies against ochratoxin A and its detection in chilies by ELISA. Journal of Agricultural and Food Chemistry 2000, 48, 5079-5082. 40. Bier, F. F.; Stöcklein, W.; Böcher, M.; Bilitewski, U.; Schmid, R. D., Use of a fibre optic immunosensor for the detection of pesticides. Sensors and Actuators B: Chemical 1992, 7, 509-512. 41. Geng, T.; Morgan, M. T.; Bhunia, A. K., Detection of low levels of Listeria monocytogenes cells by using a fiber-optic immunosensor. Applied and Environmental Microbiology 2004, 70, 6138-6146. 42. Oh, B.-K.; Kim, Y.-K.; Park, K. W.; Lee, W. H.; Choi, J.-W., Surface plasmon resonance immunosensor for the detection of Salmonella typhimurium. Biosensors and Bioelectronics 2004, 19, 1497-1504. 43. Kim, S. J.; Gobi, K. V.; Iwasaka, H.; Tanaka, H.; Miura, N., Novel miniature SPR immunosensor equipped with all-in-one multi-microchannel sensor chip for detecting low-molecular-weight analytes. Biosensors and Bioelectronics 2007, 23, 701-707. 44. Micheli, L.; Grecco, R.; Badea, M.; Moscone, D.; Palleschi, G., An electrochemical immunosensor for aflatoxin M1 determination in milk using screen-printed electrodes. Biosensors and Bioelectronics 2005, 21, 588-596. 45. Piermarini, S.; Micheli, L.; Ammida, N. H. S.; Palleschi, G.; Moscone, D., Electrochemical immunosensor array using a 96-well screen-printed microplate for aflatoxin B1 detection. Biosensors and Bioelectronics 2007, 22, 1434-1440. 46. Karaseva, N.; Ermolaeva, T., A piezoelectric immunosensor for chloramphenicol detection in food. Talanta 2012, 93, 44-48. 85  47. Su, X.; Li, Y., A self-assembled monolayer-based piezoelectric immunosensor for rapid detection of Escherichia coli O157: H7. Biosensors and Bioelectronics 2004, 19, 563-574. 48. He, L.; Rodda, T.; Haynes, C. L.; Deschaines, T.; Strother, T.; Diez-Gonzalez, F.; Labuza, T. P., Detection of a foreign protein in milk using surface-enhanced Raman spectroscopy coupled with antibody-modified silver dendrites. Analytical Chemistry 2011, 83, 1510-1513. 49. He, L.; Deen, B.; Rodda, T.; Ronningen, I.; Blasius, T.; Haynes, C.; Diez-Gonzalez, F.; Labuza, T. P., Rapid detection of ricin in milk using immunomagnetic separation combined with surface-enhanced Raman spectroscopy. Journal of Food Science 2011, 76, N49-N53. 50. Zhu, G.; Hu, Y.; Gao, J.; Zhong, L., Highly sensitive detection of clenbuterol using competitive surface-enhanced Raman scattering immunoassay. Analytica Chimica Acta 2011, 697, 61-66. 51. Zhou, Y.; Li, C.; Li, Y.; Ren, H.; Lu, S.; Tian, X.; Hao, Y.; Zhang, Y.; Shen, Q.; Liu, Z.; Meng, X.; Zhang, J., Monoclonal antibody based inhibition ELISA as a new tool for the analysis of melamine in milk and pet food samples. Food Chemistry 2012, 135, 2681-2686. 52. Garber, E. A. E., Detection of melamine using commercial enzyme-linked immunosorbent assay technology. Journal of Food Protection 2008, 71, 590-594. 53. Lu, Y.; Xia, Y.; Pan, M.; Wang, X.; Wang, S., Development of surface plasmon resonance immunosensor for detecting melamine in milk products and pet foods. Journal of Agricultural and Food Chemistry 2014. 54. Wu, J.; Xu, F.; Zhu, K.; Wang, Z.; Wang, Y.; Zhao, K.; Li, X.; Jiang, H.; Ding, S., Rapid and sensitive fluoroimmunoassay based on quantum dots for detection of melamine in milk. Analytical Letters 2012, 46, 275-285. 86  55. Le, T.; Yan, P.; Xu, J.; Hao, Y., A novel colloidal gold-based lateral flow immunoassay for rapid simultaneous detection of cyromazine and melamine in foods of animal origin. Food Chemistry 2013, 138, 1610-1615. 56. Kim, B.; Perkins, L. B.; Bushway, R. J.; Nesbit, S.; Fan, T.; Sheridan, R.; Greene, V., Determination of melamine in pet food by enzyme immunoassay, high-performance liquid chromatography with diode array detection, and ultra-performance liquid chromatography with tandem mass spectrometry. Journal of AOAC International 2008, 91, 408-413. 57. Lei, H.; Shen, Y.; Song, L.; Yang, J.; Chevallier, O. P.; Haughey, S. A.; Wang, H.; Sun, Y.; Elliott, C. T., Hapten synthesis and antibody production for the development of a melamine immunoassay. Analytica Chimica Acta 2010, 665, 84-90. 58. Li, X.; Luo, P.; Tang, S.; Beier, R. C.; Wu, X.; Yang, L.; Li, Y.; Xiao, X., Development of an immunochromatographic strip test for rapid detection of melamine in raw milk, milk products and animal feed. Journal of Agricultural and Food Chemistry 2011, 59, 6064-6070. 59. Li, W.; Meng, M.; Lu, X.; Liu, W.; Yin, W.; Liu, J.; Xi, R., Preparation of anti-melamine antibody and development of an indirect chemiluminescent competitive ELISA for melamine detection in milk. Food and Agricultural Immunology 2014, 25, 498-509. 60. Ye, L.; Haupt, K., Molecularly imprinted polymers as antibody and receptor mimics for assays, sensors and drug discovery. Analitical and Bioanalitical Chemistry 2004, 378, 1887-1897. 61. Haupt, K., Molecularly imprinted polymers: the next generation. Analytical Chemistry 2003, 75, 376 A-383 A. 62. Spivak, D. A., Optimization, evaluation, and characterization of molecularly imprinted polymers. Advanced Drug Delivery Reviews 2005, 57, 1779-1794. 87  63. Vasapollo, G.; Sole, R. D.; Mergola, L.; Lazzoi, M. R.; Scardino, A.; Scorrano, S.; Mele, G., Molecularly imprinted polymers: Present and future prospective. International Journal of Molecular Sciences 2011, 12, 5908-5945. 64. Lok, C.; Son, R., Application of molecularly imprinted polymers in food sample analysis—a perspective. International Food Research Journal 2009, 16, 127-140. 65. Yang, H.; Zhang, S.; Yang, W.; Chen, X.; Zhuang, Z.; Xu, J.; Wang, X., Molecularly imprinted sol−gel nanotubes membrane for biochemical separations. Journal of the American Chemical Society 2004, 126, 4054-4055. 66. Wackerlig, J.; Lieberzeit, P. A., Molecularly imprinted polymer nanoparticles in chemical sensing – Synthesis, characterisation and application. Sensors and Actuators B: Chemical 2015, 207, Part A, 144-157. 67. Bai, H.; Wang, C.; Chen, J.; Peng, J.; Cao, Q., A novel sensitive electrochemical sensor based on in-situ polymerized molecularly imprinted membranes at graphene modified electrode for artemisinin determination. Biosensors and Bioelectronics 2015, 64, 352-358. 68. Mujahid, A.; Lieberzeit, P. A.; Dickert, F. L., Chemical sensors based on molecularly imprinted sol-gel materials. Materials 2010, 3, 2196-2217. 69. Huang, X.; Zou, H.; Chen, X.; Luo, Q.; Kong, L., Molecularly imprinted monolithic stationary phases for liquid chromatographic separation of enantiomers and diastereomers. Journal of Chromatography A 2003, 984, 273-282. 70. Alvarez-Lorenzo, C.; Concheiro, A., Molecularly imprinted polymers for drug delivery. Journal of Chromatography B 2004, 804, 231-245. 88  71. Liu, P.; Liu, R.; Guan, G.; Jiang, C.; Wang, S.; Zhang, Z., Surface-enhanced Raman scattering sensor for theophylline determination by molecular imprinting on silver nanoparticles. Analyst 2011, 136, 4152-4158. 72. He, C.; Long, Y.; Pan, J.; Li, K.; Liu, F., Application of molecularly imprinted polymers to solid-phase extraction of analytes from real samples. Journal of Biochemical and Biophysical Methods 2007, 70, 133-150. 73. Yang, H.; Zhou, W.; Guo, X.; Chen, F.; Zhao, H.; Lin, L.; Wang, X., Molecularly imprinted polymer as SPE sorbent for selective extraction of melamine in dairy products. Talanta 2009, 80, 821-825. 74. Feng, S.; Gao, F.; Chen, Z.; Grant, E.; Kitts, D. D.; Wang, S.; Lu, X., Determination of α-tocopherol in vegetable oils using a molecularly imprinted polymers-surface-enhanced Raman spectroscopic biosensor. Journal of Agricultural and Food Chemistry 2013, 61, 10467-10475. 75. Zhang, C.; Cui, H.; Cai, J.; Duan, Y.; Liu, Y., Development of fluorescence sensing material based on CdSe/ZnS quantum dots and molecularly imprinted polymer for the detection of carbaryl in rice and Chinese cabbage. Journal of Agricultural and Food Chemistry 2015. 76. Liu, H.; Fang, G.; Wang, S., Molecularly imprinted optosensing material based on hydrophobic CdSe quantum dots via a reverse microemulsion for specific recognition of ractopamine. Biosensors and Bioelectronics 2014, 55, 127-132. 77. Yang, J. C.; Shin, H.-K.; Hong, S. W.; Park, J. Y., Lithographically patterned molecularly imprinted polymer for gravimetric detection of trace atrazine. Sensors and Actuators B: Chemical 2015, 216, 476-481. 89  78. Ebarvia, B. S.; Sevilla Iii, F., Piezoelectric quartz sensor for caffeine based on molecularly imprinted polymethacrylic acid. Sensors and Actuators B: Chemical 2005, 107, 782-790. 79. Bates, F.; del Valle, M., Voltammetric sensor for theophylline using sol-gel immobilized molecularly imprinted polymer particles. Microchimica Acta 2015, 182, 933-942. 80. He, L.; Su, Y.; Shen, X.; Zheng, Y.; Guo, H.; Zeng, Z., Solid-phase extraction of melamine from aqueous samples using water-compatible molecularly imprinted polymers. Journal of Separation Science 2009, 32, 3310-3318. 81. Liu, J.; Song, H.; Liu, J.; Liu, Y.; Li, L.; Tang, H.; Li, Y., Preparation of molecularly imprinted polymer with double templates for rapid simultaneous determination of melamine and dicyandiamide in dairy products. Talanta 2015, 134, 761-767. 82. Fan, W.; Gao, M.; He, M.; Chen, B.; Hu, B., Cyromazine imprinted polymers for selective stir bar sorptive extraction of melamine in animal feed and milk samples. Analyst 2015, 140, 4057-4067. 83. Zhang, Z.; Cheng, Z.; Zhang, C.; Wang, H.; Li, J., Precipitation polymerization of molecularly imprinted polymers for recognition of melamine molecule. Journal of Applied Polymer Science 2012, 123, 962-967. 84. Figueiredo, L.; Santos, L.; Alves, A., Synthesis of a molecularly imprinted polymer for melamine analysis in milk by HPLC with diode array detection. Advances in Polymer Technology 2015, in press. DOI: 10.1002/adv.21506. 85. Yu, J.; Zhang, C.; Dai, P.; Ge, S., Highly selective molecular recognition and high throughput detection of melamine based on molecularly imprinted sol-gel film. Analytica Chimica Acta 2009, 651, 209-214. 90  86. Liu, Y. T.; Deng, J.; Xiao, X. L.; Ding, L.; Yuan, Y. L.; Li, H.; Li, X. T.; Yan, X. N.; Wang, L. L., Electrochemical sensor based on a poly(para-aminobenzoic acid) film modified glassy carbon electrode for the determination of melamine in milk. Electrochimica Acta 2011, 56, 4595-4602. 87. Guven, B.; Basaran-Akgul, N.; Temur, E.; Tamer, U.; Boyac, I. H., SERS-based sandwich immunoassay using antibody coated magnetic nanoparticles for Escherichia coli enumeration. Analyst 2011, 136, 740-748. 88. Gao, F.; Feng, S.; Chen, Z.; Li-Chan, E. C.; Grant, E.; Lu, X., Detection and quantification of chloramphenicol in milk and honey using molecularly imprinted polymers: Canadian penny-based SERS nano-biosensor. Journal of Food Science 2014, 79, N2542-N2549. 89. Gao, F.; Hu, Y.; Chen, D.; Li-Chan, E. C.; Grant, E.; Lu, X., Determination of Sudan I in paprika powder by molecularly imprinted polymers-thin layer chromatography-surface enhanced Raman spectroscopic biosensor. Talanta 2015, 143, 344-352. 90. He, L.; Haynes, C. L.; Diez-Gonzalez, F.; Labuza, T. P., Rapid detection of a foreign protein in milk using IMS-SERS. Journal of Raman Spectroscopy 2011, 42, 1428-1434. 91. Sanles-Sobrido, M.; Rodríguez-Lorenzo, L.; Lorenzo-Abalde, S.; González-Fernández, Á.; Correa-Duarte, M. A.; Alvarez-Puebla, R. A.; Liz-Marzán, L. M., Label-free SERS detection of relevant bioanalytes on silver-coated carbon nanotubes: the case of cocaine. Nanoscale 2009, 1, 153-158. 92. Baniukevic, J.; Hakki Boyaci, I.; Goktug Bozkurt, A.; Tamer, U.; Ramanavicius, A.; Ramanaviciene, A., Magnetic gold nanoparticles in SERS-based sandwich immunoassay for antigen detection by well oriented antibodies. Biosensors and Bioelectronics 2013, 43, 281-288. 91  93. Chen, S.; Li, X.; Guo, Y.; Qi, J., A Ag-molecularly imprinted polymer composite for efficient surface-enhanced Raman scattering activities under a low-energy laser. Analyst 2015, 140, 3239-3243. 94. Holthoff, E. L.; Stratis-Cullum, D. N.; Hankus, M. E., A nanosensor for TNT detection based on molecularly imprinted polymers and surface enhanced Raman scattering. Sensors 2011, 11, 2700-2714. 95. Xue, J.; Li, D.; Qu, L.; Long, Y., Surface-imprinted core–shell Au nanoparticles for selective detection of bisphenol A based on surface-enhanced Raman scattering. Analytica Chimica Acta 2013, 777, 57-62. 96. Lu, X.; Al-Qadiri, H.; Lin, M.; Rasco, B., Application of mid-infrared and Raman spectroscopy to the study of bacteria. Food and Bioprocess Technology 2011, 4, 919-935. 97. Mecker, L. C.; Tyner, K. M.; Kauffman, J. F.; Arzhantsev, S.; Mans, D. J.; Gryniewicz-Ruzicka, C. M., Selective melamine detection in multiple sample matrices with a portable Raman instrument using surface enhanced Raman spectroscopy-active gold nanoparticles. Analytica Chimica Acta 2012, 733, 48-55. 98. Zhang, X.; Zou, M.; Qi, X.; Liu, F.; Zhu, X.; Zhao, B., Detection of melamine in liquid milk using surface‐enhanced Raman scattering spectroscopy. Journal of Raman Spectroscopy 2010, 41, 1655-1660. 99. Hu, Y.; Feng, S.; Gao, F.; Li-Chan, E. C.; Grant, E.; Lu, X., Detection of melamine in milk using molecularly imprinted polymers-surface enhanced Raman spectroscopy. Food Chemistry 2015, 176, 123-129. 100. Asensio, L.; González, I.; García, T.; Martin, R., Determination of food authenticity by enzyme-linked immunosorbent assay (ELISA). Food Control 2008, 19, 1-8. 92  101. Kumar, S.; Rai, A. K.; Singh, V.; Rai, S., Vibrational spectrum of glycine molecule. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2005, 61, 2741-2746. 102. Badawy, A. M. E.; Luxton, T. P.; Silva, R. G.; Scheckel, K. G.; Suidan, M. T.; Tolaymat, T. M., Impact of environmental conditions (pH, ionic strength, and electrolyte type) on the surface charge and aggregation of silver nanoparticles suspensions. Environmental Science and Technology 2010, 44, 1260-1266. 103. Tyan, Y.-C.; Yang, M.-H.; Jong, S.-B.; Wang, C.-K.; Shiea, J., Melamine contamination. Analitical and Bioanalitical Chemistry 2009, 395, 729-735. 104. Liu, Y.; Todd, E. E.; Zhang, Q.; Shi, J. R.; Liu, X. J., Recent developments in the detection of melamine. Journal of Zhejiang University. Science. B 2012, 13, 525-32. 105. Tran, B. N.; Okoniewski, R.; Storm, R.; Jansing, R.; Aldous, K. M., Use of methanol for the efficient extraction and analysis of melamine and cyanuric acid residues in dairy products and pet foods. Journal of Agricultural and Food Chemistry 2009, 58, 101-107. 106. Curcio, M.; Puoci, F.; Cirillo, G.; Iemma, F.; Spizzirri, U. G.; Picci, N., Selective determination of melamine in aqueous medium by molecularly imprinted solid phase extraction. Journal of Agricultural and Food Chemistry 2010, 58, 11883-11887. 107. Pietrzyk, A.; Kutner, W.; Chitta, R.; Zandler, M. E.; D’Souza, F.; Sannicolo, F.; Mussini, P. R., Melamine acoustic chemosensor based on molecularly imprinted polymer film. Analytical Chemistry 2009, 81, 10061-10070. 108. Baggiani, C.; Anfossi, L.; Giovannoli, C., Solid phase extraction of food contaminants using molecular imprinted polymers. Analytica Chimica Acta 2007, 591, 29-39. 109. Pichon, V., Selective sample treatment using molecularly imprinted polymers. Journal of Chromatography A 2007, 1152, 41-53. 93  110. Qian, K.; Fang, G.; He, J.; Pan, M.; Wang, S., Preparation and application of a molecularly imprinted polymer for the determination of trace metolcarb in food matrices by high performance liquid chromatography. Journal of Separation Science 2010, 33, 2079-2085. 111. Wang, X.; Fang, Q.; Liu, S.; Chen, L., The application of pseudo template molecularly imprinted polymer to the solid-phase extraction of cyromazine and its metabolic melamine from egg and milk. Journal of Separation Science 2012, 35, 1432-1438. 112. Koglin, E.; Kip, B. J.; Meier, R. J., Adsorption and displacement of melamine at the Ag/electrolyte interface probed by surface-enhanced Raman microprobe spectroscopy. The Journal of Physical Chemistry 1996, 100, 5078-5089. 113. Chen, L.; Liu, Y., Surface-enhanced Raman detection of melamine on silver-nanoparticle-decorated silver/carbon nanospheres: effect of metal ions. ACS Applied Materials and Interfaces 2011, 3, 3091-3096. 114. Chen, D.; Grant, E., Evaluating the validity of spectral calibration models for quantitative analysis following signal preprocessing. Analitical and Bioanalitical Chemistry 2012, 404, 2317-2327. 115. Kantarovich, K.; Tsarfati, I.; Gheber, L. A.; Haupt, K.; Bar, I., Reading microdots of a molecularly imprinted polymer by surface-enhanced Raman spectroscopy. Biosensors and Bioelectronics 2010, 26, 809-814. 116. He, L.; Lamont, E.; Veeregowda, B.; Sreevatsan, S.; Haynes, C. L.; Diez-Gonzalez, F.; Labuza, T. P., Aptamer-based surface-enhanced Raman scattering detection of ricin in liquid foods. Chemical Science 2011, 2, 1579-1582. 94  117. Chen, S.; Li, X.; Zhao, Y.; Chang, L.; Qi, J., High performance surface-enhanced Raman scattering via dummy molecular imprinting onto silver microspheres. Chemical Communications 2014, 50, 14331-14333.  95  Appendices  Appendix A  Chapter 3 supplementary information   A.1 Confirmation of the formation of melamine hapten-ovalbumin immunogen  Figure A.1 Ultraviolet spectrum of melamine hapten-ovalbumin immunogen Source: Shi X.W. Development of enzyme-linked immunosorbent assay for the determination of melamine. M.S. thesis, Tianjin University of Science and Technology, 2012  The changes in UV absorbance represent the successful formation of hapten-OVA complex.   96  Appendix B  Chapter 4 supplementary information  B.1 Influence of flow rate on the recovery of melamine from NISPE Table B.1 The recovery of 0.005 mmol L-1 melamine in methanol from NISPE Flow rate Recovery of melamine in methanol 1.5 mL min-1 60.3% 2.0 mL min-1 69.0%  The increased flow rate could increase the recovery of melamine in methanol from NISPE, due to the fewer interaction opportunities between washing solvent and molecules adsorbed on the sorbent.   

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0166414/manifest

Comment

Related Items