Spoilage Indicators for Determining Tuna and Mahi-Mahi Quality and Safety

Home , Isobutylamine, Trimethylamine, Valine decarboxylase

SPOILAGE INDICATORS FOR DETERMINING TUNA AND MAHI-MAHI QUALITY AND SAFETY

JING BAI

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2018

To my family, faculty advisors, and my friends who teach me how to love this wonderful world

ACKNOWLEDGMENTS

First of all, I would like to express my sincere appreciation to my committee chair,

Dr. Paul J. Sarnoski for his patient advice and encouragement in the past three years.

He puts his trust in students and can see the true potential of his students. He cares so much about students’ projects and provides insightful advice about the research. I also sincerely appreciate my committee members, Dr. Renée M. Goodrich-Schneider, Dr.

Shirley M. Baker, Dr. Naim Montazeri, and Dr. George L. Baker for spending their precious time to provide valuable guidance and aid through this process. I would like to thank all the faculty members for imparting knowledge to me, all the staff members for helping me prepare paperwork, and order laboratory supplies in the Food Science and

Human Nutrition Department at UF.

I would especially like to thank the Yeoman Fellowship Fund and the Seafood

Industry Research Fund (SIRF) for supporting my research.

I would like to express special thanks to my great family, including my husband, my parents, my parents in-law and my little son, for their support and constant encouragement. I express my gratitude to Yangyang Song, my husband, for supporting every decision I make and helping me solve any problems I meet. I am extremely grateful to my parents for teaching me to develop integrity and telling me how to face the challenge in the life. I am not afraid of difficulties in the life because I understand my family always stands behind me and silently support me.

I would like to thank my colleagues La’Oshiaa Reed, Stephen Koltun, Robert

Nusbaum, Yaozhou Zhu and Ying Fan in our lab for providing aid in my projects and making my stay in UF much more pleasurable. I also would like to express my gratitude to all my friends in my life for their warm love and endless encouragement.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 7

LIST OF FIGURES ...... 8

ABSTRACT ...... 9

CHAPTER

1 INTRODUCTION ...... 11

2 LITERATURE REVIEW ...... 14

Economics of Tuna and Mahi-Mahi...... 14 Aquaculture of Tuna and Mahi-Mahi ...... 15 Fish Spoilage ...... 16 Amino Acids and Fish Spoilage ...... 19 Biogenic Amine Concerns ...... 20 Histamine in Tuna and Mahi-Mahi ...... 22 Effect of Spoilage on Fish Volatile Compounds ...... 24 Colorimetric Strips for the Detection of Volatile Amine...... 27 GC Methods to Determine the Aroma Profile of Seafood as Chemical Indicators of Spoilage ...... 30 HPLC and UHPLC Methods for Biogenic Amine Detection ...... 34 ELISA Detection of Histamine ...... 37

3 A RAPID UHPLC METHOD FOR THE SIMULTANEOUS DETERMINATION OF AMINO ACIDS AND BIOGENIC AMINES IN TUNA AND MAHI-MAHI ...... 40

Digest...... 40 Background Information and Objectives ...... 40 Materials and Methods...... 43 Fish Samples and Preparation ...... 43 Standards and Reagents ...... 43 Extraction and Derivatization ...... 44 Determination of Biogenic Amines ...... 45 Method Validation ...... 46 Histamine ELISA Test Kit ...... 46 Results and Discussion...... 47 Method Development ...... 47 Linearity and Sensitivity ...... 49 Recovery and Repeatability ...... 49 Resolution and Theoretical Plates ...... 51

Application to Different Spoilage Grade of Mahi-Mahi (Coryphaena hippurus) and Yellowfin Tuna (Thunnus albacares) ...... 51 Summary ...... 58

4 AROMA PROFILE CHARACTERIZATION OF MAHI-MAHI AND TUNA FOR DETERMINING SPOILAGE USING PURGE AND TRAP GAS CHROMATOGRAPHY-MASS SPECTROMETRY (PT-GC-MS) ...... 66

Digest...... 66 Background Information and Objectives ...... 67 Materials and Methods...... 70 Fish Samples and Preparation ...... 70 Standards and Reagents ...... 70 Extraction Procedures ...... 71 Purge and Trap Conditions ...... 72 GC-MS Analysis ...... 72 Calculations ...... 72 Results and Discussion...... 73 Summary ...... 82

5 DETERMINING QUALITY ATTRIBUTES OF MAHI-MAHI AND TUNA BY OPTIMIZED COLORIMETRIC STRIPS ...... 90

Digest...... 90 Background Information and Objectives ...... 91 Materials and Methods...... 93 Fish Samples and Preparation ...... 93 Standards and Reagents ...... 94 Colorimetric Strip Method ...... 95 Determination of Biogenic Amines and Free Amino Acids by UHPLC (Conducted in Chapter 3) ...... 96 Determination of Aroma Profile by PT-GC-MS (Conducted in Chapter 4) ...... 96 Statistical Analysis ...... 97 Results and Discussion...... 97 Summary ...... 106

6 CONCLUSION ...... 112

APPENDIX: EXTERNAL STANDARD PREPARATION FOR UHPLC METHOD ...... 115

LIST OF REFERENCES ...... 116

BIOGRAPHICAL SKETCH ...... 133

LIST OF TABLES

Table page

3-1 Linearity for different amino acids and amines analyzed by UHPLC ...... 59

3-2 Percentage recoveries of three biogenic amines ...... 60

3-3 The number of theoretical plates, and resolution of biogenic amines ...... 60

3-4 Amino acids and biogenic amines (mg/kg) in seven grades of mahi-mahi (M) ... 61

3-5 Amino acids and biogenic amines (mg/kg) in seven grades of tuna (T) ...... 62

3-6 ELISA results of mahi-mahi (M) and tuna (T) calculated by using the standard provided in kit and standard prepared in lab...... 63

3-7 Pearson correlation coefficients (r) between ELISA results of mahi-mahi (M) and tuna (T) calculated by using the standard provided in the kit, standard prepared in lab and histamine results from UHPLC method...... 64

4-1 Volatile compounds associated with spoilage in seven grades of mahi-mahi calculated by internal standard method, (ng/g) fish sample...... 85

4-2 Volatile compounds associated with spoilage in seven grades of tuna calculated by internal standard, (ng/g) fish sample...... 86

4-3 Biogenic amines contents in seven grades of mahi-mahi and tuna sample (ng/kg) calculated by spiking standard method and external standard method...... 87

4-4 Flavor descriptors of volatile compounds associated with spoilage in mahi- mahi and tuna samples. Pearson correlation coefficients between levels of volatile compounds with increasing spoilage grade of mahi-mahi and tuna ...... 89

5-1 Linearity of rose bengal strips and BPB strips ...... 109

5-2 Volatile biogenic amines in seven grades of mahi-mahi samples detected by rose bengal and BPB strips for fish samples (n=5) ...... 109

5-3 Volatile biogenic amines in seven grades of tuna sample calculated by rose bengal and BPB strips for fish samples (n=5) ...... 109

5-4 Pearson correlation coefficients (r) between methods for mahi-mahi (n=3)...... 110

5-5 Pearson correlation coefficients (r) between methods for tuna (n=3) ...... 111

A-1 Concentrations of each amino acid in five levels of external standard solutions and stock solution, mg/L solution...... 115

LIST OF FIGURES Figure page

2-1 Global catches of albacore, bigeye, skipjack and yellowfin data from 1960 to 2016. Data from WCPFC (2016)...... 39

3-1 Chromatographic separations of Mahi-mahi grade 1 sample spiked with 10ppm of each amino acid and biogenic amine standards...... 65

4-1 Example of a chromatogram (mahi-mahi grade 7)...... 84

5-1 Biogenic amine cocktail standard solutions reacted with rose bengal strips. Photo courtesy of author...... 108

5-2 Biogenic amine cocktail standard solutions reacted with BPB strips. Photo courtesy of author...... 108

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

SPOILAGE INDICATORS FOR DETERMINING TUNA AND MAHI-MAHI QUALITY AND SAFETY

Jing Bai

August 2018

Chair: Paul J. Sarnoski Major: Food Science

The consumption of spoiled fish containing high levels of histamine result in the highest incidence of illness from fish poisoning. Tuna (Thunnus albacares) and mahi- mahi (Coryphaena hippurus) are two major fish species responsible for histamine poisoning in the United States. The main purpose of this research was to develop spoilage indicators for determining tuna and mahi-mahi quality and safety. A reversed- phase ultra-high performance liquid chromatography (UHPLC) method, purge and trap gas chromatography-mass spectrometry (PT-GC-MS), and two color strip methods were developed and optimized to be used for determining fish spoilage. The rapid

UHPLC method developed in this study could identify and quantify dansylated amino acids, histamine and other biogenic amines that can act as co-indicators of histamine

(scombroid) poisoning in tuna and mahi-mahi fish sample simultaneously within 17.5 minutes. This UHPLC method showed good linear response, sensitivity, resolution, percentage recovery, repeatability, and number of theoretical plates. Twenty aroma compounds in mahi-mahi and sixteen volatile compounds in tuna associated with fish spoilage could be determined by this purge and trap GC-MS method without a derivatization procedure. Volatile compounds identified as key spoilage indicators of

tuna and mahi-mahi were amines (dimethylamine, trimethylamine, isobutylamine, 3- methylbutylamine, and 2-methylbutanamine), alcohols (2-ethylhexanol, 1-penten-3-ol and isoamyl alcohol, ethanol), aldehydes (2-methylbutanal, 3-methylbutanal, benzaldehyde), ketones (acetone, 2,3-butanedione, 2-butanone, acetoin) and dimethyl disulfide. A rose bengal strip, and a bromophenol blue strip created in this study produced standard curves with good linearity and also showed uniform colorimetric response to volatile amines. The colorimetric strips were validated by investigating the correlation of the results obtained by colorimetric strips with the increasing spoilage grades of fish, and results obtained by histamine-specific ELISA kit, UHPLC and PT-

GC-MS and satisfactory correlations were obtained. The three detection methods developed in this study can be used to monitor the quality changes of mahi-mahi and tuna.

CHAPTER 1 INTRODUCTION

Tuna is one of the top five consumed seafood species and accounts for around 5% of fisheries and aquaculture production for human consumption in the world (Paquotte,

2003). The annual global catch of tuna in the ocean (wild caught) increased from

698,260 tonnes in 1960 to 4,857,709 tonnes in 2016 (WCPFC, 2016). Mahi-mahi

(Coryphaena hippurus) is found mostly in tropical regions and most of the catch occurs in the Pacific Ocean. The annual landings of mahi-mahi have increased 7.5 folds in last

60 years (Whoriskey et al., 2011). One of the largest consumers of mahi-mahi is the

United States (Hunter, 2013).

Fish spoilage means any change in the condition of fish that leads to fish becoming less palatable or even toxic. These changes include off-flavors formation, amino acids changing, texture deterioration, discolorations, the decrease of nutritional value and other alterations in fish (Ashie et al., 1996). During fish spoilage, toxic biogenic amines, including histamine, cadaverine, and putrescine may be produced in certain fish species (Bulushi et al., 2009). The highest incidence of illness from fish poisoning is associated with the consumption of time and temperature abused scombroid fish that contain significant amounts of histamine (Morrow et al. 1991). Two major fish species responsible for histamine poisoning in the United States are from tuna and mahi-mahi (Ahmed, 1991).

The colorimetric strip is a cost-effective method that has been widely used in food analysis. This method is based on the principle that chemical indicators or bioactive sensors adhered to the treated papers could change color when reacting with the specific compound in food. The pH paper is used in a wide range of food laboratory

to industrial applications and the paper changes color under the influence of hydroxide or hydrogen ions in the food system. The other major applications of colorimetric strips in food science are the detection of heavy metals, testing milk pasteurization, determination of toxins and foodborne pathogens. In 2016, a new kind of indicator strip combining bromophenol blue (BPB) was developed to detect degradation levels of seafood (Dole et al., 2016).

Gas chromatography-mass spectrometry (GC/MS) is an instrument that is mostly used to identify and quantify volatile and semi-volatile compounds in seafood (Duflos et al., 2006; Grimm et al., 2000; Wong et al., 1967). Separation of components is based on the principle that different compounds have different strengths of interaction with the stationary phase. Mass spectrometry can sensitively identify molecular weight of fragment molecules. Volatile compounds have been used as indicators for the quality assessment of seafood products (Wierda et al., 2006; Soncin et al., 2008; Alasalvar et al., 2005).

Ultra-high performance liquid chromatography (UHPLC) is an advanced type of separation technology. UHPLC has the same principle as high-performance liquid chromatography (HPLC) in that the separation of compounds is dependent on the compound affinity between mobile and stationary phases. Comparing with regular LC system, UHPLC can be operated under pressure as high as 120 MPa, be packed with silica column with smaller particle size, separates molecules faster with a higher resolution, and needs less mobile phase (Wu et al., 2001). UHPLC has been widely used in determining free amino acids and biogenic amines associated with fish spoilage

(Jia et al. 2012; Simat et al., 2011).

The main purpose of this research was to develop new or refined spoilage indicators for determining tuna and mahi-mahi quality and safety. A rapid UHPLC method was developed to identify and quantify amino acids, histamine and other biogenic amines that can act as co-indicators of histamine (scombroid) poisoning in tuna and mahi-mahi fish samples. A GC-MS method was set up to determine the aroma profile of mahi-mahi and tuna for chemical indicators of spoilage. Rapid detection strips were developed and optimized to change color uniformly and give good linearity. The correlations between the colorimetric strips with UHPLC, GC, and ELISA were reported.

CHAPTER 2 LITERATURE REVIEW

Economics of Tuna and Mahi-Mahi

Tuna is a kind of saltwater fish and belongs to Scombridae family. Around 50 species fall into the category of tuna and the five major species of tuna for consumption are albacore (Thunnus alalunga), bigeye (Thunnus obesus), bluefin (Thunnus thynnus), skipjack (Katsuwonus pelamis) and yellowfin (Thunnus albacares) (Vinas, 2009). Tunas are widely distributed in oceans around the world. Tuna plays an important role in seafood international trade and accounts for around 5% of fisheries and aquaculture production for human consumption in the world (FAO, 2010). The annual global catch of tuna has tended to increase and the global catch of four major tuna species from 1960 to 2014 are shown in Figure 2-1 (WCPFC, 2016). There are three major parts of the global tuna market: sashimi market, fresh and frozen tuna, as well as canned tuna

(Jimenez-Toribio et al., 2010). The United States is the second largest importer of tuna after Japan (FAO, 2010). The top five species of fresh tuna imported to the USA are yellowfin, bigeye, albacore, bluefin, and skipjack. Fresh and frozen tuna, canned tuna, are imported into the United States with an estimated amount of 287,440 tonnes per year (FAO, 2017).

Mahi-mahi, which also named as common dolphinfish (Coryphaena hippurus), is a migratory pelagic fish and is found mostly in the tropical regions of the world. About sixty countries are known for mahi-mahi landings and most of the catch occurs in the

Pacific Ocean. Peru, Taiwan Province of China, Indonesia and Ecuador are the major countries where mahi-mahi are landed the most. Nearly 60% of mahi-mahi imported into the United States are from Peru and Ecuador. Global landings of mahi-mahi increased

from 7000 tonnes in 1950 to 103,000 tonnes in 2013 (FAO, 2016). The United States is one of the largest consumers of mahi-mahi (Hunter, 2013).

Aquaculture of Tuna and Mahi-Mahi

During the last decades, aquatic farming has been a fast-growing food production industry powered by technological impulsion. Currently, bluefin tuna

(Thunnus thynnus) is the dominate species in tuna aquaculture. Bluefin tuna is a valuable tuna species that can be sold through the selective sushi and sashimi market

(Tseng et al., 2012). The global catch of wild caught bluefin tuna decreased from 89,000 tones in 1980 to 42,000 tonnes in 2011 due to unsustainable fishing decreasing wild stocks (Metian et al., 2014). To meet the continuous high demand of bluefin tuna, aquaculture of this species has been under development over the past thirty years.

Large-scale commercial bluefin tuna aquaculture started in the 1980s, and now accounts for 18% of global bluefin tuna production (Metian et al., 2014). The

Mediterranean region, Mexico, Australia, and Japan are major regions that perform bluefin tuna aquaculture. Japan is the largest importer of farmed bluefin tunas. Before

2007, almost all of the Mediterranean farmed Atlantic bluefin were exported to the

Japanese market. However, after Mexico began to farm Pacific bluefin at a lower production cost, Mexican farmed bluefin has been competitive in the Japanese market

(FAO 2010).

Mahi-mahi has been considered as a promising candidate species for commercial aquaculture production due to its world-wide consumption. The aquaculture development of the mahi-mahi began in 1981, and circular tanks were used to stock the fish (Lee and Ostrowski, 2001). However, raising mahi-mahi on a large commercial- scale is not achieved by far due to the technological challenges. Mahi-mahi would be

sexually mature after six months, and mono-sex culture is best to achieve a high growth rate of fish. However, there is no current solution to this problem. The size of mahi-mahi harvested at six months is not comparable with the wild fish, which weight is generally above 5 kg. New techniques are needed to make the farmed mahi-mahi to have equivalent quality with the wild mahi-mahi already in the market.

Fish Spoilage

Fish spoilage can happen rapidly after fish landing and various components break down or new compounds form during this process. The three main types of fish spoilage mechanisms are microbial, enzymatic, and chemical (Ghaly et al., 2010). Lipid oxidation, protein degradation and the decrease of other valuable molecules are major concerns of fish spoilage (Clancy et al. 1995; Ghaly et al., 2010). Every year, almost 10 to 12 million tonnes of fish are lost due to spoilage accounting for ten percent of the total fish production (FAO, 2010). Fish spoilage is considered as an important aspect of food safety.

The growth and metabolism of microbes is considered as a main reason leading to the spoilage of fish. Higher levels of free amino acids and trimethylamine oxide exist in fish materials than other types of meat and these substances are good microbial substrates (Gram, 2002). Fish has an immune system to prevent bacteria from invading the fish tissue but the immune system stops after fish death. Bacteria then can enter the fish through the skin and contaminate the body cavity, belly, gill tissue, and kidney and proliferate freely (Fraser and Sumar, 1998). Microorganisms with amino acid decarboxylase activity can produce biogenic amines, and sulphides with unpleasant odors (Ghaly et al., 2010). Histamine is produced in raw fish by the reaction of the bacterial histamine decarboxylase. Histamine is produced by gram-positive lactic acid

bacteria in fermented products, such as wine, aged cheeses, and fish sauce. However, in raw fish tissue, histamine is produced by gram-negative enteric bacteria such as

Enterobacter aerogenes, Morganella morganii and Pseudomonas aeruginosa

(Hungerford, 2010). These bacteria, which produce histamine in fish, generally exist in the saltwater environment and are naturally present on the external surfaces and inside of live fish (Visciano et al., 2012). There are two routes of synthesis of putrescine in fish.

Arginine can indirectly produce putrescine by arginine decarboxylase via agmatine. Also, putrescine can be formed from arginine by arginine deiminase, ornithine carabamoyltransferase and ornithine decarboxylase (Prester 2011). The genus

Staphylococcus is the major bacteria producing putrescine in fish tissue (Wunderlichova et al., 2014). Free lysine can produce cadaverine by lysine decarboxylase and it has been known that various species of bacteria have lysine decarboxylase activity. A research study showed that ninety-two percentage of mesophilic bacteria with decarboxylase activity isolated from mahi-mahi had lysine decarboxylase activity (Frank et al., 1985). Shewanella putrifaciens, Photobacterium phosphoreum and Vibrionacaea are the major bacteria responsible for the production of trimethylamine (TMA), which has a strong fishy odor (Ghaly et al., 2010). Beyond producing amines, the growth of microbes, such as Shewanella putrifaciens and Pseudomonas perolens can also lead the formation of short-chain carbonyls, alcohols, esters, sulfur compounds and others with unpleasant odors (Duflos et al., 2010; Jorgensen et al., 2001).

Enzymatic spoilage is another basic mechanism of fish spoilage. The muscle tissue and gut of fish contain endogenous enzymes and these enzymes lead to autolytic reactions in spoiled fish. Autolytic enzymes in fish influence the texture of fish tissues,

such as tenderization and flesh softening. The changes of texture shorten the lifetime of the fish product and have negative effects on the quality. The accumulations of hypoxanthine and formaldehyde as a result of postmortem endogenous enzymatic activity influence the textural quality of fish. After fish death, glycogen is hydrolyzed to lactic acid and this glycolysis process is a result of endogenous enzymatic activity. Due to the accumulation of lactic acid, the pH of fish meat falls, and the fish tissue is more susceptible to bacterial growth. Proteolytic enzymes widely exist in muscle and the viscera of fish, and belly bursting can be caused by these enzymes (Ghaly et al., 2010).

Muscle proteins can be hydrolyzed by endogenous proteolytic enzymes after the death of fish and the rigor mortis was observed. As the process of fish spoilage continues, rigor resolves, and the muscle of fish become becomes limp (Cheret et al., 2007;

Olafsdottir et al., 1997). The degradation of proteins produces free amino acids and peptides, which can be used as the nutrition source for microbial growth. Lipid oxidation in fish tissue is also influenced by endogenous enzymes, such as lipoxygenase and peroxidase, and unpleasant odors are formed during this process as well (Hultmann et al., 2004).

Chemical spoilage is the third mechanism of fish spoilage and mainly includes oxidative rancidity and non-enzymatic browning. Fish tissue is rich in unsaturated fatty acids that can be oxidized during fish spoilage. Hydroperoxides are produced during the propagation procedure in lipid oxidation and then break down to form compounds with unpleasant flavors. Research showed that the non-enzymatic lipid oxidation of tuna meat continued during frozen storage, and frozen tuna had higher peroxide values than fresh tuna (Tanaka et al., 2016). The Maillard reaction involving amino acids, peptides

and reducing sugars in fish is a non-enzymatic browning and leads color changes and formation of specific flavors in spoiled fish that has been cooked (Ashie et al., 1996).

The autoxidation of myoglobin to metmyoglobin, is also responsible for the browning discoloration in spoiled fish (Genigeorgis, 1985).

Amino Acids and Fish Spoilage

Seafood materials are rich in free amino acids and peptides that are produced from autolysis of fish muscle proteins and are important substrates or catalysts for reactions pertaining to fish spoilage (Fraser and Sumar, 1998). The protein content in tuna is around twenty-three percent of the wet weight basis and is the source of free amino acids (Peng et al., 2013). Essential amino acids are those amino acids that cannot to be synthesized in humans and thus need to be obtained from diet. Free tyrosine, tryptophan, threonine, cystine, valine, lysine, methionine, isoleucine, leucine, phenylalanine are essential amino acids that have been identified in tuna (Sen, 2005).

Mahi-mahi is a rich source of protein and contains eighteen amino acids, including alanine, arginine, aspartic acid, cystine, glutamic acid, glycine, histidine, isoleucine, leucine, leucine, methionine, phenylalanine, proline, serine, threonine, tryptophan, tyrosine, valine, which have been determined in mahi-mahi to participate in building muscle protein in fish (Ostrowski et al., 1989). Free histidine, ornithine, lysine, and glutamine was identified and quantified in mahi-mahi by Antoine et al. (2002). However, research about investigating the other major free amino acids in mahi-mahi has not been studied.

Amino acids are essential components and play the central role in metabolic pathways. Oxidative rancidity is a non-enzymatic mechanism that produces hydroperoxides and has been considered as a major cause of fish spoilage for a long

time (Ghaly et al., 2010). In this process, unsaturated fatty acids or triglycerides in fish are oxidized and "rancid" odors, flavors are released. Amino acids have been found to catalyze this oxidation reaction alone or in association with specific trace metal ions

(Ashie, 1996). Amino acids and peptides in fish also participate in non-enzymatic browning, such as the Maillard reaction, and cause discoloration of fish muscle (Ocano-

Higuera, 1992). Amino acids are also precursors of some substances, such as biogenic amines. Biogenic amines are organic bases and are produced in fish by microbial decarboxylation of amino acids or by transamination of amino acids (Zhai et al., 2012).

Many biogenic amines have biological activity and can influence physiological functions in human body. Histamine, cadaverine, and putrescine are biologically active amines and have been widely studied due to their toxicity (FDA, 2011). During spoilage, deamination of amino acids produce ammonia, and deamination of sulfur-containing amino acids form sulfur compounds which give unpleasant off-odors to seafood

(Herbert and Shewan, 1975).

Biogenic Amine Concerns

Among the biogenic amines found to occur during the spoilage process of fish, only histamine, cadaverine, and putrescine are considered as significant markers of food quality (Bulushi et al., 2009). It has been found that though decarboxylation, histamine, cadaverine, and putrescine are produced from the free amino acids histidine, lysine, and arginine, respectively (Prester, 2011). Histamine has a heterocyclic structure, cadaverine and putrescine have an aliphatic structure (Mohamed et al., 2009). Although the toxicological levels of individual biogenic amines are difficult to establish, a maximum level of total amines has been proposed as 750–900 mg/kg (Ladero et al.,

2010).

Histamine is identified as the major natural chemical responsible for fish poisoning. Histamine intoxication was first found originating from the family Scombridae

(Lehane and Olley, 2000). One characteristic of scombroid fish is that these fish contain a high amount of histidine, which is the precursor of histamine. Other non-scombroid fish species, such as mahi-mahi, are also implicated in scombroid poisoning (FDA,

2005). Some bacteria have been found to produce histamine in fish samples at a temperature as low as 0 °C and this phenomenon makes it difficult to prevent histamine formation in fish products (Hungerford, 2010). Research pointed out that histamine did not distribute uniformly in spoiled fish (Lehane and Olley, 2000). Histamine content that exceeds a concentration of 50 ppm (5 mg/100g) in tuna and mahi-mahi represented the decomposition in these fish (FDA, 2005). According to the fish and fishery products hazards and controls guidance (FDA, 2011), illness-causing fish mostly contains more than 200 mg/kg histamine. Putrescine and cadaverine can potentiate histamine toxicity by inhibiting the intestinal histamine-metabolizing enzymes and diamine oxidase

(Bulushi et al., 2009; Visciano et al., 2012).

Scombroid poisoning can cause allergy-like symptoms and the onset of scombroid poisoning is rapid, which range is from several minutes to 3 hours (Bulushi et al., 2009). After ingestion of spoiled fish containing more than 100 ppm of histamine, the person might have symptoms including oral numbness, headache, dizziness, palpitations, difficulty in swallowing and some allergy-like symptoms (Bulushi et al.,

2009; Hungerford, 2010). However, some people are sensitive to the biogenic amines and even ingesting a low amount of histamine can lead to the onset of symptoms.

Histamine, putrescine, and cadaverine are considered essential spoilage markers of fish products due to ability to show the microbial contamination and degradation reactions during storage. The amount of histamine and putrescine of mackerel (S. scombrus) increased after storage of fish samples at 22 °C for 12 hours.

These two biogenic amines were considered as quality markers of mackerel (Prester et al., 2009). The formation of histamine, putrescine, and cadaverine in herring was observed after the fish samples were stored at 10 °C for two days (Mackie et al., 1997).

Histamine, putrescine, and cadaverine also accumulated in sardines after a storage period at 4 °C and these three biogenic amines were identified as quality indicators of sardines (Sardina pilchardus) (Ozogul et al., 2006). Histamine content in sardine increased to 620 ppm after the sardine samples stored at 25 °C for 24 hours (Visciano et al., 2007).

Histamine in Tuna and Mahi-Mahi

Tuna and mahi-mahi have been considered as two major sources of scombroid poisoning (Ahmed, 1991; FDA, 2011). Histidine is a precursor of histamine formation and the muscle tissue of tuna and mahi-mahi contains large amounts of histidine

(Bulushi et al., 2009). Free histidine levels in tuna are around 7 g/kg and in mahi-mahi are around 5 g/kg (Antoine et al., 1999). Histamine production in fish depends on the level of endogenous histidine in the fish, the presence of bacterial histidine decarboxylase, and the environmental conditions (Visciano et al., 2012). The decarboxylase enzymes produced in spoiling fish by certain bacteria can convert amino acids to biogenic amines (Lehane and Olley, 2000). Freshly caught fish have the low level of histamine, usually are less than 0.1 mg/100g (Auerswald et al., 2006).

Histamine is formed from free histidine by bacterial histidine decarboxylases in fish

tissue, usually when fish is exposed to elevated temperatures after the catch (Visciano et al., 2012). There are 112 species of bacteria that have been identified as histidine- decarboxylating bacteria (Taylor et al., 1978). The family Enterobacteriaceae, the genera Clostridium and Lactobacillus are the major bacteria families responsible for histidine decarboxylation (Lehane and Olley, 2000). Histidine-decarboxylating bacteria are present at a great proportion of the microbial population when fish spoil. Lehane and

Olley (2000) found that 31% of isolates from decomposing skipjack tuna and 7% of isolates from spoiled mahi-mahi growing at warm temperature were histidine- decarboxylating bacteria. Decarboxylase enzymes produced by endogenous bacteria are insignificant when compared with those produced by exogenous sources (Lehane and Olley, 2000). Due to histidine-decarboxylating bacteria growing rapidly when the temperature is near 32.2 °C, high temperature spoilage is identified as the main reason leading to accumulation of histamine in fish (FDA, 2011).

Research showed that after twelve-day storage at 7 °C, the histamine concentration in mahi-mahi increased from 0 mg/100g to 160 mg/100g, while the histidine concentration decreased from 400 mg/100g to 180 mg/100g (Antoine et al.,

2002). Histamine amount in both red and white muscle of tuna increased after storage in a controlled environment for 33 days, and histidine amount of these fish samples decreased (Ruiz-Capillas and Moral, 2004). The amount of histamine, cadaverine, putrescine of tuna increased after a storage period, and these biogenic amines were considered as hygienic quality markers of tuna (VecianaNogues et al., 1997). Rossi et al. (2002) reported that after 48 hours storage at room temperature, the levels of

histamine, cadaverine, and putrescine in Skipjack (Katsuwonus pelamis) increased to

1533, 649 and 60 mg/kg, respectively.

Effect of Spoilage on Fish Volatile Compounds

The volatile profile is one essential quality parameters of fish meat, and it can reflect the organoleptic characteristic of the fish product (Edirisinghe et al., 2007).

Volatile compounds produced in fish meat are mainly based on microbial action, enzymatic action, lipid oxidation and other chemical reactions. Several specific alcohols, carbonyls, acids, amines, sulfur compounds, aldehydes, and ammonia have been identified as spoilage indicators of fish products due to the content of these compounds changing during fish spoilage (Ashie et al.,1996; Duflos et al., 2006).

Short-chain carbonyls, alcohols, and esters can generate due to the microbial spoilage, enzymatic or non-enzymatic lipid oxidation. Ethanol, 2,3-butanediol 1-penten-

3-ol, 3-methyl-1-butanol, 1-butanol, and 1-octen-3-ol were alcohols, which content in tested fish samples increased during fish spoilage and response for the pungent, alcoholic and creamy odor (Leduc et al., 2012; Duflos et al., 2005; Olafsdottir et al.,

2005). The level of ethanol reached 314 mg/kg in pink salmon (Oncorhynchus gorbuscha) after three days storage at 10 °C due to the microorganisms utilizing carbohydrates (Himelbloom et al., 2013). The accumulation of branched-chain alcohols in spoiled fish is reported as the result of degradation of amino acids (Rehbein and

Oehlenschlager, 2009). The aldehydes 3-methylbutanal and 2-methylbutanal the fishy odor and have been identified as fish spoilage indicator due to their formation in whiting, mackerel and cod flesh during spoilage (Duflos et al., 2005). Ethyl acetate and ethyl butanoate are two esters determined in spoiled fish and contribute fruity and sweet odors (Iglesias et al., 2009; Olafsdottir et al., 2005).

Volatile amines, such as trimethylamine (TMA), dimethylamine (DMA), and isobutylamine, are identified as essential spoilage indicators due to their gradual accumulation during fish spoilage and the contribution of characteristic fishy odor.

Trimethylamine oxide (TMAO) has been widely found in marine fish and is used as an osmoregulant by fish to avoid dehydration and balance changing salt levels from the environment (Gram et al., 2002). Gram-negative bacteria can reduce TMAO to trimethylamine (TMA) to obtain energy. TMAO is non-odorous, however, TMA is a volatile compound with a very low odor threshold and a stale fish odor. The level of TMA in fish has been identified as an indicator of microbial deterioration in fish (Fraser and

Sumar, 1998). TMA has been used as an effective marker to distinguish the fresh and spoiled fish samples (Leduc et al., 2012; Bene et al., 2001; Ghaly et al., 2010).

Trimethylamine oxide in fish fillets can be also reduced to dimethylamine (DMA) during spoilage of fish. The concentration of DMA in freshly caught fish is as low as 2 ppm.

DMA starts to accumulate automatically from TMAO by the activity of endogenous enzymes in the very early stage of spoilage (Chan et al., 2006). DMA has an ammonia- like odor, and the amount of DMA has been accepted as an index for fish freshness.

Research showed that the content of DMA increased in albacore tuna after frozen storage (Ben-gigirey et al., 1999).

Other major volatile amines produced during spoilage of fish are ammonia and isobutylamine. Ammonia usually already present in freshly caught fish and accumulates during fish storage by deamination of amino acids. Isobutylamine has fishy type odor and is commonly considered as the microbial degradation product. This amine is

produced by decarboxylation of valine during fish spoilage (Gruger et al., 1972; Eskin,

2013; Gill et al., 1983).

Sulfur compounds are considered as another key component of the volatile compounds formed in fish spoilage process (Gram et al., 2002). Sulfur compounds are originally in low concentrations in the fish body and accumulate after fish landing (Duflos et al., 2006). Sulfur compounds have extremely low thresholds and give out very unpleasant odors. The generation of sulfur compounds in fish is mainly by microbial enzymatic activity (Ashie et al., 1996). Plenty of sulfur-containing amino acids, peptides exist in fish tissue and the degradation of these compounds during spoilage process produce the odorous sulfur compounds. Hydrogen sulfide, dimethyl disulfide, dimethyl trisulfide and methanethiol are sulfur compounds commonly found in spoiled fish (Kawai et al., 1996). Methylethyl disulfide, 3-(methylthio)-propanal, 1-(methylthio)-propane, 2- methyl-3-furanthiol are sulfur compounds that also be found in spoiled tuna (Varlet and

Fernandez, 2010).

Volatile acids, such as formic acid, acetic acid, and propionic acids, are produced from the breakdown of certain amino acids and atmospheric oxidation of lipids

(Olafsdottir et al., 2005; Koutsoumanis et al., 1999). These volatile acids also contribute to the odor of spoiled fish. The formation of acetic acid in cod (Gadus morhua) was observed after a ten-day spoilage process, and this formation was associated with microbial activity (Duflos et al., 2006). The concentration of formic acid and acetic acid in smoked salmon increased with storage time (Hansen et al.,1995). Terpenes are compounds already exist in freshly caught fish, and some specific terpenes accumulate during the spoilage process and contribute to odor change of fish products. Limonene

concentration in seabream and Baltic herring increased during frozen storage (Iglesias et al., 2009; Aro et al., 2003). Beyond the limonene concentration increasing, α-pinene,

3-carene also accumulated during the spoilage of seabream (Alasalvar et al., 2005).

Alkanes and alkenes also produced in some specific fish types, such as mackerel, as the storage time increases (Dulos et al., 2006). The volatile profile is a typical feature of food, and the volatile compounds difference between fresh and spoiled fish can be used as chemical fingerprints to reflect the relative compounds changes.

Colorimetric Strips for the Detection of Volatile Amine

More research has been focused on using several amine sensitive dyes to determine volatile amines (Rakow et al., 2005; Steiner et al., 2010; Kuchmenko et al.,

2011). Amine sensitive dyes are colorimetrically responsive to volatile bases produced during food spoilage and simultaneously change their color. Metalated tetraphenylporphyrins, pH indicators and highly solvatochromic dyes are three families of chemically responsive dyes, which have been used to determine biogenic amines

(Rakow et al., 2005). Solvatochromic dyes are a category of chemical compounds that change color depending on the polarity of the solvent dissolving the dye. The electronic structure of solvatochromic dyes usually contains a strong zwitterionic component, and electron donating and withdrawing groups are at the opposite ends of the molecule

(Reichardt, 1994). As the solvent polarity increases, a bathochromic shift occurs with positive solvatochromic dyes and a hypos-chromic shift occurs to negative solvatochromic dyes (Cartwright, 2016). The commonly reported class of solvatochromic dyes includes azobenzenes, thiazines, pyridinium N-phenolate betaine dyes, and merocyanines (Atwood et al., 2017).

Bromophenol blue is sulfonated hydroxy-functional triphenylmethane dye and is a commonly used pH indicator dye with a low pKa (Mills et al., 1995). As an acid-base indicator, bromophenol blue will lose a proton when the pH of the environment around this indicator is higher than the pKa of the dye. This displacement changes the electronic distribution within the molecule and the indicator changes its color from yellow to blue (Flores, 1978). Bromophenol blue has been known can react with basic amines

(Kuchmenko et al., 2011). Bromophenol blue was found to react with volatile biogenic amines produced by a cod sample and showed a dramatic color change from yellow to blue (Miller et al., 2006). Bromophenol blue also has been used as a color indicator to assess the freshness of guava. As the volatile compounds produced during developing of guava, pH in the package headspace decreased and bromophenol blue changed its color from blue to green (Kuswandi et al., 2012).

Rose bengal is a xanthene dye with photophysical properties and has played a significant role in photobiology and dye-sensitized oxygenation (Lamberts and Neckers,

1984). Rose bengal changes its color based on a protonation and deprotonation reaction. The dye is in the lactone form and transparent when exposed to a low pH environment; while it changes to its quinoid form with pink color as pH increases

(Akerlind et al., 2011; Schoolaert et al., 2016). A rapid colorimetric method was built by using rose bengal to quantify anhydrous caffeine and chlorphenoxamine hydrochloride simultaneously in a pharmaceutical (Amin et al., 1995). Rose bengal has been used to detect the existence of amines and ammonia based on the principle that amines have basicity and can neutralize rose bengal (Paczkowski et al., 1985). A monitoring tape using rose bengal as the indicator to determine the ammonia gas content in the air has

been developed (Nakano et al., 1994). Rose bengal was used as a xanthene dye in an indicator device to react with the volatile amines produced by biological agents in food.

The principle of this indicator device was that rose bengal would change its color from transparent to pink when exposed to the volatile amines in tested food (Miller et al.,

2006).

Filter paper can be used to absorb dye solution to make colorimetric strips. After the strips are exposed to the headspace of samples, dye on the strips will react with volatile biogenic amines and change color (Dole et al., 2016). A colorimeter is an instrument that measures color, and it is used to measure the indicator strips. A L*a*b* system is a cylindrical coordinate system and is used to quantify the color of strips. b* is the yellow/ blue coordinate and the more positive the b* value means more yellow hue is present; the more negative the b* value means more blue hue is present. a* value presents the red/green coordinate, which negative value means green and positive value means red. A series of standard cocktails with different concentrations can be used to quantify the volatile amines in tested samples.

In previous research, BPB strips were found to correlate volatile biogenic amine content with quality grades for mahi-mahi (Dole et al., 2016). However, the BPB strip method is a broad detection method for the class of volatile biogenic amines and as a result the BPB strips were not in total agreement with the results from ELISA, which measure a specific analyte, in this case histamine (Dole et al., 2016). The BPB strips developed in Phase I did not have a good uniformity of color change and the linearity of the standard curve from BPB method was low (Dole et al., 2016). The indicator device containing rose bengal, which used to detect the biogenic amines in food, was more

complicated and with a higher cost than colorimetric strips (Miller et al., 1999). Rapid assays should be developed and optimized to produce an accurate, sensitive, low cost and timesaving method to detect fish spoilage.

GC Methods to Determine the Aroma Profile of Seafood as Chemical Indicators of Spoilage

Gas Chromatography-Mass Spectrometry (GC-MS) is a highly effective analytical instrument to separate, identify and quantify volatile and semi-volatile chemicals from a complex food matrix (Lambropoulou et al., 2007; Sandra et al., 2003; Wang et al., 1999).

In gas chromatography, the mobile phase is the gas that moves through the column, and the stationary phase is a polymer film that coats the column filling or the column wall (Abraham et al., 1999). Compounds passing through the column have different strengths of interaction with the stationary phase. The compound having a larger interaction with the stationary phase needs a longer time to interact with the column and migrates through the column later than other compounds (Sneddon et al., 2007). Mass spectrometry can provide the “fingerprint” information of a molecule including its molecular weight, structure or elemental composition. This principle of this technique is that molecules in samples are converted into ions as in the gas phase with or without fragmentation and then are distinguished by their mass-to-charge ratio (m/z). The major application of GC-MS includes identification and quantification of food composition, food additives, aroma components, transformation products (Simko et al., 2002; Wishart,

2008; Bianchi et al., 2007). It can also detect a variety of contaminants, such as pesticides, packaging materials, and toxins (Tanaka et al., 2000).

Purge and trap method is a dynamic headspace extraction method commonly connected to GC-MS or gas chromatography-olfactometry (GC/O). The purge and trap

method was established and firstly applied in analytical techniques in the 1970s (Snow and Slack, 2002). For the purge and trap process, inert gas goes through the sample and volatile analytes are stripped from the sample. Volatile compounds are then re- focused on a trap and then thermally desorbed onto a GC. The development of this dynamic extraction method improves the detection levels of analytical instrumentation and provide accurate and precise analysis. Comparing with the static headspace method, purge and trap method has a lower limit of detection (LOD) value and can be more sensitive (Lucentini et al., 2005; Beltran et al., 2006). This dynamic extraction method was able to extract volatile compounds in a higher amount than solid-phase microextraction (SPME) (Povolo et al., 2003). A trap containing certain adsorbent resins is able to remove water from the volatile compounds introduced onto GC. A purge and trap method was equipped to GC/MS to concentrate volatile compounds of menhaden fish oil and twenty-nine compounds, including aldehydes, ketones, and carboxylic acids, were able detected (Hsieh et al., 1989). A Tenax column was used to trap volatile compounds after the fish sauce was purged for sixteen minutes and twenty-three compounds were determined by GC-MS (Fukami et al., 2002). Purge and trap is widely used in collecting and concentrating volatile components from fish meat, such as gilthead sea bream, pink salmon, sockeye salmon, and atlantic salmon (Girard et al.,

2000; Alexi et al., 2017; Jonsdottir et al., 2008).

Amine columns usually have low or mid polarity phases and are designed for determining amines and other specific basic chemical compounds without any complex derivatization procedure. Using amine columns in GC-MS can improve the response for the basic compounds and also prevent tailing of these analytes. Amine columns are

also able to separate the natural compounds having the oxygen groups easily influenced by hydrogen bonding. A SPME-GC-MS method with Rtx-Volatile amine column was able to determine trimethylamine (TMA) and dimethylformamide in the headspace above solid hexamethylene triperoxide diamine (HMTD) without a deactivation procedure (Steinkamp et al., 2016). Trimethylamine and dimethylsulfide in marine sediments were also analyzed by using a GC-MS method with Rtx-Volatile amine column, which was base-deactivated (Zhuang et al., 2017). GC with flame ionization detector (GC-FID) using cold on-column injection with an Rtx-5 amine column was able to identify and quantify putrescine and cadaverine in a standard solution without any derivatization procedure (Bonilla et al., 1997). A Rtx-5 amine column was also used in a GC system coupled to a nitrogen-phosphorus to detect the ephedrines in urine samples simultaneously (Eenoo et al., 2001). Volatile amines C1 to C9, including dimethylamine, trimethylamine, monoethylamine, isopropylamine and others, in standard solution was able separated and detected by GC-FID with Rtx-5 Amine column or PoraPLOT Amines column (Abalos et al., 2001).

Over the years, different samples extraction methods and column types have been used to detect spoilage of seafood products by GC-MS. Volatile compounds considered as spoilage indicator of cold smoked salmon were identified by using three different GC-MS methods. All these three GC-MS methods used purge and trap method as dynamic headspace collection and DB-5 MS column or DB-1701 column were used to separate volatile compounds (Jorgensen et al., 2001; Joffraud et al., 2001; Jonsdottir et al., 2008). A SPME-GC-MS method using a ZB-Wax column was developed to determine the storage influence on volatile compounds of fresh king salmon (Wierda et

al., 2006). GC-MS method equipped with CAR/PDMS fiber to extract volatile compounds and Rtx-WAX column was able to investigate the spoilage indicators of sea bream and prawn (Soncin et al., 2008). Effect of storage on sea bream was identified by a GC-MS method using a Tenax trap and a WCOT fused silica column and spoilage markers were determined (Alasalvar et al., 2005). Volatile compounds of European seabass produced during storage were able to be detected by using the GC-MS method with dynamic headspace extraction and RTX-5 column (Leduc et al., 2012). Freshness markers of whiting were detected using a GC-MS method containing CAR/PDMS fiber and a BPX5 capillary column (Duflos et al., 2010). Volatiles identified as spoilage markers of yellowfin tuna was also detected by an SPME-GC-MS method (Edirisinghe et al., 2007). The GC-MS methods mentioned above were able to detect alcohols, acids, aldehydes, alkanes, ketones, trimethylamine and sulfur compounds.

Amines in seafood, such as isobutylamine, 3-methylbutylamine, and 2- methylbutylamine, have been identified as spoilage markers of seafood product and give off fishy odor (Gill et al., 1983; Eskin, 2013; Mayr and Schieberle, 2012). However, these amines have not been able to be detected by using GC-MS without derivatization because short chain amines have high polarity, basic character, and high aqueous solubility. Preparation of amine derivatives is usually a necessary step to increase the volatility of compounds when using GC to analyze amines (Staruszkiewicz et al., 1981;

Rogers et al., 1997; Du et al., 2001). For example, amines including putrescine and cadaverine in salmon were able detected by using an SPME-GC-MS method with on- fiber derivatization procedures (Awan et al., 2008). A new simplified and accurate GC-

MS method should be developed to separate and determine amines and other spoilage markers of fish products without any complex derivatization step.

HPLC and UHPLC Methods for Biogenic Amine Detection

Ultra-high performance liquid chromatography (UHPLC) is another used instrument to measure amino acids and biogenic amines in seafood products. When the mobile phase passes through the column, components in mobile phase have varying strengths of interaction with the stationary phases. During an LC separation run, the composition of the mobile phase is often changed to alter the phase partitioning of each compound between the mobile phases and stationary phases. This is called gradient elution. The elution time of each compound is dependent on the relative strengths of its interaction with the mobile and stationary phase.

Reversed-phase columns (typically C18) are usually used as stationary phases and polar mobile phases, such as aqueous ammonium acetate, acetonitrile, formic acid, or mixtures of these solvents, are usually used for the separation of biogenic amines

(Erim 2013). For example, an HPLC with C18 column and fluorescence detection was used study biogenic amines in canned tuna fish, and mackerel (Peng et al., 2008). An

HPLC method with fluorescence derivatization used the C18 column to separate eight different amines, including histamine, putrescine, cadaverine and others, in wines

(Busto et al., 1997). HPLC with C18 column was used to detect seven biogenic amines in beer (Tang et al., 2009). A C18 column with 1.8 μm particles was applied in a UHPLC method and was able to separate putrescine, cadaverine, histamine, tyramine and other four biogenic amines in fish and chicken samples (Dadakova et al., 2009).

For fish and fish products, aqueous trichloroacetic acid (TCA) is a major solvent used to extract free amino acid and biogenic amines due to its good protein precipitation

capacity (Hwang et al., 1997). Research showed that five biogenic amines were able to be extracted from homogenized fish and fishery products by using a 5% TCA solution

(Shakila et al., 2001). Histamine, putrescine, tyramine, and spermidine were extracted from canned tuna sample by TCA solution, and then their concentrations in canned tuna were calculated (Zarei et al., 2011). Five free amino acids and six biogenic amines were extracted by TCA solution from tuna muscle tissue and quantified by HPLC (Ruiz-

Capillas et al., 2004). TCA solution was used to extract biogenic amines from fish tissue and eight biogenic amines, including putrescine, cadaverine, histamine, were able to be determined in this extraction solution (Sagratini et al., 2012).

A derivatization step is required for HPLC because most of the biogenic amines are lack of chromophore. Dansyl chloride and o-phthaldialdehyde (OPA) are two major derivatives used in HPLC methods (Malle et al. 1996; VecianaNogues et al., 1997;

Salazar et al., 2000). Amino groups of free amino acids and biogenic amines can react with dansyl chloride and stable derivatives with a chromophore can be formed. Due to their fluorescent characteristics, the dansyl derivatives can be determined using UV detection. Dansyl chloride was used for derivatize biogenic amines in the extraction solution of fish and fishery product (Shakila et al., 2001). Eight biogenic amines in several types of fish and fish products were derivatized by dansyl chloride and then identified by HPLC system equipped with fluorescence detector (Zhai et al., 2012). The reagent o-phthaldialdehyde can react with amines and free amino acids to produce fluorescent products. Histamine, putrescine, and cadaverine in tuna fish reacted with o- phthaldialdehyde and produced stable derivatives (Rossi et al., 2002). Free amino acids in fish, including lysine, histidine, and others, were reacted with o-phthaldialdehyde

(OPA) to be able detected by a fluorescence detector (Antoine et al., 1999). Nine biogenic amines in canned yellowfin tuna, including histamine, putrescine, and cadaverine, were derivatized by dansyl chloride in pre-column derivatization method and were derivatized by o-phthaldialdehyde (OPA) in a post-column derivatization method (Simat and Dalgaard, 2010).

The HPLC technique is sensitive, reproducible and the most useful for simultaneous detection of free amino acids and biogenic amines related to fish spoilage indicators by far (Onal et al., 2007). Ten biogenic amines in tuna considered as hygienic quality indicators were monitored by using HPLC method equipped with a C18 column and involving a post-column derivatization procedure (VecianaNogues et al., 1997).

Four free amino acids in mahi-mahi, which are considered as the precursors of biogenic amines, were reacted with o-phthaldialdehyde (OPA) and then analyzed by HPLC method with C18 column (Antoine et al., 2002). Dansylated amines in fish and fishery products were separated and detected by an HPLC method with a good linearity and sensitivity (Shakila et al., 2001). Eight biogenic amines as spoilage markers in fermented fish products reacted with dansyl chloride and were able to be determined by

HPLC (Kose et al., 2012). Thirteen amino acids in mountain trout reacted with 9- fluorenylmethyl chloroformate were detected, and quantified by HPLC system (Gunlu et al., 2014). The simultaneous detection of biogenic amines and amino acids can also be achieved by HPLC technique. Histamine, cadaverine, tryptamine, tyramine and their precursor amino acids were derivatized by dansyl chlorides and were able detected simultaneously by an HPLC-UV detection method (Mazzucco et al., 2010).

The UHPLC technique is an evolved instrument that can be operated under higher pressure than a regular HPLC system. This characteristic of UHPLC is that columns with smaller particle sizes are used than in regular HPLC. Due to the smaller particle size packed in the LC column, the UHPLC has better efficiency and resolution than regular HPLC. UHPLC has been used in biogenic amines detection to decrease analysis time. UHPLC was applied to determine seven biogenic amines in Bokbunja wines and produced good linearity for calibration curves of standards (Jia et al. 2012).

However, a UHPLC method that can detect the major free amino acids and biogenic amines simultaneously has not been developed. A new UHPLC method is needed to be able to investigate the content change of biogenic acids and their precursor amino acids during fish spoilage. Also, the spoilage effect on the major amino acids in fish should be studied to monitor the quality change of fish product.

ELISA Detection of Histamine

Enzyme-linked immunosorbent assay (ELISA) is a biochemical assay technique used to detect substances such as peptides and antibodies. An antigen (the specific antigen is usually proprietary knowledge) is immobilized on a solid surface and a specific antibody binds to the antigen. A substance is added (usually a chromophore) to give a detectable signal after the preceding reaction. ELISA has been used to measure histamine in food, and this method was based on a color-change reaction (Serrar et al.,

1995). The AOAC (No. 070703), The Neogen Veratox® test kit (Neogen Corp, Lansing,

MI) has been validated as a quantitative ELISA test to determine histamine in tuna.

The AOAC (No. 070703) method has good reproducibility, and can be rapidly, and sensitively performed in the research laboratory (Lupo et al., 2011). It is known that

ELISA test kits are good technology for disease outbreak investigatory studies. The

detection ranges for this AOAC (No. 070703) is up to 50 ppm and no false positive or negative result was found when using this kit to test a wide range of histamine standard solutions (Hungerford et al., 2012). Histamine content in several fish products, including bonito, salmon, mackerel, herring and others, were quantified by the Neogen Veratox®

ELISA kit and good recoveries of histamine were observed (Kose et al., 2011). The

AOAC (No. 070703) was also used to detect the histamine concentration in tuna and mahi-mahi samples and the ELISA results correlated with the sample spoilage grade.

However, using an ELISA kit in routine safety detection of large number samples in an industrial process can be expensive (Lehane and Olley, 2000). Also, the reagents used for ELISA kit are easily degraded and should be stored refrigerated at 2-8 ºC no longer than twelve months.

6,000,000

5,000,000

4,000,000

3,000,000

2,000,000

Global catches (tones) catches Global 1,000,000

0 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2014 2016 Year

Figure 2-1. Global catches of albacore, bigeye, skipjack and yellowfin data from 1960 to 2016. Data from WCPFC (2016).

CHAPTER 3 A RAPID UHPLC METHOD FOR THE SIMULTANEOUS DETERMINATION OF AMINO ACIDS AND BIOGENIC AMINES IN TUNA AND MAHI-MAHI

Digest

Tuna and mahi-mahi are two major fish species responsible for histamine poisoning in the United States. The purpose of this research was to develop a rapid

Ultra-High Performance Liquid Chromatography (UHPLC) method to identify and quantify amino acids, histamine and other biogenic amines that can act as co-indicators of histamine (scombroid) poisoning in tuna and mahi-mahi. In this reversed-phase

UHPLC method, amino acids and biogenic amines were extracted from homogenized mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) using aqueous 5% trichloroacetic acid (TCA) and were derivatized with dansyl chloride. The dansylated compounds were separated using a C18 reversed phase column with 1.3

µm particle size and then detected by an ultraviolet (UV) detector. The modified UHPLC method could determine ten amino acids and four biogenic amines simultaneously in mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) within 17.5 minutes. This UHPLC method showed good linear response, sensitivity, resolution, recovery, repeatability, and the number of theoretical plates. The UHPLC method developed in this study is a rapid and accurate method to monitor quality changes of mahi-mahi and tuna by inspecting the changes of amino acids and biogenic amines.

Background Information and Objectives

Fish spoilage is defined as any undesirable changes of fish that happen rapidly after fish landing and involves the breakdown of various organic compounds or formation of new molecules (Ashie et al., 1996; Ghaly et al., 2010). The major concern of fish spoilage includes lipid oxidation, protein degradation and the decline of other

nutritional components in fish (Clancy et al. 1995; Prester, 2011). Fish spoilage is considered as the main issue that causes fish loss and each year there are around 10 to 12 million tonnes of fish that are lost due to the spoilage (FAO, 2010). The autolysis of fish muscle proteins results in the formation of a large number of free amino acids and polypeptides, which are essential substrates or catalysts for reactions pertaining to fish spoilage (Ghaly et al., 2010). Studies also reported that several free amino acids were related to the flavor of fish (Fraser and Sumar, 1998; Ghaly et al., 2010; Ruiz-

Capillas and Moral, 2004). Arginine, glutamic acid, glycine, alanine, phenylalanine, isoleucine, leucine, lysine, histidine, and tyrosine have been identified as major free amino acids in tuna and mahi-mahi (Ruiz-Capillas and Moral, 2004; Sen, 2005; Chong,

2014).

Biogenic amines are organic bases and can be produced during the process of fish spoilage by microbial decarboxylation of free amino acids or by transamination of free amino acids (Zhai et al., 2012). Even though several biogenic amines are produced during fish spoilage, only histamine, cadaverine, and putrescine are identified as the chemical indicators of fish quality and safety (Bulushi et al., 2009). Histamine

(scombroid) poisoning, which is associated with the consumption of spoiled scombroid fish containing significant amounts of histamine, is identified as the highest incidence of illness from fish poisoning (Morrow et al. 1991). Histamine toxicity can be potentiated by cadaverine and putrescine due to their inhibiting ability to the intestinal histamine- metabolizing enzymes and diamine oxidase (Bulushi et al., 2009; Visciano et al., 2012).

Histamine, cadaverine, and putrescine can be synthesized by the decarboxylation of free histidine, lysine, and arginine, respectively (Prester, 2011). The formation of these

three amines in spoiled fish depends on the content of their endogenous precursor free amino acids, the presence of bacterial decarboxylase and the environmental conditions

(Visciano et al., 2012). Tuna and mahi-mahi are two major sources of histamine poisoning in the United States due to high levels of histidine existing in their muscle tissue (Ahmed, 1991; Bulushi et al., 2009).

High-performance liquid chromatography (HPLC) is a sensitive, reproducible instrument to identify and quantify free amino acids and biogenic amines associated with fish spoilage (Onal et al., 2007; VecianaNogues et al., 1997; Antoine et al., 2002;

Shakila et al., 2000; Kose et al., 2012; Gunlu et al., 2014; Mazzucco et al., 2010). Ultra- high performance liquid chromatography (UHPLC) is an evolved separation technology that has the same principle with HPLC that the compound affinity between mobile and stationary phases determine the separation of compounds on the column. The UHPLC instrument can be operated under very high pressures produced by a column with small particle size and has better efficiency and resolution than traditional HPLC. The UHPLC technology has been applied in simultaneous detection of biogenic amines and amino acids in fermented food products, such as wine and cheese, to reduce the elution time and improve the resolution (Jia et al., 2011). A UHPLC-MS/MS method was developed by He et al. (2016) to simultaneously determine nine biogenic amines and their precursors in cheese, red wine, and fish meat.

However, a rapid and simple UHPLC method that could simultaneously detect the major free amino acids and biogenic amines related to fish spoilage has not been developed. A new UHPLC method was needed to be able to simultaneously investigate the content change of biogenic acids and the major amino acids in fish products.

Materials and Methods

Fish Samples and Preparation

The mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) analyzed in this research were caught commercially from South Pacific waters. More than five sensory experts in the Food and Drug Administration (FDA) and National

Marine Fisheries Service (NMFS) applied the sensory grading system provided in the

FDA Office of Regulatory Affairs (ORA) Laboratory Manual (FDA, 2013) to evaluate the fish filets of mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) into seven grades. The grading system used by FDA/NMFS experts depended on olfaction and were graded 1 to 7 to represent their quality. Grade 1 represented high quality, while grade 7 represented very poor quality fish.

The individually packaged and graded fish filets from FDA/NMFS were shipped overnight, received frozen on dry ice, and then stored in a -20 ºC freezer until analysis was performed. For each grade of fish samples, vacuum packaged frozen samples were defrosted overnight at room temperature, and then were chopped and homogenized by a blender (Total Blend Classic, Blendtec, Orem, UT) to perform chemical analysis.

Standards and Reagents

All chemicals used in this study were of analytical grade or higher. Amino acid mix standards, L-Alanine, L-Arginine, L-Glutamic acid, Glycine, L-Isoleucine, L-Leucine,

L-Lysine, L-Phenylalanine, L-Histidine, L-Tyrosine, histamine, cadaverine, sodium bicarbonate, and formic acid solution were supplied by Sigma–Aldrich (St. Louis, MO).

Putrescine was purchased from MP Biomedicals (Santa Ana, CA). Dimethylamine, dansyl chloride were obtained from ACROS Organics (Geel, Belgium). Sodium

hydroxide, HPLC-grade acetonitrile, HPLC-grade water, ammonium hydroxide, trichloroacetic acid, and hydrochloric acid were supplied by Fisher Chemical (Pittsburgh,

PA).

Individual free amino acid and biogenic amine solutions were prepared separately by dissolving reagent into 0.1M HCl. Then these individual standard solutions were diluted at various levels for separation of the compounds on the UHPLC for determination of retention time.

Individual free amino acids were 2.5 μmoles/L in 0.1M HCl in the purchased amino acids mix standard solution (Sigma–Aldrich, St. Louis, MO). The amino acid mix standard was diluted to achieve an approximately 200 ppm stock solution (amino acid quantities differed because of molarity to ppm conversion). The stock biogenic amine cocktail was 600 mg/L for each amine and was prepared by dissolving cadaverine, putrescine, histamine, and dimethylamine in 0.1M HCl. The stock solutions were then diluted to produce concentrations suitable for UHPLC analysis.

For quantitation, external calibration curves were utilized. Five different concentrations of biogenic amines cocktail containing cadaverine, putrescine, histamine were used: 5, 10, 50, 100, 200 mg/L. Five different concentration levels of the amino acids mix standard with dimethylamine (5, 10, 50, 100, 200 mg/L) were used for quantitation.

Extraction and Derivatization

The extraction procedure of Zhai et al. (2012) with modifications was used in this study. In brief, 3 g homogenized fish sample was added in a 15 mL centrifuge tube with

10 mL of 5% trichloroacetic acid (TCA) and then was vortexed for 15min. The centrifuge tube spun at 5000 g at 4 ºC for ten minutes. After the extract was removed, the

remaining solid was extracted using 10 mL of 5% TCA again by the same procedures as above, and the supernatant was collected. Both supernatants were combined and passed through a Whatman No. 1 filter paper.

The derivatization procedure of Simat et al. (2011) with modifications was applied. An amount of 300 μL 2 mol/L NaOH solution and 300 μL of saturated NaHCO3 solution were added to 1 mL of filtered fish extract or 1 mL of standard solution. Dansyl chloride was dissolved in acetone to archive 1% (w/v) concentration, and 2 mL of this solution was added the resulting solution and then was protected from light and incubated for 60 min at 40 ºC in a water bath (Isotemp 220, Fischer Scientific,

Pittsburgh, PA). Excess dansyl chloride was removed by adding 120 μL of an NH4OH solution (4 mol/L) and then the solution was stored away from light for 1 hour. The solution was collected and filtered by a 0.2 μm PTFE membrane (Phenomenex,

Torrance, CA) before LC injection. For each grade of mahi-mahi or tuna, the extraction and derivatization procedure were performed in triplicate.

Determination of Biogenic Amines

Quantification of the amino acids and biogenic amines was carried out by using an Agilent 1290 Infinity Series UHPLC System. A Kinetex® 1.3 µm C18 Column, 50 x

2.1 mm was used for separation. The mobile phase for this UHPLC method was water with 0.1% formic acid(A) and acetonitrile with 0.1% formic acid(B). The flow rate was 0.5 mL/min and the column temperature was 30 ºC. The elution gradient was: 0 min 5% B;

1 min 5% B; 5 min 41% B; 11 min 68% B; 14 min 95% B; 14.5 min 95% B; 14.6 min 5%

B; 17.5 min 5% B. The injection volume was 10 μl. The analytes were detected at a wavelength of 254 nm.

Method Validation

The sensitivity of this optimized UHPLC method was investigated by determining the limit of detection (LOD) and the limit of quantification (LOQ). For the peak of each amino acid and biogenic amine, the calculated signal to noise ratio was determined by the software of this UHPLC instrument. The LOD value was 3 times the calculated signal to noise ratio and was converted to concentration units by using the concentration of the compound. The calculation of LOQ was similar to LOD and was calculated as 10 times of the signal to noise ratio.

The percentage recoveries of cadaverine, putrescine, and histamine were determined by spiking biogenic amine cocktails in triplicate into homogenized grade 1 of mahi-mahi and tuna samples verified to contain biogenic amines below detable levels, to achieve the final concentration of each amine as 50 mg/kg fish. Then the spiked fish samples were extracted, derivatized, and the detected biogenic amines levels were compared with the theoretical amount.

To determine the repeatability of this UHPLC method, six injections of a single fish extraction with 10 mg/kg of each free amino acid and biogenic amine were compared, and the percentage relative standard deviation (%) for each amine and free amino acids were reported.

The resolution (R) and the number of theoretical plates for histamine, cadaverine, and putrescine were calculate using the formulas provided by Moldoveanu and David

(2017).

Histamine ELISA Test Kit

An AOAC-validated ELISA test kit supplied from Neogen Corporation (Lansing,

MI) was also used to quantify histamine levels in mahi-mahi (Coryphaena hippurus) and

yellowfin tuna (Thunnus albacares) in this study. Fish sample preparations, dilutions, and test kit procedures followed the kit instructions. The standard provided by the

ELISA kit and histamine standards prepared in the lab were used to build calibration curves separately to quantify histamine levels in fish samples. For each grade of mahi- mahi or tuna, the ELISA test was performed in triplicate.

Results and Discussion

Method Development

To obtain a rapid, sensitive UHPLC method with good resolution, different extraction procedures, components of the mobile phase, elution gradients, column temperatures and the wavelengths of the detector were compared. Figure 3-1 shows the chromatographic separations of the fourteen dansylated compounds in a mahi-mahi grade 1 sample spiked with 10 mg/kg of each amino acid and biogenic amine standard.

The reversed-phase UHPLC method developed in this study was able to separate ten amino acids and four biogenic amines from tuna and mahi-mahi samples within 17.5 minutes.

The HPLC instrument has been reported as a sensitive and accurate tool for the determination of free amino acids in mahi-mahi (Coryphaena hippurus), bigeye tuna

(Parathunnas mebachi) and flounder (common flounder) (Antoine et al., 1999; Ruiz-

Capillas and Moral, 2004). Detection of derivatized biogenic amines in fish products by using HPLC or UHPLC with various detectors have been documented (Veciana-Nogues et al., 1997; Shakila et al., 2001; Kose et al., 2012; Simat et al., 2011). He et al. (2016) reported a UHPLC-MS/MS method that could simultaneously extract and analyze tyramine, histamine, tryptamine, putrescine, agmatine, spermidine, cadaverine, spermine, phenylethylamine and their precursor amino acids in red wine, cheese, and

fish. Among the biogenic amines detected in the study conducted by He et al. (2016), only histamine, putrescine and cadaverine were considered as fish spoilage indicators and also other major amino acids in fish were not determined in this method. The reversed phase UHPLC method developed in our study focused on rapidly separating and determining the major amino acids in fish and the biogenic amines as fish spoilage indicators without the use of costly tandem mass spectrometry. Leucine and dimethylamine (DMA) could not be separated and quantified by this modified UHPLC method. Due to its high-volatility, DMA could be determined by other analytical instruments such as gas chromatography-mass spectrometry (Chan et al., 2006).

LC columns with small particle size (1.3 μm) used in this reversed-phase UHPLC method improved resolution, prevented peak broadening and reduced the elution time.

The particle size of the LC column is of primary importance when choosing a stationary phase and decreasing particle size can increase separation efficiency (Tuzimski et al.,

2015). Several studies have focused on optimizing chromatographic separation of biogenic amines by reducing the particle size of LC column. Previously, Simat and

Dalgaard (2011) reduced the particle size of the C18 column from 5 µm to 1.8 µm in a pre-column derivatization HPLC method and was able to reduce the elution time used for separating nine amines in seafood from 29 minutes to 12 minutes. Jia et al. (2011) reduced the particle size of the C18 column to 1.7 µm in an LC-Q-TOFMS method to lower the elution time for separating 23 amino acids and 7 biogenic amines in beer, cheese, and sausage. The method reported by Cai et al. (2016) used a C18 column with 1.8 µm particle size in HPLC–MS/MS to separate five isotope-coded derivatized biogenic amines in rice wine within 8 minutes. However, column back pressure is

inversely proportional to the square of the column particle diameter (Harris et al., 2017).

Due to that the UHPLC instrument can be operated under higher pressure, a UHPLC instrument is needed when using LC column with very small particle size.

Linearity and Sensitivity

For this developed UHPLC method, the linearity of each amino acid and amines was studied by using the calibration curve obtained from 5 different concentration levels. Calibration curves of the dansylated amino acids and amines showed excellent linearity with a coefficient of determination (r2) higher than 0.99. Glutamic acid was one exception that the r2 value for its standard curve was 0.9504. All calibration curves obtained by this UHPLC method had wide linear concentration ranges (Table 3-1).

The UHPLC method modified in this study showed appropriate sensitivity. The

LOD value refers to the smallest quantity or concentration of a compound obtained from the tested sample that is able to be detected by a specific analytical method. The limit of quantitation (LOQ) refers to the level of the analyte in a sample where quantitative results with a high degree of accuracy are achievable. The LOD values for histamine, cadaverine, and putrescine were between 0.0164 to 0.0196 mg/kg and for the nine determined amino acids were 0.0217 to 0.2850 mg/kg. The corresponding LOQ values for these three biogenic amines were between 0.0546 to 0.0653 mg/kg and for the nine amino acids were 0.0723 to 0.9500 mg/kg. This modified UHPLC method was more sensitive than other HPLC and UHPLC methods used for determining biogenic amines in fish products (Sagratini et al., 2012; Tahmouzi et al., 2011; Simat et al., 2011).

Recovery and Repeatability

The accuracy of this UHPLC method was determined by the spike recovery method. To calculate the percentage recoveries of histamine, cadaverine, and

putrescine, a known amount of biogenic amines mixtures was spiked into the homogenized tuna and mahi-mahi samples before the extraction procedure and then the percentage of the theoretical amount was determined. This UHPLC method had satisfactory recoveries for cadaverine, putrescine, and histamine in mahi-mahi and tuna samples. The percentage recoveries for these three biogenic amines were in the range of 95.19 to 110.4% in mahi-mahi and were 115.95 to 141.26% in tuna samples (Table

3-2). The percentage recoveries of each biogenic amine in tuna were higher than those in mahi-mahi. These results showed that the matrix of different types fish influenced the recoveries of biogenic amines. This was consistent with the results published by Simat and Dalgaard (2011). The percentage recoveries of cadaverine, putrescine, and histamine were 95 to 114% in lean tuna and were 97 to 111% in fatty herring when determined by four different HPLC methods (Simat and Dalgaard, 2011). A UHPLC method developed by Redruellon et al. (2013) used to simultaneously quantify amino acids and biogenic amines in cheese also had the similar percentage recoveries of cadaverine, putrescine, and histamine which were 106 to 140%. The percentage recoveries of biogenic amines greater than 100% were likely due to the lower instrument responses for the external calibration curves than the theoretical value reducing the slope of calibration curves and making the calculated concentrations of biogenic amines larger than their theoretical value.

The precision was determined by six UHPLC injections of a mahi-mahi grade 1 sample spiked with a known amount of amino acids and biogenic amines. For the separated amino acids and biogenic amines, the relative standard deviation (RSD) of the instrumental repeatability ranged from 0.2272 to 3.9039% and were shown in Table

3-1. As the RSD values of the instrument repeatability below than 5%, the precision of this UHPLC method was acceptable (Taverniers et al., 2004).

Resolution and Theoretical Plates

Excellent resolution between putrescine, cadaverine, and histamine were obtained by reducing the particle size of the C18 column, evaluating different elution programs, and optimizing the column temperature (Table 3-3). A resolution value of greater than 1.5 ensures that sample components are well separated (Corradini et al.,

2011). The number of theoretical plates exceeding 200,000 was achieved using the C18 column with 1.3 μm particles (Table 3-3), and this modified UHPLC method had good elution efficiency. A reduction in the particle size of the column was the main factor that increased the number of theoretical plates (Hayes et al., 2014).

Application to Different Spoilage Grade of Mahi-Mahi (Coryphaena hippurus) and Yellowfin Tuna (Thunnus albacares)

The concentrations of free amino acids and biogenic amines in tested mahi-mahi

(Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) are shown in Table 3-4 and Table 3-5, respectively. Similar levels of tested amino acids and biogenic amines in tuna and mahi-mahi were reported by several references and the ranges from references were as follows: 12.5 to 139 mg/kg for arginine, 32 to 1400mg/kg for glutamic acid, 0 to 840 mg/kg for glycine, 0 to 415 mg/kg for alanine, 16.5 to 157 mg/kg for phenylalanine, 0 to 930 mg/kg for isoleucine, 0 to 716 mg/kg for lysine, 530 to 7080 mg/kg for histidine, 0 to 124mg/kg for putrescine, 0 to 930 mg/kg for cadaverine, 0 to

4500 mg/kg for histamine, 0 to 113 mg/kg for tyrosine. (Ruiz-Capillas and Moral, 2004;

Antoine et al., 1999; Rossi et al., 2002; Antoine et al., 2002; Stockemer, 1982). The fish filets of mahi-mahi and tuna were graded by trained sensory panels in FDA and

National Marine Fisheries Service (NMFS) using seven grades to evaluate the fish quality. Table 3-4 and Table 3-5 showed that the concentrations of histamine, cadaverine, and putrescine were generally lower in higher quality samples for both mahi-mahi and tuna (eg. Grade 1 and 2).

For mahi-mahi samples, histamine was only detected in grade 5 and grade 6 samples. The histamine concentrations for these two grades of mahi-mahi samples were above 2000 mg/kg. Grade 7 mahi-mahi was evaluated with lower quality than grade 5 and 6 mahi-mahi based on the sensory grading system. However, the histamine level of grade 7 mahi-mahi was anomalous that histamine was not detected in this grade of mahi-mahi by using the modified UHPLC method. The histamine concentration detected by the ELISA test kit of mahi-mahi grade 5 to 7 samples (Table

3-6) is consistent with the UHPLC results in that grade 7 by ELISA only showed a very low level of histamine. Dole et al. (2016) reported similar results that grade 7 of tuna contained the lower amount of histamine than grade 5 and grade 6 tuna samples. This result might because the tested fish filets were graded based on olfaction and the formation rate of histamine was different than the formation rate of unpleasant odors in fish. For tuna samples, the histamine level was a maximum of 3696 mg/kg in grade 6 tuna, then followed by 2922 mg/kg in grade 7 tuna, 696 mg/kg in grade 4, and 209.05 mg/kg in grade 5. No histamine was detected in grade 1 to 3 of tuna samples by this

UHPLC method. An ELISA histamine test kit agreed with the histamine levels in fish samples determined by the UHPLC (Table 3-6).

Histamine standards obtained from different commercial brands can influence the

ELISA test kit response due to the different purity and stability of reagent (Crowther,

2000; Shi, 2010). To be consistent with the developed UHPLC method, the same histamine standard reagent used to produce the standard curve in UHPLC method was also prepared in dilutions to set up a calibration curve in the ELISA method. Histamine amount in fish samples calculated by using the histamine standard provided in the kit, and the standard prepared in the lab were shown in Table 3-6. For each concentration of histamine standard curve, the ELISA response for histamine standard prepared in lab was higher than that for the standard provided by ELISA kit. The higher ELISA kit response for the prepared standard curve led to the smaller calculated histamine level in fish samples (Table 3-6). Histamine contents in mahi-mahi and tuna determined by

ELISA kit using these two different standard curves significantly correlated with histamine results from UHPLC (Table 3-7). Histamine amounts in fish calculated by standard curve prepared in the lab were more consistent with results from UHPLC

(Table 3-4, 3-5 and 3-6).

Among the free amino acids quantified in this UHPLC method, histidine levels in all tested fish samples were higher than the other free amino acids. The levels of histidine were detected as 654 to 2941 mg/kg in mahi-mahi and 1523 to 5837 mg/kg in tuna (Table 3-4 and Table 3-5). Histidine levels in good quality mahi-mahi (Grade 2, 3,

4) were at least 1000 mg/kg higher than histidine levels in low-quality mahi-mahi. The histidine level in grade 1 mahi-mahi was anomalous and was lower than those of any other grades. Similar result was reported by Antoine et al. (2002) that the level of histidine in mahi-mahi (Coryphaena hippurus) after storage at 7 °C for 2 days was lower than the histidine level in mahi-mahi after four-day or six-day storage at 7 °C. For tuna

samples, in general the histidine levels decreased as the histamine levels detected in tuna increased.

In general, the level of histamine in good quality fish, such as the freshly caught snoek, swordfish, tuna and yellowtail, is below than 10 mg/kg (Auerswald et al., 2006).

Histamine in raw fish is formed by free histidine under the reaction of the bacterial histamine decarboxylase during fish spoilage (Prester 2011; Lehane and Olley, 2000).

Gram-negative enteric bacteria naturally exist as the microflora of live fish and are responsible for histamine production during fish spoilage (Hungerford, 2010; Ghaly et al.,2010). Tuna and mahi-mahi are considered as two major sources of scombroid poisoning due to their muscle tissue containing high levels of free histidine, which is the precursor of histamine (Ashie et al., 1996; Prester 2011). Antoine et al. (1999) detected histidine levels as 1829 to 5415 mg/kg in mahi-mahi (Coryphaena hippurus) and 2209 to 7080 mg/kg in bigeye tuna (Parathunnas mebachi), and these levels were significantly higher than histidine levels in flounder, which was as low as 90 to 150 mg/kg. Antoine et al. (2002) reported that histidine levels declined from 4000 mg/kg to

1800 mg/kg with increasing levels of histamine from 0 to 1600 mg/kg in mahi-mahi after twelve-day storage at 7 °C. Ruiz-Capillas and Moral (2004) observed the correlation between free histidine and histamine in tuna that histidine levels decreased as well as the histamine levels increased in white muscle of bigeye tuna (Thunnus obesus) after storage in a controlled environment for 33 days.

Putrescine and cadaverine were detected in all grades of mahi-mahi in this study and were in the concentration range of 1.51 to 29.03 mg/kg and 2.11 to 253 mg/kg, respectively. As the grade of mahi-mahi increased, the levels of putrescine and

cadaverine in fish filets generally increased (Table 3-4). For all grades of tuna, putrescine was only detected in grade 4 and 6 of tuna with levels as low as 0.03 and

0.47 mg/kg, respectively. The cadaverine levels in the different tuna grades ranked in the same order as the ranking of histamine levels. The maximum cadaverine level was

343 mg/kg in grade 6, then followed by 140 mg/kg in grade 7 tuna, 107 mg/kg in grade

4 tuna, and 73 mg/kg in grade 5 tuna. Cadaverine was also detected as 2.72 mg/kg in grade 1 of tuna.

During fish spoilage, putrescine can be formed in two routes of synthesis, but both start from arginine. Putrescine can be formed from arginine by arginine deiminase, ornithine carabamoyl-transferase and ornithine decarboxylase. Putrescine can also be indirectly synthesized from arginine by arginine decarboxylase via agmatine (Prester,

2011; Wunderlichova et al., 2014). The microbial formation of putrescine in fish was likely influenced by factors including microbial activity, temperature, food matrix and the amount of precursor (Wunderlichova et al., 2014). Levels of arginine, which is the precursor of putrescine, were slightly higher in tuna samples than in mahi-mahi (Table

3-4 and Table 3-5). The levels of putrescine in poor grade tuna were lower than those in mahi-mahi possibly due to weaker microbial activity and a different food matrix

(Wunderlichova et al., 2014). Putrescine in fish tissue is mainly generated by the genus

Staphylococcus (Wunderlichova et al., 2014). The different compositions of free amino acids and peptides, which are the important nutrient source for bacteria, in mahi-mahi and tuna and other factors of the fish, such as levels of contamination, pH, sodium chloride content, water activity, influence the growth of the genus Staphylococcus may

have resulted in different levels of putrescine between tuna and mahi-mahi (Chirife and

Buera,1996).

For both tested mahi-mahi and tuna, lysine levels generally declined with increasing cadaverine levels in fish (Table 3-4 and Table 3-5). Cadaverine can be synthesized from lysine by lysine decarboxylase and various species of bacteria in fish have been identified producing lysine decarboxylase (Mayr and Schieberle, 2012; Onal,

2007; Hungerford, 2010, Ababouch et al., 1991).

Even though the formations of several biogenic amines have been observed during fish spoilage, only putrescine, cadaverine, and histamine are considered as significant markers of fish quality and safety due to their ability to show the microbial growth and reactions during storage (Bulushi et al., 2009). Histamine is the major natural chemical responsible for scombroid poisoning and can lead to symptoms such as oral numbness, headache, dizziness, palpitations and other allergy-like symptoms

(Bulushi et al., 2009). Studies have reported that putrescine and cadaverine, which are produced by decarboxylase activities of bacteria during fish spoilage, are able to potentiate histamine toxicity by inhibiting the intestinal histamine-metabolizing enzymes and diamine oxidase (Bulushi et al., 2009; Visciano et al., 2012). Other authors reported similar results that the levels of histamine, putrescine, and cadaverine generally were higher in the lower quality fish (Prester et al., 2009; Mackie et al., 1997;

Ozogul et al., 2006; Visciano et al., 2007; VecianaNogues et al., 1997). In addition, for each poor grade mahi-mahi and tuna tested sample, histamine was highest, then followed by cadaverine, and putrescine. These differences might be due to the variability of precursor free amino acids concentration in fish (Table 3-4 and Table 3-5).

Antoine et al. (2002) reported a similar ranking that the levels of putrescine, cadaverine, and histamine in mahi-mahi (Coryphaena hippum) increased to 30 mg/kg, 150 mg/kg and 1600 mg/kg respectively after twelve-day storage at 7 °C. Results reported by

Rossi et al. (2002) showed that putrescine, cadaverine, and histamine accumulated to

60 mg/kg, 649 mg/kg and 1533 mg/kg in Skipjack tuna after storage at 21 °C for 48 hours.

Tyrosine can produce tyramine by bacterial decarboxylase activities during fish spoilage (Prester, 2011) and this reason can be used to explain the decrease of tyrosine in the lower quality mahi-mahi and tuna samples. Ababouch et al. (1991) reported tyrosine levels declined in sardine (Sardina pilchardus) after storage in ice for 8 days. Tyrosine levels in grade 5 and 6 mahi-mahi, and grade 6 and 7 tuna samples were unable to be determined because histamine co-eluted with tyrosine when histamine levels were above 2000 mg/kg. The FDA (2005) released guidance that the levels of histamine in mahi-mahi and tuna above 50 mg/kg represented the decomposition in these fish and above 500 mg/kg should be toxic to humans. The histamine level, from which the co-elution started, was greatly above the toxic level for histamine for mahi-mahi and tuna fish. Therefore, this method remains valid because it is able to separate and detect biogenic amines over the useful concentration range, and co-elution only becomes an issue at clearly unacceptable levels of spoilage.

Isoleucine and phenylalanine are precursors of 2-methylbutylamine and phenylethylamine, respectively (Mayr and Schieberle, 2012). The correlations between these two free amino acids with the quality of mahi-mahi and tuna were not observed in this study. Similar results were reported by Ababouch et al. (1991) and Ruiz-Capillas

and Moral (2004). Studies have shown that some free amino acids, such as alanine, glycine, and glutamic acid are associated with fish flavor and the loss of these amino acids should be monitored due to their impact on fish aroma (Ghaly, 2010; Ruiz-Capillas and Moral, 2004). From the result of this study, the changing of these three free amino acids levels in tested mahi-mahi and tuna did not have a clear correlation with the fish quality grades.

Summary

The reversed-phase UHPLC method modified in this study using column particles of 1.3 μm was able to determine fourteen dansylated amino acids and biogenic amines in mahi-mahi and tuna simultaneously within 17.5 minutes. This developed

UHPLC method provided satisfying linear responses, sensitivity, resolution, percentage recoveries, repeatability, and the number of theoretical plates. This UHPLC method could be a valuable method to monitor the changes of amino acids and biogenic amines in tuna and mahi-mahi.

Table 3-1. Linearity for different amino acids and amines analyzed by UHPLC Amino acids and Regression Linear range Coefficient of %RSD LOD (PPM) LOQ (PPM) Amines equation (μg/mL) determination (r2) Arginine y=6.257x+33.11 0.285-287.4 0.9979 1.645 0.285 0.95 Glutamic acid y=11.72x+496.6 0.0385-242.8 0.9504 0.2296 0.0385 0.1284 Glycine y=8.923x+0.1631 0.2445-123.8 1 0.3619 0.2445 0.8151 Alanine y=21.06x-3.774 0.2166-146.9 1 0.5871 0.2166 0.7219 Phenylalanine y=16.87x-75.57 0.0647-272.6 0.9980 1.374 0.0647 0.2156 Isoleucine y=13.65x-99.37 0.0768-216.4 0.9967 0.4278 0.0768 0.256 Lysine y=39.28x+83.06 0.02169-241.2 0.9989 0.2596 0.02169 0.07231 Histidine y=31.64x+170.2 0.02176-256 0.9994 3.904 0.02176 0.07253 Putrescine y=65.26x+190.2 0.01638-200 0.9987 0.3505 0.01638 0.0546 Cadaverine y=59.29x+198.8 0.01717-200 0.9989 1.924 0.01717 0.05722 Histamine y=52.39x+93.88 0.01959-200 0.9993 0.2272 0.01959 0.06529 Tyrosine y=33.72x+95.06 0.02147-299 0.9981 0.9432 0.02147 0.07156

Table 3-2. Percentage recoveries of three biogenic amines Compound Name mahi-mahi tuna Putrescine 95.19±1.12 124.2±4.35 Cadaverine 91.57±3.18 116±3.32 Histamine 110.4±4.48 141.3±4.87

Table 3-3. The number of theoretical plates, and resolution of biogenic amines Putrescine Cadaverine Histamine N 214500 231300 316200 Resolution 7.624 (putrescine/cadaverine) 3.014 (cadaverine/histamine) ------

Table 3-4. Amino acids and biogenic amines (mg/kg) in seven grades of mahi-mahi (M) Compound M1 M2 M3 M4 M5 M6 M7 name Arginine 97.48±16.3 96.63±28.2 135.7±19 128.8±23.2 123.6±7.5 169±7.45 108.3±10.5 Glutamic acid 312.5±73.7 1646 ±92.2 1633±98.1 1812±130 226.9±77.7 220.9±35.4 499.5±53.6 Glycine 46.44±23 321.3±17.5 338.1±73.7 226.3±27.5 361.3±154 115.7±8.51 190.7±3.43 Alanine 186.9±23.6 189±6.92 132.5±6.69 186.2±39.7 341.1±14.9 286.4±24.1 296±2.02 Phenylalanine 31.18±1.4 31.56±1.6 31.32±2.18 28.17±0.74 30.21±3.73 36.91±2.15 24.5±0.596 Isoleucine 171.8±6.96 397.8±27.4 292.3±10.3 372±11 152.5±1.62 157.6±7.54 102.8±1.04 Lysine 23.11±15 93.45±3.87 41.15±9.93 63±6.71 50.38±8.23 17.47±1.35 13.4±1.22 Histidine 654.5±78.9 2656±140 2941±212 2703±97.8 1503±135 14500±123 1192±21.8 Putrescine 1.515±0.443 4.17±0.964 5.47±1.08 6.454±1.08 8.375±1.56 29.03±2.91 29.01±0.835 Cadaverine 2.113±0.471 20.82±1.99 17.74±1.48 16.59±0.94 124.7±1.3 233±12.6 253.2±7.42 Histamine n.d. n.d. n.d. n.d. 2183±57* 2310±71.8* n.d. Tyrosine 44.22±3.83 31.64±2.43 23.92±2.67 25.98±1.32 * * 9.253±0.75 *Tyrosine in mahi-mahi 5 and 6 co-eluted with histamine, thus the results are semi-quantitative n.d. indicates not detected

Table 3-5. Amino acids and biogenic amines (mg/kg) in seven grades of tuna (T) Compound T1 T2 T3 T4 T5 T6 T7 name Arginine 138.2±25.5 179.8±8.9 310.7±81.1 168.6±8.35 164.4±7.52 212.7±79.3 263.4±37.5 Glutamic acid 1986±467 2045±135 1727±56.5 2090±417 2321±283 1537±72.9 2041±19.4 Glycine 34.92±0.776 33.47±1.86 32.61±2.49 39.66±1.4 37.23±1.36 48.76±1.63 33.09±3.11 Alanine 97.85±50 163.1±9.21 143.5±1.41 115.8±66.9 183.1±80.1 288.6±8.19 239.7±4.23 Phenylalanine 43.36±3.2 48.99±3.32 45.96±0.598 86.49±29.5 56.35±5.65 84.5±6.27 74.28±2.43 Isoleucine 284.6±104 184.6±25.5 373.5±6.59 452.3±184 307.8±32.7 275.9±47.2 358.8±39.6 Lysine 51.07±22.3 98.56±12.7 140.2±15.5 37.18±6.98 29±23.7 3.59±1.9 6.56±7.16 Histidine 5837±410 3279±758 4782±132 4893±234 3903 ±116 1523±44.9 2902±59.1 Putrescine n.d. n.d. n.d. 0.0298±0.0516 n.d. 0.4689±0.812 n.d. Cadaverine 2.72±0.09 n.d. n.d. 107.5±17.8 73.73±2.14 343.0±3.81 140.9±4.86 Histamine n.d. n.d. n.d. 696.4±172 209.1±18.2 3696 ±41.7* 2922±61.7* Tyrosine 23.88±22.1 47.79±8.46 16.54±3.71 n.d. n.d. * * *Tyrosine in tuna 6 and 7 co-eluted with histamine, thus the results are semi-quantitative n.d. indicates not detected

Table 3-6. ELISA results of mahi-mahi (M) and tuna (T) calculated by using the standard provided in kit and standard prepared in lab. Grade Histamine (ppm) - standard Histamine (ppm) - standard provided in kit prepared in lab M1 0.466±0.264 0.158±0.104 M2 8.324±1.25 4.767±0.833 M3 2.236±0.77 1.126±0.45 M4 3.184±0.23 1.687±0.141 M5 2822±280 2122±234 M6 3256±135 2487±114 M7 11.68±1.18 7.961±0.89 T1 1.249±0.966 0.5407±0.494 T2 0.9753±0.45 0.3863±0.216 T3 0.6467±0.179 0.4797±0.135 T4 785.1±108 637.8±89 T5 254.5±14.34 204.7±11.79 T6 4371±187 3406±158 T7 3205±66.5 2435±54.6

Table 3-7. Pearson correlation coefficients (r) between ELISA results of mahi-mahi (M) and tuna (T) calculated by using the standard provided in the kit, standard prepared in lab and histamine results from UHPLC method. Histamine results from UHPLC mahi-mahi tuna Histamine from ELISA - 0.9958 0.9985 standard provided in kit Histamine from ELISA - 0.9948 0.9976 standard prepared in lab

Figure 3-1. Chromatographic separations of Mahi-mahi grade 1 sample spiked with 10ppm of each amino acid and biogenic amine standards. The dansylated compounds are: (1) L-Arginine; (2) L-Glutamic acid; (3) Glycine; (4) L-Alanine; (5) L-Phenylalanine; (6) L-Isoleucine; (7) L-Leucine; (8) Dimethylamine; (9) L- Lysine; (10) Histidine; (11) Putrescine; (12) Cadaverine; (13) Histamine; (14) L-Tyrosine.

CHAPTER 4 AROMA PROFILE CHARACTERIZATION OF MAHI-MAHI AND TUNA FOR DETERMINING SPOILAGE USING PURGE AND TRAP GAS CHROMATOGRAPHY- MASS SPECTROMETRY (PT-GC-MS)

Digest

Alcohols, aldehydes, ketones, amines, and sulfur compounds are essential aroma compounds identified in spoiled fish products. Gas chromatography-mass spectrometry (GC-MS) is an instrument that is widely used to identify and quantify volatile and semi-volatile compounds in fish products. In this research, a simplified and accurate GC-MS method was developed to determine the aroma profile of mahi-mahi and tuna for chemical indicators of spoilage. In the developed GC-MS method, trichloroacetic acid (TCA) solution was used to extract analytes from homogenized fish samples. The purge and trap system was used for sample introduction, and the GC-MS with RTX-Volatile Amine column was able to determine compounds without a derivatization procedure. The created purge and trap gas chromatography-mass spectrometry (PT-GC-MS) method could identify and quantify twenty aroma compounds in mahi-mahi (Coryphaena hippurus) and sixteen volatile compounds in yellowfin tuna

(Thunnus albacares) associated with fish spoilage. The amines (dimethylamine, trimethylamine, isobutylamine, 3-methylbutylamine, and 2-methylbutanamine), alcohols

(2-ethylhexanol, 1-penten-3-ol and isoamyl alcohol, ethanol), aldehydes (2- methylbutanal, 3-methylbutanal, benzaldehyde), ketones (acetone, 2,3-butanedione, 2- butanone, acetoin) and dimethyl disulfide strongly correlated with spoilage of tuna and mahi-mahi and were considered as the key spoilage indicators.

Background Information and Objectives

Fish spoilage means any change in the condition of fish resulted from microbial, enzymatic, and chemical reactions that leads to fish becoming less palatable or even poisonous (Ghaly et al., 2010; Prester et al., 2011). In the process of fish spoilage, complex physical and chemical changes are involved and the primary concern of fish spoilage including the degradation of protein, lipid oxidation and hydrolysis of carbohydrate (Hungerford, 2010; Ashie et al., 1996). Each year, the total waste of spoiled fish is around 10 to 12 million tonnes and the economic impact due to fish spoilage cannot be ignored (FAO, 2010). Toxic biogenic amines, including histamine, cadaverine, and putrescine, can be formed in fish during spoilage and consumption of spoiled fish that contain significant amounts of histamine lead to fish poisoning (Bulushi et al., 2009; Prester, 2011; Visciano et al., 2012). Tuna and mahi-mahi are identified as two major fish species responsible for histamine poisoning in the United States (Ahmed,

1991; Bulushi et al., 2009; Antoine et al., 1999). Therefore, monitoring the quality of tuna and mahi-mahi is helpful to prevent histamine poisoning.

The effect of fish spoilage on the volatile profile of fish meat has been well documented. Microbial action, enzymatic action, lipid oxidation and other chemical reactions lead to the change of volatile components in fish meat and influence the organoleptic characteristic of fish products (Edirisinghe et al., 2007). The changes of specific alcohols, carbonyls, acids, amines, sulfur compounds, aldehydes, and ammonia during fish spoilage have been reported (Ashie et al., 1996; Duflos et al., 2006). These volatile compounds are considered as quality indicators of fish product and can be used to determine fish spoilage. Short-chain carbonyls, alcohols, and esters, such as ethanol,

1-penten-3-ol, 3-methyl-1-butanol, 1-butanol, and 1-octen-3-ol, 3-methylbutanal, 2-

methylbutanal, ethyl acetate and ethyl butanoate, accumulate in spoiled fish due to microbial spoilage, enzymatic or non-enzymatic lipid oxidation and are responsible for the pungent, alcoholic, creamy, and fishy odors of spoiled fish (Leduc et al., 2012;

Duflos et al., 2006; Olafsdottir et al., 2005; Iglesias et al., 2009). Several studies have identified volatile amines, including trimethylamine (TMA), dimethylamine (DMA), and isobutylamine, as critical markers of fish freshness due to their gradual accumulation during the spoilage process and the contribution of characteristic fishy odor (Leduc et al., 2012; Bene et al., 2001; Ghaly et al., 2010; Gill et al., 1983). Sulfur compounds accumulate after fish landing due to microbial enzymatic activity and can give off unpleasant odors at extremely low concentration (Gram et al., 2002; Duflos et al., 2006;

Ashie et al., 1996; Kawai et al., 1996). The variations of volatile acids, terpenes, alkanes, and alkenes have also been observed in spoiled fish (Olafsdottir et al., 2005;

Koutsoumanis et al., 1999; Iglesias et al., 2009; Aro et al., 2003).

Gas Chromatography-Mass Spectrometry (GC-MS) is a highly effective analytical instrument that is widely used to separate, identify and quantify volatile and semi- volatile compounds in fish products (Sandra et al., 2003; Wang et al., 1999). In the GC-

MS system, analytes are separated due to their different strengths of interaction with the stationary phase, and then the “fingerprint” information of molecules is sensitively identified by mass spectrometry (Abraham et al., 1999; Sneddon et al., 2007). Over the years, GC-MS methods equipped with different sample extraction methods and column types have been applied to determine the quality indicators of spoiled fish (Hsieh et al.,

1989; Girard et al., 2000; Alexi et al., 2017; Jonsdottir et al., 2008). A GC-MS method equipped with CAR/PDMS fiber and a BPX5 capillary column was able to investigate

the freshness markers of whiting (Merlangius merlangus) (Duflos et al., 2010). An

SPME-GC-MS method using a CP-Wax 52 CB GC column was developed to detect the volatiles identified as spoilage indicators of yellowfin tuna (Thunnus albacares)

(Edirisinghe et al., 2007). An SPME-GC-MS method by Wierda et al. (2006) using a ZB-

Wax column was used to analyze the effect of storage on the volatile profile of fresh king salmon (Oncorhynchus tshawytscha). The application of the GC-MS method in determination of volatile compounds used as spoilage indicators of fish spoilage, including alcohols, acids, aldehydes, alkanes, ketones, trimethylamine, and sulfur compounds, has also been reported in other studies (Soncin et al., 2008; Leduc et al.,

2012; Alasalvar et al., 2005).

Most amines in fish products, such as isobutylamine, 3-methylbutylamine, and 2- methylbutylamine, are identified as important spoilage indicators and contribute to the fishy odor of spoiled fish (Gill et al., 1983; Eskin, 2013; Takahashi et al., 2004; Mayr and

Schieberle, 2012; Prester, 2011). However, using GC-MS instrument to determine short chain amines has some inherent challenges because of their high polarity, basic character, and high aqueous solubility. Complicated derivatization steps usually are needed to increase the volatility of amines when using GC (Staruszkiewicz et al., 1981;

Rogers et al., 1997; Du et al., 2001). The objective of this study was to develop a simplified and accurate GC-MS method to determine amines and other important aroma compounds for mahi-mahi and tuna to serve as chemical indicators of spoilage without any complex derivatization step.

Materials and Methods

Fish Samples and Preparation

The mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) analyzed in this research were caught from South Pacific waters. More than five sensory experts in FDA and National Marine Fisheries Service (NMFS) applied the sensory grading system provided in the FDA ORA Laboratory Manual (FDA, 2013) to grade the fish filets of mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) into seven grades. The grading system used by FDA/NMFS experts depended on olfaction and were graded 1 to 7 to represent quality. Grade 1 represented high quality, while grade 7 represented very poor quality fish.

Standards and Reagents

All chemicals used in this study were of analytical grade or higher. Isobutylamine,

3-methylbutylamine, and 2-methylbutylamine, trimethylamine, 1-nonanol, sodium bicarbonate, were supplied by Sigma–Aldrich (St. Louis, MO). Dimethylamine was obtained from ACROS Organics (Geel, Belgium). Sodium hydroxide, ammonium hydroxide, trichloroacetic acid were supplied by Fisher Chemical (Pittsburgh, PA).

For preparing the internal standard solution, 1-nonanol was dissolved in HPLC grade methanol at a concentration of 100 mg/L. For preparing external calibration curves of biogenic amines, three different concentrations of a mixture of biogenic

amines containing isobutylamine, 3-methylbutylamine, 2-methylbutylamine, trimethylamine and dimethylamine were used: 10, 50, 100 mg/L in HPLC grade water.

In addition, spiking known amounts of biogenic amines was applied to determine the effect of fish matrix on headspace volatile biogenic amines. A biogenic amine mixture containing isobutylamine, 3-methylbutylamine, 2-methylbutylamine, trimethylamine and dimethylamine were spiked in triplicate into homogenized grade 1 of tuna samples to achieve the final concentration of each amine as 5, 10, 50 and 100 mg/kg fish.

Extraction Procedures

The extraction procedures of Zhai et al. (2012) and Ruiz-Capillas et al. (1999) with modifications were used in this study. In brief, 6g homogenized fish sample was added to a 50 mL centrifuge tube with 20 mL of 5% trichloroacetic acid (TCA) and then was vortexed for 15min. Next the extract was centrifuged at 5000 g for ten minutes at 4

ºC. After centrifugation, the remaining solid was extracted using 20 mL of 5% TCA again by the same procedures as above, and the supernatant was collected. Both supernatants were combined and passed through a Whatman No. 1 filter paper. Fish extract was stored in a refrigerator at 4 °C and should be analyzed by GC-MS within 24 hours. To prepare samples injected into PT-GC-MS, an amount of 4.5 mL 2 mol/L

NaOH solution, 6.75 mL of saturated NaHCO3 solution (adjusted the solution to basic pH to increase volatility of amines) and 22.5 mL of filtered fish extract or standard solution were added into a 40 mL amber headspace vial (Thermo Scientific; Sunnyvale,

CA) and vortexed for 20 seconds. For each grade of mahi-mahi or tuna, the extraction and dilution procedures were performed in triplicate.

Purge and Trap Conditions

Volatile compounds in fish extract were collected and concentrated by the automated purge and trap unit AQUATek 100 (Teledyne Tekmar, Mason, OH) prior to

GC-MS analysis. The autosampler pulled 5 mL prepared fish extraction and 2 μL 100 mg/L 1-nonanol into the purge and trap unit. Fish extraction spiked with the internal standard was purged with ultrahigh purity helium gas at a flow rate 20 mL/min for 30 min at a purge temperature of 55 ºC. The headspace volatiles were adsorbed to a

Tenax trap No. 1 (Teledyne Tekmar, Mason, OH). Then the trapped compounds were desorbed at 180 °C for 3 minutes to the GC injection port.

GC-MS Analysis

Identification and quantification of analytes were conducted on an Agilent

Technologies 7890A Gas Chromatograph with 5975C Mass Spectrometer (Agilent,

Santa Clara, CA). The injected volatile compounds (split at 20:1) were separated on a

60 m Rtx-volatile Amine column (I.D. 0.32 mm) (Restek Corporation, Bellefonte, PA).

The carrier gas was helium operated at a rate of 2 mL/min. The oven temperature program was as follows: hold 6 min at 40ºC; 40-175ºC at 10ºC/min; 175-240ºC at

15ºC/min; 240-260ºC at 20ºC/min; hold 5.2min at 260ºC. The separated volatile compounds were transferred from GC to the MS to be scanned in the m/z range from

20 to 350.

Calculations

The percentage recoveries of isobutylamine, 3-methylbutylamine, 2- methylbutylamine, trimethylamine and dimethylamine for spiking standard method, external standard method and the internal standard method were quantified by dividing the back-calculated concentration of each amine by its theoretical amount. The true

concentrations of five amines in fish samples were quantified by using back-calculated concentrations from three methods (spiked standard curves, external standard curves, and addition of internal standard) divided by the percentage recoveries of each amine for each method, respectively.

To compare the amount of each compound in different grades of mahi-mahi or tuna, One-way analysis of variance (one-way ANOVA) was carried out with SAS statistical software (SAS, Cary, NC). The significance level (α) of the ANOVA test was set at 0.05. A Fisher’s Least Significant Difference (LSD) test was used to assess which means were significantly different. Pearson's correlation coefficients (r) was determined using SAS to investigate the relationships between determined volatile components and different grades of fish samples.

Results and Discussion

To obtain a simplified and accurate GC-MS method that can determine the aroma profile of mahi-mahi and tuna without derivatization, different extraction methods, split ratios, types of GC columns, temperature programs, carrier gas flow rates, were compared. The optimized purge and trap gas chromatography-mass spectrometry (PT-

GC-MS) method created in this study was applied to the analysis of volatile compounds of different grades mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) and several volatile compounds were able to be determined in these two species of fish samples. To identify spoilage indicators, peak responses of each compound of different grades of mahi-mahi or tuna were compared to determine which compounds increased, declined or remained the same during fish spoilage. Table 4-1 and Table 4-2 show that this PT-GC-MS allowed for the identification and quantification of twenty-three volatile compounds in mahi-mahi samples and nineteen volatile

compounds in tuna associated with spoilage, respectively. The flavor descriptors of volatile compounds associated with spoilage in tested mahi-mahi and tuna sample are shown in Table 4-4. Figure 4-1 shows the separation for relevant chemical indicators of spoilage in a grade 7 mahi-mahi sample.

The purge and trap method applied in this GC-MS method provided accurate and precise analysis. It has been reported that the purge and trap method can lower the limit of detection (LOD) value, extract volatile compounds in a higher amount and is more sensitive than a static headspace method (Lucentini et al., 2005; Beltran et al., 2006;

Povolo et al., 2003). The Tenax trap No. 1 used in this dynamic extraction method works well to trap nonpolar compounds and is able to remove water from the volatile compounds introduced onto GC due to its hydrophobic property. Fukami et al. (2002) used Tenax to trap volatile compounds after the fish sauce was purged for sixteen minutes, and twenty-three compounds were determined by this GC-MS method. The

Tenax trap No. 1 has also been applied in a GC-MS method to concentrate volatile compounds of Sea Bream (Sparus aurata) and seventy-eight compounds, including aldehydes, ketones, carboxylic acids, terpenes, and amines were able detected

(Alasalvar et al., 2005). The application of purge and trap to concentrate volatile compounds from fish meat, such as sea bream and sockeye salmon, has been well documented (Girard et al., 2000; Alexi et al., 2017; Jonsdottir et al., 2008; Leduc et al.,

2012).

The Rtx-volatile amine column applied in this PT-GC-MS worked well for the separation of biogenic amines. Dimethylamine, trimethylamine, isobutylamine, 3- methylbutylamine, 2-methylbutanamine were able to be separated, identified, and

quantified in tested fish samples by this developed PT-GC-MS method without any sample derivatization procedures. Determination of small chain amines by gas chromatography (GC) has inherent problems due to their physicochemical properties, including high polarity, basic character, and high aqueous solubility (Abalos et al., 2001; de Zeeuw et al., 2011). Strong adsorption of amines on the column leads to tailing peaks. In order to overcome this problem, complicated derivatization steps are commonly needed to convert amines to less polar derivatives to be more appropriate for the GC determination. Stationary phase deactivation technologies have been applied to modify the fused silica surface of amine column to make the columns more stable for determination of volatile amines (Abalos et al., 2001). Amine columns usually have low or mid polarity phases and can be determined without any complicated derivatization steps due to its inertness, water tolerance, and loadability (de Zeeuw et al., 2011).

Applying amine columns in GC-MS can improve the separation of basic compounds and avoid tailing of these compounds. A GC-MS equipped with a base-deactivated Rtx-

Volatile amine column has been used to quantify the amount of trimethylamine in standard solutions or marine sediments without any derivatization procedure

(Steinkamp et al., 2016; Zhuang et al., 2017). A SPME-GC-FID method with Rtx-5 amine column or PoraPLOT amine column was able to analyze short-chain volatile amines C1 to C9, including dimethylamine, trimethylamine, monoethylamine, isopropylamine and others, in standard solution without derivatization (Abalos et al.,

2001).

Table 4-3 shows the biogenic amines content in different grades of mahi-mahi and tuna samples calculated by the spiking standard method and the external standard

method. To calculate a biogenic amine in one fish sample, the same peak response for that amine in fish was calculated by using the standard curves provided by the spiking standard method and external standard method and then dividing by the percentage recovery of the individual amine in each method, respectively. The quantification results of each biogenic amine in fish samples by spiking standard method were higher than those calculated by external standard method (Table 4-3). These observed differences were due to the matrix effect of the fish. It was observed that for each biogenic amine concentration level of standard curve, the GC-MS peak response of each amine for the spiking standard method was lower than that for the external standard method, which led to the slopes of the spiking standard curves being lower than those of external standard curves. These observations were due to the matrix of fish changing the equilibrium distribution of an analyte between the sample phase and the gas phase, which led to the reduction of amines in the headspace. The lower instrument response for the standard curves in the spiked standard method led to the larger calculated amine amount in fish samples.

As the fish spoilage grade increased, the detected levels of trimethylamine (TMA) and dimethylamine (DMA) in both mahi-mahi and tuna samples increased significantly

(Table 4-1, Table 4-2 and Table 4-3). The accumulation of DMA and TMA in spoiled fish have been well documented by several studies (Leduc et al., 2012; Bene et al., 2001;

Prester, 2011). Significant positive correlations between TMA, and DMA concentrations with increasing spoilage grade of fish were observed for mahi-mahi and tuna (Table 4-

4). Both DMA and TMA are potent odorants giving off a characteristic fishy odor and contribute to the off-flavor of spoiled fish (Prester, 2011; Olafsdottir et al., 2005). The

odor threshold values of DMA and TMA are 0.0846 mg/L and 0.0008 mg/L, respectively

(Ruth, 1986). The precursor of DMA and TMA is identified as trimethylamine oxide

(TMAO), which is an odorless osmolyte naturally presenting in marine fish and protecting fish from dehydration (Ghaly et al., 2010). TMA can be produced from TMAO by gram-negative bacteria during the process of fish spoilage, and the maximum level of

TMA in fish was found to be 680 mg/kg (WHO, 2006). The level of TMA in fish has been considered as a marker of microbial deterioration in fish (Chan et al., 2006; Leduc et al., 2012; Bene et al., 2001; Ghaly et al., 2010; Zhang et al., 2010; Alexi et al., 2017;

Olafsdottir et al., 2005; Jorgensen et al., 2001; Alasalvar et al., 2005). Dimethylamine

(DMA) can be produced from the degradation of trimethylamine oxide (TMAO) by intrinsic enzyme activity during the spoilage (Chan et al., 2006; Ashie et al., 1996).

Chan et al. (2006) observed that there was an increase in the level of DMA in freshwater grouper (Siniperca chuatsi) and mangrove snapper (Lutjanus griseus) from zero mg/kg in fresh fish to 5.6 and 2.98 mg/kg after a twenty-eight days storage at 0 °C, respectively. The amount of DMA in fish tissue has been reported as an accepted measure of fish deterioration. Based on the TMA and DMA content detected in the seven grades of mahi-mahi and tuna (Table 4-1 and 4-2), the level of TMA above 20 mg/kg or DMA above 6 mg/kg in mahi-mahi or tuna means likely indicates poor quality.

Isobutylamine, 3-methylbutylamine, and 2-methylbutylamine were not detected in good quality mahi-mahi (e.g. Grade 1 and 2) and their level significantly increased in spoiled mahi-mahi (e.g. Grade 5, 6, 7). For all grades of tuna, these three biogenic amines were only detected in grade 4 of tuna. So correlations between the levels of these three amines with the increasing spoilage grade of tuna were zero.

Isobutylamine can give off an intense fishy odor even at low concentration and accumulation of this compound is as a result of microbial degradation of fish tissue (Gill et al., 1983). Free valine can produce isobutylamine in meat by the reaction of the bacterial valine decarboxylase during spoilage, and this reaction was determined to occur in spoiled marine fish (Eskin, 2013; Gill et al., 1983; Gruger et al., 1972). The compound 3-methylbutylamine, which also named as isopentylamine, can be formed in spoiled meat by microbial decarboxylation of leucine with the formation of fatty, unpleasant ammonia odor (Takahashi et al., 2004; Mayr and Schieberle, 2012).

Isoleucine can be decarboxylated by decarboxylase activities of bacteria and produce 2- methylbutylamine that has a fishy odor (Mayr and Schieberle, 2012; Zufall and Munger,

2016). Tuna and mahi-mahi containing any of these three amines indicates spoilage occurred in the fish sample (Table 4-1 and 4-2).

Five alcohols detected in mahi-mahi and three alcohols detected in tuna changed as spoilage progressed (Table 4-1 and Table 4-2). The increasing spoilage grade of mahi-mahi was negatively correlated with the concentrations of 2-ethylhexanol, methanol, tert-butanol and was closely positively correlated with concentrations of 1- penten-3-ol and isoamyl alcohol in fish (Table 4-4). For tuna samples, the increasing spoilage grade of fish was negatively correlated with levels of 2-ethylhexanol and was positively correlated with levels of ethanol in fish (Table 4-4). The changes of alcohols during fish spoilage are mainly associated with the microbial activity and influence the organoleptic characteristic of the fish product (Girard et al., 2000; Duflos et al., 2006).

Leduc and et al. (2012) reported that the level of 2-ethylhexanol, which compound gives off pleasant fatty and fruity odor, in European sea bass (Dicentrarchus labrax) and

Gilthead sea bream (Sparus aurata) decreased to zero after ninety days frozen storage.

Iglesias et al. (2009) also reported similar observation that 2-ethylhexanol decreased in gilthead sea bream during a 266 days frozen storage period. The loss of the compound

2-ethylhexanol may be useful as a quality marker to signify a decline in freshness of fish. Ethanol leads to a strong alcoholic smell and the formation of ethanol in spoiled fish is associated with the microbial decomposition of fish. The increase of ethanol in spoiled yellowfin tuna (Thunnus albacares), whiting (Merlangius merlangus), cod

(Gadus morhua) and mackerel (Scomber scombrus), have been reported in several studies (Edirisinghe et al., 2007; Soncin et al., 2008; Duflos et al., 2006). The accumulation of isoamyl alcohol in spoiled Baltic herring (Clupea harengus membras) has been reported and isoamyl alcohol was identified as an indicator of microbial spoilage with fusel and alcoholic odor (Aro et al., 2003). The degradation of linoleic or arachidonic acids has been reported as the major reason leading to the formation of 1- penten-3-ol with pungent odor in spoiled fish (Girard et al., 2000). For tert-butanol, pleasantly fruity, creamy, buttery odors have been identified. Based on the mean separation results (Table 4-1 and 4-2), Pearson correlations (Table 4-4), and the flavor descriptor of each compound for three alcohols (2-ethylhexanol, 1-penten-3-ol and isoamyl alcohol) in mahi-mahi and two alcohols (2-ethylhexanol, and ethanol) in tuna were used to identify key indicators that could contribute to spoilage. Based on the volatile alcohols determined in poor grades of fish (Table 4-1 and Table 4-2), the existence of isoamyl alcohol in mahi-mahi and the ethanol level in tuna above 0.1 mg/kg indicates the quality of fish is poor.

Four aldehydes, including 2-methylpropanal, 2-methylbutanal, 3-methylbutanal, benzaldehyde, significantly increased in spoiled mahi-mahi and tuna samples. These four aldehydes significantly positively correlated with spoiled fish, except 2- methylpropanal only significantly correlated with tuna. The detected level of hexanal in mahi-mahi declined as the spoilage grade increased. The odor thresholds of 2- methylpropanal, 2-methylbutanal, and 3-methylbutanal are low and the range of their thresholds is from 2.24 to 40.7 ng/g (Shimoda et al., 1996). The increase of these three aldehydes in spoiled fish is associated with the lipid oxidation, and deamination of amino acids due to microbial catabolic activities (Joffraud et al., 2001; Shimoda et al.,

1996). Duflos et al. (2006) reported a similar trend to this study in that the level of 2- methylpropanal, 2-methylbutanal, 3-methylbutanal in whiting (Merlangius merlangus), cod (Gadus morhua) and mackerel (Scomber scombrus) increased after a ten days storage under 4ºC. Significant increases of 2-methylpropanal, 2-methylbutanal, and 3- methylbutanal in European sea bass (Dicentrarchus labrax), gilthead seabream (Sparus aurata), cod (Gadus morhua) and salmon (Salmo salar) were reported after a ninety days frozen storage (Leduc et al., 2012). These three aldehydes have been identified as good markers of fish spoilage in several studies (Duflos et al., 2010; Olafsdottir et al.,

2005; Jorgensen et al., 2001; Joffraud et al., 2000; Alasalvar et al., 2005; Mace et al.,

2013). Benzaldehyde is an aroma with almond, fruity odor and Mace et al. (2013) observed the accumulation of benzaldehyde in P. phosphoreum-inoculated raw salmon

(Salmo salar) after an eight days storage at 8 °C. Hexanal is a carbonyl compound with a “green” plant-like odor and is a volatile commonly present in fresh fish. Wierda et al.

(2006) reported the decline of hexanal in spoiled king Salmon (Oncorhynchus

tshawytscha) and identified hexanal as an indicator for salmon freshness. The level of hexanal in both Whiting (Merlangius merlangus) and Cod (Gadus morhua) decreased after stored the fish samples for 10 days at 4 °C (Duflos et al., 2006).

Three ketones, including acetone, 2,3-butanedione, 2-butanone, generally increased as the spoilage grade of mahi-mahi and tuna increased (Table 4-1 and Table

4-2). Significant positive correlations between levels of these three ketones with the spoiled grade of mahi-mahi and the significant correlation between 2-butanone level and grade of tuna were observed (Table 4-4). The residual glycogen catabolism and fatty acid catabolism are major reasons lead to the accumulation of these three ketones in spoiled fish and release a buttery, green odor (Joffraud et al., 2001). Wierda et al.

(2006) reported the accumulation of acetone in spoiled king salmon. Significant increases of acetone and 2-butanone in sea bream (Sparus aurata) were observed after the freshly caught fish were stored on ice for six days (Soncin et al., 2008). Duflos et al.

(2006) observed that the level of acetone in Cod (Gadus morhua), the levels of 2,3- butanedione, 2-butanone in Mackerel (Scomber scombrus) and Cod (Gadus morhua) increased after the fish samples was stored at 4 °C for ten days. Similar results have also been reported by several studies and acetone, 2,3-butanedione, 2-butanone are identified as good markers to characterize freshness of fish due to their accumulation during fish spoilage (Leduc et al., 2012; Alasalvar et al., 2005; Prost et al., 2004;

Olafsdottir et al., 2005).

Dimethyl disulfide is a sulfur compound with sulfurous odor and the formation of this compound was observed only in spoiled mahi-mahi samples. The level of dimethyl disulfide is in initially low in freshly caught fish and accumulates after fish landing

(Duflos et al., 2006). Dimethyl disulfide has an extremely low odor threshold, and it can importantly affect overall aroma of spoiled fish. The generation of dimethyl disulfide is associated with the degradation of sulfur-containing amino acids and peptides. After fish samples were stored in ice for 23 days, the level of dimethyl disulfide increased from 20 to 560 ng/g in cultured sea bream (Sparus aurata) and from 0 to 617 ng/g in wild sea bream (Sparus aurata), respectively (Alasalvar et al., 2005). Similar results were also reported by Olafsdottir et al. (2005) and Leduc et al. (2012) that dimethyl disulfide increased in cod (Gadus morhua) as the spoiled degree increased. Any mahi-mahi sample containing dimethyl disulfide is of poor quality (Table 4-1).

The decline of pyridine in the tested spoiled mahi-mahi samples was observed in this study. Acetoin was only detected in grade 4, 5, and 7 tuna samples and the detected level of acetoin was highest in grade 5 tuna, and then followed by grade 4 tuna, grade 7 tuna. Acetoin is also named as 3-hydroxy-2-butanone and is formed by microbial activities occurring the fish spoilage imparting a buttery odor (Emborg et al.,

2006). It also has been reported that acetoin can be reduced to 2,3-butanediol as the spoilage progresses (Duflos et al., 2006). The relative level of acetoin in yellowfin tuna

(Thunnus albacares) increased after fish samples were stored in ice for twenty days

(Edirisinghe et al., 2007). Acetoin has been identified as a useful marker to characterize freshness of fish in previous studies (Duflos et al., 2010; Olafsdottir et al., 2005;

Jonsdottir et al., 2008). The cut off level of acetoin for good quality tuna should be zero based on the GC-MS results of the different grades of tuna samples (Table 4-2).

Summary

A PT-GC-MS method with RTX-Volatile Amine column was developed for the analysis of mahi-mahi and tuna volatiles for chemical indicators of spoilage. Twenty

aroma compounds in mahi-mahi and sixteen aroma compounds in tuna associated with fish spoilage were determined. Five biogenic amines (dimethylamine, trimethylamine, isobutylamine, 3-methylbutylamine, and 2-methylbutanamine) strongly correlated with spoilage of both mahi-mahi and tuna. Three alcohols (2-ethylhexanol, 1-Penten-3-ol and isoamyl alcohol), three aldehydes (2-methylbutanal, 3-methylbutanal, benzaldehyde) 3 ketones (acetone, 2,3-butanedione, 2-butanone), and dimethyl disulfide had good correlation with spoilage of mahi-mahi. Two alcohols (2-ethylhexanol, ethanol), four aldehydes (2-methylpropanal, 2-methylbutanal, 3-methylbutanal, benzaldehyde) and three ketones (2,3-butanedione, 2-butanone, acetoin), strongly correlated with tuna spoilage. The purge and trap system applied in this method enriched the concentration of the analytes injected into GC-MS instrument. The application of RTX-volatile amine column made it possible to identify and quantify these five amines without any sample derivatization. This simplified and accurate PT-GC-MS method could be used to monitor the markers of tuna and mahi-mahi spoilage.

Figure 4-1. Example of a chromatogram (mahi-mahi grade 7). Nineteen chemical indicators were found: 1. Methanol 2. DMA 3. TMA 4. Ethanol 5. Acetone 6. tert-Butanol 7. 2-Methylpropanal 8. 2,3-Butanedione 9. 2-Butanone 10. Isobutylamine 11. 3-Methylbutanal 12. 2-Methylbutanal 13.1-Penten-3-ol 14. 3-Methylbutylamine 15. 2-Methylbutylamine 16. Isoamyl Alcohol 17. Dimethyl Disulfide 18. Benzaldehyde 19. 2-Ethylhexanol

Table 4-1. Volatile compounds associated with spoilage in seven grades of mahi-mahi calculated by internal standard method, (ng/g) fish sample. Compound name M1 M2 M3 M4 M5 M6 M7 Methanol 1696±338 2942±609 2532±833 3365±2800 1305±394 1296±349 1093±374 DMA n.d. 4086±7070C 5277±232C n.d. 26410±4970AB 23500±5860B 33850±4730A TMA 14800±4140C 18020±589C 14830±2520C 14800±4850C 138100±16200B 119700±10800B 243000±48100A Acetone 206.8±123B 231.8±6.18B 884.7±337A 1038±193A 907±87.9A 1011±114A 764.9±30A tert-butanol 1045±179 1129±54 1003±309 1014±60.4 994.5±91.4 996.2±162 888.9±113 2-methylpropanal 71.19±9.25B n.d. 80.21±31.3B 61.65±14.8B 116.5±28.5A 79.24±3.04B 73.89±4.3B 2,3-Butanedione 152.1±41.6C 404.3±81.5A 97.45±17.8CD 132.1±41CD 243.3±66.9B 92.89±3.99CD 53.99±11.8D 2-Butanone 100.6±9.37 125.9±26.6 136.4±7.11 145.7±24.3 191±52.1 145.6±33.1 138.9±15.4 Isobutylamine n.d. n.d. n.d. n.d. 953.8±1030B 2394±565A 975±242B 3-methylbutanal 173.3±15.6C 123.2±26.4C 174.5±16.7C 147.4±11.2C 274.9±61.6B 240.6±9.66B 473.9±64.8A 2-methylbutanal n.d. n.d. 153±23AB 107.6±15.7B 129±116B 151.1±9.38AB 229.6±30.5A 1-Penten-3-ol n.d. 455.9±47.5B n.d. 87.82±11.9D 182.5±16C 690.6±32.7A 52.51±11.4E 3-methylbutylamine n.d. n.d. n.d. n.d. 2108±1750B 7369±1160A 1200±249BC 2-methylbutylamine n.d. n.d. 335.2±380C 364.3±631C 1333±1080B 3343±464A 756.4±126BC isoamyl alcohol n.d. n.d. n.d. n.d. n.d. n.d. 238.7±9.35 Pyridine 36.2±4.89B 72.48±16.4A n.d. n.d. n.d. n.d. n.d. Dimethyl disulfide n.d. n.d. n.d. 24.33±21.1B n.d. 3.34±5.79C 128.4±10.8A Hexanal 22.96±20.8C 131.4±20.4A n.d. 10.15±17.6C 32.89±32.9C 97.21±16.2B n.d. Benzaldehyde 88.92±5.93E 256.2±15.7B 122.6±3.21D 186.6±16.5C 162.9±8.88C 344.6±9.43A 273.8±35.7B 2-ethylhexanol 316.5±10.7A 300.9±4.12A 60.03±34.8D 86.63±11.5CD 109.6±56BC 136.1±8.2B 129.7±1.47BC *Different letters within the same row indicate significant differences according to an LSD means separation test (p<0.05). n.d. indicates not detected

Table 4-2. Volatile compounds associated with spoilage in seven grades of tuna calculated by internal standard, (ng/g) fish sample. Compound Name T1 T2 T3 T4 T5 T6 T7

DMA 3988±747BC 2133±1030B 3201±1120BC 5729±2420AB 8209±4150A 1768±927 C 4141±1040BC

TMA 12390±3770D 16470±1220D 15260±932D 23730±3450C 36180±4740A 28890±2640B 33620±613AB

Ethanol 33.77±6.86 46.9±6.21 39.4±4.96 118.1±13.5 174.3±36.2 1211±1510 486.5±29.9

Acetone 1389±1720 2839±687 1399±706 1902±1460 1268±119 700±598 2619±272 tert-butanol 646.7±111 467.6±105 650±14.2 534.4±151 686.1±136 559.4±62.6 566.1±19.2

2-methylpropanal 7.983±7.04E 19.97±1.88DE 30.92±17.3CD 52.91±2.17B 54.27±6.28B 103.5±17.7A 45.59±4.65BC

2,3-Butanedione n.d. n.d. n.d. 404.4±45.3B 637.8±13.7A 30.4±11.5D 64.19±5.8C

2-Butanone 24.59±11.7C 40.64±4.92B 27.76±6.32C 49.32±4.94AB 57.02±5.13A 43.18±9.63B 46.59±3.57AB

Isobutylamine n.d. n.d. n.d. 677.8±161 n.d. n.d. n.d.

3-methylbutanal 71.3±30.2D 93.76±20.1D 100.9±20.2D 205.7±18.8B 168.2±7.62C 288.4±26.7A 149±8.95C

2-methylbutanal 26.27±17.1D 40.23±10.4D 45.26±5.44D 96.45±10.2B 76.14±8.98C 157.9±13A 75.15±7.04C

Acetoin n.d. n.d. n.d. 935.2±88.5B 1567±69.4A n.d. 51.11±5.47C

3-methylbutylamine n.d. n.d. n.d. 1820±365 n.d. n.d. n.d.

2-methylbutylamine n.d. n.d. n.d. 751.4±172 n.d. n.d. n.d.

Benzaldehyde 23.97±14.6E 78.54±13.4D 23.82±3.66E 106.1±4.64B 123.7±3.56A 88.61±11.8CD 94.82±11.8BC

2-ethylhexanol 53.73±26.5AB 71.76±8.75A 53.65±2.17AB 50.11±4.63BC 52.34±5.3AB 52.63±2.35AB 31.47±1.55C

*Different letters within the same row indicate significant differences according to an LSD means separation test (p<0.05). n.d. indicates not detected

Table 4-3. Biogenic amines contents in seven grades of mahi-mahi and tuna sample (ng/kg) calculated by the spiked standard method and an external standard method. Sample DMA TMA Spiked External Spiked External M1 n.d. n.d. 28700±8030 13350±3740 M2 4098±7100 432.3±749 34940±11400 16250±5310 M3 5293±2330 558.4±246 28760±4900 13370±2270 M4 n.d. n.d. 28710±9410 13350±4380 M5 26490±4980 2794±525 267700±31400 124500±14600 M6 23570±5870 2486±620 232200±20900 108000±9730 M7 33960±4750 3582±501 471300±93300 219200±43400 T1 4000±749 422±79 24020±7310 11170±3400 T2 2139±1030 225.7±108 31940±2370 14860±1100 T3 3211±1120 338.7±118 29600±1810 13770±841 T4 5746±2430 606.1±256 46020±6690 21400±3110 T5 8234±4160 868.6±439 70180±9190 32640±4270 T6 1773±931 187.1±98.1 56020±5120 26050±2380 T7 4154±1040 438.2±110 65210±1190 30330±553

Table 4-3. Continued Sample Isobutylamine 3-methylbutylamine 2-methylbutylamine Spiked External Spiked External Spiked External M1 n.d. n.d. n.d. n.d. n.d. n.d. M2 n.d. n.d. n.d. n.d. n.d. n.d. M3 n.d. n.d. n.d. n.d. 336.4±381 30.92±35 M4 n.d. n.d. n.d. n.d. 365.7±633 33.61±58.2 M5 958.1±1040 83.71±90.7 21050±1740 147±122 1338±1080 123±99.4 M6 24050±568 210.1±49.6 7361±1150 514±80.6 3356±465 308.5±42.8 M7 979.4±243 85.57±21.2 1199±249 83.69±17.4 759.2±127 69.79±11.6 T1 n.d. n.d. n.d. n.d. n.d. n.d. T2 n.d. n.d. n.d. n.d. n.d. n.d. T3 n.d. n.d. n.d. n.d. n.d. n.d. T4 680.8±162 59.48±14.2 1818±365 127±25.5 754.2±172 69.33±15.8 T5 n.d. n.d. n.d. n.d. n.d. n.d. T6 n.d. n.d. n.d. n.d. n.d. n.d. T7 n.d. n.d. n.d. n.d. n.d. n.d. *S: calculated levels of amines using spiking standard method (ppm in fish) *E: calculated levels of amines using external standard method (ppm in fish) *n.d: indicates not detected

Table 4-4. Flavor descriptors of volatile compounds associated with spoilage in mahi- mahi and tuna samples. Pearson correlation coefficients between levels of volatile compounds with increasing spoilage grade of mahi-mahi and tuna Compound Name Mahi-mahi Tuna Flavor descriptor Pearson Pearson coefficient (r) coefficient (r) Methanol -0.3571 n.f. Alcoholic (Ruth, 1986) DMA 0.849* 0.1289 Fishy (Ruth, 1986) TMA 0.855* 0.8654* Fishy (Olafsdottir et al., 2005) Ethanol n.f. 0.4451* Alcoholic (Ruth, 1986) Acetone 0.6569* -0.0382 Apple pear (Ruth, 1986) tert-butanol -0.362 -0.0148 Camphor-like odorW 2-methylpropanal 0.4099 0.7246* Fresh, floral (Girard et al., 2000) 2,3-Butanedione -0.4688* 0.2689 Sweet, butteryW 2-Butanone 0.4391* 0.587* Fruity, greenW Isobutylamine 0.6751* 0 Fishy (Gill et al., 1983) 3-methylbutanal 0.7689* 0.685* Sweet, caramel, fishy (Olafsdottir et al., 2005) 2-methylbutanal 0.7967* 0.7021* Musty (Girard et al., 2000) 1-Penten-3-ol 0.2353 n.f. Pungent (Girard et al., 2000) Acetoin n.f. 0.2094 Buttery (Xiao et al., 2014) 3-methylbutylamine 0.5644* 0 Unpleasant ammonia (Takahashi et al., 2004) 2-methylbutylamine 0.6043* 0 Fishy (Zufall and Munger, 2016) Isoamyl alcohol 0.612* n.f. Fusel, alcoholic (Komes et al., 2006) Pyridine -0.673* n.f. FishyW Dimethyl disulfide 0.6258* n.f. Sulfurous (Alasalvar et al., 2005) Hexanal -0.1482 n.f. Green (Girard et al., 2000) Benzaldehyde 0.6487* 0.6414* Almond, fruity (Girard et al., 2000) 2-ethylhexanol -0.6144* -0.5408* Fatty, fruity (Mahmoud et al., 2017) * Significant below 0.05 level. n.f. signifies not found. W Flavor descriptor obtained from the website reference: http://www.thegoodscentscompany.com/

CHAPTER 5 DETERMINING QUALITY ATTRIBUTES OF MAHI-MAHI AND TUNA BY OPTIMIZED COLORIMETRIC STRIPS

Digest

Tuna (Thunnus albacares) and mahi-mahi (Coryphaena hippurus) are two major fish species responsible for histamine poisoning in the United States. The purpose of this research was to develop a low-cost and easily operated colorimetric strip method for the rapid determination of biogenic amines in mahi-mahi and tuna with high sensitivity. The color strip method was developed by investigating different types of dyes, filter papers, sample volume, water bath temperature and other parameters.

Ultimately rose bengal and bromophenol blue (BPB) dyes were chosen. These two dyes produced standard curves with good linearity and uniformity of color change. The r2 values for the standard curves of the rose bengal and BPB were 0.9535 and 0.8883, respectively. Significant positive Pearson correlations coefficients (r) between the volatile biogenic amine levels detected by these two colorimetric strip methods with increasing spoilage grade of mahi-mahi (rose bengal: r=0.8907, p<0.0001; BPB: r=

0.8711, p<0.0001) and tuna (rose bengal: r= 0.8351, p<0.0001; BPB: r= 0.7362, p=0.0001) were observed. For mahi-mahi, the volatile biogenic amines detected by the colorimetric strips correlated positively with increasing levels of eight biogenic amines

(histamine, putrescine, cadaverine, trimethylamine, dimethylamine, isobutylamine, 3- methylbutylamine, and 2-methylbutylamine), free alanine, four aldehydes (2- methylpropanal, 2-methylbutanal, 3-methylbutanal, benzaldehyde), isoamyl alcohol, two ketones (acetone, 2-butanone), and dimethyl disulfide. For tuna, the results determined by colorimetric strips positively correlated with three biogenic amines (histamine, putrescine, cadaverine, trimethylamine), three free amino acids (glycine, alanine,

phenylalanine), four aldehydes (2-methylpropanal, 2-methylbutanal, 3-methylbutanal, benzaldehyde), and ethanol.

Background Information and Objectives

Due to the high moisture content, rich free amino acids, nucleotides and trimethylamine oxide, raw fish is easily perishable (Gram and Huss, 1996; Prester,

2011). The combined effects of microbial, enzymatic, and chemical reactions result in the degradation of fish and lead to the generation of undesirable metabolites in spoiled fish (Ghaly et al., 2010). The mechanisms used by fish to prevent bacteria invading the tissue cease after fish death and bacteria start to contaminate the internal tissues of fish

(Fraser and Sumar, 1998). Biogenic amines are organic bases and can be produced in fish by microbial decarboxylation of free amino acids during the fish spoilage (Zhai et al.,

2012; Bulushi et al., 2009; Ashie et al., 1996). Histamine transformed from free histidine by bacterial histamine decarboxylase has been reported as the principal toxin responsible for fish poisoning and histamine poisoning occurs globally (Hungerford,

2010; Visciano et al., 2012). Putrescine and cadaverine are the biogenic amines generated from decarboxylation of arginine and lysine respectively, and are identified as significant indicators of fish quality due to their ability to potentiate histamine toxicity

(Prester 2011; Bulushi et al., 2009; Visciano et al., 2012). Tuna and mahi-mahi are two major fish species most commonly associated with histamine poisoning in the United

States due to the high levels of free histidine in their muscles (Ahmed, 1991).

The organoleptic characteristic of fish significantly changes during fish spoilage due to the influence of microbial action, enzymatic action, and lipid oxidation

(Edirisinghe et al., 2007). Some biogenic amines have major effects on fish quality by contributing to the typical fishy odor (Fraser and Sumar, 1998; Takahashi et al., 2004;

Olafsdottir et al., 2005). Trimethylamine oxide (TMAO) is non-odorous and exists at high levels in marine fish to help prevent fish from becoming dehydrated (Gram et al., 2002).

Trimethylamine (TMA) has a stale fish odor and has been used as a quality marker to investigate the freshness of fish due to the transformation of TMAO to TMA during fish spoilage (Ghaly et al., 2010; Leduc et al., 2012; Bene et al., 2001). Dimethylamine

(DMA - ammonia-like odor) can also be reduced from TMAO and is considered as an index for fish freshness based on its significant influence on fish odor (Chan et al., 2006;

Loughran et al., 2000). Specific alcohols, carbonyls, acids, sulfur compounds, aldehydes, and ammonia accumulating in fish during spoilage also release the fishy, sour, or ammonia-like odors characteristic of spoiled fish (Leduc and et al., 2012;

Iglesias et al., 2009; Duflos et al., 2005; Olafsdottir et al., 2005; Ashie et al., 1996;

Kawai et al., 1996).

The quality of tuna and mahi-mahi currently is graded by Food and Drug

Administration (FDA) / National Marine Fisheries Service (NMFS) experts using olfaction (Pivarnik et al., 2001). The fish sample is evaluated by analysts using standardized sensory criteria and is then assigned a sensory score between 0 to 100 points, where 0 points means no deterioration and 100 points presents severe decomposition. Fish samples assigned sensory scores are then ranked into one of seven grades. Grade 1 represents high quality, while grade 7 represents very poor quality fish (Pivarnik et al., 2001). In this subjective sensory grading system, fish can be recognized as stale, bad or putrid fish by sight and smell. The information provided by this subjective sensory method is close to what consumers experience. However, sensory panels need to be trained to understand how to use this standardized sensory

grading system. In addition, human variation cannot be ignored in this subjective method. Therefore, a rapid, easily operated and objective analytical method is needed to analyze fish spoilage (Murray et al., 2001).

Several amine sensitive dyes have been investigated and applied for the determination of volatile amines (Rakow et al., 2005; Steiner et al., 2010; Kuchmenko et al., 2011). Biogenic amines generated in fish during spoilage react with amine sensitive dyes and result in the color change of these dyes. Three families of amines sensitive dyes are metalated tetraphenylporphyrins, pH indicators, and highly solvatochromic dyes (Rakow et al., 2005). In previous research, a new type of colorimetric strip containing bromophenol blue, which is a sulfonated hydroxy-functional triphenylmethane dye, was developed by Dole et al. (2016) and was colorimetrically responsive to volatile amines produced by tuna and mahi-mahi. The volatile biogenic amines in mahi-mahi detected by this BPB strip correlated with the quality grades of fish

(Dole et al., 2017). However, the linearity of the standard curve from this BPB method was low, and the color change of the BPB strip was not uniform. A rapid colorimetric strip should be developed and optimized to maximize strip sensitivity and uniformity of color change. In this study, colorimetric strip methods were validated by investigating the correlation of the volatile amine content in fish samples detected by the colorimetric strip method with the spoilage grades of fish, and results obtained by UHPLC, ELISA, and GC-MS methods.

Materials and Methods

Fish Samples and Preparation

The mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) analyzed in this study were caught from South Pacific waters. More than five sensory

experts in FDA and National Marine Fisheries Service (NMFS) applied the sensory grading system provided in the FDA ORA Laboratory Manual (FDA, 2013) to grade the fish filets of mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) into seven grades. The grading system used by FDA/NMFS experts depended on olfaction, and were graded 1 to 7 to represent their quality. Grade 1 represented high quality, while grade 7 represented very poor quality.

The individually packaged and graded fish filets from FDA/NMFS were shipped overnight, received frozen on dry ice, and then stored in a -20 ºC freezer until analysis was performed. For each grade of fish samples, vacuum packaged frozen samples were defrosted overnight at room temperature, and then were cut into pieces and homogenized by a blender (Total Blend Classic, Blendtec, Orem, UT) to perform chemical analysis.

Standards and Reagents

All chemicals used in this study were of analytical grade or higher.

Trimethylamine, histamine, cadaverine, rose bengal were supplied by Sigma–Aldrich

(St. Louis, MO). Dimethylamine and bromophenol blue were obtained from ACROS

Organics (Geel, Belgium). Putrescine was purchased from MP Biomedicals (Santa Ana,

CA).

The stock biogenic amine cocktail was 10 mg/L for each biogenic amine and was prepared by dissolving cadaverine, putrescine, histamine, dimethylamine, and trimethylamine in Milli-Q water (EMD Millipore Synergy Ultrapure Water Systems,

Darmstadt, Germany). For quantitation, an external calibration curve was utilized. Five different concentrations of biogenic amines cocktail were prepared by diluting the stock

standard cocktail solution with Milli-Q water: 10, 20, 30, 40, 50 mg/L for the total biogenic amine concentration.

Colorimetric Strip Method

Three-quarter-inch squares of Whatman No. 4 filter paper were soaked in a 0.2%

BPB solution in 70% ethanol or 0.24% rose bengal solution in 70% ethanol for one minute. Then the strips were allowed to dry for an hour. The dried rose bengal strips were acidified by 0.001M hydrochloric acid (HCl) solution to convert the color of rose bengal strip from pink to transparent color and then were allowed to dry for another hour. The dried BPB and rose bengal strips were then ready to measure the volatile amines of the fish samples or the standard cocktail. Homogenized fish filet (50 grams) or 50 mL biogenic amine cocktail was put on the bottom of a 250 mL glass jar and BPB or rose bengal strip was suspended in the sealed jars. The BPB strip or rose bengal strip exposed to 50 mL pure Milli-Q water was used as the blank strip for each colorimetric strip method. The sealed jars were put in a 45 ºC water bath for one hour in

BPB strip method and for 45 minutes in rose bengal strip method. The strip was removed, and a colorimeter (Chroma Meter CR-400/410, Konica Minolta, Tokyo, Japan) was used to measure the color of the strip based on the Hunter L*a*b* color system. For each strip, five L*a*b measurements were conducted. The color of the four corners and the center of each filter paper were measured and averaged. For each grade of mahi- mahi or tuna, the color strip detection procedures were performed in five repetitions.

The b* value difference between the blank strip and exposed strip for BPB strip method, and the a* value difference between the blank strip and exposed strip for rose bengal strip method were used to back calculate the volatile biogenic amines levels in samples by using external standard curves.

Determination of Biogenic Amines and Free Amino Acids by UHPLC (Conducted in Chapter 3)

In brief, 3g homogenized fish sample was added in a 15 mL centrifuge tube with

10 mL of 5% trichloroacetic acid (TCA) and vortexed for 15min. The centrifuge tube spun at 5000 g at 4 ºC for ten minutes. After the extract was removed, the remaining solid was extracted using 10 mL of 5% TCA again by the same procedures as above, and the supernatant was collected. Both supernatants were combined and passed through a Whatman No. 1 filter paper.

An amount of 300 μL 2 mol/L NaOH solution and 300 μL of saturated NaHCO3 solution were added to 1 mL of filtered fish extract or 1mL of standard solution. Dansyl chloride was dissolved in acetone to archive 1% (w/v) concentration, and 2 mL of this solution was added the resulting solution and then was protected from light and incubated for 60 min at 40 ºC water bath (Isotemp 220, Fischer Scientific, Pittsburgh,

PA). Excess dansyl chloride was removed by adding 120 μL of an NH4OH solution (4 mol/L) and then the solution was stored away from light for 1 hour. The solution was collected and filtered by a 0.2 μm PTFE membrane (Phenomenex, Torrance, CA) before LC injection. For each grade of mahi-mahi or tuna, the extraction and derivatization procedure were performed in triplicate. Quantification of the amino acids and biogenic amines was carried out by the reversed phase UHPLC method developed in Chapter 3.

Determination of Aroma Profile by PT-GC-MS (Conducted in Chapter 4)

In brief, 6 g homogenized fish sample was added to a 50 mL centrifuge tube with

20 mL of 5% trichloroacetic acid (TCA) and then was vortexed for 15min. Next the extract was centrifuged at 5000 g for ten minutes at 4 ºC. After centrifugation, the

remaining solid was extracted using 20 mL of 5% TCA again by the same procedures as above, and the supernatant was collected. Both supernatants were combined and passed through a Whatman No. 1 filter paper. An amount of 4.5 mL 2 mol/L NaOH solution, 6.75 mL of saturated NaHCO3 solution (adjusted the solution to basic pH to increase volatility of amines) and 22.5 mL of filtered fish extract or standard solution were added into a 40 mL amber headspace vial (Thermo Scientific; Sunnyvale, CA) and vortexed for 20 seconds. For each grade of mahi-mahi or tuna, the extraction and dilution procedures were performed in triplicate. Volatile compounds in fish were determined by the purge and trap gas chromatography-mass spectrometry (PT-GC-MS) developed in Chapter 4.

Statistical Analysis

Pearson's correlation coefficients (r) were determined using SAS to investigate the relationships between the results of colorimetric strip methods with the histamine- specific ELISA kit results obtained in Chapter 3, the contents of biogenic amines and free amino acids obtained by UHPLC in Chapter 3, and the contents of volatile compounds associated with fish spoilage obtained by GC-MS in Chapter 4 for fish samples. To compare the amount of each compound in different grades of mahi-mahi or tuna using the strips, one-way analysis of variance (one-way ANOVA) was carried out with SAS statistical software (SAS, Cary, NC). The significance level (α) of the ANOVA test was set at 0.05. A Fisher’s LSD procedure was used to assess which means were significantly different.

Results and Discussion

To obtain accurate, sensitive, and rapid colorimetric strip methods to detect degradation levels of mahi-mahi and tuna, different dye reagents, concentrations of dye

solution, concentrations of acid used to acidify the colorimetric strip, types of cellulose filter papers, size of filter papers, headspace volumes jars, the methods of exposing strips to sample, and water bath treatment time were investigated. Some optimizing procedures had a negative influence on the linearity of the standard curve and the uniformity of strip color change. For example, increasing concentration of dye solution increased the linearity of the standard curve but made the strip color change unevenly when reacted with the fish samples. The optimized colorimetric strip should achieve the balance to produce biogenic amine standard curves with good linearity and also change color uniformly when exposed to fish samples. The r2 value for the standard curve of

Dole et al. (2016) was only 0.6657, and that value increased to 0.9535 and 0.8883 in rose bengal strip and BPB strip method after optimization, respectively (Table 5-1).

Figure 5-1 and Figure 5-2 show the colorimetric response of the rose bengal strip and

BPB strip to biogenic amine cocktail standards, respectively. Comparing with the BPB strip developed in Dole et al. (2016), the optimized rose bengal strip and BPB strip increased the uniformity of color change. Also, the rose bengal strip was more sensitive to the low concentration of biogenic amines samples (Figure 5-1 and Figure 5-2).

Bromophenol blue is a sulfonated hydroxy-functional triphenylmethane dye that is commonly used as an acid-based indicator with a low pKa of 4.1 (Mills et al., 1995). If the pH of the environment around bromophenol blue is higher than the pKa value of this dye, bromophenol blue will lose a proton and become deprotonated in an aqueous solution. This displacement changes the electron distribution within the molecule and also leads to shifting in the characteristic absorption spectra of BPB. In an acidic solution, the UV-blue region is strongly absorbed by BPB and the color of this dye

appears as yellow; while increasing pH of the solution leads BPB to mainly absorb in the red region and show blue color (Chattopadhyaya et al., 2016; Henari et al., 2012).

Kuchmenko et al. (2011) reported that bromophenol blue was an amine indicator and changed color from blue to yellow when reacted with amines in standard solutions containing triethylamine, N,N-dimethylformamide dimethyl acetal, piperidine, cyclohexylamine, aniline, N-methylaniline, and benzylamine. Miller et al. (2006) also used bromophenol blue as an indicator that was able colorimetrically responsive to the volatile biogenic amines produced in spoiled cod samples. An indicator strip combining bromophenol blue (BPB) developed by Dole et al., (2017) showed good correlation between volatile biogenic amine content with quality grades for mahi-mahi.

Rose bengal is an anionic water-soluble xanthene dye with photophysical properties and has played a significant role in microscopy staining, and chemosensing

(Lamberts and Neckers, 1984; Jiang et al., 2013). Rose bengal can be used as an acid- base indicator and changes its color with pH based on a protonation and deprotonation reaction. When rose bengal strip was acidified by an aqueous hydrochloric acid (HCl) solution in this study, the pH of the environment decreased, and this dye was in lactone

(colorless) form. Rose bengal is able to transform back to its quinoid form developing pink color due to base exposure (Akerlind et al., 2011). Amines, such as piperidine, diethylamine, triethylamine, trimethylamine, have basicity and can neutralize rose bengal with color change (Paczkowski et al., 1985). Rose bengal has been used as the indicator in several colorimetric methods to quantify the basic amines in the sample rapidly. A porous cellulose tape containing rose bengal developed by Nakano et al.

(1994) was found to be a highly sensitive and rapid tool to quantify ammonia levels in

the air. Miller et al. (2006) also used rose bengal as a spoilage indicator to detect the biogenic amines content in frozen cod samples based on the colorimetrically responsive of rose bengal to volatile bases generated by fish decomposition.

Dole et al. (2016) reported that BPB strip method was a non-specific assay and reacted with a broad class of volatile biogenic amines. Rather than using one specific biogenic amine to create the calibration curve, the standard cocktail containing the equal concentration of histamine, cadaverine, putrescine, trimethylamine (TMA), dimethylamine (DMA), which are identified as major biogenic amines in spoiled fish, were used to build the calibration curve. Table 5-1 shows the regression results for the biogenic amine cocktails for rose bengal and BPB method. For the rose bengal strip method, the x-axis was the total concentration of biogenic amine cocktail (ppm) and y- axis was the a* score. The a* score is the absolute difference between the a* value of the blank strip and the exposed strip. The a* value is the red/green coordinate of the

L*a*b* color system, and the more negative a* value means a greener hue is present; the more positive a* value means a redder hue is present. As the total concentration of biogenic amines increases, the rose bengal strip becomes pinker and the a* value difference increases. For the BPB strip method, the x-axis was still the total concentration of biogenic amines (ppm), while the b* score was plotted along the y-axis.

The b* score presents the b* value difference between unexposed and exposed strips.

The b* value in the L*a*b* system presents the blue/yellow coordinate, which a negative value means blue and positive means yellow. In the absence of biogenic amines, a yellow-blue color is seen after the strip was exposed to the biogenic amines solution and higher amines concentration yielded deeper blue color.

100

The levels of volatile biogenic amines in mahi-mahi (Coryphaena hippurus) and yellowfin tuna (Thunnus albacares) were detected by rose bengal strips and BPB strips, and the calculated results are shown in Table 5-2 and Table 5-3. The fish filets of mahi- mahi and tuna were graded 1 to 7 by trained sensory experts of the FDA and National

Marine Fisheries Service (NMFS). Pearson correlation coefficients (r) were calculated among spoilage grades of fish and results obtained by different analytical methods to investigate the linear relationship between variables. The range of the Pearson correlation coefficient is between -1 to 1. A correlation closing to -1/or 1 indicates there is good negative/or positive linear relation between two variables, and a correlation of 0 means the two variables are not correlated. For both the rose bengal strip method and

BPB strip method, the calculated volatile biogenic amines in fish samples generally increased as the spoilage grade of mahi-mahi and tuna increased (Table 5-2 and Table

5-3). Significant positive Pearson correlations between rose bengal results and BPB results with the increasing spoiled grade of mahi-mahi (rose bengal: r=0.8907, p<0.0001;

BPB: r= 0.8711, p<0.0001) and tuna (rose bengal: r= 0.8351, p<0.0001; BPB: r= 0.7362, p=0.0001) were observed (Table 5-4 and Table 5-5). Given that the grade of fish filets is assigned based on the olfaction, these two types of colorimetric strips developed in this study have good correlation with the organoleptic characteristic of fish products.

The Pearson correlations between the Gas Chromatography-Mass Spectrometry

(GC-MS) results and the colorimetric strips results can be used to explain the close correlation between colorimetric strip response and olfactory response to fish (Table 5-4 and Table 5-5). Among the volatile compounds detected in mahi-mahi and tuna associated with fish spoilage by the GC-MS method, the concentration of

101

trimethylamine (TMA) in fish samples was at least ten times higher than the concentrations of other volatile compounds detected in fish for both tuna and mahi-mahi

(Table 4-1 and Table 4-2 in Chapter 4). Trimethylamine (TMA) was significantly positively correlated with the volatile amine content detected by two types of colorimetric strip methods for both mahi-mahi (rose bengal: r=0.862; BPB: r=0.901) and tuna samples (rose bengal: r=0.596; BPB: r=0.484) (Table 5-4 and Table 5-5). TMA has a prominent fishy odor, and is known to accumulate with spoilage (Prester, 2011; Leduc et al., 2012; Ghaly et al., 2010). The level of TMA can be as high as 680 mg/kg in fish and TMA has been considered as an essential marker of microbial deterioration in fish

(WHO, 2006; Leduc et al., 2012; Bene et al., 2001; Ghaly et al., 2010; Chan et al., 2006;

Zhang et al., 2010; Olafsdottir et al., 2005). Dimethylamine, isobutylamine, 3- methylbutylamine, and 2-methylbutylamine are volatile biogenic amines with intense fishy odor and were also determined by the GC-MS method in this study.

Trimethylamine oxide (TMAO) is also the precursor of DMA and produces DMA by the endogenous enzymes during the process of fish spoilage (Ashie et al., 1996; Chan et al.,

2006; Ben-gigirey et al., 1999). Isobutylamine, 3-methylbutylamine, and 2- methylbutylamine are formed in spoiled fish by microbial decarboxylation of valine, leucine, and isoleucine, respectively (Gill et al., 1983; Gruger et al., 1972; Takahashi et al., 2004; Mayr and Schieberle, 2012; Zufall and Munger, 2016). The concentration of

DMA detected by GC-MS in poor grade of mahi-mahi (e.g. Grade 6 and 7) was ten times higher than those tuna samples (Table 4-1 and Table 4-2 in Chapter 4). Also isobutylamine, 3-methylbutylamine, and 2-methylbutylamine significantly increased in spoiled mahi-mahi but were not detected in tuna samples. So these four biogenic

102

amines were only significantly and positively correlated with the colorimetric strips results for mahi-mahi samples (Table 5-4 and Table 5-5).

For tuna samples, results of rose bengal strips and BPB strips were closely positively correlated with ethanol and other four aldehydes, including 2-methylpropanal,

2-methylbutanal, 3-methylbutanal, benzaldehyde, detected by GC-MS in this study

(Table 5-5). For mahi-mahi samples, results of both colorimetric strip methods were significantly positively correlated with isoamyl alcohol, four aldehydes (2-methylpropanal,

2-methylbutanal, 3-methylbutanal, benzaldehyde), ketones (acetone, 2-butanone), dimethyl disulfide; and were negatively correlated with 2-ethylhexanol and pyridine from the GC-MS data. The accumulation or decline of volatile compounds during fish spoilage lead to the close positive or negative correlations among results of colorimetric strip methods and levels of volatile compounds. The lipid oxidation and the deamination of amino acids under microbial catabolic activities have been well documented as the major reasons for the accumulation of 2-methylpropanal, 2-methylbutanal, 3- methylbutanal, benzaldehyde in spoiled fish (Joffraud et al., 2001; Shimoda et al., 1996;

Mace et al., 2013; Duflos et al., 2006; Leduc et al., 2012; Olafsdottir et al., 2005;

Jorgensen et al., 2001). These four aldehydes have low odor thresholds and contribute the fishy, musty, and almond odor to spoiled fish (Shimoda et al., 1996; Girard et al.,

2000). During the fish spoilage, changes of alcohols are mainly due to the microbial decomposition of fish and have effects on the organoleptic characteristic of the fish product (Duflos et al., 2006). The accumulation of ethanol and isoamyl alcohol have been reported in spoiled fish with fusel and alcoholic odor (Aro et al., 2003; Soncin et al.,

2008; Duflos et al., 2006). The alcohol 2-ethylhexanol is pleasant with fatty and fruity

103

odor and the decline of this compound has been observed in spoiled sea bass

(Dicentrarchus labrax) and sea bream (Sparus aurata) (Leduc and et al., 2012; Iglesias et al., 2009). Acetone and 2-butanone give off sweet, sharp odor and accumulate in spoiled fish as the results of the residual glycogen catabolism and fatty acid catabolism

(Joffraud et al., 2001). Significant increases of acetone and 2-butanone have been reported in spoiled fish, such as sea bream (Sparus aurata), cod (Gadus morhua), mackerel (Scomber scombrus), and influence the organoleptic qualities of fish (Soncin et al., 2008; Duflos et al., 2006; Alasalvar et al., 2005; Prost et al., 2004). Dimethyl disulfide with extremely low thresholds is generated in fish during the spoilage process due to the degradation of sulfur-containing amino acids, and peptides existing in fish tissue and releases strong sulfurous odor (Ashie et al., 1996). Because dimethyl disulfide was only detected in spoiled mahi-mahi sample and not in poor grades of tuna sample (Table 4-1 and Table 4-2 in Chapter 4), dimethyl disulfide only correlated with the colorimetric strip results for mahi-mahi.

Positive correlations between rose bengal results and BPB results with the histamine-specific ELISA results, and the histamine level detected by UHPLC were observed for mahi-mahi and tuna samples (Table 5-4 and Table 5-5). Histamine is reported as the major pathogenic substance responsible for fish poisoning and generates from free histidine in spoiled fish by the bacterial histamine decarboxylase.

Gram-negative enteric bacteria, which participate in the transformation of free histidine to histamine in fish tissue, are widely distributed in organisms that live in the saltwater environment and also present on the external surfaces and inside of live fish. The consumption of spoiled scombroid fish containing the high level of histamine can cause

104

allergy-like symptoms and lead to the highest incidence of illness from fish poisoning

(Morrow et al. 1991). Tuna and mahi-mahi contain high levels of free histidine, the precursor for histamine formation, as 7 g/kg and 5 g/kg respectively have been reported as two major sources of scombroid poisoning in the United States (Ahmed, 1991;

Antoine et al., 1999). From the ELISA results and the histamine level detected by

UHPLC, the maximum histamine level in poor grade tuna was around 1000 mg/kg higher than histamine level in spoiled mahi-mahi due to tuna containing larger amounts of histidine (Table 3-4 to Table 3-6 in Chapter 3). As a result, the Pearson correlation coefficients (r) between the colorimetric strips results and the histamine content detected by ELISA (rose bengal: r=0.9691; BPB: r=0.9802) and UHPLC (rose bengal: r=

0.967; BPB: r=0.9758) for tuna samples are higher than those for mahi-mahi samples

(Table 5-4 and Table 5-5).

For both tuna and mahi-mahi, cadaverine and putrescine detected by UHPLC were positively correlated with the results from colorimetric strip methods due to the accumulation of cadaverine and putrescine in spoiled fish. Cadaverine and putrescine are two biogenic amines generated from free lysine, and arginine respectively by bacterial decarboxylase activities in spoiled fish and can potentiate histamine toxicity

(Bulushi et al., 2009; Visciano et al., 2012; Prester, 2011). The accumulation of cadaverine and putrescine have been reported in spoiled fish, such as mahi-mahi, tuna, mackerel, herring, sardines, and are identified as significant indicators present signifying microbial contamination and reactions in fish products (Antoine et al., 2002; Rossi et al.,

2002; Prester et al., 2009; Mackie et al., 1997; Ozogul et al., 2006).

105

The results of rose bengal strips and BPB strips were negatively correlated with levels of free lysine, histidine, tyrosine and were positively correlated with free alanine detected by UHPLC for both tuna and mahi-mahi. The results of colorimetric strips were also significantly and positively correlated with levels of free glycine and alanine in tuna, and were negatively correlated with levels of free glutamic acid and isoleucine in mahi- mahi. Lysine, histidine, tyrosine, and isoleucine are precursors of cadaverine, histamine, tyramine and 2-methylbutylamine, respectively (Mayr and Schieberle, 2012; Prester,

2011; Hungerford, 2010). The negative correlations between these free amino acids with the colorimetric strips results are likely due to these amino acids transforming to biogenic amines by bacterial decarboxylase activities during the process of fish spoilage

(Prester, 2011; Ashie et al., 1996).

Summary

The rose bengal strips and BPB strips were two rapid, easily operated and low- cost colorimetric strip methods developed and optimized in this study for the determination of volatile biogenic amines in spoiled mahi-mahi and tuna. The optimized strip methods produced regression equations with good linearity and strip color change had better uniformity than previous similar studies. Rose bengal strips showed more sensitive color response for low concentrations of biogenic amines than the BPB strips.

Volatile biogenic amine levels in mahi-mahi and tuna detected by these two colorimetric strip methods were statistically significant and positively correlated with the spoilage grade of fish assigned by FDA/NMFS sensory experts. The correlation between colorimetric strip results and the results from the histamine-specific ELISA kit, UHPLC and GC-MS were likely influenced by the matrix of mahi-mahi and tuna. These two non- specific colorimetric strip methods optimized in this study successfully evaluated the

106

spoilage of mahi-mahi and tuna samples and could be valuable methods to monitor the quality of mahi-mahi and tuna.

107

Figure 5-1. Biogenic amine cocktail standard solutions reacted with rose bengal strips. Photo courtesy of author.

Figure 5-2. Biogenic amine cocktail standard solutions reacted with BPB strips. Photo courtesy of author.

108

Table 5-1. Linearity of rose bengal strips and BPB strips Color strip Regression equation Coefficient of Determination (r2) Rose Bengal Strip y=0.8206x-0.168 0.9535 BPB Strip y=1.477x-0.3453 0.8883

Table 5-2. Volatile biogenic amines in seven grades of mahi-mahi samples detected by rose bengal and BPB strips for fish samples (n=5) Grade Volatile Biogenic Volatile Biogenic Amines (ppm)-BPB Amines (ppm)-Rose bengal M1 1.752±1.02C 1.572±1.31F M2 1.749±1.33C 1.387±0.31F M3 3.101±2.56C 21.39±3.67D M4 0.1037±0.169D 6.391±2.91E M5 33.96±0.346B 25.33±1.35C M6 34.89±0.614B 30.58±2.33B M7 36.53±0.776A 38.62±4.97A

Table 5-3. Volatile biogenic amines in seven grades of tuna sample calculated by rose bengal and BPB strips for fish samples (n=5) Grade Volatile Biogenic Volatile Biogenic Amines (ppm)-BPB Amines (ppm)-Rose bengal T1 n.d. 1.015±0.846E T2 0.07973±0.178C 1.012±0.958E T3 n.d. 2.026±2.37DE T4 0.7554±0.497C 8.79±2.73C T5 0.7683±0.866C 6.965±1.52CD T6 28.23±5.32A 33.04±6.96A T7 17.94±4.06B 23.62±7.94B n.d. indicates not detected Different letters within the same column indicate significant differences according to an LSD means separation test (p<0.05).

109

Table 5-4. Pearson correlation coefficients (r) between methods for mahi-mahi (n=3) BPB strip Rose bengal strip r p-value r p-value Fish spoilage grade 0.8711 <.0001 0.8807 <.0001 Histamine from ELISA 0.6916 0.0005 0.4019 0.0709 Putrescine from UHPLC 0.817 <.0001 0.8189 <.0001 Cadaverine from UHPLC 0.9393 <.0001 0.8698 <.0001 Histamine from UHPLC 0.6927 0.0005 0.3988 0.0733 Glutamic acid from UHPLC -0.721 0.0002 -0.4689 0.0320 Alanine from UHPLC 0.8868 <.0001 0.6221 0.0026 Isoleucine from UHPLC -0.779 <.0001 -0.6967 0.0004 Lysine from UHPLC -0.5294 0.0136 -0.6348 0.002 Histidine from UHPLC -0.4983 0.0215 -0.2766 0.2249 Tyrosine from UHPLC -0.7874 0.0005 -0.8683 <.0001 DMA from GC 0.942 <.0001 0.857 <.0001 TMA from GC 0.901 <.0001 0.862 <.0001 Isobutylamine from GC 0.767 <.0001 0.5918 0.0047 3-methylbutylamine from GC 0.665 0.001 0.473 0.0304 2-methylbutylamine from GC 0.6746 0.0008 0.5032 0.0201 Acetone from GC 0.3986 0.0735 0.5189 0.0159 2-Methylpropanal from GC 0.506 0.0193 0.5129 0.0174 2-Butanone from GC 0.4373 0.0474 0.3325 0.1408 3-Methylbutanal from GC 0.7806 <.0001 0.8322 <.0001 2-Methylbutanal from GC 0.63 0.0022 0.8463 <.0001 Isoamyl alcohol from GC 0.5112 0.0179 0.6616 0.0011 Pyridine from GC -0.4999 0.021 -0.6656 0.001 Dimethyl disulfide from GC 0.4514 0.04 0.6048 0.0037 Benzaldehyde from GC 0.5727 0.0067 0.4503 0.0405 2-Ethylhexanol -0.3438 0.1271 -0.5836 0.0055 Only indicators significantly correlated with strips (p< 0.05) are reported here.

110

Table 5-5. Pearson correlation coefficients (r) between methods for tuna (n=3) BPB strip Rose bengal strip r p-value r p-value Fish spoilage grade 0.7362 0.0001 0.8351 <.0001 Histamine from ELISA 0.9802 <.0001 0.9691 <.0001 Putrescine from UHPLC 0.4705 0.0314 0.2434 0.2878 Cadaverine from UHPLC 0.9205 <.0001 0.9032 <.0001 Histamine from UHPLC 0.9758 <.0001 0.967 <.0001 Glycine from UHPLC 0.6471 0.0015 0.5728 0.0067 Alanine from UHPLC 0.7887 <.0001 0.7583 <.0001 Phenylalanine from UHPLC 0.5636 0.0078 0.6842 0.0006 Lysine from UHPLC -0.619 0.0028 -0.6797 0.0007 Histidine from UHPLC -0.8162 <.0001 -0.7568 <.0001 Tyrosine from UHPLC -0.5826 0.0227 -0.7412 0.0016 TMA from GC 0.484 0.026 0.596 0.0044 Ethanol from GC 0.5837 0.0055 0.6314 0.0021 2-methylpropanal from GC 0.752 <.0001 0.7607 <.0001 3-methylbutanal from GC from GC 0.7034 0.0004 0.7414 0.0001 2-methylbutanal from GC 0.7713 <.0001 0.7904 <.0001 Benzaldehyde from GC 0.2491 0.2762 0.3968 0.0749 Only indicators significantly correlated with strips (p< 0.05) are reported here.

111

CHAPTER 6 CONCLUSION

Tuna (Thunnus albacares) and mahi-mahi (Coryphaena hippurus) are two major fish species that lead to histamine poisoning in the United States, Italy, France, and

Spain due to the high level of histidine in their muscle. Currently, the spoilage grade of tuna and mahi-mahi is assigned by a subjective sensory method, which is conducted by

Food and Drug Administration (FDA)/National Marine Fisheries Service (NMFS) experts. However, human variation cannot be ignored in this subjective method. In the present research, three different analytical methods were developed and used to determine tuna and mahi-mahi qualify and safety.

A reversed phase UHPLC method for the simultaneous determination of amino acids (arginine, glutamic acid, glycine, alanine, phenylalanine, isoleucine, leucine, lysine, histidine, and tyrosine), histamine and other biogenic amines (cadaverine, putrescine, dimethylamine) that can act as co-indicators of histamine (scombroid) poisoning in tuna and mahi-mahi was developed and validated. Resolution of this method was improved, and the elution time was reduced to 17.5 minutes by using a

Kinetex® C18 column (50 x 2.1mm I.D, 1.3 µm particle size). The developed UHPLC method was validated by confirming linearity, LOD, LOQ, resolution, recovery, repeatability, the number of theoretical plates, and satisfactory results were obtained.

The developed rapid UHPLC method is an accurate method and is suitable to be adopted for inspecting the changes of amino acids and biogenic amines. The UHPLC technique is sophisticated and skilled technicians are needed to perform this method.

This method can be used as a confirmatory method by labs to check fish samples after a first screening test is performed.

112

A simplified and rapid purge and trap gas chromatography-mass spectrometry

(PT-GC-MS) method developed in this research was able to identify and quantify twenty aroma compounds in mahi-mahi and sixteen volatile compounds in yellowfin tuna associated with fish spoilage. The concentrations of the analytes injected into GC-MS instrument were enriched by applying the purge and trap system. The base-deactivated

Rtx-Volatile amine column applied in this PT-GC-MS could determine five amines, including dimethylamine, trimethylamine, isobutylamine, 3-methylbutylamine, 2- methylbutanamine, without any complicated derivatization procedure. The aroma compounds, including amines (dimethylamine, trimethylamine, isobutylamine, 3- methylbutylamine, and 2-methylbutanamine), alcohols (2-ethylhexanol, 1-penten-3-ol, isoamyl alcohol, ethanol), aldehydes (2-methylbutanal, 3-methylbutanal, benzaldehyde), ketones (acetone, 2,3-butanedione, 2-butanone, acetoin) and dimethyl disulfide, were identified as key spoilage indicators of tuna and mahi-mahi based on their close correlation with spoilage grade of fish. The rapid PT-GC-MS method developed in this study is an efficient analytical method for determining volatile profiles of fish samples in analytical labs of the seafood industry or the government. The identified quality markers can be used to monitor the spoilage level of tuna and mahi-mahi. Skilled technicians are required for this method as well.

An rose bengal strip, and on optimized BPB strip were optimized in this research by changing parameters of dye solution, headspace volume above samples, parameters of filter papers, etc. The two types of strips developed in this research were colorimetrically responsive to volatile amines produced by tuna and mahi-mahi, and were rapid, easily operated, and is a low-cost method to determine fish spoilage. These

113

colorimetric strips produced standard curves with satisfactory linearity and also change color uniformly when exposed to fish samples. The volatile amine levels in mahi-mahi and tuna detected by these two colorimetric strip methods significantly correlated with increasing spoilage grade of fish sample and were also closely correlated with the contents of biogenic amines and free amino acids obtained by the developed UHPLC, histamine level detected by ELISA, and the aroma compounds identified as spoiled indicators of mahi-mahi and tuna. These correlations prove that the rose bengal strips and BPB strips can be applied to the routine analysis by seafood industry or the government for the quality and safety assessment of mahi-mahi and tuna and could replace the current olfactory method applied for grading fish spoilage. The bromophenol blue reagent and rose bengal reagent are usually under ten dollar per gram. Skilled technicians are not required for this method due to its simple principle.

In the further research, the rose bengal strip acidification process should be simplified. In the meantime, the two color strips developed in current research are needed to be used within 12 hours after preparing the strips. The future study should investigate different methods to extend the expiration period of these two colorimetric assays and make them shelf stable.

114

APPENDIX EXTERNAL STANDARD PREPARATION FOR UHPLC METHOD

Table A-1. Concentrations of each amino acid in five levels of external standard solutions and stock solution, mg/L solution. Compound name Level 1 Level 2 Level 3 Level 4 Level 5/ Stock solution Arginine 7.186 14.37 71.86 143.7 287.4 Glutamic acid 6.069 12.14 60.69 121.4 242.8 Glycine 3.094 6.188 30.94 61.9 123.8 Alanine 3.671 7.343 36.71 73.4 146.9 Phenylalanine 6.815 13.63 68.15 136.3 272.6 Isoleucine 5.411 10.82 54.11 108.2 216.4 Leucine 5.411 10.82 54.11 108.2 216.4 Lysine 6.03 12.06 60.3 120.6 241.2 Histidine 6.4 12.8 64 128 256 Tyrosine 7.474 14.95 74.74 149.5 299

115

LIST OF REFERENCES

Ababouch, L., Afilal, M. E., Benabdeljelil, H., & Busta, F. F. (1991). Quantitative changes in bacteria, amino-acids and biogenic-amines in sardine (Sardina- pilchardus) stored at ambient-temperature (25-28 °C) and in ice. International Journal of Food Science and Technology, 26(3), 297-306.

Abalos, M., Mas, R., Suc, V., & Bayona, J. M. (2001). Evaluation of capillary gas chromatography columns for the determination of free volatile amines after solid- phase microextraction. Chromatographia, 54(1-2), 109-113.

Abraham, M. H., Poole, C. F., & Poole, S. K. (1999). Classification of stationary phases and other materials by gas chromatography. Journal of Chromatography A, 842(1-2), 79-114.

Ahmed, F. E. (1991). Naturally-occurring seafood toxins. Journal of Toxicology-Toxin Reviews, 10(3), 263-287.

Akerlind, C., Arwin, H., Jakobsson, F. L. E., Kariis, H., & Jarrendahl, K. (2011). Optical properties and switching of a Rose Bengal derivative: A spectroscopic ellipsometry study. Thin Solid Films, 519(11), 3582-3586.

Alasalvar, C., Taylor, K. D. A., & Shahidi, F. (2005). Comparison of volatiles of cultured and wild sea bream (Sparus aurata) during storage in ice by dynamic headspace analysis gas chromatography mass spectrometry. Journal of Agricultural and Food Chemistry, 53(7), 2616-2622.

Alexi, N., Fountoulaki, E., & Grigorakis, K. (2017). Quality of reared gilthead sea bream (Sparus aurata) during ice storage, as affected by dietary fish oil substitution; an instrumental and sensory designation approach. Aquaculture Research, 48(7), 3817-3828.

Amin, A. S., & Elhenawee, M. M. (1995). Colorimetric method for the simultaneous determination of chlorphenoxamine hydrochloride and anhydrous caffeine in pure and dosage forms with rose-bengal. Mikrochimica Acta, 118(3-4), 177-183.

Antoine, F. R., Wei, C. I., Littell, R. C., & Marshall, M. R. (1999). HPLC method for analysis of free amino acids in fish using o-phthaldialdehyde precolumn derivatization. Journal of Agricultural and Food Chemistry, 47(12), 5100-5107.

Antoine, F. R., Wei, C. I., Otwell, W. S., Sims, C. A., Littell, R. C., Hogle, A. D., & Marshall, M. R. (2002). Analysis of biogenic amines and their precursor free amino acids in mahi-mahi (Coryphaena hippurus). Journal of Food Biochemistry, 26(2), 131-152.

116

Antoine, F. R., Wei, C. I., Otwell, W. S., Sims, C. A., Littell, R. C., Hogle, A. D., & Marshall, M. R. (2002). Gas chromatographic analysis of histamine in mahi-mahi (Coryphaena hippurus). Journal of Agricultural and Food Chemistry, 50(17), 4754-4759.

Aro, T., Tahvonen, R., Koskinen, L., & Kallio, H. (2003). Volatile compounds of Baltic herring analysed by dynamic headspace sampling-gas chromatography-mass spectrometry. European Food Research and Technology, 216(6), 483-488.

Ashie, I. N. A., Smith, J. P., & Simpson, B. K. (1996). Spoilage and shelf-life extension of fresh fish and shellfish. Critical Reviews in Food Science and Nutrition, 36(1- 2), 87-121.

Awan, M. A., Fleet, I., & Thomas, C. L. P. (2008). Optimising cell temperature and dispersion field strength for the screening for putrescine and cadaverine with thermal desorption-gas chromatography-differential mobility spectrometry. Analytica Chimica Acta, 611(2), 226-232.

Atwood, J. L., Gokel, G. W., Barbour, L. J., Rissanen, K., Jayawickramarajah, J., Wilson, A. J., Dalgarno, S., MacGillivray, R., Glass, E., Raston, C. (2017). Comprehensive supramolecular chemistry II. Amsterdam: Elsevier.

Auerswald, L., Morren, C., & Lopata, A. L. (2006). Histamine levels in seventeen species of fresh and processed South African seafood. Food Chemistry,98(2), 231-239.

Beltran, J., Serrano, E., Lopez, F. J., Peruga, A., Valcarcel, M., & Rosello, S. (2006). Comparison of two quantitative GC-MS methods for analysis of tomato aroma based on purge-and-trap and on solid-phase microextraction. Analytical and Bioanalytical Chemistry, 385(7), 1255-1264.

Ben-gigirey, B., De Sousa, J., Villa, T. G., & Barros-velazquez, J. (1999). Chemical changes and visual appearance of albacore tuna as related to frozen storage. Journal of Food Science, 64(1), 20-24.

Bene, A., Fornage, A., Luisier, J. L., Pichler, P., & Villettaz, J. C. (2001). A new method for the rapid determination of volatile substances: the SPME-direct method - Part I: Apparatus and working conditions. Sensors and Actuators B-Chemical, 72(2), 184-187.

Bianchi, F., Careri, M., Musci, M., & Mangia, A. (2007). Fish and food safety: Determination of formaldehyde in 12 fish species by SPME extraction and GC- MS analysis. Food Chemistry, 100(3), 1049-1053.

117

Bonilla, M., Enriquez, L. G., & McNair, H. M. (1997). Use of cold on-column injection for the analysis of putrescine and cadaverine by gas chromatography. Journal of Chromatographic Science, 35(2), 53-56.

Bulushi, I. A., Poole, S., Deeth, H., & Dykes, G. (2009a). Biogenic Amines in Fish: Roles in Intoxication, Spoilage, and Nitrosamine FormationA Review. Critical Reviews in Food Science and Nutrition, 49(4), 369-377.

Busto, O., Miracle, M., Guasch, J., & Borrull, F. (1997). Determination of biogenic amines in wines by high-performance liquid chromatography with on-column fluorescence derivatization. Journal of Chromatography A, 757(1-2), 311-318.

Cai, Y. P., Sun, Z. W., Chen, G., Liu, X. M., You, J. M., & Zhang, C. Q. (2016). Rapid analysis of biogenic amines from rice wine with isotope-coded derivatization followed by high performance liquid chromatography-tandem mass spectrometry. Food Chemistry, 192, 388-394.

Cartwright, S. J. (2016). Solvatochromc dyes detect the presence of homeopathic potencies. Homeopathy, 105(1), 55-65.

Chan, S. T., Yao, M. W. Y., Wong, Y. C., Wong, T., Mok, C. S., & Sin, D. W. M. (2006). Evaluation of chemical indicators for monitoring freshness of food and determination of volatile amines in fish by headspace solid-phase microextraction and gas chromatography-mass spectrometry. European Food Research and Technology, 224(1), 67-74.

Chattopadhyaya, M., Murugan, N. A., & Rinkevicius, Z. (2016). Origin of the Absorption Band of Bromophenol Blue in Acidic and Basic pH: Insight from a Combined Molecular Dynamics and TD-DFT/MM Study. Journal of Physical Chemistry A, 120(36), 7175-7182.

Cheret, R., Delbarre-Ladrat, C., de Lamballerie-Anton, M., & Verrez-Bagnis, V. (2007). Calpain and cathepsin activities in post mortem fish and meat muscles. Food Chemistry, 101(4), 1474-1479.

Chong, C. Y., Abu Bakar, F., Rahman, R. A., Bakar, J., & Zaman, M. Z. (2014). Biogenic amines, amino acids and microflora changes in Indian mackerel (Rastrellinger kanagurta) stored at ambient (25-29 A degrees C) and ice temperature (0 A degrees C). Journal of Food Science and Technology-Mysore, 51(6), 1118-1125.

Corradini, D., & Phillips, T. M. (2011). Handbook of HPLC. Boca Raton, FL: CRC Press.

Clancy, S., Beames, R., Higgs, D., Dosanjh, B., Haard, N., & Toy, B. (1995). Influence of spoilage and processing temperature on the quality of marine fish protein sources for salmonids. Aquaculture Nutrition, 1(3), 169-177.

118

Crowther, J. R. (2000). The ELISA guidebook. Totowa, NJ: Humana.

Dadakova, E., Krizek, M., & Pelikanova, T. (2009). Determination of biogenic amines in foods using ultra-performance liquid chromatography (UPLC). Food Chemistry, 116(1), 365-370. de Zeeuw, J., Stricek, R., & Stidsen, G. (2011). An Advanced Base-Deactivated Capillary Column for Analysis of Volatile Amines, Ammonia, and Alcohols. American Laboratory, 43(2), 37-40.

Dole, T., Koltun, S., Baker, S. M., Goodrich-Schneider, R. M., Marshall, M. R., & Sarnoski, P. J. (2016). The matrix effect of tuna and mahi-mahi on biogenic amine detection. Journal of Food and Health Science,74-81.

Dole, T., Koltun, S., Baker, S. M., Goodrich-Schneider, R. M., Marshall, M. R., & Sarnoski, P. J. (2017). Colorimetric Evaluation of Mahi-Mahi and Tuna for Biogenic Amines. Journal of Aquatic Food Product Technology, 26(7), 781-789.

Du, W., Huang, T., Kim, J., Marshall, M. R., & Wei, C. (2001). Chemical, Microbiological, and AromaScan Evaluation of Mahi-mahi Fillets under Various Storage Conditions. Journal of Agricultural and Food Chemistry,49(1), 527-534.

Duflos, G., Coin, V. M., Cornu, M., Antinelli, J. F., & Mallel, P. (2006). Determination of volatile compounds to characterize fish spoilage using headspace/mass spectrometry and solid-phase microextraction/gas chromatography/mass spectrometry. Journal of the Science of Food and Agriculture, 86(4), 600-611.

Duflos, G., Leduc, F., N'Guessan, A., Krzewinski, F., Kol, O., & Malle, P. (2010). Freshness characterisation of whiting (Merlangius merlangus) using an SPME/GC/MS method and a statistical multivariate approach. Journal of the Science of Food and Agriculture, 90(15), 2568-2575.

Duflos, G., Moine, F., Coin, V. M., & Malle, P. (2005). Determination of volatile compounds in whiting (Merlangius merlangus) using headspace-solid-phase microextraction-gas chromatography-mass spectrometry. Journal of Chromatographic Science, 43(6), 304-312.

Edirisinghe, R. K. B., Graham, A. J., & Taylor, S. J. (2007). Characterisation of the volatiles of yellowfin tuna (Thunnus albacares) during storage by solid phase microextraction and GC-MS and their relationship to fish quality parameters. International Journal of Food Science and Technology, 42(10), 1139-1147.

Eenoo, P. V., Delbeke, F., Roels, K., & Backer, P. D. (2001). Simultaneous quantitation of ephedrines in urine by gas chromatography–nitrogen–phosphorus detection for doping control purposes. Journal of Chromatography B: Biomedical Sciences and Applications,760(2), 255-261.

119

Emborg, J., & Dalgaard, P. (2006). Formation of histamine and biogenic amines in cold- smoked tuna: An investigation of psychrotolerant bacteria from samples implicated in cases of histamine fish poisoning. Journal of Food Protection, 69(4), 897-906.

Erim, F. B. (2013). Recent analytical approaches to the analysis of biogenic amines in food samples. Trac-Trends in Analytical Chemistry, 52, 239-247.

Eskin, N. A., & Shahidi, F. (2014). Biochemistry of Foods. Saint Louis: Elsevier Science.

FAO (2010) Fisheries and aquaculture topics: Food security and fisheries. Topics Fact Sheets. Text by Peter Manning. Fisheries and Aquaculture Department, Rome.

Flores, R. (1978). Rapid and reproducible assay for quantitative estimation of proteins using bromophenol blue. Analytical Biochemistry, 88(2), 605-611.

Food and Drug Administration (FDA). (2011). Fish and Fishery Products Hazards and Controls Guidance, 4th Edn. Washington, DC: Department of Health and Human Services, Food and Drug Administration, Center for Food Safety and Applied Nutrition.

Frank, H. A., Baranowski, J. D., Chongsiriwatana, M., Brust, P. A., & Premaratne, R. J. (1985). Identification and decarboxylase activities of bacteria isolated from decomposed mahimahi (coryphaena-hippurus) after incubation at 0 °C and 32 °C. International Journal of Food Microbiology, 2(6), 331-340.

Fraser, O., & Sumar, S. (1998). Compositional changes and spoilage in fish ‐ an introduction. Nutrition & Food Science,98(5), 275-279.

Fukami, K., Ishiyama, S., Yaguramaki, H., Masuzawa, T., Nabeta, Y., Endo, K., & Shimoda, M. (2002). Identification of distinctive volatile compounds in fish sauce. Journal of Agricultural and Food Chemistry, 50(19), 5412-5416.

Genigeorgis, C. A. (1985). Microbial and safety implications of the use of modified atmospheres to extend the storage life of fresh meat and fish. International Journal of Food Microbiology, 1(5), 237-251.

Ghaly, A. E., Dave, D., Budge, S., Brooks, M.S. (2010). "Fish Spoilage Mechanisms and Preservation Techniques: Review." American Journal of Applied Sciences 7.7 (2010): 859-77. Print.

Globefish highlights: A quarterly update on world seafood markets. (2017). Rome: Food and Agriculture Organization of the United Nations.

120

Gill, C. O. (1983). Meat spoilage and evaluation of the potential storage life of fresh meat. Journal of Food Protection, 46(5), 444-452.

Girard, B., & Durance, T. (2000). Headspace volatiles of sockeye and pink salmon as affected by retort process. Journal of Food Science, 65(1), 34-39.

Gram, L., & Dalgaard, P. (2002). Fish spoilage bacteria - problems and solutions. Current Opinion in Biotechnology, 13(3), 262-266.

Gram, L., & Huss, H. H. (1996). Microbiological spoilage of fish and fish products. International Journal of Food Microbiology, 33(1), 121-137.

Grimm, C. C., Lloyd, S. W., Batista, R., & Zimba, P. V. (2000). Using microwave distillation-solid-phase microextraction-gas chromatography-mass spectrometry for analyzing fish tissue. Journal of Chromatographic Science, 38(7), 289-296.

Gruger, E. H. (1972). Chromatographic analyses of volatile amines in marine fish. Journal of Agricultural and Food Chemistry, 20(4), 781-&.

Gunlu, A. and Gunlu, N. (2014). Taste activity value, free amino acid content and proximate composition of Mountain trout (Salmo trutta macrostigma Dumeril, 1858) muscles. Iranian Journal of Fisheries Sciences, 13(1)58-72

Hansen, L. T., Gill, T., & Huss, H. H. (1995). Effects of salt and storage-temperature on chemical, microbiological and sensory changes in cold-smoked salmon. Food Research International, 28(2), 123-130.

Harris, D. C., & Lucy, C. A. (2017). Quantitative chemical analysis. New York, NY: Macmillan Higher Education / Worth Publ.

Hayes, R., Ahmed, A., Edge, T., & Zhang, H. F. (2014). Core-shell particles: Preparation, fundamentals and applications in high performance liquid chromatography. Journal of Chromatography A, 1357, 36-52.

He, Y. R., Zhao, X. E., Wang, R. J., Wei, N., Sun, J., Dang, J., Chen, G., Liu, Z.Q., Zhu, S.Y., You, J. M. (2016). Simultaneous Determination of Food-Related Biogenic Amines and Precursor Amino Acids Using in Situ Derivatization Ultrasound- Assisted Dispersive Liquid-Liquid Microextraction by Ultra-High-Performance Liquid Chromatography Tandem Mass Spectrometry. Journal of Agricultural and Food Chemistry, 64(43),

Himelbloom, B. H., Oliveira, A. C. M., Chantarachoti, J., & Crapo, C. A. (2013). Ethanol development in tissues of spoiling whole pink salmon (oncorhynchus gorbuscha). Journal of Food Quality, 36(4), 263-268.

121

Henari, F. Z., Al-Saie, A., & Culligan, K. G. (2012). Optical limiting behavior of bromophenol blue and its dependence on pH. Journal of Nonlinear Optical Physics & Materials, 21(1), 9.

Herbert, R. A., & Shewan, J. M. (1975). Precursors of volatile sulfides in spoiling north- sea cod (gadus-morhua). Journal of the Science of Food and Agriculture, 26(8), 1195-1202.

Hsieh, T. C. Y., Williams, S. S., Vejaphan, W., & Meyers, S. P. (1989). Characterization of volatile components of menhaden fish (brevoortia-tyrannus) oil. Journal of the American Oil Chemists Society, 66(1), 114-117.

Hungerford, J., & Wu, W. H. (2012). Comparison study of three rapid test kits for histamine in fish: BiooScientific MaxSignal enzymatic assay, Neogen Veratox ELISA, and the Neogen Reveal Histamine Screening test. Food Control, 25(2), 448-457.

Hungerford, J. M. (2010). Scombroid poisoning: A review. Toxicon, 56(2), 231-243.

Hwang, D. F., Chang, S. H., Shiua, C. Y., & Chai, T. J. (1997). High-performance liquid chromatographic determination of biogenic amines in fish implicated in food poisoning. Journal of Chromatography B, 693(1), 23-29.

Iglesias, J., & Medina, I. (2008). Solid-phase microextraction method for the determination of volatile compounds associated to oxidation of fish muscle. Journal of Chromatography A, 1192(1), 9-16.

Iglesias, J., Medina, I., Bianchi, F., Careri, M., Mangia, A., & Musci, M. (2009). Study of the volatile compounds useful for the characterisation of fresh and frozen-thawed cultured gilthead sea bream fish by solid-phase microextraction gas chromatography-mass spectrometry. Food Chemistry, 115(4), 1473-1478.

Jimenez-Toribio, R., Guillotreau, P., & Mongruel, R. (2010). Global integration of European tuna markets. Progress in Oceanography,86(1-2), 166-175.

Jia, S., Kang, Y. P., Park, J. H., Lee, J., & Kwon, S. W. (2012). Determination of biogenic amines in Bokbunja (Rubus coreanus Miq.) wines using a novel ultra- performance liquid chromatography coupled with quadrupole-time of flight mass spectrometry. Food Chemistry, 132(3), 1185-1190.

Jiang, J. X., Li, Y. Y., Wu, X. F., Xiao, J. L., Adams, D. J., & Cooper, A. I. (2013). Conjugated Microporous Polymers with Rose Bengal Dye for Highly Efficient Heterogeneous Organo-Photocatalysis. Macromolecules, 46(22), 8779-8783.

122

Joffraud, J. J., Leroi, F., Roy, C., & Berdague, J. L. (2001). Characterisation of volatile compounds produced by bacteria isolated from the spoilage flora of cold-smoked salmon. International Journal of Food Microbiology, 66(3), 175-184.

Jonsdottir, R., Olafsdottir, G., Chanie, E., & Haugen, J. E. (2008). Volatile compounds suitable for rapid detection as quality indicators of cold smoked salmon (Salmo salar). Food Chemistry, 109(1), 184-195.

Jorgensen, L. V., Huss, H. H., & Dalgaard, P. (2001). Significance of volatile compounds produced by spoilage bacteria in vacuum-packed cold-smoked salmon (Salmo salar) analyzed by GC-MS and multivariate regression. Journal of Agricultural and Food Chemistry, 49(5), 2376-2381.

Kawai, T. (1996). Fish flavor. Critical Reviews in Food Science and Nutrition, 36(3), 257-298.

Kose, S., Kaklikkaya, N., Koral, S., Tufan, B., Buruk, K. C., & Aydin, F. (2011). Commercial test kits and the determination of histamine in traditional (ethnic) fish products-evaluation against an EU accepted HPLC method. Food Chemistry, 125(4), 1490-1497.

Kose, S., Koral, S., Tufan, B., Pompe, M., Scavnicar, A., & Kocar, D. (2012). Biogenic amine contents of commercially processed traditional fish products originating from European countries and Turkey. European Food Research and Technology, 235(4), 669-683.

Koutsoumanis, K., & Nychas, G. J. E. (1999). Chemical and sensory changes associated with microbial flora of Mediterranean boque (Boops boops) stored aerobically at 0, 3, 7, and 10 degrees C. Applied and Environmental Microbiology, 65(2), 698-706.

Kuchmenko, T. A., and Mishina, A. A. (2011). "Sorption Specifics of Volatile Amines on Thin Films of Acid-Base Indicators." Journal of Analytical Chemistry, 66.8: 701-08. Print.

Kuswandi, B., Maryska, C., Jayus, Abdullah, A., & Heng, L. Y. (2013). Real time on- package freshness indicator for guavas packaging. Journal of Food Measurement and Characterization, 7(1), 29-39.

Ladero, V., Calles-Enriquez, M., Fernandez, M., & Alvarez, M. A. (2010). Toxicological Effects of Dietary Biogenic Amines. Current Nutrition & Food Science,6(2), 145- 156.

Lamberts, J. J. M., & Neckers, D. C. (1984). Rose-bengal and non-polar derivatives - the birth of dye sensitizers for photooxidation. Zeitschrift Fur Naturforschung Section B-a Journal of Chemical Sciences, 39(4), 474-484.

123

Lambropoulou, D. A., & Albanis, T. A. (2007). Methods of sample preparation for determination of pesticide residues in food matrices by chromatography-mass spectrometry-based techniques: a review. Analytical and Bioanalytical Chemistry, 389(6), 1663-1683.

Leduc, F., Krzewinski, F., Le Fur, B., N'Guessan, A., Malle, P., Kol, O., & Duflos, G. (2012). Differentiation of fresh and frozen/thawed fish, European sea bass (Dicentrarchus labrax), gilthead seabream (Sparus aurata), cod (Gadus morhua) and salmon (Salmo salar), using volatile compounds by SPME/GC/MS. Journal of the Science of Food and Agriculture, 92(12), 2560-2568.

Leduc, F., Tournayre, P., Kondjoyan, N., Mercier, F., Malle, P., Kol, O., Berdague, J., Duflos, G. (2012). Evolution of volatile odorous compounds during the storage of European seabass (Dicentrarchus labrax). Food Chemistry, 131(4), 1304-1311.

Lee, C. S., & Ostrowski, A. C. (2001). Current status of marine finfish larviculture in the United States. Aquaculture, 200(1-2), 89-109.

Lehane, L., & Olley, J. (2000). Histamine fish poisoning revisited. International Journal of Food Microbiology, 58(1-2), 1-37.

Loughran, M., & Diamond, D. (2000). Monitoring of volatile bases in fish sample headspace using an acidochromic dye. Food Chemistry, 69(1), 97-103.

Lucentini, L., Ferretti, E., Veschetti, E., Sibio, V., Citti, G., & Ottaviani, M. (2005). Static headspace and purge-and-trap gas chromatography for epichlorohydrin determination in drinking water. Microchemical Journal, 80(1), 89-98.

Lupo, A., & Mozola, M. (2011). Validation study of a rapid ELISA for detection of histamine in tuna. Journal of Aoac International, 94(3), 886-899.

Mace, S., Joffraud, J. J., Cardinal, M., Malcheva, M., Cornet, J., Lalanne, V., Chevalier, F., Serot, T., Pilet, MF., Dousset, X. (2013). Evaluation of the spoilage potential of bacteria isolated from spoiled raw salmon (Salmo salar) fillets stored under modified atmosphere packaging. International Journal of Food Microbiology, 160(3), 227-238.

Mackie, I. M., Pirie, L., Ritchie, A. H., & Yamanaka, H. (1997). The formation of non- volatile amines in relation to concentrations of free basic amino acids during postmortem storage of the muscle of scallop (Pecten maximus), herring (Clupea harengus) and mackerel (Scomber scombrus). Food Chemistry, 60(3), 291-295.

124

Mahmoud, M. A. A., & Buettner, A. (2017). Characterisation of aroma-active and off- odour compounds in German rainbow trout (Oncorhynchus mykiss). Part II: Case of fish meat and skin from earthen-ponds farming. Food Chemistry, 232, 841- 849.

Matsumoto, M., & Yamanaka, H. (1990). Postmortem biochemical-changes in the muscle of kuruma prawn during storage and evaluation of the freshness. Nippon Suisan Gakkaishi, 56(7), 1145-1149.

Mayr, C. M., & Schieberle, P. (2012). Development of Stable Isotope Dilution Assays for the Simultaneous Quantitation of Biogenic Amines and Polyamines in Foods by LC-MS/MS. Journal of Agricultural and Food Chemistry, 60(12), 3026-3032.

Mazzucco, E., Gosetti, F., Bobba, M., Marengo, E., Robotti, E., & Gennaro, M. C. (2010). High-Performance Liquid Chromatography-Ultraviolet Detection Method for the Simultaneous Determination of Typical Biogenic Amines and Precursor Amino Acids. Applications in Food Chemistry. Journal of Agricultural and Food Chemistry, 58(1), 127-134.

Metian, M., Pouil, S., Boustany, A., & Troell, M. (2014). Farming of bluefin tuna- reconsidering global estimates and sustainability concerns. Reviews in Fisheries Science & Aquaculture, 22(3), 184-192.

Miller, D. W., Wilkes, J. G., & Conte, E. D. (2006). U.S. Patent No. US7014816B2. Washington, DC: U.S. Patent and Trademark Office.

Mills, A., Wild, L., & Chang, Q. (1995). Plastic colorimetric film sensors for gaseous ammonia. Mikrochimica Acta,121(1-4), 225-236.

Miyake, M., Guillotreau, P., Sun, C-H., & Ishimura, G. (2010). Recent developments in the tuna industry: Stocks, fisheries, management, processing, trade and markets. Rome: FAO.

Mohamed, R., Livia, S. S., Hassan, S., Soher, E. S., & Ahmed-Adel, E. B. (2009). Changes in free amino acids and biogenic amines of Egyptian salted-fermented fish (Feseekh) during ripening and storage. Food Chemistry, 115(2), 635-638.

Moldoveanu, S. C., & David, V. (2017). Selection of the HPLC method in chemical analysis. Amsterdam; Boston ; Heidelberg ; London ; New York: Elsevier.

Morrow, J. D., Margolies, G. R., Rowland, J., & Roberts, L. J. (1991). Evidence that histamine is the causative toxin of scombroid-fish poisoning. New England Journal of Medicine, 324(11), 716-720.

Murray, J. M., Delahunty, C. M., & Baxter, I. A. (2001). Descriptive sensory analysis: past, present and future. Food Research International, 34(6), 461-471.

125

Nakano, N., Kobayashi, Y., & Nagashima, K. (1994). Development of a monitoring tape for ammonia gas in air using rose-bengal. Analyst, 119(9), 2009-2012.

Office of Regulatory Affairs. "Compliance Policy Guides - CPG Sec. 540.525 Decomposition and Histamine Raw, Frozen Tuna and Mahi-Mahi; Canned Tuna; and Related Species." U S Food and Drug Administration Home Page. Office of Regulatory Affairs.

Ocano-Higuera, V. M., Marquez-Rios, E., Canizales-Davila, M., Castillo-Yanez, F. J., Pacheco-Aguilar, R., Lugo-Sanchez, M. E., Garcia-Orozco, K.D., Graciano- Verdugo, A. Z. (2009). Postmortem changes in cazon fish muscle stored on ice. Food Chemistry, 116(4), 933-938.

Olafsdottir, G., Jonsdottir, R., Lauzon, H. L., Luten, J., & Kristbergsson, K. (2005). Characterization of volatile compounds in chilled cod (Gadus morhua) fillets by gas chromatography and detection of quality indicators by an electronic nose. Journal of Agricultural and Food Chemistry, 53(26), 10140-10147.

Olafsdottir, G., Martinsdottir, E., Oehlenschlager, J., Dalgaard, P., Jensen, B., Undeland, I., Mackie, I.M., Henehan, G., Nilsen, H. (1997). Methods to evaluate fish freshness in research and industry. Trends in Food Science & Technology, 8(8), 258-265.

Onal, A. (2007). A review: Current analytical methods for the determination of biogenic amines in foods. Food Chemistry, 103(4), 1475-1486.

Ostrowski, A. C., & Divakaran, S. (1989). The amino-acid and fatty-acid compositions of selected tissues of the dolphin fish (coryphaena-hippurus) and their nutritional implications. Aquaculture, 80(3-4), 285-299.

Ozogul, F., & Ozogul, Y. (2006). Biogenic amine content and biogenic amine quality indices of sardines (Sardina pilchardus) stored in modified atmosphere packaging and vacuum packaging. Food Chemistry, 99(3), 574-578.

Paczkowski, J. (1985). Photophysical properties of rose bengal and its derivatives (XII). Advances in Free Radical Biology & Medicine,1(5-6), 341-351. doi:10.1016/8755-9668(85)90075-4

Peng, J. F., Fang, K. T., Me, D. H., Ding, B., Yin, J. Y., Cui, X. M., Zhang, Y., Liu, J. F. (2008). Development of an automated on-line pre-column derivatization procedure for sensitive determination of histamine in food with high-performance liquid chromatography-fluorescence detection. Journal of Chromatography A, 1209(1-2), 70-75.

126

Peng, S., Chen, C., Shi, Z, Wang, L. (2013). Amino Acid and Fatty Acid Composition of the Muscle Tissue of Yellowfin Tuna (Thunnus Albacares) and Bigeye Tuna (Thunnus Obesus). Journal of Food and Nutrition Research, Vol. 1, No. 4, 42-45

Pivarnik, L., Ellis, P., Wang, X., & Reilly, T. (2001). Standardization of the ammonia electrode method for evaluating seafood quality by correlation to sensory analysis. Journal of Food Science, 66(7), 945-952.

Povolo, M., & Contarini, G. (2003). Comparison of solid-phase microextraction and purge-and-trap methods for the analysis of the volatile fraction of butter. Journal of Chromatography A, 985(1-2), 117-125.

Prester, L. (2011). Biogenic amines in fish, fish products and shellfish: a review. Food Additives and Contaminants Part a-Chemistry Analysis Control Exposure & Risk Assessment, 28(11), 1547-1560.

Prester, L., Macan, J., Varnai, V. M., Orct, T., Vukusic, J., & Kipcic, D. (2009). Endotoxin and biogenic amine levels in Atlantic mackerel (Scomber scombrus), sardine (Sardina pilchardus) and Mediterranean hake (Merluccius merluccius) stored at 22 degrees C. Food Additives and Contaminants Part a-Chemistry Analysis Control Exposure & Risk Assessment, 26(3), 355-362.

Prost, C., Hallier, A., Cardinal, M., Serot, T., & Courcoux, P. (2004). Effect of storage time on raw sardine (Sardina pilchardus) flavor and aroma quality. Journal of Food Science, 69(5), S198-S204.

Rakow, N. A., Sen, A., Janzen, M. C., Ponder, J. B., & Suslick, K. S. (2005). Molecular Recognition and Discrimination of Amines with a Colorimetric Array. Angewandte Chemie,117(29), 4604-4608.

Rehbein, H., & Oehlenschlager, J. (2009). Fishery products: Quality, safety and authenticity. Chichester (UK): Wiley-Blackwell.

Redruello, B., Ladero, V., Cuesta, I., Alvarez-Buylla, J. R., Martin, M. C., Fernandez, M., & Alvarez, M. A. (2013). A fast, reliable, ultra high performance liquid chromatography method for the simultaneous determination of amino acids, biogenic amines and ammonium ions in cheese, using diethyl ethoxymethylenemalonate as a derivatising agent. Food Chemistry, 139(1-4), 1029-1035.

Reichardt, C. (1994). Solvatochromic dyes as solvent polarity indicators. Chemical Reviews, 94(8), 2319-2358.

Rossi, S., Lee, C., Ellis, P. C., & Pivarnik, L. F. (2002). Biogenic amines formation in Bigeye tuna steaks and whole Skipjack tuna. Journal of Food Science, 67(6), 2056-2060.

127

Ruiz-Capillas, C., & Horner, W. F. A. (1999). Determination of trimethylamine nitrogen and total volatile basic nitrogen in fresh fish by flow injection analysis. Journal of the Science of Food and Agriculture, 79(14), 1982-1986.

Ruiz-Capillas, C., & Moral, A. (2004). Free amino acids and biogenic amines in red and white muscle of tuna stored in controlled atmospheres. Amino Acids, 26(2), 125- 132.

Ruth, J. H. (1986). Odor thresholds and irritation levels of several chemical-substances - a review. American Industrial Hygiene Association Journal, 47(3), A142-A156.

Rogers, P. L., and W. Staruszkiewicz (1997). "Gas Chromatographic Method for Putrescine and Cadaverine in Canned Tuna and Mahimahi and Fluorometric Method for Histamine (Minor Modification of Aoac Official Method 977.13): Collaborative Study." Journal of Aoac International 80.3: 591-602. Print.

Benford, D.J., DiNovi, M., Kawamura, Y., Leblanc, J.C., Schlatter, J.R. (2012). Safety evaluation of certain food additives. Geneva: World Health Organization.

Sagratini, G., Fernandez-Franzon, M., De Berardinis, F., Font, G., Vittori, S., & Manes, J. (2012). Simultaneous determination of eight underivatised biogenic amines in fish by solid phase extraction and liquid chromatography-tandem mass spectrometry. Food Chemistry, 132(1), 537-543.

Sandra, P., Tienpont, B., & David, F. (2003). Multi-residue screening of pesticides in vegetables, fruits and baby food by stir bar sorptive extraction-thermal desorption-capillary gas chromatography-mass spectrometry. Journal of Chromatography A, 1000(1-2), 299-309.

Salazar, M. T., Smith, T. K., & Harris, A. (2000). High-performance liquid chromatographic method for determination of biogenic amines in feedstuffs, complete feeds, and animal tissues. Journal of Agricultural and Food Chemistry, 48(5), 1708-1712.

Serrar, D., Brebant, R., Bruneau, S., & Denoyel, G. A. (1995). The development of a monoclonal antibody-based elisa for the determination of histamine in food - application to fishery products and comparison with the hplc assay. Food Chemistry, 54(1), 85-91.

Schoolaert, E., Steyaert, I., Vancoillie, G., Geltmeyer, J., Lava, K., Hoogenboom, R., & De Clerck, K. (2016). Blend electrospinning of dye-functionalized chitosan and poly(epsilon-caprolactone): towards biocompatible pH-sensors. Journal of Materials Chemistry B, 4(26), 4507-4516.

Sen, D. P. (2005). Advances in fish processing technology. New Delhi: Allied.

128

Shakila, R. J., Vasundhara, T. S., & Kumudavally, K. V. (2001). A comparison of the TLC-densitometry and HPLC method for the determination of biogenic amines in fish and fishery products. Food Chemistry, 75(2), 255-259.

Shi, S. (2010). Antigen retrieval immunohistochemistry based research and diagnostics. Hoboken: Wiley.

Shimoda, M., Peralta, R. R., & Osajima, Y. (1996). Headspace gas analysis of fish sauce. Journal of Agricultural and Food Chemistry, 44(11), 3601-3605.

Simat, V., & Dalgaard, P. (2011). Use of small diameter column particles to enhance HPLC determination of histamine and other biogenic amines in seafood. LWT - Food Science and Technology,44(2), 399-406.

Simko, P. (2002). Determination of polycyclic aromatic hydrocarbons in smoked meat products and smoke flavouring food additives. Journal of Chromatography B- Analytical Technologies in the Biomedical and Life Sciences, 770(1-2), 3-18.

Sneddon, J., Masuram, S., & Richert, J. C. (2007). Gas chromatography-mass spectrometry-basic principles, instrumentation and selected applications for detection of organic compounds. Analytical Letters, 40(6), 1003-1012.

Snow, N. H., & Slack, G. C. (2002). Head-space analysis in modern gas chromatography. Trac-Trends in Analytical Chemistry, 21(9-10), 608-617.

Soncin, S., Chiesa, L. M., Panseri, S., Biondi, P., & Cantoni, C. (2009). Determination of volatile compounds of precooked prawn (Penaeus vannamei) and cultured gilthead sea bream (Sparus aurata) stored in ice as possible spoilage markers using solid phase microextraction and gas chromatography/mass spectrometry. Journal of the Science of Food and Agriculture, 89(3), 436-442.

Staruszkiewicz, W. F., and Bond, J. F. (1981). "Gas-Chromatographic Determination of Cadaverine, Putrescine, and Histamine in Foods." Journal of the Association of Official Analytical Chemists 64.3: 584-91. Print.

Steiner, M., Meier, R. J., Duerkop, A., & Wolfbeis, O. S. (2010). Chromogenic Sensing of Biogenic Amines Using a Chameleon Probe and the Red−Green−Blue Readout of Digital Camera Images. Analytical Chemistry,82(20), 8402-8405.

Steinkamp, F. L., DeGreeff, L. E., Collins, G. E., & Rose-Pehrsson, S. L. (2016). Factors affecting the intramolecular decomposition of hexamethylene triperoxide diamine and implications for detection. Journal of Chromatography A, 1451, 83-90.

129

Stockemer, J. (1982). Determination of Amino Acids and Biogenic Amines in the Light and Dark Muscle Meat of Tuna by Means of the Amino Acid Analyser and HPLC. Z Lebensm Unters Forsch, 174:108-113

Tahmouzi, S., Khaksar, R., & Ghasemlou, M. (2011). Development and validation of an HPLC-FLD method for rapid determination of histamine in skipjack tuna fish (Katsuwonus pelamis). Food Chemistry, 126(2), 756-761.

Takahashi, Y. K., Nagayama, S., & Mori, K. (2004). Detection and masking of spoiled food smells by odor maps in the olfactory bulb. Journal of Neuroscience, 24(40), 8690-8694.

Tanaka, R., Nakazawa, N., Maeda, T., Fukushima, H., Wada, R., Sugiura, Y., Matsushita, T., Hatate, H., Fukuda, Y. (2016). Effects of Chilled Storage, Freezing Rates, and Frozen Storage Temperature on Lipid Oxidation in Meat Blocks from Cultured Bluefin Tuna Thunnus thynnus. Journal of Aquatic Food Product Technology, 25(7), 1073-1085.

Tanaka, T., Yoneda, A., Inoue, S., Sugiura, Y., & Ueno, Y. (2000). Simultaneous determination of trichothecene mycotoxins and zearalenone in cereals by gas chromatography-mass spectrometry. Journal of Chromatography A, 882(1-2), 23- 28.

Tang, T., Shi, T. Y., Qian, K., Li, P. L., Li, J. Q., & Cao, Y. S. (2009). Determination of biogenic amines in beer with pre-column derivatization by high performance liquid chromatography. Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences, 877(5-6), 507-512.

Taverniers, I., De Loose, M., & Van Bockstaele, E. (2004). Trends in quality in the analytical laboratory. II. Analytical method validation and quality assurance. Trac- Trends in Analytical Chemistry, 23(8), 535-552.

The state of world fisheries and aquaculture 2016: Contributing to food security and nutrition for all. (2016). Rome: Food and Agriculture Organization of the United Nations.

Tseng, M., Jean, C., J., P., & Hung, Y. (2012). Interspecific and Intraspecific Genetic Diversity of Thunnus Species. Analysis of Genetic Variation in Animals.

Tuna fishery yearbook 2016. (2017). Western and Central Pacific Fisheries Commission

Tuzimski, T., & Sherma, J. (2015). High performance liquid chromatography in pesticide residue analysis. Boca Raton, FL: CRC Press.

130

Varlet, V., & Fernandez, X. (2010). Review. Sulfur-containing Volatile Compounds in Seafood: Occurrence, Odorant Properties and Mechanisms of Formation. Food Science and Technology International, 16(6), 463-503.

VecianaNogues, M. T., MarineFont, A., & VidalCarou, M. C. (1997). Biogenic amines as hygienic quality indicators of tuna. Relationships with microbial counts, ATP- related compounds, volatile amines, and organoleptic changes. Journal of Agricultural and Food Chemistry, 45(6), 2036-2041.

Vinas, J., & Tudela, S. (2009). A Validated Methodology for Genetic Identification of Tuna Species (Genus Thunnus). PLoS ONE,4(10).

Visciano, P., Campana, G., Annunziata, L., Vergara, A., & Ianieri, A. (2007). Effect of storage temperature on histamine formation in Sardina pilchardus and Engraulis encrasicolus after catch. Journal of Food Biochemistry, 31(5), 577-588.

Visciano, P., Schirone, M., Tofalo, R., & Suzzi, G. (2012). Biogenic amines in raw and processed seafood. Frontiers in Microbiology, 3, 10.

Wang, G. D., Lee, A. S., Lewis, M., Kamath, B., & Archer, R. K. (1999). Accelerated solvent extraction and gas chromatography mass spectrometry for determination of polycyclic aromatic hydrocarbons in smoked food samples. Journal of Agricultural and Food Chemistry, 47(3), 1062-1066.

Wierda, R. L., Fletcher, G., Xu, L., & Dufour, J. P. (2006). Analysis of volatile compounds as spoilage indicators in fresh king salmon (Oncorhynchus tshawytscha) during storage using SPME-GC-MS. Journal of Agricultural and Food Chemistry, 54(22), 8480-8490.

Wishart, D. S. (2008). Metabolomics: applications to food science and nutrition research. Trends in Food Science & Technology, 19(9), 482-493.

Wong, N. P., Damico, J. N., & Salwin, H. (1967). Investigation of volatile compounds in cod fish by gas chromatography and mass spectrometry. Journal of the Association of Official Analytical Chemists, 50(1), 8.

Wu, N. J., Lippert, J. A., & Lee, M. L. (2001). Practical aspects of ultrahigh pressure capillary liquid chromatography. Journal of Chromatography A, 911(1), 1-12.

Wunderlichova, L., Bunkova, L., Koutny, M., Jancova, P., & Bunka, F. (2014). Formation, Degradation, and Detoxification of Putrescine by Foodborne Bacteria: A Review. Comprehensive Reviews in Food Science and Food Safety, 13(5), 1012-1030.

Xiao, Z. J., & Lu, J. R. (2014). Generation of Acetoin and Its Derivatives in Foods. Journal of Agricultural and Food Chemistry, 62(28), 6487-6497.

131

Zarei, M., Najafzadeh, H., Enayati, A., Pashmforoush, M. (2011). Biogenic Amines Content of Canned Tuna Fish Marketed in Iran. American-Eurasian Journal of Toxicological Sciences, 3 (3): 190-193

Zhai, H. L., Yang, X. Q., Li, L. H., Xia, G. B., Cen, J. W., Huang, H., & Hao, S. X. (2012). Biogenic amines in commercial fish and fish products sold in southern China. Food Control, 25(1), 303-308.

Zhang, Z. M., Li, G. K., Luo, L., & Chen, G. N. (2010). Study on seafood volatile profile characteristics during storage and its potential use for freshness evaluation by headspace solid phase microextraction coupled with gas chromatography-mass spectrometry. Analytica Chimica Acta, 659(1-2), 151-158.

Zhuang, G. C., Lin, Y. S., Bowles, M. W., Heuer, V. B., Lever, M. A., Elvert, M., & Hinrichs, K. U. (2017). Distribution and isotopic composition of trimethylamine, dimethylsulfide and dimethylsulfoniopropionate in marine sediments. Marine Chemistry, 196, 35-46.

Zufall, F., & Munger, S. D. (2016). Chemosensory transduction: The detection of odors, tastes, and other chemostimuli. Amsterdam: Elsevier/AP, Academic Press is an imprint of Elsevier.

132

BIOGRAPHICAL SKETCH

Jing Bai is originally from China and has lived in several places including Kaifeng,

Zhengzhou, Chicago, and Gainesville, FL. She graduated with a Bachelor‘s degree of

Engineering in food quality and safety from Henan Agricultural University in 2012. Then she received her Master of Science in food safety and technology from Illinois Institute of Technology in 2014. In August 2015, she entered the Food Science program at the

University of Florida and started the pursuit of her doctoral degree. Jing Bai has a strong passion for food flavor and intends to be a research scientist in this research field.

133